1

Inhibitory synapses are repeatedly assembled

and removed at persistent sites in vivo

by

Katherine L. Villa

B.S., M.S. Molecular and Cellular Biology

The Johns Hopkins University, Baltimore, 2007, 2008

Submitted to the Department of Biology in

Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy in Biology at

The Massachusetts Institute of Technology

February 2016

© Massachusetts Institute of Technology

All rights reserved

Signature of Author: ____________________________________________________________

Department of Biology

November 6, 2015

Certified by: __________________________________________________________________

Elly Nedivi

Professor of Brain & Cognitive Sciences and Biology

Thesis Supervisor

Accepted by: _________________________________________________________________

Michael Hemann

Associate Professor of Biology

Co-Chair, Biology Graduate Committee

2

3

Inhibitory synapses are repeatedly assembled

and removed at persistent sites in vivo

by

Katherine L. Villa

Submitted to the Department of Biology

on November 6th, 2015 in Partial Fulfillment of the

Requirements for the Degree of

Doctor of Philosophy in Biology

Abstract

Structural plasticity, the rewiring of synaptic connections, occurs not only during development, but is prevalent in the adult brain and likely represents the physical correlate of learning and memory. Removal or addition of excitatory and inhibitory synaptic inputs onto a neuron can affect their relative influence on excitation in specific dendritic segments, and ultimately regulate neuronal firing. However, the structural dynamics of excitatory and inhibitory synapses in vivo, and their relation to each other, is not well understood. To gain insight into synaptic remodeling in the adult brain in vivo, we used dual- and triple- color two-photon imaging to track the dynamics of all inhibitory and excitatory synapses onto a given neuron in the cerebral cortex at different timescales. By studying synaptic changes over 4-day or 24-hour intervals we were able to determine that inhibitory synapses are remarkably dynamic in vivo. We found that Inhibitory synapses occur not only on the dendritic shaft, but also a significant fraction is present on dendritic spines, alongside an excitatory synapse. Inhibitory synapses on these dually innervated spines are remarkably dynamic and in stark contrast to the stability of excitatory synapses on the same spines. Many of the inhibitory synapses on dendritic spines repeatedly disappear and reappear in the same location. These reversible structural dynamics indicate a fundamentally new role for inhibitory synaptic remodeling – flexible, input-specific modulation of stable excitatory connections. To determine whether synapse dynamics are regulated by experience-dependent plasticity, we performed monocular deprivation, finding that an ocular dominance shift reduces inhibitory synaptic lifetime and increases recurrence. To investigate the molecular mechanism of rapid inhibitory synapse appearance and removal, I am currently testing molecular interventions that influence the clustering of gephyrin, a scaffolding molecule that anchors inhibitory receptors at postsynaptic sites.

Thesis Supervisor: Elly Nedivi

Title: Professor of Brain & Cognitive Sciences and Biology

4

Acknowledgements

I first and foremost would like to thank my advisor, Elly Nedivi, who is quite literally the

best PhD advisor I could have asked for. Elly has taught me so much about being a top-caliber

scientist, a great writer, and balancing a successful career with being a great mother. I would

like to thank her for her hands on mentoring, especially for helping me get through writing, which

has always been my least favorite part.

I would also like to thank my committee Troy Littleton, Martha Constantine-Patton, and Jeffrey

Lichtman, for their helpful scientific input, for sharing equipment, and also to Troy for organizing

the MCN program which has brought so many interesting talks to our building.

I would like to thank my undergraduate advisor, Victor Corces, one of the kindest, most

supportive people I have ever met, who first opened my eyes for loving science and exploring

interesting questions.

I will of course always be indebted to my great colleagues in the Nedivi lab, especially Jerry

Chen, who taught me everything I know about 2 photon imaging, and Sven Loebrich, who

helped me learn cloning, cell culture, and anything else I ever asked him about. I also want to

thank the rest of the imaging team, Ronen Eavri, Jaichandar Subramanian, Peter Wang, and

Aygul Balcioglu-Dutton, for their help, support, and troubleshooting along the way. Also, to

everyone else who overlapped with me in the Nedivi lab: Jenifer Leslie, Genevieve Flanders,

Zachary Tong, Marc Benoit, Mette Rathje, and Charles Moss, who created such a friendly and

entertaining lab environment.

Most importantly, I need to thank Kalen Berry, without whose programing knowledge, tinkering

skills, and helpful attitude, our paper never would have been completed. Thank you Kalen for

making up for the holes in my skillset and for being such a pleasure to work with every single

day.

To my collaborators Jae Won Cha and Peter So, thank you for building the microscope that I

used on a daily basis to perform my research. To Yoshiyuki Kubota, thank you for completing

the electron microscopy that was essential for validating our work and asking interesting follow-

up questions, and for many stimulating conversations. To Won Chan Oh and Hyung-Bae Kwon,

who I met recently, but who contributed such an interesting electrophysiology experiment to our

paper.

5

Thank you to the funding agencies who have supported my graduate education and provided

funding to complete this work: MIT’s pre-doctoral training grant through the NIH, my personal

NRSA fellowship, and to the National Eye Institute, which continues to support the Nedivi lab.

To all my friends at MIT, in Boston, and in New Jersey: Daisy, Izarys, Katie G (aka Katie #2),

Rini, Ann, Jim, Angie, Kristina, Catie Thompson (Catie with a C), Dave Hall!, Bill, Leah, and

everyone who made my time at MIT so much fun.

Of course, I owe infinite gratitude to my family: To my parents for always always supporting me

and all my crazy hard-headed ideas; to my late grandparents, Nanny and Poppy, for instilling a

love of learning when I was just a baby; to my current gray haired ladies, Ginga, Aunt Mickey,

and Aunt Isabell who kept the pressure on for me to finish grad school and get a real job, and

for always reminding me what is really important in life, to my beloved tios, Auntie, Uncle, Aunt

Pammy, and Uncle Billy, for their constant love and support, to my cousin Gina for being such

an inspiration to me, and to my younger sister Rachel who inspired me to be someone worthy of

looking up to (and for being so beautiful ;P).

Finally I want to thank my husband and soul mate, Juan Alvarez. Even though I never believed

in soul mates before meeting you, every day I am with you provides more evidence against my

prior hypothesis. You complement me in every way imaginable and I am a better person

because of you. I am so lucky to have found someone who not only makes me happier each

and every day, but who also understands the demands of a scientific career.

6

Table of Contents

Chapter 1 - Introduction .......................................................................................................... 8

Organization of synapses across the dendritic arbor .............................................................. 9

First views through Golgi staining and electron microscopy .............................................. 9

Synaptic visualization by new in vivo fluorescent labeling methods ..................................12

Circuit rewiring through experience dependent plasticity .......................................................14

Conclusion ............................................................................................................................15

Chapter 2 - Clustered dynamics of inhibitory synapses and dendritic spines in the adult

neocortex .................................................................................................................................17

Abstract .................................................................................................................................17

Introduction ...........................................................................................................................18

Results ..................................................................................................................................19

Simultaneous In Vivo Imaging of Inhibitory Synapses and Dendritic Spines .....................19

Teal-Gephyrin Puncta Correspond to Inhibitory Synapses ...............................................22

Differential Distribution of Inhibitory Spine and Shaft Synapses .......................................24

Inhibitory Spine and Shaft Synapses are Kinetically Distinct ............................................26

Inhibitory Synapse and Dendritic Spine Changes Are Locally Clustered ..........................28

Discussion .............................................................................................................................31

Supplementary Figures .........................................................................................................36

Experimental Procedures ......................................................................................................39

Chapter 3 - Inhibitory synapses are repeatedly assembled and removed at persistent

sites in vivo .............................................................................................................................47

Abstract .................................................................................................................................47

Introduction ...........................................................................................................................48

Results ..................................................................................................................................50

Triple-color labeling of pyramidal cells .............................................................................50

Teal-gephyrin and PSD-95-mCherry puncta represent inhibitory and excitatory synapses,

respectively ........................................................................................................................52

Teal-gephyrin puncta on dually innervated spines represent functional GABAergic

synapses............................................................................................................................54

The majority of excitatory synapses and the spines that bear them are stable .................56

Inhibitory synapses disappear and appear again in the same location .............................58

Kinetics of inhibitory synaptic dynamics are altered in experience dependent plasticity ...60

Inhibitory axons appear to remain in place after loss of a recurrent inhibitory synapse .....63

7

Discussion .............................................................................................................................65

Supplemental Data ................................................................................................................69

Experimental Procedures ......................................................................................................78

Chapter 4 - Probing the molecular mechanism of rapid inhibitory dynamics ....................88

Abstract .................................................................................................................................88

Introduction ...........................................................................................................................89

Results ..................................................................................................................................91

Generation of gephyrin S268A and S270A expression plasmids ......................................91

In vivo 2-photon imaging of L2/3 pyramidal cells expressing mutant gephyrin ..................91

Density of spines and synapses of gephyrin mutants .......................................................92

Dynamics of spines and synapses in gephyrin mutants ....................................................93

Discussion .............................................................................................................................95

Methods ................................................................................................................................96

Chapter 5 – Conclusion ........................................................................................................ 102

Inhibitory synapses are repeatedly assembled and removed from persistent sites ......... 105

Functional implications of inhibitory synaptic placement across the dendritic arbor ........ 107

Molecular mechanism of rapid gephyrin dynamics ......................................................... 110

Future Experiments ........................................................................................................ 110

References ............................................................................................................................ 113

8

Chapter 1 - Introduction

Parts of this chapter will be published in:

KL Villa, E Nedivi. “Excitatory and inhibitory synaptic placement and functional implications.”

Dendrites: development and disease. Ed. Kazuo Emoto, Rachel Wong, Eric Huang and Casper

Hoogenraad. 2015. Springer. In press.

Synaptic transmission between neurons is the basic unit of communication in neural

circuits. The neurotransmitter synthesized and released at the synapse is the basis for

classification of neurons as excitatory or inhibitory. Glutamatergic neurons, releasing glutamate,

excite postsynaptic cells, while GABAergic neurons, releasing GABA (γ-aminobutyric acid),

inhibit them. The relative number and distribution of excitatory and inhibitory synaptic inputs

across individual dendrites and neurons are the hardware of local dendritic and cellular

computations and therefore key to information processing in the brain. The delicate balance

between excitation and inhibition is essential for precise nervous system function and plasticity,

as its perturbation has been associated with disorders ranging from epilepsy (Mohler et al.,

2004), to autism (Rubenstein and Merzenich, 2003), and mental retardation (Dani et al., 2005;

Kleschevnikov et al., 2004), as well as neuropsychatric disorders such as schizophrenia (Lewis

et al., 2005). Here, we discuss the structural and functional observations that have guided the

understanding of excitatory and inhibitory synaptic organization across the neuronal arbor, and

the subcellular targeting properties of both excitatory and inhibitory subtypes. We focus primarily

on the adult mammalian cortex, where excitatory and inhibitory cell types, their connectivity, and

its functional implications have been best characterized. Deep understanding of the structural

distribution of synaptic inputs, and how those structures change over time, is important for the

accurate modeling of neurons and for an understanding of how summing inputs across the

dendritic arbor can engender the firing patterns of a particular cell.

9

Organization of synapses across the dendritic arbor

First views through Golgi staining and electron microscopy

Pyramidal neurons

In the late 19th century, development of the Golgi stain, a silver stain that sparsely but intensely

labels neurons, enabled the first clear discrimination of individual neuronal morphology (Figure

1A-C). In his meticulously detailed drawings Ramon y Cajal characterized the elaborate

dendritic arbors and axonal processes of neurons with a wide range of morphologies. One

obvious distinction was that some neurons have smooth dendrites while others have many

spiny protrusions. In general, most neurons with spiny dendrites were later revealed to be

glutamatergic and excitatory, while neurons with smooth dendrites for the most part release

GABA and are inhibitory (Gabbott and Somogyi, 1986; Kubota, 2014; Kubota et al., 2011b;

Thomson and Deuchars, 1997). Elegant electron microscopy (EM) studies on Golgi stained

cells observed that each spine contains a synapse characterized by round presynaptic vesicles

Figure 1. Visualization of L2/3 pyramidal neurons. A) Golgi stained pyramidal neuron (image kindly provided by Dr.

Terry Robinson, University of Michigan). B and C) High magnification views of the two dendrites marked by arrows

in panel A. Note dendritic spines. D) 3D volume projection of a pyramidal neuron imaged in vivo using 2 photon

microscopy with a pseudo-colored YFP fill. E) An individual dendrite expressing YFP (red) and teal-gepyrin (green)

to label inhibitory synapses. F) The same dendrite imaged 8 days later. Yellow arrows indicate dynamic inhibitory

synapses and red arrows indicate dynamic spines. Filled arrows show when structures are present and open arrows

indicate their absence.

10

and a robust postsynaptic density (Hersch and White, 1981; LeVay, 1973; Parnavelas et al.,

1977). These asymmetric synapses, classified as type 1 synapses, are innervated by axons of

glutamatergic neurons (Baude et al., 1993). While there are a few reports of specific cell

regions on particular types of neurons where type 1 synapses are localized directly on the

dendritic shaft (Megias et al., 2001; Parnavelas et al., 1977), this is generally not the case.

Given that the vast majority of spines, with the exception of some very thin spines (about 2-4%

of total cortical spines) (Arellano et al., 2007; Hersch and White, 1981; White and Rock,

1980), have a single type 1 excitatory synapse (Harris et al., 1992; LeVay, 1973) it was

reasonable to assume that each spine can serve as a readily identifiable structural surrogate for

an excitatory synapse. Thus, for spiny pyramidal neurons, dendritic spine distributions as seen

by Golgi stain or by filling cells with a fluorescent dye could inform as to the placement of

excitatory synapses across the dendritic arbor (Elston and Rosa, 1997; Larkman, 1991;

Trommald et al., 1995).

There is a great variety in the density of spines between cells, and between different

branches on the same cell. To give a general sense of spine distributions, spine density is low

within 40 µm of the cell body, reaching a maximum density in a region 40-130 µm from the

soma, and then gradually decreasing toward a dendrite’s distal tips (Megias et al., 2001). On the

distal branches of pyramidal cells, spine density can range from 0 to 70 spines per 10 µm

(Megias et al., 2001) with the highest spine density often found on the thickest dendrites, usually

the primary apical dendrite. The total number of spines on spiny excitatory neurons typically

ranges from 5,000 – 35,000 spines per cell, ultimately depending on the total dendritic length,

with L5 pyramidal cells generally having longer total dendrite length and therefore more

synapses than L2/3 pyramidal cells, for example (Larkman, 1991).

Inhibitory synapses innervated by the axons of GABAergic neurons, classified by EM as

type 2, or symmetric synapses, are typified by a symmetric synaptic cleft, due to a minimal

11

postsynaptic density, as well as flattened presynaptic vesicles (Davis and Sterling, 1979).

Unfortunately, since type 2 inhibitory synapses lack a morphological surrogate, their distribution

patterns cannot be discriminated by Golgi staining alone. Laborious EM reconstructions of

relatively small neuronal regions have shown excitatory to inhibitory synaptic ratios along

pyramidal dendrites ranging from 6.5 - 12.5 to 1 (Davis and Sterling, 1979). As previously

mentioned, the cell body and proximal dendrites of excitatory neurons typically lack spines.

However, they are often densely innervated by type 2 synapses. In the aspiny region of the

proximal dendrites, inhibitory synapse density can be as high as 17 synapses per 10 µm

(Megias et al., 2001). EM studies also show that on excitatory dendrites, the majority of type 2

inhibitory synapses are located on the dendritic shaft at a density of about 3 per 10 µm (Hersch

and White, 1981). Inhibitory synapses can also be located on dendritic spines adjacent to

excitatory synapses (Parnavelas et al., 1977), as well as on the axon initial segment (Hersch

and White, 1981; Westrum, 1966). It was later found that inhibitory axons from different

inhibitory neuron subtypes specifically target discrete regions of their postsynaptic pyramidal cell

partners (this is discussed in greater detail further in the chapter).

Non-pyramidal neurons

The dendrites of inhibitory neurons, in general, do not contain spines. However, a small

subset of inhibitory neurons have dendritic spines (Azouz et al., 1997; Feldman and Peters,

1978), with densities that range from 0.3 to 7 spines per 10 µm (Kawaguchi et al., 2006; Keck et

al., 2011). Like spiny pyramidal cells, the cell body and most proximal dendrites of spiny

interneurons within 30 µm of the soma lack spines (Kawaguchi et al., 2006).

Immunohistochemistry experiments show that the majority of spines on these interneurons

colocalize with Vglut1 but not VGAT, indicating that they mostly harbor excitatory synapses

(Keck et al., 2011). The proportion of excitatory synapses located along the shaft of spiny

12

inhibitory dendrites has not been established. We also know little about the distribution of

inhibitory synapses on spiny interneurons (Gulyas et al., 1992).

Aspiny interneurons receive both excitatory and inhibitory inputs onto their soma and

proximal dendrites (Davis and Sterling, 1979), with a higher density along their distal dendrites

(Parnavelas et al., 1977). Similarly to inhibitory synapses on pyramidal cells, both excitatory

synapses and inhibitory synapses onto dendrites of aspiny inhibitory neurons cannot be

visualized through a morphological surrogate and could initially be examined only by EM on

relatively small arbor segments or by immunohistochemistry. EM reconstructions of isolated

branches from different inhibitory neuron types provides anecdotal evidence related to their

synaptic densities (Gulyas et al., 1999; Gulyas et al., 1992).

Synaptic visualization by new in vivo fluorescent labeling methods

Golgi staining gave us a first view of dendritic spine distributions, and EM studies were

first to shed light on the fundamental layout of excitatory and inhibitory synaptic distributions on

different neuronal types. Yet, both methods have inherent limitations. As a cell fill, Golgi stain

can at best identify spines on cells with defined morphology and location, while all inhibitory

synapses as well as excitatory synapses on aspiny dendrites remain invisible. EM is limited by

the difficulty of reconstructing large dendritic segments. EM reconstruction of an entire cell

would require a heroic effort, one rarely attempted (see (White and Rock, 1980) for a

reconstruction of an entire spiny stellate neuron). Both Golgi staining and EM reconstructions

necessitate tissue fixation and cannot be used for visualizing structural dynamics. Thus, our

initial view of synaptic distributions was constrained by the limitations of the Golgi and EM

methods.

Recently, imaging of fluorescently-labeled cells in vivo has provided a new view not only

of dendritic spine distributions, but also of excitatory and inhibitory synapses as well as the

13

dynamics of these structures (Figure 1D-F). This is accomplished using two-photon imaging,

which has several advantages over conventional light microscopy. Because two-photon

excitation occurs when two photons simultaneously excite a fluorescent molecule, the

wavelength of excitation is twice that of conventional microscopy, and these longer wavelengths

are more effective at penetrating deep into tissue with less scattering. Thus the use of two-

photon microscopy allows imaging of structures up to 400 µm deep within the cortex.

Additionally, two-photon microscopy uses focal point scanning of a laser, which allows high

resolution imaging and reduces background noise, as only a single point is excited at a time.

This also reduces photobleaching.

The first studies using two-photon imaging on neurons fluorescently labeled with green

fluorescent protein (GFP) showed that in adult animals dendrites of pyramidal neurons are very

stable (Grutzendler et al., 2002; Mizrahi and Katz, 2003; Trachtenberg et al., 2002), but

dendritic spines are highly dynamic (Trachtenberg et al., 2002), implying a capacity for synaptic

removal and addition. Spine dynamics of excitatory as well as spiny inhibitory neurons can be

further increased upon sensory deprivation (Hofer et al., 2009; Holtmaat et al., 2009; Holtmaat

et al., 2006; Trachtenberg et al., 2002). Dynamics differ dependent on deprivation protocol and

cortical lamina, consistent with the view that they reflect specific circuit alterations. Although it

has been shown that spines that persist for more than 4 days always acquire an excitatory

synapse, specific labeling of the excitatory synapse is necessary to observe synaptic dynamics

at shorter timescales. Though spines serve as a rough structural surrogate for excitatory

synapses, there is no structural marker of inhibitory synapses, rendering them invisible.

Synaptic labeling of inhibitory synapse is necessary in order to classify their distribution and

dynamics in vivo.

In contrast to the stability of excitatory dendritic branches, inhibitory dendrites are

capable of growth and retraction in vivo (Brown et al., 2011; Holtmaat et al., 2009; Lee et al.,

14

2006) and their dynamics are influenced by sensory manipulations (Brown et al., 2011). The

boutons of inhibitory neurons are also capable of remodeling. Thus, synaptic distributions are

not necessarily rigid. Rather, both excitatory and inhibitory synapses are dynamic structures and

their remodeling potentially underlies functional plasticity.

Circuit rewiring through experience dependent plasticity

While the overall structure of synaptic inputs is important for understanding the

integration of signals at any particular point in time, a key feature of the brain is its ability to

learn and adapt in response to experience. Learning directly results in structural changes in the

connections between neurons, and the brain is capable of remodeling these connections

throughout life. One well-established system used to study cortical plasticity is by using

Monocular Deprivation (MD), the closure of one eye, to induce an ocular dominance shift in

binocular visual cortex. When the eye is closed for a period of time, there is a reduction in the

cortical response to the closed eye, and an increase in the response to the open eye. In some

mammals a MD only affects the cortex early in development during a critical period, but in mice

an MD causes an ocular dominance shift even in adults (Frenkel et al., 2006; Sato and Stryker,

2008).

Inhibition plays a fundamental role in ocular dominance both during development and in

adult mice. During development, the onset and closure of the critical period for ocular

dominance plasticity is triggered by the maturation of GABAergic interneurons (Fagiolini and

Hensch, 2000; Huang et al., 1999). Loss of visual input through MD results in a decreased

expression of GABA and GAD, the GABA synthesizing enzyme, from the deprived eye column

in visual cortex (Hendry and Jones, 1988). Thus, loss of visual experience, such as through MD

or dark rearing, results in a loss of inhibition, which has been shown to restore a juvenile form of

15

plasticity in the adult cortex (He et al., 2006). The structural consequence of the reduction of

inhibition seen after MD or sensory deprivation, is that inhibitory neurons themselves remodel

both their axons and dendrites (Brown et al., 2011; Chen et al., 2011; Keck et al., 2011; Pieraut

et al., 2014; Schuemann et al., 2013; Wierenga et al., 2008). This remodeling is also seen in

inhibitory neurons after treatment with the selective serotonin reuptake inhibitor, fluoxetine,

which has also been shown to restore a juvenile state of plasticity in adults (Chen et al., 2011;

Maya Vetencourt et al., 2008). This suggests that the reduction of inhibition is sufficient to

create a permissive environment allowing for increased plasticity of excitatory neurons.

However, it was not known if these effects occur within a single cell or only in the context of a

larger circuit.

Conclusion

The precise way in which excitatory and inhibitory inputs are integrated within a single

neuron, and how this integration supports computation in functioning networks, are still critical

questions in neuroscience today. Despite advances in our knowledge, we still have limited

experimental evidence of the fine-scale synaptic architecture across the neuronal arbor of

individual cells of different types, and we know even less of the functional interactions between

synaptic excitation and inhibition (Chadderton et al., 2014).

When we began this work, there was no way to study the distribution or dynamics of

inhibitory synapses in vivo. By tagging the postsynaptic scaffolding protein, gephyrin, with a

florescent marker we were able to observe the changes in inhibitory synapses in a living mouse

brain. We also used a fluorescently labeled PSD-95 to simultaneously track all inhibitory and

excitatory synapses on a single cell. We focused on how these synapses change under

baseline conditions and after Monocoular Deprivation (MD) as a model of experience-

16

dependant plasticity. My research has provided important insights on how inhibitory synapses

rapidly remodel in vivo, with implications for the regulation of cortical plasticity at the level of

individual cells and circuits.

17

Chapter 2 - Clustered dynamics of inhibitory synapses

and dendritic spines in the adult neocortex

Portions of this chapter have been published in:

JL Chen, KL Villa, JW Cha, P So, Y Kubota, E Nedivi 2012. “Clustered Dynamics of Inhibitory

Synapses and Dendritic Spines in the Adult Neocortex.” Neuron 74, 361-373.

Contributions: Figures and published manuscript prepared by Jerry Chen. I assisted with sample preparation for Fig. 2 and data collection and analysis for Figs. 3-5 (numbered 2.2-2.5 in the thesis), and S1-S3 in the paper, numbered 2.6-2.8 in this thesis. Immuno EM and reconstruction performed by Yoshiyuki Kubota.

Abstract

An important feature of brain function is the capacity to dynamically adapt in response to

the environment, which is achieved by the remodelling of synaptic connections. In the case of

excitatory synapses on excitatory neurons, dendritic spine dynamics have served as a structural

surrogate for monitoring synaptic rearrangements in vivo. For inhibitory synapses, the lack of

such a surrogate necessitates their direct visualization. We performed dual color two-photon

imaging to simultaneously monitor inhibitory synapse and dendritic spine rearrangement onto

pyramidal neurons in the adult mammalian cortex. This is the first instance of inhibitory synapse

imaging in vivo. We found that 1/3 of inhibitory synapses are located not on the dendritic shaft,

as commonly thought, but on dendritic spines, and that spine synapses are more prevalent at

apical dendrites. Further, chronic monitoring of these two synapse classes showed that they

have different capacities for turnover during normal and altered sensory experience. Monocular

deprivation results in a loss of inhibitory synapses on the shaft and on spines. Inhibitory

synapse dynamics are temporally and spatially clustered with turnover of dendritic spines. Our

findings provide in vivo evidence for local coordinated rearrangements between diverse sets of

synaptic inputs.

18

Introduction

The ability of the adult brain to change in response to experience arises from coordinated

modifications of a highly diverse set of synaptic connections. These modifications include the

strengthening or weakening of existing connections, as well as new synapse formation or

elimination. The persistent nature of structural synaptic changes make them particularly

attractive as cellular substrates for long-term changes in connectivity, such as might be required

for learning and memory. Sensory experience can produce parallel changes in excitatory and

inhibitory synapse density in the cortex (Knott et al., 2002), and the interplay between excitatory

and inhibitory synaptic transmission serves an important role in adult brain plasticity (Spolidoro

et al., 2009). Excitatory and inhibitory inputs both participate in the processing and integration of

local dendritic activity (Sjostrom et al., 2008), suggesting that they are coordinated at the

dendritic level. However, the manner in which these changes are orchestrated and the extent to

which they are spatially clustered are unknown.

Evidence for the gain and loss of synapses in the adult mammalian cortex has

predominantly used dendritic spines as a proxy for excitatory synapses on excitatory pyramidal

neurons. The vast majority of excitatory inputs to pyramidal neurons synapse onto dendritic

spine protrusions that stud the dendrites of these principal cortical cells (Peters, 2002) and to a

large approximation are thought to provide a one-to-one indicator of excitatory synaptic

presence, when the spine persists for more than 4 days (Holtmaat and Svoboda, 2009).

Inhibitory synapses onto excitatory neurons target a variety of subcellular domains, including the

cell body, axon initial segment, and dendritic shaft, as well as some dendritic spines. Unlike

monitoring of excitatory synapse elimination and formation on neocortical pyramidal neurons,

there is no morphological surrogate for the visualization of inhibitory synapses. Inhibitory

synapse dynamics has been inferred from in vitro and in vivo monitoring of inhibitory axonal

bouton remodeling (Chen et al., 2011; Keck et al., 2011; Wierenga et al., 2008). However,

imaging of presynaptic structures does not provide information regarding the identity of the

postsynaptic cell or their subcellular sites of contact. In addition, monitoring of either dendritic

19

spine or inhibitory bouton dynamics has thus far utilized a limited field of view and has not

provided a comprehensive picture of how these dynamics are distributed and potentially

coordinated across the entire arbor.

Here, we simultaneously monitored inhibitory synapse and dendritic spine remodeling

across the entire dendritic arbor of cortical L2/3 pyramidal neurons in vivo during normal and

altered sensory experience. We found that inhibitory synapses on dendritic shafts and spines

differ in their distribution across the arbor, consistent with different roles in dendritic integration.

These two inhibitory synapse populations also display distinct temporal responses to visual

deprivation, suggesting different involvements in early versus sustained phases of experience-

dependent plasticity. Finally, we find that the rearrangements of inhibitory synapses and

dendritic spines are locally clustered, mainly within 10 µm of each other, the spatial range of

local intracellular signaling mechanisms, and that this clustering is influenced by experience.

Results

Simultaneous In Vivo Imaging of Inhibitory Synapses and Dendritic Spines

In order to label inhibitory synapses for in vivo imaging, we generated a Cre-dependent

plasmid expressing Gephyrin, a post-synaptic scaffolding protein exclusively found at

GABAergic and glycinegic synapses (Craig et al., 1996), fused to Teal flourescent protein (Teal-

Gephyrin) (Fig. 2.1A). This construct was co-electroporated with two additional plasmids: a Cre-

dependent enhanced yellow fluorescent protein (eYFP) plasmid as a dendritic and spine fill and

limiting amounts of a Cre construct. Since the Teal-Gephrin and the eYFP expression cassettes

were inserted in the antisense orientation each flanked by two sets of Cre recombination sites,

only cells expressing both constructs as well as Cre could show labelling, ensuring all labelled

cells were dually labelled for Teal-Gephrin and eYFP. Limiting Cre amounts allowed sparse

labeling for single cell imaging and reconstruction. Electroporations were performed in utero on

E16 embryos of pregnant C57Bl6 mice, targeting the lateral ventricle to label cortical progenitors

20

at the time L2/3 pyramidal neurons are born (Fig. 2.1B). Following birth mice were reared to 6-8

weeks of age, then implanted with bilateral cranial windows over the visual cortices.

Figure 2.1. Chronic in vivo two-photon imaging of inhibitory synapses and dendritic spines in L2/3 pyramidal

neurons. (A) Plasmid constructs for dual labeling of inhibitory synapses (Teal-GPHN) and dendritic spines (eYFP)

in cortical L2/3 pyramidal neurons. Cre/loxP system was used to achieve sparse expression density. (B)

Experimental time course. (C) CCD camera image of blood vessel map with maximum z-projection (MZP) of

chronically imaged neuron (white arrow) superimposed over intrinsic signal map of monocular (yellow) and

binocular (red) primary visual cortex. (D) Low-magnification MZP of acquired two-photon imaging volume. (E and

G) High-magnification view of dendritic segments (boxes in D) with labeled dendritic spines (red) and Teal-GPHN

puncta (green). (F and H) Examples of dendritic spine and inhibitory synapse turnover (boxes in E and G,

respectively). Dual color images (top) along with single-color Teal-GPHN (middle) and eYFP (bottom) images are

shown. Dendritic spines (squares), inhibitory shaft synapses (arrows), and inhibitory spine synapses (triangles) are

indicated with stable (white) and dynamic (yellow) synapses or spines identified. An added inhibitory shaft synapse

is shown in (F). An added inhibitory spine synapse and eliminated dendritic spine is shown in (H). Scale bars: (C),

200 µm; (D), 20 µm; (E and G), 5 µm; and (F and H), 2 µm.

21

Allowing 2-3 weeks for recovery (Lee et al., 2008), labeled neurons were identified and 3D-

volume images were acquired using a custom built two-channel two-photon microscope.

Imaging of eYFP-labeled neuronal morphology and Teal-labeled Gephyrin puncta was

Figure 2.2. Teal-gephyrin puncta correspond to inhibitory synapses. (A) In vivo image of an eYFP- (red) and Teal-

Gephyrin-labeled (green) dendrite. Letters indicate identified dendritic spines; numbers indicate identified inhibitory

synapses. (B) Reidentification of the same imaged dendrite in fixed tissue after immunostaining for eYFP. (F)

Serial-section electron microscopy (SSEM) reconstruction of the in vivo imaged dendrite (in green) with identified

GABAergic synapses (in red), non-GABAergic synapses (in blue), and unidentified spine-synapses (arrows). (D–

F) High-magnification view of region outlined in (C) with merged (top-left panels), eYFP only (top-middle panels),

Teal-Gephyrin only (top-right panels) in vivo images and SSEM reconstruction (bottom panel). (G) Electron

micrograph of inhibitory shaft synapse “1” in (D) identified by a GABAergic presynaptic terminal visualized by

postembedding GABA immunohistochemistry with 15 nm colloidal gold particles (black circles identify GABAergic

presynaptic terminal) contacting eYFP-labeled dendritic shaft (DAB staining, red arrows mark synaptic cleft). (H)

Electron micrograph of doubly innervated dendritic spine ‘d2’ in (E) with inhibitory synapse (red arrows mark

synaptic cleft) and excitatory synapse (blue arrows mark synaptic cleft). Scale bars: (A–C), 1 µm; (D–F, top panels),

1 µm; (D–F, bottom panels), 500 nm; and (G and H), 100 nm.

22

performed by simultaneous excitation of eYFP and Teal and resolution by wavelength

separation into two detection channels followed by post-hoc spectral linear un-mixing. In

addition, cortical maps of monocular and binocular primary visual cortex were obtained by

optical imaging of intrinsic signals and blood vessel maps were used to identify the location of

imaged cells with respect to these regions (Fig. 2.1C). Cells in binocular visual cortex were

imaged at 4 day intervals, initially for 8 days of normal visual experience followed by 8 days of

monocular deprivation by eyelid suture (MD) with an intermediate imaging session at 2 days

MD. In vivo imaging of electroporated neurons shows distinct labeling of neuronal morphology

by eYFP with clear resolution of dendritic spines. Teal-Gephyrin was visualized as clear

punctate labeling along the dendritic shaft and on a small fraction of dendritic spines (Fig. 2.1D,

E). The majority of Teal-Gephyrin puncta were stable and could reliably be re-identified over

multiple days and imaging sessions. Examples of puncta turnover was also observed over

multiple imaging sessions wherein individual puncta appeared and persisted, or disappeared

and remained eliminated.

Teal-Gephyrin Puncta Correspond to Inhibitory Synapses

To demonstrate that Teal-Gephyrin puncta visualized in vivo correspond to inhibitory

synapses, we performed serial section immunoelectron microscopy (SSEM) on an in vivo

imaged L2/3 pyramidal neuron dendrite labeled with eYFP/Teal-Gephyrin (Fig. 2.2A).

Immediately after two-photon imaging, the brain was fixed, sectioned, and stained with an

antibody to eYFP followed by a biotin-conjugated secondary and detected with nickel-

diaminobenzidine (DAB; Fig. 2.2B). A 30 µm dendritic segment with strong DAB staining was

relocated and then further cut into serial ultrathin sections and processed for postembedding

GABA immunohistochemistry to discriminate between inhibitory and excitatory presynaptic

terminals.

23

The robustness of the nickel-DAB staining was such that it frequently obscured most of

the postsynaptic dendritic compartment, including the postsynaptic density, of many synaptic

contacts. Visualization of the postsynaptic density is considered an important criterion for

identifying synapses, and in some cases, a postsynaptic membrane specialization could be

discerned despite the DAB staining, but other important criteria include aggregation of synaptic

small vesicles at the presynaptic junction, and a clear synaptic cleft structure between the pre-

and postsynaptic junction. Contacts were categorized as synapses only if all or at least two of

these criteria were present in at least a few serial ultrathin sections. The densities of GABA

marker colloidal gold particles were clearly different between GABA-positive and GABA-

negative presynaptic terminals, and a terminal was categorized as GABA-positive if a high

particle density was found in the presynaptic terminal across multiple serial ultrathin sections.

The reconstructed segment contained 26 dendritic spines observed in vivo, which were

reidentified after SSEM-reconstruction and all found to bear synaptic contacts (Fig. 2.2C). Six

additional spines, each with a single excitatory synapse, were identified by SSEM but not

visualized in vivo, likely because of their orientation perpendicular to the imaging plane. Eleven

filopodia-like structures without synaptic contacts were also found. These possessed very thin

necks (50–250 nm in width) and were also unresolved by two-photon microscopy. This suggests

that whereas dendritic spines imaged in vivo indeed closely represent excitatory synaptic

contacts, in vivo imaging potentially underestimates their true number by as much as 20%.

All ten of the Teal-Gephyrin puncta visualized in vivo corresponded with GABAergic

synapses found by SSEM. Six were localized on the dendritic shaft while four were located on

dendritic spines (Figures 2.2C–2.2G). Three out of the four dendritic spines bearing inhibitory

synapses were found to be co-innervated with an excitatory synapse (Figures 2H). Although a

coinnervated excitatory synapse was not found on the remaining spine, this is likely due to

known limitations of the SSEM reconstruction (Kubota et al., 2009). The proportion of doubly

innervated dendritic spines observed on this segment is comparable to previously reported

24

results (Kubota et al., 2007). Further SSEM reconstruction of the surrounding neuropil revealed

additional GABAergic processes touching the imaged dendrite without forming synaptic contact.

No Teal-Gephyrin puncta were observed in vivo at these points of contact. These results

confirm that imaged Teal-Gephyrin puncta correspond one-to-one with GABAergic inhibitory

synapses.

Differential Distribution of Inhibitory Spine and Shaft Synapses

To date, data regarding inhibitory synapse distribution on L2/3 pyramidal cell dendrites

and its relation to dendritic spine distribution has only been estimated from volumetric density

measurements (DeFelipe et al., 2002). We first used the Teal-Gephrin labeling to characterize

the distribution of inhibitory spine and inhibitory shaft synapses as well as dendritic spines along

the dendritic arbor of imaged L2/3 pyramidal neurons. The density of dendritic spines was 4.42

± 0.27 per 10 µm length of dendrite (Fig. 2.3A). From these spines, a small fraction (13.60 ±

1.38%) bore inhibitory synapses with a density of 0.71 ± 0.11 per 10 µm. Inhibitory synapses

along the dendritic shaft were more abundant than inhibitory spine synapses with a density of

1.68 ± 0.08 per 10 µm. While there was no difference in dendritic spine and inhibitory shaft

synapse density in apical versus basal dendrites, apical dendrites contained a higher density of

inhibitory spine synapses compared to basal dendrites (Mann-Whitney U test, P < 0.05) (Fig.

2.3B). When spine and synapse distribution were measured along the dendrite as a function of

distance from the cell soma, their density along both apical and basal dendrites was found to be

constant regardless of proximal or distal location (Fig. 2.3C). In contrast, the density of

inhibitory spine synapses on apical dendrites increased with increasing distance from the cell

soma suggesting that a larger fraction of dendritic spines bear inhibitory synapses at distal

apical dendrites. Inhibitory spine synapse density was two-fold higher at locations along the

apical dendrites greater than 125µm from the cell soma as compared to proximal locations

along the same dendritic tree (Mann-Whitney U test, P < 0.05) (Fig. 2.3D), resulting in a two-fold

increase in the ratio of inhibitory spine synapses to dendritic spines (Mann-Whitney U test, P <

25

0.05) (Fig. 2.3E). Identical analysis performed ex vivo in 50 µm coronal sections yielded the

same results, validating the reliability of our in vivo imaging-based quantifications and showing

that imaging depth does not diminish the fidelity of synapse scoring in the depth range that we

are imaging (Fig. 2.6). These findings demonstrate that while the distribution of inhibitory shaft

synapses is constant throughout the dendritic field, inhibitory spine synapse distribution is non-

uniform with highest density at distal apical dendrites.

Figure 2.3. Dendritic distribution of inhibitory shaft and spine synapses. (A) Dendritic density of dendritic spines,

inhibitory shaft synapses, and inhibitory spine synapses per cell. (B) Density per dendrite in apical versus basal

dendrites of dendritic spines (left), inhibitory shaft synapses (middle), and inhibitory spine synapses (right). (C)

Dendritic density as a function of distance from the cell soma in apical (black) and basal (red) dendrites for dendritic

spines (top panel), inhibitory shaft synapses (middle panel), and inhibitory spine synapses (bottom panel). (D)

Density of inhibitory spine (left) and inhibitory shaft (right) synapses in proximal (0–125 µm from soma) versus

distal (125–200 µm from soma) apical dendrites. (E) Ratio of inhibitory spine (left) and inhibitory shaft (right)

synapses to dendritic spines in proximal (0–125 µm from soma) versus distal (125–200 µm from soma) apical

dendrites. n = 14 cells from 6 animals for (A, C–E); n = 43 apical dendrites, 40 basal dendrites for (B); p < 0.05.

Error bars, SEM.

26

Inhibitory Spine and Shaft Synapses are Kinetically Distinct

Given their differential anatomical distribution, we next asked if inhibitory spine and shaft

synapses also differ in their capacities for synaptic rearrangement during normal experience or

MD (Fig. 2.1B, 2.4A). The majority of inhibitory synapse rearrangements observed were

persistent (persisting for at least two imaging sections), with only a small fraction of events

transiently lasting for only one imaging session, 4.20% ± 2.56% of all events in the case of

inhibitory shaft synapses and 9.00% ± 3.97% for inhibitory spine synapses (Fig. 2.7A and

2.7B). During normal experience, inhibitory shaft synapses and dendritic spines showed similar

fractional turnover rates with 5.36±0.97% shaft synapses and 5.26±0.89% dendritic spines

remodeling over an 8 day period (Fig. 2.4B). Inhibitory spine synapses, whether stable or

dynamic, were exclusively located on stable, persistent spines. These synapses were

fractionally, highly dynamic compared to both dendritic spines and shaft synapses with

18.84±5.50% of inhibitory spine synapses remodeling over an 8 day period of normal vision

(dendritic spines versus inhibitory spine synapses, Wilcoxon rank sum test, p < 0.05; inhibitory

shaft synapses versus inhibitory spine synapses, Wilcoxon rank sum test, p < 0.05).

In the adult mouse, a prolonged MD produces an ocular dominance (OD) shift in the

binocular visual cortex, characterized by a slight weakening of deprived-eye inputs and a

strengthening of non-deprived eye inputs (Frenkel et al., 2006). We observed no increase in

spine gain or loss on L2/3 pyramidal neurons during MD (Fig. 2.4C). For inhibitory shaft

synapses, MD increased the fractional rate of synapse loss to 1.05±0.42% per day during the

first two days of MD (Wilcoxon rank sum test, P < 0.05). This increase in synapse loss

persisted throughout the entire 8 days of MD (control vs. 2-4d MD, P < 0.01; control vs. 4-8d

MD, p < 0.001, Wilcoxon rank sum test). A much larger increase in the fractional rate of

synapse loss was observed for inhibitory spine synapses to 6.52±1.98% per day from 0-2d MD

(control versus 0-2d MD, Wilcoxon rank sum test, p < 0.01). However, this fractional increase

27

occurred only during the first two days of MD. Given the higher density of inhibitory shaft

synapses as compared to inhibitory spine synapse, despite the higher rate of inhibitory spine

synapse elimination, in response to MD, the overall net change in the two synapse populations

per length of dendrite were similar (Fig. 2.4D). These findings demonstrate that changes in

experience can drive inhibitory shaft and synapse remodeling in distinct ways.

Figure 2.4. Inhibitory spine and

shaft synapses form two kinetic

classes. (A) Example of dendritic

spine and inhibitory

synapse dynamics of L2/3

pyramidal neurons in binocular

visual cortex during monocular

deprivation. Dual color images

(left) along with single-color Teal-

GPHN (middle) and eYFP (right)

images are shown. Dendritic

spines (squares), inhibitory shaft

synapses (arrows), and inhibitory

spine synapses (triangles) are

indicated with stable (white) and

dynamic (yellow) synapses or

spines identified.(B) Fraction of

dynamic dendritic spines,

inhibitory shaft synapses, and

inhibitory spine synapses during

control conditions of normal

vision. (C) Fraction of additions or

eliminations of dendritic spines

(top), inhibitory shaft synapses

(middle), and inhibitory spine

synapses (bottom) at 4 day

intervals before and during

monocular deprivation. (D)

Fraction of eliminations of

inhibitory spine inhibitory and

inhibitory shaft synapses at 0–

2 days MD and 2–4 days MD. n =

14 cells from 6 animals; ∗, p <

0.05; ∗∗, p < 0.01; and ∗∗∗, p <

0.005. Error bars, SEM.

28

Inhibitory Synapse and Dendritic Spine Changes Are Locally Clustered

Long-term plasticity induced at one dendritic spine can coordinately alter the threshold

for plasticity in nearby neighboring spines (Govindarajan et al., 2011; Harvey and Svoboda,

2007). Electrophysiological studies suggest that plasticity of inhibitory and excitatory synapses

may also be coordinated at the dendritic level. Calcium influx and activation of calcium-

dependent signaling molecules that lead to long-term plasticity at excitatory synapses can also

induce plasticity at neighboring inhibitory synapses (Lu et al., 2000; Marsden et al., 2010).

Conversely, inhibitory synapses can influence excitatory synapse plasticity by suppressing

calcium-dependent activity along the dendrite (Miles et al., 1996). Given the limited spatial

extent of these signaling mechanisms (Harvey and Svoboda, 2007), we looked for evidence of

local clustering between excitatory and inhibitory synaptic changes.

We first looked at the distribution of dynamic events resulting in persistent changes (both

additions and eliminations) on each dendritic segment (68.1 ± 2.9 µm in length) as defined by

the region from one branch point to the next or from branch tip to the nearest branch point.

During normal visual experience, 58.2% ± 7.6% of dendritic segments per cell contained both a

dynamic inhibitory (spine or shaft) synapse and a dynamic dendritic spine (Fig. 2.5A). On these

dendritic segments, a large fraction of dynamic inhibitory synapses and dendritic spines were

found to be located within 10 µm of each other, suggesting that these changes were clustered

(dynamic spines to nearby dynamic inhibitory synapses, repeated-measures ANOVA, p < 1 ×

10−10; dynamic inhibitory synapses to nearby dynamic spines, repeated-measures ANOVA, p <

0.0005; Fig. 2.5B and 2.8A). To determine whether clustering of dynamic inhibitory synapses

with dynamic spines were merely a reflection of the dendritic distribution of inhibitory synapses

and spines, we performed nearest neighbor analysis between every monitored dynamic and

stable inhibitory synapse and every dynamic and stable spine (Fig. 2.5C). We found that

inhibitory synapse changes occur in closer proximity to dynamic dendritic spines as compared

29

to stable spines (K-S test, p < 2.0 × 10−6; Fig. 2.5D). Conversely, dendritic spine changes occur

in closer proximity to dynamic inhibitory synapses as compared to stable inhibitory synapses (K-

S test, p < 2.0 × 10−4; Fig. 2.5E). Interestingly, dendritic spine changes were not clustered with

each other and indeed occurred with less proximity to neighboring dynamic spines as compared

to stable spines (stable spines versus dynamic spines, K-S test, p < 0.05; Fig. 2.5F). We

observed no difference in nearest neighbor distribution between dynamic inhibitory synapses

Figure 2.5. Inhibitory synapse and dendritic spine dynamics are spatially clustered. (A) Distribution of dendritic

segments with no dynamic events, only dynamic spines, only dynamic inhibitory synapses, and both dynamic

spines and inhibitory synapses. (B) Fraction of dynamic inhibitory synapses with nearby dynamic spines as a

function of proximal distance from dynamic inhibitory synapse. (C) Simplified diagram of possible clustered events

between dynamic inhibitory synapses and dynamic dendritic spines. For the purpose of illustration, only a sample

of clustered events are shown; however, for quantifications in (D–H), all dynamic events were scored, including

inhibitory spine and shaft synapses, additions, and eliminations. “d” illustrates a stable inhibitory synapse (light

green arrow) and dynamic inhibitory synapse (dark green arrow) with neighboring dynamic spine (purple square).

“e” illustrates a stable spine (pink arrow) and dynamic spine (purple arrow) with neighboring dynamic inhibitory

synapse (dark green square). “f” illustrates a stable spine (pink arrow) and dynamic spine (purple arrow) with

neighboring dynamic spine (purple square). “g” illustrates a stable spine (light green arrow) and dynamic spine

(dark green arrow) with neighboring dynamic inhibitory synapse (dark green square). (D–G) Cumulative probability

distribution (CPD) of nearest neighbor distances comparing stable and dynamic counterparts and their nearest

dynamic spine or inhibitory synapse. Stable versus dynamic inhibitory synapse to nearest dynamic dendritic spine

for (D). Stable versus dynamic dendritic spine to nearest dynamic inhibitory synapse for (E). Stable versus dynamic

dendritic spine to nearest dynamic dendritic spine for (F). Stable versus dynamic inhibitory synapse to nearest

dynamic inhibitory synapse for (G). (H) Comparison of clustered events (within 10 µm) between dynamic spines

and inhibitory synapses before and during MD. Frequency of events is shown in the left panel. Fraction of dynamic

inhibitory synapses participating in clustered events is shown in the center panel. Fraction of dynamic spines

participating in clustered events is shown in the right panel. ∗, p < 0.05; n = 14 cells from six animals for (A) and

(H); n = 83 dendrites for (B); n = 2,230 dendritic spines, 1,211 inhibitory synapses for (D–G). Error bars, SEM.

30

and their dynamic or stable inhibitory counterparts (Fig. 2.5G). These results demonstrate that

dendritic spine-inhibitory synapse changes are spatially clustered along dendritic segments,

whereas dendritic spine-dendritic spine changes and inhibitory synapse-inhibitory synapse

changes are not. Clustered dynamics were the same for inhibitory shaft or spine synapses in

relation to the nearest dynamic dendritic spine (Fig. 2.8B).

We next asked how altering sensory experience through MD affects clustering of

inhibitory synapse and dendritic spine changes. We found that clustering between dynamic

inhibitory synapses and dendritic spines persisted during MD (Fig. 2.8C) with a similar spatial

distribution compared to control conditions (Fig. 2.8D). We compared the frequency of clustered

events during normal vision and MD by quantifying the number of inhibitory synapses and

dendritic spine changes occurring within 10 µm of each other. MD increased the frequency of

clustered events from 0.013 ± 0.004 to 0.020 ± 0.003 per µm dendrite (Wilcoxon rank-sum test,

p < 0.05; Fig. 2.5H). Since MD increases inhibitory synapse but not dendritic spine dynamics,

we asked how an increase in clustered events could occur without a concurrent change in

dendritic spine remodeling. We found that whereas the fraction of dynamic spines did not

increase in response to MD (Fig. 2.4B–2.4D), the fraction of dynamic spines participating in

clustered events increased from 38.4% ± 9.0% to 59.0% ± 7.7% during MD (Wilcoxon rank-sum

test, p < 0.05).

To rule out the possibility that the increased clustering during MD is simply the result of

the increased presence of dynamic inhibitory synapses, we calculated the likelihood that a

dendrite with a dynamic spine and dynamic inhibitory synapse would be located within 10 µm of

each other assuming these events were not clustered. Based on the density of dendritic spines,

8.3 ± 0.5 spines are located within 10 µm of a dynamic inhibitory synapse. During MD, 6.4% ±

1.0% of spines are dynamic, 88.9% ± 8.3% of which are located on dendrites with dynamic

inhibitory synapses. If changes are not clustered, we calculated a 44.1% ± 7.2% probability that

a dynamic spine would be within 10 µm of a dynamic inhibitory synapse. However, we find that

31

a significantly larger number, 74.3% ± 7.6% of dynamic spines are within 10 µm of a dynamic

inhibitory synapse (Wilcoxon rank-sum test, p < 0.005). Conversely, 4.8 ± 0.3 inhibitory

synapses are located within 10 µm of a dynamic spine. During MD, 13.6% ± 2.0% of inhibitory

synapses are dynamic, 83.1% ± 4.8% of which are located on dendrites with dynamic spines.

We calculated a 50.5% ± 5.0% probability that a dynamic spine would be within 10 µm of a

dynamic inhibitory synapse if changes are unclustered. Here again, we find a significantly larger

number of 75.2% ± 5.1% of dynamic inhibitory synapses located within 10 µm of a dynamic

spine (Wilcoxon rank-sum test, p < 0.001). These results demonstrate that the percent of

clustered dynamic spines and inhibitory synapses in response to MD is significantly higher than

would be expected simply based on the increased fraction of dynamic inhibitory synapses. This

suggests that whereas MD does not alter the overall rate of spine turnover on L2/3 pyramidal

neurons, it does lead to increased coordination of dendritic spine rearrangement with the

dynamics of nearby inhibitory synapses.

Discussion

By large-volume imaging of inhibitory synapses directly on a defined cell type, L2/3

pyramidal neurons, we have characterized the distribution of inhibitory spine and shaft synapses

across the dendritic arbor and measured their remodeling kinetics during normal experience and

in response to MD. We have validated that gephyrin is a faithful marker of inhibitory synapses

through validation by SSEM. Surprisingly we found that inhibitory synapses on dendritic spines

are highly prevalent. Dually innervated spines harboring both an excitatory and an inhibitory

synapse had previously been reported by EM (Fifkova et al., 1992; Kisvarday et al., 1985; Knott

et al., 2002; Kubota et al., 2007; Megias et al., 2001; Parnavelas et al., 1977). However, the

fluorescent labeling provided an orders of magnitude larger sample size as compared to EM,

leading to the realization that on average a third of inhibitory synapses on L2/3 pyramidal

32

neurons reside on spines. Our large volume imaging of entire L2/3 pyramidal neurons labeled

with fluorescent gephyrin in addition to a cell fill, allows the first quantitative look at inhibitory

synapse distributions across the dendritic arbors of this cell type: on average 5 spines and 2

inhibitory synapses per 10 µm. Both spines and inhibitory shaft synapses are distributed

relatively uniformly across the spiny region of the dendritic arbor. In contrast, inhibitory spine

synapses were found to be denser on the apical tuft than on basal and proximal dendrites.

We find that inhibitory synapses targeting dendritic spines and dendritic shafts are

uniquely distributed and display distinct temporal kinetics in response to experience. In addition,

by simultaneous monitoring of inhibitory synapses and dendritic spines across the arbor, we

found that their dynamics are locally clustered within dendrites and this clustering can be further

driven by experience.

We speculate that the differential distribution of inhibitory spine and shaft synapses may

reflect differences in connectivity patterns across dendritic compartments as well as different

roles in the processing of local dendritic activity. Functionally, dendritic inhibition has been

shown to suppress calcium-dependent activity along the dendrite (Miles et al., 1996), originating

from individual excitatory synaptic inputs as well as back-propagating action potentials (bAPs)

from the soma. Local excitation arising from dendritic and NMDA spikes can spread for 10–20

µm and evoke elevated levels of calcium along the dendrite (Golding et al., 2002; Schiller et al.,

1997). Our finding that shaft inhibitory synapses are uniformly distributed across dendrites,

whereas inhibitory spine synapses are twice as abundant along distal apical dendrites

compared to other locations suggest that these two types of synapses have different roles in

shaping dendritic activity. The regular distribution of inhibitory shaft synapses may reflect their

ability to broadly regulate activity from multiple excitatory synaptic inputs and from bAPs,

influencing the integration of activity from mixed sources.

The nonuniform distribution of inhibitory spine synapses may reflect differences in the

relative sources of calcium influx at their respective locales. For example, the amplitude of bAPs

33

along dendrites decreases with increasing distance from the soma. Whereas bAPs can routinely

produce calcium influx into the most distal parts of basal dendrites, detectable calcium influx into

the more distal regions of apical dendrites has only been demonstrated under the most stringent

conditions (Larkum and Nevian, 2008). The increased density of inhibitory spine synapses at

distal apical dendrites, a region in which calcium activity is likely to be more dominated by

synaptic inputs than bAPs may reflect an increased relevance in the modulation of individual

synaptic inputs. Indeed, we—along with others (Knott et al., 2002; Kubota et al., 2007)—have

shown that dendritic spines with inhibitory synapses are co-innervated with an excitatory

synapse, suggesting that they may gate synaptic activity of individual excitatory synaptic inputs.

Recently, it has been shown that inhibitory synapses on dendritic spines are capable of blocking

excitatory currents on the spines which they are located (Chiu et al., 2013). The distribution of

inhibitory spine synapses may also relate to the different sources of excitatory connections onto

the apical dendrite, suggesting they may be involved in gating specific types of inputs. The

apical tuft of L2/3 pyramidal neurons receives a larger proportion of excitatory inputs from more

distant cortical and subcortical locations compared to other parts of the dendritic arbor

(Spruston, 2008). Subcortical afferents have been identified as the excitatory input that co-

innervates spines with inhibitory synapses (Kubota et al., 2007), suggesting that these inhibitory

contacts are ideally situated to directly modulate feed-forward sensory-evoked activity in the

cortex. Interestingly, we find that all of these co-innervated spines are stable, both during normal

experience and MD, regardless of the dynamics of the inhibitory spine synapse. This suggests

that subcortical inputs entering the cortex onto dually innervated spines are likely to be directly

gated by inhibition at their entry level, the spine, but because of the structural stability of these

feed forward inputs, their functional modification would have to rely on removal/addition of the

gating inhibitory input. This particular type of excitatory synapse may be much more directly

influenced by the inhibitory network than excitatory synapses on singly innervated spines that

are exposed to the inhibitory network only at the level of the dendrite.

34

Inhibitory synapses are quite responsive to changes in sensory experience. Focal retinal

lesions have been shown to produce large and persistent losses in axonal boutons in the adult

mouse visual cortex (Keck et al., 2011). Our ability to distinguish inhibitory spine and shaft

synapses provide insight into the degree of inhibitory synapse dynamics in the adult visual

cortex. We find that in binocular visual cortex, MD produces a relatively large initial increase in

inhibitory spine synapse loss. Acute changes in inhibitory spine synapse density have also been

observed in the barrel cortex after 24 hr of whisker stimulation (Knott et al., 2002), further

supporting the notion that these synapses are highly responsive and well suited to modulate

feed-forward sensory-evoked activity. Whereas inhibitory spine synapses are responsive to the

initial loss of sensory input, the sustained increase in inhibitory shaft synapse loss we observe

parallels the persistent absence of deprived-eye input and may serve the broader purpose of

maintaining levels of dendritic activity and excitability during situations of reduced synaptic drive.

These losses in inhibitory synapses are consistent with findings that visual deprivation produces

a period of disinhibition in adult visual cortex (Chen et al., 2011; He et al., 2006; Hendry and

Jones, 1988; Keck et al., 2011) that is permissive for subsequent plasticity (Chen et al., 2011)

(Harauzov et al., 2010) (Maya Vetencourt et al., 2008).

Finally, models for synaptic clustering have been proposed as a means to increase the

computational capacity of dendrites (Larkum and Nevian, 2008) and as a form of long-term

memory storage (Govindarajan et al., 2006). These models have generally been derived from

evidence of coordinated plasticity between excitatory synapses (Govindarajan et al., 2011;

Harvey and Svoboda, 2007). We find that clustered plasticity at the level of synapse formation

and elimination can also occur between excitatory and inhibitory synapses and that these

changes occur mainly within 10 µm of each other. This is a distance at which calcium influx and

calcium-dependent signaling molecules from individual excitatory inputs can directly influence

the plasticity of neighboring excitatory synapses (Govindarajan et al., 2011; Harvey and

Svoboda, 2007). These findings provide evidence that experience-dependent plasticity in the

35

adult cortex is a highly orchestrated process, integrating changes in excitatory connectivity with

the active elimination and formation of inhibitory synapses.

36

Supplementary Figures

Figure 2.6. Ex vivo analysis of dendritic spine and inhibitory synapse distribution. (A) Density (left) and ratio to

dendritic spines (right) of inhibitory spine and shaft synapses in proximal (0- 125µm from soma) versus distal (125-

200µm from soma) apical dendrites measured in 50 µm coronal sections of L2/3 pyramidal dendrites expressing

eYFP/Teal-Gephyrin, counterstained with anti-VGAT (n = 5 cells from 5 mice). (B) Two-photon imaging of

eYFP/Teal-Gephyrin labeled dendrite taken in vivo at ~200 µm below the pial surface and ex vivo in a fixed 50 µm

slice section. Letters indicate re-identified Teal-Gephyrin puncta. (C) Quantification of the 9 fraction of

corresponding dendritic spines and inhibitory synapses along individual dendritic segments in one cell imaged in

vivo and ex vivo in 50 µm slice sections as a function of in vivo imaging depth. (D) Fraction of corresponding

dendritic spines and inhibitory synapses for one cell imaged in vivo and ex vivo in 50 µm slice sections binned by

in vivo imaging depth (n = 157 dendritic spines, 95 inhibitory synapses). (*P < 0.05) Scale bar: 5 µm.

37

Figure 2.7. Most inhibitory synapse dynamics are persistent and uniformly distributed across imaging intervals (A)

Fraction of transient changes out of total number of inhibitory spine or shaft synapses before and after MD is

negligible. (B) Fraction of transient changes out of total dynamic events for inhibitory spine or shaft synapses before

and after MD is small. (C) Dendritic spine, inhibitory spine synapse, and inhibitory shaft synapse dynamics

monitored across 16 days of normal experience at 4 day intervals, shows no bias towards increased eliminations

with consecutive imaging sessions. (n = 3 cells).

38

Figure 2.8. Additional analysis of clustered events. (A) Fraction of dynamic spines with nearby dynamic inhibitory

spines as a function of proximal distance from dynamic spine. (B-D) Cumulative probability distribution (CPD) of

nearest neighbor distances comparing (B) Dynamic inhibitory shaft synapse vs. dynamic inhibitory spine synapse

to nearest dynamic dendritic spine. (C) Stable vs. dynamic inhibitory synapse to nearest dynamic dendritic spine

during MD. (D) Dynamic inhibitory synapse to nearest dynamic dendritic spine during control vs. MD. (n = 2230

dendritic spines, 1211 inhibitory synapses).

39

Experimental Procedures

Generation of Expression Plasmids

For construction of the Cre expression plasmid (pFsynCreW), a Cre insert with 5′ NheI and 3′

EcoRI restriction sites was generated by PCR amplification from a WGA-Cre AAV vector

(Gradinaru et al., 2010) and subcloned into a pLL3.7syn lentiviral expression plasmid (Rubinson

et al., 2003). The Cre-dependent eYFP expression plasmid (pFUdioeYFPW) was constructed

by subcloning a “double” floxed inverse orientation (dio) eYFP expression cassette (a gift from

K. Deisseroth) into the pFUGW lentiviral expression plasmid (Lois et al., 2002), replacing

the GFP coding region between the 5′ BamHI and 3′ EcoRI restriction sites. The Cre-dependent

Teal-gephyrin expression plasmid (pFUdioTealGephyrinW) was constructed as follows. First, a

Teal insert was generated by PCR amplification of Teal (Allele Biotech, San Diego, CA, USA)

with added 5′ NheI and 3′ EcoRI restriction sites and subcloned into the pLL3.7syn lentiviral

expression plasmid. Next, Gephyrin with 5′ BsrGI and 3′ MfeI restriction sites was generated by

PCR amplification from a GFP-Gephyin expression plasmid (Fuhrmann et al., 2002) and

subcloned into the Teal expression plasmid using the BsrGI and EcoRI sites to generate a Teal-

Gephryin fusion protein. Finally, Teal-Gephyrin with 5′ BsiWI and 3′ NheI restriction sites

was PCR amplified from this plasmid and subcloned into the Cre-dependent eYFP expression

plasmid described above, replacing eYFP in the dio expression cassette.

In Utero Electroporation

All animal work was approved by the Massachusetts Institute of Technology Committee on

Animal Care; it conforms to the National Institutes of Health guidelines for the use and care of

vertebrate animals. L2/3 cortical pyramidal neurons were labeled by in utero electroporation on

E16 timed pregnant C57BL/6J mice (Charles River, Wilmington, MA, USA) as previously

described (Tabata and Nakajima, 2001). pFUdioeYFPW,

40

pFUdioTealGephyrinW, pFUCreW plasmids were dissolved in 10 mM Tris ± HCl (pH 8.0) at a

10:5:1 molar ratio for a final concentration of 1 µg/µl along with 0.1% of Fast Green (Sigma-

Aldrich, St. Louis, MO, USA). The solution, containing 1-2 µl of plasmid, was delivered into

the lateral ventricle with a 32 gauge Hamilton syringe (Hamilton Company, Reno, NV, USA).

Five pulses of 35–40 V (duration 50 ms, frequency 1 Hz) were delivered, targeting the visual

cortex, using 5 mm diameter tweezer-type platinum electrodes connected to a square wave

electroporator (Harvard Apparatus, Holliston, MA, USA).

Cranial Window Implantation

Mice born after in utero electroporation were bilaterally implanted with cranial windows at

postnatal days 42–57 as previously described (Lee et al., 2008). Sulfamethoxazole (1 mg/ml)

and trimethoprim (0.2 mg/ml) were chronically administered in the drinking water through the

final imaging session to maintain optical clarity of implanted windows.

Optical Intrinsic Signal Imaging

For functional identification of monocular and binocular visual cortex, optical imaging of intrinsic

signal and data analysis were performed as described previously (Kalatsky and Stryker, 2003).

Mice were anesthetized and maintained on 0.5%–0.8% isofluorane supplemented by

chloroprothixene (10 mg/kg, i.m.) and placed in a stereotaxic frame. Heart rate was continuously

monitored. For visual stimuli, a horizontal bar (5° in height and 73° in width) drifting up with a

period of 12 s was presented for 60 cycles on a high refresh rate monitor positioned 25 cm in

front of the animal. Optical images of visual cortex were acquired continuously under 610 nm

illumination with an intrinsic imaging system (LongDaq Imager 3001/C; Optical Imaging Inc.,

New York, NY, USA) and a 2.5×/0.075 NA (Zeiss, Jena, Germany) objective. Images were

spatially binned by 4×4 pixels for analysis. Cortical intrinsic signal was obtained by extracting

the Fourier component of light reflectance changes matched to the stimulus frequency, whereby

41

the magnitudes of response in these maps are fractional changes in reflectance. The magnitude

maps were thresholded at 30% of peak response amplitude to define a response

region. Primary visual cortex was determined by stimulation of both eyes. Binocular visual

cortex was determined by stimulation of the ipsilateral eye. Monocular visual cortex was

determined by subtracting the binocular visual cortex map from the primary visual cortex map.

Monocular Deprivation

Monocular deprivation was performed by eyelid suture. Mice were anesthetized with 1.25%

avertin (7.5 ml/kg IP). Lid margins were trimmed and triple antibiotic ophthalmic ointment

(Bausch & Lomb, Rochester, NY, USA) was applied to the eye. Three to five mattress stitches

were placed using 6-0 vicryl along the extent of the trimmed lids. Suture integrity was inspected

directly prior to each imaging session. Animals whose eyelids did not seal fully shut or had

reopened were excluded from further experiments.

Serial Section Immunoelectron Microscopy

For post hoc localization of previously in vivo imaged dendrites, blood vessels were labeled

with a tail vein injection of fixable rhodamine dextran (5% in PBS, 50 µl; Invitrogen, Carlsbad,

CA, USA) delivered 30 min prior to perfusion. Animals were fixed and perfused with an initial

solution of 250 mM sucrose, 5 mM MgCl2 in 0.02 M phosphate buffer (PB; pH 7.4), followed by

4% paraformaldehyde containing 0.2% picric acid and 0.5% glutaraldehyde in 0.1 M PB.

Following perfusion and fixation, cranial windows were removed and penetrations of DiR

(Invitrogen) were made into cortex around the imaged region. Brains were removed, 50 µm thin

sections were cut parallel to the imaging plane and visualized with an epifluorescence

microscope. The brain section containing the branch tip of interest was identified by combining

in vivo two-photon images and blood vessel maps with post hoc blood vessel labeling and DiR

penetrations. The identified section was prepared for immunoelectron microscopy as previously

42

described (Kubota et al., 2009). Imaged dendrites were stained by immunohistochemistry using

an antiserum against eGFP (1: 2,000; kind gift from Dr. Nobuaki Tamamaki, Kumamoto

University, Japan), followed by biotin-conjugated secondary antiserum (1:200; BA-1000, Vector

Laboratories, Burlingame, CA, USA) and then the ABC kit (PK-6100, Vector Laboratories). The

neurons were labeled with 0.02% DAB, 0.3% nickel in 0.05 M Tris-HCl buffer (pH 8.0). Prepared

sections were then serially resectioned at 50 nm thickness using an ultramicrotome (Reichert

Ultracut S, Leica Microsystems, Wetzlar, Germany). The ultrathin sections were incubated with

an antiserum against GABA (1:1,000; A-2052, Sigma-Aldrich) in 0.1% Triton X-100, 0.05 M Tris-

HCl buffer, followed by 15 nm colloidal gold conjugated secondary antiserum (1:100;

EM.GAR15, British Biocell International, Cardiff, UK; (Kawaguchi and Kubota, 1998). Whereas

DAB staining can obscure postsynaptic structures, postembedding GABA immunoreactivity has

been demonstrated to detect 100% of inhibitory presynaptic terminals (Kawaguchi and Kubota,

1998). The stained dendrite was imaged from 136 serial ultra thin sections using TEM (Hitachi

H-7000 equipped with AMT CCD camera XR-41, Hitachi, Japan). Image reconstruction and

analysis was performed with Reconstruction (http://synapses.clm.utexas.edu/tools/index.stm).

Two-Photon Imaging

Starting at three weeks after cranial window surgery, allowing sufficient time for recovery, adult

mice were anesthetized with 1.25% avertin (7.5 ml/kg IP). Anaesthesia was monitored by

breathing rate and foot pinch reflex and additional doses of anesthetic were administered during

the imaging session as needed. In vivo two-photon imaging was performed using a custom-built

microscope, including a custom-made stereotaxic restraint affixed to a stage insert and custom

acquisition software modified for dual channel imaging. The light source for two-photon

excitation was a commercial Mai Tai HP Ti:Sapphire laser (Spectra-Physics, Santa Clara, CA,

USA) pumped by a 14 W solid state laser delivering 100 fs pulses at a rate of 80 MHz with the

power delivered to the objective ranging from approximately 37–50 mW depending on imaging

43

depth. Z-resolution was obtained with a piezo actuator positioning system (Piezosystem Jena,

Jena, Germany) mounted to the objective. The excitation wavelength was 915 nm, with the

excitation signal passing through a 20x/1.0 NA water immersion objective (Plan-Apochromat,

Zeiss, Jena, Germany). After exclusion of excitation light with a barrier filter, emission photons

were spectrally separated by a dichroic mirror (520 nm) followed by bandpass filters (485/70

and 560/80 nm) and then collected by two independent photomultiplier tubes. An initial low-

resolution imaging volume (500 nm/pixel XY-resolution, 4 µm/frame Z-resolution) encompassing

a labeled cell was acquired to aid in selecting the region of interest for chronic imaging. All

subsequent imaging for synapses and dendritic spine monitoring was performed at higher

resolution (250 nm/pixel XY-resolution, 0.9 µm/frame Z-resolution). Two-photon raw scanner 16

bit data was processed for spectral linear unmixing and converted into an 8 bit RGB image z-

stack using Matlab and ImageJ (National Institutes of Health, Bethesda, MD, USA).

Spectral Linear Unmixing and Image Processing

Spectral linear unmixing is based on the fact that the total photon count recorded at each pixel

in a given channel is the linear sum of the spectral contribution of each fluorophore weighted by

its abundance. For a dual channel detection system, the contribution of two fluorophores can be

represented by the following equations:

J 1 (x,y)=s 1 , 1×I 1 (x,y)+s 1 , 2×I 2 (x,y) , (1)

J 2 (x,y)=s 2 , 1×I 1 (x,y)+s 2 , 2×I 2 (x,y) , (2)

where J is the total signal per channel, I is the fluorophore abundance, and S is the contribution

of that fluorophore. These equations can be expressed as a matrix:

[J ]= [S] [ I ] , (3)

whereby the unmixed image [I] can be calculated using the inverse matrix of S:

[ I ]=[S] − 1 [J] . (4)

44

Assuming the detected signal in both channels represents the total spectral contribution for both

fluorophores:

s 1 , 1+s 2 , 1=1, (5)

s 1 , 2+s 2 , 2=1. (6)

[S] was determined experimentally by dual channel acquisition of single excitation two-photon

images of cell culture with single fluorophore expression, adjusting laser power and dwell time

to achieve photon count levels approximating in vivo signal intensity. The mean contribution for

each fluorophore into each channel representing the reference spectra from the acquired

images was calculated using Matlab (Mathworks, Natick, MA, USA). These values were

subsequently used for spectral linear unmixing of dual channel 16 bit two-photon raw scanner

data into an 8 bit RGB image z-stack using Matlab and ImageJ (National Institutes of Health). S

measured from in-vivo-labeled samples was similar to the in vitro determined value, and

spectral unmixing with either the in vitro or in vivo values yielded essentially the same result.

Simulations were performed to validate that the 200–250 µm imaging depth used for our data

acquisition is well within the signal intensity range, where spectral unmixing can work reliably.

Data and Statistical Analysis

For whole-cell dendritic arbor reconstruction and analysis of dendritic morphology, 3D stacks

were manually traced in Neurolucida (MicroBrightField, Inc., Williston, VT, USA). The

main apical trunk of each cell was excluded from analysis as its orientation was perpendicular to

image stacks and thus could not be reconstructed at high resolution. Dendrites are defined as

dendritic segments stretching from one branch point to the next branch point or from one branch

point to the branch tip. Dendritic spine and inhibitory synapse tracking and analysis was

performed using V3D (Peng et al., 2010). Dendritic spine analysis criteria were as previously

described (Holtmaat et al., 2009). Using these scoring criteria, the lack of image volume rotation

from imaging session to session may have resulted in some z-projecting dendritic spines being

45

left unscored. This did not influence quantification of spine dynamics due to their low incidence

and the fractional scoring. Inhibitory synapses were identified as puncta colocalized to the

dendrite of interest with a minimal size of 3×3 or 8–9 clustered pixels (0.56 µm2) with a minimal

average signal intensity of at least four times above shot noise background levels. Comparing

the measured pixel dimensions of two-photon imaged synapses with their size measured after

EM reconstruction, shows that the pixel dimensions of these synapses range from 8–31 pixel

clusters and correlate with their true physical dimensions. This confirms that a threshold of 3×3

pixels would be sufficient for identifying virtually any synapse.

Transient changes were defined as dendritic spines or synapses that appeared or

disappeared for only one imaging session. For persistent changes, spines or synapses that

appeared and persisted for at least two consecutive imaging sessions were scored as additions,

with the exception of additions occurring between the second-to-last and last imaging session.

Spines or synapse that disappeared and remained absent for at least two consecutive imaging

sessions were scored as eliminations, with the exception of eliminations occurring between the

first and second imaging session. Only persistent spine and synapse changes were used for

measuring turnover rates and clustered dynamics. In total, 2,230 dendritic spines and 1,211

inhibitory synapses from 83 dendritic segments in 14 cells from 6 animals were followed over 6

imaging sessions.

For each cell, the fractional rate of additions and eliminations were defined as the

percentage of dendritic spines or inhibitory synapses added or eliminated, respectively, between

two successive imaging sessions out of the total number of dendritic spines or inhibitory

synapses divided by the number of days between imaging sessions. The Wilcoxon rank-sum

test, Mann-Whitney U-test, or repeated-measures ANOVA and Tukey's post hoc test were used

for statistical analysis of time course data or dendritic density, where n indicates the number of

cells or dendritic segments. The Kolmogorov-Smirnov test was used for statistical analysis of

46

nearest neighbor distance distributions, where n indicates the number of dendritic spines or

inhibitory synapses. All error bars are SEM.

47

Chapter 3 - Inhibitory synapses are repeatedly

assembled and removed at persistent sites in vivo

Portions of this chapter will be published in:

KL Villa, KP Berry, J Subramanian, JW Cha, WC Oh, HB Kwon, PTC So, Y Kubota, E Nedivi. "Inhibitory synapses are repeatedly assembled and removed at persistent sites in vivo."Submitted.

Contributions: 2 photon image acquisition and analysis for Figures 3.1,3.4-6, 3.10-12 were done by myself in collaboration with Kalen Berry. Yoshiyuki Kubota performed EM analysis and reconstruction (Figure 3.2, 3.7, 3.9, 3.13, T1, T2) and Won Chan Oh performed slice culture electrophysiology (Figure 3.3), Jaichandar Subramanian cloned the PSD95-mCherry construct and contributed to data analysis.

Abstract

Older concepts of a hard-wired adult brain have been overturned in recent years by imaging

studies visualizing synaptic remodeling in vivo, remodeling thought to mediate rearrangements

in microcircuit connectivity. Using three-color labeling and spectrally resolved two-photon

microscopy, we monitor in parallel the structural dynamics (assembly or removal) of excitatory

and inhibitory postsynaptic sites on the same individual neurons in mouse visual cortex in vivo.

We found that on a daily basis, inhibitory synapses on dendritic spines are exceptionally

dynamic, largely because many disappear and reappear in the same location. In contrast,

excitatory synapses on the same spines are remarkably stable. Monocular deprivation, a model

of sensory input-dependent plasticity, shortens inhibitory synapse lifetimes and lengthens

intervals to recurrence, resulting in a new dynamic state with reduced inhibitory synaptic

presence. Reversible structural dynamics indicates a fundamentally new role for inhibitory

synaptic remodeling – flexible, input-specific modulation of stable excitatory connections.

48

Introduction

Historically, inhibitory synapses onto pyramidal cell dendrites were thought to reside

mostly on the dendritic shaft, serving to locally modulate excitability within short dendritic

segments. According to computational models of cortical circuit function, placement of inhibitory

synapses along the dendritic shaft is strategic (Gidon and Segev, 2012), in that rearrangement

of a small number of well-placed inhibitory synapses can significantly influence segment- or

branch-specific computations (Chklovskii et al., 2004; Poirazi and Mel, 2001). Recently, the

ability to directly visualize inhibitory synapses in vivo revealed that approximately 30% reside on

dendritic spines, rather than on the dendritic shaft (Chen et al., 2012). Inhibitory synapses on

dendritic spines are always adjacent to an excitatory synapse on the same spine, and can

directly shunt excitation onto that synapse (Chiu et al., 2013). The prevalence of these dually

innervated spines (DiS) with a compartmentalized mode of inhibitory modulation raises

questions regarding how structural plasticity of this new class of inhibitory synapse could

effectively act to alter local excitatory circuit properties. Yet, it is unknown how rapidly inhibitory

synapses on dendritic spines or shafts can be assembled or removed in vivo, and how such

changes in inhibitory synapses relate to changes in neighboring excitatory synapses.

Fluorescent proteins fused to postsynaptic scaffolding molecules have recently

enabled direct visualization of synaptic remodeling in vivo (Cane et al., 2014; Chen et al.,

2012; Kelsch et al., 2008; van Versendaal et al., 2012). However, the difficulty of

simultaneously imaging both inhibitory and excitatory synapses in vivo has precluded a side-by-

side comparison of their dynamic properties across subcellular compartments. At best, inhibitory

synapses have been directly tracked in combination with a cell fill, where spine dynamics were

assumed to represent excitatory synapse dynamics (Chen et al., 2012; van Versendaal et al.,

2012). Here, we utilize in vivo, triple-color, two-photon microscopy for side by side imaging of

excitatory and inhibitory synapses on individual cortical pyramidal neurons tracked daily for

eight consecutive days. This allowed the resolution of spines lacking synapses, spines

49

innervated by a single excitatory synapse, termed singly innervated spines (SiS), and spines

innervated by both an excitatory and an inhibitory synapse, termed dually innervated spines

(DiS), as unique populations with different dynamic properties. Monitoring the short-term

dynamics of excitatory synapses on SiS vs DiS showed that excitatory synapses on DiS were

remarkably stable as compared to those on SiS. Further, side-by-side comparison of excitatory

and inhibitory synapses residing on the same DiS, revealed a stark contrast between the

stability of the excitatory synapse and the unusually dynamic nature of the inhibitory synapse on

the same spine.

The percentage of dynamic structures seen with daily interval imaging were surprisingly

high in comparison to previous studies imaging at 4-day intervals (Chen et al., 2012), indicating

that many dynamic events are quickly reversed and therefore undetected with less frequent

imaging. This was most pronounced for inhibitory synapses on DiS, which frequently

disappeared and then reappeared in the same location. EM reconstruction of spines after the

postsynaptic element of a recurrent inhibitory synapse disappeared, in some cases revealed

that the presynaptic inhibitory axon remained in place, potentially serving as a placeholder for

the synapse’s return. This new, recurrent type of synaptic structural dynamics provides a

potential mechanism for reversible gating of specific excitatory connections.

Monocular deprivation (MD), a model of sensory input-dependent plasticity, significantly

increased the number of dynamic inhibitory events, compared to normal experience (NE). After

MD, dynamic inhibitory synapses are present for fewer consecutive days and inhibitory spine

synapses are gone longer before reoccurring. These results show that the observed disinhibition

previously seen by our lab and others (Chen et al., 2011; Chen et al., 2012; van Versendaal et

al., 2012) in response to adult MD is not just a matter of a one-time inhibitory synaptic loss upon

deprivation. Rather, MD results in a transition to a new dynamic state for inhibitory synapses

50

where their net presence is reduced throughout the deprivation period.

Results

Triple-color labeling of pyramidal cells

To visualize the full complement of inhibitory and excitatory postsynaptic sites onto

individual L2/3 pyramidal cells in mouse V1, we extended our previous strategy for monitoring

inhibitory synapse dynamics in vivo (Chen et al., 2012), to include excitatory synapses labeled

with a third color. In our previous imaging studies, conducted at 4 or 7 day intervals, it was

reasonable to assume that a dendritic spine that lasted for 2 consecutive imaging sessions

contained an excitatory synapse. However, this is not necessarily the case when imaging at

shorter intervals. Therefore, we co-electroporated three ‘double-floxed' inverted open reading

frame (dio) based plasmids (Atasoy et al., 2008; Dhande et al., 2011), expressing either eYFP

as a cell fill, Teal-gephyrin as a postsynaptic marker of inhibitory synapses (Essrich et al.,

1998), or PSD-95-mCherry as a postsynaptic marker of excitatory synapses, at high molar

ratios into E15.5 C57BL/6 mice, together with limiting amounts of a Cre construct (Fig. 3.1A).

This favors high incidence of fluorophore co-expression, while providing the sparse labeling

required for single neuron imaging and reconstruction.

In vivo triple color imaging of eYFP-labeled neuronal morphology, PSD-95-mCherry

puncta, and Teal-gephyrin puncta in individual L2/3 neurons within adult V1 was performed

through a cranial window with a custom built two-photon microscope. All three fluorophores

were simultaneously excited using two excitation wavelengths and their emissions spectrally

separated and collected with three PMTs. Bleed-through photons, due to overlapping emission

spectra, were reassigned to their appropriate color channel by post-hoc spectral linear un-

mixing. To monitor synapse dynamics, neurons were imaged 9 times at 24-hour intervals (Fig.

51

3.1B), across a 200x200x200 µm volume capturing a large portion of the dendritic arbor at

250nm/pixel XY and 0.9µm/frame Z-resolution. Triple-labeled neurons showed clear arbor

morphology with distinct resolution of dendritic spines, Teal-gephyrin puncta, and PSD-95-

mCherry puncta (Fig. 3.1C-F).

The linear density per 10 µm of spines (4.62 ± 1.09), inhibitory shaft synapses (1.84 ±

0.36), and inhibitory synapses on DiS (0.79 ± 0.26) (Fig. 3.2A) closely matched our two-color

imaging data (Chen et al., 2012), and spine density was unaltered by co-expression of Teal-

gephyrin and PSD-95-mCherry synaptic markers (Fig. 3.8A-C). Additionally, since in this case

PSD-95-mCherry served as an excitatory synaptic marker independent of the spine fill, we could

discern that an average 21.6 ± 8.5% (335/1380) of spines lacked PSD-95-mCherry labeling

(Fig. 3.2A). This is consistent with reports from dual color imaging of PSD-95-eGFP against

Figure 3.1. Triple color labeling of L2/3 pyramidal cells in vivo. (A) Plasmid combination for labeling cell fill

(eYFP), inhibitory synapses (Teal-gephyrin), and excitatory synapses (PSD-95-mCherry). (B) Experimental time

course. (C) Low-magnification maximum z-projection (MZP) of cell fill pseudo-colored red. (D) MZP of dendritic

segment from (C) with labeled inhibitory synapses (cyan) and excitatory synapses (pseudo-colored green). (E and

F) Examples of dynamic synapses from boxed regions in (D) on indicated days. Upper, middle, and lower panels

show 3 channel merge, Teal-gephyrin alone, and PSD-95-mCherry alone, respectively. Arrows denote dynamic

synapses. Inhibitory synapses in (E) appear on days 8 and 9. Excitatory synapse in (F) disappears on day 4 and

its spine is removed on day 6. Scale Bars in µm: (C) 10; (D) 5; (E and F) 2.

52

DsRed Express fills of L2/3 pyramidal neurons imaged in vivo (Cane et al., 2014). These data

indicate that adding PSD-95-mCherry did not alter either inhibitory or excitatory synaptic

numbers.

Teal-gephyrin and PSD-95-mCherry puncta represent inhibitory and excitatory synapses,

respectively

Teal-gephyrin puncta had been previously validated as a faithful marker of inhibitory

synapses (Bloodgood et al., 2013; Chen et al., 2012; Isshiki et al., 2014; van Versendaal et al.,

2012), and PSD-95 fused to GFP has been used in vivo as a marker of excitatory synapses

(Cane et al., 2014; Isshiki et al., 2014; Kelsch et al., 2008). To further demonstrate that Teal-

gephyrin and PSD-95-mCherry puncta visualized in vivo correspond to inhibitory and excitatory

synapses, respectively, we performed Focused Ion Beam / Scanning Electron Microscope

(FIB/SEM) on in vivo imaged L2/3 pyramidal neuron dendrites (Figs. 2B-F and S2). Immediately

after the last two-photon imaging session, brains were fixed, sectioned, and stained with an

antibody to eYFP followed by a biotin-conjugated secondary and detected with nickel-

diaminobenzidine (DAB). Five dendritic segments totaling 90 µm of dendritic length with

observable DAB staining were relocated and then further investigated using a combined

FIB/SEM (Kubota et al., 2011a) or postembedding immuno-SSEM (Serial Section Electron

Microscopy) (Chen et al., 2012). The FIB/SEM provides a larger field of view, allowing sampling

of more synapses. However, image quality is not as high as in SSEM. Its relative fuzziness in

combination with the DAB staining makes it difficult to distinguish between symmetric and

asymmetric synapses. While the FIB/SEM is sufficient for identifying a synaptic junction

(Kubota, 2015), the designation of whether it is inhibitory therefore relies on other criteria (see

below).

53

In agreement with our previous findings (see Supplementary Figures S1 & S3 in (Chen

et al., 2012)), Teal-gephyrin puncta on the final imaging session proved a faithful marker of

symmetric synaptic presence with no false negatives found by EM. One hundred percent of

Teal-gephyrin puncta visualized on the last session of two-photon imaging were identified with a

synaptic cleft and vesicles by EM. From a total of 28 teal-gephyrin positive synapses, 27 were

Figure 3.2. PSD-95-mCherry and Teal-gephyrin puncta represent excitatory and inhibitory synapses,

respectively. (A) Density of dendritic spines, inhibitory synapses on the dendritic shaft, and inhibitory

synapses on dually innervated spines (DiS), per 10 µm dendritic length. 1555 spines were tracked from 63

dendritic segments from 8 animals. Error bars represent SEM. (B) Nonconsecutive z-stacks from branch

‘ka’ (Supplemental Table 1) session 9, chosen to optimally display all synapses. Letters indicate spines,

numbers indicate inhibitory synapses. (C) 3D reconstructed dendrite from FIB/SEM image stacks shows

perfect correspondence with session 9. (D) Excitatory synapse ‘e’ on SiS (yellow arrows). (E) Inhibitory

shaft synapse ‘8’ (blue arrows). (F) Nonconsecutive micrographs of DiS ‘g4’. The large perforated synapse

seen in both sections was designated excitatory (pink arrows), and the smaller synapse seen in the left

panel inhibitory (red arrows). Arrows flank edges of synaptic clefts, seen as evenly spaced regions between

pre and post-synaptic membranes with synaptic vesicle clusters in the pre-synaptic terminal. Scale Bars in

µm: (A,B) 1.5; (D-F) 0.2.

54

confirmed as inhibitory according to at least one of the following criteria: 1) they were located on

the dendritic shaft; 2) they stained positive for GABA thorough post-hoc immuno-EM; 3) we

were able to trace the axon to another symmetric synapse formed onto a cell lacking DAB

staining; 4) Two terminals were clearly visible on a DiS and the excitatory terminal was validated

by other criteria (Fig. 3.2B-F, Fig. S3.2, Table T1). Inhibitory synapses on spines were

exclusively found on spines that also contained an excitatory synapse (DiS).

PSD-95-mCherry puncta visualized on the last imaging session in vivo corresponded to

asymmetric (excitatory) synapses, with no false positives (Fig. 3.2B-F, Fig. S3.2, Table T1).

From a total of 17 PSD-95-mCherry puncta, 100% were validated as excitatory synapses by the

following criteria: 1) Location on a SiS; 2) clear postsynaptic density observed despite DAB

staining; 3) perforated synapse; 4) backtracing the excitatory axon to its nearest asymmetric

synapse onto a non-DAB stained cell. Some very small or newly formed excitatory synapses

were seen by EM, but not observed with PSD-95-mCherry in vivo (6 in total), potentially due to

the presence of MAGUKs other than PSD-95. Immature synapses containing MAGUKs such as

MAP102 or SAP97 and lacking PSD-95 have been reported early in development (Elias et al.,

2008; Sans et al., 2000) and by EM in the adult rat visual cortex (Aoki et al., 2001).

Teal-gephyrin puncta on dually innervated spines represent functional GABAergic synapses

To confirm that Teal-gephyrin puncta represent functional GABAergic synapses and that

lack of a Teal-gephyrin punctum indicates the absence of a functional GABAergic synapse, we

performed two-photon GABA uncaging and electrophysiological recordings in organotypic slice

cultures of L2/3 pyramidal neurons expressing tdTomato and Teal-gephyrin. CDNI-GABA was

uncaged adjacent to spines with or without Teal-gephyrin puncta (Fig. 3.3A). Spines containing

55

Teal-gephyrin puncta evoked an average uncaging-evoked inhibitory postsynaptic current

(uIPSC) of 5.80 ± 0.21 pA, and spines lacking Teal-gephyrin puncta evoked an average uIPSC

of 1.92 ± 0.13 pA (Fig. 3.3B-C). Spines with and without gephyrin showed little overlap in terms

of uIPSC amplitudes (Fig. 3.3D). The uIPSCs evoked by GABA uncaging adjacent to spines

Figure 3.3. Teal-gephyrin puncta on dually innervated spines represent functional GABAergic synapses.

(A) Representative image of a dendritic segment from an L2/3 pyramidal neuron in organotypic slice culture

expressing tdTomato (red) and Teal-gephyrin (green). Arrowhead marks a gephyrin positive spine. Crosses

indicate two-photon GABA uncaging spots. Scale bar= 1µm. (B) Representative uIPSC traces evoked from a spine

positive for gephyrin (1, green) and its neighboring gephyrin-negative spine (2, red). Arrowheads indicate onset of

GABA uncaging. (C) Summary graph of uIPSC amplitudes from gephyrin positive spines (green bar; n = 42 spines,

18 cells) and spines lacking gephyrin puncta (red bar; n = 44 spines. **p < 0.01 by two-tailed student’s t-test). (D)

Segregation of uIPSC amplitudes between spines with gephyrin puncta (green circles) and lacking puncta (red

circles). (E) Gephyrin negative spines (red) clearly segregate from gephyrin positive spines (green). Positive

correlation between uIPSCs and Teal-gephyrin signal intensity (r = 0.53 by Pearson’s correlation; p < 0.01) and no

correlation between uIPSC and background gephyrin intensity of gephyrin negative spines (r=-0.09 by Pearson’s

correlation; p=0.56; n = 42 spines, 18 cells). Error bars are S.E.M.

56

lacking Teal-gephyrin puncta were smaller than 4 pA, possibly due to nonsynaptic GABA

receptors, known to be diffusely distributed across the cell membrane (Nusser et al., 1998;

Thomas et al., 2005; Tretter et al., 2008; Yeung et al., 2003). The uIPSCs evoked from Teal-

gephyrin containing spines were mostly 5 pA or higher, likely representing synaptic currents

from clustered GABA receptors. Teal-gephyrin puncta size correlated with uIPSC amplitude,

with larger synapses producing stronger uIPSC currents, while Teal background fluorescence

on spines without teal-gephyrin puncta showed no correlation with uIPSC amplitude (Fig. 3.3E).

Thus, gephyrin positive and gephyrin negative spines segregate as two distinct populations in

terms of GABA evoked uIPSCs. Teal-gephyrin fluorescence imaged in vivo displayed a

similarly clear threshold separating between spines with and without Teal-gephyrin puncta (Fig.

3.10A), with the size distributions of teal-gephyrin puncta comparable between the in vivo

imaging and organotypic slice data (Fig. 3.10B). These data are consistent with the requirement

for gephyrin for synaptic GABA(A) receptor clustering (Kneussel et al., 1999) and indicate that

Teal-gephyrin puncta on spines correspond to functional inhibitory synapses.

The majority of excitatory synapses and the spines that bear them are stable

We then asked whether the presence of an excitatory synapse influenced spine

dynamics. Our triple-color labeling strategy now allows the classification of spines into 3 distinct

categories, spines lacking PSD-95, singly innervated spines (SiS) which contain a PSD-95

puncta alone, and DiS that contain an excitatory synapse as well as an inhibitory synapse (Fig.

3.4A). In agreement with (Cane et al., 2014), we found that the stable spine population was

mostly comprised of spines with PSD-95-mCherry puncta. Spines that never acquired a PSD-

95-mCherry punctum were highly dynamic, with 80.7 ± 2.9% (288/359) dynamic over the entire

imaging period (Fig. 3.4B). Only 12.6 ± 3.7% (104/848) of SiS were dynamic during this period.

57

Many first lost their PSD-95-mCherry punctum and then disappeared (Fig. 3.1F), or appeared

and then later gained PSD-95-mCherry (Fig. 3.4E). DiS were the most stable category of

spines, with 99.0 ± 0.5% (345/348) stable over the entire imaging period (Fig. 3.4B). We next

Figure 3.4. Most excitatory synapses and the spines that bear them are stable. (A) Proportion of spines

without PSD-95, singly innervated spines (SiS) containing only PSD-95, and dually innervated spines (DiS)

containing both PSD-95 and gephyrin. (B) Fraction of dynamic and stable spines normalized by subclass. (C)

Fraction of dynamic PSD-95 puncta on subclasses of spines, compared to dynamic gephyrin puncta on DiS (n=8

cells and 1555 spines. *p<0.05, **p<0.001 ***p<0.0001 by ANOVA in A-C). (D and E) Examples of dynamic spines.

Arrows denote dynamic spines, filled when spine is present, empty when spine is absent. Scale bars = 2 µm. (F)

Stable spines are larger than dynamic spines. Cumulative probability distribution (CPD) comparing between size

of dynamic and stable spines, assessed as the ratio of YFP cell fill intensity in the spine head to the average

intensity of the dendritic branch (n = 3 cells with 367 stable, 54 dynamic. p= 4.2x10-13 by KS test). (G) CPD

comparing sizes of spines with PSD-95 to spines without PSD-95 (n= 3 cells with 362 with PSD-95, 59 without

PSD-95. p=6.2x10-17 by KS test). (H) CPD comparing spine size between the 3 spine categories. DiS spines are

larger than SiS and spines without PSD-95 (n= 3 cells with 119 DiS, 243 SiS, 59 no PSD-95. p= 6.4x10-13 by KS

58

examined the dynamics of PSD-95 puncta on spines and asked, ‘how do the dynamics of

excitatory synapses differ from the dynamics of inhibitory synapses on spines?’ We found that,

while PSD-95 puncta on stable SiS are occasionally dynamic (26.2 ± 4.7% dynamic, 177/744),

PSD-95 puncta on DiS are extremely stable (97.8 ± 1.0% stable, 340/348), in stark contrast to

inhibitory synapses on DiS, the majority of which (64.8 ± 5.3%, 215/348) are dynamic (Fig.

3.4C, p<0.0001 by one-way ANOVA with Tukey’s Multiple Comparison Test).

Other studies have shown that larger spines are more likely to be stable (Holtmaat et al.,

2005; Trachtenberg et al., 2002). Consistent with these results, we found that dynamic spines

are significantly smaller than stable spines (Fig. 3.4F, p= 4.2 x10-13 by KS test), and that spines

lacking PSD-95 are smaller than spines containing PSD-95 (Fig. 3.4G, p=6.2x10-17 by KS test).

With the addition of the Teal-gephyrin label, we were able to separate spines with PSD-95 into

two populations, distinguishing SiS from DiS. This revealed that SiS are significantly smaller

than DiS (Fig. 3.4H, p= 6.4x10-13 by KS test). Thus, the stability of DiS could be due to

increased stabilization resulting from either the presence of two synapses or the larger size of

DiS, although these two factors are likely related.

Inhibitory synapses disappear and appear again in the same location

We find that inhibitory synapses on DiS are by far the most dynamic population,

significantly more dynamic than either spines or inhibitory synapses on the dendritic shaft (Fig.

3.5A, p<0.001 by ANOVA with Tukey’s Multiple Comparison Test). The percentage of dynamic

inhibitory synapses seen with daily interval imaging were surprisingly high in comparison to

previous studies imaging at 4-day intervals (Chen et al., 2012), with 25.7 ± 3.9% (335/1380) of

spines, 19.0 ± 2.3% (89/534) of inhibitory synapses on the dendritic shaft, and 64.2 ± 5.3%

(175/304) of inhibitory synapses on DiS dynamic. This was not due to increased

59

photobleaching, since dynamics were not significantly different in the first 2 days of imaging

compared to the last 2 days of imaging (Fig. 3.8D). To further rule out technical issues related to

the shorter imaging interval, we restricted our analysis to data exclusively from day 0, 4, and 8,

as if we had imaged at 4-day intervals, and found that dynamics were comparable to our

previous study on a per day basis (Chen et al., 2012), and significantly lower than for the same

cells analyzed daily (Fig. 3.5B also see Fig. 3.11A for further clarification). This difference was

Figure 3.5. Increased dynamics of inhibitory synapses on DiS seen with short interval imaging is due to

their repeated disappearance/reappearance in the same location. (A) Comparison of % dynamic vs stable

dendritic spines, inhibitory synapses on DiS, and inhibitory synapses on the dendritic shaft (***p<0.001 by ANOVA).

(B) Comparison of % dynamic structures per day with daily imaging vs 4-day imaging intervals (*p<0.02 **p<0.002

by two tailed student’s T-test). (C) Percentage of one time dynamic, transient and recurrent synapses. (n= 8 cells

and 1555 spines, 955 inhibitory synapses from 63 dendrites in A-C. ***p<0.0001 by ANOVA). (D and E) Daily

imaging sessions showing examples of recurrent dynamic gephyrin puncta on DiS. Arrows denote dynamic

synapses, filled when synapse is present, empty when synapse is absent. Note that the PSD-95 puncta are stable

over all sessions on these spines. Scale bars = 2 µm. All error bars represent SEM. Tukey’s Multiple Comparison

Test was used for all ANOVA comparisons.

60

particularly robust for inhibitory synapses on DiS, with 28.8 ± 7.7% dynamic per day when

imaged and analyzed at daily intervals vs. 11.8 ± 2.9% dynamic per day when analyzed only for

4-day intervals. The increase in dynamic events when imaging at short intervals indicates that

many dynamic events are quickly reversed and therefore undetected with less frequent imaging

(examples in Fig 5D-E).

Given the stunning prevalence of short-lived dynamics for the inhibitory synapse

population on DiS, we classified synapses by their dynamic history across the entire 9-day

imaging period. We defined synapses that change once (appear or disappear) and don’t change

again within our imaging time frame as one-time dynamic, that appear once and disappear as

transient, and those that disappear and reappear at least once as recurrent. Approximately half

of the dynamic inhibitory synapses on dendritic shafts, 9.2 ± 1.3% (44/517), were in the one-

time dynamic category, 2.2 ± 0.9% (8/517) transient, and 7.7 ± 1.1% (36/517) recurrent (Fig.

3.5C). Inhibitory synapses on DiS exhibited significantly more of each dynamic category, with

28.8 ± 2.4% (98/348) one-time dynamic, 16.0 ± 3.5% (50/348) transient, and as many as 19.5 ±

2.5% (67/348) of total synapses exhibiting recurrent dynamics. The rapid insertion and removal

of inhibitory synapses on DiS is in striking contrast to the stability of excitatory synapses on the

same spines, which when broken down into these categories are 0.004 ± 0.004% (1/348) one-

time dynamic, 0.008 ± 0.008% (2/348) transient and 0.008 ± 0.008% (4/248) recurrent (Fig.

3.5C, p<0.0005 by ANOVA with Tukey’s Multiple Comparison Test).

Kinetics of inhibitory synaptic dynamics are altered in experience dependent plasticity

We next investigated whether the rapid inhibitory synaptic dynamics revealed by short

interval imaging reflect circuit modifications in response to visual experience. In the adult

mouse, MD of at least 7 days produces a functional ocular dominance (OD) shift in binocular

61

V1, a paradigm for sensory experience-dependent plasticity (Frenkel et al., 2006; Sato and

Stryker, 2008). We performed a MD immediately following the first imaging session and

monitored synaptic dynamics at daily intervals for a total of 7 days. MD animals were compared

Figure 3.6. Kinetics of inhibitory synapse dynamics are altered by an experience dependent plasticity

paradigm. (A) MD increases the % of dynamic inhibitory shaft, but not inhibitory spine synapses. (B) Dynamic

inhibitory shaft synapses are more recurrent after MD. (C) Dynamic inhibitory synapses on DiS are more recurrent

after MD. (D) MD causes an increase in the number of dynamic events per location for both inhibitory shaft and

inhibitory synapses on DiS. (E) MD decreases the average number of consecutive days present for both dynamic

inhibitory shaft and dynamic inhibitory spine synapses. (F) MD increases the number of days between synapse

reappearance for inhibitory synapses on DiS. (A-G) Error bars = SEM. *p<0.05, **p<0.01, ***p<0.0005. All p-values

from 2 tailed student’s T-test; n=7 cells NE (304 inh on DiS, 534 inh shaft), 7 cells MD (556 inh on DiS, 717 inh on

shaft). (G) Survival fraction of observed synaptic events (points), fit with the exponential and constant

term, SF = fe−t/τ + s (lines), allowing visualization of how MD affects the survival of each population.

62

to animals who had normal vision (normal experience NE) for the same period. We found that

MD caused a significant increase in the percent of dynamic inhibitory synapses on shafts (NE

16.9% ± 2.3% (89/534) vs. MD 25.8% ± 1.7% (186/717); p<.01 by two-tailed student’s t-test),

but not on DiS (NE 61.3% ± 6.6% (175/304) vs. MD 58.7% ± 6.6% (283/556); p>.05 by two-

tailed student’s t-test) (Fig 6A) or spines with or without PSD-95 (Fig. 3.12). The PSD-95-

mCherry puncta on DiS are also unaffected by MD.

After MD, both shaft and spine inhibitory synapse populations showed a significant

increase in the number of recurrent events (Fig 6B-C) (% recurrent inhibitory synapses on DiS:

NE 22.8 ± 2.8% (40/175) vs. MD 41.2 ± 1.8% (115/283); p<0.0002; on shaft: NE 25.9 ± 6.5%

(24/89) vs. MD 50.2 ± 6.3% (91/186); p<0.02 by two-tailed student’s t-test). This was at the

expense of one-time changes, which were significantly lower for both populations (% one-time

dynamic inhibitory synapses on DiS: NE 57.0 ± 6.2% (99/175) vs. MD 40.3 ± 4.1% (125/283);

p<0.05; on shaft: NE 63.1 ± 8.4% (52/89) vs. MD 39.4 ± 6.7% (74/186); p<0.05 by two-tailed

student’s t-test). Thus, inhibitory shaft synapses from a stable pool are recruited during MD to

the dynamic population, while inhibitory synapses on DiS, which are already quite dynamic

under naïve conditions, are not. For both inhibitory synapse types, MD shifts dynamics event

from the one-time dynamic to the recurrent category. This increase in recurrence is reflected in

more dynamic events per synaptic location (Fig 6D) (inhibitory synapses on DiS: NE 1.6 ± 0.1

vs. MD 1.9 ± 0.06; p<0.03 by two-tailed student’s t-test; Inhibitory synapses on Shaft: NE 1.5 ± 0.2 vs. MD 1.9 ± 0.1; p<0.05 by two-tailed student’s t-test). Additionally, the average number

of days dynamic synapses are present after MD is decreased in both synapse populations (Fig

6E) (inhibitory synapses on DiS: NE 2.9 ± 0.2 vs. MD 2.3 ± 0.1; p<.05 by two-tailed student’s

t-test; inhibitory synapses on shaft: NE 3.1 ± 0.3 vs. MD 2.5 ± 0.09 ; p<.05 by two-tailed

student’s t-test). Recurrent inhibitory synapses on DiS were also absent for more days before

63

reoccurring after MD (Fig 6F) (NE 1.5 ± 0.1 vs. MD 1.8 ± 0.09; p<0.03 by two-tailed student’s

t-test).

To quantify how the differences in inhibitory synapse dynamics after MD impacted their

overall lifetime, we calculated the survival fractions for each population, and fit an exponential

curve to these values (Fig 6G). The exponential fit is a convenient metric for the mean lifetime

(τ) of each inhibitory synapse population. MD decreased the lifetime of inhibitory synapses on

DiS (NE τ = 3.62 MD; τ = 2.15). The same is true for inhibitory synapses on the shaft (NE τ =

6.69 MD; τ = 2.40) (Fig 6E, G). We hypothesized that the increase in the dynamic inhibitory

shaft population after MD, combined with the shorter mean lifetimes of both dynamic inhibitory

synapse populations, would result in a net disinhibition after MD. When we pool the number of

imaging sessions each inhibitory synapse was present, there is a significant decrease in

inhibitory synapse presence after MD (NE mean time present 5.95; N=838 inhibitory synapses;

MD mean time present 5.75; N=1273 inhibitory synapses; p=0.0058 Wilcoxon Rank-Sum test,

one tailed). Overall, these data show that rapid inhibitory synapse dynamics are responsive to

sensory manipulations and likely reflect circuit adaptations to the visual environment. Further,

disinhibition caused by MD, is not simply a result of one-time inhibitory synaptic loss, but rather

a destabilization of inhibitory synapses, resulting in a new dynamic state where they are more

dynamic and less likely to be present.

Inhibitory axons appear to remain in place after loss of a recurrent inhibitory synapse

The ability of inhibitory synapses on DiS to repeatedly return to the same location within

24-48 hours suggested that other elements might remain stable at the same locale to serve as a

landmark or placeholder. To explore this, we fixed brains for EM immediately following the last

in vivo imaging session. We specifically searched for DiS that had lost a recurrent inhibitory

64

synapse within 24-48 hours prior to fixation to determine if any synaptic structural elements

were visible by EM. Five inhibitory synapses on DiS that were present on multiple non-

consecutive imaging sessions, but absent on session 9 (three shown in Fig. 3.5D-E), had no

Figure 3.7. EM reconstruction of recurrent inhibitory synapses lost within 24-48 hours of fixation and their

presynaptic afferents. (A) 2-photon image from session 9 of the region in Fig. 3.5D. (B) 3D reconstructions of this

region. Left panel shows the spine ‘a’ which lacks an inhibitory synapse on session 9. Right panel shows the

presynaptic inhibitory axon (red) which makes a symmetric synapse (purple arrow) on a neighboring dendrite (sky

blue). (C) EM micrograph of spine ‘a’ (left) with presynaptic excitatory terminal in blue and adjacent inhibitory

terminal in red. Yellow arrowhead indicates probable location of the recently disappeared inhibitory synapse. Bold

white arrow indicates excitatory synapse. Right panel shows clear symmetric inhibitory synapse (purple arrow) onto

neighboring (sky blue) dendrite. (D) 2-photon image from session 9 of the region in Fig. 3.5E. (E) Reconstruction

of spines ‘j’ and ‘k’ which both lacked inhibitory synapses on the final imaging session. Note that spine ‘k’ is a z-

projecting spine. Putative inhibitory axon terminals in red. (F) EM micrograph of spine ‘k’, blue arrowhead indicates

where inhibitory synapse may have previously been located. (G) 3D reconstruction of another dendrite (Fig. 3.9E,

Movie M1), yellow and black arrows indicate the inhibitory synapses on DiS and shaft, respectively. Reconstructed

presynaptic inhibitory axon colored red. (H) A multisynaptic bouton (MSB) contacts both the DiS and the inhibitory

dendrite (yellow). (I) Left panel shows complete reconstruction of all 4 dendrites targeted by the MSB. Right panel

shows this region with red and yellow segments removed to visualize the synapses onto excitatory dendrites. (J)

EM micrograph with presynaptic excitatory axon in dark blue, DiS in black, presynaptic inhibitory axon (MSB) in

red, postsynaptic dendrites in light blue and purple. Bold white arrow indicates excitatory synapse (for more detail

see Fig. 3.12B). Scale Bars in µm: (A,B,D,E) 1.25; (C,F,J) 0.2; (G,H,I) 0.5.

65

discernable inhibitory synaptic junction visible by EM (Fig. 3.7, Fig. 3.13A, Table T2). However,

in four cases a putative inhibitory presynaptic terminal with a few vesicles was present adjacent

to an excitatory synapse on the same spines. Two of these axon terminals could be traced to

symmetric synapses formed onto a non-DAB stained cell confirming they were in fact inhibitory

afferents. (Fig. 3.7B-C,E-F).

For one of the recurrent inhibitory synapses, present on sessions 4-6 and not 9, a very

tiny synaptic contact was visible, although its size, 0.010 square microns on one section, would

not normally be scored positive by EM (Fig. 3.7G-J, Fig. 3.9E, Fig. 3.13B). This site also

retained its presynaptic terminal, a rather promiscuous multi-synaptic bouton, which contacted

the DiS of interest, and also formed large symmetric synapses on two other excitatory dendrites

and on one inhibitory dendrite. This bouton is particularly interesting because it shows that the

inhibitory axon is able to form synapses onto a diverse range of targets: onto the dendritic shaft

and the DiS of pyramidal neurons, but also onto inhibitory dendrites. This axon also formed a

large inhibitory shaft synapse next to the DiS, on the same dendrite. This large multisynaptic

bouton is highly unlikely to retract when the synapse is lost from the DiS, considering it is

anchored by three other synaptic connections. Taken together, these data suggest that loss of

gephyrin from inhibitory synapses on DiS, even for a short interval, represents loss of a synaptic

junction. Further, even after the postsynaptic gephyrin has disappeared, we have observed 4/5

cases where the putative inhibitory axon remains in place, potentially enabling the rapid

reappearance of Teal-gephyrin puncta and the reformation of the inhibitory synapse.

Discussion

The loss and gain of spines observed in vivo has long been assumed to represent the

remodeling of excitatory connections: specifically, a permanent loss or gain of contact with the

66

presynaptic cell. When a new spine is formed, if it lasts for at least 4 days, it will gain an

excitatory synapse and persist for months (Cane et al., 2014; Holtmaat et al., 2005). In

agreement with recent observations (Cane et al., 2014), we find that the majority of spine

dynamics are due to a short lived population of spines that never acquire PSD-95 puncta, and

thus, are unlikely to represent functional changes in circuitry. Our three-color imaging has also

enabled further distinction of spines with PSD-95 into SiS or DiS. We find that DiS are the

largest and most stable of all spine classes. Since spine size is well known to correlate with

stability (Prange and Murphy, 2001) (De Roo et al., 2008; Ehrlich et al., 2007) as is synaptic

presence (Cane et al., 2014), the increased stability of DiS could be a function of their large size

or the presence of two synapses, although these factors are likely related.

We have previously shown that inhibitory synapses on DiS are more dynamic than shaft

inhibitory synapses or spines (Chen et al., 2012). Here we show that they are also more

dynamic than excitatory synapses, and in particular when compared against the extreme

stability of excitatory synapses on DiS. Our short interval imaging protocol reveals that a large

proportion of dynamic inhibitory synapses, in particular inhibitory synapses on DiS, are transient

or recurrent. Even this large fraction may be an underestimate as events currently scored as

one-time dynamic could potentially resolve as transient or recurrent if we were able to image

longer than 9 daily sessions, or potentially when imaging at even shorter intervals. This

repeated insertion and removal of inhibitory synapses at stable sites is unlikely to serve the

purpose of rearranging circuitry in terms of testing or exchanging presynaptic partners, as has

been the model for excitatory synapse remodeling (Chen and Nedivi, 2010; Nimchinsky et al.,

2002). Inhibitory synaptic remodeling on DiS may serve a fundamentally different purpose. It

has been shown, in vivo, that GABA uncaging can reduce Ca2+ transients within spines (Chiu et

al., 2013). Inhibitory synapse removal and insertion on DiS could enable flexible, input specific

modulation of a stable excitatory synapse. Further, the size of individual inhibitory synapses can

vary up to 3 fold across imaging sessions (Figure 11B and C), suggesting that inhibitory spine

67

synapses can potentially fine tune their excitatory partner on a DiS beyond a yes/no switch.

Considering that DiS often receive direct excitatory input from thalamus (Kubota et al., 2007;

van Versendaal et al., 2012), feed forward connections are likely to be particularly influenced by

this previously unsuspected mode of inhibitory modulation.

Rapid inhibitory synapse dynamics are responsive to sensory manipulations. However,

rather than abrupt inhibitory synaptic loss followed by gradual recovery (Chen et al., 2011; Chen

et al., 2012; van Versendaal et al., 2012), we find that MD causes a destabilization of inhibitory

synapses, and transition to a new dynamic state where they are more dynamic and less likely to

be present. Inhibitory shaft synapses are recruited from the stable population into the dynamic

pool, both dynamic inhibitory synapse populations show shorter mean lifetimes, and inhibitory

synapses on DIS show longer time to recurrence. Disinhibition through inhibitory synaptic loss

after sensory deprivation is consistent with studies showing that whisker stimulation leads to an

increase in the number of inhibitory synapses on spines (Knott et al., 2002). Inhibitory synapse

dynamics are clearly sensitive to activity levels in the excitatory circuits they modulate.

EM reconstruction of sites previously occupied by recurrent gephyrin, suggests that

nearby inhibitory axons are well situated to serve as placeholders or nucleation sites for the

return of inhibitory synapses. This is consistent with cell culture studies showing that two days of

TTX treatment results in the loss of postsynaptic, but not presynaptic inhibitory synapses

(Kilman et al., 2002), and this loss of inhibition is completely and rapidly reversible (Rutherford

et al., 1997). This ability of inhibitory synapses to rapidly remodel in response to changes in

activity levels in cell culture (Kilman et al., 2002; Marty et al., 1997; Rutherford et al., 1997), and

in vivo (Chen et al., 2012; van Versendaal et al., 2012) is an important homeostatic mechanism

for maintaining a stable excitatory/inhibitory balance in cortical circuits (Nelson and Turrigiano,

2008; Turrigiano, 2011). Thus, the rapid addition and removal of inhibitory synapses at stable

sites that we show for the first time in vivo may serve to not only to reversibly regulate individual

68

synaptic contacts on DiS and dendritic excitability locally, but to more broadly regulate

homeostatic drive onto individual pyramidal neurons (Kilman et al., 2002; Turrigiano and

Nelson, 2004). Since these dynamics occur without complete dismantling of the inhibitory

afferent, recruitment of an inhibitory synapse may be guided by activity of the nearby excitatory

synapse.

It is interesting to note that for inhibitory synapses imaged in vivo, as well as in

organotypic slice culture there is a clear threshold for scoring Teal-gephyrin positive puncta that

corresponds to the emergence of GABA evoked uIPSCs. This suggests that GABA receptor

clustering by gephyrin may require a critical aggregation step that specifies emergence of an

inhibitory synapse.

69

Supplemental Data

Figure 3.8: PSD-95-mCherry and Teal-gephyrin expression does not alter spine density. (A)

Representative image of a cell expressing eYFP alone. (B) Representative image of cell expressing

eYFP, PSD-95-mCherry and Teal-gephyrin. (C) Quantification of spine density, PSD-95-mCherry and

Teal-gephyrin expression does not alter spine density. (D) Fraction of additions or eliminations of

dendritic spines (left), inhibitory synapses on DiS (middle), and inhibitory shaft synapses (right) at 2

day intervals, indicating that the dynamics of spines and inhibitory synapses are evenly spaced

throughout the imaging sessions. Error bars represent SEM.

70

Figure 3.9: Reconstructions and EM micrographs from additional branches. (A) Reconstruction

from branch ‘be’ (Supplemental Table 1). (B) Excitatory synapse ‘l’. (C) Inhibitory shaft synapse ‘6’.

(D) Nonconsecutive z-stacks of dually innervated spine ‘c4’. (E) Reconstruction of branch ‘kb’, an

additional branch from the same cell as branch ‘ka’ (Figure 2). (F) Excitatory synapse ‘r’. (G)

Inhibitory shaft synapse ‘12’. (H) Previously unpublished micrographs of an excitatory synapse and

inhibitory synapse from branch ‘jc’, originally published in (Chen et al., 2012). For this branch, post-

hoc anti-GABA colloidal gold immunolabeling was performed to label inhibitory presynaptic

terminals (visualized as dark dots in the presynaptic inhibitory axon, red arrow). Scale Bars in µm:

(A,E) 2; (B-D, F,G) 0.2; (H) 0.5

71

Figure 3.10: Size distributions of imaged teal-gephyrin puncta. (A) Relative teal-

gephyrin fluorescence levels for SiS and DiS spines imaged in vivo show a clear distinction between

these populations. Fluorescent levels of each spine head were normalized against background levels

on each branch shaft for both teal-gephyrin and the YFP cell fill. Normalized teal-gephyrin was then

divided by the normalized YFP signal to control for spine size. Data are from the first imaging session

of 3 WT cells. SIS=259 spines (empty circles), DIS=71 spines (filled circles), 38 stable (magenta), 33

were later dynamic (cyan), and of those 14 were recurrent (yellow). (B) Cumulative probability

distribution showing normalized teal-gephyrin fluorescence levels of puncta on DIS in vivo (magenta)

and in organotypic slices (cyan). Synapse size distributions are not significantly different in the two

preparations. p=.12 by KS-Test. In vivo = 71 inhibitory synapses. Organotypic=42 inhibitory synapses.

72

Figure 3.11: Synapse dynamics are stochastic and the size of inhibitory synapses vary over

time. (A) Percentage of dynamic events when comparing: (a) session 1 to session 2 (daily change

with one day imaging), (b) session 1 to session 5 (as if imaging had been done once every four days),

(c) cumulative changes from session 1 to session 5 under conditions of daily imaging. (d) ‘b’ divided

by 4 to calculate daily dynamics under conditions of 4 day imaging. (e) ‘c’ divided by 4 to calculate

daily dynamics in conditions of daily imaging. ‘d’ and ‘e’ are represented in Figure 5b. (B) Inhibitory

synapses on DiS that went away and returned again change size over time. n=8 synapses from 2

cells. (C) Stable inhibitory synapses on DiS that were present in every imaging session still fluctuate

in size over time. n=11 synapses from 2 cells. Synapse size was calculated by summing the intensity

of a teal-gephyrin puncta in a 3x3 pixel ROI from the Z frame where the signal was brightest,

normalizing to background shaft gephyrin signal and then to normalized YFP signal to control for

spine size.

73

Figure 3.12: Spine dynamics are not altered by MD. Comparing the percentage of dynamic SiS

and spines without PSD-95 between the MD and NE reveals no significant difference in spine

dynamics. DiS are not dynamic under NE (Fig. 3.4D) or MD conditions. P>.05 by two-tailed student’s

t-test.

74

Supplemental Figure 3.13: Series of consecutive EM micrographs for DiSs showing loss of

inhibitory synaptic contact. (A) EM micrograph from Figure 7C as a series of consecutive z stacks

with 12 nm step between stacks. (B) EM micrograph from Figure 7J as a series of consecutive z

stacks with 12 nm step between stacks.

75

76

Table T1: EM validation of all gephyrin puncta seen in last 2 photon imaging session.

Reconstructed dendrites are from 5 different branches from 3 independent cells and 2 different

animals, and encompass a total length of 90 µm. Branch ‘jc’ from (Chen et al., 2012). Branch ‘ka’ and

‘kb’ are from the same cell (Fig 2C, Fig 9E). Branch ‘be’ is 2 adjacent branches from a third cell (Fig

9A). All synapses visualized on the last session of two-photon imaging were designated positive by

EM if they had a clear synaptic cleft apposed by clustered vesicles (third column from left). Given the

difficulty of scoring excitatory vs inhibitory junctions in DAB stained tissue, especially using FIB/SEM,

confirmation of 27/28 teal-gephyrin positive synapses as inhibitory was done using other criteria listed

in the right-most column. Of 18 PSD95-mCherry puncta visualized by two-photon imaging, 100%

were found by EM; however 5 additional excitatory synapses were found by EM which were not

visualized through 2 photon imaging.

77

Table T2: EM tracking of recently disappeared teal-gephyrin puncta. Post hoc EM of 5 spines

that lost a teal-gephyrin puncta within 24-48 hours of the last imaging session also shows loss of

inhibitory synapse structure (no cleft with apposed vesicle clusters). In 4/5 cases an axon was found

adjacent to the spine, which in 2/4 cases could be traced back to a neighboring symmetric synapse,

thus confirming that in many cases an inhibitory axon could be identified adjacent to a spine that had

a recently disappeared teal-gephyrin puncta.

78

Experimental Procedures

Generation of Expression Plasmids

The Cre-dependent eYFP and Teal-gephyrin plasmids (pFUdioeYFPW, pFUdioTealGephyrinW)

have been described previously in (Chen et al., 2012), and the Cre plasmid in (Subramanian et

al., 2013). To generate a Cre-dependent PSD-95-mCherry within a dio cassette, PSD-95

lacking a stop codon was first amplified with an added 5’Nhe1 site from pFuPSD-95-TealW (gift

from Jerry Chen), then cloned into pcDNA3 (Invitrogen) between the KpnI and EcoRI sites, to

create pcDNA3-PSD-95. mCherry was amplified from pcDNA3.3-mCherry (Addgene) with an

added 3’AgeI site, then cloned into pcDNA3-PSD-95 between the EcoRI and XhoI sites to make

pcDNA3-PSD-95-mCherry. PSD-95-mCherry was then removed by NheI and AgeI digestion

and subcloned into the Cre dependent plasmid pFudio-AscI-AgeI-NheIW, generated by

replacing eYFP in pFudio-eYFPW with a linker sequence containing the AgeI restriction site.

In Utero Electroporation

All animal experiments were approved by the Massachusetts Institute of Technology Committee

on Animal Care and meet the National Institute of Health guidelines for the use and care of

vertebrate animals. In utero electroporation on E15.5 timed pregnant C57BL/6J mice was

performed to label L2/3 cortical pyramidal neurons, as previously described (Tabata and

Nakajima, 2001). Animals were co-electroporated with Cre-dependent constructs expressing

eYFP, PSD-95-mCherry, or Teal-gephyrin along with a plasmid expressing Cre recombinase at

a ratio of 10:10:5:1, respectively (total DNA concentration 2 µg/µL), with 0.1% Fast Green for

visualization. A total of 0.75 µl of the plasmid solution was injected into the right lateral ventricle

with a 32 gauge Hamilton syringe (Hamilton Company, Reno, NV, USA). Five pulses of 36V

(duration 50 ms, frequency 1 Hz) targeting the visual cortex were delivered from a square wave

79

electroporator (ECM830; Harvard Apparatus, Holliston, MA, USA) using 5 mm diameter

platinum electrodes (Protech International, Boerne, TX, USA).

Cranial Window Implantation

After in utero electroporation, pups were reared to adulthood (P42-57) and implanted with a 5

mm cranial window over the right hemisphere, as described (Lee et al., 2008).

Sulfamethoxazole (1 mg/ml) and trimethoprim (0.2 mg/ml) were chronically administered in the

drinking water to maintain optical clarity of implanted windows.

Optical Intrinsic Signal Imaging

To identify binocular visual cortex, optical imaging of intrinsic signal and data analysis were

performed as described previously (Kalatsky and Stryker, 2003). Mice were anesthetized and

maintained on 0.25%-0.75% isofluorane and secured in a stereotaxic frame. A horizontal bar (5o

in height and 73o in width) drifting upward with a periodicity of 12 s was presented for 60 cycles

on a high refresh rate monitor 25 cm in front of the animal. Optical images of visual cortex were

acquired continuously under 610 nm illumination with an intrinsic imaging system (LongDaq

Imager 3001/C; Optical Imaging Inc., New York, NY, USA) through a 2.5x/0.075 NA objective

(Zeiss, Jena, Germany). Images were spatially binned by 4x4 pixels for analysis and cortical

intrinsic signal was computed by extracting the Fourier component of light reflectance changes

matched to stimulus frequency. Response magnitude was the fractional change in reflectance,

and the magnitude maps were thresholded at 30% of peak response amplitude. Binocular visual

cortex was delineated upon stimulation of the ipsilateral eye only.

80

Two-photon imaging

Animals were allowed 2 weeks for recovery after cranial window surgery. When windows

cleared, labeled cells in binocular visual cortex were screened for the presence of all 3

fluorescent labels. Cells that exhibited high background labeling with synaptic labels or any

ectopic clumps of synaptic proteins were not used for experiments. This selection process

ensured that all imaged cells had similar levels of fluorescent labeling. Only one cell was

imaged per animal. For eight mice with normal visual experience in vivo two-photon imaging

was performed daily for 9 consecutive sessions using a custom-built microscope with custom

acquisition software to enable triple color imaging. For each imaging session, mice were

anesthesized with isofluorane (0.75%-1.25%) and secured in a stereotaxic frame. Fluorophores

were simultaneously excited with a commercial Mai Tai HP Ti: Sapphire laser (Spectra-Physics,

Santa Clara, CA, USA) at 915 nm to excite eYFP and Teal, and a Chameleon Compact OPO

(Coherent, Santa Clara, CA, USA) at 1085 nm to excite mCherry. The outputs of the two

excitation lasers were combined at a polarized beam splitter with orthogonal polarization

directions, and delivered to the two-photon microscope through the same beam path. After

scanning with a galvonometric XY-scanning mirrors (6215H, Cambridge Technology) and a

piezo actuator Z-positioning system (Piezosystem Jena, Jena, Germany), the two laser beams

were focused by a 20x/1.0 NA water immersion objective lens (W Plan-Apochromat; Zeiss,

Jena, Germany) to the same focal volume location in the specimen, within one pixel size

accuracy. These lasers produce ~100 fs unsynchronized pulses at a rate of 80 MHz. The power

delivered by each laser to the specimen ranged from approximately 35-50 mW depending on

imaging depth. The emission signals from the three fluorophores were collected by the same

objective lens, passed through an IR blocking filter (E700SP; Chroma Technology, Bellows

Falls, VT, USA), and were separated according to their emission spectra by dichroic mirrors at

520 nm and 560 nm. After passing through three independent bandpass filters (485/70m,

81

550/100m, and 605/75m) the three emission signals were collected simultaneously onto 3

separate PMTs. Imaging for synapse monitoring was performed at high resolution (250 nm/pixel

XY-resolution, 0.9 µm/frame Z-resolution). Two-photon raw scanner data was processed for

spectral linear unmixing and converted into a RGB image z-stack using Matlab and ImageJ

(National Institutes of Health, Bethesda, MD, USA).

Monocular Deprivation

Monocular deprivation was performed by eyelid suture immediately after the first imaging

session. Mice were anesthetized with 2% isoflurane, lid margins were trimmed and triple

antibiotic ophthalmic ointment (Bausch & Lomb, Rochester, NY, USA) was applied to the eye.

Four to five individual stitches were placed using 6-0 vicryl along the extent of the trimmed lids.

Daily imaging sessions were performed for six additional sessions after MD. Suture integrity

was inspected directly prior to each imaging session. Animals whose eyelids did not seal fully or

had reopened were excluded from further experiments. A total of seven mice with successful

MD and no photobleaching were used for analysis.

Spectral Linear Unmixing and Image Processing

Spectral linear unmixing for three colors was performed using a similar approach as previously

described for two color unmixing (Chen et al., 2012). Briefly, spectral linear unmixing is based

on the fact that the total photon count at each pixel in a given channel is the linear sum of the

spectral contribution of each fluorophore weighted by its abundance. For a triple channel

detection system, the contribution of three fluorophores can be represented by the following

equations.

82

J1(x,y)=s1.1 x I1 (x,y) + s1.2 x I2 (x,y) + s1.3 x I3 (x,y)

J2(x,y)=s2.1 x I1 (x,y) + s2.2 x I2 (x,y) + s2.3 x I3 (x,y)

J3(x,y)=s3.1 x I1 (x,y) + s3.2 x I2 (x,y) + s3.3 x I3 (x,y)

Where J is the total signal per channel, I is the fluorophore abundance, and S is the contribution

of that fluorophore. These equations can be expressed as a matrix:

[J]=[S][I],

whereby the unmixed image [I] can be calculated using the inverse matrix of S:

[I]=[S]-1 [J].

Assuming the detected signal in both channels represents the total spectral contribution for all

three fluorophores:

s1.1 + s2.1 + s3.1 = 1

s1.2 + s2.2 + s3.2 = 1

s1.3 + s2.3 + s3.3 = 1

[S] was determined experimentally from two-photon images of HEK cell cultures expressing

single fluorophores and excited by both lasers at the same wavelengths used in vivo. Average

laser power was adjusted to achieve photon count levels approximating in vivo signal intensity.

We have previously shown that S derived from cell culture and in vivo data is interchangeable

(Chen et al., 2012). The mean contribution for each fluorophore into each channel (s1.1 – 3.3) was

calculated using Matlab (Mathworks, Natick, MA, USA). These values were subsequently used

for spectral linear unmixing of triple channel two-photon raw scanner data into a RBG image z-

stack using Matlab and ImageJ (National Institutes of Health). For all figures, images are

83

filtered and interpolated for optimal visualization. 3D image stacks are converted into maximum

intensity z-projections where noted in figure legends.

Data Analysis

Dendritic spines, PSD-95 containing excitatory synapses, and inhibitory synapses were scored

manually with a custom-written 4D point tracking system implemented in Fiji (Schindelin et al.,

2012) using a modified version of the ObjectJ plugin

(https://sils.fnwi.uva.nl/bcb/objectj/index.html). To avoid individual scoring bias, each cell was

independently scored by two investigators. Dendritic spine analysis criteria was defined as

previously described (Chen et al., 2012; Holtmaat and Svoboda, 2009). Because of the bright

labeling on the soma and axon initial segment, individual contacts in these regions could not be

resolved and analysis was restricted to dendrites starting approximately 40 microns from the

soma to the most distal tips. Because z-projecting spines lacking PSD-95 could not be

visualized, for consistency, all z-projecting spines were excluded from analysis, even those

which contained PSD-95. Gephyrin puncta were scored as synapses if they were at least 3x3

pixels, or 8–9 clustered pixels (0.56 µm2) in size, with a minimal average signal intensity of at

least four times above shot noise background levels, and were present in 2 consecutive z

planes. PSD-95 puncta were scored as synapses if they were least 2x2 pixels, or 4-5 clustered

pixels (0.27 µm2) in size with a minimal average signal intensity of at least four times above shot

noise background levels and were present in 2 consecutive z planes. EM validation confirmed

that these criteria represent inhibitory and excitatory synapses, respectively, with no false

positives. In general, the PSD-95-mCherry fluorophore was more prone to bleaching than YFP

or Teal-gephyrin. Cells that lost PSD-95-mCherry were not analyzed further. One-time dynamic

changes were defined as any structure that changed one time and never changed again.

84

Changes were scored as transient if a structure was added and later removed in a subsequent

imaging session. Changes were scored as recurrent if structures appeared, then disappeared,

and then appeared again at the same site; or conversely, ever disappeared and then

reappeared more than once in the same location. At least 25 spines and synapses were

counted per branch and at least 200 structures were counted per animal. Dendritic arbors were

manually traced in Neurolucida (MicroBrightField, Inc., Williston, VT, USA) to quantify the length

of scored branches.

In the normal experience dataset, we tracked a total of 1555 spines and 955 inhibitory

synapses on 62 dendritic segments from 8 animals. These included 30 basal, and 32 apical

dendrites, with a combined branch length of 3.01 mm, 1.63mm basal and 1.42mm apical.

Analysis for 1 animal was restricted to sessions 1-7 and another to sessions 1-6 due to photo-

bleaching of PSD-95-mCherry in later sessions. The animal with 6 sessions was not included in

the comparisons with the MD cells.

In the MD dataset, 1500 spines and 1273 inhibitory synapses were tracked over 7

consecutive imaging sessions. In all, 43 dendritic segments from 7 animals, 22 basal and 21

apical dendrites, with a combined branch length of 2.83mm, 1.54mm basal and 1.29mm apical.

All comparisons between MD and Normal Experience datasets were done using only the first 7

imaging sessions of all cells.

Spine size calculations were performed by placing a 5x5 pixel box in the most prominent

z frame around the spine head, subtracting background fluorescent levels and normalizing to

the average dendritic shaft intensity, as performed in (Holtmaat et al., 2005). Cumulative

probability distributions were calculated in Matlab and significance was determined by a two-

sample Kolmogorov-Smirnov test.

85

Focused Ion Beam/ Scanning Electron Microscopy

To demonstrate that Teal-gephyrin and PSD-95-mCherry puncta visualized in vivo correspond

to synapses, we performed serial section immuno-EM on in vivo imaged L2/3 pyramidal neuron

dendrites labeled with eYFP, Teal-gephyrin, and PSD-95-mCherry. Immediately after two-

photon imaging, the brain was fixed with 4% paraformaldehyde, 1% glutaraldehyde and 0.2 %

picric acid in 0.1 M PB, cut to 50 µm sections, and stained with an antiserum to GFP (1:2000

dilution, rabbit antiserum, gift from Dr. Tamamaki, Kyoto University) followed by a biotin-

conjugated secondary and then detected with nickel diaminobenzidine (Ni-DAB). Tissue was

then flat-embedded in Epon. Four dendritic segments with identifiable DAB staining were

relocated and then further investigated using a combined FIB/SEM (Focused Ion Beam /

Scanning Electron Microscope)(Kubota et al., 2011a). For EM observation, the Epon block

containing the dendritic segments was glued to a stainless-steel sample holder using silver

paste to avoid charging the epoxy. The top surface was coated by several 10 nm thick layers of

iridium using a sputter coater. The mounted block was transferred in a FIB/SEM, (Hitachi

MI4000L, Tokyo, Japan) which contains two beams that intersect at a right angle.The dendritic

segments were identified by SEM imaging on the top surface of the block using guidance lines

intersecting at right angles that were etched shallowly using the FIB, and by comparing this to

light microscopy images obtained earlier. The top region was protected by ion beam induced

deposition of platinum (52um*20um area, 1nA ion beam current, 900 sec deposition time). After

the coarse milling process by the FIB, the freshly exposed surface of the block was imaged at 1

kV acceleration potential and 1nA beam current using the in-lens detector with 10 µsec dwell

time/pixel. High resolution imaging was achieved by low kV imaging and detection of the

secondary electrons of the stained tissue surface. Using contrast inversion, TEM-like contrast

and comparable imaging information was obtained. Using the ‘Multi-Cut & See’ function, seven

adjacent images of 2000 x 2000 store resolution were acquired serially after milling the block

86

surface at 12 nm z-steps. Between 1183 and 1641 serial section images were acquired for each

image stack. The milling time was 26 seconds/slice and, the imaging time was about 40

sec/image, or 280 seconds for all 7 images. In total, the image acquisition took 7 days. The

serial image alignment was done using a homemade script for Matlab (kindly provided by Dr.

Shawn Mikula, Max Planck Institute, Heidelberg, Germany). The dendritic and synaptic

structures were rendered using the 3D reconstruction software, Reconstruct(Fiala, 2005)

(available at http://synapses.clm.utexas.edu/tools/index.stm). Synapses were scored according

to 3 criteria, the presence of a postsynaptic density, the aggregation of small synaptic vesicles

within the presynaptic terminal, and a clear synaptic cleft structure between the pre and

postsynaptic membranes at a distance of approximately 20 nm between two parallel

membranes. Contacts with DAB staining obscuring the postsynaptic compartment were only

categorized as inhibitory synapses if they were validated though 1) location on the dendritic

shaft, 2) backtracing the axon to another symmetric synapse formed onto a cell lacking DAB

staining, or 3) if two terminals were clearly visible on a DiS and the excitatory terminal was

validated by other criteria.

Preparation of organotypic slice cultures and DNA transfection

Organotypic slice cultures from mouse visual cortex were prepared from P3-P4 C57BL/6 mice

(Stoppini et al., 1991), and transfected 3-4 days before imaging/uncaging experiments using

biolistic gene transfer (180 psi)(Woods and Zito, 2008). The same Cre-dependent Teal-

gephyrin, and Cre plasmids used for the in vivo studies were coated onto 6-7 mg of gold

particles together with a tdTomato plasmid (Kwon et al., 2012) for cell fill (12 µg of tdTomato, 18

µg of Teal-gephyrin, and 16 µg of Cre).

87

Two-photon slice imaging and GABA uncaging

Imaging and uncaging were performed at 19-22 days in vitro (DIV) on transfected layer 2/3

pyramidal neurons within 40 µm of the slice surface at room temperature in recirculating artificial

cerebrospinal fluid (ACSF; in mM: 127 NaCl, 25 NaHCO3, 1.25 NaH2PO4, 2.5 KCl, 25 D-

glucose, aerated with 95% O2 /5 % CO2) in the presence of 2 mM CaCl2, 1 mM MgCl2, 1 mM

CDNI-GABA, and 0.001 mM TTX. For each neuron, image stacks (512 x 512 pixels; 0.035 µm /

pixel) with 1 µm z-steps were collected from one segment of secondary or tertiary apical

dendrites 30-50 µm from the soma using a two-photon microscope (Prairie Technologies, Inc)

with a pulsed Ti::sapphire laser (Mai Tai DeepSee, Spectra Physics) tuned to 930 nm (2-2.5

mW at the sample). To record uncaging-evoked inhibitory postsynaptic currents (uIPSCs), layer

2/3 pyramidal neurons were patched in voltage-clamp configuration (electrode resistances 5-8

MΩ, Vhold = +10 mV) using cesium-based internal solution (in mM: 135 Cs-methanesulfonate,

10 HEPES, 10 Na2 phosphocreatine, 4 MgCl2, 4 Na2-ATP, 0.4 Na-GTP, 3 Na L-ascorbate,

0.02 Alexa 594, ~300 mOsm, ~pH 7.25) in ACSF. For GABA uncaging, 720 nm light was

delivered 0.5 µm away from the target spine with a power of 18~20 mW for 3 ms. uIPSC

amplitudes from individual spines were quantified as the average of 8-10 pulses at 0.15 Hz.

Teal-gephyrin expression level in individual spines was measured from background-

subtracted and bleed-through-corrected green fluorescence intensities using the integrated pixel

intensity of a boxed region of interest (ROI) surrounding the spine head, as described (Woods et

al., 2011). In brief, relative Teal-gephyrin enrichment in spines was calculated by normalizing

the green fluorescence intensities (as described above) for each individual spine to the mean

green fluorescence intensities measured from four ROIs on the dendritic shaft: “spine with

gephyrin” (expression level > mean ) vs. “spine lacking gephyrin” (expression level < mean).

88

Chapter 4 - Probing the molecular mechanism of rapid

inhibitory dynamics

Contribution: All experiments and figures by Katherine Villa.

Abstract

We have previously shown that inhibitory synapses are remarkably dynamic in vivo and

are capable of disappearing and reappearing at stable sites on a timescale of days. The

mechanism for rapid synapse assembly and disassembly is likely to depend on postsynaptic

signaling cascades that regulate aggregation of the inhibitory post-synaptic scaffolding protein

gephrin required for synaptic clustering of inhibitory receptors. Activation of these postsynaptic

signaling cascades is potentially triggered by changes in membrane potential, influenced by

local synaptic activity. To examine the molecular mechanisms of rapid gephyrin dynamics

observed in vivo, I have created serine to alanine mutations in residues 268 and 270 of gephyrin

(S268A, S270A) by site-directed mutagenesis and assessed their effect on inhibitory synapse

dynamics. The S268 and S270 sites are phosphorylated by intracellular signaling cascades and

degraded by the calcium-activated protease, calpain, which may indicate that inhibitory synaptic

removal is mediated by neuronal activity. Preliminary results indicate that gephyrin S270A

mutants may display higher density of inhibitory synapses on dually innervated spines and more

recurrent dynamics of inhibitory shaft synapses. This could indicate that the phosphorylation at

residue 270 of gephyrin inhibits the rapid re-formation of inhibitory synapses.

89

Introduction

Inhibitory GABA-A receptors are clustered at inhibitory synapses in the cortex by the

postsynaptic protein gephyrin, a protein which self-assembles into a scaffold and interacts with

the cytoskeleton (Tyagarajan and Fritschy, 2014). Interestingly, gephyrin likely evolved from the

molybdenum cofactor (MoCo) synthesizing enzyme in bacteria, and the mammalian gephyrin

protein still possesses MoCo synthesis functions. Mice deficient in gephyrin die early postnatally

and display a loss of GABA-A receptor subtypes from postsynaptic sites. Restoration of MoCo

synthesis in gephyrin knockout mice does not rescue lethality, indicating that it is gephyrin

clustering of GABA receptors (and glycine receptors outside the cortex) which is essential for

life (Grosskreutz et al., 2003).

Gephyrin is post-translationally modified at several residues, as identified by mass-

spectroscopy (Schwarz et al., 2001; Tyagarajan et al., 2011; Zita et al., 2007). The

phosphorylation of gephyrin plays an important role in the regulation of its localization, stability,

and interactions with other proteins (Tyagarajan et al., 2013; Wuchter et al., 2012). Gephyrin

serine residues 270 and 268 are phosphorylated by glycogen synthase kinase 3β (GSK-3β) and

extracellular signal-regulated kinases 1 and 2 (ERK1/2), respectively.

Both GSK-3β and ERK1/2 have been shown to play important roles at synapses. GSK-

3β, which is abundantly expressed in neurons, is inhibited by phosphorylation at serine 9, which

is performed by many regulatory proteins, including PI3K-Akt, protein kinase A (PKA), and

protein kinase C, among others. This inhibition of GSK-3β occurs during the induction of LTD

(Hooper et al., 2007), and in dendrites during synaptogenesis and leads to dendritic growth

during development (Rui et al., 2013). Dysregulation of GSK3 is linked neurological conditions

such as Alzheimer's disease and other forms of neurodegeneration, and over-activation of GSK-

3β leads to neuronal apoptosis (Kaytor and Orr, 2002). ERK1/2 is activated by neurotrophins,

neuronal activity, BDNF, or cAMP and plays roles in the differentiation, survival, and adaptive

90

responses of neurons in vivo (Cavanaugh et al., 2001). BDNF activation causes an increase in

dendritic spine density which is mediated through the activation of ERK1/2 (Alonso et al., 2004).

The phosphorylation of gephyrin at S268 is therefore likely to increase upon neuronal

activity (through the activation of BDNF and ERK1/2), while phosphorylation of S270 is likely to

be inhibited by neuronal activity, as GSK-3β is inhibited by activity. Once gephyrin is

phosphorylated at either of these residues, it is more likely to be degraded by calpain, a

calcium-activated protein kinase (Tyagarajan et al., 2011). The degradation of gephyrin after

phosphorylation reduces amplitude and frequency of miniature GABAergic postsynaptic

currents (Tyagarajan et al., 2011).

Phosphorylation at the S268 and S270 sites has different effects on gephyrin clustering

in neuronal cell culture. Expression of gephyrin with S270 mutated to alanine (S270A), making it

phosphorylation incompetent at this site, increases the density of gephyrin clusters in cultured

neurons (Tyagarajan et al., 2011). When S270 is mutated to glutamic acid (S270E), which acts

as a phospho-mimic, there is a decrease in the density of gephyrin clusters.

Similar experiments with a S268A mutation results in a higher density of gephyrin

clusters that are larger in size than wild type gephyrin (Tyagarajan et al., 2013). While in the

presence of endogenous gephyrin the S268E mutation showed no significant effect on the

density or size of gephyrin clusters, when endogenous gephyrin was knocked down in

combination with expression of the mutants, the phenotypes of the S268A and S268E were

altered. Through experiments combining the S268 and S70 mutants, it was determined that the

two phosphorylation sites interact with each other and that the S268 location determines

gephyrin cluster size and the S270 location determines cluster density. Dephosphorylation at

91

S268 leads to larger size but lower density, whereas dephosphorylation at S270 leads to

smaller size and increased density in culture.

To determine if phosphorylation of gephyrin at the S268 and S270 sites regulates the

density and size of gephyrin clusters in vivo, and whether this plays a role in recurrent gephyrin

dynamics, I will mutate these sites from serines to alanines, rendering them phospho-

incompetent, and abolishing the degradation of gephyrin by calpain. We hypothesize that

S268A and S270A mutants would be more stable in vivo, thus preventing rapid inhibitory

synapse assembly and disassembly.

Results

Generation of gephyrin S268A and S270A expression plasmids

The gephyrin coding cDNA was PCR amplied from pFUdioTealGephyrinW (Chen et al.,

2012) and inserted into the pENTRY vector (ThermoFisher). The gephyrin cDNA was seperately

mutated to create S268A or S270A mutants by PCR mutagenesis using the GeneArt Site-

directed mutagenesis system (ThermoFisher), and sequenced to confirm correct mutagenesis.

The two mutated teal-gephyrin in the pENTRY vectors were cloned back into the

pFUdioTealGephyrinW vector using the Asc1 and Nhe1 restriction sites.

In vivo 2-photon imaging of L2/3 pyramidal cells expressing mutant gephyrin

E16.5 mice were in utero electroporated with three ‘double-floxed’ inverted open reading

frame (dio) plasmids, expressing either eYFP as a cell fill, PSD-95-mCherry as a postsynaptic

marker of excitatory synapses, or teal-gephyrin S268A or S270A as a mutant marker of

inhibitory synapses, with limiting amounts of a Cre plasmid.

92

In vivo triple color imaging of eYFP-labeled neuronal morphology, PSD-95-mCherry

puncta, and Teal-gephyrin puncta in individual L2/3 neurons within adult V1 was performed

through a cranial window with a custom built two-photon microscope. All three fluorophores

were simultaneously excited using two excitation wavelengths and their emissions spectrally

separated and collected with three PMTs. Bleed-through photons, due to overlapping emission

spectra, were reassigned to their appropriate color channel by post-hoc spectral linear un-

mixing. To monitor rapid synaptic dynamics, neurons were imaged 9 times at 24-hour intervals

across a 200x200x200 µm volume capturing a large portion of the dendritic arbor at 250nm/pixel

XY and 0.9µm/frame Z-resolution. Triple-labeled neurons showed clear arbor morphology with

distinct resolution of dendritic spines, Teal-gephyrin puncta, and PSD-95-mCherry puncta.

Density of spines and synapses of gephyrin mutants

. Two gephyrin S268A and two gephyrin S270A cells have been analyzed so far. What is

presented here is preliminary data which does not lend itself to drawing strong conclusions

Figure 4.1. Density of spines and inhibitory synapses of wild type and gephyrin mutant cells. (A) Quantification of

density of synaptic structures. n= 7 cells WT, 2 cells Gephyrin S268A, 2 cells Gephyrin S270A. * = p<0.05, **

p<0.01 by ANOVA. (B) Representative images of dendritic branches. YFP cell fill pseudo colored red, teal-gephyrin

WT and mutant constructs in cyan. Scale bar = 2.5 um.

93

about the nature of these mutations in vivo. We have found that S270A mutants have a higher

density of dually innervated spines compared to wild type gephyrin cells (Fig. 4.1A). This

matches with the phenotype of gephyrin S270A mutants seen in vitro. Of the two gephyrin

S268A cells analyzed, one had a very low density of spines (most of which contained inhibitory

synapses) and the other had a spine density similar to wild type. It may be possible that

gephyrin S268A mutants will show a lower spine density than WT cells, but analysis on more

cells is needed. From in vitro data, we would expect the S268A mutant to exhibit a higher

density of inhibitory synapses than in cells only expressing wild type gephyrin. Gephyrin S270A

cells display a trend of having a higher density of spines, which may become significant when

more data is analyzed. The reason that we expected gephyrin S268A and S270A to have higher

densities of inhibitory synapses is because these mutations make it impossible for the mutant

gephyrin protein to be phosphorylated and then degraded by calpain, potentially reducing the

instances of rapid inhibitory synapse disassembly. The possible trend of S268A mutants to have

lower spine density, and S270A mutants to have a higher spine density is an unexpected result

and could be indicative of the relationship between gephyrin dynamics and spine plasticity,

though again, we need to analyze more cells before we can draw significant conclusions.

Dynamics of spines and synapses in gephyrin mutants

Based on the in vitro results and proposed mechanism of gephyrin phosphorylation and

degradation, we hypothesized that gephyrin S268A and S270A mutants would be more stable

than wild type gephyrin. However, this is not what we have observed thus far. There is a

statistically significant increase in the dynamics of inhibitory shaft synapses in gephyrin S270A

mutants, from 19 ± 2% in wildtype to 36 ± 3% in S270A (p<0.05 by ANOVA). The gephyrin

S268A mutant is trending to have lower dynamics of inhibitory synapses on dually innervated

spines, but this is not yet statistically significant with n=2 cells.

94

We further classified the dynamic inhibitory synapses as one time dynamic, transient, or

recurrent (Fig. 4.2). There was no statistically significant difference in the percent of dynamic

synapses that were in each of these three categories for inhibitory synapses that are located on

dually innervated spines, but again, there was a large variance between the two gephyrin

S268A mutant cells, which may show a difference when more data is added. We did, however,

find that the percentage of recurrent inhibitory shaft synapses increased from 23 ± 3% in wild

type to 58 ± 3% in S270A mutants. Again, this was unexpected because we expected more

stability of inhibitory synapses, both on the shaft and on spines, in the two gephyrin mutants.

This result seems to indicate that the phosphorylation of gephyrin does not affect stable

clusters, but does influence how quickly gephyrin monomers can form clusters. When gephyrin

is dephosphorylated at S270, it is able to reform clusters more quickly than wild type gephyrin.

Figure 4.2. Inhibitory shaft synapses are more dynamic in gephyrin S270A mutants. (A) There is an

increase in the % of inhibitory shaft synapses that are dynamic in the gephyrin S270A mutant (p<0.05

by ANOVA), but no differences in gephyrin S268A mutants. (B) No statistically significant differences

in the % of dynamic inhibitory synapses that are one time dynamic, transient, or recurrent in the

gephyrin mutants. (C) There is a statistically significant increase in the percentage of recurrent shaft

inhibitory synapses in the gephyrin S270A mutant (p<0.05 by ANOVA).

95

Discussion

We have found that by mutating the serine 270 residue to alanine, and making a

phospho-incompetent gephyrin, there is an increased density of gephyrin clusters on dually

innervated spines as well as an increase in recurrent dynamics of inhibitory synapses on the

dendritic shaft. We still need to acquire and analyze more data on the S268A mutants to reach

conclusions regarding the role of phosphorylation at this residue in vivo.

It is possible that the dephosphorylation of gephyrin at S270 has no effect on stable

gephyrin clusters but does influence how quickly gephyrin monomers are able to assemble into

clusters to form inhibitory synapses. A major caveat to these results is that the mutant gephyrin

constructs are expressed in cells that still express unlabeled endogenous wild type gephyrin

protein. This could be obscuring the phenotype of the gephyrin mutants and making the

changes in their dynamics less pronounced. One way to overcome the effect of endogenous

wild type gephyrin would be to co-electoporate our constructs with a construct containing an

shRNA against wild type gephyrin.

It also remains to be tested whether the increased density of gephyrin clusters seen in

vivo represent functional synapses. This would need to be tested by localized GABA uncaging

and electrophysiological recordings in slice culture.

96

Methods

Generation of Expression Plasmids

The Cre-dependent eYFP, Teal-gephyrin, and PSD-95-mCherry plasmids (pFUdioeYFPW,

pFUdioTealGephyrinW, pFUdioPSD-95-mCherryW) have been described previously in (Chen et

al., 2012), and the Cre plasmid in (Subramanian et al., 2013). To generate Cre-dependent

Teal-gephyrin plasmids with the S268A and S270A mutations within a dio cassette, the gephyrin

cassette of pFUdioTealGephyrinW was PCR amplified and ligated into a pEntry/d-Topo vector

(ThermoFisher Scientific). The primers for this reaction were (5’ –CACCGGCGCGCCGTACG-

3’) and (5’- AAAGCTAGCCCGCCACCATG-3’) to insert the Entry sequence and preserve the

Asc1 and Nhe1 restriction sites for later cloning. The p-EntryTealGephyrin vector was

mutagenized by PCR mutagenesis using the following primers: S268A sense (5’-

TCCCTTAGCACTACTCCTGCCGAGTCGCCCCGTGCCCAGGCTACA- 3’),

Antisense (5’- TGTAGCCTGGGCACGGGGCGACTCGGCAGGAGTAGTGCTAAGGGA -3’),

S270A sense (5’- TCCCTTAGCACTACTCCTTCAGAGGCCCCCCGTGCCCAGGCTACA -3’),

S270A antisense (5’- TGTAGCCTGGGCACGGGGGGCCTCTGAAGGAGTAGTGCTAAGGGA -

3’). After sequencing to confirm successful mutagenesis and no insertion of unwanted

mutations, the mutants were cloned back into the pFUdioTealGephyrinW plasmid using the

Asc1 and Nhe1 restriction sites.

In Utero Electroporation

All animal experiments were approved by the Massachusetts Institute of Technology Committee

on Animal Care and meet the National Institute of Health guidelines for the use and care of

vertebrate animals. In utero electroporation on E15.5 timed pregnant C57BL/6J mice was

performed to label L2/3 cortical pyramidal neurons, as previously described (Tabata and

97

Nakajima, 2001). Animals were co-electroporated with Cre-dependent constructs expressing

eYFP, PSD-95-mCherry, or Teal-gephyrin mutants S268A or S270A along with a plasmid

expressing Cre recombinase at a ratio of 10:10:5:1, respectively (total DNA concentration 2

µg/µL), with 0.1% Fast Green for visualization. A total of 0.75 µl of the plasmid solution was

injected into the right lateral ventricle with a 32 gauge Hamilton syringe (Hamilton Company,

Reno, NV, USA). Five pulses of 36V (duration 50 ms, frequency 1 Hz) targeting the visual

cortex were delivered from a square wave electroporator (ECM830; Harvard Apparatus,

Holliston, MA, USA) using 5 mm diameter platinum electrodes (Protech International, Boerne,

TX, USA).

Cranial Window Implantation

After in utero electroporation, pups were reared to adulthood (P42-57) and implanted with a 5

mm cranial window over the right hemisphere, as described (Lee et al., 2008).

Sulfamethoxazole (1 mg/ml) and trimethoprim (0.2 mg/ml) were chronically administered in the

drinking water to maintain optical clarity of implanted windows.

Two-photon imaging

Animals were allowed 1-2 weeks for recovery after cranial window surgery. When windows

cleared, labeled cells in binocular visual cortex were screened for the presence of all 3

fluorescent labels. Cells that exhibited high background labeling with synaptic labels or any

ectopic clumps of synaptic proteins were not used for experiments. This selection process

ensured that all imaged cells had similar levels of fluorescent labeling. Only one cell was

imaged per animal. For eight mice with wild type gephyrin, 2 mice with S268A, and 2 mice with

S270A, in vivo two-photon imaging was performed daily for 9 consecutive sessions using a

custom-built microscope with custom acquisition software to enable triple color imaging. For

each imaging session, mice were anesthesized with isofluorane (0.75%-1.25%) and secured in

98

a stereotaxic frame. Fluorophores were simultaneously excited with a commercial Mai Tai HP

Ti: Sapphire laser (Spectra-Physics, Santa Clara, CA, USA) at 915 nm to excite eYFP and Teal,

and a Chameleon Compact OPO (Coherent, Santa Clara, CA, USA) at 1085 nm to excite

mCherry. The outputs of the two excitation lasers were combined at a polarized beam splitter

with orthogonal polarization directions, and delivered to the two-photon microscope through the

same beam path. After scanning with a galvonometric XY-scanning mirrors (6215H, Cambridge

Technology) and a piezo actuator Z-positioning system (Piezosystem Jena, Jena, Germany),

the two laser beams were focused by a 20x/1.0 NA water immersion objective lens (W Plan-

Apochromat; Zeiss, Jena, Germany) to the same focal volume location in the specimen, within

one pixel size accuracy. These lasers produce ~100 fs unsynchronized pulses at a rate of 80

MHz. The power delivered by each laser to the specimen ranged from approximately 35-50 mW

depending on imaging depth. The emission signals from the three fluorophores were collected

by the same objective lens, passed through an IR blocking filter (E700SP; Chroma Technology,

Bellows Falls, VT, USA), and were separated according to their emission spectra by dichroic

mirrors at 520 nm and 560 nm. After passing through three independent bandpass filters

(485/70m, 550/100m, and 605/75m) the three emission signals were collected simultaneously

onto 3 separate PMTs. Imaging for synapse monitoring was performed at high resolution (250

nm/pixel XY-resolution, 0.9 µm/frame Z-resolution). Two-photon raw scanner data was

processed for spectral linear unmixing and converted into a RGB image z-stack using Matlab

and ImageJ (National Institutes of Health, Bethesda, MD, USA).

Spectral Linear Unmixing and Image Processing

Spectral linear unmixing for three colors was performed using a similar approach as previously

described for two color unmixing (Chen et al., 2012). Briefly, spectral linear unmixing is based

on the fact that the total photon count at each pixel in a given channel is the linear sum of the

spectral contribution of each fluorophore weighted by its abundance. For a triple channel

99

detection system, the contribution of three fluorophores can be represented by the following

equations.

J1(x,y)=s1.1 x I1 (x,y) + s1.2 x I2 (x,y) + s1.3 x I3 (x,y)

J2(x,y)=s2.1 x I1 (x,y) + s2.2 x I2 (x,y) + s2.3 x I3 (x,y)

J3(x,y)=s3.1 x I1 (x,y) + s3.2 x I2 (x,y) + s3.3 x I3 (x,y)

Where J is the total signal per channel, I is the fluorophore abundance, and S is the contribution

of that fluorophore. These equations can be expressed as a matrix:

[J]=[S][I],

whereby the unmixed image [I] can be calculated using the inverse matrix of S:

[I]=[S]-1 [J].

Assuming the detected signal in both channels represents the total spectral contribution for all

three fluorophores:

s1.1 + s2.1 + s3.1 = 1

s1.2 + s2.2 + s3.2 = 1

s1.3 + s2.3 + s3.3 = 1

[S] was determined experimentally from two-photon images of HEK cell cultures expressing

single fluorophores and excited by both lasers at the same wavelengths used in vivo. Average

laser power was adjusted to achieve photon count levels approximating in vivo signal intensity.

We have previously shown that S derived from cell culture and in vivo data is interchangeable

(Chen et al., 2012). The mean contribution for each fluorophore into each channel (s1.1 – 3.3) was

calculated using Matlab (Mathworks, Natick, MA, USA). These values were subsequently used

for spectral linear unmixing of triple channel two-photon raw scanner data into a RBG image z-

100

stack using Matlab and ImageJ (National Institutes of Health). For all figures, images are

filtered and interpolated for optimal visualization. 3D image stacks are converted into maximum

intensity z-projections where noted in figure legends.

Data Analysis

Dendritic spines, PSD-95 containing excitatory synapses, and inhibitory synapses were scored

manually with a custom-written 4D point tracking system implemented in Fiji (Schindelin et al.,

2012) using a modified version of the ObjectJ plugin

(https://sils.fnwi.uva.nl/bcb/objectj/index.html). To avoid individual scoring bias, each cell was

independently scored by two investigators. Dendritic spine analysis criteria was defined as

previously described (Chen et al., 2012; Holtmaat and Svoboda, 2009). Because of the bright

labeling on the soma and axon initial segment, individual contacts in these regions could not be

resolved and analysis was restricted to dendrites starting approximately 40 microns from the

soma to the most distal tips. Because z-projecting spines lacking PSD-95 could not be

visualized, for consistency, all z-projecting spines were excluded from analysis, even those

which contained PSD-95. Gephyrin puncta were scored as synapses if they were at least 3x3

pixels, or 8–9 clustered pixels (0.56 µm2) in size, with a minimal average signal intensity of at

least four times above shot noise background levels, and were present in 2 consecutive z

planes. PSD-95 puncta were scored as synapses if they were least 2x2 pixels, or 4-5 clustered

pixels (0.27 µm2) in size with a minimal average signal intensity of at least four times above shot

noise background levels and were present in 2 consecutive z planes. EM validation confirmed

that these criteria represent inhibitory and excitatory synapses, respectively, with no false

positives. In general, the PSD-95-mCherry fluorophore was more prone to bleaching than YFP

or Teal-gephyrin. Cells that lost PSD-95-mCherry were not analyzed further. One-time dynamic

changes were defined as any structure that changed one time and never changed again.

Changes were scored as transient if a structure was added and later removed in a subsequent

101

imaging session. Changes were scored as recurrent if structures appeared, then disappeared,

and then appeared again at the same site; or conversely, ever disappeared and then

reappeared more than once in the same location. At least 25 spines and synapses were

counted per branch and at least 200 structures were counted per animal. Dendritic arbors were

manually traced in Neurolucida (MicroBrightField, Inc., Williston, VT, USA) to quantify the length

of scored branches.

102

Chapter 5 – Conclusion

Inhibitory synapses dynamically regulate the integration of neuronal signals and the

firing of individual neurons, and a proper balance between excitation and inhibition is critical for

normal brain function. The precise timing of inhibitory up- and down- regulation during

development has been implicated in the opening and closure of critical periods (Fagiolini and

Hensch, 2000; Huang et al., 1999), and in adults, there is evidence that reduction in inhibitory

tone can open the door for increased plasticity (Chen et al., 2011; Harauzov et al., 2010). Thus

gaining deeper insight into the distribution and dynamics of inhibitory synapses, and how they

relate to excitatory synapses, is essential for understanding the function, not only of a given

neuron within its neural circuit, but also of plasticity mechanisms. Only recently has the

expression of fluorescent proteins fused to postsynaptic scaffolding molecules enabled direct

synaptic visualization in vivo. Fluorescent proteins fused either to PSD-95, as a post synaptic

marker of excitatory synapses, or gephyrin, as a post synaptic marker of inhibitory

synapses, together with a separate cell fill to outline their location on the arbor has allowed

a direct assessment of excitatory and inhibitory synaptic distribution patterns on cortical

pyramidal cells, and their dynamics at different cellular locales.

Clustered dynamics of inhibitory synapses and dendritic spines in the adult neocortex

Using dual-color two photon microscopy, we labeled inhibitory synapses and used a cell fill

to track spines on L2/3 pyramidal neurons. Although inhibitory synapses on spines had been

anecdotally observed before, through electron microcopy studies on short dendritic segments,

by labeling inhibitory synapses in vivo across the entire dendritic arbor, we observed that

inhibitory synapses on spines account for 1/3 of the total population of inhibitory synapses, a

much greater proportion than had previously been thought. Further, we found these synapses to

103

be non-uniformly distributed across the arbor, with the highest density at the apical tuft.

Inhibitory synapses on dendritic spines have been shown to specifically gate depolarization by

the excitatory synapse located on the same spine (Chiu et al., 2013), suggesting they play a

different inhibitory role than shaft synapses. While inhibitory synapses on spines specifically

inhibit the excitatory synapse on that same spine, inhibitory synapses on the shaft have a

broader effect on the integration of many excitatory synapses along that branch. By showing the

prevalence of inhibitory synapse population and their non-uniform distribution, we introduce a

new player with a previously unrecognized role to the modeling of synaptic integration along the

dendritic arbor.

The discovery that 29.7% of inhibitory synapses are located on spines was very surprising,

since data from electron microscopy studies show less frequent incidence of these synapses. In

a complete reconstruction of all synaptic connections from an approximately 20 µm length (1500

µm3 volume) area in L5, It was found that 81% of inhibitory synapses are located on the shaft,

and only 19% of inhibitory synapses are located on spines (Kasthuri et al., 2015). This

discrepancy could be due to the small number of inhibitory synapses counted in the EM dataset

(86 inhibitory synapses), or could potentially be because there are fewer inhibitory synapses on

spines in L5 as compared to L2/3. Indeed, we have seen that there is a greater density of

inhibitory synapses on spines in the more superficial region of L2/3 than the deeper region of

L2/3, suggesting that different layers of cortex have a different distribution of inhibitory synapses

on spines. Although it is unlikely that the exogenous Teal-gephyrin expression is inducing the

formation of additional inhibitory synapses, we cannot formally rule it out. Why do we believe

that our teal-gephyrin plasmid is not inducing an increase in the number of inhibitory synapses

due to increased concentration of gephyrin in the cell? The dio lentiviral plasmids are known to

be low expressing, and unlikely to cause overexpression. Expression of PSD-95-mCherry,

which had previously been shown to cause an increase in spine density, did not do so in our

104

experiments using the same expression plasmid as for the Teal-gephyrin. Other studies have

used tagged gephyrin in vivo and seen no evidence that its expression causes an increase in

the density of inhibitory synapses (Chen et al., 2015; van Versendaal et al., 2012). Most

importantly, studies using in vivo electrophysiology have reported that 51% of randomly imaged

spines on L/2 neurons showed significant inhibition of calcium signaling upon GABA uncaging

(Chiu et al., 2013), numbers that are closer to ours.

By imaging cells every four days, we found that inhibitory synapses on dendritic spines and

shafts display different kinetics of dynamics in vivo. Inhibitory synapses on dendritic spines are

much more dynamic than those located on the shaft during normal experience. Monocular

deprivation (MD), a model of experience-dependent plasticity, also causes differential changes

in inhibitory shaft and spine synapses. While both types of inhibitory synapses are lost after MD,

inhibitory synapses on the shaft are lost consistently over an 8-day period, while inhibitory

synapses on spines are lost primarily in the first 2 days after MD. This also implies that the

spine and shaft inhibitory synapses play different roles in the regulation of neuronal signaling

during experience-dependent plasticity. Z-projecting inhibitory synapses on spines would be

scored as shaft synapses, and this could be slightly increasing our rate of dynamics observed

for inhibitory shaft synapses.

An additional finding, enabled by our ability to visualize both spines and inhibitory synapses

was that inhibitory synapse dynamics and dendritic spine dynamics are preferentially clustered

within 10 µm on dendritic branches. It has been shown that calcium-dependent signaling

molecules can diffuse at approximately a 10 µm distance to influence the plasticity of

neighboring synapses (Govindarajan et al., 2011; Harvey and Svoboda, 2007). However, we did

not observe clustering between two dynamic excitatory synapses, or between two neighboring

inhibitory synapses, only between dynamic excitatory and inhibitory synapses. This opens

questions regarding the timescale of these clustered dynamics, as well as the order and valance

105

of the changes. Do inhibitory synapse changes precede those of excitatory synapses? If there is

a loss of an inhibitory synapse, is there a corresponding loss or gain of an excitatory synapse?

These are interesting questions that remain to be answered, but will require analysis of larger

data sets.

Inhibitory synapses are repeatedly assembled and removed from persistent sites

As a first step to investigating the temporal relationship between inhibitory synapse and

spine dynamics, we reduced our imaging interval from 4 days to 24 hours. To perform these

experiments we needed a way to also label excitatory synapses, since on short time scales one

cannot assume that newly formed spines represent excitatory synaptic presence. Dual color

studies with PSD-95-GFP labeling in addition to a cell fill in vivo shows that most stable spines

have large PSD-95 puncta, but a small population are devoid of a PSD95 label (Cane et al.,

2014). A recent EM study shows that only 5% of spines lack an excitatory synapse dendrites

(Kasthuri et al., 2015). Unlike the gephyrin, which is a faithful marker of inhibitory synapses of all

sizes, PSD-95 is only thought to be present at large, mature synapses. Smaller, immature

spines contain other scaffolding molecules, such as PSD93 or SAP 102 instead of PSD-95(Aoki

et al., 2001; Elias et al., 2008; Sans et al., 2000). The 20% of spines lacking PSD-95 that we

observe are highly dynamic and are therefore likely to contain small immature synapses that

lack PSD95. Adding a third color, PSD-95-mCherry, to our labeling allowed us to confirm that all

spines which have an inhibitory synapse also have an excitatory synapse, and were thus

termed dually innervated spines (DiS).

By imaging at 24 hour intervals, we surprisingly found that, while PSD-95 puncta are

mostly stable, inhibitory synapses rapidly appear and disappear from the same locations. Both

inhibitory shaft synapses and inhibitory synapses on DiS display this behavior, though inhibitory

106

synapses on DiS are much more dynamic. Also, as previously mentioned, some of the shaft

dynamics could be in reality due to inhibitory synapses on z-projecting spines. The observation

of recurrent dynamics represents a fundamental change in our understanding of how inhibitory

synapses regulate neuronal activity. Instead of synaptic changes representing a permanent loss

or gain of connections between neurons, we propose that rapid formation or elimination of

inhibitory connections at discrete sites can act as a reversible modulatory switch for controlling

individual excitatory synapses (on DiS) or regions of dendritic branches (on the shaft). It had

been suggested that cranial window preparations could make rotation of dendritic branches

more likely (Grutzendler et al., 2002), which could result in an increase in the dynamics

observed. It may be true that slight rotations in the dendrite would cause some spines lacking

PSD-95 to appear or disappear in the z-plane, and could thus inflate our observed dynamics of

these spines, however, this is unlikely to affect our conclusions for gephyrin or PSD-95

containing spines, as these labels are observable even if a spine has a slight rotation into the z-

plane. Our quantification of the relative size of teal gephyrin puncta shows that the recurrent

puncta are not simply the smallest class of synapses, but encompass a wide range of sizes

(Fig. 3.10). This implies that the dynamics we see of recurrent inhibitory synapses are not

simply an artefact due to the smallest synapses falling above and below our detection threshold.

Figure 5.1: Recurrent inhibitory synapses are not

smaller than stable inhibitory synapses on DiS. Size

distribution of stable, one-time dynamic, and recurrent

inhibitory synpases on DiS quantified by measuring the

intensity of pixils in the teal-gephyrin channel. This data

is not normalized by spine size, only raw data is shown.

107

To further demonstrate this point, we quantified the size of teal-gephyrin puncta without

normalization to spine size (Figure 5.1), and again we see that the size of recurrent gephyrin

puncta is no smaller than that of stable gephyrin puncta. While it is true that when we fail to see

a gephyrin puncta with two photon imaging, it may be the case that a few gephyrin molecules

are still present but fall below our detection threshold, however the electrophysiology

experiments (Fig. 3.3) argue against this as there was a clear distinction in uIPSC amplitude

between spines containing and spines lacking gephyrin puncta.

We also found that DiS are the largest spines, and the most stable. The excitatory

synapses on these spines are also extremely stable. There is some evidence that DiS may be

innervated by excitatory afferents from the thalamus (Kubota et al., 2007), suggesting that the

inhibitory synapses on these spines exert specific control over stable feed-forward sensory

inputs. The presence or absence of the inhibitory synapses on DiS is strongly affected by

sensory experience. MD causes their destabilization, and a transition to a more dynamic state.

Functional implications of inhibitory synaptic placement across the dendritic arbor

The placement and activity of inhibitory synapses plays a critical role in the integration of

signals along dendrites. Individual pyramidal cells can receive thousands of excitatory synaptic

inputs that can be functionally diverse. Yet, the action potential output of most pyramidal

neurons is sparse and narrowly tuned. This restriction of action potential firing to a defined

sensory range within a limited temporal window is brought about by the integration of excitatory

synaptic events combined with the local influence of inhibition (reviewed in (Chadderton et al.,

2014). Inhibition has a major influence on the effective output of neurons at multiple levels.

Individual inhibitory synapses on spines can veto individual spine conductances (Chiu et al.,

2013). Inhibitory shaft synapses can influence excitatory integration within local dendritic

segments (Kim et al., 1995; Perez-Garci et al., 2006; Willadt et al., 2013). On a more cellular

108

scale, PV cell innervation of the soma (Pouille and Scanziani, 2001), or chandelier cell

innervation of the axon initial segment (Woodruff et al., 2011) can cancel axonal output. In the

latter cases, the high density of inhibitory synapses on the cell body and axon initial segment (Di

Cristo et al., 2004; Miles et al., 1996; Woodruff et al., 2010) suggests that synchronous firing of

many synapses would be needed for effective inhibition in these locales, and the firing of a

single inhibitory synapse would have a negligible effect on outcome. This is in contrast to

dendritically targeted inhibition, where individual synapse changes could have a dramatic effect

on the computation of dendritic integration for specific branches.

How are excitatory and inhibitory inputs integrated functionally at the level of individual

dendritic branches and how does that influence overall neuronal firing? Clearly, local integration

of excitatory and inhibitory synaptic activity is influenced by individual synaptic strength, timing,

and location on the dendrite. Theoretical modeling predicts that activation of multiple distributed

inhibitory synapses across the dendritic arbor can result in a global inhibition (Gidon and Segev,

2012). However, experiments suggest that the influence of individual inhibitory synapses is

locally restricted to the branch they target (Stokes et al., 2014), perhaps even to individual

spines (Chiu et al., 2013). Thus, functional pairing or clustering of relevant excitatory synapses

on particular dendritic segments with their inhibitory neighbors is highly relevant to dendritic

integration. Moreover the local constellation of such pairing or clustering is likely to have a

different influence depending on their more global location along the arbor.

Dendritically targeted inhibition can limit initiation of dendritic Ca2+ spikes in both

hippocampal and cortical pyramidal neurons (Buhl et al., 1994; Karube et al., 2004; Kim et al.,

1995; Larkum et al., 1999; Perez-Garci et al., 2006; Willadt et al., 2013), so that they are

initiated only in response to inputs encoding particular functional features. Recent experiments

in slice culture have shown that single inhibitory synapses can completely inhibit a back-

propagating action potential (Mullner et al., 2015). Mathematical models focusing on spike

109

activity within dendrites have shown that “off path” inhibitory synapses, where the inhibitory

synapse is located distally to the excitatory synapse, are more effective at preventing the

initiation of dendritic spikes (Gidon and Segev, 2012). Thus, the influence of inhibition on spike

initiation can actually be more powerful when inhibitory synapses are located on distal dendrites

rather than proximal dendrites and in a constellation where the inhibitory synapse is distal to its

functionally paired excitatory synapse (Gidon and Segev, 2012; Li et al., 2014). Conversely, in

cases where a dendritic spike has already been successfully initiated, an ‘on path’ inhibitory

synapse, located between the excitatory synapse and the soma, is the most effective way to

dampen the signal (Gidon and Segev, 2012; Hao et al., 2009; Zhang et al., 2013).

Our in vivo work demonstrated that dynamic inhibitory synapses are likely to be located

within 10 µm of dynamic spines, and this clustering of dynamic events increases upon sensory

deprivation (Chen et al., 2012). It has been suggested that a loss of inhibition creates a

permissive environment for excitatory synaptic changes (Chen et al., 2011; Harauzov et al.,

2010; Maya Vetencourt et al., 2008) because it broadens the window of spike timing dependent

plasticity (Bi and Poo, 2001; Higley and Contreras, 2006; Song et al., 2000; Spruston, 2008).

The loss of an inhibitory synapse could create a permissive environment for neighboring

excitatory synaptic gain or loss, while the addition of an inhibitory synapse could suppress

plasticity in neighboring dendritic regions. Inhibitory inputs could also limit the ability to promote

excitatory synaptic clustering through plasticity mechanisms. Thus, the specific placement of

inhibitory synapses can regulate which excitatory synapses are able to undergo plasticity.

Further, a recent modeling study suggests that the placement of inhibitory synapses can result

in LTP, LTD, or a lack of plasticity in specific neighboring dendritic segments (Bar-Ilan et al.,

2012), highlighting the importance on the placement of inhibitory synapses on the valance of

plasticity at neighboring excitatory synapses.

110

Molecular mechanism of rapid gephyrin dynamics

Based on our observation that inhibitory synapses can rapidly disappear and reappear

from the same location in vivo, we wanted to explore the molecular mechanism that is the basis

for these rapid dynamics. We hypothesize that the rapid recurrence of inhibitory synapses is

due to the fact that the inhibitory axon that remains abutting the postsynaptic compartment

could serve as a placeholder for the return of the inhibitory synapse in that same location. Thus,

the decision to form or destroy a synapse may depend on the activity within the postsynaptic

cell. This suggests that there may be activity-dependent signaling mechanisms occurring within

the postsynaptic cell which regulate the stability of gephyrin.

One possibility for the regulation of gephyrin stability is the signaling pathways that result

in the phosphorylation of gephyrin at the serine 268 and 270 residues, which results in the

degradation of gephyrin by calpain, a calcium-activate protease (Fig. 5.1). Phosphorylation of

S268 is performed by ERK1/2, which is known to be activated by increased neuronal activity

and activated BDNF. Phosphorylation at S270 is performed by GSK-3β, which may be inhibited

by activity, because it is inhibited during LTD stimulation protocols. Thus the two serine residues

seem to be phosphorylated under different conditions of neuronal activity, and mutants display

different clustering phenotypes in cell culture experiments.

In order to explore the molecular alterations and pathways that could lead to the rapid

disassembly and reassembly of gephyrin in vivo, we created gephyrin S268A and S270A

phospho-abolishing mutations. Preliminary results indicate that the S270A mutation increases

the density of inhibitory synapses on DiS and increases the rapid dynamics of inhibitory

synapses on the dendritic shaft. This seems to indicate that the dephosphorylation of gephyrin

S270 does not affect established gephyrin clusters, because they do not show increased

stability as expected. However, the dephosphorylation of S270 may slow the degradation of

111

gephyrin monomers, increasing the available pool of monomeric gephyrin, and thus allow

inhibitory synapses to reform quickly.

Epilogue

We would like to continue experiments on gephyrin S268A and S270A mutants to

increase the sample size and determine if there are any significant effects of the S268A

mutation in vivo.

Figure 5.1. Gephyrin clustering at GABAergic synapses is modulated by various post-translational modifications.

Glycogen synthase kinase 3β (GSK3β) and extracellular signal-regulated kinases 1 and 2 (ERK1/2), which target

Ser270 and Ser268, respectively, cooperate with calpain to negatively modulate gephyrin clustering. The scheme

also indicates the need for calcium (from unknown sources) to activate calpain (and thereby cleave gephyrin

molecules) as well as to activate calcium-activated protein kinases (and phosphatases), such as

calcium/calmodulin-dependent protein kinase IIα (CaMKIIα), which are known to regulate GABAAR

phosphorylation. In addition, MAM domain-containing glycosylphosphatidylinositol anchor protein (MDGA) interacts

with NLGN2 to suppress the formation of GABAergic synapses, whereas immunoglobulin superfamily member 9B

(IGSF9B; which is enriched in cortical interneurons) interacts with NLGN2 through S-SCAM to facilitate GABAergic

synapse formation. PI3K, phosphatidylinositol 3-kinase. Figure from (Tyagarajan and Fritschy, 2014).

112

We also want to test the effect of neuronal activity on gephyrin dynamics. Since gephyrin

is degraded by calpain, a calcium-activated enzyme, it is likely that neuronal activity can alter

the activity of calpain, and thus the stability or rate of formation of gephyrin clusters. We would

like to see what affect altering neuronal activity would have on wild type gephyrin, and if this

effect is altered by the expression of gephyrin mutants. It is possible to activate or silence

individual postsynaptic neurons by expressing DREADD constructs or using optogenetics, and

then exploring how alterations in activity affect the dynamics of wild type gephyrin. It would be

interesting to see if activating the postsynaptic neuron has the immediate effect of increasing

the number of inhibitory synapses, as would be expected with the homeostatic hypothesis, or if

it causes the activation of calpain, which would result in the degradation of gephyrin.

Interestingly, a recent study in hippocampal slice cultures explored the activity-dependent

modulation of gephyrin dynamics (Flores et al., 2015). Carbachol treatment, a learning-related

protocol which induces structural plasticity, caused an increase in the density and size of

gephyrin clusters, and a theta-burst stimulation produced the same effect, which was prevented

by the blockade of NMDA receptors. Treatment with the GABA-A receptor blocker gabazine,

also induced and increase in the size and density of gephyrin clusters, and 24 hour optogenetic

stimulation had the same effect. These results suggest that gephyrin synaptic dynamics are

homeostatically regulated, and it would be interesting to explore these ideas further in vivo.

113

References

Alonso, M., Medina, J.H., and Pozzo-Miller, L. (2004). ERK1/2 activation is necessary for BDNF to increase dendritic spine density in hippocampal CA1 pyramidal neurons. Learn Mem 11, 172-178. Aoki, C., Miko, I., Oviedo, H., Mikeladze-Dvali, T., Alexandre, L., Sweeney, N., and Bredt, D.S. (2001). Electron microscopic immunocytochemical detection of PSD-95, PSD-93, SAP-102, and SAP-97 at postsynaptic, presynaptic, and nonsynaptic sites of adult and neonatal rat visual cortex. Synapse 40, 239-257. Arellano, J.I., Espinosa, A., Fairen, A., Yuste, R., and DeFelipe, J. (2007). Non-synaptic dendritic spines in neocortex. Neuroscience 145, 464-469. Atasoy, D., Aponte, Y., Su, H.H., and Sternson, S.M. (2008). A FLEX switch targets Channelrhodopsin-2 to multiple cell types for imaging and long-range circuit mapping. J Neurosci 28, 7025-7030. Azouz, R., Gray, C.M., Nowak, L.G., and McCormick, D.A. (1997). Physiological properties of inhibitory interneurons in cat striate cortex. Cereb Cortex 7, 534-545. Bar-Ilan, L., Gidon, A., and Segev, I. (2012). The role of dendritic inhibition in shaping the plasticity of excitatory synapses. Frontiers in neural circuits 6, 118. Baude, A., Nusser, Z., Roberts, J.D., Mulvihill, E., McIlhinney, R.A., and Somogyi, P. (1993). The metabotropic glutamate receptor (mGluR1 alpha) is concentrated at perisynaptic membrane of neuronal subpopulations as detected by immunogold reaction. Neuron 11, 771-787. Bi, G., and Poo, M. (2001). Synaptic modification by correlated activity: Hebb's postulate revisited. Annu Rev Neurosci 24, 139-166. Bloodgood, B.L., Sharma, N., Browne, H.A., Trepman, A.Z., and Greenberg, M.E. (2013). The activity-dependent transcription factor NPAS4 regulates domain-specific inhibition. Nature 503, 121-125. Brown, K.N., Chen, S., Han, Z., Lu, C.H., Tan, X., Zhang, X.J., Ding, L., Lopez-Cruz, A., Saur, D., Anderson, S.A., et al. (2011). Clonal production and organization of inhibitory interneurons in the neocortex. Science 334, 480-486. Buhl, E.H., Halasy, K., and Somogyi, P. (1994). Diverse sources of hippocampal unitary inhibitory postsynaptic potentials and the number of synaptic release sites. Nature 368, 823-828. Cane, M., Maco, B., Knott, G., and Holtmaat, A. (2014). The relationship between PSD-95 clustering and spine stability in vivo. J Neurosci 34, 2075-2086.

114

Cavanaugh, J.E., Ham, J., Hetman, M., Poser, S., Yan, C., and Xia, Z. (2001). Differential regulation of mitogen-activated protein kinases ERK1/2 and ERK5 by neurotrophins, neuronal activity, and cAMP in neurons. J Neurosci 21, 434-443. Chadderton, P., Schaefer, A.T., Williams, S.R., and Margrie, T.W. (2014). Sensory-evoked synaptic integration in cerebellar and cerebral cortical neurons. Nat Rev Neurosci 15, 71-83. Chen, J.L., Lin, W.C., Cha, J.W., So, P.T., Kubota, Y., and Nedivi, E. (2011). Structural basis for the role of inhibition in facilitating adult brain plasticity. Nat Neurosci 14, 587-594. Chen, J.L., and Nedivi, E. (2010). Neuronal structural remodeling: is it all about access? Curr Opin Neurobiol 20, 557-562. Chen, J.L., Villa, K.L., Cha, J.W., So, P.T., Kubota, Y., and Nedivi, E. (2012). Clustered dynamics of inhibitory synapses and dendritic spines in the adult neocortex. Neuron 74, 361-373. Chiu, C.Q., Lur, G., Morse, T.M., Carnevale, N.T., Ellis-Davies, G.C., and Higley, M.J. (2013). Compartmentalization of GABAergic inhibition by dendritic spines. Science 340, 759-762. Chklovskii, D.B., Mel, B.W., and Svoboda, K. (2004). Cortical rewiring and information storage. Nature 431, 782-788. Craig, A.M., Banker, G., Chang, W., McGrath, M.E., and Serpinskaya, A.S. (1996). Clustering of gephyrin at GABAergic but not glutamatergic synapses in cultured rat hippocampal neurons. J Neurosci 16, 3166-3177. Dani, V.S., Chang, Q., Maffei, A., Turrigiano, G.G., Jaenisch, R., and Nelson, S.B. (2005). Reduced cortical activity due to a shift in the balance between excitation and inhibition in a mouse model of Rett syndrome. Proc Natl Acad Sci U S A 102, 12560-12565. Davis, T.L., and Sterling, P. (1979). Microcircuitry of cat visual cortex: classification of neurons in layer IV of area 17, and identification of the patterns of lateral geniculate input. J Comp Neurol 188, 599-627. De Roo, M., Klauser, P., Mendez, P., Poglia, L., and Muller, D. (2008). Activity-dependent PSD formation and stabilization of newly formed spines in hippocampal slice cultures. Cereb Cortex 18, 151-161. DeFelipe, J., Alonso-Nanclares, L., and Arellano, J.I. (2002). Microstructure of the neocortex: comparative aspects. J Neurocytol 31, 299-316. Dhande, O.S., Hua, E.W., Guh, E., Yeh, J., Bhatt, S., Zhang, Y., Ruthazer, E.S., Feller, M.B., and Crair, M.C. (2011). Development of single retinofugal axon arbors in normal and beta2 knock-out mice. J Neurosci 31, 3384-3399. Di Cristo, G., Wu, C., Chattopadhyaya, B., Ango, F., Knott, G., Welker, E., Svoboda, K., and Huang, Z.J. (2004). Subcellular domain-restricted GABAergic innervation in primary visual cortex in the absence of sensory and thalamic inputs. Nat Neurosci 7, 1184-1186. Ehrlich, I., Klein, M., Rumpel, S., and Malinow, R. (2007). PSD-95 is required for activity-driven synapse stabilization. Proc Natl Acad Sci U S A 104, 4176-4181.

115

Elias, G.M., Elias, L.A., Apostolides, P.F., Kriegstein, A.R., and Nicoll, R.A. (2008). Differential trafficking of AMPA and NMDA receptors by SAP102 and PSD-95 underlies synapse development. Proc Natl Acad Sci U S A 105, 20953-20958. Elston, G.N., and Rosa, M.G. (1997). The occipitoparietal pathway of the macaque monkey: comparison of pyramidal cell morphology in layer III of functionally related cortical visual areas. Cereb Cortex 7, 432-452. Essrich, C., Lorez, M., Benson, J.A., Fritschy, J.M., and Luscher, B. (1998). Postsynaptic clustering of major GABAA receptor subtypes requires the gamma 2 subunit and gephyrin. Nat Neurosci 1, 563-571. Fagiolini, M., and Hensch, T.K. (2000). Inhibitory threshold for critical-period activation in primary visual cortex. Nature 404, 183-186. Feldman, M.L., and Peters, A. (1978). The forms of non-pyramidal neurons in the visual cortex of the rat. J Comp Neurol 179, 761-793. Fiala, J.C. (2005). Reconstruct: a free editor for serial section microscopy. J Microsc 218, 52-61. Flores, C.E., Nikonenko, I., Mendez, P., Fritschy, J.M., Tyagarajan, S.K., and Muller, D. (2015). Activity-dependent inhibitory synapse remodeling through gephyrin phosphorylation. Proc Natl Acad Sci U S A 112, E65-72. Frenkel, M.Y., Sawtell, N.B., Diogo, A.C., Yoon, B., Neve, R.L., and Bear, M.F. (2006). Instructive effect of visual experience in mouse visual cortex. Neuron 51, 339-349. Fuhrmann, J.C., Kins, S., Rostaing, P., El Far, O., Kirsch, J., Sheng, M., Triller, A., Betz, H., and Kneussel, M. (2002). Gephyrin interacts with Dynein light chains 1 and 2, components of motor protein complexes. J Neurosci 22, 5393-5402. Gabbott, P.L., and Somogyi, P. (1986). Quantitative distribution of GABA-immunoreactive neurons in the visual cortex (area 17) of the cat. Exp Brain Res 61, 323-331. Gidon, A., and Segev, I. (2012). Principles governing the operation of synaptic inhibition in dendrites. Neuron 75, 330-341. Golding, N.L., Staff, N.P., and Spruston, N. (2002). Dendritic spikes as a mechanism for cooperative long-term potentiation. Nature 418, 326-331. Govindarajan, A., Israely, I., Huang, S.Y., and Tonegawa, S. (2011). The dendritic branch is the preferred integrative unit for protein synthesis-dependent LTP. Neuron 69, 132-146. Govindarajan, A., Kelleher, R.J., and Tonegawa, S. (2006). A clustered plasticity model of long-term memory engrams. Nat Rev Neurosci 7, 575-583. Gradinaru, V., Zhang, F., Ramakrishnan, C., Mattis, J., Prakash, R., Diester, I., Goshen, I., Thompson, K.R., and Deisseroth, K. (2010). Molecular and cellular approaches for diversifying and extending optogenetics. Cell 141, 154-165.

116

Grosskreutz, Y., Betz, H., and Kneussel, M. (2003). Rescue of molybdenum cofactor biosynthesis in gephyrin-deficient mice by a Cnx1 transgene. Biochem Biophys Res Commun 301, 450-455. Grutzendler, J., Kasthuri, N., and Gan, W.B. (2002). Long-term dendritic spine stability in the adult cortex. Nature 420, 812-816. Gulyas, A.I., Megias, M., Emri, Z., and Freund, T.F. (1999). Total number and ratio of excitatory and inhibitory synapses converging onto single interneurons of different types in the CA1 area of the rat hippocampus. J Neurosci 19, 10082-10097. Gulyas, A.I., Miettinen, R., Jacobowitz, D.M., and Freund, T.F. (1992). Calretinin is present in non-pyramidal cells of the rat hippocampus--I. A new type of neuron specifically associated with the mossy fibre system. Neuroscience 48, 1-27. Hao, J., Wang, X.D., Dan, Y., Poo, M.M., and Zhang, X.H. (2009). An arithmetic rule for spatial summation of excitatory and inhibitory inputs in pyramidal neurons. Proc Natl Acad Sci U S A 106, 21906-21911. Harauzov, A., Spolidoro, M., DiCristo, G., De Pasquale, R., Cancedda, L., Pizzorusso, T., Viegi, A., Berardi, N., and Maffei, L. (2010). Reducing intracortical inhibition in the adult visual cortex promotes ocular dominance plasticity. J Neurosci 30, 361-371. Harris, K.M., Jensen, F.E., and Tsao, B. (1992). Three-dimensional structure of dendritic spines and synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation of synaptic physiology and long-term potentiation. J Neurosci 12, 2685-2705. Harvey, C.D., and Svoboda, K. (2007). Locally dynamic synaptic learning rules in pyramidal neuron dendrites. Nature 450, 1195-1200. He, H.Y., Hodos, W., and Quinlan, E.M. (2006). Visual deprivation reactivates rapid ocular dominance plasticity in adult visual cortex. J Neurosci 26, 2951-2955. Hendry, S.H., and Jones, E.G. (1988). Activity-dependent regulation of GABA expression in the visual cortex of adult monkeys. Neuron 1, 701-712. Hersch, S.M., and White, E.L. (1981). Quantification of synapses formed with apical dendrites of Golgi-impregnated pyramidal cells: variability in thalamocortical inputs, but consistency in the ratios of asymmetrical to symmetrical synapses. Neuroscience 6, 1043-1051. Higley, M.J., and Contreras, D. (2006). Balanced excitation and inhibition determine spike timing during frequency adaptation. J Neurosci 26, 448-457. Hofer, S.B., Mrsic-Flogel, T.D., Bonhoeffer, T., and Hubener, M. (2009). Experience leaves a lasting structural trace in cortical circuits. Nature 457, 313-317. Holtmaat, A., Bonhoeffer, T., Chow, D.K., Chuckowree, J., De Paola, V., Hofer, S.B., Hubener, M., Keck, T., Knott, G., Lee, W.C., et al. (2009). Long-term, high-resolution imaging in the mouse neocortex through a chronic cranial window. Nat Protoc 4, 1128-1144.

117

Holtmaat, A., and Svoboda, K. (2009). Experience-dependent structural synaptic plasticity in the mammalian brain. Nat Rev Neurosci 10, 647-658. Holtmaat, A., Wilbrecht, L., Knott, G.W., Welker, E., and Svoboda, K. (2006). Experience-dependent and cell-type-specific spine growth in the neocortex. Nature 441, 979-983. Holtmaat, A.J., Trachtenberg, J.T., Wilbrecht, L., Shepherd, G.M., Zhang, X., Knott, G.W., and Svoboda, K. (2005). Transient and persistent dendritic spines in the neocortex in vivo. Neuron 45, 279-291. Hooper, C., Markevich, V., Plattner, F., Killick, R., Schofield, E., Engel, T., Hernandez, F., Anderton, B., Rosenblum, K., Bliss, T., et al. (2007). Glycogen synthase kinase-3 inhibition is integral to long-term potentiation. Eur J Neurosci 25, 81-86. Huang, Z.J., Kirkwood, A., Pizzorusso, T., Porciatti, V., Morales, B., Bear, M.F., Maffei, L., and Tonegawa, S. (1999). BDNF regulates the maturation of inhibition and the critical period of plasticity in mouse visual cortex. Cell 98, 739-755. Isshiki, M., Tanaka, S., Kuriu, T., Tabuchi, K., Takumi, T., and Okabe, S. (2014). Enhanced synapse remodelling as a common phenotype in mouse models of autism. Nat Commun 5, 4742. Kalatsky, V.A., and Stryker, M.P. (2003). New paradigm for optical imaging: temporally encoded maps of intrinsic signal. Neuron 38, 529-545. Karube, F., Kubota, Y., and Kawaguchi, Y. (2004). Axon branching and synaptic bouton phenotypes in GABAergic nonpyramidal cell subtypes. J Neurosci 24, 2853-2865. Kasthuri, N., Hayworth, K.J., Berger, D.R., Schalek, R.L., Conchello, J.A., Knowles-Barley, S., Lee, D., Vazquez-Reina, A., Kaynig, V., Jones, T.R., et al. (2015). Saturated Reconstruction of a Volume of Neocortex. Cell 162, 648-661. Kawaguchi, Y., Karube, F., and Kubota, Y. (2006). Dendritic branch typing and spine expression patterns in cortical nonpyramidal cells. Cereb Cortex 16, 696-711. Kawaguchi, Y., and Kubota, Y. (1998). Neurochemical features and synaptic connections of large physiologically-identified GABAergic cells in the rat frontal cortex. Neuroscience 85, 677-701. Kaytor, M.D., and Orr, H.T. (2002). The GSK3 beta signaling cascade and neurodegenerative disease. Curr Opin Neurobiol 12, 275-278. Keck, T., Scheuss, V., Jacobsen, R.I., Wierenga, C.J., Eysel, U.T., Bonhoeffer, T., and Hubener, M. (2011). Loss of sensory input causes rapid structural changes of inhibitory neurons in adult mouse visual cortex. Neuron 71, 869-882. Kelsch, W., Lin, C.W., and Lois, C. (2008). Sequential development of synapses in dendritic domains during adult neurogenesis. Proc Natl Acad Sci U S A 105, 16803-16808.

118

Kilman, V., van Rossum, M.C., and Turrigiano, G.G. (2002). Activity deprivation reduces miniature IPSC amplitude by decreasing the number of postsynaptic GABA(A) receptors clustered at neocortical synapses. J Neurosci 22, 1328-1337. Kim, H.G., Beierlein, M., and Connors, B.W. (1995). Inhibitory control of excitable dendrites in neocortex. J Neurophysiol 74, 1810-1814. Kleschevnikov, A.M., Belichenko, P.V., Villar, A.J., Epstein, C.J., Malenka, R.C., and Mobley, W.C. (2004). Hippocampal long-term potentiation suppressed by increased inhibition in the Ts65Dn mouse, a genetic model of Down syndrome. J Neurosci 24, 8153-8160. Kneussel, M., Brandstatter, J.H., Laube, B., Stahl, S., Muller, U., and Betz, H. (1999). Loss of postsynaptic GABA(A) receptor clustering in gephyrin-deficient mice. J Neurosci 19, 9289-9297. Knott, G.W., Quairiaux, C., Genoud, C., and Welker, E. (2002). Formation of dendritic spines with GABAergic synapses induced by whisker stimulation in adult mice. Neuron 34, 265-273. Kubota, Y. (2014). Untangling GABAergic wiring in the cortical microcircuit. Curr Opin Neurobiol 26, 7-14. Kubota, Y. (2015). New developments in electron microscopy for serial image acquisition of neuronal profiles. Microscopy (Oxf) 64, 27-36. Kubota, Y., Hatada, S., Kondo, S., Karube, F., and Kawaguchi, Y. (2007). Neocortical inhibitory terminals innervate dendritic spines targeted by thalamocortical afferents. J Neurosci 27, 1139-1150. Kubota, Y., Hatada, S.N., and Kawaguchi, Y. (2009). Important factors for the three-dimensional reconstruction of neuronal structures from serial ultrathin sections. Frontiers in neural circuits 3, 4. Kubota, Y., Karube, F., Nomura, M., Gulledge, A.T., Mochizuki, A., Schertel, A., and Kawaguchi, Y. (2011a). Conserved properties of dendritic trees in four cortical interneuron subtypes. Sci Rep 1, 89. Kubota, Y., Shigematsu, N., Karube, F., Sekigawa, A., Kato, S., Yamaguchi, N., Hirai, Y., Morishima, M., and Kawaguchi, Y. (2011b). Selective coexpression of multiple chemical markers defines discrete populations of neocortical GABAergic neurons. Cereb Cortex 21, 1803-1817. Kwon, H.B., Kozorovitskiy, Y., Oh, W.J., Peixoto, R.T., Akhtar, N., Saulnier, J.L., Gu, C., and Sabatini, B.L. (2012). Neuroligin-1-dependent competition regulates cortical synaptogenesis and synapse number. Nat Neurosci 15, 1667-1674. Larkman, A.U. (1991). Dendritic morphology of pyramidal neurones of the visual cortex of the rat: III. Spine distributions. J Comp Neurol 306, 332-343. Larkum, M.E., and Nevian, T. (2008). Synaptic clustering by dendritic signalling mechanisms. Curr Opin Neurobiol 18, 321-331.

119

Larkum, M.E., Zhu, J.J., and Sakmann, B. (1999). A new cellular mechanism for coupling inputs arriving at different cortical layers. Nature 398, 338-341. Lee, W.C., Chen, J.L., Huang, H., Leslie, J.H., Amitai, Y., So, P.T., and Nedivi, E. (2008). A dynamic zone defines interneuron remodeling in the adult neocortex. Proc Natl Acad Sci U S A 105, 19968-19973. Lee, W.C., Huang, H., Feng, G., Sanes, J.R., Brown, E.N., So, P.T., and Nedivi, E. (2006). Dynamic remodeling of dendritic arbors in GABAergic interneurons of adult visual cortex. PLoS Biol 4, e29. LeVay, S. (1973). Synaptic patterns in the visual cortex of the cat and monkey. Electron microscopy of Golgi preparations. J Comp Neurol 150, 53-85. Lewis, D.A., Hashimoto, T., and Volk, D.W. (2005). Cortical inhibitory neurons and schizophrenia. Nat Rev Neurosci 6, 312-324. Li, S., Liu, N., Zhang, X.H., Zhou, D., and Cai, D. (2014). Bilinearity in spatiotemporal integration of synaptic inputs. PLoS Comput Biol 10, e1004014. Lois, C., Hong, E.J., Pease, S., Brown, E.J., and Baltimore, D. (2002). Germline transmission and tissue-specific expression of transgenes delivered by lentiviral vectors. Science 295, 868-872. Lu, Y.M., Mansuy, I.M., Kandel, E.R., and Roder, J. (2000). Calcineurin-mediated LTD of GABAergic inhibition underlies the increased excitability of CA1 neurons associated with LTP. Neuron 26, 197-205. Marsden, K.C., Shemesh, A., Bayer, K.U., and Carroll, R.C. (2010). Selective translocation of Ca2+/calmodulin protein kinase IIalpha (CaMKIIalpha) to inhibitory synapses. Proc Natl Acad Sci U S A 107, 20559-20564. Marty, S., Berzaghi Mda, P., and Berninger, B. (1997). Neurotrophins and activity-dependent plasticity of cortical interneurons. Trends Neurosci 20, 198-202. Maya Vetencourt, J.F., Sale, A., Viegi, A., Baroncelli, L., De Pasquale, R., O'Leary, O.F., Castren, E., and Maffei, L. (2008). The antidepressant fluoxetine restores plasticity in the adult visual cortex. Science 320, 385-388. Megias, M., Emri, Z., Freund, T.F., and Gulyas, A.I. (2001). Total number and distribution of inhibitory and excitatory synapses on hippocampal CA1 pyramidal cells. Neuroscience 102, 527-540. Miles, R., Toth, K., Gulyas, A.I., Hajos, N., and Freund, T.F. (1996). Differences between somatic and dendritic inhibition in the hippocampus. Neuron 16, 815-823. Mizrahi, A., and Katz, L.C. (2003). Dendritic stability in the adult olfactory bulb. Nat Neurosci 6, 1201-1207. Mohler, H., Fritschy, J.M., Crestani, F., Hensch, T., and Rudolph, U. (2004). Specific GABA(A) circuits in brain development and therapy. Biochem Pharmacol 68, 1685-1690.

120

Mullner, F.E., Wierenga, C.J., and Bonhoeffer, T. (2015). Precision of Inhibition: Dendritic Inhibition by Individual GABAergic Synapses on Hippocampal Pyramidal Cells Is Confined in Space and Time. Neuron 87, 576-589. Nelson, S.B., and Turrigiano, G.G. (2008). Strength through diversity. Neuron 60, 477-482. Nimchinsky, E.A., Sabatini, B.L., and Svoboda, K. (2002). Structure and function of dendritic spines. Annu Rev Physiol 64, 313-353. Parnavelas, J.G., Sullivan, K., Lieberman, A.R., and Webster, K.E. (1977). Neurons and their synaptic organization in the visual cortex of the rat. Electron microscopy of Golgi preparations. Cell Tissue Res 183, 499-517. Peng, H., Ruan, Z., Long, F., Simpson, J.H., and Myers, E.W. (2010). V3D enables real-time 3D visualization and quantitative analysis of large-scale biological image data sets. Nat Biotechnol 28, 348-353. Perez-Garci, E., Gassmann, M., Bettler, B., and Larkum, M.E. (2006). The GABAB1b isoform mediates long-lasting inhibition of dendritic Ca2+ spikes in layer 5 somatosensory pyramidal neurons. Neuron 50, 603-616. Peters, A. (2002). Examining neocortical circuits: some background and facts. J Neurocytol 31, 183-193. Pieraut, S., Gounko, N., Sando, R., 3rd, Dang, W., Rebboah, E., Panda, S., Madisen, L., Zeng, H., and Maximov, A. (2014). Experience-dependent remodeling of basket cell networks in the dentate gyrus. Neuron 84, 107-122. Poirazi, P., and Mel, B.W. (2001). Impact of active dendrites and structural plasticity on the memory capacity of neural tissue. Neuron 29, 779-796. Pouille, F., and Scanziani, M. (2001). Enforcement of temporal fidelity in pyramidal cells by somatic feed-forward inhibition. Science 293, 1159-1163. Prange, O., and Murphy, T.H. (2001). Modular transport of postsynaptic density-95 clusters and association with stable spine precursors during early development of cortical neurons. J Neurosci 21, 9325-9333. Rubenstein, J.L., and Merzenich, M.M. (2003). Model of autism: increased ratio of excitation/inhibition in key neural systems. Genes, brain, and behavior 2, 255-267. Rubinson, D.A., Dillon, C.P., Kwiatkowski, A.V., Sievers, C., Yang, L., Kopinja, J., Rooney, D.L., Zhang, M., Ihrig, M.M., McManus, M.T., et al. (2003). A lentivirus-based system to functionally silence genes in primary mammalian cells, stem cells and transgenic mice by RNA interference. Nat Genet 33, 401-406. Rui, Y., Myers, K.R., Yu, K., Wise, A., De Blas, A.L., Hartzell, H.C., and Zheng, J.Q. (2013). Activity-dependent regulation of dendritic growth and maintenance by glycogen synthase kinase 3beta. Nat Commun 4, 2628.

121

Rutherford, L.C., DeWan, A., Lauer, H.M., and Turrigiano, G.G. (1997). Brain-derived neurotrophic factor mediates the activity-dependent regulation of inhibition in neocortical cultures. J Neurosci 17, 4527-4535. Sans, N., Petralia, R.S., Wang, Y.X., Blahos, J., 2nd, Hell, J.W., and Wenthold, R.J. (2000). A developmental change in NMDA receptor-associated proteins at hippocampal synapses. J Neurosci 20, 1260-1271. Sato, M., and Stryker, M.P. (2008). Distinctive features of adult ocular dominance plasticity. J Neurosci 28, 10278-10286. Schiller, J., Schiller, Y., Stuart, G., and Sakmann, B. (1997). Calcium action potentials restricted to distal apical dendrites of rat neocortical pyramidal neurons. J Physiol 505 ( Pt 3), 605-616. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., et al. (2012). Fiji: an open-source platform for biological-image analysis. Nat Methods 9, 676-682. Schuemann, A., Klawiter, A., Bonhoeffer, T., and Wierenga, C.J. (2013). Structural plasticity of GABAergic axons is regulated by network activity and GABAA receptor activation. Frontiers in neural circuits 7, 113. Schwarz, G., Schrader, N., Mendel, R.R., Hecht, H.J., and Schindelin, H. (2001). Crystal structures of human gephyrin and plant Cnx1 G domains: comparative analysis and functional implications. J Mol Biol 312, 405-418. Sjostrom, P.J., Rancz, E.A., Roth, A., and Hausser, M. (2008). Dendritic excitability and synaptic plasticity. Physiol Rev 88, 769-840. Song, S., Miller, K.D., and Abbott, L.F. (2000). Competitive Hebbian learning through spike-timing-dependent synaptic plasticity. Nat Neurosci 3, 919-926. Spolidoro, M., Sale, A., Berardi, N., and Maffei, L. (2009). Plasticity in the adult brain: lessons from the visual system. Exp Brain Res 192, 335-341. Spruston, N. (2008). Pyramidal neurons: dendritic structure and synaptic integration. Nature Reviews Neuroscience 9, 206-221. Stokes, C.C., Teeter, C.M., and Isaacson, J.S. (2014). Single dendrite-targeting interneurons generate branch-specific inhibition. Frontiers in neural circuits 8, 139. Stoppini, L., Buchs, P.A., and Muller, D. (1991). A simple method for organotypic cultures of nervous tissue. J Neurosci Methods 37, 173-182. Subramanian, J., Dye, L., and Morozov, A. (2013). Rap1 signaling prevents L-type calcium channel-dependent neurotransmitter release. J Neurosci 33, 7245-7252. Tabata, H., and Nakajima, K. (2001). Efficient in utero gene transfer system to the developing mouse brain using electroporation: visualization of neuronal migration in the developing cortex. Neuroscience 103, 865-872.

122

Thomson, A.M., and Deuchars, J. (1997). Synaptic interactions in neocortical local circuits: dual intracellular recordings in vitro. Cereb Cortex 7, 510-522. Trachtenberg, J.T., Chen, B.E., Knott, G.W., Feng, G., Sanes, J.R., Welker, E., and Svoboda, K. (2002). Long-term in vivo imaging of experience-dependent synaptic plasticity in adult cortex. Nature 420, 788-794. Trommald, M., Jensen, V., and Andersen, P. (1995). Analysis of dendritic spines in rat CA1 pyramidal cells intracellularly filled with a fluorescent dye. J Comp Neurol 353, 260-274. Turrigiano, G. (2011). Too many cooks? Intrinsic and synaptic homeostatic mechanisms in cortical circuit refinement. Annu Rev Neurosci 34, 89-103. Turrigiano, G.G., and Nelson, S.B. (2004). Homeostatic plasticity in the developing nervous system. Nat Rev Neurosci 5, 97-107. Tyagarajan, S.K., and Fritschy, J.M. (2014). Gephyrin: a master regulator of neuronal function? Nat Rev Neurosci 15, 141-156. Tyagarajan, S.K., Ghosh, H., Yevenes, G.E., Imanishi, S.Y., Zeilhofer, H.U., Gerrits, B., and Fritschy, J.M. (2013). Extracellular signal-regulated kinase and glycogen synthase kinase 3beta regulate gephyrin postsynaptic aggregation and GABAergic synaptic function in a calpain-dependent mechanism. J Biol Chem 288, 9634-9647. Tyagarajan, S.K., Ghosh, H., Yevenes, G.E., Nikonenko, I., Ebeling, C., Schwerdel, C., Sidler, C., Zeilhofer, H.U., Gerrits, B., Muller, D., et al. (2011). Regulation of GABAergic synapse formation and plasticity by GSK3beta-dependent phosphorylation of gephyrin. Proc Natl Acad Sci U S A 108, 379-384. van Versendaal, D., Rajendran, R., Saiepour, M.H., Klooster, J., Smit-Rigter, L., Sommeijer, J.P., De Zeeuw, C.I., Hofer, S.B., Heimel, J.A., and Levelt, C.N. (2012). Elimination of inhibitory synapses is a major component of adult ocular dominance plasticity. Neuron 74, 374-383. Westrum, L.E. (1966). Synaptic contacts on axons in the cerebral cortex. Nature 210, 1289-1290. White, E.L., and Rock, M.P. (1980). Three-dimensional aspects and synaptic relationships of a Golgi-impregnated spiny stellate cell reconstructed from serial thin sections. J Neurocytol 9, 615-636. Wierenga, C.J., Becker, N., and Bonhoeffer, T. (2008). GABAergic synapses are formed without the involvement of dendritic protrusions. Nat Neurosci 11, 1044-1052. Willadt, S., Nenniger, M., and Vogt, K.E. (2013). Hippocampal feedforward inhibition focuses excitatory synaptic signals into distinct dendritic compartments. PLoS One 8, e80984. Woodruff, A.R., Anderson, S.A., and Yuste, R. (2010). The enigmatic function of chandelier cells. Front Neurosci 4, 201. Woodruff, A.R., McGarry, L.M., Vogels, T.P., Inan, M., Anderson, S.A., and Yuste, R. (2011). State-dependent function of neocortical chandelier cells. J Neurosci 31, 17872-17886.

123

Woods, G., and Zito, K. (2008). Preparation of gene gun bullets and biolistic transfection of neurons in slice culture. J Vis Exp. Woods, G.F., Oh, W.C., Boudewyn, L.C., Mikula, S.K., and Zito, K. (2011). Loss of PSD-95 enrichment is not a prerequisite for spine retraction. J Neurosci 31, 12129-12138. Wuchter, J., Beuter, S., Treindl, F., Herrmann, T., Zeck, G., Templin, M.F., and Volkmer, H. (2012). A comprehensive small interfering RNA screen identifies signaling pathways required for gephyrin clustering. J Neurosci 32, 14821-14834. Zhang, D., Li, Y., Rasch, M.J., and Wu, S. (2013). Nonlinear multiplicative dendritic integration in neuron and network models. Front Comput Neurosci 7, 56. Zita, M.M., Marchionni, I., Bottos, E., Righi, M., Del Sal, G., Cherubini, E., and Zacchi, P. (2007). Post-phosphorylation prolyl isomerisation of gephyrin represents a mechanism to modulate glycine receptors function. EMBO J 26, 1761-1771.