Chapter 4 Electron Transfer in Respiratory Complexes Resolved by an Ultra-Fast Freeze-Quench...

Provided for non-commercial research and educational use only. Not for reproduction, distribution or commercial use.

This chapter was originally published in the book Methods in Enzymology, Vol. 456, published by Elsevier, and the attached copy is provided by Elsevier for the author's benefit and for the benefit of the author's institution, for non-commercial research and educational use including without limitation use in instruction at your institution, sending it to specific colleagues who know you, and providing a copy to your institution’s administrator.

All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution’s website or repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier's permissions site at:

http://www.elsevier.com/locate/permissionusematerial

From: Nikolai P. Belevich, Marina L. Verkhovskaya, and Michael I. Verkhovsky, Electron Transfer in Respiratory Complexes Resolved by an Ultra-Fast Freeze-

Quench Approach. In William S. Allison and Immo E. Scheffler, editors: Methods in Enzymology, Vol. 456,

Burlington: Academic Press, 2009, pp. 75-93. ISBN: 978-0-08-087776-1

© Copyright 2009 Elsevier Inc. Academic Press.

Author's personal copyAuthor's personal copy

C H A P T E R F O U R

Methods

ISSN 0

Institute

Electron Transfer in Respiratory

Complexes Resolved by an Ultra-Fast

Freeze-Quench Approach

Nikolai P. Belevich, Marina L. Verkhovskaya, and

Michael I. Verkhovsky

Contents

1. In

in

076

of

troduction

Enzymology, Volume 456 # 2009

-6879, DOI: 10.1016/S0076-6879(08)04404-2 All rig

Biotechnology, University of Helsinki, Helsinki, Finland

Else

hts

76

1
.1. T ransient kinetics as the main tool for enzyme
catalytic cycle investigation
76
1
.2. C omparison of the different flow techniques 76
2. U
ltra-Fast Freeze-Quench Setup 79
2
.1. M ixer design 80
2
.2. F ast freezing 81
3. S
ample Preparation 82
4. C
alibration of the Freeze-Quench Setup 85
5. A
pplication of Freeze-Quench Approach for the Monitoring of
Complex I Reduction by NADH
85
Refe
rences 92
Abstract

The investigation of the molecular mechanism of the respiratory chain com-

plexes requires determination of the time-dependent evolution of the catalytic

cycle intermediates. The ultra-fast freeze-quench approach makes possible

trapping such intermediates with consequent analysis of their chemical struc-

ture by means of different physical spectroscopic methods (e.g., EPR, optic, and

Mossbauer spectroscopies). This chapter presents the description of a setup

that allows stopping the enzymatic reaction in the time range from 100 msec totens of msec. The construction and production technology of the mixer head,

ultra-fast freezing device, and accessories required for collecting a sample are

described. Ways of solving a number of problems emerging on freezing of the

reaction mixture and preparing the samples for EPR spectroscopy are proposed.

The kinetics of electron transfer reaction in the first enzyme of the respiratory

vier Inc.

reserved.

75

76 Nikolai P. Belevich et al.

Author's personal copy

chain, Complex I (NADH: ubiquinone oxidoreductase), is presented as an illus-

tration of the freeze-quench approach. Time-resolved EPR spectra indicating the

redox state of FeS clusters of the wild-type and mutant (R274A in subunit

NuoCD) Complex I from Escherichia coli are shown.

1. Introduction

1.1. Transient kinetics as the main tool for enzymecatalytic cycle investigation

The investigation of the mechanism of any complex reaction requires deter-mination of the time-dependent evolution of the molecular species involved.The typical turnover rate of the respiratory chain complexes under steady-stateconditions is approximately 100 to 200 enzyme revolutions per second. Eachturnover consists of a number of distinct intermediates, which are formed in atime range frommicroseconds to several milliseconds after the reaction initia-tion. The large number of catalytic intermediates, complicated by a hierarchyof transition rate constants between thembecause the transition ratesmay differby several orders of magnitude, makes it practically impossible to study themolecular mechanism by means of conventional steady-state techniques.Under steady-state conditions only the slowest intermediate is populated,leaving all other intermediates completely invisible. The transient kineticsapproach allows one to overcome the stated difficulties; this is why it is oneof themost fruitfulways to investigate the reactionmechanism of enzymes thatare considered as molecular machines. At the same time, such an approachrequires certain conditions that must be fulfilled, such as

1. The initial state of all the enzyme molecules in the sample should beidentical before the reaction initiation.

2. The start of the reaction (addition of substrate, inhibitor removal, etc.)has to be carried out faster than the fastest transition under the investiga-tion. This is a real challenge, because the fastest mixing time of aconventional stopped-flow apparatus is in the range of milliseconds,which is much slower than most of the intermediate transitions in thecatalytic cycle. During the last few years ultra-fast mixing devices basedon very low volumes of the mixing chambers (�1 nl) have been devel-oped (Cherepanov and de Vries, 2004, Lin et al., 2003, Tanaka et al.,2003, Verkhovskaya et al., 2008).

1.2. Comparison of the different flow techniques

Historically, the earliest rapid kinetic measurements used a continuous-flow(CF) arrangement combined with absorbance measurements of the reactionprogress at different points downstream. The drawback of this method is

Ultra-Fast Freeze-Quench 77


significant reagent consumption. This has been replaced by the morecommonly used stopped-flow (SF) method because of sample economyand the possibility to measure the kinetics over a relatively long time period(up to several minutes). However, the kinetics resolution time accessible toSF is generally limited to 1 msec or longer, because the measurements aremade only after the flow comes to a complete stop, which is inherently slowand can result in vibration and pressure artifacts (Roder et al., 2004). CFtechniques have undergone a renaissance in recent years because of advancesin mixer design and detection methods, which made it possible to push thetime resolution into the microsecond time range (Bilsel et al., 2005, Shastryet al., 1998). For the reaction, which could not be investigated by opticalspectroscopy, freeze-quench protocols can be used in combination withslower analytical techniques such as NMR, EPR, or mass spectrometry( Johnson, 1995).

In general, the freeze-quench mixing-and-sampling technique can beviewed as a three-stage process. At the first stage, the reaction is initiated byrapid mixing of the reaction components. At the second stage, the sample istransported from the mixer to the freeze-quenching device. And at the thirdstage, the sample undergoes fast freezing. Thus, the characteristic reactiontime, tr, can be described as

tr ¼ tm þ tt þ tf ð4:1Þwhere tm is the mixing time, tt is the sample-transport time, and tf is thefreezing time.

An ideal freeze-quenching instrument should have a tm þ tf value closeto zero, and in this case a reaction occurs mostly during the second stage.The sample-transport time tt is determined by the sample flow rate and thedistance between the mixer and freeze-quenching device and could bevaried during the experiment to follow the progress of a reaction. Theconventional freeze-quenching devices have minimal tmþ tf (often definedas the dead time of a setup) approximately 5 msec (Ballou, 1978), which isnot sufficient for studying most enzymatic reactions; therefore, attempts toachieve better kinetic resolution for this method concentrated on the mixerdesign, and improvements were made to the freezing technique.

The mixing time is determined by the sample flow rate, the effectivevolume of the mixer and the mixer geometry. Reduction in the mixingtime can be achieved by decreasing the dimensions of the mixing chamber.However, this advantage is compromised by a concurrent decrease in theReynolds number (Re), which characterizes the dynamics of a liquid flowinside a tube with a circular cross section. The Reynolds number is defined asRe¼ rvd/�, where v is the average flow speed, d represents the smallest crosssection in the flow channel, and r and � are the density and the viscosityof the liquid. When the Reynolds number is low, a laminar flow results.



A Reynolds number greater than 2000 is characteristic of the turbulent flowrequired for efficient mixing. To reach this condition and prevent a laminarflow in the miniaturized mixers an increase in flow speed is typically required.Shastry et al. (1998) described a device with small mixing volume and highflow rate. It is based on the design by Regenfuss et al. (1985) and uses aprinciple of coaxial reagents flow combined with ball mixing (Berger, 1978).In this device they obtained full reagents mixing within 15 msec. The mixerwas built of quartz and consists of several fine inner parts. The complicateddesign of this mixer makes it fragile for use with freeze-quenching techniquesand also it cannot easily be replicated.

Lin et al. (2003) described an ultra-fast micro fluidic mixer made ofsilicon with the use of photolithographic techniques. Its mixing chamberhas the dimensions of 50� 100� 100 mm and contains seven vertical pillarswith a diameter of 10 mm arranged perpendicularly to the flow direction in astaggered fashion. With this design, the flow velocity is modulated by thealternating passages as the mixed fluid travels down the mixing chamber.The associated changes in the Reynolds number fall within the 200 to 2000range along the flow direction. Lin et al. (2003) claim that the mixing timefor the silicon mixer is approximately 20 msec. For the sample freezing theyused a new freeze-quench device, which consisted of two rotating copperwheels maintained at 77 K. A free-flowing jet exits from the mixer andinstantly freezes on the wheels’ surface. The frozen sample is subsequentlyground into an ultra fine powder. The use of a silicone mixer in combina-tion with a new freeze-quench device enabled them to reach a timeresolution of 50 msec, although for calibration they used an uncorrectedrate constant (Cherepanov and de Vries, 2004) for azide binding tometmyoglobin, and real-time resolution is not so high.

Cherepanov and de Vries (2004) reported a freeze-quench instrumentwith a dead time of approximately 120 msec. They designed a tangentialmicro-mixer, which had a mixing time of up to 20 msec. The mixer iscomposed of four channels arranged in an X-form geometry. The oppositechannels have an offset with respect to each other by the half-width of thechannel, so that the liquid streams are premixed in the middle of ‘‘X’’ with aresulting angular momentum at relatively low Reynolds numbers (Re <1000). After premixing, the four-stream solution is forced through a narroworifice and forms a free jet in a direction perpendicular to the plane of thechannels. The flow through the orifice is accelerated to a high linearvelocity, and the mixing of the reaction components takes place atReynolds numbers of approximately 4000. For the freeze-quenching thesample jet was sprayed into the vial containing cold isopentane at 125 K.

Each of these two instruments has advantages and limitations. Thesample freezing device designed by Lin et al. (2003) is evidently moreefficient, because the metals have very high thermal conductivity, and thejet directed at the cold metal surface should freeze very fast. The fast rotation



of the wheels guarantees that the jet will have contact with an ice-free metalsurface. On the contrary, freezing the jet by spraying it into cold isopentaneshould not be so fast, because the thermal conductivity of the isopentane atroom temperature is approximately 4000 times worse than that of copper(0.11 and 401 W m�1 K� 1 , respectively).

The four jets tangential mixer designed by Cherepanov and de Vries(2004) seems to be efficient in a wider flow rate range than that presented byLin et al. (2003). This conclusion is based on the general criteria for themixer design ( Smith, 1973); in particular, the efficiency of mixing dependson the ability of the mixer to break up the streams of the two reactants intoelementary blocks of the smallest size.

In our freeze-quench system we used a mixer based on the design byCherepanov and de Vries (2004) and combined it with the freeze-quenchdevice described by Lin et al. (2003). The mixing chamber has a volume of1 nl (dimensions of 0.15 mm � 0.15 mm � 0.05 mm), which allows us toapproach the tm time of �20 msec at the typical flux rate of the experimentof 3 ml/min. The freezing time tf was minimized by an application of liquidnitrogen cooled silver—material with the highest possible thermalconductivity.

2. Ultra-Fast Freeze-Quench Setup

The freeze-quench system was built in our laboratory based on theprinciples described by Cherepanov and de Vries (2004) and Lin et al. (2003).A block diagram of the system is depicted in Fig. 4.1. It has a symmetric flowdesign and consists of two LKB 2150 high-performance liquid chromatogra-phy (HPLC) pumps (2a,b), two mechanically coupled Rheodyne 7125sample injectors (4a,b) equipped with 100-ml sample loops (5a,b), and atangential four-jet mixer (8). To increase pressure stability, pulse suppressorswere included into the flow path (3a,b). A pulse suppressor consists of asequentially connected pressure damper (LP-21 LO-Pulse, Scientific System)and a short HPLC reverse-phase column. The use of a column allowed us toget an optimal loading for the HPLC pump to improve pressure stability. Toprotect the mixer orifice from clogging, filters were inserted into the flowpath (6a,b). The filters consisted of two sequential in-line solvent filters 2 mmand 0.5 mm. All fluid transport lines were equipped with HPLC-grade 1/1600PEEK tubing (inner diameter 0.0100) and standard high-pressure HPLCconnections. Sample loops were made of PEEK tubes with inner diameter0.0500. To decrease sample dilution, the flow path after the sample injectorswas designed to be as short as possible.

The mixing device works in continuous-flow mode. In the purgingmode, the buffer (1, Fig. 4.1), which bypasses the sample loops, comes

3a 3b

4a

5a

6a 6b

4b

8

7

910

WASTE

11

1

Controlunit

5b

2a 2b

Figure 4.1 Schematic representation of the freeze-quench setup.1,The vessel with theargon saturated carrier solvent; 2a,b, HPLC pumps; 3a,b, pressure pulsation dampers;4a,b, Rheodyne-type sample injection valves; 5a,b, sample loops; 6a,b, solvent filters; 7,microcontroller-based timing control system; 8, tangential four-jet mixer; 9, low tem-perature quenching system; 10, waste container; 11, table with vertically adjustableposition.



through the mixer, and a free-flowing jet is directed at the waste container(10, Fig. 4.1). Switching mechanically coupled sample injectors triggers theoperational mode. At this moment, the sample loops filled with assaysolutions are inserted into the flow path, and a microcontroller-based timingsystem (7, Fig. 4.1) generates the time delay necessary for assay solutions toreach the mixer head. Then, the mixer head moves into the ‘‘work’’position, where the jet of the mixed solutions is directed at the freeze-quench device (FQD). At this point, the system delivers a volume equal tothe volume of both sample loops and then returns the mixing head back tothe ‘‘waste’’ position. Timing delays depending on tubing length andsolvent flow rate can be adjusted with millisecond precision.

2.1. Mixer design

The tangential four-jet mixer is shown in Fig. 4.2 (see also Cherepanov andde Vries [2004]). The mixer body is made of a stainless-steel cylinder with a6-mm outer diameter. Four holes (diameter, 1.5 mm) were drilled along thecylinder axis. Then stainless-steel tubes with 1/1600 outer and 0.00500 innerdiameters were pressed into the cylinder to create flow channels. The top

1

A A

BB

2

3

Figure 4.2 Picture and schematic drawing of the tangential four-jet mixer. 1, Stainless-steel body of the mixer with stainless steel HPLC tubing pressed in. 2, 50-mm-thickstainless-steel cover with the orifice. A and B are the entry for corresponding reactioncomponents.3,Brass holderof themixer fastenedbyscrews.Y-connectorsdistributingtworeactants through the fourchannels arranged incross geometry are seen in the picture.Thechannels etched on the surface are offset with respect to the opposite one for half-width ofthechannel.



surface of the cylinder was ground and polished. In addition, the tubes on theopposite side of the cylinder were caulked by soldering. On the top of thecylinder, crosslike channels weremade by electrochemical etching. Thewidthof the channel was �100 mm and the depth�50 mm. On the top edge of thecylinder four asymmetric guiding grooves were made. These grooves serve asguides for the exact positioning of the stainless-steel cover on the mixer body.This cover with an orifice is placed on the top of the mixer body to form amixing chamber. The cover was stamped out of a 50-mm stainless-steel sheet,and the mixer body serves as a punch head. After stamping, the cover has ashape that exactly fits themixer body surface.On the inner surface of the coveran imprint of the cross-like channels can be obtained.This imprint lets us locatean orifice position in the center of the crosslike channels. The orifice hole wasdrilled by electrochemical micromachining technology (Schuster et al., 2000).A number of covers were manufactured with an orifice diameter in the range20 to 30 mm. The mixer body and cover were placed into a brass holder andfastened by screws (Fig. 4.2).

2.2. Fast freezing

The freeze-quench device (9, Fig. 4.1) was built based on the principlesdescribed by Lin et al. (2003) and Tanaka et al. (2003). It consists of a stainless-steelDewar,drumunit, anda samplecollecting funnel attachedto anEPRtube.

The drum unit consists of two wheels and a driving motor (Fig. 4.3).The wheels are made of pure (99.99%) silver and have a cylindric shape witha 4-cm diameter. They are arranged in a side-by-side fashion. The driving

1

2

3

4

Figure 4.3 Drumunit of the freeze-quench device.1, Silver cylinders; 2, motor driver;3, lever for lifting the collecting funnel with the scrapers onto the cylinders’ surfaceduring collection of the samples; 4, liquid nitrogen refilling tube.



wheel, attached to a motor, drives the slave wheel through frictional contact.As a result, twowheels rotate in opposite directions at the same speed.Duringthe experiment the drum unit is placed on top of the Dewar filled with liquidnitrogen in such a way that the bottom halves of both rotating wheels areimmersed in liquid nitrogen to maintain low temperature. For sufficientcooling down of the wheels, a 10-min rotation is required before theexperiment together with constant refilling of the Dewar with liquid nitro-gen. For sample cryofixation, the mixed-solution jet is directed at the coldsurface of the rotating cylinder, where it freezes instantly. To improvefreezing efficiency, the mixer head executes a penduliform motion alongthe cylinder axis and directs the jet at the ice-free zone of the cylinder. Therotating cylinders grind the ice into a fine powder as it is carried downthrough the cylinder interface. The powder is collected into an EPR tubethrough a collecting funnel placed in liquid nitrogen directly below thecylinders. The top of the rectangular funnel is equipped with two scrapers.In the operationalmode, the scrapersmake contactwith the cylinder’s surfaceand improve the efficiency of the powder collection. An EPR tube (5-mmdiameter) is attached to the funnel by a heat-shrinkable connection.

3. Sample Preparation

Reaction species evolution could be monitored by changing thetransportation time tt (Eq. 4.1). This parameter is the function of thedistance between the mixer and the cooled wheel surface of the FQD.



Therefore, to obtain kinetics of the reaction, a number of samples withdifferent distances between the mixer and FQD have to be collected. Thefastest time points in the kinetics were collected with a jet speed of �80m/sec for a 28-mm orifice (flow speed 1.5 ml/min for each HPLC pump)and the distance varied from 8 to 32 mm. For all other samples the jet speedwas decreased to 26 m/sec (flow speed 0.5 ml/min per HPLC pump). Allreagent solutions and buffers were degassed on a vacuum/gas line before theexperiment. The two sample loops were filled with 100-ml solutions of thereactants under investigation by a syringe equipped with Millex-HV 0.45-mm filter. The sample collection procedure consisted of the following. Arectangle funnel with attached EPR tube was placed inside a Dewar filledwith liquid nitrogen. A dried-out drum unit was placed on top of theDewar. The FQD was placed on a vertically adjustable table (11,Fig. 4.1), and the distance between the mixer head and the drum unit wasset to a desirable value. Then rotation of the wheels was initiated, and theywere allowed to cool down for approximately 10 min. During cooling, thelevel of liquid nitrogen in theDewarwas kept constant by continuous refilling.After cooling, injection of the sample and the cryofixation procedure wereinitiated. When the desired volume of the frozen sample was ground into apowder and the mixer head had returned to the waste position (Fig. 4.1), thedrum unit was removed from the top of the Dewar. The funnel connected tothe EPR tube filled with the powdered, frozen sample and immersed in liquidnitrogenwas slowly lifted to allow evaporation of the liquid nitrogen and to letthe powder settle inside the EPR tube. This procedure has to be performedwith great care to prevent sample blowout caused by nitrogen boiling inside theEPR tube. A stick made of copper was used to accelerate the evaporation ofliquid nitrogen from the EPR tube. During this procedure, the low end of thecopper stick was manually placed at the nitrogen gas/liquid interface tointensify nitrogen evaporation. After the powder settled in the EPR tube,the funnel was removed.

Two possible complications exist with the liquefied gases in the sample.The first complication arises if all the liquid nitrogen is not completelyremoved from the sample powder. The powder, acting as a sponge, absorbsliquid very well. A small temperature rise during the transfer of the EPRtube to the cavity of the EPR spectrometer will blow the sample because ofthe immediate liquid to gas transition of the nitrogen in the sample. Thesecond complication comes from oxygen. The boiling temperature ofoxygen (90 K) is higher than that of nitrogen (77 K). Therefore, duringthe preparation of the sample, oxygen in the air could condense in liquidform inside the EPR tube. Oxygen as a biradical has a broad EPR spectrumthat could overlap the studied signals. To avoid these two complications thesample was subjected to an annealing procedure to remove all traces ofliquid gases. For annealing, the EPR tube was placed in a low-temperaturechamber (Fig. 4.4), where it was incubated for 15 min at 100 K to evaporate

1

3

2

4

6

5

Figure 4.4 The annealing device consists of a thermal insulated Styrofoam chamber(1) designed as a lid,which is placed on topof the stainless-steel Dewarwith liquid nitro-gen (2). The passive ‘‘cold’’ exchanger made of copper (3) maintains the temperatureinside the annealing chamber at approximately100 K. An additional heating element (4)placed inside the annealing chambercoupledwith a heatcontroller lets us use this deviceas a low temperature thermostat with a temperature range of 100 to 240 K. A safetyplunger (6) placed inside the EPR tube (5) protects the sample against blowout duringannealing.



any traces of liquid nitrogen and oxygen. After incubation, the sample wastamped down by a polycarbonate stick cooled with liquid nitrogen tocompletely pack it in the EPR tube with a packing factor of 0.5 to 0.6.The packing factor was calculated as a ratio between the volumes of the samplein the liquid and frozen powder states. As a final step in the sample preparationa plug made of cotton wool was inserted into the EPR tube above the samplepowder to ensure that traces of liquefied gases would not destroy the sample.Then the EPR tube, whichwas tightly closedwith a plastic lid, was transferredinto an EPR spectrometer cavity. This is a very important point. If the EPRtubes were to remain open, oxygen from the atmosphere would liquefy insidethe tube located in the helium cryostat at temperatures far below 100 K, andthe problems described previously would occur.



4. C alibr ation of the Freeze-Quench Setup

The mixing and freezing time of an ideal freeze-quench instrumentcan be assumed to be zero, and the sample-aging time can be calculated byuse of the distance between the mixer and freeze-quench device and the jetspeed. However, the following complications should be considered. First,the speed of the free-flowing jet could slow down because of air resistancethat results in an increase in sample transportation time. The braking effectof air becomes significant with acceleration of the jet speed. For example, tominimize air-braking Cherepanov and de Vries (2004) used a low-pressurechamber. Therefore, all freeze-quench instruments require direct calibra-tion to obtain a reliable sample-aging time.

To calibrate a freeze-quench system, a well-known chemical reaction isused as a molecular timer. For example, the binding of azide (N3

�) with themetmyoglobin from horse heart often is used for this purpose. (Ballou,1978, Cherepanov and de Vries, 2004, Lin et al., 2003). It is important tonote that azide binding is not a simple second-order reaction, because therate constant decreases at high azide concentrations (>1 mM ). Thesepeculiar properties of azide binding should be taken into account. A detailedanalysis of this reaction can be found in Cherepanov and de Vries (2004).We also used the binding of azide to metmyoglobin to calibrate the timescale of our instrument. The rate constants for this reaction obtained on thebasis of timing by jet length were found to be very close to those calculatedfrom the known parameters of the reaction. The obtained kinetics ofmetmyoglobin ligand binding (Fig. 4.5) allowed estimation of the deadtime, which represents the sum of mixing and freezing times. The zerotime point (100% of the signal) was obtained by mixing metmyoglobin withthe buffer without azide. The experimental points indicating the content oflow- and high-spin heme in the sample were plotted against the timeintervals calculated from the jet speed and the distance between the mixerhead and cold silver surface (Fig. 4.5). The theoretical curve, which fitsthese experimental data, represents the reaction with a rate constant of 19500 M�1 s�1 (Cherepanov and de Vries, 2004) and a dead time of 57 msec.

5. Application of Freeze-Quench Approach

for the Monitoring of Complex I Reduction

by NADH

NADH:ubiquinone oxidoreductase (Complex I) is one of the keyenzymes of the respiratory chain of mitochondria and many bacteria.Complex I couples electron transfer from NADH to ubiquinone with

100

80

60

40

20

00 1 2 3

Time, ms

High spin

Low spin

4 5

Sign

al a

mpl

itud

e, %

Figure 4.5 Kinetics of the NaN3 binding to the metmyoglobin. Metmyoglobin fromhorse heart, 1 mM, was mixed 1:1with 80 mM sodium azide (pH 5.0).The reaction wasstopped in the time range from 0.09 to 5 msec, and the kinetics of the metmyoglobinligand bindingwas followedbyX-band (9.4GHz)EPRmeasurements (10K,microwavepower 2 mW).The appearance of low-spin heme and decay of high-spin heme spectraweremonitored.



translocation of 2 H+/e� across the membrane (Galkin et al., 1999,Wikstrom, 1984). Although the enzyme has been under investigation formore than 40 years, its molecular mechanism is not yet known. Complex Iconsists of water soluble and membrane domains; the former contains FMNfollowed by a chain of 8 to 9 FeS clusters, and no redox centers are found inthe latter. So far, Complex I has not been accessible to time-resolvedmethods, because its catalytic reaction (kcat ca. 500 sec�1) is too fast to becaptured by conventional techniques. Because multiple FeS clusters ofComplex I have broad overlapped and unresolved spectra with low extinc-tion coefficient, it was impossible to follow the redox states of particularclusters in the catalytic cycle of Complex I by means of optical spectroscopy.EPR spectroscopy reveals the individual signals of FeS clusters ( see Ohnishi[1998] for a review), therefore, we investigated electron transfer in Com-plex I from Escherichia coli by use of an ultra-fast freeze-quench approach,which allowed us to follow electron distribution in the chain of FeS clustersduring the catalytic cycle.

Complex I was purified from aerobically grown bacterial cells, asdescribed in (Sinegina et al., 2005). Because purified Complex I falls apartabove pH 6.0, measurements at higher pH values require treating the



enzyme with phospholipids to secure protein stability and activity duringthe experiment (Sinegina et al., 2005). Finally, the 1.1-ml volume samplecontaining Complex I at 20 to 24 mg/ml in a buffer consisting of 50 mMMES-NaOH, pH 6.0, 100 mM NaCl, 0.1% n-dodecyl b-D-maltoside, 7%sucrose, and 1.7% glycerol (nonactivated Complex I) or 20 mg/ml azolec-tin, 30 mM MES/NaOH, 70 mM HEPES/KOH, pH 7.5, 70 mM NaCl,0.19% sodium cholate, 0.1% n-dodecyl b-D-maltoside, 7% sucrose, and1.7% glycerol (activated Complex I) was degassed and saturated withargon in a vacuum/gas line and loaded into the sample syringe. The othersyringe was filled with 1.1 ml of a solution containing 200 mM disodiumNADH also saturated with argon. A high NADH concentration wasrequired to avoid possible diffusion limitations of the substrate binding.The buffer, which was degassed and saturated with argon, consisted of50 mM HEPES-Bis tris propane, pH 7.0, 100 mM NaCl, and 0.05% DM.After mixing and obtaining frozen samples by stopping the reaction within adesirable time range, X-band (9.4 GHz) EPR measurements were per-formed with a Bruker EMX EPR spectrometer, equipped with an OxfordInstruments ESR900 helium flow cryostat with an ITC4 temperaturecontroller. The field modulation frequency was 100 kHz; the modulationamplitude was 1.27 mT. The obtained EPR spectra were normalized fortemperature, gain, microwave power, and corrected for baseline.

Mechanical grinding of the frozen protein sample at cryogenic tempera-tures by silver wheels results in a rise in the artificial radical signal thatappears regardless of the redox state of Complex I and can also be observedwith other proteins (e.g., when Complex I was replaced with albumin)(Fig. 4.6). It is probable that the signal originates from amino acid radicalsformed during cryogenic braking of protein molecules in the presence ofoxygen. This proposal is supported by the observation that degassing of allthe solutions used for freeze-quench experiments and saturating them withargon strongly decrease this artificial radical signal. When radicals aregenerated in frozen samples and most probably are not well divided spa-tially, their dismutation occurs on moderate warming of the sample. There-fore, the other way to diminish artificial signals is annealing the samples byuse of the annealing device described previously. A shift from the liquidnitrogen temperature to�80 to 90� resulted in a strong decrease in artificialradical signals (Fig. 4.7). The radicals dismutation was practically accom-plished at �60�; however, at �40 to 50�, the frozen reaction of Complex Ireduction started to proceed in a time scale of the order of minutes,indicating that the correct annealing temperature should be optimized fora particular reaction.

An example of the data obtained by freeze-quench on fast reduction ofComplex I by NADH is shown in Fig. 4.8. Wild-type and mutatedComplex I, R274A, where arginine 274 in NuoCD subunit was substitutedfor alanine (Belevich et al., 2007), were used for the experiments. Six spectra

1.0

1

2

� 10−3

0.5

0

2.1 2.05 2 1.95g-value

1.9 1.85

−0.5

−1.0

−1.5

−2

Figure 4.6 The radical spectrumappearing in a frozen albumin sample aftercryogenicgrinding with silver wheels.45 K, 2mW, albumin concentration 20mg/ml.1, EPR spec-trum of the albumin frozen on the silver wheel and ground; 2, EPR spectrum of thealbumin frozen in liquidN2.



from the top to the bottom present the redox state of FeS clusters in distinctsamples where the reaction was stopped in the time range from 90 msec to3 to 8 msec. The last (seventh) spectrum, which is designated by infinitetime, was obtained from one of the previous samples that was first meltedand then immediately frozen again. On melting, Complex I was fullyreduced with NADH. Therefore, the extent of the reduction in the FeScenters in this sample was taken as 100%. Analysis of the obtained spectraallowed the determination of the kinetics of the reduction of the particularcenters, which are shown in the bottom row of Fig. 4.8 below thecorresponding experimental spectra. Complex I contains two binuclearFeS clusters, N1a and N1b; the reduction of both of these was followed at45 �K, a temperature at which the tetranuclear clusters do not contribute tothe EPR spectra. In Fig. 4.8 only the gx and gy components of N1a (gxy ¼1.95, 1.92) and N1b (gxy ¼ 1.938, 1.930) spectra are presented. Similarkinetics of [2Fe-2S] clusters reduction was observed in wild-type andmutated complex I; the N1a cluster was rapidly reduced with a characteris-tic time close to 100 msec in wild-type and 200 msec in R274, meanwhile

1

2

1–2

2.1 2.05 2 1.95g-value

1.851.9

Figure 4.7 Disappearance of the artificial radical on annealing.The reaction of Com-plex I reduction was stopped at 160 msec.The EPR spectra (45 K, 2 mW) of the samplewere obtained immediately after the sample preparation at liquid nitrogen temperature(curve1) and after 20min incubation at�90� (curve 2).The curve1-2 showing the differ-ence between two spectra above indicates the decrease in the artificial radical signal;meanwhile the redox state of Complex I (troughs at g¼1.92 and1.93) did not change.



N1b in both cases was reduced much more slowly with a characteristic timeslightly over 1 msec. The substitution of basic arginine 274 located in theNuoCD subunit at a distance of 5 A from the tetranuclear N2 cluster forneutral alanine results in disappearance of the N2 EPR signal caused byeither loss of this center (Grgic et al., 2004) or such strong decrease of itsmidpoint redox potential that it cannot be reduced by NADH (Belevichet al., 2007). The signal of N2 (gxyz ¼ 1.895, 1.904, 2.045) is the mostprominent signal from all [4Fe-4S] clusters in E. coli Complex I; it overlapsthese other signals and makes spectra analysis difficult. It is clearly seen fromFig. 4.8, where EPR spectra taken at 10 K and 10 mW are shown (it ismainly fast relaxing tetranuclear clusters that contribute EPR spectrumunder these conditions). Again only gxy components are presented. Thestrong signal of N2 (g¼ 1.9) in wild-type Complex I appearing with a short

45K, 2mWWild-type

0

120

100

80

60

40

20

0

120

100

80

60

40

20

00 5 10

Time, ms

N1b

N1aN1a

N1b

0.04

0.02

0

−0.02

−0.04

−0.06

−0.08

−0.10

−0.121.96

0 5 10Time, ms

[2F

e-2S

] cl

uste

rs r

educ

tion

, %

[2F

e-2S

] cl

uste

rs r

educ

tion

, %

1.94 1.92 1.9 1.88 1.96 1.94 1.92g-value

1.9 1.88g-value

90ms

180ms

720ms

2000ms

4000ms

8000ms2900ms

1400ms

720ms

360ms

180ms

90ms

••

−0.05

−0.10

−0.15

−0.20

R274A

Figure 4.8 (Continued)



Willd type

0 0

−2

−4

−6

−8

−0.005

−0.010

−0.015

−0.020

−0.025

−0.030

120 120

100

80

60

40

20

0

100

80

60

N6bN6b

N2

40

[4F

e-4S

] cl

uste

rs r

educ

tion

, %

[4F

e-4S

] cl

uste

rs r

educ

tion

, %

20

00 5

Time, ms

10

Time, ms

0 5 10

g-value g-value1.95 1.9 1.85 1.95 1.9 1.85

90ms

180ms

720ms

2000ms

4000ms

8000ms

�

90ms

180ms

360ms

720ms

1400ms

2900ms

�

10K, 10mW

R274A� 10−3

Figure 4.8 Theresultsof freeze-quenchexperimentsonfast reductionofComplexIwithNADH.Purifiednonactivatedwild-type andmutatedNuoCDR274Aenzymeswere used.TheEPRspectrafromthetoptothebottompresentthe redoxstateof FeSclusters indistinctsampleswhere the reactionwas stopped at the time interval indicated above the curves.Thespectrumdesignatedby infinite timewasobtained fromoneof theprevious samplesmeltedand immediately frozen again (100%reduction).TheEPRspectra at 45Kshowthe changesin the redox state of only [2Fe-2S] clusters: N1a (trough at g ¼ 1.92) and N1b (trough atg¼1.932).The EPRspectra at10K and10mWmainly show the redox state of fast-relaxing[4Fe-4S]clusters:N2(troughatg¼1.896)andN6b(troughatg¼1.887).Theobtainedspectraallow determination of the individual reduction kinetics, which are shown below thecorrespondingsetofspectra.





characteristic time of approximately 100 msec covers the relatively broadsignal of N6b (g ¼ 1.887), the appearance of which could be noticed byonly a slight broadening of the entire band and an insignificant shift in thetrough, which makes determination of the characteristic time for N6breduction (2 msec) not fully reliable. In contrast, the spectra of R274samples lacking an N2 signal clearly show the N6b band, which appearswith a trustworthy characteristic time of 2 msec. Determination of thesequence of the redox events in FeS clusters of Complex I allows us todraw important conclusions on the molecular mechanism of this enzyme(Verkhovskaya et al., 2008).

REFERENCES

Ballou, D. P. (1978). Freeze-quench and chemical-quench techniques. Methods Enzymol.54, 85–93.

Belevich, G., Euro, L., Wikstrom, M., and Verkhovskaya, M. (2007). Role of the conservedarginine 274 and histidine 224 and 228 residues in the NuoCD subunit of complex I fromEscherichia coli. Biochemistry 46, 526–533.

Berger, R. L. (1978). Some problems concerning mixers and detectors for stopped flowkinetic studies. Biophys. J. 24, 2–20.

Bilsel, O., Kayatekin, C., Wallace, L. A., and Matthews, C. R. (2005). A microchannelsolution mixer for studying microsecond protein folding reactions. Rev. Sci. Instrum. 76,014302 (1–7).

Cherepanov, A. V., and de Vries, S. (2004). Microsecond freeze-hyperquenching:Development of a new ultrafast micro-mixing and sampling technology and applicationto enzyme catalysis. Biochim. Biophys. Acta 1656, 1–31.

Galkin, A. S., Grivennikova, V. G., and Vinogradov, A. D. (1999). !H+/2e� stoichiome-try in NADH-quinone reductase reactions catalyzed by bovine heart submitochondrialparticles. FEBS Lett. 451, 157–161.

Grgic, L., Zwicker, K., Kashani-Poor, N., Kerscher, S., and Brandt, U. (2004). Functionalsignificance of conserved histidines and arginines in the 49-kDa subunit of mitochondrialcomplex I. J. Biol. Chem. 279, 21193–21199.

Johnson, K. A. (1995). Rapid quench kinetic analysis of polymerases, adenosinetripho-sphatases, and enzyme intermediates. Methods Enzymol. 249, 38–61.

Lin, Y., Gerfen, G. J., Rousseau, D. L., and Yeh, S. R. (2003). Ultrafast microfluidic mixerand freeze-quenching device. Anal. Chem. 75, 5381–5386.

Ohnishi, T. (1998). Iron-sulfur clusters/semiquinones in complex I. Biochim. Biophys. Acta1364, 186–206.

Regenfuss, P., Clegg, R. M., Fulwyler, M. J., Barrantes, F. J., and Jovin, T. M. (1985).Mixing liquids in microseconds. Rev. Sci. Instrum. 56, 283–290.

Roder, H., Maki, K., Cheng, H., and Shastry, M. C. (2004). Rapid mixing methods forexploring the kinetics of protein folding. Methods 34, 15–27.

Schuster, R., Kirchner, V. V., Allongue, P., and Ertl, G. (2000). Electrochemical micro-machining. Science 289, 98–101.

Shastry, M. C., Luck, S. D., and Roder, H. (1998). A continuous-flow capillary mixingmethod to monitor reactions on the microsecond time scale. Biophys. J. 74, 2714–2721.



Sinegina, L., Wikstrom, M., Verkhovsky, M. I., and Verkhovskaya, M. L. (2005). Activationof isolated NADH:ubiquinone reductase I (complex I) from Escherichia coli by detergentand phospholipids. Recovery of ubiquinone reductase activity and changes in EPRsignals of iron-sulfur clusters. Biochemistry 44, 8500–8506.

Smith, M. H. (1973). Observations relating to the efficiency of mixing in rapid reaction flowdevices. Biophys. J. 13, 817–821.

Tanaka, M., Matsuura, K., Yoshioka, S., Takahashi, S., Ishimori, K., Hori, H., andMorishima, I. (2003). Activation of hydrogen peroxide in horseradish peroxidase occurswithin approximately 200 micros observed by a new freeze-quench device. Biophys. J.84, 1998–2004.

Verkhovskaya, M. L., Belevich, N., Euro, L., Wikstrom, M., and Verkhovsky, M. I. (2008).Real-time electron transfer in respiratory complex I. Proc. Natl. Acad. Sci. USA 105,3763–3767.

Wikstrom, M. (1984). Two protons are pumped from the mitochondrial matrix per electrontransferred between NADH and ubiquinone. FEBS Lett. 169, 300–304.

Chapter 4 Electron Transfer in Respiratory Complexes Resolved by an Ultra-Fast Freeze-Quench...

Documents

Transcript of Chapter 4 Electron Transfer in Respiratory Complexes Resolved by an Ultra-Fast Freeze-Quench...