Mass spectrometric investigation of the neuropeptide complement and release in the pericardial...

Mass spectrometric investigation of the neuropeptide complement

and release in the pericardial organs of the crab, Cancer borealis

Lingjun Li,*,�,�Wayne P. Kelley,�,1 Cyrus P. Billimoria,� Andrew E. Christie,§,2 Stefan R. Pulver,�Jonathan V. Sweedler� and Eve Marder�

*School of Pharmacy & Department of Chemistry, University of Wisconsin, Madison, Wisconsin, USA

�Department of Chemistry and the Beckman Institute, University of Illinois, Urbana, Illinois, USA

�Department of Biology, Volen Center, Brandeis University, Waltham, Massachusetts, USA

§Department of Neuroscience, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, USA

Abstract

The crustacean stomatogastric ganglion (STG) is modulated

by both locally released neuroactive compounds and circula-

ting hormones. This study presents mass spectrometric

characterization of the complement of peptide hormones

present in one of the major neurosecretory structures, the

pericardial organs (POs), and the detection of neurohormones

released from the POs. Direct peptide profiling of Cancer

borealis PO tissues using matrix-assisted laser desorption/

ionization (MALDI) time-of-flight (TOF) mass spectrometry

(MS) revealed many previously identified peptides, including

proctolin, red pigment concentrating hormone (RPCH), crus-

tacean cardioactive peptide (CCAP), several orcokinins, and

SDRNFLRFamide. This technique also detected corazonin, a

well-known insect hormone, in the POs for the first time.

However, most mass spectral peaks did not correspond to

previously known peptides. To characterize and identify these

novel peptides, we performed MALDI postsource decay (PSD)

and electrospray ionization (ESI) MS/MS de novo sequencing

of peptides fractionated from PO extracts. We characterized a

truncated form of previously identified TNRNFLRFamide,

NRNFLRFamide. In addition, we sequenced five other novel

peptides sharing a common C-terminus of RYamide from the

PO tissue extracts. High K+ depolarization of isolated POs

released many peptides present in this tissue, including sev-

eral of the novel peptides sequenced in the current study.

Keywords: Cancer borealis, MALDI MS, neuropeptides,

neurosecretion, pericardial organs, postsource decay peptide

sequencing.

J. Neurochem. (2003) 87, 642–656.

The crustacean stomatogastric ganglion (STG) is one of the

leading systems for studying the neural basis of motor pattern

generation and the modulatory effects of neuroactive

substances at the cellular and network levels (Marder and

Hooper 1985; Harris-Warrick et al. 1992; Marder et al.

1995; Marder and Calabrese 1996; Skiebe 2001; Nusbaum

and Beenhakker 2002). Previous studies have shown that the

STG is modulated by both neuroactive agents released

locally from input axons and circulating hormones delivered

via the hemolymph (Marder 1987; Christie et al. 1995). A

large number of studies have documented that a diverse

assortment of small molecule transmitters, amines and

Received April 30, 2003; revised manuscript received July 16, 2003;

accepted July 16, 2003.

Address correspondence and reprint requests to Dr Lingjun Li, School

of Pharmacy, University of Wisconsin, 777 Highland Avenue, Madison,

WI 53705–2222, USA. E-mail: [email protected] address: Pharmaceutical Development, GlaxoSmithKline, King

of Prussia, PA 19406, USA.2Current address: Department of Biology, University of Washington,

Seattle, WA 98195–1800, USA.

Abbreviations used: ACN, acetonitrile; AST, allatostatin; CabTRP,

Cancer borealis tachykinin-related peptide; CCAP, crustacean cardio-

active peptide; CHH, crustacean hyperglycemic hormone; CID, colli-

sional induced dissociation; DHB, 2,5-dihydroxybenzoic acid; DG

neuron, dorsal gastric neuron; dgn, dorsal gastric nerve; ESI, electro-

spray ionization; IC neuron, inferior cardiac neuron; lvn, lateral ven-

tricular nerve; LP neuron, lateral pyloric neuron; MALDI, matrix-

assisted laser desorption/ionization; MS, mass spectrometry; mvn, medial

ventricular nerve; PD neuron, pyloric dilator neuron; pdn, pyloric dilator

nerve; PO, pericardial organ; PSD, postsource decay; PY neuron, pyloric

neuron; pyn, pyloric nerve; RPCH, red pigment concentrating hormone;

RP-HPLC, reverse phase high performance liquid chromatography;

STG, stomatogastric ganglion; STNS, stomatogastric nervous system;

TFA, trifluoroacetic acid; TOF, time-of-flight; VD neuron, ventricular

dilator neuron.

Journal of Neurochemistry, 2003, 87, 642–656 doi:10.1046/j.1471-4159.2003.02031.x

642 � 2003 International Society for Neurochemistry, J. Neurochem. (2003) 87, 642–656

neuropeptides are present in the STG neuropil. Several recent

reports have shown that many of the same neuromodulatory

compounds are also present in the neurohemal organs of

crustaceans (Christie et al. 1995; Skiebe 1999; Skiebe et al.

1999; Pulver and Marder 2002). One such neurohemal

structure, the pericardial organ (PO), has long been known to

be an important source of circulating hormones. Moreover,

several studies have shown that many of the hormones

present in the POs can modulate the neural circuits in the

STG, which is situated in the ophthalmic artery anterior to

the heart (Alexandrowicz and Carlisle 1953; Stangier et al.

1986; Stangier et al. 1988; Keller 1992; Christie and

Nusbaum 1995). The current study focuses on mass spectr-

ometric characterization of neuropeptides present in and

released from the POs of the crab Cancer borealis.

Most of our knowledge concerning the peptide content of

the stomatogastric nervous system (STNS) and the POs is

based on immunocytochemistry. This technique, while a

powerful first step in determining modulator presence, leaves

the amino acid sequence identity of the native peptides

unknown. Clearly, unambiguous determination of the actual

peptide structure with accurate molecular weight and amino

acid sequence information is essential for studies of the

physiological functions of these molecules. Matrix-assisted

laser desorption/ionization mass spectrometry (MALDI MS)

is a highly accurate method that can be used to directly

profile peptides from complex biological samples such as

tissue sections or single cells with minimal sample prepar-

ation (Jimenez et al. 1994; Li et al. 1999b; Li et al. 2000a;

Li et al. 2000b; Predel 2001). Its application to peptide

identification has recently been extended to the crustacean

STNS in the identification and characterization of multiple

members of the orcokinin peptide family (Li et al. 2002;

Skiebe et al. 2002). In the current study, we aim to determine

the complement of peptides present in the POs of the crab, C.

borealis. To overcome the challenges of the high chemical

complexity of these neuropeptides, and the almost complete

lack of genomic information, we employed front-end multi-

stage microseparation to simplify cellular matrices, and

performed de novo mass spectrometric sequencing to obtain

the primary structures of several previously unknown

peptides. To explore the putative hormonal role of these

peptides, we also performed the first mass spectrometric

assay of peptide release from the POs using high K+-induced

depolarizations.

Materials and methods

Animals and dissection

Jonah’s crabs, C. borealis were obtained from Commercial Lobster

(Boston, MA, USA) and maintained without food in artificial

seawater tanks at 10–12�C. Animals were cold-anesthetized by

packing in ice for 15–30 min prior to dissection. The POs were

dissected by removing the carapace covering the heart and

subsequently removing the longitudinal body muscles, hyperdermis,

and heart. The animal was then pinned ventral side up in a Sylgard-

lined dissection dish to expose the pericardial cavity. The POs were

identified visually as an iridescent web of nerves branching over the

muscles surrounding the pericardial cavity and dissected free. All

dissection was carried out in chilled (approximately 4�C) physio-logical saline (composition in mM: NaCl, 440; KCl, 11; MgCl2, 26;

CaCl2, 13; Trizma base, 11; maleic acid, 5; pH 7.45).

Cellular sample preparation for MALDI MS analysis

Small pieces of tissues were dissected and prepared for MALDI MS

analysis. Physiological saline was replaced with an aqueous MALDI

matrix solution, 10 mg/mL of 2,5-dihydroxybenzoic acid (DHB)

(ICN Pharmaceuticals, Costa Mesa, CA, USA), to remove the

extracellular salts associated with the tissue sample (Garden et al.

1996). Tungsten needles and fine forceps were used to dissect and

transfer small pieces of tissue onto a MALDI sample plate

containing 0.5 lL of either regular aqueous DHB (10 mg/mL)

matrix solution or concentrated DHB (50 mg/mL) in acetone/water

(4 : 1) mixed solvent. Once on the sample plate, the tissue was

smashed with dissection tools and allowed to dry at ambient

temperature and then subjected to MALDI MS analysis.

Microbore reverse phase (RP)-HPLC of homogenates

The extraction and first stage separation procedure were performed

as reported previously (Li et al. 2002). Peptides were initially

extracted from 26 POs using acidified acetone (1 : 40 : 6, concen-

trated HCl/acetone/H2O) as described previously (Floyd et al.

1999). Briefly, samples were homogenized in a microhomogenizer

(Jencons Scientific Ltd, UK), and the supernatant drawn off and

centrifuged (Baxter Biofuge 15, Mcgraw Park, IL, USA). This

process was repeated several times, water was added and then the

extract was concentrated under a stream of nitrogen to approxi-

mately 300 lL.Separations were performed utilizing a microbore HPLC (Magic

2002, Michrom BioResources, Auburn, CA, USA). For the first stage

separation, an aliquot of the extract was injected onto a reverse phase

1.0 · 150 mm C-18 column (Reliasil) with a 5-lm particle size and

30 nm pore size. The column was equilibrated with solvent A at a

programmed temperature of 35�C. An aliquot of the aqueous extractwas injected onto the column at a constant flow rate of 50 lL/min anda gradient developed from 5 to 98% of solvent B in 34 min. Solvent A

consisted of 2% acetonitrile (ACN), 98% H2O and 0.1% trifluoro-

acetic acid (TFA) (v/v). Solvent B consisted of 95% ACN, 5% H2O

and 0.1% TFA (v/v). Sample peaks were detected via absorbance at

214 and 280 nm wavelengths and the eluent collected by a small

volume fraction collection system (Gilson FC 203B, Middleton, WI,

USA). To identify peptides of interest, each fraction was screened

using MALDI MS; approximately 0.25 lL of each fraction (approxi-mately 20–200 lL total) was deposited on aMALDIMS sample platefollowed by the same volume of an a-cyano-4-hydroxycinnamic acidmatrix (10 mg/mL in 6 : 3 : 1 ACN/H2O/3% TFA) (Aldrich,

Milwaukee, WI, USA). Thus more than 95% of each fraction was

available for further assays.

Second stage microbore HPLC separation

Fractions assayed by MALDI time-of-flight (TOF) MS that

contained insufficient intensity or interfering peaks from coelution

Neuropeptides in C. borealis pericardial organs 643

� 2003 International Society for Neurochemistry, J. Neurochem. (2003) 87, 642–656

were further purified by a second stage HPLC. Second stage

separations were effected by changing the column to a

1.0 · 150 mm Vydac C-18 MS column (218MS5115, Grace Vydac,

Hesperia, CA, USA) consisting of 5 lm particles with a 30-nm pore

size, changing the solvent additives, slowing the relative flow rate

and by flattening the gradient profile. Separations of aliquots from

the combined first stage fractions were performed both with on-line

electrospray ionization (ESI)-MS detection and off-line fraction

collection. Both procedures utilized identical solvents and methods.

Fractions were collected manually on a small volume fraction

collection system.

The selected fractions from the first-stage analysis were com-

bined and concentrated to an aqueous medium using a commercial

Speedvac system (Thermo-Savant, Holbrook, NY, USA). An aliquot

of the sample was then loaded onto a peptide trap inline with the

injection port of the microbore HPLC instrument. For HPLC-ESI-

MS experiment a 5-lL injection of the concentrate was used and theremainder, approximately 50 lL, was saved for the second stageseparation and fraction collection. Each aliquot was injected onto

the column at a uniform flow rate of 50 lL/min. The mobile phaseconsisted of solvent A: 95% H2O (Burdick and Jackson, Muskegon,

MI, USA), 5% ACN, 0.1% acetic acid (v/v) and 0.02% TFA (v/v)

(Sigma-Aldrich, Milwaukee, WI, USA) and B: 10% H2O, 90%

2 : 2 : 1 ACN, 0.1% acetic acid (v/v) and 0.014% TFA (v/v). A

gradient was developed from 5 to 15% solvent B in 5 min and 15–

45% B in an additional 30 min.

MALDI MS

MALDI mass spectra were obtained using a Voyager DE STR

(Applied Biosystems, Framingham, MA, USA) TOF mass spectro-

meter equipped with delayed ion extraction. A pulsed nitrogen laser

(337 nm) was used as desorption/ionization source, and positive-ion

mass spectra were acquired using both linear and reflectron mode.

Each representative mass spectrum shown is the smoothed average

of 128–256 laser pulses. Mass calibration was performed externally

using a mixture of synthetic peptide standards (PE Biosystems,

Framingham, MA, USA). Mass accuracy was typically better than

0.01% (Li et al. 1998).

PSD analysis

Semi-purified microbore LC fractions containing a peptide of

interest, or 2 pmol of synthetic peptide standards (synthesized by

either the Biotechnology Center at the University of Illinois or the

Protein Chemistry Laboratory at the University of Pennsylvania,

School of Medicine), were subjected to PSD analysis. For MALDI

PSD analysis, the matrix a-cyano-4-hydroxycinnamic acid (10 mg/mL in 6 : 3 : 1 ACN/water/3% TFA) was used. In these experi-

ments, the total acceleration voltage was 20 kV, grid voltage set at

75%, guide wire voltage set at 0.03%, and a delay time of 75 ns

used. By the use of timed ion selector, different precursor ions were

selected from a mixture of peptides and subjected to fragmentation.

Under these experimental conditions, the mass accuracy of the

precursor ion was within 30 p.p.m., and the average error on the

mass assignment of the PSD ions was less than 0.3 Da. Spectra were

obtained by accumulating data from 100 to 256 laser shots. To

obtain complete PSD spectra, a series of reflectron TOF spectral

segments were acquired, each optimized to focus fragment ions

within different m/z ranges (Kaufmann et al. 1993). Each segment

was stitched together using the Biospectrometry Workstation

software to generate a composite PSD spectrum.

On-line ESI MS/MS analysis

For on-line MS detection, the column was connected directly to the

atmospheric pressure inlet port of an LCQ Deca, ESI-ion trap mass

spectrometer (Thermo-Finnigan, San Jose, CA, USA). The transfer

volume to the MS detector was calculated to be similar to that of the

absorbance detector resulting in retention times that are similar to

the fraction collection assays. The MS tune method employed a

spray voltage of 4.3 kV, a capillary temperature of 220�C, capillaryvoltage of 21 Vand a tube lens offset of 10 V. By default, automatic

gain was used to control injection of ions into the trap. The Xcalibur

instrument setup software (Thermo-Finnigan, San Jose, CA, USA)

employed a modified triple play data dependent acquisition control.

Briefly, a full scan MS was followed by a zoomscan and then MS2

of the largest peak in the full scan. Dynamic exclusion was utilized

in order to include high resolution and MS2 information on

secondary peaks within the scan.

Nanospray ESI MS/MS

To perform collisional-induced dissociation (CID) experiments for

sequence information while utilizing minimal sample, the ion trap

instrument was fitted with a static nanospray source (Thermo-

Finnigan, San Jose, CA, USA). An aliquot of approximately 5 lLwas taken from a fraction from the second stage separation

corresponding to the mass of interest, concentrated via Speedvac

and reconstituted in 10 lL of 50/50 (v/v) CH3OH and H2O with

0.1% formic acid. The sample was loaded into a tapered, platinum

coated, borosilicate nanospray emitter, PicoTip� (New Objective,

Woburn, MA, USA) and a spray voltage of 1.5 kV was applied. The

instrument capillary temperature was set at 200�C with an applied

potential of 39 V. The optimized spray voltage and ion optic settings

were adjusted using the instrument’s auto tune function. Once a full

scan was observed in real time, MS2 and MS3 experiments were

directed by manual input of the m/z-values.

Tandem MS sequencing experiments used a mass isolation width

of 1.5 Da for the precursor ion and 1.0 Da for resulting MS2

fragment ions to be retained in the trap. An activation Q (ion

instability parameter) value of 0.25 at an activation time of 30 ms

was utilized. The normalized collisional (RF) energy was set at 35%

for both MS2 and MS3 experiments. Spectra were acquired for

approximately 2 min resulting in 164 scans. Sequence verification

was facilitated using the online Protein Prospector, MS-Product

program from the University of California, San Francisco.

MALDI assay for peptides released from the POs

To depolarize the POs, we employed a 10-fold higher concentration

of K+ (110 mM) than that in normal saline to the isolated pair of POs

from each animal (N ¼ 3). A cocktail of peptidase inhibitors

including amastatin (20 lM), leupeptin (10 lM), antipain-dihydro-chloride (71 lM), bestatin (130 lM), phosphoramidon (37 lM), andaprotinin (0.3 lM) (Roche Molecular Biochemicals, Germany) wasused throughout the experiment. All of the preparations were kept

on ice (�4�C) for the duration of the experiment. To determine if thedetected peptides were released in a depolarization and Ca2+-

dependent manner, we repeated the experiment without high

K+ depolarization in normal saline as well as in high K+/low Ca2+

644 L. Li et al.


(1.3 mM Ca2+ and 13 mM Mn2+) saline. Each preparation was

incubated for 10 min in the experimental condition with two washes

in normal saline between each condition to remove peptides

resulting from previous experimental protocol.

We sampled 10 lL of releasate from each experimental condi-

tion, desalted and concentrated the releasate with ZipTip pipette tips

(Millipore, Bellerica, MA, USA) packed with C18 reverse phase

media prior to spotting onto the MALDI sample plate. Briefly, the

ZipTip was wet with 50% ACN in Milli-Q water, equilibrated with

0.1% TFA in Milli-Q water, loaded with sample containing 0.1%

TFA by aspirating and dispensing the sample 10 times, and then

washed with 5% methanol in 0.1% TFA/water, followed by elution

directly onto a MALDI MS sample plate with 2 lL DHB (50 mg/

mL) in 50% ACN with 0.1% TFA.

Electrophysiology

The stomach was removed as previously described. The STNS was

isolated from the stomach and pinned onto a Sylgard (Dow Corning,

Midland, MI, USA) coated dish. The STG and the stomatogastric

nerve (stn) were desheathed. The STG was isolated from anterior

ganglia inputs by placing isotonic sucrose (750 mM) containing

10)6 M tetrodotoxin in a Vaseline well built around the stn.

Extracellular recordings were made from pyloric and gastric nerves

using stainless steel pin electrodes within saline-filled Vaseline

wells. Signals were amplified by an AM-Systems 1700 differential

amplifier (Carlsborg, WA, USA) and recorded using an Axon

Instruments (Foster City, CA, USA) data interface board. Spike time

data were extracted using scripts written by Dirk Bucher in Spike2

(Cambridge Electronic Design, Cambridge, UK). Two-minute

epochs of data were used in the analysis of peptide actions. The

two minutes preceding the application of peptide and the eighth to

tenth minute of application were analyzed.

Results

Direct MALDI MS profiling of peptides

in the pericardial organs reveals a multitude

of peptides present

As shown in Fig. 1, direct MALDI MS profiling of a small

piece of freshly isolated PO from C. borealis revealed the

presence of a number of previously identified peptides

including proctolin, crustacean cardioactive peptide (CCAP),

SDRNFLRFa, and Ala13-orcokinin, as well as many other

unknown peptides. Figure 1(b) shows the mass spectrum

acquired in the high mass region. A number of peaks were

detected in the mass range spanning from 2000 to 9650 Da. A

multitude ofCHHs have been previously characterized in sinus

gland (SG) neurosecretory system in the eyestalk and POs of

various species of crustacea (Keller 1992; Dircksen et al.

2001), with molecular weights ranging from 8420 to 8634 Da.

The peak atm/z 8561.70, however, does not correspond to any

of the previously identified CHHs. The detection of this peak

and a few other peaks around 8000 and 9000 Da in the mass

spectrum suggested the possible presence of CHH peptides in

the crab, C. borealis. Another interesting peak that might be

related to the CHH precursor was the mass spectral peak atm/z

4071.68, whose molecular weight was in close agreement to

that of CHH-precursor related peptide (CPRP, with 0.002%

mass measurement error) of the shore crab, Carcinus maenas

(Dircksen et al. 2001). However, without additional informa-

tion from sequencing analysis, conclusive assignment of these

mass spectral peaks was not possible.

Because peptides are distributed differentially throughout

the POs, direct tissue profiling only provides a snapshot of the

peptides present in a specific region of this neurohemal organ.

Due to the heterogeneous distribution of modulators in this

tissue, a given peptide was not always detected in the MALDI

mass spectra of individual PO tissue samples. In Table 1, we

consider a peptide to be present (+) in the POs only if we

observed the peptide signals in over 80% of the total spectra

generated (N ¼ 102). This cut-off value is determined based

on the occurrence of known peptides present in the POs and

the reproducibility of our MS detection.

Off-line coupling microbore LC fractionation

of PO tissue extract with MALDI MS detection

for peptide sequencing

To obtain a comprehensive description of peptides present in

the PO tissue and simplify the complex mixture resulting

from the tissue extract, we performed multiple stages of

microbore LC separation prior to MALDI MS analysis.

(a)

(b)

Fig. 1 Representative MALDI mass spectrum of a small piece of

freshly isolated pericardial organ from C. borealis. (a). The mass

region of 625–1525 is displayed (b). The mass range from 2000 to

9650 is displayed. Signals correspond to protonated molecular ions,

[M + H]+, where M is the molecular weight of each peptide. Several

previously identified peptides are labeled with asterisks, and repre-

sented in bold face.



Tab

le1

Peptides

identified

inth

epericard

ialorg

ans

of

Cancer

bore

alis

by

MA

LD

IM

S

[M+

H]+

Peptide

nam

eP

eptide

sequence

Imm

unore

activity

Pre

vio

usly

isola

ted

Direct

tissue

MA

LD

I*

LC

-MA

LD

IP

SD

and/o

rE

SI

MS

/MS

Rele

ase*

649.3

6P

rocto

linR

YLP

T+

(Christie

et

al.

1995)

+(M

ard

er

et

al.

1986)

++

+

766.3

8C

abT

RP

IbS

GF

LG

MR

a+

(Christie

et

al.

1995)

+(C

hristie

et

al.

1997)

––

+

784.4

1F

VN

SR

Ya

FV

NS

RY

a+

++

832.4

1F

YA

NR

Ya

FY

AN

RY

a+

+–

862.4

2F

YS

QR

Ya

FY

SQ

RY

a–

+–

899.4

4A

ST

-3G

GS

LY

SF

GLa

+(C

hristie

et

al.

1995)

––

–+

931.4

3R

PC

HpQ

LN

FS

PG

Wa

+(C

hristie

et

al.

1995)

–+

+–

956.3

9C

CA

PP

FC

NA

FT

GC

+(C

hristie

et

al.

1995)

–+

++

965.5

4F

LR

Fam

ide-r

ela

ted

pep

tid

eN

RN

FL

RF

a+

(Christie

et

al.

1995)

–+

++

976.4

6S

GF

YA

NR

Ya

SG

FY

AN

RY

a+

++

1030.5

1P

AF

YS

QR

Ya

PA

FY

SQ

RY

a+

++

1053.5

6F

LR

Fam

ide-r

ela

ted

peptide

SD

RN

FLR

Fa

+(C

hristie

et

al.

1995)

+(W

eim

ann

et

al.

1993)

++

+

1066.5

9F

LR

Fam

ide-r

ela

ted

peptide

TN

RN

FLR

Fa

+(C

hristie

et

al.

1995)

+(W

eim

ann

et

al.

1993)

––

+

1369.6

3C

ora

zo

nin

pQ

TF

QY

SR

GW

TN

a+

(Christie

and

Nusbaum

1995)

++

+

1474.6

5A

la13-O

rcokin

inN

FD

EID

RS

GF

GF

A+

(Liet

al.

2002)

–+

++

1502.6

8V

al1

3-O

rcokin

inN

FD

EID

RS

GF

GF

V+

(Liet

al.

2002)

––

+–

1547.6

8S

er9

-Orc

okin

inN

FD

EID

RS

SF

GF

N+

(Liet

al.

2002)

––

+–

*Peptides

dete

cte

din

direct

tissue

MA

LD

I(N

¼102)

and

rele

ase

experim

ents

(N¼

12)

inover

80%

of

the

spectr

aare

mark

ed

by

aplu

ssig

n;

inth

ese

two

colu

mns,

peptides

are

identified

by

mole

cula

rm

ass

measure

ment

only

,w

ith

am

ass

measure

ment

accura

cy

of50

p.p

.m.fo

rre

flectr

on

mode

and

200

p.p

.m.fo

rlin

ear

mode.P

eptide

identification

confirm

ed

by

MS

/MS

sequencin

gis

mark

ed

with

aplu

ssig

n.

CC

AP

,cru

sta

cean

card

ioactive

peptide;

RP

CH

,re

dpig

ment

concentr

ating

horm

one.

Novelpeptides

identified

inth

ecurr

ent

stu

dy

are

inbold

face.

646 L. Li et al.


A chromatogram of the first stage LC separation is depicted

in Fig. 2 (center). Several representative MALDI survey

spectra of corresponding LC fractions are shown surrounding

the LC chromatogram. In many ways MALDI-TOF MS is an

orthogonal separation technique in that it provides a further

dynamic range and resolution to the separation of the crude

material. As shown in the displayed MALDI spectra of

various LC fractions in Fig. 2, each LC fraction still

contained numerous compounds, illustrating the high com-

plexity of the crude material. With the precursor ion selection

capability, MALDI postsource decay (PSD) analysis can be

performed on semipurified LC fractions to obtain peptide

sequence information. As an example, Fig. 3 shows several

MALDI PSD fragmentation spectra obtained from a single

LC fraction, with those selected precursor ions labeled with

dots. Detailed analyses of PSD fragmentation spectra of

precursor ions at m/z 965.543, 976.464, 1030.51, and

1369.60 are described below; analysis of the fragmentation

pattern of the precursor ion at m/z 1474.6 confirms the

identity of an orcokinin peptide reported previously (Li et al.

2002).

MALDI-PSD analysis of putative corazonin peptide

As shown in Fig. 1, among the numerous peaks detected in

the crab PO tissue, the mass spectral peak at m/z 1369.62

matches the calculated protonated molecular weight of

corazonin (Mr ¼ 1369.63), an insect hormone that is

conserved in all but one species examined (Veenstra 1989,

1991). To determine if the mass spectral peak at m/z 1369.62

has an identical amino acid sequence to that of authentic

corazonin, we performed PSD analysis.

Figure 4 shows the peptide sequence of corazonin (Fig.

4a) and PSD spectra obtained from the putative corazonin

containing LC fraction (Fig. 4b, upper trace) and the

synthetic corazonin standard (Fig. 4b, lower trace). The

identity of authentic corazonin in the PO extract is strongly

supported by the virtually identical fragmentation pattern

obtained between the putative and the synthetic corazonin

PSD spectra. The complete fragmentation observed from the

precursor ion at m/z 1369.63 in the PO extract also allowed

de novo sequencing of the peptide. Thus, we deduced the

amino acid sequence of the putative corazonin peak using the

mass difference between consecutive b- or y-type ions.

The sequence is identical to authentic corazonin-

qTFQYRGWTN-amide.

To confirm the sequence of corazonin, we entered the

derived sequence into the MS-Product software program

developed by the UCSF Mass Spectrometry Facility (http://

prospector.ucsf.edu). The software calculates the possible

fragment ions resulting from PSD processes. The fragment

ions detected in the PSD spectra match the predicted

fragment ions and several additional internal fragment ions

are labeled (Fig. 4b). The measured molecular weight of the

precursor ion and the mass of the C-terminal fragment ions

Fig. 2 An illustration of MALDI MS tracking of microbore HPLC

separation of crude tissue extract of pericardial organs from C. bore-

alis. A chromatogram from the first stage HPLC separation of the

extract is shown in the center. Detection was achieved by absorbance

at a wavelength of 214 nm. Several MALDI MS survey spectra of

corresponding LC fractions are displayed surrounding the LC chro-

matogram.

Fig. 3 MALDI PSD sequencing of multiple novel peptides from a

semipurified microbore LC fraction of C. borealis pericardial organ

extract. The spectrum in the center is a MALDI mass spectrum of a

single LC fraction, with gray dots on the selected peaks indicating the

precursor ions being isolated for MALDI PSD fragmentation. The

surrounding traces are PSD composite fragmentation spectra of var-

ious precursor ions selected.



confirm the amidated C-terminus and a pGlu-modified

N-terminus in the peptide. This result confirms that authentic

corazonin is present in the POs of C. borealis. Authentic

corazonin has recently been shown to be a powerful

modulator of the neural circuits present in the STG of

C. borealis (A.E. Christie, unpublished observation).

A new member of the FMRFamide peptide family

Figure 5(a) shows the MALDI PSD fragmentation spectrum

of a precursor ion at m/z 965.54. Using the same sequencing

strategy outlined above, we determined the amino acid

sequence to be NRNFLRFamide. This represents a truncated

form of the TNRNFLRFamide peptide that was previously

isolated and sequenced from the STNS of the crab C. borealis

(Weimann et al. 1993). The derived amino acid sequence is

shown at the top of the spectrum. The sequence interpretation

was performed from both the N- (b-ions) and C-termini

(y-ions). The PSD analysis of the synthetic peptide standard

NRNFLRFamide produced an identical fragmentation pat-

tern (data not shown), further substantiating the proposed

sequence.

The similarity between this peptide and those previously

shown to be physiologically active on STG motor patterns

(Weimann et al. 1993) caused us to ask whether NRNFLRF-

amide would mimic the actions of the previously identified

extended FLRFamides. Therefore, we carried out a series of

electrophysiological experiments. Figure 5(b) shows that

bath application of 10)6 M synthetic NRNFLRFamide can

activate both pyloric and gastric rhythms in preparations in

which neither rhythm is being expressed (N ¼ 4). Shown on

the left is a schematic drawing illustrating the extracellular

recording sites and experimental conditions used to collect

the data shown on the right. The STG was isolated from

anterior ganglia inputs by placing sucrose and 10)6 M

tetrodotoxin in a Vaseline well on the stn, as indicated in

Fig. 5(b). Extracellular recording electrodes were placed on

the nerves containing the axons of motor neurons of the

STNS. The lateral ventricular nerve (lvn) carries the axons of

the lateral pyloric (LP) neuron, the pyloric (PY) neurons, and

pyloric dilator (PD) neurons, all active in the pyloric rhythm.

The pyloric nerve (pyn) and the pyloric dilator nerve (pdn)

contain axons of the PY and PD neurons, respectively. The

median ventricular nerve (mvn) carries the axons of the

ventricular dilator (VD) and inferior cardiac (IC) neurons.

The dorsal gastric nerve (dgn) carries the axon of the dorsal

gastric (DG) motor neuron, an important component of the

gastric mill network. Under control conditions (Fig. 5b, right

upper panel), the mvn and pdn were silent, and the other

nerve recordings showed some sporadic activity. Application

of 10)6 M NRNFLRFamide (Fig. 5b, right lower panel)

activated the DG neuron in four of four experiments. In all

four experiments the DG neuron was silent in control saline.

In three of the four experiments, NRNFLRFamide induced

DG bursting. In these three experiments NRNFLRFamide

increased the mean number of spikes per burst from 0 to

105 ± 16.7 (standard error of the mean, p < 0.05), increased

the mean burst duration from 0 to 7.91 ± 0.90 s (p < 0.02)

(a)

(b)

Fig. 4 MALDI PSD analysis of the putative corazonin peptide. (a) The

sequence of corazonin, with the N-terminal on the left, and the

C-terminal on the right. Single letter amino acid abbreviations are

used, with q (pGlu), T (Thr), F (Phe), Q (Gln), Y (Tyr), S (Ser), R (Arg),

G (Gly), W (Trp), and N (Asn). Observed b-type (bottom) and y-type

(top) ion pairs are indicated by arrows. (b) MALDI PSD fragmentation

ion spectra of both the ion at m/z 1369.63 from the LC fraction of the

PO extract from C. borealis (upper trace) and the synthetic corazonin

standard (lower trace). As shown in the figure, the immonium ions

observed in the low mass region of the spectrum (as highlighted in the

inset) indicates the presence of amino acids including Ser (60), Arg

(129, 70, 87, 112), Thr (74), pGlu (84), Asn (87), Gln (101), Phe (120),

Tyr (136), and Trp (159). Mass signal pairs (mb + my ¼ mprecursor + 1)

at m/z 112.1/1259.4, 213.2/1158.3, 360.4/1011.1, 488.5/883.0, 651.7/

719.8, 738.8/632.7, 895.0/476.5, 952.0/419.5, 1138.2/233.2, and

1239.3/132.1 are assigned as paired b-type and y-type ions. The

b-type ions are determined based on the detection of their corres-

ponding a-type ions (loss of CO, yielding a 28-Da lower mass ion). The

fragment ion labels used are based on the established nomenclature

(Roepstorff and Fohlman 1984). N-Terminal ion series such as a/b-

ions and their loss of neutrals, C-terminal ion series such as y-ions,

and several internal fragment ions as well as immonium ions are

labeled only in the upper trace spectrum. The masses of the fragment

ion signals labeled with asterisks in the lower trace are identical to

fragment ions detected in the upper trace.

648 L. Li et al.


with a mean cycle period of 16.0 s. In the fourth experiment

NRNFLRFamide induced tonic firing of the DG neuron.

NRNFLRFamide applications significantly increased the

mean duty cycle (fraction of the time the neuron is active

over one entire period of activity) of the LP neuron from

0.12 ± 0.05–0.23 ± 0.07 (p < 0.05). These effects are simi-

lar to those previously reported for SDRNFLRFamide and

TNRNFLRFamide (Weimann et al. 1993).

De novo sequencing of RYamide peptides

Because of the lack of genomic sequence information, partial

sequence-specific fragmentation is not sufficient for peptide

identification. Thus, only high quality PSD fragmentation

spectra resulted in confident peptide identification. Several

PSD analyses generated complete sequence specific frag-

mentation that allowed de novo sequencing of previously

unknown peptides. Figure 6 shows the PSD fragmentation

spectrum of a precursor ion at m/z 976.46. The determined

amino acid sequence was SGFYANRY-amide. We performed

the PSD analysis of the synthetic peptide of the above

sequence which showed identical fragmentation, thus con-

firming the identity of the proposed amino acid sequence

(data not shown).

Another unknown peak at m/z 1030.51 in the LC fraction

was also subjected to PSD fragmentation analysis as shown

in Fig. 7(a). Using the same procedure outlined above, we

obtained a partial sequence of P/A-FYSQRY-amide, with

(a)

(b)

Fig. 5 Characterization of NRNFLRFamide. (a) MALDI PSD frag-

mentation spectrum of a precursor ion at m/z 965.54. The derived

amino acid sequence is shown at the top of the spectrum. The frag-

ment ions are labeled using the established nomenclature. The amino

acid is deduced from both directions using the complete N- and C-

terminal ion series. Detected b/y ion pairs include 115.1/851.5, 271.3/

695.4, 385.2/581.4, 532.3/434.3, 645.3/321.2, 801.4/165.1. Immonium

ions indicating the presence of Asn (87, 70), Arg (70, 87, 112, and

129), Phe (120), and Leu (86) are detected and labeled as single-letter

amino acid abbreviations in Fig. 5(a). Single letter amino acid abbre-

viations are used, with N (Asn), R (Arg), F (Phe), and L (Leu). (b)

Actions of NRNFLRFamide on the motor patterns of the stomato-

gastric ganglion (STG) of the crab, C. borealis. Shown on the left is a

schematic diagram of the experimental configuration. The STG was

isolated by placing sucrose in a Vaseline well around the stomato-

gastric nerve (stn). Shown on the right are simultaneous extracellular

recordings from the pyloric nerve (pyn), the medial ventricular nerve

(mvn), the pyloric dilator nerve (pdn), the lateral ventricular nerve (lvn),

and the dorsal gastric nerve (dgn). The upper panel shows the

recordings in control saline, and the lower panel shows the initiation

and activation of pyloric and gastric mill rhythms upon the application

of 10)6M NRNFLRFamide.

Fig. 6 MALDI PSD de novo sequencing of an octapeptide (m/z

976.46) fractionated from C. borealis PO extract. The derived amino

acid sequence is shown at the top of the spectrum. In the low mass

region, immonium ions indicative of the presence of Gly (30), Ala (44),

Ser (60), Asn (70, 87), Arg (70, 87 112), Phe (120), and Tyr (136) were

detected in the spectrum. These ions are labeled using single letter

code, with S (Ser), G (Gly), F (Phe), Y (Tyr), A (Ala), N (Asn), R (Arg).

Starting from the high mass end, using the formula [M + H]+ )18

(H2O) –X (where X ¼ each of the 20 amino acid residue masses), the

highest b-type ion is determined to be 796.4, with y1 ion at 181.1, thus

indicating that the C-terminal residue is Tyr, with an amidation modi-

fication. Because the b-type ions were generally accompanied by their

corresponding a-type ions in MALDI PSD fragmentation process, the

b/y ion pairs were identified as following: 796.4/181.1, 640.3/337.2,

526.2/451.2, 455.2/522.3, 292.1/685.3, 145.1/832.4, 88.1/889.4. The

amino acid sequence was then deduced from the mass difference

between consecutive b- or y-ions. Several internal fragment ions are

also labeled in the spectrum.



ambiguity of the order of the first two amino acids at the

N-terminus. To obtain complementary sequence specific

fragmentation information to allow the unambiguous deter-

mination of the N-terminal sequence, and to confirm the

proposed sequence, we performed ESI MS/MS analysis. We

pooled several fractions containing the peak at m/z 1030 from

the first stage separation and performed a second stage HPLC

separation (see Materials and methods). A repeat of the

separation on a much smaller aliquot was performed with

ESI MS detection inline. This allowed the assignment of m/z-

values for each of the fractions and rapid identification of the

fraction from the larger scale separation that contained the

peak at m/z 1030.5. Figure 7(b) illustrates the ion chroma-

togram resulting from the HPLC-ESI MS separation (inset)

and the resulting full scan MS of the peak eluting at

11.3 min. High resolution, zoomscan spectra for the two

major components confirm that the peak at m/z 1030.5 was

singly charged and the peak at m/z 515.82 was the doubly

charged species.

We used nanospray ionization (NSI) mass spectrometry to

elucidate the amino acid sequence of the peak at m/z 1030.5,

detected in the HPLC fraction. A MS2 collisional spectrum of

the singly charged ion at 1030.5 is presented in Fig. 7(c). The

tandem MS spectrum corroborated the proposed peptide

sequence for the peak at m/z 1030, PAFYSQRY-amide.

Compared to the MALDI PSD data, the detection of y7 ion at

m/z 933.5 allowed the unambiguous assignment of the

N-terminal sequence as PA. MS3 experiments on the b7fragment were also performed (not shown) and gave similar

information on the b6-b3 and y2-y5 fragments. The NSI-MS2

together with the MALDI PSD spectra and identified

immonium ions provide substantive evidence for the pro-

posed sequence.

Interestingly, we also identified three other novel peptides

containing RYamide at their C-terminus using the same MS-

based sequencing and spectral interpretation procedure as

described above. The derived amino acid sequences are

FVNSRYamide, FYANRYamide, and FYSQRYamide. The

(a)

(b)

(c)

Fig. 7 Mass spectrometric characterization of a novel peptide at m/z

1030.51. (a) MALDI PSD fragmentation ion spectrum of precursor ion

at m/z 1030.51. The derived amino acid sequence is shown at the top

of the spectrum. The detected b/y ion pairs are shown on the peptide

sequence, with b-ions at the bottom and y-ions on the top. The frag-

ment ions in the spectrum are labeled using the established nomen-

clature. The immonium ions are labeled with single-letter amino acid

code, with P (Pro), A (Ala), F (Phe), Y (Tyr), S (Ser), Q (Gln), and R

(Arg). Note the ambiguity at the N-terminal sequence. (b) MS spec-

trum from the HPLC-ESI MS assay corresponding to the peak from the

ion chromatogram (inset) eluting at 11.3 min. The singly and doubly

charged species of the m/z 1030 peak are labeled. (c) MS/MS spec-

trum of the singly charged m/z 1030.5 peak. Loss of H2O from the

precursor was observable at m/z 1012.4. The singly charged b7 ion

corresponding to the loss of RY-NH2 was readily visible at m/z 850.3.

Mass spectral peaks at m/z 833.2, 832.3 and 822.3 were assigned as

b7 ion with loss of NH3, H2O (not labeled) and CO (a7 ion). The low

mass cut-off of the ion trap for this parent ion precluded inspection of

the b2 and a2 ions; however, the corresponding y6 ion can be seen at

m/z 862.3. The y6-NH3 ion was also visible. The remaining b and

related ions were readily identifiable; a3/b3 (m/z 287.9 and 316.0),

a4/b4 (m/z 450.9, 479.0), a5 (not labeled), b5-H2O and b5 (m/z 538.1,

548.1, 566.1), and b6-NH3 (not labeled), b6-H2O and b6 (m/z 676.1,

677.2, 694.3). Ions of singly charged y-type fragments likewise can be

identified; y5 and y5-NH3 (m/z 715.3, 698.4), y4 (m/z 552.3), y3 (m/z

465.3), y2 and y2-NH3 (m/z 337.2, 320.2). The tandem MS spectrum

contains the fragmentation ion information (most notably, the y7 ion at

933.5) that supports the proposed sequence of PAFYSQRY-amide.

650 L. Li et al.


sequences of these new peptides are summarized in bold face

in Table 1. We have tested the physiological actions of some

of the RYamide peptides on the STG network; so far no

apparent actions have been found. Study of their actions on

other target tissues in C. borealis is ongoing.

Peptide release from the POs

To investigate the putative hormonal roles of the peptides

present in the POs, we performed mass spectrometric assay

for the peptides released upon high K+ depolarization

(N ¼ 3). Figure 8 shows representative data from one

experiment. Figure 8(a) shows the MALDI mass spectrum

of the releasate collected in normal saline prior to high K+

depolarization. Here, a few peaks were detected, presumably

due to constitutive release. Once the preparation was

transferred to the high K+ saline solution, a dramatic increase

in the number of peptide peaks was seen (Fig. 8b). Several of

these peaks correspond to known peptides including proct-

olin, Cancer borealis tachykinin-related peptide Ib (Cab-

TRP1b), CCAP, SDRNFLRFamide, and Ala13-orcokinin. In

addition, we also detected many of the peptides identified for

the first time in this report, including corazonin, NRNFLRF-

amide, FVNSRYamide, SGFYANRYamide, and PA-

FYSQRYamide, indicating that these new peptides are

released by high K+ depolarizations. Figure 8(c) shows the

result from high K+ depolarization in low Ca2+, high Mn2+

saline (reversible calcium channel blocker); minimal peptide

release was observed, suggesting a Ca2+-dependence of

release. Finally, after several washes with normal Ca2+ saline,

we applied a second high K+ depolarization in normal saline.

Many of the same peptides were released, with slightly

reduced signal intensity (Fig. 8d). The last column in Table 1

summarizes the peptides detected in total of 12 release

experiments. If a peptide peak was detected in over 80% of

the spectra, we consider a peptide to be released (+) from the

POs in response to high K+ depolarization. Due to the low

concentrations of peptides being released, peptide analysis

was performed in linear mode only, with average mass

measurement accuracy at 200 p.p.m.

Discussion

The POs are major neurosecretory structures that can release

amines and peptides into the hemolymph and elicit a variety

of physiological actions. Berlind (1976) demonstrated that

extracts of the POs of crabs injected into intact animals

caused an increase in the frequency of scaphognathite

beating. Neuromodulators found in decapod POs also

modulate the amplitude and frequency of heart beat

(Alexandrowicz and Carlisle 1953). Additionally, it has been

shown that neuroactive substances in the POs altered

properties of cardiac ganglion (Cooke and Hartline 1975),

neuromuscular junctions, and muscle contractibility (Kravitz

et al. 1980; Lingle 1981; Mercier et al. 1990; Meyrand and

Marder 1991; Worden et al. 1995; Jorge-Rivera and Marder

1996; Jorge-Rivera et al. 1998). Furthermore, neurohor-

mones found in the POs, are capable of exerting a wide range

of modulatory effects on the neural circuits in the STG

(Nusbaum and Beenhakker 2002). For all these reasons,

considerable efforts have been made to determine the

composition of the neuroactive substances present in the

POs in various decapod species. Using immunocytochemical

and biochemical techniques, a wide array of neuromodulatory

(a)

(b)

(c)

(d)

Fig. 8 MALDI MS profiling of releasate from C. borealis PO. (a)

MALDI mass spectrum of control sample in normal saline before high

K+ depolarization. Baseline activity is observed with a few peaks

labeled. (b) MALDI mass spectrum of releasate collected from high K+

depolarization. Numerous peptide peaks are detected, with those

labeled with dots being previously known peptides or newly

sequenced peptides in the current study. The inset lists all the pep-

tides identified based on molecular weight measurement. (c) MALDI

profile collected from preparation in low Ca2+, high Mn2+ (reversible

calcium channel blocker) saline with high K+ depolarization. Minimal

peptide release was observed. (d) MALDI mass spectrum of releasate

collected from a second high K+ depolarization after switching the

preparation back in normal saline. Most of the same peptides seen in

(b) were detected. The identified peaks are listed in the inset.



substances have been identified in the POs in both adult

animals and during embryonic and larval development

(Keller 1992; Christie et al. 1995; Pulver and Marder

2002). In addition to various amines such as serotonin

(Beltz and Kravitz 1983), dopamine (Siwicki et al. 1987),

and octopamine (Evans et al. 1976), many peptides such as

proctolin, FLRFamide-like, CCAP, and orcokinin peptides

have been found in the POs of numerous decapod crustacean

species (Keller 1992; Christie et al. 1995; Pulver and Marder

2002). For example, proctolin is present in the POs of the

lobster, Homarus americanus (Schwarz et al. 1984), the

crayfish, Procamburus clarkii (Siwicki and Bishop 1986),

and the shore crab Carcinus maenas (Stangier et al. 1986).

FMRFamide-like peptides are localized in the POs of the

lobster H. americanus (Kobierski et al. 1987; Trimmer et al.

1987), the crayfish P. clarkii (Mercier et al. 1993), and

several crab species (Krajniak 1991; Christie et al. 1995).

Moreover, allatostatin-like peptides (Skiebe 1999), orcokinin

family peptides (Li et al. 2002; Skiebe et al. 2002; Skiebe

2003), crustacean hyperglycemic hormone (Keller 1992;

Dircksen et al. 2001), and CCAP (Stangier et al. 1988;

Christie et al. 1995; Skiebe et al. 1999) have also been

previously reported to be present in the POs of several

decapod species.

MS-based peptide identification strategy

The current study represents the first mass spectrometric

investigation of the neuromodulatory complement of the POs

in the crab C. borealis. As summarized in Table 1, many

previously identified peptides were detected in direct tissue

MALDI analysis. Only four of these peptides were previ-

ously isolated and sequenced in C. borealis using conven-

tional biochemical techniques that involved multiple steps of

purification of a large pool of tissue samples followed by

Edman degradation (Marder et al. 1986; Weimann et al.

1993; Christie et al. 1997). Other peptides were previously

identified based only on immunoreactivities (Christie et al.

1995; Skiebe 2001; Li et al. 2002). While useful for tissue

localization, immunocytochemistry suffers from limitations

including cross-reactivity with structurally similar peptides

thus preventing the unequivocal identification of a specific

peptide, and the small number of peptides that can be

analyzed simultaneously. As the next step after determining

the identity of a neuromodulator in a particular neuronal

circuit is often determining its physiological role using

exogenous application of synthetic peptides, knowing the

exact chemical structure of the putative hormone is important

information not provided by immunocytochemistry.

MALDI-based peptide identification allows simultaneous

detection of a full spectrum of peptides present at significant

concentrations directly from tissue samples with high mass

accuracy. As shown in the current study, direct tissue

MALDI and PSD sequencing analysis of the LC fractions

resulting from the PO tissue extract confirmed the structures

of several known peptides, such as proctolin, RPCH, CCAP,

extended FLRFamide-related peptides, and orcokinins. In the

case of peptide families, MALDI MS analysis allows

unambiguous identification of the actual forms and different

members of a peptide family. For example, several forms of

the extended FLRFamide-related peptides (including a novel

truncated form of the FLRFamide peptide) were identified in

the POs to substantiate the positive FLRFamide immunore-

activity documented previously. Similarly, Ala13-orcokinin

was detected in the PO tissue samples by MALDI MS, and

two additional forms of orcokinins (Val13- and Ser9-orcoki-

nins) were detected in LC fractions from pooled PO tissue

extract, indicating a possible differential expression of

different forms of orcokinins in the PO, as the latter two

forms of orcokinins were not observed in MALDI spectra of

the tissue samples surveyed. These two forms of orcokinins

are likely expressed in the different regions of the PO tissue

from that of Ala13-orcokinin, or at much lower level to be

detected in individual organs by MS. As we have previously

demonstrated simultaneous detection of five different forms

of orcokinins, including Ala13-, Val13-, and Asn13-orcokinins

in the PO tissue from H. americanus (Li et al. 2002), it is

unlikely that the absence of Val13- and Ser9-orcokinins is due

to the difference of ionization efficiency or analyte suppres-

sion of Ala13-orcokinin. Thus, it is advantageous to use the

combination of direct tissue profiling and HPLC fraction-

ation of pooled tissue extract to generate a more complete

characterization of peptides present in the POs.

As demonstrated in Fig. 1, there are many mass spectral

peaks that do not correspond to previously identified

peptides. To characterize several of these new peptides, we

performed de novo MALDI PSD sequencing of a number of

unknown peaks fractionated from tissue extracts. To simplify

the complex mixture resulting from the PO tissue extract and

also concentrate the peaks, microbore LC separation was

performed prior to MALDI MS analysis. While on-line LC

coupling with ESI MS/MS has been the preferred method for

large-scale peptide identification from protein digests and

other complex mixtures, the off-line coupling MALDI PSD

approach has some advantages, including the ability to select

many more peaks for sequencing without time constraints.

This is particularly useful for analyzing complex tissue

extracts, where many peptides coelute and low abundance

peptides are often missed in LC ESI-MS/MS analysis due to

the preferential selection and identification of high-intensity

peaks in the elution time window. As evident from Fig. 3,

multiple PSD sequencing analyses were performed on

several precursor ions from individual LC fractions, yielding

enhanced chemical information from the limited amount of

samples. Because no genomic sequence information is

available for C. borealis, only peptides producing complete

fragmentation allowed the derivation of full amino acid

sequence. While all of the HPLC fractions were analyzed by

mass spectrometry, many LC fractions contained larger

652 L. Li et al.


peptides (precursor mass > 2000 Da) that fragment less

efficiently, which makes it difficult to generate complete

sequence information. Furthermore, several peptides were

eluted in a few consecutive LC fractions, so we were only

able to fully sequence seven new peptides based on PSD

fragmentation analysis, but many more peptides generated

partial sequence-specific fragmentation that did not yield

complete peptide identification. We have also demonstrated

the use of combination of MALDI-PSD and nanoESI CID

fragmentation techniques to generate complementary

sequence-specific fragmentation information to allow the

complete characterization of the peak at m/z 1030.51. While

CID generates more efficient fragmentation, the low mass

cut-off limitation associated with the ion trap mass analyzer

prevents the effective observation of low mass regions. In

contrast, PSD produces abundant immonium ions in the low

mass region that are indicative of amino acid compositions,

which are especially useful for de novo sequencing of

unknown peptides.

FMRFamide-like peptides

FMRFamide-like peptides are perhaps the most widely

distributed neuropeptides in the animal kingdom. Since their

first discovery in mollusks (Price and Greenberg 1977), a

large number of related peptides have been chemically

characterized in many different phyla (De Loof and Schoofs

1990; Krajniak 1991; Mercier et al. 1991; Mercier et al.

1993; Schoofs et al. 1997; Li et al. 1999a; Sithigorngul

et al. 2001; Baggerman et al. 2002). Two extended forms of

FLRFamide, TNRNFLRFamide and SDRNFLRFamide were

previously purified and sequenced in the lobster H. americ-

anus (Trimmer et al. 1987) and the crab C. borealis

(Weimann et al. 1993). Here we report a new member of

the extended FLRFamide peptide, NRNFLRFamide in the

C. borealis based on de novo sequencing analysis of the PO

extract. This peptide was previously isolated in the POs from

crayfish, P. clarkii (Mercier et al. 1993). This result is

consistent with previous observation that additional FMRF-

amide-like immunoreactive HPLC fractions were detected in

the crab nervous system (Marder et al. 1987; Weimann et al.

1993). Physiological experiments demonstrated that

NRNFLRFamide elicits effects on the pyloric and gastric

mill rhythms similar to those seen with TNRNFLRFamide

and SDRNFLRFamide. This may suggest that only the

sequence of RNFLRFamide is important for receptor recog-

nition and binding. It is also interesting to note that

NRNFLRFamide is often the most intense peak in the mass

spectra of the PO tissue samples, suggesting a differential

expression of these peptides in the POs and the STNS, and a

potential neurohormonal role of the peptide.

Corazonin

While corazonin has been reported in a number of insect

species (Veenstra 1989, 1991, 1994; Schoofs et al. 2000;

Hansen et al. 2001), the occurrence of corazonin in crusta-

cean species is documented for the first time in the current

study. Consistent with the highly conserved amino acid

sequence of this peptide throughout arthropod species, the

de novo sequencing of the putative corazonin peak in

C. borealis PO extract revealed the authentic form of this

peptide is also found in C. borealis. With the availability of

multiple corazonin gene and precursor sequences of several

insect species, it was possible to use a multiple sequence

alignment procedure to locate the highly conserved

sequence. We then employed this template sequence as a

guide to search for putative corazonin in crustacean species.

Despite the highly conserved peptide structure, the functions

of corazonin appear to be species specific and even tissue

specific. For example, this peptide is highly effective at

stimulating the activity of the heart and hyperneural muscle,

whereas other visceral muscles are completely insensitive to

this peptide (Predel and Eckert 2000). The corazonin

myostimulatory effect is mainly restricted to the American

cockroach, Periplaneta americana (Predel and Eckert 2000;

Predel et al. 2001), while in locust His7-corazonin induces

body color pigmentation (Tawfik et al. 1999; Schoofs et al.

2000). Electrophysiological experiments have demonstrated

strong modulatory effects on the pyloric rhythm of the

isolated STG upon application of synthetic corazonin (A.E.

Christie, unpublished observation), which adds to the wide

spectrum of the functions of the corazonin peptide.

CHH and other larger peptides in the PO

Although the major focus of the current study was charac-

terizing peptides in the lower mass range, several peaks were

detected in the mass range from 2000 to 9650 Da. One

notable peptide detected in the MALDI profiling of the POs

is a peak at 8562 Da, a possible candidate for CHH in

C. borealis. CHH are involved in various physiological

processes including regulation of blood glucose and lipids.

Since the first identification of a SG-CHH in the shore crab

Carcinus maenas, more than 20 SG-derived CHHs have been

isolated and identified in various species of crustacea (Keller

1992). While the primary source of CHHs is in the SG in the

eyestalk, multiple forms of novel CHH-like peptides were

recently reported in the POs from the shore crab, C. maenas

(Dircksen et al. 2001). Given the molecular masses of these

previously identified CHHs, it is possible that the peak at

8562 Da might be the CHH present in C. borealis with

several amino acid substitutions. Due to the inefficiency of

MS/MS fragmentation for peptides larger than 2500 Da, the

MS-based sequence analysis of the intact CHH peptide was

not possible; therefore, the conclusive peptide assignment

can not be made without additional confirmation. The

detection of a mass spectral peak at m/z 4072, which

corresponds to a putative peptide encoded by the CHH

precursor, supported the presence of CHH in the POs of

C. borealis. The identical molecular weight (and likely



sequence) of CPRP in C. borealis to that of CPRP from

C. maenas supports earlier observation that PO-CHH and

SG-CHH share an identical N-terminal sequence (positions

1–40), but differ considerably in the remaining sequence.

RYamides and peptide release

Finally, we sequenced several new peptides sharing a

common C-terminal, RYamides. While the physiological

effects of these newly sequenced peptides are not known, the

fact that these peptides are C-terminally amidated and

released upon high K+ depolarization strongly suggests

neurohormonal roles for these peptides. Several of these new

RYamide peptides are coreleased with NRNFLRFamide,

corazonin, proctolin, CCAP, and Ala13-orcokinin by high K+

depolarization in a Ca2+-dependent manner. This is also the

first demonstration of assaying peptide release from

C. borealis POs using mass spectrometric techniques. While

some variability was observed between different preparations

and the first and second high K+ depolarizations, several

peptides including proctolin, CabTRP1b, FVNSRYamide,

NRNFLRFamide, SGFYANRYamide, PAFYQSRYamide,

and Ala13-ocrokinin were always released. It is interesting

to note that CabTRP1b was consistently detected in releasate,

yet no CabTRP1a was detected. Previous studies on the two

tachykinin-related peptides showed that CabTRP1a is 20

times more abundant and 500 times more potent than

CabTRP1b (Christie et al. 1997). These results led to the

speculation that CabTRP1b is a breakdown product of

CabTRP1a. However, the detection of CabTRP1b in relea-

sates in the current study suggests that this shorter form of

the peptide may be directly cleaved from the precursor

protein and could serve a physiological role in C. borealis.

In summary, the combination of both direct tissue MALDI

profiling and MS-based sequencing allows comprehensive

characterization of the peptide complement in a nervous

system at higher throughput with greater chemical details.

The large-scale mass spectrometric investigation of neuro-

peptides and hormones in the pericardial organs of C. borealis

revealed much greater diversity and complexity of the

peptide messengers than had been previously demonstrated

by immunocytochemical and electrophysiological approa-

ches. We confirmed many of the previously known peptides

and unambiguously identified different chemical forms of the

peptide families. Furthermore, we have fully sequenced and

identified several new peptides. However, it is evident from

the current study that many more peptides remain unchar-

acterized. It is worth noting that members of several well-

known peptide families visualized in immunocytochemical

studies were not observed. This is likely due to amino acid

variations from the authentic forms found in other species,

post-translational modifications and perhaps cross-immuno-

reactivity and/or insufficient sensitivity of the current MS

methods. Future work will aim to characterize these peptides

by employing the combination of immunoaffinity and

MS-based sequencing approaches. Furthermore, the de novo

sequencing methodology will be coupled with database

searching via homology from related species whose genomic

sequences are available. Such peptidomic approaches prom-

ise to significantly accelerate the discovery of new peptides

and yield a complete description of the peptide signaling

molecules involved in the crustacean nervous system and to

further increase our understanding of peptide functions at the

network level.

Acknowledgements

This work was supported by National Institute of Neurological

Disorder and Stroke grants NS17813 (EM) and NS31609 (JVS). We

thank Dr Michael Nusbaum (University of Pennsylvania School of

Medicine) for the gift of synthetic corazonin.

References

Alexandrowicz J. S. and Carlisle D. B. (1953) Some experiments on the

function of the pericardial organs in Crustacea. J. Mar. Biol. Assoc.

UK 32, 175–192.

Baggerman G., Cerstiaens A., De Loof A. and Schoofs L. (2002) Pep-

tidomics of the larval Drosophila melanogaster central nervous

system. J. Biol. Chem. 277, 40368–40374.

Beltz B. S. and Kravitz E. A. (1983) Mapping of serotonin-like im-

munoreactivity in the lobster nervous system. J. Neurosci. 3,

585–602.

Berlind A. (1976) Neurohemal organ extracts effect the ventilation

oscillator in crustaceans. J. Exp. Zool. 195, 165–170.

Christie A. E. and Nusbaum M. P. (1995) Distribution and effects of

corazonin-like and allatotropin-like peptides in the crab stomato-

gastric nervous system. Soc. Neurosci. Abstr. 21, 629.

Christie A. E., Skiebe P. and Marder E. (1995) Matrix of neuromodu-

lators in neurosecretory structures of the crab, Cancer borealis.

J. Exp. Biol. 198, 2431–2439.

Christie A. E., Lundquist T., Nassel D. R. and Nusbaum M. P. (1997)

Two novel tachykinin-related peptides from the nervous system of

the crab Cancer borealis. J. Exp. Biol. 200, 2279–2294.

Cooke I. M. and Hartline D. K. (1975) Neurohormonal alteration of

integrative properties of the cardiac ganglion of the lobster

Homarus americanus. J. Exp. Biol. 63, 33–52.

De Loof A. and Schoofs L. (1990) Homologies between the amino acid

sequences of some vertebrate peptide hormones and peptides iso-

lated from invertebrate sources. Comp. Biochem. Physiol. B 95,

459–468.

Dircksen H., Bocking D., Heyn U., Mandel C., Chung J. S., Baggerman

G., Verhaert P., Daufeldt S., Plosch T., Jaros P. P. et al. (2001)

Crustacean hyperglycaemic hormone (CHH)-like peptides and

CHH-precursor-related peptides from pericardial organ neurose-

cretory cells in the shore crab, Carcinus maenas, are putatively

spliced and modified products of multiple genes. Biochem. J. 356,

159–170.

Evans P. D., Kravitz E. A., Talamo B. R. and Wallace B. G. (1976) The

association of octopamine with specific neurones along lobster

nerve trunks. J. Physiol. 262, 51–70.

Floyd P. D., Li L., Moroz T. P. and Sweedler J. V. (1999) Characteri-

zation of peptides from Aplysia using microbore liquid

chromatography with matrix-assisted laser desorption/ionization

time-of-flight mass spectrometry guided purification. J. Chroma-

togr. A 830, 105–113.

654 L. Li et al.


Garden R. W., Moroz L. L., Moroz T. P., Shippy S. A. and Sweedler J. V.

(1996) Excess salt removal with matrix rinsing: direct peptide

profiling of neurons from marine invertebrates using matrix-

assisted laser desorption/ionization time-of-flight mass spectrome-

try. J. Mass Spectrom. 31, 1126–1130.

Hansen I. A., Sehnal F., Meyer S. R. and Scheller K. (2001) Corazonin

gene expression in the waxmoth Galleria mellonella. Insect Mol

Biol. 10, 341–346.

Harris-Warrick R. M., Marder E., Selverston A. I. and Moulins M.

(1992) Dynamic Biological Networks. The Stomatogastric Nervous

System, p. 328. MIT Press, Cambridge.

Jimenez C. R., van Veelen P. A., Li K. W., Wildering W. C., Geraerts W.

P., Tjaden U. R. and van der Greef J. (1994) Neuropeptide

expression and processing as revealed by direct matrix-assisted

laser desorption ionization mass spectrometry of single neurons.

J. Neurochem. 62, 404–407.

Jorge-Rivera J. C. and Marder E. (1996) TNRNFLRFamide and

SDRNFLRFamide modulate muscles of the stomatogastric system

of the crab Cancer borealis. J. Comp. Physiol. A 179, 741–751.

Jorge-Rivera J. C., Sen K., Birmingham J. T., Abbott L. F. and Marder E.

(1998) Temporal dynamics of convergent modulation at a crusta-

cean neuromuscular junction. J. Neurophysiol. 80, 2559–2570.

Kaufmann R., Spengler B. and Lutzenkirchen F. (1993) Mass spectro-

metric sequencing of linear peptides by product-ion analysis in a

reflectron time-of-flight mass spectrometer using matrix-assisted

laser desorption ionization. Rapid Commun. Mass Spectrom. 7,

902–910.

Keller R. (1992) Crustacean neuropeptides: structures, functions and

comparative aspects. Experientia 48, 439–448.

Kobierski L. A., Beltz B. S., Trimmer B. A. and Kravitz E. A. (1987)

FMRFamidelike peptides of Homarus americanus: distribution,

immunocytochemical mapping, and ultrastructural localization in

terminal varicosities. J. Comp. Neurol. 266, 1–15.

Krajniak K. G. (1991) The identification and structure-activity relations

of a cardioactive FMRFamide-related peptide from the blue crab

Callinectes sapidus. Peptides 12, 1295–1302.

Kravitz E. A., Glusman S., Harris-Warrick R. M., Livingstone M. S.,

Schwarz T. and Goy M. F. (1980) Amines and a peptide as neu-

rohormones in lobsters: actions on neuromuscular preparations and

preliminary behavioural studies. J. Exp. Biol. 89, 159–175.

Li L., Moroz T. P., Garden R. W., Floyd P. D., Weiss K. R. and Sweedler

J. V. (1998) Mass spectrometric survey of interganglionically

transported peptides in Aplysia. Peptides 19, 1425–1433.

Li C., Nelson L. S., Kim K., Nathoo A. and Hart A. C. (1999a) Neu-

ropeptide gene families in the nematode Caenorhabditis elegans.

Ann. NY Acad. Sci. 897, 239–252.

Li L., Garden R. W., Romanova E. V. and Sweedler J. V. (1999b) In

situ sequencing of peptides from biological tissues and single

cells using MALDI-PSD/CID analysis. Anal. Chem. 71, 5451–

5458.

Li L., Garden R. W. and Sweedler J. V. (2000a) Single-cell MALDI: a

new tool for direct peptide profiling. Trends Biotechnol. 18, 151–

160.

Li L., Romanova E. V., Rubakhin S. S., Alexeeva V., Weiss K. R., Vilim

F. S. and Sweedler J. V. (2000b) Peptide profiling of cells with

multiple gene products: combining immunochemistry and MALDI

mass spectrometry with on-plate microextraction. Anal. Chem. 72,

3867–3874.

Li L., Pulver S. R., Kelley W. P., Thirumalai V., Sweedler J. V. and

Marder E. (2002) Orcokinin peptides in developing and adult

crustacean stomatogastric nervous systems and pericardial organs.

J. Comp. Neurol. 444, 227–244.

Lingle C. (1981) The modulatory action of dopamine on crustacean

foregut neuromuscular preparations. J. Exp. Biol. 94, 285–299.

Marder E. (1987) Neurotransmitters and neuromodulators, in The

Crustacean Stomatogastric Nervous System. A Model for the Study

of Central Nervous Systems. (Selverston, A. I. and Moulins, M.,

eds), pp. 263–300. Springer-Verlag, New York.

Marder E. and Hooper S. L. (1985) Neurotransmitter modulation of the

stomatogastric ganglion of decapod crustaceans, in Model Neural

Networks and Behavior (Selverston, A. I., ed.), pp. 319–337.

Plenum Press, New York.

Marder E. and Calabrese R. L. (1996) Principles of rhythmic motor

pattern generation. Physiol. Rev. 76, 687–717.

Marder E., Christie A. E. and Kilman V. L. (1995) Functional organ-

ization of cotransmission systems: lessons from small nervous

systems. Invert. Neurosci. 1, 105–112.

Marder E., Hooper S. L. and Siwicki K. K. (1986) Modulatory action

and distribution of the neuropeptide proctolin in the crustacean

stomatogastric nervous system. J. Comp. Neurol. 243, 454–467.

Marder E., Calabrese R. L., Nusbaum M. P. and Trimmer B. A. (1987)

Distribution and partial characterization of FMRFamide-like pep-

tides in the stomatogastric nervous systems of the rock crab,

Cancer borealis, and the spiny lobster, Panulirus interruptus.

J. Comp. Neurol. 259, 150–163.

Mercier A. J., Orchard I. and TeBrugge V. (1991) FMRFamide-like

immunoreactivity in the crayfish nervous system. J. Exp. Biol. 156,

519–538.

Mercier A. J., Schiebe M. and Atwood H. L. (1990) Pericardial peptides

enhance synaptic transmission and tension in phasic extensor

muscles of crayfish. Neurosci. Lett. 111, 92–98.

Mercier A. J., Orchard I., TeBrugge V. and Skerrett M. (1993) Isolation

of two FMRFamide-related peptides from crayfish pericardial

organs. Peptides 14, 137–143.

Meyrand P. and Marder E. (1991) Matching neural and muscle oscilla-

tors: control by FMRFamide-like peptides. J. Neurosci. 11, 1150–

1161.

Nusbaum M. P. and Beenhakker M. P. (2002) A small-systems approach

to motor pattern generation. Nature 417, 343–350.

Predel R. (2001) Peptidergic neurohemal system of an insect: mass

spectrometric morphology. J. Comp Neurol. 436, 363–375.

Predel R. and Eckert M. (2000) Neurosecretion: peptidergic systems in

insects. Naturwissenschaften 87, 343–350.

Predel R., Nachman R. J. and Gade G. (2001) Myostimulatory neuro-

peptides in cockroaches: structures, distribution, pharmacological

activities, and mimetic analogs. J. Insect. Physiol. 47, 311–324.

Price D. A. and Greenberg M. J. (1977) Purification and characterization

of a cardioexcitatory neuropeptide from the central ganglia of a

bivalve mollusc. Prep. Biochem. 7, 261–281.

Pulver S. R. and Marder E. (2002) Neuromodulatory complement of the

pericardial organs in the embryonic lobster, Homarus americanus.

J. Comp. Neurol. 451, 79–90.

Roepstorff P. and Fohlman J. (1984) Proposal for a common nomen-

clature for sequence ions in mass spectra of peptides. Biomed.

Mass Spectrom. 11, 601.

Schoofs L., Veelaert D., Vanden Broeck J. and De Loof A. (1997)

Peptides in the locusts, Locusta migratoria and Schistocerca gre-

garia. Peptides 18, 145–156.

Schoofs L., BaggermanG., Veelaert D., BreuerM., Tanaka S. andDe Loof

A. (2000) The pigmentotropic hormone [His(7)]-corazonin, absent

in a Locusta migratoria albino strain, occurs in an albino strain of

Schistocerca gregaria. Mol. Cell Endocrinol. 168, 101–109.

Schwarz T. L., Lee G. M., Siwicki K. K., Standaert D. G. and Kravitz

E. A. (1984) Proctolin in the lobster: the distribution, release, and

chemical characterization of a likely neurohormone. J. Neurosci. 4,

1300–1311.

Sithigorngul P., Saraithongkum W., Longyant S., Panchan N.,

Sithigorngul W. and Petsom A. (2001) Three more novel FMRF-



amide-like neuropeptide sequences from the eyestalk of the giant

freshwater prawn Macrobrachium rosenbergii. Peptides 22, 191–

197.

Siwicki K. K. and Bishop C. A. (1986) Mapping of proctolinlike im-

munoreactivity in the nervous systems of lobster and crayfish.

J. Comp. Neurol. 243, 435–453.

Siwicki K. K., Beltz B. S. and Kravitz E. A. (1987) Proctolin in iden-

tified serotonergic, dopaminergic, and cholinergic neurons in the

lobster, Homarus americanus. J. Neurosci. 7, 522–532.

Skiebe P. (1999) Allatostatin-like immunoreactivity within the stoma-

togastric nervous system and the pericardial organs of the crab

Cancer pagurus, the lobster Homarus americanus, and the crayfish

Cherax destructor and Procambarus clarkii. J. Comp. Neurol. 403,

85–105.

Skiebe P. (2001) Neuropeptides are ubiquitous chemical mediators:

using the stomatogastric nervous system as a model system. J. Exp.

Biol. 204, 2035–2048.

Skiebe P. (2003) Neuropeptides in the crayfish stomatogastric nervous

system. Microsc. Res. Tech. 60, 302–312.

Skiebe P., Dietel C. and Schmidt M. (1999) Immunocytochemical

localization of FLRFamide-, proctolin-, and CCAP- like peptides

in the stomatogastric nervous system and neurohemal structures of

the crayfish, Cherax destructor. J. Comp. Neurol. 414, 511–532.

Skiebe P., Dreger M., Meseke M., Evers J. F. and Hucho F. (2002)

Identification of orcokinins in single neurons in the stomatogastric

nervous system of the crayfish, Cherax destructor. J. Comp. Neu-

rol. 444, 245–259.

Stangier J., Dircksen H. and Keller R. (1986) Identification and

immunocytochemical localization of proctolin in pericardial organs

of the shore crab, Carcinus maenas. Peptides 7, 67–72.

Stangier J., Hilbich C., Dircksen H. and Keller R. (1988) Distribution of

a novel cardioactive peptide (CCAP) from pericardial organs of the

shore crab, Carcinus maenas. Peptides 9, 795–800.

Tawfik A. I., Tanaka S., De Loof A., Schoofs L., Baggerman G.,

Waelkens E., Derua R., Milner Y., Yerushalmi Y. and Pener M. P.

(1999) Identification of the gregarization-associated dark-pigmen-

totropin in locusts through an albino mutant. Proc. Natl Acad. Sci.

USA 96, 7083–7087.

Trimmer B. A., Kobierski L. A. and Kravitz E. A. (1987) Purification

and characterization of FMRFamide-like immunoreactivity sub-

stances from the lobster nervous system: isolation and sequence

analysis of two closely related peptides. J. Comp. Neurol. 266, 16–

26.

Veenstra J. A. (1989) Isolation and structure of corazonin, a cardio-

active peptide from the American cockroach. FEBS Lett. 250,

231–234.

Veenstra J. A. (1991) Presence of corazonin in three insect species, and

isolation and identification of [His7]corazonin from Schistocerca

americana. Peptides 12, 1285–1289.

Veenstra J. A. (1994) Isolation and structure of the Drosophila corazonin

gene. Biochem. Biophys. Res. Commun. 204, 292–296.

Weimann J. M., Marder E., Evans B. and Calabrese R. L. (1993) The

effects of SDRNFLRFamide and TNRNFLRFamide on the motor

patterns of the stomatogastric ganglion of the crab Cancer borealis.

J. Exp. Biol. 181, 1–26.

Worden M. K., Kravitz E. A. and Goy M. F. (1995) Peptide F1, an

N-terminally extended analog of FMRFamide, enhances contractile

activity in multiple target tissues in lobster. J. Exp. Biol. 198, 97–

108.

656 L. Li et al.


Mass spectrometric investigation of the neuropeptide complement and release in the pericardial...

Documents

Transcript of Mass spectrometric investigation of the neuropeptide complement and release in the pericardial...