Content uploaded by Susanne Neupert
Author content
All content in this area was uploaded by Susanne Neupert on May 08, 2020
Content may be subject to copyright.
Enhanced Coverage of Insect Neuropeptides in Tissue Sections by
an Optimized Mass-Spectrometry-Imaging Protocol
Alice Ly,*
,†,⊥
Lapo Ragionieri,*
,‡,⊥
Sander Liessem,
‡
Michael Becker,
†,§
Soren-Oliver Deininger,
†
Susanne Neupert,*
,‡
and Reinhard Predel*
,‡
†
Bruker Daltonik GmbH, Fahrenheitstraße 4, 28359 Bremen, Germany
‡
Department for Biology, Institute of Zoology, University of Cologne, 50674 Cologne, Germany
*
SSupporting Information
ABSTRACT: Mass spectrometry imaging (MSI) of neuropeptides has become a
well-established method with the ability to combine spatially resolved information
from immunohistochemistry with peptidomics information from mass spectro-
metric analysis. Several studies have conducted MSI of insect neural tissues;
however, these studies did not detect neuropeptide complements in manners
comparable to those of conventional peptidomics. The aim of our study was to
improve sample preparation so that MSI could provide comprehensive and
reproducible neuropeptidomics information. Using the cockroach retrocerebral
complex, the presented protocol produces enhanced coverage of neuropeptides at
15 μm spatial resolution, which was confirmed by parallel analysis of tissue extracts
using electrospray-ionization MS. Altogether, more than 100 peptide signals from
15 neuropeptide-precursor genes could be traced with high spatial resolution. In
addition, MSI spectra confirmed differential prohormone processing and distinct
neuropeptide-based compartmentalization of the retrocerebral complex. We believe that our workflow facilitates incorporation
of MSI in neuroscience-related topics, including the study of complex neuropeptide interactions within the CNS.
Neuropeptides are structurally diverse signaling molecules
that control and regulate essential physiological
functions in vertebrates and invertebrates, including growth,
feeding, reproduction, and environmental-stress tolerance. A
major source of neuropeptides is the central nervous system
(CNS), where neuropeptides can act as transmitters or
neuromodulators. Alternatively, neuropeptides can be pro-
duced in neurosecretory cells within the CNS and released as
peptide hormones into circulation, mostly from neurohemal
organs which function as hormone repositories. The large
number of neuropeptides and neuropeptide receptors generally
hamper decoding of coordinated peptide actions. Some
neuropeptide precursors may result in mature peptides that
activate different receptors (e.g. melanocyte-stimulating-
hormone precursors of vertebrates and CAPA precursors of
insects),
1−3
which further complicates the full recognition of
neuropeptide actions.
Mass spectrometry has increasingly been used to analyze the
neuropeptidome of the CNS, even to the single-cell level.
4,5
Although the aim of many of these approaches is to decipher
neuropeptide relationships or compensation strategies, a
number of limitations persist in the study of such complex
neuropeptide interactions. In insects, which include notable
model organisms in neuropeptide research, such as the fruit fly
Drosophila melanogaster and the honeybee Apis mellifera, small
tissue sizes, low peptide abundances, and complex cellular
patterns of peptidergic neurons still necessitate the extensive
use of immunohistochemistry (IHC) to complement neuro-
peptidomic studies. IHC has traditionally been used to
investigate neuropeptide distributions in the CNSs of insects,
but it has limited abilities for visualizing neuropeptides from
different precursors in the same sample, even when using
fluorochrome-coupled secondary antisera. IHC also usually
fails to discriminate between sequence-related precursor
products. In insects, this problem is evident for numerous
RFamides, which are both processed from single precursors
(up to 20 extended FMRFamides) and encoded by additional
genes, such as myosuppressin,sulfakinin,short neuropeptide F,
and long neuropeptide F. In this context, matrix-assisted laser
desorption/ionization mass spectrometry imaging (MALDI-
MSI) could be an ideal alternative technique for studying the
spatial distribution of neuropeptides in the nervous system.
6−8
MALDI-MSI been successfully employed for analyzing the
brain peptides of crustaceans.
9,10
Few studies have so far used
MALDI-MSI for the detection of insect neuropeptides, and the
reported tissue preparations and spatial resolutions (30 μm
and greater) are usually not sufficient to discriminate small
structures within the insect nervous system.
11,12
A study using
prototype MSI instrumentation examined lipid distributions in
20 μmD. melanogaster whole-body sections with pixel sizes
ranging between 5 and 10 μm and also reported several mass
matches with neuropeptides.
13
In all of these studies, however,
Received: September 20, 2018
Accepted: January 3, 2019
Published: January 3, 2019
Article
pubs.acs.org/ac
Cite This: Anal. Chem. 2019, 91, 1980−1988
© 2019 American Chemical Society 1980 DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
This is an open access article published under an ACS AuthorChoice License, which permits
copying and redistribution of the article or any adaptations for non-commercial purposes.
Downloaded via 134.214.214.72 on August 27, 2019 at 11:58:49 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
mass spectra did not detect neuropeptide contents in a manner
comparable to that obtained with the analysis of single
dissected neurons or with direct tissue profiling.
5,14
We investigated the suitability of MALDI-MSI for the
analysis of neuropeptide distributions in the retrocerebral
complex (RCC) of the American cockroach, Periplaneta
americana, a model organism in neuropeptide research, and
developed an optimized MSI protocol to analyze as much of
the neuropeptidome as possible. The RCC is the major
neuroendocrine organ in insects, comparable to the pituitary
gland of vertebrates. A recent MALDI-MSI study was able to
detect several neuropeptides in 30 year old paraffin-embedded
RCC samples.
15
In comparison with insect brains, whose
complexity hampers easy recognition of specific areas by MSI,
the organization of the RCC is easier to reconstruct and
facilitates reproducibility.
Figure 1 is an overview of the P. americana RCC and its
connections to key neurological structures. The RCC consists
of a pair of corpora cardiaca, which are fused posteriorly to a
pair of corpora allata. Whereas the corpora allata synthesize
sesquiterpenoids (juvenile hormones), the corpora cardiaca
exclusively release peptide hormones. Each corpus cardiacum
is subdivided antero-dorsally into a glandular part that
produces the insect equivalent of glucagon, the adipokinetic
hormones (AKHs);
16
the remaining parts of the corpora
cardiaca serve as neurohemal release sites of numerous peptide
hormones from the brain and subesophageal ganglion (SEG).
These hormones reach the RCC via different nervi corporis
cardiaci (NCC) and nervi corporis allati-2 (NCA-2).
17,18
Axons from NCA-2 as well as a number of axons from
neurosecretory cells of the brain cross the corpora allata and
contribute to neuropeptide detection along these glands. The
RCC is connected to the stomatogastric nervous system (SNS)
by way of the nervi cardiostomatogastrici (NCS). The
products of a large number of neuropeptide genes of insects
can be found in the RCC,
18−21
but the exact neuropeptide-
based compartmentalizationoftheRCCisstilllargely
unknown.
Our data demonstrate the extent to which MALDI-MSI with
commercially available instrumentation can be used to
reconstruct the distribution of neuropeptides in an insect
nervous system. With the described protocol, we obtained
good coverage of the neuropeptides expected to be present in
the RCC. In addition, the spatial distributions of further
neuropeptides could be verified for the first time in the RCC.
■EXPERIMENTAL SECTION
Chemicals and Reagents. α-Cyano-4-hydroxycinnamic
acid (CHCA) and peptide calibration standard II were
purchased from Bruker Daltonik GmbH (Bremen, Germany).
HPLC-grade ethanol and acetonitrile were obtained from
Honeywell (Seelze, Germany). Trifluoroacetic acid (TFA) was
purchased from Merck (Darmstadt, Germany). Standard food-
grade gelatin purchased from local supermarkets (Dr. Oetker
Gelatin, white; Bielefeld, Germany) was used in this study. An
ELGA Purelab flex system (Veolia; Celle, Germany) was used
to generate deionized water.
Animal Model and Sample Preparation. The animals in
this study were treated pursuant to the Declaration of Helsinki.
Cockroaches were raised and maintained at a constant
temperature (28 ±1°C) under a 12 h light−dark cycle with
free access to food and water. For experiments, adult
cockroaches were kept at 4 °C for 30 min before RCCs
were dissected in insect saline (126 mM NaCl, 5.4 mM KCl,
0.17 mM NaH2PO4, 0.22 mM KH2PO4; pH 7.4), rinsed in
deionized water, and embedded in 100 mg/mL gelatin/water.
The gelatin was dissolved in deionized water, heated to 80 °C
for 5 min to ensure dissolution, and then kept at 50 °Cto
maintain viscosity. For tissue embedding, 400 μL of dissolved
gelatin was poured into a handmade aluminum-foil mold with
an 8 mm internal diameter and allowed to solidify at room
temperature for at least 30 min. Subsequently, the RCCs were
placed horizontally on the solidified gelatin and then slowly
covered with 200 μL of gelatin at approximately 30 °C. The
embedded tissue was snap-frozen at −50 °C immediately after.
RCC samples were cryosectioned (−10 °C) at 14 or 20 μm
thickness with a 10-degree blade angle on a Microm 550
cryostat (Thermo Fisher; Walldorf, Germany), and thaw-
mounted onto indium tin oxide (ITO)-coated glass slides
(Bruker Daltonik). Finally, the samples were stored at −80 °C
until MSI measurement.
Prior to matrix application, the samples were removed from
the freezer, brought to room temperature, dried in a nitrogen-
rich environment using an ImagePrep device (Bruker
Daltonik), and stored under vacuum (300 mbar). Following
this, samples were either not washed at all or washed at room
temperature in 70% (v/v) ethanol/water and 100% ethanol for
20 s each with an interval of five seconds drying time between
each wash. The latter probes were dried again under vacuum
(300 mbar) for 1 h at room temperature. After optical scanning
(TissueScout; Bruker Daltonik), the sections were coated with
5 mg/mL CHCA dissolved in different ratios of acetonitrile/
water/TFA using a SunCollect Dispenser System (SunChrom;
Friedrichsdorf, Germany). The matrix was sonicated and
filtered before being sprayed. SunChrom spray control
software v2.5 was used to deposit eight layers of matrix
using variable spray rates (10, 20, 30, 40, and 50 μL/min; three
layers at 60 μL/min) at a speed of 900 mm/min. The line
Figure 1. Overview of the P. americana RCC (dorsal view) and its
junctions with brain and stomatogastric nervous system (SNS). The
black line indicates the area of the brain from which come the nerves
that supply the RCC with neurosecretion. Neurosecretory cells in the
pars intercerebralis and pars lateralis of the protocerebrum are
indicated by green and blue circles, respectively. Dotted lines
represent the respective pathways leading to the nervi corporis
cardiaci. NCC, nervus corporis cardiaci; NCA, nervus corporis allati;
NCS, nervus cardiostomatogastricus; SEG, subesophageal ganglion.
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1981
distance (Y-direction) was set to 2 mm, and the spray-nozzle
height (Z-position) was 25 mm.
Immunohistochemistry. Samples were fixed in 4%
paraformaldehyde diluted in phosphate-buffered saline (PBS,
pH 7.2) at 4 °C for 30 min and subsequently washed three
times in PBS for 30 min. The samples were preincubated with
5% normal goat serum dissolved in PBS for 30 min and then
incubated for 12 h at 4 °C in rabbit anti-P. americana
corazonin serum (1:4000, kindly provided by J. Veenstra) and
mouse anti-Diploptera punctata allatostatin A-7 serum (1:200,
5F10 kindly provided by B. Stay) diluted in PBS. Following
washing (3 ×30 min), the samples were incubated with goat
anti-mouse Cy2 (1:500)- and goat anti-rabbit Cy3 (1:3000)-
tagged secondary antibodies (Jackson Immuno Research; West
Grove, PA) at 4 °C for 12 h. Finally, samples were mounted in
Entellan and stored at 4 °C.
Image Processing. Immunostainings were examined with
a confocal laser-scanning microscope (ZEISS LSM 510 Meta
system; Jena, Germany), equipped with an Apochromat 10×/
0.45W (NA 0.45) objective using the multitrack mode. Cy2
was excited at 492 nm and emission collected with a BP 505−
530 filter, and Cy3 was excited at 543 nm and emission
collected via a LP 560 BP filter. Serial optical sections, each 0.9
μm thick, were analyzed and assembled into combined images
using the Zeiss LSM 5 image browser version 3. The final
figures were exported and processed to adjust brightness and
contrast with Adobe Photoshop 7.0 software (Adobe Systems;
San Jose, CA).
Preparation of RCC Extracts. Extracts of P. americana
RCC were prepared as described.
22
Briefly, a single P.
americana RCC was extracted in 20 μL of solution containing
50% methanol/water and 1% formic acid. Extracts were
sonicated for a few seconds and then centrifuged for 15 min at
13 000 rpm. Supernatants were transferred to fresh sample
tubes (0.5 mL) and either used for Quadrupole Orbitrap MS
or MALDI-TOF MSI. For MSI experiments, 0.3 μLof
supernatant was deposited on an ITO glass slide and allowed
to dry. Matrix was applied using the SunCollect sprayer as
described for imaging experiments. This procedure was
repeated three times to reduce batch effects, and the results
were compared. The extract was measured as a single spectrum
using the MSI-acquisition parameters as an imaged area
(average of 50 pixels per area).
MALDI-Mass-Spectrometry Imaging. MALDI-MSI was
performed using a rapifleX MALDI-TOF Tissuetyper mass
spectrometer (Bruker Daltonik) in positive-ion-reflector mode
over a mass range of m/z600−3200, with a 15 μm laser-spot
size and a 15 μm lateral step size. For each measurement
position, 500 laser shots were accumulated using a Smartbeam
3D Nd:YAG (355 nm) at a frequency of 5000 Hz and a sample
rate of 1.25 GS/s with baseline subtraction (TopHat) during
acquisition. The instrument was calibrated using peptide-
calibration standard II spotted onto the matrix-coated ITO
glass slide, taking care that the spot did not obscure the tissue.
Ion images were generated using flexImaging v. 5.0 and SCiLS
Lab MVS software version 2018a (Bruker Daltonik) with the
data normalized to the total ion count (TIC). Reduced data
(Bruker DAT files) were uploaded and preprocessed for a
time-of-flight (TOF) instrument in SCiLS and underwent
spatial-segmentation analysis using a bisecting-k-means-with-
correlation-distance approach.
23
The default pipeline was used
with the following modifications: medium denoising and ±0.30
Da interval width. Ten RCC preparations with the
corresponding sections (3−5 sections of each RCC prepara-
tion) were analyzed using the MSI protocol.
Quadrupole Orbitrap Mass Spectrometry. The RCC
extract was analyzed with a Q-Exactive Plus (Thermo Fisher
Scientific; Waltham, MA) as described.
24
Prior to injection, the
sample was desalted using self-packed Stage Tip C18 spin
columns. Using PEAKS 8.5 (PEAKS Studio, BSI; Waterloo,
Canada) and MaxQuant (v. 1.6.2.10, MPI; Martinsried,
Germany), MS2spectra were compared to an internal database
containing known P. americana neuropeptides
14
and the newly
annotated neuropeptide-like precursor 1 (NPLP1,
MH837510). For both pipelines, the maximum number of
PTMs (sulfation of Tyr, C-terminal amidation, cystine
formation, oxidation of Met and Trp, pyroglutamyl formation
on Glu and Gln, and N-terminal acetylation of Lys) per
peptide was set at five and no digestion mode was selected. For
analyses using MaxQuant, the first-search peptide tolerance
was set at 20 ppm, and the main-search peptide tolerance was
set at 4.5 ppm. The false-discovery rate (FDR) was set to 0.01
for peptide-spectrum match, and only peptides with a P-score
>60 were considered for manual inspection. For the peptide
search using PEAKS, the parent-mass error tolerance was set at
10 ppm, and the fragment-mass error tolerance was set at 0.05
Da. FDR below 1% and fragment spectra with a peptide score
(−10 log P) equivalent to a P-value of about 1% were selected
and manually reviewed.
Statistics. Paired ttests were used to calculate the effects of
different method parameters (GraphPad Prism (v. 5.04); San
Diego, CA).
■RESULTS AND DISCUSSION
Conceptualization of an Imaging Protocol for Insect-
Neuropeptide Analysis. To yield as much neuropeptidomic
information as possible, we tested different approaches for
sample preparation utilized throughout the MALDI-MSI field
(for a review, see Buchberger et al.).
7
As a guide, we used the
neuropeptide complement from an RCC extract detected with
the same setup for sample preparation (including the matrix
sprayer, matrix composition, and matrix-application proce-
dure) and MALDI-TOF (the same mass analyzer, ionization
technique, sample-target device, and instrument settings) that
were used for the imaging experiments. No significant
degradations of peptides in any of our experiments were
found, an obvious advantage of using insect-tissue samples.
Analysis of the tissue extracts revealed 60 mature neuro-
peptides that could be assigned to 15 neuropeptide precursor
genes of P. americana (Table S1). Assignment of ion signals to
neuropeptides of P. americana was supported by MS2data from
quadrupole orbitrap analyses of an RCC extract (Figure S1).
For our first MSI experiments with RCC sections, we adapted
a protocol for MALDI imaging of the honeybee brain
11
but
with a matrix sprayer for deposition of CHCA instead of a
matrix spotter. This approach revealed peptidomic information
that was much less comprehensive than that obtained from
RCC extracts. In fact, only few abundant peptides were
detected with weak spatial distributions (data not shown).
Therefore, we re-evaluated each experimental step to reach an
optimized MSI protocol suitable for comprehensive analysis of
neuropeptides in RCC tissue sections (see also Figure S2).
Step 1: Dissection. To avoid excessive release of peptides
during dissection, we used a cold saline solution during
preparation of the RCC. Before embedding, the isolated RCC
was washed in ice-cold deionized water for a few seconds to
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1982
remove the saline solution and to avoid salt-crystal formation
during the embedding−freezing steps. As a general rule, short
dissection times of less than 5 min and strict avoidance of
direct contact of the RCC tissue with the forceps resulted in
more consistent mass spectra along the complete tissue
sections. For the transfer, we used the nerves that leave the
RCC toward the periphery.
Step 2: Embedding. The small size of the RCC
(approximately 0.5 ×1 mm) made it necessary to embed
the tissue prior to sectioning. Gelatin is an embedding
substrate known to be compatible with MALDI-MSI.
9,25
Chen et al. used 100 mg/mL gelatin in water for MALDI-
MSI of crustacean-brain neuropeptides;
9
this concentration
also worked with the much smaller RCC tissue. We also tested
gelatin concentrations ranging between 80 to 120 mg/mL
without obtaining better results; sectioning quality decreased
with lower gelatin concentrations, whereas the increased
density of more concentrated gelatin reduced the ability to
properly embed the samples. Accurate horizontal placement of
the RCC, which consists of two mirror-imaged parts, facilitated
quality control by comparison of peptide distribution in the
two parts. We found that optimal positioning was achieved by
using two layers of gelatin. The RCC was placed and oriented
on the solid lower layer; any remaining water was carefully
removed around the RCC by using a glass capillary before the
sample was slowly covered with the fluid (warmer) gelatin.
Step 3: Cryosectioning. MSI of whole RCC has previously
been reported,
12
but thicker tissues are not highly electrically
conductive, which can result in poor spectra.
26
In addition,
only peptides located at the outer margin of the RCC are likely
to be analyzed when performing whole-tissue profiling
combined with matrix spraying. In order to obtain uniform
tissue sections with reproducible mass-spectrometry profiles,
tests were performed using different section thicknesses (5−20
μm), cutting temperatures (−20 to −10 °C), and blade cutting
angles (5−20°). The best section quality without folding or
squeezing tissue sections was achieved with a blade angle of
10°, a temperature of −10 °C, and a tissue thickness of 14−20
μm; although cutting thinner sections was possible, it was
more difficult to maintain tissue integrity. Tissue sections were
serially collected on ITO glass slides using manual control of
cutting pace and stored at −80 °C. For optimal peptide
coverage in mass spectra, the samples were dried under
vacuum at about 300 mbar for at least 12 h after defrosting.
Shorter drying times (e.g., 1 h) decreased the peptide coverage
significantly (p= 0.0006, Figure S3).
Step 4: Ethanol Washes of Tissue Sections. Neuropeptide
analysis by MSI from crustacean neuronal tissues was reported
without washing for whole tissues and sections prior to matrix
application;
9
MSI of neuropeptides from whole Drosophila
sections was performed without washing the tissue sections,
but the complete animals were immersed in ethanol prior to
sectioning.
13
The use of ethanol washes has been reported to
remove lipids and salts that can interfere with peptide and
protein signals.
11,27
We observed statistically significant lower
neuropeptide coverage and strong interference from lipid
species when washing was omitted (ttest, p= 0.0004, Figure
S4). Two consecutive ethanol washes using first 70% (v/v)
ethanol/water followed by absolute ethanol for 20 s each
provided the best coverage of neuropeptides in the samples
(Figure S4). After washing, the sections were dried again for at
least 1 h at 300 mbar to ensure removal of ethanol. The
ethanol concentrations used in our MSI experiments have been
reported for the detection of intact proteins
28
and slightly
modified from those used in the detection of A. mellifera brain
neuropeptides.
11,29
For the comparison of peptide coverage,
we selected regions of interest (ROIs; 200 ×200 μm) within
corpora allata tissue that showed more uniform peptidome in
consecutive sections compared with other parts of the RCC.
The average number of peptides identified within the ROIs
was significantly higher for washed samples (10.63 ±1.133, N
= 8) than for samples prepared without washing steps (4.625 ±
0.7055, N= 8). The good neuropeptide coverage in the mass
spectra obtained from the washed samples was accompanied
by a slight decrease in the resolution of MSI ion maps. For
those peptides that were detectable in sections without
washing (e.g., pyrokinins), we therefore used both approaches
in parallel. The obtained differences in resolution indicate a
certain degree of delocalization of peptides during the washing.
Therefore, further decreasing the laser-spot size or the size of
the matrix crystals might have little effect in these samples.
Step 5: Matrix Application. We used CHCA, which is a
matrix preferred for the detection of peptides and small
proteins
30
and has been reported for detecting neuropeptides
in MSI experiments on honeybee brains
11,29
and flatworms.
31
For high-spatial-resolution measurements, smaller matrix-
crystal sizes are desirable as large crystals can lead to analyte
spread and require more energy for ionization.
32
The spraying
device employed in this study has previously been used for
detecting small molecules,
33,34
N-glycans, tryptic peptides,
34
and insect neuropeptides.
11,29
The chosen parameters (spray
rate, spray-head speed, spray-head distance from sample, and
number of cycles) were selected on the basis of a combination
of visual inspection of matrix deposition (size, even
distribution, no convergence of droplets) during the cycles
and analysis of peptide yields in subsequent mass-spectrometry
experiments. For example, when altering spray speeds and
rates, we ensured that the sprayed layers were completely dry
before starting the next layer and compared how many
neuropeptide signals were detected. In order to ensure
reproducible results, an ITO glass slide was always coated
with matrix to test the functionality of the sprayer. For that, the
glass slide was weighed, coated with matrix, and weighed again
to estimate the amount of deposited matrix. Coating of
samples commenced if the matrix evenly covered the slide and
weighed between 0.9−1.0 mg. The initial matrix composition,
5 mg/mL CHCA dissolved in 70% ACN/H2O with 0.1% TFA,
was successively modified to 5 mg/mL in 50% ACN/H2O with
2% TFA. Increasing TFA concentrations resulted in higher
neuropeptide signal intensities in the MSI spectra and a
significant increase in peptide coverage (p= 0.0167, Figure
S5). Using these spraying conditions, we obtained matrix
crystals of about 20 μm, which corresponds roughly to the 15
μm laser-spot size used in our analyses. Smaller laser-spot sizes
(5−10 μm) were tested but failed to generate a full peptidome
in subsequent mass spectra.
High-Spatial-Resolution MALDI-MSI of Multicopy
Peptides in the RCC. A number of mature insect neuro-
peptides are processed as multiple copies (paracopies) from
precursor proteins. These paracopies are usually processed in
equimolar ratios. Among the known cockroach neuropeptides
present in the RCC, allatostatin A (AstA) peptides, extended
FMRFamides (FMRFs), myoinhibitory peptides (MIPs),
sulfakinins (SKs), kinins, and pyrokinins (PKs) have several
paracopies ranging in number from two (SKs) to 21
(FMRFs).
14
These paracopies ideally show (1) constant
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1983
relative signal intensities in the MSI spectra and (2) identical
spatial distributions. Hence, analysis of paracopies in MSI
spectra provides information regarding spectra quality.
MSI ion maps from a single section consistently verified
similar distributions of the different FMRF paracopies, which
all have specific sequences in P. americana (Figure 2A). The
presence of FMRF ion signals was consistent in the mass
spectra of the MSI tissue sections (Figure 2B) and the extract
samples that had been spotted on ITO glass slides before being
analyzed with the same MALDI-TOF equipment (Figure 2C).
Analysis of all sections of single RCCs also revealed the overall
distribution of these peptides along the RCC. It has to be
noted that the distribution of FMRFs in the RCC was resolved,
although these peptides showed low signal intensities in the
mass spectra from the RCC extracts (Figure 2C).
Differential Distribution of Neuropeptides in the
RCC−SNS. Subsequent to the confirmation that MSI spectra
show reliable spatial distributions of neuropeptide paracopies,
we analyzed the distributions of all mature neuropeptides
detected in our MSI spectra. Altogether, we observed ion
signals of 57 mature neuropeptides, a number that matches
well the number of neuropeptide ion signals obtained by
extract analysis (Table S1). The number of observed peptides
exceeded 100 if additional precursor peptides (cleavage
products without known functions) were included.
Figure 3 exemplarily shows neuropeptide distributions in
two RCC sections. The local accumulation of peptides from
various neuropeptide genes within the RCC differed
dramatically and recalled some old neuroanatomical studies
that described distinct axonal pathways within the seemingly
uniform RCC.
35,36
The ion maps only partially matched the
few available immunostainings depicting the distribution of
neuropeptides in the RCC, but they correspond to the data
obtained by direct tissue profiling of parts of the RCC and
SNS.
37
A brief summary of the distribution patterns of
neuropeptides in the RCC−SNS, as revealed by MSI, is given
hereafter.
Myosuppressin and sk genes are both expressed in neuro-
secretory cells of the pars intercerebralis in the protocerebrum
with projection into the RCC via NCC-1. The neuropeptides
from these genes showed distinct accumulations within the
RCC (Figure 3A,B) that were different from each other and
from those of the other RFamide peptides, such as short
neuropeptide F (Figure 3C), and FMRFs (Figure 3F). For
kinins and MIPs, which were found by IHC in cells of the pars
lateralis and pars intercerebralis,
18,38
we also observed different
distribution patterns. Kinin accumulation was restricted to the
corpora cardiaca, whereas MIPs were more abundant along the
SNS (Figure 3D,E). As shown for PK-3, PKs were most
abundant around the corpora allata (Figure 3G); a detailed
description of the distribution of PKs is given in the following
section on prohormone processing. The two AKH peptides,
which are products of different genes, were restricted to the
glandular antero-dorsal part of the RCC (Figure 3J). The only
peptide entirely restricted to the SNS was proctolin (Figure
3K), whereas CCAP was observed in the neurohemal part of
the corpora cardiaca only (Figure 3L).
We also observed the spatial distributions of neuropeptides
not previously described by mass spectrometry in the RCC of
P. americana, such as allatotropin (AT; Figures 3I and S6) and
FMRFs (see above, Figure 2). AT showed a distribution
different from all other neuropeptides. Prominent ion signals
were detected both in the SNS and in the corpora cardiaca but
rarely in the corpora allata (Figure 3I). It is unknown how AT
enters the RCC, but on the basis of the MSI information, it
seems possible that AT reaches the corpora cardiaca through
NCC-1. This assumption was substantiated by direct peptide
profiling of isolated NCC-1 using conventional MALDI-TOF
mass spectrometry (Figure S7). In addition to known
cockroach peptides, multiple products of the neuropeptide-
like precursor 1 (NPLP1), not reported in P. americana so far,
could be identified. Peptides from NPLP1 precursors have
been found in the CNS and RCC in several insects.
39−41
The
mass signals of 13 NPLP1 peptides were detected with
distributions mostly restricted to the neurohemal part of
corpora cardiaca (Figures 3H and S1 and Table S1).
Corazonin and AstA peptides, both of which were detected
using IHC in cells of the pars lateralis of the protocerebrum
Figure 2. FMRF paracopies in mass spectra from RCC preparations.
(A) MSI from a single tissue section showing the distributions of four
FMRFs, suggesting identical spatial distributions of these peptides in
the RCC. Section: 20 μm, scale bar: 200 μm, ion-intensity bar: 100−
20%. (B) Mass spectrum obtained by means of MSI. The analyzed
spot is indicated in (A) by an arrow. (C) Mass spectrum obtained by
means of MSI of an aliquot of an RCC extract spotted on an ITO
glass slide. The matrix-spraying and MALDI-TOF equipment were
exactly the same as those as used for (B). The accuracy of mass
matching for peptide assignment was settled at ±0.25 Da.
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1984
with projection via NCC-2 into the RCC,
42,43
showed different
distribution patterns. Corazonin signals were highly abundant
in the neurohemal part of the corpora cardiaca, the nervus
cardiostomatogastricus, and adjacent parts of the SNS (Figure
S6) but mostly not detectable along the corpora allata. This
was the other way around with AstA signals, which were weak
in the mass spectra of the neurohemal part of the corpora
cardiaca but, in most preparations, distinct around the corpora
allata and the posteriorly directed part of the SNS (nervus
esophageus). These differences were not expected according to
IHC analyses. We therefore performed AstA−corazonin IHC
double staining on peripheral RCC sections with less complex
axon pathways and compared the staining patterns with the
MSI images from consecutive sections (Figure 4). The
resulting data confirmed that AstA and corazonin indeed
have different spatial distributions along the RCC (see also
Figure S6).
Differential Prohormone Processing. An advantage of
MSI is the capability of detecting differential prohormone
processing. In P. americana,differential processing is only
demonstrated for the PK precursor. Although the full set of
PKs is processed in cell clusters of the SEG with projections to
Figure 3. MALDI-MSI ion maps confirming the differential distribution within the RCC−SNS of neuropeptides from 12 different genes. (A) Pea-
SK, m/z1443.6 ±0.25 Da, ion-intensity bar: 100−20%. (B) Myosuppressin (pQ), m/z1257.6 ±0.25 Da, ion-intensity bar: 100−20%. (C) Short
neuropeptide F, m/z1315.7 ±0.25 Da, ion-intensity bar: 100−20%. (D) Kinin-1, m/z949.5 ±0.25 Da, ion-intensity bar: 100−40%. (E) MIP-2,
m/z1389.6 ±0.25 Da, ion-intensity bar: 100−35%. (F) FMRF-15, m/z1159.6 ±0.25 Da, ion-intensity bar: 100−20%. (G) PK-3, m/z996.6 ±
0.25 Da, ion-intensity bar: 100−20%. (H) NPLP-1, m/z1585.8 ±0.25 Da, ion-intensity bar: 100−20%. (I) Allatotropin, m/z1366.7 ±0.25 Da,
ion-intensity bar: 100−20%. (J) AKH-1, m/z973.5 ±0.20 Da, ion-intensity bar: 100−10%. (K) Proctolin, m/z649.4 ±0.25 Da, ion-intensity bar:
100−20%. (L) CCAP, m/z956.5 ±0.25 Da, ion-intensity bar: 100−40% (see Figure 1 for an overview of the architecture of RCC−SNS). Scale bar
(white): 600 μm, section thicknesses: (A−I) 20 μm and (J−L) 14 μm.
Figure 4. Distribution of corazonin and AstA analyzed in serial RCC sections by (A) immunohistochemistry and (B) MSI (the more peripheral
section). Data obtained by both methods confirmed the different spatial distributions of corazonin and AstA, which are produced in cells of the pars
lateralis of the brain, along the RCC. Labeling on the RCC margin is likely due to autofluorescence (detached gelatin). Scale bar: 200 μm, section
thickness: 20 μm. Ion-intensity bar: 100−20%. The accuracy of mass matching for peptide assignment was settled at ±0.25 Da.
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1985
the RCC via NCA-2, few cells in the brain with projections
into the RCC via NCC-1 do not process PK-1, and mass
spectra of NCC-1, therefore, did not show ion signals of this
PK.
44
MSI spectra from the RCC confirmed these data (Figure
5). In the anterior corpora cardiaca near the junction with
NCC-1, all PKs except PK1 were detectable, whereas all PKs
were prominent in the neurohemal part of the RCC near the
entrance of NCA-2 and in the corpora allata.
Two of the PKs (PK-1 and PK-3) have ion signals with
similar masses to those of sodium- and potassium-adduct ions
of AKHs. In MALDI-TOF mass spectrometry, AKHs are
represented only by these adducts, but ion maps verified that
even a mass difference of only 0.2 Da was sufficient to
discriminate between these peptides (Figure S8). A previous
MSI study of whole RCC tissue showed similar ion-signal
discrimination.
12
Once the accuracy of our MSI analyses was confirmed, we
tested a bioinformatic approach based on the information
obtained in MSI experiments. Application of spatial-segmenta-
tion analysis to a single RCC section enabled the assignment of
the main compartments within the RCC−SNS (Figure 6).
These regions correspond to the corpora allata and adjoining
nervi corporis allati-1, the corpora cardiaca glandular area, the
corpora cardiaca neurohemal area, and the SNS. This result
demonstrates statistical discrimination among the different
areas, independent from any a priori knowledge. Interestingly,
the neurohemal area within the corpora cardiaca is further
differentiated in three subcompartments; an anterior part
surrounding the glandular tissue of the corpora cardiaca, a
posterior portion with nervi cardiostomatogastrici and
adjoining tissue, and the portion of the corpora cardiaca
located between these parts. The respective dendrogram shows
that the latter two compartments are more closely related to
each other than to the neuroglandular area or the SNS.
■CONCLUSIONS
This study provides an MSI workflow for analysis of
neuropeptides in insect neuroendocrine tissues that results in
a comprehensive neuropeptidome with high reproducibility,
ion-signal quality, and spatial resolution. Seemingly minor
changes of established protocols produced an overall view of
neuropeptide distributions with high-spatial resolution using
conventional MALDI-TOF mass-spectrometry equipment.
Novelties included the distinct accumulation of different
neuropeptides in the RCC−SNS, which even holds for
neuropeptides produced in different cell populations within a
cell cluster.
MSI experiments can potentially be incorporated in
neuroscience-related topics such as complex changes in the
neuropeptidome of insects that might be associated with
development or adaptations induced by environmental stress
(e.g., xenobiotics). The presented sample-preparation protocol
can certainly be used for other MALDI-MSI instrumentation,
including those with higher mass or spatial resolutions (e.g.,
MALDI-FT-ICR and AP-SMALDI-Orbitrap). Enhanced later-
al resolution may be possible particularly in combination with
spraying devices that result in smaller matrix-crystal sizes or
sublimation/re-extraction procedures.
45−47
Commonly used
washing steps prior to matrix application potentially result in
analyte spreading and therefore might neutralize attempts to
obtain better lateral resolution. For RCC tissue, alternative
Figure 5. (A) Ion maps of four PKs indicating differential processing of the PK precursor. (B) Four PKs detected in the posterior part of the RCC,
which mostly contains PKs processed in cells of the SEG. (C) Anterior corpus cardiacum tissue, which receives neuropeptides from the brain,
showing no PK-1 ion signals. Section thickness: 20 μm; scale bar: 200 μm; ion-intensity bar: 100−20%, except for m/z883.5 (100−35%). The
accuracies of mass matching for peptide assignment were settled at ±0.25 Da for PK-2, -3, and -4 and at ±0.001 Da for PK-1. Tissue sections were
not washed with ethanol prior to matrix spraying.
Figure 6. Spatial segmentation analysis of MSI data from a single
RCC section. Different levels in the segmentation dendrogram
represent distinct regions of the RCC corresponding to the corpora
allata (CA) and nervi corporis allati-1 (NCA-1) and the glandular and
neurohemal corpora cardiaca (CC). The neurohemal part of the
corpora cardiaca is further subdivided into three subcompartments.
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1986
tests without washes are rewarding, given that the peptide of
interest is detectable with spatial resolution. If peptidomics
information needs to include more extensive neuropeptide
complement, washing steps are indispensable. In this context,
the experiments presented in this study may serve as a guide
when starting with other tissue preparations.
■ASSOCIATED CONTENT
*
SSupporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.anal-
chem.8b04304.
List of mature neuropeptides from 15 precursor genes of
P. americana; quadrupole orbitrap MS2spectra of P.
americana neuropeptides; workflow for MSI sample
preparation optimized for insect neuroendocrine tissue
(RCC); comparison of peptide coverage in tissue
sections dried for 1 or 12 h before washing; comparison
of peptide coverage in tissue sections with and without
successive ethanol washes; comparison of peptide
coverage in tissue sections after matrix spraying (5
mg/mL CHCA in 50% ACN/H2O) with matrix solution
containing 0.1 or 2% TFA; distribution of allatostatinA-
11, corazonin, and allatotropin in consecutive RCC
sections; MALDI-TOF direct tissue profiling of a
dissected nervus corporis cardiaci 1; and discrimination
between mass-similar neuropeptides (PDF)
■AUTHOR INFORMATION
Corresponding Authors
*E-mail: rpredel@uni-koeln.de. Tel.: +49-221-470-5817
(R.P.).
*E-mail: mail@susanne-neupert.de. Tel.: +49-221-470-8267
(S.N.).
*E-mail: alice.ly@bruker.com. Tel.:+49-421-220-54782 (A.L.).
*E-mail: lapo.ragionieri@uni-koeln.de. Tel.: +49-221-470-
8592 (L.R.).
ORCID
Lapo Ragionieri: 0000-0003-0099-2719
Sander Liessem: 0000-0002-7073-2659
Susanne Neupert: 0000-0003-1562-5743
Present Address
§
M.B.: Boehringer Ingelheim Pharma GmbH & Company KG,
88397 Biberach an der Riss, Germany
Author Contributions
⊥
A.L. and L.R. contributed equally. All authors contributed to
the writing and have given approval to the final version of the
manuscript.
Notes
Theauthorsdeclarethefollowingcompetingfinancial
interest(s): A.L. and S.O.D. were employees of Bruker
Daltonik GmbH for the duration of this study. M.B. was an
employee of Bruker for part of this study.
■ACKNOWLEDGMENTS
This project was supported by a European Commission
Horizon2020 Research and Innovation Grant 634361 (nEU-
ROSTRESSPEP); the German Research Foundation (PR 766/
11-1); and the Graduate School for Biological Sciences,
Cologne (DFG-RTG 1960: Neural Circuit Analysis of the
Cellular and Subcellular Level). We thank Susanne Hecht
(Bruker Daltonik GmbH) for help in sample preparation.
■REFERENCES
(1) Cazzamali, G.; Torp, M.; Hauser, F.; Williamson, M.;
Grimmelikhuijzen, C. J. P. Biochem. Biophys. Res. Commun. 2005,
335,14−19.
(2) Iversen, A.; Cazzamali, G.; Williamson, M.; Hauser, F.;
Grimmelikhuijzen, C. J. P. Biochem. Biophys. Res. Commun. 2002,
299, 628−633.
(3) Smith, M. E. POMC Opioid Peptides. In Handbook of
Biologically Active Peptides; Kastin, A. J., Ed.; Elsevier: San Diego,
2006; pp 1325−1331.
(4) Rubakhin, S. S.; Sweedler, J. V. Nat. Protoc. 2007,2, 1987−97.
(5) Neupert, S.; Johard, H. A. D.; Nassel, D. R.; Predel, R. Anal.
Chem. 2007,79, 3690−3694.
(6) Chen, R.; Li, L. Anal. Bioanal. Chem. 2010,397, 3185−3193.
(7) Buchberger, A. R.; DeLaney, K.; Johnson, J.; Li, L. Anal. Chem.
2018,90, 240−265.
(8) Spengler, B. Anal. Chem. 2015,87,64−82.
(9) Chen, R.; Cape, S. S.; Sturm, R. M.; Li, L. Methods Mol. Biol.
2010,656, 451−463.
(10) Chen, R.; Jiang, X.; Prieto Conaway, M. C.; Mohtashemi, I.;
Hui, L.; Viner, R.; Li, L. J. Proteome Res. 2010,9, 818−832.
(11) Pratavieira, M.; da Silva Menegasso, A. R.; Garcia, A. M. C.; dos
Santos, D. S.; Gomes, P. C.; Malaspina, O.; Palma, M. S. J. Proteome
Res. 2014,13, 3054−3064.
(12) Verhaert, P. D. E. M.; Pinkse, M. W. H.; Strupat, K.; Conaway,
M. C. P. Methods Mol. Biol. 2010,656, 433−449.
(13) Khalil, S. M.; Pretzel, J.; Becker, K.; Spengler, B. Int. J. Mass
Spectrom. 2017,416,1−19.
(14) Neupert, S.; Fusca, D.; Schachtner, J.; Kloppenburg, P.; Predel,
R. J. Comp. Neurol. 2012,520, 694−716.
(15) Paine, M. R. L.; Ellis, S. R.; Maloney, D.; Heeren, R. M. A.;
Verhaert, P. D. E. M. Anal. Chem. 2018,90, 9272−9280.
(16) Gade, G.; Hoffmann, K. H.; Spring, J. H. Physiol. Rev. 1997,77,
963−1032.
(17) Willey, R. B. J. Morphol. 1961,108, 219−261.
(18) Predel, R.; Rapus, J.; Eckert, M. Peptides 2001,22, 199−208.
(19) Clynen, E.; Schoofs, L. Insect Biochem. Mol. Biol. 2009,39,
491−507.
(20) Siegmund, T.; Korge, G. J. Comp. Neurol. 2001,431, 481−491.
(21) Wegener, C.; Reinl, T.; Jansch, L.; Predel, R. J. Neurochem.
2006,96, 1362−1374.
(22) Vogt, M. C.; Paeger, L.; Hess, S.; Steculorum, S. M.; Awazawa,
M.; Hampel, B.; Neupert, S.; Nicholls, H. T.; Mauer, J.; Hausen, A.
C.; et al. Cell 2014,156, 495−509.
(23) Trede, D.; Schiffler, S.; Becker, M.; Wirtz, S.; Steinhorst, K.;
Strehlow, J.; Aichler, M.; Kobarg, J. H.; Oetjen, J.; Dyatlov, A.; et al.
Anal. Chem. 2012,84, 6079−6087.
(24) Liessem, S.; Ragionieri, L.; Neupert, S.; Buschges, A.; Predel, R.
J. Proteome Res. 2018,17, 2192−2204.
(25) Altelaar, A. F. M.; van Minnen, J.; Jime
nez, C. R.; Heeren, R.
M. A.; Piersma, S. R. Anal. Chem. 2005,77, 735−741.
(26) Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. J. Mass Spectrom.
2003,38, 699−708.
(27) Seeley, E. H.; Oppenheimer, S. R.; Mi, D.; Chaurand, P.;
Caprioli, R. M. J. Am. Soc. Mass Spectrom. 2008,19, 1069−1077.
(28) Meding, S.; Walch, A. Methods Mol. Biol. 2012,931, 537−546.
(29) Pratavieira, M.; Menegasso, A. R. da S.; Esteves, F. G.; Sato, K.
U.; Malaspina, O.; Palma, M. S. J. Proteome Res. 2018,17, 2358−2369.
(30) Cohen, S. L.; Chait, B. T. Anal. Chem. 1996,68,31−37.
(31) Ong, T.-H.; Romanova, E. V.; Roberts-Galbraith, R. H.; Yang,
N.; Zimmerman, T. A.; Collins, J. J.; Lee, J. E.; Kelleher, N. L.;
Newmark, P. A.; Sweedler, J. V. J. Biol. Chem. 2016,291, 8109−8120.
(32) Sadeghi, M.; Vertes, A. Appl. Surf. Sci. 1998,127−129, 226−
234.
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1987
(33) Dekker, T. J. A.; Jones, E. A.; Corver, W. E.; van Zeijl, R. J. M.;
Deelder, A. M.; Tollenaar, R. A. E. M.; Mesker, W. E.; Morreau, H.;
McDonnell, L. A. Anal. Bioanal. Chem. 2015,407, 2167−2176.
(34) Heijs, B.; Holst, S.; Briaire-de Bruijn, I. H.; van Pelt, G. W.; de
Ru, A. H.; van Veelen, P. A.; Drake, R. R.; Mehta, A. S.; Mesker, W.
E.; Tollenaar, R. A.; Bove
e, J. V. M. G.; et al. Anal. Chem. 2016,88,
7745−7753.
(35) Gundel, M.; Penzlin, H. Cell Tissue Res. 1980,208, 283−297.
(36) Lococo, D. J.; Tobe, S. S. Int. J. Insect Morphol. Embryol. 1984,
13,65−76.
(37) Predel, R. J. Comp. Neurol. 2001,436, 363−375.
(38) Nassel, D. R.; Cantera, R.; Karlsson, A. J. Comp. Neurol. 1992,
322,45−67.
(39) Predel, R.; Neupert, S.; Derst, C.; Reinhardt, K.; Wegener, C. J.
Proteome Res. 2018,17, 440−454.
(40) Schmitt, F.; Vanselow, J. T.; Schlosser, A.; Kahnt, J.; Rossler,
W.; Wegener, C. J. Proteome Res. 2015,14, 1504−1514.
(41) Verleyen, P.; Baggerman, G.; Wiehart, U.; Schoeters, E.; Van
Lommel, A.; De Loof, A.; Schoofs, L. J. Neurochem. 2004,88, 311−
319.
(42) Duve, H.; Thorpe, A.; Tobe, S. S. Cell Tissue Res. 1991,263,
285−291.
(43) Veenstra, J. A.; Davis, N. T. Cell Tissue Res. 1993,274,57−64.
(44) Predel, R.; Eckert, M.; Polla
k, E.; Molna
r, L.; Scheibner, O.;
Neupert, S. J. Comp. Neurol. 2007,500, 498−512.
(45) Hankin, J. A.; Barkley, R. M.; Murphy, R. C. J. Am. Soc. Mass
Spectrom. 2007,18, 1646−1652.
(46) Yang, J.; Caprioli, R. M. Anal. Chem. 2011,83, 5728−5734.
(47) Kompauer, M.; Heiles, S.; Spengler, B. Nat. Methods 2017,14,
1156−1158.
Analytical Chemistry Article
DOI: 10.1021/acs.analchem.8b04304
Anal. Chem. 2019, 91, 1980−1988
1988
