Analysis of Peptide Stereochemistry in Single Cells by Capillary Electrophoresis—Trapped Ion Mobility Spectrometry Mass Spectrometry

Abstract

Single cell analysis strives to probe molecular heterogeneity in morphologically similar cell populations through quantitative or qualitative measurements of genetic, proteomic, or metabolic products. Here, we applied mass analysis of single neurons to investigate cell-cell signaling peptides. The multiplicity of endogenous cell-cell signaling peptides is a common source of chemical diversity among cell populations. Certain peptides can undergo post-translational isomerization of select residues, which has important physiological consequences. The limited number of single cell analysis techniques that are sensitive to peptide stereochemistry make it challenging to study isomerization at the individual cell level. We performed capillary electrophoresis (CE) with mass spectrometry (MS) detection to characterize the peptide content of single cells. Using complementary trapped ion mobility spectrometry (TIMS) separations, we measured the stereochemical configurations of three neuropeptide gene products derived from the pleurin precursor in individual neurons (N = 3) isolated from the central nervous system of Aplysia californica. An analysis of the resultant mobility profiles indicated >98% of the detectable pleurin-derived peptides exist as the non-isomerized, all-L forms in individual neuron cell bodies. However, we observed 44% of the Plrn2 peptide from the pleurin precursor was present as the isomerized, D-residue-containing form in the nerve tissue. These findings demonstrate an unusual distribution of isomerized peptides in A. californica and establish CE–TIMS MS as a powerful analytical tool for investigating peptide stereochemistry at the single cell level.

Graphical Abstract

INTRODUCTION

The inversion of amino acid residue stereochemistry can manifest substantial alterations to the normal physiochemical or biological properties of endogenous peptides.¹ Enzymatic L-to-D residue isomerization is an unusual post-translational modification (PTM) in the secretory pathway that contributes to the diversification of peptide structures and functions.^1,2 For decades, the products of L-to-D residue isomerization, D-amino acid containing peptides (DAACPs), have been isolated from numerous animal species;² however, because the enzymes that catalyze isomerization are not known, conventional molecular biology or biochemistry approaches cannot be used to measure DAACPs in individual cells. As a result, peptide isomerization remains an underexplored PTM.² Improved methods for measuring peptide diastereomers are essential for understanding the biosynthetic pathways involved in neuropeptide isomerization.

Differentiating and quantifying peptide diastereomers in single cells requires methods that provide both low detection limits and high stereoselectivity. Mass spectrometry (MS) excels at distinguishing the individual chemical components in complex mixtures and satisfies the low detection-limit requirements needed for single cell analysis. MS has been successfully used for single cell chemical profiling,^3,4 but because MS alone is not inherently stereospecific, it must be coupled with complementary techniques to analyze endogenous DAACPs. High sensitivity and stereoselective tandem MS (MS/MS)-based methods have been developed to distinguish peptide diastereomers,^5–8 but have only occasionally been useful for detecting DAACPs in bulk tissue extracts;^2,9–12 moreover, fragmentation-based techniques have seldom been applied for the relative quantification of DAACPs in individual cells.¹³ This is in part because differentiation and quantitation of peptide diastereomers by MS/MS is not straight forward, and relies on product ion ratios rather than a physical separation of diastereomers.^5,7,13

Solution-phase separation methods such as liquid chromatography (LC) have been used to distinguish native DAACPs in biological systems,^2,11 but LC requires sample dilution and clean-up steps that tend to be incompatible with handling single cells, even when coupled with MS detection. In contrast, capillary electrophoresis (CE) is convenient for small-volume analysis, and when used in combination with MS detection, is well suited for the biomolecular characterization of individual cells. CE can accomplish efficient molecular separations on small sample-volumes (pL–nL of sample),^14,15 which minimizes sample dilution steps and improves the ionization of low-abundance analytes for subsequent MS analysis. Thus, CE–MS has been widely applied for profiling small molecules, metabolites, proteins, and peptides in animal tissues, including at or near the single cell level.^16–24

Despite the advantages of CE–MS for single cell analysis, there is still a need for further improvements to existing sampling techniques and electrophoretic separations of peptide stereoisomers.^25,26 Specifically, CE separations of peptide stereoisomers may require specialized buffer conditions and chiral selectivity reagents that are incompatible with MS detection.^27–29 As a result, characterization of DAACPs by CE–MS remains challenging.

Coupling CE with ion mobility spectrometry (IMS) MS is a promising approach that can circumvent several of the challenges facing the analysis of DAACPs in single cells by CE–MS. Electrospray ionization (ESI)–IMS MS analysis has been applied to achieve lower detection limits and increase MS spectral feature detection compared to ESI–MS alone.^30–32 This is possible because IMS serves as an additional dimension of ion separation in the gas phase, allowing for the partitioning of isomeric/isobaric species based on molecular geometry. Furthermore, IMS MS can distinguish peptide diastereomers with subtle structural variations that might not easily separate by solution-phase separation approaches.^2,33–36

Conveniently, mass spectrometers capable of IMS measurements can be outfitted with commercial or custom CE devices for single cell measurements. While there are several commercially available IMS systems, the resolving powers achievable by trapped ion mobility spectrometry (TIMS) relative to the timescale of the mobility measurements makes TIMS well suited to coupling with CE separations.^37,38 Parallel accumulation TIMS enables acquisition cycle speeds and mobility measurements on the order of tens to hundreds of milliseconds, which are compatible with high-efficiency CE separations where analytes can migrate over a narrow migration time frame (on the order of seconds). Additionally, with TIMS, ions are preconcentrated (trapped) in an electric field and this enables increased MS signal-to-noise ratios and lower detection limits with minimal ion losses.³⁹ For these reasons, TIMS is rising in popularity for the analysis of small-volume biological samples.⁴⁰ Taken together, CE in combination with TIMS MS has potential for characterizing DAACPs at the single cell level.

We created a CE–TIMS MS workflow that allowed us to accurately assign the stereochemistry of selected endogenous neuropeptides in individual neurons (Scheme 1). More specifically, sampling techniques and CE–MS approaches adapted to work with individual cells were combined with TIMS and used to characterize isomerized peptides.^2,41 CE enabled us to separate the peptides from individual neurons with minimal sample dilution, and TIMS allowed us to accurately assign the native peptide stereochemistry. The combined approaches facilitated the detection and simultaneous stereochemical characterization of three known isomerized neuropeptides derived from the pleurin precursor (NP_001191654.1) in single neurons isolated from the Aplysia californica central nervous system (CNS).¹¹

Scheme 1.

The workflow used to measure neuropeptide stereochemistry in individual neurons.

The A. californica CNS is an excellent model for developing single cell methods because the locations of individual neurons that are known to express DAACP-encoding precursors have been identified by in-situ hybridization,^42,43 and these live neurons can be manually isolated for subsequent analysis.⁴⁴ In this study, we specifically targeted neurons expressing the pleurin precursor because it was recently discovered to encode three peptides that undergo post-translational L-to-D residue isomerization: Plrn1, Plrn2, and Plrn3 (Plrn1–Plrn3; sequences are listed in Table 1).¹¹

Table 1.

Native and synthetic pleurin precursor-derived peptide sequences and m/z values monitored.

Name	Sequence	Native [m+2H]2+	Synthetic [m*+2H]2+
l-Plrnl	MFYTKGSDSDYPRI-NH2	839.9	840.4
d2-Plrnl	MdFYTKGSDSDYPRI-NH2	839.9	840.4
d3-Plrnl	MFdYTKGSDSDYPRI-NH2	839.9	840.4
l-Plrn2	SFYTTGNGNHYPRI-NH2	813.4	813.9
d2-Plrn2	SdFYTTGNGNHYPRI-NH2	813.4	813.9
d3-Plrn2	SFdYTTGNGNHYPRI-NH2	813.4	814.4
l-Plrn3	GIFTOSAYGSYPRV-NH2	772.9	773.4
d2-Plrn3	GdIFTQSAYGSYPRV-NH2	772.9	773.4
d3-Plrn3	GIdFTQSAYGSYPRV-NH2	772.9	773.9

d-residues indicated in bold font (dX). [m+2H]²⁺ is the theoretical monoisotopic m/z of the native Plrnl, Plrn2 and Plrn3 peptides. [m*+2H]²⁺ is the theoretical monoisotopic m/z of the isotope labeled synthetic ¹³C-Plrn1-Plm3 peptides. Positions of ¹³C-Gly residues in synthetic standards are underlined in the sequence (G).

Based on our prior work,¹¹ we know that in the A. californica CNS, Plrn1 exists as two possible stereoisomers, L-Plrn1 and D2-Plrn1 (D2 indicates a D-residue at the second residue from the N-terminus); whereas Plrn2 and Plrn3 each exist as three possible stereoisomers, L-Plrn2/Plrn3, D2-Plrn2/Plrn3, or D3-Plrn2/Plrn3 (D3 indicates a D-residue at the third residue from the N-terminus) (Table 1).

Using CE–TIMS MS, we sought to determine the percentage of pleurin-derived peptides that are isomerized in individual neurons. To accomplish this, live pleurin-expressing neurons were manually isolated from the A. californica CNS according to an in-situ hybridization staining map.⁴³ Individual cell extracts were analyzed by CE–TIMS MS and the mobility profiles of the native Plrn1–Plrn3 ions were compared to the mobility profiles of synthetic ¹³C-labeled Plrn1–Plrn3 standards to assess the native peptide stereochemistry. Further, we used TIMS mobility profiles to quantify the relative abundance of Plrn2 diastereomers in both neuron cell bodies, and in the nerves connecting the central ganglia (referred to as the ganglia connectives). Our findings reveal that the D-residue-containing forms of pleurin-derived peptides are not homogenously distributed between these two locations of the CNS.

EXPERIMENTAL SECTION

Reagents and Standards.

All solvents and reagents, LC–MS grade or higher, used for the CE– TIMS MS measurements were purchased from Thermo Fisher Scientific (Waltham, MA). Unless otherwise specified, reagents used for A. californica dissections and neuron isolations were >98% purity and purchased from MilliporeSigma (St. Louis, MO) or Thermo Fisher Scientific. Synthetic peptide standards of ¹³C-labeled Plrn1, Plrn2, and Plrn3 were previously synthesized and characterized,¹¹ with the sequences provided in Table 1.

Animals and Single Cell Isolation.

A. californica (100–200g) were obtained from the National Institutes of Health/University of Miami National Resource for Aplysia (Miami, FL). Animals were kept in an aquarium containing aerated and filtered Instant Ocean seawater (Aquarium Systems Inc., Mentor, OH) at 14°C until use. Prior to dissection, the animals were weighed and anesthetized by injecting 50% (volume/body weight) of 333 mM MgCl2 into the body cavity. To aid the isolation of individual neurons, CNS tissues were incubated for 45–60 min at 34°C in a solution containing 1% neutral protease, Dispase (100 units/mL, Worthington Biochemical Corp., Lakewood, NJ), and antibiotic cocktail (penicillin G, 100 μg/mL streptomycin, and 100 μg/mL gentamicin) in artificial sea water (ASW: 460 mM NaCl, 10mM KCl, 11 mM CaCl₂, 55 mM MgCl₂, 5 mM NaHCO₃, pH 7.8). Following incubation, CNS tissues were rinsed in ASW/antibiotic cocktail for 20 min at 4°C, and the ganglionic sheaths surgically removed. Individual neurons from the pleural ganglia were manually isolated with electrolytically sharpened tungsten needles using pleurin in-situ staining images as a guide.^42,43 For sampling ganglia connectives, approximately 50–100 μm-long portions were cut using microdissection scissors. Scheme 2 depicts the sampling of pleurin neurons and ganglia connectives. The isolated neurons and ganglia connectives were transferred individually onto the inner side of a 0.25 mL PCR tube filled with 5 μL of acidified methanol (90:9:1 methanol: water: glacial acetic acid) using lab-made plastic micropipettes (see Supporting Information, Experimental section for fabrication details) filled with deionized water. Excess liquid was aspirated and neurons were forced down the tube wall into the bottom of the tube by flowing extraction media pipetted from the same tube. Sample tubes with and without cells were kept on ice for the length of the isolation procedure. Sample tubes were stored at –20°C and analyzed by CE–TIMS MS within 24 h. After all individual neurons and ganglia connective samples were collected, 0.3 μL of the extracts were sampled for peptide validation by an orthogonal technique, matrix-assisted laser desorption/ ionization (MALDI)-time-of-flight (TOF) MS, and the remaining extract was used for CE–TIMS MS analysis.

Scheme 2.

Right pleural ganglion (not drawn to scale) illustrating the sampling of an individual neuron and cerebro-pleural ganglia connective. An isolated pleurin neuron is shown removed from the cell body. A pleurin neuronal process is depicted in red, extending through the cerebro-pleural connective

MALDI-TOF MS Analysis of Single Neurons.

Individual pleurin-expressing neurons were isolated and analyzed by MALDI-TOF MS using methods described previously,^43,44 with minor alterations. Briefly, each neuron was isolated from the A. californica pleural ganglia as described above, but instead of depositing the neuron into an extract solution, it was deposited directly onto a MALDI target plate (Bruker Corp., Billerica, MA) and excess liquid removed. The deposited neuron was then mixed with ~1 μL of 50 mg/mL 2,5-dihydroxybenzoic acid in 50% CH₃CN with 0.1% trifluoroacetic acid and allowed to air dry. Peptide profiles were measured using an ultrafleXtreme MALDI-TOF mass spectrometer (Bruker Corp.) in positive reflection mode (see Supporting Information, Experimental Section for more details).

CE Separations

For CE, the sheath liquid was 50% methanol in water solution with 0.1% formic acid (FA). The sheath liquid was supplied through the emitter at a 750 nL/min flow rate. The CE separation was performed on a fused silica capillary, 75–80 cm, 40 μm inner diameter and 105 μm outer diameter (CM Scientific, Silsden, UK), with a separation voltage of +20 kV. The background solution was 0.75% FA and 25% CH3CN in water solution.⁴⁵ Field-amplified sample injection was performed by using electrokinetic injection of the sample solution by +20 kV for 30 s. A lab-built CE–ESI source^19,46 was coupled to a timsTOF Pro mass spectrometer (Bruker Corp.). The emitter was grounded, and the ion guide capillary voltage of the mass spectrometer was set at 2000 V to establish the cone-jet spray. The dry nitrogen gas was set at 180°C with a flow rate of 3 L/min. The single cell extracts (~5 μL) were dried using vacuum centrifugation and reconstituted in 2 μL of acidified methanol; 1 μL of extract was deposited into lab-made stainless steel nano-vials^46,47 and injected into the capillary by electrokinetic injection, as described above. At the end of each set of single cell measurements, 1 μL solutions of 10 nM ¹³C-Plrn1–Plrn3 were electrokinetically injected into the capillary to provide a comparison of the 1/k0 values of the synthetic standards to the native peptides.

TIMS MS Parameters for Single Neurons and Ganglia Connectives.

The timsTOF Pro calibration was performed using a low-concentration ESI tune mix (P/N G1969–85010, Agilent Technologies, Santa Clara, CA,) as described previously.^39,48 Nitrogen was the buffer gas for all measurements. Following the instrument calibration, the standard ESI source was replaced by a custom, lab-built CE–ESI source and the acquisition was adjusted to the optimized TIMS MS method. Figures S1 and S2 show the results from the TIMS parameter optimization, and the optimization procedures are described in the Supporting Information, Experimental Section. For data acquisition on single cells, the optimized TIMS MS method was operated with the PASEF mode off. Imex settings were set to custom resolution, set 1/k₀ End for accumulation to 2.00 V·s·cm⁻², ramp time was 350 ms and lock duty cycle to 100% was selected. The mobility range was optimized by narrowing the mobility range to 0.9 < 1/k₀ < 1.2 V·s·cm⁻². For single cells, the mass spectrometer was operated in full scan mode with an m/z range of 100– 1700. For TIMS MS measurements on ganglia connectives, the Plrn1–Plrn3 peptide [m+2H]²⁺ ions were not detected using the full m/z scan mode that was used for single cells. We therefore analyzed the ganglia connectives using multiple reaction monitoring (MRM) mode, specifically focusing on detecting the Plrn2 peptide, and this resulted in improved signal. For MRM measurements of Plrn2, an m/z window of 813.4 ± 0.2 was monitored, corresponding to the [m+2H]²⁺ of native Plrn2. Additional relevant instrument operation parameters are provided in the Supporting Information.

RESULTS AND DISCUSSION

Characterizing Plrn1—Plrn3 Standards.

Before analysis of the stereochemistry of Plrn1–Plrn3 in single cells, the synthetic ¹³C-labeled Plrn1–Plrn3 standards were characterized by CE and TIMS MS to determine the effectiveness of our experimental conditions for separating the diastereomers. Injections of individual and combined ¹³C-Plrn1–Plrn3 diastereomers showed Plrn1, Plrn2, and Plrn3 sequences migrate differently,but the L-,D2-, and D3-forms of each peptide co-migrate (FigureS3). ¹³C-Plrn1–Plrn3 had long migration times (20–30 min), which may have contributed to peak broadening and reduced the CE separation efficiency of the diastereomers. Overall, the CE conditions used in this study were suitable for separating peptides, but did not allow for the separation of ¹³C-Plrn1–Plrn3 diastereomers. The subsequent TIMS separations compensated for the sub-optimal CE separations of stereoisomers.

In contrast to CE, TIMS allowed us to better distinguish between ¹³C-Plrn1–Plrn3 diastereomers. Extracted ion mobilograms (EIMs) obtained from individual injections of the synthetic ¹³C-Plrn1–Plrn3 standards are shown in Figure 1. For TIMS separations, we achieved peak resolving powers ranging from 106 to 162 for ¹³C-Plrn1 and ¹³C-Plrn2 diastereomers, and 37 to 142 for ¹³C-Plrn3 diastereomers (Table S1). The greatest resolution was achieved for ¹³C-Plrn2 diastereomers; Δcollisional cross section (CCS) values as low as 5 Ų could be differentiable (Figure 1B).⁴⁹ The structural changes to a peptide resulting from the presence of a D-residue may differ depending on the sequence or length of the peptide, and the position and structure of the D-residue⁵⁰ In the cases of ¹³C-Plrn1–Plrn3, each diastereomer had a unique mobility fingerprint with varying degrees of complexity. The presence of multiple peaks in the EIMs suggests that these peptides can adopt more than one stable gas phase conformation during TIMS,^38,51 but exploring the exact physical explanation for why these features arose was beyond the scope of this work. The important takeaway is that we expected the native Plrn1–Plrn3 EIMs to exhibit similar features to the synthetic standards.

Figure 1.

Differentiating synthetic neuropeptide stereoisomers by TIMS. Stacked EIMs of the [m*+2H]²⁺ ions of pure (A) ¹³C-L-Plrn1, ¹³C-D2-Plrn1 and ¹³C-D3-Plrn1; (B) EIMs of ¹³C-L-Plrn2, ¹³C-D2-Plrn2 and ¹³C-D3-Plrn2; (C) EIMs of ¹³C-L-Plrn3, ¹³C-D2-Plrn3 and ¹³C-D3-Plrn3. The CCS values (in Ų) are displayed above each peak. Individual electropherograms corresponding to each peptide are shown in Figure S3.

MALDI-TOF and CE—TIMS MS of Individual Neurons.

Locating and isolating the target neurons for analysis was challenging. For cell isolation, pleurin neurons were located in live ganglia by referencing previously published images of pleurin neurons in preserved ganglia.^41,42 The available reference images show that the expression of the pleurin gene transcript is localized to clusters of ~10 neurons in the pleural ganglia.^42,43 But in the live CNS, there were no morphological features unique to pleurin neurons that distinguished them from the hundreds of surrounding neurons. Manual manipulation ensured precision isolation of solitary cells.

Precise and careful manual cell isolation was also important for reducing sample complexity and mitigating sample loss. The neurons were isolated from live ganglia in ASW, so it was important to minimize the transfer of excess ASW to avoid potential complications with electrophoretic separations and the electrospray process. Live dissected neurons were manipulated using a micropipette filled with deionized water. When performed quickly, this provided a means to wash away salts while preserving cellular integrity. Careful manual manipulation allowed us to isolate neurons without the need for any subsequent desalting procedures, thereby reducing sample handling and associated sample losses.

We initially characterized pleurin neurons by MALDI-TOF MS to determine the peptide fingerprint expected from a typical pleurin neuron. MALDI-TOF MS spectra showed that Plrn1–Plrn3 [m+H]⁺ ions were among the most-abundant ions detected in the cells, but peptides derived from the urotensin II precursor (ApUII),⁵² Aplysia mytilus inhibitory peptide-related peptides (AMRPS) precursor,⁵³ and thymosin beta (TMSB),⁵⁴ were also detected (Figure 2A). Full sequences and observed m/z values of peptides colocalizing with Plrn1–Plrn3 in neurons are reported in Table S2. These results establish that the peptide fingerprint of pleurin neurons is largely consistent between animals, and from cell-to-cell (N ≥ 10 neurons from N ≥ 2 animals).

Figure 2.

Single cell analysis of pleurin neurons by MALDI-TOF MS and CE–TIMS MS. (A) Representative MALDI-TOF MS spectrum of an individual pleurin neuron and native peptide identifications. Inset in panel A is the zoomed view of peaks between m/z 3500–5500. (B) Representative CE–TIMS MS base peak electropherogram (m/z 550–1200, mobility range 0.9–1.2 V·s·cm⁻²) obtained from a different single pleurin neuron. Peaks matching to peptides also identified by MALDI MS are numbered according to Panel A. Unlabeled peaks correspond to unknown ions. (C–E) Assignment of native Plrn1–Plrn3 stereochemistry in the pleurin neuron by TIMS. Plots show overlaid EIMs of native (C) Plrn1, (D) Plrn2, and (E) Plrn3 with the corresponding synthetic ¹³C-labeled all-L standard. Note that the ¹³C-L-Plrn1–Plrn3 EIMs were replotted from Figure 1 for comparison purposes. (F–H) EIMs of 1:1:1 mixtures of synthetic ¹³C-labeled L-, D2-, D3- (F) Plrn1, (G) Plrn2 and (H) Plrn3. EIMs in panels C–H are of the [m+2H]²⁺ or [m*+2H]²⁺ ions. The vertical bar in (H) serves to highlight the differences between the mobility fingerprints of the pure L-Plrn1 in panel C, and the combined L-, D2-, and D3-Plrn1.

For CE-TIMS MS, isolated neurons were analyzed using the same CE conditions and TIMS MS parameters that were used for the synthetic peptides. Although the TIMS parameters were tuned for the optimal resolution of Plrn1–Plrn3 [m+2H]²⁺ ions, the simultaneous detection of other colocalizing peptides was still possible. This allowed us to directly compare peptide profiles observed by CE-TIMS MS and MALDI-TOF MS. A representative base peak electropherogram from an individual pleurin-containing neuron is shown in Figure 2B. Including Plrn1–Plrn3, a total of 8 peptides were identified that matched peptides also detected by MALDI-TOF MS (Figure 2A, B; Table S2). In contrast to MALDI TOF MS, however, Plrn1–Plrn3 were not the most-abundant ions detected in neurons by CE-TIMS MS, but this can be explained by differences in the ionization techniques (MALDI vs. ESI).

We isolated a total of N = 7 neurons for CE-TIMS MS analysis, and N = 3 showed the characteristic peptide profile expected for a pleurin neuron. The detection of multiple expected colocalizing neuropeptides helped validate that the individual cells were isolated from the intended cell population. But there was a trade-off between the number of peptides that could be simultaneously analyzed by CE-TIMS MS vs. the TIMS resolution. Ions with mobilities outside the set mobility range of our experiment (0.9 < 1/k₀ < 1.2 V·s·cm⁻²) were filtered out by the TIMS analyzer, and thus, did not appear in the final mass spectrum. Expanding the mobility range of our measurement may have resulted in the detection of more peptides, but at the expense of TIMS resolution.

Assignment of Native Plrn1—Plrn3 Stereochemistry.

To determine the stereochemistry of Plrn1– Plrn3 in individual neurons, the mobility profiles of synthetic and native Plrn1–Plrn3 peptides were directly compared; all showed multiple isomeric peaks and closely resembled those obtained from the pure synthetic ¹³C-L-Plrn1–Plrn3 standards (Figure 2C–E). An overlay of the intensity-normalized EIMs of the synthetic ¹³C-L-Plrn1–Plrn3 and native peptides showed significant congruency between the traces (Figure 2C–E). The native Plrn1–Plrn3 mobility profiles did not resemble those expected for D2- or D3-Plrn1–Plrn3 (Figure 1A–C), or mixtures of synthetic diastereomers (Figure 2F–H). Further, even though we observed the L-, D2-, and D3-¹³C-Plr1—plrn3 stereoisomers were not fully resolved by TIMS, the EIMs of the synthetic diastereomer mixtures were still distinct from the pure all-L peptides (Figure 2F–H). The slight variations between the mobility fingerprints of the pure synthetic peptides and mixtures of synthetic diastereomers were essential for assigning the stereochemistry of the native neuropeptides. We emphasize the importance of our initial analysis of the synthetic ¹³C-Plrn1–Plrn3 standards for assigning the native neuropeptide stereochemistry. The EIMs of the native Plrn1–Plrn3 showed multiple isomeric peaks, which could have easily been misinterpreted as resulting from different native peptide diastereomers. By analyzing the synthetic peptides, we established the isomeric peaks in the Plrn1–Plrn3 EIMs are likely gas phase conformers.

Based on the significant agreement between the mobility profiles of synthetic 13C-L-Plrn1—Plrn3 and the native peptides (N = 3 sampled neurons), we concluded that negligible quantities of D2-orD3-Plrn1–Plrn3 contributed to the native peptide EIMs. In a separate experiment, we determined a minimum percentage of isomerized peptide that can be detected from an EIM by spiking solutions of 100 fmol/μL synthetic 13C-L-Plrn2 with synthetic ¹³C-D2-Plrn2 at concentrations ranging from 0–25 fmol/μL (see 13C-Supporting Information, Experimental Section). We were able to detect < 2% of ¹³C-D2-Plrn2 for an amount of 120 fmol ¹³C-L-Plrn2 injected(Figure S4). Similar limits were expected for the ¹³C-D3-Plrn2, as well as the ¹³C-Plrn1 and ¹³C-Plrn3 diastereomers. We concluded that > 98 %of the Plrn1–Plrn3 detected in pleurin neurons exist in the all-L configurations.

We had initially predicted the all-L forms of pleurin-derived peptides would colocalize with the D2 and/or D3 forms in individual neurons, but our stereochemical analysis did not support our initial expectation. This expectation was based on our previous work, where we reported that more than 50% of Plrn1–Plrn3 exist as either the D2 or D3 forms in whole ganglia homogenates.¹¹ But neural ganglia are complex anatomical structures having a conglomerate of cell types encircling a large, synaptically dense region called the neuropil. Thus, the chemical complexity of an entire ganglion homogenate is substantially greater than an individual neuron. Consequently, the relative abundance of DAACPs in whole ganglia homogenates is expected to differ compared to single cells. For example, the relative abundance of another known DAACP in A. californica, NdWF-NH₂, has been previously quantified in both single cells and bulk tissue extracts. It was found that the relative abundance of NdWF-NH₂ in whole ganglia homogenates was between 60–80%,⁵⁵ whereas in single cells it was closer to 90%.^13,27

Our finding that D2- or D3-Plrn1–Plrn3 were not detectable in individual pleurin neurons is distinct from previous observations indicating significant conversion of NWF-NH₂ to NdWF-NH₂ in individual neurons from A. californica.^13,27 We considered whether experimental factors such as ion suppression, matrix effects, electrokinetic injection bias, or selective adsorption of DAACPs to the surfaces of the capillary, tubes, pipette tips etc., prevented the detection of isomerized Plrn1–Plrn3 in single cell samples. However, our parallel analysis on synthetic ¹³C-Plrn1–Plrn3 diastereomers excluded these factors.

A possible explanation for the absence of isomerized Plrn1–Plrn3 from individual pleurin neurons is that the isomerized peptides localize in the neuronal processes (axons) rather than the cell bodies. Immunohistochemical staining with stereospecific antibodies has shown the subcellular distribution of DAACPs in neurons from the crayfish species, Orconectes limosus; these studies revealed the isomerization of the crustacean hyperglycemic hormone (CHH) to D3-CHH increased distally from cell bodies to the axon terminal at the release site.⁵⁶ These observations have shaped the current understanding of L-to-D residue isomerization as a late stage PTM that occurs gradually during anterograde neuropeptide transport.^1,2 In A. californica, neuropeptides can be transported several centimeters from cell bodies along cellular processes to the sites of release, at rates of 100 mm/day.^57,58 Because L-to-D residue isomerization has previously been shown to occur gradually during neuropeptide transport in crustaceans, it is plausible that the isomerization of pleurin-derived peptides in A. californica occurs predominantly along cellular processes. When neurons are isolated from A californica, the fragile neuronal processes detach from the cell body and remain in the ganglion. This is depicted in Scheme 2, which shows an isolated pleurin neuron and its neuronal process (in red) remaining in the ganglion. By isolating the cell bodies without the cellular processes, we sampled and detected newly synthesized all-L peptides as opposed to the slower forming isomerized products, D2-and/or D3-Plrn1–Plrn3.

CE—TIMS MS of Ganglia Connectives.

Our anatomical knowledge of pleurin neurons in the A. californica CNS allowed us to design an experiment to measure the Plrn1–Plrn3 stereochemistry in cellular processes. It has been shown that the cellular processes from pleurin neurons located in the pleural ganglia project to adjacent ganglia (either cerebral, abdominal, or pedal) via ganglia connectives.⁵⁹Scheme 2 illustrates a neuronal process extending to a neural connective. These ganglia connectives serve as “highways” for delivering secretory vesicles containing neuropeptides to their release sites.⁵⁸ Theorizing that the isomerization of pleurin-derived peptides occurs in the cellular processes, we expected the isomerized pleurin-derived peptides to accumulate in the ganglia connectives. To validate our hypothesis, we used CE–TIMS MS to characterize the stereochemistry of pleurin-derived peptides in ganglia connectives. Cerebro-pleural connectives were chosen for CE–TIMS MS analysis because sample screening by MALDI-TOF MS indicated that the pleurin-derived peptides were among the most abundant ions in these samples (Figure S5). However, while Plrn1, Plrn2 and Plrn3 were all detected in the cerebro-pleural connectives by MALDI-TOF MS, only the stereochemistry of Plrn2 was successfully measured by CE–TIMS MS in the samples (see the Experimental Section above: TIMS MS Parameters for Single Neurons and Ganglia Connectives).

EIMs obtained from synthetic ¹³C-L-Plrn2 and ¹³C-D2-Plrn2 standards (Figure 3A) were compared to an EIM obtained for native Plrn2 in the cerebro-pleural connective (Figure 3B). As expected, native D2-Plrn2 was colocalized with all-L Plrn2 in extracts from the cerebro-pleural connectives (N = 3 sampled sections), indicated by the presence of an additional peak at 1/k₀ = 1.06 V·s·cm⁻² that was not observed in the EIMs obtained from single neuron samples (Figure 3). No peaks were detected at the 1/k₀ values expected for D3-Plrn2 in the EIM from the cerebro-pleural connectives (N = 3 sampled sections), but this was not surprising because in our previous work,¹¹ D3-Plrn2 was only detected in buccal ganglia extracts but not in either pleural or cerebral ganglia extracts. Using the ¹³C-L-Plrn2 and ¹³C-D2-Plrn2 TIMS measurements shown in Figure 3A, we estimated the ratio of D2-Plrn2 to L-Plrn2 in the connectives to be approximately 0.8:1. (for details on this calculation, see the Supporting Information, Experimental Section). This corresponds to 44% D2-Plrn2 and 66% L-Plrn2 in the ganglia connective. The presence of D2-Plrn2 in the cerebro-pleural connective supports our revised hypothesis that isomerized forms of pleurin peptides accumulate in cellular projections rather than cell bodies, and this in agreement with published findings from the crayfish.⁵⁶

Figure 3.

EIMs obtained by CE-MRM-TIMS MS showing the localization of D2-Plrn2 to the cerebro-pleural connective. (A) EIMs of synthetic ¹³C-L-Plrn2 (solidtrace) and ¹³C-D2-Plrn2 [m*+2H]²⁺ ions (dashed trace). (B) EIM of native Plrn2 [m+2H]²⁺ detected in the cerebro-pleural ganglia connective.

Our results show that the L-Plrn1–Plrn3 peptides in the A. californica CNS can be spatially separated from their DAACP counterparts. The localization of isomerized Plrn1–Plrn3 within the central nerves indicates that in our previous work,¹¹ a majority of the D2-and/or D3-Plrn1–Plrn3 we detected in ganglia homogenates may have been from the ganglia connectives and/or neuropil rather than the neuron cell bodies. This biological insight demonstrates the significant advantages of performing measurements at the single cell level as opposed to bulk tissue.

CONCLUSIONS

This work has established CE–TIMS MS as an effective technique for the analysis of peptides and peptide diastereomers at the single cell level. The combination of CE and TIMS MS enabled the stereochemical characterization and relative quantitation of peptide diastereomers in single cells and nerve tissue. Direct comparisons of the mobility profiles of synthetic and endogenous peptides allowed us to unambiguously assign native neuropeptide stereochemistry. Our stereochemical analysis of Plrn1–Plrn3 in individual neurons indicated L-Plrn1–Plrn3, but not D2- and/or D3-Plrn1–Plrn3, are localized in the cell bodies. In contrast, D2-Plrn2 comprised 44% of the Plrn2 found within the cerebro-pleural ganglia connectives. These results are consistent with DAACP localization patterns observed in crustacean neurons. Future CE–TIMS MS studies will focus on applying the protocols reported here to investigate the effects of external chemical stimuli on the levels of DAACPs in individual pleurin neurons in cell culture. CE–TIMS MS may also be useful for measuring other known DAACPs in A. californica or identifying novel, low-abundance DAACPs present only in rare cell-types. The extension of these CE–TIMS MS protocols to studying DAACPs in single cells from a variety of species could shed further light on the molecular mechanisms underlying L-to-D residue isomerization.