Microanalysis of Brain Angiotensin Peptides Using Ultrasensitive Capillary Electrophoresis Trapped Ion Mobility Mass Spectrometry

Abstract

While the role of the renin-angiotensin system (RAS) in the peripheral circulation is well-characterized, we still lack in-depth understanding of its role within the brain. This knowledge gap is sustained by lacking technologies for trace-level angiotensin detection throughout tissues, such as the brain. To provide a bridging solution, we enhanced capillary electrophoresis (CE) nano-flow electrospray ionization (ESI) with large-volume sample stacking and employed trapped ion mobility time-of-flight (timsTOF) tandem HRMS detection. A dynamic pH junction helped stack approximately 10-times more of the sample than optimal using the field-amplified reference. In conjunction, the efficiency of ion generation was maximized by a cone-jet nanospray on a low sheath-flow tapered-tip nano-electrospray emitter. The platform provided additional peptide-dependent information, the collision cross section, to filter chemical noise and improve sequence identification and detection limits. The lower limit of detection reached sub-picomolar, or ~30 zmol (~18,000 copies) level. All nine targeted angiotensin peptides in mouse tissue samples were detectable and quantifiable from the paraventricular nucleus (PVN) of the hypothalamus, even after removal of circulatory blood components (perfusion). We anticipate CE-nanoESI with timsTOF HRMS to be broadly applicable for the ultrasensitive detection of brain peptidomes in pursuit of better understanding of the brain.

Graphical Abstract

INTRODUCTION

The renin-angiotensin system (RAS) is a widely studied circulatory peptide system essential to cardiovascular and thirst homeostasis and involved in cardiovascular health and disease progression.¹ As illustrated in Figure 1A, the canonical components of the RAS peptide system are made up of the pro-hormone precursor protein angiotensinogen (AGT), which is cleaved by a series of enzymes to 8 different angiotensin (Ang) peptides.^2-3 While the cardiovascular and thirst mechanisms of this peptide system continue to be well studied and at the focus of active research,^4-6 there remains a large gap in technologies for investigating Ang peptide synthesis, detection, and function, specifically within the brain.⁷

Figure 1.

Our approach for ultrasensitive detection of the renin-angiotensin system (RAS) in the brain. (A) Angiotensin (Ang) peptides and protein players of the pathway. Red marks those of interest for this study. Key: ACE = angiotensin-converting enzyme; ACE2 = ACE homologue; AP-A = aminopeptidase A; AP-N = aminopeptidase N; Carb-P = carboxypeptidase; PO = prolylendopeptidase. (B) Workflow extracting the peptides from the paraventricular nucleus (PVN) of the hypothalamus after removal of circulatory blood components (perfusion). (C) Main analytical steps integrating LVSS by a dynamic pH junction, efficient ionization by stable cone-jet CE nano-flow electrospray ionization (nanoESI), and trapped ion mobility time-of-flight (timsTOF) tandem high-resolution mass spectrometry (HRMS) for detection. Supplementary Movie 1 (S1) shows the stable Taylor cone (Tc) electrospraying droplets into the orifice of the mass spectrometer (OR). Scale bar, 50 μm.

Although it is well known that all the components of the RAS, required for the de novo synthesis of active Ang peptides, are present in the brain, quantification and analysis of these peptides has been analytically challenging. The brain RAS is well studied, primarily at its G-protein coupled receptor (GPCR) level, in brain regions critical for cardiovascular, autonomic, and thirst homeostasis, such as the paraventricular nucleus (PVN).^8-11 The brain RAS, however, was found to extend beyond this role (i.e., classic thirst, cardio-autonomic homeostasis), as there is evidence for its function in regulating psychological and emotional stress, learning, memory, and cognition,^12-13 all critical aspects for brain health and homeostasis.^{3, 14-16} Further understanding the complete dynamics of the brain RAS, including a potential interplay between local synthesis in distinct brain nuclei and uptake from circulation, is currently equivocal and limited by available technologies for Ang peptide detection and quantification.^{8, 14} Development of new technologies enabling trace-level Ang detection hold a key to critical information for further understanding the role of the RAS in brain health and disease.

Many analytical challenges hinder molecular characterization of the brain RAS. Sample collection must tailor to limited sample amounts in brain tissue nuclei, complex brain physiology at the blood-brain barrier, and complex molecular composition affecting peptide detection. The Ang peptides are produced at widely different concentrations in circulation (~1–100 pM in plasma) and tissue (~5–250 fmol/g).⁷ Conventional assays for the RAS, such as radioimmunoassays, have sufficient sensitivity but lack the specificity to distinguish the different peptides of the RAS, all of which possess highly similar sequences.^17-20 High-resolution mass spectrometry (HRMS) addresses this limitation by integrating high chemical and structural specificity with ultrahigh detection sensitivity. HRMS detection is rapid, can be made quantitative, and does not require the use of functional probes or chemical labeling. Recent developments in HRMS achieved sub-attomole lower limits of detection (LLOD) for peptides from limited sample amounts.^21-22 Chemical and spectral interferences were, however, found to challenge peptide identification and quantification, especially on low-abundance signals that are common to trace-level analyses.^23-24

Gas-phase separation via ion mobility HRMS offers remedy, without risking loss or dilution to the sample.²⁵ As the collision cross section (CCS) is an intrinsic property of peptides and CCS can be derived from measured ion mobility, ion mobility HRMS enhances the fidelity of identifications with another dimension of compound-dependent property, the CCS. This information can aid molecular identifications directly from cells and tissues using technologies that exchange traditional liquid-phase separation for throughput, such as direct-infusion ESI^26-27, MALDI imaging²⁸, and ambient ionization HRMS²⁹. When hyphenated to high-performance liquid chromatography, ion mobility substantially enhanced the detectable portion of proteomes by enhancing detection of peptides resulting from bottom-up proteomics.³⁰

Quantification of Ang peptides traditionally benefited from separation prior to HRMS detection. NanoLC HRMS was used to approximate endogenous Ang levels in plasma and bulk tissue,^31-33 but detection limits were challenged for detection in limited tissue nuclei, such as the PVN. Recently, capillary electrophoresis (CE) emerged as alternative for trace detection. Sample amounts analyzed by CE are orders of magnitude smaller than nanoLC, usually ~1–10 nL vs. 100 nL–1 μL, respectively. In-capillary preconcentration methods, such as field-amplified sample stacking (FASS), boosts sensitivity by increasing the amount of sample analyzed.³⁴ High theoretical plates tighten peptide zones into ion signals of high signal-to-noise ratio (S/N). A low electroosmotic flow naturally adapts to nano-flow ESI for efficient ion generation.

We and others recently advanced CE-ESI to trace levels of Ang peptides. Microanalytical platforms were built for CE to enable handling nanoliters of samples. Low sheath-flow CE-ESI interfaces capable of subattomole-to-zeptomole detection were hyphenated to quadrupole Orbitrap³⁵, time-of-flight (TOF)^35-36, and quadrupole-orbitrap-ion trap tribrid^37-38 mass spectrometers for detection. The technology was able to quantify the RAS peptides in microscale biopsies (punches) from (control) PVN, in ca. 10-times better LLOD than nanoLC.³⁷ The limits of detection, however, were still insufficient for the detection of Ang peptides in perfused systems,^39-40 likely because infusion with phosphate-buffered saline (PBS) removed the circulatory peptides.

Here, we extended CE-HRMS to be able to detect RAS peptides at tissue-endogenous amounts. We proposed to improve detection by addressing current limitations in sampling, separation, ionization, and HRMS-based detection (Figure 1B). Figure 1C illustrates the integration of strategically chosen technologies to this end. (1) To increase detection sensitivity, we proposed to adapt large-volume sample stacking and a custom-designed ultrasensitive tapered-tip CE-ESI interface to a commercial CE instrument. (2) To enhance the specificity of molecular identifications, we aimed to employ trapped ion mobility time-of-flight (timsTOF) tandem HRMS for detection. (3) We sought to establish proof-of-principle by enabling the detection and quantification of the Ang peptides in perfused PVN biopsies. Knowledge of the identity and tissue-bound amount of the peptides in finite tissue regions, such as the PVN, would not only help to better understand the brain RAS physiology but also may generate new understanding for Ang peptidomes across distinct neural circuits as well as interactions with other neuropeptide modulators (Fig. 1A).

METHODS

Materials.

Fused silica capillaries (20/90 μm inner/outer diameter and 75/360 μm inner/outer diameter) were purchased from Polymicro Technologies (Phoenix, AZ). The tapered-tip metal emitter (100/100 μm inner/outer diameter at taper, 300 μm outer diameter at distal end) was produced by New Objective (Littleton, MA). ZipTip kits for desalting (10 μL C18) were obtained from MilliporeSigma (Burlington, MA).

Solvents and Standards.

All solvents used were LC-MS grade from Fisher Scientific (Hampton, NH). Peptide standards of AngT, Ang 1-9, Ang I, Ang II, and Ang III, and Ang IV were from MilliporeSigma (Burlington, MA). The peptide extraction solution consisted of 75% (v/v) acetonitrile (ACN) with 1% (v/v) formic acid (FA) solution. For C18 ZipTip desalting, samples were dissolved in 0.1% (v/v) FA and eluted in the elution solution consisting of 50% (v/v) ACN with 0.1% FA. For CE, the background electrolyte (BGE) consisted of 50% (v/v) ACN with 4% (v/v) FA (resulting pH 2.3). The peptide standards and extracts were dissolved in 50 mM ammonium bicarbonate (AmBic) (resulting pH 8.2). The CE-ESI sheath solution was 50% (v/v) methanol (MeOH) with 0.1% (v/v) FA.

Animal Care and Handling.

All protocols regarding the humane treatment of animals were approved by the Institutional Care and Use Committee of The George Washington University (IACUC no. A2021-009). Adult (3–4 months old) male C57BL/6J mice were purchased from Jackson Laboratory (Bar Harbor, ME) and were housed in temperature- and humidity-controlled polyethylene cages on a 12-h light/dark cycle. Leading up to the experiment, water and food were supplied to all animals ad libitum.^41-42

Tissue Preparation.

The mice (n = 3/group) were anaesthetized with urethane and perfused trans-cardially with PBS to remove the blood and its components. The whole brains were then dissected, snap frozen, and stored at −80 °C. Using a Cryostat (CyroStar NX50), tissue punches were collected from the PVN (~3.1 mm³) at the following coordinates specified by the Mouse Brain Atlas⁴³: −0.7 mm caudal, ±0.25 mm lateral to bregma, and 4.7 mm below the skull surface.

Sample Preparation.

The peptides were extracted from individual tissue punches in 5 μL of peptide extraction solution, facilitated by homogenization in a bath sonicator with periodic vortex mixing. Cell debris was pelleted by centrifugation at 12,000 × g for 2 min. The resulting extracts from each of 3 tissue punches were then pooled, vacuum-dried, and reconstituted in 0.1% FA. These extracts were desalted on ZipTips, eluted in the elution solution, vacuum-dried, and reconstituted in 8 μL of 50 mM AmBic.

CE-ESI-timsTOF HRMS.

The peptides were separated in a 130 cm long fused silica capillary (20/90 μm inner/outer diameter, Polymicro Technologies) on a commercial CE instrument (model CESI 8000, AB Sciex, Framingham, MA). Before injection of the sample, the CE capillary was flushed with the BGE at 100 psi for 5 min. An ~75 nL portion of the sample was pressure-injected (25 psi for 99 sec) in the capillary, followed by pressure-injection of an ~10 nL plug of BGE (10 psi for 30 sec). Electrophoresis was conducted by applying +20 kV to the inlet end of the CE capillary (vs. Earth ground at the outlet). As detailed elsewhere,³⁶ the capillary outlet was fed into a custom-built CE-ESI interface consisting of a tapered-tip emitter. This emitter was connected to a syringe pump flooding the outlet of the CE capillary end with 200 nL/min of sheath flow to help maintain the stable cone-jet spraying regime for maximal ionization efficiency.⁴⁴ As shown in Supplementary Movie 1 (S1), stability of the Taylor cone was monitored with a long-working distance camera.

Positively charged ions were detected on a timsTOF Pro mass spectrometer (Bruker Scientific, Billerica, MA) with the following experimental settings: sampling plate, −1,300 V (vs. Earth ground); MS¹ scan range, m/z 100–1,700; 1/k₀ scan range, 0.6–1.48 V s cm⁻² over a 100 ms ramp time. Each day, the ion mobility values were calibrated to <1% accuracy with a standard calibration mix. A multiple reaction monitoring (MRM) method was developed for the 6 Ang peptides, each targeting the dominant charge states with a 2 m/z isolation window for collision-induced dissociation (CID) under individually-optimized energy setting (see Table 1).

Table 1.

Estimation of lower limits of detection (LLOD). For each peptide, 75 nL of 1 pM, or 75 amol from each peptide were analyzed to measure their electrophoretic migration time (MT), accurate mass, gas-phase ion mobility (1/k₀), and collision cross section (CCS). The detection limit was estimated by extrapolating the ion signal to a S/N of 3.

Peptide	Sequence	MT (min)	Charge	Theor. m/z	m/z error (ppm)	Collision Energy (eV)	1/k0 Range (Vs/cm2)	Experimental CCS (Å2)	Estimated LLOD
									pM	zmol
AngT	DRVYIHPFHLVIH	33.7	+2	823.4517	−1.0	10	1.327–1.354	541.6	0.3	23
			+3	549.3036	−1.7	25	0.957–0.976	585. 7
			+4	412.2295	0.2	20	0.794–0.810	647.8
Ang 1-9	DRVYIHPFH	33.1	+2	592.304	−2.2	22	0.831–0.847	339.9	0.42	32
			+3	395.2051	−1.4	22	0.682–0.695	418.4
Ang I	DRVYIHPFHL	34.0	+2	648.846	−0.9	25	1.062–1.084	434.4	0.48	36
			+3	432.8998	2.4	25	0.809–0.826	496.8
Ang II	DRVYIHPF	35.9	+2	523.7745	−0.3	26	0.788–0.804	323.2	0.5	37
			+3	349.5188	−0.9	20	0.643–0.656	395.7
Ang III	RVYIHPF	34.2	+2	466.2611	−0.3	25	0.755–0.770	309.9	0.54	40
Ang IV	VYIHPF	39.4	+1	775.4137	−0.9	22	1.217–1.241	250.6	1.1	81
			+2	388.2105	0.7	15	0.713–0.727	293.6

Data analysis.

The sequences of the Ang peptides and the raw MS-MS/MS files were processed in Skyline version 20.2.^45-46 Chromatographic peak boundaries were manually confirmed. The following transition settings were used: precursor and product mass, monoisotopic; precursor charges, 1–4; ion types, b, y, p (precursor); m/z range, 100–1,700; method match tolerance, 0.055 m/z; precursor mass analyzer, TOF. Peptides were quantified by integrating and summing the under-the-curve peak areas for both precursor and product ions.

Statistical Analysis.

Quantification was performed using 2 or 3 biological replicates, with each originating from a pool of 3 tissues, each isolated from a different animal. Biological replicates were analyzed in technical triplicate. Peak areas were averaged across 3 technical replicates (i.e., replicate injections of the same sample). For development of the method, measurement error was calculated based on the standard deviation (SD). Statistical analysis was performed using Student’s unpaired two-tailed t-test with p < 0.05 indicating statistical significance after validating the null hypothesis.

Sex as a Biological Variable.

Sex differences are known to play a role in the RAS. As the goal of this study was to develop a validated instrumental platform, all measurements were exclusively performed in the male. The technology can be readily used for replicate measurements also in female mice.

Safety Considerations.

All experiments were carried out following standard safety protocols for chemical and biological materials. Gloves and safety glasses were worn, and care was taken when handling fused silica capillaries, which present a puncture hazard. All electrically-conductive components of the CE-ESI setup were Earth-grounded to protect users against electrical shock hazards.

RESULTS AND DISCUSSION

Large-Volume Sample Stacking Improved Detection.

To address lower peptide yields from tissues upon perfusion, we developed an ultrasensitive assay integrating CE, nanoESI, and timsTOF HRMS. Figure 1C illustrates steps of the analytical workflow. After loading the sample into the CE fused silica capillary, an external electric field was applied across the BGE. Peptides migrating to the capillary outlet entered our second-generation CE-ESI interface for ionization.³⁶ The interface was constructed from a grounded tapered-tip emitter that coaxially supplied the CE capillary outlet with a low flow of sheath liquid. Under visual inspection of its liquid meniscus using a long-working distance camera and the HRMS ion signal, the CE-ESI interface was operated in the cone-jet spraying regime for maximal ion yield.⁴⁴ With manual operation, this custom-built platform delivered robust performance. We previously optimized FASS to enable the detection of ~300 zmol to 5 amol of Ang II from ~10 nL of sample using tribrid HRMS.³⁷ Here, we adapted this design onto a commercial CE ESI instrument (CESI, AB Sciex). With a capability for handling microliters of sample in automation, we anticipated this method to fulfill the bioanalytical needs of this project and be broadly adoptable in other laboratories as well.

We reasoned that stacking a larger volume of the sample would improve the limit of detection, as a higher abundance of the analyte ions would be available for HRMS detection. Our established methods based on FASS provided optimal performance from sample volumes limited to ~1–20 nL (see earlier). Encouraged by recent enhancements to the capacity of sample loading,⁴⁷ we selected dynamic pH junction to implement large-volume sample stacking (LVSS). As illustrated in Figure 1C (see inset), this approach takes advantage of the pH-dependent nature of zwitterionic analytes (peptides) to control the direction of migration. We chose the pH of the sample solvent (pH 8.2) to be basic so that analytes migrated towards the inlet end of the capillary at the start of electrophoresis. On moving through progressively lowering pH environments in the BGE (pH 2.3 in the bulk), the peptides eventually become neutral, thus undergoing local enrichment, before separation toward the outlet end of the capillary by electrophoresis. In theory, this approach was scalable to larger volumes (or concentrations) of analytes.

We benchmarked large-volume sample injection against FASS, the closest neighboring technology (Fig. 2). In a series of experiments, comparable amounts of total Ang standard peptides were measured using FASS (from acidified 50% ACN) or large-volume sample stacking (from 50 mM AmBic as the sample solvent). Figure 2A exemplifies separation of ~75 amol of each peptide by analyzing different concentrations and volumes of standards. Transient peaks in selected ion traces marked the migration time of each peptide based on the summed single-stage (MS¹) and tandem MS (MS²) scans (see Table 1). The relative migration order was reproducible between the technical replicates at the tested conditions, demonstrating robust performance. Larger loading volumes appeared to influence detection amounts for select peptides. The under-the-curve peak areas are systematically evaluated, as shown in Figure 2B. Ang 1-9, Ang II, and Ang IV produced higher ion signal, with these differences amounting to statistical significance for Ang II. Despite loading 10- to 100-times more volume of the sample than during FASS, transient peaks with similar temporal widths and S/N ratio were detected. Compared to FASS, LVSS elongated electrophoresis from ~30 min to ~45 min per measurement, which we considered a modest and acceptable trade-off to realize an ~10-times enhancement in detection limits by permitting the analysis of large volumes of the sample.

Figure 2.

Signal enhancement by LVSS. (A) Evaluation of selected-ion electropherograms for 75 amol of each of the angiotensin standards using field-amplified sample stacking (FASS) in 50% acidified ACN (left and middle insets) and LVSS with dynamic pH junction in 50 mM AmBic (right inset). (B) Comparison of under-the-curve peak areas of the peptides between technical triplicates, revealing substantial enhancement to detection limits. Key: error bars, SD; n.s., not statistically significant; statistical significance, *p < 0.05 (one-way ANOVA followed by Tukey HSD post-hoc test).

While LVSS improved detection limits, it did not necessarily address the chemical complexity of the PVN extracts. In ESI-HRMS, the molecular complexity of biological samples is known to cause signal interference during detection and peptide sequencing, which in turn is expected to worsen detection limits and the confidence of peptide identification. We posed an additional dimension of separation, ideally without contributing to sample loss or lowering analytical throughput, to be able to address these challenges. As an additional layer of separation online, prior to m/z detection by TOF, we nested ion mobility into our CE-MS workflow (recall Fig. 1). Based on recent results from bottom-up proteomics using nanoLC,⁴⁸ we anticipated gas-phase separation by this technology over milliseconds to be compatible with the duration of electrophoretic peaks (several seconds), parallel accumulation-serial fragmentation to enhance the duty cycle of tandem MS, and ion enrichment prior to detection to improve signal for HRMS detection in our experiments.

The Approach Improved Identifications to High Fidelity.

We tested these hypotheses in proof-of-principle experiments on peptide extracts from the PVN. Figure 3A shows representative separation of the 6 Ang peptides in the extract of 3 pooled brain tissue punches from perfused mice, analyzed using CE-ESI-timsTOF-HRMS. The peptides were readily detectable, each transient signals with baseline separation. The inset illustrates assessment of the S/N ratios for the ion signals. Figure 3B presents the measured 1/k0 distributions for each dominant ionic species that formed from the standards during direct infusion ESI on a calibrated timsTOF high-resolution mass spectrometer. Table 1 tabulates the experimental collision cross sections (CCS) that were derived from the experimental 1/k₀ values. As expected,⁴⁹ higher charge states yielded higher CCSs for the peptides. The 1/k₀ values for each peptide standard were then measured and used to filter the ion signals in the biological samples.

Figure 3.

Proof-of-principle experiments demonstrating improved CE-ESI sensitivity using timsTOF HRMS. (A) Representative electropherogram showing baseline separation of Ang peptide signals from tissue extract (inset) and calculation of the signal-to-noise ratio (S/N). (B) Ion mobility separation of 1 μM Ang peptide standards using direct infusion ESI, allowing us to measure their CCS. (C) Comparison of S/N for peptides detected in perfused tissue extract, revealing sensitivity enhancement by supplementary ion mobility (IM) separation. Error bars, SD based on technical triplicates. (D) Linear peptide quantification between 1–100 nM with ultrasensitive detection to 1 pM concentration using standards (all R² > ~0.8). (Inset) Representative example, showing measurement of Ang II with ~30 zmol estimated lower limit of detection. Error bars, SD based on technical duplicates.

Ion selection based on CCS aided the specificity of detection. The mobility values of peptides detected in the biological samples were within 1% of the standard measurements (see Methods), providing an additional piece of molecule-dependent information to support identifications. Our CCS values, including Ang I³⁺, Ang II²⁺, and Ang IV²⁺, were within ~10% of the values reported in the literature, without any corrections applied to the measured values.⁵⁰ This CCS accuracy combined with 0.2–2.4 ppm accuracy in m/z measurements (Table 1), and matching MS/MS to spectral libraries ensured high-confidence identification of the Ang peptides in the biological samples.

The analytical figures of merits were determined. To evaluate to what extent the incorporation of ion mobility separation improved the sensitivity of CE-HRMS, we compared the average S/N across 3 replicate injections, each with TIMS turned on and off on the same sample. As shown in Figure 3C, all 6 peptides showed a substantial improvement in S/N with the incorporation of ion mobility, with fold changes ranging by ~5- to 50-times. We attributed increased S/N to improvements resulting from multiple sources. We anticipated ion accumulation by trapping to boost ion flux sensitivity and rapid removal of interfering background ions to reduce the background chemical noise prior to TOF detection.

Higher S/N promised better sensitivity and quantification. Figure 3D presents the quantification of the Ang peptides from a dilution series. For this study, the linear dynamic range was tested over 3 log-order of concentrations. For studies needing to assess a broader range of concentration, the detector of the timsTOF mass spectrometer provides an extended dynamic range. For each peptide, the LLOD was calculated by extrapolating to a S/N of 3. As shown in Figure 3D and tabulated in Table 1, most peptides were detectable at concentrations of ~1 pM, approximately a 20-fold improvement over previous CE-ESI-HRMS designs.^{35, 37} This improvement in detection expanded Ang measurements to ultrasensitive analyses.

The additional dimension of separation also raised a potential for improved identification. In timsTOF HRMS, ion mobility separation reduces chemical interference during MS¹, allowing to select precursor ions in higher purity. In turn, they produced less-convoluted tandem spectra during CID, thus elevating the fidelity of sequence assignment. Figure 4 exemplifies the case for Ang 1-9³⁺. The precursor ion (m/z 395.2051) was isolated with a 2 m/z window and was fragmented using 22 eV collision energy. The resulting spectra were filtered in Skyline to only include signals from within the mobility range of 0.683-0.689 V s cm⁻², as was experimentally determined using a synthetic peptide standard (see Table 1). To assess the quality of the MS/MS spectra, the dot-product (dotp) score was calculated using Skyline. This score quantifies the correlation between product ion signal peak intensities in the sample and library spectra, with scores of 1 indicating high correlation and 0 indicating no correlation.⁴⁶ A similar analysis of the results found that the dotp scores increased by an average of 27% for all the Ang peptides targeted in this study. Therefore, together with enhanced LLOD by LVSS CE-ESI, timsTOF HRMS effectively enhanced quantification and the fidelity of peptide identifications.

Figure 4.

Identifications with improved fidelity using timsTOF tandem HRMS. The example compares the tandem mass spectra from Ang 1-9³⁺, with (ON) and without (OFF) trapped ion mobility spectrometry (TIMS). TIMS prior to fragmentation reduced chemical interference, improving sequence assignments, as quantified by higher dot-product (dotp) scores resulting from Skyline.

The Ang Peptides Became Detectable in Mouse Brain Tissue Punches.

These analytical performance metrics were attractive for helping to characterize peptides in the PVN. With an ~300 zmol LLOD,²⁶ our recent CE-HRMS approach enabled high-sensitivity characterization of Ang peptides in limited amounts from the PVN in wild-type (control) mice. After removal of the Ang peptides interfering from circulation (PBS perfusion), the peptide ion signal levels, however, proved still challenging to detect in the tissue. To address questions on the production of the peptides, biological studies would benefit from the removal of circulatory Ang peptides (e.g., blood) via perfusion,^{31, 36} further taxing detection limits.

The CE-ESI-timsTOF HRMS method was used to detect the tissue-endogenous peptides. By perfusing the mice prior to dissecting the PVN, blood surrounding the tissues was removed, allowing us to measure the Ang peptides predominantly in the tissues, likely at reduced levels compared to the control (no perfusion). A total of 3 tissue punches were pooled, each from the PVN (biological replicates, labeled BR1–3). Each micropunch sampled an estimated ~0.5 mm (diameter) × 1 mm (height), or ~200 nL voxel of the tissues. The pooled replicates were individually processed and analyzed. The CE-nanoESI-timsTOF platform and method developed here enabled the detection of all the RAS peptides, despite perfusion. The technical reproducibility and sensitivity of the method was sufficient to support biological studies.

Our technology lent an opportunity to estimate the local peptide amounts directly in the tissue punches. Figure 5 compares the log-normalized under-the-curve peak areas of the peptide signals in the PVN. The calculated signal areas were within the linear quantitative dynamic concentration range of the technology. Using the previous concentration calibration curve (recall Fig. 3D), the detected signals approximated to 100 pM to 100 nM of peptides in the PVN extracts. For Ang II, this concentration translated to ~100 pmol, or, assuming a wet tissue density of ~1 g/mL, ~500 fmol/g (peptide/tissue) in each PVN biopsy. These amounts are ~50-times higher than those previously reported in the average brain tissue (2–15 fmol/g)⁷. With a capability for microanalysis of local tissue regions, such as the PVN, our results contribute to questions on heterogeneous tissue distribution of the peptides within the brain. Such measurements, with proper controls in future experiments, may help resolve a discrepancy on reported amounts due to the challenging nature of quantifying these Ang peptides.^{7, 51} While this study was not designed to provide absolute amounts of each peptide in the tissue, it demonstrated that the technology is capable of confidently detecting low levels of Ang peptides in limited brain tissues with sufficient sensitivity for quantification.

Figure 5.

Robust detection of all Ang peptides in the PVN of perfused mice. Error bars show SD based on technical triplicate of the biological replicates labeled BR1 and BR3. For BR2, the bar shows the average abundance from technical duplicate analyses. Detection of all the RAS Ang peptides within the linear dynamic range of quantification demonstrated sufficient analytical performance to enable local quantitative studies on the RAS in the brain.

CONCLUSIONS

We developed a CE-nanoESI-HRMS method that enhanced detection sensitivity for angiotensin (Ang) peptides in distinct hypothalamic mouse brain nuclei. The approach benefited from multiple technical advances that were strategically chosen to integrate complementary performance metrics. LVSS using a dynamic pH junction boosted the analyzable amount of peptides in CE by ca. ten-fold, and a CE-nanoESI source featuring a tapered-tip emitter in the cone-jet spraying regime ensured efficient ionization. Trapped ion mobility separation prior to TOF HRMS provided an orthogonal dimension of separation, improving the fidelity of peptide identifications. We ascribed the observed enhancements to reduced (isobaric) peptide interferences and addition of another dimension of peptide-dependent information, the gas-phase collision cross section. The technology accomplished an ~1 pM, or ~30 zmol, LLOD and 3 orders of magnitude tested dynamic range for quantification. Ion mobility separation from our study enhanced the confidence of peptide identification, raising a potential to expand ~1 zmol LLOD by commercial CE-HRMS-based bottom-up proteomics to other laboratories.³⁵

In addition to improving bioanalysis, this work also raised a potential to elucidate the biochemistry of the RAS using the ultrasensitive CE-timsTOF-HRMS method. Key peptides of the RAS became detectable in the PVN of perfused mice using this technology, including AngT, Ang I, Ang II, Ang III, Ang IV, and Ang 1–9. Quantification of higher Ang amounts in the PVN than reported in the average brain added to questions on peptide distributional heterogeneity. Sensitive and tissue-specific detection and quantification of the peptides presents a bioanalytical leap forward toward better understanding the local RAS in the brain and its specific areas, important in the physiological and pathophysiological control of cardiovascular, thirst and stress regulation.

ABBREVIATIONS

ACN

Acetonitrile

AmBic

ammonium bicarbonate

Ang

angiotensin

ACE

angiotensin converting enzyme

AngT

angiotensinogen

BGE

background electrolyte

CE

capillary electrophoresis

ESI

electrospray ionization

FASS

field amplified sample stacking

FA

formic acid

HRMS

high resolution mass spectrometry

IM

ion mobility

LC

liquid chromatography

LLOD

lower limit of detection

MS

mass spectrometry

MS/MS

tandem mass spectrometry

MeOH

methanol

MRM

multiple reaction monitoring

PVN

periventricular nucleus

RAS

renin-angiotensin system

S/N

signal-to-noise ratio

TOF

time-of-flight

tITP

transient isotachophoresis

TIMS

trapped ion mobility spectrometry