Online capillary electrophoresis – mass spectrometry analysis of histatin-5 and its degradation products

Abstract

Histatin-5 (Hst-5) is a human salivary peptide with antibacterial and antifungal activities. Thorough characterization and reliable quantification of Hst-5 and its degradation products are essential for understanding the Hst-5 degradation pathway. Due to the highly basic and strong cationic nature of the Hst-5 peptide, the quantitative analysis of Hst-5 and its degradation forms by online mass spectrometry remains challenging. Here, we adopt a recently developed electrokinetically pumped sheath liquid capillary electrophoresis – mass spectrometry (CE-MS) coupling technology, and successfully apply it for the analysis of Hst-5 and its degradation products. Our CE-MS method is demonstrated to be robust and quantitative. This novel analytical platform is reproducible and free of sample carryover. The efficacy of this method is demonstrated with a kinetic study of Hst-5 degradation by Sap9, a secreted aspartic peptidase. Our work demonstrates the potential of online CE-MS as a powerful approach for characterizing highly basic peptides.

Graphical Abstract

INTRODUCTION

Human saliva contains numerous antimicrobial antibodies, enzymes, and peptides¹ that function to defend against commonly found fungi and other foreign species in the human oral cavity. One such fungus, Candida albicans, is an opportunistic pathogen responsible for candidiasis in humans^2–4. This microbe produces a family of secreted aspartic peptidases, Saps, which include an abundant host salivary antifungal peptide, histatin-5 (Hst-5), among their substrates⁵. This proteolytic degradation of Hst-5 by Saps, along with the strong anti-Candida activity of Hst-5, suggests that Hst-5 is involved with host defense against infection from Candida albicans⁶.

Hst-5 peptide and its degradation products have previously been analyzed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry⁸, and the detection and quantification of histatins have been observed by reversed-phase high-performance liquid chromatography (HP-RPLC) with optical detection using a mobile phase incompatible with electrospray ionization^8,9. Despite the use of MALDI-TOF and HPLC-UV for the measurement of the Hst-5 peptide, direct liquid chromatography with mass spectrometry (LC-MS) for the analysis of Hst-5 and its degradants remains challenging, due to severe chromatographic peak broadening, poor separation resolution, and significant sample carryover, which are caused by the interactions between highly basic peptides and chromatographic packing materials.

In our earlier work, we tested a collection of HPLC columns including C18, C3, C8, and monolith C18. It was typical to have Hst-5 eluting over a 6 to 10-min window with severe carryover. Extracted ion chromatograms (EICs) of m/z 506.9239, corresponding to [M+6H]⁶⁺ of Hst-5 can be found in Figure S-1 (top). Not only did Hst-5 elute over a 15-min period, we also observed a carry-over of ~50%, as demonstrated in the blank injection taken after Hst-5 (Figure S1, bottom). Without an acceptable separation mechanism, we resolved to desalting followed by flow injection at the beginning of our work, knowing it was not ideal due to manual injection and potential signal suppression. Even with extensive washing in flow injection, we were observing significant carryover.

Capillary electrophoresis (CE) is a separation technique that is highly complementary to liquid chromatography. The separation principle of CE is based on the mobility of molecules under an electric field, as opposed to the interaction between analytes and chromatography packing materials¹⁰. Recent breakthroughs in CE-MS coupling technologies have overcome the historic difficulties of directly coupling CE to MS and have made CE-MS an increasingly appealing analytical platform for biological molecular analysis^11,12. In this study, we utilize an electrokinetically pumped sheath liquid CE-MS ion source (Figure S-2)^11,13,14. Briefly, a fused silica capillary is inserted into a borosilicate electrospray emitter that has been filled with sheath liquid. The borosilicate glass emitter generates an electroosmotic flow towards the mass spectrometer, producing a steady nanoflow of the sheath liquid.

We demonstrate the applications of this novel CE-MS technology on the qualitative and quantitative analysis of Hst-5 and its Sap9 cleavage products. No carryover is observed between subsequent sample injections. In addition, the system is robust and shows highly reproducible migration time with a Hst-5 linear dynamic range covering concentrations of biological interest (2.5 μM to 50 μM). The efficacy of this method is demonstrated by monitoring relative quantitative changes in Hst-5 degradation products over increasing incubation periods with Sap9. The separation effectiveness suggests the potential application of this technology to the analysis of a broad family of highly basic and cationic peptides.

EXPERIMENTAL

Reagents

Acetic acid (LC-MS grade), formic acid (LC-MS grade), acetonitrile (Optima™ LC/MS grade), and methanol (Optima™ LC/MS grade) were purchased from Fisher Scientific. The Hst-5 peptide (DSHAKRHHGYKRKFHEKHHSHRGY) was synthesized by GenScript with a purity ≥95% and trifluoroacetic acid salt removal to hydrochloride. Purified Sap9 protease (produced in Pichia pastoris without its glycosylphosphatidylinositol anchor¹⁵) was kindly provided by Prof. Bernhard Hube (Leibniz Institute for Natural Product Research and Infection Biology – Hans Knöll Institute, Germany). All other chemical reagents were purchased from Sigma Aldrich.

Sample preparation

Hst-5 peptide was dissolved in autoclaved 18 mΩ ionic purity water. For the Sap9 enzymatic kinetics study, a solution of 100 μM Hst-5 in water was mixed with 6 ng/μL Sap9 in 2 mM sodium phosphate buffer at a 1:1 ratio and incubated at 37 °C for time increments ranging from 5 min to 3 h. The enzyme was inactivated by heating at 100 °C for 5 min.

The peptide samples were desalted using C18 TopTip spin columns (Glygen, Columbia, MD). After an initial column bed wetting with 0.1% formic acid/60% acetonitrile and rinsing with two rounds of 0.1% formic acid, the peptide samples were loaded onto the spin columns, desalted by a solution of 0.1% formic acid, and eluted by 0.1% formic acid/60% acetonitrile in water. Due to the hydrophilic nature of this peptide, we acknowledge that desalting on C18 media will result in some sample loss. However, this is necessary to remove the phosphate buffer, which otherwise creates a discontinuous potential across the capillary, with a suppressed electric field strength in the sample, leading to peak broadening and splitting. Unlike a reversed-phase separation of peptides, this step only requires the separation of salt from peptides, not the separation of peptides from one another. The eluent was diluted two-fold with 0.1% formic acid. One microliter of MRFA peptide solution (191 μM in methanol) was added to each sample solution as an internal standard prior to CE-MS analysis.

CE-MS analysis

A Thermo Scientific™ Orbitrap Fusion™ Lumos™ Tribrid™ mass spectrometer was coupled with an ECE-001 capillary electrophoresis unit through an EMASS-II CE-MS ion source (CMP Scientific Corp., Brooklyn, NY 11226). A PS1 (CMP Scientific, Brooklyn, NY) neutral coated capillary (360 μm O.D., 50 μm I.D., 75 cm) was used for separation. Electrospray emitters are 1 mm O.D. and 0.75 mm I.D., with tip orifice sizes of 20–30 μm (CMP Scientific Corp.). Samples were injected at 2 psi for 7 sec (~20 nL) using the built-in autosampler. The sample tray was maintained at 10 °C and the buffer tray was maintained at 20 °C. Capillary temperature was not controlled. The background electrolyte for CE separation was 20% acetic acid. The sheath liquid for electrospray was 15% acetic acid/15% acetonitrile. Electrophoresis was carried out at 30 kV. Electrospray ionization voltage was controlled by the CMP ion source at 2.0 kV, while the source voltage of the Orbitrap Fusion Lumos mass spectrometer was set at 0 kV. The electrospray emitter was placed 3 mm away from the mass spectrometer inlet.

Mass spectra were acquired from m/z 250 to 1,400 using quadrupole isolation, with mass resolution of 120,000 at m/z 200. For identification of degradation products, data dependent HCD spectra were recorded at a resolution of 120,000 (m/z 200) under “top of speed” mode with a cycle time of 3 s and NCE of 35. All other mass spectrometer acquisition parameters such as automatic gain control, number of microscans, and maximum ion injection time, were set to the default values for the Orbitrap Fusion Lumos mass spectrometer.

Data acquisition and analysis

The MS data acquisition and analysis were performed using Thermo Scientific™ Xcalibur™ 3.0 software. For identification and verification of Hst-5 and its degradation products, HCD spectra were processed using the ProSightPD node in Proteome Discoverer software (v2.1, Thermo Scientific). Because digestion sites of Sap proteases are not well defined. We used ProsightPD (v 1.2) to identify degradation products of Hst-5 and its analogs after Sap treatment. Sequences of all Hst-5 analogs were compiled into a FASTA file, which was then used to create a ProsightPC database through the ProsightPC database wizard. The search was conducted with the ProsightPD Biomarker Node, which allowed identification of partial sequences (degradation products) that best match raw data. A 20-ppm tolerance was allowed for both precursor and fragment ion masses. Spectra for all degradation products were then manually evaluated for bond coverage and signal to noise ratio of matched fragment ions. The capillary electrophoresis method setup and operation were performed using Clarity software. For peak integration, the entire electropherogram was deconvoluted using the Xtract program in Xcalibur software such that contribution from all charge states are included in quantification.

RESULTS AND DISCUSSION

Using electrokinetically pumped sheath liquid nanospray CE-MS (Figure S-1), we successfully developed a novel approach for the characterization of Hst-5 and its degradation products. The performance of this CE-MS analytical platform was evaluated for peak shape, separation, system robustness, sample carryover, reproducibility, and linearity for quantitative analysis.

Reproducibility

Classically, CE separations have struggled with reproducible migrations times^17,18,19. Peptide accumulation on the interior walls of bare fused silica tends to lead to a steady shift in sample migration with sequential runs^20,21. While the specificity of mass spectrometry has partially removed the need for consistent migration times as a means of sample identification, it is still necessary to achieve reproducible peaks for quantitative analysis and downstream data processing. Here we chose to employ a separation capillary with a neutral coating, which reduces sample adsorption to the capillary walls.

In order to validate the reproducibility of our method, we performed six sequential injections of intact Hst-5 at 50 μM (Figure 1). Table 1 displays the highly reproducible migration times, peak heights, and peaks areas. Note that peak area and height were notably lower in injection 3. This may be the result of a small air bubble in the sample vial. Compared to LC-MS injections, sample volume in CE-MS is usually one hundred-fold lower. In our method, we apply 2 psi pressure for 7 seconds and the estimated injection volume is ~20 nL, which is <1.5% of the capillary volume. At this volume, a tiny air bubble could significantly reduce the amount of sample injected. In subsequent data, including all figure data following the runs presented in Figure 1, a standard tetrapeptide, MRFA, was added as an internal standard to normalize potential variation in sample loading. MRFA was well separated from Hst-5 and its degradation products.

Figure 1.

Total ion electropherograms of six consecutive injections of Hst-5. Normalized peak intensity and migration time are shown. Injections were manually stopped after the elution of the Hst-5.

Table 1:

Reproducibility of sequential Hst-5 injections

Injection	Migration Time (min)	Peak Height (e9)	Peak Area (e10)
1	8.60	25.21	23.27
2	8.54	26.37	27.73
3	8.49	9.87*	9.10*
4	8.45	21.89	22.79
5	8.51	21.55	21.39
6	8.42	23.58	25.06
Average	8.50	21.41**	21.56**
		23.72	24.05
Standard Deviation	0.06	5.95**	6.48**
		2.08	2.44
CV (%)	0.76	27.80**	30.07**
		8.77	10.15

^*Low values are likely a result of an air bubble during sample injection

^**Including injection 3

Dynamic range and carryover

50 μM was selected as the upper limit of concentration. Figure 2 shows the deconvoluted, extracted ion electropherograms of Hst-5 at concentrations from 2.5 to 50 μM (50 fmol to 1 pmol injected). Signal intensity is observed to increase linearly with concentration, with an R² value of ~0.995. Figure 3 plots Hst-5 concentrations against normalized peak areas. Peak areas are normalized to the area of the MRFA peak in the same injection to correct for potential variations in signal intensity. As the goal in developing this method was evaluating of Sep enzymatic digestion of Hst-5 peptide, we did not investigate the limit of detection (LOD), which may empirically be defined as 2.5 μM. However, the LOD can be theoretically calculated from Figure 3, assuming that the linear trend continues. LOD is defined as the lowest concentration yielding a standard deviation of 3.3 times the signal to noise. Based on our data, the theoretical LOD may be calculated as 0.73 μM. It is worth noting that the migration time of Hst-5 maintained reproducibility across the nine consecutive injections throughout this analysis.

Figure 2:

Extracted ion electropherograms of Hst-5 at each measured concentration with migration time reported. The last trace is a blank injection, taken immediately after analysis of the 50 μM standard.

Figure 3:

Normalized Hst-5 peak area against sample concentration. The linearity is associated with an R² value of 0.995.

In our earlier study, we observed Hst-5 strongly interacting with a multitude of common LC stationary phases, resulting in peptide elution during blanks and making quantitation impossible (Figure S-1). A blank CE-MS electropherogram taken after completion of the dynamic range analysis is shown as the last trace in Figure 2. No sample retention (carryover) was observed. Some peak tailing of Hst-5 was observed, which may be the result of minor interactions with the capillary walls. Exploring alternative capillary coatings in the future may help create a more uniform peak distribution.

Identification and validation of Hst-5 degradation products

Hst-5 was incubated with Sap9 in order to create a mixture of intact peptides and degradation products. CE-MS/MS with HCD fragmentation was used for Hst-5 degradation product identification. The resulting MS/MS spectra were matched to the Hst-5 amino acid sequence using the Prosight PD node in Proteome Discoverer software. Peptide identifications were then evaluated by manual inspection of the matched b and y ions. An example MS/MS spectrum of intact Hst-5 is shown in Figure 4.

Figure 4:

Annotated HCD MS/MS spectrum of intact Hst-5, with bond coverage and Prosight Lite score.

The migration order of degradation product traces shown in Figure 5 adds a second metric of identification confidence. In theory, CE migration time is determined by mass over charge (m/z). Peptides with lower m/z values migrate more quickly, as expected. Minimum theoretical m/z values for each degradation product (assuming all amino acids H, R, K have +1 charge) can be found in Table 3. Although smaller molecules usually migrate faster, a larger molecule with higher charge can migrate faster than a smaller molecule that has less charge. This is reflected in Figure 5, where fragment 14–24 migrated more slowly than the larger, intact peptide 1–24.

Figure 5:

Extracted ion electropherograms of major degradation products after (A) 30 and (B) 90-min incubation with normalized peak intensities, extracted mass ranges, and migration times reported. After deconvolution, the most intense mass of each isotopic profile is extracted. Peptide identities are shown to the right, using the amino acid numbering from the full, 24 AA, Hst-5 sequence.

Table 3:

Maximum theoretical charge states and observed migration times on degradation peptides of Hst-5.

Sequence	Max Theoretical Charge	Min Theoretical m/z	Migration Time (min)
6–24	+12	208.1	8.96
7–24	+11	212.8	8.97
1–17	+10	216.1	9.02
1–24	+14	216.8	9.12
13–24	+7	223.1	9.16
14–24	+6	238.8	9.59

Maximum theoretical charge is calculated by assuming that all histidines, arginines, and lysines are in a +1 charge state.

The peak widths of many fragments in our method are less than 15 s (Figure 5). An MS duty cycle of 3 s in our initial data dependent acquisition collects enough spectra for peptide identification, but will not collect enough data points across the migration profile for accurate quantification and visualization of this narrow peak shape. Therefore, once we’ve identified Hst-5 degradation products in the initial work, we decrease cycle time by removing MS/MS spectra and acquiring only full scan MS1 in subsequent analyses. Peptides were identified by their deconvoluted molecular weight using the Xtract program within Xcalibur software (Thermo). Peptide identities are shown to the right of Figure 5, using the amino acid numbering from the full Hst-5 sequence.

After deconvolution of the entire electropherogram, extracted ion electropherograms of each degradation product were reconstituted, peak area integrated, and normalized to the internal standard to determine relative quantity. Example electropherograms are displayed in Figure 5. The most intense mass from each deconvoluted peptide is traced in the figure. Most degradation products generate a symmetric peak shape. Hst-5 displays some tailing, and splitting is observed in fragment 6–24, but both are significantly improved from LC conditions. All histatin peaks migrate through the capillary in less than ten min, as a result of their extremely high charge density. As such, the separation is very efficient and amenable to high throughput analysis.

Hst-5 kinetic analysis

Here we demonstrate the efficacy of our CE-MS method by evaluating the kinetics of Hst-5 degradation. A more expansive kinetic study using this technique has been recently reported by Ikonomova et al²².

Hst-5 was incubated with Sap9 for 0, 5, 15, 30, 45, 60, 90, 120, and 180 min. The deconvoluted peak areas of major degradation products were normalized to the internal standard. As an example, two of these separations are exhibited in Figure 5A (30 min) and 5B (90 min).

Figure 6 shows the change in quantities of Hst-5 and each major degradation product over the course of the incubations. For ease of visualization, the figure is split into three different intensity ranges and the amino acid sequence of Hst-5 is included as a reference.

Figure 6:

Fragment peak areas normalized to the internal standard (MRFA) peak area against incubation time. For display clarity, traces are split into (A) high MS intensity, (B) medium MS intensity, and (C) low MS intensity. The Hst-5 amino acid sequence is DSHAKRHHGYKRKFHEKHHSHRGY. Numbers in trace labeling for A-C correspond to the positions in the Hst-5 amino acid sequence.

There are a few notable trends. We observed preferred cleavage around lysine residues, as previously reported^6,22,23. As incubation time is increased, a greater portion of Hst-5 is digested, resulting in a decrease in Hst-5 peak area (Figure 6A). In general, there is a corresponding increase in the peak areas of degradation products. The initial, rapid increase of fragment 1–17 relative to the other fragments (Figure 6A) likely indicates that cleavage first occurs between residues K17 and H18, in agreement with Bochenska et al⁵. The complementary fragment, 18–24, also increases alongside fragment 1–17, but may be complicated by further, simultaneous degradation into smaller fragments.

No products continued to increase after 120 min of digestion, indicating that Hst-5 was near complete degradation, or the degradation products have begun breaking down into smaller, undetected fragments.

Fragments 6–24 and 7–24 (Figure 6C) initially showed accumulation but began declining after 90 min. It is probable that they further degrade into fragments 13–24 and 14–24 (Figure 6A, 6B), as the peak areas of these smaller fragments continued to grow until at least 120 min. Fragments 6–24 and 7–24 are large peptides that are close in size to Hst-5 and may effectively compete with Hst-5 as Sap9 substrates.

The antifungal activity of some of the fragments we detected has been previously tested. The N-terminal fragments 1–12 (complementary to fragment 13–24)²⁴ and 1–14 were previously reported to be inactive^24,25, while C-terminal fragments 13–24 had reduced activity compared to Hst-5. While some fragments retain activity, their potency is masked when all fragments are pooled^6,22,23.

CONCLUSION

We demonstrate a simple CE-MS method for the separation and online analysis of Hst-5 and its degradation products. Separations are demonstrated to be highly reproducible with no carryover between sequential runs. A linear relationship is observed over the concentration range of biological interest. To demonstrate the method’s efficacy, relative quantities of Hst-5 degradation products are measured against increasing digestion time with the aspartic peptidase Sap9. The developed method has the potential for use in Hst-5 studies beyond the scope of this publication, including amino acid substitution activity assays, Sap enzyme competitive analyses, and fundamental drug discovery. It has also been used to determine the preferred cleavage sites of Sap enzymes. By enabling online analysis and removing the need for fraction collection and subsequent analyses of individual fractions, assay throughput and accuracy may be greatly improved, which opens the door for more extensive work in understanding the antifungal activity of Hst-5 and other highly basic peptides.

Table 2:

Normalized Hst-5 peak areas against sample concentration

Concentration (μM)	Peak Area	MRFA Peak Area	Normalized Peak Area
2.5	4.89E+07	1.63E+08	0.30
5.0	1.72E+09	1.25E+09	1.38
7.5	2.37E+09	1.43E+09	1.66
10.0	2.93E+09	1.27E+09	2.30
12.5	3.40E+09	1.21E+09	2.81
15.0	4.85E+09	1.23E+09	3.93
20.0	6.52E+09	1.34E+09	4.87
25.0	2.30E+10	3.35E+09	6.87
50.0	2.44E+10	1.92E+09	12.70