Distinguishing Histidine Tautomers in Proteins Using Covalent Labeling-Mass Spectrometry

Abstract

In this work, we use diethylpyrocarbonate (DEPC)-based covalent labeling together with LC-MS/MS analysis to distinguish the two sidechain tautomers of histidine residues in peptides and proteins. From labeling experiments on model peptides, we demonstrate that DEPC reacts equally with both tautomeric forms to produce chemically different products with distinct dissociation patterns and LC retention times, allowing the ratios of the two tautomers to be determined in peptides and proteins. Upon measuring the tautomer ratios of several histidine residues in myoglobin, we find good agreement with previous 2D NMR data on this protein. Because our DEPC labeling/MS approach is simpler, faster, and more precise than 2D NMR, our method will be a valuable way to determine how protein structure enforces histidine sidechain tautomerization. Because the tautomeric state of histidine residues is often important for protein structure and function, the ability of DEPC labeling/MS to distinguish histidine tautomers should equip researchers with a tool to understand histidine residue structure and function more deeply in proteins.

Graphical Abstract

INTRODUCTION

Characterizing protein higher-order structure is important because structure and function are closely correlated. In addition, protein structural information provides insights about their folding pathways, potential for aggregation, interactions with other compounds, and physicochemical properties. Mass spectrometry (MS) is emerging as a powerful way to study protein structure due to its high sensitivity, good structural resolution, relatively high-throughput, and ability to study proteins in mixtures. A variety of approaches, such as hydrogen/deuterium exchange (HDX), covalent labeling (CL) or footprinting, and chemical cross-linking (XL), have been combined with MS to provide information about protein higher-order structure and interactions in solution. Among these approaches, CL is emerging as a powerful tool, giving residue-level structural information via reactions of solvent exposed residues on the protein surface.^1–3 The method can be simple, and it usually has high sensitivity and good time resolution. In contrast to chemical XL methods, data interpretation is relatively straightforward. It also has some advantages over HDX such as no back-exchange and no labeling scrambling. Up to now, this method has been successfully applied for studying protein-protein interactions,^4,5 probing protein aggregation,^6,7 evaluating ligand binding sites and affinities,^8,9 and monitoring protein folding/unfolding.^10,11 CL-MS has also been explored to analyze membrane proteins,¹² for refining computer-based structural predictions,^13,14 and for studying protein structure and interactions in cells and in vivo.^15–17

Covalent labeling reagents, in general, can be classified as ones with broad specificity (e.g., hydroxyl radicals, carbenes and trifluoromethyl radicals) or ones that are residue specific (e.g., 2,3-butanedione, dimethyl(2-hydroxy-5-nitrobenzyl) sulfonium bromide). Other reagents, such as diethylpyrocarbonate (DEPC), could be considered pseudo-specific. DEPC modifies side chains of nucleophilic residues such as Cys, His, Lys, Tyr, Ser, and Thr residues and the N-termini in proteins. Compared to methods that use broadly specific reagents, such as hydroxyl radical footprinting (HRF) that often need a sophisticated apparatus (e.g. laser, flash lamp), DEPC labeling can be conducted more readily by simply adding the reagent to solution. The reaction generates a single product with a mass shift of 72.02 Da, simplifying the identification of labeled sites. Our lab and others have used this reagent to investigate functional histidines,¹⁸ protein-protein interactions,^19,20 protein-ligand interactions,^21,22 and antibody structure.^23–25

DEPC reacts most readily with histidine residues, allowing the accessibility of this residue to be effectively probed. In proteins, histidine exists in an equilibrium between two tautomers, with the hydrogen on the ε nitrogen (N^ε-H) being the preferred form over the hydrogen on the δ nitrogen (N^δ-H) (Figure 1). Even though the N^ε-H tautomer is preferred, the N^δ-H form is dominant in special cases in proteins, such as in the catalytic His at the active site of serine proteases^26,27 or other structural His residues that have unique H-bonding in protein pockets.²⁸ Exchange between these two tautomers is also a mechanism by which proton transfer can be mediated in proteins.^29,30 Distinguishing these two tautomers typically requires 2D nuclear magnetic resonance (NMR) experiments,³¹ usually along with estimations of a given His residue’s pK_a and thus its protonation state.³² Because histidine residues have two tautomers, their side chains can be potentially labeled by DEPC at two different sites, producing two isomers (Figure 1). In this work, we have investigated whether DEPC labeling and LC-MS analysis are able to distinguish these two tautomers. Identifying the ratio of these isomers can yield deeper insight into histidine residue structure and function, effectively providing higher resolution protein structural information. To measure tautomer ratios, we conducted DEPC labeling on model peptides and tryptic peptides and have analyzed the resulting products by LC-MS/MS. Upon identifying experimental conditions that enable the two isomers to be distinguished, we then use DEPC labeling to identify the histidine tautomers present in proteins. These measurements provide information about how the local structure around the histidine residues influences the formation of a given tautomer.

Figure 1:

DEPC structure and its reaction with the two different histidine tautomers.

EXPERIMENTAL SECTION

Materials

The proteins and reagents used in this work are described in the Supporting Information. The model peptides that were used are Fmoc-DGHGG-NH₂, angiotensin II, and kinetensin, and the model proteins that were studied include β-2 microglobulin (β2m), myoglobin, and carbonic anhydrase.

DEPC Labeling Reactions

Lyophilized powders of the model peptides (Fmoc-DGHGG-NH₂, angiotensin II, or kinetensin) were first reconstituted in water, and then diluted in 10 mM MOPS buffer (pH 7) to make a final solution of the desired concentration. For angiotensin II and kinetensin, the N-termini of the peptides were first acetylated for 1 h at room temperature using sulfo-NHS-acetate (NHSA) in a 10-fold molar excess before the DEPC labeling reaction. Acetylating the N-terminus ensured a high yield of modified His residues in these peptides. DEPC stock solutions were prepared in dehydrated acetonitrile. Each acetylated model peptide (50 μM) was reacted with DEPC at 37 °C for 1 min at molar excesses ranging from 1 to 18. The reaction was then quenched by the addition of imidazole at a 1:50 DEPC to imidazole molar ratio.

Model proteins (50 μM) were prepared in 10 mM MOPS buffer (pH 7). Protein labeling was initiated by adding DEPC at molar excesses ranging from 2 to 8. After 1 min at 37 °C, the DEPC labeling reactions were quenched by imidazole. The final amount of acetonitrile in the protein solution was 2% or less.

For DEPC labeling reactions on tryptic peptides, peptides from β2m, myoglobin, and carbonic anhydrase were first generated via proteolytic digestion with trypsin (see Proteolytic Digestion). N-termini blocking of the peptides was achieved by reacting NHSA for 1 h at a 10:1 molar ratio of reagent to the number of potential amine groups. The acetylated tryptic peptides were then labeled by DEPC at 37 °C for 1 min at molar excesses ranging from 50 to 150 before being quenched by imidazole.

Proteolytic Digestion

The proteins, before or after a DEPC reaction, were preconcentrated and desalted in MOPS buffer via Amicon® centrifugal filters with a 10 kDa molecular weight cutoff (MWCO). Then, 10% (v/v) acetonitrile was added to β2m and myoglobin samples, followed by incubation at 50 °C for 45 min to denature the proteins. The disulfide bond of β2m was reduced by TCEP in a 40-fold molar excess. After 3 min at room temperature, iodoacetamide was added in an 80-fold molar excess to alkylate the resulting free Cys residues. The samples were kept in the dark at room temperature for 30 min. To unfold carbonic anhydrase, 8 M urea was added to the samples. After 10 min, the samples were desalted and preconcentrated again to ensure the concentration of urea was less than 2 M. Immobilized trypsin slurry was prewashed 3 times with a MOPS buffer. The digestion was performed for 16 hours at 37 °C by adding immobilized trypsin at a 1:4 (v/w) enzyme to substrate ratio. Finally, the enzyme was separated by centrifugation.

Liquid Chromatography

Online LC-MS/MS analyses were performed on a Thermo Scientific Dionex Ultimate 3000 HPLC system (Waltham, MA) with a Thermo Scientific Acclaim^™ PepMap^™ RSLC C18 column (15 cm × 300 μm, 2 μm particle size, 100 Å pore size) at a flow rate of 4 μL/min. The mobile phases were water and acetonitrile, both containing 0.1% formic acid. For model peptides, the mobile phase started at 5% acetonitrile for the first 5 min and was then followed by an isocratic elution at 25% acetonitrile for 25 min. For protein digests, the mobile phase started at 5% acetonitrile, and after 5 min, it was increased to 50% acetonitrile over 50 min before being changed to and held at 95% to flush the column for 5 min.

Mass Spectrometry

Mass spectral data of the model peptides were acquired on a Bruker AmaZon quadrupole ion trap (Billerica, MA). The electrospray needle voltage was kept at 4.1 kV with the capillary temperature at 250 °C. To identify the modification sites, collisional-induced dissociation (CID) was conducted with a ramp of voltages ranging from 0.9 to 2.7 V. Data processing was performed on Bruker Compass^™ data analysis 4.0.

Mass spectral data of the protein digests were acquired on a Thermo Scientific Orbitrap Fusion mass spectrometer (Waltham, MA). The electrospray voltage was 4 kV, and the ion transfer tube temperature was 275 °C. Tandem MS on peptides was conducted using CID. The precursor ions (unmodified or modified peptide ions) were selected using a quadrupole mass filter. CID was performed in the linear quadrupole ion trap with collision energy of 35%. Product ions were detected by the orbitrap analyzer with a resolution of 30,000. Data processing was performed using Thermo Xcalibur.

Molecular Dynamics (MD) Simulations

See Supporting Information for details.

RESULTS AND DISCUSSION

DEPC Labeling on Model Peptides

The histidine side chain is a heterocycle having two nitrogen atoms. The non-bonding electron pair on one of the nitrogens (black in Figure S1) is part of the aromatic π-electron system, while the electron pair on the other nitrogen is not (red in Figure S1). This latter nitrogen reacts more readily with an electrophile like DEPC without disrupting the aromatic system. The nucleophilic reaction with DEPC generates a carbethoxylated product with a mass shift of 72.02 Da (Figure S1). Because histidine has two tautomers and one nitrogen in each tautomer will initially react with DEPC, distinguishing the two isomers is feasible. Longer reaction times and higher DEPC concentrations can cause a second carbethoxylation to occur on the histidine side chain (Figure S1),³³ but this reaction can be avoided by controlling the labeling reaction conditions.

To test our ability to distinguish the two tautomers via DEPC labeling, we reacted several histidine-containing model peptides. The first peptide was Fmoc-DGHGG-NH2, which included an Fmoc group on the N-terminus to ensure that only histidine was labeled. During the LC run, an isocratic elution was used to maximize the separation of the two isomers and to ensure that the mobile phase composition did not impact peptide ionization efficiency, allowing us to better quantify the ratios of each tautomer. Figure 2 shows the extracted ion chromatograms of the unmodified (blue trace) and modified (red trace) peptides. The peaks eluting at 11.4 and 13.6 min have identical m/z ratios and are both singly-modified species with the DEPC label found on the histidine residue, according to the tandem MS data (Figure 3). Furthermore, the tandem mass spectra of these isobaric ions are different, suggesting that they are isomeric forms. One of the key differences in the MS/MS data is that modified species 2 has a significantly more abundant carbethoxylated a₃ ion (i.e. a₃*) compared to modified species 1.

Figure 2:

Extracted ion chromatograms of the unlabeled (blue trace) and DEPC labeled (red trace) peptide Fmoc-DGHGG-NH₂. The concentration of DEPC was 10 times higher than the concentration of the peptide, and the reaction was conducted for 1 min.

Figure 3:

Tandem mass spectra of two modified His isomers. a) Tandem mass spectrum of modified species 1 from the chromatogram in Figure 2. b) Tandem mass spectrum of modified species 2 from the chromatogram in Figure 2. The product ions that are indicated with an asterisk (*) contain the carbethoxylated product.

Another important observation is that the abundance ratio of the two DEPC labeled forms is 3.9 ± 0.2 based on three replicate experiments. This result is important because NMR studies on histidine tautomers in the free amino acids^34,35 and unstructured polypeptides^36–38 show that at neutral pH, the N^ε-H tautomer is favored ~4:1 over N^δ-H tautomer. Thus, the ratio of peak areas of two modified species suggests that modified species 1 has the DEPC modification on N^ε2, which means labeling occurs via the N^δ-H tautomer, while modified species 2 has the DEPC label on the N^δ1, which means the labeling occurs via the N^ε-H tautomer. Further evidence for these assignments can be obtained by considering the peptide dissociation pathways and the effect of histidine modification on the dissociation patterns. Modified species 2 has a more abundant a₃* ion in its tandem mass spectrum because the location of the DEPC label on the side chain facilitates dissociation via the b_x-y_z pathway and subsequent loss of CO by the b_x-a_x pathway.³⁹ In contrast, the location of the DEPC label in modified species 1 allows the histidine pathway^39,40 and preferential formation of a more stable b₃* ion because of the favorable five-membered ring that can form (Figure 4). The same five-membered ring formation is prevented in modified species 2 with the DEPC label on N^δ1. These dissociation preferences further support the assignments of species 1 as the N^δ-H tautomer and species 2 as the N^ε-H tautomer.

Figure 4.

Peptide fragmentation pathways of two isomers during CID of DEPC-labeled histidine-containing peptides. The left pathway is for modified species 1 (N^δ-H tautomer), and the right pathway is for modified species 2 (N^ε-H tautomer).

To further assess the ability of DEPC labeling to distinguish the two histidine tautomers, we investigated other histidine-containing peptides as well. The N-termini and other DEPC-modifiable residues were acetylated before DEPC labeling to ensure only the histidine residue was modified in each peptide. As examples, LC-MS results indicate that the ratios of the areas for modified species 2 to modified species 1 are 3.3 ± 0.1 and 2.3 ± 0.2, respectively, for the peptides angiotensin II and kinetensin (Figure S2). In the case of angiotensin II, modified species 2 has a more abundant a₆²⁺* ion compared to modified species 1 (Figure S3), again suggesting that modified species 2 is the N^ε-H tautomer. Similarly, modified species 2 for kinetensin has a more abundant a₅²⁺* ion than modified species 1 (Figure S4). Together these results, along with the results with Fmoc-DGHGG-NH2, indicate that we can identify the two histidine tautomers, and the MS/MS data help confirm which peak is the N^δ-H tautomer and which is the N^ε-H tautomer. The measured tautomer ratio of 3.3 for angiotensin II is essentially the same as the previously measured value of ~4 by NMR.³⁷ The value 2.3 for kinetensin is close to the values expected for free histidine at neutral pH and for other peptides that have been measured by NMR.^36,38 The slight change from a value of ~4:1 might be expected, as the pK_a of histidine can somewhat influence the ratio of the tautomers.³² The different amino acid sequences of these two peptides almost certainly influences the pK_a of the histidine residue in each peptide.

DEPC Labeling on Tryptic Peptides at Different Concentrations

A prerequisite for accurately determining the ratio of the two tautomeric forms is that both histidine tautomers react at the same rate with DEPC. It is reasonable that the reactive nitrogens in the tautomers have similar nucleophilicity, but to test this, we reacted several peptides at different DEPC concentrations and measured their labeling ratios. Because DEPC hydrolyzes in water with a half-life of approximately 9 min at pH 7 and 25 °C,⁴¹ we assessed reactivity by increasing DEPC concentrations with a constant reaction time of 1 min. Figure 5 displays the DEPC labeling results for select peptides. In each case, the labeling ratio does not vary over about an order of magnitude of DEPC concentrations. For example, Figure 5a indicates that the ratio of two modified forms for Fmoc-DGHGG-NH₂ remains essentially constant around a value of 4 at DEPC concentrations ranging over an order of magnitude. Other peptides have slightly different ratios (Figure 5b–5d), but the labeling ratios stay constant for a given peptide as the DEPC concentrations are increased. Interestingly, all the studied peptides show a higher abundance of the N^ε-H tautomer (Table S1) as expected, although it should be noted that for some peptides the MS/MS data make it difficult to confirm the tautomer identity (e.g. Ac-HKIPIK(Ac) in Figure S5 and S6). Challenges in determining the tautomer identity with MS/MS arise primarily when the histidine residue is on the N-terminus of the peptide, as b₁ and a₁ product ions are usually not observed in tandem mass spectra. It should be noted, however, that for all other DEPC-modified peptides in which histidine is not the N-terminal residue, the a_n*/b_n* product ion ratio, where n is residue number corresponding to the modified histidine, is always higher for the N^ε-H tautomer than for the N^δ-H tautomer. Indeed, in many cases the a_n*/b_n* product ion ratio for the N^δ-H tautomer is 0. Moreover, in every peptide where MS/MS can confirm the identity of the tautomer, it is found that the N^ε-H tautomer elutes later than the N^δ-H tautomer. This predictable elution behavior allows identification of the tautomer in the absence of adequate MS/MS data. Overall, these results demonstrate that the two tautomeric forms are equally reactive, allowing us to use the labeling and MS/MS data to determine the ratio of the two tautomers in proteins.

Figure 5.

The labeling ratio of the two modified forms for select peptides at increasing DEPC concentrations. a) Fmoc-DGHGG-NH₂; b)Ac-IQVYSRHPAENGK(Ac); c) Ac-VEADIAGHGQEVLIR; d) and Ac-LFTGHPETLEK(Ac). MS/MS was used to confirm the identity of each tautomer by measuring the ratio of the a_n/b_n ions.

DEPC Labeling of Proteins to Determine Histidine Tautomer Ratios

Two intact proteins, myoglobin and β2m, were separately labeled by DEPC. After digestion, the labeling ratios of the two tautomers for various histidine residues were measured by LC-MS/MS analysis. While DEPC can also label lysine, serine, threonine, tyrosine, and cysteine residues, LC enables the separation of histidine-labeled peptides from the same peptides labeled at other residues. Moreover, MS/MS allows the histidine-labeled peptides to be readily distinguished from peptides labeled at other sites. Myoglobin is a good test protein because it has 11 histidine residues, and the tautomers for many of these histidines have been previously measured by NMR.⁴² In our LC-MS/MS experiments, we are able to measure the DEPC labeling of seven of these histidine residues, and LC is able to resolve the tautomers for most of these seven histidine residues. The ratios of the two tautomers are indicated in Table 1, along with a comparison to the previous NMR data when possible. The tautomer ratios for the corresponding free, acetylated peptides are also included, as they provide context for understanding how local sequence and three-dimensional structure influence the tautomer ratios. Measurements of these peptides also help confirm the identity of each tautomer, as the free peptides always generate products for both tautomers. This comparison is particularly important for some histidines in the proteins that exist in just in a single tautomeric state. Many of our tautomer ratios agree with the data from NMR. For example, His25 is primarily in the N^δ-H state (100%) according to previous NMR data,⁴² which matches our data showing that the N^δ-H tautomer has a much higher peak area than the N^ε-H one. Similarly, we measure a tautomer ratio of 3.8 ± 1.4 for His114, and by NMR the value is estimated to be 2.3. Interestingly, our labeling ratio for His120 is 22 ± 5, indicating the N^ε-H tautomer is primarily present, whereas the NMR data suggests 100% N^ε-H. Our ability to measure a low percentage of the N^δ-H tautomer likely reflects the greater precision of our measurements. We are also able to provide tautomer ratios for His65 and His98, which cannot be definitively measured by NMR. Our measured tautomer ratios are inconsistent with NMR data for His37 and His117. For His37, we measure a value of 28, indicating the N^ε-H tautomer is dominant, whereas the NMR data suggests the opposite. It should be noted, however, that the NMR assignment of the tautomeric state for this His residue is inferred based on ¹³C resonances rather than the more direct ¹⁵N resonances, which are not measured. It is possible that the inferred NMR data is incorrect. Another possibility is that this His residue in the sperm whale myoglobin studied by NMR⁴² has a different structure than this His residue in the equine myoglobin studied in our work, despite the 88% sequence homology of the two proteins. Our measured ratio for His117 is also somewhat different from the measured NMR data, but it should be noted that the NMR data for His117 shows significant exchange broadening, which reduces confidence in the measured ratios. Overall, our data shows reasonable agreement with previously obtained NMR data, suggesting that we can successfully measure the ratios of histidine tautomers in proteins via DEPC labeling and LC-MS/MS analysis.

Table 1.

Labeling ratios of the two histidine isomers in proteins and in their corresponding free tryptic peptides.

Protein	Residue	pKa1	Peak area Nε-H/Peak area Nδ-H (from LC-MS measurements)		Nε-H/Nδ-H Ratio measured by NMR
			Peptide	Protein
myoglobin	H25	5.5	5.6 ± 0.9	0.063 ± 0.044	0
	H37	6.3	4.7 ± 0.6	28 ± 6	0.25
	H65	4.3	1.5 ± 0.7	1.3 ± 0.6	-2
	H98	5.5	-3	-3	-2
	H114	5.6	1.1 ± 0.2	3.8 ± 1.4	2.3
	H117	6.9	6.1 ± 3.5	0.52 ± 0.2	1.54
	H120	5.2	4.7 ±1.1	22 ± 5	infinite
β2m	H13	6.2	1.6 ± 1.0	2.7 ± 1.3	-6
	H31	6.0	4.9 ± 1.6	9.2 ± 2.1
	H51	6.5	6.9 ± 1.8	05
	H84	6.3	5.2 ± 1.3	3.4 ± 1.3

¹The pK_a values for each His residue was calculated using the program PROPKA and the protein structure from MD simulations (see SI).

²The ratios of the two tautomers for this His residue were not measurable by NMR.⁴²

³These labeled peptide peaks for these two tautomers cannot be separated by LC.

⁴This value from NMR is estimated and therefore has lower confidence because of exchange broadening in the NMR experiment.⁴²

⁵Only the N^δ-H form is measured.

⁶There are no NMR results for the His tautomers of β2m.

We also used our labeling approach to determine the tautomer ratios for the protein β2m, which have not been measured by NMR. From our data, we find that the N^ε-H tautomer is favored for three (His13, His31, and His84) of the four histidine residues. For His51, we find that only the N^δ-H form is present in the protein, even though both tautomers are measured in the free peptide that contains this residue (Table 1). The reason for the prevalence of this tautomer could be due to the structure around His51. The N^δ is buried, according to MD simulations (Figure S7), forcing the N^ε to be the protonated site, which causes the N^δ-H form to be dominant. In effect, the ability to determine the tautomeric states of His residues in proteins provides more detailed structural information for proteins.

CONCLUSIONS

In this work, we have used DEPC labeling together with LC-MS/MS analysis to distinguish histidine tautomers in peptides and proteins. Using known preferences for the N^ε-H tautomer in free peptides, as well as peptide dissociation patterns during CID, we are able to distinguish the ratios of the tautomers from the extent of their reactivity with DEPC. In free peptides, the DEPC modification ratio of the two forms is close to 4:1 (N^ε-H:N^δ-H), although the exact value of this ratio is dependent on peptide sequence. Using this DEPC labeling strategy, we have determined the ratio of the two tautomers for several histidine residues in proteins, and where comparisons are possible, our measurements are consistent with previous 2D NMR data. In contrast to 2D NMR measurements, our DEPC labeling approach is simpler, faster, and is more precise. The ability to distinguish histidine tautomers via DEPC labeling indicates the sensitivity of this labeling chemistry to the local structure around the histidine residues in proteins, essentially providing high resolution structural information.