Mapping low-resolution three-dimensional protein structures using chemical cross-linking and Fourier transform ion-cyclotron resonance mass spectrometry

Authors

  • Gry H. Dihazi,

    1. Biotechnological-Biomedical Center, Faculty of Chemistry and Mineralogy, University of Leipzig, D-04103 Leipzig, Germany
    Search for more papers by this author
  • Andrea Sinz

    Corresponding author
    1. Biotechnological-Biomedical Center, Faculty of Chemistry and Mineralogy, University of Leipzig, D-04103 Leipzig, Germany
    • Biotechnological-Biomedical Center, Faculty of Chemistry and Mineralogy, University of Leipzig, Linnéstr. 3, D-04103 Leipzig, Germany.
    Search for more papers by this author

Abstract

Techniques in mass spectrometry (MS) combined with chemical cross-linking have proven to be efficient tools for the rapid determination of low-resolution three-dimensional (3-D) structures of proteins. The general procedure involves chemical cross-linking of a protein followed by enzymatic digestion and MS analysis of the resulting peptide mixture. These experiments are generally fast and do not require large quantities of protein. However, the large number of peptide species created from the digestion of cross-linked proteins makes it difficult to identify relevant intermolecular cross-linked peptides from MS data. We present a method for mapping low-resolution 3-D protein structures by combining chemical cross-linking with high-resolution FTICR (Fourier transform ion-cyclotron resonance) mass spectrometry using cytochrome c and hen egg lysozyme as model proteins. We applied several homo-bifunctional, amine-reactive cross-linking reagents that bridge distances from 6 to 16 Å. The non-digested cross-linking reaction mixtures were monitored by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS) to determine the extent of cross-linking. Enzymatically digested reaction mixtures were separated by nano-high-performance liquid chromatography (nano-HPLC) on reverse-phase columns applying water/acetonitrile gradients with flow rates of 200 nL/min. The nano-HPLC system was directly coupled to an FTICR mass spectrometer equipped with a nano-ESI (electrospray ionization) source. Cross-linking products were identified using a combination of the GPMAW software and ExPASy Proteomics tools. For correct assignment of the cross-linking products the key factor is to rely on a mass spectrometric method providing both high resolution and high mass accuracy, such as FTICRMS. By combining chemical cross-linking with FTICRMS we were able to rapidly define several intramolecular constraints for cytochrome c and lysozyme. Copyright © 2003 John Wiley & Sons, Ltd.

In recent years, the number of identified proteins has increased drastically through genomic and proteomic approaches. These rapid developments require fast methods for three-dimensional (3-D) structural elucidation of these proteins, currently primarily X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy. Both techniques yield detailed information on the protein structure; however, NMR spectroscopy requires large quantities of pure protein (10–100 mg) in a specific solvent, and for X-ray studies the protein must be crystallized. Thus, elucidation of a specific protein structure by these techniques may take months or even years.1,2

The combination of chemical cross-linking and mass spectrometry could offer an alternative approach for rapidly mapping amino acid distances within a protein. Previous studies have shown that chemical cross-linking can yield low-resolution structural information about the constraints within a molecule.3,4 Mass spectrometry is the method of choice for these studies because of its high sensitivity,5,6 enabling rapid analysis of the complex mixtures obtained from enzymatic digests of cross-linking reaction mixtures.7 The combination of chemical cross-linking, enzymatic digestion, and subsequent mass spectrometric analysis, for elucidation of 3-D protein structures, offers distinct advantages over traditional methods such as X-ray crystallography and NMR spectroscopy.8 With this method, analysis of a protein structure can be performed with minute amounts of proteins within a relatively short time.9 In contrast to NMR spectroscopy, the size of the protein does not present a problem, as the enzymatic digestion products are analyzed. In addition, the protein does not have to be crystallized, as is the case for X-ray studies, and the great variety of commercially available cross-linking reagents with different lengths and specificities offers a high degree of flexibility for experimental design.

However, despite using mass spectrometry as an analysis tool, identification of cross-linking products can be hampered by the complexity of the cross-linking reaction mixture. This problem has been partially addressed by using isotope-labeled10 or fluorescently labeled11 cross-linking reagents in order to facilitate identification of the cross-linked products. Another approach using chemical cross-linking and mass spectrometry was applied to the analysis of intermolecular interfaces by labeling the interacting proteins with 14N/15N.12 Recently, a ‘top-down approach’ was presented, in which the crude cross-linked protein mixture is injected into an ESI-FTICR mass spectrometer and the cross-link positions are localized by multiple stages of fragmentation in the mass spectrometer.13

In this work, we describe a method for the rapid determination of 3-D protein structures by combining chemical cross-linking with high-resolution FTICR mass spectrometry, using the structurally well-characterized proteins cytochrome c and hen egg lysozyme as model proteins. With mass accuracies of lower than 3 ppm for nano-HPLC/nano-ESI-FTICRMS analysis,14 this approach offers the possibility of identifying the cross-linking products by their exact masses alone, without the need to apply labeled cross-linking reagents or proteins.

EXPERIMENTAL

Chemicals

Chemicals were purchased from Sigma (Taufkirchen, Germany) or VWR (Darmstadt, Germany) at the highest purity available. Cross-linking reagents, bis(sulfosuccinimidyl) suberate (BS3), ethylene glycolbis(sulfosuccinimidyl succinate) (sulfo-EGS), disulfosuccinimidyl tartarate (sulfo-DST), dimethyl adipimidate hydrochloride (DMA), and dimethyl suberimidate hydrochloride (DMS), were obtained from Pierce (Rockford, IL, USA). MALDI matrices and peptides for MS calibration were purchased from Sigma (Taufkirchen, Germany). Nano-HPLC solvents were spectroscopic grade (Uvasol®, VWR). Water was purified with a Direct-Q5 water purification system (Millipore, Eschborn, Germany).

Cross-linking reactions

Horse heart cytochrome c and hen egg lysozyme were purchased from Sigma and used without further purification. Cytochrome c or lysozyme was dissolved at a concentration of 3 mg/mL in 20 mM Hepes buffer, pH 7.7. The protein stock solution was diluted with 20 mM Hepes buffer to a protein concentration of 5 μM. 2–20 μL of a freshly prepared solution containing 10 mM sulfo-DST, BS3 or sulfo-EGS in dimethyl sulfoxide (DMSO) were added to the protein solution in a 10- to 100-fold molar excess over the protein to give a final volume of 400 μL. A control sample without cross-linking reagent was prepared in the same fashion. For cross-linking reactions with DMA or DMS (in 20 mM Hepes buffer, pH 7.7), the cross-linking reagent was added in a 200-fold molar excess over the protein. The samples were incubated at room temperature. Aliquots of 100 μL were taken after 15, 30, 60 and 120 min, and the reaction was terminated by adding 5 μL of a 400 mM NH4HCO3 solution.

In-solution digestion

For in-solution digestion, a solution containing 5 μL 8 M urea and 20 μL 400 mM NH4HCO3 was added to 20 μL of the cross-linking reaction mixture or the non-reacted protein solution, respectively. In the case of lysozyme samples, 5 μL of a 45 mM dithiothreitol (DTT) solution were added to the samples, and the mixtures were incubated for 15 min at 50°C. Subsequently, 5 μL of a 100 mM iodoacetamide solution were added before the mixture was incubated at room temperature for 15 min. To both cytochrome c and lysozyme samples, 60 μL of water were added before the samples were treated with 0.5 μL of a 2 μM (50 ng/μL) solution of trypsin (sequencing-grade, Roche Diagnostics, Mannheim, Germany) in 50 mM NH4HCO3 and/or 0.7 μL of a 1.5 μM (40 ng/μL) solution of endoproteinase AspN (sequencing-grade, Roche Diagnostics) in 10 mM Tris-HCl (corresponding to enzyme/substrate ratios of 1:50); digestion was performed at 37°C for 16 h. The reaction was stopped by freezing the samples at −20°C.

MALDI-TOFMS

MALDI-TOFMS measurements were performed using a Voyager-DE RP Biospectrometry Workstation (Applied Biosystems, Foster City, CA, USA) equipped with a nitrogen laser (337 nm). Data acquisition and data processing were performed using the Voyager software, version 5.1, and the Data Explorer software, version 4.0 (Applied Biosystems). All measurements were performed in positive ionization mode. Peptides were measured in reflectron mode using α-cyano-4-hydroxycinnamic acid (HCCA) as matrix, whereas proteins were measured in linear mode using sinapinic acid as matrix. The matrix was prepared in 30% acetonitrile, 70% water, and 0.1% trifluoroacetic acid (TFA). Samples were prepared using the dried droplet method, and spectra were obtained by summation of 50–100 laser shots. For measurements in reflectron mode, spectra were recorded over the range m/z 500–4000; the instrument was calibrated using [M+H]+monoisotopic ions of angiotensin I (m/z 1296.69), substance P (m/z 1347.74), and somatostatin (reduced form, m/z 1637.72). For measurements in linear mode, spectra were recorded over the range m/z 1000–30 000; calibration was performed using [M+H]+average values for cytochrome c (m/z 12361.1), lysozyme (m/z 14305.1), and myoglobin (m/z 16952.6). Before MALDI-TOFMS, the protein samples were desalted using ZipTips C4 or Microcon filtration units YM-10 (both from Millipore).

Nano-HPLC

Nano-HPLC was performed using an Ultimate Nano-LC system (LC Packings, Amsterdam, The Netherlands) equipped with a Switchos II column switching module and a Famos micro autosampler with 5-μL sample loop. Samples were injected by the autosampler and concentrated on a trapping column (PepMap, C18, 300 μm × 1 mm, 5 μm, 100 Å, LC Packings) with water containing 0.1% formic acid at flow rates of 20 μL/min. For sample volumes larger than 5 μL, the injection procedure was repeated up to three times, thus applying sample volumes of up to 20 μL. After 2 min, the peptides were loaded onto the separation column (PepMap, C18, 75 μm × 150 mm, 3 μm, 100 Å, LC Packings) which had been equilibrated with 95% A (water + 0.1% formic acid). Peptides were eluted using the following gradient: 0–30 min, 5–50% B; 30–31 min, 50–95% B; 31–35 min, 95% B (where B = acetonitrile + 0.1% formic acid) at flow rates of 200 nL/min, and detected by their UV absorption at 214 and 280 nm.

Nano-ESI-FTICRMS

The nano-HPLC system was on-line coupled to an Apex II FTICR mass spectrometer with 7 Tesla superconducting magnet (Bruker Daltonics, Billerica, MA, USA) equipped with a nano-ESI source (Agilent Technologies, Waldbronn, Germany). For nano-ESI, coated fused-silica PicoTips (tip i.d. 8 μm, New Objective, Woburn, MA, USA) were applied. The capillary voltage was set to −1400 V. Mass spectral data were acquired over the m/z range 200–2000, four scans were combined into one spectrum, and 400 spectra were recorded for each LC/MS run. Data were acquired over a time of 34.5 min. Calibration of the instrument was performed with CID fragments (capillary exit voltage 200 V) of the LHRH peptide, with m/z values calculated as: y5, 499.2987; b4, 522.2096; y6, 662.3620; y7, 749.3941; b7, 855.3784; y8, 935.4734; b8, 1011.4795, and also with the intact LHRH [M+H]+monoisotopic at m/z 1183.5643 and [M+H]2+monoisotopic at m/z 592.2358. Data acquisition and processing were performed using the XMASS software, version 5.0.10 (Bruker Daltonics). MS data acquisition was started with a trigger signal from the HPLC system 5 min after starting the LC gradient. Processing of the raw data was performed using the ‘Projection’ tool in the XMASS software.15 Identification of the digestion products was performed using the ExPASy Proteomics tool ‘PeptideMass’ in the Swiss-Prot Database16 and the Biotools software package, version 2.0 (Bruker Daltonics).

Analysis of cross-linking products

The identification of cross-linking products was performed with the GPMAW software, version 5.12 beta 3 (Lighthouse Data, Odense, Denmark) and the ExPASy Proteomics tool ‘FindMod’ in the Swiss-Prot Database.16

RESULTS AND DISCUSSION

The basis for mapping low-resolution 3-D structures of cytochrome c and lysozyme by chemical cross-linking and mass spectrometry was a preceding detailed structural characterization of these proteins. Analyses of the primary protein structures and post-translational modifications were performed by mass spectrometric peptide fingerprinting and by exact mass measurements of the intact proteins by FTICRMS. ESI-FTICRMS measurement of the exact mass of cytochrome c yielded a base peak at m/z 12359.420 after deconvolution, which corresponds well to the theoretical value of m/z 12359.393 for the acetylated, heme-bound form of cytochrome c. For lysozyme, a base peak at m/z 14304.972 was determined to be in good agreement with the theoretical value of m/z 14304.818 for the molecular mass of the reduced form of the protein in which the asparagine residue in position 103 is exchanged for aspartic acid.17 Peptide mass fingerprinting confirmed this amino acid exchange. In addition, ESI-FTICRMS of intact lysozyme yielded a weak signal at m/z 14467.046, indicating a glycosylated form of lysozyme (data not shown).18 Amino acid sequences of lysozyme and cytochrome c are presented in Fig. 1.

Figure 1.

Amino acid sequences of (A) cytochrome c and (B) lysozyme.

General strategy for determining 3-D protein structures

The analytical strategy for determining the 3-D protein structures by chemical cross-linking and mass spectrometry is depicted in Fig. 2. The proteins were incubated with different amine-reactive, homo-bifunctional cross-linking reagents (the NHS esters sulfo-DST, BS3, sulfo-EGS, and the imidoesters DMA and DMS) at room temperature for different times. An overview of the cross-linking reagents used in this work, as well as their respective spacer lengths, is given in Table 1. After the cross-linking reaction, the reaction mixture was analyzed by MALDI-TOFMS in order to check the extent of cross-linking products and thus to optimize the reaction conditions. For analysis of the cross-linking products the reaction mixture was digested enzymatically in-solution and the digestion products were analyzed by nano-HPLC/nano-ESI-FTICRMS. Identification of the cross-linking products was performed using the GPMAW software and the ExPASy Proteomics tools, considering only signals occurring exclusively in samples treated with cross-linking reagents.

Figure 2.

General strategy for mapping low-resolution three-dimensional structures of proteins by chemical cross-linking and mass spectrometry.

Table 1. Chemical structures and spacer lengths of the amine-reactive, homo-bifunctional cross-linking reagents
NameChemical structureSpacer
Sulfo-DSTDisulfosuccinimidyl tartarate
original image
6.4 Å
BS3Bis(sulfosuccinimidyl) suberate
original image
11.4 Å
Sulfo-EGSEthylene glycolbis(sulfosuccinimidyl succinate)
original image
16.1 Å
DMADimethyl adipimidate × 2 HCl
original image
8.6 Å
DMSDimethyl suberimidate × 2 HCl
original image
11.0 Å

Calculation of distances of amino groups

In order to estimate the extent of intramolecular cross-linking in the proteins used for our studies, the distances of ε-amino groups of lysines and the free N-terminus (only in lysozyme) were calculated for cytochrome c and lysozyme from different entries in the Protein Data Bank.19 For cytochrome c an X-ray structure (pdb entry: 1HRC) and an NMR structure (pdb entry: 1OCD) were used as the basis for structural data. X-ray data describe a crystallized form of the protein, thus yielding structural information on the most stable conformation of the protein, whereas NMR data represent the solution structure. As proteins can adopt different conformations in the crystal and in solution, protein structures and thus calculated distances of ε-amino groups can vary between the NMR and X-ray data. For lysozyme, an X-ray structure (pdb entry: 1HEL) was used to calculate the distances between the ε-amino groups of lysines and the free N-terminus, respectively, using the RasMol 2.7 program.20 We considered all amino groups located within distances of up to 20 Å to be suitable for cross-linking, as the longest cross-linking reagent used in this study (sulfo-EGS) possesses a spacer length of 16.1 Å. From the NMR and X-ray structures of cytochrome c we detected 51 and 59 possibilities, respectively, for amino groups that could possibly form cross-linking products. For lysozyme, a calculation of the distances between the ε-amino groups of lysines and the free N-terminus yielded seven possible cross-linking combinations.

MALDI-TOFMS

MALDI-TOFMS served as a rapid method to establish the optimum cross-linking reaction conditions for the different cross-linking reagents. When using amine-reactive cross-linking reagents, special care has to be taken not to disturb the 3-D structure of the proteins due to the loss of positive charge at the lysines. On the other hand, sufficient amounts of cross-linking products have to be created to allow mass spectrometric detection.

For the NHS esters, the optimum reaction conditions were determined using cytochrome c and BS3, which was applied in a 10-, 20- 50, and 100-fold molar excess over the protein. The reaction was conducted at room temperature, aliquots were taken after 15, 30, 60, and 120 min, and after 20 h, and the cross-linking reaction mixtures were analyzed by MALDI-TOFMS. Not surprisingly, the extent of cross-linking product formation increased with growing amounts of cross-linking reagent (Fig. 3(A)) as well as with longer incubation times (Fig. 3(B)). The peak spacings in the mass spectra exhibit a mass difference of 138 u corresponding well to a theoretical mass increase of 138.068 u caused by the reacted cross-linker BS3. Interestingly, the peaks became broader with increasing incubation time, which might indicate an increased formation of cross-linking products with hydrolyzed cross-linking reagent. After incubating cytochrome c with a 10-fold molar excess of BS3 for 30 min, one to three cross-linking molecules had reacted with the protein, while, with a 100-fold molar excess of BS3, four to six cross-linking molecules had attached to cytochrome c. Using a 20-fold molar excess of BS3, two to four cross-linking molecules had reacted with the protein after a reaction time of 15 min, while, after 120 min reaction time, three to five cross-linker molecules had reacted with the protein.

Figure 3.

MALDI-TOFMS analysis of the reaction mixture of cytochrome c (solid line) and BS3. XL: cross-linking reagent. (A) Incubation time 30 min. Molar excess of BS3: dashed line, 10-fold; bold line, 20-fold; dotted line, 50-fold; grey line, 100-fold. (B) 20-fold molar excess of BS3. Dashed line, incubation time 15 min; bold line, incubation time 30 min; dotted line, incubation time 60 min; grey line, incubation time 120 min.

The modification of lysozyme with BS3 showed a similar time-dependence with respect to the extent of cross-linking product formation. Using a 20-fold excess of BS3, two to five cross-linking products had been formed after an incubation time of 15 min, while, after 120 min, three to six BS3 molecules had reacted with lysozyme (data not shown).

For the cross-linking experiments with the NHS esters sulfo-DST and sulfo-EGS, we applied the reaction conditions established with BS3, as all NHS esters exhibited the same reaction pattern. Thus, cross-linking reagents were added in a 20-fold molar excess over the proteins. For sulfo-EGS, two to six molecules of cross-linker reacted with lysozyme as well as with cytochrome c depending on the reaction time (15–120 min). In contrast, sulfo-DST did not show any time-dependence; at incubation times of 15, 30, 60 and 120 min, between one and four cross-linking molecules had reacted with cytochrome c, while for lysozyme the formation of only a cross-linking product with one sulfo-DST molecule was detected (data not shown). This difference in the number of cross-linking products created may be explained by the number of lysine residues in the proteins as well as by the distances between their amino groups. Cytochrome c possesses 19 lysines, for which six combinations (Lys 5-Lys 8, Lys 7-Lys 8, Lys 22-Lys 27, Lys 25-Lys 27, Lys 73-Lys 86, and Lys 99-Lys 100) of ε-amino groups in the NMR structure, and eight combinations (Lys 5- Lys 8, Lys 7- Lys 100, Lys 13- Lys 86, Lys 13- Lys 87, Lys 25- Lys 27, Lys 39- Lys 60, Lys 53- Lys 79, and Lys 86-Lys 87) of ε-amino groups in the X-ray structure, are within distances of 10 Å and could theoretically be cross-linked by sulfo-DST bridging a distance of 6.4 Å. Lysozyme, on the other hand, possesses the free N-terminus and six lysines, for which two combinations, N-terminus-Lys 1 and Lys 96-Lys 97, are within a distance of under 10 Å.

The reaction patterns of the imidoesters DMA and DMS (Table 1) are different from that of the NHS esters. As imidoesters react more slowly than the NHS esters, competition reactions due to partial hydrolysis of cross-linking reagents are observed more frequently for imidoesters than for NHS esters. Therefore, a higher molar excess of cross-linking reagent was applied for DMA and DMS in comparison with sulfo-DST, BS3, and sulfo-EGS. However, we detected almost no cross-linking product formation for both cytochrome c and lysozyme.

Identification of cross-linking products

After the cross-linking reaction and subsequent analysis of the reaction mixtures by MALDI-TOFMS, the cross-linking reaction mixtures were enzymatically digested in-solution (Fig. 2) with trypsin or endoproteinase AspN. Additionally, the cross-linking mixtures were treated with a combination of both proteases in order to keep the size of the created cross-linking products in the optimum ESI-FTICRMS detection range of lower than m/z 2000. When applying cross-linking reagents for determination of the 3-D structures of proteins, only intramolecular cross-linking products are of interest. As high molecular weight aggregates can be formed by intermolecular cross-linking reactions, it is especially important to control the reaction conditions in order to suppress formation of intermolecular cross-linking products. For detection of intermolecular cross-linking products, the cross-linking reaction mixtures were subjected to 1-D gel electrophoresis (SDS-PAGE). For protein concentrations of 5 μM, which were applied in our experiments, we did not observe any high molecular weight aggregates.

After enzymatic digestion, the digestion mixtures were analyzed by nano-HPLC/nano-ESI-FTICRMS.14 Figure 4 shows the 2-D plot of an LC/MS separation of the reaction mixture of cytochrome c with sulfo-EGS, and the corresponding deconvoluted mass spectrum is presented in Fig. 5. The mass spectrum in the 2-D plot represents the mass-to-charge ratios of multiply charged ions of three cross-linking products created after the reaction of cytochrome c with sulfo-EGS (Fig. 4). The cross-linking products comprising amino acids 80–88 and 39–55 with one sulfo-EGS molecule attached are both present as doubly charged ions, while the cross-linking product comprising amino acids 23–38 with one sulfo-EGS molecule attached is present as a doubly and triply charged ion (Fig. 4). The peptide consisting of amino acids 80–88 contains three lysines at positions 86, 87, and 88. As trypsin cleaves only with a low probability at modified lysines, only one possible cross-linking product between Lys 86 and Lys 87 can be derived. The peptide fragment comprising amino acids 39–55 contains three lysines at positions 39, 53, and 55, and because of the reasons mentioned above the cross-linking product can only be formed between Lys 39 and Lys 53. As both cross-linking products exhibit a free Lys residue in addition to the N-terminus, which can bear a positive charge, both peptides are present as doubly charged ions. The peptide fragment consisting of amino acids 23–38 contains two lysines in positions 25 and 27; therefore, only one possibility exists for the cross-linking reaction. As this peptide contains two histidine residues in addition to the free N-terminus, up to three protons can be attached; therefore, this cross-linking product is present as doubly and triply charged ions.

Figure 4.

Two-dimensional plot (ESI mass spectrum) and total ion current (TIC) of a nano-HPLC/nano-ESI-FTICRMS analysis of a tryptic digest of cytochrome c cross-linked with sulfo-EGS.

Figure 5.

Deconvoluted ESI-FTICR mass spectrum of a tryptic digest of cytochrome c cross-linked with sulfo-EGS. The three detected cross-linking products of cytochrome c with sulfo-EGS are marked with a square. The signal at m/z 2066.964 comprising amino acids 39–55, cross-linked with one molecule sulfo-EGS, is shown enlarged.

For the cross-linking reaction of lysozyme with BS3, the signal of the doubly charged ion of the peptide comprising amino acids 87–100 with one BS3 molecule attached was detected. This peptide contains two lysines in positions 96 and 97, thus leaving only one possibility for cross-linking. Another cross-linking product of lysozyme and BS3 was detected comprising amino acids 1–5, which contains the lysine in position 1 in addition to the free N-terminus, and again there is only one possibility for the cross-linking reaction.

A summary of the identified cross-linking products is shown in Tables 2 and 3 for cytochrome c and lysozyme, respectively. For cytochrome c, cross-linking products were identified with all three NHS esters, BS3, sulfo-DST, and sulfo-EGS. Four different cross-linking products were identified, formed by the cross-linking of Lys 25 and Lys 27, Lys 39 and Lys 53, Lys 86 and Lys 87, and Lys 99 and Lys 100. The cross-linking product between Lys 99 and Lys 100 was only observed in experiments in which both trypsin and endoproteinase AspN were the digestion enzymes. This may be a result of the size of the peptide fragment obtained after tryptic cleavage: after digestion with trypsin alone a peptide fragment comprising amino acids 89–104 is created, which possesses a mass of over 2000 u without the possibility for multiple protonations, as the only basic amino acids in this peptide, Lys 99 and Lys 100, are both modified with the cross-linking reagent. Therefore, such a peptide would be very difficult to detect in the present ESI-FTICRMS measurements.

Table 2. Identified cross-linking products of cytochrome c with the NHS esters sulfo-DST, BS3, and sulfo-EGS
Cross-linking reagentCross-linking productDistance (Å) X-ray/NMRObserved mass [M+H]+Sequence (amino acids)
  1. XL = cross-linking reagent.

Sulfo-DSTK25–K275.26/8.171789.90023–38 + XL
 K86–K874.72/10.521149.63680–88 + XL
 K99–K10013.56/7.741492.75292–103/93–104 + XL
BS3K25–K275.26/8.171813.98423–38 + XL
   1942.07923–39 + XL
 K86–K874.72/10.521173.70780–88 + XL
 K99–K10013.56/7.741516.83092–103/93–104 + XL
Sulfo-EGSK25–K275.26/8.171901.96323–38 + XL
 K39–K5315.62/16.402066.96439–55 + XL
 K86–K874.72/10.521261.68580–88 + XL
 K99–K10013.56/7.741604.80892–103/93–104 + XL
Table 3. Identified cross-linking products of lysozyme with the NHS ester BS3
Cross-linking reagentCross-linking productDistance (Å) X-rayObserved mass [M+H]+Sequence (amino acids)
  1. XL = cross-linking reagent.

BS3N-Terminus-K17.44744.4421–5 + XL
 K96–K979.431643.86887–100 + XL

In the case of cytochrome c, identical cross-linking products were identified for cross-linking experiments with the three NHS esters used in our experiments. The only exception was the cross-linking product between Lys 39 and Lys 53, which was detected exclusively in the cross-linking reaction with sulfo-EGS, bridging the longest distance of all cross-linking reagents used for our experiments (spacer length 16.1 Å). This finding is in good agreement with the theoretical distance between Lys 39 and Lys 53 (NMR structure: 16.4 Å, X-ray structure: 15.6 Å). Additionally, we observed that very short distances, e.g., between the ϵ-amino groups of Lys 25 and Lys 27 (NMR structure: 5.3 Å, X-ray structure: 8.2 Å), are bridged by all three NHS esters (Table 2), indicating a high degree of flexibility between these two lysine residues. The cross-links between Lys 25 and Lys 27 and between Lys 39 and Lys 53 had also been detected in previous cross-linking experiments of cytochrome c with the isotopically labeled, amine-specific cross-linking reagent disuccinimidyl adipate (DSA).10

For lysozyme, however, a completely different picture was obtained. Cross-linking reactions with sulfo-DST and sulfo-EGS yielded no cross-linking products, while the cross-linking reaction with BS3 gave two cross-linking products (Table 3). In the two products formed by cross-linking between the N-terminus and Lys 1, and Lys 96/Lys 97, respectively, the amino groups are within distances of 7.4 and 9.4 Å (X-ray structure), which is in good agreement with a spacer length of 11.4 Å for BS3. However, we also expected to find cross-linking products for sulfo-DST in addition to BS3. One explanation for the lack of cross-linking product formation of sulfo-DST with the N-terminus and the ε-amino group of lysine 1 could be that both reactive centers of the N-terminal amino acid are relatively rigid and therefore cannot be cross-linked by a short reagent, such as sulfo-DST with a spacer length of 6.4 Å. In the case of the ε-amino groups of Lys 96 and Lys 97, a cross-linking reaction by sulfo-DST would be more probable. It is possible, however, that lysozyme adopts different conformations in solution than the one perceived by the static X-ray structure, causing the ε-amino groups of the lysines in positions 96 and 97 to be too far apart from one another in order to be cross-linked by sulfo-DST. Interestingly, the BS3 cross-linking product of lysozyme between Lys 96 and Lys 97 was exclusively detected in reaction mixtures digested with endoproteinase AspN and trypsin, analogous to the cross-linking product in cytochrome c between Lys 99 and Lys 100.

Table 4 shows more identified possible cross-linking products. With the cross-linking reagent sulfo-DST, a peptide fragment comprising amino acids 1–13 (acetylated N-terminus) and one sulfo-DST molecule was identified in addition to the unambiguously identified cross-linking products. As this peptide contains lysines in positions 5, 7, 8 and 13, there are several possibilities for cross-linking products. Cross-linking of Lys 13 is excluded as trypsin rarely cleaves after modified lysines, as mentioned above. This demonstrates that in this case the high mass accuracy obtained in FTICRMS is not sufficient to locate the actual cross-linking sites, but that tandem mass spectrometry (MS/MS) experiments are needed. With sulfo-EGS, we detected a cross-linking product between Lys 5 and Lys 88; however, as the signals corresponding to this cross-linking product were only of weak intensity, we considered it to be only tentatively identified.

Table 4. Possible cross-linking products of cytochrome c with the NHS esters sulfo-DST and sulfo-EGS
Cross-linking reagentCross-linking productsDistance (Å) X-ray/NMRObserved mass [M+H]+Sequences (amino acids)
  1. XL = cross-linking reagent.

Sulfo-DSTK5–K7 or13.95/13.22  
 K5–K8 or9.36/9.091631.8631–13(acet) + XL
 K7–K810.81/8.88  
Sulfo-EGSK5–K8816.51/17.021219.61688–91 + 4–7 + XL
   1532.74488–91 + 1–7 + XL

Tables 5 and 6 show a summary of identified peptide fragments that were modified by hydrolyzed cross-linking reagents. For lysozyme, reaction products with the hydrolyzed cross-linking reagents sulfo-DST and BS3 were detected for four out of six lysines (Lys 1, Lys 13, Lys 97, and Lys 116) (Table 6), while for the lysines in positions 33 and 96 no modifications with hydrolyzed cross-linking reagents were found. For the cross-linking reactions of cytochrome c with BS3, several modifications with hydrolyzed cross-linking reagent were identified (Table 5). As cytochrome c is a lysine-rich protein with a total of 19 lysine residues, modified peptides can contain several lysine residues making it difficult to unambiguously assign the modified lysine residue.

Table 5. Peptides of cytochrome c modified with hydrolyzed cross-linking reagent (XL-OH)
Cross-linking reagentModified lysineObserved mass [M+H]+Sequence (amino acids)
Sulfo-DSTK391302.59039–49 + XL-OH
 K53 or K552009.94050–65 + XL-OH
 K87, K88 or K991867.01787–100 + XL-OH
  1738.92087–99/88–100 + XL-OH
 K99 or K1001510.76092–103/93–104 + XL-OH
BS3K391326.65834–49 + XL-OH
 K39 or K531996.99739–55 + XL-OH
 K721795.84861–73 + XL-OH
 K99 or K1001534.83892–103/93–104 + XL-OH
Sulfo-EGSK5, K7 or K81906.9532–13 + XL-OH
 K99 or K1001622.82592–103/93–104 + XL-OH
Table 6. Peptides of lysozyme modified with hydrolyzed cross-linking reagent (XL-OH)
Cross-linking reagentModified lysineObserved mass [M+H]+Sequence (amino acids)
Sulfo-DSTK1738.3811–5 + XL-OH
 K131181.5346–14 + XL-OH
 K971936.88697–112 + XL-OH
 K1161465.674115–125 + XL-OH
 K1161735.818113–125 + XL-OH
BS3K1762.4511–5 + XL-OH
 K131205.6036–14 + XL-OH
 K971960.94997–112 + XL-OH
 K1161489.747115–125 + XL-OH
Sulfo-EGSK11094.4801–5 + XL-OH

CONCLUSIONS

In this work we have demonstrated that a combination of chemical cross-linking with high-resolution FTICR mass spectrometry provides a promising method for rapid mapping of low-resolution three-dimensional protein structures. The greatest limitation for a broad applicability of this technique is currently the lack of suitable software programs for the analysis of the cross-linking products. At present, deficits in commercially available computer software persist in the calculation of intrapeptidal cross-linking products as well as in the calculation of cross-linking products obtained after enzymatic digestion with a mixture of proteases. For a thorough characterization of the three-dimensional structures of proteins, the application of different cross-linking reagents with different lengths and various specificities is needed. For the present work, we applied exclusively amine-reactive, homo-bifunctional cross-linking reagents. Our experiments revealed that, for correct assignment of cross-linking products, the use of a mass spectrometric method providing high mass measurement accuracy and high mass resolution, such as FTICR mass spectrometry, is an absolute requirement. By combining chemical cross-linking with FTICRMS, we were able to rapidly define several intramolecular constraints for both cytochrome c and lysozyme. Our experiments clearly indicate, however, that further work is needed to evaluate different cross-linking reagents with different specificities and to improve existing computer software to automate data analysis.

Acknowledgements

The authors thank Dr. Marcus Bantscheff for critical reading of the manuscript and valuable suggestions. The junior research group of A. S. is funded by the Saxon State Ministry of Higher Education, Research and Culture.

Ancillary