INTRODUCTION
Protein–protein interactions (PPIs) are essential for the proper functions of proteins (Laet al. 2013). The strength of a PPI is denoted by the equilibrium dissociation constant (KD), which is equal to koff/kon, with koff and kon being the kinetic dissociation/association rate constants, respectively (Duet al. 2016). The window of biologically relevant KD values is extremely wide, spanning over 12 orders of magnitude. The PPIs are often classified as transient or stable based on their binding affinities (Perkinset al. 2010). Characterizing transient PPIs has been technically challenging (Perkinset al. 2010; Qin and Gronenborn 2014). Among the existing techniques, NMR is an exquisite tool for the examination of transient PPIs under physiological conditions (Liuet al. 2016; Vaynberg and Qin 2006; Xinget al. 2014), but an effective analysis can be hampered by the complexity of NMR spectra that increases with protein size. Other methods such as yeast two-hybrid analysis (Berggardet al. 2007), phage display (Smith 1985), affinity-based pull-down (Berggardet al. 2007), and fluorescence titration (Acuner Ozbabacanet al. 2011) are more successful when applied to the stable interactions than to the transient ones, although they do not provide direct structural information about the binding interfaces.
Chemical cross-linking of proteins coupled with mass spectrometry analysis (abbreviated as CXMS, XL-MS, or CLMS) has become a valuable tool for investigating the structures of protein complexes and PPIs in recent years (Chavez and Bruce 2019; Liu and Heck 2015; O'Reilly and Rappsilber 2018; Wheatet al. 2021; Yanget al. 2012; Yu and Huang 2018). In general, two residues spatially close to each other can be covalently linked by chemical cross-linkers under mild conditions. Following protease digestion, cross-linked peptide pairs can be identified by tandem mass spectrometry, which then enables the positioning of cross-linked residues in the protein sequences (Herzoget al. 2012; Lvet al. 2020; Wuet al. 2017; Zhaoet al. 2018, 2019). Both the intra- and inter-protein interaction regions can be further identified by imposing a distance restraint between each pair of cross-linked residues that is no more than the maximal allowable distance of a cross-linker (Tang and Gong 2020). Additionally, CXMS has the advantage of being fast and sensitive and requiring no sample purification to homogeneity (Fanet al. 2014; Yu and Huang 2018).
Two interacting proteins can form an ensemble of on-pathway or off-pathway encounter complexes (ECs), which undergo rapid exchange with the stereospecific complex (SC). The SC is mediated by characteristic short-range specific interactions, typically involving hydrogen bonding, salt bridges, hydrophobic stacking, and π-stacking interactions (Fawziet al. 2010; Kozakovet al. 2014). In comparison, ECs are formed mainly through nonspecific electrostatic interactions, and are short-lived, lowly populated, and highly heterogeneous (Anthis and Clore 2015; Schilder and Ubbink 2013; Tanget al. 2006). Yet by forming the ECs, two binding partners are kept close to each other, facilitating the formation of the SC (Donget al. 2022; Tanget al. 2006). Conventional chemical cross-linkers can capture transient PPIs, including both ECs and SC. For example, studies have succeeded in using disuccinimidyl suberate (DSS) and glutaraldehyde to fix transiently formed complexes before affinity purification (Shiet al. 2015) or single particle cryo-electron microscopy (Kastneret al. 2008).
Cross-linking products are also determined by the physiochemical properties of a chemical cross-linker. These properties include the specificity of the reactive groups of a cross-linker (Belsom and Rappsilber 2021), and length, rigidity, and hydrophobicity/hydrophilicity of the spacer arm (Dinget al. 2016; Hofmannet al. 2015; Yuet al. 2020). Also important is the preferred conformation(s) of a cross-linker, which is related to the chemical composition of the spacer arm and local dielectric constant at the nearby protein surface (Gonget al. 2020). Furthermore, it has been tentatively suggested that the reaction kinetics of the reactive group of a cross-linker governs the capturing of a transient versus stable PPI (Belsom and Rappsilber 2021; Yanget al. 2018; Ziemianowiczet al. 2019), but this remains to be investigated in a formal way.
We recently reported a class of non-hydrolyzable amine-selective di-ortho-phthalaldehyde (DOPA) cross-linkers, one of which is DOPA2 (Wanget al. 2022). The maximally allowed Cα–Cα distance of DOPA2 is 30.2 Å, slightly longer than that of the widely used N-hydroxysuccinimide (NHS) ester cross-linker DSS (24.0 Å) (Fig. 1A ). Importantly, cross-linking of proteins by DOPA2 is 60–120 times faster than that by DSS. We then asked whether DOPA2 outperforms DSS in capturing transient PPIs. In the current study, we compared the cross-linking effects of these two cross-linkers on two heterodimeric complexes. Unexpectedly, we found that whether or not DOPA2 outperforms DSS depends on the nature of a protein complex. The SC was better depicted by the DOPA2 cross-links whereas the more heterogenous and short-lived ECs were better depicted by the DSS cross-links.
1 Performance of DOPA2 and DSS on weak protein complexes. A The chemical structures of DOPA2 and DSS. B Conceptual diagram of carbohydrate transport and phosphorylation by the phosphotransferase system. C,D SDS-PAGE of DOPA2 or DSS-cross-linked EIN/HPr, EIIAGlc/EIIBGlc complexes, with the protein concentrations comparable to the respective equilibrium dissociation constant. Heterodimer bands with the expected molecular weights were marked by square brackets. Cross-linking of EIN/HPr with 0.8 mmol/L DOPA2 resulted in high Mw products that did not enter the separating gel |
RESULTS
DOPA2-linked protein complexes were more visible on SDS-PAGE than the DSS-linked counterparts
In the bacterial glucose phosphotransferase system (Deutscheret al. 2006; Kotrbaet al. 2001), the phosphate group from phosphoenolpyruvate is transferred to glucose via Enzyme I (EI), the phosphocarrier protein (HPr), and subsequently, Enzyme II (EII). EI is composed of an N-terminal domain (EIN) and a C-terminal domain (EIC). EII comprises three functional subunits EIIAGlc, EIIBGlc, and EIICGlc (Deutscheret al. 2006; Kotrbaet al. 2001). Transient interactions are responsible for the transferring of a phosphoryl group from EIN to HPr (Garrettet al. 1997, 1999) and from EIIAGlc to EIIBGlc (Caiet al. 2003). The KD values of these two protein complexes, EIN/HPr and EIIAGlc/EIIBGlc, are ~7 μmol/L (kon ≈ 1.57 × 108 L/mol·s, koff ≈ 1100 s−1) (Garrettet al. 1997; Suhet al. 2007) and ~25 μmol/L (kon ≈ 3.20 × 107 L/mol·s,koff ≈ 800 s−1) (Caiet al. 2003; Reizeret al. 1992), respectively (Fig. 1B ).
Consistent with the transient and weak interactions between EIN and HPr and between EIIAGlc and EIIBGlc, DSS cross-linking yielded no discernable covalent dimer bands on the SDS-PAGE at a protein concentration of 0.25 mg/mL (for EIN/HPr) or 0.63 mg/mL (for EIIAGlc/EIIBGlc) (Fig. 1C and 1D). Note that the protein concentrations used here are in the range for a typical cross-linking experiment. Interestingly, DOPA2 cross-linking at the same protein concentrations generated a faint covalent dimer band for either complex with the expected molecular weight (Fig. 1C and 1D).
In the above experiments, the protein concentrations were set to about 1 × KD, i.e., 7 µmol/L for EIN and HPr (0.06 and 0.19 mg/mL, respectively), and 25 µmol/L for EIIAGlc and EIIBGlc (0.40 and 0.23 mg/mL, respectively). Under these conditions, the equilibrium concentrations of the heterodimers, EIN/HPr and EIIAGlc/EIIBGlc, were 2.67 and 9.55 μmol/L, respectively. In other words, only 38% of the protein subunits were in the complex form. When the protein concentrations were increased to 10 × KD, the complex percentage increased to 73%, corresponding to an equilibrium concentration of 51.09 μmol/L for the EIN/HPr heterodimer and 182.46 μmol/L for EIIAGlc/EIIBGlc. Thus, there is a 19-fold increase in the quantity of the protein complexes relative to that at 1 × KD. Indeed, at 10 × KD concentration, cross-linking reactions with DOPA2 generated more prominent dimer bands on SDS-PAGE than with DSS (Fig. 2A –2D). Therefore, it seems that DOPA2 is better than DSS at capturing transient PPIs.
2 In comparison to DSS, DOPA2 captured more non-covalent dimers readily visible on SDS-PAGE. A,B SDS-PAGE of cross-linked EIN/HPr and EIIAGlc/EIIBGlc complexes. Captured protein heterodimers were marked by square brackets. C,D The relative intensity of the heterodimer band, normalized by the total intensity. E,F The inter-molecular or intra-molecular residue pairs identified in the covalent dimer bands cross-linked by DOPA2 or DSS. G,H DOPA2 cross-linked or DSS cross-linked residue pairs identified from excised dimer bands were mapped to the primary sequences of the respective complexes (visualized using xiNET (Combeet al. 2015)). Cross-links were filtered by requiring FDR < 0.01 at the spectra level, E-value < 1 × 10 −3 and spectral counts ≥ 3 |
Liquid chromatography coupled with tandem mass spectrometry analysis (LC-MS/MS) of the cross-linked dimers from the gel bands indicated that the DOPA2 samples contained more intra- and inter-molecular cross-linked residue pairs (referred to as cross-links hereafter) than the DSS counterparts (Fig. 2E –2H). For example, the number of inter-molecular cross-links between EIN and HPr was 31 for DOPA2 and 21 for DSS, and that between EIIAGlc and EIIBGlc was 37 for DOPA2 and 26 for DSS (Fig. 2E and 2F). An analysis of the Euclidean distance of each cross-linked residue pair revealed that a higher percentage of the DOPA2 cross-links (63% for EIN/HPr, 64% for EIIAGlc/EIIBGlc) are compatible with the known complex structures than the DSS cross-links (28% for EIN/HPr, 51% for EIIAGlc/EIIBGlc) (Table 1 ). The same observation holds for the solvent accessible surface distance (SASD) (Table 1 ).
1 Structural compatibility rate of residue pairs (inter-links plus intra-links) |
| Cα–Cα distance (Å) | EIN/HPr | EIIAGlc/EIIBGlc | ||||
| Euclidean distance | SASD | Euclidean distance | SASD | |||
| * Spacer arm with MM2 minimization SASD is short for Solvent Accessible Surface Distance. Structural models of EIN/HPr (PDB code: 3EZA) and EIIAGlc/EIIBGlc (PDB code: 1O2F) were used as reference. Both were treated as stereospecific complexes | ||||||
| DOPA2 | 30.2* | 63% (36/57) | 46% (24/52) | 64% (35/55) | 35% (18/52) | |
| DSS | 24 | 28% (11/39) | 18% (6/34) | 51% (18/35) | 30% (10/33) | |
Transient protein complexes cross-linked by DSS were heterogeneous and spreading widely on SDS-PAGE
In parallel, we analyzed the same cross-linking reactions without SDS-PAGE separation (Fig. 3A ). Interestingly, when digested in solution, the samples cross-linked with DSS yielded more cross-link identifications than the samples cross-linked with DOPA2 for both intra- and inter-molecular cross-links (i.e., 40 and 13 for DSS, 24 and 6 for DOPA2) (Fig. 3B ). This is the opposite of the in-gel digestion result, which contained fewer DSS cross-links than DOPA2 cross-links. Furthermore, unlike the samples digested in-solution, the excised, in-gel digested covalent dimers produced more inter- than intra-molecular cross-links, either with DOPA2 or with DSS (Fig. 3B ). When the inter-molecular cross-links were mapped to the known structure of EIN/HPr, we found that the cross-links bridging the interface between EIN and HPr are more for DOPA2 than for DSS (Fig. 3C –3F, denoted by orange-colored lines). Of note, the cross-links examined here all passed a stringent filter of the search results (FDR < 0.01, at least four MS2 spectra at E-value < 1 × 10 −8).
3 Contrasting cross-link identification results before and after SDS-PAGE purification. A Schematic diagram of in-gel digestion or in-solution digestion of EIN/HPr complex after DOPA2 or DSS cross-linking. B The inter-molecular or intra-molecular residue pairs identified in the EIN/HPr complex cross-linked by DOPA2 or DSS with purification by SDS-PAGE or not. The number of spectra identified is shown below. C,D The inter-molecular residue pairs identified by DOPA2 in the dimer band or in solution were mapped to the primary sequence of EIN/HPr (visualized using xiNET (Combeet al. 2015)). E,F As in panels C and D, but for cross-links identified by DSS. The stereospecific interface between EIN and HPr was indicated by light orange color. The green lines and the orange lines denoted the cross-links that were assigned to ECs and SC, respectively. The numbers of the corresponding spectra of each cross-link were indicated by the thickness of the line. Cross-links were filtered by requiring FDR < 0.01 at the spectra level, E-value < 1 × 10 -8 and spectral counts > 3 |
To account for the discrepancy, we systematically analyzed the cross-linking products following SDS-PAGE separation. The most prominent covalent dimer band (L3) and the gel slices above (denoted as L1 and L2) and below (denoted as L4, L5, and L6) were all analyzed (Fig. 4A ). For intra-molecular cross-links (intra-EIN and intra-HPr), more cross-links were identified for DSS than for DOPA2 (53 versus 41) from these gel slice samples (Fig. 4B and 4C). For inter-molecular cross-links between EIN and HPr, the opposite was observed: fewer inter-molecular cross-links were identified with DSS than with DOPA2 (29 versus 33 residue pairs, or 679 versus 1455 spectra) (Fig. 4B and 4C, and Fig. 4D and 4E). Notably, the inter-molecular cross-links by DOPA2 were all identified from either the visible EIN/HPr dimer band (L3) or above (L1–L2), whereas those by DSS were detected throughout the lane (L1–L5) (Fig. 4B and 4C). These results suggest that, when compared to the dimeric complexes cross-linked by DOPA2, DSS-linked complexes are more heterogeneous, spreading over a wider molecular weight range and affording a less distinct dimeric band.
4 Systematic analysis of cross-linked protein species separated by SDS-PAGE. A SDS-PAGE of DOPA2 or DSS-cross-linked EIN/HPr and the demarcation range for systematic excision of the gel slides, arbitrarily denoted as L1 through L6. B,C The number of inter- or intra-molecular residue pairs identified for DOPA2 or DSS cross-linking in L1–L6, respectively. EIN(30)-EIN(30), the only unambiguously identified homodimeric cross-link suggestive of higher-order interactions between the EIN/HPr heterodimers, was included in this graph. It came from L1 and L2 of the DOPA2 treated sample or L1, L2 and L4 of the DSS treated sample. D,E The number of inter- or intra-molecular spectra identified for DOPA2 or DSS cross-linking in L1–L6, respectively. Cross-links were filtered by requiring FDR < 0.01 at the spectra level, E-value < 1 × 10 -8 and spectral counts in each sample ≥ 2 |
DOPA2 cross-linking favored the stereospecific complex whereas DSS cross-linking favored the encounter complexes
We showed in our previous work (Gonget al. 2015) that a vast majority of the EIN/HPr inter-molecular cross-links generated by BS2G or BS3 (both are NHS ester cross-linkers) featured ECs. Only one out of the 13 inter-molecular BS2G or BS3 cross-links comes from SC (Gonget al. 2015). Fleeting ECs cannot be captured by crystallization, nor by the standard NMR technique, but are detectable with the use of paramagnetic NMR (Tanget al. 2006). In contrast, the SC of EIN/HPr has been determined to be atomic resolution using the standard NMR method (Garrettet al. 1999; Tanget al. 2006).
In the current study, we mapped inter-molecular cross-links on the representative structures of the ECs and the SC of EIN/HPr (Fig. 5A ). When the EIN/HPr samples were analyzed by LC-MS/MS without SDS-PAGE separation (in-solution samples) as done before (Gonget al. 2015), one of the six inter-molecular DOPA2 cross-links is consistent with the SC (17%). In contrast, none of the 13 inter-molecular DSS cross-links is consistent with the SC (Fig. 5B , Table 2 and the supplementary Table S1A and S1B). From the gel band of the covalently linked EIN/HPr dimer, four (27%) out of the 15 inter-molecular DSS cross-links represented the SC (Fig. 5C , Table 2 and the supplementary Table S1C). As the SC cross-links are already favored by DOPA2, further increase in SC became less obvious when analyzing the gel band (Fig. 5C and Table 2 ). Nevertheless, LC-MS/MS analysis of the gel bands containing DOPA2-linked EIN/HPr showed that 8 out of 25 (32%) inter-molecular cross-links correspond to the SC (Fig. 5C , Table 2 and the supplementary Table S1D). In brief, cross-links representing the SC were found enriched among the DOPA2 cross-links and the cross-links identified from the exercised covalent complex bands.
5 Analysis of the heterodimeric cross-links with respect to the conformations of the stereospecific and encounter complexes. A Three representative encounter structures (EC-I, EC-II, and EC-III) and stereospecific complex (SC, PDB code: 3EZA) (Fawziet al. 2010; Garrettet al. 1999; Tanget al. 2006) for EIN/HPr complex. B The number of DOPA2 or DSS cross-linked inter-links of EIN/HPr compatible with SC and ECs w/t or w/o gel purification. C The relative percentages of DOPA2 or DSS cross-linked inter-links of EIN/HPr compatible with SC and ECs w/t or w/o gel purification. D The number of DOPA2 or DSS cross-linked spectra of EIN/HPr compatible with SC and ECs in L1–L3, respectively. E The relative percentages of DOPA2 or DSS cross-linked spectra of EIN/HPr compatible with SC and ECs in L1–L3, respectively. Cross-links were filtered by requiring FDR < 0.01 at the spectra level, E-value < 1 × 10 −8 and spectral counts in each sample ≥ 2 |
2 Inter-molecular cross-links identified from the EIN/HPr complex |
| Cross-linker | Encounter complexes (ECs) | Stereospecific complex (SC) | Total X-links | Total spectra of inter-molecular X-links | Total spectra of intra-molecular X-links | |||||
| # of X-links | # of spectra | # of X-links | # of spectra | |||||||
| The cross-links identification results were filtered by requiring FDR < 0.01 at the spectra level, E-value < 1 × 10 −8 and spectral counts > 3. The estimated FDR at the residue pair level is zero. The cross-links were classified according to their structural compatibility with either the stereospecific complex or the encounter complexes. The cross-linking data of 0.05, 0.2, and 0.8 mmol/L DOPA2 or DSS were combined | ||||||||||
| In-solution | DSS | 13 (100%) | 498 (100%) | 0 | 0 | 13 | 498 | 5007 | ||
| DOPA2 | 5 (83%) | 168 (97%) | 1 (17%) | 5 (3%) | 6 | 173 | 2504 | |||
| In-gel | DSS | 11 (73%) | 1515 (60%) | 4 (27%) | 1027(40%) | 15 | 2542 | 1187 | ||
| DOPA2 | 17 (68%) | 2186 (56.6%) | 8 (32%) | 1676 (43.4%) | 25 | 3862 | 4926 | |||
We also examined the inter-molecular cross-links identified in gel slices L1–L3. Starting with the DOPA2- or DSS-linked covalent dimer bands at L3, more DOPA2 cross-links (six residue pairs, 337 spectral counts) were consistent with the SC than DSS cross-links (five residue pairs, 150 spectral counts) (Fig. 5D and 5E and the supplementary Table S2). This is consistent with the result shown above (Fig. 5B and 5C) and with that from L2 (Fig. 5D and 5E, and the supplementary Table S2). The highest molecular weight region L1 had only a small number of cross-link spectra identified, diminishing but not eliminating the said difference between DOPA2 and DSS (Fig. 5D and the supplementary Table S2). Together, these data indicate that DOPA2 captures the SC more effectively than DSS.
A proposed model for the differential preference of SC and ECs by DOPA2 and DSS cross-linking
One intriguing result in this study is the differential preference of DOPA2 and DSS towards SC and EC, respectively. Namely, for the weak EIN/HPr complex, the products generated by the relatively slow cross-linking reagent DSS predominantly represent the fleeting ECs, while the longer-lived SC is better represented among the products of the faster cross-linking reagent DOPA2. Similar to this observation, a previous study also demonstrated that NHS esters DSS/BS3 trapped dynamic states in the ensemble while the fast photoactivatable diazirine-containing SDA cross-linker had better performance than DSS/BS3 for accurate modeling (Ziemianowiczet al. 2019).
The process of capturing a protein–protein interaction by cross-linking may be conceptualized as follows. We start from the intermediate state when one end of the cross-linker is already covalently attached to either protein (A or B) in a transient complex A•B. For the subsequent inter-molecular cross-linking reaction, the likelihood of the formation of an inter-molecular cross-link will be determined by the reaction rate of the amine-targeting group as well as the half-life of the A•B complex.
On the other hand, if the reaction rate is not fast enough in comparison to the half-life of the A•B complex, no inter-molecular cross-link will form in this round of association. However, this “planted” cross-linker still has a chance to form an inter-molecular cross-link in the next round of association before it loses reactivity following an attack from an adjacent intra-molecular amine group or a quenching reagent.
As shown in Fig. 6 , we propose that DSS or BS3 cross-linking reactions are too slow to capture either the SC or the fleeting ECs in a single association/dissociation round. As a result, the observed cross-links may come from the subsequent association events. As the SC and the ECs undergo rapid interconversion, it is not surprising that a DSS or BS3 mono-linked protein can be attached to its partner protein in conformations different from the SC. This is in line with an earlier work demonstrating that a slow reacting sulfonyl fluoride group can capture weak and transient interactions once planted onto a protein (Yanget al. 2018). This explains the observation that the number of inter-molecular DSS or BS3 cross-links identified from EIN/HPr is small and most of them represent ECs.
6 Proposed model for the conformational preference by DOPA2 and DSS in cross-linking. (1) An EC is too transient for either DOPA2 or DSS to capture immediately. (2) Compared to DSS, the fast-reacting cross-linker DOPA2 has a higher chance of capturing the stereospecific complex of EIN/HPr, by forming an inter-molecular cross-link before the two subunits dissociate. (3) Cross-linking by DSS is too slow to capture an EIN/HPr complex in one shot. DSS cross-linking probably takes place in two steps: first a DSS mono-link forms, then in one or more subsequent dissociation/association cycles, the mono-link planted on a subunit turns into a cross-link. Between the mono-link and the cross-link, the protein has many opportunities to sample various alternative conformational states, including the ECs |
In contrast, DOPA2 has a faster reaction rate and therefore can be more efficient in capturing the SC before the two subunits dissociate. Our recent NMR analysis showed that the actual dissociation constant koff of EIN/HPr complex can reach 8900 s−1, corresponding to a lifetime of ~100 µs of the complex (Donget al. 2022). Thus, it is remarkable that DOPA2, or DSS to a lesser extent, still captures the stereospecific EIN/HPr complex. As ECs exist as an ensemble of conformations while the SC represents the final lowest-energy conformational state, cross-linked ECs are more heterogeneous, and therefore more spread-out and less visible on SDS-PAGE, as in the case of DSS cross-links.
DISCUSSION
In this study, we have shown that under near-physiological conditions, DOPA2 is advantageous over DSS in capturing transiently interacting proteins for the structural characterization of a stereospecific complex. It makes intuitive sense that a faster cross-linker is advantageous in capturing a transient PPI, but the truth is more nuanced than expected. “Fast” and “slow” are relative terms. Although faster than DSS in protein cross-linking, DOPA2 is probably not fast enough to cross-link the SC of EIN/HPr whose lifetime is ~100 µs in a single association-dissociation cycle, let alone the fleeting ECs. Counterintuitively, the slower DSS captures more ECs, probably by going through more protein association-dissociation rounds than DOPA2. This could explain the observation that although most intermolecular cross-links by either DOPA2 or DSS were consistent with the ECs, more DOPA2 intermolecular cross-links than DSS intermolecular cross-links could be assigned to the SC (see Table 2 ). It is worth noting that for a stable protein-protein interaction, slow NHS ester cross-linkers can capture the SC effectively (Gong 2015).
Though the structures of ECs of EIIAGlc/EIIBGlc are not yet available from other characterization methods and cannot be compared against, the cross-links found outside the interface region of the SC are in highly charged patches, characteristic of ECs. Further, DSS and DOPA2 displayed the same contrasting behaviors on EIIAGlc/EIIBGlc as they did on EIN/HPr. For either heterodimeric complex, more DOPA2 cross-links (>60%, with intra- and inter-molecular cross-links combined) were consistent with the structure of the SC than did the DSS cross-links (28%–51%) (Table 1 ). Moreover, DOPA2 cross-linking generated a covalent heterodimer band that is much more distinct on SDS-PAGE than did DSS cross-linking for both EIN/HPr and EIIAGlc/EIIBGlc (Fig. 1 and Fig. 2 ). Thus, our findings lead to the conclusion that gel purification is generally a good practice for structural modeling of the SC.
We also noticed that in our previous analysis of six model proteins (aldolase, BSA, catalase, GST, lysozyme, and myosin), the DOPA2 cross-links have a higher compatibility rate with the determined structures of these monomeric or homo-dimeric proteins than the DSS cross-links (Wanget al. 2022). Based on our current findings, we propose that the difference in the compatibility rate can be accounted for, at least in part, by the fact that faster DOPA2 cross-linking is more effective at capturing the lowest-energy conformational state, thereby accommodating intramolecular dynamics.
EXPERIMENTAL SECTION
Materials and reagents
DSS, tris(2-carboxyethyl) phosphine (TCEP), Dithiothreitol (DTT), and 2-Iodoacetamide (IAA) were purchased from Pierce Biotechnology (Thermo Scientific). Dimethylsulfoxide (DMSO), HEPES, NaCl, urea, CaCl2, and methylamine were purchased from Sigma-Aldrich. Acetonitrile (ACN), formic acid (FA), acetone, and ammonium bicarbonate were purchased from J.T. Baker. Trypsin and Asp-N (gold mass spectrometry grade) were purchased from Promega. DOPA2 was synthesized as previously described (Wanget al. 2022).
Preparation of protein samples
Protein cross-linking
The KD value of EIN/HPr complex was ~7 μmol/L in 20 mmol/L Tris-HCl buffer, pH 7.4, 150 mmol/L NaCl at 40 ˚C (Suhet al. 2007). EIN/HPr complexes were diluted to 1 × KD (0.25 mg/mL) and 10 × KD (2.5 mg/mL), and then were cross-linked with DOPA2 at the final concentration of 0.05, 0.2 and 0.8 mmol/L at room temperature for 10 min, respectively. EIN/HPr complexes were also cross-linked with DSS at the final concentration of 0.05, 0.2 and 0.8 mmol/L at room temperature for 1 h.
The KD value of EIIAGlc/EIIBGlc complex was estimated to be ~25 μmol/L. According to a previous study, EIIAGlc/EIIBGlc has a Km value of 1.7–25 μmol/L (Reizeret al. 1992). We did not succeed in obtaining an accurate KD value by either surface plasmon resonance assay or ELISA. However, it is clear that the binding is weak, and the KD value is likely greater than 10 μmol/L. EIIAGlc/EIIBGlc complexes were diluted to 1 × KD (0.63 mg/mL) and 10 × KD (6.3 mg/mL), and then were cross-linked with DOPA2 at the final concentration of 0.005, 0.01, 0.05, 0.2 and 0.8 mmol/L at room temperature for 10 min, respectively. EIIAGlc/EIIBGlc complexes were also cross-linked with DSS at the final concentration of 0.05, 0.2 and 0.8 mmol/L at room temperature for 1 h.
In-solution digestion
Cross-linked proteins were precipitated with the 4-fold volume of acetone for at least 30 min at −20 °C. The pellets were air dried and then dissolved, assisted by sonication, in 8 mol/L urea, 20 mmol/L methylamine, 100 mmol/L Tris, pH 8.5. After reduction (5 mmol/L TCEP, RT, 20 min) and alkylation (10 mmol/L IAA, RT, 15 min in the dark), the samples were diluted to 2 mol/L urea with 100 mmol/L Tris, pH 8.5. Denatured proteins were digested by trypsin alone or trypsin plus Asp-N at a 1/50 ( w/w) enzyme/substrate ratio at 37 °C for 16–18 h, and the reactions were quenched with 5% formic acid (final conc.).
In-gel digestion
The target bands in the one-dimensional glycine gel were excised manually from the gel slab and cut into pieces. Briefly, after fully washed with the destaining solution and ddH2O, the gel pieces were in-gel reduced (10 mmol/L DTT, 56 °C, 40 min) and alkylated (55 mmol/L IAA, RT, 60 min in the dark) and then dehydrated by 100% ACN. Gel pieces were further rehydrated (50 mmol/L NH4HCO3, 10 ng/μL trypsin) and digested for 16–18 h. The peptides were twice extracted from the gel by extraction solution I (50% ACN, 5% FA) and extraction solution II (75% ACN, 5% FA), respectively. The extracted digests were combined and the sample volume was reduced to about 10 μL in SpeedVac for MS analysis.
LC-MS analysis
All proteolytic digestions of proteins were analyzed using an EASY-nLC 1000 system (Thermo Fisher Scientific) interfaced with an HF Q-Exactive mass spectrometer (Thermo Fisher Scientific). Peptides were loaded on a pre-column (75 μm ID, 4 cm long, packed with ODS-AQ 12 nm–10 μm beads) and separated on an analytical column (75 μm ID, 12 cm long, packed with Luna C18 1.9 μm 100 Å resin). Slight modifications to the separation method were made for different samples. EIN/HPr and EIIAGlc/EIIBGlc complexes were injected and separated with a 75 min linear gradient at a flow rate of 200 nL/min as follows: 0–5% B in 1 min, 5%–35% B in 59 min, 35%–100% B in 5 min, 100% B for 10 min (A = 0.1% FA, B = 100% ACN, 0.1% FA). The top fifteen most intense precursor ions from each full scan (resolution 60,000) were isolated for HCD MS2 (resolution 15,000; NCE 27) with a dynamic exclusion time of 30 s. Precursors with 1+ , 2+ , more than 6+ , or unassigned charge states were excluded.
Identification of cross-links with pLink 2
The search parameters used for pLink 2 (Chenet al. 2019) were as follows: instrument, HCD; precursor mass tolerance, 20 ppm; fragment mass tolerance 20 ppm; cross-linker DOPA2 (cross-linking sites K and protein N-terminus, cross-link mass-shift 334.084, mono-link w/o hydrazine mass-shift 352.096, mono-link w/t hydrazine mass-shift 348.111), cross-linker DSS (cross-linking sites K and protein N-terminus, cross-link mass-shift 138.068, mono-link mass-shift 156.079); fixed modification Carbamidomethyl[C]; variable modifications Deamidated[N], Deamidated[Q], and Oxidation[M]; peptide length, minimum 6 amino acids and maximum 60 amino acids per chain; peptide mass, minimum 600 and maximum 6,000 Da per chain; enzyme, trypsin, with up to three missed cleavage sites per cross-link. Protein sequences of model proteins were used for database searching. The results were filtered by requiring a spectral false identification rate < 0.01.
Cα–Cα distance calculations
The Cα–Cα Euclidean distances were measured using PyMOL for each PDB file. The PDB files we use are as follows: EIN/HPr (3EZA (Garrettet al. 1999)), EIIAGlc/EIIBGlc (1O2F (Caiet al. 2003)). The Cα–Cα Solvent Accessible Surface Distance (SASD) was calculated using Jwalk (Matthew Allen Bullocket al. 2016). If the SASD of cross-linked residue pairs cannot be calculated due to a lack of surface accessibility, these residue pairs are excluded from the calculation. When calculating structural compatibility, the distance cut-offs are 30.2 Å for DOPA2, and 24.0 Å for DSS.
Classification of inter-molecular cross-links
The ensemble structure refinement based on the intermolecular CXMS restraints of BS2G and BS3 was performed in our previous study (Gonget al. 2015). The ensemble structures of EC-I, EC-II, and EC-III from the previous report (Tanget al. 2006) and the SC from the PDB structure (accession code 3ZEA) were used as the reference structures. If the Euclidean distance of cross-linked residue pairs was within the range of the cross-linker, the cross-link was assigned to the SC. The cross-link that would be over-length in the SC was assigned to the ECs.