Saša Kazazić*a,
Zrinka Karačićb,
Igor Sabljića,
Dejan Agićc,
Marko Tominb,
Marija Abramićb,
Michal Dadlezd,
Antonija Tomićb and
Sanja Tomić*b
aRuđer Bošković Institute, Institute of Biochemistry and Biophysics Polish Academy of Sciences, Croatia. E-mail: sasa.kazazic@irb.hr
bRuđer Bošković Institute, Institute of Biochemistry and Biophysics Polish Academy of Sciences, Croatia. E-mail: sanja.tomic@irb.hr
cJosip Juraj Strossmayer University of Osijek, Faculty of Agriculture, Croatia
dInstitute of Biochemistry and Biophysics Polish Academy of Sciences, Poland
First published on 10th April 2018
The hydrogen deuterium exchange (HDX) mass spectrometry combined with molecular dynamics (MD) simulations was employed to investigate conformational dynamics and ligand binding within the M49 family (dipeptidyl peptidase III family). Six dipeptidyl peptidase III (DPP III) orthologues, human, yeast, three bacterial and one plant (moss) were studied. According to the results, all orthologues seem to be quite compact wherein DPP III from the thermophile Caldithrix abyssi seems to be the most compact. The protected regions are located within the two domains core and the overall flexibility profile consistent with semi-closed conformation as the dominant protein form in solution. Besides conservation of conformational dynamics within the M49 family, we also investigated the ligand, pentapeptide tynorphin, binding. By comparing HDX data obtained for unliganded protein with those obtained for its complex with tynorphin it was found that the ligand binding mode is conserved within the family. Tynorphin binds within inter-domain cleft, close to the lower domain β-core and induces its stabilization in all orthologues. Docking combined with MD simulations revealed details of the protein flexibility as well as of the enzyme–ligand interactions.
Protein purity was confirmed by SDS-PAGE according to Laemmli.34 Protein concentrations were determined by the Bradford method.35
Enzyme inhibition studies with tynorphin were performed using the Arg2-2NA as a substrate, according to Baršun et al.36 The inhibition constant, Ki, was calculated according to the equation:
Ki = i/(v0/vi − 1)·Km/(Km + s) | (1) |
Separate hydrogen deuterium exchange experiments were carried out for six unliganded DPP III enzymes, human V412I mutant, and their tynorphin complexes. The inhibition constant value (Table 1 ) was used to estimate proper tynorphin/enzyme ratio (T/E) present in stock solution in order to ensure that high percentage of the enzyme stays bound in complex under H/D exchange conditions. Those ratios were: 4.4 for hDPP III (99.76% bound); 3.3 for protein variant V412I hDPP III (99.91% bound); 10.7 for yDPP III (98.97% bound); 20 for BtDPP III (98.98% bound); 11.1 for PgDPP III (98.44% bound); 13.6 for PpDPP III (98.60% bound) and 120 for CaDPP III (97.26% bound). The reaction buffer was prepared using D2O (99.8% Armar Chemicals, Switzerland), in which the pH (uncorrected meter reading) was adjusted using DCl or NaOD (Sigma). After mixing 5 μL of protein stock with 45 μL of D2O exchange reaction was started at room temperature. Exchange reaction was started and followed for five time periods (10 s, 1 min, 20 min, 1 h and 4 h) each carried out in triplicate. The exchange reaction was quenched by reducing the pH to 2.5 by adding the reaction mixture to a plastic tube containing stop buffer (2 M glycine buffer, pH 2.5) cooled on ice. Immediately after quenching, the sample was manually injected into the nanoACQUITY (Waters, Milford, MA) UPLC system. Further pepsin digestion, LC, and MS analysis were carried out exactly as described for the non-deuterated sample. Relative deuterium uptake values are calculated without correction for back exchange. Peak deconvolution of the yeast DPPIII 407VRLKIGFKNVSLGNIL422 bimodal peak envelopes was carried out with MultiPeakFit package of the Igor Pro software (Igor Pro v 6.372, Multi-peak fit v2, Gauss fit, WaveMetrics, Inc.). Differential comparisons between HDX results for unliganded and tynorphin complex for each DPP III orthologue were carried out following the procedure described in Houde et al.37
Enzyme | Ki tynorphin (μM) |
---|---|
hDPP III | 0.03 |
yDPP III | 0.47 |
BtDPP III | 0.98 |
PgDPP III | 0.72 |
PpDPP III | 0.66 |
CaDPP III | 16.75 |
After 10 seconds | After 4 hours | |
---|---|---|
>30% | <20% | |
hDPP III | 12% | 18% |
yDPP III | 26% | 15% |
BtDPP III | 13% | 22% |
PgDPP III | 19% | 25% |
PpDPP III | 19% | 25% |
CaDPP III | 13% | 16% |
The proteins and protein-substrate complexes were placed into a truncated octahedron box filled with TIP3P water molecules,45 and Na+ ions46 were added in order to neutralize the systems.
Before running productive molecular dynamics simulations, the protein geometry was optimized in three cycles (every 1500 steps) and the system was equilibrated. In the first cycle of optimization, water molecules were relaxed, while the rest of the system was harmonically restrained with a force constant of 32 kcal mol−1 Å−1. In the second and third cycle, the same force constant (32 kcal mol−1 Å−1) was applied to the zinc ion, while the protein backbone was restrained with force constants of 12 and 2 kcal mol−1 A−1, respectively. The energy minimization procedure, consisting of 470 steps of steepest descent followed by conjugate gradient optimization for the remaining steps, was the same in all cycles. During the first period of equilibration (200 ps of gentle heating from 0 to 300 K with a time step of 1 fs), the NVT ensemble was used, while all of the following simulations were performed at constant temperature and pressure (300 K and 1 atm, 2 fs time step, the NpT ensemble). During the equilibration, the zinc ion and/or its ligands were weakly restrained. The temperature was held constant using Langevin dynamics47 with a collision frequency of 1 ps−1. The pressure was regulated by a Berendsen barostat.48 Bonds involving hydrogen atoms were constrained using the SHAKE49,50 algorithm. The ligand free proteins for which the crystal structure was available were equilibrated for 1 ns and in the case of the initial structure derived by comparative modelling 50 ns of equilibration was performed. For each orthologue at least 150 ns of productive MD simulations was accomplished.
Fig. 1 Conformational flexibility of unliganded hDPP III enzyme. Deuterium uptake values for the peptides are expressed as the percentage of the maximum incorporation measured in control experiment. Relative percentage values for each incubation time period are color-coded and mapped onto the crystal structure of the unliganded hDPP III enzyme (PDB_code 3FVY) and presented in front (A) and back view (B). Parts of the structure colored white are not covered with peptic peptides (detailed data about sequence coverage in Fig. S1A†). Protein structure regions discussed in the text are indicated. For the labeled peptides identification (sequences) see Table S1.† |
In other words for all hDPP III peptides, local unfolding/folding is faster than it is their exchange reaction rate. Broadening of the peptide isotopic envelope due to deuterium incorporation follows a binomial distribution. Agreement between the measured deuterium uptake and the uptake determined from the amide hydrogen bond analysis during MD simulations (Fig. S2†) was obtained with a correlation coefficient of 0.766. Calculation of the theoretical values is carried out according to the similar procedure already published.52 Apart from the limitations of MD simulation, i.e. the conformations sampled during 150 ns present only a part of the protein conformational space, the correlation between the experimental and theoretical data depends on the quality of the HDX prediction model and approximations that are taken into account as well.53–55
Peptides with deuterium uptake larger than 30% of their available amide hydrogens after the shortest (10 seconds) exchange time period indicate region permanently exposed to water or region that often becomes exposed to solvent under conformational changes of the protein. On the other hand, peptides having less than 20% of their available amide hydrogens exchanged with deuterium after the longest (4 hours) exchange time period correspond to the protected, hydrophobically shielded protein regions with amide hydrogens involved in the internal hydrogen bond network. The ratio of fast and slow exchanging peptides can be considered a measure of enzyme conformational plasticity. For human DPP III the fast exchanging peptides cover just 12% of the amino acid sequence. The comparable level of H/D exchange was determined for all unliganded DPP III orthologues wherein the largest share of the fast exchanging peptides was found in yeast DPP III (Table 2 ). More detailed insight into the local conformational behavior of the unliganded DPP III enzymes was provided by analyzing relative fractional uptake values for the peptic peptides. The values measured for human and yeast orthologues are shown in Fig. 2. It can be noticed that patterns of the peptides with high H/D exchange determined for these two enzymes are similar. The peptides with the highest exchange ratio are part of the protein regions that visit the exchange-competent state many times during the labelling periods. Region I includes the amino acid residues 263–443 and 270–453 in the hDPP III and yDPP III, respectively and comprises several secondary structure elements, equivalent in hDPP III and yDPP III wherein the largest (five strands) β-sheet defines the bottom of the active site cavity. This region also includes the hinge region and several helices. In hDPP III these are helices 17, 18 and 19 in the hairpin, and helices 14, 15, 16 in the lower domain. Region II covers the unstructured loop in the upper domain between amino acid residues His450 and Tyr506. Region III covers motifs located in the upper domain; the beta hairpin (E), helices 31, 32 and 33 and the hinge region between Arg669 and Val673; a second strand of the beta sheet (A), the beta hairpin (F) and helix 34 (see Fig. S1B†). Peptides with minimum exchange indicate the highly protected protein regions with the amide hydrogens involved in stable internal hydrogen bonds. Such regions are helices 2, 3, 6, 35 and partially helices 12 and 14 in the lower domain of hDPP III and helices 23, 25 and 27 comprising the upper domain core. The substantial diversity in the enzymes amino acid composition (Table S2 and Fig. S3†) gives an explanation for the variability of the HDX profiles determined for different members of the M49 enzyme family. As can be seen by comparing the flexibility profiles of different M49 orthologues (Fig. S4†) the highest similarity between the deuterium uptake determined in different orthologues is in the region denoted as Region I in hDPP III. This common HDX pattern is related to the DPP III inter-domain dynamics. By comparing the structure of the unliganded hDPP III with the hDPP III structure in its complex with tynorphin Bezerra et al.18 found that the long range conformational change of the protein structure could be described as domain rotation around the ‘peptide hinge’ 409LGNVLAVAYATQ420. For all DPP III orthologues the HDX study revealed EX2 exchange regime except for the yeast DPP III peptide 407VRLKIGFKNVSLGNIL422 which exhibited mixed EX1/EX2 kinetics (Fig. 3). This peptide comprises the hinge residues 418LGNIL422 corresponding to residues 409LGNVL413 of hDPP III. Bimodal broadening of the isotope envelope (Fig. 3) observed in this case indicates that inter-domain motion which exposes this peptide to interact with deuterated solvent is slower than chemical exchange rate of the exposed amide hydrogen's so that several of them exchange at once. Such behavior is opposite to that represented with binomial broadening (Fig. 3) were inter-domain motion is fast enough so that hDPP III peptide 399RQTEGFKNVSLGNVL413 visits exchange competent conformation many times before one amide hydrogen is exchanged. Replacement of valine at position 412 with more bulky isoleucine has not resulted with the EX1/EX2 HDX kinetics in the human V412I mutant. Namely, the peptide 399RQTEGFKNVSLGNIL413 bearing isoleucine instead of valine still exhibited binomial exchange profile (EX2 exchange regime, Fig. 3) although overall exchange was higher than it was for wild type hDPP III (Fig. 3). Conformational space sampled during MD simulation is close to fully folded protein form56 covering fast local structural fluctuations which correspond to peptides that follow EX2 HDX kinetics and have binomial isotopic envelope broadening. Yeast DPPIII 407VRLKIGFKNVSLGNIL422 is the only peptide in this study that was exhibiting mixed EX1/EX2 exchange kinetics indicating existence of population with different conformational states where one of them is populated as a result of larger inter-domain motion occurring on relatively slower timescale (EX1) which was not identified in MD simulation. The conformation of the peptide (VRLKIGFKNVSLGNIL) itself has not changed significantly during the simulations, but its accessibility to solvent did. Namely, we traced two different forms of its surrounding; one in which the peptide's backbone is mostly solvated with SASA of the peptide nitrogen atoms (Val407N–Leu422N) being 6.1 Å2 and the other in which it is more protected with SASA of N407–N422 of 1.6 Å2 (Fig. S5†). According to this result, we could assume that these two forms of yDPPP III are (more or less) equally populated in the solution during the HDX experiment.
Fig. 2 Fractional uptake of deuterium in peptides of (A) hDPP III, and (B) yDPP III, obtained by pepsin hydrolysis during five exposure periods: 10 s-yellow trace, 1 min-red trace, 20 min-light blue trace, 1 hour-dark blue trace and 4 hours-black trace. For the peptides identification (sequences), see Table S1.† |
The kinetic measurements of the enzyme inactivation by tynorphin revealed the different degree of inactivation in studied orthologues. As shown in Table 2, all six studied DPP III enzymes were inhibited by tynorphin. However, inhibitory potency differed significantly. Human DPP III was the most inhibited while four enzymes yDPP III, BtDPP III, PgDPP III and PpDPP III have had inhibition constant values 10–30 times higher, and the Ki value for CaDPP III was 500-fold increased, compared to hDPP III.
Fig. 4 Graphical representation of the deuterium uptake differences found for peptic peptides of human and yeast DPP III by comparing two states, unliganded enzyme and its tynorphin complex. Presented are statistically significant values calculated as the difference in the area under deuterium uptake curves of two protein states (unliganded enzyme and its tynorphin complex) which is normalized to a maximal possible deuterium uptake curve area for that specific peptide. All differences are shown in Fig. S6.† Light blue bars denote positive differences, meaning that area under uptake curve was larger for unliganded enzyme than for its complex with tynorphin. Positive difference colored black is for slow exchanging conformer and dark blue for fast exchanging conformer. Pink bars denote negative differences indicating that deuterium uptake in the region of protein structure covered by the corresponding peptide was higher upon tynorphin binding. Such color-coded values for the labelled peptides are mapped onto a corresponding enzyme–tynorphin structure. (A) Ribbon representation of the hDPP III structure bound to the tynorphin inhibitor from a front (left) and back (right) view of the active site cleft. (B) Ribbon representation of the yDPP III structure bound to the tynorphin inhibitor from a front (left) and back (right) view of the active site cleft. In both structures, tynorphin was presented by colored sphere atoms (orange-C, white-H, red-O and blue-N). For the peptides' identification (sequences), see Table S1.† |
It is interesting to notice that most of the peptides showing an increase of the deuterium uptake upon ligand binding, like p2, p34 and p55/p56, are situated on the “back”, convex, side of the protein for which MD simulations had already reported21 the increase in residue based solvent accessible surface area during protein closure. Agreement with experiment was also obtained for the peptides p13, p34 and p55/p56 for which we have determined smaller mean number of hydrogen bonds per residue in the liganded enzyme structure than in the unliganded one (Table S3†). Consequently, the same four peptides from the lower domain (p29, p38, p39 and p40) and p62 peptide from the upper hDPP III domain interact with tynorphin and establish larger mean number of hydrogen bonds per residue in the complex structure than in the ligand-free structure obtained by the MD simulations (Table S3†).
Although tynorphin binding did not induce the deuterium uptake increase in any region of yDPP III, its binding induced the deuterium uptake decrease in the upper part of the lower domain five-stranded β-core (Fig. 4B) as it was determined in the case of hDPP III. We assumed that the tynorphin binding modes in human and yeast DPP III are similar and built the yDPP III–tynorphin complex accordingly. During 100 ns of MD simulations the tynorphin remained bound in the form of a β-strand into the enzyme active site, close to the lower domain β-core. The representative orientation of the ligand is shown in Fig. 6, and the RMSD plot of the tynorphin backbone, as well as the overlay of three conformations generated during MD simulations are given in Fig. S8.† This is in agreement with the HDX results which revealed deuterium uptake decrease in the peptides from the predicted tynorphin binding site: 290YINHFVTGSSQAHKEAQKL308 (p33), 372YEKPIFNPPDF382 (p40), 387VLTFTGSGIPAGINIPNYDD406 (p41), 407VRLKIGFKNVSLGNIL422 (p42). Peptide p42 exhibits bimodal exchange kinetics and deuterium incorporation decrease in tynorphin complex. Decrease of less than 1% was measured for fast exchanging conformation indicating very weak protection from tynorphin. For slow exchanging conformation deuterium incorporation decrease is more than 22% indicating much stronger protection of the bound tynorphin (Fig. 3 and 4B).
Beside the ligand binding region, the reduced deuterium uptake was also observed in several other peptides from both lower domain: 160IGIYHVEEKAAL171 (p17), 172LGFPSQGYTSA182 (p20), 223QIWVASE229 (p26) and the upper domain: 498YKVGETWGSKFGQL511 (p52) and 630YLKHLHVYKCSG641 (p65). MD simulations revealed that ligand binding boosts the protein closure, i.e. the radius of gyration (Rgyr) of yDPP III decreases faster during MD simulation of the complex with tynorphin than of the ligand free protein (Fig. S8†), indicating that these peptides are either approaching the bound substrate (p52) or protein residues from the other domain (p17 and p20) in the yDPP III–tynorphin complex.
Differential HDX experiment with the human V412I mutant was carried out to check the importance of the amino acid composition within the hinge region for the conformation dynamics of the complex. By changing valine at position 412 in hDPP III to isoleucine the same amino acid composition in the hinge region of human (403GFKNVSLGNVL413) and yeast (412GFKNVSLGNIL422) orthologue was achieved. The differential HDX data analysis shows that tynorphin binding reduces deuterium uptake within the active site cavity in the regions of the V412I mutant which are getting in close contact with the inhibitor. Those are the parts of the β-core structure covered by peptides: 285FTQGSIEAHKRGSRF299 (p29), 310SYIGF314 (p31), 312IGFIES317 (p32), 381FAGSGIPAGINIPNYDDL398 (p38) and 399RQTEGFKNVSLGNIL413 (p40). Similarly to differential HDX experiment for yDPP III protein, the reduced deuterium uptake was observed for the upper domain peptide 489YRSGETWDSKF499 (p51) and for the lower domain peptide 152FSLEPRLRHLGLGKEGITT170 (p16) which are brought closer to each other by conformational change during the transition from open into the closed conformation. The position of these two peptides suggests that in the closed conformation the two domains are placed relative to each other like it was observed for yDPP III complex indicating a significant impact of the amino acid composition of the hinge region to the mode of closure and the inter-domain position. The H/D exchange curve observed for the mutant p40 peptide (Fig. 3) indicates faster deuterium exchange in the V412I mutant than in the wild-type hDPP III. This finding together with the absence of deuterium uptake increase in mutated enzyme suggests that the protein closure upon ligand binding is less pronounced in the V412I mutant than in the wild-type hDPP III.
Differential analysis of the HDX data obtained for two bacterial DPP III homologs, from Bacteroides thetaiotaomicron and Porphyromonas gingivalis, revealed only a weak influence of the tynorphin binding on the HDX kinetics (Fig. 7 and 8). Such findings are in agreement with the recently published crystal structure of the ligand free BtDPP III, which is significantly more compact than human DPP III,59 and MD simulations results of the BtDPP III which showed that the amplitude of the interdomain separation in hDPP III is significantly larger than in BtDPP III.58 In the case of BtDPP III the slight decrease in the deuterium uptake was observed for three peptides, 382IGINLPNAN390 (p42), 520LVRIEPGNN528 (p56) and 204YGAMKDPKDETPVSY218 (p26). The first is part of the lower domain beta strand and a loop within the active site cavity while the other two are located in the upper and lower domain, respectively. In accord with the observed decrease in deuterium uptake, molecular dynamics simulations of the BtDPP III–tynorphin complex showed the substrate interactions with the amino acids from peptides p42 and p56 (Fig. S9 and Table S4†). The peptide p26 consists mostly of a single loop, which makes it susceptible to large conformational motions. During the BtDPP III–tynorphin complex closure it translocated and established hydrogen bonds with the upper domain amino acids. Similar behavior has been previously reported for the BtDPP III–Arg2-2NA complex.58 However, this conformational transition has not been traced during MD simulations of the ligand-free enzyme (Fig. S10†).
Fig. 7 Graphical representation of the deuterium uptake differences found by comparing two states of the BtDPP III (unliganded enzyme and its complex with tynorphin). Ribbon representation of the BtDPP III structure (PDB accession number 5NA7) bound to the tynorphin inhibitor from a front (left) and side (right) view of the active site cleft is shown. For the peptides identification (sequences), see Table S1.† |
Fig. 8 Graphical representation of the deuterium uptake differences found by comparing two states of the PgDPP III (unliganded enzyme and its complex with tynorphin). Ribbon representation of the modelled PgDPP III structure bound to the tynorphin inhibitor from a front (left) and side (right) view of the active site cleft is shown. For the peptides identification (sequences), see Table S1.† |
PgDPP III is significantly larger (886 amino acid residues) than human, yeast, and B. thetaiotaomicron DPP III. It has all the evolutionarily conserved regions of the DPP III family but, differently from the other characterized DPP III orthologues, it possesses a C-terminal extension containing an armadillo (ARM) type of fold similar to that of the AlkD family of bacterial DNA glycosylases.
However, complementation assays in a DNA-repair-deficient Escherichia coli strain indicated the absence of alkylation repair function of this enzyme. On the other hand, its peptidase activity is comparable to that of BtDPP III.33 Weak changes of the HDX profiles upon the enzyme complexation with tynorphin were detected in the DPP III part of the protein region represented by peptides p18, p37, p39, and p45, (161IIKASSVNF169, 355LTIAGDSYPATPIG365, 375WIRAEHGSKSVT386, and 432HECLGHGSGQLLPGVPGDALGEHAST457, respectively) and in the peptide p74 from the ARM fragment (Fig. 8). According to the computational results, the peptides p37 and p45 interact directly with peptide ligand during MD simulations of the PgDPP III–tynorphin complex (Fig. S11 and Table S4†). Like in the case of human and yeast orthologue, ligand binding into the enzyme active site boosts the protein closure.12 In the case of PgDPP III this resulted in concurrent movements of both the lower DPP III domain and ARM fragment in the direction of the upper DPP III domain as it was discussed in our recent publication.33 This could serve as an explanation for the measured HDX decrease in the regions distant from the tynorphin binding site.
The plant DPP III from moss Physcomitrella patens (PpDPP III), differently from the other DPP III orthologues, contains the so-called NUDIX motif on the N-terminal part of the sequence, a characteristic of Nudix hydrolases (Fig. 9).8 Changes in HDX kinetics were detected only in part of the sequence covering DPP III domains, but not in the NUDIX domain (Fig. 9 and S6E†). All peptides with the reduced deuterium uptake upon complexation are located within the active site cleft indicating that tynorphin interacts with larger part of the beta core from the lower domain and so significantly reduces conformational dynamics of the active site cleft region.
Fig. 9 Graphical representation of the deuterium uptake differences found by comparing two states of the PpDPP III (unliganded enzyme and its complex with tynorphin). Ribbon representation of the modelled PpDPP III structure bound to the tynorphin inhibitor from a front (left) and side (right) view of the active site cleft is shown. For the peptides identification (sequences), see Table S1.† |
Peptides with reduced deuterium uptake, p50, p53, p57, p60 and p65 (448VTIGPYETYE457, 471IGIRDDEATQRLKL484, 518LYNSGDVKGPQTVAF532, 549VMLKNISQAKF559 and 571VEASQRGAVDF581, respectively) contain amino acid residues from the ligand binding subsites and according to the molecular modelling results they take part in ligand stabilization (Fig. S12†).
Comparing the HDX data obtained for unliganded DPP III from Caldithrix abyssi (CaDPP III) and its complex with tynorphin revealed significant changes of the HDX kinetics in the upper domain induced by tynorphin binding. Even though the β-sheet in the lower domain was again identified as a potential substrate binding place (moderately small decrease of the deuterium uptake was observed in one peptide covering part of the beta strand 307SAGDTKAGVQTLA319 (p39)) the significant decrease in the deuterium uptake was found in distant peptides of the upper domain: 345AKFDKLLKPIAE356 (p44), 351LKPIAE356 (p45), 352KPIAEKVL359 (p46), 361AEQLPLVT368 (p48), 421YNNLF425 (p54), 425FMIEKGVYPPEFEKQIY441 (p55) and 472LEKGAY477 (p60) (Fig. 10).
Fig. 10 Graphical representation of the deuterium uptake differences found by comparing two states of the CaDPP III (unliganded enzyme and its complex with tynorphin). Ribbon representation of the CaDPP III structure bound to the tynorphin inhibitor from a front (left) and back (right) view of the active site cleft is shown. For the peptides identification (sequences), see Table S1.† |
The other three peptides 379HEISHGLGPGKIVL392 (p50), 393NGRQTEVKKELKETYSSIEE412 (p51) and 450RTIRFGIN457 (p58) in the CaDPP III upper domain show an increase in deuterium uptake while the peptide 89RASSDPLDQLRL100 (p11), covering a small region in lower domain, experienced a decrease in deuterium uptake upon complexation. In summary, the binding of tynorphin significantly changes flexibility of the upper domain of Caldithrix abyssi DPP III. It should be noted that peptides p50 and p51 comprise the amino acids from the conserved regions, pentapeptides HEISH and EECK(R)A. MD simulations revealed a decrease of the H-bond population upon the substrate binding in the regions comprising peptides p50, p51 and p58, clearly showing that the tynorphin binding destabilizes the catalytically relevant amino acids (see Table S5† for the hydrogen bond population in the relevant peptides). However, like in the mesophylic orthologues, tynorphin binds in the form of a β-strand into the CaDPP III active site, close to the lower domain β-core (Fig. 11).
Both X-ray diffraction and MD simulations showed that the structure of the ligand free CaDPP III is much more compact than the ligand free structures of the other DPP III orthologues, which might be the reason for the significantly higher Ki value determined for tynorphin.12,18,19,59
Inspection of the H/D exchange in conserved protein regions revealed that the region containing the hexapeptide (or pentapeptide in the case of CaDPP III and PpDPP III) signature motif is, in general, among the most protected, i.e. the most rigid, regions within the considered orthologues, with exception of CaDPP III. The relative protection is highest in the human orthologue, in agreement with its high activity. In addition, results of H/D exchange experiment agree with the results of MD simulations which showed that in solution the semi-closed human DPP III form is the most populated one.
The similarities in the human and yeast DPP III flexibility profiles are closely correlated with the similarity of their 3D structures (RMSD of about 2 Å).
Besides the flexibility conservation, we also studied the possible differences in the ligand, pentapeptide tynorphin, accommodation into the enzyme and its influence on the protein conformation and local flexibility.
Despite differences in orthologues' structure and flexibility, we found that in all cases tynorphin binds to the upper part of the lower domain β-sheet.
It seems that the more structured enzymes (i.e. enzymes with small share portion of unstructured regions like loops), such as human and C. abyssi DPP III, are more sensitive to ligand binding than those whose structure is more disordered, e.g. BtDPP III and PgDPP III. The tynorphin binding mode previously determined in human DPP III by X-ray diffraction is, according to the differential H/D exchange study and MD simulations, preserved in complexes with different DPP III orthologues.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ra13059g |
This journal is © The Royal Society of Chemistry 2018 |