pmc.ncbi.nlm.nih.gov

Temperature dependence of protein motions in a thermophilic dihydrofolate reductase and its relationship to catalytic efficiency

Abstract

We report hydrogen deuterium exchange by mass spectrometry (HDX-MS) as a function of temperature in a thermophilic dihydrofolate reductase from Bacillus stearothermophilus (Bs-DHFR). Protein stability, probed with circular dichroism, established an accessible temperature range of 10 °C to 55 °C for the interrogation of HDX-MS. Although both the rate and extent of HDX are sensitive to temperature, the majority of peptides showed rapid kinetics of exchange, allowing us to focus on plateau values for the maximal extent of exchange at each temperature. Arrhenius plots of the ratio of hydrogens exchanged at 5 h normalized to the number of exchangeable hydrogens vs. 1/T provides an estimate for the apparent enthalpic change of local unfolding, ΔH°unf(avg). Most regions in the enzyme show ΔH°unf(avg) ≤ 2.0 kcal/mol, close to the value of kT; by contrast, significantly elevated values for ΔH°unf(avg) are observed in regions within the core of protein that contain the cofactor and substrate-binding sites. Our technique introduces a new strategy for probing the temperature dependence of local protein unfolding within native proteins. These findings are discussed in the context of the demonstrated role for nuclear tunneling in hydride transfer from NADPH to dihydrofolate, and relate the observed enthalpic changes to two classes of motion, preorganization and reorganization, that have been proposed to control the efficiency of hydrogenic wave function overlap. Our findings suggest that the enthalpic contribution to the heavy atom environmental reorganizations controlling the hydrogenic wave function overlap will be dominated by regions of the protein proximal to the bound cofactor and substrate.

Keywords: dihydrofolate reductase, hydrogen deuterium exchange, hydrogen tunneling, protein dynamics, thermophilic protein

Dihydrofolate reductase (5,6,7,8-tetrahydrofolate: NADP+ oxidoreductase, EC 1.5.1.3) (DHFR) is a ubiquitous enzyme found in prokaryotes and eukaryotes. It catalyzes the stereospecific transfer of a hydride ion from NADPH cofactor to 7,8-dihydrofolate (DHF) to yield 5,6,7,8-tetrahydrofolate (THF) and oxidized nicotinamide adenine dinucleotide phosphate (NADP+). DHFR is the only source of THF, a one-carbon donor/acceptor unit vital in the biosynthesis of purines, pyrimidines, and amino acids (1). DHFR’s role in maintaining the cellular pools of THF has been exploited in the development of pharmacological targets that inhibit cell growth and proliferation by disrupting DNA synthesis (e.g. anticancer and antibacterial agents) (e.g., refs. 2–4).

The best-studied DHFR is from Escherichia coli (Ec-DHFR), where studies of the hydride transfer step have established the size and temperature dependence of the intrinsic deuterium kinetic isotope effect (KIE) (5, 6). Elegant NMR (7–9) and FRET (10, 11) experiments have further implicated loop closures in catalysis (12), as well as the sampling of higher energy protein substates within each catalytic complex that are characteristic of downstream chemical complexes (13). In a previous study, we presented the kinetic and structural properties of a homologous, monomeric, thermophilic DHFR from Bacillus stearothermophilus (Bs-DHFR) (14), with the goal of comparing its kinetic and dynamical properties to the canonical Ec-DHFR. The ability to access a larger temperature range using a thermophilic variant of DHFR is a distinctive advantage when interrogating the temperature dependence of the hydride transfer step and its relationship to protein motions.

The hydride transfer reaction catalyzed by Ec-DHFR shows a presteady state rate constant of 220 s-1 at neutral pH and 25 °C and rate-limiting product release (12 s-1) (15). The faster rate of hydride transfer is consistent with a steady state deuterium KIE of unity, emphasizing the importance of presteady state studies to characterize the H-transfer step at this pH. Stopped-flow studies have provided a deuterium (H/D) KIE of 2.9 at pH 7, 25 °C, whereas a comparison of H/T and D/T KIEs at the same temperature indicated an intrinsic KIE of 3.5 ± 0.2 (5). Tunneling has been established for Ec-DHFR from the observation that Arrhenius prefactor isotope effects greatly exceed semiclassical limits, with reported values for AH/AD = 4.0(1.5) (5). These parameters, together with energy of activation estimates of 3.7–5.6 kcal/mol (5, 6), implicate a tunneling mechanism that is controlled by the heavy atom motions of the protein.

Perhaps not unexpectedly, based on reported comparisons of other homologous classes of mesophilic and thermophilic proteins (e.g., refs. 16–18), X-ray structures for the Ec- and Bs-DHFRs are nearly identical (14). Stopped-flow studies of the chemical step in Bs-DHFR as a function of temperature yielded an enthalpy of activation of 5.5 kcal/mol and an inverse AH/AD ratio of 0.57; i.e., an increased temperature dependence of the KIE compared to Ec-DHFR. Though both proteins have been concluded to use a tunneling mechanism for hydride transfer, the value of AH/AD for Bs-DHFR is closer to low activity mutants of Ec-DHFR (6, 19) implicating a less optimized active site in the thermophilic enzyme. In light of the high structural similarity between Ec- and Bs-DHFRs, we hypothesize that their observed kinetic differences are likely to arise from subtle differences in protein dynamics.

Hydrogen/deuterium exchange (HDX) offers an excellent probe of the impact of extrinsic parameters on protein flexibility (20, 21), in particular under conditions of EX-2 exchange, which reflects the transient equilibration between closed and open forms of a protein within its native folded structure. Whereas studies have been carried out on native proteins by NMR (22, 23), the use of HDX coupled to mass spectrometric analysis of protein-derived peptides provides complementary information (24). Here, we use hydrogen exchange by mass spectrometry (HDX-MS) to examine protein-breathing motions within segments of Bs-DHFR as a function of temperature. A major finding is that the majority of Bs-DHFR undergoes a local unfolding with small enthalpies of activation that are close to the value of kT; the only regions that show elevated enthalpic barriers are those that bridge the cofactor and substrate-binding sites. These results are relevant to the two classes of protein motion, preorganization and reorganization, that have been proposed to control wave function overlap in enzymatic H-tunneling processes (25, 26).

Results

Sequence Coverage of Bs-DHFR.

Eleven peptides were identified after pepsin proteolysis of Bs-DHFR (Table S1), ranging in length between 7 and 20 residues with an average length of 13 residues, and accounting for 91% coverage of the primary sequence. The peptides are annotated onto the primary sequence and mapped onto the X-ray crystal structure of Bs-DHFR in Fig. 1 A and B.

Fig. 1.

Fig. 1.

Peptic fragments of Bs-DHFR. (A) Eleven nonoverlapping fragments of Bs-DHFR generated after proteolysis by pepsin at pH 2.4. (B) The pepsin-generated fragments are mapped onto the X-ray structure of Bs-DHFR (PDBID:1ZDR), according to colors in (A). (C) X-ray structure of Ec-DHFR, indicating bound dihydrofolate (DHF) (rose) and cofactor NADPH (blue).

Thermal Stability of Bs-DHFR.

Circular dichroism (CD) experiments performed on Bs-DHFR in D2O indicated complete stability at 10 °C for up to 4 h, but significant unfolding after 11 min at 65 °C (Fig. S1 A and B). Thermal stability was further examined by monitoring the CD spectrum at 222 nm between 10 °C and 95 °C (Fig. S1C). Signals between 20–60 °C and 75–95 °C represent native and unfolded states, respectively, whereas the region between 60–75 °C highlights cooperative protein unfolding. A melting temperature of 67 °C was estimated from the transition midpoint, revealing enhanced stability compared to Ec-DHFR whose melting temperature was reported as ca. 50 °C (27). In this manner, 55 °C was established as the upper temperature limit for extended incubation of Bs-DHFR in D2O for HDX-MS experiments. The lower temperature limit was 10 °C, to avoid freezing of D2O and the possibility of protein cold denaturation.

Time and Temperature Dependence for HDX in Bs-DHFR.

Bs-DHFR was incubated in D2O between 0 and 5 h at varying temperatures (10, 25, 35, 40, 50, 55 °C) and HDX measurements were made on eleven peptides. Table S2 summarizes the peptides, the maximal number of exchangeable amides (N∞), and the observed deuteration after 10 s and 300 min, at 10 °C and 55 °C. Primary mass spectra are shown for the example of peptide 3 at 10 °C and 55 °C (Fig. S2), illustrating the increased weighted average mass as a function of time of HDX. Similar data quality was observed for all Bs-DHFR-derived peptides, and no peptide showed any evidence of bimodal patterns that would be indicative of global protein unfolding.

We define the fractional deuteration as (NT,t/N∞) where NT,t equals the number of deuterons detected in each peptide at temperature T and time t, and N∞ equals the number of exchangeable amides that should be deuterated at infinite time. From the fractional deuteration (parentheses in Table S2) we calculate the values averaged over all peptides at 10 °C and 55 °C. At 10 s, the averaged fractional deuteration across the protein was 28% at 10 °C, increasing to 44% at 55 °C. At 300 min, the averaged deuteration was 62% at 10 °C increasing to 87% at 55 °C. In contrast, HDX measurements of full-length apo Ec-DHFR showed deuteration at 85% of amides by 30 min at 15 °C (28). Thus, it is clear that in Bs-DHFR, the temperature must be elevated to 55 °C to approach the HDX behavior seen in Ec-DHFR at 15 °C, indicating the lower conformational mobility for the Bs-DHFR at equivalent temperatures.

DHFR consists of two subdomains (adenosine-binding subdomain and loop subdomain), each of which behaves as a rigid body. The adenosine-binding subdomain is formed from strands βB, βC, βD, and βE and helices αC and αE with the intervening CD loop, and provides the binding site for the adenosine moiety of the cofactor NADPH (Fig. 1C). The loop subdomain consists of strands βA, βF, βG, and βH and helices αB and αF with loops M20, FG, and GH. The nicotinamide moiety of cofactor binds to a deep pocket formed by αF, βA, and M20, whereas the pteridine ring of DHF/THF interacts with residues in a cleft formed by αB, αC, and βA. Thus, strand βA bridges the cofactor and substrate, bringing these reactants in close proximity to allow hydride transfer via nuclear tunneling (Fig. 1C). Much structural evidence shows that the M20 loop adopts open, closed, and occluded conformations, respectively associated with cofactor binding, catalysis, and product release (8, 9, 29). Thus, the dynamics of the M20 loop is associated with catalytic turnover, and its closed conformation in the ternary complex, which shields reactants from solvent, is stabilized by interactions with FG loop residues.

Time courses for HDX at each temperature are presented in Fig. 2, and fitted by nonlinear least squares to a two-exponential equation (see Eq. S3 in SI Text) to determine best fit parameters [A, k1, B, k2 (Table S3)]. Due to the rapidity of exchange, the values for k1 and k2 appear random, whereas the applitudes A and B indicate a regular trend. Deuteration, which plateaued at 5 h in all cases, was used to determine the extent of HDX in localized regions of the protein (Table S2) and mapped against the backbone structure of Bs-DHFR (Fig. 3). At 10 °C (Fig. 3A), extensive HDX was observed throughout the adenosine-binding subdomain including the CD loop (peptides 4–6), reaching 80–91% deuteration. Extensive HDX was also observed in helix αB (peptide 3), which normally is associated with rigid body movements in the loop subdomain. Thus, the entire adenosine-binding subdomain and the αB helix of the loop subdomain are highly accessible to exchange with solvent, suggesting significant backbone flexibility within secondary structure elements in these regions. The remainder of the loop subdomain showed lower HDX, reaching 37–66% at 5 h, 10 °C. This included surface loops M20, FG and GH (peptides 2, 9, 10), which in X-ray structures show high exposure to solvent. The significant protection from exchange in the loop subdomain suggests interactions within the secondary structures that confer local rigidity in the absence of bound substrates. Finally, pronounced protection from HDX was observed in the βA strand (peptide 1), which reached < 10% deuteration after 5 h at 10 °C, consistent with its location within the protein core. Taken together, the findings show that βA forms a solvent-inaccessible core and suggest that the conformational mobility of helix αB is linked more closely to the adenosine-binding subdomain, despite its structural association with the loop subdomain.

Fig. 2.

Fig. 2.

HDX time courses of Bs-DHFR peptides as a function of temperature. Extent of deuteration in peptides of Bs-DHFR, after incubating protein in D2O for 10 s–5 h at 10 (green), 25 (orange), 35 (black), 40 (purple), 50 (red), and 55 (blue). The top of each graph indicates the maximal number of exchangeable amides corrected to 100% D2O.

Fig. 3.

Fig. 3.

Regional variations in HDX temperature effects. (A) HDX at t = 300 min and 10 °C shows high solvent exchange in the adenosine-binding subdomain and αB of Bs-DHFR. In contrast, most of the loop subdomain including M20, FG and GH loops shows intermediate accessibility to solvent, and there is strong protection from exchange in the protein core (βA and βF). (B) At 55 °C, deuteration increases, reaching maximal levels in parts of the loop subdomain, although still submaximal in the M20, FG, and GH loops. The largest temperature dependence of deuteration occurs in βA and αF.

The effect of temperature on the HDX behavior of Bs-DHFR was next evaluated over different regions of the enzyme. The extent of hydrogen exchange increased throughout the molecule with temperature, reaching 90–100% at 55 °C in the adenosine-binding subdomain and αB (peptides 3–6), as well as the loop between αF–βF and strand βH in the loop subdomain (peptides 8, 11) (Fig. 3B). In other parts of the loop subdomain, deuteration reached submaximal levels of 66–80% (peptides 1, 2, 7, 9, 10). Deuteration of peptides 1 and 8 increased by 3-fold or more between 10 °C and 55 °C, whereas temperature had much lower effects on the rest of the molecule (peptides 2, 4, 7, 9, 10, and 11). Effects of temperature on peptides 3, 5, and 6 were difficult to evaluate due to limitations in dynamic range, because these reached near maximal deuteration at the lowest temperatures. Taken together, the results revealed distinct variations in temperature effects in different regions of Bs-DHFR, indicating regional differences in the enthalpy of hydrogen exchange.

Determination of ΔH°unf(avg) for Temperature-Dependent Changes in Local Protein Flexibility/Unfolding.

In an EX-2 exchange experiment, the rate of exchange is represented as the product of the equilibrium constant for formation of a transiently open state (Eopen), Kopen = kopen/kclosed, multiplied by the intrinsic rate constant for exchange for each amide hydrogen, kexch:

graphic file with name pnas.1003678107eq4.jpg [1]

Raising the temperature would be expected to increase both Kopen and kexch, with previous studies indicating a ΔH‡ for kexch of ca. 17 kcal/mol (20). Because Kopen is quite unfavorable, with an anticipated ΔH° that is positive, the net ΔH‡ for kobs is anticipated to be in excess of 17 kcal/mol. Such an analysis can be applied either to individual residues, via NMR, or to the aggregate of the residues within a given region of protein via HDX-MS. Importantly, these individual analyses of HDX assume the same physical state for Eopen, with temperature affecting Kopen and kexch, but not the nature of Eopen.

Our data, however, reveal an additional process that leads to measurable changes in the structure of Eopen. In virtually every instance, the time courses reached distinctive plateau values at 5 h, implying populated states with different endpoints for deuterium exchange at different temperatures. Such behavior was reminiscent of the temperature-dependent fraying of H-bonds within DNA duplexes (30–33) and offered a means of examining the local fraying/unfolding of proteins as a function of temperature. As illustrated in Scheme 1, raising the temperature would be expected to break an increasing number of noncovalent interactions in certain regions of the protein. This is distinct from simple transient fluctuations from closed to open states, because the open state in these studies shows different properties at each temperature.

Scheme 1.

Scheme 1.

Schematic illustrating the temperature-dependent disruption of local secondary structures within a native folded protein. According to this model, temperature affects not only the equilibrium constant, Kopen, but also the structure of the opened state of protein. T1 is the lowest initial temperature and T2, an elevated temperature.

To analyze this behavior, we defined an apparent ΔH°unf(avg) for local unfolding, which represents the average property of the Inline graphic at each site within a given peptide. For a peptide of length i amino acids, each residue provides an energetic contribution to ΔH°unf with the total apparent enthalpy of a peptide defined as shown in Eq. 2 where ni represents the amide hydrogen that exchanges at residue i and nT is the sum of all n’s from n1 to ni*:

graphic file with name pnas.1003678107eq5.jpg [2]

The magnitude of the observed enthalpy ΔH°unf(avg) for a given peptide will reflect the behavior of its component residues. For instance, if a peptide is composed of residues that are almost completely exposed to solvent within each excursion from Eclosed to Eopen, the in-exchange at various temperatures will increase, but converge toward a similar value. On the other hand, if a peptide is in a folded or structurally stable region, only a fraction of its residues will display in-exchange at very long times within a given temperature regime. As the degree of local unfolding increases with temperature, a new and distinct value for maximal in-exchange becomes apparent, reflecting the disruption of a greater number of interactions within the region of the protein that is being monitored in the representative, isolated peptides.

To estimate values for ΔH°unf(avg) among the various peptides of Bs-DHFR, we first normalized the observed number of exchanged amide hydrogens to the total exchangeable amides, yielding Nt=300 min/N∞. Plots of ln (Nt=300 min/N∞) as a function of 1/Temperature (K-1) for each of the peptides are given in Fig. 4. Two classes of behavior were discerned in which the majority of the peptides (2–7, 9–11) showed relatively small changes in deuteration. For most peptides values for ΔH°unf(avg) were < 2 kcal/mol (Fig. 4 and Table S4), in comparison to an energy of ca. 0.7 kcal/mol at the functional temperature (ca. 70 °C) for B. stearothermophilus (14). Peptides 10 and 11 are represented by somewhat elevated values for ΔH°unf(avg), with an average value of 2 kcal/mol. For peptides 3, 5, and 6, the slope was clearly low but the high degree of error precluded accurate measurements. The implication is that regions in the adenosine-binding subdomain and most of the loop subdomain, including the M20 and FG loops, show low ΔH°unf(avg), reflecting facile changes in local flexibility.

Fig. 4.

Fig. 4.

Regional variations in the enthalpy of HDX. Plots of ln (Nt,t=300 min/N∞) vs. 1/T show the linear dependence, with Inline graphic indicated in the frame of each peptide.

The important exceptions were peptides 1 and 8, representing βA and αF. These regions, represented by stiffer behavior and resistance to local changes in protein flexibility at low temperatures, showed significantly larger effects of temperature and higher estimates of ΔH°unf(avg)†. In every instance, enthalpies for local unfolding estimated from HDX measurements were far below the ΔH° for global unfolding of Ec-DHFR (34), which is likely to set the lower limit for the global unfolding of the more stable Bs-DHFR.

Discussion

A major challenge in enzymology is our ability to relate the conformational landscape sampled by a protein to its catalytic function. Hydrogen tunneling has played a major role in linking protein motions to chemical catalysis (25, 26). In this study, our methodology of measuring HDX against varying temperatures on a thermophilic enzyme provides a means of assessing values for the enthalpic barriers controlling local unfolding within native proteins. Applying the strategy to Bs-DHFR reveals that regions which support and bridge the cofactor/substrate-binding sites within the protein core are characterized by elevated values for ΔH°unf(avg) that bracket ΔH‡ for catalysis, in contrast to the small values for ΔH°unf(avg) across the rest of the protein.

The enthalpic barrier controlling protein motions in Bs-DHFR is inherently related to the heavy atom barriers that control the efficiency of hydride transfer from NADPH to DHF. Hydrogen tunneling efficiency can be represented in a modified Marcus equation (termed reorganization) and consists of a vertical coordinate that determines the energy needed to achieve transient degeneracy between reactant and product, together with a horizontal coordinate that represents the adjustment of the distance between the donor and acceptor atoms to one that permits efficient tunneling (ca. 2.7–2.8 Å), (Eq. 3 and refs. 35 and 36).

graphic file with name pnas.1003678107eq6.jpg [3]

Although formally derived in the context of nonadiabatic H-transfer processes, Eq. 3 illustrates the physical parameters that are expected to control hydrogenic wave function overlap. The vertical coordinate is strictly analogous to the Marcus expression for electron tunneling and is represented in the first exponential term of Eq. 3; this contains the sum of the inner sphere and outer sphere environmental reorganization (λ) and the reaction driving force (ΔG°). This first exponential, which will generally be strongly temperature-dependent but only weakly isotope-dependent, may be expected to dominate the enthalpic barrier to catalysis. The horizontal component, shown in the second and third exponentials of Eq 3, is defined by the integration of a Frank–Condon wave function overlap term, dependent on the mass (mH), frequency (ωH), and distance traveled (rH) by the tunneling particle, over a family of H-donor/acceptor distances. The latter, third exponential represents an oscillator whose energy (Ex) is dependent on the initial distance between the H-donor and acceptor well as the mass and frequency of the vibrating (distance sampling) protein unit (35, 36). For a tunneling particle as large as hydrogen, the distance between the H-donor and acceptor can play an important role in wave function overlap, with relatively small changes in the distance between the donor and acceptor atoms having a significant impact on tunneling efficiency. In the course of distance sampling, the reaction barrier is greater for D- than H-transfer, a consequence of the twofold greater mass of the heavy isotope (and its attendant smaller wavelength). This horizontal coordinate, thus, becomes the determinant of the temperature dependence of the KIE.

One property that has emerged among a very large number of distinct enzyme classes is a kinetic isotope effect that shows little or no temperature dependence; i.e., an Ea(D) ≃ Ea(H) (25, 26). This unexpected property implies a protein active site in which the hydrogen donor and acceptor can approach one another at a distance of ca. 2.7–2.8 Å (37), ca. 0.5 Å shorter than their ground state van der Waals distance. The achievement of such short internuclear distances implies that another type of motion must interface with the barrier introduced by the reorganization term, attributed to extensive sampling of the protein conformational landscape to access the catalytically relevant conformational substates (termed preorganization) (25, 26).

Previous studies of a thermophilic alcohol dehydrogenase, ht-ADH, have used HDX to yield insight into the relationship of such conformational sampling motions to the reorganization barrier (38, 39). HDX and kinetic measurements on ht-ADH revealed a temperature transition at 30 °C, with the following properties: In the high-temperature regime, the protein showed increased flexibility accompanied by a temperature-independent KIE for hydrogen transfer, whereas in the low-temperature regime, ht-ADH underwent a rigidification, which was accompanied by an increase in the temperature dependence of the KIE. These studies clearly demonstrated a requirement for long-range protein flexibility to achieve a very close approach between active site H-donor and acceptor. They also support a general picture for enzyme-mediated hydrogen tunneling (25, 26) that involves extensive sampling among protein conformational substates with only a small family of substates capable of achieving the very close distances that will support efficient H-tunneling. For native proteins under optimal conditions, the portion of the activation barrier (ΔG‡ = ΔH‡ - TΔS‡) that dominates this preorganization process is expected to be TΔS‡, representing the probability that a given protein can achieve the family of catalytically enabled conformers.

Our current studies with thermophilic DHFR provide a unique way to linking the energetic barriers that control preorganization and reorganization to the hydride transfer step. Although Bs-DHFR does not display a temperature break in HDX or catalytic properties, as observed for ht-ADH, the HDX measurements provide an ability to estimate apparent enthalpic barriers for changes in local protein flexibility that can be related to enzyme function. The fact that the majority of the peptides interrogated undergo local unfolding with values for ΔH°unf(avg) ≤ 2 kcal/mol conforms to the expectation of a smooth conformational landscape that prevents the trapping of protein into local energy minima. The contrasting finding of elevated values for ΔH°unf(avg) that are restricted to regions of DHFR that bridge the substrate and cofactor binding sites can be interpreted in the context of the contribution of the reorganization parameter, λ, to Eq 3. We propose that the major portion of this barrier for Bs-DHFR derives from a requisite transient restructuring of the protein’s catalytic core, with the catalytic enthalpic barrier of ΔH‡ = 5.5 kcal/mol falling between the values of ΔH°unf(avg) of 3.8 (0.6) and 8.9 (1.2) observed in peptides 8 and 1, respectively.

One aspect of these studies is their focus on the apo-form of Bs-DHFR. This was done with the recognition that the binding of NADPH and DHF to Bs-DHFR would afford preferential protection against HDX within the core regions of protein, limiting interpretation of positional differences in protein flexibility. Dynamical studies of other proteins have indicated the presence of similar conformational substates both in the presence and absence of protein ligands, with ligands functioning to shift equilibria rather than to generate unique states (40, 41). The application of HDX-MS to native state carboxypeptidase B has been reported to lead to undistinguishable exchange when carried out in the presence or absence of the substrate hippuryl-Arg (42). For the future, it would be of interest to quantify the impact of bound cofactor and substrate on the apparent enthalpies of local unfolding in the peripheral regions of Bs-DHFR, as well as the temperature dependence of HDX-MS in other H-transfer enzymes to see if similar patterns for regional protein unfolding can be discerned.

The model for Bs-DHFR presented is distinct from previous studies of Ec-DHFR, which have proposed a role for motions in the M20 and FG loops as the dominant source of the protein reorganization barrier (7, 8, 43–45). In our model of Bs-DHFR catalyzed H-transfer, we attribute the strong temperature-dependent motions needed for reorganization to those found in βA and αF, whereas the motions within the M20 and FG loops that contribute to hydrogen tunneling are ascribed to a low enthalpy conformational sampling process that contributes dominantly to the preorganization process. Maintenance of low enthalpic differences among substates ensures a smooth conformational landscape that prevents the trapping of protein into catalytically nonproductive energy wells (45). Our findings suggest that the detection of heavy atom motions in the M20 loop of Ec-DHFR, which occur with a rate constant similar to that for catalysis, may represent the accommodation of the protein to downstream intermediate and product complexes (13, 40), in contrast to functioning as the causative heavy atom motions that control the hydride transfer converting bound NADPH/DHF to NADP+/THF.

Materials and Methods

Cloning and Enzyme Purification.

Bs-DHFR was cloned, expressed, purified, and assayed as previously described (14) with minor modifications. Further details are available in SI Text, as are the protocols for circular dichroism spectroscopy.

Hydrogen Exchange (HDX-MS) Measurements.

The general protocols for HDX-MS on protein-derived peptides have been described (compare refs. 38 and 46). The data identifying and confirming the peptide studies are in Table S1, whereas the more particular experimental conditions are in SI Text.

Supplementary Material

Supporting Information

Acknowledgments.

We thank Dr. S. Sharma for assistance with data fitting, Prof. D. Wemmer (University of California, Berkeley) and Prof. C. Perrin (University of California, San Diego) for valuable discussions, and Mae Tulfo for technical assistance. This work was supported by National Institutes of Health Grant GM039296 and National Science Foundation Grant MCB0446395 (J.P.K.), National Institutes of Health Grant GM074134 (N.G.A.), and National Institutes of Health Training Grant T32 GM08295-15 (O.A.O.).

Footnotes

The authors declare no conflict of interest.

*An alternative analysis to Eq. 2 would be to normalize the weighted enthalpy terms by the number of hydrogens actually exchanged within the experimental temperature range:

graphic file with name pnas.1003678107eq7.jpg

However, because nT is very close to the Inline graphic [compare (A + B) to N in Table S3] this analysis yields almost identical results to Eq 2 with greater error.

†The probability of being in a locally unfolded state at 70 °C can be estimated from P = e-ΔHunf/kT, yielding a value of ca. 6% when ΔH°unf = 2 kcal/mol vs. ca. 1 × 10-5% when ΔH°unf = 9 kcal/mol.

References

  • 1.Blakley R. In: Folates and Pterins. Blakley RA, Benkovic SJ, editors. New York: Wiley; 1984. pp. 191–253. [Google Scholar]
  • 2.Peppard WJ, Schuenke CD. Iclaprim, a diaminopyrimidine dihydrofolate reductase inhibitor for the potential treatment of antibiotic-resistant staphylococcal infections. Curr Opin Invest Dr. 2008;9:210–225. [PubMed] [Google Scholar]
  • 3.Gangjee A, Kurup S, Namjoshi O. Dihydrofolate reductase as a target for chemotherapy in parasites. Curr Pharm Design. 2007;13:609–639. doi: 10.2174/138161207780162827. [DOI] [PubMed] [Google Scholar]
  • 4.Soeiro MNC, de Castro SL. Trypanosoma cruzi targets for new chemotherapeutic approaches. Expert Opin Ther Tar. 2009;13:105–121. doi: 10.1517/14728220802623881. [DOI] [PubMed] [Google Scholar]
  • 5.Sikorski RS, et al. Tunneling and coupled motion in the E. coli dihydrofolate reductase catalysis. J Am Chem Soc. 2004;126:4778–4779. doi: 10.1021/ja031683w. [DOI] [PubMed] [Google Scholar]
  • 6.Wang L, Goodey NM, Benkovic SJ, Kohen A. Coordinated effects of distal mutations on environmentally coupled tunneling in dihydrofolate reductase. Proc Natl Acad Sci USA. 2006;103:15753–15758. doi: 10.1073/pnas.0606976103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Osborne MJ, Schnell J, Benkovic SJ, Dyson HJ, Wright PE. Backbone dynamics in dihydrofolate reductase complexes: Role of loop flexibility in the catalytic mechanism. Biochemistry. 2001;40:9846–9859. doi: 10.1021/bi010621k. [DOI] [PubMed] [Google Scholar]
  • 8.Venkitakrishnan RP, et al. Conformational changes in the active site loops of dihydrofolate reductase during the catalytic cycle. Biochemistry. 2004;43:16046–16055. doi: 10.1021/bi048119y. [DOI] [PubMed] [Google Scholar]
  • 9.Boehr DD, Dyson HJ, Wright PE. Conformational relaxation following hydride transfer plays a limiting role in dihydrofolate reductase catalysis. Biochemistry. 2008;35:9227–9233. doi: 10.1021/bi801102e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Zhang ZQ, Rajagopalan PTR, Selzer T, Benkovic SJ, Hammes GG. Single-molecule and transient kinetics investigation of the interaction of dihydrofolate reductase with NADPH and dihydrofolate. Proc Natl Acad Sci USA. 2004;101:2764–2769. doi: 10.1073/pnas.0400091101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Antikainen NM, Smiley RD, Benkovic SJ, Hammes GG. Conformation coupled enzyme catalysis: Single-molecule and transient kinetics investigation of dihydrofolate reductase. Biochemistry. 2005;44:16835–16843. doi: 10.1021/bi051378i. [DOI] [PubMed] [Google Scholar]
  • 12.Falzone CJ, Wright PE, Benkovic SJ. Dynamics of a flexible loop in dihydrofolate reductase from E. coli and its implication for catalysis. Biochemistry. 1994;33:439–442. doi: 10.1021/bi00168a007. [DOI] [PubMed] [Google Scholar]
  • 13.Boehr DD, McElheny D, Dyson HJ, Wright PE. The dynamic energy landscape of dihydrofolate reductase catalysis. Science. 2006;313:1638–1642. doi: 10.1126/science.1130258. [DOI] [PubMed] [Google Scholar]
  • 14.Kim HS, Damo SM, Lee SY, Wemmer D, Klinman JP. Structure and hydride transfer mechanism of a moderate thermophilic dihydrofolate reductase from B. stearothermophilus and comparison to its mesophilic and hyperthermophilic homologues. Biochemistry. 2005;44:11428–11439. doi: 10.1021/bi050630j. [DOI] [PubMed] [Google Scholar]
  • 15.Fierke CA, Kuchta RD, Johnson KA, Benkovic SJ. Implications for enzymic catalysis from free-energy reaction coordinate profiles. Cold Spring Harb Sym. 1987;52:631–638. doi: 10.1101/sqb.1987.052.01.072. [DOI] [PubMed] [Google Scholar]
  • 16.Wintrode PL, Zhang DQ, Vaidehi N, Arnold FH, Goddard WA. Protein dynamics in a family of laboratory evolved thermophilic enzymes. J Mol Biol. 2003;327:745–757. doi: 10.1016/s0022-2836(03)00147-5. [DOI] [PubMed] [Google Scholar]
  • 17.Wallon G, et al. Crystal structures of E. coli and Salmonella typhimurium 3-isopropylmalate dehydrogenase and comparison with their thermophilic counterpart from Thermus thermophilus. J Mol Biol. 1997;266:1016–1031. doi: 10.1006/jmbi.1996.0797. [DOI] [PubMed] [Google Scholar]
  • 18.Britton KL, et al. Structure determination of the glutamate dehydrogenase from the hyperthermophile Thermococcus litoralis and its comparison with that from Pyrococcus furiosus. J Mol Biol. 1999;293:1121–1132. doi: 10.1006/jmbi.1999.3205. [DOI] [PubMed] [Google Scholar]
  • 19.Wang L, Goodey NM, Benkovic SJ, Kohen A. The role of enzyme dynamics and tunnelling in catalysing hydride transfer: Studies of distal mutants of dihydrofolate reductase. Philos T R Soc B. 2006;361:1307–1315. doi: 10.1098/rstb.2006.1871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Bai Y, Milne JS, Mayne L, Englander SW. Primary structure effects on peptide group hydrogen exchange. Proteins. 1993;17:75–86. doi: 10.1002/prot.340170110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Englander SW, Downer NW, Teitelba H. Hydrogen exchange: The modern legacy of Linderstrom–Lang. Protein Sci. 1997;6:1101–1109. doi: 10.1002/pro.5560060517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Bai Y, Sosnick TR, Mayne L, Englander SW. Protein folding intermediates: Native-state hydrogen exchange. Science. 1995;269:192–196. doi: 10.1126/science.7618079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Chamberlain AK, Handel TM, Marqusee S. Detection of rare partially folded molecules in equilibrium with the native conformation of RNaseH. Nat Struct Biol. 1996;3:782–787. doi: 10.1038/nsb0996-782. [DOI] [PubMed] [Google Scholar]
  • 24.Hoofnagle AN, Resing KA, Ahn NG. Protein analysis by hydrogen exchange mass spectrometry. Annu Rev Bioph Biom. 2003;32:1–25. doi: 10.1146/annurev.biophys.32.110601.142417. [DOI] [PubMed] [Google Scholar]
  • 25.Klinman JP. An integrated model for enzyme catalysis emerges from studies of hydrogen tunneling. Chem Phys Lett. 2009;471:179–193. doi: 10.1016/j.cplett.2009.01.038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Nagel ZD, Klinman JP. A 21st century revisionist's view at a turning point in enzymology. Nat Chem Biol. 2009;5:543–551. doi: 10.1038/nchembio.204. [DOI] [PubMed] [Google Scholar]
  • 27.Swanwick RS, Daines AM, Tey LH, Flitsch SL, Allemann RK. Increased thermal stability of site-selectivity glycosylated dihydrofolate reductase. Chem Biochem. 2005;6:1338–1340. doi: 10.1002/cbic.200500103. [DOI] [PubMed] [Google Scholar]
  • 28.Yamamoto T, Izumi S, Gekko K. Mass spectrometry of hydrogen/deuterium exchange of E. coli dihydrofolate reductase: effects of loop mutations. J Biochem-Tokyo. 2004;135:487–94. doi: 10.1093/jb/mvh056. [DOI] [PubMed] [Google Scholar]
  • 29.Sawaya MR, Kraut J. Loop and subdomain movements in the mechanism of E. coli dihydrofolate reductase: Crystallographic evidence. Biochemistry. 1997;36:586–603. doi: 10.1021/bi962337c. [DOI] [PubMed] [Google Scholar]
  • 30.Danell AS, Parks JH. Fraying and electron autodetachment dynamics of trapped gas phase oligonucleotides. J Am Soc Mass Spectr. 2003;14:1330–1339. doi: 10.1016/S1044-0305(03)00578-6. [DOI] [PubMed] [Google Scholar]
  • 31.Milev S, et al. Energetics of sequence-specific protein-DNA association: Conformational stability of the DNA-binding domain of integrase Tn916 and its cognate DNA duplex. Biochemistry. 2003;42:3492–3502. doi: 10.1021/bi026936x. [DOI] [PubMed] [Google Scholar]
  • 32.Lam SL, Ip LN. Low temperature solution structures and base pair stacking of double helical d(CGTACG)(2) J Biomol Struct Dyn. 2002;19:907–917. doi: 10.1080/07391102.2002.10506793. [DOI] [PubMed] [Google Scholar]
  • 33.Hochstrasser RA, Carver TE, Sowers LC, Millar DP. Melting of a DNA helix terminus within the active-site of a DNA-polymerase. Biochemistry. 1994;33:11971–11979. doi: 10.1021/bi00205a036. [DOI] [PubMed] [Google Scholar]
  • 34.Luo J, Iwakura M, Matthews CR. Detection of a stable intermediate in the thermal unfolding of a cysteine-free form of dihydrofolate reductase from Escherichia coli. Biochemistry. 1995;34:10669–10675. doi: 10.1021/bi00033a043. [DOI] [PubMed] [Google Scholar]
  • 35.Knapp MJ, Klinman JP. Environmentally coupled hydrogen tunneling: Linking catalysis to dynamics. Eur J Biochem. 2002;269:3113–3121. doi: 10.1046/j.1432-1033.2002.03022.x. [DOI] [PubMed] [Google Scholar]
  • 36.Kuznetsov AM, Ulstrup J. Proton and hydrogen atom tunneling in hydrolytic and redox enzyme catalysis. Can J Chem. 1999;77:1085–1096. [Google Scholar]
  • 37.Hammes-Schiffer S. Comparison of hydride, hydrogen atom, and proton-coupled electron transfer reactions. Chemphyschem. 2002;3:33–42. doi: 10.1002/1439-7641(20020118)3:1<33::AID-CPHC33>3.0.CO;2-6. [DOI] [PubMed] [Google Scholar]
  • 38.Liang ZX. Thermal-activated protein mobility and its correlation with catalysis in thermophilic alcohol dehydrogenase. Proc Natl Acad Sci USA. 2004;101:9556–9561. doi: 10.1073/pnas.0403337101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Nagel ZD, Klinman JP. Tunneling and dynamics in enzymatic hydride transfer. Chem Rev. 2006;106:3095–3118. doi: 10.1021/cr050301x. [DOI] [PubMed] [Google Scholar]
  • 40.Eisenmesser EZ, et al. Intrinsic dynamics of an enzyme underlies catalysis. Nature. 2005;438:117–121. doi: 10.1038/nature04105. [DOI] [PubMed] [Google Scholar]
  • 41.Hanson JA, et al. Illuminating the mechanistic roles of enzyme conformational dynamics. Proc Natl Acad Sci USA. 2007;104:18055–18060. doi: 10.1073/pnas.0708600104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Liu Y-H, De La Rosa M. Enzyme conformational dynamics during catalysis and in the “resting” state monitored by hydrogen/deuterium exchange mass spectrometry. FEBS Lett. 2006;580:5137–5142. doi: 10.1016/j.febslet.2006.08.042. [DOI] [PubMed] [Google Scholar]
  • 43.Agarwal PK, Billeter SR, Rajagopalan PT, Benkovic SJ, Hammes-Schiffer S. Network of coupled promoting motions in enzyme catalysis. Proc Natl Acad Sci USA. 2002;99:2794–2799. doi: 10.1073/pnas.052005999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Wong KF, Selzer T, Benkovic SJ, Hammes-Schiffer S. Impact of distal mutations on the network of coupled motions correlated to hydride transfer in dihydrofolate reductase. Proc Natl Acad Sci USA. 2005;102:6807–6812. doi: 10.1073/pnas.0408343102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Arora K, Brooks CL. Functionally important conformations of the Met20 Loop in dihydrofolate reductase are populated by rapid thermal fluctuations. J Am Chem Soc. 2009;131:5642–5647. doi: 10.1021/ja9000135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Resing KA, Hoofnagle AN, Ahn NG. Modeling deuterium exchange behavior of ERK2 using pepsin mapping to probe secondary struture. J Am Soc Mass Spectr. 1999;10:685–702. doi: 10.1016/S1044-0305(99)00037-9. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information