A Rapid Analysis of Variations in Conformational Behavior during Dihydrofolate Reductase Catalysis

: Protein ﬂ exibility is central to enzyme catalysis, yet it remains challenging both to predict conformational behavior on the basis of analysis of amino acid sequence and protein structure and to provide the necessary breadth of experimental support to any such predictions. Here a generic and rapid procedure for identifying conformational changes during dihydrofolate reductase (DHFR) catalysis is described. Using DHFR from Escherichia coli (EcDHFR), selective side-chain 13 C labeling of methionine and tryptophan residues is shown to be su ﬃ cient to detect the closed-to-occluded conformational transition that follows the chemical step in the catalytic cycle, with clear chemical shift perturbations found for both methionine methyl and tryptophan indole groups. In contrast, no such perturbations are seen for the DHFR from the psychrophile Moritella profunda , where the equivalent conformational change is absent. Like EcDHFR, Salmonella enterica DHFR shows experimental evidence of a large-scale conformational change following hydride transfer that relies on conservation of a key hydrogen bonding interaction between the M20 and GH loops, directly comparable to the closed-to-occluded conformational change observed in EcDHFR. For the hyperthermophile Thermotoga maritima , no chemical shift perturbations were observed, suggesting that no major conformational change occurs during the catalytic cycle. In spite of their conserved tertiary structures, DHFRs display variations in conformational sampling that occurs concurrently with catalysis.

P rotein motions play important roles in enzyme catalysis, and they are often necessary for progression through the catalytic cycle.Originally postulated more than half a century ago in Koshland's "induced fit" model, 1 conformational changes have only more recently been shown to be widely important for substrate binding and product release and to provide an optimal environment for catalysis. 2Indeed, multiple enzyme families that display conserved motions during catalysis have been discovered, which indicates evolutionary pressures on both enzyme tertiary structures and the associated dynamics. 3,4−11 Understanding this diversity of protein motions will in principle help to reveal their importance for the progression through the catalytic cycle and the influence of motions on catalytic proficiency.
DHFRs catalyze transfer of the pro-R hydride of reduced nicotinamide adenine dinucleotide phosphate (NADPH) to 7,8-dihydrofolate (DHF) to form 5,6,7,8-tetrahydrofolate (THF).THF acts as a one-carbon shuttle in the metabolism of purines, thymidylate, and certain amino acids and as a result is critical for cellular proliferation.As DHFR is paramount in maintaining the intracellular pool of THF, it has been a longstanding target for both antibacterial and antineoplastic drugs. 12he canonical DHFR catalytic cycle, derived from studies of Escherichia coli DHFR (EcDHFR), exhibits five intermediate complexes 13 that adopt two distinct conformations, namely, closed and occluded (Figures 1 and 2). 7EcDHFR contains three mobile loops within its structure: the highly mobile M20 loop (residues 9−24 by EcDHFR numbering), the FG loop (residues 116−132), and the GH loop (residues 142−149). 7he holoenzyme (E:NADPH) first binds DHF to form the Michaelis complex (E:NADPH:DHF).In both these intermediates, a network of hydrogen bonds between residues in the M20 and FG loops stabilizes a closed conformation, 7 in which the enzyme possesses the optimal electrostatic environment for hydride transfer. 7,14After product formation, these interactions are disrupted and new hydrogen bonds are established between the M20 and GH loops, stabilizing an occluded conformation, which is conserved in the three remaining intermediates (E:NADP + :THF, E:THF, and E:THF:NADPH) of the catalytic cycle.Release of NADP + and rebinding of NADPH precede the release of THF from the active site, which at pH 7 is the ratelimiting step of the reaction. 7Interchange between the closed and occluded conformations has been shown to modulate the affinity for the nicotinamide cofactor and to be important for product release. 7,13,15ariations in the conformational changes seen in EcDHFR have been identified in other DHFR homologues.For the DHFR from the psychrophile Moritella prof unda (MpDHFR), which shares 55% sequence identity with EcDHFR, 16 the catalytic cycle progresses without the enzyme adopting the occluded conformation following hydride transfer, 10 even though the tertiary structures of EcDHFR and MpDHFR are nearly superimposable (Figure 2). 7,17Crucially, hydrogen bonding between Ser148 and Asn23 is important for stabilizing the occluded conformation; it is abolished in MpDHFR because of a proline residue in place of Ser148. 10Nonetheless, kinetically MpDHFR follows a catalytic cycle similar to that of EcDHFR, where release of product is the rate-limiting step under steady state conditions at pH 7. 17 The homodimeric DHFR from the hyperthermophile Thermotoga maritima (TmDHFR) exhibits a catalytic activity much lower than that of EcDHFR. 6,18,19Residues corresponding to the FG loop in TmDHFR form part of the dimer interface, and TmDHFR is thought to proceed through the catalytic cycle in an open conformation, although fully conclusive experimental evidence is still lacking. 6,18,19hylogenetic analysis indicates that among bacterial DHFRs there is diversity in whether the hydrogen bond donor and acceptor necessary to permit formation of the occluded conformation are present. 20Meanwhile, for mammalian DHFRs, a proline rich region (PWPP) is inserted into the M20 loop, resulting in diminished flexibility. 11,21,22This has been demonstrated to affect the stability of the occluded conformation, and the N23PP and N23PP/S148A variants of EcDHFR remain in the closed conformation after the chemical step and show reduced catalytic activity (diminished k cat ), with dissociation of NADP + becoming rate-limiting. 21In contrast, EcDHFR-S148P cannot form an occluded conformation and displays an increased level of product inhibition, but the rate of hydride transfer is not significantly different from that of wildtype EcDHFR. 10Together, these results indicate that while the occluded conformation has little influence on hydride transfer, it is important for modulating cofactor binding affinities and minimizing product inhibition. 10,21Bacterial DHFRs that conserve the Asn23−Ser148 interaction or possess an equivalent stabilizing interaction should adopt conformations similar to that of the occluded conformation in EcDHFR. 10though a wealth of kinetic and structural data is available for a number of DHFRs, 5 there currently exists no rapid and cost-effective method for assessing enzyme conformational states and loop dynamics for diverse DHFR orthologs.Notably, there are crystal structures of DHFRs from >20 organisms, 20 but these structures rarely describe the entire catalytic cycle and are often determined for apoenzymes or solely with the inhibitors methotrexate and trimethoprim.Moreover, for those DHFR structures that are anticipated to lack the occluded conformation, such as those for the enzymes from Bacillius anthracis 8 and Lactobacillus casei, 9,25 one can infer differences from only the model conformational cycle, because the absence of structural evidence of an occluded conformation does not necessarily imply that the conformation is not adopted.Here we demonstrate that by selective 13 C labeling of methionine and tryptophan side chains only, conformations in solution can be probed rapidly and highly cost-effectively by nuclear magnetic resonance (NMR) spectroscopy at different stages of the catalytic cycle for different DHFR orthologs.The reduced spectral complexity compared to those of more conventional backbone NMR methods aids interpretation and avoids more cumbersome assignment steps.We show that loop motions essential for progression through the catalytic cycle in EcDHFR are not necessary for efficient catalysis in other orthologs [7][8][9]13,25 and present the first clear experimental evidence in solution that TmDHFR remains in an open conformation throughout the catalytic cycle.Furthermore, we show that DHFR from Salmonella enterica (SeDHFR) exhibits conformational behavior similar to that of EcDHFR, in that it adopts an additional conformation after the chemical step. This wok confirms the importance of hydrogen bonding interactions between the M20 and GH loops in stabilizing alternative occluded-like conformations within the product complex and demonstrates that EcDHFR is not unique in its ability to switch between conformations.

■ MATERIALS AND METHODS
Indole (2-13 C) was purchased from Cambridge Isotope Laboratories.NADP + and NADPH were purchased from Melford.DHF was produced by sodium dithionite reduction of folate as described previoulsy. 26All other chemicals, including L-methionine-(methyl-13 C), were from Sigma-Aldrich.DHFRs were grown in minimal medium and prepared as previously reported, 17 with the exception of the addition of either Lmethionine-(methyl-13 C) 27 (80 mg L −1 ) or indole-(2-13 C) 28 (50 mg L −1 )a ta nO D 600 of 0.4.EcDHFR, MpDHFR, and SeDHFR were purified via anion exchange chromatography (Q-Sepharose) followed by size exclusion chromatography. 29or SeDHFR, the procedure was adjusted from that published for EcDHFR and is outlined in Figure S1.TmDHFR was purified as previously described. 19The SeDHFR encoding gene (Uniprot entry P12833) was purchased from Genscript within the pET-11b vector.The SeDHFR-S150A variant was prepared by standard site-directed mutagenesis methods with primers (Fwd, 5′ GCGGACGATAAGAACGCGTATGCGTGCGAG-TTTG 3′; Rev, 5′ CAAACTCGCACGCATACGCGTTCTT-ATCGTCCGC 3′) that were purchased from Sigma-Aldrich and expressed and purified as for wild-type SeDHFR.
All NMR experiments were performed on a Bruker AVANCE III 600 MHz ( 1 H) spectrometer equipped with a QCI-P cryoprobe and final enzyme concentrations of approximately 250 μM.EcDHFR, MpDHFR, and SeDHFR were prepared in 50 mM potassium phosphate buffer (pH 7.0)  7,23,24 containing 1 mM NaCl and 10 mM β-mercaptoethanol.TmDHFR was prepared in 50 mM Tris buffer (pH 7.0) containing 1 mM NaCl and 10 mM β-mercaptoethanol.Ligand stocks were pre-prepared at a concentration of 100 mM and pH 7, using extinction coefficients of 6200 cm −1 M −1 at 339 nm, 30 17800 cm −1 M −1 at 260 nm, 31 and 28000 cm −1 M −1 at 282 nm 32 for NADPH, NADP + , and folate/DHF, respectively.A 10-fold excess of ligands was used in each NMR experiment and the sample adjusted to pH 7.0 before measurement of the spectra, ensuring saturation (Table S1).To form the E:NADP + :THF complex, NADPH and DHF were incubated with the enzyme for 20 min at 37 °C, with reaction completion and the absence of oxidation products confirmed by 1 H NMR. D 2 O (10%) was added to all NMR samples.Spectra for MpDHFR, EcDHFR, SeDHFR, and TmDHFR were recorded at 7, 25, 25, and 40 °C, respectively, to maintain sample integrity and to be close to physiological conditions for each enzyme.Spectra were processed with NMRPipe 33 and analyzed using CcpNmr Analysis 2.4.1. 34ith the exception of Met94, all methionine methyl resonances of MpDHFR had been published previously. 35he assignment of the Met94 resonances was confirmed, and tryptophan indole-δ 1 resonances were assigned by re-evaluation of the three-dimensional (3D) 13 C-and 15 N-edited NOESY spectra acquired previously. 35Each tryptophan indole NH resonance showed an intense NOESY crosspeak to a unique indole-δ 1 1 H resonance observed here, readily allowing assignment.For EcDHFR, methionine methyl resonances and tryptophan indole-δ 1 resonances for EcDHFR were similarly assigned using 3D 13 C-edited NOESY, CCH-TOCSY, and HCCH-TOCSY spectra acquired on a Varian INOVA 800 MHz ( 1 H) spectrometer equipped with a cryogenically cooled HCN probe, using the 3D 15 N-edited NOESY spectrum acquired previously, 36 and by reference to published data. 37he equation Δδ =[ 1 / 2 (δ H 2 + δ C 2 /4)] 1/2 was used to obtain weighted average chemical shift perturbations. 38

■ RESULTS AND DISCUSSION
To investigate the distribution of protein conformations during the catalytic cycle of different DHFR orthologs, EcDHFR and MpDHFR were first probed as models to identify whether systematical labeling of the proteins with [methyl- 13 C]methionine 27 and [indole-δ 1 - 13 C]tryptophan 28 could identify catalytic cycles that include or omit the occluded conformation.TmDHFR and SeDHFR were investigated to elucidate their conformational behavior, which had not been explored previously.Methionine comprises 3.1% of amino acid residues in MpDHFR and EcDHFR, 2.4% in TmDHFR, and 1.8% in SeDHFR, while tryptophan makes up 3.1% of the tryptophan amino acid residues in EcDHFR, 1.9% in MpDHFR and SeDHFR, and 0.6% in TmDHFR.This density ensured that a good description of protein conformational space was achieved without significant spectral crowding (Figure 2 and Figure S2). 6,7,17,18In addition, a number of methionine and tryptophan residues are conserved in structurally significant positions. 6,7,16Most notably, Trp22 in the M20 loop (Figure S2) is found in a wide range of DHFR orthologs and has been shown to be important for EcDHFR catalysis. 39Similarly, Met42 is strongly conserved.In total, four of the five methionine residues in EcDHFR are conserved in MpDHFR (Met1, Met20, Met42, and Met92), and three of the five are conserved in SeDHFR (Met1, Met42, and Met92), although only Met44 is conserved in TmDHFR (Figure S2).The two remaining methionines in TmDHFR are located within the dimer interface.In addition to Trp22, EcDHFR, MpDHFR, and SeDHFR possess a Trp133 equivalent, although other tryptophan residues are well-dispersed throughout the protein structure.Met1 in TmDHFR is mostly removed during expression by an aminopeptidase present in E. coli (Figure S3).S2 and S3) exhibit five distinct cross-peaks corresponding to the five methionine and five tryptophan residues in the presence of ligands.The quality of the spectrum for [methyl- 13 C]methionine-labeled apoenzyme was lower, with multiple minor cross-peaks observed for some residues (Figure S5).This is characteristic of spectra of apo-EcDHFR and apo-MpDHFR where the enzyme is in slow exchange between multiple conformational states. 36,41Addition of ligands (NADP + and folate to model the Michaelis complex or NADP + and THF to form the product ternary complex) resulted in improved spectral quality, reduction in the number of cross-peaks, and increased resolution. 36Complex formation also led to significant changes in chemical shift, indicative of the enzyme undergoing a transition to a new conformation.Dramatic changes in chemical shift were also observed between the E:NADP + :folate and E:NADP + :THF complexes, as the enzyme progressed to an occluded conformation after the chemical step.The largest chemical shift perturbations were exhibited by Met20 and Trp22, located in the highly flexible M20 loop; Met16, Met92, and Trp133 also showed notable changes.As with the spectrum of [methyl- 13 C]methionine-labeled apo-EcDHFR, the 1 H− 13 C HSQC spectra of apo-MpDHFR labeled with [methyl- 13 C]methionine or [indole-δ 1 -13 C]tryptophan exhibited clear evidence of conformational heterogeneity (Figure S5).Again, addition of ligands led to an increase in spectral quality, indicative of the enzyme adopting one major conformation. 1H− 13 C HSQC spectra of the two complexes showed the expected five methionine and three tryptophan cross-peaks (Figure 3C,D).No significant change in chemical shift was observed between the Michaelis and product complexes for methionine or tryptophan residues (Tables S4  and S5).Notably, small changes in proton chemical shifts for Met21 and Trp23 were observed, but as these residues are directly adjacent to the site of chemistry, these most likely occur because of the differences in ligands, highlighting how it is important to differentiate between small changes that describe the changes at the site of chemistry and more widespread chemical shift deviations, indicative of protein conformational changes as observed for EcDHFR.
The extensive differences in chemical shifts seen in the EcDHFR:NADP + :folate and EcDHFR:NADP + :THF complexes, as had previously been observed for [ 15 N]alaninelabeled EcDHFR, 42 provide clear evidence that methionine and tryptophan labeling is sufficient to observe the closed-tooccluded conformational change in EcDHFR.As expected, residues in the flexible M20 loop (Met16, Met20, and Trp22) all show large changes.Met16 in the apoenzyme and Michaelis complex appears to be solvent-exposed.However, in the product complex, the residue is located directly within the active site, 7 drastically changing the immediate molecular environment of the 13 C methyl group.Again, Met20 in the Michaelis complex points toward the FG loop, whereas in the product complex, it is more solvent-exposed, thus accounting for the observed downfield shift.Importantly, in EcDHFR, chemical shift changes are widely observed for all methionine and most tryptophan resonances, allowing binding effects of different ligands and deviations in protein conformation to be differentiated.No such widespread and significant changes in chemical shift were observed between the MpDHFR:NADP + :folate and MpDHFR:NADP + :THF complexes, confirming earlier work that suggested that MpDHFR does not adopt an occluded conformation after hydride transfer. 10The results gained from EcDHFR and MpDHFR confirm that selective side-chain labeling can be used to identify whether an occluded-like conformation is adopted by a specific DHFR.
It has so far been challenging to investigate TmDHFR by conventional NMR methods because of poor tumbling in solution and fast T 2 relaxation as a consequence of its dimeric nature with a total molecular mass of 38k and its elongated shape. 18Perdeuteration of TmDHFR to improve relaxation issues proved to be ineffective as the high thermal stability of TmDHFR prevented complete proton back-exchange of backbone amide groups in non-solvent-exposed regions (Figure S4).Consequently, three-dimensional NMR spectra of TmDHFR are of limited quality, and no triple-resonance assignment data for TmDHFR have been published.However, TmDHFR contains only three methionines and one tryptophan. 18Because these side chains display superior relaxation properties, selective side-chain labeling can be used to explore the conformational dynamics of TmDHFR. 1 H− 13 C HSQC spectra of TmDHFR exhibit four distinct methionine cross-peaks and one tryptophan cross-peak (Figure 3E,F).Met1 could be assigned as the intensity of this resonance was considerably lower than the others, as expected for a residue present in only ∼10% of the protein molecules because of aminopeptidase activity in E. coli (Figure S3).No significant changes in [methyl- 13 C]methionine chemical shifts were observed between the apoenzyme and either complex (Tables S6 and S7).Trp23 displays a proton shift difference of 0.07 ppm between the Michaelis and product complexes, similar to that observed for MpDHFR.This difference is most likely due to different ligands, which are directly proximal to Trp23.These results suggest that TmDHFR adopts only a single major conformation in solution.
The occluded conformation has so far only been observed in EcDHFR, although some evidence of conformational flexibility following the chemical step has been reported for bacterial DHFRs. 11We therefore tested our proposal that DHFR homologues that can form a hydrogen bond similar to the crucial Ser148−Asn23 interaction in EcDHFR will adopt a conformation similar to the occluded conformation found in EcDHFR.SeDHFR contains residues His24 and Ser150 in the equivalent positions (Figure S2), and the replacement of asparagine by histidine is not likely to have a negative effect because the hydrogen bonds are formed to the backbone amide and carbonyl of Asn23.SeDHFR should therefore adopt an occluded-like conformation following the chemical step.No NMR assignments or structural information currently exists for SeDHFR, although tentative assignment of methionine residues could be performed via sequence analysis, indicating that conserved residues are likely to reside in electronic environments similar to those in EcDHFR and MpDHFR.Like EcDHFR and MpDHFR, the quality of methyl- 13 C spectra of the methionine-labeled SeDHFR apoenzyme was low as a consequence of conformational heterogeneity (Figure S5).SeDHFR exhibits three well-defined cross-peaks when in the E:NADP + :folate complex.However, unlike EcDHFR, SeDHFR shows two distinct species in the E:NADP + :THF complex; both species seem to be equally populated (Figure 3G,H).Half of the cross-peaks present in the product complex overlap with those present in the Michaelis complex, indicating that in the product complex SeDHFR is characterized by an equilibrium between the Michaelis conformation and a second conformational state (Tables S8 and S9).It is worth noting that changing Ser148 in EcDHFR leads to a destabilization of the occluded conformation. 21To confirm whether the two observed species in SeDHFR related to similar conformational behavior, SeDHFR-S150A was prepared.In the product complex, two species were present in this variant, but the relative population of the two states was drastically shifted to favor the Michaelis complex (Figure 3I,J and Tables S10 and S11).This reinforced the idea that the chemical shift perturbations observed in Met94 and Met42 are likely to be caused by a conformational change resembling that found in EcDHFR and again confirmed the significance of the hydrogen bond interaction between the M20 and GH loops in its formation.No significant chemical shift perturbations were observed in the [indole-δ 1 -13 C]tryptophan-labeled SeDHFR, which in this instance proved not to be diagnostic of conformational change.
SeDHFR, unlike EcDHFR and MpDHFR, lacks methionine residues within the M20 loop, which directly participates in the closed-to-occluded conformational change.To provide further evidence that the second conformation present in the product complex of SeDHFR resembles the occluded conformation observed for EcDHFR, 1 H− 15 N HSQC spectra of both the Michaelis and product complexes were recorded (Figures S6  and S7).As observed through the [methyl- 13 C]methionine resonances, widespread chemical shift differences are found between the complexes, again justifying that simple side-chain labeling is sufficient to identify conformational change within DHFRs, even when the probes are not located directly within the site of conformational change.This suggested that SeDHFR does indeed form an occluded-like conformation.
In general, we found [methyl- 13 C]methionine chemical shifts more diagnostic of DHFR conformational changes than those of [indole-δ 1 -13 C]tryptophan.In all four DHFRs, Trp22 forms a hydrogen bond to bound folates.It would therefore be expected to show a difference in chemical shift between the a p o e n z y m ea n dt h et w oc o m p l e x e s .I nM p D H F Ra n d TmDHFR, Trp23 does indeed display small proton chemical shifts for each separate complex, while in contrast, no significant differences are observed for SeDHFR.A large deviation is found for the Michaelis complex of EcDHFR, but apoenzyme and product complexes are similar.Taken together, this illustrates how ligand binding effects dominate for this residue because of its location adjacent to the site of chemistry, obscuring any chemical shift perturbations in response to protein conformational changes.Similarly, for the spectra of apo-EcDHFR, MpDHFR, and SeDHFR, the tryptophan data show no evidence of the conformational heterogeneity obviously present upon observation of [methyl- 13 C]methionine spectra and previously reported. 28,43Conformational changes can nonetheless be followed through [indole-δ 1 -13 C]tryptophan labeling, because there is clear evidence for the closed-tooccluded transition in EcDHFR, which may simply be more evident for EcDHFR than for SeDHFR because of the larger number of [indole-δ 1 - 13 C]tryptophans present within its sequence.
Importantly, by combination of the [indole-δ 1 -13 C]tryptophan labeling with [methyl- 13 C]methionine labeling, which could in principle have been performed in a single protein, clear evidence of large-scale protein conformational changes was found for EcDHFR and SeDHFR.In EcDHFR, the ability to switch to an occluded conformation is essential for progression through the catalytic cycle, but this had not previously been observed in other DHFRs. 10,11Ser148 in EcDHFR, which forms two hydrogen bonds to the M20 loop, is replaced with proline in MpDHFR. 10Consequently, M20 loop motions appear not to play a significant role in MpDHFR catalysis, and it was known that MpDHFR remains in a closed conformation throughout the catalytic cycle. 10,44Switching from the closed to the occluded immediately follows hydride transfer in EcDHFR, before release of NADP + from the product ternary complex. 45Switching to the occluded conformation in EcDHFR lowers the binding affinity of NADP + , mitigates possible product inhibition effects, and allows efficient progress through the catalytic cycle. 10,46owever, similar single-turnover rate constants for EcDHFRand MpDHFR-catalyzed hydride transfer at pH 7 show that the ability to form the occluded conformation does not have a major effect on the chemical step.Hydride transfer in the occluded conformation is not possible because of the distance between the reactants. 10,36espite the lack of assignment data for both Tm-and SeDHFR, important conclusions can be drawn from the data presented here.Side-chain labeling of methionine and tryptophan yields simple and intuitively useful data.Previous reports that TmDHFR remains in a fixed open conformation during catalysis were based upon two X-ray single-crystal structures, and neither of them effectively modeled the Michaelis or product complex. 18Hence, here the first direct evidence that TmDHFR remains fixed in an open conformation throughout the catalytic cycle is reported.Any further emphasis is however limited for TmDHFR because of the different distribution of methionine residues to Ec-and MpDHFR and highlights how the distribution and conservation of labels can be potentially limiting when establishing conformational changes and comparing orthologs.
The ease of labeling meant that the conformational behavior of SeDHFR, for which no structural data exist, could be analyzed rapidly, and predictions based upon primary sequence analysis were verified.SeDHFR exists in multiple solution forms, and their assignment could be based on similarity to other DHFRs.Like EcDHFR, SeDHFR showed two conformations in the product complex, but the relative stability of two conformations was evidently different.The occluded conformation dominates in the product complex of EcDHFR, while in SeDHFR, the Michaelis conformation is still present at an approximately equimolar concentration.This difference may influence the kinetics of product release.
Principally selective side-chain 13 C labeling provides a rapid and cost-effective approach for surveying protein conformations during catalysis for multiple proteins, allowing more traditional and detailed analysis to be adjudged worthwhile.The cost of production of proteins labeled with [methyl- 13 C]methionine (£11 per liter of culture) and [indole-δ 1 -13 C]tryptophan (£96 per liter of culture) is considerably lower than that of triply ( 2 H, 13 C, 15 N, £622 per liter of culture) or doubly ( 13 C, 15 N, £262 per liter of culture) labeled proteins required for full NMR assignments (when accounting for 4 g L −1 [U-2 H, 13 C]-or [U-1 H, 13 C]glucose, respectively).Selective labeling also provides greatly simplified spectra, exhibiting only a fraction of the number of cross-peaks seen for fully labeled proteins, and the data can be acquired over much narrower spectral windows (10−14 ppm in 13 C rather than ∼70 ppm for a full aliphatic 1 H− 13 C HSQC spectrum).In addition, 13 CH 3 groups (as in [methyl- 13 C]methionine) typically show high sensitivity in NMR experiments, because of the three equivalent protons, 43 and aliphatic 13 C shifts are most sensitive to conformational changes. 38For DHFR, selective side-chain 13 C labeling is clearly sufficient to gain insight into large-scale protein motions that are important for progression through the catalytic cycle and expands upon previous work that demonstrated for EcDHFR the effectiveness of side-chain labeling methods in measuring picosecond to nanosecond time-scale protein dynamics. 47n conclusion, we have demonstrated here the utility of selective side-chain labeling to rapidly probe conformational behavior during DHFR catalysis.EcDHFR is known to undergo large-scale conformational changes along the catalytic cycle and shows clear chemical shift perturbations of its methionine methyl and tryptophan indole groups following the chemical step.MpDHFR, on the other hand, an enzyme known to maintain a single major conformation during catalysis, does not display such perturbations.Thermostable, homodimeric TmDHFR does not show any significant chemical shift perturbations following the chemical step, corroborating crystallographic evidence and providing the first direct evidence that this enzyme also adopts a single major conformation in solution.Conformational sampling is, however, not unique to EcDHFR, and DHFR from S. enterica exhibits similar conformational behavior, most likely adopting an occludedlike conformation.Clearly, amino acid sequence analysis allows the prediction of large-scale protein conformational behavior in DHFRs, and spectra obtained with site selectively labeled proteins may be readily interpreted even in the absence of resonance assignments.However, the scope of selective sidechain 13 C labeling is not limited to an analysis of structural events during DHFR catalysis and can be used to track conformational changes in a wide range of proteins.

* S Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.7b00045.
Purification procedure for SeDHFR and SDS-PAGE gel; amino acid sequence analysis of multiple DHFRs; mass spectrum of TmDHFR highlighting the presence of Met1; a modified version of Figure 3

Figure 1 .
Figure 1.Catalytic cycle of EcDHFR under steady state conditions at pH 7. The major protein conformations adopted by the complexes in the catalytic cycle are indicated: open/disordered (green), closed (red), and occluded (blue).7,23,24 40  1  H−13 CH S Q Cs p e c t r ao fE c D H F Rl a b e l e dw i t h [methyl-13 C]methionine or [indole-δ 1 -13 C]tryptophan (Figure 3A,B and Tables