Quantifying the Relationship between Conformational Dynamics and Enzymatic Activity in Ribonuclease HI Homologues

Abstract

Ribonuclease HI (RNHI), a ubiquitous, non-sequence-specific endonuclease, cleaves the RNA strand in RNA/DNA hybrids. RNHI functions in replication, genome maintenance, and retroviral reverse transcriptases contain an essential ribonuclease H domain. NMR spectroscopy combined with molecular dynamics (MD) simulations suggest a model in which the extended handle region domain of Escherichia coli RNHI populates (substrate-binding competent) “open” and (substrate-binding incompetent) “closed” states, while the thermophilic Thermus thermophilus RNHI mainly populates the closed state at 300 K (Stafford, K. A., Robustelli, P., and Palmer, A. G. PLoS Computational Biology 2013, 9, 1–10). In addition, an in silico designed mutant E. coli Val98Ala RNHI was predicted to populate primarily the closed state. The present work validates this model and confirms the predicted properties of the designed mutant. MD simulations suggest that the conformational preferences of the handle region correlate with the conformations of Trp85, Thr92, and Val101. NMR residual dipolar coupling constants, three-bond scalar coupling constants, and chemical shifts experimentally define the conformational states of these residues and hence of the handle domain. These NMR parameters correlate with the Michaelis constants for RNHI homologues, confirming the important role of the handle region in modulation of substrate recognition and illustrating the power of NMR spectroscopy in dissecting the conformational preferences underlying enzyme function.

Graphical Abstract

Comparison between homologous enzymes has long been employed to discern sequence and structural regions critical to activity because homologues exhibit similar activity, but under different conditions.(1–4) Ribonuclease HI (RNHI) is an endonuclease that cleaves the RNA strand in RNA/DNA duplexes(5) and plays roles in DNA replication,(6) genome maintenance(7) and retroviral replication.(8–9) Structural, dynamical, and functional relationships have been investigated for RNHI homologues from the mesophile Escherichia coli (EcRNHI), thermophile Thermus thermophilus (TtRNHI), and psychrotroph Shewanella oneidensis (SoRNHI)(10–14). Based on NMR spin relaxation spectroscopy(15) and molecular dynamics (MD) simulations,(16) a two-state kinetic model was proposed for the handle region (residues 81–101; EcRNHI sequence numbering is used herein), an extended domain necessary for substrate binding(17) (Figure S1).

In this model, the handle region (Figure 1A) dynamically populates “open” (substrate-binding competent) or “closed” (substrate-binding incompetent) conformations (Figure 1B). Thus, the increased Michaelis constant for TtRNHI, compared to EcRNHI, is associated with an increased population of closed conformations and hence reduced affinity for substrate and lower activity. Furthermore, an in silico designed mutant Val98Ala (V98A) EcRNHI was predicted to populate primarily closed states and therefore to exhibit an increased Michaelis constant.(16)

Figure 1. (A) Handle region, (B) Two-state kinetic model (C) Handle distance distribution (D, E) T92 RDC distributions.

Distance metric or T92 ¹H-¹⁵N RDC for T92 distributions derived from MD simulations are shown as solid lines for wild-type proteins and dashed lines for V98A EcRNHI. In (E), the T92 RDCs for EcRNHI are separated into separate distributions for closed and open states using the handle distance metric; V98A EcRNHI has T92 RDC profiles for open and closed states. (C) Green diamonds are distances measured from x-ray crystal structures (insets on top right) for SoRNHI (2E4L), EcRNHI (2RN2), and TtRNHI (1RIL). (D, E) Green diamonds are experimental RDCs for wild-type proteins and the yellow diamond is for V98A EcRNHI. (*) are RDCs obtained from the respective crystal structures using the experimentally derived alignment tensor.

The present study provides experimental confirmation of the proposed two-state model and of the predicted properties of the V98A EcRNHI mutant. Measured NMR residual dipolar coupling constants (RDCs), three-bond scalar coupling constants, and isotropic chemical shifts report on the conformational state of the handle domain and are highly correlated with Michaelis constants for RNHI homologues and V98A EcRNHI.

MD simulations were examined to identify NMR observable parameters sensitive to the conformational state of the handle region. All-atom 100-ns MD simulations have been reported previously for the RNHI homologues and mutant.(16) Open and closed states in individual MD frames were characterized by the distance from the Ca of Trp 85 (W85) that forms part of a DNA-binding channel (Figure S1C) (18) to the Ca of Thr 92 (T92) at the tip of the dynamic handle loop (this metric produces better resolution between conformations than previous measures).(16) Distance distributions are shown in Figure 1C, with open conformations having distance ≥ 7.8 Å and closed conformations having distances < 7.8 Å)

¹H-¹⁵N RDCs(19–20) were predicted for individual MD frames using PALES.(21) Backbone amide ¹H-¹⁵N RDC distributions for T92 are shown in Figure 1D. Separating the RDCs by the handle distance metric (Figure 1E) shows that more negative RDCs are associated with closed conformations of the handle region. The distribution of predicted Ne1-He1 RDCs for the indole group of W85 also correlates with the handle distance metric, although not as strongly as for T92 (Figure S2).

Experimental ¹H-¹⁵N RDCs measured in anisotropic media using the IPAP technique(22) agree well with values back calculated from x-ray crystal structures for RNHI proteins (Figure S3 and Table S1). Measured values for T92 N-H and W85 Ne1-He1 are negative with ordering TtRNHI < V98A EcRNHI < SoRNHI < EcRNHI (the SoRNHI T92 RDC could not be measured because of resonance overlap and the TtRNHI W85 RDC is not measurable because of exchange broadening). RDCs for T92 N-H and W85 Nε1-Hε1 calculated from experimentally determined alignment tensors showed similar ordering. These data indicate that EcRNHI has the largest population of the open conformation and TtRNHI the smallest.

Valine 98 (V98) in EcRNHI forms the border of a hydrophobic spine with V101 that links helix C and D of the handle region to DNA binding channel residues W81 and W85 through side chain stacking interactions; residue 101 is Arg in TtRNHI, but similar stacking interactions are observed (Figure S4). Chemical exchange line broadening implicates the sidechains of V98 and V101 in handle region conformational dynamics, possibly arising from χ1 rotameric jumps.(23) ¹H-¹⁵N-NOESY-HSQC spectra show NOE contacts between T92 H^N and V98 Hβ and Hγ1 atoms in EcRNHI but not between T92 H^N and A98 Hβ atoms in the mutant (Figure S5). MD simulations suggest that the distance between V98 Cγ1 and T92 N display discrete distance profiles for open and closed frames (Figure S6). MD simulations also suggest that χ1 rotameric combinations V98 trans/V101 gauche+ or V98 gauche–/V101 gauche+ are correlated with closed or open conformations, respectively (Figure S7). In addition, the closed conformation of the handle region is preferentially populated when residue 101 adopts the trans χ1 rotamer (Figure 2).

Figure 2. Handle distance profiles correlated to rotameric state of residue 101.

A) EcRNHI (B) V98A EcRNHI (C) SoRNHI (D) TtRNHI. (blue) trans, (black) gauche+, and (pink) gauche–populations of the χ1 dihedral angle.

Rotamer distributions for Val residues in RNHI homologues were obtained from ¹³CO-¹³Cγ and ¹⁵N-¹³Cγ three-bond scalar coupling constants and ¹³Cγ chemical shifts.(24) Experimental rotamer populations agree well with estimates from MD simulations (Table S2 and S3). The low (~10%) population of trans χ1 rotamers for V101 in EcRNHI suggests averaging of the handle region between open and closed conformations. The increased (>30%) population of trans χ1 rotamers for V101 in V98A EcRNHI suggests an increased population of the closed conformation.

In addition, the largest chemical shift perturbations in V98A EcRNHI occur around T92 and V101, suggesting a conformational difference in the handle region for the mutant (Figure S8). The difference in one-bond ¹H-¹⁵N scalar coupling constants for the W85 amide in EcRNHI and V98A EcRNHI of −2.6 Hz is among the largest observed (Table S4), suggesting a conformational change in the handle region. Zweckstetter and coworkers have suggested Δ¹J_NH ≥ |1.6 Hz| reflect changes in hydrogen bonding.(25)

The Michaelis constant K_M measured for EcRNHI and V98A EcRNHI are reported in Figure S11 and Table S5. K_M for the mutant is approximately four-fold greater than for the wild-type protein. Michaelis-Menton enzyme kinetic parameters have been reported for the other RNHI homologues.(26–27) Each of these studies also reported kinetic parameters for EcRNHI measured under identical conditions. Thus, to control for different experimental protocols, K_M for each RNHI homolog was normalized by the corresponding K_M for EcRNHI. The normalized K_M ratios are plotted versus W85 RDCs, T92 RDCs and residue 101 trans populations in Figure 3A–C respectively. Weighted conformer scores were calculated from these three data sets using Equation S1 and strongly correlate with K_M ratios (Figure 3D).

Figure 3. Experimental RDCs and residue 101 trans percentage vs respective K_M Ratios.

Square: EcRNHI, Circle: SoRNHI, Triangle: V98A EcRNHI, Diamond: TtRNHI; RDC experiments were done in triplicate and error bars indicate propagated standard error; error bars for back calculated RDCs are the rms between experimental values (A) Experimental (black) and back calculated from alignment tensor (pink) RDCs for W85 Nε1-Hε1 RDCs plotted against K_M ratios from Table S5. (B) Same as A but for T92 N-H RDC (C) Calculated % (trans) from scalar coupling constants (black), chemical shifts (pink) and simulations (yellow) plotted against K_M ratios from Table S5. (D) Weighted Conformer scores (calculated using the data from A, B and C with Eq. S1) plotted against K_M ratios from Table S5.

Discussion

NMR spin relaxation measurements have identified ms-ms time scale conformational dynamic processes in the handle region of RNHI enzymes.(15) MD simulations suggest that these observations reflect transitions between open and closed conformations of the handle region.(16) This model predicts that closed conformations would disfavor substrate binding, leading to elevated K_M without affecting the chemical step of the reaction. An in silico designed mutant, V98A EcRNHI, was predicted to have an increased population of the closed state and consequently larger K_M. However, prior NMR and biochemical experiments did not specifically detect an equilibrium between open and closed states and did not characterize the properties of the mutant.

The present work identified three structural features differentiating open and closed conformations of the handle region: orientation of T92 N-H bond, orientation of W85 Nε1-Hε1 bond, and rotamer preferences of residue 101. These three structural features were quantified using NMR RDCs, scalar coupling constants and chemical shifts. Each of these measurements represents the weighted average over the distribution of conformations of the proteins in solution. Not all parameters could be measured for all proteins, but at least two parameters were characterized experimentally for each protein.

These data are summarized in Figure 3 and show a remarkable linear correlation with the K_M for the enzyme reaction, including for the in silico designed mutant. These results provide strong evidence in support of the two-state model for handle region conformational dynamics and for the hypothesis that substrate recognition is strongly favored by open conformations of the handle region.

Structures of closed and open structures of EcRNHI taken from MD simulations are superposed in Figure 4. These structures have predicted T92 N-H RDCs of −4 Hz and 4 Hz, respectively, typical of the distributions shown in Figure 1E. The angle between the N-H vectors in these two structures is 46.3° (Figure S9). In closed conformations, the T92 N-H bond vector points upward towards the outermost tip of the loop, while in open structures, the bond vector points inwards toward the center of the loop. This difference allows T92 O to be in a better position/range to interact with R88 Hε in closed structures, possibly through water-mediated interactions(16) (Figure S10); previous research(16) also has elaborated the importance of Arg 88 (R88) in regulating handle region dynamics.

Figure 4. Geometry and interactions of T92.

Superposed EcRNHI MD simulated structures with predicted T92 N-H RDCs that fall at the center of the closed (pink: −4 Hz) and open (cyan: 4 Hz) distributions. 180° views of the handle region are shown. T92 N is in blue and the distance from T92 O (red) to R88 Hε (white) is provided.

Tryptophan 85 is highly conserved and plays a major role in the stability of the hydrophobic core and forms part of the DNA strand substrate-binding channel. The V98 sidechain stacks directly against the W85 indole ring and modulates this hydrophobic network through rotameric transitions with residue 101. Rotameric transitions appear to be concerted with handle region conformational dynamics bringing T92 closer to W85 and R88. The handle region strongly favors the closed conformation when residue 101 adopts the trans rotamer and measured values of the trans population are highly correlated with K_M for the RNHI enzymes.

Conclusion

Previous NMR spin relaxation spectroscopy and MD simulations hypothesized a two-state model for conformational dynamics of the handle region of ribonuclease HI enzymes. Based on this model, an in silico mutant V98A EcRNHI was designed and predicted to have an increased population of the closed conformation of the handle region and consequently an increased Michaelis constant. This model has been validated by NMR measurements for W85, T92, and residue 101, whose conformational preferences are strongly correlated with the open-closed transition in the two-state model. The experimental measurements, including RDCs, scalar coupling constants, and chemical shifts, validate this model, including the properties of the V98A mutant. An unexpectedly strong correlation between the NMR parameters and the Michaelis constant for the enzymatic reaction indicates the important role of the handle region for tuning of substrate recognition in the RNHI enzymes.