Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Biochemistry. Author manuscript; available in PMC 2010 April 28.
Published in final edited form as:
PMCID: PMC2860867

Structure, Stability, and Folding of Ribonuclease H1 from the Moderately Thermophilic Chlorobium tepidum: Comparison with Thermophilic and Mesophilic Homologues


Proteins from thermophilic organisms are able to function under conditions that render a typical mesophilic protein inactive. Pairwise comparisons of homologous mesophilic and thermophilic proteins can help to identify the energetic features of a protein’s energy landscape that lead to such thermostability. Previous studies of bacterial ribonucleases H (RNases H) from the thermophile Thermus thermophilus and the mesophile Escherichia coli revealed that the thermostability arises in part from an unusually low change in heat capacity upon unfolding (ΔCp) for the thermophilic protein. Here, we have further examined how nearly identical proteins can adapt to different thermal constraints by adding a moderately thermophilic homologue to the previously characterized mesophilic and thermophilic pair. We identified a putative RNase H from Chlorobium. tepidum and demonstrated that it is an active RNase H and adopts the RNase H fold. The moderately thermophilic protein has a melting temperature (Tm) similar to that of the mesophilic homologue yet also has a surprisingly low ΔCp, like the thermophilic homologue. This new RNase H folds through a pathway similar to that of the previously studied RNases H. These results suggest that lowering the ΔCp may be a general strategy for achieving thermophilicity for some protein families and implicate the folding core as the major contributor to this effect. It should now be possible to design RNases H that display the desired thermophilic or mesophilic properties, as defined by their ΔCp values, and therefore fine-tune the energy landscape in a predictable fashion.

Proteins from thermophilic organisms have the remarkable ability to remain folded and functional at temperatures where proteins from mesophiles are usually unfolded and inactive. Understanding the mechanisms by which these proteins have been optimized to function under such extreme conditions would provide great insight into the rules governing protein stability and thermodynamics. An obvious first step toward this goal is to compare the similarities and differences between sets of homologous proteins from thermophilic and mesophilic organisms. Such structural comparisons reveal that, as expected, the homologous proteins show nearly identical folds, and while specific changes in the side chains and their interactions can be seen, there are few generalizable rules (1-7). Similarly, sequence-based comparisons have been unable to determine any single model for thermal adaptation (8, 9), concluding that residue-level control of energetics does not obey any obvious rules. Furthermore, the sequence similarity between thermophilic and mesophilic homologues is generally the same as that between two mesophilic homologues. Thus, there is no easy way to computationally determine whether a given protein originates from either a thermophilic or mesophilic organism.

Despite the difficulties in determining thermophilicity computationally, bioinformatic studies have found some general structural and sequence-based features that tend to correlate with the optimal growth temperature of organisms (6, 9-11). None of these features are present in all thermophilic proteins, and there are exceptions to every rule that has been proposed. The simplest experimental approach employed to probe the basis of thermal stability in proteins has been to make point substitutions in mesophilic proteins that lead to an increased melting temperature (Tm) of the protein. The Tm of a specific protein can be naively increased by a simple residue replacement, resulting in introduction of specific stabilizing interactions. This approach, however, rarely results in a truly thermophilic protein but rather creates a hyperstable mesophilic protein (12, 13).

Alternatively, thermodynamic characterization can usually differentiate between homologous thermophilic and mesophilic proteins. While the melting temperature of the thermophilic protein is generally higher than that of the mesophilic counterpart, studies on multiple pairs of thermophilic and mesophilic proteins show that they use various methods, including a change in enthalpy (ΔH), a change in heat capacity upon unfolding (ΔCp),1 or a change in overall stability (ΔG)(14-18), to tune their thermal stabilities.

Ribonuclease H (RNase H) has proven to be an excellent model protein for such thermodynamic studies. RNase H is a well-characterized nuclease that specifically digests the RNA strand of RNA–DNA hybrids. It is found in many different organisms, and in all cases examined, it folds into a single α+β domain ranging from 130 to 166 amino acids displaying the so-called RNase H fold (1, 19). The energy landscapes of two RNases H, those from Thermus thermophilus and Escherichia coli (TthRNH and EcoRNH, respectively), have been extensively studied via thermal and chemical denaturation as well as hydrogen exchange (16, 20, 21). The stability curves (temperature vs ΔGunf profile) show that, in addition to a higher melting temperature (Tm), TthRNH has a dramatically lower change in heat capacity upon unfolding (ΔCp) and a higher overall stability (ΔGunf) than EcoRNH (19). While the different ΔGunf value was not unexpected, the variation in ΔCp was quite surprising. Previous studies have determined that ΔCp correlates strongly with the change in solvent-exposed surface area upon unfolding (ΔASA) (22), and because the two proteins bury a similar amount of nonpolar surface area when folded, it was expected that they would share similar ΔCp values. After this was determined to not be the case, further studies revealed that the folding core of TthRNH is largely responsible for its low ΔCp (23) and that substituting a single polar residue into the core region results in a more mesophile-like ΔCp (24). Differential scanning calorimetry assigns the unusually low ΔCp to an unexpectedly low absolute Cp of the unfolded state (25). All of these data can be explained by the hypothesis that TthRNH buries some hydrophobic surface area in the unfolded state (23, 25). The residual structure is thought to arise from the core of TthRNH, but the nature of this residual structure and how it is encoded in the sequence is unknown. Experimentally, this has been particularly challenging, as the unfolded state is not significantly populated under native conditions. Moreover, whether this unusually low ΔCp is a general feature of thermophilic proteins, or even thermophilic RNases H, is unclear.

To further probe the role of such residual structure in the unfolded state of thermophilic proteins, as well as to determine whether such behavior can be generalized to other thermophilic proteins, we have characterized a putative RNase H from the moderately thermophilic Chlorobium tepidum. C. tepidum has an optimal growth temperature (48 °C) between those of E. coli (37 °C) and T. thermophilus (68 °C) (26). These three RNases H have pairwise sequence identities greater than 50% and similarities greater than 70% (Figure 1). We have demonstrated that the putative protein is indeed an RNase H, determined the crystal structure of C. tepidum RNase H* (the asterisk denotes a cysteine-free variant of the protein), and determined that it shows few changes from the structures of the other two molecules, with an α-carbon root-mean-square deviation (rmsd) of <1.5 Å across all three proteins. We describe structural, thermodynamic, and kinetic analyses of the moderately thermophilic RNase H homologue, which are suggestive of a low ΔCp as a general property of thermophilic RNases H.

Figure 1
(A) Sequence alignment of T. thermophilus, C. tepidum, and E. coli RNases H* with conserved thermophilic residues colored red, conserved mesophilic residues blue, and invariant residues gray. Former cysteines in C. tepidum and E. coli RNases H are marked ...


Cloning and Expression of C. tepidum RNase H

The ribonuclease H gene was amplified from C. tepidum genomic DNA (obtained from ATCC) by PCR using primers designed to add flanking KpnI and NdeI restriction sites, which were used to insert the gene into a pAED4 vector (27). Both strands of the coding region in the resulting plasmid (pKR203) were sequenced. Standard QuikChange (Stratagene) mutagenesis was used to create a cysteine-free variant in which all three cysteines were replaced with alanines (pKR201). The gene for wild-type CtepRNH was subcloned into a pET27 vector (Novagen) containing a thrombin site followed by an N-terminal six-His tag for ease of purification (pKR200). Plasmids encoding CtepRNH or CtepRNH* were transformed into E. coli BL21DE3 Codon+ cells (Novagen), and the cells were grown at 37 °C in Luria broth with 200 μg/mL ampicillin and 25 μg/mL chloramphenicol. Expression was induced at an OD600 of 0.6 by the addition of 1 mM IPTG, and cells were grown for an additional 3 h before being harvested by centrifugation.

Purification of C. tepidum RNases H

Purification of wild-type CtepRNH was conducted at pH 8.0 in the presence of fresh 1 mM tris(2-carboxyethyl)phosphine (TCEP). Cell pellets were resuspended in 50 mM Tris (pH 8.0), 20 mM NaCl, and 1 mM TCEP and lysed by sonication. The soluble fraction was loaded onto a HiTrap Heparin HP column from GE Healthcare and eluted with a linear NaCl gradient (from 0.02 to 0.8 M). Fractions containing RNase H were pooled, loaded onto a HisTrap nickel column from GE Healthcare {in nickel buffer A [20 mM NaPO4 (pH 8.0), 0.5 M NaCl, 30 mM imidazole, and 1 mM TCEP]}, and eluted with a step to 500 mM imidazole. Pooled fractions were incubated overnight with thrombin to cleave the histidine tag and dialyzed back into nickel buffer A. His tag-free protein was reloaded onto the nickel column, the flow-through collected, and the resulting protein judged to be >98% pure by SDS–PAGE. CtepRNH* variants were purified similarly; however, in the sonication and heparin column purification steps, TCEP was omitted from all buffers and the nickel column was replaced with a Source 15S column. Fractions from the heparin column containing cysteine-free RNase H were pooled, loaded onto the Source 15S column [in 20 mM NaAc (pH 5.5), 200 mM NaCl, and 0.1 mM EDTA], and eluted with a linear NaCl gradient (from 0.2 to 0.6 M). Fractions were pooled and judged to be >98% pure by SDS–PAGE.

GdmCl-Induced and Thermal Denaturation

Circular dichroism (CD) measurements were taken on an Aviv 62DS spectropolarimeter with a Peltier temperature-controlled cell holder. For all samples described below, each CD measurement is an average of the signal at 222 nm for 1 min, in a 1 cm path length cuvette.

For experiments with CtepRNH, samples containing 50 μg/mL protein, 20 mM NaAc (pH 5.5), 50 mM KCl, 1 mM TCEP, and the appropriate concentration of GdmCl were allowed to equilibrate overnight before the CD signal was measured. For thermal denaturation, the CD signal was recorded every 3 °C, with a 5 min equilibration time and a 1 min averaging time at each temperature. At the end of the melt, reversibility was determined by returning to the beginning temperature and comparing the CD spectrum from 200 to 300 nm to the premelt spectrum. Experiments with the cysteine-free mutant C. tepidum RNase H* were conducted under identical conditions, but lacking TCEP.

GdmCl-induced denaturation of CtepRNH* was also performed by the serial addition of a solution containing protein at a high GdmCl concentration, 50 μg/mL protein, 20 mM NaAc, and 50 mM KCl (pH 5.5) to a starting solution containing the above without GdmCl. For each protein, the same starting and titrant solutions were used to measure the GdmCl-induced equilibrium unfolding at 10 different temperatures ranging from 5 to 50 °C. Additional data were collected by preparing 25 samples at varying concentrations of GdmCl and collecting temperature melts every 5 °C from 5 to 60 °C. Data from temperature melts at varying concentrations of GdmCl and data from GdmCl-induced melts at varying temperatures were both used for generation of the stability curves.

Denaturation and Stability Curve Data Analysis

Free energies of unfolding (ΔGunf) were determined from GdmCl-induced denaturation of CtepRNH* at several temperatures, assuming a two-state model and a linear relationship between ΔGunf and concentration of GdmCl (28). To determine the Tm for each protein, thermal melts were fit using a two-state model and the Gibbs–Helmholtz relationship between ΔGunf and temperature (29). The resulting free energies of unfolding obtained from the denaturant melts were plotted as a function of temperature, with a single point added for the TmGunf = 0). These data (ΔGunf vs T) were fit to a modified Gibbs–Helmholtz equation (eq 1) using the Tm as a reference temperature:


Crystal Structure of CtepRNH*

CtepRNH* was crystallized by the hanging-drop method. Lyophilized protein was dissolved in water, resulting in a final protein concentration of 10 mg/mL. Crystals of CtepRNH*, were grown in 20% PEG 4000, 0.1 M Na HEPES (pH 7.5), and 10% 2-propanol, cryoprotected under the same conditions but with an additional 30% 2-propanol, flash-frozen in liquid nitrogen, and screened for diffraction. A full X-ray diffraction data set was collected at Stanford Synchrotron Radiation Laboratory beamline 11-1 from a crystal that diffracted to 1.6 Å.

To determine the crystal structure, more protein was expressed in minimal medium incorporating selenomethionine and again crystallized, this time from a final protein concentration of 5 mg/mL. The hanging-drop method was used to set new trays with SeMet CtepRNH*, resulting in crystals from a drop containing 26% PEG 4000, 0.2 M NH3Ac, and 0.1 M NaAc (pH 5.4). Crystals were cryoprotected by addition of 30% MPD to crystallization conditions and flash-frozen in liquid nitrogen. Crystals were screened for diffraction and anomalous signal at LBNL ALS beamline 8.3.1. A SeMet crystal diffracted to 2.0 Å and was used for experimental phasing. The crystal structure from the native data set was determined, and models for each of the two molecules in the asymmetric unit were manually rebuilt using O (30). The model was further refined using the experimental phase restraints and RefMac (31). Data collection and refinement statistics are listed in Table 1. The structure has been deposited in the PDB as entry 3H08.

Table 1
Data Collection and Refinement Statistics

Solvent-accessible surface area calculations were performed using AREAIMOL for each molecule in the asymmetric unit cell of CtepRNH*, EcoRNH*, TthRNH*, and a chimera of EcoRNH* and TthRNH*, TCEO (PDB entries 1F21, 1RIL, and 1JL2, respectively) (32). Figures were made using PyMOL (33).

Refolding Kinetic Experiments

Folding kinetics of CtepRNH* were monitored at 25 °C by following the CD signal at 222 nm at various final urea concentrations (0.6–4.6 M). All folding data were collected on an Aviv 202 stopped-flow instrument. Folding was initiated by a 1:11 dilution of unfolded CtepRNH* [8 mg/mL protein, 6.3 M urea, 20 mM NaOAc, and 50 mM KCl (pH 5.5)] into folding conditions [20 mM NaOAc, 50 mM KCl (pH 5.5), and 0.6–4.6 M urea]. An average of seven kinetic traces (CD signal vs time) were collected for each condition i. The kinetic traces were fit to single-exponential equations (A = Bekt + C) using SigmaPlot (SPSS, Chicago, IL). The burst phase CD signal was determined by calculating the signal at time zero using the parameters determined by the fit (B and C). The final signal was determined by the fit parameter C. Both the burst phase and final signals at various urea concentrations were fit to obtain free energies of unfolding for the intermediate (ΔGui) and native (ΔGun) states, respectively, assuming a two-state model and a linear dependence of ΔG on urea concentration (28).


Identification, Cloning, and Purification of a Putative RNase H from C. tepidum

A putative ribonuclease H gene from the bacterium C. tepidum was identified via a BLAST search (7) through homology to both E. coli and T. thermophilus RNases H. This gene was found to be 53% identical (65% similar) to EcoRNH and 57% identical (70% similar) to TthRNH. The gene was isolated from purified C. tepidum genomic DNA obtained from ATCC and subcloned into a pET27 vector. The resulting protein was overexpressed and purified via a two-column procedure (see Experimental Procedures). CtepRNH contains three cysteine residues at positions homologous to those of the three cysteines in E. coli RNase H. All three of these residues were mutated to alanine and subcloned into a pAED4 vector (27). The resulting protein, designated CtepRNH* [similar to the RNase H variants studied from T. thermophilus and E. coli (16, 34)], was purified via heparin and ion exchange columns, as described in Experimental Procedures (16).

C. tepidum RNase H* Folds into an Active Ribonuclease H

Basic biophysical characterization revealed that both CtepRNH and CtepRNH* are folded and active. Circular dichroism (CD) spectra and denaturation profiles confirmed that both proteins have a high degree of secondary structure, suggesting a well-folded protein. The CD spectra are also consistent with the RNase H fold and similar to those found for other RNases H (16). RNase H activity assays (23, 35) were performed, and the proteins were found to be active under standard assay conditions (Figure 2).

Figure 2
Biophysical attributes of C. tepidum RNase H*: (A) Activity assay of 20 and 50 nM C. tepidum RNase H* under standard assay conditions, as described in refs (21) and (27). (B) CD spectra of C. tepidum RNase H* taken before (●) and after (○) ...

The Structure of C. tepidum RNase H* Determined by X-ray Crystallography Reveals an RNase H Fold

To confirm that CtepRNH* adopts the RNase H fold, the crystal structure was determined by single-wavelength anomalous dispersion (SAD) methods from selenomethionine-labeled protein to 1.6 Å resolution (see Experimental Procedures). The structure contains two molecules in the asymmetric unit, both of which adopt a fold identical to those of most other known RNase H structures. In fact, alignment of the α-carbons of CtepRNH* with those of EcoRNH*, TthRNH*, and a chimera of EcoRNH* and TthRNH* (TCEO) (23) yields pairwise rmsds of 1.288, 1.450, and 1.224 Å, respectively (Figure 3). In comparison, the two molecules in the CtepRNH* asymmetric unit (chain A and chain B) gave an rmsd of 0.440 Å. From our examination of the crystal structure, it is clear that there are no major structural differences between C. tepidum and the other RNases H*.

Figure 3
Superposition of the E. coli (blue), C. tepidum (green for chain A and aqua for chain B), TCEO (yellow), and T. thermophilus (red) RNases H*. The α-carbon backbones of all four proteins are aligned and have an rmsd of 1.45 Å.

Thermodynamic Characterization of C. tepidum RNase H*

The free energies of unfolding (ΔGunf) of both wild-type and cysteine-free C. tepidum RNase H were determined by monitoring GdmCl-induced denaturation following both tryptophan fluorescence and the CD signal at 222 nm. The data were fit to a two-state model with a linear dependence of ΔGunf on GdmCl concentration (22, 36), resulting in a ΔGunf at 25 °C of 8.1 ± 0.5 kcal/mol with an m value of 4.1 ± 0.5 kcal mol−1 M−1 for CtepRNH*. The results obtained via fluorescence and CD overlap well (Figure 4), indicating apparent two-state folding. The denaturant dependence, or m value, is similar to that obtained for other RNases H, as expected for homologous proteins (22). The ΔGunf at 25 °C for CtepRNH is slightly higher (10.8 ± 0.9 kcal/mol). This stability difference between the wild-type and cysteine-free versions is comparable to the energetic difference between TthRNH and TthRNH* (13.3 and 12.2 kcal mol−1, respectively). Thermal denaturation following the CD signal at 222 nm demonstrated reversible unfolding (Figure 2) with a Tm of 68.5 °C for wild-type and 66.5 °C for cysteine-free C. tepidum RNase H. Again, this difference is similar to that seen between T. thermophilus RNase H (89 °C) and its cysteine-free variant (86 °C) (16). In all of the following thermodynamic studies, the cysteine-free version was used and the results for it were compared to the results from previous studies using cysteine-free variants of E. coli and T. thermophilus RNase H (where wild-type cysteines were replaced with alanines in the case of EcoRNH or a combination of alanines and serines in the case of TthRNH*).

Figure 4
GdmCl-induced equilibrium denaturation monitored by CD (○) and tryptophan fluorescence (●). Data were taken for 50 μg/mL C. tepidum RNase H* at 25 °C in 50 mM KCl and 20 mM NaAc (pH 5.5) with varying concentrations of GdmCl. ...

The ΔGunf of CtepRNH* at various temperatures was determined to generate a protein stability curve, which plots stability as a function of temperature (29). To generate this profile, a minimum of three GdmCl-induced denaturant melts were obtained at each temperature, and the resulting ΔGunf values for each melt were averaged. These averaged ΔGunf values, along with a single point generated from the thermal denaturation data for the TmGunf = 0), were fit to the Gibbs–Helmholtz equation (eq 1) (29, 36). The resulting curve and the individual stability measurements are plotted in Figure 5, along with the results from the previous studies on E. coli and T. thermophilus RNases H*. CtepRNH* clearly has a broader stability curve than EcoRNH* as reflected in a smaller change in heat capacity upon unfolding (Table 2). Surprisingly, the value obtained for the ΔCp of CtepRNH* is very similar to that of TthRNH* while most other thermodynamic parameters are closer to those of EcoRNH* (Table 2).

Figure 5
Protein stability curves of E. coli (–––), C. tepidum (—●—), and T. thermophilus (–·–) RNases H*. Each point represents the average of two or three ΔGunf values determined ...
Table 2
Thermodynamic Parameters of E. coli, C. tepidum, T. thermophilus, and a quintuple mutant of E. coli RNases H*a

Probing the Folding Pathway of C. tepidum RNase H*

To evaluate the energy landscape of CtepRNH* and determine if, like the well-characterized EcoRNH* and TthRNH*, the protein has a fast folding core, we evaluated its folding kinetics using stopped-flow circular dichroism. The time course of folding and unfolding was followed by monitoring the CD signal at 222 nm at various final urea concentrations. As has been seen with all other RNases H studied in this manner, at low concentrations of denaturant (<3 M urea), CtepRNH* has a large burst phase upon refolding, with >75% of the CD signal lost in the dead time of the instrument (12 ms), followed by the remaining folding process, which occurs on a slower, measurable, time scale (Figure 6A). Below ~2 M urea, the amplitude of this burst phase dominated the signal change to the extent that we could no longer confidently fit the remaining data (<10% signal change in the observable phase).

Figure 6
Kinetics of C. tepidum RNase H*. (A) Refolding of C. tepidum RNase H* in 2.8 M urea. Refolding was initiated by rapid dilution of unfolded protein (6.3 M urea) into folding conditions [20 mM NaOAc and 50 mM KCl (pH 5.5) with various urea concentrations] ...

These data suggest that, like EcoRNH* and TthRNH*, CtepRNH* folds in a three-state manner, populating a refolding intermediate at low concentrations of denaturant. The folding process was modeled as a three-state, on-pathway process, in which the intermediate is an obligatory step in the folding process. Indeed, single-molecule refolding data from EcoRNH* have shown that the intermediate of RNase H is on-pathway (37). The amplitude of the burst phase signal was fit independently (U and I) to determine the free energy of unfolding for the intermediate, ΔGui, and its linear dependence on urea, mui. On the basis of this analysis, we determined the stability of the intermediate of CtepRNH* to be 4.1 ± 1 kcal/mol, with an mui value of 1.0 ± 0.1 kcal mol−1 M−1. The stability of the intermediate falls directly between those of the intermediates of EcoRNH* and TthRNH* (3.5 and 4.5 kcal/mol, respectively), with similar m values (1.2 and 1.0 kcal mol−1 M−1). Data points taken from the final amplitude change observed in these kinetic studies were also fit to a two-state model (U and N), and the resulting ΔGunf is 7.5 ± 0.9 kcal/mol with an m value of 1.6 ± 0.2 kcal mol−1 M−1 (Figure 6B). As expected from other studies on RNases H, the ΔGunf determined from urea is slightly lower than that determined by GdmCl-induced denaturation studies (7.5 ± 0.9 kcal/mol vs 8.1 ± 0.5 kcal/mol).

The rates for each final concentration of denaturant were fit and plotted to create a chevron plot (Figure 6, inset). In other RNases H, we observed a rollover in the folding limb, which is attributed to the presence of a folding intermediate. In contrast, in the case of CtepRNH*, because of the dominating signal from the burst phase in low concentrations of urea, the observed rate was not fit at low denaturant concentrations, and therefore, there is no obvious rollover in the folding limb in this plot. On the other hand, we observed a distinct rollover in the CtepRNH* unfolding limb, suggesting the presence of an unfolding intermediate, that is not seen in other RNases H. Studies on other RNases H were limited in the range of denaturants accessible, but in this case, we were able to obtain data at denaturant concentrations much higher than the midpoint of the transition. Thus, it is not clear if this potential unfolding intermediate is a unique or general feature for this family of enzymes. Further investigation probing the possibility of an unfolding intermediate is underway. In sum, the kinetic refolding study on CtepRNH* shows that, much like EcoRNH*, it folds through an apparent three-state pathway. By analogy to the other RNases H, it is likely that this intermediate is composed of the same core region known to be a dominant factor in encoding the unusually low ΔCp of thermophilic RNase H.


Our previous studies on the stability profiles and folding pathways of two homologous RNases H (one from the mesophile E. coli and the other from the thermophile T. thermophilus) indicated that the ability of the TthRNH to function at high temperatures may depend on its unusually low change in heat capacity upon unfolding (ΔCp) (16) and that the hydrophobic core of the thermophilic protein was the region responsible for the low ΔCp. To probe the generality of these features, we identified and characterized a third homologous RNase H from the moderately thermophilic organism C. tepidum, with an optimal growth temperature between those of the two previously studied RNases H. All three proteins show pairwise sequence identities of >50%.

The three RNases H studied contain several free cysteines in their wild-type sequences [E. coli (three), T. thermophilus (four), and C. tepidum (three)]. Our previous studies on the energetics and folding of both E. coli RNase H and T. thermophilus RNase H were conducted on cysteine-free variants. Therefore, for ease of comparison, purification, and data collection, we created a similar variant for C. tepidum RNase H-CtepRNH*. Since the three free cysteines in CtepRNH are found in positions homologous to the positions of the three free cysteines in the E. coli protein, all three were mutated to alanine, as in EcoRNH*. As with the other two proteins, the cysteine-free variants are all active and maintain wild-type-like stability and melting temperatures. While it would be most informative to ascertain the activity of these RNases H at varying temperatures, because of the instability of the substrate at higher temperatures, the activity of each protein at the organism’s growth temperature cannot be determined and compared.

The crystal structure of CtepRNH* revealed the typical RNaseH-like fold. The similarity of the α-carbon rmsds and the solvent-accessible surface areas (SASA) of RNases H from E. coli, C. tepidum, and T. thermophilus and TCEO (a chimera of T. thermophilus and E. coli) indicates that any differences among these three proteins are not the result of any major structural feature and depend on either more subtle structural differences or features of their energy landscapes (1, 19, 23). We did not directly compare the details of our structure to any additional RNase H structures, such as those from humans, Sulfolobus tokodaii, Bacillus halodurans, and some retroviruses, since these have notable structural differences, often lacking specific secondary structural elements such as the basic loop (38-42).

Analysis of the protein stability curve shows a ΔCp for CtepRNH* that is much lower than both that of EcoRNH* and the value predicted on the basis of our crystal structure for CtepRNH*. Previous studies have experimentally determined that the change in heat capacity upon unfolding correlates with the change in hydrophobic solvent-accessible surface area upon unfolding (22, 43). Calculating the expected change in SASA upon unfolding based on the crystal structure predicts a ΔCp for CtepRNH* between 2.1 and 2.4 kcal mol−1 K−1, while our stability curve shows the ΔCp to be 1.7 kcal mol−1 K−1. The ΔCp determined by our data is quite close to that of TthRNH* (1.8 kcal mol−1 K−1), while that of the mesophilic protein EcoRNH* (2.7 kcal mol−1 K−1) is very close to the calculated value. On the basis of the crystal structures, it is predicted that these three homologous RNases H* should have quite similar ΔCp values. The ΔCp values from the stability curve data, however, differ by nearly 50%, leading to the hypothesis that both thermophilic RNases H* may have some residual structure in the unfolded state (24) and that a low ΔCp may be a common feature of RNases H from even moderately thermophilic organisms.

Previous work has shown that the hydrophobic core of RNase H has a major role in determining the ΔCp. A sequence alignment of all three RNases H shows that the core of CtepRNH* is more similar to the core of TthRNH* than to that of EcoRNH* (62% identical vs 54% identical). This similarity may indicate that the hydrophobic core of CtepRNH*, like that of TthRNH*, retains some structure in the unfolded state, resulting in a surprisingly low ΔCp and perhaps playing a key role in determining the protein’s thermophilic behavior.

While CtepRNH* shares some important similarities with TthRNH*, including a low ΔCp and reversible heat denaturation, both its maximum stability and melting temperature are very close to those of EcoRNH*. It appears that CtepRNH* represents a case in which these thermodynamic features (melting temperature, change in heat capacity upon unfolding, or maximum stability) have evolved independently from one another. Despite the differences in thermodynamic parameters, the stabilities of all three proteins at the optimal growth temperatures of each organism are remarkably similar (5.6 kcal mol−1 for T. thermophilus, 6.8 kcal mol−1 for E. coli, and 4.6 kcal mol−1 for C. tepidum).

The folding pathway of CtepRNH* appears to share features with both previously studied RNases H*. This moderately thermophilic protein populates a stable, partially folded intermediate very early in the folding pathway. Analysis of the burst phase signal indicates that the stability of the intermediate of CtepRNH* is 4.1 kcal/mol, while the intermediates of EcoRNH* and TthRNH* have stabilities of 3.6 and 4.5 kcal/mol, respectively. Previous results suggest that the stability of the intermediate of CtepRNH* should result in an increased stability of the overall protein relative to EcoRNH*, as the stabilities of the intermediates of EcoRNH* and TthRNH* account for ~35% of their overall stability. We have found that despite the increased stability of the intermediate, the overall stability of CtepRNH* is very close to that of EcoRNH*, meaning that the intermediate of CtepRNH* accounts for ~50% of the global stability of the protein.

Because of the increased relative stability of the intermediate, we had concerns that the protein was populating the intermediate in our experiments, invalidating the two-state approximation used to fit our equilibrium denaturation and obtain the ΔGunf values. However, this does not appear to be the case for several reasons. First, our denaturant data obtained via different methods are coincident, suggestive of a two-state transition. Second, a breakdown of the two-state assumption due to a notable population of an intermediate is known to result in an artificially low m value (44, 45). Homologous proteins are expected to exhibit similar m values (22), and our calculated m value is within error of those obtained for other RNases H where the two-state assumption has been validated by modeling based on extensive kinetic data, including the transient intermediate (45). Finally, modeling of our system, including the small population of the intermediate, does not significantly alter the stability profile we obtain. We believe, therefore, that the unusually low ΔCp is not a direct consequence of the population of the intermediate. The low ΔCp likely arises from residual structure in the unfolded state localized to this same core region, as demonstrated for TthRNH*.

The differences in the energetic properties of the RNases H underscore the subtlety of interactions leading to thermal stability. While it may be possible to increase the melting temperature of a protein through simple substitutional mutation, maintaining the finely tuned relationship among enzymatic activity, dynamics, and energetics that is required by biology at different temperature ranges clearly necessitates a more complex approach. Indeed, mutagenesis studies in which residues in E. coli RNase H were replaced with their counterparts from T. thermophilus RNase H have resulted in the classification of some residues as “thermophilic” due to the increased Tm of the protein upon substitution (12, 46, 47). These stabilizing mutations do not, however, recapitulate the thermodynamic profile we observe for thermophilic proteins. In fact, a recent structural thermodynamic study of one such quintuple mutant of E. coli RNase H reveals a maintained mesophile-like ΔCp with an increase in Tm (13), indicating that while the mutant protein is more thermally stable, it is not actually recapitulating the “thermophilic behavior” characteristic of the other RNases H (Table 2). Interestingly, many of these so-called thermophilic residues are also found in C. tepidum RNase H, including an inserted Gly present in T. thermophilus and C. tepidum but not in E. coli RNase H (Figure 1) (48). However, C. tepidum RNase H behaves quite differently from the mutant E. coli RNase H, serving as a strong confirmation that simple residue substitution cannot fully govern thermal stability.

Our studies indicate that the RNase H family may employ an unusually low ΔCp as a general strategy for achieving thermophilicity and that this likely represents a mechanism utilized by other protein families. Our results show that thermal stability (as defined by Tm) does not necessarily correlate with thermophilicity, as shown by the similarity in Tm values between CtepRNH* and EcoRNH*. The similarities in the folding pathway we observe among the different RNases H also suggest that all members of this family encode a marginally stable core of the protein that encompasses an early folding intermediate. The folding intermediate observed for RNase H from C. tepidum is likely composed of the same core region as seen by native state and pulsed labeling hydrogen exchange for both the E. coli and T. thermophilus RNases H. By analogy, we expect that the low ΔCp of C. tepidum RNase H arises from residual structure in the unfolded state encoded by this same core region. On the basis of this work, it should be possible to design sequences of RNases H that display the desired thermophilic or mesophilic properties as defined by their ΔCp and, therefore, fine-tune the energy landscape in a predictable fashion.


We thank Rachel Bernstein for critical reading of the manuscript and James Berger for crystallographic supplies and help.


Research in S.M.’s laboratory is supported by Grant GM50945 from the National Institues of Health.


circular dichroism
RNase H
ribonuclease H
change in heat capacity upon unfolding
change in free energy upon unfolding (GuGf)
ribonuclease H from Chlorobium tepidum
ribonuclease H from Escherichia coli
ribonuclease H from Thermus thermophilus
cysteine-free version of ribonuclease H from C. tepidum
cysteine-free version of ribonuclease H from E. coli
cysteine-free version of ribonuclease H from T. thermophilus
Protein Data Bank


1. Ishikawa K, Okumura M, Katayanagi K, Kimura S, Kanaya S, Nakamura H, Morikawa K. Crystal structure of ribonuclease H from Thermus thermophilus HB8 refined at 2.8 Å resolution. J. Mol. Biol. 1993;230:529–542. [PubMed]
2. Goedken ER, Keck JL, Berger JM, Marqusee S. Divalent metal cofactor binding in the kinetic folding trajectory of Escherichia coli ribonuclease HI. Protein Sci. 2000;9:1914–1921. [PubMed]
3. Knapp S, de Vos WM, Rice D, Ladenstein R. Crystal structure of glutamate dehydrogenase from the hyperthermophilic eubacterium Thermotoga maritima at 3.0 Å resolution. J. Mol. Biol. 1997;267:916–932. [PubMed]
4. Knapp S, Kardinahl S, Hellgren N, Tibbelin G, Schafer G, Ladenstein R. Refined crystal structure of a superoxide dismutase from the hyperthermophilic archaeon Sulfolobus acidocaldarius at 2.2 Å resolution. J. Mol. Biol. 1999;285:689–702. [PubMed]
5. Korndorfer I, Steipe B, Huber R, Tomschy A, Jaenicke R. The crystal structure of holo-glyceraldehyde-3-phosphate dehydrogenase from the hyperthermophilic bacterium Thermotoga maritima at 2.5 Å resolution. J. Mol. Biol. 1995;246:511–521. [PubMed]
6. Szilagyi A, Zavodszky P. Structural differences between mesophilic, moderately thermophilic and extremely thermophilic protein subunits: Results of a comprehensive survey. Structure. 2000;8:493–504. [PubMed]
7. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J. Mol. Biol. 1990;215:403–410. [PubMed]
8. Spassov VZ, Karshikoff AD, Ladenstein R. The optimization of protein-solvent interactions: Thermostability and the role of hydrophobic and electrostatic interactions. Protein Sci. 1995;4:1516–1527. [PubMed]
9. Berezovsky IN, Shakhnovich EI. Physics and evolution of thermophilic adaptation. Proc. Natl. Acad. Sci. U.S.A. 2005;102:12742–12747. [PubMed]
10. Zeldovich KB, Berezovsky IN, Shakhnovich EI. Protein and DNA sequence determinants of thermophilic adaptation. PLoS Comput. Biol. 2007;3(e5) [PubMed]
11. Tanner JJ, Hecht RM, Krause KL. Determinants of enzyme thermostability observed in the molecular structure of Thermus aquaticusd-glyceraldehyde-3-phosphate dehydrogenase at 25 Å Resolution. Biochemistry. 1996;35:2597–2609. [PubMed]
12. Kimura S, Nakamura H, Hashimoto T, Oobatake M, Kanaya S. Stabilization of Escherichia coli ribonuclease HI by strategic replacement of amino acid residues with those from the thermophilic counterpart. J. Biol. Chem. 1992;267:21535–21542. [PubMed]
13. Haruki M, Tanaka M, Motegi T, Tadokoro T, Koga Y, Takano K, Kanaya S. Structural and thermodynamic analyses of Escherichia coli RNase HI variant with quintuple thermostabilizing mutations. FEBS J. 2007;274:5815–5825. [PubMed]
14. Bae E, Phillips GN., Jr. Structures and analysis of highly homologous psychrophilic, mesophilic, and thermophilic adenylate kinases. J. Biol. Chem. 2004;279:28202–28208. [PubMed]
15. Christodoulou E, Vorgias CE. The thermostability of DNA-binding protein HU from mesophilic, thermophilic, and extreme thermophilic bacteria. Extremophiles. 2002;6:21–31. [PubMed]
16. Hollien J, Marqusee S. A thermodynamic comparison of mesophilic and thermophilic ribonucleases H. Biochemistry. 1999;38:3831–3836. [PubMed]
17. Topping TB, Gloss LM. Stability and folding mechanism of mesophilic, thermophilic and hyperthermophilic archael histones: The importance of folding intermediates. J. Mol. Biol. 2004;342:247–260. [PubMed]
18. Razvi A, Scholtz JM. Lessons in stability from thermophilic proteins. Protein Sci. 2006;15:1569–1578. [PubMed]
19. Katayanagi K, Miyagawa M, Matsushima M, Ishikawa M, Kanaya S, Nakamura H, Ikehara M, Matsuzaki T, Morikawa K. Structural details of ribonuclease H from Escherichia coli as refined to an atomic resolution. J. Mol. Biol. 1992;223:1029–1052. [PubMed]
20. Hollien J, Marqusee S. Structural distribution of stability in a thermophilic enzyme. Proc. Natl. Acad. Sci. U.S.A. 1999;96:13674–13678. [PubMed]
21. 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. [PubMed]
22. Myers JK, Pace CN, Scholtz JM. Denaturant m values and heat capacity changes: Relation to changes in accessible surface areas of protein unfolding. Protein Sci. 1995;4:2138–2148. [PubMed]
23. Robic S, Berger JM, Marqusee S. Contributions of folding cores to the thermostabilities of two ribonucleases H. Protein Sci. 2002;11:381–389. [PubMed]
24. Robic S, Guzman-Casado M, Sanchez-Ruiz JM, Marqusee S. Role of residual structure in the unfolded state of a thermophilic protein. Proc. Natl. Acad. Sci. U.S.A. 2003;100:11345–11349. [PubMed]
25. Guzman-Casado M, Parody-Morreale A, Robic S, Marqusee S, Sanchez-Ruiz JM. Energetic evidence for formation of a pH-dependent hydrophobic cluster in the denatured state of Thermus thermophilus ribonuclease H. J. Mol. Biol. 2003;329:731–743. [PubMed]
26. Eisen JA, Nelson KE, Paulsen IT, Heidelberg JF, Wu M, Dodson RJ, Deboy R, Gwinn ML, Nelson WC, Haft DH, Hickey EK, Peterson JD, Durkin AS, Kolonay JL, Yang F, Holt I, Umayam LA, Mason T, Brenner M, Shea TP, Parksey D, Nierman WC, Feldblyum TV, Hansen CL, Craven MB, Radune D, Vamathevan J, Khouri H, White O, Gruber TM, Ketchum KA, Venter JC, Tettelin H, Bryant DA, Fraser CM. The complete genome sequence of Chlorobium tepidum TLS, a photosynthetic, anaerobic, green-sulfur bacterium. Proc. Natl. Acad. Sci. U.S.A. 2002;99:9509–9514. [PubMed]
27. Doering DS, Matsudaira P. Cysteine scanning mutagenesis at 40 of 76 positions in villin headpiece maps the F-actin binding site and structural features of the domain. Biochemistry. 1996;35:12677–12685. [PubMed]
28. Santoro MM, Bolen DW. A test of the linear extrapolation of unfolding free energy changes over an extended denaturant concentration range. Biochemistry. 1992;31:4901–4907. [PubMed]
29. Becktel WJ, Schellman JA. Protein stability curves. Biopolymers. 1987;26:1859–1877. [PubMed]
30. Jones TA, Zou JY, Cowan SW, Kjeldgaard M. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr.A. 1991;47(Part 2):110–119. [PubMed]
31. Murshudov GN, Vagin AA, Dodson EJ. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr.D. 1997;53:240–255. [PubMed]
32. Lee B, Richards FM. The interpretation of protein structures: Estimation of static accessibility. J. Mol. Biol. 1971;55:379–400. [PubMed]
33. DeLano WL. The PyMOL Molecular Graphics System. DeLano Scientific; San Carlos, CA: 2002.
34. Kanaya S, Kimura S, Katsuda C, Ikehara M. Role of cysteine residues in ribonuclease H from Escherichia coli. Site-directed mutagenesis and chemical modification. Biochem. J. 1990;271:59–66. [PubMed]
35. Black CB, Cowan JA. Magnesium activation of ribonuclease H: Evidence for one catalytic metal ion. Inorg. Chem. 1994;33:5805–5808.
36. Pace CN, Laurents DV. A new method for determining the heat capacity change for protein folding. Biochemistry. 1989;28:2520–2525. [PubMed]
37. Cecconi C, Shank EA, Bustamante C, Marqusee S. Direct observation of the three-state folding of a single protein molecule. Science. 2005;309:2057–2060. [PubMed]
38. You DJ, Chon H, Koga Y, Takano K, Kanaya S. Crystal structure of type 1 ribonuclease H from hyperthermophilic archaeon Sulfolobus tokodaii: Role of arginine 118 and C-terminal anchoring. Biochemistry. 2007;46:11494–11503. [PubMed]
39. Nowotny M, Gaidamakov SA, Crouch RJ, Yang W. Crystal structures of RNase H bound to an RNA/DNA hybrid: Substrate specificity and metal-dependent catalysis. Cell. 2005;121:1005–1016. [PubMed]
40. Davies JF, II, Hostomska Z, Hostomsky Z, Jordan SR, Matthews DA. Crystal structure of the ribonuclease H domain of HIV-1 reverse transcriptase. Science. 1991;252:88–95. [PubMed]
41. Das D, Georgiadis MM. The crystal structure of the monomeric reverse transcriptase from Moloney murine leukemia virus. Structure. 2004;12:819–829. [PubMed]
42. Nowotny M, Gaidamakov SA, Ghirlando R, Cerritelli SM, Crouch RJ, Yang W. Structure of human RNase H1 complexed with an RNA/DNA hybrid: Insight into HIV reverse transcription. Mol. Cell. 2007;28:264–276. [PubMed]
43. Gomez J, Hilser VJ, Xie D, Freire E. The heat capacity of proteins. Proteins. 1995;22:404–412. [PubMed]
44. Mayne L, Englander SW. Two-state vs. multistate protein unfolding studied by optical melting and hydrogen exchange. Protein Sci. 2000;9:1873–1877. [PubMed]
45. Spudich G, Marqusee S. A change in the apparent m value reveals a populated intermediate under equilibrium conditions in Escherichia coli ribonuclease HI. Biochemistry. 2000;39:11677–11683. [PubMed]
46. Ishikawa K, Kimura S, Kanaya S, Morikawa K, Nakamura H. Structural study of mutants of Escherichia coli ribonuclease HI with enhanced thermostability. Protein Eng. 1993;6:85–91. [PubMed]
47. Kimura S, Kanaya S, Nakamura H. Thermostabilization of Escherichia coli ribonuclease HI by replacing left-handed helical Lys95 with Gly or Asn. J. Biol. Chem. 1992;267:22014–22017. [PubMed]
48. Ishikawa K, Nakamura H, Morikawa K, Kimura S, Kanaya S. Cooperative stabilization of Escherichia coli ribonuclease HI by insertion of Gly-80b and Gly-77 → Ala substitution. Biochemistry. 1993;32:7136–7142. [PubMed]