PEDS Advance Access published online on December 19, 2007
Protein Engineering Design and Selection, doi:10.1093/protein/gzm072
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
An adaptive mutation in adenylate kinase that increases organismal fitness is linked to stability–activity trade-offs
Department of Biochemistry and Cell Biology, Rice University, 6100 Main Street, MS-140 Houston, TX, USA
3 To whom correspondence should be addressed. E-mail: shamoo{at}rice.edu
 |
Abstract |
|---|
 Top Abstract Materials and methodsResultsDiscussionFundingAcknowledgementReferences |
|---|
Protein function is a balance between activity and stability. However, the relevance of stability–activity trade-offs for protein evolution and their impact on organismal fitness have been difficult to determine. Previously, we have linked organismal survival at increasing temperatures to adaptive changes to a single protein sequence through allelic replacement of an essential gene, adenylate kinase (adk), in a thermophile. In vivo continuous evolution of the temperature-sensitive thermophile has shown that the first step toward increased organismal fitness is mutation of glutamine-199 to arginine in the mesophilic enzyme (AKsub Q199R). Here, we show that although substitution of Arg-199 did confer a modest increase in stability (0.6 kcal mol–1at 20°C;
Tm = 3.0°C), it is a large change in the activity profile of the enzyme that is responsible for its exceptional robustness during the earlier experimental evolution study. Kinetic studies of AKsub Q199R show that it has a strong loss of enzymatic activity (>50%) at lower temperatures (20–45°C) and a subsequent increase at elevated temperatures. The stability–activity trade-off observed for AKsub Q199R was linked to the rigidification of the overall structure through stabilization of a polypeptide loop containing Arg-199 that is part of the ATP-binding site of the enzyme. Structural analysis revealed the formation of new ionic interactions facilitated by Arg-199. Our results suggest that stability–activity trade-offs are employed readily as an evolutionary strategy during natural selection to increase organismal fitness.
Keywords: adaptive evolution/adenylate kinase/molecular evolution/thermostability/trade-offs
Protein flexibility is crucial for enzyme catalysis (Creighton, 1993
). During the course of evolution, natural selection is likely to have favored protein sequences that could reach an optimal compromise between stability and function without decreasing the fitness of the organism (DePristo et al., 2005). The increased stability of many thermophilic proteins is accompanied by loss of protein flexibility and reduced enzymatic activity at low temperatures (Jaenicke, 1991; Somero, 1995; Zavodszky et al., 1998; Wolf-Watz et al., 2004). Moreover, changes in catalytic residues that increase protein stability typically decrease activity and suggest that function often comes with a substantial penalty to stability (Yutani et al., 1987; Meiering et al., 1992; Beadle and Shoichet, 2002; Mukaiyama et al., 2006). Although protein function appears to be a compromise between activity and stability, it has been difficult to establish the extent (if any) to which stability–activity trade-offs limit protein adaptation.
In vivo experimental evolution of a single gene provides an opportunity to investigate the functional intermediates of protein adaptive evolution under natural selection (Couñago et al., 2006). To identify these intermediates, we developed the ‘weak link’ approach to study adaptation of a single gene within a bacterial population (Couñago et al., 2006). In the weak link approach, survival of a genetically modified thermophilic organism (Geobacillus stearothermophilus) at elevated temperatures is directly linked to changes in the essential enzyme adenylate kinase derived from the mesophile Bacillus subtilis (AKsub). In vivo, adenylate kinase (AK) (E.C. 2.7.4.3) function is essential for adenylate homeostasis and energy metabolism (Noda, 1973). Recombinant G.stearothermophilus cells expressing AKsub are unable to grow at temperatures higher than 55°C due to heat inactivation of the mesophilic enzyme and consequent disruption of adenylate homeostasis and energy metabolism (Couñago and Shamoo, 2005).
Bacillus subtilis AK is a small (217 amino acids) monomeric protein (Fig. 1). Structural and biochemical data for AK orthologs isolated from mesophilic, thermophilic and psychrophilic bacteria have suggested that AK activity is dependent on the maintenance of a flexible active site within the temperature ranges of the parent organism (Glaser et al., 1992; Perrier et al., 1994; Berry and Phillips, 1998; Bae and Phillips, 2004). During catalysis, the LID and AMP-binding domains of the enzyme undergo large hinge-bending motions over the more static core domain (Vonrhein et al., 1995; Muller et al., 1996). As in many thermostable proteins, the greater stability of G.stearothermophilus AK is facilitated by an increase in the number of ionic interactions that enhances protein thermostability but reduces flexibility and activity at lower temperatures (Berry and Phillips, 1998; Bae and Phillips, 2004). Together, these characteristics make AKsub an excellent subject for exploration of the constraints underlying the landscape of protein evolution.
|
(distance: 3.1Å) and NH1 (distance: 2.8Å) from Arg-199; hydrogen bonding ion pairs (alternating red and black dashed lines) are seen between O
1 of Asp207 and the NH1 (distance: 2.6Å) and NH2 (distance: 2.8Å) of Arg-199; and an ion pair (red dashed line) is formed between OMutation of glutamine-199 to arginine (AKsub Q199R) originated spontaneously in a large population of recombinant G.stearothermophilus cells subjected to a continuous temperature gradient (55–70°C) for
1500 generations. Cells expressing AKsub Q199R appeared during the first 50 generations and dominated the bacterial population for the ensuing 750. During this period, mutant AKsub Q199R cells drove all competition to extinction. Eventually, AKsub Q199R itself was driven to extinction but not before out competing an estimated 12 million mutant and wild-type cells during the course of the experiment (Couñago et al., 2006). In our original study, we had hypothesized that a modest increase in stability was responsible for the success of the strains carrying the mutation but could not explain how the mutation facilitated such a broad range of success in vivo (9°C) when the change in in vitro stability of the mutant protein was only 3°C. In contrast, strains with wild-type AK do not survive above the temperature at which AK unfolds. The current study uses a combination of enzyme kinetics, folding studies and a more detailed evaluation of the AKsub Q199R structure to elucidate the mechanisms responsible for the increased fitness observed for strains carrying AKsub Q199R. AKsub Q199R is just one mutation away from its ancestral gene. As the smallest unit of change to a protein molecule, the single mutation Q199R offers a unique opportunity to study the structural mechanisms behind protein function and stability in a physiologically relevant scenario. Here, we investigated the physical origins of the adaptive success of AKsub Q199R. Our results suggest that stability–activity trade-offs are employed readily by natural selection during protein adaptive evolution. |
Materials and methods |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionFundingAcknowledgementReferences |
|---|
Protein expression and purification
Point mutations to the wild-type B.subtilis adk gene were made by site-directed mutagenesis to the parental gene in vector pET11a (Novagen, Gibbstown, NJ, USA), which was the generous gift of Dr G.N. Phillips (University of Wisconsin at Madison). All nucleotide changes were confirmed by DNA sequencing. Proteins were overexpressed and purified from Escherichia coli as previously described (Couñago et al., 2006) and quantitated by a modified Bradford assay (Bio-Rad, Hercules, CA, USA) and UV absorbance.
Measurements of the velocity of the forward reaction (MgATP+AMP
MgADP+ADP) were performed by an end-point assay at the indicated temperatures (Saint Girons et al., 1987; Glaser et al., 1992). The reaction mixture consisted of 4.0 ml of reaction buffer (25 mM phosphate buffer pH 7.2, 5 mM MgCl2, 65 mM KCl, 1 mM DTT, 1.4 mM ATP and various AMP concentrations 5, 10, 30, 75, 150, 300, 700 or 1000 µM). This mixture was kept at the desired temperatures for 5 min in a BoekerGrant BD120l water bath prior to addition of AK to a final concentration of 1 nM. The reaction was stopped at various time points by transferring 300 µl aliquots from the reaction mixture to an ice-cold tube containing the AK inhibitor P1,P5- di(adenosine-5’)pentaphosphate (Ap5A) to a final concentration of 0.2 mM. The amounts of ADP produced at various time points were estimated using a secondary assay. In the secondary reaction, 100 µl of a solution containing 0.3 mM NADH, 0.5 mM phosphoenolpyruvate and 5 units of both lactate dehydrogenase and pyruvate kinase were added to the various time point aliquots and the amounts of ADP produced estimated by the conversion of NADH to NAD+ as indicated by absorbance at 340 nm. Initial velocities (v0) for various AMP concentrations and temperatures were estimated by progression curves of ADP production versus time. KM and Vmax were estimated in a v0 versus [AMP] plot by fitting the data to the Michaelis–Menten equation using a non-linear least square fitting routine in SigmaPlot 10.0 (Systat, San Diego, CA, USA). Data shown are the average of two independent experiments, performed in triplicate, plus standard deviations. Pyruvate kinase, lactate dehydrogenase, phosphoenolpyruvate, NADH, AMP, ADP, ATP and Ap5A were purchased from Sigma- Aldrich (St Louis, MO, USA).
Thermodynamic parameters for activation for the forward reaction of AK were calculated as described (Lonhienne et al., 2000) using the following equations:
|
| (1) |
|
| (2) |
|
| (3) |
Far-UV circular dichroism (CD) traces were acquired for 20 µM protein in 10 mM potassium phosphate pH 7.2. Solvent screening experiments were performed at 500 mM NaCl and 10 mM potassium phosphate pH 7.2. The stability of the ligand bound, closed conformation was estimated in the presence of 100 µM Ap5A. Thermodenaturation of protein samples was found to be more than 90% reversible when samples were heated to temperatures that resulted in denaturation of 90% of the protein and then returned to 20°C. The transitions showed no dependence on scan rate or protein concentration within the 10–100 µM protein range tested. Fraction of unfolded protein was calculated by following the changes in ellipticity at 220 nm as a function of temperature (1°C per minute) from 20°C to 90°C with three measurements per minute. To minimize errors due to baseline correction, the thermal unfolding mid-points (Tm) were calculated using the first derivative of the CD signal versus temperature (John and Weeks, 2000). Data shown are the average of two independent experiments.
Time-resolved folding and unfolding were probed by far-UV CD (at 220 nm) using an Applied Photophysics Pi-Star stopped-flow mixer. Both wild-type and Q199R AK were mixed in a 1:10 ratio with appropriate GuHCl/buffer solutions (10 mM potassium phosphate pH 7.2). Six kinetic traces were averaged and fit to monophasic decay equations. The observed kinetic traces for the folding and unfolding rate (kF and kU, respectively) were devoid of missing amplitudes (within the 2–3 ms dead time) and free of any apparent protein concentration dependence, within the 10–100 µM protein range tested. Unfolding and refolding rate constants at different GuHCl concentrations were fit assuming standard linear dependence of ln kF and ln kU on GuHCl concentration. Equilibrium GuHCl titrations were performed for each protein to determine the appropriate GuHCl conditions to use in the kinetic experiment.
PDB entries for the analyzed adenylate kinase structures were: AKsub 1P3J
[PDB]
(Bae and Phillips, 2004), AKsub Q199R 2EU8 (Couñago et al., 2006) and AKste 1ZIO (Berry and Phillips, 1998). All properties were analyzed for residues 1–212, as the last five residues are missing from the atomic coordinates of AKsub. Structures were aligned with PyMOL (DeLano, 2002). Two oppositely charged residues were identified as an ion pair if their closest oppositely charged atoms were within a 4.0Å cutoff. Carboxylic oxygen atoms of Asp, Glu and the C-terminal residue were considered as negatively charged atoms; amino nitrogen atoms of Arg, Lys, His and the N-terminal residue as positively charged atoms. The HB2 options in WHAT IF (Vriend, 1990) were used to define hydrogen bonds. In HB2, the total hydrogen bonding energy is optimized to find the optimal positions for all hydrogen atoms simultaneously, and hydrogen bonds are determined from the donor/acceptor types, the H-acceptor distance, the donor-H-acceptor angle and the position of the hydrogen with respect to the acceptor. Polar and apolar, exposed and buried surface areas were calculated with WHAT IF using a probe radius of 1.4Å. Nitrogen and oxygen atoms were considered polar and carbon and sulfur atoms apolar. All figures with molecular structures were generated using PyMOL (DeLano, 2002).
|
Results |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionFundingAcknowledgementReferences |
|---|
The mutant enzyme trades off activity at lower temperatures for increased activity at higher temperatures
The increased stability of the mutant enzyme is accompanied by a reduction in its catalytic efficiency at 20–60°C. In vivo, adenylate kinase regenerates ADP from AMP and ATP at near equilibrium rates maintaining the cellular homeostasis of adenylate species (Noda, 1973). Many thermophilic enzymes display reduced catalytic activity below physiologically relevant temperatures, an observation usually associated with increased protein stability and ‘rigidification’ at lower temperatures (Jaenicke, 1991; Somero, 1995; Zavodszky et al., 1998; Wolf-Watz et al., 2004). In order to investigate if this phenomenon was also observed for AKsub Q199R, we have estimated the temperature dependence of the steady-state kinetic parameters (KM, Vmax and kcat) for AMP utilization (Fig. 2 and Table I). AKsub Q199R displayed a 2-fold increase in KM (AMP) at all investigated temperatures compared with the wild-type enzyme (Table I). Although the Vmax and kcat for AKsub Q199R is lower than that for the parent enzyme at lower temperatures (27–45°C), they exceed the wild-type at elevated temperatures (55–60°C). As expected for a mesophilic enzyme, AKsub displayed higher catalytic efficiency (kcat/KM) at physiological temperatures for B.subtilis (20–40°C). The catalytic efficiency (kcat/KM) of AKsub Q199R was found to be 2-fold lower than the wild-type at reduced temperatures (20–45°C). Nevertheless, the difference in catalytic efficiency between the mutant and the wild-type enzyme is reduced at elevated temperatures (55–60°C) (Table I) and suggests that only the catalytic efficiency at lower temperatures has been compromised. The KM for ATP was also measured at 20°C, 30°C and 37°C for both AKsub and AKsub Q199R. The KM for ATP of AKsub and AKsub Q199R was comparable (9 ± 1, 10 ± 2 and 8 ± 2 versus 8 ± 1, 10 ± 2 and 9 ± 2 µM at 20°C, 30°C and 37°C, respectively) and had less temperature dependence than that of AMP.
|
|
The catalytic rate of AKsub Q199R displays stronger temperature dependence than that of the wild-type enzyme
The temperature dependence of the kinetic parameters KM and kcat for AMP are shown in Fig. 2C and D, and are based on values shown in Table I. The KM for both enzymes displays a linear relationship with temperature (Fig. 2C). The slope and intersection of ln KM versus 1/T was used to calculate Van't Hoff's enthalpy (HKM) and entropy (TSKM) and showed that the affinity for AMP for the mutant (HKM=13.1 ± 0.1 and TSKM=1.3 ± 0.1 kcal mol–1) and wild-type (HKM=11.0 ± 0.1 and TSKM=1.3 ± 0.1 kcal mol–1) have similar dependence on temperature. The catalytic rate for both enzymes also displays a linear relationship with temperature; however, as seen on a plot of ln kcat versus 1/T (Fig. 2D), the kcat of AKsub Q199R displays stronger temperature dependence than that of the wild-type enzyme. The Ea for wild-type and mutant enzymes calculated from the slope of ln kcat versus 1/T plot are 7.7 ± 0.1 and 14.8 ± 0.1 kcal mol–1, respectively.
We have calculated the thermodynamic activation parameters for the reaction (G#, H# and TS#) for both AKsub and AKsub Q199R from their kcat and Ea values (Tables II and III) according to Eqs (1–3) (Lonhienne et al., 2000). The free energy of activation (G#) for both enzymes was found to be of similar magnitude. However, AKsub Q199R displays a2-fold increase in activation enthalpy (H#) and a 3-fold reduction in the magnitude of its activation entropy term (TS#) when compared with AKsub (Table II). Absolute values of the thermodynamic parameters should be interpreted cautiously; however, the relative changes of enthalpy (H#) and entropy (S#) of AKsub and AKsub Q199R suggest significant changes during adaptation to higher temperatures (Lonhienne et al., 2000).
|
|
Disruption of the Arg-199
Asp-207 electrostatic interaction by mutagenesis destabilizes AK
We have used site-directed mutagenesis to confirm the role of ion pair Arg-199Asp-207 in the increased stability of the evolved mutant enzyme. We introduced mutation D207N into the AKsub Q199R background. The mid-point denaturation temperatures (Tm) for wild-type and mutant adenylate kinases were estimated by following the changes in molar ellipticity at 220 nm upon thermal denaturation (Table III). The resulting double mutant can no longer form the ionic interactions facilitated by residue Asp-207 and, as expected, mutant AKsub Q199R/D207N displays reduced stability when compared with AKsub Q199R (Fig. 3A and B and Table III).
|
Solvent screening disrupts the increased thermostability of mutant Q199R
The contributions of solvent-exposed ionic interactions to protein stability were assessed by estimating the Tm of the protein at increased salt concentrations (Lee et al., 2005). At low ionic strength (10 mM potassium phosphate pH 7.2), the difference in the Tms of wild-type and mutant AKsub Q199R was found to be 3.0°C (Fig. 3A and B and Table III). In the presence of 500 mM NaCl, the contributions of solvent-exposed ionic interactions to protein stability were greatly reduced due to solvent screening, and under these high salt conditions both wild-type and mutant proteins displayed equivalent Tms (Tm=–0.8°C) (Fig. 3B and Table III).
Owing to the stabilizing Hofmeister effect at high ionic strengths (Collins and Washabaugh, 1985), both wild-type and mutant enzymes showed a >15°C increase in stability in 500 mM NaCl. In order to confirm that the stabilizing Hofmeister effect contributes similar to the stability of both proteins, we measured the Tm of a different thermostable adenylate kinase mutant, Q199R/T179I. Mutation of Thr-179 to isoleucine did not result directly in the formation of any new ionic interactions (to be published). Therefore, the difference in Tm (Tm) between Q199R and Q199R/T179I should be the same, regardless of the salt concentration. Indeed, the Tm between Q199R and Q199R/T179I at low and high ionic strengths was found to be 2.3°C and 2.9°C, respectively (Fig. 3B and Table III).
Mutant Q199R is only marginally more thermostable than wild type in the presence of a substrate analog
The atomic structures for both wild-type and mutant enzymes revealed that the main-chain carbonyl oxygen of residue 199 interacts with the N6 amino group of the adenine ring from inhibitor Ap5A bound to the ADP/ATP binding site (Fig. 1B). Since position 199 is implicated in substrate binding and a substantial amount of AK will be bound to substrate in vivo, we have measured the stabilities of wild-type and mutant enzymes in their ligand bound form. It has been previously shown that AKsub and AKste are greatly stabilized when bound to Ap5A (Bae and Phillips 2004). Likewise, saturating concentrations of Ap5A stabilized AKsub Q199R by >15°C (Fig. 3C and Table III). These experiments showed that in the Ap5A bound, closed conformation, the mutant enzyme is only marginally more thermostable than wild-type. The Tm between the two proteins was reduced from 3.0°C to 0.3°C in the presence of inhibitor (Fig. 3C and Table III).
Increased stability of the mutant enzyme is accompanied by a reduction in its folding and unfolding rates
In order to examine the contribution of mutation Q199R under a less static perspective than X-ray diffraction or equilibrium methods, we have investigated the folding kinetics for both AKsub and AKsub Q199R. Time-resolved folding and unfolding experiments were monitored by changes in far-UV CD at 220 nm at 20°C as a function of the chemical denaturant guanidine hydrochloride (GuHCl). All kinetic measurements were single-exponential processes and devoid of missing amplitudes within the dead time (2–4 ms) of the instrument. The semi-logarithmic plot (Chevron plot) of folding (kF) and unfolding rate (kU) constants versus GuHCl concentration exhibited the characteristic V-shape typical of two-state kinetic behavior (Fig. 4). The folding speeds for both wild-type and Q199R mutant in water at 20°C were found to be slow (0.030 and 0.015 s–1, respectively, Table IV). The ‘rigidification’ of the mutant enzyme caused by mutant residue Arg-199 observed in the crystal structure is consistent with the 2.0 and 5.3-fold reduction in folding and unfolding rates, respectively (Fig. 4 and Table IV). Protein flexibility is crucial for enzyme catalysis (Creighton, 1993), and adenylate kinases have been shown to undergo dramatic domain movements upon substrate binding, catalysis and product release (Muller et al., 1996; Wolf-Watz et al., 2004).
|
|
The free energy of unfolding in water,
GU(H2O), was derived from the unfolding and folding rate constants in 0 M GuHCl (Pace and Shaw, 2000) and was found to be 3.4 and 4.0 kcal mol–1 for wild-type and mutant proteins, respectively (Table IV). These values are in excellent agreement with equilibrium experiments in GuHCl (Fig. 4). The meq-value is a measure of the dependence of GU on GuHCl concentration and is thought to reflect the change in solvent exposed surface area of the protein upon folding (Pace and Shaw, 2000). The calculated meq-values (meq=mF – mU) for both wild-type and mutant enzymes were found to be similar. However, mF is larger for AKsub resulting in the folding transition state being more native like for the wild-type protein (βF=0.45) than for the mutant (βF=0.26) (Table IV). Increased thermostability of mutant Q199R is the result of a new interaction network
Structural features such as reduced solvent exposed area and increased number of ion pairs and hydrogen bonds have been implicated with the increased stability of enzymes isolated from thermophilic organisms (Jaenicke, 1991; Bae and Phillips, 2004; Berezovski and Shakhnovich, 2005; Lee et al., 2005). In order to establish the structural basis for the overall increase in the thermostability of mutant AKsub Q199R, we compared the structures of the mutant enzyme with that of its parent, the mesophilic protein from B. subtilis (AKsub) and the thermophilic ortholog from G. stearothermophilus (AKste) (Berry and Phillips, 1998; Bae and Phillips, 2004; Couñago et al., 2006). All three enzymes have been previously crystallized with the inhibitor Ap5A bound to the active site and have nearly identical overall structures (Table V). By chance, mutant AKsub Q199R crystallized with two protein molecules (A and B in Tables V and VI) in the crystallographic asymmetric unit. In contrast, AKsub and AKste crystals have only one protein molecule per asymmetric unit. This difference in crystallographic packing provided an additional level of validation for our structural analyses since any significant differences would likely be observed in both copies of AKsub Q199R. Surprisingly, many changes observed between AKsub and AKsub Q199R were of comparable magnitude or number to that seen between copies A and B of AKsub Q199R (Table VI) and were useful for identification of the adaptive changes in the structure of AKsub Q199R.
|
|
Arg-199 in the mutant protein is directly involved in facilitating ionic interactions with Asn-196 and Asp-207 (Fig. 1B). Interestingly, the Q199R mutation also indirectly facilitated the formation of additional solvent-exposed ion pairs not found in the wild-type enzyme (Table VI). Nevertheless, only three of these new ion pairs were present in both copies of the mutant protein in the asymmetric unit. Although crystallization conditions may affect the distribution of ion pairs on the surface of a protein, both AKsub and AKsub Q199R were crystallized at similar ionic strengths. We found no significant difference in the accessible surface areas or the number of hydrogen bonds between AKsub and the AKsub Q199R structures beyond the variability observed between copies A and B of AKsub Q199R in the asymmetric unit (Table VI).
|
Discussion |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionFundingAcknowledgementReferences |
|---|
Mutation Q199R originated spontaneously in a large population of cells undergoing selection for increased AK thermostability. Cells expressing AKsub Q199R dominated the bacterial population for over 750 generations and display a significantly broader range of success (9°C) than would have been anticipated from preliminary characterization of AKsub Q199R in vitro (Couñago et al., 2006
).
At the structural level, the adaptive success of AKsub Q199R is due to the formation of new interactions between Arg-199 and residues Asp-207 and Asn-196 (Fig. 1). If the Arg-199Asp-207 interaction is abolished by solvent screening or site-directed mutagenesis, the stability of the enzyme is reduced to that of wild-type AKsub (Table III). The modest 3.0°C increase in stability conferred by an additional ion pair between Arg-199Asp-207 and hydrogen bonds between Arg-199Asn-196 is well within the expected value for such interactions. For example, the transfer of individual solvent exposed ion pairs found in AKste was shown to increase the stability of the mesophilic AKsub enzyme between 0.1°C and 3.3°C (Bae and Phillips, 2005). The Q199R mutation also facilitates an increase in the overall number of electrostatic interactions throughout the core of the structure (Table VI). Most of these interactions are not new but are the product of modest changes in the positions of side-chain rotamers that bring interactions within the 4Å cut-off used in the analysis. An increase in the overall number of interactions was quite unexpected and supports the notion that thermostability can be facilitated readily by changes to the network of electrostatic interactions on the protein surface that increase their number and complexity (Kumar and Nussinov, 1999).
The most likely mechanisms behind the increased stability of AKsub Q199R are the interactions formed between Arg-199 and Asn-196 and Asp-207 at the local level that in turn propagate through the core domain. The loop where Arg-199 is located connects the last two secondary structures in AKsub: β-strand 8 and 
-helix 9 (Fig. 1). Crystallographic temperature factors report on many sources of protein disorder, including the intrinsic flexibility of proteins (Wray et al., 1999). For AKsub, -helix 9 has the highest main-chain temperature factors of the entire structure and residues 213–216 are completely disordered (average B-factor for residues 202–212=42.3 Å2 versus 25.1 Å2 for AKsub overall). In contrast, temperature factors for the same helix in both copies of AKsub Q199R show nearly comparable thermal motions with the overall structure and are very well ordered (average B-factor for residues 202–216=14.4 Å2 versus 13.3 Å2 for AKsub Q199R overall). It is therefore possible that Arg-199 serves as a ‘staple’ to favor interaction of the C-terminal -helix 9 with the twisted β-sheet made up of β-strands 1, 4 and 8, thus promoting the ‘rigidification’ of the protein.
The Gibbs free energy of unfolding [GU(H2O)] describes the net effect of folding and unfolding rates (for a two-state process: GU(H2O)=RTln[kF/kU]). Mutations can stabilize proteins by slowing unfolding, speeding folding or some combination of the two. For a related eubacterial adenylate kinase (E.coli AK, 47% identical, 1.0 Å overall r.m.s.d.), the refolding reaction involved a fast collapse of the hydrophobic core domain followed by slow formation of secondary elements, including β-8 (Ratner et al., 2005). Consistent with the idea that its structure is rigidified by the Arg-199 ‘staple’, mutant AKsub Q199R displays both slower folding and unfolding rate constants compared with wild type. It is interesting to note that thermostable proteins consistently display slower unfolding rates when compared with their mesophilic counterparts (Ogasahara et al., 1998; Dams and Jaenicke, 1999; Mukaiyama et al., 2004), suggesting that introduction of a kinetic barrier to unfolding might be an evolutionary strategy to increase protein function at higher temperatures (Wittung-Stafshede, 2004).
Recently, in vitro directed evolution has demonstrated that it is possible to decouple catalytic efficiency from stability (Miyazaki et al., 2000; Sriprapundh et al., 2003). It has been difficult, however, to establish the extent to which stability–activity trade-offs constrain the adaptive landscape during protein evolution. Loss of catalytic efficiency by thermophilic proteins at low temperatures as well as the stability of psychrophilic enzymes at high temperatures may reflect evolution and not the necessity to trade-off the characteristics of activity and stability. Combined with our structural and stability data, the enzyme kinetics studies show that AKsub Q199R is a protein where a stability–activity trade-off has occurred. At the temperature growth range for B.subtilis (27–35°C), AKsub displays a catalytic efficiency almost 2.5-fold higher than the one observed for mutant AKsub Q199R. At temperatures where cells expressing AKsub Q199R dominated the bacterial population during the evolution experiment (55–60°C), the difference in catalytic efficiency between parent and evolved enzymes is reduced, but is still higher for the wild-type protein (Table I). Although analysis of activity data alone would predict a slight advantage for the wild-type protein at the 50–60°C range, the in vivo concentration of substrate is in the millimolar range, and as a consequence the 2-fold difference in KM will not be limiting, and Vmax at high temperature will most likely play the determining role in fitness (using the measured KM for AMP as the equilibrium constant). The energy charge (EC=[ATP]+1/2[ADP]/[ATP+ADP+AMP]) has been measured to be 0.4 for G.stearothermophilus and strongly support earlier work suggesting that in vivo concentrations of adenylate species are well above the KM of either wild type or AK Q199R (Atkinson, 1968; Counago and Shamoo, 2005). Although AMP levels are thought to be limiting in vivo, it is also possible that changes in ATP affinity could also contribute to the observed trade-off; however, two observations argue against this idea: (i) the temperature dependence of the KM for ATP were less than that of AMP and (ii) the KM for AKsub and AKsub Q199R are both greater than that for AMP. In contrast to the KM, the temperature dependence of Vmax is greatly accentuated for the mutant enzyme and is reflected in its catalytic rate (kcat). The kcat of the mutant enzyme is increased by mutation Q199R. Stabilization of the protein in the loop adjoining the C-terminal and near the ATP binding site has decreased the catalytic rate of the enzyme at low temperatures, whereas increasing it in the temperature range where cells expressing AKsub Q199R out-competed the wild type. Moreover, the temperature at which kcat for AKsub Q199R superseded the one for the wild-type enzyme, 48°C (Fig. 2), is in excellent agreement with the observation that cells expressing wild-type AKsub were lost very quickly during the evolution experiment (Couñago et al., 2006). Our stability studies also suggest that at higher temperatures, the amount of properly folded enzyme is considerably larger for mutant AKsub Q199R (Fig. 3B). In the absence of substrate at 55°C, nearly 40% of the wild-type enzyme was unfolded when compared with 25% of AKsub Q199R. Thus, in addition to trading overall substrate affinity, as seen by the KM, the mutant enzyme has also traded activity at low temperatures for increased stability and activity at higher temperatures.
The lowered catalytic efficiency of AKsub Q199R is likely to be a direct consequence of the rigidification induced by the Arg-199 ‘staple’. For adenylate kinases, the large movements of the AMP-binding and LID domains during catalysis depend on hinge-like movements and expose a large fraction of the enzyme to the solvent (Vonrhein et al., 1995; Muller et al., 1996). Such large movements of active site-related domains are likely to be detrimental to the overall protein stability. Nevertheless, increasing the stability of the activity site may impact the catalytic efficiency of the enzyme. Studies on several mesophilic enzymes have shown that changes in active site residues that increase stability usually result in reduced enzymatic activity (Yutani et al., 1987; Meiering et al., 1992; Beadle and Shoichet, 2002; Mukaiyama et al., 2006). Although not involved directly in catalysis, position 199 does make direct main-chain interactions to the N6 atom of Ap5A found in the ATP binding pocket of both wild-type and mutant enzymes. Presumably, additional ‘rigidification’ of the core of the protein may have had adverse effects on substrate binding as indicated by an increase in KM for AMP relative to AKsub while increasing the overall stability of the enzyme.
Enzymes adapted to function at elevated temperatures often display higher activation enthalpies (H#) and lower activation entropies (TS#) than enzymes adapted to work at lower temperatures (Lonhienne et al., 2000). These findings provide a thermodynamic explanation for the stronger temperature dependences and lower activities at room temperatures observed for thermophilic enzymes (Low et al., 1973; Fields and Somero, 1998; Lonhienne et al., 2000; D'Amico et al., 2003). Likewise, analysis of the thermodynamic activation parameters for wild-type and mutant AKs has revealed that mutation Q199R significantly increases the activation enthalpy (H#) of the reaction while lowering its activation entropy (TS#), resulting in a stronger temperature dependence and lower activity at room temperatures (Tables I and II). The higher enthalpy of activation often observed for thermophilic enzymes has been ascribed to an increased number of enthalpy-driven interactions that must be broken to reach the transition state during catalysis (Lonhienne et al., 2000). In the case of AKsub Q199R, the increase in the activation enthalpy observed for the mutant enzyme may be further evidence that the ionic interactions facilitated by Arg-199 stabilize the active site of the enzyme.
The in vivo adaptation of B.subtilis AK in a thermophilic genome is reminiscent of colonization of hot environments by mesophilic bacteria. Phylogenetic studies on thermophilic proteins isolated from archaeal- and eubacterial organisms suggest that for the latter, natural selection favored sequence-centered strategies to increase protein stability without introducing major structural changes. Proteins of eubacterial organisms that managed to colonize hot environments display sequence substitutions that favor the formation of a number of restricted interactions, among which ion pairs are the most salient (Berezovsky and Shakhnovich, 2005). Interestingly, the first adaptive step toward increased AKsub stability was the formation of ionic interactions facilitated by mutant residue Arg-199.
The ability to identify and isolate functional intermediates during protein adaptive evolution and directly link the impact of changes in protein structure to reproductive success of an organism (fitness) allows the exploration of the adaptive landscape that constrains protein evolution and the identification of evolutionary strategies behind increased protein stability and function. Taken together, our results have revealed the adaptive strategy employed by mutant AKsub Q199R that allowed the host organism to reach fixation in a large bacterial population under selection. Mutation Q199R led to a stability–activity trade-off in B.subtilis AK. The most likely mechanism behind the trade-off was a ‘rigidification’ of secondary structures surrounding the active site through new ionic interactions facilitated by Arg-199. Mutant Q199R also provided an example of how natural selection can produce some of the trends observed in naturally thermostable enzymes using experimental evolution. Furthermore, ionic interactions are the most salient feature of many thermostable proteins, and in the case of AKsub Q199R leads to features usually observed in thermophilic proteins such as ‘rigidification’ of the protein structure, reduced unfolding rates and stability–activity trade-offs. It is also interesting to note that AKsub Q199R was able to expand the temperature niche of the organism by 9°C which suggests that large benefits at the organism level can be brought up by modest changes in protein structure and function.
|
Funding |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionFundingAcknowledgementReferences |
|---|
National Science Foundation (0212417/0641792); the Welch Foundation (C1584 to Y.S. and C1588 to P.W.S.); W.M. Keck Foundation to the Gulf Coast Consortia through the Keck Center for Computational and Structural Biology (R.C.).
|
Footnotes |
|---|
1 Present address: Department of Biochemistry, University of Otago, New Zealand
2 Present address: Division of Biology, California Institute of Technology, Pasadena, CA 91125, USA
|
Acknowledgement |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionFundingAcknowledgementReferences |
|---|
We thank Drs K.R.MacKenzie, J.S.Olson and an anonymous referee for help and advice.
|
References |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionFundingAcknowledgementReferences |
|---|
Atkinson D.E. Biochemistry (1968) 7:4030–4034.[CrossRef][Web of Science][Medline]
Bae E., Phillips G.N. Jr. J. Biol. Chem. (2004) 279:28202–28208.
Bae E., Phillips G.N. Jr. J. Biol. Chem. (2005) 280:30943–30948.
Beadle B.M., Shoichet B.K. J. Mol. Biol. (2002) 321:285–296.[CrossRef][Web of Science][Medline]
Berezovsky I.N., Shakhnovich E.I. Proc. Natl Acad. Sci. USA (2005) 102:12742–12747.
Berry M.B., Phillips G.N. Jr. Proteins (1998) 32:276–288.[CrossRef][Web of Science][Medline]
Collins K.D., Washabaugh M.W. Q. Rev. Biophys. (1985) 18:323–422.[Medline]
Couñago R., Shamoo Y. Extremophiles (2005) 9:135–144.[CrossRef][Medline]
Couñago R., Chen S., Shamoo Y. Mol. Cell. (2006) 22:441–449.[CrossRef][Web of Science][Medline]
Creighton T.E. Proteins: Structures and Molecular Properties (1993) New York: W. H. Freeman and Company.
D'Amico S., Marx J.C., Gerday C., Feller G. J. Biol. Chem. (2003) 278:7891–7896.
Dams T., Jaenicke R. Biochemistry (1999) 38:9169–9178.[CrossRef][Web of Science][Medline]
DeLano W.L. The PyMOL Molecular Graphics System (2002) San Carlos, CA, USA: DeLano Scientific.
DePristo M.A., Weinreich D.M., Hartl D.L. Nat. Rev. Genet. (2005) 6:678–687.[CrossRef][Web of Science][Medline]
Fields P.A., Somero G.N. Proc. Natl Acad. Sci. USA (1998) 95:11476–11481.
Glaser P., Presecan E., Delepierre M., Surewicz W.K., Mantsch H.H., Barzu O., Gilles A.M. Biochemistry (1992) 31:3038–3043.[CrossRef][Web of Science][Medline]
Jaenicke R. Eur. J. Biochem. (1991) 202:715–728.[Web of Science][Medline]
John D.M., Weeks K.M. Protein Sci. (2000) 9:1416–1419.[Web of Science][Medline]
Kumar S., Nussinov R. J. Mol. Biol. (1999) 293:1241–1255.[CrossRef][Web of Science][Medline]
Lee C.F., Allen M.D., Bycroft M., Wong K.B. J. Mol. Biol. (2005) 348:419–431.[CrossRef][Web of Science][Medline]
Lonhienne T., Gerday C., Feller G. Biochim. Biophys. Acta (2000) 1543:1–10.[CrossRef][Medline]
Low P.S., Bada J.L., Somero G.N. Proc. Natl Acad. Sci. USA (1973) 70:430–432.
Meiering E.M., Serrano L., Fersht A.R. J. Mol. Biol. (1992) 225:585–589.[CrossRef][Web of Science][Medline]
Miyazaki K., Wintrode P.L., Grayling R.A., Rubingh D.N., Arnold F.H. J. Mol. Biol. (2000) 297:1015–1026.[CrossRef][Web of Science][Medline]
Mukaiyama A., Takano K., Haruki M., Morikawa M., Kanaya S. Biochemistry (2004) 43:13859–13866.[CrossRef][Web of Science][Medline]
Mukaiyama A., Haruki M., Ota M., Koga Y., Takano K., Kanaya S. Biochemistry (2006) 45:12673–12679.[CrossRef][Web of Science][Medline]
Muller C.W., Schlauderer G.J., Reinstein J., Schulz G.E. Structure (1996) 4:147–156.[Medline]
Noda L. The enzymes—Boyer P.D., ed. (1973) 8. New York: Academic. 279–305.
Ogasahara K., Nakamura M., Nakura S., Tsunasawa S., Kato I., Yoshimoto T., Yutani K. Biochemistry (1998) 37:17537–17544.[CrossRef][Web of Science][Medline]
Pace C.N., Shaw K.L. Protein Struct. Funct. Genet. (2000) 4:1–7.
Perrier V., Surewicz W.K., Glaser P., Martineau L., Craescu C.T., Fabian H., Mantsch H.H., Barzu O., Gilles A.M. Biochemistry (1994) 33:9960–9967.[CrossRef][Web of Science][Medline]
Ratner V., Amir D., Kahana E., Haas E. J. Mol. Biol. (2005) 352:683–699.[CrossRef][Web of Science][Medline]
Saint Girons I., Gilles A.M., Margarita D., Michelson S., Monnot M., Fermandjian S., Danchin A., Barzu O. J. Biol. Chem. (1987) 262:622–629.
Somero G.N. Annu. Rev. Physiol. (1995) 57:43–68.[CrossRef][Web of Science][Medline]
Sriprapundh D., Vieille C., Zeikus J.G. Protein Eng. (2003) 16:683–690.
Vonrhein C., Schlauderer G.J., Schulz G.E. Structure (1995) 3:483–490.[Medline]
Vriend G. J. Mol. Graph. (1990) 8:52–56, 29.[CrossRef][Web of Science][Medline]
Wittung-Stafshede P. Biochim. Biophys. Acta (2004) 1700:1–4.[Medline]
Wolf-Watz M., Thai V., Henzler-Wildman K., Hadjipavlou G., Eisenmesser E.Z., Kern D. Nat. Struct. Mol. Biol. (2004) 1:945–949.
Wray J.W., Baase W.A., Lindstrom J.D., Weaver L.H., Poteete A.R., Matthews B.W. J. Mol. Biol. (1999) 292:1111–1120.[CrossRef][Web of Science][Medline]
Yutani K., Ogasahara K., Tsujita T., Sugino Y. Proc. Natl Acad. Sci. USA (1987) 84:4441–4444.
Zavodszky P., Kardos J., Petsko G.A. Proc Natl Acad. Sci USA (1998) 95:7406–7411.
Received May 21, 2007; revised November 9, 2007; accepted November 12, 2007.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||