Thermophilic Enzyme or Mesophilic Enzyme with Enhanced

Subscriber access provided by UNIVERSITY OF MICHIGAN LIBRARY

Article

Thermophilic Enzyme or Mesophilic Enzyme with Enhanced Thermostability: Can We Draw a Line? Xiaomin Jing, Wilfredo Evangelista Falcón, Jerome Yves Baudry, and Engin H. Serpersu J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.7b04519 • Publication Date (Web): 08 Jul 2017 Downloaded from http://pubs.acs.org on July 10, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Thermophilic Enzyme or Mesophilic Enzyme with Enhanced Thermostability: Can We Draw a Line? Xiaomin Jing1,^, Wilfredo Evangelista Falcon1,3,^, Jerome Baudry1,3,*, Engin H. Serpersu2,%,* 1: Department of Biochemistry and Cellular and Molecular Biology, 2: Graduate School of Genome Science and Technology, the University of Tennessee and Oak Ridge National Laboratories, Knoxville, Tennessee 37996 3: UT/ORNL Center for Molecular Biophysics, Oak Ridge National Laboratory, Oak Ridge, TN. ^

These authors contributed equally

*

Corresponding authors

%

Current Address, National Science Foundation, 4201 Wilson Blvd. Arlington, VA 22230

Address for correspondence: Engin Serpersu, National Science Foundation, 4201 Wilson Blvd. Arlington, VA 22230 ([email protected]) and Jerome Baudry, BCMB Department, M407 Walters Life Sciences 1414 Cumberland Avenue , Knoxville, TN 37996 ([email protected])

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

Abstract The aminoglycoside nucleotidyltransferase 4' (ANT) is a homodimeric enzyme that modifies the C4'-‐OH site of aminoglycoside antibiotics by nucleotidylation. A few single and double residue mutants of this enzyme (T130K, D80Y and D80Y/T130K) from Bacillus stearothermophilus show increased thermostability. This paper investigates how such residue replacements, which are distant from active site and monomer-‐monomer interface, result in various changes of the thermostability of the enzyme. In this work, we show that thermodynamic properties of enzyme–ligand complexes and protein dynamics may be indicators of thermophilic behavior. Our data suggests that, one of the single site mutants of ANT, D80Y, may be a thermophilic protein and the other thermostable mutant T130K is actually a more heat-‐stable variant of the mesophilic WT with a higher Tm. Our data also suggest that T130K and D80Y adopt different global dynamics strategies to achieve different levels of thermostability enhancement, and that the differences between the properties of the species can be described in terms of global dynamics rather than in terms of specific structural features. Thermophilicity of the D80Y comes at the cost of less favorable thermodynamic parameters for ligand binding relative to WT. On the other hand, the T130K species exhibits the same affinity to ligands and the same thermodynamic parameters of complex formation as the WT enzyme. These observations suggest that a quantitative characterization of ligand binding and protein dynamics can be used to differentiate thermophilic proteins from their simply more heat stable version of mesophilic counterparts.

2


Page 3 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction Enzymes from thermophilic organisms play an important role in biotechnology and industry. Their thermostability and optimal activity at high temperatures makes them more suitable for many processes compared to their mesophilic counterparts. Thermophilic proteins may adopt various strategies to achieve thermal adaptation. Therefore, there has been many studies aimed to understand molecular mechanisms that increase thermal stability of proteins. However, the molecular and thermodynamic basis of thermophilicity of proteins remains largely undetermined. Earlier work suggested that thermophilic variants of enzymes have several differences from mesophilic enzymes, such as increased polar interactions1-‐4 or more Hydrogen bonds or increased hydrophobic interactions due to a better packing of the hydrophobic core5-‐8 or some combination of these properties. In addition, other studies also suggested that distributed effects and dynamic properties of proteins may play significant roles in attaining thermostability 9-‐12 and indeed justification of thermophilicity based solely on amino acid sequences do not allow a clear rationalization of thermophilic behavior13-‐15.

Figure 1. Crystal Structure of D80Y16 with bound ligands. MgATP (purple) and kanamycin A (red) are bound to the active site, which is formed at the interface of monomers. Residues D80 (green) and T130 (yellow) are shown as ball and stick model. Attempts at engineering thermostable proteins, using either evolutionary strategies or structure-‐ based approaches, have worked on a case by case basis. Most of the current efforts appear to be directed towards the improvement of thermostability17-‐19. These types of approaches may yield enzymes with higher thermostability for practical applications; however, as indicated in 20 they may not be thermophilic proteins. Instead, they may be mesophilic proteins with increased melting temperatures (Tm)20-‐22. In this work, we attempted to show that increased Tm alone does not render a protein to be considered as “thermophilic” and there are molecular properties that are specific for thermophilic proteins. By using single mutant variants of the aminoglycoside nucleotidyltransferase(4ʹ′), we show that dynamic properties of apo-‐enzymes and thermodynamics of enzyme-‐ligand interactions are among the parameters that allow discrimination between thermophilic proteins and those simply with increased Tm. In this work, for simplicity, we will use the term “thermostable” to denote the variant with higher Tm but otherwise have identical protein dynamics and thermodynamics of ligand-‐protein interaction to 3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 26

the mesophilic WT enzyme and reserve the term “thermophilic” for the variant that show significantly different properties in these aspects that we attribute to be indicators of thermophilic behavior. The aminoglycoside nucleotidyltransferase (4ʹ′) (ANT) catalyzes the transfer of the AMP group from ATP to the 4'-‐OH site of aminoglycosides (Fig S1), which results in elimination of their effectiveness as antibiotics. The enzyme was originally isolated from S aureus as a mesophilic protein23. Thermostable variant was isolated from Bacillus stearothermophilus24. Of the two thermostable, single mutants of the enzyme, T130K displays dynamic and thermodynamic properties identical to that of the mesophilic WT while, the other mutant, D80Y, has significantly different properties. It is not clear how a single residue replacement, distant from active site and the monomer-‐monomer interface (Figure 1), yield global stabilization of enzyme at elevated selection temperatures. This work describes thermodynamic and computational data leading to differentiation of thermophilic variant from more heat stable mesophilic variant of ANT. Experimental and Computational Methods Chemicals and Reagents High-‐performance Ni-‐Sepharose resin and IPTG were obtained from GE Healthcare (Pittsburg, PA) and Inalco Spa (Milan, Italy) respectively. Ion exchange matrix, Macro Q, was purchased from Bio-‐Rad Laboratories (Hercules, CA). Dr. E. Fernandez of the University of Tennessee generously provided the purified thrombin. Aminoglycosides and all other chemicals were purchased at the highest purity available from Sigma, Aldrich (St. Louis, MO). As a standard procedure in this laboratory, aminoglycosides were purchased as sulfate salts and desulfated by ion exchange chromatography for use as base. Site-‐directed mutagenesis T130K variant cloned in a pET15b vector was used as the template in PCR amplification to create wild type and double variant clones. For D80Y variant, wild type clone was used as the template in PCR reaction. All the mutagenesis was performed using the Quikchange Lightning site-‐directed mutagenesis kit (Stratagene, La Jolla, CA). PCR products were transformed into competent XL10 cells. pET15b plasmids with wild type/each variant were isolated by Miniprep (Qiagen, Germantown, MD) and confirmed by sequencing check (Molecular Biology Resources Facility, University of Tennessee, Knoxville). Positive clone was transformed into E.coli BL21 (DE3) Gold competent cells. Positive clones were confirmed by sequencing check. Purity of wild type enzyme and variants (>95%) were visualized by 10% SDS-‐PAGE and further validated by analytical ultracentrifugation tests. Overexpression and purification Overexpression and purification of wild type enzyme, D80Y and double variants are performed similarly with T130K variant as previously described25, except cell lysis was performed by three cycles of French press. Enzymes remained active for at least a month at -‐20°C. Protein concentrations were determined by absorbance at 280nm using extinction coefficient of 49390 M-‐1cm-‐1 for wild type enzyme and T130K and 50880 M-‐1cm-‐1 for all other variants. Concentrations of proteins were reported as monomer. Analytical ultracentrifugation 4


Page 5 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Beckman XL-‐1 analytical ultracentrifuge was used to perform sedimentation velocity experiments. 400 μL samples were loaded into double-‐sector cells in an An-‐50Ti rotor. Sample temperature was equilibrated at 25°C for an hour before the run. To achieve linearity protein concentrations were detected at 230 nm for low concentrations (1-‐3 μM) while measurements at 280 nm was used for concentrations 5-‐20 μM. For enzyme at concentration higher than 40 µM, the interference optical system was used for detection. All the scans were collected at a rotor speed of 50,000 rpm at 25°C. SEDFIT (version 12.44)26 was used to fit sedimentation data by using continuous (c(s)) distribution model. Protein partial specific volume, buffer density and buffer viscosity were calculated using SEDNTERP27. The enzymes were dialyzed extensively in a buffer solution of 50 mM PIPES pH 7.5 containing 0 or 100 mM NaCl and used in experiments. All ligands were also dissolved in the same dialysis buffer. The weight-‐average sedimentation coefficients (sw(S)) were determined at each concentration by integrating the peaks from the c(s) distributions using SEDFIT26. The monomer-‐dimer self-‐association model in SEDPHAT28 was used to determine dissociation constants for the dissociation of dimers into monomers by using sw(S) as a function of protein concentration. Isothermal titration calorimetry Titrations of substrate binding were performed as previously described29. Titrations of dimer dissociation were performed similarly, except each titration consisted of 4 injections of 65 μL and were separated by 360 seconds. Apo-‐enzyme with concentration above dissociation constant was titrated into buffer. As a result, a certain fraction of dimeric enzyme dissociated into monomeric form, accompanied with the observed heat (ΔH0). Known dissociation constants of dimers for each variant, determined by AUC, allowed calculation of fractions of dimer and monomer in both syringe and cuvette, which, in turn, allowed calculation of the heat of dimer dissociation. Circular Dichroism CD was performed on a Jasco J-‐815 spectrometer using a cuvette with a path length of 2 mm. Spectra were recorded several times with a rates of 1 and 1.5°C/min increments in temperature while monitoring the changes at 222 nm. Protein concentration was 18uM (monomer) for all the runs. Molecular Dynamic simulations Systems were constructed based on the crystal structure of D80Y species (Protein Data Bank ID: 1KNY) using Molecular Operating Environment (MOE, version 2012, Chemical Computing Group, Ltd, Montréal, Canada). The co-‐crystalized ligand and cofactor were removed from the model such that the apo form of the wild type, and T130K species were built. Each structure was explicitly solvated in a TIP3P water in a cubic box of 8 nm x 8 nm x 8 nm. Periodic Boundary Conditions in all directions were applied with electrostatic type Fast smooth Particle Mesh Ewald, PME. A 18,000 steps energy minimization was performed using the steepest decent algorithm and maximum force equal to 10 KJ mol-‐1 nm-‐1. Molecular Dynamics simulations were carried out using the Gromacs 4.6.130-‐31 simulation engine and the AMBER-‐f99sb32 force field. For each species and temperature studied here, a 50 ns equilibration with a 2fs integration timestep was performed in the NPT thermodynamic ensemble, and 1 bar pressure using the Nosé-‐Hoover and Berendsen weak coupling. Finally, a 100 nanoseconds production was run using the Panirello-‐Rahman algorithm. Atomic coordinates were saved on disk every 5 ps. 5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 26

The trajectories obtained from the production time were analyzed using Principal Component Analysis from Gromacs tools, eigenvectors and eigenvalues were calculated to identify the main dynamic modes for the three species, WT, T130K, and D80Y at three different temperatures 300K, 322K, and 330K. Results A couple of single mutations on ANT increased the thermostability of the enzyme without affecting the catalytic activity of the enzyme where each mutation provided a different degree of thermostability 33 . Thermal denaturation of wild type and thermostable variants, determined by CD, confirmed these observations (Figure S2). The melting temperatures increase in the order of wild type (40.9 ± 0.5°C), T130K (49.1±0.6°C), D80Y (56.2±0.2°C) and D80Y/T130K (62.6±0.1°C). While D80Y displays a higher thermostability relative to T130K, the double variant T130K/D80Y shows an additive effect of individual mutations on Tm. The additivity of the increments in Tm values is consistent with two different mechanisms providing an increase in thermostability to the double variant. Thermal denaturation of all four enzymes was irreversible and aggregation was observed post denaturation. However, identical denaturation profiles were obtained with each variant at different heating rates, which indicates that thermal denaturation of enzymes were not perturbed by kinetics of aggregation34. Subunit-‐subunit interaction is affected by single amino acid replacements ANT is a homodimeric enzyme with ligand binding sites at the interface of monomers. Even though the mutation sites are away from the subunit-‐subunit interface, they affected the dimer formation as detected by analytical ultracentrifugation (AUC). As shown in Figure 2, monomer-‐dimer equilibrium was concentration dependent with all variants and there was a progressive increase in tendency for dimerization with increasing thermostability. Dissociation constants (Kd) of dimers for all enzymes were determined as previously described29. The curves shown in Figure S3 represent the best fit to the data points for all the tested enzymes. The dissociation constants for homodimers of wild type, T130K, D80Y and D80Y/T130K were estimated to be within the range of 28-‐32 μM, 1.6-‐1.8 μM, 0.8 -‐ 0.9 μM and 0.40 -‐ 0.5 μM respectively. In all cases, Kd increased steadily from the most thermophilic to mesophilic variants reaching to a ~70-‐fold difference between the mesophilic WT and the most thermostable variant D80Y/T130K. Our previous study on T130K showed that increasing salt concentration favored the formation of dimer indicating that hydrophobic interactions are the major cause of dimerization29. Similar sedimentation velocity experiments were performed for wild type, D80Y and D80Y/T130K variants in the absence of added salt. On the average, there was an almost an order of magnitude increase in Kd with all enzymes indicating that hydrophobic interactions are the main contributors of dimer formation with all variants. Dimer dissociation, detected by isothermal titration calorimetry (ITC), showed an endothermic behavior with all four enzymes in the presence and absence of 100 mM NaCl. The heat of dissociation varied within 3.1 -‐ 5.5 kcal/mol for all variants and didn’t appear to follow the order of thermostability. We note that these determinations were based on dilution of the enzyme from a concentrated stock to the ITC chamber and errors were larger for D80Y and D80Y/T130K due to low Kd for dimer dissociation. 6


Page 7 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Antibiotic selection separates two thermostable variants carrying a single mutation Thermodynamic properties of enzyme–ligand complexes with all variants were determined by ITC. Tobramycin and neomycin were used as representatives of kanamycins and neomycins respectively (Figure S1). The formation of the aminoglycoside-‐enzyme complexes are typically results in changes in pKas of functional groups in the enzyme and in the ligand causing a net proton uptake or release35-‐38. Therefore, buffer solutions with different heats of ionization (ΔHion) were used in titrations Figure 2. Size distribution of ANT and its variants determined by sedimentation velocity experiments.

For all plots, red, blue and green lines represent 5, 10 and 20 μM protein respectively. Panels are WT, T130K, D80Y and T130K/D80Y from top to bottom respectively. to determine the intrinsic binding enthalpies (∆Hint) and the change in net protonation (∆n) for wild type and thermostable variant of ANT as described in detail earlier36. Figure 3 shows an example of such data as a plot of observed enthalpy (ΔHobs) vs. ΔHion where the intercept on y-‐axis is ΔHint and ∆n is obtained from the slope of the line. Experiments performed with T130K yielded identical results to those published earlier29. As it can be seen in Figure 3, slopes of lines determining the value of Δn display a clear separation between the pairs of D80Y/double mutant and WT/T130K when neomycin is the ligand.

7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 26

Figure 3. Effect of the heat of ionization on the observed enthalpy of binding. The binding enthalpy of neomycin to WT (○), T130K (●), D80Y (▫) and D80Y/T130K (◊) is plotted against the heat of ionization of buffers (PIPES, HEPES and Tris with 2.7, 4.9 and 11.4 kcal/mol of heat of ionization respectively). The dissociation constants for enzyme–neomycin complexes showed remarkable differences; while the WT and T130K have nanomolar affinity, binding of neomycin to D80Y and D80Y/T130K was weaker by almost three orders of magnitude. Large differences were also visible in the other thermodynamic properties of enzyme–ligand complexes (Table 1). Binding of neomycin to WT and T130K was highly exothermic with similar ΔHint values but the binding enthalpy of neomycin to D80Y and D80Y/T130K was much less exothermic with less unfavorable entropy relative to WT/T130K pair. Overall, WT and T130K showed highly similar thermodynamic properties of complex formation with neomycin, which were significantly different than D80Y and D80Y/T130K. Table 1. Thermodynamic parameters for the formation of ANT–Neomycin complexesa

Wild Type

T130K

D80Y

T130K/D80Y

ΔHint (kcal/mol)

-30.0 ± 0.8

-32.8 ± 0.6

-17.2 ± 0.6

-15.6 ± 0.4

990 ± 200

990 ± 150 -8.4 ± 0.2

KD (nM)

40 ± 6

24 ± 6

ΔG (kcal/mol)

-10.1 ± 0.1

-10.4 ± 0.1

-8.3 ± 0.3

TΔS (kcal/mol)

-19.9

-22.4

-8.9

-7.2

Δn

1.5 ± 0.1

1.95 ± 0.1

1.0 ± 0.1

0.95 ± 0.1

a

Standard error of the mean, based on three trials, are shown. Least square fits from ΔHobs vs ΔHion plots were used to determine errors in the intrinsic enthalpy and in net protonation. 8


Page 9 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2. Thermodynamic parameters for the formation of ANT–Tobramycin complexesa

Wild Type

T130K

D80Y

ΔHint (kcal/mol)

-19.5 ± 0.9

-19.3 ± 1.4

-13.9 ± 2.1

KD (µM)

2.5 ± 0.4

1.1 ± 0.2

1.7 ± 0.6

2.0 ± 0.7

ΔG (kcal/mol)

-7.7 ± 0.1

-8.1 ± 0.1

-8.0 ± 0.2

-7.8 ± 0.2

TΔS (kcal/mol)

-11.8

-11.2

1.1 ± 0.1

1.4 ± 0.2

Δn

T130K/D80Y -13.3 ± 0.8

-5.9 0.8 ± 0.3

-5.5 0.9 ± 0.10

a

Standard error of the mean, based on three trials, are shown. Least square fits from ΔHobs vs ΔHion plots were used to determine errors in the intrinsic enthalpy and in net protonation. Contrary to the binding of neomycin, the binding affinity of tobramycin to all variants was similar (Table 2). However, just like in the complex formation with neomycin, the enthalpy of formation of WT-‐ and T130K–tobramycin complexes were more exothermic than D80Y and D80Y/T130K variants without a clear separation in Δn values. Furthermore, the change in heat capacity (ΔCp), as determined from the slopes of lines of temperature dependence of the binding enthalpy, displayed much smaller range of variation with D80Y as compared to those observed with the WT and T130K when different aminoglycosides are used (Figure 4).

Figure 4. Temperature dependence of binding enthalpy. Data for each enzyme are shown with 4 aminoglycosides representing the widest range of binding affinity to each variant. Top panel, WT: neomycin (●), tobramycin (▪), kanamycin B (Δ) and ribostamycin (○); Middle panel, T130K: neomycin (●), paromomycin (◊), tobramycin (▪) and kanamycin A (▫); Lower panel, D80Y neomycin (●), kanamycin A (▫) kanamycin B (Δ) and ribostamycin (○). Scale of y-‐axis is matched to aid in visual comparison. Distinct dynamic properties of T130K and D80Y. The effect of mutations on the dynamics properties of WT, T130K, and D80Y was characterized from the 100 ns MD trajectories at three different temperatures, 300K, 322K, and 330K (melting temperature of D80Y). The trajectories were analyzed via Principal Component Analysis (PCA) to identify the main 9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 26

motions of the protein species at each temperature39, as previously shown by us 40. Analysis of 300K data suggests that the global dynamics of D80Y is dominated by the first dynamic mode (Figure 5-‐right panel), whereas at 322K and 330K the first two modes contribute both significantly to the global protein dynamics. The dynamics of the WT and T130K species are mainly due to the contribution of the first two modes at all temperatures studied here as shown in Figure 5 (left and middle panels), although the first mode of motion contributes more at 330K. The orientation and amplitude of the first two modes for WT, T130K, and D80Y at 300K are shown by vectors in Figure 6. The origin of these vectors indicate the region of the protein undergoing motion in those particular vectors’ direction. The left panels of these figures show the motion due to the first mode of the protein. The first mode corresponds to the same dominant movement in the three species: an open/close “breathing” motion. The difference between the dynamics of the three species at different temperatures originates from the second dynamics mode, right panels of Figure 6, which becomes significantly contributing in the D80Y variant only at 322K and 330K. These results suggest that the first mode of motion has the same direction at all temperatures for all the three species (left panels of Figures S4 and S5). As it occurs at 300K, the difference seems to arise from the second mode, whose vectors have different origins and directions depending on the species (right panels of Figures S4 and S5).

Figure 5. Principal modes of motion of ANT projected onto the first 25 eigenvalues calculated from PCA in MD trajectories at different temperatures; Left to right are WT, T130K, and D80Y, respectively. Filled circles (blue), squares (green) and diamonds (red) represent 300K, 322K and 330K respectively. Insets show expanded regions of the several initial modes of motion. Discussion D80Y and T130K adopt different structural strategies in enhancing enzyme thermostability Thermostability of proteins can be increased by residue replacements to introduce specific stabilizing interactions, which may produce a thermostable protein. However, such studies may generate simply a variant of mesophilic protein with higher Tm rather than being a thermophilic variant. In this work, we determined dynamic and thermodynamic properties of two single mutants of ANT, both of which have increased thermostability however only one of them, D80Y, appeared to be thermophilic enzyme. Data shown in this work revealed significant clues on the identification of molecular properties that separate thermophilic protein from a thermostable mesophilic counterpart. In this case, T130K shows a 7°C increased Tm relative to WT, however, its dynamic properties and thermodynamics of enzyme–ligand complexes are identical to the mesophilic WT. D80Y variant, on the other hand, has 14°C higher Tm and has different protein dynamics and thermodynamics parameters of enzyme–ligand interactions. 10


Page 11 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Moreover, findings from MD simulations point out to a temperature-‐ and sequence of mode-‐

dependence of the ANT’s protein dynamics. The three protein species characterized here all exhibit the Figure 6. The arrows, obtained from PCA, represent the first (left panel) and second mode (right panel) of motions respectively. Top to bottom are: WT, T130K, and D80Y. C-‐α atoms of Asp80 and Thr130 are represented by orange and purple spheres, respectively. same first mode of motion at the three temperatures tested here. However, the second mode of motion is different in various species and temperatures. The WT and T130K species exhibit very similar magnitude of contributions from the second mode, but the regions affected by this second mode and its directions are different between the two species, as shown in Figures 6 and 5S. On the other hand, D80Y exhibits different relative contributions from the first and second mode with respect to the other two species at 300K (Figure 5). At 330K, this variant shows a significantly different second dynamics mode, both in magnitude and direction, when compared to that of WT or T130K (Figure S5). These differences suggest that mesophilic/thermophilic properties can be characterized by global dynamic properties, rather than just by structurally-‐localized differences between species. 11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 26

Crystal structures of WT and T130K are not available, however, the structure of D80Y is available16. CD spectra (210-‐280 nm) suggests that structures of all variants are similar and, furthermore, the effect of salt on homodimer formation also indicates that similar interactions are driving the dimer formation in all mutants. Therefore, all interpretations, made based on the available structure of D80Y, which we believe, are reasonably justified. Both residues T130 and D80 are exposed to solvent, are away from the monomer-‐monomer interface and the active site. Yet, the mutation of these residues increases Tm. The crystal structure of D80Y shows that the 130th residue is positioned at the beginning of the fifth major alpha-‐helical region. E127 and D133 are within 4.5Å of T130. Thus, the replacement of T with K may create strong polar interactions with either E127 or E133. Positions of these residue are shown in a close up of the structure in Figure 7. Thus, it is likely that the new interactions raise the melting temperature of the mutant by affecting local interactions and structure without affecting the overall dynamic properties of the enzyme and thermodynamics of enzyme-‐ligand interactions.

Figure 7. Close-‐up of T130 and its surrounding residues. T130 (blue), E127 and D133 (purple) are shown in ball and stick manner with several other neighboring residues. Kanamycin A is shown in white.

Figure 8. Close-‐up of the position of residue 80. The surroundings of position 80 is shown in the crystal structure of D80Y16 . D80Y is shown in yellow. Kanamycin A is shown in white. There are no positively charged residues within 4.5Å of D80. One of the neighboring residues of D80, N78, forms hydrogen bonds with E76 and E67, which are the key residues involved in electrostatic 12


Page 13 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


interactions with antibiotics. (Figure 8) In addition, D80 toY80 replacement may also benefit from the internal hydrophobic packing of enzyme by the introduction of an aromatic residue at this site. Thus, it appears that D80 acts a node in the connectivity network of residues of ANT and have significant role in determining protein dynamics. This, in turn, affects alters thermodynamics of ligand binding and makes it thermodynamically less favorable. The behavior of the double variant D80Y/T130K suggests that D80Y mutation has the dominant effect while the effect of mutation at T130 is limited to increasing Tm confirming that thermostability achieved by different mechanisms by these variants. We should also mention that thermophilic behavior of D80Y is contrary to current paradigm, which suggests that thermophilic proteins contain more polar residues relative to their mesophilic counterparts. Replacement of the ionic side chain of aspartate with phenolic side chain of tyrosine in D80Y contrasts this notion and suggests that alteration of dynamic properties of the protein may be more important in achieving thermophilicity. This observation, again, suggests that comparisons based on protein sequences alone need to be interpreted cautiously to differentiate molecular properties of thermophilic proteins from mesophilic ones. Various structural and dynamic features of thermophilic and mesophilic enzymes were compared in numerous studies. These efforts, however, may be complicated by the fact that they included all thermostable enzymes, some of which may simply converted to thermostable versions of mesophilic enzymes by the introduction of stabilizing interactions such as disulfide bonds or salt bridges. Furthermore, these comparisons included proteins that have differences ranging from a few residues to tens of residues in their primary sequences. Some of these variations may introduce changes that are not necessarily specific for thermophilicity. In this work, such differences are minimized to a single residue replacement between the variants. Ligand binding to the thermophilic variants are enthalpically less favored relative to WT and T130K. However, relatively less unfavorable entropic contribution renders ligand binding still thermodynamically favorable (ΔG

Thermophilic Enzyme or Mesophilic Enzyme with Enhanced

Recommend Documents