Molecular Gating of an Engineered Enzyme Captured in Real Time

Subscriber access provided by YORK UNIV

Article

Molecular Gating of an Engineered Enzyme Captured in Real Time Piia Kokkonen, Jan Sykora, Zbynek Prokop, Avisek Ghose, David Bednar, Mariana Amaro, Koen Beerens, Sarka Bidmanova, Michaela Slanska, Jan Brezovsky, Jiri Damborsky, and Martin Hof J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b09848 • Publication Date (Web): 03 Dec 2018 Downloaded from http://pubs.acs.org on December 3, 2018

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 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 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.

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 31 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

Journal of the American Chemical Society

TOC 84x45mm (300 x 300 DPI)

ACS Paragon Plus Environment

Journal of the American Chemical Society 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 31

Molecular Gating of an Engineered Enzyme Captured in Real Time Piia Kokkonen#1,2, Jan Sykora#3, Zbynek Prokop#1,2 Avisek Ghose3, David Bednar1,2, Mariana Amaro3, Koen Beerens1, Sarka Bidmanova1,2, Michaela Slanska1, Jan Brezovsky1, Jiri Damborsky1,2*, Martin Hof3* Loschmidt Laboratories, Department of Experimental Biology and Research Centre for Toxic Compounds in the Environment RECETOX, Faculty of Science, Masaryk University, Kamenice 5/A13, 625 00 Brno, Czech Republic 1

International Clinical Research Center, St. Anne's University Hospital Brno, Pekarska 53, 656 91 Brno, Czech Republic 2

J. Heyrovsky Institute of Physical Chemistry of the ASCR, v. v. i., Dolejskova 3, 182 23 Prague 8, Czech Republic 3

# Equal contribution *To whom correspondence should be addressed. E-mail: [email protected], [email protected]

Current affiliation AG: Center for Nanobiology and Structural Biology of the Institute of Microbiology ASCR, v.v.i., Zámek 136, Nove Hrady 373 33, Czech Republic; JB: Laboratory of Biomolecular Interactions and Transport, Institute of Molecular Biology and Biotechnology, Faculty of Biology, Adam Mickiewicz University, Umultowska 89, 61-614 Poznań, Poland

ABSTRACT Enzyme engineering tends to focus on the design of active sites for the chemical steps, while the physical steps of the catalytic cycle are often overlooked. Tight binding of a substrate in an active site is beneficial for the chemical steps, whereas good accessibility benefits substrate binding and product release. Many enzymes control the accessibility of their active sites by molecular gates. Here we analyzed the dynamics of a molecular gate artificially introduced into an access tunnel of the most efficient haloalkane dehalogenase using pre-steady-state kinetics, a single-molecule fluorescence spectroscopy and molecular dynamics. Photoinduced electron-transfer – fluorescence correlation spectroscopy (PET-FCS) has enabled real-time observation of molecular gating at single molecule level with the rate constants (kon = 1822 s-1, koff = 60 s-1) corresponding well with those from the pre-steady-state kinetics (k-1 = 1100 s-1, k1 = 20 s-1). PET-FCS technique is used here to study the conformational dynamics in a soluble enzyme, thus demonstrating an additional application for this method. Engineering dynamical molecular gates represents a widely applicable strategy for designing efficient biocatalysts.

1 ACS Paragon Plus Environment

Page 3 of 31 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 are highly efficient biocatalysts with outstanding specificity and selectivity. Enzymatic reactions take place in their active sites which are frequently hidden inside the protein core. The active sites are connected to the bulk solvent through the transport tunnels. The importance of tunnels in enzymatic activity, substrate specificity, and stereoselectivity has been explored in a number of studies using X-ray crystallography, steady-state and pre-steady-state kinetics and molecular modeling.1,2 Beyond complementarity between the substrate and binding site, tunnels offer another level of discrimination for molecules going in and out of the enzyme active cavity. Tunnels discriminate the preferred substrates or cofactors, reduce the access of solvents that may disturb the catalysis, and prevent the escape of reactive intermediates before the reaction is finished. The orchestration of enzymatic reactions commonly requires a precise order of several individual steps and the dynamic character of transport tunnels can play a key role in the catalytic efficiency. Many enzymes employ molecular gates to control the access of small molecules into their catalytic sites, in a similar fashion as ion channels control the access of ions into the cell. These gates can adopt open and closed states and thus grant or deny the access of substrates, ions and water molecules into the catalytic site, reflecting the requirements of the catalytic cycle.2 The low efficiency of de novo designed biocatalysts relative to those that have evolved naturally demonstrates that our understanding of enzymatic catalysis is still incomplete.3 Especially, the dynamic processes that control the function of enzymes are only recently being taken into account in the enzyme design protocols.4,5 Great progress has been achieved in computational protein design of the substrate/intermediate complementarity with the active site.6,7 Considerably less attention has been paid to the role of ligand transport and the dynamics of the access pathways, even though many times the ratelimiting step of enzymatic catalysis is not the chemical one. In the past few years, the design of access pathways has gained popularity and as it is offering a complementary strategy to the design of the catalytic



Page 4 of 31

sites.1–2,8–11 Rational designing of the tunnels and dynamical dates in enzymes is still challenging, however, because their motions are currently insufficiently understood. Previously we have described the most active haloalkane dehalogenase LinB86, carrying four mutations Trp140Ala+Phe143Leu+Leu177Trp+Ile211Leu.12,13 The variant was constructed from one of the less active variants, LinB32, which had the main access tunnel blocked by an introduced tryptophan residue Leu177Trp.12 Even though the kinetics of LinB32 indicated improvement in a chemical step, the poor product transport substantially hindered its catalytic activity.13 The mutation of three additional residues (Trp140Ala+Phe143Leu+Ile211Leu) opening a novel auxilliary tunnel lead to LinB86 which has the highest catalytic efficiency with the substrate 1,2-dibromoethane ever reported for this enzyme family. The newly opened de novo tunnel is used for ligand transportation to a lower extent than the natural one, but the decreased intramolecular contacts enabled the helix surrounding the main tunnel to move in a gate-like fashion. To verify the enhanced dynamics of this region and to elucidate the mechanism of the superior catalytic performance of LinB86, we combined the data from pre-steady-state kinetics, single-molecule fluorescence spectroscopy and molecular modelling. The kinetic analyses identified the effects on the individual catalytic steps of 1,2-dibromoethane conversion by all three enzymes: wild-type LinBwt, LinB32 and LinB86.A The detailed kinetics of enzymeproduct interactions predicted a significant role of isomerization dynamics involved in the superior catalytic behavior of LinB86. Subsequently, photoinduced electron transfer - fluorescence correlation spectroscopy (PET-FCS) and molecular simulations provided real-time observation and a structural view of the gating dynamics. In this study, the PET-FCS was used to monitor the conformational dynamics of a soluble enzyme; previously the methodology has been utilized to study transmembrane proteins or protein folding.14,15 In the broader context of protein science, this study demonstrates how PET-FCS technique can In our previous work, LinB32 was referred as LinB-ClosedW and LinB86 as LinB-OpenW.13 The naming scheme has been changed due to the new observations reported in this study which show that LinB86 adopts both open and closed states and thus referring to it as LinB-OpenW would be confusing for the readers. 3

A


Page 5 of 31 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


be used to better understand the impact of structural dynamics on function or selectivity of a protein. The results led to a detailed insight into the molecular gating, which is instrumental for future de novo design of biocatalysts with balanced chemical and physical steps of the catalytic cycle. RESULTS Pre-steady-state kinetics. The kinetic model for the reaction of LinBwt, LinB32 and LinB86 with 1,2-dibromoethane was globally fitted to pre-steady-state kinetic data including rapid quench burst, multiple turnover stopped-flow fluorescence analyses and steady-state initial rate kinetic measurements (Supplementary Information I). The analysis provided individual rate and equilibrium constants for the minimal kinetic pathway including: (i) substrate binding, (ii) cleavage of carbon-halogen (SN2 reaction) bond leading to the formation of an alkyl-enzyme intermediate, (iii) hydrolysis of the alkyl-enzyme intermediate (and reaction) and (iv) product release (Figure 1A). In comparison to LinBwt, LinB32 displays improvement of the first (SN2) chemical step, but decreased affinity for substrate and a significant slowdown of product release. The LinB86 variant reconstituted the substrate binding equilibrium to the wild-type level and simultaneously retained the improvement of the first chemical step introduced in LinB32 by the Leu177Trp mutation. Strikingly, the LinB86 acquired additional efficiency for the second (AdN) chemical step and extensively accelerated the transport of the product out of the active site (Supplementary Figures 1-3). The four mutations introduced to LinB86 provided effective balance among all steps involved in the catalytic cycle with 1,2-dibromoethane. The product release remains the rate-limiting step for all three variants (Supplementary Table 1).



Page 6 of 31

Figure 1: Pre-steady-state kinetics. (A) The equilibrium and rate constants for individual catalytic steps for the reaction of LinBwt (black), LinB32 (red) and LinB86 (blue) with 1,2-dibromoethane at 37oC and pH 8.6. The scheme depicts the minimal kinetic model, where S stands for a substrate, P for a product, E for an enzyme, ES for a non-covalent enzyme-substrate complex, EI for a covalently bound alkyl-enzyme intermediate and EP for an enzyme-product complex. KA is equilibrium association constant for the ES complex formation, KD is equilibrium dissociation constant (KD = 1/KA), k2 is rate constant for cleavage of carbon-halogen bond (SN2), k3 is rate constant for hydrolysis of the alkyl-enzyme intermediate (AdN) and k4 is rate constant for product release. The catalytic mechanism is adopted from Verschueren et al., 1993.16 (B) Schematic illustration of mutations and conformational states of LinB variants (top) and kinetics of close-open enzyme isomerization (bottom). The bulky “W” introduced to the main access tunnel of LinB32 (red) stabilizes the closed state. For LinB86 (blue), the mutations depicted as “LAL” opened de novo auxilliary tunnel that primarily does not maintain ligand transport, but greatly enhances the gating dynamics. The dependence of observed rates (kobs) on ligand concentration recorded upon binding and dissociation experiments of bromide ion with LinB32 (red) and LinB86 (blue) is plotted in the graph. The rate of enzyme close-open isomerization representing the steps k1 and k-1 were derived from the global fit of binding and dissociation data to the kinetic model described in the Supplementary Scheme 2.

The kinetics of bromide binding and dissociation provided information about the dynamical behavior of individual mutants. The binding kinetics of LinB32 and LinB86 show a major difference in comparison to the wild-type enzyme: both variants show complex kinetic behavior (Supplementary information II), while a single step binding of bromide reaching rapid equilibrium has been described for LinBwt.17 The decrease in observed rates (kobs) with increasing ligand concentration determined for LinB32 5 ACS Paragon Plus Environment

Page 7 of 31 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


and LinB86 (Figure 1B) is indicative of the presence of a slow conformational step preceding the binding. The kinetic model suggests that LinB32 and LinB86 exist in solution as a mixture of two conformers.18 Interestingly, LinB86 is a highly flexible protein and undergoes the open-close transition more than two orders of magnitude faster than the rather rigid LinB32 (Figure 1B, Supplementary Figure 4). For both proteins, the rate of forward isomerization is significantly slower than the rate for the reverse step, favoring the closed conformation of the free enzyme. The equilibrium ratio between the two conformational states changes in the presence of halide ions. Bromide stabilizes the open conformation, which represents approximately 20% and 50% of the total enzyme population for LinB32 and LinB86, respectively. Interestingly, the apparent dissociation constants (290 ± 20 mM for LinBwt, 210 ± 20 mM for LinB32 and 180 ± 60 mM for LinB86) determined from the steady-state fluorescence show that only the kinetics, not the overall affinity, has been altered by modification of the access tunnels (Supplementary Figure 5). To test the hypothesis that Leu177Trp mutation is important for the SN2 chemical step of the catalytic cycle, we conducted a back-mutation experiment of LinB86+Trp177Leu (Supplementary information section IV). The steady-state turnover obtained for LinB86+Trp177Leu (kcat = 3.8 ± 0.1 s-1) is significantly decreased in comparison to the turnover number observed for LinB86 (kcat = 57 ± 3 s-1) and even LinBwt (kcat = 12 ± 4 s-1).13 Thus, the Leu177Trp mutation blocking main tunnel is important for the catalytic efficiency, which is further improved by the mutations in the engineered auxilliary tunnel that lead to enhanced dynamics in LinB86. A single-molecule PET-FCS. We have applied the concept of PET–FCS14,19–21 to observe the openclose gating dynamics in LinB86. FCS is a single-molecule fluorescence fluctuation spectroscopy method that provides information on protein dynamics on nanosecond to millisecond time-scales (Figure 2A). The photochemical process of PET is used as an “on and off” switch for the fluorescence signal of a fluorophore, thus creating the fluorescence fluctuations detected by FCS. The switch is activated by the structural 6 ACS Paragon Plus Environment


Page 8 of 31

changes provoked by the dynamical gate, and consequently the FCS data reflects the rate of the gating process (Figure 2BC). The prerequisite for a successful PET-FCS study is proper design of the “on and off” switch. The chosen Atto-oxazine fluorophores Atto655 and AttoOxa12 can be quenched, i.e., turned off, via PET from tryptophan.22,23 This enabled us to study the dynamic gating of LinB86 in relation the introduced Trp177. A computational study has been performed to design and construct the best system for the single molecule PET-FCS. The modelling identified suitable sites along the LinB86 sequence for the cysteine replacement, which can be subsequently labelled with thiol reactive Atto variants (Supplementary Information III and IV). The following parameters were evaluated for the mutable residues: (i) solvent accessibility, (ii) backbone flexibility and (iii) stability. According to these parameters, four LinB86 mutants were constructed, purified and experimentally characterized (Asp149Cys, Gln172Cys, Leu179Cys and Ala269Cys). The protein stability tests identified LinB92 variant (Gln172Cys) as the most suitable candidate for the PET-FCS study. An additional variant of LinB92 with back-mutated Trp177 and denoted as LinB99 (Gln172Cys+Trp177Leu) was prepared for control experiments to exclude non-specific contributions to the PET-FCS data. Two thiol reactive dyes Atto655 (Atto 655-Maleimide) and AttoOxa12 (Atto Oxa12-Maleimide) were selected for labelling of both LinB92 and LinB99 (Supplementary Table 4).


Page 9 of 31 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


Figure 2: Single molecule spectroscopy. (A) Scheme depicting the principle of single molecule FCS experiment. The fluorescence of freely diffusing molecule (orange) is monitored during its transit through the observation volume (in green). (B) Schematic illustration of the dynamic behavior of LinB92-Atto655, i.e., LinB86 labelled with Atto dye at the position 172, which can adopt open and closed conformations. When labelled with Atto655 or AttoOxa12, the closed conformation of LinB92-Atto655 maintains the fluorescence dye away from Trp177 (depicted as W) preventing the PET. Thus, the closed state of LinB92Atto655 is fluorescent. In contrast, the open state of LinB92-Atto655 enables the Atto dye to get into close proximity of Trp177, resulting in PET and quenching of the dye. The open state is therefore non-fluorescent. The respective rate constants determined by the single molecule experiments are depicted. (C) Simplified cartoon of time trace of the fluorescence intensity fluctuations recorded in FCS experiment (raw data trace is depicted as Supplementary Figure 8 in the Supporting information). Transition to the open state leads to a reduction in the fluorescence intensity (in green), while the opposite transition recovers the fluorescence intensity (in red). These fluctuations caused by the conformational transition are projected in the recorded autocorrelation function G(τ), which enables the determination of the rate constants of the transitions between the closed and open states. (D) FCS autocorrelation curves G(τ) obtained for LinB92-Atto655 (black) and LinB99-Atto655 (red). LinB99 carries out back mutation (Trp177Leu) and serves as the control sample to exclude non-specific contributions to the PET-FCS data. The inset in linear scale highlights the 8 ACS Paragon Plus Environment


Page 10 of 31

observed difference in the PET component between LinB92-Atto655 and LinB99-Atto655 on the (sub)millisecond time scale. (E) PET relaxation time constant τ1 obtained for LinB92-Atto655 (black) and LinB99-Atto655 (red) at different bromide concentrations.

The FCS autocorrelation functions G(τ) obtained for the enzymes labelled with Atto655 (Figure 2D) show a complex behavior with a dominant component on the millisecond time scale that is attributed to the diffusion of the protein. For LinB92-Atto655 an additional component, τ1, is detected on the (sub)millisecond time scale, while the control variant LinB99-Atto655 lacks this contribution to G(τ) (Figures 2D and 2E, Supplementary Table 5). Therefore, it is rational to causally link the τ1 component in LinB92Atto655 to the transition between the fluorescent and the non-fluorescent states (induced by PET due to the conformational change). Single-molecule PET-FCS experiments carried out in solutions of the labelled LinB92-Atto655 and LinB99-Atto655 at various bromide concentrations (Figure 2E, Supplementary Table 5) confirmed the conclusion drawn from the pre-steady-state kinetic data: the presence of bromide ions changes the kinetics of transition between the open and closed states of LinB86. For low bromide concentrations a slowing down of τ1 is apparent, however, at bromide concentrations higher than 50 mM, τ1 is apparently reduced until the difference between LinB92-Atto655 and its control variant LinB99Atto655 is no longer visible (Figure 2E). Careful inspection of the autocorrelation functions and kinetic data indicates that the conformational dynamics is shifted to longer millisecond timescales at elevated bromide concentrations. This overlays the component τ1 with the diffusion part of G(τ), τD, leading to an apparent speeding up of τ1. Analogous behavior is also observed for LinB92-AttoOxa12 and LinB99AttoOxa12 (Supplementary Table 6). The reproducibility with a different dye supports the conclusion that the τ1 component of the FCS data reflects the “on and off” switch of fluorescence caused by the PET process, which itself is brought upon by the conformational change of the enzyme during the molecular gating. Molecular dynamics simulations. The molecular dynamics simulations of LinB86 and LinB92Atto655 revealed that the gating mechanism includes a large movement of one of the cap domain helices 9 ACS Paragon Plus Environment

Page 11 of 31 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


(Figure 3). Since both LinB86 and LinB92-Atto655 displayed very similar opening in the simulations, we conclude that the Atto655 dye does not disturb the dynamics of the gating process (Supplementary Table 9). In the open conformation, the Atto655 probe binds to the mouth of the access tunnel leading to the active site (Figure 3). Consequently, the fluorophore of the Atto655 probe is in a close vicinity of Trp177 promoting the PET process. This open conformation corresponds to the non-fluorescent state. In contrast, the closed state of LinB92-Atto655 reduces the interaction possibilities for Atto655 and Trp177, which hinders PET and leads to the fluorescent state of the probe. Thus, a two-state model can be applied for the equilibrium between the two distinct conformations of open non-fluorescent and closed fluorescent protein states (Figure 2B). The opening of the enzyme increases the main access tunnel bottleneck from 1.3 Å to 1.9 Å, and the passage of the ligands through the open state displays lower energy barriers (Supporting Information Table 9 and Section VII).



Page 12 of 31

Figure 3: Molecular dynamics simulations. Individual snapshots of the two states observed for the LinB92-Atto655 complex in the molecular dynamics simulations. (A) The fluorescent probe Atto655 (cyan) is attached to the residue Cys172 and in the closed conformation it cannot reach the Trp177 (blue). (B) In the open conformation the fluorescence of the Atto655 probe is quenched by the interaction with tunnelblocking Trp177 (blue). A flexible helix of the cap domain (in red) represents the gating element, which is undertaking periodical conformational transitions leading to open and closed states. The conformations are defined by the distance of the Cα atoms of residues Trp177 and Asp147 (start of the red helix), which is the metrics used in the Markov state model construction. The closed state resembles the crystal structure of LinB86 (PDB ID 5LKA) and is the predominant conformation in the simulations with the equilibrium probability >98% in both the LinB92-Atto655 and LinB86 simulations (Supplementary Table 9). Application of the two-state model from the molecular dynamics simulations to the PET-FCS results allowed the calculation of the koff and kon rate constants (Table 1), which coincided well with the k-1 and k1 rate constants determined by kinetic measurements (Supplementary Table 2). These experimental results show that the fluorescent closed state is preferred over the non-fluorescent open one. This corresponds also to the molecular dynamic simulations where the equilibrium probability of the open state was 1.2 ± 0.5% 11 ACS Paragon Plus Environment

Page 13 of 31 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


in the LinB92-Atto655 simulations and 2.2 ± 0.1% in the LinB86 simulations (Supplementary Table 9). The kon and koff values obtained from the PET-FCS experiments are dependent on bromide concentrations in good correspondence with the pre-steady-state kinetic data (Supplementary Table 2). Taken together, these data confirm dynamics of a molecular gate is not significantly disturbed by the attachment of the PET probe and that the molecular gating is an essential part of the catalytic cycle of LinB86. Table 1. Estimated kon and koff rate constants characterizing the transition between the non-fluorescent and fluorescent state of LinB92-Atto655. The values were calculated from parameters α1 and τ1 (Supplementary Table 5) assuming a two-state model. The data for AttoOxa12 are summarized in Supplementary Table 7. LinB92-Atto655 LinB92-Atto655 + 50 mM BrLinB92-Atto655 + ≥ 100 mM Br-

kon (s-1) 1822±488 899±283 ≤ 600*

koff (s-1) 60±31 28±16 ≤ 20*

*The lower threshold of the rate constants was qualitatively estimated by assuming the PET component is overlapping with the diffusion, i.e., being on the order of 2 ms.

DISCUSSION The modifications of the LinB variants presented in this study are in the access tunnels of the enzyme and they affect the transport of the ligands. The kinetic results revealed that LinB32 and LinB86 have a more complex dynamical profile than LinBwt. Leu177Trp mutation in LinB32 causes a tighter packing of the tunnels that reduces its dynamics while enhancing the chemical step of the catalysis. On the other hand, the additional mutations Trp140Ala+Phe143Leu+Ile211Leu in LinB86 lowered interactions in the cap domain of the enzyme. This engineering step re-established the dynamical gate-like movement of the cap domain helix that controls the transport of substrates, products and solvent molecules. Such molecular gating elements and large structural conformational changes are common among many proteins.24–26 The rate constants that characterize the equilibrium between the two conformational states of LinB86 agree for all the three different methodologies applied. Closed state of the enzyme LinB86 is the predominant one. Both the pre-steady-state kinetics and the PET-FCS experiments displayed slowing of the 12 ACS Paragon Plus Environment


Page 14 of 31

enzyme’s dynamical behavior at elevated concentrations of bromide ions. The obtained data indicates that the gating dynamics is affected significantly by the concentration of this reaction product and confirms that both methods are indeed looking at the same phenomenon. Single-molecule experiments have been previously conducted to study the conformations of transmembrane proteins and chaperones, and to observe protein folding and denaturation pathways.19,20,27– 34

Here we have shown that the single-molecule PET-FCS is also an excellent method for studying gating

and ligand transport in enzymes. Using protein engineering, we have tuned the dynamics of the main tunnel (occurring in micro- to millisecond time scale) which turned out to be crucial for the ligand transport in LinB32 and LinB86. Would it be possible to modify the enzyme structure so that the open state is more predominant in order to further increase accessibility of buried active site, as the product release is still the rate-limiting step? There is a good chance that increasing the dynamics of the protein would increase the activity even further, however, it is known that promoting flexibility of enzymes generally leads to problems with their stability.35 We have increased the flexibility of the cap domain from LinB32 to LinB86 by mutating larger amino acid residues into smaller ones with the intention of creating a de novo access tunnel.7 Based on the results reported in our previous study, the newly opened auxilliary tunnel is used in ligand transport in LinB86, even though the main tunnel is still the preferred transport pathway.13 Here we show that the unintentional facilitation of the cap domain movements and the enhanced dynamics of the main tunnel explain the increased enzymatic efficiency more than auxiliary tunnel, even though this new tunnel is functional and used by the ligands. Excellent agreement of kinetic data and PET-FCS experiments in the transitions between the open and closed states let us believe that modified conformational dynamics is the mechanism for improved catalytic efficiency of LinB86. Back mutation introduced to LinB86+Trp177Leu protein resulted in a significant drop in activity, suggesting that introduced gating element is composed of both Trp177 and a flexible helix of the cap domain. This gating element is undertaking periodic open-closed 13 ACS Paragon Plus Environment

Page 15 of 31 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


transitions, thus balancing physical and chemical steps of the catalytic cycle. While the current trend in enzyme engineering is to design the best environment for stabilization for the transition state, the access tunnels and their gating elements represent a natural hotspots for mutagenesis, especially in enzymes with buried active sites and the rate-limitation in the physical steps of the catalytic cycle.1,2,36

CONCLUSIONS We have recently demonstrated that introduction of a de novo tunnel in the enzyme with buried active site can provide a biocatalyst with superior catalytic performance.7 The main objective of this study was to carry out in-depth analysis of the effect of mutations on functional properties and protein dynamics using several experimental and theoretical techniques. The introduction of a bulky tryptophan residue in the access tunnel of LinB lead to the variant LinB32, which displayed blocked access tunnel and slower dynamical movements of the cap domain. This engineering step had an adverse effect on the product release, while it at the same time increased the rates of the chemical steps with the model substrate 1,2dibromoethane. The mutation of three additional residues in de novo tunnel resulted in LinB86, which is the most catalytically efficient haloalkane dehalogenase variant ever. This variant benefited both from the increased rate of the chemicals step, as LinB32, and the re-established dynamical movements of the cap domain that allows better transportation of substrates and products. Here, we have captured the dynamical reasons for the superior activity of LinB86 using single-molecule PET-FCS in combination with pre-steadystate kinetics and molecular dynamics simulations. We have revealed the molecular mechanisms of the excellent catalytic behavior of LinB86 and the usability of the PET-FCS methodology for the determination of dynamical movements of proteins. Moreover, we demonstrate for the first time that PET-FCS is a powerful tool for monitoring the conformational dynamics of a soluble enzyme. Since the dynamical cap domain is one of the determining factors for the activity and specificity shared within the haloalkane dehalogenase enzyme family,37,38 we believe that presented data can initiate a future single-molecule studies 14 ACS Paragon Plus Environment


Page 16 of 31

deciphering the role of dynamics to enzymatic catalysis. Moreover, our results illustrate the effective use of the combination of methodologies to study protein dynamical behavior and offer a solid base for studying protein structures with engineered dynamics. Engineering the gating mechanisms in enzymes with buried active sites opens new possibilities in protein engineering outside perfecting the active site complementarity.

METHODS Rapid-quench analysis: Rapid quench flow experiments were performed at 37°C in a 100 mM glycine buffer pH 8.6 using the rapid quench flow instrument model QFM 400 (BioLogic, France). The reaction was started by rapid mixing of 75 µl enzyme with 75 µl substrate solution and quenched with 100 µl 0.8 M H2SO4 after time intervals ranging from 5 ms to 1 s. The quenched mixture was directly injected into 0.5 ml of ice-cold diethyl ether with 1,2-dichloroethane as the internal standard. After extraction, the diethyl ether layer containing non-covalently bound substrate and alcohol product was collected, dried on a short column containing anhydrous Na2SO4 and analyzed on gas chromatograph Agilent 7890 (Agilent, USA) equipped with capillary column DB-FFAP (30m x 0.25mm x 0.25μm, Phenomenex) and connected with mass spectrometer Agilent 5975C (Agilent, USA). The amount of halide in the water phase was measured by ion chromatograph 861 Advanced Compact IC equipped with METROSEP A Supp 5 column (Metrohm, Switzerland). Stopped-flow analysis: The experiments were performed by using the stopped-flow instrument SFM-300 (BioLogic, France) combined with MOS-200 spectrometer equipped with a Xe arc lamp. The dead time of the instrument was 0.4 ms. Fluorescence emission from tryptophan residues was observed through a 320 nm cut-off filter upon excitation at 295 nm. All reactions were performed at 37°C in a 100 mM glycine buffer of pH 8.6. Multiple turnover fluorescence traces were recorded upon rapid mixing of an 15 ACS Paragon Plus Environment

Page 17 of 31 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


enzyme with 0-4 mM 1,2-dibromoethane, each trace shows the average of seven individual experiments. Bromide binding was measured upon rapid mixing of enzyme with bromide to a final concentration of 0-1 M. The bromide dissociation was analyzed under rapid dilution of enzyme pre-incubated with 50-300 mM bromide and consequently mixed 1:1 with pure buffer. To detect the rapid equilibrium phase during bromide dissociation, enzyme pre-incubated with 150 mM bromide was rapidly diluted by mixing 1:1 with a buffer containing 0, 50, 100 and 150 mM bromide. Steady-state kinetics: The initial rates of reaction of LinB variants with 1,2-dibromoethane were measured using a VP-ITC isothermal titration microcalorimeter (MicroCal, USA) at 37°C in a glycine buffer pH 8.6. The micro-calorimeter reaction mixture vessel was filled with 1.4 mL of enzyme solution at a concentration of 0.004-0.01 mg/mL. The substrate solution was titrated at 150 s intervals in the reaction mixture. Each injection increased the substrate concentration until the reaction rate became saturated. A total of 28 injections were carried out during the titration. The reaction rates reached after every injection (in units of thermal power) were converted to enzyme activity. The substrate concentration was verified by the gas chromatography (Finnigan, USA). Data analysis and statistics: The kinetic data were fitted using dynamic kinetic simulation program KinTek Global Kinetic Explorer version 6.0 (KinTek, USA). Data fitting used numerical integration of rate equations from an input model searching a set of parameters that produce a minimum χ2 value using nonlinear regression based on the Levenberg-Marquardt method.39,40 The residuals were normalized by sigma value for each data point. The thermodynamic cycle was enforced to maintain a net free energy change of zero when fitting data. The standard error was calculated from the covariance matrix during nonlinear regression in globally fitting all fluorescence traces for each binding and dissociation experiment.41 Detailed information for particular kinetic data analysis and solving specific kinetic models are described in Supplementary Information.



Page 18 of 31

Construction of the mutants for the labelling: Expression of the different LinB variants was achieved using E. coli BL21(DE3) containing the created expression vectors. LB medium with ampicillin (1 l) was inoculated with 5 ml freshly grown pre-culture. The bacteria were grown at 37 °C, till optical density measured at 600 nm reached approximately 0.6. Then, the cells were cooled below 20 °C and 0.5 mM IPTG was added and growth continued at 20 °C. After overnight growth, cells were harvested using centrifugation for 10 min at 8000 rpm. A washing step was carried out using approximately 125 ml Wash Buffer. Cells were harvested by centrifugation (10 min at 8000 rpm). Cell pellets were dissolved in approximately 40 ml of Purification Buffer A and then frozen at -80 °C. Disruption was achieved by sonication (4*8 min, 0.3 cycle, 85 amplitude). Crude extract was centrifuged (20-60 min at 14000 rpm) to remove cell debris. Cell free extracts were filtered through a Rotilabo-Spritzenfilter (0.45 μm) to remove remainders of cell debris. Samples were purified using metal affinity chromatography with a small adaptation of step wise increased imidazole concentration for washing (10 ml 4%, 10 ml 6% and 15 ml 10%). Finally, imidazole was removed by a double dialysis step in 50 mM phosphate buffer, pH 7.5. Protein samples were filtered through a sterile 0.22 μm Rotilabo-Spritzenfilter for storage. Concentration of prepared LinB variants was determined in triplicates using Bradford reagent and the purity was checked by SDS-PAGE. Enzyme labelling: Labelling of enzymes (LinB92 and LinB99) with Atto-655-Maleimide and AttoOxa12-Maleimide (Atto-Tec GmbH, Germany) was performed in native buffer (50 mM potassium phosphate buffer, pH 7.4 at 20°C). The protein solution (0.2 mg/ml) was mixed with cysteine-reactive Atto analogues added from their lyophilized stock at molar ratio of 1:1. After incubation at 10oC for 1 h, the unreacted label was removed by the desalting column (Zeba Spin Column, Fisher Scientific). The collected fraction was then further dialyzed overnight against 50 mM phosphate buffer at 6°C (MWCO 14 kDa) to remove traces of the free dye. In parallel, for FCS measurement, labelled enzyme was further diluted, generally 10000x, in phosphate buffer and dialyzed against 100 mM Glycine buffer (pH 8.6 at 20 oC) overnight at 6°C (MWCO 14 kDa).42 17 ACS Paragon Plus Environment

Page 19 of 31 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


Kinetics of the labelled enzymes: The activity of LinB variants was tested with 1,2-dibromoethane at 37°C. The activity was examined in 10 ml 100 mM glycine buffer, pH 8.6 with 10 μl of substrate in 25ml Reaction Flasks closed by Mininert Valves. The reactions were initiated by addition of 100 μl enzyme with concentration of approximately 0.2 mg/ml. Samples (1 ml) were taken at 3-5 min time intervals until 20 min and immediately mixed with 35% nitric acid to stop the enzymatic reaction. Increasing concentration of reaction product was measured spectrofotometrically at 460 nm using the Iwasaki protocol with mercuric thiocyanate and ferric ammonium sulphate using a microplate reader Sunrise (Tecan, Switzerland).43 Dehalogenation activities were quantified by the slope of the relationship between the product concentration and the time. UV-VIS absorption and steady-state fluorescence measurement: Absorption spectra were recorded on Shimadzu UV-2600 spectrophotometer at the speed of 150 nm/min. Steady state fluorescence spectra were measured on Jobin-Yvon Fuorolog 3 spectrometer at the speed 100 nm/min with the emission and excitation bandwidth set equally to 2 nm. Excitation wavelength coincided with FCS experiment (635 nm). Signal was corrected for background noise. Fluorescence steady state anisotropy was measured with inserted polarizers. Four intensities of polarized fluorescence (IVV, IVH, IHV, IHH) were recorded, where "V" and "H" in the subscripts represent the vertical and horizontal directions, and the first and second subscript denotes the direction plane of the polarization in the excitation and emission arm, respectively. The steadystate anisotropy is calculated as follows: 𝐼𝑉𝑉 ― 𝐺𝐼𝑉𝐻

𝑟𝑠𝑡 = 𝐼𝑉𝑉 +

(1)

2𝐺𝐼𝑉𝐻

in which G is the instrumental correction factor given by the observed ratio, IHV/IHH. Far-UV Circular dichroism (CD) spectra: Far-UV CD spectra of the labelled and non labelled enzymes were measured in native buffer in 0.2 mm quartz cuvette at 10°C on Jasco 1500 CD spectrophotometer. Concentration of enzyme was adjusted to avoid pile up effect on the detector. Measured 18 ACS Paragon Plus Environment


Page 20 of 31

CD data were expressed as mean residue elipticity [Θ]MR bearing the units of deg.cm2.dmol-1 according to equation (2): [𝛩]𝑀𝑅 =

100𝛩

(2)

𝑙𝑐𝑛

where Θ is acquired ellipticity (mdeg), l is cuvette path length (mm), c is molar protein concentration, and n is the number of peptide bonds in the protein.44 FCS setup and data analysis: FCS data acquisitions were performed in glass-bottom Lab-Tek chambers (ThermoFisher). To minimize adsorption of proteins on glass surface, chambers were incubated with 0.01% poly-L-lysine solution (15-20 min) followed by washing 20 times with deionized water. Prior to measurements, labelled protein was further diluted to 1 nM concentration in 100 mM glycine buffer, pH = 8.6. All experiments were carried out at 20°C. FCS measurements were performed on a standard inverse fluorescence microscope (Olympus IX71, combined with Microtime 200, PicoQuant, Germany) equipped with a diode laser as the continuous excitation source (635 nm, LDH-D-C635 PicoQuant,Germany). Collimated laser beam was guided into a water-immersion 60x objective (NA 1.2, Olympus, Japan) by a dichroic mirror (z635, Chroma filters, USA). The back aperture of the objective was underfilled to achieve enlarged focal volume (its FWHM approximately 8 times larger than the Airy disk). The average laser power was adjusted to be 185 μW leaving the objective plane. The fluorescence signal was collected by the same objective passing through the dichroic mirror onto a 150 μm pinhole. The emission signal was further guided through a band-pass filter (Semrock, FF01-697/58-25, USA), separated by 60/40 beam splitter plate and imaged onto the active area of two avalanche photodiodes (Perkin-Elmer APD, SPCM-AQRH, Canada), which were synchronized to circumvent the dead-time of the electronics and after-pulsing effects. The signal was recorded in the pt2 mode using a Picoharp 300 device (PicoQuant, Germany).45 The data were analyzed using PicoQuant Symphotime software yielding the cross-correlation curves G(τ), which


Page 21 of 31 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


were further fitted, using in-house written Origin and Matlab scripts, with equation (3) describing both diffusion and PET components:

(

𝜏

𝐺(𝜏) = (1 + 𝜏𝐷)

―1

) (1 + ∑ 𝛼 exp ( ― )) 𝜏

𝑛 𝑛

𝜏𝑛

(3)

where, ρ and τD represent the diffusion time and corresponding amplitude; and αn and τn stand for the PET relaxation amplitudes and corresponding PET time constants. The data were further evaluated assuming a two-state transition between fluorescent and non-fluorescent conformational states. Validity of the assumption was supported by molecular dynamics simulations. Rate constants kon and koff were calculated according to: 𝛼1 =

𝑘𝑜𝑓𝑓

(4)

𝑘𝑜𝑛 1

(5)

𝜏1 = 𝑘𝑜𝑓𝑓 + 𝑘𝑜𝑛

where, kon corresponds to the transition from the non-fluorescent form to the fluorescent one, while rate constant koff describes the opposite process. Computational design of a system for single molecule analysis: Sequences of six experimentally characterized haloalkane dehalogenases DhaA, LinB, DrbA, DmbC, DhlA and DmbB, were used as queries for PSI-BLAST searches against nr database of NCBI (for details see Supplementary Information IV).46,47 A multiple sequence alignment was constructed and used for estimation of the level of conservation at individual sites within the haloalkane dehalogenase HLD-II subfamily. Evolutionary variable spots for Cys replacement were visually inspected and their solvent accessibility, a fluctuation quantified by B-factors and the distance to Trp177 were assessed. Computational parameters for the Cys-Atto-655 complex: Using Avogadro 1.0.3 program,48 the structure of the Cys-Atto655 residue (Supplementary Figure 9) was prepared in the extended α-helical conformation of backbone atoms, and the eight lowest energy conformations of PET probe were identified 20 ACS Paragon Plus Environment


Page 22 of 31

for each backbone conformation with weighted rotor search keeping the backbone atoms fixed. Next, the Cys-Atto655 residue was capped with N-methylamide (NME) and acetyl (ACE) residues and the geometry of all 16 selected conformations were energy minimized with b3lyp/6-31G* using Gaussian09 program revision D.01.49 The two lowest energy conformations for each backbone conformations were used for fitting the partial atomic charges of the novel Cys-Atto655 residue using RESP ESP charge derive (R.E.D.) server 2.050,51 with HF/6-31G* level of theory and Gaussian09 program included in the R.E.D. server. The charges on the Cys-Atto655 residue were derived using RESP-A1A charge model and multi-conformation multi-orientation RESP fit.52 During the fitting procedure, the charges on the capping NME and ACE residues were constrained to zero and the charges on the four atoms of peptide bond (Atom IDs 1, 2, 4 and 5) were constrained to the values corresponding to the respective atoms of the electro-neutral residues from the Cornell et al. force field.53 The atom types were derived by analogy with the Cornell et al. force field with the following exceptions: (i) the parameters for the aromatic oxygen (atom ID 84) were obtained from the study of VanBeek et al. 54, (ii) the parameters for S6 atom type from the GAFF force field55 were used for the sulfonate sulfur (Atom ID 50), and (iii) the missing parameters for the sp2 nitrogens (Atom IDs 20, 32, 58 and 72) were generated by parmchk2 program form AMBERTools 16 suite.56 The assigned atom types and derived partial charges of Cys-Atto655 residue are summarized in Supplementary Table 8. Molecular dynamics simulations: The crystal structure of LinB86 (PDB ID 5LKA) was downloaded from RSCB Protein Data Bank.13 The hydrogen atoms were added to the structure with the H++ web server.46 The water molecules from the crystal structure, which did not overlap with the protonated structure, were retained. The Cys-Atto655 residue was manually inserted to the position 172 using PyMol so that the fluorophore was not interacting with any protein atoms at the start of the simulations. The resulting structure was solvated using the tLeap module of AmberTools 14 in a cubic water box of TIP3P water molecules so that all protein atoms were at least 10 Å from the surface of the box.56,57 Cl- and Na+ ions were added to neutralize the charge of the protein and to get a final concentration of 0.1M. The system 21 ACS Paragon Plus Environment

Page 23 of 31 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


was equilibrated using the Equilibration_v1 module of High Throughput Molecular Dynamics (HTMD) and the simulations were carried out in AceMD. The system was first minimized using conjugate-gradient method for 500 steps. Then the system was heated and minimized as follows: (i) 500 steps (2 ps) of NVT thermalization to 310 K with the Berendsen barostat with constraints on all heavy atoms of the protein, (ii) 625000 steps (2.5 ns) of NPT equilibration with Langevin thermostat with 1 kcal*mol-1*Å-2 constraints on all heavy atoms of the protein and (iii) 625000 steps (2.5 ns) of NPT equilibration with the Langevin thermostat without constraints. Holonomic constraints were set on all hydrogen-heavy atom bond terms and the mass of hydrogen atoms was scaled with factor 4, enabling the 4 fs timestep.58–61 The simulations at 310 K employed periodic boundary conditions, using the particle mesh Ewald method for treatment of interactions beyond 9 Å cut-off, electrostatic interactions suppressed >4 bond terms away from each other and the smoothing and switching functions of van der Waals and electrostatic interactions were enabled to avoid drastic changes at the cut-off from 7.5 Å to 9 Å. HTMD with AceMD were used to perform adaptive sampling of the LinB92 system.62 The 20 ns production runs were started with the files resulting from the equilibration and they employed the same settings as the last step of the equilibration. Adaptive sampling was run first using the distance of the PET probe and the Trp177 as the reaction coordinate, then sampling the χ2 dihedral angle of the Trp177, the binary contact map between Atto655 probe and Trp177 atoms and finally the RMSD of the backbone Cα atoms. The total simulation time used in calculations was 8720 ns saved with 0.1 ns intervals. The distance between Trp177 and Asp147 was used as the metric for the Markov model construction. The data was clustered using MiniBatchKmeans algorithm to 200 clusters. The implied timescale plots of Markov state model at various lag times were constructed and a lag time of 10 ns was selected for the two-state Markov model construction (Supplementary Figure 10). To estimate the errors in the data, 1000 individual 2-state Markov models were constructed with random 50% bootstrapped data using the previous parameters. Simulations and analysis with similar parameters were calculated for LinB86 without the Atto-655 probe to ensure that observed protein conformational change is not caused by the attachment of PET probe. These simulations were started from LinB86 crystal structure (PDB ID 5LKA),13 22 ACS Paragon Plus Environment


Page 24 of 31

followed the previously described equilibration schemes and the adaptive sampling was conducted on the RMSD of the backbone Cα atoms. Total simulation time was 3920 ns and the construction of the Markov state models and bootstrapping were performed similarly to those described for LinB92-Atto655 complex (Supplementary Table 9). ACKNOWLEDGEMENTS The authors thank Gaspar Pinto and Ondrej Vavra (Masaryk University) for conducting CaverDock calculations. The authors further thank the Grant Agency of the Czech Republic (GA16-07965S, and GA1606096S), the Czech Ministry of Education (LQ1605, LM2015051, CZ.02.1.01/0.0/0.0/16_013/0001761, LM2015047, and LM2015055) and the European Union (720776). Computational resources were supplied by the Ministry of Education, Youth and Sports of the Czech Republic under the Projects CESNET (Project No. LM2015042) and CERIT-Scientific Cloud (Project No. LM2015085). SUPPORTING INFORMATION Details of pre-steady-state kinetics of 1,2-dibromoethane conversion, pre-steady state kinetics of bromide binding and dissociation, stopped-flow fluorescence analysis of 2-bromoethanol conversion, steady-state kinetics of LinB86+W177L, single molecule PET-FCS results of enzyme dynamics and computational design, molecular dynamics simulations and ligand binding simulations. REFERENCES (1) (2) (3) (4)

Prokop, Z.; Gora, A.; Brezovsky, J.; Chaloupkova, R.; Stepankova, V.; Damborsky, J. Engineering of Protein Tunnels: Keyhole-Lock-Key Model for Catalysis by the Enzymes with Buried Active Sites. In Protein Engineering Handbook, Vol. 3.; Weinheim: Wiley-VCH, 2012; pp 421–464. Gora, A.; Brezovsky, J.; Damborsky, J. Gates of Enzymes. Chem. Rev. 2013, 113 (8), 5871–5923. Huang, P.-S.; Boyken, S. E.; Baker, D. The Coming of Age of de Novo Protein Design. Nature 2016, 537 (7620), 320–327. Maria-Solano, M. A.; Serrano-Hervás, E.; Romero-Rivera, A.; Iglesias-Fernández, J.; Osuna, S. Role of Conformational Dynamics in the Evolution of Novel Enzyme Function. Chem. Commun.. 2018, 54 (50), 6622–6634.


Page 25 of 31 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

(5) (6) (7) (8) (9) (10) (11)

(12) (13) (14) (15) (16) (17) (18) (19) (20) (21)


Campbell, E. C.; Correy, G. J.; Mabbitt, P. D.; Buckle, A. M.; Tokuriki, N.; Jackson, C. J. Laboratory Evolution of Protein Conformational Dynamics. Curr. Opin. Struct. Biol. 2018, 50, 49– 57. Chen, Z.; Zeng, A.-P. Protein Engineering Approaches to Chemical Biotechnology. Curr. Opin. Biotechnol. 2016, 42, 198–205. Swint-Kruse, L. Using Evolution to Guide Protein Engineering: The Devil Is in the Details. Biophys. J. 2016, 111 (1), 10–18. Romero‐Rivera, A.; Iglesias‐Fernández, J.; Osuna, S. Exploring the Conversion of a D-Sialic Acid Aldolase into a L-KDO Aldolase. Eur. J. Org. Chem. 2018, 2018 (20–21), 2603–2608. Liskova, V.; Bednar, D.; Prudnikova, T.; Rezacova, P.; Koudelakova, T.; Sebestova, E.; Smatanova, I. K.; Brezovsky, J.; Chaloupkova, R.; Damborsky, J. Balancing the Stability-Activity Trade-Off by Fine-Tuning Dehalogenase Access Tunnels. ChemCatChem 2015, 7 (4), 648–659. Li, G.; Yao, P.; Gong, R.; Li, J.; Liu, P.; Lonsdale, R.; Wu, Q.; Lin, J.; Zhu, D.; Reetz, M. T. Simultaneous Engineering of an Enzyme’s Entrance Tunnel and Active Site: The Case of Monoamine Oxidase MAO-N. Chem. Sci. 2017, 8 (5), 4093–4099. Yan, X.; Wang, J.; Sun, Y.; Zhu, J.; Wu, S. Facilitating the Evolution of Esterase Activity from a Promiscuous Enzyme (Mhg) with Catalytic Functions of Amide Hydrolysis and Carboxylic Acid Perhydrolysis by Engineering the Substrate Entrance Tunnel. Appl. Environ. Microbiol. 2016, 82 (22), 6748–6756. Biedermannová, L.; Prokop, Z.; Gora, A.; Chovancová, E.; Kovács, M.; Damborsky, J.; Wade, R. C. A Single Mutation in a Tunnel to the Active Site Changes the Mechanism and Kinetics of Product Release in Haloalkane Dehalogenase LinB. J. Biol. Chem. 2012, 287 (34), 29062–29074. Brezovsky, J.; Babkova, P.; Degtjarik, O.; Fortova, A.; Gora, A.; Iermak, I.; Rezacova, P.; Dvorak, P.; Smatanova, I. K.; Prokop, Z., Chaloupkova, R.; Damborsky, J. Engineering a de Novo Transport Tunnel. ACS Catal. 2016, 6 (11), 7597–7610. Sauer, M.; Neuweiler, H. PET-FCS: Probing Rapid Structural Fluctuations of Proteins and Nucleic Acids by Single-Molecule Fluorescence Quenching. In Fluorescence Spectroscopy and Microscopy; Methods in Molecular Biology; Humana Press, Totowa, NJ, 2014; pp 597–615. Li, X.; Xing, J.; Qiu, Z.; He, Q.; Lin, J. Quantification of Membrane Protein Dynamics and Interactions in Plant Cells by Fluorescence Correlation Spectroscopy. Mol. Plant 2016, 9 (9), 1229–1239. Verschueren, K. H.; Seljée, F.; Rozeboom, H. J.; Kalk, K. H.; Dijkstra, B. W. Crystallographic Analysis of the Catalytic Mechanism of Haloalkane Dehalogenase. Nature 1993, 363 (6431), 693– 698. Prokop, Z.; Monincová, M.; Chaloupková, R.; Klvaňa, M.; Nagata, Y.; Janssen, D. B.; Damborský, J. Catalytic Mechanism of the Haloalkane Dehalogenase LinB from Sphingomonas Paucimobilis UT26. J. Biol. Chem. 2003, 278 (46), 45094–45100. Fersht, A. Measurement and Magnitude of Individual Rate Constants. In Structure and mechanism in protein science; W. H. Freeman and Co.: New York, NY, 1998; pp 147–149. Doose, S.; Neuweiler, H.; Sauer, M. Fluorescence Quenching by Photoinduced Electron Transfer: A Reporter for Conformational Dynamics of Macromolecules. Chemphyschem 2009, 10 (9–10), 1389–1398. Schulze, A.; Beliu, G.; Helmerich, D. A.; Schubert, J.; Pearl, L. H.; Prodromou, C.; Neuweiler, H. Cooperation of Local Motions in the Hsp90 Molecular Chaperone ATPase Mechanism. Nat. Chem. Biol. 2016, 12 (8), 628–635. Ries, J.; Schwarze, S.; Johnson, C. M.; Neuweiler, H. Microsecond Folding and Domain Motions of a Spider Silk Protein Structural Switch. J. Am. Chem. Soc. 2014, 136 (49), 17136–17144.



Page 26 of 31

(22) Vaiana, A. C.; Neuweiler, H.; Schulz, A.; Wolfrum, J.; Sauer, M.; Smith, J. C. Fluorescence Quenching of Dyes by Tryptophan: Interactions at Atomic Detail from Combination of Experiment and Computer Simulation. J. Am. Chem. Soc. 2003, 125 (47), 14564–14572. (23) Doose, S.; Neuweiler, H.; Sauer, M. A Close Look at Fluorescence Quenching of Organic Dyes by Tryptophan. Chemphyschem 2005, 6 (11), 2277–2285. (24) Hanoian, P.; Liu, C. T.; Hammes-Schiffer, S.; Benkovic, S. Perspectives on Electrostatics and Conformational Motions in Enzyme Catalysis. Acc. Chem. Res. 2015, 48 (2), 482–489. (25) Narayanan, C.; Bernard, D. N.; Doucet, N. Role of Conformational Motions in Enzyme Function: Selected Methodologies and Case Studies. Catalysts 2016, 6 (6):81, 1-17. (26) Kohen, A. Role of Dynamics in Enzyme Catalysis: Substantial versus Semantic Controversies. Acc. Chem. Res. 2015, 48 (2), 466–473. (27) Ding, B.; Hilaire, M. R.; Gai, F. Infrared and Fluorescence Assessment of Protein Dynamics: From Folding to Function. J. Phys. Chem. B 2016, 120 (23), 5103–5113. (28) Kleckner, I. R.; Foster, M. P. An Introduction to NMR-Based Approaches for Measuring Protein Dynamics. Biochim. Biophys. Acta 2011, 1814 (8), 942–968. (29) Orozco, M. A Theoretical View of Protein Dynamics. Chem. Soc. Rev. 2014, 43 (14), 5051–5066. (30) Lu, H. P. Sizing up Single-Molecule Enzymatic Conformational Dynamics. Chem. Soc. Rev. 2014, 43 (4), 1118–1143. (31) Yang, H.; Luo, G.; Karnchanaphanurach, P.; Louie, T.-M.; Rech, I.; Cova, S.; Xun, L.; Xie, X. S. Protein Conformational Dynamics Probed by Single-Molecule Electron Transfer. Science 2003, 302 (5643), 262–266. (32) Blank, K.; De Cremer, G.; Hofkens, J. Fluorescence-Based Analysis of Enzymes at the SingleMolecule Level. Biotechnol. J. 2009, 4 (4), 465–479. (33) Rogers, J. M. G.; Polishchuk, A. L.; Guo, L.; Wang, J.; DeGrado, W. F.; Gai, F. Photoinduced Electron Transfer and Fluorophore Motion as a Probe of the Conformational Dynamics of Membrane Proteins: Application to the Influenza A M2 Proton Channel. Langmuir 2011, 27 (7), 3815–3821. (34) Neuweiler, H.; Johnson, C. M.; Fersht, A. R. Direct Observation of Ultrafast Folding and Denatured State Dynamics in Single Protein Molecules. Proc. Natl. Acad. Sci. U. S. A. 2009, 106 (44), 18569–18574. (35) Miller, S. R. An Appraisal of the Enzyme Stability‐activity Trade‐off. Evolution 2017, 71 (7), 1876–1887. (36) Kingsley, L. J.; Lill, M. A. Substrate Tunnels in Enzymes: Structure-Function Relationships and Computational Methodology. Proteins 2015, 83 (4), 599–611. (37) Koudelakova, T.; Chovancova, E.; Brezovsky, J.; Monincova, M.; Fortova, A.; Jarkovsky, J.; Damborsky, J. Substrate Specificity of Haloalkane Dehalogenases. Biochem. J. 2011, 435 (2), 345– 354. (38) Tyzack, J. D.; Furnham, N.; Sillitoe, I.; Orengo, C. M.; Thornton, J. M. Understanding Enzyme Function Evolution from a Computational Perspective. Curr. Opin. Struct. Biol. 2017, 47, 131–139. (39) Levenberg, K. A Method for the Solution of Certain Non-Linear Problems in Least Squares. Q. Appl. Math. 1944, 2 (2), 164–168. (40) Marquardt, D. An Algorithm for Least-Squares Estimation of Nonlinear Parameters. J. Soc. Ind. Appl. Math. 1963, 11 (2), 431–441. (41) Johnson, K. A.; Simpson, Z. B.; Blom, T. Global Kinetic Explorer: A New Computer Program for Dynamic Simulation and Fitting of Kinetic Data. Anal. Biochem. 2009, 387 (1), 20–29. (42) Kim, Y.; Ho, S. O.; Gassman, N. R.; Korlann, Y.; Landorf, E. V.; Collart, F. R.; Weiss, S. Efficient Site-Specific Labeling of Proteins via Cysteines. Bioconjug. Chem. 2008, 19 (3), 786–791. (43) Iwasaki, I.; Utsumi, S.; Ozawa, T. New Colorimetric Determination of Chloride Using Mercuric Thiocyanate and Ferric Ion. Bull. Chem. Soc. Jpn. 1952, 25 (3), 226–226. 25 ACS Paragon Plus Environment

Page 27 of 31 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


(44) Kelly, S. M.; Price, N. C. The Use of Circular Dichroism in the Investigation of Protein Structure and Function. Curr. Protein Pept. Sci. 2000, 1 (4), 349–384. (45) Macháň, R.; Kapusta, P.; Hof, M. Statistical Filtering in Fluorescence Microscopy and Fluorescence Correlation Spectroscopy. Anal. Bioanal. Chem. 2014, 406 (20), 4797–4813. (46) Altschul, S. F.; Madden, T. L.; Schäffer, A. A.; Zhang, J.; Zhang, Z.; Miller, W.; Lipman, D. J. Gapped BLAST and PSI-BLAST: A New Generation of Protein Database Search Programs. Nucleic Acids Res. 1997, 25 (17), 3389–3402. (47) NCBI Resource Coordinators. Database Resources of the National Center for Biotechnology Information. Nucleic Acids Res. 2018, 46 (D1), D8–D13. (48) Hanwell, M. D.; Curtis, D. E.; Lonie, D. C.; Vandermeersch, T.; Zurek, E.; Hutchison, G. R. Avogadro: An Advanced Semantic Chemical Editor, Visualization, and Analysis Platform. J. Cheminformatics 2012, 4 (1), 17. (49) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox D. J. Gaussian 09 Rev. D.01; Wallingford, CT, 2009. (50) Dupradeau, F.-Y.; Pigache, A.; Zaffran, T.; Savineau, C.; Lelong, R.; Grivel, N.; Lelong, D.; Rosanski, W.; Cieplak, P. The R.E.D. Tools: Advances in RESP and ESP Charge Derivation and Force Field Library Building. Phys. Chem. Chem. Phys. 2010, 12 (28), 7821–7839. (51) Vanquelef, E.; Simon, S.; Marquant, G.; Garcia, E.; Klimerak, G.; Delepine, J. C.; Cieplak, P.; Dupradeau, F.-Y. R.E.D. Server: A Web Service for Deriving RESP and ESP Charges and Building Force Field Libraries for New Molecules and Molecular Fragments. Nucleic Acids Res. 2011, 39, W511-517. (52) Bayly, C. I.; Cieplak, P.; Cornell, W.; Kollman, P. A. A Well-Behaved Electrostatic Potential Based Method Using Charge Restraints for Deriving Atomic Charges: The RESP Model. J. Phys. Chem. 1993, 97 (40), 10269–10280. (53) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A. A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids, and Organic Molecules. J. Am. Chem. Soc. 1995, 117 (19), 5179–5197. (54) VanBeek, D. B.; Zwier, M. C.; Shorb, J. M.; Krueger, B. P. Fretting about FRET: Correlation between κ and R. Biophys. J. 2007, 92 (12), 4168–4178. (55) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and Testing of a General Amber Force Field. J. Comput. Chem. 2004, 25 (9), 1157–1174. (56) Case, D. A.; Betz, R. M.; Botello-Smith, W.; Cerutti, D. S.; Cheatham III, T. E.; Darden, T. A.; Duke, R. E.; Giese, T. J.; Gohlke, H.; Goetz, A. W.; Homeyer, N.; Izadi, S.; Janowski, P. A.; Kaus, J.; Kovalenko, A.; Lee, T. S.; LeGrand, S.; Li, P.; Lin, C.; Luchko, T.; Luo, R.; Madej, B.; Mermelstein, D.; Merz, K. M.; Monard, G.; Nguyen, H.; Nguyen, H. T.; Omelyan, I.; Onufriev, A.; Roe, D. R.; Roitberg, A.; Sagui, C.; Simmerling, C. L.; Swails, J.; Walker, R. C.; Wang, J.; Wolf, R. M.; Wu, X.; Xiao, L.; York, D. M.; Kollman, P. A., AMBER16. University of California, San Francisco, 2016. 26 ACS Paragon Plus Environment


Page 28 of 31

(57) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79 (2), 926. (58) Harvey, M. J.; De Fabritiis, G. An Implementation of the Smooth Particle Mesh Ewald Method on GPU Hardware. J. Chem. Theory Comput. 2009, 5 (9), 2371–2377. (59) Feenstra, K. A.; Hess, B.; Berendsen, H. J. C. Improving Efficiency of Large Time-Scale Molecular Dynamics Simulations of Hydrogen-Rich Systems. J. Comput. Chem. 1999, 20 (8), 786– 798. (60) Hopkins, C. W.; Le Grand, S.; Walker, R. C.; Roitberg, A. E. Long-Time-Step Molecular Dynamics through Hydrogen Mass Repartitioning. J. Chem. Theory Comput. 2015, 11 (4), 1864– 1874. (61) Naritomi, Y.; Fuchigami, S. Slow Dynamics in Protein Fluctuations Revealed by Time-Structure Based Independent Component Analysis: The Case of Domain Motions. J. Chem. Phys. 2011, 134 (6), 065101. (62) Doerr, S.; De Fabritiis, G. On-the-Fly Learning and Sampling of Ligand Binding by HighThroughput Molecular Simulations. J. Chem. Theory Comput. 2014, 10 (5), 2064–2069.


Page 29 of 31 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


Figure 1 338x190mm (96 x 96 DPI)



Figure 2 1940x1746mm (30 x 30 DPI)


Page 30 of 31

Page 31 of 31 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


Figure 3 355x266mm (300 x 300 DPI)


Molecular Gating of an Engineered Enzyme Captured in Real Time

Recommend Documents