1 Mechanism of Competitive Inhibition and ... - ACS Publications

*Corresponding Author: Joel L. Kaar. University of Colorado Boulder. Department of Chemical and Biological Engineering. Campus Box 596. Boulder, CO 80...
2 downloads 0 Views 2MB Size
Subscriber access provided by READING UNIV

Article

Mechanism of Competitive Inhibition and Destabilization of Acidothermus Cellulolyticus Endoglucanase 1 by Ionic Liquids Samantha R. Summers, Kayla G Sprenger, Jim Pfaendtner, Jan Marchant, Michael Finley Summers, and Joel L Kaar J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.7b08435 • Publication Date (Web): 09 Nov 2017 Downloaded from http://pubs.acs.org on November 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 36

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

Mechanism of Competitive Inhibition and Destabilization of Acidothermus Cellulolyticus Endoglucanase 1 by Ionic Liquids Samantha R. Summers1, K. G. Sprenger2, Jim Pfaendtner2, Jan Marchant3, Michael F. Summers3, Joel L. Kaar1* 1

Department of Chemical and Biological Engineering, University of Colorado, Boulder, CO 80309 2 Department of Chemical Engineering, University of Washington, Seattle, WA 98195 3 Howard Hughes Medical Institute and Department of Chemistry, University of Maryland Baltimore County, Baltimore, MD 21250

*Corresponding Author: Joel L. Kaar University of Colorado Boulder Department of Chemical and Biological Engineering Campus Box 596 Boulder, CO 80309 Tel: (303) 492-6031 Fax: (303) 492-4341 Email: [email protected]

1

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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 36

ABSTRACT The ability of ionic liquids (ILs) to solubilize cellulose has sparked interest in their use for enzymatic biomass processing. However, this potential has yet to be realized, primarily because ILs inactivate requisite cellulases by mechanisms that have yet to be identified. We used a combination of enzymology, circular dichroism (CD), nuclear magnetic resonance (NMR), and molecular dynamics (MD) methods to investigate the molecular basis for the inactivation of the endocellulase 1 (E1) from Acidothermus cellulolyticus

by

the

imidazolium

IL

1-butyl-3-methylimidazolium

chloride

([BMIM][Cl]). Enzymatic studies revealed that [BMIM][Cl] inactivates E1 in a biphasic manner that involves rapid, reversible inhibition followed by slow, irreversible deactivation. Backbone NMR signals of the 40.5 kDa E1 were assigned by triple resonance NMR methods, enabling monitoring of residue-specific perturbations. 1H-15N NMR titration experiments revealed that [BMIM][Cl] binds reversibly to the E1 active site, indicating that reversible deactivation is due to competitive inhibition of substrate binding. Prolonged incubation with [BMIM][Cl] led to substantial global changes in the 1

H-15N HSQC NMR and CD spectra of E1 indicative of protein denaturation. Notably,

weak interactions between [BMIM][Cl] and residues at the termini of several helices were also observed which, together with MD simulations, suggest that E1 denaturation is promoted by [BMIM][Cl]-induced destabilization of helix capping structures. In addition to identifying determinants of E1 inactivation, our findings establish a molecular framework for engineering cellulases with improved IL compatibility.

2

ACS Paragon Plus Environment

Page 3 of 36

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

INTRODUCTION Since the discovery that ionic liquids (ILs) can solubilize cellulose,1 there has been widespread interest in exploiting ILs for the biocatalytic hydrolysis of biomass.2 This interest has stemmed primarily from the potential of ILs as a cleaner alternative to caustic conditions currently employed for biomass processing. Notably, ILs have high thermal stability, negligible vapor pressure, are recyclable, and could potentially be used in a relatively clean, “one-pot” process for both pretreatment (solubilization) and biocatalytic hydrolysis of cellulose - a long-sought approach for generation of alternative fuel sources.3 Although ILs can efficiently solubilize biomass, efforts to develop IL-based conversion approaches have been largely unsuccessful due to their inhibitory influence on cellulase activity. Even at relatively low concentrations (10-15 v/v%), ILs are capable of deactivating a variety of cellulases, including those from microbial and fungal organisms4 and commercial cocktails.5 Efforts to improve IL tolerance using screening and directed evolution approaches have been met with limited success.6 Cellulases from thermophilic and halophilic organisms with somewhat reduced sensitivity to ILs have been identified,4,7 and ILs that dissolve biomass and are less inactivating towards cellulases have also been reported.8–10 However, while these approaches have afforded incremental improvements in cellulase-IL compatibility, cellulase activity required for practical utility has yet to be achieved. Studies aimed at understanding the mechanism of IL-dependent inactivation have led to speculation that ILs either promote enzyme denaturation and/or aggregation, competitively inhibit substrate binding, alter enzyme dynamics, or promote some

3

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

combination of these activities.5,11–14 Since knowledge of the molecular-level mechanism of IL-induced inactivation could facilitate efforts to engineer enzymes with improved compatibility, we herein combined enzyme activity assays with high-resolution NMR, circular dichroism (CD), and molecular dynamics (MD) simulations to probe the molecular mechanism of IL-dependent inactivation of the catalytic domain of endocellulase Cel5A from the hyperthermophile Acidothermus cellulolyticus, which is highly active and thermostable (Topt = 83 °C).15,16 Our studies focused specifically on the Y245G variant of E1 (40.5 kDa) that exhibits reduced sensitivity to product inhibition by cellobiose during cellulose hydrolysis,17 and its interactions with the imidazolium based IL, [BMIM][Cl]. To permit analysis of residue-specific interactions and structural perturbations, we assigned the backbone 1H,

15

N, and

13

C NMR signals for most residues with non-exchanging amide

protons using triple resonance NMR methods. Our studies provide the first NMR assignment of a cellulase catalytic domain, establish the mechanism of IL-induced enzyme inactivation, and suggest a strategy for the rational engineering of cellulases with improved IL compatibility.

METHODS E1 Expression and Purification Escherichia coli cells containing the PET21-b plasmid with the gene for E1 were generously provided by Michael Himmel (National Renewable Energy Laboratory). For activity and stability assays as well as circular dichroism experiments, the cells were grown in Lysogeny broth at 37 °C and induced with IPTG for 3 h. For NMR

4

ACS Paragon Plus Environment

Page 4 of 36

Page 5 of 36

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

experiments, the cells were grown in M9 minimal media at 37 °C with either (for 1H-15N HSQC and 2D-NOESY experiments) or

15

15

NH4Cl

NH4Cl and D-glucose-13C6 (for

triple resonance experiments) as the sole nitrogen and carbon sources, respectively, and were induced with IPTG for 16 h. Following induction, the cells were harvested via centrifugation and E1 was purified as described previously by Baker and co-workers.15 This process of purification involved hydrophobic and anion exchange chromatography followed by incubation at 65 °C for 10 min, which resulted in the precipitation of nonthermostable impurities. The non-thermostable impurities were removed via filtration using a 0.22 um PES membrane. Upon filtration, E1 was dialyzed for 1 h into 10 mM citrate buffer (pH 5.0), resulting in an end sample with a pH of ~7.2. The final purity of labeled and unlabeled E1 for all experiments was greater than 90% as determined by SDS-PAGE. For storage, E1 was kept at 4 °C prior to use for all experiments, at which temperature the enzyme was determined to be stable.

E1 Activity Measurements The activity of E1 was determined via monitoring the hydrolysis of 4-nitrophenyl β-D-cellobioside (5 mM) in 50 mM citrate buffer (pH 5.0) at 65 °C; aliquots of the reaction were periodically quenched with base (0.5 M sodium carbonate, pH ~11.5), facilitating measurement of the colorimetric product, 4-nitrophenolate at 405 nm. For stability measurements, E1 was incubated in 0, 2.5, 5, 10, and 20 w/v% [BMIM][Cl] at 65 °C for 0-24 h, after which the residual relative activity of E1 was assayed using 4nitrophenyl β-D-cellobioside as described above. Although the addition of [BMIM][Cl] altered the pH of the assay solution slightly in a concentration-dependent manner (the

5

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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 36

final pH of the assay solution with 20 w/v% [BMIM][Cl] was 4.7), control experiments showed that such changes in pH had a negligible impact on the activity of E1. For timedependent deactivation measurements, relative activity was determined by normalizing the activity to that immediately after the addition of [BMIM][Cl] at the same concentration without incubation in the IL. For inhibition measurements, relative activity was determined by normalizing activity upon addition of [BMIM][Cl] without incubation against the activity of E1 in the absence of the IL (50 mM citrate, pH 5.0). ILs were obtained from Sigma Aldrich and used without further purification.

CD Measurements of E1 in ILs Near-UV (260-320 nm) spectra of E1 were collected on an Applied Photophysics Chirascan-plus CD spectrometer. Spectra were measured using ~1-3 mg/mL of E1 at 50 °C with addition of 0, 1, 1.5, 2, 3, 4, 5, or 20 w/v% [BMIM][Cl] in a sample cuvette with a 1 mm path length. For time-dependent analysis, E1 was incubated in 5 w/v% or 20 w/v% [BMIM][Cl] at 50 °C for 0-23 h prior to collection of near-UV spectra. For each sample, five spectra were recorded with a 0.5 nm step-size, time-per-point of 0.5 s, and 1 mm bandwidth, and subsequently averaged. Blank reference spectra for each concentration of IL (without enzyme) were subtracted from each sample with E1, thereby accounting for background signal from the IL.

NMR Characterization E1 was partially assigned via a combination of two- and three-dimensional experiments,18

including

CBCA(CO)NH,19,20

6

HNCACB,20,21

ACS Paragon Plus Environment

HNCO,22–24

Page 7 of 36

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

HN(CA)CO,24,25 HN(CO)CA,22,24 and HNCA.22–24 For these experiments, spectra of E1 (~100-200 µM) at pH 7.20 were acquired in a shaped tube at 50 °C with 5 v/v% D2O using a Bruker Avance III 800 MHz spectrometer with a 5 mm 1H{13C/15N} cryogenic probe. Assignments were obtained in a semi-automated fashion. Briefly, tables of spinsystem data were input into the Probabilistic Interaction Network of Evidence Algorithm (PINE)26 and MARS27 for automated assignments. Manual assignments were carried out where automated assignments failed. Secondary structure was determined from a consensus chemical shift index28 approach using chemical shift data for Cα, Cβ, and C’ atoms. All chemical shifts were deposited in the BioMagResBank (accession number to be added prior to publication). Following the assignment of E1, IL-induced chemical shift perturbations were measured by recording 1H-15N HSQC spectra for E1 in the presence of 0-20 w/v% [BMIM][Cl], 0-5 w/v% [EMIM][Cl], or urea (8 M). 1H-15N HSQC spectra in the presence of low concentrations of ILs (0-5 w/v%), high concentration of ILs (20 w/v%), and urea were acquired using a Varian Inova 800 MHz spectrometer, Bruker Avance III 800 MHz spectrometer, and a Varian 900 MHz spectrometer, respectively, with a cryogenic probe under similar conditions as for the assignment experiments (~100-200 µM E1 with 5 v/v% D2O at 50 °C). Chemical shift perturbations were quantified via determination of the Euclidean distance between each peak and the corresponding peak in 2

the reference spectra (without IL) using the formula: ∆δED = (0.5 ∗ (∆δ1H + 0.127 ∗ 2

∆δ15N )). .29 For 1H-15N HSQC experiments with 20 w/v% [BMIM][Cl], data were acquired in 1.5 h increments for 24 h after addition of the IL to E1. For 1H-15N HSQC

7

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

experiments with urea, E1 was incubated with urea at 50 °C for 30 min prior to collection of the spectra.

Molecular Dynamics Simulations The structure of E1 simulated in classical molecular dynamics (MD) simulations was obtained via direct mutation of the 1ECE crystal structure to generate the Y245G variant of the catalytic domain of endocellulase Cel5A from Acidothermus cellulolyticus used in the experiments. Two trials of E1 in an aqueous solution of 20 w/v% [BMIM][Cl] were simulated with the GROMACS 5.1.2 MD engine.30 Packmol31 was used to generate cubic simulation boxes of ~8.8 nm in length with ~65,000 atoms. Based on our previous work,32,33 the IL was modeled using the general AMBER force field (GAFF)34 with point charges determined via the RESP35 calculation method and scaled down by 20% to improve the simulated dynamical properties of the IL.36 The TIP3P37 water model was used to represent water, and the Amber14SB force field38 was used to model both the E1 enzyme and sodium ions that were added to each system to achieve overall charge neutrality. Following a steepest descent energy minimization, the systems were simulated in the NVT ensemble at high temperature to eliminate any artifacts from the initial packing configurations. The enzyme and crystallographic waters were frozen in place, and the Bussi-Donadio-Parrinello thermostat39 (used in all subsequent simulations) was employed to maintain the temperature at 227 °C for 1 ns to randomize the solvent structure. The systems were then quenched in a second NVT simulation at 27 °C for 1 ns. A final 1 ns equilibration step was performed in the NPT ensemble using an isotropic pressure

8

ACS Paragon Plus Environment

Page 8 of 36

Page 9 of 36

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

coupling scheme with a time constant of 1 ps and the Berendsen barostat40 to maintain a system pressure of 1 bar; the constraints on the enzyme and crystallographic waters were removed at this point. Finally, production simulations were performed in the NPT ensemble at 50 °C for 500 ns, using the Parrinello-Rahman barostat.41 Bonds between hydrogen and heavy atoms were constrained using the LINCS algorithm42 to employ the 2 fs time step used in all simulations, and full periodic boundary conditions (PBC) were applied in all directions. Lennard-Jones and Coulombic interactions were explicitly calculated up until a cutoff value of 1.0 nm, and the particle mesh Ewald (PME)43 summation method was used to account for long-range electrostatics. To study the unfolding pathways of E1 in [BMIM][Cl], additional NPT simulations of E1 in both pure water and in aqueous 20% w/v% [BMIM][Cl] were performed at 225 °C; this has been demonstrated to be an effective way to accelerate the unfolding process without changing the underlying physics of the system (i.e., the unfolding pathways that would be observed at lower temperatures).44 It is necessary to employ high temperatures even for some miniature peptides in ILs that typically unfold beyond the timescales that are reasonably accessible with classical MD,45 but is especially necessary when simulating a large, thermophilic enzyme like E1. In each solvent, 10 independent simulations were initiated from the crystal structure and were carried out for 100 ns each. The PLUMED 2.0 simulation plugin46 was used to track the root mean squared deviation (RMSD) of the enzyme’s C  atoms from the crystal structure over time in each simulation. Visual Molecular Dynamics (VMD)47 was used for visual analysis of the simulation trajectories.

9

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

RESULTS AND DISCUSSION Biphasic inactivation of E1 by [BMIM][Cl] The effects of [BMIM][Cl] on E1 activity were assayed in aqueous solutions containing 0-20 w/v% [BMIM][Cl] at 65 °C. As shown in Figure 1A, activity decreased significantly as a function of increasing IL concentration, with ~50% of the initial enzyme activity lost upon addition of only 2.5 w/v% [BMIM][Cl]. This immediate inhibition was reversible, as evidence by activity assays performed immediately after dilution of the IL at low concentrations of [BMIM][Cl]. Incubation of E1 in the presence of higher IL concentrations (≥ 5 w/v%; 65 °C) over longer time periods led to irreversible loss of enzyme activity (Figure 1B). At the highest concentration of IL used (20 w/v%), the enzyme was nearly completely inactivated after 1 h incubation at 65 °C, whereas E1 retained 100% of its initial, reversibly inhibited activity after incubation with 2.5 w/v% [BMIM][Cl] for 24 h. At intermediate IL concentrations of 5 w/v% and 10 w/v% [BMIM][Cl], E1 retained approximately 67% and 15% of its initial activity, respectively, upon prolonged incubation (10 h). This slower loss of enzyme activity was not reversible, as E1 precipitated during incubation, which could not be reversed by removal of [BMIM][Cl] by dialysis. These findings indicate that E1 is inactivated by [BMIM][Cl] in a biphasic manner, in which enzyme activity is rapidly but reversibly inhibited by [BMIM][Cl] at all concentrations employed. At higher [BMIM][Cl] concentrations (≥ 5 w/v%; 65 °C), this initial inhibition is followed by slower and irreversible loss of activity. The extent of initial, reversible inhibition, and subsequent irreversible inhibition, are both dependent on [BMIM][Cl] concentration. These findings are consistent with earlier observations that

10

ACS Paragon Plus Environment

Page 10 of 36

Page 11 of 36

[BMIM][Cl] inhibits E1 in a concentration-dependent manner, although previous timedependent inactivation was not observed for E1 with imidazolium ILs in the same study.48 Earlier studies also showed that [BMIM][Cl] reversibly inhibits a cellulase from Trichoderma reesei at low IL concentrations (0-5 wt/wt%).11 Interestingly, the activities of GH11 xylanase enzymes from Trichoderma longibrachiatum and Dictyoglomus thermophilum were similarly found to be inhibited by ILs in a biphasic manner that is dependent on both IL concentration and incubation time,13,14 suggesting that cellulase and xylanase may be inhibited by [BMIM][Cl] via a similar mechanism.

120

Activity Relative to Buffer (%)

A

100

B Activity Relative to Initial Activity (%)

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

80 60 40 20 0

0

5 10 15 [BMIM][Cl] (w/v%)

20

10 15 20 Incubation Time (h)

25

120 100 80 60 40 20 0 0

5

Figure 1. (A) Immediate and (B) time-dependent effect of [BMIM][Cl] on the activity of E1 at 65 °C. In panel B, [BMIM][Cl] concentrations are 0.0 (buffer only, orange triangle), 2.5 (green diamond), 5.0 (purple square), 10.0 (blue circle), and 20.0 (red triangle) w/v%. Error bars, which in some cases are smaller than the symbols, were determined from the mean of two independent experiments. The lines connecting the symbols were added for clarity to follow the trends in the data.

11

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

Irreversible Deactivation is due to Protein Denaturation To understand the mechanistic origin of the apparent biphasic inactivation of E1 by [BMIM][Cl], we investigated the impact of [BMIM][Cl] on E1 structure. Of specific interest was to determine if the immediate and time-dependent effects of [BMIM][Cl] on E1 activity were related to IL-induced denaturation of E1. The immediate and timedependent effects of [BMIM][Cl] on E1 structure were determined via monitoring changes in the tertiary structure of E1 by CD. Such changes were monitored by measuring spectra of E1 in the near-UV wavelength region (260-320 nm) at 50 °C while varying both IL concentration and incubation time with the IL (0-23 h). Because the chloride anion and [BMIM] absorb strongly in the far-UV region, we were unable to also monitor changes in secondary structure in the presence of the IL. Analysis of near-UV spectra of E1 from CD showed that the loss of tertiary structure of E1 immediately upon addition of [BMIM][Cl] was negligible (Figure 2A). As evidence of this, the spectra of E1 in the presence of 1-5 w/v% [BMIM][Cl] was nearly identical to that of E1 in the absence of [BMIM][Cl] (i.e., in buffer alone). Although the spectra for E1 in the presence of [BMIM][Cl] were shifted slightly from that of E1 in the absence of [BMIM][Cl], this shift was presumably insignificant given that this shift was apparent over the entire wavelength range, and that the characteristic shape of the spectra in the presence and absence of the IL were virtually the same. Interestingly, at 5 w/v% [BMIM][Cl], minimal changes in the tertiary structure of E1 were also observed upon incubation in the IL at elevated temperature for 23 h (Figure 2B). This observation is consistent with the high degree of activity retention of E1 when incubated with 5 w/v% [BMIM][Cl] for 24 h in stability studies as described above.

12

ACS Paragon Plus Environment

Page 12 of 36

Page 13 of 36

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

Conversely, when incubated with 20 w/v% [BMIM][Cl], a dramatic loss in signal over time was observed, indicating the loss of tertiary structure of E1 during incubation (Figure 2C). In this case, the loss of tertiary structure over time was rapid with changes in the near-UV spectra of E1 becoming apparent within as little as 1.5 h. The rapid changes in tertiary structure were also consistent with the rapid inactivation of E1 in stability studies. Notably, the spectrum for E1 with 20 w/v% [BMIM][Cl] before 1.5 h appeared similar to that for E1 with 1-5 w/v% [BMIM][Cl] without incubation. As with the time-dependent effects at 5 w/v% IL, the rapid decrease in tertiary structure with 20 w/v% IL is consistent with the rapid inactivation of E1 in stability studies. The results of tertiary structure analysis suggest that the immediate inactivation of E1 upon addition of IL is not due to denaturation, but rather some other mechanism. Prior studies have suggested that imidazolium ILs, including [BMIM][Cl], may inactivate cellulase via binding in the cellulase binding tunnel and thus active site of cellulases. Of direct relevance to our work, Johnson and Snow49 reported that 1-ethyl-3methylimidazolium associated with the binding pocket of E1 in MD simulations. Li and co-workers50 similarly found in in silico studies that the butyl tail of 1-butyl-3methylimidazolium bound near the active site of CBHI from Trichoderma reesei. Furthermore, it was also shown that 1-ethyl-3-methylimidazolium bound to the active site of the hemicellulose degrading enzyme, GH11 xylanase,13 which is highly similar to that of cellulases. In light of these reports, we hypothesized that the immediate inactivation of E1 in the presence of [BMIM][Cl] was due to competitive inhibition of E1 via binding of the [BMIM] cation. Such a mechanism would explain the rapid loss of activity upon addition of the IL in the absence of gross structural changes without incubation. At the

13

ACS Paragon Plus Environment

The Journal of Physical Chemistry

same time, the results of our structural analysis by circular dichroism indicate that the slow loss of activity as a function of time may be attributed to IL-induced denaturation of E1, thereby explaining the biphasic nature of the inactivation of E1 as observed in activity and stability studies.

[θ] (deg cm2 dmol-1)

A

50 0 -50 -100 -150 -200 -250

B

50

[θ] (deg cm2 dmol-1)

-300 260

0

270

280 290 300 310 Wavelength (nm)

320

270

280 290 300 310 Wavelength (nm)

320

270

280 290 300 310 Wavelength (nm)

320

-50 -100 -150 -200 -250 -300 260

C

50

[θ] (deg cm2 dmol-1)

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 14 of 36

0 -50 -100 -150 -200 -250 260

Figure 2. (A) Immediate and (B) and (C) time-dependent effects of [BMIM][Cl] on the tertiary structure of E1 at 50 °C. Panel A shows an overlay of the near-UV circular dichroism spectra of E1 immediately after addition of the various [BMIM][Cl] solutions. Overlaid spectra in panel A depict [BMIM][Cl] of 0 (tan), 1.0 (burgundy), 1.5 (green), 2 (red), 3 (orange), 4 (blue), and 5 w/v% (violet). Panels B and C show overlays of the

14

ACS Paragon Plus Environment

Page 15 of 36

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

near-UV spectra of E1 over time after the addition of 5 w/v% and 20 w/v% [BMIM][Cl], respectively. Overlaid spectra in panel B and C depict incubation times of 0 (red), 1.5 (burgundy), 3 (orange), 5 (green), 7 (blue), and 23 h (black).

Structural Analysis of the Impact of [BMIM][Cl] on E1 by NMR To further investigate the structural basis for the inactivation of E1 by [BMIM][Cl], IL-induced changes in the structure of E1 were probed by NMR. The 1H15

N HSQC spectrum of E1 in the absence of ILs shows well-dispersed peaks, indicating a

compact, folded state (Figure 3A). Using standard triple resonance techniques we were able to assign the backbone resonances for 247 out of 340 (73%) non-proline residues, including many of the residues in the active site region and substrate-binding tunnel (Figure 3C,D). The majority of peaks for which assignments could not be obtained corresponded to residues within highly flexible loop regions, based on the crystal structure of E1 (Figure 3C). A comparison of the secondary structure of E1 from chemical shift indexing28 and from the crystal structure of E1 is shown in Figure S1. The close agreement suggests the crystal structure is maintained under our solution conditions. Some observable NH signals could not be unambiguously assigned due to signal overlap and/or lack of redundant triple resonance connectivities; for these signals, potential residue identities, based on

13

C chemical shifts and non-redundant

connectivities, are reported in Table S1.

15

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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. Backbone chemical shift assignment of E1. (A) and (B) 1H-15N HSQC spectrum of E1 annotated with assignments. “W” represents unassigned tryptophan sidechain amides and dotted lines connect unassigned glutamine and asparagine side-chain amides associated with the same residue. Panel A is a truncated version of the entire spectra collected, and omits more unassigned tryptophan side-chain amides. (B) For added clarity, an enlarged spectrum of the crowded region in panel (A) is shown. (C) The crystal structure of E1 (PDB accession number 1ECE), showing assigned (purple) and not assigned (red) residues. (D) The E1 active site, colored as in panel C. The Y245G mutation is denoted by an asterisk. Cellotetraose is depicted by green lines. To elucidate the immediate inhibitory effect of [BMIM][Cl] on E1, 1H-15N HSQC spectra were obtained as a function of increasing IL concentration. At low [BMIM][Cl]

16

ACS Paragon Plus Environment

Page 16 of 36

Page 17 of 36

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

concentrations (0-5 w/v%) a limited number of chemical shift perturbations are seen, indicating local effects from specific interactions, but no global changes to the overall structure (Figures 4A and B). These chemical shift perturbations were reversible by dilution of [BMIM][Cl] (Figure S2) and are not associated with sample pH changes upon IL addition (Figure S3), suggesting a direct interaction between the IL and E1. As in our previous work,51 the perturbations were presumably the result of direct ion interactions between either the cation or anion of the IL and the perturbed residue. Interestingly, the two residues that experienced the largest chemical shift perturbation, W213 and E282, are located in the active site.52 W213, which forms interactions with cellulose when in complex with E1, is situated near the acid/base catalytic residue E162. E282 is involved in the catalytic mechanism and acts as a nucleophile by attacking the anomeric carbon of the Glc3 ring during cellulose hydrolysis. The perturbation of the peak for W213 was particularly large (∆δ1H ~0.79 ppm in 5 w/v% [BMIM][Cl]), suggesting that this residue experienced a dramatic alteration in local chemical environment. Additionally, the magnitude of the perturbations of these signals appeared to plateau (i.e., reach a maximum value) at the highest concentration of IL unlike for other sites where the magnitude of the perturbations continued to increase even at the highest concentrations of IL. Residues that exhibited chemical shift perturbations in the presence of 5 w/v% [BMIM][Cl] are mapped on the crystal structure of the enzyme in Figures 4C and D. The significant, saturatable, and reversible perturbation of W213 and E282 signals by [BMIM][Cl] provides compelling evidence that reversible inactivation of E1 by [BMIM][Cl] is due to competitive binding to the E1 active site. In the case of W213, the

17

ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

imidazolium cation may form a cation-π interaction with the aromatic ring of the tryptophan through the permanent positive charge of the cation. Such an interaction would be consistent with prior crystallographic evidence that the interaction of [BMIM] with tryptophan may be stabilized by cation-π interactions.53 In the case of E282, the interaction of the IL with this site likely also involves the positively-charged cation and is electrostatic in nature. Notably, although other exposed tryptophans, including W42 and W319, are present near the active site, these residues, which have a similar residue environment as W213, were not perturbed. The lack of perturbation of these residues was surprising given that [BMIM] may presumably interact with these residues in a similar manner to W213. To confirm that the inhibition of E1 by [BMIM][Cl] was competitive, the impact of the IL on the Michaelis-Menten parameters of E1 was analyzed. Lineweaver-Burk analysis of this impact showed that the impact of [BMIM][Cl] on the apparent Vmax of E1 was negligible while the value of apparent Km increased with increasing IL concentration, thereby confirming the inhibition was competitive (Figure S4). Additionally, analogous 1H-15N HSQC experiments with [EMIM][Cl], which is chemically similar to [BMIM][Cl], also found the same residues were perturbed (Figure S5). This suggests that other imidazolium ILs may similarly inhibit cellulases through a common mechanism involving competitive inhibition.

18

ACS Paragon Plus Environment

Page 18 of 36

Page 19 of 36

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

Figure 4. NMR chemical shift perturbations of E1 with low concentrations of [BMIM][Cl]. (A) and (B) 1H-15N HSQC overlays of E1 with increasing titration concentrations of [BMIM][Cl] (light blue, 0 w/v%; pink, 0.5 w/v%; red 1.0 w/v%; orange 1.5 w/v%; green, 2.0 w/v%; blue, 3.0 w/v%; purple, 4.0 w/v%; black, 5.0 w/v%). Residues are labeled and arrows indicate the direction of significantly shifted residues. An asterisk (*) indicates that a residue is either in the active site or substrate-binding tunnel. (C) and (D) Perturbation heat maps of E1 when exposed to 5 w/v% [BMIM][Cl] compared against buffer alone. Tan represents unassigned residues, whereas other colors map the weighted chemical shift (∆δED ) with standard deviations (σ) from the mean chemical shift of