Improved thermostability of Clostridium thermocellum endoglucanase Cel8A by using consensus-guided mutagenesis

Abstract

The use of thermostable cellulases is advantageous for the breakdown of lignocellulosic biomass toward the commercial production of biofuels. Previously, we have demonstrated the engineering of an enhanced thermostable family 8 cellulosomal endoglucanase (EC 3.2.1.4), Cel8A, from Clostridium thermocellum, using random error-prone PCR and a combination of three beneficial mutations, dominated by an intriguing serine-to-glycine substitution (M. Anbar, R. Lamed, E. A. Bayer, ChemCatChem 2:997-1003, 2010). In the present study, we used a bioinformatics-based approach involving sequence alignment of homologous family 8 glycoside hydrolases to create a library of consensus mutations in which residues of the catalytic module are replaced at specific positions with the most prevalent amino acids in the family. One of the mutants (G283P) displayed a higher thermal stability than the wild-type enzyme. Introducing this mutation into the previously engineered Cel8A triple mutant resulted in an optimized enzyme, increasing the half-life of activity by 14-fold at 85°C. Remarkably, no loss of catalytic activity was observed compared to that of the wild-type endoglucanase. The structural changes were simulated by molecular dynamics analysis, and specific regions were identified that contributed to the observed thermostability. Intriguingly, most of the proteins used for sequence alignment in determining the consensus residues were derived from mesophilic bacteria, with optimal temperatures well below that of C. thermocellum Cel8A.

Fig 1

Protein sequence of Cel8A from C. thermocellum. Consensus mutations are indicated (in bold).

Fig 2

Distribution of consensus mutations in the shuffled library. Several isolates contained between 1 and 2 additional missense mutations that were not included in the analysis.

Fig 3

Frequency of mutations in thermostable mutants. Ten of the most stable mutants were sequenced and mapped.

Fig 4

Kinetics of thermal inactivation of Cel8A variants at 85°C. The residual activities were measured at different time points. Shown are Cel8A mutants G283P, TM, and QM (see Table 3). Wild-type Cel8A (WT) served as a control. The activity of unheated enzymes was taken as 100%. Each point represents the mean of duplicate determinations.

Fig 5

Specific activities of wild-type Cel8A (WT) and thermostable mutants on CMC and PASC. Enzymes were incubated with 0.5% (wt/vol) solutions at 65°C for 1 h.

Fig 6

The positions of the residues that were replaced in the QM in the Cel8A structure (PDB code 1CEM) are shown.

Fig 7

Contour plots of RMSD (Angstrom) values for the Cα atom of residues (33 to 395) during simulation time (ns) for the wild type (WT) (A) and the quadruple mutant (QM) (B). Graphs are grayscaled from white (0) to black (20) with increments of 2. The regions between amino acids 261 and 278 and between 311 to 333 are labeled. Significant destabilizations of these loops are observed after a 1-ns simulation time for the WT structure at 450 K (dark-shaded regions). On the other hand, the cumulative effect of the given mutations for the QM structure significantly enhances the stabilization of these loops (lighter shades of gray) at 450 K.

Fig 8

Sampling of loop conformations from MD trajectory. Cartoon representations of wild-type (WT) and quadruple-mutant (QM) structures from MD trajectory are superimposed on the starting conformation (colored blue) at 450 K. To demonstrate the increased rigidity of the loop residues 261 to 278 (A) and 311 to 333 (B) for QM relative to WT, these residues are colored according to time scale (blue for 0 ns, white for 2 ns, and red for 4 ns), and representative snapshots from the trajectory (every 80 ps) are superimposed.