FoldX stability analysis — candidate mutations ranked

Contributors: Chester

FoldX run

Ran FoldX on all candidate single mutations from the HotSpot Wizard list (22 Jun) against the GH10 PDB. 5 repair runs per mutation (testing different residue conformations — runs limited by compute, not methodology).

Output files: PdbList_GH10_Repair.fxout and Raw_GH10_Repair.fxout.

[Attachments: Raw_GH10_Repair.fxout, PdbList_GH10_Repair.fxout]

Stability metric

ΔΔG = WT_entropy − Repair_entropy. Negative ΔΔG = more stable than WT; positive = less stable (more disorder).

Top stabilising mutations (ranked by ΔΔG)

(Full ranked table of 25 mutations available in Raw_GH10_Repair.fxout. Mutant PDB files follow the pattern GH10_Repair_[PDB#]_0.pdb through _4.pdb for each mutation’s 5 repair runs.)

These mutations are candidates for targeting in the epPCR library or for site-directed mutagenesis, prioritising those with the most negative ΔΔG (greatest predicted stabilisation).

Docking re-screen of stabilised mutants

Converted the stabilised-mutant PDBs to .pdbqt and re-ran AutoDock Vina — this time scripted from the terminal rather than docked manually one at a time, to process the full candidate set. Output: a sorted summary of predicted binding free energy (more negative = better binding) per mutation.

[Attachment: vina_summary_sorted.csv]

Result: 8 of the stability-screened single-residue mutants also showed improved cellobiose binding scores relative to WT (−6.6 kcal/mol), with the best two at −6.7 and −6.8 kcal/mol.

The improvement is small. Even if wet-lab work confirms these single mutants bind better, the activity increase is likely minor and may be difficult to distinguish from noise in an assay.

Next step under consideration: combining multiple stabilising/binding-improving mutations, since single substitutions alone are unlikely to produce a wet-lab-detectable change.

GH1 glucose tolerance — literature review

Background reading on a known limitation of most β-glucosidases: product inhibition by glucose, which caps activity at the high glucose concentrations industrial cellulose degradation would produce. Some GH1 β-glucosidases specifically (relevant here since GH1 is our fusion track’s target domain, and structurally simpler than other GH families) have engineered or natural glucose-tolerant variants. Proposed mechanisms:

• Decoy glucose-binding sites that reduce the fraction of glucose occupying the true active site.
• Glucose disrupting the active-site water matrix, perturbing hydrogen-bond formation and substrate binding/stability.

Examples from the literature:

• Thr228, Gln301, Phe302 mutations increased glucose tolerance via transglycosylation of glucose into usable intermediates (acting as decoy binding sites).
• “Gate-keeper” mutants L167W and P172L (tested individually, not combined) increased glucose tolerance roughly 3-fold (0.1–0.25 M) and maintained activity at 1.0 M glucose, likely by narrowing the substrate channel. Both also raised thermostability (optimum temperature 40 °C → 50 °C, lower Km).

Further reading:

• Notenboom et al. — crystallography/mutation study of Cex cellulose/xylan specificity (already cited above).
• “An engineered GH1 β-glucosidase displays enhanced glucose tolerance and increased sugar release from lignocellulosic materials,” Scientific Reports.
• “Structural basis for glucose tolerance in GH1 β-glucosidases,” IUCr.
• “A mechanism of glucose tolerance and stimulation of GH1 β-glucosidases,” Scientific Reports.