Machine learning reveals sequence-function relationships in family 7 glycoside hydrolases

Abstract

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Virtual Bioprospecting of Interfacial Enzymes: Relating Sequence and Kinetics

“…One of the modular CBHs in group C is Pc Cel7D (labeled in Figure a), which is known to have a relatively open substrate-binding cleft, because of truncation of some peripheral loops . This supports the idea that the peripheral loops, which are severely truncated in the GH7 EGs (see Figure b), can tune the substrate-binding strength of the CD, as demonstrated in recent studies. , …”

“…The average binding energy was shifted by ∼8 kJ/mol in enzymes with a CBM, but there was a large overlap of distribution functions for enzymes with and without CBM in Figure c, and this shows that stronger ligand interaction in the CD can at least in part compensate for the lack of a separate CBM. This observation is consistent with a recent report by Gado and co-workers …”

“…We note in passing that this grouping of CBM possession occurred although the tree was based solely on CD sequences. This aligns well with a recent observation that the presence of a CBM can reliably be predicted from the CD sequence alone …”

“…This observation is consistent with a recent report by Gado and co-workers. 35 Prediction of Kinetic Parameters. To illustrate the relationship between enzyme sequence and kinetics, we combined a maximum-likelihood tree for the CDs of fungal GH7, and the binding data from Figure 2.…”

“…This aligns well with a recent observation that the presence of a CBM can reliably be predicted from the CD sequence alone. 35 We partitioned the tree into arbitrary groups with related sequences and similar binding energy as indicated in the figure. Groups A and B were solely composed of EGs and showed relatively weak substrate binding.…”