Enzyme activity is traditionally studied with a major focus on the active site and how its structure contributes to the enzyme’s function. However, dynamics has more recently been implicated in many aspects of enzyme function including binding and sequestration of the substrate from solvent (Joseph et al., 1990, Knowles, 1991), the catalytic step (Eisenmesser et al., 2005, Clarkson et al., 2009), and product release (Cui et al., 2019). For many enzymes such as Xylanase A (XylA), there are indications that dynamics plays an important role and efforts have been made to influence that through mutation (Bourgois et al., 2007, Kazuyo et al., 2014, Pollet et al., 2009). However, there remain many unanswered questions, warranting further exploration into how mutations affect the structure and/or enzymatic activity.

GH11 xylanases are a class of secreted xylanases, which endolytically cleave β-1,4-xylosidic bonds in xylan polymers (Paës et al., 2012). They are prevalent across many different organisms in eukaryota and prokaryota. All GH11 xylanases show high specificity for the xylan polymer, its native substrate, and high sequence conservation (Sapag et al., 2002). GH11 xylanases have a beta-jelly roll structure with a loop termed the thumb region covering the active site (Wakarchuk et al., 1994). The active site comprises parts of the thumb, palm, and fingers regions (Fig. 1A). There are two catalytic residues: E78, which is a nucleophile that starts in a deprotonated state at the beginning of the catalytic cycle and E172, which is the acid/base hydrolyzing residue and starts the reaction in a protonated state (Wakarchuk et al., 1994, Joshi et al., 2001).

The overall catalytic mechanism is an acid/base hydrolysis in two steps: enzyme glycosylation followed by deglycosylation. E78 initiates the hydrolysis mechanism with a nucleophilic attack to carbon-1 in the 1,4-β-xylosidic bond. The oxygen of carbon-1 in the xylosidic linkage then abstracts the proton from E172 (Wakarchuk et al., 1994). This results in the glycosylated intermediate of E78 covalently bonded to the substrate and E172 in the deprotonated state (Wakarchuk et al., 1994, McIntosh et al., 1996). Deglycosylation is then initiated by the now deprotonated, negatively charged carboxylate of E172 activating the oxygen on a coordinated water molecule, through hydrogen bonding or abstraction, for nucleophilic attack on the xylosidic carbon-1 attached to E78 (McIntosh et al., 1996). Given the acid/base nature of the reaction, the pKa values for the catalytic residues are critical for their ability to perform the mechanism (Wakarchuk et al., 1994, Joshi et al., 2001, McIntosh et al., 1996, Joshi et al., 2000). During catalysis, E172 undergoes a pKa shift from 6.7 at the start of glycosylation to 4.2 for deglycosylation, while E78 remains at a pKa of 4.6 (McIntosh et al., 1996). It is also known that various active site residues, including N35 and Y80 help ensure the pKa of the two side chains are appropriate for performing the full mechanism (Joshi et al., 2000).

Xylanase A has many regions implicated in the activity of the enzyme. The thumb region, residues Y113-T124, is thought to aid in substrate acquisition, positioning, catalytic mechanism, and product ejection, potentially relying on mobility to allow for this behavior (Mhlongo et al., 2015, Vieira and Ward, 2012, Vieira et al., 2009, Paës et al., 2007). There is high sequence conservation, particularly at the tip of the thumb in the PSIXG sequence motif, spanning residues 116–120. While deletion of the thumb region renders the enzyme inactive, it does not affect binding of the native xylan substrate but does allow for other non-native polymers to bind (Paës et al., 2012, Sapag et al., 2002, Paës et al., 2007). One study also found that the P116G mutation enhanced millisecond dynamics in the thumb region and was associated with the emergence of a new transglycosylation reaction rather than just the native glycosylation/deglycosylation reaction (Marneth et al., 2021). Molecular dynamics simulations run at elevated temperatures showed an increase in the rate of thumb opening and closing (Vieira and Ward, 2012, Vieira et al., 2009, Marneth et al., 2021, Murakami et al., 2005, Muilu et al., 1998). All of these suggest a possible role of dynamics in catalysis, specifically in thumb region mobility impacting reaction turnover (see Table 2 and Table 3).

From a structural dynamics point of view, the D11F/R122D (Pollet et al., 2009) double mutation is among the most interesting. It was originally designed to allow resistance to a Xylanase A competitive inhibitor, Triticum aestivum xylanase inhibitor (TAXI) (Sørensen and Sibbesen, 2006, Rasmussen et al., 2010). In the original design process they found that D11F was resistant to TAXI (Sørensen and Sibbesen, 2006) but decreased enzymatic activity. R122D, however, conferred no resistance but had an increase in activity from wild-type (WT) (Sørensen and Sibbesen, 2006). Of the three D11F/R122D crystal structures solved to date, one (PDB 3EXU) has been crystallized in a thus far unique thumb open conformation. All other Xylanase A crystal structures show the thumb region in a closed position. The other two crystal structures with the double mutation (PDB 2B46 and 2B45) show the thumb closed but have ligands in the active site. This has led to some debate over the thumb positioning in this structure and whether this is due to crystal contacts or, as Pollet et al. suggests (Pollet et al., 2009), a change in the thumb opening thermodynamics of Xylanase A, with the ligands contributing the closed conformations observed in 2B46 and 2B45. Only chain A in the 3EXU structure has a resolvable thumb region, which makes crystal contacts with chain B (Pollet et al., 2009). The chain B thumb region is disordered with no resolvable electron density, which the authors suggest is caused by the double mutation (Pollet et al., 2009).

Molecular dynamics simulations have also given evidence that this is not just a consequence of crystal lattice contacts stabilizing the open form but mostly a consequence of R122D allowing for variation in intramolecular contacts to help stabilize the open state (Pollet et al., 2009, Vieira and Ward, 2012). Vieira et al. proposed that the R122D mutation, located on the thumb, allows for a more extensive intramolecular hydrogen bond network in the open structure, with D122 hydrogen bonding to Y113 and S117, and R112 being positioned such that instead of hydrogen bonding with solvent alone as in WT, it hydrogen bonds with E78 and Y80 (Vieira and Ward, 2012). The D11F mutation is in the fingers region while R122D is in the thumb region and does not directly interact with the substrate in either the WT or the double mutant. While D11F/R122D appears to allow for increased access of the open state when not bound, it retains activity, albeit reduced, on the native polymer, xylan (Bourgois et al., 2007).

The mechanism behind how the double mutation changes the protein is still not well understood (Bourgois et al., 2007, Pollet et al., 2009). Here we use a combination of kinetics measurements, NMR spectroscopy, and computational simulation to gain further insight into how the D11F/R122D mutation affects the structure and function of xylanase A. Analysis of available crystal structures suggests that structural heterogeneity in the thumb and fingers region is uncorrelated with being ligand bound or free. Our kinetics analysis shows the double mutation decreases catalytic activity with the native polymer while unexpectedly increasing activity on the short substrate ONPX2. NMR experiments reveal structural perturbations in both the thumb and fingers regions and increased slow-timescale dynamics. Finally, alchemical free energy simulations suggest that the double mutation indeed destabilizes the closed state of the thumb region.