Discussion
The vast majority of microbial cellulases and other cellulose- degrading enzymes such as xylanases are organized into a modular format which can be divided into catalytic and non-catalytic domains. One or a few carbohydrate - binding motifs (CBMs), responsible for binding to different substrates, are associated with the catalytic domain (CD) of many microbial cellulases. Each cellulase possesses its own unique assortment of CBMs. CBMs are often separated from other cellulase modules by linker molecules, which are believed to assist in the process of targeting the CD to different types of substrate (Boraston et al., 1998).
Cel9B of T. fusca follows this modular organization by including both CBM2 and CBM4, which are separated from the CD by a fibronectin-like domain (Fn3) and an Eset domain, respectively. The arrangement of the catalytic domain followed by a CBM2 that is found in Cel9B is also common in a number of enzymes produced by bacteria as well as other organisms, suggesting a common gene transfer event may have taken place in the distant past (Gao et al., 2010, Li et al., 2009).
Experiments carried out in this paper using truncation mutants and different substrates have indicated that mutants keeping the CBM2 domain bind substrates at higher yield that the other mutants, and increase cellulase activity against both crystalline and amorphous cellulose. This was clearly demonstrated using wild type Cel9B as well as truncation mutant M2, which lacks the CBM4. These results agree with for the work performed for other cellulases which exhibit higher levels of activity in the presence of CBM2, including Pseudomonas ,Cellulomonas fini , and Cellvibrio japonicus (Boraston et al., 1998, Tomme et al., 1996, Gilbert et al., 2013, Attia et al, 2016).   CBM2 is one of the most prevalent carbohydrate binding modules and is over- represented in thermophilic microorganisms. It is possible that under thermophilic conditions, CBM2 binding to cellulose may be more stable than those of other CBMs (Moser et al., 2008). Xylan-specific binding of CBM2 has been demonstrated in the past by Simpson et al., (2000). The fact that CBM2 can bind to both amorphous as well as crystalline cellulose could imply that there are differences in CBM2 function on these vastly different substrates. It has been suggested that CBM2s may enhance cellulose activity through its binding mechanism, either by bringing the catalytic domain proximal to the substrate itself or by disrupting and displacing recalcitrant substrate so that more of it is made to be accessible to the catalytic domain (Knowles et al., 1988; Din et al., 1991, Tomme et al., 1996).
Truncation mutant ΔFn3, which lacks an intact Fn3 domain, displayed reduced substrate binding and cellulase activity. It is possible that this protein is misfolded. During the production strain grew slowly (approximately 8 hours to reach OD 0.6 rather than 3-4 hours for the others) and the protein bound poorly to the nickel column. Unlike the other truncation mutant proteins, all of Fn3 came out in the flowthrough and did not bind to the nickel column, as determined by activity assay (Table 2). Brunecky et al., 2012, suggested that the Fn3 domain aids in the thermostability of the protein and may in fact function as a spacer between other functional domains. It is possible, therefore, that the Fn3 domain may have some indirect stabilizing role in the protein that was disrupted upon deletion, or perhaps it aids in the folding of the native protein structure. Fn3 domains have also been thought to play roles in regulating CBM interactions with substrates (Kataeva et al., 2002, Cadena et al., 2010). Although removal of the Fn3 domain in the ΔFn3 mutant appears to have impeded substrate binding, as shown in Table 3, it is more likely that this is really the result of improper protein folding, as truncation mutant Fn3 (containing solely the catalytic domain and the Fn3 domain) does not bind to substrate. Other research groups have shown that Fn3 domain does not affect enzyme binding, rather it can disrupt the surface of substrate so that single cellulase chains are directed to the catalytic site, and thus play a role in regulating catalysis on specific substrates (Kataeva et al., 2002, Guillen et al., 2010).
Truncation mutants M1 and ΔE9B, containing CBM4, exhibited pronounced activity on xyloglucan and BMCC that was missing from the other truncation mutants. Truncation mutant ΔE9B lacks the Eset domain, an Ig-like domain that may play a role in enzyme folding (Kataeva et al., 2005). This Eset domain is retained in the M1 mutant (Figure 1). CBM4 is known to bind to crystalline cellulose as well as amorphous cellulose (Alahuhta et al., 2010). In this study, truncation mutant M1 exhibited binding to BMCC, albeit at almost half of the yield of the M2 mutant that contains the CBM2 domain. Truncation mutant ΔE9B was not able to bind strongly to BMCC (5% binding yield) yet exhibited activity on this substrate. Furthermore, both M1 and ΔE9B showed no ability to bind to xyloglucan (data not shown) yet displayed significant activity on this substrate. These results bring into question the role of CBM4 in Cel9B. The relatively high activity of these two mutants on xyloglucan (with respect to the wild type) and their poor ability to bind to substrates would suggest that CBM4 plays a role that differs in general from CBM2. Blake et al., (2006) have demonstrated that CBM4 binding to cellulose chains is limited but requires access to the entire width of a cellulose chain.  Indeed, this different binding of CBM4 may be necessary for Cel9B to efficiently catalyze areas of the cell wall that cannot be targeted effectively by CBM2. The poor binding ability to various substrates may indicate that CBM4 plays a role at a later stage of cell wall degradation, where substrates that are available to interact with CBM4 differ from those that are found early on in the degradation process.
Truncation mutants cd9B, E9B, and Fn3 contain the catalytic domain and no CBMs. These mutants exhibited very low yield binding (below 10%) to any of the substrates tested in this study. While all were able to display catalytic activity, cd9B displayed the most, suggesting that retention of the Eset domain and fibronectin-like domain (in E9B and Fn3 mutants, respectively) may have a deleterious effect on the folding of cd9B that impacts its catalytic activity.
WT, M1, and M2 have the highest and similar activities on PASC while the other mutants have activities which are similar to each other, but somewhat lower than WT. This shows that either CBM can improve activity on this substrate. PASC is an amorphous insoluble substrate with much higher surface area than BMCC. On the other hand, the activity results on BMCC show that deleting CBM4 (M1 and ΔE9B) causes a large increase in activity for M1 and a small increase for ΔE9B over time with respect of the wild type. This suggests that for Cel9B, CBM4 is not important for crystalline cellulose activity. Furthermore, deleting the Eset domain decreases the activity of M1 for reasons that would require further research.
Since CMC is a soluble substrate, removing the CBMs would not be expected to affect this activity, and the removal of CBMs presents increased catalytic activity on CMC for several other cellulases. The activities on PASC and xylan are very similar to the activities on CMC and this may be because both are partially soluble. In addition, these two substrates are not as highly ordered as cellulose, so that binding domains may not be required for their hydrolysis.