Previous studies have shown that dockerin-bearing enzymes in solution are multi-modular objects with substantial flexibility of the linker that separates the dockerin from the other modules, notably the catalytic domain [19,20,22,23]. Significantly, no measurable intermolecular interactions have been revealed in any of the studied cellulosomal enzymes [26]. This is also the case for the solution structure of dockerin-bearing exocellulase Cel48S-t and endocellulase Cel8A-b in our study (Figures 2 and 3).
Interestingly, the processive endoglucanase Cel9A-r, that contains a CBM3c module in addition to the catalytic domain and the dockerin, does not display the same features. As indicated by the smaller Dmax and Rg than expected (Table 1), this multi-modular enzyme is much more compact and does not appear to reach very extended conformations in solution, in stark contrast to the other two enzymes. Crystal structures of homologous Cel9 enzymes devoid of their dockerins have highlighted that for this type of enzymes the adjacent CBM3c is tightly tethered to the catalytic domain, with essentially no flexibility in their linker [80–83]. Nevertheless, flexibility would be expected for the linker between the CBM and the dockerin. This is not what we observe for the solution structure of Cel9A-r (Figure 3c); here, the linker seems to be pleated against the CBM. It could thus be speculated that the hydrophobic character of the substrate-binding surface of the CBM3c module might be concealed by the linker residues owing to unspecific interactions, such as those observed in ‘fuzzy complexes’ of intrinsically disordered proteins [84–86]. Indeed, pleating of linkers upon increasing the molecular mass of these enzymes has previously been documented for bi-modular enzymes composed of a catalytic domain and a dockerin in complex with their cognate cohesin [20].
Notably, CBM3c- containing GH9 processive cellulases are recurrent and important enzymes in cellulosomal complexes [77] that might play a key role in further interaction of the overall complex with the insoluble substrate. As such, they are generally present in cellulosomal complexes in higher abundance than other enzymes [87]. In addition, a molecular modeling study involving the self-assembly of the cellulosome enzyme complex [88] has revealed that the binding mechanism of enzymes is dependent on mass and flexibility: larger, multimodular and flexible enzymes (a GH9 homolog in that particular study) exhibit increased binding propensities, compared to smaller quickly diffusing enzymes, thus physically controlling the stoichiometry of integration. Consequently, the more compact form of the Cel9A-r observed here might be a minor state, artificially stabilized by the experimental conditions that lead to the pleating of the linker to cover the exposed hydrophobic surface of the CBM3c, and this conformation might be released upon contact with scaffoldins.
Genome mining of cellulosome-producing bacteria has revealed a large variety of cellulosomal systems [89] that potentially are linked to the natural habitats of the microorganisms [90]. The encountered diversity raises the question whether the composition and spatial organization follows a general rule, or if the diversity also reflects the need to vary the connected biophysical properties, to adapt to specific habitats or substrate sources. In this context, it remains crucial to understand the link between the architecture of cellulosomal systems and their efficiency remains of growing interest. SAXS measurements on several scaffoldins [20,22,23,28,29], most of them being chimeric constructions, revealed differences in flexible behavior, depending on where the adjacent cohesins are situated within the sequence, with N-terminal cohesins and linkers being more flexible than central ones [28]. In our present study, we expand the SAXS studies of these objects in solution to include three original scaffoldins, which are ScaA, ScaH and ScaK, found in the human gut bacterium R. champanellensis [80]. This bacterium is to date the only human colonic bacterium so far reported to efficiently degrade recalcitrant plant polysaccharides, such as crystalline cellulose and xylan [91]. Interestingly, while ScaA can be considered one of the smallest “classical” scaffoldins, consisting of 2 cohesins with an X domain and a dockerin, the other two scaffoldin proteins, ScaH and ScaK, contain catalytic modules within their primary sequences [80]. Since no structural homologues of these modules were available, molecular modeling was not possible for these macromolecules. Nevertheless, Rg and Dmax values (Table 1), as well as the P(r) function (Figure 4c), derived from the scattering curves of these proteins in solution, are consistent with rather extended, flexible and multimodular components. Moreover, the Kratky-plots reveal a larger globular object, combined with substantial disordered regions (additional Figure S2). These results are in agreement with the suggestion that these scaffoldins reflect a naturally occurring expansion or diversification of strategies for cohesin–dockerin interactions [92]. These architectural data need now to be completed by single molecule force spectroscopy experiments to demonstrate possible implications of these variations on the complex mechanostability of these interacting proteins [93]. In particular, more work is needed to assess how the balance between compaction and flexibility may be fine-tuned in response to the nature and recalcitrance of the substrate that is targeted and the environment of action. In this context, the presence of unconventional scafoldins, containing peptidases and oxidative enzymes, have been found in C. alkalicellulosi, which appear to be associated with both cell-associated and cell-free systems, and might be linked to their occurrence in alkaline soda lake ecosystem [73].
As a next step, the study of artificial designer cellulosomes offers a valuable tool for unraveling synergy-connected architectural features of the complexed cellulosomal enzymes, and may produce to guidelines for design of more efficient and more stable complexes. In the light of the detailed biochemical study of various designer cellulosomes and their efficiency [14] that demonstrated the outstanding performance of Scaf20L in complex with three enzymes, we have explored the overall structural arrangement in solution of this particular cellulosomal complex using the dissect and build strategy with SAXS. Our results on Scaf20L alone, in complex with one single enzyme and in complex with three different enzymes again highlight that ‘loading’ the scaffoldins with enzymes influences the flexibility of the linker regions; the more the complex is loaded, the more compact the overall spatial arrangement becomes. The data clearly show that multiple conformers exist in solution, varying between compact forms with pleated linkers and extended conformations, in which the enzymes point away from each other. This spatial arrangement and variability might lay the basis for the mechanics of their plastic action adapted to heterologous catalysis, where the extended conformers are those that stabilize interaction with the (solid) substrate, and the more compact forms maintain the integrity of the complexes in the free and substrate-unbound state, as has been previously proposed [20,22]. Our findings on the biophysical values of Rg and Dmax for CipA and its enzyme-complex support this hypothesis. They also confirm the existence of galleries of “loose cellulosome” conformations (additional Figure S3) that have been depicted way back in 1987 by Mayer et al. [27]. The next step would be to further probe the spatial arrangements of these large multi-enzyme complex structures in interaction with a natural, complex substrate, from meso to atomistic scale.