Turning a potent family-9 free cellulase into an operational cellulosomal component and vice versa.

Ruminiclostridium cellulolyticum and Lachnoclostridium phytofermentans are cellulolytic clostridia either producing extracellular multienzymatic complexes termed cellulosomes or secreting free cellulases respectively. In the free state, the cellulase Cel9A secreted by L. phytofermentans is much more active on crystalline cellulose than any cellulosomal family-9 enzyme produced by R. cellulolyticum. Nevertheless, the incorporation of Cel9A in vitro in hybrid cellulosomes was formerly shown to generate artificial complexes with altered activity, whereas its incorporation in vivo in native R. cellulolyticum cellulosomes resulted in a strain displaying a weakened cellulolytic phenotype. In this study, we investigated why Cel9A is so potent in the free state but functions poorly as a cellulosomal component, in contrast to the most similar enzyme synthesized by R. cellulolyticum, Cel9G, weakly active in the free state but whose activity on crystalline cellulose is drastically increased in cellulosomes. We show that the removal of the C-terminal moiety of Cel9A encompassing the two X2 modules and the family-3b carbohydrate binding module (CBM3b), reduces its activity on crystalline cellulose. Grafting a dockerin module further diminishes the activity, but this truncated cellulosomal form of Cel9A displays important synergies in hybrid cellulosomes with the pivotal family-48 cellulosomal enzyme of R. cellulolyticum. The exact inverse approach was applied to the cellulosomal Cel9G. Grafting the two X2 modules and the CBM3b of Cel9A to Cel9G strongly increases its activity on crystalline cellulose, to reach Cel9A activity levels. Altogether these data emphasize the specific features required to generate an efficient free or cellulosomal family-9 cellulase.