Here we report the expression of an engineered CipA* under the control of a novel inducible promoter in T. saccharolyticum which allowed for the assembly of active cellulosomes when co-cultured with a cipA deletion strain of C. thermocellum.
The wild-type cipA gene has multiple sections with essentially identical DNA sequences, corresponding to the 9 type-I dockerin regions . These repeated sequences can be problematic by complicating sequence verification via routine sequencing technology and could also lead to unwanted partial gene deletion via homologous recombination. Many strains of E. coli used to heterologously express cellulosomal proteins for biochemical studies are recA- [24, 27, 38]. However, studies such as this one which seek to integrate cellulase genes into the chromosome via native host machinery must use recA+ strains thereby exacerbating the challenge of homologous recombination. By using cipA*, designed to avoid repeated sequences, routine sequencing proceeded without difficulty and homologous recombination was not observed.
Toxicity has been observed by other groups working with heterologous expression of cipA, and were solved, at least in part, by using inducible promoters in E. coli and Lactococcus lactis[23, 27]. For this and other reasons we wished to express cipA* under the control of an inducible promoter. As there have been no inducible promoter systems described for T. saccharolyticum we designed and tested one based on xynA’s promoter. We chose this promoter due to the fact that XynA has been shown to be non-essential  and thus could be replaced with a gene of interest. We found that the xynA promoter avoided toxicity effects in both E. coli and T. saccharolyticum, although apparent toxicity was encountered using the pta/ack promoter.
In past in vitro heterologous cellulosome expression reports significant hydrolysis of microcrystalline cellulose was either not achieved or not tested, with one exception reaching 45% hydrolysis [15, 17, 18, 23, 25, 28–31]. In the data reported here if we remove the contribution of cell mass to the dry weights and compare that to the uninoculated bottles we see that total cellulose solubilization is between 98 and 100 percent for the co-culture of the CipA* expressing T. saccharolyiticum and the wild-type C. thermocellum, and between 71 and 93 percent for the CipA* expressing T. saccharolyiticum and C. thermocellum ΔcipA. While the trans-complementation co-culture can achieve close to wild type co-culture solubilization in some cases, it is rather variable. The lower hydrolysis from the CipA* expressing T. saccharolyticum and C. thermocellum ΔcipA co-culture, as compared to that of wild-type C. thermocellum, may be the result of one or more effects. First, while the cohesin dockerin interaction is quite robust, it is possible that when produced by two different strains the assembly of mature cellulosomes occurs less efficiently than if the components are being secreted simultaneously from the same cell [39, 40]. Second, the assembled cellulosomes are incapable of adhering to the surface of the more metabolically active T. saccharolyticum, and thus are either present free in the media or bound to the surface of C. thermocellum’s via the native anchor proteins [41–43]. This could result in local product inhibition, and may contribute to the lower than expected hydrolysis [3, 44–46].
Krauss et al. found that CipA purified from E. coli was populated with cellulosomal enzymes present in C. thermocellum supernatants and had near wild-type activity on microcrystalline cellulose . These results led the authors to conclude that the cohesin dockerin interactions are the primary means of cellulosome assembly, an interpretation which our work also supports. Whereas Krauss et al. reported cellulosome assembly in vitro, we demonstrate here assembly from components produced by growing cultures with heterologous production using a host that has a temperature optimum compatible with that of the C. thermocellum cellulosome.
In both Krauss et al. and this study, cellulosomes derived from components produced in separate organisms but otherwise unmodified are found to exhibit similar, although somewhat lower, activity on crystalline cellulose as compared to controls with components produced by a single organism. By contrast, studies involving “designer cellulosomes”, in which specific catalytic components bind in a specific order to chimeric scaffoldins, either report several-fold lower activity on crystalline cellulose compared to controls or do not report activity on crystalline cellulose at all. This difference could be because of the importance of a diverse population of cellulosomes with randomly-combined catalytic components, the smaller size of chimeric scaffoldins used in designer cellulosome work, or a combination. The model system reported here, featuring full-sized engineered scaffoldins, is a promising platform for understanding these effects.
An unexpected result of the expression and purification of CipA* in T. saccharolyticum was the observed instability upon high temperature treatment with SDS-PAGE loading buffer present. While a similar effect has been reported by Morag and Lamed in C. thermocellum, the conditions used to achieve this result are markedly different [36, 37]. In the previous reports, purified protein was subjected to low pH (3.5) or low ionic strength (dialyzed against double distilled water overnight) which resulted in the cleavage of an Asp-Pro peptide bond present in the cohesin domain of CipA. In contrast, no harsh conditions were applied to CipA* from T. saccharolyticum with supernatant pHs staying above 5.8 and with no dialysis treatment applied. This may indicate a considerable difference between the extracellular environment developed by cultures of C. thermocellum compared to that of T. saccharolyticum, and could be important in future attempts at heterologous cellulosome expression.