Functional heterologous expression of an engineered full length CipA from Clostridium thermocellum in Thermoanaerobacterium saccharolyticum

Background Cellulose is highly recalcitrant and thus requires a specialized suite of enzymes to solubilize it into fermentable sugars. In C. thermocellum, these extracellular enzymes are present as a highly active multi-component system known as the cellulosome. This study explores the expression of a critical C. thermocellum cellulosomal component in T. saccharolyticum as a step toward creating a thermophilic bacterium capable of consolidated bioprocessing by employing heterologously expressed cellulosomes. Results We developed an inducible promoter system based on the native T. saccharolyticum xynA promoter, which was shown to be induced by xylan and xylose. The promoter was used to express the cellulosomal component cipA*, an engineered form of the wild-type cipA from C. thermocellum. Expression and localization to the supernatant were both verified for CipA*. When a ΔcipA mutant C. thermocellum strain was cultured with a CipA*-expressing T. saccharolyticum strain, hydrolysis and fermentation of 10 grams per liter SigmaCell 101, a highly crystalline cellulose, were observed. This trans-species complementation of a cipA deletion demonstrated the ability for CipA* to assemble a functional cellulosome. Conclusion This study is the first example of an engineered thermophile heterologously expressing a structural component of a cellulosome. To achieve this goal we developed and tested an inducible promoter for controlled expression in T. saccharolyticum as well as a synthetic cipA. In addition, we demonstrate a high degree of hydrolysis (up to 93%) on microcrystalline cellulose.


Background
A long sought goal in the cellulosic ethanol field is onestep solubilization and fermentation without added enzymes [1,2]. Such consolidated bioprocessing, or CBP, is considered to be the ultimate low cost approach for cellulose hydrolysis and fermentation [2]. A successful CBP organism must be capable of solubilizing both cellulose and hemicellulose, and also fermenting the resulting sugars to a useful product (e.g., ethanol) at high yield and titer. Unfortunately, no single organism has yet been found or developed that combines these two essential characteristics [3].
Two saccharolytic bacteria of interest for development of CBP-enabling microbes are Clostridium thermocellum and Thermoanaerobacterium saccharolyticum, both Grampositive, thermophilic anaerobes. C. thermocellum exhibits among the highest growth rates on cellulose among described microbes [1], but lacks the ability to ferment hemicellulose. C. thermocellum's ability to solubilize crystalline cellulose, as well as other insoluble components of plant biomass, results from its elaborate, multi-protein cellulase complex or cellulosome [4][5][6][7]. T. saccharolyticum readily solubilizes hemicellulose and ferments all common sugars in biomass, but does not solubilize cellulose. This bacterium is highly amenable to genetic manipulation, indeed exhibiting natural competence, and has been engineered to make ethanol at high yields and titers [8][9][10][11].
The component of the C. thermocellum cellulosome with the highest molecular weight is the scaffoldin protein, CipA, which has been implicated in mediating the enzymatic synergy seen in the cellulosome [7,12,13]. The structure of the C. thermocellum CipA includes one cellulose binding domain or CBD, one type II dockerin which is used to associate with cell wall anchoring proteins, and 9 highly conserved type I cohesins interspaced by flexible linker regions [7]. The type 1 cohesin domains bind with high affinity to type 1 dockerin domains present in over 70 catalytically-active enzymes [14].
In C. thermocellum the presence or absence of CipA has little effect on activity on phosphoric acid swollen cellulose (PASC), carboxymethyl cellulose (CMC), or β-Glucan, but when absent, results in over an order of magnitude decrease in activity on microcrystalline cellulose [27,33]. With this in mind, microcrystalline cellulose was chosen as a test substrate for cellulosome assembly and complementation.
Heterologous expression of a functional cellulosome system in T. saccharolyticum is of interest both for fundamentally-oriented studies of microbial cellulose utilization and as a strategy for developing a CBP-enabling microorganism. The logical point of departure for this endeavor is expression of CipA. Here we endeavor to develop an inducible gene expression system in T. saccharolyticum, synthesize a gene (cipA*) coding for the same amino acid sequence as CipA but with more diversity in the DNA sequences for ease of use, express CipA* in T. saccharolyticum, and demonstrate functionality by complementing ΔcipA mutants of C. thermocellum.

Construction and testing of xylose/xylan inducible promoter system
We hypothesized that the promoter and leader sequences of the family 10 endoxylanase xynA (Tsac_1459) might be useful in expressing and secreting heterologous proteins in T. saccharolyticum. To test this hypothesis, we performed RT-PCR to determine gene transcription during growth on cellobiose, xylose, and xylan ( Figure 1). The results show that xynA is indeed expressed during growth on xylose or xylan and not glucose. In addition it was also observed to be catabolite repressed when glucose and xylose were both present (data not shown). Next, pMC200, which was designed to remove the xynA coding region, was transformed into T. saccharolyticum M1442 to confirm that the regions of homology and resistance cassette were sufficient to drive integration and the removal of the xynA open reading frame. XynA mutants were tested for growth on xylan and xylose with no substantial defect in growth, in agreement with previous results [34].
A second test of the system was performed by replacing xynA with the His tagged coding region for the secreted C. thermocellum family 48 cellulase cel48S (Clo1313_2747), via pMC212. Expression was confirmed via RT-PCR ( Figure 2) and was shown to still be regulated by xylose. Unfortunately, the strain did not produce detectable quantities of Cel48S, and as a result no further work was done with the strain. From these data we concluded the xynA promoter would function well as an inducible promoter for heterologous gene expression. Furthermore, it allowed cloning of potentially toxic gene products in E. coli. For example, it was found that when cipA*, discussed in detail below, was placed downstream of the native T. saccharolyticum pta/ack promoter, this resulted in substantial toxic effects to E. coli. However, no toxicity issues were seen with the xynA promoter upstream of cipA*. This indicates that unlike the pta/ack promoter the regions of the promoter found on pMC200 were insufficient to drive expression in E. coli in the absence of inducer (xylose).

Design and synthesis of CipA*
The native cipA sequence has large repeated regions likely to be problematic for sequencing as well as genetic stability once introduced into T. saccharolyticum. To optimize and homogenize cipA for expression and sequencing in T. saccharolyticum, we first identified repeated regions, the largest 4 of these being identical repeats of almost 500 base pairs ( Figure 3). To introduce silent mutations to remove this homology we utilized alternate codons with a minimum usage of 15% for a given amino acid in predicted T. saccharolyticum open reading frames. By manually and iteratively randomizing which alternate codon was used in a given location we were able to break up the repeated regions without generating extended new homologous sequences. It should also be noted that a standard codon optimization would, rather than rectifying the problem, only serve to enhance the redundancy. The result is cipA* which has 1325 nucleotide mismatches to the wild type and no unbroken repeats longer than 19 nucleotides while maintaining a wild type amino acid sequence ( Figure 3).

Expression and localization of CipA*
CipA, like other cellulosomal components, is secreted via the sec pathway which utilizes an N-terminal signal peptide which is cleaved to liberate the mature protein.
In order to assure the His tag's presence in the mature protein, cipA* was tagged at the C-terminus with a 10X His tag via a linker (GGGTGHHHHHHHHHH) for detection via western blot.
A number of methods to concentrate T. saccharolyticum supernatant were tested including His purification with Ni beads and FPLC. However, the best results were obtained using molecular weight cut off spin columns. Initial attempts included bacterial or mammalian protease inhibitors, but after it became clear that proteolysis was not an issue, their inclusion was discontinued. Xylose was used as an inducer since it was as effective as xylan ( Figure 1) but was more practical to use with spin columns. Unlike in E. coli no negative cellular effects were seen as the result of the presence, or induction, of cipA*. The concentrated protein was washed with 20 mM sodium citrate buffer (pH 5.7) to remove residual sugars to prevent heavy warping of the protein bands during migration on a SDS-PAGE gel.
Protein isolated from the supernatant and pellet was subject to western blot analysis, probing for the 10X-His tags. CipA* was observed in both the cytoplasm and in the supernatant under inducing conditions ( Figure 4) and was absent when grown on glucose (data not shown). To further confirm active secretion and to rule out the possibility of cell lysis as a means of generating CipA* in the supernatant, a second strain was constructed lacking the predicted native sec tag [35] (Additional file 1: Figure S2) and the localization of cipA* in the two strains was compared. The CipA* derivative lacking a sec tag was only detected in the pellet fraction whereas the CipA* with its sec tag intact was found in both the pellet and the supernatant suggesting lysis was not responsible for presence of CipA* in the supernatant (Figure 4).
Our initial attempts using the standard 100°C denaturation in preparation for running an SDS-PAGE resulted in protein cleavage as others have also observed for cipA [36,37]. However, when the denaturation temperature was dropped, or time was decreased, we obtained a single intact band.

Interspecies complementation
Co-cultures of C. thermocellum and T. saccharolyticum were used to test the functionality of heterologously expressed CipA*. We utilized a cipA deletion strain of C. thermocellum that shows markedly impaired growth on microcrystalline cellulose but still produces cellulosomal components other than CipA when cultivated on cellobiose. By co-culturing C. thermocellum ΔcipA with the strain of T. saccharolyticum that expresses CipA*, we hypothesized that CipA*, if functional, would complement the deletion in C. thermocellum and act as a scaffold thus restoring the ability to hydrolyze cellulose. To evaluate the co-culture's ability to hydrolyze cellulose we determined residual cellulose/dry weight and product formation ( Figure 5).
The wild type C. thermocellum performed well with respect to cellulose hydrolysis as measured by dry weight. The observed low product formation and unfermented sugars seen with wild type C. thermocellum are most likely the result of the low starting pH and the lack of pH control resulting in a discontinuation of metabolic, but not enzymatic hydrolysis activity later in growth. No strain of T. saccharolyticum alone appeared to have any effect on cellulose, but grew entirely on the supplied xylose, nor did the wild type strain of T. saccharolyticum rescue the cellulose hydrolysis defect in C. thermocellum strain DS11. Only T. saccharolyticum expressing cipA* (strain DHC15) was able to restore cellulose hydrolysis functionality to the C. thermocellum ΔcipA (strain DS11).
Finally, we wished to confirm that populated cellulosomes were being formed. Cellulosomes were purified via affinity digestion from co-cultures between DHC15 and DS11 and compared via native PAGE to those from wild type C. thermocellum and concentrated supernatants from DHC15 and DS11 grown indivigually (Additional file 2: Figure S3). As expected neither DHC15 nor DS11 were able to form cellulosomes when grown independently. When DHC15 and DS11 were grown together cellulosomes were produced with an identical native PAGE migration as those from wild type C. thermocellum.
These data demonstrate that the T. saccharolyticumproduced CipA* is capable of gathering and displaying functional cellulosomal enzymes. In addition, the appearance of the pellet further supported the removal of cellulose from the wild type C. thermocellum and the co-cultures of cipA* and ΔcipA as these dry pellets were nearly translucent, suggesting only cell debris, rather than the white cellulose observed in the other samples.

Discussion
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 [7]. 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 [34] 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 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][42][43]. This could result in local product inhibition, and may contribute to the lower than expected hydrolysis [3,[44][45][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 [27]. 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.

Conclusion
Combined with the native ability of T. saccharolyticum to utilize hemicellulose and the availability of engineered strains that produce ethanol at high yield and titer [8,11], a strain of T. saccharolyticum with the ability to solubilize cellulose would be a strong candidate organism for CBP. Our results, including expression and secretion of a functional, engineered, full-length CipA, represent a step toward developing such an organism. In addition, we demonstrate a model system in which understanding cellulolytic organisms and their enzyme systems can be tested by systematically reconstructing them.

Microorganisms and growth media
The parent strain for all T. saccharolyticum strains is M1442 [47], engineered with deletions of the genes for phosphotransacetylase, acetate kinase, and lactate dehydrogenase and expressing genes for urea utilization from C. thermocellum which serve to buffer acid production [11]. C. thermocellum DSM1313 was obtained from DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Germany). C. thermocellum strain DS11, a cipA deletion mutant, was generated in our laboratory and is derived from C. thermocellum DSM1313 Δhpt and was supplied by D. G. Olson [48]. All T. saccharolyticum strains were grown in modified DSMZ M122 medium [9] with 10 g/l xylose or xylan where noted, 0.5 g/l urea at pH 6.3 and 55°C unless otherwise stated. C. thermocellum strains were grown in M122 with 10 g/l cellobiose, pH 7.0 at 55°C. All cultures were grown in a Coy anaerobic chamber under a nitrogen, carbon dioxide, and hydrogen gas mix unless otherwise noted.

Plasmid and strain construction
All plasmids were constructed in Saccharomyces cerevisiae FY2 [49] via yeast mediated homologous recombination [50], isolated from yeast with a Zymoprep Yeast Plasmid Miniprep II (Zymo Research, Irvine, CA) and transformed into chemically competent Invitrogen E. coli TOP10 (Invitrogen Corp, Carlsbad, CA) to generate sufficient quantities of plasmid for transformation into T. saccharolyticum. Plasmids were then isolated from E. coli with the QIAGEN Plasmid Mini kit (QIAGEN Inc, Hilden, Germany). T. saccharolyticum was transformed with plasmid DNA as previously described in a Coy anaerobic chamber [9]. Plasmids and strains are listed in Table 1. Genotypic confirmation for modified strains was obtained via PCR and sequencing. CipA* was synthesized by GeneArt (Additional file 3: Figure S1) [51].

Reverse transcription PCR
RNA was isolated from cultures of T. saccharolyticum incubated overnight at 55°C in modified M122 [9] with Co-cultures 50 ml co-cultures were grown in 115 ml nitrogen flushed anaerobic serum bottles agitated in an incubator at 55°C. The medium for the co-cultures was modified DSMZ M122 [9] with 10 g/l xylose and 10 g/l Sigmacell 101 (a microcrystalline cellulose similar to avicel) and 0.5 g/l cellobiose to assist in the initial growth of C. thermocellum. The initial pH was pH 6.3, previously demonstrated to be suitable for co-cultures between these two organisms [52]. Co-cultures were allowed to grow for 5 days. Cellulosomes were purified via affinity digestion of PASC [27].

Product formation and dry weight
Dry weight was determined by pelleting the remaining cellulose via centrifugation for 10 minutes at 8000 g, washing twice with deionized water, and drying at 45°C under vacuum for 2 days. Dried pellets were then weighed to determine residual solids. The contribution of cell mass was taken into account with controls containing 10 g/l cellobiose in place of Sigmacell. Residual sugars and ethanol in the supernatants were quantified by HPLC using an Aminex HPX-87H column (Bio-Rad Laboratories, Hercules, CA

Additional files
Additional file 1: Sequence of cipA* minus the predicted signal peptide with a C-terminal 10X his tag and linker region.
Additional file 2: Coomassie stained native PAGE of trans-species formed cellulosomes.
Additional file 3: Sequence of cipA* with a C-terminal 10X his tag and linker region.

Competing interests
This research was supported in part by Mascoma Corporation, Lebanon NH, with which authors DHC, CDH, DAH and LRL are affiliated. Mascoma Corporation has a commercial interest in the organisms used in this study. The authors DHC and AMG are listed on the international and national pending patents for the sequence of cipA* [52].
Authors' contributions DHC performed the work presented herein and drafted the manuscript. DGO assisted in drafting the manuscript. CDH, AMG, DAH and LRL supervised the work and assisted in drafting the manuscript. All authors read and approved the final manuscript.
favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof."