ABSTRACT
Further understanding of the plant cell wall degradation system of Clostridium cellulolyticum and the possibility of metabolic engineering in this species highlight the need for a means of random mutagenesis. Here, we report the construction of a Tn1545-derived delivery tool which allows monocopy random insertion within the genome.
The economic feasibility and sustainability of lignocellulosic ethanol production are dependent on the development of robust microorganisms which can efficiently degrade and/or convert plant biomass to ethanol (5). The anaerobic, mesophilic, Gram-positive bacterium Clostridium cellulolyticum is a candidate microorganism, as it is capable of hydrolyzing plant cell wall polysaccharides and fermenting the hydrolysis products to ethanol and other metabolites (7). C. cellulolyticum achieves this efficient hydrolysis by using multiprotein extracellular enzymatic complexes, termed cellulosomes (13). As plant cell walls consist of several intertwined heterogeneous polymers, primarily composed of cellulose, hemicellulose, and pectin, cellulosomes contain many subunits (cellulosomal enzymes) with diverse and complementary enzymatic properties (2). Thus, this model organism is also a good candidate for the development of novel and efficient cellulases and hemicellulases for the saccharification of plant biomass.
Gene transfer has been successfully carried out in C. cellulolyticum (8, 12). This possibility has allowed the in vivo function of cellulosomal enzymes in C. cellulolyticum to be examined by overexpression (9) or down expression (11) of targeted genes. However, random mutagenesis of the entire chromosome and screening of mutants to identify key components for plant cell wall degradation have never been described. Conjugative transfer of Tn1545 from Enterococcus faecalis to C. cellulolyticum has been described but is limited by low transfer frequency and poor reproducibility (8). To improve transposon mutagenesis of C. cellulolyticum, we exploited the two-plasmid Tn1545 delivery system described by Trieu-Cuot et al. (15). In this system, the Tn916 integrase-encoding gene is carried by an expression vector, whereas the attachment site of Tn1545 is carried by a suicide vector. Tn916 and Tn1545 being closely related (4), integration of the Tn1545 derivative occurs in the genome after transformation of the strain with both vectors (15).
Development of a functional Tn1545-based transposon delivery system for C. cellulolyticum.
For the construction of the Tn916 integrase expression vector (pJint5), two cloning steps in Escherichia coli DH5α were necessary (Fig. 1 A). First, the integrase-encoding gene int was amplified with intbam (5′-CCCGTAGGATCCAGAATTTAAAAGGAGGGATTAAAATGATAAAATAGTATTAAGTCGTATCAAGGCT-3′) and intnar (5′-CATTAAGTGGCGCCTACTAAGCAAACAAGACGCTCCTGT-3′) primers by using pAM120 (6) as a template. These primers harbor tails (underlined) for cloning under the control of the thiolase gene promoter (Pthl, a constitutive promoter in C. cellulolyticum) and the translation initiation signals from C. cellulolyticum in pSOS952 as described previously (9, 10). Second, a restriction fragment containing Pthl-int was cloned in the clostridial replicative vector pJIR750 (1) to give pJint5 (Fig. 1A). pJIR750 and pJint5 were transferred by electrotransformation into C. cellulolyticum H10 (ATCC 35319) and both types of transformants were selected in basal medium supplemented with cellobiose (4 g liter−1) and thiamphenicol (10 μg ml−1) as previously described (12). In the first trial of mutagenesis, the pACYC184 derivative pAT112 (15) was used to transform H10(pJIR750) and H10(pJint5) strains. pAT112 harbors an erythromycin-resistant cassette and the attachment site of Tn1545 (attTn1545) recognized by the integrase (Fig. 1B). In both cases, we obtained low transfer frequencies of about 10−9 per recipient. Molecular characterization of transformants showed that pAT112 replicated in C. cellulolyticum H10 and that transposition events had not occurred (data not shown). To overcome this problem of the pACYC184 derivative replication, attTn1545 was cloned in a suicide vector in C. cellulolyticum harboring an erythromycin-resistant cassette (pMEM2). The construction of pMEM2 (Fig. 1B) involved the cloning of ermB obtained from pAT18 (14) with primers emup (5′-CCCTATATGCTTAGAAGCAAACTTAAGAGTC-3′) and emdown (5′-CCCTATAGGTACCATCGATACAAATTCCCCGAT-3′) in the EcoRV linear pMOSBlueT (Amersham Biosciences, Saclay, France), a pUC derivative which does not replicate in clostridia. attTn1545, carried by a KpnI-EcoRI fragment, was obtained from pAT112 with primers ISLKpnI (5′-GGGGTACCAGGAGCGTCTTGTTGCTTAG-3′) and ISREcoRI (5′-GGAATTCGGATTAAATCGTCGTACAAAGG-3′) and was cloned in pMEM2 to give pMIS1545 (Fig. 1B). Mutagenesis by transposition was then undertaken by using pMIS1545 to transform H10(pJIR750) and H10(pJint5) in three independent experiments. Whereas no transformants were obtained with the control strain H10(pJIR750), erythromycin- and thiamphenicol-resistant clones were obtained with H10(pJint5) on basal medium supplemented with cellobiose (4 g liter−1), thiamphenicol (5 μg ml−1), and erythromycin (2.5 μg ml−1). The transfer frequencies of about 10−5 per recipient suggested that the production of Tn916 integrase, directed by pJint5, allowed the integration of pMIS1545 into the C. cellulolyticum H10 genome.
Flow diagram of plasmid constructions. (A) Construction of the integrase expression vector (pJint5). (B) Construction of the Tn1545 derivative (pMIS1545). The origin of replication is given for each plasmid; plasmids harboring oriR pBR322, oriR pUC, and oriR pACYC184 replicate in E. coli, and plasmids harboring oriR pIM13, oriR pAMβ1, and oriR pACYC184 replicate in C. cellulolyticum. The genes bla, cat, ermB, and aphA3 encode resistance to ampicillin, chloramphenicol/thiamphenicol, erythromycin, and kanamycin, respectively. Pthl is the promoter of the thiolase-encoding gene in C. acetobutylicum, int is the gene encoding the Tn916 integrase, and attTn1545 is the attachment site of Tn1545. Amplification steps are described, and restriction sites of enzymes used for digestion are given. Δ NarI-BamHI indicates the deletion of a NarI-BamHI fragment present in pSOS952.
Molecular characterization of the transposition mutants.
The molecular characterization of six erythromycin- and thiamphenicol-resistant clones was performed by Southern blotting of HindIII-digested genomic DNA and hybridization with a probe generated by PCR labeling with digoxigenin-dUTP (DIG DNA labeling mix; Roche Applied Science, Meylan, France) and primers emup and emdown. The Tn1545 derivative was detected in all transconjugants, while it was not detected in the wild-type strain (Fig. 2). The sites of transposition for some clones were examined by inverse PCR. After digestion and ligation of genomic DNA of 14 clones, DNA fragments carrying pMIS1545/chromosome junction sequences were amplified by PCR with two sets of divergent primers hybridizing with left and right extremities of Tn1545 carried by pMIS1545. Primers 1545EcoD (5′-ATGAATGAGCTTTGATACGACGAT-3′) and 1545EcoR (5′-TTGACCTTGATAAAGTGTGATAAGTCC-3′) and primers 1545KpnD (5′-AATACTCGAAAGCACATAGAATAAGGC-3′) and 1545KpnR (5′-TAGCTGTCAGAAGTGGTAAATAAGTAGTAAAT-3′) were used to determine the left and right junctions, respectively. As shown in Table 1, only one insertion has been identified for each clone, whereas analysis of Fig. 2 reveals two possible insertions of pMIS1545. Theoretical restriction analyses of these particular loci suggest that for both clones, the bands of about 7 kb correspond to partial digestions rather than to multicopy insertions. Analyses of the target sites (Table 1) revealed that pMIS1545 transposed randomly in regions where stretches of T are separated from stretches of A by the coupling sequence as previously described for Tn1545 (16). The insertion in the mutant If generated a 6-bp duplication of the target DNA. This result is not typical of the insertion of elements from the Tn916 and Tn1545 family (3) and cannot be explained. The insertions in the 14 clones analyzed appeared to be randomly distributed throughout the genome of C. cellulolyticum, in open reading frames (ORFs) or in intergenic regions. Insertion in ORFs, i.e., gene disruption, should be frequent, as the percentage of GC (GC%) in the genome is 37.4 (http://genome.ornl.gov/microbial/ccel/ ). Promoter predictions (http://linux1.softberry.com/ ) at the junctions of pMIS1545 and the insertion sites revealed that putative −10/−35 promoter sites were found in 12 of the 14 sequences analyzed (data not shown). Thus, as for Tn916 (4), insertions in intergenic regions, or in a particular gene of an operon, would not lead to the absence of downstream gene expression. Furthermore, stability of the insertions was assessed for four clones after growth for approximately 100 generations in broth with and without erythromycin.
Southern blot analysis of mutants, using a probe generated from pMIS1545. Numbers indicate DNA fragment sizes (in kb) of HindIII-digested chromosomal DNA. Lane 1 contains HindIII-linearized pMIS1545 (4.4 kb), lane 2 DNA from the wild-type H10 strain, and lanes 3 to 8 DNA from the mutant H10::pMIS1545 strains (Ia, Ib, Ic, Id, Ie, and If). DNA fragments labeled with an asterisk were identified by inverse PCR.
Insertion sites from 14 clones containing pMIS1545 integrated in the genome
Conclusion.
We have adapted the two-plasmid Tn1545 delivery system described by Trieu-Cuot et al. (15) to perform transposon mutagenesis of C. cellulolyticum. In the presence of Tn916 integrase produced in the strain, the Tn1545 derivative (pMIS1545) has been successfully integrated into the genome. Analyses of the insertion mutants obtained showed that the insertions are random in AT-rich sites and in monocopy in the genome. This random mutagenesis system should facilitate the identification of essential genes involved in biomass utilization and thus the development of new strategies for strain engineering for biofuel production from biomass.
ACKNOWLEDGMENTS
This research was supported by grants from the Centre National de la Recherche Scientifique, the Agence Nationale de la Recherche, and the Université d'Aix-Marseille.
We thank M. Abou Arraj and L. Cordiez for technical assistance and S. Pagès, H.-P. Fierobe, and S. Perret for fruitful discussions.
FOOTNOTES
- Received 5 October 2009.
- Accepted 22 April 2010.
- Copyright © 2010 American Society for Microbiology
