The resistance of cellulases and xylanases to proteolytic inactivation

The sensitivity of a range of cellulases and xylanases to proteolytic inactivation was investigated. The xylanases, all the Clostridium thermocellum cellulases and cellulase E from Pseudomonas fluorescens subsp. cellulosa exhibited no decrease in catalytic activity during a 3-h incubation with proteinases of the small intestine. Under these conditions, the control Escherichia coli enzymes analysed had half-lives of 4.3-13.5 min. The addition of substrate significantly decreased the sensitivity of proteinase-labile enzymes to inactivation. The significance of these data in relation to the use of cellulases and xylanases for improving animal nutrition is discussed.

Specificity of an esterase (XYLD) from Pseudomonas fluorescens subsp. cellulosa

Activity of an esterase from Pseudomonas fluorescens subsp. cellulosa (XYLD) on an insoluble feruloylated hemicellulose substrate (de-starched wheat bran) was dependent on the source of added endo-xylanase. The esterase exhibited high selectivity for the nature, position of linkage and size of the feruloylated oligosaccharides generated by hydrolysis of the hemicellulose. Increased affinity of XYLD with increasing size of the oligosaccharide substrate suggests that optimal activity is observed on substrates with at least 4 sugars.

A non-modular endo-beta-1,4-mannanase from Pseudomonas fluorescens subspecies cellulosa

Pseudomonas fluorescens subsp. cellulosa when cultured in the presence of carob galactomannan degraded the polysaccharide. To isolate gene(s) from P. fluorescens subsp. cellulosa encoding endo-beta-1,4-mannanase (mannanase) activity, a genomic library of Pseudomonas DNA, constructed in lambda ZAPII, was screened for mannanase-expressing clones using the dye-labelled substrate, azo-carob galactomannan. The nucleotide sequence of the pseudomonad insert from a mannanase-positive clone revealed a single open reading frame of 1257 bp encoding a protein of M(r) 46,938. The deduced N-terminal sequence of the putative polypeptide conformed to a typical prokaryotic signal peptide. Truncated derivatives of the mannanase, lacking 54 and 16 residues from the N- and C-terminus respectively of the mature form of the enzyme, did not exhibit catalytic activity. Inspection of the primary structure of the mannanase did not reveal any obvious linker sequences or protein motifs characteristic of the non-catalytic domains located in other Pseudomonas plant cell wall hydrolases. These data indicate that the mannanase is non-modulator, comprising a single catalytic domain. Comparison of the mannanase sequence with those in the SWISSPROT database revealed greatest sequence homology with the mannanase from Bacillus sp. Thus the Pseudomonas enzyme belongs to glycosyl hydrolase Family 26, a family containing mannanases and endoglucanases. Analysis of the substrate specificity of the mannanase showed that the enzyme hydrolysed mannan and galactomannan, but displayed little activity towards other polysaccharides located in the plant cell wall. The enzyme had a pH optimum of approx. 7.0, was resistant to proteolysis and had an M(r) of 46,000 when expressed by Escherichia coli.

Epithelial sorting of a glycosylphosphatidylinositol-anchored bacterial protein expressed in polarized renal MDCK and intestinal Caco-2 cells

To evaluate whether a glycosylphosphatidylinositol (GPI) anchor can function as a protein sorting signal in polarized intestinal epithelial cells, the GPI-attachment sequence from Thy-1 was fused to bacterial endoglucanase E' (EGE') from Clostridium thermocellum and polarity of secretion of the chimeric EGE'-GPI protein was evaluated. The chimeric EGE'-GPI protein was shown to be associated with a GPI anchor by TX-114 phase-partitioning and susceptibility to phosphoinositol-specific phospholipase C. In polarized MDCK cells, EGE' was localized almost exclusively to the apical cell surface, while in polarized intestinal Caco-2 cells, although 80% of the extracellular form of the enzyme was routed through the apical membrane over a 24 hour period, EGE' was also detected at the basolateral membrane. Rates of delivery of EGE'-GPI to the two membrane domains in Caco-2 cells, as determined with a biotinylation protocol, revealed apical delivery was approximately 2.5 times that of basolateral. EGE' delivered to the basolateral cell surface was transcytosed to the apical surface. These data indicate that a GPI anchor does represent a dominant apical sorting signal in intestinal epithelial cells. However, the mis-sorting of a proportion of EGE'GPI to the basolateral surface of Caco-2 cells provides an explanation for additional sorting signals in the ectodomain of some endogenous GPI-anchored proteins.

Structure of the catalytic core of the family F xylanase from Pseudomonas fluorescens and identification of the xylopentaose-binding sites

BACKGROUND

Sequence alignment suggests that xylanases evolved from two ancestral proteins and therefore can be grouped into two families, designated F and G. Family F enzymes show no sequence similarity with any known structure and their architecture is unknown. Studies of an inactive enzyme-substrate complex will help to elucidate the structural basis of binding and catalysis in the family F xylanases.

RESULTS

We have therefore determined the crystal structure of the catalytic domain of a family F enzyme, Pseudomonas fluorescens subsp. cellulosa xylanase A, at 2.5 A resolution and a crystallographic R-factor of 0.20. The structure was solved using an engineered catalytic core in which the nucleophilic glutamate was replaced by a cysteine. As expected, this yielded both high-quality mercurial derivatives and an inactive enzyme which enabled the preparation of the inactive enzyme-substrate complex in the crystal. We show that family F xylanases are eight-fold alpha/beta-barrels (TIM barrels) with two active-site glutamates, one of which is the nucleophile and the other the acid-base. Xylopentaose binds to five subsites A-E with the cleaved bond between subsites D and E. Ca2+ binding, remote from the active-site glutamates, stabilizes the structure and may be involved in the binding of extended substrates.

CONCLUSIONS

The architecture of P. fluorescens subsp. cellulosa has been determined crystallographically to be a commonly occurring enzyme fold, the eight-fold alpha/beta-barrel. Xylopentaose binds across the carboxy-terminal end of the alpha/beta-barrel in an active-site cleft which contains the two catalytic glutamates.

Xylanase B from Neocallimastix patriciarum contains a non-catalytic 455-residue linker sequence comprised of 57 repeats of an octapeptide

A Neocallimastix patriciarum cDNA library was screened for xylanase-expressing clones, which were distinct from the previously characterized N. patriciarum xynA cDNA encoding xylanase A. A single cDNA, designated xynB, which did not exhibit homology with xynA, was isolated. Northern-blot analysis of mRNA from Avicel-grown N. patriciarum showed that xynB hybridized to a 3.4 kb mRNA species. The nucleotide sequence of xynB revealed a single open reading frame of 2580 bp coding for a protein designated xylanase B (XYLB), of M(r) 88,066. The primary structure of XYLB was comprised of a 21-residue N-terminal signal peptide, followed by a 304-amino acid sequence that exhibited substantial homology with the catalytic domains of family F xylanases. The N-terminal domain was linked to a C-terminal 70-residue sequence by a putative linker region, comprising 12 tandem repeats of a sequence containing TLPG as the core sequence, followed by an octapeptide XSKTLPGG where X can be S, K or N, which was repeated in tandem 45 times. Truncated derivatives of xynB encoding the N-terminal 338 residues directed the synthesis of a functional xylanase, confirming that the region of XYLB, which exhibited homology with family F xylanases, constitutes the catalytic domain. To investigate the catalytic properties of XYLB, the catalytic domain was fused to the Escherichia coli maltose-binding protein, and the fusion protein purified by amylose affinity chromatography. The purified enzyme hydrolysed oat, rye and wheat arabinoxylan releasing primarily xylobiose, xylotriose and some xylose. The XYLB fusion did not cleave any cellulosic substrates. The data presented in this report suggest that the multiple xylanases of N. patriciarum arose, not through the duplication of a single gene, but by the transfer of distinct xylanase-encoding DNA sequences into the anaerobic fungus. The possible origin of the xynB gene is discussed.

Unmodified and recombinant strains of Lactobacillus plantarum are rapidly lost from the rumen by protozoal predation

A genetically-manipulated strain of Lactobacillus plantarum and the unmodified parent strain were introduced into the rumen of sheep at an initial inoculum level of 1 x 10(7) cfu ml-1 of rumen fluid. There were no significant differences between the viable counts of the two inoculants throughout a 24 h sampling period. The rates of loss were 0.36 and 0.29 h-1 (proportion of colony-forming units lost, measured over the first 2 h) for the parent strain and recombinant strain respectively, and within 24 h of inoculation neither of the strains were detectable in rumen fluid. Further experiments in vitro revealed that the inoculants persisted in sterile rumen fluid with a loss rate of 0.044 and 0.057 h-1 for the parent strain and the recombinant strain respectively. Incubations with rumen fluid alone, protozoa-free rumen fluid and protozoa-enriched rumen fluid revealed that protozoal predation was the most significant factor in the loss of the introduced population. The loss rates from protozoa-free rumen fluid were not significantly different (P < 0.05) from those observed in sterile rumen fluid.

Intronless celB from the anaerobic fungus Neocallimastix patriciarum encodes a modular family A endoglucanase

The cDNA designated celB from the anaerobic rumen fungus Neocallimastix patriciarum contained a single open reading frame of 1422 bp coding for a protein (CelB) of M(r) 53,070. CelB expressed by Escherichia coli harbouring the full-length gene hydrolysed carboxymethylcellulose in the manner of an endoglucanase, but was most active against barley beta-glucan. It also released reducing sugar from xylan and lichenan, but was inactive against crystalline cellulose, laminarin, mannan, galactan and arabinan. The rate of hydrolysis of cellulo-oligosaccharides by CelB increased with increasing chain length from cellotriose to cellopentaose. The predicted structure of CelB contained features indicative of modular structure. The first 360 residues of CelB constituted a fully functional catalytic domain that was homologous with bacterial endoglucanases belonging to cellulase family A, including five which originate from three different species of anaerobic rumen bacteria. Downstream from this domain, and linked to it by a serine/threonine-rich hinge, was a non-catalytic domain containing short tandem repeats, homologous to the C-terminal repeats contained in xylanase A from the same anaerobic fungus. Unlike previous fungal cellulases, genomic celB was devoid of introns. This lack of introns and the homology of its encoded product with rumen bacterial endoglucanases suggest that acquisition of celB by the fungus may at some stage have involved horizontal gene transfer from a prokaryote to N. particiarum.

Evidence for a general role for high-affinity non-catalytic cellulose binding domains in microbial plant cell wall hydrolases

Cellulases expressed by Cellulomonas fimi consist of a catalytic domain and a discrete non-catalytic cellulose-binding domain (CBD). To establish whether CBDs are common features of plant cell-wall hydrolases from C. fimi, the molecular architecture of xylanase D (XYLD) from this bacterium was investigated. The gene encoding XYLD, designated xynD, consisted of an open reading frame of 1936 bp encoding a protein of M(r) 68,000. The deduced primary sequence of XYLD was confirmed by the size (64 kDa) and N-terminal sequence of the purified recombinant xylanase. Biochemical analysis of the purified enzyme revealed that XYLD is an endoacting xylanase which displays no detectable activity against polysaccharides other than xylan. The predicted primary structure of XYLD comprised an N-terminal signal peptide followed by a 190-residue domain that exhibited significant homology to Family-G xylanases. Truncated derivatives of xynD, encoding the N-terminal 193 amino acids of mature XYLD directed the synthesis of a functional xylanase, confirming that the 190-residue N-terminal sequence constitutes the catalytic domain. The remainder of the enzyme consisted of two approximately 90-residue domains, which exhibited extensive homology with each other, and limited sequence identity with CBDs from other polysaccharide hydrolases. Between the two putative CBDs is a 197-amino-acid sequence that exhibits substantial homology with Rhizobium NodB proteins. The four discrete domains in XYLD were separated by either threonine/proline-or novel glycine-rich linker regions. Although full-length XYLD adsorbed to cellulose, truncated derivatives of the enzyme lacking the C-terminal CBD hydrolysed xylan but did not bind to cellulose.(ABSTRACT TRUNCATED AT 250 WORDS)

Cellulase Ss (CelS) is synonymous with the major cellobiohydrolase (subunit S8) from the cellulosome of Clostridium thermocellum

The controversy regarding the identity of a major cellulosomal component type from two different strains of Clostridium thermocellum has been resolved. The principal cellobiohydrolase, subunit S8, from the cellulosome of strain YS has been demonstrated to be synonymous with cellulase component Ss (CelS) from the cellulosome of ATCC strain 27405. This component is not related to any other cellulosomal subunit or cloned endoglucanase in this organism.

A modular esterase from Pseudomonas fluorescens subsp. cellulosa contains a non-catalytic cellulose-binding domain

The 5' regions of genes xynB and xynC, coding for a xylanase and arabinofuranosidase respectively, are identical and are reiterated four times within the Pseudomonas fluorescens subsp. cellulosa genome. To isolate further copies of the reiterated xynB/C 5' region, a genomic library of Ps. fluorescens subsp. cellulosa DNA was screened with a probe constructed from the conserved region of xynB. DNA from one phage which hybridized to the probe, but not to sequences upstream or downstream of the reiterated xynB/C locus, was subcloned into pMTL22p to construct pFG1. The recombinant plasmid expressed a protein in Escherichia coli, designated esterase XYLD, of M(r) 58,500 which bound to cellulose but not to xylan. XYLD hydrolysed aryl esters, released acetate groups from acetylxylan and liberated 4-hydroxy-3-methoxycinnamic acid from destarched wheat bran. The nucleotide sequence of the XYLD-encoding gene, xynD, revealed an open reading frame of 1752 bp which directed the synthesis of a protein of M(r) 60,589. The 5' 817 bp of xynD and the amino acid sequence between residues 37 and 311 of XYLD were almost identical with the corresponding regions of xynB and xynC and their encoded proteins XYLB and XYLC. Truncated derivatives of XYLD lacking the N-terminal conserved sequence retained the capacity to hydrolyse ester linkages, but did not bind cellulose. Expression of truncated derivatives of xynD, comprising the 5' 817 bp sequence, encoded a non-catalytic polypeptide that bound cellulose. These data indicate that XYLD has a modular structure comprising of a N-terminal cellulose-binding domain and a C-terminal catalytic domain.

Secretion of a prokaryotic cellulase in bacterial and mammalian cells

The catalytic domain of mature Clostridium thermocellum endoglucanase E (EGE') and derivatives of the enzyme fused to prokaryote and eukaryote signal peptides (SP), were produced in Chinese hamster ovary (CHO) cells and Escherichia coli. All three forms of the endoglucanase were secreted into the periplasm of Escherichia coli, but only derivatives of the enzyme containing an N-terminal SP were exported from CHO cells. Extracellular EGE', purified from E. coli and CHO cultures, displayed similar properties suggesting that glycosylation of the enzyme in the eukaryote did not significantly alter the protein's properties. Data presented in this report indicate that mature EGE' contains secretion signals which are recognised only by the E. coli protein export apparatus, suggesting that there are differences in the recognition of certain secretion signals in eukaryotes and prokaryotes. As mature EGE' does not contain secretion signals recognised by the mammalian cell, membrane translocation of the bacterial cellulase in a higher eukaryote is directed by an N-terminal prokaryotic SP.

Manipulation of the repertoire of digestive enzymes secreted into the gastrointestinal tract of transgenic mice

In non-ruminant livestock the energy which can be derived from dietary cellulose and xylan is limited by the inefficient microbial fermentation of these polymers in the hind-gut. Furthermore, in poultry, cereal-derived plant structural polysaccharides impair normal digestive function through the formation of gel-like structures, which trap nutrients rendering them unavailable to the animal. The nutrition of non-ruminant livestock could be significantly improved by the depolymerization of plant structural polysaccharides, through the introduction of cellulase activity into the small intestines of these animals. Here we describe the expression of Clostridium thermocellum endoglucanase E in the exocrine pancreas of transgenic mice. A non-glycosylated active enzyme is secreted into the small intestines, and is resistant to proteolytic inactivation, demonstrating the feasibility of generating non-ruminant animals with the endogenous capacity to depolymerize plant structural polysaccharides in the small intestines.

Gene sequence and properties of CelI, a family E endoglucanase from Clostridium thermocellum

The Clostridium thermocellum celI gene, coding for endoglucanase I (CelI), consists of an open reading frame (ORF) of 2640 nucleotides and codes for a protein of M(r) 98531. The ORF was confirmed as celI by comparing the N-terminal sequence of purified recombinant CelI with that deduced from the nucleotide sequence. CelI hydrolysed lichenan and carboxymethylcellulose, but was principally active against barley beta-glucan. It exhibited significant sequence identity with subfamily E2 endoglucanases, and by analogy with others in this group contains a catalytic domain of around 500 residues located in the N-terminal half of the protein. The C-terminal region of CelI was highly homologous with the cellulose-binding domain of the non-catalytic cellulosome subunit, S1. A repeated segment, previously shown to be highly conserved in xylanase Z and in other endoglucanases from C. thermocellum, was absent from CelI. Antiserum raised against purified recombinant CelI cross-reacted with proteins contained in the cellulosomes of two strains of C. thermocellu, suggesting that CelI is either a component of the cellulosome or is homologous to other cellulosome proteins. A second gene, located upstream of celI, consisted of an ORF of 1671 nucleotides, coding for a protein of M(r) 61042. Based on its homology with the Escherichia coli tar gene product, the polypeptide encoded by the second gene is tentatively identified as a sensory transducer.

Crystallization and preliminary X-ray analysis of the catalytic domain of xylanase a from Pseudomonas fluorescens subspecies cellulosa

The catalytic domain of the xylan-degrading enzyme xylanase A, from Pseudomonas fluorescens subspecies cellulosa, has been expressed in Escherichia coli and crystallized. The crystals are well ordered and diffract to 1.8 A using X-rays generated at the Photon Factory in Japan. The crystals are orthorhombic, space group P2(1)2(1)2(1) with a = 95.7 A, b = 97.1 A and c = 149.8 A (all +/- 0.2 A). The similarity of the a and b cell edges, the intensity of the reflections along c* and the self rotation function results suggest a pseudo-tetragonal arrangement of molecules in the unit cell. There are probably four molecules in the asymmetric unit.

The role of conserved tryptophan residues in the interaction of a bacterial cellulose binding domain with its ligand

The five conserved tryptophan residues in the cellulose binding domain of xylanase A from Pseudomonas fluorescens subsp. cellulosa were replaced with alanine and phenylalanine. The mutated domains were fused to mature alkaline phosphatase, and the capacity of the hybrid proteins to bind cellulose was assessed. Alanine substitution of the tryptophan residues, in general, resulted in a significant decrease in the capacity of the cellulose binding domains to bind cellulose. Mutant domains containing phenylalanine substitution retained some affinity for cellulose. The C-terminal proximal tryptophan did not play an important role in ligand binding, while Trp13, Trp34 and Trp38 were essential for the cellulose binding domain to retain cellulose binding capacity. Data presented in this study suggest major differences in the mechanism of cellulose attachment between Pseudomonas and Cellulomonas cellulose binding domains.

Identification of the cellulose-binding domain of the cellulosome subunit S1 from Clostridium thermocellum YS

The 3' region of a gene designated cipB, which shows strong homology with cipA that encodes the cellulosome SL subunit of Clostridium thermocellum ATCC 27405, was isolated from a gene library of C. thermocellum strain YS. The truncated S1 protein encoded by the cipB derivative bound tightly to cellulose. The cellulose-binding domain in this polypeptide consisted of a C-terminal proximal 167 residue sequence which showed complete identity with residues 337-503 of mature SL from C. thermocellum strain ATCC 27405. The cellulose-binding domain interacted with both crystalline and amorphous cellulose, but not with xylan.

Homologous catalytic domains in a rumen fungal xylanase: evidence for gene duplication and prokaryotic origin

A cDNA (xynA), encoding xylanase A (XYLA), was isolated from a cDNA library, derived from mRNA extracted from the rumen anaerobic fungus, Neocallimastix patriciarum. Recombinant XYLA, purified from Escherichia coli harbouring xynA, had a M(r) of 53,000 and hydrolysed oat-spelt xylan to xylobiose and xylose. The enzyme did not hydrolyse any cellulosic substrates. The nucleotide sequence of xynA revealed a single open reading frame of 1821 bp coding for a protein of M(r) 66,192. The predicted primary structure of XYLA comprised an N-terminal signal peptide followed by a 225-amino-acid repeated sequence, which was separated from a tandem 40-residue C-terminal repeat by a threonine/proline linker sequence. The large N-terminal reiterated regions consisted of distinct catalytic domains which displayed similar substrate specificities to the full-length enzyme. The reiterated structure of XYLA suggests that the enzyme was derived from an ancestral gene which underwent two discrete duplications. Sequence comparison analysis revealed significant homology between XYLA and bacterial xylanases belonging to cellulase/xylanase family G. One of these homologous enzymes is derived from the rumen bacterium Ruminococcus flavefaciens. The homology observed between XYLA and a rumen prokaryote xylanase could be a consequence of the horizontal transfer of genes between rumen prokaryotes and lower eukaryotes, either when the organisms were resident in the rumen, or prior to their colonization of the ruminant. It should also be noted that Neocallimastix XYLA is the first example of a xylanase which consists of reiterated sequences. It remains to be established whether this is a common phenomenon in other rumen fungal plant cell wall hydrolases.

Growth and Survival of Genetically Manipulated Lactobacillus plantarum in Silage

The growth and persistence of two genetically manipulated forms of Lactobacillus plantarum NCDO (National Collection of Dairy Organisms) 1193 have been monitored in grass silage. Both recombinants contained pSA3, a shuttle vector for gram-positive organisms that encodes erythromycin resistance. In one of the recombinants, pSA3 was integrated onto the chromosome, whereas in the other, a pSA3 derivative designated pM25, which contains a Clostridium thermocellum cellulase gene cloned into pSA3, was maintained as an extrachromosomal element. This extrachromosomal element is a plasmid. Rifampin-resistant mutants were selected for the recombinants and the parent strain. When applied to minisilos at a rate of 10 6 CFU/g of grass, both the recombinants and the parent strain proliferated to dominate the epiphytic microflora and induced an increase in the decline in pH compared with that of the noninoculated silos. The presence of extra genetic material did not appear to disadvantage the bacterium in comparison with the parent strain. The selective recovery of both strains by using rifampin and erythromycin was confirmed by Southern hybridization. Interestingly, the free plasmid (pM25) appeared more stable in silage than was expected from studies in MRS broth. The plasmid was retained by 85% of the rifampin-resistant L. plantarum colonies isolated from a day 30 silo. These data answer an important question by showing that genetically manipulated recombinants of L. plantarum can proliferate and compete with epiphytic lactic acid bacteria in silage.

Characterization of the gene celD and its encoded product 1,4-beta-D-glucan glucohydrolase D from Pseudomonas fluorescens subsp. cellulosa

A genomic library of Pseudomonas fluorescens subsp. cellulosa DNA constructed in pUC18 and expressed in Escherichia coli was screened for recombinants expressing 4-methylumbelliferyl beta-D-glucoside hydrolysing activity (MUGase). A single MUGase-positive clone was isolated. The MUGase hydrolysed cellobiose, cellotriose, cellotetraose, cellopentaose and cellohexaose to glucose, by sequentially cleaving glucose residues from the non-reducing end of the cello-oligosaccharides. The Km values for cellobiose and cellohexaose hydrolysis were 1.2 mM and 28 microM respectively. The enzyme exhibited no activity against soluble or insoluble cellulose, xylan and xylobiose. Thus the MUGase is classified as a 1,4-beta-D-glucan glucohydrolase (EC 3.2.1.74) and is designated 1,4-beta-D-glucan glucohydrolase D (CELD). When expressed by E. coli, CELD was located in the cell-envelope fraction; a significant proportion of the native enzyme was also associated with the cell envelope when synthesized by its endogenous host. The nucleotide sequence of the gene, celD, which encodes CELD, revealed an open reading frame of 2607 bp, encoding a protein of M(r) 92,000. The deduced primary structure of CELD was confirmed by the M(r) of CELD (85,000) expressed by E. coli and P. fluorescens subsp. cellulosa, and by the experimentally determined N-terminus of the enzyme purified from E. coli, which showed identity with residues 52-67 of the celD translated sequence. The structure of the N-terminal region of full-length CELD was similar to the signal peptides of P. fluorescens subsp. cellulosa plant-cell-wall hydrolases. Deletion of the N-terminal 47 residues of CELD solubilized MUGase activity in E. coli. CELD exhibited sequence similarity with beta-glucosidase B of Clostridium thermocellum, particularly in the vicinity of the active-site aspartate residue, but did not display structural similarity with the mature forms of cellulases and xylanases expressed by P. fluorescens subsp. cellulosa.

Constitutive secretion of a bacterial enzyme by polarized epithelial cells

The constitutive (or default) pathway for protein secretion was investigated in two epithelial cells, Madin-Darby canine kidney (MDCK) and human colonic adenocarcinoma (Caco-2), using a bacterial enzyme. The choice of a bacterial protein was based on the requirement to identify a protein devoid of sorting signals. The sorting of a bacterial endoglucanase derived from Clostridium thermocellum, endoglucanase E, from stably transfected MDCK and Caco-2 cells was examined. The choice of a bacterial endoglucanase for these studies has advantages of simple, sensitive and quantitative detection, while higher eukaryotic cells do not express endoglucanase activity. Both cell lines secreted a 50 kDa form of the bacterial protein, while smaller intracellular forms were also observed. In polarized layers of MDCK cells the endoglucanase was secreted into both membrane domains in the ratio 62% apical and 38% basolateral. In Caco-2 cells secretion was predominantly, 70%, through the basolateral membrane. These results define the constitutive pathway for protein secretion in these two model epithelial cells.

The use of chimeric gene constructs to express a bacterial endoglucanase in mammalian cells

The synthesis and secretion of a truncated Clostridium thermocellum endoglucanase (EGE') encoded by the celE' gene was investigated in Chinese hamster ovary (CHO) cells. Fusion genes consisting of the human growth hormone (hGH) gene and celE', transcribed from the SV40 early enhancer/promoter, were constructed and stably transfected into CHO cells. A gene consisting of celE' inserted into the first exon of the hGH gene resulted in the synthesis of truncated proteins (less than or equal to 22 kDa) lacking endoglucanase activity. Cloning celE' into the second exon of the hGH gene, resulted in the synthesis and secretion of a 50 kDa protein with endoglucanase activity. A 50 kDa protein was also synthesised by cells transfected with celE' cloned into the fifth exon of the hGH gene. However, despite a 5-fold increase in enzyme activity compared to the exon 2 transfected cell line less than 40% of the protein was secreted. Constructs devoid of introns, in which celE' was fused to the SV40 early promoter and to the rabbit beta-globin polyadenylation sequence resulted in a 2-18-fold increase in endoglucanase activity compared to the constructs containing introns. In addition more than 75% of the synthesised protein was secreted. Analyses of EGE' encoded mRNA from the transfected cell lines suggests that the presence of introns results in the aberrant splicing of message by the use of cryptic splice sites in the celE' gene. These results demonstrate that introns are not required for the efficient expression of a bacterial endoglucanase in mammalian cells, rather introns appear to reduce expression of the encoded protein.