Production and secretion of heterologous proteins by Lactococcus lactis

Lactococcus lactis strains have been used for centuries in food fermentation, now appreciated as traditional biotechnology. They have been applied in the cheesemaking process and for the manufacturing of other dairy products. Years of experience with these lactic acid bacteria have led to a profound understanding of the microbiological and technological aspects of L.lactis. Recent progress in the genetics of L. lactis made this organism a suitable candidate for the use in modern biotechnology as a host for the production of homologous, heterologous, or engineered proteins. The purpose of the research, described in this thesis, was to investigate the capacity of L. lactis to produce and, in particular, to secrete heterologous proteins.

Chapter 1 presents a brief overview of the knowledge about gene expression and secretion systems in procaryotes, with specific attention to heterologous gene expression in L. lactis.

In Chapter 2 the characterization of Usp45, the major extracellular protein of L. lactis is described. The aim of isolating the usp45 gene was to use its promoter, ribosome binding site and export signal for the development of an expression and secretion system for L.lactis. Determination of the nucleotide sequence of the gene revealed several specific characteristics of the Usp45 protein and its gene. i) The primary amino acid sequence of the protein revealed an unusual amino acid composition and did not show homology with any protein of known function. ii) The mature protein starts at amino acid residue 28. It is preceded by a leader peptide which contains 3 possible signal peptidase I cleavage sites, after ala ¹⁹, ala ²⁰and ala ²⁷. iii) The open reading frame encoding Usp45 is preceded by two Shine and Dalgarno (SD) sequences, both with uncommon properties. iv) The A/T content of the region 100 nucleotides upstream of the usp45 promoter is significantly (10%) higher than that found in L. lactis coding sequences. All these characteristics have been investigated.

No function could be postulated for Usp45. Its deduced amino acid sequence showed 30% homology with that of P54 of Enterococcus faecium, but the function of this protein is also unknown. However, a weak cross reaction has been observed of antibodies raised against surface layer (SL) proteins from Lactobacillus helveticus with the Usp proteins from several L. lactis strains, and of the Usp45 antibodies with the surface proteins of Lactobacillus plantarum NCFB 1988. This could indicate that L. lactis Usp proteins are related to SL-proteins.

Several attempts to inactivate the chromosomal usp45 gene have been performed. (15). Various plasmids have been constructed in which the usp45 homologous region carried deletions or several mutations in the 3' and 5' region of the structural gene. Campbell-like integration of these plasmids would result in truncation of the usp45 gene. Furthermore, plasmids have been constructed for the inactivation of the usp45 gene by replacement recombination. Although various studies have shown that homologous recombination can be used for gene inactivation in L. lactis none of the tested strategies resulted in the inactivation of the usp45 gene, suggesting that usp45 is an essential gene in L. lactis.

Chapter 3 describes the construction of an expression and secretion system, based on the usp45 gene, which was evaluated with the prtP gene of L. lactis SK1 1 and the amyS gene of B. stearothermophilus as reporter genes. Fusions in which the 27 amino acid leader peptide directed the secretion of the homologous proteinase, were used to show that the usp45 leader is sufficient for the efficient secretion of PrtP. In addition, to examine the functionality of two shorter leader peptides (aa 1-19 and aa 1-20) as signal peptides, plasmids encoding the first 19, 20 or 27 residues of Usp45 fused to the mature α-amylase were introduced in E. coli and L. lactis. In E. coli these plasmids resulted in secretion of active α-amylase into the periplasm and even into the external growth medium. However, in L. lactis only the 27-residue leader peptide functioned as an export signal. These results provide experimental evidence for the postulated difference in length between signal peptides from Gram-positive and Gram- negative organisms. When the Usp45 signal peptide was used for the secretion of the α-amylase, less than 50% of this reporter protein was located in the extracellular medium. The remaining fraction was present as an unprocessed, inactive precursor in the intracellular fraction.

Analysis of the usp45 promoter region is described in Chapter 4 . Several usp45 - amyS gene fusions were constructed and introduced on plasmids or in the chromosome of L. lactis . These gene fusions were used to demonstrate the role in transcription of the A/T-rich region immediately upstream of the usp45 -35 region. The highest levels of α-amylase production were obtained when the usp45 - amyS fusion was located at the position of the usp45 gene in MG1363. These data showed that a large DNA region of more than 200 bp upstream of the -35 region was of major importance for expression. Similar results have been reported by Van Rooijen et al. (20) for the lac promoter. This could indicate a general role in transcription activation for the region upstream of lactococcal promoters. The L. lactis strains harboring the various usp45 - amyS fusions produced sufficient α-amylase activity to allow growth on media containing starch as a sole energy source.

Chapter 5 describes the effect of the translational initiation region on expression of the usp45 - amyS gene fusion. Both postulated SD regions SD1, an extremely weak ribosome binding site 6 bp upstream of the ATG start codon, and SD2, with a higher complementarity to the 3' end of the 16S rRNA, but 21 bp upstream of the ATG codon, were altered in the usp45 - amyS secretion vector. These studies revealed that translation of the usp45 - amyS fusion is possible from both SD regions. Deletion of either one of the SD regions resulted in normal (x-amylase expression, suggesting that L. lactis tolerates SD-sequences with low complementarity to the 3' end of 23S rRNA, and ribosome binding sites with a 21 bp spacing. A reduction of the spacing between SD2 and the AUG start codon, resulted in a ribosome binding site which corresponds to the consensus in L. lactis, with respect to both spacing and free energy. However, this ribosome binding site did not lead to an significant increase in α-amylaseproduction.

To enhance the level of heterologous gene expression, a search for stronger promoters was initiated. Several studies have been performed to isolate strong lactococcal promoter sequences, using promoter probe vectors. One of the isolated promoters resulting from these studies, has already been used in a heterologous expression system for the production of hen egg lysozyme and B. subtilis neutral protease. We have used different approaches to isolate strong promoters.

 
Based on the N-terminal sequence of an abundant intracellular 30 kDa protein, oligonucleotide probes were designed and used to screen a genomic library from L. lactis. This resulted in the isolation of the 1-kb Hind III -Pst I fragment containing an open reading frame of 612 nucleotides that could encode a polypeptide of 204 aa (Fig. 1). The determined N-terminus was found at at position 2-11 of the deduced amino acid sequence, indicating that the N-terminal methionine had been removed in L. lactis . Homology analysis revealed that the amino acid sequence showed considerable homology with the aminoterminal part of glyceraldehyde 3-phosphate dehydrogenase from B. subtilis , E. coli , yeast, mouse and humans. Based on this homology and on the molecular weight of the L. lactis protein it could be calculated that the cloned fragment lacked approximately 350 nucleotides of the 3' part of the glyceraldehyde 3-phosphate dehydrogenase gene. Upstream of the ATG start codon a putative SD sequence was located. The complementarity of this SD sequence to the 3' end of the L. lactis 16S rRNA was 74.76 kJ/mol, which is the highest value found untill now in L. lactis . The start of transcription was determined and revealed that the promoter (TTTGCA-16bp-TAAAAT-7bp-T) differed at 3 positions with the consensus for lactococcal promoters (Fig. 1). A small inverted repeat is present in the nontranslated 5' part of the mRNA. It remains unclear whether this repeat plays a role in the high expression of the gene.

Another approach to isolate strong promoters was based on the screening of a L. lactis genomic library with antibodies raised against abundant intracellular proteins. Interestingly this resulted in the cloning of the dnaJ gene as described in Chapter 6 . The dnaJ gene is one of the first heat shock genes characterized in L. lactis. Investigation of the promoter region showed that heat shock regulation in L. lactis, and very likely in other Gram-positive organisms, is not achieved by an alternative sigma factor as is the case in E. coli . An inverted repeat which is highly conserved in the promoter region of heat shock genes from Gram-positive organisms, is responsible for the repression of transcription of the dnaJ gene at non-stress conditions. The dnaJ promoter was used in the usp45-amyS fusion, and a 2-4 fold induction of α-amylase was accomplished after heat shock. However, the dnaJ promoter did not result in higher expression of the α-amylase, as compared to the usp45 promoter.

In the last decennium a lot of research has been performed on the exploration of several microorganisms, such as yeast, fungi, E. coli and B. subtilis as a potential host for heterologous protein production and secretion. Initially, B.subtilis appeared to be a good organism for this purpose. It is a Gram-positive organism, capable of secreting large amounts of proteins into the extracellular medium. In addition, it can be cultivated in large amounts, at low costs. One of the main obstacles that limits the application of B. subtilis as a production host is its high extracellular proteolytic activity, resulting in degradation of the heterologous proteins of interest. L. lactis could be a suitable alternative. It is also a Gram-positive organism. However, it has a low extracellular proteolytic activity and Prt strains are available. Furthermore, it possesses the additional desired features which make this organism a suitable candidate for heterologous gene expression. It is a safe, non-pathogenic organism and has a widespread use in the food industry. These properties have stimulated investigation on the use of L. lactis as a production organism. The overproduction of several homologous proteins has now been accomplished in L. lactis and the production of several heterologous proteins has been established (Chapter I). However, the secretion of heterologous proteins in L. lactis is inefficient (Chapter III). The investigation on the secretion of proteins in L. lactis has been initiated (Chapter II and Chapter III) and the results invite for continuing these investigations.

Besides a view on the several approaches which can be used in these investigations, the work presented in this thesis has yielded a set of expression and secretion vectors, which could be employed to express heterologous genes in L. lactis. Furthermore they can be used in the further unravelling of the expression and secretion mechanism of L. lactis.