Replication of the linear plasmids of Kluyveromyces lactis: Analysis of the putative DNA polymerases.

The dairy yeast Kluyveromyces lactis contains two cytoplasmic linear plasmids pGKL1 (K1) and pGKL2 (K2) which are associated with a killer and immunity phenotype. DNA sequence analyses have revealed that both K1 and K2 contain ORFs (ORFs 1 and 2, respectively) with the potential to encode Class B DNA polymerases (Dpols). This is consistent with a growing body of evidence which indicates that many 'linear plasmids' are dependent for their replication upon self-encoded DNA and RNA polymerases. Attempts to over-express the native Dpol genes of both K1 and K2, in E. coli, were largely unsuccessful. Fragments, ranging from 390 bp to the entire 3 kb gene, were cloned into a variety of expression vectors, but no protein products were observed. The lack of expression arose due to problems at the transcriptional, translational and post-translational level, reflecting the difficulty E. coli probably experiences expressing DNA of such a high A+T content. Experiments using gene fusions revealed a dramatic reduction in the level of product yield when the native K-plasmid DNA was coupled to the highly expressed amino terminal stabilising moiety. In vitro transcription data also revealed that transcription of the native genes appeared to be prematurely terminating, probably due to the occurrence of fortuitous transcriptional terminator sequences in the A+U rich mRNA. A 168 bp fragment of the extreme 5' end of the putative K2 Dpol was chemically resynthesised, incorporating an optimal codon bias for high level expression in E. coli. The gene fragment was designed such that a second section of the gene could easily be added at a later date, doubling the size of the potential product. This fragment was cloned into a fusion vector which directed the expression of heterologous genes as C-terminal fusions with the 27.5 kDa enzyme glutathione S-transferase (GST). Upon induction, strains bearing this plasmid expressed the GST-Dpol fusion protein to over 10% of total cellular protein. The addition of the extra 54 amino acids to the GST was, however, sufficient to render most of the fusion product insoluble. In the absence of a stabilising conjugant peptide, the small resynthesised gene was transcribed, but a protein product failed to accumulate. Peptides corresponding to two potentially antigenic sites, within the N-terminus of the K2 Dpol, were chemically synthesised. The two 8-residue peptides were coupled to tetravalent Multiple Antigen Peptide cores, and were used to immunise chickens. The peptides, in this form, however, failed to elicit an immune response from the recipients. The presence of covalently attached terminal proteins and the cytoplasmic location of these plasmids has severely impaired the ability to manipulate these plasmids. However, homologous recombination in vivo has been developed and used to incorporate selectable markers, to disrupt plasmid-borne genes and to re-introduce modified versions of endogenous genes. An affinity tag, consisting of six adjacent histidine residues was incorporated into the 3' end of the putative K1 Dpol gene. The modified gene was then coupled to a selectable marker and successfully re-introduced into K1. A K. lactis nuclear vector was constructed which expressed an anti sense RNA complementary to the extreme 5' end of the putative K1 Dpol gene. Despite constitutive high level expression of this RNA, the replication of plasmid K1 appeared to be unaffected.