How Escherichia coli can bias the results of molecular cloning: preferential selection of defective genomes of hepatitis C virus during the cloning procedure

Abstract

Cloned PCR products containing hepatitis C virus (HCV) genomic fragments have been used for analyses of HCV genomic heterogeneity and protein expression. These studies assume that the clones derived are representative of the entire virus population and that subsets are not inadvertently selected. The aim of the present study was to express HCV structural proteins. However, we found that there was a strong cloning selection for defective genomes and that most clones generated initially were incapable of expressing the HCV proteins. The HCV structural region (C-E1-E2-p7) was directly amplified by long reverse transcription-PCR from the plasma of an HCV-infected patient or from a control plasmid containing a viable full-length cDNA of HCV derived from the same patient but cloned in a different vector. The PCR products were cloned into a mammalian expression vector, amplified in Escherichia coli, and tested for their ability to produce HCV structural proteins. Twenty randomly picked clones derived from the HCV-infected patient all contained nucleotide mutations leading to absence or truncation of the expected HCV products. Of 25 clones derived from the control plasmid, only 8% were fully functional for polyprotein synthesis. The insertion of extra nucleotides in the region just upstream of the start codon of the HCV insert led to a statistically significant increase in the number of fully functional clones derived from the patient (42%) and from the control plasmid (72-92%). Nonrandom selection of clones during the cloning procedure has enormous implications for the study of viral heterogeneity, because it can produce a false spectrum of genomic diversity. It can also be an impediment to the construction of infectious viral clones.

Figure 1

Sequence analysis of eight clones derived from the patient. Amplification was performed with primer A0 (45-mer). Black boxed regions indicate putative translated HCV products. Continuous lines indicate nonsilent mutations, and discontinuous lines indicate silent mutations, compared with the consensus sequence of HCV H77 (24).

Figure 2

Sequence analysis of clones derived from the control plasmid. (A) Eight clones derived from the control plasmid with primer A0 (38-mer). (B) Ten clones derived from the control plasmid with primer A+1. Black boxed regions indicate putative translated HCV products. Continuous lines indicate nonsilent mutations, and discontinuous lines indicate silent mutations, compared with the control plasmid sequence (24). The last clone of A contains the mutation within the start codon (AUG to GUG). This mutation, theoretically, could result in a low level of translation.

Figure 3

Transcription–translation of clones derived from the control plasmid. Amplification with primer A0 (38-mer) (A) or primer A+3 (B). First lane, positive control (a clone with the consensus amino acid sequence of H77). Remaining lanes, representative clones. Arrow, position of the expected full-length product (≈85 kDa).