Vector for Positive Selection of In-Frame Genetic Sequences

Polymerase chain reaction (PCR)based approaches are becoming increasingly important for the identification of members of extended multigene families as well as homologous gene structures present in phylogenetically divergent species (2,5,10,11). Many of these approaches rely on the use of highly degenerate primers and/or reduced priming stringencies that can generate a broad range of products, including significant numbers of amplification products that contain frameshift(s) and/or termination codon(s). Recently, we introduced the use of short, minimally degenerate primers complementing conserved structural motifs for PCR amplification of homologs of antigenbinding receptor genes in phylogenetically diverse species (6,8–10). This approach is also associated with the generation of amplification artifacts that require DNA sequencing to be distinguished from products that warrant further study. To facilitate identification of amplification products containing open reading frames, we have engineered a vector, pGFPfs, that affords positive selection of recombinants based on the continuity of coding sequence within a lacZ:insert:GFP (green fluorescent protein) fusion construct that is expressed in E. coli. pGFPfs was derived from pGFPuv (CLONTECH Laboratories, Palo Alto, CA, USA) by ligating a linker (formed by annealing the partially complementary oligonucleotides: 5′-GATCGATATCTCGAGT-3′ and 5′-CTAGACTCGAGATATCGATCTGCA-3′) into the multiple cloning site (MCS) of PstI/ XbaI-digested pGFPuv. Incorporation of this linker, which was confirmed by DNA sequencing, disrupts the GFPuv reading frame and introduces additional unique restriction sites (Figure 1A). The GFP variant in these vectors, GFPuv, differs from wild-type GFP (7) in that it is optimized for bacterial expression, solubility and fluorescence of the isopropyl-β-D-thiogalactopyranoside (IPTG)-inducible lacZ:GFPuv fusion protein (3). Colonies producing GFPuv fusion proteins are identified by viewing under longwave UV light. Positive selection of recombinants containing in-frame inserts is based on correction of a frameshift that has been engineered within the 5′ MCS of pGFPfs. PCR products were generated using methods that have been described previously by our laboratory (8–10) and were ligated into HindIII/XbaI-digested pGFPfs. Ligation mixtures were used to transfect E. coli DH5αF′ (Life Technologies, Gaithersburg, MD, USA). Cells were plated on LB plates containing 10 mM IPTG and 100 μg/mL ampicillin, and colonies were grown at 30°C for 48 h. Plasmid DNA was isolated for sequencing using a QIAprep Spin Miniprep Kit (Qiagen, Chatsworth, CA, USA) in accordance with the manufacturer’s recommended protocol. Sequencing was performed with a 4000L Automated Sequencer (LI-COR, Lincoln, NE, USA) using a SequiTherm LongRead Cycle Sequencing Kit (Epicentre Technologies, Madison, WI, USA). Plates were viewed using a UVL 56 Blak-Ray (366 nm) longwave ultraviolet lamp (UVP, Upland, CA, USA). IPTG induction of pGFPfs generates a 17-amino acid protein that is encoded by a fragment of lacZ, the contiguous cloning sites and a 3′ termination codon (Figure 1A); GFPuv is not expressed. Because cloning sites are 5′ of the GFPuv coding sequence, expression of GFPuv requires frame continuity within an insert. Therefore, directional cloning of PCR-derived products that are devoid of stop codons can be used to correct the reading frame to that of the GFPuv coding sequence (Figure 1B). Primers incorporating restriction sites can be designed to either remove the frameshift (e.g., HindIII/KpnI cloning) or introduce a second frameshift within the 3′ primer (e.g., HindIII/PstI cloning) to reestablish the correct reading frame and allow expression of GFPuv. The capacity of pGFPfs to discriminate between coding and noncoding amplification products is dependent on noncoding products having an internal stop codon(s) or shifted reading frame; i.e., ±1 base. It is apparent that 66% of noncoding products will not have the