Ley 387 de 1997

The λ Red recombinase system was employed, in an attempt to introduce the fusA(G1478T)

mutation into a naïve E. coli MDS42 background. E. coli MDS42 is a stable strain with a

low propensity for recombination (Chakiath & Esposito, 2007); therefore, the first test was to determine whether the λ Red system would work in this strain. To do this, the metC gene

in E. coli MDS42 and in another strain, E. coli BW25113, which served as a positive control

was knocked out (Section 6.4.1.1). E. coli strain BW25113 was selected as the positive

control because a previous study had successively inactivated genes using the O Red system in this strain (Baba et al., 2006). To disrupt metC, a 1.4 kb kanamycin resistance cassette

derived from the E. coliBW25115 ∆metC knockout strain from the Keio collection (Baba

et al., 2006) that had been amplified by PCR was used. The kanamycin resistance marker was flanked by 54 bp and 47 bp sequences that were homologous to the 5´ and 3´ ends of the metC gene, respectively.

The metC gene knockout experiments were performed and the Red recombination

protocol yielded 41 and ~ 500 colonies of E. coli MDS42 and BW25113, respectively. To

confirm that the colonies were true positives, colony PCR was performed on three colonies randomly selected off each plate using primers that bound upstream of the metC gene and

to the kanamycin resistance marker. PCR yielded a 750 bp PCR product confirming that the metC gene had successfully been replaced with the kanamycin marker (Figure 16).

Figure 16. Successful recombination using the kanamycin resistance marker

The λ Red recombinase system was used to introduce the kanamycin resistance marker into E. coli MDS42

(lanes 1-3) and BW25113 cells (lanes 7-9). A successful recombination event yielded a 750 bp product after performing PCR. Control samples (lanes 4-6 and 10) represent cells that have undergone the same procedure but in the absence of the kanamycin resistance marker. A no template control (NTC) was included in the PCR reaction.

Since recombination was successful in E. coliMDS42 using the λ Red system, the

next step was to proceed with constructing the fusA(G1478T) mutation in this strain. In

this approach, using two primer pairs specific for the fusA gene, fusA_A and fusA_B (Table

I.3, Appendix I.5), two regions of different sizes were amplified by PCR, 584 bp and 816 bp, respectively, from the E. coli MDS42 yejG (HTD) clone. Two regions of different sizes

were amplified in order to investigate whether the length of the DNA would affect recombination efficiency. Both PCR products contained the fusA(G1478T) mutation.

The PCR products were introduced into E. coliMDS42 cells expressing the λ Red

proteins. For the negative control, E. coli MDS42 strain expressing the Red proteins were

electroporated in the absence of DNA. It was possible to select for clones that had acquired the mutation because E. coli MDS42 cells that harbour the fusA(G1478T) mutation are

resistant to higher concentrations of tobramycin than cells that do not carry the mutation (MIC values of 24 μg.mL-1 _{and 12 μg.mL}-1_{, respectively; Section 4.3.3). The next day, 2}

candidate colonies were obtained for primer fusA_A and 17 colonies were obtained when primer fusA_B was used. However, the negative control plate contained 7 colonies. Therefore, the next step was to check whether any of the colonies contained the

To check whether recombination had been successful, a PCR discrimination assay was used (Figure 17). Two primer pairs were designed that were specific for the mutated

fusA allele (fusA_m) and wild-type fusA (fusA_w). For the fusA_m primer pair, the 3´ end

of the reverse primer ended with base A, at a position that matched in sequence to the G1478T mutation (Figure 17). If the clone carried the mutation in the fusA gene, then the

PCR would yield a 477 bp PCR product. For the fusA_w primer pair, the 3´ end of the forward primer ended with G. A 399 bp PCR product was amplified if the gene was wild- type. All of the colonies that grew on the non-control plates were tested by PCR discrimination using the fusA_m and fusA_w primers. The results showed that only the primer pair specific for the wild-type fusA gene yielded bands for all of the clones. For the

positive control, the PCR discrimination assay was also performed on the E. coli MDS42 yejG (HTD) clone whose DNA was used as a template to generate the PCR products used

in the recombination reaction. This PCR yielded a band with the expected size (~ 500 bp). This suggested that recombination had not been successful in any of the 19 colonies tested.

The experiment was repeated but this time using fragments of fusA that were similar

or longer in DNA size to the kanamycin resistance marker (1.4 kb). Another set of primers specific to fusA, fusA_C and fusA_D (Table I.3, Appendix I.5), were used to amplify 1,482

bp and 2,211 bp PCR products, respectively from the E. coli MDS42 yejG (HTD) clone.

When the recombination experiment was performed, two colonies were present on one (out of two) of the control plates but there was no growth on the sample plates.

Figure 17. Designing primers for PCR discrimination to detect the fusA(G1478T) mutation

The diagram shows the design of the fusA_w (19 bp) and fusA_m (19bp) primer pairs that are specific for the wild-type and mutated fusA gene (2115 bp), respectively. The fusA_m primer pairs generated a 477 bp

As a result of the negative outcome, attempts were made to optimise the Red recombination protocol. One such change was decreasing the recombination time from 3 h to 2.5 h, 40 min and 20 min. The reason for this step was to decrease the time available for cells that had not recombined from outgrowing the minority that had successfully recombined. This way, it was more likely to select for the desired clone from the aliquots of cells spread on plates. Another modification was to try a different tobramycin concentration (20 μg.mL-1) for the selection stage. This was to inhibit the growth of cells that had acquired a chromosomal mutation that conferred a lower level of tobramycin resistance than that conferred by the fusA(G1478T) mutation (24 μg.mL-1). Finally, clones were selected in the presence of the different aminoglycosides (sisomicin and apramycin). Despite these efforts, none of a total of 76 clones analysed had successfully incorporated the G1478T mutation.

As a last resort, the ‘Quick & Easy E. coli Gene Deletion Kit’ (Gene Bridges

GmbH), which uses the Red/ET system, was employed. As mentioned earlier, the Red/ET system is similar to the λ Red system. This kit was used because according to the manufacturers, it provided an optimized system for disrupting and modifying genes on the

E. coli chromosome. As before, a test was carried out to confirm that the Red/ET system

could work in E. coli MDS42. Again, the metC gene was disrupted in the E. coli MDS42

and BW25113 strains with the 1.4 kb kanamycin resistance marker. PCR screening of three randomly picked colonies of each strain confirmed that all six colonies had successfully recombined.

A different strategy was used for this experiment to see whether this would yield the desired result. Ellis and colleagues (2001) showed that it was possible to create point mutations with the Red recombinase system using single-stranded DNA. They and others (Wang et al., 2009; Costantino & Court, 2003) have also shown that higher recombination efficiencies can be obtained with the oligonucleotide that corresponds in sequence to the lagging strand. The model for single-stranded recombination is that upon entering the cell, the RecT protein binds to the oligonucleotide to protect it from nuclease degradation (Karakousis et al., 1998). The binding of RecT also mediates the annealing of the oligonucleotide to the lagging strand at transiently single-stranded regions at the replication

including DNA replication (Ellis et al., 2001). The oligonucleotide is then joined to the surrounding Okazaki fragments by DNA ligase.

For this experiment, a 71 nucleotide synthetic single-stranded oligonucleotide, including the G1478T mutation at base 36 flanked by 35 nucleotides of wild-type fusA was

designed as the substrate for recombination. A 71-base DNA sequence was chosen because oligonucleotides of similar length have been successfully used for recombination in other studies (Costantino & Court, 2003; Ellis et al., 2001). To increase the recombination efficiency, the oligonucleotide included four phosphorothioate bonds located at the 5´ terminus to prevent in vivo degradation of the synthetic DNA by endogenous exonucleases

(Wang et al., 2009). To determine the ‘lagging strand’ sequence, the direction of transcription of the fusA gene relative to the direction of replication in E. coli from the

origin towards the terminus was identified (Figure 18).

Figure 18. Designing single-stranded oligonucleotides that correspond to the lagging strand

E. coli chromosome (represented as a circle) is replicated bi-directionally from the origin of replication (oriC)

toward the terminus (ter), as indicated by black arrows. The direction of transcription of the fusA gene (red

arrow) is shown. Thus the leading and lagging strand sequences can be inferred from the orientation of fusA

relative to OriC.

E. coli MDS42 cells expressing the Red proteins were transformed with the

phosphorothioate-containing single-stranded oligonucleotide. A control expressing the Red proteins was electroporated in the absence of the phosphorothioate-containing DNA. Following recombination and overnight incubation on tobramycin-containing plates, there were 4, 0 and 1 colony on the sample plates and 6 colonies on the control plate. Despite the

low number of colonies that grew, all of the 5 colonies that grew on the sample plates were screened using the PCR discrimination assay. However, recombination had not been successful in any of the clones. The experiment was repeated several times with many of the steps altered but none of the clones analysed had successfully incorporated the mutant sequence. Attempts were also made to introduce the mutation into E. coli strains DH5α-E

and BW25113 using this system. Over 200 colonies were analysed, however, none of clones investigated had acquired the fusA(G1478T) mutation. The PCR discrimination

assay was performed successfully on clone E. coli MDS42 yejG (HTD), indicating that the

assay was working (Figure 19).

Finally, the experiment was repeated using the λ Red system with the 584 PCR product [that included the fusA(G1478T) mutation] in naïve E. coli MDS42 cells that were

expressing YejG. This was to test whether YejG might facilitate acquiring the fusA

mutation. However, the clones that appeared on the plates containing tobramycin were false positives.

Figure 19. Gel of PCR discrimination assay

PCR discrimination assay was performed on E. coli MDS42 cells after attempted recombination with the

single-stranded oligonucleotide, using the Red/ET system (lanes 1-3). A 477 bp product was present after PCR if the fusA(G1478T) mutation was present in the chromosome, indicating that recombination was