Ley 975 de 2005

This section explains the rationale behind the construction of specific clones that were used in the fitness assays, to investigate whether the fusA(G1478T) mutation imparts a

fitness cost and to assess the role of YejG in cell fitness. Specifically, the following questions were investigated:

i) Does harbouring the fusA(G1478T) mutation incur a fitness cost?

ii) Is there a benefit to over-expressing YejG in a clone with the fusA(G1478T)

mutation?

iii) Does over-expressing YejG alleviate any fitness cost the fusA(G1478T)

mutation may impart?

iv) Does over-expressing the mutated version of YejG [YejG(HTD)] improve or alleviate any fitness cost the fusA(G1478T) mutation may impart?

i) Does harbouring the fusA(G1478T) mutation incur a fitness cost?

To address these questions, a clone with the fusA(G1478T) mutation and another with

the wild-type fusA gene were required. As it was not possible to construct a clone with the fusA mutation (Section 6.2), fitness assays were performed with the E. coli MDS42 yejG

(HTD) clone [carries the fusA(G1478T) mutation and the pCA24N-yejG(HTD) plasmid].

However, the aim was to investigate whether the fusA(G1478T) mutation incurred a fitness

cost in the following conditions; (i) in the absence of aminoglycoside and (ii) without the presence of plasmid-encoded YejG or the mutated version of YejG. Therefore, experiments were carried out to replace pCA24N-yejG(HTD) with pBlueScript, since it had also proven

impossible to cure cells of the pCA24N plasmid (Section 6.4.3.3).

The pBlueScript plasmid (Table I.2, Appendix I.4) was useful for replacing pCA24N for several reasons: it did not contribute to aminoglycoside resistance; it carried a different

antibiotic marker (ampicillin) to the pCA24N plasmid; and it turned colonies blue in the presence of Xgal, allowing for blue/white screening. If pBlueScript were to be used in the fitness experiment, it would have to be carried by both the strains with wild-type fusA,and

the fusA(G1478T) allele. However, this would mean that all of the cells would form blue

colonies on Xgal plates, making it impossible to distinguish between the two clones. To solve this problem a pBlueScript plasmid with a non-functional lacZ gene (pWhite) was

constructed.

An arbitrarily chosen non-functional fragment of the tdcD gene (535 bp) was cloned

into the middle of lacZD located on the pBlueScript plasmid (Section 6.4.3.2) to create

pWhite (Table I.2, Appendix I.4). As a result, the lacZD fragment was disrupted and

became out of frame, no longer enabling the formation of blue colonies in the presence of Xgal. I now had a plasmid that did not contribute to aminoglycoside resistance, had a different resistant marker to pCA24N, did not have a functional lacZ and could displace

pCA24N-yejG(HTD). Now it was trivial to mark one strain with a functional copy of lacZ.

As described above, if both the clones were transformed with pWhite, then all of the cells would form white colonies on Xgal plates, again making it impossible to differentiate between the two clones. To resolve this issue, one of the clones (the E. coli MDS42 strain

with wild-type fusA) was marked with a functional chromosomal copy of lacZ (Section

6.4.3.1) before being transformed with the pWhite plasmid. In the end, to investigate whether the fusA(G1478T) mutation incurred a fitness cost, the following clones were used: E. coli MDS42 with fusA(G1478T) mutation + pWhite versus E. coli MDS42 with wild-

type fusA + lacZ + pWhite.

The results revealed that the fusA(G1478T) mutation imparts a statistically significant

reduction in fitness (W = 0.93 ± 0.01; P < 0.01). This implies that harbouring this mutated fusAallele in the absence of aminoglycoside is detrimental to the cell’s growth. Although

it was not surprising that this mutation incurred a fitness cost, given the difficulty in creating a fusA mutated clone, a larger fitness decrease was expected.

ii) Is there a benefit to over-expressing YejG in a clone with the fusA(G1478T)

mutation?

To determine whether there was a benefit to over-expressing YejG in a clone with the

fusA(G1478T) mutation, the E. coli MDS42 yejG (HTD) clone [with the fusA(G1478T)

mutation + pCA24N-yejG(HTD)] was first transformed with pBlueScript, to displace the

existing plasmid. After confirming that the cells had lost pCA24N-yejG(HTD) (the cells

were ampicillin resistant and chloramphenicol sensitive), they were transformed with either pCA24N-yejG or pCA24N-GFP. To check whether the clones had lost the pBlueScript

plasmid, the cells were spread on LB agar containing ampicillin to confirm that they were susceptible. The strain expressing GFP was marked with a functional copy of lacZ so that

the two clones could be distinguished.

The fitness assays showed that there was a significant fitness benefit to over-expressing YejG in a strain with the fusA(G1478T) mutation (W = 1.18 ± 0.03; P < 0.01). To follow

on from this, it was investigated whether this benefit might include alleviating the fitness cost imparted by the fusA(G1478T) mutation.

iii) Does over-expressing YejG alleviate any fitness cost the fusA(G1478T)

mutation may impart?

The developing hypothesis was that YejG could alleviate the fitness cost associated with the fusA(G1478T) mutation. This was tested by performing a competitive fitness assay

between cells with the fusA(G1478T) mutation + pCA24N-yejG (the same clone used in

(ii) above) and a strain with the wild-type fusA gene + pCA24N-GFP. The strain with the

wild-type fusA gene was marked with a functional copy of lacZ.

Surprisingly, the results revealed that the strain harbouring the fusA(G1478T) mutation

and over-expressing YejG was significantly fitter than the strain with the wild-type fusA

gene and no YejG over-expression (W = 1.46 ± 0.05; P < 0.01). It would appear that YejG

not only alleviates the fitness cost incurred by the fusA(G1478T) mutation, but it also

improves the fitness. However, since these results relate to wild-type YejG, but the E. coli

mutated version of YejG (Section 4.3.4), I investigated whether harbouring this YejG variant would change the fitness.

iv) Does over-expressing the mutated version of YejG [YejG(HTD)] improve or alleviate any fitness cost that the fusA(G1478T) mutation may impart?

Next, it was investigated whether a clone with the fusA(G1478T) mutation combined

with the mutated version of YejG [which was the plasmid carried by the E. coli MDS42 yejG (HTD) clone] was fitter than when the mutation is combined with wild-type YejG. A

similar experiment to that performed in (iii) was carried out, but this time the E. coli

MDS42 with the fusA(G1478T) mutation harboured pCA24N-yejG(HTD), instead of wild-

type yejG.

Over-expressing the mutated YejG variant in a strain with the fusA(G1478T) mutation

resulted in no reduction in fitness (W = 1.00 ± 0.01; P > 0.01). The results suggest that

although this YejG variant may alleviate the fitness cost incurred by the chromosomal mutation, it is not able to enhance the cell’s fitness to the extent that wild-type YejG can.

Finally, a fitness assay was performed to determine whether YejG would also improve the fitness of a strain with the wild-type fusA. To examine this, a YejG-expressing strain

and a clone marked with a functional copy of lacZ over-expressing GFP were employed.

The results revealed that in the absence of aminoglycoside, the YejG-expressing strain was significantly fitter (W = 1.16 ± 0.03; P < 0.01) than the control strain.

The fitness experiments revealed that YejG was able to improve the fitness of cells whether or not they carried the fusA mutation. In the next set of experiments, growth assays

were performed in order to investigate whether YejG expression was affecting a particular stage of cell growth.