This is a revised version of a previously submitted manuscript by Metelkina et al. which reports a small series of dual inhibitors of the Pseudomonas aeruginosa virulence
factors LecA and LasB. The authors have addressed most, if not all, of the issues highlighted by the reviewers. The most significant change concerns the addition of new data from two anti-virulence assays:

The authors now show that one of the new dual inhibitors (cmpd 12) is superior to both the individual inhibitors 1 and 2, and a combination of 1 + 2, in inhibiting LasB-dependent cytotoxicity of P. aeruginosa PAO1 culture supernatant on human A549 cells. They also show that cmpd 12 at 100uM almost completely inhibits adhesion of fluorescein-labelled LecA to human A549 cells, whereas, 1, 2, and 1 + 2 have no effect. These results go some way to addressing my request for “relevant functional data from anti-microbial or anti-virulence assays” and make this a much more complete, significant, and compelling story overall.

p8 – “The IC50 values obtained for LecA inhibition…” My understanding is that these data come from a ligand binding assay. It seems more appropriate therefore to refer to LecA binding rather than inhibition (which to me implies inhibition of a biochemical reaction).

Fig. 4 – why is reference cmpd 1 included in graph B, but reference cmpds 2 & 3 are not included in graph A? If possible, these results should be included, as they allow an immediate comparison of the dual inhibitors with the requisite parent inhibitor.

Fig. 7 – I agree that cmpd 12 shows statistically significant activity in the LasB-dependent cytotoxicity assay, whereas cmpds 1, 2, and 1 + 2 do not. However, the effect seems quite modest, increasing cell viability from about 20% in the negative control, to about 40%. To give the reader a sense of how significant this change is, it would be extremely useful to include a positive control for direct comparison. Is this possible? I understand that these are first-in-class inhibitors, and that a direct reference cmpd with the same MoA may therefore not be available, but even a positive control with a different MoA but the same phenotype would be extremely useful.

On balance, I support publication of this revised manuscript in Chemical Science, if the few remaining issues above can be addressed.

In my previous review of this revised manuscript, I had raised three remaining issues, which the authors have now addressed in their response, if not in the manuscript itself.

1 – I accept the authors’ explanation for the use of IC50’s, to describe results from a competitive ligand binding assay, even if I don’t necessarily agree with it.

2 – The authors write that they routinely include an internal standard in the LecA assay (Fig. 4, panel B) but haven’t done the same for the LasB assay (Fig. 4, panel A) as these “compounds have been previously evaluated multiple times”. But this is not the same as including known inhibitors as an internal standard? They may at least want to change the wording in the main text to “in the same range as the previously observed activities of compounds 2 and 3 (IC50 = 0.40 and 1.2 µM, respectively)” and add the relevant references there.

3 – With regard to Fig. 7, the authors state that “the relevant control [for the effect of cmpd 12] is the lasB deficient bacterial strain compared to the wildtype strain.” This should be discussed briefly in the main text, including the fact that cmpd 12 leads to only about 50% recovery of A549 viability compared to the lasB-deficient strain. The caption to Fig. 7 also needs an explanation for columns 5-7 (i.e. the positive and negative controls).