On 2017 May 20, Jonathan Lalsiamthara commented:

Dear Jean-Jacques Letesson,

Thank you for your concerns on this manuscript. We could have replied to you sooner if the queries were forwarded to the given corresponding author and also updates regarding this topic could have been promptly relayed to you.

1 Our interest was to generate O-PS shortened S19 strain. Interestingly, this particular strain (IVRIGEBJ7) which has putative perosamine synthetase gene knocked-out has shown truncated O-PS profile. We also reported that it is rough intermediate (by comparing it with smooth and complete-rough strains) and proved that it is on the surface rather than inside cytoplasm. While we could have communicated with the strain name IVRIGEBJ7, to make it more meaningful we go by the target gene name 'per'. It can still be useful to report a potential vaccine candidate without knowing the actual genetic makeup, because historically S19 and RB51 have many unknown mutations when they were reported. Nevertheless, the whole-genome sequence (http://genomea.asm.org/content/3/6/e01336-15.full.pdf) of the mutant strain is now available on GenBank and complementation eventually confirmed the veracity of the mutation.

2 Indeed there are two enzymes encoding perosamine synthetase. This manuscript reported knock-out of a putative one (which some workers may annotate it as wbkB). This might explain your concerns regarding “so called "per" mutant”, only that the enzymes are not exclusive to S19, but both are present across Brucella spp.

3 It was discussed that similar “OPS profiles” i.e truncated OPS in PAGE- silver or -western blots were also observed in two given studies, delta pgm and RB51. The discussion did not claim anything about their phenotypes but related to their OPS properties. The typographical error should be “.. RB51 strain with wboA gene complemented”.

On 2017 Mar 23, Jean-Jacques Letesson commented:

The basic scientific requirement to characterize a mutant is to prove that the phenotype characterized can be reverted to the WT phenotype by bringing in trans the WT gene.

This has not be done in this study as no complementation was even evoked. Therefore the phenotype observed cannot be claimed as caused by the "per" deletion.

There are in addition several fact in this paper that were either omitted, overinterpreted or even mistaken.

1- Concerning the roughness of the so called "per" mutant

-the O-PS of Brucella is an homopolymer of N-formylperosamine (meaning that the sole an unique sugar used to build the O-chain is this one) and that N-formylperosamine is the biosynthetic result of three sequential reactions catalyzed by enzymes (Gmd, per, Wbkc) encoded by three unique genes.

-without "per" it is impossible to build a "bona fide" perosamine and thus an O-PS chain. In none of the Brucella species in which the "per" mutation has been tested, has a partial/shortened or whatsoever modified O-chain been described.

-Considering that the S19 per mutant described in this paper is "mid-smooth" will thus need another gene encoding the "same" enzymatic activity that would only exist or be expressed in B. abortus S19 and not in any other Brucella species.

-the epitopic composition of the O-PS should have been explored with a panel of MoAbs specific for Brucella O-PS and not with an uncharacterized polyclonal antibody that give a VERY POOR immunoreactivity on the O-PS of the WT S19

2- To explain the (mid-smooth phenotype) and the truncated O-PS observed in the S19 per mutant, the authors evoke, in the discussion, that a similar observation was made for delta pgm of B. abortus S2308 and also for the B. abortus RB51 strain with wboA gene deletion mutation. THIS CLAIM IS ABSOLUTELY FALSE

-the pgm mutant is completely ROUGH (but has an O-chain in the cytoplasm and is unable to link it to the core)

-the wboA is completely ROUGH and in the reference they cite the strain is a complemented one that is still ROUGH but produce an O-chain in the cytoplasm. (Vemulapalli R, He Y, Buccolo LS, Boyle SM, Sriranganathan N, Schurig GG. Com-plementation of Brucella abortus RB51 with a Functional wboA Gene Results inO-Antigen Synthesis and Enhanced Vaccine Efficacy but No Change in RoughPhenotype and Attenuation. Infect Immun 2000;68:3927–32.)

On 2017 Mar 23, Jean-Jacques Letesson commented:

The basic scientific requirement to characterize a mutant is to prove that the phenotype characterized can be reverted to the WT phenotype by bringing in trans the WT gene.

This has not be done in this study as no complementation was even evoked. Therefore the phenotype observed cannot be claimed as caused by the "per" deletion.

There are in addition several fact in this paper that were either omitted, overinterpreted or even mistaken.

1- Concerning the roughness of the so called "per" mutant

-the O-PS of Brucella is an homopolymer of N-formylperosamine (meaning that the sole an unique sugar used to build the O-chain is this one) and that N-formylperosamine is the biosynthetic result of three sequential reactions catalyzed by enzymes (Gmd, per, Wbkc) encoded by three unique genes.

-without "per" it is impossible to build a "bona fide" perosamine and thus an O-PS chain. In none of the Brucella species in which the "per" mutation has been tested, has a partial/shortened or whatsoever modified O-chain been described.

-Considering that the S19 per mutant described in this paper is "mid-smooth" will thus need another gene encoding the "same" enzymatic activity that would only exist or be expressed in B. abortus S19 and not in any other Brucella species.

-the epitopic composition of the O-PS should have been explored with a panel of MoAbs specific for Brucella O-PS and not with an uncharacterized polyclonal antibody that give a VERY POOR immunoreactivity on the O-PS of the WT S19

2- To explain the (mid-smooth phenotype) and the truncated O-PS observed in the S19 per mutant, the authors evoke, in the discussion, that a similar observation was made for delta pgm of B. abortus S2308 and also for the B. abortus RB51 strain with wboA gene deletion mutation. THIS CLAIM IS ABSOLUTELY FALSE

-the pgm mutant is completely ROUGH (but has an O-chain in the cytoplasm and is unable to link it to the core)

-the wboA is completely ROUGH and in the reference they cite the strain is a complemented one that is still ROUGH but produce an O-chain in the cytoplasm. (Vemulapalli R, He Y, Buccolo LS, Boyle SM, Sriranganathan N, Schurig GG. Com-plementation of Brucella abortus RB51 with a Functional wboA Gene Results inO-Antigen Synthesis and Enhanced Vaccine Efficacy but No Change in RoughPhenotype and Attenuation. Infect Immun 2000;68:3927–32.)

On 2017 May 20, Jonathan Lalsiamthara commented:

Dear Jean-Jacques Letesson,

Thank you for your concerns on this manuscript. We could have replied to you sooner if the queries were forwarded to the given corresponding author and also updates regarding this topic could have been promptly relayed to you.

1 Our interest was to generate O-PS shortened S19 strain. Interestingly, this particular strain (IVRIGEBJ7) which has putative perosamine synthetase gene knocked-out has shown truncated O-PS profile. We also reported that it is rough intermediate (by comparing it with smooth and complete-rough strains) and proved that it is on the surface rather than inside cytoplasm. While we could have communicated with the strain name IVRIGEBJ7, to make it more meaningful we go by the target gene name 'per'. It can still be useful to report a potential vaccine candidate without knowing the actual genetic makeup, because historically S19 and RB51 have many unknown mutations when they were reported. Nevertheless, the whole-genome sequence (http://genomea.asm.org/content/3/6/e01336-15.full.pdf) of the mutant strain is now available on GenBank and complementation eventually confirmed the veracity of the mutation.

2 Indeed there are two enzymes encoding perosamine synthetase. This manuscript reported knock-out of a putative one (which some workers may annotate it as wbkB). This might explain your concerns regarding “so called "per" mutant”, only that the enzymes are not exclusive to S19, but both are present across Brucella spp.

3 It was discussed that similar “OPS profiles” i.e truncated OPS in PAGE- silver or -western blots were also observed in two given studies, delta pgm and RB51. The discussion did not claim anything about their phenotypes but related to their OPS properties. The typographical error should be “.. RB51 strain with wboA gene complemented”.