Referee 1's Review:

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.
**********

The manuscript pursues the genotypic characterization of Campylobacter jejuni (CJ) isolates from South Africa, with a focus on possible differentiation between isolates involved in gastroenteritis and those with autoimmune sequelae such as Guillain-Barre Syndrome (GBS) syndrome. There is a continuing need for identification and characterization of potential genetic determinants associated with CJ's ability to cause such autoimmune sequelae. In agreement with several previous investigations, the findings from this manuscript did not identify genetic markers unique to GBS-associated isolates.

The work is done by a team with proven expertise in the MLST- and microarray-based subtyping of Campylobacter. I have full confidence in the implementation of the methodologies. The key findings basically support findings from several other investigations, and therefore the report's key findings are largely confirmatory. However, considering the importance of Campylobacter involvement in GBS, the additional information in this manuscript is still useful, because in confirming previous results is also alerts the community on the need to look elsewhere for further understanding of why certain Campylobacter infections are followed by GBS (as authors discuss on p. 17).
My main concern is that a key conclusion of the authors, i.e. the potential for CJE1 and CJE2 to differentiate South African HS:41 isolates form those from other parts of the world is not warranted, due to the very small number of isolates from pother parts of the world. The latter are a total of three isolates! Two from Mexico, and one from Canada; year of isolation is "unknown" for all three; without further analysis such as epidemiological background or additional genomic subtype information, the two isolates from Mexico, both of which have the same ST, could be a single strain.

Specific comments
1. In Table 1, please include a column with CJE1-4 content. On the other hand, listing of alleles may not be necessary, since they can be accessed from the MLST database; a column indicating STs could be sufficient instead.
2. The finding on divergence of the putative patnothenate and biotin biosynthesis genes, and the molybdenum ABC transport, should be further discussed in RESULS and in DISCUSSION. Please, elaborate on divergence, e.g. was genomic location of the genes conserved in those isolates harboring divergent cassettes? How often has absence or divergence of these genes been noted in other studies? It is difficult to determine on the basis of the information provided (p. 12, L20-23) to what extent absence or divergence of these genes is characteristic of serotype HS:41 in general (extensively documented to be clonal). In this context, since the GBS isolate RM3193 has been sequenced (at least this is how I understand the description of this strain as "reference strain" on p. 13, last paragraph, it would be good to elaborate on presence or divergence of these genes in RM3193.
3. If RM3193 has indeed been sequenced, and used as reference in array construction, the actual data should be included in the paper, not just indicated as "data not shown" (p. 13, L23); they would be highly relevant and useful to the readers.
4. p. 14, L2-3. The term "unique differences" needs clarification. Are the authors referring to differences between GBS and other isolates? Or train-specific differences?
5. Abstract needs some additional information in Background that summarizes the key findings (clonality) among South African strains, and describes the key reasons for the current study. Briefly describe please why this particular population of isolates from south Africa was chosen for the study.
6. Again in Abstract, please include total numbers of isolates that were investigated (indicate total HS:41 vs. others, and total number of GC|BS vs. gastroenteritis isolates).
7. p. 2, Abstract, L12-18, I suggest great caution here, since only three HS:41 isolates from other locations were studied. If the section is included, p. 2 L16 should read "...was absent in two closely related HS41 isolates from Mexico".
8. Abstract: Conclusions need some additional information on the lack of detected genomic content differences between GBS and other HS:41 isolates.
9. p. 3, L12 "...may be due...hosts" reads awkward, please re-format; many other reasons may underlie the range of disease outcomes.
10. p. 4, L3, briefly summarize the LOS classes here.
11. p. 4, L6-7, re-work; it can be either "a high frequency of" or "all"-not both.
12. p. 4, L8. The word "clonal" here is awkward, unless you provide further information that justifies it-this would be a good place to summarize the previous investigations e.g. ref. 17,51, etc.
13. p. 5, L2. The PFGE and temperate phage impact would be better placed in a separate sentence; please include the recent reference by the Connerton group-- PLoS Pathog. 2007 Aug 24;3(8):e119.--, as well.
14. p. 5, L5. This section could be preceded by a couple of sentences describing the need to study the South African isolates.
15. p. 7, L9, add "(Table 1)" after RM3193
16. p. 11, L10, I suggest " ...observed for three serotype HS:41 isolates from ..." isolates may be a better use for the Mexican cultures, unless you have evidence that "strains" is justified.
17. p. 11, L17, I suggest "... as the two ST-1672 isolates from ..."; similarly, p. 12, L7, "...were the two ...isolates form Mexico..."
18. p. 12, L14, I suggest to delete "intraspecies"
19. p. 14, L20-22. This conclusion is technically true, but it is really based on three isolates "from other geographical locations". I would suggest to either delete it or re-format it by including the numbers of isolates that were tested from these other locations. The burden will be on the reader to evaluate how meaningful the observed differences involving three isolates really are; at the same time, this might prompt readers with access to HS:41 isolates from other locations to have them evaluated in terms of CJE1-4 content.
20. p. 15, L9, please delete or re-format, as described above. Hard to conclude that "...they are also distinct from other HS:41 strains from other parts of the world..." when you included just three isolates from these other parts of the world.
21. p. 15, L16-17, I suggest "...strains from the two closely related HS:41 isolates from Mexico" .
22. p. 15, L20-21, It would be good to briefly discuss the findings from the earlier Parker et al. publication (ref 38) here; could similar conclusions be drawn from that earlier study as well?
23. p. 16, L14. I suggest replacing "Nevertheless" with "Furthermore"
24. p. 16, L2022. Weren't oligos derived from the genome of the GBS strain RM3193 also used (p. 13-14, see also point 3 above)?
25. p. 17, L2-3, I suggest replacing "...for resolving genetic differences in..." with "...for high-resolution subtyping of..." or something equivalent. Genetic differences are not resolved by SNP analysis.