Brucella neotomae Infection in Humans, Costa Rica

Several species of Brucella are known to be zoonotic, but B. neotomae infection has been thought to be limited to wood rats. In 2008 and 2011, however, B. neotomae was isolated from cerebrospinal fluid of 2 men with neurobrucellosis. The nonzoonotic status of B. neotomae should be reassessed.


The Study
In 2008, a Brucella sp. isolate was submitted to the Tropical Diseases Research Center at the Universidad de Costa Rica. This isolate (denoted strain bneohCR2) was cultured from a cerebrospinal fluid sample obtained from a 64-yearold male patient at one of the main hospitals in San José, Costa Rica. In 2011, another isolate (denoted strain bneo-hCR1) was recovered from a cerebrospinal fluid sample from a 51-year-old male patient at a regional hospital in Costa Rica. As is common for other patients with brucellosis, the blood leukocyte count for each patient was almost within the reference range, and C-reactive protein level was within reference range. Both patients showed clinical signs compatible with neurobrucellosis (6), had positive Rose Bengal test results, and recovered after receiving 1 month of streptomycin (750 mg/d) and 3 months of doxycycline (100 mg/12 h).
We performed whole-genome sequencing of bneohCR1 and bneohCR2 and resequencing of reference strain B. neotomae 5K33. Sequencing data were deposited at the European Nucleotide Archive (http://www.ebi.ac.uk/ena/) under accession codes ERS1563929 (bneohCR1), ERS1563928 (bneohCR2), and ERS1624467 (SK33). To place the bneo-hCR1 and bneohCR2 in a phylogenetic context, publicly available reads from 51 Brucella whole-genome sequences (online Technical Appendix Table 2) were aligned and then mapped to B. suis 1330 by using SMALT version 0.5.8 (ftp:// ftp.sanger.ac.uk/pub/resources/software/smalt/). Reads from bneohCR1 and bneohCR2 genomes mapped to 98.6% of the B. suis 1330 genome. SNPs were called from the alignment by use of Samtools (http://samtools.sourceforge.net/), and 34,307 variable sites across all isolates were extracted by using SNP sites (13). The resulting alignment of SNPs was used for maximum-likelihood phylogenetic reconstruction by use of RAxML version 7.0.4 (https://github.com/ stamatak/standard-RAxML). The generated phylogenetic tree confirmed that the bneohCR isolates clustered together with B. neotomae reference strain 5K33 (ENA accession no. JMSC01, assembly accession no. GCA_00742255.1) (Figure 2). Isolates bneohCR1 and bneohCR2 differed from the reference genome by 174 and 160 SNPs, respectively. This number of SNPs is smaller than that between B. abortus 9-941 and B. abortus 2308 (214 SNPs), which are 2 wellrecognized strains of the same Brucella species.
Analysis of 23 previously reported genomic islands or anomalous genomic loci (14) was performed for both bneo-hCR genomes. For this analysis, a "genomic-island pseudomolecule" was constructed by concatenation of 23 genomic regions obtained from different Brucella genomes. BLAST (https://github.com/sanger-pathogens/Farm_blast) comparison of this pseudo-molecule and the bneohCR draft genomes, generated by assembly with Velvet (15), showed that the genomic loci known as 26.5 kb, 12 kb, and GI-6 that are absent in B. neotomae (14) are also absent in the queried genomes.

Conclusions
This report of B. neotomae as a cause of zoonotic disease raises questions about possible underrepresentation of reported cases. This study also has implications for brucellosis diagnosis. Specifically, the differences among B. neotomae and the other Brucella species at the biochemical level are subtle. The major difference between B. neotomae and B. abortus, the main cause of human brucellosis in Costa Rica, is the presence of oxidase activity in B. abortus, which is assessed subjectively (7,8). Because B. neotomae has not, until now, been considered zoonotic, some cases of brucellosis reported as being caused by atypical zoonotic classical Brucella might have been misdiagnosed cases of B. neotomae infection. The introduction of whole-genome sequencing into the clinical field will thus improve diagnosis.
A lack of epidemiologic information with regard to the 2 patients reported here precluded the investigation of exposure or contact with hosts known to harbor B. neotomae. Further studies are needed to establish which animals may act as reservoirs for B. neotomae in Costa Rica.
In summary, B. neotomae should be considered a zoonosis risk for infection in humans. Incorporation of molecular techniques for diagnosis will help resolve the

Brucella neotomae Infection in Humans
Brucella genus homogeneity obtained when only biochemical assays are used.