Identification of emerging Brucella species: new threats for human and animals

The IDEMBRU project focuses on new brucellosis threats, aiming to develop a toolkit to detect and characterize emerging Brucella species and reservoirs. These threats include both emerging atypical Brucella strains, and classical species identified in atypical animal hosts (wild mammals, amphibians, exotic pets), human populations, as well as re-emerging classical pathogenic Brucella linked with new consumption or movement patterns (e.g. raw milk products or imported animal-derived products). This objective will be achieved through a “One Health” approach, adaptation of existing -omics methods, including next generation sequencing (NGS), proteomics and metabolomics in the context of emerging Brucella species. This project will not perform extensive investigations on Brucella melitensis and Brucella abortus, which are already an integral part of EU sanitary control and surveillance plans unless clearly new epidemiological situations/risks are identified. Identification and quantification of emerging animal reservoirs and their surrounding environment will improve networks for monitoring new Brucella threats across European countries. Up to now, emerging Brucella have not been extensively investigated within different epidemiological contexts in terms of natural landscape, livestock demographics, and wild animal populations in Europe . Epidemiological data (human and animal health), phenotypic and genomic data, new biomarkers and specific characteristics will be integrated to compare emerging Brucella to classical species, thus supporting a better elucidation of differences and specificities of these new Brucella species and, importantly, facilitating an assessment of zoonotic potential and associated threats for public health.

The results obtained at the BfRshort forGerman Federal Institute for Risk Assessment all account to point ii) “characterization of the emerging Brucella species by identification of genomic and phenotypic variability via high throughput methods” of the IDEMBRU project. In general, the tasks addressed at the BfRshort forGerman Federal Institute for Risk Assessment are divided into three topics: A. Metabolic phenotyping, B. RNA-Seq as a novel method to identify regulatory processes of Brucella spp., and C. MALDI-TOF based Brucella identification and discrimination.
A. Metabolic phenotyping
The Brucella BfRshort forGerman Federal Institute for Risk Assessment MICRONAUT™ system is an established and well performing metabolic phenotyping platform. This phenotyping system in a 96-well format provides information about enzymatic activities, e.g. peptidases, and substrate utilization, e.g. carbohydrates and amino acids, of Brucella spp., subsequently enabling their identification and discrimination upon metabolic characteristics. Our work with the MICRONAUT™ system clearly revealed that the atypical Brucella species, such as B. microti and B. inopinata but also Brucella isolates from frogs and the former Ochrobactrum species O. anthropi and O. intermedium exhibit clearly expanded and distinguishable metabolic activities compared to classical Brucella species like B. suis and B. melitensis. In contrast, the reduced metabolic activity of the atypical B. vulpis and the Brucella strains isolated from rodents makes the differentiation from B. suis and B. melitensis more difficult. Unfortunately, the production of this commercial MICRONAUT™ system for Brucella phenotyping has been discontinued during the first quarter of the IDEMBRU project. Hence, we focused on the BIOLOG™ phenotype microarray system to pursue our goal to characterize the metabolic properties of atypical Brucella species.The BIOLOG™ phenotype microarrays do not directly examine bacterial enzymatic activities and growth, but rather measure metabolic activity upon substrate utilization independent from growth. For the slow-growing isolates of B. papionis and B. vulpis, we observed distinct substrate utilization patterns, with substrates catabolized or non-catabolized in a similar (e.g. glucose, succinate, citrate) or dissimilar (e.g. glutamate) fashion. However, the BIOLOGTM phenotype microarray analysis was not applicable for the highly metabolically active and fast-growing B. microti, since negative and positive controls could not be distinguished in the experimental setup. We observed unspecific positive reactions in all 96-wells of the micro titer plate. Thus, the BIOLOG™ system requires additional optimization for testing of metabolically highly active Brucella spp. However, these comprehensive modifications within the commercial test system could not be established and implemented during the funding period. The metabolic phenotyping of atypical Brucella species is a prerequisite for a subtask within the RNA-Seq approach (B). To fulfill this task, we complemented our MICRONAUT™- and BIOLOG™-based metabolic comparison of typical and atypical Brucella species with an alternative experimental approach. Hence, we performed substrate utilization experiments in defined minimal medium supplemented with a single carbon or other energy source. This minimal medium was modified from the medium published by Plommet containing solely salts as well as vitamins and did not significantly promote the growth of any tested Brucella strain [doi: 10.1016/s0934-8840(11)80165-9]. A set of 40 substrates, including amino and organic acids, their respective derivatives as well as sugars or sugar alcohols, were individually tested for their growth promoting properties with typical and atypical Brucella spp., and former Ochrobactrum isolates. This approach underlined the capacity of novel Brucella isolates to grow on single energy sources that distinguish them from classical Brucella spp. with their typically fastidious growth phenotype. The novel atypical Brucella spp. show a large utilization spectrum of carbohydrates and organic acids; with greatest diversity in the utilization of amino acids and their derivatives. Strikingly, no strain of the classical Brucella species was able to proliferate with any of the used substrates under these restricted nutritional conditions. Similar growth characteristics showed B. vulpis belonging to the none-core Brucella species. Brucella sp. NF2627 isolated from an Australian rodent showed a highly restricted substrate utilization pattern but was able to grow well on certain sugars, erythritol, pyruvate, lactate and the amino acids proline and lysine.As demonstrated by our Brucella BfRshort forGerman Federal Institute for Risk Assessment MICRONAUT™ phenotyping assay [doi: 10.1186/1471-2180-10-269; doi: 10.1038/srep44420], other non-core Brucella species as well as the former O. anthropi and O. intermedium in contrast to the typical classical Brucella species make use of a wide array of substrates for their growth. Our analysis revealed a highly variable substrate utilization pattern for the tested B. microti isolates, supporting a previous study about the genetic and phenotypic diversity in B. microti [doi: 10.1128/aem.06351-11]. Interestingly, some B. microti isolates could catabolize and grow on ectoine, previously described only for Brucella isolates form Pac-Man frog and former O. anthropi [doi: 10.3389/fcimb.2016.00116].The analyzed Brucella isolates from frogs proliferated with an extended (relative to classical Brucella spp.) but less diverse number of substrates compared to B. microti, most similar to the growth phenotypes of B. inopinata BO1 and B. inopinata-like BO2. Unexpectedly, not all Brucella frog isolates were able to catabolize rhamnose as first described for the B. sp. B13-0095 isolated from a Pac-Man frog [doi: 10.3389/fcimb.2016.00116].The former O. anthropi and O. intermedium species were able to use most of the substrates and grew on 35 and 34 of the 40 tested substances, respectively. Our experimental setup identified several substrates that allowed the distinction between the former Ochrobactrum and the atypical Brucella species, a feature that could potentially be used for diagnostics. Only O. anthropi and O. intermedium but none of the tested atypical Brucella isolates proliferated with citric acid, gluconic acid and mannitol as sole carbon and energy source in the defined minimal medium. In contrast, all tested atypical Brucella isolates but not the former O. anthropi and O. intermedium species were able to grow on adipic acid as sole carbon and energy source.
B. RNA-Seq as a novel method to identify regulatory processes of Brucella spp.We evaluated several RNA-seq kits for various Brucella species. Based on our experiments a suitable protocol has been developed. This protocol can be used for Brucella cultivation, RNA extraction, quality control, library preparation and sequencing.Our pilot experiment addressed environmental conditions that emerging Brucella spp. might be exposed to [doi: 10.3389/fmicb.2021.794535]). Overnight cultures of B. vulpis F60 and B. microti CCM4915 were used to inoculate Gerhardt’s minimal medium at pH7 and pH4.5 in triplicates. Growth phase dependent comparative RNA-seq-based transcriptome analysis was performed on samples that reached an OD600 >0.5. Samples were processed according to the developed protocol. For transcript abundance determination and exploratory differential data analyses, we applied the bioinformatics tools kallisto and sleuth, respectively [doi: 10.1038/nbt.3519; doi: 10.1038/nmeth.4324]. Volcano plots for B. vulpis and B. microti indicated intra-species regulation, since many transcripts were differentially expressed (up- or down-regulated) in response to acidic stress in both species. Furthermore, the complexity of the two species-related datasets were then reduced by means of a principle component analysis (PCA). The samples of the B. vulpis dataset showed clear clustering of triplicates according to their experimental conditions (i.e. pH7 vs. pH4.5) in PCA, with slightly more heterogeneity within the replicates of low pH than neutral pH. The differences between these clusters are larger on principle component 1 (PC1) than on principle component 2 (PC2). In contrast, PCA of the B. microti dataset was less conclusive. Here, the replicates of neutral pH correlate with another and formed a cohesive cluster, whereas, replicates of low pH showed large variation and thus weak correlation among the technical replicates. Two samples correlated and clustered closely together, while the other sample showed differences placing it apart on PC1 and PC2. This heterogeneity within the technical replicates of low pH of the B. microti dataset made the discrimination of experimental conditions based on clusters less clear than for the B. vulpis dataset. In general, these diverging datasets highlight that RNA-seq experiments cannot be easily implemented even by a specialist, who initially performed these experiments in the lab and did the bioinformatics analyses. In the middle of the IDEMBRU project, the responsible employee left our lab, which resulted in a massive loss of key knowledge in RNA-seq and necessary bioinformatics analyses (brain drain). SOPs and bioinformatics pipelines could not be established before this change and without novel expertise. In addition, the SARS-CoV-2 pandemic negatively influenced the recruitment of new RNA specialists. Finally, the metabolic phenotyping required tremendous resources since well-established, commercially available assays were withdrawn from the market.
C. MALDI-TOF based Brucella identification and discrimination

This topic was highly dependent on cooperation and strain exchange among IDEMRBU project partners to gather a variety of emerging species and their proteomic fingerprints. The FLIshort forFriedrich Loeffler Institute provided ten novel atypical Brucella isolates, which have been sequenced and analyzed by MALDI-TOF MS. . Our in-house spectral library, predominantly consisting of classical Brucella isolates, was extended by integrating the processed novel Brucella isolates from FLIshort forFriedrich Loeffler Institute, as well as by all novel Brucella isolates from our own strain collection. For discriminatory power, we also integrated the spectra of former Ochrobactrum species, namely O. intermedium and O. anthropi. We analyzed all newly integrated spectra and identified species-specific biomarker peaks for the novel species and isolates by comparison of all mass spectral patterns. The previously reported genus-specific peak for Brucella was confirmed in our dataset [doi: 10.1371/journal.pntd.0006874]. In addition, a biomarker peak that is absent in former O. anthropi and O. intermedium isolates but present in the established genus of Brucella – including all core and non-core Brucella species – was identified within the expanded dataset. Samples belonging to the former genus Ochrobactrum revealed unique peaks that were absent in all core and non-core Brucella species. Furthermore, discrimination of the two Ochrobactrum species within the extended Brucella genus was achieved with additional biomarker peaks. In addition to the biomarkers for former Ochrobactrum spp., specific peaks for all novel Brucella species were identified. Thus, with our spectral library we are able to identify and discriminate all currently integrated Brucella species. A manuscript addressing the differentiation of novel and atypical Brucella with biomarker signals in mass spectral libraries is currently in progress [Kühn et al.short foret alii (lat. "and others"), in progress].