This research was carried out in accordance with protocols approved by an Institutional Animal Care and Use Committee of the National Institutes of Health (study protocol #'s 2008-1, 2010-78 and 2011-48). Animal work was conducted adhering to the institution's guidelines for animal use, and followed the guidelines and basic principles in the United States Public health Service Policy on Humane Care and Use of Laboratory Animals, and the Guide for the Care and Use of Laboratory Animals. Residents in the villages gave informed consent prior to our setting traps in their houses.

Personal protective equipment (PPE) utilized in these studies was in accordance with established institutional guidelines to prevent exposure to rodent-borne pathogens. In areas of known or suspected LASV circulation, additional PPE was utilized including double gloves, Tyvek coveralls and gowns and HEPA filtered personal powered air purifying respirators with full head covers. Each day at the conclusion of sample processing, traps were disinfected in a mild bleach solution for a minimum of 10 minutes and triple-rinsed with clean water. The work station and all equipment were similarly disinfected. Animal carcasses were incinerated on site.

Between December 2007 and March 2012, six field expeditions were conducted to capture and sample small mammals in peridomestic settings from across Mali for the purposes of testing them for zoonotic pathogens [17], [18](Figure 1, Table 1). Trapping was conducted in various ecozones of sub-Saharan Mali including the grasslands and croplands south of the Sahel to the open and wooded savannas of Southern Mali [19]. Small mammals were live-trapped at each location for one to three consecutive nights using Sherman traps (H.B. Sherman Traps) baited with a mixture of onions and roasted peanuts and set in areas with known or suspected rodent activity. The following morning traps were collected and captured animals immediately processed onsite. Small mammals were anaesthetized by inhalation of isoflurane and exsanguinated via cardiac puncture. The weight, sex and relative age (inferred from body weight and developmental stage) were recorded for each rodent and an ear punch collected and stored in 70% ethanol for cytochrome B sequencing.

At some locations (Figure 1) necropsies were performed on animals and liver samples collected for virus detection. Tissue samples were divided into two pieces; one was immediately frozen and stored in liquid nitrogen (for virus isolation), while the other half was submerged in 600 µl of buffer RLT (Qiagen), mechanically homogenized using a bullet blender (Next Advance) and frozen in liquid nitrogen. Immediately after collection, 140 µl of whole blood was inactivated with lysis buffer AVL (Qiagen) and subsequently frozen. The remaining blood was allowed to clot, spun-down and sera removed while in the field. At the conclusion of the field work, samples were stored at −80°C and shipped on liquid nitrogen to the BSL-4 facility at the Rocky Mountain Laboratories (RML) of the Division of Intramural Research, National Institute of Allergy and Infectious Diseases, NIH.

Genomic DNA was extracted from 3 mm ear punches using DNeasy kits (Qiagen) and speciation of rodents was genetically confirmed by cytochrome B sequence analysis using primers L14723 and H15915, essentially as previously described [18], [20].

Serum samples were tested for the presence of anti-LASV IgG antibodies using an enzyme linked immunosorbent assay (ELISA) based on a recombinant nucleocapsid protein antigen derived from LASV Josiah [21] and using a mixture of anti-rat and anti-Peromyscus leucopus secondary antibodies (each at 1∶2,000 dilution, KPL). Samples were initially screened at a 1∶100 dilution and were considered seropositive if they yielded an OD₄₀₅ value greater than three times the standard deviation of seronegative controls. Reactive samples were titered using 4-fold dilutions.

Total RNA was extracted from inactivated tissue and blood samples using RNeasy and QIAamp viral RNA kits (Qiagen), respectively. These extractions were screened for the presence of LASV RNA using real-time and conventional RT-PCR assays as previously described [17], [22].

Tissue homogenates were prepared from selected seronegative rodents which had detectable LASV RNA in liver and blood samples and passaged twice in Vero E6 cells. For the initial passage, homogenates were diluted 1∶2,500 and incubated on cells for 4 days. Supernatant from p1 was diluted 1∶1,000 and passaged a second time (p2) on nearly confluent monolayers of Vero E6 cells. Cells were monitored daily for CPE and supernatant was harvested on day 5 post-infection. LASV isolation was confirmed by RT-PCR as outlined above.

The full-length genomic sequence of five LASV isolates from Mali, including the original isolate (Soromba-R) obtained in 2009 [17], was determined using high throughput sequencing technologies. Approximately 28 mL of p2 supernatant of each isolate was centrifuged at 25,000×G for 2 hrs in a SW-28 rotor. The resulting viral pellets were lysed in buffer RLT containing 145 mM β-mercaptoethanol and RNA was extracted and purified using RNeasy mini-columns according to the manufacturer's protocol (Qiagen). cDNA was synthesized and libraries generated and quantified as previously described [23]. The cDNA libraries were normalized to 1×10⁷molecules/µL and pooled. Preparation of templated beads for 454 sequencing followed the emPCR Method Manual-Lib-L-LV (Roche Applied Science). Library pools were added to DNA Capture Beads at a target of 1.0 copy per bead. Enriched DNA Capture Beads were then sequenced on a 454 Titanium instrument (Roche Applied Science) per the manufacturer's instructions using a 2-region gasket.

Genomic viral sequences on the Genome Sequencer FLX generated 18,000 usable fragment reads with 275-fold coverage. De Novo genome assembly was performed using GS De Novo Assembler v2.6 (454 Life Sciences) and CLC Genomics Workbench 4.0 (CLC Bio). Translated BLAST (blastx) was performed to eliminate non-viral contaminants and assembly was performed using Sequencher v5.0 (Gene Codes). Assembled contigs were refined by mapping the 454 reads using GS Reference Mapper v2.6 (454 Life Sciences). Full length sequences of the nucleocapsid protein, glycoproteins and polymerase genes were aligned using Clustal 2.1 multiple sequence alignment program (Conway Institute UCD) with the ClustalW algorithm and phylogenetic trees were constructed using Geneious Tree builder v6.51 (Biomatters Ltd.) with the Jukes-Cantor Neighbor-joining method with bootstrapping to 10,000 iterations.

Between December 2007 and March 2012, a total of 793 small animals were captured in sub-Saharan Mali for the purpose of screening for zoonotic pathogens. As previously described, 14 different species of rodents and shrews were captured in Mali [18]. Mastomys natalensis was the most frequently captured small mammal at the majority of locations (Table 1).The diversity of captures was greatest in the northern regions, especially around the Niger inland delta. In contrast, little or no rodent diversity was observed in the southern regions of Mali around the village of Soromba where, during three field expeditions, all but two animals captured in these areas were M. natalensis (246/248, 99.2%). Representative cytochrome b sequences (deposited to GenBank) as well as skull voucher specimens (deposited to the Smithsonian Institution) for these animals are described elsewhere [17], [18].

Serum samples from 715 rodents, including 511 from M. natalensis, were tested for the presence of antibodies reactive to a recombinant LASV nucleocapsid antigen by standard ELISA methodologies. Shrew samples were not evaluated due to a lack of specificity of the secondary antibodies and sample volumes from some (mostly juvenile) rodents were of insufficient quantity to perform serological analysis. Overall, 203 of the 204 serum samples tested from rodents other than M. natalensis were seronegative. The lone sample that was reactive had a serological titer of 1600 and was collected from an adult male African grass rat (Arvicanthis niloticus) captured in Belenikegny, a small fishing village on the Bani River (Figure 1). Although we were unable to determine the infecting arenavirus, it was most likely Ippy virus which is antigenically related to LASV and is maintained in nature in Arvicanthis species [24], [25]. Of the 511 serum samples from M. natalensis, a total of 35 were positive for an overall seroprevalence of 6.8%. Geographically, serological evidence of LASV infection was limited to southern Mali, specifically the villages near Soromba, where the overall prevalence increased to 14.2% (35 of 246). The seroprevalence rates for individual villages in this region varied from 0 to 48% (Table 2).

Tissue samples were collected from 591 rodents and tested for the presence of LASV RNA by real-time and conventional RT-PCR assays. Similar to the serological results, molecular evidence of LASV infection was only found in rodents collected in southern Mali with a total of 19 positive animals identified for an overall prevalence of 7.7% (19 of 246). Prevalence rates for individual villages ranged from 0 to 24% (Table 2). Of these 19 infected animals, 6 were also seropositive. Interestingly, the majority of seropositive animals (29 of 35, 82.9%) did not have detectable viral RNA in blood or liver samples (Table 2).

Combining the results of serological and molecular testing, a total of 48 (19.4%) rodents, all M. natalensis, demonstrated evidence of LASV infection. Twenty-nine animals were seropositive and RT-PCR negative; six animals were both seropositive and RT-PCR positive; and the remaining thirteen animals were seronegative but RT-PCR positive (Table 2). The majority of these animals were adult (41 of 48, 85.4%), with roughly an equal distribution of males and females (22 of 48, 45.8% compared with 26 of 48, 54.2%, respectively) (Table 3).

The full-length genomic sequences of five LASV isolates were determined using next generation sequencing technologies (accession numbers KF478760-KF478769). In addition to the original isolate (Soromba-R) [17], four new LASV isolates were obtained from tissues of rodents captured in March 2012 from the villages of Soromba, Komina, Bamba and Ouoma (designated Soromba-R30, Komina-R16, Bamba-R114 and Ouoma-R123, respectively). Phylogenetic trees constructed with the nucleotide sequences of the nucleocapsid protein (Figure 2), glycoproteins (Figure 3) and polymerase (Figure 4) all demonstrated similar topologies. The Malian isolates grouped closely with strain AV within the LASV lineage and were most divergent from isolates originating from Nigeria. The nucleocapsid gene sequence of the prototype Malian LASV strain Soromba-R differed from AV by 14% at the nucleotide level and 4.2% at the amino acid level. The glycoprotein gene sequences differed by 15.4% at nucleotide level and 4.4% at the amino acid level.