In a survey for rodentborne hemorrhagic fever viruses, 612 small rodents representing 17 different genera (most abundant were Mastomys [n = 325], Praomys [n = 95], and Nannomys [n = 83]) were trapped in Guinea from 2002 to 2004 and screened for hantavirus RNA by reverse transcription–polymerase chain reaction (RT-PCR). We used a molecular genetic approach to screen the rodent population because hantavirus RNA (as shown for SNV) can be amplified from the blood of persistently infected mice by RT-PCR over a long period (6). For this purpose, we developed a nested RT-PCR assay to detect currently known and possible novel members of the genus Hantavirus. The assay was based on degenerated primers (HAN-L-F1: 5´-ATGTAYGTBAGTGCWGATGC-3´ and HAN-L-R1: 5´-AACCADTCWGTYCCRTCATC-3´ for primary PCR, HAN-L-F2: 5´-TGCWGATGCHACIAARTGGTC-3´ and HAN-L-R2: 5´-GCRTCRTCWGARTGRTGDGCAA-3´ for nested PCR) designed from an alignment of all available nucleotide sequences of the highly conserved L segment. For the RT-PCR, total RNA was extracted from wild-trapped rodent blood (preserved in liquid nitrogen) with the Blood RNA kit (Peqlab, Erlangen, Germany) and reverse transcribed with random hexamers as primers.

A sample (designated SA14) obtained from 1 of 4 investigated African wood mice (H. simus) generated an L segment–derived PCR product of expected size. This rodent was trapped in January 2004 in a forest habitat near the village of Sangassou, near Macenta, Guinea (8°36´49´´N, 9°28´27´´W). Its karyotype was determined (2n = 48, fundamental no. = 74, autosomal fundamental no. = 70) and the complete cytochrome b gene was sequenced and compared with the genes of other Hylomyscus species recognized in the most recent revision of the genus (7) (GenBank accession nos. DQ212188 and DQ078229-DQ078245).

The 412-nucleotide (nt) sequence of the first PCR product was determined by amplification, cloning, and sequencing of overlapping fragments generated by 2 seminested PCRs (GenBank accession no. DQ268652). Additional S and M segment–specific nested PCR assays were developed to further characterize the novel virus. PCR fragments of 837 nt and 694 nt could be analyzed (GenBank accession nos. DQ268650 and DQ268651, respectively).

The Table shows nucleotide sequence identity comparisons between SA14 and other members of the genus Hantavirus. Murinae-associated hantaviruses (HTNV, DOBV, SEOV) showed the highest similarity to the SA14 sequence in all 3 genomic segments (71.3%–77.1% for S, 72.9%–77.9% for M, and 72.3%–75.9% for L). This similarity is consistent with the evolutionary relationship of their putative hosts. On the amino acid level, corresponding sequences of deduced viral proteins showed highest similarity with those of other Murinae-associated hantaviruses (81.7%–88.5% for S, 82.2%–89.6% for M, and 85.4%–87.5% for L). The amino acid sequence divergence between SA14 and most related hantaviruses corresponds to that typically found between different virus species, e.g., SNV and ANDV.

The S-, M-, and L-segment–derived nucleotide sequences of SA14 were subjected to maximum likelihood (ML) and neighbor-joining (NJ) phylogenetic analyses with available nucleotide sequences of other Murinae-associated hantaviruses. PUUV, TULV, SNV, and ANDV sequences were used as outgroups. In the S segment ML phylogenetic tree (Figure), SA14 clustered with other Murinae-associated viruses. As expected, 3 clades were formed by members of the 3 established hantavirus species (HTNV, DOBV, and SEOV). Within this well-supported cluster, the SA14 sequence is most closely related to the DOBV clade. The M segment analysis showed an identical placement of SA14 with strong statistical support (PUZZLE [8] and bootstrap values above the threshold value of 70%, data not shown). In L-segment phylogeny, the resolution of the tree was decreased. The SA14 L sequence did not join with statistical support any of the 3 groups but formed a fourth clade within the cluster of Murinae-associated hantaviruses. (PUZZLE and bootstrap values above the threshold for the placement of SA14 within the Murinae-associated viruses but <50% in both analyses for its clustering with any of these viruses, data not shown).

Extended fragments of novel hantavirus S, M, and L genome segments were recovered from an arboreal African rodent. They clearly represent genetic material of a novel hantavirus species because their amino acid sequence is significantly (≈15%) divergent from those of other hantaviruses, they form a distinct clade in phylogenetic trees, and they were detected in a rodent species previously not recognized as a natural host of hantaviruses. We propose to name the new species Sangassou virus (SANGV) after the locality where it was detected.

Although hantaviruses are emerging viruses circulating in Asia, Europe, and the Americas, our study represents the first genetic evidence for hantaviruses in Africa. Suspected human hantavirus infections have been reported in various African countries (10–13). Most of these are seroepidemiologic studies reporting antibodies reacting with HTNV antigen. However, Apodemus agrarius, the natural host of HTNV, is not found in Africa. Based on the putative cross-reactivity of antigens from HTNV, SANGV, and other Murinae-associated viruses, human infections, at least in tropical forest parts of Africa where Hylomyscus species are prevalent, could be caused by SANGV or other Murine-associated hantaviruses.

To our knowledge, 1 case of HFRS has been reported in central Africa (14). Although HFRS is not a known disease in West or central Africa, one cannot ignore the potential pathogenicity of SANGV or other African hantaviruses. HFRS may be confused with other severe diseases (leptospirosis, rickettsiosis, other viral hemorrhagic fevers, plague, severe pneumonia, sepsis) or may be unrecognized because of poor health care. One should remember that HPS and Sigmodontinae-associated hantaviruses were not recognized until 1993, even in such a highly developed country as the United States.

Further studies are needed to verify the presence and distribution of hantaviruses in Africa and their potential impact on human health. These studies should focus on areas with forest activities, such as logging, which may bring humans into contact with viral reservoirs (15). Our data justify inclusion of hantavirus infection in the differential diagnosis of patients from Africa with unexplained febrile nephropathies or noncardiogenic pulmonary edema.