Sangassou Virus, the First Hantavirus Isolate from Africa, Displays Genetic and Functional Properties Distinct from Those of Other Murinae-Associated Hantaviruses

Abstract

We have discovered the first indigenous African hantavirus, Sangassou virus (SANGV). The virus was isolated from an African wood mouse (Hylomyscus simus), trapped in a forest habitat in Guinea, West Africa. Here, we report on the characterization of the genetic and functional properties of the virus. The complete genome of SANGV was determined and showed typical hantavirus organization. The small (S), medium (M), and large (L) genome segments containing genes encoding nucleocapsid protein, two envelope glycoproteins, and viral polymerase were found to be 1,746, 3,650, and 6,531 nucleotides long, respectively. The exact 5′ and 3′ termini for all three segments of the SANGV genome were determined and were predicted to form the panhandle structures typical of bunyaviruses. Phylogenetic analyses of all three segment sequences confirmed SANGV as a Murinae-associated hantavirus most closely related to the European Dobrava-Belgrade virus. We showed, however, that SANGV uses β₁ integrin rather than β₃ integrin and decay-accelerating factor (DAF)/CD55 as an entry receptor. In addition, we demonstrated a strong induction of type III lambda interferon (IFN-λ) expression in type I IFN-deficient Vero E6 cells by SANGV. These properties are unique within Murinae-associated hantaviruses and make the virus useful in comparative studies focusing on hantavirus pathogenesis.

INTRODUCTION

Despite hantaviruses being a well-recognized human pathogen in most of the continents in the world, indigenous African representatives have only very recently been identified. Novel hantavirus genome sequences were obtained from an African wood mouse, Hylomyscus simus (recognized as a species in the most recent revision of the genus [32] and previously designated a subspecies of Hylomyscus alleni [30]). The virus was named Sangassou virus (SANGV) after the village in Guinea where the animal carrying the virus had been trapped (17). Later, genome sequences of another unique hantavirus were found in Therese's shrew (Crocidura theresae), and it was named Tanganya virus (TANGV), again after the village where it was found (18). These findings stimulated a recent seroepidemiological study in the forest region of Guinea which clearly showed that local hantaviruses infect humans (seroprevalence of 1.2% in the general population). Moreover, a 4.4% seroprevalence of hantavirus-specific antibodies in patients with fever of unknown origin and detection of SANGV-specific IgM and IgG antibodies in one patient indicate that hantaviruses may constitute a medical problem in Guinea (19). Very recently, additional hantaviruses, Azagny virus harbored by the West African pygmy shrew (Crocidura obscurior) and Magboi virus detected in the slit-faced bat (Nycteris hispida) have been found in the neighboring countries Cote d'Ivoire (15) and Sierra Leone (44), respectively, further confirming that West Africa is a region where hantavirus infections need to be considered.

Hantaviruses (Bunyaviridae family) have been known for decades in other parts of the world. In Asia and Europe, hantaviruses cause hemorrhagic fever with renal syndrome (HFRS). Prominent representative hantaviruses are Hantaan virus (HTNV) and Seoul virus (SEOV), which are prevalent mainly in Korea and China, as well as Puumala virus (PUUV) and Dobrava-Belgrade virus (DOBV), which are present in Europe. So-called New World hantaviruses such as Sin Nombre virus (SNV) and Andes virus (ANDV) are the causative agents of hantavirus cardiopulmonary syndrome (HCPS). Despite its severity and high fatality rate (up to 50%), the disease and the New World hantaviruses were recognized in the United States only in 1993 (31), indicating that hantaviruses may escape human attention even in countries with developed medical health care systems.

Both diseases, HFRS and HCPS, are acute febrile infections with similar initial symptoms, such as the abrupt onset of a high fever, malaise, myalgia, back and abdominal pain, and other flu-like symptoms, and are associated with acute thrombocytopenia and increased vascular permeability. HFRS is mainly characterized by renal failure, while pulmonary and cardiovascular dysfunctions are more characteristic of HCPS. The pathogenesis of HFRS and HCPS is assumed to be a complex, multifactorial process which includes T-cell-mediated endothelial damage, immune effectors, and β₃ integrin dysfunction-mediated increase of vascular permeability (6, 16, 25, 29, 38).

Hantaviruses form a unique genus Hantavirus within the Bunyaviridae family. The virus genome consists of three segments of negative-stranded RNA; the large (L) segment encodes the viral RNA-dependent RNA polymerase, the medium (M) segment encodes the envelope glycoproteins Gn and Gc (cotranslationally cleaved from a glycoprotein precursor), and the small (S) segment encodes the nucleocapsid (N) protein (25, 34).

Hantaviruses are transmitted to humans by aerosolized excreta of their natural hosts, small mammals. For many years, they were considered rodent-borne viruses, but recently many new distinct hantaviruses of currently unknown pathogenic potential have been discovered in shrews and moles (order Soricomorpha). In their natural hosts, hantaviruses produce chronic infections with no apparent harm. They are strictly associated with one (or few closely related) small mammal species as their natural reservoir hosts, which is also reflected in their phylogeny. Rodent-borne hantaviruses form three major evolutionary clades corresponding to the subfamilies of their rodent hosts; HTNV, SEOV, and DOBV are examples of Murinae-associated hantaviruses, PUUV and Tula virus (TULV) belong to the Arvicolinae-associated hantaviruses, and SNV and ANDV are representatives of Neotominae- and Sigmodontinae-associated hantaviruses, respectively. The Soricomorpha-associated viruses seem to form at least two main additional clusters but without a clear association with the phylogeny of their hosts (2, 14). Surprisingly, the most recently found mole-associated Rockport virus was shown to share a most recent common ancestor with Arvicolinae- and Sigmodontinae-associated hantaviruses (13).

Here we report the isolation of SANGV in cell culture. Analysis of the complete virus genome confirmed its phylogenetic placement within the Murinae-associated hantaviruses. SANGV seems to differ from other Murinae-associated hantaviruses in the biological properties analyzed, such as receptor usage or the induction of interferon (IFN).

MATERIALS AND METHODS

Cells and viruses.

Vero E6 cells (Vero C1008, ATCC CRL 1586) were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS), 100 IU penicillin, and 100 μg/ml streptomycin. A549 cells (human epithelial lung cell line; ACC 107, German Collection of Microorganisms and Cell Cultures, Braunschweig, Germany) were cultured in minimal essential medium (MEM) with Earle's salt supplemented with 5% FCS, 25 mM HEPES, 1% glutamine, 1% sodium pyruvate, 1% nonessential amino acids, and 0.1% gentamicin sulfate. CHO-K1 cells stably transfected with pcDNA 3.1 vector (Invitrogen) expressing α_V integrin and pZeoSV vector (Invitrogen) expressing β₃ integrin as well as with empty control vectors were cultured and maintained in selection medium (50% Ham's F-12 medium, 50% DMEM, 10% FCS, 1% penicillin-streptomycin, 1% l-glutamine, 350 μg/ml G418, and 250 μg/ml zeocin) (1). Integrin expression was monitored by fluorescence-activated cell sorting (FACS) analysis.

Human umbilical vein endothelial cells (HUVEC) were isolated from umbilical cord veins and identified by the method of Jaffe et al. (12). Briefly, cells were obtained after exposure to 0.1% collagenase solution and washed, resuspended, and cultivated in endothelial basal medium (EBM) supplemented with 10% endothelial cell growth medium (ECGM), 1% l-glutamine, 1.4% penicillin-streptomycin, 0.2% amphotericin, and 200 μl/liter endothelial growth factor and seeded into 6-well plates. Confluent monolayers of primary cells were used directly for the experiments.

HTNV strain 76-118, Prospect Hill virus (PHV) strain 3571, and SANGV strain SA14 (isolated within this study) were used in the experiments. Virus stocks were prepared in Vero E6 cells cultivated in 75-cm² cell culture flasks. The cells were infected at a multiplicity of infection (MOI) of 0.1 for 1 h at 37°C in a 5% CO₂ humidified atmosphere (standard cell culture conditions), and then fresh medium was added. Seven days postinfection (p.i.), cell culture supernatants were collected, centrifuged to remove cell debris, aliquoted, and stored at −80°C. Virus stocks and cells were determined to be free of mycoplasma contamination by using the PCR-based VenorGeM mycoplasma detection kit (Minerva Biolabs).

For measurement of lambda interferon (IFN-λ) content by an enzyme-linked immunosorbent assay (ELISA) outside a biosafety level 3 (BSL-3) facility, virus stocks were UV inactivated. Virus stock solution (0.5 ml) was transferred to a small plastic petri dish and placed directly on the workspace of the UV transilluminator equipped with 8-W tubes (Vilber Lourmat). Inactivation was performed by UV irradiation for 3 min at 312 nm, corresponding to 1.4 J/cm² (23).

SANGV isolation.

Spleen tissue from a Hylomyscus simus animal trapped near Sangassou village in a forest in Guinea was used for virus isolation attempts by the method of Klempa et al. (20). Briefly, the tissue was triturated in a closed mechanical blender FastPrep Instrument (Bio 101 Systems) as a 10% tissue suspension in Dulbecco's medium supplemented with 0.2% bovine serum albumin (BSA). The suspension was briefly centrifuged at low speed, inoculated (0.4 ml/flask) onto cultures of confluent Vero E6 cells in 25-cm² flasks, and incubated at 37°C. The cell culture medium was changed for the first time after 90 min and then weekly. Cells were passed at 2-week intervals with the addition of the same amount of fresh uninfected cells. During this step, several slides were prepared and examined for characteristic hantavirus antigen expression following immunofluorescence assay (IFA) techniques (4).

The experiments were performed under biosafety level 3 containment conditions in the Institute of Virology, Charité Medical School.

Virus ultracentrifugation.

For production of high-titer, IFN-free virus stocks, 175-cm² cell culture flasks were infected and incubated for 7 days at culture conditions. After 2 freeze-thaw cycles, cells were scraped from the bottom of the culture vessel and exposed to sonication. Cell debris was removed by centrifugation. The supernatant was transferred into sealed tubes and ultracentrifuged for 3 h at 28,000 × g and 4°C. Virus pellets were resolved in fresh culture medium by repeating the vortexing and sonication steps.

Virus titration.

The viral stocks were titrated by using the chemiluminescence focus assay of Heider et al. (11). Briefly, 10-fold serial dilutions of viral stock were inoculated into six-well plates with nearly confluent monolayers of Vero E6 cells. After an adsorption period for 1 h at 37°C, the cells were overlaid with a mixture of 1% agarose and Eagle basal medium. The plates were then incubated for 12 days. Virus-infected cells were detected with anti-SANGV rabbit hyperimmune serum, followed by peroxidase-labeled goat anti-rabbit IgG and chemiluminescence substrate Super Signal West Dura (Thermo Scientific).

Sequencing the complete genome of SANGV.

To determine the nucleotide sequences of all three genomic segments, a series of overlapping PCR fragments was generated for every segment and subjected to cloning and sequencing. Hantaviral RNA was first extracted from cell culture supernatant by using the QIAamp viral RNA minikit (Qiagen). Total RNA was reverse transcribed using SuperScript II (Invitrogen) and random hexamers. The primer sequences for reverse transcription-PCR (RT-PCR) were designed from published Murinae-associated hantavirus sequences (Table 1). RT-PCRs were performed using the TripleMaster PCR system (5 Prime). The amplified products were then cloned into pSCA vector using a StrataClone PCR cloning kit (Stratagene), and at least three clones were sequenced from both directions.

Table 1.

Primers used to sequence the complete genome of SANGV

Primera	Primer sequenceb (5′ to 3′)
Alex-HAN	gac cat cta gcg acc tcc acT AGT AGT AKR CNC C
MURS-1246R	GGR TCC ATR TCA TCI CC
MURS-374F	WGG ICA RAC IGC WGA YTG G
MURS-598F	TGA ARG CWG AIG ARA TIA CAC
MURM-215F	CCI GAR AGY TCI TGY ARY ATG GA
SA14 M-1496R	GTT GCC CAC CCA TGA A
MURM-1472F	GGI TTY CAY GGI TGG GC
SA14 M-2313R	GCC CCA ACT ATT TTC ATA CT
MURM-2269F	AAR TAY SAR TAY CCI TGG CA
MURM-2999R	CCC CAI GCI CCY TYW AT
MURM-2500F	ATW GAY ATG AAY GAY TGY TTY GT
MURM-2822R	TGR AAI GAR TCA ATI GTI GC
MURL-1F	TAG TAG TAG ACT CCS KAA
MURL-145R	TCC ATY TGR TCW ACA ATR TCA
SA14L-113F	GAT TAT CTT GAT CGG CTT TAT
SA14L-1458R	CCC TTA TCA TAT GCC CTA TGT
MURL-1439F	AAR AAR ACI ACW GCA TGG CA
SA14L-3049R	CGGAGTGCATCTATCACA
SA14L-3230F	TCT TCC CTT TTT GGT GTT GCC
SA14L-6307R	CTC TCA CCC CAA CCC TTA ACG
MURL-5972F	GAT GAT GAR TTY ACA ATW GAC
SA14S-1520Fc	ATT CCG TTG AAT TAC TCA GG
SA14S-174Rc	AAT TCC TTC TCT ATC ACT GAG
SA14M-2876Fc	GGG ATG CTT AGG GAC C
MURM-491Rc	CAR CTY TTC ATC ATR TTR CA
SA14L-6387Fc	GGA TTT GCC CAG AGG ATG TTG
SA14L-74Rc	CAA TGA GTT TAT GTG GTG AGC

^aMUR or SA14 in the primer name indicates whether the primer was designed from multiple-sequence alignment of Murinae-associated hantavirus (MUR) or from the available SANGV/SA14 sequence (SA14). The next letter (S, M, or L) indicates the targeted genome segment. The number indicate the binding position. The last letter indicates the forward (F) or reverse (R) orientation of the primer.

^bThe lowercase letters in the primer sequence indicate the 5′ tail of the heterologous sequence used as a stand-alone primer in the subsequent PCR reamplification. R = A or G; Y = C or T; M = A or C; S = G or C; K = G or T; W = A or T; N = A, G, C, or T; I = inosine.

^cPrimers for determination of exact 5′ and 3′ termini of the genome segments.

To determine the exact 5′ and 3′ termini of the SANGV genome segments, viral RNA was first ligated using T4 RNA ligase (Fermentas), and PCR fragments spanning the ligated ends were amplified (Table 1) and sequenced as described above.

Small mammal trapping and screening.

Small mammals were trapped in Guinea using Sherman LFA live traps (H.B. Sherman Traps, Inc.) by the method of Fichet-Calvet et al. (5). Total RNA was extracted from blood samples (preserved in liquid nitrogen) with the Blood RNA kit (Peqlab). For hantavirus screening, the L-segment-based nested reverse transcription-PCR assay was used as described previously (17).

Sequence and phylogenetic analyses.

The overlapping nucleic acid sequences obtained were combined for analysis and edited with the aid of the SEQMAN program from the Lasergene software package (DNASTAR). The sequence data were further analyzed by using the BioEdit software package (10). Multiple-sequence alignments were constructed using the MUSCLE program (3) implemented in MEGA5 (42). Evolutionary analyses were also conducted by using MEGA5. The evolutionary histories were inferred by using the maximum likelihood method based on the Tamura-Nei model (41) using a discrete Gamma distribution (+G) with 5 rate categories and by assuming that a certain fraction of sites are evolutionarily invariable (+I).

Antibody blocking experiments.

Antibodies against α₅β₁ integrin (mouse monoclonal antibody [MAb] MAB1969; Millipore), α_Vβ₃ integrin (mouse MAb MAB1976; Millipore), and decay-accelerating factor (DAF)/CD55 (rabbit polyclonal antibody H319; Santa Cruz) were added to confluent Vero E6 cells and HUVEC. The cells were treated with 40 μg/ml of antibodies for 1 h at 4°C. The hantavirus inocula (multiplicity of infection of 0.05) were then added to the monolayer. After incubation for 1 h at 37°C, the cells were washed with medium and incubated for 24 h at 37°C. Samples were taken for reverse transcription-quantitative PCR (RT-qPCR) and Western blot analyses. The degree of antibody inhibition of infection was calculated in comparison to untreated, infected cells.

Receptor binding experiments.

CHO and stably transfected CHO cells stably expressing β₃ integrins (CHO-β₃ cells) were grown until confluence and cooled down to 4°C. Virus suspension (MOI of 0.5) kept at 4°C was added to the cells and incubated for 1 h in a refrigerator. After incubation, the cells were washed five times with cold medium, and bound virus was lysed by RLT buffer from QIAamp RNeasy minikit (Qiagen) and used for RNA extraction according to the manufacturer's specifications. The binding affinity of virus particles to CHO-β₃ cells was measured by RT-qPCR as a ratio between the number of virus genome equivalents detected on CHO-β₃ cells in comparison to those detected on nontransfected CHO cells.

Hantavirus quantitative RT-PCR detection.

RNA samples from cell culture supernatants were extracted by using the QIAamp viral RNA minikit (Qiagen), while samples from infected cells were extracted by using the RNeasy minikit (Qiagen).

For virus binding experiments with CHO cells, total RNA was reverse transcribed with random hexamers in order to measure the number of viral particles. RT-qPCR was performed with virus-specific primers and probes for HTNV using a Light Cycler system by the method of Kramski et al. (22). For SANGV and PHV, the following primers and probes were used: SANGV F (F stands for forward), AGGCTGTCAGACAACAAGCA; SANGV R (R stands for reverse), GCTCCTGCAAATACCCAAAT; SANGV TAQ, 6FAM-TGGACCACATTGACTCACCATCATCA-TAMRA; PHV F, AGGAAGAGATCACTCGCCAT; PHV R, TCCAATGTTGACACTGCTGA; and PHV TAQ, 6FAM-CATTGCCCGGCAGAAGCTCA-TAMRA.

For antibody blocking experiments with Vero E6 cells, cDNA was synthesized with positive-strand-specific primers in order to monitor de novo-synthesized viral mRNA. Briefly, specific reverse primers from RT-qPCR described above were modified by the addition of a 5′ anchor sequence gACCATCTAgCgACCTCCAC. In the case of the SANGV reverse primer, the addition of the anchor sequence led to the prediction of strong secondary structure, and therefore, a completely new primer containing the anchor sequence was designed and used (SANGV RT, gACCATCTAgCgACCTCCAC-ACAgCgATCAggTgCTCC). In the successive RT-qPCR, primer containing the anchor sequence was used instead of the specific reverse primers to ensure that only molecules transcribed by the positive-strand-specific primer (and therefore containing the anchor sequence) are detected. The results obtained from cellular RNA were presented as the fold change in gene expression (33) normalized to phosphobilinogen deaminase (PBGD) as the housekeeping gene (36).

Western blot analysis of N-protein expression.

Vero E6 cells and HUVEC were harvested at the indicated time points after infection. Protein extracts were separated in a 10% SDS-polyacrylamide gel and blotted onto a nitrocellulose membrane (Whatman). Hantavirus nucleocapsid (N) protein was detected with N-protein-specific polyclonal rabbit serum and β-actin-specific mouse MAb SC69879 (Santa Cruz). Signals were visualized by using Chemiluminescence Super Signal West Dura kit according to the protocol supplied by the manufacturer (Thermo Scientific).

The densities of N-protein and β-actin (reference protein) bands on Western blots were quantified by using the ImageJ 1.41o program (Wayne Rasband, National Institutes of Health, Bethesda, MD). The expression of N protein was normalized to the expression of β-actin. The percentage of antibody-mediated inhibition of viral infection (N-protein expression) was calculated in comparison to untreated but infected cells.

ELISA of IFN-λ1 and antibody blocking studies.

Virus stocks derived from Vero E6 cells and Vero E6 cell conditioned medium (served as a negative control) were exposed to UV irradiation (described above). The amount of IFN-λ1 was measured by ELISA using the human interleukin 29 (IL-29) DuoSet ELISA development system (DY1598; R&D Systems), following the manufacturer's instructions and the modifications of Prescott et al. (35).

For antibody blocking studies, virus or recombinant proteins were incubated with 1 μg/well of anti-IFN-λ1 or anti-IFN-β antibodies (catalog no. AF1598 or AF814, respectively; R&D Systems) for 1 h at room temperature. The resulting mixture was then added to A549 cells and incubated for 16 h prior to the measurement of MxA mRNA expression. Recombinant proteins used included 2 ng/well of recombinant human IFN-λ1 (rhIFN-λ1) or rhIFN-β (catalog no. 1598-IL-025 or 11415-1, respectively; R&D Systems).

Monitoring of MxA RNA expression by RT-qPCR.

A549 cells were seeded in 12-well plates at a density to achieve 90 to 95% confluence after overnight incubation at culture conditions. The cells were infected with an MOI of 1, and 16 h postinfection, RNA was isolated using the RNeasy kit (Qiagen). Extracted RNA was subjected to DNase digestion following the protocol provided by the manufacturer. Purified RNA was reverse transcribed by using a Moloney murine leukemia virus (M-MLV) reverse transcription kit with random hexamer primers (Invitrogen). MxA-relative mRNA expression was quantified by QuantiTest Sybr green PCR kit (Qiagen) using the manufacturer's protocol with mxA gene-specific primers (MxA F [gAggAgATCTTTCAgCACCTgAT] and MxA R [CTggATgATCAAAgggATgTggC]) and an annealing temperature of 55°C. Data are presented as the fold change in gene expression normalized to PBGD as described above for hantavirus quantitative RT-PCR detection.

Nucleotide sequence accession numbers.

All sequences obtained in this study have been submitted to the GenBank database under accession numbers JQ082300 to JQ082305.

RESULTS

Virus isolation and whole-genome characterization.

In the initial study, SANGV RNA was detected in the blood of an African wood mouse (Hylomyscus simus) and designated SA14 (17). A suspension of spleen tissue from this animal was used for isolation attempts on the Vero E6 cell line. Hantavirus antigen was detected in the inoculated cells 42 days p.i. (3rd passage) in immunofluorescence tests using human DOBV-specific convalescent-phase serum. Successful isolation of the virus was then confirmed by RT-PCR. The first virus stock was prepared from the 6th passage of the virus, and its infectious virus titer was determined by a chemiluminescence focus assay to be 3 × 10⁴ focus-forming units (FFU)/ml. Similarly low virus titers were observed after further consecutive virus passages. Virus growth curves were prepared in Vero E6, A549, and Huh7 cells. The highest titers of up to 10⁵ FFU/ml were observed in Vero E6 cells 7 days p.i. (data not shown).

Availability of the cell culture isolate enabled determination of the complete viral genome sequence. The SANGV S segment was found to be 1,746 nucleotides (nt) long and to contain a single open reading frame (ORF) of 1,290 nt (positions 46 to 1335) that encodes the 429-amino-acid (aa)-long putative nucleocapsid (N) protein. The complete M segment consists of 3,650 nt and again carries a single ORF (3,408 nt; positions 41 to 3448) of the putative 1,135-aa-long glycoprotein precursor (GPC). The complete L-segment sequence was also determined and was found to be 6,531 nt long with a single ORF (6,456 nt; positions 38 to 6493) which encodes the putative 2,151-aa-long RNA-dependent RNA polymerase (L protein). Functionally relevant and highly conserved regions such as the RNA binding domain of the N protein (amino acid positions 175 to 217) and the WAASA motif determining the cleavage of GPC (amino acid positions 656 to 660) were identified in SANGV amino acid sequences. Highly conserved amino acids in the A to E motifs of the L-protein polymerase domain (amino acid positions 964 to 1149) were also found in the SANGVL amino acid sequences.

In order to directly infer the 5′ and 3′ termini of the SANGV genome, RNA ligation of viral RNA was performed, and PCR fragments spanning the ligated ends were prepared and sequenced for all three segments. Although slightly different from HTNV and other hantaviruses, the sequences obtained were predicted to form panhandle-like structures of 17 to 19 bp typical of bunyaviruses. As for other hantaviruses, the complementarity is incomplete with a mismatch at position 9 and a noncanonical U-G pair at position 10 in all three segments (Fig. 1). In the M-segment termini of SANGV, one additional noncanonical pair instead of a perfect match found in HTNV (34) was detected at position 15.

Fig 1.

Panhandle-forming terminal nucleotides (17 to 19 bp long) of the Sangassou virus (SANGV) RNA genomic segments. Complementary pairing (|) and noncanonical U-G pairs (:) are shown.

Comparison of the obtained complete nucleotide and amino acid sequences with other hantavirus representatives revealed rather low percentage identity values (Table 2). SANGV shares the highest similarity with DOBV, followed by SEOV and HTNV. For all three segments, even the smallest amino acid sequence difference is markedly higher than 7% (11.5%, 19.6%, and 13.3% for DOBV N, GPC, and L proteins, respectively). In the case of the S-segment 3′ noncoding region (which is usually highly conserved within a hantavirus species but shows dramatic differences in length and nucleotide composition between the species), sequence percent identity values dropped notably to values around 30 to 60% (data not shown).

Table 2.

Complete nucleotide and amino acid sequence identities of the SANGV isolate SA14 compared with other rodent-borne hantavirus representatives

Hantavirus	% identitya
	S segment		M segment		L segment
	1,746 nt	429 aa	3,650 nt	1,135 aa	6,531 nt	2,151 aa
DOBVSK/Aa	74.6	88.5	71.6	80.4	75.0	86.7
HTNV76–118	67.8	81.8	68.7	76.2	73.9	84.3
SEOV80–39	71.1	82.7	69.8	77.7	74.9	85.2
PUUVCG1820	54.8	61.2	56.6	52.9	65.7	68.3
TULVMoravia	54.1	61.7	56.7	53.7	65.2	68.5
SNVNM H10	47.9	60.8	56.2	52.8	65.4	69.2
ANDVChile	51.0	62.2	56.3	53.3	65.1	68.2

^ant, nucleotides; aa, amino acids.

Phylogenetic analysis.

During the course of our project to detect additional African hantaviruses, 340 additional samples (spleen tissue samples) from small mammals belonging to 16 rodent and shrew genera captured on 17 trapping sites across Guinea (West Africa) were screened for the presence of hantavirus RNA by the genus-specific RT-PCR previously used to initially detect SANGV (17). A second sample, designated SA22, was found to be positive and again was from a Hylomyscus simus animal that was trapped near Sangassou village. In addition to the partial L-segment sequence (347 nt) obtained from the screening PCR, complete S-segment (1,747-nt) and partial M-segment (533-nt) sequences from SA22 were prepared and analyzed. In nucleotide sequence comparisons, these sequences were found to be highly similar but not identical to the sequences of SANGV SA14 isolate (nucleotide sequence identity values of 97.8%, 95.1%, and 96.3% for S-, M-, and L-segment sequences, respectively).

In addition, in the maximum likelihood (ML) phylogenetic tree based on complete S-segment ORF sequences, SA22 was shown to be very closely related but not identical to SA14 and thus represented the second strain of SANGV (Fig. 2A). The same position of SA22 could also be found in phylogenetic trees constructed from partial M- and L-segment sequence data sets (data not shown).

Fig 2.

Maximum likelihood trees showing the phylogenetic placement of SANGV within the Murinae-associated hantaviruses constructed on the basis of complete S (A), M (B), and L (C) segment coding sequences. Evolutionary analyses were conducted with MEGA5 (37). The evolutionary histories were inferred by using the maximum likelihood method based on the Tamura-Nei model (36) using a discrete Gamma distribution (+G) with 5 rate categories and by assuming that a certain fraction of sites are evolutionarily invariable (+I). Bars indicate evolutionary distance of 0.1 substitution per position in the sequence. Bootstrap values of ≥70%, calculated from 1,000 replicates, are shown at the tree branches. SANGV is marked by a dark gray box and an arrow. Other Murinae-associated hantaviruses are marked by light gray boxes. Virus abbreviations: ANDV, Andes virus; BAYV, Bayou virus; DOBV, Dobrava-Belgrade virus; HTNV, Hantaan virus; PHV, Prospect Hill virus; PUUV, Puumala virus; SANGV, Sangassou virus; SEOV, Seoul virus; SNV, Sin Nombre virus; THAIV, Thailand virus; TULV, Tula virus.

The results of this analysis including two complete SANGV S-segment sequences confirmed the initial classification of SA14 based on partial S-segment sequences (17), showing that SANGV belongs to the Murinae-associated hantaviruses. SANGV and DOBV, which is present in Europe, share a common ancestor. HTNV is the most closely related virus within the Asian hantaviruses, while SEOV and Thailand virus occupy more ancestral positions within the group of Murinae-associated hantaviruses (Fig. 2A). In the evolutionary trees based on complete M-segment ORF sequences (Fig. 2B) and complete L-segment ORF sequences (Fig. 2C), the position of SANGV as a sister group of DOBV remained unchanged, although the resolution and statistical support for the overall tree topology of Murinae-associated hantaviruses were not stable for all three data sets.

SANGV receptor usage.

Antibody blocking and receptor binding experiments were performed in order to determine SANGV cellular receptor usage. First, antibodies against α₅β₁ integrin, α_Vβ₃ integrin, and DAF/CD55 were used to treat Vero E6 cells at 4°C 1 h prior to infection. The level of infection was then determined by Western blotting (staining for hantavirus N protein) and RT-qPCR (measurement of S-segment mRNA copies) 1 day p.i. HTNV and PHV were used as controls, since they were reported to be inhibited by β₃ and β₁ integrin-specific antibodies, respectively (7). The efficiency of antibody blocking was evaluated as a percentage of virus infection inhibition in comparison to untreated infected cells. The results of both RT-qPCR (Fig. 3A) and Western blotting (Fig. 3B) clearly indicated that SANGV infection can be efficiently blocked only by anti-α₅β₁ integrin MAb. In agreement with previous studies, HTNV infection was more efficiently blocked by β₃- integrin-specific MAb, while PHV was more efficiently blocked by β₁ integrin-specific MAb. Interestingly, although infection by both control viruses could also be efficiently inhibited by anti-DAF/CD55 MAb, this was not the case for SANGV, which could be significantly inhibited exclusively by β₁ integrin-specific antibodies. These findings were further confirmed on HUVEC, which are often used as a model representing a natural target cell of hantaviruses (Fig. 3C). In HUVEC, the differences between SANGV inhibition by β₁- and β₃ integrin-specific antibodies was not as pronounced as in Vero E6 cells.

Fig 3.

Inhibition of SANGV infection in vitro by application of specific monoclonal antibodies against β₁ integrin, β₃ integrin, and decay-accelerating factor (DAF) in Vero E6 cells. Efficiency of inhibition was monitored in comparison to HTNV and PHV by hantavirus-specific RT-qPCR (A) and Western blot analysis of the hantavirus N protein (B). In addition, the efficiency of inhibition of SANGV infection in human umbilical vein endothelial cells (HUVEC) was monitored by Western blot analysis (C). The cells were pretreated with 40 μg/ml of antibodies for 1 h at 4°C and then infected (MOI of 0.05). After incubation for 1 h at 37°C, the cells were washed with medium and incubated for 24 h at 37°C. The degree of antibody inhibition of infection was calculated in comparison to untreated, infected cells. Error bars represent standard deviations of the means from three experiments. (D) Binding of SANGV to CHO cells stably expressing β₃ integrins (CHO-β₃ cells) in comparison to normal CHO cells. Virus binding was performed at 4°C for 1 h. HTNV and PHV were used as controls. The amount of bound virus was measured through detection of viral RNA by specific qPCR. The binding affinity of virus particles to CHO-β₃ cells is expressed as a ratio between virus genome equivalents detected on CHO-β₃ cells and virus genome equivalents detected on normal CHO cells. Error bars represent standard deviations of the means from three experiments.

To unequivocally exclude β₃ integrin as a potential SANGV receptor, we performed an additional experiment where virus binding to β₃ integrins was examined by comparison of virus binding to CHO cells (expressing no integrins on their surface) and CHO-β₃ cells stably expressing β₃ integrins (Fig. 3D). In this experimental setup, for HTNV, around 7-fold-higher level of virus genome copies was detected on CHO-β₃ cells than on CHO cells. In contrast, no significant difference in the amount of viral RNA detected was observed for both SANGV and PHV. Unfortunately, recombinant CHO cells expressing β₁ integrins were not available for this experiment.

In aggregate, the data clearly indicate that SANGV does not bind to β₃ integrins and therefore does not use them as a receptor. Instead, SANGV infection was efficiently blocked by anti-β₁ integrin antibodies, suggesting the usage of β₁ integrin as a receptor for SANGV.

Induction of IFN-λ in response to SANGV infection.

Recently, it was shown by others (35, 40) that hantaviruses can also induce lambda interferon (IFN-λ) in IFN type I-deficient Vero E6 cells. Thereby, IFN-λ might be present in hantavirus stocks generated in these cells. To evaluate the capability of SANGV to induce IFN-λ secretion in Vero E6 cells, we determined IFN-λ concentration by ELISA in regular SANGV stocks prepared from cell culture supernatants of infected Vero E6 cells (SANGV-E6). Indeed, we measured a high IFN-λ1 concentration of about 2 ng/ml in several SANGV-E6 stocks (Fig. 4A). On the other hand, no detectable amounts of IFN-λ1 were found in stocks where the Vero E6 cell-derived supernatants were replaced by fresh cell culture medium after ultracentrifugation (SANGV-ucf). Interestingly, a comparable concentration of IFN-λ1 was measured in analyzed PHV stocks derived from Vero E6 cells, while HTNV stocks did not contain detectable amounts of IFN-λ1 (Fig. 4A).

Fig 4.

Sangassou virus induces IFN-λ expression in IFN type I-deficient Vero E6 cells. IFN-λ is responsible for early activation of the MxA gene in infected A549 cells. (A) Concentration of IFN-λ1 present in virus stocks measured by ELISA. Virus stocks prepared from cell culture supernatants of infected Vero E6 cells are indicated by “E6” at the end of the virus name. Ultracentrifuged stocks, where the supernatants derived from Vero E6 cells were replaced by fresh cell culture medium, are marked with “ ucf” at the end of the virus name. Vero E6 cell-derived medium served as a negative control (NC). Data are presented as the means plus standard deviations of the means (error bars) from three independently prepared stocks. (B) SANGV-E6 virus induces early MxA mRNA expression in A549 cells which can be prevented by preincubation of the stock with blocking anti-IFN-lambda antibody. SANGV stocks (∼150 μl; MOI of 1) or recombinant proteins rhIFN-λ1 and rhIFN-β (2 ng) were preincubated with 1 μg of the corresponding blocking antibody (anti-IFN-λ1 or anti-IFN-β). One hour after incubation, the resulting mixture was added to the top of A549 cells. Sixteen hours postinfection, MxA mRNA expression was measured by RT-qPCR, and expressed as fold induction in comparison to untreated A549 cells taken as a negative control (NC). Error bars represent standard deviations of the means from three experiments.

High concentrations of IFN-λ1 detected in SANGV-E6 stocks were shown to have substantial consequences during in vitro studies of cellular innate immunity responses. In contrast to SANGV-ucf, the SANGV-E6 stock showed high-level induction of MxA protein, encoded by interferon-stimulated mxA gene, in A549 cells 16 h p.i. Moreover, this activity could be almost completely abolished by pretreatment of the stock with anti-IFN-λ1 MAb, while pretreatment with IFN-β-specific antibody had no effect (Fig. 4B). In the control experiments, both recombinant IFN-λ1 and IFN-β showed similar levels of MxA mRNA induction in A549 cells, which could be significantly reduced by their preincubation with specific anti-IFN-λ1 and anti-IFN-β antibodies, respectively.

Altogether, these data showed that SANGV strongly induces IFN-λ expression in IFN type I-deficient Vero E6 cells. Early activation of the mxA gene in A549 cells infected with SANGV-E6 is due to IFN-λ present in the viral preparation.

DISCUSSION

We detected SANGV, the first indigenous African hantavirus, 5 years ago by nested RT-PCR screening of small mammal samples from Guinea, followed by sequencing the amplified product but without isolating the virus (17). In this study, we report isolating the virus in Vero E6 cells. The availability of a cell culture isolate allowed us to perform further genetic and functional characterization of the new virus.

Although membership of SANGV within the Murinae-associated hantaviruses could clearly be demonstrated by phylogenetic analysis of partial virus sequences in the primary report on SANGV discovery (17), its detection in 1 out of 4 African wood mouse (Hylomyscus simus) animals did not provide irrefutable proof that H. simus was the natural host of SANGV. The possibility of random “spillover” infection from other rodent species of the Murinae subfamily could not be excluded. Here we found SANGV sequences (SA22) in 1 out of 2 additionally investigated H. simus animals. Altogether, SANGV has been detected in 2 out of 6 H. simus individuals in two independent screenings involving a total of 1,649 animals from 19 rodent (mostly of the Murinae subfamily) and shrew genera trapped in Guinea. Given the large number of other Murinae rodents tested, it is highly unlikely that H. simus was only randomly (transiently) infected and that the real natural host of the virus remained unknown. Detection of SA22 as a second SANGV strain, again in H. simus, considerably increased the probability that H. simus is a natural reservoir of SANGV. Nevertheless, further studies need to be performed to more definitely identify the reservoir species. Species of the genera Mastomys, Praomys, Nannomys, Crocidura, Lophuromys, Hybomys, and Paraxerus have been trapped in the same locality (Sangassou village) in Guinea and therefore represent additional putative SANGV hosts.

Determination of the full genome sequence of SANGV allowed a more accurate and comprehensive phylogenetic analysis based on the coding sequences of all three genomic segments. In all cases, SANGV was shown to be most closely related to DOBV, the most virulent European hantavirus. In S-segment analysis, where the largest data set of complete sequences is currently available, SANGV and DOBV share an ancestor with HTNV, suggesting that DOBV and SANGV evolved when the pre-HTNV ancestor expanded in its rodent host from Asia to Europe and Africa, respectively. Notably, in a recent phylogenetic study of African Murinae rodents, lineages including Hylomyscus and Apodemus genera were placed in the same evolutionary clade (26). Given the large biodiversity of Murinae rodents in Africa, it is likely that several African Murinae-associated hantaviruses are yet to be discovered.

A SANGV isolate was used to develop diagnostic tools that were used to assess public health relevance of hantaviruses in West Africa. An enzyme-linked immunosorbent assay (ELISA) based on an immunofluorescence assay using SANGV recombinant nucleocapsid protein was established. These assays and the application of viable virus in a focus reduction neutralization test (FRNT) allowed the use of the full spectrum of serodiagnostic methods on the basis of SANGV antigen. The data obtained clearly showed that hantaviruses infect humans and might be an unrecognized medical problem at least in Guinea (19).

Given the preliminary evidence of human SANGV infections in Guinea based on seroprevalence data, we analyzed properties that are considered to play a role in hantavirus pathogenesis. Most hantaviruses (HTNV, SEOV, PUUV, SNV, and New York virus [NYV]) are reported to use β₃ integrins as a receptor for virus entry (7, 8). β₃ integrins have primary roles in regulating vascular integrity, endothelial cell permeability, and platelet function. Pathogenic hantaviruses were reported to bind plexin, semaphorin, and integrin (PSI) domains on inactive (bent) conformations of β₃ integrin subunits, block endothelial cell migration, and enhance the permeability of endothelial cells in response to vascular endothelial growth factor (VEGF) (9, 27, 37). On the other hand, PHV and TULV using β₁ integrins as receptors had no effect on the ability of endothelial cells to migrate on either β₃ or β₁ integrin ligands (9) and on endothelial cell permeability (6). Our blocking and binding studies indicated that β₁ integrin is most likely used for entry, as shown also for PHV and TULV (8, 28). SANGV thereby seems to be the first Murinae-associated virus shown to be using β₁ integrins as a receptor.

In addition to β₁ and β₃ integrins, DAF/CD55 was identified as a coreceptor at least for HTNV and PUUV to enter the apical surface of polarized cells (24). In our experimental system, both control viruses, HTNV and PHV, could be efficiently inhibited by DAF/CD55-specific MAb. This confirmed the findings of Krautkrämer and Zeier (24) for HTNV and extended them also to PHV. Intriguingly, SANGV could not be efficiently blocked by anti-DAF/CD55 MAb. β₁ integrins therefore seem to be exclusive receptors for SANGV. Regarding receptor usage, SANGV clearly differs not only from the β₃ integrin-using viruses but also from PHV. Whether these differences have some consequences for virus pathogenicity remains to be determined.

According to the use of β₁ integrin, but not β₃ integrin, as a receptor, SANGV might be assumed to be a nonpathogenic hantavirus. However, hantavirus pathogenesis is a multifactorial process that includes contributions from immune responses (immune complexes, complement activation, cytotoxic T cells, etc.), platelet dysfunction, and the dysregulation of endothelial cell barrier functions (reviewed in reference 27). It must be clarified how strictly hantavirus pathogenicity is dependent on β₃ integrin dysfunction-mediated increase of vascular permeability. Increased capillary permeability may also be caused by hantavirus-specific cytotoxic T cells attacking endothelial cells which present viral antigens on their surface (reviewed in reference 43). Nevertheless, our recent seroepidemiological study showed the presence of SANGV-specific neutralizing antibodies in patients with fevers of unknown origin, suggesting that SANGV is pathogenic for humans (19), but final evidence of detection of SANGV RNA in specimens from patients has not yet been obtained.

Type III IFNs (IFN-λ1 to -λ3 also designated IL-29, IL-28A, and IL-28B, respectively) were recently discovered and shown to have antiviral effects (21, 39). Type III IFNs are transcribed independently of type I IFNs (IFN-α/β) and use a different heterodimeric receptor. However, both IFNs activate the Jak/STAT (signal transducer and activator of transcription factor) signal transduction cascade and therefore can induce a similar subset of interferon-stimulated genes (ISGs). Recently, Prescott et al. (35) reported that IFN type I-deficient Vero E6 cells are able to secrete IFN-λ in response to infection by SNV, ANDV, and PHV. This response can in turn elicit downstream biological processes. Moreover, induction of IFN-λ in Vero E6 cells (and A549 cells) was also shown on the transcriptional level for HTNV (40). In agreement with these studies, we detected comparably large amounts of IFN-λ in PHV stocks and barely detectable amounts of IFN-λ in HTNV preparations. In contrast to HTNV, another Murinae-associated hantavirus, SANGV can elicit strong IFN-λ excretion in Vero E6 cells. The range of hantaviruses reported to induce secretion of abundant IFN-λ in Vero E6 cells by SANGV is extended also to the Murinae-associated hantaviruses. Strong IFN-λ expression might also explain low virus titers which SANGV regularly reaches in Vero E6 cells.

SANGV is the first indigenous hantavirus discovered in Africa. Although its pathogenic potential and public health relevance have yet to be determined, it shows some unique properties within Murinae-associated hantaviruses, such as the use of β₁ integrins as receptors and strong induction of IFN-λ, which warrant comparative studies focusing on hantavirus pathogenesis.