Divergent Simian T-Cell Lymphotropic Virus Type 3 (STLV-3) in Wild-Caught Papio hamadryas papio from Senegal: Widespread Distribution of STLV-3 in Africa

Abstract

Among eight samples obtained from a French primatology research center, six adult guinea baboons (Papio hamadryas papio), caught in the wild in Senegal, had a peculiar human T-cell leukemia virus type 2 (HTLV-2)-like Western blot seroreactivity (p24⁺, GD21⁺, K55^+/−). Partial sequence analyses of the tax genes (433 bp) indicated that these baboons were infected by a novel divergent simian T-cell lymphotropic virus (STLV). Analyses of the complete proviral sequence (8,892 bp) for one of these strains (STLV-3/PPA-F3) indicate that this STLV was highly divergent from the HTLV-1 (61.6% of nucleotide similarity), HTLV-2 (61.2%), or STLV-2 (60.6%) prototype. It was, however, much more closely related to the few other known STLV-3 strains, exhibiting 87 and 89% of nucleotide similarity with STLV-3/PHA-PH969 (formerly PTLV-L/PH969) and STLV-3/CTO-604, respectively. The STLV-3/PPA-F3 sequence possesses the major HTLV or STLV open reading frames corresponding to the structural, enzymatic, and regulatory proteins. However, its long terminal repeat comprises only two 21-bp repeats. In all phylogenetic analyses, STLV-3/PPA-F3 clustered together in a highly supported single clade with the other known strains of STLV-3, indicating an independent evolution from primate T-cell lymphotropic virus type 1 (PTLV-1) and PTLV-2. The finding of a new strain of STLV-3 in a West African monkey (Guinea baboon) greatly enlarges the geographical distribution and the host range of species infected by this PTLV type in the African continent. The recent discovery of several different STLV-3 strains in many different African monkey species, often in contact with humans, strongly suggests potential interspecies transmission events, as it was described for STLV-1, between nonhuman primates but also to humans.

The primate T-cell lymphotropic viruses (PTLVs), which include human T-cell leukemia virus type 1 (HTLV-1) (21), simian T-cell lymphotropic virus type 1 (STLV-1) (I. Miyoshi, S. Yoshimoto, M. Fujishita, H. Taguchi, I. Kubonishi, K. Niiya, and M. Minezawa, Letter, Lancet ii:658, 1982), HTLV-2 (9), STLV-2 (5, 34), and STLV-3 (6), constitute a group of related human and simian retroviruses, sharing common biological and molecular features. Nevertheless, their origin and evolutionary relationship, as well as their modes of dissemination, are still unclear and a matter of discussion (2, 4, 10, 26, 29). While STLV-1 has been detected and characterized in many different African and Asian monkey and ape species (Miyoshi et al., letter; 3, 7, 8, 11-16, 18-20, 22-25, 27, 37-41), including several isolates from wild-caught animals, STLV-2 has been found only in two colonies of bonobos (Pan paniscus) kept either in a zoo (33) or in a primate center colony (5). As for STLV-3, only one strain of its type was known until 2001. This strain, PHA-PH969 (originally called PTLV-L/PH969), was isolated in 1994 from an Eritrean baboon (Papio hamadryas) kept in a captive colony in Leuven, Belgium (6). Very recently, a few other STLV-3 strains (all different) were characterized in different wild-caught monkey species living either in East (Ethiopia) or Central (Cameroon) Africa. Regarding the latter country, these divergent viruses were found in two Cercocebus torquatus species (17) (strains STLV-3/CTO-604 and STLV-3/CTO-602) and in two Cercopithecus nictitans species (36) (strains STLV-3 CNI-227 and STLV-3 CNI-217). Regarding East Africa, a serological survey in wilderness-dwelling nonhuman primates performed in Ethiopia reported a high prevalence of STLV-3, confirmed by PCR, among P. hamadryas (22 of 40 animals tested) and P. hamadryas/anubis hybrid baboons (19 of 50) (30). In this study, the authors have only sequenced a small part of the tax gene (180 bp) in a P. hamadryas hamadryas (PHA-7550) and in two baboon hybrids, HYB-2210 and HYB-2220.

The goal of this study was to search for novel divergent STLV strains among African monkeys or apes, especially those originating from West Africa, and from species for which no data on STLV were available. In September 2001, we had the opportunity to test a small colony of eight P. hamadryas papio. These animals (Table 1) were kept since 1988 in a French primatology center for ethological research (Joël Fagot, Centre National de la Recherche Scientifique cognitive center, Marseille, France). They all originated from West Africa (Senegal) where they were all caught in the wild. Since then, they have never been in close contact with any other monkey species or apes and have never been inoculated with any human or simian materials.

TABLE 1.

HTLV and STLV serological (immunofluorescence assay and Western blotting) and molecular (tax and pol PCR) data of the eight wild-caught P. hamadryas papio from Senegal^a

Animalc	Sex	Immunofluorescence assay titerd		Western blot patterne	Results of:
					tax PCRb		pol PCR
		MT2	C19		KKPX1-KKPX2	SK43-SK44	SK110-SK111
PPA-F3	Male	160	160	GD21, p24, p36f, p53f, K55, MTA-1	+	−	−
PPA-F4	Female	20	160	GD21, p24	+	−	−
PPA-F5	Male	80	160	GD21, p24, p36f, p53f	+	−	−
PPA-F6	Female	—	—	Negative	−	−	−
PPA-F7	Male	40	80	GD21, p24, p36f, p53f	+	−	−
PPA-F8	Female	320	320	GD21, p24, p36, p53, K55	+	−	−
PPA-F9	Male	80	160	GD21, p24, p36, p53	+	−	−
PPA-F11	Male	—	—	Negative	−	−	−

^aAmplification products were hybridized in Southern blotting with γ-³²P-radiolabeled probes SK45 (tax) and SK112 or SK188 (pol).

^b+, present; −, absent.

^cAll animals were more than 10 years old.

^d—, negative.

^ef, faint.

Serological and molecular studies were performed to search for the presence of HTLV or STLV infection in these eight animals. Among them, six scored positive with the screening assay (immunofluorescence) on both HTLV-1 (MT2)- and HTLV-2 (C19)-producing cell lines with slightly higher titers on C19 cells (Table 1). Furthermore, these sera exhibited an HTLV-2-like Western blot (HTLV2-4; Diagnostic Biotechnology) pattern, with strong reactivities against GD21 and p24 that were associated or not associated with faint reactivities against p53 (Fig. 1). Two sera (F3 and F8) exhibited reactivities against the HTLV-2 gp46 specific peptide, K55 (Fig. 1). There was a good correlation between the antibody titer level and the intensity of the Western blot reactivity (Table 1). Peripheral blood mononuclear cells (PBMCs) obtained from three animals were kept in culture in the presence of interleukin 2 for 4 to 5 weeks with a low growth rate, but no long-term culture could be established. No fluorescent cells could be detected in these cultured cells by immunofluorescence assay using either different polyclonal HTLV-1 and/or HTLV-2 sera or plasma of the infected animals or monoclonal antibodies directed against HTLV-1 p19 and p24 antigens. DNAs extracted from uncultured PBMCs were tested by three PCR experiments with HTLV-1 and/or HTLV-2 pol (SK110-SK111) or tax (SK43-SK44 and KKPX1-KKPX2) primers and probes using previously described PCR conditions (15). As seen in Table 1, solely the KKPX1 and -2 set of tax primers gave a positive signal that hybridized specifically by Southern blot only with a highly conserved probe (SK45) that could detect HTLV-1 and HTLV-2 as well as STLV-3. This suggested that these baboons were infected by a divergent STLV strain. By use of highly conserved primers (KKPX1-MacArc4AS) (16), a 433-bp sequence (tax sequence) was obtained and found nearly identical (only a 1-bp difference in one strain) for the four baboons studied (PPA-F3, PPA-F4, PPA-F5, and PPA-F9). Comparison with all the available PTLV prototype sequences indicated that this 433-bp fragment exhibited high similarities with the two other known STLV-3 sequences (96% with STLV-3/CTO-604 and 95.1% with STLV-3/PHA-PH969), while it was more divergent but roughly equidistant from all the other PTLV prototypes, including HTLV-1 ATK (79.6%) and HTLV-2 MO (79.6%).

FIG. 1.

Western blot serological pattern of the eight wild-caught P. hamadryas papio from Senegal using the Western blot from Diagnostic Biotechnology (HTLV blot 2.4 version).

The complete sequence (8,892 bp) was then obtained for one strain (STLV-3/PPA-F3) after PCR amplification on cellular DNA from uncultured PBMCs and cloning of the nine proviral fragments, as previously described (17). Five amplification products were obtained with primers that have been previously used for amplifying the STLV-3/CTO-604 sequence (fragments B, C, D, E, and I). The four other proviral fragments (A, F, H, and G) were obtained, with both specific and consensus primers: F3As (sense, 5′-GGGGGGCTTTGTTCCGCTCGG-3′) and F3Aas (antisense, 5′-CCTTCCTTGGGGCAAGGGCCG-3′) at 61°C (annealing temperature); F3Fs (sense, 5′-CGTCACTCACCTCCAGTACAAGCG-3′) and F3Fas (antisense, 5′-GGGCATTGTATAGCGCAGGGG-3′) at 58°C; F3Hints (sense, 5′-CCCAGCAGAAGCAGCCTATCC-3′) and F3Has (antisense, 5′-CGGAGAGCGAGATAGAGCTGG-3′) at 57°C; and F3G2s (sense, 5′-GGCTTATTCCCAAAAACTACCCGGG-3′) and F3G2as (antisense, 5′-GGAGTGCTTGAATGCTAACTGGGG-3′) at 61°C. The comparison of the PPA-F3 complete proviral sequence (8,892 bp) with the other PTLV prototypes indicated that, although unique, this new strain was more closely related to the STLV-3/PHA-PH969 (87.9% similarity on the complete sequence) or to STLV-3/CTO-604 (89.4%) than to HTLV-1 (61.6%), HTLV-2 (61.2%), or STLV-2 (60.6%) prototypes (Table 2). Sequence alignment showed that, despite being quite divergent from them, the overall genetic organization of the new STLV-3/PPA-F3 provirus was similar to those of the other PTLVs, with the existence of the major open reading frames (ORFs) that encode the structural Gag (nucleotides [nt] 743 to 2011) and Envelope (nt 5056 to 6531) polyproteins and the third exon of the regulatory proteins Tax (nt 7232 to 8280) and Rex (nt 7232 to 7717). As described for other strains of PTLV, the protease and polymerase proteins of STLV-3/PPA-F3 are encoded by two overlapping ORFs (nt 1963 to 2497 and nt 2382 to 5063, respectively), via one or two successive −1 ribosomal frameshifts that align the different ORFs. Although the STLV-3/PPA-F3 sequence was related to STLV-3/PHA-PH969 and STLV-3/CTO-604, the overall divergence between the three strains in the major ORFs was equivalent to or higher (8 to 13%) than those between the different subtypes of HTLV-1 (A, B, C, and D) or HTLV-2 (A, B, and D) (7 to 9%) (data not shown). Interestingly, the long terminal repeat (LTR) of STLV-3/PPA-F3 presented a high degree of similarity (85.5 and 88.1%) to those of STLV-3/PHA-PH969 and STLV-3/CTO-604, with a similar length (684 bp versus 695 and 694 bp, respectively). Its overall organization is also identical to those present in the HTLV-bovine leukemia virus genus, with highly conserved regions (position of U3/R and R/U5 boundaries, TATA box, polyadenylation site, and signal and potential splice donor site). Interestingly, the STLV-3/PPA-F3 LTR was found to be shorter than those of HTLV-1 (756 bp) and HTLV-2 (764 bp). This was mainly due to the presence of only two 21-bp repeats (the middle and the proximal ones) in the STLV-3/PPA-F3 LTR. This result has also been previously reported for the STLV-3/PHA-PH969 (31) and STLV-3/CTO-604 strains (17). A comparison of the protein sequences of the different STLV strain is given in Table 3. The 422-amino-acid Gag precursor was cleaved in three core proteins: p19, p24, and p15, in analogy to the HTLV-1 protein nomenclature. The p24 protein sequence is the most conserved among the different STLV-3 strains (range, 96% to 98% similarity) and among the other PTLVs (17, 31). The highest divergence among the PTLV Gag proteins is observed in p19, with a hypermutated region in the carboxy-terminal part of the protein (with a 6-residue deletion in the PTLV-3 protein) that corresponds to the immunodominant epitope in HTLV-1 ATK (residues 102 to 117: PPSSPTHDPPDSDPQI). This is in complete agreement with the Western blot pattern, which shows strong reactivities against the conformational p24 epitope and no reactivity against p19 HTLV-1 protein (Fig. 1). The Env surface protein sequence is the most divergent among the PTLVs, but there is no clear correlation with the immunoblot pattern. Indeed, the HTLV-specific peptides K55 (HTLV-2) and MTA-1 (HTLV-1) present 64 and 65% similarity with STLV-3/PPA-F3, respectively, but only two out of seven P. hamadryas papio sera (PPA-F3 and PPA-F8) exhibit antibodies that recognize K55. Among these two sera, one showed also a faint reactivity against MTA-1 (PPA-F3) (Fig. 1).

TABLE 2.

Nucleotide sequence comparisons of the principal ORFs between the STLV-3 prototype subtype strains and STLV-3/PPA-F3

Comparison	% Similarity
	General	gag	pro	pol	env	rex	tax
STLV-3/PPA-F3 vs STLV-3/CTO-604	89.4	90.2	86.5	89.6	88.2	95.8	92.6
STLV-3/PPA-F3 vs STLV-3/PHA-PH969	87.9	88.3	84.6	87.8	88.6	95.4	91
STLV-3/PHA-PH969 vs STLV-3/CTO-604	87.3	87.3	84	88	85.1	94.9	92.5

TABLE 3.

Protein sequence comparisons between the different STLV-3 prototype subtype strains and STLV-3/PPA-F3

Comparison	% Similarity
	Gag			Pro	Pol	Env		Px
	p19	p24	p15			Surface	Transmembrane	Rex	Tax
STLV-3/PPA-F3 vs STLV-3/CTO-604	95.9	98.1	91.8	90.4	92.6	92.4	97.2	91.7	95.7
STLV-3/PPA-F3 vs STLV-3/PHA-PH969	95.1	96.7	91.8	89.8	92.2	89.8	98.8	87.4	96
STLV-3/PHA-PH969 vs STLV-3/CTO-604	94.3	96.7	92.94	87.6	91.7	88.2	98.3	87.9	95.7

Although previous studies (32) have shown the existence of four putative ORFs in the pX region of STLV-3/PHA-PH969, only three alternatively spliced messengers could be detected by reverse transcriptase PCR on the RNA extracted from the PH969 cell line. These doubly spliced messengers encoded the Tax and Rex proteins and a putative protein called RORFI. This protein is related at the amino acid level (leucine rich) to p12^I and p10^I of HTLV-1 and HTLV-2, respectively. The sequence analyses of the new STLV-3/PPA-F3 Px regions show only two ORFs that correspond to the third Tax and Rex exons. Regarding the possible equivalent of the STLV-3/PHA-PH969 RORFI sequence in our novel strain, we found a shorter ORF (nt 6823 to 6954) that could encode a putative leucine-rich protein with one mutation that replaces a glutamine at position amino acid 65 (CAG) with a stop codon (TAG). However, another mutation in the splice acceptor site replacing AG with AA could probably abolish the splice junction. To get new insights on the genomic organization of the Px region and to search for the presence of singly or doubly spliced viral messengers previously described in the other human T-cell leukemia or simian T-cell lymphotropic retroviruses, we performed a series of reverse transcriptase PCRs on the total RNA extracted from short-term PPA-F3-cultured PBMCs. We did not amplify any cDNAs that could correspond to singly or doubly spliced viral mRNAs, probably due to a low viral messenger expression. However, we found consensus splice acceptor and donor sites that are identical to those described for STLV-3/CTO-604 and STLV-3/PHA-PH969. The potential splice junction positions are the LTR donor site sd-LTR at nt 409, the sd-Env at nt 5059, and the exon 3 Tax/Rex acceptor sa-T/R at nt 7231.

Finally, phylogenetic analyses were performed for both the small tax fragment (180 bp) alignment whose sequences are available for all the published STLV-3 strains and for the complete env polyprotein gene. Our new strains were compared with all the PTLV representative strains of the PTLV-1 and PTLV-2 subtypes. The consensus unrooted tree generated by the neighbor-joining method (1,000 replicates) gives a tree topology with three major PTLV groups supported by high bootstrap values (>98% by neighbor joining) in both the env and tax analyses (Fig. 2A and B). Regarding the tax analyses (Fig. 2A), while the PTLV-1 and PTLV-2 groups comprise both human and simian strains, the PTLV-3 type comprises only simian strains that cluster in three distinct groups or clades supported by significant bootstrap values (>84% by neighbor joining). The first group comprises the strains from Ethiopia and Eritrea (except the strain PHA-7550), the second comprises our novel strains from P. hamadryas papio (PPA-F3, PPA-F4, PPA-F5, and PPA-F9) and the CTO-604/CTO-602 strains from Cameroon, and the third comprises the two C. nictitans strains from Cameroon (CNI-217 and CNI-227). We decided to name provisionally these groups according to the geographical origin of the strains: East African STLV-3 (or STLV-3 A) with STLV-3/PHA-PH969 as the original prototype sequence, West and Central African STLV-3 (or STLV-3 B), and the Central African STLV-3 (STLV-3 C) for the C. nictitans divergent strains. Although the same tree topology was conserved in the env polyprotein analysis (Fig. 2B), the bootstrap value for West and Central African STLV-3 was less significant (57.3), related to the low, equidistant similarity that exists between the STLV-3 strains. Despite this fact, the evolutionary distance observed between the three STLV-3 strains was higher than the distance observed between the PTLV-1 and PTLV-2 subtypes, suggesting a long, independent evolution.

FIG. 2.

(A) Unrooted phylogenetic tree generated by the neighbor-joining method (NJ) on the consensus published STLV-3 tax sequence (nt 7467 to 7646 of STLV-3/PPA-F3). Bootstrap support (1,000 replicates) for the NJ tree is noted on the branches of the tree. The STLV-3/PPA-F3, -F4, -F5, and -F9 sequences were analyzed with HTLV or STLV prototype sequences available from the GenBank database (26). Branch lengths are proportional to the evolutionary distance (scale bar) between the taxa. (B) Unrooted phylogenetic tree generated by the neighbor- joining (NJ) method on the env polyprotein nucleotide sequence (nt 1 to 1467 of the HTLV-1A ATK prototype sequence). Bootstrap support (1,000 replicates) for the NJ tree is noted on the branches of the tree. The STLV-3/PPA-F3 strain was analyzed with HTLV or STLV prototype sequences available from the GenBank database (26). Branch lengths are proportional to the evolutionary distance (scale bar) between the taxa.

Recent data (26) from a molecular clock analysis of PTLV evolution, in which the third codon position was used, led to a PTLV evolutionary rate estimated to be no higher than 1.67 × 10⁻⁶ ± 0.17 × 10⁻⁶ nucleotide substitutions by site and per year. The clock was calibrated by employing the earliest migration from Asia to Melanesia 60,000 years ago as the lower limit for the node separating HTLV-1 Melanesian subtype C from the other HTLV-1 subtypes. A more recent study (17), performed on the third position codon of the Env polyprotein and Env gp21, estimated the evolutionary rate to be around 1.9 × 10⁻⁶ ± 0.1 × 10⁻⁶ nucleotide substitutions per site per year. To better estimate the separation between the STLV-3 strains, we tested the molecular clock hypothesis for PTLV with our new strain on the third codon position of the Env polyprotein using the TN93 model implemented in the Puzzle 4.0.2 program. The molecular clock model was indeed found appropriate, and we calibrated the evolutionary rate to be around 1.67 × 10⁻⁶ ± 0.18 × 10⁻⁶ substitutions per site per year. Using this value, we estimated that the separation between STLV-3/PHA-PH969 and STLV-3/CTO-604 occurred 145,200 ± 13,800 years ago and that the separation from strain STLV-3/PPA-F3 (to STLV-3/PHA-PH969) occurred 109,100 ± 12,500 years ago. These dates were more ancient than the estimated separation dates between divergent PTLV-1 strains (STLV-1 TE4 and HTLV-1 MEL5, around 90,000 ± 8,900 years) or the divergent HTLV-2 subtypes (HTLV-2 EFE2 and HTLV-2 Gab, around 65,680 ± 8,400 years).

In this paper, we report the detection, complete nucleotide sequence, and genomic organization of a novel highly divergent STLV strain that is related to but distinct from the few other known strains of the STLV-3 type (17, 30, 31, 36). The novel STLV-3 strains reported here (STLV-3/PPA-F3, -F4, -F5, and -F9) were present in P. hamadryas papio, large primates whose habitat is mostly restricted to the woodlands, savannas, and Sahelian steppes, within reach of water, of West Africa. The eight studied baboons indeed originated from West Africa (Senegal), where they were all caught in the wild. Since then, they have never been in close contact with other monkeys or apes and have never been inoculated with any human or simian materials. These data suggest strongly that these novel viruses naturally infect P. hamadryas papio in the wild. Furthermore, there are, to our best knowledge, no other seroepidemiological studies on wild-caught P. hamadryas papio and, at the viromolecular level, only one STLV partial env sequence has been obtained from a P. hamadryas papio (11). This is a typical STLV-1 (strain PPA-5X28), and the exact origin of the animal, which was kept in captivity in the United States, was unknown. At the molecular level, HTLV-1 and HTLV-2 promoters are constituted by three 21-bp repeated elements that contain a core element essential for the LTR transcriptional activation (28). The analysis of the STLV-3/PPA-F3 LTR shows a smaller U3 region than those of HTLV-1 or HTLV-2, due mainly to the deletion of the TATA distal 21-bp repeat. These HTLV-1 and -2 promoter elements play an important role in basal transcription in absence of Tax, but mutating one of them unequally reduces the basal HTLV-1 transcription level (1). Therefore, the potential transformant capacity of the different STLV-3 Tax, including the new ones reported here, is now under investigation. Interestingly, it is also worthwhile to note that, while several adult T-cell leukemia- and lymphoma-like diseases have been reported in STLV-1-infected monkeys or apes, there are no reports yet on the diseases possibly associated with STLV-3 infection. Based on our new data and on the few other available serological features of STLV-3-related monkeys, it appears now clearly that STLV-3 infection elicits in its natural hosts the following common pattern that associates an HTLV-2-like serology (strong p24 associated with GD21 and little or no p19 seroreactivity) and a roughly equivalent seroreactivity on both HTLV-1 (MT2)- and HTLV-2 (C19)-producing cells. In addition to STLV-3 infection, such a serological pattern has been reported only in the few P. paniscus specimens infected by STLV-2 strains (5, 34) and in one P. anubis baboon (Bab 503) infected with an African STLV-1 strain (14). Among the 1,100 samples of African and Asian monkeys and apes that we have tested so far in our laboratory, we never observed such a seroreactivity except in the sole P. anubis baboon cited above and in the rare STLV-3-infected animals. It is, however, worthwhile to note that most of the STLV seroepidemiological surveys conducted in monkeys cannot really be considered representative of the situation in the wild, since a large proportion of samples were obtained from captive animals. The biodiversity of such STLV-3 strains in the wild, especially in Africa, is thus far from being known. Multiple episodes of interspecies transmission of PTLV-1 (STLV-1 or HTLV-1), have occurred between different primates, including humans in Central Africa (11, 13, 15, 20, 35). It is thus very tempting to speculate that some other STLV-3 strain or related viruses may exist in other monkeys species but also that HTLV strains related to STLV-3 may exist in human populations living in such areas. The presence of such viruses in baboons, which are known to live often in close contact with humans, reinforces such a possibility. It opens the possibility of a human counterpart of these viruses in African inhabitants exhibiting an HTLV-2-like seroreactivity.

In conclusion, the discovery of this novel STLV-3 strain in West Africa greatly enlarges the geographical distribution of this PTLV type in the African continent, as it is now present in three totally different ecosystems of East (desert), Central (rain forest), and West (savanna) Africa (Fig. 3). Furthermore, the presence of a novel, highly divergent STLV-3 strain in an evolutionarily distant African monkey species (Guinea baboon) also enlarges the host range of species infected by such viruses and reinforces the possible African origin of PTLV.

FIG. 3.

Actual geographical distribution of STLV-3.

Nucleotide sequence accession number.

The complete genome of STLV-3/PPA-F3 has been registered under GenBank accession number AF 517775.