Identification of Significant Genomic Changes and Compartmentalization of Simian Foamy Virus in a Human Zoonotically Infected by a Chimpanzee (Pan troglodytes troglodytes)

TL;DR

This study examines how simian foamy virus adapts in a human infected by a chimpanzee, revealing genomic changes and compartmentalization that may limit transmission and pathogenicity.

Contribution

The study identifies specific genomic changes and compartmentalization patterns in SFV following zoonotic transmission to a human.

Findings

SFV genomes in the human and chimpanzee were nearly identical, but the human had significant deletions in the LTR region.

Genetic stability was observed in bet sequences across body compartments, with evidence of compartmentalization in urine and semen.

Low proviral loads and undetectable SFV in most urine specimens suggest limited viral persistence in the human host.

Abstract

Despite increasing reports of zoonotic simian foamy virus (SFV) infections globally, knowledge of its genetic adaptation in humans and impact on viral transmission and pathogenicity remains limited. We obtained complete SFV genomes using metagenomics analysis of viral isolates from peripheral blood lymphocytes (PBLs) and throat specimens from a worker (Case 6) and source chimpanzee (B1) that bit him. We analyzed viral diversity in three genomic regions (LTR, tas, and bet) involved in replication and latency using longitudinal specimens (PBLs, throat, saliva, urine, and semen) from Case 6 over five years, and PBLs from B1 and five additional chimpanzees over three years. Proviral loads were measured using a validated qPCR assay. Phylogenetic analysis revealed nearly identical SFV genomes in Case 6 and B1. Overall, bet sequences exhibited high genetic stability across body compartments and over time, with evidence of compartmentalization in Case 6 urine and semen specimens. G→A substitutions in GG and GA motifs in bet indicated heterogeneous APOBEC-associated editing across hosts and anatomical compartments following zoonotic transmission. Case 6 had significant deletions in the LTR region that were absent in B1 and other chimpanzees. Length variation in tas, including truncated forms, was observed across longitudinal specimens from Case 6, B1, and other chimpanzees. Proviral loads were consistently low and undetectable in most Case 6 urine specimens. Together, analysis of this SFV transmission pair identifies genomic changes likely to affect viral replication and persistence, highlighting mechanisms that may limit secondary transmission and pathogenicity of SFV in humans.

Figures

Full text

1. Introduction

Simian foamy viruses (SFVs) are complex retroviruses ubiquitous in nonhuman primates (NHPs) that have co-evolved with their hosts for millions of years [1,2,3,4,5]. Despite this long evolutionary history in primates, humans are not naturally infected with a human-specific foamy virus but are susceptible to zoonotic infection [6,7,8,9,10,11,12,13,14,15]. Zoonotic SFV transmission occurs frequently among people exposed to NHPs at work and in natural settings across North and South America, Africa, and Asia, with higher prevalence in those reporting severe bites [6,7,8,9,10,11,12,13,14,15]. SFV infection in humans is characterized by persistent seropositivity, often accompanied by molecular detection or isolation of SFV from peripheral blood lymphocytes (PBLs) [6,7,8,9,10,11,12,13,14,15]. Phylogenetic analyses have identified diverse SFV sources for human infection, including chimpanzees, gorillas, baboons, mandrills, African green monkeys (AGMs), colobus monkeys, guenons, and macaques [6,7,8,9,10,11,12,14]. In contrast, zoonotic SFV infections from New World monkeys in South America and the United States have only been identified serologically, with no information available on the specific SFV strains in these infections [13,15,16].

Although SFV is not known to be pathogenic in NHPs, cross-species transmission can alter the pathogenicity of simian retroviruses, as seen with human immunodeficiency viruses (HIVs) and human T-cell lymphotropic viruses (HTLVs), which emerged from benign simian immunodeficiency virus (SIV) and simian T-lymphotropic virus (STLV) infections, respectively [17,18,19,20]. To date, SFV-infected people have not shown evidence of disease, possibly due to the generally healthy populations studied [6,8,9,12,14,21]. Two studies looked at disease in SFV-infected people [22,23]. One study followed seven men in the United States infected with baboon and chimpanzee SFV for one to five years, reporting mostly nonspecific symptoms and common age-related disease; one participant had mild thrombocytopenia and asymptomatic, nonprogressive monoclonal natural killer cell lymphocytosis of unclear relation to SFV [22]. A case control study of 24 apparently healthy Cameroonian hunters infected with gorilla-type SFV found anemia and hematological abnormalities of uncertain significance [23]. Additional studies are needed to define disease potential in SFV-infected people, as the incidence of disease may be low, follow long latency periods, or be associated with specific SFV variants [8].

Viruses, especially RNA viruses with high mutation and recombination rates, readily adapt to new hosts following cross-species infection, as seen in the SARS-CoV-2 pandemic [24]. Host-specific adaptation of SIV first in chimpanzees and subsequent HIV adaptation in humans enabled increased viral fitness, person-to-person transmission, and pathogenicity [20,25]. However, comprehensive characterization of SFV genomes following zoonotic infection is limited, and it is unclear how viral adaptation affects viral replication, transmission, or pathogenicity. A partial genome from PBLs of a worker infected with SFV from an AGM revealed a deleterious mutation in the bet gene (between envelope [env] and transcriptional transactivator [tas] genes) associated with viral persistence and was subsequently confirmed by complete genome sequencing [26,27]. Deletions in the unique 3-prime (U3) region of the genome flanking long terminal repeats (LTRs) and the tas gene were observed in tissue culture isolates from the first reported human infection with chimpanzee SFV but are believed to be in vitro adaptations [28,29]. We also identified an in-frame 303-bp tas deletion in a PBL tissue culture isolate from a worker infected with baboon SFV [30]. These variants, termed Δtas genomes, are thought to replicate poorly due to the lack of Tas production essential for viral replication, contributing to SFV persistence [31,32]. A recent study of five hunters in Cameroon infected with SFV originating from chimpanzee, guenon, and gorilla found no Δtas, Δbet, or ΔLTR variants in tissue culture isolates, though the Δtas deletion was present in PBLs from one hunter with chimpanzee SFV, and ΔLTR deletions were found in PBLs from two hunters with gorilla SFV [31]. However, these variants were not detected in expanded screening of PBLs from 10 additional chimpanzee and 30 gorilla SFV-infected hunters in Cameroon [31]. In addition, the ΔLTR variants were not detected in nine captive chimpanzees, suggesting the occurrence of these mutations in vivo is rare [29].

In each of these previous investigations, direct genomic comparisons of SFV from source animals and the zoonotically infected human have been limited, restricting insights into viral evolution after transmission. We previously identified a veterinarian (Case 6) infected with SFV from a central African chimpanzee (Pan troglodytes troglodytes) and confirmed the genetic and epidemiological link to the P. t. troglodytes (B1) that bit him [33]. This unique pair enabled evaluation of viral changes post-transmission. We sequenced and analyzed full-length SFV genomes from PBLs of Case 6 (SFVptr_hu6) and the source chimpanzee (SFVptr_B1). Additional SFVptr_hu6 genomes were obtained from PBLs and throat specimens collected nearly two years later to assess genomic stability and compartmentalization. Viral diversity and tissue compartmentalization were evaluated through phylogenetic analysis of divergent bet sequences from DNA specimens from Case 6 across multiple body sites over several years. Longitudinal proviral loads in these compartments were measured to test whether mutations correlated with changes in proviral load, potentially affecting transmission and pathogenicity. Finally, we screened these body fluids for Δtas and ΔLTR variants, previously observed in other human SFV infections, that may influence replication, transmissibility, and pathogenicity.

2. Materials and Methods

2.1. Exposure History for Case 6. Case 6, a Veterinarian, Reported a Severe Chimpanzee Bite in 1977 That Required Surgery [33]

The captive chimpanzee (B1) was a P. t. troglodytes, confirmed by genetic testing [33]. Serum samples archived from B1 since 1978, approximately four months post-injury, and sera archived from Case 6 since 1981, all tested positive for SFV antibodies, indicating a minimal infection duration of 44 years as of 2025 [33]. Phylogenetic analysis of proviral SFV integrase (IN) sequences from Case 6, B1, and other chimpanzees that Case 6 worked with confirmed B1 as the source of his SFV infection [33]. The wife of Case 6 was SFV-negative despite decades of intimate contact [22].

2.2. Human and Chimpanzee Specimen Collection and Processing

EDTA-treated whole blood and body fluid specimens (parotid saliva, throat swab, saliva, urine, and semen) were collected from Case 6 during voluntary participation in a long-term follow-up study. Parotid saliva was collected in intraoral cups and immediately transferred to cryovials. Throat and saliva swabs were collected using viral culturettes (Becton/Dickinson). EDTA-treated whole blood (and saliva from one chimpanzee) was collected from B1 and six other captive chimpanzees during annual physical examinations in accordance with the guidelines of animal care and use committees at each institution. Non-blood specimens were shipped on wet ice to CDC immediately after collection; whole blood was shipped at room temperature. Case 6 samples used for sequence generation were collected in 1998 and 2000, and the B1 sample was collected in 1998.

At CDC, PBLs were isolated by Ficoll-Hypaque centrifugation of the whole blood, and DNA lysates were prepared as described previously [33]. Parotid saliva was centrifuged for 2 min at 1000× g, and the cell pellets and supernatant were aliquoted and frozen at −80 °C until tested. Throat and saliva swabs were placed in 2 mL phosphate-buffered saline (PBS), vortexed, and centrifuged for 5 min at 1000× g to pellet any cells. The cell pellet was washed twice with PBS and divided equally for PCR testing and tissue culture. Urine and semen samples were centrifuged for 10 min at 800× g to pellet any cells, washed twice with PBS, aliquoted, and stored at −80 °C.

2.3. Tissue Culture

Viral isolation was performed by co-cultivating equal numbers of PBLs and canine thymocytes (Cf2Th) on the day the PBLs were isolated from whole blood, as reported previously [33]. Throat and saliva specimens were also co-cultivated with Cf2Th cells the day they were prepared. Cultures were monitored every three to four days for syncytial cytopathic effect (CPE) typical of FV. When at least 50% of the cultures showed CPE, the cells were trypsinized, and DNA lysates were prepared and screened for proviral SFV IN sequences by PCR to confirm infection. Supernatants from positive tissue cultures were aliquoted and stored at −80 °C.

2.4. PCR Amplification of Subgenomic Regions

We first tested all specimens for β-actin sequences by PCR to confirm the integrity of the extracted nucleic acids as previously described [7]. Using 0.5–1.0 µg of DNA from body and tissue culture specimens, we amplified SFV IN sequences with generic nested primers and conditions previously described to confirm tissue culture infection [5,33]. To evaluate SFV intra-host compartmentalization and viral diversity, we used nested PCR to generate bet sequences. The first-round bet primers CPZO2F1 (5′-CCA AAT TTT AAC TTG CTG (T/C)CA GGC-3′) and CPZO2R1 (5′-CTC TCT GAG GTC CAT AAG CTT CCA-3′) produced a 470-bp fragment, while the second-round primers CPZO2F2 (5′-CCT GAT GTT TGG TGT ACC CC(T/C) TCT-3′) and CPZO2R2 (5′-TGT GTG GAA CTG CAG AGC TTT GAA-3′) generated a 288-bp sequence. PCR conditions were 94 °C for 5 min, followed by 35 cycles of 94 °C for 1 min, 50 °C for 1 min, and 72 °C for 1 min, with a final extension of 72 °C for 7 min, and storage at 4 °C.

To check for ΔLTR variants, we used first-round primers 11662F (5′-GCT GGA AAG TAG TTA CTG AGG CA-3′) and 12937R (5′-GGA GAG TCT CAT ACG CTC TCG ACG-3′) and nested primers 11772F (5′-TGT GAC CCT TTC ATT GAC TCA GGA-3′) and 12793R (5′-AAA AGC TGC CTG CGT TAA GGA-3′) to generate 1275-bp and 1021-bp fragments, respectively, with final sizes depending on the absence or presence of deletions. PCR conditions included 94 °C for 5 min, followed by 4 cycles of 94 °C for 1 min, 37 °C for 1 min, and 72 °C for 1 min, then 36 cycles of 94 °C for 1 min, 50 °C for 1 min, and 72 °C for 1 min, with a final extension of 72 °C for 7 min, and storage at 4 °C.

For Δtas variants, we used nested primers TAScpzFout (5′-GGA TTC CTA CCA AGA AGA AG-3′) and TAScpzRout (5′-CTG AAT GTT CAC CTG ACC C-3′) along with TAScpzFin (5′-CTC TTA GTG AGC TTG TTG G-3′) and TAScpzRin (5′-CAC CTG ACA TGT GAA CTT CC-3′), generating 892-bp and 765-bp fragments, respectively, with final sizes depending on the absence or presence of deletions and PCR primers used. These PCR assays are described in detail elsewhere [31]. The nested tas primers are upstream of the nested bet primers. The sizes and quality of PCR amplicons were verified by gel electrophoresis in 1.0% agarose gels and visualized with ethidium bromide staining.

2.5. Cloning of bet Sequences for Quasispecies Analysis

We cloned PCR-amplified bet sequences into the pCR2.1 TOPO vector according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). Positive clones were identified by colony PCR with the inner bet PCR primers CPZO2F2 and CPZO2R2 and visualized by ethidium bromide staining and agarose gel electrophoresis. We Sanger-sequenced between five and 12 positive clones for each sample examined, as next-generation sequencing (NGS) was not available at that time.

2.6. Proviral Load Measurement

Limited information exists on viral loads in SFV-infected humans and NHPs [11], which are associated with increased pathogenicity and transmissibility of other human retroviruses. To investigate proviral loads in Case 6 and SFV-infected chimpanzees, we developed a generic quantitative PCR test (qPCR) targeting pol sequences using the new complete SFV genomes obtained in our study. The qPCR assay used the pol primers SIF4O (5′CTM ACT AGT ATT GCW RTT CCR AAG GT3′) and SIR1N (5′GTT TTA TIT CAC TAT TTT TCC TTT CCA C3′), along with double-quenched probes SIP4O (FAM5′ CAC TCY GAT <ZEN> YAR FFK RCD GCA TTC AC 3′IBFQ) and SIP5RON (FAM5′ CTT TGR GGR <ZEN> TGR TAA GGA GTA CTG WAT TCC 3′IBFQ). The inclusion of ZEN and IBFQ quenchers in the probes improve signal-to-noise ratios and assay sensitivity. This qPCR assay was validated to detect 1–5 copies of chimpanzee SFV per 100,000 cells, equivalent to about 1 µg DNA [7]. We used 200 ng DNA and the AmpliTaq Gold system (Applied Biosystems/Roche, Branchburg, NJ, USA) on a Bio-Rad CFX96 iCycler (Hercules, CA, USA) with the following conditions: 95 °C for 10 min, followed by 55 cycles of 95 °C and 52 °C for 15 s each, and 62 °C for 30 s.

SFV proviral load was normalized by DNA input measured with ribonuclease P subunit 30 (RPP30) qPCR using the primers RPP30FM 5′-GCA GAT TTG GAC CTG CGA GCG-3′ and RPP30RM 5′-GTG AGC GGC TGT CTC CAC AAG-3′, along with the probe RPP30PM 5′-HEX TTC TGA CCT GAA GGC “T”CT GCG CGG-SpC6-3′, where “T” is BHQ1 modified. The reaction contained 900 nM of primers, 250 nM of probe, 1× AmpliTaq Gold buffer, 400 nM dNTPs, 3 mM MgCl2, and 2.5 U of AmpliTaq Gold enzyme in a final volume of 20 ul. Cycling included 95 °C for a hot start, followed by 55 cycles of 95 °C for 15 s and 62 °C for 30 s on a Bio-Rad CFX96 iCycler. We used 40 ng DNA as input for the RPP30 qPCR assay.

2.7. Sanger Sequencing of PCR Products from Subgenomic Regions

PCR products were purified using the Qiaquick gel extraction kit (Qiagen Inc., Valencia, CA, USA) and sequenced in both directions with a Big Dye terminator cycle kit (ThermoFisher Scientific, Waltham, MA, USA) using nested PCR primers. Additional internal sequencing primers for the LTR region included: 11795F (5′-AGT GAC TCA GAT GGA CCC CTG TA-3′), 12065F (5′-GAT GAT GCG TTA GTT CGA ATC CCA-3′), 12025R (5′-TGC TGT TCC AGC TTA CAG AGC CA-3′), and 12744R (5′-GGG CAA CTT TAG AGG TGT TGT GTC-3′).

2.8. Next-Generation Sequencing and Genome Assembly

To obtain complete genomes from viral isolates, we employed a metagenomics approach as described elsewhere [30]. Briefly, we centrifuged 0.5–1.0 mL of tissue culture supernatant at 43,000 rpm at 4 °C for 30 min and re-suspended the viral pellet in 165 μL of supernatant. The sample was then treated with a cocktail of DNase enzymes (Turbo DNase, Ambion, Austin, TX, USA; Baseline-ZERO DNase, Biosearch Technologies, Middlesex, UK). Viral nucleic acids were isolated using the Qiagen QIAamp MinElute virus spin kit (Qiagen, Hilden, Germany), omitting carrier RNA. The eluted RNA was converted to double-stranded cDNA using random hexamers with the Superscript double-stranded cDNA Synthesis kit (Invitrogen, Carlsbad, CA, USA). cDNA was purified with the Agencourt Ampure XP system (Beckman Coulter, Brea, CA, USA), and approximately 1 ng of cDNA underwent simultaneous fragmentation and tagmentation using the Illumina DNA Prep Kit (Illumina Systems, San Diego, CA, USA). The resulting DNA library was sequenced on an Illumina MiSeq (MiSeq Reagent Kit V2, 500 cycle, Illumina Systems, San Diego, CA, USA) and analyzed in a 2 × 150-bp mode.

Raw data were de-hosted by removing reads aligned to the canine reference genome (Illumina iGenomes, Canis familiaris NCBI build 3.1) using Bowtie2 v2.5.1 [34]. Remaining reads were processed using the nf-core/viralrecon pipeline (v2.6.0, nf-core/viralrecon: nf-core/viralrecon v2.6.0—Rhodium Raccoon) in de novo assembly mode. Quality control (FastQC) v 0.11.9 and genome assembly using SPAdes v 3.15.5 [35], Unicycler v 0.4.8 [36], and Minia v 3.2.5 [37] were conducted, and an assembly report was created using QUAST v 5.2.0 [38]. We used Blastn v 2.13.0 to identify sequences closest to the assembled SFV contigs, evaluating them based on the BLAST v 2.42 output, coverage, and length [39]. Contigs of appropriate size representing SFV genomes were selected for further analysis.

2.9. Sequence Analysis of Complete SFV Genomes

We identified protein-coding reading frames in the SFVptr_hu6 and SFVptr_B1 genomes using the website insilico.ehu.es/translate/, accessed March-June 2024. The positions of the complete 5′ and 3′ LTRs were determined manually with reference to previously published chimpanzee SFV genomes. Potential splice donor and acceptor positions were inferred using neural network predictions implemented in NetGene2 (https://services.healthtech.dtu.dk/services/NetGene2-2.42/), accessed March 2024. We identified potential nuclear localization signals in the Tas protein using NucPred (https://nucpred.bioinfo.se/cgi-bin/single.cgi), accessed March 2024 and PSORTII (https://psort.hgc.jp/form2.html), accessed March 2024.

Using Geneious v2025.0.3, we extracted the five protein-coding regions (group-specific antigen [gag], polymerase [pol], env, tas, and bet) and aligned them with representative ape SFVs with complete genomes for comparison. We also created a concatemer—of joined sequences from three gene regions of the major FV coding regions (gag, pol, and env)—for robust phylogenetic analysis.

Codon-based nucleotide alignments of the concatemers and the bet gene sequences were performed using MAFFT v7.0.26 [40], followed by manual adjustments and gap stripping. We determined the best-fitting nucleotide substitution model using MEGA v7.0.26, which was inferred to be the general time-reversible (GTR) model with gamma (G) distribution and invariable sites (GTR + G + I). Likelihood mapping of quartet topologies was conducted in IQ-TREE v1.6.12 to assess the phylogenetic signal in the alignment [41]. We also checked the phylogenetic signal and evidence of nucleotide substitution saturation in the alignments using DAMBE v7.0.35 (http://dambe.bio.uottawa.ca/DAMBE/dambe.aspx) and identified saturation at the 3rd codon position in the alignment, thus only using the 1st and 2nd codon positions for the concatemer phylogenies [30].

We inferred phylogenies for the concatemer sequences using Bayesian inference with BEAST v.10.5.0 [30]. We included SFV from 18 other simians and three non-simians (feline, bovine, equine) using 100 million Markov Chain Monte Carlo (MCMC) iterations with a 10% burn-in. We ensured parameter convergence with effective sampling size (ESS) values > 600 using Tracer v1.7.2. Trees were logged every 10,000 generations, and two independent BEAST runs were performed to ensure convergence and reliability of results. We used TreeAnnotator v10.5.0 to select the maximum clade credibility tree from the posterior distribution of 10,001 sampled trees, applying a burn-in of 1000 trees.

We conducted variant calling analysis of the complete genomes using the Saref v 3.4.4 Nextflow pipeline (https://doi.org/10.5281/zenodo.3901628) with the bioinformatic tools DeepVariant ≤, Haplotypecaller (GATK) [42], BCFtools [43], and Strelka2 [44]. The results from each tool were evaluated individually for evidence of genomic variation. Highlighter plots (https://www.hiv.lanl.gov/content/sequence/HIGHLIGHT/highlighter_top.html), accessed June–August 2025 were used to visualize comparisons of the human and chimpanzee SFV bet sequences. Potential apolipoprotein B mRNA editing enzyme catalytic polypeptide-like (APOBEC) signatures in the bet sequences were inferred using the B1 PBL sequence from 9-18-98, and to define the local dinucleotide context at each position. APOBEC-relevant contexts were defined by the reference dinucleotide starting at the focal G: GG (consistent with APOBEC3G-like targeting) and GA (consistent with TC-family APOBEC targeting). Least-squares regression analysis was conducted to compare enrichment of GG versus GA dinucleotides.

We performed sequence analysis of subgenomic bet, tas, and LTR regions. The bet gene phylogenies were built using BEAST after removing duplicate sequences from each body site using the ELIMDUPES tool (https://www.hiv.lanl.gov/content/sequence/elimdupesv2/elimdupes.html), accessed May 2025. We defined clusters in the tree using posterior probability values ≥0.7, with gorilla SFV (SFVggo) sequences available in GenBank as the outgroup. The bet trees were visualized using FigTree v1.4.4 (https://tree.bio.ed.ac.uk/software/figtree/). We also analyzed clustering of SFV bet sequences from different body compartments of Case 6, PBLs, and saliva of chimpanzees, and PBLs of gorilla (SFVggo) as a control group using MicrobeTrace and a Tamura Nei 1993 (TN93) genetic distance threshold of 1.5 percent [45].

We analyzed the tas and LTR sequences by aligning with MAFFT and performing manual editing in MEGA. Graphical representations of the alignments were prepared using BioEdit (BioEdit Sequence Alignment Editor for Windows 95/98/NT/XP/Vista/78/10) and finalized in MS Word to highlight deletions and important motif locations.

2.10. GenBank Accession Numbers

The SFVptr_hu6 and SFVptr_B1 PBL genomes from 1998 are assigned the accession numbers PV567535 and PV567536, respectively. The SFVptr_hu6 PBL and throat genomes from 2000 have accession numbers PV976853 and PV976854, respectively. The 240-nt bet, LTR, and tas sequences have the accession numbers PV606080-PV606364, PX234069-PX234103, and PX314827-PX314863, respectively.

2.11. Ethics Approval

Case 6 consented to participate in the initial “Voluntary Seroprevalence Study of Retrovirus Infections Among Persons Exposed to Nonhuman Primates” (CDC IRB protocol #1200) and the subsequent voluntary “Long Term Follow up of Persons Infected with Unusual Retroviruses” studies (CDC IRB protocol #1678), both approved by institutional review boards at CDC and participating institutions. Whole blood specimens and saliva were collected from chimpanzees during annual physical exams, following animal care and use committee guidelines at the respective institutions.

3. Results

3.1. SFVptr_hu6 and SFVptr_B1 Genome Assembly

Complete SFV genomes were assembled using a metagenomics NGS approach from RNA extracted from Case 6 PBL tissue culture supernatants collected on 16 September 1998 and 29 November 2000 and a throat specimen collected on 29 November 2000. For comparison, we also generated the complete SFV genome from a B1 tissue culture isolate from a PBL specimen collected on 12 August 1998. All tissue cultures were PCR-positive for IN sequences by day seven, with cytopathic effect observed from day 17. NGS of SFV-positive samples generated 4.3–7.7 million paired-end reads per sample, yielding assemblies with about 10,000× coverage, and the mean mapping quality averaged 48.8. The Case 6 PBL sample had a lower coverage of 226× due to host-read predominance and filtering-induced reduction in fragment size. All assemblies achieved greater than 100× coverage with complete genome breadth, indicating extremely high-quality results. Mapping raw reads back to the assemblies confirmed mean depths of roughly 7300×–39,500×, again with the Case 6 PBL sample having the lowest mean depth coverage but well above quality thresholds.

3.2. Case 6 and B1 Have Nearly Identical SFV Genomes

Both SFVptr_hu6 and SFVptr_B1 genomes contain all expected structural (gag and env), enzymatic (pol), and regulatory (tas and bet) coding regions, flanked by LTRs (Figure S1). Gene and genome lengths were comparable to other chimpanzee SFVs, except for the shorter ΔLTR variants of SFVpsc_huHSRV (Table 1). All SFVptr genes and LTRs were intact in the three Case 6 and single B1 genomes, except for bet, which had a premature stop codon, truncating Bet by nine amino acids (Figure S1), consistent with previous reports [31].

SFVptr_hu6 was nearly identical to SFVptr_B1 (99.94% nucleotide [nt] and 99.85% amino acid identity), both with a genome length of 13,228 nt (Table 2). Phylogenetic analysis showed robust clustering of FV gag-pol-env concatemers by species, supporting the co-evolutionary hypothesis of FV. SFVptr_hu6 and SFVptr_B1 gag-pol-env concatemers clustered strongly with each other and other chimpanzee SFV in the Old-World ape SFV clade (Figure 1). In addition, SFVptr_B1 was ancestral to the SFVptr_hu6 genomes, reinforcing B1 as the origin of the Case 6 infections.

Eight point mutations distinguished the SFVptr_hu6 and SFVptr_B1 PBL genomes from 1998, including T > C mutations at positions 1147 and 12,616 in the U3 region of the LTR, T > A and T > G mutations in pol at positions 5505 and 6053, G > A mutations in env and tas at positions 9827 and 10,445, and an A > G mutation in the tas/bet coding region at position 10,849 (Figure S1). In contrast, only 1–2 single nucleotide polymorphisms (SNPs) were found between the SFVptr_hu6 PBL genomes from 1998 and 2000 and the SFVptr_hu6 PBL and throat genomes from 2000, indicating high genomic stability over two years and across different body compartments. Variant analysis of NGS reads from SFVptr_hu6 and SFVptr_B1 PBL tissue culture isolates showed no significant post-transmission evolutionary changes in coding regions or transcription sites essential for viral replication.

3.3. Viral Diversity and Compartmentalization of SFVptr_hu6 in Urine and Semen Specimens While Maintaining Stable Proviral Loads

Specimens from multiple body compartments were available from Case 6 (1998–2003), and PBLs were collected from B1 in 1998, 2001, and 2002 (Table 3). Case 6 throat and saliva swabs from 2002 were combined and referred to as an oral cavity specimen. For comparison, we included PBL samples from three additional chimpanzee species: P. t. troglodytes (1040; collected in 1999 and 2002), P. t. schweinfurthii (1016; collected in 1999), and three P. t. verus (A101, A136, A182; collected in 1998 and 2001). A 2001 saliva specimen from A101 was also tested. PCR testing confirmed all Case 6 specimens were positive for IN and β-actin sequences, except three urine samples (1/12/00, 5/15/02, and 3/13/03) and throat and parotid saliva specimens (3/13/03). Among these five IN-negative specimens, two urine samples (1/12/00 and 3/13/03) and the throat sample were also β-actin PCR-negative, indicating poor DNA integrity. Overall, our results demonstrate the wide distribution of SFV in the body of an infected human. IN and β-actin sequences were detected in all chimpanzee specimens.

Proviral loads in Case 6 tissue specimens ranged from 1.34 × 10^2^ to 1.11 × 10^4^ copies/µg DNA over five years, with the lowest levels in PBLs and urine specimens from 2002 and the highest levels in the oral cavity in 2002 (Figure 2, Table 3). No significant differences in proviral loads were observed among Case 6 specimens. Similarly, B1 proviral loads ranged from 1.5 × 10^3^ to 5.83 × 10^4^ over three years, with comparable levels in the other chimpanzee specimens. These results indicate stable proviral loads over time in Case 6 and chimpanzees, with no significant differences between them. SFV RNA levels could not be assessed due to insufficient material.

Next, we investigated viral diversity by analyzing 284 clones containing bet sequences (240-nt) from different body compartments of Case 6, and from PBLs of different chimp species and saliva from one chimp (Table 3). Clonal sequence analysis revealed little genetic variation within each compartment of Case 6, with a mean nucleotide identity of at least 99.5% for most sites. However, PBL specimens from 2000 and 2003 and the oral cavity specimen from 2002 exhibited lower mean nucleotide identities, ranging from 97.0 to 98.0%. Mean nucleotide identities of PBL specimens from B1 and other chimp subspecies were similar, ranging from 97.2 to 99.7%. In contrast, the mean nucleotide identity of sequences from the P. t. verus A101 saliva sample was lower at 96.1% due to one highly divergent clone with 80% nucleotide identity to other clones in this sample.

To further investigate the evolutionary relationships of the bet sequences, we conducted three additional analyses. Using MicrobeTrace with a 1.5% TN93 genetic distance threshold, we identified eight clusters among all 294 bet sequences, comprising 12,177 links, indicating very close genetic relationships within each cluster (Figure 3). Cluster sizes ranged from three to 154 sequences. Two clusters showed compartmentalization: one contained all eight urine-derived bet sequences from Case 6 (2001), and the other included 36 members comprising all seven Case 6 semen bet sequences, three Case 6 oral cavity sequences, and one PBL sequence from Case 6 in 2001 and 2003, suggesting likely seeding of the semen compartment from these other body sites. The remaining 24 bet sequences in this cluster were from B1 PBLs (1998, n = 13; 2001, n = 11). Our findings confirm a molecular link between Case 6 and B1 and provide evidence of compartmentalization in urine and, to a lesser extent, semen.

Most bet sequences formed one large cluster of 154 nodes consisting of Case 6 PBL, throat, oral cavity, and PBL and throat tissue culture sequences from 1998 to 2003. One B1 PBL sequence from 1998 and 2001 and eight B1 PBL tissue culture sequences from 1998 were also members of this large cluster. All 20 SFV PBL sequences from P. t. troglodytes 1040 collected in 1999 and 2002 were members of the large cluster, indicating the close relationships among P. t. troglodytes SFV strains. In contrast, bet sequences from the P. t. verus (A101, A136, and A182) and P. t. schweinfurthii (1016) formed three distinct clusters. All eight sequences from P. t. schweinfurthii 1016 clustered tightly. The 22 A182 and 25 A101 (seven 1998 PBL, nine 2001 PBL, and nine 2001 saliva) sequences formed a single cluster, linked by three A182 PBL sequences. One A101 saliva sequence clustered with the largest P. t. troglodytes cluster, suggesting possible infection with a minority P. t. troglodytes variant. Most A136 PBL sequences collected (1998, n = 6; 2001, n = 19) clustered with four A101 PBL sequences, supporting a shared ancestry among A101, A136, and A182, consistent with their shared subspecies and institutional origin. These results further document chimpanzee subspecies specificity of the bet sequences, as previously shown for IN and gag sequences [33].

Seven sequences were singletons: one from Case 6 (PBL, 2003), one B1 PBL (1998), one A136 PBL from each year (1998, 2001), and three A101 sequences from 2001 (two PBL and one saliva). Lastly, SFVggo bet sequences formed two small clusters, one with three sequences and another with seven.

Next, we used Bayesian phylogenetics to confirm relationships identified in the network cluster analysis. After removing duplicate sequences, 102 bet sequences were analyzed to reduce the computational complexity. We found that the results mirrored the MicrobeTrace findings (Figure 4), showing clustering by chimpanzee subspecies and gorilla hosts with strong posterior probability support. Key confirmations included the Case 6 urine sequence clade, clustering of the Case 6 semen sequence with B1 PBL sequences, and the large clade containing most other Case 6 sequences from various body sites, and the P. t. troglodytes B1 and 1040 PBL sequences.

We then utilized highlighter plots to display the nucleotide substitutions across an alignment of 56 bet sequences, excluding duplicates and eight SFVggo sequences (Figure S2). These plots also identified APOBEC (apolipoprotein B mRNA-editing enzyme, catalytic polypeptide-like)-induced mutations, which were more frequent in the urine sample from 2020 (nine sites) compared to P. t. verus specimens (four to seven sites), but mostly at different bet positions. Other Case 6 and chimpanzee specimens showed fewer than four APOBEC sites (Figure S2). The urine clones also had a G > A mutation not seen in any of the other specimens from Case 6 and B1 (Figure S2). Using the B1 PBL bet sequence from September 1998 as the reference, GG and GA dinucleotide counts were comparable across all sequences, indicating that observed signatures were not attributable to underlying sequence composition. Linear regression analysis revealed a positive association (R^2^ = 0.71) between GG- and GA-context presence (slope = 0.90, p-value < ×10^−15^), indicating correlated but non-identical APOBEC editing patterns across SFV sequences. When stratified by host, GG-context differed between sequences derived from B1 and those obtained from Case 6, consistent with host-dependent variation in APOBEC-associated editing. Within the human recipient, GG-context also varied across anatomical compartments, including blood, oral cavity, saliva, throat, semen, and urine, indicating that APOBEC editing intensity is not uniform across specimen sites.

Overall, our detailed analysis of bet sequences demonstrates SFV nucleotide sequence stability over many years and strong molecular linkage between Case 6 and B1, evidenced by clustering of their PBL SFV sequences and clustering of the Case 6 semen and oral cavity specimens with the B1 PBL SFV sequences from multiple timepoints. These findings further support B1 as the source of the Case 6 SFV infection and provide evidence of compartmentalization of SFV sequences in the urine and semen of Case 6.

3.4. Analysis of In Vivo Specimens Shows ΔLTR and Δtas in Case 6 but Not B1

To further investigate the genetic differences between the SFV genomes of Case 6 and B1, we checked for ΔLTR U3 and Δtas variants previously reported in other SFV infections. Starting with 9/16/1998, we identified ΔLTR U3 sequences with one or two deletions (282–479 nt) in Case 6 PBLs, alongside wild-type sequences (977 nt; Table 3, Figure 5), suggestive of a mixed viral population. Conversely, these ΔLTR variants were absent in other Case 6 body sites and tissue culture isolates, all B1 PBL specimens and tissue culture isolates, and all other chimpanzee specimens. These U3 deletions eliminated three conserved ETS proto-oncogene 1 (ETS-1) transcription factor sites and a TATATA motif (Figure S3). Interestingly, the longer 5′ U3 deletions were near a splice donor site (5′-AATAGGTTAA^GTAAGTTAC-3′) at nucleotide positions 497–516 (Figure S3). One ΔLTR variant with a 479-nt deletion from the Case 6 PBLs (9/16/98) contained a unique 12-nt insertion (5′-CATGAGGAGTTT-3″) not found in the Case 6 SFV genome or other GenBank SFV sequences.

Analysis of tas revealed length variation consistent with truncated forms across multiple longitudinal specimens from Case 6, the source chimpanzee (B1), and other chimpanzees, with both full-length and shorter tas variants detected primarily in peripheral blood and selected tissue culture samples (Table 3, Figure 5). The 301-nt Δtas variant was found in three Case 6 body sites: semen (1/12/20), PBLs (5/15/02 and 3/11/03), and oral cavity (5/15/02). It was absent in the Case 6 PBL specimen from 11/29/00 but present in culture from that date, although the 11/29/00 Case 6 PBL sequence had an internal stop codon at position 275. Conversely, all B1 PBL specimens (1998, 2001) had wild-type tas, but the B1 PBL tissue culture from 12/8/98 showed both intact and Δtas forms (Table 3). Two P. t. verus PBLs (A101 and A182) had the Δtas variant, which was absent in a third P. t. verus and in PBLs from P. t. troglodytes (1016) and P. t. schweinfurthii (1040). The 301-nt tas deletion precisely matched the bet gene splice donor and acceptor sites (Figure S4), removing the Bet intron and the tas gene but preserving bet coding. A single APOBEC site was found in three Case 6 tas sequences (PBLs from 11/29/20 and 5/15/02 and semen from 1/12/20) but not in B1 PBL sequences.

In summary, we detected ΔU3 LTR variants in multiple Case 6 PBL specimens but not directly in B1 PBLs, nor in other body sites in Case 6. These results suggest de novo genetic changes in Case 6 following transmission and compartmentalization of the ΔU3 LTR in PBLs of Case 6. We found Δtas variants in Case 6 across various timepoints and body specimens, while B1 PBLs remained wild-type. Most chimpanzee PBL samples lacked the Δtas variant. These findings suggest that tissue culture may induce the deleted tas forms. Following zoonotic transmission to humans, SFV replication may be downregulated due to the loss of transcription sites in the U3 LTR region and disruption of tas-mediated activation of SFV promoters in the PBLs, potentially contributing to viral latency.

4. Discussion

Detailed molecular characterization of zoonotic infections is essential for understanding adaptation, transmission, and disease potential in new human hosts [46]. We present the first genetic analysis of complete and partial SFV genomes from an infected worker (Case 6) and the source chimpanzee (B1) of this infection. Metagenomic sequencing of tissue culture isolates showed nearly identical SFV genomes in Case 6 and B1. Phylogenetic analysis of partial and complete genomes confirmed the molecular link between Case 6 and B1. However, PCR and sequencing of longitudinal specimens from Case 6 showed significant deletions in the U3 LTR region, which were only present in the B1 PBL tissue culture isolate. We also found Δtas variants in PBLs from both Case 6 and B1 and in semen and the oral cavity of Case 6. Analysis of bet sequences identified unique variants in Case 6 urine and possibly semen, suggesting tissue compartmentalization. Overall, our findings indicate substantial proviral changes in SFV from Case 6 after transmission from B1, likely affecting viral replication in certain compartments and promoting latency through adaptation in the new host. These genetic changes may contribute to the apparent lack of person-to-person transmission and diseases associated with zoonotic SFV infections despite persistent infection.

Previous studies investigating the genetic consequences of zoonotic transmissions of SFV to humans have often lacked specimens from the source animal [26,28,29,30,31,47]. Our study is unique in including longitudinal specimens collected from both the source chimpanzee and the infected human. We show that the intact SFV genomes from Case 6 and B1 were nearly identical, sharing high nucleotide identity in infectious virus isolated from tissue cultures of PBLs and a throat specimen from Case 6 specimens collected three months to two years apart. Despite Case 6 being bitten in 1977, with a seropositive specimen in 1981, this high genetic stability likely persisted over two decades for both individuals. Similarly, a prior study found minimal genetic change in SFV from a hunter and the mandrill that bit him ten years earlier [14]. These findings indicate that SFV can remain genetically stable for at least a decade following zoonotic infection. Indeed, phylogenetic analyses of natural infections further show that SFV is highly stable over millions of years, maintaining species-specific lineages [1,5,30].

Targeted PCR and sequence analysis of the LTR and tas regions from multiple body specimens, known to have deletions that impact viral replication [26,27,29,31,47,48], revealed nearly identical deletions in vivo for Case 6 but not always for B1. In Case 6, the ΔLTR variants appeared before Δtas variants, with Δtas first detected in semen and later in throat culture and oral cavity specimens, while ΔLTR variants were restricted to PBLs, which suggests compartmentalization and reduced replication fitness in PBLs compared to other tissues. The absence of ΔLTR variants in B1 suggests these changes in Case 6 are likely arising de novo rather than being transmitted. We detected both wild type and two ΔLTR variants in Case 6 PBLs in 1998, but only the wild-type variant was present in the corresponding tissue culture, supporting reports that in vitro PBL culture favors transition to latency via mutations [29,48,49]. In contrast, in B1, both regions were intact except for the Δtas deletion in the PBL tissue culture, likely induced during culturing. Previous studies found the 301-bp Δtas deletion in only one of twelve (8.3%) SFVptr-infected hunters, and the ΔLTR deletion was absent in all, with neither mutation observed in five P. t. trogolodytes tested [31]. The presence of both full-length and truncated tas variants across Case 6, B1, and other chimpanzees indicates that tas length variation is not unique to zoonotic infection but instead represents a recurring feature of SFV genomes, potentially reflecting constraints on viral replication or transcription rather than host-specific adaptation. The presence of the premature stop in the bet gene, with subsequent truncation of the Bet protein, could lead to substantially reduced replication efficiency. Our findings for Case 6 demonstrate the presence of in vivo SFV deletion mutants following zoonotic transmission, detected in PBLs as well as in additional sampled compartments, including semen and oral specimens. In contrast, analysis of the source chimpanzee B1 was limited to longitudinal PBL specimens, in which these deletions were largely absent. Because direct host comparisons are therefore restricted to PBLs, the observed differences in this compartment suggest post-zoonotic viral evolution, potentially reflecting a transmission bottleneck or host-specific selective pressures resulting in more restrictive viral replication in the human host.

Proviral loads, which can increase with viral replication, remained low and stable in both Case 6 and B1 over several years, including in semen and oral samples from Case 6. Combined with the LTR and tas deletions and bet truncation, these results may help explain the lack of transmission from Case 6 to his spouse despite decades of unprotected sexual contact [22]. Our study provides the first measurements of proviral loads in multiple body sites of humans infected with chimpanzee SFV. Previous studies, which mainly tested PBLs from gorilla SFV-infected humans, reported lower proviral loads by about two log10 copies/µg DNA compared to our findings [21,50]. Differences in proviral loads between studies may be due to distinct SFV variants, variations in qPCR assays targeting different IN gene regions, and the use of SYBR green quantification versus fluorogenic probes used in our assay [21,31]. Fluorogenic qPCR assays are generally more sensitive than SYBR green-based methods [51]. Our assay detected 1–5 copies of SFVptr in 100,000 cells, compared to one copy per 50,000 cells in the earlier study [50]. Additionally, our qPCR probes included dual quenchers, enhancing assay sensitivity and specificity [52].

Our assessment of viral diversity within Case 6 and naturally infected chimpanzees revealed intra-host SFV diversity was very low over the 3–5-year period examined, with low genetic drift in bet sequences in vivo. This finding is consistent with prior longitudinal studies reporting similarly low SFV genetic drift in African green monkeys and mandrills over a decade-long period, based on clonal analysis of env or IN sequences, respectively [14,53]. Unlike those studies, we also quantified proviral loads and found them to be low in longitudinal specimens, consistent with the observed low viral genetic diversity and suggesting low viral replication is likely associated with minimal genetic drift. Despite this overall stability, we observed SFV bet compartmentalization in Case 6’s urine and semen, suggesting localized selective pressures and possible seeding of the semen compartment from the oral cavity and peripheral blood. Notably, we identified multiple APOBEC3 sites within the Case 6 urine and P. t. verus PBL bet sequences. APOBEC3 proteins inhibit replication of human SFV and other retroviruses by inducing G > A hypermutations throughout the viral genome, an effect generally less pronounced in simian hosts [54,55,56]. Our finding of undetectable proviral loads in this urine sample supports APOBEC3-mediated inhibition of SFV replication and infectivity in humans, while proviral loads in P. t. verus PBLs were unaffected. Furthermore, our analysis showed G→A substitutions in both GG and GA dinucleotide contexts across sequences from this chimpanzee–human transmission pair, indicating variable APOBEC-mediated editing across hosts and anatomical compartments. This pattern aligns with previous studies demonstrating that APOBEC3 activity imposes strong but heterogeneous restriction on SFV following zoonotic transmission [56,57]. The presence of both GG- and GA- signatures is consistent with the activity of multiple APOBEC3 enzymes, while the strong correlation between these contexts suggests that variation among sequences primarily reflects differences in overall APOBEC exposure or survival of edited proviruses rather than mutually exclusive enzyme activity. Together, these findings support a model in which SFV proviral diversity following zoonotic transmission is shaped by low-level replication combined with heterogeneous, host- and compartment-dependent APOBEC-mediated restriction.

Our study has several limitations. Our findings are based on a single SFV-infected worker, limiting generalizability; further studies are needed to confirm these results in other humans with SFV infection originating from chimpanzees and other NHP species. NGS was not available during our clonal sequence analysis, which would have allowed more sensitive detection of viral variants in body specimens. However, our approach and results are comparable with previous studies using PCR clones to investigate viral diversity [14,53]. Although NGS can detect rare variants, in our study, it did not identify the Δtas and ΔLTR strains detected with nested PCR. These results likely reflect detection by NGS of replication-competent SFV in viral RNA from tissue culture, whereas nested PCR of genomic DNA detects latent, defective SFV. Our finding of Δtas and ΔLTR proviruses across five years and anatomical compartments indicate that they represent a stable component of the Case 6 proviral population rather than transient or rare variants. Finally, our in silico APOBEC analyses are limited to context-based inference in a single genomic region and transmission pair but nonetheless reveal consistent, biologically plausible patterns of APOBEC-mediated restriction across hosts and compartments.

In summary, we identified significant in vivo genetic differences in SFV between Case 6 and the source chimpanzee B1, including large deletions in the U3 LTR and tas genes of Case 6 across multiple timepoints and body sites. These deletions in regions with known regulatory functions may likely restrict viral replication in Case 6. We also found evidence of APOBEC restriction and compartmentalization in Case 6 urine, which may further limit viral replication and transmission. Our findings provide a better understanding of viral adaptation following zoonotic SFV transmission to humans.