Evolutionary characterization of antiviral SAMD9/9L across kingdoms supports ancient convergence and lineage-specific adaptations

Legrand, Alexandre; Demeure, Rémi; Chantharath, Amandine; Rey, Carine; Baltenneck, Julie; Gilchrist, Cameron L. M.; Rocha, Joana L.; Loyer, Clara; Picard, Léa; Cimarelli, Andrea; Steinegger, Martin; Rousset, Francois; Sudmant, Peter H.; Etienne, Lucie

doi:10.1038/s41559-025-02845-x

Download PDF

Article
Open access
Published: 11 November 2025

Evolutionary characterization of antiviral SAMD9/9L across kingdoms supports ancient convergence and lineage-specific adaptations

Nature Ecology & Evolution volume 9, pages 2206–2222 (2025) Cite this article

7281 Accesses
3 Citations
25 Altmetric
Metrics details

Subjects

Abstract

Human SAMD9 and SAMD9L are duplicated genes that encode innate immune proteins restricting poxviruses and lentiviruses, such as human immunodeficiency virus (HIV). Mutations in these genes are implicated in genetic diseases and cancer. Here we combined structural similarity searches, phylogenetics and population genomics with experimental assays of SAMD9/9L functions to resolve the evolutionary and functional dynamics of these immune proteins, spanning from prokaryotes to primates. We discovered structural analogues of SAMD9/9L in the antibacteriophage defence system, Avs, resulting from convergent evolution. Further, the predicted nuclease site was conserved in bacterial analogues and was essential for cell death, suggesting a fundamental role in defence across different life kingdoms. Despite this shared immunity, we identified genomic signatures of evolutionary arms races in mammals, with remarkable gene copy number variations. We further unveiled that the absence of SAMD9 in bonobos corresponds to a recent gene loss still segregating in the population. Finally, we found that chimpanzee and bonobo SAMD9Ls have enhanced anti-HIV-1 functions compared with the human orthologue. These SAMD9/9L adaptations probably resulted from strong viral selective pressures, including by lentiviruses, and could contribute to lentiviral resistance in bonobos. Evolutionary characterization of Avs9 and SAMD9/9L provides a deeper understanding of how the immune system adapts to fight viruses over billions of years of evolution.

SAMD9 senses cytosolic double-stranded nucleic acids in epithelial and mesenchymal cells to induce antiviral immunity

Article Open access 22 April 2025

SAMD4 represses VGLL4 mRNA to activate TEAD and promote cancer progression

Article Open access 21 April 2026

Birth of protein folds and functions in the virome

Article Open access 26 August 2024

Main

The continuous arms races between viruses and their hosts have driven the evolution of several immune defences across all life forms^1,2. In humans, an emerging actor of cell-autonomous antiviral immunity is the gene family composed of the paralogues sterile alpha motif (SAM) domain-containing proteins 9 and 9L (SAMD9/9L). These interferon-stimulated genes are in tandem on human chromosome 7 and encode large, multidomain proteins with antiviral properties against poxviruses and lentiviruses^3,4,5. SAMD9 and SAMD9L inhibit cellular and viral protein translation^6,7,8, acting through an essential ribonuclease site in the AlbA_2 domain^5,9. They are also modulators of endosomal trafficking^10,11 and SAMD9 was recently identified as a nucleic acid sensor¹². Deleterious germline mutations in human SAMD9 or SAMD9L dysregulate their activity leading to severe life-threatening genetic syndromes, such as MIRAGE (myelodysplasia, infection, restriction of growth, adrenal hypoplasia, genital phenotypes and enteropathy)¹¹, SAMD9L-associated autoinflammatory disease¹³, ataxia-pancytopaenia¹⁴ and normophosphataemic familial tumoural calcinosis¹⁵. Although the SAMD9/9L gene family extends beyond humans, its evolutionary history and functional diversification remain largely unexplored.

Recently, some human immune genes were shown to have a deep evolutionary origin in bacterial defences against phages¹⁶. For example, major eukaryotic antiviral immune sensors or effectors, such as cGAS or Viperin, have originated from prokaryotic antiviral systems^16,17,18,19. Ancestral immunity therefore broadly defines shared immune defences between prokaryotes and eukaryotes through the presence of conserved immune modules (domains or proteins)¹⁸, that results from horizontal gene transfer, convergent evolution or vertical inheritance.

SAMD9 and SAMD9L are large multidomain proteins, which are presumed members of the STAND (signal transduction ATPases with numerous domains) superfamily²⁰. SAMD9/9Ls are composed of (from N to C terminus): a SAM, a Schlafen (SLFN)-like AlbA_2 domain with nuclease site, a SIR2, a P-loop NTPase, tetratricopeptide repeats (TPRs) and an oligonucleotide/oligosaccharide-binding OB-fold. All are predicted domains, except the AlbA_2 for which the structure was recently solved²¹. Computational analyses suggested the presence of homologues of SAMD9/9L domains in other animals and bacteria²⁰. However, the ancient evolutionary and functional history of SAMD9/9Ls and its potential link to immunity outside mammals remain unknown.

In mammals, the gene family consists of two paralogues, SAMD9 and SAMD9L, which originally duplicated in placentals and have evolved under positive selection²², suggesting past molecular arms races with pathogens. Interestingly, these mammalian paralogues exhibit both functional redundancy and divergence in their antiviral functions. Both are restriction factors against poxviruses, but with species-specific variations in their susceptibility to poxviral countermeasures⁴. In humans, they have different functions in human immunodeficiency virus (HIV) and lentiviral infections: SAMD9L is antiviral, while SAMD9 has no, or a proviral, effect⁵.

Lentiviruses naturally infect most African non-human primate species (simian immunodeficiency viruses (SIVs)), with the notable exception of some species, such as bonobos^23,24,25,26. SIVcpz from chimpanzees is at the origin of the HIV-1 group M, responsible for the AIDS pandemic²⁷. Lentiviruses have co-evolved with primates for millions of years and have been selective drivers of adaptation in primate antiviral defence factors, such as APOBEC3G, Tetherin/BST-2 or TRIM5 (refs. ^28,29,30). These species- and lineage-specific adaptations have further shaped host molecular barriers to cross-species transmissions^31,32.

In this study, we combined AI-predicted structural similarity searches, phylogenetics and population genomics with experimental assays of SAMD9/9L functions in prokaryotes and great apes to resolve the evolutionary and functional dynamics of this antiviral system across scales. We notably found that SAMD9/9Ls and some prokaryotic antiviral STAND (Avs) originated from convergent evolution and depend on AlbA_2 domain for their activity, and that SAMD9/9Ls have evolved under recurrent diversifying evolution in mammals by genomic structural variation, notably gene losses and adaptive polymorphisms to lentiviruses in great apes.

Results

SAMD9/9L bacterial structural analogues with conserved multidomain architecture and predicted antiphage activity

Benefiting from recent advances in structural similarity methods, we investigated the evolutionary history of SAMD9/9Ls across the tree of life. We first used Foldseek³³, which allows fast protein structure alignment and search, with the AlphaFold-predicted SAMD9/9L structures and amino acid sequences as inputs (Methods). We identified 238 hits that share high structural similarity, mainly belonging to bacteria and metazoa (30% template modelling (TM) score and 80% coverage cut-off; Supplementary Table 2). Analysis of bacterial hits with DefenseFinder³⁴, which performs hidden Markov model (HMM) searches against a database of prokaryotic antiviral systems, showed that some of these hits belong to the Avs family of antiphage immune proteins^35,36,37,38. Avs proteins are nucleotide-binding oligomerization domain-like receptor (NLR)-like proteins that sense viral phage infections by their C-terminal sensor domains and perform specific antiviral functions through their N-terminal effector domains. The effector domain may have diverse specific activities, such as ATP degradation via a PNP domain or NAD⁺ depletion via a SIR2 domain^17,37,39,40. Here we found that SAMD9/9L shares strong structural similarity with Avs5 from Desulfobacula sp. and from a newly identified protein, which we named ‘Avs6’, from Labedella endophytica (Uniprot IDs A0A1F9N8W4 and A0A3S0VH60, respectively) (Fig. 1a). However, Avs5 and Avs6 lack the N-terminal AlbA_2 domain from SAMD9/9L, with Avs6 instead encoding a PNP domain predicted to degrade ATP molecules (Fig. 1a). Remarkably, our analyses also identified other bacterial proteins sharing strong structural similarity with human SAMD9/9L on up to 88% of the protein length (for example, 1,407 of 1,589 amino acids total in SAMD9 and 1,400 of 1,584 in SAMD9L for A0A7T4VS34 from Pseudomonas fluorescens). This similarity covers all its functional domains, except the first N-terminal SAM, which seems exclusive to eukaryotes⁴¹. Despite the low amino acid sequence identity (around 15%), the structural similarity is highly significant (TM score up to 0.56; Supplementary Fig. 1a)⁴². We therefore propose ‘Avs9’ as a name for bacterial Avs proteins harbouring the same domain composition as SAMD9/9L at the exception of the SAM domain, referring to the latter gene family name (Fig. 1a). Furthermore, our analysis identified hundreds of uncharacterized proteins across various domains of life: metazoa, bacteria and archaea. It is however notable that, although STAND proteins are also present in Plantae and fungi^43,44, we did not retrieve any significant hits in these kingdoms of life, neither in algae nor in protozoa. Further searches with various inputs, including Avs9, and more relaxed parameters did not recover any.

Fig. 1: Structural homology analyses show strong similarity between SAMD9/9L gene family and Avs antiphage proteins. — **Fig. 1: Structural homology analyses show strong similarity between *SAMD9*/9L gene family and Avs antiphage proteins.**

To investigate the evolutionary history of the structural analogues, we performed maximum likelihood phylogenetic analyses, using IQ-TREE, of both the sequence and the structural alignments of these proteins (from Muscle and FoldMason, respectively) (Supplementary Fig. 1b). We found that SAMD9/9Ls and Avs9s were in two distinct clades falling within the same lineage. However, because the proteins bear different domain compositions, we next performed analyses for specific domains, individually (Supplementary Fig. 1c). Starting with the central P-loop NTPase domain, which is a domain common to all hits, we similarly found that Avs9 and SAMD9/9L clades fell into the same lineage in the phylogeny (Supplementary Fig. 1d). Next, we only kept proteins encoding for an AlbA_2 effector domain, resulting in 23 hits (the small number of hits are due to the AlphaFold Database clustered at 50% identity): 13 eukaryotic SAMD9/9Ls and 10 bacterial Avs9s. The resulting tree topology also showed two clades corresponding to SAMD9/9Ls and Avs9s (Fig. 1b). We further observed scattered absence of SAM, OB-fold or SIR2 domains (Fig. 1b). While this suggests that these proteins have a propensity for domain modularity and that these domain losses could be tolerated, the functional impact of such modulation would require further investigations. Overall, the extensive structural similarity of SAMD9/9L- and Avs9-like proteins observed across such a vast evolutionary range is remarkable. It suggests a strong selective advantage driving the structural and functional conservation of this putative immune antiviral system from bacteria to mammals.

Eukaryotic SAMD9/9Ls and bacterial Avs9s result from convergent evolution

To determine whether eukaryotic SAMD9/9Ls result from the ancestral acquisition of a full-length bacterial analogue (such as Avs9) or emerged through convergent evolution, we performed new structural similarity searches for AlbA_2-containing proteins and studied the phylogenetic distribution of this effector domain across the tree of life (Fig. 2a,b). First, we found that AlbA_2 was not only present in Avs9-like prokaryotic proteins, but was widely represented in bacteria. Concerning eukaryotes, well-known AlbA_2-containing proteins included members of SAMD9/9L and SLFN antiviral immune factors, which, in our analysis, fell in distinct clades (except for SLFNL1, which is distantly related to other SLFNs⁴⁵ and whose evolutionary history was uncertain), suggesting different evolutionary origins (Fig. 2b and Supplementary Fig. 2b,c). As for the ten identified Avs9s, they all clustered within a single clade that was distant from eukaryotic SAMD9/9Ls (Fig. 2b and Supplementary Fig. 2b,c), showing that full-length Avs9- and SAMD9/9L-like proteins did not originate from a single common ancestor. Overall, the presence of evolutionarily distant AlbA_2 domains together with structural similarity and a similar domain organization strongly suggest that bacterial Avs9s and eukaryotic SAMD9/9Ls evolved similar domain architectures through convergent evolution.

**Fig. 2: Avs9 is part of prokaryotic defence systems and induces cell death in bacteria, through its SAMD9/9L/SLFN-analogous active site in the AlbA_2 domain.**

Antiphage defence systems tend to physically cluster in defence islands of bacterial genomes⁴⁶. To investigate the possible immune function of bacterial AlbA_2 domains, we used DefenseFinder to evaluate their propensity to be encoded in genomic neighbourhoods of known defence systems. We found a significant association of bacterial AlbA_2 domains with defence systems in all tested clades, including Avs9 (Fig. 2b), strongly suggesting that their primary function is immune defence.

AlbA_2 domain is a shared effector determinant of the activity of SAMD9/9Ls and Avs9s

To investigate the function of bacterial Avs9 proteins, we first looked for the conservation of the AlbA_2 catalytic site. This active ribonuclease site is composed of three to four negatively charged residues, which are necessary for SAMD9/9L and SLFN11/12/13/14 antiviral activities^{5,9,47,48,49,50,51,52}. Here we found similar residues and structure in Avs9 AlbA_2 (E14, D19, D45, E65; Fig. 2c and Supplementary Fig. 2a), suggesting the presence of similar catalytically active site and function.

Second, to investigate Avs9 activity in vivo, we selected and synthesized an Avs9 from P. fluorescens (A0A7T4VS34), and we cloned its full-length gene, or the AlbA_2 domain alone, under an inducible promoter into Escherichia coli. Expression of full-length Avs9 (Fig. 2d), or the AlbA_2 domain alone (Supplementary Fig. 2d), led to major cell death upon mild induction at 25 °C. Interestingly, the P. fluorescens Avs9 cell death induction was abolished by introducing a single-residue mutation in the AlbA_2 predicted catalytic site (D45N; Fig. 2d and Supplementary Fig. 2d). Of note, an equivalent mutation in human SAMD9/9L and SLFN AlbA_2 also abolishes their activity. Therefore, our results strongly suggest a cell-killing defence function of bacterial Avs9, dependent on AlbA_2 and its predicted nuclease active site.

Ancient duplication followed by frequent CNVs of the SAMD9/9L gene family in mammals

Although SAMD9/9Ls are present in several vertebrate hosts (Fig. 1b), the SAMD9/9L duplication at the origin of the human paralogous SAMD9 and SAMD9L genes seems more recent. This duplication was previously described in placental eutherians, after the divergence of marsupials from placentals²². To address the evolutionary history of the duplicated SAMD9/9L, we first performed genomic analyses of the SAMD9/9L gene family locus in various vertebrate species focusing on mammals: 38 ungulates, 22 chiropterans, 45 carnivores, 34 primates, 41 rodents and 27 additional species from other mammalian and non-mammalian vertebrate orders (Fig. 3a). We found that all these analysed vertebrate species, except monotremes (n = 2), exhibited at least one gene from the SAMD9/9L gene family (Fig. 3a).

Furthermore, thanks to genomic sequence advances and contrary to previous findings²², our observations indicated that marsupials also possess two copies and therefore the gene duplication of SAMD9/9Ls was not restricted to eutherians (Fig. 3a,b). To elucidate the origin of SAMD9/9L in mammals, we performed phylogenetic analyses from 301 homologue sequences of 189 mammalian species spanning 160 million years of divergence (using the National Center for Biotechnology Information (NCBI) blastn implemented in DGINN⁵³) and from 18 non-mammalian vertebrate species (12 aves and 6 amphibians) as outgroups (Fig. 3b with selected species and IQ-TREE tree, Supplementary Fig. 3a,b for the complete IQ-TREE and PhyML trees using sequences from Supplementary Table 3; of note, PhyML trees were similar to IQ-TREE trees at key branches). Despite the presence of two copies in marsupial genomes (Fig. 3a), one of them, named SAMD9m, did not group with the eutherian SAMD9 or SAMD9L clade in the homologous gene tree, but branched outside (Fig. 3b, statistical significance assessed from 1,000 bootstrap replicates). Therefore, it is likely that the marsupial SAMD9m resulted from an ancestral independent duplication predating the divergence of marsupials and eutherians (Fig. 3b). Following this divergence, one copy was potentially lost in eutherians, followed by a subsequent duplication event (Fig. 3c). Alternatively, other hypotheses may involve a single duplication event followed by gene conversion within the two eutherian copies, or loss of a copy through incomplete lineage sorting.

In placentals, following the duplication event that gave rise to SAMD9 and SAMD9L orthologues, we identified at least ten independent losses of one of the two paralogues throughout mammalian evolution (five losses in the main represented lineages of Fig. 3a, as well as additional losses within orders: Fig. 3d and Supplementary Fig. 3c). Notably, artiodactyls experienced the loss of SAMD9L, while carnivores lost SAMD9. Furthermore, although the synteny and the copy numbers remained largely conserved within the divergence of the last two groups, radiations of primates, bats and rodents, exhibited many SAMD9/9L gene losses and some duplication events (Fig. 3d and Supplementary Fig. 3c; Data availability).

Ancient and recent unfixed gene losses during primate SAMD9/9L evolution, through different genetic and genomic mechanisms

Genomic structural changes, such as gene duplication and loss (copy number variations (CNVs)), alongside mutation and recombination events, are an important source of genetic diversity upon which natural selection can act, and thus have strong adaptive potential during virus–host arms races^54,55. Such genomic adaptations during mammalian evolution have been rampant in response to past lentiviral or poxviral epidemics^30,56. Because of (1) the very high rate of CNVs in primate SAMD9/9Ls, (2) the co-evolution of primates with lentiviruses for millions of years³² and (3) the potential lentivirus-driven adaptation of SAMD9/9L (ref. ⁵), we performed in-depth phylogenomic, genetic and positive selection analyses in primates.

We identified at least four independent gene loss events during primate evolution (Fig. 4a). SAMD9L was lost in prosimians, while bonobos and the common ancestors of Platyrrhini and Colobinae experienced the independent loss of SAMD9. Genomic analyses showed that SAMD9 loss in Platyrrhini resulted from the complete loss of the SAMD9 genomic locus in the common ancestor. In contrast, SAMD9 loss in Colobinae occurred through a different genetic mechanism by single nucleotide changes introducing several premature stop codons in the coding sequence (Fig. 4a and Supplementary Fig. 4a).

**Fig. 4: Ancient and recent unfixed gene losses in primates, through different genetic and genomic mechanisms.**

We next specifically tested whether episodes of positive selection occurred in SAMD9 or SAMD9L in some primate lineages that experienced paralogue loss. We performed targeted branch-specific analyses, using the adaptive branch-site random effects likelihood (aBSREL), testing specifically lineages associated with gene loss or paralogue retention (‘tested branch’, also known as ‘foreground’ branches) in the SAMD9 or the SAMD9L gene trees (others branches were set as ‘background branches’)⁵⁷ (Supplementary Fig. 4b,c). Our analysis suggested that several primate lineages that lost either SAMD9 or SAMD9L might have been the targets of episodic positive selection (Supplementary Fig. 4b,c). These losses could result from adaptation to ancient pathogen challenges, cost of maintaining the two copies or relaxation of functional constraints that occurred on specific branches.

Bonobos and chimpanzees are human’s closest living relatives with a genetic divergence of approximatively 1.3% with humans and only of 0.4% amongst themselves⁵⁸. Despite this strong proximity, bonobos possess a unique genetic distinction, with a 41.46 kilobase (kb) deletion in the SAMD9/9L locus. In fact, they stand out as the sole hominid with a single copy of the SAMD9/9L gene family, retaining only SAMD9L (ref. ⁵⁹). To characterize the recent loss of SAMD9, we investigated the prevalence of the 41.46 kb deletion in the SAMD9/9L locus in bonobos at a population level (Pan paniscus; n = 13) in a joint analysis that additionally included all currently recognized chimpanzee subspecies or populations (Pan troglodytes spp.; n = 59) (Fig. 4b). We found that, among 13 individuals, 10 bonobos exhibited the same deletion with the complete absence of the SAMD9 gene, indicating common homozygous genomic deletion (Fig. 4c). However, the remaining three bonobos presented this deletion on a single chromosome, representing a heterozygous genomic absence of SAMD9 (Fig. 4c). Importantly, the analysed bonobo individuals are not related⁶⁰, suggesting that, despite the small sample size, the heterogeneous genomic makeup in SAMD9/9L may be representative of the population. This was in sharp contrast with chimpanzees, which had SAMD9^+/+ present in all 59 individuals, a pattern shared with humans (Fig. 4c). This suggests a recent SAMD9 genomic loss event specific to the bonobo lineage still segregating in the population, potentially impacting the immunity of bonobos.

Chimpanzee and bonobo (Pan) SAMD9Ls have an increased anti-HIV-1 activity compared with human SAMD9L

Beyond the genomic loss that occurred in the SAMD9/9L locus during hominid evolution, we investigated the evolutionary divergence of SAMD9 and SAMD9L at the genetic level. We therefore analysed the non-synonymous single nucleotide polymorphisms (SNPs) within the coding sequences of both genes for the 72 Pan individuals, using panTro6 genome as reference, as well as for more than 4,000 humans (Fig. 5a and Supplementary Fig. 5a,b). For SAMD9, we found very few non-synonymous SNPs amongst Pan. This amino acid conservation was particularly exemplified in the 12 Eastern chimpanzees (Pan troglodytes verus) and in the 3 bonobos encoding a single SAMD9 copy, as they encoded an identical protein sequence (Fig. 5a). In SAMD9L, we found a more widespread distribution of missense SNPs in chimpanzees and bonobos (Fig. 5a). Remarkably, we found two specific variants in SAMD9L that stood out because of their high frequency in the bonobo population. All bonobos (n = 13) encoded for a homozygous serine (S) at position 90 and 9 out of 13 bonobos encoded an arginine (R) at position 1,446 (either homozygous or heterozygous) (Fig. 5a and Supplementary Fig. 5a). Of note, there were no SAMD9L SNPs differentiating SAMD9^+/− and SAMD9^−/− bonobos. Intriguingly, chimpanzees and all other primates analysed to date, including the 4,099 human genomes, encoded a leucine L90 and a lysine K1446 (Fig. 5a and Supplementary Fig 5a,c), suggesting that these variants are specific to bonobos amongst primates. Furthermore, compared with bonobos, Eastern and Central chimpanzees showed a more widespread distribution of SNPs within SAMD9L, and no missense polymorphisms appeared at high frequency in the chimpanzees (Fig. 5a and Supplementary Fig. 5a).

Fig. 5: Chimpanzee and bonobo (Pan) SAMD9L major variants have an increased restrictive activity against HIV-1 pWITO, but not SIVcpz EK505, compared with human SAMD9L. — **Fig. 5: Chimpanzee and bonobo (*Pan*) SAMD9L major variants have an increased restrictive activity against HIV-1 pWITO, but not SIVcpz EK505, compared with human SAMD9L.**

Human SAMD9L inhibits cellular protein synthesis and restricts lentiviral HIV-1 infection^5,8,61,62. This is particularly interesting in the context of natural lentiviral infections in hominids, where humans and two chimpanzee subspecies are infected by HIV-1 and SIVcpz, respectively. Yet, two chimpanzee subspecies (Eastern and Nigerian-Cameroon chimpanzees) and bonobos have no evidence of modern natural lentiviral infections^25,26,27,63 (Fig. 4b). Furthermore, pandemic HIV-1 in humans originated from cross-species transmission of SIVcpzPtt from Central chimpanzees (Pan troglodytes troglodytes)^27,64.

We therefore determined the functional consequences on lentiviral infections of the genotypic differences between human, chimpanzee and bonobo SAMD9Ls. We cloned the native chimpanzee SAMD9L and the two major variants of bonobo SAMD9L (bonobo R1446 and bonobo K1446) in an expression plasmid to compare them with human SAMD9L (Fig. 5b,c). Of note, the chimpanzee SAMD9L has eight amino acid changes compared with the human one (Fig. 5b). We investigated their functions in the context of HIV-1 replication, as in ref. ⁵, as well as SIVcpzPtt replication (SIVcpz EK505, a kind gift from B. Hahn^27,65). Briefly, we cotransfected 293 T cells with the mCherry-SAMD9L plasmids along with an infectious molecular clone (IMC) encoding a transmitted/founder HIV-1 natural strain (pWITO, as in ref. ⁵) or an SIVcpzPtt strain (EK505). Two days later, we measured viral and cell protein expression in the producer cells (Fig. 5c–e and Supplementary Fig. 5d) and the infectious virus yield by Tzm-bl reporter assay (Fig. 5c–e). Remarkably, although all ectopic SAMD9Ls were expressed at similar levels, we found that chimpanzee and bonobo SAMD9Ls had a significant increase in anti-HIV-1 pWITO activity compared with human SAMD9L (Fig. 5d, P < 0.005), suggesting some species-specificity.

By testing effects on SIVcpzPtt EK505, we first showed that human SAMD9L was also restrictive against this SIVcpzPtt strain. Yet, chimpanzee and bonobo SAMD9Ls did not present an increased anti-SIVcpz activity compared with human SAMD9L, suggesting possible lentiviral-strain specificity. Therefore, chimpanzee and bonobo SAMD9Ls seem to have an increased anti-HIV-1 activity, as compared with human SAMD9L (Fig. 5e). It is further possible that SIVcpz, naturally circulating in chimpanzee populations, adapted to the Pan SAMD9L increased antiviral function.

Bonobo-specific polymorphisms confer an increased anti-HIV-1 activity to human SAMD9L without compromising cellular protein synthesis

We wondered whether minimal changes in human SAMD9L informed from some of these natural Pan variants could impact human SAMD9L functions. We specifically tested the bonobo SNPs, which could constitute species-specific adaptations with functional implications. We therefore cloned L90S and/or K1446R variants in the context of the human SAMD9L plasmid and investigated their effects on two key functions of human SAMD9L: antilentiviral function as well as cellular protein synthesis shutdown (Fig. 6).

**Fig. 6: Bonobo-specific polymorphisms enhance human SAMD9L anti-HIV-1 activity, without affecting its translation shutdown effect.**

First, we investigated the effect of these bonobo-specific variants on human SAMD9L function in the context of lentiviral replication, as in Fig. 5. Interestingly, we found that ectopic SAMD9L-L90S/K1446R and the single variants were expressed at a similar level to wild-type SAMD9L, but had a significant twofold increase in anti-HIV-1 pWITO activity (Fig. 6a). The double-mutant SAMD9L-L90S/K1446R appeared the most restrictive, while the single mutants had intermediate effects, suggesting additive functions (Fig. 6a). The increased effects of the human SAMD9L mutant with L90S and/or K1446R seemed independent of HIV-1 protein translation shutdown (Supplementary Fig. 6a).

Second, we tested the activity of SAMD9L-L90S/K1446R on SIVcpzPtt EK505 replication and found that, similarly to wild-type Pan SAMD9Ls, the specific SNPs in the context of the human SAMD9L did not increase its anti-SIVcpz function (Fig. 6b).

Lastly, we assessed the extent of whole cellular translation in human 293 T cells transfected with high doses of wild-type human mCherry-SAMD9L or mCherry-SAMD9L-L90S/K1446R or the single mutants. We used Click-iT l-homopropargylglycine (HPG) synthesis assays, which measures the incorporation of HPG an analogue of methionine into newly synthesized proteins by flow cytometry Supplementary Fig. 6b). As previously shown^5,8,61,62, we found a dose-dependent shutdown of cellular translation in cells ectopically expressing human SAMD9L (mCherry⁺) normalized to the control mCherry⁻ cells (Fig. 6c and Supplementary Fig. 6b). Importantly, we found that hSAMD9L-L90S/K1446R and the single variants showed no differences in the cellular translation repression compared with the human wild-type SAMD9L (Fig. 6c). This suggests that the non-synonymous bonobo variants in SAM and TPR domains do not change the activity of human SAMD9L on cellular protein synthesis.

Overall, bonobo-specific polymorphisms specifically enhance human SAMD9L antiviral function against HIV-1, without affecting the translation shutdown function.

Discussion

This study unveils key aspects of the functional evolution of SAMD9/9Ls at different timescales, highlighting multidomain and functional convergence between metazoan SAMD9/9Ls and prokaryotic Avs9s, as well as recurrent genetic and genomic adaptations in mammals from ancient to very recent times. On one hand, we identified SAMD9/9L structural analogues in bacterial defence systems that induce cell death with similar AlbA_2 effector determinants, suggesting a remarkable shared immune factor across billions of years. On the other hand, our analyses of mammalian SAMD9/9L revealed dynamic and episodic adaptations, notably in primates, probably in response to epidemics, including from lentiviruses. By an ‘immuno-evo’ framework, we bring key insights to understand the duality between the maintenance of a key antiviral shared immunity and its constant adaptations to pathogens.

SAMD9/9L as shared immunity in prokaryotes and metazoans

The SAMD9/9L gene family may be part of the ‘shared immunity’, in which antiviral mechanisms are similar between life kingdoms, either resulting from (1) horizontal gene transfer of bacteria, (2) from vertical inheritance originating from LUCA (last universal common ancestor) or (3) from convergent evolution^16,18,66. The list of shared immune defence systems is rapidly expanding with currently about a dozen identified antiviral systems, including cGAS, Viperin and TLRs¹⁸. Here we revealed striking structural similarity between human antiviral SAMD9/9L and prokaryotic Avs proteins³⁷, specifically with a new Avs protein family (Avs9). Notably, they both share the key nuclease domain, AlbA_2, which, in human SAMD9/9L, is responsible for transfer RNA^Phe cleavage and viral and cellular translational inhibition^5,9. In Avs9 from P. fluorescens, we uncovered cell-killing activity, which also depended on AlbA_2 and its predicted nuclease site. Future biochemical and functional investigations will determine if Avs9 has bone fide nuclease activity—and which substrate—as well as identify if Avs9 has specific antiphage functions. Furthermore, similar to the Avs system, in which infected bacteria use an altruistic self-killing mechanism for the benefit of the colony³⁷, tight regulation of SAMD9/9L is probably essential and is certainly possible thanks to the intermediate and C-terminal domains⁸.

Uncovering phylogenetic relationships of prokaryotic Avs9 and metazoan SAMD9/9L protein families strongly suggest that they result from evolutionary convergence, possibly driven by analogous viral selective pressure. Moreover, domain shuffling is an important driver of protein evolution and can give rise to similar advantageous multidomain combinations. This work exemplifies a surprising evolutionary and functional parallel through the independent emergence of analogous antiviral proteins across vast evolutionary timescales.

Extensive, and probably adaptative, SAMD9/9L CNVs

Despite its simultaneous presence in evolutionarily distant organisms, we globally found a patchy distribution of SAMD9/9L homologues and structural analogues across domains of life. For example, plants, fungi, algae and protozoa do not seem to harbour complete SAMD9/9L homologues. Furthermore, we found a very rapid evolution at the genomic and genetic levels during mammalian evolution in SAMD9/9L. Therefore, this ancient conservation is concomitant with rapid evolution, almost certainly as the result of virus–host arms races. Gene loss and duplication may, for example, provide an advantage, similar to observations in other innate immune genes, such as the APOBEC3 family^31,67,68,69 and many other factors¹. It is noteworthy that most genomic variations were gene losses, rather than extensive duplications, at least in most analysed mammalian species. These may be the result of evolutionary trade-offs, where the benefits of losing a gene (potentially escaping viral antagonism or hijacking) outweighed the costs.

At the mechanistic level, while most SAMD9/9L losses occurred by genomic loss of a chromosomal region, Colobinae primates seem to have lost SAMD9 by early stop codons and pseudogenization. However, it cannot be excluded that, in some species, this may lead to the expression of a truncated SAMD9 retaining the AlbA_2 effector function but lacking the crucial regulatory intermediate and C-terminal domains, similar to some SAMD9L autoinflammatory gain-of-function variants, for example^8,61. Further study on its selective pressure, and on messenger RNA transcript and protein expression in natural tissues from Colobinae species under immune stimulation would resolve this question.

The specific case of SAMD9 unfixed loss in bonobos and adaptive chimpanzee and bonobo SAMD9Ls: modern implications, functions and potential past lentiviral drivers

Bonobos harbour a recent unfixed loss of SAMD9, which occurred through a large chromosomal deletion. Despite bonobos, chimpanzees and humans being closely related, the variability observed at this locus is intriguing. Two bonobo-specific missense polymorphisms in SAMD9L that confer an increased antiviral activity against HIV-1 are located in the SAM and TPR domains, which could be involved in protein–protein or protein–RNA interactions^37,70,71,72. One possibility is that the variants may modulate SAMD9L sensing, especially for the TPR domains, which have been reported to act as (viral) sensors in Avs and IFIT proteins (interferon-induced protein with TPRs)^37,73. Further, although most human deleterious gain-of-function mutations in SAMD9/9L associated diseases are in the P-loop NTPase domain, some are described in the TPR or SAM domains⁷⁴. However, we did not observe a gain-of-function phenotype on global cellular translation for the bonobo-specific variants. This therefore suggests that the variants may not destabilize the inactive closed form of the protein, nor change SAMD9L basal activities. Instead, it might modify its specificity and sensitivity in viral sensing, potentially adapting its interface with viruses. Otherwise, it may impact other potential functions of SAMD9L, for example in endosomal trafficking, increasing specific anti-HIV functions^5,10,11. It would also be interesting to determine if, and how, chimpanzee and bonobo SAMD9Ls restrict other infections from poxviruses or other RNA viruses^3,12, and whether those may have driven some of the adaptations.

Our data show that chimpanzee and bonobo SAMD9Ls have an increased anti-HIV-1 phenotype compared with human. The additional loss of the pro-HIV-1 SAMD9 (ref. ⁵) may have been particularly advantageous (overall increased fitness) during past lentiviral infections in bonobo ancestors (that is, increased antiviral SAMD9L and loss of prolentiviral SAMD9). The presence of 3 of 13 unrelated bonobos with 1 SAMD9/SAMD9L allele and 1 SAMD9L-only allele suggests that SAMD9 loss is unfixed, and probably recent. If selection favours individuals without SAMD9, the gene could eventually be completely lost in the bonobo population. The modern genomic makeup of the SAMD9/9L locus in chimpanzees, and even more so in bonobos, therefore suggests adaptation to lentiviral-like epidemics that occurred in Pan, as well as since the bonobo–chimpanzee divergence. In the future, performing this evo-functional study with a larger bonobo sample size, genomic phased data and long-read sequencing would enable robust determination of the exact selective pressures shaping antiviral defence mechanisms in this species. It would also help in determining the haplotype structures on which the SAMD9 gene was lost and the SAMD9L substitutions occurred, potentially providing further insight into the epistatic interactions between these genetic changes.

Unlike some chimpanzees and humans that are infected by SIVcpz and HIVs, respectively, and suffer from AIDS symptoms^75,76,77, modern bonobos are not known to be naturally infected by any lentiviruses^25,26. Overall, SAMD9/9L adaptation may nowadays participate, with other factors⁷⁸, in bonobo population resistance against lentiviral/SIV infections.

Finally, it is noteworthy that in this study SIVcpzPtt EK505 did not show an increased sensitivity to chimpanzee and bonobo SAMD9Ls, or to Pan SNPs in the context of human SAMD9L. This may be the result of virus–host co-evolution^1,30, where SIVcpz has adapted to the natural genetic makeup of its host, particularly of chimpanzee antiviral innate immunity.

Altogether, our findings highlight the strength of evo-immuno approaches in unravelling links between the evolutionary history of innate immunity and contemporary challenges in human health. The identification of SAMD9/9L homologues and structural–functional analogues across diverse taxa, as prokaryotes and primates, shows a shared convergent immunity. Common challenges, such as fighting viral infections, drive both conservation or convergence of key immune systems as well as their rapid evolution through arms races. In this regard, using diverse models (human, diverse eukaryotic cells and bacteria) and natural variants in closely related species for functional studies can bring valuable insights with broader medical applications, such as the incorporation of potentiator mutations in antiviral factors (protein engineering) or the use of bacterial antiviral proteins that could act against human viruses.

Methods

Comparative genomics, phylogenetics and positive selection analyses in mammals

To obtain the coding sequences of the SAMD9 and SAMD9L homologues in bats, rodents, primates, ungulates and carnivores, we used the detection of genetic innovations (DGINN) pipeline⁵³ with, respectively, Myotis myotis, Rattus norvegicus, Homo sapiens, Hippopotamus amphibius and Phoca vitulina Refseq SAMD9 and SAMD9L, as queries. Briefly, the coding sequences from each group were automatically retrieved with NCBI blastn^79,80, cleaned and aligned with MAFFT⁸¹. Homologous sequences from marsupials, aves and amphibians were retrieved using NCBI blastn. Of note, these analyses are based on publicly available genome annotations (not necessarily genes annotated as ‘SAMD9/9L-like’ but regions annotated as coding regions), so it is not excluded that some unannotated genes were not analysed. The species and accession numbers are presented in Supplementary Table 2. Nucleotide alignments of SAMD9 or SAMD9L from each mammalian group were manually curated before being used as input in DGINN for automatic codon alignment using the probabilistic alignment kit (PRANK)⁸² and phylogenetic tree building using PhyML⁸³ (with default settings in DGINN).

Furthermore, all codon alignments, as well as outgroup sequences, were realigned in a three-step fashion to obtain a high-quality mammalian-wide codon alignment (1) using Muscle⁸⁴, (2) manually curating the sequences, (3) codon aligned with PRANK. A phylogenetic tree was inferred from this alignment using IQ-TREE webserver (GTR + F + I + G4 identified as the best substitution model by ModelFinder implemented in IQ-TREE)⁸⁵. We also performed analyses with PhyML (best model estimated from Smart model selection, SMS: GTR + R). These analyses allowed us to attribute the phylogenetically aware ‘SAMD9’ or ‘SAMD9L’ nomenclature.

We next tested for positive selection occurring at specific branches using the aBSREL program on the DataMonkey webserver⁵⁷. We used as inputs the codon alignments of (1) primate SAMD9 with Cricetulus griseus (criGri) and Tupaia chinensis (tupChi) SAMD9s as outgroups and (2) primate SAMD9L with Cricetulus griseus (criGri) SAMD9L as outgroup. Each branch that we tested for evidence of positive selection was defined as ‘tested branch’ (that is, ‘foreground’) and the remaining as ‘background branches’. Two models were fit to each tested branch: one that allows for episodic diversifying selection (with ω > 1) and one that does not. A likelihood ratio test is then used to compare these models and assess whether the tested branch shows evidence of positive selection. For branches where selection is detected, aBSREL also estimates the proportion of codon sites that are subject to positive selection.

For cases in which we suspected gene losses, we confirmed the absence of coding genes by several methods. We analysed the SAMD9/9L genomic locus (between HEPACAM2 and CDK6 syntenic genes) on NCBI genome data viewer of specific species. We verified that there were no missing data or low sequence quality in this genomic region. Pseudogenes, here identified by several and very early stop codons, were only analysed systematically in primates. For the two monotreme species, in which no SAMD9/9L homologues could be retrieved by genome-wide blast, the HEPACAM2-CDK6 genomic locus was retrieved. We found no missing data (no ‘N’) and no homology using blast or alignments (with relaxed parameters) of non-annotated regions with human SAMD9/9L.

Genome alignment and SNPs analyses in hominids

The genomic sequences of 13 bonobos, 59 chimpanzees and 4,099 humans were retrieved from public online databases: 1000 Genomes Project, Human Genome Diversity Project (HGDP), NCBI bioprojects PRJNA189439, SRP018689 and PRJEB15083 (refs. ^60,86) (The 1000 Genomes Project Consortium 2015). DNA sequences from Pan individuals were aligned to Clint_PTRv2/panTro6 reference genome using BWA-MEM⁸⁷. Then, variant calling was done using FreeBayes⁸⁸ to obtain a vcf file. The SAMD9 and SAMD9L locus regions (chromosome 7: 88599930–89350079 in panTro6) were extracted and parsed using an ad hoc R script using VariantAnnotation, GenomicFeatures, AnnotationHub, org.Pt.eg.db, ggplot2, R packages. This script was used to identify non-synonymous variants among SNPs and to visualize them. The equivalent genomic region in human was retrieved by using the LiftOver tool of the UCSC genome browser (chromosome 7: 92540798–93289621 in the human reference GRCh38/hg38). These coordinates were then used to subset the HGDP + 1000 Genomes Project vcf for this region.

Structure similarity search, structurally aware alignment and phylogenetic analyses across kingdoms of life

Protein structures were obtained from AlphaFold DB^89,90 and RCSB PDB⁹¹. Foldseek³³ was used for detection of structural similarity. We used a sequential strategy. First, we queried Foldseek v.427df8a with SAMD9 (PDB ID Q5K651) and SAMD9L (Q8IVG5) against the AlphaFold database clustered at 50% sequence identity (AFDB50), using a 30% TM-score threshold and a maximum E-value of 0.0001 (default settings). Then, we constructed a query database consisting of SAMD9 (Q5K651), SAMD9L (Q8IVG5) and bacterial top hits—AVAST type V (A0A1F9N8W4) and Avs9 (A0A100VJR7, A0A2S4Y961 and A0A7T4VS34)—and used it for searches under the same settings. Search hits were subsequently filtered for a minimum 80% query coverage to cover at least 1,000 amino acids of the query structures, resulting in 238 analogue structures. FoldMason v.333d54c (ref. ⁹²) was used to generate a multiple structure alignment (MSTA) of the identified structural analogues and MUSCLE v.5 was used to generate a multiple amino acid sequence alignment (MSA). The domain coordinates used for presence/absence analysis of each domain or for the extraction of a given domain MSTA are presented in Supplementary Fig. 1c and are based on the predicted SAMD9 three-dimensional structure. Additionally, using FoldMason, an MSTA was generated on the human SLFNs (SLFNL1, SLFN5 and SLFN11–14, respectively, corresponding to PDB IDs Q499Z3, Q08AF3, Q7Z7L1, Q8IYM2, Q68D06 and P0C7P3) with the 23 hits over 238 containing an AlbA_2 domain. Phylogenetic trees were constructed from both the MSA and MSTA using IQ-TREE 2.3.0 (ref. ⁹³) with the LG + F + G4 substitution model and 1,000 bootstrap replicates, and visualized using the ggtree R library⁹⁴ or interactive tree of life (iTol)⁹⁵.

Phylogenetic analysis of prokaryotic and eukaryotic AlbA_2 domains

The HMM profile of the Pfam AlbA_2 (PF04326) domain was retrieved from the Pfam database⁹⁶. This profile was searched against a custom protein database combining: (1) 41,150 complete bacterial genomes downloaded from Refseq in August 2024, filtered for redundancy using the clusthash function of MMseqs2 (v.13.45111) using default parameters⁹⁷; (2) 455 complete archaeal genomes downloaded from Refseq in August 2024; (3) 993 representative eukaryotic genomes from the EukProt database⁹⁸, filtered for redundancy using the clusthash function of MMseqs2 (13.45111) using default parameters. The AlbA_2 HMM profile was searched into this combined protein database using hmmsearch (v.3.3.2) with default parameters⁹⁹. Hits with at least 90 covered profile residues were selected and the amino acid sequences of the aligned regions complemented with ten residues on each side were extracted. Sequences were clustered using the easy-cluster function of MMseqs2 (v.13.45111) with parameter --min-seq-id 0.8 (ref. ⁹⁷). Cluster representatives were aligned with Clustal-Omega¹⁰⁰ with default parameters and the alignment was trimmed using ClipKit¹⁰¹. The trimmed alignment was used to compute a tree using IQ-TREE with parameters -m L -bb 10000 -nm 10000, which was then visualized using iTOL.

For each genome, defence systems were detected using DefenseFinder (v.1.3)³⁴. For each clade, we calculated a defence score as the fraction of bacterial genes found within ten genes upstream or downstream of a defence protein as detected by DefenseFinder in their genome of origin. As a control, we calculated for each clade the defence score expected by chance as the fraction of all genes (from the same genomes) that are found within ten genes upstream or downstream of a defence protein. This value was used to assess for each clade whether AlbA_2-containing genes colocalize with antiphage systems more frequently than expected by chance, using a binomial test and false discovery rate (FDR) correction (Supplementary Table 1).

Bacterial strains and plasmids

The codon-optimized open reading frame encoding Avs9 from P. fluorescens (Uniprot ID A0A7T4VS34) was ordered as a gene fragment from Twist Bioscience, cloned into the pBbA6c vector¹⁰² by T5 exonuclease-dependent assembly cloning¹⁰³ and transformed into E. coli DH5α λpir. Constructions were sequence-verified by Sanger sequencing (Microsynth). The complete Avs9 gene fragment sequence synthesized and used in this study is provided in Data availability. We further made the Avs9-D45N mutant by site-directed mutagenesis using polymerase chain reaction amplification with Q5 polymerase (New England Biolabs) and KLD cloning (New England Biolabs).

Bacterial drop assays

E. coli DH5α λpir cells carrying pBbA6c, pBbA6c-Avs9 or pBbA6c-Avs^D45N were grown for 6 h at 37 °C and 180 rpm in Luria–Bertani (LB) medium supplemented with chloramphenicol at 20 μg ml⁻¹ and glucose at 1 g ml⁻¹. After tenfold serial dilutions, 5 µl of each dilution were spotted on LB agar plates supplemented with 20 µg ml⁻¹ of chloramphenicol and either 1% glucose or 100–500 µM isopropyl β-d-1-thiogalactopyranoside (IPTG). Plates were incubated at 37 and 25 °C for 24 and 48 h, respectively. Strong toxicity of pBbA6c-Avs9 was observed upon incubation at 25 °C.

Plasmids for expression in human cells

HIV-1 T/F pWITO (human immunodeficiency virus 1 pWITO.c/2474, ARP-11739) encoding a full-length IMC was contributed by J. Kappes and C. Ochsenbauer through the National Institutes of Health (NIH) AIDS repository programme. IMC for SIVcpz EK505 was a gift from B. Hahn^27,65. The pMT06-Flag-mCherry-SAMD9L plasmid was constructed by cloning the synthetized human SAMD9L gene into a pMT06-Flag-mCherry backbone, from the original RRL.sin.cPPT.CMV/Flag-E2-crimson.IRES-puro.WPRE (MT06, a gift from C. Goujon: Addgene plasmid no. 139448; http://n2t.net/addgene:139448; RRID: Addgene_139448)¹⁰⁴. The pMT06-Flag-mCherry-chimp SAMD9L plasmid (chimpanzee SAMD9L) was synthesized and cloned by Azenta Genewiz. The bonobo SAMD9L plasmids were generated through site-directed mutagenesis of the pMT06-FLAG-mCherry-chimp SAMD9L plasmid. Human SAMD9L-L90S/K1446R, SAMD9L-L90S and SAMD9L-K1446R (double and single mutant) plasmids were generated from pMT06-Flag-mCherry-SAMD9L plasmid, using the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent) following the manufacturer’s instructions. Sequences were confirmed through full-length plasmid and/or Sanger sequencing (Microsynth).

Cell lines and culture

Human embryonic kidney 293 T cell lines (ATCC, catalogue no. CRL-3216) and TZM-bl (NIH AIDS Research and Reference Reagent Program, catalogue no. 8129) were grown in Dulbecco modified Eagle medium containing 10% fetal calf serum (Sigma catalogue no. F7524) and 100 U ml⁻¹ of penicillin/streptomycin. TZM-bl cells express the cell-surface proteins CD4, CCR5 and CXCR4 and encode for luciferase and β-galactosidase under the control of the long-terminal repeat (LTR) promoter. They are commonly used for lentiviral titration of culture supernatants (Tzm-bl assays).

Production and quantification of replication-competent lentivirus

A total of 293 T cells were initially seeded in 6-well plates at a density of 0.2 M cells ml⁻¹ (400,000 cells total per well). After 24 h, the cells were cotransfected using TransIT-LT1 (Mirus) with a plasmid encoding a fully replication-competent lentivirus (IMCs), alongside either a plasmid encoding SAMD9L or an empty control. The quantity of DNA used was 250 ng for host plasmids and 1,200 ng for virus plasmids. Subsequently, 48 h post-transfection, cells were harvested for western blot. The supernatants were collected and stored at −80 °C for further titration of infectious virus yield via TZM-bl cells. For titration, TZM-bl cells were plated in 96-well plates and exposed to serial dilutions of viral supernatant. Following 48 h of infection, cell lysis was performed using BrightGlow Lysis Reagent (Promega E2620) and relative light units were measured using the Tecan Spark Luminometer. Infectious virus yields under various conditions were consistently expressed as fold-change compared with paired viral infection conditions in the absence of SAMD9L.

Western blot analysis

Cells were harvested and lysed using ice-cold RIPA buffer (composed of 50 mM Tris pH 8, 150 mM NaCl, 2 mM EDTA and 0.5% NP40) supplemented with protease inhibitors (Roche), followed by sonication. Proteins from cell lysates or supernatants were separated by electrophoresis and transferred onto a PVDF membrane via overnight wet transfer at 4 °C. Stain-Free gel (BioRad) was used for loading and protein transfer controls. Following blocking in TBS-T 1X solution (Tris buffer saline, consisting of Tris HCl 50 mM pH 8, NaCl 30 mM and 0.05% Tween 20) with 5% powdered milk, the membranes underwent incubation with primary antibodies for a duration ranging from 1 h to overnight, followed by subsequent 1-h incubation with secondary antibodies. Detection was carried out using SuperSignal West Pico Chemiluminescent Substrate (ThermoFisher Scientific) and imaged using the Chemidoc Imaging System (BioRad). Antibodies used included anti-SAMD9L (Proteintech, 25173-1-AP), anti-Gag (NIH HIV Reagent Program, 183-H12-5C), anti-HIV-1-gp120 (Aalto, D7324; NIH HIV Reagent Program, 16H3), as well as secondary IgG-peroxidase conjugated anti-mouse (Sigma, catalogue no. A9044) and anti-rabbit (Sigma, catalogue no. AP188P). ‘Total proteins were used as a loading control with BioRad stain-free gel.

Protein synthesis assay

A total of 293 T cells were seeded at 0.2 M cells ml⁻¹ in 12-well plates (200,000 cells total per well). At 24 h after seeding, cells were transfected with 250 or 500 ng of host DNA plasmid, using TransIT-LT1 (Mirus) following the manufacturer’s instructions. At 48 h post-transfection, cells were incubated in HPG for 30 min at 37 °C. Medium was discarded and cells were harvested and fixed with PFA 4%. Cells were then washed with PBS BSA 3% and permeabilized in PBS 0.5% Triton X-100 for 15 min. Click-iT Plus Alexa Fluor Picolyl Azide assay was then performed and cells were analysed on MACSQuant VYB Cytometer (Miltenyi Biotec, SFR BioSciences).

Other softwares and statistical analyses

DefenseFinder³⁴ was used to analyse sequences from Foldseek analyses³³. Sequencing analyses and representations were conducted using Geneious (Biomatters), ESPript 3.0 https://espript.ibcp.fr (ref. ¹⁰⁵) and UGENE v.52.0 (ref. ¹⁰⁶). R scripts were used to conduct analyses of genomic data. Graphic representations and statistical analyses were carried out using GraphPad Prism 9 and R scripts. In the figures, data are presented as mean ± s.d. and each point correspond to an independent biological replicate. Statistics were performed using the two-ratio paired t-test, except where indicated in the figure legends. P values are according to figure legends and exact P values are reported in Supplementary Table 1 and Source data.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data and reagents are available in this manuscript (including Supplementary Information, Supplementary Tables and Supplementary Data) or accessible upon request to the corresponding author. The alignments and phylogenetic trees are all openly available via figshare (phylogenetics associated: with Fig. 1 and Supplementary Fig. 1 at https://doi.org/10.6084/m9.figshare.29082710.v1 (ref. ¹⁰⁷); with Fig. 2 and Supplementary Fig. 2 at https://doi.org/10.6084/m9.figshare.29082710 (ref. ¹⁰⁸); with Fig. 3 and Supplementary Fig. 3 at https://doi.org/10.6084/m9.figshare.29042762 (ref. ¹⁰⁹); with Fig. 4 and Supplementary Fig. 4 at https://doi.org/10.6084/m9.figshare.29087987 (ref. ¹¹⁰); and with Supplementary Fig. 5c at https://doi.org/10.6084/m9.figshare.29088050 (ref. ¹¹¹). The Avs9 sequence used for experimental assays in Fig. 2d and Supplementary Fig. 2d is available via figshare at https://doi.org/10.6084/m9.figshare.29082719 (ref. ¹¹²). Source data are provided with this paper.

Code availability

The scripts for the analyses of the polymorphic sites are available via GitLab at http://gitbio.ens-lyon.fr/ciri/lp2l/vcf_to_nice_figures.git.

References

Tenthorey, J. L., Emerman, M. & Malik, H. S. Evolutionary landscapes of host–virus arms races. Annu. Rev. Immunol. 40, 271–294 (2022).
Article CAS PubMed Google Scholar
Georjon, H. & Bernheim, A. The highly diverse antiphage defence systems of bacteria. Nat. Rev. Microbiol. 21, 686–700 (2023).
Article CAS PubMed Google Scholar
Liu, J. & McFadden, G. SAMD9 is an innate antiviral host factor with stress response properties that can be antagonized by poxviruses. J. Virol. 89, 1925–1931 (2014).
Article PubMed PubMed Central Google Scholar
Meng, X. et al. A paralogous pair of mammalian host restriction factors form a critical host barrier against poxvirus infection. PLoS Pathog. 14, e1006884 (2018).
Article PubMed PubMed Central Google Scholar
Legrand, A. et al. SAMD9L acts as an antiviral factor against HIV-1 and primate lentiviruses by restricting viral and cellular translation. PLoS Biol. 22, e3002696 (2024).
Article CAS PubMed PubMed Central Google Scholar
Meng, X. & Xiang, Y. RNA granules associated with SAMD9-mediated poxvirus restriction are similar to antiviral granules in composition but do not require TIA1 for poxvirus restriction. Virology 529, 16–22 (2019).
Article CAS PubMed Google Scholar
Sivan, G., Glushakow-Smith, S. G., Katsafanas, G. C., Americo, J. L. & Moss, B. Human host range restriction of the vaccinia virus C7/K1 double deletion mutant Is mediated by an atypical mode of translation inhibition. J. Virol. 92, e01329-18 (2018).
Russell, A. J. et al. SAMD9L autoinflammatory or ataxia pancytopenia disease mutations activate cell-autonomous translational repression. Proc. Natl Acad. Sci. USA 118, e2110190118 (2021).
Article CAS PubMed PubMed Central Google Scholar
Zhang, F. et al. Human SAMD9 is a poxvirus-activatable anticodon nuclease inhibiting codon-specific protein synthesis. Sci. Adv. 9, eadh8502 (2023).
Article CAS PubMed PubMed Central Google Scholar
Nagamachi, A. et al. Haploinsufficiency of SAMD9L, an endosome fusion facilitator, causes myeloid malignancies in mice mimicking human diseases with monosomy 7. Cancer Cell 24, 305–317 (2013).
Article CAS PubMed Google Scholar
Narumi, S. et al. SAMD9 mutations cause a novel multisystem disorder, MIRAGE syndrome, and are associated with loss of chromosome 7. Nat. Genet. 48, 792–797 (2016).
Article CAS PubMed Google Scholar
Hou, G. et al. SAMD9 senses cytosolic double-stranded nucleic acids in epithelial and mesenchymal cells to induce antiviral immunity. Nat. Commun. 16, 3756 (2025).
Article CAS PubMed PubMed Central Google Scholar
de Jesus, A. A. et al. Distinct interferon signatures and cytokine patterns define additional systemic autoinflammatory diseases. J. Clin. Invest. 130, 1669–1682 (2020).
Article PubMed PubMed Central Google Scholar
Gorcenco, S. et al. Ataxia-pancytopenia syndrome with SAMD9L mutations. Neurol. Genet. 3, e183 (2017).
Article CAS PubMed PubMed Central Google Scholar
Topaz, O. et al. A deleterious mutation in SAMD9 causes mormophosphatemic familial tumoral calcinosis. Am. J. Hum. Genet. 79, 759–764 (2006).
Article CAS PubMed PubMed Central Google Scholar
Wein, T. & Sorek, R. Bacterial origins of human cell-autonomous innate immune mechanisms. Nat. Rev. Immunol. 22, 629–638 (2022).
Article CAS PubMed Google Scholar
Rousset, F. Innate immunity: the bacterial connection. Trends Immunol. 44, 945–953 (2023).
Article CAS PubMed Google Scholar
Bernheim, A., Cury, J. & Poirier, E. Z. The immune modules conserved across the tree of life: towards a definition of ancestral immunity. PLoS Biol. 22, e3002717 (2024).
Article CAS PubMed PubMed Central Google Scholar
Ledvina, H. E. & Whiteley, A. T. Conservation and similarity of bacterial and eukaryotic innate immunity. Nat. Rev. Microbiol. 22, 420–434 (2024).
Article CAS PubMed PubMed Central Google Scholar
Mekhedov, S. L., Makarova, K. S. & Koonin, E. V. The complex domain architecture of SAMD9 family proteins, predicted STAND-like NTPases, suggests new links to inflammation and apoptosis. Biol. Direct 12, 13 (2017).
Article PubMed PubMed Central Google Scholar
Peng, S. et al. Structure and function of an effector domain in antiviral factors and tumor suppressors SAMD9 and SAMD9L. Proc. Natl Acad. Sci. USA 119, e2116550119 (2022).
Article CAS PubMed PubMed Central Google Scholar
Lemos de Matos, A., Liu, J., McFadden, G. & Esteves, P. J. Evolution and divergence of the mammalian SAMD9/SAMD9L gene family. BMC Evol. Biol. 13, 121 (2013).
Article CAS PubMed PubMed Central Google Scholar
Peeters, M. et al. Risk to human health from a plethora of simian immunodeficiency viruses in primate bushmeat. Emerg. Infect. Dis. 8, 451–457 (2002).
Article PubMed PubMed Central Google Scholar
Chahroudi, A., Bosinger, S. E., Vanderford, T. H., Paiardini, M. & Silvestri, G. Natural SIV hosts: showing AIDS the door. Science 335, 1188–1193 (2012).
Article CAS PubMed Google Scholar
Li, Y. et al. Eastern chimpanzees, but not bonobos, represent a simian immunodeficiency virus reservoir. J. Virol. 86, 10776–10791 (2012).
Article CAS PubMed PubMed Central Google Scholar
Van Dooren, S. et al. Lack of evidence for infection with simian immunodeficiency virus in bonobos. AIDS Res. Hum. Retrovir. 18, 213–216 (2002).
Article PubMed Google Scholar
Keele, B. F. et al. Chimpanzee reservoirs of pandemic and nonpandemic HIV-1. Science 313, 523–526 (2006).
Article CAS PubMed PubMed Central Google Scholar
Neil, S. J. D., Zang, T. & Bieniasz, P. D. Tetherin inhibits retrovirus release and is antagonized by HIV-1 Vpu. Nature 451, 425–430 (2008).
Article CAS PubMed Google Scholar
Sawyer, S. L., Wu, L. I., Emerman, M. & Malik, H. S. Positive selection of primate TRIM5α identifies a critical species-specific retroviral restriction domain. Proc. Natl Acad. Sci. USA 102, 2832–2837 (2005).
Article CAS PubMed PubMed Central Google Scholar
Duggal, N. K. & Emerman, M. Evolutionary conflicts between viruses and restriction factors shape immunity. Nat. Rev. Immunol. 12, 687–695 (2012).
Article CAS PubMed PubMed Central Google Scholar
Etienne, L. et al. The role of the antiviral APOBEC3 gene family in protecting chimpanzees against lentiviruses from monkeys. PLoS Pathog. 11, e1005149 (2015).
Article PubMed PubMed Central Google Scholar
Compton, A. A., Malik, H. S. & Emerman, M. Host gene evolution traces the evolutionary history of ancient primate lentiviruses. Philos. Trans. R. Soc. B 368, 20120496 (2013).
Article Google Scholar
van Kempen, M. et al. Fast and accurate protein structure search with Foldseek. Nat. Biotechnol. 42, 243–246 (2024).
Article PubMed Google Scholar
Tesson, F. et al. A comprehensive resource for exploring antiphage defense: DefenseFinder Webservice, Wiki and Databases. Peer Community J. 4, e91 (2024).
Article Google Scholar
Wang, Y., Tian, Y., Yang, X., Yu, F. & Zheng, J. Filamentation activates bacterial Avs5 antiviral protein. Nat. Commun. 16, 2408 (2025).
Article CAS PubMed PubMed Central Google Scholar
Béchon, N. et al. Diversification of molecular pattern recognition in bacterial NLR-like proteins. Nat. Commun. 15, 9860 (2024).
Article PubMed PubMed Central Google Scholar
Gao, L. A. et al. Prokaryotic innate immunity through pattern recognition of conserved viral proteins. Science 377, eabm4096 (2022).
Article CAS PubMed PubMed Central Google Scholar
Gao, L. et al. Diverse enzymatic activities mediate antiviral immunity in prokaryotes. Science 369, 1077–1084 (2020).
Article CAS PubMed PubMed Central Google Scholar
Rousset, F. et al. A conserved family of immune effectors cleaves cellular ATP upon viral infection. Cell 186, 3619–3631 (2023).
Article CAS PubMed Google Scholar
Garb, J. et al. Multiple phage resistance systems inhibit infection via SIR2-dependent NAD+ depletion. Nat. Microbiol. 7, 1849–1856 (2022).
Article CAS PubMed Google Scholar
Schultz, J., Ponting, C. P., Hofmann, K. & Bork, P. SAM as a protein interaction domain involved in developmental regulation. Protein Sci. Publ. Protein Soc. 6, 249–253 (1997).
Article CAS Google Scholar
Xu, J. & Zhang, Y. How significant is a protein structure similarity with TM-score = 0.5? Bioinformatics 26, 889–895 (2010).
Article CAS PubMed PubMed Central Google Scholar
Leipe, D. D., Koonin, E. V. & Aravind, L. STAND, a class of P-loop NTPases including animal and plant regulators of programmed cell death: multiple, complex domain architectures, unusual phyletic patterns, and evolution by horizontal gene transfer. J. Mol. Biol. 343, 1–28 (2004).
Article CAS PubMed Google Scholar
Loquet, A. & Saupe, S. J. Diversity of amyloid motifs in NLR signaling in fungi. Biomolecules 7, 38 (2017).
Article PubMed PubMed Central Google Scholar
Mordier, J., Fraisse, M., Cohen-Tannoudji, M. & Molaro, A. Recurrent evolutionary innovations in rodent and primate Schlafen genes. Preprint at bioRxiv https://doi.org/10.1101/2024.01.12.575368 (2024).
Doron, S. et al. Systematic discovery of anti-phage defense systems in the microbial pan-genome. Science 359, eaar4120 (2018).
Article PubMed PubMed Central Google Scholar
Valenzuela, C., Saucedo, S. & Llano, M. Schlafen14 impairs HIV-1 expression in a codon usage-dependent manner. Viruses 16, 502 (2024).
Article CAS PubMed PubMed Central Google Scholar
Kobayashi-Ishihara, M. et al. Schlafen 12 restricts HIV-1 latency reversal by a codon-usage dependent post-transcriptional block in CD4+ T cells. Commun. Biol. 6, 487 (2023).
Metzner, F. J., Huber, E., Hopfner, K.-P. & Lammens, K. Structural and biochemical characterization of human Schlafen 5. Nucleic Acids Res. 50, 1147–1161 (2022).
Article CAS PubMed PubMed Central Google Scholar
Garvie, C. W. et al. Structure of PDE3A-SLFN12 complex reveals requirements for activation of SLFN12 RNase. Nat. Commun. 12, 4375 (2021).
Article CAS PubMed PubMed Central Google Scholar
Yang, J.-Y. et al. Structure of Schlafen13 reveals a new class of tRNA/rRNA-targeting RNase engaged in translational control. Nat. Commun. 9, 1165 (2018).
Article PubMed PubMed Central Google Scholar
Li, M. et al. DNA damage-induced cell death relies on SLFN11-dependent cleavage of distinct type II tRNAs. Nat. Struct. Mol. Biol. 25, 1047–1058 (2018).
Article PubMed PubMed Central Google Scholar
Picard, L. et al. DGINN, an automated and highly-flexible pipeline for the detection of genetic innovations on protein-coding genes. Nucleic Acids Res. 48, e103 (2020).
Article CAS PubMed PubMed Central Google Scholar
Wellenreuther, M., Mérot, C., Berdan, E. & Bernatchez, L. Going beyond SNPs: the role of structural genomic variants in adaptive evolution and species diversification. Mol. Ecol. 28, 1203–1209 (2019).
Article PubMed Google Scholar
Radke, D. W. & Lee, C. Adaptive potential of genomic structural variation in human and mammalian evolution. Brief. Funct. Genomics 14, 358–368 (2015).
Article CAS PubMed PubMed Central Google Scholar
Daugherty, M. D. & Malik, H. S. Rules of engagement: molecular insights from host–virus arms races. Annu. Rev. Genet. 46, 677–700 (2012).
Article CAS PubMed Google Scholar
Smith, M. D. et al. Less is more: an adaptive branch-site random effects model for efficient detection of episodic diversifying selection. Mol. Biol. Evol. 32, 1342–1353 (2015).
Article CAS PubMed PubMed Central Google Scholar
Prüfer, K. et al. The bonobo genome compared with the chimpanzee and human genomes. Nature 486, 527–531 (2012).
Article PubMed PubMed Central Google Scholar
Mao, Y. et al. A high-quality bonobo genome refines the analysis of hominid evolution. Nature 594, 77–81 (2021).
Article CAS PubMed PubMed Central Google Scholar
Prado-Martinez, J. et al. Great ape genetic diversity and population history. Nature 499, 471–475 (2013).
Article CAS PubMed PubMed Central Google Scholar
Allenspach, E. J. et al. Germline SAMD9L truncation variants trigger global translational repression. J. Exp. Med. 218, e20201195 (2021).
Article CAS PubMed PubMed Central Google Scholar
Thomas, M. E. et al. Pediatric MDS and bone marrow failure-associated germline mutations in SAMD9 and SAMD9L impair multiple pathways in primary hematopoietic cells. Leukemia 35, 3232–3244 (2021).
Article CAS PubMed PubMed Central Google Scholar
Locatelli, S. et al. Why are Nigeria-Cameroon chimpanzees (Pan troglodytes ellioti) free of SIVcpz infection? PLoS ONE 11, e0160788 (2016).
Article PubMed PubMed Central Google Scholar
Gao, F. et al. Origin of HIV-1 in the chimpanzee Pan troglodytes troglodytes. Nature 397, 436–441 (1999).
Article CAS PubMed Google Scholar
Barbian, H. J. et al. Neutralization properties of simian immunodeficiency viruses infecting chimpanzees and gorillas. mBio 6, e00296-15 (2015).
Article PubMed PubMed Central Google Scholar
Burroughs, A. M. & Aravind, L. Identification of uncharacterized components of prokaryotic immune systems and their diverse eukaryotic reformulations. J. Bacteriol. https://doi.org/10.1128/jb.00365-20 (2020).
Sawyer, S. L., Emerman, M. & Malik, H. S. Ancient adaptive evolution of the primate antiviral DNA-editing enzyme APOBEC3G. PLoS Biol. 2, e275 (2004).
Article PubMed PubMed Central Google Scholar
Münk, C., Willemsen, A. & Bravo, I. G. An ancient history of gene duplications, fusions and losses in the evolution of APOBEC3 mutators in mammals. BMC Evol. Biol. 12, 71 (2012).
Article PubMed PubMed Central Google Scholar
Ito, J., Gifford, R. J. & Sato, K. Retroviruses drive the rapid evolution of mammalian APOBEC3 genes. Proc. Natl Acad. Sci. USA 117, 610–618 (2020).
Article CAS PubMed Google Scholar
Kim, C. A. & Bowie, J. U. SAM domains: uniform structure, diversity of function. Trends Biochem. Sci. 28, 625–628 (2003).
Article CAS PubMed Google Scholar
Knight, M. J., Leettola, C., Gingery, M., Li, H. & Bowie, J. U. A human sterile alpha motif domain polymerizome. Protein Sci. 20, 1697–1706 (2011).
Article CAS PubMed PubMed Central Google Scholar
Johnson, B. et al. Human IFIT3 modulates IFIT1 RNA binding specificity and protein stability. Immunity 48, 487–499 (2018).
Article CAS PubMed PubMed Central Google Scholar
Fleith, R. C. et al. IFIT3 and IFIT2/3 promote IFIT1-mediated translation inhibition by enhancing binding to non-self RNA. Nucleic Acids Res. 46, 5269–5285 (2018).
Article CAS PubMed PubMed Central Google Scholar
Sahoo, S. S. et al. Clinical evolution, genetic landscape and trajectories of clonal hematopoiesis in SAMD9/SAMD9L syndromes. Nat. Med. 27, 1806–1817 (2021).
Article CAS PubMed PubMed Central Google Scholar
Keele, B. F. et al. Increased mortality and AIDS-like immunopathology in wild chimpanzees infected with SIVcpz. Nature 460, 515–519 (2009).
Article CAS PubMed PubMed Central Google Scholar
Etienne, L. et al. Characterization of a new simian immunodeficiency virus strain in a naturally infected Pan troglodytes troglodytes chimpanzee with AIDS related symptoms. Retrovirology 8, 4 (2011).
Article CAS PubMed PubMed Central Google Scholar
Greenwood, E. J. D. et al. Simian immunodeficiency virus infection of chimpanzees (Pan troglodytes) shares features of both pathogenic and non-pathogenic lentiviral infections. PLoS Pathog. 11, e1005146 (2015).
Article PubMed PubMed Central Google Scholar
Pawar, H., Ostridge, H. J., Schmidt, J. M. & Andrés, A. M. Genetic adaptations to SIV across chimpanzee populations. PLoS Genet. 18, e1010337 (2022).
Article CAS PubMed PubMed Central Google Scholar
Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinf. 10, 421 (2009).
Article Google Scholar
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).
Article CAS PubMed Google Scholar
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Article CAS PubMed PubMed Central Google Scholar
Löytynoja, A. Phylogeny-aware alignment with PRANK. Methods Mol. Biol. 1079, 155–170 (2014).
Article PubMed Google Scholar
Guindon, S. et al. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst. Biol. 59, 307–321 (2010).
Article CAS PubMed Google Scholar
Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004).
Article CAS PubMed PubMed Central Google Scholar
Trifinopoulos, J., Nguyen, L.-T., von Haeseler, A. & Minh, B. Q. W-IQ-TREE: a fast online phylogenetic tool for maximum likelihood analysis. Nucleic Acids Res. 44, W232–W235 (2016).
Article CAS PubMed PubMed Central Google Scholar
Bergström, A. et al. Insights into human genetic variation and population history from 929 diverse genomes. Science 367, eaay5012 (2020).
Article PubMed PubMed Central Google Scholar
Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. Preprint at https://doi.org/10.48550/arXiv.1303.3997 (2013).
Garrison, E. & Marth, G. Haplotype-based variant detection from short-read sequencing. Preprint at https://doi.org/10.48550/arXiv.1207.3907 (2012).
Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021).
Article CAS PubMed PubMed Central Google Scholar
Varadi, M. et al. AlphaFold protein structure database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res. 50, D439–D444 (2022).
Article CAS PubMed Google Scholar
Berman, H. M. et al. The protein data bank. Nucleic Acids Res. 28, 235–242 (2000).
Article CAS PubMed PubMed Central Google Scholar
Gilchrist, C. L. M., Mirdita, M. & Steinegger, M. Multiple protein structure alignment at scale with FoldMason. Preprint at bioRxiv https://doi.org/10.1101/2024.08.01.606130 (2024).
Minh, B. Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol. 37, 1530–1534 (2020).
Article CAS PubMed PubMed Central Google Scholar
Yu, G., Smith, D. K., Zhu, H., Guan, Y. & Lam, T. T.-Y. ggtree: an R package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods Ecol. Evol. 8, 28–36 (2017).
Article Google Scholar
Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res. 52, W78–W82 (2024).
Article PubMed PubMed Central Google Scholar
Paysan-Lafosse, T. et al. The Pfam protein families database: embracing AI/ML. Nucleic Acids Res. 53, D523–D534 (2025).
Article PubMed Google Scholar
Steinegger, M. & Söding, J. MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nat. Biotechnol. 35, 1026–1028 (2017).
Article CAS PubMed Google Scholar
Richter, D. J. et al. EukProt: a database of genome-scale predicted proteins across the diversity of eukaryotes. Peer Community J. 2, evae231 (2022).
Article Google Scholar
Eddy, S. R. Accelerated profile HMM searches. PLoS Comput. Biol. 7, e1002195 (2011).
Article CAS PubMed PubMed Central Google Scholar
Sievers, F. et al. Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol. Syst. Biol. 7, 539 (2011).
Article PubMed PubMed Central Google Scholar
Steenwyk, J. L., Iii, T. J. B., Li, Y., Shen, X.-X. & Rokas, A. ClipKIT: a multiple sequence alignment trimming software for accurate phylogenomic inference. PLoS Biol. 18, e3001007 (2020).
Article CAS PubMed PubMed Central Google Scholar
Lee, T. S. et al. BglBrick vectors and datasheets: a synthetic biology platform for gene expression. J. Biol. Eng. 5, 12 (2011).
Article CAS PubMed PubMed Central Google Scholar
Xia, Y. et al. T5 exonuclease-dependent assembly offers a low-cost method for efficient cloning and site-directed mutagenesis. Nucleic Acids Res. 47, e15 (2019).
Article PubMed Google Scholar
Doyle, T. et al. The interferon-inducible isoform of NCOA7 inhibits endosome-mediated viral entry. Nat. Microbiol. 3, 1369–1376 (2018).
Article CAS PubMed PubMed Central Google Scholar
Robert, X. & Gouet, P. Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res. 42, W320–W324 (2014).
Article CAS PubMed PubMed Central Google Scholar
Okonechnikov, K., Golosova, O., Fursov, M. & UGENE team Unipro UGENE: a unified bioinformatics toolkit. Bioinformatics 28, 1166–1167 (2012).
Article CAS PubMed Google Scholar
Etienne, L. Alignments and trees associated with Fig. 1 and S1: structural homology analyses show strong similarity between SAMD9 gene family and Avs antiphage proteins. figshare https://doi.org/10.6084/m9.figshare.29082653.v2 (2025).
Etienne, L. Alignments and trees associated with Fig. 2 and S2: Avs9 is part of prokaryotic defense systems and induces cell death in bacteria, through its SAMD9/SLFN-analogous active site in the AlbA_2 domain. AlbA_2 effector domain is shared between human SLFNs, SAMD9s and bacterial Avs9, with a conserved catalytic site. figshare https://doi.org/10.6084/m9.figshare.29082710.v1 (2025).
Etienne, L. Alignments and trees associated with Fig. 3 and S3: SAMD9/9L gene tree in mammals (with some other vertebrates as outgroups). figshare https://doi.org/10.6084/m9.figshare.29042762.v1 (2025).
Etienne, L. Alignments and trees associated with Fig. 4 and S4: primate phylogenetic analyses of SAMD9 and SAMD9L. Codon alignments used as inputs. figshare https://doi.org/10.6084/m9.figshare.29087987.v1 (2025).
Etienne, L. Alignments associated with S5c: SAMD9L primate amino acid sequence alignment showing the bonobo-specific variants. figshare https://doi.org/10.6084/m9.figshare.29088050.v1 (2025).
Etienne, L. Avs9 gene fragment sequence synthesized and used in this study, associated with Fig 3 and S3. figshare https://doi.org/10.6084/m9.figshare.29082719.v1 (2025).

Download references

Acknowledgements

We thank F. Fiorini (MMSB), F. Roucher-Boulez (HCL), A. Bernheim and H. Vaysset (Pasteur Institute) for feedback on the project, LP2L past and present laboratory members for support and discussion, L. Guéguen (LBBE) for providing guidance in the positive selection analyses in DGINN. We thank A. Le Corf (CIRI), C. Goujon (IRIM), L. Guiguettaz and E. Ricci (LBMC), B. Hahn (University of Pennsylvania, Perelman School of Medicine), M. Emerman (Fred Hutch) and the NIH AIDS reagent programme for sharing reagents. We acknowledge the contribution of SFR Biosciences (Université Claude Bernard Lyon 1, CNRS UAR3444, Inserm US8, ENS de Lyon) ANIRA platform, especially V. Barateau for her assistance, as well as D. Decimo of the BSL-3 ENS-Lyon platform. We also thank all the contributors of the publicly available bioinformatic programmes and genomic sequences. This work was funded by grants from: the French Research Agency on HIV and Emerging Infectious Diseases ANRS/MIE (grant nos ECTZ19143 and ECTZ245897) to L.E., the Sidaction (grant no. 23-1-AEQ-13601 to L.E. and 2020—grant no. 12673 and 2023—grant no. 13574 PhD fellowships to A.L.), a 2021 France-Berkeley Fund Award to L.E. and PHS, Agence Nationale de la Recherche under France 2030 bearing the reference ANR-24-RRII-0005 on funds administered by Inserm (EvoCure) to L.E. and F.R., Institute of General Medical Sciences (grant no. R35GM142916) to P.H.S., the Vallee Scholars Award to P.H.S., the Weill Neurohub Award to P.H.S. L.E. and P.H.S. are further supported by CNRS International Research Project IRP RAPIDvBAT. L.E. and A.C. are supported by the CNRS. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author information

These authors contributed equally: Rémi Demeure, Amandine Chantharath.

Authors and Affiliations

Centre International de Recherche en Infectiologie (CIRI), Inserm U1111, UCBL1, CNRS UMR5308, ENS de Lyon, Université de Lyon, Lyon, France
Alexandre Legrand, Rémi Demeure, Amandine Chantharath, Carine Rey, Julie Baltenneck, Clara Loyer, Léa Picard, Andrea Cimarelli, Francois Rousset & Lucie Etienne
School of Biological Sciences, Seoul National University, Seoul, South Korea
Cameron L. M. Gilchrist & Martin Steinegger
Department of Integrative Biology, University of California, Berkeley, CA, USA
Joana L. Rocha & Peter H. Sudmant

Authors

Alexandre Legrand
View author publications
Search author on:PubMed Google Scholar
Rémi Demeure
View author publications
Search author on:PubMed Google Scholar
Amandine Chantharath
View author publications
Search author on:PubMed Google Scholar
Carine Rey
View author publications
Search author on:PubMed Google Scholar
Julie Baltenneck
View author publications
Search author on:PubMed Google Scholar
Cameron L. M. Gilchrist
View author publications
Search author on:PubMed Google Scholar
Joana L. Rocha
View author publications
Search author on:PubMed Google Scholar
Clara Loyer
View author publications
Search author on:PubMed Google Scholar
Léa Picard
View author publications
Search author on:PubMed Google Scholar
Andrea Cimarelli
View author publications
Search author on:PubMed Google Scholar
Martin Steinegger
View author publications
Search author on:PubMed Google Scholar
Francois Rousset
View author publications
Search author on:PubMed Google Scholar
Peter H. Sudmant
View author publications
Search author on:PubMed Google Scholar
Lucie Etienne
View author publications
Search author on:PubMed Google Scholar

Contributions

A.L. and L.E. conceived the project. L.E. supervised the project. A.L., R.D., A. Chantharath, C.L.M.G., M.S., F.R., P.H.S. and L.E. participated in the development or design of the methodology. A.L. and C.L.M.G. performed Foldseek analyses and downstream phylogenetics, with supervision from L.E. and M.S. A.L. and F.R. performed phylogenetic analyses across the tree of life. J.B. and F.R. performed experiments in bacteria. A. Chantharath and L.E. performed phylogenetics, genomics and positive selection analyses on mammals and primates, with help from L.P. who performed preliminary analyses. J.L.R. generated geographic maps and provided support for the population genetic and genomic analyses. P.H.S. performed the genomic locus alignment and supervised A.L. in genomic analyses in Pan individuals. C.R. generated the initial script for polymorphism analyses and supported A.L. in biocomputational analyses. C.L. provided technical support for molecular biology assays. A.L., R.D. and L.E. performed lentiviral infection experiments. A.L. performed translational shutoff assay, with help from A. Chantharath. A.L., A. Chantharath, R.D., C.R., C.L.M.G., A. Cimarelli, M.S., F.R., P.H.S., L.E. provided resources for this work. A.L., A. Chantharath and L.E. wrote the original draft of the manuscript, with help from R.D., J.L.R., C.R. and F.R. for the visualization. A.L., A. Chantharath, R.D., C.R., J.B., C.L.M.G., J.L.R., C.L., L.P., A. Cimarelli, M.S., F.R., P.H.S. and L.E. reviewed and edited the manuscript.

Corresponding author

Correspondence to Lucie Etienne.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Ecology & Evolution thanks the anonymous reviewers for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Supplementary Figs. 1–7.

Reporting Summary (download PDF )

Supplementary Tables (download XLSX )

Supplementary Tables 1–3.

Supplementary Data (download PDF )

Flow cytometry gating strategy.

Source data

Source Data (download XLSX )

Source data and exact P values for Figs. 5 and 6.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, and provide a link to the Creative Commons license. You do not have permission under this license to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Legrand, A., Demeure, R., Chantharath, A. et al. Evolutionary characterization of antiviral SAMD9/9L across kingdoms supports ancient convergence and lineage-specific adaptations. Nat Ecol Evol 9, 2206–2222 (2025). https://doi.org/10.1038/s41559-025-02845-x

Download citation

Received: 15 September 2024
Accepted: 28 July 2025
Published: 11 November 2025
Version of record: 11 November 2025
Issue date: December 2025
DOI: https://doi.org/10.1038/s41559-025-02845-x