The genome of human immunodeficiency virus (HIV) has an average nucleotide composition strongly biased as compared to the human genome. The consequence of such nucleotide composition on HIV pathogenicity has not been investigated yet. To address this question, we analyzed the role of nucleotide bias of HIV-derived nucleic acids in stimulating type-I interferon response in vitro. We found that the biased nucleotide composition of HIV is detected in human cells as compared to humanized sequences, and triggers a strong innate immune response, suggesting the existence of cellular immune mechanisms able to discriminate RNA sequences according to their nucleotide composition or to detect specific secondary structures or linear motifs within biased RNA sequences. We then extended our analysis to the entire genome scale by testing more than 1300 HIV-1 complete genomes to look for an association between nucleotide composition of HIV-1 group M subtypes and their pathogenicity. We found that subtype D, which has an increased pathogenicity compared to the other subtypes, has the most divergent nucleotide composition relative to the human genome. These data support the hypothesis that the biased nucleotide composition of HIV-1 may be related to its pathogenicity.
Citation: Vabret N, Bailly-Bechet M, Najburg V, Müller-Trutwin M, Verrier B, Tangy F (2012) The Biased Nucleotide Composition of HIV-1 Triggers Type I Interferon Response and Correlates with Subtype D Increased Pathogenicity. PLoS ONE7(4): e33502. https://doi.org/10.1371/journal.pone.0033502
Editor: Fabrizio Mammano, Inserm, France
Received: August 26, 2011; Accepted: February 14, 2012; Published: April 18, 2012
Copyright: © 2012 Vabret et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the Centre National de la Recherche Scientifique (PEPS-60593a), the Agence Nationale de la Recherche contre le SIDA (ANRS-2010-311), and the European Commission (Europrise, LSHP CT-2006-037611). Dr. Vabret is a fellow from Ecole Normale Supérieure (ENS), Lyon. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Without treatment, most individuals infected with HIV will develop AIDS as a consequence of years of chronic and excessive activation of their immune system . The generalized immune activation in AIDS consists in over expression of activation markers, excessive proliferation and apoptosis of CD4 and CD8 T cells, NK cells, B cells and macrophages, in addition to a strong type I interferon response, which is maintained throughout the chronic phase of infection . Because the virus replicates preferentially in activated CD4 T cells, chronic activation supports viral replication by providing targets. A number of mechanisms have been suggested to contribute to immune activation in AIDS pathogenesis , , such as i) the innate and adaptive response to HIV and its antigens, ii) the direct effect of viral proteins on cell receptors like Env-CD4/CCR5 or TCR down-regulation by Nef, iii) the activation of TLRs by microbial products leaking from the damaged gut mucosa, iv) the frequency of opportunistic infections, v) increased levels of proinflammatory and pro-apoptotic cytokines, vi) depletion of CD4 T regulatory cells.
Viral nucleic acids trigger an antiviral innate immune response in infected cells, which contributes to activating the immune system , , . Cytoplasmic accumulation of HIV nucleic acids has been shown to be responsible for the majority of T-cell apoptosis during HIV infection . IFN-α/β is the principal mediator of antiviral innate immunity. Different forms of viral RNA molecules can trigger IFN-α/β stimulation, either single or double stranded, naked or combined with capsid proteins , , , . Several cytosolic receptors have been identified that recognize HIV-1 patterns , . Plasmacytoid dendritic cells (pDCs) produce large amounts of IFN-α/β in response to microbial stimuli. During the acute phase of HIV infection, pDCs are the main producers of IFN-α/β. They recognize viral ssRNA via TLR7, and dsDNA via TLR9 , . This antiviral response contributes to containing the acute infection. In primate models of lentiviral infection, pDCs produce large amounts of IFN-α/β during the acute phase of infection, whatever a pathogenic or non-pathogenic outcome , . However, during the chronic phase of infection, viral RNA, which is still produced in infected cells, stimulates IFN-α/β production on the long term. A recent study reported that in chronically HIV-infected patients IFN-α colocalized only with few pDCs, but rather with other TLR7 negative cells .
The genomes of HIV and most lentiviruses present a particularly biased nucleotide composition compared to that of their primate hosts, with as much as 35% adenosine in HIV-1 RNA genome. As a consequence, the average amino acid composition is different and the synonymous codon usage is also modified , . Lentiviral biased nucleotide composition has been explained by dNTP pool imbalance during reverse transcription ,  and by the antiviral activity of the cellular Apobec 3G (A3G) cytidine deaminase, which mutates G to A in HIV provirus  thus impacting HIV-1 evolution and drug resistance , , although the counteracting activity of the viral Vif protein reduces Apobec impact on the nucleotide bias of HIV genome .
Such biased nucleotide composition obviously has direct consequences on genome structure and stability, but its effects on viral biology have not been investigated yet. We wondered whether the strongly divergent nucleotide composition of HIV-1 nucleic acids, as compared to host, plays a role in triggering antiviral innate response. To address this question, we compared the capacity of HIV-1 RNA molecules corresponding to the three major viral genes, either wild type or artificially unbiased, to modulate IFN-α/β synthesis in a sensitive in vitro assay in human cells. We also tested a set of small RNA fragments covering the entire HIV-1 genome for their capacity to activate IFN-α/β response according to their nucleotide composition. We then performed a large-scale bioinformatics study to compare the divergence in nucleotide composition between all subtypes of HIV-1 group M viruses. Altogether, this work shows that AU-rich biased HIV-1 sequences, mainly located within gag, pol, env genes, strongly stimulate IFN-α/β response, and that HIV-1 subtype D, which has an increased pathogenicity, has a more biased sequence.
The Biased Nucleotide Composition of HIV-1 RNA Stimulates IFN-α/β Synthesis
To first analyze the divergence between HIV and human nucleotide composition, we plotted the distribution of human coding sequences according to their length and nucleotide divergence, as measured by the Chi-square distance relative to average nucleotide A/C/G/T frequency of all human coding sequences (Figure 1). This analysis shows that the majority of human coding sequences are gathered in a group of relatively small size sequences with homogenous nucleotide composition. Comparing in the same analysis HIV-1 nucleotide composition to the human genome highlights that gag, pol, env and vpu genes are highly divergent from human sequences (black dots in Figure 1). To investigate the role of such biased nucleotide composition of HIV genome on immune activation, we analyzed the ability of HIV-1-derived RNA fragments, whose nucleotide sequence was either wild type (WT) or artificially optimized to human codon usage, to induce IFN-α/β synthesis in vitro in epithelial-like HEK 293 T cells (ATCC, CRL-11268). We chose these cells because they are easily transfectable and produce IFN-α/β in response to HIV-1 infected cells although they lack TLR7 . The three major HIV-1 genes gag, pol, env (gag, pol from hxb2 sequence and env from a primary isolate sequence) and their corresponding versions optimized towards human codon bias (Gag opt, Pol opt, Env opt) were in vitro transcribed with T7 RNA polymerase into uncapped and unpolyadenylated RNA fragments. These RNA fragments were co-transfected into HEK 293 T cells together with a reporter plasmid containing the luciferase gene fused downstream of five interferon-stimulated response elements (ISRE-luciferase). A low amount of a plasmid harboring a thymidine kinase (Tk) promoter just upstream of the renilla luciferase gene was also co-transfected to allow normalization between wells. In a first set of experiments, we performed a dose-response analysis for each RNA molecule tested to determine the highest non-toxic dose that did not affect the level of renilla expression. A unique dose of 20 ng that was not toxic and did not affect, or only moderately, transfection efficiency was chosen for transfecting all RNA fragments (Figure S1). The reporter activity, reflecting the induction of IFN-α/β response, was calculated as the ratio of firefly luciferase activity to reference renilla luciferase activity measured in cell lysates 20 hours after transfection (Figure 2A). While the WT versions of Gag, Pol or Env RNAs induced a strong IFN-α/β expression, this capacity was abolished in their humanized versions, which induced very low levels of IFN-α/β. Although both type of sequences shared the same coding information, transfecting uncapped and unpolyadenylated RNA molecules in this experiment allowed evaluating the role of RNA molecules in the absence of expressed proteins. Moreover, to ensure that no protein is expressed from RNA, we deleted the first nucleotide of the AUG codon in each construct. The nucleotide composition of wild type and humanized HIV-1 RNA fragments used in this experiment reveals that humanization results essentially in reducing A/U and increasing G/C contents (Figure 2B). Our observation demonstrates that the biased nucleotide composition of HIV synthetic RNA molecules, or its structural consequence, is detected in human cells and triggers a strong innate immune response.
The distribution of human coding sequences is represented according to their length (log scale) and nucleotidic divergence, as measured by the Chi-square distance relative to average composition of all human coding sequences. The colored background shows the density of human coding sequences as indicated on the z-scale. Black dots represent either single HIV-1 genes or the entire HIV-1 genome. Note the distinguishable positions of the gag, pol, env and vpu genes.
(A) Type-I interferon stimulation was measured as the ISRE-dependant expression of luciferase activity in HEK 293 T cells upon transient transfection of HIV-1-derived RNA molecules. The ISRE reporter activity is the ratio of firefly luciferase activity to control renilla luciferase activity in triplicate experiments. In vitro transcribed RNA molecules corresponding to the wild type sequence of the three HIV-1 major genes induced a strong response while their humanized version induced a tenfold lower response. (B) Nucleotide composition of HIV-1 RNA fragments expressed as the A/C/G/U frequencies.
A-rich Regions of HIV-1 Genome are Responsible for IFN-α/β Stimulation
To investigate whether shorter specific regions within HIV-1 genome are responsible for IFN-α/β stimulation, we first looked for hyper-biased region on the whole genome. For that, we computed the Chi-square distance between the A/C/G/U frequencies of a sliding window 500 nt wide along the genome of HIV-1 clone hxb2 and the corresponding frequencies of the entire coding sequences of the human genome (Figure 3A). As expected from Figure 1, we found the most divergent regions located within the three large viral genes (gag, pol, env), while overlapping coding regions and cis-active regulatory sequences are unbiased. Since adenosine enrichment contributes to most of the divergence (red line in Figure 3A), these “hot spots” of divergence are likely the footprint of adenosine accumulation during low-fidelity reverse transcription and by the action of cytidine deaminases. This highlights that the long open reading frames (ORF) of lentiviruses (gag, pol, env) can accumulate more synonymous mutations than the rest of the genome without impairing viral fitness. They reflect the tendency of lentiviral genomes to become A-rich through evolution .
A) The black line shows the nucleotide divergence of HIV-1 compared to human genome, as measured by a Chi-square distance in a 500 bp sliding window along the HIV-1 hxb2 genome. The individual contributions of A, C, G, and U nucleotides to this divergence are shown respectively in red, blue, yellow and green. The HIV-1 genome map is schematically represented above the figure. B) Histogram represents the ISRE reporter activity as the ratio of firefly luciferase activity to control renilla luciferase activity determined in HEK 293 T cells cotransfected with HIV-1 RNA fragments and pISRE-Luc reporter. C) Correlation between the nucleotide divergence (x-axis) and the luciferase activation (y-axis) of the 38 RNA fragments from HIV-1 hxb2 genome. Each point corresponds to the average of four replicates and the correlation coefficient was computed only on these averages to not artefactually lower the p-value.
To determine then the local distribution of immune activating sequences along the viral genome, we measured the ability of a set of RNA fragments covering the entire genome of HIV-1 clone hxb2 to induce IFN-α/β synthesis in vitro in the same system as used above. We generated 38 PCR fragments of approximately 500 bp, with overlaps of 250 bp, and covering the entire genome of HIV-1 hxb2 (primers used are listed in Table S1). These DNA fragments were in vitro transcribed with T7 RNA polymerase into a set of overlapping uncapped and unpolyadenylated RNA fragments. The quality and integrity of RNA fragments was controlled by migration on electrophoresis chips (Figure 4). These RNA fragments were co-transfected into HEK 293 T cells together with the reporter ISRE-luciferase plasmid, and luciferase activity was measured 20 h after transfection. Again, we performed a dose-response analysis for each RNA molecule to determine the highest non-toxic dose that did not affect the level of renilla expression. A unique dose of 12 ng that was not toxic and did not affect transfection efficiency was chosen for transfecting all RNA fragments (Figure S1). The ability of each fragment to induce an interferon response, calculated as the ratio of firefly luciferase activity to reference renilla luciferase activity (Figure 3B), was compared to its divergence in nucleotide composition with the human genome composition (Figure 3A, black line). The comparison highlights that RNA fragments that have the greatest divergence in nucleotide composition (mainly due to adenosine enrichment) are the most potent stimulators of IFN-α/β response (Figure 3C, p = 1.3 10−5, Pearson correlation test, R2 = 0,41). Although the p-value of this correlation is very strong, some fragments do not follow the global trend, particularly within the pol gene where 4 fragments located in regions of high nucleotide divergence appear to be low inducers of IFN-α/β response. Because the design of primers for the synthesis of RNA fragments was arbitrary (approximately every 500 bp), outliers may result from the elimination of either a secondary structure or a linear motif necessary for recognition in the host cell. Altogether, these results demonstrate that the most biased HIV RNA sequences are responsible for IFN-α/β stimulation.
. Overlapping RNA fragments covering the entire genome of HIV-1 hxb2. PCR fragments of approximately 500 bp, with overlaps of 250 bp, and covering the entire genome of HIV-1 hxb2 were in vitro transcribed with T7 RNA polymerase into a set of uncapped and unpolyadenylated RNA fragments. After purification RNA fragments were migrated on Agilent-2100 Small RNA chips. A ladder on the left indicates the size in nucleotides.
Subtype D is the Most Divergent Among HIV-1 Group M Viruses
To investigate the possible influence of the biased nucleotide composition of HIV RNA in AIDS pathogenesis, we looked for an association between nucleotide composition of HIV-1 group M subtypes and their pathogenicity. Different studies recently reported that among HIV-1 group M subtypes, infection with subtype D is associated with a significantly faster CD4 cell decline than other subtypes , , , , thus indicating an increased pathogenicity of subtype D viruses. Many reasons may account for this increased pathogenicity, such as virus replicative capacity or escape to immune responses. To look for an association with nucleotide bias, we measured the divergence of nucleotide composition between all HIV-1 subtypes and human genome by computing the Chi-square distance between the A/C/G/U frequencies of more than 1300 HIV-1 complete genomes and the entire coding sequences of human genome (Figure 5). All HIV-1 group M “pure” subtypes, for which more than 20 genomic sequences are available in Los Alamos database (clade A, B, C, D, F & G), were used in this analysis. We first evaluated the divergence between all subtypes by using ANOVA test. Looking blindly among the subtypes highlights that at least one subtype is significantly more biased and more divergent relative to human genome than the others (p = 2.10−6). We then used the Student t-test to compare subtype D divergence relatively to all other subtypes pooled together. It appears that subtype D is significantly more divergent to human genes than all other subtypes (Bonferroni-corrected Student t-test, p = 3.10−5). The observation that the most pathogenic HIV-1 subtype is also the most divergent in nucleotide composition with human genome is a preliminary indication of a possible contribution of HIV-specific nucleotide bias in AIDS pathogenesis. Detecting this effect through large-scale comparative genomic analysis indicates that the causative mechanism involved is disseminated along the HIV-1 sequence.
Boxplots show the median and dispersion of the nucleotide divergence of individual HIV-1 genomes from different subtypes to the average of human coding sequences. Red stars indicate that D subtype is significantly more divergent than other subtypes (p = 2×10−6 ANOVA test, p = 5.8×10−6 Student t-test, p = 2.8×10−7 Wilcoxon test). The random 10 kb column shows the median and standard deviation of the divergence between 10 kb of random human coding sequences and the average of all human coding sequences; 95% of the points have a divergence below 0.03, and less than 1% of the points are above 0.05, ie comparable to HIV divergence. Standard deviation gives a natural scale of variation to the graph. These data were not used in the statistical procedure to compare subtypes.
Very little is known about the consequences of the biased nucleotide composition of human immunodeficiency viruses on their pathogenicity. In this report we addressed this question by analyzing the role of nucleotide composition of HIV-1-derived nucleic acids in stimulating IFN-α/β response, then by correlating the pathogenicity of HIV-1 M subtypes with the divergence of their nucleotide composition with human coding sequences.
To measure type-I IFN stimulation, we used an in vitro assay allowing to discriminating the role of RNA nucleotide composition in the absence of viral proteins. Our experiments show that the biased nucleotide composition of HIV RNA is detected in human cells and triggers a strong innate immune response. We evidenced that HIV-1-derived RNA molecules that are the most biased in their nucleotide composition as compared to human coding sequences are the most potent IFN-α/β stimulators. They correspond to the three main viral genes, which are the most abundant viral RNA molecules in the cytoplasm of infected cells during HIV viral production. Optimization of these sequences to human codon bias, and therefore to human A/C/G/U frequency abolished their stimulating capacity. Similarly, discrete fragments of 500 bp within HIV-1 genome stimulate type-I IFN according to their nucleotide composition, the most biased being the best stimulators. These results indicate the existence of a molecular mechanism capable of sensing the biased nucleotide composition of RNA molecules or its structural consequence, or the presence of A-rich linear motifs, such as instability elements for example. Several receptors may discriminate nucleic acids according to their nucleotide composition. HIV encodes multiple U-rich sequences recognized by Toll-like receptor 7/8 . Cytosolic AT-rich DNA can be transcribed by RNA pol-III into RNA molecules that induce IFN-α/β through the RIG-I pathway . In the HCV genome, RIG-I preferentially binds poly-U/UC, poly-U/UG and poly-A RNAs . RIG-I is also able to trigger IFN-α/β response in the presence of genomic RNA extracted from HIV-1 particles . Other hypotheses can be formulated, such as the formation of RNA secondary structures specifically recognized by innate immunity sensors, or the recognition of specific AU-rich linear motifs. Finally, sequence specific RNAse such as RNAse L could also increase the innate immune response against AU-rich RNAs .
Whether this nucleotide bias is detected in infected natural target cells and contributes to activating immune system in HIV patients is unknown. Therefore, we looked for an association between nucleotide composition of HIV-1 subtypes and their pathogenicity. To investigate whether HIV-1 subtype nucleotide composition, and particularly its divergence with the human genome, is associated with differential pathogenicity, we combined information from numerous HIV-1 complete genomes and from large-scale epidemiological studies , , , . We observed that subtype D, which cause a faster CD4 cell decline, is significantly more divergent from the human genome than other subtypes. This observation supports the assumption that the higher capacity of biased HIV-1 nucleic acids to stimulate IFN-α/β synthesis may be linked with immune activation and pathogenesis. Although several other factors may determine the pathogenicity of a given HIV-1 subtype, such as replicative capacity, ability to escape adaptive immune response, restriction factors , , our data suggest that the strength of host innate immune response and its adaptation to viral nucleotide composition is critical.
To further extend this observation and determine whether the aggressive nucleotide composition of other lentiviruses than HIV-1 is linked to pathogenicity, it is of interest to look at non-human primate models for which SIV chronic viral infection is associated with either pathogenic or non-pathogenic outcome. Indeed, persistent infection by a highly replicating lentivirus does not necessarily lead to AIDS, and natural hosts such as African green monkeys (AGM), sooty mangabeys (SM) and mandrills (Mnd), never develop chronic immune activation and do not progress to AIDS, despite chronic high viremia  . Conversely, experimental transmission of the SM type of SIV (SIVsm) to rhesus macaques (RM), or of SIVagm.ver90 and SIVagm.sab92018 to pigtailed macaques causes immune over-activation and progression to AIDS . It may thus be hypothesized that interspecies transmission of SIV results in immune hyper-activation within new hosts, whereas natural hosts, in which the infection has long been established, tolerate the virus and control the homeostasis of their immune system. Analyzing the nucleotide divergence of different SIV strains with their primate hosts would be useful in this context. However, although numerous SIV complete genomes are available, only three primates complete genomes have been published (human, chimpanzee and rhesus macaque) and large-scale genomic sequences from natural SIV hosts primates, in which infection is non-pathogenic, are quite rare while they would provide a crucial input. We are currently working on such analyses.
In conclusion, our observations suggest for the first time that the specific nucleotide composition of HIV-1 genome might have a consequence on its pathogenicity. They suggest the existence of cellular immune mechanisms able to discriminate RNA sequences according to their nucleotide composition or to detect specific secondary structures or linear motifs within biased RNA sequences. These observations are valid at intra-genome and between-genome scales, indicating that the causative mechanism they involve is disseminated along the HIV-1 sequence, and therefore can only be detected through large-scale comparative genomic studies. Our observations also lead to readdress the antiviral role of RNA editing enzymes such as cytidine deaminases . In the case of A3G, increasing the A/U content of lentiviral genomes makes the viral RNAs more divergent. The resulting sequences, if still able to replicate, would activate the innate immune response to a higher extent. During co-evolution between a lentivirus and its host, this mechanism might help the immune system to reinforce the discrimination between self and non-self to preserve the organism, counterbalancing the humanization of viral genome due to nucleotidic pool constraints in the host cell , . Intriguingly, this mechanism might also contribute to maintaining or increasing the virus pathogenicity. Although further work is necessary to confirm this observation in non-human primate models of SIV infection, highlighting the role of lentiviral nucleotide composition may help in understanding lentiviral pathogenesis and in AIDS vaccine design.
Materials and Methods
Primers (Table S1) were designed to amplify 38 overlapping fragments of around 500 pb long by PCR reaction (Enzyme Phusion, 35 cycles, Ta = 60°C, 30 sec. elongation). Primers were also designed to amplify sequences (1 min. elongation) from Gag, Pol or Env from HIV-1 clone hxb2 or from plasmid expressing codon-optimized sequences of the genes. Gag optimized gene was purchased from Genecust, Pol and Env genes were purchased to Geneart. A 5′ tail containing the sequence of T7 promoter was included in every forward primer to allow the subsequent in vitro transcription reaction. PCR products were purified and used as template for T7 RNA synthesis according to manufacturer’s instructions (T7 RiboMA Express, Promega). Resulting RNA were purified using RNeasy mini kit (Qiagen) and concentration was determined by nanodrop measurement. RNA quality and integrity was controlled using Agilent-2100 Bioanalyzer prior to cell transfection.
Luciferase Reporter Gene Assay
Expression of IFN-α/β was determined by transient transfection of reporter plasmids pISRE-Luc containing five ISRE enhancer elements upstream of the firefly luciferase gene (Stratagene). HEK 293 T cells (ATCC, CRL11268) were plated in 24-well plates (2×105 per well). One day later, cells were transfected using 1 µl of Lipofectamine 2000 (Invitrogen) with pISRE-Luc reporter plasmid (250 ng/well), a plasmid harboring a thymidine kinase (Tk) promoter just upstream of the renilla luciferase gene (25 ng/well), and 12 ng of each RNA fragment (500 pb) or 20 ng of Gag, Pol or Env (hxb2 or optimized). After 20 h, cells were lysed, and the firefly and renilla luciferase activities were measured in cell lysates using the Dual-luciferase Reporter Assay System (Promega) according to manufacturer’s instructions. Reporter activity was calculated as a triplicate of the ratio of firefly luciferase activity to reference renilla luciferase activity.
Data and Statistical Analyses
Chi-square distances between the nucleotide frequencies of HIV-1 and the human genome were computed by:
The estimate of the nucleotidic composition of the human genome was computed on all coding sequences. For HIV-1, the entire genome of the hxb2 clone (HIVHXB2CG) was used for the frequencies estimation. All HIV-1 sequences from different subtypes (more than 1300) were downloaded from the HIV database (www.hiv.lanl.gov). Statistical procedure in the analysis of these data involved first an ANOVA to detect a difference between subtypes, followed by a Student t-test to compare subtype D relatively to all others subtypes pooled together. Bonferroni multiple test correction was applied. The correlation between Luciferase reporter activity and host/virus nucleotide divergence was computed using the nucleotide frequencies of each 500 bp RNA fragment using standard linear correlation test. Density of human genes of given length and nucleotidic divergence relative to the average of all human coding sequences (Fig. 1) was computed using a standard 2-dimensional gaussian kernel. All analyses were done using the R software.
Efficiency of HEK 293T cells transfection in presence of HIV-1 RNA fragments. RNA was co-transfected with a reporter ISRE-firefly luciferase plasmid and a plasmid harboring a thymidine kinase promoter upstream the renilla luciferase gene in triplicate experiments. As a control of transfection efficiency, the renilla luciferase activity was measured in cell lysates 20 hours after transfection. The different RNA fragments used are indicated.
We thank Patrick Wincker (CEA, Genoscope, Evry, France) for providing the nucleotide composition of EST sequences from AGM. We are very grateful to Alice Lepelley, Olivier Schwartz, Massimo Vergassola and Françoise Barre-Sinoussi (Institut Pasteur, Paris) for valuable discussions and critical reading of the manuscript.
Conceived and designed the experiments: NV MBB FT. Performed the experiments: NV MBB VN. Analyzed the data: NV MBB FT BV. Contributed reagents/materials/analysis tools: MMT BV. Wrote the paper: NV FT. Designed all softwares used in this work: MBB.
- 1. Grossman Z, Meier-Schellersheim M, Sousa AE, Victorino RM, Paul WE (2002) CD4+ T-cell depletion in HIV infection: are we closer to understanding the cause? Nat Med 8: 319–323.Z. GrossmanM. Meier-SchellersheimAE SousaRM VictorinoWE Paul2002CD4+ T-cell depletion in HIV infection: are we closer to understanding the cause?Nat Med8319323
- 2. Pandrea I, Sodora DL, Silvestri G, Apetrei C (2008) Into the wild: simian immunodeficiency virus (SIV) infection in natural hosts. Trends Immunol 29: 419–428.I. PandreaDL SodoraG. SilvestriC. Apetrei2008Into the wild: simian immunodeficiency virus (SIV) infection in natural hosts.Trends Immunol29419428
- 3. Sodora DL, Silvestri G (2008) Immune activation and AIDS pathogenesis. AIDS 22: 439–446.DL SodoraG. Silvestri2008Immune activation and AIDS pathogenesis.AIDS22439446
- 4. Sodora DL, Allan JS, Apetrei C, Brenchley JM, Douek DC, et al. (2009) Toward an AIDS vaccine: lessons from natural simian immunodeficiency virus infections of African nonhuman primate hosts. Nat Med 15: 861–865.DL SodoraJS AllanC. ApetreiJM BrenchleyDC Douek2009Toward an AIDS vaccine: lessons from natural simian immunodeficiency virus infections of African nonhuman primate hosts.Nat Med15861865
- 5. Jacquelin B, Mayau V, Targat B, Liovat AS, Kunkel D, et al. (2009) Nonpathogenic SIV infection of African green monkeys induces a strong but rapidly controlled type I IFN response. J Clin Invest 119: 3544–3555.B. JacquelinV. MayauB. TargatAS LiovatD. Kunkel2009Nonpathogenic SIV infection of African green monkeys induces a strong but rapidly controlled type I IFN response.J Clin Invest11935443555
- 6. Harris LD, Tabb B, Sodora DL, Paiardini M, Klatt NR, et al. (2010) Downregulation of robust acute type I interferon responses distinguishes nonpathogenic simian immunodeficiency virus (SIV) infection of natural hosts from pathogenic SIV infection of rhesus macaques. J Virol 84: 7886–7891.LD HarrisB. TabbDL SodoraM. PaiardiniNR Klatt2010Downregulation of robust acute type I interferon responses distinguishes nonpathogenic simian immunodeficiency virus (SIV) infection of natural hosts from pathogenic SIV infection of rhesus macaques.J Virol8478867891
- 7. Doitsh G, Cavrois M, Lassen KG, Zepeda O, Yang Z, et al. (2010) Abortive HIV Infection Mediates CD4 T Cell Depletion and Inflammation in Human Lymphoid Tissue. Cell 143: 789–801.G. DoitshM. CavroisKG LassenO. ZepedaZ. Yang2010Abortive HIV Infection Mediates CD4 T Cell Depletion and Inflammation in Human Lymphoid Tissue.Cell143789801
- 8. Yoneyama M, Fujita T (2010) Recognition of viral nucleic acids in innate immunity. Rev Med Virol 20: 4–22.M. YoneyamaT. Fujita2010Recognition of viral nucleic acids in innate immunity.Rev Med Virol20422
- 9. Diebold S (2010) Innate recognition of viruses. Immunol Lett 128: 17–20.S. Diebold2010Innate recognition of viruses.Immunol Lett1281720
- 10. Rehwinkel J (2010) Exposing viruses: RNA patterns sensed by RIG-I-like receptors. J Clin Immunol 30: 491–495.J. Rehwinkel2010Exposing viruses: RNA patterns sensed by RIG-I-like receptors.J Clin Immunol30491495
- 11. Bauer S, Muller T, Hamm S (2009) Pattern recognition by Toll-like receptors. Adv Exp Med Biol 653: 15–34.S. BauerT. MullerS. Hamm2009Pattern recognition by Toll-like receptors.Adv Exp Med Biol6531534
- 12. Manel N, Hogstad B, Wang Y, Levy DE, Unutmaz D, et al. (2010) A cryptic sensor for HIV-1 activates antiviral innate immunity in dendritic cells. Nature 467: 214–217.N. ManelB. HogstadY. WangDE LevyD. Unutmaz2010A cryptic sensor for HIV-1 activates antiviral innate immunity in dendritic cells.Nature467214217
- 13. Lepelley A, Louis S, Sourisseau M, Law HK, Pothlichet J, et al. (2011) Innate sensing of HIV-infected cells. PLoS Pathog 7: e1001284.A. LepelleyS. LouisM. SourisseauHK LawJ. Pothlichet2011Innate sensing of HIV-infected cells.PLoS Pathog7e1001284
- 14. Beignon AS, McKenna K, Skoberne M, Manches O, DaSilva I, et al. (2005) Endocytosis of HIV-1 activates plasmacytoid dendritic cells via Toll-like receptor-viral RNA interactions. J Clin Invest 115: 3265–3275.AS BeignonK. McKennaM. SkoberneO. ManchesI. DaSilva2005Endocytosis of HIV-1 activates plasmacytoid dendritic cells via Toll-like receptor-viral RNA interactions.J Clin Invest11532653275
- 15. Mandl JN, Barry AP, Vanderford TH, Kozyr N, Chavan R, et al. (2008) Divergent TLR7 and TLR9 signaling and type I interferon production distinguish pathogenic and nonpathogenic AIDS virus infections. Nat Med 14: 1077–1087.JN MandlAP BarryTH VanderfordN. KozyrR. Chavan2008Divergent TLR7 and TLR9 signaling and type I interferon production distinguish pathogenic and nonpathogenic AIDS virus infections.Nat Med1410771087
- 16. Nascimbeni M, Perie L, Chorro L, Diocou S, Kreitmann L, et al. (2009) Plasmacytoid dendritic cells accumulate in spleens from chronically HIV-infected patients but barely participate in interferon-alpha expression. Blood 113: 6112–6119.M. NascimbeniL. PerieL. ChorroS. DiocouL. Kreitmann2009Plasmacytoid dendritic cells accumulate in spleens from chronically HIV-infected patients but barely participate in interferon-alpha expression.Blood11361126119
- 17. Berkhout B, Grigoriev A, Bakker M, Lukashov VV (2002) Codon and amino acid usage in retroviral genomes is consistent with virus-specific nucleotide pressure. AIDS Res Hum Retroviruses 18: 133–141.B. BerkhoutA. GrigorievM. BakkerVV Lukashov2002Codon and amino acid usage in retroviral genomes is consistent with virus-specific nucleotide pressure.AIDS Res Hum Retroviruses18133141
- 18. Berkhout B, van Hemert FJ (1994) The unusual nucleotide content of the HIV RNA genome results in a biased amino acid composition of HIV proteins. Nucleic Acids Res 22: 1705–1711.B. BerkhoutFJ van Hemert1994The unusual nucleotide content of the HIV RNA genome results in a biased amino acid composition of HIV proteins.Nucleic Acids Res2217051711
- 19. Vartanian JP, Meyerhans A, Sala M, Wain-Hobson S (1994) G–>A hypermutation of the human immunodeficiency virus type 1 genome: evidence for dCTP pool imbalance during reverse transcription. Proc Natl Acad Sci U S A 91: 3092–3096.JP VartanianA. MeyerhansM. SalaS. Wain-Hobson1994G–>A hypermutation of the human immunodeficiency virus type 1 genome: evidence for dCTP pool imbalance during reverse transcription.Proc Natl Acad Sci U S A9130923096
- 20. Deforche K, Camacho R, Laethem KV, Shapiro B, Moreau Y, et al. (2007) Estimating the relative contribution of dNTP pool imbalance and APOBEC3G/3F editing to HIV evolution in vivo. J Comput Biol 14: 1105–1114.K. DeforcheR. CamachoKV LaethemB. ShapiroY. Moreau2007Estimating the relative contribution of dNTP pool imbalance and APOBEC3G/3F editing to HIV evolution in vivo.J Comput Biol1411051114
- 21. Bishop KN, Holmes RK, Sheehy AM, Malim MH (2004) APOBEC-mediated editing of viral RNA. Science 305: 645.KN BishopRK HolmesAM SheehyMH Malim2004APOBEC-mediated editing of viral RNA.Science305645
- 22. Wood N, Bhattacharya T, Keele BF, Giorgi E, Liu M, et al. (2009) HIV evolution in early infection: selection pressures, patterns of insertion and deletion, and the impact of APOBEC. PLoS Pathog 5: e1000414.N. WoodT. BhattacharyaBF KeeleE. GiorgiM. Liu2009HIV evolution in early infection: selection pressures, patterns of insertion and deletion, and the impact of APOBEC.PLoS Pathog5e1000414
- 23. Mulder LC, Harari A, Simon V (2008) Cytidine deamination induced HIV-1 drug resistance. Proc Natl Acad Sci U S A 105: 5501–5506.LC MulderA. HarariV. Simon2008Cytidine deamination induced HIV-1 drug resistance.Proc Natl Acad Sci U S A10555015506
- 24. van Hemert FJ, Berkhout B (1995) The tendency of lentiviral open reading frames to become A-rich: constraints imposed by viral genome organization and cellular tRNA availability. J Mol Evol 41: 132–140.FJ van HemertB. Berkhout1995The tendency of lentiviral open reading frames to become A-rich: constraints imposed by viral genome organization and cellular tRNA availability.J Mol Evol41132140
- 25. Easterbrook PJ, Smith M, Mullen J, O’Shea S, Chrystie I, et al. (2010) Impact of HIV-1 viral subtype on disease progression and response to antiretroviral therapy. J Int AIDS Soc 13: 4.PJ EasterbrookM. SmithJ. MullenS. O’SheaI. Chrystie2010Impact of HIV-1 viral subtype on disease progression and response to antiretroviral therapy.J Int AIDS Soc134
- 26. Kaleebu P, French N, Mahe C, Yirrell D, Watera C, et al. (2002) Effect of human immunodeficiency virus (HIV) type 1 envelope subtypes A and D on disease progression in a large cohort of HIV-1-positive persons in Uganda. J Infect Dis 185: 1244–1250.P. KaleebuN. FrenchC. MaheD. YirrellC. Watera2002Effect of human immunodeficiency virus (HIV) type 1 envelope subtypes A and D on disease progression in a large cohort of HIV-1-positive persons in Uganda.J Infect Dis18512441250
- 27. Kiwanuka N, Robb M, Laeyendecker O, Kigozi G, Wabwire-Mangen F, et al. (2010) HIV-1 viral subtype differences in the rate of CD4+ T-cell decline among HIV seroincident antiretroviral naive persons in Rakai district, Uganda. J Acquir Immune Defic Syndr 54: 180–184.N. KiwanukaM. RobbO. LaeyendeckerG. KigoziF. Wabwire-Mangen2010HIV-1 viral subtype differences in the rate of CD4+ T-cell decline among HIV seroincident antiretroviral naive persons in Rakai district, Uganda.J Acquir Immune Defic Syndr54180184
- 28. Bousheri S, Burke C, Ssewanyana I, Harrigan R, Martin J, et al. (2009) Infection with different hiv subtypes is associated with CD4 activation-associated dysfunction and apoptosis. J Acquir Immune Defic Syndr 52: 548–552.S. BousheriC. BurkeI. SsewanyanaR. HarriganJ. Martin2009Infection with different hiv subtypes is associated with CD4 activation-associated dysfunction and apoptosis.J Acquir Immune Defic Syndr52548552
- 29. Meier A, Alter G, Frahm N, Sidhu H, Li B, et al. (2007) MyD88-dependent immune activation mediated by human immunodeficiency virus type 1-encoded Toll-like receptor ligands. J Virol 81: 8180–8191.A. MeierG. AlterN. FrahmH. SidhuB. Li2007MyD88-dependent immune activation mediated by human immunodeficiency virus type 1-encoded Toll-like receptor ligands.J Virol8181808191
- 30. Chiu YH, Macmillan JB, Chen ZJ (2009) RNA polymerase III detects cytosolic DNA and induces type I interferons through the RIG-I pathway. Cell 138: 576–591.YH ChiuJB MacmillanZJ Chen2009RNA polymerase III detects cytosolic DNA and induces type I interferons through the RIG-I pathway.Cell138576591
- 31. Saito T, Owen DM, Jiang F, Marcotrigiano J, Gale M (2008) Innate immunity induced by composition-dependent RIG-I recognition of hepatitis C virus RNA. Nature 454: 523–527.T. SaitoDM OwenF. JiangJ. MarcotrigianoGale M Jr2008Innate immunity induced by composition-dependent RIG-I recognition of hepatitis C virus RNA.Nature454523527
- 32. Solis M, Nakhaei P, Jalalirad M, Lacoste J, Douville R, et al. (2010) RIG-I Mediated Antiviral Signalling is Inhibited in HIV-1 Infection by a Protease-Mediated Sequestration of RIG-I. J Virol. M. SolisP. NakhaeiM. JalaliradJ. LacosteR. Douville2010RIG-I Mediated Antiviral Signalling is Inhibited in HIV-1 Infection by a Protease-Mediated Sequestration of RIG-I.J Virol
- 33. Malathi K, Dong B, Gale M , Silverman RH (2007) Small self-RNA generated by RNase L amplifies antiviral innate immunity. Nature 448: 816–819.K. MalathiB. DongGale M JrRH Silverman2007Small self-RNA generated by RNase L amplifies antiviral innate immunity.Nature448816819
- 34. Sauter D, Schindler M, Specht A, Landford WN, Munch J, et al. (2009) Tetherin-driven adaptation of Vpu and Nef function and the evolution of pandemic and nonpandemic HIV-1 strains. Cell Host Microbe 6: 409–421.D. SauterM. SchindlerA. SpechtWN LandfordJ. Munch2009Tetherin-driven adaptation of Vpu and Nef function and the evolution of pandemic and nonpandemic HIV-1 strains.Cell Host Microbe6409421
- 35. Brenchley JM, Silvestri G, Douek DC (2010) Nonprogressive and progressive primate immunodeficiency lentivirus infections. Immunity 32: 737–742.JM BrenchleyG. SilvestriDC Douek2010Nonprogressive and progressive primate immunodeficiency lentivirus infections.Immunity32737742
- 36. Silvestri G, Fedanov A, Germon S, Kozyr N, Kaiser WJ, et al. (2005) Divergent host responses during primary simian immunodeficiency virus SIVsm infection of natural sooty mangabey and nonnatural rhesus macaque hosts. J Virol 79: 4043–4054.G. SilvestriA. FedanovS. GermonN. KozyrWJ Kaiser2005Divergent host responses during primary simian immunodeficiency virus SIVsm infection of natural sooty mangabey and nonnatural rhesus macaque hosts.J Virol7940434054
- 37. Smith HC (2011) APOBEC3G: a double agent in defense. Trends Biochem Sci. HC Smith2011APOBEC3G: a double agent in defense.Trends Biochem Sci