Abstract
Introduction
The vast proliferation of human endogenous retroviruses, and their derived and dependent products, in the human genome is increasingly recognized to be important in human embryology, normal physiology and disease. I introduced this subject in a preliminary paper, in 2004, 1 but progress during the ensuing five years demands a wholesale reappraisal. As with mitochondrial symbiosis, we find profound differences, conceptually, and mechanistically, between viral symbiogenesis and what we have come to expect of mutation per se. Moreover, while there are helpful commonalities with the symbiotic mechanisms of mitochondrial genetics, there are also important differences. These differences, which have evolutionary and medical implications, derive from the quintessential nature of viruses, and their evolved genetic and genomic behavioural patterns.
Retroviral symbiosis
We are familiar with exogenous retroviruses, such as the human T-cell leukaemia viruses, HTLV-I and HTLV-II, and the AIDS pandemic, caused by the human immunodeficiency viruses, HIV-1 and HIV-2. All such retroviruses enter into a state of ‘persisting infection’ with the host population, and this has symbiogenetic implications. 2 For example, even in the prevailing pandemic of AIDS, there is intense co-evolution between HIV-1 and the human major histocompatibility antigens, typifying the early stage of ‘aggressive symbiosis’. 3 When the virus infects a new host it first discovers its target cells, which appears to be the CD4+T lymphocyte and macrophage. 4 Here, employing its retrovirus-specific enzyme reverse transcriptase, it convert its RNA-based viral genome to the homologous DNA before uniting with the host genome, subverting its normal controls to convert it into a factory for the production of virus. This means that the virus is pre-evolved to manipulate its host genome, albeit at peripheral cell level. Many retroviruses have the additional, remarkable, ability to invade the germ-lines of their hosts, a process known as ‘endogenization’, which gives rise to a new symbiogenetic evolutionary entity – the holobiontic union of virus and host.
To date, the HIV-1 virus has not been shown to endogenize. Until recently many virologists believed that the lentivirus genus of retroviruses, which includes HIV-1, were unable to do so. But lentivirus endogenization has now been confirmed in European rabbits 5 and an endogenized immunodeficiency lentivirus, thought to be ancestral to all primate lentiviruses, has been found in the genome of the grey mouse lemur, a basal primate that inhabits Madagascar. 6 The latter may reflect an integration event dating back at least 14 million years, suggesting that lentiviruses have been co-evolving with primates for much longer than was previously considered.
Given the structure of the human genome, it would appear that our ancestors suffered many other retroviral epidemics throughout our human and primate history, progressing to germ-line union. If we consider that each invading virus was a complete, evolutionarily-competent, entity, and that it already possessed the pre-evolved capacity to manipulate the genome it was invading, we can anticipate that such unions would have resulted in major symbiogenetic potential. We can also extrapolate that such viral symbiogenesis would differ from what we saw with mitochondria, since bacterial genes and sequences have no such capacity for host genomic manipulation.
There are a few simple terms we need to define. When an ‘exogenous’ retrovirus becomes a permanent component of the human genome it is called a ‘human endogenous retrovirus’ (HERV). Unlike mitochondrial genes, HERVs are inherited in classical Mendelian fashion, as integral components of the new holobiontic nuclear genome. In Part 1, I referred to the retroviral epidemic affecting the koala population of Australia, which is following the typical pattern of ‘aggressive symbiosis’, with endogenization well under way within 100 years or so of the beginning of the epidemic. There is also evidence of a recent endogenization in humans, where HERV-K113, found on chromosome 19 in just 29% of people of mainly African, Asian and Polynesian descent, appears to have entered the human genome after the last great migration from Africa, less than 150,000 years ago. 7 We are only beginning to appreciate the evolutionary importance of such retroviral invasions. For example, the evolutionary virologist, Villarreal, has concluded that there was a major explosion of new retroviruses with the origins of jawed vertebrates, another with the origins of mammals and again with the origins of the primates. 8 These are likely to have played an important role in host evolution. The vertebrate-associated explosion gave rise to large-scale endogenous colonisation with resultant genomic expansion. This was coincident with the origins of adaptive immunity, leading Villarreal to propose a complex step-by-step scenario in which retroviral-host interaction gave rise to adaptive immunity as an integral part of the evolution of a highly sophisticated system of self-identity. We shall return to Villarreal's hypothesis in considering the viral role in the auto-immune diseases in a subsequent paper. Meanwhile, we shall focus on the retroviral invasions of the vertebrate, mammalian, primate and hominid genomes that have played an important role in our pre-human and human evolution.
All retroviruses have a similar basic genomic structure ( Basic genome of retrovirus
The first step in the symbiotic adaptation of an endogenized virus will be the loss of
infectious behaviour. This has been achieved with the most recently endogenized human
virus, HERV-K113, mentioned above, through the intervention of a single stop codon in
its gag domain.
13
From this point on selection working at holobiontic level will adapt both the
viral and original vertebrate elements to the new partnership, so that
mutation-plus-selection is now fulfilling an editorial role. Selfish elements will be
silenced in the short term by epigenetic mechanisms, such as methylation, histone
modification or RNAi, and in the longer term through genetic mechanisms such as mutation
of their DNA sequences, or through elimination from the genome through chromosome breaks
in cell division, or through deletions during homologous sexual recombination. But where
viral integrity, in whole or in part, is important to the holobiont, the viral sequences
will be conserved. These symbiogenetic mechanisms will apply to the elements of both
host and viral genomes, and they will continue to apply long after the initial
virus-host fusion, with positive selection preserving viral or host
genes, viral or host translational sequences and viral
or host control and developmental sequences that contribute to the
evolving holobiont. This offers a comprehensive, and testable, explanation for the
proliferation of whole viruses, and the so-called ‘defective HERVs’, and HERV products,
such as LINEs and SINEs ( HERVs and products in the human genome
Some 9% of the human genome consists of complete human endogenous retroviruses, or
HERVs, and their LTRs, and if we extend this to HERV genes, fragments and derivatives,
such as LINEs and SINEs, the viral component amounts to roughly 43% of our DNA ( DNA breakdown of human genome
The implications for human development and normal physiology
Two endogenous retroviruses, HERV-W and HERV-FRD, play an important role in the construction and physiological function of the human placenta, their env genes coding for the proteins syncytin-1 and syncytin-2, respectively, which fuse the trophoblast cells of the placenta into the confluent multinucleated syncytial layer. 19, 20, 21 Each virus fulfils a specific and different role in placental construction, working in a complementary way with a third virus, ERV-3 (also known as HERV-R), in a complex developmental coordination. 22 However, 1% of Caucasians may have a mutation that precludes the fusion and putative immunosuppressive functions of ERV-3, yet this does not appear to prevent pregnancy, 23 suggesting that there may be some degree of redundancy, or flexibility, in the remarkable multi-HERV coordination. Only a minority of mammals has been assessed to date, but it would appear that the major mammalian groups have different patterns of placentation, involving different endogenous retroviruses, for example primates, mice and sheep, 24, 25, 26, 24–27 suggesting that mammalian placentation is either of polyphyletic origins or that newly incoming exogenous viruses may compete with and replace established viruses. Syncytin-2, unlike syncytin-1, has also been shown to have immunosuppressive properties that may play a role in the maternal tolerance to foetal antigens. 28 Humans share their placental viruses with the great apes, and possibly old world, but not new world, monkeys. 29 The exogenous retrovirus that gave rise to the HERV-W endogenous family is believed to have entered the ancestral genome almost 40 million years ago. When Bonnaud and his colleagues tracked the action of natural selection on the HERV-W locus ERVWE1 (whose env gene codes for syncytin-1) in chimpanzee, gorilla, orang-utan and gibbon, they found that the genetic signature crucial to the gene's fusogenic action had been conserved over the tens of millions of years of primate divergence. 30 This is what one would anticipate from the symbiogenetic perspective, with selection operating at the level of the holobiontic genome.
There is growing evidence that many retroviruses and their products are active during human germ cell formation and embryogenesis. Based on the fact that mature spermatozoa are endowed with reverse transcriptase, Spadafora and associates have blocked its action, using the anti-retroviral drug Nevirapine, to discover that this caused an irreversible arrest of development up to the four-cell stage of mouse development. 31 They further demonstrated that reverse transcriptase inhibition caused a substantial reprogramming of gene expression in arrested embryos, involving both developmental and translational genes. 32 This suggests that endogenous retroviruses, or LINE type products, are playing an important role at the earliest stages of mammalian development. Larsson and colleagues have pioneered the study of HERVs in human embryogenesis, showing that ERV-3, one of the three viruses involved in placentation, is also highly expressed in many human foetal tissues, including adrenal cortex, kidney tubules, tongue, heart, liver and central nervous system. 33 The same researchers have also shown that ERV-3 is highly expressed in the sebaceous glands of normal skin. 34 De Parseval and colleagues searched the human genome for retroviral envelope genes with ‘open reading frames’ – genes available for transcription to proteins – to discover 16 HERV env genes, all of which were expressed in healthy tissues. 35 Three of these were expressed in placenta, as seen above. One, a newly identified env gene, was expressed at a high level in the fully developed thyroid gland, and appeared to be exclusive to this organ, with a possible link to hormone secretion. Another HERV gene, envR, was expressed in the adrenal, where the association may again be linked to hormone secretion. It is significant that all 16 were expressed in the testis. This is congruent with the fact that endogenous retroviruses are a common finding in the male reproductive tract from Drosophila to mammals. 36
De Parseval also found that, despite the fact that some of these genes were only transcribed at low level, all of the promoters were active, very likely in germ line expression. Others have detected HERV-W RNA in the human testis and HERV-R env mRNA appears to be expressed in the first phases of spermatogenesis but not in the Sertoli or Leydig cells. 37 This expression appears to be linked to steroid hormones since, as Larsson and colleagues have shown, HERV-R contains androgen receptor sites in its 5'LTR. Proteins coded by HERV-derived L1 retrotransposons have also been found in prespermatogonia of foetal testis, in germ cells of adult testis, and in Leydig and Sertoli cells, as well as vascular endothelial cells, the latter suggesting a possible role in vasculogenesis. 38
HERVs and their products appear to play an equally pervasive role in normal adult structure and physiology. Perhaps most dramatically of all, a number of researchers, including Larsson and colleagues at Uppsala, have shown expression of HERV env and gag genes in key tissues and structures of the human brain, indicating what appears to be structural or physiological function, or both. 39 . 40 In particular, the Swedish group have demonstrated widespread and dense expression of the env proteins syncytin-1 and syncytin-2, presumably different variants of the proteins from those found in the placenta, which is undergoing further evaluation. 41 Other researchers are studying the control and promotional sequences of HERVs in human physiology. For example Mager and colleagues have shown that the LTR of ERV-L controls most of the gene transcripts of the human gene β3GAL-T5 in the human colon, 42 meanwhile Sverdlov and colleagues have shown two examples of HERV LTRs that are participating in the specific antisense regulation of the human genes SLC4A8 (sodium bicarbonate co-transporter) and IFT172 (intraflagellar transport protein 172). 43 Previously the same researchers had shown that at least 50% of the human-specific HERV-K LTRs are active promoters for non-viral DNA transcription. 44 For example, the LTR of ERV-9 has been shown to play a key role in transcriptional control of the β-globin gene cluster of humans. 45, 46 This key viral promoter has also been conserved throughout at least 15 million years of primate evolution, and appears to have displaced several other non-viral promoters within the locus control region. A range of virus-derived elements, including SINEs (Alus), LINEs, and LTRs, are found in a large number of human protein-coding genes, where most are inserted into introns – the non-coding regions between the coding regions, known as exons – and where they appear to influence the function of some 533 genes. 47 Other LTRs have taken on the role of alternative promoters, or splice receptors, for example in the control of the endothelin B receptor and apolipoprotein C-I genes, 48 and in the control of the human leptin receptor, which is involved in energy expenditure, production of sex hormones and activation of haemopoietic cells. 49 Endogenous retroviruses have also been involved in the evolution, and tissue specific expression, of the enzyme amylase in humans. 50, 51 HERV sequences have also contributed functional genes, or parts of genes, to the human genome, including integrase 52 and transaldolase. 53
Viruses also have the ability to create entirely new genes (neogenes) through stitching together fragments of other genes, 54 and HERVs have retained this capability. For example a novel gene, PIPSL, common to chimpanzee and human testis, appears to have been created by ‘exon-shuffling’. 55 Another family of neogenes, called Mart, are located on the X chromosomes of various mammals, including humans. 56 They play ubiquitous roles during embryogenesis of the mouse, with important function in the nervous system. Although largely unevaluated in humans, at least one of the Mart genes (Mart 8) is amplified in the human genome. It seems likely that, in time, more such virally-created neogenes will be discovered as part of our viral-symbiogenetic inheritance.
Virus-derived transposable elements have also been found to contribute to a substantial fraction of human regulatory sequences. 57 LINEs are HERV-derived long interspersed repetitive elements that have massively reproduced themselves throughout the human genome. Although they have lost much of the original HERV genetic structures, they retain the enzymes necessary to replicate and move to new positions within the genome. This is known as ‘retrotransposition’. SINES are short interspersed repetitive elements that have also massively dispersed themselves through the genome, but they have lost the enzymes necessary for retrotransposition and so they must borrow these from other sources, most likely from LINEs. 58 Recent work by Ohshima and colleagues has demonstrated evolutionary links between LINEs and SINEs, which, between them, comprise some 34% of the human genome. 59 Mammalian specific LINES are known as LINE-1s (often abbreviated to L1s). Most of these have been inactivated by selection, but about 100 or so remain highly active, copying themselves and inserting elsewhere in the genome, where they play a significant role in the structure and regulation of the genome. One important role for L1s may be DNA repair. 60, 61, 62 A subsection specific to humans has been specified variously as LINE-1H, or LIH, or sometimes Ta, and there is growing evidence that the rate of L1 amplification has been increasing during recent human evolution. 63, 64
Alu elements, also known as Alu repeats, are specific to primates and comprise a short sequence of approximately 300 nucleotides, which tend to be loosely included with SINEs. Like SINEs, they require the presence of HERV or LINE enzymes to insert themselves into a new location. 65, 66 They are so efficient at this that they have amplified to more than one million copies in the human genome, including 2000 or so that are specific to humans, and they continue to amplify at the rate of about one new insertion every 200 births, so that these more recent insertions have been used to track and survey human evolutionary origins. 67, 68 Where HERVs and LINEs tend to transpose themselves randomly within the genome, Alus appear to hone into regions that are particularly rich in vertebrate genes. This, as we shall see in the ensuing paper, results in a mutation-like propensity to disrupt normal gene expression, and thus give rise to a wide range of diseases. Nevertheless, these too have played an important role in the evolution of the human genome, through such insertions as well as recombination between elements, gene conversion and alterations in gene expression, 69 and possibly in the evolution of the primate transcriptome. 70
Geneticists are becoming increasingly aware of the importance of viral lineages, including genes, gene fragments, LTRs, LINES, SINES, and dependents such as Alus, on genetic and whole genomic function in humans. The observed viral roles fulfil the predictions of symbiogenetic evolution, as do the large repertoire of viral elements that code for a multiplicity of functions essential to the evolution and normal working of the genome, including the contribution of unique coding sequences, organisational roles during genomic duplication, accurate transmission to progeny cells, and a fundamental role in the cooperative molecular interactions intrinsic to nucleoprotein complexes. 71, 72 Indeed the viral roles in human genomic function are so widespread, yet still underestimated, that Flockerzi and colleagues have suggested the need for a specialised human endogenous retrovirus transcriptome project. 73 A holistic long-term consequence of the symbiotic viral component of the genome is an increase in ‘genetic plasticity’, which has important implications for medicine as well as evolutionary biology. 74 This is likely to contribute to the high level of genetic variation currently being observed between individual human genomes. 75
Part 3 of this series will examine the role of HERVs and their products in miscellaneous human diseases and the autoimmune diseases.
Footnotes
DECLARATIONS
