The following article is excerpted from a seminar delivered to the University of Washington for the Minority Scientist Seminar Program in December and in July of 1997 to the Department of Genetics in the University of Valencia.

This year, 1997, marks the one hundredth year anniversary for the discovery of virus as a filterable causal agent of plant and animal disease. During the ensuing 100 years we have witnessed the growth of biological thought that has included the synthesis between genetics and evolutionary biology and the subsequent advent of molecular biology. These fields encompass broad themes in biology that link the study of all living things. However, in spite of the apparent long history of the study of viruses, the modern concept and definition of a virus as a molecular genetic parasite was not clearly presented until Salvador Lauria published his first essay of on this topic in 1950, well after the development of much of the thought behind our understanding of evolutionary mechanisms. We currently think of a virus as an agent that necessarily reduce host fitness and generally cause disease, together with other pathogenic microorganisms, such as bacteria and fungi. In this presentation I will develop the idea that in addition to this role, viruses can also invent systems of molecular genetic identity and superimpose a new combined identity onto the infected host. In so doing, a virus can allow to host itself to adapt to the environment and evolve quickly, providing a creative force that the host may further develop into systems of identity and immunity that can contribute directly to host evolution. Along these lines, in this lecture I will examine the evidence that endogenous retroviruses might be crucial to the evolution of placental orders. Because placental species give live birth, they face an immunological dilemma in that their highly adaptive immune systems fail to recognize their own embryos which have ÔforgienÕ father derived (allogeneic) antigens. As the genomes of placental mammals are also highly infected with retroviruses found only in their genomes (endogenous) and because retroviruses are generally immunosuppressive, I examine the possibility that the embryo is acting like an infectious agent that produces virus to suppress the motherÕs immune system. In addition, I will compare this situation to that of another genomic virus, the polydnaviruses which are symbiotic in parasitoid wasp species and are also immunosuppressive. I will conclude by arguing that the persistence of parasitic viral-like genomes may represent one of the primary mechanisms for the evolution of higher order living systems.

Our current and prevailing view of the relationship of a virus to its host is similar to how we think of a predator and its prey. The virus can be considered like a predator and the host its prey in terms of their interactions as well as how we model this dynamic using mathematical models of disease. This predator-prey model assumes that the affect of virus infection on host is to decrease host fitness via disease. In this model, efficient virus replication is often linked to host induced disease. This type of model appears to work well when applied to numerous acute viral infections of human and animal disease, such as smallpox, measles, influenza and more recently HIV and AIDS. This type of model also seems applicable to other microbiological agents, such as pathogenic bacteria and fungi, in terms of affects on host fitness. Thus it appears that we have a well established and accepted explanation of virus host dynamics that fits known human viral epidemics.

The view that viruses are principally major agents of disease is richly deserved. In human history, viral epidemics have accounted for more human deaths than all known wars and famine combined. This especially evident in the new world following the introduction of smallpox, then measles, influenza, and mumps into the then naive native american population from Europe. In Europe, these disease had already established a childhood pattern of infection. Essentially all adult Europeans were literally the survivors of childhood infections with smallpox and measles. Smallpox was particularly significant in New World demographics. The first epidemic on the mainland was to hit the Aztec population around the time of the infamous Ônoche tristeÕ on June 20th 1520, killing all the leaders and many warriors of the Aztec revolt that expelled Cortez. This revolt, one that killed many of CortezÕs conquistadors and drove them from Tinochtitlan (the Aztec capital), was to ultimately fail in spite of the seemingly enormous numerical advantage of the Aztecs. This smallpox epidemic and resulting social chaos was to deliver a death blow to this numeric superiority and clear the way for the successful return Cortez and his allies the following year. This and subsequent epidemics were to continue their inexorable march through the Inca and Mayan civilizations then later into the North and South American continents, including Indians from the Mississippi valley, the Eastern seaboard, then into California and the Columbia river valley resulting in the greatest demographic catastrophe in human history. Viral epidemics continue to threaten the human population. In this century the great influenza pandemic of 1918 accounts for the most human deaths of any event, while the new HIV pandemic is the newest worldwide epidemic threat. Clearly, viral epidemics seem to prey on human populations. However, when we consider virus host relationships during evolutionary time scales, none of these acute epidemic infections were likely to have been prevalent during most of human evolution for they are not able to establish stable persistent infections needed to survive in small groups of hunter gatherers that characterized much of early human history. Yet other viruses were almost certainly prevalent.

It is clear that there exist another relationship of virus to host that does not fit the predator-prey model noted above and works well in host that live in small populations. This other relationship is characterized by a stable persistence of the virus once the host has been infected. The stability of the virus-host linkage is often seen even on an evolutionary time scale. In the vertebrates, this relationship is easily apparent in the viruses I study, the small DNA viruses such as papillomavirus and polyomavirus. Once the host becomes infected by such viruses, they are usually present for the life of the host. These and many other viral infections spend the great majority of their existence as persistent non-pathogenic infections with little apparent effect on host fitness. As these agents can be highly prevalent in their host species and can be found in most related species, they are often conserved phylogenetically. That is, these types of virus appear to be co-evolving with their host and diverge from each other at the same rate that the host species themselves have diverged. In addition, such viruses are generally host species specific., which raises the question about the mechanisms of species specificity. In any human audience, for example, it is highly likely that the great majority will be infected with an array of such viruses, including human papilloma virus, human polyomavirus, human adenovirus as well as several members of the herpesvirus family. Persistent infections are common in nature and can readily be found in most organisms, from bacteria to mammals. Some of these persistent infections also appear to be genomic in that the virus is usually and sometimes always found in the host genome. For example, most bacteria have one or several prophage in their genomes. In mammals, it appears all species have endogenous retroviruses as part of the host DNA. In spite of how common such infections may be, we do not have a well developed mathematical model of persistent virus-host interaction for it draws much less attention and study then disease causing infections. Only when such infections occasionally lead to disease, such as in the association of otherwise inapparent Human Herpes virus type 8 with KaposiÕs sarcoma in AIDS patients to we pay attention to such persistent agents.

Persistent infections by viruses and virus like parasites, such as bacterial plasmids, are known to often provide genes to the host that can be of adaptive value. For example, bacterial pathogenesis, is often due to the ability of certain bacteria to adapt to specific environments of a host such as a human, but in so doing also cause disease. Many of the genes that make bacteria pathogenic are due to persistent viral infections or infections with virus-like plasmids, including the ability to produce diphtheria and botulinum toxins. Those genes that are not directly part of a virus usually reside in clusters know as pathogenic islands that appear to have arisen from horizontal transmission events and thus resemble viral transmission, as recently reviewed by Stanley Falkow. It therefore seems curious that one of the most adaptable organisms we know, bacteria, use horizontally transmitted elements as a major mechanism of adaption. Why donÕt all organisms use this highly successful process?

However, if we examine what happens to genomic DNA as it has evolved to higher order, certain trends become apparent. Most notably is the clear trend to accumulate of virus-like or parasitic DNA as part of the genome. Although the human chromosome is about 1000 times the size of the smallest bacteria, the human genome only has about 50 times as many genes as bacteria have. However, as all organisms evolve to higher order, they appear to decrease the density of genes within DNA from about 1.2 kb per gene in bacteria, 4 kb per gene in drosophila to about 29 kb per gene in the human genome, and not simply accumulate more genes. There is no accepted evolutionary theory that accounts for this pattern of genome evolution. Thus in mammals about 95% of the DNA is non-coding. In the case of mammals, a great part of this noncoding DNA is made up various families of repeated DNA such as Alu, SINES, LINES, retroposons and endogenous retroviruses. Placental mammals are especially prone to the presence of these types DNA, relative to avian and reptile species. In fact each mammal has it own unique version of LINE DNA present at up to a million copies per genome. All of these ÔparasiticÕ DNAÕs appear to derive from the action of reverse transcriptase on various cellular and retroviral RNAÕs and thus they have precise 5Õ ends, a part of the reverse transcriptase coding sequence and DNA copies of the poly-A tail. Yet we know of no selective force that should conserve the presence of reverse transcriptase to generate such DNA families. Avian and reptile species have much less if any of this type of DNA. Although avians do have some families of repeated DNAÕs, these avian repeat sequences lack these features seen in LINES. There is currently no explanation for the presence of so much of this DNA in mammalian genome. In addition, placental species all appear to have also conserved intact version (with all the genes for making a virus) of endogenous retrovirus that is co-evolving (phylogenetically congruent) with the host species. However, these intact copies, unlike the retroposons noted above are present at a very low copy level, sometimes only one copy per haploid genome. There is no explanation currently available to account for such conservation especially given that these endogenous retroviruses bear no resemblance to existing free retroviruses and would not likely provide possible protection against free living versions of such retroviruses.

The genesis of placental species is a relatively new development in the evolution of organisms and dates to the Cretaceous-Paleocene boundary about 65 million years ago. Although not well appreciated, the first mammals, know as the Multituberculates, came into being well before this and even before the dinosaurs during the Triassic period about 210 million years before the present. These mammals were most likely egg laying monotreme like creatures that resembled rodents, thus they most likely did not face the immunological dilemma of modern placental species. Early mammals were to become very well distributed, surviving past the demise of the dinosaurs but mysteriously becoming extinct about 35 million years ago. The closest relative to the placental mammal are the marsupials. Yet marsupials have only limited immunological contact between the fetus and the mother. Their embryos are much more egg like than those of placental species. A main distinction between marsupial embryos and placental embryos is the presence of the outer cell layer of the early placental embryo known as the trophectoderm. This cell layer is to only one to expressing paternal genes and is involved directly in implantation into the uterus then goes on to develop into the placenta. This tissue is the first cell type to differentiate in the placental embryo, yet was also the most recently evolved relative to early mammals. It therefore appears that the trophectoderm is crucial for the biology of placental life strategy.

In terms of implantation and escape from immunological rejection, the trophectoderm appears central to the ability of a placental embryo to prevent immunological recognition. Unlike most any other tissue, mouse trophectoderm can be implanted across strain barriers without being rejected. In addition, the trophectoderm can protect the inner embryo from attack by macrophages. However, it has been unclear what aspect of the trophectoderm protects the embryo. Various models have been proposed including altered expression of antigen presenting molecules (MHC) but these models all have significant problems. However, one activity that is rather unique to the trophectoderm (syncytiotrophoblast) is remarkable; they express extremely large quantities of endogenous retrovirus genes and retroviral particles, which include the envelope gene. The envelope gene is deleted from most of the defective copies of retroviruses found in the genome but has been conserved in these intact copies. In addition, the envelope gene is generally responsible for the ability of many retroviruses to suppress the immune system of the host. Other endogenous retroviral genes, such as gag, also appear to be able to modulate the immune response. The retrovirus that are being expressed in the embryoÕs trophectoderm are also highly conserved in all placental species examined so far. It therefore seems possible that this endogenous retrovirus may be providing protection to the embryo from the mothers immune system.

The model being proposed is as follows. Placental embryos make an extraembryonic tissue, the trophectoderm, which produces and endogenous retrovirus at high levels. This retrovirus, however, is not fully competent to replicate, but instead it can only infect local immune cells (such as macrophages) of the uterus, thereby compel those cells to express viral genes and prevent them from initiating an immune response. Thus, the motherÕs immune system will thus remain competent to respond to other infections but is specifically prevented from mounting an immune response to the embryo. Only the presence of the embryo itself prevents an immune reaction.

However, there are numerous consequences expected from such a situation. For one, the placental embryo must conserve this endogenous retrovirus in order to allow live birth. Therefore, early embryos will always be expressing the reverse transcriptase enzyme. The required expression of this ÔviralÕ enzyme can have other effects, such as allowing a high rate of reverse transcription to occur with various cellular or retroviral RNA transcripts of genes. Thus placentals will be prone to accumulate retroposons, such as LINE elements. However, given the immunosuppressing activity of the envelope gene, it is expected that most displaced copies of endogenous retrovirus will have deleted this gene. In addition, these endogenous retroviruses may also on occasion lead to the generation of free living retroviruses following recombination with various cellular or other viral genes. Placentals species should thus be prone to retrovirus infections.

Placentals therefore are proposed to maintain a genomic virus, an endogenous retrovirus, that is expressed in the early embryo. This virus is not competent to replicate but will infect local immune cells of the uterus and suppress the initiation of immune recognition. However, as remarkable as this proposal may seem, this type of relationship is not unique to placental mammals. There exists another completely different genomic virus in hymenoptera insects that is also expressed along with an egg whose role is also to suppress the immune recognition of a host in addition to other effects. This virus is known as polydnavirus which are made in the ovaries of most parsitoid wasp species. This virus is injected along with the wasp egg into host larvae. The virus infects the parasitized larvae and prevents phagocytic cells from attacking the egg. The virus does not replicate in the larvae but does cause larval cells to express polydnaviral genes. In a very real sense, polydnaviruses are a natural system of gene therapy. In addition, there are no free living versions of polydnaviruses viruses known, they are all genomic in the hymenoptera species, suggesting that their relationship with the parasitoid wasp host dates to the beginnings of this life style of wasp.

These two situations appear to be examples of a persistent virus that can provide a new function for the host and allow new life strategies. I have called such a virus a ÔmetavirusÕ for it transcends the distinction of virus from its host. This concept raises a broader question about how viruses and their more defective cousins, such as plasmids or retroposons, can allow rapid host adaption. Viruses can pose a major evolutionary dilemma for the host. In some cases they can evolve at rates up to a million fold greater then that of the host. Given this enormous advantage in adaption, how is the host ever to stay ahead of potential viral parasites? This seems an impossible task. However, if we consider the strategy of persistence of parasitic agents as a very selected or fit relationship, we can offer a solution to this dilemma. Viral like parasites may themselves be inventing the new identities that allow host evolution. Consider for example the evolution of bacterial systems of antiviral immunity. This is principally the restriction modification system found in most bacteria in which one modification enzyme chemically alters the DNA while another restriction enzyme will degrade unmodified ÔforenÕ DNA. How was this system invented as it requires the simultaneous creation of two independent enzyme activities? We now know that there exist parasitic and persisting plasmids of E. Coli, that code for both restriction and modification enzymes. Together, these two genes compel the infected bacteria to persistently maintain the plasmid. If this parasitic plasmid is lost from the host bacteria, the remaining restriction activity will degrades the now unmodified bacterial DNA and destroy the bacteria. In a sense, this parasitic plasmid provides both a genetic poison and the antidote that forces persistence. However, once thee bacterial host is persistently infected with such a parasitic plasmid, it has acquired a new molecular genetic system of DNA identification and will be immune to the infection of other viruses or genomic parasites that are not similarly modified. Thus a persistent infection can provide a critical adaptive function (immunity) to the host. It seems likely to me that if we look carefully, we will see the footprints of persistent viruses that were involved in the invention of the amazingly adaptive immune system of vertebrates.

It therefore appears that virus and virus-like systems may be contributing in a most profound way to evolution of their host. An early view from Salvador Lauria had envisioned a similar conclusion. When considering the issue of how a virus might contribute to the host, he wrote Ò... may we not feel that in the virus, in their merging with the cellular genome and their re-emerging from them, we observe the units and process which, in the course of evolution, have created the successful genetic patterns that underlie all living cells?Ó. Thus we may wish to reconsider the architecture of the tree of life. Rather then thinking of it as a tree with the usual structure of unconnected tips from a common branch perhaps a structure more reminiscent of that done by Mexican artisans rendering of the Ôarbol de a vidaÕ in which horizontal elements, bearing images of both life and death, are connecting the tips may better account for the contributions of viral agents to host evolution