Pa­le­ovi­rol­ogy

by Welkin John­son

How does one even be­gin to in­ves­ti­gate the nat­ural his­tory of viruses? The di­nosaurs be­queathed a mot­ley as­sort­ment of bones, teeth, foot­prints strid­ing 'cross an­cient riverbeds, fos­silized eggs, the oc­ca­sional co­pro­lite. The tiny trilo­bite left last­ing and ubiq­ui­tous im­pres­sions, find­ing its way into text­books and mu­seum gift shops. Even pre­his­toric cyanobac­te­ria, minis­cule and bone­less, are abun­dantly memo­ri­al­ized by gath­er­ings of statue-like stro­ma­to­lites.

Stro­ma­to­lites from the Pro­tero­zoic (2.3 bil­lion years ago) found in the An­des of Bo­livia. Source

As a sci­en­tist fas­ci­nated with the evo­lu­tion­ary in­ter­play be­tween viruses and their hosts, I ad­mit to con­sid­er­able pro­fes­sional envy. The pa­le­on­tol­o­gists have it good. What, if any­thing, does a virus leave be­hind? My study sub­jects are ut­terly lack­ing in bony, fos­siliz­able ma­te­r­ial, are too tiny to leave in­for­ma­tive im­pres­sions in stone, and, un­like bac­te­ria, pro­duce no tell­tale geo­chem­i­cal sig­na­tures. By ne­ces­sity, vi­ral pre­his­tory is tra­di­tion­ally in­ferred in­di­rectly from phy­lo­ge­netic re­con­struc­tion, typ­i­cally based on aligned se­quences of highly con­served sub­do­mains shared by many vi­ral poly­merases. But these are ge­netic se­quences, ob­tained from mod­ern vi­ral species, and the in­ferred an­ces­tors aren't "real;" they are sim­ply av­er­ages, each one a best-guess con­sen­sus. More im­por­tantly, this ap­proach is lim­ited to viruses with liv­ing mod­ern de­scen­dents; it tells us noth­ing about ex­tinct vi­ral lin­eages. (Most likely T. Rex had its own con­tin­gent of ob­lig­ate in­tra­cel­lu­lar par­a­sites?).

In this re­gard, the retro­viruses are the no­table ex­cep­tion.

Re­verse tran­scrip­tase con­verts the retro­vi­ral RNA genome into dou­ble stranded DNA, which is then in­te­grated ir­re­versibly into the ge­nomic DNA of the in­fected cell to form the DNA provirus. This process has oc­ca­sion­ally re­sulted in the de­po­si­tion of proviruses in the germlines of their an­i­mal hosts. Over the ex­panse of evo­lu­tion­ary time, the genomes of vir­tu­ally every an­i­mal species have be­come rid­dled with these provi­ral se­quences, the so-called en­doge­nous retro­viruses (ERVs). Most ERV se­quences have been de­graded by the ac­cu­mu­la­tion of mu­ta­tions but are still rec­og­niz­able as retro­vi­ral in ori­gin. The hu­man genome alone con­tains hun­dreds of thou­sands of HERVs (Hu­man ERVs), out­num­ber­ing our genes. Ex­trap­o­late these num­bers across the en­tirety of the an­i­mal king­dom, and col­lec­tively ERV loci may well com­prise a "fos­sil" col­lec­tion num­ber­ing in the hun­dreds of mil­lions of spec­i­mens. (Now who's jeal­ous?)

"Res­ur­rected" viri­ons bud­ding from a trans­fected cell. Bar = 500 nm (left) and 100 nm (right). Im­age cour­tesy of Paul Bi­eni­asz. Source

ERVs have at­tracted their own ded­i­cated "fos­sil" hunters. Jonas Blomberg's lab at Up­p­sala Uni­ver­sity, and their col­lab­o­ra­tors, have even de­vel­oped a nifty tool, called Retro­Tec­tor, for dig­i­tally sift­ing through bil­lions of base-pairs of ge­nomic DNA the way a Leakey might sift through sed­i­ment for a tooth or a bone chip. Like some­thing out of a Michael Crich­ton novel, the labs of Thierry Hei­d­mann and Paul Bi­eni­asz have gone so far as to use con­sen­sus se­quences to re­con­struct an­cient retro­vi­ral genomes that then pro­duced par­ti­cles in trans­fected cells, thus es­sen­tially com­plet­ing a vi­ral repli­ca­tion cy­cle ini­ti­ated more than a hun­dred thou­sand years ago.

The lentiviruses, a genus of the mod­ern retro­viruses, in­cludes two groups: the pri­mate lentiviruses, in­clud­ing the hu­man im­mun­od­e­fi­ciency viruses (HIV‑1 and HIV‑2) and the myr­iad Simian Im­mun­od­e­fi­ciency Viruses (SIV) of African pri­mates, and the non-pri­mate lentiviruses, in­clud­ing viruses of goats, sheep, cows, cats, horses, and oth­ers. Re­mark­ably, one team of ERV hunters re­cently un­earthed lentivi­ral ERVs from both of these mod­ern groups.

One was dis­cov­ered in the genome of the Eu­ro­pean brown rab­bit, and for this the dis­cov­er­ers coined the name RELIK (Rab­bit En­doge­nous Lentivirus type‑K). RELIK se­quences are most closely re­lated to the non-pri­mate lentiviruses, shar­ing sim­i­lar genome struc­tures and sig­nif­i­cant se­quence sim­i­lar­ity. They have now been found in other lep­orid species, in­di­cat­ing that the virus (and by ex­ten­sion the Lentivirus genus) is at least 12 mil­lion years old.

Gray Mouse Lemur. Source

Pri­mate lentivi­ral ERVs were found in the genomes of Gray Mouse Lemurs (Mi­cro­ce­bus mur­i­nus), pint-sized pri­mates found ex­clu­sively on the is­land of Mada­gas­car. They called these SIVgml (for Simian Im­mun­od­e­fi­ciency Virus of Gray Mouse Lemurs), sig­ni­fy­ing their close sim­i­lar­ity to the ex­tant SIVs en­demic among mod­ern apes and mon­keys in Africa. SIV-like se­quences have since been found in sev­eral ad­di­tional species of lemur. Given that the Mada­gas­car pri­mates have been ge­o­graph­i­cally sep­a­rated from main­land rel­a­tives for the past 75 mil­lion years, ex­plain­ing the pres­ence of ex­tant SIV on the main­land and the SIVgml ERVs on Mada­gas­car presents a gen­uine puz­zle to pa­le­ovi­rol­o­gists.

SIVgml is also some­thing rare that pa­le­on­tol­o­gists are of­ten hard pressed to come by: a tran­si­tional form. Pri­mate and non-pri­mate lentiviruses dif­fer by, among other things, the pres­ence of a dUT­Pase gene in the non-pri­mate ver­sion. While SIVgml has sev­eral genes unique to the mod­ern pri­mate lentiviruses, it also has a dUT­Pase, thus es­tab­lish­ing an evo­lu­tion­ary link be­tween the two branches. Where the lentiviruses orig­i­nated is a mys­tery, but if we ac­cept that RELIK and SIVgml must have shared a com­mon an­ces­tor, then at some point at least one an­ces­tral lentivirus must have mi­grated across much of the African con­ti­nent.

Thir­teen-lined ground squir­rel. "What? Viruses in MY genome?" Source

In 2008 while at­tend­ing a meet­ing on Awaji Is­land, Japan, I ran into my for­mer col­league, Dr. Keizo Tomanaga. Keizo is a vi­rol­o­gist in the Re­search In­sti­tute for Mi­cro­bial Dis­eases, Os­aka Uni­ver­sity where he and his col­leagues study Borna Dis­ease Virus (BDV). As we sam­pled hors d'oeuvres, Keizo ca­su­ally men­tioned that they had found Bor­navirus se­quences in the hu­man genome. Seek­ing clues to the func­tion of Bor­navi­ral pro­teins, they had done a BLAST search of the hu­man genome look­ing for cel­lu­lar pro­teins that were struc­turally sim­i­lar to Bor­navirus pro­teins. What they ac­tu­ally found were (at least at one time) bona fide Bor­navi­ral genes. Since these el­e­ments were de­rived from one par­tic­u­lar seg­ment of the Bor­navi­ral genome, that be­ing the gene for the vi­ral N pro­tein (N for nu­cle­o­pro­tein), they named them "En­doge­nous Borna-Like N" el­e­ments, or EBLNs. Keizo and his col­leagues ul­ti­mately un­earthed a trove of EBLNs sand­wiched into the genomes of mul­ti­ple pri­mate species, as well as other mam­mals in­clud­ing ele­phants and thir­teen-lined ground squir­rels. With the pub­li­ca­tion of these find­ings in Na­ture, retro­viruses no longer hold the dis­tinc­tion of be­ing the only an­i­mal viruses with a rich fos­sil record.

This find­ing was un­ex­pected since, un­like retro­viruses, Bor­naviruses repli­cate with­out us­ing a DNA in­ter­me­di­ate. So how did their N‑gene se­quences wind up be­ing part of hu­man ge­nomic DNA? The Bor­navirus genome is a sin­gle neg­a­tive-sense RNA mol­e­cule used as the tem­plate for both tran­scrip­tion of vi­ral mes­sen­ger RNAs and for pro­duc­tion of the pos­i­tive-sense antigenomes (used in turn to pro­duce more neg­a­tive-sense genomes for the prog­eny viri­ons). Uniquely among the RNA viruses, much of this process oc­curs in the nu­cleus and nu­cle­o­lus of the in­fected host cell, where it takes full ad­van­tage of host-cell RNA pro­cess­ing ma­chin­ery. This puts the Bor­navi­ral se­quences in the right place to ac­cess the cel­lu­lar genome, but still does not ex­plain how the in­for­ma­tion gets con­verted to dou­ble-stranded, in­te­grated DNA.

No­tably, some of the ELBNs con­tain poly‑A runs, con­sis­tent with an mRNA tem­plate, and some of the in­te­gra­tions are flanked by short du­pli­ca­tions of the site in ge­nomic DNA where in­ser­tion occurred—a hall­mark of retro­trans­po­son-me­di­ated in­ser­tion. Putting it all to­gether, a likely sce­nario for for­ma­tion of EBLN's is that the vi­ral mRNA is re­verse-tran­scribed and in­te­grated by a cel­lu­lar retro­trans­po­son, such as the long-in­ter­spersed nu­cleotide el­e­ments (LINEs). In the same pa­per, the Os­aka Uni­ver­sity team also re­ports de­tect­ing newly-in­serted Bor­navirus DNA se­quences in chron­i­cally in­fected, cul­tured cells, and in the brains of per­sis­tently-in­fected mice, pro­vid­ing proof-in-prin­ci­ple that this can oc­cur as a byprod­uct of Bor­navi­ral repli­ca­tion. It re­mains a mys­tery as to why only the vi­ral N pro­tein cod­ing se­quences are found in EBLNs, but the au­thors sug­gest that a chance affin­ity be­tween the N‑protein mRNA and the retro­trans­po­son repli­ca­tion ma­chin­ery prob­a­bly plays a role.

De­spite their abun­dance, the ERVs and EBLNs are un­likely to be a rep­re­sen­ta­tive sam­pling of their con­tem­po­raries. To be­gin with, the in­fec­tious event has to oc­cur within a germline cell. Thus, tro­pism plays a large role; viruses in­fect­ing the ap­pro­pri­ate tis­sues have the great­est prob­a­bil­ity of leav­ing their mark. More­over, the rare newly-in­serted vi­ral se­quence is un­likely to per­sist in the pop­u­la­tion and, if detri­men­tal to the host, may by culled by neg­a­tive se­lec­tion from the host gene pool. The process of mol­e­c­u­lar fos­sil for­ma­tion thus un­folds over the course of many host gen­er­a­tions, with those ERVs and EBLNs found in mod­ern genomes rep­re­sent­ing the im­prob­a­ble few that squeaked through. In­cred­i­bly, the hun­dreds of thou­sands of ERVs present in the hu­man genome may rep­re­sent but a tiny frac­tion of the to­tal im­pact of retro­vi­ral epi­demics on pri­mate evo­lu­tion.

That an­cient en­doge­nous el­e­ments like RELIK, SIVgml and the EBLNs are eas­ily rec­og­niz­able as rel­a­tives of mod­ern RNA viruses is more than a lit­tle sur­pris­ing. We are used to hear­ing about the con­se­quences of er­ror-prone tran­scrip­tion and lack of proof-read­ing as­so­ci­ated with poly­merases of the RNA viruses, and how these prop­er­ties fa­cil­i­tate enor­mous vari­a­tion and adapt­abil­ity. Yet mod­ern Bor­naviruses still re­sem­ble EBLNs, in some cases af­ter more than 40 mil­lion years! Such long-term sta­bil­ity, for which these fos­sils pro­vide the first di­rect ev­i­dence, high­lights the pit­falls of ap­ply­ing se­quence di­ver­gence and mu­ta­tion rates to es­ti­mate rates of RNA virus evo­lu­tion.

Don't I know you from some­where?" Source

From the per­spec­tive of host evo­lu­tion, these el­e­ments (ERVs and EBLNs) can be sources of ge­netic vari­a­tion, pro­vid­ing ad­di­tional fod­der for nat­ural se­lec­tion. There are sev­eral well-doc­u­mented ex­am­ples of ERV con­tribut­ing to for­ma­tion of new cel­lu­lar func­tions. In mice, two genes that for­merly en­coded retro­vi­ral pro­teins (Fv1 and Fv4) were con­scripted by evo­lu­tion into the ser­vice of the host and now pro­tect the cell against ex­oge­nous in­fec­tion by in­hibit­ing retro­vi­ral repli­ca­tion. Hu­man Syn­cytin, a cel­lu­lar pro­tein in­volved in pla­cen­tal mor­pho­gen­e­sis, evolved from the fu­sion pro­tein of a long-ex­tinct retro­virus. It re­mains to be seen whether any of the EBLNs, too, have taken on new roles in their hosts.

All of this leaves one won­der­ing what other vi­ral fos­sils may lurk in all that ge­nomic DNA spread across the mighty Tree of Life. It's a good bet they are there, but how do we find them? And more im­por­tantly, will we know them when we see them? Show some­one a trilo­bite fos­sil for the first time, and they will im­me­di­ately rec­og­nize what was once a liv­ing crea­ture; most will even guess ac­cu­rately that it was some sort of arthro­pod. Like trilo­bite fos­sils, ERVs and ELBNs are rec­og­niz­able for what they are, based on se­quence sim­i­lar­ity with mod­ern retro­viruses and Bor­naviruses, re­spec­tively. But what about ex­tinct viruses, viruses snuffed out long be­fore hu­mans walked the earth? What about viruses with novel genome struc­tures, or never-be­fore-seen repli­ca­tion strate­gies? In such cases, where noth­ing re­mains but a bit of DNA se­quence, might we be star­ing right at one with­out rec­og­niz­ing it for what it is (or was)?

Ref­er­ence

Horie M, Honda T, Suzuki Y, Kobayashi Y, Daito T, Os­hida T, Ikuta K, Jern P, Go­jo­bori T, Cof­fin JM, & Tomon­aga K (2010). En­doge­nous non-retro­vi­ral RNA virus el­e­ments in mam­malian genomes. Na­ture, 463 (7277), 84–7 PMID: 20054395

Welkin is As­sis­tant Pro­fes­sor of Mi­cro­bi­ol­ogy and Mol­e­c­u­lar Ge­net­ics at Har­vard Med­ical School, and an As­so­ciate Blog­ger for Small Things Con­sid­ered.