Sex And Evo­lu­tion, Vib­rio Style

by J. Jef­frey Mor­ris

Fam­ily trees: Ancestry.com vs. the mi­cro­bial di­as­pora

Fig­ure 1. Two ways of think­ing about fam­ily trees. Both trees rep­re­sent 3 gen­er­a­tions of the Simp­son fam­ily. In A, we see a Bart-cen­tric tree, show­ing Bart and all of his di­rect-line an­ces­tors. In B, we see a com­mon-an­ces­tor ori­ented tree, show­ing all the de­scen­dants of Jacque­line Bou­vier (in­clud­ing Selma's adopted baby Ling). Source

My wife is an avid ge­neal­o­gist and I'm an evo­lu­tion­ary bi­ol­o­gist, so fam­ily trees come up for dis­cus­sion from time to time around the din­ner ta­ble. Laura al­ways ima­gi­nes fam­ily trees as ex­tend­ing up­ward from a cur­rently liv­ing per­son to in­clude all their an­ces­tors, whereas I al­ways think of them as point­ing down­ward from a com­mon an­ces­tor to all their des­cen­dants (Fig 1). Ob­vi­ously, I think my way is the right way – just try to fig­ure out what a "sec­ond cousin thrice re­moved" is with­out con­sid­er­ing com­mon an­ces­try – but in the end they both re­flect the fact that we are all the prod­ucts of a ver­ti­cal line of des­cent from a se­ries of moth­ers and fa­thers. Our most dis­tant di­rect line an­ces­tors weren't hu­man of course, but even those an­cient cho­a­no­fla­gel­late moms and dads 500 mil­lion years ago were mak­ing ba­bies in a man­ner that was very sim­i­lar (at least on a mo­le­cu­lar level) to how we do it to­day.

When sci­en­tists first started piec­ing to­gether the evo­lu­tion­ary re­la­tion­ships be­tween mi­crobes, they as­sumed that the same kind of ver­ti­cal de­scent hap­pened with those tiny crit­ters as well. The first mol­e­c­u­lar at­tempts to­wards un­der­stand­ing the mi­cro­bial fam­ily tree used ri­bo­so­mal RNA genes, which hap­pened to fit nicely into the par­a­digm of ver­ti­cal de­scent – the rRNA genes of sim­i­lar bac­te­ria where more sim­i­lar to each other than they were to more dis­sim­i­lar bac­te­ria, just as we would ex­pect if the se­quences be­came more di­ver­gent as time passed af­ter two species di­verged from a com­mon an­ces­tor.

Fig­ure 2. If Wood­land Crit­ters Un­der­went Ho­ri­zon­tal Jean (sp) Trans­fer. Art­work by Sarah J. Ad­kins.

But it turned out that the choice of rRNA as an evo­lu­tio­nary marker was some­what for­tu­itous in this re­gard, be­cause many other genes have fam­ily trees very dif­fe­rent from what we would ex­pect based on the rRNA phy­logeny. As we looked at more and more genes, it be­came clear that, in the mi­cro­bial world, hor­i­zon­tal trans­mis­sion of genes was hap­pen­ing all the time. This process, al­most com­pletely ab­sent in big or­gan­isms like us, fun­da­men­tally changes how evo­lu­tion works in mi­crobes. It's re­ally a fun­da­men­tally weird idea – just con­si­der what your lo­cal for­est might look like if ho­ri­zon­tal ge­netic trans­fer (HGT) was com­mon­place in the an­i­mal king­dom (Fig. 2)!

There are lots of ways that mi­crobes can pick up for­eign DNA. Some mi­crobes can "eat" naked DNA from the en­vi­ron­ment, and some­times viruses ac­ci­den­tally carry mi­cro­bial DNA in­stead of their own genomes. Per­haps the most well-known vec­tors of HGT are plas­mids, small cir­cu­lar chro­mo­somes that repli­cate in­de­pen­dently of the microbe's pri­mary chromosome(s). We know a lot about these mech­a­nisms of HGT be­cause re­searchers have har­nessed them for ge­netic en­gi­neer­ing and syn­thetic bi­ol­ogy, but we know much less about HGT in the wild, where most of our ef­forts have fo­cused on plas­mids that move an­tibi­otic re­sis­tance genes be­tween path­o­genic mi­crobes. It's still an open ques­tion how of­ten HGT oc­curs in the wild, and what the most im­por­tant vec­tors are for mov­ing genes around.

Prob­ing the mys­ter­ies of mi­cro­bial pop­u­la­tion ge­net­ics with ma­rine vib­rios

A decade-long se­ries of ex­per­i­ments led by Mar­tin Polz of the MIT has pro­vided some of the first steps to­ward un­der­stand­ing HGT in na­ture. Polz and co-work­ers iso­lated sev­eral hun­dred strains of Vib­rio bac­te­ria from a sin­gle wa­ter sam­ple taken from the coastal At­lantic Ocean. These strains were as­so­ci­ated with dif­fer­ent size par­ti­cles con­tained in the wa­ter sam­ple, rep­re­sent­ing ei­ther free-liv­ing bac­te­ri­o­plank­ton or else or­gan­isms liv­ing on pieces of de­tri­tus (e.g. al­gal slime or fish poop) sus­pended in the wa­ter (Fig. 3). In other words, the strains came from very dif­fer­ent eco­log­i­cal set­tings – imag­ine the dif­fer­ence be­tween free-swim­ming squid and bot­tom-dwelling clams.

Fig­ure 3. Dif­fer­ent size par­ti­cles have fun­da­men­tal­ly dif­fer­ent mi­cro­bial ecolo­gies. Pro­duc­t­ion of iron-bind­ing siderophores, a leaky pub­lic good, was un­evenly par­ti­tioned be­tween dif­fe­rent size par­ti­cles in the same wa­ter sam­ple. "Chea­ters" were more com­mon on big­ger par­tic­les, pre­sum­ably be­cause they could mooch off of more neigh­bors than they could if they were on smaller par­ti­cles. Source

All of Polz's Vib­rio iso­lates were very closely re­lated to each other based on their ri­bo­so­mal RNA se­quences, but care­ful analy­sis of other genes re­vealed that strains from the same eco­log­i­cal set­ting were more re­lated to each other than they were to strains from dif­fer­ent set­tings (Hunt et al., 2008). Ex­am­i­na­tion of whole genome se­quen­ces of se­lect iso­lates pre­sented an even clearer pic­ture: al­though the genomes of all the iso­lates were very sim­i­lar, cer­tain mu­tant ver­sions of genes ex­isted in all free-liv­ing iso­lates but no par­ti­cle-as­so­ci­ated iso­lates (and vice versa), sug­gest­ing that these "ecoS­NPs" were adap­tive in one but not the other eco­log­i­cal set­ting. In con­trast, many more gene vari­ants were found dis­tri­bu­ted be­tween mem­bers of both size frac­tions. The sim­plest way to ex­plain this pat­tern is that HGT was oc­cur­ring rapidly be­tween these Vib­rio strains, but it was more rapid be­tween in­hab­i­tants of the same eco­log­i­cal set­ting (Fig. 4).  In fact, HGT was so com­mon in these or­gan­isms that their genomes looked as one would ex­pect from sex­u­ally re­pro­duc­ing or­gan­isms that un­dergo chro­mo­so­mal re­com­bi­na­tion dur­ing meioi­sis!

Tiny chro­mo­somes shed new light on hor­i­zon­tal gene trans­fer

A re­cent pa­per by Hong Xue, a grad­u­ate stu­dent in the Polz lab, pro­vides new ev­i­dence about how HGT oc­curs in this com­mu­nity.  Xue and co-au­thors ex­tracted and se­quenced all the small, self-repli­cat­ing chro­mo­somes (epi­somes) from the PolzVib­rio col­lec­tion and found many po­ten­tial vec­tors of HGT. The ma­jor­ity of epi­somes ap­peared to be cir­cu­lar genomes from lyso­genic viruses, but plas­mids were also abun­dant. Ev­i­dence sug­gests that these epi­somes spread even more rapidly by HGT than the ge­nomic mu­ta­tions dis­cussed pre­vi­ously. When plas­mids were sorted into groups based on gene con­tent, the se­quences of the plas­mids within a given group were > 98% sim­i­lar. Since DNA ac­cu­mu­lates cer­tain kinds of mu­ta­tions at a steady, clock-like rate, this means that not much time has passed since these plas­mids first ap­peared in the pop­u­la­tion. It might have only taken a few months, maybe a cou­ple of years, for these plas­mids to in­fect these pop­u­la­tions. In­ter­est­ingly, the virus genomes had much more se­quence di­ver­sity, sug­gest­ing that they only be­come vir­u­lent – and thus spread be­tween cells – in­fre­quently.

Fig­ure 4. Hor­i­zon­tal gene flow is slower bet­ween than within eco­logically dif­fer­ent po­pu­lations. At some point, an an­ces­tral pop­u­la­tion (pur­ple) split into two dif­fer­ent pop­u­la­tions (green and red), for in­stance by col­o­niz­ing se­pa­ra­te size par­ti­cles. As they evolve in sep­a­rate habi­tats, they fre­quently ex­change DNA with their own pop­u­la­tion, but only rarely with the other pop­u­la­tion. Over time, this leads to clearer and clearer se­pa­ra­tion be­tween gene con­tent in the two po­pu­la­tions. Source

It's pretty ob­vi­ous how the virus genomes spread bet­ween cells – they oc­ca­sion­ally be­come vir­u­lent, killing their hosts and in­fect­ing neigh­bor­ing cells. The story is a lit­tle more com­plex with the plas­mids, how­ever.  Only about 1/6 of the plas­mids were "con­jugal" – i.e., they con­tained all the genes nec­es­sary to trans­mit them­selves from one cell to an­other. An­other 1/5 of the plas­mids had se­quences that would make them "mo­bi­liz­able" as long as they hap­pened to be in the same cell as a con­ju­gal plas­mid. How­ever, the ma­jor­ity of the plas­mids had no clear mech­a­nism for cell-to-cell trans­mis­sion. There were no sig­nif­i­cant dif­fer­ences in within-group se­quence di­ver­gence be­tween these classes of plas­mid, though, mean­ing that the "non-trans­miss­able" plas­mids move around just as fast as the trans­mis­si­ble ones. This means we still don't know ex­actly how HGT op­er­ates in these pop­u­la­tions, and there are more big dis­cov­er­ies to be made in the pop­u­la­tion ge­nomics of nat­ural mi­cro­bial com­mu­ni­ties.

What do epi­somes do for their hosts?

The­o­reti­cians of­ten as­sume that HGT is more rapid in dense pop­u­la­tions be­cause di­rect con­tact be­tween cells should hap­pen more of­ten. Thus, one might ex­pect to see more epi­somes in par­ti­cle-as­so­ci­ated Vib­rios than in the free-liv­ing groups. But the op­po­site was true: there were sig­nif­i­cantly more epi­somes amongst free-liv­ing iso­lates. Why? Plas­mids of­ten car­ried com­plete sets of gene for use­ful meta­bolic func­tions such as amino acid me­tab­o­lism and iron ac­qui­si­tion. These path­ways yield prod­ucts that are po­ten­tially avail­able to nearby or­gan­isms, open­ing the pos­si­bil­ity that "cheaters" who don't pay the cost to make the prod­ucts can co­ex­ist with pro­duc­ers. Pre­vi­ous work from the Polz lab showed that iron ac­qui­si­tion cheaters did ex­ist amongst these Vib­rio iso­lates, and they were more com­mon in the par­ti­cle-as­so­ci­ated strains. Thus, it might be the case that eas­ily-gained and eas­ily-lost epi­somes carry con­di­tion­ally use­ful leaky path­ways – things that might be use­ful un­der one eco­log­i­cal set­ting, but dis­pens­able in an­other.

An on­go­ing re­source for ma­rine mi­cro­bi­ol­ogy

Big an­i­mals like us are bi­ased to­ward think­ing about the nat­ural world and its dy­nam­ics on a big an­i­mal scale.  But mi­crobes don't play by our rules, and their rapid evo­lu­tion­ary changes wreak havoc on our de­sire to squeeze them into clas­si­cal schemes of tax­o­nomic or­ga­ni­za­tion. Mod­ern se­quenc­ing tech­nolo­gies are only just now start­ing to pro­vide the tools nec­es­sary to study the pop­u­la­tion ge­net­ics of mi­crobes. The PolzVib­rio col­lec­tion is one of the first at­tempts to probe a sin­gle well-de­fined mi­cro­bial com­mu­nity at the depth nec­es­sary to fig­ure out what's re­ally go­ing on out there. The num­ber of in­sights that have come from this col­lec­tion in the first 10 years of its ex­is­tence – and the fact that you can ac­tu­ally go back out to the col­lec­tion site and find the "same" or­gan­isms again – sug­gests that Polz has cre­ated a re­source that will con­tinue to im­prove (dare I say rev­o­lu­tion­ize?) our un­der­stand­ing of mi­cro­bial evo­lu­tion and ecol­ogy.

 

Jeff Mor­ris is As­sis­tant Pro­fes­sor at the Uni­ver­sity of Al­abama, Birm­ing­ham. His re­search in­ter­ests are pretty much coter­mi­nous with those of Mor­ris­lab. He's also very in­ter­ested in the his­tory of the hu­man species, and sus­pects that if he had it all to do over again he might have cho­sen to study mol­e­c­u­lar an­thro­pol­ogy in­stead of mi­cro­bi­ol­ogy.