Mi­cro­tubules in the Ver­ru­comi­cro­bial Closet

by Daniel P. Haeusser

Gen­eral bi­ol­ogy text­books, like the one I use for teach­ing, of­ten de­pict the prokary­otic cell as an oval of ho­moge­nous-ap­pear­ing cy­to­plasm sur­rounded by two mem­branes. This sim­pli­fied im­age of a Gram-neg­a­tive bac­te­r­ial cell con­trasts with the de­pic­tion of a well-or­ga­nized, fea­ture-laden eu­kary­otic cell with its mem­brane-bound or­ganelles. While many would agree that the ab­sence of a mem­brane-bound nu­cleus de­fines the prokaryotic–eukaryotic bound­ary, the re­cent decade of re­search has given us a glimpse of the ac­tual or­ga­ni­za­tional com­plex­ity of bac­te­ria, which in­cludes en­cod­ing cy­toskele­tal el­e­ments. Just as eu­kary­otes have their tubu­lin, actin, and in­ter­me­di­ate fil­a­ments, the prokary­otes have FtsZ, MreB, and cres­centin (al­though the oc­cur­rence of each varies greatly).

Fig­ure 1. Global phy­logeny of the tubu­lin su­per­fam­ily. Source

The first of these prokary­otic cy­toskele­tal pro­teins to be dis­cov­ered, FtsZ, is a dis­tantly-re­lated mem­ber of the tubu­lin fam­ily (Fig­ure 1). FtsZ is a highly con­served pro­tein in­volved in cy­toki­ne­sis in most bac­te­ria, cer­tain ar­chaea, chloro­plasts, and in the mi­to­chon­dria of some pro­tists. Al­though FtsZ shows only ~17% amino acid iden­tity with tubu­lin, their crys­tal struc­tures are re­mark­ably sim­i­lar (a fine ex­am­ple of struc­ture trump­ing se­quence).  No­tably, FtsZ, un­like tubu­lins, is un­able to form mi­cro­tubule-like struc­tures; in­stead it just self-as­so­ciates into loosely bun­dled protofil­a­ments. Al­though a re­cent pa­per shows that FtsZ forms tubule-like struc­tures in vitro in the pres­ence of a bind­ing partner/assembly reg­u­la­tor, these tubules are quite dis­tinct from eu­kary­otic mi­cro­tubules in both size and struc­ture, and their in vivo rel­e­vance is un­cer­tain.

For many years FtsZ re­mained the only mem­ber of the tubu­lin fam­ily to be iden­ti­fied in prokary­otes. How­ever, re­ports of mi­cro­tubule-like struc­tures ob­served in spe­cific bac­te­r­ial species have con­tin­ued to ap­pear (re­viewed here). As fur­ther ev­i­dence of their tubu­lin na­ture, many of them have been found to be sen­si­tive to tubu­lin de­poly­mer­iza­tion agents and to re­act with anti-tubu­lin an­ti­bod­ies. Un­for­tu­nately, none of these ob­served struc­tures have been matched to par­tic­u­lar genes, so their ge­netic iden­tity re­mains a mys­tery.

Fig­ure 2. TEM of P. de­jongeii show­ing a pros­the­cae (PT) with in­set high­light­ing the in­tra­cy­to­plas­mic mem­brane (ICM): an ex­am­ple of prokary­otic com­part­men­tal­iza­tion. Source

The study of bac­te­r­ial tubu­lin re­ceived a tremen­dous boon in 2002 with the dis­cov­ery of true α- and β‑tubulin ho­mo­logues (named BtubA and BtubB) in bac­te­ria. These pro­teins share 37% se­quence iden­tity with tubu­lin al­though they clus­ter phy­lo­ge­net­i­cally in their own group. They were found in the Pros­the­cobac­ter, a free-liv­ing genus in the Planc­to­mycetes-Ver­ru­comi­cro­bia-Chlamy­diae (PVC) su­per­phy­lum that has been de­bated as be­ing an evo­lu­tion­ary link to eu­kary­otes. Pros­the­cobac­ter are mor­pho­log­i­cally de­fined (and named) by an elon­gated po­lar ap­pendage (the pros­the­cae) (Fig­ure 2).

The dis­cov­ery of BtubA/B nat­u­rally led re­searchers to in­ves­ti­gate whether the pro­teins formed mi­cro­tubule-like struc­tures in a Pros­the­cobac­ter cell. Dis­ap­point­ingly, no such struc­tures were seen ei­ther in vivo or with in vitro-as­sem­bled BtubA/B. Thus, de­spite the ho­mol­ogy be­tween BtubA/B and tubu­lin, it ap­peared as if they were as in­ca­pable of form­ing true mi­cro­tubules as FtsZ.

Fig­ure 3a. BtubA/B as­sem­ble into five-protofil­a­ment mi­cro­tubules. Shown here is an 11 nm thick cry­oto­mo­gram slice of P. van­neer­venii with mi­cro­tubule struc­ture bun­dles (ar­rows), and in 3D re­con­struc­tion. Bar = 100 nm. Source

This has now changed. As is of­ten the case, new tech­niques and imag­ing ca­pa­bil­i­ties have re­vealed struc­tures that pre­vi­ous stud­ies failed to vi­su­al­ize, specif­i­cally BtubA/B struc­tures that may have been de­stroyed by con­ven­tional elec­tron mi­croscopy sam­ple prepa­ra­tion. Last month, re­searchers us­ing elec­tron cry­oto­mog­ra­phy (ECT) pub­lished a re­port of BtubA/B mi­cro­tubule-like struc­tures in sev­eral Pros­the­cobac­ter species (Fig­ure 3). These mi­cro­tubules run par­al­lel to the mem­brane, in­di­vid­u­ally or in small bun­dles, pre­dom­i­nantly in the cell stalk or in the re­gion where the stalk tran­si­tions into the cell body. The au­thors would have liked to demon­strate that the mi­cro­tubules con­sisted of BtubA/B by knock­ing out the btub lo­cus and ob­serv­ing the loss of the tubules, but the re­quired ge­netic sys­tem of ma­nip­u­la­tion is not yet avail­able. For­tu­nately, an­other species could sub­sti­tute for the nonex­is­tent mu­tants. This species, P. flu­vi­atilis, lacks the btub genes, makes no BtubA/B, and pos­sesses no dis­cernible mi­cro­tubules. More ev­i­dence: mi­cro­tubules formed when BtubA/B were ex­pressed in E. coli, and they could also be as­sem­bled in vitro.

In ECT to­mo­grams, the Pros­the­cobac­ter mi­cro­tubules are ~200 – 1200 nm long, with an av­er­age di­am­e­ter of 7.6 nm and a re­peat dis­tance of 4.4 nm (Fig­ure 3a). No­tably, these di­men­sions are sim­i­lar whether the mi­cro­tubules are pro­duced by Pros­the­cobac­ter species, ex­pressed in en­gi­neered E. coli, or as­sem­bled in vitro. How does their fine struc­ture com­pare to canon­i­cal eu­kary­otic mi­cro­tubules? The pub­lished crys­tal struc­ture of the BtubA/B dimer fits best with a five-protofil­a­ment tubule model with 4.6 nm spac­ing (Fig­ure 3b). In con­trast, the num­ber of protofil­a­ments in eu­kary­otic mi­cro­tubules is usu­ally 13. Spac­ing, how­ever, is con­served (~5 nm in eu­kary­otes), which sug­gests that in both groups the fil­a­ments form one-start he­li­cal tubules.

Fig­ure 3b. Pseudo-atomic model of five protofil­a­ment bac­te­r­ial mi­cro­tubule (blue; built from Pro­tein Data Bank ) su­per­im­posed on the im­age of a cryo-sec­tioned BtubA/B tube (black). Bar = 10 nm. Source

One might ex­pect that BtubA/B or re­lated genes could be found in other bac­te­ria as well, par­tic­u­larly con­sid­er­ing that mi­cro­tubule-like struc­tures have been ob­served in sev­eral other species. For ex­am­ple, con­sider the un­cul­tured ec­tosym­bionts of Eu­plo­tid­ium cil­i­ates called epixeno­somes (mean­ing ex­ter­nal alien body), an amaz­ing de­fense sys­tem fea­tured in this blog sev­eral years ago. The epixeno­somes are also mem­bers of the phy­lum Ver­ru­comi­cro­bia, and they con­tain ob­serv­able mi­cro­tubules. How­ever, the lack of a ge­netic sys­tem again makes ge­netic ver­i­fi­ca­tion im­prac­ti­cal, thus their ac­tual re­la­tion to BtubA/B, or the tubu­lin fam­ily in gen­eral, is still un­clear.

So, where do BtubA/B fit evo­lu­tion­ar­ily within the tubu­lin fam­ily (Fig­ure 1)? On the one hand, they share traits with FtsZ that tubu­lin lacks, namely chap­er­one-in­de­pen­dent fold­ing and weak dimer­iza­tion. How­ever, BtubA/B can form het­erodimers, some­thing that is gen­er­ally not pos­si­ble for FtsZ, ex­cept in rare species or in chloro­plasts that en­code more than one type of FtsZ. The au­thors of the ECT re­port agree with the sug­ges­tion that BtubA/B rep­re­sent an an­cient form of tubu­lin that pre­ceded the larger, 13-protofil­a­ment mi­cro­tubules seen in eu­kary­otes. The unique pres­ence of BtubA/B in Pros­the­cobac­ter and re­lated species sug­gests that they orig­i­nated via hor­i­zon­tal gene trans­fer, but from whence they came re­mains enig­matic.

Fig­ure 4. Im­muno­flu­o­res­cent stain­ing of BtubA/B co­ex­pressed in E. coli. Bar = 2 μm. Source

The big ques­tion re­main­ing for BtubA/B and the mi­cro­tubules they make is what func­tion they serve. An early hy­poth­e­sis, given the re­duced genomes of many PVC su­per­phy­lum mem­bers, was that they func­tioned in cell di­vi­sion in place of FtsZ. Sub­se­quent work showed, how­ever, that Pros­the­cobac­ter do en­code ftsZ, mak­ing a role for BtubA/B in cell di­vi­sion less likely. How­ever, Pros­the­cobac­ter FtsZ di­verges in key residues im­por­tant for as­sem­bly and GTP hy­drol­y­sis. Sim­i­larly, the Thau­mar­chaeota are known to en­code ftsZ with a highly di­ver­gent tubulin/FtsZ sig­na­ture se­quences, and they uti­lize the Cdv sys­tem for cy­toki­ne­sis in­stead. Al­ter­na­tively, given the lo­cal­iza­tion of BtubA/B ob­served in Pros­the­cobac­ter, one might ex­pect they con­tribute to the struc­ture or for­ma­tion of the pros­the­cae, but sim­i­larly-shaped species ex­ist with­out the btub genes or ob­serv­able mi­cro­tubu­lar struc­tures (e.g. P. flu­vi­atilis). When co-ex­pressed in E. coli, BtubA/B form mi­cro­tubules that run par­al­lel to the mem­brane the en­tire length of the cell (Fig­ure 4), but I know of no spe­cific phe­no­type that they con­fer out­side of those that could oc­cur from gen­eral for­eign pro­tein over­ex­pres­sion.

The de­vel­op­ment of a ge­netic sys­tem in a species that en­codes BtubA/B would be ideal, but un­til then the study of ex­ist­ing model sys­tems may be the only route for fur­ther ex­plo­ration of BtubA/B as­sem­bly and func­tion. Re­gard­less, the unique prop­er­ties of BtubA/B in shar­ing char­ac­ter­is­tics of both tubu­lin and FtsZ make even their in vitro study highly rel­e­vant to the un­der­stand­ing of this cy­toskele­tal fam­ily. Such stud­ies may an­swer pend­ing ques­tions about the reg­u­la­tion of bac­te­r­ial mi­cro­tubule as­sem­bly, as well as their di­rec­tion­al­ity, dy­namic in­sta­bil­ity, and ac­ces­sory pro­tein requirements—all prop­er­ties that the more di­ver­gent tubu­lin-fam­ily mem­ber FtsZ ap­pears to lack. An­other dis­tant tubu­lin-fam­ily mem­ber, TubZ (Fig­ure 1) from a vir­u­lence plas­mid of B. thuringien­sis, does dis­play tread­milling with­out any ap­par­ent tubule for­ma­tion, and will aid such com­par­a­tive stud­ies within the tubu­lin fam­ily.

Fi­nally, it leads one to won­der: What other eu­kary­otic gar­ments are re­sid­ing uniden­ti­fied and un­char­ac­ter­ized in the prokary­otic closet?
 

Ref­er­ence

Pil­hofer M, Ladin­sky MS, Mc­Dowall AW, Petroni G, Jensen GJ (2011). Mi­cro­tubules in bac­te­ria: an­cient tubu­lins build a five-protofil­a­ment ho­molog of the eu­kary­otic cy­toskele­ton. PLoS Bi­ol­ogy, 9 (12). PMID 22162949

Daniel is a post­doc­toral fel­low in the Mar­golin Lab in the De­part­ment of Mi­cro­bi­ol­ogy & Mol­e­c­u­lar Ge­net­ics at the Uni­ver­sity of Texas, Hous­ton Med­ical School. He also teaches as an ad­junct pro­fes­sor at the Uni­ver­sity of Hous­ton-Down­town in the De­part­ment of Nat­ural Sci­ences. In ad­di­tion to sci­ence, he en­joys read­ing, writ­ing, and film. He can be found on Google+.