A Bac­terium Learns Long Di­vi­sion

by Nanne Nan­ninga

The com­mon pic­ture of a di­vid­ing rod-shaped bac­terium en­com­passes the po­si­tion­ing of the di­visome, in­clud­ing an FtsZ-ring, in the cell cen­ter. This oc­curs af­ter the cell has dou­bled its length with­out in­creas­ing its di­am­e­ter. Con­versely, in­crease in di­am­e­ter with­out cell elon­ga­tion would seem highly un­likely in a rod-shaped or­gan­ism. Yet, this hap­pens.

Fig­ure 1: SEM of tightly ap­posed ecto­symbionts on the sur­face of L. oneis­tus. Elec­tron mi­cro­graph by N. Leisch.

In fact, this is the nor­mal con­di­tion for an ec­tosym­bi­otic bac­terium that lines the sur­face of the ma­rine ne­ma­tode Laxus oneis­tus. The sym­biont is a g‑proteobacterium like E. coli, but un­like E. coli it has not been cul­tured out­side its nat­ural habi­tat. As orig­i­nally de­scribed by Polz et al. in the early nineties, the bac­te­ria are po­si­tioned perpendi­cular to the sur­face of L. oneistus. They are glued to the sur­face of the ne­ma­tode by a C‑type lectin. The orig­i­nal pa­per al­ready in­di­cated that di­vi­sion takes place longi­tu­dinally, with one of the daugh­ters pre­sumed to re­main at­tached to the ne­ma­tode. This makes sense be­cause oth­er­wise the daugh­ter cells would get lost to the en­vi­ron­ment. Re­call that this phe­nom­e­non lies at the ba­sis of the Helm­stet­ter-Cooper baby ma­chine, whereby one of a pair of new­born cells is se­lected to start a syn­chro­nous cul­ture.

Vol­umes plot­ted ver­sus length (A) or width (B) of L. oneis­tus ec­tosym­biont (black) or E. coli (red) cells. (A) Cell length shows a di­rect lin­ear cor­re­la­tion to cell vol­ume in E. coli (n = 5,021), but not in the ec­tosym­biont (n = 4,004). (B) Cell width shows a di­rect lin­ear cor­re­la­tion to cell vol­ume in the ec­tosym­biont, but not in E. coli. (C–H) Con­fo­cal laser scan­ning mi­croscopy im­ages show­ing FtsZ (green) lo­cal­iza­tion in non-di­vid­ing © and di­vid­ing (D–H) L. oneis­tus ec­tosym­biont cells. For each di­vi­sion stage the rep­re­sen­ta­tive lo­cal­iza­tion pat­tern ob­served in >25 cells is dis­played (bot­tom), with the cor­re­spond­ing dif­fer­en­tial con­trast im­age (top). In an ec­tosym­biont not dis­play­ing any mem­brane con­stric­tion ©, FtsZ flu­o­res­cence con­cen­trates at one pole. In a sym­biont with po­lar in­den­ta­tions, FtsZ ap­pears as a me­di­ally lo­cal­ized el­lip­ti­cal ring (D), which con­stricts at later fis­sion stages (E–G) and to a max­i­mum im­me­di­ately be­fore sep­a­ra­tion of cell daugh­ters (H). In­sert in (G) shows a lat­eral view of the FtsZ ring. Ar­row­heads point to slight po­lar in­den­ta­tions, ar­rows to read­ily vis­i­ble con­stric­tions, as­ter­isks to deep con­stric­tions. Dashed lines rep­re­sent cell out­line. Scale bar is 1 μm. (I) Cell length and lon­gi­tu­di­nal FtsZ flu­o­res­cence pro­files of 2,358 ec­tosym­biont cells sorted by in­creas­ing cell width, from left to right. The green out­line rep­re­sents the cell length and the heat map rep­re­sents the FtsZ flu­o­res­cence in­ten­sity of each cell along its length. The left­most bar rep­re­sents the thinnest cell, the right­most bar the widest cell. The cal­i­bra­tion bar dis­plays FtsZ flu­o­res­cence in­ten­sity in ar­bi­trary units (a.u.). Source

Re­cently, Leisch and col­leagues have ex­tended these ob­ser­va­tions by care­fully de­ter­min­ing cellu­lar di­men­sions and vi­su­al­iz­ing the FtsZ di­vi­sion pro­tein with flu­o­res­cent E. coli mon­o­clonal anti­bodies. The re­sults can be com­pared with E. coli data (Fig­ure 2A, B). Whereas E. coli elon­gates as ex­pected, this is not what hap­pens with the sym­biont. The sym­biont does not change its length (Fig­ure 2A) but in­creases its di­am­e­ter (Fig­ure 2B). Consequently,the sym­biont di­vides longitudi­nal­ly, (Fig. 2 C‑H). Im­munos­tain­ing of FtsZ re­veals that FtsZ po­si­tion­ing cor­re­lates with the cell con­striction. In fact, an el­lip­soidal FtsZ-ring is ob­served stretched along the length of the ecto­sym­biont.

These re­mark­able ob­ser­va­tions pro­vide more ques­tions than an­swers. For in­stance, how does the di­am­e­ter of the sym­biont in­crease? Is the in­crease the same in all direct­ions or is it po­lar­ized? Though, the au­thors do not dis­cuss this point, it would seem that the bac­terium has the shape of a shoe­box. Thus, in­crease in width im­plies wi­den­ing of the shoe­box with­out al­ter­ing its height. How is the shoe­box-shape, if ap­plic­a­ble, main­tained? Further­more, how is DNA seg­re­ga­tion car­ried out in the ab­sence of cell elon­ga­tion? How are the nu­cleoids arranged? It would seem that widen­ing of the shoe­box is suf­fi­cient. Again the spa­tial mech­a­nism is not known. As pointed out by the au­thors, genes en­cod­ing Min, C, D and E are also present. Do they func­tion in the sym­biont, and if so how do they con­form to dif­fer­ent geom­e­try? An­swers to these ques­tions are likely to elu­ci­date the de­grees of free­dom in bac­te­r­ial cell di­vi­sion. But that's what we learned in our arith­metic class about the long di­vi­sion. Un­less one deals with a whole num­ber, there is al­ways a re­main­der.

Prof. Nan­ninga is Emer­i­tus Pro­fes­sor of Mol­e­c­u­lar Cy­tol­ogy at the Uni­ver­sity of Am­s­ter­dam Swammer­dam In­sti­tute for Life Sci­ences