A Fas­tid­i­ous Bac­te­rio­phage

by Michael Yarmolin­sky

Pa­trons of up­scale seafood restau­rants are given the op­por­tu­nity to see that the un­for­tu­nate crea­tures des­tined for the lob­ster pot are wav­ing their an­ten­nae about. Savvy cus­tomers at down­scale seafood mar­kets eval­u­ate ques­tion­able claims of fresh­ness by smell. A fas­tid­i­ous bac­te­rio­phage would wel­come the op­por­tu­nity to gauge the qual­ity of a po­ten­tial meal, if only it could make that as­sess­ment. I was re­cently re­minded, in the course of dis­pos­ing of old reprints, that a bac­te­rio­phage named Chi can do so. It at­tacks only motile strains of bac­te­ria, and then only if the fla­gella are ac­tive. How is this cir­cum­spect ap­praisal ac­com­plished?

Phage Chi (X 220,000). Source

It has been sev­eral years since the pub­li­ca­tion of con­vinc­ing sup­port for a me­chan­i­cal model that can ac­count for the dis­crim­i­na­tion ex­hib­ited by Chi. Why re­visit it now? First, for the ben­e­fit of those who have over­looked the re­mark­able story of Chi's fas­tid­i­ous be­hav­ior and sec­ond be­cause I sus­pect that we have learned only the half of it.

What sort of phage is Chi?

Chi is a vir­u­lent, dou­ble-stranded DNA phage with a long, un­even his­tory and a long, ta­pered tail end­ing in a sin­gle, long, kinky tail fiber. It was first char­ac­ter­ized in 1936 in the lab­o­ra­tory of Félix d'Herelle, the ad­ven­turer-sci­en­tist who is jointly cred­ited with the dis­cov­ery of the viruses he des­ig­nated "bac­te­rio­phages." Chi's genome (60 kb) has re­cently been se­quenced (An­drew Kropin­ski, Kelly Hughes and Roger Hen­drix, in prepa­ra­tion).

The fla­gel­lar or­ganelle. Source

The ini­tial re­port showed that Chi at­tacks only fla­gel­lated bac­te­ria. Growth of sen­si­tive strains on agar con­tain­ing phe­nol, at a con­cen­tra­tion known to pre­vent the de­vel­op­ment of fla­gella, ren­dered the bac­te­ria Chi-re­sis­tant. Chi-re­sis­tance could be used to se­lect mu­tants de­fec­tive in fla­gel­la­tion. Mu­tants al­tered in sev­eral of the 44 known fla­gel­lar genes of Sal­mo­nella were sub­se­quently se­lected in this way.

What sort of or­ganelle is a fla­gel­lum?

Fla­gella are long, thin, he­li­cal fil­a­ments com­monly much longer than the bac­te­r­ial body from which they emerge at var­i­ous sites. Their some­what flex­i­ble, si­nu­soidal ap­pear­ance was in­ter­preted, un­til the early 70s, as ev­i­dence that they act like whips, but their he­lic­ity is in­trin­sic and their ac­tion that of a pro­peller ro­tated by a mo­tor em­bed­ded in the cell body. Nor­mal ro­ta­tion speeds for Es­cherichia coli fla­gella are around 6000 rpm, but a record speed, set by a vig­or­ous Vib­rio, is 100,000 rpm. Each fla­gel­lum is both a re­versible mo­tor or­ganelle and a pro­tein ex­port and as­sem­bly ap­pa­ra­tus that fab­ri­cates the ex­ter­nal fil­a­ment by ex­trud­ing fla­gellin monomers through a cen­tral chan­nel and adding them to the grow­ing fla­gel­lum at its very tip.

The base of a fla­gel­lum. Com­puter-gen­er­ated im­age by Kei­ichi Namba. Source

The walls of a fla­gel­lum (in Es­cherichia and Sal­mo­nella) are com­posed of 11 he­li­cal protofil­a­ments; each can as­sume one of two con­for­ma­tions, L and R, re­sult­ing in 12 poly­mor­phic al­ter­na­tives. The nor­mal ra­tio is 2 R to 9 L forms, but the pro­por­tion is al­tered by the tor­sion gen­er­ated dur­ing ro­ta­tion. Any protofil­a­ment in­ho­mo­gene­ity con­torts the fil­a­ment into a he­lix (a su­per­he­lix). Mu­ta­tion­ally al­tered fla­gella with all L or all R protofil­a­ments are straight and their ro­ta­tion pro­duces no motil­ity. Two right-handed he­li­cal grooves run the length of the fla­gel­lum, ex­tend­ing into the basal hook struc­ture that at­taches the fil­a­ment to the ro­tor. The pitch of these grooves tight­ens as the pro­por­tion of R fil­a­ments in­creases. (We'll see the im­por­tance of this in a mo­ment.)

Ab­sent any stim­u­lus, each fla­gel­lum al­ter­nates sto­chas­ti­cally be­tween in­ter­vals of coun­ter­clock­wise (CCW) and clock­wise (CW) ro­ta­tion. When ro­tat­ing CCW, the fla­gella of a cell form a bun­dle and pro­pel the bac­te­ria for­ward; when ro­tat­ing CW, fla­gella leave the bun­dle and in­duce tum­bling.

Movies of motile E. coli. Over­all source. a. Flu­o­res­cent fil­a­ment leav­ing bun­dle. Source. b. Flu­o­res­cent bun­dles 500 rev/sec. Source. c. Teth­ered E. coli al­ter­nat­ing di­rec­tions of ro­ta­tion. Source

A typ­i­cal run of smooth swim­ming lasts about a sec­ond or two and speeds the bac­terium along at ten to twenty body lengths per sec­ond. Dur­ing this "run," a small change in an en­vi­ron­men­tal stim­u­lus can al­ter the prob­a­bil­ity of con­tin­u­ing the run for sev­eral sec­onds (fa­vored by at­trac­tants) or tum­bling (fa­vored by re­pel­lents). Thus, the ba­sis for chemo­tac­tic be­hav­ior is the mod­u­la­tion by en­vi­ron­men­tal stim­uli of a bi­nary switch con­trol­ling a bal­ance be­tween or­der and chaos. Po­tent pa­ra­me­ters those.

What is it about ac­tive fla­gella that Chi senses?

A ma­jor study of host sen­si­tiv­ity to Chi in­fec­tion was pub­lished by Eli­nor Meynell, 1961. Her analy­sis showed that fla­gel­lated but non-motile strains of Sal­mo­nella are Chi-re­sis­tant, and that the phage ad­sorbs very poorly to var­i­ously par­a­lyzed bac­te­ria and not at all to iso­lated fla­gella. Elec­tron mi­cro­graphs in­cluded in her ar­ti­cle showed Chi phages ad­sorbed to dis­tal parts of fla­gella and gave no ev­i­dence of their drift to­ward the cell body with in­creas­ing time post ad­sorp­tion. There­fore, Meynell fa­vored the hy­poth­e­sis that the ad­sorbed phages in­ject their DNA into the fla­gel­lar chan­nel, thence into the bac­terium.

left: In­ter­ac­tion of phage Chi with E. coli. Chi at­tached to fla­gel­lar fil­a­ments via their tail fibers (X 120,000). right: Chi (heads emp­tied) at­tached at the base of a fla­gel­lum (X 150,000). Source

Sub­se­quent EM stud­ies by Sil­via Schade, Julius Adler and Hans Ris, 1967, strongly sup­ported an al­ter­na­tive idea, i.e., that the phage slide down the sur­face of a fla­gel­lum and in­ject their DNA at the base. Chi phages, each with its kinky tail fiber wrapped around a fla­gel­lum, were de­tected at 0.5 min post-in­fec­tion; by five min the phages, now with empty heads, were po­si­tioned at the base of the fla­gella, the pre­sumed in­jec­tion site.

The about-to-be-dis­carded reprint that caught my at­ten­tion was a study by Shoshana Ravid and Michael Eisen­bach, 1983. These au­thors tested a model pro­posed a decade ear­lier (Berg and An­der­son, 1973) in which CCW fla­gel­lar ro­ta­tion was en­vi­sioned dri­ving an ad­sorbed Chi phage to­wards the cell body, like a nut be­ing screwed along a bolt. Un­for­tu­nately for the model, their data sug­gested that in­ces­sant ro­ta­tion in ei­ther di­rec­tion fa­vored ir­re­versible ad­sorp­tion.

Fil­a­ment sur­face lat­tice. Both fig­ures are from slide 17 of an in­for­ma­tive pre­sen­ta­tion on fla­gel­lar ro­ta­tion avail­able here. (Left) Nor­mal fla­gel­lum. Num­bered sub­units are el­e­ments of the ba­sic he­lix, shown here with 5½ sub­units per turn. The sub­units (shaded) of the re­sul­tant al­most hexag­o­nal, cylin­dri­cal lat­tice are in close con­tact in three he­li­cal di­rec­tions (named 11‑, 6‑, & 5‑start) cor­re­spond­ing to how many he­li­cal lines would be needed to tra­verse all sub­units of the lat­tice. As there are not ex­actly 5½ subunits/turn, the protofil­a­ments (one in black) are at a slight an­gle to the cylin­der axis (11'), which is dis­torted into a he­li­cal form. Orig­i­nal source: Mac­nab, R.M., Fla­gella and Motil­ity, Part 2, pp 123–145, Chap­ter 10 Fig 4 in Nei­d­hardt, F.C. et al. (eds.) Es­cherichia coli and Sal­mo­nella, 2nd ed. (1999) ASM Press. (Right) Straight (all L & all R) fla­gella. Mea­sured at a ra­dius of 4.5 nm, the in­ter­sub­unit spac­ing along an R protofil­a­ment (11-start R he­lix) is 0.07 nm less than for cor­re­spond­ing sub­units of an L protofil­a­ment. Orig­i­nal source

Thus the nut & bolt model ap­peared mori­bund un­til 16 years later—26 years af­ter the model had been proposed—when mem­bers of the Howard Berg lab­o­ra­tory res­ur­rected it (Ar­avinthan Samuel et al., 1999). Their study con­firmed the early ob­ser­va­tions that sen­si­tiv­ity to Chi re­quires fla­gel­lar ro­ta­tion (al­though the ro­ta­tion need not pro­duce motil­ity). Fur­ther­more, in their ex­per­i­ments, the di­rec­tion of the ro­ta­tion does mat­ter. Work­ing with E. coli in which the CCW ro­ta­tional bias could be ex­per­i­men­tally in­creased or de­creased by al­ter­ing the ex­pres­sion of a sin­gle gene, the au­thors showed that a strong CW bias re­sults in both in­ces­sant tum­bling and Chi-re­sis­tance, whereas a strong CCW bias re­sults in con­tin­u­ous smooth swim­ming and Chi-sen­si­tiv­ity. Fur­ther sup­port for the nut & bolt model was pro­vided by tam­per­ing with the bolt. Re­call that the usual ra­tio of protofil­a­ments is 2R to 9L. As­says for the Chi sen­si­tiv­ity of S. ty­phimurium strains whose protofil­a­ments are locked by mu­ta­tion into dif­fer­ent poly­mor­phic forms re­vealed a telling speci­ficity. Fla­gella with 0, 1, or 2 R protofil­a­ments con­fer sen­si­tiv­ity to Chi; fla­gella with 5, 6, or 11 do not. As the pro­por­tion of R fil­a­ments in­creases, the pitch of the he­li­cal grooves that run the length of each fla­gel­lum tight­ens. Just as you can screw a nut onto a bolt only if the thread­ing matches, these re­sults sug­gest that the vi­ral tail fiber (folded to form the "nut") fits only into the rel­a­tively wide grooves found on fla­gel­lar "bolts" with few R protofil­a­ments.

The di­rec­tional thread­ing of Chi phage along a spin­ning fla­gel­lum is an ex­otic ex­am­ple of a di­rected mol­e­c­u­lar move­ment along a fil­a­ment and, to my knowl­edge, has not been stud­ied fur­ther. More fa­mil­iar sit­u­a­tions in which nut & bolt mech­a­nisms have been in­voked are di­rec­tional translo­ca­tions in which the "bolt" is a DNA he­lix, e.g., DNA translo­ca­tion by cer­tain he­li­cases. There is also ev­i­dence to sug­gest that non-specif­i­cally bound pro­teins dif­fus­ing along DNA fol­low a spi­ral path. As a macro­scopic ob­server, I view the abil­ity of Chi to lasso and ride astride a spin­ning fil­a­ment as an ac­ro­batic feat out of the Wild West. The de­tails of this rodeo could be in­ter­est­ing.

What else might Chi sense?

I have no idea what else Chi might sense, but, tak­ing a clue from phage T4, I would not be sur­prised if its long, kinky tail fiber was not some­how pro­tected while the prob­a­bil­ity of find­ing a suit­able host was low. Like T4, Chi at­tacks en­teric bac­te­ria. Their in­tra­colonic life is rich; their out-of-body ex­pe­ri­ences are gen­er­ally im­pov­er­ished. T4 re­quires tryp­to­phan (or, in much higher con­cen­tra­tions, other amino acids) to ex­tend its rigid, jointed tail fibers for ad­sorp­tion to its host. Mono­va­lent cations in high con­cen­tra­tion are also re­quired for ad­sorp­tion. Both re­quire­ments are read­ily met in the colonic en­vi­ron­ment, rarely else­where. Chi does not have the lux­ury, avail­able to T4, of los­ing a tail fiber, hav­ing but one to be­gin with. Might Chi hold that one close to the chest (or tail) un­til con­di­tions were right? That would re­quire a chem­i­cal sense and a re­lease mech­a­nism. Has any­body looked?

Per­haps the ca­pac­ity for cir­cum­spect ap­praisal by a phage should not be viewed as ex­cep­tional. The process of phage ad­sorp­tion to its re­cep­tor site is not nec­es­sar­ily slam-bang, but (as Roger Hen­drix pointed out to me) may in­clude a re­versible ini­tial step that is sen­si­tive to en­vi­ron­men­tal con­di­tions. This is true of the well-stud­ied lambda phage that, de­spite the lack of tail fibers, may be said to pal­pate its prospec­tive vic­tims be­fore it com­mits to in­fect­ing them. The bet­ter en­dowed an­ces­tor of lambda, Ur-lambda (with tail-fibers), is pre­sum­ably more so­phis­ti­cated at pal­pa­tion; it has a larger host range and can come to a de­ci­sion faster.

In praise of a road not taken

The study of bac­te­r­ial chemo­taxis has of­fered the kinds of chal­lenges that the math­e­mati­cian David Hilbert, in his his­tor­i­cal 1900 ad­dress, char­ac­ter­ized as a good prob­lem: dif­fi­cult in or­der to at­tract us, yet not com­pletely in­ac­ces­si­ble, lest it mock our ef­forts. It should be a guide­post on the labyrinthine paths to hid­den truths, and ul­ti­mately a re­minder of our plea­sure in the suc­cess­ful so­lu­tion. The study of chemo­taxis has re­lied on guid­ance from a wide range of dis­ci­plines and has con­tributed guide­posts to them. It has pro­vided a par­a­digm of adap­tive be­hav­ior on the time scales of both on­togeny and phy­logeny and has led to a model of ha­bit­u­a­tion with po­ten­tially wide ap­plic­a­bil­ity (Bray, Levin and Mor­ton-Firth, 1998). The fla­gel­lum is a mar­vel of nano-tech­nol­ogy, a pro­to­type for nano-en­gi­neers, in­deed so mar­velous as to have be­come a "poster child" for pro­po­nents of "in­tel­li­gent de­sign." For­tu­nately, there are more sa­vory re­wards for suc­cess in un­rav­el­ing Nature's hid­den truths, not least of which is a re­minder of the plea­sure in the suc­cess­ful so­lu­tion.

Ref­er­ence

Samuel AD, Pitta TP, Ryu WS, Danese PN, Le­ung EC, Berg HC. (1999). Fla­gel­lar de­ter­mi­nants of bac­te­r­ial sen­si­tiv­ity to chi-phage. Proc. Natl. Acad. Sci. USA, 96 (17), 9863−9866. PMID 10449785

Two brief re­views:

Mac­nab R (1999). The bac­te­r­ial fla­gel­lum: Re­versible ro­tary pro­peller and Type III ex­port ap­pa­ra­tus (Guest Com­men­tary). J Bact, 181, 7149–7153

Berg HC (2008). Quick Guide: Bac­te­r­ial fla­gel­lar mo­tor. Curr Biol, 18 (16), R689-R691

P.S. An­other re­mark­able fea­ture of some phage tail fibers (Chi's?) is their en­larged host range reper­toire, achieved by re­com­bi­na­tion within the tail fiber genes. The phe­nom­e­non is well known in phages Mu and P1, but par­tic­u­larly dra­matic in a Bor­de­tella phage, dis­cussed by Merry and Elio in an ear­lier STC blog post.

Michael Yarmolin­sky is Sci­en­tist Emer­i­tus in the Lab­o­ra­tory of Bio­chem­istry and Mol­e­c­u­lar Bi­ol­ogy of the Cen­ter for Can­cer Re­search, Na­tional Can­cer In­sti­tute, NIH, Bethesda, MD.

An af­ter­word by Merry & Elio

phage PBS1 ad­sorbed to the fla­gella of B.subtilis SBI9. Neg­a­tively stained with 2% uranyl ac­etate (left: X 120,000, right: X 61,000).

Michael Yarmolinsky's in­trigu­ing post prompted a com­ment by Stan Zahler call­ing our at­ten­tion to an­other phage that ad­sorbs to the fla­gella of its host: the Bacil­lus sub­tilis trans­duc­ing phage PBS1. So off we went to see what we could see about this one. The pri­mary source for its story seems to be a 1968 re­port by Rai­mondo, Lundh, and Mar­tinez. Whereas Chi has one long kinky tail fiber, PBS1 has three he­li­cal ones that can be seen wrapped around the fla­gel­lum af­ter ad­sorp­tion.

Is PBS1 as fas­tid­i­ous as Chi, with sim­i­lar cri­te­ria for a suit­able fla­gel­lum? A read of the pa­per cited and oth­ers yields sev­eral clues sug­gest­ing that a dif­fer­ent mech­a­nism is at play here. But these re­ports leave us with more ques­tions than an­swers. And, as far as we can tell, un­like the case for Chi, no one has fol­lowed up on this story. Any vol­un­teers? We guar­an­tee your find­ings a place on the pages of Small Things Con­sid­ered.