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 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.

The flagellar organelle. Source.

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.)

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.

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. 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.

Interaction of phage Chi with
E. coli. Chi attached to flagellar
filaments via their tail fibers
X 120,000.

   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.

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, H.C. 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

Michael Yarmolinsky's intriguing post prompted a comment by Stan Zahler (see Comments) 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.

(Right) PBS1 adsorbed to the flagella of B. subtilis SBI9. Negatively stained with 2% uranyl acetate. X 120,000. (Click image to enlarge.)

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.

(Right) PBS1 adsorbed to the flagella of B. subtilis SBI9. Negatively stained with 2% uranyl acetate. X 61,000.

Samuel AD, Pitta TP, Ryu WS, Danese PN, Leung EC, & Berg HC (1999). Flagellar determinants of bacterial sensitivity to chi-phage. Proceedings of the National Academy of Sciences of the United States of America, 96 (17), 9863-6 PMID: 10449785

Comments

[Mark O. Martin]

Also, Dr. Yarmolinsky's story reminds me of Archimedes' Screw:

http://en.wikipedia.org/wiki/Archimedes_screw

Then, for some reason (I'm writing and trying to be thoughtful as I do so), I am reminded of a couple of truly remarkable articles about flagellar rotation and signal transduction:

http://jb.asm.org/cgi/content/abstract/167/1/210;

and

http://www.cell.com/abstract/0092-8674(88)90197-3

Aaaaaannnddd...for a more recent review of this kind of work:

http://www.cell.com/trends/microbiology/abstract/S0966-842X(09)00240-6

Short version: things that restrict free rotation of the polar flagellum create a signal that alters gene expression. V. parahaemolyticus, in fact, measures viscosity directly by this mechanism, and thus can use that information to alter gene regulation on surfaces!

Beauty in small things....

[Mark O. Martin]

Wow. Simply wow. The problem with teaching microbiology, despite its wonders, is finding "nuggets of awesomeness" (as a student put it) that could be introduced once a week. No textbook can keep up! Seriously, I think Elio and Merry could produce a book of their "all star" hits for students every year. I know I would use it in class.

Thanks to Dr. Yarmolinsky for this story!

[mschaech@sunstroke.sdsu.edu]

For what it's worth, Phage PBS1 of Bacillus subtilis, a widely-used transducing phage, also requires that its host be flagellate, and, usually, motile.
-- Stan Zahler