© Nature Publishing Group

Bordetella species, which cause respiratory infections in mammals, use a complex program of gene expression, mediated by the BvgAS phosphorelay, to change the cell surface and by extension infectivity. Bordetella bacteriophages use a diversity-generating system that changes receptor molecules for attachment and infection at specific frequencies. Thus BPP phage preferentially infect infectious, Bvg+ phase bacteria, BMP phage preferentially infect avirulent, Bvg-bacteria, and BIP are indiscriminate to bacterial phase. (From S. Doulatov et al., Nature, 431:476-81, Sept. 23, 2004.)

With inquisitive minds and tools as simple as a Waring blender, the work of early phage researchers such as Max Delbruck, Seymour Benzer, and Alfred Hershey generated much of the knowledge underlying contemporary molecular biology. But in a decision that echoes current debates about focusing too narrowly on canonical model organisms,1 Delbruck insisted that research be done only on the T-phages that infect...


Like all phages, Bordetella bacteriophage infects its bacterial host through ligand/receptor interaction. If the phage's tail-fiber proteins locate host receptors, the bacterium is doomed. But Bordetella bacteria, which switch between a virulent (Bvg+) and nonvirulent (Bvg-) phase, present a special challenge: "It's like a shape-shifter," says Young. "It makes it difficult for bacteriophage; they're not used to bacteria being able to change their complete outside surface whenever they want to." If phage were to mutate at a normal pace, they simply could not keep pace with Bordetella.

In response, the phage has developed the ability to switch tail-fiber protein tropisms, resulting in multiple states. When Miller and his collaborators compared the viral sequences, they discovered that all the tropisms contained a 134-basepair region of variability (VR1) and an almost 90% identical template repeat (TR). Adjacent to the TR was the unexpected Bordetella reverse transcriptase (brt), which led to the supposition that brt might aid in diversity generation, since it would be otherwise superfluous in a dsDNA phage.

Experiments in which brt was inactivated by mutations showed that the phage could still replicate. But, they could not change tropism in their tail fibers, supporting RT's role in diversity generation. Recent experiments have additionally shown that for as yet unknown reasons, this function is surprisingly specific: Mutagenesis takes place only at adenines. "If we compare variable repeats, we see nucleotide changes occur at certain positions relative to the template," says Miller. "Every position in the template at which there's an adenine is a position where we see variability in different phage."

To understand the mechanism, Miller's group is attempting to knockout the repair enzyme MutY. "If you line up an incorrect base across from an [adenine], MutY will take out the A at different efficiencies depending on which base it is mispaired with. Other enzymes then affect excision and repair synthesis to result in the complementary base being inserted across from the base that was mispaired with A," says Miller's nearly identically named UCLA colleague, Jeffrey H. Miller, who discovered the MutY gene.4 "Our model is that ... at the end of the variable repeat ... there's a single-strand nick that occurs, and that essentially is what determines specificity for the mechanism," says J.F. Miller.



© 2004 Nature Publishing Group

Putative DGRs are shown alongside Bordetella phage DGR. Dotted lines represent regions not shown to scale. (From S. Doulatov et al., Nature, 431:476-81, Sept. 23, 2004.)

J.F. Miller has dubbed the actors in this process diversity-generating retroelements (DGRs). His UCLA colleague compares this mechanism to the mammalian adaptive immune system, pointing out that phage are not alone in cleverly achieving targeted genomic diversity. "It's an additional example of a mechanism for mutagenizing a portion of a gene or only one gene instead of mutagenizing everything," says J.H. Miller.

In a recent Nature paper J.F. Miller and collaborators, including biology professor Steven Zimmerly from the University of Calgary, discovered that the mechanism extends beyond phage.5 Their paper describes an entire family of DGRs in various bacteria. They found it notable, says J.F. Miller, "that in every case when we compared the analog of our template with the analog of our variable region, they differ at positions corresponding at adenine residues in the template, so it looked like the same mechanism was conserved." (Also notable, first author Sergei Doulatov was an undergraduate when he and graduate student Asher Hodes designed the multiple substitution experiments at the heart of the paper.)

The mechanism appears in the genomes of everything from mammalian commensal bacteria to photosynthetic marine cyanobacteria. "This variability mechanism has found its way into other organisms to do other things," says J.F. Miller. "Now the question is: What [is] the universe of these bacterial elements doing for their bacterial hosts?"

Luis Villarreal, professor of molecular biology and biochemistry at UC, Irvine, finds the evolutionary implications particularly exciting, "It's a biological peculiarity, but to me it still makes a big point," says Villarreal. "If you open up a book on phage biology, this should be prominent. It's the link towards a way to solve the problem of creating genetic diversity in the prokaryotic world and creating diversity via a similar solution in eukaryotes, such as in our adaptive immune system." Incidentally, Villarreal is working on a book about viral evolution.


J.F. Miller, who partnered in a biotech startup called Avidbiotics in Delaware, sees two developmental pathways: "We're simultaneously pursuing two major applications. The first involves antimicrobials/phage therapy, and the other is focused on diversifying proteins of interest as a means to find new drugs."

Avidbiotics intends to create an effective means of phage therapy that would overcome the fears that phage treatment could lead to the unintended transfer of genomic material. Rather than using whole phage, they would create dozens, if not hundreds, of tail fibers, each with a different protein, and combine them to use against bacteria. Tail fibers alone are enough to kill bacteria by piercing and depolarizing the cell membrane, but since they do not contain genomic material, they would not replicate, solving not just the transfer issue, but enabling accurate dosing. "The FDA is adamant about understanding the dosing of therapeutics," says David Martin, J.F. Miller's partner and a former head of R&D at Genentech, "If you've got a therapeutic that replicates exponentially, you cannot control that dose."

No matter how fast the target bacteria might mutate, this "cocktail of tails" would likely have enough different proteins to find a receptor to bind. Eventually, of course, the bacteria would mutate beyond whatever was in the mix, so the plan is to survey isolates and keep pace by developing new cocktails – still a faster process than traditional antibiotic development, according to J.F. Miller.

Martin tempers optimism with a realistic assessment of the challenges involved to receive federal approval in the United States. "There's a lot of naïveté among companies that are trying to develop bacteriophage therapy ... about what it's going to take to even start a clinical trial in the United States, much less wind up with an approved product," says Martin.

The DGR itself could also become a bench tool, says J.F. Miller. "If a virus can use this to diversify a receptor, maybe if we understand the mechanism, we can transplant into other proteins to diversify them," he says. If it works, phage research will have come a long way from Waring blenders.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?