by Roberto
What with the all the talk about CRISPR, the renewed interest in phage therapy, the widespread occurrence of huge phages (to mention but a few of the reasons phage are constantly in the news these days), it's no exaggeration to say that once again in biology "phage are all the rage." After they burst into the scene to launch molecular biology, phage went into a relative eclipse period, just to explode into their current full glory (puns intended). Yet, amidst all the very exciting current news about phage, what most tickles my fancy about them is their ability to mediate horizontal gene transfer via transduction. Hatfull and Hendrix start their landmark review on phages and their genomes – published ten years ago almost to the day – with a couple of mindboggling numbers: "there are more than 1031 phage particles in the biosphere" and "there are about 1023 phage infections per second on a global scale." While they focus on the facts that the phage population is very large and highly dynamic, I marvel at the evolutionary implications of 'oh so many' transduction events! Even if you take those numbers with a grain of salt (as well you should) and give them a leeway of a few orders of magnitude either way, their unimaginably large magnitude is cause for excitement. But I must admit, I have a very soft spot in my heart for transduction.
The very first experiments I did as an undergraduate were P1 transductions. The beauty of doing those experiments was that they left me plenty of time to think. Why did the phage sometimes make the mistake of filling its capsid with host DNA? And how did we know that those transducing particles were filled with just host DNA? The search for answers led to me one of my early favorite papers, Ikeda and Tomizawa 1965, elegantly demonstrating that transducing particles contained only host DNA.
Then I hit a wall. I tried reading the first reference of Ikeda and Tomizawa's paper: Zinder and Lederberg 1952, Genetic Exchange in Salmonella. This classic, describing the discovery of transduction, was a difficult read back then and remains so to this day. Fortunately, there are two lovely pieces that greatly aid in understanding this discovery. First, there is Sandy Parkinson's 2016 "Classic Spotlight," that nicely explains the experiments and the logic of Zinder and Lederberg. Second, there is Norton Zinder's 1992 personal retrospective of the events leading to the 1952 publication. It is sure to keep you at the edge of your seat! Here are two teasers from it, both related to the 1951 Cold Spring Harbor Symposium, where his early results were presented. "If ever a science was in prerevolutionary crisis (Kuhn 1970) it was genetics in 1951. The symposium was opened by Richard Goldschmidt, who proclaimed that there was no such thing as a gene." "Lederberg gave a paper from our laboratory which I believe has won all competition for incomprehensibility. He spoke for more than six hours." Can you really resist the urge to read it?
Fig. 1. Electron micrograph of bacteriophage P22. Source. Frontispiece.
By the time Zinder went to Cold Spring Harbor in 1951, he already had results showing that a "filterable agent" mediated recombination between some Salmonella strains. But exactly what was going was still a mystery. Zinder himself states: "Recall that this was all before the Hershey-Chase Experiment and the Watson-Crick model for DNA," so it's no surprise that the mechanism was not readily apparent. But, at the meeting Zinder had a Eureka moment when Harriett Ephrussi-Taylor said that the genetic exchange was due to a phage with some DNA stuck onto it. Within a couple of months – August to October 1951, precisely seventy years ago – Zinder had figured it out: Salmonella phage P22 mediated the observed genetic exchange, the phenomenon was named transduction. This also marks the birth of Salmonella genetics.
Fig. 2. Top panel: Swarm of a motile transductant emanating from a streak of non-motile cells. Source. Bottom panel, trails of microcolonies coming out of a point of inoculation of non-motile cells. Source.
Let me summarize. In P22-mediated transduction, first the phage lyses a cell. In the process some of the lysed cell's DNA is packaged into phage capsids. Those transducing particles inject the prior's host DNA into a new host and recombinants form. Is that all there is? No, not at all. There is another pathway. Around the time Zinder discovered the filterable agent, Bruce Stocker was working on Salmonella motility and flagellar variation. No surprise that Stocker decided to visit Lederberg's lab. In collaboration with Zinder, Stocker worked on the genetics of motility using a decades-old assay: in medium containing low agar concentration, motile cells can swim and eventually can occupy an entire Petri plate. In contrast, non-motile cells grow only at the site of inoculation. This proved ideal to obtain motile recombinants after mixing non-motile strains with a P22 lysate from motile strains. Flares (or swarms) of motile transductants emanate out of the streaks of non-motile strains (Figure 2 top panel). Now, here's the surprise. At frequencies far greater than motile transductants, Stocker observed trails of microcolonies emanating from the sites of inoculation of a non-motile strain after it has been exposed to a P22 lysate from a motile strain (Figure 2 bottom panel). Now, reader, if you are unfamiliar with these trails, stop reading and think about the phenomenon for a bit. What causes it? The explanation follows.
In their paper describing transduction of flagellar characters, Stocker, Zinder and Lederberg 1953, from simply looking at the trails offer a mechanism. I find their words, 68 years later, still awe-inspiring: "Many combinations of phage and recipient non-motile cells produced not only swarms but also groups of micro-colonies deep in the agar and at a distance from the site of inoculation (trails). These trails were generally more numerous than the swarms. The number of colonies in a trail varied from 4 to 50, or even more in plates incubated for 48 hr. Subculture from these colonies has yielded only non-motile forms. Yet the distance of the microcolonies from the site of inoculation, and their position in the depths of the medium, make it clear that the cell from which each colony grew must have reached the site of the colony by active movement. We suggest that trails result from abortive transduction, arising as follows: if a gene carried into an non-motile cell is dominant to that responsible for the absence of flagella, the cell will become motile; but if the exogenous gene fails to replace the homologous gene in the (hypothetical) chromosome, it will not be replicated, so that at cell division there will be produced one cell containing the foreign gene and therefore motile, and one cell lacking it, and therefore non-motile. Thus at any time there will be only one motile cell containing the foreign gene, and its path through the agar will be marked by a trail of micro-colonies growing from the non-motile daughter cells produced each time it divides." Both Lederberg and Stocker pursued this phenomenon and by 1956 both had reached similar interpretations on the nature of abortive transduction. It resulted from non-replicated DNA from the prior host, transmitted through many generations to a single descendant.
Fig. 3. The fate of P22-derived transducing DNA. While some of the DNA can recombine into the genome, at much greater frequencies the incoming DNA from the prior host is abortively transduced. Source.
For decades, I have remained fascinated by those trails and abortive transduction. Recently, I thought of a possible Talmudic Question: If the incoming transducing fragment must be attacked by nucleases to recombine, how does it survive 50 or more generations during abortive transduction? Then I did my due diligence and found the mechanism had been figured out! In 1997, Benson and Roth showed that the incoming transducing fragment enters one of two pathways (Figure 3). In one, the less frequent one, the linear fragment is attacked by the RecBCD nuclease, leading to recombination. In the other, by far the more frequent one, three phage proteins circularize the transducing fragment, render it nuclease-resistant and non-replicating. Thus, DNA from the prior host is now inherited by only one daughter cell after every division.
Now comes my own epiphany on abortive transduction. Normally, when we think of transduction and horizontal gene transfer, we think of stable recombinants that will inherit the DNA, thus having important evolutionary consequences. I want you now to think a bit outside that box. Go back to those 1023 phage infections happening every second. At much greater frequency than stable horizontal gene transfer, there will be myriad cells in every microbial community on Earth, carrying circles of stable, non-replicating DNA from a prior host. The phenotype of those cells will be altered, not for all progeny but for only one cell after each division. Those cells will have the capacity to exploit new niches for many generations. The ecological consequences of abortive transduction are enormous! When I read microbial ecology papers, I've yet to come across this idea. I hope that some of our STC readers will consider exploring this exciting possibility.
Comments