Barbara Sherry

rueckert.gif (2119 bytes)

The following reminiscences are from Barbara Sherry who is presently an Associate Professor in the Department of Microbiology, Pathology, and Parasitology North Carolina State University, College of Veterinary Medicine.

When I started my graduate work in Roland Rueckert's lab, I was offered the opportunity to tag after a student who would graduate soon: Joe Icenogle. He was wonderfully patient, taught me a lot, and got me very interested in a project he had initiated with poliovirus examining mechanisms of antibody neutralization. Roland suggested I continue the investigations using rhinovirus 14. So at my oral preliminary exam, I proposed to determine how many antibodies had to bind a single rhinovirus 14 virion to neutralize it. I did fine in that examination; there were no problems. But Roland could see that my heart wasn't really in it, and he asked if that project was really the one I wanted to pursue. Well, I realized that I really wanted to address another hypothesis that Roland had tossed around. Perhaps (he had suggested), poliovirus had only three serotypes while rhinovirus had over one hundred, because poliovirus had many more sites than rhinovirus for antibody neutralization. Therefore (he continued), poliovirus might mutate at one site to escape antibody neutralization, but that would leave many other sites unchanged and still subject to neutralization. Therefore, generation of new serotypes would be almost unachievable. Conversely, when rhinovirus mutated, it would change one of the few sites available for neutralization, and it would therefore be relatively easy for that virus to mutate to an entirely new serotype. I thought this was unbelievably clever, and I couldn't wait to determine the number of antibody neutralizing sites on rhinovirus 14.

Anne Mosser and I generated a large panel of monoclonal antibodies specific for rhinovirus 14, and I used them to generate virus "escape mutants" (viruses selected to grow in the presence of the antibody, and therefore, by definition, mutated at the antibody-binding site or at another site that alters the antibody-binding site). I then tested each virus escape mutant for its resistance to all the other monoclonal antibodies, and in this way, segregated the viruses (and antibodies) into four independent groups. Virus escape mutants in one group were always neutralized by antibodies in that group, but resistant to antibodies in any of the other three groups. This, we thought, would correspond to four independent sites on the rhinovirus 14 capsid. I then used isoelectric focusing to determine which of the four viral capsid proteins was altered in each of the viral mutants. Now, if you think about it, this didn't appear to be a very logical approach. After all, isoelectric focusing will only detect charge changes in amino acids, and many mutations will change the amino acid but not the charge. But Roland was always thinking. He had me calculate what fraction of single nucleotide changes would result in an amino acid charge change (a very reasonable calculation), and the fraction was surprisingly high! So, having a dozen or more viral mutants for each of my four groups meant that it was highly likely that several of them would have altered isoelectric points in the affected capsid protein. I did the experiment, and the results were very gratifying. Most mutants in a group were altered in a single viral capsid protein. But, there were these troubling exceptions. A handful of mutants were altered in a viral capsid protein other than their grouping suggested. How many times do you think I re-checked? It was the fly in the ointment. Never mind, Roland assured me, go on and sequence all of the mutants. So, thanks to an ideal collaboration with Rich Colonno at Merck, Sharpe & Dohme, we were provided with primers to sequence our rhinovirus 14 mutants. Let me remind you that this was in about 1984, and you didn't just log in to a company or contact your local university primer-synthesizing group. Primers were gold back then. So this was an example where the work simply would not have been done if we didn't have that collaboration going.

To my dismay, several mutants that were altered in the "correct" viral capsid protein for their group had mutations that lay hundreds of bases away from the rest of the mutants in that group. Of course we knew that this could reflect protein secondary structure, but there just seemed to be too many of these "exceptions". Roland coughed a little, but remained calm. Meanwhile, Michael Rossmann and his group were working on the rhinovirus 14 structure by x-ray diffraction. We had been meeting regularly as a newly formed "WISPUR" group (Wisconsin - Purdue collaboration), which really moved everything along. One of the more remarkable aspects of this whole project was the perfect timing of it all. Michael and his group were struggling with fitting the rhinovirus 14 sequence into their x-ray diffraction results, and it was a tremendous help to them (or so they told me) to know which amino acids were likely to be external on the capsid. Bingo! My escape mutants told them which amino acids were altered by antibody selection, and therefore which were likely to be on the outside of the virion. It fit beautifully!

And then, Roland and I flew down to Purdue with the last of my sequencing information: those pesky virus "exceptions"; the ones that had mutations way too far away from the rest of their group or who were altered in the wrong capsid protein; the ones Roland had told me not to fret about. The tension was phenomenal. We sat in a room while I called out the amino acid number and the supposed antibody group. Eddie Arnold and his colleagues searched through the stacks of transparencies that defined the structure (again, this was back in the stone-age!). The first mutant we checked was the least risky: an amino acid change in the "right" capsid protein, but a little far away from the cluster that defined that group. Audible (very audible!) excitement when it turned out to be located directly adjacent to the other amino acids in that group in the three dimensional structure of the capsid. On to the next mutant: similar but with an amino acid change even further away. Giddiness as it too fell in an appropriate location on the three dimensional structure. With each successive mutant falling into place, there was this unbelievable incredulity. And while we were getting increasingly confident, we would all hesitate before we checked the next one; kind of pausing in the overwhelming relief, and worrying that the next mutant would be the one that destroyed the mood. Not a single mutant let us down. Every "exception" in the linear sequence was perfectly grouped in the three dimensional structure. I had never before experienced such resonance, such harmony in the data, and I haven't since.

The rest happened very quickly. We presented the data at a meeting on Viral Entry in Philadelphia, where I met a number of "bigwigs" in virology. I was truly shocked that any of them had any interest in talking to a peon like me (and learned a lot from those conversations). Eddie Arnold and I presented back-to-back talks to a huge, packed room at ASV. I was petrified, but thrilled. Time magazine called Roland when he was away, and did a telephone interview with me instead. Again: petrified, but thrilled. Roland, in his own special way, told me (at all of 27 years old) that I had probably reached a high point of my scientific career. He knew that this kind of synchrony is rare. And then it was time to move on. What became of Roland's hypothesis about picornavirus serotype diversity? Anne Mosser continued to work with poliovirus, and demonstrated that the original hypothesis was simply wrong. And doesn't that just show that it doesn't matter why you're asking the questions, you just have to keep asking them and work as a team. And it certainly framed the way I look at all results now: those "pesky exceptions" may turn out to be your next Nature paper.