Bacteriophage P22 Scaffolding Protein (the Ghost in the Shell)

Barrie (Greene) Bell
Graduate Student, 1990-1995

I first met Professor Jonathan King when I took his class on “The Protein Folding Problem” during my first year of graduate school at MIT. It was an intriguing class, and Jon was an engaging teacher who successfully introduced us students to both the idea that how proteins got to their final folded structure was indeed a problem, and the strategies that researchers were using to begin solving it. At that time, the King Lab was studying protein folding pathways using the tailspike protein of bacteriophage P22. But what I was actually more interested in was the King Lab’s other area of research, virus assembly.

The question of virus assembly appealed to my artistic as well as scientific interests. I had spent time drawing icosahedral shapes as well as constructing them from paper. So I understood that sixty viral coat proteins could form twelve identical pentamers that would make the twelve vertices of an icosahedral viral capsid. In order to form a capsid the size of bacteriophage P22, however, the coat protein subunits would have to make hexamers as well as pentamers. How did the coat proteins know to adopt the correct “quasi-equivalent” conformations that would allow them to form hexamers only at the correct positions?

With Jon’s guidance, I decided to focus my thesis work on the P22 scaffolding protein. Previous work in the lab had determined that the assembly of the virus capsid involved an intermediate, a “procapsid”. Instead of viral DNA, the procapsid contained a “scaffolding” protein essential for virus assembly, but completely absent in the final DNA-filled virus structure. Similar scaffolding proteins were involved in the assembly of many other viruses, including some that infected humans, such as herpesviruses. There were many interesting questions about viral scaffolding proteins. What was the structure of the scaffolding? How did it interact with coat protein and other capsid components, such as the portal through which the DNA was packaged? By what means did it come out of the capsid to allow the DNA to go in? Which regions of the molecule were required for specific functions?

I thus became one of the “heads,” as the members of the lab who worked on P22 capsid proteins were called, in contrast to the “tails” whose projects focused on the tailspike protein. At this time the “tails” included postdocs Bao-Lu Chen and Anna Mitraki, graduate student Susan Sather, and technician Cammie Haase-Pettingell. The “heads” included postdoc Peter Prevelige, graduate student Carl Gordon, and visiting postdoc Marisa Galisteo. Also joining in lab meetings was Patricia Reilly, who managed the electron microscopy lab.

I had been a physics major as an undergraduate, and although I had done some work in microbiology labs, I was still unfamiliar with many techniques. Accordingly, my initial impression of the King Lab was a cluttered jumble of equipment of largely indiscernible purposes. Much of this was initially deciphered for me by Cammie, who showed me how to grow and work with bacteriophages. I much appreciated her patience as I gradually learned how to carry out these processes without contaminating all my pipettes with phage.

Senior graduate students Susan and Carl also both helped me to find my way around the lab. I was particularly impressed with Susan’s invariably immaculate workspace, covered in pristine lab soaker and arrayed with neatly labeled bottles. While I was quite proud of my very own lab bench, which had a traditional stone top distinct from the other benches in the lab, I was never able to make it look quite as neat. From Carl I learned phage genetics techniques used to map mutants and cross them into desired backgrounds, and I eventually inherited his own stocks of phage strains. Regretfully, I had to discard a number of these as I was unable to decipher Carl’s writing on the tubes to figure out what the strains were.

My main mentor, though, was Peter, who had developed many of the methods used to purify procapsid proteins and to study procapsid assembly in vitro. I dutifully read up on the various analytical equipment used in the lab so I could explain to Peter’s satisfaction that I understood the physical principles behind how they worked. I then followed Peter about the lab as I learned protein purification and analysis techniques, as well as the use of ingenious contraptions such as the sucrose gradient mixer and fractionator.

I still recall the distinct scent of phage growing in bacterial cells, from the many hours spent cultivating liters of phage-infected cells, the first step in producing purified phage proteins. This was done in the back room, where I kept the door closed to keep the smell from invading the rest of the lab space. I grew to find the concentrated phage scent sort of pleasant. This opinion was not shared by Mario, the janitor, who accused me of attempting to knock him out with the stench when he opened the door to collect the trash.

My most tragic moment in the lab was the day when, at the end of a two week long purification process, I misread a container label and precipitated my scaffolding protein with aluminum sulfate instead of ammonium sulfate. Precipitation by ammonium sulfate can be reversed to regain soluble protein. As I learned to my dismay, precipitation by aluminum sulfate could not be reversed by any methods I knew or that Peter suggested, turning all of my carefully purified protein into clumpy garbage.

The ultimate purpose of learning all these methods was of course to begin doing original work on my own research problem. The central basis of my thesis work involved the production and analysis of mutant scaffolding proteins. The phage assembly process was believed to have multiple steps, from initiation, regulation of correct assembly, scaffolding release, DNA entry, and the structural change from the round procapsid to the clearly icosahedral mature phage form. Yet these processes occurred quickly enough that it was hard to separate them out either in vitro or in vivo. The thought was that scaffolding mutants would affect or block distinct steps in the assembly pathway, such as initiation, capsid growth, or scaffolding release, allowing the phage assembly process to be broken down into further discrete steps.

I recall that my thesis committee suggested that this plan was a “fishing expedition.” I will admit that I never quite understood why this was supposed to be a bad thing. While I appreciated the scientific formalism of generating hypotheses, it seemed to me that sometimes you needed to first generate more data in order to have something to hypothesize about. There was a legitimate concern here though, in that previous rounds of mutant generation had yielded good collections of temperature-sensitive mutants in both the coat and tailspike proteins, but only one lonely scaffolding protein mutant. Furthermore, the temperature sensitive coat proteins all had defects in protein folding. This was of course quite interesting for the members of the lab who studied protein folding pathways, but not really relevant to the phage assembly process.

Still, evidence suggested that the scaffolding protein had an uncomplicated, primarily helical structure, so it seemed likely that mutants here might affect protein function, not folding. I pored through the theses of former graduate students to find any mentions of previously isolated scaffolding or coat mutant strains and the methods by which they had been obtained. In order to find mutants with a phenotype, I selected revertants of amber mutations in scaffolding that had a temperature sensitive or cold sensitive phenotype, as well as selecting suppressors of cold sensitive coat protein mutants. The final result of various mutant production methods as well as excavations through the old phage stocks in the cold room were P22 strains with mutations at four different sites in the scaffolding protein.

Now I could begin the exciting work of figuring out what effect these mutations had on scaffolding functions. Two of the temperature sensitive scaffolding mutants turned out to assemble procapsids that lacked the DNA packaging portal, and thus were unable to package DNA. These mutants confirmed that scaffolding was required for incorporation of the portal, and further defined the scaffolding region involved.

Two mutants in a different region of the sequence were impaired in scaffolding release. One of these mutants, selected as a suppressor of a cold sensitive mutation in the coat protein, produced procapsids that did not appear to release scaffolding, and from which the scaffolding was much harder to extract in vitro. The other mutant had a similar phenotype at high temperature, although the scaffolding was not quite as difficult to extract in vitro. Much to my fascination, this mutant also generated aberrantly shaped procapsid structures when grown at high temperature. These were not the spiral structures or small capsids made in the absence of scaffolding protein, but capsids that were larger or more oblate than normal, many of which seemed to not quite fully close.

This mutation, as well as a different substitution at the same site, also had a cold sensitive phenotype, which resembled the phenotype of the cold sensitive coat protein mutant. Phage growth was blocked at the procapsid stage, and the procapsids produced seemed to have lost much of their scaffolding protein, as though it had prematurely leaked from the procapsid. This was an intriguing result since the scaffolding protein normally had to exit in order to allow DNA to enter the procapsid, so it was unclear why this would pose a problem for phage maturation. Other experiments involving proteolysis of scaffolding protein confirmed that the essential coat binding region was at the c-terminus of the scaffolding protein, while these mutations were in the middle of the scaffolding protein sequence, so they were not involved in directly mediating the binding of scaffolding to the coat protein shell.

During my final months in the lab, I selected second site suppressors of the cold sensitive scaffolding mutation, which when sequenced were found to be in several different sites within the middle region of the scaffolding sequence, further suggesting this region was involved in scaffolding-scaffolding interactions. Given the phenotypes of the mutants in this region, these results suggested that interactions between scaffolding proteins were involved in both regulation of correct capsid assembly and in the release of scaffolding from procapsids.

By the time the scaffolding mutants had been isolated, Carol Teschke, a postdoc with lots of experience in protein folding, had joined the lab and showed me how to properly conduct and interpret protein folding experiments. The unfolding of scaffolding protein turned out not to be simple two-state mechanism as for a typical globular protein, but a complex process involving the sequential denaturation of multiple domains. The scaffolding protein was notably unstable, to the extent that some domains are probably largely unstructured at physiological temperatures. The less stable domains included the regions involved in coat binding, portal insertion and scaffolding release, suggesting that many critical scaffolding functions actually required a high degree of conformational flexibility. Interestingly, both of the mutations affecting scaffolding release resulted in severe destabilization of part of the protein to thermal denaturation, possibly supporting the notion of scaffolding release as an active function requiring a properly folded domain.

In addition to this work in Jon’s lab, I also had the unusual opportunity to spend several weeks away in not just one, but two other labs, as part of collaborations to obtain structures of P22 procapsids. I learned about protein crystallography in Professor Jack Johnson’s lab at Purdue, where I succeeded in growing beautiful crystals of P22 procapsids. These crystals sadly did not diffract x-rays at all, thus dashing our hopes of being the first to solve the crystal structure of a procapsid. Subsequent attempts to generate crystals of scaffolding protein never succeeded in producing crystals of any kind. Jon was convinced that these failures in themselves said something important about the scaffolding protein structure. I was admittedly reluctant to pursue this because I did not see how it would lead to results I could put in a thesis, but in retrospect Jon was almost certainly correct that this was some of the first evidence that the folded structure of the scaffolding protein was not sufficiently stable as to allow it to form an organized crystal lattice.

I also visited Professor Was Chiu’s lab at Baylor College of Medicine to learn about cryo-electron microscopy and image reconstruction. I personally did not show much talent at either freezing samples for cry-electron microscopy or computationally processing the resulting images. Luckily, graduate student Pamela Thuman-Commike was happy to take care of the computational side of things while allowing me to attend to what she considered to be the unpleasantly messy wet lab work of producing the samples. This was the start of a multi-year collaboration between the two labs that led to several papers describing the structures of P22 procapsids containing wild-type and mutant scaffolding proteins, mature virions, and the small procapsids formed in the absence of scaffolding protein. We were able to see the structure of the procapsid at sufficient resolution to see the clearly skewed herons with quasi-equivalent coat subunits. The scaffolding protein itself remained elusive, since it was apparently not arranged within the procapsids with icosahedral symmetry matching that of the coat protein shell, and thus did not show up in the image reconstructions.

One of the things I particularly appreciated about being a graduate student in the King Lab was that Jon took his role as a mentor seriously. He made sure that students in his lab learned not not only techniques, but how to formulate a problem and strategy for attacking it. He also coached me to develop my presentation skills, which admittedly started out as rather awful.

It was thus a proud moment when, during my fourth year as graduate student, I did my ten- minute research presentation at the summer Phage Assembly conference. I received several compliments on the conciseness and clarity of my presentation, which I had carefully rehearsed multiple times as Jon had taught me.

Because Jon believed in providing opportunities for students to experience doing actual experiments, there were always undergraduates working in the lab, both MIT students in the undergraduate research program and students from various summer programs. I had the chance to mentor three students, Morgan Kelly, Raka Mustaphi, and Joel Johnson. Defining a project of suitable scope for each of them to work on, and training them in the techniques they needed in order to carry it out, was as much a form of education for me as for them.

After graduating from MIT, I went to USCF to do postdoctoral work on clathrin and adaptor proteins, still pursuing an interest in self-assembling structures. I eventually ended up leaving lab research entirely for work in patent law. However, many of the skills I developed in the King Lab including practical problem solving, technical writing, presenting, and training, remained valuable throughout the rest of my career.

I appreciate that being asked to write this short memoir inspired me to reread my old papers after many years and to review the latest research on viral scaffolding proteins. It is satisfying to see that further work, including by King Lab alums, has shown the three-dimensional structures of scaffolding protein coat binding domains, and clarified how these domains interact with coat proteins. It is notable that the structure of scaffolding proteins within capsids and the nature of scaffolding-scaffolding interactions have still not been determined for P22 or for any other virus. Even after all these years, the scaffolding protein still retains some of its mystery.