The Orange and the Circus Tent

Illustrations by Grady McFerrin

agnified one million times, a virus is the size of an orange, while a human host cell is as large as a circus tent. Yet, within hours, the handful of genes in a single virus particle takes control of the cell with its 22,000-genes, and turns it into a virus factory.

How is the invasion possible? I've spent the larger part of my career trying to find out.

Although a biochemist by training, my fascination with viruses has spanned nearly five decades. I can trace the very moment it began to a conversation in the spring of 1969. At the time, I was a third year undergraduate in biochemistry, working in the lab of my brother-in-law, Kai Simons. Kai had just returned to Helsinki, Finland, from a postdoctoral stint at Rockefeller University in New York, and was full of ideas and plans.

LESSONS FROM VIRAL SPIES

Our lab has followed viruses through the steps outlined below. In the process, we've established mechanisms by which a cell performs many of its normal functions.

(1) Binding - Binding generally occurs to specific cell surface carbohydrates, lipids, and proteins different for different viruses.

(2) Lateral diffusion - The attached virus particles slide along the cell surface.

(3) Internalization - The viruses exploit endocytic pathways such as clathrin-mediated endocytosis that cells use to remodel their plasma membrane and bring in nutrients from the outside.

(4) Membrane penetration -The viruses or their capsids penetrate through the limiting membrane of these organelles into the cytosolic compartment.

(5) Intracytosolic transport along microtubules - The virus or viral capsid makes use of the microtubules and molecular motors for transport to the nucleus or other locations in the cell.

(6) Nuclear import - To deliver their genes into the nucleus for replication, many viruses make use of the import machinery and nuclear pore complexes that the cell normally uses for the transport of proteins and other macromolecules into the nucleus.

We met in the cafeteria of the Theoretical Institutes of Medicine (now The Haartman Institute) at the University of Helsinki to chat with Kai's friend, Leevi Kääriänen, who was just back from his own sabbatical at Sloan Kettering Institute in New York. Leevi told us of an animal virus he worked on called the Semliki Forest virus (SFV), named after a jungle area in Uganda where it had been first discovered. It was simple and easy to propagate in tissue culture; the particles were small and highly uniform in size with a central capsid containing a single RNA molecule, surrounded by a lipid bilayer. The virus exited its host cells by budding off a piece of the plasma membrane, containing both the lipid bilayer and viral surface glycoproteins.

The simplicity of SFV, and the possibilities it presented, hit us like an electric shock. Kai's main focus at the time was to study lipid-protein interactions using serum lipoproteins as a model system. The lipoproteins were isolated from serum acquired by plasmaphoresis performed on Kai, me, and other 'volunteers.' My project was to purify the protein moiety of low density lipoprotein (LDL) using organic solvents, keep it soluble using detergents, and characterize it. Being able to test our hypotheses regarding lipid-protein interactions using a real bilayer model -the SFV - and stopping the regular blood-letting sessions, was exciting for all involved.

Virtually overnight, Kai converted his operation to a SFV lab. Since we did not have a tissue culture incubator, his office was converted to a tissue culture room a few times a year: Books, papers, and Kai himself were thrown out and the heating set on max. Surfaces were disinfected with ethanol, and tables were stacked with 200 flat-bottom, glass prescription bottles. The bottles were used to grow BHK-21 cells, which were then infected according to a half-mystical ritual performed by Kai's technician Hilkka Virta. This 'low-tech' protocol provided us with purified virus that was shared between lab members. Kai struck up longstanding and highly productive collaborations not only with Kääriänen and his group one floor above us, but also our neighbor on the floor, Ossi Renkonen, who was an excellent lipid and carbohydrate chemist.

SFV turned out to be a true bonanza for all of us. My PhD thesis dealt with the mechanisms by which detergents solubilize biological membranes and with the methodology required to work with detergent-solubilized integral membrane proteins.¹

In those days, my interest was largely biochemical, particularly in the properties of membrane proteins, although I did also spend a lot of time trying to take the virus apart to its individual components, in an attempt to recreate the infectious particle from scratch. Needless to say, that project never panned out. While fascinated by the simplicity of the virus and its conceptual beauty, my interest at that time was not in the biology of the virus, in viral multiplication in the host cells, or in disease. That would come later.

In 1975 both Kai, Henrik Garoff, and I moved with our families to Heidelberg, Germany, to start working in the newly-founded European Molecular Biology Laboratory. EMBL enjoyed a wonderful mix of nationalities and experts, and a great spirit of collaboration and sharing. The question I wanted to address was simple: How did SFV, an animal virus, enter the host cell and establish infection? Viruses have no metabolism, no means of locomotion, and no molecular machinery that would allow independent replication. To multiply they have to force living cells to produce copies of themselves, as instructed by the viral genes. The mission of a virus particle is thus quite straightforward: To mediate the transfer of the viral genome in a replication competent state from an infected cell to a non-infected cell, and protect it in transit.

We knew how some bacteriophages entered bacteria, but the task for animal viruses is altogether different. The process had been studied extensively using electron microscopy resulting in a bitter controversy² between leading groups. Static pictures of viral entry could not resolve how the genetic material was released during productive infection. Were viruses passing directly through the plasma membrane into the cytosol, or were they internalized by endocytosis into intracellular vacuoles first?

Together with Jürgen Kartenbeck, an electron microscopist and close friend, Erik Fries, a postdoctoral student, and Kai, we employed a combination of old and new tools to probe the SFV-cell interaction. By tagging viruses with fluorescence, and following their interaction with cells, we could see them bind and move on the cell surface. By electron microscopy, we could see viruses enter indentations in the plasma membrane of the host cell and then into internal vacuoles. Using biochemical assays, we could confirm that the viruses were indeed rapidly internalized presumably by endocytosis. We found that we could inhibit infection by adding weak bases such as ammonia to the culture. We also tried mixing viruses with artificial membranes hoping that the virus would 'enter,' but nothing happened. We had a lot of data, many pieces of the story, but we didn't understand how they all fit together. I was deeply frustrated.

In 1978, two years into the project, I was invited to a Dahlem Conference in Berlin where the world leaders in cell biology discussed secretion, membrane fusion, endocytosis and other topics in the emerging field of membrane traffic. Among others, I had the opportunity to talk with Michael Brown and Joseph Goldstein from the University of Texas about their work on cholesterol metabolism, for which they would win the Nobel Prize in 1985. Together with Richard Anderson, they had recently shown that LDL, a carrier of cholesterol in serum, is taken up via receptors into clathrin-coated vesicles, which then fuse with lysosomes. Our lab had taken electron micrographs showing SFV stuck in clathrin-coated pits and vesicles (Click here for Image), but we had not fully understood the significance.³

Suddenly, our observations began to make sense. Like LDL, SFV was entering the cell by receptor-mediated endocytosis in clathrin-coated pits. We realized that the reason ammonia prevented infection was because it was elevating the pH inside lysosomes and other acidic compartments. The acidic environment of endocytic vacuoles was required to change the conformation of the viral spike glycoproteins, activating a latent membrane fusion activity resulting in the fusion of the viral membrane with that of the limiting membrane of the vacuole. Once conceived, this hypothesis seemed obvious. We were able to rapidly confirm it using both in vivo and in vitro experiments.⁴

We also discovered how to mix the viruses with artificial lipid bilayers (liposomes) and induce virus fusion in vitro. The trick was simply to drop the pH. This allowed us to study the acid-induced conformational changes in the viral membrane proteins, and to describe some fundamental features of protein-mediated membrane fusion.

It was a turning point in my scientific work and career. Joined by postdocs Judith White, Mark Marsh, and Karl Matlin, we established viral fusion as a general mechanism for the entry of most enveloped viruses including vesicular stomatitis (VSV) and influenza viruses. We also defined the organelle where fusion occurred as a pre-lysosomal compartment, an elusive compartment distinct from lysosomes that we called the endosome.⁵ Pre-lysosomal compartments were just starting to be identified and characterized by several cell biology groups.

Although many of the molecular details had yet to be worked out, we had solved the mechanism by which most animal viruses enter host cells at a conceptual level. The scientific community began to take note. In 1981 George Palade, whom I had first met at the Dahlem Conference, came through Heidelberg. He was seeking to expand the Section of Cell Biology at Yale University Medical School, and recruit new faculty working on membrane trafficking. I felt extremely pleased and flattered when he invited me to join, and also relieved because finding a job had not proven to be so easy.

By then, I had worked with Kai Simons for 13 years. Our families had lived together in a shared house for more than 10 years and I had come to think of our lab research as a small family enterprise. Taking the position at Yale would mean moving my family again to a new country, with a new language and culture. However, I was more than ready to start my own lab, and who could refuse an invitation to join what was considered by many to be the leading institute of cell biology at the time.

I was accompanied in the move by postdocs White and Marsh who worked with me in EMBL, and two technicians Eva Bolzau and Jenny Wellsteed, who made the start in New Haven, Conn., efficient and rapid. Also, I was able to join a long-lasting and highly productive partnership with Ira Mellman, who joined Palade's section at the same time. The two of us and our families became good friends. We fused our labs together, and enjoyed a productive partnership and daily working relationship for the next 16 years.

As part of the Section of Cell Biology (later a department), my interests naturally shifted towards basic functions of the cell. It was clear that viruses could be used as tools to analyze cellular processes - you could simply send a virus into the cell and follow its course. It would serve as a professional spy reporting back to you. It knows more about the target than you do. It 'speaks' the language of the host. It knows the passwords and PIN-codes, and it faithfully follows all the rules. In fact, many of the basic concepts of molecular biology, such as splicing, capping, polyA tails, promotors, oncogenes, and enhancers, were discovered by analyzing viruses. Why not use them in the same way to study cell biological functions and processes?

Although fewer and fewer labs continued using viruses as probes in the 1980s when it became possible to clone, sequence, mutate, and express cellular genes, we stayed true to the idea of using viruses to reveal the cell's inner machinery. Over many years, we used different viruses, and studied many different topics including endocytosis, membrane traffic, organelle acidification, membrane fusion, glycosylation and other protein modifications, in vivo protein folding and quality control, nuclear import, and microtubule-mediated transport. (See graphic above)

One of my long-standing projects focused on the cell biology of protein folding and maturation in the endoplasmic reticulum (ER) of the living cell. The project provided a central theme during many years of work both at Yale and later, when I moved to ETH Zurich. It contributed several concepts of fundamental significance, demonstrating the mechanisms by which a cell performs "quality control" of newly formed proteins.⁶ In the past, protein folding had been analyzed almost exclusively in vitro using the refolding approach pioneered by Nobel Prize winner Christian Anfinsen. Little was known about how proteins fold in living cells. The synthesis of viral glycoproteins in the infected cell provided us with a powerful system to study cell biological aspects of protein folding in situ.

Like cellular glycoproteins, viral spike glycoproteins are synthesized in the ER, where most acquire disulfide bonds and multiple N-linked glycans.⁷ They are then transported to the plasma membrane or to different organelles of the secretory pathway where virus budding occurs. They are often produced in large quantities under conditions where the cell's own protein synthesis is suppressed. This makes them excellent experimental models to study protein synthesis and maturation in vivo.

Using a pulse-chase technique with very short pulses, we analyzed the translation, folding, and subsequent fate of influenza hemaglutinin (HA) and vesicular stomatitis virus (VSV) G-protein. We could show that protein folding starts inside the ER lumen already before the polypeptide is completely translated. It involves the orderly oxidation of specific intra-molecular disulfide bonds, and it depends on the addition of N-linked glycans to the growing nascent chain, as well as the interaction with ER chaperones. Folding is completed after release from the ribosomes and followed by assembly of oligomers. Only when the proteins had acquired a folded, correctly assembled conformation are they transported to the Golgi complex and the cell surface. Of the various chaperones in the ER, we showed that calnexin and calreticulin interact with the newly added N-linked glycans, and that they are part of an elaborate quality control system that includes folding sensors and retention mechanisms, making sure that only correctly folded proteins are able to leave the ER. (See graphic below)

GLYCOPROTEIN FOLDING QUALITY CONTROL

Like luggage that must be trafficked through airline hubs, (1) new proteins are tagged as they enter the endoplasmic reticulum (ER ) with sugar moieties that direct them through the folding process. (2) The folding chaperones, calreticulin and ER p57, which recognize the sugar tags, assist the protein during folding. By removing a sugar from the moiety, an enzyme releases the protein from the chaperones. (3) Another enzyme that acts as a folding sensor, checks the new protein before its (4) release from the ER . (5) If the protein is still incompletely folded the protein is marked for a second pass through the folding cycle by re-addition of a sugar. (6) Permanently misfolded proteins loose another sugar moiety, and this directs them for degradation.

These observations of viral glycoproteins brought us into the field of glycobiology;⁷ it was now possible to better understand the role of N-linked glycans and their complex trimming program in the ER. Together with colleagues and group members at ETH Zurich, where I moved in late 1997, we continued to determine the cell biological, biochemical and structural characteristics of the calnexin/calreticulin cycle.

Through all these years of exploiting viruses to acquire insider-information about cells, I paid scant attention to the fact that viruses are pathogens. Millions of people die from viral diseases such as AIDS, hepatitis, polio, cancer, and less deadly viruses cause severe morbidity and economic loss. Aside from vaccination, there is little that can be done against most viral diseases. Antiviral drugs are problematic in that viruses rapidly acquire resistance against them by mutating their genes. Experts in infectious diseases are urging scientists to take the threat of emerging and re-emerging infectious agents more seriously, and to start investigating new approaches against viruses.

In my case, a more medically oriented new direction has come quite naturally. For more than 10 years, my lab at ETH Zurich has joined forces with two former lab alumni: Urs Greber, who was part of the lab at Yale, and Lucas Pelkmans, who worked with us at ETH Zurich as a graduate student a few years ago. Jointly, our goal is to analyze different viruses and identify cell proteins needed to support infection using a systems biology approach. The information can then be used to develop antiviral agents that target critical cellular proteins rather than viral proteins. This approach should diminish the potential for the emergence of resistant strains, and might provide drugs effective against multiple viruses that use a common pathway.

To identify potential host cell targets, we are using sequence information from the human genome project alongside automated, high throughput siRNA silencing screens in which we test infection by viruses after silencing single human genes one by one. The approach is yielding hundreds of 'hits,' allowing us to reconstruct in detail which cellular functions and pathways are critical for a given virus. Some are common to many viruses, others are unique. This technology provides a wealth of new information about host cell-virus interactions, and is, as we hoped, revealing many potential host cell targets for novel antivirals. We have founded a start-up company, 3-V Biosciences, to bring the work to the level of actual lead compounds and eventually to a new generation of antiviral drugs.

After watching the performances that these 'oranges' put up in the cellular 'circus tent' for all these years, I have gained deep respect for their clever acrobatics and their deceptive strategies. I would be more than pleased if our work could lead the way to an antiviral that, with the swing of a cape and a grand poof of smoke, would put an end to their spectacular but fatal acts.

Ari Helenius is a professor of biochemistry at ETH, the Swiss Federal Institute of Technology, in Zurich.