As hot as GPCRs were as an object of research, the one thing that researchers had failed to achieve, despite several decades of work, was to elucidate an accurate high-resolution structure of such a receptor. That changed, however, in the summer of 2000 when a team of researchers led by University of Washington biochemist Krzysztof Palczewski published the structure of rhodopsin in the 4 August 2000 issue of Science. Palczewski’s Science paper promptly became a fixture in Science Watch’s Biology Top Ten and has now been cited more than 600 times in less than three years. Palczewski’s impact in the field is also demonstrated by four other papers on visual transduction that have each been cited over 100 times.

            Palczewski, 45, received his Master of Science degree from the University of Wroclaw in Poland in 1980 and his Ph.D. in biochemistry from the same institution in 1986. He then spent four years at the University of Florida, the first two as a post-doc with Paul Hargrave, before moving on to Oregon Health &  Sciences University in 1990. In 1992, he joined the faculty at the University of Washington , where he now holds professorships in Ophthalmology, Pharmacology, and Chemistry.

From his office in Seattle, Palczewski spoke to Science Watch correspondent Gary Taubes

  How did you originally get involved in studying the signaling pathways of vision?

When I was a graduate student in Poland I came to the United States for a short period to work in Paul Hargrave’s lab. This was in 1983, and he had been working for 10 years on the sequencing of rhodopsin. His lab was considered one of the best in the world in which to learn about protein sequencing. I had been interested in visual processes since high school and now started learning more about it. When I finished my doctorate in 1986, I came back to the U.S. to work with Hargrave and study rhodopsin phosphorylation, and that was the beginning.

  It seems that this same subject, in effect, has kept you busy for 15 years. What is it that makes it so appealing to you?

This is what I’m interested in: the question of how light is changed into biochemical signals in photoreceptor cells. What I’ve wanted to do since the beginning is understand all the elements of visual transduction. Close your eyes now and you’ll immediately perceive how much information you lose about your environment. You can feel how much information comes through this visual transduction system. It doesn’t happen because of some mysterious events—it happens on a chemical level. It’s the chemistry, and rhodopsin is at the heart of it. Of course, rhodopsin is equally important now because it is a prototypical membrane protein from a family of proteins called seven transmembrane receptors, or G-protein-coupled receptors. These constitute a huge class of cell surface receptors. Virtually every physical process in the body is controlled to one extent or another by GPCRs, and about 50% of all medications work to modulate the activity of GPCRs. To name a few, if you think about blood or ocular pressures, addiction, ovulation, or energy balance—all of those are very much regulated by some type of GPCR.

As a result, whenever we learn something new about rhodopsin, the next step is confirming that other GPCRs have similar properties. That’s really put this receptor at the forefront of biology for many years. Beginning in the mid-1970s, there was this surge of information showing the similarity between visual transduction and hormonal signal transduction. Many laboratories, including Hargrave’s, identified the common elements of visual and hormonal transduction. Now it’s almost trivial to talk about it, but at the time I started in Hargrave’s lab, that was when we were just building this library of information suggesting the similarity of these parallel pathways.

  What is it about rhodopsin that makes it such a model paradigm for studying GPCRs in general?

It’s the most abundant of these receptors. You can isolate about a milligram from a single bovine retina and activate it by light in a very precise manner. In comparison, you might get nanograms of an adrenergic receptor from a native source. It’s the most concentrated receptor and the most accessible. And that’s why, in addition to the importance of vision, it’s the most extensively studied. Before the 1980s, before molecular cloning, rhodopsin was the only accessible receptor on which you could do biochemical and biophysical studies, rather than simple binding assays, which was what you had for all other receptors. Rhodopsin was at the forefront of working out the methodologies for studying these receptors. For example, my lab was the first to study rhodopsin phosphorylation in vivo by using mass spectrometry, because we were able to isolate a sufficient amount of material to do that. So rhodopsin was the first to be cloned, to have its protein sequencing done, and to have its structure elucidated. Now we’re pursuing its structure in native membranes, and that will most likely be done before any other GPCRs. The abundance of this protein and the quality of material have really been crucial.

  Why did you think you were ready to take on the challenge of determining the structure of rhodopsin, and why had no one ever done it before?

My lab was interested in structural studies of visual proteins for quite some time. We had started our work on arrestin, which is a protein involved in visual transduction, and my colleagues and I were very quickly successful with crystallizing that protein. That’s the first step toward determining the structure. But then we were beaten to publication by two other labs. About that time I was in at a meeting in Japan and met a very talented post-doc named Tetsuji Okada, and he very much wanted to come work with us on rhodopsin. So he did, and he developed a method of purification of rhodopsin that took advantage of the very high rhodopsin expression level in photoreceptor cells. The second very important aspect was that rhodopsin could be purified by simple extraction, and that gave us a very high concentration of rhodopsin protein that could be useful for crystallization. These proteins are very difficult to crystallize, or if they do crystallize, they are usually not well-ordered so they won’t diffract. But Tetsuji was very devoted and persistent, and he modified and tuned the method of extraction and then crystallization until he could grow relatively large crystals. Once we had those crystals, we worked in close collaboration with two other labs: Ronald Stenkamp’s here at the University of Washington , and Masashi Miyano’s at the Harima Institute in Japan . The crystal growth was done in my lab and the diffraction was done at synchrotron facilities around this country and in Japan.

  What have you learned from the structure, and why is it important enough to be cited more than 600 times in less than three years?

It’s important because all these GPCRs share similar folds and very similar structure—although this is only an assumption, of course, because we know the structure of only one of them: rhodopsin. But based on biochemical studies and on computational work, we believe that all the other receptors will have very similar if not identical structures. More than anything, having the structure told us exactly what it was. It gave us certainty. This receptor has been studied for 120 years. There were very good prediction models about what the structure would look like. There were a lot of biochemical and structural prediction methods developed to give us an idea what it would look like. Of course, the resolution of those predictions was relatively low. We knew there were seven transmembrane helices, and indeed that’s what the structure reveals. But the structure gives us certainty about the orientation of those helices, about the foundation of the environment of the ligand-binding site. This is critical for other GPCRs, particularly when you’re talking about designing drugs. You need to know some molecular details—where it binds and how one can best block that binding site to make a better drug.

  What’s the next step for your lab?

We now have the structure, but we don’t have it in the native membranes. For a long time now, we’ve been developing a very nice method of isolating those membranes from photoreceptor cells, and what we would like to do is use atomic force microscopy to effectively touch the rhodopsin molecule in its native membrane to see how it is organized and how it interacts with partner proteins. Again, very instrumental are collaborations with leaders in the field. To this end, we, Dr. Yan Liang, and Dr. David Saperstein at the University of Washington, were very lucky to work with Prof. Andreas Engel from Basel, Switzerland, and his associate, the extraordinarily talented Dr. Dimitrios Fotiadis. The first fruits of this work have just been published (D. Fotiadis, et al., Nature, 421:127-8, 2003). In this study, for the first time, native ROS-disk membranes have been imaged at sufficient resolution to reveal each individual rhodopsin molecule. The distinct, densely packed double rows clearly demonstrate the dimeric nature of rhodopsin in native membranes where it resides, supporting previous biochemical and pharmacological analyses that proposed dimerization and higher-order oligomerization for other GPCRs. The impressive organization of signaling GPCR molecules has important implications for G-protein recognition, binding kinetics, signal amplification, and signal termination. This new information runs against a 30-year-old dogma within the rhodopsin field, where it was believed that rhodopsin is a monomeric, highly mobile molecule. More work needs to be done, to extend this initial report and to validate our observation. Clearly, it is an area of research worth our focus and energy.

  Now that you have the structure, how do you apply that knowledge?

You can use the structure for multiple purposes. If you want to study the protein organization on the membrane, for instance, you can use the detailed atomic structure to help orient the molecule in the membrane. It’s recently been discovered that GPCRs aggregate, that they form a dimer. That may be very important because it means a large number of combinations can be generated by this dimeric association. Knowing the crystal structure will give us important information about the orientation of the GPCRs in the dimers. You can also use the structure to predict the arrangement of helices in other GPCRs, and you can then model those using the tools of computational biology. And then, this gives you the information you need to predict binding sites, which you can then use in rational drug design. Finally, you can use the structural information to understand GPCR mutations that lead to disease. There are about 100 or so mutations of rhodopsin, for instance, that lead to retinitis pigmentosa, a blinding disease. The structure provides you with information that might tell you why this mutant protein may be misfolded and how that can lead to disease. You can also start thinking about the ligand and the pharmacological chaperones that might restore proper folding and restore function of these mutated rhodopsin molecules. This has already been materialized to some degree (see U.E. Petaja-Repo, et al., Embo J., 21:1628-37, 2002; or N.M. Syed, et al., J. Biol. Chem., [in press]).

Finally, there is simple curiosity. Without knowing the structure, you cannot really envision the molecular mechanism through which rhodopsin interacts with other receptors and other proteins.