Dr. Wlodawer,National Cancer Institute in Frederick, MD.

Alex Wlodawer received his Ph.D. from the University of California, Los Angeles, in 1974. Having completed his postdoctoral training at Stanford University, he joined the National Bureau of Standards in 1976 and in 1987 moved to the National Cancer Institute, where he is Chief of the Macromolecular Crystallography Laboratory. In 1998-1999 he was an elected Visiting Fellow of Sidney Sussex College, University of Cambridge, UK. He is a member of the American Crystallographic Association and the Protein Society, and has been an elected officer in both organizations. He serves on the editorial boards of the Journal of Biological Chemistry, Protein Science, The FEBS Journal, and Acta Biochimica Polonica. He was appointed an Adjunct Professor of Biochemistry and Molecular Biology and an Associate Member of the George Washington University Institute of Biomedical Sciences, Washington DC. He was awarded a degree of Doctor Honoris Causa of the Technical University of Lodz, Poland, was elected a Foreign Member of the Polish Academy of Sciences, and is a recipient of the 2006 NCI Mentor of Merit Award.

In Acta D, you recently expressed the point of view that experimental data for structures solved by X-ray and NMR should be deposited and released under the same policies as coordinate files. ¹   Why do you think this is so important?

The question of which crystallographic results should be deposited in PDB and on what schedule has been asked many times, but still does not have the final answer. Here, the experimental data refer to the processed data, e.g ., the structure factors in X-ray diffraction, not the raw images. The rules changed very substantially about 8 years ago, when the International Union of Crystallography modified its deposition regulations, and their recommendations became generally accepted by most funding agencies and by scientific journals. The coordinates of published structures must now be deposited in PDB and released upon publication of the relevant papers. However, although structure factors must also be deposited, their release can be delayed by up to 6 months. In a recent Letter to Editor published in Acta Cryst. D, ¹ I proposed that such a delay should be disallowed. I feel very strongly that the coordinates and structure factors are a matched pair, and one needs the other. The heart of the matter is that scientific results should be useful for the community (I consider description of a structure without the availability of coordinates to be advertising and not science), and verifiable (how can we prove that the structure is correct if not by comparison with the structure factors?). Let me give an example from a paper which I recently reviewed. The authors presented a series of structures of enzyme-inhibitor complexes, with one of the structures repeating a previously published experiment. However, the conformation of the inhibitor reported in the new paper was very different, changing in a substantial way the interpretation of the enzymatic mechanism. Unfortunately, with the diffraction data for the original structure never deposited (against the journal rules!), it was not possible to verify if the differences were real (and thus significant for the understanding of how the enzyme works) or due to errors in the interpretation present in the original paper. This is just one example, but I could cite many more. With the acceleration of the process of structure solution we should not have to wait half a year to verify what we read in the papers, if any doubts are raised. And let us remember that the most interesting results are often the ones that are most controversial.   

As a member of several editorial boards, what types of information are you looking for when reviewing macromolecular structure papers?  Has a journal published a paper, only to be surprised by the validation remarks in the corresponding PDB file?   What types of information do you think would be valuable to a reviewer of a paper describing a macromolecular structure?

Oh boy, have we been surprised… I have seen many papers, often published in Science and Nature (these journals seem to care more about getting the scoop than getting it right) where a look at the PDB files would bring a very unpleasant surprise about the quality of structural work. Whenever I review a paper that describes a structure I look first to see if the coordinates have already been deposited. However, that still tells me very little of what is hidden beyond the accession code. There has been much discussion of whether the reviewers should be given the actual coordinates. In an ideal world they should, but even I am a realist, and I know what is possible, and what is not. However, a minimum of what I would like to see as an Editor is the header of the PDB file and a brief version of the validation report. The former will tell me if the authors were lazy and did not bother to calculate the rmsd’s, or disclose what programs were used to solve the structure, where data were collected, etc . Many PDB data sets have all such records populated by a uniform answer “NULL”. A short validation report would tell me if the structure might have some serious problems. If I see a D amino acid in an otherwise normal protein, or interatomic distances of 0.1 Å, I would like at least to ask the authors some questions before accepting their otherwise brilliant paper.

Compared to pharmaceutical companies, what is the National Cancer Institute’s approach to focusing structural studies to cancer?

I am not allowed to talk about the policy of NCI without obtaining all sorts of permissions, so I better not delve too deeply into this matter. In general, it is not the mission of NCI to create drugs, but rather to create knowledge that might be the basis for drug development by pharmaceutical companies. We have fewer chemists than even some startup biotech/pharma companies, so not too much should be expected of us in this area. However, we do have some superb scholars doing fundamental research who generate data allowing understanding of the basis of cancer, delineating novel drug targets, creating new treatment methods and protocols, etc. Thus our role and that of the pharmaceutical companies should be considered to be different, but complementary.

What are your thoughts on the current state of crystallographic education?

What education? As far as I know, there is none. I do not believe that crystallography is still taught as a discipline, at least in the United States. Whether students will be exposed to it rigorously depends entirely on a good will of a faculty member old enough to know what he/she teaches. I am afraid that the education of most young people that actually solve crystal structures is limited to reading the manuals for HKL2000, CCP4, SHELX, COOT, or other black boxes. I am really afraid that when my generation retires, there will be few of the younger people who will be able not only to solve structures, but also understand the methods and develop them further. Hopefully I am wrong (happens often to me), but certainly I am not optimistic in this respect.

A. Wlodawer (2007) Deposition of structural data redux. Acta Crystallogr. D63:421-423.