S. PEREZ, C. VERGELATI AND V. H. TRAN 267 by van der Waals forces only; no short non-bonded distances are observed. The authors thank Professor A. Sarko for providing the initial version of the CRYSP program. The data collection was performed within the 'Groupe Grenob- lois de Diffractomftfie'. References ARNOTT, S. & SCOTT, S. E. (1972). J. Chem. Soc. Perkin Trans. 2, pp. 324-335. Box, M. J. (1965). Comput. J. 8, 42-52. BUSING, W. R., MARTIN, K. O. & LEVY, H. A. (1962). ORFLS. Report ORNL-TM-305. Oak Ridge National Laboratory, Ten- nessee, USA. COIRO, V. M., GIGLIO, E. & QUAGLIATA, C. (1972). Acta Cryst. B28, 3601-3605. CROMER, D. T. & WABER, J. T. (1965). Acta Cryst. 18, 104-109. DAUBER, P. & HAGLER, A. T. (1980). Acc. Chem. Res. 13, 105-112. DHEU, M. L. & P~REZ, S. (1981). J. Chem. Educ. 58, 552-553. KITAIGORODSKII, A. I. (1973). Molecular Crystals and Molecules. New York: Academic Press. LESER, J. (cid:127) RABINOVICH, D. (1978). Acta Cryst. B34, 2250-2252. PI~REZ, S. & BRISSE, F. (1978). Biopolymers, 17, 2083-2096. PI~REZ, S. & MARCHESSAULT, R. H. (1978). Carbohydr. Res. 65, 114-120. PI~REZ, S., SARKO, A. & TRAN, V. (1985). In preparation. PI~REZ, S., ST. PIERRE, J. & MARCHESSAULT, R. H. (1976). Can. J. Chem. 56, 2866-2871. RAMACHANDRAN, G. N. & SHAMALA, N. (1976). Acta Cryst. A32, 1008-1009. SCOT'F, R. A. & SCHERAGA, H. A. (1966). J. Chem. Phys. 45, 2091-2101. SIMONETTA, M. (1974). Acc. Chem. Res. 7, 345-350. SMITH, P. J. C. & ARNOTT, S. (1978). Acta Cryst. AJ4, 3-11. STEWART, R. F., DAVIDSON, E. R. & SIMPSON, W. T. (1965). J. Chem. Phys. 42, 3175-3187. TAKEDA, H., KAIYA, T., YASUOKA, N. & KASAI, N. (1978). Carbohydr. Res. 62, 27-37. TAKEDA, H., YASUOKA, N. & KASAI, N. (1977). Carbohydr. Res. 53, 137-152. WILLIAMS, D. E. (1969). Acta Cryst. A25, 464-470. WILLIAMS, D. E. (1972). Acta Cryst. A28, 629-635. ZUGENMAIER, P. t~ SARKO, A. (1972). Acta Cryst. B28, 3158- 3166. ZUGENMAIER, P. t~ SARKO, A. (1976). Biopolymers, 15, 2121- 2136. Acta Cryst. (1985). B41, 267-269 Use of the Multiwire Area Detector Diffractometer as a National Resource for Protein Crystallography BY NGUYEN HUU XUONG, DONALD SULLIVAN, CHRIS NIELSEN AND RONALD HAMLIN Department of Biology, Chemistry and Physics, University of California at San Diego, La Jolla, CA 92093, USA (Received 14 November 1984; accepted 25 February 1985) Abstract After many years of development, a high-speed data- collection system for protein crystallography using multiwire area detectors called Mark II has been dedicated as a national resource at the University of California, San Diego. Protein crystallographers can apply for time and come to this resource to collect data. During the last twelve months, this system has been used to collect more than 5 million intensity measurements in 17 different projects. The quality of the data was excellent with data reproducibility around 6% in intensity. The data were of such quality that many groups with good heavy-atom derivatives were able to have the protein structure solved within two months after the data collection. The Mark II system which can collect data at 100 times the rate of a standard diffractometer consists mainly of a rotating-anode X-ray generator (Elliot GS6) with a graphite monochromator and two multi- 0108-7681/85/040267-03501.50 wire X-ray area detectors (Cork, Hamlin, Vernon, Xuong & Perez-Mendez, 1975). The crystal is oriented by a three-circle goniostat and can be cooled to 243 K by a dry nitrogen cooling device. The whole system is controlled by a VAX 11/750 computer. This com- puter not only controls the orientation of the crystal but also analyzes the data from the area detectors to extract fight away the integrated intensity of the recor- ded reflections as the data are being collected. Both the hardware and software of the Mark II have been described in detail elsewhere (Cork et aS., 1975; Ham- lin, 1982; Xuong, Freer, Hamlin, Nielsen & Vernon, 1978; Howard, Nielsen & Xuong, 1985). We have now published a method or strategy on how to set up the data-collection runs depending on the space group and cell dimensions of the protein crystal (Xuong, Nielsen, Hamlin & Anderson, 1985). The system is kept in running condition by a small crew consisting of one physicist, one electronic technician and one computer programmer. The first-time user is helped by 'in-house' protein crystallographers but can learn to operate the system in a day or two. O 1985 International Union of Crystallography
`
`NOVARTIS EXHIBIT 2051
`Par v Novartis, IPR 2016-00084
`Page 1 of 3
`
`

`
`268 USE OF THE MULTIWIRE AREA DETECTOR DIFFRACTOMETER Table 1. Protein crystallography data collected at the UCSD resource during a 12-month period (1983-1984) Crystal DNA polymerase Histidine deearboxylase Aspartate transearbamylase, active (R) form Manganese superoxide dismutase Histone ot~.amer Trimethylamine dehydrogenase L-3-Hydroxyacyle coenzyme A dehydrogenase Iron superoxide dismutase Tu-elongation factor Erythrocyte nucleosome Glutamine synthetase Ribulose biophosphate carboxylase Glycogen phosphorylase Chicken dihydrofolate Cytochrome c peroxidase at 258 K Compound 1 of cytochrome c peroxidase Phospholipase Group leader(s) Steitz, Yale Univ. Hackert Univ. of Texas Lipscomb, Harvard Univ. Ludwig, Univ. of Michigan Love, Johns Hopkins Univ. Matthews, Washington Univ. Banaszak, Washington Univ. Ludwig, Univ. of Michigan Jurnak, Univ. of California, Riverside Buniek, Oak Ridge National Laboratory Eisenberg, Univ. of California, Los Angeles Eisenberg, Univ. of California, Los Angeles Fletterick, Univ. of California, San Francisco Kraut Univ. of California, San Diego Kraut-Xuong, Univ. of California, San Diego Kraut-Xuong, Univ. of California, San Diego Xuong, Univ. of California, San Diego Number Number Number of of of Space Cell axes (/~) intensity data crystals Resolution Rsy m Structure group a b c measurements sets* used (A) (%) solutioh P43 104 104 86 320 000 3 4 2.8 6 Solved I422 222 222 107 724 000 3 3 3.0 6 Solved P3121 122 122 156 840000 3 8 2-8 6 Solved P4122 146 146 56 340 000 3 3 2"5 3 Solved P3~21 119 119 103 384 000 3 3 3.0 6 Solved P2t 147 72 84 1 000 000 3 5 2-5 5 Good electron (/3 = 98*) density map C222 t 227 82 124 438 000 4 5 2-8 6 Good electron density map P2t212 t 82 75 71 101 000 1 1 2.0 6 In progress P222 71 74 171 350000 3 3 2.8 6 In progress P21 166 216 66 52 000 1 1 7 10 In progress (/3 = 96 o) C2 236 135 201 113 000 2 4 3.5 6 In progress (/3 = 103") 1422 149 149 138 150 000 2 3 2.8 7 In progress P43212 129 129 119 238 000 1 3 2-8 6 Structure refinement C2 89 48 64 78 000 1 2 1.6 3 Structure (/3 = 125 °) refinement P2t2~2 ~ 108 77 52 237 000 1 2 2.0 5 Structure refinement P2t2~2 ~ 108 77 52 53 000 1 6 2.4 5 Structure refinement P412t2 88 88 105 400 000 2 2 2.8 6 In progress * The number of data sets includes at least one native data set. The remaining data sets, if any, are for different heavy-atom derivatives. Once a year (in about July), the resource sends out a request for proposals (to US crystallographers). The proposals, usually in a form of a two- or three-page letter, are examined by an advisory committee con- sisting of five distinguished protein crystallographers and biochemists. The advisory committee rates these proposals and those with high rating are selected. After an initial adaptive period, we have settled on the selection of one outside group every month. Each outside group will have two weeks to collect data. The remaining time is devoted to preventive mainten- ance, system improvement and data collection for 'in-house' protein crystallographers. Since each data set (to 3 or 2.5 A resolution) usually takes about 2 to 4 d, a period of two weeks will allow a group to collect data on a native crystal and many heavy-atom derivatives. It is interesting to notice that each full data set can usually be obtained with only one crystal. The Mark II system hardware and software have been designed to extract the best intensity measure- ments out of each crystal. Since reflections from pro- tein crystals usually have a relatively small ratio of intensity over background, we have tried to reduce the background by using a graphite monochromator. The background for each reflection is measured at the same detector coordinates as the reflections but in a nearby range of the to-scanning angles (Xuong et al., 1978; Howard et al., 1985). To reduce statistical fluctuation, the background counts are accumulated 16 times as long as the time used to measure a reflec- tion intensity (Xuong et al., 1978; Howard et al., 1985). Due to the measurement of hundreds of simul- taneous reflections at the same time, we can afford to expose each reflection for a long time (from 150 to 300s for an average intensity measurement as compared with 10 to 20 s on a standard diffractometer with a rotating-anode generator). To correct for absorption and crystal decay, every data-collection 'run' is divided into scaling shifts of 5 ° rotation in the to-scanning angle and there are a sufficient number
`
`NOVARTIS EXHIBIT 2051
`Par v Novartis, IPR 2016-00084
`Page 2 of 3
`
`

`
`N. H. XUONG, D. SULLIVAN, C. NIELSEN AND R. HAMLIN 269 of duplicate measurements from the same reflection or symmetry-related ones to permit the precise deter- mination of the scale constants by least squares (Xuong et al., 1978; Howard et al., 1985). Since each scale constant applies to a group of reflections collec- ted on each detector that is localized both in reciprocal space and in measurement time, this method is very similar to the local scaling method mentioned by Matthews & Czerwinski (1975), but it also compensates for crystal decay due to radiation damage. However, the main reason for the quality of the intensity data is that each reflection (or its sym- metry-related ones) is measured on average from 5 to 10 times. Therefore, if one defines the data repro- ducibility to be N N Rsym = )-', )-', I(I(h,k,l))-I(h,k,l),[/E E I(h,k,l),, h,k~l i=1 h, lgl i=1 where the inner summation is over all measurements of each reflection or its symmetry-related one, then the precision on the average intensity which will be used for phase refinement or electron-density-map calculation will be Rsym/x/N. For example, if Rsy m = 6% and N =9 then the average intensity precision will be 2%. As shown in Table 1, the Mark II system was used during the last twelve months to collect more than 5 million intensity measurements in 17 different pro- jects. The quality of the data was excellent with data reproducibility around 6% (in intensity) for most crystals. The data were of such quality that many groups, with good heavy-atom derivatives, were able to have the protein structure solved within two months after the data collection. Often the electron density map was computed using phase information derived from only one heavy-atom derivative. In conclusion, our experience has shown that a high-speed data-collection system using a rotating- anode X-ray generator and two multiwire area detec- tors can be used quite effectively as a national resource. We are planning to add three more cham- bers to the system making it very efficient to collect data even for crystals with cell dimensions as large as 300/~. A few practical details we would like to mention are that the monochromator effectively reduces X-ray damage on the crystal and that the cooling device can sometimes dramatically prolong the life of a crystal (using 278 K instead of room temperature, for example). During the last twelve months, the down- time of the system was kept to less than 20%, most of which was due to troubles with the rotating-anode X-ray generator. Because the data are available immediately in the final form of integrated intensity measurements, a resource using a rotating-anode X-ray generator and multiwire area detectors is, we feel, preferable to the one with a storage ring and a rotation camera set-up (Greenbough & Helliwell, 1983). It takes time to scan and analyze films from a camera; also the film method is much less precise than a method using a photon- counting detector. Of course one could and should use area detectors at a storage ring too (Greenbough & Helliwell, 1983). We feel, however, that a resource with a rotating-anode and multiwire area detectors is optimum for data collection for most protein structure solutions because it can usually extract more than enough data from a single crystal and because it is much less expensive to operate. A resource with area detectors at a storage ring should be reserved for special projects like time-resolved crystallography (Bartunik, 1983; see also Bildenback, Moffat & Szebenyi, 1984) or phase determination using data from X-rays of different wavelengths (Phillips & Hodgson, 1980; Arndt, Greenbough, Helliwell, Howard, Rule & Thompson, 1982). This project was funded by grants from the National Institutes of Health (RR01644 and GM 20102) and from the National Science Founda- tion (PCM 8021653). References ARNDT, W. W., GREENBOUGH, T. J., HELLIWELL, J. R., HOWARD, J. A. K., RULE, S. A. & THOMPSON, A. W. (1982). Nature (London), 298, 835-838. BARTUNIK, H. D. (1983). Nucl. Instrum. Methods, 208, 523-533. BILDENBACK, D. H., MOFFAT, K. 86 SZEBENYI, D. (1984). Nucl. lnstrum. Methods. In the press. CORK, C., HAMLIN, R., VERNON, W., XUONG, NG. H. 86 PEREZ- MENDEZ, V. (1975). Acta Cryst. A31,702-703. GREENBOUGH, T. J. 86 HELLIWELL, J. R. (1983). Prog. Biophys. Mol. Biol. 41, 67-123. HAMLIN, R. (1982). Trans. Am. Crystallogr. Assoc. 18, 95-123. HOWARD, A., NIELSEN, C. 86 XUONG, NG. H. (1985). Methods Enzymol. In the press. MATTHEWS, B. 86 CZERWINSKI, E. (1975). Acta Cryst. A31,480- 487. PHILLIPS, J. C. 86 HODGSON, K. O. (1980). In Synchrotron Radi- ation Research, edited by H. WINICK 86 S. DONIACH. New York: Plenum. XUONG, NG. H., FREER, S. T., HAMLIN, R., NIELSEN, C. 86 VERNON, W. (1978). Acta Cryst. A34, 289-296. XUONG, NG. H., NIELSEN, C., HAMLIN, R. 86 ANDERSON, D. (1985). J. Appl. Cryst. In the press.
`
`NOVARTIS EXHIBIT 2051
`Par v Novartis, IPR 2016-00084
`Page 3 of 3