HI
`
`53706
`
`MADISON
`1215 H DAYTON ST
`UNIV HISCONSIN
`GEOLOGY GEOPHY LIB
`00000905836SC Ol/03/92N9119
`17
`************ CAR-RT SORT ** CR03
`
`Par Pharm., Inc.
`Exhibit 1029
`Page 001
`
`

`
`AMERICAN
`AssociATION FOR THE
`. ADVANCEMENT OF
`SCIENCE
`
`SciENCE
`
`ISSN 0036-8075
`!0 MAY 1991
`VOLUME 252
`NuMBER 5007
`
`755 This Week in Science
`
`M DtJni.JQij- 757 Technology for America's Future
`Letters
`
`763 Science in the Persian Gulf: T. M. BoYCE • Math Problems: T. M. MuRPHY;
`M. W. LEVINE; R. D. HANsoN; C. B. HATFIELD; D. E. KosHLAND, JR. • Energy
`Savings: A. B. LoviNS; G. M. BARNWELL • EPA Committee: M. E. O'CONNOR •
`Crystal Stmcture of Bee-Venom Phospholipase A2 : Correction: D. L. ScoTT,
`Z. OTWINOWSKI, M. H . GELB, P. B. SIGLER • Membrane-Bound Phosphotyrosine
`Phosphatases: A. F. WILLIAMS
`
`ScienceSco e
`News & Con1ment
`
`Research News
`
`Articles
`
`Research Articles
`
`767 Sweeping overhead rates under the mg; gambling with Poker Flat science; etc.
`
`768 Baltimore Throws in the Towel • David Baltimore's Mea Culpa
`771 The True Source of HIV?
`772 Science Under Wraps in Prince William Sound
`773 Science Academy Elects New Members
`774 Briifrngs: Hidden Costs of the Space Station • A Big Gift from Big Oil • A Billion
`Bucks for Materials • Congressional Day • Ten Years for the Brain • Cuban AIDS
`Control • Biotechnology Execs Earn More • Correction
`
`776 Engineering Dogma Gives Way to Chaos • Flying High with Chaos Control
`778 A New Ball Game in Nuclear Physics
`779 How Peptide Hormones Get Ready for Work
`781 Praying Mantises Play Top Gun
`782 Sex and the Single Gene
`783 Deep Rocks Stir the Mantle Pot
`
`789 Reproductive Behavior and Health in Consanguineous Marriages: A. H . BITTLES,
`W. M. MAsoN, J. GREENE, N. A. RAo
`795 Neutron Scattering: Progress and Prospects: J.D. AxE
`802 Diversity of G Proteins in Signal Transduction: M. I. SIMoN, M.P. STRATHMANN,
`N. GAUTAM
`
`809 Zinc Finger-DNA Recognition: Crystal Stmcture of a Zif268-DNA Complex at
`2.1 A: N. P. PAVLETICH and C. 0. PABo
`817 A New Cofactor in a Prokaryotic Enzyme: Tryptophan Tryptophylquinone as the
`Redox Prosthetic Group in Methylamine Dehydrogenase: W. S. MciNTIRE,
`D. E. WEMMER, A. CHISTOSERDOV, M. E. LIDSTROM
`
`•
`
`•
`
`SCIENCE (ISSN 0036-8075) Is published weekly on Friday, except the last week In December, by the Am~rican
`Association for the Advancement of Science, 1333 H Street, NW, Washington, DC 20005. Second-class postage
`(publication No. 484460) paid at Washington, DC, and additional mailing offices. Copyright © 1991 by the American
`Association for the Advancement of Science. The title SCIENCE is a registered trademark of the AAAS. Domestic
`Individual membership and subscription (51 Issues): $82 ($50 allocated to subscription). Domestic institutional ·
`subscription (51 issues): $150. Foreign postage extra: Canada $46, other (surface mail) $46, air freight $90. First class,
`airmail, school-year, and student rates on request. Change of address: allow 6 weeks, giving old and new addresses
`and 11-digit account number. Postmaster: Send change of address to Science, P.O. Box 1723, Riverton, NJ 08077.
`Single copy sales: $6.00 per issue prepaid includes surface postage; Guide to Biotechnology Products and
`Instruments, $20. Bulk rates on request. Authorization to photocopy material for internal or personal use under
`circumstances not falling within the fair use provisions of the Copyright Act is granted by AAAS to libraries and other
`users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that the base
`fee of $1 per copy plus $0.10 per page is paid directly to CCC, 27 Congress Street, Salem, Massachusetls 01970. The
`Identification code for Science Is 0036-8075/83 $1 + .to. Science is indexed in the Reader's Guide to Periodical
`Literature and In several specialized indexes.
`The American Association for the Advancement of Science was founded In 1848 and incorporated in 1874. Its
`objectives are to further the work of scientists, to facilitate cooperation among them, to foster scientific freedom and
`responsibility, to Improve the effectiveness of science in the promotion of human welfare, to advance education in
`science, and to increase public understanding and appreciation of the importance and promise of the methods of
`science in human progress.
`
`752
`
`SCIENCE, VOL. 252
`
`Par Pharm., Inc.
`Exhibit 1029
`Page 002
`
`

`
` 7KLV PDWHULDO PD\ EH SURWHFWHG E\ &RS\ULJKW ODZ 7LWOH  86 &RGH

`
`Our results show that the presence of the
`female-specific tra and the tra-2 products
`promote female-specific splicing of dsx pre(cid:173)
`mRNA and that the 13··nt sequences in the
`female exon and the female-specific acceptor
`sequence participate in regulation of dsx
`expression. The tra and tra-2 products may
`interact directly with the 13-nt sequences in
`the female-specific exon, and such interac(cid:173)
`tions may allow the suboptimal female-spe(cid:173)
`cific polypyrimidine stretch to be recognized
`as a splicing signal, thus resulting in en(cid:173)
`hancement of the use of the female-specific
`acceptor site.
`
`REFERENCES AND NOTES
`
`l. R. Nothiger and M. Steinmann-Zwicky, Trends
`Genet. 1, 209 (1985); T. W. Cline, in Origin and
`Evolution cif Sex, H. 0. Halvorson and A. Monroy,
`Eds. (Liss, New York, 1985), pp. 301-327; B. S.
`Baker, Nature 340, 521 (1989).
`2. B. S. Baker and K. A. Ridge, Genetics 94, 383
`(1980).
`3. B. S. Baker and M. F. Wolfner, Genes Dev. 2, 477
`(1988); K. C. Burtis and B. S. Baker, Cell 56, 997
`(1989).
`4. R. N. Nagoshi, M. McKeown, K. C. Burtis, J. M.
`Belote, B. S. Baker, Cell 53, 229 (1988).
`5. M. McKeown, J. M. Belote, R. T. Boggs, ibid., p.
`887; B. A. Sosnowski, J. M. Belote, M. McKeown,
`ibid. 58, 449 (1989).
`6. K. Inoue, K. Hoshijima, H. Sakamoto, Y. Shimura,
`Nature 344, 461 (1990).
`7. T. Schupbach, Dev. Bioi. 89, 117 (1982); J. M.
`Belote and B.S. Baker, Proc. Nat/. Acad. Sci. U.S.A.
`79, 1568 (1982); Dev. Bioi. 95, 512 (1983).
`8. H. Amrein, M. Gorman, R. Nothiger, Cell 55,
`1025 (1988); T. J. Goralski, J.-E. Edstrom, B. S.
`Baker, ibid. 56, lOll (1989).
`9. C. C. Query, R. C. Bentley, J.D. Keene, ibid. 57, 89
`(1989); I. W. Mattaj, ibid., p. l.
`10. J, M. Belote, M. McKeown, R. T. Boggs, R.
`Ohkawa, B. A. Sosnowski, Dev. Genet. 10, 143
`(1989).
`ll. The region from and including the entire third exon
`to the Pvu II site located 1284 bp downstrean1 of the
`female acceptor site, and the region from the Ace I
`site located 250 bp upstream of the male acceptor
`site to the site located 64 bp downstream of the
`donor site of tl1e fifth exon, were joined and inserted
`into the copia vector.
`12. M. McKeown, J. M. Belote, R. T. Boggs, Cell 53,
`887 (1988).
`13. K. Hoshijima et a/., unpublished observations.
`14. R. N. Nagoshi and B. S. Baker, Genes Dev. 4, 89
`(1990).
`15. The dsx fragment that extends from the third exon to
`1128 bp downstream of the female acceptor site and
`which excludes tl1e polyadenylation signal (Af) was
`fused to the _fiz fragment that contains a portion of
`the intron and the following second exon {corre(cid:173)
`sponding to bases 825 to 2534 of the published
`sequence [A. Laughon and M.P. Scort, Nature 310,
`25 (1984)]}. The resulting fragment was inserted
`into the copia vector.
`16. The region from the third exon (138 bp) to the site
`48 bp downstream of the donor site of the third
`exon, and the region from the site 250 bp upstream
`of the male acceptor site to the site 64 bp down(cid:173)
`stream of the donor site of the fifth exon, were
`joined and inserted into the copia vector.
`17. The copia-dsx deletion mutants were constmcted as
`follows:
`the regions between 345 to 460 bp
`(Rl-5·6), 234 to 459 bp (R5·6), 190 to 480 bp
`(R6), 344 to 599 bp (R1), or 234 to 599 bp (RO)
`downstream of the female acceptor site were deleted
`from copia-dsx, and Kpn I linkers were inserted. A
`polymerase chain reaction (PCR) fragment tl1at
`contained tl1e region 268 to 627 bp downstream of
`the female-specific acceptor site was inserted at the
`Kpn I linker of the RO constmct in the correct
`
`836
`
`orientation ( SR) or in the opposite orientation
`(ASR). PCR was performed essentially as described
`[R. K. Saiki eta/., Science 239,487 (1988)].
`18. 'M. R. Green,A1111t1. Rev. Genet. 20, 671 (1986).
`19. Oligonucleotide-directed mutagenesis was per(cid:173)
`formed essentially as described [M. J. Zoller and M.
`Smitl1, Methods Enzymol. 100, 468 (1983)].
`20. The region 340 to 1128 bp downstream of the
`female acceptor site was deleted from copia-dtz,
`
`copia-dtz(Pyl8), and copia-dtz(Pyl3). Pre·mRNA
`expressed from the deletion construct of copia-dtz
`was spliced at the _fiz acceptor site even in the
`presence of the Ira and lra-2 products in Kc cells.
`21. We thank M. McKeown for Ira eDNA and H. Amrein
`tra-2 eDNA. Supported by Grant-in-Aid
`for
`62065009 for specially promoted research from the
`Ministry of Education, Science and Culture ofJ a pan.
`l November 1990; accepted l March 1991
`
`Solution Structure of FKBP, a Rotamase Enzyme and
`Receptor for FI(506 and Rapamycin
`STEPHEN W. MICHNICK, MICHAEL K. ROSEN, THOMAS J. WANDLESS,
`MARTIN KARPLUS, * STUART L. SCHREIBER*
`
`Immunophilins, when complexed to immunosuppressive ligands, appear to inhibit
`signal transduction pathways that result in exocytosis and transcription. The solution
`structure of one of these, the human FK506 and rapamycin binding protein (FKBP),
`has been determined by nuclear magnetic resonance (NMR). FKBP has a previously
`unobserved antiparallel 13-sheet folding topology that results in a novel loop crossing
`and produces a large cavity lined by a conserved array of aromatic residues; this cavity
`serves as the rotamase active site and drug-binding pocket. There are other significant
`structural features (such as a protruding positively charged loop and an apparendy
`flexible loop) that may be involved in the biological activity of FKBP.
`
`F KBP IS A SOLUBLE, CYTOSOLIC RE(cid:173)
`
`ceptor (1, 2) for the immunosuppres(cid:173)
`sants FK506 and rapamycin (3). Both
`FKBP and cyclophilin (4-6), which is a
`receptor for the immunosuppressant cyclo(cid:173)
`sporin A (CsA), catalyze the interconversion
`of cis- and trans-rotamers of the peptidyl(cid:173)
`prolyl amide bond of peptide and protein
`substrates. These rotamases are inhibited by
`their respective immunosuppressive ligands.
`Mechanistic studies suggest that a complex of
`immunophilin (immunosuppressant binding
`protein) and ligand interferes with the intracel(cid:173)
`lular transport of proteins involved in signal
`transduction pathways associated with both
`exocytosis (7) and transcription (3, 8).
`In this report we present the solution
`structure of human FKBP obtained by nu(cid:173)
`clear Overhauser effect (NOE)-restrained
`molecular dynamics (rMD) simulations (9).
`The NMR structures satisfYing the NOB
`and empirical energy function restraints
`have backbone root-mean-square deviations
`(RMSDs) from the refined average structure
`in the range of 0.80 to 1.4 A for the 13
`strands, 0.19 to 0.40 A for the a helix, and
`1.02 A to 1.72 A for all residues except 83
`to 90 (see below). Many side chains, partic(cid:173)
`ularly aromatics, are well defined. FKBP has
`a novel folding topology in .that two loops
`that connect the strands of an anti parallel 13
`sheet cross one another. Hydrophobic side
`
`Department of Chemistry, Harvard University, Cam(cid:173)
`bridge, MA 02138.
`
`*To whom correspondence should be addressed.
`
`chains in the core of the protein form a
`large, deep pocket that includes the rota(cid:173)
`mase active site and the drug-binding site.
`The structure of the binding site appears to
`be highly conserved in related FK506- and
`rapamycin-binding proteins.
`Sequence-specific assignments of 92% of
`the observable 1 H resonances in FKBP have
`been made by use of a combination of
`homonuclear and heteronuclear two-dimen(cid:173)
`sional NMR techniques as reported earlier
`(10). Structural restraints for the dynamics
`simulations were obtained through empiri(cid:173)
`cal calibration of cross-peak magnitudes in
`nuclear Overhauser effect spectroscopy
`(NOESY) spectra recorded with mixing times
`
`107
`
`Fig. 1. a-Carbon ribbon drawing of human
`FKBP; the five-stranded 13 sheet, a helix, and
`connecting loops are indicated, as well as the
`COOH- and NH2 -termini and certain residues of
`interest.
`
`SCIENCE, VOL. 252
`
`Par Pharm., Inc.
`Exhibit 1029
`Page 003
`
`

`
`Fig. 2. (A) The a-carbon traces of 13 structures
`generated by restrained molecular dynamics are
`shown in blue. (SA)ref> which includes side chains
`of aromatic residues in the drug-binding pocket,
`is shown as the heavy line. Aromatic side chains
`are color-coded: Trp, yellow; Phe, orange; and
`Tyr, green. The NH2 - and COOH-termini are
`indicated with green and red dot surfaces, respec(cid:173)
`tively. (B) (SA),cr, showing location of the ligand
`binding site formed by the 13 sheet and the a helix;
`certain residues are numbered. (C) Expanded
`view of (SA)ref> showing crossing of the loops
`cmmecting the first and fourth (relevant residues
`in orange) and the fifth and second (relevant
`residues in yellow) strands of the 13 sheet; known
`interstrand hydrogen bonds are indicated. Rele(cid:173)
`vant residues in the third strand are shown in
`green. (D) Hydrophobic residues in the drug(cid:173)
`binding pocket of (SA),er· Side chains are color(cid:173)
`coded as in (A), with additional aliphatic residues
`in purple.
`
`of 50, 100, 150, and 200 ms versus known
`distances in regular secondary structural el(cid:173)
`ements (11 ). Of particular use in unambig(cid:173)
`uously determining long-range NOEs were
`two FKBP samples biosynthetically deuter(cid:173)
`ated at the methyl protons of Leu and Ile
`and of Val and Ile (12, 13). A total of 860
`NOB-derived distance constraints was used
`in the simulations. In addition, 43 Xi dihe(cid:173)
`dral angle restraints (in the range ±60° or
`± 120°) and 44 backbone<!> angle restraints (in
`the range ± 60°, ± 50°, or ± 40°) were used; the
`angle constraints are based on analysis of cou(cid:173)
`pling constants and cross-peak magnitudes
`measured in correlated spectroscopy (COSY)
`spectra, and on NOB data (14). In 32 cases, Xi
`could be restricted to a ±60° range, allowing
`stereospecific assignment of side chain protons
`based on intraresidue NOEs from NH and
`CaR. Hydrogen bond constraints were used
`for 25 slowly exchanging amide protons. Sev(cid:173)
`enty-nine NOEs, six hydrogen bonds, and six
`backbone dihedral angles were determined in
`an iterative marmer based on unrefined, re(cid:173)
`strained structures. The rMD simulations
`were carried out by using a simulated anneal(cid:173)
`ing protocol with the program X-PLOR (9).
`Statistics indicative of the accuracy and preci(cid:173)
`sion of the experimentally determined struc(cid:173)
`ture of FKBP and a description of the proto-
`
`REPORTS
`
`837
`
`10 MAY 1991
`
`Par Pharm., Inc.
`Exhibit 1029
`Page 004
`
`

`
`col used to generate these structures are
`provided in Table l.
`The solution structure of FKBP (Figs. 1
`and 2) is characterized by a large am(cid:173)
`phiphilic, antiparallel five-stranded ~ sheet
`with +3, + 1, -3, + 1 topology (15). The
`strands of the sheet are composed of resi(cid:173)
`dues 1 to 7, 20 to 29, 46 to 49, 7l to 77,
`and 96 to 107. An amphiphilic a helix,
`formed from residues 59 to 65, packs
`against the hydrophobic face of the sheet at
`an angle of 60° with respect to the long axis.
`The helix is tethered to the fifth and second
`strands of the sheet by nine- and five-residue
`loops, respectively. The sheet has a right(cid:173)
`handed twist and wraps around the helix to
`form a well-ordered hydrophobic core (Fig.
`1); this structure is in accord with a large
`number of long-range side chain NOE in(cid:173)
`teractions. Backbone hydrogen-bond pro(cid:173)
`pensities are satisfied for the ~ sheet by
`interstrand amide-carbonyl contacts and for the
`a helix by interresidue amide-carbonyl i, i + 3
`contacts. There are no side chain hydrogen
`bonds between the helix and the sheet, indicat(cid:173)
`ing stabilization of these structures through
`hydrophobic (van der Waals) interactions. The
`loops connecting the ~ strands contain second(cid:173)
`ary structural elements (including several ~
`turns) that satisfY some of the backbone hydro(cid:173)
`gen bond propensities. The loops are well
`defined by medium- and long-range NOE con(cid:173)
`tacts except for two regions, residues 37 to 43
`and 83 to 90 (see Fig. 2, A and B).
`A notable feature of FKBP resulting from
`the + 3, + 1, -3, + l topology of the ~ sheet is
`a topological crossing of the loops Ser8-Gly19
`and Leu50-Gln70. A view of this region of the
`protein is shown in Fig. 2C. Although crossing
`topologies have been obsetved in proteins con(cid:173)
`taining parallel ~ sheets (16), they were pre(cid:173)
`sumed to be forbidden in antiparallel sheets
`(17). The absence of loop crossings has been
`attributed to the difficulties of obtaining effi(cid:173)
`cient side chain packing, of satisfYing the hy(cid:173)
`drogen bond propensities of the backbone
`amides of both segments, and of describing a
`simple folding pathway (18). In FKBP, the
`structure of the crossing segments is deter(cid:173)
`mined by NOEs including Thr14-Ser67, Glyl2-
`Ser67 Arg13-Ser67 Pro16-Val68 and Pro16-
`Leu1~6. There are ~an der Waal~ contacts be(cid:173)
`tween the side chains of Pro16 and Leu106,
`Val68 and Leu103, and Thr14 and Val68. The
`hydrogen-bonding propensities of many resi(cid:173)
`dues in the crossing region (Pro9, Asp11
`Thr14, Met66, Ser67, and Val68) are satisfied
`through backbone-backbone and backbone(cid:173)
`side chain interactions. All amide protons in
`the inner strand and two amide protons in the
`outer strand (Lys17 and Gln20) exchange slow(cid:173)
`ly when the protein is dissolved in D 20; this
`result is indicative of hydrogen bonds stabiliz(cid:173)
`ing this topology.
`
`,
`
`838
`
`In a preliminary study of the FKBP(cid:173)
`FK506 and FKBP-rapamycin complexes we
`noted NOEs from the indole ring of Trp59
`and several other unidentified aromatic pro(cid:173)
`tons to the pipicolinyl ring of the bound
`drugs (13). Based on the solution conforma(cid:173)
`tion of FKBP, it was possible to define the
`binding region as an extensive aromatic clus(cid:173)
`ter composed of Tyr26 Phe46 Phe48
`Trp59, Tyr82, and Phe99: These' residue~
`pack together with a number of aliphatic
`residues to form a large, well-defined hydro(cid:173)
`phobic core (Fig. 2D). Thus, FK506 and
`rapamycin bind in the hydrophobic pocket
`that consists of a twisted and curved anti par(cid:173)
`allel ~-sheet platform with walls formed by
`an a helix and a long loop (Fig. 2B). The
`location and detailed structure of the bind(cid:173)
`ing site has been confirmed by the subse(cid:173)
`quent x-ray structure of the FKBP-FK506
`complex (19). The architecture of the bind(cid:173)
`ing site, whose natural ligands are not
`
`known but may include peptides, is very
`different from those formed by ~ barrels in a
`number of proteins that bind hydrophobic
`ligands [such as P2 myelin protein (20) and
`bilin binding protein (21)]. However, al(cid:173)
`though the overall structures are very differ(cid:173)
`ent, the FKBP binding site has elements in
`common with that of the class I major
`histocompatibility complex (MHC) glyco(cid:173)
`protein HLA-A2 (22), which has a flatter
`antiparallel ~-sheet platform with walls
`formed by two long a helices. Peptide seg(cid:173)
`ments that form loops at the open end of the
`pocket, specifically, Asp37 to Asn43 and
`Gly82 to Ile90, are not well defined (see Fig.
`2, A and B). The lack of long-range NOEs
`in both segments coupled with previous
`evidence of conformational exchange in the
`second (10) indicate that they may be flexi(cid:173)
`ble in solution. A comparison of 1 H and
`15N chemical shifts obtained from single(cid:173)
`quantum coherence (SQC) spectra of 15N-
`
`Table 1. Structural and energetic statistics for FKBP. Structure determination was achieved with a
`simulated annealing protocol consisting of three stages: (i) a conformational search phase; (ii) an
`annealing phase; and (iii) a refinement phase (9, 30). Numbers in parentheses are the number of
`restraints in each class or the number of bonds, angles, and dihedrals in the structure geometry
`statistics. SA; is the ith structure obtained by the protocol given above. The SA; column gives the
`average and standard deviations for the indicated variables obtained from the 21 structures with the
`lowest empirical energies and residual violations of experimental restraints. (SA),.r represents the
`average structure of SA; least-square fit to each other including all atoms (except for residues 83 to
`90) and refined with 1000 steps of steepest-descent energy minimization. The RMSDs are from the
`upper or lower bounds of the distance restraints; none of the structures showed deviations greater
`than 0.25 A. Restraints were classified into the following categories based on strong, medium, and
`weak NOEs (11): backbone, 2.5 ± 0.5 A, 2.75 ± 0.75 A, 3.5 ± 1.5 A; side chain, 2.5 ± 0.5 A,
`3.0 ± 1.0 A, 3.5 ± 1.5 A; 0.5 A was added to the upper limits for distances involving methyl
`protons. Average RMSDs and standard deviations of the final structures SA; against the refined
`average structure (SA),.r· For [3 sheet: backbone, 1.01 ± 0.16; all atoms, 1.92 ± 0.14. For ex helix:
`backbone, 0.31 ± 0.07; all atoms, 1.05 ± 0.16. For all atoms except residues 83 to 90: backbone,
`1.45 ± 0.20; all atoms, 2.49 ± 0.17.
`
`Parameter
`
`RMSDsftotll experimellfal distance restrai11ts (A)
`(910)
`0.024 ±
`
`(246)
`(254)
`(39)
`(146)
`
`(12)
`(5)
`(25)
`(133)
`(50)
`
`0.011 ±
`0.024 ±
`0.041 ±
`0.031 ±
`
`0.008 ±
`0.012 ±
`0.018 ±
`0.023 ±
`0.034 ±
`
`Total
`Peptide backbone restraints
`Intraresidue
`Interresidue sequential (li-jl = 1)
`Interresidue short-range (li-jl ,.; 5)
`Interresidue long-range (li-jl > 5)
`Side-chain restraints
`Intraresidue
`Interresidue sequential (li-jl = 1)
`Interresidue short-range (li-jl s 5)
`Interresidue long-range (li-jl > 5)
`Hydrogen-bond restraints*
`Deviations from idealized geometryt
`Bonds (A)
`Angles (degrees)
`Imp ropers (degrees)
`
`(SA),. r
`
`0.008
`
`0.022
`0.008
`0.021
`0.051
`
`0.024
`0.003
`0.000
`0.020
`0.037
`
`0.002
`
`0.003
`0.003
`0.008
`0.004
`
`0.010
`0.018
`0.008
`0.004
`0.008
`
`0.000
`0.064
`0.030
`
`0.013
`3.240
`0.490
`
`SCIENCE, VOL. 252
`
`Etotal
`Ebond
`Bangle
`Edih~dral
`Eimpro per
`Evdw
`
`0.013 ±
`(1682)
`3.360 ±
`(3047)
`0.460 ±
`(498)
`Potetltial enetgy statistics (kcal/mol)
`± 120
`-2280.0
`-2170.0
`±
`3.1
`47.8
`49.9
`± 20
`486.0
`652.0
`± 18
`512.0
`532.0
`1.8
`15.8
`16.5
`±
`± 15
`-358.0
`-336.0
`-3200.0
`± 120
`-2920.0
`Eelectrosratic
`1.7
`5.9
`5.6
`±
`Echide
`3.9
`22.3
`27.0
`±
`ENOE
`*Backbone hydrogen bond restraints were assigned to ranges 'NH- O = 1.9 ± 0.5 A and ' N -o = 2.8 ± 0.5
`A.
`tidealized geometries based on CHARMM parameters (31) .
`
`Par Pharm., Inc.
`Exhibit 1029
`Page 005
`
`

`
`labeled FKBP and the 15N-labeled FKBP(cid:173)
`FK506 complex suggests that significant
`changes occur in these loop regions follow(cid:173)
`ing drug binding. Thus, the distinct inhibi(cid:173)
`tory effects that result from FK506 and
`rapamycin binding ( 8) may involve their
`influence on the geometry of these loops.
`We have recently provided evidence sug(cid:173)
`gesting that the mechanism of rotamase
`catalysis is due to noncovalent stabilization
`of a twisted amide in the transition state of
`the reaction (23, 24), rather than formation
`of a covalent tetrahedral intermediate. A
`similar mechanism has been proposed for
`the rotamase cyclophilin (25, 26). All Lys,
`Ser, and Thr side chains in FKBP are direct(cid:173)
`ed away from the active site, and the Ca of
`Cys22 is ll A from that ofTrp59 toward the
`narrow end of the molecule. The pipecolinyl
`ring ofFK506, which contacts Trp59
`, prob(cid:173)
`ably mimics the proline ring of a natural
`peptide substrate that is subject to rotamase
`catalysis. Thus, the aforementioned residues
`are too far from the active site to add to the
`peptidyl-prolyl amide carbonyl and facili(cid:173)
`tate rotation about the C-N bond. Site(cid:173)
`directed mutagenesis is being used to deter(cid:173)
`mine the role of the tyrosines and other key
`residues found in the active site (Fig. 2D).
`Several higher molecular weight FK506-
`and rapamycin-binding proteins have re(cid:173)
`cently been reported (27). The immunophil(cid:173)
`ins of molecular weight 13,000 and 27,000
`contain FKBP-like domains of -llO amino
`acids that share high sequence identity to
`FKBP (28). Aromatic residues that corre(cid:173)
`, and Phe99
`, Tyr82
`spond to Trp59
`, which
`line the drug-binding pocket, are conserved in
`these proteins and in all FIZBPs identified from
`different organisms to date (8), suggesting that
`the ligand-binding pocket is similar in all FK(cid:173)
`BPs. Thus, the present structure of human
`FIZBP may be relevant to understanding not
`only its own enzymatic and drug-binding
`properties, but those of all members of this
`emerging family of proteins.
`Note added in proif: A report has appeared
`noting the identity of FIZBP to inhibitor-2 of
`protein kinase C (PKCI-2) (29); however, we
`find that FIZBP does not inhibit the kinase
`activity of isolated protein kinase C or protein
`kinase C-mediated events in cells (32).
`
`REFERENCES AND NOTES
`
`l. M. W. Harding et al., Nature 341, 758 (1989).
`2. J, J. Siekierka eta/., ibid., p. 755.
`3. B. E. Bierer eta/., Proc. Nat/. Acad. Sci. U.S.A. 87,
`9231 (1990).
`4. R. E. Handschumacher eta/., Science 226, 544 (1984).
`5. G. Fisher eta/., Nature 337, 476 (1989).
`6. N. Takahashi, T. Hayano, M. Suzuki, ibid., p. 473.
`7. T. Hultsch eta/., Proc. Nat/. Acad. Sci. U. S. A., in
`press.
`8. S. L. Schreiber, Science 251, 283 (1991).
`9. A. T. BriingerandM. Karplus,Acc. Chem. Res. 24,54
`(1991) for review; R. Kaptein, E. R. P. Zuiderweg, R.
`M. Sd1eek, R. Boelens, W. F. van Gunsteren, ]. Mol.
`
`10MAY 1991
`
`Bioi. 182,179 (1985); A. T. Btiinger, G. M. Clore, A.
`M. Gronenborn, M. Katpllh,, Proc. Nat/. Acad. Sci.
`U.S.A. 83, 3801 (1986); M. Nilges, G. M. Clore, A.
`M. Gronenborn, FEBS Lett. 229, 317 (1988); M.
`Nilges, in Computational Aspeds qfthe Study q[Bio/ogical
`Macromolecules by NMR, J, C. Hoch, Ed. (Plenum,
`New Y ark, in press); A. T. Briinger, X-PLOR Manual,
`Version2.1, (Yale University, New Haven 1990).
`10. M. K. Rosen, S. W. Michnick, M. Karplus, S. L.
`Schreiber, Biochemistry, in press.
`11. A. Kumar, G. Wagner, R. R. Ernst, K. Wuthrich, ].
`Am. Chem. Soc. 103, 3654 (1981); M.P. WilliatnSon
`eta/.,]. Mol. Bioi. 182, 295 (1985).
`12. R. F. Standaert, A. Galat, G. L. Verdine, S. L.
`Schreiber, Nature 346, 671 (1990).
`13. T. J. Wandless, S. W. Michnick, M. K. Rosen, M.
`Karplus, S. L. Schreiber, ]. Am. Clwn. Soc. 113,
`2339 (1991).
`14. A. Pardi, M. Billeter, K. Wuthrich, ]. Mol. Bioi.
`180, 741 (1984); D. Neuhaus, G. Wagner, M.
`Vasak, J. H. R. Kagi, K. Wuthrich, Em.]. Biochem.
`151, 257 (1985); G. Wagner et al.,]. Mol. Bioi.
`196, 611 (1987); E. R. P. Zuiderweg, R. Boelens,
`R. Kaptein, Biopolymers 24, 601 (1985).
`15. J. S. Richardson, Nature 268, 495 (1977).
`16. For example, gyceraldehyde phosphate dehydrogenase
`[D. Moras et al.,]. Bioi. Chem. 250,9137 (1975)].
`17. C. Chothia, Amw. Rev. Biochem. 53, 537 (1984);
`Y. N. Chirgadze,Acla Crystallogr. A43, 405 (1987).
`18. 0. B. Ptitsyn and A. V. Finkelstein, Q. Rev. Bio(cid:173)
`phys. 13, 339 (1980).
`19. G. D. Vm Duyne, R. F. Standaert, P. A. Karplus, S.
`L. Schreiber, J. Clardy, Scieuce, 252, 839 (1991).
`20. T. A. Jones, T. Bergfors, J. Sedzik, T. Unge, EMBO
`]. 7, 1597 (1988).
`21. R. Huber eta/.,]. Mol. Bioi. 198, 499 (1987).
`22. P. J. Bjorkmanel a/., Nature 329,506 (1987); ibid.,
`p. 512.
`23. M. K. Rosen, R. F. Standaert, A. Galat, M. Nakat(cid:173)
`suka, S. L. Schreiber, Scieuce 248, 863 (1990).
`24. M. W. Albers, C. T. Walsh, S. L. Schreiber, ]. Org.
`Chem. 55, 4984 (1990).
`25. J. Liu, M. Albers, C. Chen, S. L. Schreiber, C. T.
`Walsh, Proc. Nat/. Acad. Sci. U.S.A. 87, 2304 (1990).
`26. R. K. Harrison and R. L. Stein, Biochemistry 29,
`1684 (1990).
`27. H. Pretzel al.,]. Am. Chem. Soc. 113, 1409 (1991).
`28. Y.-J. Jin, M. W. Albers, B. E. Bierer, S. L. Schreiber,
`S. J, Burakoff, Proc. Nat/. Acad. Sci. U. S. A., in
`press. A. Galat, W. S. Lane, R. F. Standaert, S. L.
`Schreiber, in preparation.
`29. M.G. Goebl, Cel/64, 1051 (1991).
`30. Starting strucrures for the simulated annealing pro(cid:173)
`tocol consisted of FKBP polypeptide chains with
`random backbone dihedral angles. The conforma(cid:173)
`tional search phase consisted of 20 ps of molecular
`
`dynamics at l 000 K under the influence of a force
`field devised for simulated annealing srudies. A
`quartic repulsion potential was used in the force field
`in place of the standard CHARMM van der Waals
`potential for purposes of computational efficiency.
`The force constant of the repulsive potential (ERE(cid:173)
`PEL) was scaled to 0.002 kcal mol- 1 A - 4 during the
`conformational search to allow atoms to pass freely
`through each other. The distance restraint potential
`target function (ENoE) was a square well with
`harmonic walls at the upper and lower bounds of the
`distance restraints and a harmonic plus linear switch(cid:173)
`ing function at 0.5 A above the upper bound. The
`harmonic force constant was 50 kcal mol- 1 A - 2 and
`the linear component had a slope of 5 kcal mol- 1
`A -l. The purpose of this phase was to obtain an
`unbiased search of conformational space with a force
`field dominated by the experimental target function.
`The annealing phase consisted of two stages. First,
`during a 10-ps interval the slope of the linear com(cid:173)
`ponentofENoE was increased to 50 kcal mol- 1 A- 1
`and the scale of EREPEL was raised to 0.1 kcalmol- 1
`A - 4 . Second, the distance restraint potential was
`made harmonic at both the lower and upper distance
`bounds, experimental dihedral restraints were intro(cid:173)
`duced, the scale of EREPEL was increased to 4.0 kcal
`mol- 1 A -4, md the temperarure was adjusted from
`1000 K to 300 K in steps of 25 K over 2.8 ps. The
`experimental dihedral potential (Ecdihc) was a square
`well with harmonic walls at the upper and lower
`bounds with a force constant of 200 kcal mol- 1
`rad- 2
`. The refinement phase consisted of 1000 steps
`of steepest descent minimization in which the full
`CHARMM force field including electrostatic and
`van der Waals nonbonded terms were used (31).
`Statistics were calculated with the program X-PLOR
`(9). Energy analysis was performed with the stan(cid:173)
`dard CHARMM force field (31).
`31. B. R. Brooks eta/.,]. Comp11t. Chem. 4, 187 (1983).
`32. M. W. Albers eta/., Biomed. Chem. Lett., in press.
`33. Coordinates of the 21 final strucrures and of the mini(cid:173)
`mized average strucrure · will be deposited in the
`Brookhaven Protein Data Bank along with a list of all
`experimental restraints. Supported by the National In(cid:173)
`stirute of General Medical Scien= (GM-38627,
`awarded to S.L.S; GM-30804, awarded to M.K.).
`National Science Foundation predoctoral fellowships to
`M.K.R. and T.J,W. are gratefully acknowledged. We
`thankS. Shambayati for pointing out the novelty of the
`loop crossing topology in FKBP. NMR spectra were
`obtained through the auspices of the Harvard Univer(cid:173)
`sity Department of Chemistry Instrumentation Center,
`which was supported by NlH grant 1-Sl0-RR04870
`and NSF grant CHE 88-14019.
`
`29 January 1991; accepted 1 April1991
`
`Atomic Structure of FI<BP-FI(506, an
`Immunophilin-Immunosuppressant Complex
`GREGORY D. VAN DUYNE, ROBERT F. STANDAERT,
`P. ANDREW KARPLUS, STUART L. SCHREIBER,* JON CLARDY*
`
`The structure of the human FK506 binding protein (FKBP), complexed with the immu(cid:173)
`nosuppressant FK506, has been determined to 1.7 angstroms resolution by x-ray crystal(cid:173)
`lography. The conformation of the protein changes little upon complexation, but the
`conformation ofFK506 is markedly different in the bound and unbound forms. The drug's
`association with the protein involves five hydrogen bonds, a hydrophobic binding pocket
`lined with conserved aromatic residues, and an unusual carbonyl binding pocket. The nature
`of this complex has implications for the mechanism of rotarnase catalysis and for the
`biological actions of FK506 and rapamycin.
`
`I N AN ACCOMPANYING REPORT (1), Tim
`
`function of FIZBP (2, 3) was discussed,
`and its structure, determined by nuclear
`
`Overhauser effect (NO E)-restrained molecular
`dynamics, was described. After this structure
`had been determined, experiments were