Post-synaptic Density Disc Large Zo-1 (PDZ) Domains: From Folding and Binding to Drug Targeting

Full text

(1)

(2)

(3)

(4)

(5)

(6) .

(7)

(8)

(9)

(10)

(11)

(12) ! .

(13)

(14)

(15) . ! "! #$%&%'%##%().

(16) Dissertation presented at Uppsala University to be publicly examined in B22, BMC, Uppsala, Friday, September 3, 2010 at 10:15 for the degree of Doctor of Philosophy (Faculty of Medicine). The examination will be conducted in English. Abstract Chi, C. 2010. Post-synaptic Density Disc Large Zo-1 (PDZ) Domains. From Folding and Binding to Drug Targeting. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 573. 42 pp. Uppsala. ISBN 978-91-554-7836-0. Understanding how proteins fold and bind is interesting since these processes are central to most biological activity. Protein folding and protein-protein interaction are by themselves very complex but using a good and robust system to study them could ease some of the hurdles. In this thesis I have tried to answer some of the fundamental questions of protein folding and binding. I chose to work with PDZ domains, which are protein domains consisting of 90-100 amino acids. They are found in more than 400 human proteins and function mostly as proteinprotein interaction units. These proteins are very stable, easy to express and purify and their folding reaction is reversible under most laboratory conditions. I have characterized the interaction of PSD-95 PDZ3 domain with its putative ligand under different experimental conditions and found out that its binding kinetics is sensitive to salt and pH. I also demonstrated that the two conserved residues R318 and H372 in PDZ3 are responsible for the salt and pH effect, respectively, on the binding reaction. Moreover, I determined that for PSD 95 PDZ3 coupling of distal residues to peptide binding was better described by a distance relationship and there was a very weak evidence of an allosteric network. Further, I showed that another PDZ domain, SAP97 PDZ2 undergoes conformational change upon ligand binding. Also, I characterized the binding mechanism of a dimeirc ligand/PDZ1-2 tandem interaction and showed that despite its apparent complexity the binding reaction is best described by a square scheme. Additionally, I determined that for the SAP 97 PDZ/HPV E6 interaction that all three PDZ domains each bind one molecule of the E6 protein and that a set of residues in the PDZ2 of SAP 97 could operate in an unexpected long-range manner during E6 interaction. Finally, I showed that perhaps all members in the PDZ family could fold via a three state folding mechanism. I characterized the folding mechanism of five different PDZ domains having similar overall fold but different primary structure and the results indicate that all five fold via an intermediate with two transition states. Transition state one is rate limiting at low denaturant concentration and vice versa for transition state two. Comparing and characterizing the structures of the transition states of two PDZ domains using phi value analysis indicated that their early transition states are less similar as compared to their late transition states. Keywords: protein-protein interaction, protein folding, and drug design Celestine Chi, Department of Medical Biochemistry and Microbiology, Box 582, Uppsala University, SE-75123 Uppsala, Sweden © Celestine Chi 2010 ISSN 1651-6206 ISBN 978-91-554-7836-0 urn:nbn:se:uu:diva-126129 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-126129).

(17) To my mother.

(18)

(19) List of Papers. This thesis is based on the following papers, which are referred to in the text by their Roman numerals. I. Chi, C. N. , Engstrom, A., Gianni, S., Larsson, M., and Jemth, P. (2006) Two conserved residues govern the salt and pH dependences of the binding reaction of a PDZ domain. J. Biol. Chem. 281, 36811-36818. II. Chi, C. N., Elfstrom, L., Shi, Y., Snall, T., Engstrom, A., and Jemth, P. (2008) Reassessing a sparse energetic network within a single protein domain. Proc. Natl. Acad. Sci. USA. 105, 4679-4684. III. Chi C. N., Bach A., Engstrom A., Wang H.,Stromgaard K., Gianni S., Jemth P. (2009) A sequential binding mechanism in a PDZ domain. Biochemistry. 48, 7089-7097. IV. Chi, C. N., Gianni, S., Calosci, N., Travaglini-Allocatelli, C., Engstrom, A., and Jemth, P. (2007) A conserved folding mechanism for PDZ domains. FEBS Lett. 581, 1109-1113. V. *Calosci N., Chi C. N., Richter B., Camilloni C., Engstrom A., Eklund L., Travaglini-Allocatelli C., Gianni S., Vendruscolo M., Jemth P.,. (2008) Comparison of successive transition states for folding reveals alternative early folding pathways of two homologous proteins. Proc. Natl. Acad. Sci. USA. 105, 19240-19245. VI. Bach A., Chi C. N., Pang F. G., Olsen L., Kristensen S. A., Jemth P., and Stromgaard K.,. (2009) Design and Synthesis of Highly Potent and Plasma-Stable Dimeric Inhibitors of the PSD-95/NMDA Receptor Interaction. Angewandte Chemie, DOI: 10.1002/anie.200904741.

(20) VII. Chi C. N., Bach A., Gottschalk M., Kristensen S. A., Strømgaard K. and Jemth P. (2010) Deciphering the kinetic binding mechanism of dimeric ligands, using a potent plasma-stable dimeric inhibitor of postsynaptic density protein-95 as an example. Submitted. VIII. Chi C. N., Bach A., Engstrom, A., Strømgaard K, Lundström P., Ferguson N., and Jemth P. Conformational change and noncanonical interactions in the complex between human papillomavirus E6 protein and Synapse-Associated Protein 97. Submitted. * Contributed equally to the work Reprints were made with permission from the respective publishers..

(21) Contents. Introduction..................................................................................................... 9 Protein-protein interaction and protein folding.......................................... 9 Background................................................................................................... 13 Binding mechanisms ................................................................................ 13 Theory ................................................................................................. 13 Folding and stability ............................................................................ 16 Theory ................................................................................................. 17 Present work ................................................................................................. 20 Aims ......................................................................................................... 20 Results and Discussion ................................................................................. 21 Conclusion ............................................................................................... 30 Future Perspectives....................................................................................... 31 Populärvetenskaplig sammanfattning ........................................................... 33 Acknowledgements....................................................................................... 34 References..................................................................................................... 37.

(22) Abbreviations. Dlg GABA GK GRIP. Disc large -aminobutyric acid guanylate kinase glutamate receptor. interacting. protein HPV. human papilloma virus. MAGUK. membrane associated guanylate kinase like N-methyl-D-aspartate Post-synaptic density-95 Disc large Zo-1 protein interacting with C-kinase post-synaptic density src homology-3 synapse associated protein zonula occludens-1. NMDA PDZ PICK PSD SH3 SAP Zo-1.

(23) Introduction. Information transfer in biological systems is a fundamental aspect of life that is central to all biological processes and, thus, for the proper functioning of an organism. The source of information and the point of execution are often located at different parts in living organisms. Therefore, biological systems have had to derive a means of transfering this information from the site of input to the site of action. This information transfer process in living systems is known as signaling and a lot of different specialized molecules have evolved to carry out this process. The most common biological signaling molecules are proteins. Signaling through proteins often occurs via proteinprotein contacts. Proteins are dynamic molecules that constantly adopt different conformations. The conformation a protein will adopt at one point will depend on several factors, for example its contact partner (other biological molecules) or conditions such as pH, salt concentration and temperature. Some proteins have an average conformation that is regarded as stable and will not change much even in the presence of its interacting partner. For some there is no fixed conformation (intrinsically unstructured) and they only adopt a well define form in the presence of their interacting partners. In the process of adopting this stable conformation (folding) the amino acids are rearranged in the right orientation and new contacts are formed and broken. Understanding how proteins bind and/or fold can be important for targeting them during disease.. Protein-protein interaction and protein folding Protein-protein interaction is the process whereby a protein comes into contact (directly or indirectly) with another protein for a particular function. It could be for structural reasons, such as maintenance of cell rigidity, cell motion or the signaling of information between neighboring molecules in order to carry out an important biological process such as replication or repair. Central to performing these biological functions proteins have to be able to recognize their interacting partners. Groups of proteins that recognize similar binding partners (such as proteases) can share general properties such as protein fold type or have similar. 9.

(24) amino acid groups at the binding site. A protein sequence and structure that can evolve, function and exist as an independent unit is called a domain. Post-synaptic density Disc large Zo-1 (PDZ) domains are the most common protein-protein interaction units in eukaryotes, for example they are found in more than 300 proteins in the mouse (1). One or more PDZ domains may be present within a single protein molecule. They function mostly as proteinprotein interaction units and hold receptors and signaling molecules together within large protein complexes. PDZ domains are organized differently in various proteins. Some consist only of PDZ domains, such as syntenin, or of PDZ domains together with either a Src Homology 3 (SH3) domain, a guanylase kinase like domain or a Lin2 domain, as in MAGUK proteins (for example PSD 95 and SAP 97). PDZ domains are also found together with other protein domains such as L27, LIM, WW etcetera (2). PDZ domains often recognize the C-termini of their targets. Some also recognize internal peptide sequences (3-5). Sequence alignment of different PDZ domains demonstrates that they share less than ~40% similarity (Figure 1). Yet they fold into the same globular structure that consists of two alpha helices and four to six beta strands (Figure 2). PDZ domains are broadly divided into classes but strictly speaking are evenly distributed throughout the selectivity space (1). Isolated PDZ domains are usually very stable and their folding/unfolding reactions are reversible under most experimental conditions, enabling the use of PDZ domains for the study of the molecular principles governing protein folding and binding. In this thesis, I will explore different PDZ domains from a variety of proteins and compare their binding/folding mechanisms. I will start by providing a background on protein binding and folding. 10.

(25) Figure 1 Sequence alignment of nine PDZ domains included in this study. Regions representing the helices and strands are shown. A star denotes identical or similar regions * (40% similarity within this subset of PDZ domains).. Figure 2. Structures of PDZ domains from PSD 95. Left panel PDZ2 PDB code: 2OQS(6). Structural elements are depicted in different colors, red: helices, yellow: strands, green: loops, blue: peptide ligand. Right panel PDZ3 PDB code: 1BE9 (7). The peptide ligand is shown in red and numbered p-3, p-2, p-1 and p0 from the carboxylate terminus.. 11.

(26) Table 1. Computationally predicted allosteric networks in PDZ domains. Colour codes represent the network residues; Red and green for PDZ3 (8,9), blue, purple, orange and brown for PTP-BL PDZ2 (10-12). The last column on the right in red represents all the five networks of residues merged as if they were from a single study. PDZ3 PSD95. PTP PDZ2 (ID5G). Residue numbeResidue type REGION. Study 1. Study 2. PTP PDZ2 (ID5G) residue type. region Study 3. Study 3. Study 3. Study 4. Study 5 Study 5. 300 301 PHE 302 LEU 303 GLY 304 GLU 305 GLU 306 ASP 307 ILE 308 PRO 309 ARG 310 GLU. E310. 311 PRO 312 ARG. S. 313 ARG. S. 314 ILE. S. 315 VAL. S. 316 ILE. S. 317 HIS. S. 318 ARG 319 GLY. R318. 320 SER. S319. 321 THR. T320. 322 GLY. G322A (G26A). 323 LEU. L323. 324 GLY. S. 325 PHE. S. 326 ASN. S. 327 ILE. S. 328 ILE. S. 329 GLY. S. G329A (G33 G329. 330 GLY. S. G330A (G34A). F325K (F29K) I327. 331 GLU. E331. 332 ASP. D332. 333 GLY. G333. 334 GLU. E334. 335 GLY. S. 336 ILE. S. 337 PHE. S. 338 ILE. S. 339 SER. S. 340 PHE. S. 341 ILE. S. 342 LEU. S. G335E (G39E). 343 ALA 344 GLY 345 GLY 346 PRO. H. 347 ALA. H. 348 ASP. H. 349 LEU. H. 350 SER 351 GLY 352 GLU 353 LEU 354 ARG 355 LYS 356 GLY. S. 357 ASP. S. 358 GLN. S. 359 ILE. S. 360 LEU. S. 361 SER. S. 362 VAL. S. V362I (V66I). 363 ASN. S. N363A (N67A). 364 GLY 365 VAL. S. 366 ASP. S. 367 LEU. S. 368 ARG 369 ASN 370 ALA 371 SER 372 HIS. H. 373 GLU. H. 374 GLN. H. 375 ALA. H. 376 ALA. H. 377 ILE. H. 378 ALA. H. 379 LEU. H. 380 LYS. H. His372. H372. A376V (A80VA376. K380A (K84AK380. 381 ASN 382 ALA 383 GLY 384 GLN. S. 385 THR. S. 386 VAL. S. 387 THR. S. 388 ILE. S. 389 ILE. S. 390 ALA. S. 391 GLN. S. 392 TYR. S. 393 LYS. S. 394 PRO. H. 395 GLU. H. 396 GLU. H. 397 TYR. H. 398 SER. H. 399 ARG. H. Q384 V386I (V90I). 400 PHE 401 GLU 402 ALA. 12. F400. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27. k P G D I F E V E L A K N D N S L G I S V T G G V N. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90. T S V R H G G I Y V K A V I P Q G A A E S D G R I H K G D R V L A V N G V S L E G A T H K Q A V E T L R N T G Q V V H L L L E. 91 92 93 94 95 96. K G Q S P T. 3. S S S S S S. 12 13 14 15 16 17. 17. 18. 18 20. 18. 20. S S. 22. S. 23. 23. 24 25 26. 26 27. 28 29. 29 30. 30. 33 34. 34 35. S. 37. S S. 39. S. 41. S S S. 43. 43. 44 45. 45. 46 H. 47. H. 48. 46. 46. 52. 52. H H H H. 52 53 54. S. 58. S. 59. 58. 58 59. S S. 61. 61. S. 62. S S. 64. S. 66. S. 67. 66. 66. 68 69. 69 71. H H. 74. H. 75. H H H. 78. 78. 81. 81. 78. H H H. 82 S S S. 85. 85. S. 86. S. 87. S S S S S. 89 90. 87. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34. K P G D T F E V E L A K T D G S L G I S V T G G V N. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97. T S V R H G G I Y V K A I I P K G A A E S D G R I H K G D R V L A V N G V S L E G A T H K Q A V E T L R N T G Q V V H L L L E. 98 99 100 101 102. K G Q V P. Merged.

(27) Background. Binding mechanisms One of the objectives of a kinetician is to determine the sequence of events in molecular interactions and unravel important steps in the reaction pathway. The results of such investigation could be beneficial in drug design and drug discovery. For example, a drug molecule could be designed to bind only a particular conformation of its desired target, thus shifting equilibrium towards this form (for example in Alzheimers disease, drugs can be designed that target only the native conformations of amyloid beta and prevent it from forming fibrils).. Theory Consider the following bimolecular interaction schemes; let k1 be the forward rate constant and k-1 the reverse rate constant for the initial binding in Scheme I. Scheme I Scheme II Scheme III. A + B <=> AB A + B <=> AB <=> AB* A + B <=> B* + A <=> AB*. The formation of product AB in scheme I with time is given as in (Eq.1). and can be re-written as was determined elsewhere (13) AB = ABEQ (1 - e-kobs t )/(1 + e-kobs t ). (Eq.1). where AB is the amount or concentration of product formed at time t, ABEQ is the total product formed at equilibrium, is the parameter describing the deviation of the experimental setup from pseudo-first order conditions and varies between -1 and 1. kobs is the observed rate constant and depends on the reversible association of A and B If the concentration of A is in excess, such that [A] >> [B] (Scheme I; pseudo-first order conditions), then the observed rate constant for the molecular interaction is given in (Eq.2) kobs = k1[A] + k-1. (Eq.2) 13.

(28) Under second order conditions where [A] is similar to [B], the dependence of the observed rate constant on the forward and reverse rates is now different (Eq.3). kobs = (k12([A]0-[B]0) + k-12 + 2k1k-1 ([A]0-[B]0))1/2. (Eq.3). k1 is the on-rate constant, k-1 is the off-rate constant, and [A]0 and [B]0 are the initial concentrations of A and B, respectively. If, as in scheme II, AB forms a product AB* we expect at least two relaxation times (kobs1 and kobs2), since there are two steps involved. Also, if there is a rapid equilibrium followed by a slower conformational change i.e. k-1 >> k2 as in scheme II, then under pseudo-first order conditions kobs1 = k1[A] + k-1. (Eq.5). kobs2 = k-2 + k2 [A]/(K + [A]). (Eq.6). K is the equilibrium constant for the fast phase. However, if there is slower conformational change followed by a rapid binding event, as in scheme III, then kobs2 = k1 + k-1 K/(K + [A]). (Eq.7). kobs1 = k-2 + k2 [A]. (Eq.8). In most experimental setups, as with the stopped-flow machine, only one relaxation time is often measured and this jeopardizes interpretation. Distinguishing scheme II from III is not always easy but measurements at higher concentrations of the varied species and with faster techniques can provide a clue. As depicted by (Eq.6), kobs2 in scheme II will increase with increasing [A], whilst, kobs2 will decrease with increasing [A] in scheme III (Eq.7). Dissecting the molecular mechanism of interactions between PDZ domains and their binding partners can be challenging, one of the reasons being that PDZ domains are often found in tandem or in arrays and sometimes interact with identical ligands or have overlapping specificities. Another situation, as depicted by schemes II and III, arises when there is a rate limiting conformational change in the binding reaction (14,15). The situation becomes even more complex in living cells wherein other domains may influence the binding mechanisms through, for example, allostery and competitive binding. The strength or affinity (KD) with which PDZ domains interact with Cterminal peptides has been determined from in solution assays to be in the. 14.

(29) micromolar range (1-50M) (16) using both fluorimetric (10,17-21) and microcalorimetric (22,23) techniques. It has been suggested that the micromolar affinity of PDZ-ligand interactions is reasonable since PDZ domains are mostly found in regions that require rapid adaptation to external stimuli and should therefore respond within seconds (16). The binding kinetics at 10 °C and pH 7.5 of two members of the PDZ family, PTPBL PDZ2 and PDZ3 PSD95, binding to C-terminal peptides have been well characterized (14,18). At the time of initiation of this thesis, relatively little was known about the interaction between PDZ domains and a whole domain ligand, for example the E6 C-terminal domain (E6-C); this question is addressed in the present study. Structural data indicate that PDZ domains maintain degenerate specificity and that the ligand interaction depends on the actual ligand chemical structure. For example, when PDZ domains interact with the putative class I, II or III motifs (as previously classified), the peptide binding takes place in an elongated groove between the 2 helix and 2 strand, forming an additional -strand (7,24,25) (Figure 2). For class I motif the hydroxyl of Thr or Ser at peptide position2 forms a hydrogen bond to the N-3 nitrogen of His at 2 and the carboxylate of the terminal hydrophobic residue is coordinated via a water molecule to an Arg or a Lys residue. The class II domains form an extra hydrophobic-binding pocket composed of residues from the 2 strand and 2 helix and a hydrophobic residue at peptide position2 of the peptide ligand interacts with a hydrophobic residue in the 2 helix (26,27). For class III a tyrosine in the 2 position forms a hydrogen bond to an aspartic/glutamic acid at peptide position2 (24). This mode of recognition that is maintained by C-terminal peptides is canonical, with the most important positions being position 0 and 2. (see Figure 2). In a few PDZ domains, such as dishevelled PDZ and nNOS PDZ, an extended conformation known as a -finger interacts with the corresponding PDZ in a canonical fashion (24,28). Other PDZ’s interact with internal sequences in a non-canonical mode. For example, the PDZ domain of the cell polarity protein Par-6 can bypass the canonical mode by binding to Pals-1 (lin7) (25). In the bound conformation, the ligand pocket of Par-6 PDZ becomes deformed due to an additional interaction beyond the p0 position. This prevents Par-6 PDZ from undergoing affinity activation by the CRIB domain as is the case for Cterminal peptides (25,29). Another non-canonical mode of interaction is exhibited by PDZ1 of syntrophin when it binds the peptide TNEFYF and the peptide binds nearly perpendicular to that expected for the canonical binding mode, with only the p0 and p-1 position making contacts to the PDZ (30). PDZ can also form homo or hetero-oligomers. For example, PDZ1 and PDZ2 of ZO proteins can oligomerize in a new interaction mode that is known as domain swapping (31). This likely increases the ligand binding. 15.

(30) affinity since there is a greater ligand-binding surface created by the oligomers compared to the monomer. PDZ domains often occur concatenated and/or together with other domains. A question one will pose is what is the structural basis for this arrangement? PDZ domains have been shown to have overlapping specificities (17,32) and bind similar ligands with different/similar affinities yet are placed side by side with each other in the same protein molecule. Several studies have shown that this type of arrangement is optimal for PDZ-ligand interaction as exemplified by the PDZ1-2 of syntenin-syndecan interaction where PDZ2 acts as the high affinity binder and strongly requires the presence of the PDZ1 in tandem for its interaction (33). In the case of PDZ1-2 of GRIP1, PDZ1 binds Fras1 but requires PDZ2 in tandem for its stability (34). All three PDZ1-3 of SAP97 and PDZ1-3 of PSD95 bind to the cytoplasmic tail of Kir2.1 inwardly rectifying potassium channels and control their localization and expression (35,36). The linker region between PDZ2 and PDZ3 was recently shown to be highly flexible and the PDZ2 and PDZ3 domains found to make interdomain contacts and may be relevant for intradomain regulation within PDZ2 and PDZ3 (35).. Folding and stability In studying protein folding the chemist aims at characterizing the events that results in the three dimensional structure of the protein, starting from its amino acid sequence. Each amino acid in an unfolded polypeptide can theoretically make millions of interaction en route to the final protein fold. This poses a serious problem since it will mean that the time for a protein to fold from its unfolded polypeptide will be extremely long. However, we know that this is not the case (37). Several different concepts have been put forward in an attempt to solve this enigma behind protein folding. 1) The framework model, which postulated that proteins fold in a stepwise manner via a series of partially formed intermediates. This model has been challenged since it is known that small proteins could fold without detectable intermediates (38). 2) The nucleation condensation mechanism that claims the protein could fold via simultaneous formation of secondary and tertiary structure (39). 3) The hydrophobic collapse hypothesized that a protein will collapse rapidly around its hydrophobic side chains and then rearrange from limited conformations to form an intermediate, which will then propagate to the native structure (40). Finally, the statistical thermodynamic approach with the folding funnel (Figure 3) (41), which considers that a typical polypeptide chosen at random has a “rugged” energy landscape. The protein is “steered” towards the native conformation from a restricted conformational space. An outcome of this theory is that, many folding routes are possible and only those with minimal energy will proceed to the native state. It should be born in mind that these folding mechanisms are not discrete but also overlapping in some cases. 16.

(31) Width of the landscape. Unfolded polypeptide. Increase in native character. Intermediate state Free energy. Folded polypeptide. Native state. Figure 3 The energy landscape theory of protein folding. The width of the funnel indicates the possible states of an unfolded polypeptide. This figure was redrawn based on the picture in (42).. Theory Protein folding is often studied by unfolding a protein using chemical denaturants, temperature, or pressure and then refolding its back to it native conformation. Most small proteins (50-150 amino acids) fold in an apparent two-state manner and their kinetics can be described by a simple D N, where D denotes the denatured state and N is the native state. If ku represents the unfolding rate constant and kf the folding rate constant then the observed rate constant kobs is given as ku + kf. It is known that at [denaturant] > D50% for many small proteins ln ku is linear with [denaturant] (Eq.10) (Figure 4). D50% represents the concentration of the denaturant at which 50% of the protein is in the denatured state. ln ku = ln kuH2O + mku [denaturant]. (Eq.10). where kuH2O is the folding rate constant in water and mku is a constant of proportionality. A similar relationship for kf at [denaturant] < D50% is given in (Eq.11) (Figure 4) ln kf = ln kfH2O - mkf [denaturant]. (Eq.11). It follows that mkf + mku = mT: The m-value is related to the change in accessible surface area upon denaturation (43). ln kobs = ln kfH2O -mkf [denaturant] + ln kuH2O + mku[denaturant] (Eq.12). 17.

(32) A plot of this form gives a V-shaped curve, which is sometimes called a chevron plot (43). For a two state behaviour the change in free energy is as given in (Eq.13) GD-N = GD-NH2O - mD-N [denaturant]. (Eq.13). where GD-NH2O is the value in water and mD-N is the slope and measures the change in accessible surface area on denaturation (44).. Figure 4 Variation of observed rate constants with denaturant concentration for a two-state folding protein (left panel) and a three state or broad transition state (right panel.. The folding kinetics of some small proteins however, do not proceed as Eq. 12 and instead a plot of the observed rate constant with [denaturant] follows either Eq. 14 or 15 (Fig.4) (43). ln ku = ln kuH2O + mku [denaturant] x 1/(1+ ln Kp (Eq.14). + mp [denaturant]). ln kf = ln kfH2O - mkf [denaturant] x 1/(1+ ln Kp (Eq.15). + mp [denaturant]). In Eq.14 Kp is the equilibrium (or partitioning) constant between the ratelimiting transition states and mp is the associated m-value. In (Eq.15) Kp is the equilibrium constant between the denatured and intermediate populations and mp is the associated m-value This means that there is a change in rate limiting step in the folding kinetics at high [denaturant] or low [denaturant]. This is manifested as a roll over in the chevron plot (Figure 4). 18.

(33) Figure 5 Free energy diagram for a three state folder (e.g PDZ3) at low denaturant concentration(16). TS1 is rate determining at lower denaturant concentration. At least two main hypotheses have been used to explain the roll over effect. 1) The broad barrier, which states that different regions of a broad transition barrier becomes rate-limiting at different denaturant concentrations in the same manner as the Hammond theory (45-47). 2) The sequential barrier, postulates a switch between two or more transition states (48,49) (Figure 5). The two theories can explain well the deviation from linearity in the plots of kobs with [denaturant] but have different implications on the nature of the energy landscape (50). A recent analysis based on the dependence of the T (Tanford -value) to denaturant concentration is used to distinguish the two mechanisms. T was constructed over a wide range of denaturant concentrations for two proteins displaying curved chevron plots. It was observed that the change T on denaturant concentration was constant for the broad barrier and sigmoidal or bell shaped for the two transition state, and turns to zero at situations where the protein is infinitely stable and unstable (51).. 19.

(34) Present work. PDZ domains are one of the most studied (in terms of binding partners) and well-characterized domains of the MAGUK family. Their binding partners have been well described and over 217 X-ray and solution structures are available in the PDB database. Understanding the binding and folding mechanisms of PDZ domains could be beneficial for a number of studies, especially in computer simulations and in drug design.. Aims 1) To investigate the binding mechanisms of PDZ domains (Paper I, II & III) 2) To investigate the folding mechanism of PDZ domains. Is folding topology dependent and/or determined by sequence? (Paper IV & V) 3) Targeting PDZ domains in disease. Are PDZ domains druggable? (VI, VII & VIII). 20.

(35) Results and Discussion. In the study of protein-protein and protein folding interactions, the kinetician intends to find the simplest routes and pathways towards the final protein complex and/or native state. The study of protein folding in general terms is an example of a curiosity driven science. This notwithstanding, the knowledge gained from such studies could serve as important guidelines in a number of computer assisted predictions of protein structures and functions. Similar basic physico-chemical rules are applied in the study of both proteinprotein interaction and protein folding; in protein-protein interactions the solvent contributes immensely to the general interaction process. In protein folding it is the interaction between the protein and the solvent that is the point of focus. A useful and instructive approach towards understanding the principles governing protein-protein interaction and protein folding is to study the binding and folding of a family of proteins under various experimental conditions (16,52,53), because differences and similarities in the binding and folding pathways can be appreciated at the levels of both sequence and topology. In this thesis, I studied the binding and folding mechanisms of different members of the PDZ domain family. As mentioned in the introduction PDZ domains generally bind to a short stretch of the C-terminus of proteins and this “peptide binding” takes place between the 2 helix and 2 strand of the PDZ. This is sometimes called augmentation and the peptide makes an extra anti-parallel strand (Figure 1) (3-5). Both structural and mutagenesis studies have confirmed that only a few residues in both the binding pocket of the PDZ domain and in the peptide ligand are important for binding (1,7,23,54). Under the orthodox classification, PDZ domains were classified into three main classes based on the peptide ligands that they recognize. For example, the class one PDZ domains recognize peptides of the form -X-T/S-X-Vcoo- (numbered -3, -2, -1, 0), where X refers to any amino acid. However, a more complete scheme of PDZ recognition modes in a genome-wide scale of the mouse proteome shows that PDZ domains do not fall into discrete classes; instead they are evenly distributed in the selectivity space (1). It is puzzling how only a few residues in the binding site of the PDZ domain could determine the specificity and affinity of PDZ-peptide ligand interactions. Structural data of the complex between the third PDZ domain of PSD 21.

(36) 95 with its peptide show that the C-terminal carboxylate at the 0 position of the peptide is coordinated via a water molecule to residue R318 and the T residue at -2 position of the peptide forms a hydrogen bond with H372 in the 2 helix (7). Kinetic ligand binding experiments demonstrated that salt and pH could influence the binding reaction of the third PDZ domain of PSD 95 and the CRIPT peptide, but the basis of this was not clear (18). Therefore, we explored the binding reaction further by mutagenesis and found that two conserved residues (H372 and R318) were responsible for the pH and salt dependence in the PDZ-ligand interaction (Paper I). We evaluated the effect of different anions on the on-rate constant and found that chloride competes out the peptide ligand. Mutation of R318 to Ala abolished the effect of chloride on the on-rate constant. Chloride binds specifically to R318 and competes with the peptide ligand in a concentration dependent manner. We found that this effect is not long-range, as was previously suggested (18). The effect of chloride on the on-rate constant was confirmed by timeresolved urea denaturation experiments that found that the Arg to Ala mutant was less stabilized by chloride than the wild type (Figure 6A and B). We also showed that the increased off-rate constant of dissociation at low pH was due to protonation of H372 as mutation to Ala abolished this dependence on pH (Figure 6C and D). It appears that all PDZ domains of this type might respond in a similar mechanism to salt and pH. However, mutation of Lys-86 to Met in an analogous PDZ, 1syntrophin PDZ, resulted in a loss of affinity but did not abolish the effect of NaCl on the equilibrium dissociation constant (55). The basis of salt dependence on the PDZ-peptide interaction made by the 1syntrophin PDZ domain is thus distinct. Having established the residues responsible for the specificity and affinity in the third PDZ domain of PSD 95, we now wondered whether or not allosteric regulation and cooperativity are operational in PDZ domains. Indeed all proteins are dynamic in the picosecond-to-microsecond time scale, as shown by NMR experiments (56) and sometimes this property is important for function. An evolutionarily conserved network of residues that could operate in unexpected long-range interactions was identified in the PDZ domain family of proteins (8). In fact several different network of residues have been identified in a number of PDZ domains (9-12). The implication of this is that in most PDZ domains energy can propagate from the ligand binding site to another distal site through a set of residues i.e. allostery. However, structural and kinetic data indicate that the binding mechanism could range from a simple lock and key to a sequential binding mechanism in some members of the PDZ domain family (18,57). Furthermore, sequence alignment of the five different networks identified (mentioned above) by various methods shows poor correlation amongst them (Table 1). A pretty straightforward procedure to determine if two positions in a protein cooperate energetically to ligand binding is to perform what is known as double mutant cycle analysis (43). We applied this technique to the most studied PDZ domain, PDZ3 of PSD 22.

(37) 95 in order to see if indeed the computationally identified networks could be confirmed with experiments and if these residues are energetically linked (Paper II). However, we found that coupling was better described by a distance relationship (Figure 6E). These results together with the analysis in Table 1 argue against an evolutionary conserved network of residues. Thus, I will suggest that each PDZ domain has its own allosteric network, which are almost independent of each other. A. 10. B. R318A no salt. 100. R318A no salt. WT no salt. WT no salt. R318A 0.6 M KCl. R318A 0.6M KCl WT 0.6 M KCl. kobs (s-1). kobs (s-1). WT 0.6 M KCl. 1 6.5. 7. 10. 7.5. 8. 1.6. 1.8. 2. urea (M). [Urea] (M). C. D. Fluorescence. Fluorescence 0.6. 0.6. 0.5. 0.5. 0.4. pH=5.7 pH=8.6. 0.4. pH=5.7 pH=8.6. 0.3. 0.3. 0.2. 0.2 0. 50. 100. 150. E. 0. 200. 400. 600. 800. 1000. [H372A] (M). [WT] (M). IGCI (kcal/mol). 2. 1.5. 1. 0.5. 0 0. 5. 10. 15. 20. 25. Distance (Å). 23.

(38) Figure 6 Dependences of PDZ3-peptide interactions. A) and B) dependence of the folding B) and unfolding A) rate constants for PDZ3 wildtype and R318A mutant on chloride concentration. C) and D) pH dependence of the equilibrium constant for PDZ3-peptide interaction, C) wildtype-peptide and D) H372A-peptide interaction(14). E) Binding coupling energies of PDZ3 as a function of distance (58). The two main classical binding mechanisms in proteins that undergo structural changes upon binding are the Monod-Wyman-Changeaux concerted mechanism (59) and the Koshland-Nemethy-Filmer induced-fit or sequential mechanism (60). These are indeed extremes of the very complex process of binding and it is true that both mechanisms can operate in the same system (61). Recent analysis using various methods has uncovered that most if not all proteins exist in multiple conformations and that only one of these forms is competent in binding the ligand (62-65). This is indeed a new formalism of an old concept (Monod-Wyman-Changeaux model), which is referred to as conformational sampling, population shift, selected fit etcetera and has been considered a general mechanism of allostery. To see which of the binding modes is operating in PDZ domains we decided to perform ligand binding on SAP 97 PDZ2. We used three different mutant proteins and performed ligand binding with three different peptides (Paper III). We showed that the initial binding step is followed by a fast conformational change with a forward rate constant of 5000 s-1 (Figure 7). These results do not rule out the presence of a faster initial event (e.g. fast sampling of various conformations) but simply indicates that for some PDZ domains the preferred path is induced-fit. Based on these results and previous results from PTP-BL PDZ 2 (57), it is tempting to speculate that for most PDZ domains the preferred mechanism of interaction is the induced-fit.. 24.

(39) Figure 7 Free energy diagram and reaction schemes for SAP97 PDZ2-peptide binding (15). The energy diagram depicts the case where k-1 >> k2 and is the likely scenario for SAP 97 PDZ2-peptide interaction. PDZ domains fold into a compact globular structure that is made up of two alpha helices and 4-6 beta strands. Pair-wise sequence comparison and overall sequence comparison between several members of the PDZ domain family show very little identity amongst them (Figure 1)(53). It is very striking how such a low identity within the family, which is not very different from random sequence comparison, could give rise to similar folds with similar biological properties. One strategy to understand this is to study the folding reaction of several members of a given family of proteins and, as such, general correlations between amino acid sequences and folding pathways could be extrapolated. This principle has been applied in a number of studies (39,52,66-69) and they reveal that the overall folding mechanism is conserved within a fold, and hidden common features can be unveiled (70). Studies on the folding kinetics of PTPBL PDZ2 (18) and PSD95 PDZ3 (71) indicate that PDZ domains fold in a sequential manner via an intermediate. Detailed analysis reveals that the intervening folding intermediate is of high energy and is obligatory for PDZ domains (72). To see if this is an inherent property within the PDZ domain family, we studied the folding kinetics of five members of the PDZ domain family under various conditions (Paper IV). The results showed that the folding kinetics of all five PDZ displayed a characteristic three state behavior and could be described by a model with two transition states separated by an intermediate. TS1 was rate limiting at low denaturant while TS2 was rate limiting at high denaturant (Figure 5). In addition, the T values indicated that the position of the transition states 25.

(40) among members is conserved. Studies on PSD95 PDZ3 using hydrogen exchange shows that the intervening intermediate is of low energy and resembles the native state (71). Importantly, the PDZ domain used by Bai and coworkers (71) contained an extra C terminus not present in our constructs and the intermediate detected included the non-native C-terminus and is distinct from the one discussed here. Therefore, based on the following, 1) analysis of the unfolding/refolding amplitudes (which show no indication of any additional ultra-fast burst phase undetectable for the stopped flow apparatus) and 2) binding induced folding experiments (72), we concluded that the folding mechanism of PDZ domains is explained by a three state model, where the intermediate is of high energy and obligatory. The folding process is often described in terms of the energy landscape of protein folding, where an unfolded polypeptide is considered to occupy an huge number of trajectories (Figure 3) (41,73-75). The folding process starts with the descent towards the final native fold (in the funnel theory of protein folding). This process is not smooth and, as described elsewhere (72) both obligatory and non-obligatory intermediates are encountered during the process (76). Separating these intermediates are transition states and, as shown for PDZ domains, two main transition states are visible one at low and one at high denaturant concentrations, making PDZ domains three state folders. Since the transitions states (specifically for the PDZ domains studied here) lie on the pathway that leads to the native fold, the mapping of the structures of the transitions states could reveal details that are important in the folding process (77). In order to gain insights into the structures of the late and early transitions states of PDZ domains, we determined phi-values () and used them as restraints in molecular dynamics simulations on two members of PDZ domain family of proteins, PTP BL PDZ2 (77) and PSD 95 PDZ3 (Paper V). We found a good agreement between the -values of the late transition states (correlation factor of 0.72) and poor for the early transition states (correlation factor of 0.5) for the two PDZ domains (Figure 8). This result indicates that while the native fold guides the late folding stages there is much freedom in the early folding process. Structural calculations with molecular dynamics showed that the late transition states were very similar, while the early transition states were less similar. In terms of the energy landscape of protein folding, the kinetic paths occupied by the two PDZ domains during folding seem to fall on separate trajectories and only come together at a very late stage en route to the bottom of the funnel.. 26.

(41) Figure 8 Chevron plots and phi-phi plots for the folding of PDZ domains. All five chevrons (left) reveals changes in rate limiting step (53). Phi-phi plots for the early transition states (middle) and the late transition states (right) for PTP-BL PDZ2 and PSD 95 PDZ3 (78). PDZ domains interact directly with glutamate receptors (GluRs) and control their localization and signaling along the post-synaptic density. This process has a very important physiological function in learning, memory (long term potentiation, a form of synaptic strengthening following brief, high frequency, stimulation) and development (79). Glutamate or other excitatory amino acids that potentiate plasticity, become toxic at high concentrations (80). The over/under stimulation of GluRs have been implicated in neuronal degenerative diseases (2,81,82), including Schizophrenia and Alzheimer’s, wherein the PDZ proteins PICK1 and XII interact with C-kinase 1 and amyloid precursor protein respectively (83). PDZ have also be implicated in Huntington’s disease and Parkinson’s disease, wherein CASK/lin2 PDZ interacts with perkin (84). It should be noted that PDZ domains are implicated only indirectly in these disease states via their interaction with glutamate receptors. Because of their involvement in several pathological situations they are seen as possible therapeutic targets. However, this is not generally encouraging as most of the drug-like molecules developed are peptides with poor pharmaco-dynamic properties. An attractive approach is peptidomimetics (85). In this strategy, the starting point is a known peptide ligand that is slowly “chopped off” and into which is simultaneously introduced a non-peptide moiety. It is often hard to maintain both the affinity and specificity of the peptide in such an approach. However, certain aspects of the system understudy can be use to overcome this, eg PDZ domains often occur concatenated and an appealing strategy is to design di- or tri substituted inhibitors. Further, the use of a suitable linker (in this case polyethylene glycol) could impose favorable pharmacokinetic properties (86). Based on this, we designed a high affinity dimeric plasma stable peptapeptide against a PDZ1-2 tandem of PSD 95 (Paper VI) and determined the kinetics and, more importantly the contributions of the binding rate constant to the binding free energy (Paper VII). The binding kinetics of a dimeric ligand binding to two or more binding sites is complex and it is difficult to predict the overall affinity enhancement resulting from linking two ligands together 27.

(42) (87-89). We determined that for a dimeric ligand binding to two binding sites, the overall kinetic scheme describing the process is a square. There is an initial binding event for the dimeric ligand to either of the two PDZ domains in the PDZ1-2 tandem (Figure 9). The subsequent intramolecular binding step of the second peptide then increases the affinity of the interaction by almost two orders of magnitude.. Figure 9 A kinetic scheme for the binding reaction of a dimeric ligand binding to PDZ1-2 tandem. The peptide is depicted as A-A and binding can proceed in either sides of the square (15). One of the commonest direct involvements of PDZ domains in disease is in cervical cancer, which is caused by the Human Papilloma Virus (HPV) (90). HPV is classically subdivided into two main groups, “high” and “low” risk types, based on their prevalence ratio in cervical cancers (83,91). The two early gene products E6 and E7 of the high-risk types have been termed oncogenes since they can transform and immortalize cells in vitro (92,93). About 99% of cervical cancers contain DNA of the high-risk types with type 16 HPV being the most prevalent (83). The E6 protein of high-risk types binds to PDZ domains of SAP97/Dlg and other MAGUK proteins and targets them for degradation (93,94). SAP97 is a human homologue of the Drosophila disc large tumor suppressor protein Dlg and is found at tight junctions and functions in cell-cell contact and cell rigidity. The E6 proteins of high-risk HPV types interact with PDZ via the canonical class I binding motif, X-S/T- COOH, while E6 proteins of the low-risk HPV types (e.g. HPV6) do not possess this motif and do not interact with PDZ domains (94). An understanding of how the HPV E6 protein interacts with the PDZ domains holds an important remedy in the targeting of these proteins in cervical cancer that is caused by HPV. Structural and kinetic studies on the interaction between a short C-terminal E6 peptide and a single PDZ domain have been carried out (6,15,95). While this approach is very attractive because of the simplicity of the PDZ/ligand interaction, several details are left out, which is especially important if such studies are to serve as templates of drug design. A reasonable solution is to carry out binding reactions with the whole C-terminal domain of E6 proteins, which we did in Paper VIII. For the E6-SAP 97 PDZ interaction, first we determined that all PDZ domains could bind an E6 molecule independently in solution. Second, we showed 28.

(43) that that one molecule of SAP 97 PDZ region could bind three E6 molecules. Furthermore, we found that the saturated complex (SAP 97 PDZ/E6 complex) exhibited structural changes and compaction compared to free SAP 97 PDZ, which appears to involve rearrangement of the entire SAP 97 PDZ molecule upon E6 binding. Finally, we detected a few residues that exhibited unexpected long-range interactions in the second PDZ domain (Figure 11). This is very encouraging, at least from a therapeutic point of view, as those residues, which exhibited long-range effects, could serve as allosteric sites to disrupt the E6-PDZ interaction and could also act as templates for developing high affinity binders that could be further processed to create PDZ antibodies against the E6 molecule.. Figure 10 Structures of PDZ1-2 tandem of PSD 95 (PDB code 2KA9) (96) and a dimeric inhibitor. 29.

(44) 10. 9. 105. 8. 7. 6 105. G330 G336. G335 G328. 110. 110. 1 - 15N. (ppm). T383. 115. G356. E385. G325. 115. S332 F397. T351. 120. 120. N393 E380. 125. N339. H360 H341. 125. L371 I354. 130. 130 E355. 10. 9. 8. 2 - 1H. 7. 6. (ppm). Figure 11 Heteronuclear single quantum correlation spectroscopy of the SAP 97 PDZ2-E6 interaction. PDZ in complex with E6 (blue) and free (red). Arrows indicate residues that moved upon E6 interaction (Paper VIII). Conclusion I set out to understand the principles governing protein folding and binding in a broader perspective. I chose to work with a small protein domain; the PDZ domain, which has a simple binding partner and for which many threedimensional structures have been characterized. I found out that for one PDZ domain the main mode of interaction occurs via the sequential mechanism, whilst in others the lock and key is still in operation (57). Nowadays it is common to use computational algorithms with or without restraints from NMR to predict allostery in protein domains. While this is very attractive it appears that for PDZ domains allostery cannot be generalized. Interestingly, PDZ domains fold via two-transition states and a highenergy intermediate. In all PDZ domains studied, the position of the transition state appears to be conserved. Furthermore, the structures of the late transition states are similar between PDZ domains and late folding events are therefore determined by the native topology. However, the early transition states are less similar and seem to lie on different trajectories along the folding path. The development of drugs against PDZ domains and the use of PDZ domains as drugs are still in their infancy. In this work, much progress was made towards the design of inhibitors against PDZ domains by introducing bidentate ligands with much improved affinity and pharmaco-kinetic properties. Moving one step further, I utilized full protein domains as ligands for PDZ domains and in so doing identified a network of residues, which can operate in an unexpected long-range manner. This set of residues could be further exploited as allostric control sites. 30.

(45) Future Perspectives. The significance of dynamic allostery in small domains and in PDZ domains in particular has increased over the years. With the development of new methodologies to probe dynamics in single protein domains, it is tempting to speculate that even PDZ domains, which operate in the traditional lock and key fashion, could have an operational allosteric network. As discussed, the use of computer algorithms to predict such networks often ends up with more false positives. A general strategy is to make single and double amino acid side chain deletions to the whole protein and/or ligand and determine their effects on binding free energy (i.e. coupling energy). While such an approach is very attractive, it is equally expensive and time consuming. However, this will yield a more quantitative picture of the contribution of each residue in the protein. Two PDZ domains (e.g. PSD 95 PDZ3 and SAP 97 PDZ2) that exhibit completely different mechanism of interaction could be selected for such a study. Until recently, protein-folding studies have always been carried out in test tubes and in vitro settings. This is often an oversimplification of the process and it is likely that some details are missed out in the process. A more detailed description of the process would be gained from performing folding in living cells. In their recent work, Ebbinghaus et al (97) utilized a technique called fast relaxation imagining (FREI) to probe the folding of phosphoglycerate kinase in two human cell lines and concluded that cell environment influences protein stability. It will be interesting to apply such a technique to the study of PDZ domain folding and stability. In the PDZ-E6 study, a small set of residues were identified that could operate in an unexpected long-range effect (paper VIII). It will be interesting to confirm this finding by measuring the binding kinetics of mutants in which conservative mutations have been introduced into these positions. Higher affinity binders could then be selected by phage display for those residues that exhibit changes in E6 affinity. These higher affinity binders will then be optimized to produce PDZ antibodies against E6. A more elusive goal is the determination of what comes first, folding or binding? This is a vital question that stems from the fact that a great deal of proteins are secreted unstructured and only adopt their three dimensional conformation in the presence of ligands they interact with (98). The nNOS 31.

(46) PDZ interacts with PDZ2 of PSD95 through a beta finger that is unstructured in solution. Upon binding, this finger adopts a folded conformation. The question is what comes first, binding or folding. Unveiling such a molecular mechanism is a glorious goal that will shed light on questions such as does DNA sequence direct protein structure or is protein three-dimensional structure determined by its environment?. 32.

(47) Populärvetenskaplig sammanfattning. Proteiner är en av de huvudsakliga beståndsdelarna i allt levande. Proteiner är livsviktiga: så gott som alla processer i levande organismer dikteras av proteiner som interagerar med andra molekyler. Den minsta byggstenen i proteiner är aminosyran. Aminosyror sätts ihop på ett speciellt sätt så att proteinmolekylen får en funktionell tredimensionell struktur. Ibland har proteiner samma tredimensionella struktur fast deras sammansättning av aminosyror skiljer sig åt. Den process där en kedja av aminosyror bildar den tredimensionella strukturen kallas proteinveckning. Cellens funktion beror alltså i slutänden på veckade proteiner och hur de växelverkar med andra proteiner och molekyler. Följaktligen är proteinveckning och interaktioner mellan proteiner nödvändiga för nästan allt som sker i levande celler. Både ur grundforsknings- och medicinsk synvinkel är det viktigt att förstå hur proteiner veckas, att känna till deras struktur och hur de interagerar med varandra. Ibland kan till exempel en sjukdom bero på att vissa proteinproteininteraktioner inte fungerar eller på att ett protein veckas på ett felaktigt sätt. För att studera dessa processer behövs ett robust och enkelt experimentellt system, eftersom reaktionerna kan vara väldigt komplexa. I den här avhandlingen har jag använt mig av en proteinfamilj som kallas PDZ för att undersöka olika aspekter av proteiners interaktioner: (i) växelverkan mellan protein och dess ligand (“målmolekyl”), (ii) proteinveckning och (iii) oönskade protein-proteininteraktioner vid sjukdom. Jag fann att PDZ-domäner använder olika mekanismer när de växelverkar med andra proteiner. Vissa ändrar sin struktur för att bindningen ska bli starkare medan andra inte gör det. En annan upptäckt var att alla PDZ-domäner som vi undersökte, trots att de har delvis olika sekvenser av aminosyror, följer samma väg till sin tredimensionella struktur när de veckas. När jag tittade på veckningen i detalj såg jag att framförallt den senare delen av veckningsprocessen är väldigt lika för PDZ-domäner. Slutligen så upptäckte jag att ett “elakt” protein (E6 från papillomavirus som orsakar livmoderhalscancer) interagerar med ett viktigt cellulärt protein (SAP97) så att detta ändrar form och kanske rentav skyddar E6 från att upptäckas av immunsystemet. Men den goda nyheten är att jag också upptäckte några små förändringar i SAP97 som vi kanske kan utnyttja för att designa ett läkemedel mot E6-proteinet.. 33.

(48) Acknowledgements. A lot of people have been very helpful to me from the very beginning right to this point, the space available in this book will not be enough to express my sincere thanks to all of you. Which names should I call and which ones should I leave out? However, as you read through this section bear in mind that your name is mentioned in an invisible ink. I thank you very much for all your support Especially: Associate Professor Per Jemth, my supervisor for excellent supervision, introducing me into the world of science, sharing great and wonderful scientific ideas, always having time for my questions, patience, support and encouragement, in fact Per you are a better runner who beat me in all the races Maria Selmer, my co-supervisor and for always trying to be there for me Professor Ulf Lindahl, for encouraging my stay in the lab during the first year of my research Stefano Gianni, you have been there from the start to the finish, what else can I say “pronto”! Assoc. Professor Åke Engström for all the “mass specs” and the words of wisdom Patrik Lundstöm for introducing me into the world of protein NMR and for your continuous guidance Professor Hans Ronne for being such a good examiner All present and past members of the Jemth lab especially, Lisa Elfström for the wonderful time we had in the lab. Huiqun, Yao for being a wonderful. 34.

(49) student and Raza Haq soon is your turn, Maike a wonderful and great labmate, all the “fikas”, “hmmm” but just more appreciation, Aziz and Andreas, Nicoletta Calosci, a collegue who became a good friend indeed. All my co-authors and collaborators especially Anders Bach, and Neil Ferguson for the excellent work in the last paper All past and present members of the B9:4 corridor, Erik, Sophia, Wei, Ieva, Birgitta, Sandra, Pia, Karin you name the rest Past and present members at IMBIM for making the place such a wonderful one for me and without you I will not have been able to do this, especially to Eva G. and Sabrina my sincere thanks for your support during my first year at IMBIM, Vahid buddy what would I have done with you, Anh-Tri “hello small guy”, Marisa, Kathrin always nice and eager to know, Amani and Nizar (my “African” community). To the technical staff at IMBIM especially Olav, Kerstin, Marianne, Barbro, Erika, Rehne for all your assistance To my friends and people I have met throughout the years, seriously guys I will not have been here without you. Special thank you to Dr. Trivial man (Nyadong) you have a special place in my heart, Noela, Banabas and Bertile, Adiaba, Karl, Alvaro, Ylva I, JJ, My fellow Cameroonians in Uppsala and in particular those “belonging”, to all my football friends for always being there, Anne-Sophie for always being nice and happy. For all those whose names have not been mentioned I owe you more than words can offer A big hug and thank you to Liz (Elizabeth Morris) for taking time of your busy schedule to go through my thesis you are a friend indeed A special thank you to Florence for your support and kindness; words alone cannot express how nice you are in times of need A big thanks to my mother for having made it possible that I am who I am And finally to Henry for being such a great, wonderful and happy kid 35.

(50)

(51) References. 1.. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.. 16. 17.. Stiffler, M. A., Chen, J. R., Grantcharova, V. P., Lei, Y., Fuchs, D., Allen, J. E., Zaslavskaia, L. A., and MacBeath, G. (2007) Science 317, 364-369 Nourry, C., Grant, S. G. N., and Borg, J.-P. (2003) Sci. STKE 2003, re7Kim, E., Niethammer, M., Rothschild, A., Jan, Y. N., and Sheng, M. (1995) Nature 378, 85-88 Kornau, H. C., Schenker, L. T., Kennedy, M. B., and Seeburg, P. H. (1995) Science 269, 1737-1740 Niethammer, M., Kim, E., and Sheng, M. (1996) J Neurosci 16, 2157-2163 Liu, Y., Henry, G. D., Hegde, R. S., and Baleja, J. D. (2007) Biochemistry 46, 10864-10874 Doyle, D. A., Lee, A., Lewis, J., Kim, E., Sheng, M., and MacKinnon, R. (1996) Cell 85, 1067-1076 Lockless, S. W., and Ranganathan, R. (1999) Science 286, 295-299 Liu, Z., Chen, J., and Thirumalai, D. (2009) Proteins 77, 823831 Fuentes, E. J., Gilmore, S. A., Mauldin, R. V., and Lee, A. L. (2006) J Mol Biol 364, 337-351 Kong, Y., and Karplus, M. (2009) Proteins 74, 145-154 Dhulesia, A., Gsponer, J., and Vendruscolo, M. (2008) J Am Chem Soc 130, 8931-8939 Malatesta, F. (2005) Biophys Chem 116, 251-256 Chi, C. N., Engström, Å., Gianni, S., Larsson, M., and Jemth, P. (2006) J Biol Chem 281, 36811-36818 Chi, C. N., Bach, A., Engström, Å., Wang, H., Strmgaard, K., Gianni, S., and Jemth, P. (2009) Biochemistry 48, 70897097 Jemth, P., and Gianni, S. (2007) Biochemistry 46, 8701-8708 Lim, I. A., Hall, D. D., and Hell, J. W. (2002) J Biol Chem 277, 21697-21711. 37.

(52) 18.. 19. 20.. 21. 22. 23. 24. 25. 26. 27. 28.. 29. 30.. 31.. 32.. 33. 34.. 38. Gianni, S., Engström, Å., Larsson, M., Calosci, N., Malatesta, F., Eklund, L., Ngang, C. C., Travaglini-Allocatelli, C., and Jemth, P. (2005) J Biol Chem 280, 34805-34812 Harris, B. Z., and Lim, W. A. (2001) J Cell Sci 114, 32193231 Madsen, K. L., Beuming, T., Niv, M. Y., Chang, C.-w., Dev, K. K., Weinstein, H., and Gether, U. (2005) J Biol Chem 280, 20539-20548 Novak, K. A. P., Fujii, N., and Guy, R. K. (2002) Bioorg & Med Chem Lett 12, 2471-2474 Milev, S., Bjelic, S., Georgiev, O., and Jelesarov, I. (2007) Biochemistry 46, 1064-1078 Saro, D., Klosi, E., Paredes, A., and Spaller, M. R. (2004) Org. Lett. 6, 3429-3432 Tochio, H., Zhang, Q., Mandal, P., Li, M., and Zhang, M. (1999) Nat Struct Biol 6, 417-421 Penkert, R. R., DiVittorio, H. M., and Prehoda, K. E. (2004) Nat Struct Mol Biol 11, 1122-1127 Dev, K. K., van der Putten, H., Sommer, B., and Rovelli, G. (2003) Neuropharmacology 45, 1-13 Daniels, D. L., Cohen, A. R., Anderson, J. M., and Brunger, A. T. (1998) Nat Struct Biol 5, 317-325 Wong, H.-C., Bourdelas, A., Krauss, A., Lee, H.-J., Shao, Y., Wu, D., Mlodzik, M., Shi, D.-L., and Zheng, J. (2003) Molecular Cell 12, 1251-1260 Peterson, F. C., Penkert, R. R., Volkman, B. F., and Prehoda, K. E. (2004) Mol Cell 13, 665-676 Grembecka, J., Cierpicki, T., Devedjiev, Y., Derewenda, U., Kang, B. S., Bushweller, J. H., and Derewenda, Z. S. (2006) Biochemistry 45, 3674-3683 Wu, J., Yang, Y., Zhang, J., Ji, P., Du, W., Jiang, P., Xie, D., Huang, H., Wu, M., Zhang, G., Wu, J., and Shi, Y. (2007) J Biol Chem 282, 35988-35999 Im, Y. J., Park, S. H., Rho, S. H., Lee, J. H., Kang, G. B., Sheng, M., Kim, E., and Eom, S. H. (2003) J Biol Chem 278, 8501-8507 Grootjans, J. J., Reekmans, G., Ceulemans, H., and David, G. (2000) J Biol Chem 275, 19933-19941 Long, J., Wei, Z., Feng, W., Yu, C., Zhao, Y.-x., and Zhang, M. (2007) J Mol Biol 375, 1457-1468.

(53) 35.. 36. 37. 38. 39. 40. 41. 42. 43.. 44. 45. 46. 47. 48. 49. 50. 51. 52.. 53. 54.. Goult, B. T., Rapley, J. D., Dart, C., Kitmitto, A., Grossmann, J. G., Leyland, M. L., and Lian, L. Y. (2007) Biochemistry 46, 14117-14128 Pegan, S., Tan, J., Huang, A., Slesinger, P. A., Riek, R., and Choe, S. (2007) Biochemistry 46, 5315-5322 Levinthal, C. (1968) J Chim Phys 85 Jackson, S. E., and Fersht, A. R. (1991) Biochemistry 30, 10428-10435 Itzhaki, L. S., Otzen, D. E., and Fersht, A. R. (1995) J Mol Biol 254, 260-288 Ptitsyn, O. B. (1995) Trends in Biochemical Sciences 20, 376379 Leopold, P. E., Montal, M., and Onuchic, J. N. (1992) Proc Natl Acad Sci U S A 89, 8721-8725 Clementi, C., Nymeyer, H., and Onuchic, J. N. (2000) J Mol Biol 298, 937-953 Fersht, A. (1998) Structure and Mechanism in Protein Science; A Guide to Enzyme Catalysis and Protein Folding, W.H Freeman and Company) Myers, J. K., Pace, C. N., and Scholtz, J. M. (1995) Protein Sci 4, 2138-2148 Hammond, G. S. (1955) J. Am. Chem. Soc. 77, 334-338 Silow, M., and Oliveberg, M. (1997) Biochemistry 36, 76337637 Oliveberg, M., Tan, Y.-J., Silow, M., and Fersht, A. R. (1998) J Mol Biol 277, 933-943 Walkenhorst, W. F., Green, S. M., and Roder, H. (1997) Biochemistry 36, 5795-5805 Kiefhaber, T., Bachmann, A., Wildegger, G., and Wagner, C. (1997) Biochemistry 36, 5108-5112 Sanchez, I. E., and Kiefhaber, T. (2003) J Mol Biol 325, 367376 Gianni, S., Brunori, M., Jemth, P., Oliveberg, M., and Zhang, M. (2009) Biochemistry 48, 11825-11830 Wensley, B. G., Batey, S., Bone, F. A., Chan, Z. M., Tumelty, N. R., Steward, A., Kwa, L. G., Borgia, A., and Clarke, J. Nature 463, 685-688 Chi, C. N., Gianni, S., Calosci, N., Travaglini-Allocatelli, C., Engström, Å., and Jemth, P. (2007) FEBS Lett 581, 1109-1113 Saro, D., Li, T., Rupasinghe, C., Paredes, A., Caspers, N., and Spaller, M. R. (2007) Biochemistry 46, 6340-6352. 39.

(54) 55. 56. 57.. 58. 59. 60. 61. 62. 63. 64.. 65.. 66. 67. 68. 69. 70.. 71. 72.. 73.. 40. Harris, B. Z., Lau, F. W., Fujii, N., Guy, R. K., and Lim, W. A. (2003) Biochemistry 42, 2797-2805 Mittermaier, A., and Kay, L. E. (2006) Science 312, 224-228 Gianni, S., Walma, T., Arcovito, A., Calosci, N., Bellelli, A., Engstrom, A., Travaglini-Allocatelli, C., Brunori, M., Jemth, P., and Vuister, G. W. (2006) Structure 14, 1801-1809 Chi, C. N., Elfstrom, L., Shi, Y., Snall, T., Engstrom, A., and Jemth, P. (2008) Proc Natl Acad Sci U S A 105, 4679-4684 Monod, J., Wyman, J., and Changeux, J. P. (1965) J Mol Biol 12, 88-118 Koshland, D. E., Jr., Nemethy, G., and Filmer, D. (1966) Biochemistry 5, 365-385 Lancet, D., and Pecht, I. (1976) Proc Natl Acad Sci U S A 73, 3549-3553 Gunasekaran, K., Ma, B., and Nussinov, R. (2004) Proteins 57, 433-443 Pan, H., Lee, J. C., and Hilser, V. J. (2000) Proc Natl Acad Sci U S A 97, 12020-12025 Gsponer, J., Christodoulou, J., Cavalli, A., Bui, J. M., Richter, B., Dobson, C. M., and Vendruscolo, M. (2008) Structure 16, 736-746 Lange, O. F., Lakomek, N.-A., Fares, C., Schroder, G. F., Walter, K. F. A., Becker, S., Meiler, J., Grubmuller, H., Griesinger, C., and de Groot, B. L. (2008) Science 320, 14711475 Daggett, V., and Fersht, A. R. (2003) Trends Biochem Sci 28, 18-25 Friel, C. T., Capaldi, A. P., and Radford, S. E. (2003) J Mol Biol 326, 293-305 Ferguson, N., Capaldi, A. P., James, R., Kleanthous, C., and Radford, S. E. (1999) J Mol Biol 286, 1597-1608 Karplus, M., and Weaver, D. L. (1994) Protein Sci 3, 650-668 Gianni, S., Guydosh, N. R., Khan, F., Caldas, T. D., Mayor, U., White, G. W., DeMarco, M. L., Daggett, V., and Fersht, A. R. (2003) Proc Natl Acad Sci U S A 100, 13286-13291 Feng, H., Vu, N.-D., and Bai, Y. (2005) J Mol Biol 346, 345353 Ivarsson, Y., Travaglini-Allocatelli, C., Jemth, P., Malatesta, F., Brunori, M., and Gianni, S. (2007) J Biol Chem 282, 85688572 Wolynes, P. G., Onuchic, J. N., and Thirumalai, D. (1995) Science 267, 1619-1620.

(55) 74. 75. 76.. 77.. 78.. 79. 80. 81. 82. 83. 84.. 85.. 86.. 87. 88. 89. 90.. 91. 92.. Dinner, A. R., Sali, A., Smith, L. J., Dobson, C. M., and Karplus, M. (2000) Trends Biochem Sci 25, 331-339 Bryngelson, J. D., and Wolynes, P. G. (1987) Proc Natl Acad Sci U S A 84, 7524-7528 Jemth, P., Gianni, S., Day, R., Li, B., Johnson, C. M., Daggett, V., and Fersht, A. R. (2004) Proc Natl Acad Sci U S A 101, 6450-6455 Gianni, S., Geierhaas, C. D., Calosci, N., Jemth, P., Vuister, G. W., Travaglini-Allocatelli, C., Vendruscolo, M., and Brunori, M. (2007) Proc Natl Acad Sci U S A 104, 128-133 Calosci, N., Chi, C. N., Richter, B., Camilloni, C., Engstrom, A., Eklund, L., Travaglini-Allocatelli, C., Gianni, S., Vendruscolo, M., and Jemth, P. (2008) Proc Natl Acad Sci U S A 105, 19241-19246 Malenka, R. C., and Nicoll, R. A. (1993) Trends in Neurosciences 16, 521-527 Olney, J. W. (1990) Annual Review of Pharmacology and Toxicology 30, 47-71 Bordi, F., and Ugolini, A. (1999) Progress in Neurobiology 59, 55-79 Pisani, A., Bonsi, P., Centonze, D., Gubellini, P., Bernardi, G., and Calabresi, P. (2003) Neuropharmacology 45, 45-56 Boulet, G., Horvath, C., Vanden Broeck, D., Sahebali, S., and Bogers, J. (2007) Int J Biochem Cell Biol 39, 2006-2011 Kang, B. S., Cooper, D. R., Jelen, F., Devedjiev, Y., Derewenda, U., Dauter, Z., Otlewski, J., and Derewenda, Z. S. (2003) Structure 11, 459-468 Bach, A., Chi, C. N., Olsen, B. T., Pedersen, W. S., Roder, U. M., Pang, F. G., Clausen, R. P., Jemth, P. and Strmgaard K. (2008) J Med Chem 51, 6450-6459 Bach, A., Chi, C. N., Pang, G. F., Olsen, L., Kristensen, A. S., Jemth, P., and Stromgaard, K. (2009) Angew Chem Int Ed Engl 48, 9685-9689 Jencks, W. P. (1981) Proc Natl Acad Sci U S A 78, 4046-4050 Zhou, H. X. (2001) Biochemistry 40, 15069-15073 Zhou, H. X. (2003) J Mol Biol 329, 1-8 Bosch, F. X., Manos, M. M., Munoz, N., Sherman, M., Jansen, A. M., Peto, J., Schiffman, M. H., Moreno, V., Kurman, R., and Shah, K. V. (1995) J Natl Cancer Inst 87, 796-802 zur Hausen, H. (2002) Nat Rev Cancer 2, 342-350 Sedman, S. A., Hubbert, N. L., Vass, W. C., Lowy, D. R., and Schiller, J. T. (1992) J Virol 66, 4201-4208 41.

(56) 93. 94. 95. 96. 97. 98.. 42. Kiyono, T., Hiraiwa, A., Ishii, S., Takahashi, T., and Ishibashi, M. (1994) J Virol 68, 4656-4661 Lee, S. S., Weiss, R. S., and Javier, R. T. (1997) Proc Natl Acad Sci U S A 94, 6670-6675 Zhang, Y., Dasgupta, J., Ma, R. Z., Banks, L., Thomas, M., and Chen, X. S. (2007) J Virol 81, 3618-3626 Wang, W., Weng, J., Zhang, X., Liu, M., and Zhang, M. (2009) J Am Chem Soc 131, 787-796 Ebbinghaus, S., Dhar, A., McDonald, J. D., and Gruebele, M. Nat Methods 7, 319-323 Tompa, P. (2002) Trends Biochem Sci 27, 527-533.

(57)

(58) *+,-)(%,,-./(.

(59)

(60)

(61)

(62)

(63) ,%&0( (1,0(2+$/,314( +( * +,%/ (%,,1%,0(2+$/,314( +(5-../ -)(%,35$$//3$%31$'(%1..(%6*1(7 +.(1,0(+./(,( (%,,%(8(.,,9%7( 0 %((%+0/'%%(570/(,0($%3/( ,%'$,( ,(%,//3,0%$:0,0((%( :,/.%(0()( $%(1-../ (%,,1%,0(2+$/,31 4( +(6 %%,;$%355,0((%(7.$'/0( $ (%,0(,,/(<.%(0()($%(1-../ (%,,1%,0(2+$/,314( +(=6

(64). ,%'$,&.$'/+,6$$6( $%&'&(&$$& )!.

(65)

(66)

(67) .

(68)

No results found