A conserved folding mechanism for PDZ domains

FEBS Letters 581 (2007) 1109–1113

A conserved folding mechanism for PDZ domains Celestine N. Chia,1, Stefano Giannib,*,1, Nicoletta Caloscib, Carlo Travaglini-Allocatellib, ˚ ke Engstro¨ma, Per Jemtha,* A b

a Department of Medical Biochemistry and Microbiology, Uppsala University, BMC Box 582, SE-75123 Uppsala, Sweden Dipartimento di Scienze Biochimiche ‘‘A. Rossi Fanelli’’ and Istituto di Biologia e Patologia Molecolari del CNR, Universita` di Roma ‘‘La Sapienza’’, Piazzale A. Moro 5, 00185 Rome, Italy

Received 13 November 2006; revised 22 January 2007; accepted 6 February 2007 Available online 15 February 2007 Edited by Gianni Cesareni

Abstract An important question in protein folding is whether the folding mechanism is sequence dependent and conserved for homologous proteins. In this work we compared the kinetic folding mechanism of five postsynaptic density protein-95, disc-large tumor suppressor protein, zonula occludens-1 (PDZ) domains, sharing similar topology but having different primary structures. Investigation of the different proteins under various experimental conditions revealed that the folding kinetics of each member of the PDZ family can be described by a model with two transition states separated by an intermediate. Moreover, the positions of the two transition states along the reaction coordinate (as given by their bT-values) are fairly constant for the five PDZ domains.  2007 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Protein folding; Folding mechanism; PDZ domain; Kinetics

1. Introduction At first sight the protein folding reaction appears enormously complex with hundreds of non-covalent interactions forming simultaneously. But, protein folding can be dissected like any chemical reaction: identify and characterize substrates, products and intermediates, and determine their pathway of interconversion and the associated rate constants. A powerful approach to elucidate the relationships between sequence information and folding mechanism is to study proteins which differ in sequence but share the same overall fold. The strategy assumes that general correlations between aminoacid sequences and folding pathways may be extrapolated by comparing folding processes for different members of a given protein family. Such studies have shown that the overall folding mechanism is generally conserved within a fold family, and hidden common features may be unveiled even when apparently different folding mechanisms are observed. For example, in the case of the homeodomain superfamily *

Corresponding authors. Fax: +46 18 471 4673 (P. Jemth). E-mail addresses: [email protected] (S. Gianni), [email protected] (P. Jemth). 1

These authors contributed equally to the work.

Abbreviations: PDZ, postsynaptic density protein-95, disc-large tumor suppressor protein, zonula occludens-1; TS, transition state

[1], where the folding processes may span from pure nucleation–condensation [2] to framework [3] mechanisms, the structure of the transition state is rather conserved [4]. Similarly for the bacterial immunity proteins Im9 and Im7, which appear to fold by two-state or three-state mechanisms [5], the late transition state ensembles of the two proteins have similar properties [6]. It has been shown previously that several proteins, including small single domain proteins, may accumulate obligatory folding intermediates [7,8]. Under such circumstances it is interesting to compare the structural properties of intermediate states, as well as the intervening transition states, when the amino acid sequence is varied but the three-dimensional structure is relatively unchanged [9]. In this study we have analysed and compared the kinetic folding mechanisms of four related but distinct postsynaptic density protein-95, disc-large tumor suppressor protein, zonula occludens-1 (PDZ) domains together with one previously published [10,11]. All the proteins considered display the canonical PDZ fold [12]. Pairwise comparisons between the sequences of the PDZ domains give 25– 50% identity but only 12 residues are conserved among all five domains (Fig. 1). We found that, despite this low identity and an apparent folding complexity, the folding reactions for PDZ domains can be explained by a model with an intermediate and two transition states that are rather conserved with regard to their positions along the folding reaction coordinate.

2. Materials and methods 2.1. PDZ constructs The following PDZ domains were expressed as His-tagged proteins (MHHHHHPRGS-etc): PSD-95 PDZ1 (61–151), PSD-95 PDZ2 (155– 249), and PSD-95 PDZ3 (309–401) (numbers in parenthesis refer to residue numbers in the parental human full-length PSD-95a without exon 4b, i.e. the same numbering as used in Doyle et al. [12] and Tochio et al. [13]). The PSD-95 PDZ domains will be referred to as PDZ1, PDZ2 and PDZ3, respectively. The pdb codes for the solved structures of these three PDZ domains are: PDZ1, 1IUO [14] (the numbering in the pdb file is 1–91), PDZ2, 1QLC [13], and PDZ3, 1BE9 and 1BFE [12]. For human neuronal nitric oxide synthase (nNOS) PDZ, a construct expressing residues 1–132 of nNOS was used (pdb codes: 1QAU [15] and 1B8Q [16]). The typical PDZ domain consists of six b-strands, bA–bF and two a-helices, aA and aB [12–16]. For each of the PDZ domains a Trp residue was introduced as a fluorescent probe [10,17–19] for the folding studies: PDZ1 F95W; PDZ2 Y190W, and PDZ3 F337W. These three Trps were situated in the bC strand of the respective PDZ domain, in the corresponding position to that used before in folding studies of PTP-BL PDZ2 [10,11] and PDZ3 [17]. The corresponding mutant of nNOS PDZ, I42W (numbering according to full-length nNOS, used in Hillier et al. [15]), precipitated easily and

0014-5793/$32.00  2007 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2007.02.011

1110

C.N. Chi et al. / FEBS Letters 581 (2007) 1109–1113

Fig. 1. Sequence comparison of the five PDZ domains included in the study. All constructs also contained the His tag MHHHHHPRGS at the Nterminus, except PTP-BL PDZ2 which started with MHHHHHHM. The residues in bold were mutated to Trp to obtain a fluorescent label for the folding studies. Identical conserved residues are denoted by a star, *. See Section 2 for details on the numbering.

another Trp mutant, F21W, was used in the folding experiments. Also, the nNOS PDZ contained two additional strands in the C-terminus as compared to PDZ1, 2 and 3. The reason was that the shorter nNOS PDZ 1–99 construct expressed poorly and was also precipitating easily. 2.2. Protein purification The cDNAs encoding PDZ1, PDZ2, PDZ3 and nNOS PDZ were cloned into the BamH1 and EcoR1 restriction sites of a modified pRSET vector (Invitrogen), resulting in clones expressing His-tagged PDZ domain. Protein expression and purification were as described [18,20] but the final purification step was different for PDZ2 and nNOS PDZ after nickel (II)-charged chelating sepharose column chromatography: A concentrated sample of nNOS PDZ was passed through a QSepharose column equilibrated with 50 mM Tris–HCl pH 8.5 and the unbound fraction contained pure nNOS PDZ. For PDZ2 the most concentrated samples after the nickel column were pooled together and loaded on a Superdex-75 (GE healthcare) gel filtration column equilibrated with 20 mM Tris–HCl pH 7.5. Fractions containing pure PDZ as judged from SDS–PAGE were pooled. MALDI-TOF mass spectrometry was used to confirm the identity of the purified PDZ domains and their concentrations were determined from extinction coefficients calculated from amino acid analysis or standard values of the amino acid residues [21]. 2.3. Stopped-flow spectroscopy Kinetic folding and unfolding rate constants were measured on SX17 and SX18 stopped-flow spectrometers from Applied Photophysics (Leatherhead, UK). PDZ domain in buffer with or without urea was mixed to final concentrations of 2–3 lM with buffer-urea solutions to induce either refolding or unfolding of the protein. Excitation was at 280 nm and the change in Trp emission with time was recorded either above 320 nm using a cut-off filter or around 330 nm using an interference emission filter. 2.4. Data analysis The kinetic data were analyzed assuming a linear dependence of the logarithm of the rate constants on urea concentration [22]. The linear version of the equations are

2O expðmD-z ½urea=RT Þ kf ¼ kH f

ð1Þ

k u ¼ k uH2 O expðmz-N ½urea=RT Þ

ð2Þ

mD-à and mà-N are the respective slopes of the dependence of log kf and log ku on [urea]. The equation describing a two-state mechanism is the sum of Eqs. (1) and (2). When there was a change in rate-limiting step in either the refolding or unfolding arm the data were analyzed according to a three-state model and the sum of Eqs. (2) and (3) (refolding) or Eqs. (1) and (4) (unfolding) were used [23]. 2O expðmD-z ½urea=RT Þ  1=f1 þ K p expðmp ½urea=RT Þg kf ¼ kH f

ð3Þ

k u ¼ k uH2 O expðmz1-N ½urea=RT Þ  1=f1 þ K p expðmp ½urea=RT Þg

ð4Þ

In Eq. (3), Kp is the equilibrium (or partitioning) constant between the denatured and intermediate populations and mp the associated m value. In the case of Eq. (4), Kp is the partitioning constant between the two transition states and mp the associated m value. Global fit of folding kinetics to common bT values was performed assuming: bT1 ¼ mD-z1 =mD-N

ð5Þ

bT2 ¼ mD-z2 =mD-N ¼ 1  ðmz1-N  mp Þ=mD-N

ð6Þ

Data were analyzed with Kaleidagraph (Synergy Software) and Prism (GraphPad Software, Inc.).

3. Results 3.1. Stability of PDZ constructs His-tagged versions of the following PDZ domains were expressed and purified as described in Section 2: PDZ1, PDZ2 and PDZ3 from PSD-95, and nNOS PDZ. All four PDZ domains were folded as judged by far-UV circular dichroism measurements (not shown) and fluorescence-monitored denaturation (Fig. 2). Equilibrium denaturation experiments were performed in 50 mM potassium phosphate pH 7.45 by increas-

C.N. Chi et al. / FEBS Letters 581 (2007) 1109–1113

1111

1

0.8

10

0.7

kobs (s-1)

Fluorescence (relative units)

0.9

PDZ1 PDZ2 PDZ3

100

PDZ1 PDZ2 PDZ3 nNOS PDZ

0.6

1

0.5 0.1 0.4

0.3 0

2

4

6

0

8

2

4

6

8

[Urea] (M)

[Urea] (M)

Fig. 3. Observed rate constants for folding of PSD-95 PDZ domains versus [urea] (chevron plots), at pH 7.4 in 50 mM KPi. Data were fitted to a two state model (solid lines, cf Eqs. (1) and (2)). See Table 1 for fitted parameters.

Fig. 2. Equilibrium denaturation of PSD-95 PDZ domains at pH 7.4 and nNOS PDZ at pH 4.0. Data were fitted to the equation describing solvent denaturation of a two state system (solid line) [22]. See Table 1 for fitted parameters.

pH) conditions. Our results are summarized in Fig. 4 and they demonstrate that the folding experiments can be conducted under conditions such that all PDZ domains display kinetics characteristic of a three-state mechanism, where transition state 1 (TS1) is rate limiting at lower urea and TS2 at higher urea concentrations. Clear deviations from linearity were detected in the unfolding arm for PDZ1, 2 and 3, and in the refolding arm for nNOS PDZ. For PDZ1, the change in fluorescence signal upon denaturation disappeared at low pH and the more powerful guanidine HCl was used as denaturant to expose the curvature of the unfolding arm (Fig. 4). Experiments on PDZ1 at 40 C with urea gave a similar result (not shown). Observed rate constants for PDZ1, 2, 3, and PTP-BL PDZ2 were fitted simultaneously to the sum of Eqs. (1) and (4), and nNOS PDZ to the sum of Eqs. (2) and (3), where the kinetic m values were replaced by functions of bT1, bT2 and mD-N (cf. Eqs. (5) and (6) (Fig. 4). The position of the two transition states (bT1 and bT2) were shared among the data sets. The data for PDZ1, 2, 3, and PTP-BL PDZ2 were all well described by the two bT values 0.66 and 0.93 suggesting that even the positions of the transition states may be conserved in the folding reaction of most PDZ domains. It should also be noted that global fit of data for all five PDZ domains gives a reasonable fit with bT values of 0.3 and 0.8 for TS1 and TS2, respectively,

ing the urea concentration at a constant PDZ concentration and record the fluorescence at 340 nm upon excitation at 280 nm. The data were fit to the general equation for solvent denaturation of a two-state protein [22] to generate equilibrium constants for denaturation (Table 1). 3.2. Folding of PDZ domains We measured the refolding and unfolding rate constants for PDZ1, 2 and 3 at different urea concentrations, in 50 mM potassium phosphate, pH 7.45 (so-called chevron plots, Fig. 3, Table 1). Experimental kinetic traces were satisfactorily described by single exponential kinetics for all PDZ domains. Whilst application of the two-state assumption (Eqs. (1) and (2)) yielded good fits to data for PDZ1 and 3, in the case of PDZ2 there was a minor deviation of the unfolding rate constants from linearity, as mirrored by analysis of curve fit residuals (data not shown). Folding and unfolding rate constants for nNOS PDZ were difficult to collect at neutral pH due to low kobs values. Previous experiments [10] suggested that two different transition states are rate limiting under different conditions for PTP-BL PDZ2. To investigate whether this behaviour is generic for PDZ domains we subjected PDZ1, 2, 3 and nNOS PDZ to denaturant-induced folding experiments under various stabilising (addition of sulfate) or destabilising (low

Table 1 Equilibrium (Fig. 2) and kinetic (Fig. 3) data for PDZ domains at pH 7.4 PDZ

Equilibrium

Kinetics 1

mD-N (kcal mol PSD-95 PDZ1 PSD-95 PDZ2 PSD-95 PDZ3 nNOS PDZa PTP-BL PDZ2b

1.42 ± 0.08 1.05 ± 0.05 1.02 ± 0.07 1.8 ± 0.5 1.2 ±0.1

1

M )

[Urea]50% (M)

2O DGH D-N

3.31 ± 0.03 2.89 ± 0.04 5.36 ± 0.06 3.0 ± 0.1

4.7 ± 0.2 3.0 ± 0.2 5.5 ± 0.4 5.4 ± 1.5 2.9 ± 0.2

1

ðkcal mol Þ

Fitting errors are shown. The experiment was performed at pH 4.0. See Table 2 for kinetic mD-N and [Urea]50% values. b Data from Gianni et al. [10], pH 7.2. a

mD-N (kcal mol1 M1)

[Urea]50% (M)

1.34 ± 0.02 1.19 ± 0.01 1.11 ± 0.01

3.09 ± 0.03 2.66 ± 0.02 5.35 ± 0.01

1112

C.N. Chi et al. / FEBS Letters 581 (2007) 1109–1113

showing that a large range of bT1 that can describe the data accurately. bT values for fitting of individual data sets are given in Table 2. It is clear that the exact position of TS1 may differ among the PDZ domains (0.21–0.66) whereas the position of TS2 is very conserved (0.81–0.88).

2

PDZ1 PDZ2 PDZ3 PTP-BL PDZ2 nNOS PDZ

1

log kobs/s-1

4. Discussion An increasing number of studies have shown that, despite the inherent complexity of the protein folding reaction, the denatured chain folds to its native conformation in a surprisingly simple way, often without detectable intermediates [24]. Identification of the minimal kinetic mechanism describing the folding of a protein is a crucial task in addressing the molecular details of the folding reaction. In this work we compared the folding mechanism of five members of the PDZ domain family. A comparison of their sequences shows that the sequence identity among all five domains is approximately 13% (Fig. 1). Despite the low identity, observed folding kinetics are consistent with a three-state model with two transition states and an intermediate, where the first transition state TS1 is rate limiting in water for all PDZ domains tested, and the native-like TS2 becomes rate limiting at different denaturant concentrations for the different PDZ domains (resulting in an unfolding or refolding roll over). Furthermore, the unfolding/refolding amplitudes show no indication of any additional ultra-fast burst phases undetectable for a stopped flow appara-

0

-1

-2 0

2

4

6

8

[Denaturant] (M)

Fig. 4. Chevron plots of five PDZ domains under conditions that reveal changes in rate-limiting step. PDZ1, 50 mM KPi, pH 7.4 with guanidine HCl as denaturant; PDZ2, 50 mM K-acetate, pH 5.45; PDZ3, 100 mM K-formate, pH 2.85; nNOS PDZ, 50 mM K-formate, pH 4.0; PTP-BL PDZ2, 50 mM NaPi, pH 7.2. All experiments were performed at 25C. The data for PDZ1, 2, 3 and PTP-BL PDZ2 were fit simultaneously to a common set of bT values (bT1 = 0.66 ± 0.01 and bT2 = 0.93 ± 0.04, fitting errors). See Table 2 for fits to individual data sets.

Table 2 Kinetic parameters for the folding of PDZ domains in Fig. 3 (see figure legend for experimental conditions) PDZ

kf (s1)

ku1 (s1)

bT1

bT2

mD-N (kcal mol1 M1)

H2 O DGD-N ðkcal mol1 Þ

PSD-95 PDZ1a PSD-95 PDZ2 PSD-95 PDZ3 nNOS PDZ PTP-BL PDZ2b

3.5 ± 0.1 1.4 ± 0.1 14 ± 0.3 6.3 ± 0.9 2.8

(6.8 ± 0.7) · 104 (1.4 ± 0.5) · 103 0.75 ± 0.01 (16 ± 3) · 103 0.020

0.66 ± 0.01 0.51 ± 0.03 0.53 ± 0.01 0.21 ± 0.01 0.53

0.84 ± 0.01 0.81 ± 0.02 0.87 ± 0.01 0.88 ± 0.01 0.89

3.16 ± 0.03 1.36 ± 0.06 1.41 ± 0.02 1.96 ± 0.09 1.34

5.05 ± 0.06 4.1 ± 0.2 3.54 ± 0.06 6.7c 2.9

Fitting errors are shown. a Guanidine–HCl was used as denaturant. b Data from Gianni et al. [10], pH 7.2. bT values and mD-N for PTP-BL PDZ2 were from a global fit of 10 chevrons measured at different pH values. c Very large error due to long extrapolation of the second folding rate constant.

TS2

TS1 TS1

TS2

I

I

G

D

N

0

N

G

0.5

βT

1

D

0

0.5

1

βT

Fig. 5. Energy diagram for the folding reactions of PDZ domains at low (left panel) and high (right panel) urea concentration. The m values for protein folding are related to the compactness of the molecules [25] and their ratios (e.g. mD-à1/mD-N = bT1) can therefore be used to define the progress of the folding reaction along the reaction coordinate.

C.N. Chi et al. / FEBS Letters 581 (2007) 1109–1113

tus, suggesting that the intermediate is a high energy species (Fig. 5). Remarkably, the positions of the two transition states along the reaction coordinate change little within the family as shown by the global fit of folding data with shared bT values. This similarity is reflected in the m-values, i.e. the slopes of the observed rate constants versus [denaturant] as seen in Fig. 4. The position of TS2 is particularly conserved whereas TS1 appears more malleable in terms of compactness. The lower bT for TS1 for nNOS PDZ may be partly due to the higher mD-N value of nNOS PDZ (Tables 1 and 2) which would give a lower bT1 if the extra two strands in the C-terminal extension are not folded in TS1 (cf Fig. 1). The folding pathway of PDZ3 from PSD-95 has been studied recently by native hydrogen exchange [17]. Quantitative analysis of exchange kinetics revealed the presence of an intermediate in PDZ3 folding. The authors proposed that the structure of the folding intermediate resembles that of the native PDZ fold, except for the N- and C-terminal regions which are unfolded. Importantly, the intermediate detected by Bai and co-workers involves strands that are C-terminal to the canonical PDZ domain and not present in our PDZ constructs. In conclusion, the high energy intermediate identified here is common to all the PDZ domains considered in this work and it may be a generic species in the folding of PDZ domains. Our data thus support the notion that topology governs folding mechanism. The major difference in the folding pathway between members of the same family simply lies in the relative stability, and perhaps structure, of the intermediate species and the intervening transition states, which may thus result in apparently different kinetic behavior. Acknowledgements: This work was supported by grants from the Swedish Research Council, Magnus Bergvall’s foundation, the Royal Swedish Academy of Sciences, Linne´stiftelsen, La¨ngmanska Kulturfonden, the Medical Faculty, Uppsala University and the Department of Medical Biochemistry and Microbiology, Uppsala University (to PJ), and by a grant from the Italian Ministero dell’ Istruzione dell’ Universita` e della Ricerca 2005027330_005 to (CTA).

References [1] Daggett, V. and Fersht, A.R. (2003) Is there a unifying mechanism for protein folding? Trends Biochem. Sci. 28, 18–25. [2] Itzhaki, L.S., Otzen, D.E. and Fersht, A.R. (1995) The structure of the transition state for folding of chymotrypsin inhibitor 2 analysed by protein engineering methods: evidence for a nucleation–condensation mechanism for protein folding. J. Mol. Biol. 254, 260–288. [3] Karplus, M. and Weaver, D.L. (1994) Protein folding dynamics: the diffusion–collision model and experimental data. Protein Sci. 3, 650–668. [4] 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) Unifying features in protein-folding mechanisms. Proc. Natl. Acad. Sci. USA 100, 13286–13291. [5] Ferguson, N., Capaldi, A.P., James, R., Kleanthous, C. and Radford, S.E. (1999) Rapid folding with and without populated intermediates in the homologous four-helix proteins Im7 and Im9. J. Mol. Biol. 286, 1597–1608. [6] Friel, C.T., Capaldi, A.P. and Radford, S.E. (2003) Structural analysis of the rate-limiting transition states in the folding of Im7

1113

[7]

[8] [9] [10] [11]

[12]

[13] [14]

[15]

[16]

[17] [18]

[19]

[20]

[21] [22] [23] [24] [25]

and Im9: similarities and differences in the folding of homologous proteins. J. Mol. Biol. 326, 293–305. Jemth, P., Gianni, S., Day, R., Li, B., Johnson, C.M., Daggett, V. and Fersht, A.R. (2004) Demonstration of a low-energy onpathway intermediate in a fast-folding protein by kinetics, protein engineering, and simulation. Proc. Natl. Acad. Sci. USA 101, 6450–6455. Roder, H., Maki, K. and Cheng, H. (2006) Early events in protein folding explored by rapid mixing methods. Chem. Rev. 106, 1836– 1861. Travaglini-Allocatelli, C., Gianni, S. and Brunori, M. (2004) A common folding mechanism in the cytochrome c family. Trends Biochem. Sci. 29, 535–541. Gianni, S., Calosci, N., Aelen, J., Vuister, G.W., Brunori, M. and Travaglini-Allocatelli, C. (2005) Kinetic folding mechanism of PDZ2 from PTP-BL. Protein Eng. Des. Sel. 18, 389–395. Gianni, S., Geierhaas, C.D., Calosci, N., Jemth, P., Vuister, G.W., Travaglini-Allocatelli, C., Vendruscolo, M. and Brunori, M. (2007) A PDZ domain recapitulates a unifying mechanism for protein folding. Proc. Natl. Acad. Sci. USA 104, 128–133. Doyle, D.A., Lee, A., Lewis, J., Kim, E., Sheng, M. and MacKinnon, R. (1996) Crystal structures of a complexed and peptide-free membrane protein-binding domain: molecular basis of peptide recognition by PDZ. Cell 85, 1067–1076. Tochio, H., Hung, F., Li, M., Bredt, D.S. and Zhang, M. (2000) Solution structure and backbone dynamics of the second PDZ domain of postsynaptic density-95. J. Mol. Biol. 295, 225–237. Long, J.F., Tochio, H., Wang, P., Fan, J.S., Sala, C., Niethammer, M., Sheng, M. and Zhang, M. (2003) Supramodular structure and synergistic target binding of the N-terminal tandem PDZ domains of PSD-95. J. Mol. Biol. 327, 203–214. Hillier, B.J., Christopherson, K.S., Prehoda, K.E., Bredt, D.S. and Lim, W.A. (1999) Unexpected modes of PDZ domain scaffolding revealed by structure of nNOS-syntrophin complex. Science 284, 812–815. Tochio, H., Zhang, Q., Mandal, P., Li, M. and Zhang, M. (1999) Solution structure of the extended neuronal nitric oxide synthase PDZ domain complexed with an associated peptide. Nat. Struct. Biol. 6, 417–421. Feng, H., Vu, N.D. and Bai, Y. (2005) Detection of a hidden folding intermediate of the third domain of PDZ. J. Mol. Biol. 346, 345–353. ˚ ., Larsson, M., Calosci, N., Malatesta, F., Gianni, S., Engstro¨m, A Eklund, L., Ngang, C.C., Travaglini-Allocatelli, C. and Jemth, P. (2005) The kinetics of PDZ domain-ligand interactions and implications for the binding mechanism. J. Biol. Chem. 280, 34805–34812. Gianni, S., Walma, T., Arcovito, A., Calosci, N., Bellelli, A., ˚ ., Travaglini-Allocatelli, C., Brunori, M., Jemth, P. Engstro¨m, A and Vuister, G.W. (2006) Demonstration of long-range interactions in a PDZ domain by NMR, kinetics and protein engineering. Structure 14, 1801–1809. ˚ ., Gianni, S., Larsson, M. and Jemth, P. Chi, C.N., Engstro¨m, A (2006) Two conserved residues govern the salt and pH dependences of the binding reaction of a PDZ domain. J. Biol. Chem. 281, 36811–36818. Gill, S.C. and von Hippel, P.H. (1989) Calculation of protein extinction coefficients from amino acid sequence data. Anal. Biochem. 182, 319–326. Fersht, A. (1999). Structure and Mechanism in Protein Science. W.H. Freeman and Co., New York, pp. 508–572. Parker, M.J., Spencer, J. and Clarke, A.R. (1995) An integrated kinetic analysis of intermediates and transition states in protein folding reactions. J. Mol. Biol. 253, 771–786. Jackson, S.E. and Fersht, A.R. (1991) Folding of chymotrypsin inhibitor 2.1. Evidence for a two-state transition. Biochemistry 30, 10428–10435. Myers, J.K., Pace, C.N. and Scholtz, J.M. (1995) Denaturant m values and heat capacity changes: relation to changes in accessible surface areas of protein unfolding. Protein Sci. 4, 2138–2148.