The Transition State of Coupled Folding and Binding for a Flexible β-Finger

doi:10.1016/j.jmb.2012.01.042

J. Mol. Biol. (2012) 417, 253–261 Contents lists available at www.sciencedirect.com

Journal of Molecular Biology j o u r n a l h o m e p a g e : h t t p : / / e e s . e l s e v i e r. c o m . j m b

The Transition State of Coupled Folding and Binding for a Flexible β-Finger O. Andreas Karlsson†, Celestine N. Chi†, Åke Engström and Per Jemth⁎ Department of Medical Biochemistry and Microbiology, Uppsala University, BMC Box 582, SE-75123 Uppsala, Sweden Received 1 December 2011; received in revised form 13 January 2012; accepted 25 January 2012 Available online 30 January 2012 Edited by C. R. Matthews Keywords: intrinsically disordered proteins; PDZ domain; binding kinetics; phi binding; flexible proteins

Flexible and fully disordered protein regions that fold upon binding mediate numerous protein–protein interactions. However, little is known about their mechanism of interaction. One such coupled folding and binding occurs when a flexible region of neuronal nitric oxide synthase adopts a β-finger structure upon binding to its protein ligand, a PDZ [PSD-95 (postsynaptic density protein-95)/Discs large/ZO-1] domain from PSD-95. We have analyzed this binding reaction by protein engineering combined with kinetic experiments. Mutational destabilization of the β-finger changed mainly the dissociation rate constant of the proteins and, to a lesser extent, the association rate constant. Thus, mutation affected late events in the coupled folding and binding reaction. Our results therefore suggest that the native binding interactions of the β-finger are not present in the rate-limiting transition state for binding but form on the downhill side in a cooperative manner. However, by mutation, we could destabilize the β-finger further and change the rate-limiting step such that an initial conformational change becomes rate limiting. This switch in rate-limiting step shows that multistep binding mechanisms are likely to be found among flexible and intrinsically disordered regions of proteins. © 2012 Elsevier Ltd. All rights reserved.

Introduction Intrinsically disordered proteins are involved mainly in protein–protein and protein–DNA interactions. 1–4 These disordered proteins do not adopt a well-defined conformation when free in solution, as shown by NMR, small-angle X-ray scattering and other techniques. 5,6 The whole protein can sometimes be disordered, but more commonly protein regions of varying length, ranging from domains to shorter stretches, display different degree of disorder. 2 Upon binding to their cognate ligand, the disordered regions fold

*Corresponding author. E-mail address: [email protected]. † O.A.K. and C.N.C. contributed equally to this work. Abbreviation used: nNOS, neuronal nitric oxide synthase.

into ordered structures that can be solved by, for example, X-ray crystallography. 7,8 These bound conformations may be quite extended over the surface of the binding partner 8 or, sometimes, more compact with a defined hydrophobic core. 9 Several explanations as to why intrinsically disordered proteins are so common, in particular, in signaling, have been proposed. 10 For example, if the unbound protein is unfolded, it is possible to decrease the affinity of the protein–ligand complex while retaining specific interactions (low affinity combined with high specificity). Further, disorder may promote allostery. 11 It is also argued that disordered proteins are more easily degraded such that a short intracellular half-life is achieved, but this is not necessarily the case. 12 Finally, a number of theoretical studies have highlighted possible mechanistic benefits of being unfolded prior to binding, with the fly-casting mechanism as the most prominent example. 13,14 A recent paper appraises these proposed advantages of being disordered, from a

0022-2836/$ - see front matter © 2012 Elsevier Ltd. All rights reserved.

Coupled Folding and Binding for a β-Finger

254 kinetic point of view, and reaches the conclusion that the dissociation rate constants of disordered proteins are optimized for signaling; that is, a relatively high dissociation rate is desirable. 15 However, despite their pivotal roles in signaling, few kinetic studies (e.g., Refs. 8,16 and 17) have addressed the binding mechanism of intrinsically disordered proteins. 5 There is currently an interest in structural extensions from canonical PDZ [PSD-95 (postsynaptic density protein-95)/Discs large/ZO-1] domains and their functional roles. 18–20 One of these noncanonical extensions, the one from neuronal nitric oxide synthase (nNOS) PDZ, is 32 amino acid residue long and contains segments of disordered amino acids that become ordered upon interaction with either of its ligands, PDZ2 of PSD-95 or syntrophin PDZ, as shown by NMR 21–23 and crystallography 24 (Fig. 1). This plastic C-terminal extension folds into a wellstructured β-finger in the ligand-bound form and offers an experimental system to test the binding mechanism of flexible proteins and intrinsically disordered regions. The interaction differs from that of canonical PDZ interactions, which involve the Cterminus of the protein ligand. When nNOS PDZ is the ligand, the backbone of the turn in the β-finger forms hydrogen bonds to the so-called carboxylate binding loop in PSD-95 PDZ2 (or syntrophin PDZ), which normally binds the C-terminus of the peptide ligand. Thus, this β-finger is needed for proper signaling involving nNOS. However, the mechanism of interaction for the β-finger of nNOS PDZ has not been looked at, for example, whether it involves an initial encounter complex or if there exists a small population of binding-competent forms that is in

(a)

equilibrium with an ensemble of β-finger conformations that cannot bind PSD-95 PDZ2 (conformational selection). To address these questions, we introduced mutations in the β-finger and investigated their effects on the kinetics of binding and on the stability of the bimolecular complex between nNOS PDZ and PSD-95 PDZ2. With this approach, we demonstrate that native interactions form on the downhill side of the rate-limiting barrier, similarly to what we recently observed with short disordered peptides. 26 Our data therefore suggest a generic binding mechanism for disorderto-order transitions that involve β-strands.

Results Choice of mutations We wanted to test the role of stability of a flexible region of a protein in the binding reaction. The idea was thus to modulate the stability (which should be related to its flexibility) of the β-finger of nNOS PDZ by mutation without disturbing the binding interface with PSD-95 PDZ2. We included six mutants of nNOS PDZ in this study (Table 1), all of which were situated in the flexible β-finger but not making extensive direct interactions to the PSD-95 PDZ2 protein. Because neither of the PDZ domains contains a suitable probe for fluorescence experiments, Trp residues were engineered in both PSD-95 PDZ2 (I195W) 27,28 and nNOS PDZ (V122W, in the β-finger). The six β-finger mutations were introduced

(b) I120 G113 G115

syntrophin PDZ

L107 2.7Å R121 V122 nNOS PDZ 2.9Å

D62

T123

Fig. 1. Crystal structure of nNOS PDZ bound to syntrophin PDZ. 24 (a) The residues of the β-finger mutated in this study are shown as sticks. The main-chain residues of syntrophin PDZ and nNOS PDZ are shown as surface. (b) A closeup showing some of the mutated residues (L107, I120, R121, V122 and T123) together with the salt bridge between R121 in the β-finger and D62 from the canonical part of nNOS PDZ. The picture was drawn in PyMOL. 25

Coupled Folding and Binding for a β-Finger

255

Table 1. Binding rate constants for the interaction between nNOS PDZ and PSD-95 PDZ2 and binding Φ values for mutations in the β-finger of nNOS PDZ nNOS PDZ (varied)+PSD-95 PDZ2 I195W

nNOS PDZ mutant

kon (μM− 1 s− 1 )

koff (s− 1 )

0.40 ± 0.01 0.13 ± 0.003 0.40 ± 0.01 0.18 ± 0.005 0.19 ± 0.005 — 0.31 ± 0.008

1.6 ± 0.02 9.8 ± 0.13 10.5 ± 0.3 0.21 ± 0.005 1.1 ± 0.03 0.6 ± 0.01 1.7 ± 0.10

Wild type L107A G113A G115A I120G R121A T123A a b

nNOS PDZ V122W+PSD-95 PDZ2 (varied) Φa

0.38 0.0 − 0.64b

kon (μM− 1 s− 1 )

koff (s− 1 )

0.41 ± 0.01 0.21 ± 0.005 0.37 ± 0.01 0.31 ± 0.008 0.28 ± 0.007 — 0.38 ± 0.01

0.68 ± 0.02 4.7 ± 0.24 2.6 ± 0.35 0.08 ± 0.001 0.72 ± 0.01 — 0.59 ± 0.004

Φa 0.26 0.07 − 0.15

Φ values are only reported for absolute ΔΔGEq values higher than 0.35 kcal mol− 1 [see Eq. (3)]. Two kinetic phases were observed. The major one is tabulated and used to calculate the Φ value.

both in wild-type nNOS PDZ and in the V122W pseudo wild type.

used to analyze binding traces [Eq. (5)]. The observed rate constants were plotted versus concentration of the varied species to analyze the binding mechanism (Fig. 2b). The association rate constant for nNOS PDZ was roughly 2 orders of magnitude lower than that of PDZ–peptide interactions measured by us previously (0.4 μM − 1 s − 1 for nNOS PDZ compared to around 30 μM − 1 s − 1 for both SAP97 PDZ2 and PTP-BL PDZ2 at 25 °C, with short peptides as ligands 29,30). For SAP97 PDZ2 and PTPBL PDZ2, we found a fast conformational change by fluorescence measurements in a continuous-flow spectrometer. The much lower association rate constant for nNOS PDZ suggests that a rate-limiting

Binding kinetics We performed binding experiments using the stopped-flow technique, where nNOS PDZ was mixed with PSD-95 PDZ2, and the change in Trp fluorescence over time was recorded. In most cases, the binding traces were monophasic, and data were fitted to a single-exponential equation to obtain observed rate constants kobs [Eq. (4) and Fig. 2a]. However, for some mutants, biphasic kinetics were observed, and a double-exponential equation was

(a)

(b)

5.8

30 200 PSD-95 PDZ2 varied

150

obs

-1

(s )

25

100

k

5.6

20

kobs (s-1)

Fluorescence

5.7

5.5

50

0

15

0

100 200 300 400 500 600 [PSD-95 PDZ2] (µM)

5.4 0

0.5

1

1.5

2

10

Residuals

Time (s) 0.2 0.1 0 -0.1 -0.2

nNOS PDZ varied PSD-95 PDZ2 varied

5

0 0

0.5

1

Time (s)

1.5

2

0

10

20

30

40

50

60

[PDZ] (µM)

Fig. 2. Binding kinetics of the interaction between nNOS PDZ and PSD-95 PDZ2. (a) Example of a binding trace between nNOS PDZ (5 μM) and PSD-95 PDZ2 I195W (20 μM). The residuals are from a fit to a single-exponential equation. (b) Observed rate constant plotted versus PSD-95 PDZ2 and nNOS PDZ concentrations, respectively. The inset shows kobs as a function of [PSD-95 PDZ2] at an extended concentration range.

Coupled Folding and Binding for a β-Finger

256

(a)

40 wild type nNOS PDZ L107A G113A G115A I120G T123A

35 30

20

k

obs

-1

(s )

25

15 10 5 0 0

(b)

10

20

30

40

50

60

50

60

[nNOS PDZ] (µM) 40 V122W L107A/V122W G113A/V122W G115A/V122W fast phase I120G/V122W fast phase T123A/V122W

35 30

20

k

obs

-1

(s )

25

15 10 5 0 0

(c)

10

20

30

40

[PSD-95 PDZ2] (µM) 40 1.4

35 kobs (s-1)

1.2

30

I120G/V122W

obs2 obs2

1 0.8 0.6

0.2

-1

(s )

G115A/V122W

k

0.4

25

0

20

5

10 15 20 25 30 [PSD-95 PDZ2] (µM)

35

k

obs

k

15

conformational change might be involved after association (induced fit) but that it is much slower than that involving a C-terminal peptide as ligand. Another scenario would be a very fast pre-equilibrium where a high-energy excited state of nNOS PDZ is the binding-competent conformation. In either case, the slow association is most likely related to conformational transition(s) of the β-finger of nNOS PDZ, since on-rate constants of PSD-95 PDZ2 with short peptides are similar to those of other PDZ domains. 28 For example, only certain conformation(s) of the β-finger might allow productive binding (conformational selection). One mechanism does not rule out the other, and both conformational selection and induced fit may be present. A conformational change, be it fast or slow, will eventually result in a nonlinear behavior of kobs at high enough concentration of the molecular species varied in the reaction. We looked for such deviation from linearity but could not detect any up to 500 μM PSD-95 PDZ2 (inset, Fig. 2b). In a resistive joule heating temperature jump system, we attempted even higher concentrations of PSD-95 PDZ2, but the signal-to-noise ratio was too low to detect any trace (the equilibrium of this binding reaction is only perturbed to a very small extent by a temperature increase, in particular, under these highly saturating conditions). Thus, the rate constant of any first-order step (i.e., a conformational change prior to or after the initial encounter) must be ≫ 200 s − 1 , the highest measured observed rate constant. Binding experiments were performed both by varying wild-type nNOS PDZ and following the fluorescence of PSD-95 PDZ2 I195W and by varying wild-type PSD-95 PDZ2 and following the fluorescence of nNOS PDZ V122W. These binding kinetics were not dependent on which Trp was used for detection of kinetic traces, nor the varied species (Fig. 2b and Table 1). This similarity suggests that the respective engineered Trp did not affect association kinetics (kon). Dissociation rate constants were determined in displacement reactions, which give highly accurate and precise values of koff. 28,29,31 Here, the nNOS PDZ/PSD-95 PDZ2 complex was mixed rapidly with high concentrations of a competing peptide ligand such that the complex was dissociated. This dissociation was monitored by the change in fluorescence as the competing ligand

R121A mutant

10

PSD-95 PDZ2 I195W varied nNOS PDZ R121A varied

5 0 0

10

20

30

[PDZ] (µM)

40

50

60

Fig. 3. Binding kinetics of nNOS PDZ variants with point mutations in the β-finger. Association kinetics were analyzed by plotting kobs versus (a) nNOS PDZ or (b) PSD95 PDZ2 concentration and fitting Eq. (6) to the data. In (c), data for association kinetics between nNOS PDZ R121A and PSD-95 PDZ2 I195W are shown. Both species were varied in separate experiments, while the other one was kept constant. Inset, slow phases detected for G115A/ V122W and I120G/V122W, respectively.

Coupled Folding and Binding for a β-Finger

257

A = DDG‡ = DDGEq

off

nNOS PDZ varied PSD-95 PDZ2 varied

y = -0.46582 + 0.92338x R= 0.9878

1

0.5

0

-0.5

-1

-1.5 -1

-0.5

0

0.5

1

1.5

2

log K /µM d

log k

(b)

log k

on on

nNOS PDZ varied PSD-95 PDZ2 varied

1 y = -0.56044 - 0.054x R= 0.16998 y = -0.46477 - 0.078006x R= 0.47463

0.5

ð1Þ

on

ð2Þ

-1

0




and DDGEq = RTln Kdmut = KdWT

off

y = -0.56118 + 0.94656x R= 0.94936

log k /s

DDG = RTln

WT mut kon = kon

log k

1.5

where ‡

log k

(a)

-1

We then measured kinetics for the β-finger mutants, either by varying wild-type or mutant nNOS PDZ (Fig. 3a) or wild-type PSD-95 PDZ2 (Fig. 3b) or in displacement experiments to get koff (Table 1). Most mutants displayed similar binding kinetics as the wild type, but with slightly decreased kon and increased or decreased koff values. However, the kobs values for one of the β-finger mutants, R121A, did indeed display a nonlinear dependence, a decrease of kobs on increasing ligand concentration (Fig. 3c). Such behavior is indicative of a slow conformational change before (fast) binding. 32 This R121A mutation apparently changes the pathway of binding. Interestingly, the minor phase (kobs2) detected for the mutants G115A/V122W and I120G/V122W also displayed a similar behavior (Fig. 3c, inset), suggesting that a small population of these mutants populate a conformation that is not binding competent, possibly similar to that for R121A. To get a picture of the transition state for the binding reaction, we calculated Φ values for binding 33,34 for each mutated position (Table 1). In such analysis, the change in free energy for the binding kinetics upon mutation (ΔΔG ‡) is related to the change in free energy for the overall binding reaction (ΔΔGEq).

off

Binding kinetics for the mutants

constant (Fig. 4). This correlation suggests that the conformations that were destabilized (or interactions broken) by mutations in the β-finger form late and cooperatively in the binding reaction. The data set for wild-type nNOS PDZ

log k /s

trapped PSD-95 PDZ2 (see Materials and Methods). The dissociation rate constant was decreased by the V122W mutation, some 2-fold (Table 1).




ð3Þ

A Φ value of 1 thus means that the change in Kd is equal to the change in kon. The implication is that the native interaction(s), which is deleted upon mutation, is present already in the transition state of the binding reaction. On the other hand, a Φ value of 0 means that the change in Kd is equal to the change in koff and that the native interaction(s) of the mutated residue forms after the rate-limiting step. The destabilization of the protein complex upon mutation, ΔΔGEq, was too low 35,36 to calculate accurate Φ values for each position (Table 1). However, a Brønsted or Leffler type of plot 34,37 clearly demonstrates that changes in Kd are determined mainly by changes in the dissociation rate constant and not in the association rate

-0.5

-1

-1.5

-2 -1

-0.5

0

0.5

1

1.5

2

log K /µM d

Fig. 4. Linear free-energy relationships of the rate and equilibrium constants of binding. (a) Plot of log koff versus log Kd for data where either PSD-95 PDZ2 I195W was held constant and nNOS PDZ was varied or nNOS PDZ V122W was held constant and PSD-95 PDZ2 was varied. Both approaches yielded a similar result, namely, that the variation in affinity, Kd, is governed by N90% by koff, upon mutational modulation of stability. (b) Similar plot as in (a), but with log kon versus log Kd.

258 agreed very well with that of the V122W mutant, both for koff (Fig. 4a) and for kon (Fig. 4b). (The data for R121A were not included since the ratelimiting transition state is different for this mutant.) Nonetheless, Φ values for some positions could be calculated (Table 1). Both G113A and G115A in the turn display Φ values close to 0, or even negative in the case of G115A, which shows that their native interactions in the β-turn are not formed in the rate-limiting transition state. Note that the G115A mutation results in stabilization of the bimolecular complex through a decrease in koff. L107A, on the other hand, had a positive Φ value significantly different from 0, 0.3–0.4.

Discussion Our mutational and kinetic analysis of the β-finger of nNOS PDZ highlights two mechanistic aspects that may be general for the binding reactions of flexible or intrinsically disordered protein regions. Firstly, the linear free-energy diagrams (Fig. 4) show that native interactions/conformations in the bimolecular complex form cooperatively after the main transition state. Secondly, one mutation (R121A) modulated the pathway of the reaction such that a first-order step (a conformational change) was rate limiting for the overall reaction. This switch shows that the energy landscape is fairly “plastic”, for binding reactions of disordered protein regions. These two conclusions are discussed in more detail below. The linear free-energy relationships in Fig. 4 clearly show that changes in affinity (Kd) on mutation are due to changes in koff and not in kon. This relationship suggests that native interactions form after the rate-limiting transition state in analogy with folding Φ values. 38 The linear dependence further suggests that mutational destabilization has a similar effect at different positions along the β-finger. In other words, the native interactions are formed in an all or none fashion, that is, cooperatively (cf. Brønsted plots from protein folding studies, e.g., for the classical twostate folder CI2 39). This finding is similar to the results we obtained with a simpler model system for intrinsically disordered proteins, namely, disordered peptides. 26 In the current study, we have a larger set of point mutations in a larger flexible region, as compared with the disordered peptides. In addition, the mutations in the present study target residues outside of the canonical binding pocket, with the aim to modulate the stability of the β-finger while retaining the interaction surface, whereas our previous study looked at the sidechain interactions formed within the binding pocket. Data for all mutants of the β-finger fall close to the fitted line in a log koff versus log Kd plot

Coupled Folding and Binding for a β-Finger

(R 2 = 0.90 and 0.97 for our two different data sets, respectively). The fact that we observe a similar behavior with this flexible β-finger as with peptides suggests that the result is general, at least upon formation of β-sheets. The binding Φ value for L107A was around 0.3–0.4 (Table 1), suggesting that its side chain may be the nucleus for the cooperative folding and binding of the rest of the β-finger. The side chain of L107 points into the core of nNOS PDZ and acts as an anchor for the β-finger. The β-turn of the finger allows binding to the carboxylate binding loop of PSD-95 PDZ2, thus mimicking the C-terminus of peptide ligands in canonical PDZ–peptide interactions (Fig. 1). This turn has not formed in the transition state, as shown by low and negative binding Φ values for G113A and G115A, respectively. From the data, we may speculate that the binding reaction involves a weak pre-complex: the kinetics show that a second-order event, binding, giving a bimolecular complex, is rate limiting under the experimental conditions; yet, several native interactions/conformations form after the main barrier along the reaction coordinate. A pre-complex has been observed directly by NMR for one of the previously studied reactions of intrinsically disordered proteins, that of pKID and KIX. 8 Our data are also in-line with a recent study 34 (as well as earlier studies 40–42) on the S-peptide from ribonuclease S binding to the S-protein, where the coupled folding and binding reaction displays low Φ values. Like for the S-peptide/S-protein interaction, direct kinetic evidence for an intermediate in the form of a hyperbolic dependence of kobs on ligand concentration was found for the interaction between the disordered protein WASp and the ordered Cdc42. 43 This association was governed by electrostatic steering, involving charged groups that interacted in the transition state but did not form salt bridges in the bimolecular ground-state complex, suggesting a pre-complex with nonnative interactions. Nonnative interactions were also found to be important by computer simulations for the HIF1α/CBP interaction 44 and that between IA3 and YPrA. 45 In a previous study, the R121Q mutation in nNOS PDZ resulted in a severely reduced affinity such that no binding was detected by surface plasmon resonance. 21 In our study, we observed binding between an R121A mutant and PSD-95 PDZ2 using stopped-flow fluorimetry. We could estimate from kinetic amplitudes that the Kd for the interaction is N 30 μM (data not shown), and it is likely that the concentration of R121Q used in the surface plasmon resonance experiment was too low to give a measurable signal. The mutation used in this study, R121A, switched the mechanism such that an initial conformational change was an obligatory step before association with PSD-95 PDZ2. To accommodate a more disordered β-finger in its

Coupled Folding and Binding for a β-Finger

binding pocket, we cannot, however, rule out that the slow conformational change occurs in PSD-95 PDZ2, but we find this scenario less likely. Note that only one of the expected two kinetic phases of a two-step mechanism was observed for the R121A mutation. However, this is common, in particular, when a slow conformational change is followed by a faster binding event. The observed slow firstorder step for R121A does not rule out any subsequent induced-fit step. However, looking at the major transition state, the mechanism has changed to one involving a slow conformational transition. The fact that less severe mutations (G115A and I120G) also shift the equilibrium such that a slow phase appears in binding kinetics with the V122W pseudo wild type demonstrates that, in general, conformational equilibria may govern the association kinetics for intrinsically disordered proteins. The simplest explanation of our data is that the β-finger of nNOS PDZ may exist in highenergy conformations. Upon mutation of R121 to Ala, the predominant binding-competent “natively flexible” conformation is destabilized to a “mutationally disordered” high-energy species. For the mutants G115A/V122W and I120G/V122W, the natively flexible binding-competent conformation is still the predominant species, but for R121A, the equilibrium has shifted such that the majority of molecules must undergo a slow conformational transition before association with PSD-95 PDZ2. Thus, we probably trapped a high-energy offpathway intermediate by mutation, again demonstrating the plasticity of the energetic landscape for folding and binding of PDZ domains. 46

Materials and Methods Expression and purification of mutants PSD-95 PDZ2 (residues 155–249 of human PSD-95) and nNOS PDZ (residues 1–132 of human nNOS) were previously expressed and purified in our laboratory as part of a protein folding study. 27 In the present study, the proteins were purified in essentially the same way. In short, the cDNA for the respective PDZ domains, cloned into the pRSET A expression vector, were transformed into Escherichia coli BL21 DE3 plysS. The cells were grown to an optical density of 0.6–1 at 37 °C in 2TY medium. Protein expression was induced with 1 mM isopropyl-βthiogalactopyranoside, and the cells were grown overnight at 30 °C (25 °C for the R121A mutant). Following centrifugation, the pellets containing expressed PDZ domains were sonicated, centrifuged, filtered and purified first using nickel affinity chromatography (by an N-terminal His-tag), followed by anion-exchange chromatography in 50 mM Tris–HCl (pH 8.5). The PDZ variants were collected as the flow through and were pure as judged from SDS-PAGE. The identity of the PDZ wild types and mutants were confirmed with matrix-assisted

259 laser desorption/ionization–time of flight mass spectrometry, and their concentrations were determined by calculated extinction coefficients based on amino acid sequence. The mutants were folded as judged by far-UV circular dichroism and urea denaturation experiments. Binding kinetics PSD-95 PDZ2 and nNOS PDZ binding kinetics were performed on an SX-20MV stopped-flow spectrometer (Applied Photophysics, Leatherhead, UK) at 25 °C in 50 mM potassium phosphate (pH 7.5). Fluorescence was monitored using the increase in tryptophan emission (excitation at 280 nm; emission at 330 ± 30 nm). In order to determine the rate constants for the nNOS PDZ/PSD-95 PDZ2 interaction, we varied the PSD-95 PDZ2 or nNOS PDZ concentration at constant concentration of the respective interaction partner, nNOS V122W (5 μM) or PSD-95 PDZ2 I195W (5 μM), respectively. These Trp residues were introduced to obtain large fluorescence changes upon binding, and they did not significantly affect the binding kinetics as shown in Fig. 2b where the slopes of the curves, the association rate constants, are similar for the two pseudo-wild-type PDZs. Further, since we make our conclusions from comparison with the respective pseudo-wild-type PDZ, any absolute changes related to the Trp residues will cancel out. Kinetic traces from time-resolved nNOS PDZ/PSD-95 PDZ2 binding experiments were fitted to single- or double-exponential functions [Eqs. (4) and (5)].
A = DAEq 1 − e − kobs t + C ð4Þ

A = DAEq 1 − e − kobs1 t + DBEq 1 − e − kobs2 t + C

ð5Þ

where A is the signal recorded with the time t; ΔAEq and ΔBEq are the amplitudes of the respective phase; and kobs is the observed rate constants. The kobs values were plotted versus PSD-95 PDZ2 or nNOS PDZ concentration and fitted to the general equation for reversible association of two molecules [Eq. (6)]. 47 
2
0:5 2 2 kobs = kon n− ½A0 + koff + 2kon koff n + ½A0 ð6Þ kon is the association or on-rate constant, koff is the dissociation or off-rate constant and [A]0 and n are the initial concentrations of the varied and constants species, respectively. Displacement kinetics To determine the dissociation rate constant koff, the nNOS PDZ/PSD-95 PDZ2 complex was premixed in one syringe in the stopped flow (5 μM V122W with 5 μM PSD95 PDZ2 or 3 μM wild-type nNOS PDZ with 1.5 μM PSD95 PDZ2 I195W). The respective complex was rapidly mixed (1:1) with a peptide ligand for PSD-95 PDZ2 (unlabeled YKQTSV or Dansyl-YKQTSV, respectively). At high concentrations of this peptide, any dissociated PSD95 PDZ2 will be trapped, resulting in a virtually irreversible reaction, limited by the rate constant for

Coupled Folding and Binding for a β-Finger

260 dissociation of the premixed complex. For the nNOS PDZ V122W/PSD-95 PDZ2 complex, the non-dansylated YKQTSV was used, and Trp fluorescence was used to monitor the dissociation. For the wild-type nNOS PDZ/ PSD-95 PDZ2 I195W complex, dansylated peptide was used, and the change in Trp (330-nm band-pass filter) or dansyl (420-nm-cutoff filter) fluorescence was used to monitor the reaction. In all cases, the experimental traces were monophasic. The observed rate constant at high peptide is equal to the koff for the respective bimolecular complex. 28,29

12.

13.

14. 15.

Acknowledgements

16.

This work was supported by the Swedish Research Council (grant 2009-5659) and the Human Frontiers Young Investigator Science Program. We thank Dr. Jakob Dogan for insightful comments on the manuscript.

17.

References

18.

1. Dyson, H. J. & Wright, P. E. (2005). Intrinsically unstructured proteins and their functions. Nat. Rev., Mol. Cell Biol. 6, 197–208. 2. Dunker, A. K., Silman, I., Uversky, V. N. & Sussman, J. L. (2008). Function and structure of inherently disordered proteins. Curr. Opin. Struct. Biol. 18, 756–764. 3. Uversky, V. N. & Dunker, A. K. (2010). Understanding protein non-folding. Biochim. Biophys. Acta, 1804, 1231–1264. 4. Vuzman, D. & Levy, Y. (2012). Intrinsically disordered regions as affinity tuners in protein–DNA interactions. Mol. Biosyst. 8, 47–57. 5. Wright, P. E. & Dyson, H. J. (2009). Linking folding and binding. Curr. Opin. Struct. Biol. 19, 31–38. 6. Bernado, P. & Svergun, D. I. (2012). Structural analysis of intrinsically disordered proteins by small-angle X-ray scattering. Mol. Biosyst. 8, 151–167. 7. Radhakrishnan, I., Perez-Alvarado, G. C., Parker, D., Dyson, H. J., Montminy, M. R. & Wright, P. E. (1997). Solution structure of the KIX domain of CBP bound to the transactivation domain of CREB: a model for activator:coactivator interactions. Cell, 91, 741–752. 8. Sugase, K., Dyson, H. J. & Wright, P. E. (2007). Mechanism of coupled folding and binding of an intrinsically disordered protein. Nature, 447, 1021–1025. 9. Kjaergaard, M., Teilum, K. & Poulsen, F. M. (2010). Conformational selection in the molten globule state of the nuclear coactivator binding domain of CBP. Proc. Natl Acad. Sci. USA, 107, 12535–12540. 10. Tompa, P. (2005). The interplay between structure and function in intrinsically unstructured proteins. FEBS Lett. 579, 3346–3354. 11. Hilser, V. J. & Thompson, E. B. (2007). Intrinsic disorder as a mechanism to optimize allosteric

19. 20.

21.

22.

23.

24.

25. 26.

27.

coupling in proteins. Proc. Natl Acad. Sci. USA, 104, 8311–8315. Tompa, P., Prilusky, J., Silman, I. & Sussman, J. L. (2008). Structural disorder serves as a weak signal for intracellular protein degradation. Proteins, 71, 903–909. Shoemaker, B. A., Portman, J. J. & Wolynes, P. G. (2000). Speeding molecular recognition by using the folding funnel: the fly-casting mechanism. Proc. Natl Acad. Sci. USA, 97, 8868–8873. Trizac, E., Levy, Y. & Wolynes, P. G. (2010). Capillarity theory for the fly-casting mechanism. Proc. Natl Acad. Sci. USA, 107, 2746–2750. Zhou, H. X. (2012). Intrinsic disorder: signaling via highly specific but short-lived association. Trends Biochem. Sci. 37, 43–48. Chemes, L. B., Sanchez, I. E. & de Prat-Gay, G. (2011). Kinetic recognition of the retinoblastoma tumor suppressor by a specific protein target. J. Mol. Biol. 412, 267–284. Narayanan, R., Ganesh, O. K., Edison, A. S. & Hagen, S. J. (2008). Kinetics of folding and binding of an intrinsically disordered protein: the inhibitor of yeast aspartic proteinase YPrA. J. Am. Chem. Soc. 130, 11477–11485. Wang, C. K., Pan, L., Chen, J. & Zhang, M. (2010). Extensions of PDZ domains as important structural and functional elements. Protein Cell, 1, 737–751. Petit, C. M., Zhang, J., Sapienza, P. J., Fuentes, E. J. & Lee, A. L. (2009). Hidden dynamic allostery in a PDZ domain. Proc. Natl Acad. Sci. USA, 106, 18249–18254. Zhang, J., Petit, C. M., King, D. S. & Lee, A. L. (2011). Phosphorylation of a PDZ domain extension modulates binding affinity and interdomain interactions in the PSD-95 MAGUK. J. Biol. Chem. 286, 41776–41785. Tochio, H., Mok, Y. K., Zhang, Q., Kan, H. M., Bredt, D. S. & Zhang, M. (2000). Formation of nNOS/PSD-95 PDZ dimer requires a preformed β-finger structure from the nNOS PDZ domain. J. Mol. Biol. 303, 359–370. Tochio, H., Zhang, Q., Mandal, P., Li, M. & 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. Wang, P., Zhang, Q., Tochio, H., Fan, J. S. & Zhang, M. (2000). Formation of a native-like β-hairpin finger structure of a peptide from the extended PDZ domain of neuronal nitric oxide synthase in aqueous solution. Eur. J. Biochem. 267, 3116–3122. Hillier, B. J., Christopherson, K. S., Prehoda, K. E., Bredt, D. S. & Lim, W. A. (1999). Unexpected modes of PDZ domain scaffolding revealed by structure of nNOS–syntrophin complex. Science, 284, 812–815. DeLano, W. L. (2002). The PyMOL Molecular Graphics System. DeLano Scientific, San Carlos, CA. Haq, S. R., Chi, C. N., Bach, A., Dogan, J., Engström, Å., Hultqvist, G. et al. (2012). Side-chain interactions form late and cooperatively in the binding reaction between disordered peptides and PDZ domains. J. Am. Chem. Soc. 134, 599–605. Chi, C. N., Gianni, S., Calosci, N., TravagliniAllocatelli, C., Engström, Å. & Jemth, P. (2007). A conserved folding mechanism for PDZ domains. FEBS Lett. 581, 1109–1113.

Coupled Folding and Binding for a β-Finger 28. Chi, C. N., Bach, A., Gottschalk, M., Kristensen, A. S., Strømgaard, K. & 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. J. Biol. Chem. 285, 28252–28260. 29. Chi, C. N., Bach, A., Engström, Å., Wang, H., Strømgaard, K., Gianni, S. & Jemth, P. (2009). A sequential binding mechanism in a PDZ domain. Biochemistry, 48, 7089–7097. 30. Gianni, S., Walma, T., Arcovito, A., Calosci, N., Bellelli, A., Engström, Å. et al. (2006). Demonstration of long-range interactions in a PDZ domain by NMR, kinetics and protein engineering. Structure, 14, 1801–1809. 31. Schreiber, G. & Fersht, A. R. (1995). Energetics of protein–protein interactions: analysis of the barnase– barstar interface by single mutations and double mutant cycles. J. Mol. Biol. 248, 478–486. 32. Fersht, A. (1999). Structure and Mechanism in Protein Science. W. H. Freeman and Co., New York, NY. 33. Schreiber, G., Haran, G. & Zhou, H. X. (2009). Fundamental aspects of protein–protein association kinetics. Chem. Rev. 109, 839–860. 34. Bachmann, A., Wildemann, D., Praetorius, F., Fischer, G. & Kiefhaber, T. (2011). Mapping backbone and side-chain interactions in the transition state of a coupled protein folding and binding reaction. Proc. Natl Acad. Sci. USA, 108, 3952–3957. 35. Fersht, A. R. & Sato, S. (2004). ϕ-Value analysis and the nature of protein-folding transition states. Proc. Natl Acad. Sci. USA, 101, 7976–7981. 36. Sanchez, I. E. & Kiefhaber, T. (2003). Origin of unusual ϕ-values in protein folding: evidence against specific nucleation sites. J. Mol. Biol. 334, 1077–1085. 37. Leffler, J. E. (1953). Parameters for the description of transition states. Science, 117, 340–341. 38. Fersht, A. R., Matouschek, A. & Serrano, L. (1992). The folding of an enzyme. I. Theory of protein engineering

261

39.

40. 41.

42.

43.

44.

45.

46.

47.

analysis of stability and pathway of protein folding. J. Mol. Biol. 224, 771–782. Itzhaki, L. S., Otzen, D. E. & 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. Goldberg, J. M. & Baldwin, R. L. (1998). Kinetic mechanism of a partial folding reaction. 2. Nature of the transition state. Biochemistry, 37, 2556–2563. Goldberg, J. M. & Baldwin, R. L. (1998). Kinetic mechanism of a partial folding reaction. 1. Properties of the reaction and effects of denaturants. Biochemistry, 37, 2546–2555. Goldberg, J. M. & Baldwin, R. L. (1999). A specific transition state for S-peptide combining with folded Sprotein and then refolding. Proc. Natl Acad. Sci. USA, 96, 2019–2024. Hemsath, L., Dvorsky, R., Fiegen, D., Carlier, M. F. & Ahmadian, M. R. (2005). An electrostatic steering mechanism of Cdc42 recognition by Wiskott–Aldrich syndrome proteins. Mol. Cell, 20, 313–324. De Sancho, D. & Best, R. B. (2012). Modulation of an IDP binding mechanism and rates by helix propensity and non-native interactions: association of HIF1α with CBP. Mol. Biosyst. 8, 256–267. Wang, J., Wang, Y., Chu, X., Hagen, S. J., Han, W. & Wang, E. (2011). Multi-scaled explorations of bindinginduced folding of intrinsically disordered protein inhibitor IA3 to its target enzyme. PLoS Comput. Biol. 7, e1001118. Haq, S. R., Jürgens, M. C., Chi, C. N., Koh, C. S., Elfström, L., Selmer, M. et al. (2010). The plastic energy landscape of protein folding: a triangular folding mechanism with an equilibrium intermediate for a small protein domain. J. Biol. Chem. 285, 18051–18059. Malatesta, F. (2005). The study of bimolecular reactions under non-pseudo-first order conditions. Biophys. Chem. 116, 251–256.