Detection of a Hidden Folding Intermediate of the Third Domain of PDZ Hanqiao Feng†, Ngoc-Diep Vu† and Yawen Bai* Laboratory of Biochemistry National Cancer Institute NIH, Building 37, Room 6114E Bethesda, MD 20892, USA The folding pathway of the third domain of PDZ from the synaptic protein PSD-95 was characterized using kinetic and equilibrium methods by monitoring the fluorescence signal from a Trp residue that is incorporated at a near-surface position. Kinetic folding of this domain showed multiple exponential phases, whereas unfolding showed a single exponential phase. The slow kinetic phases were attributed to isomerization of proline residues, since there are five proline residues in this domain. We found that the logarithms of the rate constants for the fast phase of folding and unfolding are linearly dependent on the concentrations of denaturant. The unfolding free energy derived from these rate constants at zero denaturant was close to the value measured using the equilibrium method, suggesting the absence of detectable sub-millisecond folding intermediates. However, native-state hydrogen exchange experiments detected a partially unfolded intermediate under native conditions. It was further confirmed by a protein engineering study. These data suggest that a hidden intermediate exists after the rate-limiting step in the folding of the third domain of PDZ. Published by Elsevier Ltd. Keywords: protein folding; hydrogen exchange; PDZ domain; hidden intermediate *Corresponding author Introduction To understand how proteins fold, it is necessary to characterize the folding free energy landscape in detail, including unfolded (U), intermediate (I), and native (N) states. Since the structure of native state can be determined with high resolution using NMR and X-ray crystallography, characterization of other states becomes the major focus of protein folding studies. Among these states, folding intermediates are not observable for all proteins. While partially unfolded intermediates are found commonly in larger proteins in traditional kinetic folding experi- ments, 1,2 they are generally not detectable in smaller proteins (!120 amino acid residues). 3,4 In the few cases in which several groups have claimed to detect early intermediates in the sub-millisecond range in the folding of small proteins, stabilizing salt (0.4 M Na 2 SO 4 ) has been used. 5–8 Even for such cases, the conclusions still appear to be questionable. 4 In addition, no structure of such an early folding intermediate has been characterized in sufficient detail. Although early folding intermediates are not detectable in the kinetic folding experiments for small proteins in the absence of a stabilizing reagent, in several cases, including cytochrome c, 9 Rd-apocyto- chrome b 562 , 10–12 and barnase, 13 hidden intermediates after rate-limiting transition states have been identi- fied using a native-state hydrogen exchange (HX) method. 9,14,15 Moreover, a high-resolution structure of the hidden intermediates of Rd-apocytochrome b 562 has been determined. 12 PDZ-3 is a small modular interaction domain that binds in a sequence-specific fashion to a short C-terminal peptide of K C channel subunits. 16 This small protein module has a significant portion of b-strands in the native structure (see Figure 1). So far, no folding study has been performed on PDZ-3. Here, we have characterized the folding pathway of a PDZ-3 construct with 115 amino acid residues using a number of experimental tools, including stopped-flow fluorescence, native-state hydrogen exchange, and protein engineering. The results suggest that a partially unfolded intermediate exists after the rate-limiting step in the folding of this small protein module. 0022-2836/$ - see front matter Published by Elsevier Ltd. † H.F. & N.-D.V. contributed equally to this work. Abbreviations used: U, unfolded; I, intermediate; N, native; HX, hydrogen exchange; GdmCl, guanidinium chloride; WT, wild-type; HSQC, heteronuclear single quantum coherence. E-mail address of the corresponding author: yawen@helix.nih.gov doi:10.1016/j.jmb.2004.11.040 J. Mol. Biol. (2005) 346, 345–353