• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. May 29, 2007; 104(22): 9272–9277.
Published online May 18, 2007. doi:  10.1073/pnas.0703434104
PMCID: PMC1890484

The helix–turn–helix motif as an ultrafast independently folding domain: The pathway of folding of Engrailed homeodomain


Helices 2 and 3 of Engrailed homeodomain (EnHD) form a helix–turn–helix (HTH) motif. This common motif is believed not to fold independently, which is the characteristic feature of a motif rather than a domain. But we found that the EnHD HTH motif is monomeric and folded in solution, having essentially the same structure as in full-length protein. It had a sigmoidal thermal denaturation transition. Both native backbone and local tertiary interactions were formed concurrently at 4 × 105 s−1 at 25°C, monitored by IR and fluorescence T-jump kinetics, respectively, the same rate constant as for the fast phase in the folding of EnHD. The HTH motif, thus, is an ultrafast-folding, natural protein domain. Its independent stability and appropriate folding kinetics account for the stepwise folding of EnHD, satisfy fully the criteria for an on-pathway intermediate, and explain the changes in mechanism of folding across the homeodomain family. Experiments on mutated and engineered fragments of the parent protein with different probes allowed the assignment of the observed kinetic phases to specific events to show that EnHD is not an example of one-state downhill folding.

Keywords: barrier-limited, NMR, protein, temperature jump

Protein structural motifs are simple combinations of a few secondary structure elements with a specific geometric arrangement that occur frequently in protein structures. The simplest motif with a specific function consists of two α-helices connected by a turn region, the helix–turn–helix (HTH) motif (1). Several motifs often combine to form compact, globular structures, which are called domains (1, 2). There can be an obvious difference in polypeptide chain size between a motif and a domain, but, more importantly, a motif is believed not to be able to fold to a stable structure, whereas a domain should.

The HTH motifs involved in DNA binding are composed of “stabilization” and a “recognition” helixes, connected by a sharp turn that typically does not tolerate insertions or distortions (36). They have high structural similarity (7) and are very readily identifiable (8). They also are found in DNA repair and replication and RNA metabolism proteins, in catalytic domains of diverse enzymes, and involved in protein–protein interactions. Monomeric HTH motifs are unknown. The simplest proteins with an HTH motif are the dimeric Arc repressor (9, 10) and the three-helix bundle homeodomains (4). The HTH motif thus is regarded as a motif rather than a domain.

Homeodomains are small, conserved three-helix bundle proteins of ≈60 residues (1, 11). They have a common fold where helix 1 (H1) (residues 10–22) and helix 2 (H2) (residues 28–37) lie antiparallel to each other with helix 3 (H3) lying across both (residues 42–56). The Engrailed homeodomain (EnHD) has biphasic folding kinetics: formation of an intermediate in microseconds followed by 10-fold slower folding to the native state (1218). The mutation L16A reduces the stability of the native state, so that a highly helical compact intermediate/denatured state is populated (17). L16A has an HTH motif but loses its interactions with H1. H1 has exceptionally high helical propensity as an independent sequence (12).

Here, we investigated the equilibrium and kinetic properties of folding of the 16–59 fragment of EnHD, which lacks the N terminus and half of H1, with the loop region and the HTH motif being intact. We found that this fragment satisfies the properties commonly expected from a “self-folding domain,” such as multiple secondary structure elements in close contact, notable chemical shift dispersion, some degree of folding cooperativity, and resistance to thermal unfolding (19). Kinetic measurements on the folding of the HTH motif allowed us to assign with confidence the previous tentative assignment of the very fast phase in the folding of EnHD (14, 17).


The Fragment Is Monomeric in Solution.

Both NMR 1D 1H T2 measurements and analytical ultracentrifugation at 25°C indicated that the fragment was monomeric. A value of 75 ms for T2 at 25°C obtained from the 1D 1H spin-echo experiment (20) on the samples used for structure determination is consistent with Mr ≈ 5,500. Sedimentation equilibration gave an Mr of 5,600, which is within error for a monomeric protein, and with no systematic deviations in the fitted residuals.

NMR Chemical Shifts of Residues 24–59 Similar to Those in L16A.

The 1H-15N heteronuclear single quantum correlation (HSQC) spectrum was similar to that of L16A, which was confirmed by full reassignment of the fragment (Fig. 1). There was a significant systematic difference between backbone chemical shifts (and Cβ) of the fragment and full-length protein for residues <24, consistent with loss of helicity in the fragment for these residues [residue numbering is the same as in the full-length protein, Protein Data Bank (PDB) ID codes 1ENH (21) or 1ZTR (17)]. The chemical shifts for residues ≥24 were very alike for both proteins, with the largest difference, 0.6 ppm, occurring for R31 Cα. There were similar small changes in the 1H-13C HSQC spectrum, with most of the peaks being close to their counterparts in the full-length protein. Overall, these chemical shift changes were insignificant, and it is unlikely that the hydrophobic core of the protein changed, which is consistent with the NOESY spectrum.

Fig. 1.
Comparison of chemical shifts between EnHD L16A and its 16–59 fragment. (a) The 1H-15N HSQC spectra of EnHD residues 16–59 (red) overlayed with that of L16A (blue). Peaks for L16A residues up to 15 are shown as black contours. (b) The ...

Because the chemical shift changes between the full-length protein and the fragment were small, it is unlikely that H1 interacts with H2/H3 in the L16A mutant, as previously suggested by lack of NOEs between H1 and the HTH motif (17). We note that the NMR experiments for L16A were under conditions where it was still possible to populate the native state, albeit with a population of <5% (17). The overall changes in chemical shifts were very small and not clustered at any specific location within H2/H3. Accordingly, there is no evidence that H1 in L16A interacts noncovalently with H2/H3.

Backbone Structure of the HTH Motif Is the Same as in EnHD Crystal Structure.

Comparison of the values of the determined HNAn external file that holds a picture, illustration, etc.
Object name is cjs0807.jpgN residual dipolar couplings (RDCs) with the corresponding RDCs predicted from the crystal structure (PDB ID code 1ENH) is an exquisite test of structure (22, 23). Typically, in high-resolution crystal structures, the Cornilescu Q < 0.25 (24), indicating that the crystal structure is an accurate representation of the average HNAn external file that holds a picture, illustration, etc.
Object name is cjs0807.jpgN bond vector orientation within the molecular frame (and, by implication, of the protein backbone) in solution (22). The longest stretch of residues that has Q < 0.25 (Pearson's correlation coefficient of 0.95) is residues 28–52 (25 measured RDCs), showing that the HTH motif is preserved in the 16–59 fragment of the protein. Because the HNAn external file that holds a picture, illustration, etc.
Object name is cjs0807.jpgN bond vector orientations from the crystal structure are consistent with the solution conformation of the HTH motif in the fragment. they were constrained in this orientation by using [var phi] /ψ dihedral angles from the crystal structure during the CNS structure calculation.

Structure of the HTH Motif.

The structure of the 16–59 fragment (PDB ID code 2P81) was determined by using ≈700 intraresidue NOEs and 50 dihedral angle restraints obtained from the crystal structures. It has a typical HTH motif arrangement of helices with the two helices [for consistency with the previous homeodomain work, called H2 (stabilization) and H3 (recognition)] fraying at the ends. As in the structure of L16A, the solvent-exposed aromatics in H3 bend back toward the leucines in H2 and the turn region, protecting the small hydrophobic core of the protein (Fig. 2).

Fig. 2.
Structure of the HTH motif and summary of sequential NOEs and secondary chemical shift values. (a) Lowest-energy structure of HTH motif. (b) Bundle of the 25 lowest-energy conformers with backbone atoms N, Ca, and C of residues 28–53 superimposed ...

The protein structure is well defined with a backbone root mean square deviation (rmsd) of 0.34 Å for residues 28–53 and an rmsd relative to the crystal structure of 1.3 Å for residues 28–52 [supporting information (SI) Table 3]. The apparent slight bend in the stabilization helix might be just a structure calculation artifact because the measured RDCs suggest that the backbone structure for residues 28–52 should be the same as that of the WT protein. The agreement of the predicted RDCs with the determined ones yielded Q = 0.28 (r = 0.94) for residues 28–53 and Q = 0.34 (r = 0.92) for residues 24–53. The former agreement is almost as good as that obtained for the WT protein, but the backbone dihedral angles were constrained to resemble those of the WT protein. The agreement for the structured part of the protein (residues 24–53) is better than in the corresponding fragment from the crystal structure of PDB ID code 1ENH (Q = 0.54, r = 0.87).

We know of no structure calculation methods that would allow for the separation of NOE restraints coming from the (possible) multiple conformations of the fragment when no significant restrain violations occur during the calculation (25). However, it is possible to satisfy all of the NOEs (≈700 intraresidue, ≈16 NOEs per residue) with a single structure that has the same backbone dihedral angles as the WT protein (±1°) for residues 28–53. The structure had the same backbone fold and arrangement of the side chains when no crystal-structure-derived dihedral angle restraints were used, when only TALOS dihedral angle restraints were incorporated, or when ensemble averaging (25, 26) was used.

Good Dispersion of Chemical Shifts and Preservation of the HTH Motif Structure at High Ionic Strength.

The stability of EnHD increases at high ionic strength, which screens unfavorable electrostatic interactions in the HTH motif (17). The stability of fragment HTH motif conformation should behave similarly. There were significant changes in the methyl region of the 1H spectrum (Fig. 3) with increasing ionic strength at 150 mM NaCl, in which the most upfield-shifted resonance of L34 Hδ had a chemical shift of 0.58 ppm, and 2 M NaCl, in which the most upfield-shifted resonance had a chemical shift of 0.28 ppm. Such a wide dispersion of chemical shifts in the methyl region is indicative of a folded protein (27).

Fig. 3.
The influence of increasing ionic strength on the chemical shifts of the fragment. (a) Proton 1D spectra of EnHD residues 16–59 at increasing salt concentrations at 5°C. The spectra were acquired in 50 mM acetate buffer, pH 5.7, and the ...

There were significant peak movements in 1H-15N HSQC spectra at the higher salt concentrations. On average, the amide peaks moved upfield, which is consistent with an increase in helical content on salt addition. The most upfield-shifted amide group at ≈2 M NaCl had a chemical shift of 7.35 ppm, which is almost as large as the most upfield-shifted L40 HN (7.29 ppm) in the WT protein.

We assigned the Hα chemical shifts at 500 mM NaCl (50 mM acetate, pH 5.7) to observe the trend in the secondary structure or population change upon the addition of salt (see Fig. 3). Consistent with the CD measurements, the Hα secondary shifts became more negative at high salt for residues in H2 and H3. For the residues in the turn region (residues 38–41), the Hα chemical shifts became more turn-like. Concurrently, the helical secondary shifts present for residues 16–27 were shifted toward the random coil values, although most were within experimental error (≈0.005 ppm). Notably, the significant change in the Hα chemical shift of T27, which participates in the N-capping of H2 in the WT structure, and also Y25 might indicate a decrease in the population of helical structure for this chameleon sequence (28). The lack of increase of helical content for the residues in WT H1 implied that, as expected, the homeodomain-like native state did not become populated.

The 2D NOESY spectra at 100 and 500 mM NaCl had peaks in similar places. Because of the higher dispersion of chemical shifts, a number of them could be unambiguously assigned. In particular, the nonnative long-range NOEs between the Leu methyl groups and W48 Hε1 still were present as well as the nonnative NOEs between the T27 Hγ2 and the aromatic ring of F49, implying no change in structure upon increase in ionic strength.

Sigmoid Folding Transition at High Ionic Strength.

Because of broad unfolding transitions and marginal stability of early folding intermediates, it is impossible, in most cases, to distinguish between early folding intermediates and denatured states of varying compaction (17). Like EnHD L16A, the 16–59 fragment had a folding transition that lacked the native baseline upon thermal denaturation under physiological ionic strength (Fig. 3). But at higher ionic strength, there was a sigmoid curve, although the transition was much broader than that of WT protein.

The Tm and ΔHD–N at the highest concentration of NaCl, which had the best baselines (see Fig. 4), were 45 ± 4°C and 14 ± 1 kcal·mol−1, respectively, by using a free fit to the standard two-state equation (29) and ΔCp of 350 cal·mol−1·K−1. In 2 M [NaCl] at 25°C, the HTH motif is ≈75% folded, and at 5°C, it is ≈85% folded. The Tm at physiological ionic strength is ≈25°C lower than at 2 M [NaCl].

Fig. 4.
Relaxation kinetics of EnHD residues 16–59 monitored by fluorescence (a, b, and f) and IR (c and d) and equilibrium properties (e). (a) Kinetic time course for EnHD residues 16–59 at 25°C, 2 M [NaCl]. The raw data (red) were fitted ...

Microsecond Folding by IR and Fluorescence.

T-jump kinetic experiments to 25°C used IR detection as a global probe for formation of secondary structure (IR, 1,634 cm−1 helical structure) and fluorescence detection of local tertiary interactions (Trp) for the HTH motif fragment and L16A mutant (Fig. 4 and Tables 1 and and2).2). Both probes yielded similar relaxation times of ≈2.2 μs for L16A and 2.7 μs for the fragment at low salt concentrations, indicating concurrent formation of secondary and tertiary interactions. The presence of this 2- to 3-μs phase in both the native protein and its fragment suggested that the residues within the HTH motif are responsible for the ultrafast phase observed in the folding of the WT protein as well as the L16A mutant. An additional phase, with τ1/2 ~ 100 ns, was detected by IR but not by fluorescence, has been observed in all helical proteins studied so far, and might be related to helix–coil transition (30).

Table 1.
Relaxation kinetics of EnHD residues 16–59 and L16A monitored by IR
Table 2.
Relaxation kinetics of EnHD residues 16–59 and L16A measured by fluorescence

Both EnHD WT and the L16A mutant have biphasic folding kinetics at high salt concentrations (17). The fragment at 500 mM NaCl, monitored by IR, and at 2 M [NaCl], monitored by fluorescence (Fig. 4), had only monophasic kinetics with relaxation times of 3 and 4.8 μs, respectively, with only weak temperature dependence. The absence of the slow phase is consistent with the docking of H1 being responsible for the slow kinetic phase in the folding of WT protein. In a control experiment on L16A at high salt, we observed biphasic kinetics and were able to extract the two rates from a biexponential fit for only the lower temperatures. The rates for the fast phase are in very good agreement with the ones obtained for the fragment, suggesting that they might be reporting on the same process.


Structure of Fragment at Physiological Ionic Strengths.

The 16–59 fragment of EnHD (L16A) was stabilized by high concentrations of salt, just as is L16A, because of screening of its positive charges. At high salt, the fragment had a well defined folding transition. At lower ionic strengths, we could not follow the full transition because the native state was not populated sufficiently to give good baselines. The structure was determined at physiological ionic strength where the protein is only ≈60–70% folded. As is usual, however, for a protein in fast exchange, the NMR observables could still be explained by using a single backbone structure. Truncation of L16A before position 16 resulted in disruption of H1 but retention of the HTH motif. Although there were small perturbations of backbone and side-chain chemical shifts, in the fragment, its overall NOE pattern, chemical shifts, and other NMR observables were nearly identical to those observed for EnHD L16A for the HTH motif and part of the WT-turn region. Increasing the NaCl concentration, to stabilize the native HTH fold, resulted in an increase in the chemical shift dispersion. All of the properties of this protein fragment, e.g., having a defined structure with good chemical shift dispersion and a thermal transition, are consistent with the notion of a self-folding domain, making the EnHD HTH motif an ultrafast-folding protein domain. We could further truncate the protein to the sequence 24–59, and the structure was still folded (T.L.R., unpublished data).

Motifs and Domains.

The HTH motif has not been found independently folded in solution, and so it is thought (4) that the smallest domain that contains the HTH motif consists of a three-helical bundle, like any homeodomain protein. The present work shows that the HTH motif per se can fold and is stable in solution, making the simplest motif a domain. Multidomain proteins are created by duplication, divergence, and recombination of domains (31). Formation of new domains is more difficult to explain (32). The presence of stable motifs would support the hypothesis that the motifs constitute “the atoms of evolution” (33, 34) and explains the wide range of proteins that have an HTH motif (4). The ability of such a structured segment to act as a template for the folding of the rest of the resulting polypeptide would increase the likelihood of successful nonhomologous recombination in vivo, particularly if this segment contains all of the essential functional determinants of a parent protein (35).

Filling in the EnHD Folding Pathway.

The mutant L16A has native structure and native folding kinetics at high ionic strength but denatures at physiological ionic strength (14, 17). At lower ionic strengths, L16A is a genuine denatured state/folding intermediate of the high ionic strength native structure and also a model for the WT nonnative state under physiological conditions. Previously, we were unable to distinguish definitively whether the nonnative state of L16A is an intermediate on the folding pathway separated by an energy barrier from a denatured state or a denatured state of varying compaction (12, 17). As we had used just fluorescence as a probe in ultrafast-folding kinetics, we could only tentatively assign the ultrafast-folding phase of EnHD to the formation of the HTH motif. Here, the simultaneous monitoring of IR changes as a measure of global secondary structure formation and fluorescence for local tertiary interaction in ultrafast T-jumps showed concomitant formation of secondary and tertiary structure (Fig. 4 and Tables 1 and and2).2). Because the fragment has all of the properties of a folded protein, especially with all probes of folding having the same folding kinetics, it is most likely that EnHD L16A at low ionic strength is a folding intermediate. Furthermore, at high ionic strength where the L16A mutant folds by three-state kinetics, just like the parent EnHD, the fast step is clearly global formation of secondary structure and the slow step the final docking, as also evinced by simulation (12, 13, 15). Small size, fast folding, and availability of the structure for the HTH domain should make it an excellent target for further testing of simulation and folding algorithms, both as an independent domain as well as in the context of the full-length EnHD protein.

Proof of an On-Pathway Intermediate.

Classical evidence for the presence of an on-pathway intermediate requires three criteria to be fulfilled: (i) the intermediate forms rapidly enough to be on the pathway; (ii) it reacts (or interconverts) rapidly enough to be on pathway; and (iii) it can be isolated and characterized (29). We have isolated and solved the structures of good models for the intermediates in the folding of EnHD and now have measured the relaxation kinetics for their formation. The HTH motif of EnHD (and its L16A mutant at high salt) is a well defined entity that is formed fast enough to be on the reaction pathway. H1 of L16A docks onto the HTH motif fast enough to be on the overall folding pathway.

Downhill Folding?

The rate constants for the folding of EnHD are well above the values proposed for the onset of barrier-free downhill folding (36). Our interpretation of the folding pathway of EnHD WT as a three-state model has been contested, and it is claimed that biexponential relaxation kinetics with two phases that differ by approximately one order of magnitude is the trademark of folding with a marginal barrier, and so EnHD is an example of one-state downhill folding (36). Elsewhere (37), it is contended that the fast phase of folding of lambda6–85 results from reaction of a “preactivated” transition state population on the “molecular time scale” and that the fast phase of EnHD might have the same origin. But Hagen (38) recently has shown that the kinetic signatures proposed for downhill folding are unreliable, including the probe-dependence of kinetics (39). We also have shown that it is not possible to distinguish between downhill and conventional barrier-limited folding for rapidly interconverting protein states by equilibrium methods, and it is necessary to invoke single-molecule studies or a denaturant dependence of folding rate constants that signifies a large change in accessible surface area between ground and transition states (40). Small domains are particularly difficult to study because of their very broad denaturation transitions (17). Accordingly, it is most likely to be a futile exercise to validate mechanisms by conventional kinetics and equilibrium measurements on a single species of protein.

To distinguish between mechanisms, it is necessary to assign observed rate constants to specific processes. The use of mutated and engineered constructs with different properties is invaluable in such assignments. For example, the suggested reaction of a preactivated transition state population (37) for the fast phase is unlikely because that phase is found for the EnHD 16–59 fragment that does not populate the native state. Furthermore, by using both IR and fluorescence probes, we can assign the fast phase to the process of concurrent formation of secondary and tertiary structures in both the fragment and the first step of the full-length protein. The concurrent formation of secondary and tertiary structure is straightforwardly consistent with a cooperative transition but would require more tortuous explanation by downhill folding. Furthermore, the finding of the fast phase in the fragment and the same fast phase preceding the slow phase for the full-length protein shows that the two exponentials are related to two different steps and not the manifestation of complex kinetics proposed as a signature of downhill folding.

The different temperature dependences of the fast and slow kinetic phases in EnHD could provide an explanation for the observation of the “strange,” stretched exponentials in folding kinetics of a number of proteins (39) as they merge and also would predict a more complex relaxation decay. Because the slower kinetic phase has higher temperature dependence than the fast phase (as observed in Fig. 4), at higher temperatures, where their rates become similar, it is not possible to fit both of them with confidence, and the resulting trace also will fit to a stretched exponential.

More General Implications.

There are four other examples of proteins with HTH motifs that display biphasic folding kinetics or show signature of a structured denatured state: c-Myb (41, 42), computationally redesigned EnHD with 25% sequence identity to the original EnHD (43), FF domain (44, 45), and lambda6–85-repressor (37). It is likely that the HTH motif also will be structured to various extents in the denatured state of some of these proteins. For example, removal of a Pro residue from the turn region of c-Myb increased the HTH motif propensity and caused the intermediate to be experimentally visible (41, 42).

It was suggested in the initial formulation of the nucleation-condensation mechanism (46) that there is a transition from concerted to stepwise folding as the stability of substructures increases so that they can fold independently. This view is increasingly supported by experiment and simulation (42, 4749). The finding that the HTH domain can have an independent existence confirms the underlying principle of independent stability in the slide from nucleation-condensation to framework mechanisms in the folding of the EnHD family (42).


Protein Expression.

The L16A mutant of EnHD was expressed as described in ref. 17. The 16–59 fragment was obtained by overnight thrombin cleavage of 1 mM solution of EnHD L16A mutant at 37°C by using 100 units·ml−1 of the protease in 50 mM potassium phosphate buffer, pH 8.0. The fragment was purified by using reverse phase chromatography on a Vydac (Deerfield, IL) C18 238TP101522 semipreparative column.


All NMR experiments were performed in 100 mM sodium chloride, 50 mM deuterated acetate buffer (pH 5.7), and 7% D2O at 5°C on a 500 μM sample. The other experiments were performed by using a nondeuterated version of the above solution. The 2D NOESY spectrum (150-ms mixing time) for distance constraints in structure calculation was acquired on Bruker (Billerica, MA) DRX600 spectrometer with others acquired on Bruker DRX500 spectrometer. The chemical shifts of the fragment were assigned by using standard triple-resonance methods (50).

For structure determination, the volumes of the peaks from the 2D NOESY spectra were converted into distances based on sequential NOEs from regular secondary structure elements and put into four categories: 1.8–2.7 Å, 1.8–3.6 Å, 1.8–5.5 Å, and 1.8–7.0 Å. The backbone [var phi]/ψ dihedral angle restraints were taken from the x-ray crystal structure of PDB ID code 1ENH for residues 28–52. The bounds for the dihedral restraints were equal to twice the standard deviation of the angles derived from the nine crystal structures of DNA-unbound EnHD WT and its mutants (PDB ID codes 1P7I/1P7J/1ENH) (21, 51). These bounds were similar to the ones obtained from TALOS (52).

Structures were calculated with CNS 1.1 (53) by using anneal.inp. Twenty-five (of 50) lowest-energy structures were taken as the representative of the ensemble. The Ramachandran plot analysis shows 75.7% residues in most-favored regions, 20.1% in additional allowed, 3.6% in generously allowed, and 0.6% in disallowed regions. The HNAn external file that holds a picture, illustration, etc.
Object name is cjs0807.jpgN RDCs were measured in 7% acrylamide gels, radially compressed with a gel-stretching apparatus (54) and measured by using DSSE (55). The rmsd between back-to-back measurement of RDC was <0.2 Hz. Comparison of the predicted and determined RDCs was carried out with PALES (56).


Degassed and filtered samples of EnHD at 200–300 μM were pumped slowly through the 0.5 × 2-mm quartz flow cell at 5 μl·min−1 during data acquisition with the temperature jump apparatus (57) (described in detail in SI Text). We estimate the overall bandwidth of the detection electronics to be ≈20 MHz.

Equilibrium Measurements.

Sedimentation equilibrium was performed on a Beckman (Fullerton, CA) XL-I analytical ultracentrifuge on a 15N-labeled protein at a concentration of 500 μM in the NMR buffer at 25°C at 36,000 rpm and 45,000 rpm. The data were fitted to an ideal, single-species model by using the in-house software Ultraspin 2.3, assuming a solvent density of 1.0 g·ml−1 and a partial molar volume of 0.73 ml·g−1. CD measurements were recorded on a Jasco (Easton, MD) J-720 spectrometer with a heating rate of 60–80°C/h and a protein concentration 25–50 μM.

Supplementary Material

Supporting Information:


We thank Dmitry Veprintsev for help with analytical ultracentrifugation. T.L.R. was supported by a Trinity College Internal Graduate Studentship and the Medical Research Council.


Engrailed homeodomain
helix 1
helix 2
helix 3
Protein Data Bank
heteronuclear single quantum correlation
residual dipolar coupling.


The authors declare no conflict of interest.

This article contains supporting information online at www.pnas.org/cgi/content/full/0703434104/DC1.


1. Branden E, Tooze J. Introduction to Protein Structure. New York: Garland Publishing; 1999.
2. Doolittle RF. Annu Rev Biochem. 1995;64:287–314. [PubMed]
3. Harrison SC, Aggarwal AK. Annu Rev Biochem. 1990;59:933–969. [PubMed]
4. Aravind L, Anantharaman V, Balaji S, Babu MM, Iyer LM. FEMS Microbiol Rev. 2005;29:231–262. [PubMed]
5. Ohlendorf DH, Anderson WF, Matthews BW. J Mol Evol. 1983;19:109–114. [PubMed]
6. Sauer RT, Yocum RR, Doolittle RF, Lewis M, Pabo CO. Nature. 1982;298:447–451. [PubMed]
7. Ferrer-Costa C, Shanahan HP, Jones S, Thornton JM. Bioinformatics. 2005;21:3679–3680. [PubMed]
8. Dodd IB, Egan JB. Nucl Acids Res. 1990;18:5019–5026. [PMC free article] [PubMed]
9. Breg JN, Vanopheusden JHJ, Burgering MJM, Boelens R, Kaptein R. Nature. 1990;346:586–589. [PubMed]
10. Anderson TA, Cordes MH, Sauer RT. Proc Natl Acad Sci USA. 2005;102:18344–18349. [PMC free article] [PubMed]
11. Gehring WJ, Qian YQ, Billeter M, Furukubo-Tokunaga K, Schier AF, Resendez-Perez D, Affolter M, Otting G, Wuthrich K. Cell. 1994;78:211–223. [PubMed]
12. Mayor U, Grossmann JG, Foster NW, Freund SM, Fersht AR. J Mol Biol. 2003;333:977–991. [PubMed]
13. Mayor U, Johnson CM, Daggett V, Fersht AR. Proc Natl Acad Sci USA. 2000;97:13518–13522. [PMC free article] [PubMed]
14. Mayor U, Guydosh NR, Johnson CM, Grossmann JG, Sato S, Jas GS, Freund SM, Alonso DO, Daggett V, Fersht AR. Nature. 2003;421:863–867. [PubMed]
15. Hubner IA, Deeds EJ, Shakhnovich EI. Proc Natl Acad Sci USA. 2006;103:17747–17752. [PMC free article] [PubMed]
16. Beck DA, Daggett V. Methods. 2004;34:112–120. [PubMed]
17. Religa TL, Markson JS, Mayor U, Freund SM, Fersht AR. Nature. 2005;437:1053–1056. [PubMed]
18. DeMarco ML, Alonso DO, Daggett V. J Mol Biol. 2004;341:1109–1124. [PubMed]
19. Neidigh JW, Fesinmeyer RM, Andersen NH. Nat Struct Biol. 2002;9:425–430. [PubMed]
20. Sklenar V, Bax A. J Mag Ree. 1987;74:469–479.
21. Clarke ND, Kissinger CR, Desjarlais J, Gilliland GL, Pabo CO. Protein Sci. 1994;3:1779–1787. [PMC free article] [PubMed]
22. Bax A. Protein Sci. 2003;12:1–16. [PMC free article] [PubMed]
23. Prestegard JH, al-Hashimi HM, Tolman JR. Q Rev Biophys. 2000;33:371–424. [PubMed]
24. Cornilescu G, Marquardt JL, Ottiger M, Bax A. J Amer Chem Soc. 1998;120:6836–6837.
25. Bonvin AM, Brunger AT. J Biomol NMR. 1996;7:72–76. [PubMed]
26. Bonvin AM, Brunger AT. J Mol Biol. 1995;250:80–93. [PubMed]
27. Wuthrich K. NMR of Proteins and Nucleic Acids. New York: Wiley; 1986.
28. Mezei M. Protein Eng. 1998;11:411–414. [PubMed]
29. Fersht AR. Structure and Mechanism in Protein Science: A Guide to Enzyme Catalysis and Protein Folding. New York: Freeman; 1998.
30. Brewer SH, Song B, Raleigh DP, Dyer RB. Biochemistry. 2007;46:3279–3285. [PubMed]
31. Vogel C, Bashton M, Kerrison ND, Chothia C, Teichmann SA. Curr Opin Struct Biol. 2004;14:208–216. [PubMed]
32. Lupas AN, Ponting CP, Russell RB. J Struct Biol. 2001;134:191–203. [PubMed]
33. Liu J, Rost B. Curr Opin Chem Biol. 2003;7:5–11. [PubMed]
34. Rost B. Curr Opin Struct Biol. 2002;12:409–416. [PubMed]
35. Andreeva A, Murzin AG. Curr Opin Struct Biol. 2006;16:399–408. [PubMed]
36. Naganathan AN, Doshi U, Fung A, Sadqi M, Munoz V. Biochemistry. 2006;45:8466–8475. [PMC free article] [PubMed]
37. Yang WY, Gruebele M. Nature. 2003;423:193–197. [PubMed]
38. Hagen SJ. Proteins. 2007 Mar 23; doi: 10.1002/prot.21342. [PubMed] [Cross Ref]
39. Ma H, Gruebele M. Proc Natl Acad Sci USA. 2005;102:2283–2287. [PMC free article] [PubMed]
40. Huang F, Sato S, Sharpe TD, Ying L, Fersht AR. Proc Natl Acad Sci USA. 2007;104:123–127. [PMC free article] [PubMed]
41. White GW, Gianni S, Grossmann JG, Jemth P, Fersht AR, Daggett V. J Mol Biol. 2005;350:757–775. [PubMed]
42. Gianni S, Guydosh NR, Khan F, Caldas TD, Mayor U, White GW, DeMarco ML, Daggett V, Fersht AR. Proc Natl Acad Sci USA. 2003;100:13286–13291. [PMC free article] [PubMed]
43. Gillespie B, Vu DM, Shah PS, Marshall SA, Dyer RB, Mayo SL, Plaxco KW. J Mol Biol. 2003;330:813–819. [PubMed]
44. Allen M, Friedler A, Schon O, Bycroft M. J Mol Biol. 2002;323:411–416. [PubMed]
45. Jemth P, Gianni S, Day R, Li B, Johnson CM, Daggett V, Fersht AR. Proc Natl Acad Sci USA. 2004;101:6450–6455. [PMC free article] [PubMed]
46. Itzhaki LS, Otzen DE, Fersht AR. J Mol Biol. 1995;254:260–288. [PubMed]
47. Brockwell DJ, Radford SE. Curr Opin Struct Biol. 2007;17:30–37. [PMC free article] [PubMed]
48. Daggett V, Fersht AR. Trends Biochem Sci. 2003;28:18–25. [PubMed]
49. Gianni S, Geierhaas CD, Calosci N, Jemth P, Vuister GW, Travaglini-Allocatelli C, Vendruscolo M, Brunori M. Proc Natl Acad Sci USA. 2007;104:128–133. [PMC free article] [PubMed]
50. Bax A, Ikura M, Kay LE, Barbato G, Spera S. Ciba Found Symp. 1991;161:108–119. 119–135. discussion. [PubMed]
51. Stollar EJ, Mayor U, Lovell SC, Federici L, Freund SM, Fersht AR, Luisi BF. J Biol Chem. 2003;278:43699–43708. [PubMed]
52. Cornilescu G, Delaglio F, Bax A. J Biomol NMR. 1999;13:289–302. [PubMed]
53. Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS, et al. Acta Crystallogr D Biol Crystallogr. 1998;54:905–921. [PubMed]
54. Chou JJ, Gaemers S, Howder B, Louis JM, Bax A. J Biomol NMR. 2001;21:377–382. [PubMed]
55. Cordier F, Dingley AJ, Grzesiek S. J Biomol NMR. 1999;13:175–180. [PubMed]
56. Zweckstetter M, Bax A. J Amer Chem Soc. 2000;122:3791–3792.
57. Brewer SH, Vu DM, Tang Y, Li Y, Franzen S, Raleigh DP, Dyer RB. Proc Natl Acad Sci USA. 2005;102:16662–16667. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem Compound links
  • MedGen
    Related information in MedGen
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles
  • Structure
    Published 3D structures
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...