Binding Affinity and Function of the Extremely Disordered Protein Complex Containing Human Linker Histone H1.0 and Its Chaperone ProTα

Abstract

It was recently reported that human linker histone H1.0 and its chaperone prothymosin-α (ProTα) form an extremely disordered 1:1 complex with an ultrahigh affinity (equilibrium dissociation constant K_D of ∼2 × 10⁻¹² M) measured using a single-molecule Förster resonance energy transfer method. It was hypothesized that the ultrahigh affinity and extreme disorder may be required for the chaperone function of ProTα, in which it displaces the linker histone from condensed chromatin. Here, we measure the binding affinity for the ProTα–H1.0 complex using isothermal titration calorimetry and report a K_D value of (4.6 ± 0.5) × 10⁻⁷ M. In addition, we show that ProTα facilitates the formation of the H1.0–nucleosome complex in vitro. The results of our study contrast with those of the previous report and provide new insights into the chaperone function of ProTα. Possible causes for the observed discrepancy in binding affinity are discussed.

Molecular communication in biology is facilitated by protein interactions. The specificity of the interactions is thought to be determined by the precise structural complementarity at the binding interface of the interacting partners. This paradigm of structural biology has held up very well, even for proteins that are intrinsically disordered or contain intrinsically disordered regions, simply because they still form specific structures at the binding interface in the complex. However, a recent challenge to the paradigm has emerged. Borgia et al.¹ reported that human linker histone H1.0 and its chaperone ProTα form a 1:1 complex that displays extreme disorder. The absence of a well-structured interface between H1.0 and ProTα is characterized by the small chemical shift changes upon binding of the two proteins as well as fast dynamic motions throughout the binding residues in the complex. Despite its disordered features, the ProTα–H1.0 complex was reported to have a surprisingly ultrahigh affinity, with a K_D value of ∼2 × 10⁻¹² M as measured using the single-molecule Förster resonance energy transfer (smFRET) method. There has been speculation that the ultrahigh affinity with extreme disorder is required for the chaperone function of ProTα, displacement of the linker histone from condensed chromatin.¹

To understand the molecular mechanism of the chaperone function of ProTα, we investigated the interactions between ProTα and H1.0. In contrast to the previous study, we used conventional isothermal titration calorimetry (ITC) to measure the binding affinity. Surprisingly, we obtained a K_D value that is nearly 5 orders of magnitude larger than the previously reported value measured using smFRET under the physiologically relevant ionic strength (IS). In addition, we found that ProTα facilitates the formation of the H1.0–nucleosome complex in vitro. Here we report our experimental results and discuss possible causes for the observed discrepancy in binding affinity.

We measured the binding affinities of ProTα and H1.0 at various ionic strengths by titrating ProTα into H1.0 in solution or vice versa (Figure 1 and Figure S1). The change in measured heat as a function of the molar ratio of ProTα to H1.0 fits a 1:1 binding model, which is consistent with the previous smFRET result. However, the K_D values as measured using ITC were larger than those previously measured using smFRET (Figure 1b). At the physiologically relevant ionic strength (165 mM), our K_D value was (4.6 ± 0.5) × 10⁻⁷ M, which is nearly 5 orders of magnitude larger than the previously reported value at the same ionic strength. Only at a much higher ionic strength did we obtain a binding affinity that is only slightly larger than the corresponding value measured using smFRET.

Figure 1.

Binding affinity of H1.0 and ProTα measured using ITC. (a) Illustration of a representative titration profile at an ionic strength (IS) of 165 mM. The filled circles represent the experimental data. The solid line is the fitting curve. (b) Comparison of the ITC results (●) with those measured using the smFRET method (□).¹ The solid lines are the fitting curves: a linear function to the ITC data and an exponential function to the smFRET data. The table shows the ionic strengths and measured K_D values from the ITC experiments. The K_D value is the average of the results from two ITC experiments. The uncertainty shows the deviation from the average.

We also found that the logarithm of our K_D values versus the logarithm of ionic strengths fits to a linear function that predicts the number of released encounter ions to be [d(ln K_D)/d(ln IS)] ≈ 6.² In contrast, fitting of the corresponding data as measured using the smFRET method to a linear function in the previous study led to the number of released encounter ions being ∼18.¹ However, we found that the data measured using the smFRET method better fit an exponential function (Figure 1b). These results suggest that the structural state of the complex as characterized by ITC is independent of ionic strength, whereas the structural state revealed by the smFRET method may change as a function of ionic strength.

In addition, the ultrahigh-affinity binding appears to be inconsistent with the previously reported nuclear magnetic resonance (NMR) results (see Extended Data Figure 2 of ref 1). The NMR study showed only one set of peaks for both H1.0 and ProTα when either is present in excess, suggesting that H1.0 and ProTα associate and dissociate on the NMR time scale (milliseconds) or faster. However, using the ultrahigh K_D value (∼2 × 10⁻¹² M) and the association rate constant (k_on of 3.1 × 10⁹ M⁻¹) measured from the stopped-flow experiment,¹ the calculated off-rate constant (k_off) is 6.5 × 10⁻³ s⁻¹. With such a slow off-rate constant, the dissociation from the complex to the free protein state would occur in the slow exchange regime of the NMR experiment and lead to two sets of NMR peaks for the species that is in excess: one set would correspond to the free form and the other to that in the complex. By contrast, using the K_D value measured from the ITC experiment, the calculated k_off is ∼1.3 × 10³ s⁻¹, which is consistent with the observation of one set of NMR peaks for both H1.0 and ProTα.

To test whether ProTα prevents linker histone H1.0 from forming a complex with the chromatin because of its ultrahigh affinity as previously postulated, we examined the effect of ProTα on the formation of the H1.0–nucleosome complex. We mixed H1.0 and the nucleosome in a 1:1 ratio in the presence of various amounts of ProTα that ranged from 0 to 3-fold greater than the nucleosome concentration. The nucleosome consists of 197 bp DNA centered with the Widom 601 DNA sequence that is used for nucleosome positioning, leaving ∼25 bp linker DNA on each side of the nucleosome.³ By conducting an electrophoretic mobility shift assay (EMSA), we found that H1.0 forms the complex with the nucleosome more efficiently in the presence of ProTα (Figure 2).

Figure 2.

ProTα facilitates the formation of the H1.0–nucleosome complex. (a) A 5% native polyacrylamide gel electrophoresis EMSA on the free nucleosome, H1.0, and ProTα mixed at the indicated ratios. (b) Plots of the integrated intensities within the corresponding rectangular boxes (marked with i–iii) on the gel. Note that the nucleosome–H1.0 complex is more compact than the nucleosome and runs faster in the gel.

Previous studies have shown that linker histone H1.0 binds to the nucleosome with an affinity of ∼5 pM.⁴ Thus, the submicromolar binding affinity is sufficient for ProTα to be able to deposit H1.0 to the nucleosome while also competing with nonspecific binding of H1.0 to DNA. These results suggest that ProTα functions as a chaperone to facilitate the formation of the H1.0–nucleosome complex rather than to remove linker histones from condensed chromatin.

It is not clear what causes the large discrepancy in the measured K_D values determined by the ITC and smFRET methods. ITC is a well-established method for measuring the binding affinity of proteins, and the experimental procedure involves simple mixing. By contrast, smFRET is an indirect method for measuring binding affinity and assumes that binding of ProTα and H1.0 can be described with a Langmuir-type isotherm at ionic strengths of ≥200 mM and model fitting to derive an effective concentration of H1.0 in solution.¹ In addition, the smFRET experiments used ProTα that is labeled with two fluorescent dyes. Borgia et al.¹ justified the use of dye labeling by showing that the locations of the dyes in ProTα had only weak effects on binding affinity. However, the effect of the dyes on binding affinity could be nonspecific and independent of their locations in the protein. For example, association of additional dye-labeled monomers to the preformed ProTα–H1.0 complex is observed in the smFRET experiment¹ (Extended Data Figure 4d,e of ref 1) but not in the ITC or NMR experiments using proteins without dye labeling.

We performed size-exclusion experiments for the mixture of ProTα and H1.0 at a 1:1 ratio at a concentration of ∼10 μM for each protein and compared its elution profile with those of mixtures containing additional H1.0 (Figure 3). We observed no larger complex than the ProTα–H1.0 complex at a 1:1 ratio, i.e., no formation of higher-order oligomers. The 1:1 stoichiometry observed in the size-exclusion and ITC experiments at micromolar concentrations excludes the possibility that the ultrahigh and moderate binding affinities represent two different binding events as suggested by Borgia et al.¹

Figure 3.

Size-exclusion experiments showed no formation of higher-order oligomers when H1.0 is in excess. Elution profiles of the ProTα–H1.0 complex (black solid line) at concentration of ∼10 μM for each of the proteins and the same complex with additional amounts of H1.0.

Finally, fluorescent dye labeling was previously suggested to cause compaction of some intrinsically disordered proteins.^5–10 For protein NLS,^5,6 it was attributed to the existence of a large number of charged residues. This result is consistent with the observation that the discrepancies in binding affinity measured using ITC and smFRET are dependent on ionic strength (Figure 1b), with a larger difference at lower ionic strengths.

In summary, our study shows that ProTα binds to H1.0 with submicromolar affinity and functions as a chaperone to promote deposition of H1.0 to form a specific complex with the nucleosome in vitro. Fluorescent dye labeling is a likely cause of the dramatic discrepancy in binding affinities as measured using the smFRET and ITC methods.