Tissue-Specificity of Dystrophin-Actin Interactions: Isoform-Specific Thermodynamic Stability and Actin-Binding Function of Tandem Calponin-Homology Domains.

Genetic mutations in Duchenne muscular dystrophy (DMD) gene affecting the expression of dystrophin protein lead to a number of muscle disorders collectively called dystrophinopathies. In addition to muscle dystrophin, mutations in brain-specific dystrophin isoforms, in particular those that are expressed in the brain cortex and Purkinje neurons, result in cognitive impairment associated with DMD. These isoforms carry minor variations in the flanking region of the N-terminal actin-binding domain (ABD1) of dystrophin, which is composed of two calponin-homology (CH) domains in tandem. Determining the effect of these sequence variations is critical for understanding the mechanisms that govern varied symptoms of the disease. We studied the impact of differences in the N-terminal flanking region on the structure and function of dystrophin tandem CH domain isoforms. The amino acid changes did not affect the global structure of the protein but drastically affected the thermodynamic stability, with the muscle isoform more stable than the brain and Purkinje isoforms. Actin binding investigated with actin from different sources (skeletal muscle, smooth muscle, cardiac muscle, and platelets) revealed that the muscle isoform binds to filamentous actin (F-actin) with a lower affinity compared to the brain and Purkinje isoforms, and a similar trend was observed with actin from different sources. In addition, all isoforms showed a higher affinity to smooth muscle actin in comparison to actin from other sources. In conclusion, tandem CH domain isoforms might be using minor sequence variations in the N-terminal flanking regions to modulate their thermodynamic stability and actin-binding function, thus leading to specificity in dystrophin-actin interactions in various tissues.

Introduction

Duchenne muscular dystrophy (DMD) is an X-linked genetic disorder caused by the absence or loss of function of dystrophin protein (UniProtKB P11532).[1−4] It is characterized by progressive muscle weakness and degeneration, finally leading to death at an early age due to cardiac or respiratory failure.[5] One of the major functions of dystrophin includes binding to actin cytoskeleton using its N-terminal actin-binding domain (N-ABD or ABD1).[6,7] At least seven different promoters exist on the DMD gene that lead to tissue-specific expression of dystrophin isoforms with different lengths.[2] Transcription from the first three promoters leads to the expression of full-length (427 kDa) dystrophin isoforms: muscle dystrophin (in skeletal and cardiac muscles), brain dystrophin (in the hippocampus and the cortex in the brain), and Purkinje dystrophin (in cerebellar Purkinje neurons), named based on their major site of localization.[8,9] Other promoters lead to the expression of progressively shorter isoforms from the N-terminus (260, 140, 116, and 71 kDa) lacking the ABD1 (Figure a). Each of the three full-length isoforms [muscle (DP427M or Dys-M), brain (DP427B or Dys-B), and Purkinje (DP427P or Dys-P)] is encoded by a unique first exon and a set of common exons from 2 to 79, leading to variations in the N-terminal flanking region. In this study, we examined how these minor sequence variations at the N-terminus affect the structure and function of dystrophin ABD1.

Figure 1

Dystrophin isoforms resulting from different promoters and sequence comparison of actin-binding tandem calponin-homology (CH) domains of muscle, Purkinje, and brain isoforms. (A) A schematic representation of the dystrophin gene and isoforms resulting from transcription initiating from different promoters. The position of promoters is represented on the gene with curved arrows as muscle (M), Purkinje (P) neurons, brain (B), retina (R), brain, kidney (B3), peripheral nervous system (S), and ubiquitous (G). Transcription initiation from M, P, and B promoters utilizes a unique first exon and produces protein products with a molecular weight of ∼427 kDa, which are represented as DP427M (Dys-M), DP427P (Dys-P), and DP427B (Dys-B) respectively. Transcription from R, B3, S, and G promoters initiates with exon 30, 45, 56, and 63 and produces protein products with molecular weights of 260 kDa (DP260), 140 kDa (DP140), 116 kDa (DP116), and 71 kDa (DP71), respectively. (B) Sequence alignment of the tandem CH domains of Dys-M, Dys-P, and Dys-B. The amino acid sequence translated from the unique first exon is shown in bold letters in black (Dys-M), red (Dys-P), and blue (Dys-B). The rest of the amino acid sequence is identical for the three isoforms. The figure also shows the three actin-binding surfaces (ABS1, ABS2, and ABS3) through which dystrophin tandem CH domain interacts with actin. (C) Multiple sequence alignment of six actin isoforms at the amino acid level. Sequence differences are shown in bold letters. Differences in the N-terminal acidic residues are shown in green. Unique amino acid substitutions in a single isoform are represented by cyan, and substitutions in two isoforms are represented by red. Amino acids present in the majority of isoforms are represented by black.

Dystrophin ABD1 is composed of two calponin-homology (CH) domains in tandem. Similar tandem domains are present in a number of actin-binding proteins including utrophin, α-actinin, spectrin, and fimbrin.[10−12] Previous studies have shown that the N-terminal (CH1) and C-terminal (CH2) domains contribute differentially to the stability and function of dystrophin and utrophin tandem CH domains, with CH1 domains being unstable but with a higher binding affinity for filamentous actin (F-actin).[13−15] CH2 domains do not bind to actin but provide stability to ABD1.[13,15] Utrophin tandem CH domain contains a long 23-amino-acid flanking region compared to dystrophin tandem CH domain, and deleting this region impacts the stability and actin-binding affinity of the tandem CH domain, making the truncated utrophin tandem CH domain behave similar to that of dystrophin with a shorter flanking region.[16] A similar role for the N-terminal flanking regions has been recently proposed for other tandem CH domains.[17−19] The ABD1 of the three dystrophin isoforms (muscle, Purkinje, and brain) is composed of 246, 242, and 238 amino acid residues, respectively, and the only difference lies in their N-terminal flanking regions before the first actin-binding site (ABS1) (Figure b). An 11 amino acid stretch in muscle dystrophin is replaced by 7 amino acids in Purkinje dystrophin and 3 amino acids in brain dystrophin. We examined how these three naturally occurring isoforms use minor sequence variations in their flanking regions to modulate the thermodynamic stability and actin-binding function, and whether they lead to tissue specificity of dystrophin–actin interactions.

Similar to dystrophin, actin also exists in multiple isoforms with nearly identical amino acid sequences but coded by six different genes (Figure c).[20] Actin isoforms have a conserved amino acid sequence across species, and the six genes encoding them are also evolutionary conserved. Like dystrophin, actin isoforms show tissue-specific expression, with β- and γ-actin ubiquitously expressed, while other actin isoforms expressed in specific muscle tissues. This indicates the importance of different isoforms of actin and led to the hypothesis that actin isoforms perform nonredundant functions in respective tissues. Animal knockout studies have shown that specific actin isoforms perform specialized functions, which cannot be taken over by concomitant up-regulation of other isoforms.[20,21] Biochemical and biophysical studies have shown that despite a difference of a few amino acids mainly in their N-terminal flanking regions (Figure c), distinct actin isoforms have differing properties in terms of their polymerization and depolymerization potential,[22,23] Thus, similar to dystrophin, minor variations in the amino acid sequence and tissue specificity of actin isoforms may be an important factor in determining actin–dystrophin interactions. In addition, specificity might exist in actin–dystrophin interactions, i.e., a particular dystrophin isoform may be able to distinguish different actin isoforms in terms of actin-binding affinity, which has not been examined before.

In this study, we examined how the tandem CH domains of the three naturally occurring dystrophin isoforms use their minor variations in the N-terminal flanking regions to modulate their stability and actin-binding function. In addition to determining the amino acid sequence of the protein (Figure a), transcription from different promoters can also control the levels of transcript and protein levels, leading to tissue specificity.[8,24] Since dystrophin isoforms are expressed in different tissues and the actin isoform expression pattern also varies with the tissue type, we also examined whether different isoforms of the dystrophin tandem CH domain can differentiate between actin isoforms in terms of their binding affinity.

Results

Purification and Characterization of Proteins

Tandem CH domains of the three dystrophin isoforms were expressed in high yield in Escherichia coli, using the protocols described in the Materials and Methods section. Proteins were obtained from soluble fraction and purified to homogeneity using nickel-nitrilotriacetic acid (Ni-NTA) affinity chromatography. Figure a shows the sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of the three purified isoforms. Native-PAGE analysis (Figure b) shows the three isoforms as homogenous preparations. The tandem CH domain of Dys-M migrates slightly faster toward the positively charged electrode compared to the other two proteins because of the presence of two additional acidic residues in its N-terminal flanking region.

Figure 2

Purification and characterization of tandem CH domains of dystrophin isoforms. (A) SDS-PAGE and (B) native-PAGE of purified dystrophin tandem CH domains. Lanes labeled Dys-M, Dys-P, and Dys-B correspond to the tandem CH domains of the skeletal muscle, Purkinje neurons, and brain cortex isoforms of dystrophin, respectively. Lane labeled M represents the molecular weight markers (180, 130, 100, 70, 55, 40, 35, 25, 15, and 10 kDa from top to bottom). (C) Far-UV circular dichroism (CD) and (D) intrinsic tryptophan fluorescence spectra of dystrophin tandem CH domain isoforms: Dys-M (black), Dys-P (red), and Dys-B (blue). Solid and dashed lines in panel D represent the fluorescence spectra of native and denatured states, respectively.

Secondary and tertiary structure characterization of the three tandem CH domains was performed by circular dichroism (CD) and fluorescence spectroscopy (Figure c,d). CD spectra of the proteins showed a characteristic signature of α-helical proteins with major negative bands at 222 and 208 nm.[25] The three spectra were indistinguishable suggesting similar content of the secondary structural components in the three isoforms. This indicates that differences in the N-terminal flanking region did not affect the secondary structure of dystrophin tandem CH domain isoforms, consistent with earlier studies on other tandem CH domains.[16,17] Fluorescence spectrum of the tandem CH domain of Dys-M in its native state showed emission maximum red-shifted compared to those of Dys-P and Dys-B. Fluorescence emission after excitation at 295 nm gives selective information on the environment of tryptophan residues.[26] Fluorescence emission spectra of exposed tryptophan residues are red-shifted compared to buried tryptophan residues. The tandem CH domain of Dys-M contains eight tryptophan residues as opposed to six in the tandem CH domains of Dys-P and Dys-B. The two additional tryptophans in the tandem CH domain of Dys-M are present in its unstructured N-terminal flanking region, which explains why its fluorescence spectrum is relatively red-shifted compared to those of Dys-P and Dys-B. The urea denatured state of all of the three isoforms showed similar emission spectra with maxima at 354 nm, suggesting similar exposure of tryptophan residues upon denaturation.

Tandem CH Domain of Muscle Dystrophin Is More Stable Than Those of Purkinje and Brain Isoforms

The stability of the three isoforms was compared using urea and thermal denaturation. The CD signal at 222 nm and fluorescence emission at 320 nm were used as probes for the secondary and tertiary structures of the proteins to record the urea denaturation profiles (Figure a,b). For thermal denaturation, in addition to CD and fluorescence (Figure c,d), optical density (OD) at 350 nm was used as a probe to monitor protein aggregation (data is not shown). Urea and thermal denaturation profiles were fit to a two-state unfolding model (eqs , 2, and 4 in the Materials and Methods section) to obtain the free energy of unfolding at zero denaturant concentration (ΔG°unf), the slope of linear variation of ΔGunf with denaturant concentration (m-value), midpoint denaturant concentration of the melt (Cm), and midpoint temperature of the thermal melt (Tm) (Tables and S1).

Figure 3

Stability comparison of the tandem CH domains of dystrophin isoforms using urea denaturation and temperature melts. (A) Change in the CD signal at 222 nm and (B) change in the fluorescence intensity at 320 nm as a function of urea concentration. (C) Change in the CD signal at 222 nm and (D) change in the fluorescence intensity at 355 nm plotted against temperature. Black, red, and blue colors represent the data corresponding to the tandem CH domains of Dys-M, Dys-P, and Dys-B, respectively.

Table 1

Thermodynamic Parameters of Tandem CH Domains Obtained from Urea and Temperature Denaturation Melts (Figure )a

	urea denaturation (global fit of CD and fluorescence)			thermal denaturation
tandem CH domain isoform	ΔG°unf (kcal/mol)	Cm (M [urea])	m-value (kcal/mol M [urea])	Tm (°C) (CD)	ΔHm, app (CD) (kcal/mol)	Tm (°C) (fluorescence)	ΔHm,app (fluorescence) (kcal/mol)
Dys-M	11.86 ± 0.34	5.56 ± 0.01	–2.13 ± 0.06	60.66 ± 0.03	5.88 ± 0.13	61.55 ± 0.11	9.91 ± 1.16
Dys-P	6.30 ± 0.30	5.26 ± 0.03	–1.20 ± 0.06	58.03 ± 0.02	5.26 ± 0.05	59.96 ± 0.12	8.07 ± 0.95
Dys-B	6.82 ± 0.26	5.16 ± 0.02	–1.32 ± 0.05	58.54 ± 0.02	5.59 ± 0.06	59.63 ± 0.11	9.62 ± 1.21

At least two data sets from two independent protein preparations were used for the data analysis.

Urea denaturation profiles with both fluorescence and CD showed that the tandem CH domain of Dys-M is more stable compared to those of Dys-P and Dys-B. ΔG°unf, Cm, and m-values obtained for a given protein by individually fitting fluorescence and CD denaturant melts to a two-state equilibrium unfolding model (eqs and 2 in the Materials and Methods section) agreed well with each other (Table S1 in the Supporting Information). Therefore, the fluorescence and CD data were analyzed globally to fit to a two-state equilibrium unfolding model (eqs and 2) by sharing the ΔG°unf, Cm, and m-values between the data sets. At least two data sets (from different batches of protein purification) of each protein were analyzed for each probe, i.e., CD and fluorescence. ΔG°unf for Dys-M tandem CH domain was 11.86 ± 0.34 kcal/mol, whereas that for the tandem CH domains of Dys-P and Dys-B were 6.30 ± 0.30 and 6.82 ± 0.26 kcal/mol, respectively, indicating that the tandem CH domain of Dys-M is stable by ∼5 kcal/mol compared to those of the other two isoforms. Similarly, the m-value for Dys-M tandem CH domains was −2.13 ± 0.06 kcal/(mol M) [urea] compared to −1.20 ± 0.06 and −1.32 ± 0.05 kcal/(mol M) [urea] for the tandem CH domains of Dys-P and Dys-B, respectively. The m-value is a measure of the change in the accessible surface area upon denaturation and thus is also a measure of the compactness of a protein in the native state in the case of two-state folders.[27] However, since the dystrophin tandem CH domain contains two domains in tandem and folds in a non-two-state manner,[28] decreases in m-value might represent an increased population of partially unfolded state(s) between native and unfolded states.[29]

Temperature denaturation profiles recorded using CD at 222 nm reflect secondary structural changes with temperature (Figure c), while those recorded using fluorescence emission at 355 nm (Figure d) give information about changes in the tertiary structure of the protein with temperature. Thermal denaturation recorded using OD at 350 nm showed an increase in light scattering, implying aggregation of these proteins with temperature (data is not shown). Since the thermal denaturation was irreversible, the data was analyzed only qualitatively for Tm comparison. Like the results obtained above for chemical denaturation, thermal denaturation also reveals higher stability for the tandem CH domain of Dys-M compared to those of Dys-P and Dys-B as reflected in the Tm values measured using both CD and fluorescence as probes (Table ).

Isoform Specificity and Actin-Binding Function of Dystrophin Tandem CH Domain Isoforms

Tandem CH domains of dystrophin isoforms were tested for binding to actin obtained from different sources (skeletal muscle, smooth muscle, cardiac muscle, and platelets), using the co-sedimentation assays.[13,14,16,30] Like dystrophin, actin also exists as six distinct isoforms with well-defined tissue specificity.[20] Whether these differences in the tissue-specific expression of actin or dystrophin isoforms impact actin–dystrophin interactions has not been examined before. Isolation of pure actin isoforms from natural sources is challenging due to small differences in their amino acid sequences (Figure c), which leads to co-purification of isoforms with one another. Thus, in this study, actin isolated from four different natural sources was used. Their composition in terms of actin isoforms is shown in Table .

Table 2

Actin Composition from Different Tissues

actin source	composition
skeletal muscle	α-skeletal actin
smooth muscle	γ-smooth muscle actin (80%), β-cytoplasmic actin (20%)
cardiac	α-cardiac actin
non-muscle	β-cytoplasmic actin (85%), γ-cytoplasmic actin (15%)

Actin co-sedimentation assays with tandem CH domains of dystrophin isoforms were performed for skeletal muscle actin (Figure S1 in the Supporting Information), using the protocols described earlier.[14,31] The data was fit using eq with Bmax fixed to 1, as observed previously for dystrophin tandem CH domains.[13,30,31] The calculated Kd values were 130 ± 7, 87 ± 4, and 109 ± 3 μM for Dys-M, Dys-P, and Dys-B respectively (Table S2 in the Supporting Information). To determine whether there exists a statistically significant difference in the actin-binding affinity of different isoforms, the co-sedimentation assays were performed with a fixed concentration of F-actin (7 μM) incubated with 60 μM of dystrophin tandem CH domains. The bar chart in Figure was obtained after repeating the co-sedimentation assays at least five to seven times. The binding data reflects that the tandem CH domain of muscle dystrophin binds weakly to all of the actin isoforms compared to those of the Purkinje and brain isoforms. Purkinje and brain isoforms do not show such distinction among themselves, similar to no changes in their stabilities (Figure and Table ). Also, smooth muscle actin shows stronger binding to all dystrophin isoforms compared to the skeletal muscle, cardiac, and non-muscle actin.

Figure 4

Actin co-sedimentation assays of tandem CH domains of dystrophin isoforms with actin from different tissues. The molar ratio of bound dystrophin to actin calculated from the Coomassie blue-stained band intensities from SDS-PAGE. Actin from different tissues is represented as skeletal muscle actin (Sk-M), smooth muscle actin (Sm-M), cardiac muscle actin (Car-M), and non-muscle actin from human platelets (Non-M). Data labeled as Dys-M, Dys-P, and Dys-B correspond to the tandem CH domains of muscle, Purkinje, and brain isoforms. The error bars represent one standard error of the mean. p-Values ≤ 0.01 are represented as (**), p ≤ 0.001 are represented as (***), and p ≤ 0.0001 are represented as (****).

To test the statistical significance of the results obtained from co-sedimentation assays, data presented in Figure were analyzed using a two-way analysis of variance (ANOVA) with Tukey’s test for comparison of the means. The two-way ANOVA tested the impact of two independent variables, dystrophin isoform and actin source, on the dependent variable, which is the amount of dystrophin tandem CH domain bound to actin (Table ). This test also provided information on whether or not the interaction between the two independent variables affects the dependent variable. The results obtained from the ANOVA are summarized in Table . Results show that the effect of type of dystrophin isoform on actin–dystrophin binding was not dependent on the source of actin, i.e., the interaction effect from the ANOVA was not significant (p = 0.82). However, both variables, i.e., the type of dystrophin isoform and the tissue source of actin, influenced actin–dystrophin binding independently (p-value = 4.6 × 10–11 and 2.6 × 10–4, respectively). A comparison of means analysis showed that the tandem CH domain of Dys-M shows significantly weaker binding to F-actin compared to those of Dys-P and Dys-B (p-value <1 × 10–6 for both). Also, binding of dystrophin isoforms to smooth muscle actin is significantly stronger from that of skeletal muscle actin (p-value = 0.003), cardiac actin (p-value = 5.4 × 10–4), and non-muscle actin (p-value = 0.006) (Table ).

Table 3

Parameters Obtained from Two-Way ANOVA Analysis of Actin-Binding Assays (Figure )

variable	degrees of freedom (n – 1)	sum of squares	mean square	f-value	p-value
dystrophin tandem CH domain isoforms	2	0.21	0.11	37.53	4.58 × 10–11
actin source	3	0.06	0.02	7.53	2.55 × 10–4
interaction	6	0.01	0.001	0.48	0.82

Table 4

Parameters Obtained from Tukey’s Test for Comparison of Means across (a) Dystrophin Tandem CH Domain Isoforms and (b) Actin Source at 95% Confidence Intervala

(a)
groups compared	mean difference	standard error of the mean	q-value	p-value	lower control limit	upper control limit
Dys-P Dys-M	0.11	0.02	9.84	<1.00 × 10–5	0.07	0.15
Dys-B Dys-M	0.12	0.02	10.73	<1.00 × 10–5	0.08	0.16
Dys-B Dys-P	0.01	0.02	0.83	0.83	–0.03	0.05

In the table, Dys-M, Dys-P, and Dys-B correspond to the data on the tandem CH domains of muscle, Purkinje, and brain isoforms, respectively. Sk-M, Sm-M, Car-M, and Non-M correspond to actin from skeletal muscle actin, smooth muscle actin, cardiac muscle actin, and non-muscle actin from human platelets.

Discussion

Minor Variations in the N-Terminal Flanking Region Determine the Thermodynamic Stability of Dystrophin Tandem CH Domain Isoforms

According to the results presented in this study, tandem CH domains of dystrophin isoforms have different stabilities, with the tandem CH domain of Dys-M being more stable by ∼5 kcal/mol than those of Dys-P and the Dys-B. In terms of the length of the protein, a stretch of 11 amino acid residues at the N-terminus of muscle isoform is replaced with 7 and 3 different amino acids in Purkinje and brain isoforms, respectively (Figure b). Changes in the length of protein by introduction or deletion of amino acid residues at the N or C-terminus can affect the protein stability.[16,32−35] Also, another important determinant for protein stability, in general, is the N and C-terminal contact.[36,37] Dystrophin tandem CH domain exists in solution in a closed conformation with its N- and C-termini closer.[38] Thus, in the closed conformation, any residue changes at the termini of dystrophin tandem CH domain might impact protein stability by directly influencing the N–C contact, and our experimental results presented here on the dystrophin tandem CH domain support this hypothesis. Consistently, deletion of the N-terminal flanking region in the case of utrophin tandem CH domain, which exists in an open conformation,[39] results in marginal stability changes.[16]

Protein Conformation Defines Stability–Function Relationship in Dystrophin Tandem CH Domain Isoforms

The variations in the N-terminal flanking regions of dystrophin tandem CH domain isoforms are neither part of the known actin-binding sites (ABSs) on dystrophin (Figure b) nor they are part of any stable secondary structure.[40−44] Still, the results obtained in this study indicate that minor amino acid sequence variations at the N-terminus can significantly impact the stability and actin-binding function of tandem CH domains of dystrophin isoforms. The variations in these isoforms are part of the N-terminal flanking region of the CH1 domain (Figure b). Previous studies have shown that CH1 and CH2 domains of dystrophin contribute differentially toward the stability and function of its tandem CH domain.[13] Actin binding is primarily determined by the unstable CH1 domain, whereas stability is determined primarily by the CH2 domain, which does not bind to actin.[13] This seems like a general strategy adopted by other tandem CH domains, like utrophin and α-actinin, to fine-tune their stability and function by modulating inter-CH-domain interactions.[13−15,45] The full-length tandem CH domain of dystrophin is stabilized because of the tethering of an unstable CH1 domain to a stable CH2 domain and also due to N–C-terminal contact in a closed conformation.[38] A likely example of the importance of inter-CH-domain interactions and N–C contact in stabilizing tandem CH domains comes from utrophin, which, unlike dystrophin, exists in an open conformation.[39] Correspondingly, the utrophin tandem CH domain binds to F-actin with a higher affinity than that of dystrophin.[13,30] This leads to the hypothesis that the closed conformation provides stability to the tandem CH domain at the expense of its actin-binding function. Since the utrophin tandem CH domain exists in an open conformation, deleting the N-terminal flanking region in utrophin caused lesser stability changes (∼1.4 kcal/mol)[16] compared to the dystrophin tandem CH domain (∼5 kcal/mol) (Table ) that exists in a closed conformation.

The results obtained in this study agree with the above mentioned earlier studies. The Dys-M tandem CH domain is more stable than those of Dys-P and Dys-B (Figure and Table ) but has a lesser affinity toward binding to F-actin (Figure ). Lower stability of Dys-P and Dys-B tandem CH domains could stem from decreased N–C contact leading to more open conformation, which increases their actin-binding affinity. Comparison of m-values in urea denaturation profiles also showed lower m-values for tandem CH domains of Dys-P and Dys-B (Table ), suggesting a more open conformation for these isoforms as compared to that of Dys-M. A similar inverse relationship between the thermodynamic stability and actin-binding function has been observed for tandem CH domains of utrophin,[30,46] α-actinin,[47] filamin,[48] and spectrin.[49] Many mutations that destabilize the tandem CH domains have increased actin-binding affinity compared to their corresponding wild-type proteins.

An inverse relationship between thermodynamic stability and actin-binding function of tandem CH domains can be quantitatively correlated only when we can estimate the thermodynamic stability of the CH1 domain in the tandem CH domain, since CH1 controls the actin-binding function of tandem CH domains.[13,14] Denaturant melts of tandem CH domains, such as those in Figure A,B, should in principle be analyzed using a three-state equation with an intermediate state where only the CH1 domain is unfolded in between the native state (with both CH domains folded) and unfolded state (with both CH domains unfolded).[16,28] In such a three-state analysis, the first transition corresponds to the unfolding of the CH1 domain and the breaking of the CH1–CH2 interactions, whereas the second transition corresponds to the unfolding of the CH2 domain. In the case of the utrophin tandem CH domain, which exists in an open conformation with minimal inter-CH-domain interactions,[13,30,39,46] the difference in free energy of the first transition upon N-terminal modifications was nearly the same as the change in actin-binding free energy.[16] Such analysis cannot be done on the dystrophin tandem CH domain, which exists in a closed conformation with significant inter-CH-domain interactions,[13,30,38] because the first transition in the denaturant melt corresponds to the sum of the stabilities of the CH1 domain and the CH1–CH2 interactions that differ between the three isoforms (see the Supporting Information).

Tissue-Specificity of Tandem CH Domain–Actin Interactions

Dystrophin tandem CH domain isoforms bind to smooth muscle actin with a higher affinity than actin from other tissues. Actin from smooth muscles is composed mostly of γ-smooth muscle actin (85%) and β-actin (15%). Non-muscle actin is composed of β-actin (85%) and γ-cytoactin (15%). Actin from other tissues is composed of majorly α-actin. The results presented here indicate that dystrophin isoforms are able to distinguish γ-smooth muscle actin from other actin isoforms. Previous studies with the utrophin tandem CH domain have shown that it can distinguish β-actin from α-actin and binds more efficiently to β-actin.[50] Similar studies showed a broader difference between muscle and non-muscle actins binding to profilin,[51] thymosin β4,[52] ezrin,[53] and plastin.[54]

In summary, the results obtained in this study suggest that naturally occurring tandem CH domains of dystrophin isoforms utilize minor sequence variations in their N-terminal flanking regions to modulate their stability, actin-binding function, and tissue specificity of actin–dystrophin interactions. Brain and Purkinje isoforms have lesser stability compared to the muscle isoform. Stability and actin-binding affinity follow an inverse correlation with less stable brain and Purkinje isoforms binding with a higher affinity to actin. Protein conformation, probably dictated by the N–C contact, seems to play a major role in defining the stability and actin-binding efficiency of tandem CH domains. Proteins in closed conformation have higher stability and a lower affinity for actin, while proteins in open conformation have lower stability and a higher affinity for actin. In addition, tandem CH domains of dystrophin isoforms were also able to distinguish smooth muscle actin from other isoforms of actin. This differential interaction with actin isoforms could be useful in understanding diverse symptoms of dystrophinopathies, which affect tissues other than muscle and could help in designing therapeutics better suited for such symptoms.

Materials and Methods

Cloning, Expression, and Purification of Tandem CH Domains of Dystrophin Isoforms

Genes encoding the tandem CH domains of dystrophin isoforms (muscle, brain, and Purkinje) were synthesized by Gene Universal and cloned into a pET-SUMO expression vector using BamH1 and Xho1 sites. The original pET-SUMO vector leaves an additional serine residue at the N-terminus after the expressed protein was cleaved with the Ulp1 protease[55] and hence the codon corresponding to the serine residue was deleted using site-directed mutagenesis. The cloned genes were transformed into E. coli BL21 (DE3) cells. The sequence was confirmed after isolating the plasmid from the transformed cells using a plasmid miniprep kit from Qiagen. The bacterial constructs encoding the three proteins were grown in lysogeny broth (LB) media at 37 °C, and protein expression was induced with 0.5 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) at a cell density corresponding to OD600 nm of 0.5–0.8. All of the three proteins were expressed in soluble fraction and purified using a nickel-nitrilotriacetic acid (Ni-NTA) affinity column. The N-terminal polyhistidine-SUMO tag was cleaved using Ulp1 protease leaving no additional amino acids at the N-terminus, and pure proteins were eluted from the Ni-NTA column in the flow-through. Pure proteins were dialyzed in phosphate-buffered saline (PBS) (0.1 M sodium phosphate, 0.15 M NaCl, pH 7.0). The homogeneity of the proteins was checked using SDS-PAGE and native-PAGE.

Fluorescence and Circular Dichroism (CD) Spectroscopy

The fluorescence spectra of tandem CH domains of three dystrophin isoforms in PBS were recorded at a concentration of 1 μM on a Quantamaster, PTI fluorimeter. The samples were excited at 295 nm, and the emission was recorded from 305 to 400 nm. CD spectra for the three proteins were recorded at a concentration of 10 μM on an Applied Photophysics Chirascan Plus spectrometer.

Urea Denaturation Melts

Urea denaturation melts were performed at a protein concentration of 1 μM using both CD and fluorescence measurements. To obtain the Gibbs free energy of unfolding in the absence of the denaturant (ΔG°unf) and the linear slope of the variation of Gibbs free energy ΔGunf with denaturant concentration (m-value), changes in ellipticity at 222 nm and fluorescence emission at 320 nm (295 nm excitation) as a function of urea concentration were fitted to a two-state equilibrium unfolding model,[56] using the equationwhere SD is the measured signal as a function of denaturant concentration [D], SN and SU are the intrinsic signals corresponding to the native and unfolded states in the absence of the denaturant, mN and mU are the slopes of linear dependence of SN and SU on denaturant concentration, R is the universal gas constant, and T is the absolute temperature in kelvin.

Midpoint denaturant concentration, Cm, in the denaturant melt was estimated by modifying the above eq as

Thermal Denaturation Melts

Temperature denaturation melts for tandem CH domains were recorded at a protein concentration of 1 μM by monitoring the changes in ellipticity at 222 nm with an increase in the solution temperature. The temperature ramp rate was kept at 1 °C/min. To calculate the midpoint melting temperature (Tm), data were fitted to a two-state unfolding equation[57,58]where ΔHm is the enthalpy change at Tm and ΔCp is the heat capacity change between the native and unfolded states. ΔCp was set to zero for the calculation of Tm from thermal melts, which simplifies the above equation to

Since thermal melts of dystrophin tandem CH domains are irreversible because of aggregation at higher temperatures, only Tm values were considered for comparison between the three isoforms.

Actin-Binding Assays

Actin from different sources (skeletal muscle, smooth muscle, cardiac muscle, and platelets) was purchased from Cytoskeleton Inc. For obtaining binding curves, skeletal muscle actin was polymerized in polymerization buffer (50 mM KCl, 2 mM MgCl2, 1 mM adenosine 5′-triphosphate (ATP), and 10 mm Tris, pH 7.5), and 7 μM was incubated with dystrophin tandem CH domains ranging in concentration from 2 to 60 μM. The reaction mixture was centrifuged at 100 000g for 25 min. Pellets were resuspended in 30 μL of SDS-PAGE loading buffer, 15 μL of which was used for gel electrophoresis. Bands of actin and dystrophin tandem CH domains were stained with Coomassie brilliant blue R-250. The amount of dystrophin bound to actin was calculated by measuring the ratio of band intensities of dystrophin to actin using the Quantity One software on Bio-Rad Gel Doc XR. The band intensities were corrected for differential staining of proteins with Coomassie blue, as described before.[14,59] For determining the maximal number of binding sites (Bmax) and the dissociation constant (Kd), actin-binding assays were fit to the equationwhere y is the molar ratio of the bound tandem CH domain to actin and x is the unbound tandem CH domain concentration. The data fitted well with Bmax fixed to 1, as observed previously for dystrophin tandem CH domains. For statistical analysis, 7 μM polymerized actin from different sources was incubated at room temperature with 60 μM of tandem CH domains in a reaction volume of 100 μL. The reaction mixture was processed as described above to obtain the amount of dystrophin bound to actin. The data was presented as the mean of at least five to seven independent experiments and was analyzed by two-way ANOVA using the Origin Pro version 8 software to find the statistical significance of the effect of two variables, actin tissue source and dystrophin isoform, on actin–dystrophin binding affinity. Tukey’s test was used for comparing the means of individual groups.