Structural Basis for the Effective Myostatin Inhibition of the Mouse Myostatin Prodomain-Derived Minimum Peptide

Abstract

Myostatin inhibition is one of the promising strategies for treating muscle atrophic disorders, including muscular dystrophy. It is well-known that the myostatin prodomain derived from the myostatin precursor acts as an inhibitor of mature myostatin. In our previous study, myostatin inhibitory minimum peptide 1 (WRQNTRYSRIEAIKIQILSKLRL-amide) was discovered from the mouse myostatin prodomain. In the present study, alanine scanning of 1 demonstrated that the key amino acid residues for the effective inhibitory activity are rodent-specific Tyr and C-terminal aliphatic residues, in addition to N-terminal Trp residue. Subsequently, we designed five Pro-substituted peptides and examined the relationship between secondary structure and inhibitory activity. As a result, we found that Pro-substitutions of Ala or Gln residues around the center of 1 significantly decreased both α-helicity and inhibitory activity. These results suggested that an α-helical structure possessing hydrophobic faces formed around the C-terminus is important for inhibitory activity.

Myostatin, a transforming growth factor β (TGF-β) superfamily protein, is a promising target for treating muscle atrophic disorders such as muscular dystrophy, cancer cachexia, and sarcopenia. As previously reported, the lack of myostatin leads to a significant increase in muscular mass, whereas its overexpression induces cachexia.^1,2 Based on these findings, several strategies to induce muscular growth have been investigated. In particular, much attention has been focused on myostatin inhibitory molecules, including neutralizing antibodies,³ prodomain proteins,^4,5 soluble decoys of active type II receptors (ActRII),⁶ interacting proteins (growth and differentiation factor-associated serum protein, GASP),⁷ follistatin,⁸ and follistatin-related protein.^5,9 Treatment with neutralizing antibodies increased muscle mass and strength in Duchenne muscular dystrophy model mdx mice,³ while decoy receptors prevented muscle wasting and cancer-induced cardiac atrophy in tumor-bearing mice.⁶

It is well-known that the prodomain of myostatin interacts with mature myostatin to form an inactive complex in the extracellular matrix and in serum.^2,5,10 In a previous structural study of TGF-β1, the N-terminal α-helical region of its prodomain was found to be buried in the type I receptor-binding pocket of the mature domain.¹¹ Indeed, Walton et al. reported that aliphatic (Ile and Leu) residues in the α-helical region of the prodomain plays a key role in TGF-β1 inactivation.¹² This α-helical region, which includes aliphatic residues, is highly conserved throughout the TGF-β superfamily. In 2004, Jiang et al. reported that the human myostatin prodomain-derived fragment, which consists of 74 residues (positions: 19–92 in the prodomain), shows the significant myostatin inhibitory activity as a glutathione S-transferase fusion protein.⁴ As shown in Figure 1, our group recently identified a mouse myostatin prodomain-derived minimum peptide 1 (23 residues, including a conserved α-helical region spanning positions 21–43)¹³ through the earlier discovery of core inhibitory fragment (p29) consisting of 29 residues (position: 19–47).¹⁴ Intramuscular injection of these prodomain-derived peptides, including p29, significantly increased muscle mass in muscular dystrophy model mdx mice. Additionally, we determined the importance of Trp residue (position 21) and the mouse-derived Arg-Tyr sequence (positions 26 and 27) in the synthetic peptide inhibitor 1 using a cell-based luciferase reporter assay. Next, we carried out a structure–activity relationship (SAR) study of 1 in order to develop a more potent derivative inhibitor. Our first SAR study focused on the N-terminal Trp of minimum peptide 1 yielded the derivative 2, which is three times as potent as 1, registering an IC₅₀ value of 1.2 μM on the reporter assay (Figure 1).¹⁵ However, the structural basis for effective myostatin inhibition by the mouse-derived minimum peptide 1 is still unclear. Here, we performed Ala scanning of 1 to identify other key residues and investigated the relationship between secondary structure and inhibitory activity by synthesizing Pro-substituted peptides.

Figure 1.

Structures of myostatin inhibitory peptides p29, 1, and 2, derived from the prodomain of mouse myostatin.¹³⁻¹⁵ Numbers added to the top of amino acid indicate the position in the prodomain sequence of mouse myostatin.

In this study, all peptides were synthesized by the 9-fluorenylmethoxycarbonyl (Fmoc)-based solid-phase peptide synthesis (Fmoc-SPPS) method as previously reported.¹³ Prepared resins bearing protected peptides were treated with trifluoroacetic acid (TFA)-m-cresol-thioanisole-1,2-ethanedithiol (EDT) (4.0 mL, 40:1:1:1) for 150 min at room temperature. The resulting crude peptides were purified by preparative reversed-phase high-performance liquid chromatography (RP-HPLC). All peptides were characterized by electrospray ionization time-of-flight mass spectrometry (ESI-TOF MS) (see Supporting Information). Each peptide was at least 95% pure, as determined by RP-HPLC analysis at 230 nm.

The myostatin inhibitory activities of the peptides were measured by a luciferase reporter assay as previously reported (see Supporting Information).¹⁵ Briefly, premixed 8 ng/mL (0.32 nM) recombinant human myostatin, with or without synthetic peptides, was added to human embryonic kidney 293 (HEK293) cells for 4 h. These cells were transiently transfected with myostatin-responsive (SBE)₄-luc reporter beforehand. The myostatin-dependent reporter activity was measured by luminometer. Recombinant protein of mouse myostatin prodomain at 10 nM was used as a positive control. The results for each peptide are presented as a relative value compared to the reporter activity of myostatin alone (=100).

Secondary structure analysis of peptides was performed by measuring circular dichroic (CD) spectra as previously reported (see Supporting Information).¹³ Briefly, peptides were dissolved to yield a final concentration of 5 μM in 20 mM sodium phosphate buffer (pH 7.4) containing 10% 2,2,2-trifluoroethanol (TFE). CD spectra of the synthetic peptides were measured at 25 °C.

To identify the key residues for the effective myostatin inhibition of 1, we synthesized 22 different Ala-substituted peptides (Figure 2A) and evaluated their inhibitory activities (Figure 2B). Peptide 1 at 10 μM blocked the myostatin activity completely, performing similarly to the prodomain protein. Similarly, R22A, Q23A, N24A, T25A, R26A, S28A, R29A, E31A, K34A, Q36A, S39A, K40A, and R42A also suppressed the myostatin-dependent reporter activity, suggesting that side chain structures of these residues are less relevant to the inhibitory activity. However, the inhibitory activity of W21A was not observed as expected from our previous report.¹³ Focusing on the mouse-derived Arg-Tyr sequence at positions 26 and 27 of 1, Ala substitution of Tyr at position 27 (Y27A) only resulted in a significant decrease in myostatin inhibitory activity (Figure 2B). This decrease suggests that the mouse-derived Tyr enhanced the inhibitory activity compared with human-derived Ser. Interestingly, sequence alignment focused on peptide 1 showed that rat and rabbit, in addition to mouse among several species (Figure S1), only possess Tyr at position 27, implying that searching in the rodent sequence with this specific Tyr led us to the successful identification of synthetic peptide inhibitor 1 in our previous study.¹³

Figure 2.

(A) Structures of Ala-substituted peptides. The numbers above each amino acid indicate its position in the prodomain sequence of mouse myostatin. (B) Luciferase reporter assay to determine the ability of Ala-substituted peptides to inhibit myostatin relative to peptide 1. Cell line, HEK293; peptide concentration, 10 μM; positive control (prodomain, mouse recombinant) concentration, 10 nM; myostatin concentration, 8 ng/mL (0.32 nM); incubation, 4 h. Values represent means ± SD (n = 3).

In the C-terminal region, the importance of all aliphatic residues (Ile and Leu) for the effective myostatin inhibition of 1 was demonstrated by the results of I30A, I33A, I35A, I37A, L38A, L41A, and L43A (Figure 2B). Walton et al. had already reported that the five hydrophobic residues corresponding to positions 30, 37, 38, 41, and 43 in 1 contribute to TGF-β1 inactivation.¹² In the present study, we discovered the necessity for other residues, namely, Ile at positions 33 and 35. The inhibitory activities of I33A and I35A were quite weak at 10 μM, suggesting that Ile residues at positions 33 and 35 are indispensable for the myostatin inhibition by peptide 1.

Next, we focused on the rodent-specific Tyr at position 27 and synthesized a series of peptide derivatives to perform a SAR study (Figure 3A). As shown in Figure 3B, Y27F, Y27y (small letter: d-form), and Y27W at 10 μM were able to significantly inhibit myostatin activity, whereas Y27H, Y27Q, Y27R, and Y27E showed weak inhibition. Moreover, the potency of the former three peptide derivatives is almost identical since no significant difference was observed among them at lower concentrations (3 μM) (data not shown). These results suggested that the position 27 can accommodate aromatic hydrophobic amino acids bearing phenyl or indolyl group at side chain, i.e., Tyr, Phe, and Trp,¹⁶ without any restriction on either l- or d-isomers.

Figure 3.

(A) Structures of Tyr-substituted peptides. The numbers above each amino acid indicate its position in the prodomain sequence of mouse myostatin. (B) Luciferase reporter assay to determine the ability of of Tyr-substituted peptides to inhibit myostatin relative to peptide 1. Cell line, HEK293; peptide concentration, 10 μM; positive control (prodomain, mouse recombinant) concentration, 10 nM; myostatin concentration, 8 ng/mL (0.32 nM); incubation, 4 h. Values represent means ± SD (n = 3).

To investigate the relationship between secondary structure and inhibitory activity, we measured the CD spectra of these peptide derivatives. We previously reported that peptide 1 tends to form an α-helix structure in the presence of TFE.¹³ Similarly, a series of peptide derivatives also displayed the characteristic absorptions at 208 and 222 nm in the presence of 10% TFE (Figure S2), indicating that the secondary structure of 1 is not greatly affected by the amino acid substitution at position 27. Previously, we also reported that the N-terminal fragment of 1 forms a random coiled structure.¹³ This finding and our present results, including Y27y, suggest that the myostatin inhibitory activity of peptide is dependent on the side chain structure at position 27.

Moreover, to examine the relationship between the secondary structure and the inhibitory activity, we synthesized five different peptide derivatives, N24P, S28P, A32P, Q36P, and K40P, each with a potentially α-helix-breaking Pro substitution targeting every fourth residue in peptide 1 from position 24 to 40. These five residues were selected from among Ala-substitutable amino acids (Figure 4A).¹⁷ As shown in Figure 4B,C, the myostatin inhibitory activities and secondary structures of these peptides were evaluated similarly using the reporter assay and CD spectra measurement, respectively. In the N-terminal random coiled region of 1, Pro substitutions at positions 24 and 28 did not significantly decrease their α-helicity, whereas the latter led to a moderate loss of inhibitory activity. These results implicate that the side chain orientation of Tyr at position 27 in S28P may be improperly fixed by the introduction of Pro at the neighboring position 28. In the C-terminal α-helical region, Pro substitutions at positions 32, 36, and 40 broke the α-helix structure of 1 (Figure 4C). However, a significant decrease in inhibitory activity was observed in A32P and Q36P although the influence of broken secondary structure on the inhibitory activity was limited in K40P (Figure 4B). These results suggested that the α-helix structure formed around position 32–36 of peptide 1 plays an especially important role in effective myostatin inhibition.

Figure 4.

(A) Structures of Pro-substituted peptides. The numbers above each amino acid indicate its position in the prodomain sequence of mouse myostatin. (B) Luciferase reporter assay to determine the ability of of Pro-substituted peptides to inhibit myostatin relative to peptide 1. Cell line, HEK293; peptide concentration, 10 μM; positive control (prodomain, mouse recombinant) concentration, 10 nM; myostatin concentration, 8 ng/mL (0.32 nM); incubation, 4 h. Values represent means ± SD (n = 3). (C) CD spectra of Pro-substituted peptides in the presence of 10% TFE solution prepared in 20 mM sodium phosphate buffer (pH 7.4); peptide concentration, 5 μM.

Based on these findings, we finally drew a predicted α-helical 3D model (MOE, Molecular Operating Environment System) and 2D wheel focused on positions 26–43 of 1 similarly to a previous report concerning TGF-β1.¹² Additionally, we highlighted the aliphatic residues that form the major and minor hydrophobic faces in blue and red, respectively (Figure 5). Interestingly, Leu at position 43 (highlighted in purple) was not included in both hydrophobic faces, although it was determined as one of key residues in Ala scanning as shown in Figure 2. Hence, this residue may relate to other interaction with myostatin. Additionally, although Leu at position 41 seems to participate in forming the major hydrophobic face, the myostatin inhibition by K40P only slightly decreased compared with 1 without the structural influence of breaks in its α-helical nature as mentioned above (Figure 4). These considerations imply that a real major hydrophobic face in an active form of 1 is composed of three Ile residues at positions 30, 33, and 37 and that the spatial restriction of Leu residues at positions 41 and 43 is not so strong. Furthermore, the necessity for Ile at position 35 of 1 was first determined in this study since it corresponds to Gly in TGF-β1. Hence, it raises the possibility that the second (minor) hydrophobic phase is formed by Leu at position 38 as shown in Figure 5.

Figure 5.

(A) Predicted 3D helical model and (B) 2D helical wheel focused on positions 26–43 in peptide 1. The hydrophobic face formed by aliphatic residues 30, 33, 37, and 41 is highlighted in blue; another hydrophobic face based on residues 35 and 38 is highlighted in red. Leu at position 43 is highlighted in purple.

In conclusion, Ala scanning of a mouse-derived myostatin inhibitory minimum peptide 1 afforded the key residues at positions 21, 27, 30, 33, 35, 37, 38, 41, and 43 for the effective inhibition. In the N-terminal random coiled region, the rodent-specific Tyr at position 27 could be substituted with other aromatic hydrophobic residues such as Phe and Trp and the d-form of Tyr without affecting the secondary structure. In the C-terminal region, the α-helix formed around positions 32–36 was especially important for effective myostatin inhibition by 1. These findings would be valuable for future SAR studies to discover and develop more potent myostatin inhibitors.

Glossary

ABBREVIATIONS

ActRII

activin type II receptor

CD

circular dichroism

DIPCI

N,N′-diisopropylcarbodiimide

DMEM

Dulbecco’s modified Eagle’s medium

DMF

N,N-dimethylformamide

DMSO

dimethyl sulfoxide

EDT

1,2-ethanedithiol

ES

electrospray

FBS

fetal bovine serum

Fmoc

9-fluorenylmethoxycarbonyl

GASP

growth and differentiation factor-associated serum protein

HEK293

human embryonic kidney 293

HOBt

1-hydroxybenzotriazole

MOE

Molecular Operating Environment

MS

mass spectrometry

RP-HPLC

reverse-phase high-performance liquid chromatography

SAR

structure–activity relationship

SBE

Smad binding element

SPPS

solid-phase peptide synthesis

TFA

trifluoroacetic acid

TFE

2,2,2-trifluoroethanol

TGF-β

transforming growth factor-β

TOF

time-of-flight

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.6b00420.

• Materials, experimental procedure, analytical data for all peptide derivatives, analytical HPLC chromatograms, and Figures S1–S2 (PDF)

The authors declare no competing financial interest.