Discovery of Hydrocarbon-Stapled Short α-Helical Peptides as Promising Middle East Respiratory Syndrome Coronavirus (MERS-CoV) Fusion Inhibitors.

The hexameric α-helical coiled-coil formed between the C-terminal and N-terminal heptad repeat (CHR and NHR) regions of class I viral fusion proteins plays an important role in mediating the fusion of the viral and cellular membranes and provides a clear starting point for molecular mimicry that drives viral fusion inhibitor design. Unfortunately, such peptide mimicry of the short α-helical region in the CHR of Middle East respiratory syndrome coronavirus (MERS-CoV) spike protein has been thwarted by the loss of the peptide's native α-helical conformation when taken out of the parent protein structure. Here, we describe that appropriate all-hydrocarbon stapling of the short helical portion-based peptide to reinforce its bioactive secondary structure remarkably improves antiviral potency. The resultant stapled peptide P21S10 could effectively inhibit infection by MERS-CoV pseudovirus and its spike protein-mediated cell-cell fusion; additionally, P21S10 exhibits improved pharmacokinetic properties than HR2P-M2, suggesting strong potential for development as an anti-MERS-CoV therapeutic.

Introduction

Class I viral fusion proteins, such as those of the corona-, retro-, filo-, orthomyxo-, and paramyxoviruses, share similarities in their apparent use of a trimer of α-helical hairpins, or six-helix bundle (6-HB), to mediate membrane fusion processes of virus to target cell.[1,2] In the 6-HB viral fusion apparatus, the C-terminal heptad repeat (CHR) segment of the trimeric virus glycoproteins zips up into three α-helices along the conserved hydrophobic grooves on the periphery of an internal three-stranded coiled coil formed by the N-terminal heptad repeat (NHR) region of those fusion proteins.[3,4] Despite the topologically similar conformation of the fusion-promoting hexameric helical scaffold, their α-helical portion of the CHR domain displays dramatic differences in length. For example, the fusion core structure of HIV, a Retroviridae family member, contains regular α-helical CHR domains involving approximately 10 helical turns that bind to the central trimeric NHR helices.[5,6] In contrast, some other class I enveloped viruses, including Middle East respiratory syndrome coronavirus (MERS-CoV) of the Coronaviridae family, encode a notable dimorphism in the length of its NHR and CHR helices.[7−10] In the MERS-CoV 6-HB assembly, the NHR trimers have ∼21 helical turns, whereas the central CHR helices are limited to ∼4.5 turns within the longer heptad-repeat sequence (Figure A).[11]

Figure 1

Crystal structure of the MERS-CoV six-helix bundle (6HB) fusion core and the design of peptides based on the interaction between the NHR and CHR domains. (A) Cartoon representations of the MERS-CoV and HIV 6HBs, in which the NHR trimers and CHR segments are colored in gray and green, respectively. In the HIV 6HB structure (PDB 1AIK), the CHR segments form regular helices that pack against the central NHR core. In contrast, only a short helical domain is found within the MERS-CoV CHR (PDB 4NJL). (B) Helical wheel representation of the 6HB. In the CHR, the residues at the a–d positions (yellow) in direct contact with the NHR domains are buried in the 6HB. (C) Interaction between the NHR and CHR peptides, as well as the designed P21 peptide. The dashed lines between the NHR and CHR domains indicate the interaction between the residues located at the a–d positions in the CHR and the e–g positions in the NHR to form the 6HB. The helical domain sequence in MERS-CoV CHR is highlighted in red.

Blocking α-helix-mediated NHR/CHR interactions using CHR-derived peptides that target the transiently exposed NHR coiled coil is a promising approach to inhibit membrane fusion and viral infection.[12,13] With this strategy, a family of decoy α-helices structurally mimicking the lengthy C-terminal helices of HIV was identified as low-nanomolar inhibitors of viral entry.[14,15] It was disappointing that a synthetic peptide spanning the α-helical region of the MERS-CoV CHR domain sequence showed no anti-MERS-CoV activity.[11,16] Although the typical α-helical region is an important recognition motif for MERS-CoV fusion core formation, this short CHR peptide alone does not retain its native conformation owing to the lack of structural reinforcement provided by the parent protein, which, in turn, retards its binding to partner NHR helices.[11]

Peptide stapling based on different macrocyclization chemistry is a key technique for constraining short peptides in α-helical structures and thereby preorganizing them into their bound conformations to modulate helix-mediated protein–protein interactions with reduced entropic penalty.[17,18] Among numerous stapling chemistries, ruthenium-catalyzed RCM is one of the most noteworthy methods.[19,20] This method of producing all-hydrocarbon-stapled α-helical peptides affords high levels of α-helix induction and nucleation, thus directly resulting in peptide target-binding proclivity, cell permeability, and serum stability.[21,22] Furthermore, the therapeutic potential of stapled peptides has been showcased in a growing diversity of biological settings. Two hydrocarbon-stapled peptides developed by Aileron Therapeutics have reached clinical trials, epitomizing the early indications regarding the clinical translational potential of these stabilized α-helical peptidomimetics.[23,24]

In the present study, we employed hydrocarbon stapling to recapitulate the topography of α-helical motifs found in the MERS-CoV CHR region to inhibit MERS-CoV infection and its spike (S) protein-mediated cell–cell fusion. In the process of developing these stapled α-helical peptides, we performed a detailed study to identify the optimal stapling sites and the effect of hydrophobic amino acid incorporation to generate promising therapeutics.

Design

The primary structure of the MERS-CoV S protein hexameric coiled-coil fusion complex exhibits a heptad repeat pattern usually denoted as a–b–c–d–e–f–g.[11,16] In structural terms, three NHR helices form a trimeric coiled-coil inner core using their a–d residues for self-association, while those at the e–g positions participate in the interhelical knob-in-hole packing interaction with residues at the a–d positions of CHR helices, which are considered critical for stabilizing the 6-HB structure (Figure B).[25,26] According to the crystal structure of the MERS-CoV fusion core, residues L1262 at position d to Y1280 at position a in the CHR domain adopt a canonical α-helical region, binding to the deep grooves of the NHR trimers. Seminal work investigating the stability of the coiled coil suggests that helices consisting of fewer than three heptads generally do not create enough opportunities for hydrophobic interactions in the core to favor superhelical assembly.[27,28] On the basis of the X-ray crystallography and coiled-coil sequence-to-structure relationship study, we engineered a peptide containing the three integral heptads, designated as P21, spanning residues 1260–1280 of the MERS-CoV S protein CHR as a template for inhibitor design (Figure C). Analysis of the crystal structure of CHR in complex with NHR helices shows that six residues at the a and d positions (L1262, M1266, L1269, V1273, L1276, and Y1280) are buried in the 6-HB. In addition to the hydrophobic interactions contributed by their side chains, residues L1262, M1266, V1273, and Y1280 form four hydrogen bonds with NHR-helical trimer residues, including Q1023, N1016, Q1009, and K1000, via their main-chain amino or oxygen groups. Meanwhile, the hydroxyl group of the Y1280 side chain accepts a hydrogen bond from the NH group of Q1009 in NHR helices.[11,16] Together with critical hydrophobic stacking forces, these hydrogen bonds constitute a polar contact network that further tightly ties CHR to NHR helix. We initially utilized the “staple scanning” approach along the length of the peptide helix, but away from the critical a–d residues, to identify the optimal stapling sites. We substituted pairs of (S)-2-(4-pentenyl)alanine (abbreviated as S5 residues) at the i and i+4 positions because it is the most widely used unnatural amino acid for such type of staple.[21] Accumulated evidence that underline the structure and stability of coiled-coils has shown that the superhelix stability can generally be improved by inserting residues with more hydrophobic side chains at the a–d positions.[28,29] Among the six buried a–d residues, Leu and Val, who possess high hydropathy indices,[30] can strongly bind to the grooves on the surface of the N-helix trimer through their fully hydrophobic side chain interactions but Met and Tyr are polar residues. On the basis of the scanning study results, we elected to create stapled structures with hydrophobic mutations in the buried a–d residues through individual substitution of M1266 or Y1280 with l-norleucine, an isosteric Met analogue with less polar and more hydrophobic properties,[31] or F, respectively, or a combination of both replacements, thus expecting to form a potential hydrophobic binding core.

Results and Discussion

Effect of Stapling on Inhibition of MERS-CoV S Protein-Mediated Cell–Cell Fusion

First, peptides P21S1–P21S10 were tested for their ability to inhibit MERS-CoV fusion with its target cell using our previously developed MERS-CoV S protein-mediated cell–cell fusion assay.[11] As shown in Table , the linear wild-type P21 exhibited no inhibition at a concentration of up to 50 μM but when the peptides were cyclized markedly improved inhibitory activity was demonstrated. Among these α-helical mimetics, P21S8 and P21S10 strongly inhibited S protein-mediated cell–cell fusion, with EC50 values of 0.26 and 0.33 μM, respectively, which are even more potent than the most active MERS-CoV fusion inhibitor so far, namely the 36-mer peptide HR2P-M2.[11] Importantly, their unmetathesized analogues, i.e., peptides containing un-cross-linked S5 residues had a dramatically decreased membrane fusion inhibitory potency (Table ). These results reveal the critical contribution of hydrocarbon stapling to replicate the local topography of MERS-CoV short C-terminal helices for therapeutic benefit. In addition, shifting the position of the all-hydrocarbon staple had a significant effect on anti-MERS-CoV activity. Peptides P21S2, P21S4, P21S5, and P21S9, with a 2/6, 5/9, 8/12, and 15/19 staple, respectively, exhibited only similar activity compared to their relevant unstapled peptides, whereas peptides P21S1, P21S3, P21S6, and P21S7 had no activity. Compared to P21S8 whose cross-link had the S,S-configuration, its R,R-configured counterpart, P21R8, provided 62-fold weaker anti-MERS-CoV activity, consistent with previous publications revealing the requirement of S-form building blocks for an i, i+4 staple type.[21] Finally, we explored whether the incorporation of hydrophobic mutations at the binding interface could lead to an improvement in biological activity. In our assay, replacing M1266 with l-norleucine (i.e., Z), which led to P21S8Z, retained the high activity of P21S8 whereas Y1280F alone and the combination of M1266Z and Y1280F showed decreased potencies compared to that of P21S8, indicating the critical role of hydrogen bond formation between Y1280 and Q1009 of the S protein NHR trimer. Although buried hydrophobic residues may contribute more energy toward stabilization of a coiled-coil structure than polar interactions,[28,32] subtle specific interacting networks are required for hydrocarbon-stapled peptides to stop the MERS-CoV–cell fusion process efficiently.

Table 1

Inhibitory Activities of Stapled Peptides on MERS-CoV S Protein-Mediated Cell–Cell Fusion

compd	sequencea	EC50 (μM)e
Hydrocarbon Stapled Peptidesb
P21S1	*DLT*EM LSLQQVV KALNESY	>50
P21S2	L*LTY*M LSLQQVV KALNESY	3.90 ± 1.1
P21S3	LDL*YEM *SLQQVV KALNESY	>50
P21S4	LDLT*EM L*LQQVV KALNESY	7.14 ± 0.7
P21S5	LDLTYEM *SLQ*VV KALNESY	10.7 ± 2.6
P21S6	LDLTYEM L*LQQ*V KALNESY	>50
P21S7	LDLTYEM LSL*QVV *ALNESY	>50
P21S8	LDLTYEM LSLQ*VV K*LNESY	0.26 ± 0.05
P21S9	LDLTYEM LSLQQVV *ALN*SY	14.1 ± 2.3
P21S10	LDLTYEM LSLQQVV K*LNE*Y	0.33 ± 0.04
Unstapled Peptidesc
P21L2	LXLTYXM LSLQQVV KALNESY	10.9 ± 1.1
P21L4	LDLTXEM LXLQQVV KALNESY	8.21 ± 0.9
P21L5	LDLTYEM XSLQXVV KALNESY	4.49 ± 0.6
P21L8	LDLTYEM LSLQXVV KXLNESY	20.6 ± 3.3
P21L9	LDLTYEM LSLQQVV XALNXSY	10.9 ± 1.0
P21L10	LDLTYEM LSLQQVV KXLNEXY	3.55 ± 0.2
P21S8 Mutated Peptidesd
P21R8	LDLTYEM LSLQ∧VV K∧LNESY	16.3 ± 1.1
P21S8Z	LDLTYEZ LSLQ*VV K*LNESY	0.63 ± 0.05
P21S8F	LDLTYEM LSLQ*VV K*LNESF	2.16 ± 1.1
P21S8ZF	LDLTYEZ LSLQ*VV K*LNESF	3.89 ± 0.8
Control Peptides
P21	LDLTYEM LSLQQVV KALNESY	>50
HR2P-M2	SLTQINTTLLDLEYEMKKLEEVVKKLEESYIDLKEL	0.75 ± 0.09

These peptides have an acetyl group at the N-terminus and carboxyamide at the C-terminus.

Asterisks indicate the positions of the S5 residues, which react to form the all-hydrocarbon staple.

X indicates the positions of the S5 amino acids left uncyclized.

denotes the positions of the R5 amino acids, which react to form staples.

EC50 data were derived from the results of three independent experiments and are expressed as the mean ± standard deviation.

Stapled Peptides As Inhibitors of Pseudotyped MERS-CoV Infection

Subsequently, we tested the potential inhibitory activities of the P21S8, P21S10, and P21S8Z peptides, which exhibited promising inhibitory potency in the cell–cell fusion assay, on MERS-CoV pseudovirus infection in Huh-7 cells. As shown in Table and Supporting Information, Figure S1, P21S8, P21S10, and P21S8Z could significantly inhibit the MERS-CoV pseudovirus carrying the wild-type S protein, with EC50 values of 3.03, 0.97, and 2.80 μM, respectively. Although P21S8 exhibited 2.8-fold greater potency than the positive control HR2P-M2 in inhibiting cell–cell fusion, it was 2.8-fold less active than HR2P-M2 against wild-type MERS-CoV pseudovirus. Notably, P21S10 displayed potency similar to that of HR2P-M2 in both cell–cell and virus–cell fusion inhibition. These results suggest that, unlike P21S10, P21S8 does not have good correlation between its blockage of S protein-mediated membrane fusion and inhibition of wild-type MERS-CoV infection. One of the possible explanations for this discrepancy is that P21S8 may have lower target binding capability than P21S10 during the inhibition of virus infection. In addition, none of these viral fusion blockers displayed significant cytotoxicity to Huh-7 cells up to 100 μM (Table and Supporting Information, Figure S2). The selectivity index (SI = CC50/EC50) of P21S10 was greater than 103. Among clinical MERS-CoV strains, S protein amino acid changes of Q1020H and Q1020R located in the NHR domain were under positive selection and present in nearly all variants.[33] Therefore, we next investigated whether P21S8, P21S10, and P21S8Z would be effective against MERS-CoV strains with the Q1020H and Q1020R mutations. Strikingly, these peptides were effective for inhibiting infection by MERS-CoV pseudoviruses with mutated S protein, with EC50 values similar to those for inhibiting pseudovirus carrying the wild-type S protein (Table and Supporting Information, Figure S1). P21S8 and P21S8Z displayed about 3-fold less potency than HR2P-M2 for inhibiting infection by pseudoviruses carrying S protein with Q1020H or Q1020R mutation; however, P21S10 possessed potency similar to that of HR2P-M2. Because MERS-CoV is primarily infecting the human respiratory tract, we further tested the inhibitory activities of these stapled peptides on MERS-CoV pseudovirus infection in a human lung-derived cell line Calu-3. Strikingly, P21S8, P21S10, and P21S8Z could inhibit MERS-CoV pseudovirus infection in Calu-3 cells with EC50 values of 2.21, 1.58, and 2.57 μM, respectively, in an agreement with result observed in experiment using Huh-7 cells (Supporting Information, Table S1 and Figure S1). They also had no significant cytotoxicity to Calu-3 cells at the concentrations as high as 100 μM (Supporting Information, Figure S3). Moreover, we evaluated the aqueous solubility of these stapled peptides. P21S8 and P21S10 had aqueous solubilities of 6.47 and 10.5 mg/mL at pH 7.4, respectively, which were 14-fold and 22-fold higher than that of P21S8Z (Supporting Information, Table S2), suggesting their good potential for further development.

Table 2

Inhibitory Activity of P21S8, P21S10, and P21S8Z on Infection by Pseudotyped MERS-CoVa

	EC50 (μM) for inhibiting
compd	WT MERS-CoV pseudovirus	Q1020H-MERS-CoV pseudovirus	Q1020R-MERS-CoV pseudovirus	CC50 (μM)	SIb
P21S8	3.03 ± 0.29	4.06 ± 0.34	1.98 ± 0.28	>100	>33
P21S10	0.97 ± 0.08	1.82 ± 0.28	0.89 ± 0.07	>100	>103
P21S8Z	2.80 ± 0.74	4.15 ± 0.25	2.49 ± 0.18	>100	>36
HR2P-M2	1.07 ± 0.21	1.25 ± 0.18	0.64 ± 0.16	>100	>93

Data were derived from the results of three independent experiments and are presented as the mean ± standard deviation.

SI (selectivity index) = CC50/EC50 for inhibiting WT MERS-CoV pseudovirus infection.

Effect of Stapling on Helical Propensity

Next, we sought to determine whether hydrocarbon stapling could effectively recapitulate the native conformation of the α-helical region in the MERS-CoV CHR domain. CD spectroscopy provides a typical signature for α-helices with a maximum near 190 nm and double minima at 208 and 222 nm. CD spectra showed that P21 displayed a random coil conformation in phosphate buffer. Strikingly, with the exception of P21S2 that was too insoluble to make accurate measurements, the other hydrocarbon-stapled peptides exhibited α-helical structure with 15.6–54.5% helicity. Among them, the highly effective stapled peptides P21S8 and P21S10 were found to have greatly improved helical propensity, compared to the linear wild-type peptide, with helical contents of 53.7% and 46.5%, respectively (Figure ). Furthermore, these stapled peptides were indeed more helical than their correponding uncyclized linear precursors (Supporting Information, Table S3–S4). These data further confirmed the successful mimicry of the CHR helix via cyclization using the stapling strategy. Interestingly, although the CD spectra for peptides P21S8 and P21S10 were similar to those of P21S6 and P21S7, a difference in inhibitory potency of more than 193-fold was observed between them, suggesting that the much higher biological activity cannot be merely attributed to the helical propensity. Hence, the following target-binding studies were performed to further elucidate their mechanisim of action.

Figure 2

CD spectra of the isolated peptides P21S8 and P21S10. The corresponding unstapled counterparts and linear peptide P21 are included for structural comparison. The final concentration of each peptide in PBS (pH 7.2) was 50 μM.

Binding of Stapled Peptides to Their Complementary S Protein NHR Domain

Previous studies have shown that peptides derived from the MERS-CoV S protein CHR domain can bind to their NHR target to form a stable α-helix hexamer.[11] Here, we first sought to determine the interaction of the stapled helical peptides with their native NHR ligand, designated HR1P, to form stable α-helical complexes using CD spectroscopy. The significant increase over the sum of the individual spectra of two peptides mixed together in equimolar concentrations indicates that the interaction between them led to enhanced helical structures.[34,35] The isolated HR1P displays a random coil structure in phosphate buffer (pH 7.2) as judged by CD spectroscopy, which was consistent with previous reports.[11] Both P21S8 and P21S10 interacted with HR1P, resulting in 35.5% and 45.6% α-helical content in the complex, respectively. The CD signals of the P21S8/HR1P mixture and the P21S10/HR1P mixture at 222 nm were dramatically greater than the mathematical sum of those of the isolated peptides, indicating the induction of a large α-helical structure upon their interaction (Figure A). Strikingly, the thermal unfolding transition (Tm) values of P21S8/HR1P and P21S10/HR1P reached 71.1 and 76.4 °C, respectively (Figure B). Together, these CD data are consistent with P21S8/HR1P and P21S10/HR1P folding to form thermally stable and cooperatively folded helical bundles. Stapled peptides with moderate activity, including P21S4, P21S5, and P21S9, resulted in a less stable interaction with Tm values ranging from 56.6 to 66.6 °C while peptides with very poor antiviral activity showed no, or weak, coiled-coil interactions (Supporting Information, Table S4). These results demonstrated a strong correlation between the biological activities and the thermal stabilities of the stapled peptides/HR1P complexes.

Figure 3

Binding of P21S8 and P21S10 to the NHR peptide HR1P. (A) CD spectrum of peptide mixtures (Spec(C+N), solid symbols) and the sum of the spectra of the related isolated peptides (Spec(C) + Spec(N), open symbols) are shown for comparison. The hydrocarbon-stapled peptide/HR1P interaction induces more α-helix content than the sum of the single peptides. The final concentration of each peptide in PBS was 50 μM. (B) The thermostability of the complexes formed by P21S8 or P21S10 and the target mimic NHR peptide HR1P was determined by CD spectroscopy. The final concentration of each peptide in PBS was 50 μM. (C) N-PAGE analysis of the stapled peptides, HR1P, and their complexes. Lane 1, HR1P; lane 2, P21S8 + HR1P; lane 3, P21S8; lane 4, P21S10 + HR1P; lane 5, P21S10; lane 6, P21S9 + HR1P; lane 7, P21S9; lane 8, P21S3 + HR1P; lane 9, P21S3. (D) Sedimentation velocity analysis of the P21S8/HR1P mixture and the P21S10/HR1P mixture. The sedimentation coefficient (s) and the observed molecular mass (kDa) of each peak are indicated in parentheses.

The binding of the stapled peptides to the NHR-derived peptide HR1P was further confirmed using native (N)-PAGE. In N-PAGE, owing to its net positive charge, HR1P did not appear on the gel as it migrated into the solvent under the native electrophoresis conditions; these results are consistent with previous observations.[11] The respective stapled helical peptides with various antiviral potencies were mixed with HR1P and subjected to N-PAGE analysis. As shown in Figure C, both P21S8 and P21S10 formed stable complexes with HR1P, as evidenced by new bands at the upper position in the gel concomitant with the fading or disappearance of the stapled peptide bands, suggesting that oligomeric complexes formed via coiled-coil interactions. The N-PAGE results also indicated a stronger binding of either P21S8 or P21S10 to HR1P than the bindings of P21S4, P21S5, and P21S9 to HR1P based on the densities of the residue-stapled peptide bands observed in the gel (Figure C and Supporting Information, Figure S4). Consistent with the cell–cell fusion assay and CD spectroscopy results, the inactive peptide P21S1, P21S3, and P21S6 exhibited no NHR binding. Interestingly, the mixture of P21S7/HR1P did not show new band at the upper position in the gel, while the P21S7 band in the mixture had a much lower intensity than the corresponding P21S7 band alone (Supporting Information, Figure S4). This is in agreement with the CD data, where the signal of the P21S7/HR1P mixture was dramatically smaller than that of the mathematical sum of the isolated peptides, suggesting that addition of P21S7 to HR1P distorts the α-helical conformation of P21S7 due to their interaction. The combined CD and N-PAGE data indicate that P21S7 may associate with the viral N-helices to form nonhelical complexes or aggregates that cannot migrate in the gel, thus resulting in no inhibitory activity against MERS-CoV infection.

The sizes of the complexes formed between HR1P and P21S8 or P21S10 were further confirmed by sedimentation velocity analysis (SVA). As analyzed by SVA, the sedimentation coefficients of the P21S8/HR1P complex and the P21S10/HR1P complex were 2.27 and 1.94 s, corresponding to 22.6 and 22.3 kDa, respectively. Compared to the expected molecular masses of 7.07 kDa for the P21S8/HR1P heterodimer and 7.10 kDa for the P21S10/HR1P heterodimer, we concluded that P21S8 and P21S10 could both associate with HR1P to form a heterogeneous 6HB complex, thus blocking MERS-CoV entry into the host cells (Figure D and Supporting Information, Table S5).

Pharmacokinetic Studies

The pharmacokinetic behavior of P21S10, P21S8, and HR2P-M2 was comparatively evaluated in rats. Sprague–Dawley rats (200 ± 10 g) were injected intravenously with the three peptides (4 mg/kg). Seven blood samples were collected sequentially over a 240 min period (three animals at each time point), and the concentrations of the intact peptides in plasma were determined by LC/MS/MS analysis. As shown in Figure and Table , although these three peptides exhibited similar in vivo half-lives, the mean value of the maximum blood concentration (Cmax) for P21S10 was 3.6–5.8-fold greater than those for P21S8 and HR2P-M2. Meanwhile, we observed dramatic differences among the areas under the plasma concentration–time curve extrapolated to the last time point (AUC0–), a measure of the total systemic exposure to a drug, for the three compounds. The AUC0– was 91.3 μg/(mL·h) for P21S10 as compared to 44.1 μg/(mL·h) for P21S8 and 3.36 μg/(mL·h) for HR2P-M2. Consequently, P21S10 showed the lowest systemic clearance (0.041 (L/h)/kg), which was approximately 25-fold less than that of HR2P-M2 (1.04 (L/h)/kg). Thus, the current results indicate that P21S10 should be further developed as a potential drug candidate on the basis of its high potency in inhibiting infection of divergent MERS-CoV strains and its improved pharmacokinetics compared with HR2P-M2.

Figure 4

Pharmacokinetic profiles of P21S10, P21S8, and HR2P-M2 in plasma following the administration of a single intravenous dose (4 mg/kg) to Sprague–Dawley rats (n = 3).

Table 3

Pharmacokinetic Parameters of P21S10, P21S8, and HR2P-M2 in Rats Following a Single Dose iv Administration Calculated by Noncompartmental Analysis by Using DAS, version 3.2.8a

compd	AUC (0–t) (μg/(mL·h))	MRT (0–t) (h)	t1/2 (h)	CL ((L/h)/kg)	Cmax (μg/mL)	Vd (L/kg)
P21S10	91.3 ± 9.8	1.08 ± 0.04	1.15 ± 0.15	0.041 ± 0.006	98.7 ± 6.5	0.066 ± 0.004
P21S8	44.1 ± 1.0	1.03 ± 0.04	1.32 ± 0.49	0.082 ± 0.004	61.8 ± 13.6	0.156 ± 0.052
HR2P-M2	3.35 ± 0.5	1.34 ± 0.14	1.54 ± 0.73	1.04 ± 0.26	17.0 ± 1.3	2.188 ± 0.603

MRT, mean residence time; CL, clearance; Cmax, maximum drug concentration; Vd, volume of distribution.

Conclusions

We provide the first example of all-hydrocarbon-stapled peptides mimicking the native conformation of the C-terminal short α-helical region within the MERS-CoV S protein. These stapled peptides were able to block the formation of the hexameric coiled-coil fusion complex and thus inhibit viral–host cell membrane fusion. Our data showed that one of these stapled peptides, P21S10, not only preserves the biologically active α-helical conformation, as found in the MERS-CoV fusion machinery, but also has good potential to be developed into anti-MERS therapeutics based on its promising antiviral activity, effective native ligand-binding capability, and favorable pharmacokinetics. Formation of MERS-CoV 6-HB fusion core between CHR and NHR is basically the result of α-helix-mediated protein–protein interactions wherein the CHR helix can be considered to have two distinct faces, i.e., a buried hydrophobic surface and a solvent-exposed surface.[29] The established rules of thumb that underline the sequence and structure in coiled coils reveal that stabilization of individual helices will enhance the overall stability of the superhelical protein structural motif.[28] Therefore, strategies used for reinforcing the secondary structure of constituent α-helices in a coiled coil such as introducing intrahelical salt bridges at solvent accessible site and substitution of noncore residues with helix-favoring amino acids have been successfully applied to HIV fusion inhibitor design.[36] We envisioned that these strategies would also be applicable to the further optimization of our lead compounds found in this study. In addition, rather than generating hydrophobic mutations in buried residues, for future design one should also consider the place of polar amino acids at the interactive site of stapled peptides, which could establish attractive interhelical electrostatic interactions or hydrogen bonds, thus increasing their aqueous solubility and target-binding affinity.[37,38] Overall, considering the universal 6HB fusion mechanism employed by the class I enveloped viruses, our study establishes the paradigm that hydrocarbon stapling may be an attractive strategy for developing viral entry inhibitors designed to control viral epidemics through recapitulating certain α-helical peptides of the viral fusion core structures so that NHR/CHR coiled coil-mediated interactions are inhibited.

Experimental Section

General Peptide Synthesis

Peptides were prepared using a CEM Liberty automated microwave peptide synthesizer together with the Fmoc solid-phase peptide synthesis protocol on Rink-Amide resin with a loading capacity of 0.44 mmol/g. Briefly, the Fmoc protecting group of the resin was removed with 20% piperidine/DMF (2 × 30 min). Natural amino acids were coupled using O-benzotriazol-1-yl-N,N,N′,N′-tetramethyl-uronium hexafluorophosphate (3 equiv) in DMF and diisopropylethylamine (DIEA) (6 equiv) in NMP as coupling reagents, 3 equiv of Nα-protected amino acid, and 3 equiv of 1-hydroxy-benzotriazole. After each coupling, or Fmoc removal, the resin was washed with DMF (5 × 1 min) and DCM (3 × 1 min). For the assembly of olefin-containing amino acids, a double coupling method was performed to achieve a complete reaction, as described previously.[39] After final deprotection, the peptide-bound resin was treated with acetic acid anhydride–DIEA (1:1, v/v) (2 × 30 min). For synthesis of the all-hydrocarbon-stapled peptides, the RCM reactions of the protected peptides containing S5 ((S)-2-(4-pentenyl)alanine) or R5 ((R)-2-(4-pentenyl)alanine) were performed on-resin using the first-generation Grubbs catalyst. First, the reaction bottle containing peptide-bound resin was kept under lucifugal conditions by silver paper and then protected under nitrogen. Second, Grubbs I catalyst (0.3 equiv) in dichloroethane was added to the resin, the mixture was stirred for 6 h at room temperature, and then repeated treatments with fresh catalyst were carried out for 6 h to obtain more complete conversion. After the reaction solution was drained, the resin was washed with DMF (2 × 1 min) and DCM (2 × 1 min). The resin cleavage was performed by treatment with a cleavage cocktail containing 85% TFA, 5% thioanisole, 5% m-cresol, and 5% water. After cleavage from resin, crude peptides were precipitated using cold diethyl ether. The crude products were purified by preparative reversed-phase (RP) HPLC, and the purity of each compound was confirmed to be ≥95% by analytical RP-HPLC. Such information is provided in the Supporting Information (Tables S6–S7). The molecular weight of the peptides was confirmed by Q-FT-ICR-MS (Apex Qe, Bruker, Germany).

Inhibition of MERS-CoV S-Protein-Mediated Cell–Cell Fusion Assay

MERS-CoV S protein-mediated cell–cell fusion was assessed, as described previously.[40,41] In brief, 293T cells expressing MERS-CoV S protein and enhanced GFP (EGFP) were used as the effector cells (293T/MERS/EGFP), and Huh-7 cells expressing the MERS-CoV receptor DPP4 were used as the target cells. 293T/EGFP cells, which express only EGFP, were used as negative control cells. Huh-7 cells were plated in a 96-well plate (5 × 104 per well) and cultured at 37 °C for 5 h. Inhibitors at the indicated concentrations were then added, followed by the addition of 293T/MERS/EGFP cells or 293T/EGFP cells (1 × 104 per well). After coculture at 37 °C for 4 h, the 293T/MERS/EGFP cells and 293T/EGFP cells fused or unfused with Huh-7 cells were counted under an inverted fluorescence microscope (Nikon Eclipse Ti–S).

Inhibition of Pseudotyped MERS-CoV Infection

MERS-CoV pseudovirus was generated via cotransfection of 293T cells with the plasmid pNL4-3.luc.RE as well as the pcDNA3.1-MERS-CoV-S plasmid, as previously described.[42] Peptides at graded concentrations were mixed with the pseudovirus and then incubated for 1 h at room temperature. The mixture was added to Huh-7 or Calu-3 cells, fresh medium was added 12 h later, and the cells were incubated for an additional 48 h at 37 °C. Fluorescence was determined immediately using a luciferase kit (Promega) and an Ultra 384 luminometer (Tecan) after the addition of luciferase substrate (Promega).

Cytotoxicity Assessment

The cytotoxicity of compounds on Huh-7 cells used for testing pseudotyped MERS-CoV infection was measured using a cell counting kit (CCK-8, Dojindo, Kumamoto, Japan). Briefly, 5 μL of test peptide at graded concentrations were added to 100 μL of cells (1 × 105 per well) in wells of a 96-well plate incubated at 37 °C for 12 h before the addition. After incubation at 37 °C for 48 h, 100 μL of CCK-8 were added. The absorbance at 450 nm was measured 2 h later.

Aqueous Solubility Determination

Solubility was measured by using an HPLC-UV method. Stapled peptides (∼2 mg) were added to 1.5 mL Eppendorf tubes and either pH 7.4 phosphate buffer (100 μL) or double-distilled H2O (ddH2O) (100 μL) was added for dissolution with shaking for 24 h at 25 °C, followed by centrifugation of the mixture at 10000 rpm for 15 min. The saturated supernatant solution was filtered through a 0.45 μm filter membrane and then transferred to other vials for analysis by HPLC with UV detection. Each sample was assayed in triplicate. For quantification, analytical RP-HPLC was used with a Zorbax Eclipse XDB-C8 column (4.6 mm × 150 mm, 5 μm). Solvent A, 0.1% TFA in H2O; Solvent B, 0.1% TFA in 70% CH3CN/H2O; flow rate, 1 mL/min; gradient, 5–100% solvent B in solvent A over 25 min. The aqueous concentration was determined by comparison of the peak area of the saturated solution with a standard curve plotted for the peak area versus known concentrations, which was prepared by solutions of test compound in PBS or ddH2O at 20, 10, 5, 2.5, 0.5, and 0.05 mg/mL.

CD Spectroscopy

Peptides were dissolved in PBS (50 mM, pH = 7.2) to a final concentration of 50 μM. Equimolar mixtures of HR1P, and test peptides were incubated at 25 °C for 30 min at a concentration of 50 μM. The CD spectra were obtained on a Chirascan Plus spectropolarimeter (Applied Photophysics Ltd., Leatherhead, Surrey, UK) at 4 °C. The measurement parameters were set up as follows: wavelength, 190–260 nm; step resolution, 0.5 nm; speed, 20 nm/min. The CD data are shown as the mean residue ellipticity, and the [θ]222 value of −33000° cm2/dmol was taken to correspond to 100% α-helicity. Thermal denaturation was monitored at 222 nm by implementation of a thermal gradient of 2 °C/min from 10 to 90 °C.

N-PAGE

Stapled peptides and HR1P were dissolved in 50 mM PBS and ddH2O at a concentration of 200 μM, respectively. Each stapled peptide was incubated with HR1P at 25 °C for 30 min, respectively. After the addition of Tris-glycine native sample buffer (BioRad, Hercules, CA), the samples were loaded onto an 18% Tris-glycine gel with a Tris-glycine sample buffer (pH 8.3). Gel electrophoresis was carried out at a constant voltage of 120 V at room temperature for 3 h, and then the gel was stained with Coomassie Blue R250. The images were obtained by the ChampGel 6000 Imaging System (SageCreation Ltd., Beijing, China).

SVA

SVA was performed using a ProteomelabTMXL-A/XL-I analytical ultracentrifuge (Beckman Coulter, Fullerton, CA) equipped with a three-channel cell in an An-60 Ti rotor. All samples were prepared at a final concentration of 250 μM in 50 mM PBS. The HR1P/P21S8 and HR1P/P21S10 mixtures were incubated at 25 °C for 30 min and were initially scanned at 3000 rpm for 10 min. Data were obtained at 60000 rpm at a wavelength of 280 nm and 20 °C for 7 h. Weight-averaged molecular weights were calculated and fitted by processing with the SEDFIT program.

Pharmacokinetic Assessments

Sprague–Dawley rats weighing 200 ± 10 g each were obtained from the Animal Center of Beijing Institute of Pharmacology and Toxicology and were used for pharmacokinetic assessments. Animals were treated in accordance with the Animal Welfare Act and the “Guide for the Care and Use of Laboratory Animals” (NIH Publication 86-23, revised 1985). Complete pharmacokinetic experimental procedures are provided in the Supporting Information.