Discovery of peptidomimetic inhibitors of CREB/CBP by targeting hydrophobic grooves on the surface of the CBP KIX domain

Abstract

Cyclic AMP-response element binding protein (CREB), a downstream transcription factor of multiple signaling pathways, is overexpressed in many different types of cancers. Thus, targeting CREB has great potential for the development of antitumor agents. Peptidic foldamers have emerged as a powerful tool to disrupt disease-related protein–protein interactions (PPIs) with chemodiversity and high stability towards enzymatic degradation. Herein, we harnessed several hydrophobic groups of helical sulfonyl-γ-AApeptide foldamer targeting the hydrophobic grooves on the surface of the KIX domain of CREB binding protein (CBP), to disrupt CREB/CBP PPI. We showed that several stapled sulfonyl-γ-AApeptides could suppress CREB-mediated gene transcription and exhibit effective antiproliferative activity in cell-based assays and demonstrate its potency in inhibiting tumor growth in vivo. Our studies suggest that sulfonyl-γ-AApeptides as a class of helical foldamer could mimic the helical kinase-inducible activation domain of CREB (KID) to target the hydrophobic grooves on the surface of CBP KIX domain, and thereby inhibiting KIX–KID interaction, which provides a new strategy for the development of antitumor agent by targeting PPIs involving intrinsically disordered proteins (IDPs).

Graphical abstract

Stapled sulfonyl-γ-AApeptides selectively inhibit CREB/CBP interaction by targeting the KIX domain, offering a promising strategy to block transcription driven by intrinsically disordered proteins.

1. Introduction

Using helical foldamers to target protein–protein interactions (PPIs) involving intrinsically disordered proteins (IDPs) is a promising strategy for developing molecular tools to dissect protein function and the potential discovery of therapeutics for many diseases1, 2, 3, 4, 5, 6. As one of the most well-known IDPs and a member of the family of basic leucine zipper-containing transcription factors⁷, the cyclic AMP-response element binding protein (CREB) is a downstream transcription factor of multiple signaling pathways. Unphosphorylated CREB binds to the cAMP-response elements (CRE) but is transcriptionally inactive. Following stimulation by various extracellular signals including hormones, growth factors, and neuronal activity⁸, CREB is activated by phosphorylation^9,10. Once phosphorylated, CREB can regulate the gene expression by associating with transcription coactivators such as CREB-binding protein (CBP) and its paralog p300 to form CREB/CBP complex through the kinase-inducible activation domain (KID) of CREB and the KID-interacting domain (KIX) in CBP^11,12, leading to the recruitment of other transcriptional machinery to the gene promoter to activate CREB-dependent gene expression. CREB is overexpressed in many different types of cancers ranging from prostate cancer, breast cancer, non-small-cell lung cancer to medulloblastoma13, 14, 15, 16. As such, discovering novel inhibitors of the KIX–KID interaction that block CREB-mediated gene expression may facilitate anticancer therapeutic development. It is noted that in addition to CREB, the KIX domain of CBP also recognizes the transactivation domains of other nuclear factors involved in cancer development, suggesting targeting the KIX domain of CBP is a useful strategy for cancer-related drug discovery17, 18, 19.

Herein, we report our discovery of sulfonyl-γ-AApeptide foldamers as cell-permeable CREB inhibitors by recognizing the interface of KIX–KID. Several stapled peptidomimetics could inhibit the CREB/CBP PPI, suppress CREB-mediated gene transcription, and exhibit effective antiproliferative activity in cell-based assays and xenograft models (77% inhibition for MDA-MB-231 xenografts, i.p. at 10 mg/kg, QD, for 21 days) (Fig. 1A). Our findings suggested that sulfonyl-γ-AApeptides as a new class of helical foldamers could mimic KID domain and target the hydrophobic grooves on the surface of CBP KIX domain, which provides a new strategy for the development of antitumor agent by targeting PPIs involving intrinsically disordered proteins.

Figure 1.

Structural information of KID domain of CREB and sulfonyl-γ-AApeptide. (A) Mechanism of inhibiting tumor growth by stapled sulfonyl-γ-AApeptide. (B) Interactions between KIX and pKID (PDB: 1KDX). Residues of pKID that interact with KIX are labeled. (C) Helical-wheel representations of αB helix of KID. (D) Crystal structure of L-sulfonyl-γ-AApeptide (Left) and the schematic representation of the distribution of side chains in l-sulfonyl-γ-AApeptide (Right). (E) Design principle. (F) An overlay of sulfonyl-γ-AApeptide with the αB helix of KID.

2. Results and discussion

2.1. Rational design of CREB/CBP inhibitors that recognize the hydrophobic grooves of CBP KIX domain by sulfonyl-γ-AApeptides

In 1997, a solution structure of the complex of phosphorylated KID (pKID) and KIX reveals that pKID will undergo a coil-helix folding transition and form two α helices upon binding to the KIX domain (Fig. 1B, PDB: 1KDX)²⁰. The residues in the amphipathic αB helix of pKID mediate most interactions with KIX by contacting with the hydrophobic grooves formed by the side chains of residues Leu-599, Leu-603, Lys-606, Tyr 650, Leu 653, Ala-654, IIe-657 and Tyr-658 in the KIX. These hydrophobic interactions are essential for the formation of the pKID–KIX complex, and a mutation of hydrophobic residues in either KID or KIX will decrease the binding severely²¹. Detailly, in the αB helix of pKID, the side chain of Leu-141 projects deeply into the hydrophobic groove, and residues of IIe-137 and Leu-138 pack against a shallow region of the hydrophobic groove. All these residues also have extensive van der Waals contacts with the residues of KIX domain. Additional hydrophobic contacts are observed near the end of the binding groove formed by the side chains of Ala-145 and Pro-146 of KID, which primarily interact with the Leu-603 of KIX. Overall, these hydrophobic groups are located on one face of the αB helix of KID, roughly on two lines parallel to the helix axis (Fig. 1C). We, therefore, speculated that molecules bearing hydrophobic groups would be able to target the hydrophobic surface of KIX, leading to the discovery of novel KID–KIX PPI inhibitors. However, the development of peptide-based inhibitors is challenging due to the intrinsically disordered nature of the KID peptide as well as poor solubility and stability²².

Peptidomimetic foldamers based on the unnatural framework have shown promise in interrogating PPIs as they are resistant to enzymatic hydrolysis and are capable of folding into the well-defined helical secondary structure23, 24, 25, 26. We recently developed the molecular framework based on sulfonyl-γ-AApeptides, which was introduced as a class of unnatural peptidomimetic foldamer that offers unique advantages such as robust helical folding propensity, stability, diversity, and bioavailability, leading to applications in modulating a number of PPIs27, 28, 29, 30, 31, 32, 33. The crystal structures of homogeneous L-sulfonyl-γ-AApeptides reveal that they adopt 4₁₄ left-handed helices with a helical pitch of 5.1 Å, highly analogous to that of α-helix (5.4 Å) (Fig. 1D)³³. The robust folding conformation is presumably stabilized by both intramolecular hydrogen bonding and the restraining nature of sulfonamide moieties in the molecular backbone. We speculated that since sulfonyl-γ-AApeptides possess the excellent folding capability in the presence or absence of protein targets, they could be designed to mimic the folded helical domain of KID, which is otherwise disordered in the absence of KIX. In the structure of the sulfonyl-γ-AApeptide, the chiral side chains 1a, 3a, 5a, and 7a are on the same face of the helical scaffold; sulfonyl side chains 1b, 3b, 5b, and 7b are on the adjacent face (Fig. 1D). They were thus chosen to assemble the hydrophobic groups. For side chains of sulfonyl-γ-AApeptides not directly involved in contact with the KIX, we employed some positively or negatively charged hydrophilic side chains, in the hope of stabilizing the helical scaffold, as well as enhancing the solubility and cell permeability (Fig. 1E).

An overlay of sulfonyl-γ-AApeptides with the αB helix of KID (Fig. 1F) suggested that the side chains of 1a, 3a, and 5a derived from the residue of Tyr, Phe and Leu would take a similar position as Tyr-134, Leu-138, and Leu-141 in αB helix. The isobutyl and methyl groups placed at positions 3b, 5b, and 7b could mimic the residues of IIe-137 and Ala-145 well. Thus, we initially designed six γ-AA peptide sequences L1 to L6 (Fig. 2A). It is well known that stapled helical peptides have a more α-helical folding propensity and exhibit improved membrane permeability and proteolytic stability34, 35, 36. Although sulfonyl-γ-AApeptides were shown to be resistant to enzymatic degradation, stapling through macrocyclization was believed to enhance their cell permeability^32,37. In our initial fluorescence polarization (FP) assays, we found that all the designed sulfonyl-γ-AApeptides can bind to KIX (Supporting Information Fig. S1). Among these, L3 showed the best binding affinity towards KIX domain, we thus designed L3 derivatives (S1–S3) with staples crosslinking various AApeptide residues which are theoretically on the side opposite the binding interface with the KIX domain of CBP (Fig. 2B).

Figure 2.

Structures of designed linear and stapled peptides. (A) Linear l-sulfonyl-γ-AApeptides. Residues involved in the interaction with KIX are highlighted in blue. (B) Stapled l-sulfonyl-γ-AApeptides.

2.2. Inhibition of CREB-mediated gene transcription by L1–L6 and S1–S3

We first assessed the potential inhibitory activities of the designed sulfonyl-γ-AApeptides in inhibiting CREB-mediated gene transcription using a CREB Renilla luciferase (RLuc) reporter assay in HEK 293T cells³⁸. In this assay, HEK 293T cells were transfected with an RLuc reporter plasmid under the control of a synthetic CREB promoter containing three copies of cAMP-response element (CRE). Then the cells were treated with different sulfonyl-γ-AApeptides followed by forskolin (10 μmol/L, cAMP pathway activator) stimulation. We found linear peptides L1 to L6 did not show appreciable inhibitory effect toward CREB-mediated gene transcription in HEK 293T cells (Fig. 3A). Excitingly, two stapled peptides S2 and S3 could inhibit the CREB-mediated gene transcription with an IC₅₀ of 20.37 and 66.02 μmol/L, respectively (Fig. 3B). While no obvious inhibition of CREB's activity was observed with the stapled peptide S1 even though it was also developed based on the same sequence as linear peptide L3 (Fig. 3B), suggesting that the nature of the staple can have a dramatic effect on the cell permeability.

Figure 3.

Inhibition of CREB-mediated gene transcription. HEK 293T cells were transfected with CRE-RLuc. Then the cells were treated with increasing concentrations of linear peptides (A) or stapled peptides (B) for 30 min before the addition of Fsk (10 μmol/L) for 5 h. The Renilla luciferase activity was normalized to the protein concentration and presented as percentage of CREB activity. The errors are SEM from one representative experiment in triplicates.

2.3. Binding affinities and cell permeabilities of the peptide L3 and its stapled derivatives

To investigate the reason leading to the different inhibitory activities, we first conducted an FP assay to measure the binding affinity of linear peptide L3, stapled peptides S1, S2 and S3 to the KIX domain, respectively. As shown in Fig. 4A and Supporting Information Figs. S1 and S2 have a good K_d value of 0.84 μmol/L for binding to the KIX domain, while L3, S1 and S3 show a weaker binding affinity with a K_d value of 3.83, 4.15 and 3.26 μmol/L, respectively.

Figure 4.

Binding affinity and cell uptake assays. (A) Binding affinity of N-terminally FITCylated S2 and S3 bound to KIX domain. Binding affinity was assessed by FP assay. Error bars = standard deviation from three independent experiments. (B) Confocal microscopy imaging of HEK293T cells treated with FITC-labeled peptides L3, S1, S2, and S3. DAPI (top), FITC (middle), Merge (bottom). (C) Flow cytometry analysis. Cells were incubated with 1, 10, and 20 μmol/L FITC-labeled peptides, respectively. Mean fluorescence from two independent experiments was normalized against the fluorescence of peptide S3 at 20 μmol/L.

We then evaluated the cell-penetrating properties of FITC-conjugated L3, S1, S2, and S3, using N1 (αB helix of KID, sequence: YRKILNDLSSDA) as a control. The cell uptake was determined by confocal microscopy at first. Incubation of HEK-293T cells with FITC-labeled peptides followed by confocal imaging showed significant cell uptake for stapled peptides S2 and S3 at 10 μmol/L, whereas no intracellular signal was detected for linear L3, stapled S1 and N1 at the same concentration (Fig. 4B and Supporting Information Figs. S2–S4). In addition, flow cytometry-based assay provided further support for cell uptake. The peptides S2 and S3 showed about a 2- and 4-fold increase of cell uptake compared to liner L3 and stapled S1 (Fig. 4C and Supporting Information Table S1). Taken together, our results showed that the linker for stapling played a vital role in cell permeability. We speculated that perfluoroaryl stapling might largely promote cell penetration and this is consistent with other stapled peptides where perfluoration of aromatic staples can enhance cell permeability^39,40.

2.4. Structure–activity relationship (SAR) analysis

To further improve cellular activity, we set out to prepare additional derivatives of S2. The replacement of NH₂ groups of side chains of S2 with guanidine groups led to S2-1 showing a more than 3-fold decrease of activity in CREB inhibition (IC₅₀ = 67.86 μmol/L) as compared to S2 (Table 1), indicating in our case guanidine groups did not improve cell permeability. We next investigated the effect of stapling position on the activity. Four sequences (S2-2 to S2-5) with different positions of cyclization were synthesized and tested for their ability to inhibit the CREB-mediated gene transcription. As shown in Table 1, all these peptides have decreased activities compared to S2, suggesting the position of stapling between 4a and 6a was optimal.

Table 1.

Structures and CREB inhibition activities of peptides S2-1 to S2-5. Residues involved in the hydrophobic interaction with KIX domain were shown in blue.

Compound	Structure	CREB inhibition IC50 (μmol/L)a
S2-1		67.86 ± 15.19
S2-2		82.24 ± 71.78
S2-3		47.84 ± 23.31
S2-4		>100
S2-5		>100

^aCREB inhibition refers to inhibition of CREB-mediated gene transcription in HEK 293T cells using a CREB reporter assay.

We subsequently investigated the effect of hydrophobic groups on the recognition of the KIX surface. To do that, we assigned four hydrophobic areas H1 to H4 in our sequences and explored different hydrophobic groups in each area for their activities for CREB inhibition. Particularly, in H1, considering the residue of Tyr makes extensive van der Waals contacts with the side chain of residues in KIX, we only evaluated the substituents on the sulfonyl side chain. To this end, S2-6 was prepared by changing the methyl group of S2 to a benzyl group. In H2, we replaced the chiral side chain with the residue of Leu (S2-7). For H3, considering that the side chain of Leu-141 deeply projects into the hydrophobic pocket, we investigated the effect of different sizes of the hydrophobic groups on the activity by using methyl (S2-8), phenyl (S2-9, S2-10), parachlorobenzyl (S2-11, S2-12), benzyl (S2-13) and 2-naphthyl groups (S-14). In H4, we introduced parachlorobenzyl (S2-15), phenyl (S2-16), and isobutyl groups (S2-17). In total, thirteen peptides were prepared (Fig. 5), and the biological results of these peptides are summarized in Table 2. Evaluation of these peptides in the CREB inhibition assay led to the design and synthesis of a new peptide S2-18 with the preferred group on each hydrophobic position. Specifically, modifications at the H1 and H2 positions showed that replacing the original methyl and phenyl groups led to decreased activity (S2-6 and S2-7, respectively), indicating that these substituents were optimal and should be retained. At H3, substituting the leucine side chain with a phenyl group (S2-9) slightly improved activity, supporting this modification. In contrast, variations at H4—including para-chlorobenzyl (S2-15), phenyl (S2-16), and isobutyl (S2-17)—all led to reduced activity, suggesting that the small methyl group of alanine was favorable at this position. S2-18 indeed showed increased CREB inhibition activity with IC₅₀ of 18.89 μmol/L in the luciferase assay compared to S2.

Figure 5.

Structures of all peptides for the structure–activity analysis and mutation assays. Residues involved in the hydrophobic interaction with KIX domain were shown in blue and maroon. Modified residues were emphasized in maroon.

Table 2.

CREB inhibition activities and antiproliferative activities.

Compound	CREB inhibition IC50 (μmol/L)a	GI50 (μmol/L)b
		MCF-7	MDA-MB-231	MDA-MB-468
N1	>100	–	–	–
S2	20.37 ± 8.44	0.83 ± 0.11	4.59 ± 1.37	5.63 ± 1.70
S2-6	54.05 ± 18.21	–	–	–
S2-7	>100	–	–	–
S2-8	>100	–	–	–
S2-9	18.22 ± 2.57	10.08 ± 3.91	7.95 ± 2.97	4.95 ± 0.86
S2-10	68.87 ± 25.89	–	–	–
S2-11	72.29 ± 44.84	–	–	–
S2-12	80.34 ± 13.22	–	–	–
S2-13	32.50 ± 8.47	7.42 ± 1.62	8.70 ± 1.62	4.63 ± 0.26
S2-14	38.09 ± 7.02	–	–	–
S2-15	23.43 ± 7.97	7.63 ± 2.44	19.20 ± 4.02	5.54 ± 0.69
S2-16	43.15 ± 19.83	–	–	–
S2-17	37.42 ± 14.78	–	–	–
S2-18	18.90 ± 12.40	11.0 ± 4.89	14.47 ± 2.98	10.11 ± 2.19
S2-19	>100	–	–	–
S2-20	>100	–	–	–
S2-21	63.96 ± 22.32	–	–	–

^aCREB inhibition refers to inhibition of CREB-mediated gene transcription in HEK 293T cells using a CREB reporter assay.

^bGI₅₀ is the concentration required to inhibit the cancer cells' growth by 50% as evaluated by the CCK-8 assay. The compounds were incubated with cells for 72 h. Each compound was tested in triplicate in two independent experiments.

2.5. Specificity evaluation

To further evaluate the specificity of our peptidic foldamers for the KIX domain, we first determined the binding stoichiometry between S2 and KIX using isothermal titration calorimetry (ITC). The thermodynamic analysis yielded a binding stoichiometry (N) of 1.24, consistent with a 1:1 interaction between S2 and the KIX (Supporting Information Fig. S5).

We then performed structure–activity relationship (SAR) studies by modifying the hydrophobic group at the H4 position of peptide S2. Specifically, we either replaced the hydrophobic side chain with hydrophilic moieties or deleted it entirely. The chemical structures and CREB inhibitory activities of these analogs are shown in Fig. 5 and Table 2. Notably, the peptides S2-19 and S2-20 with the hydrophobic H4 residue replaced with Lys residue or guanidine residue became completely inactive at the highest tested concentration of 100 μmol/L. For the peptide S2-21, its inhibitory activity decreases more than 2-fold compared to the parent peptide S2. These results suggested our peptidomimetics specifically interacted with the hydrophobic groove on the surface of the KIX domain to disrupt the KIX–KID PPI.

To further evaluate the specificity of S2, we investigated if it would affect the transcription driven by other transcription factors. CREB activates its transcription through its interaction with CBP/p300 upon phosphorylation at Ser133. To investigate if S2 affects other transcription factors that interact with CBP/p300 for transcription activation, we tested heterologous transcription activators Gal4-MLL (2840–2858; DCGNILPSDIMDFVLKNTP). MLL (2840–2858) encodes the transcription activation domain from mixed lineage leukemia (MLL) that binds to CBP at an allosteric site to CREB binding site. We then fused human MLL (2840-2858) in-frame with the DNA-binding domain of yeast Gal4 transcription factor. In these transcription reporter assays, HEK293T cells were transfected with Gal4-fusions along with a Gal4-responsive luciferase reporter. Then the cells were treated with increasing concentrations of S2. In contrast to S2's effect on CREB, no stimulation on Gal4-MLL was observed at lower concentrations (10 and 50 μmol/L), and only a little bit stimulation of 1.47-fold at 100 μmol/L (P = 0.017) was seen (Fig. 6A).

Figure 6.

Specificity evaluation of S2 and S2-18 and Circular Dichroism spectra of some peptides. (A) The effect of S2 on MLL driven transcription (P = 0.017). (B) Binding affinity of N-terminally FITCylated S2 bound to KIX domain in presence of MLL peptide (2845–1858). (C) Overlay of ¹H–¹⁵N HSQC spectra for before (magenta) and after (blue) the addition of increasing concentrations of S2-18 to KIX domain. (D) Average chemical shift changes in ppm (parts per million) of the KIX domain (residues 587–673) when bound to S2-18. (E) Ribbon structure (left) and molecular surface (right) of KIX domain upon binding with S2-18. Residues colored red are chemical shifts above the average when peptide S2-18 was added. (F) CD spectra of linear peptide L3, stapled peptides S2 and S2-18, and control peptide N1 (sequence: YRKILNDLSSDA), as measured at 100 μmol/L, room temperature in H₂O.

Furthermore, based on the previous finding that the MLL activation domain facilitated the binding of CBP to CREB⁴¹, we examined the interaction between S2 and the CBP KIX domain in the presence of the human MLL peptide (2845-2858; LPSDIMDFVLKNTP). As illustrated in Fig. 6B, the inclusion of the MLL peptide significantly increased the binding affinity between S2 and the CBP KIX domain, with a three-fold enhancement from 0.84 to 0.27 μmol/L. Given that S2 mimics the CREB KID domain, this result provides compelling evidence about the action site of S2.

2.6. 2D NMR study of S2-18 with KIX

To gain insight into the interaction of sulfonyl-γ-AApeptides with KIX, we compared ¹H–¹⁵N HSQC spectra for ¹⁵N-labeled KIX in the presence and absence of S2-18, one of the most potent inhibitors of CREB-mediated gene expression based on its IC₅₀ value. Small to moderate chemical shift changes were observed for KIX residues involved in the formation of the hydrophobic groove upon binding S2-18. This represents the major binding site for the nonpolar face of KID, especially residue Tyr-658, which is necessary for the binding WT KID residue Ser-133 and contains the interaction-stabilizing residue Ala-654. As shown in Fig. 6C–E, Leu-603, Leu-653, Ala-654, and Tyr-658 showed obvious chemical shifts, which overlapped with the change of chemical shifts of residues when the KID peptide was added to KIX¹², suggesting that sulfonyl-γ-AApeptides targeted the hydrophobic groove on KIX.

2.7. Peptide aggregation propensity

Peptide aggregation is one of the major challenges in the development of therapeutic peptides, as it can adversely affect formulation stability, patient safety, and manufacturing efficiency^42,43. To assess the aggregation propensity, we selected peptides S2 and S2-18 and evaluated their behavior by ¹H NMR⁴⁴, using the natural peptide N1 as a control. ¹H NMR experiments were performed at 25 °C with the peptide concentration of 0.5, 1 and 2 mmol/L in D₂O. For both S2 and S2-18, signal shift occurred to a similar degree for all signals in the respective NMR spectra and the relative intensities of signals within each spectrum were practically unchanged (Supporting Information Fig. S6A). Thus, a single and well-separated signal could be used for straightforward comparison of key trends in the NMR result.

The spectrum of S2 and S2-18 shows a practically unchanged general appearance throughout the investigated concentration and time ranges, respectively, as natural peptide N1 (Fig. S6B). Besides, an essential linear increase in absolute integral with increasing concentration of S2 and S2-18 suggested they both predominantly present as monomers throughout the investigated time ranges (Fig. S6C).

2.8. Non-specific membrane disruption and selectivity assay

We performed a hemolytic assay of peptides S2 or S2-18 against human red blood cells (hRBCs) to rule out the potential side effect of the peptides in disrupting cell membranes. As shown in Supporting Information Fig. S7, our peptides didn't disrupt the red blood cell membrane up to a concentration of 450 μmol/L.

2.9. Circular dichroism (CD) measurements

To assess the conformational properties of our sulfonyl-γ-AApeptide-based peptidomimetics, CD studies were then conducted to analyze the helical propensity. The studies were carried out in deionized water and recorded between 190 and 260 nm. For comparison, the natural peptide N1 (structure was shown in Supporting Information Table S2) was included as a control. As shown in Fig. 6F, N1 exhibited a characteristic minimum around 197 nm, consistent with a random coil conformation, reflecting its intrinsically disordered nature that only adopts a helical structure upon binding to KIX. In contrast, sulfonyl-γ-AApeptides S2 and S2-18 showed a single maximum at ∼200 nm, revealing their left-handed helical conformation. It should be noted that the CD signatures of helical α-peptide and l-sulfonyl-γ-AApeptides are completely different due to their distinct molecular scaffold.

2.10. Inhibition of cancer cell proliferation by several peptides

The promising results from the CREB reporter assay prompted us to evaluate the antiproliferative activity of the most active peptidomimetics in three different cancer cell lines: MCF-7 (breast cancer), MDA-MB-231 (breast cancer) and MDA-MB-468 (breast cancer) using the Cell Counting Kit 8 (CCK-8) assay. The concentration required to inhibit 50% of the cancer cell growth (GI₅₀) is shown in Table 2 and presented in Fig. 7A. Among them, compound S2 presented a robust antitumor activity towards MCF-7 cells with a GI₅₀ of 0.83 μmol/L, and the GI₅₀ is 4.59 and 5.63 μmol/L for MDA-MB-231, and MDA-MB-468 cells, respectively.

Figure 7.

Peptides inhibit tumor cell growth. (A) Proliferation of MCF-7, MDA-MB-231, and MDA-MB-468 was inhibited by l-sulfonyl-γ-AApeptides in a dose-dependent manner (72 h treatments). (B) Compound S2 selectively inhibited tumor cell growth. Shown are antiproliferative dose response curves of S2 in breast cancer MCF-7, MDA-MB-468 and MDA-MB-231 cells as well as normal HEK 293T cells (72 h treatments).

We then tested if it was toxic to normal cells. Intriguingly, no significant inhibition of growth was observed in normal cell lines HEK 293T, up to 50 μmol/L concentration, which is more than 10-fold higher than its GI₅₀ in MCF-7 and MDA-MB-231 breast cancer cells (Fig. 7B).

2.11. In vivo PK/PD evaluation of peptide S2

We next evaluated S2's PK properties in C57BL/6 mice following intraperitoneal and oral administration (Fig. 8A). The data demonstrated that S2 possessed an acceptable half-life (T_1/2 = 3.0 h) and an acceptable oral bioavailability of ∼5.8% for oral administration. For IP administration, S2 demonstrated a longer half-life (T_1/2) of 6.0 h and a high peak concentration (C_max) of 38,948.5 μg/L.

Figure 8.

In vivo PK/PD evaluation of peptide S2. (A) PK profile of S2, i.p. = intraperitoneal administration, p.o. = oral administration. (B) Tumor growth curve of different treatment groups in female BALB/c nude mice bearing MDA-MB-231 tumors. Data points represent group mean. Error bars represent standard error of the mean (SEM). ∗∗∗∗P < 0.0001. (C) Tumors were removed from 5 mice in each group at study termination and are shown in the pictures. The number above each tumor represents the numbering of the mice. (D) Body weight changes in different treatment groups in female BALB/c nude mice bearing MDA-MB-231 tumors. Data points represent group mean body weight. Error bars represent standard error of the mean (SEM).

Having demonstrated antitumor activity in vitro, S2 was further evaluated in an MDA-MB-231 cell-derived xenograft mouse model. Mice were treated via intraperitoneal injection with S2 at a dose of 10 or 50 mg/kg once daily for 21 days, while saline was used as a negative control. As shown in Fig. 8B and C, the tumor size in the S2-treated group was smaller than that in the control group, indicating that the growth of xenograft tumors was significantly inhibited by S2, with tumor growth inhibition rate TGI (%) values of 77.42% and 76.46% at the dosage of 10 and 50 mg/kg, respectively. In addition, no significant toxicity was observed at this dosage, as monitored by the changes in body weight (Fig. 8D). Taken together, S2 was effective at inhibiting MDA-MB-231 tumor growth in vivo.

2.12. Discussion

In cancer cells, many signaling pathways are dysregulated leading to alterations of gene transcription. CREB is a downstream transcription factor of a multitude of signaling pathways and is overexpressed in many different types of cancers. Discovering novel inhibitors of KIX–KID interaction hold great promise to block CREB-mediated gene expression, thereby facilitating anticancer therapeutic development⁴⁵. So far, several small-molecule inhibitors of CREB-mediated transcription have been reported, including KG-501 and 666-15. KG-501 exhibited an IC₅₀ of 5.4 μmol/L in a CREB Renilla luciferase reporter assay, while 666-15 showed stronger inhibition with an IC₅₀ of 81.0 nmol/L in the same cellular assay⁴⁶. However, both compounds act through mechanisms that are not strictly dependent on direct KIX domain engagement. Notably, KG-501 demonstrated no significant inhibition of the KIX/KID interaction up to 50 μmol/L^38,47, and 666-15 disrupted this interaction only at relatively high concentrations (IC₅₀ of ∼18 μmol/L) in a split RLuc complementation assay⁴⁶. Additionally, 666-15 suffers from limited oral bioavailability, which remains a challenge for further development⁴⁸.

Sulfonyl-γ-AApeptides, as a new class of helical foldamers, were successfully employed to mimic the protein helical domain and modulate a series of medicinally relevant PPIs. In this study, we employed the strategy of l-sulfonyl-γ-AApeptide helical foldamers to mimic the αB-helix of KID based on the structure of CREB KID domain bound to the KIX domain of CBP. Although the αB peptide alone did not form a helical structure due to its intrinsically disordered nature, our sulfonyl-γ-AApeptides could adopt robust helical conformation in solution, which enabled us to introduce function groups at the desired positions when interacting with the KIX domain. As the hydrophobic residues in the amphipathic αB helix of pKID are involved in the most interactions with KIX by contacting with the hydrophobic grooves on the surface of the KIX, a few sulfonyl-γ-AApeptides were designed with hydrophobic side chains. For side chains of sulfonyl-γ-AApeptides not directly involved in the interaction of KIX, we still designed residues to engage in the interactions, as well as contribute to the stability and cell permeability. The design was considerably successful as some stapled peptidomimetics were discovered and demonstrated effective and selective inhibitory activity toward CREB-mediated gene transcription and antiproliferative activity in cell-based assays. The interaction with KIX is selective and specific, as demonstrated by FLuc/RLu assay, 2D-NMR studies, and the findings that in the presence of the MLL peptide, the binding affinity of our CREB mimetics revealed enhanced binding affinity toward CBP KIX. Furthermore, the lead sulfonyl-γ-AApeptide S2 shows its efficacy to inhibit the tumor growth in a MDA-MB-231 xenograft mouse model. Our approach of sulfonyl-γ-AApeptides provides an alternative strategy for the development of antitumor agents by targeting PPIs involving intrinsically disordered proteins (IDPs).

3. Conclusions

In summary, we report a series of novel CREB inhibitors that recognize the hydrophobic grooves of the CBP KIX domain by sulfonyl-γ-AApeptides, which could be defined as the mimic of αB helix of the KIX domain. These unnatural helical peptidomimetics can disrupt cancer-related KIX/KID PPIs with noticeable potency and selectivity. Cell-based studies indicated that our peptides are cell-permeable and can inhibit the growth of cancer cells. In the MDA-MB-231 xenograft models, our compound S2 showed about 77% inhibition towards tumor growth. An NMR study and specificity evaluation confirmed they have inhibitory activities by interacting with the hydrophobic groove on the surface of KIX to disrupt the KIX–KID interaction. This work also represents the successful application of unnatural peptidomimetics in disrupting KIX/KID PPI involving intrinsically disordered proteins (IDPs), which has long been considered a challenging target. Our approach provides a practical method for the development of novel foldamer peptidomimetics with proteolytically stable and cell-penetrating propensity. We believe this work will lead to the discovery of more CREB inhibitors by targeting the hydrophobic grooves on the surface of the KIX domain.

4. Experimental

4.1. Reagents

Fmoc-protected amino acids were purchased from Chem-Impex (Wood Dale, IL). Rink-Amide-MBHA resin (loading 0.623 mmol/g) was purchased from GL Biochem and was used for the solid phase synthesis of sulfonyl-γ-AApeptides. All solvents and other chemical reagents used for building block synthesis were obtained from commercial suppliers and used without purification unless otherwise indicated. The sulfonyl-γ-AA building blocks were synthesized based on previous reports. The sulfonyl-γ-AApeptides were purified and analyzed on a Waters Breeze 2 HPLC system installed with both an analytic module (1 mL/min) and a preparative module (16 mL/min) by employing a method using 5%−100% linear gradient of solvent B (0.1% TFA in ACN) in solvent A (0.1% TFA in H₂O) over 45 min, followed by 100% solvent B over 5 min. All compounds are >95% pure by analytical HPLC. The molecular weight of each peptide was confirmed by high-resolution mass spectrometry obtained from Agilent LC–MS QTOF 6540. HEK 293T, MDA-MB-231, MDA-MB-468, and MCF-7 cell lines were obtained from Moffitt Cancer Center in Tampa and were maintained in DMEM medium (high glucose, supplemented 10% FBS). Human red blood cells (hRBC) were obtained from Moffitt Cancer Center in Tampa. MDA-MB-231 cells used for xenograft model were maintained in vitro as adherent culture in L-15 medium supplemented with 15% fetal bovine serum and 1% penicillin-streptococci, at 37 °C in an atmosphere of 0% CO₂ in air. Cells in exponential growth phase were collected and quantified before inoculation.

4.2. Preparation of αB helix of KID domain

αB helix of the KID domain was synthesized on 150 mg of Rink-Amide resin (0.623 mmol/g) at room temperature. A scheme can be found in Supporting Information Fig. S8A. Briefly, the resin was soaked in DMF for 10 min, then the Fmoc group was removed by 20% piperidine in DMF solution (15 min × 2). After a thorough wash with DCM and DMF, the Fmoc-protected amino acid (2 equiv.) was coupled to the resin under the condition of DIC (4 equiv.) and HOBt (4 equiv.) in DMF for 4 h. Then, the resin was washed with DMF and DCM, and treated with 20% piperidine in DMF solution again, followed by the coupling with the next amino acid. This reaction cycle was repeated until desired peptides were synthesized. After removing the last Fmoc group, the N-terminus of the sequence was capped with acetic anhydride (2 mL) in pyridine (4 mL) twice. Then the peptides were cleaved from the resin by the solution of TFA/DCM (4 mL, 1:1, v/v) and purified by the Waters HPLC system.

4.3. Preparation of linear sulfonyl-γ-AApeptides

Sulfonyl-γ-AApeptides were prepared on 150 mg of Rink-Amide resin (0.623 mmol/g) at room temperature. A scheme can be found in Fig. S8B. Briefly, the resin was soaked in DMF for 10 min, then the Fmoc group was removed by 20% piperidine in DMF solution (15 min × 2). After a thorough wash with DCM and DMF, the sulfonyl-γ-AA building block (2 equiv.) was coupled to the resin under the condition of DIC (4 equiv.) and HOBt (4 equiv.) in DMF for 4 h. Then, the resin was washed with DMF and DCM, and treated with 20% piperidine in DMF solution again, followed by the coupling with the next sulfonyl-γ-AA building block. These cycles were repeated until desired sulfonyl-γ-AApeptides were synthesized. After removing the last Fmoc group, the N-terminus of the sequence was capped with acetic anhydride (2 mL) in pyridine (4 mL) (10 min × 2). Then the peptides were cleaved from the resin by the solution of TFA/DCM (4 mL, 1:1, v/v) and purified by the Waters HPLC system.

4.4. Preparation of stapled sulfonyl-γ-AApeptides

Stapled sulfonyl-γ-AApeptides were prepared on 150 mg of Rink-Amide resin (0.623 mmol/g) at room temperature. As shown in Fig. S8C, after capping the N-terminus of the sequence with acetic anhydride, the O-NBS was removed by DBU (10 equiv.) and 2-mercaptoethanol (5 equiv.) in DMF for 30 min, followed by the addition of a solution of terephthaloyl chloride (1 equiv.) and DIPEA (10 equiv.) in DCM or perfluoroaryl linker (1 equiv.) and DIPEA (20 equiv.) in DMF. The mixture was shaken for 1 h, then washed with DMF and DCM, and cleaved by the solution of TFA/DCM (4 mL, 1:1, v/v), and purified by the Waters HPLC system.

4.5. Preparation of FITC-labeled αB helix and FITC-labeled sulfonyl-γ-AApeptides

After attaching the last amino acid or sulfonyl-γ-AA building block, the Fmoc protecting group was then removed and Fmoc-protected β-Ala-OH (2 equiv.), DIC (4 equiv.) and HOBt (4 equiv.) in DMF were added. The mixture was shaken for 2 h, then the Fmoc group was removed, and FITC (1.2 equiv.) and DIPEA (10 equiv.) in 3 mL of DMF was added to the resin. After the mixture was shaken overnight, the resin was washed with DMF and DCM, then the FITC-labeled peptide was cleaved from the resin by 1:1 (v/v) DCM/TFA. The crude was purified by Waters HPLC, and the detailed structures can be found in Table S2, and they were confirmed by QTOF–MS (Supporting Information Table S3).

4.6. Preparation of KIX protein

The KIX domain of CBP (mouse residues 586–672) was purchased from Eurofins (Lancaster, PA) was inserted into a pET28a vector with a six-histidine tag and thrombin cut site N-terminal to the gene using BamHI and EcoRI cut sites and then transformed into E. coli BL21 (DE3) cells. Bacteria were grown at 37 °C to an OD₆₀₀ of 0.6, at which point the cells were transferred to 15 °C, induced with 1 mmol/L IPTG, and grown for 22 h. Cells were pelleted at 8200×g and stored at −80 °C. Cells were resuspended in 25 mL of lysis buffer containing one Pierce Protease Inhibitor tablet (Thermo Fisher, Waltham, MA, USA), 50 mmol/L NaH₂PO₄, 300 mmol/L NaCl, 0.02% NaN₃, and 10 mmol/L imidazole, pH 8.0. Cells were lysed using a French press at psi >1000 and lysate was centrifuged at 38,000×g. The soluble fraction of lysate was loaded onto a Ni-NTA column (Qiagen, Germantown, MD, USA) and eluted in buffer containing 250 mmol/L imidazole. Protein was then dialyzed overnight into buffer containing 50 mmol/L NaH₂PO₄, 300 mmol/L NaCl, and 0.02% NaN₃, pH 7.0, and the histidine tag was cleaved using a Thrombin CleanCleave Kit (Sigma–Aldrich, Burlington, MA, USA) for 2 h at room temperature. Cleaved protein was further purified using a 16/600 mm Superdex 75 size exclusion column (Cytiva, Marlborough, MA, USA). Protein purity was verified using SDS-PAGE and concentration was determined via spectrometry.

4.7. FP assays measure the binding affinity of αB helix or sulfonyl-γ-AApeptides to the KIX domain

The binding affinity (K_d) of the peptide to the KIX domain was measured by fluorescence polarization (FP) assays at 25 °C. Briefly, a constant amount of the 100 nmol/L FITC-labeled peptide was incubated with a serial dilution of protein in PBS buffer with 1% Pluronic F68. The K_d values were calculated using Eq. (1), in which the $L_{st}$ and $x$ refer to the concentration of peptide and protein, respectively.

$$y={\text{FP}}_{\min }+\left({\text{FP}}_{\max }-{\text{FP}}_{\min }\right)\times \frac{\left(K_{\mathrm{d}}+L_{\text{st}}+x\right)-\sqrt{{\left(K_{\mathrm{d}}+L_{\text{st}}+x\right)}^2-4L_{\text{st}}\times x}}{2L_{\text{st}}}$$ (1)

4.8. Cell permeability assay

HEK 293T cells were cultured in 6-well plates and allowed to attach to the bottom of the plate for overnight. Then, appropriate amounts of FITC-labeled peptides L3, S1, S2 and S3 dissolved in DMEM (0.1% DMSO) were added to the cells to final concentrations of 1, 10 and 20 μmol/L. The cells were incubated with the peptides for 4 h at 37 °C and 5% CO₂. After incubation, cells were washed three times with PBS and stained with DAPI (10 μg/mL) for 10 min, then it was washed with PBS twice and fixed with methanol for 5 min. They are then washed twice with PBS and imaged using Olympus FV1000 MPE multiphoton laser scanning microscope.

4.9. Flow cytometry assay

HEK 293T cells were seeded up to 3 × 10⁵ cells into centrifuge tubes with DMEM culture medium (supplemented with 10% FBS and 1% penicillin–streptomycin solution). Then, appropriate amounts of FITC-labeled peptides L3, S1, S2 and S3 dissolved in DMEM (0.1% DMSO) were added to the cells to make the final concentrations of 1, 10 and 20 μmol/L, respectively. The mixture was incubated for 30 min at room temperature in the dark, followed by removing any unbound drugs with centrifugation at 500×g for 3 min. The cells were then washed with staining buffer and the suspended cells were centrifuged at 500×g for 5 min. After decanting the buffer. the cells were resuspended in 200 μL of staining buffer for final flow cytometric analysis by Guava® easyCyte™ Flow Cytometer.

4.10. Inhibition of CREB-mediated gene transcription

HEK 293T cells in a well of 6-well plate were transfected with a plasmid (2 μg) encoding Renilla luciferase under the control of three copies of CRE (pCRE-RLuc) using Lipofectamine²⁰⁰⁰ for 3 h. Then the cells were replated into a 96-well plate, and the cells were allowed to be attached to the bottom of the plate overnight. The cells were treated with different concentrations of peptides for 30 min, followed by the addition of Fsk (10 μmol/L). The cells were further incubated for 5 h at 37 °C. The media were removed, and the cells were lysed in 30 μL of 1 × Renilla luciferase lysis buffer (Promega) and the Renilla luciferase activity was measured using the Renilla luciferase assay system (Promega). The protein concentration in each well was determined using the Protein Assay Dye Reagent Concentrate (Biorad). The luciferase activity in each well was normalized to the protein concentration and expressed as relative luciferase unit (RLU)/μg protein. The IC₅₀ was calculated in Prism 9.0 using the nonlinear regression analysis.

4.11. 2D NMR study of peptide S2-18 with KIX domain

NMR experiments were performed at 25 °C on a Varian VNMRS 800-MHz spectrometer with a triple resonance pulse field Z-axis gradient cold probe. ¹H–¹⁵N HSQC spectra were collected to assess binding, using 425 μmol/L samples of ¹⁵N-labeled KIX in the apo form or in complex with unlabeled peptide S2-18 at a concentration of 1700 μmol/L⁴⁹. Experiments were conducted in buffer consisting of 90% H₂O, 10% D₂O, 10 mmol/L NaH₂PO₄, 1.8 mmol/L KH₂PO4, 2.7 mmol/L KCl, 137 mmol/L NaCl, 3% DMSO, and 0.02% NaN₃, pH 7.4. HSQC used sweep widths of 9689.22 and 2200 Hz for ¹H and ¹⁵N dimensions, respectively, and complex points 2048 and 256 for ¹H and ¹⁵N dimensions, respectively. External water was used as reference for the ¹H, and 118.143 ppm was used as reference for the ¹⁵N dimension. Data was processed using NMRFx and analyzed using NMRViewJ^50,51. Chemical shifts were calculated as Eq. (2):

$$\text{Chemica}\text{shifts}={[\left({\left({\Delta }^1\mathrm{H}\right)}^2+({\Delta }^{15}\mathrm{N}/{5)}^2\right)/2]}^{1/2}$$ (2)

For backbone assignments of the apo KIX protein, we collected 3D HNCO, HNCA, and HNCACB dataof ¹³C–¹⁵N labeled KIX at 425 μmol/L. HNCO experiments used sweep widths of 9689.22 Hz (¹H), 2200 Hz (¹⁵N), and 2412.096 Hz (¹³CO) and complex points of 2048 (¹H), 60 (¹⁵N), and 112 (¹³CO)^52,53. The reference for the carbon dimension was 174.095 ppm. HNCA experiments used sweep widths of 9689.22 Hz (¹H), 2200 Hz (¹⁵N), and 6031.345 Hz (¹³C_α) and complex points of 2048 (¹H), 80 (¹⁵N), and 256 (¹³C_α). The reference for the carbon dimension was 565.133 ppm. HNCACB experiments used sweep widths of 9689.22 Hz (¹H), 2200 Hz (¹⁵N), and 14073.137 Hz (¹³C_α) and complex points of 2048 (¹H), 80 (¹⁵N), and 220 (¹³C_α). The reference for the carbon dimension was 46.136 ppm. 3D NMR spectra were processed and analyzed using POKY⁵⁴. Surface models that show chemical shifts >0.02 ppm of KIX residues in response to S2-18 were made using VMD 1.9.1 overlaid on a complex of KIX (mouse residues 586–666) and the KID domain of CREB (rat residues 119–146) (PDB ID: 1KDX)^55,56.

4.12. Cell growth inhibition assay

The cells were plated in 96-well plates and allowed to attach to the bottom of the plate overnight. Three replicates were made for each measurement. Then the cells were treated with serial dilution of peptides for 72 h at 37 °C. At the end of treatment, the viable cells were determined by incubating cells with CCK-8 reagent (10 μL) for 2 h. Then the media were removed, and the formed purple formazan was dissolved into DMSO for quantification by absorbance at 450 nm using a microplate reader. The cell viability percentage was calculated using Eq. (3):

$$\text{Cell}\text{viability}\text{percentage}=100\times \left[\frac{A_{\text{treated}}-A_{\text{medium}}}{A_{\text{control}}-A_{\text{medium}}}\right]$$ (3)

where A_treated represents absorbance in well treated with a compound, A_medium represents the absorbance of medium, and A_control denotes media-treated cells. The GI₅₀ was then calculated in Prism 9.0 using the nonlinear regression analysis.

4.13. In vivo pharmacokinetics study

In two separate experiments, the compound S2 is administered either p.o. or i.p. to mice at a dose of 50 mg/kg (volume 150 μL). Following administration, 75 μL blood samples are collected at 10, 20, 30 min, 1, 2, 4, 8, 16, 24, 36, and 48 h (n = 3 per time point, and each mouse is used for three time points, thus 12 mice are used for either p.o. or i.p. to make a total of 24 mice) after drug administration. Blood is collected into 1.5-mL Eppendorf tubes containing 30 μL disodium EDTA (0.5 mol/L, pH 8.0) and kept on ice until plasma collection (<30 min), and then centrifuged at 4000 rpm for 10 min at 4 °C (Thermo Scientific, Legend Micro 21R centrifuge, Waltham, MA, USA). The supernatants (serum) are collected. The 100 μL serum samples are added to the mixture of 270 μL acetonitrile and 30 μL glacial acetic acid. The samples are allowed to rest on ice for 15 min and centrifuged at 10,000 rpm at 4 °C for 15 min (Thermo Scientific). The clarified supernatants are transferred to vials and analyzed by Agilent LC–MS QqQ 6460.

4.14. MDA-MB-231 xenograft tumor model

This assay was conducted by the GenScript Biotechnology Co., Ltd. All the procedures related to animal handling, care and treatment in the study were performed according to the guidelines approved by the Institutional Animal Care and Use Committee (IACUC) following the guidance of the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) with the ethical approval number GS-PIVP21-001-04-0008. Specifically, MDA-MB-231 cells (cell viability >90%) were collected and resuspended in serum-free L-15 medium to 5 × 10⁷/mL, then an equal volume of Matrigel was added and mixed. Each mouse (BALB/c nude mice, female, 6–8 weeks, weighing approximately 18–22 g, purchased from Beijing Vital River Laboratory Animal Technology) will be injected subcutaneously with 0.2 mL of MDA-MB-231 cell suspension above the right scapula. Afterwards, the tumor volume of the mice will be monitored using vernier calipers. When the average tumor volume of the mice reaches about 100–150 mm³, they will be randomly divided into groups according to the tumor volume. Tumor volumes and body weight will be measured twice per week after group division. Tumor volumes were calculated as Eq. (4):

$$V=(L\times W\times W)/2$$ (4)

where L is tumor length and W is tumor width. The tumor inhibitory effect of the test substance is evaluated by the tumor growth inhibition rate TGI (%) or the relative tumor proliferation rate T/C (%). The calculation formula of TGI (%) is shown in Eq. (5):

$$\text{TGI}(\%)=\left[1–(\text{Average}\text{tumor}\text{volume}\text{at}\text{the}\text{end}\text{of}\text{treatment}\text{in}\mathrm{a}\text{certain}\text{treatment}\text{group}–\text{Average}\text{tumor}\text{volume}\text{at}\text{the}\text{beginning}\text{of}\text{treatment}\text{in}\text{the}\text{treatment}\text{group})/\left(\text{Average}\text{tumor}\text{volume}\text{at}\text{the}\text{end}\text{of}\text{treatment}\text{in}\text{the}\text{control}\text{group}\right)–\left(\text{Average}\text{tumor}\text{volume}\text{at}\text{the}\text{beginning}\text{of}\text{treatment}\text{in}\text{the}\text{control}\text{group}\right)\right]\times 100\%$$ (5)

Relative tumor proliferation rate T/C (%): the calculation formula is shown in Eq. (6):

$$T/C(\%)=T_{\text{RTV}}/C_{\text{RTV}}\times 100$$ (6)

T_RTV: the average tumor volume of the treatment group; C_RTV: the average tumor volume of the negative control group), T_RTV take data from the same day as C_RTV.

After the experiment, the tumor weight will be detected and the T/C_weight (%) value will be calculated. The calculation formula is shown in Eq. (7):

$$T/C_{\text{weight}}(\%)=T_{\text{weight}}/C_{\text{weight}}\times 100$$ (7)

T_weight: the average tumor weight of the treatment group; C_weight: the average tumor weight of the negative control group value). All data was expressed as mean ± standard error of mean (SEM). The two-way ANOVA test in GraphPad will be used to compare the differences between each test substance group and the control group. P < 0.05 is considered to have a significant difference.

4.15. Hemolytic assays

Hemolytic activity was determined by incubating suspensions of human red blood cells with serial dilutions of each peptide. Human red blood cells (hRBC) were rinsed several times in PBS, followed by centrifugation at 4 °C/700×g for 10 min. Then the supernatants were removed and the remaining hRBC was diluted with PBS buffer in a 1:20 ratio. Then 50 μL of the diluted hRBC was added to the prepared 96-well microplate, which contains different concentrations of each peptide. The PBS-treated hRBC is the negative control, and hRBC treated with 1% Triton was the positive control. Then, the 96-well microplate was incubated at 37 °C for 1 h and centrifuged at 4 °C/3500 rpm for 10 min (Eppendorf 5810R, Eppendorf AG, Hamburg, Germany). 100 μL of PBS buffer was added to a new 96-well microplate, and 30 μL of the supernatant from the centrifuged solution was transferred to the new 96-well microplate. The absorbance was measured at 540 nm. The hemolysis percentage was calculated as Eq. (8):

$$\text{Hemolysis}\text{percentage}(\%)=({\text{Abbs}}_{\text{Sample}}–{\text{Abbs}}_{\text{buffer}})/({\text{Abbs}}_{\text{Triton}}–{\text{Abbs}}_{\text{buffer}})\times 100$$ (8)

where Abbs is absorbance detection at 540 nm. The experiment was performed in duplicate and repeated independently three times.

4.16. CD measurement

CD spectra of peptides were measured in H₂O at a concentration of 100 μmol/L by using JASCO Model J-1500 Circular Dichroism Optical Rotatory Dispersion (CD/ORD) spectrometer with a 1 mm path length quartz cuvette. Ten scans were averaged for each sample and three independent experiments were conducted. The final spectra were normalized by subtracting the average blank spectra. Molar residue ellipticity (MRE) [θ] (deg/cm²/dmol) was calculated by Eq. (9):

$$\left[\theta \right]={\theta }_{\text{obs}}/\left(n\times \mathrm{I}\times c\times 10\right)$$ (9)

where ${\theta }_{\text{obs}}$ is the measured ellipticity in millidegree, $n$ is the number of side groups, I is the path length in centime (0.1 cm), and $c$ is the centration of peptide in molar unit.

Author contributions

Jianfeng Cai, Gary Daughdrill, and Xiangshu Xiao conceived the idea, Bo Huang prepared all the peptidomimetics and conducted binding affinity assays, and CD spectra assays. Emily Gregory-Lott conducted the NMR study and prepared KIX domain. Bingbing X. Li assayed the peptide activities in CREB Renilla luciferase (RLuc) reporter assay. Menglin Xue did the membrane disruption assay, and cell growth inhibition assay. Anabanadam Asokan assigned the KIX protein sequence. Sihao Li helped with the preparation of building blocks. Timothy H. Tran helped with binding study, Shaohui Wang and Xingming Sun prepared samples for PK test. Ning Shen and Chuanhai Cao helped with the flowcytometry assay. Bo Huang, Emily Gregory-Lott, Bingbing X. Li, Xiangshu Xiao, Gary Daughdrill, and Jianfeng Cai analyzed data and interpreted results. Bo Huang, Emily Gregory-Lott, Bingbing X. Li, Xiangshu Xiao, Gary Daughdrill, and Jianfeng Cai wrote and revised the manuscript. All authors reviewed and approved the final version of the manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Appendix A. Supporting information

The following is the Supporting Information to this article: