Human Insulin-like Peptide 5 (INSL5). Identification of a Simplified Version of Two-Chain Analog A13

Abstract

The receptor for insulin-like peptide 5 (INSL5), RXFP4, is a potential pharma target for treating human conditions such as constipation, anorexia, and obesity. However, since INSL5 has a complex structure of two chains and three disulfide bonds, its synthesis has proven to be extremely difficult via either chemical or recombinant approaches. Previous studies led to the engineering of a high yielding simplified INSL5 analog, named analog 13 (A13), which retains native INSL5-like activity. The focus of this study is to further simplify the structure of A13 by truncating the N-terminal residues of the B-chain. We have found that the first six residues at the N-terminus of A13 are not important for RXFP4 binding and cAMP potency. The most minimized active structure of INSL5 identified in this study is A13: B7–24 which will be an important research tool to study the physiological role of RXFP4 and a template for further modification to improve its pharmacokinetic properties.

The gut endocrine system is a rich source of hormones with therapeutic potential. For example, glucagon-like peptide-1 (GLP1) is a well-known gut hormone and several GLP1 analogs are approved drugs for the treatment of obesity and diabetes.^1,2 Insulin-like peptide 5 (INSL5; Figure 1) is a gut hormone that is expressed by colonic L-cells.³ This peptide exerts its physiological functions through the G protein-coupled receptor known as relaxin family peptide receptor 4 (RXFP4). Recent studies suggest that INSL5 is involved in the regulation of insulin secretion,⁴ food intake,³ or colon motility.⁵ Therefore, RXFP4 is a potential drug target for the treatment of metabolic (diabetes, obesity, anorexia) and colon motility disorders (e.g., constipation and diarrhea).

Figure 1.

Summary of SAR of human INSL5: (A) Primary sequence of human INSL5 and analog 13 (important residues are highlighted in red and mutated residue in purple). Z represents pyruglutamic acid. (B) NMR structure of human INSL5 (PDB: 2KBC) with important residue side chains in the B-chain (residue number labeled in red color).

INSL5 belongs to the relaxin family of peptides, which comprises of seven members, including relaxin 1–3 and insulin-like peptides 3–6 (INSL3–6). All seven members are structurally related to insulin, which contains two chains (A and B) cross-braced by two interdisulfide bonds between the chains and one intradisulfide bond within the A-chain⁶ (Figure 1). Relaxin family peptides act through a group of relaxin family peptide receptors 1–4 (RXFP1–4), which are G-protein coupled receptors.⁷ The overlapping expression of INSL5 and RXFP4 in the colon and in vitro pharmacological analysis indicated that RXFP4 is the cognate receptor of INSL5.⁸ The synthesis of INSL5 (45 amino acids, two chains, 3 disulfide bonds, Figure 1A) is extremely challenging by either chemical or recombinant synthesis approaches.^9,10 Both A- and B-chains are aggregating in nature, and thus synthesis and postsynthesis purification and handling are difficult which leads to very poor recovery. The yield of human INSL5 through chemical synthesis methods is only 0.8%⁹ and the lack of significant quantities of the ligand has limited the validation of RXFP4 as a therapeutic target. Recently, a simplified analog of INSL5, called analog 13 (A13), has been identified, which retains similar biological activities as human INSL5 based on cell-based assays¹¹ and also demonstrates potent activity in vivo on colon motility in mice.⁵ Compared with human INSL5, A13 has a truncated N-terminus for the first seven residues on its A-chain and has no intra-A-chain disulfide bond. In addition, the K^A16 residue of mouse INSL5 was introduced into the A13 A-chain to enhance the binding and activation at RXFP4 (Table 1). The synthesis yield of A13 was increased to 14% which is 17.5-fold higher than that of human INSL5¹¹.

Table 1. Primary Structure of Human INSL5, R3/I5, A13, A13: B3–24, A13: B5–24, and A13: B7–24, and Their Binding Affinity (pKi) and Potency (pEC50) at the RXFP4 Receptor.

Z represents pyroglutamatic acid. The important amino acids are highlighted in red color. #p < 0.001, ^p < 0.01 vs R3/I5.

^*Due to insufficient amount of sample, A13:B7–24 could not be repeated for RXFP4 binding assays.

The structure activity relationship (SAR) studies of INSL5¹¹⁻¹⁵ (Figure 1A) and the NMR structure (Figure 1B)¹⁶ suggested that multiple critical residues are located in the B-chain that adopted a long α-helical structure in the NMR studies. The A-chain of INSL5 has two α-helices and is considered to play a significant role in maintaining the B-chain structure. There are several residues in the A-chain that form hydrogen-bonds with B-chain residues. Altogether, the three helices enclose the hydrophobic core in INSL5, which are joined together by three disulfide bonds. Like other relaxin family peptides, the N-terminus of the INSL5 B-chain has limited structure with the NMR solution structure showing a disordered region (B1–B3) followed by an extended β-sheet (B4–B7).¹⁶ Thus, we hypothesized that the N-terminus residues of the B-chain might not be crucial for RXFP4 binding and can be truncated without altering activity. Therefore, the focus of this paper is to study the functions of the N-terminal residues of the INSL5 B-chain with the purpose to further simplify the structure of INSL5.

To determine the functions of the N-terminus of the INSL5 B-chain, three analogs, A13: B3–24, A13: B5–24, and A13: B7–24 were designed and synthesized (Table 1) based on the structure of A13 which was reported to be as active as INSL5.¹¹ All of the three A13-based truncated peptides have similar disulfide-bonded structures with the difference in the length of the B-chain N-terminus (Table 1).

The linear single A- and B-chains of each analog were synthesized by our optimized Fmoc solid-phase peptide synthesis protocol⁹ (Scheme 1). Briefly, due to the challenges in the SPPS of the INSL5 A-chain, pseudoproline Ler-Ser(φ^MeMePro)–OH was used to enable its successful synthesis. The Fmoc deprotection of the B-chain was carried out using 2% 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in 20% piperidine in DMF instead of standard 20% piperidine/DMF which improved the quality of the B-chain by decreasing aggregation. After purification of the A and B chains by reverse-phase high-performance liquid chromatography (RP-HPLC), two disulfide bonds between the chains were formed in a stepwise manner as we reported recently¹¹ (Scheme 1). The purity of the final peptides (88.4%–92.4%) was confirmed by analytical RP-HPLC peak area integration (SI: Table S1). The identities of the peptides were confirmed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) by a Bruker Ultraflex II instrument (Bruker Daltonics, Bremen, Germany). Analogue 13 was previously made and used for this study. A13: B3–24: m/z 4005.448 [M + H]⁺, calcd 3999.650; A13: B5–24: m/z 3818.308 [M + H]⁺, calcd 3813.490; A13: B7–24: m/z 3549.713 [M + H]⁺, calcd 3544.1429. The detailed characterization (analytical HPLC, retention time t_R, void time T₀, capacity/retention factor [k′ = (t_R – T₀)/T₀], MALDI-TOF MS, amino acid content, purity, and yield) of these analogues is provided in the Supporting Information (SI: Table S1).

Scheme 1. Schematic Diagram of Synthesis of Analogue 13, A13: B3–24, A13: B5–24, and A13: B7–24. (i) TFA:DODT:H₂O:TIPS (94:3:2:1, v/v), 120 min. (ii) DPDS (4 equiv); anisole/TFA mixture (1:9); 45 min. (iii) 6 M guanidium HCl buffer (pH = 8.5); 15 min. (iv) Iodine oxidation; 1 h. See ref (11) for detailed protocol.¹¹.

The Direct Detect assay-free cards and the Direct Detect spectrometer were used to measure the actual peptide contents (SI: Table S2). The overall yields of all three A13-based truncated peptides (A13: B3–24, A13: B5–24, and A13: B7–24) were 13–14%, calculated from the purified B-chains (SI: Table S1, Scheme S1). The yield of each peptide is comparable (∼13%) to that of A13 (14%) and is significantly higher than that of human INSL5 (0.8%)¹¹ (SI: Table S1).

Each peptide (A13: B3–24, A13: 5–24, and A13: 7–24) was tested in CHO-K1-RXFP4 cells for binding affinity against europium-labeled chimeric peptide R3/I5. R3/I5 has been previously reported as a high-affinity ligand for both RXFP4 and RXFP3.¹⁷ All three truncated analogs exhibited comparable RXFP4 binding affinity to A13, albeit with slightly reduced affinity compared with R3/I5 (Figure 2A, Table 1). Then, we tested these peptides for their ability to activate the RXFP4 receptor by inhibiting forskolin-stimulated cAMP production in CHO-K1-RXFP4 cells (Figure 2B, Table 1). Consistent with the result of competition binding assays, all three A13-based truncated peptides were full agonists and showed comparable potency to A13, although both A13 and the three truncated peptides have a slightly reduced potency compared with human INSL5 (Figure 2B, Table 1). Overall, the deletion of the first six residues on the N-terminus of the INSL5 B-chain did not reduce binding affinity or potency at human RXFP4. This result fits well with the SAR study of relaxin family peptides in that the N-terminus of the B-chain of relaxin 2, 3 and INSL3 is dispensable for their binding affinity and activity to their receptors.¹⁸⁻²⁰

Figure 2.

In vitro activity of Analogue13 (A13), A13: B3–24, and A13: B5–24, and A13: B7–24 in comparison to R3/I5 or hINSL5. (A) Competition binding assay using europium-labeled R3/I5; (B) cAMP inhibition activity of peptides in CHO-K1-RXFP4 cells.

The analog A13: B7–24 is therefore the simplest RXFP4 peptide agonist developed so far with 32 amino acids and two disulfide bonds compared with A13 with 38 amino acids and human INSL5 with 45 amino acids and three disulfide bonds. Previous studies indicated that a number of residues (e.g., R13, Y17, R23, W27) and the rigid α-helix structure within the C-terminus and midregion of INSL5 B-chain are important for its activity.^13,15,16,21 In addition, the A-chain of INSL5 is suggested to provide structural support for maintaining the correct conformation of the INSL5 B-chain via the two interdisulfide bonds and multiple hydrogen-bonds. The structure of A13: B7–24 has no extra residues at the termini beyond disulfide bonds except at the C-terminus of the B-chain which contains critical activation residues (e.g., R23 and W24).^13,15,21 Thus, A13: B7–24 is considered to be the most minimized two-chain structure of human INSL5.

The conformational properties of two of the three truncated analogs (A13: B3–24, A13: B5–24) were assessed by CD spectroscopy and compared with that of A13. Due to insufficient amount of A13: B7–24, its CD spectroscopy was not performed. A13 and its two truncated analogs (A13: B3–24, A13: B5–24) showed random coil structures in 10 mM phosphate buffer (PB) (Figure 3A). As previous study of INSL5-related peptides revealed that the α-helical structure could be induced by trifluoroethanol (TFE),²² thus, we further studied the CD spectroscopy of A13: B3–24 and A13: B5–24 in phosphate buffer containing 20% TFE. TFE is known to improve hydrogen bonding of peptides and proteins and thus stabilizes the α-helical structure of peptide. In the CD spectra obtained in phosphate buffer containing 20% TFE, all peptides, including A13, A13: B3–24, and A13: B5–24, showed a typical α-helical pattern with double minima at 208 and 222 nm (Figure 3B) suggesting that they had a tendency to form α-helical structure.

Figure 3.

Circular dichroism spectra of Analogue 13, A13: B3–24, and A13: B5–24 in (A) phosphate buffer (pH 7.5) and (B) 20%TFE/10 mM phosphate buffer at 25 °C.

Helix content of peptides is directly proportional to mean residue ellipticity (MRE) at 222 nm, [θ]₂₂₂. Therefore, [θ]₂₂₂ was used to calculate the helix content.²³ 100% helicity was calculated by using the formula ^max[θ]₂₂₂  =  – 40000 × [(1 – 2.5/n)] + (100 + T), where n equals to number of amino acid residues and T =  temperature of the peptide solution in °C. Percentage helicity was then calculated as 100 × [θ]₂₂₂/^max[θ]₂₂₂. The [θ]₂₂₂ value for A13 in 20% TFE is −11880.6, which corresponded to an α-helix content of 34.1% (Table 2). The α-helical content of A13 in TFE (34%) is similar to that of native INSL5 (38%)⁹ in aqueous solution (phosphate buffer) suggesting that A13 has a tendency to form native INSL5-like helical structure in an appropriate environment.

Table 2. α-Helix Contents of A13, A13: B3–24, and A13: B5–24 in 10 mM PB and 20%TFE/10 mM PB.

A13 and analogs	10 mM PB	20%TFE/10 mM PB
A13	7%	34%
A13: B3–24	4%	19%
A13: B5–24	5%	19%

Truncation of the N-terminal B-chain in A13: B3–24 led to a significant decrease in the [θ]₂₂₂ value, −6424.4, corresponding to 18.5% helicity (Table 2). A13: B5–24 has a similar [θ]₂₂₂ value to A13: B3–24, −6602.4, which corresponded to 19% helicity (Table 2). Hence, although A13: B3–24 and A13: B5–24 demonstrated reduced helicity in the CD spectra, they still retained high affinity and potency at RXFP4, indicating that the induced secondary structure of A13: B7–24 upon binding to the receptor is sufficient for full RXFP4 binding and activity.

The current study further minimized the structure of INSL5 by truncating six residues from the B-chain N-terminus of A13. Altogether, three novel peptides (A13: B3–24, A13: B5–24, and A13: B7–24) were developed which displayed comparable binding affinity and cAMP activation at RXFP4. Our study suggested that the N-terminal residues of INSL5 B-chain are dispensable for RXFP4 activity and A13: B7–24 is the most truncated active analog that is an important tool for future validation of RXFP4 in vivo.

Glossary

Abbreviations

INSL5

insulin-like peptide 5

A13

analog 13

GLP-1

glucagon-like peptide-1 (GLP-1)

RXFP4

relaxin family peptide receptor 4 (RXFP4)

INSL3–6

insulin-like peptide 3–6

RXFP1–4

relaxin family peptide receptors 1–4

Fmoc

9-fluorenylmethoxycarbonyl

HCTU

O-(1H-6-chlorobenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate

DMF

N,N-dimethylformamide

TFA

trifluoroacetic acid

DPDS

2,2′-dipyridyl disulfide

tBu

tetra-tert-butyl

DIEA

diisopropylethylamine

TIPS

triisopropylsilane

DODT

3,6-dioxa-1,8-octanedithiol

RP-HPLC

reversed phase high performance liquid chromatography

MALDI- TOF MS

matrix-assisted laser desorption ionization time-of-flight mass spectrometry

DBU

1,8-diazabicyclo[5.4.0]undec-7-ene

CD

circular dichroism;

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.0c00435.

• Schematic diagram of peptide synthesis; HPLC purity and characterization data including MALDI-TOF MS and direct detect analysis of INSL5 analogues (PDF)

Author Contributions

All authors have given approval to the final version of the manuscript. Conceived and designed the experiments: MAH. Performed the experiments: XZ. Analyzed the data: XZ, RADB. Contributed reagents/materials: MAH, RADB. Wrote the paper: XZ, RADB, MAH.

This research was funded by NHMRC (Australia) IDEAS grant #APP1182996 to MAH. During these studies, RADB was the recipient of an NHMRC Research Fellowship (1135837). Studies at the Florey Institute were supported by the Victorian Government’s Operational Infrastructure Support Program.

The authors declare no competing financial interest.