Study of Chemical Ligation Via 17-, 18- and 19-Membered Cyclic Transition States

Authors

  • Siva S. Panda,

    1. Center for Heterocyclic Compounds, Department of Chemistry, University of Florida, Gainesville, FL 32611-7200, USA
    Search for more papers by this author
    • These authors contributed equally to this paper.

  • Claudia El-Nachef,

    1. Center for Heterocyclic Compounds, Department of Chemistry, University of Florida, Gainesville, FL 32611-7200, USA
    Search for more papers by this author
    • These authors contributed equally to this paper.

  • Kiran Bajaj,

    1. Center for Heterocyclic Compounds, Department of Chemistry, University of Florida, Gainesville, FL 32611-7200, USA
    Search for more papers by this author
  • Abdulrahman O. Al-Youbi,

    1. Chemistry Department, King Abdulaziz University, Jeddah 21589, Saudi Arabia
    Search for more papers by this author
  • Alexander Oliferenko,

    1. Center for Heterocyclic Compounds, Department of Chemistry, University of Florida, Gainesville, FL 32611-7200, USA
    Search for more papers by this author
  • Alan R. Katritzky

    Corresponding author
    1. Center for Heterocyclic Compounds, Department of Chemistry, University of Florida, Gainesville, FL 32611-7200, USA
    2. Chemistry Department, King Abdulaziz University, Jeddah 21589, Saudi Arabia
    Search for more papers by this author

Corresponding author: Alan R. Katritzky, katritzky@chem.ufl.edu

Abstract

Unprotected S-acylated cysteine isopeptides containing α-, β- or γ-amino acid units have been synthesized, and their conversion to native hexapeptides by S- to the N-terminus ligations involving 17-, 18- and 19-membered cyclic transition states have been demonstrated both experimentally and computationally to be more favorable than intermolecular cross-ligations.

Synthetic methods for peptides are of great interest: native chemical ligation (NCL) that chemically connects two unprotected peptides to yield a long-chain polypeptide has been intensively studied since first developed by Kent and co-workers (1).

Alternative approaches to overcome the requirement in classical NCL of an N-terminal cysteine residue have included (i) traceless Staudinger ligation (2,3), (ii) NCL with a phenylalanine and valine analog bearing a sulfhydryl group at the β-position followed by removal of the sulfanyl group (4–6), (iii) NCL followed by the conversion of cysteine to serine (7), (iv) sugar-assisted ligation (8–10), and (v) cysteine-free ‘direct aminolysis’ methods (11). However, new ligation strategies are still in demand for the synthesis of underivatized and post-translational modified peptides and proteins.

Native chemical ligation has been extensively studied in peptidic compounds bearing a cysteine residue at the N-terminus (12). Modified cysteine scaffolds have also been incorporated for the syntheses of novel peptides and proteins (13–16) and for surface immobilization (17). For instance, N-acylcysteines have found utility as photoactivatable analogs of glutathione (18) and for the synthesis of oxytocin-like peptides (19).

Haase and Seitz showed that an internal cysteine with glycine or alanine as the N-terminal residue enhanced the rate of NCL. Ligation rate also depends on the ring size formed during the SN acyl transfer (20). Wong and coworkers (9,21) reported that N-terminal glycine favors second-generation sugar-assisted ligation in cysteine-containing and cysteine-free glycopeptides over sterically encumbered l-amino acids like valine, leucine, isoleucine, and proline. According to a generally accepted notion, the subsequent acyl shift should proceed via entropically favored 5- and 6-member ring intermediates to achieve useful rates. However, our group has tested the feasibility of chemical ligation of S-acylated cysteine peptides via 5-, 8-, 11-, and 14-membered cyclic transition states (22–24). We now report migration of the acyl group from S-acylated cysteine isopeptides containing α-, β- and γ-amino acids via 17-, 18- and 19-membered cyclic transition states.

Methods and Materials

Melting points were determined on a capillary point apparatus equipped with a digital thermometer. NMR spectra were recorded in CDCl3, DMSO-d6, acetone-d6, or methanol-d4 on Mercury or Gemini NMR spectrometers operating at 300 MHz for 1H (with TMS as an internal standard) and 75 MHz for 13C. Elemental analyses were performed on a Carlo Erba-EA1108 instrument. All microwave-assisted reactions were carried out with a single mode cavity Discover Microwave Synthesizer (CEM Corporation Matthews, NC, USA). The reaction mixtures were transferred into a 10-mL glass pressure microwave tube equipped with a magnetic stir bar. The tube was closed with a silicon septum, and the reaction mixture was subjected to microwave irradiation (Discover mode; run time: 60 seconds; PowerMax-cooling mode). HPLC-MS analyses were performed on reverse phase gradient Phenomenex Synergi Hydro-RP (2.1 × 150 mm; 5 μm) + guard column (2 × 4 mm) or Thermoscientific Hypurity C8 (5 μm; 2.1 × 100 mm + guard column) using 0.2% acetic acid in H2O/methanol as mobile phases; wavelength = 254 nm; and mass spectrometry was carried out with electrospray ionization (ESI).

General procedure, experimental details for the preparation of compounds S1–S6, 1–3 are given in the Appendix S1. In addition, Tables S1 and S2 that contain details for the characterization of compounds S1a–S1e (Table S1), and compounds S3–S6, 1 (Table S2) are also included in the Appendix S1.

General procedure for the preparation of mono-isohexapeptides (5, 10, 15)

To the solution of compound 3 (0.20 g, 0.339 mmol) in water and THF, Et3N (0.086 mL, 0.618 mmol) was added and stirred at 0 °C for 5 min until dissolved. Boc-protected α-, β- or γ-amino acid (0.339 mmol) was added slowly, and then the mixture was left to stir at 0 °C for 5 h (monitored by TLC). Upon completion, THF was evaporated under reduced pressure and the residue was dissolved in ethyl acetate. The organic layer was extracted with 2N HCl (3 × 10 mL) and water (1 × 20 mL) then washed with brine (1 × 20 mL). The organic layer was dried over magnesium sulfate and evaporated under reduced pressure then washed with diethyl ether to yield Boc-protected mono-isohexapeptides (4, 9, 14). Then the Boc-protected mono-isohexapeptide was stirred in dioxane/HCl for 1 h. Dioxane was evaporated under reduced pressure, and the residue was treated with diethyl ether. The resulted turbid solution was kept at −5 °C overnight. The solid obtained was filtered and washed with diethyl ether to yield mono-isohexapeptides (5, 10, 15).

(2S,5S)-14-amino-2-((((S)-2-(((Benzyloxy)carbonyl)amino)propanoyl)thio)methyl)-5-isobutyl-4,7,10,13-tetraoxo-3,6,9,12-tetraazatetradecan-1-oic acid hydrochloride (5)

White microcrystals (78%); mp 86–88 °C; 1H NMR (DMSO-d6) δ 8.75 (t, = 5.1 Hz, 1H), 8.44 (d, = 7.5 Hz, 1H), 8.30–8.26 (m, 1H), 8.21–8.16 (m, 2H), 8.09 (d, = 7.5 Hz, 1H), 7.99 (d, = 8.0 Hz, 1H), 7.38–7.34 (m, 5H), 5.07–5.00 (m, 2H), 4.40–4.33 (m, 1H), 4.25–4.17 (m, 2H), 3.82 (d, = 5.4 Hz, 2H), 3.77–3.74 (m, 2H), 3.64–3.59 (m, 2H), 3.26 (dd, J = 13.4, 5.9 Hz, 1H), 3.08 (dd, = 13.2, 8.4, Hz, 1H), 1.61–1.52 (m, 1H), 1.48–1.40 (m, 2H), 1.25 (d, = 6.9 Hz, 3H), 0.89–0.81 (m, 6H). 13C NMR (DMSO-d6) δ 201.8, 172.1, 171.4, 168.5, 168.3, 166.4, 155.8, 136.7, 128.4, 127.9, 127.8, 65.8, 62.8, 56.7, 51.6, 50.6, 42.0, 41.8, 41.2, 29.2, 24.1, 23.2, 21.6, 17.3. HRMS Calcd for C26H38N6O9SNa [M + Na]+ 633.2313; Found 633.2305.

(2S,5S)-15-Amino-2-((((S)-2-(((benzyloxy)carbonyl)amino)propanoyl)thio)methyl)-5-isobutyl-4,7,10,13-tetraoxo-3,6,9,12-tetraazapentadecan-1-oic acid hydrochloride (10)

White microcrystals (74%); mp 80–82 °C; 1H NMR (DMSO-d6) δ 8.54–8.47 (m, 1H), 8.43–8.36 (m, 1H), 8.23–8.16 (m, 2H), 8.10 (d, = 7.5 Hz, 1H), 8.04–7.92 (m, 2H), 7.40–7.31 (m, 5H), 5.11–5.02 (m, 2H), 4.40–4.32 (m, 1H), 4.28–4.17 (m, 2H), 3.73 (d, = 5.1 Hz, 4H), 3.26 (dd, = 13.4, 5.9 Hz, 1H), 3.17–3.06 (m, 1H), 3.04.–2.92 (m, 2H), 2.57 (t, = 6.9 Hz, 2H), 1.61–1.54 (m, 1H), 1.45 (t, = 6.9 Hz, 2H), 1.25 (d, = 7.5 Hz, 3H), 0.88–0.82 (m, 6H). 13C NMR (DMSO-d6) δ 201.8, 172.1, 171.4, 170.0, 169.1, 168.4, 155.8, 136.7, 128.4, 127.9, 127.8, 65.8, 56.7, 51.6, 50.6, 42.2, 41.9, 41.1, 35.2, 32.1, 29.2, 24.1, 23.2, 21.6, 17.3. HRMS Calcd for C27H40N6O9SNa [M + Na]+ 647.2470; Found 647.2485.

(2S,5S)-16-amino-2-((((S)-2-(((benzyloxy)carbonyl)amino)propanoyl)thio)methyl)-5-isobutyl-4,7,10,13-tetraoxo-3,6,9,12-tetraazahexadecan-1-oic acid hydrochloride (15)

White microcrystals (72%); mp 78–80 °C; 1H NMR (DMSO-d6) δ 8.43–8.39 (m, 1H), 8.35–8.29 (m, 1H), 8.21–8.17 (m, 1H), 8.12–7.97 (m, 4H), 7.45–7.32 (m, 5H), 5.12–5.03 (m, 2H), 4.41–4.35 (m, 1H), 4.27–4.19 (m, 2H), 3.76–3.65 (m, 4H), 3.27 (dd, = 13.2, 5.4 Hz, 1H), 3.17–3.05 (m, 1H), 2.86–2.73 (m, 2H), 2.28 (t, = 6.8 Hz, 2H), 1.84–1.78 (m, 2H), 1.62–1.56 (m, 1H), 1.47 (t, = 6.9 Hz, 2H), 1.20 (d, = 7.2 Hz, 3H), 0.89–0.83 (m, 6H). 13C NMR (DMSO-d6) δ 201.8, 172.1, 171.9, 171.4, 169.3, 168.4, 155.8, 136.7, 128.4, 127.9, 127.8, 66.4, 65.8, 65.0, 56.7, 50.7, 42.1, 41.9, 40.3, 31.9, 29.2, 24.1, 23.2, 21.6, 17.3, 15.2. HRMS Calcd for C28H42N6O9SNa [M + Na]+ 661.2626; Found 661.2629.

Results and Discussion

We synthesized the intermediate mono-isopentapeptide 3 as starting material to study the possibility of long-range S- to N-acyl migration via 17-, 18- and 19-membered cyclic transition states. Compound 3 on coupling with α-, β-, or γ-amino acids gave the starting mono-isohexapeptides (5, 10, 15) needed for the ligation studies. To enhance migration rates, we used a glycine unit at the N-terminus of mono-isohexapeptide 5 and β- and γ-amino acid units in mono-isohexapeptide 10 and 15, respectively.

Preparation of the mono-isopentapeptide 3

The benzotriazolide S1a of Fmoc-protected glycine was coupled with free leucine S2a at 20 °C in the presence of base to give Fmoc-protected dipeptide S3 (87%). Compound S3 was converted into the benzotriazolide S4 at −15 °C, and S4 was then reacted with S2b at 20 °C in the presence of triethylamine to generate protected tripeptide Fmoc-Gly-Leu-Cys-OH S5 (91%). Tripeptide S5 was S-acylated by Cbz-Ala-Bt S1b in the presence of KHCO3 to provide N-protected mono-isotetrapeptide S6, which, after deprotection by DBU, yielded the mono-isotetrapeptide 1 in 85% yields (Scheme S1 in Appendix S1; 22, 25–27).

Mono-isotetrapeptide 1 on coupling with Boc-Gly-Bt S1c at −10 °C in THF-water in the presence of 2 equiv. of triethylamine gave mono-isopentapeptide 2 (65%). Compound 2 on deprotection of the Boc group with dioxane/HCl solution afforded unprotected mono-isopentapeptide 3 (81%), which was used as the common intermediate to prepare mono-isohexapeptides (5, 10, 15) for ligation studies (Scheme 1).

Figure 
                Scheme 1:
              .

 Synthesis of mono-isopentapeptide 3.

Study of the feasibility of SN acyl migration via a 17-membered cyclic transition state

Boc-protected mono-isohexapeptide 4 was synthesized in 74% yield in solution phase, by coupling benzotriazolide of Boc-protected glycine S1c with unprotected mono-isopentapeptide 3 at −10 °C. The Boc group of compound 4 was deprotected by stirring with a concentrated solution of HCl in dioxane for 1 h to afford the HCl salt of unprotected mono-isohexapeptide 5. Compounds 4 and 5 were fully characterized by 1H, 13C NMR, and HRMS analysis (Scheme 2).

Figure 
                Scheme 2:
              .

 Synthesis of mono-isohexapeptides 5, 10, 15.

Chemical ligation via a 17-membered cyclic transition state was investigated by subjecting mono-isohexapeptide 5 to microwave irradiation at 50 °C, 50 W irradiation power for 3 h using 1 m NaH2PO4/Na2HPO4 phosphate buffer to maintain pH 7.3 (Scheme 3) (28). The reaction acidified with 2 N HCl to pH = 1. The mixture was extracted with ethyl acetate (3 × 20 mL), the combined organic extracts were dried over MgSO4, and the solvent was removed under reduced pressure. The ligation mixture was weighed and then a solution in methanol (1 mg/mL) was analyzed by HPLC-MS (Figures S3–S8 in Appendix S1).

Figure 
                Scheme 3:
              .

 Chemical Ligation reaction of mono-isohexapeptides 5, 10, 15.

HPLC-MS (ESI) analysis of the ligated mixture showed the expected migration product 6 (rt 25.47, m/z 611.1) together with intermolecular transacylation product 7 (rt 30.96, m/z 816.0) and the disulfide dimer of the expected ligated product 8 (rt 29.04, m/z 1219.1). It is well known (29) that small cysteine-containing peptides are easily oxidatively dimerized in solution when exposed to air in the absence of a reducing agent.

A small amount of reactant 5 (rt 23.25, m/z 611.1) was also present in the ligation mixture. HPLC-MS, via (-) ESI-MS/MS, confirmed that the ligated product 6 (rt 25.47, m/z 611.1) and starting mono-isohexapeptide 5 (rt 23.25, m/z 611.1) produced different MS fragmentation patterns (Figs S2, S4 and S5). The relative abundance of the crude ligated mixture as analyzed by analytical HPLC is shown in Table 1 (Figure S3 in Appendix S1). The result indicates the formation of ligated product (+ 8) with 60% relative abundance, which infers that S- to N-acyl group migration via a 17-membered transition state is preferred over intermolecular acylation. Thus, long-range acyl migration via a 17-membered cyclic transition state is feasible and may afford a promising approach for the synthesis of native peptide analogs (Table 1).

Table 1.   Chemical ligation of S-acyl isohexapeptides 5, 10, and 15
ReactCyclic TS sizeTotal crude yield (%) of products isolatedRelative amounts of each product (%)a,b,cProduct characterization by HPLC-MS
Ligated monomer (LM)Ligated dimer (LDTransacylation (TA)
RLMLDTALM[M + H]+LD[M + H]+TA[M + H]+
  1. aDetermined by HPLC-MS semiquantitative. The area of ion-peak resulting from the sum of the intensities of the [M + H]+ and [M + Na]+ ions of each compound was integrated.

  2. bR, recovered reactant; LM, ligation monomer; LD, ligation dimer; TA, transacylation.

  3. cAmounts are corrected for LD = 2 mmol, LM = 1 mmol.

5 17782185129 6 611.0 8 1219.1 7 816.0
10 1885111382 11 623.2 13 1245.5 12 828.3
15 19801121967 16 639.1 18 1275.3 17 844.1

This result is consistent with our previous work in which we have shown similar SN acyl migration by an intramolecular mechanism (23,24) and this was supported by competitive experiments (23) and computational arguments (30).

Study of SN acyl migration via an 18-membered cyclic transition state

We next investigated the possibility of SN acyl migration via an 18-membered cyclic transition state in mono-isohexapeptide 10, which was synthesized by coupling of Boc-β-Ala-Bt S1d with mono-isopentapeptide 3 under the conditions previously used for the formation of 4, followed by the deprotection of Boc protecting group. Compound 9 was isolated and fully characterized by 1H, 13C NMR, and HRMS analysis (Figures S11–S15 in Appendix S1) (Scheme 2).

The ligation experiment was conducted by subjecting mono-isohexapeptide 10 in 1 m phosphate buffer (pH = 7.3) to microwave irradiation at 50 °C, 50 W for 3 h (Scheme 3). After work-up, the crude ligated mixture and starting mono-isohexapeptide 10 were analyzed by HPLC-MS (ESI) (Figure S11 in Appendix S1).

The results showed almost complete consumption of 10 (rt 20.62, m/z 623.2) to produce the expected ligated product 11 (rt 18.06, m/z 623.2), together with its disulfide dimer 13 (rt.18.67, m/z 1245.5) and the intermolecular transacylation product 12 (rt. 20.75, m/z 828.3). HPLC-MS, via (-) ESI-MS/MS, confirmed that the ligated product 11 (rt 18.06, m/z 623.2) and starting mono-isohexapeptide 10 (rt. 20.62, m/z 623.2) produced different MS fragmentation patterns (Figures S10, S12, and S13 in Appendix S1). The relative abundance of transacylated product 12 is 82%, suggesting that intramolecular acyl migration 1011 through an 18-membered transition state is less favorable than 56 through a 17-membered transition state.

Study of SN acyl migration via a 19-membered cyclic transition state

Boc-protected mono-isohexapeptide 14, containing a γ-amino acid unit, was synthesized by coupling N-Boc-γ-aminobutyroyl benzotriazolide S1e with mono-isopentapeptide 3 at −10 °C to give 14 in 76% yield. Compound 14 was treated with dioxane/HCl mixture for 1 h to afford the hydrochloride salt of unprotected mono-isohexapeptide 15 (72%). Compounds 14 and 15 were fully characterized by 1H, 13C NMR, and HRMS analysis (Figures S18–S22 in Appendix S1) (Scheme 2).

The chemical ligation experiment on mono-isohexapeptide 15 was carried out under similar experimental conditions to those used for the ligation of 5 and 6 (Scheme 3).

On comparing the HPLC-MS (ESI) analyses of the starting mono-isohexapeptide 15 with crude ligated mixture (Figures S17 and S19 in Appendix S1), it was observed that the peak corresponding to starting 15 (rt 26.00, m/z 639.2) was diminished in the crude reaction mixture but ligated product 16 (rt 28.29, m/z 639.1), transacylated product 17 (rt 33.76, m/z 844.1), and disulfide dimer 18 (rt 32.17, m/z 1275.3) were present. The relative abundance of crude ligated mixture is shown in Table 1. HPLC-MS, via (−) ESI-MS/MS, confirmed that the ligated product 16 and starting mono-isohexapeptide 15 produced distinctively different MS fragmentation patterns (Figures S17 and S20 in Appendix S1). The ratio 16 + 18: 17 was 31: 67 (Table 1).

Computational rationalization

We rationalized the trends observed in the relative yields of internal ligation computationally using the same methodology (30) previously applied successfully to similar ligations, including full conformational searches and conformer scoring against a scoring function. As the first necessary step in internal ligation is bringing the terminal amino group into the proximity of the thioester carbon atom, it is logical to use the conformer in which the geometrical distance between the reactive centers is small enough as a prerequisite for higher yields. This distance b(N-C) is employed here as a simple but relevant scoring function to prioritize the conformers.

A full conformational search of isopeptide structures 5, 10, and 15 was performed using the MMX force field, as implemented in pcmodel v. 9.3 Serena software, Bloomington, IN, USA. The resultant conformers were ranked in the ascending order of the b(N-C) scoring function, and the best preorganized (with the smallest b(N-C) values) are shown in Figure 1. The scoring function values b(N-C) for structures 5, 10, and 15 are 3.55, 4.15, and 3.61 Å, respectively. These values introduce a relative steric preference of structures 5, 10, and 15 to give internal ligation products rather than transacylation products: > 15 > 10.

Figure 1.

 Best preorganised conformer of 5, 10, 15.

The three peptides under study have six of seven common amino acid units in the sequence X-Gly-Gly-Leu-Cys-(Cbz-Ala)-OH. Only X is variable, standing for Gly, β-Ala, and GABA in 5, 10, and 15, respectively. In turn, Gly (2-aminoacetic acid), β-Ala (3-aminopropionic acid), and GABA (4-aminobutyric acid) differ only in the number of carbon atoms separating the N-terminus from the C-terminus, which exerts varying steric influences to form the 17-, 18- and 19-membered ring transition states needed for the internal ligation. One of the local effects not accountable by molecular mechanics may be the possibility of forming an intramolecular hydrogen bond N-H…O in β-alanine which is missing in glycine and γ-aminobutyric acid. To check this possibility, quantum chemical calculations at the HF/6-31G* level of theory were carried out for the transoid and cisoid conformations of Gly, β-Ala, and GABA. As a result, only in β-Ala was the cisoid conformer slightly more stable by 2 kT than the sterically less demanding transoid conformer. The calculated higher stability of the cisoid conformer cannot be explained in any other way other then hydrogen bonding, and this allows one to hypothesize that the amino group in β-Ala is conformationally locked to some extent and thus is less available for ligation. This may be a reason for the preference of intermolecular acylation over internal ligation in isopeptide 10.

The potential energy surface for structure 5 allows a conformer in which the b(N-C) score is smallest among the three peptides under study. In addition, this structure affords a transannular hydrogen bond, which is absent in structures 10 and 15. Inspecting the best conformer shown in Figure 1A, one finds a close NH-O contact (2.39 Å, 162°) that can be classified as a hydrogen bond. It forms a 13-membered ring by connecting the Cbz carbonyl oxygen atom and the leucine’s N-H and is capable of stabilizing the conformation favorable for internal ligation.

The above conformational analysis, supported by the force field and quantum chemical calculations, delivers a valid theoretical explanation why the long-range acyl migration via a 17- or 19-membered cyclic transition state is preferred over intermolecular acylation, while for the isopeptide 10, the reverse is true.

Conclusion

In conclusion, our methodology affords the successful synthesis of cysteine-containing isopeptides having α-, β-, or γ-amino acid units. We studied acyl migration from S to the N-terminus of NH2 group under microwave irradiation and found the intramolecular acyl transfer through 17-TS appeared to be favored over those via 18-TS and 19-TS, which has also been supported by ab initio and force field calculations. In agreement with our previous work (15), the cyclic transition states studied in this study compared with previous are in the order of their sizes 5 > 17∼14 > 11 > 19 > 18 > 8. Thus, long-range chemical ligation via a 17-membered cyclic transition state may be a promising approach for the synthesis of native peptide analogs.

Acknowledgments

We thank Dr C. D. Hall for useful suggestions and English checking and Dr Jodie Johnson for his help in HPLC-MS analysis. This work was supported by King Abdulaziz University under grant No D006/431; we also thank the University of Florida and the Kenan Foundation for financial support.

Ancillary