Kristina
Klinker
a,
Regina
Holm
b,
Philipp
Heller
a and
Matthias
Barz
*b
aGraduate School Materials Science in Mainz, Staudinger Weg 9, 55128 Mainz, Germany
bInstitute of Organic Chemistry, Johannes Gutenberg-Universität Mainz, Duesbergweg 10-14, 55128 Mainz, Germany. E-mail: barz@uni-mainz.de; Fax: +49 6131-39-24778; Tel: +49 6131-39-26256
First published on 1st May 2015
In this work, we describe the synthesis of block copolypeptides, polypeptoids and block copolypept(o)ides by chemical ligation techniques. Polysarcosine (PSar), poly(N-ε-trifluoroacetyl-L-lysine) (PLys(TFA)) and poly(γ-benzyl-L-glutamate) (PGlu(OBzl)) homopolymers of different polarities and end group functionalities but with similar average degrees of polymerization (Xn = 50 and 100) could be obtained by ring opening polymerization (ROP) of α-amino acid N-carboxyanhydrides (NCA) and postpolymerization modification reactions. In the next step, these polymers were applied to copper(I)-catalyzed azide–alkyne coupling (CuAAC), strain-promoted azide–alkyne coupling (SPAAC) and native chemical ligation (NCL). Our results suggest that all the employed ligation techniques can be used for the synthesis of block copolypeptides, polypeptoids and block copolypept(o)ides. SPAAC displayed, for most conditions, the highest ligation efficiencies (up to 86%) and, from a practical point of view, is the most feasible method. NCL, however, performed very well for short hydrophilic polymers (up to 88%) and is favourable for the ligation of peptides from solid phase peptide synthesis (SPPS) to polysarcosine. As a proof of principle, we report a protocol for an efficient NCL coupling of polysarcosine to the T-cell receptor core peptide (TCR CP), which is known to inhibit IL-2 production in antigen-stimulated T cells and, therefore, to suppress inflammation. In a comparative study, the ligation method, which directly relates to the chemical nature of the ligation site, neither influenced cytotoxicity nor complement activation. To conclude, chemical ligation techniques represent a complementary synthetic approach to the well-established sequential ring opening polymerization of NCAs.
These polymers – unlike most other biocompatible synthetic polymers – have the potential to combine the hydrolytic stability with in vivo degradability. In the case of enzymatic degradation, polypeptides can decompose into α-amino acids, which may be further metabolized in the organism, avoiding design limitations due to renal excretion or storage diseases.7,8 For these reasons, polypeptides have become intriguing candidates as materials in drug delivery applications.9 Although polymer–protein conjugates have already entered clinical practice, polymeric micelles, polyplexes and nanohydrogels are expected to shape the future of nanomedicine. These systems combine a functional core with a stealth-like corona and are formed by electrostatic/entropic or polarity driven self-assembly of block copolymers. While in the first case, the cargo (siRNA, mRNA or pDNA) shapes the properties of the system, the size and shape of the second are mainly determined by the hydrophobic–hydrophilic ratio and the chemical nature of the individual polymer blocks within the amphiphilic block copolymer.10,11
Recently, Birke et al. and Heller et al. have introduced polypept(o)ides as a class of functional block copolymers based on the polypeptoid polysarcosine (PSar) and functional polypeptides.12–15 PSar – poly(N-methyl glycine) – seems to be a particularly interesting alternative to polyethylene glycol (PEG) since it shows very similar physicochemical properties. It is soluble in aqueous solution and in organic solvents, e.g., DMSO or chloroform, and is exclusively a weak hydrogen bond acceptor, which prevents the formation of secondary structures other than random coils.16,17 In addition, a similar protein resistance to PEG has been reported.16,18 Kimura and coworkers as well as our research group have investigated PSar as a hydrophilic block in amphiphilic block copolymers as an alternative to PEG.19,20
Block copolypept(o)ides are readily synthesized by the sequential polymerization of NCAs, where the first block acts as the macroinitiator for the second block. However, chemical ligation techniques are thought to present an important alternative pathway to block copolypept(o)ides whenever a functional block is required, which cannot be derived by NCA polymerization. This can be the case if the second block needs to possess a specific sequence and is, thus, synthesized by solid phase peptide synthesis (SPPS). Also, the second block might not tolerate deprotection conditions that are mandatory for the other. The class of chemical ligation reactions mostly relies on the selective reaction between two (bio)orthogonal groups, which are ideally absent from peptides and proteins (or biomolecules in general). Amongst these are azides, alkynes, tetrazine and many others that have been widely used for different in vitro and in vivo applications. They range from the labeling of biomolecules with fluorescent probes, affinity tags and isotope labels to the synthesis of glycopeptides and biohybrids such as polymer–protein conjugates, cell-surface labeling and the immobilization of proteins.21–24
Efficient ligation methods are characterized by a very high selectivity for product formation, easy purification and mild reaction conditions (room temperature and physiological pH), which is essential for many proteins in order to maintain their activity.22,25
Probably one of the best known chemical ligation techniques is the classical “click” reaction: CuI-catalyzed cycloaddition of azides and terminal alkynes (CuAAC) developed by Sharpless26 and separately by Meldal and co-workers.27 In contrast to Huisgen's 1,3-dipolar cycloaddition,28,29 only the 1,4-substituted stereoisomer is obtained under much milder reaction conditions. CuAAC can be conducted in different solvents, including water and buffered saline at physiological pH. To avoid the use of copper(I) as a catalyst, especially with regard to in vivo applications, the strain-promoted alkyne–azide cycloaddition (SPAAC) has been developed by Bertozzi et al.30 The initial idea dates back to discoveries of Wittig and Krebs who reported that cyclooctynes react with azides to form triazoles.30,31 For non-substituted cyclooctynes, the reaction rates are relatively slow (0.0012 M−1 s−1), but the attachment of electron-withdrawing groups or additional ring-strain has been shown to significantly improve reaction rates.32 Azide–alkyne coupling reactions have been widely used in very different fields, including chemical biology, drug discovery, materials science and polymer science.33–35 There are many examples where click chemistry has been successfully employed in the synthesis of nanocarriers for drug delivery.36,37
Native chemical ligation (NCL) is another interesting ligation technique since it results in an amide bond at the ligation site instead of an electron-poor aromatic system. The method was developed by Dawson and Kent in the early 90s.38,39 Today, it is one of the most effective practical methods for the total synthesis of proteins by using fully unprotected peptide fragments. This is an important aspect when considering the fact that coupling of protected peptide fragments is often limited by the insolubility in aqueous media of longer peptide chains.40 It has recently been shown by Dittmann et al. that NCL can be performed in polar aprotic organic solvents like DMF with the addition of a base without racemization, which is potentially very useful for the ligation of hydrophobic peptide segments.41,42
While a lot of research has been devoted to different kinds of ligation strategies, a systematic comparison of such methodologies for the synthesis of block copolymers is missing. Here, we report a comparative study of chemical ligation techniques, namely CuAAC, SPAAC and NCL, for the synthesis of block copolypeptides, polypeptoids and block copolypept(o)ides. Polymers used for the individual ligations differ in the chain length, polarity and chemical nature of end groups. We report the optimized conditions for the individual reactions and provide coupling efficiencies. Furthermore, we describe the synthesis and purification of a polysarcosine–peptide conjugate derived by the described methodology, which is of amphiphilic nature. The employed T-cell receptor core peptide (TCR CP) is known to inhibit IL-2 production in antigen-stimulated T cells.43
Infrared spectroscopy was performed on a Jasco FT/IR-4100 with an ATR sampling accessory (MIRacle, Pike Technologies) using 16 scans per measurement. IR spectra were analyzed using Spectra Manager 2.0 (Jasco). GPC (Gel permeation chromatography) in HFIP was performed with 3 g l−1 potassium trifluoroacetate at 40 °C. The columns were packed with modified silica (PFG columns, particle size: 7 μm, porosity: 100 and 1000 Å). A refractive index detector (G 1362A RID, Jasco) and a UV/vis detector (UV-2075 Plus, JASCO) was used to detect the polymer. Molecular weights were calculated using calibration performed with PMMA standards (Polymer Standards Services GmbH). As the internal standard, toluene was used. Melting points were determined on Mettler Toledo FP62 melting point apparatus at a heating rate of 5 °C min−1. For the MTT assay, a CellTiter 96® Non-Radioactive Cell Proliferation Assay kit was purchased from Promega (Madison, USA). Adsorption was measured using a Dynex MRX TC Revelation microtiter plate reader (Dynex Technologies, Chantilly, VA).
A Complement Component 5a (C5a) ELISA kit was obtained from antibodies-online GmbH (Aachen, Germany).
13C NMR (400 MHz, CDCl3) δ [ppm] = 49.13 (CH2–N3), 39.31 (CH2–NH2), 32.43 (CH2–CH2–CH2).
IR azide ν [cm−1] = 2088.5.
Sarcosine (15.16 g, 170.2 mmol, 1 eq.) was weighed into a pre-dried three-necked flask and dried under vacuum for 1 hour. 300 mL absolute (abs.) THF was added under a steady flow of nitrogen. The apparatus was connected to two gas washing bottles filled with an aqueous sodium hydroxide solution. Diphosgene (16.26 ml, 134 mmol, 0.8 eq.) was added slowly via a syringe. The colorless suspension was heated to 70 °C, yielding a clear solution after 3 hours of stirring. The solvent was evaporated under reduced pressure, yielding a brown oil as the crude reaction product. The oil was heated to 50 °C and dried under reduced pressure to obtain an amorphous solid. The crude reaction product was redissolved in 60 mL THF and precipitated with 300 mL abs. n-hexane. The precipitate was filtered off under a N2 atmosphere and dried with a stream of dry nitrogen for 60–90 minutes to remove the residual traces of solvents. The next day, the product was dried under high vacuum for 2 hours in the sublimation apparatus and subsequently sublimated at 80–85 °C and <1 × 10−2 mbar. The product was collected from the sublimation apparatus in a glove box on the same day. Colorless crystals were obtained (50–67%). mp = 104.3 °C; 1H NMR (300 MHz, CDCl3) δ[ppm] = 2.86 (s, 3H, NH–CH3), 4.22 (s, 2H, NH–CH2–CO).
9.75 g (41.1 mmol, 1 eq.) of H-Glu(OBzl)-OH were weighed into a three-necked round-bottomed flask equipped with a magnetic stir bar, a condenser, a rubber septum and a glass stopper and dried under vacuum. After 30 minutes, 100 mL abs. tetrahydrofuran (THF) was added to the solid, resulting in a colorless suspension. The suspension was heated to 70 °C, and diphosgene (4 mL, 32.9 mmol, 0.8 eq.) was added to the mixture and stirred for 1.5 h. The suspension turned into a clear, yellow solution. After the reaction mixture had cooled down, a stream of dried N2 was passed through the mixture overnight to remove the excess of HCl and diphosgene. The next day, the resulting orange oil was evacuated to dryness. The product was purified by recrystallization from THF and n-hexane twice. The colorless precipitate was isolated by filtration under N2. 7.93 g (0.03 mmol, 73%) of colorless crystals were obtained. mp = 93.7 °C; 1H NMR (400 MHz, CDCl3) δ [ppm] = 2.23–2.00 (m, 1H, β-CH2), 2.35–2.22 (m, 1H, β-CH2), 2.60 (t, 2H, 3JH,H = 6.7 Hz, γ-CH2), 4.45–4.30 (m, 1H, α-CH), 5.14 (s, 2H, –OCH2), 6.3 (s, 1H, CH–NH–CO), 7.42–7.32 (m, 5H, –benzyl).
Polymerization using 3-azido-1-aminopropane, propargyl-amine and DBCO-amine was carried out in the same way. Yields were 91 and 96% (3-azido-1-aminopropane), 88 and 91% (propargylamine) and 99 and 100% (DBCO-amine). Initiator signals used for integration: 3-azido-1-aminopropane 1H NMR (400 MHz, DMSO-d6) δ [ppm] = 1.71–1.61 (m, 2H, –CH2–CH2–N3). DBCO-amine 1H NMR (400 MHz, DMSO-d6) δ [ppm] = 7.86–7.10 (m, 8H, benzylic protons); initiator signals from propargylamine were not visible due to overlap with polymer backbone signals.
Polymerization using propargylamine, 3-azido-1-amino-propane and DBCO-amine as initiators was carried out in the same way. Yields were 88 and 77% (3-azido-1-aminopropane), 83 and 95% (DBCO-amine), 89 and 74% (propargylamine) for different chain lengths. Initiator signals used for integration: 3-azido-1-aminopropane 1H NMR (400 MHz, DMSO-d6) δ [ppm] = 1.73–1.65 (m, 2H, CH2–CH2–CH2), 3.18–3.10 (m, 2H, CH2–NH2); propargylamine 1H NMR (400 MHz, DMSO-d6) δ [ppm] = 3.07 (m, 1H, –CCH), initiator signals from DBCO-amine could not be detected due to overlap with polymer backbone signals.
Polymerization using 3-azido-1-aminopropane was carried out in the same way as described above. Yields were 73 and 77% for different chain lengths. Initiator signals from 3-azido-1-aminopropane could not be detected due to overlap with polymer backbone signals.
183.0 mg (0.073 mmol, 1 eq.) of PSar were weighed into a flame-dried Schlenk flask and dried under vacuum for 30 minutes. The polymer was then dissolved in 2 mL of dry DMF. In another Schlenk flask, 163.9 mg (0.731 mmol, 10 eq.) S-benzyl-thiosuccinic acid, together with 305.4 mg (0.805 mmol, 11 eq.) N,N,N′,N′-tetramethyl-O-(1H-benzotriazol-1-yl)uronium hexafluorophosphate (HBTU) and 108.4 mg (0.802 mmol, 11 eq.) hydroxybenzotriazole (HOBt) were dissolved in 2 mL of dry DMF while cooling the flask in an ice bath. 273.1 μl (207.6 mg, 1.606 mmol, 22 eq.) of DIPEA were directly added and the mixture was stirred for 30 minutes with cooling. The resulting activated ester was subsequently added to the polymer solution and stirred at room temperature for 65 h.
For purification, the reaction mixture was precipitated in diethyl ether and centrifuged (4500 rpm at 4 °C for 15 min). After discarding the liquid fraction, new Et2O was added and the polymer was resuspended using sonication. The suspension was centrifuged again and the procedure was repeated. The polymer was then dissolved in H2O and dialyzed for 4 days against MilliQ-water (MWCO = 1000 g mol−1 or 3500 g mol−1, depending on the molecular weight) and lyophilized. For different chain lengths, yields ranged from 67 to 90%. Degrees of end group modification were quantitative according to 1H NMR (additional signal in 1H NMR (400 MHz, DMSO-d6) δ [ppm] = 7.19–7.34 (m, 5nH, –benzyl)). End group modification with Boc-L-Cys(Trt) was carried out in the same way. Additional signals in 1H NMR (400 MHz, DMSO-d6) δ [ppm] = 7.24–7.34 (m, 15nH, –trityl), 1.36 (s, 9nH, –C(CH3)3).
PLys(TFA) polymers were precipitated in diethyl ether and extensively dialyzed against H2O/MeOH 1:9. Polymers were freeze-dried from H2O/MeOH. For different chain lengths, yields ranged from 49 to 71%. Degrees of end group modification ranged from 94% to quantitative according to 1H NMR (additional signals from Boc-L-Cys(Trt) modification in 1H NMR (400 MHz, DMSO-d6) δ [ppm] = 7.17–7.39 (m, 15nH, –trityl)).
While C-terminal azide or alkyne functionalities were introduced via primary amine initiators in the ring opening polymerization of α-amino acid N-carboxyanhydrides, N-terminal functionalities, namely cysteine and a thioester moiety, have been introduced by postpolymerization modification using peptide-coupling chemistry.
The polymers have been synthesized from the corresponding NCAs as previously reported (see Scheme 2).12,48 The employed NCAs, namely SarNCA, Lys(TFA)NCA and Glu(OBzl)NCA, were synthesized according to the Fuchs–Farthing method.49,50 Melting points (Glu(OBzl)NCA: mp = 93.7 °C, Lys(TFA): mp = 101 °C, SarNCA: mp = 104.3 °C) were in good agreement with the literature or even higher than those reported which accounts for high purity.51–53 All the synthesized NCAs were stored at −80 °C under a nitrogen atmosphere over several months without any detectable degradation or oligomerization.
Scheme 2 (A) Homopolymer synthesis by NCA polymerization; (B) end group functionalization of PSar (with 1 and 2) and of PLys(TFA) (only with 1). |
While heterotelechelic homopolymers with alkyne or azide end groups have been synthesized directly using 3-azido-1-aminopropane (APA), propargylamine (PA) and DBCO-amine as the initiators for the ROP, polypept(o)ides for NCL were further modified at the N-terminus to incorporate S-benzyl thiosuccinic acid and Boc-L-cysteine(Trt). Polymers were characterized by HFIP-GPC and 1H NMR. Analytical data of homopolymers are summarized in Table 1.
Polymer | Initiator | X n(NMR)a | M n(GPC)b(kg mol−1) | Đ GPC | |
---|---|---|---|---|---|
a X n was determined by 1H NMR in DMSO-d6 by relating end group signals to those of the repeating unit. b M n was determined relative to PMMA standards. | |||||
P1 | PSar | NPA | 58 | 16.1 | 1.06 |
P2 | PSar | NPA | 117 | 24.3 | 1.09 |
P3 | PSar | NPA | 46 | 13.3 | 1.05 |
P4 | PSar | APA | 44 | 11.8 | 1.05 |
P5 | PSar | APA | 96 | 20.6 | 1.08 |
P6 | PSar | PA | n.d. | 14.9 | 1.08 |
P7 | PSar | PA | n.d. | 25.0 | 1.10 |
P8 | PSar | DBCO | 50 | 14.7 | 1.06 |
P9 | PSar | DBCO | 119 | 24.6 | 1.06 |
P10 | PGlu(OBzl) | APA | 64 | 14.8 | 1.05 |
P11 | PGlu(OBzl) | APA | 108 | 21.7 | 1.07 |
P12 | PGlu(OBzl) | PA | 72 | 14.7 | 1.05 |
P13 | PGlu(OBzl) | PA | 117 | 19.9 | 1.06 |
P14 | PGlu(OBzl) | DBCO | n.d. | 14.6 | 1.07 |
P15 | PGlu(OBzl) | DBCO | n.d. | 24.9 | 1.07 |
P16 | PLys(TFA) | NPA | 49 | 9.0 | 1.14 |
P17 | PLys(TFA) | NPA | 111 | 18.5 | 1.08 |
P18 | PLys(TFA) | APA | n.d. | 8.8 | 1.11 |
P19 | PLys(TFA) | APA | n.d. | 17.5 | 1.11 |
The degree of polymerization was calculated by 1H NMR by comparing the signal of the end group with those of the repeating units. Monomodal Poisson-like molecular weight distributions were ensured by HFIP GPC. Unfortunately, for some polymer–end group combinations, the degree of polymerization (Xn) could not be obtained by 1H NMR due to an overlap of end group proton signals with polymer backbone signals (see the Experimental section). A comparison of Mn(GPC) and ĐGPC values of these polymers with their homologues with different end groups ensured that even in the absence of NMR studies, these polymers are of comparable chain length.
In general, homopolymers possess low dispersities (Đ ≤ 1.1) and high end-group integrities (see Table 1). This is remarkable since especially in the case of PLys(TFA), the dispersities determined by GPC are even overestimated due to the coexistence of secondary structures differing in hydrodynamic volume as previously reported by Huesmann and coworkers.48 For PSar and PGlu polymers, dispersities are below 1.1 and for PLys(TFA), below 1.2 (see Table 1) and GPC elugrams indicate Poisson-like molecular weight distributions (see the ESI†).
Neopentyl-terminated polymers were further modified at the N-terminus by standard peptide coupling chemistry using HOBt/HBTU/DIPEA (see Fig. 1). PSar was functionalized with S-benzyl thiosuccinic acid (BTS) and Boc-L-Cys(Trt) (Cys) respectively, and PLys(TFA) was functionalized with Cys. Modification of polypeptides with BTS did not give reliable degrees of end group modification. Although it has been shown that N-terminal thioester synthesis can be achieved for low molecular weight peptides,47 the required long reaction times (2 days) for quantitative N-terminal modification led probably to aminolysis of the thioester, resulting in an inactivation of the end group by succinimide formation. Cyclisation cannot occur for PSar because of its methyl-substituted N-terminus. The degree of end group modification was determined by comparing the integral of the neopentylamine initiator's t-butyl group at 0.83–0.85 ppm with the integral of the benzylic protons’ signal of S-benzyl thiosuccinic acid (5H, 7.2–7.3 ppm) and of the trityl protons’ signal of Boc-L-Cys(Trt), respectively (15nH, 7.2–7.3 ppm and 9nH, 1.35 ppm), in DMSO-d6. For modification of PLys(TFA), only the trityl protons’ signals were clearly detectable and used for the determination of the degree of end group modification (see Fig. 1A). The determined degrees of modification and analytical data of end group modified polymers can be found in Table 2. Within the limits of accuracy, NMR data suggest nearly quantitative modification for the synthesized polymers (see Table 2). In addition, MALDI-TOF MS analysis of PSar polymers (Xn = 50) containing five different functionalities (PA, DBCO, APA, NPA + BTS, NPA + Boc-L-Cys(Trt)) needed for ligation reactions further confirms the NMR data (see the ESI†).
Fig. 1 (A) 1H NMR spectrum of PLys(TFA)100-Boc-L-Cys(Trt) before (red) and after (green) deprotection with TFA using TIPS as the scavenger for 30 minutes; (B) DOSY 1H NMR spectrum of PSar50-BTS. |
Polymer | End group | X n(NMR)a | Đ GPC | M n(GPC)b(kg mol−1) | % Mod.c |
---|---|---|---|---|---|
a X n was determined by 1H NMR in DMSO-d6 by relating end group signals to those of the repeating unit. b M n was determined relative to PMMA standards. c % Mod. was determined by 1H NMR in DMSO-d6. | |||||
PSar (P1) | BTS | 60 | 1.06 | 14.9 | Quant. |
PSar (P2) | BTS | 124 | 1.09 | 22.5 | Quant. |
PSar (P3) | Cys | 44 | 1.12 | 11.9 | Quant. |
PSar (P2) | Cys | 117 | 1.25 | 18.8 | Quant. |
PLys(TFA) (P16) | Cys | 53 | 1.12 | 8.9 | 94 |
PLys(TFA) (P17) | Cys | 99 | 1.11 | 18.5 | Quant. |
The covalent attachment of the respective end groups to the polymer was verified by diffusion ordered spectroscopy (DOSY) (see Fig. 1B).
Finally, polymers modified with Boc-L-Cys(Trt) were subsequently deprotected in TFA for 30 minutes at room temperature using TIPS as a scavenger. The complete deprotection was monitored by 1H NMR (see Fig. 1A).
Deprotection in TFA was chosen since PSar and PLys(TFA) are both soluble in TFA, and for purification the reaction mixture was precipitated in diethyl ether to easily remove the residual scavenger and the resulting hydrophobic products of the removed protecting groups (Trt).
Ligation experiments were performed for hydrophilic, hydrophilic/hydrophobic (amphiphilic) and hydrophobic conditions to provide coupling protocols for the synthesis of block copolymers consisting of blocks with different polarities. The employed polymer combinations are listed in Table 3. Two different degrees of polymerization (M/I ratio = 50, 100) were designed to study the influence of the chain length on the ligation efficacy. PSar was employed as a hydrophilic and non-ionic block, while PLys(TFA) and PGlu(OBzl) served as hydrophobic blocks, with PLys(TFA) being slightly more polar than PGlu(OBzl) due to the amide bond in the side chain. Ligations were carried out between polypept(o)ides with similar chain lengths (50/50 and 100/100) to facilitate the analysis of coupling product formation by HFIP GPC.
Ligation type | Hydrophilic | Amphiphilic | Hydrophobic |
---|---|---|---|
CuAAC | PSar/PSar | PSar/PGlu(OBzl) | PLys(TFA)/PGlu(OBzl) |
SPAAC | PSar/PSar | PSar/PGlu(OBzl) | PLys(TFA)/PGlu(OBzl) |
NCL | PSar/PSar | PSar/PLys(TFA) | Not accessible |
NCL was only performed for hydrophilic and amphiphilic conditions for the reasons explained above and GPC samples were pretreated with TCEP·HCl to avoid misinterpretation by oxidation (dimer formation). All ligations were performed at 1:1 molar ratio of polymers and 40 °C for 70 h to ensure comparability between individual ligation reactions.
To quantify relative coupling efficiencies, GPC traces were fitted by Gaussian distributions (see Fig. 2), from which the peak areas were used to calculate the respective contents. This methodology is enabled since the GPC plots of homopolymers themselves can be fitted with negligible deviation by Gaussian distributions as expected for polymers having a Poisson-like molecular weight distribution. Ranges of the determined efficiencies are shown in Fig. 2 as well as elugrams of ligations between short polypept(o)ides (50/50) under hydrophilic conditions for CuAAC, SPAAC and NCL. The corresponding fit curves used for efficiency determination are also displayed.
Fig. 2 HFIP GPC elugrams of ligation products from CuAAC, SPAAC and NCL (Xn = 50, hydrophilic) and Gauss-fits used to determine efficiencies displayed in the table below. |
The determined efficiencies for individual ligation experiments ranged from 53 to 88% (see the ESI†). It has to be noted that the residual unreacted homopolymers can also be attributed to a non-equimolar ratio of the two employed homopolymers, since due to the intrinsic uncertainty of the determined chain lengths, an exact 1:1 ratio of functional end groups is hardly achievable. Coupling efficiencies are likely to be slightly overestimated for amphiphilic CuAAC ligation products, since copper was removed by precipitation in water/MeOH to enable NMR analysis. This was done in order to prove triazole formation (see the ESI†). In general, precipitation of the coupling reaction is not expected to significantly effect the molecular weight distribution of ligation products, since both individual homopolymers already precipitate under the reported conditions. We want to emphasize here that coupling efficiencies were primarily determined for better comparability of the different techniques. Nevertheless, these results indicate that block copolypept(o)ide synthesis is feasible with all three presented ligation techniques with good efficiencies.
As expected, in general, ligations between shorter polypeptides are easier to realize than ligations between longer polypeptides/polysarcosine. This is likely due to the increased steric demand and the decreased probability of meeting between reactive chain ends. Generally, azide–alkyne coupling reactions gave better results than NCL even if NCL showed very high (up to 88%) coupling efficiencies for hydrophilic conditions. When the two different types of azide–alkyne coupling reactions are directly compared, SPAAC seems to be slightly superior to CuAAC. Not only are coupling efficiencies slightly better compared to CuAAC, especially for amphiphilic and hydrophobic conditions, but also, for SPAAC, no additional catalyst is necessary. Therefore, the reaction is very simple to conduct and it is not required to work in dry solvents or under inert conditions. When functionalities for click chemistry can be easily introduced into both blocks as shown for homopolymers in this study, SPAAC appears to be the method of choice. On the other hand, chain length determination by 1H NMR is a lot more reliable when neopentylamine (9 equivalent protons) or other initiators with clearly detectable signals are used.
In the case of a peptide already bearing an N-terminal Cys residue (synthesized either by SPPS or by biotechnological means), NCL can become the preferred method. However, NCL is optimized for biological applications and is only very easy to conduct in aqueous buffered solutions. In organic solvents, however, the reaction demands working under absolutely dry conditions to avoid hydrolysis of the thioester. If no reducing agent is added, it is also necessary to work under oxygen free conditions to avoid cystine formation, which inhibits the cysteine to further take part in chemical ligation. For the synthesis of amphiphilic conjugates, also mixtures of organic solvents and aqueous buffered solutions are an option.
Nevertheless, it needs to be kept in mind that end group modification might not always be quantitative and is also dependent on the chain length, which further limits this method for block copolymer synthesis.
In this work, the main focus was set on a direct comparison between different ligation techniques and for the purpose of feasible monitoring. Purification of block copolymers, e.g. block copolypept(o)ides, however, can be easily conducted whenever two polymers or peptides, which differ significantly in chemical properties like polarity (solubility) or in size, are used as for example shown by Lecommandoux and coworkers.44
Thus, the ligation approach is also a complementary pathway to the sequential ring opening polymerization for the synthesis of amphiphilic block copolypept(o)ides.
Fig. 3 Concentration of C5a determined by ELISA after incubation of L1(CuAAC), L2(SPAAC) and L3(NCL) with human serum for 1 h at 37 °C. |
Hence, these preliminary experiments suggest that the nature of the ligation site neither affects toxicity nor complements activation of these polymers, which is a first crucial requirement for application in nanomedicine. Nevertheless, future research needs to address different types of immune responses apart from complement activation.
PSar124-BTS was conjugated to TCR CP in a mixture of 4 M urea buffer/THF 1:1 in the presence of MPAA and TCEP·HCl, using an excess of TCR CP (2 eq.). After purification by precipitation and dialysis, the conjugate was analysed by 1H NMR, DOSY and HFIP GPC.
DOSY data (see Fig. 4) clearly demonstrate covalent attachment to the polymer since the corresponding signals appear as one diffusing species. After the successful synthesis of the conjugate, the HFIP GPC elugram displayed bimodal molecular weight distribution with a second peak at double molecular weight (see Fig. 4). This second peak disappears after treatment with TCEP·HCl overnight in H2O. The higher molecular weight peak can thus be attributed to dimer formation, which only becomes possible once the free thiol moiety is incorporated into the polymer, further proving the effective coupling of the decapeptide to polysarcosine.
Fig. 4 (A) DOSY 1H NMR of PSar124-TCR CP; (B) HFIP GPC elugram before and after conjugation with TCR CP; (C) reaction scheme of conjugation. |
In conclusion, a well-defined conjugate can be obtained under reductive conditions, which can be easily purified by precipitation and dialysis. The conjugate is currently under investigation with respect to inhibition of IL-2 production in different inflammation models.
Footnote |
† Electronic supplementary information (ESI) available: 1H NMR spectra of homopolymers, S-benzyl thiosuccinic acid, 3-azido-1-aminopropane, and IR spectra of azide polymers. 1H NMR spectra of ligation products by CuAAC for short chain lengths (Xn = 50). HFIP-GPC elugrams and fits of all the performed ligation experiments (and of the employed homopolymers respectively), and MALDI-TOF spectra for PSar50 with all five end group modifications. See DOI: 10.1039/c5py00461f |
This journal is © The Royal Society of Chemistry 2015 |