Copolymerization of Styrene and a Cyclic Peptide

Published Date: 24 Jan 2018

Putting peptides into the backbone chain of polyolefins: the radical copolymerization of styrene and a cyclic peptide containing the disulfide bond

Anja C. Pauly^a, Daniel Rentsch^b and Fabio di Lena*^a.

Supporting Information

ABSTRACT: For the first time, a vinyl monomer such as styrene has been radically copolymerized with a cyclic peptide containing the disulfide bond. A new class of bio-hybrids is obtained in which the amino acid sequence is statistically distributed within the polymer’s backbone chain. The structure of the copolymer has been confirmed by means of conventional as well as diffusion-edited ¹H NMR, MALDI FT-ICR mass spectrometry, FT-IR spectroscopy, TGA, DSC, and a series of control experiments.

With the aim to combine the advantageous properties of biological macromolecules such as, for example, the biological function, molecular recognition, and chirality, with the solution properties, processability, etc. of synthetic macromolecules, polymer chemists have started to develop the so-called bio-hybrid polymers. Bioconjugates are the most studied class of bio-hybrids.¹ These are block copolymers in which a protein, polysaccharide or nucleotide is chemically linked to a synthetic polymer such as a polyolefin, polyether or polyester. In this type of structures, the constituent blocks maintain their individual properties, which make them, in many ways, similar to polymer mixtures. At odds with block copolymers, statistical copolymers do not exhibit the characteristics of polymer mixtures but behave like homogeneous materials with peculiar physical and chemical properties. Here we report the preparation of a new class of bio-hybrids in which, much like in statistical copolymers, an amino acid sequence is incorporated directly into the backbone chain of a polyolefin like polystyrene. The polymers are prepared by the radical ring-opening copolymerization²of a cyclic peptide containing the disulfide (S-S) bond and styrene. Cycles containing the S-S bond are known to undergo radical copolymerization with vinyl monomers such as methyl acrylate, vinyl acetate, acrylonitrile and styrene.³

The driving force behind the research is our interest in finding new, simple and industrially friendly ways to turn commodity polymers into specialty polymers with high added value. To our knowledge, the only examples of polyolefins containing amino acids in the backbone chain have been prepared by Wagener and co-workers by means of acyclic diene metathesis (ADMET)⁴ polymerization of dienes containing a single amino acid residue conducted in the presence of a ruthenium carbene catalyst.⁵ The approach we describe here is metal-free, enables the incorporation of sequences of amino acids and employs radical polymerization, a process with which more than 50% of all the polymers produced worldwide are made.

Scheme 1. Radical copolymerization of styrene with the cyclic tripeptide cCLC.

Styrene and the cyclic peptide S¹,S³-cyclo(L-cysteinyl-L-leucyl-L-cysteine), from now on referred to as cCLC (or CLC when ring-opened), were chosen as model monomers. They were reacted with a molar ratio of 94:6 in dimethyl sulfoxide

Table 1. Polymerization conditions, yield, number average molecular weight, polydispersity index, degradation temperatures, glass transition temperatures and CLC content of the copolymers.

1. 1M in DMSO.
2. After precipitation in water and dialysis in MeOH.
3. Determination by SEC in THF on the basis of polystyrene calibration.
4. Determination by comparison of the integrated peaks in the ¹H-NMR spectra of the isopropyl unit in CLC and the phenyl unit in polystyrene.

(DMSO) at 70 ºC for 12h with two different amounts of azobisisobutyronitrile (AIBN) affording the copolymers P1 and P2 (Scheme 1, Table 1). The copolymers were purified by precipitation in water and dialysis in methanol so as to remove, among the other possible impurities, unreacted cCLC and/or cCLC-derived by-products. The overall yield was equal to 40% for P1 and 43% for P2. When analysed by means of size exclusion chromatography (SEC), the copolymer P1, obtained by using a higher amount of AIBN, resulted to have a number average molecular weight () of 2,500 and a polydispersity index (PDI) of 1.79. On the other hand, P2, synthesized by using a smaller amount of AIBN, turned out to have a higher molecular weight () and a comparable PDI of 1.64 (Table 1). The SEC traces of both copolymers are shown in the Supporting Information (Figure S1).

The signals in the ¹H NMR spectra of P1 (Figure S2) and P2 (Figure 1A) could be assigned to both styrene and CLC units. On the one hand, the peaks at 0.87 ppm and 1.10 ppm, visible also in ¹H NMR spectrum of unreacted cCLC (Figure S3), could be assigned to the iso-propyl residue of CLC. On the other hand, the two groups of peaks at 1.54 and 1.92 ppm, and at 6.55 and 7.05 ppm correspond to the aliphatic and the aromatic protons of polystyrene, respectively. The remaining proton signals of CLC could be assigned with a lower degree of confidence due to the overlapping signals of solvent and/or polystyrene. By comparing the area underneath the peak at 0.87 ppm relative to the iso-propyl group of CLC with the area underneath the peak around 7 ppm relative to the phenyl ring of styrene, it was calculated that the peptide makes up 6 mol% of copolymer P1 and 9 mol% of P2. A different degree of co-monomer incorporation is not odd if one considers that the composition, like other properties of a polymer, is function of the chain length up to a critical value that depends on the specific system. It is then reasonable to assume that such critical value for had not been reached in the present case. The topic has been extensively investigated and the interested reader is referred to the literature for details.⁶ In the diffusion-edited mode, in which the ¹H NMR spectra were recorded applying a flow-compensated double-stimulated-echo with a gradient strength up to 40%,⁷ a similar set of signals were found for the styrene and CLC units (Figure 1B and S2). By exploiting the fact that the translational diffusion in solution is size-dependent, the diffusion-edited NMR is able to discriminate between signals relative to low and high molecular weight species.⁸ Since only the solvent signals disappeared, the NMR data are a strong indication that the peptide is incorporated into polystyrene rather than forming a physical blend with it. It is worth noting that the diffusion-edited NMR is not quantitative and thus the molar composition of the copolymers could be determined only from the conventional ¹H-NMR spectra. The analysis by MALDI FT-ICR mass spectrometry⁹ substantiates these conclusions. A mass distribution (Figure 2) that accurately matches that of monocharged polystyrene chains each containing one CLC moiety and AIBN-derived isobutyronitrile groups as both α- and ω-chain ends was indeed obtained.

Figure 1. ¹H-NMR spectra of the copolymer P2 (A), ¹H-diffusion edited ¹H-NMR spectra of the copolymer P2 with gradient strength of 40% (B) in THF-d⁸, and the corresponding chemical structure (C).

The results of all the other analytical techniques used to characterize the copolymers are in line with what found above. In the FT-IR spectra, for example, signals belonging to both styrene and amino acid moieties could be detected (Figure 3), which are: (i) the bands at 1735 cm^-1 (carboxylic group) and 1654 cm^-1 (amide group) of CLC, which are also present in the FT-IR spectrum of unreacted cCLC; and (ii) the signals of the aromatic carbon-carbon bonds (1492 and 1452 cm^-1) and carbon-proton bond of the phenyl rings (736 and 696 cm^-1) of

Figure 2. MALDI FT-ICR spectrum of the copolymer P2 in the positive mode (A), the magnification of the spectrum in the mass range 4600 – 5000 with the comparison of the theoretical and observed m/z (B), and the corresponding chemical structure (C).

polystyrene. Furthermore, two distinct mass losses, one around 200 ºC and the other at 417 ºC, can be seen in the thermogravimetric (TGA) traces of the copolymers P1 and P2 (Table 1). By direct comparison with the TGA of the constituting materials, which show a mass loss at 208 °C for unreacted cCLC and one at 418 °C for pristine polystyrene, the two steps observed in the TGA of both copolymers could be assigned to the degradation of the CLC and styrene units, respectively (Figure S4). The differential scanning calorimetry (DSC) thermogram of P1 displayed a glass transition occurring around 66 ºC, which is identical to the glass transition temperature (T_g) of a polystyrene of prepared in our lab (66 ºC). Therefore, the amount of CLC incorporated in the polymer turned out to be too low to produce a measurable effect on the glass transition. In contrast, the amount of CLC in the copolymer P2 turned out to be sufficient to produce a change in the glass transition temperature, which was measured to be 54 ºC (Table 1). This is significantly lower than T_g of polystyrenes with (75 ºC) and (89 ºC) prepared in our lab. The DSC scans of the two copolymers P1 and P2 in comparison with polystyrenes with similar molecular weight are shown in Figure S5. The relatively high T_g of polystyrene is classically rationalized in terms of a reduced chain flexibility due to the bulky phenyl groups that hinder the rotation of the backbone’s carbon-carbon bonds. We surmise that CLC increases the chain flexibility by acting as a spacer between the styrene units, which results in the lowering of the glass transition temperature. It is worth noting that the T_g and the of (atactic) polystyrene are positively correlated up to , after which the T_g reaches a stationary value of ca. 108 ºC.¹⁰ Hence, the use of polymers with similar molecular weights is essential for comparing, meaningfully, the glass transition temperatures.

In absence of cCLC, the polymerization of styrene under the same experimental conditions afforded polymers with in 76% yield and in 73% yield for the

Figure 3. FT-IR spectra of the cyclic tripeptide cCLC, the copolymer P2 and Polystyrene.

lower and higher amounts of AIBN, respectively. In both cases, the molecular weights and reaction yields for pristine polystyrene were higher than those of the relative copolymers. This is not surprising since disulfides are known to act as chain transfer agents in and to produce a certain retardation effect on radical polymerization.³ When the polymerization was repeated omitting the styrene from the reaction mixture, no polymer was obtained. Hence, cCLC, like other cyclic disulfides,² does not homopolymerize in the presence of a radical initiator. This control experiment suggests that the peptide should not be blockily distributed along the polymer chain. Moreover, the possibility that the copolymer could be alternating is ruled out by the fact that the degree of peptide incorporation is well below 50 mol%. It is therefore reasonable to assume that both P1 and P2 are statistical copolymers of styrene and CLC.

Peptides like cCLC are peculiar in that they bear unbound amine and carboxyl groups while being cyclic. This makes them and their copolymers either cationic or anionic or zwitterionic depending on the pH. Charge-bearing polymers are often reported as bioactive, e.g., hemostatic¹¹ and/or antimicrobial¹². Consequently, the class of materials here described might show bioactivity without containing intrinsically bioactive, amino acid sequences. Furthermore, apart from the specific functionalities, the peptide is likely to confer improved degradability on the polyolefin. Experiments in both directions are presently ongoing and will be the subject of another publication.

In conclusion, we have shown that a peptide sequence can be incorporated into the backbone chain of a polyolefin via radical polymerization. Styrene and a cyclic tripeptide containing the disulfide bond were chosen as model monomers. Although cyclic disulfides are known to ring-open via the homolytic cleavage of the S-S bond in the presence of certain radicals, the result reported in this work is not trivial since the efficiency of such a reaction depends significantly on the disulfide used. Investigations are presently underway in order to explore the monomer scope, in terms of both the olefin and the peptide, the bioactivity and degradability of the copolymers, as well as the possibility to extend the process to reversible-deactivation radical polymerizations¹³ such as ATRP¹⁴. The preparation of a whole new range of functional and degradable materials is anticipated.

ASSOCIATED CONTENT

Supporting Information

Detailed experimental procedures as well as spectroscopic, thermal and chromatographic data. This material is available free of charge via the Internet at http://pubs.acs.org.

REFERENCES

1.Lutz, J.-F.; Börner, H. G., Modern trends in polymer bioconjugates design. Prog. Polym. Sci. 2008, 33 (1), 1-39.

2.Sanda, F.; Endo, T., Radical ring-opening polymerization. J. Polym. Sci. A Polym. Chem. 2001, 39, 265–276.

3.(a) Stockmayer, W. H.; Howard, R. O.; Clarke, J. T., Copolymerization of vinyl acetate with a cyclic disulfide. J. Am. Chem. Soc. 1953, 75 (7), 1756-1757; (b) Tobolsky, A. V.; Baysal, B., The Reaction between styrene and ring disulfides: copolymerization effected by the chain transfer reaction. J. Am. Chem. Soc. 1953, 75 (7), 1757-1757; (c) Suzuki, T.; Nambu, Y.; Endo, T., Radical copolymerization of lipoamide with vinyl monomers. Macromolecules 1990, 23, 1579-1582.

4.Baughman, T. W.; Wagener, K. B., Recent advances in ADMET polymerization. Adv. Polym. Sci. 2005, 176, 1-42.

5.Hopkins, T. E.; Pawlow, J. H.; Koren, D. L.; Deters, K. S.; Solivan, S. M.; Davis, J. A.; Gomez, F. J.; Wagener, K. B., Chiral polyolefins bearing amino acids. Macromolecules 2001, 34, 7920-7922.

6.(a) Mirabella Jr, F. M.; Barrall Ii, E. M.; Jordan, E. F., Jr.; Johnson, J. F., Copolymer composition as a function of molecular weight and the effect of conversion on this relationship. J. Appl. Polym. Sci. 1976, 20 (3), 581-589; (b) Mirabella Jr, F. M.; Barrall Ii, E. M., Determination of copolymer composition as a function of molecular weight by preparative gel permeation chromatography and comparison to the rapid stop-and-go gpc/ir method. J. Appl. Polym. Sci. 1976, 20 (4), 959-965; (c) Mirabella Jr, F. M., Monte Carlo simulation of copolymerization and compositional inhomogeneity of copolymers: comparison to experimental data. Polymer 1977, 18 (7), 705-711.

7.Jerschow, A.; Müller, N., Suppression of convection artifacts in stimulated-echo diffusion experiments. Double-stimulated-echo experiments. J. Magn. Reson. 1997, 125 (2), 372-375.

8.(a) Chen, A.; Wu, D.; Johnson, C. S., Determination of Molecular Weight Distributions for Polymers by Diffusion-Ordered NMR. J. Am. Chem. Soc. 1995, 117 (30), 7965-7970; (b) Lucas, L. H.; Larive, C. K., Measuring ligand-protein binding using NMR diffusion experiments. Concept. Magn. Reson. A 2004, 20A (1), 24-41.

9.Zhang, L.-K.; Rempel, D.; Pramanik, B. N.; Gross, M. L., Accurate mass measurements by Fourier transform mass spectrometry. Mass Spec. Rev. 2005, 24 (2), 286-309.

10.Claudy, P.; Létoffé, J. M.; Camberlain, Y.; Pascault, J. P., Glass transition of polystyrene versus molecular weight. Polym. Bull. 1983, 9 (4-5), 208-215.

11.di Lena, F., Hemostatic polymers: The concept, state of the art and perspectives. J. Mater. Chem. B 2014, 2 (23), 3567-3577.

12.Muñoz-Bonilla, A.; Fernández-García, M., Polymeric materials with antimicrobial activity. Prog. Polym. Sci. 2012, 37 (2), 281-339.

13.Shipp, D. A., Reversible-Deactivation Radical Polymerizations. Polym. Rev. 2011, 51 (2), 99-103.

14.Matyjaszewski, K.; Tsarevsky, N. V., Macromolecular engineering by atom transfer radical polymerization. J. Am. Chem. Soc. 2014, 136 (18), 6513-6533.

ToC

1