Synthesis of thermo-responsive glycopolymersviacopper catalysed azide–alkyne ‘click’ chemistry for inhibition of ricin: the effect of spacer between polymer backbone and galactose

Abstract

Block copolymers poly(diethylene glycol methyl ether methacrylate)-block-poly(2-hydroxy ethyl methacrylate) (PDEGMA-b-PHEMA) and poly(diethylene glycol methyl ether methacrylate)-block-poly(ethylene glycol methyl ether methacrylate) PDEGMA-b-PPEGMA as well as homopolymers of PPEGMA and PHEMA were generated via Reversible Addition–Fragmentation chain Transfer (RAFT)polymerisation. The terminal hydroxyl moieties on the HEMA and PEGMA blocks were alkyne functionalised and conjugated to azide functionalised galactoseviaCuAAC leading to the glycopolymers PDEGMA-b-P(HEMA-Gal) and PDEGMA-b-P(PEGMA-Gal). Both dynamic light scattering (DLS) and transmission electron microscopy (TEM) prove the thermo-responsive characteristics of the polymers, due to the PDEGMA blocks, with the formation of polymeric micelles above a temperature of 40 °C. These glycopolymers were tested for their binding efficiency to ricin, a dangerous plant toxin reportedly weaponised by terrorist organisations. The binding experiments establish an order of ricin inhibition of P(HEMA-Gal) < P(PEGMA-Gal) < PDEGMA-b-P(HEMA-Gal) ≤ PDEGMA-b-P(PEGMA-Gal) showing that micelles are advantageous, but also that the introduction of a flexible spacer between polymer and galactose can enhance binding to a certain extent.

---

Introduction

Carbohydrates hold great significance to life due to their role in important biological functions such as immune response and cell recognition.^1–3

In recent times, the scientific community has shown that synthetic glycopolymers are a class of polymers that have high potential for biomedical applications due to their ability to interact with carbohydrate moieties and lectins in a manner similar to natural glycoproteins.³ The key to successful carbohydrate–lectin interaction lies in multivalent interactions, which result in specificity and high affinity. This effect, possible by having multiple saccharide functionalities on polymers, is known as the ‘cluster glycoside effect’.⁴

Ricin is a well-known plant-derived toxin that has reportedly been weaponised for use in terrorism.⁵ It can be easily isolated from castor beans and there are presently no known countermeasures for it.^5,6 Binding to galactose can block its cytotoxic action,² although bonds with single galactose moieties are generally weak and ineffective. As with all carbohydrate–lectin interactions the multivalency effect is crucial for biologically relevant binding.^7,8 This suggests a potential application for galactose based glycopolymers.

There are generally two approaches in synthesizing glycopolymers, one of which is polymerizing a monomer that holds saccharide functionality. Various polymerisation techniques can be employed to generate glycopolymers from these monomers including living free radical techniques such as RAFT.⁹ The major drawback of this approach, however, is that it has been found that a number of these sugar monomers have the tendency to self-polymerise.¹⁰

The alternative technique involves post-functionalising a polymer with a saccharide.¹¹ Post-functionalisation allows for greater control and variability of the glycopolymer as different saccharides can be conjugated to the same polymeric architecture. The post-functionalisation route has the added advantage of not requiring exotic or new monomer species, which could in turn also affect the control in RAFT polymerisations resulting in long inhibition periods or non-living characteristics.

Fernández-García and coworkers have reported conjugating amino-saccharides to polymer containing activated esters forming amide linkages.^12,13 A technique that has been more widely used is the copper catalysed azide–alkyne Huisgen 1,3-dipolar cycloaddition (CuAAC) reaction. This is due to the flexibility and versatility of the reaction—possible under mild conditions in both aqueous and organic medium.¹⁴CuAAC conjugation techniques are widely used for various different functionalisations and have consistently reported high yields and orthogonality.¹⁵ Haddleton and co-workers have successfully reported a number of applications for the use of reactions to post-functionalise linear polymers with sugars and showed effective binding of the generated glycopolymers to lectins.^16–18

Similar to CuAAC are other ‘click’ chemistry techniques such as the thiol–alkene conjugation. This involves the reaction of a thiol and an alkene, by either a radical approach or alternatively viaMichael Addition.¹⁹ Chen et al. synthesized glucose based glycopolymer micelles via this approach in a copolymer with a thermoresponsive poly(diethylene glycol methyl ether methacrylate) (PDEGMA) block. The studies found that there was an enhancement of glycopolymer interaction with the Concanavalin A lectin when micelles were formed.²⁰ The increased binding to lectins upon micelle formation was also observed with mannose based thermoresponsive glycopolymers²¹ as well as galactose based micelles.^22,23

Binding efficacy is optimised when the distance between saccharide units is close to or the same as the spacing between binding sites in a lectin. As a result, polymer rigidity as well as the distance of the sugar moiety to the polymer backbone may have an effect on protein binding.²⁴ A more flexible polymer, or a polymer with pendant sugars linked with a spacer away from the backbone, would therefore be able to conform to the distance between binding sites. In the case of a rigid polymer, the only way to achieve optimum binding would be if the spaces between the sugars on the polymer matched the binding sites on the lectin exactly.²⁵

In this study, we synthesize a series of thermoresponsive block copolymers containing PDEGMA and either poly 2-hydroxy ethyl methacrylate (PHEMA) or poly(polyethylene glycol methacrylate) (PPEGMA). The PHEMA and PPEGMA blocks are functionalised with terminal alkynes and then glycosylated with an azide-galactose. The purpose of the use of two different monomers for this study is to investigate how the introduction of a flexible poly(ethylene glycol) PEG spacer affects binding to ricin. (Scheme 1)

Scheme 1 Overview of the different polymeric designs.

Experimental procedures

Materials

The RAFT agent cumyl dithiobenzoate (CDB),²⁶4-oxo-4-(prop-2-ynyloxy)butanoic anhydride²⁷ and azide-galactose²⁸ were prepared according to the procedure described in previous publications. Diethylene glycol methyl ether methacrylate (DEGMA), 2-hydroxy ethyl methacrylate (HEMA) and polyethylene glycol methacrylate of M_n = 360 (PEGMA₃₆₀) were all purchased from Aldrich and destabilized by passing them over a column of basic alumina. 2,2-Azobisisobutyronitrile (AIBN), (Fluka 98%) was recrystallised twice from methanol. All other chemicals were used as supplied by manufacturers unless otherwise stated.

Syntheses

(a) Synthesis of PDEGMA. DEGMA (6.0 g, 3.2 × 10⁻² mol), CDB (0.088 g, 3.2 × 10⁻⁴ mol) and AIBN (0.0053 g, 3.2 × 10⁻⁵ mol) were sealed in a Schlenk flask. The reaction mixture was subjected to 4 freeze–thaw degassing cycles. The flask was immersed in an oil bath at 65 °C for 6 h. This bulk polymerisation was terminated by cooling the mixture to 0 °C and exposing it to air. The resultant polymer was precipitated twice in diethyl ether and dried under vacuum to remove unreacted monomer and AIBN. The polymer was then analysed via¹H NMR and GPC to assess its size. The theoretical M_n obtained viaNMR is 7900 g mol⁻¹ while the M_n obtained viaGPC is 8200 g mol⁻¹ with a PDI of 1.25. ¹H NMR (300 MHz, CDCl₃, δ, ppm): 0.90 (6H, (CH₃)₂C-(C₆H₅)); 1.50–2.10 (5n₁H, CH₃CCH₂ (backbone)); 3.40 (3n₁H, CH₃OCH₂CH₂); 3.50–3.80 (6n₁H, CH₃OCH₂CH₂OCH₂CH₂); 4.10 (2n₁H, OCH₂CH₂COO); 7.20–7.50 (10H, aromatic peaks). n₁ is the degree of polymerisation of DEGMA.

(b) Synthesis of PDEGMA-b-PHEMA. PDEGMA obtained from part (a) with M_n = 8200 g mol⁻¹ was used as a macroRAFT agent for chain extension with HEMA. PDEGMA (0.30 g, 3.7 × 10⁻⁵ mol), AIBN (0.00060 g, 3.7 × 10⁻⁶ mol) and HEMA (0.476 g, 3.66 × 10⁻³ mol) were combined with N,N-dimethylacetamide (DMAc) to make up a total volume of 9 mL and separated into 3 reaction vessels. The vessels were sealed and then degassed by nitrogen sparging for 30 minutes at 0 °C. The reaction vessels were then placed in an oil bath at 60 °C and removed at 30, 60 and 90 minutes. The polymerisation was terminated by cooling the mixture to 0 °C and exposing it to air. The product was precipitated in diethyl ether and dried under vacuum. PHEMA homopolymers were also generated using the same RAFT, monomer and AIBN concentrations and experimental conditions. ¹H NMR (300 MHz, CDCl₃, δ, ppm): 0.90 (6H, (CH₃)₂C-(C₆H₅)); 1.50–2.10 (5n₁H + 5n₂H, CH₃CCH₂ (backbone)); 3.40 (3n₁H, CH₃OCH₂CH₂); 3.50–3.80 (6n₁H, CH₃OCH₂CH₂OCH₂CH₂, 2n₂H, CH₂CH₂OH); 4.10–4.30 (2n₁H + 2n₂H, OCH₂CH₂COO); 7.20–7.50 (10H, aromatic peaks). n₁ and n₂ are the degree of polymerisation of DEGMA and HEMA respectively.

(c) Synthesis of PDEGMA-co-PPEGMA. PDEGMA obtained from part (a) with M_n = 8200 g mol⁻¹ was used as a macroRAFT agent for chain extension with PEGMA₃₆₀. PDEGMA (0.2 g, 2.4 × 10⁻⁵ mol), AIBN (0.00040 g, 2.4 × 10⁻⁶ mol) and PEGMA₃₆₀ (4.39 g, 1.22 × 10⁻² mol) were combined with 6 mL of toluene and separated into 2 reaction vessels. The vessels were sealed and then degassed by nitrogen sparging for 30 minutes at 0 °C. The reaction vessels were then placed in an oil bath at 60 °C and removed at 90 and 120 minutes. The polymerisation was terminated by cooling the mixture to 0 °C and exposing it to air. The product was precipitated in diethyl ether and dried under vacuum. PPEGMA homopolymers were also generated using the same RAFT, monomer and AIBN concentrations and experimental conditions. ¹H NMR (300 MHz, CDCl₃, δ, ppm): 1.50–2.10 (5n₁H + 5n₃H, CH₃CCH₂ (backbone)); 3.40 (3n₁H, CH₃OCH₂CH₂); 3.50–3.80 (6n₁H, CH₃OCH₂CH₂OCH₂CH₂, 4n₃(m − 1), COOCH₂CH₂O(CH₂CH₂O)_m−1H); 4.10 (2n₁H + 2n₃H, OCH₂CH₂COO). n₁ and n₃ are the degree of polymerisation of DEGMA and PEGMA₃₆₀ respectively. m is the degree of polymerisation of ethylene glycol on PEGMA₃₆₀.

(d) Alkyne functionalisation of block copolymers. Polymers generated in parts (b) and (c) were analysed for the quantity of hydroxyl functions via both NMR and size exclusion chromatography. The polymers were combined with 4-oxo-4-(prop-2-ynyloxy)butanoic anhydride, triethyl amine (TEA), and 4-dimethylaminopyridine (DMAP) in a molar ratio of 3, 3 and 0.15 respectively to every hydroxyl function. The mixture was dissolved in a minimal amount of N,N-dimethyl formamide (DMF) so as to just dissolve the solids, and left to react at 50 °C over 18 hours. The product was then precipitated twice in diethyl ether so as to remove all traces of free alkyne. For PDEGMA-b-PHEMA: ¹H NMR (300 MHz, DMSO-d₆, δ, ppm): 1.50–2.10 (5n₁H + 5n₂H, CH₃CCH₂ (backbone)); 2.70 (4n₂H, COOCH₂CH₂COOCH₂CCH); 3.40 (3n₁H, CH₃OCH₂CH₂); 3.50–3.80 (6n₁H, CH₃OCH₂CH₂OCH₂CH₂, 2n₂H, CCOOCH₂CH₂OOC); 4.10–4.30 (2n₁H + 2n₂H, OCH₂CH₂COO); 4.70 (4n₂, COOCH₂CCH). n₁ and n₂ are the degree of polymerisation of DEGMA and HEMA respectively. For PDEGMA-b-PPEGMA: ¹H NMR (300 MHz, DMSO-d₆, δ, ppm): 1.50–2.10 (5n₁H + 5n₃H, CH₃CCH₂ (backbone)); 2.70 (4n₃H, COOCH₂CH₂COOCH₂CCH); 3.40 (3n₁H, CH₃OCH₂CH₂); 3.50–3.80 (6n₁H, CH₃OCH₂CH₂OCH₂CH₂, 4n₃(m − 1), COOCH₂CH₂O(CH₂CH₂O)_m−1OC); 4.10 (2n₁H + 2n₃H, OCH₂CH₂COO); 4.70 (4n₃, COOCH₂CCH). n₁ and n₃ are the degree of polymerisation of DEGMA and PEGMA₃₆₀ respectively. m is the degree of polymerisation of ethylene glycol on PEGMA₃₆₀.

(e) Glycosylation of alkyne functionalised block copolymersvia click chemistry. The alkyne functionalised polymers from part (d) were combined with a 1.2 molar equivalent of azide functionalised galactose to alkyne functions, equimolar ratio of CuSO₄ to alkyne functions, and 5 molar equivalents of sodium ascorbate to CuSO₄. This mixture was dissolved in DMF with 10% of distilled water and placed in an oil bath at 45 °C for 3 days. The product was then dialysed against water with 4 water changes over 24 h to remove impurities and excess galactose, copper sulfate and sodium ascorbate. The dialysis product was then reacted with pentaerythritol tetrakis(3-mercaptopropionate) to remove any remnant copper that could be ligated to the resultant triazole ring, filtered and then freeze-dried. For PDEGMA-b-PHEMA: ¹H NMR (300 MHz, D, δ, ppm): 1.50–2.10 (5n₁H + 5n₂H, CH₃CCH₂ (backbone)); 2.70 (4n₂H, COOCH₂CH₂COOCH₂CCH); 3.30 (3n₁H, CH₃OCH₂CH₂); 3.40–4.30 (6n₁H, CH₃OCH₂CH₂OCH₂CH₂; 2n₂H, CCOOCH₂CH₂OOC; 2n₁H + 2n₂H, OCH₂CH₂COO; 11n₂, galactose); 5.20 (4n₂, COOCH₂C=CHN₃); 5.60 (n₂H, d, anomeric glucose proton); 8.20 (1n₂, COOCH₂C=CHN₃). n₁ and n₂ are the degree of polymerisation of DEGMA and HEMA respectively. For PDEGMA-b-PPEGMA: ¹H NMR (300 MHz, DMSO-d₆, δ, ppm): 1.50–2.10 (5n₁H + 5n₃H, CH₃CCH₂ (backbone)); 2.70 (4n₃H, COOCH₂CH₂COOCH₂CCH); 3.30 (3n₁H, CH₃OCH₂CH₂); 3.40–4.30 (6n₁H, CH₃OCH₂CH₂OCH₂CH₂; 4n₃(m − 1), COOCH₂CH₂O(CH₂CH₂O)_m−1OC; 2n₁H + 2n₃H, OCH₂CH₂COO; 11n₂, galactose); 5.20 (4n₂, COOCH₂C=CHN₃); 5.60 (n₂H, d, anomeric glucose proton); 8.20 (1n₂, COOCH₂C=CHN₃). n₁ and n₃ are the degree of polymerisation of DEGMA and PEGMA₃₆₀ respectively. m is the degree of polymerisation of ethylene glycol on PEGMA₃₆₀.

Analysis

Nuclear Magnetic Resonance (NMR) spectroscopy. ¹H and ¹³C NMR spectra were recorded on a Bruker ACF300 (300 MHz) spectrometer, with CDCl₃, D₂O or d-DMSO used as solvents.

Size exclusion chromatography (SEC). Molecular weight distributions of the block copolymers were determined by size-exclusion chromatography (SEC) using a Shimadzu modular system, comprising of an auto-injector, a Phenomenex Phenogel 5.0 μm bead-size guard column (50 × 7.5 mm), four linear Phenomenex columns (10⁵, 10⁴, 10³ and 500 Å), and a differential refractive index detector. The eluents used were either N,N-dimethylacetamide (DMAc) (0.05% w/v LiBr, 0.05% w/v BHT) at 50 °C or tetrahydrofuran (THF) at 40 °C with a flow rate of 1 mL min⁻¹. The system was calibrated using narrow polystyrene standards ranging from 162 to 2 × 10⁶ g mol⁻¹. Mark–Houwink parameters for polyethylene glycol (K = 11.0; α = 0.725) were used for samples analysed in THF.

Dynamic light scattering (DLS). Dynamic light scattering studies of the polymers at 1 mg mL⁻¹ in an aqueous medium were conducted using a Malvern Instruments Zetasizer Nano ZS instrument equipped with a 4 mV He–Ne laser operating at λ = 633 nm, an avalanche photodiode detector with high quantum efficiency, and an ALV/LSE-5003 multiple tau digital correlator electronics system.

Transmission electron microscopy (TEM). The sizes and morphologies of the polymers were observed using a transmission electron microscope JEOL1400 TEM at an accelerating voltage of 100 kV. The polymers were dissolved in water (1 mg mL⁻¹) and deposited onto 200 mesh, holey film, copper grid (ProSciTech). For the preparation of samples above the LCST, the solution was heated to 40 °C, and at the same time, a copper grid was also equilibrated at this temperature. A drop of the heated polymer solution was then placed on the copper grid and dried in the oven at 40 °C.

Ricin binding assays. The procedure as described by Dawson et al.²⁹ for an asialofetuin competition assay was followed. Nunc-Immuno Maxisorp 96-well 0.35 mL microtitration plates were used. Polymers were incubated with ricin for 1 h before transfer to the plate. Absorbance at 405 nm was measured on a Bio-tek Synergy HT plate reader every minute for 15 min. A ricin standard curve was run in triplicate on each plate to enable calculation of residual ricin values. Galactose inhibition of ricin was tested (in duplicate) on every plate as a control to enable direct comparison with the polymers. Each polymer was tested in duplicate.

Inhibition of ricin was also tested at 40 °C. For these assays the plate was incubated with blocking buffer in a Ratek orbital mixer incubator at 40 °C with gentle shaking for 1 h. The ricin and polymer/galactose mixtures were also incubated at 40 °C with gentle shaking for 1 h before being transferred to the plate at 40 °C and subsequently incubated for 1 h on the plate. The rest of the assay was performed at room temperature.

Results and discussion

The synthesis of different PDEGMA-b-PHEMA and PDEGMA-b-PPEGMA was achieved by RAFT polymerization using cumyl dithiobenzoate (CDB) as the RAFT agent and AIBN as a thermal initiator (Scheme 2). PDEGMA was generated first, but since both blocks are based on methacrylates the reverse approach would be possible. Synthesis of PDEGMA macroRAFT agent was carried out in bulk so as to keep the reaction time as low as possible and avoid unnecessary side reactions, thus maximising the percentage of RAFT capped chains. The M_n of the carefully purified polymer was obtained viaSEC and ¹H NMR and was found to be 8200 g mol⁻¹ and 7900 g mol⁻¹ respectively.

Scheme 2 Outline of synthesis route of block copolymers. (a) The synthesis of PDEGMA macroRAFT agent; (b) the synthesis of PDEGMA-b-PHEMA and (c) of PDEGMA-b-PPEGMA.

In a subsequent step, PDEGMA was used as the macroRAFT and was then chain extended with either HEMA or PEGMA₃₆₀ using now DMAc as solvent to reduce viscosity. The same-sized PDEGMA was used for all copolymerizations to generate a constant block size of the stimuli-responsive block while the glycopolymer block was varied in chain length. Samples were then drawn at different time intervals and the block copolymers were analyzed viaGPC as well as ¹H NMR.

Previous studies³⁰ involving the polymerization of PEGMA₃₆₀ recommended to keep the conversion of the reaction below 20% as there is evidence of gelation. It is suggested that the monomer as supplied by the manufacturer contains small amounts of poly(ethylene glycol) dimethacrylate, which could polymerise and cross-link the propagating chains.

The SEC traces of the copolymer show a narrow distribution with no shoulders at higher molecular weights or lower molecular weights tailings and a low PDI. Table 1 summarises the different homo and block (co)polymers prepared with their respective PDIs (Table 1, column: M_n (PDI) unfunct).

Table 1 Summary of all polymers made with their M_n and PDI reported from GPC measurements including a theoretical M_n (calculated from the degree of polymerisation found viaNMR) reported for the unfunctionalised (co)polymers

No	DEGMA units	HEMA units	PEGMA360 units	M n ^Theo unfunct	M n (PDI) unfunct	M n (PDI) alkyne	M n (PDI) glyco
a Commercial polymer purchased from Sigma-Aldrich. Mv ≈ 20 000.
0	43	—	—	7900	8200 (1.25)	—	—
1	—	—	24	8600	9400 (1.10)	11 300 (1.23)	34 800 (1.52)
2	—	90	—	12 000	11 800 (1.18)	24 600 (1.22)	35 900(1.23)
3^a	—	154	—	20 000	17 200 (1.80)	34 000 (1.88)	45 000 (1.84)
4	43	—	20	15 000	15 700 (1.11)	16 900 (1.16)	58 400 (1.82)
5	43		16	13 800	14 200 (1.17)	16 000 (1.23)	53 400 (1.75)
6	43	20	—	10 400	11 000 (1.13)	12 500 (1.15)	13 700 (1.14)
7	43	79	—	18 000	18 400 (1.16)	23 600 (1.21)	26 900 (1.28)
8	43	68	—	16 500	17 000 (1.16)	21 800 (1.20)	24 100 (1.35)

---

PEGMA₃₆₀ and HEMA homopolymers were also generated with CDB as the RAFT agent. The purpose of this is to serve as a basis of comparison between non-thermoresponsive homopolymers and the thermoresponsive DEGMA containing copolymers.

The terminal hydroxyl functionalities of the (co)polymers were then converted to alkynes with 4-oxo-4-(prop-2-ynyloxy)butanoic anhydride in DMF in the presence of DMAP and TEA. The product of this functionalisation was precipitated in cold ether 3 times to ensure that all reactants are removed from the polymer. Full alkyne functionalisation was verified via¹H NMR in DMSO-d₆ by the appearance of peaks at 2.60 and 4.69 ppm, with integrals that verify the correct ratios of the associated protons. The NMR also serves to identify any excess 4-oxo-4-(prop-2-ynyloxy)butanoic anhydride that still might be present (Fig. 1).

Fig. 1 ¹H-NMR (from bottom to top): (I) pre-functionalised block copolymer (d-DMSO); (II) alkyne functionalisation of terminal hydroxyls on the PPDEGMA block (d-DMSO); (III) after ‘Clicking’ of azide galactose on polymer (D₂O).

Alkyne functionalized polymer was then reacted with azide functionalized galactosevia a copper catalysed azide–alkyne cycloaddition using copper(II) sulfate and sodium ascorbate. The reaction was left to run over 72 hours to ensure that all alkynes were reacted with the azide containing galactose molecules since their close proximity to one another could be sterically hindered. Scheme 3 depicts the various stages of functionalisation of the (co)polymers. The product of the ‘click’ reaction was dialysed against water for 2 days with water changes every 6 hours to remove excess galactose, copper sulfate, sodium ascorbate and solvent. However, traces of the catalyst were often still present and had to be removed using pentaerythritol tetrakis(3-mercaptopropionate) as a strong ligand for copper ions. Copper has been known to ligate to the triazole rings formed in azide–alkyne coupling reactions.^15,31 To ensure that most of the copper was removed from the gly(co)polymer product, pentaerythritol tetrakis(3-mercaptopropionate) was added dropwise to an aqueous solution of the product. The mercaptan forms a white precipitate as it preferentially ligates to the copper. The mixture was filtered, the excess mercaptan separated, and then freeze-dried before any further analyses were carried out. The product was analysed via¹H NMR. The disappearance of the peak at 2.5 ppm corresponding to the CH from the alkyne together with the appearance of the peak at 8.28 ppm due to the presence of a triazole ring suggests that all alkyne had reacted as illustrated in Fig. 1. In addition the methylene group adjacent to the ester functionality (E in Fig. 1) shifts from 4.7 ppm to 5.2 ppm. The signals 8.28, 5.60 and 5.20 ppm are with an intensity ratio of 1 : 1 : 2 in good agreement with the expected value. The M_ns of the various glycopolymers generated at each stage of functionality are also reported in Table 1.

Scheme 3 Postfunctionalization of the terminal hydroxyl moieties on the polymers. For the PHEMA based (co)polymers, m′ = 1.

Fig. 2 and 3 show the evolution of SEC traces at each step of polymerisation and functionalisation for PDEGMA-b-PPEGMA and PDEGMA-b-PHEMA copolymers respectively.

Fig. 2 SEC traces for the stages of functionalisation for polymer 4.

Fig. 3 SEC traces for the stages of functionalisation for polymer 7.

There is a noticeable high molecular weight hump in the SEC of all the glycosylated polymers. Lau et al.³² report artifacts similar to this where they found that triazole rings can have a synergistic effect on the interchain hydrogen bonding to the point of gelation. A phenomenon such as this one reasonably explains the high molecular weight humps seen in the SECs, especially due to the close proximity of the galactose moieties to the triazole rings on the polymers. To exclude the possibility of the formation of disulfide formation, which could be formed from the RAFT endgroup after hydrolysis, the SEC experiments were carried out in the presence of a small amount of DTT to reduce the disulfide to thiol. Changes in the intensity of the high molecular weight shoulder were however negligible. Another reason could be the occurrence of some transesterification reactions taking place under the conditions of the click reaction.

The glycopolymers (Table 3) were tested for the thermoresponsive characteristics by measuring their average size at various temperatures by dynamic light scattering (DLS). The polymers were dissolved in de-ionised water at a concentration of 1 mg mL⁻¹. These polymers were left in water for 96 hours at room temperature to ensure complete dissolution since dissolution could be hindered by inter- and intra-molecular hydrogen bonding. The polymer solution was filtered using a 0.45 μm filter and placed in a 1 cm disposable plastic cuvette for DLS analysis.

It is expected that the polymers dissolve in water unimolecularly at 20 °C, while above the LCST of PDEGMA, micelle formation occurs resulting in aggregates with a PDEGMA core and aglycopolymer shell. The number-average and volume-average hydrodynamic diameter at 20 °C and 40 °C are displayed in Table 2. The strong discrepancy between both values at 20 °C is synonymous with a broad and multi-modal size distribution indicating the formation of unexpected aggregates. The particle size recorded for all polymers at a temperature of 20 °C is much larger than the average size of linear polymers in solution, which are often merely a few nanometres. Strong aggregation caused by the presence of carbohydrates, PEG chains and triazole (see above) may hamper the full dissolution of the theoretically fully water-soluble polymer. Aggregation of polysaccharides and glycopolymers has been observed earlier and investigated in detail.^33,34 Once at 40 °C, above the LCST of PDEGMA of 29 °C, all the measured distributions have similar mean-values indicative for a better defined product with a smaller particle size distribution. A notable increase in the recorded number-mean particle diameter and count rate confirms that these polymers do indeed hold some thermoresponsive characteristic. The drop in mean volume and intensity diameters suggests that at higher temperatures, the hydrogen bonds break and allow the polymers to self assemble into micelles.

Table 2 Summary of the 5 thermoresponsive copolymers with their composition and their number mean and volume mean particle diameter, as well as DLS count-rates at 20 °C and 40 °C

No	Polymer	Number mean/nm		Volume mean/nm		Count rate
		20 °C	40 °C	20 °C	40 °C	20 °C	40 °C
4	PDEGMA43-b-P(PEGMA-Gal)20	50	136	357	174	180	373
5	PDEGMA43-b-P(PEGMA-Gal)16	92	145	526	198	175	338
6	PDEGMA43-b-P(HEMA-Gal)20	84	114	558	116	98	252
7	PDEGMA43-b-P(HEMA-Gal)79	60	119	199	140	110	261
8	PDEGMA43-b-P(HEMA-Gal)68	81	158	414	198	161	418

---

Images of the polymer samples were also obtained through TEM. The samples were prepared at a room temperature of 20 °C as well as at a temperature of 40 °C and stained with phosphor tungstic acid. The TEM images at 20 °C do not have any characteristically significant features, however the images at 40 °C show particles which are not present in the samples prepared at 20 °C (Fig. 4). The shape of the particles obtained in the TEM samples prepared at 40 °C appears to be round which further suggests that there is a formation of micelles although the formation of large compound micelles could be possible considering the measured large sizes. The particle size through TEM appears to be comparable with the sizes reported through DLS (Table 3).

Fig. 4 (Clockwise from top left) Polymer 7 prepared on the copper grid at 20 °C (a); polymer 8 prepared on the copper grid at 20 °C (b); polymer 8 prepared on the copper grid at 40 °C (c); polymer 7 prepared on the copper grid at 40 °C (d). All samples negatively stained with phosphor tungstic acid.

Table 3 Comparison of particle sizes at 40 °C including the measured diameter from TEM images, and the number average and volume average diameters as reported by the DLS

No	Polymer	TEM/nm	Num. avg. diam./nm	Vol. avg. diam./nm
4	PDEGMA43-b-P(PEGMA-Gal)20	130	136	174
5	PDEGMA43-b-P(PEGMA-Gal)16	200	145	198
6	PDEGMA43-b-P(HEMA-Gal)20	130	114	116
7	PDEGMA43-b-P(HEMA-Gal)79	150	119	140
8	PDEGMA43-b-P(HEMA-Gal)68	200	158	198

---

Ricin binding assays

Galactose is a highly bioactive carbohydrate, which plays a major role in many recognition events including specific binding to plant and animal lectins including ricin.⁹Ricin, a highly toxic lectin isolated from castor bean (Ricinus communis), is a hetero lectin consisting of two parts, ricin A and ricin B chains. The B chain has two or three galactose binding sites^35,36 and therefore a multivalent galactose system may provide a means of neutralisation. The focus of interest in this study is how the polymers prepared may increase binding to ricin compared to free galactose thus displaying a multivalency effect. The polymers were tested by asialofetuin competition assay in comparison to galactose at room temperature and at 40 °C. The concentrations were standardised as galactose equivalents. Only the IC50 value—the concentration of galactose that inhibits 50% of ricin—is listed in Table 4. It should be noted here that the concentrations are only approximate since solubility issues in the media led to uncertainties.

Table 4 Ricin inhibition assay IC50 ratios

No	Sample/polymer	Approximate IC50 ratios
	Galactose	0.75	1
1	P(PEGMA-Gal)24	—	>0.4
2	P(HEMA-Gal)90	0.2	1.2
3	P(HEMA-Gal)154	—	>0.75 (likely actual ratio of 1, tracks along with galactose)
4	PDEGMA43-b-P(PEGMA-Gal)20	—	>0.175
5	PDEGMA43-b-P(PEGMA-Gal)16	<0.5, >0.19	<0.3, >0.125
6	PDEGMA43-b-P(HEMA-Gal)20	—	>0.125
7	PDEGMA43-b-P(HEMA-Gal)79	0.2	1
8	PDEGMA43-b-P(HEMA-Gal)68	—	>0.3

---

All polymers were tested at 40 °C and three polymer samples were tested additionally at 20 °C and their performance were compared to galactose. Increase in temperature meant a decrease of IC50 for all the four samples tested (Table 4, middle column). While galactose alone acted only slightly better with increasing temperature, the polymers led to noticeable better binding at higher temperature. This includes the tested homo glycopolymer (2) for which thermoresponsive self-assembly is not possible. It is not known how the buffer system used in these assays may affect the self-assembly of the polymers. An explanation for the increase in binding for the three polymers can be found in the changes in aggregate formation with temperature. It is possible that at room temperature the polymer is too closely associated with itself to participate fully in ricin binding. The DLS studies, which indicate strong aggregation and also the difficulties in solubilising these polymers may be an indication of tight inter- and intra-molecular binding. However on increase of temperature these associations would lessen, potentially freeing the galactose moieties for ricin binding. Block copolymers in addition will form more ordered structures at higher temperature with a high surface concentration of galactose. The increased binding here is in good agreement with earlier studies using thermo-responsive block copolymer, which show that upon micelle formation binding is more efficient.^21,23 However, the absolute inhibition of ricin was only conclusively better for sample 5 (PDEGMA₄₃-b-P(PEGMA-Gal)₁₆) compared to monovalent Galactose. For polymers 2 and 7 the increase at higher temperature only brings them into line with what is seen for monovalent galactose, whereas polymer 5 is better than galactose at room (IC50 is half that of galactose) and elevated temperature.

Comparison of the performance of all the polymers with monovalent galactose at 40 °C shows that P(HEMA-Gal) has the weakest binding barely showing any improvement to galactose. Introduction of the flexible PEG spacer is indeed beneficial and the amount necessary to inhibit 50% of ricin is reduced by more than half. This would not be entirely unexpected due to the added flexibility this polymer has in its side-chains with the PEG spacer. Mammen and coworkers²⁴ discuss the importance of matching the distance between two saccharides on a polymer to binding sites on a lectin to obtain good binding. It is evident from these data that having a PEG spacer in a polymer could indeed give the polymer the flexibility it requires to achieve good protein binding. Even better inhibition was observed by the micellar system (with exception of sample 7, whose bad performance cannot currently be explained). Upon micelle formation, the glycopolymers take on a more brush-like structure; the shorter the hydrophilic block, the more stretched are the glycopolymers, and thus the more brush-like they become. Comparing the glycopolymer block length with the values displayed in Table 4, last column, there seems to be a trend between the potential correlation between block length and performance with shorter block (therefore more brush-like structures) exhibiting better inhibition. In addition, a slight improvement may also be observed when a PEG spacer is employed, although this may be within errors.

These initial experiments establish an order of ricin inhibition of P(HEMA-Gal) < P(PEGMA-Gal) < PDEGMA₄₃-b-P(HEMA-Gal) ≤ PDEGMA₄₃-b-P(PEGMA-Gal) with better binding observed in shorter glycopolymer blocks.

The IC50 values reported here show only a modest enhancement in binding between galactose and galactose polymers. The IC50 is proportional to the inhibitor dissociation constant by the following equation: K_i = IC₅₀/(1 + R/K_r) where R is ricin conc. and K_r is the ricin–asialofetuin dissociation constant.³⁷ The ratio of IC50 (multivalent polymer)/IC50(galactose) is equivalent to K_i(polymer)/K_i(galactose) and is a measure of enhancement (as a reduction in dissociation constant). However Lundquist and Toone⁴ caution against making such interpretations of these types of assays and relying on Scatchard plots (where this equation is derived from) is problematic for covalently tethered ligands. More than just the kinetics affects the IC50 determined in an assay.³⁸ This type of assay is useful as a guide only and to quantify any kinetic or thermodynamic enhancements due to strict multivalency effects techniques such as SPR and ITC should be used. However kinetic data may not account for all mechanisms of inhibition in real biological systems such as clustering or steric hinderance.

Conclusion

Block copolymers of DEGMA and HEMA as well as DEGMA and PEGMA were synthesized by RAFT polymerisation. The terminal hydroxyl moieties on the HEMA and PEGMA blocks were alkyne functionalised with an alkyne anhydride and then later conjugated to azide functionalised galactoseviaCuAAC in mild conditions. DLS studies show that raising a solution of the glycopolymers to 40 °C has an effect on its particle size. Below this temperature, the particle size is generally found to be larger, however this phenomenon can be attributed to inter-molecular hydrogen bonding. TEM images taken of the polymers prove that micelles form above the LCST, while those samples prepared below the LCST had no significant characteristics. This confirms that the glycopolymers do exhibit thermoresponsive characteristics and form micelles at temperatures above 40 °C.

The purpose of these glycopolymers was to investigate the effect of the distance of galactose from the polymer backbone on lectin/protein and glycopolymer interaction. Studies were carried out to assess the binding efficiency of the glycopolymers to the plant toxin, ricin. PDEGMA-b-PPEGMA glycopolymers showed the highest degree of binding with the polymer with the shorter glycopolymeric chain having a better inhibition of ricin as compared to monovalent galactose. We have established an order of ricin inhibition of P(HEMA-Gal) < P(PEGMA-Gal) < PDEGMA₄₃-b-P(HEMA-Gal) ≤ PDEGMA₄₃-b-P(PEGMA-Gal) with better binding observed in shorter glycopolymer blocks and P(HEMA-Gal) not showing any significant improvement in ricin inhibition over monovalent galactose.

Acknowledgements

The authors would like to thank the Australian Research Council (ARC) for financial support. The authors would like to acknowledge the Mark Wainwright analytical centre at UNSW for help in NMR analysis and microscopy analysis.

Notes and references

1. Y. Miura, J. Polym. Sci., Part A: Polym. Chem., 2007, 45, 5031–5036.
2. J. M. Lord, L. M. Roberts and J. D. Robertus, FASEB J., 1994, 8, 201–208.
3. S. G. Spain and N. R. Cameron, Polym. Chem., 2011, 2, 60–68.
4. J. J. Lundquist and E. J. Toone, Chem. Rev. (Washington, DC, U. S.), 2002, 102, 555–578.
5. J. Audi, M. Belson, M. Patel, J. Schier and J. Osterloh, JAMA, J. Am. Med. Assoc., 2005, 294, 2342–2351.
6. S. Lundberg, L. Melin, C. Nilsson and P. von Schoenberg, Ricin. Threat, effects and protection, User Report FOI-R–1261–SE, FOI-Swedish Defence Research Agency, Umea, 2004.
7. L. L. Kiessling and N. L. Pohl, Chem. Biol., 1996, 3, 71–77.
8. V. Ladmiral, E. Melia and D. M. Haddleton, Eur. Polym. J., 2004, 40, 431–449.
9. S. R. S. Ting, G. Chen and M. H. Stenzel, Polym. Chem., 2010, 1, 1392–1412.
10. S. G. Spain, M. I. Gibson and N. R. Cameron, J. Polym. Sci., Part A: Polym. Chem., 2007, 45, 2059–2072.
11. M. A. Gauthier, M. I. Gibson and H.-A. Klok, Angew. Chem., Int. Ed., 2009, 48, 48–58.
12. G. Martínez, M. Fernández-García and M. Sánchez-Chaves, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 18–27.
13. M. L. Cerrada, M. Sánchez-Chaves, C. Ruiz and M. Fernández-Garcia, Biomacromolecules, 2009, 10, 1828–1837.
14. R. A. Evans, Aust. J. Chem., 2007, 60, 384–395.
15. W. Binder and R. Sachsenhofer, Macromol. Rapid Commun., 2007, 28, 15–54.
16. V. Ladmiral, G. Mantovani, G. J. Clarkson, S. Cauet, J. L. Irwin and D. M. Haddleton, J. Am. Chem. Soc., 2006, 128, 4823–4830.
17. J. Geng, G. Mantovani, L. Tao, J. Nicolas, G. Chen, R. Wallis, D. A. Mitchell, B. R. G. Johnson, S. D. Evans and D. M. Haddleton, J. Am. Chem. Soc., 2007, 129, 15156–15163.
18. L. Nurmi, J. Lindqvist, R. Randev, J. Syrett and D. M. Haddleton, Chem. Commun., 2009, 2727–2729.
19. A. B. Lowe, Polym. Chem., 2010, 1, 17–36.
20. G. Chen, S. Amajjahe and M. H. Stenzel, Chem. Commun., 2009, 1198–1200.
21. M. Hetzer, G. Chen, C. Barner-Kowollik and M. H. Stenzel, Macromol. Biosci., 2010, 10, 119–126.
22. S. R. S. Ting, E. H. Min, P. Escalé, M. Save, L. Billon and M. H. Stenzel, Macromolecules, 2009, 42, 9422–9434.
23. S. R. S. Ting, E. H. Min, P. B. Zetterlund and M. H. Stenzel, Macromolecules, 2010, 43, 5211–5221.
24. M. Mammen, S. K. Choi and G. M. Whitesides, Angew. Chem., Int. Ed., 1998, 37, 2754–2794.
25. T. Hasegawa, S. Kondoh, K. Matsuura and K. Kobayashi, Macromolecules, 1999, 32, 6595–6603.
26. L. Albertin, M. Stenzel, C. Barner-Kowollik, L. J. R. Foster and T. P. Davis, Macromolecules, 2004, 37, 7530–7537.
27. P. Antoni, Y. Hed, A. Nordberg, D. Nyström, H. von Holst, A. Hult and M. Malkoch, Angew. Chem., Int. Ed., 2009, 48, 2126–2130.
28. J. Geng, J. Lindqvist, G. Mantovani, G. Chen, C. Sayers, G. Clarkson and D. Haddleton, QSAR Comb. Sci., 2007, 26, 1220–1228.
29. R. M. Dawson, M. R. Alderton, D. Wells and P. G. Hartley, J. Appl. Toxicol., 2006, 26, 247–252.
30. L. Zhang, T. Nguyen, J. Bernard, T. Davis, C. Barner-Kowollik and M. Stenzel, Biomacromolecules, 2007, 8, 2890–2901.
31. T. R. Chan, R. Hilgraf, K. B. Sharpless and V. V. Fokin, Org. Lett., 2004, 6, 2853–2855.
32. K. Lau, H. Chow, M. Chan and K. Wong, Angew. Chem., Int. Ed., 2008, 47, 6912–6916.
33. D. Balasubramanian, B. Raman and C. S. Sundari, J. Am. Chem. Soc., 1993, 115, 74–77.
34. L.-C. You, F.-Z. Lu, Z.-C. Li, W. Zhang and F.-M. Li, Macromolecules, 2003, 36, 1–4.
35. Y. P. Venkatesh and J. M. Lambert, Glycobiology, 1997, 7, 329–335.
36. E. Rutenber and J. D. Robertus, Proteins: Struct., Funct., Genet., 1991, 10, 260–269.
37. R. M. Dawson, B. M. Paddle and M. R. Alderton, J. Appl. Toxicol., 1999, 19, 307–312.
38. S. M. Dimick, S. C. Powell, S. A. McMahon, D. N. Moothoo, J. H. Naismith and E. J. Toone, J. Am. Chem. Soc., 1999, 121, 10286–10296.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c1py00048a

---