Hyperbranched Glycopolymers of 2-(α-d-Mannopyranose) Ethyl Methacrylate and N,N’-Methylenebisacrylamide: Synthesis, Characterization and Multivalent Recognitions with Concanavalin A

A series of novel hyperbranched poly[2-(α-d-mannopyranosyloxy) ethyl methacrylate-co-N,N’-methylenebisacrylamide] (HPManEMA-co-MBA) are synthesized via a reversible addition fragmentation polymerization (RAFT). The dosage ratios of linear and branch units are tuned to obtain different degree of branching (DB) in hyperbranched glycopolymers. The DB values are calculated according to the content of nitrogen, which are facilely determined by elemental analysis. The lectin-binding properties of HPManEMA-co-MBA to concanavalin A (ConA) are examined using a turbidimetric assay. The influence of defined DB value and molecular weight of HPManEMA-co-MBA on the clustering rate is studied. Notably, HPManEMA-co-MBAs display a low cytotoxicity in the MTT assay, thus are potential candidates for biomedical applications.


Introduction
Carbohydrates and their corresponding conjugated glycoproteins are key participants in a measureless array of biological functions, and can also be used as the carrier or drug in the treatment of diseases, such as cancer, cytotoxic chemotherapy or radiotherapy [1][2][3][4][5][6][7][8]. However, the monovalent interaction between the monosaccharide units and proteins is relatively weak due to their low affinity [9][10][11][12][13][14]. Compared with monosaccharide, polysaccharides being of high molecular weight show strong affinity due to the "cluster glycoside effect" [15][16][17][18]. Apart from the effect of molecular weight, different architectures and topological expressions also have important influence on recognition event. For example, multi-block glycopolymer with a degree of monomer sequence control is a potent competitive inhibitor of the HIV envelope glycoprotein gp120 by interacting with human dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN) [19,20].
With the combination of living radical polymerization and click chemistry, Perrier and coworkers found a new strategy to synthesize hyperbranched glycopolymers with highly degree of branching (DB) [29]. It should be noted that the DB value played an important role in the recognition event. Technologies, such as nuclear magnetic resonance (NMR), viscosity measurements and light-scattering methods, are often used to evaluate the DB value of hyperbranched polymers [30][31][32][33]. However, in some specific conditions, these technologies are difficult to use for determining the DB value of hyperbranched glycopolymers. For instance, when NMR technology is used, the signals of proton belonging to the linear and branch unit might overlap, making it difficult to distinguish them. Aware of the importance of "cluster glycoside effect", it would be highly desired to prepare hyperbranched glycopolymers whose DB values can be determined using a facile method.
The DB values of the hyperbranched glycopolymers were facilely determined via elemental analysis because the nitrogen atoms are merely contained in branch units. Moreover, the obtained hyperbranched glycopolymers bearing defined DB and molecular weight were used to investigate the binding behavior of Con A. Furthermore, the resulting hyperbranched glycopolymers were studied by MTT assay to test their cytotoxicity.
Number-average molecular weight (M n ), weight-average molecular weight (M w ), polydispersity, and [η] were obtained by 270-doul detector-Size Exclusion Chromatography (MDSEC) (Malvern Instruments Ltd., Malvern, PA, USA), equipped with RI detector, two-angle light scattering detector and viscosity detector. The columns (15 cm viscotek G2500PW×l and 15 cm Waters Wat011545) were eluted with 0.1 M NaNO 3 solution and calibrated with Polyethylene oxide std-PEO22K. The calibration was performed at 25 • C and a flow rate of 1 mL·min −1 . The product dissolved in 0.1 M NaNO 3 solution, and passed through 0.2 µm filter before injection. Recycling preparative gel permeation chromatography (GPC) purifications were performed on a Shimadzu HPLC system equipped with a model SPD-20A absorbance detector, a model RID-10A differential refractometer, an in-line degasser, a model LC-6AD pump, a model CBM-20A controller and a Shodex KF-802 preparative GPC column. The samples were cycled over the column three times before separation (eluent, THF; flow rate, 3.0 mL·min −1 ) THF. The binding behavior of HPManEMA-co-MBA with Con A was examined by measuring the turbidity at 420 nm [36], which were performed on a Shimadzu WV-2550 spectrophotometer (Shimadzu, Kyoto, Japan). Con A was diluted to 1 mM (based on Con A tetramer) with 10 mM HEPES-buffered (pH = 7.4) saline containing 150 mM NaCl and 1 mM CaCl 2 (hereinafter referred to as HBS). The Con A solution (2 mL) was transferred to a 1 cm cuvette. After incubation at 25 • C, 200 µL of HBS solution of HPManEMA-co-MBA was injected to the cuvette, in which the final concentration of ManEMA unit was 500 µM. HeLa cells were obtained from Shanghai Institutes for Biological Sciences of the Chinese Academy of Sciences (CAS). The Cell viability was used to evaluate the cytotoxicity of HPManEMA-co-MBA. In brief, HeLa cells (5 × 10 3 cells) in 100 µL of DMEM containing 10% FCS were plated in a 96-well plate and incubated for 24 h in a humidified atmosphere of 5% CO 2 in air at 37 • C (Sanyo, Model MCO-18AIC, Tokyo, Japan). After 100 µL of a HPManEMA-co-MBA in DMEM containing 10% FCS and 2% DMSO was added, the Hela cells were incubated, and the final DMSO content was 1% in all cases. Cells incubated without HPManEMA-co-MBA were used as an untreated control group. After 24 h incubation, MTT dye solution (20 µL, 5 mg·mL −1 ) was added to each well, and the cells were incubated for another 4 h. The cells were analyzed using a microplate spectrophotometer (BioRad Model 3550, Richmond, CA, USA) at 570 nm, and the cell survival rate was calculated by normalization with respect to the value of control group.

Synthesis of HPManEMA-co-MBA
Hyperbranched glycopolymers HPAcManEMA-co-MBA was synthesized via RAFT polymerization. In brief, AcManEMA (1 mmol, 458 mg), a certain amount of MBA as cross-linker, and CTA (0.02 mmol, 2.77 mg), were dissolved in 4 mL of DMF, and the solution was ultrasonicated until completely dissolved. Then, oxygen was removed by repeated vacuum-nitrogen cycles (three times). After the AIBN (0.01 mmol, 1.64 mg) initiator was added into the system, the polymerization was conducted at 70 • C in an oil bath for 25 h. Afterwards, the product was precipitated by dropping the solution into a large excess of diethyl ether. The precipitated polymer was separated by centrifugation, and redissolved in acetone. The solution was precipitated again. Finally, the product was dried in vacuum to get a white solid.
HPManEMA-co-MBA was obtained by the deprotection of acetyl in HPAcManEMA-co-MBA. The hyperbranched polymers HPAcManEMA-co-MBA (200 mg) were dissolved in 8 mL CH 3 Cl:CH 3 OH (1:1), under nitrogen gas atmosphere at room temperature. After 15 min, fresh CH 3 ONa (1 mL, 1 M) was added and kept for 1 h, and the white solid was precipitated from the solution. After filtration, the white solid was dissolved in water, purified with cation-exchange resins, and finally dialyzed for 5 days, lyophilized to get white water-soluble products HPManEMA-co-MBA. The hyperbranched glycopolymers so obtained are denoted as S1, S2, S3 and S4, and the figure refers to the feed ratio of AcManEMA and MBA (100:3, 100:5, 100:10 and 100:20, respectively). The synthetic procedure of linear polymer PManEMA was the same as that of HPManEMA-co-MBA, but the CTA and MBA were not used.

Reaction Mechanism of HPAcManEMA-co-MBA
The mechanism for synthesis of HPAcManEMA-co-MBA is outlined in Scheme 1. In the initiation stage, the initiator 2,2 -azobis (2-methylpropionitrile) (AIBN) was decomposed into primary radicals, which initiated vinyl groups to obtain segments bearing radicals named chain radicals. In the chain growth stage, the RAFT polymerization mechanism illustrated that CTA reacted with chain radicals, and the new formation (dormant species) lost the ability to initiate vinyl groups. This reaction was reversible. Thus, the molecular chain would propagate until the chain radicals (P m+n *) or the new radicals (R 2 *) leave from the dormant species. It is worth noting that MBA contains two vinyl groups so that the monomers can easily form a network structure. Thus, in the branched chain growth stage, the other vinyl group of MBA unit was initiated by R 2 *, P m+n * or primary radical, forming branched chain under the help of CTA which inhibited the gelation formation [37]. In the last stage, after the oxygen was introduced in the reaction system, the chain termination was achieved by disproportionation reaction or coupling reaction. Polymers 2018, 10, x FOR PEER REVIEW 4 of 11 Scheme 1. Mechanism of the polymerization of HPAcManEMA-co-MBA.

Determination of DB Value
The structure of HPManEMA-co-MBA (S1) was characterized by 1 H NMR. Compared with the spectrum of AcManEMA (Figure 1a), the chemical shift of HPAcManEMA-co-MBA (Figure 1b) was almost unchanged, but the signals of vinyl group disappeared, indicating that there was no vinyl group in the polymer structure. Meanwhile, the peaks at 0.75-1.25 ppm appeared, suggesting the formation of main chain of polymer. After the deprotection of HPAcManEMA-co-MBA, the peaks at 1.9-2.2 ppm, assigned to acetyl group, decreased significantly in Figure 1c, suggesting that almost all acetyl groups were removed to form HPManEMA-co-MBA. Therefore, the results of 1 H NMR confirmed the successful incorporation of ManEMA and MBA units into the HPManEMA-co-MBA via RAFT polymerization. Nevertheless, the key parameter of DB was difficult to elucidate due to the overlapping of signals deriving from the proton in the ManEMA and MBA units.

Determination of DB Value
The structure of HPManEMA-co-MBA (S1) was characterized by 1 H NMR. Compared with the spectrum of AcManEMA (Figure 1a), the chemical shift of HPAcManEMA-co-MBA (Figure 1b) was almost unchanged, but the signals of vinyl group disappeared, indicating that there was no vinyl group in the polymer structure. Meanwhile, the peaks at 0.75-1.25 ppm appeared, suggesting the formation of main chain of polymer. After the deprotection of HPAcManEMA-co-MBA, the peaks at 1.9-2.2 ppm, assigned to acetyl group, decreased significantly in Figure 1c, suggesting that almost all acetyl groups were removed to form HPManEMA-co-MBA. Therefore, the results of 1 H NMR confirmed the successful incorporation of ManEMA and MBA units into the HPManEMA-co-MBA via RAFT polymerization. Nevertheless, the key parameter of DB was difficult to elucidate due to the overlapping of signals deriving from the proton in the ManEMA and MBA units. The DB value of HPManEMA-co-MBA was tuned by the feed ratio of AcManEMA and MBA. The molecular weight and α value were determined by MDSEC, which was equipped with the differential refractive index (RI), viscometer, and two-angle light scattering (LS) triplet detectors. The parameters of HPManEMA-co-MBA (S1, S2, S3 and S4), including yield, molecular weight, α and DB value, are shown in Table 1. The polymers showed approximately equal yields, which illustrated the activities of AcManEMA and MBA were nearly. The molecular weight increased from 8.8 × 10 4 Dalton to 16.8 × 10 4 Dalton with decreasing the feed ratio of AcManEMA and MBA. This might be due to the double vinyl groups of MBA that increased the probability of polymer chain re-propagating. If the resulting polymers were reinitiated from the MBA units, branch chains were formed, giving rise to a higher DB value. This surmise is also supported by the descending trend of α values from S1 to S4 in Table 1, suggesting that the corresponding DB value increased and the topological structures of HPManEMA-co-MBA were more compact. As presented in Table 1, the α values of HPManEMA-co-MBA were much smaller than that of linear PManEMA (0.77), which further supported the formation of hyperbranched topological structure. Nevertheless, the specific DB value was not obtained yet. To obtain the specific DB value of HPManEMA-co-MBA, elemental analysis was used to measure the content of elemental C, N and H, and the results are listed in Table 2. Comparing the The DB value of HPManEMA-co-MBA was tuned by the feed ratio of AcManEMA and MBA. The molecular weight and α value were determined by MDSEC, which was equipped with the differential refractive index (RI), viscometer, and two-angle light scattering (LS) triplet detectors. The parameters of HPManEMA-co-MBA (S1, S2, S3 and S4), including yield, molecular weight, α and DB value, are shown in Table 1. The polymers showed approximately equal yields, which illustrated the activities of AcManEMA and MBA were nearly. The molecular weight increased from 8.8 × 10 4 Dalton to 16.8 × 10 4 Dalton with decreasing the feed ratio of AcManEMA and MBA. This might be due to the double vinyl groups of MBA that increased the probability of polymer chain re-propagating. If the resulting polymers were reinitiated from the MBA units, branch chains were formed, giving rise to a higher DB value. This surmise is also supported by the descending trend of α values from S1 to S4 in Table 1, suggesting that the corresponding DB value increased and the topological structures of HPManEMA-co-MBA were more compact. As presented in Table 1, the α values of HPManEMA-co-MBA were much smaller than that of linear PManEMA (0.77), which further supported the formation of hyperbranched topological structure. Nevertheless, the specific DB value was not obtained yet. To obtain the specific DB value of HPManEMA-co-MBA, elemental analysis was used to measure the content of elemental C, N and H, and the results are listed in Table 2. Comparing the structure of ManEMA and MBA, elemental N was only contained in the MBA unit, thus the content of elemental N could represent the number of branch point. According to the structural formula of repeat units ManEMA (C 12 H 20 O 8 ) and MBA (C 7 H 10 N 2 O 2 ), the ratio of ManEMA and MBA could be calculated from Equation (1). For the units MBA, the mass ratio of C:N should be 3:1. Therefore, in Equation (1), 3N% represented the elemental C content of MBA, and the (Total C%-3N%) was the elemental C content of ManEMA. After normalization, the content ratio of ManEMA and MBA units R (ManEMA:MBA) is shown in Table 2. In HPManEMA-co-MBA, ManEMA was the linear unit (L) and terminal unit (T), and MBA acted as branch unit (D), supported by the results of 1 H NMR. In the low molar mass region, the number of the branch unit (D) is approximating the number of the terminal unit (T) [38], thus the DB could be written as Equation (2), and the corresponding DB values of HPManEMA-co-MBA were displayed in Table 2. Therefore, the results of elemental analysis were the direct evidence that DB value of samples increased from 12.4% to 30.6%.

Effect of DB and M n on The Binding Behavior between HPManEMA-co-MBA and Con A
To obtain the influence of DB on the recognition event, HPManEMA-co-MBA with the same M n but different DB values were used to investigate the binding behavior between Con A and HPManEMA-co-MBA. Firstly, the samples of S1, S2, S3 and S4 were fractionated by preparative gel permeation chromatography to yield purified HPManEMA-co-MBA fractions. The obtained fractions (S1-X, S2-X, S3-X and S4-X, X = 1, 2, 3, 4 or 5) were measured by MDSEC and labeled with M n and α value, and the data were listed in the Table S1. As shown in Figure 2, the M n of S1-5, S2-4 and S3-4 were nearly equal (M n = 2.7 × 10 4 Dalton), whereas the α values were 0.71, 0.63 and 0.54, respectively. According to Equation (2), the DB values of S1-5, S2-4 and S3-4 were 12.4%, 17.8% and 23.2%, respectively. Thus, these fractions S1-5, S2-4 and S3-4 were employed to investigate the influence of DB on the aggregation rate between Con A and HPManEMA-co-MBA, and the results were shown in Figure 3. The affinity of Con A and HPManEMA-co-MBA was evaluated in terms of the initial clustering rate (k i /∆A 420 s −1 ), which was obtained from the slope of the tangent at the early stage (A 420 = 0.04) [39]. The corresponding k i of S1-5, S2-4 and S3-4 were 9.0 × 10 −4 , 4.5 × 10 −4 and 2.6 × 10 −4 , respectively, while the k i of linear PManEMA was 1.4 × 10 −2 , asreported by Obata [39]. These results revealed that the affinity of HPManEMA-co-MBA decreased when the DB value of HPManEMA-co-MBA increased from 12.4% to 23.2%. This tendency was opposite to that reported by Fernandez-Megia [40]. Note that the low epitope density of the multivalent ligand may lead to a lower rate in the initial aggregate of ligand and lectin (i.e., PManEMA and Con A) [41]. Thus, the opposite trend can be attributed to the lower content of ManEMA units in the HPManEMA-co-MBA.  After sample S4 was fractionated, the serial fractions S4-X (X = 1, 2, 3, 4, 5) would possesses the same DB value, which was supported by the results of α values (Table S1) and elemental analysis (Table S2). The Mn of fractions S4 decreased from S4-1 to S4-5, as shown in Table S1. Therefore, the serial fractions S4-X (X = 1, 2, 3, 4, 5) with the same DB (30.6%) but different Mn could be used to investigate the influence of Mn on the binding behavior between Con A and HPManEMA-co-MBA. As presented in Figure 4, the absorbance at 420 nm increased with decreasing of Mn, and the corresponding ki were 2.4 × 10 −4 , 2.0 × 10 −4 , 1.6 × 10 −4 , 0.9 × 10 −4 and 0.8 × 10 −4 , respectively. As reported by Cairo [41], low epitope density of the multivalent ligand causes the low ki. The concentration of multivalent ligand can be estimated by the ManEMA unit, thus the mole concentration of HPManEMA-co-MBA with lower Mn should be higher than those of higher Mn. Therefore, the high mole concentration means that the sample has more opportunities to form the cross-linked complexes leading to a higher ki.  After sample S4 was fractionated, the serial fractions S4-X (X = 1, 2, 3, 4, 5) would possesses the same DB value, which was supported by the results of α values (Table S1) and elemental analysis (Table S2). The Mn of fractions S4 decreased from S4-1 to S4-5, as shown in Table S1. Therefore, the serial fractions S4-X (X = 1, 2, 3, 4, 5) with the same DB (30.6%) but different Mn could be used to investigate the influence of Mn on the binding behavior between Con A and HPManEMA-co-MBA. As presented in Figure 4, the absorbance at 420 nm increased with decreasing of Mn, and the corresponding ki were 2.4 × 10 −4 , 2.0 × 10 −4 , 1.6 × 10 −4 , 0.9 × 10 −4 and 0.8 × 10 −4 , respectively. As reported by Cairo [41], low epitope density of the multivalent ligand causes the low ki. The concentration of multivalent ligand can be estimated by the ManEMA unit, thus the mole concentration of HPManEMA-co-MBA with lower Mn should be higher than those of higher Mn. Therefore, the high mole concentration means that the sample has more opportunities to form the cross-linked complexes leading to a higher ki. After sample S4 was fractionated, the serial fractions S4-X (X = 1, 2, 3, 4, 5) would possesses the same DB value, which was supported by the results of α values (Table S1) and elemental analysis (Table S2). The M n of fractions S4 decreased from S4-1 to S4-5, as shown in Table S1. Therefore, the serial fractions S4-X (X = 1, 2, 3, 4, 5) with the same DB (30.6%) but different M n could be used to investigate the influence of M n on the binding behavior between Con A and HPManEMA-co-MBA. As presented in Figure 4, the absorbance at 420 nm increased with decreasing of M n , and the corresponding k i were 2.4 × 10 −4 , 2.0 × 10 −4 , 1.6 × 10 −4 , 0.9 × 10 −4 and 0.8 × 10 −4 , respectively. As reported by Cairo [41], low epitope density of the multivalent ligand causes the low k i . The concentration of multivalent ligand can be estimated by the ManEMA unit, thus the mole concentration of HPManEMA-co-MBA with lower M n should be higher than those of higher M n . Therefore, the high mole concentration means that the sample has more opportunities to form the cross-linked complexes leading to a higher k i .

Cytotoxicity Test of HPManEMA-co-MBA
In the biomedical field, cell viability is often used to determine the cytotoxicity and biocompatibility of biomaterials. The cytotoxicity of HPManEMA-co-MBA against HeLa cells was studied using the MTT assay. Figure 5 shows the cell viability after incubation with S4 at different concentrations. In the scope of the testing concentration, the cell viabilities were around 100% after 6 h incubation with S4. In the case of 0.15 mg·mL −1 , the cell viability was higher than 90% even when the cells were incubated for 24 h. These results showed that the cytotoxicity of HPManEMA-co-MBA is very low in a short period of time. The good viability of obtained polymers satisfied the demand of common-used drug carriers. Owing to the low cytotoxicity, HPManEMA-co-MBA is proposed as a potential system for drug delivery applications.

Conclusions
In conclusion, novel hyperbranched glycopolymers HPManEMA-co-MBA are successfully synthesized via RAFT polymerization using the MBA as the branch unit. The DB value is facilely determined via elemental analysis because the elemental nitrogen only exists in branch units. The HPManEMA-co-MBA samples with the same defined DB value or molecular weight are employed to investigate the binding behavior between HPManEMA-co-MBA and Con A. In the case of the same Mn (Mn = 2.7 × 10 4 Dalton) but different DB values, the clustering rate of HPManEMA-co-MBA and Con A decreases with the increasing of DB value. In contrast, in the case of the same DB value

Cytotoxicity Test of HPManEMA-co-MBA
In the biomedical field, cell viability is often used to determine the cytotoxicity and biocompatibility of biomaterials. The cytotoxicity of HPManEMA-co-MBA against HeLa cells was studied using the MTT assay. Figure 5 shows the cell viability after incubation with S4 at different concentrations. In the scope of the testing concentration, the cell viabilities were around 100% after 6 h incubation with S4. In the case of 0.15 mg·mL −1 , the cell viability was higher than 90% even when the cells were incubated for 24 h. These results showed that the cytotoxicity of HPManEMA-co-MBA is very low in a short period of time. The good viability of obtained polymers satisfied the demand of common-used drug carriers. Owing to the low cytotoxicity, HPManEMA-co-MBA is proposed as a potential system for drug delivery applications.

Cytotoxicity Test of HPManEMA-co-MBA
In the biomedical field, cell viability is often used to determine the cytotoxicity and biocompatibility of biomaterials. The cytotoxicity of HPManEMA-co-MBA against HeLa cells was studied using the MTT assay. Figure 5 shows the cell viability after incubation with S4 at different concentrations. In the scope of the testing concentration, the cell viabilities were around 100% after 6 h incubation with S4. In the case of 0.15 mg·mL −1 , the cell viability was higher than 90% even when the cells were incubated for 24 h. These results showed that the cytotoxicity of HPManEMA-co-MBA is very low in a short period of time. The good viability of obtained polymers satisfied the demand of common-used drug carriers. Owing to the low cytotoxicity, HPManEMA-co-MBA is proposed as a potential system for drug delivery applications.

Conclusions
In conclusion, novel hyperbranched glycopolymers HPManEMA-co-MBA are successfully synthesized via RAFT polymerization using the MBA as the branch unit. The DB value is facilely determined via elemental analysis because the elemental nitrogen only exists in branch units. The HPManEMA-co-MBA samples with the same defined DB value or molecular weight are employed to investigate the binding behavior between HPManEMA-co-MBA and Con A. In the case of the same Mn (Mn = 2.7 × 10 4 Dalton) but different DB values, the clustering rate of HPManEMA-co-MBA and Con A decreases with the increasing of DB value. In contrast, in the case of the same DB value

Conclusions
In conclusion, novel hyperbranched glycopolymers HPManEMA-co-MBA are successfully synthesized via RAFT polymerization using the MBA as the branch unit. The DB value is facilely determined via elemental analysis because the elemental nitrogen only exists in branch units. The HPManEMA-co-MBA samples with the same defined DB value or molecular weight are employed to investigate the binding behavior between HPManEMA-co-MBA and Con A. In the case of the same M n (M n = 2.7 × 10 4 Dalton) but different DB values, the clustering rate of HPManEMA-co-MBA and Con A decreases with the increasing of DB value. In contrast, in the case of the same DB value (30.6%) but different M n , the clustering rate increases with the decreasing of M n . In addition, the obtained HPManEMA-co-MBA shows a low cytotoxicity, and exhibits potential biomedical applications as drug delivery systems.