Polymerization and characterization of HEMA-TMS macromonomer and

85

the clearer results. The BSA in HEPES buffer (10 mg/mL) was injected and the protein exposure was interrupted by exchanging the protein solution for a pure buffer solution.

3.2.6 Bacterial adhesion on polymer-immobilized HAp surfaces

S. epidermidis was cultured in the nutrient medium (polypeptone 5 g/L, beef extract 3 g/L, and NaCl 5 g/L) and constantly shaken at 100 rpm (Shaker NTS-2100, Tokyo Rikakikai Co., Ltd, Tokyo, Japan) until the concentration was approximately 10⁸ CFU/mL. Bacterial suspensions were centrifuged and washed with PBS buffer (NaCl 137 mM, KCl 2.7 mM, Na2HPO4 10 mM, KH2PO4 1.8 mM) three times to remove the culture medium.

The polymer-immobilized HAp plates were placed in 24 well plates. A 0.5 mL S.

epidermidis PBS suspension was added to each well and then incubated at 37 °C for 2.5 h to imitate the initial step of bacteria adhesion. After cultivation, to evaluate the adhesion amount on the surfaces, the cells were fixed for SEM (SU8000, Hitachi High-Technologies Co., Tokyo, Japan) observation. The plates were washed with PBS buffer and water to remove the non-adherent fractions. A 2.5% glutaraldehyde PBS solution was added to each well and incubated for 24 h at 37°C to fix the bacterial cell shape.

After fixation, the plates were gently washed with PBS buffer and dehydrated with 25, 30, 50, 70, 80, 90 and 100% ethanol for 5, 5, 10, 10, 10, 10 and 10 min (thrice), respectively. The number of adhered cells on the HAp plates was counted based on at least 25 images from five separate areas on each plate (n = 3).

86

The RAFT polymerization has been commonly used to directly prepare linear polymers with controlled molecular weights and low polydispersities. To prepare a more well-defined macroinitiator, I used the hydroxyl-group-protected monomer, HEMA-TMS, which has a lower polarity and tolerable solubility, and therefore, can be readily polymerized ¹³. A densely grafted brush polymer was synthesized by chain extension of PEGMA from the macroinitiator precursor. The polymerization produced the desired polymers in good yield (approximately 87%), showing the high blocking efficiency of PEGMA on the HEMA-TMS macromonomer backbone. With the control block polymers in hand, phosphate-functionalized polymers were prepared by reaction of phosphorus oxychloride with the hydroxyl groups on the side chains. After purification by dialysis, the polymer structure was confirmed by ¹H NMR. The peak shifts in SEC were used to confirm the chain growth of the polymers and evaluate the molecular weight. To prove the presence of phosphate groups, EDX of the purified polymers was performed, which revealed that the weight of P increased with increasing hydroxyl group content in the polymers. The yield of phosphorylation was more than 95%. The synthetic approaches for the preparation of the macroinitiator, block copolymers, and phosphorylation were shown in Scheme 3-1.

87

Scheme 3-1. Synthesis of the HEMA-TMS macroinitiator (A), poly(HEMA-TMS-b- PEGMA) by RAFT polymerization (B), phosphorylation of block polymers (C), and polymers after dialysis (D).

The ¹H NMR peak assignment was shown in Figure 3-1. The peak at approximately 4.75 ppm assigned to the hydroxyl groups of HEMA was clearly distinguishable from the backbone signals in this well-resolved spectrum. This indicated the hydrolysis of the TMS protecting groups during the polymerization of the macro-initiator (A) and block copolymers (B). Because the raw materials and solvents were not dehydrated before use, the presence of residual water in the reaction mixture might have resulted in the instability and hydrolysis of the TMS groups. The phenyl group of the CAPAB terminal, ppm: 7.85-7.81 (m, 2H, CHC=S), 7.63-7.65 (m, 1H, CHCHCH), 7.45-7.50 (m, 2H, CHCHC=S), was clearly discernable, which confirmed that the macroinitiator was effectively able to transfer radicals in the subsequent polymerization. As the chain of PEGMA extended, the peaks corresponding to the proton signals of the ethyl groups in PEGMA and HEMA brushes partially overlapped.

88

After phosphorylation, the peak assigned to the hydroxyl group disappeared completely (D), implying the desired preparation of phosphorylated block copolymers.

Figure 3-1. ¹H NMR spectrum (400 MHz, DMSO) of the HEMA-TMS macroinitiator (A) after precipitation, block copolymer poly(HEMA23-b-PEGMA104) (B), and phosphorylated block copolymer (D) after dialysis purification.

The SEC showed the complete attachment of the block polymers to the HEMA macroinitiator with a clean shift of the distribution towards a lower retention time (Figure 3-2). The polydispersities of the block polymers were less than 1.35, which were narrow enough to prove the completion of the living radical block polymerization.

89

The molar mass distribution of the block copolymers was monomodal but not perfectly symmetric, indicating that the samples showed small shoulder peaks at a lower retention time (higher molecular weight). This was attributed to the complete deprotection of the hydroxyl groups in HEMA during the polymerization of the block polymers, resulting in higher interaction with the column owing to the higher polarity. In consideration of the signal of the hydroxyl groups and the disappearance of TMS groups in ¹H NMR (Figure 3-1), the polarity of the polymers was presumed to change during the reaction.

PolyHEMA is soluble only in extremely polar solvents such as DMF and might be comparatively less soluble in the eluent used in the SEC. Another possibility could be excessively high conversion causing over propagation and chain-chain coupling during the latter period of the reaction. Even there was undesired propagation, the molecular weight distribution was still acceptable. The molar mass evaluation of the macroinitiator and the corresponding block polymers is summarized in Table 3-1.

Figure 3-2. SEC of the TMS macroinitiator (A) and the corresponding HEMA-PEGMA block copolymers (B) (in 10 mM LiBr DMAc solution, the determination was calibrated with PMMA standards). The SEC was performed on a HLC-8320 GPC Eco-SEC equipped with a TSKgel Super AW guard column and TSKgel Super AW (4000 and 2500) columns (TOSOH, Tokyo Japan).

90

Table 3-1. Summary of polymerization and the physical characterization of polymers.

Polymer

Macro-HEMA Mn(×10³) PDI Block polymer Mn (×10⁵) PDI

1 23 5.0 1.14 HEMA23

-b-PEGMA104

5.5 1.22

2 50 10.5 1.02 HEMA50

-b-PEGMA111

6.0 1.32

3 81 16.8 1.05 HEMA81

-b-PEGMA115

6.7 1.31

4 105 22.0 1.08 HEMA105

-b-PEGMA108

7.2 1.30

The existence of phosphorus in the phosphorylated block polymers was determined by EDX. After sufficient dialysis in distilled water, the polymers should be purified and hydrolyzed completely. Thus, the P constituent can be attributed to the polymers after phosphorylation. The phosphorus content increased as the phosphorylated HEMA content in the polymers increased. The calibrated (with nonfreezing water) value of phosphorus was consistent with the theoretical one, which confirmed the phosphorylation reaction and the existence of phosphorus in polymers (data shown in Figure 3-3). The measured total phosphorus proportion was approximately 1 wt%, which was lower than the theoretical value. Carbon, hydrogen, and oxygen were the major constituents of the polymer. These constituents included the hydrated water, the sorbed water (so-called nonfreezing water) penetrating the polymer matrix. The sorbed water can interact with the carbonyl groups in PEGMA and HEMA through hydrogen bonding ¹⁴.

91

Figure 3-3. Summary of quantitative phosphorus content in polymers determined by EDX. (1: polymer 1, 2: polymer 2, 3: polymer 3, 4: polymer 4) (n = 3 samples of each condition)

3.3.2 Polymer immobilization on HAp surfaces

The XPS analyses were carried out on the unmodified and polymers-modified HAp surfaces to prove the immobilization of the polymers. Figure 3-4 showed the corresponding regions of the spectrum measured with the unmodified and polymer-immobilized HAp surfaces. The single peak at 285.0 eV was ascribed to alkyl type carbon (C-C, C-H). The second peak, which was 1.5 eV higher than the main peak and had the same FWHM, was ascribed to the ester (C-O-C). The peak at a binding energy of approximately 288 eV higher than the main peak was assigned to the carbonyl (C=O).

92

Figure 3-4. XPS C(1s) spectra of (A) UV/O3 treated raw HAp surface, polymer 1- immobilized HAp surface, polymer 2-immobilized HAp surface, polymer 3-immobilized HAp surface, and polymer 4-3-immobilized HAp surface.

After immobilization, the peaks that were derived from the polymers increased significantly. These results indicated that the polymers were deposited onto the HAp surfaces. The peaks that were detected on the raw surface were presumed to be the adventitious carbon produced by the short exposures to the atmosphere or carbon contamination within the vacuum of the XPS chambers ^15,16. Despite the presence of peaks assigned to impurities, the immobilization of polymers was evident.

The relative amount of immobilized polymer was calculated according to the XPS results. Because the peaks assigned to C=O were considered to predominantly result from monomeric units in the polymer, whereas the Ca(2p) peaks arise entirely from HAp, the relative number of immobilized polymer could be calculated as the proportion of the atomic concentration of C=O to Ca ¹⁷. To estimate the durability of the polymers,

93

coated HAp plates were immersed into water for 2.5 h and 18 h and examined using the same measurements and calculations. All the results were summarized in Figure 3-5.

Figure 3-5. Relative immobilization amounts of polymer units on the HAp surfaces and after 2.5 h and 18 h immersion. (control: raw UV/O3 retreated HAp surface, 1: polymer 1, 2: polymer 2, 3: polymer 3, 4: polymer 4) (n = 3 samples of each condition)

All the polymers were confirmed to immobilize on the HAp surfaces, even polymer 1, which contains very short bonding segment. Polymer 2 showed the highest number of immobilized monomeric units. Compared with the primary data, no obvious reduction in the number of immobilized units on HAp was observed, which indicated that the polymers were still very stable even after long-term soaking. In detail, a short phosphorylated HEMA segment (approximately 20 mer) of the polymer was enough for immobilization onto the HAp surface. As the segment was extended to a 50 mer, the immobilization amount on the surface increased to reach the maximum value. However, the opposite tendency was observed with increasing bonding segment (81 mer to 105 mer) in the polymers. Because of the longer bonding sites, the intervals between the polymers increased, which resulted in a lower density of the polymers attached on the

94 surfaces.

These results suggested that the immobilization of polymers on the HAp surface was achieved with considerable durability. Even a short bonding segment (23 mer) was capable of coating onto the HAp surfaces. Because the polymer immobilization was investigated with change in time in the last project, the detailed process was not repeated here. The most efficient self-etching enamel-dentin adhesives are dependent on their strongly acidic adhesive monomers ¹⁸. Therefore, a high immobilization amount can be attributed to the strong affinity between the organic phosphates or phosphonates in the polymers to HAp. Interactions including electrostatic bonding, covalent bonding, and hydrogen bonding contribute to the high affinity. Specifically, dissolved Ca²⁺ ions just above the most surficial layer on HAp and the fixed ions in the apatite lattice presumably act as the connection to firmly combine the phosphate sites in the polymers to the surfaces through electrostatic interactions ¹⁹. Meanwhile, phosphate groups attached to the polymer chains may exchange or substitute the phosphate ions in the lattice, causing the rearrangement in the crystalline network for further phosphate adsorption ²⁰. Other theories claim that comparably modest interactions such as hydrogen bonding and some nonspecific intermolecular bonding may also play a role in the immobilization process ¹⁸. Thus, multiple interactions contribute to the high affinity of block polymers to hydroxyapatite.

3.3.3 QCM measurement of polymer immobilization and protein resistance on HAp-functionalized chips

QCM is a well-established technique for investigating the immobilization of polymers and further protein adsorption on modified monolayer surfaces. The polymer immobilization and protein adsorption were monitored through the change in the resonance frequency. The frequency curves were summarized in Figure 3-6. The curves reached saturation quickly, indicating that the adsorption of polymers and protein

95 completes in few minutes.

Figure 3-6. QCM measurement of polymers immobilization (pale grey area) on the HAp functionalized surfaces and then BSA adsorption (pale pink area) on the raw (red) and modified surfaces. Curves labeled with numbers present F on the surfaces modified with polymers: 1: polymer 1, 2: polymer 2, 3: polymer 3, 4: polymer 4.

For the polymer immobilization step, all the polymers were immobilized on the surfaces rapidly and the immobilized polymer weights decreased as the length of the phosphorylated HEMA segment in the polymers increase. Frequency shifts increased from approximately 39 Hz to 51 Hz as the bonding sites reduced from 105 mer (polymer 4) to 23 mer (polymer 1). Qualitatively, even shorter phosphorylated HEMA segments in the polymer were able to provide sufficient bonding sites for larger PEGMA grafting density on the HAp surfaces, which was consistent with the results of XPS. Because the total monomeric units immobilized on the surfaces could be estimated as (m + n)×weight / (m×MnHEMA + n×MnPEGMA), polymer 1 may be less favorable to be evaluated in the form of monomeric units as shown in Figure 3-5.

96

Furthermore, the combination of water with the dense PEG brushes of polymer 1 may increase the frequency shift. These may be the cause of the inconsistent result compared with those of the XPS analysis and other reasons are under investigation.

For the protein adsorption step, to obtain a clearer result under the effect of buffer on protein adsorption ¹², the experiments were conducted with HEPES buffer.

Compared with the adsorption curve of BSA on the raw HAp (F of approximately 45 Hz), the F was reduced to 14, 5, 4 and 2 Hz on polymer 4, 3, 2, and 1 coated HAp surfaces, respectively. A dramatic decrease of F was achieved on the different polymer immobilized surfaces. The results suggested that polymer 1 showed superior performance to that of the other polymers and the protein resistance depended on the polymer and PEG density on the surfaces. The PEG density is known to determine the degree of protein resistance ^9,21. The grafting density of the polymers, meaning the surface density of the PEG brushes changes according to the length of the bonding sites on the polymers. Thus, the shorter the bonding sites, the better the PEG brushes surface density and protein adsorption resistance. Specifically, polymer 4 contained the longest bonding sites, which resulted in the lowest surface grafting density and eventually to a large F, indicating large protein adsorption and lower resistant efficacy compared with the other polymers.

3.3.4 Bacterial adhesion on polymer-immobilized HAp plates

The antibacterial properties of different polymers were investigated by S.

epidermidis adhesion experiments on the polymer-modified HAp surfaces. The number of adhered bacteria per surface area are summarized in Figure 3-7 and representative SEM images are shown in Figure 3-8. The results indicate that the polymers show considerable antibacterial efficacy for S. epidermidis adhesion. Compared with the number of adhered bacteria on unmodified HAp surfaces, adhered bacteria on the polymer-immobilized surfaces were reduced effectively. The inhibitory behavior varied

97

depending on the different polymers composition. Polymer 1 displayed superior bacterial inhibition efficacy to the other polymers, with a bacterial adhesion inhibition of 70%. Polymers 2 and 3 showed some bacterial inhibition, whereby the number of adhered bacteria were reduced to 50%. Polymer 4 on the HAp had a small effect and no significant difference compared with the bare surface, with only a 20% reduction in bacterial attachment.

Figure 3-7. The amounts of adhered bacteria on the raw and polymer-immobilized HAp surfaces after short-term exposure (2.5 h) to the S. epidermidis (10⁸ CFU/mL) suspensions. (n = 25 images of 5 separate areas × 3 samples). Student’s t-test for paired data was used for statistical comparison between the control and inhibition conditions. P values < 0.05 were considered as significant and are marked by an asterisk above the column.

98

Figure 3-8. Representative SEM images of adhered bacteria on the raw and polymers-immobilized HAp surfaces. Bacteria were marked with red circles. Scale bar, 10 m.

The plausible explanation could be that the shorter intervals between the polymers after immobilization of polymer 1 contributed to the most compact brushes. The adequate brushes grafting density on the modified surfaces resulted in effective anti-biofouling performance. On the contrary, insufficient grafting density obtained by using polymer 4 showed unsatisfactory bacterial inhibition behavior. The dense PEG chains