Restriction of CaCO₃ polymorph by NH⋯O hydrogen-bonded poly(methacryloylaminocarboxylate) ligands: induced polymorph change by strength and/or formation manner of hydrogen bond

Abstract

Novel poly(methacryloylaminocarboxylate) ligands were synthesized as models for an acidic peptide in CaCO₃ biominerals. Syndiotactic poly[3-methyl-2-(methacryloylamino)butyrate], isotactic-rich poly[3-(methacryloylamino)propionate], and syndiotactic poly[4-(methacryloylamino)butyrate] form 5-, 6-, and 7-membered-ring intra-side-chain NH⋯O hydrogen bonds, respectively, between a carboxylate oxyanion and the neighboring amide NH proton, while syndiotactic poly[3-(methacryloylamino)propionate] forms intermolecular NH⋯O hydrogen bonds. These polymer ligands strongly bind to CaCO₃ crystals because of the NH⋯O hydrogen bonds. In polymer ligand–CaCO₃ composites, the strength and/or mode of hydrogen bonding in the polymer ligands affects the CaCO₃ polymorphs. Five-membered-ring intra-side-chain and intermolecular hydrogen-bonded polymers have weak hydrogen bonds and produce calcite crystals, whereas other ligands with 6- and 7-membered-ring intra-side-chain hydrogen bonds have strong hydrogen bonds and yield vaterite crystals. Polymorph change of CaCO₃ crystal is induced by differences of strength and/or mode of the NH⋯O hydrogen bond.

---

Introduction

Biominerals occurring in nature, such as those that make up mollusk shells, teeth, and magnetotactic bacterial magnetosomes, are usually composed of biopolymers and inorganic crystals, e.g., CaCO₃, Ca₁₀(PO₄)₆(OH)₂, Fe₃O₄, etc.¹ These biomineralization processes are often mediated and regulated by a small amount of a biopolymer ligand.² When shells biomineralize, for example, oriented CaCO₃ crystals grow on a nucleating protein sheet.^3–6 Through this process, a small amount of organic material binds to the crystal face and controls the crystal growth, and the resulting biominerals possess highly controlled shapes and sizes under ambient conditions.⁶ Actually, our recent report indicates that each aragonite brick in the nacreous layer of Pinctada fucata (Japanese pearl oyster) is not a single crystal but a composition of nano-ordered aragonite crystals and that biopolymers participate in the generation and orientation of aragonite nanocrystals.⁷

In CaCO₃ architectures, unusually acidic proteins are found in aragonite- (stable form of CaCO₃) or calcite- (the most stable form of CaCO₃) containing mineralized tissues of invertebrates, which are rich in Asx (i.e., Asp or Asn) and/or Glx (i.e., Glu or Gln).^8–11 For example, the soluble protein from the mollusk Atrina rigida contains about 40 mol% of acidic residues (Asx and Glx).¹⁰ The protein components of the biopolymer ligands from the ascidian spicules, which are extracted from the amorphous CaCO₃ produced by sponges,⁹ are rich in Glu and/or Gln.¹¹ Interestingly, Falini et al. demonstrated that when the assemblage of unusually acidic proteins from an aragonite layer is added, aragonite crystal forms, and when the proteins are derived from a calcite layer, calcite crystal forms.¹² Acidic residues of most of the proteins affect the mineralization of CaCO₃ architectures.

Artificial carboxylate or poly(carboxylate) ligand–CaCO₃ composites were used as models for biominerals that contain acidic peptides in the previous reports.^5,13–18 Some reports treated unique ligands having a extremely elaborate structure, such as anionic porphyrins,¹⁴ dendrimers,¹⁵ helical poly(isocyanide),¹⁶ calixarenes,¹⁷ aerosol thin films,¹⁸etc. It has been shown that the carboxylate produces a specific surface during CaCO₃ crystallization.

The strong binding of biopolymer ligands to a CaCO₃ crystal is one of the puzzles of mimicking biominerals. In this respect, we have discussed the scenario of an NH⋯O hydrogen bonding between a Ca(II)-coordinated oxygen atom and the neighboring amide NH proton. This NH⋯O hydrogen bond prevents the metal–oxygen bond from dissociating^19–21 like the NH⋯S hydrogen bond, which shifts the pK_a value of a thiol²² and stabilizes the metal–thiolate complexes.^23,24 The cytochrome P-450 model (porphinato)(thiolato)iron(III) complexes with single or double NH⋯S hydrogen bonds are stable against O₂ or H₂O,²³ although all the known iron(III) porphyrin thiolato complexes have been reported to be unstable in the presence of air and moisture.²⁵ Furthermore, 2,6-diacetylaminobenzoate has a higher formation constant toward Ca²⁺ ion than does nonsubstituted benzoate because of the formation of an intramolecular NH⋯O hydrogen bond.²¹ In biominerals that contain acidic peptides, these effects of NH⋯O hydrogen bonds probably contribute to the stabilization of a Ca–carboxylate oxygen bond.

Previously, we synthesized poly(carboxylate) ligands that had an NH⋯O hydrogen bond between a carboxylate oxygen and the neighboring amide NH proton; we used this construct as an artificial model for acidic peptides in CaCO₃ biominerals.^26–28 These polymer ligands strongly bind to CaCO₃ crystals through the hydrogen bond. In the absence of such a hydrogen bond, the metal–oxygen bond is readily hydrolyzed under neutral conditions. Addadi et al. also reported the difference between poly(acrylate) and native acidic peptides.⁵ However, our polymer ligands have problems in terms of stereoselectivity or conformation of a polymer main-chain, because these polymers were prepared by polymer reactions that include poly(carboxylic anhydride) with alkylamine,^26,27 or poly(allylamine) with carboxylic anhydride.²⁸ Additionally, introduction of a hydrogen-bonded unit to a polymer side-chain was not in so high yield.

In the present study, we synthesized novel poly(methacryloylaminocarboxylate) ligands having 5-, 6-, or 7-membered-ring intra-side-chain NH⋯O hydrogen bonds (Chart 1). These polymer ligands were prepared from the corresponding monomer units and the stereochemistry of the polymer main-chain was controllable. Furthermore, all side-chains contain a hydrogen-bonded unit. In this paper, we demonstrate that these polymer ligands restrict CaCO₃ morphologies by utilizing ¹³C cross-polarization/magic angle spinning (CP/MAS) NMR, X-ray powder diffraction (XRD), and field emission/scanning electron microscopy (FE/SEM) techniques.

Chart 1

Experimental

Materials

Methacryloyl chloride was obtained from Tokyo Kasei Kogyo Co., Ltd. (Tokyo, Japan) and distilled before use. Synthetic procedures of monomers, MA-Val-OPac, MA-β-Ala-OBzl, and MA-γ-Abu-O^tBu, are described in the ESI.† Other reagents were commercially obtained and solvents were used after distillation. All synthetic procedures were performed under Ar atmosphere.

Poly(MA-Val-OPac)

A mixture of MA-Val-OPac (761 mg, 2.5 mmol) and 2,2′-azobisisobutyronitrile (AIBN) (3.2 mg, 20 µmol) was stirred at 353 K. After 2 days, 2 mL of CH₂Cl₂ was added into the mixture and the solution was added dropwise into n-hexane. The precipitate was collected by filtration and dissolved into 2 mL of CH₂Cl₂. The solution was added dropwise into MeOH to give a white powder. The white powder was collected by filtration and dried over P₂O₅ under reduced pressure. Yield, 21%. M_n = 54000 Da. M_w/M_n = 1.29.

Poly(MA-Val-OH) (1H)

To poly(MA-Val-OPac) (160 mg, 530 µmol) in 20 mL of AcOH was added Zn powder (10.0 g) and the mixture was stirred at room temperature. After 4 hours, the solution was filtered to remove Zn powder and filtrate was concentrated. The residue was washed with CH₂Cl₂ and 2% HCl aqueous solution. Yield, 45 mg (46%). Stereoselectivity (diad): isotactic/syndiotactic = 23/77. The deprotection of carboxyl groups in the polymer side-chain proceeded in 50% yield from ¹H NMR spectra.

Poly(MA-β-Ala-OBzl)

A mixture of MA-β-Ala-OBzl (1.0 g, 4.0 mmol) and benzoyl peroxide (BPO) (9.8 mg, 40 µmol) was stirred at 353 K. After one day, 10 mL of CH₂Cl₂ was added into the mixture and the solution was added dropwise into n-hexane to give a white powder. The white powder was collected by filtration and dried over P₂O₅ under reduced pressure. Yield, 690 mg (69%). M_n = 3100 Da. M_w/M_n = 1.48.

Poly(MA-β-Ala-OH) (2H)

To poly(MA-β-Ala-OBzl) (1.32 g, 5.3 mmol) in 10 mL of MeOH was added 1 M NaOH aqueous solution (5.3 mL, 5.3 mmol) and the solution was stirred at room temperature. After 12 hours, the solution was concentrated and to the residue was added excess 3.5% HCl aqueous solution. The solution was concentrated to dryness. The residue was dissolved in 10 mL of MeOH and the solution was added dropwise into 100 mL of ether to give a white powder. The white powder was collected by centrifugation and dried over P₂O₅ under reduced pressure. Yield, 498 mg (59%). Stereoselectivity (diad): isotactic/syndiotactic = 16/84.

Isotactic-rich poly(MA-β-Ala-OBzl)

A mixture of MA-β-Ala-OBzl (2.2 g, 9.0 mmol), AIBN (15 mg, 90 µmol) and Y(OTf)₃ (480 mg, 0.90 mmol) in 2 mL of MeOH was stirred at 333 K. After 8 hours, 50 mL of CH₂Cl₂ was added into the mixture and the solution was added dropwise into n-hexane (1 L) to give a white powder. The white powder was collected by filtration and dried over P₂O₅ under reduced pressure. Yield, 2.1 g (95%). M_n = 910. M_w/M_n = 1.03.

Isotactic-rich poly(MA-β-Ala-OH) (2^isoH)

2^isoH was obtained by hydrolysis of the isotactic-rich poly(MA-β-Ala-OBzl). This procedure was the same as the method for the hydrolysis of syndiotactic poly(MA-β-Ala-OBzl). Stereoselectivity (diad): isotactic/syndiotactic = 69/31.

Poly(MA-γ-Abu-O^tBu)

A mixture of MA-γ-Abu-O^tBu (830 mg, 3.7 mmol) and AIBN (6.0 mg, 37 µmol) was stirred at 353 K. After one day, 5 mL of CH₂Cl₂ was added into the mixture and the solution was added dropwise into n-hexane to give a white powder. The white powder was collected by filtration and dried over P₂O₅ under reduced pressure. Yield, 378 mg (46%). M_n = 1700 Da. M_w/M_n = 1.19.

Poly(MA-γ-Abu-OH) (3H)

Poly(MA-γ-Abu-O^tBu) (1.235 g, 5.4 mmol) was dissolved into 10 mL of THF at 353 K. A mixture of NaOH (450 mg, 11 mmol) aqueous solution (2 mL) and 3 mL of EtOH were added to the solution. The solution was stirred at 353 K for one day and then THF and EtOH were removed under reduced pressure. The residue was suspended in water and the solution was stirred at room temperature. After one day, 3.5% HCl aqueous solution was added to the solution to give a white powder. Yield, 664 mg (71%). Stereoselectivity (diad): isotactic/syndiotactic = 22/78.

Crystallization of CaCO₃ composites in the presence of polymer ligands

Ten mL of an aqueous calcium chloride solution (1 M) was added to 10 mL of a methanol solution of a polymer ligand (10 mM per repeated unit). After stirring for 3 hours, 10 mL of an aqueous ammonium carbonate solution (1 M) were dropped into the mixed solution and then it was stirred for 3 hours. CaCO₃ crystallization was carried out by mixing these solutions at 303 K under acidic conditions (pH 3.8–4.0). CaCO₃–polymer composites were obtained at pH 7.2–7.4. The final CaCO₃/polymer ligand ratio was 100/1. The obtained CaCO₃ crystals were washed with methanol and distilled water to remove unbound polymer ligand, and dried over P₂O₅ under reduced pressure.

Physical measurements

Solution-state Fourier transform infrared (FT-IR) spectra in chloroform-d were obtained with a Jasco FT-IR 8300. 400 MHz ¹H NMR spectra were taken on a JEOL JNM-GSX400 in dimethyl sulfoxide-d₆ or chloroform-d, at 303 K. Tetramethylsilane was used as a standard on ¹H NMR measurements. 500 MHz ¹H NMR and 1D TOCSY spectra were recorded by a JEOL JNM-LA500 in chloroform-d at 303 K. 75 MHz solid-state ¹³C CP/MAS spectra were recorded on a Chemagnetics CMX-300 spectrometer at room temperature. The chemical shifts of CP/MAS spectra were calculated from the methyl group of hexamethylbenzene (17.36 ppm). The rotating speed was 4000 rps, and contact time was 1.00 ms in the CP/MAS measurements. Gel permeation chromatographic (GPC) analyses were carried out at 313 K using Tosoh MultiporeHXL-M columns connected to a Tosoh UV detector and RI detector. THF was used as an eluent at a flow rate of 0.8 mL min⁻¹. Samples were dissolved in THF (1.0 mg mL⁻¹) and were filtered in order to remove particulates before injection. The molecular weight was calibrated according to poly(styrene) standards. FE/SEM images were obtained with a Hitachi S-5000 at 20 kV acceleration voltages. Samples for FE/SEM analysis were vapor deposited with osmium oxide. XRD patterns were taken on a Rigaku RAD system at room temperature. The radiation used was CuKα monochromatized with graphite.

Results

Polymer ligand preparation

The three types of polymer ligands, poly[3-methyl-2-(methacryloylamino)butylic acid] (1H), poly[3-(methacryloylamino)propionic acid] (2H), and poly[4-(methacryloylamino)butyric acid] (3H), were synthesized by the reactions as shown in Scheme 1. The monomer preparations for 1H, 2H, and 3H were performed by the reaction of methacryloyl chloride with the corresponding C-terminal-protected amino acids, H-Val-OPac, H-β-Ala-OBzl, and H-γ-Abu-O^tBu, respectively (see ESI†). The polymerizations were carried out by using AIBN or BPO to induce free radical polymerization. After the polymerization, the carboxylic acid group in a polymer side-chain was treated by a typical deprotection procedure. The synthesis of these polymer ligands was confirmed by ¹H NMR spectra. Fig. 1 shows the ¹H NMR spectra of the polymer ligands in Me₂SO-d₆ at 303 K. Amide NH ¹H NMR signals in these polymer ligands were observed at 8.00, 7.40, and 7.28 ppm, for 1H, 2H, and 3H, respectively. The ¹H NMR signal of the COOH group in each polymer ligand appeared at levels between 12 and 13 ppm. Other ¹H NMR signals of the polymer main-chain and the side-chain were also observed at a high magnetic field above 4.0 ppm (detailed signal assignments are described in the ESI†).

Fig. 1 ¹H NMR spectra of (a) 1H, (b) 1NMe₄, (c) 2H, (d) 2NMe₄, (e) 2^isoH, (f) 2^isoNMe₄, (g) 3H and (h) 3NEt₄ (303 K, Me₂SO-d₆).

Scheme 1 Synthetic procedure of polymer ligands.

Polymer ligands 2 and 3 completely undergo the deprotection procedure, while the conversion of 1 was 50% from the ¹H NMR spectrum. The remaining C-terminal-protected group, the ester group, does not affect the crystallization of polymer ligand–CaCO₃ composites, because the formation constant of the ester group toward the Ca²⁺ ion is lower than that of the carboxylate group. Additionally, the ester group is not able to neutralize the charge of the Ca²⁺ ion on the surface of CaCO₃ crystals.

The number-average molecular weight (M_n) of each polymer ligand was determined by GPC in THF solution (Table 1). The M_n value of the polymer ligand was obtained in the COOR form, because the polymer ligands in the COOH form associate with each other in THF solution. The obtained M_n values of these polymer ligands were 54000, 3100, and 1700 Da, for 1, 2, and 3, respectively. Interestingly, the polydispersities of the polymer ligands appear in a relatively narrow range (M_w/M_n = 1.19 to 1.48), whereas similar amino acid-containing polymers obtained by free radical polymerization have a wide range of values as reported previously.²⁹

Table 1 Number average molecular weight and tacticities of poly(carboxylate) ligands (1–3)

	M n ^a	M w/Mn^a	mm	mr + rm	rr	m	r
a Determined by GPC (THF, PS standard) with ester form of polymer. b Estimated with ester form by ^13C NMR spectra in chloroform-d at 303 K. c Estimated with COOH form by ^13C NMR spectra in Me2SO-d6 at 353 K.
1	54000	1.29	5^b	37^b	58^b	23^b	77^b
2	3100	1.48	0^c	33^c	67^c	16^c	84^c
2^iso	910	1.03	37^c	63^c	0^c	69^c	31^c
3	1700	1.19	3^c	38^c	59^c	22^c	78^c

---

Polymer tacticity was determined utilizing a ¹³C NMR measurement in Me₂SO-d₆ at 353 K as listed in Table 1. Estimation for tactic triad, that is meso–meso (mm), racemo–racemo (rr), meso–racemo (mr), or racemo–meso (rm), was performed based on quaternary carbon signals of the polymer ligand.³⁰ Generally, a stereospecific polymer is rarely obtained by a radical polymerization and only atactic or slightly syndiotactic-rich polymer was produced.³¹ In our case, the results indicate that polymer ligands 1, 2, and 3 hardly form the mm triad. The obtained values of syndiotactic diad content are 77, 84, and 78% for 1, 2, and 3, respectively. Therefore, the polymer ligands have a highly controlled syndiotactic tacticity with a narrow polydispersity.

Isotactic-rich poly[3-(methacryloylamino)propionic acid] (2^isoH) was synthesized in order to examine the effects of the polymer tacticity on CaCO₃ crystallization. Preparation of isotactic-rich polymer was performed by free radical polymerization with a Lewis acid (Scheme 1). We obtained the syndiotactic polymers (1H, 2H, and 3H) without the Lewis acid. Recently, several reports described the synthesis of stereocontrolled polymer by using various Lewis acids or amide NH-protected monomer.^32,33 Because acrylamide and methacrylamide monomers easily undergo proton-transfer polymerizations in an anionic polymerization, that is usually used in order to obtain an isotactic polymer. For example, a Lewis acid, Y(OTf)₃, was employed as a catalyst for a stereospecific radical polymerization in order to obtain an isotactic-rich polymer. Use of the AIBN–Y(OTf)₃–MeOH system for radical polymerization controls thestereoregularity, and the obtained polymer has a high isotactic content (80%) at mild temperature (333 K).³³ In the present study, 2^isoH includes isotactic content, which is about 70% on the basis of the ¹³C NMR spectral data, and has a narrow polydispersity (M_w/M_n = 1.03) (Table 1).

Formation of an NH⋯O hydrogen bond in polymer ligands

Formation of an NH⋯O hydrogen bond in the polymer side-chain was examined with simple model ligands, i.e., 3-methyl-2-(2,2,2-triphenylacetylamino)-butyrate (4), 3-(2,2,2-triphenylacetylamino)-propionate (5), and 4-(2,2,2-triphenylacetylamino)-butyrate (6) (Chart 2), for polymer-repeated units of 1, 2, and 3, respectively. These model ligands have a bulky triphenylacetylamino group to avoid polynuclear complexation and will be discussed in detail. The ¹H NMR spectra of these model ligands in chloroform-d are shown in ESI† Fig. S1a–f. In the carboxylic acid form, the amide NH ¹H NMR signals of 3-methyl-2-(2,2,2-triphenylacetylamino)-butyric acid (4H), 3-(2,2,2-triphenylacetylamino)-propionic acid (5H), and 4-(2,2,2-triphenylacetylamino)-butyric acid (6H) were observed at 6.22, 6.31, and 5.96 ppm, respectively. On the other hand, the amide NH ¹H NMR signals of tetramethylammonium carboxylate of 4 (4NMe₄), 5 (5NMe₄), and 6 (6NMe₄) were observed at 6.64, 7.07, and 6.13 ppm with downfield shifts of 0.42, 0.76, and 0.17 ppm, respectively, from those of the corresponding carboxylic acid. Thus, relatively downfield shifts of amide NH ¹H NMR signals indicate the formation of an intramolecular NH⋯O hydrogen bond.

Chart 2

The temperature coefficient was obtained by the variable-temperature (243 to 303 K) measurements of ¹H NMR spectra in chloroform-d, to confirm the intramolecular hydrogen bonding. The temperature coefficients of the amide NH signals of 4H, 4NMe₄, 5H, 5NMe₄, 6H, and 6NMe₄ are −0.37, −1.2, −2.7, −1.2, −2.2, and −1.7 ppb K⁻¹, respectively. These low coefficients of the carboxylate anions (4NMe₄, 5NMe₄, and 6NMe₄) indicate that an intramolecular NH⋯O hydrogen bond forms in each carboxylate anion. In general, the formation of a stable intramolecular NH⋯O hydrogen bond in linear peptides gives a more positive temperature coefficient than the value of −2.4 ppb K⁻¹ found in chloroform-d.³⁴ The model ligands for polymer ligands also have a more positive temperature coefficient than the value in the corresponding carboxylic acid. Thus, these models form the intramolecular NH⋯O hydrogen bond.

Intramolecular or intermolecular hydrogen bond formation was also established by a FT-IR measurement in chloroform-d (Fig. 2). The amide NH IR band of 4NMe₄ was observed at 3377 cm⁻¹, although that of 4H was observed at 3455 cm⁻¹. Similar low wavenumber shifts were found in the other carboxylate anions, 5NMe₄ and 6NMe₄, by 418 and 424 cm⁻¹, respectively. The large low-wavenumber shifts in the amide NH band between the carboxylic acid and the carboxylate anion forms indicate the formation of the intramolecular NH⋯O hydrogen bonds. The degree of shifts of the amide NH bands in the FT-IR corresponds to the strength of the NH⋯O hydrogen bond.^20,35 Therefore, 5NMe₄ and 6NMe₄ form strong NH⋯O hydrogen bonds, whereas 4NMe₄ forms a weak NH⋯O hydrogen bond. These amide NH IR bands do not shift in dilute solution (2 mM solution concentration), which excludes the possibility of formation of intermolecular NH⋯O hydrogen bonds. These results for 4, 5, and 6 in chloroform-d demonstrate that each model ligand forms an intramolecular NH⋯O hydrogen bond as shown in Chart 2.

Fig. 2 FT-IR spectra in 10 mM chloroform solution of (a) 4H, (b) 4NMe₄, (c) 5H, (d) 5NMe₄, (e) 6H and (f) 6NMe₄. The assignments of NH IR bands of 5NMe₄ and 6NMe₄ were determined from deuterium derivatives 7 and 8, deuterated 5NMe₄ and 6NMe₄, respectively (detailed synthetic procedures are shown in ESI†). The ND vibration bands for 7 and 8 are at 2199 cm⁻¹ and 2197 cm⁻¹, respectively.

In a polar solvent, such as Me₂SO-d₆, these model ligands form the intramolecular NH⋯O hydrogen bond similar to those observed in the less polar solvent, chloroform-d. The amide NH ¹H NMR signals of 4, 4NMe₄, 5, 5NMe₄, 6, and 6NMe₄ in Me₂SO-d₆ appear at 6.42, 6.94, 7.10, 8.08, 7.17, and 7.19 ppm, respectively (ESI† Fig. S1g-l). The amide NH ¹H NMR signals of these model ligands in the carboxylate anion form also show downfield shifts of 0.52, 0.98, and 0.02 ppm for 4NMe₄, 5NMe₄, and 6NMe₄, respectively, compared with those in the carboxylic acid form. Thus, these results indicate that the model ligands form the intramolecular NH⋯O hydrogen bond in both the polar and less polar solvents.

In polymer ligands, 2NMe₄ exhibits a different trend in the ¹H NMR spectra (Fig. 1) compared with the others, although all polymer ligands, 1NMe₄, 2NMe₄, 2^isoNMe₄, and 3NEt₄, are suggested to form intra-side-chain NH⋯O hydrogen bonds from the results of the model ligands. The amide NH ¹H NMR signal of 2NMe₄ appears at 7.38 ppm with upfield shifts of 0.02 ppm (Fig. 1d), while the signals of 1NMe₄, 2^isoNMe₄, and 3NEt₄ are observed with 1.50, 1.40, and 1.02 ppm downfield shifts, respectively (Fig. 1: b, f, and h). Additionally, the amide NH ¹H NMR signal of 2NMe₄ is relatively broad. This result gives the possibility of intermolecular NH⋯O hydrogen bonds in 2NMe₄. Previously, we have demonstrated that an amide NH ¹H NMR signal of intermolecular NH⋯S hydrogen-bonded tetrapeptide was broadened by formation of an intermolecular hydrogen-bonded polymeric structure.³⁶ In this case, 2NMe₄ also forms intermolecular NH⋯O hydrogen bonds. In contrast, other polymer ligands, 1NMe₄, 2^isoNMe₄, and 3NEt₄, cannot form such intermolecular interactions, because their side-chains hinder the formation of intermolecular hydrogen bonds. Thus, three polymer ligands, 1NMe₄, 2^isoNMe₄, and 3NEt₄, form intra-side-chain NH⋯O hydrogen bonds as well as these models. In contrast, syndiotactic-rich 2NMe₄ forms intermolecular NH⋯O hydrogen bonds, as shown in Fig. 3. However, isotactic-rich 2^isoNMe₄ allows the formation of intra-side-chain NH⋯O hydrogen bonds even in the solid-state. This possibility is discussed later.

Fig. 3 Schematic drawings of hydrogen-bonded polymers: (a) 2NMe₄ and (b) 2^isoNMe₄.

Observation of polymer ligands in CaCO₃ composites by ¹³C CP/MAS NMR

To observe the polymer ligands in the CaCO₃ composites, we measured the ¹³C CP/MAS spectra of polymer ligand 1–CaCO₃ composite (1CaCO₃), 2–CaCO₃ composite (2CaCO₃), 2^iso–CaCO₃ composite (2^isoCaCO₃), and 3–CaCO₃ composite (3CaCO₃) (Fig. 4) and compared these with the spectra of these poly(carboxylic acid) and poly(carboxylate) anion without CaCO₃ (Fig. 5). The ¹³C CP/MAS NMR measurement is useful to obtain chemical information of the polymer ligand in a CaCO₃ composite.^7,37,38 The chemical shift of each ¹³C signal of the carbonyl carbons was estimated by using curve-fitting analysis. The ¹³C signals of CONH and COOH of 1H are observed at 177 and 180 ppm, respectively (Fig. 5a), and those signals for 3H also appear at the same chemical shifts for 1H (Fig. 5g). The ¹³C signals of CONH and COOH of 2H are observed at 174 and 179 ppm, respectively (Fig. 5c), and those in 2^isoH appear at 175 and 178 for CONH and COOH, respectively (Fig. 5e). On the other hand, the ¹³C signals of COO⁻ of 1NMe₄, 2NMe₄, 2^isoNMe₄, and 3NEt₄ show downfield shifts by 1, 2, 2, and 2 ppm, respectively (Fig. 5: b, d, f, and h), from those of the corresponding poly(carboxylic acid)s. The ¹³C signals of CONH of 2NMe₄, 2^isoNMe₄, and 3NEt₄ also show downfield shifts of 5, 2, and 2 ppm, respectively (Fig. 5: d, f, and j). These results indicate that environments around the carbonyl carbon in the polymer ligands change between the carboxylic acid and the carboxylate anion forms.

Fig. 4 Solid-state ¹³C CP/MAS spectra of (a) 1CaCO₃, (b) decombined analysis of carbonyl region of 1CaCO₃, (c) 2CaCO₃, (d) decombined analysis of carbonyl region of 2CaCO₃, (e) 2^isoCaCO₃, (f) decombined analysis of carbonyl region of 2^isoCaCO₃, (g) 3CaCO₃ and (h) decombined analysis of carbonyl region of 3CaCO₃ (where: * is a spinning side band; # is a background peak).

Fig. 5 Solid-state ¹³C CP/MAS spectra of (a) 1H, (b) 1NMe₄, (c) 2H, (d) 2NMe₄, (e) 2^isoH, (f) 2^isoNMe₄, (g) 3H, and (h) 3NEt₄ (where: * is a spinning side band).

In the polymer ligand–CaCO₃ composites, the ¹³C NMR signals of the polymer ligands appear near the carbonate ¹³C signal (¹³CO₃²⁻) in ¹³C CP/MAS spectra (Fig. 4). The curve-fitting shows the ¹³C signals of CONH and COO⁻ of 1CaCO₃ are at 179 and 184 ppm, respectively (Fig. 4b). The ¹³C signals of CONH and COO⁻ of 2CaCO₃ appear at 175 and 181 ppm, respectively (Fig. 4d), and those of CONH and COO⁻ of 2^isoCaCO₃ are also observed at 173 and 181 ppm, respectively (Fig. 4f). The ¹³C signals of CONH and COO⁻ of 3CaCO₃ appear at 177 and 183 ppm, respectively (Fig. 4h). The observation of these ¹³C CP/MAS signals indicates that the polymer ligand in CaCO₃ composites remained after washing. No ¹³C signal of a polymer ligand should be observed in the ¹³C CP/MAS spectrum if the polymer ligand is easily removed from the CaCO₃ crystals upon washing.³⁸ Thus, these polymer ligands still bind to CaCO₃ crystals as the poly(carboxylate) anion form after the washing process.

Morphologies of poly(carboxylate) ligand–CaCO₃ composites

Morphologies of polymer ligand–CaCO₃ composites were estimated via XRD measurements (Fig. 6). It is known that the main morphologies of CaCO₃ crystal are calcite, vaterite, and aragonite.¹ Our polymer ligand–CaCO₃ composites did not contain aragonite, as was confirmed by XRD observation. The value of calcite content can be obtained from the relative intensity I_c/(I_c + I_v), where I_c is the peak intensity of the calcite (104) plane (2θ = 29.3°) and I_v is that of the vaterite (101) plane (2θ = 27.0°).³⁹ The calibration curve was obtained with variations of the ratio of calcite to vaterite content (see ESI†). The results of these measurements indicate that the calcite content levels of 1CaCO₃ and 2CaCO₃ were 59 and ∼100%, respectively. 2^isoCaCO₃ and 3CaCO₃ were mainly composed of vaterite crystals (vaterite contents were 97 and 68%, respectively). Thus, polymer ligands 1 and 2 predominantly form calcite crystals, while 2^iso and 3 produce vaterite crystals.

Fig. 6 XRD patterns of (a) 1CaCO₃, (b) 2CaCO₃, (c) 2^isoCaCO₃, (d) 3CaCO₃, and (e) calcite crystals obtained in the absence of polymer.

Fig. 7 shows the FE/SEM images of polymer ligand–CaCO₃ composites. The FE/SEM image of CaCO₃ crystals obtained in the absence of any polymer ligand is also depicted therein. It is known that a mixture of calcium chloride and ammonium carbonate without any polymer ligand results in the formation of the most stable polymorph, calcite, at room temperature.⁴ In the presence of our polymer ligands, 1CaCO₃ and 2CaCO₃ also yield calcite crystals (Fig. 7a and b). However, the crystal surface of these composites is different from calcite crystals in the absence of any ligand (Fig. 7d). The surfaces of 1CaCO₃ and 2CaCO₃ are craggy, although pure calcite crystals show a smooth surface. These differences are probably due to polymer binding to the CaCO₃ crystal. On the other hand, 2^isoCaCO₃ and 3CaCO₃ show vaterite crystals, which are rarely formed in nature. This finding also supports the polymer binding.

Fig. 7 FE/SEM images of (a) 1CaCO₃, (b) 2CaCO₃, (c) 2^isoCaCO₃, (d) 3CaCO₃, and (e) a calcite crystal in the absence of polymer (20 kV acceleration voltages). Scale bars in these images represent 3 µm for (a), (b), (c), and (d), and 600 nm for (e).

Discussion

Effects of molecular weights of polymer ligands on CaCO₃ crystallization

The number average molecular weights of polymer ligands, 1, 2, 2^iso, and 3, are different. We think that the effects of molecular weights of polymers are negligible, because the effects of the NH⋯O hydrogen bonds are enough to affect the Ca-binding abilities of polymers on the Ca-binding site, and we also think some functional groups in biopolymers regulate the Ca-binding abilities in Ca-biominerals. The solution structures of high and low molecular weight polymer ligands are probably different from each other. However, if a polymer ligand forms a random coil, this polymer probably exhibits similar behavior to that of linear, zigzag chain, or helix coil polymers that are structure-controlled polymers. In this case, hydrogen-bonded units would exist at an effective position for Ca-binding. On the other hand, we suggest that it is enough for the Ca-binding that some hydrogen-bonded units exist in the Ca-binding site. Actually, only half of the carboxylate residues affect hydroxyapatite-binding in osteocalcin, which is a Ca-binding protein in bone.⁴⁰ Therefore, the molecular weight of the polymer ligand does not need to be taken into consideration.

Intra-side-chain NH⋯O hydrogen bonds in polymer ligands on crystallization of CaCO₃ composites

The three types of polymer ligands, 1, 2^iso, and 3, form the intra-side-chain NH⋯O hydrogen bonds in their carboxylate anion forms. The formation of the NH⋯O hydrogen bond stabilizes the Ca–O bond on the surface of the CaCO₃ crystal and prevents its dissociation. These effects of the hydrogen bond on the Ca–O bond have been discussed in detail in our previous works.^19,21 NH⋯O hydrogen bonds in carboxylic acid derivatives having neighboring amide NH groups lower the pK_a value. Because of lowering the pK_a value, the carboxylate anion is hardly protonated by water under neutral conditions on the surface of the CaCO₃ composite. Actually, the intramolecular NH⋯O hydrogen-bonded benzoate derivative has a higher formation constant toward Ca²⁺ ion than the nonsubstituted benzoate without the NH⋯O hydrogen bond.²¹ Furthermore, our recent report indicates that the adhesion force toward calcium phosphate crystals with the intramolecular NH⋯O hydrogen-bonded polymer ligand, poly[1-carboxylate-2-(1-ethoxycarbonyl-3-methylsulfanylpropylcarbamoyl)-ethylene-alt-ethylene], is 5 times higher than the non-NH⋯O hydrogen-bonded poly(acrylate) ligand.⁴¹ Thus, the metal–oxygen bond is protected from hydrolysis by the intramolecular (or intra-side-chain) NH⋯O hydrogen bond.

The polymer ligands hardly dislodge from CaCO₃ crystals during the washing and still bind to CaCO₃ crystals after the washing. On the other hand, when the CaCO₃ is crystallized in the presence of three model ligands, 4, 5, and 6, the ligands are involved in the composites, and the model ligands are easily removed from CaCO₃ crystals during the washing process (¹³C CP/MAS spectra and FE/SEM images were shown in the supplemental information). Our polymer ligands clearly bind to CaCO₃ crystals. Thus, the intra-side-chain NH⋯O hydrogen bond in the polymer ligand plays an important role in the crystallization of CaCO₃.

Effects of hydrogen bond strength toward CaCO₃ polymorphs

The extract number of atoms which the intra-side-chain NH⋯O hydrogen bond ring comprises, affects to CaCO₃ polymorphs. The alignment of the intra-side-chain in the polymer also influences the polymorphs. In the presence of syndiotactic polymers, the 5-membered ring hydrogen-bonded polymer 1 makes calcite crystals, whereas the 6-membered ring hydrogen-bonded polymer, 2^iso, and the 7-membered one, 3, form vaterite crystals. These differences are due to the strength of the hydrogen bond. Polymer ligand 1 forms weak intra-side-chain NH⋯O hydrogen bonds just as its model ligand, 4; however, polymer ligands 2^iso and 3 form strong hydrogen bonds similar to the models, 5 and 6, respectively. Formation of weak NH⋯O hydrogen bonds in a polymer ligand results in the relatively low stabilization of Ca–O bonds and restriction of CaCO₃ crystal growth, and the obtained CaCO₃ crystal is the most stable polymorph, calcite, without any restriction by the ligands. Actually, poly(acrylate) ligand without the hydrogen bond forms the most stable polymorph, calcite.⁵ In contrast, strong NH⋯O hydrogen-bonded polymer ligands stabilize one of the metastable polymorphs, vaterite, which is rarely found in nature.^27,28 Thus, strong intra-side-chain NH⋯O hydrogen-bonded polymer ligands, such as 2^iso and 3, restrict CaCO₃ crystal growth to obtain vaterite crystals, although a weak intra-side-chain NH⋯O hydrogen-bonded ligand, such as 1, yields calcite crystals without any restricting crystal growth.

Relationships between polymer tacticities and morphologies of CaCO₃ composites

In the present study, the syndiotactic polymer ligand, 2, yields calcite crystals, although the isotactic-rich polymer ligand, 2^iso, forms vaterite crystals. Polymer ligand 2^iso forms strong intramolecular hydrogen bonds similar to its model, 5, because the degree of the downfield shift of the amide NH ¹H NMR signal of 2^iso from the COOH to the COO⁻ forms is similar to that of 5 in Me₂SO-d₆. However, 2 does not exhibit a similar downfield shift. The hydrogen bond in 2 is weak and formed intermolecularly. Polymer ligand 2 clearly shows a different tendency from 2^iso. Although 2 forms intermolecular NH⋯O hydrogen bonds in solution, some parts of the side-chains of 2 probably form intra-side-chain NH⋯O hydrogen bonds in the solid state, such as 2CaCO₃. Actually, 2 kept the strong bonds to CaCO₃ crystals after the washing process. This strong binding is probably due to the formation of the intra-side-chain NH⋯O hydrogen bonds.^28,38 The hydrogen bond strength in 2 is weak, as in 1. Therefore, calcite crystals are obtained in the presence of polymer 2.

In contrast, polymer 2^iso hardly forms intermolecular hydrogen bonds because of the intra-side-chain static hindrance. Thus, 2^iso strongly binds to CaCO₃ crystal in the same way as was found in 3 by the strong formation of NH⋯O hydrogen bonds, and thus restricts the crystal growth to give vaterite crystals.

Conclusion

Novel poly(methacryloylaminocarboxylate) ligands were synthesized as model ligands of the acidic peptides in CaCO₃ biominerals. These polymer ligands form 5-, 6-, or 7-membered-ring intra-side-chain or intermolecular NH⋯O hydrogen bonds between the carboxylate oxyanion and the neighboring amide NH. The strength of the NH⋯O hydrogen bonds in the polymer ligands restricts CaCO₃ morphologies. The formation of weak hydrogen bonds in the polymer ligand does not affect the CaCO₃ crystal growth; however, strong hydrogen bonds restrict crystal growth. Five-membered-ring intra-side-chain and intermolecularly NH⋯O hydrogen-bonded polymer ligand–CaCO₃ composites form calcite crystals due to the formation of the weak hydrogen bond, while 6- or 7-membered-ring intra-side-chain NH⋯O hydrogen-bonded polymer ligand–CaCO₃ composites form vaterite crystals because of the formation of the strong hydrogen bond. In conclusion, various CaCO₃ morphologies can be regulated by slightly changing the strength of the hydrogen bonds in a polymer ligand.

Acknowledgements

We acknowledge the support of this work by a Research Fellowship of the 21st Century Center Of Excellence Program “Integrated EcoChemistry” for Young Scientists [for K. Takahashi, 2002–2004] and a Grant-in-Aid for Scientific Research (A) (No. 12304040) from the Ministry of Education, Culture, Science, Sports, and Technology, Japan.

References

1. S. Mann, in Inorganic Materials—2nd Edition, ed. D. W. Bruce and D. O'Hare, John Wiley & Sons Ltd., Chichester, 1996, p. 255.
2. (a) S. Weiner, CRC Crit. Rev. Biochem., 1986, 20, 365; (b) S. Mann, D. D. Archibald, J. M. Didymus, T. Douglas, B. R. Heywood, F. C. Meldrum and N. J. Reeves, Science, 1993, 261, 1286.
3. (a) S. Mann, Struct. Bonding, 1983, 54, 125; (b) D. Jacquemain, S. G. Wolf, F. Leveiller, M. Deutsch, K. Kjaer, J. Als-Nielsen, M. Lahav and L. Leiserowitz, Angew. Chem., Int. Ed. Engl., 1992, 31, 130; (c) R. B. Frankel and S. Mann, Encyclopedia of Inorganic Chemistry, John Wiley & Sons Ltd., Chichester, 1994; (d) A. M. Belcher, X. H. Wu, R. J. Christensen, P. K. Hansma, G. D. Stuky and D. E. Morse, Nature, 1996, 381, 56; (e) J. Aizenberg, J. Hanson, T. F. Koetzle, S. Weiner and L. Addadi, J. Am. Chem. Soc., 1997, 119, 881.
4. L. Addadi and S. Weiner, Mol. Cryst. Liq. Cryst., 1986, 134, 305.
5. L. Addadi, J. Moradian, E. Shay, N. G. Maroudas and S. Weiner, Proc. Natl. Acad. Sci. USA, 1987, 84, 2732.
6. L. Addadi and S. Weiner, Angew. Chem., Int. Ed. Engl., 1992, 31, 153.
7. K. Takahashi, H. Yamamoto, A. Onoda, M. Doi, T. Inaba, M. Chiba, A. Kobayashi, T. Taguchi, T. Okamura and N. Ueyama, Chem. Commun., 2004, 996.
8. S. Albeck, J. Aizenberg, L. Addadi and S. Weiner, J. Am. Chem. Soc., 1993, 115, 11691.
9. J. Aizenberg, G. Lambert, L. Addadi and S. Weiner, Adv. Mater., 1996, 8, 222.
10. B.-A. Gotliv, L. Addadi and S. Weiner, ChemBioChem, 2003, 4, 522.
11. S. Raz, P. C. Hamilton, F. H. Wilt, S. Weiner and L. Addadi, Adv. Funct. Mater., 2003, 13, 480.
12. G. Falini, S. Albeck, S. Weiner and L. Addadi, Science, 1996, 271, 67.
13. (a) S. Mann, J. M. Didymus, N. P. Sanderson, B. R. Heywood and E. J. A. Samper, J. Chem. Soc., Faraday Trans., 1990, 86, 1873; (b) T. Kato, T. Suzuki, T. Amamiya, T. Irie and M. Komiyama, Supramol. Sci., 1998, 5, 411; (c) I. Lee, S. W. Han, H. J. Choi and K. Kim, Adv. Mater., 2001, 13, 1617; (d) L. A. Estroff, C. D. Incarvito and A. D. Hamilton, J. Am. Chem. Soc., 2004, 126, 2.
14. G. Xu, N. Yao, I. A. Aksay and J. T. Groves, J. Am. Chem. Soc., 1998, 120, 11977.
15. K. Naka, Y. Tanaka, Y. Chujo and Y. Ito, Chem. Commun., 1999, 1931.
16. J. J. J. M. Donners, R. J. M. Nolte and N. A. J. M. Sommerdijk, J. Am. Chem. Soc., 2002, 124, 9700.
17. D. Volkmer, M. Fricke, D. Vollhardt and S. Siegel, J. Chem. Soc., Dalton Trans., 2002, 4547.
18. D. Rautaray, S. R. Sainkar and M. Sastry, Chem. Mater., 2003, 15, 2809.
19. (a) A. Onoda, Y. Yamada, M. Doi, T. Okamura and N. Ueyama, Inorg. Chem., 2001, 40, 516; (b) A. Onoda, Y. Yamada, T. Okamura, M. Doi, H. Yamamoto and N. Ueyama, J. Am. Chem. Soc., 2002, 124, 1052.
20. A. Onoda, Y. Yamada, J. Takeda, Y. Nakayama, T. Okamura, M. Doi, H. Yamamoto and N. Ueyama, Bull. Chem. Soc. Jpn., 2004, 77, 321.
21. A. Onoda, Y. Yamada, Y. Nakayama, K. Takahashi, H. Adachi, T. Okamura, H. Yamamoto, N. Ueyama, D. Vyprachtiky and Y. Okamoto, Inorg. Chem., 2004, 43, 4447.
22. N. Ueyama, S. Moriyama, T. Ueno and A. Nakamura, in Peptide Science: Present and Future, Proceeding of the International Peptide Symposium, 1st, Kyoto, Nov. 30-Dec. 5, 1997, ed. S. Yasutsugu, Kluwer, Dordrecht, 1999, p. 288.
23. N. Ueyama, N. Nishikawa, Y. Yamada, T. Okamura and A. Nakamura, J. Am. Chem. Soc., 1996, 118, 12826.
24. (a) T. Ueno, Y. Kousumi, K. Yoshizawa-Kumagaya, K. Nakajima, N. Ueyama, T. Okamura and A. Nakamura, J. Am. Chem. Soc., 1998, 120, 12264; (b) T. Okamura, S. Takamizawa, N. Ueyama and A. Nakamura, Inorg. Chem., 1998, 37, 18.
25. (a) S. C. Tang, S. Koch, G. C. Papaefthymiou, S. Foner, R. B. Frankel, J. A. Ibers and R. H. Holm, J. Am. Chem. Soc., 1976, 98, 2414; (b) W. R. Scheidt, D. K. Geiger, Y. J. Lee, C. A. Reed and G. Lang, J. Am. Chem. Soc., 1985, 107, 5693.
26. N. Ueyama, H. Kozuki, M. Doi, Y. Yamada, K. Takahashi, A. Onoda, T. Okamura and H. Yamamoto, Macromolecules, 2001, 34, 2607.
27. K. Takahashi, H. Kozuki, A. Onoda, T. Okamura, H. Yamamoto and N. Ueyama, J. Inorg. Organomet. Polym., 2002, 12, 99.
28. K. Takahashi, M. Doi, A. Kobayashi, T. Taguchi, A. Onoda, T. Okamura, H. Yamamoto and N. Ueyama, J. Cryst. Growth, 2004, 263, 552.
29. F. Sanda, T. Abe and T. Endo, J. Polym. Sci., Part A: Polym. Chem., 1997, 35, 2619.
30. A. Gallardo and J. S. Román, Polymer, 1993, 34, 394.
31. (a) T. Nakano, M. Mori and Y. Okamoto, Macromolecules, 1993, 26, 867; (b) A. Matsumoto, T. Matsumura and S. Aoki, Macromolecules, 1996, 29, 423; (c) A. Matsumoto, K. Yokoi, S. Aoki, K. Tashiro, T. Kamae and M. Kobayashi, Macromolecules, 1998, 31, 2129.
32. (a) A. Dasgupta and S. Sivaram, Macromolecules, 1994, 27, 1665; (b) S. Habaue, T. Uno, H. Baraki and Y. Okamoto, Macromolecules, 2000, 33, 820; (c) C. L. Mero and N. A. Porter, J. Org. Chem., 2000, 65, 775; (d) T. Kitayama, W. Shibuya and K.-i. Katsukawa, Polym. J., 2002, 34, 405.
33. Y. Isobe, D. Fujioka, S. Habaue and Y. Okamoto, J. Am. Chem. Soc., 2001, 123, 7181.
34. E. S. Stevens, N. Sugawara, G. M. Bonora and C. Toniolo, J. Am. Chem. Soc., 1980, 102, 7048.
35. T. Okamura, K. Sakaue, N. Ueyama and A. Nakamura, Inorg. Chem., 1998, 37, 6731.
36. M. Kato, T. Okamura, H. Yamamoto and N. Ueyama, J. Inorg. Biochem., 2003, 96, 165.
37. N. Ueyama, T. Hosoi, Y. Yamada, M. Doi, T. Okamura and A. Nakamura, Macromolecules, 1998, 31, 7119.
38. K. Takahashi, M. Doi, A. Kobayashi, T. Taguchi, A. Onoda, T. Okamura, H. Yamamoto and N. Ueyama, Chem. Lett., 2004, 33, 192.
39. A. Katsifaras and N. Spanos, J. Cryst. Growth, 1999, 204, 183.
40. Q. Q. Hoang, F. Slcherl, A. J. Howard and D. S. C. Yang, Nature, 2003, 425, 977.
41. K. Takahashi, M. Doi, H. Mohri, T. Okamura, H. Yamamoto and N. Ueyama, Chem. Lett., 2004, 33, 1528.

---

Footnote

† Electronic supplemental information (ESI) available: NMR signal assignments of polymer ligands, synthetic procedures of monomers for polymer ligands (MA-Val-OPac, MA-β-Ala-OBzl, and MA-γ-Abu-O^tBu) and model ligands for the polymer repeated unit (4H, 5H, and 6H), ¹H NMR spectra of 4H, 4NMe₄, 5H, 5NMe₄, 6H, and 6NMe₄, ¹³C CP/MAS spectra of model ligand–CaCO₃ composites, FE/SEM images of model ligand–CaCO₃ composites, and the calibration curve of the ratio of calcite to vaterite content on an XRD. See http://www.rsc.org/suppdata/jm/b4/b415692g/

---