Introduction

Precise synthesis of optically active polymers is a challenging topic in polymer chemistry, because naturally occurring polymers, such as polypeptides, are formed by controlling the absolute stereogenic centers in the main chain [1,2,3,4,5]. The stereoselectivity of the main chain represents a significant foundation for three-dimensional structures with biological foundations. On this point, the synthesis of optically active polymers, which are controlled by the absolute chiral configuration of the main chain, has attracted a great deal of attention. The current goal is not only to understand the structures and properties of optically active polymers but also to develop a novel foundation for the polymeric synthesis of characteristic functional materials. Asymmetric chirogenic polymerization [6] is a promising method for synthesizing optically active polymers in which the stereogenic centers on the main chain can be controlled. Brookhart [7], Nozaki [8, 9], and Sen [10] reported the asymmetric alternating polymerization of α-olefins and CO using a Pd complex with chiral ligands. To date, several efficient asymmetric chirogenic polymerizations have been reported [11,12,13,14]. However, in most asymmetric chirogenic polymerization systems, the main chain structures of the polymer are less amenable to design. This is because chain polymerization, which is difficult to introduce in various designs, occurs on the main chain.

In the past decade, a number of asymmetric reactions have been reported [15]. If such enantioselective reactions can be applied to asymmetric polymerization and repetitive asymmetric reactions are maintained throughout the propagation step of the polymerization reaction, the synthesis of new polymers possessing various backbones and functionalities in the main chain would be possible. Thus, asymmetric reactions may facilitate the tailor-made synthesis of optically active polymers. Nevertheless, asymmetric chirogenic polymerization based on organic reactions has not been studied extensively to date. In common polymerization systems, highly enantioselective reactions cannot readily be maintained from the beginning to the end of polymerization, since it is difficult for asymmetric catalysts to recognize the reactive terminal moiety of the polymer due to steric hindrance.

Recently, we reported a new method for achieving asymmetric chirogenic polymerization based on our development of asymmetric allylic substitutions catalyzed by planar–chiral cyclopentadienyl–ruthenium (Cp’Ru) complexes [16,17,18,19,20,21]. This review focuses on our recent development of a novel synthetic approach for optically active polymers through the combination of asymmetric chirogenic polymerization and postpolymerization.

Asymmetric allylic substitutions catalyzed by Cp’Ru complexes

The catalysis of asymmetric allylic substitutions by organometallic complexes represents a powerful method for creating carbon–carbon and carbon–heteroatom bonds with high enantioselectivity [22,23,24,25]. Previously, we have shown that cyclopentadienyl–ruthenium (Cp’Ru) complexes are proficient catalysts for regio- and enantioselective allylic substitutions of monosubstituted allylic chlorides with types of nucleophile (Nu) containing carbon, oxygen, and nitrogen atoms (Fig. 1a). Therefore, since the Cp’Ru-catalyzed allylic substitutions exhibit high reactivity and selectivity, we seek to extend this system to an asymmetric chirogenic polymerization reaction via polycondensation.

Fig. 1
figure 1

a Asymmetric allylic substitutions catalyzed by the Cp’Ru complex. b Asymmetric chirogenic polymerization using asymmetric allylic substitutions catalyzed by (S)-I

Asymmetric chirogenic polymerization: asymmetric allylic substitutions catalyzed by a Cp’Ru complex

Previously, we reported the asymmetric allylic substitution with N-alkoxyamide as a nucleophile to afford carbon–nitrogen bonds with high enantioselectivity [19]. To this end, we designed an achiral AB-type bifunctional monomer (1a) bearing an allylic chloride and N-hexyloxybenzamide [26] (Fig. 1b-(a)). The polymerization of monomer 1a was carried out based on previously reported methods employing (S)-I as the catalyst. The polymerization of monomer 1a proceeded smoothly with almost quantitative monomer conversion after 36 h. Purification by column chromatography on silica gel provided the polymer poly-1a at a 70% yield (Mn = 17,000; Mw/Mn = 1.5). Poly-1a was then characterized by 1H NMR analysis and compared to the branched allylic compound 3; the analysis clearly showed that the polymerization proceeded with high regioselectivity (x/y = 98/2) and that the polymer was almost completely composed of chiral branched allylic structures (Fig. 2a). Therefore, the catalytic system employing I can be utilized to introduce regular arrangements of chiral carbons into the main chain of the resulting polymers.

Fig. 2
figure 2

Characterization of poly-1a. a 1H NMR (500 MHz, in CDCl3 at 298 K) spectra of a model compound and b poly-1a. b CD and UV spectra of poly-1a and model compound 4 (in CHCl3 at 298 K). The vertical axis is normalized to the benzene unit. c 1H NMR (400 MHz, CDCl3, 298 K) spectra of a poly-1a-(7j)-((rac)-I), b poly-1a-(7j)-((R)-I), and c poly-2a-(7j)-((S)-I)

The enantioselectivity of the asymmetric chirogenic polymerization was estimated by ultraviolet (UV) and circular dichroism (CD) spectroscopic techniques. The UV and CD spectra of poly-1a and model compound 4 (98% ee) are shown in Fig. 2b. The model compound exhibited a positive Cotton effect in the region of 240–260 nm in the CD spectrum, which is attributed to the π–π* transition of the benzene chromophore. The spectrum of poly-1a also displayed a CD spectral pattern quite similar to that of 4. The CD spectra are reflective of the local conformation of poly-1a; therefore, the new stereocenters of poly-1a exhibit the same absolute configuration as those of 4. Thus, the stereochemistry at the asymmetric carbon atoms of the polymer main chain is strictly controlled during asymmetric chirogenic polymerization.

The selectivity of the polymer was also estimated using another method. As will be described later, the polymer can be modified by chiral substitution with a 2,3,4,6-tetra-O-acylated-β-d-glucopyranose-1-thio group on the terminal olefinic moiety of the monomer unit using the thiol–ene reaction (Fig. 2c). The 1H NMR spectra of poly-1a-(7j)-((rac)-I) (Mn = 12 000; Mw/Mn = 1.4), which was synthesized from poly-1a-((rac)-I) (Mn = 11000; Mw/Mn = 1.5) polymerized using the racemic catalyst (rac)-I, exhibited two peaks at 4.57 (d, –SCHCH(OAc)–, 1 H) and 4.53 ppm (d, –SCHCH(OAc)–, 1 H) with a ratio of 1:1 in the region representing the methine proton on the C1 carbon of the pyranoside according to the chirality of the main chain (Fig. 2c-(a)). In contrast, poly-1a-(7j)-((S)-I) and poly-1a-(7j)-((R)-I), which were polymerized using (S)-I and (R)-I, respectively, each exhibited only one peak in the corresponding region (at 4.53 or 4.57 ppm), indicating that the stereocenters of the polymer were strictly controlled during asymmetric chirogenic polymerization when an optically active catalyst was employed (Fig. 2c-(b), (c)). These results show that the planar–chiral Cp’Ru complex functions as a highly efficient catalyst for stereochemical control even in polymerization reactions.

The asymmetric polymerization reaction developed herein is applicable to a variety of monomers bearing different arylene backbones (1bd) (Fig. 1b-(a)). All of these reactions also proceeded with high regio- and enantioselectivity to afford the desired polymer (poly-1bd).

The polymerization mechanism is shown in Fig. 3; the reaction involves the oxidative addition of an allylic moiety to the (S)-I complex, affording a π-allyl intermediate with high diastereoselectivity at the chiral metal center and planar chirality at the π-allyl ligand. Subsequent inside attack of the nucleophile via the Ru complex and substitution of the chloride with the nucleophile affords a branched allylic compound with high enantioselectivity. Thus, it is likely that the asymmetric chirogenic polymerization proceeds through the same pathway described above with control over the metal-center chirality during polymerization, making control of the chirality of the metal-center essential for the success of the polymerization.

Fig. 3
figure 3

Plausible reaction mechanism of asymmetric chirogenic polymerization using (S)-I

Enantio- and diastereoselective polymerization: asymmetric allylic alkylation catalyzed by a Cp’Ru complex

Asymmetric allylic alkylation is one of the most promising methods for the construction of carbon–carbon bonds. Therefore, the application of these reaction systems for the synthesis of optically active polymers is of great interest. However, when the aforementioned asymmetric chirogenic polymerization reaction is applied to carbon nucleophiles, the carbon nucleophile exhibits a prochiral difference from a nitrogen nucleophile; therefore, it is necessary to control the two vicinal stereocenters in the monomer unit. However, the control of vicinal stereocenters through allylic substitutions in organic synthesis is challenging [27,28,29,30,31,32,33]. Previously, asymmetric allylic alkylation catalyzed by I was reported [16]. The reaction proceeded with quantitative conversion under mild conditions yielding vicinal stereocenters with regio-, diastereo-, and enantioselectivity. The asymmetric alkylation reaction was attempted for asymmetric chirogenic polymerization [34].

An achiral AB-type bifunctional monomer 2 bearing allylic chloride and prochiral β-diketone moieties was designed and synthesized (Fig. 1b-(b)). The polymerization was conducted under optimized conditions, proceeding with almost quantitative monomer conversion to afford the desired polymer. After purification, poly-2 was obtained in 75% yield (Mn = 7000, Mw/Mn= 2.6). The chemical structure of poly-2 was confirmed by NMR analysis, which indicated that the polymerization of 2 catalyzed by (S)-I proceeded with high regioselectivity. Additionally, the diastereoselectivity was estimated by comparison to the model compound, which gave a diastereo-ratio (dr) = 8:1. Therefore, the polymerization of monomer 2 proceeded with high diastereoselectivity, greater than that of the model reaction (dr = 4:1). The enantioselectivity of the resulting polymer was estimated using UV and CD spectrometry, which were assigned by density functional theory (DFT) and time-dependent (TD)-DFT. The results indicated that the polymerization proceeded in a highly enantioselective manner to afford an optically active polymer.

Synthesis of optically active polymer using postpolymerization

In our asymmetric polymerization method, the spacer group and the side chain of the N-alkoxide group can be modified. Additionally, the resulting optically active polymers retain one reactive terminal double bond per monomeric unit, which can be used for further transformations. Recently, the postpolymerization modification method for preparing new functional polymers has become a powerful tool in polymer synthesis [35,36,37,38,39,40,41]. This method can be used to obtain a series of polymers from a single parent polymer. Therefore, upon the postpolymerization conversion of terminal olefins, our polymerization methods are expected to expand the versatility of optically active polymer synthesis. Herein, new methods for the synthesis of optically active polymers through the combination of asymmetric polymerization and postpolymerization modification are described.

Synthesis of optically active polymer bearing chiral cyclic architecture

Natural polymers, such as nucleic acids, starch, and cellulose, possess chiral cyclic structures in their main chains. Such cyclic structures restrict the conformation of the main chain bonds to form the appropriate chiral foundation. Therefore, intelligently designing the main chain to restrict the local conformation, as well as to control the resulting stereocenters, plays an important role in the continued development of artificial chiral foundations, although it is quite difficult to synthesize optically active polymers bearing such chiral cyclic architecture in the main chain.

Previously, we focused on the ring-closing metathesis (RCM) reaction as a candidate for the transformation of terminal olefins. This reaction is a powerful tool in organic synthesis for the preparation of cyclic compounds [42, 43]. We accomplished the synthesis of optically active γ- and δ-lactones using a combination of asymmetric allylic carboxylation and RCM reaction [21]. Based on these results, we designed a new synthetic approach for the synthesis of optically active polymers bearing chiral cyclic architecture in the main chain using a combination of asymmetric allylic amidation and RCM reactions [44].

An achiral AB-type monomer (5a) bearing an allylic chloride and an N-alkoxyamide group possessing an olefinic moiety was designed. Upon polymerization, the resulting polymer presents two olefinic moieties per monomer unit for postpolymerization RCM. The asymmetric chirogenic polymerization catalyzed by (S)-I was carried out under optimized conditions. After 36 h, 5a was quantitatively converted to afford the corresponding polymer poly-5a in 80% yield, maintaining the unreacted terminal olefin on the N-alkoxide moiety (Table 1a). The resulting polymer was analyzed by 1H NMR, CD, and UV spectroscopic methods, which indicated that the polymerization proceeded in a highly regio- and enantioselective manner to afford an optically active polymer. The absolute molecular mass of the resulting polymer was determined by a Viscotek TDA 305 triple detection system. Additionally, monomers bearing 1,3-phenylene (5b), 2,6-naphthylene (5c and 5d), or 2,7-naphthylene (5e) as a spacer group (Ar) were also polymerized with quantitative monomer conversion and high regioselectivities.

Table 1 (A) Asymmetric chirogenic polymerization using asymmetric allylic substitution of monomer 5 catalyzed by (S)-Ia. (B) Ring-closing metathesis reaction of poly-5b

The resulting polymer (poly-5a) underwent RCM using a Hoveyda–Grubbs II catalyst under optimized conditions (Table 1b). The olefinic components of poly-5a were fully consumed through the RCM. After purification to remove the catalyst, the polymer poly-6a (Mn = 17000; Mw/Mn = 1.5) was obtained in a 60% yield. Since the degree of polymerization (n) calculated by the absolute molecular mass was not affected after RCM (poly-5a: n = 63; poly-6a: n = 61), the evidence indicated that the intermolecular reaction of the polymer dominated. Additionally, the postpolymerization conversion is applicable to various monomer residues, potentially affording optically active polymers possessing several types of main chain and side chain structures.

Postpolymerization modification of the side chain in optically active polymers by thiol–ene reaction

The thiol–ene reaction is one of the most reliable methods for postpolymerization modification and has facilitated the functionalization of unsaturated aliphatic hydrocarbons of polymer side chains with complete chemo- and regioselectivity [45,46,47,48,49,50,51,52]. Additionally, the thiol–ene reaction can tolerate many functional groups under mild conditions. These advantages suit our envisaged synthetic approach [53].

The thiol–ene reaction of poly-1a (Mn = 16 000; Mw/Mn = 1.4) with 1-dodecanethiol (7a) was conducted in the presence of 2,2-dimethoxy-2-phenylacetophenone (DMPA) as a photoinitiator (entry 1 in Table 2). The reaction mixture was irradiated at room temperature using a 500 W high-pressure mercury lamp (centered at 350 nm). The reaction proceeded with quantitative conversion to give poly-1a-(7a) (Mn = 27000; n = 58), which is appended with a dodecylthio group at the terminal olefin moiety of poly-1. After purification, the functionalized polymer was obtained in 90% isolated yield. These postpolymerization modification conditions can be applied with various thiols (7bl) (entries 2–12). All of the tested reactions proceeded with almost quantitative conversion regardless of the polarity or bulkiness of the substituents, and the solubility of each resulting optically active polymer poly-1a-(7) can be altered depending on the substituent that is introduced.

Table 2 Modification of the side chain in poly-1 by the thiol–ene reactiona

Synthesis of helical poly-arylopeptide through reductive cleavage of N–O bond

Chiral helical polymers have received a great deal of attention in polymer chemistry because naturally occurring polymers often form sophisticated helical structures. In artificial polymers, the precise synthesis of helical polymers has also been reported for the preparation of potential functional materials [54,55,56]. Although we have explained the synthesis of various types of optically active polymers based on asymmetric chirogenic polymerization, the poly-N-alkoxyamide polymers (poly-1 and poly-1-(7)), unlike proteins, do not form stable secondary structures in solution even though the asymmetric centers of the monomer unit are strictly controlled. This is because the rotational barrier of the N–C bond of the N-alkoxyamide moiety is relatively low compared to that of peptides. If the flexible alkoxyamide bonds of the polymer main chain could be converted to planar amide bonds, one would expect the creation of a secondary structure (Fig. 4a). Thus, a new synthetic approach to nonnatural helical polypeptides using our previously presented asymmetric polymerization method and N–O bond reduction is presented here (Fig. 4) [57].

Fig. 4
figure 4

Novel synthetic approach to poly-arylopeptides using reductive cleavage of N–O bond of poly-1

In organic synthesis, the N-alkoxyamide group is used as a precursor to the amide moiety; they can be transformed without racemization by the reductive cleavage of the N–O bonds [58,59,60,61,62,63]. Through optimization studies, samarium diiodide was chosen as the N–O bond reducing reagent [62, 63]. The treatment of poly-1a-(7a) (Mw = 51000; Mw/Mn = 2.5; n = 110) with SmI2 in THF resulted in quantitative conversion of the N-alkoxyamide moieties into amide groups without racemization (Fig. 4b). The resulting polymer, poly-8a-(7a) (Mw = 39000; Mw/Mn = 2.6), was completely characterized by 1H NMR and IR spectroscopy as well as size exclusion chromatography. This is the application of the SmI2 complex in postpolymerization modification. To distinguish the resulting polyamides from simple polyamides, poly-8-(7) was classified as a poly-arylopeptide. This designation was used because the monomer units of the polymer were composed of amino acids containing arylene spacers, and the polymer directionality of the main chain mimics that of natural polypeptides. This synthetic method was applied to various types of poly-1a-(7) polymers with different side chains, including oligoetherthio and N-(tert-butoxycarbonyl)aminoethylthio substituents.

The polymer conformation of poly-8a-(7a) was investigated by UV and CD spectroscopic methods. The CD spectrum of poly-8a-(7a) showed a large bisignate Cotton effect that differed from that of the corresponding monomer, 9 (n = 1) (Fig. 5a). As additional chirality was present, derived from the asymmetric carbons of the main chain, a one-handed helical structure was suggested for poly-8a-(7a). In the case of dimer 10 (n = 2), a new bisignate Cotton effect was observed, similar to that of poly-8a-(7a), which was shifted to a longer wavelength as the monomer unit was elongated (from n = 2 to n = 4 (11)) (Fig. 5b). To reveal the origin of the Cotton effect of poly-8a-(7a), DFT and TD-DFT calculations were carried out for 10 based on the crystal structure of the model compound [64]. In this case, since the carbonyl group and benzene rings are generally regarded as coplanar, two possible conformations based on the optimized structure of 9 can be assumed according to the orientation of the vicinal unit: the turn (10turn) and zigzag (10zig) types. The calculated UV and CD spectra of 10turn were in close agreement with the observed spectra of 10 (Fig. 5c, d). The results indicated that longer oligomer lengths form a stable turn-type conformation. The proposed structure of poly-8a-(7a) was constructed based on the optimized structure of 10turn, which exhibited a right-handed (P) 31-helical structure (Fig. 5e).

Fig. 5
figure 5

CD and UV spectra of a poly-8a-(7a) and 9 in THF at 25 °C (The vertical axis is normalized on the basis of the absorption at 236 nm) and b 9-11 at THF at 25 °C. Simulated CD and UV spectra of c 10zig, and d 10turn by TD-DFT calculation (B3LYP/6-31**). e Plausible helical conformation of poly-8a′ (n = 12, R = Et) based on the conformation of 10turn

Side chain-driven dual structural system of poly-arylopeptide; selective helical formation derived from aromatic ring flips on the backbone

To develop a dual structural architecture, we used 2,6-naphthalene rings as the axially unsymmetrical spacer in the main chain [65]. Aromatic-ring flips give the syn and anti geometrical isomers of the 2,6-naphthalene spacer where, in solution, the energy barrier between the isomers is negligible. In this work, the minute energy difference between the two isomers is amplified by incorporating 2,6-naphthalene spacers into the polypeptide backbones, which creates a thermodynamic driving force for the formation of two distinct secondary structures, (i.e., 31- and 41-helices) biased toward one geometrical isomer depending on the side chain (Fig. 6a).

Fig. 6
figure 6

a Geometrical properties of 2,6-naphthylene amide and this concept. b CD spectra of poly-8b-(7a) and poly-8b-(7c) in THF 25 °C. Concentration: [poly-8b-(7a)] = 0.32 mM and [poly-8b-(7c)] = 0.30 mM. c Two plausible helical conformations of poly-8b (n = 15) based on the syn and anti results. The upper structure corresponds to poly-8b-(7a), and the lower structure corresponds to poly-8b-(7c). Side chains are omitted for clarity

The CD spectra of poly-8c-(7a) and poly-8c-(7c) in THF at 25 °C revealed very different Cotton effects (Fig. 6b). The CD spectrum of poly-8c-(7a) showed a plus-to-minus bisignate Cotton effect between 252 and 231 nm, which was similar to that of poly-8a-(7a). Conversely, poly-8c-(7c) presented a CD spectrum with a minus-to-plus bisignate curve exhibiting a negative Cotton effect at 257 nm and a positive Cotton effect at 238 nm. These results indicated that poly-8c-(7a) and poly-8c-(7c) adopted different specific structures in solution, whereas poly-8a-(7), which contains p-phenylene spacers with no geometrical isomers, formed a stable 31-helix regardless of the side chain. Based on the DFT and TD-DFT calculations, it was found that the dodecyl side chain induced successive syn units (syn, syn, syn, etc), while the oligoether group induced successive anti units (anti, anti, anti, etc), leading to the proposal of two global conformations (Fig. 6c). Poly-8c-(7a) exhibited a right-handed 41-helix due to the successive formation of syn units, while poly-8c-(7c) formed a right-handed 31-helix in THF. This geometrical control of the local structure is due to the relative thermodynamic stability of the polymers.

Conclusions

This review summarizes a new method for achieving asymmetric chirogenic polymerization based on asymmetric allylic substitutions catalyzed by a planar–chiral Cp’Ru complex. The presented polymerizations proceed with high regio- and enantioselectivity. The presence of asymmetric carbons in the main chain of the resulting polymers was precisely controlled. These results indicated that the planar–chiral Cp’Ru complex functions as a highly efficient catalyst to control stereochemistry, even in polymerization reactions. Additionally, the resulting polymer exhibited terminal double bonds at each monomer unit in the main chain, which can be easily modified using RCM and thiol–ene reactions as a postpolymerization method to enable the quantitative conversion of the main chain or modification of the side chains. Furthermore, we applied our polymerization system to the synthesis of poly-arylopeptides, which involved the use of asymmetric chirogenic polymerization and reductive cleavage of the N–O bonds in the N-alkoxyamide of the main chain. This method does not require enantiomerically pure amino acids; the use of a condensation reaction and stepwise synthetic chemistry affords a long peptide chain with controlled asymmetric centers.

Because these polymerization reactions are polycondensation reactions, optically active polymers possessing various main chain structures can be synthesized by polymerizing monomers exhibiting various main chain skeletons. In addition, by incorporating postpolymerization conversion, the flexible design of three-dimensional structures based on the precise tailored synthesis of the primary structures is expected. These studies are based on our organic and organometallic chemistry investigations and our understanding of the reaction mechanisms and reactive species involved and their application to the polymerization reactions. From this work, the future development of asymmetric chirogenic polymerization reactions based on new catalyst design and the synthesis of optically active polymers with unique physical properties are expected.