Functionalizable Stereocontrolled Cyclopolyethers by Ring‐Closing Metathesis as Natural Polymer Mimics

Abstract Whereas complex stereoregular cyclic architectures are commonplace in biomacromolecules, they remain rare in synthetic polymer chemistry, thus limiting the potential to develop synthetic mimics or advanced materials for biomedical applications. Herein we disclose the formation of a stereocontrolled 1,4‐linked six‐membered cyclopolyether prepared by ring‐closing metathesis (RCM). Ru‐mediated RCM, with careful control of the catalyst, concentration, and temperature, selectively affords the six‐membered‐ring cyclopolymer. Under optimized reaction conditions, no metathetical degradation, macrocycle formation, or cross‐linking was observed. Post‐polymerization modification by dihydroxylation afforded a novel polymer family encompassing a poly(ethylene glycol) backbone and sugar‐like functionalities (“PEGose”). This strategy also paves the way for using RCM as an efficient method to synthesize other stereocontrolled cyclopolymers.

Control over the absolute configuration of as ynthetic polymer main chain remains as ignificant challenge, [1,2] especially in light of the importance of this regularity in natural polymers. [3] This stereoregularity of the main chain plays as ignificant role in shaping the three-dimensional structure,a nd in-turn influencing the biological function of the natural macromolecules.Inaddition, rings are embedded in the backbones of many natural polymers,w hich restricts the bond rotation around the stereogenic centers. [4] These local conformational restrictions result in as pecific compact structure along the polymer backbone:t hat is,t he sixmembered cyclic structures of cellulose and amylose backbones ( Figure 1) form linear and helical structures,r espectively.H owever,s ynthetic polymers that are made of similar 1,4-linked six-membered rings that would mimic these secondary structures present au nique challenge to synthetic chemists,e specially if conventional polymerization techniques are employed. [5] Ring-closing metathesis (RCM) has been predominantly used to prepare small cyclic molecules,i ncluding phosphineboranes,s ulfides,a mines,p henols,a nd oxazolines. [6] Theu se of RCM in polymer synthesis,h owever,r emains rare, [7] with only af ew examples of the preparation of polymeric nanoparticles, [8] cyclic polymers, [9] and cyclopolymers. [10] We hypothesized that this under-utilized post-polymerization technique could be employed for the synthesis of cyclopolyethers to mimic the topology of polysaccharides,w here the configuration of all the stereogenic centers present in the polymers is controlled (Figure 1). While the broad functional group tolerance of metathesis catalysts suggests ab road reaction scope,p oly(ethylene glycol) backbones are especially interesting given their role as gold standard stealth polymers in drug delivery. [11] We thus envisaged asequence of ring-opening polymerization (ROP), ring-closing metathesis (RCM), and dihydroxylation (DH), as shown in Figure 1. An initial ROPo f3 ,4-epoxy-1-butene (EB) would afford polyepoxybutene (PEB), with the chirality of the parent epoxide leading to stereogenic control in the linear polymer.R ingclosing metathesis would then give a1 ,4-linked functionalizable cyclopolyether (FCPE). Further functionalization, specifically diatereoselective dihydroxylation (DH), would produce an ew stereocontrolled polymer that we have called "PEGose", as it has the structural features of both sugars and PEG,with the glycosidic bond of amylose replaced by astrong ether link.
Fort his structural control, it is imperative to start with enantiopure EB.A tactic PEB would lead to an on-stereocontrolled cyclopolymer, where the 1,4-links would be indiscriminately cis (amylose-like) or trans (cellulose-like; Figure 2). Furthermore,the subsequent dihydroxylation reaction would not be diastereoselective on the trans 1,4disubstituted six-membered rings.O nt he other hand, dihydroxylation of the cis-cyclopolymer will occur exclusively from the top face,w hich is not hindered by the two substituents.
Ring-opening and ring-closing reactions were first explored with the commercially available racemic 3,4epoxy-1-butene monomer.Although PEB has been prepared by an umber of routes from the EB monomer, [12,13] our optimized reaction conditions used tetraphenylporphyrin aluminum chloride [(TPP)AlCl] as an initiator [14] for the bulk polymerization of EB at ambient temperature [Eq. (1), see Table S1 in the Supporting Information].
The 13 CNMR spectrum of the resultant PEB showed two signals for the stereogenic carbon atom, thus confirming the expected atacticity of the produced PEB (a-PEB), as the catalyst is not stereoselective.C ontrolled low-molecularweight polymers were produced, with M n,GPC values of 2100 and 3200, respectively,and < 1.2. Although RCM on a-PEB will not give astereocontrolled cyclopolyether,itwas used to establish the optimum RCM conditions.A lthough little difference in molecular weight was observed by gel-permeation chromatography (GPC) for the lowest molecular weight PEB samples,b ecause of overlap with eluent peaks,t he higher molecular weight samples (M n,GPC 3200) showed ac lear loss of molecular weight on ring closing,c orrelating well with the loss of asingle ethylene molecule per repeat unit (Table 1). Optimal ring-closing conditions for PEB included the use of 5mol %o ft he second-generation Hoveyda-Grubbs (HG2) catalyst at high concentrations of the polymer (! 0.2 m with respect to the monomer unit) in 1,2-dichloroethane (1,2-DCE;T able 1). Note that poor conversions were achieved with the first-generation Grubbs catalyst, likely because of the lower reactivity and thermal stability (see Table S2). [15] When concentrations were kept at 0.2 m ( Table 1, entry 1), no cross-linking was observed, and the polymer dispersity and viscosity remained similar. However,athigher concentrations (0.4 m), competing ring-closing and polymer cross-linking occurs,a se videnced by the formation of ah igh-molecularweight shoulder in the GPC traces (see Figure S1). [16] This new polymer represents the first synthetic cyclopolyether prepared, but its inherent atacticity prevents any overall topological control. Thus,i sotactic-rich PEB (i-PEB) was synthesized by ROPo ft he R enantiomer of EB (95:5 e.r.), which was prepared from racemic EB by Jacobsensh ydrolytic kinetic resolution (see page 3i nt he Supporting Information). [17] Theb ehavior of the isotactic polymer towards RCM was directly compared to that of its atactic derivative under the previously optimized conditions ( Table 2).
Thek inetics of the cyclization of the enantiomerically pure monomer were significantly slower than for the racemic monomer.P lotting the reaction kinetics (Figure 3a nd Figure S2) showed that the cyclization reaction progressed quickly in the beginning, with 94 %o fthe pendent vinyl groups forming cross-links within 30 min for both the atactic and isotactic derivatives.T he metathesis reaction then significantly slowed down, especially for the more conforma-  tionally rigid isotactic derivative.This profile suggests amechanism originally proposed by Coates and Grubbs: [10] 1) afast stage when the catalyst randomly closes adjacent olefins until only isolated olefins remain, and 2) as low stage when the rings rearrange along the chain until all the olefins are cyclized. This requires that the cyclized olefins can undergo further metathetic reactions,t hus enabling ar eopening and exchange of the product rings. [18] Thet wo-stage reactivity is showcased through the RCM optimization, with the first stage completed in as imilarly short time regardless of the solvent used (1,2-DCE, DCM, THF,C HCl 3 )o rc atalyst loading (2-5%;s ee Tables S3-S6). An illustration of this mechanism can be observed in the 13 CNMR spectra of both the atactic and isotactic FCPE and PEB (Figure 4), which demonstrate that greater than 99 %of the olefins of PEB are cyclized ( Figure 4B,E). In the atactic FCPE ( Figure 4B), the new olefin peaks appear as broad, overlapping resonances (d = 125-132 ppm), which reflects the different ring configurations along the polymer backbone.On the other hand, i-FCPE showed only two sharp olefin resonances ( Figure 4E), thereby confirming the stereocontrolled structure of the polymer (Figure 2, cis-cyclopolymer). However,t he spectrum of i-FCPE after 94 %c onversion ( Figure 4D), which was taken after 30 min, showed the 6%of uncyclized isolated olefin signals (d = 118.3 and 135.6 ppm) and several cyclic olefin signals (d = 125-132 ppm).
Most of these cyclic olefin peaks were not observed at the end of the reaction (after 7days), which purports that in the initial metathesis stage when the catalyst randomly closes the olefins,d ifferent ring sizes were formed ( Figure 5). In the subsequent slow stage,t he rings rearrange along the chain until only the most thermodynamically stable six-membered rings are present.
Fixing the free rotation of the pendent olefins through RCM has an impact on the glass transition temperature (T g ) in both a-and i-PEB.T he organized structure of i-FCPE has as ignificantly higher T g value than a-FCPE (À11 8 8Cf rom À26 8 8C). This is consistent with the presence of cycles hindering segmental chain mobility in both structures.
To prepare the PEGose polymer, i-FCPE was dihydroxylated under mild conditions using N-methylmorpholine Noxide (NMO) and OsO 4 as ac atalyst. This second postpolymerization functionalization was diastereoselective,a s OsO 4 attacks on the less hindered side of the ring (Figure 2), as demonstrated by 13 CNMR spectroscopy ( Figure 6). The dihydroxylation produces aunique stereocontrolled polymer structure,w ith ah ydrophilic surface (cis-diols) opposite the hydrophobic backbone.T his distinctive structure could have potential applications in biomaterials,b lood storage,o rd rug    delivery,with face polarity shaping the surface chemistry and self-assembly. [19,20] While amylose,( C 6 H 10 O 5 ) n ,a nd PEGose, (C 6 H 10 O 4 ) n have similar monomer units,PEGose is connected with an additional methylene bridge,w hich gives more flexibility to the polymer backbone.C ircular dichroism (CD) was used to determine the influence of this CH 2 unit on the secondary structure of the PEGose.I ndeed, PEGose and amylose have the same prominent negative bands at l = 182 nm ( Figure S25), which shows that this new PEGose has an extended pseudohelical structure similar to amylose. [21] Efforts to gain complementary X-ray characterization of this self-assembly are ongoing. Although excess NMO affords complete dihydroxylation of the double bonds,t he reaction also offers the ability to adjust the polymer polarity by limiting this co-oxidizing reagent. Reducing the NMO loading from 1.1 to 0.8 equivalents dramatically alters the polarity and solubility of the resultant polymer (Table 3), thereby offering as econdary tuning for biomedical applications and leaving sites remaining for further functionalization or drug conjugation. [22] Whereas the parent polymer is soluble in organic solvents and the fully dihydroxylated polymer is freely soluble in water and DMSO, this strategy allows for abroad range of polymer polarities to be accessed.
In conclusion, we have shown that RCM of linear, stereoregular polymers with pendent olefins can be used to prepare cyclopolymers with excellent control over the ring size.F urther functionalization of the latent olefin groups by dihydroxylation provides sugar-like structures with ap oly-(ethylene glycol) backbone that leads to an ew PEGose architecture.T he isotactic linear PEB leads,a fter RCM, to ac yclic polymer with well-defined cis substitution patterns. By taking advantage of the diastereoselectivity of the subsequent dihydroxylation reaction, we were able to create acyclopolymer where the configuration of all the stereogenic centers is controlled, and which mimics the natural amylose. This new platform offers significant potential for future functionalization, drug conjugation, and biomedical mimicry, and is as ignificant focus of our future work, as is expanding this idea to other polymer backbones. [a] Determined by 1 HNMR spectroscopy through integration of the olefin signals relative to the polymer signals.