Degradable Polyphosphoramidate via Ring-Opening Metathesis Polymerization

Abstract

We report the synthesis of a degradable polyphosphoramidate via ring-opening metathesis polymerization (ROMP) with the Grubbs initiator (IMesH₂)(C₅H₅N)₂(Cl)₂Ru═CHPh. Controlled ROMP of a low ring strain diazaphosphepine-based cyclic olefin was achieved at low temperatures to afford well-defined polymers that readily undergo degradation in acidic conditions via the cleavage of the acid-labile phosphoramidate linkages. The diazaphosphepine monomer was compatible in random and block copolymerizations with phenyl and oligo(ethylene glycol) bearing norbornenes. This approach introduced partial or complete degradability into the polymer backbones. With this chemistry, we accessed amphiphilic poly(diazaphosphepine-norbornene) copolymers that could be used to prepare micellar nanoparticles.

Graphical Abstract

Degradable polymers are attractive in a host of applications, including drug delivery, tissue engineering, and fabricating electronic devices and recyclable materials.^1–6 Traditionally, ring-opening polymerization (ROP) of cyclic monomers, such as cyclic ketene acetals, lactides, and lactones, is harnessed to synthesize well-defined degradable polyesters.^7–11 Recently, polymers containing a variety of hydrolytically and redox degradable moieties have been prepared via ruthenium-based metathesis polymerizations, including acyclic diene metathesis (ADMET) polymerization,^12,13 cascade enyne metathesis polymerization,^14,15,16 and ring-opening metathesis polymerization (ROMP).

ROMP is known as a powerful tool for the synthesis of polymers with predictable molecular weights, narrow molecular weight distributions, and complex architectures.^17–22 In particular, the development of well-defined ruthenium carbene initiators has enabled efficient polymerization with excellent functional group tolerance.^23–25 Despite the tremendous diversity of functional nondegradable polymers accessed via ROMP,^17–25 examples of well-defined high molecular weight polymers consisting of fully degradable backbones remain rare. In 2013, Kiessling and co-workers demonstrated the controlled ROMP of bicyclic oxazinones, resulting in polymers that degrade under both acidic and basic conditions.²⁶ Recently, Johnson and co-workers reported the statistical copolymerization of silyl ether-based cyclic olefins with norbornene-based macromonomers via ROMP to generate backbone degradable bottlebrush and star copolymers.²⁷ Other monomers, including dioxepins,²⁸ cyclic phosphates,²⁹ disulfide containing cyclic olefin,³⁰ levoglucosenol,³¹ cyclic carbonate,³² and cyclic enol ethers,³³ have been reported in the preparation of degradable polymers via ROMP. However, in these cases, some fundamental limitations were observed, including poor control over the polymer molecular weight and dispersity, as well as being limited to producing high molecular weight fragments upon the degradation of copolymers consisting of both degradable and nondegradable monomer units.

In 2012, Kilbinger and co-workers reported the synthesis of a diazaphosphepine-based cyclic olefin, 2-phenoxy-1,3,4,7-tetrahydro-1,3,2-diazaphosphepine 2-oxide (PTDO), which was used as a sacrificial agent for amine end-functionalization of living ROMP polymers.³⁴ The researchers claimed the formation of an unreactive ruthenium carbene post PTDO incorporation. We envisioned that PTDO should be a good candidate for the preparation of fully degradable polymers via ROMP under the appropriate conditions. It is well-known that the release of ring strain of cyclic monomers serves as the main driving force in ROMP, compensating for entropy loss during polymerization.²⁴ However, an evaluation of the ring strain energy of PTDO via a density functional theory calculation revealed a low ring strain of 10.86 kcal/mol (Figure S1). This number is markedly lower than the strain energy of norbornene (27.2 kcal/mol), the most widely used monomer class employed for ROMP.^35,36 In the case of low ring strain monomers such as cyclopentene, one approach to achieve high monomer conversion is to lower the reaction temperature.^25,37–40 Therefore, we hypothesized that ROMP of PTDO would be achieved at low reaction temperatures to afford polymers with controlled molecular weight and low dispersity. Furthermore, upon hydrolysis, the polymer would degrade into a phosphoric acid and 1,4-diamino-2-butene, which is an unsaturated analog of biogenic amines involved in cell growth and differentiation, leading to a potentially biocompatible material.⁴¹

To start the investigation, the seven membered cyclic monomer PTDO was prepared via a two-step synthesis with an overall yield of 44% (Figure 1).^34,42 The monomer was then subjected to the Grubbs initiator (IMesH₂)-(C₅H₅N)₂(Cl)₂Ru═CHPh (I) at different reaction temperatures (Table 1). This initiator was chosen due to its fast initiation rate and high reactivity toward less active olefins.^25,43 Initial screening was performed with 0.3 M PTDO (Table 1, entries 1 and 2). Despite the nearly 50% monomer conversion at 20 °C, we only observed the formation of oligomers with a broad dispersity of 1.62. A secondary peak at 25.8 min on size-exclusion chromatography (SEC) indicated the presence of side reactions such as adverse backbiting and early termination (Figure S2). By decreasing the reaction temperature to 2 °C, a high molecular weight polymer with narrow dispersity was obtained (M_n = 19.3 kDa, Đ = 1.10). An increase in initial monomer concentration to 0.5 M did not lead to changes in M_n or Đ, but significantly reduced side reactions (Figure S2). As such, ROMP with 0.5 M PTDO at 2 °C proved to be optimal, resulting in 55% PTDO conversion in 5 h (M_n = 14.2 kDa, Đ = 1.16, Table 1, entry 5). Further decreases in reaction temperature led to a reduction in monomer conversion (Table 1, entries 6 and 7). At −20 °C, more than 95% of PTDO was retrieved after quenching the reaction with ethyl vinyl ether (EVE), indicating a failure to initiate.²⁵ Polymerizations with the bromopyridine-modified Grubbs initiator⁴³ (IMesH₂)-(C₅H₄NBr)₂(Cl)₂Ru═CHPh were also performed, affording a similar trend with respect to the effect of temperature on the polymerization (Table S1 and Figure S3).

Figure 1.

Synthesis and ring-opening metathesis polymerization of 2-phenoxy-1,3,4,7-tetrahydro-1,3,2-diazaphosphepine 2-oxide (PTDO) to afford polymers bearing phosphoramidate linkages.

Table 1.

ROMP of PTDO at Varying Temperatures^a

entry	T (°C)	[M]0 (M)	conv.b (%)	Mn,theoc (kDa)	Mn,MALSd (kDa)	Đ d
1	20	0.3	48	10.7	5.1	1.62
2	2	0.3	58	13.1	19.3	1.10
3	20	0.5	42	9.4	7.3	1.26
4	10	0.5	47	10.5	14.6	1.19
5	2	0.5	55	12.4	14.2	1.16
6	−5	0.5	44	9.9	12.6	1.16
7	−20	0.5	<5	N/A	N/A	N/A

^aROMP was performed in 10/90 v/v MeOH in DCM under N₂ for 5 h and quenched with ethyl vinyl ether. A feed ratio of 100:1 [M]₀:[I] was used for all reactions.

^bMonomer conversion was determined by ¹H NMR spectroscopy.

^cTheoretical molecular weight M_n,theo = DP_monomer × MW_monomer where DP = feed ratio × conversion.

^dMolecular weight and dispersity were determined by size-exclusion chromatography with multiangle light scattering (SEC-MALS) with a measured dn/dc of 0.1174 mL/g (Figure S4).

To examine the chain growth nature of PTDO polymerization, parallel reactions were setup and quenched at different times (Figure 2, Table 2, entries 1–5). The polymer molecular weight increased linearly with respect to monomer conversion and maintained a low Đ (1.07–1.26), indicative of good control over the polymerization (Figures 2A and S5–S7). A plot of ln([M]₀/[M]_t) versus time revealed a linear relationship to 3 h, indicating pseudo first-order kinetics consistent with a controlled polymerization (Figure S6). At 5 h, no significant increase in PTDO conversion and M_n was observed as the reaction reached equilibrium between monomer and polymer (Figure 2B). However, broadening of the SEC peaks was noticed at equilibrium, indicating chain transfer and backbiting events at high monomer conversion. Throughout the polymerizations, both E and Z alkene configurations were observed according to ¹H NMR analysis. An E/Z ratio of 5:1 was determined for the backbone olefins based on the integration of olefin signals, suggesting the predominant formation of trans-olefin under the investigated polymerization conditions (Figures S5 and S7). By increasing the [M]₀:[I] ratio, a series of PPTDOs with different molecular weights and consistently low Đ were synthesized (Figure 2C, Table 2, entries 5–7). The theoretical molecular weights were in good agreement with the measured M_n from SEC-MALS. These results unequivocally indicate that PTDO undergoes controlled ROMP before reaching equilibrium. Terminating the reaction at early times (3–5 h) enables access to well-defined PPTDO with a predictable molecular weight.

Figure 2.

Synthesis of PPTDOs by ROMP in accordance with Table 2. (A) Plot of M_n and Đ vs monomer conversion, obtained by a combination of SEC-MALS and ¹H NMR analysis. The dotted line represents the theoretical M_n. (B) SEC traces (normalized RI) of PPTDOs quenched at different reaction times (correlated to Table 2, entries 1–5). (C) SEC traces (normalized RI) of PPTDOs of different molecular weights (correlated to Table 2, entry 5: PTDO₅₅; entry 6: PTDO₉₄; entry 7: PTDO₂₁₆).

Table 2.

Kinetics of PTDO Homopolymerization and Synthesis of Different Molecular Weight PPTDO^a

entry	[M]0:[I]	time (h)	conv.b (%)	Mn,theoc (kDa)	Mn,MALSd (kDa)	Đ d
1	100:1	0.5	21	4.8	3.8	1.13
2	100:1	1	30	6.7	7.9	1.07
3	100:1	2	44	9.8	11.7	1.12
4	100:1	3	53	11.9	13.3	1.20
5	100:1	5	55	12.4	14.2	1.26
6	200:1	5	47	21.1	19.4	1.20
7	400:1	5	54	48.4	44.3	1.17

^aPolymerization condition: 2 °C under N₂.

^bConversion was determined by ¹H NMR spectroscopy.

^cTheoretical M_n was calculated from monomer conversion.

^dMolecular weight and dispersity were determined via SEC-MALS.

The thermal properties of PPTDO (DP = 94, E/Z = 7:1) were then investigated. At room temperature, PPTDO appeared as a yellow, brittle solid. The polymer became rubbery when left open to air, possibly due to the plasticizing effect of moisture. Thermogravimetric analysis revealed that the polymer decomposes at 235 °C (Figure S8). By differential scanning calorimetry, only the glass transition temperature of 42 °C was observed, indicating the amorphous and noncrystalline nature of the polymer (Figure S9). In combination, the thermal analyses show PPTDO is thermally stable and should have a long shelf life at room temperature or lower, under anhydrous conditions.

To understand the degradation mechanism of the material, both PTDO and PPTDO were subjected to a mixture of 0.25 M HCl in DMSO-d₆. Both ¹H and ³¹P NMR spectroscopies were used to monitor the degradation kinetics (Figures S10–S15). Over 48 h, the PTDO monomer completely degraded, as indicated by the disappearance of the ³¹P resonance at 18.07 ppm (Figure 3A). The appearance of a peak at 3.20 ppm attributed to be the single phosphoramidate linkage cleavage product. The signal at −6.72 ppm is well aligned with phenylphosphoric acid (δ = −6.23 ppm). The minor difference in chemical shift is caused by the different protonation state. The full PTDO to phenylphosphoric acid conversion was achieved in 192 h. ¹H NMR spectroscopy further confirmed the identities of the degraded products as phenylphosphoric acid and 1,4-diamino-2-butene (Figure S13). Notably, the degradation of PPTDO was slower than that of the monomer (Figure S10). At 240 h, more than 80% of the polymer degraded into phenylphosphoric acid as the ³¹P resonance at 13.00 ppm disappeared and a signal at −6.72 ppm emerged (Figure 3B). The peaks between 2.8 and 3.5 ppm represent the intermediate degraded polymers with different degrees of backbone hydrolysis. Finally, no high molecular weight species were detected by SEC-MALS, further confirming complete backbone degradation (Figure 3C). We expected that by modifying the molar mass of the polymer and solution pH, different backbone degradation rates could be achieved. Motivated by these results, we then moved on to investigate the possibility of using PTDO as a comonomer to introduce degradability into other ROMP derived polymers (Figure 4).

Figure 3.

Degradation of PTDO and PPTDO (DP = 94) in 0.25 M HCl in DMSO-d₆ at room temperature. ³¹P NMR spectra of (A) PTDO degradation and (B) PPTDO degradation at different times. (C) SEC traces of PPTDO before and after 240 h acid treatment. Note: ³¹P NMR spectroscopy of phenylphosphoric acid was measured to give a signal at −6.23 ppm.

Figure 4.

Synthesis and degradation of NB-PTDO copolymers. Synthetic scheme for the preparation of (A) random copolymers, and (B) block copolymers of NB and PTDO. SEC traces (normalized RI) of random copolymers with (C) NBPh, and (D) NBOEG before and after acid treatment. SEC traces (normalized RI) of block copolymers with (E) NBPh, and (F) NBOEG before and after acid treatment. Degradation condition: 0.5 M HCl in DMSO for 24 h at room temperature.

Random copolymerization of PTDO with norbornenes bearing phenyl and oligo(ethylene glycol) (NBPh and NBOEG) were performed (Figure 4A). A 1:1 ratio of PTDO: NB was dissolved in a mixture of MeOH/DCM and cooled to 2 °C before the addition of 0.02 equiv. of I. The reaction was quenched after 5 h with EVE. ¹H NMR spectroscopy revealed an 84% conversion of NBPh and a 64% conversion of PTDO, yielding the polymer NBPh₄₂-co-PTDO₃₂ with Đ of 1.13 (Figure S16). Similarly, NBOEG copolymerized with PTDO to give the polymer NBOEG₃₈-co-PTDO₃₂ with Đ of 1.13 (Figure S17). In both cases, PTDO conversion (~64%) was higher than that for PTDO homopolymerization (vide supra), possibly because of a more active chain end post norbornene incorporation. Both copolymers and their corresponding norbornene homopolymers were analyzed by SEC-MALS before and after 0.5 M HCl treatment (Figure 4C,D). In both cases, the copolymers completely degraded into small molecules, indicating a statistical incorporation of PTDO. The norbornene homopolymers remained unchanged post acid treatment. Next, two block copolymers of PTDO and norbornene were synthesized (Figure 4B). Norbornene was first polymerized with I at room temperature to yield the first block. Then the reaction mixture was cooled to 2 °C, followed by PTDO addition for chain extension. ¹H NMR spectroscopy revealed a PTDO conversion between 40~50% (Figures S18–S19). After treating the block copolymers NBPh₅₀-b-PTDO₂₅ and NBOEG₄₀-b-PTDO₃₃ with acid, SEC traces shifted toward longer retention times overlapping with signals from the polynorbornene controls (Figure 4E,F). These results show that, by controlling the PTDO feed ratio and position in the backbone, we can achieve controlled degradation, either partial or complete, of ROMP-derived polymers.

A potential application of PTDO containing amphiphilic copolymers is to produce biodegradable nanoparticles for biomedical applications, such as drug delivery. Both the block and random copolymers NBOEG₄₀-b-PTDO₃₃ and NBOEG₃₈-co-PTDO₃₂ self-assembled into spherical nanoparticles in aqueous solution, as evidenced by dry state transmission electron microscopy.⁴⁴ The nanoparticles assembled from the block copolymer NBOEG₄₀-b-PTDO₃₃ were uniform in size with a diameter of 24 ± 4 nm, whereas two separate populations were observed in NBOEG₃₈-co-PTDO₃₂ based nanoparticles: 55 ± 8 nm and 133 ± 27 nm (Figure S20A,B). Cytotoxicity of the NBOEG-PTDO nanoparticles against human cervical carcinoma HeLa cells was evaluated by a CellTiter-Blue assay. No appreciable cytotoxicity was observed even at a high concentration of nanoparticles at 150 μg/mL, highlighting the potential of these nanoparticles as nontoxic nanocarriers (Figure S20C).

In summary, we demonstrate an approach to generate degradable polymers featuring phosphoramidate linkages in their backbones. ROMP of the low strain diazaphosphepine based monomer, PTDO, at low temperatures proceeded in a controlled manner, affording high molecular weight polyphosphoramidates that undergo complete backbone degradation upon acid treatment. We also showed that the PTDO monomer copolymerized efficiently with norbornenes, introducing partial or complete degradability into those polymer backbones. Furthermore, the amphiphilic copolymers of PTDO and oligo(ethylene glycol) bearing norbornene self-assembled into micellar nanoparticles and showed no discernible cell cytotoxicity. Moving forward, we expect a plurality of cyclic diazaphosphepine based monomers can be generated via ring-closing metathesis and subsequently polymerized via ROMP.⁴⁵ Future studies, including the synthesis and polymerization of functionalized phosphoramidates, modifying the degradation rates of polymers and nanoparticles, as well as applications in drug delivery and recyclable materials are underway.