Enzymatic Synthesis and Antimicrobial Activity of Oligomer Analogues of Medicinal Biopolymers from Comfrey and Other Species of the Boraginaceae Family

This study reports the first enzymatic synthesis leading to several oligomer analogues of poly[3-(3,4-dihydroxyphenyl)glyceric acid]. This biopolymer, extracted from plants of the Boraginaceae family has shown a wide spectrum of pharmacological properties, including antimicrobial activity. Enzymatic ring opening polymerization of 2-methoxycarbonyl-3-(3,4-dibenzyloxyphenyl)oxirane (MDBPO) using lipase from Candida rugosa leads to formation of poly[2-methoxycarbonyl-3-(3,4-dibenzyloxyphenyl)oxirane] (PMDBPO), with a degree of polymerization up to 5. Catalytic debenzylation of PMDBPO using H2 on Pd/C yields poly[2-methoxycarbonyl-3-(3,4-dihydroxyphenyl)oxirane] (PMDHPO) without loss in molecular mass. Antibacterial assessment of natural polyethers from different species of Boraginaceae family Symhytum asperum, S. caucasicum, S. grandiflorum, Anchusa italica, Cynoglossum officinale, and synthetic polymers, poly[2-methoxycarbonyl-3-(3,4-dimethoxyphenyl)oxirane (PMDMPO) and PMDHPO, reveals that only the synthetic analogue produced in this study (PMDHPO) exhibits a promising antimicrobial activity against pathogenic strains S.aureus ATCC 25923 and E.coli ATCC 25922 the minimum inhibitory concentration (MIC) being 100 µg/mL.


Introduction
Several plants from the Boraginaceae family such as Symphytum officinale L. (S. officinale, Figure 1a), comfrey, and Anchusa italica (A. italica) synonymous for Anchusa azurea (A.azurea) are known to exhibit a broad pharmacological activity [1][2][3][4][5]. Comfrey was used in traditional medicine for bone breakages, sprains and rheumatism, liver problems, gastritis, ulcers, skin disorders, joint pain and contusions, wounds, gout, hematomas, and thrombophlebitis [1], while A. italica-as stimulant, tonic, demulcent, in bilious complaints, fever, cough, asthma and as diuretic in bladder and kidney stones, as diaphoretic, narcotic, hypnotic, antiarthritis, antirheumatic, and cathartic [4]. Extracts of comfrey and A. italica showed potential antimicrobial effects on several bacterial strains tested, especially against Staphylococcus aureus ATCC 25923 (S. aureus) [1,5], as well as antifungal activity against Bipolaris oryzae (B.oryzae) (comfrey) [1]. It was highlighted that the effect of the investigated extracts could be attributed to a mix of components including phenolic compounds (caffeic and chlorogenic acids, luteolin glycoside) and allantoin, possibly due to the synergy between compounds, as well as with other bioactive molecules present in the whole plants [1,5]. Since recent research revealed that allantoin is not responsible for the antimicrobial activity of the aqueous extract of comfrey root (Figure 1b) [7], preclinical studies regarding antimicrobial activities of comfrey extracts [1] demonstrates the need for deep investigations of extracts from all parts of plants, isolation, and chemical structure determination of bioactive compounds.
Medicinal effects of Symphytum and Anchusa plants used in folk medicine could be attributed not just to some low molecular mass phenolic compounds and the abovementioned bioactive molecules but to PDHPGA as well. PDHPGA is a unique natural polyether, as it contains aliphatic ether groups in its polymer backbone. Thus, PDHPGA and its analogues can be promising candidates for many biomedical applications based on their outstanding bioactive characteristics. The formation of those analogues by avoiding toxic catalysts is a challenging but highly beneficial undertaking, and that is the focus of the current study where the major aims are: a) the preparation of PDHPGA synthetic analogues by using enzyme-mediated polymerization and b) the evaluation of antimicrobial activity of the polymers formed.

2-Methoxycarbonyl-3-(3,4-Dibenzyloxyphenyl)-Oxirane (MDBPO) (4)
A solution of 3,4-dibenzyloxybenzaldehyde (0.47 g, 1.5 mmol) and methyl chloroacetate (0.24 g, 2.25 mmol) in 3 mL THF was added to a suspension of Na (0.05 g, 2.25 mmol) in dry methanol (1.0 mL) at −10 • C during a period of 30 min. The mixture was stirred at −5 • C for 2 h and then at room temperature for 3 h. The mixture was poured into 10 mL ice water containing acetic acid (0.01 mL). Crude products precipitated as a white powder, which was collected by filtration, washed with water, and dried in vacuo.  (7) In a typical polymerization, methyl 3-(3,4-dibenzyloxyphenyl)glycidate 4 (50 mg, 0.128 mmol) was dissolved in anhydrous toluene (500 µL). To this solution, lipase from Candida rugosa (20 mg) was added forming a suspension. Polymerization was initiated by heating the suspension in an oil bath at 80 • C and was allowed to react for 7 days under vigorous stirring. Polymerization was halted by addition of THF (5 mL) and subsequent filtration through a 0.2 µm PTFE filter to remove the biocatalyst. The product (7) was concentrated in a rotary evaporator and obtained as a viscous yellow oil.

Methods
1 H NMR and 13 C NMR spectra were recorded using CDCl 3 or acetone-d 6 as a solvent at 22 • C with a Bruker AVANCE 600 MHz instrument (Bruker Co., Billerica, MA, USA).
Fourier transform infrared spectroscopy (FT-IR) spectra were obtained on a Bruker Tensor 27 spectrophotometer with an MIR source and a DLaTGS detector all from Bruker Co., USA. Spectra were recorded under ambient conditions at a resolution of 4 cm −1 . A total of 64 scans were recorded for each spectrum in addition to the background.
Size exclusion chromatography (SEC) was conducted on a system consisting of an M510 pump, a U6K universal injector (both from Waters Co., Milford, MA, USA), three 5 µm PL Gel columns (50 Å, 500 Å, and Mixed C), and a Viscotek 250 dual refractive index/viscometry detector (Malvern Panalytical, Malvern, UK). All analyses were conducted at 45 • C with freshly distilled THF at a flow rate of 0.

Determination of Minimum Inhibitory Concentration (MIC) and Minimum Bactericidal Concentration (MBC)
In order to ascertain that the possible antimicrobial activity is due to the effect of polymeric substances, pure DMSO (used to dissolve PMDMPO and PMDBPO polymers) was preliminarily tested alone for antimicrobial activity by growth inhibition assay at the maximal concentration (2% v/v) used for polymer treatment of bacterial tester cells. The results revealed that pure DMSO does not affect bacterial growth at the concentration assayed.
MICs were determined by a microdilution method [23]. Briefly, a series of polymer solutions (i.e., PDHPGA-SA, PDHPGA-SC, PDHPGA-SG, PDHPGA-AI, PDHPGA-CO, PMDMPO, PMDBPO, and PMDHPO) were prepared with a range of concentrations from 100 to 0.75 µg/mL obtained by two-fold serial dilution [24]. The serial dilutions were made in Mueller-Hinton broth (MH) (Sigma Aldrich) in a 96-well plate, starting from a stock solution of 5 mg/mL in DMSO or water. To each well was added 10 µL of a bacterial suspension, obtained from a 24 h culture grown at 37 • C on Tryptic Soy Agar (TSA), containing~10 6 cfu/mL. A positive and negative control, consisting respectively of bacterial strains in the medium without extract, and the medium without both extract and inoculum were also included in the 96-well plate. A substance control, consisting only of the substance solution without bacterial inoculum, was added to evaluate the absorbance of the substance. The plate was incubated at 37 • C for 24 h. The MIC was determined by a microplate reader (Glomax®-Multi Detection System TM297, Promega, Milano, Italy) as the lowest concentration of compound whose optical density (OD) at 570 nm, was comparable with the negative control wells (broth only without inoculum). To determine the MBC, 100 µL of subculture from each negative well and from the positive control of MIC determination was put onto a TSA plate for 24 h at 37 • C. The MBC is defined as the lowest concentration of substance allowing microbial growth up to a maximum of three colonies.

Results and Discussion
Many biocatalytic polymerizations have been exploited for the synthesis of polymers in vitro [25][26][27] including some recent research from our group [28][29][30]. Furthermore, the biocatalytic synthesis of ethers in vitro and in vivo is also known [31,32]. Naturally occurring ethers include small molecules such as antibiotics, or aromatic polymers such as lignin. In the latter case, peroxidases initiate the radical coupling of monolignols to lignin, yielding ether links between two aromatic rings or between an aromatic ring and an aliphatic moiety [32]. However, reports on the enzymatic synthesis of polymers that contain aliphatic ethers as repeating unit are sparse. These include the ring-openingpolymerization (ROP) of glycidol by lipases which yielded linear poly(glycidol) [33][34][35]. Additionally, lipase-catalyzed co-polymerizations of glycidol and acid anhydrides have been reported [34][35][36]. On the other hand, the biosynthesis of the rather rare class of natural poly(ether)s, that PDHPGA belongs to, has not been attempted.

Enzymatic Polymerization
In this study, we attempted the polymerization of glycidates (4-6) under mild and more environmentally-friendly conditions using lipases from two different sources-Candida rugosa (C. rugosa) and Penicillium camemberti (P. camemberti). In our previous study [28], they showed high catalytic activity towards various substrates. The polymerization was carried out in anhydrous toluene at 80 • C for 7 days using a biocatalyst concentration of 4%, Scheme 1a. We also found lipase from C. rugosa was the most efficient in inducing the ringopening polymerization of MDBPO while the enzyme from P. camemberti was not active under the polymerization conditions chosen. SEC ( Figure 4) and MALDI-TOF analyses ( Figure 5) showed the formation of poly(MDBPO) (7) with a degree of polymerization up to 5. The mass spectrum contained well resolved signals spaced with mass numbers, which were close to the molar mass of the expected repeating unit (390.1); the observed mass numbers of each individual peak roughly corresponded to molar masses of polymer structures with a -OH group at the terminal (7). Due to the nature of this analysis, the existence of higher oligomers could not be confirmed or excluded. The reactions with BDBPO and TBDBPO did not show any trace of polymer, most probably due to the steric hindrance by the benzyloxy-and t-butyloxy-groups. Catalytic debenzylation of PMDBPO (7) using H 2 on Pd/C yielded the synthetic analogue of poly(MDHPO) (2) at~80% yield and without loss in molecular mass (Scheme 1b).

Antimicrobial Activity
The substances to be tested by microbiological assay were initially solubilized in distilled water or DMSO (5 mg/mL) and were sterilized by microfiltration.
The samples were subjected to an evaluation of antimicrobial activity against the following important human pathogenic bacterial strains: Staphylococcus aureus ATCC 25923, Pseudomonas aeruginosa ATCC 15442, Enterococcus faecalis ATCC 29212, Escherichia coli ATCC 25922. Antimicrobial activity against free-living strains (i.e., planktonic cell cultivation) was evaluated using Muller-Hinton broth (MHB) growth medium and was expressed in terms of minimum inhibitory concentration (MIC). For each substance, the antimicrobial activity was assessed at a range of concentrations from 100 to 0.75 µg/mL (obtained by two-fold serial dilution). Among the polymers evaluated (PDHPGA-SA, PDHPGA-SC, PDHPGA-SG, PDHPGA-AI, PDHPGA-CO, PMDMPO, PMDBPO, and PMDHPO) only PMDHPO showed antimicrobial activity at the tested concentration, Table 1. In particular, PMDHPO was active against S. aureus ATCC 25923 and E. coli ATCC 25922 in MHB growth medium at the maximum concentration used. Interestingly, the MICs against S. aureus ATCC 25923 and E. coli ATCC 25922 also coincided with the minimum bactericidal concentration (MBC) of the PMDHPO as it was revealed by counting colony forming units (CFU) of bacterial cultures obtained by plating aliquots of treated cultivations on TSA growth medium.

Conclusions
This is the first attempt to produce a synthetic analogue of natural polymer poly[2methoxycarbonyl-3-(3,4-dihydroxyphenyl)oxirane], isolated from the roots and stems of AI and SG. The novelty of the synthetic strategy is that the targeted compound was prepared by enzymatic ROP of methyl 3-(3,4-dibenzyloxyphenyl)glycidate using lipase from C. rugosa. It should be noted that the oligomers formed are the first enzymatically obtained analogues of the natural polyether-PDHPGA.
Antibacterial assessment of natural polyether from different species of Boraginaceae family PDHPGA-SA, PDHPGA-SC, PDHPGA-SG, PDHPGA-AI, PDHPGA-CO, and synthetic polymers PMDMPO, PMDBPO, and PMDHPO revealed that only PMDHPO showed antimicrobial activity against pathogenic strains S.aureus ATCC 25923 and E.coli ATCC 25922 at the concentrations used.
The synthetic PMDHPO would be an interesting target for diverse biological tests, and the potential enzymatic demethylation of this product could lead to a truly synthetic analogue of PDHPGA.  Data Availability Statement: All data obtained by this study are included in this article.