Antibacterial activity of the nitrovinylfuran G1 (Furvina) and its conversion products

2-Bromo-5-(2-bromo-2-nitrovinyl)furan (G1 or Furvina) is an antimicrobial with a direct reactivity against thiol groups. It is active against Gram-positive and Gram-negative bacteria, yeasts and filamentous fungi. By reacting with thiol groups it causes direct damage to proteins but, as a result, is very short-living and interconverts into an array of reaction products. Our aim was to characterize thiol reactivity of G1 and its conversion products and establish how much of antimicrobial and cytotoxic effects are due to the primary activity of G1 and how much can be attributed to its reaction products. Stability of G1 in growth media as well as its conversion in the presence of thiols was characterized. The structures of G1 decomposition products were determined using NMR and mass-spectroscopy. Concentration- and time-dependent killing curves showed that G1 is bacteriostatic for Escherichia coli at the concentration of 16 μg/ml and bactericidal at 32 μg/ml. However, G1 is inefficient against non-growing E. coli. Addition of cysteine to medium reduces the antimicrobial potency of G1. Nevertheless, the reaction products of G1 and cysteine enabled prolonged antimicrobial action of the drug. Therefore, the activity of 2-bromo-5-(2-bromo-2-nitrovinyl)furan is a sum of its immediate reactivity and the antibacterial effects of the conversion products.

For compound characterization, the 1 H, 13 C and 15 N NMR spectra were recorded at 700.1 ( 1 H), 176.0 ( 13 C) and 70.9 ( 15 N) MHz on a Bruker AVANCE III 700 NMR spectrometer. The 15 N NMR chemical shifts were indirectly obtained from 1 H detected 1 H-15 N gs-HMBC or 1 H-15 N gs-HSQC spectra. All NMR measurements were carried out at +20 °C. For 1 H NMR spectra, watersuppression pulse sequences using excitation sculpting or presaturation were used. Chemical shifts ( 1 H and 13 C) were indirectly referenced to TMS via the residual solvent signal (CDCl 3 : 7.26 and 77.0 ppm, respectively; CD 3 CN: 1.94 and 1.32 ppm, respectively; CD 3 OD: 3.31 and 49.0 ppm, respectively). The 15 N chemical shifts were referenced externally to the signal of neat nitromethane (381.7 ppm). NMR signals were assigned from gs-HSQC, gs-HMBC, gs-NOESY and TOCSY spectra. The reported 13 C NMR data corresponds to the 1 H-decoupled 13 C NMR data (noted on spectra as " 13 C{ 1 H}"), where the 1 H-13 C couplings have been removed.
Abbreviations used to define NMR spectral multiplicities are as follows: d = doublet; m = multiplet; bs = broad signal; vt = triplet with virtual couplings.
HRMS measurements were performed on a Thermo Electron LTQ Orbitrap mass spectrometer.
The mammalian cytotoxicity was assayed using the xCELLigence RTCA DP Instrument (Roche).
OD 390 were recorded with 1 s intervals using spectrophotometer Ultrospec 7000 (GE Healthcare Life Sciences).

HPLC-HRMS analysis of reactivity of G1 towards cysteine
In order to study the reactivity of G1 towards cysteine at pH 6.0, 10 µL G1 stock solution (10 mM in DMSO) and 10 µL cysteine hydrochloride stock solution (10 mM in H 2 O) were dissolved in 980 µL sodium citrate buffer (50 mM, pH 6.0). Aliquots (100 µL) of the reaction mixture were withdrawn every 60 s over a period of 7 min and the reaction in these aliquots was then stopped by quenching with 1 µL of orthophosphoric acid (85%). Samples were analyzed by HPLC ( Figure 1). Analytical HPLC (C 18 column, 150×4.6 mm, particle size 3 µm; mobile phase: 10% to 95% gradient over 9 minutes of 0.1% formic acid in H 2 O and 0.1% formic acid in acetonitrile; flow-rate 1.0 mL/min) chromatogram after 4 minutes incubation with cysteine.
The fractions corresponding to te peaks 2 and 4 were collected separately and the contents were analyzed by HRMS (FTMS+pNSI).  After separation of the insoluble cystine by filtration using a plug of cotton wool, an analytical sample of 3 was obtained by purification with preparative HPLC (R t =21.5-23.5 min, λ max =350 nm).

Case II:
In a cation-adjusted Mueller-Hinton II broth (MHB, 20 mg) aqueous (0.4 mL H 2 O and 0.4 mL D 2 O) medium, G1 (0.40 mg, 1.3 µmol) solution in DMSO-d 6 (40 µL) was added, followed by addition of cysteine hydrochloride (0.46 mg, 2.9 µmol) solution in H 2 O (20 µL). Appearance of products 3 and 4 signals were observed by 1 H NMR spectroscopy. The appeared 1 H NMR signals observed in the MHB medium were the same as were present in the reaction mixture, where inorganic base and organic co-solvent were used.

Case I:
To the same reaction mixture in CD 3 CN/H 2 O (1:1), where compound 3 was obtained and determined, excess of cysteine hydrochloride solution in H 2 O was added until there were no signals in the 1 H NMR spectrum corresponding to G1 or compound 3. Diastereoisomeric 1:1 mixture of the obtained product 4 was determined by NMR.

Case II:
In a cation-adjusted Mueller-Hinton II broth (MHB, 20 mg) aqueous (0.5 mL H 2 O and 0.1 mL D 2 O) medium, cysteine hydrochloride (1.8 mg, 11.4 µmol) solution in H 2 O (100 µL) was added, followed by addition of G1 (15 µg, 0.050 µmol) solution in DMSO-d 6 (10 µL). Appearance of product 4 signals were observed by 1 H NMR spectroscopy, where the characteristic signals at 6.31 and 6.40 ppm, corresponding to the product's furane-ring hydrogens, were clearly visible and well-separated from all other signals in the 1 H NMR spectrum.

Case III:
In a cation-adjusted Mueller-Hinton II broth (MHB, 20 mg) aqueous (0.4 mL H 2 O and 0.4 mL D 2 O) medium, G1 (0.40 mg, 1.3 µmol) solution in DMSO-d 6 (40 µL) was added, followed by addition of cysteine hydrochloride (1.4 mg, 8.9 µmol) solution in H 2 O (60 µL). Although an excess of CysHCl was used, this time a mixture of unreacted G1 and product 4 was obtained, indicating that in an unbuffered solution and with higher G1 loading, the product 4 formation can be slow.

H )ethyl]-L-(N,N,O-H 3 )cysteine
(3.4 mg, 11.5 µmol) was dissolved in CD 3 CN (0.3 mL), followed by addition of D 2 O (0.3 mL). Saturated solution of NaHCO 3 in D 2 O (15 µL) was added via a syringe, followed by addition of cysteine hydrochloride (7.3 mg, 46.3 µmol) solution in D 2 O (105 µL). The formed cystine precipitate was filtered off using a plug of cotton wool. The filtrate composition was analyzed by NMR, which showed the presence of 1:1 diastereomeric mixture of the deuterated product 4 and cysteine.