Organocatalytic activity of granaticin and its involvement in bactericidal function

We previously discovered that actinorhodin, a benzoisochromanequinone antibiotic produced by Streptomyces coelicolor A3(2), serves as a catalyst facilitating the oxidation of ascorbic acid and cysteine (PNAS 48:17,152, 2014). In the present study, we screened for similar ascorbic acid-oxidizing activity in the culture broth of various Streptomyces spp., and discovered marked activity in the culture broth of Streptomyces vietnamensis. The principle active compound was granaticin, a pigmented antibiotic that is structurally related to actinorhodin. The absence of any metals in the purified granaticin fraction indicated that granaticin was an organocatalyst. Granaticin catalyzed the oxidation of L-ascorbic acid, generating L-dehydroascorbic acid and hydrogen peroxide (H2O2) at a 1:1 stoichiometric ratio, with 15 times higher reactivity than that of actinorhodin at an optimum pH of 7.0. Granaticin also oxidizes sulfhydryl compounds, including L-cysteine and glutathione. Growth inhibitory assays demonstrated that knockout mutants of the catalase gene exhibit high sensitivity to granaticin. The results suggest that the bactericidal activity of granaticin is exerted by the oxidation of sulfhydryl groups of cellular components and the toxicity of H2O2 generated during the oxidation reaction.

www.nature.com/scientificreports/ organocatalysts especially in the enantio-selective synthesis has increased remarkably, reflecting their growing potential 7 . Proline and imidazolidinones are well-characterized organocatalysts [8][9][10] . In 2021, the Nobel Prize in Chemistry was awarded to Drs MacMillan and List for the development of asymmetric organocatalysts 11 . Our previous discovery of actinorhodin was the first instance of an organocatalyst produced by microbial secondary metabolism.
In the present study, we screened for additional Streptomyces metabolites exhibiting ascorbic acid-oxidizing activity to further explore the occurrence of organocatalysts in microbial metabolites. We discovered that granaticin, a benzoisochromanequinone antibiotic produced by Streptomyces vietnamensis, is catalytically active. This discovery not only highlights the use of microbial products as organocatalysts, which may be important in industrial applications, but also deepens the insights into the diversity of action mechanism of antibiotics and their role in the natural environment.

Results
Identification of granaticin as a catalytic compound. To study the occurrence of organocatalytic activity in Streptomyces metabolites, we assessed the L-ascorbic acid-oxidizing activity in the culture supernatant of various Streptomyces strains (see "Methods"), and discovered that the culture broth of S. vietnamensis exhibited marked O 2 consumption when it was mixed with 10 mM L-ascorbic acid ( Fig. 2A). The same level of O 2 -consuming activity was observed even after heat treatment at 95 °C for 30 min (data not shown), indicating that the catalyst is a heat-stable molecule.
The culture broth of S. vietnamensis was purple in color, which was likely due to the production of a purple benzoisochromanequinone antibiotic, known as granaticin ( Fig. 1) 12 . Previously, we proposed that the catalytic activity of actinorhodin is based on its quinone structure, since a weak but distinct L-ascorbic acid oxidizing activity was observed with several quinolic compounds of plant origin 4 . These previous observations led us to speculate granaticin-induced catalytic activity was present in the culture supernatant of S. vietnamensis. Therefore, the principle active compound was isolated using a previously published purification method for granaticin 13 . As expected, the L-ascorbic acid-oxidizing activity was fractionated into the same fraction as that of granaticin.
LC-MS analysis of the resultant purified active fraction showed a single peak at 443 m/z (ESI, negative) (Fig. 3A). The UV-visible absorption spectrum of this fraction containing peaks at 220, 285, 490, 525 and 570 nm (Fig. 3B) was identical to that of granaticin previously reported by James and Edwards 14 . These results confirm that the isolated fraction contained purified granaticin. The metallic ion content of the fraction was determined using an inductively coupled plasma optical emission spectrometer (ICP-OES), but none of the 71 metallic ion species that could be identified by the analytical method were detected.
Catalytic activity of granaticin.

(i) Specific activity
The L-ascorbic acid oxidizing activity of purified granaticin was determined. As shown in Fig. 2B, the dissolved O 2 concentration in the reaction solution started to decrease immediately after mixing with granaticin. The addition of catalase reduced the ratio of O 2 consumption (Fig. 2B dashed line), indicating the occurrence of H 2 O 2 in the reaction mixture. LC-MS analysis of the reaction mixture revealed the presence of granaticin at the starting concentration, and the absence of any derivatives (Supporting  www.nature.com/scientificreports/ information figure S1), indicating that granaticin catalyzes the oxidation of L-ascorbic acid without changing its structure or being destroyed. The specific activity of granaticin in L-ascorbic acid oxidation was 53.7 U/μmol, which is approximately 15-fold that of actinorhodin (3.5 U/μmol) 4

. (ii) Reaction products and stoichiometry
Our previous work demonstrated that actinorhodin catalyzes the oxidation of L-ascorbic acid, generating L-dehydroascorbic acid and H 2 O 2 4 . To determine whether this was the case with granaticin, the content of the reaction mixture was analyzed by HPLC. The two peaks generated by the reaction with granaticin corresponded to those of L-dehydroascorbic acid and H 2 O 2 (Fig. 4). This indicates that granaticin catalyzes the same reaction as actinorhodin.
The stoichiometry of granaticin was determined (Fig. 5). The substrates (L-ascorbic acid and O 2 ) and products (L-dehydroascorbic acid and H 2 O 2 ) were present at a 1:1:1:1 stoichiometric ratio. (iii) Dependence on pH and temperature The pH and temperature dependence of the catalytic activity of granaticin was assessed. The highest activity was observed at pH 7.0 (Fig. 6A). The oxidation of L-ascorbic acid occurred spontaneously at pH 10 and higher, as described previously 4 . The activity increased with increasing temperature from 25 to 40 °C (Fig. 6B). The activity at temperatures higher than 40 °C could not be measured because of the heat sensitivity of the electrode system. Although data were slightly fluctuated, preincubation at temperatures between 30 and 100 °C for 1 h did not significantly affect the catalytic activity (Fig. 6C).
(iv) Substrate specificity The substrate specificity of granaticin was determined by reacting granaticin with various substrates. The substrate selectivity profile was similar to that of actinorhodin 4 , but the reactivity was different (Table 1). Compared to L-ascorbic acid (100%), the reaction of its stereoisomer D-isoascorbic acid was relatively weak (77.3%), indicating that the catalysis occurs asymmetrically. Much weaker but distinctive reactions were observed with L-cysteine methyl ester (1.5%), L-cysteine (0.6%), dithiothreitol (0.4%), D-cysteine (0.4%), and glutathione (0.1%). The narrow but distinctive difference in the reactivity with cysteine isomers also indicates the asymmetricity of the reaction. Oxidation of N-acetyl-L-cysteine, DLpenicillamine, DL-homocysteine, lipoamide, and α-lipoic acid was not observed when O 2 consumption was monitored for 30 min. www.nature.com/scientificreports/ Growth inhibition activity of granaticin. We previously hypothesized that the antibiotic activity of actinorhodin is at least partly due to the toxicity of H 2 O 2 generated along with the catalytic reaction, based on the observation that the addition of catalase rescued the growth of S. aureus exposed to actinorhodin on solid media 4 . The same assay using the filter disc diffusion method showed that the growth inhibition effect of granaticin on S. aureus was also ameliorated by the addition of catalase (data not shown). We next studied the sensitivity of catalase-deficient strains of E. coli and B. subtilis (Table 2). Figure 7 shows the dose effect of granaticin on the early growth (5 h cultivation) of these mutants in Luria-Bertani (LB) liquid medium. Granaticin is known to be an antibiotic that statically inhibits the growth of gram-positive bacteria 15 . As expected, the sensitivity of E. coli was far lower than that of B. subtilis (ca, 100-foled difference). However, in both bacteria, the catalase-deficient mutants exhibited higher sensitivity to granaticin than the control strain. In E. coli, ME8302 lacking the two catalase genes (katF and katG) was more sensitive than ME8303, lacking active katG but retaining katF (Fig. 6A). We also performed the assay with an E. coli mutant (ME6259) for recA, the master regulator of the SOS response system 16 , and found that it had a higher sensitivity than the control strain (ME6111) (Fig. 6A). A similar profile was obtained for JM109 (recA-) and JM105 (recA +) strains (data not shown). Similarly, the mutant for katA (BKE08820) and katE (BKE39050) of B. subtilis exhibited higher sensitivities than the control strain (Fig. 6B). These results support our hypothesis that the toxicity of granaticin is based www.nature.com/scientificreports/ on the effect of H 2 O 2 that is generated by its catalytic reaction. The high sensitivity of recA mutants could be due to their disability to repair the damage caused in DNA structure by the oxidative agent.

Discussion
In the present study, we determined that granaticin serves as a catalyst to facilitate the oxidation of L-ascorbic acid to L-dehydroascorbic acid. The lack of structural changes or depletion during this reaction, as well as the lack of metallic content, indicate that granaticin is an organocatalyst. The relatively narrow but distinctive difference in the reactivity with the stereoisomers of ascorbic acid and cysteine (Table 1) indicates that the catalytic reaction occurs asymmetrically. This finding reinforces the notion that these types of functional molecules are present in microbial products. Granaticin and actinorhodin belong to the same class of compounds, termed benzoisochromanequinones, which share a naphthoquinone structure (Fig. 1). We previously showed that some plant-derived natural products, which commonly retain a quinone structure, also exhibit catalytic activity, facilitating the oxidation of ascorbic acid 4 . Thus, we speculate that the quinone structure is fundamental to the catalytic function of this class of organocatalysts. We hypothesize that the structural interconversion between quinone and quinol within the catalytic molecule mediates the oxidation of ascorbic acid to dehydroascorbic acid and the reduction of O 2 to H 2 O 2 (Fig. 8). Granaticin differs from actinorhodin in terms of its specific activity, optimum pH, and substrate specificity. It is likely that the differences in the structure surrounding the quinone backbone affect the catalytic properties of these molecules. A detailed understanding of the correlation of their chemical structure with the reactivity will broaden the field of organocatalysts.
Granaticin was first identified in the culture broth of Streptomyces olivaceus based on its antibiotic activity against gram-positive bacteria 17 . It was also cytotoxic to mammalian cells. Chang et al. 18 discovered that an antibiotic termed litmomycin, which is produced by Streptomyces litmogenes, exhibited antitumor activity against leukemia in mice, and that the chemical structure of litmomycin was identical to that of granaticin. Early studies on the bactericidal mechanisms of granaticin suggest that its interaction with RNA or related molecules    19 , and maturation of ribosomal RNA 20 . Independently of its possible interaction with RNA or related molecules, Sturdik and Drobnica 21 determined that granaticin is reactive to sulfhydryl groups by observing its effect on sulfhydryl compounds and sulfhydryl enzymes, including glyceraldehyde-3-phosphate dehydrogenase, and proposed that this kind of activity is fundamental to the cytotoxicity of granaticin. This is consistent with the report by Gibson-Clay et al. 22 , who observed the inhibitory effect of granaticin on the activity of pyruvate decarboxylase, the central metabolic enzyme, depending on its activity on the oxidation and reduction of the SH group of lipoamide, which is associated with the E2 domain.
The ambiguity of the action mechanism of granaticin can be explained by its organocatalytic activity. Previously, we hypothesized that the bactericidal activity of actinorhodin is, at least in part, due to the H 2 O 2 produced during catalysis, based on the observation that the addition of catalase reduced the level of growth inhibition 4 . H 2 O 2 generated via the catalytic reaction exerts toxic effects on cellular functions. This is also supported by the antibiotic activity measurements performed in the present study (Fig. 7). The high sensitivity of the catalase mutants to granaticin suggests that the generation of H 2 O 2 within or on the surface of the target cell, along with A B C Figure 6. Effect of pH and temperature on granaticin activity. The catalytic activity of granaticin was measured at various pH values at 30 °C (A) and at various temperatures at pH 7.0 (B). The activity at temperatures higher than 40 °C could not be measured because of the operation limits of the oxygen electrode. The thermostability of the compound (C) was also studied by incubating granaticin solution at various temperatures for 1 h prior to the activity measurement. A representative result of repeated experiments is shown.   www.nature.com/scientificreports/ the oxidation of cellular material, is fundamental to the antibiotic and cytotoxic activity of this compound. The similar feature of the recA mutants is interpreted as the oxidative effect of H 2 O 2 damages the chromosomal DNA and therefore significantly affects the viability of the recA mutant strains, which lack the ability to express the DNA repair system 16 . The oxidative damage to nucleic acid structures may be correlated with the abovementioned hypothesis that granaticin exerts its cytotoxicity by introducing damage to RNA and related functions.
Recently, Nass et al. 23 reported that γ-actinorhodin, an actinorhodin derivative, exhibits unique antibacterial activity. The bactericidal activity of γ-actinorhodin was supposed to be due to the generation of reactive oxygen species (ROS), as the S. aureus mutant for superoxide dismutase was highly sensitive to the substance. Similar observations have been reported for the bactericidal activity of a plant-derived bioactive quinone termed thymoquinone 24 , a chemically synthesized anticancer and antibacterial compound YM155 with a naphthoimidazole structure 25 , and 1,4-naphtoquinones exhibiting antibacterial activity and cytotoxicity 26 . These compounds may also exert their toxicity by generating ROS. The mode of action of these other bioactive quinolic substances reinforces the view that the toxicity of granaticin is also based on the generation of ROS.
In addition to ROS generation, inactivation of proteins due to the oxidation of sulfhydryl groups could be the basis of the observed toxicity of granaticin and related compounds. Although the specific activity was relatively low, in the present study we determined that granaticin promotes the oxidation of cysteine. This suggests that the sulfhydryl group, which is exposed to the protein surface, could be the target of oxidation by granaticin. This agrees with the abovementioned observation that granaticin inhibits sulfhydryl enzymes. Overall, the lines of evidence strongly suggest the existence of a novel mechanism for the action of antibiotics and bioactive substances based on their catalytic activity.
The reactivity of granaticin at neutral pH and moderate temperature indicates that its catalytic function can proceed under standard conditions. The granaticin biosynthetic gene cluster exists in the genome of various Streptomyces and other actinomycetes 27,28 . Deng et al. 29 suggested that the ability to produce granaticin is distributed via horizontal gene transfer. Thus, the newly discovered function of this natural product is widespread in the natural environment, and it is readily available for various oxidative reactions. Taken together, these features with regard to granaticin support the occurrence of an unprecedent role, function, and utilization of microbial products.

Methods
Bacterial strains and media. The Streptomyces strains used for catalytic activity screening were obtained from the National Institute of Technology and Evaluation (NITE), Chiba, Japan (Table 3). Staphylococcus aureus NBRC100910 was used for the growth inhibition assay, and was also obtained from the NITE. Escherichia coli strains used for the growth inhibition assay (Table 2) were obtained from the National Bioresource Project (NBRP) (https:// nbrp. jp/ en/). JM105 and JM109 were purchased from Takara Bio Inc. (Kyoto, Japan). Bacillus subtilis strains 168, BKE08820 (katA-deficient), and BKE39050 (katE-deficient) were also obtained from the NBRP.
Streptomyces strains were cultured at 28 °C in Bennett's/maltose (BM) medium containing 1 g/L yeast extract (Difco Laboratories, Detroit, Michigan), 1 g/L meat extract (Kyokuto, Tokyo, Japan), 2 g/L NZ amine (Wako Pure Chemical Industries Ltd., Osaka, Japan), and 10 g/L maltose (pH 7.2). For the antibiotic assay, E. coli and B. subtilis strains were cultured in Luria-Bertani (LB) medium containing 10 g/L tryptone (Difco), 5 g/L yeast extract (Difco), and 5 g/L NaCl. Agar was added to LB broth to a final concentration of 1.5% to prepare solid media. Chemicals were purchased from Kokusan (Tokyo, Japan), unless otherwise indicated. www.nature.com/scientificreports/ containing 10 ml BM liquid medium and cultured at 28 °C for 5 days with reciprocal shaking at 300 rpm. Then, 1 ml of this preculture was inoculated into a baffled 500 ml Erlenmeyer flask containing 100 ml of BM medium. The culture was incubated at 28 °C with rotary shaking at 160 rpm for 1 day. The culture supernatant was obtained by centrifugation at 10,400×g at 4 °C for 15 min. The supernatant was added to an equivalent volume of ethyl acetate/methanol/acetic acid (80/20/5) to extract the active compounds. The organic phase was collected, evaporated to dryness, and dissolved in acetonitrile, and then fractionated using a Shimadzu LC-10Avp highperformance liquid chromatography (HPLC) system equipped with a COSMOSIL 5C18-MS-II column (4.6 mm i.d. × 150 mm; Nacalai Tesque, Inc., Kyoto, Japan) under the following conditions: column temperature, 40 °C; flow rate, 1 ml/min; photodiode array detector, 190-800 nm; mobile-phase solvent, 55% methanol; and isocratic elution for 15 min. The eluate corresponding to the isolated peak at 10.6 min was obtained and evaporated to dryness using a rotary evaporator. The dried material was then dissolved in methanol: water at a ratio of 55:45 due to its high solubility and stability, and was henceforth termed 'granaticin solution' . The solution was maintained at 4 °C to avoid evaporation of methanol. The molecular mass was determined by liquid chromatography electrospray ionization mass spectrometry (LC-ESI-MS; negative mode) with a Shimadzu LCMS-2010 system equipped with a COSMOSIL 5C 18 -PAQ column (4.6 mm i.d. × 150 mm; Nacalai Tesque) under the following conditions: column temperature, 40 °C; flow rate, 1 ml/min; and photodiode array detector, 190-800 nm; mobile phase, 0.5% (v/v) acetic acid in 50% methanol; isocratic elution for 40 min.
Characterization of the catalytic activity of granaticin. The catalytic activity of granaticin was studied using an oxygen electrode under the same conditions as those used for activity screening. The dissolved O 2 concentration in a 1 ml reaction mixture containing 20 µl of 1 µM granaticin solution was monitored for 5 min using an oxygen electrode. One hundred units of catalase (Nacalai Tesque, Inc., Kyoto, Japan) were added if needed. For the negative control measurement, 20 µl of 55% methanol was added in place of the granaticin solution to confirm that the presence of methanol does not affect the accurate measurement. The O 2 consumption was also measured using a non-contact O 2 sensor chip system (OXY-1 SMA; PreSensPrecision Sensing GmbH, Regensburg, Germany) for reactions where oxidation occurred slowly. The reaction mixture was prepared under the same conditions and enclosed in a glass vial containing a sensor chip attached to its inner surface. The dissolved O 2 concentration was recorded every 1 min for up to 30 min by scanning the chip with the chip reader from the outside of the vial. One unit of catalytic activity was defined as the amount of the compound that catalyzed the consumption of 1 µmol of O 2 per min.
To identify the reaction products generated by the catalytic oxidation of L-ascorbic acid, the reaction mixture was incubated at 30 °C for 5 min, and formic acid was added to a final concentration of 1% to stop the reaction. The temperature dependence of the catalytic activity of granaticin was determined by measuring the reduction in O 2 in the reaction mixture incubated at various temperatures between 25 and 40 °C. The composition of the reaction mixture was the same as that above, except for the addition of 2 mM L-ascorbic acid. pH dependence was analyzed by adding 2 mM L-ascorbic acid to 100 mM of the following buffers: glycine/HCl buffer (pH 2.0-3.0), citate/NaOH buffer (pH 3.0-6.0), potassium phosphate buffer (pH 6.0-7.5), Tris-HCl buffer (pH 7.5-9.5), and glycine/NaOH buffer (pH 9.5). Heat tolerance was investigated by incubating the granaticin solution at various temperatures for 1 h prior to activity measurement using the oxygen electrode as before. The reaction mixture conditions were the same as those used for the temperature dependence analysis.
Growth inhibition analysis. E. coli and B. subtilis strains were precultured for 16 h with shaking at 135 rpm in glass tubes containing 5 ml LB liquid medium at 37 °C and 30 °C, respectively, and then inoculated to a final concentration of 0.5% into the culture medium under the same conditions as the preculture, to which granaticin was added at 0-100 μM (E. coli) and 0-1.0 μM (B. subtilis). After 5 h of cultivation, growth was quantified by measuring the optical density at 600 nm.

Data availability
All data are contained within the manuscript.