In the context of research for new cytotoxic compounds, obtaining bioactive molecules from renewable sources remain a big challenge. Microorganisms and more specifically Actino- bacteria from original sources are well known for their biotechnological potential and are hot- OPENACCESS spots for the discovery of new bioactive compounds. The strain DP94 studied here had Citation: Noe¨l A, Van Soen G, Rouaud I, Hitti E, shown an interesting cytotoxic activity of its culture broth (HaCaT: IC = 8.0± 1.5μg/mL; Tomasi S (2020) Optimization of cytotoxic activity B16: IC = 4.6± 1.8μg/mL), which could not been explained by the compounds isolated in of Nocardia sp culture broths using a design of experiments. PLoS ONE 15(1): e0227816. https:// a previous work. The increase of the cytotoxic activity of extracts was investigated, based doi.org/10.1371/journal.pone.0227816 on a Taguchi L9 orthogonal array design, after DP94 culture in TY medium using two differ- Editor: Yoshihiro Uesawa, Meiji Pharmaceutical ent vessels (bioreactor or Erlenmeyer flasks). Various culture parameters such as tempera- University, JAPAN ture, pH and inoculum ratio (%) were studied. For experiments conducted in a bioreactor, Received: July 4, 2019 stirring speed was included as an additional parameter. Significant differences in the cyto- toxic activities of different extracts on B16 melanoma cancer cell lines, highlighted the influ- Accepted: December 30, 2019 ence of culture temperature on the production of cytotoxic compound(s) using a bioreactor. Published: January 14, 2020 A culture in Erlenmeyer flasks was also performed and afforded an increase of the produc- Peer Review History: PLOS recognizes the tion of the active compounds. The best conditions for the highest cytotoxicity (IC on B16: 6 benefits of transparency in the peer review ± 0.5μg/mL) and the highest yield (202.0 mg/L) were identified as: pH 6, temperature 37˚C process; therefore, we enable the publication of all of the content of peer review and author and 5% inoculum. responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0227816 Copyright:© 2020 Noe¨l et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which Introduction permits unrestricted use, distribution, and reproduction in any medium, provided the original Some antibiotics such as actinomycin, streptothricin and streptomycin have been discovered author and source are credited. since the forties . Actinobacteria also produce a wide variety of secondary metabolites with interesting biological activities such as insecticidal (e.g. avermectins from S. avermitilis ), Data Availability Statement: All relevant data are within the paper and its Supporting Information antifungal (e.g. amphotericin B from S. nodosus ), antiviral (e.g. hygromycin from S. hygro- files. scopicus ), and cytotoxic activities (e.g. bleomycin from Streptomyces verticillus  or astero- bactin from Nocardia asteroides ). These bacteria are also well described for their interesting Funding: The Funder “Ligue contre le Cancer 35” gave financial support (named BaLI) for the biotechnological applications . Actinobacteria are widespread in environment making them PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 1 / 14 Design of experiments for optimization of Nocardia culture purchase of a Bioreactor BioFlo 115®. The funders accessible sources of novel bioactive metabolites as we can see with Streptomyces and Nocardia had no role in study design, data collection and strains isolated from underground lake and moonmilk . analysis, decision to publish, or preparation of the Microbial population associated with lichens have been recently characterized in the lichen manuscript. symbiosis classically defined like a dual symbiosis between a green alga (or a cyanobacterium) Competing interests: The authors have declared and a fungus . Indeed, a plethora of bacteria have been identified from lichens using cul- that no competing interests exist. ture-independent (reviewed by Suzuki et al ) or culture-dependent approaches [11,12] and Abbreviations: OSMAC, One Strain Many their bacterial community seems to be specific to the lichen species . This long-lasting eco- Compounds; TY, Tryptone Yeast medium; LBm, logical niche is a novel reservoir for interesting bacterial strains. The three main bacterial modified Lysogeny Broth; MB, Marine Broth; ISP2, phyla associated with lichens are Proteobacteria, Firmicutes and Actinobacteria [12–14]. The International Streptomyces Project 2; IC , half chemical production of several strains associated to lichens has already been described [10,15] maximal inhibitory concentration; ANOVA, Analysis highlighting their ability to produce active compounds. of Variance; DO, Dissolved Oxygen; DOE, Design of Experiments; DMSO, Dimethylsulfoxyde; NaCl, In this study, an Actinobacterium isolated from a terrestrial lichen species Lathagrium auri- sodium chloride; MeOH, methanol; EtOAc, ethyl forme  (ex Collema auriforme), close to Nocardia sp. (100% 16S rRNA sequence similarity acetate; RE, resin extract; SE, supernatant extract; with Nocardia soli, N. cummidelens and N. salmonicida and 98.59% with N. ignorata), was NaOH, sodium hydroxide; H SO , sulfuric acid; OD, 2 4 selected. To our knowledge only our previous study has described the chemical production of Optical density; RPMI, Roswell Park Memorial this Nocardia strain . In our attempts to focus on the discovery of novel agents to treat mel- Institute medium; FCS, fetal calf serum; MTT, 3- anoma, one of the most fastest growing forms of cancer, we have selected the culture broth [4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; HPLC, High Performance Liquid extract of this bacterium which showed an interesting cytotoxic activity against murine B16 Chromatography; vvm, volume of gas per volume melanoma cell lines (IC = 23 ± 3 μg/mL). In the previous work, the cytotoxic compound(s) of liquid per minute. involved in the biological activity of this strain were not identified. In this context, the optimi- zation of culture conditions leading to the increase of the cytotoxicity of the bacterial broth was necessary to highlight bioactive compound(s). The aim of this study was to estimate the impact of different culture parameters on this cytotoxicity. We have chosen herein to evaluate various parameters: pH, temperature, stirring speed and the inoculum ratio (%) (v/v). The evaluation of all the possible combinations of these parameters requires 81 experiments, which is very time consuming. A Taguchi L9 orthogonal array design, which is a robust statistical DOE (Design of Experiments) method, allowing to study a set of variables with a limited num- ber of trials, was employed to perform 9 experiments instead of 81 . At the end of the sta- tionary growth phase the cultures were stopped and we determined the yield of extraction and the cytotoxic activity of the extracts against a cancer cell line (B16) and a non-cancer cell line (HaCaT) which will eventually led us to determine a selectivity of action. ANOVA (Analysis Of Variance) test was used to highlight which parameters exhibited a significant influence on the cytotoxicity of the bacterial extracts. Finally, the chemical analyses of the most active extracts were also carried out using HPLC (High Performance Liquid Chromatography) in order to highlight the common metabolites. Materials and methods Microorganism The strain used for this study was isolated from a terrestrial lichen Lathagrium auriforme (ex Collema auriforme) collected in Kesselfallklamm in Austria (47˚12’21.26” N, 15˚23’57.27” E) in November 2012 by Parrot et al . Its 16S rRNA gene was then sequenced using Sanger sequencing (874 pb) and the close phylogenetic neighbors of our strain were identified as Nocardia ignorata DQ659907 at 98.59% sequence identity by comparison of these data with sequences in the Eztaxon server type strain database . A second sequencing of 993 pb fol- lowing the same protocol as described by Parrot et al  allowed the identification of three closer phylogenetic neighbors Nocardia soli, N. cummidelens and N. salmonicida at 100% sequence identity using the Eztaxon server. The strain was stored after growth in ISP2 medium  with 50% v/v glycerol or 5% v/v DMSO (Dimethylsulfoxyde) at −80˚C and referenced as PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 2 / 14 Design of experiments for optimization of Nocardia culture DP94 . This strain was deposited at the Institute of Plant Sciences, University of Graz, Austria. Preliminary assays Precultures. The precultures were obtained by inoculation of 30 mL of medium with an isolated colony collected on an agar plate. The culture was incubated until the optical density reached 0.6 for further uses. A purity control was performed on agar plate to check the absence of contamination. Small-scale cultures. Four media cited below were used for preliminary assays: • TY (Tryptone Yeast medium): 10 g/L yeast extract (Sigma-Aldrich, St Louis, MO, USA), 16 g/L tryptone (Sigma-Aldrich, St Louis, MO, USA), and 5 g/L NaCl (Sigma-Aldrich, St Louis, MO, USA), • Modified LB (Lysogeny Broth) (LBm): 5 g/L peptone (Sigma-Aldrich, St Louis, MO, USA), 5 g/L NaCl (Sigma-Aldrich, St Louis, MO, USA) and 3 g/L yeast extract (Sigma-Aldrich, St Louis, MO, USA) • ISP2 (International Streptomyces Project medium 2): 4 g/L yeast extract (Sigma-Aldrich, St Louis, MO, USA), 10 g/L malt extract (Sigma-Aldrich, St Louis, MO, USA) and 4 g/L glucose • MB (Marine Broth): 37.4 g/L of commercial marine broth (Difco 1 2216, ThermoFisher, Waltham, MA, USA) 30 mL of each medium were inoculated in duplicate with the bacterial strain and incubated at 25˚C and at 110 rpm in an incubator New Brunswick1 Innova 42 (New Brunswick1, Edi- son, NJ, USA). After 8, 11 or 14 days of incubation the culture broth was centrifuged at 3000 rpm (Thermo scientific Sorvall ST40R, ThermoFisher scientific, Waltham, MA, USA), at 4˚C for 15 min. The supernatant was extracted with 2 x 30 mL of ethyl acetate and the organic layer was dried on anhydrous sodium sulfate and the solvent was evaporated in vacuo leading to a raw extract. Scale-up process. 5.4 L of TY medium and 7.2 L of LBm medium were divided in 500 mL Erlenmeyer flasks and inoculated with 1% of preculture of the bacterial strain. The cultures were incubated at 25˚C, 110 rpm in an incubator New Brunswick1 Innova 42 (New Bruns- wick1, Edison, NJ, USA). After 11 days of incubation the culture broth was extracted as described below. The monitoring of the bacterial growth of these scale-up cultures is reported in S1 Fig. Culture extraction After fermentation, the broth was collected and centrifuged at 3000 rpm (Thermo scientific Sorvall ST40R, ThermoFisher scientific, Waltham, MA, USA), at 4˚C for 15 min. The superna- tant was collected and 40 g of a resin Amberlite1 XAD 7HP (Sigma-Aldrich, St Louis, MO, USA) were added per liter. After 4h of stirring at 150 rpm the resin was filtered and desorbed three times successively with 400 mL of a mixture of MeOH/acetone (1/1, v/v) per 40 g of resin. The solvents were evaporated in vacuo and the residue was dissolved in 100 mL of water and extracted with 2 x 100 mL of EtOAc. The organic layer was then dried with anhydrous sodium sulfate, filtered and evaporated in vacuo leading to the resin extract (RE). The superna- tant previously treated with the resin was extracted two times with EtOAc (1/2, v/v) and the organic layer was dehydrated with anhydrous sodium sulfate, filtered and dried in vacuo PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 3 / 14 Design of experiments for optimization of Nocardia culture leading to the supernatant extract (SE). An extraction yield was defined as follows: masse of the extract ðmgÞ Yield ¼ volume of culture mediumðLÞ Culture in bioreactor Culture was performed in 2.5 L of TY medium in a bioreactor BioFlo/Celligen1 115 (New Brunswick1, Edison, NJ, USA). The oxygenation rate of the medium was regulated at 0.4 vvm (volume of gas per volume of liquid per minute) and the percentage of dissolved oxygen was maintained up to 40% by using the cascade mode with the stirring. Temperature, pH, stirring and inoculum ratio (%) (v/v) were selected according to the design of experiments (Tables 1 and 2). The cultures were stopped after 5 days of fermentation. An example of monitoring of culture parameters and of bacterial growth has been reported in S2 and S3 Figs. Culture in Erlenmeyer flasks Culture of the strain was performed in 300 mL of TY medium in 500 mL Erlenmeyer flask. The pH of the medium was adjusted with a solution of NaOH 5 M or H SO 0.5 N. Flasks 2 4 were then incubated in an incubator New Brunswick1 Innova 42 (New Brunswick1, Edison, NJ, USA) with a stirring fixed at 150 rpm and a temperature fixed at a value according to the design of experiments (See Tables 3 and 4). The cultures were stopped after 11 days of fermen- tation. The monitoring of bacterial growth was performed by measurement of OD (Optical Density) as shown in S4 Fig. Taguchi experimental design The Taguchi method was selected to establish the experimental design and led to 9 experi- ments either in bioreactor (Tables 1 and 2) nor in Erlenmeyer flask (Tables 3 and 4). For each experiment, the responses evaluated were the yield and the cytotoxicity of extracts against B16 and HaCaT cell lines. For the fermentation in bioreactor we selected 4 independent parameters and attributed 3 levels for each. The four parameters selected were temperature (25˚C, 30˚C and 37˚C), pH (6, 7 and 8), percentage of inoculum (v/v) (1%, 2% and 5%) and stirring speed (150, 200 and 250 rpm) (Table 1). The experimental design and ANOVA analysis were per- formed using Excel 2016. For the fermentation in Erlenmeyer flasks the stirring was fixed at 150 rpm and only 3 parameters were selected with 3 levels for each (Tables 3 and 4). Cytotoxic assays The cytotoxic activity of the resin extracts and the supernatant extracts was evaluated for each condition of the experimental design with a standard tetrazolium assay (MTT) . Two cell lines (B16: murine melanoma, HaCaT: human keratinocytes) were cultivated in RPMI (Ros- well Park Memorial Institute) 1640 medium (Thermo Fisher Scientific, USA) supplemented Table 1. Assignment of factors and level setting of the orthogonal array design L9 (3 ) for cultures in bioreactor. Factors Level 1 Level 2 Level 3 [A] Temperature (˚C) 25 30 37 [B] pH 6 7 8 [C] inoculum ratio % 1 2 5 [D] stirring speed (rpm) 150 200 250 https://doi.org/10.1371/journal.pone.0227816.t001 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 4 / 14 Design of experiments for optimization of Nocardia culture Table 2. Assignment of the experimental conditions in the orthogonal design L9 (3 ) for cultures in bioreactor. Exp n˚ Factor [A] [B] [C] [D] B1 1 1 1 1 B2 2 1 2 2 B3 3 1 3 3 B4 1 2 2 3 B5 2 2 3 1 B6 3 2 1 2 B7 1 3 3 2 B8 2 3 1 3 B9 3 3 2 1 [A]: Temperature, [B]: pH, [C]: inoculum ratio (%), [D]: stirring https://doi.org/10.1371/journal.pone.0227816.t002 with 5% fetal calf serum (FCS) and 1% antibiotics (penicillin 10 000 UI/ml and streptomycin 10 000 μg/ml, Eurobio1, Les Ulis, France). Cells were then seeds in 96-wells plate (B16: 6000 cells/well and HaCaT 10,000 cells/well) at day 0 and incubated for 24h at 37˚C and 5% CO . Extracts were then added at different concentrations (1, 10, 50, 100, and 200 μg/mL) for all the experiments. For experimental design in bioreactor and Erlenmeyer flasks the extracts were tested a second time at 5, 50, 100, 250 and 500 μg/mL to expand the range of IC values. After 24h of incubation (37˚C, 5% CO ) the cell viability was measured at 540 nm using a MTT (3- [4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay . Each experiment was repeated three times. Chemical analyses Chemical analyses of the most active extracts were carried out on a HPLC system—Diode Array Detector (LC-DAD) (Shimadzu, Marne La Vallee, France). A Prevail C18 column (5 μm, 250 × 4.6 mm, GRACE, Columbia, MD, USA) was used and a gradient system was applied: A (0.1% formic acid in water) and B (0.1% formic acid in acetonitrile). The following gradient was applied at a flow rate of 0.8 mL/min in the HPLC system: initial: 100% (A); from 0 to 5 min: 100% (A); from 5 to 35 min: 100% (A)/0% (B) to 0% (A)/100% (B); from 35 to 45 min: 100% B; from 45 to 50 min: 0% (A)/100% (B) to 100% (A)/0% (B); from 50 to 55 min: 100% (A). Samples were prepared by dissolving extracts in MeOH at 1 mg/mL and 20 μL were injected after filtration (45 μm). The LabSolutions software (Shimadzu, Marne La Vallee, France) was used for data analyses. Results Preliminary assays The Nocardia sp. strain used in this study achieved the stationary growth phase after 7 days of growth. We have cultivated this strain using 30 mL of four different media in 50 mL tubes, Table 3. Assignment of factors and level setting of the orthogonal array design L9 (3 ) for cultures in Erlenmeyer flasks. Factors Level 1 Level 2 Level 3 [A] Temperature (˚C) 25 30 37 [B] pH 6 7 8 [C] inoculum ratio % 1 2 5 https://doi.org/10.1371/journal.pone.0227816.t003 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 5 / 14 Design of experiments for optimization of Nocardia culture Table 4. Experimental conditions in the orthogonal design L9 (3 ) for cultures in Erlenmeyer flasks. Exp n˚ Factor [A] [B] [C] E1 1 1 1 E2 1 2 2 E3 1 3 3 E4 2 1 2 E5 2 2 3 E6 2 3 1 E7 3 1 3 E8 3 2 1 E9 3 3 2 [A]: Temperature, [B]: pH, [C]: inoculum ratio (%) https://doi.org/10.1371/journal.pone.0227816.t004 with a 7 days of exponential growth phase followed by 1, 4 or 7 days of stationary phase (respectively 8, 11 and 14 days in total). The best cytotoxic activity against B16 cells was achieved with the culture in LBm medium for 11 days (IC = 8 ± 2 μg/mL) and with the ISP2 medium after 14 days of culture (IC = 7.5 ± 0.5 μg/mL). Similar activities have been shown on HaCaT cells mentioning no selectivity of action. The highest amount of extract was obtained when the strain was grown in TY medium (Table 5). A scale-up assay with 300 mL of media in 500 mL Erlenmeyer flasks was then performed using LBm and TY for 11 days of cul- ture (4 days of stationary phase). In order to have sufficient amount of LBm extracts and based on the results obtained with the 30 mL process, we have inoculated 24 Erlenmeyer flasks with LBm medium (total volume 7.2 L) and 18 Erlenmeyer flasks with TY medium (total volume 5.4 L). The culture extract from TY medium gave the best activity against B16 cell line (IC = 4.6 ± 1.8 μg/mL) even though the amount obtained was low (yield = 16.1 mg/L) but higher than for the culture in LBm (Table 6). These culture conditions (TY medium and 4 days of sta- tionary phase) were applied to the Taguchi design experiments. Experimental design in bioreactor An experimental design was applied in order to find the best culture conditions to improve the yield and the cytotoxic activity of extracts. Varying temperature [A] (25, 30 and 37˚C), pH [B] (6, 7 and 8), percentage of inoculum ratio (% v/v) [C] (1, 2 and 5%) and stirring speed [D] (150, 200 and 250 rpm), 9 experiments were carried out. In all experiments, the yield of the resin extracts (RE) and the supernatant extracts (SE) were determined and their cytotoxicity was measured against B16 and HaCaT cell lines (Table 7). When similar activities have been exhibited against the two cell lines, we have focused on activities against B16. The mean IC values of RE on B16 cells, were classified into three groups: a first group with IC > 300 μg/mL (exp. B5 and B8); a second group with 100 < IC < 300 μg/mL (exp. 50 50 B1, B2 and B7) and the last group with IC < 100 μg/mL (exp. B3, B4, B6 and B9) (Table 7). In a similar manner, the results were classified into three groups for SE extracts based on their activity on B16: the first group with IC > 400 μg/mL (exp. B1, B2 and B7); a second group with 100 < IC < 400 μg/mL (exp. B4, B6 and B9) and the last group with IC < 100 μg/mL 50 50 (exp. B3, B5 and B8) (Table 7). Most of RE extracts exhibited cytotoxic activities whereas only a few of SE extracts were active. Main-effect plots (Fig 1) indicated that the IC values decreased with the maximum PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 6 / 14 Design of experiments for optimization of Nocardia culture Table 5. Weight and cytotoxicity of extracts in preliminary assays. Days of culture Medium Raw extract amount (mg) IC (μg/mL) B16 HaCaT 8 ISP2 2.3 ± 0.0 45 ± 5 59 ± 5 TY 5.2 ± 0.1 41 ± 5 72 ± 7 MB 2.2 ± 0.3 105 ± 20 115 ± 15 LBm 2.5 ± 0.4 25 ± 4 95 ± 30 11 ISP2 4.0 ± 0.7 31 ± 10 64 ± 28 TY 5.4 ± 0.9 25 ± 4 47 ± 13 MB 2.1 ± 0.7 57 ± 18 120 ± 15 LBm 2.5 ± 0.1 8 ± 2 31 ± 12 14 ISP2 3.6 ± 0.2 7.5 ± 0.5 41 ± 5 TY 6.4 ± 0.3 35 ± 8 86 ± 42 MB 3.8 ± 0.2 85 ± 10 79 ± 9 LBm 4.2 ± 0.2 14 ± 6 70 ± 4 ISP2 = International Streptomyces Project 2, TY = Tryptone Yeast, MB = Marine Broth, LBm = modified Lysogeny Broth https://doi.org/10.1371/journal.pone.0227816.t005 level of factor B at 37˚C. It was also noticed that the IC values increased in culture extracts at 30˚C, highlighting the importance of culture temperature on the cytotoxicity of the extracts. Finally, the experiment with the highest cytotoxic activity was the experiment B3 (RE, IC = 53 ± 16 μg/mL and SE IC = 41 ± 8 μg/mL on B16) with the following culture condi- 50 50 tions: pH 6, temperature 37˚C, an inoculum of 5% and a stirring of 250 rpm. Surprisingly, the experiment B2 (culture conditions: pH 6, temperature of 30˚C, inoculum of 2% and stirring speed of 200 rpm) gave the highest supernatant extraction (SE) yield. Based on the cytotoxic activity against B16 and the extraction yield, ANOVA test was used to highlight which parameter(s) could significantly influence the production of active com- pound(s). Statistical analysis. The F-ratio and the p-value determined from the ANOVA analysis on the cytotoxicity of the resin extracts showed that the temperature significantly influenced the activity (p < 0.05) (Table 8 and Fig 1). Conversely, none of the parameters selected for this analysis showed a significant effect on the extraction yield as well as for the cytotoxicity of the supernatant extracts (Table 9). Validation of the optimal conditions. The experiment B3 with the following parameters, pH 6, temperature 37˚C, inoculum 5% and stirring 250 rpm, was repeated once using bioreac- tor to confirm the previous results on HaCaT and B16. The repetition of this experiment showed close values of IC for SE on B16 and HaCaT cell lines (IC = 33 ± 8 μg/mL on B16 50 50 and IC = 23 ± 5 μg/mL on HaCaT) and a decrease of the activity of RE against B16 (IC > 50 50 200 μg/mL). This experiment confirms nevertheless that the culture conditions used are opti- mal for the cytotoxicity of the extracts from bioreactor. Table 6. Extraction yield and cytotoxicity of extracts from scale up in Erlenmeyer flasks. Medium Extract Yield (mg/L) IC (μg/mL) B16 HaCaT TY Resin extract 101.4 >200 >200 Supernatant extract 16.1 4.6 ± 1.8 8 ± 1.5 LBm Resin extract 60.3 >200 >200 Supernatant extract 3.0 52 ± 20 73 ± 20 https://doi.org/10.1371/journal.pone.0227816.t006 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 7 / 14 Design of experiments for optimization of Nocardia culture Table 7. Extraction yield and cytotoxicity of the extracts from the experimental design using bioreactor. Exp n˚ RE SE Yield (mg/L) IC (μg/mL) Yield (mg/L) IC (μg/mL) 50 50 B16 HaCaT B16 HaCaT B1 146.1 157 ± 77 92 ± 29 46.3 >500 460 ± 140 B2 160.5 205 ± 100 185 ± 65 1859.3 >500 >500 B3 216.6 53 ± 16 152 ± 28 478.6 41 ± 8 20 ± 8 B4 158.0 92 ± 19 70 ± 8 152.3 250 320 B5 177.8 345 ± 95 120 ± 40 506.9 100 ± 30 150 B6 223.5 45 ± 15 80 ± 20 36.1 270 ± 35 225 ± 45 B7 126.8 270 ± 30 185 ± 15 66.5 430 ± 90 180 ± 20 B8 703.0 340 ± 25 397 ± 27 18.4 67 ± 4 38 ± 19 B9 318.0 90 ± 50 110 ± 25 43.9 350 270 https://doi.org/10.1371/journal.pone.0227816.t007 Experimental design in Erlenmeyer flasks The data obtained in Erlenmeyer flasks were analyzed in a same manner using a Taguchi array and the yields and cytotoxic activities of the both extracts (RE and SE) were measured on B16 and HaCaT cell lines (Table 10). The most important cytotoxic activity was demonstrated with experiment E7 (RE, IC = 6 ± 0.5 μg/mL on B16) with the following culture conditions: pH 6, temperature 37˚C and 5% inoculum and led to a sufficient amount of extract (202.0 mg/L) (Table 10). Interestingly, these conditions were close to those obtained for the culture in bioreactor (excepted for the value of stirring fixed at 150 rpm for Erlenmeyer process), confirming that these parameters are the best to improve the cytotoxicity of the extracts. Conversely, the highest extraction yield (840.3 mg/mL) was obtained with E8 and the conditions (37˚C, pH 7 and inoculum 1%, Table 10) were completely different than those obtained for culture in bioreactor. Nevertheless, the ANOVA analysis of the results obtained from these Erlenmeyer cultures showed that none of the selected parameters had a significant effect on the cytotoxicity of the extracts on B16 cell lines or on the extraction yield. Fig 1. Main-effect plots for optimization of culture conditions in bioreactor. https://doi.org/10.1371/journal.pone.0227816.g001 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 8 / 14 Design of experiments for optimization of Nocardia culture Table 8. Summary of analysis of variance in the ANOVA test for cytotoxicity on B16 of RE from bioreactor. Factors Sum of squares f Average of square F-ratio F0.95(2,2) p A 14804.2 2 7402.1 7.46 19 0.118 B 82222.9 2 41111.4 41.44 19 0.024 C 1984.2 2 992.1 1.00 19 0.500 D 13206.9 2 6603.4 6.66 19 0.131 Error 1984.2 2 992.1 Total 112218.2 [A]: Temperature, [B]: pH, [C]: inoculum ratio (%), [D]: stirring https://doi.org/10.1371/journal.pone.0227816.t008 Impact of the culture vessel on the activity of extracts In order to determine the effect of the culture conditions on the production of cytotoxic metabolites, we compared the results obtained for three processes performed using either bio- reactor or Erlenmeyer flasks and possessing similar culture conditions with a stirring speed fixed at 150 rpm (Table 11). Among them, the most active extracts were obtained at 37˚C, pH 6 and a 2% inoculum ratio (experiments B9 and E9). The yields were also higher in these experiments in comparison to the other ones. Comparison of HPLC profiles of the most active extracts The HPLC chromatograms of these two experiments B9 and E9 were compared in Fig 2. Those of the most active extracts obtained during this study (SE of the large scale culture in TY shown in preliminary assays, SE from B3 and RE from E7) were reported in Fig 3. The com- pounds previously isolated from this Nocardia strain  and known compounds such as cyclo (L-Ala-L-Phe) , adenine, adenosine and 1-(5-deoxy-β-D-erythro-pent-4-enofurano- syl)  were identified by comparison with a standard and data from literature (Figs 2A, 2B and 3). The main differences between the production of metabolites by the strain in the RE of B9 and E9 are due to the presence of an unidentified compound with a r.t. of 19 min and of a diketopiperazine cyclo-(L-Ala-L-Phe) which were produced in large amount in bioreactor B9 but not in Erlenmeyer flask E9 (Fig 2A). Variability was also observed for the production of compounds eluted between 26 and 32 min (Fig 2A). In the SE of E9 a compound eluted at the beginning of the gradient appeared. Moreover, a difference of the production of compounds eluted between 22 and 28 min was also highlighted between the two culture vessels (Fig 2B). The production of adenine was most important in bioreactor than in Erlenmeyer reaction set- up (Fig 2). Despite these differences in the chemical profiles, the activity of the extracts were Table 9. Summary of p-value from the ANOVA test on all extracts from bioreactor. Factors p RE SE Effect on IC on B16 Effect on extraction yield Effect on IC on B16 Effect on extraction yield 50 50 A 0.118 0.408 0.500 0.303 B 0.024 0.475 0.333 0.312 C 0.500 0.490 0.192 0.500 D 0.131 0.500 0.388 0.385 [A]: Temperature, [B]: pH, [C]: inoculum ratio (%), [D]: stirring https://doi.org/10.1371/journal.pone.0227816.t009 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 9 / 14 Design of experiments for optimization of Nocardia culture Table 10. Extraction yield and cytotoxicity of the extracts on B16 and HaCaT cell lines for the experimental design in Erlenmeyer flasks. Exp n˚ RE SE Yield (mg/L) IC (μg/mL) Yield (mg/L) IC (μg/mL) 50 50 B16 HaCaT B16 HaCaT E1 306.7 280 ± 160 110 ± 60 199.3 375 ± 110 40 ± 15 E2 279.7 96 ± 9 122 ± 4 77.9 73 ± 20 77 ± 16 E3 218.7 70 ± 8 71 ± 19 66.0 74 ± 34 56 ± 24 E4 96.3 145 ± 15 122 ± 22 74.0 380 ± 105 75 ± 50 E5 184.0 130 ± 4 127 ± 13 57.3 390 ± 100 240 ± 150 E6 305.7 64 ± 8 72 ± 1 58.3 170 ± 10 152 ± 18 E7 202.0 6 ± 0.5 30 ± 7 583.3 480 ± 230 140 ± 70 E8 176.7 81 ± 10 50 ± 7 840.3 >500 260 ± 110 E9 489.0 63 ± 7 79 ± 15 39.7 137 ± 7 32 ± 10 https://doi.org/10.1371/journal.pone.0227816.t010 similar between bioreactor and Erlenmeyer flasks (except for the SE on B16 cells) meaning that the compounds highlighted above are not implicated in the cytotoxicity of the extracts. Finally, the comparison of the most active extracts led to the highlighting of unidentified compounds common in all the three active extracts which were indicated with black squares in Fig 3. Discussion Lichens are a novel ecological niche of considerable interest for the discovery of bacterial strains with biotechnological potential . We selected and studied a lichen-associated Acti- nobacterium, Nocardia sp that demonstrated interesting cytotoxic activities. The results detailed herein showed that the production of cytotoxic compounds by this Nocardia sp. could be modulated by culture conditions according to the OSMAC (One Strain/Many Compounds) approach previously described . The data obtained in preliminary assays are in accordance with our expectations. The marine broth currently used in our experiments performed on strains isolated from marine lichens [12,24] is not suitable to optimize the cytotoxicity of our Nocardia strain isolated from a terrestrial lichen. Moreover, the amount of production of metabolites increases with rich medium such as TY. This observation is close to those of Abdelmohsen et al  who indicated that the production of actinosporin by a marine sponge associated-Actinokineospora was widely affected depending on the culture medium used. Our results suggest in the same manner that the production of the compound(s) involved in the cytotoxicity of the extracts is influenced by the difference in the medium growth used. Table 11. Comparison of the results obtained for cultures in bioreactor and in Erlenmeyer flasks with the same culture conditions. Experiences Factors RE SE Temperature (˚C) pH Inoculum ratio (%) Yield (mg/L) IC (μg/mL) Yield (mg/L) IC (μg/mL) 50 50 B16 HaCaT B16 HaCaT B1 25 6 1 146.1 157 ± 77 92 ± 29 46.3 >500 460 ± 140 E1 25 6 1 306.7 280 ± 160 110 ± 60 199.3 375 ± 110 40 ± 15 B5 30 7 5 177.8 345 ± 95 120 ± 40 506.9 100 ± 30 39 ± 10 E5 30 7 5 184.0 130 ± 4 127 ± 13 57.3 390 ± 100 240 ± 150 B9 37 8 2 318.0 90 ± 50 110 ± 25 43.9 390 ± 90 45 ± 2 E9 37 8 2 489.0 63 ± 7 79 ± 15 39.7 137 ± 7 32 ± 10 https://doi.org/10.1371/journal.pone.0227816.t011 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 10 / 14 Design of experiments for optimization of Nocardia culture Fig 2. Comparison of chemical profiles of B9 (red) and E9 (blue). A. Resin extracts, B. Supernatant extracts. All samples were analysed at 220 nm on Prevail reversed phase C18 column with a gradient of H O (A)/acetonitrile (B) (10 min 100% of A, 30 min from 0% of B to 100% of B, 10 min 100% of B). https://doi.org/10.1371/journal.pone.0227816.g002 A significant effect of temperature on cytotoxicity of the resin extracts (p < 0.05) with an improvement of the activity at 37˚C was exhibited by ANOVA analysis of the data from exper- imental design in bioreactor. The importance of culture parameters such as temperature on increasing of the biomass and of the activity has been already demonstrated for Actinobacteria Fig 3. Comparison of the chemical profiles of the most active extracts. SE of the large scale culture in TY shown in preliminary assays (red), SE of B3 (blue) and RE of E7 (black). All samples were analysed at 220 nm on Prevail reversed phase C18 column with a gradient of H O (A)/acetonitrile (B) (10 min 100% of A, 30 min from 0% of B to 100% of B, 10 min 100% of B). https://doi.org/10.1371/journal.pone.0227816.g003 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 11 / 14 Design of experiments for optimization of Nocardia culture strains. Thus, the effect of the temperature has been described on the biomass and geosmin production by strains of N. cummidelens and N. fluminea. The highest biomass production was thus obtained at the warmest temperature tested, 30˚C and 25˚C, for the both strains and the production of geosmin was optimal at 25˚C . Similar effects has been reported on the production of rapamycin by a Streptomyces strain, where the optimal temperature was estab- lished at 23˚C, the lowest tested value . In spite of employing similar culture conditions (temperature 37˚C, pH 6 and inoculum 5%,) in both the bioreactor (B3) and Erlenmeyer flasks (E7), the best activity (IC = 6 ± 0.5 μg/mL on B16) was displayed by extracts obtained from Erlenmeyer flasks (E7) highlighting the importance of the culture process used. Various compounds have been previously isolated from the culture broth of this bacterial strain . Compounds identified were mainly diketopiperazines. An auxin derivative and purine derivatives (e.g. adenine) have also been isolated (data not published) but none of these compounds explained the cytotoxic activity of the bacterial extracts. The comparison of the chemical profiles of extracts obtained by culture in bioreactor and in Erlenmeyer flasks highlighted several differences which also revealed the impact of the sys- tem used. These variations of production could be explained by the different aeration and the stirring system used  leading to different stress conditions for bacteria. The comparison of the chemical profiles of the most active extracts (Fig 3) showed a very similar production and the major compounds already isolated possess no cytotoxic properties on B16 cell line. Nevertheless, some zones of the chromatogram corresponding to unidentified compounds have been highlighted (Fig 3). Conclusion In conclusion, the Taguchi method for design of experiments combined with an ANOVA test were shown to be powerful tools for the optimization of culture parameters of lichen-associ- ated Nocardia sp. for the production of cytotoxic compounds. Further separative experiments will be carried out with a focus on the zones highlighted on HPLC profiles during this study in order to identify compound(s) involved in the cytotoxic activity of these bacterial extracts. Supporting information S1 Fig. Monitoring of bacterial growth of scale-up in TY and LBm. (TIF) S2 Fig. Monitoring of culture parameters of B1. (TIF) S3 Fig. Monitoring of bacterial growth of B1. (TIF) S4 Fig. Monitoring of bacterial growth of E1. (TIF) Acknowledgments We sincerely thank L. Intertaglia (LBBM, Banyuls/mer) for the identification of the strain DP94 and Dr S. Sabbani for the reading and the improvement of English langage. PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 12 / 14 Design of experiments for optimization of Nocardia culture Author Contributions Conceptualization: Alba Noe ¨l, Sophie Tomasi. Formal analysis: Eric Hitti. Investigation: Alba Noe ¨l, Gwendoline Van Soen, Isabelle Rouaud. Methodology: Isabelle Rouaud. Supervision: Sophie Tomasi. Validation: Eric Hitti, Sophie Tomasi. Writing – original draft: Alba Noe ¨l, Sophie Tomasi. Writing – review & editing: Sophie Tomasi. References 1. Barka EA, Vatsa P, Sanchez L, Gaveau-Vaillant N, Jacquard C, Klenk H-P, et al. Taxonomy, Physiol- ogy, and Natural Products of Actinobacteria. Microbiol Mol Biol Rev. 2016; 80: 1–43. https://doi.org/10. 1128/MMBR.00019-15 PMID: 26609051 2. Burg RW, Miller BM, Baker EE, Birnbaum J, Currie SA, Hartman R, et al. Avermectins, New Family of Potent Anthelmintic Agents: Producing Organism and Fermentation. Antimicrob Agents Chemother. 1979; 15: 361–367. https://doi.org/10.1128/aac.15.3.361 PMID: 464561 3. Linke HA, Mechlinski W, Schaffner CP. Production of amphotericin B-14C by Streptomyces nodosus fermentation, and preparation of the amphotericin B-14C-methyl-ester. J Antibiot (Tokyo). 1974; 27: 155–160. 4. Lacal JC, Va ´ zquez D, Fernandez-Sousa JM, Carrasco L. Antibiotics that specifically block translation in virus-infected cells. J Antibiot (Tokyo). 1980; 33: 441–446. 5. Umezawa H, Ishizuka M, Maeda K, Takeuchi T. Studies on bleomycin. Cancer. 1967; 20: 891–895. https://doi.org/10.1002/1097-0142(1967)20:5<891::aid-cncr2820200550>3.0.co;2-v PMID: 5337399 6. Nemoto A, Hoshino Y, Yazawa K, Ando A, Mikami Y, Komaki H, et al. Asterobactin, a new siderophore group antibiotic from Nocardia asteroides. J Antibiot (Tokyo). 2002; 55: 593–597. 7. Bilyk O, Luzhetskyy A. Metabolic engineering of natural product biosynthesis in actinobacteria. Curr Opin Biotechnol. 2016; 42: 98–107. https://doi.org/10.1016/j.copbio.2016.03.008 PMID: 27058643 8. Axenov-Gibanov DV, Voytsekhovskaya IV, Tokovenko BT, Protasov ES, Gamaiunov SV, Rebets Y, et al. Actinobacteria Isolated from an Underground Lake and Moonmilk Speleothem from the Biggest Conglomeratic Karstic Cave in Siberia as Sources of Novel Biologically Active Compounds. PLOS ONE. 2016; 11. https://doi.org/10.1371/journal.pone.0149216 PMID: 26901168 9. Aschenbrenner IA, Cernava T, Berg G, Grube M. Understanding Microbial Multi-Species Symbioses. Front Microbiol. 2016; 7. https://doi.org/10.3389/fmicb.2016.00180 PMID: 26925047 10. Suzuki MT, Parrot D, Berg G, Grube M, Tomasi S. Lichens as natural sources of biotechnologically rele- vant bacteria. Appl Microbiol Biotechnol. 2016; 100: 583–595. https://doi.org/10.1007/s00253-015- 7114-z PMID: 26549239 11. Selbmann L, Zucconi L, Ruisi S, Grube M, Cardinale M, Onofri S. Culturable bacteria associated with Antarctic lichens: affiliation and psychrotolerance. Polar Biol. 2010; 33: 71–83. https://doi.org/10.1007/ s00300-009-0686-2 12. Parrot D, Antony-Babu S, Intertaglia L, Grube M, Tomasi S, Suzuki MT. Littoral lichens as a novel source of potentially bioactive Actinobacteria. Sci Rep. 2015; 5: 15839. https://doi.org/10.1038/ srep15839 PMID: 26514347 13. Bates ST, Cropsey GWG, Caporaso JG, Knight R, Fierer N. Bacterial Communities Associated with the Lichen Symbiosis. Appl Environ Microbiol. 2011; 77: 1309–1314. https://doi.org/10.1128/AEM.02257- 10 PMID: 21169444 14. Grube M, Cardinale M, de Castro JV, Mu ¨ ller H, Berg G. Species-specific structural and functional diver- sity of bacterial communities in lichen symbioses. ISME J. 2009; 3: 1105–1115. https://doi.org/10.1038/ ismej.2009.63 PMID: 19554038 15. Parrot D, Legrave N, Delmail D, Grube M, Suzuki M, Tomasi S. Review–Lichen-Associated Bacteria as a Hot Spot of Chemodiversity: Focus on Uncialamycin, a Promising Compound for Future Medicinal PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 13 / 14 Design of experiments for optimization of Nocardia culture Applications. Planta Med. 2016; 82: 1143–1152. https://doi.org/10.1055/s-0042-105571 PMID: 16. Noe ¨ l A, Ferron S, Rouaud I, Gouault N, Hurvois J-P, Tomasi S. Isolation and Structure Identification of Novel Brominated Diketopiperazines from Nocardia ignorata—A Lichen-Associated Actinobacterium. Molecules. 2017; 22: 371. https://doi.org/10.3390/molecules22030371 PMID: 28264516 17. Bhattacharya A, Das S, Majumder P, Batish A. Estimating the effect of cutting parameters on surface finish and power consumption during high speed machining of AISI 1045 steel using Taguchi design and ANOVA. Prod Eng. 2009; 3: 31–40. 18. Kim O-S, Cho Y-J, Lee K, Yoon S-H, Kim M, Na H, et al. Introducing EzTaxon-e: a prokaryotic 16S rRNA gene sequence database with phylotypes that represent uncultured species. Int J Syst Evol Microbiol. 2012; 62: 716–721. https://doi.org/10.1099/ijs.0.038075-0 PMID: 22140171 19. Shirling EB, Gottlieb D. Methods for characterization of Streptomyces species1. Int J Syst Evol Micro- biol. 1966; 16: 313–340. https://doi.org/10.1099/00207713-16-3-313 20. Be ´ zivin C, Tomasi S, Lohe ´ zic-Le De ´ ve ´ hat F, Boustie J. Cytotoxic activity of some lichen extracts on murine and human cancer cell lines. Phytomedicine. 2003; 10: 499–503. https://doi.org/10.1078/ 094471103322331458 PMID: 13678234 21. Cho JY, Kang JY, Hong YK, Baek HH, Shin HW, Kim MS. Isolation and Structural Determination of the Antifouling Diketopiperazines from Marine-Derived Streptomyces praecox 291–11. Biosci Biotechnol Biochem. 2012; 76: 1116–1121. https://doi.org/10.1271/bbb.110943 PMID: 22790932 22. Perrone P, Daverio F, Valente R, Rajyaguru S, Martin JA, Le ´ vêque V, et al. First Example of Phosphor- amidate Approach Applied to a 4‘-Substituted Purine Nucleoside (4‘-Azidoadenosine): Conversion of an Inactive Nucleoside to a Submicromolar Compound versus Hepatitis C Virus. J Med Chem. 2007; 50: 5463–5470. https://doi.org/10.1021/jm070362i PMID: 17914786 23. Bode HB, Bethe B, Ho ¨ fs R, Zeeck A. Big Effects from Small Changes: Possible Ways to Explore Nature’s Chemical Diversity. ChemBioChem. 2002; 3: 619–627. https://doi.org/10.1002/1439-7633 (20020703)3:7<619::AID-CBIC619>3.0.CO;2-9 PMID: 12324995 24. Parrot D, Intertaglia L, Jehan P, Grube M, Suzuki MT, Tomasi S. Chemical analysis of the Alphaproteo- bacterium strain MOLA1416 associated with the marine lichen Lichina pygmaea. Phytochemistry. 2018; 145: 57–67. https://doi.org/10.1016/j.phytochem.2017.10.005 PMID: 29091816 25. Abdelmohsen UR, Cheng C, Viegelmann C, Zhang T, Grkovic T, Ahmed S, et al. Dereplication Strate- gies for Targeted Isolation of New Antitrypanosomal Actinosporins A and B from a Marine Sponge Associated-Actinokineospora sp. EG49. Mar Drugs. 2014; 12: 1220–1244. https://doi.org/10.3390/ md12031220 PMID: 24663112 26. Schrader KK, Harries MD, Page PN. Temperature effects on biomass, geosmin, and 2-methylisobor- neol production and cellular activity by Nocardia spp. and Streptomyces spp. isolated from rainbow trout recirculating aquaculture systems. J Ind Microbiol Biotechnol. 2015; 42: 759–767. https://doi.org/ 10.1007/s10295-015-1600-2 PMID: 25724337 27. Dutta S, Basak B, Bhunia B, Sinha A, Dey A. Approaches towards the enhanced production of Rapamy- cin by Streptomyces hygroscopicus MTCC 4003 through mutagenesis and optimization of process parameters by Taguchi orthogonal array methodology. World J Microbiol Biotechnol. 2017; 33: 90. https://doi.org/10.1007/s11274-017-2260-3 PMID: 28390015 28. Jia B, Jin ZH, Mei LH. Medium Optimization Based on Statistical Methodologies for Pristinamycins Pro- duction by Streptomyces pristinaespiralis. Appl Biochem Biotechnol. 2008; 144: 133. https://doi.org/10. 1007/s12010-007-8012-3 PMID: 18456945 PLOS ONE | https://doi.org/10.1371/journal.pone.0227816 January 14, 2020 14 / 14
PLoS ONE – Pubmed Central
Published: Jan 14, 2020
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera