ARTICLE IN PRESS

Systematic and Applied Microbiology 32 (2009) 56–64 www.elsevier.de/syapm

Isolation of two strains producing pseudomonic acid A Eva Fritza, Agnes Feketeb, Jutta Lintelmannb, Philipe Schmitt-Kopplinb, Rainer U. Meckenstocka,Ã aInstitute of Groundwater Ecology, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolsta¨dter Landstrasse 1, 85764 Neuherberg, Germany bInstitute of Ecological Chemistry, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolsta¨dter Landstrasse 1, 85764 Neuherberg, Germany

Received 16 June 2008

Abstract

Two novel Pseudomonas strains were isolated from groundwater sediment samples. The strains showed resistance against the antibiotics tetracycline, cephalothin, nisin, vancomycin, nalidixic acid, erythromycin, lincomycin, and penicillin and grew at temperatures between 15 and 37 1C and pH values from 4 to 10 with a maximum at pH 7 to 10. The 16S ribosomal RNA gene sequences and the substrate spectrum of the isolates revealed that the two strains belonged to the Pseudomonas fluorescens group. The supernatants of both strains had an antibiotic effect against Gram-positive and one Gram-negative strain. The effective substance was produced under standard cultivation conditions without special inducer molecules or special medium composition. The antibiotically active compound was identified as pseudomonic acid A by off-line high performance liquid chromatography (HPLC) and Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). The measurement on ultra performance liquid chromatography (UPLC, UV–vis detection) confirmed the determination of pseudomonic acid A which was produced by both strains at 1.7–3.5 mg/l. Our findings indicate that the ability to produce the antibiotic pseudomonic acid A (Mupirocin) is more spread among the pseudomonads then anticipated from the only producer known so far. r 2008 Elsevier GmbH. All rights reserved.

Keywords: Mupirocin; Pseudomonas fluorescens; Environmental isolates; FT-ICR-MS; UPLC; HPLC; Antibiotic

Introduction Pseudomonas fluorescens strain NCIB 10586 produces the antibiotic pseudomonic acid that was introduced to The ability of Pseudomonas strains to inhibit other the market in 1986 as mupirocin for human use. The bacteria was well known already at the end of the 19th effective substance pseudomonic acid A was described century [2], and turned out to play an important role [13] and later it could be shown that derivatives of especially in the case of bacterial biocontrol in the pseudomonic acid are produced simultaneously by the rhizosphere of plants [31]. As the only known isolate, same organism [5,6,12]. The most important and major form is pseudomonic acid A also known as mupirocin ÃCorresponding author. Tel.:+49 89 3187 2561; fax:+49 89 3187 3361. and used as a topical antibiotic [34]. This polyketide E-mail address: [email protected] antibiotic targets isoleucyl-tRNA synthase, where it (R.U. Meckenstock). competes with isoleucine for the binding site and inhibits

0723-2020/$ - see front matter r 2008 Elsevier GmbH. All rights reserved. doi:10.1016/j.syapm.2008.11.001 ARTICLE IN PRESS E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64 57 the protein synthesis of sensitive bacterial strains. It is from a pellet of 5 ml culture. Full length 16S rRNA also useful against methicillin-resistant Staphylococcus genes were PCR-amplified using primers Ba27f aureus (MRSA). For all antibiotics known so far, (sequence: AGA GTT TGA TCM TGG CTC AG) bacteria developed a resistance and thus also mupirocin and 907r (sequence: CCG TCA ATT CMT TTR AGT resistant strains of MRSA were detected [7]. The TT) [21,35]. The PCR mixture (50 ml) consisted of 5.0 ml resistance mechanism is given by a mutation of the 10 Â PCR-buffer (Fermentas), 3.0 ml 25 mM MgCl2, isoleucyl-tRNA synthetase or by an additional new 0.5 ml20mg/ml BSA, 0.5 ml 10 mM dNTPs, 0.5 ml50mM isoleucyl-tRNA synthetase [7]. The producer strain f-primer and 0.5 ml50mM r-primer, 0.25 ml5U/mlMBI Pseudomonas fluorescens NCIB10586 has two different Taq, and 1 ml DNA-template. The temperature profile isoleucyl-tRNA synthases and one enzyme is highly for the thermocycler Mastercyclers Ep Gradient resistant against the produced pseudomonic acid A (Eppendorf, Hamburg, Germany) was as follows: [17,36]. 94 1C for 5 min, 94 1C for 30 s, 20–25 cycles (depending Because of a fast rupture of the epoxide ring, on DNA concentration) at 52 1C for 30 s, 70 1C for 60 s, mupirocin is probably not stable in the environment followed by an elongation cycle at 70 1C for 5 min at the over longer time frames [8]. end. Direct sequencing using amplicon primers (Ba27f, In times of rising resistances of pathogens, new Ba519r, Ba1492r) was performed on an ABI 3730 strategies seem to be necessary to find new and effective sequencer (Applied Biosystems) using the Big antibiotics. We, therefore, performed a systematic study Dye-terminator v3.1 chemistry as specified by the manu- of antibiotic producers in groundwater. A high through- facturer. The 16S rRNA gene sequence reads were put cultivation study with an artificial freshwater manually assembled and checked for quality using the medium (AFM) led to several bacterial cultures SeqMan II software module (Lasergene 6 suite, DNAs- which were tested for antibiotic production. Here, we tar, Madison, USA). For phylogenetic analysis, the 16S present two novel pseudomonic acid A producing rRNA gene sequences were integrated into an ARB strains and characterised them indicating that the ability database of 16S rRNA gene sequences by the use of the to produce pseudomonic acid is spread more among the alignment tool of the ARB software package (http:// pseudomonads. www.arb-home.de). Phylogenetic analyses based on nucleotide sequences were performed and verified by applying maximum parsimony and neighbour-joining Material and methods methods by use of the respective tools in the ARB software package. The 16S gene sequences of the two new antibiotic- Cultivation of microorganisms producing strains were submitted to GeneBank with the accession numbers EU680857 and EU680856. Two microbial strains, D7 and G11, were isolated from sediment of a pristine aquifer located at Scheyern, Germany with serial dilution in microtiter plates. Growth temperature/pH determination Cultivation was performed in nutrient broth medium (NB, peptone 5 g/l, meat extract 5 g/l, pH 7) or LB The optimal growth temperature for the strains was medium (10 g tryptone /l, 5 g yeast extract /l, 10 g NaCl/l). tested for agar and liquid cultures with NB at The cultures were kept by streaking on NB agar plates temperatures between 4 and 60 1C. The optimal pH every weak and stored at 4 1C. For growth in liquid conditions for strains were determined with AFW cultures 50 ml of medium in an Erlenmeyer flask were medium set to pH from 3 to 11 with 2 M NaOH or inoculated with cells of an overnight culture to an 2 M HCL. OD595 nm of 0.01 and shaken at 200 rpm. The same cultivation was used for the test strain Bacillus subtilis subspecies subtilis (DSMZ 10) at 30 1C (see below). For Test for catalase, oxidase activity, and growth on solidified medium, the strains were streaked metabolic properties out on the agar plates and incubated at the given temperature. The test for catalase activity was performed with a 3% aqueous solution of H2O2. A grown colony was Identification of antibiotic-producing picked from an agar plate and mixed with the solution. bacterial strains Frothing indicated catalase activity. A solution of 0.1 g of citric acid and 1.0 g 4-amino- DNA extraction for sequencing of the unknown N,N-dimethylanilin-dihydrochlorid was prepared and a bacterial strains was done with the Fast DNAs Spin soaked filter spread with a grown colony. A blue for Soil Kit (MP Biomedicals, Eschwege, Germany) colouring showed the existence of cytochrom c-oxidase. ARTICLE IN PRESS 58 E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64

To test for metabolic properties of D7 and G11, the solvent strength gradient 0.1% formic acid (A) and standard test kit apis 20NE from Biome´rieux (Marcy 100% acetonitrile containing 0.06% formic acid (B) was l’Etoile, France) was used. used. The following gradient profile was used: 0–15 min from 0% to 100% B, 15–25 min 100% B, 25–30 min Identification of antibiotic compounds from 100% to 0% B, and 30–45 min 0% B. The flow rate was 1 ml/min causing pressure drop on the column For identification of the antibacterial active com- between 95 and 160 bar. The injection volume was 20 ml. pounds, strains D7 and G11 were grown in 50 ml of NB The eluent was fractioned from 0 until 30 min at the rate medium or 50 ml of AFM [3]. AFM as a mineral of 0.5 min/tube with Gilson FC203B fraction Collector medium was chosen to have less disturbing components (Limburg, Germany). in the chromatographic analysis. Nevertheless, both media were compared for antibiotic production, because Fourier transform ion cyclotron resonance mass a better bacterial growth was found in the rich medium. spectrometry (FT-ICR-MS) analysis After 18 h of growth, the cells were separated from the supernatant by centrifugation at 4800 rpm in a High-resolution mass spectra were acquired on a FT- Megafuge 1.0R centrifuge (Fisher Scientific, Schwerte, ICR-MS (Bruker, Bremen, Germany), equipped with a Germany). The supernatant was filtered through a 12 T superconducting magnet and an Apollo II ESI 0.22 mm filter (Millipore, Schwalmbach, Germany), source. Samples containing 75% methanol were infused frozen at À80 1C and lyophilised in a freeze drier (VirTis with the micro-electrospray source at a flow rate of Company, New York, USA). After freeze drying, the 120 ml/h, with a nebulizer gas pressure of 20 psi and a residues were dissolved in 2 ml of ultra pure water and drying gas pressure of 15 psi at 250 1C. Negative and used for further analysis. positive electrospray ionisation was used. Spectra were externally calibrated on clusters of arginine (10 mg/l in Antibiotic susceptibility test methanol); calibration errors in the relevant mass range were always below 0.1 ppm. In the mass range of In order to identify antibiotically active high perfor- 150–2000 m/z 1 MW time domain was applied. The ion mance liquid chromatography (HPLC) fractions, sam- accumulation time in the ion source was set to 2 s, and ples were transferred into a 96-well microtiter plate and 128 scans were accumulated per sample. Before Fourier dried at 50 1C, until all liquid was evaporated. The test transformation of the time-domain transient, a sine for antibiotic activity was conducted with B. subtilis apodization was performed. For fragmentation experi- subspecies subtilis (DSMZ 10) as a very sensitive control ments, collision-induced dissociation (CID) was carried strain. B. subtilis was grown in 5 ml overnight cultures out. The mass window of 1 Da of the parent ions was and transferred into 50 ml Mueller Hinton Broth selected between the ion source and the analyzer cell (Meat infusion 2.0 g/l; casein hydrolysate 17.5 g/l; starch with the quadrupole and were fragmented by varying 1.5 g/l), and shaken for 5 h at 200 rpm. The OD595 nm of the excitation power (15, 20, and 25 eV). Argon as a the culture was measured and a sample was mixed with collision gas was pulsed into the chamber at a pressure À8 Mueller Hinton Agar (0.8% agar) and adjusted to a final of 1 Â 10 Torr, with 50 ms gas pulse duration. On OD595nm of 0.01. Fifty microlitre of this mixture were resonance, CID provided abundant structural informa- added with a multichannel pipette to each well to cover tion. The ion accumulation time in the ion source was the dried culture supernatants and allowed to solidify. set to 0.2 s and 64 scans were accumulated for samples. Before and after growth of 18 h, the OD595 nm was measured with a Victor3 microtiter plate reader (Perkin- Ultra performance liquid chromatography (UPLC) Elmer, U¨ berlingen, Germany). The difference in optical analysis densities was very low when a bioactive fraction inhibited growth of the test organism. Tetracycline An UPLC Acquity System (Waters, Darmstadt, dissolved in ethanol (30 mg/well) served as a control. Germany) equipped with a 2996 PDA detector was applied for analysis. A reversed phase column with Semi-preparative HPLC analysis dimensions of 2.1 Â 100 mm, filled with BEH C18 packing material with 1.7 mm particle sizes (Waters, The resolved supernatants were fractionated on a Darmstadt, Germany, Milford, MA, USA) was used for Perkin-Elmer Series HPLC (U¨ berlingen, Germany) the separation. The column was thermostated at 40 1C, equipped with a diode array detector 235C set and the autosampler at 27 1C. Five microlitre of sample to 215 nm. The separation was performed on a were injected via a partial loop with needle overfill LiChrospher 100 RP-18 column (150 Â 4.6 mm id, injection. The optimised system was run with a linear 5 mm, Merck KGaA, Darmstadt, Germany). For linear gradient starting with water containing 20–100% ARTICLE IN PRESS E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64 59 acetonitrile in 2 min. The flow rate was set to be detected within a broad pH range from nearly 4.0–10.0. 0.6 ml/min that resulted in a maximum system pressure The metabolic properties of both strains were tested of 750 bars. Detection was performed at 220 nm at a with apis 20NE kits and indicated a broad substrate scan rate of 40 Hz. spectrum with the ability to assimilate glucose, arabi- nose, mannose, mannitol, N-acetyl-glucosamine, potas- sium gluconate, capric acid, malic acid, and citrate (Table 1). The two strains only differed in the lack of Results arginine dihydrolase for strain D7. The substrate spectrum of the two strains in the apis 20NE test Description of antibiotic producers indicated a typical metabolic profile of the genus Pseudomonas for both. No difference in metabolic Two novel antibiotic-producing strains, D7 and G11, properties compared to the P. fluorescens type strain were isolated from aquifer sediment. Both were Gram- ATCC 13525 could be found (Table 1). Both strains negative, catalase, and oxidase positive, and did not were motile and showed weak fluorescence at the typical grow in the absence of molecular oxygen. Both strains wavelength of fluorescein (excitation 485 nm, emission grew with AFW, nutrient broth, and LB medium in 535 nm) increasing with higher cell numbers (data not liquid media and on agar plates. The temperature range shown). The supernatants of the two isolated strains for growth of both strains was 15–34 1C with an also showed haemolysis activity on blood agar plates optimum at 28 1C. Growth of strains D7 and G11 was (Merck, Darmstadt, Germany). The producers were

Table 1. Properties of the new bacterial isolates (strains D7 and G11), +enzyme or (metabolic) reaction present,Àenzyme or (metabolic) reaction not present.

Strain D7 Strain G11 P. fluorescens type strain ATCC13525(a)

Property Cell morphology Rod Rod Rod Growth optimum 28 1C281C 28–30 1C Motility + + + Fluorescens + + + Spore formation ÀÀ À Oxidase + + + Catalase + + + Gram stain Negative Negative Negative GC content 59.9 59.9 59–61% pH range 4.5–10 4.5–10 N.r. Nitrate reduction ÀÀ À Indole production ÀÀ À Glucose fermentation ÀÀ À Arginine dihydrolase À ++ Urease ÀÀ À ß-glucosidase ÀÀ À Protease + + + ß-galactosidase ÀÀ À Assimilation of D-glucose + + + L-arabinose + + + D-mannose + + + D-mannitol + + + N-acetyl-glucosamine + + + D-maltose ÀÀ À Potassium gluconate + + + Capric acid + + + Apidic acid ÀÀ À Malic acid + + + Citrate + + + Phenylacetic acid ÀÀ À

N.r., not reported. (a) Data after [20]. ARTICLE IN PRESS 60 E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64

Pseudomonas sp. EP25 (AM403527) bacterium KVD-1700-26 (DQ490316) Pseudomonas sp. MPU 101 (AB334527) Pseudomonas sp. IC038 (U85869) Pseudomonas sp. D7 Pseudomonas sp. G11 Pseudomonas azotoformans IAM1603T (D84009) Pseudomonas synxantha DSM13080 (AF267911) Pseudomonas reactans LMG5329 (AF255337) bv. C PC24 (AF228367) Pseudomonas gessardii CIP105469 (AF074384) Pseudomonas libaniensis CIP105460T (AF057645) Pseudomonas synxantha ATCC9890T (D84025) Pseudomonas mucidolens IAM12356 (D84017) Pseudomonas cedrella CFML96-198T (AF064461) Pseudomonas fluorescens ATCC13525T (AF094725) Outgroup

0.10

Fig. 1. Phylogenetic tree of Pseudomonas species related to D7 and G11. For tree calculation, the out group consisted of a broad collection of organisms that belong to the kingdoms Archea and Bacteria. The bar indicates 10% estimated sequence divergence. The culture collection numbers of the strains is given after the species name. Type strains are marked with T. resistant against many commercially available antibio- tics (tetracycline, cephalothin, nisin, vancomycin, nali- 2.5x106 dixic acid, erythromycin, lincomycin, and penicillin) 2.0x106 (data not shown) and were only sensitive to streptomy- cin and chloramphenicol. 1.5x106 16S rRNA gene analysis revealed members of the D7 grown in NB 6 genus Pseudomonas as closest relatives with 99% 1.0x10 5 identity to Pseudomonas sp. EP25, Pseudomonadaceae Absorbance [mAU] 5.0x10 bacterium KVD-1700-26, Pseudomonas sp. MPU101, NB control 0.0 and Pseudomonas sp. IC038 for both strains D7 and 0 5 10 15 20 25 30 G11 (Fig. 1). Time [min]

Detection of the inhibitory substance 2.5x106

6 The antibacterial substance produced was effective 2.0x10 against the described B. subtilis strain and other Gram- 1.5x106 positive microorganisms (data not shown). D7 grown in AFW 6 Since cell-free supernatants of strains G11 and D7 1.0x10 5 showed antibacterial activity, a strategy for the identi- Absorbance [mAU] 5.0x10 fication of bioactive component(s) was designed based AFW control on liquid chromatographic separation and mass spectro- 0.0 0 5 10 15 20 25 30 metric detection. Several peaks were identified in super- Time [min] natants of AFW grown cultures (Fig. 2b), and were tested for antimicrobial activity and compared with the Fig. 2. HPLC-chromatogram of bacterial supernatants of same fractions from the sterile medium control. An strain D7 grown in two different media (a) nutrient broth antibacterial activity against the test strain B. subtilis medium (NB) and (b) artificial fresh water (AFW). The boxes was detected in fraction 25 of the HPLC-chromatogram mark peaks that only occurred in the chromatogram of bacterial supernatant. (Fig. 2) with a retention time of 12.5 min.

Identification of organic compounds range were observed (data not shown). The mass spectra of the same sample measured at a negative mode were Fraction 25 exhibiting antimicrobial activity was less complex with a dominant peak at 499.2905 m/z subjected to FT-ICR-MS analysis. The samples were (Fig. 3). The elemental composition considering C, H, measured in both positive and negative ionisation mode. N, and O was determined as C26H44O9 with a relative In positive mode, several m/z peaks over the whole mass error of 0.3 ppm, and indicated pseudomonic acid A ARTICLE IN PRESS E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64 61

– 13 – [C26H43O9] [ C1C25H43O9]

AFW D7 Fraction 25 AFW D7 Fraction 25

6 6 499.2904 499.2904 4 4

2 2 500.2938

0 0 AFW G11 Fraction 25 AFW G11 Fraction 25

6 6 499.2905 499.2905

8 4 4

2 2 500.2939

0 0 Mupirocin reference material Mupirocin reference material

6 6 499.2905 499.2905 Relative intensity x 10 4 4

2 2 500.2939 501.2968 0 0 AFW control Fraction 25 AFW control Fraction 25

6 6

4 4

2 2

0 0 500 1000 1500 498 499 500 501 502 m/z m/z

Fig. 3. Dominant peaks of the FT-ICR mass spectrum at a range of 150–2000 m/z (left column) and 498–503 m/z (right column) from HPLC fraction 25 of (a) culture supernatant of strain D7, (b) culture supernatant of strain G11, (c) standard pseudomonic acid A as reference material, and (d) sterile AFM control.

(Fig. 4) as a possible structure (http://chem.sis.nlm.nih. OH gov/chemidplus/) that is known to have antibacterial OH O OH activity. After identification of the substance, commer- OH O cial mupirocin (Sigma, Hamburg, Germany) was tested O O O and showed the same inhibitory effect on growth of the test strain B. subtilis in the susceptibility test (data not shown). OH To obtain a saver identification of the putative OH O OH pseudomonic acid A structure by FT-ICR-MS, frag- OH OH mentation with different CID (15, 20, and 25 eV) was O O O applied. Mass spectra of fraction 25 were compared to the ones of purchased mupirocin reference material Fig. 4. Molecular structures of pseudomonic acid A with (data not shown). Mupirocin dissociated into two (a) closed epoxy ring and (b) derivative with opened epoxy ring. fragments with m/z of 299.1081 and 255.1600 and the ARTICLE IN PRESS 62 E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64 second one fragmented further at the ester group into The 16S rRNA gene sequencing of D7 and G11 showed m/z of 173.1182. These fragments were also observed in 99% similarity to several Pseudomonas strains including the supernatant fractions of D7 and G11 when CID was the P. fluorescens type strain. The closest relatives of D7 applied. and G11 were represented by Pseudomonas sp. EP25, The same supernatant fractions of D7 and G11 were Pseudomonadaceae bacterium KVD-1700-26, Pseudomonas separated by UPLC after the optimisation of the sp. MPU101, and Pseudomonas sp. IC038 (Fig. 1). method for the determination of pseudomonic acid A. These strains were isolated from deep sea sediment Injecting the standard solution, two peaks were ob- (Pseudomonas sp. EP25), a Hawaiian volcanic deposit served and their resolution increased with the decrease (Pseudomonadaceae bacterium KVD-1700-26), and ant- of the pH of the mobile phase (data not shown). On the arctic sea ice (Pseudomonas sp. IC038). The relationship other hand, low pH of the separation solvent had to be to all other strains of the Pseudomonas fluorescence used to get the pseudomonic acid A at neutral form, cluster was also around 99%. A direct comparison of since its predicted pKa was 4.88 (Pallas, Software the 16S rRNA gene sequences to the known mupirocin predicting pKa, logP, logD values and metabolites producer Pseudomonas fluorescens NCIB 10586 was not based on the structural formula of compounds), and possible, because neither the strain nor the 16S rRNA thus a pH o3 was required. Presumably, during the gene sequence were available. separation and/or storage, the epoxy ring of pseudo- Because a huge number of strains have been monic acid A might open resulting in a hydroxyl group accommodated in the genus Pseudomonas seven sub- (Fig. 2). The new metabolite had different retention clusters in the genus were formed [1]. The closely related behaviour which could be visualised by UPLC due to its strains P. azotoformans, P. synxantha, P. fluorescens, higher resolution efficiency compared to HPLC P. gessardii, P. libaniensis, P. mucodolens,and [10,28,30]. At the retention time of pseudomonic acid P. cedrella are all members of the P. fluorescens A (1.34 min), a peak was observed in the fractionated subcluster (Fig. 1). Strains belonging to the Pseudomonas samples with an identical UV–vis spectrum to the fluorescens subcluster are mostly known as bacterial bio reference (data not shown). Thus, the results obtained control organisms or plant pathogens [14,27,31,32],and from the FT-ICR-MS, MS/MS, and UPLC reflect the often produce polyketides (like mupirocin) as bioeffective high reliability of pseudomonic acid A identification. compounds [4]. Since two peaks of the reference material were observed, The interesting fact that we found pseudomonic acid semi-quantification on the amount of the pseudomonic A producers in groundwater demonstrated that even in acid A was possible. The amount produced in culture pristine aquifers, we have to face a certain abundance of supernatants of both strains D7 and G11 were in the low antibiotic producers. This also implies the natural mg/l range (1.8–3.5 mg/l) independent of the type of the abundance of antibiotic resistant organisms in such growth medium used, although the amount of the open habitats. In fact, our isolates were multiresistant against form of pseudomonic acid A was higher in NB medium. tetracycline, cephalothin, nisin, vancomycin, nalidixic acid, erythromycin, lincomycin, and penicillin indicating that multiresistant microorganisms have to be expected also in habitats not impacted by anthropogenic influ- Discussion ences. An identification strategy for the determination of bioactive compounds applying different analytical tech- In this study, we isolated two new pseudomonic acid niques has been built up. First, semi-preparative HPLC A producing Pseudomonas strains from a pristine was applied as a clean-up process for reducing the aquifer. The occurrence of Pseudomonads is ubiquitous number of compounds in the bioactive samples. HPLC in the environment. Special habitats of Pseudomonads has been applied earlier for the determination of are, for example, soil and rhizosphere [26]. Pseudomonads antibiotics in liquid samples such as surface water and are also often found in contaminated aquifers, because groundwater and several HPLC [26] or LC-MS proto- they are able to use a high number of substances as cols exist [29] for the detection of macrolide antibiotics energy or carbon source and can often resist toxic belonging to the polyketides like mupirocin [9,19,24,33]. compounds [11,15,18]. The fractions were then checked for antibiotic activity The bacterial genus Pseudomonas is classified by and the ones showing positive results were further negative Gram stain, the absence of gas formation from studied by FT-ICR-MS. FT-ICR-MS, especially with glucose, no photosynthetic pigments, a positive oxidase high magnetic field (12 Tesla), provides ultrahigh and catalase test, no spore formation, a negative indole resolution of the masses as their m/z and thus allows test and motility [20]. Most other Pseudomonads can the accurate determination of the elemental composition reduce nitrate what distinguishes the species P. fluorescens even from complex samples [16]. Nevertheless, one from the other species of the genus [25]. These elemental composition may represent more than one characteristics were all found for our isolated strains. structure and additional methods such as MS/MS, ARTICLE IN PRESS E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64 63 chromatography, or capillary electrophoresis have to be [7] B.D. Cookson, The emergence of mupirocin resistance: a applied in combination with reference materials for challenge to infection control and antibiotic prescribing comparison purpose [22,23]. In this case, pseudomonic practice, J. Antimicrob. Chemother. 41 (1998) 11–18. acid A was predicted from a database search of the m/z [8] P.J. Crowley, Excipients as stabilizers, Pharm. Sci. data from FT-ICR-MS. UPLC has been used for the Technol. Today 2 (1999) 237–243. confirmation study which differs in the analysis base [9] M.J.G. de la Huebra, U. Vincent, Analysis of macrolide compared to MS that provides faster separation than antibiotics by liquid chromatography, J. Pharm. Biomed. Anal. 39 (2005) 376–398. HPLC (1–5 min instead of 15–45 min), and higher [10] M. Englmann, A. Fekete, C. Kuttler, M. Frommberger, resolution of the peaks. The two possible forms of X. Li, I. Gebefugi, J. Fekete, P. Schmitt-Kopplin, The pseudomonic acid A (open and closed) could be baseline hydrolysis of unsubstituted N-acylhomoserine lactones to separated. Since the reference material also provided their homoserine metabolites. Analytical approaches two peaks, the method enabled semi-quantification of using ultra performance liquid chromatography, pseudomonic acid A, which was detected by UPLC from J. Chromatogr. A 1160 (2007) 184–193. the culture supernatants without further clean-up [11] A. Fahy, T.J. McGenity, K.N. Timmis, A.S. Ball, procedures at concentrations in the low mg/l range Heterogeneous aerobic benzene-degrading communities (1.7–3.5 mg/l). The first description of a pseudomonic in oxygen-depleted groundwaters, FEMS Microbiol. acid A producer reported amounts of 10–20 mg/l, Ecol. 58 (2006) 260–270. which would be 5–10 times more than produced by [12] T.C. Feline, R.B. Jones, G. Mellows, L. Phillips, E.B. our strains [13]. Chain, G. Mellows, Pseudomonic acid. Part 2. Biosynth- With our studies, we could extend the range of known esis of pseudomonic acid A, J. Chem. Soc. [Perkin 1] antibiotic producers in the P. fluorescens cluster and (1977) 309–318. [13] A.T. Fuller, G. Mellows, M. Woolford, G.T. Banks, K.D. provide indications that the production of pseudomonic Barrow, E.B. Chain, Pseudomonic acid: an antibiotic acid A might be a wider spread feature in this group produced by Pseudomonas fluorescens, Nature 234 (1971) than thought before. 416–417. [14] P. Garbeva, K. Voesenek, J.D. van Elsas, Quantitative detection and diversity of the pyrrolnitrin biosynthetic locus in soil under different treatments, Soil Biol. Acknowledgements Biochem. 36 (2004) 1453–1463. [15] B. Hendrickx, H. Junca, J. Vosahlova, A. Lindner, I. We thank Basem Kanawati for FT-ICR-MS support, Ruegg, M. Bucheli-Witschel, F. Faber, T. Egli, M. Mau, Hans-Peter Fiedler for providing the HPLC protocol, M. Schlomann, M. Brennerova, V. Brenner, D.H. Pieper, and Drazˇenka Selesi for creating the phylogenetic tree. E.M. Top, W. Dejonghe, L. Bastiaens, D. Springael, Alternative primer sets for PCR detection of genotypes involved in bacterial aerobic BTEX degradation: distribu- tion of the genes in BTEX degrading isolates and in References subsurface soils of a BTEX-contaminated industrial site, J. Microbiol. Methods 64 (2006) 250–265. [1] Y. Anzai, H. Kim, J.Y. Park, H. Wakabayashi, H. [16] N. Hertkorn, C. Ruecker, M. Meringer, R. Gugisch, Oyaizu, Phylogenetic affiliation of the pseudomonads M. Frommberger, E.M. Perdue, M. Witt, P. Schmitt- based on 16S rRNA sequence, Int. J. Syst. Evol. Kopplin, High-precision frequency measurements: indis- Microbiol. 50 (2000) 1563–1589. pensable tools at the core of the molecular-level analysis [2] A. Baader, C. Garre, U¨ ber antagonisten unter den of complex systems, Anal. Bioanal. Chem. 389 (2007) bakterien, Corresp.-Bl. Schweiz. A¨ erzte 13 (1887) 1311–1327. 385–392. [17] J. Hughes, G. Mellows, S. Soughton, How does [3] K. Bartscht, H. Cypionka, J. Overmann, Evaluation of Pseudomonas fluorescens, the producing organism of the cell activity and of methods for the cultivation of bacteria antibiotic pseudomonic acid A, avoid suicide?, FEBS from a natural lake community, FEMS Microbiol. Ecol. Lett. 122 (1980) 322–324. 28 (1999) 249–259. [18] A. Iribar, J.M. Sanchez-Perez, E. Lyautey, F. Garabetian, [4] C.L. Bender, V. Rangaswamy, J. Loper, Polyketide Differentiated free-living and sediment-attached bacterial production by plant-associated Pseudomonads, Annu. community structure inside and outside denitrification Rev. Phytopathol. 37 (1999) 175–196. hotspots in the river–groundwater interface, Hydrobiolo- [5] E.B. Chain, G. Mellows, Pseudomonic acid. Part 1. The gia 598 (2008) 109–121. structure of pseudomonic acid A, a novel antibiotic [19] I. Kanfer, M.F. Skinner, R.B. Walker, Analysis of produced by Pseudomonas fluorescens, J. Chem. Soc. macrolide antibiotics, J. Chromatogr. A 812 (1998) [Perkin 1] (1977) 294–309. 255–286. [6] E.B. Chain, G. Mellows, Pseudomonic acid. Part 3. [20] N.R. Krieg, J.G. Holt (Eds.), Bergey’s Manual of Structure of pseudomonic acid B, J. Chem. Soc. [Perkin 1] Systematic Bacteriology, Wiliams & Wilkins, Baltimore, (1977) 318–324. 1984. ARTICLE IN PRESS 64 E. Fritz et al. / Systematic and Applied Microbiology 32 (2009) 56–64

[21] D.J. Lane, B. Pace, G.J. Olsen, D.A. Stahl, M.L. Sogin, [29] W.M.A. Niessen, Analysis of antibiotics by liquid N.R. Pace, Rapid-determination of 16 s ribosomal-RNA chromatography mass spectrometry, J. Chromatogr. sequences for phylogenetic analyses, in: Proceedings of A 812 (1998) 53–75. the National Academy of Sciences of the United States of [30] L. Novakova, D. Solichova, P. Solich, Advantages of America 82, 1985, pp. 6955–6959. ultra performance liquid chromatography over high [22] X. Li, A. Fekete, M. Englmann, M. Frommberger, S. Lv, performance liquid chromatography: comparison of G. Chen, P. Schmitt-Kopplin, At-line coupling of UPLC different analytical 2 approaches during analysis of to chip-electrospray-FT-ICR-MS, Anal. Bioanal. Chem. diclofenac gel, J. Sep. Sci. 29 (2006) 2433–2443. 389 (2007) 1439–1446. [31] J.M. Raaijmakers, D.M. Weller, L.S. Thomashow, [23] X.F. Li, J. Fekete, A. Englmann, M. Rothballer, Frequency of antibiotic-producing Pseudomonas spp.in M. Yang, F. Hartmann, A. Chen, G.P. Schmitt-Kopplin, natural environments, Appl. Environ. Microbiol. 63 A practical approach to gradient elution in UPLC and (1997) 881–887. off-line confirmation of the results by FT-ICR-MS, J. Liq. [32] J.M. Raaijmakers, I. de Bruijn, M.J.D. de Kock, Cyclic Chromatogr. Relat. Technol. 30 (2007) 1–17. lipopeptide production by plant-associated Pseudomonas [24] B. Lingerfelt, W.S. Champney, Macrolide and ketolide spp.: diversity, activity, biosynthesis, and regulation, Mol. antibiotic separation by reversed-phase high performance Plant-Microbe Interact. 19 (2006) 699–710. liquid chromatography, J. Pharm. Biomed. Anal. 20 [33] M.P. Schlu¨sener, K. Bester, M. Spiteller, Determination (1999) 459–469. of antibiotics such as macrolides, ionophores, and [25] M.T. Madigan, John M. Martinko, in: G. Carlsons (Ed.), tiamulin in liquid manure by HPLC-MS/MS, Anal. Brock Biology of Microorganisms, Vol. 12.7, Bioanal. Chem. 375 (2003) 942–947. Pearson Prentice Hall, Upper Saddle River, 2006, [34] R. Sutherland, R.J. Boon, K.E. Griffin, P.J. Masters, pp. 345–348. B. Slocombe, A.R. White, Antibacterial activity of [26] A. Marzo, L. Dal Bo, Chromatography as an analytical mupirocin (pseudomonic acid), a new antibiotic for tool for selected antibiotic classes: a reappraisal addressed topical use, Antimicrob. Agents Chemother. 27 (1985) to pharmacokinetic applications, J. Chromatogr. A 812 495–498. (1998) 17–34. [35] W.G. Weisburg, S.M. Barns, D.A. Pelletier, D.J. Lane, [27] O.V. Mavrodi, B.B.M. Gardener, D.V. Mavrodi, 16 s Ribosomal DNA Amplification for Phylogenetic R.F. Bonsall, D.M. Weller, L.S. Thomashow, Genetic Study, J. Bacteriol. 173 (1991) 697–703. diversity of phlD from 2,4-diacetylphloroglucinol-produ- [36] T. Yanagisawa, M. Kawakami, How does Pseudomonas cing fluorescent Pseudomonas spp, Phytopathology 91 fluorescens avoid suicide from its antibiotic pseudomonic (2001) 35–43. acid?: evidence for two evolutionarily distinct isoleucyl- [28] B.J. Murphy, UPLC outperforms HPLC in metabolite tRNA synthetases conferring self-defense, J. Biol. Chem. screening, Trac-Trends Anal. Chem. 25 (2006) Vii–Viii. 278 (2003) 25887–25894.