Novel Natural Parabens Produced by a Microbulbifer Bacterium in Its Calcareous Sponge Host Leuconia Nivea

Environmental Microbiology (2009) doi:10.1111/j.1462-2920.2009.01880.x

Novel natural parabens produced by a Microbulbifer bacterium in its calcareous sponge host Leuconia nivea

Elodie Quévrain,1 Isabelle Domart-Coulon,2 parabens in a sponge host, which may have an eco- Mathieu Pernice2 and logical role as chemical mediators. Marie-Lise Bourguet-Kondracki1* 1Laboratoire de Chimie et Biochimie des Substances Introduction Naturelles UMR 5154 CNRS, Muséum National d’Histoire Naturelle, 57 rue cuvier (C. P 54), Paris, Although the Porifera phylum continues to be an excellent France. source of bioactive and original natural products (Blunt 2Laboratoire de Biologie des Organismes Marins et des et al., 2008), sponge-associated microorganisms have Ecosystèmes UMR 5178 MNHN-CNRS-UPMC, Muséum emerged as promising candidates for the production of National d’Histoire Naturelle, 57 rue cuvier (C. P 51), natural compounds (Moore, 1999; Piel, 2006; Blunt et al., Paris, France. 2007). Recently, many studies revealed that metabolites previously ascribed to sponges were in fact synthesized by microorganisms (Proksch et al., 2002; Hildebrand Summary et al., 2004; Piel, 2004; Salomon et al., 2004), while A broad variety of natural parabens, including four others demonstrated structural similarities between novel structures and known ethyl and butyl parabens, natural products isolated from sponges and bacterial were obtained from culture of a Microbulbifer sp. bac- metabolites (Crews and Bescansa, 1986; Zabriskie et al., terial strain isolated from the temperate calcareous 1986; Jansen et al., 1996; Erickson et al., 1997; Kunze marine sponge Leuconia nivea (Grant 1826). Their et al., 1998). These data explain the considerable atten- structures were elucidated from spectral analysis, tion paid by scientists over the last decade (Bewley and including mass spectrometry and 1D and 2D nuclear Faulkner, 1998; Hentschel et al., 2006; Taylor et al., 2007) magnetic resonance. Their antimicrobial activity to the diverse microorganisms associated with marine evaluated against Staphylococcus aureus was char- sponges. Electron microscopy has shown that in some acterized by much higher in vitro activity of these demosponges these microorganisms constitute up to natural paraben compounds 3–9 than commercial syn- 60% of the biomass (Vacelet and Donadey, 1977; Wilkin- thetic methyl and propyl parabens, usually used as son, 1978). In contrast to siliceous sponges, very few antimicrobial preservatives. Compounds 4 and 9 chemical and microbiological studies have so far been revealed a bacteriostatic effect and compounds 6 and carried out on calcareous sponges. Those which do exist 7 appeared as bactericidal compounds. Major paraben are restricted to the Calcinea subclass, with to our know- compound 6 was also active against Gram positive ledge no studies of the Calcaronea subclass (Schreiber Bacillus sp. and Planococcus sp. sponge isolates and et al., 2006; Muscholl-Silberhorn et al., 2008). was detected in whole sponge extracts during all In a programme devoted to studying the role of sponge- seasons, showing its persistent in situ production associated bacteria and especially the chemical media- within the sponge. Moreover, Microbulbifer sp. bacte- tors of the interactions between microbial associates in ria were visualized in the sponge body wall using their sponge host, the cultivable heterotrophic bacterial fluorescence in situ hybridization with a probe specific biota was isolated from the temperate calcareous sponge to L4-n2 phylotypes. Co-detection in the sponge host Leuconia nivea collected off Concarneau (Northeast of both paraben metabolites and Microbulbifer sp. Atlantic, France). The sponge L. nivea (Grant 1826), L4-n2 indicates, for the first time, production of natural which belongs to the Baeriidae family in the Calcaronea subclass, was chosen because its chloromethylenic

(CH2Cl2) crude extract revealed a signiﬁcant and annually persistent antimicrobial activity against Staphylococcus Received 29 August, 2008; accepted 16 December, 2008. *For cor- respondence. E-mail [email protected]; Tel. (+33)140793135; aureus. This activity was localized in the bacterial fraction Fax (+33)140795606. of the sponge obtained by differential sedimentation

Table 1. Antimicrobial activity of the sponge Leuconia nivea throughout 2005–2007, extraction yield and presence of compound 6 in the CH2Cl2 crude extracts.

Wet weight Dry weight CH2Cl2 mass CH2Cl2 extract’s yield Activity against Compound 6 Season Collection date (g) (g) extract (mg) (% of dry weight) S. aureus ATCC 6538a detection (LC/MS)

Spring 30 March 2006 21.9 5.2 13.5 0.26 + (7.5 mm) ND 27 April 2006 517.0 13.7 26.9 0.20 + (7.5 mm) + Summer 24 June 2005 4.5 1.1 18.0 1.67 + (11 mm) ND 24 July 2005 27.0 6.5 55.9 0.86 + (7 mm) + Fall 09 September 2006 ND ND 12.3 ND + (8 mm) + 27 September 2007 5.3 1.3 11.0 0.87 + (10 mm) + 17 October 2005 24.8 5.9 22.7 0.38 + (9 mm) + Winter 31 January 2006 45.6 10.9 13.3 0.12 + (8 mm) + 19 February 2007 12.1 2.9 3.3 0.11 + (8 mm) ND

a. Inhibition’s diameter of the CH2Cl2 extract measured by the disc diffusion assay method for 1 mg per disc. ND, not determined.

following the method of Richelle-Maurer and colleagues to three different seasons (data not shown). These results (2001). The cultivable heterotrophic bacterial biota asso- prompted us to investigate the microbial community associated with L. nivea was screened to isolate bioactive ciated with this marine calcareous sponge. strains, with antimicrobial activity spectra that matched whole sponge extracts. This provided a novel model to investigate the contribution of bacterial metabolites to the Description of the strain L4-n2 chemical profile and antimicrobial activity of their sponge Among the cultivable heterotrophic bacterial biota asso- host. ciated with L. nivea, one of the most active strains, named The antimicrobial activity of heterotrophic bacteria iso- L4-n2, was isolated in January 2006 on Marine Agar lated from L. nivea was examined. CH2Cl2 crude extract supplemented with nalidixic acid and grown in pure culture investigations of the most active strains, Microbulbifer in marine agar or marine broth. Bacteria of the strain L4-n2 sp., led to the isolation of a broad variety of natural para- were Gram negative, rod-shaped, and approximately bens. In addition to the known ethyl 1 and butyl 2 para- 0.3–0.5 mm wide by 2–4 mm long. They were aerobic and bens, seven natural parabens including four novel formed smooth, convex and mucoid colonies on marine structures 5, 6, 7 and 9 were identified. We also con- agar, producing a non-diffusible brown pigment. They were firmed the in situ production and annual persistence of resistant to ampicillin, penicillin and oxacillin and sensitive the major bacterial compound 6 in the L. nivea sponge to kanamycin (30 UI) and ceftazidime (30 mg). through liquid chromatography mass spectrometry (LC/ MS) analysis. This study provides the first chemical report of a family Phylogenetic affiliation of L4-n2 of bioactive compounds from a bacterium associated to a calcareous sponge belonging to the Baeriidae family, and Strain L4-n2 was identified by sequencing the partial 16S proves its in situ production within the sponge host. rRNA gene. Sequences were compared with those in the database by using FASTA (Pearson and Lipman, 1988). The FASTA search affiliated L4-n2 isolated from L. nivea Results with the RSBr-1 type strain of marine bacterium Micro- bulbifer arenaceous previously isolated from red sand- Antimicrobial activity of the sponge L. nivea stone off the coast of Scotland (Tanaka et al., 2003) with

We established that CH2Cl2 extracts of the sublittoral 99.8% 16S rDNA sequence identity. A 16S rRNA-based sponge L. nivea collected off Concarneau inhibited the phylogenetic tree constructed using the maximum likeli- growth of the clinical strain S. aureus (ATCC 6538) in the hood algorithm (Felsenstein, 1981) affiliated L4-n2 with agar diffusion assay. Collection at three month intervals the genus Microbulbifer in the Alteromonadales order of from 2005 to 2007 showed that this activity was retained the Gammaproteobacteria subdivision. Within the Microb- throughout the year (Table 1). Furthermore, the anti- ulbifer genus, the L4-n2 isolate formed a highly supported S. aureus activity was reproducibly localized in the bacte- clade with the marine sandstone derived species rial fraction of the sponge, obtained by differential M. arenaceous and with the halophile environmental sedimentation of mechanically dissociated sponge sus- Microbulbifer sp. strain YIM C306, with a bootstrap conﬁ- pension in three independent experiments corresponding dence level of 99.8% (Fig. 1).

Bacillus algicola SC3 (DQ001308)

Moritella marina NCIMB 1144T (X82142)

99 Alteromonas macleodii DSM 6062 (Y18228)

77 Pseudoalteromonas flavipulchra NCIMB 2033 (AF297958)

Microbulbifer maritimus TF-17 (AY377986) (intertidal)

L4-n2 (FM200853) (intertidal)

100 Microbulbifer sp. YIM C306 (EU135714)

98 Microbulbifer arenaceous RSBr-1T (AJ510266) (intertidal)

Microbulbifer sp. KBB-1 (DQ412068) (deep sea) 100 Microbulbifer agarilyticus JAMB A3 (AB158515) 50

Bacterium QM46 (DQ822531)

51 Microbulbifer elongatus JAMB A7 (AB107975) (deep sea)

73 Microbulbifer elongatus ATCC 10144T (AB021368) 58 (intertidal) 88 Microbulbifer elongatus DSM 6810 (AF500006)

Microbulbifer salipaludis SM-1 (AF479688)

Microbulbifer sp. Th/B/38 (AY224196) (intertidal) 98 Microbulbifer hydrolyticus DSM 11525T (AJ608704)

Microbulbifer sp. A4B-17 (AB243106)

sp. 17x/A02/240 (AY576773) 100 Microbulbifer

93 Microbulbifer mediterraneus UST 040317-067 (DQ096579) 70 Microbulbifer cystodytense C1 (AJ620879)

Microbulbifer sp. CJ11049 (AF500206) (deep sea) 100 Microbulbifer thermotolerans JAMB A94 (AB124836) 0.1

Fig. 1. Phylogenetic affiliation of the strain L4-n2 within the genus Microbulbifer in the Alteromonadales order of the Gammaproteobacteria, based on maximum likelihood algorithm using the 16S rRNA gene sequence. The tree was constructed by analysis of 1304 bp of the 1500 bp gene sequence, using the maximum likelihood method from the PHYLIP package (fastdnalml software at http://bioweb.pasteur.fr). Bootstrap values (calculated as percentage from 1000 replications) higher than 50% are reported at branch nodes and show the robustness of observed branching patterns. Scale bar represents 0.1 substitutions per nucleotide position and indicates estimated sequence divergence. Bacillus algicola (DQ001308) was used as outgroup. Information on the natural environment of the strains is indicated in brackets.

Bacterial paraben compounds against S. aureus, yielded nine compounds belonging to the paraben series: the known ethyl- and butyl-parabens

Cultures of the L4-n2 strain were extracted with CH2Cl2. 1–2 and the seven natural analogues 3–9, including the As extracts of both culture supernatant and pellet dis- four new natural parabens 5, 6, 7 and 9 (Fig. 2). Com- played identical chemical proﬁle and activity, they were pounds 1, 3, 4, 8 appeared for the ﬁrst time as natural combined to constitute the CH2Cl2 crude extract of the products. The structures of these isolated natural com- culture, which was further fractionated using a bio-guided pounds were determined by mass and nuclear magnetic assay for its antimicrobial activity. Successive chromato- resonance (NMR) (1D and 2D) spectral analysis. graphies of the CH2Cl2 crude extract on silica gel and Compound 1 was obtained as a white powder. Electro- reverse-phase columns, guided by antimicrobial activity spray ionization–mass spectrometry (ESI-MS) esta-

© 2009 The Authors Journal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology 4 E. Quévrain, I. Domart-Coulon, M. Pernice and M.-L. Bourguet-Kondracki

Fig. 2. Structures of the paraben compounds 1 R = -CH2-CH3 O 1–9. 2 R = -(CH2)3-CH3 3 R = -(CH2)7-CH3 R 4 R = -(CH ) -CH 7 2 9 3 2 O 3 1 5 R = -(CH2)2-CH(OH)-(CH2)6-CH3 6 R = -(CH2)8-CH(CH3)2 4 6 5 7 R = -(CH2)2-CH(OH)-(CH2)5-CH(CH3)2 HO 8 R = -(CH2)11-CH3 9 R = -(CH2)4-CH=CH-(CH2)5-CH3 blished the protonated molecular ion [M+H]+ at m/z Compound 5 was obtained as a pale yellow oil. The 167.0708 (D-0.019 mmu), indicating the molecular ESI-MS showed the [M+H]+ ion at m/z 295.1913 (calc. formula C9H11O3. Detailed examination of spectroscopic 295.1923 for C17H26O4) suggesting the molecular formula 1 13 data (UV, H and C NMR spectra) readily established C17H26O4 for compound 5, which differed from that of 4 the identity of compound 1 with the ethyl ester of only by an oxygen atom more. The presence of a parahy- 4-hydroxybenzoic acid named ethylparaben, a petro- droxybenzoate was easily recognized by characteristic chemical antimicrobial compound used as preservative signals from its 1H NMR spectrum. In the 1H NMR spec- for pharmaceuticals. trum, we also observed additional signals indicating the 1H NMR data of the following compounds 2–9 revealed presence of one methine proton at d 3.71 (m) and four that they share two identical sets of coupled protons at d methylene signals at d (4.58–4.33, 1.92–1.76, 1.45–1.30, 7.94 (d, 8.8) and d 6.83 (d, 8.8) and one deshielded 1.40–1.28). The COSY spectrum allowed to delineate the methylene proton signal at d 4.25 (t, 6.7). These data partial structure C1′–C4′. Finally, HMBC correlations identified them as analogues of the paraben 1. between the deshielded methylene protons at d Electrospray ionization–mass spectrometry of com- (4.58–4.33) to carbons at d 36.5 (C2′) and 68.6 (C3′) and pound 2, a white powder, showed the [M+H]+ ion at m/z between the methylene protons at d (1.93–1.76) to 195.1028 (D-0.6 mmu), leading to the molecular formula carbons at d 62.0 (C-1′) and 68.6 (C-3′) located the 1 C11H15O3. Examination of spectroscopic data (UV, H and hydroxyl function in position 3′ of the alkyl chain. The final 13C NMR spectra) identified compound 2 with the butyl structure identified the new natural paraben 5 as the ester of the 4-hydroxybenzoic acid named butylparaben, 4-hydroxybenzoic acid 3-hydroxy-decyl ester also another petrochemical antimicrobial compound also used named hydroxydecylparaben. The very low yield of this as preservative in foods and cosmetics. compound precluded determination of its absolute Compound 3 was obtained as colourless oil. Electro- configuration. spray ionization–mass spectrometry analysis suggested Compound 6 was isolated as a colourless oil. Its the molecular formula C15H23O3, for the pseudomolecular molecular formula C18H28O3 was deduced from ESI-MS of [M+H]+ ion peak at m/z 251.1652. In comparison with 1H the protonated molecular ion [M+H]+ at m/z 293.2115, NMR data of the compound 2, we observed four addi- indicating the presence of five unsaturations in the mo- tional methylene proton signals at d 1.34 (m), and 1.26 (m) lecule. Inspection of the 1H NMR spectrum indicated that suggesting the presence of an octyl alkyl chain in the compound 6 was another 4-hydroxybenzoate (4HBA) molecule, which was confirmed with HMBC and COSY alkyl ester. Additional signals for one methine proton at d spectra. Hence, compound 3 was easily identified as 1.53 (m) and two methyl doublets at d 0.88 (d, 6.6) were being the 4-hydroxybenzoic acid octyl ester or octylpara- observed in the 1H NMR spectrum. 1H COSY correlations ben, already known as a synthetic compound but isolated between this methine proton at d 1.53 and both methyl for the first time as natural product. groups at d 0.88 established the presence of an isopropyl Compound 4 was obtained as a pale yellow oil. group. HMBC correlations, as illustrated in Fig. 3, Electrospray ionization–mass spectrometry analysis supported the identification of compound 6 as the new suggested the molecular formula C17H27O3 for the methyldecylparaben. pseudomolecular [M+H]+ ion peak at m/z 279.1962. Its 1H Compound 7 was obtained as white needles. The NMR spectrum displayed analogies with compound 3. [M+H]+ ion at m/z 309.2055 suggesting for compound 7 1 The most significant differences observed in the H NMR the molecular formula C18H28O4, which differed from that spectrum concerned the alkyl chain. Additional methylene of 6 only by an oxygen atom more. However, detailed proton signals at d 1.24 (m) and 1.26 (m) suggesting the examination of spectral data indicated similarities with presence of a decyl chain in compound 4, which was compound 5. The 1H NMR spectrum also showed the easily identified as being the decyl hydroxybenzoate, a presence one methine proton at d 3.72 (m) and four new natural paraben. methylene signals at d (4.57–4.33, 1.93–1.76, 1.47–1.47,

→ H O Fig. 3. Selected COSY (–) and HMBC ( ) correlations of compound 6.

H O

HO H

1.46–1.31), indicating that 7 possessed the same C-1′– carbon resonances at d 130.6 and 129.2 respectively, C-4′ as compound 5. 1H COSY correlations between the suggesting the presence of one additional unsaturation methine proton at d 1.26 and both methyl groups at d 0.83 in the alkyl chain. Its localization was determined by established the presence of an isopropyl group, identify- HMBC correlations from the both methylene protons at d ing the new paraben 7 as the 4-hydroxybenzoic acid 2.10 (H-4′,m)andd 2.00 (H-7′, m) to the carbons reso- 3-hydroxymethyldecyl ester, also named hydroxymethyl- nating at d 129.2 (C-5′) and 130.6 (C-6′). Key COSY decylparaben. Its absolute configuration could not correlations from the both methylene protons at d 2.10 be examined due to limited amount of the isolated (H-4′,m)andd 2.00 (H-7′, m) to the methine protons at compound. d 5.35 (2H, H-5′,H-6′) delineated the partial structure Compound 8 was isolated as a white amorphous solid C4′–C7′ and confirmed the localization of the unsatura- of the molecular formula C19H30O3 deduced from ESI-MS tion. Compound 9 was identified as a new natural of the protonated molecular [M+H]+ ion at m/z 307.2278 paraben named dodec-5-enylparaben. The geometry of 13 (C19H31O3). Detailed interpretation of NMR data easily the double bond could be determined neither by C revealed the structure of the dodecylparaben. NMR nor by NOESY spectrum. Compound 9 was isolated as a white amorphous solid. Electrospray ionization–mass spectrometry analysis sug- Activity of bacterial parabens against S. aureus gested the molecular formula C19H29O3 for the pseudomolecular [M+H]+ ion peak at m/z 305.2102, differing Antimicrobial activity of the paraben family isolated from from that of 8 by two hydrogens and indicating the pre- Microbulbifer L4-n2 strain was confirmed towards sence of six unsaturations in the molecule. The main S. aureus (Gram+) bacterial strain and compared with the differences were observed in their NMR data. In 1H and antimicrobial activity of the commercial methylparaben 13C NMR spectra of compound 9, we observed the pres- and propylparaben (Interchimie). Results of this evalua- ence of two proton signals at d 5.35 (2H, m) and two tion are summarized in Table 2.

Table 2 MICs and MBCs of the compounds 1–9 against Staphylococcus aureus.

Compound MIC (mM)a,b MIC (mgml-1)a,b MBC (mgml-1)a Ratio MBC/MIC

Methylparaben 328.9–657.8 50–100 > 100 C. Dc Propylparaben 277.7–555.5 50–100 > 100 C. Dc 1 > 602.0 > 100 > 100 C. Dc 2 128.8–257.7 25–50 > 100 C. Dc 3 100–200 25–50 50–100 2 = bactericidal 4 2.8–5.6 0.78–1.56 12.5–25 16 = bacteriostatic 5 85–170 25–50 > 100 C. Dc 6 42.8–85.6 12.5–25 25–50 2 = bactericidal 7 40.5–81.1 12.5–25 25–50 2 = bactericidal 8 81.6–163.3 25–50 > 100 C. Dc 9 20.5–41.1 6.25–12.5 50–100 16 = bacteriostatic a. Compounds were tested in triplicate. b. By turbidity method. c. C. D: Cannot be determined because MBC is above detection range. Minimal inhibitory concentration (MIC) was expressed as the interval a-b, where a is the highest concentration tested at which microorganisms are growing and b the lowest concentration that causes 100% growth inhibition. Minimal bactericidal concentration (MBC) was deﬁned as the lowest concentration of compound yielding colony counts < 0.1% of the initial inoculum. MBC/MIC revealed the nature of compound. If MBC/MIC is less than 4, the compound is bactericidal, If MBC/MIC is between 8 and 16, the compound is bacteriostatic.

© 2009 The Authors Journal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology 6 E. Quévrain, I. Domart-Coulon, M. Pernice and M.-L. Bourguet-Kondracki

All of the parabens isolated from the culture of Micro- In situ chemical detection of compound 6 in the sponge bulbifer sp. strain L4-n2 exhibited significant antimicrobial CH2Cl2 extract activity. Compound 4 exhibited the greatest efficiency Liquid chromatography-mass spectrometry (LC-MS) was against S. aureus with minimal inhibitory concentration carried out to look for compound 6, the major bacterial (MIC) values of 2.8–5.6 mM. paraben produced by the Microbulbifer sp. L4-n2 strain In order to determine whether the growth inhibition cultures. Compound 6 was detected in CH Cl sponge observed was reflected a bactericidal or bacteriostatic 2 2 extracts at every season (Table 1 and Fig. 4), indicating effect, minimal bactericidal concentrations (MBCs) were persistent annual production of this bacterial metabolite determined and the ratio MBC/MIC was calculated. Using within the sponge and suggesting a contribution of this this method, compounds 4 and 9 appeared as bacterio- metabolite to the permanent antimicrobial activity of static and compounds 3, 6 and 7 appeared as bacteri- L. nivea. cidal. In order to search for a possible synergistic mode of action, antimicrobial assays were simultaneously performed with mixtures of compounds at equimolar ratio but In situ localization of the Microbulbifer strain L4-n2 by neither synergistic effects, nor additional effects were catalysed reporter deposition fluorescence in situ observed against S. aureus (data not shown). hybridization in the sponge body wall

A probe was designed against a variable region of the 16S rRNA gene specific to the clade of our Microbulbifer sp. Antimicrobial activities of the L4-n2 CH2Cl2 crude extract and purified compound 6 strain L4-n2, M. arenaceous and Microbulbifer sp. strain YIM C306 (Fig. 5). The catalysed reporter deposition fluo- The CH2Cl2 crude extract of the strain L4-n2 proved to be rescence in situ hybridization (CARD-FISH) signal was active against marine Gram positive strain isolated from positive only for the L4-n2 bacteria (Fig. 5B) and negative the sponge L. nivea and identified as Bacillus sp. L6-3e for control Alteromonadales Moritella sp. L7-2nh (Fig. 5C) and also against Gram negative Escherichia coli (ATCC and Pseudoalteromonas sp. L7-3a (data not shown) as 8739) and the environmental bacterium Vibrio splendidus well as Alphaproteobacteria Pseudovibrio sp. L4-8 (data (LGP32) as well as sponge-derived Alphaproteobacteria not shown) strains also isolated from L. nivea. Tissue Pseudovibrio sp. L4-8 (Table 3). These results revealed sections obtained from L. nivea specimens collected in the existence of various cross-interactions between strain four seasons were used: July 2005 (two sections), L4-n2 and other sponge-derived isolates. Compound 6, October 2005 (two sections), January 2006 at the time the which was the major bacterial paraben produced by L4-n2 strain was isolated (two sections) and September L4-n2, inhibited the growth of the marine L. nivea-derived 2006 when compound 6 detection by LC/MS in whole Gram positive strains Bacillus sp. L6-3e and Planococcus sponge extracts was highest in two consecutive years sp. L5-4b, but it did not inhibit Gram negative bacteria (three sections). Examination of sponge tissue sections such as E. coli (ATCC 8739), nor marine environmental after hybridization with the specific probe confirmed the V. splendidus (LGP32) and sponge-derived Pseudovibrio persistent presence of strain L4-n2 with a positive signal sp. L4-8 (Table 3). This indicated the presence of antimi- observed in these four seasons, marking a rod-shaped crobial compounds other than parabens in the CH2Cl2 bacteria, in at least five different optical field zones per extract of L4-n2 Microbulbifer strain, and that these yet section. Bacterial distribution was restricted to the extra- undetermined metabolites were being responsible for the cellular mesohyl and preferentially to the ectosome (outer activity against Gram negative bacteria. part of the body wall), although some positive bacteria

Table 3. Antimicrobial activity of the L4-n2 bacterial strain.

Gram positive strains Gram negative strains

S. aureus Bacillus sp. Planococcus sp. E. coli V. splendidus Pseudovibrio sp. Reference strains ATCC 6538 L6-3ea L5-4ba ATCC 8739 LGP32 L4-8a

b CH2CL2 crude extract +17 mm +11 mm ND +11 mm +7mm +9mm Compound 6c +13 mm +9mm +7.5 mm 0 0 0 a. Leuconia nivea isolates. b. CH2Cl2 crude extract was tested by disc diffusion assay (1 mg per disc). c. Pure compound 6 was tested by disc diffusion assay at 100 mg per disc. ND, not determined.

Fig. 4. Detection of compound 6 by LC/MS in the CH2Cl2 crude extract of L. nivea (January 2006). A. HPLC profile total ion current chromatogram selecting for retention time: 18 min. B. Positive ESI mass spectrum revealing the presence of compound 6. were also observed in the mesohyl surrounding the cho- determined as alkyl esters of the 4HBA, also named para- anocyte chambers. Signal abundance was less than bens. Their antimicrobial activity was evaluated, revealing 5% relative to the total number of bacteria stained with that natural parabens 3–9 have better in vitro activity than the non-specific DAPI DNA fluorochrome or with the synthetic methylparaben and propylparaben and than non-specific eubacterial probe EUB338, indicating a natural ethylparaben and butylparaben 1–2, which are the low number of L4-n2 affiliated representatives in the total four parabens usually used as antimicrobial preserva- bacterial community of L. nivea. Although at low relative tives. We found that compounds 3, 4, 6, 8 revealed a abundance, Microbulbifer sp. L4-n2 was persistently higher potency and appeared about 100-fold more active present in situ within its sponge host, throughout the year. than methyl and propyl synthetic parabens as well as ethyl and butyl natural parabens (1–2). Bacteria of the L4-n2 16S rDNA phylotype were loca- Discussion lized by fluorescence in situ hybridization with a specific

The CH2Cl2 crude extracts of the marine sponge L. nivea, probe in the sponge body wall, in summer, early fall, and selected after antimicrobial screening, revealed perma- winter, indicating the persistent presence of these bacte- nent activity against Gram positive (S. aureus) bacteria in ria in the L. nivea sponge. Furthermore, the in situ detec- all seasons from June 2005 through February 2007. This tion of compound 6 by LC/MS in the host sponge extract activity was localized in the bacterial fraction of the at every season indicates production of bacterial paraben sponge separated by differential sedimentation of the dis- compounds within its calcareous sponge host. sociated sponge cell suspension. Among the bacterial Recently, Xue Peng and colleagues (2006) reported isolates comprising the cultivable heterotrophic ﬂora, one that a strain of Microbulbifer sp. (strain A4B-17), isolated of the most active strains against S. aureus wasaGram from a tropical ascidian, was able to produce 4HBA and negative strain which was phylogenetically affiliated with its butyl-, heptyl- and non-yl esters. With the exception the genus Microbulbifer. of butylparaben 2, these compounds are different from The chemical nature of the compounds responsible for those produced by our strain L4-n2 derived from a cal- the antimicrobial activity against S. aureus has been careous temperate sponge. These authors also reported

© 2009 The Authors Journal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology 8 E. Quévrain, I. Domart-Coulon, M. Pernice and M.-L. Bourguet-Kondracki

Fig. 5. CARD-FISH hybridization of bacteria in culture (A, B, C) and in the sponge host Leuconia nivea (D, E, F). Culture of Microbulbifer sp. L4-n2 bacterial cells were hybridized with the universal bacterial probe EUB338 (A) and probe Ma445 specific to the Microbulbifer sp. L4-n2 phylotype (B) which does not hybridize with bacterial cells of another sponge-associated Alteromonadales bacteria Moritella sp. L7-2nh (C, and insert C visualized with EUB338 eubacterial probe). Probe NW442 specific to sponge Pseudovibrio sp. Alphaproteobacteria was used as internal negative control and showed no positive signal with L4-n2 bacterial cells (insert B). Another internal negative control, probe nonEUB, was also used (data not shown). Leuconia nivea tissue sections were hybridized with the universal probe EUB388 (D) and specific probe Ma445 with positive signal confirming the presence of Microbulbifer sp. L4-n2 phylotype (arrows) in both samples collected in January 2006 (E) and September 2006 (F). In blue, nuclei of L. nivea sponge cells are stained with DAPI. (A, B, C) Scale bar, 10 mm; (D, E, F) Scale bar, 20 mm. that out of a total of 23 strains tested, belonging to the marine bacteria belonging to the genus Microbulbifer are genus Microbulbifer and derived from various marine able to produce natural parabens. Furthermore, we have sources and geographic areas, mostly tropical, none of shown in preliminary studies that these natural parabens them except the A4B-17 strain was able to produce can inhibit other bacterial strains from the sponge. These parabens. Strain A4B-17 shared 95% 16S rRNA gene results suggest that they might play a role in the interac- sequence identity with Microbulbifer hydrolyticus IRE- tions between bacterial associates within their sponge 31T and Microbulbifer salipaludis SM-1 T and 94% 16S host and/or might contribute to the chemical defence rRNA gene sequence identity with Microbulbifer elonga- strategy of the littoral sponges against opportunistic Gram tus ATCC10144 T. Our strain L4-n2 forms a highly sup- positive environmental bacteria. ported clade with M. arenaceous, in a subgroup of We have shown that strains affiliated to Microbulbifer Microbulbifer sp. strains distant from the above men- sp. L4-n2 strains were present but not abundant in the tioned strains, which reinforces its specific interest. To sponge L. nivea and mostly restricted to the ectosome of our knowledge the antimicrobial activities of M. arena- the sponge body wall. Our study confirms previous ceous previously isolated from red sandstone off reports of Microbulbifer sp. bacteria in the cultured Scotland (Tanaka et al., 2003) and of halophile environ- fraction and among the phylotypes of sponge bacterial mental Microbulbifer sp. strain YIM C306 (EMBL communities: it adds to the recent discovery in the EU135714; X.L. Cui, Y.G. Chen, W.J. Li, L.H. Xu and Mediterranean Clathrina clathrus sponge of the Calcinea C.L. Jiang, unpublished) have not been evaluated. subclass (Muscholl-Silberhorn et al., 2008) of a Microbul- All the Microbulbifer sp. strains tested by Peng and bifer sp. isolate, based on partial (400 nt) 16S rRNA colleagues were able to produce 4HBA, although in lower sequence. We report for the first time the visualization of amounts (less than 20%) than the A4B-17 strain. Com- these bacteria within a sponge body wall with a specific pound 4HBA which is only soluble in MeOH or H2Owas oligonucleotidic probe. Future studies targeting other bac- not detected in the CH2Cl2 crude extract of our L4-n2 terial groups will be carried out to determine the relative Microbulbifer sp. strain but might be present in extracts abundance in the L. nivea sponge host of the different based on more polar solvents. Our study confirms that cultivable bacteria in order to determine which microor-

© 2009 The Authors Journal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology Parabens from sponge associated bacteria 9 ganisms represent stable members of the community. It spectra were recorded on an API Q-STAR PULSAR I of opens the way to the isolation and elucidation of other Applied Biosystem. 13 chemical mediators involved in the interactions between C NMR spectra were obtained on a Bruker AC300 at 1 sponge bacterial associates. 75.47 MHz, H NMR spectra 1D and 2D (COSY, HSQC, HMBC, NOESY) were obtained on a Bruker AVANCE 400. Although parabens are well-known compounds (syn- HMQC and HMBC experiments were acquired at 400.13 MHz thetic molecules) widely used as preservatives in cosmet- using a 1H-13C Dual probehead. The delay preceding the 13C ics, food, pharmaceutical forms, their natural role in the pulse for the creation of multiple quanta coherences through marine environment has not yet been explored and their several bounds in the HMBC was set to 70 ms. supply from natural bacterial sources could open new perspectives in the field of marine natural products Sponge material and preparation of crude extract chemistry. Live specimens of the calcareous sponge L. nivea (Porifera, Calcarea, Calcaronea, Baeriidae) were collected off Concar- Conclusion neau (France) from 2005 to 2007. Samples of fresh sponge (4–40 g wet weight corresponding to 1–10 g dry weight) were From the sponge L. nivea,aMicrobulbifer sp. strain L4-n2 immersed after collection in a CH2Cl2/MeOH 1/1 mixture at was isolated and shown to produce novel parabens with room temperature for 2 days, concentrated under reduced antibacterial activity against Gram positive reference pressure and extracted with CH2Cl2. bacteria S. aureus as well as against marine Bacillus sp. and Planococcus sp. isolates derived from the sponge. Its Antimicrobial assays major paraben metabolite was also detected in the calcareous sponge at all seasons, contributing to its year- They were performed by the disk diffusion assay method: round antimicrobial activity. The Microbulbifer strain L4-n2 1 mg of each extract (from bacteria or sponge) was solubi- was localized by fluorescence in situ hybridization in the lized in CH2Cl2 and applied directly onto a 6 mm paper disk or deposited in 6 mm wells cut out from agar plates seeded sponge body wall. Our study is the first to show that a with reference S. aureus (ATCC 6538), E. coli (ATCC 8739), paraben metabolite is persistently produced by a Micro- marine V. splendidus LGP32 (provided by Dr F. Leroux 2005) bulbifer bacterial strain within its sponge host. strains. Native Gram positive marine strains, Bacillus sp. and These results suggest that the L4-n2 Microbulbifer Planococcus sp. isolated from the cultured fraction of L. nivea strain contributes to the chemical profile of the Calci- heterotrophic bacterial flora were also tested. Antimicrobial sponge L. nivea, and that natural parabens might have an activity was determined by measuring the diameter of the ecological role in the balance between bacterial popula- inhibition zones after 24 h incubation at 37°C for S. aureus and E. coli or 23°C for the marine strains. Activity testing was tions within the sponge. repeated in at least two to three independent experiments to Currently, studies are being carried out in order to cha- ensure reproducibility of results. racterize the interactions between the L4-n2 Microbulbifer strain and the other strains in the sponge-associated cultivable heterotrophic microbiota, especially to elucidate Isolation of the L4-n2 marine bacterial strain the structure of active compounds involved in the mainte- Sponge-associated bacteria were isolated following the nance of the diversity of the microbial community charac- method outlined by Hentschel and colleagues (2001). Briefly, teristic of marine sponges. a1¥ 1 cm section of sponge was rinsed three times in fil- tered (0.2 m) sea water and pressed through a 40 mm pore- size plankton net in calcium and magnesium-free artificial sea Experimental procedures water. Serial dilutions at 10-2 and 10-4 of the maceration were spread onto Marine Agar Difco 2216 supplemented with nali- General experimental procedures dixic acid (10 mgml-1) in order to inhibit fast growing Vibrios Optical rotations were obtained on a Perkin Elmer 341 pola- and increase the bacterial diversity recovered. All the plates rimeter. UV spectra were obtained in EtOH, using a Kontron- were incubated between 12°C and 20°C. After 14 days of type Uvikon 930 spectrophotometer. Silica gel column incubation, a brown pigmented spherical colony (L4-n2) was -2 dilution and propagated in pure culture chromatographies were carried out using Kieselgel 60 (230– isolated from the 10 on marine agar. 400 mesh, E. Merck). Fractions were monitored by thin-layer chromatography (TLC) using aluminium-backed sheets (Si gel 60 F254, 0.25 mm thick) with visualization under UV (254 Phylogenetic affiliation of strain L4-n2 and 365 nm) and Liebermann spray reagent. Analytical reversed-phase high-performance liquid chromatography PCR-based analyses of bacterial 16S ribosomal RNA gene. (HPLC) (Kromasil RP18 column K2185, 4.6 ¥ 250 mm) were Bacterial DNA was obtained from the strain isolated in pure performed with a L-6200 A pump (Merck-Hitachi) equipped culture with reproducible antibacterial activity in the soft agar with a UV-vis detector (l=254 nm) L-4250C (Merck-Hitachi) diffusion assay. Following the method modified from Sritha- and a chromato-integrator D-2500 (Merck-Hitachi). Mass ran and Barker (1991), colonies grown on Marine Agar were

© 2009 The Authors Journal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology 10 E. Quévrain, I. Domart-Coulon, M. Pernice and M.-L. Bourguet-Kondracki suspended in 60 ml sterile water, boiled for 5 min, centrifuged Extraction and isolation of active compounds (15 000 g, 5 min), and stored in aliquots at 4°C or -20°C prior to 16S rRNA gene PCR amplification, purification and direct The CH2Cl2 extract (100 mg) was chromatographed on a sequencing. silicagel (Merck silica gel 70–230 mesh) column using cyclo- Bacterial 16S rRNA gene amplification was performed hexane with increasing amounts of AcOEt as eluent. using PCR primers: 27F-1385R pair (respectively E. coli Bioassay-guided fractionation retained a maximum of activity position 9:5′-GAGTTTGATCCTGGCTCA-3′ and position in fractions eluted with 20% and 30% AcOEt. 1385:5′-CGGTGTGTRCAAGGCCC-3′). The reaction mixture Fraction eluted with 20% AcOEt (17 mg) was subjected contained 50 pmol of each primer, 2.5 mmol of each of deoxy- to reversed-phase HPLC column (C18 Uptisphere nucleoside triphosphate, 10¥ ATGC SuperTaq buffer, and 250 ¥ 7.8 mm) with increasing amount of CH3CN/H2Oas -1 0.4 U of ATGC SuperTaq and the volume was adjusted with eluent (flow rate: 1 ml min , wavelength: of 254 nm) to yield sterile water to 50 ml. PCR reactions were conducted in a compound 1 (retention time: 9 min 79, 1.2 mg), compound 2 Biotherma PCR system with an initial denaturing step (94°C (retention time: 12 min 38, 1.0 mg), compound 3 (retention for 5 min) followed by 32 cycles of 94°C for 1 min, 52°C for time: 20 min 98, 0.4 mg), compound 4 (retention time: 30 min 30 s and 72°C for 30 s and a final elongation step at 72°C for 78, 1.2 mg), compound 6 (retention time: 33 min 68, 2.0 mg), 7 min. Amplified DNA was checked by 1.5% agarose gel compound 8 (retention time: 39 min 74, 0.6 mg), compound 9 electrophoresis and purified with a QIAquick PCR purification (retention time: 31 min 94, 0.8 mg). kit (Qiagen). In the same manner, fraction eluted with 30% AcOEt (4.6 mg) was subjected to reversed-phase HPLC (C18 Upti- sphere 250 ¥ 7.8 mm) using increasing amount of CH CN/ Gene sequencing and phylogenetic affiliation. Purified 16S 3 H O as eluent (flow rate: 1 ml min-1, wavelength of 254 nm) to rRNA gene amplicons were directly sequenced in both direc- 2 yield compound 5 (retention time: 6 min 54, 0.6 mg), com- tions by Genome Express (Meylan, France). After sequence pound 7 (retention time: 7 min 54, 0.6 mg). assembly, almost full-length sequences (1304 nt) were com- Four cultures of 2.5 l were required to obtain enough quan- pared with those in the 16S DNA database (EMBL Prokary- tities of pure compounds for structural elucidation and evalu- ote) by using FASTA (Pearson and Lipman, 1988) and highly ation of antimicrobial activities. similar sequences of marine and sponge origin were included in the analysis. The 16S rRNA gene sequence data were The gradient of elution used to purify parabens was applied edited using the BioEdit software and aligned with the for LC/MS studies to detect compound 6 (retention time: Dialign2 software (http://mobyle.pasteur.fr). A phylogenetic 18 min) using a C18 Kromasil (250 ¥ 4.6 mm) column. Mass tree was then constructed by Maximum Likelihood (dnaml) spectra were obtained in positive ESI mode. methods in the Phylip software package (Felsenstein, 2002) Compound 1. Ethylparaben, 4-hydroxybenzoic acid ethyl with bootstrap analyses (1000 replicates) to test the robust- ester (0.64 mg per lire of culture, 0.80% of dry weight CH2Cl2 ness of each topology. crude extract of the bacteria culture), white powder; mp 114– + 115°C; UV (EtOH) lmax nm (e): 286 (364); ESI-MS [M+H] found at m/z 167.0708 (D-0.019 mmu), calc. 167.0708 for Media and growth conditions of the strain C9H11O3. Microbulbifer sp. Compound 2. Butylparaben, 4-hydroxybenzoic acid butyl ester (0.436 mg per litre of culture, 1.78% of dry weight The bacterial strain was grown in Marine Broth Difco 2216 CH Cl crude extract of the bacteria culture), white powder; (37.4 g l-1) in the dark and shaken at 350 r.p.m. at tempera- 2 2 mp 68–69°C; UV (EtOH) l nm (e): 285 (575); ESI-MS ture ranging from 8°C to 30°C for 1–3 days. The production of max [M+H]+ found at m/z 195.1028 (D-0.6 mmu), calc. 195.1093 bioactive compounds was maximized between 12°C and for C H O . 20°C in pigmented culture arrested at the beginning of the 11 15 3 Compound 3. Octylparaben, 4-hydroxybenzoic acid octyl stationary phase. ester (0.109 mg per litre of culture, 0.43% of dry weight

CH2Cl2 crude extract of the bacteria culture), white amor- Preparation of bacterial crude extract phous solid; UV (EtOH) lmax nm (e): 283 (175); ESI-MS [M+H]+ found at m/z 251.1652 (D-0.48 mmu), calc. 251.1647

A 25 ml inoculum, prepared from ﬁve strains in marine broth for C15H23O3. and incubated during 2 days between 12°C and 20°C with Compound 4. Decylparaben, 4-hydroxybenzoic acid decyl shaking, was transferred in three 2 l Erlenmeyer ﬂasks. Each ester (0.254 mg per litre of culture, 1.04% of dry weight bacterial culture was incubated for 2 days between 12°C and CH2Cl2 crude extract of the bacteria culture), pale yellow oil; + 20°C with shaking. Cells were separated by centrifugation at UV (EtOH) lmax nm (e): 285 (966); ESI-MS [M+H] found at

4°C during 15 min at 10 000 r.p.m. The supernatant was m/z 279.1962 (D-0.18 mmu), calc. 279.1960 for C17H27O3. immediately poured in CH2Cl2 overnight. The pellet was sub- Compound 5. Hydroxydecylparaben, 4-hydroxybenzoic acid mitted to six runs of freezing and sonication before to be 3-hydroxy-decyl ester (0.145 mg per litre of culture, 0.59% of poured in CH2Cl2 for extraction after 48 h. The organic layers dry weight CH2Cl2 crude extract of the bacteria culture), pale 20 were concentrated in vacuum. Examination by TLC revealed yellow oil, [a]D =-18° (c 0.14, MeOH), UV (EtOH) lmax nm identical chemical proﬁles for both supernatant and pellet (e): 286 (1378); ESI-MS [M+H]+ found at m/z 295.1913 (D

CH2Cl2 extracts of the culture. As they also showed identical -0.4 mmu), calc. 295.1923 for C17H27O4. activity, they were combined to constitute the CH2Cl2 crude Compound 6. Methyldecylparaben, 4-hydroxybenzoic acid extract of the culture, which was further extracted. methyl-decyl ester (0.509 mg per litre of culture, 2.08% of dry

© 2009 The Authors Journal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology Parabens from sponge associated bacteria 11 weight CH2Cl2 crude extract of the bacteria culture), pale preheated chambers wetted with hybridization buffer (0.9 M + yellow oil; UV (EtOH) lmax nm (e): 286 (687); ESI-MS [M+H] NaCl, 0.02 M Tris-HCl, 0.01% SDS, 1% blocking reagent, found at m/z 293.2115 (D-0.07 mmu), calc. 293.2116 for 10% dextran sulfate, 35–50% formamide). Temperature and

C18H29O3. stringency were evaluated at different formamide concentra- Compound 7. Hydroxymethyldecylparaben, 4-hydroxyben- tions (35–50%), to optimize the signal (high speciﬁc signal zoic acid 3-hydroxy-methyl-decyl ester (0.218 mg per litre of and low background ﬂuorescence) for each oligonucleotidic culture, 0.89% of dry weight CH2Cl2 crude extract of the probe. The eubacterial probe EUB338 (5′-GCT GCC TCC 20 bacteria culture), white amorphous solid; [a]D =-7° (c 0.10, CGT AGG AGT-3′) (Amann et al., 1990) was used as a + MeOH); UV (EtOH) lmax nm (e): 287 (2000); ESI-MS [M+H] positive control for a major part of the eubacterial biota found at m/z 309.2055 (D-0.5 mmu), calc. 309.2000 for (Daims et al., 1999). A 18 nt probe, named Ma445 (5′-AGC

C18H29O4. TTA TAG CCT TCC TCC-3′, on position 445 of E. coli 16S Compound 8. Dodecylparaben, 4-hydroxybenzoic acid rRNA gene) was designed for the Microbulbifer phylogenetic dodecyl ester (0.109 mg per litre of culture, 0.43% of dry cluster (L4-n2 strain and M. arenaceous AJ510266), using weight CH2Cl2 crude extract of the bacteria culture),white the PROBE_DESIGN function of ARB and the 16S rDNA amorphous solid; UV (EtOH) lmax nm (e): 285 (966); ESI-MS accessibility information provided by Behrens and colleagues [M+H]+ found at m/z 307.2278 (D-0.4 mmu), calc. 307.2273 (2003). Its specificity was checked with BLAST (Altschul et al., for C19H31O3. 1994), FASTA and probecheck (http://www.microbial- Compound 9. Dodec-5-enylparaben, 4-hydroxybenzoic acid ecology.net/probecheck) yielding two hits M. arenaceous dodec-5-enyl ester (0.145 mg per litre of culture, 0.59% of dry AJ510266 and the recently published Microbulbifer sp. YIM weight CH2Cl2 crude extract of the bacteria culture), white C306 EU135714. Specificity of Ma445 to Microbulbifer L4-n2 amorphous solid, UV (EtOH) lmax nm (e): 285 (334); ESI-MS was checked with several Gammaproteobacteria strain [M+H]+ found at m/z 305.2102 (D-1.4 mmu), calc. 305.2116 smears, as negative controls, including two Alteromonadales for C19H29O3 Moritella sp. L7-2nh and Pseudoalteromonas sp. L7-3a and with one Alphaproteobacteria Pseudovibrio sp. L4-8 strains isolated from L. nivea. For the internal negative controls we Determination of MIC and MBC used the 17 nt nonEUB probe (5′-CTC CTA CGG GAG GCA GC-3′) (Amann et al., 1990) and the 24 nt NW442 probe Minimal inhibitory concentration determination was carried (5′-AGT TAA TGT CAT TAT CTT CAC TGC-3′) raised against out by using a minor modification to the method described by Alphaproteobacteria strain NW001 which cross-reacts with Casteels and colleagues (1993). Minimal inhibitory concen- L4-8 and other demosponge-derived Pseudovibrio sp. strains trations were determined by a twofold dilution method (Enticknap et al., 2006; Taylor et al., 2007). Reactivity of -1 from the started concentration of 100 mgml of compound. Ma445 to L4-n2 was tested on pure bacterial smears of the -1 Hence, each compound was tested at 100 mgml through L4-n2 strain, with EUB338 as internal positive control and -1 0.048 mgml . Plates were incubated during 24 h at 30°C with nonEUB probe and NW442 probe as internal negative con- shaking (500 r.p.m.). trols. All probes were labelled with Alexa488 fluorochrome Minimal bactericidal concentrations were determined by (Invitrogen, Carlsbad, CA). Each section was covered with subculturing onto agar plates each well with no visible growth 50 ng of horseradish peroxidase-labelled probe in 150 mlof after 24 h incubation. The MBC was defined as the lowest hybridization buffer and then hybridized as described by < concentration of compound yielding colony counts 0.1% of Pernice and colleagues (2007). Stringency conditions were the initial inoculums, determined by colony counts from the 35% formamide for the EUB338, nonEUB338 and NW442 growth control well immediately after inoculation (Fuchs probes and 50% for the Ma445 probe and temperature was et al., 2002). Minimal bactericidal concentration or minimal lowered from 46°C to 42°C to improve signal with the strain- inhibitory concentration revealed the nature of compound. If specific probes. MBC/MIC is less than 4, the compound is bactericidal, if The slides were washed with 2¥ PBS for 10 min, MilliQ MBC/MIC is between 8 and 16, the compound is bacterio- water for 1 min, 96% ethanol for 1 min, air dried, and then static (Bartlett, 2008, http://www.medscape.com/viewarticle/ coverslipped in Prolong Antifade mounting medium with DAPI 478151_6). (Promega). Two sections per tissue in two different seasons of collection were visually examined for bacterial distribution and relative abundance using a DMLB epifluorescence In situ localization of a Microbulbifer strain with microscope (Leica Microsystèmes SAS, France). antimicrobial activity by CARD-FISH

Fragments 1 cm3 from L. nivea specimens were ﬁxed for Nucleotide sequence accession numbers CARD-FISH in Bouin ﬁxative overnight, then washed exten- sively and stored at 4°C in 70% ethanol. Fixed tissues were The EMBL accession number of the 16S rRNA sequence of dehydrated in an increasing series of ethanol and xylene and L4-n2 strain with antimicrobial activity described in this paper embedded in paraffin. Histological sections (4 mm) were col- is: FM200853 (Gammaproteobacteria Microbulbifer L4-n2). lected on gelatine coated slides and deparaffinized with xylene and rehydrated in a decreasing ethanol series. Sec- Supplementary data tions were pretreated with HCl, lyzozyme and proteinase K, to provide accessibility to the complementary 16S rRNA oli- 1H and 13C NMR data for all compounds are available on gonucleotide probes. Hybridization was performed for 3 h in request (Tables S1 and S2).

© 2009 The Authors Journal compilation © 2009 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology 12 E. Quévrain, I. Domart-Coulon, M. Pernice and M.-L. Bourguet-Kondracki

Acknowledgements biont common to many marine sponges and evidence for vertical transmission via sponge larvae. Appl Environ This work is part of Elodie Quévrain’s thesis, supported by a Microbiol 72: 3724–3732. grant from the French Ministère de l’Enseignement Supérieur Erickson, K.L., Cardellina, J.H., and Boyd, M.R. (1997) Sali- et de la Recherche. Funding was partly provided by a B.Q.R. cylihalamides A and B, novel cytotoxic macrolides from the grant from the MNHN (2005–2006). We thank Alain Blond, sponge Haliclona sp. J Org Chem 62: 8188–8192. Alexandre Deville, Arul Marie and Lionel Dubost (MNHN, Felsenstein, J. (1981) Evolutionary trees from DNA Paris) for NMR and MS measurements, Arlette Longeon, sequences: a maximum likelihood approach. J Mol Evol Gérard Gastine and Manon Vandervennet (MNHN, Paris) for 17: 368–376. their assistance in the antibacterial tests and Madeleine Felsenstein, J. (2002) PHYLYP: Phylogeny Inference Martin for preparation of histological sections. The authors Package, Ver 3.6. Seattle, WA, USA: University of also thank the director and personnel of the Laboratoire de Washington. Biologie Marine du MNHN, Concarneau, for providing facili- Fuchs, P.C., Barry, A.L., and Brown, S.D. (2002) In vitro ties during their stays and especially Dr S. Auzoux- bactericidal activity of daptomycin against staphylococci. Bordenave for help in field work. J Antimicrob Chemother 49: 467–470. Hentschel, U., Schmid, M., Wagner, M., Fieseler, L., Gernert, C., and Hacker, J. (2001) Isolation and phylogenetic References analysis of bacteria with antimicrobial activities from the Mediterranean sponges Aplysina aerophoba and Aplysina Altschul, S.F., Boguski, M.S., Gish, W., and Wootton, J.C. cavernicola. FEMS Microbiol Ecol 35: 305–312. (1994) Issues in searching molecular sequence databases. Hentschel, U., Usher, K.M., and Taylor, M.W. (2006) Marine Nat Genet 6: 119–129. sponges as microbial fermenters. FEMS Microbiol Ecol 55: Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W., 167–177. Devereux, R., and Stahl, D.A. (1990) Combination of 16S Hildebrand, M., Waggoner, L.E., Liu, H., Sudek, S., Allen, S., rRNA-targeted oligonucleotide probes with flow cytometry Anderson, C., et al. (2004) BryA: an unusual modular for analyzing mixed microbial populations. Appl Environ polyketide synthase gene from the uncultivated bacterial Microbiol 56: 1919–1925. symbiont of the marine bryozoan Bugula neritina. Chem Bartlett, J.G. (2008) Antibiotic selection for infections involv- Biol 11: 1543–1552. ing methicillin-resistant Staphylococcus aureus; bacteri- Jansen, R., Kunze, B., Reichenbach, H., and Höfle, G. (1996) cidal Chondramide A-D, new cytostatic and antifungal cyclodep- vs bacteriostatic agents [WWW Document]. URL http:// sipeptides from Chondromyces crocatus (myxobacteria): www.medscape.com/viewarticle/478151_6. isolation and structure elucidation. Liebigs Ann 2: 285–290. Behrens, S., Fuchs, B.M., Mueller, F., and Amann, R. (2003) Kunze, B., Jansen, R., Sasse, F., Höfle, G., and Reichen- Is the in situ accessibility of the 16S rRNA of Escherichia bach, H. (1998) Apicularens A and B, new cytostatic coli for Cy3-labeled oligonucleotide probes predicted by a macrolides from Chondromyces species (myxobacteria): three-dimensional structure model of the 30S ribosomal production, physico-chemical and biological properties. subunit? Appl Environ Microbiol 69: 4935–4941. J Antibiot 51: 1075–1080. Bewley, C.A., and Faulkner, J.D. (1998) Lithistid sponges: Moore, B. (1999) Biosynthesis of marine natural products: star performers or hosts to the stars. Angew Chem Int Ed microorganisms and macroalgae. Nat Prod Rep 16: 653– 37: 2163–2178. 674. Blunt, J.W., Copp, B.R., Hu, W.P., Munro, M.H.G., Northcote, Muscholl-Silberhorn, A., Thiel, V., and Imhoff, J. (2008) P.T., and Prinsep, M.R. (2007) Marine natural products. Nat Abundance and bioactivity of cultured sponge-associated Prod Rep 24: 31–86. bacteria from the Mediterranean Sea. Microb Ecol 55: Blunt, J.W., Copp, B.R., Hu, W.P., Munro, M.H.G., Northcote, 94–106. P.T., and Prinsep, M.R. (2008) Marine natural products. Nat Pearson, W.R., and Lipman, D.J. (1988) Improved tools for Prod Rep 25: 35–94. biological sequence comparison. Proc Natl Acad Sci USA Casteels, P., Ampe, C., Jacobs, F., and Tempst, P. (1993) 85: 2444–2448. Functional and chemical characterization of Hymenoptae- Peng, X., Adachi, K., Chen, C., Kasai, H., Kanoh, K., Shizuri, cin, an antibacterial polypeptide that is infection-inducible Y., and Misawa, N. (2006) Discovery of a marine bacterium in the honeybee (Apis mellifera). J Biol Chem 268: 7044– producing 4-hydroxybenzoate and its alkyl esters, para- 7054. bens. Appl Environ Microbiol 72: 5556–5561. Crews, P., and Bescansa, P. (1986) Sesterterpenes from a Pernice, M., Wetzel, S., Gros, O., Boucher-Rodoni, R., and common marine sponge, Hyrtios erecta. J Nat Prod 49: Dubilier, N. (2007) Enigmatic dual symbiosis in the excre- 1041–1052. tory organ of Nautilus macromphalus (Cephalopoda: Nau- Daims, H., Brühl, A., Amann, R., Schleifer, K.H., and Wagner, tiloidea). Proc Biol Sci 274: 1143–1152. M. (1999) The domain-specific probe EUB338 is insuffi- Piel, J. (2004) Metabolites from symbiotic bacteria. Nat Prod cient for the detection of all bacteria: development and Rep 21: 519–538. evaluation of a more comprehensive probe set. System Piel, J. (2006) Bacterial symbionts: prospects for the sustain- Appl Microbiol 22: 434–444. able production of invertebrate-derived pharmaceuticals. Enticknap, J., Kelly, M., Peraud, O., and Hill, R.T. (2006) Curr Med Chem 13: 39–50. Characterization of a culturable alphaproteobacterial sym- Proksch, P., Edrada, R.A., and Ebel, R. (2002) Drugs from the

sea – current status and microbiological implications. Appl Vacelet, J., and Donadey, C. (1977) Electron microscopy Microbiol Biotechnol 59: 125–134. study of the association between some sponges and bac- Richelle-Maurer, E., Braekman, J.C., De Kluijver, M., Gomez, teria. J Exp Mar Biol Ecol 30: 301–314. R., Van de Vyver, G., Van Soest, R., and Devijver, C. Wilkinson, C.R. (1978) Microbial associations in sponges. I. (2001) Cellular location of (2R, 3R, 7Z) -2-aminotetradec- Ecology, physiology and microbial populations of coral reef 7-ene-1, 3-diol, a potent antimicrobial metabolite produced sponges. Mar Biol 49: 161–167. by the Caribbean sponge Haliclona vansoesti. Cell Tissue Zabriskie, T.M., Klocke, J.A., Ireland, C.M., Marcus, A.H., Res 306: 157–165. Molinski, T.F., Faulkner, D.J., et al. (1986) Jaspamide, a Salomon, C.E., Magarvey, N.A., and Sherman, D.H. (2004) modiﬁed peptide from a Jaspis sponge, with insecticidal Merging the potential of microbial genetics with biological and antifungal activity. J Am Chem Soc 108: 3123– and chemical diversity: an ever brighter future for marine 3124. natural product drug discovery. Nat Prod Rep 21: 105–121. Schreiber, A., Wörheide, G., and Thiel, V. (2006) The fatty acids of calcareous sponges (Calcarea, Porifera). Chem Supporting information Phys Lipids 143: 29–37. Sritharan, V., and Barker, R.H., Jr (1991) A simple method for Additional Supporting Information may be found in the online diagnosing M. tuberculosis infection in clinical samples version of this article: using PCR. Mol Cell Probes 5: 385–395. Table S1. 1H NMR data of compounds 1–9 recorded in Tanaka, T., Yan, L., and Burgess, J.G. (2003) Microbulbifer CDCl [d in ppm (multiplicity, J in Hz)]. arenaceous sp. nov., a new endolithic bacterium isolated 3 H Table S2. 13C NMR data of compounds 1–9 in CDCl (d in from the inside of red sandstone. Curr Microbiol 47: 412– 3 C p.p.m.). 416. Taylor, M.W., Hill, R.T., Piel, J., Thacker, R.W., and Please note: Wiley-Blackwell are not responsible for the Hentschel, U. (2007) Soaking it up: the complex lives of content or functionality of any supporting materials supplied marine sponges and their microbial associates. ISME J 1: by the authors. Any queries (other than missing material) 187–190. should be directed to the corresponding author for the article.