Homogenates from 11 taxonomically different marine sponges collected from Ras Mohamed, Egypt and Rovinj, Croatia were plated on a range of selective media for the isolation of actinomycetes. Callyspongia aff. implexa yielded the highest number of isolates (22), followed by Aplysina aerophoba (16), Spheciospongia vagabunda (11), Hemimycale culumella (9), Hyrtios erecta (8), Dysidea tupha (6), Callyspongia sp. (6), Dysidea avara (4), Amphimedon sp. (4), Negombata magnifica (4), while no actinomycete strains were cultivated from Ircinia fasciculata (Figure 1A). A comparison of the nearly complete 16S rRNA gene sequences of 52 out of 90 putative actinobacterial strains isolated from these sponges against sequences in the NCBI GenBank database revealed the phylogenetic affiliations to 18 different genera representing 14 families and seven suborders (Supplementary Table 1). The highest number of isolates were found to be affiliated with the genus Mycobacterium (13), followed by Micrococcus (12), Micromonospora (11), Microbacterium (11), Brevibacterium (9), Kocuria (6), Corynebacterium (5), Rhodococcus (5), Brachybacterium (3), Rubrobacter (2), Streptomyces (2), Dietzia (2), Salinispora (2), Actinokineospora (2), Gordonia (2), Arthrobacter (1), Nocardiopsis (1) and Rothia (1) (Figure 1B). It has been reported that the genera belonging to Micromonospora, Rhodococcus and Streptomyces are the dominant actinobacterial genera in marine environments [25]. Representatives of all three taxa were isolated from the marine sponges studied here although other genera such as Mycobacterium, Micrococcus, Microbacterium were also represented by high numbers of isolates.

Given that less than 1% of bacteria associated with sponges can be cultured, the use of appropriate isolation media is critical for improving the recovery of actinobacteria [14]. The isolation media of this study were chosen to select for actinobacteria. M1 exhibited the highest recovery producing 27 isolates and MA recovered only two isolates (Figure 1C). In terms of diversity, M1 also produced isolates belonging to the highest number of genera (13), followed by ISP medium 2 (11 genera), R2A (eight genera), AIA (six genera), M1 plus (five genera), OLIGO (four genera), GAA (two genera), and the lowest diversity was observed with MA which recovered only one genus. This variation likely reflects the effect of media composition, which is consistent with previous observations [14,25]. Addition of aqueous sponge extract to M1 medium resulted in the isolation of a putatively new genus with Rubrobacter as its closest relative. This shows that variation from standard protocols is a worthwhile procedure. It was reported that media without added sponge extract produced the largest number of different colony morphotypes and the addition of sponge extract resulted in an increase in the number of novel cultivated morphotypes [26].

Interestingly, fourteen strains exhibited only 92.9–98.2% sequence similarities with validly described species. The low sequence similarities suggest that these strains belong to novel actinomycete taxa which are further supported by phylogenetic analysis [27] (Figures 2 and 3). These putatively novel isolates were affiliated to the genera Rothia, Kocuria, Micrococcus, Brachybacterium, Brevibacterium, Microbacterium (suborder Micrococcineae) (Figure 2). The isolates EG4 and EG62 from Callyspongia sp. and Amphimedon sp., respectively, formed a separate cluster from close relatives of the genus Microbacterium. The high sequence similarity (99.8%) of the isolates with only three nucleotide differences suggests that they are the same strain although they were isolated from different sponge species. The other putatively novel isolates were affiliated with the following genera: Mycobacterium and Rhodococcus (suborder Corynebacterineae), Actinokineospora (suborder Pseudonocardineae), Streptomyces (suborder Streptomycineae) and Rubrobacter (suborder Rubrobacterineae) (Figure 3). Interestingly, isolate RV113 from Aplysina aerophoba exhibited 92.9% sequence similarities to Rubrobacter radiotolerans as well as to several uncultured clones. Phylogenetic analysis revealed that the isolate RV113 forms a distinct clade and the low sequence similarity values further suggest that RV113 possibly belongs to a novel genus with Rubrobacter as its closest relative. Phenotypic and genotypic characterization will be pursued to validate the taxonomic position of this strain as well as the remainder of putatively novel actinomycete isolates cultivated in this study.

Several studies have previously reported on the isolation and diversity of actinobacteria from marine sponges. Montalvo et al. [13] showed that actinobacteria are major components of the microbial communities of Xestospongia muta and X. testudinaria. Zhang et al. [15] isolated 106 actinomycete strains representing seven genera from the sponge Hymeniacidon perleve. Additionally, 181 culturable actinomycetes affiliated with three genera were recovered from five sponges offshore China [25]. The microbial diversity of two Red Sea sponges, Hyrtios erectus and Amphimedon sp., were recently inspected using cultivation and cultivation-independent analyses. Focused cultivation on actinobacteria yielded 35 actinomycetes represented by four genera. [28]. Specifically, strains belonging to Actinoalloteichus, Brachybacterium, Brevibacterium, Curtobacterium, Gordonia, Kocuria, Micrococcus, Micromonospora, Nocardiopsis, Rhodococcus, Salinispora and Streptomyces were previously isolated from marine sponges and may therefore be common components thereof [13,15,28–30]. However, none of the strains isolated in our study showed sequence similarities with those previously reported from sponges, thus rendering the existence of sponge-specific actinomycete clades in the sense of Hentschel et al. [10] unlikely.

A major goal of our study is to establish new sources of antibiotics and other bioactive compounds. In order to prevent reisolation of already known compounds, which is a common problem in natural products discovery, the 14 putatively novel strains plus six additional ones based on their moderate relatedness to known metabolite producers were selected for bioactivity screening. The discovery rate of novel compounds is indeed much higher in taxonomically novel than in known strains [31]. The antimicrobial activity was assessed by the disc diffusion assay while the anti-parasitic activity was measured in multi-well plates and is represented as the percentage of growth inhibition in comparison to a reference compound. Of the twenty strains tested, ten produced antimicrobially active metabolites inhibiting at least one of the test pathogens. Five isolates were capable of inhibiting the growth of Gram-positives only, one isolate was active against C. albicans only and one isolate showed activity against both groups of pathogens (Table 1). The inhibition was greatest with the Streptomyces sp. RV15 against Staphylococcus aureus. Gram-negative bacteria were not susceptible to our isolates. In terms of the anti-parasitic activity, four isolates exhibited activity against T. brucei and two isolates showed activity against L. major (Table 2). The novel Actinokineospora sp. EG 49 showed highest antitrypanosomal potential at 48% growth inhibition. Anti-parasitic activities from marine-derived actinomycetes are reported only sporadically. For example, a new thiolactone antibiotic, thiolactomycin, from Nocardia strain No. 2–200 showed potent antimalarial and antitrypanosomal activities [32]. Additionally, the macrolide antibiotic amphotericin B produced by Streptomyces nodosus was used for treatment of leishmaniasis [33].

Because crude extracts were tested for bioactivities, the chemical nature of the bioactive compound is currently unknown. Nevertheless, some genera, to which the novel isolates belong to, are well known and prolific metabolite producers. For example, the genus Streptomyces accounts for about 80% of all natural products recovered from actinomycetes to date [34]. As one example, Streptomyces sannanensis produces new aminoglycoside antibiotics sannamycins A, B and C with potent activities against Gram-positive bacteria [35]. First organic chemistry analyses of the novel isolate RV15 points to entirely novel cyclic peptides which are currently being characterized further. Rhodococcus species, of which one putatively novel and bioactive strain was isolated in this study, produce a number of commercially interesting and potentially useful products including various types of steroids and peptides [36]. For example, two antimycobacterial cyclic peptides, lariatins A and B, were isolated from the culture broth of Rhodococcus sp. K01-B0171 [37]. The genus Microbacterium, of which two putatively new species were isolated in this study, produces different glycolipids. For example, five anti-tumor glycoglycerolipids were obtained from Microbacterium sp., which was isolated from the sponge Halichondria panicea [38]. Antimicrobial activities have, to our knowledge, not been reported for the genera Actinokineospora, Kocuria and Rubrobacter and the identification of metabolites produced by the putative new species of these genera is currently underway.

In conclusion, considerable actinobacterial diversity was recovered from various marine sponges. Altogether 90 actinomycete isolates were affiliated with 18 different genera including 14 putatively novel species. One possibly new genus related to Rubrobacter was isolated on M1 agar amended with sponge extract, thus highlighting the need for innovative cultivation media. Antibacterial, antifungal, antitrypanosomal and antileishmanial activities were reported for 10 of the isolates. Marine sponges represent therefore a still largely untapped resource for novel actinomycete diversity as well as for new secondary metabolites of therapeutic value.

The first group of sponges (Aplysina aerophoba, Dysidea avara, D. tupha, Hemimycale culumella, Ircinia fasciculata) was collected by SCUBA diving at depths of 3–20 m in the Mediterranean Sea (Rovinj, Croatia, (GPS: 27°47.655 N; 34°12.904 W) in August 2008. Taxonomic identification was performed by W.E.G. Müller and I. Müller (University of Mainz, Germany). The second group (Amphimedon sp., Callyspongia sp., C. aff. implexa, Hyrtios erecta, Negombata magnifica, Spheciospongia vagabunda) was collected at a depth of 10 m in the Red Sea (Ras Mohamed, Sinai, Egypt; (GPS: 27°47.655 N; 34°12.904 W) in August 2006. Amphimedon sp. and Hyrtios erecta were identified by M. Kelly (National Institute of Water and Atmospheric Research (NIWA) Auckland, New Zealand) and the remaining sponges by R.W.M. van Soest (University of Amsterdam, Netherlands). Sponges were transferred to plastic bags containing seawater and transported to the laboratory. Sponge specimens were rinsed in sterile seawater, cut into pieces of ca. 1 cm³, and then thoroughly homogenized in a sterile mortar with 10 volumes of sterile seawater. The supernatant was diluted in ten-fold series (10⁻¹, 10⁻², 10⁻³) and subsequently plated out on agar plates.

Eight different media [M1 [19], ISP medium 2 [39], Oligotrophic medium (OLIGO) [40], M1 plus [14], Actinomycete Isolation Agar (AIA) [41], Marine Agar (MA) [42], Glycerol Asparagine Agar (GAA) [41] and R2A Agar [43]] were used for the isolation of actinobacteria. All media were supplemented with 0.2 μm pore size filtered cycloheximide (100 μg/mL), nystatin (25 μg/mL) and nalidixic acid (25 μg/mL) to facilitate the isolation of slow-growing actinobacteria. Cycloheximide and nystatin inhibit fungal growth, while nalidixic acid inhibits many fast-growing Gram-negative bacteria [14]. All media contained Difco Bacto agar (18 g/L) and were prepared in 1 L artificial sea water (NaCl 234.7 g, MgCl₂.6 H₂O 106.4 g, Na₂SO₄ 39.2 g, CaCl₂ 11.0 g, NaHCO₃ 1.92 g, KCl 6.64 g, KBr 0.96 g, H₃BO₃ 0.26 g, SrCl₂ 0,24 g, NaF 0.03 g and ddH₂O to 10.0 L) [44]. To promote the growth of selected sponge-associated actinobacteria, 1% “aqueous sponge extract” was added to the autoclaved medium. Aqueous sponge extract was prepared by grinding 20 g of sponge biomass in a mortar containing 20 mL of sterile seawater followed by centrifugation (5000 rpm, 10 min) and sterilized by filtration through a 0.2 μm pore size filter. The freshly prepared supernatant served as aqueous sponge extract. The inoculated plates were incubated at 30 °C for 6–8 weeks. Distinct colony morphotypes were picked and re-streaked until visually free of contaminants. Isolates were inoculated into liquid media (M1 or the medium on which colonies were initially isolated). The isolates were maintained on plates for short-term storage and long-term strain collections were set up in medium supplemented with 30% glycerol at −80 °C. The isolates from Egypt are abbreviated as “EG” and from Rovinj as “RV”.

16S rRNA gene amplification, cloning and sequencing were performed according to Hentschel et al. [45] using the universal primers 27F and 1492R [46]. Chimeric sequences were identified by using the Pintail program [47]. The genus-level affiliation of the sequences was validated using the Ribosomal Database Project Classifier [48]. Sequence alignment and phylogenetic analysis were performed using the ARB software package [49]. Tree construction was conducted using neighbour-joining algorithm (Jukes-Cantor correction) with bootstrap values based on 1000 replications. The 16S rRNA gene sequences of the putatively novel isolates were deposited in GenBank under the accession numbers indicated in parentheses: EG4 (GU318354), EG7 (GU318355), EG33 (GU318356), EG36 (GU318357), EG37 (GU318358), EG45 (GU318359), EG47 (GU318360), EG49 (GU318361), EG62 (GU318362), EG69 (GU318363), RV15 (GU318364), RV113 (GU318365), RV13 (GU318366) and RV89 (GU318367).

Fourteen strains selected based on phylogenetic novelty and six selected based on their affiliation to known metabolite-producers were cultured in 100 mL Erlenmeyer flasks containing 50 mL of 5 different production media (M1, ISP2, OLIGO, AIA and R2A) for each isolate. The liquid cultures were grown for 7–14 days depending on their growth rate at 30 °C while shaking at 150 rpm. An equal volume of methanol was added to the liquid cultures for cell lysis and shaking was continued (150 rpm, 1 h at room temperature; Shaker SM 30, E. Bühler). The broth was centrifuged in 50 mL falcon tubes (5000 rpm, 15 min at room temperature; Megafuge 1.0R, Heraeus) and the supernatant was stored at 4 °C.

The in vitro antimicrobial activity testing was carried out using the standard disk diffusion assay [50] against pathogenic bacteria (Staphylococcus aureus strain 8325, Enterococcus faecalis strain JH212, Escherichia coli strain 536, Pseudomonas aeruginosa strain Nr. 3) and yeast (Candida albicans strain S314). Sterile filter disks (6 mm) impregnated with actinomycete extracts were placed on agar plates that had been inoculated with the test pathogen. After 24 h incubation at 37 °C (bacteria) and 30 °C (yeast), the antimicrobial potential was quantitatively assessed as diameter of the inhibition zone (n = 2). Anti-leishmanial activity was tested following the method of Ponte-Sucre et al. [51]. Briefly, this involved the incubation of Leishmania major promastigotes for 24 h at 26 °C, 5% CO₂, and 95% humidity in the absence or presence of the extracts. Following the addition of Alamar Blue, the plates were incubated again and the optical densities were determined after 48 h with a Multiskan Ascent enzyme-linked immunosorbent assay (ELISA) reader (Thermo Electron Corporation, Dreieich, Germany). Absorbance in the absence of compounds was set as 100% of growth. Amphotericin B was used as a reference compound and positive control. Each extract was assayed in duplicate from two independent experiments. Anti-trypanosomal activity was tested following the protocols of Huber and Koella [52]. Following cultivation of Trypanosoma brucei brucei strain TC 221 in Complete Baltz Medium, a defined number of parasites were exposed in 96-well plate test chambers to the extracts. The plates were then incubated at 37 °C in an atmosphere of 5% CO₂ for 24 h. After addition of Alamar Blue, the activity of the extracts was measured by light absorption using MR 700 Microplate Reader after 48 h. Absorbance in the absence of compounds was set as 100% of growth. Each extract was assayed in duplicate from two independent experiments.