Mn(II)-oxidizing bacteria were isolated from tissue of S. domuncula during a selection procedure on Mn agar plates in modified K-medium, containing 100 μM MnCl₂, as described under “Experimental Section”. The isolated strain BAC-SubDo-3 was found to oxidize MnCl₂ (present in the K-medium) or MnCO₃ (in the agar) as deduced from the color change. In the absence of Mn in the culture system the medium (Figure 1A), as well as the colonies on the agar (Figure 1C), remained whitish. In contrast, in Mn-containing liquid cultures (Figure 1B) or on agar (Figure 1D) a shift from whitish to brownish was seen. The morphology of the bacterial streaks on agar changed during the process of Mn precipitation. While in the absence of Mn in the agar, the streaks were more or less continuous (Figure 1E) the colonies on Mn-containing agar formed islands of sizes of around 50 μm² surrounded by Mn deposits (Figure 1F).

The growth rate of strain BAC-SubDo-3 depended on the presence of MnCl₂ in the medium. In the absence of any Mn a longer lag phase of 20 to 30 hours was observed [the optical density of the 30 hours-old cultures reached a value of 0.8 units] followed by a short (from 30 to 40 hours) increase in cell density, prior to the stationary phase which was reached after 40 hours (Figure 2A). Addition of 100 μM MnCl₂ shortened the lag phase to 6 to 10 hours (the optical density after 30 hours: 2.95 units); the stationary phase was reached already after 16 to 20 hours (Figure 2B).

The Mn(II)-oxidizing Bacillus strain BAC-SubDo-03 was phylogenetically grouped by analyzing its 16S rDNA (16S-BAC-SubDo-03) by comparison with the data bases. Highest sequence similarity was found with the Mn(II)-oxidizing spores isolated from the Guaymas Basin [11], and most closely related to strain GB02-14C (Figure 3).

In previous studies it was indicated that primarily spores have the capacity to oxidize Mn(II) to Mn(IV) [21,22]. Therefore, we determined in the BAC-SubDo-03 cultures, the fraction of spores throughout the growth cycle of the bacteria. The spores were identified by staining with malachite green followed by counterstaining with safranin. During these reactions the spores stained dark-red/green (Figure 4A to C).

In the absence of MnCl₂, first spores could be detected after 20 hours of incubation in K-medium, and the fraction increased steadily until the maximum with 20% was reached after 60 hours (Figure 2A). In contrast, in the presence of 100 μM of MnCl₂ spore-formation started already after 10 hours and reached the maximum of 40% already after 40 hours of incubation (Figure 2B).

SEM images were taken from BAC-SubDo-03 cultures grown for 10 hours in the absence or presence of MnCl₂. In the absence of MnCl₂ all bacteria have a barrel-like morphology (Figure 4D to F). In contrast, cultures that were incubated for 10 hours in the presence of 100 μM of MnCl₂ contained bacteria with a different morphology. Those microorganisms had distinct elongated projections that are labeled here as endospores/spores (Figure 4G to I). In parallel, the cells were stained with malachite green/safranin (Figure 4A to C).

The Mn-oxidizing activity was determined in cell wall extracts from BAC-SubDo-03 endospores after cultivation of the bacteria in K-medium in the presence of 100 μM MnCl₂ for 24 hours. The outermost layers (cell walls) from the spores were obtained by physical treatment, as described in “Experimental Section”, and the proteins were analyzed by SDS-PAGE. The gels were either directly stained with Coomassie brilliant blue (Figure 5; left panel) or incubated in the presence of 100 μM MnCl₂ for the identification of the Mn-oxidizing activity in situ (Figure 5; right panel). For this in situ determination assay, the samples were either left unheated (n-ht) or were heated at 96°C for 5 minutes (ht). Evidently a 110-kDa protein band developed only from the unheated sample (Figure 5, right panel [n-ht], whereas no band became visible from the heated sample (Figure 5, right panel [ht]). In an additional control, it was shown that no 110-kDA band appeared when the gels were incubated in the presence of 100 μM MnCl₂ plus 10 mM Na-azide (data not shown). This finding is consistent with experimental data that indicated the involvement of a metalloprotein (multicopper oxidase) in Mn-oxidation [16].

Cultures of BAC-SubDo-03 grown in K-medium supplemented with different concentrations of MnCl₂ for periods of six to 24 hours were analyzed for their capacity to oxidize Mn(II) by the leucoberbelin blue reagent assay. The data revealed that cultures grown in the absence of MnCl₂ showed a low Mn(II)-oxidizing activity that ranged between 0.5 to 1.0 μM of oxidized Mn(II) (Figure 6). However, when the bacteria were grown in the presence of MnCl₂ a time- and concentration-dependent increase in the Mn(IV)-product formation was measured. While the increase of the Mn(II)-oxidizing activity at a concentration of 50 μM of MnCl₂ was not significant during the 6 to 24 hours incubation period a significant increase in oxidation was seen in cultures that were incubated with 100 μM MnCl₂ for 24 hours. Higher concentrations of MnCl₂ (150 μM and 200 μM MnCl₂) resulted in a further increase of the activity. At the highest concentration tested (200 μM MnCl₂), the cells formed 2.8 μM oxidized Mn after incubation for only 12 hours (Figure 6). This value even increased when the incubation of the bacteria was extended to 24 hours (5 μM).

Based on this finding, we conclude that Mn(II) in the culture medium induced the expression of the MCO in a time- and concentration-dependent manner.

In view of earlier studies [23] we performed experiments to also determine the amount of manganese bound to or oxidized by the microorganisms BAC-SubDo-03. Therefore these bacteria were cultivated in medium containing radiolabeled manganese for up to 50 minutes (Figure 7). Then aliquots were taken from the cultures and the radioactivity was determined. The data revealed that the total amount of manganese immobilized on the bacteria increased almost linearly during the 50 minute incubation. After 50 minutes 10.7 ± 1.9 nmoles of Mn were associated with bacteria based on mg of protein.

In order to test whether copper ions are able to stimulate Mn-oxidation activity of BAC-SubDo-03, the bacterial cells were incubated in the presence of 200 μM of MnCl₂ (like in the experiment summarized in Figure 6) together with increasing concentrations of Cu ions (1.0 to 6.0 μM CuCl₂). As expected, the Mn-oxidizing activity in parallel cultures incubated in the presence of 200 μM of MnCl₂ varied between 4 to 10 μM during the 24 hours incubation (Figure 8). However, if Cu ions (1.0 μM) were added, the activity increased drastically to 90 μM during the 24 hours of incubation. This level rose to 138 μM when the concentration of Cu ion was increased to 6 μM (Figure 8). This result supports further the notion that a MCO exists on the surface of the bacterial endospores.

The published sequences from the multicopper oxidases were used to design degenerate primers [11,12], which then were applied successfully to identify the corresponding oxidase of the BAC-SubDo-03 strain. The sequence was termed mnxG-SubDo-03. The deduced sequence fragment was 235 amino acids in length and spanned the central part of the putative MnxG Cu-binding regions D and A/B [13]; Figure 9A. The consensus amino acids involved in the binding of the MCO to Cu are highlighted under the respective domains C/D and A/B (Figure 9A). To elucidate the relationship with other enzymes the MnxG-SubDo-03 polypeptide deduced from the mnxG-SubDo-03 sequence was compared with other oxidases involved in Mn-oxidation [11–13]. This analysis revealed that the sponge/microorganism-related oxidase (MnxG-SubDo-03) shared the highest sequence relationship with the oxidase from the Bacillus strain GB02-14C (Figure 9B). These two sequences formed one branch together with the PL-12/PL-16 cluster.

The MCO-specific primers were also used to check if, in the seawater in which the S. domuncula specimens live, BAC-SubDo-03 or related strains could be identified. Seawater was taken, centrifuged, and the sediment was suspended in the specific PCR reaction mixture supplemented with the specific primers for the MnxG gene. However, no bacteria containing the MCO gene were found in any one of the six separate experiments. In continuation—and in order to get a hint as to whether those bacteria were absent in the surrounding seawater—we applied the cetyltrimethylammonium bromide (CTAB) extraction procedure, described originally by us [24]. Again no signals were obtained after applying the PCR technique to screen for the existence of the MCO gene (data not shown).

In a semi-quantitative approach, the expression level of the mnxG gene in BAC-SubDo-03 (mnxG-SubDo-03) was determined by RT-PCR in the absence or presence of Mn. The schedule of Mn treatment is outlined in Figure 2C. In the absence of Mn in the culture medium, no signals for mnxG-SubDo-03 transcripts were detected during an incubation period of 4 to 32 hours (Figure 10; lanes 1 to 3). However, when cultures that had been grown in the absence of Mn for 32 hours were then exposed to 100 μM MnCl₂, a strong 119 bp long band reflecting the expected size of the amplified stretch of the mnxG-SubDo-03 transcripts could be identified after ethidium bromide staining (Figure 10; lane 4). The level of transcripts could still be recognized in BAC-SubDo-03 after a MnCl₂ pulse of 1 hour (lane 5), but not anymore after 18 hours of exposure (lane 7). In parallel, amplifications the assays contained either the mnxG-SubDo-03 containing plasmid (positive control [pc]) or the empty plasmid (negative control [nc]); lanes 9 and 10. Controls were performed with samples not subjected to RT reaction, but no amplification signals were obtained (not shown).

In order to analyze the (potential) effect of the Mn-oxidizing bacteria BAC-SubDo-03 on sponge cell metabolism, the sponge 3D-cell cultures were inoculated with the microorganisms, as described under “Experimental Section”. After dissociation of the sponge tissue, the single cell suspension was allowed to re-aggregate to clumps of approximately 400–600 μm in diameter for one day in the absence or presence of MnCl₂ (Figure 11). The sizes of the cell clumps in both assays, in the absence or the presence of 50 μM MnCl₂, did not differ (6 parallel experiments have been performed); Figure 11A and C. Then the 3D-cell culture samples remained non-inoculated or were inoculated with BAC-SubDo-03 bacteria, and were incubated in the absence or presence of 50 μM MnCl₂ for an additional three days. During that period the 3D-cell clumps formed primmorphs with smooth surfaces. While primmorphs not incubated with BAC-SubDo-03 grown in the absence or presence of MnCl₂, or primmorphs inoculated with BAC-SubDo-03 and growing in MnCl₂-free medium, reached only diameters between 2 and 3 mm (Figure 11B); those primmorphs which had been inoculated with BAC-SubDo-03 and cultured in the presence of 50 μM MnCl₂ reached sizes of 5–6 mm (Figure 11D).

Histological analyses ensured that the bacteria had colonized the primmorphs. The primmorph slices were stained with hematoxylin/eosin in order to visualize and contrast the canal-like structures [25]. With this procedure no bacteria could be identified in primmorphs which had not been inoculated with BAC-SubDo-03 (Figure 11E). This was in contrast to primmorphs which had been co-incubated with BAC-SubDo-03; slices from those 3D-cell aggregates comprised deeply stained band-like colonies (Figure 11F). It should be mentioned here that the sponge S. domuncula does not abundantly contain large communities of bacteria in its body [26]. In contrast to other sponge-associated bacteria that are in S. domuncula “encapsulated” in special sponge cells, bacteriocytes [26], the BAC-SubDo-03 bacteria exist extracellularly.

Primmorphs were incubated in the absence or presence of 50 μM MnCl₂. In parallel assays those 3D-cell aggregates remained either non-inoculated with Mn-oxidizing bacteria, or were inoculated with BAC-SubDo-03, as described under “Experimental Section”. After an incubation period of three days the primmorphs were harvested and RNA was extracted. The transcripts were size-separated and probed with the cDNA encoding the cell-cycle regulated cytoplasmic thymidine kinase-1 [27,28]. This thymidine kinase-1 gene is transcriptionally regulated and its expression is limited to the S phase of the cell cycle. In parallel, the expression level of the house-keeping gene β-tubulin [29] was assessed. In turn, the level of expression of the thymidine kinase-1 gene was normalized with the expression of β-tubulin. The normalized gene expression of thymidine kinase-1 was set to 1-fold in primmorphs, incubated in the absence of MnCl₂ (Figure 12). Addition of 50 μM MnCl₂ to the cultures resulted in an increase of the expression to 2.5. If the 3D aggregates were inoculated with BAC-SubDo-03 and cultured in the absence of MnCl₂ the expression of the thymidine kinase-1 gene only slightly increased to 1.5-fold. However, a simultaneous addition of 50 μM MnCl₂ to primmorphs co-incubated with BAC-SubDo-03 resulted in a marked increase of thymidine kinase-1 expression to 5.3-fold. From these findings we conclude that the co-existence of sponge cells with Mn-oxidizing bacteria BAC-SubDo-03 and MnCl₂ resulted in a pronounced stimulation of growth of the cells in the primmorphs, an argumentation that was also supported by the obvious size increase of primmorphs that had been inoculated with BAC-SubDo-03 and cultures in MnCl₂-enriched medium (Figure 11).

Reagents were purchased as follows: Sterile seawater, Mn Agar (cat no. M2053; containing beef extract (1 g/L), yeast extract (75 mg/L), Mn-carbonate (2 g/L (10 mM)), ferrous ammonium sulfate (150 mg/L), Na-citrate (0.15 g/L) and leucoberbelin blue from Sigma, Taufkirchen (Germany); malachite green oxalate, safranin O, bacto-peptone, yeast extract, Mn(II)-chloride tetrahydrate, Triton X-100 and 2-(N-morpholino)-ethanesulfonic acid [MES] came from Carl-Roth, Karlsruhe (Germany).

Specimens of the marine sponge Suberites domuncula (Porifera, Demospongiae, Hadromerida) were collected in the Northern Adriatic Sea near Rovinj (Croatia), and then kept in aquaria in Mainz (Germany) at 17 °C. For the isolation of the Mn-oxidizing bacteria only specimens that lived at least for 6 months under these conditions were used.

Tissue samples (200 mg) were taken and minced with seawater (1:1; wt/vol). Serial dilutions of the resulting suspension (10⁻¹ to 10⁻⁵) were prepared and aliquots of 0.1 mL were placed on Mn Agar plates (100 μM of MnCO₃) [55]. Plates were incubated at 28 °C for 48 up to 72 hrs during which the bacteria were allowed to oxidize Mn(II) ions. Brown colonies, indicative for Mn(II)-oxidizing bacteria, were picked from the agar plates and transferred to a modified K-medium (pH 6.5; 1 l consisting of 75% natural seawater, 2 g bacto-peptone, 0.5 g yeast extract, and 100 μM MnCl₂) as described [21]. In parallel assays, the microorganisms were cultivated on agar plates or in modified K medium lacking Mn. Sub-cultivation was performed by 1:10 dilution with fresh medium. Among the different

Mn-oxidizing bacterial strains (termed BAC-SubDo), strain BAC-SubDo-03 was selected because of its simple growth characteristics. In one series of experiments BAC-SubDo-03 was incubated in the presence of different concentrations of MnCl₂ (0 to 200 μM) in the absence or presence of 1.3 to 5 μM cupric chloride (CuCl₂). The cell concentration was kept constant within a set of experiments.

To check if the isolate BAC-SubDo-03 existed also in the seawater of the aquarium, 3 mL water samples were taken for PCR analysis and spun down at 11,000×g for 10 min at 4 °C. The pellets were resuspended in polymerase chain reaction (PCR) buffer and analyzed for the presence of the mnxG-SubDo-03 gene by PCR; see below. In a further series of experiments DNA was isolated from the seawater, following the CTAB (cetyltrimethylammonium bromide)-NaCl extraction procedure, described by us [24,56].

Cultures of BAC-SubDo-03 were set up in modified K medium (in the presence of 100 μM MnCl₂) and incubated at 28 °C. Under those conditions BAC-SubDo-03 started to form spores (endospores) after 10 hrs, as described also for other bacteria [21]. In parallel, BAC-SubDo-03 cultures were grown in modified K medium lacking MnCl₂. In order to quantify the fraction of spores in the culture, samples were taken and stained applying a described procedure [57]. Visualization of the spores was achieved by staining with 7.6% malachite green (in distilled H₂O) and after counterstaining with 0.5% safranin. Microscopic analysis was performed with an Olympus AHBT3 light microscope.

In addition, the morphology of the spore-forming bacteria was monitored, by scanning electron microscopy (SEM) analyses. Samples were fixed in 0.1% glutaraldehyde (pH 7.9; in 75% seawater and buffered with 50 mM HEPES) overnight. Then, the samples were mounted onto aluminum stubs (SEM-Stubs G031Z; Plano, Wetzlar; Germany) and subjected to SEM with a Gemini Leo 1530 high resolution field emission scanning electron microscope (Oberkochen; Germany).

Two assay systems were applied to verify that the BAC-SubDo-03 strain exhibits Mn(II)-oxidizing activity: (i) the detection of the Mn-oxidizing activity in the cell wall of the microorganisms by the in situ “in-gel oxidation assay”, and (ii) the detection of the conversion reaction of Mn(II) to Mn(IV) by the bacteria in the growth medium, applying an optical test.

The Mn-oxidizing activity of the outermost spore layer(s) was determined by the in situ “in-gel oxidation assay” [22,58]. The bacteria were pelleted by centrifugation (10,000×g; 4 °C), washed with deionized water, and suspended in 10 mM Tris buffer (pH 7.5). Then, the cells were treated with lysozyme (50 μg/mL; 30 min at 37 °C) to obtain spores. Subsequently, the samples were washed (five times) with deionized water (supplemented with 10 mM EDTA and 5% phenylmethane sulphonyl fluoride [pH 7.5]) and the spores were collected by centrifugation (10,000×g; 4 °C) and subsequently homogenized at 6,500 rpm in a Precellys 24 homogenizer (PeqLab Biotechnologie, Erlangen; Germany). The broken spores were subjected to sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) using 10% gels. After size separation the gels were stained for total protein with Coomassie brilliant blue. To assess Mn(II) oxidation activity a described procedure was used. At first, the gels were incubated for 30 min in a 10 mM Tris buffer (pH 7.5; 0.5% Triton X-100, 10% [v/v] glycerol) to remove the NaDodSO₄ and then the gels were incubated in 10 mM HEPES buffer (pH 7.6; 200 μM MnCl₂). The process of Mn(II) oxidation could be followed visually by the formation of brown Mn oxide bands on the gels. Those bands usually appeared after 2 hrs of incubation. Previous data indicated/demonstrated that the Mn(II)-oxidizing activities, mediated by spore surface proteins, is due to the copper (Cu)-dependent MCO [12,58]. Therefore, we used the Cu-ion chelator o-phenanthroline to check if the Mn-oxidation displayed by BAC-SubDo-03 spores can be attributed to a MCO as well. Consequently, incubation of the gels was performed with 50 μM of o-phenanthroline (in 10 mM HEPES buffer at pH 7.6) for 15 min prior to the addition of 200 μM MnCl₂. Under those conditions the color reaction of the band in the gel was totally suppressed.

The Mn(II)-oxidizing reaction proceeding in the medium was followed by an optical assay that is based on leucoberbelin blue that binds to Mn(III) and Mn(IV) [21,59]. The bacterial cultures grown in K-medium (supplemented with 50 μM to 200 μM MnCl₂) for 6 to 24 hrs were centrifuged and the sediment was used for the quantitation of Mn(IV). Aliquots of 0.05 to 0.25 mL were taken from the cultures and added to 0.5 mL of a 0.04% (w/v) leucoberbelin blue reagent. In the presence of Mn(III) or Mn(IV) leucoberbelin blue is oxidized with a concomitant appearance of a blue color [59]. This shift was monitored optically at 620 nm. K-permanganate was used to establish a calibration curve.

BAC-SubDo-03 were cultivated in modified K medium (in the presence of 100 μM MnCl₂) at 28 °C for 24 hrs. Then the microorganisms were transferred to new modified K-medium (containing 100 μM MnCl₂), supplemented with radiolabeled manganese (⁵⁴Mn²⁺; Amersham, Arlington, Heights, Ill.; U.S.; specific activity of 2.66 × 10⁶ Bq/μg). To 1 mL of culture 74 × 10⁴ Bq of ⁵⁴Mn²⁺ were added and cultivation was carried on at 28 °C for 50 min. Then aliquots of 100 μL were taken, filtered through filters (Whatman GF/C glass filters), transferred to counting vials that were filled with 4.5 mL scintillation cocktail, and then radioactivity was determined as described [23,60]. By using a standard the counting efficiency was determined. The values are given in nmoles of Mn bound/oxidized/mg of protein.

Single bacterial colonies were picked and the DNA was isolated by the “AxyPrep multisource Genomic DNA Miniprep Kit” (Axygen Biosciences, Union City, CA; U.S.), as described [61].

For polymerase chain reaction (PCR) amplification of 16S rDNA fragments we applied widely used primers [62]; forward primer 27f (5′-GAGTTTGATCCTGGCTCAG-3′ together with the reverse primer 1100r (5′-GGGTTGCGCTCGTTGC-3′). The PCR product spanned the nucleotides [nt] 9–1,113 of the Escherichia coli 16S rDNA gene (E. coli 16S rDNA sequence; accession number J01859).

The conditions for the PCR were as follows. The 50 μL reaction contained the undiluted PCR buffer (Peq-Lab), 2 mM MgCl₂, 1.5 U of Taq DNA polymerase (Peq-Lab), 0.2 mM dNTPs (Roth), 10 pmol of each primer, and 1 to10 ng of template DNA. PCR cocktails were subjected to 35 cycles of amplification in a PCR thermal cycler (iCycler, Bio-Rad). The cycling conditions were as follows: initial denaturation (2 min at 95 °C), followed by 35 cycles each of denaturation (25 sec at 95 °C), annealing (30 sec at 52 °C) and extension (1 min at 72 °C), and a final extension step (7 min at 72 °C). After the cycling 5 μL aliquots were resolved in 1% agarose/TAE [Tris-acetate-EDTA] gel, stained with ethidium bromide and visualized by UV-illumination.

PCR products were purified from the reaction mixture by the “NucleoSpin Extract II” kit (Macherey-Nagel, Düren; Germany) as recommended by the manufacturer. Purified PCR products were cloned into pGEM-T vector (Promega, Madison, WI; U.S.). Plasmid DNA was isolated by “High Pure Plasmid Isolation Kit” (Roche, Mannheim; Germany). Sequencing was performed by Thermo Sequenase DNA polymerase “Fluorescent Labelled Primer Cycle Sequencing” kit with 7-deaza-dGTP (Amersham Pharmacia Biotech, Freiburg; Germany) using primers, labeled with IRD-700 or IRD-800. Sequencing ladders were resolved in SequaGel XR gel (National Diagnostics, Atlanta, GA; U.S.) using a Li-Cor Sequencer-4300.

To eliminate possible PCR errors several clones were sequenced and a consensus was compiled. The resulting sequence of the 1,114 bp long 16S rDNA fragment was termed 16S-BAC-SubDo-03.

Genomic DNA of the strain BAC-SubDo-03 was isolated as described above. In order to isolate the mnxG gene from the sponge-associated microorganism BAC-SubDo-03 degenerated primers that had been designed according to published sequences of mnxG genes were used [11–13]. The expected PCR product of approximately 700 bp spans three out of the four/six copper-binding domains that are found in the mnxG cluster of Bacillus sp. strain SG-1 and some other Mn-oxidizing bacteria [10,11,13]. The forward primer was targeted to domain D (“HCHL-F” 5′-CAYTGYCAYYTRTAYCCNCA-3′ nt pos. 4313 to 4332 in strain SG-1 [accession number U31081.1]) and the reverse primer (“HQQH-R” 5′-CCAAATAYWCCATGYTGYTGRTG-3′ nt pos. 5079 to 5057 in SG-1) was targeted to domain B. The PCR mixture was composed as described above but containing 100 pmols of each of the mnxG primers. The PCR cycling conditions were as follows: initial denaturation for 2 min at 95 °C, and then 35 cycles each of denaturation (25 sec at 95 °C), annealing (30 sec at 47 °C) and extension (1 min at 70 °C). The final extension step was performed for 7 min at 72 °C. The completed PCR reactions were resolved in 1% agarose/TAE gel, stained in ethidium bromide and visualized by UV-illumination. PCR products of the expected size (approximately 700 bp) were cut out from the gel, cloned and sequenced as described above. The consensus derived from several sequenced clones of the mnxG gene from the S. domuncula-associated Bacillus strain BAC-SubDo-03 is 708 bp long and has been termed mnxG-SubDo-03.

The expression of the mnxG-SubDo-03 gene was semi-quantitatively assessed by RT-PCR (reverse transcription-PCR), selecting a segment of the mnxG gene as a target. Bacterial cultures in K-medium (without MnCl₂) were inoculated from fresh colonies and cultivated at 28 °C with shaking. At the indicated times after inoculation cultures were supplemented with 100 μM MnCl₂. Aliquots were taken and the cell density in each assay was adjusted to the same OD (2.0 absorbance at 660 nm). Starting with 5 mL of culture, total bacterial RNA was isolated by “AxyPrep Multisource Total RNA Kit” (Axygen). Satisfactory yields of RNA were obtained from BAC-SubDo-03 by homogenization of the cells in a “Precellys-24” bead grinder (PeqLab). For the homogenization “Axygen Cell Lysis Buffer R-I” was used together with 1 mm glass beads. Grinding was performed 3 times (5 sec each at 6,000 rpm). To remove DNA, isolated RNA was treated with DNAse (Ambion Inc) at 37 °C for 30 min. Subsequently, DNAse was removed by DNAse inactivation reagent (“DNA-free” kit, Ambion). cDNA was synthesized by M-MLV Reverse Transcriptase [RT] (RNAse H Minus) primed by random decamers in 40 μL final volume, as recommended by the manufacturer (Promega). Each reaction contained 10 μg of RNA, 0.5 mM dNTPs and 400 U of the RT enzyme. The reaction mixtures were incubated at 42 °C for 1 hr; the RT was inactivated by incubation at 65 °C for 15 min. The following primer pair was used for the amplification of the mnxG gene: gmnx-F 5′-ATGTGGGGTATCAATCGGATC-3′ (nt 34 to 54 in mnxG-SubDo-03; accession number FN565386) and qmnx-R 5′-GTAGGACTATGGTGGACAAGG-3′ (nt 152 to 133). One μg of transcribed RNA was added per reaction. PCR cycling conditions were as following: 95 °C-20 sec; 56 °C-30 sec; 70 °C-45 sec for 40 cycles. The size of the PCR product was 119 bp. Aliquots of 5 μL were loaded onto high resolution gels (4 cm long gels, prepared from 3% MS500 agarose (PeqLab) in 1 × TBE [Tris base-boric acid-EDTA; pH 8.0]. As a control the PCR experiments were performed with samples that had been obtained form RNA preparations that had not been subjected to RT reaction.

The sequences were analyzed with computer programs Blast (2005; http://www.ncbi.nlm.nih.gov/blast/blast.cgi) and FASTA (2005; http://www.ebi.ac.uk/fasta33/). Multiple alignments were performed as outlined in detail [63] and phylogenetic trees were constructed based on sequence alignment by neighbor-joining. The distance matrices were calculated using the Dayhoff/PAM matrix model as described [64]. The degree of support for internal branches was further assessed by bootstrapping. The graphic presentations were prepared with GeneDoc [65].

Primmorphs were obtained from single cells that had been obtained by dissociation from sponge tissue, as described [19]. The cells were transferred to natural seawater supplemented with 0.2% RPMI1640 medium. Where indicated the medium contained 50 μM MnCl₂. After one day the cultures were split into 1 mL assays and transferred into 6-well plates (Nunc, Langenselbold; Germany) and incubation was continued either in Mn-free or in 50 μM MnCl₂ containing medium. In parallel, primmorph samples were co-incubated with the bacteria BAC-SubDo-03. Aliquots of 50 μL from an overnight culture of BAC-SubDo-03 grown in modified K-medium in the absence of MnCl₂ (OD600nm of 1.5) were added to the primmorphs. Three days later the 3D-cell aggregates, the primmorphs, reached sizes of 2–6 mm and were used for histological and gene expression analyses.

Primmorphs were fixed in paraformaldehyde, embedded in Technovit 8100 and sectioned, essentially as described [66]. The 5 μm thick slices were stained with hematoxylin/eosin [67] and inspected with an Olympus AHBT3 microscope.

RNA was extracted from liquid-nitrogen pulverized primmorphs with TRIzol Reagent (GibcoBRL Grand Island, NY; U.S.) as recommended by the manufacturer. An amount of 3 μg of total RNA was electrophoresed through formaldehyde/agarose gel and blotted onto Hybond N+ membrane as described before [68]. Hybridization experiments were performed with the following probes: the S. domuncula thymidine kinase cDNA (nt_51–162 a nucleotide segment localized within the coding region of the deduced protein; accession number CAP19646 [27]) and the S. domuncula β-tubulin cDNA fragment (nt_83–423; CAD79598 [69]). The probes were labeled with DIG-11-dUTP by the DIG DNA labeling kit (Roche Diagnostics, Mannheim; Germany). After washing, DIG-labeled nucleic acid was detected with anti-DIG Fab fragments (conjugated to alkaline phosphatase) and visualized by the chemiluminescence technique using CDP-Star. Hybridization was performed under high stringency as described [70]. For the semiquantitative analysis of the expression level, the bands on the film were scanned with the GS-525 Molecular Imager (Bio-Rad, Hercules, CA; U.S.). The signals on the blots were quantified and the intensities reflecting the thymidine kinase transcripts were correlated with those for β-tubulin. The expression fractions (ratios between the intensities of the signals for thymidine kinase and those of the corresponding signals for β-tubulin) were calculated. The expression fraction measured for primmorphs which had been incubated without MnCl₂ and had not inoculated with BAC-SubDo-03 was set to 1-fold. Five parallel experiments were performed. The values did not differ by more than 20%.

For the quantification of protein the Bradford method ([71]; Roti-Quant solution - Roth) was used.

The results were statistically evaluated using the paired Student’s t-test [72].