Emended description of the species Lampropedia hyalina

A specimen of the square-tablet-forming, coccoid, heterotrophic genus Lampropedia was described for the first time in 1886 by Schroeter in polluted water samples (Schroeter, 1886). Since then, cells similar to Schroeter's description have been observed on several occasions in aquatic systems rich in organic material, rumen or the intestinal system in some animals (Murray, 1984). Our knowledge of this genus is limited, since it is based mainly on the phenotypic characteristics of one species, Lampropedia hyalina, which is represented by four strains deposited at the ATCC. However, this collection demands further investigation, since the type strain ATCC 11041^T has apparently lost the tablet-forming ability and grows instead as single coccoid cells. Murray (1984) thus stated that a neotype strain might have to be declared. Recently, Stante et al. (1997) isolated a tablet-forming Lampropedia-like specimen (DSM 15336) from activated sludge laboratory plants treating dairy and piggery wastewater that were operated as an enhanced biological phosphorus-removing system. This strain differs in at least one metabolic aspect from the Lampropedia hyalina type strain, i.e. the ability to accumulate polyphosphate under alternating aerobic–anaerobic conditions (Stante et al., 1997). The aim of our study was to compare the phenotypic, chemotaxonomic and molecular characteristics of this isolate with two Lampropedia hyalina strains from the ATCC isolated from other environments and to design specific oligonucleotide probes for fluorescent in situ hybridization for culture-independent identification of Lampropedia hyalina.

The following strains were used: Lampropedia hyalina ATCC 11041^T (=CCUG 48573^T=DSM 16112^T), isolated in 1966 by E. G. Pringsheim from liquid manure from a dairy farm yard; Lampropedia hyalina ATCC 43383 (=CCUG 48574), isolated by J. Kirchner from the rumen of a fistulated heifer (Hungate, 1966); and Lampropedia sp. DSM 15336 (=CCUG 48575), isolated by Stante et al. (1997) from activated sludge. For standard cultivations, all three strains were grown at pH 7·2 and at 25 °C on a medium containing moderate amounts of peptone, acetate, yeast extract and soil extract [ATCC culture medium 32; green top agar (GTA)] or on a modified version of this medium made by replacing the soil extract with filtered and sterilized effluent from a diary wastewater treatment plant (Stante et al., 1997). Fresh and stained cells (Gram, Neisser, Sudan black B and Nile blue A) were viewed by light and fluorescence microscopy, respectively, as described by Stante et al. (1997). All three strains are unpigmented and produce white colonies with internal refractile structures on GTA plates at 25 °C. Strains ATCC 11041^T and ATCC 43383 have distinctive, round colonies, whereas strain DSM 15336 produces round colonies with partly square borders. Strain ATCC 43383 produces slimy, glossy, transparent colonies, whereas strains ATCC 11041^T and DSM 15336 produce dry, matt, transparent colonies. Growth was usually obtained after 1–3 days at 25 °C. All three investigated strains have been reported to be Gram-negative, coccoid cells (1·0–1·5×1·0–2·5 μm), able to produce square tablets of eight to 64 cells (Murray, 1984; Stante et al., 1997). However, both ATCC strains have apparently lost this ability, despite strict cultivation under the conditions recommended by the initial discoverers. Strain ATCC 11041^T was already reported to have lost this ability in 1963 (Murray, 1963), whereas our study is the first to report that ATCC 43383 has also lost this ability. Strain DSM 15336 is still producing square tablets when cultivated on GTA medium, on plates or broth, at neutral pH and at temperatures up to 30 °C (Fig. 1⇓). However, strain DSM 15336 may also lose the tablet-forming ability and produce coccoid cells, in some cases even abnormally enlarged (up to 3·5 μm), under certain conditions, such as when grown on GTA plates or in broth at 37 °C. Interestingly, an identical observation was also reported for ATCC 11041^T when it was still able to produce tablet structures (Lany, 1972). A similar loss of the tablet morphology of strain DSM 15336 could also be induced by growth at 30 °C on either a modified GTA medium (plates or broth) in which acetate was replaced by sucrose (GTASuc) or on Luria–Bertani (LB) medium (plates or broth). No such effect was observed when the cells were grown on GTA plates or broth at pH 9. Hybridization of cells of strain DSM 15336 cultivated on GTASuc plates or on LB plates at 30 °C or on GTA plates at 37 °C with a fluorescent oligonucleotide probe that targeted Lampropedia hyalina specifically (probe LAMP444; described below) confirmed that all cells were Lampropedia hyalina. Furthermore, when cells of strain DSM 15336 were reinoculated onto GTA plates after cultivation on GTA plates at 37 °C or on GTASuc plates as well as on LB plates cultivated at 30 °C and cultivated at 30 °C, they once again produced square tablets, thus indicating clearly that the tablet-forming ability is dependent on the cultivation conditions.

The generation times for strains ATCC 11041^T, ATCC 43383 and DSM 15336 were 27·8, 22·6 and 13·3 h, respectively, based on the measurement of cell density by spectrophotometry (Thermo Spectronic Helyos α) at 620 nm on cells cultivated in triplicate on GTA broth at pH 7·2. Tolerance tests were performed on GTA plates in triplicate at different temperatures (10, 20, 30, 40 and 50 °C), pH values (3, 5, 6, 7, 8 and 9) and sodium chloride concentrations (1, 1·5 and 2 %). Bile tolerance was determined by addition of 0·5 % bile to GTA medium (plates). Sensitivity against kanamycin (50 mg l⁻¹) was investigated on GTA plates. Production of N₂ under anoxic conditions was examined as described by Gerhardt et al. (1994). Cultivation of cells under cyclic anaerobic–aerobic conditions for stimulation of polyphosphate accumulation was performed as described by Stante et al. (1997). Two other standard media, LB plates and R2A plates (Reasoner & Geldreich, 1985), were also used to test whether these media supported growth of Lampropedia. All tests were followed for up to 7 days. Tests for oxidase, catalase and litmus milk were performed as described by Gerhardt et al. (1994). Metabolic fingerprinting (Table 1⇓) was performed by using the miniaturized assay method described by Kämpfer et al. (1991). Hydrolysis of gelatin, starch and urea were examined as described by Gerhardt et al. (1994).

In summary, the three Lampropedia strains were strongly similar in terms of general phenotypic characteristics (Table 1⇑). The major differences observed among the three strains were in the ability to accumulate polyphosphate under anaerobic–aerobic cyclic incubation conditions (Stante et al., 1996, 1997), bile tolerance (albeit weakly), the ability to hydrolyse one compound (out of 16 tested), bis-p-nitrophenyl phosphate, and some minor differences in assimilation of different substrates. Furthermore, growth measurements on GTA broth (containing acetate, one of the three compounds that could be assimilated by all three strains; Table 1⇑) showed that the rate of utilization of acetate [analysed as acetic acid by GC (DANI 8510) as described by Stante et al., 1997] differed significantly between the three strains. DSM 15336 showed the highest rate of utilization (results not shown).

The composition of fatty acids was determined as described by Kämpfer et al. (2003) on cells grown on nutrient agar plates with 1 % (w/v) sodium acetate and on GTA plates at 28 °C. The three strains showed nearly identical fatty acid compositions. The main fatty acid methyl ester components detected were C_10 : 0 3-OH (5·8–6 %), C_12 : 0 3-OH (2–2·3 %), C_14 : 0 (9·1–10·7 %), C_16 : 0 (24·4–26·2 %) and C_18 : 1ω7c (24·2–25·9 %). The presence of C₁₀ 3-OH and C₁₂ 3-OH has also been reported for other members of the Comamonadaceae (Wen et al., 1999; Spring et al., 2004).

Analysis of the production by the Lampropedia strains of three different quorum-sensing N-acylhomoserine lactone molecules, PKRC12, PASC8 and PSB404, representative for some betaproteobacteria, was performed as described by Steidle et al. (2002). However, none of the strains produced any of the N-acylhomoserine lactone molecules investigated (Steidle et al., 2002).

Genomic DNA was extracted and purified for further molecular analyses, as described by Ziemke et al. (1998). For subsequent sequencing of the 16S rRNA gene, almost complete 16S rRNA gene fragments were amplified from DNA by using the primers 27F and 1492R (Lane, 1991). For 16S rRNA gene amplification, reaction mixtures (total volume, 50 μl) containing 20 ng DNA template and each primer at a concentration of 25 pM were prepared by adding 20 mM MgCl₂ reaction buffer (Idaho Technology) and 2·5 U Taq DNA polymerase (Promega). Thermal cycling was carried out by using an initial denaturation step of 94 °C for 1 min, followed by 30 cycles of denaturation at 94 °C for 40 s, annealing at 50 °C for 40 s and elongation at 72 °C for 90 s. Cycling was completed by a final elongation step of 72 °C for 10 min. Prior to sequencing, PCR products were purified by means of the QIAquick PCR purification kit (Qiagen). Nucleotide sequences were determined by the sequencing protocol from Epicenter Technologies (Biozyme). The new 16S rRNA gene sequences (>1400 nt, <0·5 % ambiguity) were added to an alignment of about 50 000 small-subunit rRNA sequences by using the alignment tool of the arb program package (Ludwig et al., 2004). Alignments were refined by visual inspection. Phylogenetic analyses were performed by using distance-matrix, parsimony and maximum-likelihood (axml in the arb software) methods. Various datasets comparing different selections of sequences and alignment positions were used (Ludwig et al., 1998). The three Lampropedia strains used in this study showed strong similarity (98·8–99·8 %) to each other at the 16S rRNA gene level. Further sequence comparison with three additional Lampropedia hyalina 16S rRNA gene sequences available on the NCBI public database (; accession numbers AB086632, AB092691 and AB089485), previously placed within the ‘Gammaproteobacteria’ (Pseudomonadaceae) (Bergey's taxonomic outline, July 2002) but recently moved to the ‘Betaproteobacteria’ (Bergey's taxonomic outline, October 2003; Xie et al., 2003), showed strong similarity to the strains in this study (e.g. 99·8 % 16S rRNA gene similarity to the type strain ATCC 11041^T). All Lampropedia hyalina strains form a distinct group among the ‘Betaproteobacteria’, showing 94·6–93 % 16S rRNA gene sequence similarity to members of various genera such as Acidovorax, Aquaspirillum, Brachymonas, Comamonas, Delftia and Xenophilus. The closest cultured relative is most likely Brachymonas denitrificans (94·6 % 16S rRNA gene sequence similarity), as suggested by distance matrix, parsimony and maximum-likelihood treeing methods (Fig. 2⇓).

Determination of the G+C content and DNA–DNA reassociation experiments were carried out as described by Ziemke et al. (1998). The G+C contents of strains ATCC 11041^T, ATCC 43383 and DSM 15336 were respectively 63·2 mol% [slightly higher in comparison with the preliminary result of around 60 mol% cited by Murray (1984)], 66·7 and 64·4 mol%. DNA–DNA reassociation studies showed a relatively higher reassociation between DSM 15336 and ATCC 11041^T (94·7 %, sd 0·6 %) than between ATCC 11041^T and ATCC 43383 (90·4 %, sd 0·6 %). These data indicate clearly that all three strains can be classified within the same species, Lampropedia hyalina.

For culture-independent identification of Lampropedia hyalina by means of fluorescent in situ hybridization, cells were fixed in exponential phase in 4 % (w/v) paraformaldehyde for 3 h at 4 °C, as described by Amann (1995). In situ hybridization of the samples with oligonucleotides labelled with the monofunctional, hydrophilic, sulfoindocyanine dye Cy3 (purchased from Interactiva) was performed as described by Amann (1995). Samples subjected to fluorescent in situ hybridization were mounted with Citifluor and viewed using a model LSM 510 scanning confocal microscope (Carl Zeiss) equipped with a UV laser (351 and 364 nm), an Ar ion laser (450 to 514 nm) and two HeNe lasers (543 and 633 nm). Different probes were used for identification of Lampropedia hyalina. For confirmation of the phylogenetic position among the ‘Betaproteobacteria’, two general probes targeting the betaproteobacteria on the 23S rRNA gene level (BET42a; Manz et al., 1992) and on the 16S rRNA gene level (BONE, together with the competitor for BTWO; Amann et al., 1996) were used. All three investigated Lampropedia hyalina strains yielded bright signals with these probes. Probes that target Lampropedia hyalina specifically were also developed. Probe design was performed on the basis of 50 000 complete and partial 16S rRNA gene sequences present in the arb database () and updated (January 2004) with all relevant existing sequences of Lampropedia hyalina and close relatives. Three specific probes were selected on the basis of the 16S rRNA gene sequence: LAMP114 (5′-CGTTCCGATGTCTTACTC-3′), LAMP444 (5′-CCCAGTCCTTTTCGTTCT-3′) and LAMP1017 (5′-TCTTTCGAGCACAATCCC-3′). Optimization of the hybridization conditions for these probes was performed by hybridizing at different formamide concentrations, ranging from 0 to 60 %, and determining the optimal signal intensity visually by microscopy (quantitative single-cell signal intensity measurement was not possible due to the tablet clusters of DSM 15336). The Lampropedia strains investigated in this study showed no autofluorescence and showed significant hybridization signals with probes LAMP114 and LAMP444 in the absence of formamide. No signals were obtained with probe LAMP1017 in the absence of formamide, which indicates that this target region is inaccessible. Different optimal hybridization conditions were observed for probes LAMP114 and LAMP444 for the three Lampropedia strains. Probe LAMP114 yielded strong hybridization signals at a formamide concentration up to 10 % with the two strains ATCC 11041^T and ATCC 43383 and up to 30 % with DSM 15336. Probe LAMP444 yielded strong hybridization signals at a formamide concentration up to 40 % with all three strains. Hybridization with probes LAMP114 and LAMP444 with members of two distantly related genera among the ‘Betaproteobacteria’, Comamonas testosteroni and Hydrogenophaga flava, in the absence of formamide yielded no signals.