Malikia granosa gen. nov., sp. nov., a novel polyhydroxyalkanoate- and polyphosphate-accumulating bacterium isolated from activated sludge, and reclassification of Pseudomonas spinosa as Malikia spinosa comb. nov.

Since 1993 several studies applying cultivation-independent techniques have demonstrated that members of the β-Proteobacteria, especially the family Comamonadaceae, may represent a major fraction of the metabolically active microbial population in activated sludge of municipal sewage-treatment plants (Wagner et al., 1993, 1994b; Manz et al., 1994; Wallner et al., 1995; Snaidr et al., 1997; Wagner & Amann, 1997; Wagner & Loy, 2002). Several filamentous β-Proteobacteria contribute to bulking problems of activated sludge (Wagner et al., 1994a) whereas other β-Proteobacteria catalyse, for example, aerobic ammonia oxidation (Purkhold et al., 2000; Juretschko et al., 2002), denitrification (Ginige et al., 2004) or the removal of phosphorous from sewage (Lee et al., 1999, 2003; Hesselmann et al., 1999; Blackall et al., 2002). However, cultivation-independent methods that are based solely on phylogenetic sequence information are not adequate to reveal an encompassing picture of the physiology of these micro-organisms in the environment. In particular, bacteria belonging to the family Comamonadaceae and neighbouring phylogenetic groups are phenotypically highly diverse, even if they are phylogenetically closely related, such that it is very difficult to correlate a distinct phylotype found in the environment with a certain metabolic potential of a related recognized species.

Two approaches have been developed to overcome these limitations in studying the physiological potential of distinct bacterial phylotypes in sewage sludge. Lee et al. (1999) and Ouverney & Fuhrman (1999) introduced a novel tool in microbial ecology by combining the oligonucleotide-based fluorescence in situ hybridization (FISH) of single cells with a high-resolution measurement of the microbial substrate uptake pattern using microautoradiography. A less laborious approach is based on screening of novel isolates retrieved from the studied environment with oligonucleotide probes targeting a specific group of micro-organisms. Pure cultures of the strains obtained can then be studied ex situ by applying standard microbial methods. An oligonucleotide-probe-assisted directed cultivation procedure has been successfully applied to samples of activated sludge and led to the characterization of several novel species (Schulze et al., 1999; Spring et al., 2004; Kämpfer et al., 2005).

The oligonucleotide probe LDI23a (Wagner et al., 1994a), specific for Leptothrix discophora and several related species within the Comamonadaceae, was used for the targeted isolation of novel β-Proteobacteria from activated sludge of a large municipal wastewater treatment plant (München I; Großlappen, Germany). Probe LDI23a was chosen because in the high-load aeration basin of this plant abundant bacteria hybridizing with this probe in situ were observed by FISH. Screening of isolates using this probe identified two novel strains. The complete isolation procedure and characterization of one of these strains have been described previously by Spring et al. (2004). A polyphasic characterization of the second strain, designated P1^T, is presented here.

Standard microbial methods, as reported previously by Spring et al. (2001, 2004), were used for the phenotypic characterization of the newly isolated strain P1^T and of Pseudomonas spinosa ATCC 14606^T, obtained from American Type Culture Collection (ATCC; Manassas, VA, USA). P. spinosa was chosen for a comparative characterization because it represents the phylogenetically most closely related species and the description provided by Leifson (1962) is rather sparse. Unless stated otherwise, R2A medium (Reasoner & Geldreich, 1985) was used for growth experiments. Cells were Gram-negative, straight to slightly curved rods with rounded ends (Fig. 1a⇓). They occurred singly or in short chains and were motile by means of one to two polar flagella. Filamentous cells up to 50 μm in length were frequently observed in nutrient-rich media. Under conditions of nitrogen limitation highly refractile intracellular granules could be observed by phase-contrast microscopy. After prolonged incubation in media that contained only very low concentrations of bound nitrogen, cells appeared slightly swollen and were almost completely filled with these granules (Fig. 1b⇓). Storage inclusions consisted of polyhydroxyalkanoates (PHAs), which could be demonstrated by staining with Nile Blue (Ostle & Holt, 1982). An epifluorescence micrograph demonstrating the typical bright red–orange fluorescence of the PHA granules after staining with Nile Blue is available as supplementary Fig. A in IJSEM Online.

Storage of polyphosphate (poly-P) as reserve material in cells of strain P1^T was shown by Neisser staining according to the protocol of Jenkins et al. (1993). Both kinds of reserve polymers could also be detected in the closely related strain P. spinosa ATCC 14606^T.

Colonies of strain P1^T appeared after 24 h on R2A agar. They reached about 2 mm in diameter after 2 days incubation. After several days colonies appeared mucoid as a result of the formation of extracellular slime. No diffusible pigments were formed.

The newly isolated strain grew well at temperatures ranging from 5 to 40 °C (temperatures below 5 °C were not tested). The pH range for growth in buffered R2A broth was 6·0–9·0. Adaptation of strain P1^T to freshwater habitats is reflected in its low salt tolerance. NaCl is tolerated in the medium only up to a concentration of 0·5 % (w/v). Under optimal growth conditions (35 °C, pH 7·0) the doubling time of P1^T was 1·2 h. The growth characteristics of P. spinosa ATCC 14606^T were similar to P1^T with only marginal differences; these are explicitly given in the formal species descriptions below.

The spectrum of usable carbon sources was determined in a microplate assay as described by Kämpfer et al. (1991) with the modification that the final substrate concentrations were 0·1 % (w/v). Strain P1^T was able to utilize a broad range of different substrates as sole carbon sources, including carbohydrates, sugar alcohols and organic acids. Formation of PHA granules was not restricted to a specific substrate, but acetate was particularly suitable to demonstrate its accumulation in this strain. In contrast, the substrate utilization pattern of P. spinosa ATCC 14606^T was more limited and restricted to a few carbohydrates and organic acids.

Chemolithoautotrophic growth of the novel strain P1^T and P. spinosa ATCC 14606^T with hydrogen as substrate was tested under the conditions described by Malik & Schlegel (1981) for the cultivation of various Knallgas bacteria. As a trait that distinguishes both strains from most species of the closely related genus Hydrogenophaga, no aerobic autotrophic growth with hydrogen as substrate was detected. Thiosulfate was not utilized as substrate and did not support autotrophic or chemolithoheterotrophic growth. In contrast, two species of the genus Hydrogenophaga, Hydrogenophaga palleronii and Hydrogenophaga intermedia, were shown to oxidize thiosulfate to sulfate (Kämpfer et al., 2005).

Strains P1^T and P. spinosa ATCC 14606^T were both oligonitrophilic and could grow in liquid mineral medium with 0·1 g yeast extract l⁻¹ as the only source of bound nitrogen. Fixing of molecular nitrogen, however, could not be demonstrated. No growth was observed under anaerobic conditions. Nitrate was reduced to nitrite, but denitrification did not occur. Neither strain was able to grow phototrophically.

Further details of the morphological and physiological characterization of strain P1^T and P. spinosa ATCC 14606^T are given in Table 1⇓ and in the formal species descriptions below.

Analysis of the respiratory lipoquinones of strains P1^T and P. spinosa ATCC 14606^T was by HPLC and electron-impact mass spectrometry according to the methods of Monciardini et al. (2003); ubiquinone-8 (Q-8) was the predominant component in both strains. This quinone system is a characteristic feature of the β-Proteobacteria (Yokota et al., 1992). The cellular fatty acid composition of both strains was analysed after growth on R2A agar and TSA plates (Difco) and compared with the fatty acid pattern of Macromonas bipunctata DSM 12705^T, the only available pure culture of the genus Macromonas (Table 2⇓). Fatty acid methyl esters were extracted and prepared by the standard protocol of the Microbial Identification System (MIDI, Microbial ID). Extracts were analysed by GLC as described by Kämpfer & Kroppenstedt (1996) and Kämpfer et al. (1997). The fatty acid profiles of strains P1^T and P. spinosa ATCC 14606^T were very similar but distinguishable. In general, the patterns of both strains were typical for members of the Comamonadaceae and were dominated by the unbranched fatty acids 16 : 1ω7c (54–71 %), 16 : 0 (15–25 %) and 18 : 1ω7c (6–12 %). The abundance of distinct fatty acids depended largely on the incubation conditions. For example, in both strains the hydroxylated fatty acid 8 : 0 3-OH was only detected after incubation on TSA plates. One possible reason for this effect could be the much higher protein content of TSA medium compared to R2A agar. Significant differences were found for the fatty acid profile of Macromonas bipunctata DSM 12705^T, which was dominated by unsaturated fatty acids 16 : 1ω7c, 17 : 1ω6c and 18 : 1ω7c, whereas 16 : 0 was present only in minor amounts.

A comparison of the fatty acid profiles obtained with the overall fatty acid composition of the genus Hydrogenophaga is shown in Table 2⇑. The large number of type strains belonging to this genus makes detailed comparison with related taxa consisting of only one or two strains difficult. Nevertheless, the occurrence of the cyclo propane fatty acid 17 : 0 cyclo among Hydrogenophaga species could be used to distinguish it from its most closely related taxa.

The 16S rRNA gene of strain P1^T was amplified and sequenced as reported by Spring et al. (2001). A continuous stretch of 1525 nt of the 16S rRNA gene sequence of strain P1^T was determined and used for a comparative phylogenetic analysis applying the arb software package, including a current release of the small-subunit rRNA database (Ludwig et al., 2004). An arb parsimony analysis placed strain P1^T within the family Comamonadaceae of the β-Proteobacteria. In various phylogenetic trees reconstructed by using neighbour-joining, maximum-likelihood or maximum-parsimony methods, the nearest relatives of strain P1^T were Hydrogenophaga species, Macromonas bipunctata and P. spinosa. In most of the reconstructed trees these bacteria formed a coherent lineage supported by bootstrap values of 80 % or greater. A representative phylogenetic tree including bootstrap values is shown in Fig. 2⇓. In this tree, which is based on only a limited number of sequences representing mainly type species of the Comamonadaceae, some members of the genus Hydrogenophaga form a common branch with strain P1^T, P. spinosa and Macromonas bipunctata. However, when extensive datasets were used for tree reconstruction that comprise all available, nearly complete 16S rRNA gene sequences affiliated to the Comamonadaceae, strain P1^T was located on a common lineage together with P. spinosa and Macromonas bipunctata, but separated from Hydrogenophaga species (this tree is available as supplementary Fig. B in IJSEM Online). The phylogenetically most closely related species to strain P1^T, based on 16S rRNA gene sequence similarity values, was P. spinosa (98·4 %). This species can be regarded as misclassified, because it is phylogenetically only distantly related to Pseudomonas aeruginosa, the type species of the genus. 16S rRNA gene sequence similarity values between P1^T and the other closely related taxa that form a common phylogenetic branch in Fig. 2⇓ ranged from 95·3 to 96·9 %.

Genomic DNA for the determination of DNA base composition and DNA–DNA hybridization experiments was isolated using a French pressure cell and purified by chromatography on hydroxyapatite as described by Cashion et al. (1977). The G+C content of the total DNA of strain P1^T and P. spinosa ATCC 14606^T was 67 and 66 mol%, respectively. DNA–DNA hybridization experiments between both strains were carried out in 2× SSC buffer containing 10 % DMSO at 70 °C according to the method of De Ley et al. (1970). DNA renaturation rates were measured using a Gilford System 2600 spectrophotometer (Gilford Instrument Laboratories) equipped with a Gilford 2527-R thermoprogrammer and plotter. The estimated level of DNA–DNA binding between strains P1^T and P. spinosa ATCC 14606^T was 53 %, indicating that both strains represent distinct species.

A peculiarity of the novel strain P1^T and P. spinosa ATCC 14606^T was the intracellular accumulation of PHAs and poly-P. Micro-organisms that are able to store these compounds in large amounts are thought to play a major role in the enhanced biological phosphorus removal (EBPR) process of specially designed wastewater treatment plants, which treat the sewage by alternating exposure to anaerobic and aerobic conditions (Blackall et al., 2002; Serafim et al., 2002). According to current models of this process (Seviour et al., 2003) extracellular phosphorus is removed from the wastewater by phosphate-accumulating organisms (PAOs) that take up phosphorus under aerobic conditions and store it as poly-P. Under anaerobic conditions the accumulated poly-P is degraded to provide energy for the PAOs to assimilate organic substances and to synthesize PHAs, which again serve under aerobic conditions (in the absence of extracellular organic compounds) as carbon and energy source. In addition, the reserve polymer glycogen is probably used as a source for reducing equivalents under anaerobic conditions. Hence, the cycling of reserve polymers provides a selective advantage of PAOs compared to other bacteria. It is assumed that the anaerobic–aerobic cycling of biomass during the EBPR process leads to selective enrichment of PAOs under the feast–famine regime in EBPR plants (Serafim et al., 2002). Use of intracellular polymers via the process described above is probably more efficient and faster as a complete metabolic switch from aerobiosis to anaerobiosis, as observed in many facultatively anaerobic, denitrifying bacteria isolated from wastewater.

The accumulation of poly-P and PHAs in strain P1^T was studied under various conditions in order to reveal the possible mechanisms that regulate synthesis of reserve polymers in this strain. The results were compared to the accumulation of reserve polymers in phylogenetically related bacteria. Although it is difficult to identify true PAOs involved in the EBPR process based solely on ex situ experiments, the results obtained should provide valuable taxonomic information regarding the internal metabolism in strain P1^T compared to related bacteria.

In these experiments a simple freshwater medium based on the SFW medium of Spring et al. (2004) was used. The modified SFW medium contained, per litre of distilled water: 0·8 g K₂HPO₄, 0·3 g KH₂PO₄, 0·4 g MgSO₄.7H₂O, 2·0 ml trace elements solution (Vishniac & Santer, 1957), 1·0 ml selenite/tungstate solution (Tschech & Pfennig, 1984), 0·1 g yeast extract (Difco) and 10·0 ml vitamins solution (Wolin et al., 1963), to a final pH of 7·2. Substrates and a source of bound nitrogen (NH₄Cl) were added after autoclaving at the concentrations indicated below. Anoxic media were prepared by applying the anaerobe culture technique of Hungate (1950) with the modifications introduced by Bryant (1972).

No growth of strain P1^T and P. spinosa ATCC 14606^T was observed on substrates in modified SFW medium lacking vitamins or yeast extract, indicating that both strains cannot proliferate without growth factors. To study the effect of nitrogen limitation on the accumulation of reserve polymers in strain P1^T, the modified SFW medium was supplemented with 2·0 g sodium acetate l⁻¹, but not with NH₄Cl. Hence, the only source of bound nitrogen in this medium was 0·1 g yeast extract l⁻¹. A loopful of P1^T cells grown on R2A agar plates was transferred to this medium and incubated for 24 h. Cells incubated under nitrogen limitation were viewed by phase-contrast microscopy to detect visible inclusions and were tested for poly-P accumulation with Neisser staining. If visible inclusions were observed, the presence of PHA was subsequently confirmed by staining with Nile Blue. In P1^T cells grown on R2A agar, generally only low to moderate amounts of poly-P and PHA could be detected by the respective staining procedures. In contrast, incubation under nitrogen limitation resulted in the accumulation of large amounts of PHA in the form of highly refractile granules, whereas poly-P was no longer detectable within the cells. A photomicrograph illustrating the absence of poly-P in nitrogen-starved cells by Neisser staining is available as supplementary Fig. C(a) in IJSEM Online. Interestingly, no accumulation of PHAs was observed in the same medium under anoxic or micro-oxic conditions (gas atmosphere containing 1 % oxygen).

In a complementary experiment, cells of strain P1^T were incubated in modified SFW medium containing 0·4 g NH₄Cl l⁻¹ as nitrogen source, but no acetate. After 24 h incubation, intracellular PHAs were almost completely degraded and large amounts of intracellular poly-P had accumulated. A photomicrograph illustrating the storage of poly-P in carbon-starved cells by Neisser staining is available as supplementary Fig. C(b) in IJSEM Online.

The above experiments imply that the accumulation of reserve compounds in this strain depends largely on the lack of nitrogen or carbon in the medium. However, to determine if this strain could represent a typical PAO, more laborious analyses would be necessary following incubations under alternate anaerobic/oxic conditions, for example as described by Stante et al. (1997).

Similar results were obtained with P. spinosa ATCC 14606^T. The only difference was that acetate could not be utilized for the production of PHAs. However, acetate could be replaced with the substrate maltose, which was also effective in inducing the production of PHAs under nitrogen limitation.

In similar experiments with all known type strains of the genus Hydrogenophaga, the same general conditions were effective among species of this genus in the induction of reserve polymers as in strains P1^T and P. spinosa ATCC 14606^T. However, the results were not always clear-cut; several strains showed only weak growth under conditions of nitrogen limitation and PHA accumulation depended strongly on the carbon source used. In most strains a significant accumulation of PHAs was only observed with 3-hydroxybutyrate as substrate.

In Macromonas bipunctata, although closely related with the above-mentioned species, the production of intracellular reserve compounds seemed to be regulated in a different way. In this species, nitrogen limitation obviously had no effect on the production of PHAs. In contrast, the accumulation of PHAs in Macromonas bipunctata DSM 12705^T was dependent upon the carbon source, with acetate being most effective (see also Dubinina & Grabovich, 1984). Under aerobic conditions storage of poly-P was almost constitutive instead of being induced by the deprivation of nutrients. In a preliminary experiment it was shown that a shift from aerobic to anaerobic incubation conditions led to reduced amounts of intracellular poly-P. Although these experiments were only performed with cultures grown on R2A agar plates and the amount of intracellular poly-P was only estimated by visual inspection after Neisser staining, the results obtained may indicate that poly-P could be utilized by Macromonas bipunctata DSM 12705^T under conditions of low oxygen concentrations. Regardless, the internal metabolism of reserve polymers in Macromonas bipunctata DSM 12705^T is clearly different from the metabolic type found in the closely related strains P1^T and P. spinosa ATCC 14606^T and hence represents a trait of taxonomic relevance.

From the results detailed above, it can be concluded that strain P1^T represents a novel taxon within the family Comamonadaceae, for which we propose the name Malikia granosa gen. nov., sp. nov. The proposal of a new genus seems justified based on the results of a polyphasic approach including biochemical, physiological and genotypic methods. It is further proposed to transfer the misclassified species Pseudomonas spinosa to Malikia gen. nov., as Malikia spinosa comb. nov. Characteristics that differentiate the new genus Malikia from the most closely related taxa of the Comamonadaceae are listed in Table 1⇑.