Genome Analysis of Fimbriiglobus ruber SP5^T, a Planctomycete with Confirmed Chitinolytic Capability

ABSTRACT

Members of the bacterial order Planctomycetales have often been observed in associations with Crustacea. The ability to degrade chitin, however, has never been reported for any of the cultured planctomycetes although utilization of N-acetylglucosamine (GlcNAc) as a sole carbon and nitrogen source is well recognized for these bacteria. Here, we demonstrate the chitinolytic capability of a member of the family Gemmataceae, Fimbriiglobus ruber SP5^T, which was isolated from a peat bog. As revealed by metatranscriptomic analysis of chitin-amended peat, the pool of 16S rRNA reads from F. ruber increased in response to chitin availability. Strain SP5^T displayed only weak growth on amorphous chitin as a sole source of carbon but grew well with chitin as a source of nitrogen. The genome of F. ruber SP5^T is 12.364 Mb in size and is the largest among all currently determined planctomycete genomes. It encodes several enzymes putatively involved in chitin degradation, including two chitinases affiliated with the glycoside hydrolase (GH) family GH18, GH20 family β-N-acetylglucosaminidase, and the complete set of enzymes required for utilization of GlcNAc. The gene encoding one of the predicted chitinases was expressed in Escherichia coli, and the endochitinase activity of the recombinant enzyme was confirmed. The genome also contains genes required for the assembly of type IV pili, which may be used to adhere to chitin and possibly other biopolymers. The ability to use chitin as a source of nitrogen is of special importance for planctomycetes that inhabit N-depleted ombrotrophic wetlands.

IMPORTANCE Planctomycetes represent an important part of the microbial community in Sphagnum-dominated peatlands, but their potential functions in these ecosystems remain poorly understood. This study reports the presence of chitinolytic potential in one of the recently described peat-inhabiting members of the family Gemmataceae, Fimbriiglobus ruber SP5^T. This planctomycete uses chitin, a major constituent of fungal cell walls and exoskeletons of peat-inhabiting arthropods, as a source of nitrogen in N-depleted ombrotrophic Sphagnum-dominated peatlands. This study reports the chitin-degrading capability of representatives of the order Planctomycetales.

INTRODUCTION

Members of the bacterial phylum Planctomycetes are widely distributed in various aquatic and terrestrial habitats (1, 2). Three distinct orders of planctomycetes are currently recognized, namely, the Planctomycetales, Phycisphaerales, and “Candidatus Brocadiales.” Of these, a comprehensive knowledge of the metabolic potential and functional roles in the environment is available only for chemo-lithoautotrophic anammox planctomycetes of the order “Candidatus Brocadiales” (3, 4). Few currently cultured representatives of the Phycisphaerales are chemo-organotrophs, which are capable of degrading various heteropolysaccharides (5). The order Planctomycetales also accommodates chemo-organotrophic planctomycetes and includes the largest number of cultured representatives, which belong to 17 genera with validly published names. Despite the fact that this is the first described order within the phylum Planctomycetes (6), we know very little about potential functions of its members in the environment. Some of these bacteria, including Blastopirellula and Rhodopirellula-like planctomycetes, are found mostly in marine environments and represent an important part of the complex microbial biofilm community of a wide range of macroalgae (7). These planctomycetes specialize in degradation of sulfated polysaccharides produced by algae, and their genomes contain an exceptionally high number of sulfatase genes (8). Members of the family Isosphaeraceae are especially abundant in boreal peatlands (9, 10). As revealed by recent comparative genomic analysis, these peat-inhabiting planctomycetes possess extremely high but partly hidden glycolytic potential (11) and are capable of degrading various heteropolysaccharides. The ability to degrade fibrous and microcrystalline cellulose, however, has so far been demonstrated only for a single, peat-inhabiting planctomycete from the family Gemmataceae, Telmatocola sphagniphila (12). Except for participation in degradation of plant-derived polymers, planctomycetes were identified as efficient degraders of exopolysaccharides produced by other soil bacteria (13).

The ability to degrade chitin, one of the most abundant polymers in nature, has never been reported for any of the cultured planctomycetes. Given that utilization of N-acetylglucosamine as a sole carbon and nitrogen source is well recognized for these bacteria, the absence of chitinolytic capabilities looks somewhat illogical. At the same time, cells of planctomycetes have often been observed in associations with Crustacea (14–16). One of the recently described planctomycetes, “Fuerstia marisgermanicae” NH11^T, was isolated from a crab shell collected on the tidal mud flat of the German Wadden Sea (17). None of these planctomycetes was shown to degrade chitin. These cultured representatives, however, do not cover all planctomycete diversity in natural habitats. Recent molecular study of microbial degradation of chitin in an agricultural soil suggested involvement of Singulisphaera-like planctomycetes in this process (18). Additional evidence for the existence of chitinolytic planctomycetes was obtained in a metatranscriptome-based study of microbial populations driving biopolymer degradation in acidic peatlands (19). Although no changes in the relative abundance of planctomycetes within the whole bacterial community were observed in response to the amendment with chitin, some groups within uncultivated planctomycetes, in particular Gemmata-like bacteria, were clearly stimulated by the availability of this biopolymer.

Recently, we described one novel member of the family Gemmataceae from acidic peat bogs, namely, Fimbriiglobus ruber SP5^T (20). This planctomycete was isolated from the peat bog Obukhovskoye, which served as a source of peat material used in the metatranscriptomic study of Ivanova et al. (19). Although strain SP5^T was unable to degrade chitin from crab shells, weak growth was observed on amorphous chitin prepared as described elsewhere (21). Since unambiguous detection of chitin degradation capabilities in this peat-inhabiting planctomycete was complicated by its low growth rates and high cell aggregation in laboratory medium, this study was initiated in order to determine the presence of genomic determinants for chitin degradation in this bacterium.

RESULTS

Metatranscriptome-derived evidence for the presence of chitinolytic capabilities in F. ruber.

This study began with attempts to detect the substrate-induced response of F. ruber-like planctomycetes to amendment of Sphagnum-derived peat from the peat bog Obukhovskoye with chitin. For this purpose, we reanalyzed the small subunit (SSU) rRNA data set retrieved in our recent metatranscriptomic study (19), which assessed the substrate-induced response of peat-inhabiting microorganisms to amendments with various key biopolymers, including cellulose, xylan, pectin, and chitin. The 16S rRNA reads from F. ruber-like planctomycetes were sorted out using a species-level identity threshold of 97% from the sequence sets corresponding to four experimental incubations with different biopolymers and the control incubation without added substrate. As revealed by this analysis, the pool of 16S rRNA reads from F. ruber-like planctomycetes significantly increased 2.5-fold (P value of <0.01) in response to chitin availability (Fig. 1). Notably, no response was detected in incubations with pectin and xylan, while some statistically insignificant stimulation of F. ruber was also observed due to amendment of peat with cellulose. The chitinolytic potential of F. ruber, suggested by the results of metatranscriptome-based study, was further examined by the focused growth experiments with the type strain of this species, SP5^T.

FIG 1.

The relative abundance values represent average values calculated by relating the number of reads assigned to the species Fimbriiglobus ruber (blast sequence identity threshold of 97%) to the total number of SSU rRNA reads retrieved from four experimental incubations of peat samples amended with different biopolymers and from the control incubation without added substrate in the study of Ivanova et al. (19). The relative abundance values represent averages of triplicate data sets (pectin, cellulose, and xylan) or duplicate data sets (chitin and control). A significant difference in Ribo-tag abundances between control and biopolymer-amended samples is indicated by an asterisk (P < 0.01).

Growth on chitin and chitinolytic activities.

Two types of substrate utilization tests were performed with F. ruber SP5^T. One of them corresponded to the routinely used assay in which chitin is provided as a sole source of carbon, while (NH₄)₂SO₄ is supplied as a source of nitrogen. In the second type of growth experiments, chitin was added as the only source of nitrogen, while glucose served as the source of carbon. Microscopic examination of the respective cultures after 1 month of incubation revealed cells of strain SP5^T being attached to microparticles of amorphous chitin (Fig. 2); free-floating cells were absent. Only a very few rarely scattered cells of strain SP5^T were observed in cultures with amorphous chitin as a sole source of carbon (Fig. 2, frames 1a to 1c). In contrast, all microparticles of chitin were densely colonized by planctomycete cells in cultures containing chitin as a source of nitrogen (Fig. 2, frames 2a to 2c). Cell counting revealed the drastic difference by two orders of magnitude in cell abundances in these two types of incubations (Table 1), suggesting that the activity of chitin degradation is largely dependent on the presence of an easily available carbon source.

FIG 2.

Specific detection of planctomycete cells on microparticles of amorphous chitin used in growth experiments as a source of carbon (row 1) or nitrogen (row 2). Phase-contrast images (a frames), the respective epifluorescent micrographs of whole-cell hybridizations with Cy3-labeled probes PLA46-PLA886 (b frames), and DAPI staining (c frames) are shown. Bar, 10 μm.

TABLE 1.

Cell numbers of Fimbriiglobus ruber SP5^T after 30 days of incubation with chitin as a source of carbon or nitrogen

Incubation no.	Growth of F. ruber by condition (no. of cells/ml [×108])a
	Chitin as a source of C		Chitin as a source of N
	Chitin + NH4+	Control (+NH4+)b	Chitin + glucose	Control (+glucose)b
1	2.2 ± 0.6	1.1 ± 0.2	150 ± 1.6	22 ± 5
2	4.1 ± 2.3	0.9 ± 0.1	248 ± 0.3	13 ± 7
3	2.1 ± 0.9	1.8 ± 0.4	180 ± 0.2	13 ± 5

^aThe results of three independent incubations (1 to 3) are shown. Values are the average ± standard error.

^bControl experiments did not include chitin.

Chitinolytic activities of F. ruber SP5^T with synthetic soluble substrates were detected in cell extracts. Activity was maximal with the β-N-acetylglucosaminidase substrate 4-methylumbelliferyl (MU)-N-acetyl-β-d-glucosaminide (31.5 U/mg), while about 24-fold-lower activity was observed with 4-MU-diacetyl-β-d-chitobioside (1.3 U/mg), and about 40-fold-lower activity was recorded with the endochitinase substrate 4-MU-β-d-N,N′,N″-triacetylchitotriose (0.8 U/mg). Further insight into the metabolic potential of F. ruber SP5^T was made via the genome analysis.

Genome characteristics.

The genome of F. ruber SP5^T was sequenced using a combination of pyrosequencing and PacBio single-molecule real-time sequencing technique. Combined assembly of the sequencing reads obtained by two techniques yielded 20 contigs with N₅₀ contig size of 1,426,550 bp. The size of the F. ruber SP5^T genome, estimated as a total length of all contigs, is 12,363,577 bp, with a GC content of 64.2%. To the best of our knowledge, this is the largest genome of planctomycetes sequenced to date. Annotation of the genome sequence revealed 10,640 potential protein-coding genes, of which 4,693 (44%) can be functionally assigned; 2,636 genes are unique to F. ruber SP5^T with no significant similarity to any known sequences.

Metabolism and cell biology.

KEGG-based annotation of the F. ruber SP5^T genome sequence classified 2,086 proteins into 23 major functional categories. The genes encoding metabolic pathways common for chemo-organotrophic bacteria, such as glycolysis, the citrate cycle, the pentose-phosphate pathway, and oxidative phosphorylation, were present in the genome of F. ruber SP5^T. This planctomycete has the genomic potential for synthesis of all amino acids. The number of ABC-transporters in F. ruber SP5^T is 52, which is comparable to the calculated mean of 49 ABC-transporters in free-living prokaryotes (8). Also, two fructose-type sugar-specific subunits of the phosphotransferase system could be found in strain SP5^T.

The survey for genes related to cell division revealed that the FtsZ-encoding gene was absent, while two copies of the gene coding for FtsK, the DNA translocase, were present in the genome of F. ruber SP5^T. The cytoskeletal protein MreB, whose gene has a patchy presence among the planctomycetes and is absent in Gemmataceae planctomycetes (22), is present in F. ruber SP5^T. Several, but not all, genes involved in peptidoglycan biosynthesis, including murA, murB, murE, and murF, were detected (23).

Recently, the gene encoding the key enzyme for synthesis of N-methylated phosphorus-free ornithine membrane lipids, N-methyltransferase (OlsG), was identified in the genome of Singulisphaera acidiphila DSM 18658^T (Sinac_1600) (24) and later detected in genomes of other members of the family Isosphaeraceae (11). Our analysis revealed that distant homologs of Sinac_1600 are also present in the genomes of Gemmataceae planctomycetes, including F. ruber SP5^T, Zavarzinella formosa A10^T, and Gemmata sp. strain SH-PL17.

The examination of the F. ruber SP5^T genome for the presence of giant genes (25) revealed 69 genes with a size of >5 kb, among which 10 genes exceed 10 kb and 1 gene exceeds 25 kb. The functions of these giant genes remain enigmatic. So far, the highest number (60) of giant genes with a size of >5 kb was recorded for another member of the family Gemmataceae, Zavarzinella formosa A10^T (17).

Genetic background of chitinolytic capability.

Analysis of the F. ruber SP5^T genome revealed several enzymes that could potentially be involved in chitin utilization. The extracellular hydrolysis of chitin could be performed by FRUB_0131 endochitinase. This glycoside hydrolase (GH) family 18 (GH18) enzyme carries an N-terminal Sec secretion signal, suggesting its extracellular operation. GenBank searches revealed similar proteins (37 to 46% amino acid sequence identity) in only three planctomycetes, including Planctomicrobium piriforme DSM 26348^T, Planctomyces sp. strain SH-PL14 and Phycisphaerae bacterium ST-NAGAB-D1. Planctomicrobium piriforme was described as being incapable of growth on chitin (26), while the two other planctomycetes have not been characterized phenotypically.

For functional analysis of the predicted chitinase FRUB_0131, the corresponding gene coding for mature protein lacking an N-terminal signal peptide was expressed in E. coli. The recombinant enzyme was purified to homogeneity through Ni-nitrilotriacetic acid (NTA) affinity chromatography. Chitinolytic activities of recombinant FRUB_0131 were evaluated with synthetic soluble substrates. The maximal activity (4.64 U/mg) was observed with the endochitinase substrate 4-MU-β-d-N,N′,N″-triacetylchitotriose, while lower values were detected with 4-MU-diacetyl-β-d-chitobioside (0.55 U/mg) and 4-MU-N-acetyl-β-d-glucosaminide (0.16 U/mg). This substrate specificity pattern is consistent with the predicted endochitinase activity of FRUB_0131.

The search for other chitin-degrading enzymes from GH families 18, 19, and 20, carrying an N-terminal signal peptide, revealed only GH20 family protein FRUB_4986. This enzyme, annotated as β-hexosaminidase, could cleave N-acetyl-d-glucosamine (GlcNAc) from the nonreducing end of the soluble chitin oligomers (27), produced by FRUB_0131 endochitinase. A blastp search against the GenBank database revealed similar enzymes in some members of the Verrucomicrobia (up to 44% sequence identity) and only more distant homologs in the genomes of various planctomycetes. FRUB_4986 hydrolase could be responsible for the high level of acetylglucosaminidase activity described above. Concerted activity of endochitinase FRUB_0131 and putative β-N-acetylglucosaminidase FRUB_4986 could result in hydrolysis of chitin to oligomers and subsequent production of GlcNAc monomers.

One more gene, FRUB_5329, encodes a 758-amino-acid (aa) protein with a calculated molecular mass of 78.6 kDa. It comprises two copies of a carbohydrate-binding CBM2 domain and a single ChiA-like (cd06543) catalytic domain of GH18 family hydrolases at the C terminus. CBM2 typically binds cellulose but several of these modules have been shown to also bind chitin (28). Although SignalP, version 4.1, did not convincingly reveal an N-terminal secretion signal, the presence of carbohydrate-binding domains suggests that this enzyme could be also extracellular. The blastp search against GenBank revealed no proteins with the same domain architecture although hydrolases with similar catalytic domains (amino acid sequence identity up to 52%) were found among Actinobacteria. Notably, homologous enzymes are missing in other planctomycetes. Therefore, acquisition of this enzyme in F. ruber SP5^T by lateral transfer seems to be plausible.

GlcNAc may be imported into the cytoplasm by ABC-type transporters (29). In the cytoplasm, GlcNAc may be phosphorylated to yield GlcNAc-6-P by N-acetylglucosamine kinase (FRUB_5622). Then, GlcNAc-6-P could be deacetylated by N-acetylglucosamine-6-phosphate deacetylase (FRUB_4622 and FRUB_1803) yielding glucosamine 6-phosphate. The latter is converted via the action of glucosamine-6-phosphate deaminase (FRUB_0022) into fructose-6-P that enters central glycolytic pathway (30) (Fig. 3). GenBank searches revealed enzymes similar to GlcNAc kinase, GlcNAc-6-P deacetylases, and GlcN-6-P deaminase in members of the order Planctomycetales.

FIG 3.

Suggested scheme of chitin degradation in F. ruber SP5^T. Enzymes are given with gene numbers corresponding to those from the genome of F. ruber SP5^T when possible. Abbreviations: OM, outer membrane; PP, periplasm; CP, cytoplasm; GH18, GH18 family chitinase; GH20, GH20 family β-N-acetylglucosaminidase; ABC, ABC-type sugar transporter.

While analyzing the genome, we did not find enzymes responsible for intracellular metabolism of GlcNAc dimers in chitinolytic bacteria such as Chitinispirillum alkaliphilum (21), e.g., N,N′-diacetylchitobiose phosphorylase that can hydrolyze chitobiose and generate GlcNAc and GlcNAc-1-P and acetylglucosamine-1-P-mutase making GlcNAc-6-P from GlcNAc-1-P. The lack of these components of the chitinolytic pathway seems to limit the ability of F. ruber SP5^T to grow on chitin since only N-acetylglucosamine could be metabolized. Taking into account the lack of a N₂ fixation pathway in the genome of F. ruber SP5^T, one could suggest that chitin may primarily serve as a source of ammonium generated by deamination of glucosamine-6-phosphate.

Degradation of other polysaccharides.

In the original description, F. ruber was reported to be capable of growth on several polysaccharides, such as xylan, laminarin, lichenan, starch, and xanthan, while cellulose and pectin were not utilized (20). These metabolic capabilities of strain SP5^T were reevaluated by measuring the hydrolysis of the respective polysaccharides by lysed cell extracts. Carboxymethyl cellulose, xanthan, laminarin, lichenan, and xylan were efficiently hydrolyzed, while activities toward starch and microcrystalline cellulose were not observed (Fig. 4A).

FIG 4.

Hydrolytic activities of F. ruber SP5^T crude cell extract. (A) Hydrolysis of different polysaccharides evaluated by measuring the amounts of total reduced sugars released after treatment. The following substrates were tested: lane 1, microcrystalline cellulose; lane 2, starch; lane 3, carboxymethyl cellulose; lane 4, lichenan; lane 5, laminarin; lane 6, xanthan; lane 7, birchwood xylan; lane 8, beechwood xylan. (B) Hydrolysis of aryl glycosides. The following substrates were tested: lane 1, pNPGal; lane 2, pNPXyl; lane 3, pNPMan; lane 4, pNPGlu. Data represent the means of three separate experiments. The activities of the cell extract with carboxymethyl cellulose (A) and pNPGlu (B) were defined as 100%.

Hydrolytic activities in the cell extracts were also assayed using p-nitrophenyl (pNP) glycosides. The highest activity was observed with p-nitrophenyl-β-d-glucopyranoside (β-glucosidase substrate), followed by nitrophenyl-β-d-galactopyranoside (β-galactosidase substrate), p-nitrophenyl-β-d-xylopyranoside (β-xylosidase substrate), and pNP-β-d-mannopyranoside (β-mannosidase substrate), as shown in Fig. 4B.

Analysis of the F. ruber SP5^T genome revealed 11 genes encoding glycoside hydrolases predicted to have N-terminal signal peptides and thus possibly involved in extracellular hydrolysis of polysaccharides. Despite the reported inability of F. ruber to utilize cellulose, its genome encodes two signal peptide-containing GH5 family glycosyl hydrolases, annotated as endo-1,4-β-glucanases (FRUB_3928 and FRUB_8601). These enzymes could be responsible for extracellular hydrolysis of cellulose, while intracellular hydrolysis of produced cellobiose could be performed by GH94 family cellobiose phosphorylase (FRUB_8916). In contrast, no apparent endo-1,4-β-xylanases were identified although hydrolytic activity toward xylan was observed, and strain SP5^T was able to grow on this substrate. It is possible that GH5 enzymes could be responsible for hydrolysis of xylan since a number of activities were reported for this family, including endoglucanase, endomannanase, β-glucosidase, β-mannosidase, and xylanase (31). The ability to utilize xylose is consistent with the presence of the isomerase pathway of xylose metabolism, including xylose isomerase (FRUB_2606) and xylulose kinase (FRUB_3023).

Several other signal peptide-containing glycosyl hydrolases could be responsible for extracellular hydrolysis of polysaccharides. The GH99 family enzyme FRUB_7797 could exhibit endo-α-mannosidase and/or endo-α-1,2-mannanase activity, while β-galactosidase, β-glucuronidase, β-mannosidase, and exo-β-glucosaminidase activities were described for GH2 hydrolases (FRUB_5004 and FRUB_5438). Alpha-N-arabinofuranosidase FRUB_9125 could catalyze the hydrolysis of nonreducing terminal alpha-l-arabinofuranosidic linkages in arabinoxylans and arabinogalactans. Specificity of several other extracellular glycosyl hydrolases could not be reliably predicted (FRUB_0231, FRUB_0245, FRUB_2479, FRUB_3476, and FRUB_8423).

Despite the reported ability of F. ruber to utilize starch, genome analysis revealed no candidate enzymes for extracellular hydrolysis of this polysaccharide. Consistently, no hydrolytic activity toward starch was detected in cell extracts (Fig. 4A). Although the GH15 family glucoamylase (FRUB_8812) and two GH77 family amylomaltases (FRUB_5679 and FRUB_6807) are encoded, they lack a signal peptide and most likely are involved in intracellular pathways of the synthesis and degradation of storage polysaccharides.

Type IV pili.

In agreement with the morphology of F. ruber SP5^T cells, which are covered by numerous fimbriae, a complete set of genes for type IV pilus-based twitching motility is present in the genome, including pilins (Flp), assembly ATPase (PilB/TadA), inner membrane core proteins (PilC/TadB and TadC), prepilin peptidase (PilD/TadV), outer membrane secretin (PilQ/RcpA), retraction ATPase (PilT), inner membrane proteins (PilM, PilN, and PilO), and regulatory proteins. These pili may be used by F. ruber to adhere and move around chitin and possibly other polysaccharide substrates similar to the way some cellulolytic bacteria adhere to cellulose (32, 33) and Vibrio parahaemolyticus adheres to chitin (34). However, no close homologs of GlcNAc binding protein A (GbpA), which has been reported to mediate bacterial attachment to both chitin and mammalian intestinal mucin in Vibrio sp. (35), were found.

Secondary metabolite-related genes.

Genome mining of gene clusters that encode biosynthetic pathways for secondary metabolites in F. ruber SP5^T revealed 10 clusters (see Fig. S1 in the supplemental material). Five clusters encode polyketide synthases (PKS) of type I and others. The closest gene homologs to PKS clusters from F. ruber SP5^T belong to the planctomycetes Singulisphaera acidiphila DSM 18658^T, Isosphaera pallida ATCC 43644^T, Gemmata obscuriglobus UQM 2246^T, and the proteobacterium Pelobacter propionicus DSM 2379. The other five clusters encode terpene-like compounds. Their closest homologs also belong to several planctomycetes such as S. acidiphila DSM 18658^T and G. obscuriglobus UQM 2246^T and an uncharacterized member of the phylum Bacteroidetes (Fig. S1).

DISCUSSION

While the ability to utilize N-acetylglucosamine as a carbon and nitrogen source is widely distributed in members of the order Planctomycetales, the ability to degrade chitin, a natural source of this substrate, is described only for F. ruber SP5^T. These observations are consistent with the distribution of homologs of F. ruber SP5^T genes relevant to chitin metabolism in various bacterial taxa. Close homologs of enzymes required for utilization of GlcNAc (GlcNAc kinase, GlcNAc-6-P deacetylase, and GlcN-6-P deaminase) in F. ruber SP5^T are encoded in the genomes of many planctomycetes, as could be expected. In contrast, enzymes similar to GH18 family chitinases (FRUB_0131 and FRUB_5329) and GH20 family β-N-acetylglucosaminidase (FRUB_4986) are either absent in other planctomycetes or present in only a few species (see Fig. S2A to C in the supplemental material) and seem to be laterally acquired in F. ruber SP5^T. Interestingly, these three genes are located in different genome regions and, most likely, were obtained in independent transfer events. Each of these transfers is expected to enhance the chitinolytic capabilities of the recipient, facilitating extracellular hydrolysis of chitin with production of N-acetylglucosamine that enters the already existing metabolic pathway.

The ability to use chitin as a source of nitrogen is of special importance for planctomycetes that inhabit Sphagnum peat bogs. These ecosystems receive all nutrient inputs solely from the atmosphere and are, therefore, nutrient poor by nature (36, 37). Peat water usually contains very low concentrations of NH₄⁺ and NO₃⁻ (3 to 100 μM) (38–40). As a consequence, many bacteria that thrive in peat bogs are capable of N₂ fixation (41). Given the lack of dinitrogen fixation capabilities in members of the order Planctomycetales, they should either stay in a close association with N₂-fixing bacteria or rely on degradation of N-containing biopolymers. Apparently, F. ruber uses the second strategy and obtains required nitrogen from chitin, a major constituent of fungal cell walls and exoskeletons of peat-inhabiting arthropods. It may well be that many planctomycetes from Sphagnum peat bogs or other nitrogen-poor habitats possess similar metabolic capabilities. This possibility, however, cannot be verified at present since only a few genomes from peat-inhabiting planctomycetes are currently available. One suitable target for further growth experiments and enzymatic analyses is Planctomicrobium piriforme, which was isolated from a littoral wetland of a boreal lake (26). The original description of this planctomycete reports its inability to degrade chitin. This characteristic, however, was assessed in a routine test by providing chitin as a source of carbon and needs to be reevaluated using several alternative experimental settings.

As revealed in our study, the conventional growth tests for the presence of chitinolytic capabilities in bacteria may not be fully suitable for planctomycetes. In the case of F. ruber SP5^T, growth on chitin as a sole source of carbon was nearly nondetectable due to the relatively low endochitinase activity. However, growth with this biopolymer was clearly enhanced in the presence of an easily utilizable carbon substrate. It is tempting to speculate that, in habitats depleted of nitrogen but rich in organic carbon, such as peat bogs, chitin is used by planctomycetes primarily as a source of nitrogen. As becomes apparent from the recent studies, planctomycetes are more involved in the processes of biopolymer degradation in various habitats than previously thought. They also seem to possess an unusual mechanism of high-molecular-weight polysaccharide uptake, being capable of taking up and accumulating macromolecules in the periplasmic space (42, 43). This mechanism gives a clear ecological advantage to slow-growing planctomycetes by securing substantial quantities of substrate in the cell (43). Whether this adaptation feature is common for peat-inhabiting planctomycetes remains to be verified.

MATERIALS AND METHODS

Analysis of metatranscriptome-derived data.

The SSU rRNA data set retrieved in the study of Ivanova et al. (19) was used to analyze the abundance of F. ruber-like planctomycetes in Sphagnum-derived peat and the substrate-induced response of these bacteria to amendments with different biopolymers. The 16S rRNA gene sequence of F. ruber SP5^T was used as a query for blastn search via blast+ (44) among the SSU rRNA reads with an identity threshold of 97%. One-way analysis of variance (ANOVA) followed by Dunnett's multiple-comparison test (i.e., comparison of the results obtained for biopolymer-amended samples with those of the control samples) was performed using GraphPad Prism, version 6.00, for Windows (GraphPad Software, La Jolla, CA, USA). Tests were considered significant if they had a P value of <0.05.

Growth experiments and whole-cell hybridization.

F. ruber SP5^T (also LMG 29572^T or VKM B-3045^T) was grown under static conditions at 25°C in 160-ml serum bottles containing 10 ml of liquid medium M1 containing the following (per liter of distilled water): 0.1 g of KH₂PO₄, 0.1 g of MgSO₄ × 7H₂O, 0.02 g of CaCl₂ × 2H₂O, 1 ml of trace element solution “44,” and 1 ml of Staley's vitamin solution (45), pH 4.8 to 5.5. Amorphous chitin was prepared as described elsewhere (21) and added to the medium at a concentration of 0.1% (wt/vol). In one set of experimental flasks, chitin was provided as a sole source of carbon, while (NH₄)₂SO₄ (0.01%, wt/vol) was supplied as a source of nitrogen. In the second set of incubation flasks, chitin was added as the only source of nitrogen, while glucose (0.5 g liter⁻¹) served as the source of carbon. Control incubations without chitin (Table 1) were run in parallel under the same conditions. All incubations were performed in triplicate. After 30 days of incubation, the culture suspensions were fixed with 4% (wt/vol) freshly prepared paraformaldehyde solution as described by Dedysh et al. (47). A combination of two Planctomycetes-specific Cy3-labeled oligonucleotide probes, PLA46 (5′-GACTTGCATGCCTAATCC-3′) and PLA886 (5′-GCCTTGCGACCATACTCCC-3′) (46), was applied for specific detection of cells on microparticles of chitin. The oligonucleotide probes were purchased from Syntol (Moscow, Russia). Hybridization was done on gelatin-coated (0.1%, wt/vol) and dried Teflon-laminated slides (MAGV, Germany) with eight wells for independent positioning of the samples. The fixed samples were applied to these wells, hybridized to the corresponding fluorescent probes, and stained with the universal DNA stain 4′,6-diamidino-2-phenylindole (DAPI; 1 μM) as described earlier (47). The cell counts were carried out with a Zeiss Axioplan 2 microscope (Zeiss, Jena, Germany) equipped with Zeiss filters 20 and 02 for Cy3-labeled probes and DAPI staining, respectively. Cell counting was performed on 100 randomly chosen fields of view (FOV) for each test sample. The number of target cells per milliliter of culture suspension was determined from the area of the sample spot, the FOV area, and the volume of the fixed aliquot used for hybridization.

For genome analysis, F. ruber SP5^T was grown in shaking liquid cultures at 25°C in medium M31 containing the following (per liter of distilled water): 1.0 g of N-acetylglucosamine, 0.5 g of glucose, 0.1 g of peptone, 0.1 g of yeast extract, 0.1 g of KH₂PO₄, and 20 ml Hunter's basal salts, pH 5.5 to 5.8. After 3 weeks of incubation, the biomass was collected and used for DNA extraction using SDS-cetyltrimethylammonium bromide (CTAB) (48).

Genome sequencing, annotation, and analysis.

Sequencing of a shotgun genome library on a Roche GS FLX genome sequencer using the titanium protocol (Roche, Switzerland) resulted in the generation of ∼185 Mb of sequences with an average read length of 554 bp. A 10-kb genomic DNA library was prepared and sequenced on a PacBio RSII sequencer. Sequences totaling 698 Mb were obtained with an average read length of 3,655 bp. GS FLX reads were assembled using GS De Novo Assembler, version 3.0, and then a hybrid assembly of raw PacBio subreads and GS FLX large contigs was done using Canu (49). To test Canu results for misassemblies, we also generated a hybrid assembly using DBG2OLC (50). The two assemblies were compared and merged.

Gene search and annotation were performed using the RAST server (51), followed by manual correction. Signal peptides were predicted using Signal P, version 4.1, for Gram-negative bacteria (DTU Bioinformatics, Lyngby, Denmark [http://www.cbs.dtu.dk/services/SignalP/]).

The analysis of F. ruber SP5^T genome sequence using the Kyoto Encyclopedia of Genes and Genomes (KEGG) was performed by applying the GhostKOALA tool (52). Screening for secondary metabolite-related genes was performed using the online web server antiSMASH3.0.5 (Antibiotics and Secondary Metabolites Analysis SHell) (53–55).

Hydrolytic activity assays.

Chitinolytic activities of F. ruber SP5^T cell extracts were measured in a fluorimetric assay with 4-methylumbelliferyl (4-MU) derivatives using a chitinase assay kit (CS1030; Sigma). Cells were collected by centrifugation and lysed by sonication. The following substrates were used: 4-MU-β-D-N,N′,N″-triacetylchitotriose, 4-MU-diacetyl-β-d-chitobioside, and 4-MU-N-acetyl-β-d-glucosaminide. Assays were performed at 25°C in 50 mM sodium phosphate buffer (pH 6.0).

The enzymatic hydrolysis liberates 4-MU. The fluorescence of liberated 4-MU is measured using a fluorimeter with excitation at 360 nm and emission at 450 nm. One unit of activity was defined as the amount of protein (total cell extract) required to release 1 nmol of 4-MU from the appropriate substrate per minute under described conditions.

The hydrolytic activities of strain SP5^T lysed cell extracts with amorphous chitin (prepared as described by Sorokin et al. [21]), starch (S2004; Sigma), microcrystalline cellulose (310697; Aldrich), carboxymethyl cellulose (21900; Fluka), lichenan (L6133; Sigma), laminarin (L9634; Sigma), xanthan (G1253; Sigma), birchwood xylan (X0502; Sigma), and beechwood xylan (38500; Serva) were evaluated by measuring the amounts of total reduced sugars released after treatment by the DNS method (56). Cells were collected by centrifugation and lysed by sonication. Assays were performed at 25°C in 50 mM sodium phosphate buffer (pH 6.0). Substrates were used at the following concentrations (wt/vol): amorphous chitin, 1%; carboxymethyl cellulose, 1%; microcrystalline cellulose, 1%; starch, 1%; birchwood xylan, 1%; beechwood xylan, 1%; lichenan, 1%; laminarin, and 1%; xanthan, 0.5%.

The hydrolytic activities of strain SP5^T lysed cell extracts were also determined using p-nitrophenyl-β-d-galactopyranoside (pNPGal), p-nitrophenyl-β-d-xylopyranoside (pNPXyl), pNP-β-d-mannopyranoside (pNPMan), and p-nitrophenyl-β-d-glucopyranoside (pNPGlu). The reactions were performed with 2.3 mM each substrate in 50 mM sodium phosphate buffer (pH 6.0) at 25°C. The activity was measured by release of p-nitrophenyl at 420 nm.

Expression of recombinant endochitinase FRUB_0131 and verification of its functional activity.

The gene FRUB_0131 was amplified from F. ruber genomic DNA by PCR using primers FIM0131F (5′-CGGGATCCGCGGCCGAGCCGCCCG-3′) and FIM0131R (5′-GCCTGCAGTTACCGCTTTTGCCGGCTC-3′). The resulting PCR products were digested with BamHI and PstI and inserted into pQE80 (Qiagen) at BamHI and PstI sites, yielding the plasmid pQE80-FRUB_0131. This plasmid was transformed into Escherichia coli strain DLT1270. The recombinant strain was grown at 37°C in LB medium supplemented with ampicillin and induced to express recombinant enzyme by adding isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 1.0 mM at an optical density at 600 nm (OD₆₀₀) of approximately 0.5 and incubated further at 37°C for 3 h.

Cells were harvested by centrifugation at 3,500 × g for 15 min at 4°C and resuspended in 50 mM phosphate buffer (pH 7.5), 0.3 M NaCl, and 5 mM imidazole. The cell extracts after sonication were centrifuged (15,000 × g, 4°C, 30 min), and the recombinant proteins from the supernatant were purified by metal affinity chromatography using a Ni-NTA spin kit (Qiagen). Upon elution from the column, the proteins were dialyzed on a Slide-A-Lyzer Mini dialysis unit (Thermo Scientific) against 12.5 mM phosphate buffer (pH 6.0) at 4°C for 3 h (molecular mass cutoff, 3.5 kDa).

Hydrolytic activities of recombinant enzyme FRUB_0131 were analyzed using 4-MU-β-d-N,N′,N″-triacetylchitotriose, 4-MU-diacetyl-β-d-chitobioside, and 4-MU-N-acetyl-β-d-glucosaminide. Assays were performed at 25°C in 100 mM sodium phosphate buffer (pH 6.6) as described above for strain SP5^T lysed cell extracts. The protein sample obtained from E. coli strain DLT1270 following the same protocol was used as a negative control in all assays.

Accession number(s).

The annotated genome sequence of F. ruber SP5^T has been deposited in the DDBJ/ENA/GenBank database under accession number NIDE00000000.