Complete genome sequence of N2-fixing model strain Klebsiella sp. nov. M5al, which produces plant cell wall-degrading enzymes and siderophores.

The bacterial strain M5al is a model strain for studying the molecular genetics of N2-fixation and molecular engineering of microbial production of platform chemicals 1,3-propanediol and 2,3-butanediol. Here, we present the complete genome sequence of the strain M5al, which belongs to a novel species closely related to Klebsiella michiganensis. M5al secretes plant cell wall-degrading enzymes and colonizes rice roots but does not cause soft rot disease. M5al also produces siderophores and contains the gene clusters for synthesis and transport of yersiniabactin which is a critical virulence factor for Klebsiella pathogens in causing human disease. We propose that the model strain M5al can be genetically modified to study bacterial N2-fixation in association with non-legume plants and production of 1,3-propanediol and 2,3-butanediol through degradation of plant cell wall biomass.

The bacterial strain M5al (=ATCC BAA–1236 = DSM 3539 = NCIB 12204) was first designated as Aerobacter aerogenes [1], [2], then Klebsiella pneumoniae [3], and later K. oxytoca [4]. M5al was isolated from soil and is able to convert atmospheric nitrogen to ammonia [2], degrade aromatic compounds [5], [6], and produce industrial platform chemicals 1,3-propanediol and 2,3-butanediol [7]. M5al is free of capsular polysaccharides and thus was thought to be non-pathogenic and used safely [7]. M5al is easy to be genetically manipulated and has been extensively studied as a model strain in molecular genetics of N2-fixation [8], [9] and molecular engineering of microbial production of 1,3-propanediol and 2,3-butanediol [7], [10].

The whole genome sequence of the strain M5al has attracted considerable attention. The genome of an M5al mutant VJSK009 was sequenced in 2004 (http://genome.wustl.edu/genomes/view/klebsiella_oxytoca_m5al). The genome of the strain M5al was first sequenced by Solexa technology and assembled to a 114-contig draft genome (GenBank accession no. AMPJ00000000.1) [11] and later sequenced by the PacBio Single Molecule, Real-Time (SMRT) technology [12] and assembled to a nearly complete 6-contig genome (GenBank accession no. LWKU00000000.1). Here, we present the complete genome sequence and discuss M5al’s potential to colonize and provide nitrogen to plants.

The genome DNA of the strain M5al was inserted into a 15–20 kb insert library and sequenced by the PacBio SMRT technology at the Duke University Genome Sequencing & Analysis Core Resource. The sequencing resulted in 181,416 high-quality filtered reads with an average length of 13,226 bp. High-quality reads were assembled by the RS_HGAP_Assembly.3 in the SMRT analysis v2.3.0. The final assembly produced a complete genome of 307-fold coverage. The genome was annotated by the NCBI Prokaryotic Genome Annotation Pipeline [13]. The complete genome consists of a circular chromosome containing 5,800,138 bp with 55.9 mol% GC content, 5593 genes including 25 rRNA genes, 85 tRNA genes, and 7 non-coding RNA genes. The complete genome contains what appears to be a movable element (about 35 kb from locus_tag BWI76_10080 encoding a tRNA-Ser to BWI76_10240 encoding an integrase) and a frame-shifted transposase gene (BWI76_15780) inserted into a gene (split into BWI76_15775 and BWI76_15785) encoding a replication protein, which are missing from the previous 6-contig genome.

The strain M5al was submitted as K. oxytoca when the two previously sequenced genomes were deposited in GenBank and was changed to K. michiganensis in July 2016 based on overall genome relatedness. K. michiganensis is closely related to K. oxytoca [14]. However, the genome average nucleotide identities (ANIs) between M5al and the K. michiganensis type strain DSM 25444T and between M5al and the K. oxytoca type strain NBRC 105695T are 93.31% and 91.04%, respectively, lower than the threshold value of 95%–96% for species delimitation [15]. Therefore, M5al belongs to a novel species closely related to K. michiganensis.

Carbohydrate-active enzymes (CAZymes) catalyze the synthesis, degradation, and modification of complex carbohydrates [16]. CAZymes are important for the biotech industry, as the degradation of complex polysaccharides (cellulose, hemicellulose, pectin, and lignin) in the plant cell wall by microbes produces biofuels [16]. We identified 210 CAZymes in the M5al genome using the dbCAN web server (http://csbl.bmb.uga.edu/dbCAN/annotate.php). The CAZymes of M5al consist of 93 glycoside hydrolases (GH), 48 glycosyltransferases (GT), 8 polysaccharide lyases (PL), 29 carbohydrate esterases (CE), 10 auxiliary activities (AA), and 22 carbohydrate-binding modules (CBM). Notably, the CAZymes of M5al contain seven cellulases (two endoglucanases and five β-glucosidases), six xylanases, and seven pectinases (Table S1).

Most plant cell wall-degrading enzymes involved in enterobacterial phytopathogenesis are secreted through the type II secretion (T2S) system [17]. The plant cell wall-degrading enzymes to be secreted are first translocated from the cytoplasm into the periplasm through the Sec general secretory pathway or the twin-arginine translocation (Tat) system [18]. M5al contains genes encoding all the components of the T2S, Sec, and Tat systems (Table S2).

M5al grew on agar media containing cellubiose, xylan, or pectin as the sole carbon source and likely secreted β-glucosidases, xylanases, and pectinases into agar media (Fig. 1), indicating that the T2S, Sec, and Tat systems are functional. However, M5al did not show a clear endoglucanase activity for degradation of carboxymethyl cellulose (Fig. 1). We propose that M5al may produce 1,3-propanediol and 2,3-butanediol through degradation of plant cell wall biomass after activation of its endoglucanases and introduction of foreign exoglucanases.

Fig. 1

Agar plate assays showing degradation of plant cell wall polysaccharides by activities of β-glucosidase, xylanase, and pectinase from Klebsiella sp. strain M5al. (A–E) M5al (1 × 107 cells) grown on an agar medium [0.2% (w/v) NaNO3, 0.1% K2HPO4, 0.05% MgSO4, 0.05% KCl, and 1.5% purified agar] without addition of carbon sources (A), or supplemented with 0.2% cellubiose (B), 0.2% xylan (C), 0.2% pectin (D), or 0.2% carboxymethyl cellulose (CMC) (E) at 30 °C for 2 d. (F–I) Haloes, indicating activities of enzymes from M5al, formed on the agar medium containing ρ-nitrophenyl β-d-glucopyranoside (ρ-NPG, a chromogenic substrate of β-glucosidase; 0.15%) (F), 4-O-Methyl-d-glucurono-d-xylan dyed with Remazol brilliant blue R (RBB-xylan; 0.2%) (G), or pectin precipitated by hexadecyl trimethyl ammonium bromide (CTAB) (H). M5al did not use CMC (E) or degrade CMC stained by Congo red (I). RBB may inhibit M5al growth; degradation of RBB-xylan (G) was detected by putting a xylan agar plug (15 mm in diameter) grown with M5al for 2 d (C) on the center of the RBB-xylan plate.

Bacteria can use plant cell wall-degrading enzymes to weaken the plant cell walls and infect plant tissues. We introduced the plasmid pPROBE-pTetr-TT [19] carrying the gfp reporter into M5al, inoculated rice seedlings grown in a semi-solid medium with the gfp-labeled M5al cells (1 × 107cells per seedling), and visualized bacterial infection of rice roots using laser scanning confocal microscopy. The gfp-labeled M5al cells preferentially colonized the root border cells surrounding root caps, root hair zones, and the junctions between primary roots and lateral roots (Fig. 2), showing typical colonization patterns of rhizobacteria [19], [20]. The gfp-labeled M5al cells also colonized the cracks of emerging lateral roots and disrupted root epidermal cells but did not massively degrade the plant cell walls leading to root rotting symptoms (Fig. 2). Therefore, M5al controls the plant cell wall-degrading activities and is not pathogenic to the rice seedlings. Yoo et al. noted that M5al showed a low level of N2 fixation in association with rice plants [21]. Likely, M5al can be used as a model strain to study the mechanisms of bacterial N2 fixation contributing to non-legume plants.

Fig. 2

Confocal microscopic images showing gfp-labeled Klebsiella sp. M5al cells colonizing roots of rice seedlings. GFP fluorescence (500 nm–530 nm) of bacterial cells is presented in green while root autofluorescence (550 nm–650 nm) is presented in red. Arrowheads point to specific locations of bacteria that colonized root border cells surrounding root caps (A), root hair zones (B–F), junctions between epidermal cells at the root hair zone (B), junctions between primary roots (PR) and lateral roots (LR) (C), cracks caused by the emergence of lateral roots (D), and disrupted root epidermal cells (E, F). Images D and F show three-dimensional views of optical sectioning through root pieces. Image F is a magnified view of the frame in image E. Scale bars indicate 50 μm.

Siderophores are low molecular weight iron chelators produced by microbes to sequester iron and perturb host iron homeostasis [22]. Bacteria can effectively uptake the nitrogenase metal cofactors Fe and Mo via siderophores to support N2 fixation [23]. Moreover, siderophores act as both toxins and immunomodulators to the host to induce inflammatory cytokines and bacterial dissemination [22], [24]. Siderophores, especially yersiniabactin, aerobactin, salmochelin, and colibactin, are critical virulence factors of K. pneumoniae causing invasive disease in immunologically competent human hosts [25]. The M5al genome contains the gene clusters for synthesis and transport of enterobactin (Ent-Fep-Pfe cluster) and yersiniabactin (ybt cluster) and the genes encoding the aerobactin receptor (iutA) and transporter (FhuABCD) (Table 1). Moreover, M5al is able to produce and secrete siderophores into agar media (Fig. S1). Enterobactin is the most common siderophore produced by the enterobacteria and has one of the highest ferric iron affinities [22]. To oppose iron acquisition by enterobactin, host epithelial cells and neutrophils secrete the protein siderocalin to bind and sequester enterobactin. However, yersiniabactin can evade siderocalin [22]. Therefore, further study is needed regarding the virulence of the yersiniabactin from M5al, as well as the function of the siderophores in relation to N2 fixation and the production of 1,3-propanediol and 2,3-butanediol by M5al. The virulent yersiniabactin system should be knocked out for the safe use of the model strain M5al.

Table 1

Klebsiella sp. M5al genes for synthesis and transport of siderophores.

Siderophore	Locus_tag	GenPept record
Enterobactin
	BWI76_01965	TonB-dependent siderophore receptor FepA
	BWI76_07690	enterobactin synthase entD
	BWI76_07695	TonB-dependent siderophore receptor FepA
	BWI76_07705	enterochelin esterase Fes
	BWI76_07710	MbtH family protein ybdZ
	BWI76_07715	non-ribosomal peptide synthetase entF
	BWI76_07720	iron-enterobactin transporter ATP-binding protein FepC
	BWI76_07725	iron-enterobactin transporter permease FepG
	BWI76_07730	iron-enterobactin transporter FepD
	BWI76_07735	MFS transporter EntS
	BWI76_07740	Fe2+-enterobactin ABC transporter substrate-binding protein FepB
	BWI76_07745	isochorismate synthase entC
	BWI76_07750	2,3-dihydroxybenzoate-AMP ligase entE
	BWI76_07755	isochorismatase entB1, entB2
	BWI76_07760	2,3-dihydro-2,3-dihydroxybenzoate dehydrogenase entA
	BWI76_07765	thioesterase entH
Yersiniabactin
	BWI76_14505	salicylate synthase ybtS
	BWI76_14510	MFS transporter ybtX
	BWI76_14515	ABC transporter permease ybtQ
	BWI76_14520	ABC transporter ATP-binding protein ybtP
	BWI76_14525	AraC family transcriptional regulator ybtA
	BWI76_14530	non-ribosomal peptide synthetase irp2
	BWI76_14535	polyketide synthase irp1
	BWI76_14540	oxidoreductase ybtU
	BWI76_14545	thioesterase ybtT
	BWI76_14550	2,3-dihydroxybenzoate-AMP ligase ybtE
	BWI76_14555	TonB-dependent siderophore receptor fyuA
Aerobactin
	BWI76_05125	ferrichrome porin FhuA
	BWI76_05130	iron-hydroxamate transporter ATP-binding subunit FhuC
	BWI76_05135	iron-hydroxamate transporter substrate-binding subunit FhuD
	BWI76_05140	Fe3+-hydroxamate ABC transporter permease FhuB
	BWI76_11185	TonB-dependent siderophore receptor iutA

Nucleotide sequence accession number

The complete genome sequence of the strain has been deposited in GenBank under the accession number CP020657.

Author contributions

QA designed the study. ZY, SL, YL assembled and analyzed the genome. ZY, ZJ, JZ did the experiments. ZY, QA wrote the manuscript.

Conflict of interest statement

The authors declare no conflict of interest.