Whole-genome sequencing and identification of Morganella morganii KT pathogenicity-related genes

Background The opportunistic enterobacterium, Morganella morganii, which can cause bacteraemia, is the ninth most prevalent cause of clinical infections in patients at Changhua Christian Hospital, Taiwan. The KT strain of M. morganii was isolated during postoperative care of a cancer patient with a gallbladder stone who developed sepsis caused by bacteraemia. M. morganii is sometimes encountered in nosocomial settings and has been causally linked to catheter-associated bacteriuria, complex infections of the urinary and/or hepatobiliary tracts, wound infection, and septicaemia. M. morganii infection is associated with a high mortality rate, although most patients respond well to appropriate antibiotic therapy. To obtain insights into the genome biology of M. morganii and the mechanisms underlying its pathogenicity, we used Illumina technology to sequence the genome of the KT strain and compared its sequence with the genome sequences of related bacteria. Results The 3,826,919-bp sequence contained in 58 contigs has a GC content of 51.15% and includes 3,565 protein-coding sequences, 72 tRNA genes, and 10 rRNA genes. The pathogenicity-related genes encode determinants of drug resistance, fimbrial adhesins, an IgA protease, haemolysins, ureases, and insecticidal and apoptotic toxins as well as proteins found in flagellae, the iron acquisition system, a type-3 secretion system (T3SS), and several two-component systems. Comparison with 14 genome sequences from other members of Enterobacteriaceae revealed different degrees of similarity to several systems found in M. morganii. The most striking similarities were found in the IS4 family of transposases, insecticidal toxins, T3SS components, and proteins required for ethanolamine use (eut operon) and cobalamin (vitamin B12) biosynthesis. The eut operon and the gene cluster for cobalamin biosynthesis are not present in the other Proteeae genomes analysed. Moreover, organisation of the 19 genes of the eut operon differs from that found in the other non-Proteeae enterobacterial genomes. Conclusions This is the first genome sequence of M. morganii, which is a clinically relevant pathogen. Comparative genome analysis revealed several pathogenicity-related genes and novel genes not found in the genomes of other members of Proteeae. Thus, the genome sequence of M. morganii provides important information concerning virulence and determinants of fitness in this pathogen.


Background
The Gram-negative anaerobic rod Morganella morganii is the only species in the genus Morganella, which belongs to the tribe Proteeae of the family Enterobacteriaceae. The other genera in the tribe Proteeae are Proteus and Providencia. Species belonging to Morganella, Proteus, and Providencia are found in the environment and as part of the normal flora of humans. They are also important opportunistic pathogens, which cause a wide variety of nosocomial infections following surgery. Reports of individual cases of infection and nosocomialoutbreaks have revealed that infection with M. morganii can lead to major clinical problems, which are usually associated with common causes of catheter-associated bacteriuria, infections of the urinary and hepatobiliary tracts [1][2][3][4][5], wound infection, and septicaemia [6][7][8][9]. A few devastating infections with M. morganii that were associated with a high mortality rate following bacteraemia sepsis and/or nosocomial infection have also been reported, although most of such infections respond well to appropriate antibiotic therapy [3,[10][11][12][13].
Although M. morganii was formerly classified as Proteus morganii [14], it was later assigned to the genus Morganella on the basis of DNA-DNA hybridisation results [15]. Members of the genus can ferment trehalose, and express lysine decarboxylase and ornithine decarboxylase [16].
The production of urease has a fitness factor that influences bacterial growth and biofilm formation during urinary tract infections. Other virulence factors may include iron acquisition systems, type-3 secretion system (T3SS), two-component systems (TCS), proteins that function in immune evasion (IgA protease), and haemolysins [35].
The environment found in the guts of nematodes or insects may be an important determinant of bacterial pathogenicity [38]. Ethanolamine, which is abundant in human diets and the intestinal tracts of humans, can be used by gut bacteria as a source of carbon and/or nitrogen [39]. The association between the use of ethanolamine and the virulence of various pathogens has been reported [39].
Phylogenetic assessment of 16S rRNA sequences indicates that P. mirabilis is the closest relative of M. morganii. Only one complete Protues genome sequence and four draft sequences of Providencia spp. are available.
Here we report the draft genome sequence of a clinical isolate of M. morganii and the results of its bioinformatic analysis to enhance understanding of M. morganii biology. Comparative analysis of the sequences with the sequences of other Proteeae and Enterobacteriaceae genomes identifies potential virulence determinants, which may provide new drug targets.

Results
Epidemiological study of M. morganii infection Over a 6-year period (2006-2011), samples were collected from all patients at Changhua Christian Hospital, Taiwan, who presented with symptoms of Gram-negative bacterial infections. Of 82,861 samples, 1,219 (1.47%) were positive for M. morganii and 3,503 (4.23%) were positive for Proteus spp. As shown in Table 1, M. morganii was ranked between the eighth and fourteenth most prevalent Gramnegative bacterial species isolated from the hospital over 12 consecutive 6-month intervals during the 6 years of the study.
The KT strain of M. morganii was isolated from the blood of a patient who developed sepsis during postoperative care. The KT strain was found to be susceptible to amikacin, ertapenem, gentamicin, meropenem, and cefepime but resistant to ampicillin, amoxicillinclavulanate, cefazolin, cefuroxime, cefmetazole, flomoxef, and cefotaxime.

General features of the M. morganii draft genome
The genome of the M. morganii strain KT, which carries no plasmids, was assembled de novo into 58 contigs (each >200 bp long), which together comprised 3,826,919 bp with a GC content of 51.15%. The largest contig is 410,849 bp long, and the N50 statistic (the minimum contig length of at least 50% of the contigs) is 240,446 bp, with pair-end short read sequencing coverage >1,150-fold. Seven small contigs (each <13 kb) had low-depth reads (0.36-to 0.66fold), whereas two pairs and one triad shared high sequence identity with minor differences at their ends (Additional file 1). The origin of replication assigned on the basis of the GC-skew analysis together with the location of the dnaA gene and DnaA boxes of the genome lies between gidA (MM01685) and mioC (MM01686) (Additional File 1: Figure S1).
As shown in Table 2, comparison of the M. morganii sequences with the complete genome of P. mirabilis HI4320 revealed a 12.2% difference in GC-content of the two species (51.1% for M. morganii vs. 38.9% for P. mirabilis).
Sequences that encode eight 5S rRNAs, one 16S rRNA, and one 23S rRNA were identified using the ribosomal RNA scan application in RNAmmer (http://www.cbs.dtu. dk/services/RNAmmer/) ( Table 2). Further analysis of contigs revealed that the 16S rRNA had a read depth of 7.9-fold and that the 23S rRNA had a read depth of 8.2fold (Additional file 2: Supplementary table 1).

Coding DNA sequences
Database searches identified 3,565 predicted coding sequences (CDSs). Among them, 2,870 CDSs could be placed into clusters of orthologous groups with assigned biological functions ( Figure 1). The proteins annotated as pathogenicity and fitness factors are listed in Table 3, with   additional information in Tables 4, 5, 6 and Additional  Files 3

Prophages and mobile elements
Bacteriophages and transposons, which may contribute to genome plasticity [40], are, in general, laterally acquired elements [32]. In the M. morganii genome, 2 prophages (MM3229 through MM3271 and MM2276 through MM2290) and 12 degenerate prophages together comprise 236 prophage-related genes (Additional file 9: Supplementary table 8). The prophage 1 genes were also found in the genome of Proteus mirabilis, Providencia alcalifaciens, and Providencia rustigianii, whereas the prophage 2 genes are orthologous to those found in the other 14 Enterobacteriaceae genomes.

Drug resistance
A chromosomally encoded β-lactamase from M. morganii has been cloned [41]. The clinical strains with high or low levels of cephalosporinase expression were found to harbour the adjacent ampC (MM3167) and ampR The frequency of infection by 21 bacterial species was monitored over 6-month intervals (the first and second halves of each year) and ranked such that 1 denotes the species most frequently associated with bacterial infection. The column heading '# infections' represents the total number of patients infected with the species indicated. NA: data not available, not significant to the ranking list.

Motility and chemotaxis systems
Fifty three genes that encode proteins required for flagellar structure (MM1735 through MM1785, as well as MM1796 and MM1797) and eight chemotaxis-related genes (MM1786 through MM1793, including cheA, cheW, cheD, tap, cheR, cheB, cheY, and cheZ) are contained in the M. morganii genome. With the exception of the genes that encode two LysR family transcriptional regulators (MM1739 and 1765), the camphor resistance protein CrcB and related genes (MM1743 through MM1749), a short-chain dehydrogenase/reductase SDR (MM1764), and the three insecticidal toxin complex proteins encoded by xptA1A1C1 (MM1780 through MM1782), the organisation of flagellar genes is similar to that of P. mirabilis.

Zinc acquisition
The znuACB high-affinity zinc transporter system was recently shown to be a fitness factor for E. coli and P. mirabilis during experimental urinary tract infection [46,47]. The system is encoded by M. morganii gene products MM2146 through MM2148.

Type III secretion system
A T3SS, which comprises gene products MM0224 through MM0247 and has a low GC-content (43.7%), resides in a 20.8-kb pathogenicity island. This island of 24 genes, which shares homologous syntenic blocks with P. mirabilis [48], contains all the components needed to assemble a T3SS needle complex. Except for the genes that encode effector proteins, the most homologous proteins were orthologues from P. mirabilis (Additional File 5: Supplementary table 4). A putative IpaCBD operon, which encodes chaperones for three proteins (MM0244 through MM0247) was also found. The IpaC and IpaD proteins have low similarity (21% and 38%, respectively) to those of P. mirabilis, whereas the IpaB protein has the highest homology with ipaB of P. mirabilis but low similarity (20-30%) with that of Shigella spp. and Salmonella enterica subsp.

Regulation
Analysis of the M. morganii genome identified seven sigma factor subunits of RNA polymerase. These are the major sigma factor rpoD (σ 70 ,σ A ) encoded by MM1372, and six alternative sigma factors: rpoH (σ 32 , σ H ) encoded by MM1254, rpoN (σ 54 , σ N ) encoded by MM1306 to control promoters for nitrogen assimilation, rpoS (σ 38 , σ S ) encoded by MM1208 to activate stationary-phase promoters, rpoE (σ 24 , σ E ) encoded by MM1906 to regulate extracytoplasmic stresses, rpoF (σ 28 , σ F ) encoded by MM1736 to regulate flagellum-related functions, and a FecR family sigma factor encoded by MM1954. The small RNA regulatory gene ryhB found in E. coli and P. mirabilis [49], which regulates a set of iron-storage and iron-usage proteins, is present in only a single copy (contig 25).

Lipopolysaccharide and the cell capsule
Lipopolysaccharide (LPS), the main structural component of the outer membranes of Gram-negative bacteria, consists of a lipid A molecule and a variable O-antigen. Lipid A released during bacterial lysis induces endotoxic shock. Several O-specific polysaccharides of M. morganii have been investigated [50,51], and at least 55 O-antigens have been identified [52]. The genes predicted to be involved in the synthesis of LPS and the enterobacterial common antigen [53] are listed in Additional File 6: Supplementary

Immunity-like system
Clusters of regularly interspaced short palindromic repeats consist of multiple short nucleotide repeats, which are separated by unique spacer sequences flanked by characteristic sets of CRISPR-associated genes [54,55]. The CRISPR-associated proteins MM3304, MM3306, and MM3307 were identified downstream of the degenerate prophage #6. Comparison of the M. morganii genome with that of the other 14 members of the Enterobacteriaceae analysed revealed orthologues in the pathogens P. stuartii ATCC# 25827, E. coli UTI89, and E. coli APEC O1, all of which have been implicated in causing urinary tract infections [56].

Haemolysins
The gene hmpA, encoding a secreted haemolysin, was originally identified in uropathogenic isolates [57,58]. The two partner genes, hmpBA, encoding secreted proteins that are highly conserved in P. mirabilis [59], are encoded by MM2452 and MM2453 in M. morganii KT. Functional similarity of the E. coli and M. morganii homologues of HmpBA was reported previously [60].

Urease hydrolysis and putrescine production
Rapid urea hydrolysis is a prominent phenotype of Proteeae organisms [61]. The urease enzyme is believed to be a cause of urinary stone formation [62]. The urease of M. morganii, which revealed a high degree of amino acid conservation to P. mirabilis urease, has been purified and characterised [4]. Ureases from both M. morganii and P. mirabilis urease gene cluster are required for virulence [63]. The analysis revealed that the closest homologues of the members of the urease gene cluster ureABCEFGD were the orthologues from Yersinia pseudotuberculosis and Y. enterocolitica, with amino acid identities that ranged from 69% to 94%. Unlike the urease gene cluster of P. mirabilis, the M. morganii genome has a gene that encodes the transporter MM1968, in addition to MM1961 to MM1967. On the other hand, the P. mirabilis ureR gene, which encodes a transcriptional activator, was not found in the genome.
The genome of M. morganii KT contains genes for two pathways involved in putrescine production. The first of these pathways involves ornithine decarboxylase, which is encoded by speF (MM3013), and the putrescine transporter, which is encoded by potE (MM3012). The enzymes in the second pathway, which are encoded by carAB (MM2009 and MM2010), argI (MM0127), argG (MM1552), argH (MM1551), speA (MM2553), and speB (MM2554), also participate in urea production. All of the urease cluster genes that encode enzymes from both pathways have orthologues in P. mirabilis HI4320 [32].

Toxins
Several CDSs that encode potential toxins were found. These include a gene that encodes the cytotoxin RtxA (MM0676), the two XaxAB genes (MM0454 and MM0455) that encode apoptotic toxins, a putative intimin gene for host-cell invasion (MM0208), a HlyD-family toxin scretion protein (MM2481), and a transporter gene (MM2482). Among the nine insecticidal toxin-related genes (Additional File 7: Supplementary table 6), tccB, tccA, and tcdB2 (MM0965 through MM0967) are orthologous to the insecticidal toxin genes of Pseudomonas spp. The three other insecticidal toxin genes xptA1A1C1 (MM1780 through MM1782) were identified in a continuous locus between the flagellum-related genes and chemotaxis genes.

Ethanolamine utilisation system
The ethanolamine utilisation system, which is encoded by the eut operon it were found to vary substantially between species [39], is composed of the genes eutSPQTDMNEJ-GHABCLK-pduST-eutR (MM1148 through MM1130). This region carries two extra genes compared to the 17 eut genes found in other Enterobacteriaceae genomes [64]. In addition, the gene organisation is unique, with pduST located between eutK and eutR. As shown in Table 6, the sequence similarities of the three most similar proteins ranged from 51% to 87%, and species varied.
The comparison revealed that the 17-gene operon is present in Klebsiella pneumonia, E. coli, and Salmonella enterica serovars. The pdu operon, a paralogous operon required for use of propanediol that is found in these three species, was not found in the M. morganii KT genome.
Located upstream of the eut operon, MM1168 through MM1187 encode the enzymes of the cob-cbi operon, which is required for cobalamin (vitamin B 12 ) biosynthesis [65][66][67]. Under aerobic conditions, the activity of EutBC depends on the exogenous supply of cobalamin.

Other determinants of persistence of infection and fitness for infection
Pathogenic bacteria produce superoxide dismutase to protect them from being killed by the reactive oxygen species generated by their hosts [68]. As shown in Additional File 8: Supplementary table 7, the genes involved in countering superoxide stress are katA, soxS, sodC, sodB, oxyR, and sodA.
Among the orthologous genes, 2,411 CDSs were found in the genomes of the 5 Proteeae species analysed, and 1,920 CDSs were found in the 14 Enterobacteriaceae genomes analysed. As shown in Additional file 11: Supplementary table 10, of the genomes analysed, that of P. rettgeri is the most closely related genome to that of M. morganii, sharing 2,802 orthologous CDSs. The orthologous genes encode proteins for drug resistance, pathogenicity (IgA protease and LPS), motility, iron acquisition, ethanolamine use, and urease production, as well as components of an immunity-like system (CRISPR) and fimbrial adhesins, toxins, and haemolysin.

Discussion
Epidemiological studies have revealed that M. morganii is frequently isolated from clinical specimens collected from patients with nosocomial bacterial infections [3,5,11,[72][73][74]. Strains of M. morganii that confirm the chromosomal origin of the plasmid-located cephalosporinases [17] are now found throughout the world [75]. Genes from M. morganii that encode cefotaxime-hydrolysing β-lactamases have also been reported with increasing frequency [42]. Plasmidborne drug resistance factors have also increased the virulence of M. morganii [26,27]. Gene products that confer multidrug resistance, including metallo-β-lactamases and efflux pumps for tetracycline, tellurite, bicyclomycin, and kasugamycin, have been commonly reported for clinical isolates of M. morganii. In the event of an outbreak, this situation poses a potential threat owing to the absence of a proper antibiotic therapy.
Compared with other members of the genus Proteeae, which have GC contents that range from 39% to 43%, the GC content of M. morganii is 51%. which is genetically different from other species [76,77] and therefore assigned to the genus Morganella [14,15,61,76]. The plasmid-borne gene that encodes lysine decarboxylase was once used to classify bacteria [78,79]. However, the chromosomal gene that encodes ornithine decarboxylase was subsequently adopted as a characteristic feature to classify members of the genus Morganella [80,81].
Information on the 16S rRNA gene and paralogs in genome is important for evolution and bacterial population studies [40]. The P. mirabilis genome has seven rRNA operons (six 16S-23S-5S operons and one 16S-23S-5S-5S operon) [32]. Analysis of the M. morganii KT genome sequence revealed eight duplications of the 16S and 23S rRNA genes while 5S rRNA were also 8. The prophage genes that were found in the other Enterobacteriaceae species appeared to comprise 7% of the M. morganii genome.
Interestingly, the ICEPm1 and ICE/R391 genes, which are present in many P. mirabilis isolates [21], are not found in strain KT. Seventeen copies of transposase genes of the IS4 family were not present in other Proteeae genomes, which implies that different transposition events occurred during the evolution of M. morganii and of these species.
Unlike the flagellar genes of P. mirabilis, M. morganii KT has LysR family transcriptional regulatory genes (MM1739 and MM1765), genes MM1743 through MM1749 that encode the camphor resistance gene CrcB and related proteins, a short-chain dehydrogenase/reductase SDR gene (MM1764), and the insecticidal toxin complex genes xptA1A1C1 (MM1780 through MM1782).
Both M. morganii KT and P. mirabilis have duplicated MR/P fimbrial operons, albeit with different numbers of genes [32]. As shown in Figure 2, whereas three copies of the MrpI recombinase gene are found in M. morganii KT, only one copy is in the P. mirabilis genome. In M. morganii KT, the mrp' operon is oriented opposite to mrpI, whereas mrpI is not present in the P. mirabilis mrp' operon.
More than half of the pathogenicity island genes are conserved and found to show collinear synteny between M. morganii and P. mirabilis. However, genes that encode components of T3SSs have low sequence similarities.
EDTA-sensitive protease Zap gene clusters have been found in many Proteus spp. and E. coli clinical strains but are not produced by Providencia spp. and Morganella spp. [37]. The zapABCD genes found in M. morganii KT differ from those in P. mirabilis HI4320 zapEEEABCD [32] insofar as KT has an incomplete zapE gene downstream of zapABCD. Although the M. morganii urease shares a high degree of sequence similarity to P. mirabilis ureases, the presence of the unique transporter gene (MM1968) and the absence of ureR in M. morganii suggest that the two species use different systems to regulate urease transport.
Although none of the insecticidal toxin genes were found in P. mirabilis, some were found in Xenorhabdus [82,83], Pseudomonas [84], Yersinia, and Photorhabdus [85]. In Photorhabdus, the related toxin complex is released upon invasion of the nematode host [86]. Why the M. morganii clinical isolate harbours eight insecticidal toxin genes remains to be investigated.
The urinary and hepatobiliary tracts are two major portals of entry for M. morganii [3]. Compared with P. mirabilis HI4320 [32], M. morganii has fewer types of fimbriae (three vs. five) and fewer gene clusters (8 vs. 17). This implies that M. morganii may be less virulent than P. mirabilis in the context of urinary tract infections.
Given that the intestine is thought to provide a rich source of ethanolamine [39], the eut operon likely plays a critical role in enabling M. morganii, E. coli, Klebsiella spp., and Yersinia spp. to use ethanolamine as a source of carbon and/or nitrogen to colonise the intestine. However, the other Proteeae genomes lack the eut operon and pduTS, which together encode proteins that help to establish a microcompartment [87], and the cobcbi operon, which encodes the enzymes needed for cobalamin biosynthesis.
The eut operon, which includes pduST, and the cobcbi operon likely provide the fitness factors required for colonisation by intestinal bacteria. This may explain why M. morganii is more frequently associated with nosocomial bacterial infections than other members of Proteeae.
Typically antibiotics target the essential cellular function, like cell-wall synthesis, ribosomal function, or DNA replication [88]. The whole genome sequencing approaches of human hosts and pathogens facilitate the growing understanding of bacteria infectious disease mechanisms and help to reveal crucial host-pathogen interaction sites [89]. The pathogenicity genes allows usage of computation approaches to identify potential drug targets such as the conserved proteins found in common pathogens [90]. T3SSs is highly conserved in many disease-causing gram-negative pathogens and hence has been used as an alternative strategy for drug target design [90,91]. The efflux pumps which help to get rid of toxic substances also promote biofilms, thus making them attractive targets for antibiofilm measure [92,93]. In many bacteria, biofilm formation, siderophore production and adhesion activity are linked traits. Therefore, drugs that could target bacterial adhesions while colonization could be therapeutically useful [94]. Selective toxicity is another antibiotics approach, which aims to have highly effective against the microbial, but no harm to humans. The unique metabolic pathways identified in Margonella may be considered as the new drug targets.

Conclusions
The pathogenicity-related genes identified in the M. morganii genome encode drug resistance determinants and factors that influence virulence, such as fimbrial adhesins, flagellar structural proteins, components of the iron acquisition system, T3SS, and TCS, an IgA protease, haemolysins, ureases, and insecticidal and apoptotic toxins. Comparative analysis with 14 other Enterobacteriaceae genomes revealed several systems that vary between species. These include transposes of the IS4 family, insecticidal toxins, T3SS components, and proteins required for ethanolamine utilisation and cobalamin biosynthesis. It is interesting to note that neither the eut operon (which includes pduST) nor the cob-cbi operon is found in other Proteeae genomes. Nevertheless, the eut operon is also found in several other non-Proteeae enterobacteria genomes, albeit with different gene organisation.
In summary, this is the first report of an M. morganii genome sequence. Comparative genome analysis revealed several pathogenicity-related genes and genes not found in other Proteeae members. The presence of the eut operon (which includes pduST) and the cob-cbi operon in M. morganii but not in the other Proteeae genomes studied may explain why M. morganii is more frequently associated with nosocomial bacterial infections. Moreover, the evidence that M. morganii shares features with other non-Proteeae enterobacteria suggests that horizontal gene transfer has occurred between M. morganii and other intestinal bacteria.

Bacterial strains and culture conditions
With approval by the institutional biosafety committee, we isolated M. morganii strain KT (year 2009) from blood of a 57-year-old man during postoperative care. The patient had a medical history of type 2 diabetes mellitus, hepatocellular carcinoma, rectal cancer and gallbladder stone. He was admitted for rectal cancer surgery, and received chemotherapy and radiotherapy. Following surgery, bacteraemia caused sepsis.
Colony morphology, Gram staining, oxidase testing (Dry Slide; Difco Laboratories, Detroit, MI), catalase testing, and routine biochemical reactions identified M. morganii as the agent responsible for the infection. A presumptive diagnosis of infection with M. morganii was confirmed using API-20 kit reagents (Bio Merieux Vitek; Hazelwood, Mo) and the BACTEC NR-860 apparatus (Becton Dickinson Diagnostic Instrument Systems, Franklin Lake, NJ).

Genomic DNA preparation
We cultured M. morganii in trypticase soy broth or on trypticase soy agar. Bacteriological media were purchased from Biostar Inc. (Taiwan).
Genome DNA from the M. morganii strain KT was isolated using reagents from QIAGEN DNeasy kits (QIAGEN Inc., Valencia, CA).

Cloning and sequencing
Genome sequencing was performed using the whole genome shotgun strategy [95]. Genomic DNA sequencing reads of 101 bp pair-end reads, an average distance between pair reads were 200 bp, and were generated using an Illumina GA IIx (Solexa) sequencer [96].

Draft genome assembly and validation of contigs
Short sequencing reads were assembled into contigs using ABySS version 1.2.7 software [97]. Contigs were evaluated using the re-sequencing program SOAP2 [98], a tool for aligning raw reads into contigs to evaluate and validate the coverage and depths of read information for each contigs. All contigs were aligned with short reads, with a depth threshold of at least 300 reads.

Identification and annotation of coding sequences and genes encoding tRNAs and rRNAs
The draft genome of M. morganii was analysed using our own integrated annotation pipeline composed of prediction and database search tools. Glimmer (version 3.02) was used to predict assembled contig sequences for prokaryotic CDS regions [99]. Potential long CDSs were extracted using the "long-orfs" program from the Glimmer software suite. These long CDSs were then used by Glimmer to predict CDSs in all contigs.
We used BLAST to query non-redundant protein databases with all predicted CDS regions [100] and thereby find and validate significant protein identifications with E value (E < 1e-5). The annotation of validate significant protein with Clusters of Orthologous Groups functional classification were identified from protein genbank records [101]. The EMBOSS analysis package [102] was used to extract and covert sequences and get predicted open reading frame for further manually verified.
We used tRNAscan-SE to predict prokaryotic transfer RNA (tRNA) genes [103]. Ribosomal 5S, 16S, and 23S RNA (rRNA) genes predictions were performed using RNAmmer (version 1.2) [104]. Origins of replication were assigned based of the GC-skew analysis together with the location of the dnaA gene and DnaA boxes of the genome, using Ori-Finder [105]. Horizontally acquired DNA by anomalies in the G+C content was calculating by perl programming languages.
The previously published genome and protein sequences of Enterobacteriaceae genomes were downloaded from the National Center for Biotechnology Information and draft genome and protein sequences (four Providencia spp.) of Enterobacteriaceae genomes were downloaded from the Genome Institute at Washington University.
We used BLAST program and formatted proteins of other Enterobacteriaceae organisms as databases for comparison, validate orthologous protein with E value (E < 1e-4), identity (> 30%) and threshold to length percentage of alignment.

To assess if the M. morganii strain KT contained plasmids
The plasmid DNA was isolated by the procedure of alkaline denature method [106] and then separated in 1.0% agarose gel by gel electrophoresis. Plasmids presences were subsequently visualised by UV exposure of the ethidium bromide stained gel.

Nucleotide sequence accession number
The draft genome sequence of M. morganii KT has been deposited in the DDBJ/EMBL/GenBank under the accession number ALJX00000000. The version described in this paper is the first version, ALJX01000000.