Plant Growth Promoting Potentials of Beneficial Endophytic Escherichia coli USML2 in Association with Rice Seedlings

An endophytic Escherichia coli USML2 originally isolated from the inner part of an oil palm (Elaeis guineensis Jacq.) leaf tissue was inoculated to rice seedlings to investigate its ability in colonising plant inner tissues and promoting growth. Infection of E. coli USML2 was initiated by colonisation on the root surface, invasion of the interior root system followed by endophytic spreading. Inoculation of E. coli USML2 in the rice rhizosphere zone resulted in a significant increase in leaf numbers (33.3%), chlorophyll content (33.3%), shoot height (34.8%) and plant dry weight (90.4%) of 42 days old rice seedlings as compared to the control. These findings also demonstrated the ability of E. coli USML2 to spread endophytically which serves as a beneficial strategy for the bacterium to colonise the host plant and gain protection against adverse soil conditions. The genome of E. coli USML2 had also revealed predicted genes essential for endophytic bacterial colonisation and plant growth promotion which further proven potentials of E. coli USML2 as Plant Growth Promoting Endophyte (PGPE).


INTRODUCTION
Escherichia coli is a harmless natural inhabitant of the gastrointestinal tract of animals and humans (Ingerson-Mahar & Reid 2011). Its dispersal into the environment is due to release of faecal matter from the host. These continuous bulk transfers would result into stable population of E. coli in secondary habitats such as the soil (Dublan et al. 2014). Therefore E. coli was presumed not to be a normal inhabitant of the soil and has been used as an indicator of faecal pollution (Nautiyal et al. 2010;Dublan et al. 2014). However, discovery of indigenous E. coli strains in undisturbed soils of seven uninhabited areas in India with temperatures ranging from −10°C to 45°C demonstrated that E. coli is a member of the natural soil biota. Ubiquity of E. coli in these diverse soils clearly suggested that it should be treated as native soil borne bacteria (Nautiyal et al. 2010).
Occurrence of E. coli in soils is due to its ability to survive, adapt and actively grow in extra-intestinal environments (Nautiyal et al. 2010;Ingerson-Mahar & Reid 2011;Méric et al. 2013;Blount 2015). Thus, it was hypothesised that persistence of E. coli for long periods of time in tropical, subtropical and temperate environments was in response to available nutrients in soil. Substantial nutrient supplies in soil are also provided from excretion of root exudates that consist of carbohydrates, amino acids, organic acids, vitamins and phytohormones acting as chemoattractants to which bacteria respond (Narula et al. 2009;Pedraza et al. 2009;Nautiyal et al. 2010;Berg et al. 2013). Thus, this condition attracts broad diversities of soil bacteria to live in the rhizosphere zone and consequently, E. coli strains must compete for survival (Berg et al. 2013;Dudeja & Giri 2014). This is probably the reason why E. coli has been frequently found endophytically in interior plant tissues of plants where the surrounding environment is free from faecal contamination (Dublan et al. 2014). Moreover, interior plant tissues also offer a less competitive niche for their survival and persistence (Berg et al. 2013). Méric et al. (2013) claimed that presence of E. coli in inner parts of leaves is believed to gain relative protection against adverse conditions present on external surfaces of plants. In fact, interior tissues of the host plant have become a favourable habitat for bacteria due to the availability of abundant nutrients and its environmental stability (Ikeda et al. 2010). Hardoim et al. (2008) claimed that E. coli with the ability to live and persist in planta are known as endophytes.
Plant Growth Promoting Endophytes (PGPE) usually possess close relationships with their host plant (Berg et al. 2013) and are able to promote growth, 121 health and development of the host (Taghavi et al. 2010). However, knowledge about any potential beneficial effects of endophytic E. coli strains on its host plant is very much limited. In fact, the first and only report on plant growth promoting effects of E. coli was demonstrated by Nautiyal et al. (2010). It was reported that endophytic colonisation of E. coli exhibited significant growth promotion of 60-daysold maize (Zea mays cv. Arkil) with increased root length, shoot length and plant dry weight when compared with uninoculated seedlings. Hence, it revealed the potential of E. coli as a plant growth promoting endophyte. However, E. coli strains used by Nautiyal et al. (2010) were soil borne which was introduced anthropically by immersing surface sterilised maize seeds in suspension of E. coli cells (8 log10 CFU/mL). Additionally, study by Méric et al. (2013) which have isolated E. coli strains from interior plant tissues have not reported any information on the effect of its endophytic presence on growth of its host plant. However, earlier findings by Tharek et al. (2011) have successfully isolated E. coli USML2 from inner leaf tissues of surface sterilised healthy oil palm (Elaeis guineensis Jacq.). The isolate was initially identified as an Enterobacter species and it was shown to be able to infect and establish an endophytic relationship with rice seedlings Oryza sativa (MR220). The strain was selected as the most suitable candidate to investigate further on how an endophytic plant associated E. coli could potentially promote plant growth and understand the implicated responsible genes involved. Use of E. coli in this study is also an advantage due to its rapid cell division rate and rapid adaptation to the environment. Furthermore E. coli also requires simple nutritional prerequisites and has well established genetics and completed genome sequence (Brambilla 2014). Moreover, to our knowledge, no study has been conducted to discover the potential of indigenous plant-originated E. coli on plant growth promotion abilities. In addition, this is the first report on endophytic plant associated E. coli which aids in understanding the fundamental processes of plantbacterial interaction. Thus, the aims of this study are as follows: (i) to observe the ability of an E. coli USML2 in endophytic colonisation and plant growth promotion in association with rice seedlings and (ii) to predict all essential genes of E. coli USML2 which are involved in promoting endophytic colonisation and plant growth enhancement.

Non-Pathogenic Plant Associated Endophytic E. coli USML2
The strain E. coli USML2 used in this study was isolated from the inner segments of leaf tissues of surface sterilised oil palm plant (Elaeis guineensis Jacq.). The isolate, which was initially identified as an Enterobacter species, was shown to be able to infect and establish an endophytic relationship in rice seedlings Oryza sativa variety MR220 (Tharek et al. 2011). However, based on more indepth molecular identification analysis via polymerase chain reaction (PCR) and amplification of full length 16S rRNA gene fragment (~1.4 kb), the isolate was re-identified as a plant-associated E. coli USML2 (Tharek et al. 2017). The genome sequence data was deposited to NCBI with GenBank accession number of CP011124. The genome exhibited no genes involved in pathogenicity which reveals this strain as a non-pathogenic E. coli. Additionally, E. coli USML2 is in the same clade as E. coli K-12 sbstr. MG1655, E. coli K-12 DH10B and E. coli BW2952 which are strains known as non-pathogenic to humans. This strain was deposited at the Microbial Culture Collection Unit, Institute of Bioscience, Universiti Putra Malaysia, with the accession number of UPMC432. The whole genome shotgun project has been deposited at DDBJ/EMBL/Genbank under the accession number CP011124 (Tharek et al. 2017).

Predicted Genes for Promoting Endophytic Colonisation and Plant Growth Enhancement of E. coli USML2
The genome sequence data of E. coli USML2 with GenBank Accession Number CP011124 was used to annotate and predict genes which are related to colonisation and plant growth promotion abilities (Tharek et al. 2017). These genes involved in metabolic pathways were generated by Kyoto Encyclopaedia of Genes and Genomes (KEGG) embedded in RAST server (Kanehisa et al. 2016;Aziz et al. 2008;Overbeek et al. 2014).

Construction of E. coli USML2 flhC::Km r Mutant
A mutant strain of E. coli USML2 has been constructed to verify endophytic colonisation potentials of the isolate. Mutagenesis of a gene involved in flagella biogenesis of E. coli USML2 was performed and the flhC gene (located within the flhDC operon), was selected for mutation. This operon codes a vital positive regulator which controls the initiation of flagellum biosynthesis. An flhC::Km r mutant was constructed in a suicide vector (pDM4) (Yao et al. 2007). Initially, flhDC genes were amplified from the chromosomal DNA of E. coli USML2 and ligated into pLUG-Prime ® plasmid vector resulting in pPM10-flhC. Insertion of kanamycin resistant gene (Km r ) amplified from pKEM100 (Hasan et al. 2004) within pPM10-flhC resulted into construction of pPM20-flhC. Amplified flhC::Km r fragment was digested and cloned into pDM4 prior to transformation into E. coli S17-λpir. Transformants were selected on Broomfield agar supplemented with kanamycin and chloramphenicol (Blomfield et al. 1991). The resulting suicidal plasmid pDK-flhC ( Fig. 1) was transformed into E. coli USML2 and plated on Broomfield agar supplemented with the kanamycin. Selected purified colonies were then grown overnight in broth containing kanamycin to allow plasmid-chromosome recombination to occur at the homologous site. Following this, selected colonies were tested for sucrose sensitivity to screen for mutant strains. PCR primers designed from DNA sequences located outside the flhDC gene were then used. Purified PCR products were sequenced by MyTACG Bioscience Enterprise, Inc. (Malaysia) according to the Sanger sequencing procedures. The mutant strain of E. coli USML2 flhC::Km r was non-motile in contrast with the wild type strain and no flagella formation was observed.

Inoculum Preparation
Wild type strain of E. coli USML2 as well as a newly constructed E. coli USML2 flhC::Km r mutant (impaired flagella biogenesis) and Azospirillum brasilense Sp7 (ATCC 29145) (a plant growth promoting rhizosphere bacterium) (Tarrand et al. 1978) were used in this investigation. All bacterial strains tested were grown in 100 mL nutrient broth (NB)(Oxoid TM ) in a 250 mL Erlenmeyer flask and incubated at 37°C on an orbital shaker (180 rpm) for 24 h. The bacterial strains were collected by centrifugation (4 500 xg, 15 min) and washed twice with normal saline (0.15M NaCl) for inoculum preparation. A concentration of the inoculum was then adjusted to 10 8 CFU/mL with normal saline, based on optical density at 600 nm and was confirmed by plate counting as described by Compant et al. (2005).

Tissues of Rice Seedlings Grown Under Gnotobiotic Conditions
Gnotobiotic 7-day-old rice seedlings were transferred aseptically into sterilised glass tubes (20 cm in height, 2.5 cm in diameter) containing 2 mL of halfstrength N-free Yoshida medium (Yoshida et al. 1976). A total of 1 mL bacterial (10 8 CFU/mL) culture was inoculated directly onto the roots using sterilised pipette tip (10 mL). Both inoculated (+ E. coli USML2 or + E. coli USML2 flhC::Km r mutant) and uninoculated rice seedlings were incubated under 12 h photoperiods for 24 h.
Successful rhizoplane and endophytic colonisation (internal tissues of roots, stems and leaves) of both wild type and mutant strains of E. coli USML2 of rice seedlings were verified via plate counting method (Muthukumarasamy et al. 2006) and PCR analysis of two important housekeeping genes (groEL and ftsZ) and marB. Primers for these genes were designed based on sequence available in Scientific Database for the Bacterium Escherichia coli K-12 MG1655 (EcoCyc). Prior to PCR amplification of respective genes, DNA extraction was performed using G-spin™ Total DNA Extraction Mini Kit (iNtRON Biotechnology, Inc, Korea) as per manufacturer's instruction. The PCR amplification was carried out using the following primers: groEL_F (5'-ATG GCA GCT AAA GAC GTA AAA TTC GGT AAC-3'), groEL_R (5'-TTA CAT CAT GCC GCC CAT GCC ACC-3'), ftsZ_F (5'-ATG TTT GAA CCA ATG GAA CTT ACC AAT G-3'), ftsZ_R (5'-TTA ATC AGC TTG CTT ACG CAG AAT GC-3'), marB_F (5'-ATG AAA CCA CTT TCA TCC ATA-3') and marB_R (5'CTA CAT AGC GTG TTG ATT ATA ATA G-3'). Terra™ PCR Direct Polymerase Mix (Clontech® Laboratories, Inc.) was used to directly amplify the selected genes which were extracted from 2.0 mm slices of root, stem and leaf of treated rice seedlings, according to manufacturer's recommendation. PCR amplifications were carried out with a Bio-Rad thermocycler (Hercules, California, USA) under the following conditions: an initial denaturation at 98°C for 2 min, 30 cycles of denaturation at 98°C for 10 s each, annealing at 55°C for 15 s and 30 cycles, extension at 68°C for 1 min, 30 cycles and final extension at 68°C for 5 min. The purified DNA was then sent for sequencing to a service provider, MyTACG Bioscience Enterprise (Selangor, Malaysia). The sequences obtained were compared to those in the GenBank database using the National Center for Biotechnology Information (NCBI) website (http://blast.ncbi.nlm.nih.gov/ Blast.cgi) via nucleotide-nucleotide Basic Local Alignment Search Tool (BLASTn) program (Altschul et al. 1990).
Visual confirmation of rhizoplane and endophytic colonisation of inoculum tested on plant samples were also viewed under scanning electron microscope (SEM) and transmission electron microscope (TEM) after 24 h of inoculation. For SEM observations, a total of 1 cm length of the main root was excised with a sterile razor blade. Samples were mounted on stubs for SEM, coated with aurum (SEM Coating System, SC515) and viewed under a FESEM Zeiss SUPRA 50VP (Lim et al. 2016). Additionally, endophytic colonisation of inoculum tested was also observed under transmission electron microscopy (TEM). Samples of 10 mm segment from respective surface sterilised root, stem and leaf of inoculated seedlings were prepared, fixed, stained according to Fan et al. (2011) and viewed under a TEM Philips CM12 (Carl Zeiss AG Oberkochen, Germany).

Growth of Rice Seedlings Inoculated with E. coli USML2 Under Glasshouse Condition
Rice seeds (Oryza sativa L. cv MR220) were surface sterilised according to Gyaneshwar et al. (2001) prior to germination. Surface-sterilised seeds were transferred onto 1% (w/v) sterile water agar (Microbiological agar Merck, Darmstadt, Germany) and left to germinate for 3 days in the dark at room temperature followed by 12 h photoperiod incubation until day 7. The gnotobiotic 7-days-old rice seedlings were transplanted into undrained pots (1 seedling per pot) containing 0.5 kg of sterilised river sand supplemented with half-strength N-free Yoshida medium (Yoshida et al. 1976). The seedlings were maintained in the plant house of School of Biological Sciences, USM under ambient conditions. Initially, the solution level (half-strength N-free Yoshida medium) was adjusted to slightly above the sand level. Bacterial suspensions of inoculation treatments were individually pipetted onto the sand surface area closed to the seedlings. At 3 days after treatment (DAT), the level of half-strength Yoshida medium was increased to 3 cm above the sand level and maintained throughout the experiment. The treatments applied for this experiment were: (1) + inorganic N (NH 4 NO 3 ) (40mg/L, dissolved in distilled water), (2) without inorganic N (distilled water only), (3) + E. coli USML2 (overnight grown culture in nutrient broth (NB), 5 mL/pot at 10 8 cfu/mL), (4) + E. coli USML2 flhC::Km r as control inoculum (overnight grown culture in nutrient broth (NB), 5 mL/pot at 10 8 cfu/mL), and (5) + A. brasilense Sp7 (overnight grown culture in nutrient broth (NB), 5 mL/pot at 10 8 cfu/mL).
Non-destructive assessment of plant growth parameters (shoot length, leaf numbers and chlorophyll content) were observed weekly until 42 DAT. The leaf greenness (SPAD values) of youngest fully expanded leaves for each seedling was recorded using a portable leaf-chlorophyll meter (MINOLTA TM SPAD-502) (Neufeld et al. 2006). The actual leaf chlorophyll content was determined based on the standard curve of SPAD values and total leaf chlorophyll content of the seedlings (mg chlorophyll mg -1 leaf fresh weight) (Amir et al. 2001). In addition, at day harvest (42 DAT), destructive growth assessment of the seedlings which include dry weight of shoot and root was executed.

STATISTICAL ANALYSIS
The experiments were laid out in a completely randomised design (CRD) with five replications for each treatment combination. Data for microbial colonisation and plant growth observations were statistically analysed by using SAS software (SAS Institute, North Carolina, USA). Following the analysis of variance procedure, differences among treatment means were determined according to Tukey-Kramer's Honestly Significant Difference (HSD) test, P < 0.05.

Endophytic Colonisation of E. coli USML2 in Association with Rice Seedlings
E. coli USML2 was able to colonise the rhizoplane and internal tissues of root, stem and leaf (endophytically) of the host plants grown under aseptic conditions (Table 1). Viable cells of E. coli USML2 increased from 12 h to 24 h after inoculation. At 12 h of incubation, viable cell count was higher in the internal root tissues (21.0 × 10 4 CFU/g -1 (wet wt)), followed by stem (0.6 × 10 4 CFU/g -1 (wet wt)) and leaf (0.02 × 10 4 CFU/g -1 (wet wt)). Similar results were observed at 24 h after inoculation.
However, for rice seedlings inoculated with mutant strain E. coli USML2 flhC::Km r , the colonisation was observed only on the rhizoplane and internal tissues of roots. Even though colonisation of the plant parts increased from 12 h to 24 h after inoculation, the viable cells were lower as compared to those in rice seedlings inoculated with wild type strain. At 12 h after inoculation, mutant strain exhibited 44.4% and 56.2% lower viable cell count on the rhizoplane and internal root tissues, respectively. Similar trends were observed at 24 h after inoculation which showed 79.6% and 45.3% lower viable cell count in rhizoplane and internal root tissues compared to the wild type strain. However, mutants failed to colonise the internal tissues of aerial plant parts (stem and leaves). Presence of wild type and mutant strains of E. coli USML2 was further verified via PCR amplification of groEL, ftsZ and marB genes in E. coli. Results for the verification study matched results from the viable cell count where it showed that wild type strain successfully colonised all plant parts tested. While, the mutant strain only colonised the rhizoplane and internal tissues of roots within 24 h. No gene was detected in the internal tissues of aerial plant parts showing that mutant strain failed to colonise the internal tissues (endopytically) of both stem and leaves.
Microscopic analysis under SEM showed that the rhizoplane was colonised by rod shaped bacteria in the form of microcolonies (Figs. 2A and 2B). Colonisation appeared to be more abundant at the base of the lateral root junctions and on root hairs ( Fig. 2A). Layers of exopolysaccharide were also observed on the rhizoplane (Fig. 2B). The SEM images also revealed presence of mutant strains on the rhizoplane zone (Fig. 2C). Based on the micrographs, presence of bacteria was not uniform on the entire root surface. Colonisation appeared to be more at the root cracks present on the root surfaces.
Further analysis under TEM clearly showed presence of bacterial cells in the internal tissues of roots, stems and leaves of the host plants as early as 12 h of inoculation (Fig. 3). The bacterial cells were also observed in all plant tissues after 24 h of inoculation. In the root tissues, localisation of E. coli USML2 was observed mainly within intracellular and intercellular spaces of the parenchyma cells. At this stage, the parenchyma cell walls of the inoculated roots were seemed disrupted or degraded (Fig. 3B) in contrast to the intact parenchyma cell wall of the uninoculated plant (Fig. 3A). Bacterial cells were observed in both intercellular and intracellular spaces of the stem tissues (Fig. 3D). In the  128 leaf tissues, E. coli USML2 cells were localised only in the intracellular spaces (Fig. 3F). Besides that, proteins were observed in intracellular spaces of the stem tissues (Fig. 3D). These proteins are probably plant cell wall degrading enzymes (PCWDEs) secreted by E. coli USML2. Additionally, TEM images also showed the presence of bacteria in internal root tissues inoculated with mutant strains.
Besides that, structure of the cell wall was degraded probably due to PCWDEs produced by bacteria present in planta (Fig. 3G). As expected, no bacteria were observed using TEM of internal tissues of stem and leaf. This finding supported the results from the bacterial enumeration study (Table 2)

Plant Growth Promotion of Endophytic E. coli USML2 in Association with Rice Seedlings
The results showed that seedlings inoculated with E. coli USML2 and A. brasilense Sp7 showed equal good growth response (higher leaf number, shoot length and chlorophyll content) compared to the control (-N) (Fig. 4). Interestingly, seedlings inoculated with E. coli USML2 and A. brasilense Sp7 also showed significantly higher root and shoot dry weight as compared to control treatment (-N) at harvest (D 42 ). The results clearly showed that E. coli USML2 exhibited similar plant growth promoting effects as established by A. brasilense Sp7, a known diazotrophic plant growth enhancer. The seedlings inoculated with the mutant strain (E. coli USML2 flhC::Km r ) exhibited significantly poor growth compared to the wild type after 42 DAT (Fig. 4). Those inoculated with mutant strain exhibited significantly lower leaf numbers (16.7%), chlorophyll content (37.4%), shoot height (14.5%) and plant dry weight (59.3%) compared to the wild type strain of E. coli USML2.

Predicted Genes in Promoting Endophytic Colonization and Plant Growth Enhancement of E. coli USML2
Annotation generated by RAST server exhibited presence of subsystems related to motility and possible essential steps involved in endophytic bacterial colonisation. Based on the annotation, E. coli USML2 is equipped with genes predicted for its active movement towards the plant roots. This is shown by the presence of flagella biosynthesis genes in genome of E. coli USML2 which includes the following operons: flgLKJIHGFEDCBAMN, fliE, fliRQPONMLKJIHGFE, fliTSDCZY, motAB, flhBAE and flhDC (Table 3). The importance of flagella to promote plant endophytic colonisation was successfully validated via mutant strain E. coli USML2 flhC::Km r , with impaired flagella biogenesis. Additionally, bacterial movement was also due to chemotaxis activity which can be described by the presence of chemotaxis genes encoding CheA, CheW, CheR, CheM, CheY, CheZ and methyl-accepting chemotaxis protein. These che genes are found within the flagella biosynthesis gene cluster. Following movement of E. coli USML2 towards the root via motility driven chemotaxis, efficient adhesion on the root surface also plays an important role in successful endophytic colonisation. Non-specific root adhesion by endophytic bacteria usually occurs in the presence of pili. Several genes involved in biogenesis of type IV pili (pilABCMNOPQ) were present in E. coli USML2 genome. Besides that, E. coli USML2 also contains csg operon for biosynthesis of curly fibres which is involved in bacteria adhesion to surfaces. Once the bacterial cells are attached to the root surface, they form microcolonies. Bacterial cells within the microcolonies communicate to each other via quorum sensing. In E. coli USML2, genes required for quorum sensing are also present (luxR, luxS and sdiA). Communication within the microcolonies would then trigger production of extracellular polymeric substances (EPS). Genes predicted for encoding proteins involved in the production of EPS were also found in E. coli USML2. Week 2

E. coli USML2
A. brasilense Sp7 Week 2 Week 6 Week 5 Week 4 Week 3 genes discovered were phosphoglucomutase (pgm) and phosphomannomutase (manB). Presence of these genes predicted the ability of E. coli USML2 in EPS production which is essential for firm bacterial anchoring on root surfaces. In addition, the predicted genes involved in biofilm formation such as biofilm regulator protein (bssR), outer membrane secretin (pgaA), synthesis of deacetylase (pgaB), synthesis of N-glycosyltransferase (pgaC) and synthesis of auxiliary protein (pgaD) are also present. Firm attachment on the root surfaces is a precondition towards invasion of the plant host. For this purpose, E. coli USML2 must be able to degrade the plant cell wall by production of PCWDEs. Since cellulose and pectin are the major composition of the plant cell wall, therefore, production of related enzymes for degradation of both components is essential. Several predicted genes involved in degradation of cellulose and pectin are present in E. coli USML2 genome. These include kdgK, kdgR and kdgT genes, which are involved in degradation of pectin. Besides that, the presence of genes encoding for cellulose synthase enzymes (yhjO, bcsABZC) is also crucial for breakdown of cellulose which further supports the capability of E. coli USML2 in invasion of the internal tissues for in planta spreading of its host plant.
As for plant growth promoting potentials, like other non-nitrogen fixing plant growth promoting endophytes (PGPEs), E. coli USML2 is equipped with predicted genes which aid in nitrogen metabolism. It contains genes essential for dissimilatory and assimilatory nitrate reduction pathways. Among genes present include narIJHGXL, nirCBD and ntrABC. In addition, predicted genes involved in nitrate/nitrate transporter (narK) and sensor (narQ) were also found (Table 3). E. coli USML2 also has the capability to solubilise insoluble phosphate. Its capability in phosphate solubilisation is made possible by presence of genes encoding for acid and alkaline phosphatase enzymes (Table 3). These enzymes release phosphorus from organic compounds present in the soil which later will increase the availability of phosphorus. Furthermore, gene encoding pyrroloquinoline quinone (PQQ)-dependent glucose dehydrogenase (gcd) was also predicted ( Table 3). The gcd gene is essential in biosynthesis of gluconic acid (GA). In fact, secretion of GA by phosphate solubilising bacteria (PSB) is an important mechanism for mineral phosphate solubilisation in Gram-negative bacteria. This is because release of organic acid such as GA chelates the cation bound to phosphate, thus converts it into soluble forms. Additionally, E. coli USML2 genome also carries several genes involved in potassium solubilisation (Table 3). These include genes encoding trans-membrane protein (trkH and trkG), cytoplasmic membrane surface protein (trkA), and Kup system potassium uptake protein (trkD). Beside these, the genes for the two components transcriptional regulatory proteins of the potassium dihydrogen phosphate (KDP) operon (kdpE and kdpD) are also present.  narQ, narI, narJ, narH, narG, narK, narX, narL, nirC, nirD, nirB, ntrB, ntrC, ntrA Phosphate solubilisation Pyrroloquinoline quinone (PQQ)-dependent glucose dehydrogenase (gcd) Potassium solubilisation Trans-membrane protein (trkH and trkG), cytoplasmic membrane surface protein (trkA), Kup system potassium uptake protein (trkD), potassium dihydrogen phosphate (KDP) operon (kdpE and kdpD).

Siderophore production
Plant growth and morphogenesis are also affected by small molecules or volatile compounds released by microorganisms. These molecules act as phytohormones to activate plant immunity or regulate growth. The strain E. coli USML2 also harbours genes involved in the production of auxin and acetoin. The most common naturally occurring plant hormone of the auxin class is indole acetic acid (IAA). The genes for enzyme tryptophanase (tnaA) and indole-3glycerol phosphate synthase (trpC and trpF) are also present in the E. coli USML2 genome. The former is involved in reduction of tryptophan to indole, pyruvate and ammonia while the latter is involved in tryptophan-independent IAA biosynthesis pathway (Edwards & Yudkin 1984;Gaimster & Summers 2015;Panichkin et al. 2016). In addition to the genes related to IAA production, E. coli USML2 also has the capability to produce acetoin. This is indicated by presence of gene encoding pyruvate dehydrogenase (aceE) which converts a small fraction of pyruvate into acetoin.
Ability of E. coli USML2 in plant growth promotion is further supported by presence of acdS gene encoding for 1-aminocyclopropane-1-carboxylate (ACC) deaminase. This enzyme degrades ACC to form ammonium and α-ketobutyrate which respectively serve as a source of nitrogen and carbon for the host plant. Earlier findings by Yunlei et al. (2015), observed essential roles of ACC deaminase (acdS gene) in enhancing bacterial heavy metal tolerance of bacteria and plant growth promotion ability. Additionally, S-adenosylmethionine (SAM), which is converted into ACC by the enzyme ACC synthase, is also present. Other than ACC, several genes required for biosynthesis of enterobactin which is a high affinity siderophore were found. The genes entBCDEFG which are required for the biosynthesis of enterobactin were found as an operon. Enterobactin is a high affinity siderophore and the integral inner membrane protein EntS is required to export after enterobactin synthesis occurred in the cytoplasm. The iron-enterobactin complex is recovered via the ferric siderophore uptake system (ExbDB) and when the complex enters the cytoplasm, iron is extracted by a mechanism requiring the product of the iron ferric enterobactin esterase (fes) gene. In addition, E. coli USML2 genome also encompasses genes for ferrous iron uptake systems, ferric hydroxamate ABC transporters and outer membrane ferric-related siderophore receptors (TonB-dependent). Genes required for ornithine biosynthesis which is involved in increasing the rate of siderophore production rate are also present. Genes for sulphur metabolism which includes genes for assimilation of inorganic sulphate such as sulphate and thiosulfate binding protein (cysP), sulphate transport system permease protein (cysT and cysW), sulphate or thiosulfate import ATPbinding protein (cysA) and several other cys genes (cysC, cysD, cysH, cysI, cysJ, cysG) were also predicted.

DISCUSSION
The results of this research showed that endophytic E. coli USML2 has unique characteristics as a PGPE that is able to promote growth of the host plant through associative interactions. This is evidenced by the discovery of predicted PGP genes of E. coli USML2 in relation to traits of nitrogen metabolism, phosphate solubilisation, potassium solubilisation, auxin (Indole) production, acetoin production, ACCD production, siderophore production, ornithine biosynthesis and sulphur metabolism. In addition, predicted genes related to promotion of endophytic bacterial colonisation traits such as flagella biogenesis, chemotaxis type IV, pili biogenesis, curly fibres biogenesis, quorum sensing, EPS production, biofilm formation and degradation of cellulose and pectin (PCWDE) has also been found. Results also successfully highlighted endophytic E. coli USML2 colonisation of the internal tissues of root, stem and leaf, together with rhizoplane colonisation of the host plants. The inoulation treatment had subsequently promoted plant growth. However, for seedlings inoculated with mutant strain of E. coli USML2 flhC::Km r , no endophytic colonisation in the aerial tissues together with poor growth of rice seedlings were recorded.
The results clearly showed that the initial step of plant colonisation involves bacterial motility. Earlier findings by Turnbull et al. (2001) stated that motility is the initial phase of bacterial colonisation which finally contributed towards bacterial survival and attachment in soil particles and roots of the host plant. In this study, a mutant (E. coli USML2 flhC::Km r ) deficient in flagella biogenesis established the importance of flagella in endophytic colonisation of endophytes within the host plant tissues. Similarly, Hockett et al. (2013) showed that Pseudomonas syringae utilised flagellum-mediated motility to reside the phyllosphere and invade the leaf internals to survive as a pathogen. Genes involved in the biogenesis of flagella and its function consist of a highly complex regulon of at least 50 genes in 17 operons has also been widely investigated (León & Espín 2008;Lee et al. 2011;Fan & Macnab 1996). Plants treated with flagella mutant strain had significantly lower growth compared to the wild type strain indicating the importance of bacterial motility towards growth enhancement abilities. The inability of the mutant strain to colonise endophytically in the internal tissues of the aerial plant parts of rice, had affected its effectiveness as a growth promoter. Induced expression of the chemotaxis genes enabled the bacteria to swim to the secreted root exudates with the aid of flagella (Hardoim et al. 2008), resulted in bacterial attachment to the root surface followed by formation of microcolonies (Taghavi et al. 2010). In this study, efficient rhizoplane colonisation occurred on the root surface especially at the base of the lateral root junctions after 24 h of inoculation. This is probably because root exudates secreted from wounds at root junctions would provide nutrient source for the colonising bacteria and thus create a favourable condition for colonisation. Complex interactions between the bacterial cells and the host plant also resulted in the production of extracellular polymeric substances (EPS) by the microcolonies (Taghavi et al. 2010). This process of biofilm formation would also increase bacterial survivability by providing shelter to environmental stress as well as protection from adverse substances such as antibiotics and biocides (Annous et al. 2009). The ability of the colonies to produce EPS was also evident before the endophytic bacteria can finally enter the plant (Taghavi et al. 2010). Following successful colonisation on the rhizoplane and entry into the root internal tissues, E. coli USML2 had spread towards the inner parts of stem and leaves and finally promoted growth of the host plants. This serves as a beneficial strategy for the bacterium to gain protection against adverse conditions present outside the plant (Méric et al. 2013) and any potential competition against indigenous microorganism.
Following establishment of endophytic colonisation within the rice seedlings, locally isolated E. coli USML2 had successfully promoted growth and development of the host plants probably due to the expression of the PGP genes. Inoculation of E. coli USML2 and A. brasilense Sp7 showed equal good growth response of the host plants compared to the mutant E. coli USML2 flhC::Km r and control (-N). Wild type E. coli USML2 has plant growth promoting traits such as phytohormone and ACC deaminase production. Besides that, gene for siderophore, an iron-chelating complex that makes the metal available to the plant is also present. In addition, E.coli USML2 is also capable in solubilisation of phosphate and potassium. These traits stimulate plant growth either by synthesising phytohormones or by promoting nutrients availability (Liu et al. 1992;Glick 1995;Bowen & Rovira 1999;Ashrafuzzaman et al. 2009). Discovery of numerous predicted PGP genes in this study nominates E. coli USML2 as a potential endophytic plant growth promoting candidate. The mutant without flagella failed to spread into the phyllosphere, thus reduced its PGP ability. Similar results were shown earlier by Nautiyal et al. (2010) where inoculation of maize seeds with a soil originated E. coli NBRIAR3 under microplot conditions had significantly enhanced plant growth when compared to uninoculated control. However, no information on PGP genes of the isolate was highlighted in the findings.
Other findings have also successfully recorded an increase in the number of isolated PGPEs, which has the potential to enhance growth of the host plants. Such PGPEs include species of bacteria from the genera of Pseudomonas, Azospirillum, Azotobacter, Klebsiella, Enterobacter, Alcaligenes, Arthrobacter, Burkholderia, Bacillus and Serratia (Kloepper et al. 1989;Okon & Labandera-Gonzalez 1994;Glick 1995;Gururani et al. 2012). In fact, PGPE such as Pseudomonas, Bacillus, Klebsiella, Azotobacter, Azospirillum and Azomonas have been widely used as bio-inoculants to promote plant growth and development of agriculture crops (Ahemad & Kibret 2014). Application of PGPEs as inoculants for improving plant growth and yield offers an attractive way to reduce and replace chemical fertilizers, pesticides, and supplements (Stefan et al. 2008;Ashrafuzzaman et al. 2009). This extends to decrease environmental pollution and health risks originating from excessive use of the agrochemical nitrogen (N) fertilizers to achieve enhanced plant growth (Tambalo et al. 2015).

CONCLUSION
Annotation of E. coli USML2 genome sequence revealed presence of genes involved in bacterial motility towards the root, root adhesion, endophytic colonisation, and plant growth promotion. The ability of E. coli USML2 in plant growth promotion of rice seedlings and in planta spreading was demonstrated. The study also revealed potential of E. coli USML2 as a genetically amenable endophyte that can be employed in a sustainable agriculture practice. A further comprehensive transcriptomic and proteomic studies would be able to shed valuable insights towards understanding the complex regulatory pathways involved in plant growth promotion.