First report of diazotrophic Brevundimonas spp. as growth enhancer and root colonizer of potato

Rhizobacteria contain various plant-beneficial traits and their inoculation can sustainably increase crop yield and productivity. The present study describes the growth-promoting potential of Brevundimonas spp. isolated from rhizospheric soil of potato from Sahiwal, Pakistan. Four different putative strains TN37, TN39, TN40, and TN44 were isolated by enrichment on nitrogen-free malate medium and identified as Brevundimonas spp. based on their morphology, 16S rRNA gene sequence, and phylogenetic analyses. All strains contained nifH gene except TN39 and exhibited nitrogen fixation potential through acetylene reduction assay (ARA) except TN40. Among all, the Brevundimonas sp. TN37 showed maximum ARA and phosphate solubilization potential but none of them exhibited the ability to produce indole acetic acid. Root colonization studies using transmission electron microscopy and confocal laser scanning microscopy showed that Brevundimonas sp. TN37 was resident over the root surface of potato; forming sheets in the grooves in the rhizoplane. TN37, being the best among all was further evaluated in pot experiment using potato cultivar Kuroda in sterilized sand. Results showed that Brevundimonas sp. TN37 increased growth parameters and nitrogen uptake as compared to non-inoculated controls. Based on the results obtained in this study, it can be suggested that Brevundimonas spp. (especially TN37) possess the potential to improve potato growth and stimulate nitrogen uptake. This study is the first report of Brevundimonas spp. as an effective PGPR in potato.

Morphological characteristics of bacteria and 16S rRNA based identification. Four bacterial strains were obtained on semi-solid NFM medium from enrichment cultures of soil samples based on their shape, size, colony and color. All the isolates were Gram-negative, medium-sized short rods and showed round shaped colonies expect TN44, which showed irregularly shaped colonies. The isolates showed variable colony colors on LB. Isolate TN37 formed pale yellow, TN39 formed creamy, TN40 formed white while TN44 formed grayish colored colonies. The Table 2 summarizes the details of cell morphology, colony morphology and gram's reaction of these bacterial strains. In PM2A, BIOLOG microplate assay, only 9 of 95 different carbon sources were utilized by bacterial isolateTN37 showing its limited metabolic potential associated with potato roots (Supplementary Table S1).
Based on 16S rRNA sequence analysis, the bacterial isolates were identified as Brevundimonas spp., showing 99% similarity with gene sequences of different species of the genus Brevundimonas in the Genbank (Table 2). www.nature.com/scientificreports/ The sequences of 16S rRNA have been submitted in the nucleotide database of Genbank under the accession numbers LN833470, LN833471, LN833472 and LN833475. For phylogenetic analysis, neighbor-joining method was adopted with bootstrap values greater than 50 (Fig. 1). Isolate TN37 was clustered with B. naejangsanensis T (FJ544245), TN39 clustered with B. terrae T (DQ335215) while isolate TN40 and TN44 were branched with B. vancanneytii T (AJ227779) and B. naejangsanensis T (FJ544245) (Fig. 1).
NifH gene analysis and nitrogenase activity. For evaluating nitrogen-fixing potential of these isolates, genomic DNA was extracted and the nifH gene was amplified. Of four, three isolates TN37, TN40 and TN44 showed amplification of nifH gene with a product size of 300 bp while no amplification of nifH gene was observed in TN39 isolate. The 300 bp fragment was sequenced and obtained sequences were compared with already submitted sequences in the GenBank database. The BLAST and phylogenetic analyses showed that nifH gene sequence amplified from Brevundimonas sp. strain TN37 was highly similar to nifH of A. brasilense Gr42 (Acc. No. FR669137) as it made cluster with Azospirillum sp. and subcluster with A. brasilense. However, Brevundimonas spp. strains TN40 and TN44 showed 99% similarity to the nifH gene sequence of uncultured bacterium clone OTU-31 (Acc. No. KF541088) (Table 3). Though, both strains showed similarity to the same nifH gene in data bank but in phylogenetic analysis they made subcluster at 99% with each other (Fig. 2). The nifH genes of Brevundimonas spp. strains TN37, TN40 and TN44 were submitted to the database under accession numbers: LT596596, LT596597 and LT596598, respectively. The nifH gene presence shows that these isolates may have the potential of nitrogen fixation which was later validated by in vitro assay. The isolate TN37, TN39 and TN44 showed nitrogenase activity measured by ARA. Isolate Brevundimonas sp. TN37 showed the highest activity (135 nmoles mg −1 protein h −1 ) while we could not detect any ARA activity by Brevundimonas sp. strain TN40 (Table 2).
IAA production and phosphate solubilization. The results of Salkowski assay showed that bacterial strains were unable to produce IAA either in the absence or presence of l-tryptophan. All bacterial strains except TN40 showed the ability to solubilize phosphorous in the growth media. Out of which strain TN37 showed the maximum P-solubilization potential and solubilized 306.25 µg mL −1 of phosphorous while TN39 and TN44 solubilized 249.37 µg mL −1 and 272.50 µg mL −1 of phosphorous in the growth medium after 12 days with the consequent decrease in pH up to 5.5, 5.98 and 5.51, respectively (Fig. 3A,B).
Rhizosphere survival, root colonization and plant growth-promoting potential of Brevundimonas sp. TN37. Being the best among four bacteria, rhizosphere survival, root colonization and plant inoculation potential of Brevundimonas sp. TN37 was checked in vitro and in vivo. The bacterial population of strain TN37 changed dynamically in potato rhizosphere at different stages of plant growth. At day zero (1st day of inoculation), the colony forming units of Brevundimonas sp. TN37 was 7.57, which decreased gradually upto 5.56 CFU on the final harvesting (Day 60 of inoculation) (Fig. 6E). In uninoculated control plants, no bacterial cells were observed in their rhizospheric region. Electron microscopic analysis confirmed the CFU data. Observation of ultrathin sections of potato root after 30 days post-inoculation under TEM showed that Brevundimonas sp. TN37 was an inhabitant of the root surface (i.e., rhizoplane) (Fig. 4). The bacterium formed sheets in the grooves, formed by the root cells and was closely attached to the cell wall of plant ( Fig. 5A-D). No bacterial cells were observed in or on the root surface or in rhizosphere of uninoculated control plants (Fig. 5E). Confocal microscopic analysis further validated the root colonization potential of Brevundimonas sp. TN37. The potato root inoculated with YFP-labelled Brevundimonas sp. TN37, in sand culture 25 days post-inoculation showed that primary root tips and root hairs of potato were preferred sites of colonization by Brevundimonas sp. TN37. Macro-colonies of bacterial aggregates were observed on the root epidermal cells. The microscopic observations support the root colonization potential and rhizosphere competence of Brevundimonas sp. strain TN37 in potato (Fig. 5).
The potential of Brevundimonas sp. TN37 isolate in promoting plant growth was evaluated by inoculating potato plants in pots. The data showed a positive effect on biomass and nitrogen content of potato plant when inoculated with Brevundimonas sp. TN37. The effect of Brevundimonas sp. TN37 inoculation on shoot and root www.nature.com/scientificreports/  www.nature.com/scientificreports/ fresh weight, total nitrogen contents of shoot, and root length were significantly higher than negative control ( Fig. 6A-D). This increase over non-inoculated (negative) control can be attributed to biological nitrogen fixation potential of the strain.

Discussion
Nitrogen is one of the most important nutrients for potato growth because both the deficiency and the excess of N may hamper the crop cycle and yield 25 . Excessive N-fertilization results in deterioration of ecosystem and global environment. The beneficial bacteria residing in the rhizosphere can minimize the fertilizer inputs in to agriculture along with reduced cost of crop production and environmental pollution. Present study describes the potential of a nitrogen-fixing potato plant-beneficial bacterium Brevundimonas sp. TN37 for improving growth and nitrogen uptake under limited-nitrogen fertilizer. The soil sample from two different locations of Sahiwal which is a major potato growing region showed higher organic matter, nitrogen, phosphorous and pH (Table 1). Higher nitrogen and phosphorous contents shows that the soils receive excessive chemical fertilizers which ultimately degrade the soil physical and biological functioning. Characteristics of soil directly influence microbial diversity, plant-microbe interaction, roots development and plant growth [26][27][28] . Four different Brevundimonas spp. strains were isolated on NFM and further confirmed by phylogenetic analysis.
An important property of rhizobacteria is their nitrogen-fixing ability which directly improves plant growth because N is an essential macronutrient required for plant growth and development. In the present study, isolated strains Brevundimonas spp. TN37, TN39 and TN44 showed ARA activity, suggesting their ability to fix nitrogen. Furthermore, these strains showed the presence of nifH gene in their genomes. Brevundimonas sp. TN39 strain, although, showed ARA activity but did not show amplification of nifH gene with the reported universal primers. It is already reported that these universal primers do not amplify nifH gene in several strains of diazotrophs 29 . Similar discrepancies have been reported for ARA results and nifH gene profiles 30,31 . To get more clarity in such cases, genome sequencing can help to elucidate the disparity. The amplification of nifH gene in other three Brevundimonas spp. shows that the strains contain the gene and have potential for nitrogen fixation. The sequence-similarity of nifH gene of Brevundimonas sp. TN37 with Azospirillum sp. shows that the gene may be acquired from that genus through horizontal gene transfer or from a common ancestor during evolution. Higher ARA activity of Brevundimonas sp. strain TN37 might be due to the sequence similarity with Azospirillum because this genus is a well-known and widely used bacterial diazotroph 32 . On the other hand, the nifH genes of the other two Brevundimonas spp. TN40 and TN44 may represent novel genes as they do not have any cultured homologs in the database. NifH is evolutionarily conserved and could also be used for the identification of nitrogen-fixing rhizobacteria 9 but it cannot be used for specie delineation. ARA is a widely accepted test for evaluating the nitrogen-fixing potential and nitrogenase activity 33 but in present study, Brevundimonas sp. TN40 did not show any ARA activity although the strain showed the amplification of nifH gene. The reason for this may be the invitro conditions which might not be suitable for this strain to perform its nitrogen fixing activity. Study also reported that nitrogenase activity is affected due to culture medium, growth stages and conditions or possibly in some strains the activity is active only in planta 34 . Different studies reported the transfer of nitrogen between roots of different crops and nitrogen-fixing bacteria [35][36][37] . www.nature.com/scientificreports/ The evaluation of other plant growth promoting traits revealed that all strains except TN40 were capable of solubilizing phosphorous (tri-calcium phosphate) of which Brevundimonas sp. TN37 showed the maximum potential of P-solubilization (306.25 µg mL −1 ) within 12 days. Plant growth promoting rhizobacteria have the potential of P-solubilization convert the inaccessible soil P into plant-available P-form that are taken up as essential nutrient source for plant growth and development 38 . The bacteria did not show the ability to produce IAA. Although IAA-production ability is widespread among rhizobacteria but its production depends on several factors e.g., the potential of isolates to utilize tryptophan, sampling location, pH, oxygen availability, carbon conditions 39 , and the incubation time 40 . Longer incubation time causes decrease in nutrients found in growth media and possibly the amount of IAA produced consumed again by bacterial cells for their own growth. Thus, it can be suggested that may be due to one of these reasons isolates were unable to produce IAA.
Rhizosphere competence and root colonization is an important feature of any candidate strain for biofertilizer production. Brevundimonas sp. TN37 was found to be a root-colonizer of potato forming strong root-associations. The bacterial cell density was high in the rhizosphere, where root hairs were preferred sites for the early colonization. Root hairs are identified as sites of increased rhizodeposition and are supposed to be involved in specific attachment and eliciting chemotaxis response [40][41][42] . Micro and macro colonies were also observed in the junctions of primary and secondary roots which may be because it provides a better niche, support for attachment and allow nitrogen fixation to take place. Cells of Brevundimonas sp. TN37 were distributed all over the root zone which is probably due to variation in the root exudates at different places within the root system 43 . www.nature.com/scientificreports/ Being a P-solubilizing diazotroph and root colonizer Brevundimonas sp. TN37 was evaluated for its growthpromoting potential in potato. The increase in biomass and high nitrogen contents due to the inoculation of Brevundimonas sp. TN37 can be attributed to diazotrophic ability of this bacterium. The study conducted in canola 3 plant showed increased biomass when inoculated with different bacterial strains as Achromobacter, Chryseobacterium, Pantoea, Pseudomonas and Klebsiella. Similar findings were reported in wheat and Bt-cotton when inoculated with PGPR 22,44 . Thus, it can be suggested that nitrogen plays an important role in growth promotion because its deficiency influences the yield and growth of crops. This study is the first report of Brevundimonas spp. as PGPR in potato crop.

Conclusion
This study demonstrates the isolation, screening and molecular identification of Brevundimonas spp. and its role in growth promotion. Potato requires high inputs of N and P fertilizers; therefore, the presence of such P-solubilizing, nitrogen-fixing bacterium will help to increase its biomass by making nitrogen and phosphorus available from the atmosphere or rhizosphere. Our findings imply that Brevundimonas sp. TN37 has the potential of nitrogen fixation as well as P-solubilization that helps to improve plant growth and can maintain soil fertility.

Materials and methods
Soil samples from the rhizosphere of potato plants were collected in plastic bags (25 × 30 cm) from two different sites of Sahiwal, Punjab, Pakistan. Samples were kept on dry ice for further experimentation and transferred to the laboratory. All the experiments were repeated three times for validation of results. The bacterial population of rhizospheric soil was determined by the basic technique of serial dilution 45 . For isolation of nitrogen-fixing bacteria from the rhizosphere, soil with roots (0.1 g) was added to 1.5 mL Eppendorf tubes containing NFM semi-solid medium 46 and were incubated at 28 ± 2 °C for 48 h. Culture from the semi-solid medium was then streaked on NFM agar plates and LB agar plates. Single and purified colonies having different morphological characters were obtained and grown for 24 h at 28 ± 2 °C on respective media and were preserved in glycerol (20%) at − 80 °C.
Physio-chemical analysis of soil samples. The physio-chemical properties of soil samples were evaluated after drying at 40 °C for 24 h followed by sieving through 2 mm mesh. For determination of soil texture, soil sample (50 g) was soaked overnight in 40 mL of sodium hexa-metaphosphate solution (1%) with 150 mL of distilled H 2 O (dH 2 O). Bouyoucous Hydrometer was used for recording preliminary reading after 40 s of stirring while final reading was obtained after 2 h. The international textural classification system was used to assign class to soil texture 47 . Soil paste was prepared for measuring pH. 250 g of soil and dH 2 O was allowed to stand for www.nature.com/scientificreports/ 1 h. Then the pH of this paste was measured by pH meter (JENCO Model-671P) 48 . The electrical conductivity of soil was recorded from the clear extract using EC meter (Jenway). The extract was obtained from soil paste using a vacuum pump using previously described protocol of Rhoades et al. 49 . Organic matter of soil was measured by making soil solution from 1 g soil sample, 1 N potassium dichromate (10 mL) solution, conc. H 2 SO 4 (20 mL), dH 2 O (150 mL) and 0.5 N FeSO 4 solutions (25 mL). This solution was then titrated with potassium permanganate solutions (0.1 N) which gives pink color as an endpoint 50 . To measure total soil nitrogen (N), 10 g of soil was digested with conc. H 2 SO 4 (30 mL) and 10 g of digestion mixture (CuSO 4 :FeSO 4 :K 2 SO 4 = 0.5:1:10) in Kjeldahl's digestion tubes 51 . This digested material was then cooled and an aliquot of 10 mL was used for distillation of ammonia from this, with boric acid solution (4%) and methyl red and boromocresol green as indicators in a receiver. In the distillation flask, NaOH solution was added to increase the pH of the contents. Gunning and Hibbard's method as used for titration of the material using N/10 H 2 SO 4 . Titration was done after distillation in the receiver with the micro Kjeldahl apparatus 51 . Extractable phosphorus was measured by the method described by Olsen using a spectrophotometer 52 . To determine the extractable K, soil solution (5 g) was prepared with ammonium acetate solution (1 N) and the volume was made up to 100 mL. After continuous shaking, the extractable K was measured on Flame Photometer from extracts, obtained from filtering suspension by Whatman filter paper No. 1 53 .
Morphological characterization. For colony morphology, bacterial isolates were streaked onto LB agar plates followed by 24 h incubation at 28 ± 2 °C. Light microscope (Nikon LABOPHOTO-2, Japan) was used to examine the shape and motility of bacterial isolates. For Gram staining, method described by Vincent was followed 54 .
Biochemical characterization. Azospirillum brasilense strain ER20 (Accession no. HE662867) 55 was obtained from the NBRC culture collection NIBGE, Faisalabad, Pakistan and was used as a reference strain or a positive control for in vitro analysis. Nitrogen-fixing ability of bacterial strains was evaluated using ARA (Acetylene Reduction Assay) 56 . Each isolated strain was grown for 72 h in NFM semi-solid medium at 28 ± 2 °C and their nitrogenase activity was evaluated by gas chromatograph fitted with flame ionization detector and Porapak N column 57 . Activity was measured as nmoles of ethylene produced per milligram per hour of the protein. The method described by Bradford 58 was adopted for the estimation of protein concentration. IAA production was measured by calorimetric method. The test was conducted in the presence or absence of IAA precursor (l-tryptophan). Purified single colonies of bacterial isolates were added in Erlenmeyer flasks (250 mL) containing LB broth (100 mL) supplemented with 100 mg L −1 of l-tryptophan. The culture was incubated for 48 h at 28 ± 2 °C with continuous shaking, following 15 min centrifugation at 4,000×g. The collected supernatant was filtered using nylon filters (0.2 µm) and this filtered supernatant (100 µL) was mixed with Salkowski reagent (100 mL) following 20 min incubation at room temperature. The IAA production was quantified using a standard curve against known concentration of IAA at λ = 540 nm 59 . The quantitative analysis of phosphate solubilization of isolated strains was done by inoculating them in Pikovskaya's broth (100 mL) in triplicate following incubation at 28 ± 2 °C in the orbital shaker for 288 h (12 days) at 150 rpm. 20 µL of bacterial culture was harvested from each flask at 72, 168, 240 and 288 h post inoculation, following 10 min centrifugation at 13,000×g. The phosphate solubilizing activity of supernatant was determined by phospho-molybdate blue color method with the help of a spectrophotometer at λ = 882 nm 60 .

Molecular characterization.
Sequence analysis of 16S rRNA and NifH genes. DNA of the pure culture was isolated using microbial DNA isolation kit (MoBio, CA). Following the manufacturer protocol, genomic DNA was obtained and stored at − 20 °C. The amplification of 16S rRNA of bacterial isolates was done by PCR using the primers 968F 61 and 1406R 62 ensuring the conditions described by Shahid et al. 11 . A nested PCR was conducted for the amplification of nifH gene from the pure cultures 63 using primer set FGPH19 64 and PolR 65 . The 2nd PCR was done by using primers set of PolF and AQER and product of 1st PCR as the template 65 . The products of bacterial isolates and nifH gene were analyzed on ethidium bromide stained agarose TAE gel with 1 kb ladder (Fermentas, Germany). Following the manufacturer's instructions, PCR products were purified using PCR Clean-up System (Promega) and Wizard SV Gel and sent for sequencing to LGC Genomics, Berlin. The sequenced products of nifH gene and 16S rRNA gene sequences were analyzed using a sequence scanner software package.
Phylogenetic analysis. For phylogenetic analysis, MEGA6 software package was used. The sequences of isolated strains were analyzed and compared by alignment tool CLUSTAL W, using the downloaded closely related sequences from the NCBI database 66 . Phylogenetic analysis was performed using neighbor-joining method, but the bootstrap values of 70% or greater were retained for representing well-supported nodes 67 .
Phenotypic microarrays. BIOLOG micro-plates were used to evaluate the metabolic potential of bacterial isolates 68 . Strains were grown for 48 h on LB agar plates at 28 ± 2 °C. Culture was collected in Eppendorf tubes containing DEPC water (1 mL) and starved for 3 h. After that, culture was mixed with inoculation fluid IF-0 and redox indicator according to the manufacturer's instruction. In 96 wells containing carbon utilization plate (PM2A), 100 µL of culture was added following incubation for 24 h at 28 ± 2 °C and were detected on VERSA max micro-plate reader (Molecular Devices, USA) having softmax pro-software for both qualitative and quantitative evaluation 69  Plant inoculation assay. The bacterial strain was grown in 250 mL Erlenmeyer flasks containing 100 mL of LB broth and was constantly stirred at 180 rpm at 28 ± 2 °C up to 10 9 CFU mL −1 . Centrifugation at 8,000×g was done for harvesting of bacterial cultures and washed by 0.89% saline solution. After washing, the solution was again centrifuged for pelleting and the pellet was re-suspended in an equal amount of 0.89% saline solution. For evaluation of growth-promoting activity of isolates, a pot experiment under control conditions was conducted: 25 °C temperature and light and dark period of 16/8 with 400 mol m −2 S −1 photon flux density in completely randomized design (CRD) with 3 replicates of each treatment in sterilized sand. Sand was soaked for 24 h in 0.5 N Nitric acid for sterilization then washed with dH 2 O to remove acid which was further air-dried and autoclaved. Medium-sized potato (variety Kuroda) tubers (2-3 cm) were surface sterilized with sodium hypochlorite (8% v/v) for 10 min followed by extensive washing with sterilized dH 2 O. Sterilized potato tubers were immersed in bacterial inoculum for 30 min, dried for 10 min in Laminar Air flow cabinet and then sown in pots containing sterilized sand. The 2nd dose of bacterial inoculum was applied to potato roots after 7 days of sowing. The bacterial inoculum was mixed in Hoagland solution (without nitrogen source) and applied with sterilized syringe @ 5 mL/plant. There were 3 treatments with 3 replicates each, T1: un-inoculated control treatments; positive control (recommended full dose of N F ), T2: negative control; without nitrogen (N 0 ) and T3: inoculated treatment (Inoculated with bacterial isolate TN37 + N 0 = without nitrogen). Growth parameters like shoot and root length, shoot and root fresh weight and dry weight and N contents of plants was recorded after 60 days of sowing.
Statistical analysis. Analysis of variance (ANOVA) for different plant parameters was analyzed using STATISTIX8.1 software (Tallahassee, FL, USA) 74 . For comparing means of each treatment, the least significant difference test (LSD) was performed with 5% probability.