Diazotrophic bacterium Azotobacter vinelandii as a mutualistic growth promoter of an aquatic plant: Lemna minor

Lemnaceae plants, commonly referred to as duckweeds, are small planktonic terrestrial freshwater plants that live in symbiosis with various microbial communities. Azotobacter vinelandii are typical free-living nitrogen fixing soil bacteria that indirectly benefit plants by providing nitrogen compounds. In this study, Lemna minor RDSC 5512 and A. vinelandii ATCC 12837 = NBRC 13581 were co-cultured under gnotobiotic conditions. The growth of L. minor colonized by A. vinelandii was accelerated in both nitrogen-containing and nitrogen-free water conditions. The growth promotion effect is attributed to several plant growth promotion factors produced by the bacterium as well as biological nitrogen fixation in nitrogen-free condition. Moreover, L. minor elevated the nitrogen fixing activity of A. vinelandii and the cell number of A. vinelandii on L. minor increased continuously over 30 d. These observations indicated that L. minor provides a favorable environment for A. vinelandii colonization, allowing them to mutually benefit and flourish through syntrophism.


Introduction
Duckweed, family Lemnaceae, is a small floating aquatic plant, capable of growing ubiquitously and rapidly absorbing nutrient minerals from water under various climate conditions. Therefore, duckweed is regarded as an effective tool for energy saving wastewater treatment system (Bonomo et al. 1997;Yamaga et al. 2010) and subsequent biomass production (Xu and Shen 2011). Duckweed biomass has been used as an animal feed and as human food (Leng 1999) owing to its high protein content and other nutritional values (Appenroth et al. 2017). When duckweed is grown under stress conditions, its protein production level decreases and high amounts of starch accumulate, making it applicable as of duckweeds by bacterial strains are now being studied in detail (Gilbert et al. 2018;Idris et al. 2007;Ishizawa et al. 2017a;Toyama et al. 2022;Utami et al. 2018). Nitrogen is an essential minerals for plant growth. Free-living nitrogen fixing bacteria such as cyanobacteria (Duong and Tiedje 1985), Klebsiella, and unclassified aerobic diazotrophs associate with duckweed mats in ponds, where they possibly provide approximately 15-20% of the duckweed nitrogen requirement through biological nitrogen fixation (Zuberer 1982). Azotobacter is a dominant group of free-living soil diazotrophs that has been found to promote the growth and yield of terrestrial plants through non-symbiotic dinitrogen fixation (Sprent and Sprent 1990), phytohormones, cytokinin production (Abbass and Okon 1993;Taller & Wong 1989) , and exopolysaccharides production (Giti et al. 2004;Vermani et al. 1997). The beneficial effect of the phosphate solubilizing activity of Azotobacter bacteria on agriculture has also been reported (Kumar and Narula 1999). Compared to inorganic fertilizers, the inoculation of Azotobacter bacteria showed 23% more grain productivity in maize (Mahato and Kafle 2018). Considering that duckweed is a group of terrestrial plants, we sought to determine whether Azotobacter vinelandii could co-exist and affect the growth of gnotobiotic L. minor in aquatic conditions. A mutually beneficial symbiotic association between host plant L. minor and guest bacterium A. vinelandii is expected to bolster the growth of both organisms and increase duckweed biomass yield in extremely nutrient deficient situations such as nitrogen-and carbon-free water conditions, leading to the expansion of application sites for duckweed biomass production. We previously reported that Chryseobacterium sp. 27AL, indigenous to the low nitrogen food factory wastewater, promoted the growth of duckweed without competition of nitrogen minerals, Lemna gibba (Khairina et al. 2021). In contrast, A. calcoaceticus P23 originating from botanic garden pond water inhibited duckweed growth in the low nitrogen wastewater condition owing to deprivation of limited inorganic nitrogen from the host plant. This study reports the potential of free-living diazotrophic bacterium A. vinelandii to form mutualistic symbiosis and proliferate the host plant even under nitrogen-free water conditions.

Culture conditions
All the bacteria were stored at -80°C in cryotube. Liquid and solid agar of Luria (L) broth (Suzuki et al. 2014) were used for culturing A. calcoaceticus P23, P. fulva Ps6, Ensifer sp. SP4, and Burk's nitrogen-free (BS) medium (Strandberg and Wilson 1968) was used for A. vinelandii A81. A. vinelandii CA12 was grown in BS medium supplemented with 2.25 g/L ammonium acetate (CH 3 COONH 4 ) for nitrogen source. Bacterial fresh culture was prepared each time using a shaking incubator at 30˚C for 2 or 3 d depending on the growth.

Plant culture conditions
L. minor RDSC 5512 was previously sterilized via sodium hypochlorite treatment and aseptically maintained in the laboratory (Suzuki et al. 2014). The culture condition of L. minor was 28˚C, 60% relative humidity, 5,000 lx (75 µmol m − 2 s − 1 ) illumination, and 16 h-photoperiod in modified Hoagland, mH, medium (H1, Suzuki et al. 2014). Other duckweed species: Lemna gibba G3 RSDC 362, Wolffiella hyalina (provided by Graduate School of Science, Kyoto University, Oyama laboratory), and S. polyrhiza (provided by Graduate Faculty of Interdisciplinary Research, University of Yamanashi, Toyama Laboratory) were stored in an aseptic stock and cultured in the same condition. The sterility of duckweed was confirmed by the lack of bacterial colony on L agar plate after incubation at 30˚C for 3 d.

Plant growth promotion assay
Two different co-culture conditions were used in this experiment. For the "suspension experiment", fresh bacterial culture was prepared and centrifuged at 7,700 × g for 10 min at 4˚C to retrieve the bacterial cells as pellet. The pellet was washed using the mH medium, re-suspended, and diluted to make a uniform bacterial cells suspension of 0.3 OD 600 (approximately one million cells) in 50 ml mH medium, where two plant bodies (two fronds with two roots) from aseptic L. minor stock were placed and co-cultured under plant growth condition. For the "attachment experiment", two L. minor plants were cultured in bacterial cell suspension as described above for 48 h, then rinsed by submersion in sterilized distilled water twice, and introduced into bacterial free 50 ml mH medium for co-culture. Similar experiments were conducted in 50 ml nitrogen-free mH-N medium, where KNO 3 was replaced with K 2 SO 4 and BS medium when necessary. All the experiments were performed thrice (n = 3) for statistical analysis. The fronds, leaf like structures, were counted every 2 d until 10 d and the final colony forming unit (CFU) values of bacterial cells and dry weight of plants were determined.

Measurement of protein content
L. minor was harvested and fresh weight was measured before vacuum freeze drying (FDU 1110, EYELA, Tokyo, Japan). Protein fraction was prepared using Apro science protein extraction kit (Naruto, Japan) and quantified using the D C protein assay kit (Bio-Rad, Hercules, CA). Protein content, % protein/dry weight, was estimated using the standard curve plotted from different concentrations of BSA (2 mg/ml to 0.125 mg/ml).

Measurement of starch content
L. minor was harvested and fresh weight was measured before vacuum freeze-drying. The freeze-dried biomass was used for measuring the starch content, % starch/ dry weight, using Megazyme total starch assay kit (NEOGEN, Lansing, MI)

Biofilm formation assay
A 400 µL L medium was inoculated with 1% preculture in 1.5 mL polypropylene tube and left standing at 30°C for up to 5 d. After measuring the OD 600 , planktonic cells in liquid culture were carefully removed using a micropipette. After rinsing the wells with 500 µL MilliQ water, 600 µL of 0.1% crystal violet (CV) solution was added and kept in the dark for 20 min at 25˚C. CV solution was removed and rinsed with MilliQ water three times. CV dye bound with biofilm was extracted by 600 µL of 33% acetic acid. The amount of biofilm was estimated by measuring the OD 535 of CV.

Indole acetic acid (IAA) production assay
Bacterial fresh culture was prepared using a shaking incubator for 24 h at 30°C in 5 ml of liquid L medium in glass test tubes in the presence and absence of 200 mg/L tryptophan for IAA production assay (Gordon and Weber 1951). Bacterial culture was centrifuged at 4,000 x g at 4˚C for 15 min. Supernatant was diluted by 50% with MilliQ water and 200 µL Salkowski's reagent (a mixture of 50 ml of 35% perchloric acid and 1 ml of 0.5 M FeCl 3 ) was added. After storage in the dark for 25 min, OD 530 was measured. IAA production (µg/mg biomass) was quantified using the standard curve made via IAA (5-100 µg/ml).

Phosphate solubilization assay
Pikovskayas agar plate (Pikovskaya 1948) was inoculated with a drop of 10 µL of fresh bacterial culture and incubated at 30˚C for 48 h. The diameter of the clear halo around the bacterial colony was measured and expressed as the phosphate solubilization activity of the strains.

Siderophore production assay
The assay was conducted using chrome azurol S, (CAS) (Schwyn and Neilands 1987) , both qualitatively and quantitatively. Fresh bacterial culture was prepared at 30°C for 48 h in a 1.5 ml tube containing 1.0 ml L medium. The bacterial culture was centrifuged at 7,700 × g 4˚C for 10 min and 0.5 ml of supernatant was mixed with 0.5 ml of CAS reagent. After 20 min, absorbance at 630 nm was measured using the spectrophotometer. The percent siderophore unit (psu) was calculated (Arora and Verma 2017). For the qualitative detection of siderophore, a modification of simple doublelayered CAS agar (SD-CASA) diffusion assay was used (Hu and Xu 2011). Bacterial colonies were grown in the L agar media for 48 h were overlaid with 10% CAS in 1% molten agar solution and incubated at 30˚C for an additional 48 h or until an orange halo was formed around the colonies in the blue background in positive control.

Nitrogen fixing activity
Nitrogen fixing activity was measured by acetylene reduction assay (ARA) under either plant colonized or bacterial suspension conditions. ARA is based on the activity of a nitrogenase enzyme for catalyzing the reduction of acetylene to ethylene (Bergersen 1970). The amount of ethylene (C 2 H 4 ) produced in the headspace of the sample vial was measured by GC-2014-FID (Shimadzu, Kyoto, Japan) with a Shincarbon-ST 50/80 mesh (4.0 m x 3.0 mm ID stainless column, Shinwa Chemical Ind. Ltd., Tokyo, Japan). The with gold and observed using a Model S-2400 (Hitachi, Tokyo, Japan) scanning electron microscope.

Plant growth-promoting activities of A. vinelandii A81 in mH medium
The PGP activity was examined using suspension experiments against four different Lemnaceae plants. Among the plants tested in the gnotobiotic A81 co-culture condition for 10 d of incubation, Lemna minor and Wolffiella hyalina frond numbers increased up to 1.5-and 1.3-folds and dry weight up to 1.7-and 1.6-folds, respectively, compared to the no bacteria control ( Fig. 1. a-e). In addition, A81 successfully colonized on all the duckweed species: L. minor, W. hyalina, L. gibba, S. polyrhiza ( Fig. 1. f). Relatively high and low CFU versus fresh weight values in W. hyalina and S. polyrhiza were probably due to their different surface area per fresh weight. We selected L. minor for further experiments as a model plant.

Growth recovery of L. minor by A. vinelandii A81 in nitrogen-free mH-N medium
Subsequently, we examined whether A81 could recover the plant growth in nitrogen-free water condition, mH-N. Experiments were performed via attachment condition to minimize the effect of bacterial cell suspension as the simple nutrient source. This revealed that the L. minor colonized by A81 successfully increased the frond number and dry weight 2.0-and 2.2-folds, respectively, after 10 d, compared to the nitrogenase genes, nifHDK deletion mutant CA12, and no-bacteria control (Fig. 2a, b). The CFU values of A81 and CA12 in the liquid medium after 10 days were 18.73 × 10 3 and 0.03× 10 3 per plant, respectively. Change in the CFU of A81 was observed for 30 d by transferring A81 colonized L. minor plants to a new flask every 10 d. The CFU value increased for the fronds and roots in mH medium (Fig. 2c). In contrast, CFU increased significantly on the fronds compared to the roots in the mH-N medium (Fig. 2d). Significant increase in CFU of A81 on the plant for 30 d and its PGP effect demonstrates a stable mutualism developed between the guest bacteria and the host plant. Possibility that A. vinelandii could colonize inside of the plant, so called as endophyte, was denied by confirming no CFU observed after surface sterilization of the L. minor cocultured with A. vinelandii A81. standard curve of the peak area versus amount of ethylene was determined by increasing the injection volume of 803 ppm ethylene to GC.

Bacteria colonized on L. minor
ARA of bacteria on L. minor was conducted using a method described in a previous study (Zuberer 1982) with several modifications. L. minor and A. vinelandii A81, C12, or Ensifer sp. SP4 co-culture were prepared in 0.3 OD 600 bacterial suspension in nitrogen-free medium mH-N and incubated for 2 d under plant growth condition. After gentle washing with sterile MilliQ water to rinse out loosely attached bacterial cells, the initial amount of colonized bacterial cells was measured by determining the CFU values. A 3 g fresh weight sample of L. minor colonized by bacteria was placed and sealed airtight with butyl rubber stopper and aluminum crimp in a 5 ml glass vial without a liquid medium. Control vial containing aseptic L. minor was also prepared. Then 10% of the headspace gas was replaced with acetylene and incubated for 5 d by ARA measurement under plant growth condition. The final CFU of the colonized bacteria on L. minor was also measured. To measure the CFU, the aliquots of plant bodies were mashed with Nippi Biomasher II, Tokyo, Japan, and serial dilutions of the sample were enumerated on the L plate using the spread plate technique. This process was performed for the bacteria-free control experiments to ensure there was no bacterial contamination.

Bacteria in suspension co-cultured with L. minor
ARA of A. vinelandii A81 and CA12 in suspension cocultured with or without L. minor was conducted using a method described by Bergersen (1970) with several modifications. Aseptic L. minor of 5 g was placed in a 5 ml glass vial and inoculated with 2 ml bacterial cell suspension of 1.0 OD 600 , 0.8 x 10 9 CFU/ml in nitrogen-free BS medium. Control vials that contained bacterial suspension or L. minor only were also prepared. The glass vials were closed using the aeration cap (Silico-sen, Shin-etsu chemical Co. Ltd., Tokyo, Japan) and left standing for 6 h in plant culture condition. After measuring CFU, the glass vials were closed airtight, after which ARA was conducted at 30˚C for 1 h.

Scanning electron microscopy (SEM)
A sample of bacteria colonized on either polypropylene or L. minor was fixed with 5% OsO 4 followed by 2% glutaraldehyde in 0.1 M phosphate-buffered saline (pH 7.0). After fixation, the samples were dehydrated by stepwise increasing concentration of ethanol, followed by treatment of critical point carbon dioxide. The specimen was sputter coated

Nitrogen fixing activity of A. vinelandii A81 cocultured with L. minor
Nitrogen fixation by A81 was illustrated in a co-culture suspension experiment with L. minor. A81 cells produced higher amount of ethylene (349.5 µmole/h) in the vials containing L. minor than in the vials without L. minor (67.3 µmole/h) (Fig. 3). CFU/vial of A81 was also higher when co-cultured with L. minor, 3.2 x 10 9, than without L. minor, 1.1 x 10 9 after 6 h suggesting positive effect by the plant. Only a small amount of ethylene was detected in the vial of CA12 co-cultured with L. minor, 2.2 µmole/h and without L. minor, 1.1 µmole/h. de novo production of ethylene by axenic L. minor was also negligible: 2.3 µmole/h.

Nitrogen fixing activity of A. vinelandii A81 colonized on L. minor
L. minor colonized by A81 produced 293.9 µmole ethylene/g fresh ethylene, which is significantly more than that produced by bacteria-free L. minor (123.5 µmole ethylene/g fresh) (Fig. S1). L. minor with nitrogenase negative strains, CA12 and SP4, only produced similar amounts of ethylene to the bacterial-free L. minor. The initial CFU values of A81 and CA12 were 2.3×10 4 and 8.0 ×10 4 CFU/g fresh weight, respectively, while SP4 showed higher CFU of 7.0 ×10 5 CFU/g fresh weight after exactly 2 d of co-culture in mH-N. The CFU of A81 later increased from 2.3×10 4 to 5.0×10 5 after 5 d on the L. minor, and produced 170.35 µmole ethylene/g fresh plant deduced from the bacteria-free control plants. Conversely, the CFU values of CA12 and SP4 decreased or did not changed significantly with negligible levels of ethylene production: 9.0 and 7.8 µmole ethylene/g fresh plant, respectively. These observations suggest that nitrogen fixation was caused by A81 colonized on L. minor. We observed ethylene production even in the control L. minor with no bacterial inoculation, which can be attributed to the de novo production of ethylene by the plant (Yang and Hoffman 1984).  3). Asterisks indicate the significant differences between values with and without A81 (Student's t-test, * P < 0.05, ** P < 0.005)

Discussion
The plant body in duckweed provides excellent nutritious residence and an ecological niche for bacterial cells in aquatic environments, especially when the availability of organic carbon compounds is limited. As observed in this study, a group of bacteria can actively proliferate and stimulate growth promotion on the host plants, creating a mutualism that bolsters the growth of both the bacteria and the plant even under high nutrient scarcity. The growth promotion effect of A. vinelandii A81 correlates with the positive CFU change (Figs. 1 and 2), confirming that co-existent bacteria play a role in the promotion of plant growth (Idris et al. 2007;Toyama et al. 2022). The higher CFU and lower

Effect of A. vinelandii A81 on the protein and starch contents in L. minor
The protein content in L. minor increased after being co-cultured for 10 d in A81 suspension mH medium, 40.7 ± 2.2%, compared to the no bacteria control, 23.2 ± 3.4%. In contrast, to increase protein content, starch content drastically decreased to 13.6 ± 4.2%, compared to the no bacteria control 48.9 ± 6.7% (Fig. 4).  3). Asterisks indicate the significant differences between values with and without bacteria (Student's t-test, * P < 0.05, ** P < 0.005). Different alphabets indicate significant differences (Student's t-test, P < 0.05) between treatments 2022), despite these bacteria having no history of growth with duckweed. These results indicate the acceptability of wider range environmental bacteria by the host duckweed plants. We found that a soil borne bacterium of A. vinelandii A81 can have PGP effect on L. minor and W. hyalina. PGP activity of A81 and P23 on L. minor was almost the CFU values in W. hyalina and S. polyrhiza (Fig. 1f) may be due to their large and small surface area per weight, respectively, compared to those of L. minor. Growth promotion of indigenous environmental or wastewater bacteria has been found to affect L. minor (Ishizawa et al. 2017b), L. gibba (Khairina et al. 2021)d polyrhiza (Toyama et al. . Asterisks indicate the significant differences between values with and without bacteria (Student's t-test, * P < 0.05, ** P < 0.005) Fig. 3 Nitrogen fixing activity of bacterial suspension co-cultured with L. minor Closed bars, Ethylene (µmole) detected/vial. Bacterial cells were inoculated at ca. 0.8 × 10 9 CFU/ml BS medium except no bacteria control. After the preincubation for 6 h, the CFU was measured, and ARA assay was conducted for 1 h. Open triangle, CFU values in the vial after 6 h pre-incubation. All values are mean ± SD (n = 3). Different alphabets between treatments indicate significant differences (one-way ANOVA; p < 0.05, Tukey HSD as a post-hoc test). L. minor of 5 g was contained in each vial as indicated source (Bellenger et al. 2011;Danapriatna et al. 2013). We initially assumed that the nitrogenase activity, monitored by ARA assay, of A81 would be decreased in the presence of L. minor. This may be attributed to the following reason. Nitrogen compounds and the oxygenic condition resulting from the photosynthesis of L. minor have a repressing effect on the nitrogenase enzyme activity of the colonized bacteria. However, we found that A81 cells fixed nitrogen in the plant associated condition. Plant exudates have been found to support bacterial nitrogen fixation in terrestrial plants (Van Deynze et al. 2018). In addition, a recent study has found that flavone biosynthesis by plants, including Lemnaceae, enhances biofilm formation and nitrogen fixation through diazotrophic bacteria (Pagliuso et al. 2020;Yan et al. 2022).
Although the apigenin and flavonoid derivatives produced by L. minor were not measured in this study, the effect of these plant-derived secondary metabolites on the nitrogen fixation activity of A81 cannot be ruled out. Furthermore, we observed A81 cells densely colonized on L. minor and encapsulated in extracellular polymeric substances, EPS on the surface of L. minor (Fig. S3). Significant ability of biofilm formation filled with EPS upon colonization on the L. minor may enable the A81 cells to use up oxygen and shape anaerobic condition enabled nitrogen fixation. The amount of ethylene produced by L. minor/ A81 clearly exceeded those of L. minor/no bacteria and L. minor/ CA12 or L. minor/ SP4 (Fig. S2). This may nullify the possibility of increased ethylene in L. minor/A81 owing to modification of plant physiology by bacteria.
We further examined the co-culture of L. minor with A81 in the cell suspension experiment, where A81 was exposed to a large amount of L. minor biomass in nitrogen-free BS medium (Fig. 3). A pre-incubation time of 6 h was necessary to observe the changes mediated by the plant on the A81 liquid bacterial culture. The host plant did not reduce the nitrogenase activity but prompted cell proliferation, thus possibly providing bacteria with growth promoting metabolites. Although further studies are necessary before we can conclude that L. minor actively bolsters the nitrogenase activity of A81, the study findings show that a reciprocally beneficial symbiotic interaction between L. minor and Azotobacter can be developed. A. vinelandii is an industrially relevant bacterium used to produce several compounds such as poly-β-hydroxybutyrate (PHB) (García et al. 2014) and alkyl resorcinol (Funa et al. 2006). The growth of A. vinelandii cells with valuables production using photosynthesis of L. minor plant as a scaffold should lead to the construction of a sustainable industry in future. same in the mH medium (Fig. S2). SP4 is a PGPB strain for S. polyrhiza and L. minor in the genus Ensifer (previously Sinorhizobium). SP4 did not form nodule on leguminous plants such as Canavalia gladiata and Phaseolus vulgaris and no nod and nif genes were observed in its genome (data not shown). Organic nitrogen is supplied by SP4 and the host S. polyrhiza accumulates significantly high amounts of Gln upon the inoculation of SP4 (Toyama et al. 2022). Whether A81 provides inorganic NH 3 immediately after N 2 fixation or organic nitrogen to the host L. minor has yet to be determined.
The CFU of non-duckweed originated A. vinelandii A81 in this experiment was continuously increased for 30 d in both mH and mH-N medium conditions (Fig. 2c, d), while the CFU of bacterium P23 obtained from natural duckweed decreased to 16% during the 10 d co-culture (Yamakawa et al. 2018). This suggests that P23 naturally co-exists with other bacteria to share their niches with optimal population balance. The high stability of mutualism between A81 and L. minor may allow for the further development of an effective biomass production technology of L. minor. Increased CFU on the frond section compared to the root in nitrogenfree mH-N medium condition was evident. This can be attributed to the easy access of A81 to air N 2 in the former area.
We hypothesized that the nitrogenase activity of A. vinelandii A81 could play a key role in the observed PGP effect on L. minor under the nitrogen-free water condition in the mH-N medium. To verify this hypothesis, we compared the PGP activity of A. vinelandii A81 with the nitrogenase negative mutant strain of A. vinelandii CA12 where the nifHDK genes were deleted. The growth recovery of L. minor in the mH-N medium was not observed for CA12 (Fig. 2a, b). Attachment experiments were adopted in this examination where the leakage of dead cell suspension lysate containing nitrogen compounds should be minimized. The fact that nitrogenase gene deletion mutant CA12 failed to recover the growth of L. minor suggests that the direct supply of nitrogen compounds from dead bacterial cells to the plants is negligible. A. vinelandii has been utilized as a biofertilizer owing to its phosphate solubilization, siderophore production, IAA production, and nitrogen fixation capabilities for terrestrial plants (Aasfar et al. 2021). However, it is too early to conclude that these growth promoting factors are solely or jointly responsible for duckweed growth promotion, except for cytokinins (Taller & Wong 1989;Kurepa et al. 2018), as the condition of the rhizoplane of aquatic plants differs significantly from that of terrestrial plants. Exogeneous IAA has no effect on the growth of L. minor (Utami et al. 2018).
Nitrogen fixation is a complex process requiring several key metal ions and a readily available supply of carbon