Introduction

Plants of the legume family (Leguminosae or Fabaceae) are widespread all over the world. They are able to establish a nitrogen-fixing symbiosis with nodule-inducing soil bacteria, here collectively called rhizobia. As well as the rhizobia, some non-symbiotic bacteria of other genera have also been isolated from legume nodules, Rahnella (Aserse et al. 2013), Bacillus (Bai et al. 2002, 2003; Deng et al. 2011; De Meyer et al. 2015; Saini et al. 2015; Zhao et al. 2018), Enterobacter (Benhizia et al. 2004; Ibáñez et al. 2009), Agrobacterium (De Lajudie et al. 1999), Pseudomonas (Shiraishi et al. 2010; Hoque et al. 2011),Burkholderia (Diouf et al. 2007), Herbaspirillum (Weiss et al. 2012), Pantoea (Kan et al. 2007), Paenibacillus (Deng et al. 2011; Zakhia et al. 2006; Li et al. 2008, 2012), Klebsiella (Ibáñez et al. 2009), Endobacter (Ramirez-Bahena et al. 2013), Arthrobacter, Microbacterium, Curtobacterium (Sturz et al. 1997; Palaniappan et al. 2010), Xenophilus, Erwinia, Leclercia (Xu et al. 2014), Gluconacetobacter, Variovorax, Micromonospora and Hyphomicrobium (Martínez-Hidalgo and Hirsch 2017; Xiao et al. 2017; Zhang et al. 2018).

These isolates are not able to infect the plant or induce nodules; neither do they fix nitrogen symbiotically and are termed non-symbiotic endophytic bacteria. The host plant provides endophytes with supply of nutrients and shelter from most abiotic stresses. In return, plants may receive benefits from microbial associations by the enhancement of plant growth or reduction of plant stress, e.g., through the ability of bacterial 1-aminocyclopropane-1-carboxyl-ate (ACC) deaminase to modulate the level of ethylene produced by plants under stress (Lodewyckx et al. 2002; Hardoim et al. 2008), production of indoleacetic acid (IAA), P solubilization, resistance to certain pathogens (Hardoim et al. 2002; Peix et al. 2015; Tariq et al. 2014), nitrogen fixation (Andrews et al. 2010) and siderophore production (Rajendran et al. 2008), Most legumes are good pioneer species, adapted to low nutrient soils and/or environments with heavy metals (Gonzalez-Andres et al. 2005; Vidal et al. 2009; Azcon et al. 2010). In this regard, previously mentioned studies have reported that other rhizobia such as non-symbiotic bacteria may inhabit legume root nodules and have proven benefit to their legume hosts. However, non-symbiotic bacteria may inhabit legume root nodules and visibly harm the host, and their biological role in the plant is largely unknown.

As previously found, the nodules have been proposed to form an ecological niche for survival of the bacteria (Muresu et al. 2008). Assistance in nodulation ability and improved nodulation and nitrogen fixation through synergistic interactions of rhizobia with legume plants have been proven by some earlier studies (Andrews et al. 2010; Rajendran et al. 2008). Enterobacter species were once reported to have infection ability and nodulation genes. Their mode of entry is probably associated with infecting rhizobia (Benhizia et al. 2004). Moreover, recent reports suggest increase in root wet weight and nodulation when co-inoculated with nodule endophytes, compared to inoculation with rhizobia alone (Bai et al. 2002, 2003; Sturz et al. 1997; Barret and Parker 2006; Annapurna et al. 2013; Masciarelli et al. 2014). This is further supported by previous reports showing that significant increase in nodule activity and nitrogen content in soybean on co-inoculation of Bacillus megaterium LNL6 along with Bradyrhizobium japonicum MN110 (Subramanian et al. 2015).

The Leguminosae includes over 18,000 known species (http://www.Ildis.org/ Leguminosae/), but so far only a fraction of those have been investigated for nodulation (Sprent 2001) and even fewer for the occurrence of nodule endophytes. Sphaerophysa salsula is one of the wild herbaceous legumes commonly found in the northwestern arid regions of China. This deep-rooted legume has potential for vegetation of arid regions, but its bacterial partners are so far unknown. To obtain information needed to ensure good colonization and nitrogen fixation in future application, it was of interest to investigate the diversity of non-symbiotic bacteria inhabiting the root nodules of this plant. We recently isolated and characterized a number of inhabiting non-symbiotic bacteria from Sphaerophysa salsula root nodules (Deng et al. 2011). The present study is based on the same sampling sites from root nodules of Sphaerophysa salsula; however, it focuses on the non-symbiotic bacteria that are capable of atmospheric nitrogen fixation and those possessing the nifH gene. The aim of this study was to gain insights into their potential nitrogen fixation ability and to investigate the possible connections with the ecoregion and their properties in plants growing in diverse habitats. Additionally, the results on the non-symbiotic bacteria from our previous study (Deng et al. 2011) were applied and compared with the plant growth-promoting characteristics found in the present study. Thus, we hypothesized that they can enhance host and other plant growth.

To verify our speculation, we selected 20 plant growth-promoting representative bacteria, isolated from root nodules of Sphaerophysa salsula, and tested them for compatible co-inoculation with Mesorhizobium sp. Zw-19 to improve nodulation and nodule activity under pot culture conditions and field experiments, to determine their effect as PGPB on the growth of the Sphaerophysa salsula. Thus, the outcomes of this study could elucidate the relationship between them and improve their potential as a sustainable and inexpensive alternative to the use of agrochemicals.

Materials and methods

Isolation and characterization of nodule endophytes

Root nodules were collected and isolated. DNA extraction and sequence analysis of 16S rRNA gene were performed as previously reported (Deng et al. 2011).

Amplification and sequencing of the nifH gene

The primers 34F (5′-AAA GG(C/T) GG(A/T) ATC GG(C/T) AA(A/G) TCC ACCAC-3′) and 491R (5′-TTG TT(G/C) GC(G/C) GC(A/G) TAC AT(G/C) GCC ATC AT-3′), and the procedure described by Van Berkum et al. (Van Berkum et al. 1996) were used for nifH gene-specific amplification by PCR. The bands corresponding to the expected size (460 bp) were cut, purified and sequenced. Sequences were assembled and aligned, and a phylogenetic tree was constructed as described for the16S rRNA gene analysis.

Plant inoculation studies with endophytes

Sphaerophysa salsula seeds were surface sterilized with 96% sulfuric acid for 20 min, then washed ten times with sterile water to remove all traces of acid, immersed in 95% ethanol for 1 min, followed by 0.1% HgCl2 for 2 min and rinsed eight times in sterile distilled water. Surface-sterilized seeds were allowed to germinate on moist filter paper kept in sterile Petri dishes containing ten seeds each. Seedlings were coated with overnight grown bacterial cultures by incubating them in a thick suspension (approximately, 109 cfu mL−1) of the bacteria for 8 h at 28 °C. Zw-19 was isolated from Sphaerophysa salsula root nodule in this work and 16S rRNA sequence confirmed it as Mesorhizobium sp. One milliliter of inoculum from the endophytic isolates alone and the mixture containing each isolate with Mesorhizobium sp. Zw-19 (500 + 500 µL) were inoculated on the surface of sterilized Sphaerophysa salsula seeds. Seeding without any inoculation as the negative control (NC) and with Mesorhizobium sp. Zw-19 alone as a positive control (PC), which received equal volumes of inoculum, was included for comparison, respectively. The plants were then transferred to glass tubes containing nitrogen-free plant nutrient solution (Vincent 1970) sealed with cotton plugs. All nodulation tests were performed in triplicate with non-inoculated control plants included. Roots were observed for nodule formation during 6–8 weeks after inoculation. The surface-sterilized roots and stems of seedings inoculated with the endophytic bacteria alone were ground to estimate the colony-forming units (CFUs) of endophytic bacteria (Saini et al. 2015).

Chlorophyll estimation

The midrib of the leaves was removed and 1 g leaf tissue was crushed in 80% acetone; chlorophyll was measured spectrophotometrically using the specific absorption coefficients for chlorophyll a at 664 nm and chlorophyll b at 647 nm. The chlorophyll content was estimated by using the previous method (Graan and Ort 1984).

Characterization of the plant growth-promoting properties of endophytic bacteria

IAA production assay

Indoleacetic acid (IAA) production was analyzed according to the qualitative method (Glickmann and Dessaux 1995). Bacterial culture incubated in King B medium at 28 °C for 36 h was mixed with the Salkowski reagent (1:1 v/v) and incubated in darkness for 30 min. The production of IAA was recognized by the presence of red color.

Phosphate solubilization

An aliquot of 10 mL of fresh bacterial culture was spread onto TY medium supplied with 5 g L−1 of Ca3(PO4)2 and was incubated at 28 °C for 2–3 days. A clear halo around the bacterial colony indicated solubilization of mineral phosphate.

Chitinase production

The bacterial suspension was spot inoculated on chitin agar plates and the zone of clearance was recorded after incubation at 28 ± 2 °C for 48–72 h.

Siderophore detection

Bacterial cultures were multiplied in Ashby’s Mannitol Broth (AMB) for 48 h and an equal volume of culture supernatant was added to chrome azurol S (CAS) assay solution (Schwyn and Neilands 1987). A color change from blue to pink was recorded.

ACC (1-aminocyclopropane-1-carboxylate) deaminase production

ACC deaminase enzyme production capability was assessed based on the ability to take ACC as a sole nitrogen source in a minimal medium (Duan et al. 2009). Cultures were spot inoculated on plates containing minimal medium supplemented with 3 mM ACC substrate. Plates containing minimal medium without ACC served as negative control, and those with(NH4)2SO4 (2.0 g mL−1) as a nitrogen source served as positive control. The plates were incubated for 4 days at 28 ± 2 °C and the growth of the isolates on ACC-supplemented plates was compared with that of the positive and negative control plates.

Fungal inhibition growth assay

Spores of fungal cultures (Fusarium oxysporum, Alternaria burnsii and Rhizoctonia solani) were grown on potato dextrose agar (PDA) plates. An agar plug (5 mm diameter) taken from an actively growing fungal culture was placed on one side of the surface of the PDA plate. Test bacteria were streaked perpendicular to the agar plug on the opposite side toward the edge of the plates. The plate inoculated with fungal agar plugs alone was used as the control. The plates were incubated at 28 ± 2 °C until fungal mycelia completely covered the agar surface in the control plate, and the zone of inhibition was recorded. A test was considered positive when bacteria interfered with the normal spread of the fungus and presented inhibition zone in three replications.

Root hair deformation assay

Bacterial cultures were grown in 50 ml liquid media incubated at 30 °C on a shaker. After 48 h incubation, 5 ml of the culture was transferred to 95 ml fresh liquid medium and allowed to grow for another 12 h. The inoculation of bacterial culture (0.5 ml) was done by making a suspension in N-free Fahraeus medium at a density of approximately 109 cfu/ml. Sphaerophysa salsula seeds were surface sterilized with 96% sulfuric acid for 20 min. Seeds were then washed ten times with sterile water to remove all traces of acid, immersed in 70% ethanol for 1 min, followed by 0.1% HgCl2 for 2 min, and rinsed eight times in sterile distilled water. Sterilized seeds were germinated on 0.8% (w/v) agar contained in 9 cm Petri dishes for 2 days in the dark at 25 °C. Two-day-old seedings were transferred aseptically to sterilized glass slides (one per slide) containing nitrogen-free Fahraeus medium semi-solidified with 0.8% (w/v) agar. Before applying the bacterial cultures, the roots of all the plants were examined microscopically and Fahraeus slides containing plants with deformed hairs were discarded. After 1 day, seedings were inoculated with mid log phase cells of bacterial cultures. The inoculation of sterilized water and the sterilized medium served as control. Seedings were grown under day-light fluorescent tubes in a plant growth chamber with a 16 h photoperiod for 10 days at 25 °C (day) and 22 °C (night). The morphological features of primary root infection were examined by bright field microscopy at day 2 after inoculation.

Field layout

The field experiments were conducted in Yan’an University, Shaanxi, China, and designs were randomized complete blocks under field conditions. The experiment was designed as a 3 × 2×2 factorial organized in a randomized complete block split plot with three replications, and three plant samples comprised one replicate. The main plot units consisted of PGPR (plant growth-promoting rhizobacteria) strain inoculation (the mixture of endophytic isolates including Bacillus pumilus Qtx-10 and Streptomyces bottropensis Gt-10 with Mesorhizobium sp. Zw-19 by adding 100 mL of bacterial suspension 108 CFU mL−1 per plant (no-PGPR inoculation as a control). Each sub-plot was 5 × 2 m and consisted of three rows of plants with 40 cm between rows. The space between plots was 80 cm with 1 m between replications. The experiments started in early May and ended in mid-July in 2018.

Results

Isolation and characterization of nodule endophytes

Sixty-five endophytic bacteria (Deng et al. 2011) that grew on PDA, KingB or NA medium which could not induce nodules were isolated from six different samples(Table 1), and 20 representative endophytes were selected and tested for their ability to nodulate Sphaerophysa salsula and plant growth-promoting assays(Table 2). In the nodulation tests under aseptic conditions, no nodules were induced by the 20 representative endophytic bacteria. The amount of CFU, from the surface-sterilized roots and stems of the inoculated seedings varying from 1.13 × 104 to 6.52 × 105 CFU per gram of fresh root and 3.21 × 104 to 8.54 × 105 per gram of fresh stem, also fits the range of endophytic bacteria (Table 1).

Table 1 Physicochemical characteristics of the studied soils
Full size table
Table 2 Rsults of comparative growth profile of inoculation tests
Full size table

Sequence analysis of nifH gene

Amplified fragments of nifH corresponding to the expected size of 460 bp were produced in PCR from the 19 tested rhizobial DNA and from that of one single endophytic strain. The phylogenetic analysis of the sequenced fragments (Fig. 1) showed that this nifH sequence was related to that of M. amorphae. M. gobiense, M. septentrionale and M. albiziae. The nifH sequence amplified from Bacillus pumilus Qtx-10 was 100% similar to that of Rhizobium leguminosarum Qtx-10-1 and 96.14% similar to that of M. albiziae. Interestingly, on collecting them from the same sampling site, we found that Bacillus pumilus Qtx-10 co-occurred with Rhizobium leguminosarum Qtx-10-1 in the same nodule. This close relationship of the same ecological niche between them is speculated to cause the co-evolution of symbiotic genes. However, lateral gene transfer between non-symbiotic and symbiotic bacteria is rare. The positions of some isolates in this study showed aberrant features in the phylogenetic trees. Gene transfer somehow drives genetic diversity and adaptability, and lateral symbiotic gene transfer may have significant effects at the species or genus level.

Fig. 1
figure 1

Phylogenetic tree of representative strains isolated from Sphaerophysa salsula (labeled in bold) and reference strains generated by the neighbor-joining method based on nearly full length (460 bp) nifH sequences. Bootstrap values (1000 replicates) are indicated as percentages (>70%) above the branches. Scale bar indicates 2% substitution of nucleotide

Full size image

Root hair deformation assay

The root hair deformation assay performed in Sphaerophysa salsula revealed the presence of modified root hair structures in bacterial treatments. The deformation was induced only in a small zone of the root containing root hairs that had almost stopped growing. Out of 20 representative endophytic isolates, only treatment with Qtx-10, Gzn-9-1 and Zw-13-3 inoculation resulted in deformed root hairs with bulging at the root tip. Examination of the root segments under bright field microscopy showed hemispherical bulbous structures and root structures resembling the initial stages of nodule or tumor outgrowths that could not be observed in uninoculated control seedings (Fig. 2a–d).

Fig. 2
figure 2

Root hair deformation bacterial cultures-inoculated Sphaerophysa salsula seedings: a uninoculated control with normal root hairs, b microscopic view of infected root hairs in Bacillus pumilus Qtx-10 inoculated Sphaerophysa salsula seedings, c microscopic view of infected root hairs in Bacillus licheniformis Gzn-9-1 inoculated Sphaerophysa salsula seedings, d microscopic view of infected root hairs in Bacillus safensis Zw-13-3 inoculated Sphaerophysa salsula seedings. Scale bars = 200 µm in (a); 300 µm in (b, c); 200 µm in (d)

Full size image

Plant growth promoting assays

Salkowski reaction of culture supernatants revealed that of the strains Qtx-10, Gzn-9-1 and Zw-13-3 were able to synthesize IAA from tryptophan. Based on the sequencing of the 16S rRNA gene, Qtx-10 was identified as Bacillus pumilus, Gzn-9-1 was designated as Bacillus licheniformis and Zw-13-3 belonged to Lysinibacillus fusiformis. All the isolates were chitinase negative, only Zw-12-2 belonging to Bacillus safensis showed phosphate solubilization activity, having a clear halo around the colony. Out of 20 representative endophytic isolates, 11 strains gave a positive CAS assay test showing that they produced siderophores, belonging to genera Staphylococcus, Lysinibacillus, Bacillus, Paracoccus, Streptomyces, Nocardia, Mycobacterium, Paenibacillus and Pseudomonas, respectively. The capability to inhibit plant pathogenic fungi was analyzed by using Fusarium oxysporum, Alternaria burnsii and Rhizoctonia solani as target organisms, and six strains showed antifungal activity. The capability of inhibiting fungal growth was widely spread in the different genera belonging to Paracoccus, Paenibacillus, Bacillus pumilus, Inquilinus and Mycobacterium, respectively. Antifungal activity of Zy-3, Gt-1, Qtx-10, Gzn-9-1, Mq-10 and Mq-2-1 was checked against Fusarium oxysporum using potato dextrose agar (PDA) medium. The antifungal activity of strains tested varied with inhibition zones in diameter from 18.20 to 36.00 mm, strains Gt-1 and Qtx-10 induced larger inhibition zones compared to the other strains (inhibition zone more than 30.00 mm) (Fig. 3), Gt-1, Qtx-10 and Gzn-9-1 could also exhibit broad-spectrum activities against test fungi. Production of IAA, siderophore and antifungal activity was simultaneously exhibited by isolates of Qtx-10 and Gzn-9-1.

Fig. 3
figure 3

Antifungal activity of endophytes test. a Inhibition zones of strain Gt-1 against Valsa mali Miyabe et Yamada; b Inhibition zones of strain Qtx-10 against Valsa mali Miyabe et Yamada

Full size image

Under pot culture conditions, 15 out of the 20 strains showed plant growth-promoting activity with respect to various plant parameters. Among the isolates, all of the 15 isolates promoted the fresh weight and shoot length of Sphaerophysa salsula seedings when inoculated alone and showed similar effects on the fresh weight and shoot length as the rhizobial bioinoculant strain Zw-19. While Gaoshi-1, Zw-22, Zy-3, Qtx-12 and Mq-17 showed decrease in seeding fresh weight when inoculated alone, all five isolates promoted plant growth when applied as a mixture with Zw-19 (Table 2). Qtx-10 inoculation enhanced the seeding fresh weight by 87.5%, and the isolate Gt-10 enhanced the shoot length by 89.4%. Over the positive control, Gt-25 and Zy-2-1 when co-inoculated with Zw-19 showed decrease in seeding fresh weight compared to its individual application. The chlorophyll content was improved in plants inoculated with all of the 20 representative endophytic isolates, over the noninoculate control, and their chlorophyll values were also improved when co-inoculated with Zw-19. Except Gt-25, there was an increase in the average nodule fresh weight per plant when those 19 strains were inoculated in combination with Zw-19.

The results of field experiments indicated that the inoculum of strain Zw-19 combined with two kinds of bacteria (Qtx-10 and Gt-10) has a significant role in growth promoting than the single inoculum Zw-19. Under the conditions of field, the growth rate of the plant height of Sphaerophysa salsula was more than 15.99% higher than that of Zw-19. Moreover, the number of nodules was 31.89% higher than that under field conditions. The growth-promoting effect of combined Zw-19 with endophytic bacteria on to S. salsula has a significant role in various plant growth promotion parameters (Table 3).

Table 3 Rsults of comparative growth under field conditions
Full size table

Discussion

Nodules can be colonized internally by several bacterial genera unrelated to rhizobial symbiotic nitrogen fixation. Benhizia et al. (Benhizia et al. 2004) reported Pseudomonas strains to be associated with legume nodules. Actinobacteria such as streptomyces lydicus have been reported to colonize pea nodules (Tokala et al. 2002), and Bacillus thuringiensis can naturally coinhabit soybean nodules along with Bradyrhizobium japonicum. Bacillus cepacia was also recorded as a plant endophyte (Balandreau et al. 2001) and as a nodule occupant (Vandamme et al. 2002), and some other species in the genera Bacillus have been found in nodules (Bai et al. 2002; Barret and Parker 2006). Our finding (Deng et al. 2011) is consistent with previous studies showing Bacillus t be enriched in the nodules (Zakhia et al. 2006), and Bacillus and Pseudomonas were mainly endophytic bacteria in the nodules (Kan et al. 2007). As De Meyer et al. (2015) mentioned, the majority of the investigated nodules contained Bacillus (17.9%), Paenibacillus (12.5%) and Pseudomonas (15.9%) found in legume root nodules from native legume species in Flanders (De Meyer et al. 2015), which indicates that certain legume species prefer certain non-symbiotic endophytic bacteria in their root nodules. The implications of the discovery would be significant, especially if it is confirmed that the association between legume nodules and Gram-positive bacteria is common in nature and that this interaction is beneficial for plant growth.

In this study, an evidence of the high similarity (99%) sequences homologous to the nifH gene of Mesorhizobium was detected within strain Qtx-10 belonging to the genera Bacillus, and it offered strong evidence to suggest that lateral gene transfer of nifH might have happened between the symbiotic and endophytic bacteria. The frequent lateral gene transfer might help S. salsula to attract the most adapted bacteria, subsequently, and improve the diversity of bacteria within the root nodule of S. salsula. This result is similar to the previously reported sequences of nifH amplified from Bacillus spp. CCBAU15524 (EF471734) and CCBAU15518 (EF471735) were 99% similar to that of B. japonicum (AJ563961) (Li et al. 2008). Sequence analysis of the nifH gene revealed that the strains belonging to Xenophilus, Acinetobacter, Phyllobacterium, and Rhizobium had genes similar to those of Mesorhizobium and Sinorhizobium (Xu et al. 2014). Seven endophytic non-rhizobial bacteria, which belong to Enterobacter cloacae, Chryseobacterium indologenes, Klebsiella pneumoniae, Pseudomonas aeruginosa, Enterobacter ludwigii and Klebsiella variicola were isolated possessing the nifH gene (Dhole et al. 2016). Zhao et al. (2011) demonstrated that the nifH gene was amplified from representative strains Bacillus cereus and Bacillus amyloliquefaciens which were isolated from the root nodules of soybean (Glycine max L.) and had nitrogen fixation activity (Zhao et al. 2011). In the present study, nifH genes of strain Qtx-10 and other rhizobia were more close than between the 16S rRNA genes of Bacillus and rhizobia, which was consistent with the results of Zhao et al. (2011). This suggests that the presence of nifH gene in both Bacillus and rhizobia probably occurred through horizontal gene transfer. All previous studies and our study demonstrated that horizontal gene transfer could have occurred between rhizobia and non-rhizobial endophytes. They may have co-evolved. Therefore, the non-symbiotic endophytic bacteria in nodules may serve potential receptors of symbiotic genes and may be resources of novel symbiotic bacteria. Although rhizobia have been studied for more than 120 years, symbionts of less than 10% of the 750 legume genera have been fully characterized. Our work suggests that characterization of the symbionts of the yet unexplored legumes may reveal the rhizobial nature of additional members of possibly other taxonomic classes. Such a study may contribute significantly to the understanding of the origin and the evolution of the legume–rhizobia symbioses, showing that the ability to establish a symbiosis with legumes is more widespread in bacteria than anticipated to date. Based on its importance, both the horizontal transfer and the function of nifH in nodule endophytic Bacillus need to be studied further.

In the present work, we describe here non-symbiotic bacteria that are closely associated with nodules of S. salsula and their interaction with S. salsula nodulating rhizobial strain Zw-19 in planta. Our results parallel other studies, showing that the legume root nodules are known to accommodate several eubacterial genera apart from rhizobia, and their population densities are reported to be in the range of 104 viable bacteria per gram of fresh nodule tissue (Sturz et al. 1997). Although the role of such bacteria is not clear, it has been hypothesized that IAA of bacterial origin from the nodules is transported to other plant parts (Basu and Ghosh 1998) and use IAA as a part of their colonization strategy, involving phyto-stimulation and circumvention of plant defense mechanisms. Moreover, IAA might act as a signal molecule in bacteria–bacteria communication (Spaepen et al. 2007). The cooperative interactions between rhizobia and other plant root-colonizing bacteria are of relevance in the improvement of nodulation and N2 fixation in legume plants (Barea et al. 2005). Besides cellulase, protease produced by many Bacillus and Paenibacillus strains was proposed to degrade fungal cell walls, thereby inhibiting plant pathogens (Lodewyckx et al. 2002). Bacteria that produce siderophores can inhibit plant pathogens by competing with them for iron in the rhizosphere. Some plants can also bind, transport and exploit bacterial iron–siderophore complexes (Lodewyckx et al. 2002). Perhaps, some of these functions explain why the production of siderophores was common among our non-symbiotic strains. Furthermore, the re-isolation of these bacteria from stem, roots and nodules of the inoculated host plants of Sphaerophysa salsula evidenced their endophytic properties according to the definition of endophytic bacteria (Hallmann et al. 1997).

Other previous studies reported that plant growth promotion when rhizobia are co-inoculated with Bacillus spp. include studies on Rhizobium leguminosarum bv trifolii with either B. insolitus or B. brevis (Sturz et al. 1997) and with Bacillus spp. and the soybean endosymbiont, Bradyrhizobium japonicum (Bai et al. 2003; Liu and Sinclare 1993). Our results are in agreement with the above reports. To the best of our knowledge, this study is the first report to show that the effects of a Streptomyces bottropensis endophyte with Mesorhizobium spp. on Sphaerophysa salsula growth. Our results revealed that the association between legume nodules and endophytic bacteria is common in nature and that this interaction is beneficial for plant growth.

Many different mechanisms have been proposed as the basis of nodulation enhancement by epiphytic or endophytic root-associated bacteria (Barea et al. 2005). Using cell-free supernatants of bacterial cultures, it has been demonstrated that plant growth-regulating substances produced by rhizobacteria affected nodulation and nitrogen fixation, (Hardoim et al. 2002; Peix et al. 2015; Tariq et al. 2014; Andrews et al. 2010; Rajendran et al. 2008; Gonzalez-Andres et al. 2005; Vidal et al. 2009; Azcon et al. 2010; Muresu et al. 2008; Barret and Parker 2006; Annapurna et al. 2013; Masciarelli et al. 2014; Subramanian et al. 2015; Sprent 2001; Van Berkum et al. 1996; Vincent 1970; Graan and Ort 1984; Glickmann and Dessaux 1995; Schwyn and Neilands 1987; Duan et al. 2009; Tokala et al. 2002; Balandreau et al. 2001; Vandamme et al. 2002; Dhole et al. 2016; Zhao et al. 2011; Basu and Ghosh 1998; Spaepen et al. 2007; Barea et al. 2005; Hallmann et al. 1997; Liu and Sinclare 1993; AzcÓn-Aguilar and Barea 1978; Manero et al. 2003). Further work is needed in this regard.

In previous studies, endophytic bacteria contribute to plant growth, ecological performance or reduction of plant stress (Lodewyckx et al. 2002). Our work revealed that a wide variety of rhizobial and non-symbiotic endophytic bacteria can colonize S. salsula nodules; moreover, they can assist the symbiotic interaction between rhizobia and the host plant.

The previous report (Bai et al. 2003) showed that Bacillus subtilis and Bacillus thuringiensis can naturally coinhabit soybean nodules along with Bradyrhizobium japonicum, and that these Gram-positive bacteria can enhance plant productivity in co-inoculation experiments. A more recent report (Zakhia et al. 2006) described the association of 14 bacterial genera with wild legume nodules in Tunisia. In the present study, most of the root nodules of S. salsula harbor prevailingly rhizobia. Meantime, these nodules are also colonized internally by nonrhizobial endophytes. We plan to examine if undomesticated legumes differ from those cultivated in agriculture in their production of metabolites. Investigating these aspects will provide a better insight into the microbial interactions occurring in native and introduced wild legume plants and will lead to a better understanding of their nitrogen-fixing symbioses in further studies.

In sum, nodulation mainly occurred at the junction of the primary and secondary roots. The Qtx-10, Gzn-9-1 and Zw-13-3 inoculation resulted in root hair deformation in Sphaerophysa salsula. Furthermore, more lateral root formation can provide more potential rhizobial invasion sites, and thus enhance nodulation. But the tip bulging and deformation were not prominent and also uniformly distributed throughout the emerging root hairs, implying that the root infection process is a random event (Ibáñez et al. 2009; Rolfe et al. 1997).

Conclusions

Conclusively, the representative nodule endophytes improved nodulation and nodule activity in Sphaerophysa salsula on co-inoculation with Mesorhizobium sp. Zw-19. These results demonstrated the diverse non-symbiotic bacteria associated with this plant grown promotion in northwestern China and the universal existence in its nodules. These results provide valuable information about the interactions among the symbiotic bacteria, non-symbiotic bacteria and their habitats.