Rhizobacteria can be widely found in agriculture as they are symbiotically associated with legumes. Thus, they receive protection and nutrition from symbiont plant and, in turn, provide nitrogen (N), allowing in many cases the absence of mineral fertilization for this substantial nutrient.

In the 1990s and 2000s, a series of cutting-edge experiments outlined a new paradigm for the interaction between rhizobia and plants. Yanni, Rizk, Abd, El-Fattah, Squartini, Corich, Giacomini, Dazzo, F. B. (2001), are one of the first researchers to study rhizobia inoculation on cultivated non-legume plants.

Rhizobia not only act as nitrogen fixers when combined with legumes, but they also stimulate plant growth in non-legume plants. As well, they are capable of synthesizing phyto-stimulating substances, such as auxins (Osório, Gano, Binz, Lima, Martinez, Ramirez, Caballero, de Sá & Giongo, 2014; Ghosh, Kumar & Kanti, 2015), cytokinins (Senthilkumara, Madhaiyan, Sundaram & Kannaiyan, 2009), gibberellin (Miransari & Smith, 2009), ACC deaminase (Nascimento, Brígido, Alho, Glick & Oliveira, 2012), among others. The production of all these phyto-stimulating substances is intrinsically related to an increasing of non-legume crop yield and should be then explored.

As in the acceptance of the atmospheric nitrogen fixation by scientific community, evolving from the 20s to 60s and 70s, to subsequently benefit thousands of farmers throughout the world, it is a consensus that rhizobia acting as growth promoters of non-legume crops through production of phyto-stimulating substances, which are still under evaluation. We strongly believe that only through successive corroborative studies we can reach a common understanding that secondary mechanisms for plant growth promotion can also be introduced in crop succession systems, with considerable benefits.

Therefore, two valuable principles are required: an isolation of plant-bacteria and compatibility studies. In early stages of research, isolation is obtained under axenic conditions (in vitro or under greenhouse conditions) to ensure that the effect was mainly caused by plant-related bacteria. Compatibility studies are needed because rhizobia isolates are beneficial to some species or specific cultivars, but inert to others. Thus, before field studies, an extensive prospecting research both in laboratory and greenhouse are required.

One of the major phyto-stimulating substances produced by rhizobia is auxins, among which Indole Acetic Acid (IAA) is the most studied and the most produced by bacteria. IAA primarily focuses on the formation of side roots and root hairs which increases the uptake of nutrients by plant and is related to shoot dry formation (Lwin, Myint, Tar & Moe, 2012). In addition, IAA is related to colonization strategies of plants by bacteria (Spaepen, Vanderleyden & Remans, 2007), however, the effects of IAA on plant root tissue are concentration dependent and can be species specific (Gutierrez, Matsui, Lincoln & Lovell, 2009).

Different pathways for IAA biosynthesis have been identified in bacteria, being indole-3-acetamide (IAM) and indole-3-pyruvate (IpyA) tryptophan-dependent (Lambrecht, Okon, Broek & Vanderleyden, 2000). The pathway of indole-3-pyruvate (IPyA) was identified in rhizobia (Giraud, Moulin, Vallenet... Sadowsky, 2007) as well as in intermediate components of tryptamide (TRP) (Kaneko et al., 2000) and indole-3-acetonitrile (IAN) (Kaneko et al., 2002).

Millet (P. glaucum), is an exotic plant of African origin that produces high quality forage in a short time, being excellent for feeding cattle. It is widely recommended in Rio Grande do Sul and Santa Catarina states of Brazil for pastures composition and silages, according to their suitability, being highly resistant to drought (Fontanelli, 2012). Sudan grass (S. sudanense) is a forage species originating in Sudan and southern Egypt, adapted to dry and hot climates.

According to Machado, de Sá, Hahn, Oldra, Mangrich, João, Osório, Dias, Dobler & da Silva (2016), Sudan grass and millet are responsive to inoculation with rhizobia. In a study of inoculation with rhizobia in Sudan grass and millet, authors verified an increasing in total dry mass production with rhizobia inoculation.

Given these concerns, this study aimed to evaluate the effect of inoculation with plant growth promoting rhizobia in the yield of millet (P. glaucum) and Sudan grass (S. sudanense) and verify if the enrichment of culture medium with tryptophan interferes with the rhizobium-plant interaction.

Two rhizobia isolates largely studied for the promotion of grass growth, VP16 and UFRGS Lc348, were used. They were obtained from root nodules of white clover (Trifolium repens L.) and birds foot trefoil (Lotus corniculatus L.) plants, respectively. These organisms were obtained from the UFRGS Collection of Rhizobia at the Soil Microbiology Laboratory, Department of Soils, Federal University of Rio Grande do Sul, Brazil.

Rhizobia were assessed for their effect on the initial plant growth and development of adult inoculation of plant growth promoting rhizobia in Sudan grass (Sorghum * sudanense (Piper) Stapf cv. Sudanense) and millet (Pennisetum glaucum (L.) R.Br. cv. BRS1501) plants of Sudan grass and millet. Experiments were carried out in the Soil Microbiology Laboratory and in the greenhouse facilities of the Department of Soils of Agronomy School of the Federal University of Rio Grande do Sul-UFRGS, Brazil.

Isolate VP16 produced 4.0 Mg of IAA.mL^-1 in YM liquid medium and 10.9 Mg of IAA mL^-1 in YM liquid medium enriched with tryptophan. Compared to control treatment, both VP16 and VP16 + Trp, when inoculated on millet seeds, induced epicotyl and hypocotyl higher elongation (Table 1).

Table 1

Averages of four replicates, each replicate consisted of 40 seeds. Averages followed by same letter in the column do not differ by the Scott Knott test at 5% error probability. Averages without letters in the column do not differ. YM: yeast manitol, Trp.: tryptophan, % CV= coefficient of variation percentage.

These in vitro stimuli render an adaptive advantage of cultivar BRS1501 and may induce a more effective search for water and nutrients.

Isolate UFRGS Lc348 produced 1.1 Mg of IAA. mL^-1 in YM liquid medium and 60.7 Mg of IAA.mL^-1 in YM liquid medium enriched with tryptophan. Compared to both control and UFRGS Lc348 treatments, UFRGS Lc348 + Trp presented a greater epicotyl length in millet seedlings (Table 1), possibly due to high production of IAA by isolate UFRGS Lc348 in a tryptophan-enriched medium.

In vitro experiment, inoculations of the VP16 isolate in millet induced a higher epicotyl and hypocotyl length, but no direct gains were observed in plant dry mass or anticipation of germination. Thus, although they do not have allowed direct gains in dry mass, these stimuli induced by VP16 in millet can mean advantages in a dispute for nutrients in the rhizosphere, or through photosynthetic radiation, which indirectly favors the plant growth and initial development of millet.

Despite inducing greater in vitro millet seedlings length, largest production of indole substances by rhizobia in enriched substrate (VP16 + Trp and UFRGS Lc348 + Trp treatments), had achieved a decreasing in the number of germinated seeds at the end of the process (Table 1). On the other hand, the effects observed in in vitro millet seedlings were not observed in experiment with adult plants under greenhouse conditions (Table 2).

Table 2

Averages of four replicates, with two plants per replicate. Averages followed by same letter in the column do not differ by the Scott Knott test at 5% error probability. Averages without letters in the column do not differ. For statistical analysis, variables number of tillers per plant and number of leaves per plant were transformed according to the formula: ((X + 1)^0,5). %CV= coefficient of variation percentage, Control N: treatment not inoculated with nitrogen dose equivalent to 150 kg.ha^-1, Control Κ N: treatment not inoculated with nitrogen dose equivalent to 75 kg.ha^-1. Values without letters in the column did not differ.

Indole substances from the culture medium enriched with tryptophan were diluted in pots of 2 L, which is possibly associated to the inefficiency of millet inoculations under greenhouse conditions.

In experiments with in vitro Sudan grass inoculation, UFRGS Lc348 + Trp stimulated the seedling, inducing a greater length of the hypocotyl. The length of inoculated Sudan grass hypocotyl with UFRGS Lc348 was 60% higher than YM control treatment, indicating a beneficial effect to seedlings (Table 3).

Table 3

Averages of four replicates, each replicate consisted of 40 seeds. Averages followed by the same letter η the column do not differ by the Scott Knott test at 5% error probability. Averages without letters in the column do not differ. YM: yeast manitol.Trp.: tryptophan, %CV= coefficient ofvariation percentage.

However, treatments inoculated with isolate UFRGS Lc348 in a greenhouse were the worst, including control treatment / N. Indole substances produced by VP16, much lower than the one with the isolated UFRGS Lc348 enriched with tryptophan, did not affect the in vitro growth of Sudan grass. However, VP16 + Trp increased the growth of the shoot dry mass of adult plants in 13.3 %, compared to control treatment with the same dose of nitrogen (Table 4).

Table 4

Averages of four repetitions, with two plants per repetition. Averages followed by the same letter in the column do not differ by the Scott Knottiest at 5% error probability. Averages without letters in the column do not differ. For statistical analysis, the variables number of tillers per plant and number of leaves per plant were transformed according to the formula: ((X + 1)^0,5). %CV: percent coeficiente ofvariation, Control N:treatment not inoculated with nitrogen dose equivalent to 150 kg.ha^-1, Control Κ N: treatment not inoculated with nitrogen dose equivalent to 75 kg.ha^-1. Values without letters in the column did not differ.

Enriching the medium with tryptophan was essential to the success of interacting VP16 with Sudan grass, as a simple inoculation with VP16 did not surpass the control treatment 1 N. The interaction between Sudan grass and rhizobia VP16 should be tested in field experiments, and, if the effects of inoculation still persists, will be possible to reduce nitrogen fertilizers without damaging Sudan grass yields.

Another important finding was the production of Sudan grass seeds 70 days after seeding. When inoculating isolate VP16 (0.41g of seeds.plant^-1), VP16 + Trp (0.88g of seeds.plant^-1) and VP16 + UFRGS Lc348 + Trp (0.84g of seeds.plant^-1). In Control treatment / N, no seeds production during the experiment could be observed.

A deleterious effect of inoculation may be associated with a large production of the hormone, which has herbicidal action when in high concentrations (Malik & Sindhu, 2011; Senthilkumara, Madhaiyan, Sundaram & Kannaiyan, 2009). On the other hand, Schlindwein, Vargas, Lisboa, Cruz, Eichelberger, Gabiatti, Prates & Stumpf (2008), observed similar results in lettuce seedlings, in which the rhizobacteria inoculation able to synthesize a great amount of indole substances damaged the crop yield.

These results open a range of new hypotheses to further studies, for example, the exudation of tryptophan by Sudan grass roots. There are some plants able to produce tryptophan, which under field conditions, could be used by rhizobia in IAA synthesis. Kravchenko, Azarova, MakarovaI & Tikhonovich (2004), reached a similar result, when they quantified tryptophan exudation of tomato and radish aseptic roots. A tomato plant had released up to 5.3 ng of tryptophan daily, while radish released up to 390 ng of tryptophan daily. In the same study, authors observed that radish root mass increased around 36% to inoculate Pseudomonas growth promotion rhizobia, while inoculating the same bacteria in tomato, which was inefficient.

Yanni et al. (2001), and Osorio et al. (2014), reported grass growth promotion by using phyto-stimulating substances for rhizobia inoculation. If field conditions present the same results, effects of these inoculations may take into account stop using nitrogenous fertilizers without damaging crop yields. These mechanisms are additional tools to biological nitrogen fixation in legumes and should be considered for better rhizobia integration in intercropping or crop rotation systems.

Despite having a high variance associated to seed production parameter, rhizobium VP16 induced certain precocity of Sudan grass, which can be interesting in seed or Sudan grass silage production systems.

The next step in compatibility tests between pastures and plant growth promotion rhizobia should be pasture stimulus field of Sudan grass through inoculating VP16. Field conditions possibly provide tryptophan in an amount sufficient to synthetize indole substances due a complex system enriched with root exudates and decomposing tissues. Thus, there can be an increasing in Sudan grass plantation after the inoculation with VP16. Without mineral fertilization, is certainly interesting for the sustainability of pasture management systems. As it increases green mass production without increasing the costs with mineral fertilization, which reduces the risk of loss and environmental pollution by using nitrogen fertilizers.

Inoculation of millet BRS1501 with isolate VP16 cultivated in medium with and without tryptophan induces higher seedling growth, while isolate UFRGS Lc348 induces greater length for the hypocotyl of millet seedlings cv. BRS1501. Despite these stimuli in the initial phase, no effect of inoculation on millet adult plants could be observed.

In Sudan grass, the inoculation of tryptophan-enriched VP16 increases the shoot dry mass of adult plants.