Introduction

Brazil is the world's largest sugarcane producer, having harvested 665.1 million tons of sugarcane in the 2020/21 harvest. The Southeast country's main producing region reached 436 million tons harvested, remarkably in São Paulo and Minas Gerais states. Its appreciation is associated with the biofuels sector, in which its main by-products are total ethanol (29.8 billion liters), anhydrous ethanol, hydrated ethanol, and sugar (41.8 million tons) (CONAB 2020).

Sugarcane crop had shown a decline of productivity affected by the increase in harvest cycles associated with abiotic and biotic factors, including the enhanced incidence of systemic pathogens, emphasizing the bacterium Leifsonia xyli subsp. xyli (Lxx)—the causal agent of ratoon stunting disease (RSD) (Davis et al. 1980, 1984; Evtushenko et al. 2000). The Lxx bacterium is fastidious, gram-positive, coryneform, without flagella, aerobic and pleomorphic. Cultivated in vitro, it produces tiny and translucent colonies after 14 days of cultivation in Soybean Corn medium (Gillaspie et al. 1981; Cardoso 1986). The disease has no characteristic symptoms; the bacterium colonizes the xylem of the plant, obstructing the translocation of water and nutrients, especially under conditions of water restriction (Hughes and Steindl 1956; Davis and Dean 1984). Quecine et al. (2016) demonstrated that colonization of Lxx also occurs in mesophyll cells and cells in the vascular bundle's sheath. With successive harvests in susceptible varieties, we observe a delayed growth of cane sprouting, stunted stalks, and decreased production. In extreme cases, it evolves to plant death (Ricaud 1968). The symptoms are nonspecific, with cross-section stalks showing reddish-orange dots or lines in the nodal regions. They also arise with the attack of other diseases, insects, environmental factors, and mechanical damage (Gillaspie and Teakle 1989). Therefore, the diagnosis of RSD can only be confirmed through methods of detecting the presence of Lxx in the xylem sap, through serological tests (Carneiro et al. 2004; Urashima and Grachet 2012), which is a time-consuming method and does not detect low bacterial concentrations; or through molecular tests (Sun et al. 2019), which have greater sensitivity and specificity, however, are of higher costs (Wu et al. 2018) and do not detect only active colonized stalks.

RSD is spread rapidly through cutting tools during manual and mechanized cane harvesting (Damann 1992). The RSD control strategies aim to prevent the initial inoculums’ entry into the production areas, especially through heat treatment of setts or bud setts (50.5 ºC/120 min or 52 ºC/30 min). However, these seedling treatments do not eliminate Lxx (Fernandes Jr. et al. 2010; Urashima and Grachet 2012; Dias et al. 2019). Another auxiliary method, micropropagation by a meristem culture (Sreenivasan and Sreenivasan 1984; Muthukumarasamy et al. 2006), also does not guarantee the total elimination of the bacteria. Already, the use of resistant/tolerant varieties comes up against the limited availability of studies that prove such mechanisms (Gagliardi and Camargo 2009; Young 2016; Fu et al. 2019), although there is genetic variability among the clones and progenitors of sugarcane sugar in Brazil. Finally, it is recommended to reduce the spread by disinfesting agricultural equipment, using bactericidal agents, or heat, which is rarely practiced in the field due to logistical issues (Urashima et al. 2020).

In this sense, biocontrol can be used as an alternative strategy to minimize the negative impacts caused by the disease. Endophytic bacteria, such as Gluconacetobacter diazotrophicus, can compete for the same colonization sites of the pathogen, stimulate resistance mechanisms in the host and produce bacteriocins (Arencibia et al. 2006; Blanco et al. 2010; Oliveira et al. 2018). Gluconacetobacter diazotrophicus, Herbaspirillum seropedicae and Burkholderia ambifaria applied on potatoes and tomatoes in a greenhouse pot experiment were tested against Fusarium oxysporum f. sp. radicis-lycopersici and Rhizoctonia solani. The consortium was able to successfully counteract the infection of both fungal pathogens in pre-emergence (infection before germination) and post-emergence (infection after germination) (Pellegrini et al., 2020). Besides, G. diazotrophicus, H. seropedicae, H. rubrisubalbicans, Azospirillum amazonense and Burkholderia tropica colonize sugarcane roots, stalks and leaves and have been used as part of the bacterial consortium developed by the Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA) and recommended as a commercial inoculant for sugarcane due to its plant growth promoting characteristics (Reis et al. 2008; Chaves et al. 2015; Dos Santos et al. 2019).

The hypothesis of RSD biocontrol by endophytic bacteria opens the perspective of disease control, aiming at endogenous reduction of the pathogen. Besides, endophyte inoculation can result in higher productivity and lower cost for producers, especially in varieties with low resistance and/or tolerance to RSD, thus justifying the investment in seedlings treated by thermotherapy, micropropagation, or pre-inoculated with diazotrophic endophytes. In this context, the objective of the present study was to evaluate, under field conditions, the vigor, productivity, and health of sugarcane micropropagated plants inoculated with the endophytic diazotrophic bacteria Gluconacetobacter diazotrophicus, Herbaspirillum seropedicae, Herbaspirillum rubrisubalbicans and evaluate the potential of this strategy in controlling RSD, by decreasing the incidence of stalks seropositive for the presence of the pathogen Lxx on plant cane.

Materials and methods

Production of healthy Co 421 seedlings

The micropropagated sugarcane seedlings of the Co 421 variety were produced at the Meristem Culture Laboratory of the Dr Leonel Miranda Campus (UFRRJ), in which the setts were previously heat-treated at 50.5 ºC for 2 h and multiplied by meristem culture (Hendre et al. 1983). Part of the seedlings was transferred to glass flasks containing 50 mL of modified MS medium (diluted ten times, without the addition of hormones) (Murashige and Skoog 1962) and inoculated with endophytic bacteria, remaining in these conditions for seven days until transplantation nursery (Reis et al. 1999). The control condition was established without inoculation of the bacteria.

Multiplication and inoculation of the endophytic diazotrophic bacteria Gluconacetobacter diazotrophicus, Herbaspirillum seropedicae and Herbaspirillum rubrisubalbicans

The endophytic bacteria Gluconacetobacter diazotrophicus strain PAL5 (Gd) (Gillis et al. 1989), Herbaspirillum seropedicae strain HRC54 (Hs) (Baldani et al. 1986) and Herbaspirillum rubrisubalbicans strain HCC103 (Hr) (Baldani et al. 1996) were grown separately in liquid DIGS medium (Döbereiner et al. 1995) and incubated in a rotatory shaker for 24 h, at 30 ºC, under agitation at 175 rpm, obtaining a final bacteria cell suspension of 108 cells.mL−1. The mixed inoculum containing the three bacteria species was added to the seedlings in a modified MS medium (100 µL of inoculum/seedling/flask). After seven days of inoculation, the seedlings were transplanted in trays of 128 cells containing filter cake substrate, sugarcane bagasse, and soil (1:1:1) and kept in a screened nursery (50% shade) for 20 days. Then, the seedlings were transplanted in open-air beds for acclimatization for 200 days.

Pathogen inoculation

The bacterium Leifsonia xyli subsp. xyli (Lxx) (Carneiro et al., 2004) was inoculated into the seedlings eighty days before planting in the field. The inoculation was done by pruning the basal leaves with scissors pre-immersed in sap contaminated with the pathogen. The sap of infected stalks that resulted in a seropositive reaction for the presence of pathogen was obtained from the CB 49–260 variety, highly susceptible and proven to be infected by Lxx with at least 106 bacteria.mL−1, according to "Dot Blot" serological sensibility (Carneiro et al. 2004). This sap was used to inoculate all the seedlings.

Planting and experimenting in the field

The experiment was installed at the Dr Leonel Miranda Campus (UFRRJ), in a randomized block design, with four replications and nine plots each, arranged in lines of 8 m, the spacing between lines of 1.40 m and between plants of 0.5 m, 2 m firing line and distal borderline. Field planting was established with seedlings of approximately seven months. During the conduct of the experiment, irrigation or fertilization was not performed. Each treatment consisted of approximately 150 seedlings of the Co 421 variety, distributed in 1 - G. diazotrophicus (Gd); 2 - H. seropedicae (Hs); 3 - H. rubrisubalbicans (Hr); 4 - L. xyli subsp. xyli (Lxx); 5 - Gd + Lxx; 6—Hs + Lxx; 7 - Hr + Lxx; 8 - Gd + Hs + Hr + Lxx; 9 -Control (micropropagated seedlings without inoculation of bacteria).

Estimation of the endophytic bacteria population associated with the plant tissue and xylem sap

At 42 and 86 days after inoculation of the endophytic bacteria, five seedlings from each treatment were collected to quantify the population of the bacteria G. diazotrophicus, H. seropedicae, and H. rubrisubalbicans. For this purpose, 1 g of root or 5 g of leaves and stalks were washed in running tap water and partially disinfected with alcohol-soaked cotton. The roots, leaves and stalks samples were macerated in saline and diluted in series (10–2 to 10–7). Then, 100 µL of each dilution was applied in semi-solid medium LGI-P (G. diazotrophicus selective medium) and JNFb (Herbaspirillum spp. selective medium), in triplicate (Baldani et al., 2014). The Most Probable Number (MPN) of bacteria was estimated according to the McCrady Table, based on the absence or presence of a white surface pellicle into the glass vial containing the semi-solid medium, which represents the positive growth for diazotrophic bacteria (Döbereiner et al. 1995). To confirm the inoculated bacteria's identity, a platinum loop of the bacterial pellicle of the last positive dilution from JNFb and LGI-P was transferred to a new semi-solid medium. The new-formed pellicle was streaked in JNFb and LGI-P solid medium plates containing 20 mg.L−1 of yeast extract as a nitrogen source. Typical Herbaspirillum white colonies with greenest to the blue centre were transferred to JNFb semi-solid medium to confirm the diazotrophic ability by pellicle formation (Matteoli et al. 2020). Then, a portion of cells was harvested with a platinum loop, mounted in a glass slide with a coverslip and examined under a phase-contrast light microscope for the presence of small, curved rods that are fast-moving when near air bubbles. The species level confirmation was possible by contrasting the carbon source used by the bacterial re-isolates, where H. seropedicae strain HRC54 could use N-Acetyl-D-glucosamine and Myo-inositol and unable to use Meso-erythritol. H. rubrisubalbicans strain HCC103 showed an opposite pattern (Matteoli et al. 2020). Almost the same procedure was adopted for G. diazotrophicus strain PAL 5 new-formed pellicle streaked in LGI-P solid medium plates. Typical orange-coloured colonies were transferred to LGI-P semi-solid medium to confirm pellicle formation. Under a phase-contrast light microscope, Gd cells are small immobile rods (Baldani et al. 2014).

To detect the presence of endophytic bacteria in the xylem sap, a composite sample consisting of 10 stalks per plot was extracted at harvest after 16 months in the field. The xylem sap samples were obtained as described by Carneiro et al. (2004) and diluted from 10–1 to 10–4 in saline and applied in LGI-P and JNFb semi-solid medium in triplicate. The MPN was estimated as previously described.

Evaluation of agronomic characteristics

Thirty days after planting in the field, the percentage of seedling establishment, the number of seedlings, the height (cm), and the average diameter (cm) of tillers were evaluated. After 16 months of planting in the field, the experiment was harvested, first counting the number of stalks per plot (NSP). Then, 30 stalks in the sequence were removed from the center of each plot, which were used for the following evaluations: average stalk diameter (cm), measuring 10 stalks per plot at the height of the fifth basal internode with a calliper; average stalk weight (ASW) (g) estimated by the relationship between the weight of 30 stalks /30; the share weight (g), estimated by multiplying the ASW and the NSP; the Brix grade, estimated with a field refractometer and sampling five older stalks of the plot, at the height of the fifth basal internode; tons of cane per hectare (TCH) estimated by calculating the useful area of the extrapolated plot to 10,000 m2; and tons of Brix grade per hectare (TBH) estimated using the equation TCH*Brix/100.

Estimation of stalk incidence with RSD by serology

Sampling was carried out in the experimental plots, randomly removing 10 stalks per plot. Samples of xylem sap were taken from the third basal internode of the stalks for use in the "Dot Blot" serological test, as described by Carneiro et al. (2004). The mean incidence of the disease in the plot was estimated by the incidence of stalks (IS) that presented seropositive reactions for Lxx estimated by the formula: (number of positive reactions/number of samples)*100.

Statistical analysis

The results obtained from the agronomic characterization and the estimation of stalk incidence with Lxx were subjected to analysis of variance (ANOVA), followed by Scott and Knott means cluster test, at 5% probability, using the statistical program Genes (Cruz 2001). The results were presented as mean ± standard deviation. The graphs were created using the statistical program GraphPad Prism 5.0.

Results

Estimation of the endophytic diazotrophic bacteria associated with micropropagated seedlings

The MPN estimated the density of the endophytic bacteria (Table 1) in a semi-solid medium. At 42 days, G. diazotrophicus was higher in the root than in the aerial part in the treatments in which it was inoculated separately. H. seropedicae was recovered from the shoot in the Hs treatment and the root in the treatment Hs + Lxx treatment. On the other hand, H. rubrisubalbicans showed high density in the root in the Hr-treatment and the aerial part in the treatment combining Hr + Lxx. At 86 days, the density of G. diazotrophicus was also higher in the Gd-treatment, but there was a reduction in the population size related to the first sampling time for both treatments, in the root and the aerial part. There was a recovery in both tissues for H. seropedicae bacteria and H. rubrisubalbicans, regardless of inoculation with Lxx.

Table 1 Bacterial density (cells/gram of fresh biomass) in the root and shoot of sugarcane plants, at 42 and 86 days, estimated by the most probable number (MPN)
Full size table

For the treatment that received the mixed inoculum, it was observed that the population of G. diazotrophicus and Herbaspirillum spp. were not detected in the root and that only the population of Herbaspirillum spp. was detected in the aerial part at 42 days. On the other hand, at 86 days, G. diazotrophicus and Herbaspirillum spp. were detected only at the root. In the first isolation, at 42 days, microorganisms were detected both in LGI-P and in JNFb in some plots inoculated only with Lxx and in the control's JNFb. In the second isolation, at 86 days, the presence of bacteria in the Lxx treatment was not quantified, and only a small amount at the root in LGI-P medium.

In isolation performed at 16 months, a low population density of endophytes was obtained from the raw xylem sap. In the Hr treatment, low populations were detected in all blocks, ranging from 0.4 × 102 to 2.5 × 102 bacterial cells/mL of raw cane sap. In a single portion of the Hr + Lxx treatment, 2.5 × 102 bacterial cells/mL was detected. In control, only in block 4, Herbaspirillum was detected at 9.5 × 102 bacterial cells/mL and Gluconacetobacter 4.5 × 102 bacterial cells/mL.

Seedling vigor after 30 days of planting in the field

After transplanting to the field, the parameters evaluated at 30 days demonstrated that the height and average diameter of the tillers per plot showed better results for seedlings inoculated with exclusively G. diazotrophicus, while no significant differences (P ≤ 0.05) were found for the other treatments (Table 2).

Table 2 Results of the number, height (cm) and average diameter (cm) of tillers of sugarcane seedlings inoculated with endophytic bacteria and Lxx 30 days after transplanting to the field
Full size table

Agronomic characteristics of sugarcane plants

Among the agronomic parameters analyzed at the end of the harvest, only the number of stalks per plot and average stalk diameter did not show significant differences (P ≤ 0.05) between treatments (Table 3).

Table 3 Results of the characteristics associated with the vigor and productivity of the sugarcane harvest at 16 months
Full size table

For the average stalk weight, the best results were obtained from the treatments: Gd, Hr, Gd + Lxx, Hs + Lxx and Hr + Lxx. As for the stalk weight per plot, the treatments that stood out were Gd, Hr + Lxx and Hs + Lxx. In both cases, the treatment Lxx and the control presented values below the average – 1.63 and 150.42 kg, respectively. Regarding the Brix grade, the treatment in which all bacteria were inoculated differed significantly from the other treatments. The worst Brix grade was obtained from the Gd treatment, below the mean of 16.67 (Table 3).

Regarding the productivity of cane-plant in tons/hectare, at 16 months, it was found that for TCH the treatments Gd, Hr + Lxx, and Hs + Lxx stood out from the others. It is noteworthy that the Gd treatment exceeded the control by 68.5% and that the treatments Hr + Lxx and Hs + Lxx were 41.2% and 44.4% higher, respectively, superior to the treatment inoculated only with the pathogenic bacterium Lxx. Such results are equivalent to 46.6 t of stalks/ha, on average (Fig. 1).

Fig. 1
figure 1

Comparison of production in tons of cane per hectare (TCH), treatments inoculated with endophytic bacteria and with Lxx, and not inoculated. Averages ± standard deviation followed by the same letter are not significantly different at P < 0.05. Gd—G. diazotrophicus; Hs—H. seropedicae; Hr—H. rubrisubalbicans; Lxx—L. xyli subsp. xyli

Full size image

For TBH, the treatments Gd, Hr, Hr + Lxx, Hs + Lxx, Gd + Hr + Hs + Lxx are significantly higher than the others (P ≤ 0.05), presenting higher values than the general average (22.28 t of Brix/ha) (Fig. 2). Together, these treatments exceeded 7.26 t of Brix/ha of plants inoculated with Lxx only and 8.16 t of Brix/ha of plants used as control.

Fig. 2
figure 2

Comparison of production in tons of Brix per hectare (TBH) of treatments inoculated with endophytic bacteria and with Lxx and not inoculated. Averages ± standard deviation followed by the same letter are not significantly different at P < 0.05. Gd—G. diazotrophicus; Hs—H. seropedicae; Hr—H. rubrisubalbicans; Lxx—L. xyli subsp. xyli

Full size image

As for the incidence of stalks positive for Lxx, in cane-plant, the highest incidences were observed in treatments Gd, Hs, Gd + Lxx, Hs + Lxx, Hr + Lxx, and the control. The lowest incidences were observed in the treatments Hr, Lxx, and complete inoculum. The average incidence in cane-plant, was 16.39%, with the highest incidence detected in the Hs + Lxx treatment, with 32.5% (Fig. 3). The Gd, Hs and Hr treatments showed few strong positive reactions to RSD, while Hr, Lxx, and complete inoculum treatments showed very weak reactions.

Fig. 3
figure 3

Percentage of incidence of seropositive stalks for Lxx, in cane-plant, at 16 months, after transplanting micropropagated seedlings and inoculated with endophytic bacteria and Lxx and not inoculated. Averages ± standard deviation followed by the same letter are not significantly different at P < 0.05. Gd—G. diazotrophicus; Hs—H. seropedicae; Hr—H. rubrisubalbicans; Lxx—L. xyli subsp. xyli

Full size image

Discussion

The isolation procedure of endophytes from roots and shoots aimed that the bacteria inoculated have colonized the micropropagated seedlings before planting in the field, in two stages. The isolated inoculation of G. diazotrophicus revealed population densities greater than 107 cells per plant, 5 days after inoculation. However, its population has decreased about 10 times in a mixed inoculation, even though it can colonize the sugarcane tissue and other diazotrophic ones, such as H. seropedicae (Oliveira et al. 2009). This reduction was also observed in our results. Regarding Herbaspirillum spp., both species showed similar population sizes associated with the plant host; however, when inoculated together, it was not possible to identify which species was more prevalent, based on the methods that we used. In general, there was a population reduction of endophytes after 86 days, compared to 42 days after inoculation.

The detection of endophytic bacteria in the xylem sap, at 16 months, in which the stalks were mature, was reduced to a few plots, in which small densities of Herbaspirillum spp. and G. diazotrophicus, possibly because the number of endophytic bacteria in the xylem vessels, in the final stage of maturation of the sugarcane stalks, was reduced or absent. Baldani et al. (1997) mentions that H. seropedicae colonizes xylem vessels in small cell densities, while H. rubrisubalbicans colonizes the vessels in high concentrations (Olivares et al. 1997). Reis et al. (2000) reports a reduction in the G. diazotrophicus population size at the end of the culture cycle, from 15 months onwards, the same not being observed for Herbaspirillum spp. On the other hand, the pathogen population increases with maturation in the xylem vessels, with basal internodes having the greatest concentration (Harrison and Davis 1988). Thus, it was expected that the population of endophytic bacteria, in the xylem sap, at 16 months would be low. However, H. rubrisubalbicans in the inoculated plants showed that this bacterium also colonizes the xylem vessels in the basal internodes of sugarcane stalks and could compete with Lxx for the same colonization sites.

Regarding the important characteristics for agricultural production (number of tillers; stalk height and diameter up to 30 days; the number of stalks per plot (NSP); average stalk diameter; average stalk weight (ASW); plot weight; Brix grade; tons of cane per hectare (TCH); and tons of Brix per hectare (TBH)), a positive effect was observed mainly for G. diazotrophicus. When inoculated separately, the vegetative vigor induced by this bacterium led to an excessive sprouting of tillers with different maturation ages and a greater weight of the total portion, consequently causing a lower Brix degree. Even so, its inoculation produced higher TCH and TBH. The association between Gd + Lxx, on the other hand, resulted in an increase of 17.5% in HCT and 16.8% in TBH compared to treatment with Lxx alone. The high incidence of RSD in plants inoculated with Gd and Gd + Lxx demonstrated that the clonal cleaning recommended by heat treatment and meristem culture does not guarantee Lxx seedlings' cleaning/or did not prevent further reinfection. Despite this, the gains with vigor and productivity of plants inoculated with Gd, outweighed the null effects of the presence of Lxx until the time of harvest. This response is due to Gd's effects at the beginning of the plant's growth, before the increase in the population density of Lxx, which occurs with the maturation of the stalks.

The inoculation with H. seropedicae harmed the characteristics associated with productivity compared to plants inoculated with Lxx and control. Olivares et al. (1997) report different interaction levels between bacteria and plant species or an affinity between strains and cultivars. This fact explains, in part, the low initial development that later resulted in low agricultural yields in plants with Hs. H. seropedicae can also induce hypersensitivity reactions in sugarcane leaves (Olivares et al. 1997). However, in this work, the cultivar used Co 421 did not show any symptoms of mottled stripe disease, only small chlorosis in the leaves and reduced initial development in the inoculated plants. However, there is no information on disease symptoms with the inoculation of Hs or Hr in the cultivar Co 421. It was also observed that the association of Hs + Lxx promoted gains in TCH and TBH in comparison with the control and the plants inoculated with Lxx only, revealing an intriguing positive effect of the microorganism-plant interaction. The competition between bacteria (Hs and Lxx) may have stimulated the greater activity of H. seropedicae, benefiting the host plant through mechanisms of biological nitrogen fixation and production of phytohormones, for example (Serrato et al. 2010; Monteiro et al. 2012). Similary, Dall'Asta et al. (2019) evaluated the effects of H. seropedicae SmR1 strain on the maize growth and leaf anthracnose (Colletotrichum graminicola) of plants. Although the bacterium has efficiently colonized the leaf tissues and promoted maize growth, it did not affect the disease severity in leaves. Although Hs does not affect the colonization of Lxx, as the infected stalk incidence was high in plants inoculated with Hs and Hs + Lxx, there was better yield in the second treatment than treatment with the pathogen alone.

The performance of H. rubrisubalbicans in plants inoculated with Hr + Lxx was superior to that of plants inoculated with Hr only in the parameters of ASW, plot weight, and TCH, indicating a positive interaction between these bacteria, as observed for Hs + Lxx. On the other hand, unlike Hs, the isolated Hr bacterium positively affected TBH and had a lower incidence of seropositive stalks for Lxx. Hr's small densities were detected at 16 months, indicating that this bacterium was present in the xylem sap and can colonize its vessels, competing in some way with Lxx. It is noteworthy that H. rubrisubalbicans can establish beneficial non-pathogenic interactions with Poaceae, but it is also capable of causing disease in some susceptible sugarcane varieties (Schmidt et al. 2012), which was not observed for Co 421.

The effects of the association of Gd + Hs + Hr + Lxx bacteria on the Brix degree and TBH were positive, surpassing the plants inoculated with Lxx (17.2 and 36.2%, respectively) and the control (18.8 and 43.3%, respectively). These microorganisms (Gd, Hs, and Hr) were not isolated from the xylem sap at 16 months, and this treatment may be correlated with the low population density of endophytic bacteria recovered during the initial isolations. On the other hand, this treatment had a low incidence of stalks positive for Lxx. Considering that greater microbial diversity is associated with a lower individual population density due to the greater competition for nutrients, space, and oxygen (Maron et al. 2011; Van Elsas et al. 2012), the endophyte bacteria, together, promoted gains in the quality of sugar cane. G. diazotrophicus, H. seropedicae, H. rubrisubalbicans, A. amazonense and B. tropica, both individually and in a mixture, in the absence of pathogens, promote biomass gain, increase in nutrient absorption and changes in root architecture in two sugarcane cultivars (cv. RB867515 and IACSP95-5000) (Dos Santos et al. 2019).

Plants inoculated in isolation with Lxx had low agricultural yield, being below the general average for most of the evaluated parameters (except tillers height, at 30 days and average stalk diameter, at 16 months) similar as results obtained with the control - o that was not expected. Zhu et al. (2018) obtained a reduction of 12.8% in plant height, 14% in stalk diameter and 12.1% in stalk weight, on average, after 210 days of Lxx inoculation, compared to healthy plants. The low incidence of seropositive stalks with Lxx in treatments with Lxx was surprising since the inoculation of the plants was carried out simultaneously and place and with the same inoculum as the other treatments. The hypothesis that best justifies the low incidence of Lxx in the plots inoculated with the pathogen would be the variation in the explants' health used in vitro multiplication. Despite this, the low agricultural yield associated with the low incidence of RSD indicates that even small Lxx concentrations can reduce production in a variety susceptible to RSD, such as Co 421. The micro-propagated plants not inoculated (control) showed low agricultural yield and, even high incidence of sera-positive stalks for Lxx, which confirms that thermotherapy at 50.5 ºC/120 min/sample of buds, associated with the culture of meristem does not promote the eradication of RSD bacteria on xylem or sap (Damann and Benda, 1983).

Together with the plant's morphological and physiological changes, Lxx negatively regulates the cell cycle of young plants, affects hormonal defense mechanisms and modulates the activity of antioxidant enzymes, both in varieties considered susceptible and resistant to RSD (Zhang et al. 2016; Cia et al. 2018; Fu et al. 2019; Faria et al. 2020). Plants infected with Lxx have lower indole-3-acetic acid (IAA) and gibberellic acid (GA3) content and increased abscisic acid (ABA) after 180 days of emergence (Zhang et al. 2016). Considering that the hormonal balance is also important for plant growth, the endophytes used in this study may have acted in the regulation of this balance, considering that they are capable of producing phytohormones (Monteiro et al. 2012), justifying the gains in productivity before the effects of the high incidence of seropositive stalks and suppressing the negative effects of Lxx-induced RSD.

It is estimated that the RDS reaches 25 to 85% of Brazilian crops, 10% of commercial crops in the Center-South of Brazil and that 32% of the plants monitored have been using contaminated matrices (Ponte et al., 2010; Urashima et al. 2010), representing an annual economic loss of US $ 1 million (Urashima et al. 2017). In this sense, considering that the successive cuts of the cane promote a natural decline in harvest productivity even in the absence of Lxx, the gains with the inoculation of Gd in cane-plant are justified by themselves, since we obtained a 19% increase in TCH in comparison with the control treatment and the inoculated with Lxx. Although the effect on increasing productivity is reduced in successive cane crops and the negative effect of Lxx is cumulative, Gd inoculation has a significant impact on the economy. Suppose the same responses (results) were obtained with other varieties in Brazil. In that case, we can estimate that this increase in biomass in plant cane can increase gains by $ 209.76/ha, equivalent to $ 1.8 millions per harvest (according to CONAB 2020/2021, Brazil harvested 8,605 thousand ha, productivity was 77 t/ha, and the average value of cane delivered to the field was $ 14.34 – considering the current dollar exchange rate). Studies are suggested to evaluate the plant's genotypic effect on the interactions with diazotrophic bacteria in promoting growth, suppression of symptoms, and damage induced by inoculation or infestation by Lxx in sugarcane seedlings. Additional research is also needed to evaluate the effect of these endophytes on promoting plant growth, tolerance to RSD and on the spread of the pathogen in stalks after conduction of the next crop season (ratoon cane), under representative commercial crop practices and according to the harvesting method, if manual or mechanized.

Conclusions

The bacterium G. diazotrophicus provided the greatest gains in tons of cane per hectare (68.6%) and tons of Brix per hectare (59.1%). In the combined treatments of endophytes with the pathogen there was an improvement in agronomic yield, in sugarcane plant with at least 18.83 t of cane/ha, in comparison with the control, and 19.09 t of cane/ha in comparison with the treatment with Lxx, despite positive serological reactions in the plots inoculated with endophytes. The treatment with H. rubrisubalbicans alone showed a low incidence of Lxx and good agronomic yields, suggesting some negative effect of the pathogen's colonization in the xylem of cane-plant for the variety Co 421.