Yield of sugarcane varieties and their sugar quality grown in different soil types and inoculated with a diazotrophic bacteria consortium

Abstract Sugarcane is a crop of great economic, social, and environmental relevance in Brazil. The country is the largest sugar producer and the second largest bioethanol producer in the world. The goal of this study was to evaluate the efficiency of a sugarcane inoculant composed of five diazotrophic bacterial strains, as well as nitrogen fertilization of two sugarcane varieties. Two experiments were carried out on two varieties using an experimental design composed of complete randomized blocks in a factorial of two varieties and three treatments with four replicates. The treatments can be described as: inoculation with the consortium of five diazotrophic strains, or N fertilization with 120 kg ha−1, and one control treatment. The following parameters were then evaluated: stem yield, accumulation of total dry matter, nitrogen content, quality of the sugarcane juice, and 15N natural abundance on flag-leaves. Inoculation and N fertilization on the Sapucaia plantation promoted increases of stem yield equivalent to 22.3 and 26.5 Mg ha−1 in the RB867515 variety, in comparison to the control, respectively. Inoculation and N fertilizer used for the Coruripe plantation increased stem yield of 38.0 and 42.4 Mg ha−1, respectively, with the RB867515 variety, while RB72454 showed increases of 16.7 and 37.5 Mg ha−1, both compared to the control. Biological nitrogen fixation was not affected by the treatments, however, both treatments increased the total recoverable sugar yield. Benefits from inoculation appeared to promote plant growth due to the plant–bacteria interaction.

Sugarcane (Saccharum spp.) is the most important crop for Brazilian production of renewable energy. In recent years it has expanded to new agricultural areas, especially on degraded pastures, decisively influencing the economic, social, and cultural development of most regions. According to the Brazilian Geography and Statistics Institute (IBGE-SIDRA, 2015), 10.3 million hectares were harvested during the 2014/2015 crop, with an average yield of 74.3 Mg ha −1 and total production of 755 million tons. Despite all the benefits from using renewable energy, the sugarcane industry still causes several negative environmental impacts, especially greenhouse gas emissions, such as CO 2 , CH 4 , and N 2 O which result from the use of nitrogen fertilizers applied to the soil surface after harvesting Denmead et al., 2010). Thus, it is crucial to develop research in order to search for alternatives, ensuring a competitive and sustainable sugarcane industry.
Interactions between atmospheric N 2 -fixing bacteria and sugarcane have been investigated by different researchers working with such crop since 1950s (Döbereiner, 1992). Various studies to quantify the contribution of biological N 2 fixation (BNF) to the crop have shown that approximately 50% of the total N accumulated in plant tissues comes from the air Lima et al., 1987;Urquiaga et al. 2012;Yoneyama et al., 1997). An inoculant composed of five diazotrophic strains was developed, in which the BNF contribution was of approximately 30%, using the SP701143 sugarcane variety grown in pots with inoculated soil containing a consortium of N 2 -fixing bacteria . Field inoculation experiments with this consortium also showed yield increases, however, they were dependent on soil type and sugarcane variety (Oliveira et al., 2006). More recently, field inoculation studies showed that yield response varied with sugarcane varieties, and was due to both increases in the BNF input and the bacterial plant growth promotion effects (Schultz et al., , 2014. Several studies have shown that effects that promote growth due to diazotrophic bacterial inoculation are partly explained by phytohormone synthesis (especially 2003) and the species names are: Gluconacetobacter diazotrophicus (strain PAL5 T -BR11281) first described by Cavalcante and Döbereiner (1988), Herbaspirillum seropedicae (HRC54-BR11335 - Baldani et al., 1986), Herbaspirillum rubrisubalbicans (HCC103-BR11504 - Baldani et al., 1996), Nitrospirillum amazonense (Cbamc-BR11145) formerly Azospirillum described by Magalhães et al. (1983) and renamed by Lin et al. (2014); Paraburkholderia tropica (PPe8 T -BR11366) described by Reis et al. (2004) in the genus Burkholderia and renamed as new genus by Oren and Garrity (2015).
All strains were individually grown in DYGS liquid medium (Baldani et al., 2014) at 30 °C for 48 h in a rotary shaker at 150 rpm. Subsequently, the five bacterial cultures containing 10 9 cells mL −1 were mixed with the peat used as carrier in a proportion of 75 mL of DYGS culture medium mixed in 175 g of neutralized and sterilized milled peat packed in polyethylene bags. The total inoculum was composed of five packages of 250 g fresh weight containing each single strain. Three inoculant doses were diluted in 600 L of tap water and used to immerse sugarcane setts. The prepared inoculant suspension corresponded to a dose for one hectare. Pre-selected and standard size setts (with three buds) were packed for inoculation into bags according to the number of setts required per plant row (15 buds per meter), and immersed during 30 min in the inoculum suspension (600 L). Seedlings were then dried auxins) (Bashan et al., 2004;Videira et al., 2012), inorganic phosphate solubilization (Shukla et al., 2008;Singh et al., 2007), zinc compounds dissolution (Saravanan et al., 2007), increased rhizosphere retention of essential nutrients (Yadav et al., 2009), as well as biological control of plant pathogens (Spaepen et al., 2007).
The goal of this study was to evaluate the inoculation effect of the consortium composed of five diazotrophic strains over agronomic and industrial parameters of two varieties of planted sugarcane in two different soils and climate conditions, in the main sugarcane Brazilian producing regions.

General information
Tables 1 and 2 present the geographical location, soil and climate data, as well as soil.

Experimental design and varieties
Experiments were laid out in a complete randomized block factorial 2 × 3 (two varieties and three treatments) design with four replicates. RB867515 and RB72454 varieties were chosen based on their agronomic importance and contribution to the total sugarcane area cultivated in Brazil. Treatments imposed were: inoculation with the consortium of diazotrophic bacteria, N fertilization using 120 kg ha −1 , and the absolute control. The planting procedure employed stem pieces (setts) since it is a standard practice, at a density of approximately 15 buds per linear meter.

Inoculant preparation and inoculation procedure
The sugarcane inoculant was composed of five strains, isolated from different sugarcane varieties in Brazil and deposited at Embrapa Agrobiology Culture Collection (Seropédica, RJ, Brazil). The five diazotrophic bacterial strains were previously selected by Oliveira et al. (2002,  in the shade for 30 min and immediately planted. Control plants were immersed in water for 30 min as well.

Assessments, preparation, and analyses of plant samples
Agronomic evaluations were performed to determine stem yield, dry matter yield, and total N in plant shoots (stems, straw, and flag-leaves). After weighing stems, straw, and flag-leaves, subsamples were taken from each fraction and dried until constant weight in an oven adjusted at 65 °C. Sub-samples were first ground in a Wiley mill (2 mm) and then similarly finely ground as described by Arnold and Schepers (2004). Nitrogen was determined according to the semi-micro Kjeldahl method (Nogueira & Souza, 2005) and 15 N natural abundance of sub-samples of flagleaves was determined, which, according to Boddey et al. (2001), represent the whole-plant isotopic label. For analyses of samples containing between 30 and 50 μg of N, tin capsules for weighing were employed in the analyses by continuous flow isotope ratio mass spectrometry (Finnigan MAT, Bremen, Germany) at the Embrapa Agrobiology 'John Day Stable Isotope Laboratory' (Ramos et al., 2001). Sugarcane juice quality (total recoverable sugar -TRS, soluble solids percentage -Brix, and apparent sucrose percentage -Pol), as well as TRS yield (kg per hectare) were determined according to the methodology proposed by Fernandes (2000) and the Instruction Manual for the São Paulo State Producers Council for Sugarcane, Sugar, and Alcohol (CONSECANA, 2006).

N natural abundance for available N in soil
Uniformity for 15 N isotopic labeling within the total N available in the soil for plants, considering the time and depth, is a pre-requirement for the 15 N isotopic dilution technique to assess BNF, provided that control plants have similar N uptake and roots as the target plant (Unkovich et al., 2008). However, Ledgard et al. (1984) and Urquiaga et al. (2012) found that 15 N isotope distribution of the N available in the soil is not uniform according to the depth, making it difficult to compare the plant 15 N natural abundance due to different root systems and consequent withdrawal of the N available in the soil at different depths.
Based on assumptions made by Unkovich et al. (2008), uniformity of the 15 N isotope of N available in the soil was assessed by following the methodology recommended by Ledgard et al. (1984) and Urquiaga et al. (2012). Soil samples were collected at three points from each experimental area in layers described as 0-15, 15-30, 30-45, and 45-60 cm (depth that contains most of the sugarcane root system). These samples were dried and sieved through a 2 mm mesh sieve. The soil prepared was conditioned into pots containing 400 g, as well as three reference plant species, which were non-N 2 -fixing or had an insignificant fixation. Reference plants adopted for this study were Sorghum bicolor, Panicum mileaceum, and Pennisetum glaucum. The experimental design can be described by a randomized block in a factorial 3 × 4, three species of plants, four soil depths, and four replicates. Before planting, the soil was fertilized, except for N, by applying 100 mg kg −1 of P 2 O 5 (superphosphate), 100 mg kg −1 K 2 O (potassium chloride), 20 mg kg −1 of magnesium sulphate, and 50 mg kg −1 of FTE BR12 (micronutrients). Eight to ten seeds were planted per pot, with each pot containing only one plant species. Plants were grown for approximately 30 days until they presented yellowing of leaves, indicating depletion of the available N in the soil. Whole plants were harvested (root + shoot), washed, dried, weighed, ground, and analyzed as described for sugarcane samples.
The weighted average of 15 N natural abundance in the profiles was calculated through this methodology, since the N content available in the soil decreases according to profile depth, and almost all sugarcane roots are found only to a depth of 60 cm (Urquiaga et al., 2012). Thus, estimated 15 N natural abundance of soil N available to the sugarcane plant was calculated using the following Equation: Rpδ 15 N = Reference plant δ 15 N, grown in soil samples of each layer, ‰. TNrp = Total N available in each layer (15 cm), extracted by reference plant, mg. Nfdrp = N available in full depth (0-60 cm), extracted by reference plant, mg.

Statistical analysis
At first, data obtained were statistically analyzed to verify normality and homogeneity of variance errors, by applying Lilliefors and Cochran & Bartley tests, respectively, using SAEG 9.1 software (Viçosa Federal University, MG, Brazil). Then, a variance analysis was performed, followed by application of the F test using the Sisvar 4.3 software program (Lavras Federal University, MG, Brazil). Means were compared by the Scott-Knott test at 5% probability. TRS yield in kg ha −1 was compared using the same test at 10% probability. Soil δ 15 N values which were extracted by reference plants were compared through mean standard error bars.

Results
The response of the different sugarcane varieties to inoculation with the consortium composed of diazotrophic strains, as well as the application of 120 kg ha −1 of N, was When TRS yield was expressed as kg ha −1 (Table 6), a significant response to inoculation and N fertilizer was observed for both varieties cultivated in the Coruripe mill. No difference was observed at the Sapucaia mill regarding total sugar yield (TRS) recovered from both sugarcane varieties which were inoculated or N fertilized. Despite the variability in the response of the varieties to inoculation and N fertilization, increases of 10,096 and 14,853 kg ha −1 , respectively, related to the total TRS accumulated in the plants cultivated at the two mills, were observed in comparison to the control. Real TRS gains with inoculation were 68% of that observed for nitrogen fertilization. Table 7 shows delta 15 N values (δ 15 N) for inoculated and non-inoculated sugarcane varieties with the consortium of diazotrophic bacteria. No significant differences in δ 15 N variable when such varieties were grown in two diverse climatic Brazilian regions (Table 3). An expressive increase of fresh stems productivity was observed for the RB867515 variety grown in the Sapucaia mill (Southeast region), inoculated with the consortium or N fertilized. This increase was equivalent to 22.3 and 26.5 Mg ha −1 , respectively, in comparison to the control. However, there was no difference observed between the treatments for the RB72454 variety. In addition, there were significant effects on stem yields from inoculation and N fertilizer in Coruripe mill (Northeast region), with increases of 38.0 and 42.4 Mg ha −1 , respectively, in comparison to the control for RB867515. To the RB72454, the N fertilization effect was significantly higher to stem yields, in comparison to the control and inoculation (Table 3). Total dry matter accumulated presented a significant difference between the treatments, but only for the RB867515 variety in Coruripe mill, with increases of 21.5 and 18.4 Mg ha −1 by inoculation and N fertilization, respectively, in comparison to the control (Table 3).
Total N accumulated in the RB867515 variety at both mills was not significantly influenced either by inoculation or N fertilizer. A significant effect of N fertilizer over the total N accumulated in the whole plants (stems and total dry matter) was observed for RB72454 variety grown in Coruripe mill (Table 4).
Analysis of sugarcane juice quality showed that neither inoculation with the bacteria consortium nor fertilization with 120 kg ha −1 of N influenced the TRS, Brix, and Pol for the two sugarcane varieties grown in two diverse regions ( Table 5). The results indicated that inoculation and N fertilizer did not influence sugar synthesis, despite increases of stem productivity and dry matter accumulation. Table 3. Stem yield and total dry matter accumulation (stems, straw, and green leaves) of two grown sugarcane varieties inoculated with diazotrophs, fertilization with 120 kg ha −1 of n, and control in two Brazilian regions.  Table 4. total nitrogen content in stem dry matter and total dry matter (stems, straw, and green leaves) of two grown sugarcane varieties inoculated with diazotrophs, fertilization with 120 kg ha −1 of n, and control in two Brazilian regions.
notes: Means of four replicates. Means followed by different letters in the columns differ by Scott-Knott test at 5% probability c.V: coefficient of variation.  values were observed between inoculated and non-inoculated sugarcane varieties; although δ 15 N values for inoculated and non-inoculated sugarcane varieties were lower than those observed in reference plants (Panicum, Pennisetum, and Sorghum). The isotope 15 N was not evaluated for the treatment with nitrogen fertilization, since the 14 N/ 15 N ratio is altered by the introduction of the synthetic nitrogen fertilizer, being thus cannot be used for the calculation of BNF. Soil profiles for both regions showed an increase in 15 N natural abundance down to the layer between 15 and 30 cm depth, whereas available N increased within the layer between 30 and 45 cm depth for the Sapucaia plantation soil (Figure 1). On the other hand, available N in the Coruripe plantation soil was almost constant with depth, while 15 N natural abundance increased along the soil profile. Generally, 15 N natural abundance from available N found in the topsoil (0-15 cm depth) is lower than in subsoil (Unkovich et al., 2008). Results show that the soil does not provide 15 N isotope distribution uniformity according to the depth.

Discussion
Sugarcane variability in response to nitrogen fertilization, regardless of the N source, is a crop characteristic, for which there is still no conclusive explanation (Boddey et al., 2003;Franco et al., 2011;Vitti et al., 2008).
This study verified a response from both sugarcane varieties to nitrogen fertilization. Such effect was clear for RB867515 grown at Sapucaia mill, as well as both varieties grown at the Coruripe mill (Table 3). These results are distinguished from most literature reports, which usually do not show increases in productivity in the first year (planted cane) fertilized with nitrogen (Franco et al., 2010(Franco et al., , 2011Vitti et al., 2008). Several factors have been listed as being responsible for the lack of planted cane to respond Table 6. total recoverable sugar yield (trS) of two grown sugarcane varieties inoculated with diazotrophs, fertilization with 120 kg ha −1 of n, and control in two Brazilian regions. notes: Means of four replications. Values followed by different letters in the columns differ by Scott-Knott test at 10% probability. increases and the sums of the increases for treatments with inoculation and n fertilization compared to the control were not analyzed statistically. c.V: coefficient of variation.   differ between inoculated and non-inoculated plants, allowing us to infer that the benefit provided by inoculation was not due to the BNF process associated with the addition of the biofertilizer.
A similar experiment carried out on an Alfisol at Embrapa Agrobiology, RJ, showed that the RB72454 variety responded to inoculation and nitrogen fertilization according to sugarcane yield parameters in contrast to the RB867515 that was not responsive to either treatment in the plant cane or first ratoon, but only to nitrogen fertilization in the second ratoon (Schultz et al., 2014).
This study revealed RB867515 and RB72454 variability in response to inoculation and soil types. Such results are similar to those of Oliveira et al. (2006). Sugarcane variability to respond to nitrogen fertilization may be due to environmental conditions, since precipitation and temperature data described in the Materials and Methods section are related to the regional level, which may not have been the experimental area actual condition at the time of fertilizer application and during the days after application. Environmental factors, especially drought, may influence the nitrogen fertilization efficiency negatively (Gava et al., 2001). Other authors observed that good water conditions along with the correct N supply may promote sugarcane root growth (Otto et al., 2009).
TRS, Brix, and Pol similarities among inoculated plants, N fertilization (120 kg ha −1 ), and control treatments indicated that there was no effect on sugar synthesis in sugarcane plants. Among those researchers working in the area, there is no consensus concerning effects of the inoculation of diazotrophs on sugar synthesis in sugarcane plants. There are reports showing positive, negative, and non-applicable responses from nitrogen fertilization regarding sugarcane juice quality. These results have been attributed to differences between varieties, soil and climate conditions, crop management, N sources, and application procedures (Franco et al., 2010;Schultz et al., 2010).
Despite the variability from both sugarcane types to respond to inoculation and nitrogen fertilization, TRS total sum for both of them at the two regions showed a 10,096 kg ha −1 gain due to inoculation and 14,853 kg ha −1 due to nitrogen fertilization. Considering the U$ .138 (U$ 1.0 = R$ 3.323) reference value per kg of TRS paid to sugarcane producers (UNICA, 2015), inoculation provided a U$ 1393.25 gross profit, while nitrogen fertilization corresponded to U$ 2049.71. A higher economic gain (U$ 656.46) from nitrogen fertilization in comparison to inoculation represents approximately the cost of the commercial nitrogen fertilizer. Therefore, results suggest that the net return due to inoculation for these two experimental areas is the positive environmental balance which comes from inoculation, that can replace industrial nitrogen fertilizer.
A proportional δ 15 N value reduction in inoculated plants in comparison to non-inoculated plants should have been detected in case of a positive BNF contribution process (Unkovich et al., 2008). The hypothesis that the yield increase in the inoculated RB867515 variety did not result from the BNF is reinforced by the fact that N accumulated in stems and aerial part total dry matter did not differ between inoculated and non-inoculated plants, even when inoculation promoted significant increases in sugarcane yield. Very large differences between the 15 N abundance of the plant available N in the soil and that value in the sugar cane plants reinforces the belief that even without N 2 -fixing bacteria inoculation, the sugar cane plants obtained large N contributions from BNF.
These results allow us to infer that inoculation promotes other benefits for sugarcane plants, which may be associated with actions of phytohormones that are synthesized by diazotrophs (Suman et al., 2001), solubilization of phosphate and zinc compounds (Shukla et al., 2008;Singh et al., 2007), or rhizosphere retention of essential nutrients (Yadav et al., 2009). Some other studies have stated that diazotrophs can act on plants, mainly by changing root system morphology, and thus influencing positively the crop development and productivity (Bashan et al., 2004). Muñoz-Rojas and Caballero Mellado (2003) evaluated micro-propagated sugarcane (MEX 57-473 variety) inoculated with G. diazotrophicus -PAL5 T strain and grown in sterile vermiculite. The authors verified increases in weight of dry matter roots, shoots, and total N accumulation, although total N levels were lower in comparison to the control. In addition, they concluded that benefits by G. diazotrophicus inoculation may have been derived from bacterial plant growth promoting effect.
Results from the current study corroborate other studies reported in the literature using the same multiple inoculum. Silva et al. (2009) evaluated the same sugarcane varieties grown in a Ultisol in the field in Seropédica, RJ, Brazil. They found that inoculation promoted the productivity of fresh stems for both varieties but did not affect the total N accumulated in plant tissues. No response to inoculation or to nitrogen fertilization was detected for either sugarcane varieties (planted cane) grown in an Inceptisol in the field in Campos dos Goytacazes, RJ, Brazil . However, in the second ratoon, re-inoculated through aspersion on cut stems, the RB867515 variety responded to inoculation as well as with fertilization with 120 kg ha −1 of N. Nevertheless, 15 N delta values did not In summary, this study suggests that bacterial inoculants benefit sugarcane plants by increasing the stems and sugar yield (TRS). However, inoculation did not alter the BNF naturally associated with the crop, suggesting that benefits derived from the inoculant may come from plant growth promoting substances which are synthesized by diazotrophs or other growth promotion effects described for the five strains used as a mixed inoculant.