Salt stress alleviation in citrus plants by plant growth-promoting rhizobacteria Pseudomonas putida and Novosphingobium sp.

This work reveals the protective role of two rhizobacteria, Pseudomonas putida and Novosphingobium sp., on citrus plants subjected to salt stress conditions. Detrimental salt stress effects on crops are likely to increase due to climate change reducing the quality of irrigation water. Plant growth-promoting rhizobacteria (PGPRs) can mitigate stress-induced damage in plants cultivated under high salinity conditions. In this work, Citrus macrophylla (alemow) plants inoculated with the rhizobacteria Pseudomonas putida KT2440 or Novosphingobium sp. HR1a were subjected to salt stress for 30 days. Results showed that in absence of salt stress, Novosphingobium sp. HR1a induced a decrease of transpiration (E) and stomatal conductance (gs). Both rhizobacteria reduced salt stress-induced damage. Levels of abscisic acid (ABA) and salicylic acid (SA) were lower in inoculated plants under salt stress conditions. Similarly, under stress conditions maximum efficiency of photosystem II (Fv/Fm) in inoculated plants decreased to a lower extent than in non-inoculated ones. In stressed plants, Novosphingobium sp. HR1a also induced leaf accumulation of 3-indole acetic acid (IAA) and a delay in the decrease of quantum yield (ΦPSII). P. putida KT2440 inhibited root chloride and proline accumulation in response to salt stress. Although both bacterial species had beneficial effects on salt-stressed citrus plants, Novosphingobium sp. HR1a induced a better plant performance. Therefore, both strains could be candidates to be used as PGPRs in programs of inoculation for citrus protection against salt stress.


Introduction
In nature, plants are constantly subjected to a wide variety of both abiotic and biotic stress conditions which can reduce their growth and productivity.Moreover, climate change exacerbates these adverse conditions, affecting mainly to abiotic stress conditions as drought, heat stress or salinity (Zandalinas et al. 2018).In this context, salt concentration in groundwater of coastal regions is increasing because of saltwater intrusion due to overexploitation of aquifers, leading to reductions of yield and plant performance (Klassen and Allen, 2017).High substrate salinity has two different effects which trigger plant damage: the osmotic component, that appears in the early stage of stress and restricts water absorption, producing plant dehydration and turgor loss; and the ionic component, due to the accumulation of Na + and Cl -ions, that reach toxic levels in tissues (Gupta and Huang, 2014).For most species Na + appears to reach toxic concentrations in plant tissues before Cl -does.On the contrary, in the case of citrus, Cl -is considered the most toxic ion (Moya et al. 2002(Moya et al. , 2003;;Montoliu et al. 2009).Consequently, it is necessary to explore new strategies to maximize plant tolerance to this stress condition to improve plant productivity in affected zones, as occurs in the Mediterranean region where citrus is one of the main crops.
Several strategies to mitigate salt stress-induced damages in plants have been proposed, including chemical treatments, such as 24-epibrassinolide (Ekinci et al. 2012) or abscisic acid (ABA) analogues (Arbona et al. 2006), improvement of mineral fertilization (Rady, 2012), modification of gene expression (Zhao et al. 2017;Vives-Peris et al. 2018a), and modification of soil microbiota communities to potentiate plant colonization by beneficial microorganisms such as arbuscular mycorrhizal fungi or plant growth promoting rhizobacteria (PGPR) (Qin et al. 2016).PGPR can benefit plant growth and alleviate salt stress by multiple mechanisms: i) producing biofilms that favour humidity accumulation around the roots and avoid the entrance of toxic ions or pathogens (Wang et al. 2017), ii) increasing nutrient availability in the substrate due to phosphate solubilization or atmospheric nitrogen fixation (Pereira and Castro 2014;Wang et al. 2017), iii) inducing root growth and development through the release of phytohormones and secondary metabolites to the rhizosphere such as indole-3-acetic acid (IAA) or other signalling molecules as nitric oxide (Cassán et al. 2014), iv) producing siderophores (Sayyed et al. 2013) and v) decreasing 1aminocyclopropane-1-carboxylate (ACC) levels, by increasing ACC deaminase activity in the rhizosphere, which consequently derives in a decrease in ethylene concentration in plant tissues (Vacheron et al. 2013;Singh et al. 2015;Nadeem et al. 2016).It has been also reported that PGPR induces an increase in water use efficiency by modulating transpiration and stomatal conductance, and a decrease in the content of reactive oxygen species in inoculated plants (Vejan et al. 2016).The effectiveness of these stress-mitigating effects depends on the type of plants, bacteria, and their putative interaction.
Although the palliative effects of PGPR on plants subjected to abiotic stress conditions have been studied in a wide variety of herbaceous plants, including tomato, rice, lettuce wheat, potato, cotton, soybean, maize, chickpea, lentil or pea (Dimkpa et al. 2009;Nadeem et al. 2014, Vurukonda et al. 2016), the effect of these beneficial microorganisms in woody plants is less well-known and reports are limited to Vitis vinifera subjected to chilling (Barka et al. 2006), or Pinus halepensis and Quercus coccifera trees subjected to water stress (Rincón et al. 2008).In citrus, the information on the stress-mitigating role of soil microorganisms is mainly focused on the beneficial effects of mycorrhizal fungi in plants growing under different biotic or abiotic stresses such as those caused by Phytophthora (Watanarojanaporn et al. 2011), drought (Wu andZou, 2009), salinity (Satir et al. 2016;Zhang et al. 2017) and low temperatures (Wu and Zou, 2010).In relation to PGPRs, only the effect of Pseudomonas putida FCA-8 on citrus cultivated in the absence of any stress condition has been reported (Chiquito-Contreras et al. 2012).Moreover, although P. putida KT2440 has been described as a PGPR in Zea mays (Planchamp et al. 2015), the positive role on plant growth of the strain Novosphingobium sp.HR1a has not been proved, being only considered as a PGPR as with other strains in this genus (e.g., Novosphingobium oryzae sp.nov., Zhang et al. 2016).
Consequently, in this work, the putative palliative effect of two rhizobacterial strains, P. putida KT2440, and Novosphingobium sp.HR1a, on damage caused by salt stress conditions in alemow plants was evaluated.The main objective of this investigation has been to test the beneficial effect of both strains under stressful situations, regarding on different plant biochemical and physiological parameters.

Plant material and treatments
Six-month-old alemow (Citrus macrophylla Wester) plants were acclimated in a greenhouse for two months under natural photoperiod and temperatures of 25 ± 3.0°C and 18 ± 2.0°C (day/night respectively).
A non-sterilized mixture of peat moss, perlite and vermiculite (80:10:10) was used as substrate.Plants were watered with half-strength Hoagland solution three times a week (Arbona et al. 2009).
A first experiment was carried out to optimize the inoculation.Plants about 50 cm height, were transferred to plastic pots containing 400 mL of substrate (sterilized three times with an autoclave at 121º C for 30 min each time),.Inoculated plants were watered with a P. putida KT2440 solution containing the necessary bacteria to inoculate the pot volume to an OD 660nm of 0.1 (Franklin et al. 1981), while control plants were watered with the same volume of water, without inoculum.Ten days after the inoculation, salt stress was applied by adding 60 and 90 mM NaCl to the watering solution twice a week.Both, non-inoculated plants, and plants watered without NaCl were added as controls.Leaf and root samples were randomly collected 30 days after salt stress onset (Fig. 1.A).This experiment was realized with ten plants per group at the beginning of spring and it was not repeated since it was performed for stablishing the conditions for the second experiment.
In the second experiment, C. macrophylla plants (similar in height and age to those used in the first experiment) were inoculated with two different rhizobacterial strains, P. putida KT2440 and Novosphingobium sp.HR1a up to a final OD 660nm of 0.1 (Franklin et al. 1981;Segura et al. 2017).Ten days after the inoculation with bacteria, salt stress was applied, by adding 90 mM NaCl to the watering solution twice a week (Fig. 1.B).Leaf and root tissues were randomly sampled after 30 days of stress.Nondestructive analyses, including gas exchange and chlorophyll fluorescence parameters were determined at 10, 20 and 30 days after the stress onset.This experiment was performed with ten plants per group and it was repeated twice during the summer period with similar results

Visual leaf damage
In the first experiment damage apparition was determined by counting the percentage of abscessed leaves in each one of the ten plants of every group after 30 days of stress.In the second experiment different degrees of salt stress-induced damage were stablished, including non-damaged leaves, mild-damaged leaves, intermediate-damage leaves, severe-damaged leaves and abscessed leaves (Fig. 2).The percentage of apparition of each symptom was counted 30 days after salt stress imposition in ten plants per group.

Chloride analysis
Quantification of chloride ions was performed in plant tissue and soil saturated extract.Measurements in leaves and roots were performed by automatic titration with a chloride meter (Model 626, Sherwood Scientific Ltd., Cambridge, UK) as described in López-Climent et al. (2008).Three replicates of each sample were extracted by adding 25 mL of the chloride extraction buffer, consisting in 0.1 N HNO 3 (Panreac, Barcelona, Spain) and 10% glacial acetic acid (Sigma-Aldrich, St. Louis, MO, USA) to 0.25 g of fresh tissue, and were incubated for 12 hours at room temperature.Chloride concentration was measured by titrating 0.5 mL of the solution with the chloride meter Chloride measurements of soil saturated extract were performed by adding water to 2 g of soil until saturation.After 24 hours at room temperature, the water was collected from the soil with a vacuum pump, and chloride concentration was measured in the chloride meter.Chloride content was measured in three independent samples of soil for each treatment.

Determination of colonization rate
In the first experiment, plant colonization by the rhizobacteria was determined by counting colony forming units (CFU).Roots were washed with 5 mL of sterile deionized water in agitation at 200 rpm for 1 hour.
Serial dilutions of the wash water were plated on a selective lysogeny broth (LB) medium (Bertani, 1951) supplemented with chloramphenicol, an antibiotic to which P. putida KT2440 is resistant (Franklin et al. 1981).CFU were counted after 24 h of incubation at 30°C (Goldman and Green, 2008).For each treatment, three root samples from different plants were washed with sterile deionized water as described previously and three replicates of each one were plated.

Proline analysis
The concentration of proline was determined in leaf and root samples as indicated in Bates et al. (1973) with some modifications.Briefly, three replicates of fresh material was extracted by sonication in 3% sulfosalicylic acid (Panreac) in distilled water.After centrifugation, supernatant was mixed with glacial acetic acid (Sigma-Aldrich) and ninhydrin reagent (prepared as in Vives-Peris et al. 2017).Samples were incubated at 100 ºC for one hour and after centrifugation, proline concentration was spectrophotometrically determined at 520 nm.

Phytohormone analysis
Concentration of ABA, salicylic acid (SA) and IAA was determined in leaves collected after 30 days by high performance liquid chromatography coupled online to a triple quadrupole mass spectrometer (Micromass, Manchester, UK) through an orthogonal Z-spray electrospray ion source (Durgbanshi et al. 2005).Three replicates of each sample were extracted with water from 0.2 g of fresh material reduced to fine powder by using a mill ball equipment (MillMix20, Domel, Železniki, Slovenija).[ 2 H 6 ]-ABA, [ 13 C 6 ]-SA and [ 2 H 2 ]-IAA were used as internal standards.Samples were centrifuged after the extraction, and the supernatant was recovered, and pH adjusted to 2.8 to 3.2 with acetic acid.A liquid-liquid partition was performed twice with diethyl ether and the supernatant was evaporated under vacuum in a centrifuge concentrator (Speed Vac, Jouan, Saint Herblain Cedex, France).The solid residue was diluted in 0.5 mL of water:methanol 90:10 and filtered through 0.22 µM PTFE filters.Finally, 20 µL of this solution was injected into the HPLC-MS system (Acquity SDS, Waters Corp., Milford, MA, USA).
Chromatography separation was achieved by using a reversed-phase C18 column (Gravity, 50 × 2.1mm 1.8-μm particle size, Macherey-Nagel GmbH, Germany) as stationary phase, and a methanol:water gradient, both supplemented with 0.1% acetic acid, at a flow rate of 300 μL min −1 as mobile phase.Standard curves with the commercial standards of the different phytohormones were used for quantifying sample phytohormone concentrations.Results were processed using Masslynx v4.1 software.

Chlorophyll fluorescence parameters
Quantum yield (Φ PSII ) and maximum efficiency of photosystem II photochemistry, as F v /F m ratio, were measured 10, 20 and 30 days after the onstet of the experiment between 9 and 11 h AM in twelve randomly chosen undamaged leaves from three plants per treatment using a portable fluorometer (FluorPen FP-MAX 100, Photon Systems Instruments, Czech Republic).Four measurements were taken per leaf (Zandalinas et al. 2016).

Leaf gas exchange parameters
Transpiration (E) and stomatal conductance (g s ) were measured with a LCpro+ portable infrared gas analyzer (ADC Bioscientific Ltd., Hoddesdon, UK) under ambient CO 2 and humidity.Light was provided by a photosynthetically active radiation lamp at 1000 µmol m −2 s −1 photon flux density.Air flow was set at 150 µmol mol −1 and all measurements were performed between 9 and 11 h AM.Three undamaged leaves of three different plants were analyzed per treatment, and after instrument stabilization, ten measures were consecutively performed in every leaf after 10, 20 or 30 days of treatment (Zandalinas et al. 2016).

Statistical analyses
Statgraphics Plus v.5.1.Software (Statistical Graphics Corp., Herndon, VA, USA) was used for statistical analyses.Represented data are means of independent determinations and were subjected to one-or two-way analysis of variance (ANOVA) and a Tukey posthoc test (p ≤ 0.05) when significant differences were detected.

Leaf abscission
After 30 days of treatment, leaf abscission increased in non-inoculated salt-stressed plants, reaching the highest level in plants subjected to 90 mM NaCl, with a percentage of leaf abscission of 17.5%, whereas in plants inoculated with P. putida KT2440, this percentage was of 2.6%, being similar to the observed in control plants (Fig. 3).

Chloride accumulation
Chloride concentration was determined in the substrate, shoots and roots (Fig. 4).In the substrate, the presence of bacteria did not modify chloride concentration, being this value only affected by high salinity, with chloride levels 4.6 and 8.9 times higher than controls when soil was treated with 60 and 90 mM NaCl, respectively (Fig. 4A).A similar trend was observed in leaves, with an increase in the concentration of this ion depending on the stress severity exclusively, reaching values 1.8 and 3.0 times higher than those observed in leaves of non-stressed plants when they were subjected to 60 and 90 mM NaCl, respectively (Fig. 4B).The presence of the bacteria P. putida KT2440 reduced chloride accumulation in roots of plants subjected to 90 mM NaCl (24.6% reduction compared to salt-stressed non-inoculated plants, Fig. 4C).

Colonization rate
The number of CFU was analysed at the end of the experimental period.Two different levels of bacterial populations were observed (Fig. 5).In the case of non-inoculated plants, values of CFU were about 150,000 CFU g -1 root, whereas in plants inoculated with P. putida KT2440, this value was around 600,000 CFU g - 1 root.The increasing levels of NaCl did not affect the colonization rate.
Evaluation of the palliative effect of both strains in plants subjected to salt stress

Appearance of symptoms induced by salt stress
The presence of both rhizobacterial strains reduced salt stress-induced damage.Thus, whereas in noninoculated plants a 68.3% of leaves were affected, in plants inoculated with P. putida KT2440 or Novosphingobium sp.HR1a, the percentages of damaged leaves were lower (57.0 and 45.0%), respectively (Fig. 6).Moreover, in salt stress conditions, the severity of the damage in leaves from plants that had been previously inoculated with any of both PGPR strains was lower than in non-inoculated plants subjected to salt stress, with statistically significant diminutions of the appearance of severe-damaged leaves and abscessed leaves (Sup.Mat. 1).

Proline concentration
An increase of leaf proline concentration was observed in all groups of plants due to salt stress application (Fig. 7A) independently if plants had been previously inoculated or not.Proline concentration in plants under salt stress was about 4.2 times higher than in their respective non-stressed controls.In roots of plants subjected to salt stress, there was a 3.9-fold increase of proline content in salt-stressed non-inoculated plants (Fig. 7B).Proline content also exhibited an increase (2.3-fold) in roots of plants inoculated with Novosphingobium sp.HR1a.However, in roots of plants inoculated with P. putida KT2440, no statistical differences in the content of the amino acid between stressed and non-stressed plants were found.

Phytohormone concentration
Endogenous levels of the phytohormones ABA, SA and IAA were measured in leaves after 30 days of stress (Fig. 8).After this period, non-inoculated plants treated with 90 mM NaCl exhibited a 1.7-fold increase in the leaf concentration of ABA (related to non-stressed plants).Contrarily, in plants inoculated with P. putida KT2440 the concentration of this phytohormone decreased 40.4% after salt treatment.In plants inoculated with Novosphingobium sp.HR1a, salt stress did not alter leaf ABA content (Fig. 8A).
In the absence of inoculation, salt stress induced a 2.3-fold increase of SA leaf concentration in comparison with non-stressed plants.However, no differences were observed between salt-stressed plants and controls in presence of any of the bacterial strains used in the experiment (Fig. 8B).Meanwhile, increases in IAA content were only recorded in leaves of plants inoculated with Novosphingobium sp.HR1a and treated with 90 mM NaCl (2.8 fold with respect to controls, Fig. 8C).

Chlorophyll fluorescence parameters
Salt treatment and bacteria inoculation induced changes in chlorophyll fluorescence parameters (Fig. 9).Salt stress clearly reduced Φ PSII in non-inoculated plants, with a reduction of 36.4 and 55.5% related to control plants after 20 and 30 days, respectively.In plants inoculated with P. putida KT2440, this reduction was similar to that observed in non-inoculated plants, (reduction of 23.8 and 52.1% in Φ PSII values related to control after 20 and 30 of salt stress, respectively).In plants inoculated with Novosphingobium sp.HR1a, this difference of Φ PSII values between non-stressed and salt-stressed plants also decreased due to salt stress (a diminution of 32.2% respect to control after 20 days from stress onset).However, after 30 days of salt stress, the reduction of this parameter was not as marked as in non-inoculated plants (decrease of 39.0%, Fig. 9A).
In line with the Φ PSII , F v /F m decreased with the application of NaCl from 20 days of stress until the end of the experiment (28.1 and 44.3% reduction in non-inoculated plants at 20 and 30 days, respectively).
However, in inoculated plants, levels of F v /F m were similar to controls for the first 20 days of stress.After 30 days of stress F v /F m values decreased a 52.9 and 38.3% in salt-stressed plants inoculated with P. putida KT2440 and Novosphingobium sp.HR1a, respectively (in comparison with non-stressed inoculated plants, Fig. 9B).

Gas exchange parameters
Gas exchange parameters, including E and g s , were measured in leaves throughout the experimental period but no differences were observed until 20 days of stress (Fig. 10).Salt stress induced a decline of E (reduction of E values between 52.4 and 68.0% after 20 days and between 82.7 and 84.5% after 30 days, related to control).Moreover, in absence of salt stress, Novosphingobium sp.HR1a also induced a diminution of 35.5% in this parameter after 20 days in comparison with the value observed in noninoculated plants (Fig. 10A).
In addition, g s exhibited a similar trend than that observed in E. Most of the differences observed in g s were in response to salt stress, with a decrease between 58.4 and 74.0%depending on the inoculum, at 20 days, being more evident after 30 days, with values between 84.8 and 88.3% lower than those observed in nonstressed plants.g s was also influenced by the inoculation with Novosphingobium sp.HR1a in absence of salt stress, exhibiting a decrease of 47.9 and 50.9% in comparison to control plants after 20 and 30 days from the beginning of salt stress treatments respectively (Fig. 10B).

Discussion
Results presented in this work reveal that inoculation with P. putida KT2440 or Novosphingobium sp.HR1a mitigates the negative effect of salt stress on alemow plants, indicating that both rhizobacteria have a role as a PGPR.As far as we know, the beneficial effect of Novosphingobium sp.HR1a on plant performance has not been described previously.Although both strains have positive effects on plant tolerance to this stress condition, there are common and different responses depending on the inoculated rhizobacterium.
The number of soil total CFU increased because of the inoculation by P. putida KT2440, independently of the salt treatment used.Although chloramphenicol was added to the selective bacteria culture medium, a high number of CFU was also recorded in those cultures of rhizospheric soil from non-inoculated plants.
This fact could be due to the occurrence of antibiotic-resistant bacteria in soil.In fact, it has been reported the existence of different bacterial species as Bacillus subtilis, Escherichia coli, Streptococus sp., Klebsiella mobilis or Pseudomonas aeruginosa simultaneously resistant to different antibiotics including chloramphenicol in different soils (Mindlin et al. 2008;Popowska et al. 2012;Eghomwanre et al. 2016), what explains the results obtained in this work.
Salt stress had a negative effect on citrus performance, inducing leaf damage and abscission, Cl -, and proline accumulation, and a decrease in gas exchange and chlorophyll fluorescence parameters.These results are in concordance with previous reports (López-Climent et al. 2008;Hussain et al. 2012).Moreover, in noninoculated plants, salt stress induced leaf accumulation of ABA and SA, which has been widely reported in several species as Arabidopsis thaliana (Prerostova et al. 2017), Cucumis sativus (Chojak-Koźniewska et al. 2017) or the citrus rootstock Carrizo citrange (Citrus sinensis L. Osbeck x Poncirus trifoliata L. Raf., Gómez-Cadenas et al. 1998).Although the results of leaf damage appearance follow a similar trend in both experiments, the symptoms were more evident in the second experiment.This fact could be due to the increase of citrus plants metabolism during summer, absorbing higher quantities of water (and the toxic ions Na + and Cl -in salt-stressed plants), causing a higher affection of salt stress, as it has been described in other plant species as Vigna radiata (Sehrawat et al., 2015).
Among the palliative effects of the inoculation with rhizobacteria, a reduction of stress-induced proline accumulation in roots of plants inoculated with P. putida KT2440 was observed.It has been reported that levels of this amino acid generally increase in plant tissues under stress situations such as drought and salinity.It has been suggested that this compound can act as an osmoprotectant that avoids plant dehydration and turgor loss, as well as an oxidative damage inhibitor (Hayat et al. 2012).Other authors have indicated that proline should be considered a stress marker although with a marginal role in plant tolerance (Arbona et al. 2017).However, independently of its role, an increase of proline is generally associated to abioticstress induced damage.Consequently, the decrease in proline levels in salt-stressed plants inoculated with P. putida KT2440 would indicate that plants were suffering stress in a lower degree.Interestingly, whereas non-inoculated plants increased leaf contents of ABA and SA in response to salt stress, those plants inoculated with Novosphingobium sp.HR1a exhibited unaltered hormone levels in response to the adverse situation and those inoculated with P. putida KT2440 showed even a decrease in ABA levels.This lack of increase of ABA and SA levels would further support the lower impact of salt stress on inoculated plants.
It has been described previously that the ACC-deaminase activity produced by PGPRs inhibits ethylene biosynthesis in plants (Dimkpa et al. 2009).Consequently, since ethylene crosstalks with ABA, this chain could lead to a reduction of ABA levels (Arc et al. 2013).A lower increase in ABA has been also reported in salt-and osmotic-stressed cucumber plants inoculated with the PGPRs Burkholderia sp., Acinetobacter sp., and Promicromonospora sp.(Kang et al. 2014).In addition, IAA levels increased in plants inoculated with Novosphingobium sp.HR1a and subjected to salt stress.This increase could be due to the presence of the PGPR, since some rhizobacteria, including Novosphingobium genus, produce IAA (Krishnan et al. 2017).In any case, IAA has been reported as a salt stress reliever (Kaya et al. 2013) and could promote root growth and lateral root development, facilitating root exploration of new soil zones with lower contents of toxic elements or higher water availability (Bao et al. 2014).
The inoculation with both rhizobacteria allowed salt-stressed plants to keep higher F v /F m values than noninoculated ones for 20 days whereas Novosphingobium sp.HR1a also induced the maintenance of higher levels of Φ PSII in salinized plants even after 30 days of stress treatment.Although chlorophyll fluorescence parameters, as well as chlorophyll and carotenoid contents, generally decrease with the stress application (López-Climent et al. 2008), PGPRs help plants to delay the adverse effects.For example, in water stressed Ocimum basilicum plants, inoculation with Pseudomonades sp.induced higher F v /F m values (Heidari and Golpayegani, 2012).Moreover, PGPRs from Bacillus megaterium and Enterobacter sp. have been reported as inducers of chlorophyll accumulation in Abelmoschus esculentussubjected to salt stress (Habib et al. 2016).Although most studies regarding the effect of PGPRs on photosystem II are focused in herbaceous crops, Rincón et al. (2008) working with Pinus halepensis and Quercus coccifera inoculated with Pseudomonas fluorescens reported similar results as those reported here, but under drought stress conditions.
There is some controversy about the effect of PGPRs on E and g s .Whereas a positive relationship among PGPR inoculation and high E and g s values has been reported in salt-stressed Vigna radiata inoculated with Enterobacter cloacae and Bacillus drentensis (Mahmood et al. 2016), and Triticum durumtreated with PGPRs (Zhu et al. 2014), other works indicate that E and g s decrease in presence of PGPRs, improving water use efficiency and consequently improving plant tolerance to stress conditions (Bresson et al. 2013;Yasmin et al. 2013).Results obtained in this work showed that E and g s decreased in absence of stress in plants inoculated with Novosphingobium sp.HR1a, although further work would be needed to explore the metabolic consequences of this reduction.
In addition to the physiological parameters measured, chloride ion concentration decreased in roots of plants inoculated with P. putida KT2440 and subjected to salt stress.The accumulation of this ion has been previously reported as the critical component of salt stress toxicity in citrus plants, being a key marker to quantify salt stress damage in this crop (Moya et al. 2003).Therefore, reductions in the absorption of this toxic ion have been associated to a stress tolerance in citrus plants (López-Climent et al. 2008;Hussain et al. 2012).
Previous studies have reported that Novosphingobium sp.HR1a is more tolerant to high NaCl concentrations than P. putida KT2440 (Vives-Peris et al. 2018b).This better performance of Novosphingobium sp.HR1a under high salinity could explain its higher beneficial effects on salt-stressed citrus plants.
In conclusion, results presented in this work reveal that the rhizobacterial species P. putida KT2440 and Novosphingobium sp.HR1a have a palliative effect on citrus plants subjected to salt stress, reducing the damage caused by this adverse condition.Consequently, both rhizobacterial strains can be considered as PGPRs and could be used in biofertilization and bioaugmentation programs in order to promote plant growth and prevent the damage caused by salt stress.Finally, the positive effects caused by Novosphingobium sp.HR1a on plant performance were more evident although this species has not been described previously as a PGPR.

Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6
Figure 1 Experimental design of the first (A) and the second set of experiments (B)

Figure 7 Figure 8 Figure 9
Figure 7 Proline concentration in leaves (A) and roots (B) of non-inoculated plants and plants inoculated with P. putida KT2440 or Novosphingobium sp.HR1a in control conditions (white bars) and 90 mM NaCl treatments (grey bars).Values indicate the mean of three replicates ± standard error.Different letters refer to statistically significant differences at P ≤ 0.05

Figure 10
Figure 10 Gas exchange parameters in plants subjected to the different treatments.Transpiration (A) and stomatal conductance (B) of non-inoculated plants and plants inoculated with P. putida KT2440 or Novosphingobium sp.HR1a in control conditions and 90 mM treatments after 10 (white bars), 20 (light grey bars) and 30 days (dark grey bars).Values indicate the mean ± standard error.* refers to statistically significant differences among control and salt-stressed plants at P ≤ 0.05, while † refers to statistically significant differences among non-inoculated and inoculated plants at P ≤ 0.05