Drought, the main limiting factor for crop yields worldwide (Sharp et al., 2004), is considered the abiotic stress with the greatest effect on plants, since 45% of global farming lands are subjected to continuous or frequent droughts (Long and Ort, 2010). The Intergovernmental Panel on Climate Change (IPCC) has projected that the land area affected by drought will increase while water resources in the affected areas could decline by as much as 30% by mid-century (Christensen et al., 2007).

In this context, ethylene, abscisic acid (ABA) and indoleacetic acid (IAA) are phytohormones often associated with plant senescence or root control (Stella et al., 1996) and have also been associated with how plants respond to water deficit (Yamaguchi-Shinozaki and Shinozaki, 1994; Kawakami et al., 2010). Ethylene is synthesized from 1-amino cyclopropane-1-carboxylic acid (ACC) by the enzyme ACC oxidase (ACO or ethylene-forming enzyme; EC:1.14.17.4), and regulates a number of physiological mechanisms. At low concentrations, it promotes the development of adventitious roots and root hairs (Nadeem et al., 2010), and at high concentrations it inhibits root growth under stress conditions such as drying soil (Glick et al., 2007). With regard to ABA, this key phytohormone causes stomatal closure to prevent further decreases in water content, promotes root growth and the accumulation of compatible solutes, and regulates the synthesis of dehydrins as well as late embryogenesis abundant (LEA) proteins, thus coordinating various facets of the plant’s response to low water potential (ψ_w) (Pierce and Raschke, 1980; Zhu, 2002; Raschke, 2006). Some auxins, including IAA, when present in low concentrations, increase cell elongation, resulting in greater root growth and the generation of lateral roots. However, at higher concentrations IAA, inhibits root growth either directly or indirectly via the promotion of ethylene synthesis. The effect of ethylene is intertwined with the production of IAA, since on the one hand ethylene has been shown to lower endogenous IAA levels in different plant tissues, and on the other hand IAA stimulates ethylene production by promoting ACC synthase activity (Glick, 2003; Dimkpa et al., 2009).

The accumulation of xeroprotectants such as trehalose by some microorganisms and some plants enables them to withstand extreme abiotic stress such as desiccation (Chaplin, 2006; Julca et al., 2012). Efforts have been made to bolster plant drought tolerance by developing transgenic plants with altered phytohormone production (Sobeih et al., 2004) and greater trehalose concentrations (Goddijn and Smeekens, 1998; Pilon-Smits et al., 1998). In addition to transgenic plants, plant-growth-promoting rhizobacteria (PGPR) have great potential in agriculture since their presence in the roots increases crop production by enhancing the plant’s tolerance to drought and other types of environmental stress (Vilchez and Manzanera, 2011). According to Timmusk et al. (2013), these bacteria are called rhizobacterial drought-tolerance enhancers (RDTE). Over the last decade, RDTE have received particular attention (Glick et al., 2007), and since Mayak et al. (2004) reported that the strain Achromobacter piechaudii ARV8 was useful in protecting plants against drought, many other microorganisms have been described with similar properties. In general, these microorganisms are able to counteract the negative effects of drought stress in plants by producing ACC deaminase (ACCd), lowering ABA, or raising IAA production (Mayak et al., 2004). See Kaushal and Wani (2016) for a recent review on this subject.

Water stress also affects the viability of microorganisms in different ways (Manzanera et al., 2002) by compromising protein structure, stability, activity, folding, and assembly by disrupting the physical architecture and functions of nucleic acids; and by altering membrane structure and function (Chaplin, 2006; Raschke, 2006; Julca et al., 2012). A large number of RDTE have been isolated in samples from dry or drought-affected locations (Mayak et al., 2004; Dimkpa et al., 2009; Roca et al., 2013), thus suggesting that this type of microorganism might be well adapted to water stress.

Drought conditions would therefore be expected to produce the simultaneous selection of tolerant plants and microorganisms in the environment, assuming that the selection of desiccation-tolerant microorganisms counteracts the deleterious effect of drought on plants (Lau and Lennon, 2012). However, no study available has drawn a clear connection between a microbial phenotype and the ability to protect plants from drying. The present study establishes a correlation between tolerance to desiccation in a set of highly desiccation-tolerant PGPR and their closest taxonomic relatives at the species and strain level, and the level of plant protection against drought in two different crops (tomato and pepper plants). This correlation between desiccation tolerance of the microorganism and its ability to protect some plants from drought seems to depend on the microorganism’s capacity to finely regulate the concentration of trehalose in the plant as a signal of drying damage, since trehalose triggers the plant-defense system to counteract the damage caused by drought. To demonstrate the role of microbial trehalose in protecting the plant against desiccation, we isolated the otsAB genes, coding for alpha, alpha-trehalose-phosphate synthase and trehalose-6-phosphate phosphatase, from the highly desiccation-tolerant strain Microbacterium sp. 3J1, and inserted them into the desiccation-sensitive microorganism P. putida KT2440, generating P. putida KT2440 (pUCP22:otsAB) with a significant increase in the intracellular concentration of trehalose under water stress conditions. In addition, plants inoculated with the resulting trehalose-overexpressing P. putida KT2440 (pUCP22:otsAB) showed a significantly greater tolerance to drought.

Pepper and tomato plants were used to test the ability of five highly desiccation-tolerant microorganisms to protect plants against drought. All five bacterial strains were previously isolated in our laboratory: Microbacterium sp. 3J1, Arthrobacter koreensis 5J12A, A. siccitolerans 4J27, Rhodococcus sp. 4J2A2, and Leucobacter sp. 4J7B1 (Narváez-Reinaldo et al., 2010). For a negative control, we used the desiccation-sensitive PGPR P. putida KT2440, and for a positive control, we used the well-characterized RDTE A. piechaudii 366-5 (Manzanera et al., 2002; Dos Santos et al., 2004; Mayak et al., 2004). An additional set of non-inoculated plants was used as a negative control. Plants were inoculated and subjected to drying conditions as described in the experimental procedures section. After 33 days in the absence of water, both non-inoculated pepper plants and plants inoculated with desiccation-sensitive P. putida KT2440 showed signs of dying (Figure 1A). Similar results were seen when the plants were inoculated with A. siccitolerans 4J27, Leucobacter sp. 4J7B1 and Rhodococcus sp. 4J2A2. However, pepper plants inoculated with the most desiccation-tolerant isolates, Microbacterium sp. 3J1 and A. koreensis 5J12A, as well as with A. piechaudii 366-5, appeared healthy and yielded the highest values of DW (Figure 1B), FW (Figure 1C), FTW (Figure 1D), and RWC (Figure 1E), and had the longest roots (Figure 1F) and stems (Figure 1G) at the end of the experiment. Soil moisture was reduced from 71.52 to 13.75%. No statistical differences were found in the soil moisture regardless of the applied strain applied. Analogous results were found when tomato plants were used, thus validating these findings in a different plant species (Supplementary Figure S1). However, some minor but significant protection against drought was detected in tomato plants when Rhodococcus sp. 4J2A2 was used as the inoculant. These results suggest that some plant species are more sensitive than others to certain xerotolerant strains, since Rhodococcus sp. 4J2A2 showed a xeroprotectant effect in tomato but not in pepper plants under our assay conditions. In addition, these results suggest that only the strains showing the highest desiccation tolerance protect certain plants against drought.

To determine whether this protective effect was a specific characteristic of the isolate itself or was a property shared by strains that are closely related taxonomically, we compared the sequence for the 16S rRNA gene in strains Microbacterium sp. 3J1 (GenBank accession number GU815136) and A. koreensis 5J12A (GenBank accession number GU815140; Narváez-Reinaldo et al., 2010) with the sequence from the EzTaxon server corresponding to the same gene¹ in order to identify the most closely related strains. The nearly complete sequence for the 16S rRNA gene in strains Microbacterium sp. 3J1 and A. koreensis 5J12A (approximately 1,500 bp) was aligned with the sequences in closely related species of the genera Microbacterium and Arthrobacter, including A. piechaudii ATCC 43552 as an outgroup. The resulting neighbor-joining trees are shown in Figure 2.

Once, we determined their taxonomic affiliations, the bacterial strains identified as most closely related to Microbacterium sp. 3J1 were M. foliorum P333/02 (Behrendt et al., 2001) and M. phyllosphaerae P369/06 (Behrendt et al., 2001), while the strains found to be most closely related to A. koreensis 5J12A proved to be A. koreensis CA15-8 (Lee et al., 2003) and A. luteolus CF25 (Wauters et al., 2000). Further comparisons showed that the desiccation tolerance provided by Microbacterium sp. 3J1 and A. koreensis 5J12A appeared to correlate with the degree of drought protection they provided to the plant. In other words, plants were most tolerant to drought when they were inoculated with Microbacterium sp. 3J1 and A. koreensis 5J12A, according to the high values, we found for desiccation-tolerance parameters. The intermediate levels of desiccation tolerance provided by M. foliorum P333/02, M. phyllospherae P369/06, A. koreensis CA15-8, and A. luteolus CF25 were translated into intermediate levels of protection of the plant against drought.

In light of our results and earlier findings in the genera studied here, we suggest a possible correlation between the degree of desiccation tolerance in a given PGPR and the level of protection against drought it confers to the plant. To identify the molecule responsible for plant protection, we tested the strategies most commonly used by RDTE. Our panel of assays tested the in vitro production of phytohormones by the microorganisms involved in protection against desiccation (IAA, GA₃, ABA, and SA) as well as ACC consumption (Table 2), antioxidant molecules involved preventing and repairing damage caused by reactive oxygen species (ROS) [superoxide dismutase (SOD) and catalase (CAT)] (Figure 3), and trehalose as a xeroprotectant involved in the conservation of essential biomolecules (Figure 4). Our ANCOVA (one-way ANOVA, p < 0.05) results showed a correlation between plant-growth parameters, e.g., DW, FW and FTW, RWC, RL, and shoot length (SL), and the concentration of trehalose produced by the microbial cells (Figure 5). This analysis indicates that the highest trehalose production by the strain Microbacterium sp. 3J1 correlated strongly with the highest values of DW, FW, RWC, and RL in plants inoculated with this strain, as documented by coefficients of determination, R² above 0.8. Moreover, trehalose production by P. putida KT2440 (a desiccation-sensitive strain) and the rest of the analyzed strains analyzed showed similar correlations, which indicate that the values of all these parameters depend on trehalose production. This molecule which can stabilize essential biomolecules in plant cells, has been linked to the modulation of plant metabolism via the induction of stress-response genes, mostly transcriptional factors and protein kinases (Schluepmann et al., 2004). Bae et al. (2005) found that 30 mM trehalose, when added exogenously, was linked to the regulation of ethylene- and jasmonate-signaling pathways. The presence of trehalose thus appears to lead to cross-talk with ABA signaling in different physiological processes, including the regulation of stomatal aperture size (Gómez et al., 2010). Inoculation of Medicago truncatula with an IAA-overproducing strain of Sinorhizobium meliloti led to the accumulation of trehalose in the bacterial cell, and re-modulation of the plant’s phytohormone profile (Bianco and Defez, 2009). These results are evidence of a two-way interaction system between trehalose as the elicitor produced by the bacteria and the phytohormone regulatory map of the plant. In addition, exogenously added trehalose has been associated with protection against oxidative damage by reducing ROS accumulation, increasing non-enzymatic antioxidants, and co-activating the antioxidative and glyoxalase system (Mostofa et al., 2014). A more in-depth discussion of these bi-directional regulatory mechanisms was published by Lunn et al. (2014) in a recent review of trehalose metabolism in plants (Lunn et al., 2014).

Salt stress in plants is known to induce the accumulation of ROS species such as H₂O₂. To avoid or alleviate the risk of ROS accumulation, plants respond by activating different scavenging pathways (Fujita et al., 2006). Among the scavenging systems, SOD is the primary scavenger in the detoxification of ROS induced by oxidative stress in plants. Other scavenging enzymes, including the ascorbate–glutathione cycle enzymes (ascorbate peroxidase and glutathione reductase) and CAT, also play an important role in detoxifying of the toxic cellular products of SOD (i.e., H₂O₂) in plant cells (Mittler, 2002). Moreover, exogenous trehalose is able to differentially modulate antioxidant enzymes and the expression of related plant genes (Nounjan et al., 2012). To elucidate the role of trehalose production by the microorganisms in protecting the plant against desiccation applied to the plant, we assayed the effect of otsAB genes from Microbacterium sp. 3J1 that were expressed in P. putida KT2440 under the constitutive expression of the Plac promoter in the broad-host-range vector pUCP22 in order to generate the pUCP22:otsAB vector (West et al., 1994). The trehalose concentration in P. putida KT2440 (pUCP22:otsAB) increased more than one order of magnitude (from 12.12 to 134.14 ppm/CFU(10⁸)⋅ml^-1) compared to the concentration in P. putida KT2440 (pUCP22), and was nearly half the trehalose production by Microbacterium sp. 3J1 when grown in the presence of PEG 50% (-2.64 MPa). In addition, when we compared the desiccation tolerance of pepper plants inoculated with the trehalose-overproducing P. putida KT2440 (pUCP22:otsAB) to plants inoculated with P. putida KT2440 (pUCP22), we found a significant increase in desiccation tolerance in the former based on the ANOVA test (p < 0.05). DW of the plants inoculated with P. putida KT2440 (pUCP22:otsAB) was 1.13- and 1.2-fold higher than for plants inoculated with P. putida KT2440 (pUCP22) and uninoculated plants, respectively, while we observed an increase in RWC of 1.21-fold compared to plants inoculated with P. putida KT2440 (pUCP22) and 1.68-fold compared to non-inoculated plants (Figure 6).

These results point to a possible link between the degree of desiccation tolerance imparted by trehalose production by a given PGPR and the level of protection against drought that the organism confers to plants. Isolation assays have been performed to select RDTE from different locations such as quarry sand and riverside sand. All genera isolated in these studies belonged to Actinobacteria, one of the groups with the highest desiccation-tolerance values. This characteristic may reflect their natural abundance in soils, which may be explained by the broader tolerance of this genus to abiotic stress such as drought, extreme cold and starvation (Deutch and Perera, 1992; Zevenhuizen, 1992). We speculate that the high desiccation tolerance of these microorganisms enables the survival of a sufficient number of cells to colonize and establish interactions with the plant prior to the establishment of the symbiotic protection of the plant.

This is the first report available to link desiccation tolerance of a microorganism to the protection they confer to plants against drought. Plant-survival rate seems to correlate with the ability of the microbial cell to produce trehalose. The trehalose produced by the microorganism may control most of the plant’s enzymatic and non-enzymatic responses by favoring the production of the plant’s repertoire of phytohormones.

We propose that the ability of microorganisms to produce trehalose under drying conditions can facilitate their own survival and that of the plant being colonized. This makes the selection of highly desiccation-tolerant PGPR potentially useful in agriculture as biostimulants to protect plants against drought.

JV, CG-F, and DR-N performed the experimental assays. JV, JG-L, and MM have performed the statistical analysis. JV and MM have designed and written the article.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.