Trehalose in rhizobia is a key compound for signaling plant growth, yield and adaptation to abiotic stress, and its manipulation has a major agronomical impact on leguminous plant. In this work we reconstructed trehalose metabolism in R. etli, and investigated the role of trehalose in the response to high temperature and desiccation stress, as well as symbiotic performance. By using13C-NMR, we showed that besides trehalose as the major compatible solute, R. etli CE3 also amasses glutamate. In addition, it can accumulate mannitol if present in the external medium. The same compatible solute profile was recently reported for the strain R. etli 12a3, isolated from P. vulgaris nodules in Tunisian fields
Two successive genome-based metabolic reconstructions of R. etli have been reported, covering in total 405 reactions and 450 (but not trehalose-related) genes
[57, 58]. In this study, we reconstructed the metabolism of trehalose in R. etli, including trehalose uptake, degradation, and synthesis (see Figure
2). Our data suggest that uptake and catabolism of trehalose in R. etli uses the same pathways as in S. meliloti, since orthologs to the S. meliloti AglEFGK/ThuEFGK ABC trehalose/maltose/sucrose transporters
[22, 23], as well as the ThuAB catabolic route
, were found in R. etli. In addition, R. etli genome accounts for up to 3 putative copies of the trehalose-6-phosphate hydrolase (TreC). Only TreC3 was in the same group as the characterized TreC protein from E. coli, suggesting that the other copies might have a slightly different function. Interestingly, treC2 (annotated as aglA) was located upstream of the aglEFGK genes encoding the alpha-glucoside ABC transporter. In S. meliloti, aglA, encoding an alpha-glucosidase with homology to family 13 of glycosyl hydrolases, forms part of the aglEFGAK operon, suggesting a possible function in sucrose, maltose and/or trehalose catabolism. Further work is necessary to elucidate the role of the different systems involved in trehalose transport and degradation in R. etli.
Regarding trehalose synthesis, Suarez et al.
 already suggested the presence in R. etli of the three trehalose biosynthetic pathways so far known in rhizobia (OtsAB, TreS, and TreYZ). In this work, we precisely located the corresponding genes, and proposed the most plausible route of glucose synthesis from mannitol, and subsequent OtsAB-mediated trehalose synthesis (see Figure
We found that genes for trehalose metabolism were scattered in the genome, and sometimes present in more than one copy (i.e., otsA, treZ, treS, treC). This high enzyme redundancy seems to be a general characteristic of R. etli CFN 42, and was proposed to correlate with the different degrees of metabolic responses and alternative regulation necessary to cope with a challenging environment without compromising the integrity of the pathways
. Despite the active site residues were conserved in the p42a-encoded copy of OtsA, our phylogenetic analysis, together with the presence of insertion sequences flanking the gene, and its different codon used (if compared to the R. etli chromosome), strongly suggests that otsAa was acquired by lateral transfer. All these findings agree with the proposal by González et al.
 about an exogenous origin for R etli p42a.
The role of trehalose in the osmostress response has been widely demonstrated in many bacteria, including S. meliloti, B. japonicum and R. etli. In the former species, the involvement of trehalose in osmoadaptation was proposed based on three findings: (i) trehalose accumulation in the wild type was osmoregulated, (ii) an otsA mutant was osmosensitive, and (iii) overexpression of otsA led to an increased osmotolerance. Our results confirm the previous result that trehalose biosynthesis in R. etli is triggered by osmotic stress. However, the otsAch mutant reported in this work was much less affected by NaCl stress than the otsA mutant described by Suarez et al.
. These authors tested osmosensitivity in a glycerol minimal medium with 0.5 M NaCl during 48 h. In contrast, we found that the R. etli wild type strain could not grow above 0.2 M NaCl in B- mannitol minimal medium. Therefore, it is possible that the otsAch mutant described here might show an increased osmosensitivity at higher salinities. On the other hand mannitol, which was accumulated as an osmoprotectant (see Figure
4A), might have partially restored the growth of the otsAch strain when it was used as a carbon source.
Notably, extracts of otsAch cells grown with mannitol contained large amounts of glutamate, which was the predominant compatible solute (see Figure
4C). Thus, glutamate seems to be important for the long term adaptation of R. etli to osmotic stress, at least in the otsAch mutant strain describe here. Very interestingly, growth of the otsAch mutant was also affected in the absence of salinity stress (see Figure
Additional file 3: Figure S2), suggesting an important role of trehalose in R. etli physiology. Trehalose has been described to be essential as cell wall and membrane precursor
, as membrane stabilizer
, or as antoxidant
, to give some examples. This apparent essentiality of trehalose for normal growth of R. etli deserves further investigation.
A high level of trehalose accumulation is an important factor in the heat shock response in yeast
. In addition, bacteria such as E. coli and S. enterica serovar Typhimurium accumulate trehalose in response to heat stresses, and transcription of the otsAB genes for trehalose synthesis is thermoregulated
[27, 62]. In this work, we show the relevance of trehalose for R etli tolerance to high temperature. Although, trehalose content in R. etli cells grown at high temperature was very low, these levels were apparently enough as to protect wild type cells against heat stress, as growth of the trehalose deficient otsAch strain was impaired at high temperature. Similarly in E. coli, stationary phase induced thermotolerance has been shown to depend upon the rpoS regulated expression of the otsAB genes for trehalose synthesis, but the levels of trehalose synthesized on entry into stationary phase were very much lower than in osmotically stressed cells
. There is now a large body of evidence showing that the mechanisms for trehalose-mediated protection against heat and desiccation stress are different from those involved in osmoprotection, i.e., as a counteracting osmolyte. Thus, studies in vitro have shown that trehalose preserves structure and function in biomolecules and molecular assemblages, such as membranes, during drying and heat stress
Strains of R. leguminosarum bv trifolii and R. etli (this work) deficient in trehalose synthesis are more sensitive to the effects of drying, and show impaired survival upon storage. Thus, desiccation tolerance in R. etli cells was dependent of high trehalose production by osmotic pre-conditioned cells. Indeed, desiccation stress is much more harmful than heat stress for microorganisms, as it produces the accumulation of salt and solutes, hyperosmotic stress, metabolism impairment, and damage to macromolecules upon removing the aqueous monolayer
. This may explain why high trehalose content is necessary for survival of R. etli cells to drying, in order to cope with so many stresses. In agreement with this, E. coli, S. meliloti, and desert-isolated rhizobial strains nodulating acacia
 that were osmotically induced to accumulate trehalose (and also mannosucrose, in desert-isolated rhizobia), showed increased tolerance to drying and storage. Interestingly, transcriptomic analyses revealed that desiccation stress per se, if performed under controlled conditions, also induced trehalose synthesis by B. japonicum, the soil actinomycete Rhodococcus jostii and the yeast Saccharomyces cerevisiae.
It is worth mentioning that desiccation tolerance by R. etli was not improved by an increase in drying temperature. This lack of correlation has been also found in many other rhizobia
 and could be attributed, at least in R. etli, to the low induction of trehalose synthesis under high temperature. On the other hand, the survival rate of R. etli wild type strain after the vacuum-drying treatments was below 40%, and rapidly decreased after 4 days storage (see Figure
6). This differs from the high survival rates found for S. meliloti on nitrocellulose filters
 or R. leguminosarum bv trifolii on glass beads
. Rather than intrinsic tolerance to desiccation, we suggest that these differences may be related to the experimental conditions used for drying.
In rhizobia, the relationship between inactivation of a given trehalose metabolic pathway (and the resulting trehalose accumulation) and the observed symbiotic performance, seems to vary among species (see Introduction). The R. etli otsA mutant reported by Suarez et al.
 was affected in its capacity to establish an efficient symbiosis with bean plants. However, bacteroids of the R. etli otsAch mutant constructed in this work showed the same trehalose levels than those of the wild type, and were not affected in its symbiotic performance. The reasons for these differences remain to be elucidated, but it is plausible that under the conditions used in our symbiosis experiments other trehalose synthesis pathways were activated in the otsAch strain, including the otsAa copy, that may compensate the lack of otsAch. Thus, our results do not preclude a role of trehalose in the R. etli
Phaseolus vulgaris symbiosis.
In its natural habitat, soil bacteria as R. etli are subjected to fluctuating osmotic, temperature and desiccation constrains. Improving trehalose production in R etli has been shown to be a useful strategy to achieve drought tolerance of the bean plant host
. In this work, we have shown that trehalose is essential for R. etli survival to high temperature and drying under free living conditions. Thus, engineering trehalose accumulation promises to be useful to improve survival of R. etli-based inoculants during desiccation stress in storage, upon application to seeds, or once released in fields.