Effect of trehalose on survival of Bradyrhizobium japonicum during desiccation


J.G. Streeter, Department of Horticulture & Crop Science, Ohio State University/OARDC, 1680 Madison Avenue, Wooster, OH 44691 USA (e-mail: streeter.1@osu.edu).


Aims: A major reason for the ineffectiveness of legume inoculants in the field is the rapid death of rhizobia because of desiccation. The major purpose of this study was to identify conditions under which α,α-trehalose would improve survival of Bradyrhizobium japonicum during desiccation.

Methods and Results: Trehalose was added to cultures just prior to desiccation or was supplied to bacteria during the 6-day growth period. A wide variety of trehalose concentrations was tested. Trehalose added to cultures at the time of desiccation improved survival slightly, but trehalose loading during growth was much more effective in protection against desiccation. Growth of bacteria with 3 mmol l−1 trehalose increased trehalose concentration in cells by about threefold and increased survival of cells placed on soya bean [Glycine max (L.) Merr.] seeds by two- to four-fold after 2 or 24 h. Average of overall results indicate that growth of bacteria with trehalose in the medium resulted in a 294% increase in survival after 24 h of desiccation. The concentration of trehalose in cells was very highly correlated with survival of bacteria. When trehalose-loaded cells were suspended in buffer or water, 60–85% of cellular trehalose was lost in about 1 h and, in spite of these losses, survival during desiccation was not reduced.

Conclusions: Accumulation of trehalose in the cytoplasm is critical to the survival of B. japonicum during desiccation. Increasing the periplasmic concentration of trehalose is also beneficial but is not so critical as the concentration of trehalose in the cytoplasm. Because B. japonicum cannot utilize trehalose as a carbon source, cells can be loaded with trehalose by providing the disaccharide during the growth period.

Significance and Impact of the Study: Although it may not be practical to use trehalose as a carbon source in inoculant production, it may be possible to engineer greater trehalose accumulation in rhizobia. Trehalose concentration in cells should be a useful predictor of survival during desiccation.


When legume seeds are inoculated with the appropriate species of rhizobia prior to planting in the field, the vast majority of nodules formed are not formed by the bacteria in the inoculant but by indigenous strains in the soil (Miller and May 1991; Streeter 1994). Although improved inoculant formulations have improved survival of the bacteria during storage of the product (Streeter and Smith 1998; Catroux et al. 2001), these efforts have not provided much improvement in bacterial survival on the seed or in the soil (Brockwell and Bottomley 1995).

The rapid death of rhizobia after being applied to the seeds has been reported by numerous studies (Bushby and Marshall 1977a; Salema et al. 1982; Weaver and Holt 1990; Kosanke et al. 1992; Roughley et al. 1993). A typical study is that by Roughley et al. (1993), where 95% of the rhizobia died between inoculation of the seed and planting (about 4 h) and 83% of those surviving died after another 22 h in the soil. Several workers have attempted to improve survival by adding a variety of materials to the growth medium or to the inoculant at the time of application to seeds; additions include glucose, maltose, sorbitol, sucrose and cellobiose, glutamate, polyvinylpyrrolidone, montmorillonite clay and gum arabic (Vincent et al. 1962; Bushby and Marshall 1977a; Salema et al. 1982; Cliquet and Catroux 1994). Generally, either the amount required for protection is unrealistic [e.g. 9% maltose (w/v) Vincent et al. 1962] or the beneficial effect of the addition is not significant. Other strategies for improving delivery of high numbers of viable bacteria include massive inoculant doses (Dunigan et al. 1983), use of nutrient-added granules (Fouilleux et al. 1996), and preparation of inoculants such as silicate, acrylamide, starch, or cellulose gels (Jawson et al. 1989).

It is clear that the main cause of bacterial death following application to seeds is rapid desiccation. The main evidence for this are studies where inoculated seeds were maintained at high relative humidity (Mary et al. 1994) or in a moist condition (Salema et al. 1982); under these conditions, survival of applied bacteria was markedly higher. In addition, slow re-hydration of desiccated bacteria improves survival (Kosanke et al. 1992). The effectiveness of gels in delivering viable bacteria to the soil is consistent with desiccation being the major cause of cell death (Jawson et al. 1989). Studies on membrane permeability of rhizobia subject to desiccation indicate that damage to the cytoplasmic membrane may be the main cause of death (Bushby and Marshall 1977b).

The disaccharide α,α-trehalose has been reported to stabilize membrane structure under conditions of desiccation (Crowe et al. 1984). The mechanisms underlying the unique efficacy of trehalose are still under discussion and may include glass formation (Aguilera and Karel 1997) and formation of anhydrous crystals capable of reversibly absorbing water (Sussich et al. 2001). Two recent studies served as the impetus for the work reported here. The first is a report showing that addition of 100 mmol l−1 trehalose during drying of Escherichia coli or Bacillus thuringiensis improved survival markedly when compared with drying with 100 mmol l−1 sucrose (Leslie et al. 1995). In a more recent study, the gene for sucrose-6-phosphate synthase from Synechocystis sp. was transferred to E. coli and the transformed strain accumulated sucrose and survived significantly better than wild type cells following freeze drying, air drying or desiccation over P2O5 (Billi et al. 2000). The latter study suggested that internal concentrations of disaccharide might be more critical than periplasmic or extracellular disaccharide, a suggestion that is consistent with the results reported here.

Both cultured bacteria and bacteroids of Bradyrhizobium japonicum accumulate trehalose and it is somewhat peculiar in that these bacteria do not grow on trehalose as a sole carbon source (Streeter 1985). In studies reported here, accumulation of trehalose was enhanced by supplying the disaccharide in culture media, and it was found that trehalose-loaded cells survived significantly better during subsequent desiccation stress.

Materials and methods

Bacterial cultures

Bradyrhizobium japonicum USDA 110 (National Rhizobium Germplasm Resource Center, USDA/ARS, Beltsville, MD, USA) was used in all studies. This strain was chosen because it grows well (although slowly) on a defined medium (Cole and Elkan 1973). This medium contains only salts, with ammonium chloride as the N source, MES/HEPES buffer, pH 6·75, and arabinose and gluconate as carbon sources. Amendments to this medium, referred to hereafter as ‘AG’, will be described with each appropriate experiment. Stock cultures of 110 were maintained in liquid AG at 4°C and subcultured every few months. Cultures from the same stock were occasionally used for inducing nodule formation on soya beans in other unrelated studies.

For each study, replicate 5 ml cultures were grown in 18 × 150 mm test tubes slanted in test tube racks and shaken at 220 rev min−1 on an orbital shaker at 30°C. Unless indicated, cultures were used after 6 or 7 days of growth, i.e. stationary phase cultures were used for most tests of survival. Sugars added to cultures were purchased from Sigma (St Louis, MO, USA). Difco yeast extract and agar were purchased from Fisher Scientific (Pittsburgh, PA, USA). Maltoheptaose and maltopentaose were obtained from Dr Kazuhiko Murata (see Streeter and Bhagwat 1999).

Analysis of sugar composition of cells

Following sampling of cultures for survival analysis, cells were collected by centrifugation in 40 ml plastic tubes at 48 000 × g for 10 min. The supernatant was discarded and the cells were suspended in 500 μl of triple deionized (TDI) water and transferred to a 1·5 ml microfuge tube. Another 500 μl of TDI water was used to rinse the centrifuge tube and added to the microfuge tube. The cell suspension was then centrifuged at 10 000 × g for 5 min. The water wash was discarded and the cells were suspended in 500 μl of 75% (v/v) ethanol. This suspension was occasionally vortexed over a period of at least 6 h to extract the cell metabolites. The suspensions were then centrifuged again at 10 000 × g and the supernatant was transferred to a small vial and dried under a stream of air at 35°C. Carbohydrate composition of extracts was determined by gas–liquid chromatography (GLC) (Streeter 1985).

It is important to wash the cells prior to extraction in order to remove sugars present in the medium. However, this washing procedure should be performed quickly because of the release of sugar from the cells suspended in water (see Results). Typically the time between the suspension of cells in water and the suspension of the washed cells in ethanol was <6 min, during which sugar loss would be negligible. Sugar loss was reduced by suspending the cells in 11 mmol l−1 MES/HEPES buffer (the concentration of buffer in the culture medium), but the presence of residual buffer in the cell extracts complicated the sugar analysis. Using the procedure described, a single ethanol extraction essentially gave 100% recovery of sugars.

Trehalose ‘washout’ experiments

In the first experiment, cells were collected from a 100 ml culture containing 10 mmol l−1 trehalose by centrifugation at 9800 × g. After discarding the supernatant, cells were suspended in 11 mmol l−1 MES/HEPES buffer, pH 6·75 and immediately centrifuged a second time to remove medium components. Supernatant was discarded and cells were immediately suspended in 50 ml of sterile TDI water. Over time, 3·0 ml samples were taken and microfuged at 10 000 × g for 5 min. The supernatant was transferred to a GLC vial and dried for analysis of carbohydrates. Cells were suspended in 1·0 ml of 75% ethanol and handled as described above. In this experiment, survival of cells on seeds was not tested.

In the second experiment, 5·0 ml cultures were grown with 3 mmol l−1 trehalose. Cells were collected by centrifugation and washed in 11 mmol l−1 buffer as described above. After discarding the wash, each sample of cells was suspended in 5·0 ml of sterile TDI water. At selected time intervals a portion of each sample was taken for analysis of survival of bacteria on the seeds and, at the same time, cells from the remainder of the sample were collected by centrifugation and extracted for analysis of carbohydrates.

Bacterial desiccation experiments

In preliminary trials, survival of bacteria applied to seeds and various discs (paper, nylon, glass fibre) was compared. A source of soya bean seed (Renk variety RS208RR, seed lot YJBSXA) that gave minimal contamination was identified, and these seeds were used in subsequent studies in order to provide a more realistic experimental system. Sixty-eight seeds (ca 10 g) were used for each sample. Seeds were counted and placed in sterile 50 ml Erlenmeyer flasks using tweezers. For application of bacteria to seeds, 30 μl of culture was added to the flask and distributed to the surface of the seeds by rotating the flask for about a minute. This volume of liquid is based on recommendations of inoculant manufacturers for the use of liquid soya bean inoculants. With this dose, seeds appear to be uniformly wetted and very little liquid adhered to the sides of the flask. In calculating the recovery of bacteria, it was assumed that there was no loss of liquid in the flask. The glossy appearance of seeds lasted only for a few minutes because of the absorption of the liquid into the seed coats. After drying, seeds were transferred to sterile Petri dishes for incubation at room temperature. At the same time when cultured bacteria were applied to seeds, a sample of the culture was also taken for serial dilution (5 × 10−7) and three replicate 100 μl aliquots were spread on AG plates.

For analysis of survival, samples were taken after 2 and 24 h. These time periods provided an estimate of the short- and long-term death. Ten seeds were transferred to a 18 × 150 mm test tube containing 4 ml of sterile TDI water and the mixture was vortexed several times to wash off the bacteria. A sample of the washate was serially diluted to a final dilution of 10−6 and four replicate aliquots were plated on AG agar for analysis of surviving bacteria. Colonies on plates were counted after an incubation of 7 days at 30°C. Data are expressed on a CFU ml−1 basis for surviving cells. This calculation is based on the theoretical portion of the 30 μl inoculant dose on the 10-seed sample, namely 4·4 μl. This expression of the data is chosen to allow direct comparison of surviving CFU with initial CFU in the cultures. For converting the CFU ml−1 for the 2- and 24-h samples to CFU per seed, the values are multiplied by 4·41 × 10−4.

In one experiment, bacteria that had survived for 24 h on soya bean seeds were collected and put back on the seeds for a second survival analysis. After 20 seeds were taken for the 2- and 24-h analyses, the remaining 48 seeds were placed in a sterile flask containing 20 ml of AG liquid medium. After swirling for about 10 s, the liquid was poured into a sterile 40 ml centrifuge tube and the suspension centrifuged at 48 000 × g. The supernatant was discarded and the tiny pellet was suspended in 100 μl of AG medium and transferred to a sterile microfuge tube. A sample was taken for analysis of zero-time CFU and 30 μl was put on a fresh sample of 68 soya bean seeds for analysis of survival at 2 and 24 h.


In preliminary trials, the effect of 10 mmol l−1 trehalose, sucrose, or maltose in the culture medium on survival on soya bean seeds was compared. Sucrose gave small but erratic increases in survival and maltose gave no increase in survival (data not shown). Cells accumulated sucrose but did not accumulate maltose; i.e. following growth in medium containing 10 mmol l−1 maltose, no maltose could be detected in cell extracts. Maltose was chosen as a control compound to be used in subsequent studies to balance osmotic effects caused by the addition of trehalose.

Based on the report of Leslie et al. (1995) where 100 mmol l−1 trehalose was used, relatively high concentrations of trehalose were tried initially (Table 1). Growth of B. japonicum USDA 110 with 3, 6, or 9 mmol l−1 trehalose increased the concentration of trehalose in cells in a stepwise fashion. Initial CFU was not influenced by trehalose addition, but survival after 2 or 24 h was significantly increased. Six or 9 mmol l−1 trehalose showed no advantage over 3 mmol l−1 trehalose. Percentage loss data show that most bacteria died in 2 h on the soya bean seed, but the percentage loss was muted by growth with trehalose. Growth with trehalose also lowered percentage loss from 2 to 24 h. Although the proportional loss of CFU after 24 h was only lowered from 97 to 92%, the increase in surviving bacteria was threefold. Trehalose concentration in cells was significantly correlated with surviving CFU after 2 h (r = 0·758, P = 0·01) or 24 h (r = 0·630, P = 0·05).

Table 1.  Effect of trehalose added to the culture medium on survival of Bradyrhizobium japonicum USDA 110 during desiccation stress*
Trehalose in medium (mmol l−1)Trehalose in cellsCFU (×107) (ml−1)
Initial cultureDesiccation period†Loss of CFU over time (%)‡
μg ml−1μg CFU (×10−9)2 h24 h0 to 2 h2 to 24 h0 to 24 h
  1. *The medium used here contained 50 mg of yeast extract per litre. Maltose was used to balance trehalose in the cultures (see text). There were three replicate cultures per treatment and cultures were incubated for 7 days prior to analysis. The values given are mean (s.e.).

  2. †CFU values were calculated on the basis of the cultures applied to seeds.

  3. ‡[(Initial CFU − final CFU)/initial CFU] × 100.

02·1 (0·1)0·49 (0·02)416 (7)63·3 (5·9)12·1 (1·6)84·8 (1·4)81·0 (0·8)97·1 (0·4)
36·6 (0·3)1·59 (0·11)420 (27)137·8 (7·3)35·1 (2·2)66·9 (2·5)74·5 (1·3)91·5 (1·0)
610·3 (0·7)2·36 (0·21)439 (18)121·9 (1·5)32·9 (2·1)72·1 (1·4)73·0 (1·6)92·5 (0·5)
912·5 (0·1)3·16 (0·38)409 (53)137·4 (11·4)27·7 (2·5)65·9 (1·8)79·7 (2·1)93·0 (1·1)

In the second experiment, lower concentrations of trehalose were used (Table 2). Two differences between the results for the two experiments are that initial CFU were lower in the second experiment and initial CFU were increased by the addition of trehalose to the medium. These differences may be related to the fact that a low concentration of yeast extract was used in the medium for the first experiment. The yeast extract (Difco, commonly used in media) contained 4·3% trehalose, and hence the positive effect of yeast extract on initial CFU may be related, in part, to trehalose supply. In the second experiment (Table 2), surviving CFU were only about 50% of that in the first experiment; however, proportional losses were similar in the two experiments. Although 1 and 2 mmol l−1 trehalose supplied in the growth medium significantly increased survival on soya bean seeds, 3 mmol l−1 trehalose yielded slightly higher CFU (especially at 2 h) and this concentration was used in most of the subsequent experiments. The correlation between trehalose in cells and surviving CFU was very high at time 0 (r = 0·986), 2 h (r = 0·981) and 24 h (r = 0·904) (all coefficients significant at 1%).

Table 2.  Effect of lower concentrations of trehalose added to the culture medium on survival of Bradyrhizobium japonicum USDA 110 during desiccation*
Trehalose in medium (mmol l−1)Trehalose in cellsCFU (×107) (ml−1)
Initial cultureDesiccation periodLoss of CFU over time (%)
μg ml−1μg CFU (×10−9)2 h24 h0 to 2 h2 to 24 h0 to 24 h
  1. *The medium in this experiment contained no yeast extract. Other details are as in the footnotes for Table 1.

02·4 (0·2)1·01 (0·11)247 (3)32·1 (0·9)7·5 (0·1)87·0 (0·3)76·6 (0·6)97·0 (0·1)
15·3 (0·1)1·73 (0·06)309 (7)70·9 (0·6)20·4 (0·6)77·0 (0·7)71·1 (1·1)93·4 (0·1)
28·0 (0·3)2·20 (0·06)363 (5)96·0 (5·1)27·4 (2·1)73·5 (1·7)71·0 (3·5)92·4 (0·5)
311·8 (0·4)2·55 (0·09)466 (10)149·1 (2·4)29·7 (1·7)68·0 (0·8)80·0 (1·4)93·6 (0·5)

To explore the effect of yeast extract further, bacteria were grown in AG medium but with the addition 0·5, 1·0, or 1·5 g of yeast extract per litre; control medium contained no yeast extract. A typical concentration of yeast extract in media used for growing rhizobia is 1 g l−1. Although trehalose is a major constituent of yeast extract, the amount of trehalose supplied in the medium as small (0·06–0·17 mmol l−1) relative to the 3 mmol l−1 standard addition of trehalose. The addition of yeast extract to the medium increased trehalose concentration in cells to about 2·5-fold and increased survival about to 50% after 2 or 24 h for the 0·5 and 1·0 g yeast extract treatments (data not shown). These effects, although positive, were not nearly so large as those provided by adding 3 mmol l−1 trehalose to the medium (Tables 1 and 2). For the 1·5 g yeast extract treatment, CFU recovered from seed after 2 h were not different from the control and, after 24 h, CFU were actually reduced by about 40%.

Although the standard incubation time prior to analysis of on-seed survival was 6 or 7 days, it was of interest to examine log-phase cells. To perform this, bacteria were grown with 3 mmol l−1 maltose or trehalose and cultures were sampled after 2, 3, or 4 days (Table 3). Trehalose concentration in cells increased over time under both growth conditions. Survival of cells applied to seeds was very poor for 2-day cultures but, after 3 or 4 days, it was similar to that observed in stationary-phase cultures (Table 2). The positive effect of trehalose on survival was evident in 3- and 4-day cultures and was similar to that of stationary-phase cultures. Initial CFU were similar in 2-day cultures but were significantly higher in trehalose-grown cultures compared with maltose-grown cultures after 3 or 4 days. The proportional loss rates in 2-day cultures was enormous; >99% of cells were dead after 24 h on seeds, and growth with trehalose made no difference. After incubation of cultures for an additional day, cells had accumulated some trehalose and were much better capable of withstanding the effects of desiccation. Very poor survival of low-density cultures of Rhizobium leguminosarum was improved by addition of culture supernatant from high-density cultures or by addition of N-acyl homoserine lactone (Thorne and Williams 1999), indicating that very low concentrations of trehalose in 2-day-old cultures may not be the only factor that influenced survival.

Table 3.  Effect of culture time on trehalose accumulation and survival of Bradyrhizobium japonicum USDA 110 during desiccation*
Culture additionCulture time (days)Trehalose in cellsCFU (×107) (ml−1)
Initial cultureDesiccation periodLoss of CFU over time (%)
μg ml−1μg CFU (×10−9)2 h24 h0 to 2 h2 to 24 h0 to 24 h
  1. *There were three replicate cultures per treatment; AG medium (no yeast extract). Disaccharide concentrations in the medium were all 3 mmol l−1. The values are given as mean (s.e.).

  2. †Small trehalose peaks were present in chromatograms but did not integrate.

  3. ‡Very low numbers of colonies on plates prevent accurate estimates of CFU.

Maltose2<0·10†<0·10†104 (7)6·5 (1·6)∼0·6‡93·5 (1·8)91·3 (1·8)99·5 (0·2)
3 0·4 (0·1) 0·16 (0·05)251 (17)55·0 (2·7)12·1 (0·8)77·8 (2·2)77·8 (2·4)95·1 (0·4)
4 1·3 (0·1) 0·49 (0·07)263 (12)70·8 (4·5)5·9 (1·0)73·2 (0·6)91·7 (1·2)97·8 (0·3)
Trehalose2 0·9 (0·1) 0·82 (0·06)107 (2)7·4 (0·3)∼0·6‡93·0 (3·9)91·6 (6·8)99·3 (0·5)
3 3·3 (0·2) 0·97 (0·06)338 (4)95·4 (3·6)21·2 (0·5)71·7 (1·2)77·7 (1·4)93·7 (0·2)
4 5·6 (0·1) 1·38 (0·08)408 (18)99·0 (3·7)16·9 (1·5)75·7 (1·1)87·3 (3·1)95·8 (0·5)

In the experiments described so far, cultured cells were applied to seeds without first collecting and washing the cells. Thus, 3 mmol l−1 maltose or trehalose was applied to seeds along with the cells. In preliminary experiments, trehalose added to cultures just before application to seeds had only a small positive effect on survival, but this was tested again after the discovery of the much larger positive effects achieved with addition of trehalose to the growth medium. Bacteria were grown with either maltose or trehalose and, at the time of application to seeds, trehalose was also added to some cultures to give an increase in concentration of 3 mmol l−1 (Table 4). Results shown in rows one and three of Table 4 are similar to those shown in Table 2. When maltose-grown cells were amended with 3 mmol l−1 trehalose just before application to seeds, there was no effect on zero-time or 2-h CFU but there was a 75% increase in 24-h CFU. A similar increase was seen when additional trehalose was added to trehalose-grown cells. Accompanying these positive effects of trehalose addition were small but significant increases in trehalose concentration in the cells. In spite of the improved survival with trehalose addition at the time of application to cells, the multi-fold improvement in survival for trehalose-grown cells was maintained. Trehalose concentration in cells was significantly correlated (P = 0·01) with surviving CFU at time 0 (r = 0·792), 2 h (r = 0·990), and 24 h (r = 0·887).

Table 4.  Effect of trehalose addition at the time of seed application on survival of Bradyrhizobium japonicum grown with or without trehalose in the medium*
Sugar in mediumAdded before seed†Trehalose in cellsCFU (×107) (ml−1)
Initial cultureDesiccation periodLoss of CFU over time (%)
μg ml−1μg CFU (×10−9)2 h24 h0 to 2 h2 to 24 h0 to 24 h
  1. *The medium contained no yeast extract; disaccharide concentrations were 3 mmol l−1. Other details are as in Table 1.

  2. †Trehalose concentrate was added to give an increase in concentration of 3 mmol l−1. Thus, in treatment no. 4, the final trehalose concentration was 6 mmol l−1.

MaltoseWater2·5 (0·1)0·79 (0·03)317 (1)45·7 (5·8)9·6 (2·1)85·6 (1·8)79·1 (3·5)97·0 (0·7)
MaltoseTrehalose2·9 (0·1)0·94 (0·02)309 (17)46·5 (1·1)16·8 (0·5)84·9 (0·8)63·7 (1·7)94·5 (0·5)
TrehaloseWater8·1 (0·3)2·34 (0·02)344 (10)105·5 (3·0)23·3 (1·4)69·3 (0·3)77·9 (0·8)93·0 (0·2)
TrehaloseTrehalose9·3 (0·3)1·92 (0·07)483 (7)125·6 (1·3)38·2 (1·1)74·0 (0·6)69·6 (1·0)92·1 (0·3)

The above results suggest that cytoplasmic trehalose concentration may be more important than the concentration of trehalose in the periplasm. To estimate the relative concentration of trehalose in the two compartments of trehalose-loaded cells, bacteria were washed quickly and were suspended in sterile water and then sampled over a period of 3 h (details in Methods). A600 of the bacterial suspension did not change over this period (data not shown). The concentration of trehalose in cells dropped rapidly for the first 30 min and leveled off after about 1 h (Fig. 1). Trehalose concentration in the supernatant increased rapidly during the first sampling interval of 15 min but did not increase much after that time. Although data are not shown for the period from 2–3 h, trehalose concentrations in cells and supernatant were invariant during the period. Thus, it appeared that most of the trehalose in cells was in the periplasm and was lost rapidly. The peculiar thing about the result is that most of the trehalose that was lost from the cell fraction was not recovered in the supernatant fraction. No other carbohydrates – glucose in particular – were detected in any sample. Thus, periplasmic trehalose may have been hydrolysed and the resulting glucose metabolized rapidly. In a similar experiment where cells were suspended in 11 mmol l−1 MES/HEPES buffer instead of water, loss of trehalose from cells was only 60% of the total cellular trehalose compared with the 85% loss shown in Fig. 1; but most of the trehalose lost from cells was still not recovered in the supernatant.

Figure 1.

Loss of trehalose from Bradyrhizobium japonicum following suspension of trehalose-loaded cells in water. Concentration of trehalose in cells (∘) and in the supernatant (•) following collection of cells is shown. Samples consisted of 3·0 ml aliquots of a 50 ml cell suspension. The mean and s.e. of three observations were plotted and, for most data, s.e. bars are within the points

The experiment illustrated in Fig. 1 was repeated with fewer times and with application of bacteria to seeds at each time. The sampling times were 4, 20, 40 and 68 min and the trehalose concentration in cells at these times was 20·5, 12·2, 8·5, and 6·6 μg per 3·0 ml sample of cells, respectively. CFU for 0, 2, and 24 h were very similar to those shown for the 3 mmol l−1 trehalose treatment in Table 2 and CFU were not significantly different at any of the four sampling times. Thus, during the threefold drop in trehalose concentration in cells, on-seed survival of bacteria was not influenced. This result supports the importance of the cytoplasmic concentration of trehalose in improving survival of the bacteria subject to desiccation.

Rhizobia have two mechanisms for the biosynthesis of trehalose (Streeter and Bhagwat 1999) and one of these involves the use of malto-oligosaccharides (MOS) as substrates. An attempt was made to increase cytoplasmic trehalose concentration by adding 1·08 mg of maltoheptaose or maltopentaose per millilitre of medium (similar amounts of maltose or trehalose were used in control cultures). After a growth period of 7 days, trehalose concentration in cells was not increased by the MOS treatments. Initial, 2- or 24-h CFU for cells grown with MOS were also not greater than the cells grown with maltose. Survival of cells grown with trehalose was increased 3·8-fold relative to maltose-grown cells, as expected.

Two additional questions were tested. The first is whether cells that have survived on soya bean seeds represent a selective desiccation-tolerant portion of the original culture or whether death on the seed is simply a random event. The second question is whether the improved survival of trehalose-loaded cells will be amplified further when 24-h survivors are recovered and put back on seeds a second time. Surviving cells were washed off seeds using samples from the experiment described in the previous paragraph and cells were re-applied to new soya bean seeds as described in the ‘Methods’. Samples were too small to allow for analysis of trehalose in cells. The number of CFU recovered from seeds was 3·7-fold greater where bacteria had been grown with trehalose initially. Although the superior survival of the trehalose-grown cells was maintained in the second application to seeds, death of cells occurred at about the same rate regardless of the presence of trehalose in the original growth medium (Table 5). Thus, death of cells on seeds appeared to be a random process and the advantage of trehalose-grown cells was not amplified in the second cycle of desiccation.

Table 5.  Effect of trehalose in the culture medium on death rate of cells in the original culture vs cells that had survived 24 h of desiccation stress
Sugar in growth medium (3 mmol l−1)Cells on seedLoss of CFU over time (%)*
0 to 2 h2 to 24 h0 to 24 h
  1. *[(initial CFU − final CFU)/initial CFU] × 100.

MaltoseInitial culture85·3 (0·9)76·8 (2·4)96·6 (0·6)
Survivors72·3 (4·0)83·6 (5·1)95·1 (2·0)
TrehaloseInitial culture78·6 (0·4)48·6 (3·7)89·0 (0·8)
Survivors85·2 (2·0)54·7 (5·8)93·5 (0·2)


In general, inclusion of 3 mmol l−1 trehalose in a defined culture medium increased the survival of B. japonicum USDA 110 by two- to four-fold after 24 h. Addition of trehalose to the culture medium at the time of desiccation stress had a significant positive effect on survival, but the effect of trehalose ‘loading’ of cells during growth was much greater. Part of the positive effect of trehalose was because of production of higher initial CFU in cultures (Tables 2 and 4), but when death rates were normalized to initial CFU, survival was still significantly better where trehalose was provided in the medium. The concentration of trehalose in cells was significantly correlated with survival and many of the coefficients were very large. The most straightforward interpretation of these results is that cytoplasmic trehalose stabilized the cytoplasmic membrane during desiccation (Leslie et al. 1995).

Addition of yeast extract, which contains substantial amount of trehalose, also slightly improved the survival of bacteria during desiccation, but the main effect of yeast extract may be to increase maximum CFU in stationary-phase cultures (cf. Tables 1 and 2). Whenever a small amount of yeast extract (50 mg l−1) was included in the growth medium, addition of trehalose to the medium still had a strong positive effect on survival (Table 1). Survival was also increased by about twofold by the addition of 10 mmol l−1 trehalose to AG medium containing 500 mg trehalose l−1 (data not shown).

Experiments in which cells suspended in water lost their trehalose over time plus the superior effectiveness of trehalose loading during cell growth suggested that the cytoplasmic concentration of trehalose was most critical in enhancing survival during desiccation. However, addition of trehalose to cultures at the time of application to seeds also increased survival to some extent, even when cells had been pre-loaded with trehalose (Table 4). The failure to recover trehalose lost from cells in the supernatant may be explained by hydrolysis of periplasmic trehalose and rapid metabolism of the glucose formed; periplasmic disaccharidase activity has been reported in rhizobia (Streeter 1989). Alternatively, trehalose may have been directly but slowly absorbed and metabolized (Salminen and Streeter 1987).

A point not previously emphasized is that in cells cultured with maltose or with trehalose in the medium (in addition to the standard arabinose and gluconate carbon sources) contained only trehalose when extracted and analysed. (The GLC method quantifies carbohydrates with 5–12 carbons.) This result is consistent with previous results (Streeter 1985). Thus, the synthesis and retention of trehalose, in particular, appears to be important to B. japonicum. During these studies, the previous observation that B. japonicum cannot utilize trehalose as a carbon source for sustained growth (Streeter 1985) was confirmed.

Whether the results reported here can be exploited in the design of superior inoculants is not certain. A defined medium and a single strain of B. japonicum were employed here and almost endless variations in media and strains could be pursued. It is possible that similar positive effects of trehalose will not be obtained in the complex media used for inoculant production. Furthermore, the cost of including trehalose in media for the production of inoculants is probably prohibitive.

The importance of these results is that they demonstrate the effectiveness of trehalose in protection against desiccation under experimental conditions that are relevant to inoculant usage. It is possible that genetic modification of rhizobia to accumulate more trehalose or to retain more trehalose will improve the delivery of desirable bacteria to roots of legumes in the soil. A gene cluster responsible for trehalose uptake in Sinorhizobium meliloti has recently been identified (Jensen et al. 2002); production of strains with accelerated trehalose uptake but muted trehalose metabolism should be tried. In addition, analysis of trehalose accumulation by bacteria in inoculants may be useful in predicting the survival of the bacteria during desiccation stress in storage or upon application to seeds.


Salaries and research support were provided in part by State and Federal funds appropriated to the Ohio Agricultural Research and Development Center, the Ohio State University. This work was also supported in part by funds from LiphaTech, Inc. I thank Rocio Aviles-Nava for tireless effort and meticulous attention to detail, which made these results possible. I also thank Drs Stewart Smith and John Walsh for many helpful suggestions.