Most heat-tolerant rhizobia show high induction of major chaperone genes upon stress


  • Ana Alexandre,

    1. Laboratório de Microbiologia do Solo, Departamento de Biologia, ICAAM (Instituto de Ciências Agrárias e Ambientais Mediterrânicas), Universidade de Évora, Évora, Portugal
    Search for more papers by this author
  • Solange Oliveira

    1. Laboratório de Microbiologia do Solo, Departamento de Biologia, ICAAM (Instituto de Ciências Agrárias e Ambientais Mediterrânicas), Universidade de Évora, Évora, Portugal
    Search for more papers by this author

  • Editor: Kornelia Smalla

Correspondence: Solange Oliveira, Laboratório de Microbiologia do Solo, Departamento de Biologia, ICAAM (Instituto de Ciências Agrárias e Ambientais Mediterrânicas), Universidade de Évora, Apartado 94, 7002-554 Évora, Portugal. Tel.: +351 266 760 878; fax: +351 266 760 914; e-mail:


The aims of this study were to evaluate the temperature stress tolerance of chickpea rhizobia and to investigate whether tolerance is related to the species or the site of origin of the isolates. Additionally, the molecular bases of temperature stress tolerance in rhizobia were investigated, by comparing the expression of chaperone genes dnaKJ and groESL in thermotolerant and thermosensitive isolates. Tolerance to cold, heat and heat shock was evaluated for 53 mesorhizobia obtained from several provinces of Portugal and assigned to different species. Associations between isolates' tolerance phenotype and several provinces of origin were found. Some species groups were found to differ significantly in their ability to tolerate temperature stress. Analysis of the dnaK and groESL expression by Northern hybridization, using isolates from three species groups, showed an increase in the transcripts levels with heat, but not with cold stress. Interestingly, a higher induction of chaperone genes was detected in heat-tolerant isolates when compared with that of sensitive isolates of the same species. To our knowledge, this is the first analysis of chaperone genes' expression comparing tolerant and sensitive strains. The present study suggests a relationship between higher transcriptional induction of major chaperone genes and higher tolerance to heat in rhizobia.


The demand for a more effective utilization of biologically fixed nitrogen in agricultural systems has prompted studies on rhizobia diversity and tolerance to biotic and abiotic factors, such as pH, salinity and temperature (Zahran, 1999). Temperature variations, especially high temperatures, influence symbiosis in the early host–rhizobia molecular signals' exchange and also when the nodules are already established (Hungria & Vargas, 2000).

The optimum temperature range for most rhizobia is 25–31 °C (Somasegaran & Hoben, 1994) and the upper temperature limits are between 32 and 47 °C (Munevar & Wollum, 1981a; Karanja & Wood, 1988). However, these limits vary among species and strains. Rhizobia nodulating common bean are able to grow at temperatures up to 44 °C (Diouf et al., 2000), while the maximum growth temperature for chickpea rhizobia is 40 °C for both Mesorhizobium ciceri and Mesorhizobium mediterraneum (Nour et al., 1994, 1995). Besides these two species reported as specific chickpea microsymbionts, several studies have indicated that other Mesorhizobium species are able to nodulate this legume effectively, namely isolates belonging to the Mesorhizobium huakuii/Mesorhizobium amorphae species group and to the Mesorhizobium tianshanense species group (Laranjo et al., 2004; Rivas et al., 2007; Laranjo et al., 2008; Alexandre et al., 2009). However, the temperature stress tolerance of these chickpea rhizobia remains mostly unstudied.

Cellular response to heat shock is probably the most-studied stress response. Recent microarray data in Sinorhizobium meliloti reported the upregulation of 169 genes after heat shock, including genes coding for chaperones and other heat shock proteins (Sauviac et al., 2007). Chaperone systems, such as DnaK–DnaJ and GroEL–GroES, are mostly known as important components of the heat shock response, but are also induced by other stressful conditions, although constitutively expressed. These chaperones recognize exposed hydrophobic domains of target protein that would be hidden in normal protein conformation and help denatured proteins to reach its native conformation (Hartl & Hayer-Hartl, 2009).

Although the heat shock response is not studied extensively in rhizobia, several studies have addressed the multiple groEL copies typically found in these nitrogen-fixing bacteria. Bradyrhizobium japonicum shows a total of five groESL operons, which have different regulation systems and are differentially induced (Fischer et al., 1993). For example, groESL1,4,5 are heat inducible while groESL3 is induced by low-oxygen conditions (Fischer et al., 1993; Babst et al., 1996). The groESL1 is σ32 dependent, whereas the CIRCE element controls the groESL4,5 regulation (Babst et al., 1996). In Rhizobium leguminosarum, only one of the three groEL homologues is needed for normal growth and corresponds to the highly expressed one (Rodriguez-Quinones et al., 2005).

In rhizobia, as in most bacteria, dnaK is mostly a single copy gene, contrary to its co-chaperone dnaJ that, in addition to the copy found adjacent to dnaK, is often found elsewhere in the genome (Alexandre et al., 2008). In B. japonicum, dnaK seems to have an essential function, given that no dnaK knockout mutant could be obtained, while dnaJ mutants were successfully isolated, but showed a reduced growth rate, especially at high temperatures (Minder et al., 1997). groESL and dnaJ are also involved in the symbiotic performance, as already reported for several rhizobia species (Fischer et al., 1993; Labidi et al., 2000). A previous study with chickpea rhizobia indicated a consistent overexpression of a 60-kDa protein after heat shock, which might correspond to GroEL (Rodrigues et al., 2006).

The expression of groESL genes from a psychrophilic bacterium was able to increase Escherichia coli tolerance to low temperatures and decrease the lower limit temperature for growth (Ferrer et al., 2003). Another study in E. coli reported that overexpression of the native groESL system enhanced thermotolerance (Kim et al., 2009). The key roles of DnaKJ and GroESL machineries in determining temperature stress tolerance prompt us to investigate the expression of these chaperone genes, comparing thermotolerant with thermosensitive rhizobia.

The present study evaluated the temperature stress tolerance of chickpea rhizobia isolates from the entire Portuguese territory. Our aims were to identify isolates highly tolerant to heat and cold conditions and to investigate a possible relationship between stress tolerance and species or origin site of the isolates. Another aim was to investigate whether distinct temperature tolerance phenotypes were related to different induction levels of major chaperone genes. The expression of dnaK and groESL genes upon stress in thermotolerant and thermosensitive isolates belonging to different Mesorhizobium species was compared.


Evaluation of temperature stress tolerance

Bacterial strains

A total of 53 chickpea rhizobia isolates characterized previously, namely in terms of the 16S rRNA gene sequence, plasmid number and symbiotic effectiveness (Alexandre et al., 2009), were used (Table 1). These isolates were obtained from 18 soil samples, covering 10 Portuguese provinces and Madeira Island. The two chickpea Mesorhizobium type strains were also used: M. ciceri (UPM-Ca7T) and M. mediterraneum (UPM-Ca36T). All strains were routinely grown in yeast extract mannitol (YEM) broth (Vincent, 1970) and preserved in glycerol at −80 °C.

Table 1.   Mesohizobia isolates used in the present study
Province of originIsolate16S rRNA gene cluster*
  • *

    From Alexandre et al. (2009).

  • A, Mesorhizobium huakuii/Mesorhizobium amorphae species cluster; B, Mesorhizobium ciceri/Mesorhizobium loti species cluster; C, Mesorhizobium tiashanense species cluster; D, Mesorhizobium mediterraneum/Mesorhizobium temperatum species cluster.

Trás-os-Montes e Alto DouroBR-8-BragançaB
Douro LitoralPII-1-PortoB
Beira LitoralC-1-CoimbraA
Beira AltaV-5b-ViseuA
Beira BaixaCB-10-Castelo BrancoB
CB-19-Castelo BrancoB
CB-30-Castelo BrancoB
EstremaduraCR-3-Caldas da RainhaC
CR-16-Caldas da RainhaC
CR-18-Caldas da RainhaC
CR-32-Caldas da RainhaC
Alto Alentejo78-ElvasA
Baixo Alentejo6b-BejaD
MadeiraSA-9-Serra d'ÁguaA
SA-12-Serra d'ÁguaA
SA-13-Serra d'ÁguaA
SA-17-Serra d'ÁguaA

Temperature stress tolerance

The tolerance of chickpea rhizobia isolates to temperature stress was evaluated under three different conditions: heat stress (37 °C), cold stress (15 °C) and heat shock stress (48 °C for 15 min). Continuous growth at 28 °C was used as a control condition. Stress tolerance was evaluated in YEM broth, by OD540 nm readings. After overnight growth, bacterial cultures were standardized to an initial OD of 0.03 and three replicas were used for each treatment. Final OD readings were performed at the end of 48 h under each temperature condition. In heat shock stress, cultures were subjected to 48 °C for 15 min and then transferred to normal growth temperature (28 °C) for the following 48 h.

Data analysis

In order to allow comparisons between isolates, OD readings from each stress were converted into percentage values, considering growth at 28 °C as 100% growth. The average value of the three replicas and SD were calculated.

Statistical analysis was performed using spss 15.0 software (SPSS Inc., Chicago, IL). In the case of no homogeneity of variances, the Kruskal–Wallis test, as well as the Welch test, was used instead of the anova, in order to investigate a possible relationship between stress tolerance and species group of the isolates. Different post hoc tests (Tamhane, Dunnett T3 and Games-Howell) were conducted to search for categories that differ significantly from others.

Correspondence analysis (CA) was used as an exploratory data analysis technique to detect a structure in the relationships between categorical variables (Benzécri, 1973). Isolates were divided into three classes of tolerance to stress (highly tolerant, tolerant and poorly tolerant) and the relationships between these categories and provinces of origin of the isolates were investigated.

Correlations were examined between the temperature of the origin sites and the temperature stress tolerance of the isolates, using Spearman's correlation coefficient. The available temperature data (soil at 10 cm of depth and air temperatures) restricted the analysis to a subset of 28 isolates from seven provinces (Trás-os-Montes e Alto Douro, Beira Litoral, Beira Baixa, Alto Alentejo, Baixo Alentejo, Estremadura and Algarve). Temperature data were retrieved from the Portuguese Meteorological Institute (2006–2008) and the average values were calculated for summer and winter, for maximum and minimum soil and air temperatures.

Transcriptional analysis of the dnaKJ and groEL genes

RNA extraction and analysis

A total of 15 isolates were selected for transcriptional analysis of the major chaperone genes, based on their temperature stress phenotype. Total RNA was extracted from cells at control temperature (28 °C) and from cells subjected to temperature stress, according to the protocol for Rapid Isolation of RNA from Gram-negative Bacteria (Ausubel et al., 1997). A heat shock stress was performed at 48 °C for 15 min; a heat stress was performed at a lower temperature and for a longer duration of time, namely at 37 °C for 30 min, and a cold stress was performed at 15 °C for 30 min.

Total RNA was separated on a 1.5% agarose gel, in 1 × MOPS (20 mM MOPS buffer; 5 mM sodium acetate; 2 mM EDTA; pH 7.0) with a denaturing loading buffer (50% deionized formamide; 6.1% formaldehyde; 1 × MOPS). A DIG-labelled RNA molecular weight marker (Roche Applied Science) was used in order to confirm the transcripts' size. Electrophoresis was carried out at 40 V for 5 h. Total RNA was transferred overnight using the capillary transfer method into positively charged nylon membranes (Roche Applied Science) using 20 × SSC (3 M NaCl; 300 mM sodium citrate; pH 7.0). RNA was fixed to the membrane by baking at 120 °C for 30 min.

Probes synthesis

A groEL gene fragment of approximately 770 bp was amplified using the primers groEL-F (5′-GGGCCGCAACGTCGTCATCGACAA-3′) and groEL-R (5′-CTTCCAGCATGGCCTTGCGGCGAT-3′). The PCR reaction was prepared with 2 U Taq DNA polymerase (Fermentas), 1 × reaction buffer, 0.75 mM MgCl2, 0.2 mM of each dNTP (Invitrogen), 15 pmol of each primer (Stabvida) and DNA of M. mediterraneum Ca36T as a template. The amplification programme consisted of an initial denaturation step of 3 min at 95 °C, followed by 30 cycles of 2 min at 95 °C, 1 min at 68 °C and 30 s at 72 °C and a final extension step at 72 °C for 10 min. The amplified fragment was cloned using the pGEM-T Easy Vector System (Promega) following the manufacturer's instructions. The dnaKJ fragment of approximately 1600 bp was amplified as described before (Alexandre et al., 2008) and cloned using the same system mentioned above. The groEL and dnaKJ RNA probes were obtained by in vitro transcription labelling of the cloned fragment, using the DIG Northern Starter Kit (Roche Applied Science). The 16S rRNA DNA probe was synthesized using the DIG High Prime DNA Labelling and Detection Starter Kit II (Roche Applied Science) from a purified 16S rRNA gene PCR product of M. mediterraneum Ca36T, obtained as described previously (Laranjo et al., 2004).

Northern hybridization

Membranes were prehybridized in DIG Easy Hyb hybridization buffer (Roche Applied Science) for 30 min. All hybridizations were performed overnight at 68 °C. Stringency washes and immunological detection were performed according to the manufacturer's instructions. Membranes were exposed to X-ray films for 30 min.

The membranes were rehybridized with a DNA probe for the 16S rRNA gene, using the RNA–DNA hybridization temperature of 50 °C.

Data analysis

Hybridization signals were analysed using imagequant TL v7.01 (GE Healthcare). The 16S rRNA gene hybridization was used as an internal control of the amount of total RNA loaded in each sample. The ratio between the chaperone mRNA signal and the 16S rRNA gene signal was calculated and the ratio between control and temperature stress conditions yielded the number of folds by which the chaperone mRNA levels were induced after stress.

Statistical analysis was performed using spss 15.0 software (SPSS Inc.). The χ2 statistics were used to test the hypothesis that tolerant isolates show higher induction levels after a temperature upshift than sensitive isolates of the same species. In addition, the t-test was used to compare the induction levels between tolerant and sensitive isolates.


The present survey revealed that chickpea rhizobia isolates are diverse in terms of temperature stress tolerance (Fig. 1). Under 15 °C, isolates' growth ranged from 8% to 68%, while growth at 37 °C ranged from 5% to 46%. After heat shock, isolates grew between 3% and 89%. Most isolates tolerate heat shock or cold stress better than the heat stress. The two chickpea Mesorhizobium type strains (M. ciceri and M. mediterraneum) are very similar regarding their tolerance to temperature stress and were found to be highly tolerant to heat shock (>70% of growth) and tolerant to 15 °C.

Figure 1.

 Growth of chickpea rhizobia under different temperature stresses: cold (15°C), heat (37°C) and heat shock (48°C, 15 min). Percentages were calculated considering the control condition (28°C) as 100% growth. Presented values are the average of three replicas (bars show SD).

Statistical analyses indicated that there are significant differences between species groups, regarding their growth under each of the three temperature stresses (χ2=67.345, d.f.=3, P<0.001 for cold; χ2=20.963, d.f.=3, P<0.001 for heat and χ2=11.445, d.f.=3, P<0.05 for heat shock). Different post hoc tests provided consistent results in identifying the species clusters that differed significantly. For example, regarding the tolerance to 15 °C, isolates from the species cluster M. ciceri/M. loti and isolates from the species cluster M. mediterraneum/M. temperatum are significantly different from each other and from the remaining species clusters (P<0.05).

The CA revealed an association between some provinces of origin and isolates' growth under each stress. In general, isolates from the Trás-os-Montes e Alto Douro province are associated with high tolerance to cold, heat (data not shown) and heat shock (Fig. 2). Additionally, two other provinces are associated with high tolerance to heat shock, namely Algarve and Beira Litoral. On the other hand, isolates from Douro Litoral and Madeira are associated with high sensitivity to heat shock.

Figure 2.

 CA biplot of the relationship between provinces of origin and tolerance to heat shock of the isolates. (Note: the dots corresponding to provinces are occasionally overlaid.)

In order to investigate the expression of major chaperone genes in isolates with different susceptibility to temperature stress, the transcriptional levels of the dnaKJ and groESL genes were analysed by Northern hybridization. A total of 15 chickpea rhizobia isolates, comprising three species groups, were used. Isolates were selected based on the results of the temperature stress tolerance trials (Fig. 1), i.e. by their tolerant or sensitive phenotype to a given stress condition.

The dnaKdnaJ RNA probe enables the detection of three putative transcripts: the dnaK mRNA; the dnaJ mRNA; and the bicistronic dnaKJ mRNA. The most abundant transcript detected is approximately 2 kb (Fig. 3a), which is consistent with the dnaK gene size in M. loti MAFF303099 (1917 bp). The groEL RNA probe enables the detection of two possible transcripts: the groEL mRNA and the bicistronic groESL mRNA. Our results showed that the most abundant transcript is approximately 2 kb (Fig. 3b), which is consistent with the groESL bicistronic mRNA, because the groEL gene of M. loti MAFF 303099 is 1650 bp and the groES gene is 300 bp approximately.

Figure 3.

 Comparison of the transcription analysis of chaperones genes between a tolerant isolate and a sensitive isolate subjected to (a) heat and (b) cold stress. Northern blot hybridization of total RNA with probes specific for dnaKJ, groEL and 16S rRNA gene under (C) control conditions and (S) after stress.

As expected, after a temperature upshift to 37 or 48 °C, the majority of the isolates showed an increase in the chaperone transcript levels. Interestingly, after heat shock, the increase in both dnaK and groESL transcripts levels was higher in tolerant than that in sensitive isolates within the same species cluster (Fig. 4a). The only exceptions were C-1-Coimbra (tolerant) and 78-Elvas (sensitive), with low and high induction levels of dnaK, respectively. The highest induction of dnaK was detected in the tolerant isolate BR-15-Bragança (over 15-fold). Sensitive isolate 29-Beja showed repression of the dnaK gene after the temperature upshift (0.5-fold). The highest fold induction of groESL was detected in the tolerant isolate C-14-Coimbra (approximately 7.5-fold), while in sensitive isolates, almost no induction of groESL was detected following heat shock.

Figure 4.

 Induction of dnaK and groESL genes upon (a) heat shock and (b) heat stress evaluated by Northern blot analysis of total RNA from tolerant and sensitive rhizobia isolates.

Similar to the results obtained for the heat shock condition, after heat stress, a higher induction of both dnaK and groESL was detected in tolerant isolates when compared with sensitive isolates within the same species cluster (Fig. 4b). The only exception was the sensitive isolate 7a-Beja that showed high induction of groESL, contrary to the other sensitive isolate. C-14-Coimbra (tolerant) showed the highest fold induction of dnaK after stress (approximately 20-fold), while the highest groESL induction was detected in the tolerant isolate BR-15-Bragança (approximately 29-fold).

The statistical analysis confirmed the hypothesis that, after a temperature upshift, the induction levels of the major chaperone genes are higher in tolerant than in sensitive isolates of the same species (χ2=6545, d.f.=1, P<0.05). Furthermore, the induction levels of tolerant isolates are significantly different from those of sensitive isolates (t=2.189, d.f.=19, P<0.05).

The dnaK and groESL mRNAs levels following the cold stress were similar to the ones detected under normal growth conditions, regardless of the isolate phenotype or the species group, in the six isolates analysed (27-Beja, BR-9-Bragança and C-3-Coimbra, as tolerant isolates; 78-Elvas, 85-Elvas and STR-4-Santarém as sensitive isolates) (data not shown).


The present study showed that native chickpea rhizobia display a large diversity in their tolerance to temperature stress. In general, isolates grew better at low temperature than at high temperature, although most isolates endured the heat shock well. Isolates showed low percentages of growth at 37 °C, which is in agreement with previous studies in chickpea rhizobia (Maâtallah et al., 2002; Rodrigues et al., 2006). A maximum growth temperature of 40 °C was reported for both M. ciceri and M. mediterraneum type strains (Nour et al., 1994, 1995). A high tolerance to heat shock is common in rhizobia as reported previously by other studies in chickpea rhizobia and other species (Kulkarni & Nautiyal, 2000; Rodrigues et al., 2006).

In the present study, some species groups were found to differ significantly regarding their ability to grow under temperature stress. Isolates from the species groups that include the specific chickpea microsymbionts, M. ciceri/M. loti group and M. mediterraneum/M. temperatum group show the highest growth average at 15 °C and are significantly different from the other species. Regarding heat stresses, isolates from the two specific chickpea microsymbionts species groups display the two highest growth averages as well. Several studies in rhizobia have reported that tolerance seems to be species related, as for example pH tolerance (Reeve et al., 2006; Brígido et al., 2007), copper tolerance (Laguerre et al., 2006) and antibiotic resistance (Alexandre et al., 2006).

Several studies have reported that rhizobia tolerant to heat stress were mainly obtained from geographical origins with high temperatures (Zahran, 1999; Rahmani et al., 2009). However, in the present study, no positive correlation was found between rhizobia growth at 37 °C and maximum soil or air temperatures of the origin sites (data not shown), which is in accordance with other reports on chickpea rhizobia (Maâtallah et al., 2002; Rodrigues et al., 2006). Although isolates' tolerance is apparently unrelated to the temperatures of the origin site, associations were found between some provinces of origin and tolerance to temperature stress. The CA biplots clearly associate isolates from Trás-os-Montes e Alto Douro with high tolerance to all tested temperature stresses. Interestingly, this province in the north of Portugal presents the lowest air and soil temperatures during the winter and high temperatures during the summer (data not shown).

The transcriptional analysis of 15 rhizobia isolates points towards a general increase of both dnaK and groESL mRNA levels after temperature upshifts and an insignificant alteration of those levels after cold stress. Interestingly, after heat shock and heat stress, a higher increase in the amount of those chaperones transcripts was generally observed in tolerant isolates, in comparison with sensitive isolates, within the same species group. These results were obtained using isolates from three species groups and suggest that increased levels of chaperones may contribute to a higher tolerance to heat stress in rhizobia. The increase in heat tolerance caused by a higher expression of groESL genes was already reported in E. coli (Kim et al., 2009). Higher amounts of chaperones not only prevent protein denaturation more efficiently, by acting as holdases, but also allow proper protein folding, by acting as foldases. In addition, it has been shown that chaperones can increase the mRNAs' stability (Yoon et al., 2008). For these reasons, a higher expression of chaperones genes in rhizobia could contribute to higher tolerance to heat. Knockout and overexpression studies may be useful approaches to further investigation of the role of major chaperone genes in rhizobia thermotolerance. On the other hand, genome-wide techniques, such as microarray analysis for expression profiling, would provide a more complete view of the heat-inducible genes in Mesorhizobium.

The detection of the dnaK transcript alone in the studied mesorhizobia could be explained by the presence of a terminator between the dnaK and the dnaJ genes. A terminator might cause at least partial transcription termination and, consequently, low levels of dnaJ transcript. Analysis of the intergenic region of the dnaKJ operon from the completely sequenced genome of M. loti MAFF303099 shows a putative intrinsic terminator sequence, which corroborates the former hypothesis. Similar to our results, Northern hybridization experiments in Agrobacterium tumefaciens revealed no transcript corresponding to dnaJ alone, after the heat shock induction (Segal & Ron, 1995). In B. japonicum, no transcription start site was detected in the intergenic region and two putative stem-loop structures were suggested to be responsible for the transcription of dnaK alone (Minder et al., 1997).

Taking into account several studies showing that strains' endurance to high temperatures in a culture medium correlates with their symbiotic performance under heat stress (Munevar & Wollum, 1981b; Rahmani et al., 2009), the thermotolerant chickpea rhizobia identified in the present work could have potential agronomical importance. For example, isolate PMI-1-Portimão has a high symbiotic effectiveness (>75%) and is among the isolates with higher tolerance to all the tested conditions; hence, it represents a good candidate for field inoculation.

To our knowledge, this is the first study in mesorhizobia focusing on the transcriptional analysis of chaperone genes under different temperature stresses. The present results suggest the existence of a relationship between higher levels of transcriptional induction of the dnaK and groESL chaperone genes and a higher ability of chickpea rhizobia isolates to endure heat stress. Further studies are required to clarify the molecular mechanisms of temperature stress tolerance in rhizobia.


This study was supported by Programme POCTI (POCTI/BME/44140/2002) and Programme POCI 2010 (PTDC/BIO/80932/2006) financed by FCT (Fundação para a Ciência e a Tecnologia) and cofinanced by EU-FEDER. A.A. acknowledges the PhD fellowship from FCT (SFRH/BD/18162/2004). The authors thank G. Mariano for technical assistance.