DNA Research Advance Access published November 25, 2013 DNA RESEARCH pp. 1–12, (2013) doi:10.1093/dnares/dst050 Global Transcriptional Response to Heat Shock Legume Symbiont Mesorhizobium loti MAFF303099 Comprises Extensive Gene Downregulation ANA Alexandre1,2, MARTA Laranjo1,2, and SOLANGE Oliveira1,* ICAAM – Instituto Ciencias Agrarias e Ambientais Mediterranicas (Laboratorio Microbiologia do Solo), ˆ ˆ ´ ´ ´ ´ ¸˜ Universidade Evora, Nucleo da Mitra, Ap. 94, 7002-554 Evora, Portugal1 and IIFA – Instituto Investigac ao e ´ ´ ´ Formacao Avanc ada, Universidade Evora, Ap. 94, 7002-554 Evora, Portugal2 ¸˜ ¸ *To whom correspondence should be addressed. Tel. þ351-266760878. Email: ismo@uevora.pt (Received 19 August 2013; accepted 24 October 2013) Abstract Rhizobia, bacterial legume symbionts able to fix atmospheric nitrogen inside root nodules, have to survive in varied environmental conditions. The aim this study was to analyse transcriptional response to heat shock Mesorhizobium loti MAFF303099, rhizobium with large multipartite genome 7.6 Mb that nodulates model legume Lotus japonicus. Using microarray analysis, extensive transcriptomic changes were detected in response to heat shock: 30% protein-coding genes were differentially expressed (2067 genes in chromosome, 62 in pMLa and 57 in pMLb). The highest-induced genes are in same operon and code for two sHSP. Only one five groEL genes in MAFF303099 genome was induced by heat shock. Unlike other prokaryotes, transcriptional response this Mesorhizobium included underexpression an unusually large number genes (72% differentially expressed genes). This extensive downregulation gene expression may be an important part heat shock response, as way reducing energetic costs under stress. To our knowledge, this study reports heat shock response largest prokaryote genome analysed so far, representing an important contribution to understand response plant-interacting bacteria to challenging environmental conditions. Key words: stress; rhizobia; microarrays; chaperone; sHSP 1. Introduction Rhizobia are soil bacteria able to colonize legume roots and form nodules, where atmospheric nitrogen is metabolized into compounds that can be used by plant. The impact biological nitrogen fixation carried out by rhizobia in agriculture is both economic and environmental. Rhizobia may reduce use chemical N-fertilizers, which represent production cost reduction and at same time decrease in pollution resulting from N-fertilizers synthesis and from soil nitrate lixiviation.1 Rhizobia typically have large genomes, which are often composed by several replicons. These seem to be common features bacterial species that interact with host.2 This rhizobial trend to harbour large accessory genome is probably related, not only to symbiosis itself (interacting with host), but also to plasticity required to survive in complex and distinct environments. As free-living bacteria, rhizobia have to cope with changes in soil conditions and as plantsymbionts, rhizobia must overcome plant defence mechanisms and adapt to intracellular nodule environment. For all above reasons, these bacteria are particularly interesting to study stress response. The most important consequences heat stress at cellular level are protein denaturation and aggregation.3 These effects are common to other adverse # The Author 2013. Published by Oxford University Press on behalf Kazusa DNA Research Institute. This is an Open Access article distributed under terms Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided original work is properly cited. Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 Edited by Dr Naotake Ogasawara Page 2 12 Heat Shock Response Mesorhizobium loti MAFF303099 large number sHSP is common feature in rhizobia genomes.25 Some sHSP have specific regulation designated by repression heat shock gene expression (ROSE). ROSE element is posttranscriptional regulation mechanism that consists in conserved sequence downstream to promoter.26 The heat shock response has been extensively studied in bacteria, however to our knowledge, only one rhizobia strain was studied in terms heat shock transcriptome, namely S. meliloti 1021, symbiont Medicago spp.5,27 The strain analysed in present report, Mesorhizobium loti MAFF303099, is rhizobium able to establish nitrogen-fixing symbiosis with Lotus species.28,29 M. loti MAFF303099 genome comprises large chromosome (7 Mb) and two plasmids designated as pMLa (352 kb) and pMLb (208 kb). chromosomal symbiosis island (610 kb) contains most genes involved in nodulation and nitrogen fixation. previous study showed that this strain is tolerant to heat shock and cold conditions, and grows well at pH 5.30 The aim present study is to characterize transcriptional response to heat shock in resourceful rhizobium with large and complex genome. The analysis global transcriptional alterations following sudden exposure to high-temperature conditions in M. loti MAFF303099 will contribute to better understanding general stress response, in particular in symbiotic bacteria with multiple replicons and large accessory genome. 2. Materials and methods 2.1. RNA purification Overnight cultures M. loti MAFF303099 were grown in YMB31 at 288C, to final optical density 0.3 (540 nm). volume 10 ml bacterial culture was used in each treatment: 30 min at control (288C) and heat shock (488C) conditions. Cells were harvested and total RNA was purified using RNeasy Mini Kit (Qiagen). Contamination with DNA was removed by DNase digestion (Roche), followed by RNA cleanup using RNeasy Mini kit (Qiagen). Total RNA integrity was checked using RNA Nano kit and an Agilent 2100 Bioanalyser (Agilent Technologies), while RNA quantification was performed using NanoDrop ND1000 (NanoDrop Technologies). RNA was prepared from three independent cell cultures. 2.2. Microarray experiments RNA processing as well as microarrays hybridization and raw data extraction were service provided by Biocant Park—Genomics Unit (Portugal). In order to enrich RNA samples in mRNA, MICROB ExpressTM Kit (Ambion) was used to remove most rRNA. mRNA was then amplified with Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 conditions, as for example oxidative stress, so study heat stress response is also relevant in understanding tolerance to other stresses. The plasticity to respond to stressful conditions involves rapid changes in gene expression. Alternative sigma factors allow bacteria to rapidly redirect RNA polymerases pool to set genes that are required to respond to certain condition.4 Rhizobia genomes typically harbour large number alternative sigma factors, including multiple copies rpoH, which encodes s32, major sigma factor involved in heat shock response.5 s32 might be involved in response to other stresses as seen in Rhizobium etli, where rpoH2 seems to be more related to oxidative stress response.6 Furthermore, rhizobia with rpoH deletions may also be affected in their symbiotic pheno21% genes type.6,7 The transcription induced in response to temperature upshift are rpoH1 dependent in Sinorhizobium meliloti and these include chaperones, proteases and small heat shock proteins (sHSP).5 Important chaperone systems, such as GroES-GroEL and DnaK-DnaJ-GrpE, are s32-regulated in most alphaproteobacteria. Chaperones play key role in heat shock response, as they are involved in promoting acquisition native conformation by proteins that suffered denaturation and present wrong folding.8 The importance chaperonins in defining tolerance to temperature has been highlighted by several studies in E. coli.9,10 more recent study showed that high level GroESL system has fundamental role in evolution heat tolerance.11 Some important reports on functional analysis multiple groESL operons in rhizobia have been published.12 – 14 Mutational studies showed that groESL operons within same genome are induced by different stimuli and that these genes are involved not only in stress tolerance, but also in nodulation and nitrogen fixation processes.15 In Mesorhizobium spp., both dnaK and groESL genes were reported to be transcriptionally induced by temperature upshift, especially in heat tolerant isolates.16 In rhizobia, groESL operons are often CIRCE (controlling inverted repeat chaperone expression) regulated, as already reported in Bradyrhizobium japonicum, S. meliloti and Rhizobium leguminosarum.13,17,18 CIRCE is highly conserved DNA sequence that serves as binding site repressor protein HrcA.19,20 Similar to GroESL chaperonins, also DnaKJ system seems to be involved in both heat tolerance and symbiosis phenotype.21 – 23 Regarding cochaperone dnaJ, rhizobia mutants showed that both stress tolerance and symbiotic performance are affected.21,22,24 sHSP are mostly involved in preventing irreversible aggregation misfolded proteins. The presence A. Alexandre et al. MessageAmpTM II-Bacteria Kit (Ambion), with incorporation 5-(3-aminoallyl)-UTP (Ambion) for indirect labelling, which was carried out by coupling fluorescent Cy3 to amplified RNA (aRNA), following instructions Amino Allyl MessageAmpTM II aRNA Amplification Kit (Ambion). The 40K array for M. loti MAFF303099 (MYcroarray) includes probes for 7231 genes ( 99% total number protein-coding genes) with five replicates for each probe. Slide hybridization was carried out as described by microarray’s supplier, using Gene Expression Hybridization Kit (Agilent Technologies). Data were acquired using DNA Microarray B Scanner (Agilent Technologies), with an intensity 100% PTM in green channel. 2.4. Microarray data validation Validation microarray data was performed by realtime quantitative RT–PCR (qRT–PCR). cDNA was obtained by reverse transcription using Maxima First Strand cDNA Synthesis kit (Thermo Scientific) according to manufacturer’s instructions. Primers (Supplementary Table S1) were designed using Primer Express 3.0 software (Applied Biosystems). Real-time qRT– PCR reactions were prepared using 0.1 ng/ml template cDNA, SYBR Green PCR Master Mix and 0.3 mM each primer. Amplifications were carried out in 7500 Real-time PCR System (Applied Biosystems). Ct values for target genes were normalized using reference genes hisC, rpoA and sigA, which showed no variation in corresponding transcript levels for experimental conditions used (data not shown). 3. Results and discussion 3.1. Global transcriptional response Analysis M. loti MAFF303099 transcriptome allowed identification 2186 protein-coding genes that were differentially expressed after heat shock (out 7231 genes analysed), with an average false discovery rate 1.5% (accession number GSE43529). This indicates that transcript levels 30% protein-coding genes were altered by this stress. The transcriptional response included much higher number downregulated (1584) compared with upregulated (602) genes (Fig. 1). The unexpected larger proportion downregulated genes does not seem to be feature rhizobia, taking into account similar numbers induced and repressed genes reported for S. meliloti.5,27 To our knowledge, present study reports largest prokaryote genome studied so far in terms response to heat shock. To investigate influence genome size in global heat response, comparison transcriptional response to heat prokaryotes with different genome sizes was performed (Fig. 2). Strain MAFF303099 shows an unusual proportion downregulated genes in response to heat shock compared with several other bacteria and archaea that, in general, show similar number genes under- and overexpressed following temperature upshift (though different heat shock conditions are compared). Despite fact that diverse species with distinct lifestyles and subjected to different heat shock conditions are compared in Fig. 2, analysis transcriptomic data suggests general trend pronounced increase in number downregulated genes with genome size. One might speculate that many expendable genes are shutdown, so that cellular machinery can be more effective in synthesis specific functional response. Nevertheless, extensive gene downregulation is not particularly detected in accessory genome that is presumably more dispensable. Indeed, symbiosis island shows dispersed underand overexpressed genes similar to rest chromosome (Fig. 3). Furthermore, some highly induced genes are plasmid encoded, mainly in pMLb (Fig. 4). This is somewhat unexpected since symbiosis islands and plasmids are mobile elements in genome, known to be laterally transferred within soil populations and thus less expected to carry genes Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 2.3. Data analysis The microarrays data were analysed using BRB ArrayTools (version 4.2).32 The arrays were normalized using array median and genes that were differentially expressed following heat shock were identified using MeV software.33 Genes were considered differentially expressed for P  0.01 in t-test. Despite recent update on annotation MAFF303099 genome released by NCBI (October 2012), all genes differentially expressed that were annotated as ‘hypothetical protein’ were further analysed using Blast2GO software.34 This analysis included Blast, Mapping and Annotation, and allowed further annotation many genes. In order to assign highest number possible genes to clusters orthologous genes (COG) category, STRING 9.0 database (search tool for retrieval interacting genes)35 was used. MicrobesOnline Operon Predictions (www.microbesonline.org/operons/) was used for operon prediction.36 The identification putative promoter sequences was performed using BPROM-Prediction bacterial promoters software (www.softberry.com). DNAPlotter37 was used to generate circular DNA maps showing transcriptomics data. Spearman’s coefficient was used to test for correlation between genome size and number over- or underexpressed genes (IBM SPSS Statistics, version 21). Page 3 12 Page 4 12 Heat Shock Response Mesorhizobium loti MAFF303099 Figure 2. Number overexpressed (þ) and underexpressed (O) genes resulting from transcriptome studies response to heat shock 18 species Bacteria and Archaea plotted against their genome size. Trendlines are shown in grey for number overexpressed genes (R 2 ¼ 0.35; Spearman’s r ¼ 0.583, P  0.05) and in black for number underexpressed genes (R 2 ¼ 0.69; Spearman’s r ¼ 0.608, P  0.01). From smallest to largest genome size: Mycoplasma hyopneumoniae38; Tropheryma whip-plei39; Rickettsia prowazekii40; Campylobacter jejuni 41; Streptococcus thermophilus42; Achaeoglobus fulgidus 43; Bifidobacterium longum44; Xylella fastidiosa 45; Listeria monocytogenes46; Acidithiobacillus ferrooxidans 47; Corynebacterium glutamicum 48; Desulfovibrio vulgaris 49; Clostridium difficile 50; Escherichia coli 51; Methanosarcina barkeri 52; Shewanella oneidensis 53; S. meliloti5; M. loti (this study). The two rhizobia species are denoted in graphic. Note: in case multiple heat shock transcriptome datasets for same species, dataset with largest number differentially expressed genes was chosen. essential for stress survival. In addition, set 100 genes with highest M-values comprises 14 plasmid encoded genes, while 100 highly underexpressed genes are all chromosomal (Supplementary Table S2). The high number underexpressed genes may suggest that heat shock response relies on lowenergy transcriptional response. Accordingly, 40% induced genes show low increase in transcriptional levels (M , 1). This low level gene induction, commonly disregarded, may be important part cells response, as pointed before by Wren and Conway.54 Analysis location differentially expressed genes in each replicon shows an apparently random distribution over- and underexpressed genes, with exception an 200 kb-long region located in 1 000 000 – 1 200 000 (462 genes) where all differentially expressed genes are downregulated (Fig. 3). Both in chromosome and plasmids, distribution differentially expressed genes seems to be unrelated to DNA strand or GC content. Real-time qRT– PCR was used to validate microarray data. Genes were chosen based on M-values from microarrays results, in order to include overexpressed, underexpressed and not differentially expressed genes, as well as genes encoded in both DNA strands and scattered in chromosome. In general, results from real-time qRT– PCR experiments are in agreement with microarrays analysis results (Table 1), with exception dnaK gene (discussed in section ‘The DnaKJ chaperone system’). Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 Figure 1. Microarrays analysis M. loti MAFF303099 subjected to heat shock. M-values for differentially expressed genes (P , 0.01) obtained from comparison between heat shock (488C) and control (288C) conditions. Genes with increased amount mRNA following heat shock have positive M-values (overexpressed), while genes with decreased mRNA levels after heat shock show negative M-values (underexpressed). A. Alexandre et al. Page 5 12 Figure 3. Circular plots chromosome and two plasmids included in M. loti MAFF303099 genome showing, from outer to inner rings: COG group for each gene; %GC plot; and heat shock transcriptome data (M-values). The plasmids plots include two additional outer rings displaying genes encoded in plus strand (outermost ring) and minus strand. COG colours: information storage and processing—blue; cellular processes and signalling—green; metabolism—magenta; poorly characterized—yellow; more than one COG category—brown; no COG—light grey. Transcriptome data: overexpressed—black; underexpressed—grey. %GC data: above average—dark red; below average—orange. The symbiosis island (coordinates 4 644 792– 5 255 766)28 is marked in blue in chromosome plot. This figure appears in colour in online version DNA Research. Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 Protein-coding genes can be grouped into COG, according to their similarity in terms domain architecture and function.55 The present study showed that temperature stress-induced changes in expression genes belonging to all COG categories from MAFF303099 genome (Fig. 5). For all COG categories, percentage underexpressed genes is higher than that overexpressed genes (Fig. 5 and Supplementary Fig. S1). In addition to fact that high number differentially expressed genes are not in COG (1580 genes), there are also many poorly characterized genes (‘S—function unknown’ and ‘R—general function prediction only’ categories) (Supplementary Fig. S1). The COG category with highest percentage overexpressed genes is ‘L—replication, recombination and repair’ (9%). This COG category also shows lowest percentage underexpressed genes (12%). Nevertheless, percentage overexpressed genes is between 7 and 8% in nine other categories, including COG category where chaperones and other heat shock proteins are included (‘O—posttranslational modification, protein turnover and chaperones’). This suggests balanced response in terms gene induction throughout COG categories; yet, 13% overexpressed genes are not in COG. Three categories include high percentage underexpressed genes following heat shock, namely ‘D—cell cycle control, cell division, chromosome partitioning’, ‘F—nucleotide transport and metabolism’ and ‘N—cell motility’ (53, 48 and 44%, respectively). COG categories with high number overexpressed genes are ‘K—transcription’, ‘G—carbohydrate transport and metabolism’ and ‘E—amino acid transport and metabolism’ (Supplementary Fig. S1A). On other hand, COG categories E and G also show high number underexpressed genes (Supplementary Fig. S1B). This is consistent with other bacterial species for which these two COG categories also showed high number over- and underexpressed genes in response to heat shock.46,48 According to Konstantinidis and Tiedje,56 large genomes tend to have disproportional increase genes belonging to COG ‘K—transcription’, ‘T—signal transduction mechanisms’ and ‘Q—secondary metabolites biosynthesis, transport and catabolism’, which could be expected to be most underexpressed categories in large genome bacteria, nevertheless that is not observed in MAFF303099 (Fig. 5 and Supplementary Fig. S1B). Page 6 12 Heat Shock Response Mesorhizobium loti MAFF303099 Figure 4. Number and location differentially expressed genes in M. loti MAFF303099, following heat shock. Table 1. Microarrays data validation using real-time qRT– PCR Gene mll2386 – 14.6 6.6 mlr2394 groEL 11.9 5.8 mll1528 – 4.6 4.7 mll3429 clpB 7.0 2.9 mll3842 citZ 6.1 2.2 mlr5932 acdS 1.5 1.1 M-value Real-time qRT– PCR Microarrays mll3873 – 20.6 21.9 mlr0883 gcvT 20.9 22.2 mlr6118 – 22.4 22.7 mll1546 ftsZ 23.4 23.7 mll6630 – 23.8 24.0 mlr2911 flgB 23.7 24.3 mll6578 fixK 24.0 25.1 mll6432 – mll4757 mll4755 mlr7618 0.3 nde dnaK 5.5 nde dnaJ 20.2 nde greA 0.9 nde nde, not differentially expressed. 3.2. Small heat shock proteins The two most heat shock-induced genes (mll2387 and mll2386 with M-values 6.61 and 6.32, respectively) code for sHSP (Table 2). These genes are probably co-transcribed, since single putative promoter was identified upstream mll2387 (predicted promoter: -35 TTGACG and -10 ACTCATTCT). This particular sHSP operon is likely to play an important role in heat shock response, since homologous genes were also detected as most overexpressed in S. meliloti following less severe heat shock.5 Following longer heat exposure, these genes seem to be less overexpressed, yet showing an induction approximately 4-fold.27 The homologous ibpAB are also most-induced genes in heat shock response E. coli.51 Western analysis protein extracts several rhizobia species confirmed an increase amount sHPS with temperature 3.3. GroESL chaperone system Similar to other heat shock related genes, rhizobia genomes harbour several copies groESL operon, usually with different regulation mechanisms and expression kinetics.15 M. loti MAFF303099 has four groESL operons in chromosome and one in pMLa. From these five operons, only one appears to be involved in heat shock response, namely groEL gene mlr2394, which was strongly overexpressed after heat shock exposure (M ¼ 5.79). This groEL gene is highly similar to groEL5 and groEL1 from S. meliloti (87 and 83% amino acid identity, respectively), which are most heat shock-inducible copies in that species.5,27 From what is known from other rhizobia genera, only some groESL operons encoded in same genome are heat inducible and those can be regulated either by s32 or by CIRCE element.12,13,59 In case MAFF303099, CIRCE element was found upstream all groESL operons (Supplementary Fig. S2B). The same exact consensus sequence this inverted repeat is found in three operons and remaining two operons differ in only two positions. The overexpressed Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 Locus tag upshifts.25 As in many other bacteria species, in M. loti, S. meliloti and E. coli, ROSE element was identified upstream these operons26 (Supplementary Fig. S2A). Nevertheless, these sHSP genes were reported as rpoHdependent in S. meliloti,5 suggesting multiple regulation mechanisms that may allow dynamic stress response. Rhizobia genomes carry large number sHSP.25 Strain MAFF303099 has eight genes identified as sHSP, from which four were highly induced by heat shock (mll2387, mll2386, mll9627 and mll3033), two remained unaltered (mll2257 and mlr3192) and two were underexpressed (mlr4720 and mlr4721). sHSP can be divided into two classes in terms sequence: class includes sHSP similar to E. coli IbpAB, while sHSP grouped in class B are more divergent in terms sequence.25 Gene mll2387 belongs to class A, while mll2386 is more divergent and considered class B sHSP.57 According to Studer & Narberhaus58 it is improbable that mll2386 and mll2387 could form hetero-oligomers even if co-expressed, since in B. japonicum hetero-oligomers only occurred between sHSP from same class. All class sHSP from M. loti MAFF303099 (mll2387, mll3033, mlr3192 and mll9627-plasmid encoded) showed ROSE element downstream to promoter, which would confer high-temperature sensitivity to transcription these genes26 (Supplementary Fig. S2A). However, one these sHSP was not overexpressed following heat shock tested (mll3192-hspH), despite fact that its B. japonicum homolog, also regulated by ROSE element, is heat inducible.25 A. Alexandre et al. Page 7 12 groEL gene belongs to one operons regulated by slightly divergent CIRCE element. Our results suggest that presence CIRCE consensus sequence does not ensure highly efficient induction under heat stress conditions. similar situation was detected in R. leguminosarum, where putative CIRCE element was found upstream all three groESL operons, and further analysis this regulation mechanism showed that most heat-inducible operon was indeed CIRCE regulated, but second operon, less induced by heat, was not affected by CIRCE deletion or hrcA knockout.17 This second operon was rpoH regulated, suggesting an overlapping regulation mechanisms.17 Despite high M-value detected for groEL mlr2394, expression groES gene in same operon (mlr2393) following heat shock remained unaltered. Similarly, in S. meliloti, gene SMb22023 (groES5) was not induced by heat shock, despite high induction corresponding groEL5 gene (SMb21566).5,27 No promoter could be identified in 59 bp groES –groEL intergenic space using BProm, so bicistronic mRNA should be synthesized. posttranscriptional cleavage could explain why only transcript second gene in operon is highly abundant. cleavage event occurs in groESL transcript Agrobacterium tumefaciens, explaining why transcript corresponding to groEL alone is abundant mRNA detected after heat shock.60 Analysing intergenic space in MAFF303099 groES –groEL operon, using ‘KineFold Web Server’,61 stem-loop structure was found, though weaker than one described to undergo cleavage in A. tumefaciens (data not shown). GroES – GroEL complexes comprising proteins encoded by different operons tend to be less efficient than chaperonins complexes encoded by same operon.62 However, predominant GroES – GroEL complex consists single 10 kDa-heptameric ring (GroES) plus two rings seven 60 kDa-monomers (GroEL), so ratio between two is 1:2, which is consistent with lower groES transcription. 3.4. DnaKJ chaperone system The role DnaKJ chaperone system in stress response is well known in other bacteria; however, few studies address these heat shock proteins in rhizobia. In present study, dnaK (mll4757) and co-chaperone dnaJ (mll4755) were not found to be significantly heat shock induced. Nevertheless, real-time qRT– PCR results (Table 1) show that dnaK was induced by heat shock, agreeing with previous studies in Mesorhizobium.16 Approximately, 2-fold induction dnaK gene was detected in S. meliloti cells exposed to 408C for 30 min,5 while no induction was reported Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 Figure 5. Percentage genes from each COG category overexpressed and underexpressed after heat shock. Genes not in COG are also shown. The number genes included in each category is shown at right end graphic. Page 8 12 Heat Shock Response Mesorhizobium loti MAFF303099 Table 2. Overexpressed genes following heat shock, identified by microarray analysis. COG categorya Gene description M-value Chr O sHSP 6.61 mll2387 Chr O sHSP 6.32 mll3685 Chr – PRC-barrel domain-containing protein 6.08 msl2054 Chr – Hypothetical protein 5.84 mll1959 Chr – BA14k family protein 5.82 Locus tag Replicon mll2386 mlr2394 Chr O Molecular chaperone-GroEL 5.79 mll7465 Chr G ABC transporter permease 5.79 msr9689 pMLb – Hypothetical protein 5.62 msr8048 Chr – Hypothetical protein 5.61 msl2390 Chr E Usg family protein 5.54 Chr – Hypothetical protein 5.49 mlr4836 Chr HC Monooxygenase FAD-binding protein 5.27 mlr2158 Chr SR Metallo-beta-lactamase superfamily protein 5.25 mlr2234 Chr – Hypothetical protein 5.25 mll9627 pMLb O sHSP 5.23 mlr2160 Chr R Transporter component 5.17 mlr5153 Chr – Transmembrane protein 5.09 mll9357 pMLa S Domain-containing protein 5.03 mll3694 Chr T Transcriptional regulator 4.98 mll1952 Chr C Norsolorinic acid reductase 4.98 mll4827 Chr J Endoribonuclease L-PSP 4.89 mlr2159 Chr K Transcriptional regulator 4.86 msr8615 Chr R Transporter component 4.82 msl3831 Chr – Conserved hypothetical transmembrane protein 4.79 mlr9581 pMLb – PRC-barrel protein 4.78 mll4607 Chr S Ku protein 4.66 mll1528 Chr S Small integral membrane protein 4.65 mlr8230 Chr K Transcriptional regulator 4.56 mlr0408 Chr K Transmembrane anti-sigma factor 4.56 msr2497 Chr S Hypothetical protein 4.47 mll8293 Chr – Hypothetical protein 4.38 msl2212 Chr K Family transcriptional regulator 4.36 msl7604 Chr – Hypothetical protein 4.31 msl7943 Chr – Hypothetical protein 4.28 msl9358 pMLa S Transcription factor 4.27 mlr2125 Chr – Hypothetical protein 4.27 mlr3707 Chr – Hypothetical protein 4.25 mlr0407 Chr K RNA polymerase sigma factor 4.16 mll3692 Chr – Hypothetical protein 4.14 msr8675 Chr S Hypothetical protein 4.11 mlr3233 Chr N Host attachment protein 4.07 msr4317 Chr – Hypothetical protein 4.04 mll2066 Chr D Mobile mystery protein b 4.02 mll6953 Chr R Domain-containing protein 4.02 mll6858 Chr RIQ Short chain dehydrogenase 3.98 mlr1797 Chr S Conserved domain protein 3.91 Continued Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 msl1808 A. Alexandre et al. Page 9 12 Table 2. Continued COG categorya Gene description M-value Chr C Luciferase-like protein 3.91 mll2211 Chr C Morphinone reductase 3.89 msl6857 Chr R Hypothetical protein 3.88 mll8179 Chr S Family protein 3.83 Locus tag Replicon mll3445 The 50 genes with highest M-values are shown. Gene descriptions shown in bold resulted from sequence analysis using Blast2GO software. COG category letters according to NCBI functional categories (http://www.ncbi.nlm.nih.gov/COG/grace/fiew.cgi). 3.5. Sigma factors Rhizobia usually have multiple copies genes encoding same sigma factors, for example rpoH and rpoE. The M. loti MAFF303099 genome includes 25 putative sigma factors, from which four were induced by heat shock (mlr0407, mll3697, mll8140 and mlr3807). None these sigma factor-encoding genes is completely annotated; nevertheless, BLAST analysis showed that loci mlr0407 (highly induced) and mll8140 are similar to both s70 and s24, and mlr3807 is more similar to s24, while mll3697 shows high similarity to S. meliloti rpoE2 gene (76%). Sauviac and collaborators27 suggested RpoE2 as major global regulator stress response in S. meliloti, despite fact that no phenotype change was detected in rpoE2 mutant. Our results are consistent with that suggestion, since mll3697 is overexpressed in heat shock conditions with an M-value 2.4. The gene mll2869 encoding s70 was found to be underexpressed following heat shock conditions, which may contribute to extensive downregulation detected in MAFF303099 transcriptional response. Sigma factors typically related to heat shock response, as s32 (rpoH) and s24 (rpoE) that probably are encoded by mlr3741 and mlr8088 in MAFF303099, were not affected at transcriptional level by heat shock conditions applied. The gene rpoH2 (mlr3862) was also not induced in conditions ´nez-Salazar and cowused in this study. Similarly, Martı orkers6 reported that none rpoH genes were induced by heat shock in R. etli. Nevertheless, rpoH mutants are usually impaired in their stress tolerance phenotype, as is case for S. meliloti and R. etli.6,66 rpoH1 controls expression 21% heat shock-induced genes in S. meliloti and is also related to oxidative stress response, while rpoH2 seems to play minor role in heat shock response and is more involved in osmotic tolerance.5,6 In E. coli, rpoH regulation seems to be more at protein level than at transcriptional level. This control hypothesis is known as ‘unfolded protein titration model’ and involves most important chaperone systems: under normal growth conditions, s32 binds to DnaKJ and GroESL so it becomes unavailable for RNA polymerase binding; under heat stress, misfolded proteins have higher affinity for chaperone systems and s32 would be released.67 This posttranslational regulation has not been investigated in rhizobia, nevertheless fact that no rpoH induction was detected under heat stress conditions is consistent with proposed model. 3.6. Nodulation and nitrogen fixation genes Some genes involved in nodulation and nitrogen fixation were detected to be differentially expressed after heat shock. Several fix genes showed severe underexpression, especially fixK, which encodes transcriptional regulator and was most underexpressed gene following heat stress (Supplementary Table S2). Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 for shorter heat shock (428C for 15 min).27 In microarray analysis, changes in expression levels dnaK gene were considered not statistically significant due to discrepancies among replicates. It was reported for several rhizobia species that dnaJ deletions cause reduced growth at high temperatures;21,24 however, no transcriptional activation following heat shock was detected in present study or in other studies with S. meliloti.5,27 Similar to our results, no induction grpE was reported for S. meliloti by Sauviac and coworkers,27 while different study showed induction grpE by heat shock.5 Another heat shock protein that has close interaction with DnaKJ system is ClpB. The clpB gene (mll3429) was found to be overexpressed in present study, with an M-value 2.93. The clpB gene was already seen to be upregulated following heat shock in S. meliloti 5,27 and importance ClpB in rhizobia stress response, especially to heat shock, was also previously reported.63 Similar to E. coli, knockout clpB gene in Mesorhizobium ciceri led to an inability to endure high temperatures. Furthermore, in M. ciceri symbiotic performance was also negatively affected.63,64 These results are consistent with ClpB role in denatured protein disaggregation, namely by its cooperation with DnaKJ system.65 Page 10 12 Heat Shock Response Mesorhizobium loti MAFF303099 3.7. Other heat shock-inducible genes Among 50 genes with highest M-values (Table 2), there are five transcriptional regulators, one sigma factor and one anti-sigma factor, which indicates that heat shock response is complex system with relevant control at transcriptional level. Additional analysis all hypothetical proteins differentially expressed performed in this study, allowed further characterization many genes, for example mll4607, which is now annotated as Ku protein (Table 2). Together with LigD this protein is involved in DNA repair, namely in repair non-homologous end-joining double-strand DNA.70 Unlike other bacteria, rhizobial genomes encode multiple copies this Ku/LigD system, which has been further studied in S. meliloti.71 Although none ku homologues is required for symbiosis establishment, this DNA repair system is active in both free-living cells and bacteroids.71 From four ku homologs in MAFF303099 genome, three are induced by heat shock (mll4607, mlr9624 and mlr9623), as well as one three ligD homologues (mll9625). Until recently, double-stranded DNA breaks (DSB) were not thought to be consequence heat shock; however, recent study, using eukaryotic cells, showed that heat shock may in fact induce DSB on certain phases cell cycle.72 It is tempting to agree with suggestion from Kobayashi and coworkers71 that these systems do have some role under stress conditions, such as heat shock. Altogether our results suggest that in large bacterial genome, extensive gene downregulation may be an important part heat shock response. Although present study has contributed to further knowledge on rhizobia stress response, future studies are required to understand role individual genes and mechanisms regulating these molecular responses. Acknowledgements: We thank Ana Catarina Gomes from Biocant Park (Portugal) for her assistance with microarray data analysis and Owen Woody from University Waterloo (Canada) for his help with DNAplotter software. Supplementary data: Supplementary Data are available at www.dnaresearch.oxfordjournals.org. Funding ˆ This work was funded by Fundacao para Ciencia e ¸˜ Tecnologia (FCT), including research projects FCOMP-01-0124-FEDER-007091,PTDC/BIA-EVF/4158/ 2012 and strategic project PEst-C/AGR/UI0115/ 2011, that include FEDER funds through Operational Programme for Com-petitiveness Factors—COMPETE and National funds. A. A. and M. L. acknowledge postdoctoral fellowships from FCT (SFRH/BPD/73243/2010 and SFRH/BPD/27008/2006). References 1. Jensen, E.S. and Hauggaard-Nielsen, H. 2003, How can increased use biological N2 fixation in agriculture benefit environment? Plant and Soil, 252, 177– 86. 2. MacLean, A.M., Finan, T.M. and Sadowsky, M.J. 2007, Genomes symbiotic nitrogen-fixing bacteria legumes, Plant Physiol., 144, 615– 22. 3. Richter, K., Haslbeck, M. and Buchner, J. 2010, The heat shock response: life on verge death, Mol. Cell, 40, 253– 66. 4. Gruber, T.M. and Gross, C.A. 2003, Multiple sigma subunits and partitioning bacterial transcription space, Annu. Rev. Microbiol., 57, 441 –66. 5. Barnett, M.J., Bittner, A.N., Toman, C.J., Oke, V. and Long, S.R. 2012, Dual RpoH sigma factors and transcriptional plasticity in symbiotic bacterium, J. Bacteriol., 194, 4983 –94. ´nez-Salazar, J.M., Sandoval-Calderon, M., Guo, X.W., 6. Martı et al. 2009, The Rhizobium etli RpoH1 and RpoH2 sigma factors are involved in different stress responses, Microbiology, 155, 386– 97. 7. Mitsui, H., Sato, T., Sato, Y., Ito, N. and Minamisawa, K. 2004, Sinorhizobium meliloti RpoH(1) is required for effective nitrogen-fixing symbiosis with alfalfa, Mol. Genet. Genomics, 271, 416– 25. Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 The FixK is an activator for several operons, namely fiXNOQP and fixGHIS and fixK gene is upregulated by micro-oxic conditions.68 MAFF303099 genome encodes two fixNOPQ operons (encoding cytochrome oxidases), one located in symbiosis island. Interestingly, all fix genes found to be underexpressed ( fixK, fixJ, fixS, fixI, fixP, fixO, fixN) are outside symbiosis island. Uchiumi and collaborators69 suggested that rhizobia might have acquired housekeeping fixNOPQ operon before acquisition symbiosis island. Similar to present study, fix genes were previously detected to be underexpressed after heat shock in S. meliloti,5 so downregulation fixK cascade upon high-temperature conditions seems to be consistent. In S. meliloti, fixK is negatively regulated by activity fixT; however, no fixT gene is annotated in MAFF303099 genome (the most similar gene is msl5852, which is not differentially expressed in present study). From high number nodulation genes encoded in M. loti MAFF303099 genome (.40 genes) only 11 showed altered transcript levels after heat shock. The genes nodC and nodE were heat induced, while nine other nodulation genes were underexpressed. Only nodL was previously reported to be underexpressed following heat shock conditions in S. meliloti 5,27 but this gene expression remained unaltered in present study. A. Alexandre et al. 24. Nogales, J., Campos, R., BenAbdelkhalek, H., Olivares, J., Lluch, C. and Sanjuan, J. 2002, Rhizobium tropici genes involved in free-living salt tolerance are required for establishment efficient nitrogen-fixing symbiosis with Phaseolus vulgaris, Mol. Plant-Microbe Interact., 15, 225– 32. 25. Munchbach, M., Nocker, A. and Narberhaus, F. 1999, ¨ Multiple small heat shock proteins in rhizobia, J. Bacteriol., 181, 83– 90. 26. Narberhaus, F., Waldminighaus, T. and Chowdhury, S. 2006, RNA Thermometers, FEMS Microbiol. Rev., 30, 3– 16. 27. Sauviac, L., Philippe, H., Phok, K. and Bruand, C. 2007, An extracytoplasmic function sigma factor acts as general stress response regulator in Sinorhizobium meliloti, J. Bacteriol., 189, 4204– 16. 28. Kaneko, T., Nakamura, Y., Sato, S., et al. 2000, Complete genome structure nitrogen-fixing symbiotic bacterium Mesorhizobium loti, DNA Res., 7, 331– 8. 29. Laranjo, M., Alexandre, A. and Oliveira, S. Legume growthpromoting rhizobia: an overview on Mesorhizobium genus, Microbiol. Res. doi:10.1016/j.micres.2013.10 09.1012. 30. Laranjo, M. and Oliveira, S. 2011, Tolerance Mesorhizobium type strains to different environmental stresses, Antonie van Leeuwenhoek, 99, 651– 62. 31. Vincent, J.M. 1970, Manual for Practical Study Root-Nodule Bacteria. Blackwell Scientific Publications: Oxford. 32. Simon, R., Lam, A., Li, M.-C., Ngan, M., Menenzes, S. and Zhao, Y. 2007, Analysis gene expression data using BRB-ArrayTools, Cancer Inform., 3, 11– 7. 33. Saeed, A.I., Hagabati, N.K., Braisted, J.C., et al. 2006, TM4 microarray software suite, DNA Microarrays B Databases Stat., 411, 134– 93. 34. Gotz, S., Garcia-Gomez, J.M., Terol, J., et al. 2008, Highthroughput functional annotation and data mining with Blast2GO suite, Nucleic Acids Res., 36, 3420 –35. 35. Szklarczyk, D., Franceschini, A., Kuhn, M., et al. 2011, The STRING database in 2011: functional interaction networks proteins, globally integrated and scored, Nucleic Acids Res., 39, D561 – 8. 36. Price, M.N., Huang, K.H., Alm, E.J. and Arkin, A.P. 2005, novel method for accurate operon predictions in all sequenced prokaryotes, Nucleic Acids Res., 33, 880– 92. 37. Carver, T., Thomson, N., Bleasby, A., Berriman, M. and Parkhill, J. 2009, DNAPlotter: circular and linear interactive genome visualization, Bioinformatics, 25, 119 – 20. 38. Madsen, M.L., Nettleton, D., Thacker, E.L., Edwards, R. and Minion, F.C. 2006, Transcriptional profiling Mycoplasma hyopneumoniae during heat shock using microarrays, Infect. Immun., 74, 160–6. 39. Crapoulet, N., Barbry, P., Raoult, D. and Renesto, P. 2006, Global transcriptome analysis Tropheryma whipplei in response to temperature stresses, J. Bacteriol., 188, 5228– 39. 40. Audia, J.P., Patton, M.C. and Winkler, H.H. 2008, DNA microarray analysis heat shock transcriptome Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 8. Hartl, F.U. and Hayer-Hartl, M. 2009, Converging concepts protein folding in vitro and in vivo, Nat. Struct. Mol. Biol., 16, 574– 81. 9. Ferrer, M., Chernikova, T.N., Yakimov, M.M., Golyshin, P.N. and Timmis, K.N. 2003, Chaperonins govern growth Escherichia coli at low temperatures, Nat. Biotechnol., 21, 1266– 7. 10. Kim, S.Y., Ayyadurai, N., Heo, M.A., Park, S., Jeong, Y.J. and Lee, S.G. 2009, Improving productivity recombinant protein in Escherichia coli under thermal stress by coexpressing GroELS chaperone system, J. Microbiol. Biotechnol., 19, 72– 7. 11. Rudolph, B., Gebendorfer, K.M., Buchner, J. and Winter, J. 2010, Evolution Escherichia coli for growth at high temperatures, J. Biol. Chem., 285, 19029– 34. 12. Bittner, A.N., Foltz, A. and Oke, V. 2007, Only one five groEL genes is required for viability and successful symbiosis in Sinorhizobium meliloti, J. Bacteriol., 189, 1884– 9. 13. Babst, M., Hennecke, H. and Fischer, H.M. 1996, Two different mechanisms are involved in heat-shock regulation chaperonin gene expression in Bradyrhizobium japonicum, Mol. Microbiol., 19, 827– 39. 14. Gould, P.S., Burglar, H.R. and Lund, P.A. 2007, Homologous cpn60 genes in Rhizobium leguminosarum are not functionally equivalent, Cell Stress Chaperon., 12, 123– 31. 15. Alexandre, A. and Oliveira, S. 2013, Response to temperature stress in rhizobia, Crit. Rev. Microbiol., 39, 219– 28. 16. Alexandre, A. and Oliveira, S. 2011, Most heat-tolerant rhizobia show high induction major chaperone genes upon stress, FEMS Microbiol. Ecol., 75, 28– 36. 17. Gould, P., Maguire, M. and Lund, P.A. 2007, Distinct mechanisms regulate expression two major groEL homologues in Rhizobium leguminosarum, Arch. Microbiol., 187, 1– 14. 18. Bittner, A.N. and Oke, V. 2006, Multiple groESL operons are not key targets RpoH1 and RpoH2 in Sinorhizobium meliloti, J. Bacteriol., 188, 3507–15. 19. Zuber, U. and Schumann, W. 1994, CIRCE, novel heatshock element involved in regulation heat-shock operon dnaK Bacillus subtilis, J. Bacteriol., 176, 1359 –63. 20. Schulz, A. and Schumann, W. 1996, hrcA, first gene Bacillus subtilis dnaK operon encodes negative regulator class I heat shock genes, J. Bacteriol., 178, 1088 –93. 21. Minder, A.C., Narberhaus, F., Babst, M., Hennecke, H. and Fischer, H.M. 1997, The dnaKJ operon belongs to sigma(32)-dependent class heat shock genes in Bradyrhizobium japonicum, Mol. Gen. Genet., 254, 195 –206. 22. Labidi, M., Laberge, S., Vezina, L.P. and Antoun, H. 2000, The dnaJ (hsp40) locus in Rhizobium leguminosarum bv. phaseoli is required for establishment an effective symbiosis with Phaseolus vulgaris, Mol. Plant-Microbe Interact., 13, 1271– 4. 23. Gomes, D.F., da Silva Batista, J.S., Schiavon, A.L., Andrade, D.S. and Hungria, M. 2012, Proteomic profiling Rhizobium tropici PRF 81: identification conserved and specific responses to heat stress, BMC Microbiol., 12, 84. Page 11 12 Page 12 12 41. 42. 43. 44. 45. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. obligate intracytoplasmic pathogen Rickettsia prowazekii, Appl. Environ. Microbiol., 74, 7809– 12. Stintzi, A. 2003, Gene expression profile Campylobacter jejuni in response to growth temperature variation, J. Bacteriol., 185, 2009– 16. Li, J.-s., Bi, Y.-t., Dong, C., Yang, J.-f. and Liang, W.-d. 2011, Transcriptome analysis adaptive heat shock response Streptococcus thermophilus, PLoS ONE, 6, e25777. Rohlin, L., Trent, J.D., Salmon, K., Kim, U., Gunsalus, R.P. and Liao, J.C. 2005, Heat shock response Archaeoglobus fulgidus, J. Bacteriol., 187, 6046 –57. Rezzonico, E., Lariani, S., Barretto, C., et al. 2007, Global transcriptome analysis heat shock response Bifidobacterium longum, FEMS Microbiol. Lett., 271, 136–45. Koide, T., Vencio, R.Z.N. and Gomes, S.L. 2006, Global gene expression analysis heat shock response in phytopathogen Xylella fastidiosa, J. Bacteriol., 188, 5821– 30. van der Veen, S., Hain, T., Wouters, J.A., et al. 2007, The heat-shock response Listeria monocytogenes comprises genes involved in heat shock, cell division, cell wall synthesis, and SOS response, Microbiology, 153, 3593– 607. Yin, H., Tang, M., Zhou, Z., et al. 2012, Distinctive heatshock response bioleaching microorganism Acidithiobacillus ferrooxidans observed using genome-wide microarray, Can. J. Microbiol., 58, 628– 36. Barreiro, C., Nakunst, D., Hueser, A.T., Paz, H.D., Kalinowski, J. and Martin, J.F. 2009, Microarray studies reveal ‘differential response’ to moderate or severe heat shock HrcA- and HspR-dependent systems in Corynebacterium glutamicum, Microbiology, 155, 359–72. Chhabra, S.R., He, Q., Huang, K.H., et al. 2006, Global analysis heat shock response in Desulfovibrio vulgaris Hildenborough, J. Bacteriol., 188, 1817– 28. Emerson, J.E., Stabler, R.A., Wren, B.W. and Fairweather, N.F. 2008, Microarray analysis transcriptional responses Clostridium difficile to environmental and antibiotic stress, J. Med. Microbiol., 57, 757–64. Richmond, C.S., Glasner, J.D., Mau, R., Jin, H.F. and Blattner, F.R. 1999, Genome-wide expression profiling in Escherichia coli K12, Nucleic Acids Res., 27, 3821– 35. Zhang, W., Culley, D.E., Nie, L. and Brockman, F.J. 2006, DNA microarray analysis anaerobic Methanosarcina barkeri reveals responses to heat shock and air exposure, J. Ind. Microbiol. Biotechnol., 33, 784– 90. Gao, H.C., Wang, Y., Liu, X.D., et al. 2004, Global transcriptome analysis heat shock response Shewanella oneidensis, J. Bacteriol., 186, 7796– 803. Wren, J.D. and Conway, T. 2006, Meta-analysis published transcriptional and translational fold changes reveals preference for low-fold inductions, OMICS, 10, 15– 27. Koonin, E.V. 2002 (updated 2003), The Clusters Orthologous Groups (COGs) Database: Phylogenetic Classification Proteins from Complete Genomes. In: McEntyre, J. and Ostell, J. (eds), The NCBI Handbook, National Center for Biotechnology Information (US). Konstantinidis, K.T. and Tiedje, J.M. 2004, Trends between gene content and genome size in prokaryotic species with larger genomes, Proc. Natl. Acad. Sci. USA, 101, 3160 –5. 57. Nocker, A., Krstulovic, N.P., Perret, X. and Narberhaus, F. 2001, ROSE elements occur in disparate rhizobia and are functionally interchangeable between species, Arch. Microbiol., 176, 44– 51. 58. Studer, S. and Narberhaus, F. 2000, Chaperone activity and homo- and hetero-oligomer formation bacterial small heat shock proteins, J. Biol. Chem., 275, 37212– 18. 59. Rodriguez-Quinones, F., Maguire, M., Wallington, E.J., et al. 2005, Two three groEL homologues in Rhizobium leguminosarum are dispensable for normal growth, Arch. Microbiol., 183, 253– 65. 60. Segal, G. and Ron, E.Z. 1995, The dnaKJ operon Agrobacterium tumefaciens: transcriptional analysis and evidence for new heat shock promoter, J. Bacteriol., 177, 5952 –8. 61. Xayaphoummine, A., Bucher, T. and Isambert, H. 2005, Kinefold web server for RNA/DNA folding path and structure prediction including pseudoknots and knots, Nucleic Acids Res., 33, W605– 10. 62. Wang, J.D., Herman, C., Tipton, K.A., Gross, C.A. and Weissman, J.S. 2002, Directed evolution substrateoptimized GroEL/S chaperonins, Cell, 111, 1027– 39. ´gido, C., Robledo, M., Menendez, E., Mateos, P.F. and 63. Brı Oliveira, S. 2012, ClpB chaperone knockout mutant Mesorhizobium ciceri shows delay in root nodulation chickpea plants, Mol. Plant-Microbe Interact., 25, 1594– 604. 64. Squires, C.L., Pedersen, S., Ross, B.M. and Squires, C. 1991, ClpB is Escherichia coli heat-shock protein F84.1, J. Bacteriol., 173, 4254 –62. 65. Seyffer, F., Kummer, E., Oguchi, Y., et al. 2012, Hsp70 proteins bind Hsp100 regulatory M domains to activate AAA plus disaggregase at aggregate surfaces, Nat. Struct. Mol. Biol., 19, 1347 –56. 66. Ono, Y., Mitsui, H., Sato, T. and Minamisawa, K. 2001, Two RpoH homologs responsible for expression heat shock protein genes in Sinorhizobium meliloti, Mol. Gen. Genet., 264, 902– 12. 67. Guisbert, E., Yura, T., Rhodius, V.A. and Gross, C.A. 2008, Convergence molecular, modeling, and systems approaches for an understanding Escherichia coli heat shock response, Microbiol. Mol. Biol. Rev., 72, 545– 54. 68. Dixon, R. and Kahn, D. 2004, Genetic regulation biological nitrogen fixation, Nat. Rev. Microbiol., 2, 621– 31. 69. Uchiumi, T., Ohwada, T., Itakura, M., et al. 2004, Expression islands clustered on symbiosis island Mesorhizobium loti genome, J. Bacteriol., 186, 2439– 48. 70. Gong, C.L., Bongiorno, P., Martins, A., et al. 2005, Mechanism nonhomologous end-joining in mycobacteria: low-fidelity repair system driven by Ku, ligase D and ligase C, Nat. Struct. Mol. Biol., 12, 304– 12. 71. Kobayashi, H., Simmons, L.A., Yuan, D.S., Broughton, W.J. and Walker, G.C. 2008, Multiple Ku orthologues mediate DNA non-homologous end-joining in free-living form and during chronic infection Sinorhizobium meliloti, Mol. Microbiol., 67, 350–63. 72. Velichko, A.K., Petrova, N.V., Kantidze, O.L. and Razin, S.V. 2012, Dual effect heat shock on DNA replication and genome integrity, Mol. Biol. Cell, 23, 3450– 60. Downloaded from http://dnaresearch.oxfordjournals.org/ by guest on January 24, 2014 46. Heat Shock Response Mesorhizobium loti MAFF303099