Molecular Evolution Of Heat Shock Transcripiton Factor Gene Families In Legumes And Grasses

With the increase of global greenhouse effect, high temperature has become a mainfactor resulted in the reduction of both the agricultural yield and quality. Heat shocktranscription factors (Hsfs) serve as the terminal components of signal transduction and arethe central regulators of the expression of heat shock proteins and other heat shock-inducedgenes, and have important roles in improving the thermotolerance of plants. Currentprogresses mostly concentrated in the function of Hsfs in Arabidopsis and tomato, however,the genome structures and evolutionary patterns of the entire Hsf gene families are notclearly understood in plants. As more and more genomes of species were sequenced, thereis a chance to shed some light on this question. Therefore, in this study we analyzed theHsf gene families from six legume species for which substantial information aboutgenomes or transcriptomes was available, namely Lotus japonicus, Medicago truncatula,Cicer arietinum, Glycine max, Cajanus cajan and Phaseolus vulgaris. Moreover, the Hsfgene families in four grass genomes of Brachypodium distachyon, Oryza sativa, Sorghumbicolor and Zea mays were analyzed comprehensively. The origin and evolution, and geneduplication and loss of Hsf gene familes in legumes and grasses were studied based oninvestigation of intron/exon distribution patterns, protein domains and motifs, phylogeneticrelationships, intraspecies and interspecies gene colinearity (microsynteny), gene copynumber changes, environmental selection pressure as well as expression patterns of Hsfgenes. The results were as follows:1. By searching published genome and transcriptome databases, a total of11,19and13Hsfs were identified in the cool season legumes Lotus japonicus, Medicago truncatulaand Cicer arietinum, respectively, while46,22and29Hsfs were identified in the tropicalseason legumes Glycine max, Cajanus cajan and Phaseolus vulgaris, respectively. Fiveconserved domains or motifs were observed in most of the legume Hsf proteins, namelyDBD, HR-A/B, NLS, NES domains and AHA motifs. The highly structured N-terminalDBD domain of each Hsf was most conserved; it consisted of a three-helical bundle and afour-stranded antiparallel β-sheet. The AHA motifs in the C-terminus of the Class A Hsfswere highly conserved.2. The analysis of the legume Hsf gene structure in terms of intron/exon distributionpatterns revealed that among the159introns,140were phase0, and accordingly there werethe presence of an excess of symmetrical exons. Besides, in the DBD domain a highlyconservative intron insertion site was found and all with phase0. These results suggested that exon shuffling and elimination of intron may contribute to the evolution of legume Hsfgenes.3. The phylogenetic analysis showed that the140Hsf genes from the six legumespecies could be delineated into18well-supported clades, and each clade represented anancient gene lineage. Therefore, there were at least18Hsf genes in the most recentcommon ancestor of these legumes.4. By searching for intraspecies microsynteny between the genome segments oflegumes and dating the age distributions of duplicated genes, we found that the expansionof legume Hsf gene families was mainly through whole genome duplication rather thantandem duplication. The Hsf genes of Glycine max derived from the early-legume genomeduplication and the recent Glycine-lineage-specific polyploidy event. Moreover,42of46the chromosome regions hosting Hsf genes in Glycine max fell into pairs, triples orquadruples and formed paralogous groups of segments, while only a few paralogoussegments were identified in the genomes of Lotus japonicas, Medicago truncatula andCajanus cajan.5. By comparing interspecies microsynteny between the genome segments of legumes,we determined that the great majority of Hsf-containing segments in Lotus japonicas,Medicago truncatula and Cajanus cajan show extensive conservation with the duplicatedregions of Glycine max. These segments formed17groups of orthologous segments. Theseresults suggested that these regions shared ancient genome duplication with Hsf genes inGlycine max, but more than half of the copies of these genes were lost. On the other hand,the Glycine max Hsf gene family retained approximately75%and85%of duplicated genesproduced from the ancient genome duplication and recent Glycine-specific genomeduplication, respectively. Selection pressure analysis indicated that continuous purifyingselection has played a key role in the maintenance of Hsf genes in Glycine max, and theduplicated genes were subject to strong evolutionary constraints to retain the stability oftheir functions.6. The further analysis of grass genomes showed that24,25,23and25Hsf geneswere identified in Brachypodium distachyon, Oryza sativa, Sorghum bicolor and Zea mays,respectively. By searching for intraspecies gene colinearity and dating the age distributionsof duplicated genes, we found that in Brachypodium distachyon, Oryza sativa, Sorghumbicolor and Zea mays genomes more than60%Hsf-containing segments havemicrosynteny, and resulted from whole genome duplication. The Hsf gene family of Zeamays originated through the ancient whole genome duplication event occurred in the ancestor of grasses and the recent polyploidy event in the ancestor of Zea mays.7. By comparing interspecies gene colinearity between grasses, extensivemicrosynteny was also detected between Hsf-containing segments across Brachypodiumdistachyon, Oryza sativa, Sorghum bicolor and Zea mays genomes, and all94segmentsformed17groups of orthologous segments. Thus, approximately32%,29%,35%and44%of duplicated Hsf genes produced from the ancient genome duplication occurred in grassancestor were lost in Brachypodium distachyon, Oryza sativa, Sorghum bicolor and Zeamays, respectively. In addition, approximately34%of Zea mays Hsf genes, which havebeen obtained from recent genome duplication in the ancestor of Zea mays, were lost justover the past13millions years. These results suggested that in the Zea mays genome arelatively large number of ancient copies of Hsf genes have been removed, and recentcopies have a faster loss rate. Selection pressure analysis indicated that purifying selectionstill played a leading role throughout the evolution of Hsf gene families in grasses, whilestrong signatures of positive selection was detected in some parts of coding regions in theindividual genes, suggesting functional differentiation.8. The results of expression analyses of Hsf genes from Lotus japonicus and Zea maysdemonstrated that they were differentially expressed in different tissue types and abioticstresses. LjHsf-01, LjHsf-02, LjHsf-04, LjHsf-09and LjHsf-10genes of Lotus japonicus, aswell as ZmHsf-01, ZmHsf-03, ZmHsf-04, ZmHsf-23, ZmHsf-24and ZmHsf-25genes of Zeamays were significantly up-regulated by heat stress. Among these genes, microsyntenyanalysis and selection pressure analysis have proved that ZmHsf-01and ZmHsf-04ofsubclass A2were highly conserved in the evolution of grasses and the stability of thefunction. They may play an important role in the heat shock stress resistance in grasses.Therefore, we cloned the full-length genes of ZmHsf-01and ZmHsf-04from inbred lineB73and laid a foundation of further study on the function.In summary, by using the methods of comparative genomics this study demonstratedthat the evolution of legume and grass Hsf gene families were coupling with proposedwhole genome duplication events, Hsf genome segments have extensive microsyntenybetween species, and the difference of gene loss events have contributed to the divergenceof these gene families in different plant lineages. These results can serve as an importantbasis for resolving molecular evolution of the Hsf gene family on the genome-wide level.