In order to identify Hsp90 genes, the sequences of seven AtHsp90 proteins were retrieved from the A. thaliana genome,¹ and were subsequently used for determining the Hsp90 genes in the genomes of B. napus (v5),² B. rapa (Brara_Chiifu_V3.0)³ and B. oleracea (Brara_Chiifu_V3.0, Braol_JZS_V2.0; see footnote 3) via reciprocal blast with BLASTP program (Altschul et al., 1997). The default parameters with E-value less than 1E-10 were set in the BLASTP searches.

The Hsp90 protein sequences of B. napus, B. rapa and B. oleracea and A. thaliana were merged and performed multiple sequence alignment using ClustalW program (Larkin et al., 2007). Phylogenetic relationships of Hsp90 genes were constructed with alignment results through FastTree software (Price et al., 2009).

Chromosomal position information for BnHsp90 genes was extracted from generic feature format (GFF) files which were downloaded from B. napus genome website (see footnote 2). The positions of BnHsp90 genes were indicated on the corresponding chromosomes with TBtools software (Chen et al., 2020). MCscanX software (Wang et al., 2012) was used to identify the gene duplication types and collinearity relationships. Self-self comparison was performed using BLASTP program between B. napus protein sequences with the e-value under 1e-10. Duplication types of all the B. napus whole genome proteins were identified using the program incorporated in MCScanX, and then duplication information of BnHsp90 gene family was isolated. The collinear gene pairs within BnHsp90 gene family were detected using the program detect_collinearity_within_gene_families.pl. of MCscanX, and the collinearity relationships of these gene pairs were displayed using Circos software (Krzywinski et al., 2009).

The information of gene structures (exon/intron) for BnHsp90 genes was separated from GFF files and displayed using TBtools software (Chen et al., 2020). Protein sequences of BnHsp90 were submitted to MEME website⁴ for conserved motif analysis, and the display of conserved motifs were conducted in TBtools software (Chen et al., 2020). The R library ggseqlogo (Wagih, 2017) was employed to generate the sequence logos of the five most detected conserved motifs.

Promoter information of BnHsp90 genes was extracted from GFF files of B. napus, and the promoter sequences were isolated using seqtk software (Li, 2013). A total of 2,000 bp upstream sequences of the coding region were obtained and submitted to PlantCare⁵ for cis-elements analysis.

To investigate the transcriptional patterns of BnHsp90 genes in response to salt and the infection of S. sclerotiorum, SRA files were downloaded from GEO RNA-seq datasets (GSE81545) and Sequence Read Archive (SRA) database (PRJNA561674). All reads were checked the quality and cleaned using fastp software (Chen et al., 2018). Gene expression quantification was performed using a software package called Kallisto (Bray et al., 2016). The transcripts per kilobase of exon model per million mapped reads (TPM) values were separated and heat maps were constructed from relative TPM value using pHeatmap software package (Raivo, 2012).

B. napus plants are grown in an incubator at 22°C ~ 25°C, with a day and night photoperiod of 16/8 h. Variety K407 was provided by the Hybrid Rapeseed Research Center of Shanxi Province, and was used for treatment. Seeds of K407 were sterilized in 30% bleach and germinated on one-layer nylon mesh with support below, floating onto the sterile water. Seven-day-old seedlings after germination were transferred into Hoagland solution with foams as supports. The Hoagland solution was replaced every week until the plants ready for treatment. Plants in four-leaf stage were moved to the Hoagland solution containing 150 mM NaCl or without NaCl. After processing, samples were taken at 0 h, 1 h, 3 h, and 6 h, respectively. The collected true leaves were quickly frozen in liquid nitrogen. Sclerotinia sclerotiorum strains (separated from the stubbles of oilseed rape in the fields of Kaifeng, Henan, designated as Hns-1) were cultured on potato dextrose agar (PDA) plates. The actively growing mycelium at the edge of the colony on the mycelium agar block was taken out and inoculated onto the leaves of B. napus. Similarly, as a mock inoculation control, the leaves were inoculated with blank PDA agar blocks. The leaves from the inoculation area were collected at 0 h, 12 h, 24 h, and 48 h after inoculation, and immediately frozen in liquid nitrogen. All samples were stored in the refrigerator at −80°C until RNA extraction. Three independent plants produce a biological replicate, and each treatment consists of three independent biological replicates.

The materials and corresponding treatments in RNA-seq analysis from publicly available data can be found in the reports by Tan et al. (2019) and Girard et al. (2017).

RNAprep Pure Plant Plus Kit (TIANGEN, China) was used to extract total RNA. One microgram of total RNA was revers-transcribed into cDNA using HiScript II Q RT SuperMix (Vazyme, R223, Nanjing). Quantitative RT-PCR (qPCR) was performed using ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing). Relative transcript levels of BnHsp90s were analyzed by quantitative real-time PCR using actin gene (BnaA10g22340D) as an internal control. The PCR amplification was performed using LightCycler^®480 (Roche Diagnostics, United States). All experiments were performed from three biologically independent RNA samples, and each qPCR test had three replicates. Gene-specific primers were designed with Primer Premier 5.0 (PREMIER Biosoft, USA). The detailed information on the primers is shown in Supplementary Table S1.

Open reading frame of BnA03.Hsp90.2 was cloned into pEarLeyGate103 to generate a green fluorescent protein (GFP)-tagged expression vector (103-BnA03.Hsp90.2). GFP under the control of 35S promoter was used as negative control (35S-GFP), and H2B fused with mCherry was used for positive control (H2B-mCherry). Agrobacterium containing corresponding plasmids was infiltrated into Nicotiana benthamiana leaves for transient expression of fusion protein. Forty-eight hours after infiltration, the fluorescence signal was observed using a confocal laser scanning microscope (Leica TCS SP8).

The complete coding sequence of BnA03.Hsp90.2 was fused into the pGBKT7 vector to generate pGBKT7-BnA03.Hsp90.2 plasmid. The pGBKT7-BnA03.Hsp90.2, pGBKT7-AtDREB (positive control) or pGBKT7 empty vector (negative control) was transformed into the AH109 yeast strain separately. The obtained positive transformants were grown on SD/-Trp and SD/-Trp-His medium. The transcriptional activation of each transformant was assessed based on their survival after incubation at 30°C for 3 days. Primers used for constructing the plasmids were stored in Supplementary Table S1.

Expression levels were calculated using 2^−ΔΔCT method (Livak and Schmittgen, 2001). ANOVA analysis was employed to assay the difference between samples or treatments, with value of p less than 0.05 to indicate the significant differences. The office software Excel 2016 was used for data analysis. The expression levels were from three different samples or treatments. Values in the histogram were represented as average values ± standard errors.

The sequences of seven A. thaliana Hsp90 proteins were used to identify the Hsp90 family members in the B. napus genome via reciprocal BLASTP. A total of 35 candidate Hsp90 genes were detected, and were designated with their orders on each chromosome (Supplementary Table S2). The number of Hsp90s allocated on each chromosome is different: chromosome A03 contains the greatest number (6) of Hsp90 genes, chromosome C02 ranks the second, with 5 Hsp90s on it, and followed by chromosome C03, with 4 Hsp90s on it (Supplementary Figure S1). BnHsp90 genes are not randomly distributed on each chromosome, and mostly distributed at both ends of chromosomes, with many genes clustered (Supplementary Figure S1; Supplementary Table S2). On the chrA03, chrC02 and chrC03 chromosome, there are four or five Hsp90 genes distributed in clusters (Supplementary Figure S1; Supplementary Table S2). Overall, a total of 35 BnHsp90 genes were identified, with some genes clustered on ends of chromosomes.

The number of amino acids and biophysical properties vary among BnHsp90 proteins. The lengths of BnHsp90 proteins range from 63 to 872 aa (Supplementary Table S2). There are 27 BnHsp90 proteins with more than 300 aa (Supplementary Table S2). Accordingly, the molecular weights of different Hsp90s differ greatly, with fluctuations ranging from 7016.9 to 100512.9 Da (Supplementary Table S2). The number of exons encoding proteins is between 2 and 21 (Supplementary Table S2). In addition, the theoretical isoelectric point is distributed between 3.8 and 9.2, with BnA10.Hsp90.2 having the highest isoelectric point, 9.2, showing slightly alkalinity (Supplementary Table S2). Besides BnA10.Hsp90.2, the isoelectric point of other BnHsp90 proteins falls between 3.84 and 5.99, which is slightly acidic (Supplementary Table S2). To provide useful clues to reveal the functions of BnHsp90s, the subcellular locations of 35 Hsp90s proteins were predicted, and they were indicated to be in different organelles: There are 7 proteins located in cytoplasmic, 21 proteins in mitochondrial, 3 proteins in nuclear, and 4 proteins in other organelles (Supplementary Table S2).

In order to further clarify the evolutionary relationship of the Hsp90 gene family, FastTree software was used to construct a phylogenetic tree of Hsp90 proteins of A. thaliana, B. rapa, B. oleracea and B. napus. A phylogenetic tree was constructed using 7 AtHsp90s from A. thaliana, 17 BrHsp90s from B. rapa, 18 BoHsp90s from B. oleracea and 35 BnHsp90s from B. napus. According to the phylogenetic relationship of these proteins, Hsp90s were divided into five groups (groups I–V), containing 17, 11, 18, 17, and 14 proteins, respectively (Figure 1). The third group has the largest number of members, with 18. There is no Hsp90 from Arabidopsis in the fourth group. By analyzing the subcellular location, it was found that most of the genes in groups I, II, and III were located to mitochondria.

The exon-intron structure analysis can provide important evidence support for the evolution of gene families. In order to analyze the exon-intron structure of BnHsp90 coding region, the genome and coding sequences of BnHsp90 were compared. The results showed that most members of BnHsp90s contained multiple exons, and the number of introns vary greatly (Supplementary Table S2; Figure 2A). As shown in Figure 2A; Supplementary Table S2, the gene structures are complicated; BnA10.Hsp90.1, BnC02.Hsp90.1, and BnC02.Hsp90.2 contains two exons, but BnC05.Hsp90.2 contains 21 exons (the greatest number; Supplementary Table S2).

We further analyzed the conserved motifs of BnHsp90s and determined the conservative pattern of amino acid residues in their domains. The most 5 conserved motifs detected in BnHsp90 are indicated in Figure 2B, containing 29–43 amino acids. Motif 1 and 2 contain 30 amino acids, and Motif 3 and 4 contain 43 amino acids (Figure 2B). Among them, motif 5 has the lowest amino acid content, which is 29 amino acids. Among all BnHsp90 members, 25 proteins contain motif 1, 25 proteins include motif 2, 27 proteins include motif 3, 26 proteins include motif 4 and 26 proteins include motif 5 (Figure 2B). Further analysis indicated that most of the motif 2, motif 4, motif 5 are more conservative. Similar to the intron patterns observed in each gene, the number of conserved motifs was relatively consistent in a single Hsp90 gene paralog. Two-thirds of Hsp90 genes contain motifs 1–5. Among them, some very conservative sequences were discovered, similar to the results in previous studies (Zhang et al., 2017).

A syntenic analysis was performed between A. thaliana, B. napus and its ancestors to speculate the evolutionary origin of Hsp90 genes. The results indicated that many chromosomal rearrangements or gene duplications exist between B. napus and B. rapa and B. oleracea, and strong collinearity was also detected between them (Figures 3A,B). Many homologous and synchronized parts were identified in chromosomes A and C of B. napus, as well between the genome of B. rapa or B. oleracea and genome of B. napus (Figure 3A). In the syntenic analysis of the Hsp90s of B. napus and A. thaliana, five of the seven AtHsp90s were found to be syntenic gene pairs with 16 BnHsp90s (Figure 3A; Supplementary Table S4), with AT4G24190 having the highest number of syntenic genes in B. napus, 5 BnHsp90 genes, followed by AT3G07770, containing three syntenic BnHsp90 genes (Supplementary Table S4; Figure 1). This indicates an expansion of the BnHsp90 gene family in B. napus after its evolutionary separation from A. thaliana. Of all the syntenic BnHsp90 genes, some members (BnC02.Hsp90.5, BnC07.Hsp90.1, BnA03.Hsp90.6, BnC01.Hsp90.1, BnC08.Hsp90) have syntenic genes in A. thaliana, but do not have any corresponding syntenic genes in B. rapa and B. oleracea, and three genes (BnC09.Hsp90.2, BnC03.Hsp90.4, BnC05.Hsp90.1) just have corresponding collinear genes in B. oleracea but not in B. rapa (Figure 3A; Supplementary Table S3). Only one gene (BnC07.Hsp90.2) has collinear genes in B. rapa but not in B. oleracea (Figure 3A; Supplementary Table S3). Moreover, no syntenic genes were identified in A. thaliana for three BnHsp90s, (BnA03.Hsp90.1, BnC05.Hsp90.1, BnA03.Hsp90.6), however, the respective collinear genes for them can be detected in B. rapa and/or B. oleracea (Figure 3A; Supplementary Table S3).

Different gene duplication may be one of the reasons for the expansion of plant gene family. Three duplication models were detected in 35 BnHsp90 genes, including whole genome duplication (WGD), dispersed and proximal. Among them, the most detected duplication mode is WGD, and there are 25 (71.4%) genes produced by WGD, followed by dispersed and proximal, both producing 5 BnHsp90 genes (Supplementary Table S2).

The cis-acting elements in the promoter region play important roles in the plant responses to stress, and as well, participate in the responses to drought, ABA and other stresses (Yazaki and Kikuchi, 2005; Nakashima and Yamaguchi-Shinozaki, 2013). The ten most abundant cis-acting elements detected in BnHsp90s are CAAT-box, G-box, MYB, ABRE, MYC, ARE, as-1, CGTCA-motif, Unnamed_1, and Unnamed_4. CAAT-box and Unnamed_4 have the largest number of cis-acting elements identified in the BnHsp90 promoter region, followed by G-box, MYB, ABRE and MYC. Among them, a total of 143 MYB binding sites and 93 MYC binding sites were identified in the promoter region of the BnHsp90 genes, and 131 ABREs were identified, participating in the ABA response (Figure 4).

In the long-term evolutionary process, plants have formed a set of response mechanisms to abiotic stresses. It is known that Hsp90 genes will be induced in response to these abiotic stresses (Krishna, 1998). In this study, we first analyzed the transcriptome data of B. napus seedlings under salt stress in existing studies, and obtained the changes in the expression of BnHsp90 gene after salt stress (Tan et al., 2019; Figure 5). After salt stress, the expression of 9 BnHsp90 genes in B. napus seedling leaves was down-regulated, and the expression of 7 BnHsp90 genes was up-regulated (Figure 5A). Among them, the expression levels of BnA02.Hsp90.1, BnA02.Hsp90.2, BnA03.Hsp90.2, BnC02.Hsp90.1, and BnC02.Hsp90.3 changed significantly (Figure 5A). In B. napus roots, the changes in BnHsp90 gene expression were less affected by salt stress (Figure 5A). We selected 14 BnHsp90 genes with obvious changes in expression for verification (Figure 5B). In order to verify the expression changes of BnHsp90 gene under salt stress, B. napus seedlings were treated with 150 mM NaCl, and seedlings of 0 h, 1 h, 3 h, 6 h post-treatment were sampled for performing qPCR on 14 BnHsp90 genes, respectively. Following NaCl treatment, the expression of four BnHsp90 genes (BnA02.Hsp90.1, BnA02.Hsp90.2, BnA03.Hsp90.6, BnC02.Hsp90.2,) increased first and then decreased (Figure 5B). Transcripts of ten BnHsp90 genes (BnA03.Hsp90.1, BnA03.Hsp90.2, BnA03.Hsp90.4, BnA03.Hsp90.5, BnC02.Hsp90.1, BnC02.Hsp90.3, BnC02.Hsp90.5, BnC03.Hsp90.2, BnC08.Hsp90, BnC09.Hsp90.2) reduced after salt stress treatment (Figure 5B). Overall, more BnHsp90 genes were reduced by the salt treatment.

S. sclerotiorum is a severe pathogen on B. napus (Bolton et al., 2006). Previous studies have shown that the BnHsp90 gene responds to S. sclerotiorum (Cao et al., 2016). In the present study, expression of BnHsp90s were determined according to the public RNA-seq data of B. napus after treatment with S. sclerotiorum (Girard et al., 2017). As shown in Figure 6, only 27 of the 35 BnHsps displayed obvious expression in the four tissues; leaves of Westar without S. sclerotiorum inoculation (Westar_C), Westar 24 h post-inoculated with S. sclerotiorum (Westar_24 h), ZY821 without S. sclerotiorum treatment (ZY821_C), and ZY821 24 h post-inoculated with S. sclerotiorum (ZY821_24 h). Nine of these 27 genes were induced for their expression by the infection of S. sclerotiorum in leaves of both varieties, susceptible Westar and resistant ZY821 (Figure 6A). To validate these results, nine genes were selected for qPCR verification (Figure 6B). According to the results, the expression patterns of them were coincident with the RNA-seq data (Figure 6B). There are six BnHsp90 genes up-regulated (BnC01.Hsp90.1, BnA01.Hsp90, BnA03.Hsp90.1, BnC03.Hsp90.1, BnA03.Hsp90.3, BnA03.Hsp90.5; Figure 6B). Among them, the transcript level of BnC01.Hsp90.1 showed the most drastic changes (Figure 6B). Three BnHsp90 genes (BnC09.Hsp90.1, BnA06.Hsp90, BnA09.Hsp90) were down-regulated post inoculation by S. sclerotiorum (Figure 6B). These results indicate that BnHsp90 gene may be involved in the responses of B. napus to sclerotinia disease. These up-regulated genes will be our target for investigating their roles in the resistance of B. napus to S. sclerotiorum.

There are three BnHsp90 proteins (BnA03.Hsp90.2, BnA03.Hsp90.4, and BnC03.Hsp90.2) targeted to nucleus in the prediction of subcellular location. To verification of the subcellular location, one of them, BnA03.Hsp90.2 was fused with GFP to construct 103-BnA03.Hsp90.2 plasmid, which was used for transiently expressed in tobacco leaves. GFP signal was detected only in the nucleus, indicating by the mCherry signal of H2B, a nucleus-localized protein (Supplementary Figure S2A). While the GFP fluorescence (control) was observed throughout the cell (Supplementary Figure S2B). To test whether BnA03.Hsp90.2 has transcriptional activity, full length of BnA03.Hsp90.2 coding region were cloned into the pGBKT7 vector, then transformed into yeast strain AH109. Yeast cells harboring pGBKT7-BnA03.Hsp90.2, pGBKT7 (negative control) and pGBKT7-AtDREB (positive control) grew well on the SD/−Trp medium, however, only the positive control survived on the selective medium SD/−Trp-His (Supplementary Figure S2C). These results demonstrated that BnA03.Hsp90.2 was a nucleus-localized protein but did not have transactivation activity for its full-length protein in yeast.