In this study, the rice variety is Nipponbare and 9311. Seeds were rinsed with distilled water, and then germinated at 28°C under dark conditions. After 48h, seedlings with a root length of approximately 0.5 cm were moved to hydroponic culture boxes (day/night temperatures of 28°C/22°C, light/dark photoperiod of 12h/12h, and light intensity of 18,000 Lx). At the one-leaf stage, the seedlings were treated with nutrient solution. At the two-leaf stage, Cd stress experiments were performed. The CdCl₂ solutions (20 μmol/L) prepared with nutrient solution were used to simulate Cd stress, and nutrient solution without CdCl₂ was used as the control. After 3h and 6h of treatment, seedlings were selected for each sample, and quickly stored at -80°C until analysis. The experiment was performed in triplicate.

In order to identify and characterize KCS gene family in rice genome, gff3, proteins, CDS, and genome files were downloaded from Ensembl Plants (http://plants.ensembl.org/Oryza_sativa/Info/Index). Twenty-one identified KCS protein sequences of Arabidopsis thaliana were downloaded from Arabidopsis genome database TAIR (https://www.arabidopsis.org/). We performed two methods, which are Basic Local Alignment Search Tool for proteins (BLASTp) and Hidden Markov Models (HMMER) search tool, to identify KCS genes in rice genome. The BLASTp (BLAST 2.7.1+) was performed based on protein homologous alignment with default mode using the Arabidopsis KCS protein sequences to obtain the candidate KCS genes in rice genome (E-value < 10⁻¹⁰). The Hidden Markov Model files corresponding to the conserved domain (Accession No.: PF08392 and PF08541) were downloaded from database Pfam (http://pfam.xfam.org/). HMMSEARCH was used to retrieve rice candidate sequences (E-value < 10⁻²⁰) containing KCS conserved domains (FAE1_CUT1_RppA and ACP_syn_III_C). Results of these two methods were checked and merged, and the redundancies were manually removed to obtain the candidate KCS genes in rice. The candidate sequences were submitted into the NCBI Conserved Domains (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) and SMART (http://smart.embl-heidelberg.de/) to confirm the domains in rice KCS proteins (Letunic et al., 2012).

The OsKCS genes’ physical and chemical properties, namely, molecular weight (M.W.), amino acid (aa) length, isoelectric point (pl), instability index, aliphatic index (Ai), and grand average of hydropathicity (GRAVY) were evaluated by using the ExPASY-Prot (http://web.expasy.org/protparam/). The phylogenetic analysis was performed by aligning the KCS protein sequences of rice by MEGA software. The aligned sequences were subjected to neighbor-joining (NJ) tree construction using the MEGA software with 1,000 bootstrap replications and all other parameters were set to default (Hall, 2013).

The chromosome locations and structures of OsKCS genes were retrieved from rice genome annotation files, and the conserved motifs of OsKCS protein sequences were predicted by using MEME (MEME 5.1.0) (Bailey et al., 2009). The OsKCS gene structures, chromosome locations, and conserved motifs were visualized by TBtools software (Bailey et al., 2009).

The 2,000 bp genomic sequences upstream of the transcription start site of OsKCS genes were extracted from the genomic DNA sequences. Since the upstream regions of some genes overlap with other genes, the upstream regions of these genes were shortened. The promoter sequences were submitted to the PlantCARE database (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/) to predict cis-regulatory elements (Lescot et al., 2002).

The tandem and segmental duplication or whole-genome duplication (WGD) provides new insights into genes development and genome progression. The duplication events of OsKCS genes and the syntenic relationships of KCS genes between rice and maize were analyzed using MCScanX toolkit (Wang et al., 2012), and KCS relationships between the target species were visualized by using Circos (Krzywinski et al., 2009). Nonsynonymous (Ka) and synonymous (Ks) values and the Ka/Ks ratios of gene pairs were calculated by ParaAT 2.0 (Zhang et al., 2012) and KaKs_Calculator (Zhang et al., 2006), Ks value could be used as molecular clock to reckon the time after gene replication event, Ka/Ks ratio has been used to determine the type of gene selection during evolution. Ka/Ks = 1, Ka/Ks > 1, and Ka/Ks < 1 represent natural, positive, and purifying selections, respectively (Hurst, 2002). The divergence time was calculated with the formula: T = Ks/r; r = 6.5 × 10⁻⁹) (Quraishi et al., 2011).

To examine the expression pattern of KCS genes under Cd stress, three RNA samples of Nipponbare and 9311 were sequenced on the HiSeq 4000 platform (Illumina) for transcriptome analysis by Novogene Technology (Beijing, China) to obtain clean reads. Differentially Expressed Genes (DEGs) were identified by a false discovery rate ≤ 0.05 and an absolute value of the log₂ ratio ≥ 1. Using the RNA-seq data, the absolute FPKM of the OsKCS were obtained, and the R software was used for statistics and visualization. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were performed by R packages: clusterProfiler.

Total RNA was extracted from different plant tissues using Trizol reagents (Invitrogen, Carlsbad, CA, USA), RNA was reverse transcribed to cDNA using the ReverTra Ace qPCR-RT kit (Toyobo, Osaka, Japan), and qPCR was performed using SYBR Green Real-Time PCR Master Mix (Toyobo).

Twenty-two KCS genes in rice genome were identified after removing redundant and repetitive sequences from BLASTp and HMMER results. To explore the phylogenetic relationships of the KCS family, an NJ phylogenetic tree was constructed using the full-length protein sequences of 22 OsKCSs and 21 AtKCSs. Results demonstrated that KCS proteins were classified into eight clades based on phylogenetic relationship, whereas, KCS proteins in rice were divided into six subclasses: α, γ, δ, ϵ, ζ and η ( Figure 1 ). KCS genes in rice genome were named OsKCS1 to OsKCS22 according to their chromosomal locations. GO and KEGG annotation analysis of the OsKCS genes was performed to further understand the possible roles of OsKCS genes in molecular function, cellular component, and biological process at the molecular levels. GO and KEGG enrichment analysis showed that 22 OsKCS genes were all enriched in fatty acid biosynthetic process and involved in fatty acid elongation ( Supplementary Table S1 ). Furthermore, the results of physiochemical properties showed that OsKCS genes varied in their properties such as the protein length varied from 271 (OsKCS5) to 542 aa (OsKCS7), as well as the molecular weights ranged from 29.23 to 60.11 kD; the isoelectric point (pI) of OsKCS proteins also varied ranged from 7.67 (OsKCS10) to 9.81 (OsKCS15), and the protein instability indexes of 11 OsKCS proteins were smaller than 40, indicating that these proteins are stable. The remaining OsKCS proteins’ instability indexes were greater than 40; most KCS proteins are hydrophilic. The KCS genes with the highest homology to the OsKCS genes and E-values were obtained in Arabidopsis thaliana ( Table 1 ).

OsKCS proteins were classified into four main groups (KCS1-like, FAE1-like, CER6-like, and FDH-like) ( Figure 2A ). Most of the OsKCS gene family members contained 1–2 exons, of which 11 OsKCS genes did not contain introns ( Figure 2B ). With the exception of OsKCS5, which has only four motifs, most of the conserved motifs of the OsKCS gene family members have the same types, numbers, and orders ( Figure 2C ). CDD and SMART search tools for domains verification was used and found that OsKCS proteins contained two domains such as FAE1_CUT1_RppA [(PF08392) FAE1/Type III polyketide synthase-like protein domain] and ACP_syn_III_C [(PF08541) 3-Oxoacyl-acyl-carrier protein (ACP) synthase III C terminal domain], which were main conserved domains in KCS proteins ( Figure 2D ), almost all genes contain these two domains except for OsKCS12.

A total of 18 cis-regulatory elements were predicted in the upstream 2,000 bp from the transcription start sites of OsKCSs ( Figure 3 ), which were widely involved in the growth biological process, hormonal responsiveness, light responsiveness, and stress response. A list of genomic location and names/annotations of 22 OsKCSs and the respective upstream genes will be provided to know whether any of the KCS-upstream gene blocks are paralogous ( Supplementary Table S2 ). Among them, MYB transcription factors involved in plant biological process were identified in all OsKCSs. Four types of hormone-responsive elements were also found, namely, auxin-responsive elements (TGA-elements) in nine OsKCSs, MeJA-responsive elements (CGTCA-motif and TGACG-motif) in 20 OsKCSs, salicylic acid–responsive elements (TCA-element) in seven OsKCSs, and ABA-responsive elements (ABRE) in 20 OsKCSs. Furthermore, there were meristem expression-related elements (CAT-box) in seven OsKCSs and gliadin metabolism-related elements in seven OsKCSs. The results of cis-regulatory element analysis indicate that OsKCS may be expressed in different growth environments, hormones, and stress treatments. However, many motifs have not been functionally verified, and whether these motifs confer unique functions on OsKCS remains to be further studied.

Except for chromosomes 7, 8, and 12, all of the OsKCS genes were unevenly distributed on nine of 12 chromosomes ( Figure 4A ). The largest number of OsKCS genes (OsKCS5–OsKCS11) appeared on chromosome 3, followed by chromosomes 2 and 6 (three genes each), chromosomes 5, 9, and 10 (two genes each). Synteny analysis was used to understand the evolution and expansion mechanism of OsKCS gene family in the rice genome and the genomes of other species. The results of gene duplication analysis indicated that there were only two OsKCS gene pairs (OsKCS2/OsKCS17 and OsKCS3/OsKCS15), both were segmentally duplicated on chromosomes 2 and 6 ( Figure 4B ). There were no tandem duplications in tight regions on chromosomes 3 and 6. Ka/Ks value is used to evaluate the evolution of coding sequences and determine the type of selection pressure after duplication. The Ka/Ks values of the two gene pairs were smaller than 0.05, indicating that these genes had gone through purifying selection ( Supplementary Table S3 ). The results of the divergence time indicated that the duplication process between the segmental OsKCS genes was estimated to be 6 million years ago, and the evolutionary mechanism of OsKCS was conserved during evolution. In order to further understand the evolutionary origins and orthologous relationship of KCS gene family, the synteny analysis was performed among four representative plant species (two dicots: Arabidopsis and Glycine max; two monocots: Zea mays and Triticum aestivum) ( Figure 5 ). Sixteen and 40 orthologous pairs were found between Zea mays and Triticum aestivum, respectively. Two and 0 orthologous pairs were found between Glycine max and Arabidopsis, respectively. The huge differences in homologous gene pairs between dicots and monocots suggested that KCS genes may be formed after the differentiation of monocots and dicots.

To investigate the rice KCS genes expression patterns in different tissues of rice plants, we analyzed rice transcript expression (RNA-seq data) in four different tissues; this included the expression in the root, leaf, panicle, and mature seed ( Supplementary Figure S1 ). In order to further investigate the functions of OsKCS genes involved in different developing stages, qRT-PCR was performed on all KCS gene members ( Figure 6 ). These results indicated that the predictions of expression profiles of most KCS genes were consistent with the qPCR data. KCS genes were specifically expressed in leaves (two genes), stem (six genes), panicle (three genes), and seeds (11 genes), respectively, and most of the KCS genes (OsKCS4, OsKCS6, OsKCS9, OsKCS13, OsKCS16, OsKCS17, and OsKCS22) expressed in seeds were highly expressed in the early stage of growth and development, suggesting that these OsKCS genes played distinct roles during the development of grain.

Based on the analysis of cis-elements in the promoter of OsKCS genes, all of these genes were hypothesized to respond to stress. Cadmium (Cd) is considered as one of the most toxic metals for plant. In order to analyze OsKCS involved in the response to Cd stress, we analyzed the RNA-seq data to evaluate the response of 22 OsKCS to Cd treatment in indica and japonica. All OsKCSs were differently expressed under Cd stress. In japonica Nipponbare background, after 3h treatment, the expression of 12 genes was significantly down-regulated, and the expression of these genes was different between 3h and 6h treatment ( Figure 7 ). For example, OsKCS8, OsKCS13, OsKCS14, and OsKCS17 were up-regulated after 3h treatment, the expression of these genes decreased gradually in 6h but was still higher than 0h (control). In indica 9311 background, 15 genes were up-regulated and seven genes were down-regulated continuously with the extension of treatment, seven genes were down-regulated and two genes were up-regulated under 6h of Cd treatment, 13 genes were up-regulated under 3h of Cd treatment. Whether in indica or japonica, OsKCS2, OsKCS5, OsKCS6, OsKCS8, OsKCS13, OsKCS17, OsKCS18, and OsKCS21 were up-regulated gradually with the extension of treatment time. The expression of OsKCS19 and OsKCS20 was opposite in indica and japonica after Cd treatment. The results showed that most OsKCSs were sensitive to the Cd.