Genome-Wide Identification and Posttranscriptional Regulation Analyses Elucidate Roles of Key Argonautes and Their miRNA Triggers in Regulating Complex Yield Traits in Rapeseed

Argonautes (AGOs) interact with microRNAs (miRNAs) to form the RNA-induced silencing complex (RISC), which can posttranscriptionally regulate the expression of targeted genes. To date, however, the AGOs and their miRNA triggers remain elusive in rapeseed (Brassica napus). Here, we systematically performed a phylogenetic analysis and examined the collinear relationships of the AGOs among four Brassicaceae species. Their physicochemical properties, gene structures, and expression patterns among 81 tissues from multiple materials and developmental stages were further analyzed. Additionally, their posttranscriptional regulation was analyzed using psRNATarget prediction, miRNA-/mRNA-Seq analyses, and a qRT-PCR verification. We finally identified 10 AtAGOs, 13 BolAGOs, 11 BraAGOs, and 24 BnaAGOs. An expression analysis of the BnaAGOs in the B. napus cultivar ZS11, as well as genotypes with extreme phenotypes in various yield-related traits, revealed the conservation and diversity of these genes. Furthermore, we speculated the posttranscriptional regulation of the B. napus miR168a–AGO1s and miR403–AGO2s modules. Combining miRNA-Seq and mRNA-Seq analyses, we found that the B. napus miR168a–AGO1s module may play an essential role in negatively regulating yield traits, whereas the miR403–AGO2s module positively impacts yield. This is the first attempt to comprehensively analyze the AGOs and their miRNA triggers in B. napus and provides a theoretical basis for breeding high-yielding varieties through the manipulation of the miRNA–AGOs modules.


Introduction
MicroRNAs (miRNAs) are multifunctional and endogenous small nucleotide molecules, which are just 20-24 nucleotides in length but play crucial roles in the posttranscriptional regulation of the expression of target genes [1,2]. They are pivotal components of the complex network-regulating plant development, stress tolerance, and hormone signaling [3,4]. The properties of one miRNA can regulate multiple paralogous genes simultaneously, making them a research hotspot for complex quantitative traits. Many previous studies have elucidated that miRNAs induce pleiotropic effects in the regulation of complex yieldrelated traits in crops; for instance, functional analyses of transgenes in rice (Oryza sativa L.) revealed that miR396 can increase grain yields by enhancing grain size and panicle branching, and simultaneously altering the plant architecture [5,6]. The elevated expression of miR397 also promotes panicle branching and grain enlargement by targeting OsLAC, leading to an increase in grain yield of up to 25% [7]. Similarly, the overexpression of miR408 enhanced the grain yield in rice by increasing the grain number and panicle branches [8]. Many other reports also showed that miRNAs can simultaneously alter multiple traits by targeting different paralogous genes [9][10][11]. In order to execute their functions, all A syntenic analysis was performed to further determine the evolutionary relationship of the AGO family among the four species. A total of 77 homologous pairs were identified among the four species (Figure 2A, Table S2), with significantly more homologous pairs found in the A subgenome than that in the C subgenome ( Figure 2B,C). All of the AGO copies in B. napus corresponded to those of the progenitors B. oleracea and B. rapa, A syntenic analysis was performed to further determine the evolutionary relationship of the AGO family among the four species. A total of 77 homologous pairs were identified among the four species (Figure 2A, Table S2), with significantly more homologous pairs found in the A subgenome than that in the C subgenome ( Figure 2B,C). All of the AGO copies in B. napus corresponded to those of the progenitors B. oleracea and B. rapa, with the exceptions that AGO9 was missing in B. napus and AGO4 was not present in the progenitors ( Table 1). The 24 BnaAGOs were distributed across the 16 chromosomes, with the exception of chromosomes A01, C01, and C09 ( Figure 2D).

Protein Profiles, Gene Structure and Conserved Motif Analysis of the 24 BnaAGOs
To gain more insight into the structural evolution of the 24 BnaAGOs, we analyzed their physicochemical properties. Their gDNA sizes were highly similar within subgroups; the AGO1 subgroup possessed the largest gDNA lengths, from 4921 (BnaAGO10-1) to 6740 bp (BnaAGO1-4), followed by the AGO4 subgroup, while the ZIPPY subgroup genes were remarkable smaller: 3169 (BnaAGO2-3) to 3712 bp (BnaAGO2-4) ( Figure 3B, Table 2). Similarly, BnaAGO proteins varied from 867 (BnaAGO6-1 and BnaAGO6-2) to 1086 aa (BnaAGO1-4). The isoelectric points (pI) of the 24 BnaAGOs were all higher than 8.82, indicating that AGOs are basic proteins ( Table 2). The conserved motifs were highly similar among the 24 BnaAGOs ( Figure 3A). The intron number and gene structure were conserved across each AGO subfamily but varied significantly among members of different subfamilies; for instance, the AGO1, AGO4, and MEL1 subgroups possess similar gene structures and intron numbers ranging from 16 to 23, whereas the ZIPPY subgroup members contain only one to five introns ( Figure 3B, Table 2). The protein characteristics of members of the same subgroup were highly similar, but differed between subfamilies, suggesting that the AGO genes underwent frequent gene duplication and recombination during evolution.

Differential Expression of 24 BnaAGOs among Differently yielding B. Napus Cultivars
We downloaded RNA-Seq data from 64 tissues in the B. napus cultivar ZS11 [41] to analyze the expression pattern of the 24 BnaAGOs. As shown by the heatmap, the 24 BnaAGOs have their own specific expression patterns ( Figure 4); for example, most genes from the AGO6, AGO7, and AGO9 subfamilies were hardly expressed in any of the tissues. The BnaAGO4s exhibited the highest expression level on average, with all four members being expressed most strongly in the pistil and seed. Genes belonging to the AGO2 and AGO5 subfamilies had similar expression profiles, with little to no expression in the embryo. All four BnaAGO1s were expressed in all 64 tissues. For some subfamilies, the members displayed varying expression patterns; for example, BnaAGO9-3 was much more strongly expressed than its two paralogs, BnaAGO9-1 and BnaAGO9-2. In addition, three

Differential Expression of 24 BnaAGOs among Differently Yielding B. napus Cultivars
We downloaded RNA-Seq data from 64 tissues in the B. napus cultivar ZS11 [41] to analyze the expression pattern of the 24 BnaAGOs. As shown by the heatmap, the 24 BnaAGOs have their own specific expression patterns ( Figure 4); for example, most genes from the AGO6, AGO7, and AGO9 subfamilies were hardly expressed in any of the tissues. The BnaAGO4s exhibited the highest expression level on average, with all four members being expressed most strongly in the pistil and seed. Genes belonging to the AGO2 and AGO5 subfamilies had similar expression profiles, with little to no expression in the embryo. All four BnaAGO1s were expressed in all 64 tissues. For some subfamilies, the members displayed varying expression patterns; for example, BnaAGO9-3 was much more strongly expressed than its two paralogs, BnaAGO9-1 and BnaAGO9-2. In addition, three genes (BnaAGO2-1, BnaAGO2-3, and BnaAGO2-4) from the AGO2 subfamily had similar expression profiles, but BnaAGO2-2 was hardly expressed in any of the 64 tissues ( Figure 4).

Differential Expression of 24 BnaAGOs among Differently yielding B. Napus Cultivars
We downloaded RNA-Seq data from 64 tissues in the B. napus cultivar ZS11 [41] to analyze the expression pattern of the 24 BnaAGOs. As shown by the heatmap, the 24 BnaAGOs have their own specific expression patterns ( Figure 4); for example, most genes from the AGO6, AGO7, and AGO9 subfamilies were hardly expressed in any of the tissues. The BnaAGO4s exhibited the highest expression level on average, with all four members being expressed most strongly in the pistil and seed. Genes belonging to the AGO2 and AGO5 subfamilies had similar expression profiles, with little to no expression in the embryo. All four BnaAGO1s were expressed in all 64 tissues. For some subfamilies, the members displayed varying expression patterns; for example, BnaAGO9-3 was much more strongly expressed than its two paralogs, BnaAGO9-1 and BnaAGO9-2. In addition, three genes (BnaAGO2-1, BnaAGO2-3, and BnaAGO2-4) from the AGO2 subfamily had similar expression profiles, but BnaAGO2-2 was hardly expressed in any of the 64 tissues ( Figure  4). The AGOs play an important role in the posttranscriptional regulatory functions of all miRNAs, which are multifunctional factors regulating complex quantitative traits. We therefore used RNA-Seq data for the 24 BnaAGOs in materials with extremely high and low harvest indexes (HI; a quantitative yield-related trait) to further investigate whether AGOs were associated with the complex yield traits in B. napus. The RNA-Seq data from the 11 HI tissues, partially published in our previous study [42], show high expression levels only for the BnaAGO1s, BnaAGO2s, BnaAGO4s, and BnaAGO5s, while the other BnaAGOs were hardly expressed in almost any of the HI materials ( Figure 5A). Coincidently, only the AGO1, AGO2, AGO4, and AGO5 subfamilies exhibited higher expression levels in three pairs of yield-related materials, including plants with extreme phenotypes for TSW ( Figure 5B), seed number per silique (SPS; Figure 5C), and initial embryotic number (IEN; Figure 5D) (all those were unpublished data, Table S3). the 11 HI tissues, partially published in our previous study [42], show high expression levels only for the BnaAGO1s, BnaAGO2s, BnaAGO4s, and BnaAGO5s, while the other BnaAGOs were hardly expressed in almost any of the HI materials ( Figure 5A). Coincidently, only the AGO1, AGO2, AGO4, and AGO5 subfamilies exhibited higher expression levels in three pairs of yield-related materials, including plants with extreme phenotypes for TSW ( Figure 5B), seed number per silique (SPS; Figure 5C), and initial embryotic number (IEN; Figure 5D) (all those were unpublished data, Table S3). Finally, among the more highly expressed AGO1, AGO2, AGO4, and AGO5 subfamily members, we found that the BnaAGO1s and BnaAGO4s were most highly expressed in the high-HI materials, especially in the pericarp and stem tissues ( Figure 5A; indicated with red oval). On the contrary, the BnaAGO2s were significantly more highly expressed in the low-HI materials ( Figure 5A; indicated with black arrows). Only BnaAGO1s (negative) and BnaAGO2s (positive) were correlated with TSW between materials with high and low values in this characteristic ( Figure 5B). None of the genes showed a significant difference in expression between the materials with different SPS and IEN extremes ( Figure 5C,D). We therefore speculate that the AGO1, AGO2, and AGO4 subfamilies may be associated with the complex yield traits, with the BnaAGO1s and BnaAGO4s potentially making a positive contribution to yield traits and the BnaAGO2s playing the opposite function, especially with the HI and TSW traits.

MiRNA-Mediated Posttranscriptional Regulation of the BnaAGOs
To determine whether BnaAGOs are regulated by B. napus miRNAs, the psRNATarget online prediction tool was used to search the 24 full-length BnaAGO sequences against the known targets of plants miRNAs [43], using a maximum expectation of 3.5. Only B. napus miR168a and miR403 were found to have near-perfect complementarity to seven BnaAGOs: all the four BnaAGO1s (BnaAGO1-1, BnaAGO1-2, BnaAGO1-3, and BnaAGO1-4) and three BnaAGO2s (BnaAGO2-1, BnaAGO2-3, and BnaAGO2-4, but not BnaAGO2-2), which are targeted by miR168a and miR403, respectively ( Figure S1). The complementarity scores for B. napus miR168a-AGO1s and miR403-BnaAGO2s were 3 and 0, respectively, with the cleavage mode ( Table 3). The miR168 and miR403 sequences in B. napus, B. oleracea, B. rapa, and A. thaliana were downloaded from miRBase to illuminate their evolutionary relationships, and the phylogenetic tree was constructed with MEGA 7.0 ( Figure S2). We further downloaded the regulatory networks of miR168 and miR403 in the four species (https://www.mirbase.org/, accessed on 20 October 2020), and found that all seven predicted miR168 members in the four species target AGO1 genes and have known functions in the regulation of other miRNAs, whereas no related regulatory networks were available for miR403 ( Figure S2).  To clarify whether B. napus miR168a and miR403 can negatively regulate BnaAGO1s and BnaAGO2s in B. napus, we performed a qRT-PCR analysis of their expression patterns in eight tissues of the cultivar ZS11. We found that miR168a was expressed to a significantly higher level in the pericarp and the seed at 15 days after flowering (15P and 15S) than the other six tissues (the bud, flower, and the pericarp and seed at 25 or 35 days after flowering; pericarp: 25P and 35P; seed: 25S and 35S, respectively) ( Figure 6A). Coincidentally, the four BnaAGO1s are expressed at low levels in the 15P and 15S tissues. Negative expression patterns between B. napus miR168a and the BnaAGO1s can also be observed in the other six tissues. For miR403 and the BnaAGO2s, this phenomenon was even more pronounced; miR403 was only expressed in 25S and 35S, in which their three predicted targets were hardly expressed ( Figure 6B). On the contrary, the three BnaAGO2s (especially BnaAGO2-3 and BnaAGO2-4) exhibited higher expression levels in the Bu, Fl, 15S, 15P, 25P, and 35P tissues, in which miR403 expression was almost non-existent. The negative expression relationships of the miR168a-AGO1s and miR403-AGO2s were most obvious in the 25P, 25S, 35P and 35S tissues ( Figure 6). Taken together, our findings indicate that miR168a and miR403 mediate the negative posttranscriptional regulation of the BnaAGO1s and BnaAGO2s, respectively. dicted targets were hardly expressed ( Figure 6B). On the contrary, the three BnaAGO2s (especially BnaAGO2-3 and BnaAGO2-4) exhibited higher expression levels in the Bu, Fl, 15S, 15P, 25P, and 35P tissues, in which miR403 expression was almost non-existent. The negative expression relationships of the miR168a-AGO1s and miR403-AGO2s were most obvious in the 25P, 25S, 35P and 35S tissues ( Figure 6). Taken together, our findings indicate that miR168a and miR403 mediate the negative posttranscriptional regulation of the BnaAGO1s and BnaAGO2s, respectively.

miRNA-Seq and mRNA-Seq Analysis of miR168a-AGO1s and miR403-AGO2s in Differentially Yielding B. napus Materials
To determine whether the B. napus miR168a-AGO1s and miR403-AGO2s modules are associated with yield-related traits in B. napus, we performed a combined miRNA-Seq and mRNA-Seq analysis to explore their expression profiles in multiple genotypes with differences in yield-related traits, including TSW, SPS, and IEN (unpublished data, Table  S4). We revealed a spatiotemporal expression specificity for both miR168a ( Figure 7A) and miR403 ( Figure 7B) between the materials with extremely high and low TSW; miR168a was hardly expressed in the high-TSW materials but more strongly expressed in the low-TSW materials, while the four BnaAGO1s showed the opposite expression pattern ( Figure 7A). By contrast, miR403 was more highly expressed in the high-TSW materials, while the expression levels of the BnaAGO2s showed the opposite pattern ( Figure 7B). The expression levels of miR168a and miR403 showed no significant difference in the cultivars with differing SPS and IEN traits, resulting in no significant differences in the expression of the BnaAGO1s and BnaAGO2s either ( Figure 7C-F).  To determine whether the B. napus miR168a-AGO1s and miR403-AGO2s modules are associated with yield-related traits in B. napus, we performed a combined miRNA-Seq and mRNA-Seq analysis to explore their expression profiles in multiple genotypes with differences in yield-related traits, including TSW, SPS, and IEN (unpublished data, Table S4). We revealed a spatiotemporal expression specificity for both miR168a ( Figure 7A) and miR403 ( Figure 7B) between the materials with extremely high and low TSW; miR168a was hardly expressed in the high-TSW materials but more strongly expressed in the low-TSW materials, while the four BnaAGO1s showed the opposite expression pattern ( Figure 7A). By contrast, miR403 was more highly expressed in the high-TSW materials, while the expression levels of the BnaAGO2s showed the opposite pattern ( Figure 7B). The expression levels of miR168a and miR403 showed no significant difference in the cultivars with differing SPS and IEN traits, resulting in no significant differences in the expression of the BnaAGO1s and BnaAGO2s either ( Figure 7C-F).
Combining the results of our RNA-Seq analysis of the 24 BnaAGOs in ZS11 ( Figure 4); the expression analysis in the HI, TSW, SPS, and IEN materials ( Figure 5); and the miRNA-Seq and mRNA-Seq analyses of the miR168a-AGO1s and miR403-AGO2s, we can draw two main conclusions. On the one hand, the RNA-Seq data show that genes from the AGO1, AGO2, AGO4, and AGO5 subgroups have specific higher expressions in certain genotypes and developmental stages (Figures 4 and 5), and that the BnaAGO1s, BnaAGO2s, and BnaAGO4s may be associated with complex yield traits in B. napus, particularly TSW and HI ( Figure 5). On the other hand, the joint miRNA-Seq and mRNA-Seq analyses indicate that miR168a may negatively impact yield traits by the posttranscriptional regulation of the BnaAGO1s, whereas the miR403-AGO2s module plays a positive role on the complex yield traits, especially TSW (Figure 7). The negative expression patterns observed between the miRNAs and mRNAs also suggest that both miR168a-AGO1s and miR403-AGO2s are posttranscriptionally regulated, as indicated by the qRT-PCR verification results in ZS11 ( Figure 6).  Supplement Table S3  and Table S4, respectively.
Combining the results of our RNA-Seq analysis of the 24 BnaAGOs in ZS11 ( Figure  4); the expression analysis in the HI, TSW, SPS, and IEN materials ( Figure 5); and the miRNA-Seq and mRNA-Seq analyses of the miR168a-AGO1s and miR403-AGO2s, we can draw two main conclusions. On the one hand, the RNA-Seq data show that genes from the AGO1, AGO2, AGO4, and AGO5 subgroups have specific higher expressions in certain genotypes and developmental stages (Figures 4 and 5), and that the BnaAGO1s, BnaAGO2s, and BnaAGO4s may be associated with complex yield traits in B. napus, particularly TSW and HI ( Figure 5). On the other hand, the joint miRNA-Seq and mRNA-Seq analyses indicate that miR168a may negatively impact yield traits by the posttranscriptional regulation of the BnaAGO1s, whereas the miR403-AGO2s module plays a positive role on the complex yield traits, especially TSW (Figure 7). The negative expression patterns observed between the miRNAs and mRNAs also suggest that both miR168a-AGO1s and miR403-AGO2s are posttranscriptionally regulated, as indicated by the qRT-PCR verification results in ZS11 ( Figure 6).

Genome-Wide Identification, Phylogenetic Analysis of AGOs, and Their miRNA Triggers
Rapeseed (B. napus, AACC, 2n = 38) is an important oil crop that evolved from a heterogeneous hybridization of two diploid species, B. oleracea (n = 9) and B. rapa (n = 10).

Genome-Wide Identification, Phylogenetic Analysis of AGOs, and Their miRNA Triggers
Rapeseed (B. napus, AACC, 2n = 38) is an important oil crop that evolved from a heterogeneous hybridization of two diploid species, B. oleracea (n = 9) and B. rapa (n = 10). With the publication of the whole B. napus genome sequence in 2014 [44], the evolution and origin of rapeseed, as well as the genome-wide identification of key genes, have become research hotspots in recent years. Argonautes (AGOs) are important proteins that can interact with miRNAs to form RISCs, which play essential roles in regulating gene silencing [45]. Abundant AGO genes were identified in many plants [15][16][17][18][19]46]. The BnaAGO genes were predicted in a previous study [47]; however, the miRNA-AGO complexes have not yet been characterized. Here, we identified 24 BnaAGOs, as well as 13 BolAGOs and 11 BraAGOs in the progenitor species (Figure 1, Table S1). The AGOs in B. napus corresponded to orthologs in the progenitors B. oleracea and B. rapa, except that B. napus had one less AGO9 and one extra AGO4. By comparison, the AGO3 and AGO8 subfamily was smaller in both B. rapa and B. oleracea than in Arabidopsis (Figure 1, Tables 1 and S1). Taken together, the results of the AGO homolog prediction, phylogenetic analysis, and copy number analysis of the 24 identified BnaAGOs are consistent with the complex genome of B. napus as an allotetraploid.
As mentioned above, many redundant genes have been generated during the genomic evolution of the allotetraploid B. napus. The yield of this crop is a typical and complex quantitative trait [33], the improvement of which would, therefore, require altering the expression of several genes. Individual miRNAs have the ability to target multiple genes simultaneously, making them increasingly common targets for the research of complex quantitative traits. The AGO genes are important components of RISCs; thus, we used the psRNATarget prediction to search the targets of known plant miRNAs [43] to further determine whether BnaAGOs are regulated by B. napus miRNAs. Only B. napus miR168a and miR403 were found to have near-perfect complementarity to BnaAGO1s and BnaAGO2s, respectively ( Figure S1, Table 3). Through a phylogenetic analysis, we further identified seven miR168 orthologs (ath-miR168a, bra-miR168a, bna-miR168a, ath-miR168b, bra-miR168b, bna-miR168b, and bra-miR168c) and three miR403 orthologs (ath-miR403, bra-miR403, and bna-miR403) among the four Brassicaceae species ( Figure S2). As far more AGOs than corresponding miRNAs were identified in the four species, we can, therefore, conclude that the miRNAs are more evolutionarily conserved [48,49] than genes such as the AGOs.

Diversity and Conservation of the 24 BnaAGOs
The functions of genes are largely determined by their structures [50]. Our comprehensive analyses of the gene structures and physicochemical profiles of the 24 BnaAGOs also suggest that AGO genes within a subfamily are conserved and are much more similar to each other than to members of other AGO subfamilies. The ZIPPY subgroups comprised AGO genes with a remarkably smaller gDNA length than those of the other subgroups ( Figure 3B, Table 2). Consistently, the ZIPPY subgroup contains only one to five introns, whereas members of the AGO1, AGO4, and MEL1 subgroups possess similar gene structure and intron numbers, ranging from 16 to 23 ( Figure 3B, Table 2). Furthermore, the gDNA length and intron numbers of the ZIPPY subgroup were also lower than those of homologous genes in A. thaliana and Salvia miltiorrhiza [28]. These findings suggest that gene structures are conserved within the AGO subfamilies, even among different species. On the other hand, some diversity was observed even within subgroups; for example, the structure of BnaAGO9-3 differed from that of its two paralogs ( Figure 3B). Similarly, BnaAGO2-2 has lost the untranslated region (UTR) present in its three paralogs, resulting in a non-target relationship between B. napus miR403 and BnaAGO2-2 ( Figure S1).
The expression patterns also exhibited a similar phenomenon; for instance, genes from the AGO6, AGO7, and AGO9 subfamilies were hardly expressed among the 82 tissues studied in total (ZS11, HI, TSW, SPS, and IEN), whereas the BnaAGO1s, BnaAGO2s, BnaAGO4s, and BnaAGO5s exhibited significantly higher expression patterns among all the materials (Figures 4 and 5). Contrasting patterns were also observed: BnaAGO9-3 displayed spatiotemporal expression specificity in the materials, but its paralogs BnaAGO9-1 and BnaAGO9-2 were hardly expressed in any of the tissues (Figures 4 and 5). Similarly, BnaAGO2-2 was hardly expressed in the 64 ZS11 tissues examined, whereas the other three BnaAGO2s showed different degrees of spatiotemporal expression specificity (Figure 4). These expression differences may arise from the different gene structures mentioned above ( Figure 3B), as suggested previously [50]. In conclusion, the protein characters, gene structures, and expression patterns were highly conserved within the same subfamily while varying significantly across the members of different subfamilies, suggesting that the AGO genes underwent frequent gene duplication and recombination during their evolution, which resulted in their functional diversity.

Posttranscriptional Regulation by the B. napus miRNA-AGOs and the Future Regulation of Yield-Related Traits
MicroRNAs are posttranscriptional regulators of gene expression that bind to their specific target mRNAs and load on the RISC to prevent their translation through target cleavage or translation inhibition. The formation of the RISC requires the miRNA to physically interact with AGOs [3]; however, the AGOs and their miRNA triggers remain elusive in B. napus. In our study, we predicted two B. napus miRNAs mediated by seven BnaAGOs through a psRNATarget prediction [43], including four BnaAGO1s and three BnaAGO2s targeted by mi168a and miR403, respectively ( Figure S1, Table 3). A further sequence analysis showed that miR403 perfectly complements the three BnaAGO2 target sites in their UTRs ( Figure S1). In the miR168a-AGO1s, a three-base difference was present in the four target sets ( Figure S1). Additionally, the results of the qRT-PCR analysis in eight ZS11 tissues further confirmed the negative expression patterns between miR168a-AGO1s and miR403-AGO2s ( Figure 6). Consistently, AGO1 and AGO2 were found to be regulated by miR168 and miR403, respectively, in various plants, including Arabidopsis [26], rice [27], tomato [51], and Salvia miltiorrhiza [28], indicating the conserved regulatory mechanism of the miR168-AGO1 and miR403-AGO2 modules in plants.
In rice, miR168-AGO1 module was not only proven to be helpful in regulating grain yield and flowering time, but also in changing plant immunity to Magnaporthe oryzae [27]. In addition, miR403 was shown to play an essential function in regulating tomato development by affecting the expression of SlAGO2 [51]. To determine whether miR168a-AGO1 and miR403-AGO2 were associated with complex yield-related traits in B. napus, we analyzed miRNA-Seq and mRNA-Seq data in multiple genotypes with extremely high and low TSW, SPS, and IEN values. We found that miR168a was more highly expressed in low-TSW plants ( Figure 7A), whereas miR403 showed higher expression levels in the high-TSW materials ( Figure 7B). Correspondingly, the BnaAGO1s ( Figure 7A) and BnaAGO2s ( Figure 7B) displayed the opposite expression patterns to their respective miRNA triggers, indicating that the miR168a-AGO1 module may plays essential negative roles in regulating the TSW trait in B. napus, while miR403-AGO2 module have the contrasting function. No significant patterns were detected in the expression patterns of miR168a, miR403, BnaAGO1s, and BnaAGO2s between the cultivars with high and low SPS or IEN values, however ( Figure 7C-F). Despite this, we cannot conclude that the B. napus miR168a-AGO1s and miR403-AGO2 modules were irrelevant to the SPS and IEN character, because the expression of miR168 and miR403 have also been proved to reduce the AGO1 and AGO2 protein levels [52]. Furthermore, the AGO2-miR168-AGO1-miR403 loop is known to be vulnerable and easily unbalanced, so even a slight change in this loop would be amplified continuously [51].

Genome-Wide Prediction of AGOs and Their miRNA Triggers
The 10 AtAGOs protein sequences were downloaded from The Arabidopsis Information Resource database (https://www.arabidopsis.org/index.jsp, accessed on 20 June 2020) and used to perform a BLAST analysis to identify the BnaAGOs, BolAGOs, and BraAGOs [53], using an E-value of 0 to identify exact homologs. All the AGO protein sequences identified in the four species were downloaded from the Brassicaceae database (http://brassicadb.cn, accessed on 20 June 2020). MEGA 7.0 was used to perform a multiple sequence alignment using default parameters, and the neighbor-joining method was used to construct the phylogenetic tree with a bootstrap analysis of 1000 replicates [54]. The syntenic analysis was performed using TBtools (https://github.com/CJ-Chen/TBtools, accessed on 16 September 2020) to determine the syntenic relationships of the AGO genes from B. napus and the three other species.

Chromosomal Location, Gene Structure and Protein Properties
The chromosomal distribution of the AGOs was visualized with Map-Chart2.2 [55]. The gene structures and conserved motifs of the BnaAGOs were analyzed using Gene Structure Display Server (GSDS) [56] and MEME online (http://meme-suite.org/tools/ meme, accessed on 20 June 2020) [57,58], respectively, with the maximum number of predicted motifs set to 15 and default settings used for the other parameters. To further clarify the protein properties of the BnaAGOs, the isoelectric points (PI) and molecular weights of the BnaAGOs were also predicted based on the ExPASy proteomics server database (https://www.expasy.org/tools/, accessed on 20 June 2020) [59].

Identification of B. napus miRNAs with Perfect Complementarity to the BnaAGOs
To select the key B. napus miRNAs that target the BnaAGOs, a psRNATarget search (https://www.zhaolab.org/psRNATarget/, accessed on 20 October 2020) was performed using the 24 full-length B. napus sequences [43], with the maximum expectation set to 3.5, as described previously [28]. The miRBase database (http://www.mirbase.org/, accessed on 20 October 2020) [60] was searched to identify miR168 and miR403 homologs in other species. All miRNA sequences, including the precursor and mature sequences, as well as the regulatory network, were also downloaded from miRBase (http://www.mirbase.org/, accessed on 20 October 2020).

Plant Materials
All seeds of the B. napus genotypes with extreme yield-related phenotypes (HI, TSW, SPS, and IEN) were obtained from the Chongqing Rapeseed Technology Research Center, China. All plants were planted in two experimental plots in Chongqing field conditions for at least two years, and ten plants were collected for the character investigation at the harvest stage. Throughout the plant growth, different tissues were obtained and frozen at -80 • C until required. The phenotypes of CQ46 (Ningyou 12, low HI) and CQ24 (SWU47, high HI) were published in our previous report [42]. The phenotypes and statistical data for LTSW (P147, low TSW), HTSW (P478, high TSW), LSPS (P026 and P578, low SPS), HSPS (P196 and P446, SPS), LIEN (P355 and P536, low IEN) and HIEN (P300 and P456, high IEN) are presented in Table 4 and Figure 8. The data analysis were performed by a one-way ANOVA method with significant differences (p < 0.01).

miRNA-Seq and mRNA-Seq Analysis among Multiple Yield-Related Materials
To examine the expression profiles among the 24 BnaAGOs, an analysis was performed on RNA-Seq data from a typical B. napus cultivar (ZS11), which can be downloaded from https://brassica.biodb.org/, accessed on 1 October 2022 [41]. Their expression levels in 64 tissues throughout plant growth were used to construct a heatmap. To explore whether the BnaAGOs were associated with the complex yield-related traits in B. napus, mRNA-Seq data were downloaded for a pair of genotypes with extremely high and low HI, which had been published in our previous study [42]. The expression patterns of the BnaAGOs were also explored in another three pairs of yield-related genotypes, with extreme phenotypes in the TSW, SPS, and IEN traits (Table S3, unpublished data). All heatmaps were constructed using TBtools (https://github.com/CJ-Chen/TBtools, accessed on 16 September 2020). To further clarify the relationships of the B. napus miR168a-AGO1s and miR403-AGO2s, and to explore whether the miRNA-AGO modules were associated with the complex yield-related traits in B. napus, a further joint analysis of the miRNA-Seq (Table S4, unpublished data) and mRNA-Seq data (Table S3, unpublished data) from the TSW, SPS, and IEN materials  was

miRNA-Seq and mRNA-Seq Analysis among Multiple Yield-Related Materials
To examine the expression profiles among the 24 BnaAGOs, an analysis was performed on RNA-Seq data from a typical B. napus cultivar (ZS11), which can be downloaded from https://brassica.biodb.org/, accessed on 1 October 2022 [41]. Their expression levels in 64 tissues throughout plant growth were used to construct a heatmap. To explore whether the BnaAGOs were associated with the complex yield-related traits in B. napus, mRNA-Seq data were downloaded for a pair of genotypes with extremely high and low HI, which had been published in our previous study [42]. The expression patterns of the BnaAGOs were also explored in another three pairs of yield-related genotypes, with extreme phenotypes in the TSW, SPS, and IEN traits (Table S3, unpublished data). All heatmaps were constructed using TBtools (https://github.com/CJ-Chen/TBtools, accessed on 16 September 2020). To further clarify the relationships of the B. napus miR168a-AGO1s and miR403-AGO2s, and to explore whether the miRNA-AGO modules were associated with the complex yield-related traits in B. napus, a further joint analysis of the miRNA-Seq (Table S4, unpublished data) and mRNA-Seq data (Table S3, unpublished data) from the TSW, SPS, and IEN materials was performed.  (SPS; B), and initial embryonic number (IEN; C) in B. napus. LHI and LTSW and HTSW indicate high-and low-TSW materials; LSPS and HSPS indicate highand low-SPS materials; LIEN and HIEN indicate high-and low-IEN materials; The p value were performed with one-way ANOVA and ** represent significant difference at p_Value < 0.01.

RNA Extraction and qRT-PCR Verification
An RNA extraction kit (Thermo Fisher Scientific, Waltham, MA, USA) was used to extract the total RNA from the specific tissues. A reverse transcription was performed to produce cDNA using a Reverse Transcription Kit (Takara Bio, Kusatsu, Japan). A Bio-Rad CFX system (Bio-Rad Laboratories, Hercules, CA, USA) was used to perform the qRT-PCR analysis in triplicate, and the gene BnaACTIN was selected as the endogenous reference for the expression analysis. The 2 −∆∆Ct method was selected for the calculation of the relative expression level of the BnaAGOs [61]. Finally, the qRT-PCR results were visualized using Graphpad Prism 5.0 [62]. To determine the expression levels of B. napus miR168a and miR403, the gene BnaU6 was selected as the endogenous reference. All the primers used in our study were designed with Primer 5.0 [63] and are listed in Table S5.

Conclusions
In this work, a comprehensive approach includes the phylogenetic analysis of the AGOs in A. thaliana, B. napus, B. rapa, and B. oleracea, the gene structures, conserved motifs, and expression patterns of the 24 BnaAGOs in many differently yielding genotypes. Our results reveal the diversity and conservation of the sequences and functions among these genes, as well as indicating the loss of some orthologs following the cross that resulted in B. napus. In addition, a posttranscriptional regulation analysis indicated that the miR168a-AGO1 and miR403-AGO2 regulatory modules previously reported in other species were conserved in B. napus and may play roles in regulating yield-related traits in this oilseed crop, especially in TSW. Our findings, therefore, provide a new understanding of the miR168-AGO1s and miR403-AGO2s modules of B. napus, facilitating future research into the complex yield traits.