Differential Gene Expression with an Emphasis on Floral Organ Size Differences in Natural and Synthetic Polyploids of Nicotiana tabacum (Solanaceae)

Abstract

Floral organ size, especially the size of the corolla, plays an important role in plant reproduction by facilitating pollination efficiency. Previous studies have outlined a hypothesized organ size pathway. However, the expression and function of many of the genes in the pathway have only been investigated in model diploid species; therefore, it is unknown how these genes interact in polyploid species. Although correlations between ploidy and cell size have been shown in many systems, it is unclear whether there is a difference in cell size between naturally occurring and synthetic polyploids. To address these questions comparing floral organ size and cell size across ploidy, we use natural and synthetic polyploids of Nicotiana tabacum (Solanaceae) as well as their known diploid progenitors. We employ a comparative transcriptomics approach to perform analyses of differential gene expression, focusing on candidate genes that may be involved in floral organ size, both across developmental stages and across accessions. We see differential expression of several known floral organ candidate genes including ARF2, BIG BROTHER, and GASA/GAST1. Results from linear models show that ploidy, cell width, and cell number positively influence corolla tube circumference; however, the effect of cell width varies by ploidy, and diploids have a significantly steeper slope than both natural and synthetic polyploids. These results demonstrate that polyploids have wider cells and that polyploidy significantly increases corolla tube circumference.

Keywords: BIG BROTHER; GASA; Nanopore; RNA-Seq; cell size; flower size.

Figure 1

Floral morphology of diploid progenitors N. sylvestris and N. tomentosiformis, two natural N. tabacum accessions (095-55 and ‘Chulumani’), and three synthetic N. tabacum lines (QM20, QM24, and QM25): all accessions pictured here are included in the cell size data set, whereas all accessions except QM20 were used in the transcriptome analyses. Sylv = N. sylvestris, tomf = N. tomentosiformis, tab095 = N. tabacum 095-55, tabChu = N. tabacum ‘Chulumani’, QM20 = synthetic N. tabacum QM20, QM24 = synthetic N. tabacum QM24, and QM25 = synthetic N. tabacum QM25.

Figure 2

Strip plots for (A) cell width from 100 cells, (B) cell number, and (C) corolla tube circumference for all accessions, including mean and standard deviation for each accession: red dotted lines represent the progenitor average, which is the expected intermediate phenotype. Predicted corolla tube circumference based on the linear model for (D) cell width and (E) cell number: points represent actual flower data. Corolla tube circumference based on (F) ploidy including mean and standard deviation for each ploidy: points represent actual flower data. Sylv = N. sylvestris, tomf = N. tomentosiformis, tab095 = N. tabacum 095-55, tabChu = N. tabacum ‘Chulumani’, QM20 = synthetic N. tabacum QM20, QM24 = synthetic N. tabacum QM24, QM25 = synthetic N. tabacum QM25.

Figure 3

Percent complete single copy, fragmented, and missing BUSCO genes for both N. sylvestris and N. tomentosiformis across the five different assemblies performed: including the Nanopore long-reads gave an increase of complete single copy genes and decrease in missing genes compared to short-read only, regardless of assembler.

Figure 4

Subclustering of transcripts across all taxa within the 60% of anthesis developmental stage based on hierarchical clustering of differentially expressed genes: the number of transcripts found in each cluster is given. Each gray line is one gene, while the blue line is the mean expression profile for that subcluster. The order along the x-axis for each subcluster plot is synthetic polyploids (QM24 and QM25), N. sylvestris (Sylv), natural polyploids (TAB095 and TABC), and N. tomentosiformis (Tom).

Figure 5

Log₂ fold change of differentially expressed genes across comparisons within accessions between 60%, 85%, and 95% developmental time points for (A) N. tabacum 095-55, (B) N. tabacum ‘Chulumani’, (C) synthetic N. tabacum QM24, and (D) synthetic N. tabacum QM25: violin plots represent all differentially expressed genes between each comparison, whereas colorful strip plots represent differentially expressed candidate floral organ size genes. Positive logFC values represent genes upregulated in the first stage listed in the comparison, whereas negative logFC values represent genes upregulated in the second stage listed. The dashed gray lines across the plot denote the log₂ = |2| cutoff for differentially expressed genes. Sylv = N. sylvestris, tomf = N. tomentosiformis.

Figure 6

Log₂ fold change of differentially expressed genes across comparisons between natural and synthetic N. tabacum accessions at (A) 60%, (B) 85%, and (C) 95%of anthesis length: violin plots represent all differentially expressed genes between each comparison, whereas colorful strip plots represent differentially expressed candidate floral organ size genes. Positive logFC values represent genes upregulated in the first accession listed in the comparison, whereas negative logFC values represent genes upregulated in the second accession listed. The dashed gray lines across the plot denote the log₂ = |2| cutoff for differentially expressed genes. Sylv = N. sylvestris, tomf = N. tomentosiformis, tab095 = N. tabacum 095-55, tabChu = N. tabacum ‘Chulumani’, QM24 = synthetic N. tabacum QM24, and QM25 = synthetic N. tabacum QM25.