Gene dosage compensation of rRNA transcript levels in Arabidopsis thaliana lines with reduced ribosomal gene copy number

Abstract

The 45S rRNA genes (rDNA) are among the largest repetitive elements in eukaryotic genomes. rDNA consists of tandem arrays of rRNA genes, many of which are transcriptionally silenced. Silent rDNA repeats may act as 'back-up' copies for ribosome biogenesis and have nuclear organization roles. Through Cas9-mediated genome editing in the Arabidopsis thaliana female gametophyte, we reduced 45S rDNA copy number (CN) to a plateau of ∼10%. Two independent lines had rDNA CNs reduced by up to 90% at the T7 generation, named low copy number (LCN) lines. Despite drastic reduction of rDNA copies, rRNA transcriptional rates, and steady-state levels remained the same as wild-type plants. Gene dosage compensation of rRNA transcript levels was associated with reduction of silencing histone marks at rDNA loci and altered Nucleolar Organiser Region 2 organization. Although overall genome integrity of LCN lines appears unaffected, a chromosome segmental duplication occurred in one of the lines. Transcriptome analysis of LCN seedlings identified several shared dysregulated genes and pathways in both independent lines. Cas9 genome editing of rRNA repeats to generate LCN lines provides a powerful technique to elucidate rDNA dosage compensation mechanisms and impacts of low rDNA CN on genome stability, development, and cellular processes.

Figure 1

Overview of 45S rDNA loci, dynamics of transcription and ribosome assembly, and mutagenesis concept. A, The 45S rRNA gene comprises two ETSs, two ITSs, and the 18S, 5.8S, and 25S (28S in fungi and animals) ribosomal components. A single gRNA (sgRNA) was designed specifically to guide Cas9 to the 18S locus and generate DSBs across the 45S repeats. B, Transcription of 45S rRNA occurs in the nucleolus, the largest structure in the nucleus. Assembly of ribosomes requires import of the 5S transcript from the nucleoplasm, as well as import of ribosomal proteins into the nucleolus. LSU = Large Subunit; SSU = Small Subunit. C, CRISPR-Cas9 mutagenesis: Random DSBs (gray) were generated along the 45S repeats targeting the 18S locus within the 45S rRNA gene. Two outcomes were expected: deletions of large numbers of repeats or insertion of supernumerary copies (blue) by the action of NHEJ DNA repair machinery. AC, Antipodal cells; V, Vacuole; CCN, Central cell nucleus; and SC, Synergid cells.

Figure 2

Selection of Low 45S CN lines, 3'-ETS variant analysis and FISH at T7 generation. A, 45S rDNA Relative CN of T1 transformant population quantified by qPCR against the single copy genes HXK1 and TZF1. Blue bar and green bar indicate lines #289 and #236, respectively. Highlighted are LCN lines on which all experiments were performed. B, 45S rDNA Relative CN by quantification by qPCR of T4 to T7 generations. Eleven individual plants of lines #236 and #289 were selected per generation and one WT Col-0 control. Bars represent standard error of technical qPCR replicates. C, Schematic representation of 3′-allelic variants, adapted from Pontvianne, 2010. Analysis of genomic and gene dosage of 45S rDNA variants in WT and LCN at T4 and T7 generations. Percentage (%) of 45S relative CN for each was calculated by qPCR using the same DNA sample as the RT-PCR. RT-PCR analysis of 3′-EST variant expression shows qualitative differences between WT and LCN lines, indicating mutagenesis causes qualitative differences in variant expression. 45S relative CN was calculated by qPCR as described. D, Representative nuclei subjected to FISH for 45S rDNA (red) from whole cotyledon and leaf tissues in WT and line #236 (T7). DNA is counterstained with DAPI (blue). In WT seedlings, both NORs localize at the nucleolus. The diffused signal within the nucleolus suggests chromatin de-condensation. After approximately 15 DAS, NOR2 is progressively silenced and moves away from the nucleolus. NOR4 localizes at the nucleolus during vegetative development. In line #236, where 45S rDNA signal is strongly reduced, all signals remain exclusively located at the nucleolus (Scale bar: 5 μm).

Figure 3

Synthesis of 45S rRNA appears largely regulated by chromatin organization. A, 45S rDNA locus. Letters indicate the probes used for RNA gel blots (B) and transcription run-on (C). Stars represent regions amplified by ChIP-qPCR. B, RNA gel blot analysis shows accumulation of 45S and other ribosomal RNAs in WT Col-0 and the LCN lines #236 and #289, letters indicate probes used as shown in (A), Methylene Blue is shown as loading control. Worth noticing that the plastid 16S and 23S rRNAs result equally represented in the LCN mutants and show a WT-like stoichiometric ratio with the 45S-derived rRNAs. C, Absolute quantification of 18S and 25S rRNA molecules in WT and #236 (T4 generation). No difference in the accumulation of either was detected (Student t test, bars represent standard error among biological replicates, n = 3 biological replicates). D, Nuclear transcription run-on assay shows transcription rate for 45S rRNA in WT Col-0 and the LCN lines #236 and #289. ACT2 transcript was used as control. E, ChIP-qPCR shows differential enrichment of global H3, H3K9me2 (silencing mark) and H3K9Ac (active mark) across the 45S loci. Ta3 and HXK1 were used as controls for a silent retrotransposon and a transcriptionally active gene, respectively. H3 occupancy (left) was determined relative to input. Fold enrichments for H3K9me2 (middle) were normalized against heterochromatic control Ta3; fold enrichments for H3K9Ac (right) were normalized against euchromatic control HXK1. (Student t test, bars indicate standard error among biological replicates, **P < 0.01, *P < 0.05, n = 3 biological replicates, no antibody control = average of 45S no antibody amplicons).

Figure 4

Genome integrity and transcriptome effects of 45S contraction. A, DNA coverage and gene expression analysis of two independent LCN lines compared with WT. For both LCN lines, outer layer represents fold change of the DNA coverage compared with WT, inner layer the identified DEGs (q-value <0.05, fold change >1.5). Coverage >1.5 or <0.5 versus WT, representing duplications and deletions, respectively, are highlighted in red. B, Euler diagram representing the number of DEGs between the two LCN lines and WT. C, Close-up of the duplicated region identified in line 289. Y-axis represent the fold coverage versus WT, X-axis represents the position on Chromosome 4. Bars indicate genes, color coded based on their expression level versus WT. D, Coverage of 45S rDNA gene loci in LCN lines relative to WT assessed by Nanopore sequencing. E, MapMan enrichment analysis of bins significantly enriched for up or downregulated genes in both independent LCN lines. X-axis represent the fold enrichment of each significant bin, left: downregulated genes, right: upregulated genes, the size of the circles corresponds to the −log₁₀ adjusted P-value.

Figure 5

Effects of rDNA CN reduction on rRNA transcription. Analysis of rRNA transcription in our lines revealed that gene dosage may be maintained by rendering both NORs competent for transcription potentially through the removal of silencing histone marks. The rDNA LCN lines generated can be further used to advance understanding of the functional role of rRNA genes in cellular and developmental processes.