(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Circadian regulation of the transcriptome in a complex polyploid crop [1]

['Hannah Rees', 'Earlham Institute', 'Norwich Research Park', 'Norwich', 'United Kingdom', 'Rachel Rusholme-Pilcher', 'Paul Bailey', 'Royal Botanic Gardens Kew', 'Richmond', 'Surrey']

Date: 2022-10 

The circadian clock is a finely balanced timekeeping mechanism that coordinates programmes of gene expression. It is currently unknown how the clock regulates expression of homoeologous genes in polyploids. Here, we generate a high-resolution time-course dataset to investigate the circadian balance between sets of 3 homoeologous genes (triads) from hexaploid bread wheat. We find a large proportion of circadian triads exhibit imbalanced rhythmic expression patterns, with no specific subgenome favoured. In wheat, period lengths of rhythmic transcripts are found to be longer and have a higher level of variance than in other plant species. Expression of transcripts associated with circadian controlled biological processes is largely conserved between wheat and Arabidopsis; however, striking differences are seen in agriculturally critical processes such as starch metabolism. Together, this work highlights the ongoing selection for balance versus diversification in circadian homoeologs and identifies clock-controlled pathways that might provide important targets for future wheat breeding.

Funding: H.R., R.R.P. and A.H. were funded by the BBSRC Core Strategic Programme Grant (Genomes to Food Security) BB/CSP1720/1 and its constituent work package, BBS/E/T/000PR9819 (WP2 Regulatory interactions and Complex Phenotypes). B.W., R.R.P. and A.H. was supported by the BBSRC Designing Future Wheat grant BB/P016855/1; BBS/E/T/000PR9783 (DFW WP4 Data Access and Analysis). B.C and J.C were supported by the BBSRC funded Norwich Research Park Biosciences Doctoral Training Partnership grant BB/M011216/1. C.R. by a BBSRC grant BB/V509267/1 and Wave 1 of The UKRI Strategic Priorities Fund under the EPSRC Grant EP/T001569/1, particularly the “AI for Science” theme within that grant & The Alan Turing Institute. We would also like to acknowledge BBS/E/T/000PR9816 (NC1 ‐ Supporting EI’s ISPs and the UK Community with Genomics and Single Cell Analysis) for data generation and BB/CCG1720/1 for the physical HPC infrastructure and data centre delivered via the NBI Computing infrastructure for Science (CiS) group. P.B. was supported by a BBSRC TRDF grant BB/N023145/1. A.N.D., L.L.B.D. and C.A.G. are funded by BBSRC ISP Genes in the Environment (BB/P013511/1). C.A.G. was also funded by UK BBSRC SWBIO DTP (BB/M009122/1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data Availability: Data for the recreation of all main figures in this manuscript are available in S1 Data . Data for the recreation of all supplementary figures in this manuscript are available in S2 Data . Fastq data from the RNA-seq circadian time course are available to view from the Grassroots Data Repository: https://opendata.earlham.ac.uk/opendata/data/wheat_circadian_Rees_2021 . A summary csv table with expression of wheat genes (TPM), Metacycle estimates, gene annotations and triad balance classification can be viewed in S11 Table is available here: https://opendata.earlham.ac.uk/opendata/data/wheat_circadian_Rees_2021 . A TPM expression matrix at the individual replicate level with triad IDs information is also available: S12 Table : https://opendata.earlham.ac.uk/opendata/data/wheat_circadian_Rees_2021 . Code for creating Loom plots, the cross-correlation analysis and the clustering analysis are available from our groups GitHub repository: https://github.com/AHallLab/circadian_transcriptome_regulation_paper_2022/tree/main .

Copyright: © 2022 Rees et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Here, we investigate circadian balance within wheat triads to understand how circadian control is coordinated in a polyploid crop with 3 subgenomes. Second, we examine similarities and differences between the circadian transcriptome in wheat and its distant dicot relative Arabidopsis, at a global level and at the level of genes encoding key pathways such as primary metabolism and photosynthesis.

Circadian control of carbon fixation and starch metabolism are thought to form part of the selective advantage conferred to Arabidopsis by the clock [ 22 , 23 ]. This is apparent in the lhy - /cca1 - short period double mutant in Arabidopsis, where night-time starch levels reach exhaustion earlier compared to wild type, triggering early onset starvation responses that reduce plant productivity [ 23 ]. Similarly, genes encoding photosynthesis-related proteins are well-established targets of the circadian clock and include the LIGHT HARVESTING CHLOROPHYLL A/B BINDING PROTEIN genes (LHCB also known as CAB genes) and photosystem I and II reaction centre genes [ 24 , 25 ].

The circadian network in Arabidopsis comprises a series of interlocking negative transcriptional feedback loops connected by key activators [ 12 ]. Although monocots such as wheat diverged from their dicot relatives over 140 million years ago [ 13 ], many circadian oscillator components seem to have been conserved, particularly those forming the core loop network. Orthologs of TIMING OF CAB EXPRESSION 1 (TOC1) and other PSEUDO-RESPONSE REGULATOR (PRR) genes have been identified in wheat, rice, and barley, and several loci within these genes have been associated with altered flowering times, most notably the ppd-1 locus within TaPRR3/7 [ 14 – 16 ]. Likewise, mutants of orthologs of LATE ELONGATED HYPOCOTYL (LHY), GIGANTEA (GI), EARLY FLOWERING 3 (ELF3), and LUX ARRYTHMO (LUX) have been identified that alter heading dates, pathogen susceptibility, plant height, or lower grain yields [ 17 – 21 ].

(a) Circadian clocks evolved independently in the ancestors of hexaploid wheat following divergence from a common ancestor approximately 6.5 million years ago. Colours of clock icons represent theoretical differences in clock regulation integrated in the tetraploid and hexaploid hybrids either through circadian balance or through dominance of a particular homoeolog copy. Speciation and hybridisation event dates are based on estimates from [ 115 ]. (b) Histogram showing distribution of period lengths in wheat split between the A, B, and D subgenomes. Periods were measured from 24–68 h data. Dotted line indicates the mean period for the A, B, and D subgenomes. (Data_Fig 1b in S1 Data ). (c) Density plot showing the distribution of period lengths across rhythmic transcripts (BH q < 0.01) in Arabidopsis, Brassica rapa, Brachypodium distachyon, Glycine max (soybean), and wheat based on meta2d estimates on 24–68 h data following transfer to constant light. (Data_Fig 1c in S1 Data ). (d) Proportions of triads with either 0 (red segment), 1 (green segment), 2 (blue segment), or 3 (purple segment) rhythmic gene(s) out of the 16,359 expressed triads in this dataset. Lighter shading in the outer segments represents cases where 1/2 homeolog(s) have high confidence rhythmicity (BH q < 0.01) alongside an arrhythmic homeolog (BH q > 0.05). We term these genes “imbalanced rhythmicity” triads. Triad differences are based on meta2d estimates from data 0–68 h after transfer to L:L. Of the 3,448 triads with 3 rhythmic genes (represented by the purple segment in (d)), we also looked for triads with circadian imbalance in: phase ( e ), period ( f ), or relative amplitude ( g ). A total of 464 triads had homoeologs which peaked with an optimum lag of 4, 8, or 12 h following cross-correlation analysis. A total of 1,139 triads had homoeologs with period differences of more than 2 h, and 701 triads had homoeologs with a more than 2-fold difference in relative amplitude. (h) An example of a triad with imbalanced rhythmicity, where 2 homoeologs are rhythmic and 1 is arrhythmic as can be seen when the homeologs are mean normalised (i). (j) Example triad where the D-genome homeolog lags by 8 h. (k) Example triad where the A genome homoeolog has a period estimate 4 h longer than the D-genome homoeolog. (l) Example triad where the relative amplitude of the D-genome homoeolog is more than 4 times that of the A-genome homoeolog, but the A-genome homoeolog is still rhythmically expressed ( m ). In plots h, j, k, and l, error bars on each plot represent standard deviation between 3 biological replicates. (Data_Fig 1h-m in S1 Data ). Genes in example triads are: [Triad 408: TraesCS3A02G533700, TraesCS3B02G610500, TraesCS3D02G539000], [Triad 13405: TraesCS6A02G269100, TraesCS6B02G296400, TraesCS6D02G245800], [Triad 10854: TraesCS6A02G166500, TraesCS6B02G194000, TraesCS6D02G155100], and [Triad 2454: TraesCS2A02G333000, TraesCS2B02G348800, TraesCS2D02G329900]. (n) Mean expression of transcripts across all time points in the A, B, and D subgenomes within imbalanced rhythmicity triads compared with circadian balanced and arrhythmic triads. Error bars represent standard error. (Data_Fig 1n in S1 Data ).

Circadian clock homologs have been both inadvertently selected during crop domestication and identified as crop improvement targets [ 1 – 4 ]. Understanding circadian regulation of the transcriptome in crops such as bread wheat (Triticum aestivum) may provide useful insights for future crop improvement. Wheat also provides an excellent model system to explore how the circadian clock and its outputs are coordinated in a recently formed, complex allopolyploid. In Arabidopsis, circadian transcription factors act in a dose-dependent manner, with both knock-out and overexpression mutants resulting in altered function of the circadian oscillator [ 5 – 8 ]. It is not yet understood how rhythmic gene expression is balanced in species with multiple copies of the same gene. T. aestivum is a hexaploid (AABBDD) formed through interspecific hybridisation of 3 diploid ancestors around 10,000 years ago [ 9 , 10 ]. Approximately 51.7% of high-confidence wheat genes still exist in triads; sets of 3 homoeologous genes present on each of the A, B, and D genomes [ 11 ]. As these homoeologs evolved independently for several million years prior to hybridisation, it is plausible that these independent species might have been subject to different selective pressures on their clocks ( Fig 1A ).

Results

Experimental context We generated a 3-day circadian RNA-seq time course, sampling all aerial tissue from wheat seedlings cv. Cadenza, using 3 biological replicates, every 4 h, following transfer to constant light conditions (CT0-68h). Transcripts were quantified at the gene level and were averaged across biological replicates. The entire time series (from 0 to 68 h) provides information about how gene expression changes upon transfer to constant conditions and is used in our triad analysis to identify differential patterns of circadian expression between homoeologs within wheat triads. A shorter part of the dataset (24 to 68 h) is used for calculation of circadian waveform characteristics to ensure appropriate quantification of rhythms is under free-running conditions and is also used in all cross-species comparisons.

Patterns of triad circadian balance across the genome Next, we wanted to determine whether certain genomic regions incorporate a physical clustering of circadian balance in rhythmicity. We hypothesised that if blocks of sequential triads with imbalanced rhythmicity are present within particular chromosomal regions, this might indicate differential chromatin accessibility or transcriptional suppression. To investigate this, we identified regions on each set of chromosomes where there were sequential triads having the same number of rhythmic homeologs (i.e., runs of 1, 2, or 3 rhythmic gene(s)). We also looked for runs of sequential rhythmic triads specific to a particular chromosome (i.e., runs of 1, 2, or 3 rhythmic genes specifically on chromosomes A, B, or D). In both cases, we found no evidence that triads with specific numbers of rhythmic homoeologs were grouped together more often than would be expected by chance (S5 Note and S4 Table). This suggests that distributions of rhythmic balance appear to be randomly distributed across the genome (S8 Fig).

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001802

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/