Corrections EVOLUTION Correction for “Origin, inheritance, and gene regulatory con- The authors note that Fig. 2 and its corresponding legend sequences of genome dominance in polyploids,” by Margaret R. appeared incorrectly. The corrected figure and its corrected legend Woodhouse, Feng Cheng, J. Chris Pires, Damon Lisch, Michael appear below. Freeling, and Xiaowu Wang, which appeared in issue 14, April 8, 2014, of Proc Natl Acad Sci USA (111:5283–5288; first published March 24, 2014; 10.1073/pnas.1402475111). A TE NonTE Average number of RNAs Average 0.15 0.85 1.55 2.25 2.95 -5 Kb -3 Kb -1 Kb 1 Kb 3 Kb 5 Kb Distance to transcription start/stop site B 1.31 LF MFs CORRECTIONS 1.03 0.75 Average number of RNAs Average 0.47 -5 Kb -3 Kb -1 Kb 1 Kb 3 Kb 5 Kb Distance to transcription start/stop site Fig. 2. Twenty-four–nucleotide smRNAs were enriched in TEs flanking LF B. rapa genes (gene space is represented by yellow arrow). In the y axis, the number of unique, perfectly mapped 24-nt RNAs was averaged in a 100-bp sliding window moving in 10-bp increments through each flanking region of B. rapa genes. All genes have an Arabidopsis ortholog. In the x axis, kilobytes from the start and end of transcription are shown. (A) Overall targeting level distribution of 24-nt RNA molecules to flanking sequences of the average B. rapa gene, with (green) and without (red) known transposons being hard-masked. (B) Overall targeting level distribution of 24-nt RNA molecules to flanking TE sequences outside of the gene (everything but TEs masked) of B. rapa genes, with subgenome location of these genes differentiated by LF (red) or MFs (blue, MF1 and MF2). Genes from subgenome LF and MFs were calculated separately. www.pnas.org/cgi/doi/10.1073/pnas.1405833111 www.pnas.org PNAS | April 29, 2014 | vol. 111 | no. 17 | 6527–6528 Downloaded by guest on October 3, 2021 PLANT BIOLOGY PHYSIOLOGY Correction for “Natural variation of rice strigolactone biosynthesis Correction for “Ion channel-kinase TRPM7 is required for main- is associated with the deletion of two MAX1 orthologs,” by Catarina taining cardiac automaticity,” byRajanSah,PietroMesirca, Cardoso, Yanxia Zhang, Muhammad Jamil, Jo Hepworth, Tatsiana Marjolein Van den Boogert, Jonathan Rosen, John Mably, Charnikhova, Stanley O. N. Dimkpa, Caroline Meharg, Mark H. Matteo E. Mangoni, and David E. Clapham, which appeared in Wright, Junwei Liu, Xiangbing Meng, Yonghong Wang, Jiayang issue 32, August 6, 2013, of Proc Natl Acad Sci USA (110:E3037– Li, Susan R. McCouch, Ottoline Leyser, Adam H. Price, Harro J. E3046; first published July 22, 2013; 10.1073/pnas.1311865110). Bouwmeester, and Carolien Ruyter-Spira, which appeared in is- The authors note that they “inadvertently omitted crediting sue 6, February 11, 2014, of Proc Natl Acad Sci USA (111:2379– Dr. Andreas Ludwig (Friedrich-Alexander-Universität Erlangen- 2384; first published January 24, 2014; 10.1073/pnas.1317360111). Nürnberg) for providing the HCN4-CreERT2 line for sinoatrial The authors note that, in the Acknowledgments, the statement deletion of TRPM7. We apologize for this oversight.” The above “Bala sequencing was supported by European Union FP6 Project credit statement should be added to the Acknowledgments. 015468 (“CEDROME”) and a Monitoring Agricultural Resources grant” should instead appear as “Bala sequencing was supported www.pnas.org/cgi/doi/10.1073/pnas.1405727111 by European Union FP6 Project 015468 (“CEDROME”)and aMarsgrant.” www.pnas.org/cgi/doi/10.1073/pnas.1405730111 6528 | www.pnas.org Downloaded by guest on October 3, 2021 Origin, inheritance, and gene regulatory consequences of genome dominance in polyploids Margaret R. Woodhousea,1,2, Feng Chenga,b,1, J. Chris Piresc, Damon Lischa, Michael Freelinga,2,3, and Xiaowu Wangb,3 aDepartment of Plant and Microbial Biology, University of California, Berkeley, CA 94720; bInstitute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences, Beijing 100081, China; and cDivision of Biological Sciences, University of Missouri, Columbia, MO 65211 Contributed by Michael Freeling, February 11, 2014 (sent for review October 21, 2013) Whole-genome duplications happen repeatedly in a typical flow- polyploidies (probably allopolyploids) demonstrate genome domi- ering plant lineage. Following most ancient tetraploidies, the two nance (i.e., this inverse correlation), but some (probably autopo- subgenomes are distinguishable because one subgenome, the lyploids) do not (12). dominant subgenome, tends to have more genes than the other Although functional DNA in maize was deleted preferentially subgenome. Additionally, among retained pairs, the gene on the from the recessive homeolog, nonfunctional DNA (e.g., trans- dominant subgenome tends to be expressed more than its re- posons, nonconserved intron sequence) was deleted from the cessive homeolog. Using comparative genomics, we show that two subgenomes at the same frequency, leading to a working genome dominance is heritable. The dominant subgenome of one hypothesis: Genes expressed less than their homeologs are easier postpolyploidy event remains dominant through a subsequent to delete because their removal is less likely to lower fitness, polyploidy event. We show that transposon-derived 24-nt RNAs so they escape purifying selection (10). Maize genes with a doc- target and cover the upstream region of retained genes preferen- umented phenotype tend to be on the dominant subgenome (13), and conserved noncoding sequences are preferentially deleted tially when located on the recessive subgenome, and with little from the recessive subgenome of maize (14) and B. rapa (6). Plant regard for a gene’s level of expression. We hypothesize that small allotetraploids happening less than 1 Mya express genome domi- RNA (smRNA)-mediated silencing of transposons near genes causes nance (15–17), but data from synthetic allotetraploids are am- position-effect down-regulation. Unlike 24-nt smRNA coverage, biguous (18, 19). Genome dominance may take many generations transposon coverage tracks gene expression, so not all transpo- to stabilize epigenetically, like inbreeding depression (20). sons behave identically. We propose that successful ancient tet- The gene expression component of genome dominance is usu- raploids begin as wide crosses between two lines, each evolved ally determined using RNA-sequencing (RNA-seq) data from the for different tradeoffs between transposon silencing and nega- shoot, a particularly complex organ system composed of many tive position effects on gene expression. We hypothesize that organs, developmental times, and cell expression end points. following a chaotic wide-cross/new tetraploid period, genes ac- Among suitable plant polyploids with sequenced genomes, maize quire their new expression balances based on differences in trans- (Zea mays var. B73) RNA-seq data are the richest (http://qteller. poson coverage in the parents. We envision patches of silenceable com/qteller3/); maize RNA-seq data best separate out gene ex- transposon as quantitative cis-regulators of baseline transcription pression into specific end points. SI Appendix,Fig.1compares EVOLUTION rate. Attractive solutions to heterosis and the C-value paradox expression [fragments per kilobase of transcript per million are mentioned. mapped fragments (FPKM)] between a homoeologous pair of maize genes retained from the most recent WGD, using 37 Brassica | epigenetics expression datasets generated by seven different laboratories. This figure illustrates a general rule: Homeolog expression hole-genome duplications (WGDs) have occurred in many is often biased, and that bias tends to be reflected at every Weukaryotic lineages, particularly plants (1). Twenty-four ancient plant polyploidies have been documented within plant Significance genomes as of January 2014 (http://genomevolution.org/wiki/index. php/Plant_paleopolyploidy). In maize, some crucifers, and perhaps Ancient plant polyploids contain dissimilar subgenomes. Sub- all new plant polyploids, fractionation (loss of duplicated genes) genomes that have lost fewer genes are subgenomes that tend and diploidization (return to diploid meiotic behavior) happen to express their genes to higher mRNA levels: “genome domi- within a few million years of the WGD (2). In the grasses, both nance.” Genome dominance is heritable through multiple rounds fractionation and diploidization are relatively rapid (3). The of polyploidy and over tens of millions of years. Twenty-four– general mechanism of fractionation is known for the most recent nucleotide RNA coverage of noncoding, transposon DNA up- tetraploidy in the maize var. B73 lineage (4) and the paleohex- stream of genes marks the recessive subgenomes of Brassica aploidy in the Brassica rapa var. Chiifu lineage (5, 6); this mech- rapa and Arabidopsis with only a small effect of gene expres- anism is deletion following intrachromosomal recombination. sion. We hypothesize that the “diploid” parent of a tetraploidy It is useful to partition post-WGD genomes into subgenomes with the lowest transposon load was to become the dominant when discussing fractionation. By doing so, it has been found that subgenome. The balance of transposon position effects on genes subgenomes generated upon ancient WGD in both the Arabidopsis is important for the regulation of quantity (and perhaps to solve (7) and maize (4) lineages do not behave in the same way: One of the heterosis and C-value paradoxes). the two subgenomes loses significantly more genes than the other. This phenomenon, called fractionation bias (7), has been gener- Author contributions: M.R.W., F.C., and M.F. designed research; M.R.W. and F.C. per- alized to all eukaryotes (8, 9). In studies comparing gene expres- formed research; F.C. and X.W. contributed new reagents/analytic tools; M.R.W., F.C., sion from genes retained as pairs (homeologs, also known as J.C.P., D.L., M.F., and X.W. analyzed data; and M.R.W., F.C., and M.F. wrote the paper. syntenic paralogs) in maize (10) and B. rapa (11), subgenomes The authors declare no conflict of interest. that had fewer genes (more highly fractionated) are enriched for 1M.R.W.