Preferential meiotic chromosome pairing among homologous chromosomes with cryptic sequence variation in tetraploid maize

Meiotic chromosome pairing between homoeologous chromosomes was reported in many nascent allopolyploids. Homoeologous pairing is gradually eliminated and replaced by exclusive homologous pairing in well-established allopolyploids, an evolutionary process referred to as the diploidization of allopolyploids. A fundamental question of the diploidization of allopolyploids is whether and to what extent the DNA sequence variation among homoeologous chromosomes contribute to the establishment of exclusive homologous chromosome pairing. We developed aneuploid tetraploid maize lines that contain three copies of chromosome 10 derived from inbred lines B73 and H99. We were able to identify the parental origin of each copy of chromosome 10 in the materials using oligonucleotide-based haplotype-specific chromosome painting. We demonstrate that the two identical copies of chromosome 10 from H99 pair preferentially over chromosome 10 from B73 in different stages of prophase I and metaphase I during meiosis. Thus, homologous chromosome pairing is favored to partners with the most similar DNA sequences and can be discriminated based on cryptic sequence variation. We propose that innate preference of homologous chromosome pairing exists in nascent allopolyploids and serves as the first layer that would eventually block all homoeologous chromosome pairing in allopolyploids.

Phosphorylation of histone H3 by Haspin regulates chromosome alignment and segregation during mitosis in maize

In human cells, Haspin-mediated histone H3 threonine 3 (H3T3) phosphorylation promotes centromeric localization of the chromosomal passenger complex, thereby ensuring proper kinetochore-microtubule attachment. Haspin also binds to PDS5 cohesin-associated factor B (Pds5B), antagonizing the Wings apart-like protein homolog (Wapl)-Pds5B interaction and thus preventing Wapl from releasing centromeric cohesion during mitosis. However, the role of Haspin in plant chromosome segregation is not well understood. Here, we show that in maize (Zea mays) mitotic cells, ZmHaspin localized to the centromere during metaphase and anaphase, whereas it localized to the telomeres during meiosis. These results suggest that ZmHaspin plays different roles during mitosis and meiosis. Knockout of ZmHaspin led to decreased H3T3 phosphorylation and histone H3 serine 10 phosphorylation, and defects in chromosome alignment and segregation in mitosis. These lines of evidence suggest that Haspin regulates chromosome segregation in plants via the mechanism described for humans, namely, H3T3 phosphorylation. Plant Haspin proteins lack the RTYGA and PxVxL motifs needed to bind Pds5B and heterochromatin protein 1, and no obvious cohesion defects were detected in ZmHaspin knockout plants. Taken together, these results highlight the conserved but slightly different roles of Haspin proteins in cell division in plants and in animals.

An empirical bayesian approach for testing gene expression fold change and its application in detecting global dosage effects

We are motivated by biological studies intended to understand global gene expression fold change. Biologists have generally adopted a fixed cutoff to determine the significance of fold changes in gene expression studies (e.g. by using an observed fold change equal to two as a fixed threshold). Scientists can also use a t-test or a modified differential expression test to assess the significance of fold changes. However, these methods either fail to take advantage of the high dimensionality of gene expression data or fail to test fold change directly. Our research develops a new empirical Bayesian approach to substantially improve the power and accuracy of fold-change detection. Specifically, we more accurately estimate gene-wise error variation in the log of fold change. We then adopt a t-test with adjusted degrees of freedom for significance assessment. We apply our method to a dosage study in Arabidopsis and a Down syndrome study in humans to illustrate the utility of our approach. We also present a simulation study based on real datasets to demonstrate the accuracy of our method relative to error variance estimation and power in fold-change detection. Our developed R package with a detailed user manual is publicly available on GitHub at https://github.com/cuiyingbeicheng/Foldseq.

DeepGRN: prediction of transcription factor binding site across cell-types using attention-based deep neural networks

Background

Due to the complexity of the biological systems, the prediction of the potential DNA binding sites for transcription factors remains a difficult problem in computational biology. Genomic DNA sequences and experimental results from parallel sequencing provide available information about the affinity and accessibility of genome and are commonly used features in binding sites prediction. The attention mechanism in deep learning has shown its capability to learn long-range dependencies from sequential data, such as sentences and voices. Until now, no study has applied this approach in binding site inference from massively parallel sequencing data. The successful applications of attention mechanism in similar input contexts motivate us to build and test new methods that can accurately determine the binding sites of transcription factors.

Results

In this study, we propose a novel tool (named DeepGRN) for transcription factors binding site prediction based on the combination of two components: single attention module and pairwise attention module. The performance of our methods is evaluated on the ENCODE-DREAM in vivo Transcription Factor Binding Site Prediction Challenge datasets. The results show that DeepGRN achieves higher unified scores in 6 of 13 targets than any of the top four methods in the DREAM challenge. We also demonstrate that the attention weights learned by the model are correlated with potential informative inputs, such as DNase-Seq coverage and motifs, which provide possible explanations for the predictive improvements in DeepGRN.

Conclusions

DeepGRN can automatically and effectively predict transcription factor binding sites from DNA sequences and DNase-Seq coverage. Furthermore, the visualization techniques we developed for the attention modules help to interpret how critical patterns from different types of input features are recognized by our model.

Development of a Transformable Fast-Flowering Mini-Maize as a Tool for Maize Gene Editing

Maize (Zea mays ssp. mays) is a popular genetic model due to its ease of crossing, well-established toolkits, and its status as a major global food crop. Recent technology developments for precise manipulation of the genome are further impacting both basic biological research and biotechnological application in agriculture. Crop gene editing often requires a process of genetic transformation in which the editing reagents are introduced into plant cells. In maize, this procedure is well-established for a limited number of public lines that are amenable for genetic transformation. Fast-Flowering Mini-Maize (FFMM) lines A and B were recently developed as an open-source tool for maize research by reducing the space requirements and the generation time. Neither line of FFMM were competent for genetic transformation using traditional protocols, a necessity to its status as a complete toolkit for public maize genetic research. Here we report the development of new lines of FFMM that have been bred for amenability to genetic transformation. By hybridizing a transformable maize genotype high Type-II callus parent A (Hi-II A) with line A of FFMM, we introgressed the ability to form embryogenic callus from Hi-II A into the FFMM-A genetic background. Through multiple generations of iterative self-hybridization or doubled-haploid method, we established maize lines that have a strong ability to produce embryogenic callus from immature embryos and maintain resemblance to FFMM-A in flowering time and stature. Using an Agrobacterium-mediated standard transformation method, we successfully introduced the CRISPR-Cas9 reagents into immature embryos and generated transgenic and mutant lines displaying the expected mutant phenotypes and genotypes. The transformation frequencies of the tested genotypes, defined as the numbers of transgenic event producing T1 seeds per 100 infected embryos, ranged from 0 to 17.1%. Approximately 80% of transgenic plants analyzed in this study showed various mutation patterns at the target site. The transformable FFMM line, FFMM-AT, can serve as a useful genetic and genomic resource for the maize community.

A transposon surveillance mechanism that safeguards plant male fertility during stress

Although plants are able to withstand a range of environmental conditions, spikes in ambient temperature can impact plant fertility causing reductions in seed yield and notable economic losses^1,2. Therefore, understanding the precise molecular mechanisms that underpin plant fertility under environmental constraints is critical to safeguarding future food production³. Here, we identified two Argonaute-like proteins whose activities are required to sustain male fertility in maize plants under high temperatures. We found that MALE-ASSOCIATED ARGONAUTE-1 and -2 associate with temperature-induced phased secondary small RNAs in pre-meiotic anthers and are essential to controlling the activity of retrotransposons in male meiocyte initials. Biochemical and structural analyses revealed how male-associated Argonaute activity and its interaction with retrotransposon RNA targets is modulated through the dynamic phosphorylation of a set of highly conserved, surface-located serine residues. Our results demonstrate that an Argonaute-dependent, RNA-guided surveillance mechanism is critical in plants to sustain male fertility under environmentally constrained conditions, by controlling the mutagenic activity of transposons in male germ cells.

The Gene Balance Hypothesis: Epigenetics and Dosage Effects in Plants

Dosage effects in plants are caused by changes in the copy number of chromosomes, segments of chromosomes, or multiples of individual genes. Genes often exhibit a dosage effect in which the amount of product is closely correlated with the number of copies present. However, when larger segments of chromosomes are varied, there are trans-acting effects across the genome that are unleashed that modulate gene expression in cascading effects. These appear to be mediated by the stoichiometric relationship of gene regulatory machineries. There are both positive and negative modulations of target gene expression, but the latter is the plurality effect. When this inverse effect is combined with a dosage effect, compensation for a gene can occur in which its expression is similar to the normal diploid regardless of the change in chromosomal dosage. In contrast, changing the whole genome in a polyploidy series has fewer relative effects as the stoichiometric relationship is not disrupted. Together, these observations suggest that the stoichiometry of gene regulation is important as a reflection of the mode of assembly of the individual subunits involved in the effective regulatory macromolecular complexes. This principle has implications for gene expression mechanisms, quantitative trait genetics, and the evolution of genes depending on the mode of duplication, either segmentally or via whole-genome duplication.

Inbreeding Depression in Genotypically Matched Diploid and Tetraploid Maize

The genetic and molecular basis of heterosis has long been studied but without a consensus about mechanism. The opposite effect, inbreeding depression, results from repeated self-pollination and leads to a reduction in vigor. A popular explanation for this reaction is the homozygosis of recessive, slightly deleterious alleles upon inbreeding. However, extensive studies in alfalfa indicated that inbreeding between diploids and autotetraploids was similar despite the fact that homozygosis of alleles would be dramatically different. The availability of tetraploid lines of maize generated directly from various inbred lines provided the opportunity to examine this issue in detail in perfectly matched diploid and tetraploid hybrids and their parallel inbreeding regimes. Identical hybrids at the diploid and tetraploid levels were inbred in triplicate for seven generations. At the conclusion of this regime, F1 hybrids and selected representative generations (S1, S3, S5, S7) were characterized phenotypically in randomized blocks during the same field conditions. Quantitative measures of the multiple generations of inbreeding provided little evidence for a distinction in the decline of vigor between the diploids and the tetraploids. The results suggest that the homozygosis of completely recessive, slightly deleterious alleles is an inadequate hypothesis to explain inbreeding depression in general.

Rapid Birth or Death of Centromeres on Fragmented Chromosomes in Maize

Comparative genomics has revealed common occurrences in karyotype evolution such as chromosomal end-to-end fusions and insertions of one chromosome into another near the centromere, as well as many cases of de novo centromeres that generate positional polymorphisms. However, how rearrangements such as dicentrics and acentrics persist without being destroyed or lost remains unclear. Here, we sought experimental evidence for the frequency and timeframe for inactivation and de novo formation of centromeres in maize (Zea mays). The pollen from plants with supernumerary B chromosomes was gamma-irradiated and then applied to normal maize silks of a line without B chromosomes. In ∼8,000 first-generation seedlings, we found many B-A translocations, centromere expansions, and ring chromosomes. We also found many dicentric chromosomes, but a fraction of these show only a single primary constriction, which suggests inactivation of one centromere. Chromosomal fragments were found without canonical centromere sequences, revealing de novo centromere formation over unique sequences; these were validated by immunolocalization with Thr133-phosphorylated histone H2A, a marker of active centromeres, and chromatin immunoprecipitation-sequencing with the CENH3 antibody. These results illustrate the regular occurrence of centromere birth and death after chromosomal rearrangement during a narrow window of one to potentially only a few cell cycles for the rearranged chromosomes to be recognized in this experimental regime.

The deposition of CENH3 in maize is stringently regulated

The centromere, as an essential element to mediate chromosome segregation, is epigenetically determined by CENH3-containing nucleosomes as a functional marker; therefore the accurate deposition of CENH3 is crucial for chromosome transmission. We characterized the deposition of CENH3 in maize by over-expression and mutational analysis. Our results revealed that over-expressing CENH3 in callus is lethal while over-expressing GFP-CENH3 and CENH3-YFP in callus and plants is not and can be partly deposited normally. Different mutations of GFP-CENH3 demonstrated that CENH3-Thr4 in the N-terminus was needed for the deposition as a positive phosphorylation site and the last five amino acids in the C-terminus are necessary for deposition. The C-terminal tail of CENH3 is confirmed to be responsible for the interaction of CENH3 and histone H4, which indicates that CENH3 maintains deposition in centromeres via interacting with H4 to form stable nucleosomes. For GFP-CENH3 and CENH3-YFP, the fused tags at the termini probably affect the structure of CENH3 and reduce its interaction with other proteins, which in turn could decrease proper deposition. Taken together, multiple amino acids or motifs were shown to play essential roles in CENH3 deposition, which is suggested to be affected by numerous factors in maize.

Site-specific recombinase genome engineering toolkit in maize

Site-specific recombinase enzymes function in heterologous cellular environments to initiate strand-switching reactions between unique DNA sequences termed recombinase binding sites. Depending on binding site position and orientation, reactions result in integrations, excisions, or inversions of targeted DNA sequences in a precise and predictable manner. Here, we established five different stable recombinase expression lines in maize through Agrobacterium-mediated transformation of T-DNA molecules that contain coding sequences for Cre, R, FLPe, phiC31 Integrase, and phiC31 excisionase. Through the bombardment of recombinase activated DsRed transient expression constructs, we have determined that all five recombinases are functional in maize plants. These recombinase expression lines could be utilized for a variety of genetic engineering applications, including selectable marker removal, targeted transgene integration into predetermined locations, and gene stacking.

Magnitude of modulation of gene expression in aneuploid maize depends on the extent of genomic imbalance

Aneuploidy has profound effects on an organism, typically more so than polyploidy, and the basis of this contrast is not fully understood. A dosage series of the maize long arm of chromosome 1 (1L) was used to compare relative global gene expression in different types and degrees of aneuploidy to gain insights into how the magnitude of genomic imbalance as well as hypoploidy affects global gene expression. While previously available methods require a selective examination of specific genes, RNA sequencing provides a whole-genome view of gene expression in aneuploids. Most studies of global aneuploidy effects have concentrated on individual types of aneuploids because multiple dose aneuploidies of the same genomic region are difficult to produce in most model genetic organisms. The genetic toolkit of maize allows the examination of multiple ploidies and 1-4 doses of chromosome arms. Thus, a detailed examination of expression changes both on the varied chromosome arms and elsewhere in the genome is possible, in both hypoploids and hyperploids, compared with euploid controls. Previous studies observed the inverse trans effect, in which genes not varied in DNA dosage were expressed in a negative relationship to the varied chromosomal region. This response was also the major type of changes found globally in this study. Many genes varied in dosage showed proportional expression changes, though some were seen to be partly or fully dosage compensated. It was also found that the effects of aneuploidy were progressive, with more severe aneuploids producing effects of greater magnitude.

Engineered minichromosomes in plants

Artificial chromosome platforms are described in plants. Because the function of centromeres is largely epigenetic, attempts to produce artificial chromosomes with plant centromere DNA have failed. The removal of the centromeric sequences from the cell strips off the centromeric histone that is the apparent biochemical marker of centromere activity. Thus, engineered minichromosomes have been produced by telomere mediated chromosomal truncation. The introduction of telomere repeats will cleave the chromosome at the site of insertion and attach the accompanying transgenes in the process. Such truncation events have been documented in maize, Arabidopsis, barley, rice, Brassica and wheat. Truncation of the nonvital supernumerary B chromosome of maize is a favorite target but engineered minichromosomes derived from the normal A chromosomes have also been recovered. Transmission through mitosis of small chromosomes is apparently normal but there is loss during meiosis. Potential solutions to address this issue are discussed. With procedures now well established to produce the foundation for artificial chromosomes in plants, current efforts are directed at building them up to specification using gene stacking methods and editing techniques.

Meiotic crossovers characterized by haplotype-specific chromosome painting in maize

Meiotic crossovers (COs) play a critical role in generating genetic variation and maintaining faithful segregation of homologous chromosomes during meiosis. We develop a haplotype-specific fluorescence in situ hybridization (FISH) technique that allows visualization of COs directly on metaphase chromosomes. Oligonucleotides (oligos) specific to chromosome 10 of maize inbreds B73 and Mo17, respectively, are synthesized and labeled as FISH probes. The parental and recombinant chromosome 10 in B73 x Mo17 F₁ hybrids and F₂ progenies can be unambiguously identified by haplotype-specific FISH. Analysis of 58 F₂ plants reveals lack of COs in the entire proximal half of chromosome 10. However, we detect COs located in regions very close to the centromere in recombinant inbred lines from an intermated B73 x Mo17 population, suggesting effective accumulation of COs in recombination-suppressed chromosomal regions through intermating and the potential to generate favorable allelic combinations of genes residing in these regions.

Meiotic crossovers (COs) are essential for proper chromosome segregation and generating novel combinations of alleles. Here, the authors develop haplotype-specific oligos on maize chromosome 10 for fluorescence in situ hybridization and analyze CO patterns in an intermated recombinant population derived from B73 and Mo17.

Hybrid Decay: A Transgenerational Epigenetic Decline in Vigor and Viability Triggered in Backcross Populations of Teosinte with Maize

In the course of generating populations of maize with teosinte chromosomal introgressions, an unusual sickly plant phenotype was noted in individuals from crosses with two teosinte accessions collected near Valle de Bravo, Mexico. The plants of these Bravo teosinte accessions appear phenotypically normal themselves and the F1 plants appear similar to typical maize × teosinte F1s. However, upon backcrossing to maize, the BC1 and subsequent generations display a number of detrimental characteristics including shorter stature, reduced seed set, and abnormal floral structures. This phenomenon is observed in all BC individuals and there is no chromosomal segment linked to the sickly plant phenotype in advanced backcross generations. Once the sickly phenotype appears in a lineage, normal plants are never again recovered by continued backcrossing to the normal maize parent. Whole-genome shotgun sequencing reveals a small number of genomic sequences, some with homology to transposable elements, that have increased in copy number in the backcross populations. Transcriptome analysis of seedlings, which do not have striking phenotypic abnormalities, identified segments of 18 maize genes that exhibit increased expression in sickly plants. A de novo assembly of transcripts present in plants exhibiting the sickly phenotype identified a set of 59 upregulated novel transcripts. These transcripts include some examples with sequence similarity to transposable elements and other sequences present in the recurrent maize parent (W22) genome as well as novel sequences not present in the W22 genome. Genome-wide profiles of gene expression, DNA methylation, and small RNAs are similar between sickly plants and normal controls, although a few upregulated transcripts and transposable elements are associated with altered small RNA or methylation profiles. This study documents hybrid incompatibility and genome instability triggered by the backcrossing of Bravo teosinte with maize. We name this phenomenon "hybrid decay" and present ideas on the mechanism that may underlie it.

Progressive heterosis in genetically defined tetraploid maize

Progressive heterosis, i.e., the additional hybrid vigor in double-cross tetraploid hybrids not found in their single-cross tetraploid parents, has been documented in a number of species including alfalfa, potato, and maize. In this study, four artificially induced maize tetraploids, directly derived from standard inbred lines, were crossed in pairs to create two single-cross hybrids. These hybrids were then crossed to create double-cross hybrids containing genetic material from all four original lines. Replicated field-based phenotyping of the materials over four years indicated a strong progressive heterosis phenotype in tetraploids but not in their diploid counterparts. In particular, the above ground dry weight phenotype of double-cross tetraploid hybrids was on average 34% and 56% heavier than that of the single-cross tetraploid hybrids and the double-cross diploid counterparts, respectively. Additionally, whole-genome resequencing of the original inbred lines and further analysis of these data did not show the expected spectrum of alleles to explain tetraploid progressive heterosis under the complementation of complete recessive model. These results underscore the reality of the progressive heterosis phenotype, its potential utility for increasing crop biomass production, and the need for exploring alternative hypothesis to explain it at a molecular level.

Genomic Balance and Speciation

The role of genomic balance in accumulating species hybrid incompatibilities is discussed. Aneuploidy has been shown to produce more global modulations than polyploidy with the responsible genes being transcription factors and signaling components involved in molecular complexes, illustrating a stoichiometric component to gene expression. Genomic imbalance is usually detrimental to the organism and in many cases results in lethality. Here, it is proposed that once gene flow is prevented between or within populations by various speciation initiating processes, the stoichiometric relationship of members of macromolecular complexes can change via compensatory drift with the eventual result of newly established functional balances. However, when these new relationships are brought together in interspecific hybrids, detrimental consequences will occur. We suggest that these detrimental interactions contribute to hybrid incompatibilities.

Global impacts of chromosomal imbalance on gene expression in Arabidopsis and other taxa

Changes in dosage of part of the genome (aneuploidy) have long been known to produce much more severe phenotypic consequences than changes in the number of whole genomes (ploidy). To examine the basis of these differences, global gene expression in mature leaf tissue for all five trisomies and in diploids, triploids, and tetraploids of Arabidopsis thaliana was studied. The trisomies displayed a greater spread of expression modulation than the ploidy series. In general, expression of genes on the varied chromosome ranged from compensation to dosage effect, whereas genes from the remainder of the genome ranged from no effect to reduced expression approaching the inverse level of chromosomal imbalance (2/3). Genome-wide DNA methylation was examined in each genotype and found to shift most prominently with trisomy 4 but otherwise exhibited little change, indicating that genetic imbalance is generally mechanistically unrelated to DNA methylation. Independent analysis of gene functional classes demonstrated that ribosomal, proteasomal, and gene body methylated genes were less modulated compared with all classes of genes, whereas transcription factors, signal transduction components, and organelle-targeted protein genes were more tightly inversely affected. Comparing transcription factors and their targets in the trisomies and in expression networks revealed considerable discordance, illustrating that altered regulatory stoichiometry is a major contributor to genetic imbalance. Reanalysis of published data on gene expression in disomic yeast and trisomic mouse cells detected similar stoichiometric effects across broad phylogenetic taxa, and indicated that these effects reflect normal gene regulatory processes.

High-efficiency genome editing using a dmc1 promoter-controlled CRISPR/Cas9 system in maize

Previous studies revealed that the promoters for driving both Cas9 and sgRNAs are quite important for efficient genome editing by CRISPR/Cas9 in plants. Here, we report our results of targeted genome editing using the maize dmc1 gene promoter combined with the U3 promoter for Cas9 and sgRNA, respectively. Three loci in the maize genome were selected for targeting. The T0 plants regenerated were highly efficiently edited at the target sites with homozygous or bi-allelic mutants accounting for about 66%. The mutations in T0 plants could be stably transmitted to the T1 generation, and new mutations could be generated in gametes or zygotes. Whole-genome resequencing indicated that no off-target mutations could be detected in the predicted loci with sequence similarity to the targeted site. Our results show that the dmc1 promoter-controlled (DPC) CRISPR/Cas9 system is highly efficient in maize and provide further evidence that the optimization of the promoters used for the CRISPR/Cas9 system is important for enhancing the efficiency of targeted genome editing in plants. The evolutionary conservation of the dmc1 gene suggests its potential for use in other plant species.

Meiotic Studies on Combinations of Chromosomes With Different Sized Centromeres in Maize

Multiple centromere misdivision derivatives of a translocation between the supernumerary B chromosome and the short arm of chromosome 9 (TB-9Sb) permit investigation of how centromeres of different sizes behave in meiosis in opposition or in competition with each other. In the first analysis, heterozygotes were produced between the normal TB-9Sb and derivatives of it that resulted from centromere misdivision that reduced the amounts of centromeric DNA. These heterozygotes could test whether these drastic differences would result in meiotic drive of the larger chromosome in female meiosis. Cytological determinations of the segregation of large and small centromeres among thousands of progeny of four combinations were made. The recovery of the larger centromere was at a few percent higher frequency in two of four combinations. However, examination of phosphorylated histone H2A-Thr133, a characteristic of active centromeres, showed a lack of correlation with the size of the centromeric DNA, suggesting an expansion of the basal protein features of the kinetochore in two of the three cases despite the reduction in the size of the underlying DNA. In the second analysis, plants containing different sizes of the B chromosome centromere were crossed to plants with TB-9Sb with a foldback duplication of 9S (TB-9Sb-Dp9). In the progeny, plants containing large and small versions of the B chromosome centromere were selected by FISH. A meiotic "tug of war" occurred in hybrid combinations by recombination between the normal 9S and the foldback duplication in those cases in which pairing occurred. Such pairing and recombination produce anaphase I bridges but in some cases the large and small centromeres progressed to the same pole. In one combination, new dicentric chromosomes were found in the progeny. Collectively, the results indicate that the size of the underlying DNA of a centromere does not dramatically affect its segregation properties or its ability to progress to the poles in meiosis potentially because the biochemical features of centromeres adjust to the cellular conditions.

Parallel altitudinal clines reveal trends in adaptive evolution of genome size in Zea mays

While the vast majority of genome size variation in plants is due to differences in repetitive sequence, we know little about how selection acts on repeat content in natural populations. Here we investigate parallel changes in intraspecific genome size and repeat content of domesticated maize (Zea mays) landraces and their wild relative teosinte across altitudinal gradients in Mesoamerica and South America. We combine genotyping, low coverage whole-genome sequence data, and flow cytometry to test for evidence of selection on genome size and individual repeat abundance. We find that population structure alone cannot explain the observed variation, implying that clinal patterns of genome size are maintained by natural selection. Our modeling additionally provides evidence of selection on individual heterochromatic knob repeats, likely due to their large individual contribution to genome size. To better understand the phenotypes driving selection on genome size, we conducted a growth chamber experiment using a population of highland teosinte exhibiting extensive variation in genome size. We find weak support for a positive correlation between genome size and cell size, but stronger support for a negative correlation between genome size and the rate of cell production. Reanalyzing published data of cell counts in maize shoot apical meristems, we then identify a negative correlation between cell production rate and flowering time. Together, our data suggest a model in which variation in genome size is driven by natural selection on flowering time across altitudinal clines, connecting intraspecific variation in repetitive sequence to important differences in adaptive phenotypes.

Genome size in plants can vary by orders of magnitude, but this variation has long been considered to be of little functional consequence. Studying three independent adaptations to high altitude in Zea mays, we find that genome size experiences parallel pressures from natural selection, causing a reduction in genome size with increasing altitude. Though reductions in overall repetitive content are responsible for the genome size change, we find that only those individual loci contributing most to the variation in genome size are individually targeted by selection. To identify the phenotype influenced by genome size, we study how variation in genome size within a single wild population impacts leaf growth and cell division. We find that genome size variation correlates negatively with the rate of cell division, suggesting that individuals with larger genomes require longer to complete a mitotic cycle. Finally, we reanalyze data from maize inbreds to show that faster cell division is correlated with earlier flowering, connecting observed variation in genome size to an important adaptive phenotype.