The maize striate leaves2 (sr2) gene encodes a conserved DUF3732 domain and is homologous to the rice yss1 gene

Abstract

Maize striate leaves2 (sr2) is a mutant that causes white stripes on leaves that has been used in mapping studies for decades though the underlying gene has not been identified. The sr2 locus has been previously mapped to small regions of normal chromosome 10 (N10) and a rearranged variant called abnormal chromosome 10 (Ab10). A comparison of assembled genomes carrying N10 and Ab10 revealed only five candidate sr2 genes. Analysis of a stock carrying the sr2 reference allele (sr2‐ref) showed that one of the five genes has a transposon insertion that disrupts its protein sequence and has a severe reduction in mRNA. An independent Mutator transposon insertion in the gene (sr2‐Mu) failed to complement the sr2‐ref mutation, and plants homozygous for sr2‐Mu showed white striped leaf margins. The sr2 gene encodes a DUF3732 protein with strong homology to a rice gene with a similar mutant phenotype called young seedling stripe1 (yss1). These and other published data suggest that sr2 may have a function in plastid gene expression.

1. INTRODUCTION

A large number of maize nuclear genes provide products necessary for chloroplast function. Mutations in these genes result in albino, virescent, pale green, yellow, or white striped plants. In the striped class, there are at least eight different maize loci—striate leaves1, striate leaves2, striate leaves3, striate leaves4, japonica striping1, japonica striping2, iojap striping1, and iojap striping2 (Gerald Neuffer et al., 1997). These mutants impact the efficiency of plastid transcription, translation or morphogenesis such that chloroplast function is impaired but not abolished. Striping is likely caused by the sorting out of mixed populations of functional and nonfunctional chloroplasts in a way that some lineages inherit no functional chloroplasts and appear as white sectors (Birky, 1983; Coe et al., 1988). White stripes tend to be wider and more common at the margins of leaves because cells at the margin undergo more division to expand the width of the leaf than cells in the center of the leaf (Han et al., 1992; Park et al., 2000; Walbot & Coe, 1979).

The iojap striping1 (ij1) gene has been described at the molecular level (Han et al., 1992; Rhoades, 1943). Chloroplasts within white stripes of ij1 mutants are present but lack ribosomes, suggesting a failure in ribosome assembly (Shumway & Weier, 1967; Siemenroth et al., 1980; Walbot & Coe, 1979). Iojap is a member of a conserved family of ribosomal silencing factor A/DUF143‐containing proteins (Häuser et al., 2012; Wang et al., 2016) that function in bacterial and chloroplast ribosome biogenesis and protein synthesis (Trösch & Willmund, 2019; Walbot & Coe, 1979). More recently, a mutant called white and green striate leaves1 (wgsl1) was described (Li et al., 2023), which may be an allele of striate leaves4 on chromosome 6. The wgsl1 gene encodes a 16S rRNA processing protein that is thought to be required for ribosome maturation (Li et al., 2023).

Striate leaves2 (sr2) was originally identified in Waseca Minnesota as a spontaneous mutant (Joachim & Burnham, 1953). Electron microscope analysis of white tissue revealed chloroplasts with unorganized internal structure and a lack of visible ribosomes, similar to ij1 (Williams & Kermicle, 1974). The sr2 gene is located on the long arm of chromosome 10, both on the normal form of chromosome 10 (N10) and a variant of chromosome 10 known as abnormal chromosome 10 (Ab10) (Rhoades & Dempsey, 1985). Here we combine comparative genomics with genetic and molecular analyses of two alleles of the sr2 gene to demonstrate that it encodes a DUF3732 protein with homology to rice Young Seedling Stripe1 (Zhou et al., 2017), a gene that is thought to modulate gene expression by plastid‐encoded plastid RNA polymerase.

2. RESULTS

2.1. Comparative genomics of N10 and Ab10

We started our analysis with a careful comparison of two forms of chromosome 10, the normal chromosome 10 (N10) and abnormal chromosome 10 (Ab10). The two chromosomes are syntenic except at the ends of their long arms, where, on Ab10, there is a large approximately 55‐Mb meiotic drive haplotype (Dawe, 2022; Liu et al., 2020). The Ab10 meiotic drive haplotype causes the preferential transmission of Ab10 when crossed as a female (Dawe, 2022). Portions of the end of normal chromosome 10 (N10) are present within the Ab10 haplotype though the order of genes is altered by inversions and rearrangements (Rhoades & Dempsey, 1985). On N10, four genes with visible mutant phenotypes called white2 (w2), opaque7 (o7), luteus13 (l13) and striate leaves2 (sr2) occur in the order w2–o7–l13–sr2 whereas on Ab10, the gene order is l13–o7–w2–sr2 (Rhoades & Dempsey, 1985). More comprehensive mapping demonstrated that there are two separate inversions but that sr2 is not included in either one (Mroczek et al., 2006). Complete genome assembly of the Ab10 haplotype (Liu et al., 2020) and subsequent whole genome alignments confirmed the two inversions but did not identify an obvious uninverted region of homology at the ends of the shared region (Liu et al., 2020; Song et al., 2022), raising concerns about the original interpretations of gene order.

To confirm the location of sr2, we grew and analyzed a line homozygous for the Ab10 terminal deletion line known as Ab10‐Df(K). Ab10‐Df(K) had been described as having the sr2 phenotype and a reduced stature, but otherwise appearing normal, suggesting that only a small section of the Ab10 shared region, including sr2, was missing (Rhoades & Dempsey, 1985). We grew homozygous Ab10‐Df(K) plants and confirmed that they are striated (see below). In addition, we Illumina‐sequenced the genomes of plants containing Ab10‐Df(K) and two other deletion chromosomes (Ab10‐Df[L] and Ab10‐Df[M] [Hiatt & Dawe, 2003b]) that do not have the sr2 phenotype and aligned the short read data to the Ab10 reference (Liu et al., 2020). The results showed that nearly all of the shared region is present in Ab10‐Df(K), suggesting that sr2 must lie at the end of the shared region of both haplotypes, presumably within a very small region or rearrangement (Figures 1 and S1) (Rhoades & Dempsey, 1985).

FIGURE 1.

Orthologs between the N10 and Ab10 assembly. Lines indicate orthologs between N10 and Ab10 determined by OrthoFinder (Emms & Kelly, 2019). Shades of green indicate expected orthologs, red and purple indicate unexpected orthologs that could or could not be sr2, respectively. Location of genetic markers with known physical position, o7 and w2, are shown (Udy et al., 2012; Wang et al., 2011). TR1 (light blue) and knob180 (bright green) are maize knob types. Trkin (dark blue) and kindr (dark green) are kinesin‐encoding genes responsible for the preferential transmission of Ab10 (Dawe et al., 2018; Swentowsky et al., 2020). Shared (brown) indicates the regions of known homology between Ab10 and N10. The sr2 region is where the sr2 gene has been mapped on Ab10, as defined by the breakpoints of Ab10‐Df(K) and Ab10‐Df(M) (Hiatt & Dawe, 2003b) (Figure S1). Annotated genes are indicated as blue vertical bars (Hufford et al., 2021; Liu et al., 2020). The plot at the bottom shows short reads from B73 (which has N10) mapped to the Ab10 reference assembly with a mapping quality greater than or equal to 20 (Hufford et al., 2021). This alignment shows the traditionally defined shared region.

2.2. Identification of a duplicated and inverted region in Ab10

We used OrthoFinder and BLAST to detect ortholog gene pairs within the sr2 candidate region on both N10 (distal to o7, [Wang et al., 2011]) and Ab10 (between the Ab10‐Df[K] and Ab10‐Df[M] breakpoints). This approach revealed six homologous gene pairs (Figure 1, Table S1). Surprisingly, we found that this region is also present within the larger inversion proximal to the Ab10‐Df(K) breakpoint on Ab10. The results suggest that a segment carrying the six genes was duplicated and inserted distal to the Ab10‐Df(K) breakpoint (Figure 2, Table 1). The fact that it is a duplication helps to explain why it was not visible in alignments that display one‐to‐one homology relationships (Liu et al., 2020; Song et al., 2022). The OrthoFinder output, and our own analysis of the structure and transcripts of each gene indicate that only 7 of the total 12 genes are functional (Table 1). In the terminal duplicated segment, there are five sr2 candidate genes, which we will refer to in the next two sections as genes 1, 2, 4, 5, and 6. The reference names for these genes can be found in Table 1.

FIGURE 2.

Duplicated region on Ab10. Cartoon representation of the duplicated region on Ab10 identified via OrthoFinder and BLAST (Camacho et al., 2023; Emms & Kelly, 2019). Shaded regions between N10 and Ab10 indicate regions of homology where the hourglass shape indicates an inversion. TR1 and knob180 are maize knob types and r1 is a kernel and plant color locus marking the edge of the Ab10 haplotype. Location of genes with known physical position, o7 and w2, are shown (Udy et al., 2012; Wang et al., 2011). Brown squares indicate an annotated gene (Hufford et al., 2021; Liu et al., 2020). Green checks indicate genes that appear functional.

TABLE 1.

Names of genes involved in Ab10 duplication in all locations.

Gene number	B73 N10 name	Ab10 inversion name a	Ab10 duplication name b	Ortholog on Ab10
	Name	Name	Name
1	Zm00001eb434490	Zm00043a049649	Zm00043a050219	Duplication
2	Zm00001eb434500	Zm00043a049648	Zm00043a050220	Duplication
3	Zm00001eb434510	Zm00043a049638	Zm00043a050221 + Zm00043a050222	Inversion only
4	Zm00001eb434520	Zm00043a049637	Zm00043a050224	Both duplication and inversion
5	Zm00001eb434530	Zm00043a049636	Zm00043a050225	Duplication
6	Zm00001eb434540	Zm00043a049635	Zm00043a050226	Duplication

^a Inversion refers to genes annotated in their expected location in the known Ab10 inversion.

^b Duplication refers to genes annotated outside of the known Ab10 inversion.

2.3. RNA‐seq of plants homozygous for the sr2‐ref reference allele implicates gene 1 as the most likely sr2 candidate

Given that the sr2‐ref mutation has a similar phenotype to Ab10‐Df(K), it seemed possible that the sr2‐ref allele may also be associated with reduced expression. We performed a differential expression analysis using mRNA from leaf tissue. Because the genetic background of the sr2‐ref allele is not known, we used the W22 inbred as the negative control. We aligned the sr2‐ref and W22 RNA‐seq data to the B73 v5 reference genome and performed differential expression analysis. Of the six duplicated genes, only genes 1 and 2 showed significant differential expression between sr2‐ref and W22 (Figure 3). Expression of gene 2 was 44% higher in sr2‐ref than in W22 while the expression of gene 1 was dramatically reduced to only 2% of the levels observed in the W22 inbred (Figure S2).

FIGURE 3.

Differential expression of duplicated genes between sr2 and wild‐type plants. Gene numbers refer to those defined in Table 1. Color represents log2 transformed expression value for each gene. **** = less than 0.0001, * = less than 0.05, ns = not significant.

De novo assembly of the (relatively few) gene 1 (Zm00001eb434490) transcripts in sr2‐ref suggested that there is an insertion that often creates a chimeric transcript that omits the first exon (Figure S3). The insertion itself is a chimera of the second through fourth exons of the CASP‐like protein 4A2 (Zm00001eb231550) and a DNA transposon (Zm10271_AC186904_1). The transposon Zm10271_AC186904_1 is annotated as a Robertson's mutator (Mu) element, but we can find no homology to Mu in the terminal inverted repeat sequence (TIR), suggesting that it is distantly related (if at all) to canonical Mu (Lisch, 2002). If the chimeric transcript from sr2‐ref is translated, only a portion of the encoded protein would be homologous to the wild‐type gene 1 protein. There is also a very small amount of full length gene 1 transcript in sr2‐ref, but it is >99% reduced relative to W22 (Figure S3).

To confirm the presence of the insertion in the gene1 (Zm00001eb434490), we designed three pairs of primers based on transcript isoforms, each with a forward primer in the insertion and a reverse primer in the gene (Table S2). All three primer pairs produced amplicons from sr2‐ref DNA, but not from W22 control. We then Sanger‐sequenced the shortest amplicon (794 bp, forward primer matching the part of the insertion homologous to CASP like 4A2). Its sequence included not only CASP like 4A2 sequence but also the 3′ 309‐bp TIR of Zm10271_AC186904_1 that had been spliced out of all detected transcripts (the 5′ TIR is present in some transcripts, see Figure S3, isoform 5). It also revealed its precise insertion point in the first intron of gene 1 (Figure S4). These results suggest that the reduced expression in sr2‐ref is because a DNA transposon carrying a truncated piece of CASP‐like 4A2 inserted into its first intron and disrupted both transcription and splicing.

2.4. Complementation tests using transposon‐induced alleles confirm that sr2 is gene 1

To further test which of the candidate genes were sr2, we carried out complementation tests with Mu insertions for genes 1, 2, 4, and 5 from the UniformMu collection ([Settles et al., 2007], there are no mutants for gene 6). Each of the mutant alleles contained a Robertson's mutator element within the first exon (Table S3). Unfortunately the genetic background of these lines is not ideal for testing recessive alleles of sr2. All UniformMu lines carry an allele of the colored1 gene known as R1‐r:standard (McCarty et al., 2013; Settles et al., 2007). R1‐r:standard is tightly linked to a dominant allele of inhibitor of striate1 (Isr1) that inhibits the sr2 mutant phenotype (Kermicle & Axtell, 1981; Park et al., 2000). Although it is theoretically possible that Isr1 was recombined from R1‐r:standard during the preparation of the UniformMu lines (McCarty et al., 2005), it seems highly unlikely, as Isr1 (Zm00001eb429350 on chr10:141210513–141213010) is only approximately 20 kb from the P component of the complex r1 locus (Zm00001eb429330 on chr10:141187279–141196584) (Park et al., 2000; Walker et al., 1995). One copy of Isr1 reduces the striping in a homozygous sr2 line, while two copies nearly eliminate the sr2 phenotype. In plants homozygous for both Isr1 and sr2, the leaves are thinner, and white stripes are only observed at the edges of leaf margins (Park et al., 2000).

To generate material for the complementation tests, we crossed lines carrying Mu alleles of genes 1, 2, 4, and 5 to sr2‐ref. The expectation was that if one of the genes is the sr2 gene, the Mu allele/sr2‐ref heterozygote for that gene would show a striated phenotype. We also self‐crossed the UniformMu lines to obtain plants that were homozygous for Mu alleles of each of the four genes. Phenotypic analyses of the progeny of these crosses, along with positive and negative controls, were carried out in both the greenhouse and field. Homozygous sr2‐ref and Ab10‐Df(K) had white stripes on leaf sheaths and blades in both environments, with sr2‐ref being more striped in the field (Figure S5).

The complementation data indicate that gene 1 (Zm00001eb434490) is the sr2 gene. We observed leaf margin striping in the complementation tests for gene 1 and in plants homozygous for the gene 1 Mu allele (called sr2‐Mu here forward), but not in lines carrying Mu alleles of the other three genes (Figure 4, Table 2). The sr2‐Mu homozygous plants grew poorly and had thin leaves (Figure 4f), particularly in the field where they died before striping is normally visible. However, when they grew to maturity in the greenhouse, sr2‐Mu homozygous plants consistently had white stripes on the edges of sheath margins (Table 2, Figure 4).

FIGURE 4.

Phenotypes of sr2 mutants. (a) Homozygous sr2‐ref plant grown in the field. (b) Homozygous Ab10‐Df(K) plant grown in the greenhouse. (c) Plant heteroallelic for sr2‐ref/sr2‐Mu grown in the greenhouse. These plants typically had few small leaf margin stripes indicated by the blue arrow. (d) Plant heterozygous for +/sr2‐ref (where + is wild type) grown in the greenhouse. This plant is a sibling of the plant shown in (c). (e) Plant homozygous for sr2‐Mu grown in the greenhouse. (f) Sibling plants demonstrating the severe phenotype in sr2‐Mu/sr2‐Mu homozygous plants relative to siblings.

TABLE 2.

Summary of leaf margin striping phenotypes in crosses with Mu alleles. ^{a , b}

Gene number	Greenhouse fall 2023					Field summer 2023					Greenhouse summer 2023					Greenhouse spring 2023					Total
	Mu	+	Mu	Mu	+	Mu	+	Mu	Mu	+	Mu	+	Mu	Mu	+	Mu	+	Mu	Mu	+	Mu	+	Mu	Mu	+
	sr2	sr2	Mu	+	+	sr2	sr2	Mu	+	+	sr2	sr2	Mu	+	+	sr2	sr2	Mu	+	+	sr2	sr2	Mu	+	+
Gene 1 c	6	6	10 e	4	6	5	4	2 f	9	7	4	0	2	0	0	0	0	2	6	2	15	10	14	19	15
Gene 2 d	0	0	0	0	0	4	0	4	0	0	3	0	2	0	0	0	0	8	6	2	7	0	14	6	2
Gene 4 d	0	0	0	0	0	4	0	3	0	0	3	0	3	0	0	0	0	0	0	0	7	0	6	0	0
Gene 5 d	0	0	0	0	0	4	0	3	0	0	3	0	3	0	0	0	0	0	0	0	7	0	6	0	0

^a Numbers indicate the number of plants that were scored.

^b Gray shaded and bolded cells indicate that all plants had leaf margin stripes. White cells indicate that no plants had leaf margin stripes.

^c For gene 1, two crosses were used to generate the plants scored. A self‐cross of a +/Mu plant generated homozygous Mu/Mu, +/Mu, and +/+ plants. A testcross of a +/Mu X sr2/sr2 plant generated the Mu/sr2 and +/sr2 genotypes. In fall 2023, we grew a large sample of the Mu/Mu class (since previous to that we had only grown 4 to maturity). As a result, the numbers shown do not reflect Mendelian ratios.

^d For genes 2, 4, and 5, the Mu/Mu plants were obtained by self‐crossing Mu/Mu homozygotes, and the Mu/sr2 plants were obtained by crossing Mu/Mu X sr2/sr2 plants.

^e One additional plant did not show leaf margin striping. It suffered an infection that stunted its growth and prevented it from flowering. Lack of the phenotype may be a result of the combination of Isr1, stunted growth, and lack of flowering or may reflect incomplete penetrance.

^f These two plants died before flowering.

The sr2 gene is annotated as a “BTB/POZ domain protein TNFAIP protein.” However sr2 has no homology to either the BTB/POZ domain or TNFAIP1. Rather, sr2 is a DUF3732 domain protein, with strong (86%) protein homology to the rice Young Seedling Stripe1 (yss1) gene (LOC_Os04g59570, [Zhou et al., 2017]). The fact that sr2 and yss1 have similar structure and function is consistent with the fact that sr2 is a core gene in maize, found in all 26 NAM founder inbreds and is highly expressed in leaf tip tissue (Hufford et al., 2021) (Figure S2).

3. DISCUSSION

Striate leaves2 (sr2) is a morphological marker that has been known since the 1940s and has played an important role in understanding the structure of abnormal chromosome 10. Comparative genomics and analysis of two independent alleles of sr2 indicate that the causal gene is a DUF3732 domain‐containing gene that is homologous to rice yss1. The phenotypes of sr2 and yss1 are similar, except that yss1 stripes are only present in early leaves, which is not the case in sr2.

The rice YSS1 protein is localized to nucleoids (chloroplast genomes) (Zhou et al., 2017). Rice yss1 mutants display reduced expression of genes that are transcribed by plastid‐encoded RNA polymerase (Zhou et al., 2017). In Arabidopsis, the homolog of sr2 and yss1 is AT4G33480. Biochemical data show that the AT4G33480 protein physically associates with psbA mRNA, which encodes a component of the Photosystem II reaction center in chloroplasts (McDermott et al., 2019). These data are consistent with the interpretation that SR2/YSS1/AT4G33480 functions at the level of chloroplast gene expression (Zhou et al., 2017). Our data further show that SR2 is unlikely to be absolutely required for chloroplast function (at least in maize) because homozygous Ab10‐Df(K) plants lack the sr2 gene yet still grow to maturity. Similar to ij1, sr2 appears to promote faster and/or more accurate chloroplast biogenesis (Han et al., 1992).

The sr2 phenotype is highly variable. Homozygous sr2‐ref plants grown in the same environment can vary from having very few to very many stripes (Figure S5). The degree of striping is also environmentally sensitive, with full sibling sr2‐ref plants grown in the field having more severe leaf margin striping than those grown concurrently in the greenhouse (Figure S5). The powerful effects of Isr1 in reducing the sr2 phenotype illustrates that there are also genetic modifiers of sr2 (Park et al., 2000). The Isr1 gene suppresses the growth of white tissue in sr2 and other white striped mutants such as ij1. Our observation that plants homozygous for sr2‐Mu tend to have thin leaves and few stripes can be at least partially explained by the presence of Isr1 in the UniformMu background (Figure 4e,f). A perhaps related case of a chloroplast phenotype being suppressed in the UniformMu background has been described for whirly1, a mutation that causes embryo arrest in the UniformMu background but an albino phenotype in other backgrounds (Prikryl et al., 2008; Zhang et al., 2013). The heavy reliance on the UniformMu resource in recent years may have inadvertently limited the identification of the many other albino, striated leaves, iojap striping, and japonica striping loci that have yet to be described at the molecular level (Gerald Neuffer et al., 1997).

4. METHODS

4.1. Plant material and growth

Ab10‐Df(K) was obtained from Marcus Rhoades and backcrossed to the W23 inbred (Hiatt & Dawe, 2003a). Ab10‐Df(L) and Ab10‐Df(M) were identified in our lab (Hiatt & Dawe, 2003b). Seeds carrying sr2‐ref (stock X16D), mu1058934 in gene 1 (stock UFMu‐08182), mu1089572 in gene 2 (stock UFMu‐12931), mu1037653 in gene 4 (stock UFMu‐03495), and mu1037651 in gene 5 (stock UFMu‐03745) were obtained from the Maize Genetics Cooperation Stock Center in Urbana, IL, USA (Table S3). Most experiments were conducted in the UGA Botany greenhouses in Athens, GA, USA. We also grew plants in an outdoor field site in Athens, GA, USA, in April–June 2023.

4.2. Sequencing deletion lines

We collected young leaf tissue from single plants homozygous for Ab10‐Df(K), Ab10‐Df(L), and Ab10‐Df(M) and extracted DNA with the IBI Genomic DNA Mini Kit (Plant) (IB47230) using the GP1 buffer and 16,000 × g for all centrifuge steps. We used a Kapa HyperPrep Kit to prepare the sequencing library (KK8580) and adapters from Netflex DNA Barcodes (Nova‐520996). Sequencing was performed by GENEWIZ from Azenta using a HiSeq 4000. We trimmed the resulting reads using cutadapt version 2.7 (Martin, 2011), mapped the reads to the Zm‐B73_AB10‐REFERENCE‐NAM‐1.0 reference (Liu et al., 2020) using BWAmem 0.7.17 (Li & Durbin, 2009), and removed duplicates using picard version 2.16.0 (Tools, 2018). We determined the breakpoints by filtering the primary alignments with a MAPQ> = 20, counting the number of reads over each bp using IGV tools, and visually inspecting the resulting files (Robinson et al., 2011). N10 reads shown in Figures 1 and S1 are B73 ~30X Illumina short reads from the NAM project (Hufford et al., 2021). Plots were made using R v4.3.1.

4.3. Identifying and analyzing orthologs in the duplicated region of Ab10

We selected the protein sequence for the longest isoforms of the annotated genes between the colored1 (r1) gene (Zm00001eb429330, B73 v5 annotation) and the ends of the long arm of chromosome 10 in the Zm‐B73‐REFERENCE‐NAM‐5.0 (https://www.maizegdb.org/genome/assembly/Zm-B73-REFERENCE-NAM-5.0) and Zm‐B73_AB10‐REFERENCE‐NAM‐1.0 (https://download.maizegdb.org/Zm-B73_AB10-REFERENCE-NAM-1.0/) assemblies (Hufford et al., 2021; Liu et al., 2020). These files were used to run OrthoFinder v2.5.2 using default parameters (Emms & Kelly, 2019). OrthoFinder identified five genes that were duplicated at the end of the shared region. BLAST v 2.2.26 (Camacho et al., 2023) was used to identify a sixth gene within the duplication. Plots were made using R v4.3.1.

We analyzed the transcripts for all six genes in the Ab10 inverted, Ab10 duplicated and N10 regions using the multiple sequence aligner and Clustal Omega in Geneious Prime v2022.0.2 (https://www.geneious.com/). We used the coding sequence (CDS) data published with the B73 and B73‐Ab10 reference genomes (Hufford et al., 2021; Liu et al., 2020). For gene 1 within the inversion, all transcript isoforms were truncated due to an unknown insertion in exon 4, which caused a frameshift and premature stop codon in exon 5. For gene 2 within the inversion, we found that all transcript isoforms either lacked significant homology to the N10 gene or were truncated for unknown reasons. For gene 3, we used Clustal Omega to determine that the duplication homologs are missing 288 bp of coding sequence and are unlikely to be functional. For gene 4, we used Clustal Omega to determine that both the Ab10 inverted and duplicated copies are similar to the N10 homolog and are likely both functional. For gene 5, we used Clustal Omega to determine that the Ab10 inversion homolog has one isoform that is severely truncated for an unknown reason, and one that results in the insertion of a proline within the PaaI thioesterase domain (Marchler‐Bauer et al., 2017). This may or may not disrupt function. For gene 6, we used Clustal Omega to determine that both Ab10 homologs are longer than the N10 counterpart and that both the Ab10 inversion and duplication homologs have a similar amount of homology to the N10 copy. This may or may not disrupt function in the Ab10 homologs. These data in conjunction with the results from OrthoFinder indicate that genes 1, 2, 4, 5, and 6 are candidates for sr2 (Table 1).

4.4. RNA‐seq of plants homozygous for sr2‐ref

4.4.1. RNA isolation and sequencing

We collected mature leaf tissue from homozygous sr2‐ref and W22 individuals grown side by side in the UGA Botany greenhouse and immediately froze it in liquid nitrogen. RNA was extracted using the IBI Total RNA Mini Kit (Plant) (IB47340), and cDNA was prepared using BioRad iSCRIPT Transcription Supermix for reverse transcriptase‐polymerase chain reaction (RT‐PCR) (1708891). GENEWIZ from Azenta performed a library preparation with PolyA selection using an NEBNext Ultra II RNA Library Prep Kit followed by paired end 150‐bp Illumina sequencing on a NovaSeq 6000.

4.4.2. Differential expression analysis

We trimmed reads using Trimmomatic version 0.39, and checked the quality before and after using FastQC version 0.11.9 (Bioinformatics, 2014; Bolger et al., 2014). We aligned the reads to the Zm‐B73‐REFERENCE‐NAM‐5.0 reference using HISAT2 version 2.1.0 using default parameters (Hufford et al., 2021; Kim et al., 2019). Using HTSeq version 0.13.5 with default parameters, we determined the number of reads mapped to each annotated feature (Anders et al., 2015). All genes with fewer than 10 reads were removed. We used DESEQ2 version 1.38.3 to perform a differential expression analysis with default parameters. Plots were made using R v4.3.1.

4.4.3. De novo assembly of sr2‐ref transcripts

We performed a Trinity v2.10.0 de novo transcriptome assembly on pooled data for all three biological replicates of sr2‐ref (Haas et al., 2013). BLASTv 2.2.26 was used to identify isoforms with homology to the sr2 gene. Some of the isoforms also contained sequences with homology to CASP‐like protein 4A2 (B73 v5 annotation Zm00001eb231550) and L‐aspartate oxidase (B73 v5 annotation Zm00001eb231540). The relative abundance of each isoform was determined using Kallisto within Trinity v2.8.4 (Bray et al., 2016; Haas et al., 2013). All transcript assemblies were then aligned to the B73 sr2 reference gene (Zm00001eb434490) using Geneious Prime v2022.0.2 (https://www.geneious.com/) MiniMap2 with default parameters followed by minimal manual curation.

The sr2‐ref allele insertion was confirmed via PCR using primers to CASP‐like 4A2 and sr2 (Table S2). The shortest PCR product was Sanger sequenced by Eton Biosciences. Plots were made using R v4.3.1 and Geneious Prime v2022.0.2 (https://www.geneious.com/).

4.5. Genotyping

All genotyping DNA extractions were performed using a CTAB protocol (Clarke, 2009). Polymerase chain reactions were performed using Promega GoTaq Green Master Mix (M7123). The wild type locus for each gene was detected using gene specific primers (Table S2). The Mu allele for each gene was detected using a primer to the Mu terminal inverted repeat (5′‐GCCTCYATTTCGTCGAATCCS‐3′) and either a forward or reverse gene specific primer. All genotyping reactions used the following temperature profile: hold 95°C, 2 min 95°C, 30(30 s 95°C, 30 s 60°C, 45 s 72°C), 5 min 72°C.

4.6. Complementation tests using Mu‐induced alleles

For genes 2, 4, and 5, we produced a seed that was homozygous for the Mu insertion of interest and crossed these to sr2‐ref to produce a seed that was all heterozygous for the insertion and sr2‐ref. For gene 1, we generated segregating populations of mutant and wild type by self‐crossing a plant heterozygous for the Mu insertion and crossing this plant to the sr2‐ref tester.

Plants were grown in both the greenhouse and field as detailed in Table 2. In the field, we randomized plant location and surrounded experimental plants with buffer corn to limit any environmental effects. In all cases, we grew sr2‐ref and Ab10‐Df(K) plants alongside the Mu‐bearing plants to confirm that the conditions were appropriate to see the phenotype.

AUTHOR CONTRIBUTIONS

Meghan J. Brady and R. Kelly Dawe designed the research. Meghan J. Brady, Jonathan I. Gent, and Maya Cheam performed research; Meghan J. Brady analyzed data. Meghan J. Brady and R. Kelly Dawe wrote the paper.

CONFLICT OF INTEREST STATEMENT

We have no conflicts of interest to declare.

PEER REVIEW

The peer review history for this article is available in the Supporting Information for this article.

Supporting information

Table S1. Orthologs outside of regions of known Homology between Ab10 and N10.

Table S2. Gene specific primers ¹ .

Table S3. Mu alleles used in this study.

Figure S1. Ab10 deletion breakpoints. Black bar plots show average read count calculated in 1 kb windows for Illumina paired end reads mapped to the Ab10 reference using BWAmem filtered to a MAPQ of 20 and only primary alignments. Red arrows indicate the presumed breakpoint. Location of genes with known physical position, o7 and w2, are shown (Udy et al., 2012; Wang et al., 2011). TR1 and knob180 are maize knob types. trkin and kindr are kinesin‐encoding genes responsible for the preferential transmission of Ab10 (Dawe et al., 2018; Swentowsky et al., 2020). Shared indicates regions of known homology between Ab10 and N10. Blue bars indicate annotated genes (Liu et al., 2020).

Figure S2. Expression of sr2. Leaf tip expression values (bar height) and copy number (number above bar) for sr2 in all NAM founders (Hufford et al., 2021). Data to the right of the blue line are leaf tip expression data from this project, copy number is unknown for these lines.

Figure S3. The sr2‐ref allele contains an insertion. A. The first line shows the gene model from B73, where gray arrows indicate exons. Isoforms 1 through 8 are derived from a Trinity de novo transcriptome assembly of sr2‐ref (Haas et al., 2013). * indicates there is a frame shift relative to the wild type in that isoform. Pink segments indicate the first ATG, blue segments indicate the first in‐frame stop codon, gold segments indicate primers (Supplemental Table 2 and Supplemental Figure 4). Light gray boxes indicate sequence that is different from the sr2 B73 reference, black indicates strong homology to sr2. Isoform 5 includes sequence with homology to the DNA transposon Zm10271_AC186904_1. Isoforms 6 and 7 include sequence with homology to CASP‐like protein 4A2 (Zm00001eb231550). None of the isoforms include the 3’ TIR of Zm10271_AC186904_1, indicating it is reproducibly spliced out of the transcript (Supplemental Figure 4). B. Transcripts per million (TPM) for each isoform in each biological replicate of sr2‐ref and wild type leaf tissue. The number above each set of three bars represents the average.

Figure S4. The sr2‐ref insertion is in the first intron. Sanger sequenced amplicon of JIG‐419 and JIG‐420 (Supplementary Table 4) from sr2‐ref DNA. Blue solid underlined text indicates homology to CASP‐like protein 4A2 (Zm00001eb231550). Green dotted underlined text indicates homology to the DNA transposon Zm10271_AC186904_1. Orange not underlined text indicates homology to sr2 (Zm00001eb434490). Bold text indicates exonic sequence and non‐bold indicates intron sequence (matching sr2‐ref isoform 6, Supplementary Figure 3). The exons precisely match the B73 v5 canonical transcripts. Italicized text indicates sequence that was expected in the amplicon but did not appear in the Sanger sequence due to its proximity to the sequencing primer site. Gray highlighted text indicates primers.

Figure S5. The sr2‐ref phenotype varies with environment and between individuals. All individuals shown are sr2‐ref homozygous full siblings grown at the same time in Athens GA, USA. A. Plant grown in the greenhouse with weak leaf striping. This phenotype is representative of three individuals. B. Plant grown in the field with strong leaf striping. C. Plant grown in the field with moderate leaf striping. In addition to those shown in B and C, two siblings were grown in the field that had intermediate leaf striping phenotypes.

Data S1 Supporting Information

Brady, M. J. , Cheam, M. , Gent, J. I. , & Dawe, R. K. (2024). The maize striate leaves2 (sr2) gene encodes a conserved DUF3732 domain and is homologous to the rice yss1 gene. Plant Direct, 8(2), e567. 10.1002/pld3.567

DATA AVAILABILITY STATEMENT

All code is available online (https://github.com/dawelab/Striate_Leaves2). Sequencing data from this project have been deposited to the SRA under bioprojects PRJNA1019348 and PRJNA339461. Relevant gene numbers can be found in Table 1 and Figure S3.