# Source paper archive - Source title: Ontogenetic and time-course transcriptomic analyses of maize leaf primordia reveal regulators of transverse vein development - Author(s): Maria Camila Medina, Lander Geadelmann, Diana Ruggiero, Erica Unger-Wallace, Erik Vollbrecht, Penelope L Lindsay, Samuel Leiboff - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.10.28.761427v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.10.28.761427v1.full.pdf - PDF SHA-256: 11e67aaa1c98076a480ecfd4e3642f9671cf4c5cef7fbfe4e0fd00f427107dc5 - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1 **Ontogenetic and time-course transcriptomic analyses of maize leaf primordia reveal** 2 **regulators of transverse vein development** 3 Maria Camila Medina [1], Lander Geadelmann [2], Diana Ruggiero [1], Erica Unger-Wallace [2], Erik 4 Vollbrecht [2], Penelope L. Lindsay [1,*], & Samuel Leiboff [1,*] 5 1. Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 6 97331, USA 7 2. Department of Genetics, Development and Cell Biology, Iowa State University, Ames, 8 Iowa, USA 9 10 **Authors for correspondence:** 11 Samuel Leiboff 12 _E-mail:_ leiboffs@oregonstate.edu 13 _Alternate e-mail:_ sleiboff@gmail.com 14 Penelope L. Lindsay 15 _E-mail:_ penelope.lindsay@oregonstate.edu 16 17 **ORCID:** 18 M.C.M: 0000-0002-0111-526X 19 L.F.G.: 0009-0001-5753-2898 20 D.R.: 0009-0008-7795-5338 21 E.U.W.: 0000-0002-2248-9877 22 E.V.: 0000-0003-4919-1365 23 P.L.L.: 0000-0002-0958-0529 24 S.L: 0000-0002-5576-3959 25 26 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 27 **ABSTRACT** 28 Leaf vein patterning varies greatly among land plants and is linked to leaf structure and function. 29 C 4 plants, such as maize, have a denser arrangement of veins than C 3 plants, which is 30 hypothesized to be a pre-adaptive trait for C 4 photosynthesis. Therefore, understanding how leaf 31 veins are patterned has implications for maximizing photosynthetic efficiency. Major maize leaf 32 veins, called longitudinal veins, are arranged in parallel and are interconnected by transverse 33 veins. Recent work has established molecular factors controlling longitudinal vein development; 34 however, it is unknown how transverse veins are patterned. Here, we investigated how transverse 35 veins establish connections within the maize vascular network. We performed anatomical 36 analyses of transverse veins and found that despite asynchronous development, transverse veins 37 fully integrate into the longitudinal vein network. Tracing the ontogeny of transverse vein 38 development using the marker _PIN1a-YFP_, we found that two sets of two initial cells on 39 opposing sides of adjacent longitudinal veins recruit intervening ground cells to form the 40 provascular strand. Treatment with auxin inhibitors did not abolish transverse vein initiation, 41 suggesting an auxin-independent pre-pattern for transverse vein initiation. Through time-course 42 RNAseq of leaf primordia, we identified a transcriptional signature that precedes transverse vein 43 development. Two genes that belong to this gene cluster, _EREB41_ and _EREB114_, are necessary 44 for transverse vein formation. Together, our studies establish the ontogeny of maize transverse 45 vein development and reveal how new vascular connections are specified, assembled, and 46 integrated into an existing vascular network. 47 48 **Keywords:** transverse veins, development, auxin, _Zea mays_, vascular development 49 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 50 **INTRODUCTION** 51 The plant vascular network is critical for the transport of water, nutrients, and signaling 52 molecules in most land plants. The leaf vasculature, also known as leaf veins, distribute water to 53 the leaf and mobilize photosynthates to sink tissues. Vein patterning varies tremendously 54 between plant species and is intrinsically linked to leaf morphology (Robil & Scarpella, 2026). 55 C 4 plants, including maize, often have a denser arrangement of veins, which is hypothesized to 56 be a pre-adaptive trait for C 4 photosynthesis (Christin _et al._, 2013). Therefore, understanding 57 how leaf venation patterning is established has important implications for plant photosynthetic 58 efficiency and assimilation of photosynthates into sink tissues. 59 Much of our molecular understanding of vein development comes from the model plant, 60 Arabidopsis, where veins are distributed in a reticulate pattern. Auxin biosynthesis, perception, 61 transport, and signaling are essential for vein patterning in Arabidopsis (Scarpella, 2024). 62 Polarized auxin transport mediated by the PIN1 auxin transporter reinforces procambial identity, 63 priming cells for vascular development. Directed auxin transport activates expression of the 64 _ARF5 / MONOPTEROS_ _(MP)_ auxin response transcription factor and its paralog _NON-_ 65 _PHOTOTROPHIC HYPOCOTYL4_ ( _NPH4)_, which induces the expression of _PIN1_, creating a 66 positive feedback loop to reinforce vascular identity. Recent work has shown the polarity67 associated kinase PAX is also needed for proper vein patterning and is polarized independently 68 of _PIN1_, indicating multilayered control of vein development (Bassukas _et al._, 2026). 69 In contrast to the reticulate patterning of Arabidopsis, maize leaves contain a vascular network 70 composed of parallel longitudinal veins interconnected by transverse veins (TVs). The 71 longitudinal vascular system is established through the sequential initiation of multiple vein 72 classes that differ in anatomy, developmental timing, and function, generating the characteristic 73 hierarchy of vascular bundles found in the mature leaf (Sharman, 1942; Dannenhoffer & Evert, 74 1994; Kumar & Kellogg, 2019). As leaf development progresses, transverse veins are added 75 between adjacent longitudinal bundles, creating an interconnected vascular network (Sharman, 76 1942; Dannenhoffer & Evert, 1994; Nelson & Dengler, 1997; Sakaguchi & Fukuda, 2008). This 77 developmental program produces the high vein density characteristic of maize and other C₄ 78 grasses (Ueno _et al._, 2006; Kumar & Kellogg, 2019). Precise regulation of vein identity, spacing, 79 and differentiation is therefore required to establish the mature vascular architecture of the leaf. 80 Recent developmental, transcriptomic, and genetic studies have revealed that distinct vein 81 classes arise through specific developmental trajectories and molecular programs, indicating that 82 maize vascular patterning is governed by both shared and vein-type-specific regulatory 83 mechanisms (Perico _et al._, 2022, 2024; Robil & McSteen, 2023; Vlad _et al._, 2024). 84 Several mutants that affect vascular patterning have been described in maize. _tangled1_ produces 85 supernumerary bundle-sheath cells caused by perturbations in cell division planes to impact vein 86 organization (Jankovsky _et al._, 2001); _scarecrow_ ( _scr_ ) and paralogous gene _scr1h_, affect the 87 spacing between veins (Hughes _et al._, 2019); _naked endosperm (nkd)_ exacerbates the _scr_ 88 phenotype, producing vein fusions (Hughes _et al._, 2023); _too many laterals_ ( _tml_ ) specifies 89 intermediate vein identity (Vlad _et al._, 2024); and the auxin biosynthesis mutant _vanishing_ 90 _tassel2 (vt2)_ impacts the proportion of vein types in the maize leaf (Robil & McSteen, 2023). 91 Most vein patterning mutants described to date also affect other aspects of development, 92 however, making it challenging to disentangle vascular patterning mechanisms from other leaf [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 93 developmental processes. Additionally, while several mutants that impact vein development have 94 been described, it is unclear how these genes coordinate to control vein patterning. 95 Recent molecular ontogenetic, spatial transcriptomic and single-nuclei RNAseq (snRNAseq) 96 studies have provided insights into the molecular regulators and timing of maize longitudinal 97 vein development. Spatial transcriptomics revealed discrete gene expression timing, spatial 98 localization, and vein subtype specificity for patterning vascular development (Perico _et al._, 99 2024). _PIN1a_ is an early, persistent marker of maize longitudinal veins (Robil and McSteen, 100 2023; Perico et al., 2024). A suite of auxin-related genes, including _AUXIN IMPORT CARRIER 3_ 101 _(AIC3)_ and _AUXIN AMIDO SYNTHETASE 2 (AAS2),_ are expressed in procambial and middle 102 ground tissue cells of a maize leaf primordia snRNAseq atlas (Yi _et al._, 2026). Together, these 103 data support a key role for auxin in maize vascular patterning, similar to Arabidopsis. 104 Despite recent advances in the understanding of longitudinal vein development, how vascular 105 networks become fully interconnected remains poorly understood. TVs, also known as cross 106 veins or commissural veins, are understudied relative to other monocot vein types. C 4 monocots 107 tend to have more TVs than C 3 monocots, in addition to a greater number of small longitudinal 108 veins, suggesting there may be an adaptive advantage for C 4 plants to have a greater number of 109 TVs (Ueno _et al._, 2006). Previous work has shown that transverse veins develop later than 110 longitudinal veins (Sharman, 1942; Blackman, 1971; Bosabalidis _et al._, 1994; Dannenhoffer & 111 Evert, 1994; Nelson & Dengler, 1997; Sakaguchi & Fukuda, 2008); however, the developmental 112 mechanisms that specify, assemble and integrate TVs into the pre-existing vascular network 113 remain unknown. Few mutants impairing TV formation have been described beyond the rice 114 _commissural vein excessive1 (coe1)_ mutant (Sakaguchi _et al._, 2010), limiting our functional 115 genetic understanding of TV development. In this work, we used a combination of anatomy, 116 ontogenetic, transcriptomic, and genetic analyses of TVs to understand how TVs develop and 117 integrate into the maize longitudinal vein network. 118 119 **MATERIALS AND METHODS** 120 **Plant material and growth conditions** 121 _Zea mays_ inbred line B73 was used for RNA-seq and anatomical analyses. Transgenic reporter 122 lines DR5 ( _DR5::erRFP_ ) and PIN1a ( _pZmPIN1a:ZmPIN1a-YFP_ ) were acquired from the 123 Jackson Lab at Cold Spring Harbor Laboratory as reported in (Gallavotti et al., 2008) Reporter 124 lines in both the B73 and A619 genetic background were used for developmental and time125 course experiments. All plants were grown in 32-well starter trays under greenhouse conditions 126 using a 1:1 mixture of Turface MVP and Sunshine Mix #4. Plants were maintained under a 16 h 127 photoperiod at 18-24°C and 40-60% relative humidity. Natural light was supplemented with 400 128 W sodium vapor lamps or reduced with overhead shade cloths to maintain approximately 700129 1500 μmol m⁻² s⁻¹ photosynthetically active radiation (PAR) at the substrate level. 130 **Histological analysis of transverse vein differentiation** [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 131 To examine transverse vein differentiation, P8 and P9 leaf primordia were dissected under a 132 Zeiss Stemi SV11 stereomicroscope. Samples were fixed in formalin acetic acid (FAA), 133 dehydrated through a graded ethanol series, cleared with Histochoice, and embedded in Paraplast 134 Plus (Surgipath). Sections (8 μm) were obtained using a Leica HistoCore BiocutR rotary 135 microtome, deparaffinized in Histochoice, stained with Astra Blue and Safranin O, and mounted 136 in Permount. Sections were imaged using a Leica DMI8 inverted microscope to assess transverse 137 vein differentiation. 138 **Live imaging of leaf primordia** 139 Transgenic maize plants expressing _pZmPIN1a:ZmPIN1a-YFP_ were used for live imaging of 140 transverse vein initiation and development. Leaf primordia older than leaf plastochron 6 (P6) 141 were carefully removed from 2-week-old seedlings using fine forceps under a Zeiss Stemi SV11 142 stereomicroscope to expose the P6 leaf primordia. 143 Dissected shoot apices with exposed P6 primordia were transferred to μ-Slide 1 Well chamber 144 slides (Ibidi), with three shoot apices in each chamber. This experiment was performed in 145 triplicate (three independent chambers; nine shoot apices total). Owing to continued leaf growth 146 during live imaging, some transverse vein initiation sites were displaced from the imaging field 147 or were difficult to relocate, resulting in incomplete time series for a subset of samples. Each 148 chamber contained 5 mL of half-strength Murashige and Skoog (MS) medium supplemented 149 with 1.5% (w/v) agar, 1% (w/v) sucrose, and 0.1% (v/v) Plant Preservative Mixture (PPM; Plant 150 Cell Technology). The samples were imaged from the bottom of the chamber at 12 h intervals. 151 Sterile water supplemented with 0.1% (v/v) PPM was added after each imaging session. Between 152 imaging sessions, samples were maintained in a vertical orientation in the dark. 153 **Auxin perturbation treatments** 154 To investigate the role of auxin transport during transverse vein formation, exposed P6 leaf 155 primordia were treated with auxin transport inhibitors N-1-naphthylphthalamic acid (NPA) and 156 2,3,5-triiodobenzoic acid (TIBA), or with exogenous indole-3-acetic acid (IAA). Stock solutions 157 (10 mM) of NPA, TIBA, and IAA were prepared in dimethyl sulfoxide (DMSO). Appropriate 158 volumes were added to cooled, autoclaved half-strength Murashige and Skoog (MS) medium 159 supplemented with 1.5% (w/v) agar and 1% (w/v) sucrose to obtain final concentrations of 10 160 μM IAA, 10 μM NPA, 30 μM NPA, or 30 μM TIBA. 161 Transgenic maize plants expressing _pZmPIN1a:ZmPIN1a-YFP_ were used for imaging. Leaf 162 primordia older than P6 were dissected from 2-week-old seedlings under a stereomicroscope to 163 expose the P6 primordia. Dissected shoot apices were transferred to μ-Slide 1 Well chamber 164 slides (Ibidi), with three shoot apices per chamber containing 5 mL of the treatment medium. 165 Three independent chambers were prepared for each treatment (nine shoot apices in total). 166 Samples were imaged at 12 h intervals to monitor the initiation and development of transverse 167 veins. As leaf primordia continued to grow and deform during culture, transverse vein initiation 168 sites could not always be relocated at subsequent imaging sessions, resulting in incomplete time 169 series for a subset of samples. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 170 **Confocal microscopy** 171 All confocal images were acquired using a Leica Stellaris 5 inverted laser-scanning confocal 172 microscope equipped with an HC PL APO 40×/1.10 W motCORR CS2 water-immersion 173 objective. Fluorophores were excited using a tunable white-light laser (WLL) at 514 nm for YFP 174 and 558 nm for RFP. Emission was collected at 520-530 nm for YFP and 580-590 nm for RFP. 175 For live-imaging experiments, samples were maintained in the imaging chamber described above 176 and kept in a vertical orientation in the dark between imaging sessions. Z-stack images were 177 acquired through the exposed leaf primordia using a step size of 0.42 μm, with a 16-bit image 178 depth, 512 × 512 pixel resolution, and a line averaging of 4. For high-resolution descriptive 179 images, z-stacks were acquired at a resolution of 1024 × 1024 pixels. 180 **3D cell segmentation and visualization** 181 Confocal z-stacks of P6 maize leaf primordia expressing _ZmPIN1a-YFP_ were processed using 182 MorphoGraphX (version 2.0.1-382). To facilitate cell boundary detection, image stacks were 183 subjected to three rounds of Gaussian blur filtering using default parameters, followed by 184 segmentation with the Bassel Combined U-Net convolutional neural network. Segmentation 185 outputs were further smoothed using a Gaussian blur filter with a radius of 1 pixel before 186 watershed segmentation. Individual cells were segmented using the ITK Auto-Seeded Watershed 187 algorithm with a watershed level of 800. The segmentation results were manually inspected and 188 corrected, and only cells relevant to the developing transverse vein were retained for further 189 analysis. Three-dimensional cell meshes were generated from the segmented volumes using the 190 Marching Cubes 3D algorithm with a cube size of 1. Segmented cells were visualized as 191 individual 3D meshes and overlaid onto the original confocal images to enable the tracking of 192 cell recruitment and division during transverse vein formation. Segmentation was used 193 qualitatively to reconstruct cell geometry and confirm cellular relationships over time. 194 **RNA extraction, library preparation and sequencing** 195 Maize leaf primordia were collected from four independent biological replicates. Ten primordia 196 were pooled per sample for each replicate. Samples included the shoot apical meristem (SAM) 197 with P1-P3 primordia, P4, P5, and P8, and a series of substages spanning the developmental 198 interval during which transverse veins (TVs) arise (P6.0-P6.3 and P7.1-P7.3). Because TV 199 development initiates during P6, P6.2, and P6.3 primordia were further dissected into upper 200 regions containing visible transverse veins and lower regions lacking transverse veins. 201 Leaf primordia were dissected under a Zeiss Stemi SV11 stereomicroscope equipped with a 202 transmitted-light base. A subset of samples was imaged for morphological characterization and 203 measurements using a Leica M205 FCA stereomicroscope coupled to a DMC6200 camera. To 204 preserve RNA integrity, samples collected for RNA-seq were processed immediately without 205 prior imaging. Instead, they were immediately transferred to 1.5 mL tubes containing three 206 stainless-steel beads, flash-frozen in liquid nitrogen, and stored at −80°C until processing. Frozen 207 tissues were homogenized using a SPEX Geno/Grinder at 1200 rpm for 30–45 s. Total RNA was 208 extracted using TRIzol reagent (Invitrogen), precipitated with isopropanol, washed with 70% [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 209 ethanol, and resuspended in nuclease-free water. RNA concentration was determined using a 210 Qubit BR-RNA Assay Kit (Thermo Fisher Scientific), and RNA integrity was assessed by 211 electrophoresis on 1% agarose gels prepared in 1× TAE buffer. 212 RNA samples were submitted to Novogene Co., Ltd. (Sacramento, CA, USA), where sequencing 213 libraries were prepared using the NEBNext Ultra RNA Library Prep Kit and sequenced on an 214 Illumina platform. 215 **RNA sequencing alignment and normalization** 216 Quality control, read trimming, and adapter trimming were performed using fastp (Chen et al., 217 2018). Reads were aligned to the B73 NAMv5.0 genome using HiSAT2 (Kim et al., 2019). 218 Subread from the featureCounts package was used to perform union-exon read counting 219 summarized by gene model (Liao et al., 2014). Raw reads counts were analyzed with DESeq2 220 (Love et al., 2014). Genes with more than 10 reads in at least 20 samples were considered 221 expressed and retained for further analysis. Reads were normalized using variance stabilized 222 transformation (VST). Variance stabilized reads were used to perform Principal Component 223 Analysis and to calculate Fragments per Kilobase Million (FPKM) values. Data and analysis 224 outputs were further visualized using R/tidyverse (Wickham et al., 2019) and R/ggplot2 225 (Wickham et al., n.d.) packages. 226 **Gene expression analysis** 227 _Expression Time Unit analysis_ 228 The expression time unit analysis was performed as described in (Leiboff et al., 2019). Briefly, 229 the data were fit to a five-knot spline. Euclidian distance was then used to position each sample 230 transcriptome to the closest corresponding point on the spline. Distance along the spline was 231 used to determine expression time units (ETU). 232 _Phyllotimer shiny application_ 233 Raw read counts, VST, and normalized FPKM data were exported and used to construct an 234 R/Shiny (Chang et al., 2026) application that displays gene expression trends across sample 235 labels using R/tidyverse (Wickham et al., 2019) and R/ggplot2 (Wickham et al., n.d.). 236 _Clustering analysis_ 237 The clustering analysis was performed with Clust using the default parameters with FPKM238 normalized values from the RNAseq dataset (Abu-Jamous & Kelly, 2018). GO enrichment 239 analysis was performed using ShinyGO v 0.85.2 (Ge et al., 2020). Heatmaps were generated in 240 RStudio 2026.01.0+392 running R version 4.5.2 with the R/tidyheatmaps package (Engler, 241 2026). 242 **Generation of** _**ereb**_ **mutants** 243 Transposon insertion alleles into _EREB41_ and _EREB114_ in W22 were generated via the AcDs 244 project (Ahern et al., 2009; Vollbrecht et al., 2010). Stock I.S08.1846 contains a _Dissociation_ 245 ( _Ds_ ) insertion into exon two of _EREB41_, in the coding region, and was used directly. For [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 246 _EREB114_ alleles we obtained the _Ds_ B.S05.0032 in stock AcDs-00004, located 6 kb upstream 247 into the transcription start site, and remobilized it into the gene body. Briefly, plants 248 heterozygous for _Ds_ -B.S05.0032 and heterozygous for _Activator_ ( _Ac-im_ ) were crossed as males 249 to an _r1-sc:m3_ tester stock. Approximately 4,000 F1 kernels were sown in the summer nursery 250 from which we screened 3,200 plants by PCR in pools using a three-primer strategy with either 251 two outward facing _Ds_ end primers and one gene-specific primer or two gene specific primers 252 and one outward facing Ds end primer as previously described (Strable et al., 2017). We 253 identified three exonic insertions into _EREB114_ and verified them by Sanger sequencing. 254 _ereb114::Ds3_ contained a _Ds_ in exon two in the coding region. This allele was propagated by 255 outcrossing to segregate away _Ac-im_ to produce a stable allele, before crossing to _ereb41::Ds-_ 256 _I.S08.1846_ . 257 CRISPR alleles were generated by designing guide RNAs to target the first exon of _EREB41_ or 258 _EREB114_ . Transformation of maize B104 embryos was done through the Crop Bioengineering 259 Laboratory at Iowa State University. CRISPR edits generated frameshift mutations leading to 260 premature stop codons in both genes. Transformants were screened for edits and outcrossed to 261 B73 to produce a non-transgenic F1 containing the edited alleles, and phenotyping was 262 conducted on the F2 of this cross. 263 264 **Plant phenotyping** 265 The transposon _ereb_ alleles were grown and analyzed in the summer of 2026 in Corvallis, OR at 266 the Oregon State University Botany and Plant Pathology Farm. The CRISPR _ereb_ alleles were 267 grown in Ames, Iowa in summer of 2026. Transverse vein phenotyping was performed in at least 268 two separate field seasons with similar results between field seasons. 269 _Transverse vein quantification_ 270 For transverse vein phenotyping, the leaf immediately above the primary ear was collected after 271 the ear had formed, including at least 8 cm of the blade and 5 cm of the sheath. Leaves were 272 then cleared overnight at 60°C in 70% EtOH and stained in 3:1 H 2 O:IKI (Ruzin, 1999) to darken 273 the starches in the veins for imaging. Leaves were cut in half, trimmed to lay flat during 274 scanning, and imaged on an Epson Perfection V850 Pro film scanner using an aqueous mount at 275 6400 dpi (2520 dpcm). 276 For transverse vein quantification in the sheath, images of leaf halves were rotated and cropped 277 to include only the sheath region, oriented horizontally. Using Adobe Photoshop, a measurement 278 overlay including a straight line with two - 1 cm [2] sampling boxes was placed in the center of the 279 lateral-medial axis of the imaged area, extending from the auricle boundary into the sheath. The 280 measurement overlay included a first box with its midpoint 2 cm (5040 pixels) away from the 281 auricle, and a second box 4 cm away ( **Supplemental Fig. 1a** ). The Fiji Point tool was used to 282 count transverse veins within each box in the images with measurement overlays. All veins 283 beginning or ending within a box were counted. For each leaf half image, counts across the two 284 boxes were averaged. For each leaf sample, box-averaged counts across the two leaf halves were 285 then averaged. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 286 Transverse veins were counted according to a scoring system incorporating five total classes: 287 ‘Regular’ veins, and an additional four ‘Irregular’ categories (Ambiguous, Loop, Incomplete, 288 Dot) representing transverse veins with atypical morphologies ( **Supplemental Fig. 1b** ). 289 _Leaf width measurements_ 290 To measure leaf width, the widest point in the leaf blade was measured at the leaf above the ear. 291 292 **RESULTS** 293 **Transverse veins integrate longitudinal veins into a continuous vascular network, despite** 294 **their asynchronous development** 295 Since anatomical descriptions of maize transverse veins (TVs) are limited, we first asked 296 whether TVs differentiate as complete vascular bundles and how individual vascular cell types 297 establish connectivity with neighboring longitudinal veins during network formation. To 298 characterize the anatomy of differentiating TVs and determine how they connect to longitudinal 299 veins, we analyzed serial sections of leaf plastochron (P) P8 and P9 stained with Astra Blue and 300 Safranin O **(Fig. 1)** . Although transverse veins arise earlier during development, P8 and P9 301 represent the earliest stages at which all major vascular domains can be reliably resolved 302 histologically, allowing assessment of tissue continuity between transverse and longitudinal 303 veins. In longitudinal sections of P8 tissue **(Fig. 1a, b)**, transverse provascular strands were 304 present, connecting adjacent longitudinal veins in both blade and sheath, which differed in the 305 degree of vascular differentiation. Cross-serial sections from the apical blade revealed 306 established, perpendicular provascular strands arranged in a regular grid, indicating pronounced 307 spatial periodicity **(Fig. 1a)** . Cross-sections confirmed that early transverse strands comprise two 308 rows of provascular cells with meristematic features, including prominent nuclei **(Fig. 1c)**, and 309 that sheath strands contained more cells than those in the blade, consistent with wider interveinal 310 spacing **(Fig. 1d)** . By P9, TVs had matured into fully differentiated vascular bundles **(Fig. 1e–l)** . 311 Despite these differences in length and cellular composition during development, mature TVs in 312 both blade and sheath exhibited the same overall vascular organization and polarity. Cross313 sections showed continuity of all major vascular bundle cell types between transverse and 314 adjacent longitudinal veins in both blade and sheath. This included phloem and metaphloem, 315 xylem/metaxylem, and a bundle sheath lacking a sclerenchyma ring **(Fig. 1e, f)** . 316 TVs maintained canonical polarity, with phloem abaxial and xylem adaxial, in both blade and 317 sheath despite the differences in interveinal spacing, and the bundle sheath enveloping the 318 transverse strand was continuous with that of the longitudinal vein **(Fig. 1e–g)** . Longitudinal 319 sections confirmed uninterrupted connectivity along the transverse strand **(Fig. 1g–j)** . Although 320 the ladder-like geometry of mature TVs, which become increasingly oblique as a consequence of 321 leaf expansion, often prevents a single section from capturing both xylem or both phloem 322 junctions simultaneously, continuity of phloem, xylem, and bundle sheath was nevertheless 323 evident using serial sections **(Fig. 1i, j)** . At higher detail, transverse xylem elements with 324 reticulate secondary wall thickening connected directly to the metaxylem (MXy) of the 325 longitudinal vein **(Fig. 1e–i, k, l)**, whereas transverse phloem with typically one to two uniseriate [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 326 sieve elements joined the metaphloem (MPh) **(Fig. 1e–g, j)** . Polarized light imaging showed that 327 the longitudinal vein is normally surrounded by a continuous sclerenchyma ring **(Fig. 1k)** . 328 However, at the point of transverse vein insertion, the sclerenchyma ring is interrupted, allowing 329 the transverse xylem strand to establish continuity with the longitudinal vein metaxylem 330 **(Fig. 1l)** . Similar transverse vein connections were observed in the sheath **(Fig. 1m)** . 331 Together, these observations demonstrate that TVs form fully differentiated vascular bundles that 332 preserve the organization and polarity of the longitudinal veins with which they connect. 333 Consistent with classical descriptions of grass leaf anatomy, TVs establish vascular continuity 334 between neighboring longitudinal bundles through direct xylem and phloem connections. Our 335 analyses further reveal that this continuity extends beyond the conductive tissues, encompassing 336 continuous bundle sheath domains across the transverse-longitudinal junction, a feature that has 337 received little attention in previous anatomical descriptions of grass transverse veins. The local 338 interruption of the sclerenchyma ring to accommodate transverse xylem further indicates that TV 339 incorporation requires coordinated remodeling of both conductive and supportive tissues, 340 providing a more detailed anatomical framework for understanding vascular network assembly 341 in maize leaves. 342 **Transverse veins develop sequentially in a basipetal pattern during maize leaf growth** 343 Previous anatomical studies in maize have established the sequence of longitudinal vein 344 formation, with midvein and lateral veins initiating acropetally between P1 and P3, followed by 345 the basipetal initiation of intermediate veins at P5 that refines the longitudinal pattern (Sharman, 346 1942; Nelson & Dengler, 1997; Robil & McSteen, 2023) **(Fig. 2a)** . The precise sites of TV 347 initiation and the cellular mechanisms underlying their formation remain unresolved, however, 348 raising key questions about how TVs develop and integrate into the longitudinal vein network. 349 To study the onset of TV formation, we dissected _PIN1a-YFP_ -expressing leaf primordia at stages 350 P5, P6, and P7 and analyzed the apical, middle, and basal regions. We used PIN1a-YFP as a vein 351 marker, since it is an early and persistent marker of the vascular trace (Robil and McSteen, 2023; 352 Perico et al., 2024). At P5, no signs of TV initiation were detected in any region. The provascular 353 strands of the longitudinal veins were already well established, with distinct interveinal spacing 354 in the apical and middle regions **(Fig. 2b)** . By P6, TV traces became visible at the tip. In the 355 middle region, the earliest indication of TV initiation was the appearance of a pair of short, 356 rectangular cells positioned across from one another, emerging from the procambium of adjacent 357 longitudinal veins. We refer to these opposing cells as “initials” **(Fig. 2c)** . Notably, no PIN1a358 YFP signal was initially detected in the intervening ground cell between the initials. In the apical 359 region of late P6 and more prominently by P7, a greater number of transverse vascular traces 360 were observed **(Fig. 2d)** . This increase corresponded with the expression of _PIN1a-YFP_ in the 361 ground cells located between the initials, thereby completing the vascular trace. These 362 observations suggest that the initials represent the earliest specified cells in TV ontogeny, with 363 intervening ground cells subsequently recruited. In contrast, no initials were detected at the basal 364 region of the leaf blade, supporting a basipetal gradient of TV development. By P8, TVs in the 365 blade are fully established without extending beyond the anastomoses of the intermediate and 366 small veins at the base of blade. At this stage, the vascular network displays a clear ladder-like 367 organization with nearly horizontal TVs connecting adjacent longitudinal veins **(Fig. 2e).** [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 368 TVs begin in the sheath at P8 once they reach approximately 3–5 mm in length **(Fig. 2e)** . This 369 suggests that TV formation in the sheath begins only after this domain reaches a sufficient size 370 and identity. At this stage, the anastomosis of intermediate and small veins is observed at the 371 blade–pre-auricle boundary, even though the auricle is not yet morphologically distinguishable. 372 By P9 **(Fig. 2f)**, the sheath has elongated to approximately 5–10 cm. At this stage, the auricle 373 begins to expand as part of overall leaf growth, after the TV pattern in both the blade and sheath 374 has already been established. Notably, the auricle itself lacks TVs. The anastomoses between 375 intermediate and small veins at the base of the blade become more pronounced but do not extend 376 into the auricle. As the leaf continues to expand, TVs adjust their curvature and relative spacing, 377 becoming progressively more oblique relative to the longitudinal veins while maintaining the 378 established vascular network architecture, reflecting the gradient of blade expansion and TV 379 maturation. 380 **Sequential recruitment and anticlinal division of ground cells drive transverse vein** 381 **formation in maize** 382 To characterize the cellular events underlying TV initiation, we examined _PIN1a–YFP_ 383 expression in P6 leaf primordia using confocal imaging and time-lapse analysis. In our 384 observations, ground cell recruitment began with a single layer of cells positioned between the 385 initials of adjacent longitudinal veins **(Fig. 3a)** . The number of recruited cells varied with 386 interveinal spacing; we typically observed one to three cells in the blade and five or more in the 387 sheath, where spacing is wider due to the absence of small veins. Recruited ground cells initially 388 showed no detectable PIN1a–YFP signal but subsequently exhibited strong expression as they 389 became incorporated into the forming transverse strand. 390 Following recruitment, ground cells subsequently underwent anticlinal divisions, increasing the 391 number of cells within the developing procambial strand and marking the onset of procambial 392 strand formation **(Fig. 3b)** . This sequence initiated at the leaf tip and progressed basipetally, 393 consistent with the broader pattern of TV development. In regions with wider spacing, a greater 394 number of ground cells were recruited between the two initials, resulting in proportionally wider 395 developing transverse strands, indicating that TV width scales with interveinal distance. 396 Time-lapse imaging in the blade domain revealed a reproducible sequence of events **(Fig. 3c)** . 397 During the first 12 h, ground cells situated between adjacent initials were recruited into the 398 emerging transverse strand. Between 12 and 24 h, these cells divided anticlinally, increasing 399 strand thickness. By 36 h, the developing TV consisted of multiple cells arranged in a ladder-like 400 configuration. 3D segmentation of individual cells corroborated these observations and enabled 401 precise tracking of ground cell recruitment and subsequent anticlinal divisions. Together, our 402 observations support a two-step model of TV formation in maize: (1) specification of TV initials, 403 followed by (2) rapid recruitment of intervening ground cells and coordinated, oriented anticlinal 404 divisions that build the transverse procambial strand and establish the cellular framework of the 405 future TV. 406 **Transverse vein formation persists under auxin transport inhibition but shows altered** 407 **patterning** [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 408 To determine whether auxin transport is required for ground-cell recruitment and anticlinal 409 divisions during TV formation, we performed live imaging of P6 leaf primordia expressing 410 _PIN1a-YFP_ following treatment with the auxin transport inhibitors N-1-naphthylphthalamic acid 411 (NPA) and 2,3,5-triiodobenzoic acid (TIBA), or with exogenous indole-3-acetic acid (IAA) ( **Fig.** 412 **4** ). Because PIN1a-YFP accumulates in TV initials and recruited ground cells undergoing 413 anticlinal division ( **Fig. 3** ), we hypothesized that inhibition of auxin transport would impair 414 recruitment and strand formation, whereas exogenous auxin would accelerate these processes. 415 In control samples ( **Fig. 4a** ), ground cells were recruited within 12 h and underwent anticlinal 416 division by 24 h, generating a characteristic two-layered, ladder-like TV strand. Treatment with 417 10 μM IAA accelerated both recruitment and division ( **Fig. 4b** ), resulting in the formation of two 418 cell layers within 12 h. Treatment with 10 μM NPA generally did not alter the timing of 419 recruitment or anticlinal division compared to the controls ( **Fig. 4c** ). However, occasional cases 420 in which recruited cells failed to undergo subsequent anticlinal division were observed **(Fig. 4d)** . 421 In contrast, treatment with 30 μM NPA ( **Fig. 4e** ) allowed recruitment to proceed but frequently 422 inhibited anticlinal divisions, producing predominantly single-layered strands. Under both 10 μM 423 and 30 μM NPA treatments, TV initials often appeared as single cells rather than the typical 424 opposing pair, divided only after neighboring ground-cell divisions, or failed to divide altogether 425 ( **Supplemental Fig. 2** ). Treatment with 30 μM TIBA ( **Fig. 4f** ) also permitted recruitment and 426 division; however, these events occurred earlier than in the controls, and recruited cells 427 frequently displayed abnormal morphologies. 428 Together, these observations indicate that TV initiation is remarkably robust to perturbations in 429 auxin transport. Although ground-cell recruitment and anticlinal division continued under all 430 treatment conditions, alterations in the timing of these events, cell division behavior, and cell 431 morphology disrupted the organization of developing TV provascular strands. 432 **Auxin signaling is activated during the differentiation of transverse veins** 433 To further investigate the role of auxin signaling in TV development, we analyzed maize lines 434 expressing both the auxin transport reporter _PIN1a–YFP_ and the auxin response reporter, 435 _DR5::RFP_ . This approach allowed us to assess auxin transport activity and signaling across 436 successive stages of TV formation. 437 At P6, _DR5::RFP_ signal was absent from developing TVs, including those in the apical region 438 where TVs are already established and relatively older, and remained restricted to longitudinal 439 veins **(Supplemental** **Fig. 3a)** . In contrast, PIN1a–YFP was already detectable in the initials and 440 longitudinal veins **(Supplemental Fig. 3a)**, indicating that auxin transport precedes the 441 activation of auxin response. 442 At P7, _DR5::RFP_ signal became detectable in TVs located in older regions. _DR5_ activity was 443 first detected in the initials of newly established TVs, identified by PIN1a–YFP signal along the 444 full vascular trace, and subsequently expanded into the recruited ground cells located closer to 445 the tip **(Supplemental** **Fig. 3b)** . This pattern is corroborated in P8 and P9, where _DR5::RFP_ 446 appears first in the initials of recently established TVs **(Supplemental** **Fig. 3c)** and only later in 447 the recruited ground tissue of older TVs **(Supplemental** **Fig. 3d)**, reflecting the gradual [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 448 activation of auxin signaling during differentiation. In these more mature regions at later 449 developmental stages, the PIN1a–YFP signal was reduced, consistent with the downregulation of 450 auxin transport as differentiation progressed. 451 Together, these observations indicate that auxin transport is active from the earliest phases of TV 452 initiation, whereas auxin signaling appears later and expands progressively from the initials into 453 recruited cells, coinciding with the onset of vascular cell differentiation. 454 **Transcriptomic profiling of transverse vein development using a substage-based** 455 **classification framework** 456 To resolve transcriptional changes associated with TV development, we established a refined 457 developmental staging system based on morphological landmarks. Fifty leaf primordia from each 458 plastochron stage (P1–P8) were dissected and characterized under a stereomicroscope, allowing 459 subdivision of developmental intervals exhibiting major morphological transitions **(Fig. 5a)** . 460 Stages P1-P5 were included as pre-TV controls because TVs are not anatomically detectable at 461 these stages. In particular, P5 was sampled to identify transcriptional programs that may precede 462 and potentially pre-pattern TV formation before their first anatomical appearance. Based on 463 external morphology, P6 and P7 were further subdivided into finer developmental substages 464 (P6.0-P6.3 and P7.1-P7.3). During P6, trichomes (macrohairs) progressively accumulated in a 465 basipetal pattern, and chlorophyll first became apparent at the leaf apex. During P7, trichome 466 abundance increased further, chlorophyll accumulation intensified, anthocyanin pigmentation 467 appeared at the leaf tip, and blade and sheath regions became increasingly distinct. By P8, the 468 sheath had expanded substantially, the ligular region and membranous ligule were visible, and 469 anthocyanin accumulation and early tissue senescence were evident at the leaf tip **(Fig. 5a)** . 470 To distinguish transcriptional signatures associated specifically with TV initiation, P6.2 and P6.3 471 primordia were dissected into upper regions containing developing TV and lower regions lacking 472 TVs **(Fig. 5b)** . This sampling strategy generated a developmental transcriptome series that 473 captures both temporal progression and spatial variation associated with TV formation, enabling 474 the identification of gene expression programs correlated with the onset and early differentiation 475 of TV. 476 To explore major trends in our RNAseq dataset we first identified the most variable 1000 477 transcripts and performed a Principal Component Analysis (PCA). PCA successfully separated 478 leaf primordia according to their spatiotemporal stage with SAM to P3 primordia grouped 479 samples at one end of PCA space and P8 primordia at the other ( **Fig. 5c** ). Interestingly, our split480 primordium experiment showed that while Upper Leaf halves and Lower Leaf halves can be 481 projected into PCA space along the same axis as complete primordia, Upper P6.2 and P6.3 482 primordia cluster with older primordia P7.1 and P7.2 ( **Fig. 5d** ). Conversely, Lower P6.2 and P6.3 483 leaf primordia are transcriptionally similar to P6 and P6.1 sample types ( **Fig. 5d** ). This 484 transcriptional finding reflects our anatomical morphological data which show a tip-to-base 485 gradient of development where the tip of P6.2 and P6.3 are more developed than the base of the 486 primordium. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 487 Restricting our analysis to just whole primordia, we fit a 5-knot spline to summarize the 488 trajectory trend in our PCA space and then determined the closest position along the spline for 489 each RNAseq sample ( **Fig. 5e** ). We used this location to determine Expression Time Units (ETU) 490 for each sample. We found relatively even coverage across our inferred developmental trajectory, 491 with high correlation between ETU and measured primordium length, where it was available 492 (ETU ~ Millimeters, R [2] = 0.82, **Supplemental Figure 4** ). 493 Considering transcript abundance across ETU values as well as primordium sample type 494 revealed several patterns in gene expression (browsable with a R Shiny dashboard: 495 https://leiboffdoesresearch.shinyapps.io/PhylloTimer_TEST/). Expression of the auxin efflux 496 carrier, _PIN1a_, for example, was high in the meristem and young leaf primordia, dropped to a 497 minimum at the P5 primordium stage, then rose again to a new peak in P6.1-P7.2 stages, 498 correlating with the production of vein types documented by our PIN1a-YFP confocal 499 microscopy ( **Fig. 5f** ). Closer examination of multiple _PIN_ genes revealed that although 500 _PIN1/PIN1a, PIN4/PIN1d,_ and _PIN12_ exhibited maxima early and in the middle of our leaf 501 development dataset, other _PINs_ peaked specifically at early ( _PIN3, 7, 8, 9_ ) or late 502 ( _PIN10,11,5,14_ ) stages across our observed developmental axis ( **Fig. 5g** ). These unique 503 expression patterns leave the opportunity for subfunctionalization amongst _PIN_ genes either 504 temporally or, more excitingly, for the specific vascular, stomatal and other developmental events 505 that take place during each phase of development. 506 **Clustering analysis of leaf plastochron RNAseq reveals two AINTEGUMENTA-related** 507 **transcription factors important for transverse vein development** 508 To identify genes important for TV initiation, we performed a clustering analysis of our leaf 509 primordia gene expression dataset. This revealed eleven clusters with dynamic expression 510 profiles over the course of leaf development ( **Fig. 6a** ). Enriched GO terms for clusters matched 511 the timing of anatomical milestones. For example, C0 and C1 had progressively increasing 512 expression over leaf plastochron stages and had significant GO enrichment for photosynthesis513 related terms, coinciding with chlorophyll accumulation in older leaf plastochrons ( **Figs. 5a, 6a,** 514 **Table S1** ). 515 From the eleven clusters, C4, C5, and C6 expression profiles most closely matched the timing of 516 TV development ( **Fig. 6a, Table S1, Table S2** ). In C4 and C5, expression increased sharply in 517 P6.1, where TVs are first observed, and decreased sharply in P8, when most TVs are already 518 established. C4 was enriched for translation and ribosomal assembly GO terms. C5, which 519 contains _PIN1a_, was enriched for cell cycle, DNA metabolism, translation, chromatin 520 organization, and spliceosome GO terms. Other genes in C5 included the _AUXIN RESPONSE_ 521 _FACTOR (ARF)_ genes _ZmARF28_ / _Truffala_, in which a dominant mutant has pleiotropic 522 developmental effects in maize (Prigge _et al._, 2025), _ZmARF29_, a maize ortholog of _MP,_ and 523 _ZmARF20_, orthologous to _NPH4_ . C5 also contained _VT2_, which is important for vein 524 development (Robil & McSteen, 2023). Together, this gene expression signature further supports 525 a role for auxin signaling in TV development. 526 C6 gene expression profiles preceded those of C5 genes, with higher expression at earlier P 527 stages, but similar to C4 and C5, expression decreased sharply at P8. C6 was enriched for 528 chromatin organization, cell cycle, negative regulation of gene expression, DNA metabolism, and 529 ribonucleoprotein complex GO terms, indicating this module may contain important [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 530 developmental regulators. The auxin receptor _ZmTIR1_ was in this cluster, which acts upstream of 531 ARF transcription factors and is important for auxin-dependent canalization processes in 532 Arabidopsis, including vascular development (Mazur _et al._, 2020). _EREB41_ and _EREB114_ were 533 also in this module. _EREB41_ and _EREB114_, also known as _ZmANT2_ and _ZmANT3_ (Lu _et al.,_ 534 2026), are homologous to Arabidopsis AINTEGUMENTA (ANT) transcription factors, which 535 play key roles in plant development (Horstman _et al._, 2014). In the monocot _Setaria viridis_, an 536 ortholog of _EREB41_ and _EREB114_, _SvANT1,_ regulates several developmental processes, 537 including vascular development (Liu _et al._, 2020). Furthermore, in Arabidopsis, _ANT_ is 538 expressed in procambial cells prior to vascular differentiation and _ant_ mutants have a reduction 539 in minor vein density (Kang et al., 2007). Additionally, _EREB41_ and _EREB114_ are enriched in 540 middle ground tissue and procambial cell clusters in single nuclei RNAseq (snRNAseq) of P4 541 leaf primordia and are central hub genes in a gene co-expression network of procambial and 542 middle ground tissue-expressed genes (Yi _et al._, 2026). We examined expression of the _EREBs_ 543 compared to _PIN1a_ and found that _EREB41_ and _EREB114_ expression preceded _PIN1a_ 544 expression **(Figure 6b)** . Therefore, we hypothesized _EREB114_ and _EREB41_ may be important 545 for TV development. 546 To test whether _EREB114_ and _EREB41_ are needed for TV formation, we obtained transposon547 insertion and CRISPR mutants in _ereb114_ and _ereb41_ . Both CRISPR-derived and transposon 548 insertion _ereb41;ereb114_ double mutant alleles had a reduction in transverse veins ( **Fig. 6c, d,** 549 **Supplemental Figure 5, 6** ). This was particularly evident in the sheath, where there was up to a 550 95% reduction in transverse veins in the _ereb41;ereb114_ double mutant compared to WT. Of the 551 transverse veins that were formed, approximately 25% were abnormal, where they would fail to 552 connect between two longitudinal veins, or would form two connections on the same 553 longitudinal vein ( **Fig. 6e,** **Supplemental Figure 6a, c)** . In addition to defects in TV 554 development, _ereb41;ereb114_ double mutant leaves were narrower than WT ( **Supplemental** 555 **Figure 7** ). Together, through the clustering analysis of the leaf plastochron transcriptome, we 556 identified two transcription factor genes important for TV development, _EREB41_ and _EREB114_ . 557 558 **DISCUSSION** 559 **Transverse veins are unified into the longitudinal vein network, despite asynchronous** 560 **development** 561 Our findings demonstrate that transverse veins (TVs) are integrated into the established 562 developmental framework of maize leaf vascular network, while revealing a distinct 563 developmental pathway underlying their specification, assembly, and integration into the 564 vasculature. Classical studies in maize, wheat, barley, and rice demonstrated that longitudinal 565 veins are specified and differentiated before the appearance of transverse or commissural veins, 566 suggesting that these two vascular systems develop in fundamentally different cellular 567 environments (Sharman, 1942; Blackman, 1971; Bosabalidis _et al._, 1994; Dannenhoffer & Evert, 568 1994; Nelson & Dengler, 1997; Sakaguchi & Fukuda, 2008). Within this context, our 569 observations refine the timing of TV initiation, showing that TVs emerge at P6 and extend 570 basipetally across both blade and sheath domains, including late onset initiation in the sheath at 571 P8, a developmental event not previously documented. TVs form fully differentiated vascular [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 572 bundles that preserve the organization and polarity of the longitudinal veins with which they 573 connect. Notably, transverse xylem established direct continuity with longitudinal vein 574 metaxylem by locally interrupting the sclerenchyma ring surrounding the longitudinal vein, 575 providing a structural basis for integration of the two vascular systems. Although sheath TVs 576 span greater interveinal distances than blade TVs, both develop comparable vascular 577 organization and maintain the canonical adaxial-abaxial polarity of adjacent longitudinal veins. 578 Together, these observations demonstrate that TVs arise through a coordinated developmental 579 program that controls their specification, assembly, and integration into the existing vascular 580 network. 581 **A two-step cellular mechanism underlies transverse vein initiation and strand formation** 582 TV ontogeny proceeds through a stepwise assembly process in which paired initials establish the 583 position of the future transverse strand and subsequently recruit intervening ground cells that 584 contribute to vascular strand construction. Live imaging further revealed that these recruited cells 585 undergo coordinated anticlinal divisions, generating the multicellular ladder-like procambial 586 structure characteristic of developing TVs. This developmental sequence provides a mechanistic 587 framework that links classical anatomical descriptions of TV formation with contemporary 588 models of vascular specification. 589 Early anatomical studies proposed that cross veins originate from files of undifferentiated cells 590 located between neighboring longitudinal veins. (Blackman, 1971) described wheat cross veins 591 as arising from rows of mesophyll cells that subsequently divide and differentiate into vascular 592 tissues, while Dannenhoffer & Evert, (1994) documented the progressive establishment of 593 procambial strands from ground meristem cells during barley vein development. Similarly, 594 Matsukura et al., (2000) and (Sakaguchi & Fukuda, 2008) showed that rice commissural veins 595 originate from specific cell populations positioned between adjacent longitudinal bundles and 596 undergo coordinated divisions prior to vascular differentiation. Together, these studies 597 established the anatomical origin of transverse and commissural veins in grasses but could not 598 resolve the cellular dynamics underlying their assembly. Our live-imaging analyses extend these 599 observations by directly visualizing the progressive recruitment of ground cells and their 600 subsequent incorporation into a developing transverse procambial strand. 601 The distinction between paired initials and recruited ground cells may have important 602 implications for how TVs are specified. Our observations indicate that TV formation combines 603 features of the classical precursor-cell model with a recruitment-based mechanism. Vascular 604 development in grasses is generally viewed as a lineage-based process in which defined 605 procambial precursor populations arise within the ground meristem and subsequently give rise to 606 vascular and bundle-sheath tissues. In maize, vascular and bundle-sheath cells originate from 607 median ground-meristem derivatives (Bosabalidis _et al._, 1994), whereas studies in barley and 608 rice have emphasized the importance of discrete precursor cell files during vein formation 609 (Dannenhoffer & Evert, 1994; Matsukura _et al._, 2000; Sakaguchi & Fukuda, 2008). Our results 610 suggest that TV development retains an early specification step represented by the paired initials 611 but subsequently expands the developing vascular domain through recruitment of neighboring 612 ground cells. Recruited cells therefore appear to acquire a pre-procambial fate after they are [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 613 incorporated into the forming strand rather than being predetermined members of a common 614 lineage from the outset. 615 The existence of paired initials before formation of a continuous procambial strand is consistent 616 with emerging evidence that vein development proceeds through multiple intermediate cellular 617 states. Perico et al., (2024) demonstrated that developing maize veins contain molecularly 618 distinct procambial populations and identified the expression of markers such as _TML1_, _HB52_, 619 _PIN1a_, _LAX2_, and _SHR1_ in developing procambial tissues prior to anatomically recognizable 620 vascular differentiation. These observations indicate that procambial identity is established well 621 in advance of visible vascular development and suggest that the paired initials identified here 622 may represent a specialized procambial state that precedes assembly of the transverse strand. 623 This possibility is further supported by the discovery that _TOO MANY LATERALS (TML)_ is 624 expressed in a distinct subset of procambial cells during intermediate-vein development, where it 625 suppresses lateral-vein fate while promoting intermediate-vein identity (Vlad _et al._, 2024). 626 A similar developmental logic has emerged from studies of vascular patterning in Arabidopsis. 627 During vein formation, broad _PIN1_ expression domains are progressively refined into narrow 628 files of pre-procambial cells before activation of _ATHB8_ that is associated with stable vascular 629 commitment (Wenzel _et al._, 2007; Marcos & Berleth, 2014; Krishna _et al._, 2021). Consequently, 630 anatomically recognizable procambial strands emerge only after a prolonged period of molecular 631 specification and pattern refinement (Scarpella, 2024; Robil & Scarpella, 2026). Our 632 observations suggest that TVs follow a similar developmental path. Paired initials establish the 633 spatial framework of the future transverse strand, recruitment progressively expands the 634 developing vascular domain, and coordinated anticlinal divisions subsequently generate a 635 recognizable procambial strand. Future studies combining live imaging, lineage tracing, and 636 molecular characterization will be essential for determining how these initials are specified, 637 whether they represent a unique procambial identity, and which signals coordinate recruitment of 638 surrounding cells during TV assembly. 639 **Auxin transport and signaling act sequentially during TV formation** 640 Our observations indicate that the temporal relationship between auxin transport and auxin 641 signaling described during longitudinal vein development also occurs during TV formation 642 (Robil & McSteen, 2023). PIN1a accumulated first in TV initials and subsequently in recruited 643 ground cells whereas detectable _DR5_ activity appeared only after formation of a continuous 644 procambial strand and increased progressively during later stages of differentiation. This 645 sequence suggests that auxin transport precedes strong auxin-responsive transcription during TV 646 ontogeny. 647 This interpretation is consistent with increasing evidence that developmental patterning can 648 occur before the appearance of strong _DR5_ maxima. During Arabidopsis lateral root formation, 649 developmental priming and retinal-associated signaling precede the establishment of stable _DR5_ 650 expression at future sites of organogenesis (de Smet, 2012; Xuan _et al._, 2020; Dickinson _et al._, 651 2021) . Similarly, the accumulation of PIN1a before detectable _DR5_ activity suggests that TV 652 initials may represent an early specification state that precedes the onset of robust auxin653 responsive transcription associated with procambial differentiation. This is consistent with [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 654 models in which auxin transport-mediated patterning and developmental specification occur 655 prior to the establishment of strong auxin-responsive gene expression (Gallavotti et al., 2008; 656 O’Connor et al., 2014; Petrášek & Friml, 2009). 657 While the identity of these signals remains unknown, our observations support the possibility 658 that auxin-responsive transcription is activated only after a developmental pre-pattern has been 659 established. 660 **Auxin transport contributes to patterning but is not essential for transverse vein initiation** 661 Our pharmacological perturbation experiments revealed that although NPA and TIBA altered the 662 organization of TV initials, affected division behavior, and disrupted formation of the 663 characteristic ladder-like architecture, TVs continued to form under all treatment conditions. 664 Thus, while auxin transport influences TV patterning, it is not strictly required for TV initiation. 665 The persistence of TV formation following auxin transport inhibition has important implications 666 for models of vascular patterning. In classical canalization models, vascular strands are thought 667 to arise along auxin transport routes reinforced by positive feedback between auxin flow and 668 transport capacity (Wenzel _et al._, 2007; Petrášek & Friml, 2009). Within this framework, PIN669 mediated transport contributes to the spatial organization of developing vascular tissues (Wenzel 670 _et al._, 2007; Marcos & Berleth, 2014). Our observations are consistent with this role, as 671 inhibition of auxin transport disrupted the positioning, organization, and connectivity of 672 developing TVs. However, the continued formation of TVs under all treatment conditions 673 indicates that canalization alone cannot fully explain TV initiation. Instead, TV specification 674 appears to involve additional developmental processes that remain active when auxin transport675 mediated patterning is perturbed. Inhibitor-treated leaves have broadened veins, altered 676 continuity, increased vascular density, and disrupted organization, yet continue to differentiate 677 vascular tissues (Sieburth, 1999; Wenzel _et al._, 2007). Similarly, disruption of auxin transport in 678 maize is associated with aberrant vascular development, altered ligule positioning, leaf twisting, 679 and hypertrophic vascularization, while tissue differentiation itself remains intact (Tsiantis _et al._, 680 1999; Scanlon, 2003). 681 Contemporary models of vascular development describe vein formation as the outcome of 682 coordinated interactions among auxin transport, auxin signaling, procambial specification, and 683 intercellular communication (Bennett, 2015; Schuetz _et al._, 2019; Linh & Scarpella, 2022). 684 Within this framework, TV formation appears to emerge from the integration of multiple 685 developmental inputs acting across distinct stages of ontogeny. For example, Arabidopsis studies 686 have indicated that vein patterning can be shaped by the combined contributions of auxin 687 signaling, PIN-mediated transport, and plasmodesmata-based flux (Linh & Scarpella, 2022; 688 Holloway _et al._, 2025).The continued formation of TVs following auxin transport inhibition 689 indicates that acquisition of TV identity, recruitment of intervening ground cells, and assembly 690 of the procambial strand remain developmentally active under conditions that alter auxin 691 transport-mediated patterning. These observations support a model in which PIN-mediated auxin 692 transport contributes to the spatial organization and coordination of developing TVs while 693 additional developmental pathways participate in establishing the vascular connection itself. 694 O’Connor et al., (2014) proposed that PIN-dependent patterning in grasses follows [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 695 developmental trajectories that differ from those described in Arabidopsis, and Strable & 696 Nelissen, (2021) emphasized that developmental patterning and growth are tightly coordinated 697 throughout maize leaf ontogeny. Consistent with these perspectives, our findings suggest that TV 698 formation integrates positional information, auxin transport, auxin signaling, and growth 699 dynamics to generate new vascular links within an already functional transport network. 700 **Auxin-related gene expression coincides with transverse vein development** 701 Through clustering of our leaf plastochron developmental time series gene expression profiles, 702 we identified several gene expression clusters whose expression coincides with TV development. 703 This includes a cluster with _PIN1a_ and auxin response factors _ZmARF20_, _ZmARF28_, and 704 _ZmARF29_ . Intriguingly, _ZmARF28_ and _ZmARF29_ are orthologous to Arabidopsis _MP_, but a 705 different _MP_ ortholog, _ZmARF4_, is induced during longitudinal vein development, hinting at 706 potential sub-functionalization of _MP_ -related ARFs for longitudinal and transverse vein 707 development (Perico _et al._, 2024). Perico et al. identified unique transcriptional signatures 708 underling lateral and rank-1 intermediate veins, supporting the idea that a unique transcriptional 709 profile may also exist for TV development. 710 Our clustering analysis also revealed a gene expression cluster that preceded expression of 711 _PIN1a_, which included the ANT-related transcription factors _EREB41_ and _EREB114_ . 712 _ereb41;ereb114_ double mutants had a reduced number of TVs, indicating that they are important 713 regulators of TV development. ANT-related transcription factors are proposed to promote 714 meristematic competence in many tissues and have an intimate connection with auxin in 715 Arabidopsis (Horstman _et al._, 2014). _ANT_ is important for tertiary vein formation in Arabidopsis, 716 and _MP_ is bound by both ANT and its paralog, AIL6, demonstrating a conserved role for ANT717 related transcription factors in controlling vein development (Kang _et al.,_ 2007; Krizek _et al.,_ 718 2021). Additionally, the ANT-related TFs, _PLETHORA3_, _5_, and _7_ regulate auxin biosynthesis 719 and promote auxin signaling in the shoot apical meristem for proper phyllotaxis (Pinon _et al._, 720 2013). Since _EREB41_ and _EREB114_ expression precedes _PIN1a_ expression in our developmental 721 series, this suggests that these genes may act upstream of _PIN1a_ to initiate TV formation. In 722 support of this idea, exogenous auxin treatment did not affect the expression of rice _EREB_ 723 orthologs (Yi _et al._, 2026). Future work will genetically position _EREB41_ and _EREB114_ in the 724 ontogeny of transverse vein development. 725 In addition to a defect in TV development, _ereb41;ereb114_ double mutants have narrower leaves. 726 This is consistent with observations from Yi et al., where they report _ereb41_ and _ereb161_ single 727 mutants have a modest decrease in plant height and leaf width (Yi _et al._, 2026). All maize 728 mutants described to date which impact vascular development have pleiotropic developmental 729 phenotypes, suggesting that vascular development is intrinsically linked to leaf patterning. In the 730 future, employing tissue-specific CRISPR of _EREBs_ driven by procambial promoters may 731 provide a way to unlink their impact on maize vascular patterning from other aspects of leaf 732 development. 733 **Transverse veins mature into fully integrated vascular bundles** 734 The final stage of TV ontogeny is the establishment of complete vascular continuity with the pre735 existing longitudinal vascular system. By P9, TVs contained all major vascular cell types, 736 including xylem, phloem, metaxylem, metaphloem, and bundle-sheath tissues, and established [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 737 direct continuity with neighboring longitudinal veins in both blade and sheath domains. These 738 observations indicate that TV differentiation culminates in the formation of fully functional 739 vascular bundles that become integrated into the transport network of the leaf. This conclusion is 740 consistent with classical descriptions of grass cross veins, which characterized these structures as 741 complete vascular bundles that contribute to lateral transport between adjacent longitudinal veins 742 (Blackman, 1971; Bosabalidis _et al._, 1994). 743 The developmental timing of this integration is particularly noteworthy because TVs arise only 744 after neighboring longitudinal veins have already differentiated. Classical studies in wheat, 745 barley, rice, and maize demonstrated that longitudinal veins establish the primary vascular 746 scaffold before transverse or commissural veins differentiate and connect adjacent bundles 747 (Blackman, 1971; Dannenhoffer & Evert, 1994; Matsukura _et al._, 2000; Sakaguchi & Fukuda, 748 2008). In rice, commissural veins originate adjacent to differentiated vascular tissues and 749 subsequently extend toward neighboring bundles, suggesting that later-forming vascular 750 connections arise within an already organized transport system (Matsukura _et al._, 2000; 751 Sakaguchi & Fukuda, 2008). Our observations place maize TVs within this broader 752 developmental framework and provide cellular evidence of how complete vascular continuity is 753 established during their maturation. 754 A particularly striking feature of mature TVs is the preservation of vascular polarity throughout 755 the transverse connection. Transverse xylem elements connected specifically to metaxylem, 756 whereas transverse phloem joined directly to metaphloem. The characteristic adaxial xylem and 757 abaxial phloem organization of grass vascular bundles was therefore maintained across the entire 758 junction. These observations indicate that TV maturation involves precise alignment of 759 equivalent vascular domains. The maintenance of such polarity is consistent with classical 760 anatomical descriptions of grass vascular architecture and suggests that differentiating TVs 761 remain coordinated with positional information already established in adjacent longitudinal veins 762 (Blackman, 1971; Bosabalidis _et al._, 1994). 763 Integration was also evident at the level of the bundle sheath. Classical anatomical studies 764 established that transverse veins form vascular connections between neighboring longitudinal 765 veins through xylem and phloem continuity. Our observations extend these descriptions by 766 showing that bundle sheath tissue is also continuous across the transverse-longitudinal junction, 767 indicating that TV integration involves the coordinated incorporation of both conductive and 768 associated vascular tissues. The bundle sheath surrounding the transverse strand was continuous 769 with that of neighboring longitudinal veins, indicating that TV maturation coordinates the 770 differentiation of the entire vascular bundle. Because bundle sheath development is tightly 771 associated with vascular development in grasses (Bosabalidis _et al._, 1994; Kumar & Kellogg, 772 2019), the continuity of this tissue suggests that TV formation proceeds through coordinated 773 patterning of conductive and associated cell types. Similar conclusions were reached from 774 comparative studies of monocot vascular anatomy, which emphasized that vascular integration 775 involves both conductive tissues and their surrounding bundle sheath cells (Dute _et al._, 2007). 776 The anatomy of mature TVs further suggests a high degree of developmental coordination during 777 network assembly. (Sakaguchi & Fukuda, 2008) proposed that commissural veins in rice are 778 initiated adjacent to mature metaxylem and phloem tissues and may therefore respond to [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 779 developmental signals originating from established vascular bundles. Our observations are 780 consistent with this interpretation. Because TVs develop only after neighboring veins are already 781 differentiated, successful integration requires newly differentiating vascular elements to align 782 precisely with existing transport tissues. The direct continuity observed between transverse and 783 longitudinal xylem, phloem, and bundle-sheath tissues suggests that mature vascular bundles 784 contribute positional information that guides incorporation of later-forming vascular connections. 785 The direct connection between transverse xylem and longitudinal metaxylem requires local 786 interruption of the sclerenchyma ring, demonstrating that TV integration involves coordinated 787 differentiation of both conductive and supportive tissues. Such coordination allows new vascular 788 connections to be added to an already functioning transport network while preserving its overall 789 structural organization. 790 Viewed together with the developmental and molecular observations described above, these 791 anatomical findings suggest that TVs undergo the same broad progression observed for other 792 vascular tissues: specification of vascular identity precedes visible differentiation, followed by 793 tissue maturation and integration into the surrounding network (Gardiner _et al._, 2011; Krishna _et_ 794 _al._, 2021). 795 Collectively, our results identify TVs as a distinct developmental component of the maize 796 vascular system, characterized by coordinated specification, assembly, and integration into the 797 leaf-wide vascular network (Fig. 7). Early PIN1a accumulation marks nascent TV domains, 798 recruitment and anticlinal division generate the procambial strand, and subsequent activation of 799 _DR5_ accompanies of the developing vascular trace. Ultimately, mature TVs establish xylem-to800 xylem, phloem-to-phloem, and bundle-sheath continuity with neighboring longitudinal veins, 801 fully integrating into the vascular network. Together, these observations support a developmental 802 model in which paired initials specify the future TV domain, intervening ground cells are 803 recruited and incorporated into the developing strand, and subsequent differentiation generates 804 fully functional transverse vascular bundles interconnected with the pre-existing longitudinal 805 vein system **(Fig.7)** . 806 807 **ACKNOWLEDGEMENTS** 808 MCM, PLL, and SL were supported by NSF grant IOS-2211434. SL was additionally supported 809 by USDA-NIFA grant 2023-67013-44037. PLL was additionally supported by the Oregon State 810 University Botany and Plant Pathology department. DR is supported by USDA National Institute 811 of Food and Agriculture fellowship 2024-67011-42999. Work at ISU (E.V., E.U.W. and L.F.G) 812 was supported by USDA National Institute of Food and Agriculture Hatch project number 813 IOW03649 and by the ISU Crop Bioengineering Center. Confocal imaging was performed in the 814 Oregon State University Center for Quantitative Life Sciences imaging core. We wish to thank 815 Anne-Marie Girard-Pohjanpelto (Oregon State University) for extensive help with image 816 collection and management of the equipment. We extend thanks to the Oregon State University 817 Greenhouse Operations staff for the management of our plants. The DR5-RFP and Pin-YFP lines 818 were provided by David Jackson (Cold Spring Harbor Laboratory). 819 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 820 **AUTHOR CONTRIBUTIONS** 821 MCM - conceptualization, preparing figures, writing manuscript, all experiments except as 822 listed, LFG and EUW- generation of _ereb_ mutants, DR - transverse vein phenotyping and 823 quantification in _ereb_ mutants, EV - generation of _ereb_ mutants, PLL - RNAseq analysis, 824 preparing figures, writing manuscript, SL - conceptualization, RNAseq analysis, preparing 825 figures, writing manuscript 826 827 **SUPPLEMENTARY MATERIAL** 828 **Supplemental Figure 1. Transverse vein phenotyping overview.** 829 **Supplemental Figure 2. Auxin transport inhibition alters initial cell patterning.** 830 **Supplemental Figure 3. Auxin signaling is activated during transverse vein differentiation.** 831 **Supplemental Figure 4. Expression Time Units (ETUs) were strongly positively correlated** 832 **with leaf primordia size, R** **[2]** **=0.82.** 833 **Supplemental Figure 5.** _**ereb41;114**_ **transposon double mutants have fewer transverse veins** 834 **in the blade.** 835 **Supplemental Figure 6.** _**ereb41;114**_ **CRISPR double mutants make fewer transverse veins** 836 **than WT.** 837 **Supplemental Figure 7.** _**ereb41;ereb114**_ **transposon double mutants had narrower leaves** 838 **than WT plants.** 839 **Supplemental Table 1. GO enrichment for co-expressed clusters.** 840 **Supplemental Table 2. Genes in each co-expression cluster.** 841 842 **FUNDING** 843 MCM, PLL, and SL were supported by NSF grant IOS-2211434. SL was additionally supported 844 by USDA-NIFA grant 2023-67013-44037. DR is supported by USDA National Institute of Food 845 and Agriculture fellowship 2024-67011-42999. E.V., E.U.W. and L.F.G were supported by 846 USDA National Institute of Food and Agriculture Hatch project number IOW03649. 847 848 **CONFLICTS OF INTEREST** 849 The authors have no conflicts of interest to report. 850 851 **DATA AVAILABILITY STATEMENT** 852 The RNAseq data are deposited in BioProject ID: PRJNA1549843. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 853 854 **REFERENCES** 855 **Abu-Jamous B, Kelly S** . **2018** . Clust: automatic extraction of optimal co-expressed gene 856 clusters from gene expression data. _Genome Biology_ **19** . 857 **Ahern, K. R., Deewatthanawong, P., Schares, J., Muszynski, M., Weeks, R., Vollbrecht, E.,** 858 **Duvick, J., Brendel, V. P., and Brutnell, T. P.** **2009.** Regional mutagenesis using _Dissociation_ 859 in maize. _Methods_ **49** :248–254. 860 861 **Bassukas AEL, Xiao Y, Kim N, Schwechheimer C** . **2026** . PAX-regulated self-organizing 862 polarity drives procambial strand formation. _Nature Plants_ . 863 **Bennett T** . **2015** . PIN proteins and the evolution of plant development. _Trends in Plant Science_ 864 **20** : 498–507. 865 **Blackman E** . **1971** . The morphology and development of cross veins in the leaves of bread 866 wheat ( _Triticum aestivum_ L.). _Ann Bot_ **35** : 653–65. 867 **Bosabalidis AM, Evert RF, Russin WA** . **1994** . Ontogeny of the vascular bundles and 868 contiguous tissues in the maize leaf blade. _Source: American Journal of Botany_ **81** : 745–752. 869 **Chang, W., Cheng, J., Allaire, J. J., Sievert, C., Schloerke, B., Aden-Buie, G., Xie, Y., Allen,** 870 **J., McPherson, J., Dipert, A., & Borges, B. 2026.** shiny: Web Application Framework for R. 871 https://shiny.posit.co/ 872 **Chen, S., Zhou, Y., Chen, Y., & Gu, J. 2018.** fastp: an ultra-fast all-in-one FASTQ 873 preprocessor. Bioinformatics, 34(17), i884–i890. https://doi.org/10.1093/bioinformatics/bty560 874 **Christin P-A, Osborne CP, Chatelet DS, Columbus JT, Besnard G, Hodkinson TR,** 875 **Garrison LM, Vorontsova MS, Edwards EJ** . **2013** . Anatomical enablers and the evolution of C 876 4 photosynthesis in grasses. _Proceedings of the National Academy of Sciences of the United_ 877 _States of America_ **110** : 1381–1386. 878 **Dannenhoffer JM, Evert RF** . **1994** . Development of the vascular system in the leaf of barely 879 ( _Hordeum vulgare_ L.). _Int. J. Plant Sci_ **155** : 143–157. 880 **Dickinson AJ, Zhang J, Luciano M, Wachsman G, Sandoval E, Schnermann M, Dinneny** 881 **JR, Benfey PN** . **2021** . A plant lipocalin promotes retinal-mediated oscillatory lateral root 882 initiation. _Science_ **373** : 1532–1536. 883 **Donner TJ, Sherr I, Scarpella E** . **2009** . Regulation of preprocambial cell state acquisition by 884 auxin signaling in Arabidopsis leaves. _Development_ **136** : 3235–3246. 885 **Dute RR, Jackson BE, Adkins RD, Folkerts DR** . **2007** . Anatomy of the laminar organs of 886 _Commelina erecta_ (Commelinaceae). _Southeastern Naturalist_ **6** : 47–66. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 887 **Engler J. 2026.** tidyheatmaps: Heatmaps from Tidy Data. R package version 0.2.1.9000 888 **Gallavotti, A., Yang, Y., Schmidt, R. J., and Jackson, D.** (2008). The Relationship between 889 Auxin Transport and Maize Branching. _Plant Physiol_ **147** :1913–1923. 890 891 **Gardiner J, Donner TJ, Scarpella E** . **2011** . Simultaneous activation of _SHR_ and _ATHB8_ 892 expression defines switch to preprocambial cell state in Arabidopsis leaf development. 893 _Developmental Dynamics_ **240** : 261–270. 894 **Ge, S. X., Jung, D., and Yao, R. 2020.** ShinyGO: a graphical gene-set enrichment tool for 895 animals and plants. _Bioinformatics_ **36** :2628–2629. 896 **Holloway DM, Eiriksson TK, Wenzel CL** . **2025** . The role of auxin transport through 897 plasmodesmata in leaf vein canalization and patterning. _Frontiers in Plant Science_ **16** . 898 **Horstman A, Willemsen V, Boutilier K, Heidstra R** . **2014** . AINTEGUMENTA-LIKE proteins: 899 hubs in a plethora of networks. _Trends in Plant Science_ **19** : 146–157. 900 **Hughes TE, Sedelnikova O, Thomas M, Langdale JA** . **2023** . Mutations in _NAKED-_ 901 _ENDOSPERM_ IDD genes reveal functional interactions with _SCARECROW_ during leaf 902 patterning in C4 grasses. _PLOS Genetics_ **19** : e1010715. 903 **Hughes TE, Sedelnikova O V., Wu H, Becraft PW, Langdale JA** . **2019** . Redundant 904 _SCARECROW_ genes pattern distinct cell layers in roots and leaves of maize. _Development_ **146** . 905 **Jankovsky JP, Smith LG, Nelson T** . **2001** . Specification of bundle sheath cell fates during 906 maize leaf development: roles of lineage and positional information evaluated through analysis 907 of the tangled1 mutant. _Development_ **128** : 2747–2753. 908 **Kang, J., Mizukami, Y., Wang, H., Fowke, L., and Dengler, N. G.** **2007.** Modification of cell 909 proliferation patterns alters leaf vein architecture in Arabidopsis thaliana. _Planta_ **226** :1207–1218. 910 911 **Kim, D., Paggi, J. M., Park, C., Bennett, C., & Salzberg, S. L. 2019.** Graph-based genome 912 alignment and genotyping with HISAT2 and HISAT-genotype. Nature Biotechnology, 37(8), 913 907–915. https://doi.org/10.1038/s41587-019-0201-4 914 **Krishna A, Gardiner J, Donner TJ, Scarpella E** . **2021** . Control of vein-forming, striped gene 915 expression by auxin signaling. _BMC Biology_ **19** . 916 **Krizek, B. A., Bantle, A. T., Heflin, J. M., Han, H., Freese, N. H., and Loraine, A. E.** **2021.** 917 AINTEGUMENTA and AINTEGUMENTA-LIKE6 directly regulate floral homeotic, growth, 918 and vascular development genes in young Arabidopsis flowers. _J Exp Bot_ **72** :5478–5493. 919 920 **Kumar D, Kellogg EA** . **2019** . Getting closer: vein density in C4 leaves. _New Phytologist_ **221** : 921 1260–1267. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 922 **Leiboff, S., and Hake, S. 2019.** Reconstructing the Transcriptional Ontogeny of Maize and 923 Sorghum Supports an Inverse Hourglass Model of Inflorescence Development. _Current Biology_ 924 **29** :3410-3419.e3. 925 926 **Liao, Y., Smyth, G. K., & Shi, W. 2014.** featureCounts: an efficient general purpose program 927 for assigning sequence reads to genomic features. Bioinformatics. **30** (7), 923–930. 928 https://doi.org/10.1093/bioinformatics/btt656 929 930 **Linh NM, Scarpella E** . **2022** . Leaf vein patterning is regulated by the aperture of 931 plasmodesmata intercellular channels. _PLoS Biology_ **20** . 932 **Liu W-Y, Lin H-H, Yu C-P, Chang C-K, Chen H-J, Lin J-J, Lu M-YJ, Tu S-L, Shiu S-H,** 933 **Wu S-H,** _**et al.**_ **2020** . Maize ANT1 modulates vascular development, chloroplast development, 934 photosynthesis, and plant growth. _Proceedings of the National Academy of Sciences_ **117** : 21747– 935 21756. 936 **Love, M. I., Huber, W., & Anders, S. 2014.** Moderated estimation of fold change and 937 dispersion for RNA-seq data with DESeq2. Genome Biology, **15** (12), 550. 938 https://doi.org/10.1186/s13059-014-0550-8 939 **Lv Z, Wang N, Ai S, Zhang H, Lin S** . **2026** . Identification and functional validation of key 940 regulatory genes for leaf development in moso bamboo ( _Phyllostachys edulis_ ) based on 941 transcriptome data. _Plants_ **15** . 942 **Marcos D, Berleth T** . **2014** . Dynamic auxin transport patterns preceding vein formation 943 revealed by live-imaging of arabidopsis leaf primordia. _Frontiers in Plant Science_ **5** . 944 **Matsukura C, Kawai M, Toyofuku K, Barrero RA, Uchimiya H, Yamaguchi J** . **2000** . 945 Transverse vein differentiation associated with gas space formation-fate of the middle cell layer 946 in leaf sheath development of rice. _Annals of Botany_ **85** : 19–27. 947 **Mazur E, Kulik I, Hajný J, Friml J** . **2020** . Auxin canalization and vascular tissue formation by 948 TIR1/AFB-mediated auxin signaling in Arabidopsis. _New Phytologist_ **226** : 1375–1383. 949 **Nelson T, Dengler N** . **1997** . Leaf vascular pattern formation. _The Plant Cell_ **9** : 1121–1122. 950 **O’Connor DL, Runions A, Sluis A, Bragg J, Vogel JP, Prusinkiewicz P, Hake S** . **2014** . A 951 division in PIN-mediated auxin patterning during organ initiation in grasses. _PLoS_ 952 _Computational Biology_ **10** . 953 **Perico C, Tan S, Langdale JA** . **2022** . Developmental regulation of leaf venation patterns: 954 monocot versus eudicots and the role of auxin. _New Phytologist_ **234** : 783–803. 955 **Perico C, Zaidem M, Sedelnikova O, Bhattacharya S, Korfhage C, Langdale JA** . **2024** . 956 Multiplexed in situ hybridization reveals distinct lineage identities for major and minor vein [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 957 initiation during maize leaf development. _Proceedings of the National Academy of Sciences of_ 958 _the United States of America_ **121** . 959 **Petrášek J, Friml J** . **2009** . Auxin transport routes in plant development. _Development_ **136** : 960 2675–2688. 961 **Pinon V, Prasad K, Grigg SP, Sanchez-Perez GF, Scheres B** . **2013** . Local auxin biosynthesis 962 regulation by PLETHORA transcription factors controls phyllotaxis in Arabidopsis. _Proceedings_ 963 _of the National Academy of Sciences of the United States of America_ **110** : 1107–1112. 964 **Prigge MJ, Morffy N, de Neve A, Szutu W, Abraham-Juárez MJ, McAllister T, Jones H,** 965 **Johnson K, Do N, Lavy M,** _**et al.**_ **2025** . Comparative mutant analyses reveal a novel mechanism 966 of ARF regulation in land plants. _Nature Plants_ **11** : 821–835. 967 **Robil JM, McSteen P** . **2023** . Hormonal control of medial–lateral growth and vein formation in 968 the maize leaf. _New Phytologist_ **238** : 125–141. 969 **Robil JM, Scarpella E** . **2026** . Origin and diversity of leaf vein patterns. _Journal of Experimental_ 970 _Botany_ **77** : 3485–3500. 971 **Ruzin SE** . **1999** . _Plant microtechnique and microscopy_ . Oxford, New York: Oxford University 972 Press. 973 **Sakaguchi J, Fukuda H** . **2008** . Cell differentiation in the longitudinal veins and formation of 974 commissural veins in rice ( _Oryza sativa_ ) and maize ( _Zea mays_ ). _Journal of Plant Research_ **121** : 975 593–602. 976 **Sakaguchi J, Itoh JI, Ito Y, Nakamura A, Fukuda H, Sawa S** . **2010** . COE1, an LRR-RLK 977 responsible for commissural vein pattern formation in rice. _Plant Journal_ **63** : 405–416. 978 **Scanlon MJ** . **2003** . The polar auxin transport inhibitor N-1-Naphthylphthalamic acid disrupts 979 leaf initiation, KNOX protein regulation, and formation of leaf margins in maize. _Plant_ 980 _Physiology_ **133** : 597–605. 981 **Scarpella E** . **2024** . Leaf vein patterning. _Annual Review of Plant Biology_ **75** : 377–398. 982 **Schuetz M, Fidanza M, Mattsson J** . **2019** . Identification of auxin response factor-encoding 983 genes expressed in distinct phases of leaf vein development and with overlapping functions in 984 leaf formation. _Plants_ **8** . 985 **Sharman BC** . **1942** . Developmental anatomy of the shoot of _Zea mays_ L. _Annals of Botany_ **6** : 986 245–282. 987 **Sieburth LE** . **1999** . _Auxin Is Required for Leaf Vein Pattern in Arabidopsis 1_ . 988 **de Smet I** . **2012** . Lateral root initiation: One step at a time. _New Phytologist_ **193** : 867–873. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 989 **Strable, J., Wallace, J. G., Unger-Wallace, E., Briggs, S., Bradbury, P. J., Buckler, E. S.,** 990 **and Vollbrecht, E.** **2017.** Maize YABBY Genes drooping leaf1 and drooping leaf2 Regulate 991 Plant Architecture. _Plant Cell_ **29** :1622–1641. 992 993 **Strable J, Nelissen H** . **2021** . The dynamics of maize leaf development: Patterned to grow while 994 growing a pattern. _Current Opinion in Plant Biology_ **63** . 995 **Tsiantis M, Brown MIN, Skibinski G, Langdale JA** . **1999** . _Disruption of Auxin Transport Is_ 996 _Associated with Aberrant Leaf Development in Maize 1_ . 997 **Ueno O, Kawano Y, Wakayama M, Takeda T** . **2006** . Leaf vascular systems in C3 and C4 998 grasses: a two-dimensional analysis. _Annals of Botany_ **97** : 611–621. 999 **Vlad D, Zaidem M, Perico C, Sedelnikova O, Bhattacharya S, Langdale JA** . **2024** . The 1000 WIP6 transcription factor _TOO MANY LATERALS_ specifies vein type in C4 and C3 grass leaves. 1001 _Current Biology_ **34** : 1670-1686.e10. 1002 **Vollbrecht, E., Duvick, J., Schares, J. P., Ahern, K. R., Deewatthanawong, P., Xu, L.,** 1003 **Conrad, L. J., Kikuchi, K., Kubinec, T. A., Hall, B. D., et al.** **2010.** Genome-Wide 1004 Distribution of Transposed Dissociation Elements in Maize. _Plant Cell_ **22** :1667–1685. 1005 1006 **Wenzel CL, Schuetz M, Yu Q, Mattsson J** . **2007** . Dynamics of _MONOPTEROS_ and PIN1007 FORMED1 expression during leaf vein pattern formation in Arabidopsis thaliana. _Plant Journal_ 1008 **49** : 387–398. 1009 **Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan, L., François, R., Grolemund,** 1010 **G., Hayes, A., Henry, L., Hester, J., Kuhn, M., Pedersen, T., Miller, E., Bache, S., Müller,** 1011 **K., Ooms, J., Robinson, D., Seidel, D., Spinu, V., … Yutani, H. 2019.** Welcome to the 1012 tidyverse. Journal of Open Source Software, **4** (43), 1686. https://doi.org/10.21105/joss.01686 1013 **Wickham, H., Navarro, D., & Pedersen, T. L. n.d.** ggplot2: Elegant Graphics for Data 1014 Analysis (3e). Retrieved October 23, 2026, from https://ggplot2-book.org/ 1015 **Xuan W, De Gernier H, Beeckman T** . **2020** . The dynamic nature and regulation of the root 1016 clock. _Development_ **147** . 1017 **Yi J, Chen Y, Tian S, Vlad D, Yang J, Sedelnikova O, Zhao C, Su H, Dong W, Zhang Y,** _**et**_ 1018 _**al.**_ **2026** . Single-cell and tissue transcriptomes unveil factors regulating vein initiation and 1019 development in grass leaf primordia. _The Plant Cell_ **38** . 1020 1021 **FIGURE LEGENDS** 1022 **Figure 1. Transverse veins integrate longitudinal veins into a continuous vascular network** 1023 **despite asynchronous development. (a, b)** Longitudinal sections showing provascular strands 1024 of transverse veins connecting more differentiated longitudinal veins in the blade **(a)** and the [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1025 sheath **(b)** . **(c, d)** Cross‐sections showing two rows of provascular cells forming the transverse 1026 vein in the blade **(c)** and the sheath **(d)** . **(e–l)** Fully differentiated transverse veins. **(e, f, k, l)** 1027 Cross-sections and **(k, l)** polarized light images. **(g–j)** Longitudinal sections. **(e, f)** Cross-sections 1028 showing all vascular bundle cell types interconnected between transverse and longitudinal veins 1029 in the blade **(e)** and the sheath **(f),** showing conservation of vascular organization and polarity 1030 despite differences in interveinal spacing; transverse vein xylem exhibits reticulate secondary 1031 wall thickening. **(g, h)** Complete vascular continuity between transverse and longitudinal veins in 1032 the blade **(g)** and the sheath **(h)** . **(i)** Transverse vein xylem (Xy) connected the metaxylem of the 1033 longitudinal vein (LV-MXy) in the sheath. **(j)** Transverse vein phloem (Ph) connected to the 1034 metaphloem of the longitudinal vein (LV-MPh) in the sheath. **(k, l, m)** Polarized light images 1035 showing mature transverse vein xylem with reticulate secondary wall thickening, interrupting the 1036 sclerenchyma ring and connecting to the metaxylem of the longitudinal vein (LV-MXy) in the 1037 blade **(k, l)** and the sheath **(m)** . Serial sections illustrate a continuous sclerenchyma ring 1038 (asterisks) surrounding the longitudinal vein in **(k)** and the region where a transverse vein xylem 1039 strand (Xy) interrupts this ring to connect with the longitudinal vein metaxylem in **(l).** Arrow, 1040 transverse vein; LV: Longitudinal vein; Ph, phloem; Xy, xylem; LV-MPh, metaphloem of the 1041 longitudinal vein; LV-MXy, metaxylem of the longitudinal vein; BS, bundle sheath. 1042 1043 **Figure 2. Transverse veins develop sequentially in a basipetal pattern during maize leaf** 1044 **growth. (a)** Schematic and cleared leaf primordia illustrating vascular pattern establishment 1045 from plastochron (P)1 to P8. Midvein and lateral veins initiate acropetally from P1 to P3, 1046 followed by intermediate veins at P5 in a basipetal sequence. Transverse veins integrate the 1047 network beginning around P6, but the precise sites and mechanism of their initiation remain 1048 unresolved (question marks indicate unknown aspects). **(b–e)** Confocal images of leaf primordia 1049 expressing _PIN1a-YFP_ . **(b)** At P5, parallel veins are still forming through ongoing provascular 1050 specification. **(c)** At P6, transverse vein provascular traces appear at the tip (arrows). Earliest 1051 initiation involves paired initial cells from adjacent parallel vein procambium (asterisks), with no 1052 PIN1a-YFP signal in intervening ground cells. **(d)** At P7, _PIN1a-YFP_ expression emerges in 1053 ground cells between initials (arrow). Absence of initials at the base supports basipetal 1054 progression. **(e)** Epifluorescence image of a leaf primordia expressing _PIN1a-YFP_ . At P8, 1055 transverse veins are present throughout the blade but do not extend beyond the anastomoses of 1056 the intermediate and small veins, indicated by arrowheads. Transverse veins also initiate within 1057 the sheath, indicated by arrows. The auricle is not yet morphologically distinct and corresponds 1058 to the pre‑auricle region (P‑A). **(f)** At P9, transverse veins form a flexible ladder‑like pattern 1059 (arrows) as the leaf expands. The auricle is now distinguishable and expanding in the region 1060 between the anastomosis zone (arrowheads), and the ligule, which does not contain transverse 1061 veins. Arrow, transverse vein; Arrowhead, anastomosis; asterisk, initials; P-A, pre-auricle. 1062 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1063 **Figure 3. Sequential recruitment and periclinal division of ground cells drive transverse** 1064 **vein formation in maize.** All panels show leaf primordia expressing _PIN1a-YFP_ . **(a)** In the 1065 middle region of the leaf primordium, developing transverse veins show ground cell recruitment 1066 occurs as a single layer positioned between adjacent parallel veins. The number of cells in this 1067 layer varies with interveinal spacing: one to three cells in the blade, and five or more in the 1068 sheath where spacing is wider due to the absence of small veins (arrowheads). **(b)** In the leaf tip, 1069 where transverse veins are more developmentally advanced due to basipetal progression, the 1070 recruited cell layer undergoes cell division, resulting in an additional layer of ground cells 1071 (arrowheads). **(c)** Time-course imaging of transverse vein initiation in the P6 leaf primordia. 1072 Ground cell recruitment occurs within 0–12 h, followed by cell division between 12–24 h, 1073 producing two layers (arrowheads). Colored overlays highlight 3D segmentation of individual 1074 cells tracked over time. Arrowhead, ground cell recruited; Asterisks, initials. 1075 1076 **Figure 4. Transverse vein formation persists under auxin transport inhibition but shows** 1077 **altered patterning.** Time-course imaging of P6 leaf primordia expressing _PIN1a-YFP_ under 1078 control conditions and after auxin transport inhibitor treatments. Recruitment of ground cells 1079 typically occurs within the first 12 h, followed by cell division between 12–24 h (arrowheads). 1080 Colored overlays highlight 3D segmentation of individual cells tracked over time. **(a)** Control: 1081 Ground cells are recruited by 12 h and divide by 24 h, forming two layers. **(b)** At IAA 10 μM, 1082 both recruitment and cell division occurred earlier than in control between 0 -12 h. **(c–d)** Under 1083 NPA at 10 μM, recruitment occurs similarly to control (c), but cases without subsequent division 1084 were also observed (d). **(e)** At NPA 30 μM, recruitment and division occurs but exhibit altered 1085 morphology. **(f)** Under TIBA treatment, recruitment and division occur, but recruited cells 1086 display irregular shapes. Scale bar: 20µm. 1087 1088 **Figure 5. Transcriptomic profiling of transverse vein development using a substage‑based** 1089 **framework. (a)** Maize leaf primordia from P1 to P8 subdivided into substages (P6.0–P6.3 and 1090 P7.1–P7.3) based on morphological landmarks. The accompanying table summarizes the 1091 presence of transverse veins (TVs), diagnostic morphological traits, and leaf length ranges 1092 characteristic of each substage. **(b)** Confocal images illustrating P6.2 and P6.3 primordia 1093 dissected into upper regions (with TVs) and lower regions (without TVs) to capture spatial 1094 differences in transverse vein development. Arrow, trichomes; dotted vertical bar, TV‑containing 1095 region; L, ligule. **(c, d)** Principal Component Analysis (PCA) of leaf primordia transcriptome 1096 dataset. Samples are colored by plastochron stage **(c)** or leaf primordia segment collected **(d)** . **(e)** 1097 Leaf primordia data was best described by a five-knot spline. Bottom panel, Expression Time 1098 Units (ETUs) assigned to each leaf primordia stage. **(f)** Variance stabilized counts of _PIN1a_ 1099 assayed either by ETU (top panel) or morphological stage (bottom panel). **(g)** Heatmap showing 1100 relative expression of _PIN_ genes expressed during the leaf primordia stages assayed. 1101 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1102 **Figure 6. Clustering analysis of leaf primordia time-course RNAseq reveals regulators of** 1103 **transverse vein development. (a)** Clustering analysis of leaf primordia RNAseq. Using Clust, 1104 RNAseq data was subdivided into 11 clusters. Significant Biological Process (BP) GO terms are 1105 listed beneath each cluster. Green clusters have significant GO terms related to photosynthesis, 1106 yellow cluster is related to carbohydrate transport, purple clusters are related to cell cycle and 1107 gene expression, blue clusters are related to RNA regulation, and the orange cluster had diverse 1108 GO term enrichment. n= number of genes in a cluster. **(b)** Heatmap showing _PIN1a_, _EREB41_, 1109 and _EREB114_ expression by leaf primordia stage. _EREB41_ and _EREB114_ expression preceded 1110 _PIN1a_ expression. Heatmap coded by FPKM values, with warmer values representing higher 1111 expression. **(c)** _ereb114;ereb41_ double mutants had reduced transverse veins relative to the WT 1112 control W22. Images are of the sheath. Arrowheads, normal transverse veins. Scale bar, 100 µm. 1113 **(d)** Number of transverse veins/cm [2 ] in WT, _ereb41_, _ereb114_, and _ereb41;114_ mutants. n=8, 1114 different letters indicate statistically significant difference using one-way ANOVA with _post hoc_ 1115 Sidak test. Measurements are from the leaf sheath. **(e)** Number of regular and irregular transverse 1116 veins in WT, _ereb41_, and _ereb41;114_ mutants, n=8. 1117 1118 **Figure 7.** **Model of transverse vein (TV) formation in maize leaves.** _Top:_ Cellular progression 1119 of TV development. **(1)** Adjacent longitudinal veins are established and express _PIN1a_ prior to 1120 TV initiation. **(2)** A putative initiation signal is proposed to specify paired TV initials adjacent to 1121 neighboring longitudinal veins. **(3)** PIN1a accumulates in the initials. **(4)** Intervening ground 1122 cells are progressively recruited and begin expressing _PIN1a_ . **(5)** Recruited cells undergo 1123 coordinated anticlinal divisions, generating a continuous ladder-like pre-procambial strand 1124 connecting adjacent longitudinal veins. **(6)** _DR5_ signal is first detected in the initials. **(7)** _DR5_ 1125 activity expands throughout the developing procambial strand. **(8)** Mature TVs form fully 1126 differentiated vascular bundles that retain the same adaxial-abaxial organization as longitudinal 1127 veins, with xylem-to-xylem and phloem-to-phloem continuity between the two vein systems. 1128 _Bottom:_ Developmental progression of TV formation during maize leaf ontogeny. The midvein 1129 and longitudinal veins are initiated sequentially during leaf development, establishing the 1130 primary vascular scaffold of the leaf. TV initiation begins at approximately P6 and progresses 1131 basipetally through the blade during P7, while the sheath begins to elongate. By P8, TV 1132 formation is complete in the blade and continues through the sheath, showing the characteristic 1133 ladder-like vascular network. By P9, TVs are present throughout both blade and sheath domains 1134 and have differentiated into mature vascular bundles integrated into the longitudinal vein 1135 network. As leaf growth continues, transverse veins become increasingly oblique relative to the 1136 longitudinal veins. The ligule and auricle develop after TV establishment. Numbers above each 1137 leaf primordium indicate the cellular stages illustrated in the model on the top panel. Colors 1138 correspond to cell identities, reporter expression domains, and vein classes shown in the keys. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1139 1140 **Figure 1. Transverse veins integrate longitudinal veins into a continuous vascular network** 1141 **despite asynchronous development. (a, b)** Longitudinal sections showing provascular strands 1142 of transverse veins connecting more differentiated longitudinal veins in the blade **(a)** and the 1143 sheath **(b)** . **(c, d)** Cross‐sections showing two rows of provascular cells forming the transverse 1144 vein in the blade **(c)** and the sheath **(d)** . **(e–l)** Fully differentiated transverse veins. **(e, f, k, l)** 1145 Cross-sections and **(k, l)** polarized light images. **(g–j)** Longitudinal sections. **(e, f)** Cross-sections 1146 showing all vascular bundle cell types interconnected between transverse and longitudinal veins 1147 in the blade **(e)** and the sheath **(f),** showing conservation of vascular organization and polarity [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1148 despite differences in interveinal spacing; transverse vein xylem exhibits reticulate secondary 1149 wall thickening. **(g, h)** Complete vascular continuity between transverse and longitudinal veins in 1150 the blade **(g)** and the sheath **(h)** . **(i)** Transverse vein xylem (Xy) connected the metaxylem of the 1151 longitudinal vein (LV-MXy) in the sheath. **(j)** Transverse vein phloem (Ph) connected to the 1152 metaphloem of the longitudinal vein (LV-MPh) in the sheath. **(k, l, m)** Polarized light images 1153 showing mature transverse vein xylem with reticulate secondary wall thickening, interrupting the 1154 sclerenchyma ring and connecting to the metaxylem of the longitudinal vein (LV-MXy) in the 1155 blade **(k, l)** and the sheath **(m).** Serial sections illustrate a continuous sclerenchyma ring 1156 (asterisks) surrounding the longitudinal vein in **(k)** and the region where a transverse vein xylem 1157 strand (Xy) interrupts this ring to connect with the longitudinal vein metaxylem in **(l)** . Arrow, 1158 transverse vein; LV: Longitudinal vein; Ph, phloem; Xy, xylem; LV-MPh, metaphloem of the 1159 longitudinal vein; LV-MXy, metaxylem of the longitudinal vein; BS, bundle sheath. 1160 1161 1162 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1163 1164 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1179 morphologically distinct and corresponds to the pre‑auricle region (P‑A). **(f)** Epifluorescence 1180 image of a leaf primordia expressing _PIN1a-YFP_ . At P9, transverse veins form a flexible 1181 ladder‑like pattern (arrows) as the leaf expands. The auricle is now distinguishable and 1182 expanding in the region between the anastomosis zone (arrowheads), and the ligule, which does 1183 not contain transverse veins. Arrow, transverse vein; Arrowhead, anastomosis; asterisk, initials; 1184 P-A, pre-auricle. 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1195 1208 1209 1210 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1211 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1217 Ground cells are recruited by 12 h and divide by 24 h, forming two layers. **(b)** At IAA 10 μM, 1218 both recruitment and cell division occurred earlier than in control between 0 -12 h. **(c–d)** Under 1219 NPA at 10 μM, recruitment occurs similarly to control (c), but cases without subsequent division 1220 were also observed (d). **(e)** At NPA 30 μM, recruitment and division occurs but exhibit altered 1221 morphology. **(f)** Under TIBA treatment, recruitment and division occur, but recruited cells 1222 display irregular shapes. Scale bar: 20µm. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1223 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1226 P7.1–P7.3) based on morphological landmarks. The accompanying table summarizes the 1227 presence of transverse veins (TVs), diagnostic morphological traits, and leaf length ranges 1228 characteristic of each substage. **(b)** Confocal images illustrating P6.2 and P6.3 primordia 1229 dissected into upper regions (with TVs) and lower regions (without TVs) to capture spatial 1230 differences in transverse vein development. Arrow, trichomes; dotted vertical bar, TV‑containing 1231 region; L, ligule. **(c, d)** Principal Component Analysis (PCA) of leaf primordia transcriptome 1232 dataset. Samples are colored by plastochron stage **(c)** or leaf primordia segment collected **(d)** . **(e)** 1233 Leaf primordia data was best described by a five-knot spline. Bottom panel, Expression Time 1234 Units (ETUs) assigned to each leaf primordia stage. **(f)** Variance stabilized counts of _PIN1a_ 1235 assayed either by ETU (top panel) or morphological stage (bottom panel). **(g)** Heatmap showing 1236 relative expression of _PIN_ genes expressed during the leaf primordia stages assayed. [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1237 1238 **Figure 6. Clustering analysis of leaf primordia time-course RNAseq reveals regulators of** 1239 **transverse vein development. (a)** Clustering analysis of leaf primordia RNAseq. Using Clust, 1240 RNAseq data was subdivided into 11 clusters. Significant Biological Process (BP) GO terms are 1241 listed beneath each cluster. Green clusters have significant GO terms related to photosynthesis, 1242 yellow cluster is related to carbohydrate transport, purple clusters are related to cell cycle and 1243 gene expression, blue clusters are related to RNA regulation, and the orange cluster had diverse 1244 GO term enrichment. n= number of genes in a cluster. **(b)** Heatmap showing _PIN1a_, _EREB41_, 1245 and _EREB114_ expression by leaf primordia stage. _EREB41_ and _EREB114_ expression preceded [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1246 _PIN1a_ expression. Heatmap coded by FPKM values, with warmer values representing higher 1247 expression. **(c)** _ereb114;ereb41_ transposon-insertion derived double mutants had reduced 1248 transverse veins relative to the WT control W22. Images are of the sheath. Arrowheads, normal 1249 transverse veins. Scale bar, 100 µm. **(d)** Number of transverse veins/cm [2 ] in WT, _ereb41_, 1250 _ereb114_, and _ereb41;114_ mutants. n=8, different letters indicate statistically significant 1251 difference using one-way ANOVA with _post hoc_ Sidak test. Measurements are from the leaf 1252 sheath. **(e)** Number of regular and irregular transverse veins in WT, _ereb41_, and _ereb41;114_ 1253 mutants, n=8. 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 [bioRxiv preprint doi: https://doi.org/10.64898/2026.10.28.761427; this version posted November 2, 2026. The copyright holder for this preprint](https://doi.org/10.64898/2026.10.28.761427) (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made [available under a CC-BY-NC-ND 4.0 International license.](http://creativecommons.org/licenses/by-nc-nd/4.0/) 1268 1269 **Figure 7.** **Model of transverse vein (TV) formation in maize leaves.** _Top:_ Cellular progression 1270 of TV development. **(1)** Adjacent longitudinal veins are established and express _PIN1a_ prior to 1271 TV initiation. **(2)** A putative initiation signal is proposed to specify paired TV initials adjacent to 1272 neighboring longitudinal veins. **(3)** PIN1a accumulates in the initials. **(4)** Intervening ground 1273 cells are progressively recruited and begin expressing _PIN1a_ . **(5)** Recruited cells undergo 1274 coordinated anticlinal divisions, generating a continuous pre-procambial strand connecting 1275 adjacent longitudinal veins. **(6)** _DR5_ signal is first detected in the initials. **(7)** _DR5_ activity 1276 expands throughout the developing procambial strand. **(8)** Mature TVs form fully differentiated 1277 vascular bundles that retain the same adaxial-abaxial organization as longitudinal veins, with 1278 xylem-to-xylem, phloem-to-phloem and bundle-sheath continuity between the two vein systems. 1279 _Bottom:_ Developmental progression of TV formation during maize leaf ontogeny. The midvein 1280 and longitudinal veins are initiated sequentially during leaf development, establishing the 1281 primary vascular scaffold of the leaf. TV initiation begins at approximately P6 and progresses 1282 basipetally through the blade during P7, while the sheath begins to elongate. By P8, TV 1283 formation is complete in the blade and continues through the sheath, showing the characteristic 1284 ladder-like vascular network. By P9, TVs are present throughout both blade and sheath domains 1285 and have differentiated into mature vascular bundles integrated into the longitudinal vein 1286 network. As leaf growth continues, transverse veins become increasingly oblique relative to the 1287 longitudinal veins. The ligule and auricle develop after TV establishment. Numbers above each 1288 leaf primordium indicate the cellular stages illustrated in the model on the top panel. Colors 1289 correspond to cell identities, reporter expression domains, and vein classes shown in the keys. 1290 1291