Skip to main content
NCBI home page
Journal of Experimental Botany logoLink to Journal of Experimental Botany
. 2025 Jul 1;76(17):5129–5144. doi: 10.1093/jxb/eraf245

Single-nuclei sequencing of Moricandia arvensis reveals bundle sheath cell function in the photorespiratory shuttle of C3-C4 intermediate Brassicaceae

Sebastian Triesch 1, Vanessa Reichel-Deland 2, José Miguel Valderrama Martín 3, Michael Melzer 4, Urte Schlüter 5, Andreas P M Weber 6,✉,b
Editor: John Lunn7
PMCID: PMC12587421  PMID: 40590299

Abstract

Spatially confined gene expression determines cell identity and is fundamental to complex plant traits. In the evolutionary transition from C3 to the more efficient C4 photosynthesis, restricting the glycine decarboxylase reaction to bundle sheath cells initiates a carbon concentrating mechanism via the photorespiratory glycine shuttle. This evolutionary step is generally thought to play an essential role in the progression from ancestral C3 to C4 photosynthesis. Plants operating this shuttle are often referred to as C3-C4 intermediates or C2 species. Within the Brassicaceae family, which includes model and crop plants, such species have evolved independently at least five times. However, research on the biochemistry of C3-C4 intermediates in the Brassicaceae has been limited to a few case studies of differentially localized proteins between mesophyll and bundle sheath cells. Here, we leverage recent advances in single-cell transcriptome sequencing to better understand how cellular specialization affects interconnected pathways. We generated a single-nuclei RNA sequencing dataset for Moricandia arvensis, a Brassicaceae with C3-C4 intermediate characteristics, and compared it to a publicly available single-cell transcriptome of leaf tissue of the C3  Arabidopsis thaliana. We confirmed the localization of selected photorespiratory proteins by electron microscopy of immunogold-labelled leaf sections. Our analysis revealed a M. arvensis-specific shift in expression of genes directly associated with the photorespiratory reactions, including components of the glycine decarboxylase complex, glutamate:glyoxylate aminotransferase, and glycolate oxidase, suggesting a shuttle of several C2 metabolites to the bundles sheath. Additionally, associated pathways, such as ammonium assimilation, synthesis of specific amino acids, redox regulation, and transport, also showed enhanced abundance in the M. arvensis bundle sheath.

Keywords: C3-C4 intermediate photosynthesis, evolution, glycine shuttle, photorespiration, single-nucleus sequencing

Single-nucleus RNA sequencing of Moricandia arvensis revealed bundle sheath cell-specific expression of photorespiration enzymes and associated pathways beyond glycine decarboxylase, including ammonium assimilation and redox regulation. This highlights the key role of metabolic compartmentalization in supporting C3-C4 intermediate photosynthesis.

Introduction

Complex plant traits require distinct spatial gene regulation. C3-C4 intermediate photosynthesis is a distinctive trait that involves specialized leaf anatomy and metabolism (Schlüter and Weber, 2016; Lundgren, 2020; Giacomello, 2021). C3-C4 intermediate species, which have evolved multiple times in more than 50 lineages, are a valuable resource for studying the early steps of C4 evolution and the genetics underlying cell-specific gene expression (Sage et al., 2018; Lundgren, 2020; Walsh et al., 2023). The Brassicaceae family contains at least five independent origins of C3-C4 intermediate photosynthesis, in close phylogenetic proximity to well-characterized model species like Arabidopsis thaliana or Arabis alpina, and economically relevant crop species such as Brassica oleracea (cabbage), Brassica napus (rapeseed), or Diplotaxis tenuifolia (arugula) (Guerreiro et al., 2023). The close phylogenetic proximity of model, crop, and C3-C4 plants makes the Brassicaceae family an ideal system to study the causes and consequences of differential cellular specialization.

In the Brassicaceae genus Moricandia, C3-C4 intermediate characteristics include a reduced carbon compensation point (Rawsthorne et al., 1988b) and an altered leaf ultrastructure, including bundle sheath cells (BSC) with a high number of centripetally arranged chloroplasts (Schlüter et al., 2017). It was also shown that the P-protein, a subunit of the glycine decarboxylase complex (GDC), selectively localizes to the BSC mitochondria in C3-C4 intermediate Moricandia, but not in closely related C3  Moricandia species (Hylton et al., 1988; Rawsthorne et al., 1988a, 1988b). GDC is a crucial component of photorespiration, as it catalyzes the decarboxylation reaction of photorespiratory glycine to carbon dioxide (CO2) and ammonium (NH3), as well as 5,10-methyltetrahydrofolate, which is converted to serine by a serine hydroxymethyltransferase (SHM). The BSC-specific activity of the GDC P-protein has been demonstrated in several C3-C4 intermediate and C4 species, and it is widely agreed to be a critical step in the evolution of C4 photosynthesis via C3-C4 intermediate stages (Schulze et al., 2016). In the current model of the glycine shuttle, the absence of the GDC-P protein in the mesophyll cell (MC) of M. arvensis leads to the accumulation of glycine, which then passively diffuses into the BSC where it is selectively decarboxylated. This glycine shuttle and the associated BSC-specific release of CO2 leads to an elevated CO2 partial pressure around Rubisco, and the photorespiratory CO2 can be efficiently refixed (Bauwe et al., 2010). The more efficient Rubisco carboxylation reaction is one of the potential advantages of the glycine shuttle in these C3-C4 intermediate species. This is accompanied by both a high number and centripetal organization of organelles in the BSC, limiting CO2 leakage.

Glycine decarboxylation also releases nitrogen (N) in form of NH3, and products of ammonium re-assimilation must move back to the mesophyll for balancing of N metabolism (Rawsthorne et al., 1988a; Monson and Rawsthorne, 2000). Several routes for the back-shuttle of photorespiratory nitrogen have been proposed using metabolic flux modelling (Mallmann et al., 2014). Mallmann et al. (2014) propose three scenarios for NH3 shuttles between the MC and the BSC: a glutamate/2-oxoglutarate shuttle, an alanine/pyruvate shuttle, and an aspartate/malate shuttle. Since aspartate, malate, pyruvate, and alanine itself are exchanged between MC and BSC in C4 photosynthesis, the proposed nitrogen back-shuttles in C3-C4 intermediate plants could have primed primordial C4 cycles and facilitated the evolution of the additional anatomical and metabolic C4 traits. However, there is limited experimental evidence for the presence of these shuttles in Moricandia, other than increased metabolite levels of glutamate, malate and alanine, suggesting the operation of multiple or mixed shuttles in C3-C4 intermediate M. arvensis and M. suffruticosa (Schlüter et al., 2017).

The establishment of a glycine shuttle that initiates further steps of C4 evolution has been predicted by modelling studies, suggesting a smooth pathway to C4 (Heckmann et al., 2013). Previous conceptual models place anatomical preconditioning events, such as the increase of BSC organelles and their localization to the inner BSC, before the shift of GDC activity to the BSC (Sage, 2016). Not much is known about the genetic regulation underpinning the evolution of C4 anatomy. In C4 grasses such as Zea mays (maize) and Setaria, the contribution of GOLDEN2-like and SCARECROW transcription factors (TFs) has been shown to influence BSC anatomy and chloroplast biogenesis (Slewinski et al., 2012; Lambret-Frotte et al., 2024), but the target genes controlled by these TFs and their cell-specific regulation remain unknown. The significant influence of TFs on the complex development of C4 anatomy indicates that multiple genes are involved in the evolution of C4 anatomy, and that this evolution is likely mediated by changes in cis-regulatory factors (Swift et al., 2024).

We reasoned that, to understand the genetic basis of C3-C4 intermediacy, we need to elucidate gene expression dynamics in a spatial context, in particular the shift in expression between MC and BSC. Methods for single-cell transcriptome studies in C3-C4 intermediate or C4 species were, so far, limited by the physical separation of BSC from phloem tissue or even BSC from MC (Aubry et al., 2014; Burgess et al., 2019; Borba et al., 2023). To date, laser microdissection or mechanical separation of bundle sheath strands have been used to isolate MC and BSC in C4 plants. These methods have provided valuable insights into the C4-specific transcriptional control at the single-cell level (Hua et al., 2021; Liu et al., 2022; Moreno-Villena et al., 2022). Isolation of tagged nuclei from specific cell types (INTACT) allows cell-type specific transcriptome sequencing, but requires the generation of transgenic material, which is not feasible for most plant species (Deal et al., 2011). Recent advances in the development of droplet-based single-cell or single-nuclei RNA sequencing (sc/snRNA-seq) methods have opened the door to the discovery of cell-specific transcriptome patterns at unprecedented resolution (Giacomello, 2021). snRNA-seq has recently been applied to C4 sorghum, providing deep insights into the evolution of differential partitioning of gene expression to the BSC from C3 rice (Swift et al., 2024). Applied to C3-C4 intermediate photosynthesis, snRNA-seq could answer important questions regarding the interconnectivity of the glycine shuttle with other pathways and the consequences of photorespiratory specialization for the MC and BSC metabolism.

To address this, we performed snRNA-seq on leaf tissue from the C3-C4 intermediate plant M. arvensis. We established a nuclei isolation protocol for M. arvensis and utilized droplet-based snRNA-seq. The resulting dataset allowed us to distinguish specific cell types and their corresponding transcriptional profiles. To confirm that the patterns we found were associated with C3-C4 intermediate traits, we compared the dataset with a publicly available A. thaliana leaf single-cell (sc) RNA-seq dataset (Kim et al., 2021).

We examined the datasets with respect to three main research questions: (1) What is the role of the BSC in a leaf with C3-C4 intermediate photosynthesis compared to a C3 BSC? (2) How is the glycine shuttle integrated into the BSC-specific nitrogen metabolism in M. arvensis? (3) Can we infer cis-regulatory patterns underlying BSC-specific gene expression?

We determined the expression of genes involved in photorespiration and the proposed pathways for the transport of photorespiratory nitrogen. Comparison of the single-cell transcriptomes of M. arvensis and A. thaliana also revealed a remarkable recruitment of gene expression to the BSC from A. thaliana to M. arvensis. We found that this BSC functionalization was largely driven by a shift in gene expression associated with photorespiration, ammonium assimilation, redox balance, and transport to the M. arvensis BSC. It is generally accepted that the evolution of C4 photosynthesis via C3-C4 intermediate stages involves the recruitment of existing C3 genes into modified gene regulatory networks, probably via the exaptation of cis-elements (Hibberd and Covshoff, 2010). To this end, we analyzed M. arvensis upstream DNA sequences for the presence of motifs for spatially co-expressed TFs. Although limited by sequencing depth, we found the BSC-specific expression of the TGAGC-binding (TGA) and the No Apical Meristem / Arabidopsis thaliana Transcription Activation Factor1/2 /Cup-shaped Cotyledon 2 (NAC) TFs, which may mediate BSC specificity in C3-C4 intermediate M. arvensis.

Materials and methods

Plant growth

Moricandia arvensis MOR1 seeds were sterilized using chlorine gas sterilization for 1 h and germinated on ½ strength Murashige and Skoog medium with 0.4% (w/v) agar for 7 days at room temperature and a 12 h light-dark cycle. Germinated seedlings were transferred to soil and grown in a growth chamber on a 12 h light-dark cycle and 25 °C (light; 150 µmol m-2 s-1) and 22 °C (dark).

Nuclei extraction

2 g leaf material of the combined fifth and sixth leaf (counting excluding the cotyledon) were cut using scissors and placed on a petri dish containing 1 ml LB01 buffer (15 mM Tris-HCl pH=7.5, 2 mM EDTA, 80 mM KCl, 20 mM NaCl, 15 mM 2-mercaptoethanol, 0.2% (v/v) Triton X-100, 0.5 mM Spermine). Leaves were chopped on ice for 2 min with a razor blade, 1 ml LB01 buffer was added, and leaves were chopped for another 3 min. Three ml LB01 buffer was added to the chopped leaves in the petri dish, and the mixture was incubated on ice for 15 min with gentle agitation every 3 min. The mixture was filtered through a 100 µm filter and subsequently through a 20 µm filter. All filters were pre-wetted with 1 ml LB01 buffer. The filtrate was subjected to a density gradient centrifugation in density gradient centrifugation buffer (DGCB; 1.7 M sucrose, 10 mM Tris-HCl pH=8.0, 2 mM MgCl2, 5 mM 2-mercaptoethanol, 1 mM EDTA, 0.15% Triton X-100) and centrifuged at 4 °C and 1500 g for 30 min. The supernatant was discarded and the pellet was resuspended with 400 µl 10X nuclei resuspension buffer from the Epi Multiome ATAC + Gene Expression kit (10X Genomics, California, USA). A 30 µl aliquot of the nuclei suspension was stained with 1.5 µl DAPI to check nuclei integrity under a fluorescence microscope. All buffers were supplemented with 1 U/µl RNase inhibitor (Roche, Basel, CH).

Single cell library generation

A total of 5000 nuclei were used as input for the single-cell droplet library generation on the 10X Chromium Controller system utilizing the Chromium Single Cell 3’ NextGEM Reagent Kit v3.1 (10X Genomics, California, USA) according to manufacturer’s instructions. Sequencing was carried out on a NextSeq 2000 system (Illumina Inc. San Diego, USA) with a 28-10-10-90 read configuration (28 cycles read1, 10 cycles index1, 10 cycles index 2, 90 cycles read 2).

Genome annotation and orthology analysis

The Arabidopsis thaliana TAIR10 genome and annotation was obtained from www.arabidopsis.org (Lamesch et al., 2012; Berardini et al., 2015). The M. arvensis MOR1 genome was obtained from Guerreiro et al. (2023), the de novo annotation using Helixer (Stiehler et al., 2021; Holst et al., 2023) was obtained from Triesch et al. (2024). gffread 0.12.7 was used to convert gff3 to gtf files to map snRNA-seq reads. Homologs between A. thaliana and M. arvensis were identified using MMseqs2 (Steinegger and Söding, 2017). If multiple M. arvensis homologs for one A. thaliana gene were found, all were retained.

Processing of 10X genomics single cell data

Raw reads from the M. arvensis snRNA-seq experiment were processed using Cell Ranger 6.1.2 (10X Genomics, CA, USA) with the ARC-v1 chemistry option. The count matrix for the A. thaliana scRNA-seq experiment was kindly provided by Dr. Ji-Yun Kim (Kim et al., 2021). RNA velocity was determined using Velocyto 0.17.17 (La Manno et al., 2018) and scvelo 0.3.1 (https://github.com/theislab/scvelo). Further processing of the count matrix from Cell Ranger was performed using scanpy 1.9.3 (https://github.com/scverse/scanpy). Genes annotated with chloroplast and mitochondrial origin were removed. Genes expressed in fewer than 10 cells and cells with fewer than 200 genes expressed were removed. Batch effects were removed using bbknn (Polański et al., 2020). Cells were clustered using a graph-based approach using 50 principal components and a resolution of 0.25 for the integrated dataset and 0.5 for subclustering. Cross-species data integration was performed using the sc.tl.ingest function of scanpy 1.9.3.

Cluster annotation and marker gene inference

Single-cell markers genes for A. thaliana were obtained from PlantscRNAdb (Chen et al., 2021) and the M. arvensis homologs for these were used for the respective datasets. In both datasets, within each cluster, marker genes were defined as genes with a P<1e−20 and a positive log2-fold change. Each marker gene was tagged with one or more cell types from the PlantscRNAdb marker gene annotation. Fisher’s exact test was employed to find cell type annotation enrichments within the marker genes for the individual clusters. The enriched cell type annotation with the lowest p-value was naïvely selected for annotation of the respective cluster.

Cross-species cell-type comparison

Marker genes for each cluster/cell-type were defined as genes with P<1e−20 and a positive log2-fold change. Differentially expressed genes were determined from expression matrices using Mann-Whitney-U-Tests with a significance threshold of P<1e−20. For the between-species analysis, gene expression as read counts per transcript was normalized to the highest expressed marker gene of the respective cluster. MapMan (Schwacke et al., 2019) bin classifications for enrichment analyses were taken from Triesch et al. (2024). MapMan bin enrichment was performed using Fisher’s exact test, calculating the enrichment of target gene sets within the complete set of genes in the TAIR10 or M. arvensis annotation, respectively. Sankey plots were created using www.sankeymatic.com.

Cis-element detection and cross-species analysis

Transcription factor binding sites for all upstream sequences were predicted using FIMO (Grant et al., 2011) from the MEME suite 5.5.5 (Bailey et al., 2009). To this end, the 3000 bp upstream sequences for all annotated M. arvensis and A. thaliana genes were dumped using a custom python script. Motif clustering was performed on transcription factor motifs using the JASPAR database (Castro-Mondragon et al., 2022). To account for sequence composition a background model was generated using the fasta-get-markov tool from the MEME suite 5.5.5 (https://meme-suite.org/meme/). The target upstream sequence, the background model, and the database were used as input for the Find Individual Motif Occurences Tool (FIMO; https://meme-suite.org/meme/doc/fimo.html).

Immunogold labelling analysis of leaf tissue by light- and transmission electron microscopy

The central part of the fifth true leaf from M. arvensis and rosette leaves from A. thaliana of at least three independent plants were cut, chopped in 2 mm2 pieces and placed in a 2 ml tube containing the freshly prepared fixative solution (2% formaldehyde+0.5% glutaraldehyde in 0.05 M Na-cacodylate buffer, pH 7.2). Gentle vacuum infiltration of the fixative using a vacuum pump was followed by three rounds of microwave assisted fixation with a Pelco 34700 Biowave Laboratory Microwave Processor (Ted Pella, Inc., Redding CA, USA) at 150 W for 1 min with 1 min rest between rounds. Afterwards, the samples were washed once with 0.05 M Na-cacodylate buffer (0.05 M Na(CH3)2AsO2, pH 7.3) and 3 times with distilled water, applying 150 W for 1 minute after each wash step. Samples were then dehydrated using an ethanol series of 30, 40, 50, 60, 70, 80, 90% and 100%. At each dehydration step, 150 W were applied for 1 min followed by 15 min incubation on a shaker at 400 rpm at room temperature. 100% ethanol was substituted as many times as required to remove as much chlorophyll as possible. For ultrastructural analysis, sections were embedded in Spurr resin, sectioned, and analysed as described by Kavka et al. (2024).

Immunocytochemistry sectioning and immunogold labelling was performed as described in Schwarz et al. (2015). Resin infiltration samples were incubated in immediately-prepared 25, 50, 75, and 100% Lowicryl HM20 solution (Plano GmbH, Marburg, Germany) in ethanol. Samples were incubated overnight in 25% HM20, followed by 4 hours in each 50% then 75% HM20, and overnight incubation at 100% HM20 before polymerization in fill gelatine capsules at −10 °C under UV light in a Leica EM ASF2 (automated freeze substitution system). Excess resin was trimmed creating a trapezoid surface using a Leica EM Trim trimming device. Sections of 0.70 µM thickness were obtained with the Ultracut UCT ultramicrotome using a Diatome Ultra Knife. Antibodies were obtained from Peter Westhoff (Heinrich Heine University, Düsseldorf, DE) and Hermann Bauwe (University of Rostock, DE) in the case of serine hydroxymethyltransferase (SHM). Antibodies were used in a 1:50 dilution in antibody buffer (PBS buffer (137 mM NaCL, 2.7 mM KCl, 10 mM Na2HPO4; 1.8.mM KH2PO4, pH 7.4) with 0.1% BSA )and are listed in Supplementary Table S1.

Counting of gold particles and the area of the organelles were determined using ImageJ (Schneider et al., 2012). The density of gold particles was calculated by counting the gold particles on the images at a magnification of 89 000× and then calculating the number of particles per μm2. The gold particle density was measured on three different samples and a minimum of 45 BSC and MSC from the surrounding area of a minimum of 9 different veins. The labelling density was calculated as a mean of number of gold particles per µm2 from at least 13 different organelles.

Statistical analyses for the gold particle count were performed using PRISM 8 (Graphpad, https://www.graphpad.com). Differences in labelling densities of the components of the glycine decarboxylase complex (GLDP1, GLDH, GLDT), SHM and hydroxypyruvate reductase (HPR) between MC and BSC were analyzed using student’s t test. Differences were considered to be statistically significant when the P value was <0.01.

Results

Single-nuclei sequencing of Moricandia arvensis

To gain an understanding of gene expression patterns in the leaf of a C3-C4 intermediate Brassicaceae species, the fifth and sixth leaves of young M. arvensis plants were used for nuclei extraction and snRNA-seq. Across three biological replicates, a total of 11 013 nuclei were used for snRNA-seq using the 10 × single-nuclei workflow, with a median of 1070 identified genes per nucleus. After the integration of biological replicates using bbknn (Polański et al., 2020), the output was visualized using uniform manifold approximation projection (UMAP) and subjected to unsupervised clustering using scanpy, resulting in 6 distinct cell clusters (Fig. 1A). The expression levels of the closest M. arvensis homologs for single-cell marker genes from A. thaliana (Chen et al., 2021) were used for annotation of cell types (Fig. 1D).

Fig. 1.

Fig. 1.

Open in a new tab

(A) Uniform manifold approximation and projection (UMAP) of transcript profiles for Moricandia arvensis 11 013 leaf cells. Cells were grouped into clusters using scanpy. Cell identities were assigned using marker genes for different cell types (D). (B) UMAP with RNA velocity arrows modelled using velocyto and scvelo. (C) Enrichment of MapMan bins for differentially expressed genes between M. arvensis bundle sheath and mesophyll clusters. (E) UMAP with heatmap overlays indicating expression strength of selected genes.

In the resulting dataset, two MC clusters, one BSC clusters, one phloem, and two epidermis cell clusters were identified (Fig. 1A; Supplementary Table S2.1). The MC clusters were identified by the strong expression of genes encoding CARBONIC ANHYDRASE homologs, subunits of photosystem I, and RUBISCO ACTIVASE. The phloem cluster was identified by the expression of marker genes such as thioredoxin 3 TRX3, glutamine synthetase (GSR2) and repressor of GSNOR1 (ROG1). From the phloem cluster, two subclusters were annotated as phloem parenchyma and phloem companion cells. Phloem companion cells were identified using markers such as amino acid permease 1 (AAP1) and sugars will eventually be exported transporters 1 (SWEET1), and phloem parenchyma cells we identified by marker genes like regulator of bulb biogenesis1 (RBB1) and phosphoenolpyruvate carboxykinase (PCK1).

The epidermis clusters were identified by expression of epidermal marker genes such as ECERIFERUM and 3-ketoacyl-CoA synthase 6 (KCS6) (Fig. 1D). RNA velocity analysis indicated that two subclusters might represent old and young epidermal cells rather than upper and lower epidermal layers (Fig. 1B). 1687 cells formed a cluster that were annotated as BSC tissue. Previously described marker genes for the BSC cluster encode the sulfate ion transporter (SULTR) sulfate transporter homologs, the ascorbate peroxidase APX1 and the transcription factor unfertilized embryo sac (UNE12). The M. arvensis bundle sheath cluster was also marked by genes from the photorespiratory cycle, including genes encoding the GDC P-, H-, and T-protein (GLDP1, GDC-H1, GLDT), and SHM (Fig. 1D).

Comparative single-cell transcriptomics with Arabidopsis thaliana

To gain insight into cell-specific expression patterns that correlate with, and potentially underlie, traits associated with C3-C4 intermediate photosynthesis, we compared our M. arvensis snRNA-seq dataset with a publicly available scRNA-seq dataset from the C3 Brassicaceae species A. thaliana. This dataset contained expression data for 5230 vasculature-enriched leaf protoplasts from two biological replicates (Kim et al., 2021). With a median of 3342, the number of genes per cell was approximately three times higher in the A. thaliana protoplast dataset than in the M. arvensis nuclei population (980 genes per cell). To compare the datasets of both species, we assigned M. arvensis homologs of A. thaliana genes using MMseqs2 (Steinegger and Söding, 2017). In this approach, genes are clustered based on protein sequences and ortholog and paralog sequences are combined. For 18 468 A. thaliana genes, one or more homologs were found in M. arvensis. A median of two M. arvensis gene copies were found for each A. thaliana gene.

We integrated the datasets by ingesting the A. thaliana data into the M. arvensis UMAP projection. When analyzed alone, we were able to identify senescent, initial and guard cells in the A. thaliana single-cell population. However, these clusters could not be observed when integrated into the M. arvensis data, potentially because the integration only considers common genes in both datasets. The average sequencing depth reached approximately 96 000 reads per cell in the A. thaliana dataset, whereas it was limited to approximately 27 000 reads per cell in our M. arvensis dataset.

Comparison of BSC populations between the species revealed that a proportion of M. arvensis BSC marker genes were not BSC-specific in A. thaliana (Fig. 2E). 83 M. arvensis BSC marker genes were also marker genes in the A. thaliana BSC, and 110 M. arvensis BSC marker genes were not BSC-specific in A. thaliana. The 83 common BSC marker genes were enriched in the MapMan bins ‘Solute transport’ and ‘Clade-specific metabolism.Brassicaceae’, containing glucosinolate metabolism related genes. The common ‘Solute transport’ bin included experimentally verified general BSC markers such as SULTR2;2 (Kirschner et al., 2018). A. thaliana specific enrichments of the BSC cluster were not dominated by a specific pathway; among them were MapMan bins connected to ‘Vesicle trafficking’, ‘RNA processing’, ‘Protein modification’ and ‘Protein biosynthesis’.

Fig. 2.

Fig. 2.

Open in a new tab

UMAP visualizations of the integrated Moricandia arvensis single-nuclei RNA-Seq and Kim et al. (2021)  Arabidopsis thaliana single-cell RNA-Seq dataset colored by cell type (A) and species (B). (C) Heatmaps indicating normalized expression in mesophyll and bundle sheath clusters in A. thaliana and M. arvensis. Expression was normalized to the highest expressed marker gene in the respective cluster and species. (D) Enrichment of MapMan bins for 110 BSC-specific differentially partitioned genes between M. arvensis and A. thaliana. (E) Sankey diagram indicating the expression of M. arvensis marker genes (right side of the diagram) in A. thaliana clusters (left side).

The 110 genes that showed BSC-specific expression only in M. arvensis gained this expression pattern during the evolutionary progression since the last common ancestor between the two plant species. These differentially partitioned M. arvensis BSC marker genes were dominated by the MapMan bin ‘Photosynthesis.photorespiration’. It included genes encoding subunits of the mitochondrial decarboxylation system such as GLDP1, GLDT, GDC-H1 and SHM, but interestingly also two gene copies encoding the glutamate:glyoxylate aminotransferase (GGT) and one copy encoding the glycolate oxidase (GOX) (Fig. 1C, Supplementary Tables S2.2, S2.4). In A. thaliana these photorespiratory genes showed no preference for the BSC, and some parts of the GDC were even slightly more abundant in the MC (Fig. 3, Supplementary Table S2.2, S2.4).

Fig. 3.

Fig. 3.

Open in a new tab

Schematic view of the photorespiratory pathway in Arabidopsis thaliana (A) and Moricandia arvensis (B) mesophyll and bundle sheath cells. Heatmaps indicate the expression of genes encoding selected photorespiratory enzymes from the respective A. thaliana and M. arvensis single-cell RNA-Seq datasets. The expression was normalized to the highest marker genes in the respective cluster and species. The left tile of each heatmap indicates normalized expression in the mesophyll, the right tile indicates normalized expression in the bundle sheath. Multiple rows per tile indicate multiple gene copies. The nitrogen transfer reactions are shown in dashed lines. Abbreviations: CBC: Calvin-Benson-Bassham Cycle, 3-PGA: 3-Phosphoglyceric acid, Rubisco: Ribulose-1,5-bisphosphate carboxylase/oxygenase, GLYK: glycerate kinase, GOX: glycolate oxidase, GGT: glutamate:glyoxylate aminotransferase, AGT: serine:glyoxylate aminotransferase, HPR: hydroxypyruvate reductase, GLDP/GLDT/GLDH/GLDL: glycine decarboxylase P/T/H/L protein, SHMT: serine hydroxymethyltransferase, PGLP: phosphoglycolate phosphatase, GOGAT: glutamine-2-oxoglutarate aminotransferase, GS: glutamine synthetase, 2OG: 2-oxoglutaric acid.

In M. arvensis, BSC specific enrichment was also found for MapMan bins ‘Solute transport.Primary active transport’ and ‘Nutrient uptake.nitrogen assimilation’. The ‘Nutrient uptake.nitrogen assimilation’ bin contained genes in the plastidial ammonium assimilation pathway (glutamine synthetase 2 (GS2), glutamate synthase (ferredoxin) (Fd-GOGAT)) that are closely linked to recapturing of the photorespiratory released ammonium.

The enrichment of ‘Solute transport.Primary active transport’ was largely due to the BSC-specific expression of ATP-binding cassette (ABC) transporters in M. arvensis. Additional transport proteins with BSC preference only in M. arvensis included many aquaporins of the plasma membrane intrinsic protein (PIP) and tip growth defective (TIP) families and the mitochondrial A BOUT DE SOUFFLE (BOU) transporter that is closely associated to the photorespiratory GDC activity (Eisenhut et al., 2013).

Cellular preferences of potential nitrogen metabolism or C4 shuttle-associated enzymes were also investigated, but many of these showed low expression levels in the M. arvensis dataset (Supplemental Fig. S4). Of the four M. arvensis paralogs of the single plastidial A. thaliana NADP-MDH gene, one showed highly BSC-specific expression, whereas another paralog showed higher MC expression. None of the potential C4 decarboxylating enzymes were preferentially found in the M. arvensis bundle sheath, but one copy each of phosphoenolpyruvate carboxylase (PEPC) and pyruvate, phosphate dikinase (PPDK) were slightly more abundant in the MC cluster (Supplemental Fig. S4).

Cis-Sy underpinnings of differential partitioning

Next, we examined the cis-regulatory landscape of differentially partitioned genes in the M. arvensis and A. thaliana datasets. To do this, the 3000 bp upstream regions for each marker gene were analyzed using FIMO (Grant et al., 2011), which predicts TF binding sites. In contrast to the previous cross-species comparative analysis, marker genes were identified in separate, non-integrated datasets to account for the low expression of TFs in M. arvensis.

The list of putatively binding TFs based on upstream sequence elements was narrowed down using TF co-expression data from our single-cell RNAseq datasets.

In A. thaliana, binding sites for several co-expressed TFs were identified in the upstream regions of marker genes. A prime example was the accumulation of MYB30 TF binding sites in the A. thaliana epidermis marker genes, which was previously associated with epidermal wax synthesis (Zhang et al. (2019); Supplementary Fig. S6). Furthermore, DNA binding with one finger (DOF1 and DOF5) TFs were co-expressed and predicted to bind BSC marker genes in the A. thaliana data, which is supported by previous studies (Guo et al., 2009; Dai et al., 2022). The BSC-specific preference of DOF TFs was also found in a recent single-cell study of sorghum and rice (Swift et al., 2024). Recruitment of DOF TF binding sites is thought to have driven BSC repurposing in the evolution of C4 sorghum (Swift et al., 2024), for example by mediating differential partitioning of NADP-ME (Borba et al., 2023). A larger number of co-expressed and potentially binding TFs were identified in the A. thaliana dataset than in M. arvensis. This is likely an effect of the lower read coverage in the M. arvensis single-nuclei RNA-seq data, which is also likely the reason for the lack of detection of DOF TFs in the M. arvensis BSC (Supplementary Fig. S5).

Two co-expressed and putatively binding TFs were identified in the M. arvensis BSC marker genes, TGACG MOTIF-BINDING FACTOR 4 (TGA4) and NAC DOMAIN CONTAINING PROTEIN 83 (NAC083; Supplementary Fig. S6). Interestingly, these TFs were not identified as co-expressed and binding in A. thaliana BSC marker genes.

Leaf ultrastructure and protein localization through immunogold labelling

To corroborate the biological insights gained from our M. arvensis snRNA-seq dataset and the A. thaliana dataset from (Kim et al., 2021), we performed immunogold labelling to localize selected proteins predicted to be differentially distributed between the C3 and C3-C4 intermediate study systems.

In M. arvensis, BSC mitochondria were significantly labelled for GLDP, GLDH, GLDT and SHM (Fig. 4), with significant differences from MC mitochondria (Fig. 4, Table 1). For HPR, no significant differences in labelling were found between BSC and MC mitochondria for M. arvensis (Table 1). In contrast, in A. thaliana, no statistically significant differences were observed between the two cell types for the localization of GLDT, GLDH, SHM and HPR (Table 1).

Figure 4.

Figure 4.

Open in a new tab

A: Light microscopy images of cross-sections of the central region of the fifth leaf from Moricandia arvensis and Arabidopsis thaliana embedded in SPURR resin. B: Electron microscopy images showing mitochondria of BSC and MC from M. arvensis and A. thaliana. Black dots are immunogold particles labelling selected photorespiratory proteins.

Table 1.

Numbers of gold particles per unit area (µm−2) given as mean ± standard deviation.a

A. thaliana M. arvensis
Enzyme Number of gold particles µm−2
BSC MC BSC MC
GLDP 65.63±26.55 60.31±30.01 68.86±32.73* 5.34±8.12*
GLDT 36.99±19.35 35.02±17.70 109.81±43.29* 21.16±11.44*
GDLH 9.54±7.68 17.97±12.93 28.74±11.00* 6.9±6.63*
SHM 15.38±11.65 19.79±12.95 61.59±27.88* 13.71±9.51*
HPR 67.73±32 78.51±26.89 34.94±10.37 49.98±18.82
Open in a new tab

aData was obtained from different sections from at least three different leaves.

*(Asterisks) indicate a significant difference (P<0.01) between mesophyll (MC) and bundle-sheath cells (BSC) from the same species after t-test analysis.

Discussion

C3-C4 intermediate plants recapture the CO2 released by photorespiration more efficiently than C3 plants by operating a photorespiratory shuttle, also called C2 photosynthesis (Sage et al., 2012; Schlüter and Weber, 2016; Lundgren, 2020). The allocation of photorespiratory CO2 release to the BSC creates conditions that favor the carboxylation reaction of the BSC Rubisco. C2 biochemistry is mainly based on the specific localization of the GDC P-protein in BSC mitochondria (Rawsthorne et al., 1988b; Sage et al., 2012). This differential localization is observed in all C3-C4 intermediate taxa characterized to date, demonstrating that the evolution of this specialized trait relies on similar mechanisms (Hylton et al., 1988; Schlüter and Weber, 2016). Apart from the shift in GLDP protein localization, not much is known about cellular metabolic adjustments in the different evolutionary lineages. For C3-C4 Brassicaceae such as M. arvensis it has been suggested that differential gene expression is mainly limited to BSC specificity of the GLDP protein (Morgan et al., 1993). By using snRNA-seq of young M. arvensis leaves, we provide a single-cell resolution transcriptome for a C3-C4 intermediate Brassicaceae species. Our dataset covers the majority of leaf cell types and shows robustness between the biological replicates (Supplementary Fig. S1). We assigned cell identities to clusters based on marker genes, which are consistent with previous single-cell transcriptomes of plant leaves (Berrío et al., 2022; Procko et al., 2022; Wang et al., 2023; Swift et al., 2024).

To compare the single-cell transcriptome data of a C3-C4 photosynthetic leaf to a C3 species, we used publicly available scRNA-seq data from A. thaliana leaf protoplasts (Kim et al., 2021). The split between the Brassiceae and the Camelineae tribes, which contain M. arvensis and A. thaliana, respectively, dates back 25 Mya (Walden et al., 2020). However, the high number of shared marker genes between the respective cell types allows intra-species comparisons.

SnRNA-seq captures only nuclear transcripts enriched in unspliced pre-mRNAs and excludes cytoplasmic mRNAs, whereas single-cell RNA-seq (scRNA-seq) from whole protoplasts profiles both nuclear and cytosolic transcripts. Consequently, scRNA-seq typically detects more unique genes per cell. Although snRNA-seq tends to under-represent mature cytoplasmic transcripts, cell identity signatures are generally preserved across both techniques, enabling reliable clustering and annotation (Mereu et al., 2020; Conde et al., 2022; Wang et al., 2023). In our analysis, the larger number of BSC specific genes in A. thaliana (826) compared to M. arvensis BSC specific genes (110) was likely influenced by the RNA preparation method and sequencing depth of the scRNA dataset (Supplementary Table S2.3). Preferential capture of genes encoding proteins controlling RNA and protein processing and biosynthesis by scRNA seq compared to snRNA seq of the same tissue has been observed in animal material (Santiago et al., 2023), so the observed enrichment of such pathways only in A. thaliana BSC (Supplementary Table S2.4) could be related to the differences in the applied methods. To limit influences caused by differences in RNA content and count depth, batch-specific scaling was necessary to prevent clustering by library size rather than biology. We applied the scanpy ingest function (a PCA-based label transfer) to project M. arvensis single-derived transcriptomes onto the A. thaliana reference embedding, followed by batch-balanced k-nearest neighbours algorithm (bbknn) to correct batch effects, thereby aligning shared cell populations in a common manifold. Successful integration of sn- and scRNA datasets has been performed previously for different animal and plant organs (Bakken et al., 2018, Conde et al., 2022, Santiago et al., 2023; Wang et al., 2023). We observed little to no difference when comparing the expression of highly expressed genes involved in photorespiration and nitrogen shuttling between integrated and non-integrated M. arvensis and A. thaliana datasets. The M. arvensis BSC cluster could be identified by overlap with common A. thaliana BSC marker genes such as SULTR2;2 and genes involved in glucosinolate metabolism (Supplementary Table S2.4, Aubry et al., 2014; Kim et al., 2021; Procko et al, 2022). While we identified most of the cell types relevant to our study in the M. arvensis dataset we generated, guard cells were conspicuously absent, even after integration with the A. thaliana data which contain this cell type. We hypothesize that our M. arvensis nuclei isolation workflow needs to be optimized for the extraction of smaller cells, possibly by starting with more finely chopped leaf extract.

Due to the prominent role of the BSC in the C3-C4 intermediate leaf, we analyzed differential gene expression patterns between BSC and MC in M. arvensis, as well as between the M. arvensis and A. thaliana BSC clusters. In doing so, we noticed that multiple BSC marker genes in M. arvensis were not BSC-specific in A. thaliana. The majority of these differentially partitioned genes showed MC and phloem parenchyma expression in A. thaliana (Fig. 2E). The reprogramming of the BSC by recruiting gene expression from MC and phloem cells was also observed in a comparative single-cell transcriptomic study between sorghum and rice (Swift et al., 2024) and it is thought to be one of the crucial drivers of C4 evolution (Hibberd and Covshoff, 2010; Reeves et al., 2017; Singh et al., 2023).

To pinpoint genes underlying cellular specification during C3-C4 evolution in M. arvensis, we analyzed the set of differentially partitioned genes between the two species in greater detail. First, we observed a strong enrichment of photorespiratory gene expression in the M. arvensis BSC compared to A. thaliana (Fig. 1C). This was most pronounced in the differential partitioning of GDC proteins to the BSC of the C3-C4 intermediate leaf (Fig. 1E), including genes encoding the T-, L- and H- protein of the GDC, and of SHM, working together in the mitochondrial glycine to serine converting enzyme system. For GLDP, GLDT, GLDH and SHM this observation could also be confirmed on the protein level by immunolocalization (Fig. 4). The simultaneous differential partitioning of multiple parts of the GDC/SHM system is plausible because it would generate a fitness disadvantage to produce highly abundant proteins for a non-functioning GDC complex in the BSC. Loss of GDC/SHM activity in the mesophyll cells leads to the induction of a photorespiratory shuttle and increase of CO2 concentration in BSC by the enhanced glycine decarboxylation (Rawsthrone et al., 1988a, 1988b). Besides the enhanced CO2 release, the shift of the whole GDC/SHM complex to the BSC causes enhanced production of NH3, methylene-tetrahydrofolate, and NADH in this cell type and this requires adjustments in other pathways.

Consistent with this, we observed increased expression of genes involved in the GLUTAMINE SYNTHETASE/GLUTAMINE-2-OXOGLUTARATE AMINOTRANSFERASE (GS/GOGAT) cycle in the BSC, thus avoiding the loss of photorespiratory nitrogen by high ammonium fixation capacity directly in the BSC of M. arvensis leaves. In one photorespiratory cycle of C3 species like A. thaliana, two molecules of glycine are formed from glyoxylate by activity of the GLUTAMATE:GLYOXYLATE AMINOTRANSFERASE (GGT). Amino groups for this reaction are normally supplied by the SERINE:GLYOXYLATE AMINOTRANSFERASE (AGT) (Bauwe, 2023). However, since only one molecule of serine is formed from two molecules of glycine by the mitochondrial GDC/SHM, additional amino donors must be imported into the peroxisome, mainly provided by the activity of the GS/GOGAT cycle (Fig. 3). Interestingly, our snRNA dataset also indicated preferential expression of genes encoding two of the GGT gene copies and one GOX copy in M. arvensis BSC (Fig. 3). Thus, under operation of the photorespiratory glycine—serine shuttle, serine returning to the MC could provide some of the amino groups, but some of the glyoxylate, and possibly also some glycolate, could be transported to the BSC where amino donors from ammonium refixation products would be available for GGT activity (Fig. 3, Supplementary Figs S3, S4). Both peroxisomal amino transferases accept various substrates as amino donors, thereby increasing the robustness of the system (Bauwe, 2023). A partial shift of GGT and GOX activity to the BSC in C2 species would therefore significantly reduce the nitrogen imbalance between the two cell types and release the pressure on additional N balancing systems (Alvarenga et al., 2025; Borghi et al., 2022). These conclusions are well supported by previous protein activity studies where Rawsthorne et al. (1988a, 1988b) used a mixture of protoplasting and strand-separation techniques to distinguish MC from BSC cells. Higher enzyme activities were found in M. arvensis BSC fractions for GDC, GOGAT, GS, but not for HPR and GOX compared to the MC fractions. A shift of peroxisomal reaction of the photorespiratory pathway was recently also hypothesized for some grasses where the operation of a C2 cycle could be associated with an increase in BSC peroxisomes (Alvarenga et al., 2025).

However, not all photorespiratory intermediates are recycled and, in particular, glycine and serine are removed from the cycle (Busch, 2020; Walker et al., 2024). In the source leaf, the BSC would also supply metabolites to other plant parts via the adjacent vein (Supplementary Fig. S3), for which the system could be buffered by the operation of additional metabolite transport such as the proposed glutamate/2-oxoglutarate, alanine/pyruvate, and aspartate/malate shuttles (Mallmann et al., 2014). The shuttles could operate at low levels and would not require additional cell specific adaptations of the associated enzymes ALANINE AMINOTRANSFERASE (AlaAT), ASPARTATE AMINOTRANSFERASE (AspAT), and MALATE DEHYDROGENASE (MDH) (Supplementary Fig. S3). The mitochondrial GDC/SHM complex also provides backbones for the C1 metabolism, and the BSC specificity of methionine synthases in our dataset indicates that synthesis of specific amino acids is also adjusted to the spatial distribution of precursors in the C3-C4 intermediate leaf (Supplementary Fig. S3).

Interestingly, transcripts of the gene encoding the BOU transporter, linked to GDC activity, are also more abundant in the BSC. Eisenhut et al. (2013) observed that BOU expression correlates with the expression of genes encoding GDC proteins, and bou mutants exhibit decreased GDC activity. The authors suggested that BOU may transport a GDC co-factor into mitochondria. It is conceivable that BOU transports glutamate for the polyglutamylation of THF, a co-substrate for the GDC and SHM. The differential distribution of BOU in the C3-C4 intermediate BSC may ensure BSC-specific GDC activity and may influence C1 metabolism.

The distribution of the photorespiratory pathway over three different organelles requires multiple transport steps, and not all of the involved transporters have been identified (Kuhnert et al., 2021). In the M. arvensis BSCs, the shift of photorespiratory activities was accompanied by enrichment of genes encoding different solute transporters. Besides BOU, this included mainly members of the ABC transporter family and aquaporins of the plasma membrane intrinsic proteins (PIP) or the tonoplast intrinsic proteins (TIP). Members of PIP and TIP family have also been found in the chloroplast membrane, where they facilitate the diffusion of water, CO2, urea, and ammonium (Groszmann et al., 2017). In the M. arvensis BSC, fast and efficient recapture of CO2 and ammonium released in the mitochondria also depends on the permeabilities of the mitochondrial and plastidial membranes, respectively, and aquaporins could enhance conductance of these compounds. On the other hand, aquaporins can also switch between open and closed states by conformational changes, in the BSC of M. arvensis reduced CO2 conductance due to closed aquaporins in the plasma membrane could prevent leakage.

Since the mitochondrial photorespiratory reactions are also connected to NADH formation, it was plausible that changes in the BSC metabolism between the C3  A. thaliana and the C3-C4 intermediate M. arvensis require changes in redox balancing and energy homeostasis. Our dataset revealed that genes encoding proteins involved in alternative mitochondrial respiratory chains such as ALTERNATIVE OXIDASE 1 A (AOX1A) and ALTERNATIVE NAD(P)H DEHYDROGENASE 1 (NAD1) were differentially partitioned towards the M. arvensis BSC. An up-regulation of AOX gene expression was already observed in bulk transcriptome studies in M. arvensis but not in the C3 sister species M. moricandioides, indicating enhanced re-balancing of the redox metabolism due to the photorespiratory shuttle (Schlüter et al., 2017). Redox homeostasis can also be maintained by the malate valve, especially through the interplay of NADP-MDH (NADP-dependent malate dehydrogenase) enzymes (Selinski and Scheibe, 2019). Whereas the plastidial NADP-MDH is encoded by a single gene in A. thaliana, three paralogs exist in M. arvensis. Interestingly, we found a differential partitioning between these paralogs in M. arvensis. This could hint at a tight control of the malate metabolism between MC and BSC, carried out by multiple genes with potentially individual regulatory mechanisms.

Our comparative snRNA seq analysis showed that many genes classified as BSC specific only in the C3-C4 species M. arvensis can be directly or indirectly linked to the shift of the mitochondrial glycine decarboxylation reaction. Such economical metabolic integration of the photorespiratory shuttle could be linked to cell specific regulatory elements in the underlying genome (Swift et al., 2024), providing plants with the ability to avoid the production of superfluous proteins. However, also in C3 species like A. thaliana, an estimated 15% of genes were preferentially expressed in the BSC. This included genes serving specific roles in leaf hydraulics, sulfur metabolism, glucosinolate biosynthesis, trehalose metabolism, and phloem loading (Aubry et al., 2014). Similar to the situation in A. thaliana, we found a higher BSC expression of genes involved in sulfur metabolism in M. arvensis (Supplementary Fig. S2). This included the sulfate transport steps mediated by SULTR2;2 and SULTR3;3 as well as primary reductive steps of sulfate assimilation that were highly BSC-specific in both species. This specificity, however, was more pronounced in A. thaliana.

We also interrogated pathways that showed differential activity in MC and BSC of C4 species such as the Calvin-Benson cycle or sucrose and starch synthesis (Schlüter and Weber, 2020), but none of these could be assigned to BSC in our M. arvensis dataset.

Conclusions

Photorespiratory shuttle mechanisms have evolved independently in multiple plant lineages. Common features identified among C3-C4 intermediates include similar anatomical modifications and a biochemical shift of the glycine decarboxylase (GDC) reaction, primarily associated with altered positioning of the GLDP protein. Nevertheless, subsequent metabolic adjustments may have arisen through lineage-specific solutions, potentially impacting the trajectory toward the evolution of a complete C4 photosynthetic pathway (Mallmann et al., 2014; Borghi et al., 2022; Alvarenga et al., 2025).

The Asteraceae genus Flaveria contains C3 species in close phylogenetic proximity to various C3-C4 intermediates with different degree of C4 cycle contributions as well as species employing a fully developed C4 pathway. Studies in Flaveria have therefore been used for development and verification of general C4 evolution models (Mallmann et al., 2014; Adachi et al., 2023; Tang et al., 2024). Processes that address the nitrogen imbalance arising from the exchange of two molecules of glycine for one molecule of serine in the photorespiratory shuttle are thought to play a critical role in the further evolution toward C4 photosynthesis. In Flaveria, even the least advanced C3-C4 intermediates show elevated transcripts and activity levels particularly of aspartate aminotransferases (Mallmann et al., 2014; Adachi et al., 2023), suggesting that aspartate is involved in N shuttles between BCS and MC in these species. In return malate could be transported from MC to BSC for rebalancing of carbon scaffolds. In conjunction with BSC-specific NADP-ME activity, this would promote a C4-like metabolite exchange. For Flaveria, such a model is supported by increases of NADP-malic enzyme activity in optimized intermediate species (Adachi et al., 2023). Increases in activity of the decarboxylating NADP-ME in more advanced Flaveria intermediates could recently be connected to decreases in the photorespiratory shuttle, especially activity of GDC and Fd-GOGAT resulting in accumulation of alpha-ketoglutarate and further promotion of C4-like metabolite exchange (Tang et al., 2024).

The genus Moricandia on the other hand contains C3 and basic C3-C4 (also named C2) species but, so far, no closely related species operating a C4 or even C4-like cycle have been identified in the Brassicaceae (Schlüter et al., 2023). Our snRNA seq data suggest that the photorespiratory shuttle in M. arvensis is softly integrated into the cell-specific leaf biochemistry. Besides the core GDC/SHM enzyme system, shifts in cell specificity were also found for GOX and GGT, suggesting that earlier photorespiratory metabolites such as glycolate and glyocylate could contribute to metabolite exchange between the cells (Fig. 3). In turn, part of the peroxisomal transamination reaction would be directly coupled to the ammonium recapturing reaction in the BSC, thus decreasing the need for additional N-balancing metabolite exchange. In M. arvensis, the shift in GDC activity was accompanied by a parallel strong shift of the plastidial ammonium assimilating enzymes, underlining the importance of this pathway for efficient N recycling in species without additional operation of a C4-like cycle. Additionally, we found BSC-specific expression of GDC interacting pathways, such as redox and energy balancing, transport processes, and C1 metabolism, but no increase in aspartate or alanine aminotransferases or C4-like decarboxylating enzymes.

The lack of a strong C4-like N-balancing shuttle would explain the evolutionary stability of C3-C4 intermediate pathway in M. arvensis. It would be interesting to investigate if similar distribution of photorespiratory enzymes exist in other independently evolved C3-C4 Brassicaceae like Diplotaxis tenuifola. In comparison with the engineering of a C4 pathway into C3 leaves, the C2 cycle in M. arvensis requires fewer genetic adjustments and strengthens the idea that C3-C4 intermediate traits are promising targets for engineering into a C3 crop chassis (Ermakova et al., 2020).

Supplementary Material

eraf245_Supplementary_Data
eraf245_supplementary_data.zip (6.1MB, zip)

Acknowledgements

The authors thank Dr. Yi-Jun Kim for providing the A. thaliana scRNA-seq transcriptome, Samantha Flachbart for assisting in nuclei preparation, Jana Peter for assisting in plant handling and Marion Benecke and Kirsten Hoffie for assisting in immunogold labelling assays. Peter Westhoff (Universität Düsseldorf) and Hermann Bauwe (University Rostock) provided the specific antibodies for immunolocalization. The authors furthermore thank Dr. Tobias Lautwein and Anja Schuster for help and technical support with the 10 × single-nuclei workflow. Computational infrastructure and support were provided by the Centre for Information and Media Technology at Heinrich Heine University Düsseldorf.

Contributor Information

Sebastian Triesch, Institute of Plant Biochemistry, Cluster of Excellence on Plant Science (CEPLAS), Heinrich Heine University, Universitätsstraße 1, Düsseldorf D-40225, Germany.

Vanessa Reichel-Deland, Institute of Plant Biochemistry, Cluster of Excellence on Plant Science (CEPLAS), Heinrich Heine University, Universitätsstraße 1, Düsseldorf D-40225, Germany.

José Miguel Valderrama Martín, Institute of Plant Biochemistry, Cluster of Excellence on Plant Science (CEPLAS), Heinrich Heine University, Universitätsstraße 1, Düsseldorf D-40225, Germany.

Michael Melzer, Structural Cell Biology, Leibniz Institute of Plant Genetics and Crop Plant Research, OT Gatersleben, Corrensstraße 3, Seeland 06466, Germany.

Urte Schlüter, Institute of Plant Biochemistry, Cluster of Excellence on Plant Science (CEPLAS), Heinrich Heine University, Universitätsstraße 1, Düsseldorf D-40225, Germany.

Andreas P M Weber, Institute of Plant Biochemistry, Cluster of Excellence on Plant Science (CEPLAS), Heinrich Heine University, Universitätsstraße 1, Düsseldorf D-40225, Germany.

John Lunn, Max-Planck-Institut für Molekulare Pflanzenphysiologie.

Supplementary data

The following supplementary data are available at JXB  online.

Fig. S1: Replicate UMAP visualization

Fig. S2: Expression of sulfur assimilation pathway associated genes in bundle sheath and mesophyll marker genes.

Fig. S3: Schematic illustration of hypothetical photorespiratory metabolite exchange in Moricandia arvensis.

Fig. S4: Expression patterns of selected genes in Moricandia arvensis and Arabidopsis thaliana bundle sheath and mesophyll cells.

Fig. S5: Expression patterns of selected anatomical regulator genes in Moricandia arvensis and Arabidopsis thaliana bundle sheath and mesophyll cells.

Fig. S6: Transcription factor and cis-regulatory element analysis in the Moricandia arvensis and Arabidopsis thaliana single-cell sequencing data.

Table S1: Antibodies used for immunogold labelling.

Table S2.1: Mean expression count matrix of Moricandia arvensis genes in six clusters

Table S2.2: differential expression between Arabidopsis thaliana and Moricandia arvensis bundle sheath cell and mesophyll cell marker genes.;

Table S2.3: sequencing statistics

Table S2.4: enrichments of common and shared A. thaliana and M. arvensis marker genes in bundle sheath and mesophyll cells.

Author contributions

ST conducted nuclei isolation, data analysis and wrote the manuscript. VRD established the nuclei isolation workflow and assisted in nuclei preparation. JMVM and MM performed immunogold labelling assays. US contributed to writing the manuscript. US and APMW conceptualized the project and supervised the work.

Funding

This work was funded by the Deutsche Forschungsgemeinschaft (German Research Foundation) under Germany’s Excellence Strategy EXC-2084/1 (project ID 390686111) and under the CRC TRR 341 (project ID 456082119).

Data availability

Sequencing data is available in the NCBI sequence read archive (SRA) under the project ID PRJNA1186371 and https://git.nfdi4plants.org/hhu-plant-biochemistry/triesch2025_moricandia_snrna_seq.

References

  1. Adachi  S, Stata  M, Martin  DG, Cheng  S, Liu  H, Zhu  XG, Sage  RF. 2023. The evolution of C4 photosynthesis in Flaveria (Asteraceae): insights from the Flaveria linearis complex. Plant Physiology  191, 233–251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alvarenga  JP, Stata  M, Sage  RF, et al.  2025. Evolutionary diversification of C2 photosynthesis in the grass genus Homolepis (Arthropogoninae). Annals of Botany  135, 769–788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Aubry  S, Smith-Unna  RD, Boursnell  CM, Kopriva  S, Hibberd  JM. 2014. Transcript residency on ribosomes reveals a key role for the Arabidopsis thaliana bundle sheath in sulfur and glucosinolate metabolism. The Plant Journal  78, 659–673. [DOI] [PubMed] [Google Scholar]
  4. Bailey  TL, Boden  M, Buske  FA, Frith  M, Grant  CE, Clementi  L, Ren  J, Li  WW, Noble  WS. 2009. MEME suite: tools for motif discovery and searching. Nucleic Acids Research  37, W202–W208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bakken  TE, Hodge  RD, Miller  JA, et al.  2018. Single-nucleus and single-cell transcriptomes compared in matched cortical cell types. PLoS One  13, e0209648. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bauwe  H, Hagemann  M, Fernie  AR. 2010. Photorespiration: players, partners and origin. Trends in Plant Science  15, 330–336. [DOI] [PubMed] [Google Scholar]
  7. Bauwe  H. 2023. Humboldt review: photorespiration—Rubisco's repair crew. Journal of Plant Physiology  280, 153899. [DOI] [PubMed] [Google Scholar]
  8. Berardini  TZ, Reiser  L, Li  D, Mezheritsky  Y, Muller  R, Strait  E, Huala  E. 2015. The arabidopsis information resource: making and mining the “gold standard” annotated reference plant genome. Genesis  53, 474–485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Berrío  RT, Verstaen  K, Vandamme  N, et al.  2022. Single-cell transcriptomics sheds light on the identity and metabolism of developing leaf cells. Plant Physiology  188, 898–918. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Borba  AR, Reyna-Llorens  I, Dickinson  PJ, et al.  2023. Compartmentation of photosynthesis gene expression in C4 maize depends on time of day. Plant Physiology  193, 2306–2320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Borghi  GL, Arrivault  S, Günther  M, Barbosa Medeiros  D, Dell’Aversana  E, Fusco  GM, Carillo  P, Ludwig  M, Fernie  AR, Lunn  JE, Stitt  M. 2022. Metabolic profiles in C3, C3–C4 intermediate, C4-like, and C4 species in the genus Flaveria. Journal of Experimental Botany  73, 1581–1601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Burgess  SJ, Reyna-Llorens  I, Stevenson  SR, Singh  P, Jaeger  K, Hibberd  JM. 2019. Genome-wide transcription factor binding in leaves from C3 and C4 grasses. The Plant Cell  31, 2297–2314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Busch  FA. 2020. Photorespiration in the context of rubisco biochemistry, CO(2) diffusion and metabolism. The Plant Journal  101, 919–939. [DOI] [PubMed] [Google Scholar]
  14. Castro-Mondragon  JA, Riudavets-Puig  R, Rauluseviciute  I, et al.  2022. JASPAR 2022: the 9th release of the open-access database of transcription factor binding profiles. Nucleic Acids Research  50, D165–D173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Chen  H, Yin  X, Guo  L, Yao  J, Ding  Y, Xu  X, Liu  L, Zhu  Q-H, Chu  Q, Fan  L. 2021. PlantscRNAdb: a database for plant single-cell RNA analysis. Molecular Plant  14, 855–857. [DOI] [PubMed] [Google Scholar]
  16. Conde  D, Triozzi  PM, Pereira  WJ, Schmidt  HW, Balmant  KM, Knaack  SA, Redondo-López  A, Roy  S, Dervinis  C, Kirst  M.  2022. Single-nuclei transcriptome analysis of the shoot apex vascular system differentiation in Populus. Development. Nov 1;149:dev200632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Dai  X, Tu  X, Du  B, et al.  2022. Chromatin and regulatory differentiation between bundle sheath and mesophyll cells in maize. The Plant Journal  109, 675–692. [DOI] [PubMed] [Google Scholar]
  18. Deal  RB, Henikoff  S, Division  S, Hutchinson  F. 2011. The INTACT method for cell type-specific gene expression and chromatin profiling in Arabidopsis. Nature Protocols  6, 56–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Eisenhut  M, Planchais  S, Cabassa  C, et al.  2013. Arabidopsis A BOUT DE SOUFFLE is a putative mitochondrial transporter involved in photorespiratory metabolism and is required for meristem growth at ambient CO₂ levels. The Plant Journal  73, 836–849. [DOI] [PubMed] [Google Scholar]
  20. Ermakova  M, Danila  FR, Furbank  RT, von Caemmerer  S. 2020. On the road to C4 rice: advances and perspectives. The Plant Journal  101, 940–950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Giacomello  S. 2021. A new era for plant science: spatial single-cell transcriptomics. Current Opinion in Plant Biology  60, 102041–102041. [DOI] [PubMed] [Google Scholar]
  22. Grant  CE, Bailey  TL, Noble  WS. 2011. FIMO: scanning for occurrences of a given motif. Bioinformatics  27, 1017–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Groszmann  M, Osborn  HL, Evans  JR. 2017. Carbon dioxide and water transport through plant aquaporins. Plant, Cell & Environment  40, 938–961. [DOI] [PubMed] [Google Scholar]
  24. Guerreiro  R, Bonthala  VS, Schlüter  U, Hoang  NV, Triesch  S, Schranz  ME, Weber  APM, Stich  B. 2023. A genomic panel for studying C3-C4 intermediate photosynthesis in the Brassiceae tribe. Plant, Cell & Environment  46, 3611–3627. [DOI] [PubMed] [Google Scholar]
  25. Guo  Y, Qin  G, Gu  H, Qu  LJ. 2009. Dof5.6/HCA2, a Dof transcription factor gene, regulates interfascicular cambium formation and vascular tissue development in Arabidopsis. The Plant Cell  21, 3518–3534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Heckmann  D, Schulze  S, Denton  A, Gowik  U, Westhoff  P, Weber  APM, Lercher  MJ. 2013. Predicting C4 photosynthesis evolution: modular, individually adaptive steps on a mount fuji fitness landscape. Cell  153, 1579–1588. [DOI] [PubMed] [Google Scholar]
  27. Hibberd  JM, Covshoff  S. 2010. The regulation of gene expression required for C4 photosynthesis. Annual Review of Plant Biology  61, 181–207. [DOI] [PubMed] [Google Scholar]
  28. Holst  F, Bolger  A, Günther  C, et al.  2023. Helixer–de novo Prediction of primary eukaryotic gene models conbining deep learning and a Hidden Marcov Model. BioRxiv doi: 10.1101/2023.02.06.527280 [Preprint].
  29. Hua  L, Stevenson  SR, Reyna-Llorens  I, Xiong  H, Kopriva  S, Hibberd  JM. 2021. The bundle sheath of rice is conditioned to play an active role in water transport as well as sulfur assimilation and jasmonic acid synthesis. The Plant Journal  107, 268–286. [DOI] [PubMed] [Google Scholar]
  30. Hylton  CM, Rawsthorne  S, Smith  AM, Jones  DA, Woolhouse  HW. 1988. Glycine decarboxylase is confined to the bundle sheath cells of leaves of C3-C4 intermediate species. Planta  175, 452–459. [DOI] [PubMed] [Google Scholar]
  31. Kavka  M, Balles  A, Bohm  C, Dehmer  KJ, Fella  C, Rose  F, Saal  B, Schulze  S, Willner  E, Melzer  M. 2024. Phenotypic screening of seed retention and histological analysis of the abscission zone in Festuca pratensis and Lolium perenne. BMC Plant Biology  24, 577. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kim  JY, Symeonidi  E, Pang  TY, et al.  2021. Distinct identities of leaf phloem cells revealed by single cell transcriptomics. The Plant cell  33, 511–530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kirschner  S, Woodfield  H, Prusko  K, Koczor  M, Gowik  U, Hibberd  JM, Westhoff  P. 2018. Expression of SULTR2; 2, encoding a low-affinity sulphur transporter, in the Arabidopsis bundle sheath and vein cells is mediated by a positive regulator. Journal of Experimental Botany  69, 4897–4906. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Kuhnert  F, Schlüter  U, Linka  N, Eisenhut  M. 2021. Transport proteins enabling plant photorespiratory metabolism. Plants (Basel)  10, 880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. La Manno  G, Soldatov  R, Zeisel  A, et al.  2018. RNA velocity of single cells. Nature  560, 494–498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lambret-Frotte  J, Smith  G, Langdale  JA. 2024. GOLDEN2-like1 is sufficient but not necessary for chloroplast biogenesis in mesophyll cells of C4 grasses. The Plant Journal  117, 416–431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Lamesch  P, Berardini  TZ, Li  D, et al.  2012. The Arabidopsis information resource (TAIR): improved gene annotation and new tools. Nucleic Acids Research  40, D1202–D1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Liu  WY, Yu  CP, Chang  CK, et al.  2022. Regulators of early maize leaf development inferred from transcriptomes of laser capture microdissection (LCM)-isolated embryonic leaf cells. Proceedings of the National Academy of Sciences, USA  119: e2208795119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Lundgren  MR. 2020. C2 photosynthesis: a promising route towards crop improvement?  New Phytologist  228, 1734–1740. [DOI] [PubMed] [Google Scholar]
  40. Mallmann  J, Heckmann  D, Bräutigam  A, Lercher  MJ, Weber  APM, Westhoff  P, Gowik  U. 2014. The role of photorespiration during the evolution of C4 photosynthesis in the genus Flaveria. eLife  3, e02478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Mereu  E, Lafzi  A, Moutinho  C, et al.  2020. Benchmarking single-cell RNA-Sequencing protocols for cell atlas projects. Nature Biotechnology  38, 747–755. [DOI] [PubMed] [Google Scholar]
  42. Monson  RK, Rawsthorne  S. 2000. CO2 Assimilation in C3-C4 intermediate plants. In: Leegood  RC, Sharkey  TD, von Caemmerer  S, eds. Photosynthesis. Advances in Photosynthesis and Respiration, vol 9. Dordrecht: Springer, 533–550. [Google Scholar]
  43. Moreno-Villena  JJ, Zhou  H, Gilman  IS, Lori Tausta  S, Maurice Cheung  CY, Edwards  EJ. 2022. Spatial resolution of an integrated C4 + CAM photosynthetic metabolism. Science Advances  8, eabn2349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Morgan  CL, Turner  SR, Rawsthorne  S. 1993. Coordination of the cell-specific distribution of the four subunits of glycine decarboxylase and of serine hydroxymethyltransferase in leaves of C3-C4 intermediate species from different genera. Planta  190, 468–473. [Google Scholar]
  45. Polański  K, Young  MD, Miao  Z, Meyer  KB, Teichmann  SA, Park  JE. 2020. BBKNN: fast batch alignment of single-cell transcriptomes. Bioinformatics  36, 964–965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Procko  C, Lee  T, Borsuk  A, et al.  2022. Leaf cell-specific and single-cell transcriptional profiling reveals a role for the palisade layer in UV light protection. The Plant Cell  34, 3261–3279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Rawsthorne  S, Hylton  CM, Smith  AM, Woolhouse  HW. 1988a. Distribution of photorespiratory enzymes between bundle sheath and mesophyll cells in leaves of the C3-C4 intermediate species Moricandia arvensis (L.) DC. Planta  176, 527–532. [DOI] [PubMed] [Google Scholar]
  48. Rawsthorne  S, Hylton  CM, Smith  AM, Woolhouse  HW. 1988b. Photorespiratory metabolism and immunogold localization of photorespiratory enzymes in leaves of C3 and C3-C4 intermediate species of Moricandia. Planta  173, 298–308. [DOI] [PubMed] [Google Scholar]
  49. Reeves  G, Grangé-Guermente  MJ, Hibberd  JM. 2017. Regulatory gateways for cell-specific gene expression in C4 leaves with Kranz anatomy. Journal of Experimental Botany  68, 107–116. [DOI] [PubMed] [Google Scholar]
  50. Sage  RF, Sage  TL, Kocacinar  F. 2012. Photorespiration and the evolution of C4 photosynthesis. Annual Review of Plant Biology  63, 19–47. [DOI] [PubMed] [Google Scholar]
  51. Sage  RF. 2016. A portrait of the C4 photosynthetic family on the 50th anniversary of its discovery: species number, evolutionary lineages, and Hall of Fame. Journal of Experimental Botany  67, 4039–4056. [DOI] [PubMed] [Google Scholar]
  52. Sage  RF, Monson  RK, Ehleringer  JR, Adachi  S, Pearcy  RW. 2018. Some like it hot: the physiological ecology of C4 plant evolution. Oecologia  187, 941–966. [DOI] [PubMed] [Google Scholar]
  53. Santiago  CP, Gimmen  MY, Lu  Y, McNally  MM, Duncan  LH, Creamer  TJ, Orzolek  LD, Blackshaw  S, Singh  MS. 2023. Comparative analysis of single-cell and single-nucleus RNA-Sequencing in a rabbit model of retinal detachment-related proliferative vitreoretinopathy. Ophthalmology Science  3, 100335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Schlüter  U, Weber  APM. 2016. The road to C4 photosynthesis: evolution of a Complex trait via intermediary states. Plant & Cell Physiology  57, 881–889. [DOI] [PubMed] [Google Scholar]
  55. Schlüter  U, Bräutigam  A, Gowik  U, Melzer  M, Christin  PA, Kurz  S, Mettler-Altmann  T, Weber  APM. 2017. Photosynthesis in C3-C4 intermediate Moricandia species. Journal of Experimental Botany  68, 191–206. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Schlüter  U, Weber  APM. 2020. Regulation and evolution of C 4 photosynthesis. Annual Review of Plant Biology  71, 183–215. [DOI] [PubMed] [Google Scholar]
  57. Schlüter  U, Bouvier  JW, Guerreiro  R, Malisic  M, Kontny  C, Westhoff  P, Stich  B, Weber  APM. 2023. Brassicaceae display variation in efficiency of photorespiratory carbon-recapturing mechanisms. Journal of Experimental Botany  74, 6631–6649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Schneider  CA, Rasband  WS, Eliceiri  KW. 2012. NIH image to ImageJ: 25 years of image analysis. Nature Methods  9, 671–675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Schulze  S, Westhoff  P, Gowik  U. 2016. Glycine decarboxylase in C3, C4 and C3–C4 intermediate species. Current Opinion in Plant Biology  31, 29–35. [DOI] [PubMed] [Google Scholar]
  60. Schwacke  R, Ponce-Soto  GY, Krause  K, Bolger  AM, Arsova  B, Hallab  A, Gruden  K, Stitt  M, Bolger  ME, Usadel  B. 2019. MapMan4: A Refined Protein Classification and Annotation Framework Applicable to Multi-Omics Data Analysis. Molecular Plant Jun 3;12, 879–892 [DOI] [PubMed] [Google Scholar]
  61. Schwarz  N, Armbruster  U, Iven  T, Bruckle  L, Melzer  M, Feussner  I, Jahns  P. 2015. Tissue-specific accumulation and regulation of zeaxanthin epoxidase in Arabidopsis reflect the multiple functions of the enzyme in plastids. Plant & Cell Physiology  56, 346–357. [DOI] [PubMed] [Google Scholar]
  62. Selinski  J, Scheibe  R. 2019. Malate valves: old shuttles with new perspectives. Plant Biology (Stuttgart)  21 Suppl 1, 21–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Singh  P, Stevenson  SR, Dickinson  PJ, Reyna-llorens  I, Tripathi  A, Reeves  G, Schreier  TB, Hibberd  JM. 2023. C4 gene induction during de-etiolation evolved through changes in cis to allow integration with ancestral C 3 gene regulatory networks. Science Advances  9, eade9756. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Slewinski  TL, Anderson  AA, Zhang  C, Turgeon  R. 2012. Scarecrow plays a role in establishing Kranz Anatomy in Maize Leaves. Plant & Cell Physiology  53, 2030–2037. [DOI] [PubMed] [Google Scholar]
  65. Steinegger  M, Söding  J. 2017. MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nature Biotechnology  35, 1026–1028. [DOI] [PubMed] [Google Scholar]
  66. Stiehler  F, Steinborn  M, Scholz  S, Dey  D, Weber  APM, Denton  AK. 2021. Helixer: cross-species gene annotation of large eukaryotic genomes using deep learning. Bioinformatics  36, 5291–5298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Swift  J, Luginbuehl  LH, Hua  L, et al.  2024. Exaptation of ancestral cell-identity networks enables C(4) photosynthesis. Nature  636, 143–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Tang  Q, Huang  Y, Ni  X, Lyu  MJA, Chen  G, Sage  R, Zhu  XG. 2024. Increased α-ketoglutarate links the C3–C4 intermediate state to C4 photosynthesis in the genus Flaveria. Plant Physiology  195, 291–305. [DOI] [PubMed] [Google Scholar]
  69. Triesch  S, Denton  AK, Bouvier  JW, et al.  2024. Transposable elements contribute to the establishment of the glycine shuttle in Brassicaceae species. Plant Biology  26, 270–281. [DOI] [PubMed] [Google Scholar]
  70. Walden  N, German  DA, Wolf  EM, et al.  2020. Nested whole-genome duplications coincide with diversification and high morphological disparity in Brassicaceae. Nature Communications  11, 3795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Walker  B, Schmiege  SC, Sharkey  TD. 2024. Re-evaluating the energy balance of the many routes of carbon flow through and from photorespiration. Plant, Cell & Environment  47, 3365–3374. [DOI] [PubMed] [Google Scholar]
  72. Walsh  CA, Bräutigam  A, Roberts  MR, Lundgren  MR. 2023. Evolutionary implications of C2photosynthesis: how complex biochemical trade-offs may limit C4 evolution. Journal of Experimental Botany  74, 707–722. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Wang  K, Zhao  C, Xiang  S, Duan  K, Chen  X, Guo  X, Sahu  SK. 2023. An optimized FACS-free single-nucleus RNA sequencing (snRNA-Seq) method for plant science research. Plant Science  326, 111535. [DOI] [PubMed] [Google Scholar]
  74. Xia  K, Sun  H-X, Li  J, et al.  2022. The single-cell stereo-seq reveals region-specific cell subtypes and transcriptome profiling in Arabidopsis leaves. Developmental Cell  57, 1299–1310.e4. [DOI] [PubMed] [Google Scholar]
  75. Zhang  YL, Zhang  CL, Wang  GL, Wang  YX, Qi  CH, Zhao  Q, You  CX, Li  YY, Hao  YJ. 2019. The R2R3 MYB transcription factor MdMYB30 modulates plant resistance against pathogens by regulating cuticular wax biosynthesis. BMC Plant Biology  19, 362. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

eraf245_Supplementary_Data
eraf245_supplementary_data.zip (6.1MB, zip)

Data Availability Statement

Sequencing data is available in the NCBI sequence read archive (SRA) under the project ID PRJNA1186371 and https://git.nfdi4plants.org/hhu-plant-biochemistry/triesch2025_moricandia_snrna_seq.

Articles from Journal of Experimental Botany are provided here courtesy of Oxford University Press

RESOURCES