The high efficiency of C4 photosynthesis relies on spatial division of labor, classically with initial carbon fixation in the mesophyll and carbon reduction in the bundle sheath. By employing grinding and serial filtration over liquid nitrogen, we enriched C4 tissues along a developing leaf gradient. This method treats both C4 tissues in an integrity-preserving and consistent manner, while allowing complementary measurements of metabolite abundance and enzyme activity, thus providing a comprehensive data set. Meta-analysis of this and the previous studies highlights the strengths and weaknesses of different C4 tissue separation techniques. While the method reported here achieves the least enrichment, it is the only one that shows neither strong 3′ (degradation) bias, nor different severity of 3′ bias between samples. The meta-analysis highlighted previously unappreciated observations, such as an accumulation of evidence that aspartate aminotransferase is more mesophyll specific than expected from the current NADP-ME C4 cycle model, and a shift in enrichment of protein synthesis genes from bundle sheath to mesophyll during development. The full comparative dataset is available for download, and a web visualization tool (available at http://www.plant-biochemistry.hhu.de/resources.html) facilitates comparison of the the Z. mays bundle sheath and mesophyll studies, their consistencies and their conflicts.
S1, Enzyme activity and metabolite abundance
Enzyme activity and metabolite abundances from enriched bundle sheath, and mesophyll tissues along a developming Zea mays leaf. Data provided from two havests. Fast=two, 8cm slices, less than 1s to liquid nitrogen. Gradient=five, 4cm slice, ~10s to liquid nitrogen. Data is provided both in raw, and fully normalized ("deconvoluted", see methods of paper) form.
Supplemental_dataset1.xls
S2, Functional enrichments.
Significant functional enrichment in gene sets that were otherwise informative in understanding inter-study differences. First four data columns represent a contingency table with "S" standing for "set (of genes)" and "C" standing for "Category (functional)". A Fisher's exact test was used to determine significance, and the false discovery rate calculated using the Benjamini Yekutili method. Remaining columns supply information on the functional category, details of gene set, and direction of enrichment.
Supplemental_dataset2.xls
S3, Compiled RNAseq data
Quantification of gene expression for all RNAseq studies included in the manuscript. Raw unique counts, for each replicate and fully normalized mean Transcripts Per Million are provided for each sample. For the M and BS separation studies, the log2 fold change, and false discovery rate calculated with edgeR between corresponding M and BS tissues is included. For the first study of each species, functional annotations are included. Finally, the sheet Denton_2016 additionally includes the cross species mapping, with best blast hits to Arabidopsis thaliana, Setaria viridis, and Panicum virgatum from the Zea mays genes, and a boolean "RBB" column denoting whether each best blast hit was reciprocal or not.
Supplemental_dataset3.xls
S4, By gene 3' bias.
The 3' bias of RNAseq read coverage for every gene in every study calculated with PicardTools' RnaSeqMetrics, which is defined as " mean coverage of the 3' most 100 bases divided by the mean coverage of the whole transcript" (http://broadinstitute.github.io/picard/picard-metric-definitions.html#RnaSeqMetrics). Missing (NA) values occur when (whole transcript) coverage drops too low.
Supplemental_dataset4.xls
S5, Gene sets of interest
A variety of gene sets discussed in the paper either for their information content in inter-study comparisons, their current relation to C4 photosynthesis, or as strong candidate genes with multi-study support. The first sheet provides descriptions, and the remaining sheets provide gene IDs.
Supplemental_dataset5.xls
Supplemental Table1 Bioinformatis Parameters
Non-default parameters used for bioinformatics tools in this study.
Supplemental_Table1_Parameters.xls
Supplemental_Table2 Numbers of Signifiant Differences
Quantification of the number of significant differences in expression found between BS and M in each study and developmental category. Significance was calculated with edgeR for all studies, and additionally with contamDE for Denton_2016.
Supplemental_Table2_SigStats.xls
Supplemental_Table3 Growth and Harvest Comparison
Compilation of the conditions, measurement and timing used for growth and harvest of plant material as reported in each study.
Supplemental_Table3_Conditions.xls
Fig S1-7 All Supplemental Figures
Figure S1. Setup of developmental slices, enzyme/metabolite data deconvolution method, and validation of separation method.
Figure S2. Metabolite levels in BS and M of the developmental slices (full data in dataset S1).
Figure S3. Contextual data of developmental section pairing, and library complexity for inter-study comparison.
Figure S4. Transcript coverage of genes (or their orthologs) that were significantly enriched in the BS in Chang_2012, but in M in here Denton_2016.
Figure S5. Transcript coverage of examples genes from conflict set 1 (definition and gene list in dataset S5).
Figure S6. Expression of marker genes/categories used to evaluate the co-purification of additional tissues.
Figure S7. AlaAT and 40S ribosome distributions.
SupplementalFigures.pdf