3to4: Converting C3 to C4 photosynthesis for sustainable agriculture

Final Report Summary - 3TO4 (3to4: Converting C3 to C4 photosynthesis for sustainable agriculture)

Executive Summary:
Many crop plants have an inefficient pathway of CO2 assimilation (C3 photosynthesis). All plants rely upon an enzyme for fixing CO2 called Rubisco. The problem with Rubisco is that it not only fixes CO2 but also fixes O2 in a side reaction. The oxygenation reaction of Rubisco leads to the process of photorespiration that involves the light-dependent loss of CO2 from leaves. Photorespiratory losses of carbon are about 25% at 25°C, but considerably higher under warmer, tropical conditions. How can the deficiencies in Rubisco be overcome? Many algae and some land plants increase the CO2 concentration around Rubisco, outcompeting the oxygenase reaction. This can be achieved in terrestrial C4 plants by concentrating CO2 around Rubisco in the cells surrounding the vasculature, termed bundle sheath cells. If photorespiration could be reduced in current C3 crops, or if they could be converted to use C4 photosynthesis, large economic and environmental benefits would ensue because of their increased productivity and the reduced inputs per unit yield associated with the C4 pathway. C4 photosynthesis results in improved rates of carbon fixation, improved nitrogen use efficiency and improved water use efficiency. Efficient C4 photosynthesis is associated with alterations to leaf development, cell biology and biochemistry and so transferring these traits into C3 crops is a long-term undertaking. Despite its complexity, C4 photosynthesis has evolved independently over 66 times and this argues strongly for it being a tractable system to understand and manipulate. The huge advances in agricultural production associated with the Green Revolution were not associated with increases in photosynthesis, and so its manipulation remains an unexplored target for crop improvement both for food and biomass. We have undertaken fundamental research to uncover mechanisms underlying important aspects of the C4 leaf that so far have remained undefined. The aims of 3to4 were to provide leadership for a global network of laboratories that aim to underpin the feasibility of transferring C4 traits into C3 crops, allowing a 50% increase in crop yields required for 2050.

3to4 has advanced our understanding in the following areas of study:

• We isolated mutants of Arabidopsis with altered numbers or sizes of vascular cells, and mutants with smaller chloroplasts in the bundle sheath and mesophyll cells and a large range in thylakoid development. Over-expression of two of the mutants leads to more plasmodesmata between all plasmodesmata-containing cells in the leaf and to an overall increase in the size of the vasculature, which leads to more bundle sheath cells surrounding it.

• A large amount of transcriptomic data collected have identified the core C4 photosynthetic expression traits and defined the regulatory framework of C4 photosynthesis as the C3 photosynthetic regulon. We provide clear evidence that in C3 leaves, genes encoding key enzymes of the C4 pathway are already co-regulated with photosynthesis genes and are controlled by both light and chloroplast-to-nucleus signalling. The temporal, spatial, and developmental dimensions of the data represent a deep resource for the photosynthesis community for future analyses.

• In order to better understand the regulation of the cell-specific gene expression in maize leaves, one of the key features of C4 photosynthesis, we have investigated putative epigenetic and transcriptional factors involved. Altogether, we were able to identify seven novel regulators of cell-specific gene expression of C4 photosynthesis.

• We were able to identify a considerable number of new post-translational modifications occurring in C4 enzymes. We were also able to quantify their abundance over the course of the photoperiod to start addressing how the newly identified post-translational modifications regulate the cell processes during light and dark.

• We developed the first complete dynamic systems model of C4 photosynthesis, and demonstrated the importance of triose phosphate transporter in maintaining C4 flux in an NADP-ME type C4 photosynthesis under dynamic light environments.

• Species of the Brassica Moricandia were used as model system for plants using with a glycine shuttle for limited CO2 concentration in the bundle sheath cells - realization of such a shuttle is seen as the first step on the evolutionary path from C3 to C4 photosynthesis. We tested the effect of the shuttle on the anatomy, transcriptome and metabolome of the leaves under different environmental conditions. Water and radiation use efficiencies, growth and productivity improved under well-watered and water-stressed conditions with the intermediate state. This shows that any attempt to improve photosynthetic capacity of C3 crops will directly increase productivity.

• To engineer the more efficient C4 pathway into rice, the activity of multiple enzymes needs to be increased, and leaf anatomy needs to be modified. Rice lines containing a substantial proportion of the core C4 pathway were generated, and candidate genes responsible for altered leaf anatomy identified. These advances provide a platform from which photosynthesis in rice can be improved in the future.

• We have built capacity for C4 research in Europe by training future generations of researchers

Project Context and Objectives:
If photorespiration could be reduced in current C3 crops, or if they could be converted to use C4 photosynthesis, large economic and environmental benefits would ensue because of both their increased productivity and the reduced inputs per unit yield associated with the C4 pathway. It is important to note that the huge advances in agricultural production associated with the Green Revolution were not associated with increases in photosynthesis, and so its manipulation remains an unexplored target for crop improvement both for food and biomass. Efficient C4 photosynthesis is associated with alterations to leaf development, cell biology and biochemistry, and so transferring these traits into C3 crops is a long-term undertaking. Despite its complexity, C4 photosynthesis has evolved independently at least sixty times, and this argues strongly for it being a tractable system to understand. We have undertaken fundamental research to uncover mechanisms underlying important aspects of the C4 leaf that so far have remained undefined. In the medium term this will allow these important C4 traits to be transferred into C3 crops.

We estimate that the resources and increased understanding of processes underlying C4 photosynthesis generated by 3to4 provides some of the building blocks to establish C4 in C3 crops in about 15-20 years’ time. This would fit with the projected 50% increase in crop yields required for 2050. The aims of 3to4 are to provide leadership for a global network of laboratories and consortia that aim to test the feasibility of transferring C4 traits into C3 crops in order to increase yield potential. We have also built capacity for C4 research in Europe by the training of future generations of researchers.

The specific critical aims of the 3to4 project were to:
• Identify genes that control elaboration of the bundle sheath (BS) in C3 species. Currently, although C4 photosynthesis requires enhanced development of BS cells, no genes associated with BS development have been identified. This means we do not know how BS development is altered in C4 leaves, and so we cannot manipulate it. This critical aim was to provide a framework to do this.
• Define the differences in gene expression between mesophyll (M) and BS cells of closely related C3 and C4 plants. Until we know the extent to which gene expression is altered in M and BS cells of C4 plants compared with C3 plants, it is not possible to assess the extent to which we have engineered characteristics of C4 photosynthesis into C4 leaves. The critical aim here was provide the resources to allow this.
• Identify the important post-translational modifications that are associated with efficient functioning of the C4 cycle. Currently we have an incomplete knowledge of these modifications to enzymes that are important for the C4 cycle. A critical aim here was to define these modifications so that we can manipulate them as necessary.
• Increase photosynthesis using a synthetic photorespiratory by-pass, and placing elements of C4 photosynthesis into C3 crops. Using the photorespiratory bypass as proven, existing technology to enhance photosynthesis in European crops is complementary to the longer term endeavour of integrating C4 photosynthesis into C3 crops. A critical aim here was to provide resources that can be used to increase yields in the short term, as well as develop technology that will likely take 20 years to achieve, but that would have significantly larger benefits.
• Strengthen the European expertise in 3to4 conversion in order that European scientists retain their pre-eminent position in this field. A critical aim here is that 3to4 was provide international leadership in this area, both in terms of cutting edge science, and also in training the next generation of researchers. A further priority of this group was to build a vibrant interface between photosynthesis researchers and the broader community of stakeholders, including Industry.
• Harness the new fundamental knowledge generated in the programme to inform support tools on decisions of how to engineer 3to4.
• Enable the better design of experiments across globe on ways to engineer 3to4.

Work Package Objectives

WP1 Pursue forward genetics with Arabidopsis thaliana as the model species. Parenchymatous bundle sheaths are found in both C3 and C4 species, and their gene regulatory systems appear to be largely conserved. The C3 species Arabidopsis thaliana can therefore be used as the model for the identification of genes that are required for bundle sheath ontogeny and maintenance (BSOM genes). To discover these genes a forward genetic approach will be pursued. Since BS cells of C3 plants are only detectable upon microscopic examination, changes in their properties cannot be easily visualised non-destructively. We will therefore mark BS cells of Arabidopsis concomitantly by expressing firefly luciferase (LUC) and a chloroplast-located GFP (cGFP) under the control of known bundle sheath specific promoters from the glycine decarboxylase P protein gene of the C4 species F. trinervia (Asteraceae) and the sulphate transporter gene (Sult2;2) of A. thaliana. The luciferase activity imaging allows an easy estimation of quantitative deviations in reporter gene activity, but with a low spatial, i.e. cellular, resolution. The chloroplast-located GFP permits a rough estimation of chloroplast numbers. A. thaliana plants containing bundle sheath specific reporter genes will be mutagenized by ethyl methane sulfonate (EMS) or by introduction of activation tags, and mutant lines with deviating amounts of reporter gene expression will be identified and the responsible genes will be isolated.

WP2 Verify gene function in C3 dicot model (Arabidopsis) and crop plants (rapeseed) by a synthetic biology approach. BSOM genes identified in WP1 have firstly to be verified for their function in BS ontogeny and maintenance. Similarly, candidate BSOM genes that have emerged from comparative transcriptome profiling of mesophyll and bundle sheath cells of closely related C3/C4 pairs (WP4) need to be tested for their suspected function. We will pursue a two-tiered verification strategy. Firstly, we will apply gene knock-out and over-expression approaches in Arabidopsis. This allows relatively large numbers of genes to be tested and, since we are aiming to identify genes of universal and conserved function in BS ontogeny and maintenance, Arabidopsis should be a good testing model. Finally, the conserved function of BSOM genes and hence their universal application for the genetic engineering of C4 traits into current C3 crops will then be tested in the dicot crop rapeseed.

WP3 Transcriptomes of C3 and C4 pairs. The major objective of this work package is the identification of candidate genes governing the complex trait C4 photosynthesis through genome-wide comparative transcriptomic analysis of closely related pairs of C3 and C4 plants and of C4 revertants. In this context, it is important to: i) Study multiple reference points to unravel the common principles independent of the phylogenetic context. That is why we have strategically selected pairs of C3 and C4 plants from the monocotyledonous (Panicum spp.) and dicotyledonous (Atriplex spp.) groups. ii) Compare the dynamic changes in the transcriptomes over multiple stages of leaf development since it is expected that the major anatomical differences between C3 and C4 leaves, such as Kranz anatomy, are established early in leaf development, whereas the differences in biochemistry and photosynthesis are manifest later in leaf development. Hence, we will generate a differential time series of gene expression patterns during the ontogeny of leaves in C3 and C4 species to gain insights into the early events giving rise to a functional C4 leaf anatomy and into the late events that are part of maintaining these features in a mature leaf. On the basis of previous experience, we will analyse seven leaf stages, starting with emerging leaf initials and then following the leaf development over 6 consecutive days until a young photosynthetically active source leaf is established. These experiments will be conducted with 3 replicates each (i.e. 7 leaf stages times 3 replicates, 21 libraries per species to be analysed). We will do 4 contrasts: C3 Panicum vs. C4 Panicum, C3 Atriplex prostata vs. C4 Atriplex rosea, Amaranthus edulis C4 mutant, Amaranthus edulis wild type, C4 Alloteropsis semialata, C3 revertant Alloteropsis semialata. This will generate a total of 8 times 21 = 168 libraries to be sequenced, which will require 168 Solexa lanes. For Panicum ssp., we are currently planning P. maximum as the C4 species and Panicum clandestinum (= Dicanthelium clandestinum) as the C3 species. However, since the phylogenetic relationships of the grasses are frequently updated, we will re-assess the choice of species if recent phylogenies indicate more closely related species.
This strategy is complemented by comparative transcriptomic analyses of a C4 mutant in the dicotyledonous species Amaranthus edulis and a C4 revertant in the monocotyledonous species Alloteropsis semialata. These revertants and mutants can be considered as C3 plants in a “C4 body”, which indicates that C3 photosynthesis can proceed effectively in a C4 structural and enzymic background. It is highly likely that this becomes possible through acclimation, possibly by up-regulation of ‘C3’ and photorespiratory genes and down-regulation of ‘C4’ genes, but the extent of acclimation is not known, although there are substantial shifts in metabolism even in heterozygous plants. While the Amaranthus PEPC mutant requires high CO2 concentrations for growth, the Alloteropsis revertant efficiently grows at ambient CO2, which provides us with two complementary reference points for C3 photosynthesis in a C4 structural background. To obtain these datasets, we will make use of Illumina’s sequencing-by-synthesis technology. This technology now permits read lengths of 100 nts, which is sufficiently long for reliably mapping the sequence reads onto reference genomes, such as Arabidopsis or Sorghum. A single Illumina run yields close to 30 Mio sequence reads, thus providing high dynamic range and the ability to discover also low-expressed genes, such as transcription factors, which is crucial for the approach outline here.

WP4 Identify candidate genes by comparative transcriptome profiling of mesophyll and bundle sheath. A crucial feature of almost all C4 plants (i.e. those displaying Kranz anatomy) is a division of photosynthetic labour between two adjacent cell types, namely the bundle sheath and the mesophyll cells. This particular aspect, however, is one of the least understood of the C4 trait, especially since comparisons between closely related C3 and C4 species to date have not yet been possible. Thus, we will complement the time series of transcriptomic changes (WP4) with a comparative analysis of the transcriptomes of mesophyll and bundle sheath cells of closely related C3 and C4 species of Panicum and Atriplex. In principle, it is possible to separate bundle sheath and mesophyll cells after differential treatment of leaves with a mixture of cellulases and pectinases to first release the mesophyll cells. However, it is (i) difficult to achieve a reliable separation of both cell types with this method, in particular for the C3 species and (ii) the generation of protoplasts itself will lead to massive changes in gene expression, due to the release of elicitor-like molecules from the cell walls and the stress associated with removal of the cell wall. To obtain representative gene expression data, it is crucial to effectively quench RNA metabolism before separation of cell types. We will therefore use laser-dissection of fixed and embedded leaf material to generate the material for cell-type specific expression profiles. The same Panicum and Atriplex species will be used as in WP3. Three replicates will be generated for bundle sheath and mesophyll cells of each C3-C4 pair. That makes 6 times 4 = 24 libraries to be generated. Hence, together WPs 3 and 4 will generate 192 libraries for Solexa sequencing. Preparation of leaf tissue for laser dissection will be conducted using Zeiss PALM MicroBeam systems that are available at CAM and UDUS. mRNA will be isolated from pooled dissected material and amplified, and after quality control further processed for Illumina sequencing. As outlined above, bioinformatic and biostatistic data analysis will be performed by partner Nebion.

WP5 Identification and analysis of transcription factors controlling C4-specific gene expression. Functional analysis of transcription factor candidates identified in WPs 4 and 5 will provide lists of candidate genes involved in governing the C4 trait. These factors will be further characterized concerning their binding specificity. Additional factors will be identified in this WP by identification of factors that regulate the specificity of C4 promoters for mesophyll or bundle sheath cells. These analyses will facilitate control of C4 gene expression in transgenic C3 plants. WP5 combines several approaches aiming to identify regulators of C4 genes. This includes the direct identification of promoter-binding proteins and histone-modifiying enzymes. This work will be performed in maize as a C4 model to facilitate interaction with WP6. Due to budget restrictions, we cannot repeat the same work in a second species, namely Setaria. In a second independent approach, we will test transcription factors derived from comparative transcriptomics of closely related C3 and C4 pairs (WP3 and 4). Near relatives to maize or Setaria that perform C3 photosynthesis are not available. Hence, other more suitable pairs of species will be tested here. In the end, all these approaches will result in the identification of candidate gene sequences that will be tested for their ability to transfer C4 features to C3 plants.

WP6 Post-translational regulation of major C4 genes in maize and Setaria. Identifying post-translational modifications of C4 proteins that are crucial for their function. We will generate catalogues of post-translational modifications for the major C4 proteins. We will identify the functional importance of these modifications and we will test them in transgenic C3 plants. This will guide the design of gene sequences for gene transfer in WP 9. C4 plant biology has suffered from the fact that no easily manageable genetic model system has previously been available. The C4 grass Setaria promises to be such a model system. The Panicoid C4 grass Setaria viridis is a new C4 model that allows a forward-genetic approach for identifying loss-of function mutation in regulatory genes of both C4 anatomy and physiology. S. viridis is of small stature, self-compatible, has a short life cycle and can be transformed by Agrobacterium tumefaciens. The plant is a noxious weed and the wild progenitor of foxtail millet (S. italica). S. viridis has a small genome that is currently being sequenced. Because most of our agronomically important crop species are monocotyledonous plants, monocotyledonous models are required for developing the paradigms for engineering such crops. WP 6 aims at identification of post-translational modifications and testing their importance in transgenic plants. Such plants will be generated in WP9. Since Setaria is transformable with Agrobacterium and has a much shorter generation cycle than maize, it is the model of choice for the generation of transgenics to test hypotheses developed in WP6. Another practical reason for using the model Setaria is its small genome, which makes the bioinformatics analyses more straightforward because fewer gene duplications are expected. Hence unequivocal identification of target genes is facilitated.

WP7 Development of a generic model of C4 photosynthesis. We will develop and parameterize a generic C4 photosynthesis model based on large-scale determination of physiological data under different conditions. We will make use of this model to guide C4 engineering. We will develop a kinetic systems model of Kranz type C4 photosynthesis including a detailed description of photosynthetic carbon metabolism, light reactions, and the major diffusion process of CO2 from the atmosphere to the site of CO2 fixation. The model will be developed for maize. The model will first be used to identify the key features required to gain the high photosynthetic efficiency in maize, including the optimal distribution of enzymes into different enzymes, the optimal rate of fluxes for the metabolite transporters, the optimal distribution of enzymes involved in the Calvin cycle between bundle sheath and mesophyll cells. The model can help to identify bottlenecks of C4 photosynthesis and guide the design of transgenic lines in WP9 as well as the selection of putative regulatory factors in WP5.

WP8 Understanding Model Systems 1. We will transfer the photorespiratory by-pass established in one of the applicants' labs to the important European crop rapeseed. Previous work suggests that over-expression of glycollate dehydrogenase, the first enzyme of the synthetic pathway, is sufficient to enhance growth and reduce photorespiratory losses. Plant chloroplasts are capable of further converting glyoxylate, the product of the initial reaction. Lines of rapeseed containing the synthetic photorespiratory bypass will be analysed.
2. Impact of a natural photorespiratory by-pass in Brassicaceae (Moricandia). A natural photorespiratory by-pass exists in a number of so-called C3-C4 intermediate plants. In its simplest form this involves the loss of one or more glycine decarboxylase subunits from the mesophyll so that a photorespiratory glycine shuttle operates from the mesophyll cells to a bundle sheath. In these plants the bundle sheath is both enlarged and has an increased complement of organelles, particularly of the mitochondria where glycine decarboxylation occurs. By operating this shuttle these intermediates are able to recycle a portion of photorespiratory CO2, thereby lowering the CO2 compensation point. This shift in glycine decarboxylation to the bundle sheath is a feature of all C4 plants studied and has been postulated to be the first step in the evolution of the C4 pathway. Accordingly it is currently being manipulated in the C4 rice programme and in WP9 as one of the essential elements of the C4 pathway. To understand the impact of shifting the location of glycine decarboxylation on photosynthetic and non-photosynthetic processes and metabolism, plants of the genus Moricandia will be analysed in detail. Moricandia is a member of the Brassicaceae and therefore closely related to rape, which is a European crop. M. moricandioides, four accessions of M. arvensis, M. suffruticosa, and M. nitens are currently under study in Düsseldorf. Seeds of these lines will be distributed to the partners. Moricandia arvensis (C3-C4) and M. moricandioides (C3) are routinely transformable and could be complemented in Düsseldorf (i.e. GDC re-expressed in the mesophyll). Gas-exchange was measured on Moricandia leaves in the 1980s, however, now gas-exchange measurements in an eco-physiological context can be interwoven with transcriptome analyses. To address how each line partitions available carbon between organs we will investigate growth, water use and analysis of stomatal function, productivity and partitioning of dry matter.
3. Stomatal tuning in C3/C4 metabolism backgrounds using Cleome. Another advantage of C4 photosynthesis is water economy. Under both optimal and suboptimal conditions, as a consequence of efficient carbon concentrating mechanisms, C4 photosynthesis commonly operates with lower stomatal conductance than C3 photosynthesis. Moreover, stomata response occurs at different CO2 concentration in the mesophyll for C3 and C4 plants (Ci/Ca ratio is markedly lower in C4 than in C3 leaves). This implies that introducing C4 traits into C3 plants has to consider not only C4 key concentrating mechanisms but also the molecular and development basis of hydraulic control which includes guard cells behaviour and signalling in C3/C4 metabolism backgrounds. With the aim to unravel candidate genes controlling C4 stomatal response traits, we will investigate the extent to which gene expression differs in guard cells of C4 versus C3 plants in parallel with an analysis of stomatal function on two closely related species: Tarenaya hassleriana and Cleome gynandra that show C3 and C4 metabolism, respectively.
Note that WP8 objectives were modified significantly from the original proposal.

WP9 Build the C4 pathway in rice, both in terms of the metabolic pathway and in generating the Kranz anatomy 1. Generate rice lines that accumulate components of the C4 pathway. Rice lines containing 13 maize genes associated with the C4 pathway have been generated in the initial phase of the C4 rice project and are available for this WP. These lines include those with genes encoding core C4 cycle proteins (carbonic anydrase, phosphenolpyruvate carboxylase, malate dehydrogenase, NADP-dependent malic enzyme and pyruvate,orthophosphate dikinase), those with genes encoding metabolite transporters (DIT1&2), and those in which artificial microRNAs have been used to restrict the accumulation of glycine decarboxylase and Rubisco to BS cells. To build on and add value to these resources we aim to pyramid existing lines. Specifically, existing lines will be crossed to generate transgenic lines in which PEPC, MDH and PPDK are combined with low amounts of Rubisco and glycine decarboxylase in mesophyll cells, and NADP-ME and Rubisco accumulate in BS cells. We will also generate a tag that marks nuclei of rice in either M or BS cells, and so allows gene expression in these cells to be assessed. These lines in which the M or BS cells are tagged will be crossed with existing transgenics in order to follow alterations in gene expression in these cell types.
2. Characterize putative regulators of Kranz anatomy. Maize foliar leaves develop classical C4 Kranz anatomy but the husk leaves surrounding the ear display characteristics of C3 leaves (Langdale et al., 1988). As part of the C4 rice program, comparative RNA-sequencing analysis of foliar and husk leaf primordia has identified candidate positive regulators of Kranz anatomy. Depending on the filtration criteria used, the number of candidate genes ranges from 25 to 100. To characterize these genes, we will systematically knock each one down with miRNA in the C4 model grass Setaria viridis, and over-express each one in C3 rice. We will start with the 25 genes identified by the most stringent filtration criteria.

WP11 Training of researchers The overall aim of the training activities to produce a cohort of young researchers with a novel blend of applied and fundamental R&D skills that underpin future developments in improving photosynthesis. The 3to4 Training Programme will provide, through summer schools, multidisciplinary practical courses to our researchers in key state of the art techniques, including metabolomics, proteomics and bioinformatics. This will be reinforced by secondments that will enable researchers to spend more extended periods in partner laboratories both in the public and private sectors. Researchers will also be trained in complementary skills and will be strongly encouraged to participate in the organization of network meetings, training events and outreach in order to give them the opportunity to develop organizational and communication skills. They will also develop their communication skills by the opportunities for participation in, and presentation at, international conferences and by the preparation of publications for reports and scientific journals. We will utilize the substantial expertise of senior participants in the 3to4 consortium to fulfil the following specific objectives: 1. design and deliver a series of specialist summer schools for researchers. 2. organize secondments for researchers involved in 3to4 to facilitate specialist training in technologies/techniques.

Project Results:
WP1 Pursue forward genetics with Arabidopsis thaliana as the model species
Objectives. Parenchymatous bundle sheaths are found in both C3 and C4 species, and their gene regulatory systems appear to be largely conserved. The C3 species Arabidopsis thaliana can therefore be used as the model for the identification of genes that are required for bundle sheath ontogeny and maintenance (BSOM genes). To discover these genes a forward genetic approach will be pursued. Since BS cells of C3 plants are only detectable upon microscopic examination, changes in their properties cannot be easily visualised non-destructively. We will therefore mark BS cells of Arabidopsis concomitantly by expressing firefly luciferase (LUC) and a chloroplast-located GFP (cGFP) (Sattarzadeh et al., 2010) under the control of known bundle sheath specific promoters from the glycine decarboxylase P protein gene (GLDPA) of the C4 species F. trinervia (Asteraceae) (Engelmann et al., 2008) and the sulphate transporter gene (Sult2;2) of A. thaliana (Takahashi et al., 2000). The luciferase activity imaging allows an easy estimation of quantitative deviations in reporter gene activity, but with a low spatial, i.e. cellular, resolution. The chloroplast-located GFP permits a rough estimation of chloroplast numbers. A. thaliana plants containing bundle sheath specific reporter genes will be mutagenized by ethyl methane sulfonate (EMS) or by introduction of activation tags, and mutant lines with deviating amounts of reporter gene expression will be identified and the responsible genes will be isolated.

TASK 1.1: Construction of Arabidopsis thaliana reporter lines (UDUS). Since bundle sheath (BS) cells of C3 plants are only detectable upon microscopic examination, changes in their properties cannot be easily visualized non-destructively. We therefore marked BS cells of A. thaliana concomitantly by expressing firefly luciferase (LUC) and a chloroplast-located GFP (Sattarzadeh et al., 2010) under the control of the promoter from the glycine decarboxylase P protein gene (GLDPA) of the C4 species Flaveria trinervia (Asteraceae). It had been shown that this promoter is only active in the bundle sheath and vasculature of A. thaliana leaves (Engelmann et al., 2008).
Both reporter lines were transformed into A. thaliana (ecotype Col-0). Lines with single insertions and high reporter gene activities were identified, and homozygous lines were obtained by selfing. The reporter gene activity is restricted to the bundles in the leaves. In addition, the chloroplast-located GFP permits a rough estimation of chloroplast numbers, which allows us to search for mutants with an altered number of BS chloroplasts in our genetic screens.

TASK 1.2: EMS mutagenesis of Arabidopsis reporter lines and identification of mutant lines. 160,000 seeds were mutagenized with EMS (40,000 LUC reporter line, 120,000 GFP reporter line) and sown on soil in large flats under greenhouse conditions. M1 plants were harvested in pools of about 50 plants. Each M2 pool was sown individually on large flats in the greenhouse and single leaves or whole seedlings were screened for aberrant reporter gene expression, e.g. stronger or weaker reporter gene signal in the BS. A total number of 755 primary mutants were identified. The phenotype of each mutant line was checked in M3 and could be confirmed for 85 LUC mutants and 145 GFP mutant lines. Only mutants with at least 30% increased or reduced reporter gene activity in the leaf were selected, checked for point mutations in the reporter gene sequence, and used for further analysis. This resulted in 56 mutant lines. Subsequently, the most promising 26 mutant lines were selected for high-quality cross sections with the light microscope (LM). Most mutants revealed altered numbers or sizes of cells belonging to xylem or phloem. We observed mutants with smaller chloroplasts in the BS and M cells and a large range in thylakoid development. Additionally, there are a couple of mutant lines with more and smaller BS cells. In most cases this seems to be due to an increased volume of the vasculature. Ten mutant lines with different phenotypes were selected for ultrastructural imaging by the use of transmission electron microscopy (TEM) to gain further insight into the anatomy of the BS, the vasculature, and the organelles in the cells.

TASK 1.3: Positional cloning of BSOM genes (UDUS). To map the genes, we generated backcross populations for all our mutant lines in order to get a segregating F2 population in which we can screen for the particular aberrant phenotype. At least 50 individual plants were selected and DNA was isolated and pooled. We created sequencing libraries for the first eight mutant lines and they were sequenced via Illumina technology. This resulted in 2-10 potential candidate genes for each mutant line, which will now further be analyzed with CRISPR/Cas9. The sequencing of the remaining mutant lines is in progress. To conclude, we could show that using backcross population to map EMS-induced point mutations resulted in a manageable number of candidate genes.

TASK 1.4: Construction of activation constructs (UDUS). The pGLDTFt promoter from F. trinervia was used as an activation-tag to activate nearby genes. A BASTA resistance gene close to the LB site of the T-DNA allowed for rapid screening of positive transformants in the M1 generation. The activation-tag was introduced into plants of the GFP reporter line.

TASK 1.5: Isolation of BSOM genes by activation tagging (UDUS). Approximately 50,000 M1 transformants were screened for changes in reporter gene activity, which resulted in 175 primary mutants. Eight stable homozygous mutant lines were isolated of the activation tagging screen. Besides the differences in fluorescence signal, we identified two mutant lines that also contained larger bundles. We mapped the insertion sites of the activation-tag with inverse PCR and Thermal Asymmetric Interlaced PCR, which resulted in six candidate genes.

WP2 Verify gene function in C3 dicot model (Arabidopsis) and crop plants (rapeseed) by a synthetic biology approach
Objectives. BSOM genes identified in WP1 have firstly to be verified for their function in BS ontogeny and maintenance. Similarly, candidate BSOM genes that have emerged from comparative transcriptome profiling of mesophyll and bundle sheath cells of closely related C3/C4 pairs (WP4) need to be tested for their suspected function. We will pursue a two-tiered verification strategy. Firstly, we will apply gene knock-out and over-expression approaches in Arabidopsis. This allows relatively large numbers of genes to be tested and, since we are aiming to identify genes of universal and conserved function in BS ontogeny and maintenance, Arabidopsis should be a good testing model. Finally, the conserved function of BSOM genes and hence their universal application for the genetic engineering of C4 traits into current C3 crops will then be tested in the dicot crop rapeseed (B. napus).

TASK 2.1: Verification of BSOM gene function in Arabidopsis (UDUS). To verify gene functions of our first BSOM genes (BSOM 1+2), we overexpressed them in the bundle sheath and in the whole leaf of the GFP reporter line using the appropriate promoters. Additionally, we designed gene knock out constructs using the CRISPR/Cas9 system. Two of three BSOM phenotypes that have been analyzed could be reconstructed by overexpressing the candidate gene using a bundle sheath specific promoter. Both overexpressing lines also contain larger bundles as well as increased reporter gene activity in the leaf.

TASK 2.2: Analysis of candidate gene function in Arabidopsis (UDUS). LM and TEM techniques were used to better understand the effects at the cellular level in leaves of our BSOM 1+2 overexpression lines. It could be shown that the overexpression of BSOM1 leads to more plasmodesmata between all plasmodesmata-containing cells in the leaf. BSOM2 seems to have a function in vasculature differentiation since its overexpression leads to an impaired xylem formation. As a consequence, more and irregular formed parenchyma cells appear within the xylem. The overall size of the vasculature is increased as well, which leads to more bundle sheath cells surrounding it.

TASK 2.3: Verification of BSOM gene function in rapeseed (UDUS). Instead of using rapeseed, we decided to analyse the gene function of our BSOM genes in Kitaake rice instead. Therefore, we made constructs to overexpress the gene (the Arabidopsis gene and the rice orthologue) with the PEPCK promoter of Zoysia japonica, which confers bundle sheath specific expression in rice. Additionally, we made RNAi knock-down constructs. The constructs for BSOM1+2 are ready and were transformed into rice. T1 lines for all constructs will be analysed in 01/2017.

WP3 and WP4 Transcriptomes of C3 and C4 pairs and transcriptomics of bundle sheath and mesophyll
Objectives. The major objective of this work package is the identification of candidate genes governing the complex trait C4 photosynthesis through genome-wide comparative transcriptomic analysis of closely related pairs of C3 and C4 plants and of C4 revertants. In this context, it is important to: i) Study multiple reference points to unravel the common principles independent of the phylogenetic context. That is why we have strategically selected pairs of C3 and C4 plants from the monocotyledonous (Panicum spp.) and dicotyledonous (Atriplex spp.) groups. ii) Compare the dynamic changes in the transcriptomes over multiple stages of leaf development since it is expected that the major anatomical differences between C3 and C4 leaves, such as Kranz anatomy, are established early in leaf development, whereas the differences in biochemistry and photosynthesis are manifest later in leaf development. Hence, we will generate a differential time series of gene expression patterns during the ontogeny of leaves in C3 and C4 species to gain insights into the early events giving rise to a functional C4 leaf anatomy and into the late events that are part of maintaining these features in a mature leaf. On the basis of previous experience, we will analyse seven leaf stages, starting with emerging leaf initials and then following the leaf development over 6 consecutive days until a young photosynthetically active source leaf is established. These experiments will be conducted with 3 replicates each (i.e. 7 leaf stages times 3 replicates, 21 libraries per species to be analysed). We will do 4 contrasts: C3 Panicum vs. C4 Panicum, C3 Atriplex prostata vs. C4 Atriplex rosea, Amaranthus edulis C4 mutant, Amaranthus edulis wild type, C4 Alloteropsis semialata, C3 revertant Alloteropsis semialata. This will generate a total of 8 times 21 = 168 libraries to be sequenced, which will require 168 Solexa lanes. For Panicum ssp., we are currently planning P. maximum as the C4 species and Panicum clandestinum (= Dicanthelium clandestinum) as the C3 species. However, since the phylogenetic relationships of the grasses are frequently updated, we will re-assess the choice of species if recent phylogenies indicate more closely related species.
This strategy is complemented by comparative transcriptomic analyses of a C4 mutant in the dicotyledonous species Amaranthus edulis (Dever et al., 1995) and a C4 revertant in the monocotyledonous species Alloteropsis semialata (Ibrahim et al., 2009). These revertants and mutants can be considered as C3 plants in a “C4 body”, which indicates that C3 photosynthesis can proceed effectively in a C4 structural and enzymic background. It is highly likely that this becomes possible through acclimation, possibly by up-regulation of ‘C3’ and photorespiratory genes and down-regulation of ‘C4’ genes, but the extent of acclimation is not known, although there are substantial shifts in metabolism even in heterozygous plants. While the Amaranthus PEPC mutant requires high CO2 concentrations for growth, the Alloteropsis revertant efficiently grows at ambient CO2, which provides us with two complementary reference points for C3 photosynthesis in a C4 structural background. To obtain these datasets, we will make use of Illumina’s sequencing-by-synthesis technology. This technology now permits read lengths of 100 nts, which is sufficiently long for reliably mapping the sequence reads onto reference genomes, such as Arabidopsis or Sorghum. A single Illumina run yields close to 30 Mio sequence reads, thus providing high dynamic range and the ability to discover also low-expressed genes, such as transcription factors, which is crucial for the approach outline here.

Genome-wide comparative transcriptomic analysis of closely related pairs of C3 and C4 plants. Multiple reference points revealed the common principles independent of the phylogenetic context. Strategically selected C3 and C4 plants from the monocotyledonous (Megathyrsus maximus, Dicanthelium clandestinum, Panicum milliaceum, Alloteropsis semialata, Alloteropsis ecklonia, Setaria italica, Hordeum vulgare, Hymenachne amplexicaulis) and dicotyledonous (Amaranthus edulis, Atriplex rosea, Atriplex prostrata, Flaveria bidentis, Flaveria trinervia, Flaveria robusta, Flaveria pringlei, Gynandropsis gynandra, Tarenaya hassleriana) groups were sequenced and the data was integrated and analyzed with publicly available data and data generated with other funding sources (Flaveria bidentis, Flaveria trinervia, Flaveria robusta, Flaveria pringlei, Gynandropsis gynandra, Tarenaya hassleriana, Zea mays, Oryza sativa, Echinocloa glabrecence) . Since major anatomical differences between C3 and C4 leaves, such as Kranz anatomy, are established early in leaf development, while differences in biochemistry and photosynthesis are manifested later in leaf development, the dynamic changes in the transcriptomes over multiple stages of leaf development were also analyzed (M. maximus vs D. clandestinum, G. gynandra vs. T. hassleriana, Flaveria bidentis and Flaveria trinervia vs Flaveria robusta and Flaveria pringlei).
All RNA-seq data from 3to4 and additional data from public sources were analyzed using a pipeline developed by Nebion. The pipeline consists of sequencing, upload of the raw reads, data storage, quality control (FASTQC), read mapping (BLATX for cross-species mapping) and/or de-novo assembly (Trinity), expression quantification (RSEM), data analysis (for example differential expression).

The C4 cycle is evolutionarily fixed, it is modular, and it is defined by the carboxylation enzymes. Comparative transcriptomic analyses of a mutant in the dicotyledonous species Amaranthus edulis affected in the major C4 gene PPC resulting in only 5% remaining PEPC activity demonstrated that the mutant is unable to compensate for any changes in the C4 cycle. It was originally hypothesized that signs of reversion to C3 photosynthesis can be detected in the mutant leading to the potential identification of regulatory factors controlling C4 and C3 photosynthesis. The mutant and wildtype plants were grown alongside and sampled at the end of the night as well as two and seven hours into the light to capture the onset of daytime photosynthesis. Unexpectedly, transcript abundances of genes encoding core C4 cycle enzymes or central carbon metabolism were largely unaffected by the mutation in PPC and the mutant was unable to compensate the loss of PEPC through up-regulation of photorespiratory or Calvin Benson Bassham cycle gene expression in either elevated CO2 conditions or in ambient air. A disturbed starch turnover during the night led to a starvation response at the end of the night. The C4 cycle genes with the exception of the NHD showed diurnal patterns similar to other genes under diurnal control such as the photosynthetic genes. The diurnal patterns again suggested that the metabolic C4 genes employ transcriptional control mechanisms already present in C3 plants. Within the 3to4 framework, evolution was also studied based on RNA-seq data using Flaveria species (Mallmann et al., 2014), Moricandia species (Schlüter et al., 2017) (Bräutigam and Gowik, 2016).
Comparatively analysis of a limited set of monocotyledonous and dicotyledonous C4 species in comparison with sister C3 species based on RNA-seq results from the Flaverias, Cleomaceae, M. maximus and D. clandestinum, and publicly available datasets from Z. mays , S. italica, S. bicolor, Oryza sativa, and Brachypodium distachyon initially tested the subtype concept. A mixed NAD-ME and PEP-CK subtype, M. maximus, was the simplest of all subtypes in the analysis (Bräutigam et al., 2014). Modelling analysis based on the results showed that the current model does not support a PEP-CK only C4 cycle (Wang et al., 2014) Integrating all species from the 3to4 project further refined the results. The decarboxylation enzyme predicted the subcellular localization of the transfer acid generating enzymes but not necessarily whether they were amino transferases or malate dehydrogenases or both. Species with high transcript abundance of PEP-CK indicated the presence of an additional C4 cycle enzyme, PEP phosphatase, which is required to complete a C4 cycle with little or no malic enzyme activity (manuscript in preparation). PEP phosphatase resolves how species with predominantly PEP-CK activity sustain the C4 cycle (manuscript in preparation).
Within the project, the guard cell transcriptome was also analysed (Aubry et al., 2016)

Regulation of C4 cycle genes is likely shared with the photosynthesis regulon. Three leaf developmental gradients were analyzed in the context of the 3to4 project of which two were sequenced using 3to4 financing: (i) previously sequenced leaf sections of the dicotyledonous Cleomaceae were analyzed (Kulahoglu et al., 2014) (ii) for leaves of two sister pairs of the dicotyledonous Flaveriaceae (contrasting the C3 plants F. pringlei and F. robusta with the C4 plants F. bidentis and F. trinervia) a base to tip maturation gradient of leaf anatomy with six sections was defined, and RNA sequencing was undertaken along this gradient (Kümpers et al., 2017) and (iii) leaves of the monocotyledonous Dichanthelium clandestinum (C3, formerly Panicum clandestinum) and Megathyrsus maximus (C4, synonymous with Panicum maximum) were sliced into 11 sections, samples were deep-sequenced and mapped cross-species onto the Setaria italica reference.
The Cleomaceae gradient comparison showed that the overall gene expression profiles appear remarkably similar between the C3 and C4 species. The known C4 genes were recruited to photosynthesis from different expression domains in C3, including typical housekeeping gene expression patterns in various tissues as well as individual heterotrophic tissues. A structure-related module was recruited from the C3 root. Comparison of gene expression patterns with anatomy during leaf ontogeny provided insight into genetic features of Kranz anatomy. A delay in mesophyll differentiation apparent both in the leaf anatomy and the transcriptome allows for extended vein formation in the C4 leaf (Kulahoglu et al., 2014).In the analysis of the Flaveria gradient, key C4 traits including vein density, mesophyll and bundle sheath cross-sectional area, chloroplast ultrastructure, were integratively analyzed with the abundance of transcripts encoding proteins of C4 photosynthesis were quantified.
Principal components analysis indicated that leaf maturation and the photosynthetic pathway were responsible for the greatest amount of variation in transcript abundance. Photosynthesis genes were over-represented for a prolonged period in the C4 species. Through comparison with the other available data sets, a small number of transcriptional regulators were identified that have been up-regulated in diverse C4 species. The analysis identifies similar patterns of expression in independent C4 lineages and so indicates that the complex C4 pathway is associated with parallel as well as convergent evolution (Kümpers et al. 2017). The gradient in the Paniceae species showed similar results. The C4 cycle genes have expression patterns similar to the photosynthetic genes in C3 species and co-cluster with them (unpublished). Taken together the comparative leaf gradient analyses have demonstrated that the C4 cycle genes are likely hooked into the regulatory network of photosynthetic genes and thus likely share the same regulators. Candidate genes for the control of C4 photosynthesis include sigma transcription factors, BBX-type transcription factors, and a Acyl-CoA N-acyltransferases superfamily protein as well as other nucleic acid binding proteins (Külahoglu et al. 2014, Kümpers et al. 2017, Aubry et al. 2015, unpublished).
Across the monocot, dicot divide, three independent approaches identify a core set of transcription factors which co-cluster with C4 photosynthetic genes and are thus considered. Despite the wide evolutionary separation and independent origins of the C4 phenotype, these species use homologous transcription factors to both induce C4 photosynthesis and to maintain the cell specific gene expression required for the pathway to operate (Aubry et al., 2014). A core molecular signature associated with leaf and photosynthetic maturation that is likely shared by angiosperm species derived from the last common ancestor of the monocotyledons and dicotyledons was defined (Aubry et al., 2014). Cell-type specific expression was also studied in Sorghum bicolor (Covshoff et al., 2013). candidates for controlling C4 photosynthetic genes. Bundle sheath and mesophyll specific expression was analyzed together with leaf gradient data. A discontinuous leaf gradient of Gynandropsis gynandra and cell type specific data (C4, synonymous with Cleome gynandra) consisting of four major zones was compared to publicly available leaf gradients of the C4 monocotyledonous Zea mays (Li et al. 2010 and Chang et al. 2012). A new informatics framework allowed comparison of gene expression in species representing two independent C4 lineages (Cleome gynandra and Zea mays) whose last common ancestor diverged 140 million years ago. Similarly to the previous leaf gradients, we showed that genes encoding proteins of the C4 cycle are recruited into networks defined by photosynthesis-related genes.
First results of above-mentioned sub-projects proposed that C4 genes become integrated into existing photosynthesis gene regulatory networks, we thus sought to better understand how this occurs. We performed de-etiolation experiments in C3 Arabidopsis thaliana and C4 Gynandropsis gynandra to quantify the extent to which light-regulation of C4 genes evolves de novo with the appearance of the C4 trait (Burgess et al., 2016). We could provide clear evidence that in C3 leaves, genes encoding key enzymes of the C4 pathway are already co-regulated with photosynthesis genes and are controlled by both light and chloroplast-to-nucleus signalling. In C4 leaves this regulation becomes increasingly dependent on the chloroplast. The regulation of C4 cycle genes by light and the chloroplast in the ancestral C3 state has likely facilitated the repeated evolution of the complex and convergent C4 trait.
In summary, the data collected in WP3 and WP4 identified the core C4 photosynthetic expression traits and defined the likely regulatory framework of C4 photosynthesis as the C3 photosynthetic regulon. The temporal, spatial, and developmental dimensions of the data represent a deep resource for both the C3 and C4 communities for future analyses.

WP 5 Identification and analysis of transcription factors controlling C4-specific gene expression (iBET)
Objectives. Functional analysis of transcription factor candidates identified in WPs 4 and 5 will provide lists of candidate genes involved in governing the C4 trait. These factors will be further characterized concerning their binding specificity. Additional factors will be identified in this WP by identification of factors that regulate the specificity of C4 promoters for mesophyll or bundle sheath cells. These analyses will facilitate control of C4 gene expression in transgenic C3 plants. WP5 combines several approaches aiming to identify regulators of C4 genes. This includes the direct identification of promoter-binding proteins and histone-modifiying enzymes. This work will be performed in maize as a C4 model to facilitate interaction with WP6. Due to budget restrictions, we cannot repeat the same work in a second species, namely Setaria. In a second independent approach, we will test transcription factors derived from comparative transcriptomics of closely related C3 and C4 pairs (WP3 and 4). Near relatives to maize or Setaria that perform C3 photosynthesis are not available. Hence, other more suitable pairs of species will be tested here. In the end, all these approaches will result in the identification of candidate gene sequences that will be tested for their ability to transfer C4 features to C3 plants.

The task in WP5 was the identification and functionally characterization of transcription factors (TFs) binding to the ZmPEPC1 and ZmC4-NADP-ME gene promoters. In order to identify these TFs, a maize leaf cDNA expression library was constructed and screened using the Yeast-One-Hybrid (Y1H) system.
Three maize TFs belonging to different TF families were identified in the Y1H screenings as binding to the ZmPEPC1 promoter: homeobox (ZmHB87), orphan (ZmOrphan94), and cystein-rich polycomb protein (ZmCPP8). Regarding ZmC4-NADP-ME, three bHLH TFs (ZmbHLH128, ZmbHLH129, and ZmbHLH172) were found as interacting with its promoter.
In addition, knowing that ZmPEPC1 promoter confers mesophyll cell specificity and is light regulated in rice (Matsuoka et al., 1994), we searched for rice TFs binding to maize PEPC1 promoter. This was carried out using the ZmPEPC1 promoter fragments as baits to screen (Y1H) a rice cDNA expression library. This study allowed the identification of one rice OsbHLH112 TF binding to maize PEPC1 promoter. Moreover, we showed that maize OsbHLH112 orthologs, ZmbHLH80 and ZmbHLH90, also bind to the ZmPEPC1 promoter.

Characterization of maize TFs identified as binding to the ZmPEPC1 promoter. The interaction between the TFs ZmbHLH90 and ZmOrphan94 and the ZmPEPC1 promoter was validated by EMSA. The interaction between ZmbHLH90 or ZmOrphan94 and ZmPEPC1 promoter occurs in the E-box (5’-CACATG-3’) located 200bp upstream of ATG. In addition, ZmOrphan94 was also found to bind to the morning element (ME) (5’-CCACAC-3’).
In order to characterize the TF cell-specific gene expression, the expression profile of the novel TF genes identified was analysed in the mesophyll (M) and bundle sheath (BS) cells at dawn, when the ZmPEPC1 transcript level reaches its peak. All the TFs identified showed to be more expressed in the M cells with the exception of ZmOrphan94 that did not show a clear gene expression pattern. In addition, the diurnal expression pattern of the TFs was also assessed. Both ZmbHLH90 and ZmOrphan94 as well as the target gene ZmPEPC1 showed similar gene expression profiles.
Since protein – protein interaction is an important aspect influencing TF transcriptional activity, we investigated whether the maize transcription factors identified as binding to ZmPEPC1 promoter form homo and/or heterodimers. We performed direct Yeast Two-Hybrid (Y2H) and Bimolecular Fluorescence Complementation (BiFC) assays in maize mesophyll protoplasts. Six TF-TF interactions were identified: ZmbHLH90 and 80 showed to form homo- and heterodimers; and ZmOrphan94 showed to form homodimers and heterodimers with both ZmbHLHs and ZmCPP8.
To investigate whether the different TFs act as activators or repressors, trans-activation activity assays were performed in maize mesophyll protoplasts. The expression of GUS reporter gene driven by the ZmPEPC1 promoter fragment was shown to be significantly activated by ZmbHLH90 and repressed by ZmbHLH80, ZmOrphan94 and ZmCPP8. Co-expression of ZmOrphan94 or ZmbHLH80 decrease the activation induced by ZmbHLH90. ZmHB87 showed no trans-activation activity.
To functionally characterize TFs binding to ZmPEPC1 promoter, maize UFMu transposon mutant lines were acquired for ZmbHLH90 and ZmOrphan94 transcription factors. However, in most of the mutant lines there was no correlation between TF and ZmPEPC1 transcript levels with the exception of two silencing lines of ZmOrphan94. Preliminary physiological analysis showed a decrease in the CO2 assimilation rate between WT and ZmOrphan94 mutant lines. RNA samples from those lines were sent to be sequenced in order to determine the DE genes.

Characterization of maize TFs identified as binding to the ZmC4-NADP-ME promoter. The interaction between the ZmbHLH-128 and -129 TFs and N-box cis-elements (5’-CACGCG-3’) present in the ZmC4-NADP-ME promoter was validated and characterized in vitro by Electrophoretic Mobility Shift Assays (EMSA). The ZmbHLH TFs were shown to bind to two N-box motifs located very close to each other in the ZmC4-NADP-ME promoter (only 7 bp apart).
The transcript level of both ZmbHLH genes was analysed in each photosynthetic cell-type (M/BS). Although no striking cell specificity was observed, both TFs showed to be preferentially expressed in the M cells.
In order to investigate whether these maize bHLH TFs interact between them, we carried out direct Y2H and BiFC. Our results showed that ZmbHLH-128 and -129 form homo- and heterodimers.
The transcriptional regulatory activity of the ZmbHLH TFs on the ZmC4-NADP-ME promoter was assessed by trans-activation assays in vivo, in maize M protoplasts, and in planta, in tobacco leaves (in collaboration with JM Hibberd, Cambridge). ZmbHLH-128 showed to be an activator of the ZmC4-NADP-ME gene, while ZmbHLH-129 showed to have no trans activation activity. When both ZmbHLH-128 and -129 were expressed together, the no effect of the ZmbHLH-129 seemed to prevail over the activation activity of the ZmbHLH-128. Given that they form heterodimers, we hypothesize that the dimer ZmbHLH-128/ZmbHLH-129 abolishes the activator activity of ZmbHLH-128/ ZmbHLH-128.
The transcript levels of the two ZmbHLH (128 and 129) genes and ZmC4-NADP-ME target gene were analysed in a day-night cycle experiment. Both TF genes showed similar gene expression profiles, which correlates well with the ZmC4-NADP-ME expression pattern.
To characterize the biological function of ZmbHLH-128 and -129, we have ordered maize UniformMu transposon-tagged mutant lines for each TF. Preliminary molecular analyses as well as crossings between the mutant lines for the ZmbHLH128 and ZmbHLH129 were performed. However, the molecular characterization showed that these mutants did not have the insertion in the expected gene and/or the gene expression was not mis-regulated.
To investigate if the ZmbHLH TFs identified are functional and associated with cell-specific gene expression in rice, the co-transformation of the TFs and their DNA target promoter driving GUS into Kitaake rice has been initiated in collaboration with JM Hibberd (Cambridge).

TASK 3: Identification of histone-modifying enzymes regulating cell-type specific epigenetic profiles (LUH)
The aim of this task was to identify the histone methyltransferase responsible for cell-type specific promoter modifications in maize. Because of the known substrate specificity in Arabidopsis and analyses of expression profiles we identified five candidate maize genes: SDG106, SDG128, SDG108a, SDG115b, and SDG127. SDG127 was not further analysed, because no genetic resources were available. For SDG106, a RNAi-line was available. This line showed a clear suppression of the SDG106 transcript according to our data. Suppression of SDG106 reduced histone methylation on C4-genes and also disturbed C4-gene expression in comparison to non-transgenic siblings. Propagation of seed material for SDG106-RNAi lines was complicated by a strong developmental phenotype. This was also observed for some H3K4 methyltransferase knockout mutants in Arabidopsis. Instead of generating additional RNAi-lines for the other candidates, we used two alternative strategies for identification of the methyltransferase that controls cell-type specific histone modifications on C4-genes:
1. We made use of the newly available Uniform Mu transposon-tagged lines as an alternative source for mutants in methyltransferase genes. We identified mutants for SDG106 (2 lines), SDG128 (1 line), SDG115b (5 lines), and SDG108a (3 lines). All these mutants were analysed for homozygous insertion of the Mu-transposon. One mutant line for each SDG106, SDG128, SDG115a and SDG115b was positively screened for homozygous insertion of the Mu-transposon. But only for the SDG106 UniMu mutant line, a suppression of SDG106 expression could be observed. The SDG106 UniMu- line was used as an independent control for the characterization of SDG106 RNAi-lines.
2. As an alternative strategy, we decided to test whether specific methyltransferases bind to C4-genes. This approach does not interfere with the function of the methyltransferase in development. For this, we generated antibodies to the methyltransferase proteins (SDG106, SDG128, SDG108a, SDG115a and SDG115b) for chromatin immunoprecipitation or FLAG-tagged over-expressors (SDG108a). For SDG106, SDG108a, SDG115a and SDG115b, antigens for immunization were generated. For SDG128, we used a peptide antibody as an alternative, because antigen production was not successful.
The SDG106 UniMu-line showed a reduction of the SDG106 mRNA level to 5% compared to control plants. In these plants, the effects observed in the SDG106 RNAi plants could not be reproduced; the expression and histone methylation levels of analysed C4-genes were not altered relative to control plants. Additionally, the propagation of the SDG106 RNAi-lines did not produce sufficient seeds to analyse the observed effects in more detail in further experiments. The binding analyses of the SDG106 with the generated antibodies did not reveal binding of the SDG106 to C4-genes. The SDG108a UniMu-line did not show reduced expression of the SDG108a gene compared to control plants. The generated SDG108a FLAG-lines were used to analyse the binding of the SDG108a. Unfortunately, none of the analysed lines showed a positive signal in a western blot and no binding of the SDG108a could be detected at C4-genes in ChIP experiments. An additional candidate was identified during detailed analysis of the alternative transcript variant of SDG115a. Using RNA-Seq and ChIP-Seq data from another project, we were able to fix its ambiguous annotation and we identified a “new” SDG (SDG115a) that was wrongly annotated before. PCR experiments verified this new annotation. A homologue of SDG115a was found in Setaria italica. Meta-analysis of C4-ness in grasses provided by the Weber lab in pillar B indicated that the transcript of SDG115a is more abundant in C4 Alloteropsis compared to C3 Alloteropsis. We therefore included SDG115a to our list of candidate genes. The homozygous SDG115a UniMu-line did not show reduced expression compared to control plants. Additionally, no binding could be detected in ChIP analyses using the produced antibody. Because of missing resources further investigation of SDG106, SDG108a and SDG115a and their potential role in regulating C4-photoysynthesis is not possible during timeline of the project. For SDG115b and SDG128 binding analyses showed that both SDGs bind to C4-genes and this binding is much stronger than the binding to the Actin1 housekeeping gene, especially in the coding region. Furthermore, we showed binding of the closest homologues of both SDGs to C4-genes in Setaria italica which indicates a general role of both SDGs in the regulation of C4-genes. Therefore, both methyltransferases are candidates for the regulation of C4-genes (Deliverable 5.2) and were analysed for their genome-wide binding profiles in maize by ChIP-Seq analyses. In summary, our combined approach using mutant resources and in vivo binding studies revealed two candidates for methyltransferases that control C4-genes.

Contribution of LUH to TASK 1: Genome-wide mapping of transcription factor binding sites in closely related pairs of C3 and C4 plants (UDUS, LUH, NEB)
We aimed to analyse the specificity of SDG115b and SDG128 for C4-genes and their contribution to the tissue-specific expression of C4-genes by ChIP-Seq analyses. We started with a comparison of light- and dark-treated maize leaves to establish the method and analyse the C4-specificity. Three replicates for each treatment and SDG were prepared and sent to Atlas Biolabs for sequencing. The raw reads showed a good quality and 28 million reads per sample could be mapped to the maize genome in average. The quantities and qualities of the mappings were stable over all samples. We were not able to perform peak calling in a convincing way. Therefore, we asked Nebion to use their bioinformatic skills to call peaks with an alternative and more complex algorithm. These analyses are ongoing but already showed promising results that need to be optimized further. Preliminary analyses could show that both SDGs bind not only to C4-genes, but to highly transcribed genes in general. Analyses of isolated tissues indicate a preferential binding of SDG115b to C4-genes that are active in mesophyll cells and preferential binding of SDG128 to C4-genes that are active in bundle sheath cells. Thus, binding of both SDGs correlates with active transcription which was additionally shown by a decreased binding in the dark. Beside tissue specificity also high gene expression is a key feature of C4-photosynthesis.Therefore the results show an important role of SDG115b and SDG128 in regulating C4-photosynthesis.

Contribution of LUH to TASK 2: Identification and characterization of cell-type specific promoter-binding proteins (iBET)
The iBET identified transcription factors from the bHLH family by yeast one hybrid analyses that bind to the promoters of C4-PEPC (bHLH90) and C4-ME (bHLH129) in maize. We were in charge to produce antibodies directed against these transcription factors to verify the binding in vivo and evaluate the characteristics of the binding by ChIP analyses. For bHLH90 two strategies were applied to generate specific antibodies: 1. as the antigen, a partial protein was overexpressed in E. coli and sent to BioGenes, a company that produces antibodies. 2. BioGenes synthesized a peptide and used this as the antigen. For the bHLH129 only the first strategy was applied by us. All the antigens were predicted to be accessible within the active proteins and to be as unique as possible for the corresponding transcription factor. The antibodies were tested by western blotting for their specificity. Only one out of six antibodies (2 rabbits per antigen) showed a specific signal at the right size (bHLH90; protein as antigen). Unfortunately, the antibody did not reveal any binding of bHLH90 to PEPC or other prominent C4-genes in ChIP analyses. Potentially, the antibody is not able to bind bHLH90 in its native conformation or the ChIP experiments were carried out in the unsuitable situation, developmental stage or tissue. The iBET is currently testing those hypotheses.

WP 6 Post-translational regulation of major C4 genes in maize and Setaria
Objectives. Identifying post-translational modifications of C4 proteins that are crucial for their function. We will generate catalogues of post-translational modifications for the major C4 proteins. We will identify the functional importance of these modifications and we will test them in transgenic C3 plants. This will guide the design of gene sequences for gene transfer in WP 9. C4 plant biology has suffered from the fact that no easily manageable genetic model system has previously been available. The C4 grass Setaria promises to be such a model system. The Panicoid C4 grass Setaria viridis is a new C4 model that allows a forward-genetic approach for identifying loss-of function mutation in regulatory genes of both C4 anatomy and physiology. S. viridis is of small stature, self-compatible, has a short life cycle and can be transformed by Agrobacterium tumefaciens. The plant is a noxious weed and the wild progenitor of foxtail millet (S. italica). S. viridis has a small genome that is currently being sequenced. Because most of our agronomically important crop species are monocotyledonous plants, monocotyledonous models are required for developing the paradigms for engineering such crops. WP 6 aims at identification of post-translational modifications and testing their importance in transgenic plants. Such plants will be generated in WP9 (Pillar D). Since Setaria is transformable with Agrobacterium and has a much shorter generation cycle than maize, it is the model of choice for the generation of transgenics to test hypotheses developed in WP6. Another practical reason for using the model Setaria is its small genome, which makes the bioinformatics analyses more straightforward because fewer gene duplications are expected. Hence unequivocal identification of target genes is facilitated.

TASK 1: A catalogue of post-translational modifications on major C4 genes (iBET). The aim of this WP was the creation of a catalogue of new post-translational modifications (PTMs) that affect major C4 photosynthesis enzymes (Task 1) and to explore the functional meaning of some of the identified PTMs (Task 2). Because of their central function on C4 photosynthesis in maize, PEPC, PPDK and NADP-ME were the main focus of our work. At the outset of our work, the only PTMs described concerning these key C4 enzymes were the phosphorylations at S15 in PEPC and at T527 in PPDK, which were reported to act as regulators of enzyme activity [Jiao et al. Plant Physiol. 1991, 96:297; Chastain et al. Plant J. 2008, 53:854]. It was our hypothesis that phosphorylation could also be regulating other enzymes in C4 metabolism, and that other phosphorylation events could also occur in PEPC and PPDK. Thus, we first aimed at identifying and pinpointing the phosphorylation events occurring in these enzymes, and secondly, to correlate these modifications with enzyme activity variation during the light/dark cycle.
In order to determine the best protein extraction method for maize proteins and phosphoproteins, we tested five different protein extraction methods. The direct extraction in a urea-based extraction buffer revealed to be the easiest-to-perform method and the one that better ensured extraction of high protein yields with good coverage of the whole proteome and phosphoproteome, and with high reproducibility and little protein degradation.
An initial analysis of the 3rd leaf maize proteome showed the existence of several isoforms of PEPC and PPDK with different isoelectric points that can be due e.g. to phosphorylation and acetylation events. This analysis was extended and resulted in the characterization of the maize proteome from 2DE gel. The 2DE map of the maize 3rd leaf proteome clearly showed the existence and localization of several PEPC, PPDK, NADP-ME and PEPCK isoforms. Staining with Pro-Q Diamond (PQD), specific for phosphorylated proteins, revealed phosphorylation signal in several isoforms of PPDK, PEPC, and PEPCK.
Since we acquired compelling evidences of post-translational regulation in key C4 enzymes, we then extracted plants collected at seven time points (ZT 0.5 2, 4, 15.5 16.5 20, and 23.5) over a 16h/8h (day/night) photoperiod to characterize protein-isoform variation. Protein extracts were separated using 2DE (7cm IPG strips, pH 4-7) and gels were stained for both whole-proteome (Coomassie Brilliant Blue) and for phosphoproteome (Pro-Q Diamond) assessment. Gels were analyzed with Progenesis Samespots software to perform a differential analysis of each protein expression isoforms along the photoperiod. For instance, for PEPC, we could observe a dynamic setup of isoforms with: amount variation of its acidic isoforms during all the photoperiod; appearance of new isoforms upon light, co-localizing with phosphoproteome staining; and several high molecular weight isoforms during dark/light transition (data not shown). As for PEPCK, we observed a strong correlation between the maximum amounts of protein and the strongest phosphorylation signal.
Our first proteomics approach intended to identify new post-translational modifications in C4 enzymes included a 1D-PAGE separation step and the use of phosphopeptide enrichment steps prior to Orbitrap mass spectrometry (ThermoFischer Scientific Orbitrap Elite). Resultant phosphopeptides were scored by the PhosphoRS 3.1 algorithm. This tool enables automated and confident localization of phosphorylation sites within validated peptide sequences. One important finding was the phosphorylation on NADP-ME (S419) that we were able to identify. No PTM was reported for this enzyme so far. Moreover, we were able to identify 6 putative phosphorylation sites for PEPC and another 7 for PPDK.
We then focused our attention in validating this phosphorylation on NADP-ME. We made use of a TripleTOF 6600 mass spectrometer to analyze a TiO2 phosphopeptide enriched sample to successfully validate this post-translational modification on NADP-ME.
Aiming at enlarging the catalogue of PTMs, we engaged an alternative MS-based strategy termed SWATH-MS, for which we developed a new proteomics workflow. This innovative strategy was designed to enable the identification of new putative PTMs and, importantly, to relatively quantify the amount of each phosphopeptide across the time points under analysis. The validation of our previous results was also expected.
We used a Triple-TOF 5600 mass spectrometer in both IDA and SWATH acquisition modes to analyze samples collected at time points ZT4 (4h after light), ZT15.5 (30min before dark) and ZT23.5 (30min before light). This analysis provided high protein coverage, which is crucial in allowing the detection of new phosphorylation sites. We were able to cover 94.3% of PEPC, 85.8% of PPDK, 92.7% of NADP-ME and 76.0% of PEPCK sequences. This resulted in the identification of new putative phosphorylations and also in the validation of a few previously identified phosphorylations.
SWATH analysis allowed following the variation of several phosphopeptide amounts on PEPC, PPDK, and PEPCK, by performing relative quantification at ZT4, ZT15.5 and ZT23.5. As for PEPC, we could see that the phosphorylation on S15, a well-studied phosphorylation site with consequences for PEPC activity, follows the expected profile, with higher phosphorylation rate at ZT4 and lower during the night ZT23.5. Furthermore, we could detect another phosphorylation site that follows the same trend and three more that have little contribution for protein overall phosphorylation and do not respond to changes in light conditions. Overall, the quantification of phosphorylation shows a tendency which is consistent with the macroscopic phosphorylation profile observed in 2DE.
For PPDK, the newfound phosphorylations organize in clusters in the protein. Consistently, we found phosphorylation sites that do not show light-dependency and are significantly phosphorylated at all times analyzed. Nevertheless, the expected light-dependent regulation of T527 was observed and it is followed by other residues in its phosphorylation cluster (S575). In conclusion, the number of phosphorylation sites detected agrees with the ‘constitutive’ high level of PPDK phosphorylation, but the expected light-dependent phosphorylation profile is also present, suggesting other functions for this PTM.
For PEPCK, only one of the phosphopeptides could be quantified (T120). Its variation is light-dependent and varies according to the 2DE observed signal (data not shown).
In an attempt to go beyond the identification of phosphorylations and enable the identification of other PTMs such as acetylations and ubiquitinations in C4 enzymes, we performed LC-MS/MS experiments using an AB Sciex 6600 TripleTOF at iBET. We also extended the number of time points collected along the photoperiod (ZT0.5 ZT2, ZT4, ZT15.5 ZT16.5 ZT20 and ZT23.5). Over the seven time points under analysis, we were able to identify 4 putative ubiquitination sites for PEPC and PPDK and 1 putative ubiquitination site for NADP-ME. Concerning the identification of acetylations in these important enzymes, we could identify 34 putative acetylation sites for PEPC, 27 for PPDK and 5 for NADP-ME.
Over the course of this project, we were able to generate novel data that improved the knowledge on the post-translational regulation of key C4 enzymes and we will be submitting a few manuscripts to disseminate our achievements among the scientific community. Our contribution to the project have shown that proteomics and mass spectrometry toolboxes can really add essential and complementary information to what transcriptomics and metabolomics can provide and that there is still a lot more to be done to achieve complete knowledge on the PTM regulation of key C4 enzymes and, more important, to fully understand how PTMs regulate enzyme activity.

TASK 2: Significance of newly identified modifications (UDUS). Analysing the posttranslational regulation of maize C4 NADP-ME. Using nano LC-Orbital MS a phosphopeptide from the maize C4 NADP-ME (uniprot P16243) was identified. The putative residue subjected to phosphorylation is the S419.

(i) Assessment of the biochemical significance of the PTM. To mimic the constitutive phosphorylation of the S419, we have created a mutated form of the enzyme which has a glutamic acid residue instead of serine at this position (S419E) by using site-directed mutagenesis. Additionaly, a negative control was generated – we created a mutated maize C4 NADP-ME having an alanine residue instead of serine (S419A).
The wild-type enzyme and its mutated versions were heterologously expressed in E. coli and isolated from the cellular protein extracts by affinity chromatography to high purity. Subsequently, a series of kinetic measurements were performed in order to compare the main biochemical characteristics of the wild-type and mutated proteins.
The phosphomimicking protein (S419E) presented significant changes in its enzymatic properties. The optimum pH of the wild-type enzyme is 8.0 while that of the mutated protein is displaced to a value of 6.5. At pH 8.0 the pH found in the chloroplast’s stroma during the day, the mutated enzyme presented less than 10% of the wild-type activity. Furthermore, at this pH value, the mutated enzyme had much lower turn-over rate (Kcat) and affinity (higher Km) for malate than the wild-type enzyme. At pH 7, the mutated enzyme also presented less activity than the wild-type at low malate concentrations. In contrast to the wild-type, the mutated enzyme was not inhibited by high malate concentrations at pH 7. These results indicate that the phosphorylation of the S419 decrease the enzymatic activity of the maize photosynthetic NADP-ME.
The S419A mutein was generated in order to provide a negative control. The pH optimum of S419A is similar to the wild-type enzyme. The biochemical characterization of this mutated protein indicated that at the pH found in the chloroplast’s stroma this enzyme is highly active.
As S419E differs from the wild-type enzyme in terms of its biochemical characteristics, we analysed whether these differences could result from changes in the structural organization of the analysed proteins. We performed circular dichroism (CD) in order to analyse the protein’s secondary structure and sedimentation velocity centrifugation by using the analytical ultracentrifuge in order to determine the possible changes in quaternary organisation. For both sets of experiments, the proteins were analysed at pH 7.0 and 8.0. These two pH values were chosen in order to mimic the day-night change of the chloroplast stroma pH. Results of the CD-measurements have shown no differences in terms of the secondary structure between wild-type and mutated enzymes at both tested pH values. Results of the analytical ultracentrifugation have also shown no differences between all three analysed enzymes: at pH value of 8.0 they had sedimentation coefficient of approximately 10 which corresponds to the homotetrameric organization of the proteins; at pH value of 7.0 an additional peak with sedimentation coefficient of approximately 5 appeared, indicating that at this pH value some part of enzymes is additionally assembled as homodimers. Taking together, these results indicate that differences in the biochemical properties of mutated enzymes are not caused by the differences in their structural organization comparing to the wild-type NADP-ME.

(ii) Assessment of the in vivo significance of the PTM. In silico modelling of the NADP-ME variants based on the crystallographic data for the pigeon liver enzyme indicated that the S419 is spatially available for a kinase. The fact that phosphorylation of maize C4NADP-ME in S419 diminished the activity of the enzyme drastically, and this mostly happens during the first hours in light points to a regulatory effect to rapidly fine tune the activity when a full active enzyme is not necessary during the day. During the night the enzyme is mainly regulated through pH and malate inhibition.

Phosphoproteomics (USFD). Setaria viridis (Setaria) plants were grown at 28/26 °C in light/dark, in ambient CO2, maintaining 1000 μmol m-2 s-1. Leaves were collected from 7.5h darkened and 4h/15.5h illuminated plants. Whole leaf protein was extracted and resolved by SDS-PAGE, prior to tryptic digestion. Analyses were performed by biOMICS (USFD) using an uHPLC-MS/MS OrbiTrap Elite. The resulting peptides were searched against S. italica protein sequences using SEQUEST and phosphorylations were determined using phosphoRS. We identified 16 putative phosphorylation sites in alanine aminotransferase, 12 sites in aspartate aminotransferase (AspAT), 5 sites in carbonic anhydrase, 13 sites in MDH, 21 sites in NADP-ME, 7 sites in PEPC including the identified phosphorylation at Ser-15 in Zea mays (maize), and 8 sites in PPDK, including the phosphorylation identified at Thr-527 in Z. mays. PEPCK was not identified in Setaria. Further phosphorylation site assignment was carried out in Sorghum bicolor (Sorghum) and Megathyrsus maximus (guinea grass). Two putative phosphorylation sites were identified for Sorghum PEPC and 4 sites in guinea grass PEPCK. In rice, we identified 4 putative phosphorylation sites in PEPC and 4 sites in NADP-ME.

Antibodies (USFD). Work in USFD has also involved the production of suitable polyclonal (against wheat Rubisco) and monoclonal antibodies (so far against PEPC, NADP-ME). The PEPC monoclonal antibody works very well for all species tested including maize, Setaria and rice and at a high dilution. The NADP-ME antibody, though not fully tested, but works well for maize and Setaria, but not rice.

Biochemistry (USFD). The dark/light regulation of key enzymes of the C4 pathway was examined by comparing their in vitro activities using biophysical methods. We optimised the conditions for the flash extraction of leaf proteins from the mid-sections of darkened and illuminated leaves, and developed a highly reproducible microplate-based method for assaying enzyme activities in smaller reaction volumes, but involving normalised concentrations of crudes extracts across a larger number of technical and biological replicates. Testing the design of this approach, we confirmed previous observations that dark/light activity differences were present in PEPC and NADP-ME, but not found in PEPCK. The study then focused on NADP-ME and AspAT, enzymes which act on C4-acids (malate and aspartate) translocated from mesophyll to bundle sheath cells and contribute to the CO2 pool. These enzymes were extracted from three C4 monocot plants that use NADP-ME as the primary decarboxylase, which according to a previous transcriptomic analyses, the gene encoding for PEPCK is expressed at a very low level in Setaria, higher in Sorghum, and significantly more in maize. By assaying the activities of NADP-ME from dark and light periods over a pH range, we reported a shift in the pH optima in Sorghum from pH 8.2 in the dark to pH 7.4 in the light, whereas the activities of the same enzyme from maize and Setaria both peaked at pH 8.2–8.4 and remained so through dark/light transitions. We found dark-to-light increases in the rate of reaction at the pH optima in all species. By investigating the Michaelis–Menten kinetics of enzymes at their optimal pH, we found that 1) NADP-ME affinity range for malate was very high in maize, moderate in Sorghum, and much poorer in Setaria, and 2) there were higher affinities for malate in the light than in the dark in Setaria and Sorghum but not in maize (which maintained a higher substrate affinity across all time points). At the optimal pH, the presence of a reducing agent (DTT) further enhanced substrate affinity in Setaria and Sorghum, and generally had the opposite effect at sub-optimal pH. NADP-ME in Sorghum maintained a high level of activity over a broader range of Mg2+ at the light pH optimum, but was more sensitive to higher Mg2+ concentration at the dark pH optimum. In contrast, AspAT’s affinity for both aspartate (Asp) and 2-oxoglutarate (2OG) decreased in the dark-to-light transition in the three C4 species. The dark/light affinity differences for both substrates were most significant in maize, followed by Sorghum, and was the smallest in Setaria.

WP 7 Development of a generic model of C4 photosynthesis
Objectives. We will develop and parameterize a generic C4 photosynthesis model based on large-scale determination of physiological data under different conditions. We will make use of this model to guide C4 engineering. We will develop a kinetic systems model of Kranz type C4 photosynthesis including a detailed description of photosynthetic carbon metabolism, light reactions, and the major diffusion process of CO2 from the atmosphere to the site of CO2 fixation. The model will be developed for maize. The model will first be used to identify the key features required to gain the high photosynthetic efficiency in maize, including the optimal distribution of enzymes into different enzymes, the optimal rate of fluxes for the metabolite transporters, the optimal distribution of enzymes involved in the Calvin cycle between bundle sheath and mesophyll cells. The model can help to identify bottlenecks of C4 photosynthesis and guide the design of transgenic lines in WP9 as well as the selection of putative regulatory factors in WP5.

TASK 1: Kinetic systems model (SIBS)
TASK 3: Systems model of C4 photosynthesis and C4 engineering (SIBS) (in silico modelling funded from external resources, using the experimental data derived from Task 2)

C4 plants usually has higher photosynthetic CO2 uptake rate and higher light use efficiency compared with C3 species. Biochemical factors that affect this efficiency of C4 plant have been systematically evaluated. While anatomical and biochemical features that might influence mesophyll conductance and bundle-sheath conductance in C4 species have not been systematically analyzed. A 3D two cell maize reaction diffusion model was developed to explore this question. The structure of the reaction diffusion model includes a mesophyll cell connected with a bundle-sheath cell, including chloroplast, mitochondria, vacuole and cytosol, including CO2, bicarbonate and related enzymes (carbonic anhydrase (CA) and Rubisco) and also C4 related metabolites: oxaloacetic acid (OAA), malate, pyruvate, phosphoenolpyruvate (PEP) and the C4 related enzymes: PEP carboxylase (PEPC), Malate dehydrogenase (NADP-MDH), NADP-malic enzyme (NADP-ME), pyruvate and phosphate dikinase (PPDK). CO2, bicarbonate and C4 acids can diffuse and take reactions throughout the model structure. The model was parameterized with parameters from maize.
The model effectively simulated CO2 uptake rates, C4 cycle rate and metabolism response for different intercellular CO2 concentration and light intensities. Simulations showed that mesophyll conductance decreased with increasing intercellular CO2 concentration and reached stable when Ci was higher than about 30 Pa. Mesophyll conductance response curves with light intensity showed an increasing pattern when Ci equalled 15Pa. Our predicted data are broadly similar to measured data. Leakage increased and the bundle-sheath conductance decreased with elevated Ci. CO2 concentration in bundle-sheath also increased with rising Ci while the leakiness showed first decrease when Ci is low and then increase and then gradually decrease again at high Ci. The pattern and values of leakiness, leakage and bundle-sheath conductance response curves with intercellular CO2 partial pressure were broadly consistent with measurement data and C4 biochemical model. Bundle-sheath conductance, leakage and leakiness response curves with different light intensities showed complex pattern. Although the curves’ pattern is different from simulated results from C4 biochemical model, it is broadly consistent with measurement data. This model provides an effective framework to explore limiting anatomical and biochemical factors in mesophyll and bundle-sheath conductance in C4 species.

TASK 2: Model parameterization and validation (MPG). Historically C4 plants have been classified into three biochemical subtypes based on the main enzyme involved in 4-carbon metabolite decarboxylation: NADP-malic enzyme (NADP-ME), NAD-malic enzyme (NAD-ME) or phosphoenolpyruvate carboxykinase (PEPCK). However, this is now recognized to be an over-simplification as more than one decarboxylation pathway can operate in parallel within the same leaf.
Fully-expanded 4th leaves of Zea mays B73, which is classically considered to be an NADP-ME subtype species, were fed with 13CO2 for 0, 10, 15, 20, 30, 50 s and 1, 3, 5, 10, 20, 40, and 60 min in a custom-made labelling chamber that allowed rapid quenching of metabolism in the prevailing light and CO2 regime, by dropping a copper rod, pre-cooled to liquid N2 temperature, onto the leaf. Chase samples were also taken after 60 min of 13CO2 feeding followed by 5 and 20 min of non-labelled air treatment. The temporal labelling kinetics of 35 metabolites including proposed intercellular transport metabolites such as malate, aspartate, pyruvate, alanine, PEP, 3PGA and triose-phosphate, and intermediates in the Calvin-Benson cycle (CBC) and tricarboxylic acid (TCA) cycle, sucrose and starch synthesis, photorespiration pathway and amino acid metabolism were assessed by LC-MS/MS and GC-MS.
Additionally, we obtained fractions enriched in mesophyll cells (MC) and bundle sheath cells (BSC) from material labelled with 13CO2 for 3 min, allowing us to determine intercellular distributions and concentration gradients of the metabolic pools that are involved in C4 metabolism.
With these analyses we were able to define the sizes of the metabolic pools actually involved in C4 photosynthesis. For example, ~60% of the total malate pool was still unlabeled after 60 min 13CO2 treatment, indicating the presence of a large inactive pool which is probably located in the vacuole. We also confirmed that there is a concentration gradient of the active malate pool that can drive diffusion from the MC to the BSC for decarboxylation by NADP-ME. However, the concentration gradient of pyruvate, the product of malate decarboxylation in a NADP-ME subtype, cannot drive its transport from the BSC to the MC. These analyses also revealed intercellular concentration gradients of aspartate, alanine and phosphenolpyruvate that could drive the metabolite transport associated with a PEPCK subtype shuttle. Pyruvate might either need to be transported actively to the MC or transported as alanine which would also conserve the amino homeostasis between the cells. We could show that aspartate carries 10-14% of the carbon fixed in MC into the BSC in maize. There is also rapid carbon exchange between the CBC and the CO2 concentrating shuttles, probably via the transport of 3PGA/triose phosphate between BSC and MC, and this exchange is equivalent to about 10% of carbon gain. We postulate that the presence of multiple shuttles, alongside carbon transfer between them and the CBC, confers great flexibility in C4 photosynthesis, allowing maize to adjust its photosynthetic metabolism in response to changes in light, temperature, nitrogen status or other environmental factors.

WP8: Understanding Model Systems
Objectives. Assessing the impact of a photorespiratory by-pass in C3 rapeseed. We will transfer the photorespiratory by-pass established in one of the applicants' labs to the important European crop rapeseed. Previous work suggests that over-expression of glycollate dehydrogenase, the first enzyme of the synthetic pathway, is sufficient to enhance growth and reduce photorespiratory losses. Plant chloroplasts are capable of further converting glyoxylate, the product of the initial reaction (Kebeish et al., 2007 and Applicant LUH, unpublished data). Lines of rapeseed containing the synthetic photorespiratory by-pass will be analysed.
Impact of a natural photorespiratory by-pass in Brassicaceae (Moricandia). A natural photorespiratory by-pass exists in a number of so-called C3-C4 intermediate plants. In its simplest form this involves the loss of one or more glycine decarboxylase subunits from the mesophyll so that a photorespiratory glycine shuttle operates from the mesophyll cells to a bundle sheath. In these plants the bundle sheath is both enlarged and has an increased complement of organelles, particularly of the mitochondria where glycine decarboxylation occurs (Hylton et al 1988). By operating this shuttle these intermediates are able to recycle a portion of photorespiratory CO2, thereby lowering the CO2 compensation point (von Caemmerer, 1989). This shift in glycine decarboxylation to the bundle sheath is a feature of all C4 plants studied and has been postulated to be the first step in the evolution of the C4 pathway (Rawsthorne and Bauwe, 1988). Accordingly it is currently being manipulated in the C4 rice programme and in WP9 as one of the essential elements of the C4 pathway. Understanding more about the development of this natural photorespiratory by-pass is also a key element in Pillar I (Initiating and Maintaining the Bundle Sheath).
To understand the impact of shifting the location of glycine decarboxylation on photosynthetic and non-photosynthetic processes and metabolism, plants of the genus Moricandia will be analysed in detail. Moricandia is a member of the Brassicaceae and therefore closely related to rape, which is a European crop. M. moricandioides, four accessions of M. arvensis, M. suffruticosa, and M. nitens are currently under study in Düsseldorf. Seeds of these lines will be distributed to the partners. Moricandia arvensis (C3-C4) and M. moricandioides (C3) are routinely transformable (Thole and Rawsthorne, 2003) and could be complemented in Düsseldorf (i.e. GDC re-expressed in the mesophyll). Gas-exchange was measured on Moricandia leaves in the 1980s, however, now gas-exchange measurements in an eco-physiological context can be interwoven with transcriptome analyses. To address how each line partitions available carbon between organs we will investigate growth, water use and analysis of stomatal function, productivity and partitioning of dry matter.
Stomatal tuning in C3/C4 metabolism backgrounds using Cleome. Another advantage of C4 photosynthesis is water economy. Under both optimal and suboptimal conditions, as a consequence of efficient carbon concentrating mechanisms, C4 photosynthesis commonly operates with lower stomatal conductance than C3 photosynthesis (Osborne & Sack, 2012). Moreover, stomata response occurs at different CO2 concentration in the mesophyll for C3 and C4 plants (Ci/Ca ratio is markedly lower in C4 than in C3 leaves). This implies that introducing C4 traits into C3 plants has to consider not only C4 key concentrating mechanisms but also the molecular and development basis of hydraulic control which includes guard cells behaviour and signalling in C3/C4 metabolism backgrounds. With the aim to unravel candidate genes controlling C4 stomatal response traits, we will investigate the extent to which gene expression differs in guard cells of C4 versus C3 plants in parallel with an analysis of stomatal function on two closely related species: Tarenaya hassleriana and Cleome gynandra that show C3 and C4 metabolism, respectively.

Overall changes to the Programme of Work for WP8
WP8 was revised in 2014 (see revised WP8 provided in the last report). Although the rape lines with a photorespiratory by-pass are now available (after some initial problems with their regeneration in tissue culture), the evidence so far (principally from Bayer, but also from the Fraunhofer Institute) suggests that plants transformed with the by-pass do not show a sufficiently strong phenotype to warrant an extensive programme of work such as originally proposed in WP8. We did not want to be engaging four research groups (and especially PhD students) studying plants in which the phenotype is very marginal and the results are likely to be so subtle that they are difficult to detect. We discussed this in Lisbon in January 2014 and agreed (i) that further analysis of the rape transformants was a high risk option and (ii) that we should refocus the work package to Moricandia (and some stomatal work on Cleome). In addition, the fact that the Gates-funded RIPE programme in Champaign intended to work on plants with the photorespiratory by-pass would introduce an unnecessary duplication of effort. We will maintain an interest in the rape transformants and we would still propose to analyse them if initial screening at LUH shows anything of interest, but they should not be a main focus of WP8. The revised WP8 (see Appendix) involves much the same resources with the same participants and approaches. We consider that is a very low risk approach that can be fully achieved within the project timeframe. In fact we were at the point in January 2014 at which we could easily decide to switch to Moricandia from rape without any significant loss of resources (apart from having made the rape transformants).
The reasoning behind the switch to Moricandia was this that a natural photorespiratory by-pass that exists in a number of so-called C3-C4 intermediate plants. Moricandia is a brassica and is therefore closely related to rape, which is an important European crop. In its simplest form the photorespiratory by-pass in Moricandia involves the loss of one or more glycine decarboxylase subunits from the mesophyll so that a photorespiratory glycine shuttle operates from the mesophyll cells to a bundle sheath. In these plants the bundle sheath is both enlarged and has an increased complement of organelles, particularly of the mitochondria that catalyse glycine decarboxylation. By operating this shuttle these intermediates are able to recycle photorespiratory CO2, thereby lowering the CO2 compensation point. This shift in glycine decarboxylation to the bundle sheath is a feature of all C4 plants studied and has been postulated to be the first step in the evolution of the C4 pathway. Accordingly it is a step currently being manipulated in the C4 rice programme and in WP9 as one of the essential elements of the C4 pathway. Utilising Moricandia maintains the same theme as the photorespiratory by-pass in rape. Understanding more about the development of this natural photorespiratory by-pass is also a key element in Pillar I (Initiating and Maintaining the Bundle Sheath).
Although a good deal of gas-exchange was done on Moricandia in the 1980s, we now have access to modern approaches, particularly using sequencing. Further sequencing studies could be undertaken on Moricandia in the 3to4 project (these would be low risk). M. moricandioides, four accessions of M. arvensis, M.suffruticosa and M. nitens are under study in Düsseldorf and plant phenotyping and CO2 compensation point comparisons have been done. Seeds of these plants are available from Düsseldorf and have been distributed. Moricandia arvensis (C3-C4) and M. moricandioides (C3) are routinely transformable and could be complemented in Düsseldorf (i.e. GDC re-expressed in the mesophyll). However, we decided that we should not rely wholly on transgenics as these might not be available until 2015 and there is always a risk that this option fails. If available, however, they will be utilised.
In addition to Moricandia, we also proposed for WP8 using another two closely related brassica model species to study stomatal function: Tareneya (Cleome) hassleriana and C. gynandra show C3 and C4 metabolism, respectively. The aim here is to unravel candidate genes controlling C4 stomatal response traits by investigating the extent to which gene expression differs in guard cells of C4 versus C3 plants in parallel with an analysis of stomatal function. This again is low risk.

1- Assessing the impact of a photorespiratory by-pass in C3 rapeseed

TASK 1: Construction of rapeseed with the synthetic by-pass (Subcontracted, CAM and LUH)
The rapeseed constructs were generated and the production of transgenics outsourced to NIAB as planned in the proposal. NAIB generated the plants, actually twice, as the construct design was changed, however, after getting the plants made, it became clear that the constructs of synthetic by-bass related to glycollate dehydrogenase were not very successful and did not yield any stable phenotypes in C3 rapeseed. Therefore, a new venue of work was decided (see hereafter).

TASK 2: Preliminary analysis of phenotype of plants with photorespiratory by-pass (LUH)
Research suggested that phenotypes may be too subtle for further analysis, therefore, any further work on rapeseed had to be terminated. WP8 was revised to include work related to other model systems such as Moricandia and Cleome. A natural photorespiratory by-pass exists in a number of so-called C3-C4 intermediate or C2 plants such as Moricandia. The aim was to understand the mechanisms underlying the natural photorespiratory by-pass in Moricandia. All the results of the work related to these systems are detailed below.

2- Assessing carbon gain and water use efficiencies in C4 model systems (Moricandia and Cleome)

TASK 3: Assess mesophyll and diffusive limitations in Cleome (CEA/CNRS)
Mesophyll conductance to CO2 is an important factor limiting photosynthesis in C3 plants but its impact on photosynthetic rate in C4 plants is poorly understood. By using functional approaches complemented with modelling, we addressed the quantitative role of mesophyll conductance and carbonic anhydrase for C4 versus C3 photosynthesis on two closely related species: T. hassleriana and G. gynandra that show C3 and C4 metabolism, respectively. Using 18O-labelled CO2 and mass spectrometry for monitoring the rate of oxygen atoms exchange between CO2 and mesophyll water, the kinetic properties of CO2 diffusion and hydration into the leaf in vivo were investigated. Our results showed that mesophyll conductance to carbonic anhydrase sites (gmCA) was similar in the C4 G. gynandra and in the C3 T. hassleriana (gmCA = 0.5 mol m-2s-1) and markedly higher than the total mesophyll conductance (Rubisco-based) measured in the C3 T. hassleriana. By analysing the relationships between assimilation, mesophyll and stomatal conductances in the C4 G. gynandra and in the C3 T. hassleriana, we showed that mesophyll conductance has much less contribution in limiting C4 photosynthesis than C3 photosynthesis.

TASK 4: Assess stomatal response and guard cell transcriptome (CEA/CNRS, CAM)
In collaboration with CAM, we investigated the extent to which gene expression differs in guard cells of C4 versus C3 plants in two closely related species: T. hassleriana and G. gynandra that show C3 and C4 metabolism, respectively. Using gas exchange techniques, we verified that G. gynandra showed lower stomatal conductance as well as lower Ci/Ca ratio than T. hassleriana (gs = 0.18 ± 0.02 versus 0.56 ± 0.12 mol m-2s-1 and Ci/Ca = 0.40 ± 0.05 versus 0.80 ± 0.02 respectively) as commonly reported in C4 versus C3 plants. Using Laser Capture Microdissection (Arcturus XT, Life Technologies), we were able to harvest high quality RNA from guard cells (GC) and mesophyll cells from the two species and subject them to mRNA amplification. Sequencing was done at TGAC (Norwich) and data analysis of RNAseq was completed at the end of 2014. We quantified the extent to which expression of genes encoding enzymes involved in primary metabolism are re-wired in a C4 leaf and a C4 mesophyll-like blueprint was identified in GC of the C4 species. Thus, genes associated with C4 mesophyll photosynthesis were more highly expressed in guard cells of C4 compared with C3 leaves. In addition, major variations in transcripts associated with carbon-related pathways known to influence stomatal behaviour were detected. In contrast, most of ions and CO2 signalling modules appeared unchanged at the transcriptional level in GC in both C3 and C4 species. Furthermore, we detected two major patterns of cell-specific C4 gene expression within the C4 leaf. In the first, genes previously associated with preferential expression in the bundle sheath showed continually decreasing expression from bundle sheath to mesophyll to guard cell. In the second, expression was maximal in the mesophyll compared with both guard cells and the bundle sheath. These data imply at least two gene-regulatory-networks act to coordinate gene expression across the bundle sheath, mesophyll and guard cells in the C4 leaf.

TASK 5: Assess mesophyll and diffusive limitations in Moricandia (CNR)
The National Research Council of Italy (CNR) investigated the effect of drought stress on the photosynthetic and gas exchange characteristics of Moricandia species with C3 and C2 photosynthetic physiologies. Moricandia moricandioides (C3 photosynthesis), Moricandia arvensis (C2 photosynthesis) and Moricandia suffruticosa (C2 photosynthesis) were grown under controlled conditions (30˚C, 400 μmol m-2 s-1 PAR) before experiencing prolonged water-deficit. The fraction of transpirable soil water technique was used to gauge the impact of drought on the respective species. Key photosynthetic parameters were analysed at intervals in the drought response. These analyses are photosynthesis, photorespiration, light/dark respiration, performance of photosystem II, stomatal and mesophyll diffusional limitations to photosynthesis. These gas exchange measurements have been used to model mesophyll and bundle sheath conductance to CO2 and photorespiration rates using an excel based iterative model that we hope will be published alongside this data. This was completed within the 60-month milestone. At ITQB-NOVA, samples from Moricandia moricandioides (C3), M. arvensis (C2), M. suffruticosa (C2) grown in controlled environment conditions experiencing two distinct level of water availability (well-watered and drought) and under recovery were used for RNA and carbohydrate extraction. RNA samples were sent to HHU, where they will be processed for RNAseq (still pending). Carbohydrate analyses are on-going.

TASK 6: Assessing rates of photosynthesis (EMU)
Structural control over photosynthesis and photorespiratory bypass has been poorly explored in C2 and C4 plants. However, this information is highly relevant to understand the differences between C2, C3 and C4 photosynthesis, to build fully functional C4 crops and to better understand the effect of dense plant stands on future C4 crop production.
Two subtasks were fulfilled as parts of Task 6. Firstly, we explored the structure-function relationships in developing as well as fully expanded Moricandia arvensis and M. moricandioides leaves grown at different CO2 concentrations and light availabilities in order to understand how and to what extent is the C2 photosynthesis controlled by structure through the course of leaf development. Net photosynthesis rate, An, photorespiration, Rd and CO2 compensation concentration, varied widely across the treatments and this variation in physiology was strongly coupled with variation in the underlying structural and ultrastructural correlates. Light availability had a major effect on both M. arvensis and M. moricandoides, whereas the increased CO2 concentration influenced the structural correlates of M. moricandoides more. Furthermore, increased CO2 had a minor effect on M. arvensis regarding the gas-exchange parameters, while in M. moricandioides it led to a 20% increase in An, but a decreased Rd and CO2 compensation concentration. Additionally, the underlying structural and ultrastructural correlates were determined in detail.
The second subtask was to evaluate whether C4 crops have the same potential for physio-morphological shade acclimation as C3 crops. Four C4 crop species: Sorghum bicolor, Zea mays, Mischantus sinensis, Pennisetum purpureum, and a C3 crop, Triticum aestivum, were grown at three different light intensities (1200, 500 and 100 µmol m-2 s-1) for this purpose. An, CO2 compensation concentration and Rd were all affected by light availability. The most important structural and ultrastructural features underlying the modified An, Rd and were chloroplast size and individual bundle sheath cell size as well as the area of the whole Kranz complex. Contrary to the earlier reports, the results reported here suggest that C4 crops are phenotypically as plastic as the C3 T. aestivum.
Overall, we demonstrate that C2 M. arvensis is physiologically more efficient than C3 M. moricandioides. Additionally, the physiological and structural acclimations are strongly coupled in C2 Moricandia species. Furthermore, the impact of phenotypic acclimation potential of C4 plants has been underestimated as a determinant of photosynthetic rates. However, we have determined the structural correlates related to C4 shade acclimation capacity. Based on the results of Task 6, we suggest that the structural control of photosynthesis should be considered in future projects aiming to engineer the C4 pathway into C3 crops.
During the project period we 1) finalized the physiological and morphological characterization of M. moricandioides (C3) and M. arvensis (C2). Additionally, we tested the plasticity of both species by looking physio-morphological responses in response to light intensity and CO2 concentration through the course of leaf development; 2) evaluated whether C4 crops have the same potential for physio-morphological shade acclimation as C3 crops.
We completed A-Ci curves in 2% and 21% O2 and calculated CO2 compensation concentrations, net photosynthesis rates and dark respiration rates. To elucidate the structural mechanisms underlying C2 photosynthesis and photorespiration, the effect of light and CO2 concentration to a) ratio of bundle-sheath cell area to mesophyll cell area, b) leaf porosity and c) the ratio of the amount of chloroplast to amount of mitochondrion in the bundle sheath and mesophyll cells d) cell wall thickness of bundle sheath cells and mesophyll cells e) variation in mesophyll cell surface area exposed to the intercellular airspaces, f) chloroplast size in mesophyll and bundle sheath cells was studied in M. arvensis and M. moricandioides leaves grown at four different conditions: normal light (1000 µmol m-2 s-1) and ambient CO2 (400 ppm); low light (500 µmol m-2 s-1) and ambient CO2; low light and elevated CO2 (800 ppm), normal light and elevated CO2. Physiological characteristics were analysed in fully expanded leaves while structural analysis were performed during the course of leaf development (from recently unfolded –LPI1- to fully expanded stage –LPI 5-). Leaf plastochron index (LPI) was used to define the physiological age of leaves. Net assimilation rates were higher in M. arvensis leaves (15-22 µmol m-2 s-1 vs 7-15 µmol m-2 s-1). An was reduced in low light leaves in M. arvensis and M. moricandioides while CO2 didn´t have effect. CO2 compensation concentrations () were higher in M. morcandioides leaves in 2 and 21% of oxygen. In 21% of oxygen was around 10 µmol mol-1 (O2:21 %) and 1µmol mol-1 (O2: 2 %). In M. moricandioides CO2 compensation concentrations were highest in low light leaves but CO2 concentration had minor effect on M. arvensis. The ratio of the nr of mitochondria to the nr of chloroplasts per unit of leaf area was highest in M. arvensis bundle sheath cells. This ratio was affected by growth conditions and it was highest in low light leaves. While the numbr didn´t vary in M. arvensis leaves. Mesophyll and bundle sheath cell wall thickness didn´t vary in M. moricandioides leaves while in M. arvensis leaves bundle sheath cell walls were thicker than mesophyll cell walls. Bundle sheath cell walls were the thickest in high light leaves. The biggest chloroplasts were found in bundle sheath cells of M. arvensis. Chloroplasts were also bigger in M. arvensis mesophyll cells than in M. moricandoides mesophyll cells. Chloroplast area did not differ between mesophyll and bundle sheath cells of M. moricandioides. Growth conditions affected chloroplast area in M. arvensis but not in M. moricandioides leaves. Most importantly these results suggest that acclimation potential of C2 photosynthesis to CO2 and light availability is importantly controlled by structural and ultrastructural correlates.
In addition to Moricandia experiment, additional gas-exchange and structural measurements were done in C4 crops: Z. mays, S. bicolor, P. purpureum, M. sinensis and C3 reference species T. aestivum. The measurement routine contained assimilation, transpiration and fluorescence lines for CO2 concentration curves (under both ambient (21%) and low (2%) O2) and light curves (21% and 2% oxygen). Both leaf anatomy and physiology varied across light gradient. Net assimilation rates, CO2 compensation points and BC and MC anatomy differed across C4 species and was related to variation in net assimilation rate, dark respiration CO2 compensation point. Net assimilation rate, dark respiration and CO2 compensation concentration were strongly controlled by mesophyll structural and ultrastructural features, such as size of the bundle sheath cell and size of the chloroplasts in BC. Most importantly these results suggest that C4 crops are structurally as plastic as C3 crops.

TASK 7: Assessing growth, productivity, carbon partitioning and water use (KSU)
Gas exchange, growth rates, productivity and water use efficiency (WUE) were studied in closely related T. hassleriana (C3) and G. gynandra (C4) both from Cleomaceae and in M. moricandioides (C3), M. arvensis (C3-C4 intermediate or C2) and Brassica oleracea (C3) all from Brassicaceae under well-watered and water-stressed conditions. Net assimilation rate (A) versus photosynthetic active radiation (PAR), A/PAR, A versus intercellular CO2 concentration (Ci), A/Ci, and A versus temperature (T), A/T, curves were constructed for all species and with mitochondrial respiration rates, light compensation points, photorespiration rates, Vo/Vc ratios, chloroplastic CO2 concentrations and mesophyll conductance values in C3 and C2 intermediate species were modeled.

Experimental Design. Plants were grown in 10L pots at KSU. All plants were irrigated twice a week and fertilized once a week with 1X Hoagland’s solution. After 4 weeks of growth, half of the plants were well watered and the other half was water-stressed. Irrigation frequency was reduced gradually until leaf water potential of plants fell below -1.5 MPa. Gas exchange measurements were conducted on randomly selected plants from each species and for each treatment. Light response curves were constructed after plants adopted to dark and PAR was gradually increased from 0 to 2000 µmol m-2 s-1 at leaf temperatures of 25°C and 35°C, at 400 ppm ambient CO2 concentration with 55% relative humidity. WUE was calculated as dividing A by transpiration rate. A/Ci curves were constructed starting from ambient CO2 and gradually increasing and then decreasing CO2 value to compensation point. The curves were constructed for C3, C2 and C4 plants at 30°C and 35°C leaf temperatures, 1000 and 1500 PAR m-2 s-1, higher values used for C4 species. For growth analysis, about 60-day-old plants were randomly selected from each species just before flowering.

Gas-exchange. Under well-watered conditions, C4 G. gynandra showed similar A rates compared to C3 T. hassleriana up to 25 oC leaf temperatures. The C4 performed better than the C3 species at higher temperatures. Under water stress, the two species showed similar A values up to 30oC, but the C4 had higher A rates at higher leaf temperatures. Under well-watered conditions, M. arvensis (C2) showed the highest rates of A at all leaf temperatures followed by M. moricandioides (C3) and by B. oleracea (C3). M. arvensis had higher A rates than B. oleracea, which had higher rates than M. moricandioides under water stress condition.
A-PAR curves measured at 35oC showed that under drought stress C4 plants had an advantage over C3 and under well-watered conditions the C4 yielded nearly 2-fold A rates compared with the C3 plants in Cleomaceae. In Brassicales, well-watered M. arvensis showed the highest and stressed M. moricandioides the lowest values of A rates especially at higher light intensities.
In addition, stomatal conductance, transpiration rates, WUE and mitochondrial respiration were measured for a better understanding of different physiological responses of species under drought stress. Also, another aim was to understand how C2 species would respond to stress compared with sister C3 species. A-Ci curves were constructed to see the responses of plant species to elevated CO2 and their compensation points. The data showed that the C4 G. gynandra reached maximum assimilation at lower Ci values and had significantly lower CO2 compensation point compared to the C3 T. hassleriana. M. arvensis (C2) had the highest values of A up to 550 ppm Ci and the lowest CO2 compensation point compared to M. moricandioides and B. oleracea. M. moricandioides and B. oleracea had the same CO2 compensation points of 150 ppm CO2.
Moreover, Vo/Vc ratio, mesophyll conductance and chloroplastic CO2 concentrations of all species under both treatments were constructed via combination of gas exchange and chlorophyll fluorescence method and the results will be published soon in peer-reviewed journals.

Growth and Carbon Partitioning. Under drought stress C3 plants tends to produce more seeds than C4 in Cleomaceae (Harvest index: seed/shoot ratio) and this pattern can also be seen in Brassicales with C3 plants yielding higher values of harvest index compared to C2 plants under drought stress. Also, C3 plants invest more in leaves compared to other photosynthetic types. Dry weight distribution between plant parts of T. hassleriana and G. gynandra, M. arvensis, M. moricandioides and B. oleracea under drought stress was compared. Both C3 and C4 species of Cleomeceae family and C3 and intermediate plants of Brassicaceae family highly invested in reproductive parts under well-watered conditions. In all plants, a slight reduction in reproductive parts observed under water stress. In addition, under drought conditions, G. gynandra and B. oleracea increased their root and shoot contents, while T. hassleriana and both Moricandia species had shown a less investment in all parts.

Water use efficiency (WUE). The C4 plants G. gaynandra yielded higher WUE values than the C3 species T. hassleriana under well-watered and water-stressed conditions. In the control group, M. arvensis had the highest WUE at 25 and 35°C leaf temperatures. Although, B. oleracea showed a close value to M. arvensis at 25 °C, it yielded the lowest values among all plants at 35oC.

Hybridization. Hybridization attempts in Cleomeceae family and in Brassicaceae family were unsuccessful. Seeds between parental generations were infertile. All other grown seeds after hybridization process showed no sign of distinct properties from parental material. Anatomical and hydraulic studies were performed in all species and important results will be published in the near future with the acknowledgement of 3to4 and EU financial support.

TASK 8: Non-biased analysis of photosynthetic and non-photosynthetic processes (UDUS and Nebion)
Moricandia lines with operating C3 (M. moricandioides) and C3-C4 intermediate (M. arvensis; M. suffruticosa) photosynthesis were grown under controlled conditions in the greenhouse and growth chambers at DUS. Mature plants were analysed for their gas exchange features, leaf anatomy, metabolome and transcriptome. Comparison of the closely related species with differences in the photosynthetic setup aimed on one side at the identification of features that could have been beneficial for the multiple evolution of C3-C4 intermediate photosynthesis in this particular branch of Brassicaceae, and on the other side also at the identification of features which could be responsible for the lack of C4 evolution in this group of plants (summarised also in: Schlüter et al., 2017).
The initial step from C3 to C3-C4 photosynthesis is closely related to the localisation of the GLDP protein from the photorespiratory pathway. In C3 species, the protein is present in mesophyll and bundle sheath cells. In C3-C4 species, it is restricted to the bundle sheath thus initialising additional metabolite exchanges between the two cell types. Availability of transcriptome data from the Moricandia species now allowed us to compare the sequences of the GLDP with related species within a phylogenetic tree. Interestingly the subgroup of the Brassicaceae containing all C3-C4 intermediates seems to have lost the GLDP2 copy. Analysis of the promoter of the remaining GLDP1 gene copy revealed, that in the genus Moricandia a mesophyll as well as a bundle sheath specific element was present in all C3 species, but the mesophyll element was lost in the C3-C4 intermediates (Adwy et al., 2015). So evolution from C3 to C3-C4 in the Brassicaceae seems to have been facilitated first by the loss of the GLDP2 copy followed by the dominant activity of a bundle sheath specific element in promoter of the remaining GLDP1 copy.The experiments also gave hints towards two possible reasons for the lack of progression from the C3-C4 intermediate stage to C4 in Moricandia. Firstly, there is the possibility that anatomical constrains limited progression towards C4. The efficiency of the glycine shuttle and connected C- and N-balancing mechanisms depends on enhanced metabolite exchange between the mesophyll and bundle sheath cells and are therefore dependent on a limited distance between the two cell types. The importance of the narrow vein spacing increases with the increasing contribution of a C4 cycle in advanced C3–C4 intermediates and finally through to full C4. In all tested Moricandia species, the vein density was however generally low especially in comparison with plant families containing C4 species besides C3 and C3-C4 species such as Flaveria and Heliotropium. It is therefore possible that limited anatomical pre-conditions such as low vein density hampered evolution to C4 in the Moricandia.
Secondly, the Moricandia experiments also showed that environmental constrains might have contributed to the abidance of Moricandia at the C3-C4 stage. Evolution of C4 is generally driven by conditions of high photorespiration such as high temperatures and C limitation of photosynthesis. In contrast to C3–C4 lines with C4 relatives, intermediate Brassicaeae grow in colder climates (Lundgren and Christin, 2017), so pressure to reduce photorespiration might also be limited. Instead of increased input into the metabolite shuttle mechanism between the mesophyll and bundle sheath cells, the intermediate Moricandia species were characterised by reduced transcript levels of Calvin-Benson-Bassham cycle transcripts, low initial slopes of the A-ci curve indicating low Rubisco contents and high CN ratios; together these observations suggest that besides C also N metabolism might have played a major role in the evolution of C3-C4 intermediate species in the genus Moricandia.

WP 9 Build the C4 pathway in rice, both in terms of the metabolic pathway and in generating the Kranz anatomy

Objectives.
1. Generate rice lines that accumulate components of the C4 pathway. Rice lines containing 13 maize genes associated with the C4 pathway have been generated in the initial phase of the C4 rice project and are available for this WP. These lines include those with genes encoding core C4 cycle proteins (carbonic anydrase, phosphenolpyruvate carboxylase, malate dehydrogenase, NADP-dependent malic enzyme and pyruvate,orthophosphate dikinase), those with genes encoding metabolite transporters (DIT1&2), and those in which artificial microRNAs have been used to restrict the accumulation of glycine decarboxylase and RuBisCO to BS cells. To build on and add value to these resources we aim to pyramid existing lines. Specifically, existing lines will be crossed to generate transgenic lines in which PEPC, MDH and PPDK are combined with low amounts of RuBisCO and glycine decarboxylase in mesophyll cells, and NADP-ME and RuBisCO accumulate in BS cells. We will also generate a tag that marks nuclei of rice in either M or BS cells, and so allows gene expression in these cells to be assessed. These lines in which the M or BS cells are tagged will be crossed with existing transgenics in order to follow alterations in gene expression in these cell types.
2. Characterize putative regulators of Kranz anatomy. Maize foliar leaves develop classical C4 Kranz anatomy but the husk leaves surrounding the ear display characteristics of C3 leaves (Langdale et al., 1988). As part of the C4 rice program, comparative RNA-sequencing analysis of foliar and husk leaf primordia has identified candidate positive regulators of Kranz anatomy. Depending on the filtration criteria used, the number of candidate genes ranges from 25 to 100. To characterize these genes, we will systematically knock each one down with miRNA in the C4 model grass Setaria viridis, and over-express each one in C3 rice. We will start with the 25 genes identified by the most stringent filtration criteria.

TASK 1: Crossing and phenotyping lines of rice with components of the C4 pathway (CAM and IRRI)

Original goal. To build up components of the C4 cycle in an attempt to introduce elements of the C4 cycle in C3 rice. The approach was to use existing transformants for each enzyme, and iteratively cross these. Crossing allowed double transformants for five of the C4 cycle transgenics to be generated.

Project achievements. The outcome is a set of transgenic rice lines containing various components of the C4 cycle in rice. This is a major achievement due to the number of crosses involved, and the screening for presence of transgenes in progeny to take into account segregation. In essence the work-flow was to pyramid using a scheme that generated two lines that contain three components of the C4 cycle (PEPC x PPDK x MDH and PEPC x PPDK x ME), two lines that contain four components of the C4 cycle (PEPC x PPDK x MDH x ME and PEPC x PPDK x MDH x MEP1), and then two lines that each contain five components of the C4 cycle (PEPC x PPDK x MDH x ME x OMT1) and PEPC x PPDK x MDH x ME x GDCH). The most advanced lines each contain five transgenes. The first is homozygous and contains maize genes for CA, PEPC, MDH ME and PPDK. The second is segregating but contains knock-down of GDC combined with over-expression of PEPC, MDH ME and PPDK. In each case, protein levels of PEPC, PPDK, MDH and ME were verified by Western blotting. The lines have been phenotyped both for visible alterations compared with controls and alterations in labelling of 13C after fixation during photosynthesis.

Rice containing stacked transgenes being used to build the C4 pathway in rice. In addition, lines of rice were generated to allow mRNA accumulating in either the mesophyll or bundle sheath cells of rice to be isolated and sequenced. Three constructs were designed and generated and which consisting of three promoters (pZmPEPC, pZjPCK and pZmUBI) so that the tag will be expressed in mesophyll cells, bundle sheath cells and constitutively respectively. These constructs were transformed at IRRI and transgenic lines selected. Immunoblotting was used to confirm accumulation of the FLAG fusion protein, and selfing used to generate homozygous T3 lines. Immunocapture was then used to obtain ribosomes from each line, and mRNAs resident on these ribosomes purified from all lines. For both pZmPEPC and pZmUBI sufficient RNA was collected to allow deep sequencing to determine the translatome of the M compared with the whole leaf. While it was possible to isolate RNA from the BS line, there was not enough to allow sequencing, and so this experiment is currently being repeated with greater numbers of plants.

TASK 2: Placing candidate genes for Kranz anatomy into Setaria and rice (UOXF)

Original goal. To characterize genes identified as candidate regulators of Kranz anatomy, we will systematically knock each one down with RNAi in the C4 model grass Setaria viridis, and over-express each one in rice. We will start with the 25 genes identified by the most stringent filtration criteria. The transgenic lines generated will be assessed for altered leaf anatomy.

Project achievements. Twenty five maize genes were constitutively expressed in rice and transgenic lines characterized in the T1 generation where possible. If transgene expression led to severe growth perturbations such that T1 seed could not be harvested, leaf anatomy was instead characterized in multiple T0 plants. The 25 genes were part of a larger study which included a total of 60 genes. Of the 60, constitutive expression of 47 led to no phenotypic perturbations. Given that many of these genes encoded transcription factors, this observation implies either that plant development is buffered against coarse changes to these transcriptional networks, or that the maize genes could not activate downstream genes in rice. The remaining 13 genes perturbed aspects of hormone homeostasis or secondary wall formation. In the case of 10 genes, the perturbations caused were consistent with an endogenous role for the gene in vascular development. More sophisticated manipulation of gene function now needs to be carried out in rice, to fine tune spatio-temporal regulation of gene activity. More importantly, loss of function analyses need to be carried out in a C4 species. A manuscript reporting on the rice overexpression lines is about to be submitted to Scientific Reports and the 3to4 funding is acknowledged.
A year into the project it became clear, that high-throughput transformation in Setaria viridis would not be possible and thus this objective was abandoned (allowing for more genes to be characterized in overexpressing rice lines). However, over the last year a Setaria transformation service became available at the Boyce Thompson Institute, USA. As such, with permission from the programme manager, funds were used to have RNAi lines generated for 7 of the genes that had perturbed development when overexpressed in rice. These lines are currently growing and setting seed. Phenotypic characterization will be done as part of the C4 rice programme.

References
Aubry S, Aresheva O, Reyna-Llorens I, Smith-Unna RD, Hibberd JM, Genty B (2016) A Specific Transcriptome Signature for Guard Cells from the C-4 Plant Gynandropsis gynandra. Plant Physiology 170: 1345-1357
Aubry S, Kelly S, Kumpers BMC, Smith-Unna RD, Hibberd JM (2014) Deep Evolutionary Comparison of Gene Expression Identifies Parallel Recruitment of Trans-Factors in Two Independent Origins of C-4 Photosynthesis. Plos Genetics 10 https://doi.org/10.1371/journal.pgen.1004365
Bräutigam A, Gowik U (2016) Photorespiration connects C3 and C4 photosynthesis. Journal of Experimental Botany 67: 2953-2962
Bräutigam A, Schliesky S, Külahoglu C, Osborne CP, Weber APM (2014) Towards an integrative model of C4 photosynthetic subtypes: insights from comparative transcriptome analysis of NAD-ME, NADP-ME, and PEP-CK C4 species. Journal of Experimental Botany 65: 3579-3593
Burgess SJ, Granero-Moya I, Grange-Guermente MJ, Boursnell C, Terry MJ, Hibberd JM (2016) Ancestral light and chloroplast regulation form the foundations for C-4 gene expression. Nature Plants 2 16161. doi: 10.1038/nplants.2016.161
Covshoff S, Furbank RT, Leegood RC, Hibberd JM (2013) Leaf rolling allows quantification of mRNA abundance in mesophyll cells of sorghum. Journal of Experimental Botany 64: 807-813
Kulahoglu C, Denton AK, Sommer M, MaSs J, Schliesky S, Wrobel TJ, Berckmans B, Gongora-Castillo E, Buell CR, Simon R, De Veylder L, Brautigam A, Weber APM (2014) Comparative transcriptome atlases reveal altered gene expression modules between two Cleomaceae C3 and C4 plant species. The Plant Cell 26: 3243-3260
Kümpers BMC, Burgess SJ, Reyna-Llorens I, Smith-Unna R, Boursnell C, Hibberd JM (2017) Shared characteristics underpinning C4 leaf maturation derived from analysis of multiple C3 and C4 species of Flaveria. Journal of Experimental Botany 68: 177-189
Mallmann J, Heckmann D, Brautigam A, Lercher MJ, Weber APM, Westhoff P, Gowik U (2014) The role of photorespiration during the evolution of C-4 photosynthesis in the genus Flaveria. Elife 3 e02478 doi: 10.7554/eLife.02478
Schlüter U, Bräutigam A, Gowik U, Melzer M, Christin P-A, Kurz S, Mettler-Altmann T, Weber APM (2017) Photosynthesis in C3–C4 intermediate Moricandia species. Journal of Experimental Botany 68: 191-206
Wang Y, Brautigam A, Weber APM, Zhu X-G (2014) Three distinct biochemical subtypes of C4 photosynthesis? A modelling analysis. Journal of Experimental Botany 65: 3567-3578
Potential Impact:
In addition to research activities the 3to4 consortium also held an intensive training programme both for internal and external (non-consortium) researchers. This increased dissemination both within and without the consortium and was aimed at improvement of the career prospects of the young researchers. See also Template A1 (scientific publications) and Template A2 (dissemination activities).

Project Meetings and Training

List of project meetings. Agendas and minutes for these meetings as well as many presentations given at them can be found on the 3to4 website, http://www.3to4.org/. Where possible, these meetings were organised in conjunction with international conferences or training events to maximise attendance at them and also to maximise dissemination from the 3to4 consortium via presentations and posters.

Project Inception Meeting, 14-15 February 2012, Schloss Mickeln, Heinrich-Heine Universität Düsseldorf, Germany
First Annual Meeting, 27-28 June 2012, Crowne Plaza Hotel, Salzburg, Austria (prior to SEB meeting in Salzburg 29-30 June 2012 on Improving Photosynthesis)
Second Winter Meeting,20-21 January 2013, Hotel Roma, Lisbon, Portugal
Second Annual Meeting, 6 and 10 August 2013, iHotel, Champaign, Illinois, USA (in conjunction with the C4/CAM triennial meeting held in Champaign, Urbana, USA from 5-10 August 2013 and prior to the 16th International Photosynthesis Congress in St Louis, USA)
Third Winter Meeting, 19-20 January 2014, SN Hotel, Lisbon, Portugal
Third Annual Meeting, 28-30 June 2014, NH Hotel, Potsdam, Germany (in conjunction with a consortium Training Event)
Fourth Winter Meeting, 17-19 January 2015, Elite Hotel, Istanbul, Turkey
Fourth Annual Meeting, 15-17 July 2015, NH Voltaire Hotel, Florence, Italy (in conjunction with a consortium Training event)
Fifth Winter Meeting, 6-8 January 2016, Park Bosphorus Hotel, Istanbul Turkey
Fifth Annual Meeting, 3–4 August 2016, Courtyard Marriott, Düsseldorf, Germany (in conjunction with satellite workshop on C4 Photosynthesis held in Düsseldorf, on August 5-6 2016 and prior to the 17th International Photosynthesis Congress in Maastricht, Netherlands)

WP 11 Training of Researchers
TASK 1: Design and delivery of specialist training workshops in summer schools for researchers (and additional participants from outside the consortium)

First Summer School on C4 photosynthesis and complementary skills 28-30 June 2012: The first summer school on was associated with the second consortium meeting and the Society for Experimental Botany meeting that took place in Salzburg, on June 2012. It was aimed at giving a general overview on the state of the art of C4 photosynthesis to the young researchers contracted by the partner organizations and comprised the activities described below.

Introduction to the Training aims, planned workshops and secondments by Prof Manuela Chaves
General Lecture ‘Breeding success stories’ Prof Andy Greenland.
Lecture on Website skills, Paul Davis, LDM
Session of the SEB meeting on SCIENCE WITH IMPACT
Chairs: Dr Michael Berenbrink (University of Liverpool, UK) and Dr Alun Anderson (Society for Experimental Biology and former Editor-in-Chief, New Scientist)

• Dr Lynne U Sneddon (University of Chester, UK) From science, philosophy and literature to Trivial Pursuit: The impact of studying welfare in fish
• Prof Hans O Poertner (Alfred Wegener Institute for Marine and Polar Research, Bremerhaven, Germany) How animal physiology informs the Intergovernmental Panel on Climate Change (IPCC)
• Dr Rod Wilson (University of Exeter, UK) Can fish influence Earth’s climate control? The surprising role of marine fish in the global carbon cycle
• Prof Steve Long (University of Illinois, Illinois, USA) Why do we need plant biotechnology in our toolbox to feed and fuel the world?
• Dr Anthony Hall (University of Liverpool, UK) Next generation genetics: A new tool for accelerated gene discovery and breeding

2 day session of the SEB meeting on Improving Photosynthesis.
Chairs: Prof Richard Leegood (University of Sheffield, UK) and Prof Steve Long (Champaign, Illinois)
• Dr John Sheehy (IRRI) Why improve photosynthesis?
• Dr. Christine Raines (University of Exeter, UK) Metabolic engineering to enhance photosynthesis based on empirical data and in silico modelling.
• Prof Martin Parry (Rothamsted Research, UK) Rubisco activity and regulation as targets for crop improvement
• Prof Christoph Peterhänsel (LUH) Combined transcriptome, proteome and metabolite profiling reveals enhanced stress resistance of plants overexpressing a photorespiratory bypass
• Dr Julian Hibberd (CAM) Insights into the evolution of C4 photosynthesis
• Dr Andrea Bräutigam (DUS) Towards a blueprint of C4 photosynthesis in grasses
• Dr Mara Schuler (OXF) Comparative transcriptome analysis of maize leaves revealed putative regulators of C4 Kranz anatomy
• Prof Andreas Weber (DUS) Transport of photorespiratory intermediates between cellular compartments
• Prof Paul Quick (IRRI) C4 photosynthesis in rice?
• Mr Ben Williams (Cambridge University, UK) A motif present in multiple C4 genes is necessary for mesophyll specificity

Second Summer school on Bioinformatics August 27-30 2013: The second 3to4 summer school took place in ETH Zurich (Switzerland) from August 27-30, 2013. It was organized by Stefan Bleuler (Nebion) and Carla Pinheiro (ITQB). This course aimed to provide to the young researchers contracted by the partner organizations practical skills in Bioinformatics. The programme comprised: Expression data analysis using Genevestigator; Clustering and Bioclustering; Biological networks: overview and applications; Scripting, filtering and combining data from files. 15 researchers from six countries participated.

Stefan Bleuler, Philip Zimmermann (Nebion)
• Administrative issues, introductions, outlook, computer setup (Stefan Bleuler)
• Genevestigator: Introduction, workflows for different biological questions
• Clustering and bioclustering: theoretical foundations, practical issues
• Biological networks: overview and applications
Beat Gfeller, Markus Wyrss, Stefan Bleuler (Nebion)
• Unix shell, simple Unix tools, regular expressions
• Introduction to scripting (in Ruby), text file manipulation
Stefan Bleuler (Nebion)
• Scripting, filtering and combining data from files
The course allowed interactive sessions with a good working atmosphere and highly motivated participants. The course attendees were asked to score the summer school: an average score of 8 out of 10 was achieved. All participants received a certificate with the necessary information to be recognized both nationally and internationally.

Third Summer school on Metabolite Profiling July 1-4, 2014: The summer school took place at the Max-Planck-Institute of Molecular Plant Physiology (Germany), organized by Toshihiro Obata (MPIMP), Alisdair Fernie (MPIMP) and Carla Pinheiro (ITQB). This course aimed to provide to the young researchers contracted by the partner organizations the basics of metabolite profiling of primary and secondary metabolites by gas chromatography-mass spectrometry and liquid chromatography-mass spectrometry. The programme comprised:

Toshihiro Obata (Golm)
• metabolite profiling and GC-MS analysis;
• experimental design and sampling strategy;
• chromatogram evaluation;
• facility tour (greenhouse and machine park).
Takayuki Tohge (Golm)
• basics of LC-MS analysis; chromatogram evaluation for LC-MS;
• chromatogram evaluation;
• statistic analysis and biological interpretation;
• metabolic flux analysis
The course allowed interactive sessions with a good working atmosphere and highly motivated participants. In addition to the 3to4 researchers, several MPIMP researchers were also present on the theoretical sessions. The course attendees were asked to score the summer school: an average score of 6.9 out of 10 was achieved. All participants received a certificate with the necessary information to be recognized both nationally and internationally.

Ecophysiology Techniques Workshop 8-13 September 2014: The consortium sponsored attendance of its researchers at the workshop organised by a special interest group of the Society for Experimental Biology and the British Ecological Society and provided local support to the organization (Carla Pinheiro, ITQB). The workshop was held in Lisboa (Portugal). The workshop provided knowledge on photosynthesis, plant water relations, whole plant physiology and ‘-omic’ techniques.
The course attendees were asked to score the summer school: an average score of 4.5 out of 5 was achieved. All participants received a certificate with the necessary information to be recognized both nationally and internationally.

Fourth Summer school in Science Communication 13th July 2015: The fourth summer school took place at in Florence (Italy) and was organised by Melissa Marshall (Melissa Marshall Consulting LLC) and Carla Pinheiro (ITQB). This course aimed to provide to the young researchers contracted by the partner organizations with effective communication skills. The programme focused on:
• Qualities of Effective Technical Presentations
• The assertion evidence slide design
• Dynamic Delivery
• Achieving Confidence

Additionally, the researchers were asked to prepare and film a scientific presentation (5 min). The films were produced by Stream Projects (Partner STR) and presented at the project meeting in Istanbul (January 2016). The ones approved will be available at ‘Members Only’ pages of the 3to4 website.
The course attendees were asked to score the summer school: an average score of 8.2 out of 10 was achieved. All participants received a certificate with the necessary information to be recognized both nationally and internationally.

Visit to the Biochemtex Research Center (Tortona, Italy) 14 July of 2015: Biochemtex is a partner organisation of the 3to4 consortium devoted to biofuels and biochemicals production. The 3to4 researchers visited the pilot plant and became acquainted with the second generation technologies for the conversion of lignocellulosic biomass.

Fifth 3to4 summer school Discovering plant biology through RNASeq August 1-2, 2016: Held at Heinrich-Heine-University, Düsseldorf (Germany) organized by Dominik Brilhaus (HHU), Stefan Bleuler (Nebion) and Carla Pinheiro (ITQB). This course aimed to provide to the young researchers contracted by the partner organizations the basics of RNA extraction and RNAseq analysis. The course was provided by Dominik Brilhaus (HHU), Stefan Bleuler (Nebion), Jean-Marie Droz (Nebion) and Udo Gowik (HHU) and comprised lectures, hands-on sessions and a visit to the BMFZ sequencing facility. Lectures and hands-on sessions included:

• Mapping and quantification
• Differential gene expression
• Clustering, enrichment analyses, data mining
• Data presentation and interpretation

The course allowed interactive sessions with a good working atmosphere and highly motivated participants. Participants were asked to use their own data. In addition to the 3to4 researchers, several HHU students were also present. The course attendees were asked to score the summer school: an average score of 8.0 out of 10 was achieved. All participants received a certificate with the necessary information to be recognized both nationally and internationally.

TASK 2: Organisation of secondments for researchers between laboratories involved in the 3to4 project
List of secondments
• Dr Andrea Bräutigam (DUS) visited Sheffield from 25 February – 1 March 2013 to sample and extract plants for transcriptomic analysis.
• CNR Firenze hosted 3to4 consortium partner Dr Carla Pinheiro (New University of Lisbon) from 8-22 June 2014 as part of the CNR short-term mobility program. Dr Pinheiro has assisted in the design of the planned experiment to investigate the effect of drought and temperature on Moricandia and Cleome species and will be integral in the analysis of metabolites and gene expression.
• 25-28 March 2013: Andrea Bräutigam (DUS) visited University of Sheffield to sample several C4 grasses and learn about the growth conditions of the Amaranthus wild type and mutant species.
• CAM hosted two visits: one from CEA Cadarache (Olga Aresheva, 9-19 October 2012) and one from Lisbon (Isabel Abreu iBET, 26 November - 1 December 2013) to teach use of laser capture microdissection and leaf rolling to isolate mesophyll and bundle sheath cells from C4 plants.
• 21-25 October 2013: Visit by Isabel Abreu and Bruno Alexandre (iBET) to University of Sheffield (Mass Spectrometry Unit) for proteomics analysis and technology transfer.
• 5-11 May 2014 and 8-13 June 2014: Visit by Colin Osborne (USFD) to Ferit Kocacinar (KSU) for joint research (sponsored by NERC, UK).
• CNR Firenze hosted 3to4 consortium partner Dr Carla Pinheiro (New University of Lisbon) from 8th to the 22nd June 2014 as part of the CNR short-term mobility program. Dr Pinheiro has assisted in the design of the planned experiment to investigate the effect of drought and temperature on Moricandia and Cleome species and will be integral in the analysis of metabolites and gene expression.
• CNR Firenze hosted 3to4 consortium partner Dr Carla Pinheiro (New University of Lisbon) from 13-14 July 2015. Dr Pinheiro has visit the growth chambers and checked the plants. It was decided to change the growth conditions of Moricandia and Cleome species for physiological characterization and the analysis of metabolites and gene expression.
• 1.5 days visit to S. Aubry to University of Zurich for paper writing (Aubry et al. 2016).
• 30 May-17 June 2015: Visit by Emmanuel G Escobar (USFD) to Isabel Abreu (iBET) to perform 2-dimensional SDS-PAGE analysis using S. viridis leaf extracts.
• Transfer of protein sample preparation methodology from iBET to USFD (Dr Richard Leegood Lab)
• Transfer of partial catalogue with newly identified phosphorylation sites from iBET to consortium members, in particular to Dr Veronica Maurino’s Lab (DUS) for bioinformatics analysis and study of the functional significance of NADP-ME putative phosphorylation site.
• Visit from iBET to Cambridge (Dr Julian Hibberd) for technology transfer on bundle sheath and mesophyll isolation (Rita Borba, 2016).
• Visit from University of Düsseldorf (Dr Veronica Maurino) to iBET discuss ongoing collaborative work (2016).

WP 12 Dissemination and Impact
Material from the Annual Meeting in Salzburg 2012 was recorded by Stream Projects and has been posted on the 3to4 website http://www.3to4.org/. Further recordings from the Annual Meeting in Potsdam 2014 and from the Training Course in Golm 2014 have been made. Recording of material relating to training was delayed until the 2014 Summer Meeting because of the feasibility of recording the Annual Meeting at Champaign, Urbana in 2013 and of recording training material at the Nebion Training Course held in Zurich in August 2013. In addition, presentations from researchers were recorded as aids to training in presentation skills at the Summer school in Science Communication held in Florence in July 2015. Presentations from Pillar and Work Package Reports were also recorded in Florence in July 2015. All have been posted on the private pages of website. Once sensitive material has been published these will be transferred to the public pages of the website.

Co-operation with other projects/programmes
Communication has been maintained with the BMGF-funded C4 Rice programme (https://c4rice.com/ which is now in Phase 3; https://www.plants.ox.ac.uk/news/phase-iii-c4-rice-project-now-underway) at the International Rice Research Institute, Los Banõs, Philippines, both by virtue of common membership (Quick, Westhoff, Hibberd, Langdale, Zhu, Leegood, Weber) and by presentations by members of the consortium at the annual C4 Rice meeting (held at IRRI in December 2014, December 2015 and December 2016). In addition, Jane Langdale (OXF) spent 3 weeks with the BMGF RIPE (http://ripe.illinois.edu/) project at Urbana in March 2015, hosted by Steve Long.

Project Website
The project website was set up with LDM in March 2012 http://www.3to4.org/. It contains both public and private pages. The former includes recent publications from the consortium and filmed presentations, as well as lecture material from members of the consortium. Private pages include information, such as the agendas of meetings. The content of the public part of the website has been improved to appeal to the general public. There are links on the webpage to both Facebook and Twitter accounts.

The following have been used to communicate information about the project to the general public:

TV coverage/report
• Ferit Kocacinar (KSU): After initiation of the “3to4” project, the participation of KSU was announced in several local and regional newspapers. In early June 2012, Ferit Kocacinar was interviewed and filmed by Anatolian News Agency. The news was hailed as “Green Revolution in Agriculture” and “A Project That Will Excite World’s Agriculture”. http://www.aa.com.tr/tr/tag/62264--dunya-tarimini-heyecanladiracak-proje. The news was also broadcast on the web page of Agriculture TV: Ministry of Food, Agriculture and Livestock of Turkey. http://www.tarimtv.gov.tr/HD1507_dunya-tarimini-heyecanlandiracak-proje.html#.

Radio coverage/report
• Jane Langdale (OXF) Towards the Light. Interview BBC Radio 4 Series - Plants: From Roots to Riches. June 2014 (broadcast August 2014 http://www.bbc.co.uk/programmes/b04cb9gw).
• Ferit Kocacinar (June 2012) was interviewed ‘live’ by a regional Turkish Public Radio, http://www.trt.net.tr/anasayfa/canli.aspx?y=radyo&k=trtcukurova. The project was outlined and information about input of the project on the future agriculture provided.
• Jane Langdale (OXF) Guest Panel Member BBC Radio 4 ‘Infinite Monkey Cage’; January 2015

DVD /Film /Multimedia
• Julian Hibberd (CAM) Podcast for eLife (http://elifesciences.org/content/2/e00961) ‘Ordering the evolutionary events for C4 photosynthesis’.

Coverage in national press
• The 3to4 project was disseminated on several regional and national newspapers in Turkey:
http://ticaretgazetesi.com.tr/haberler_turkiye_3to4_projesinde_yetistirme_ussu-l-1-sayfa_id-666-id-105590
http://www.gidavitrini.com.tr/tarim/bu-proje-uretimi-yuzde-50-artiracak-h4317.html
http://www.kampushaber.org/isim/ferit-kocacinar.html
http://www.sondakika.com/haber/haber-dunya-tarimini-heyecanladiracak-proje-3758687/

Website for the general public/internet
• A 3to4 Facebook page was set up, particularly to facilitate communication between the younger researchers and to the public: https://www.facebook.com/3to4photosynthesis.
• A 3to4 Twitter account (@3to4project) was also set up. The two accounts are linked to the 3to4 website. The Twitter account now has 386 following.

Event targeting general public (festival, conference, exhibition, science café)
• Colin Osborne (USFD) Weston Park Museum, Sheffield (Feb 2012) - Sideshow Science "How to make a plant" A public science show for families, in which we demonstrated how photosynthesis works, and why it is important. http://www.shef.ac.uk/aps/outreach/sideshow-science
• Colin Osborne (USFD) Formby High School (April 2012) "Plants: Earth's Life Support System" A public lecture for adults and sixth formers, explaining how photosynthesis supports human society and all of life on Earth, and how research (including 3to4) is targeting this process to improve crop production.
• Colin Osborne (USFD) University of Sheffield (May 2012) "Wild about Plants" A science festival for local primary and secondary schools of demonstrations and activities showcasing plant science (including photosynthesis). Write-up and photos: http://osbornelab.group.shef.ac.uk/plant-day-2012
• Colin Osborne (USFD) Association for Science Education conference (Jan 2013) "Transforming photosynthesis: frontier biology in a changing world". This was a lecture for school science teachers at their annual conference, in the "Biology in the Real World" session. I talked about why photosynthesis is important, about C4 specifically, and how 3to4 could improve crop production. Videoed available on the BBSRC website: http://www.bbsrc.ac.uk/society/schools/biology-in-the-real-world.aspx; http://mediasite.nbi.ac.uk/Mediasite/Play/1dc818c3ee3a44ddbdfd3eda1bc39a111d?catalog=ebc3a07a-1940-4a6f-98c2-2e0a3a9a2d99
• Colin Osborne (USFD) APS Christmastime Lecture (Dec 2013) - "Sunshine for Breakfast" A lecture for 1,000 Sheffield schoolchildren, about photosynthesis and how it powers our lives via the food we eat.
Photos: https://www.facebook.com/media/set/?set=a.567305706682723.1073741833.531242640289030&type=1
https://www.facebook.com/APSoutreach
Newspaper report and video:
http://www.thestar.co.uk/news/council/education/video-pupils-enjoy-day-out-in-the-sun-1-6299941
• Peter Westhoff (UDUS) NutzPflanzen/PflanzenNutzen: Vom Urweizen der Steinzeit zu den Genpflanzen der Zukunft. Five Public Evening Lectures at Haus der Universität, Schadowplatz 14, Düsseldorf. 29 April 2014: Vom Jäger und Sammler zum Ackerbauern und Viehzüchter, 13 May 2014: Gene und Genome – Was macht ein Pflanzenzüchter?, 27 May 2014: Genpflanzen – Chance für die Landwirtschaft oder Gefahr für den Verbraucher?, 17 June 2014 Pflanzen und Mikroben – Freunde oder Feinde?, 1 July 2014: Die Herausforderungen der Zukunft.
• Peter Westhoff/Andreas Weber (UDUS) NutzPflanzen/PflanzenNutzen: a further series of public lectures given 10 May-6 June 2016.
• The 3to4 project was highlighted as one of the selected projects in an “International Brokerage Event” named “TurKey Enabling Technologies 2012”, co-financed by European Union and Republic of Turkey, with the participation of Mr. Robert Jan Smits, the Director-General for Research & Innovation, held on the 25th of May 2012, in İstanbul Turkey, with 540 participants from Europe and Turkey.
http://tket2012.meeting-mojo.com http://www.fp7.org.tr/home.do?cid=23807
• Carla Pinheiro (ITQB) presented and discussed the project in the Plant Biology classes offered to students of undergraduate program at FCT/UNL (1st cycle) and exhibited at several locations of the FCT/UNL campus.
• Emmanuel G Escobar and Colin Osborne (23-25 May 2015, USFD) participated in a public demonstration showing the importance of photosynthesis and sustainable energy.
• Carla Pinheiro (ITQB) presented and discussed the project in the Plant Biology classes offered to students of undergraduate program at FCT/UNL (1st cycle) and exhibited at several locations of the FCT/UNL campus. The 3to4 project was also presented at EXPO, the science and technology festival at FCT/UNL.
• CNR has actively participated in programs to allow schools to visit CNR research facilities and programs such as ACARISS (Increasing knowledge on environment and risks of pollution involving schools with experimental activities PAR-FAS) and Races: Raising Awareness on Climate and Energy Savings (Life+) where scientists visit local schools to explain their work.
• Andrea Bräutigam (UDUS) ‘3rd RNA-seq workshop’, RWTH Aachen, March 2015 Lectured about RNA-seq analyses in comparative analyses, co-taught the hands-on course.
• Andreas Weber (UDUS) Co-organizer: Synbio PhD Summer School: Synthetic Biology in Photosynthetic Organisms; Denmark, August 2014 (with Dario Leister and Poul Eric Jensen)
• Andreas Weber (UDUS) Co-organizer: Synbio PhD Summer School: Synthetic Biology in Photosynthetic Organisms; Denmark, August 2015 (with Poul Eric Jensen). In both summer schools APMW presented work on engineering C4 into C3, building on results and concepts of the 3to4 project
• Jane Langdale (OXF) Didcot First Café Scientifique ‘Breeding Crops for the Future’; October 2015
• Tiina Tosens (EMU) gave a TedX talk in 2015 in Tartu related to the main goals of the 3to4 project. She introduced the efforts of scientists to assure food safety in the conditions of global change to over 500 people. The talk was also available and highly watched online on the TedX Tartu webpage and has been watched about 400 times on Youtube. It received a special recognition prize from the sponsoring companies.
• Tiina Tosens (EMU) introduced the principles and main goals of 3to4 project to the public in as a part of the across-Estonian series of scientific events “Researcher’s night: Scientific festival” in 2016.

Engagement with Policy Makers
• EMU was invited to the Ministry of Rural Affairs to discuss global changes and explain the importance of the modification of C3 photosynthetic pathways in terms of food safety and creating crops more resistant to climate to Estonian policy makers.
• EMU invited members of the Ministry of Rural Affairs as well as members of other top research groups in plant sciences to our institute to introduce the importance and possibilities of the 3to4 project.
• Richard Leegood (USFD) gave a presentation on the 3to4 project at the Food Security Biotechnology & Food Security Workshop sponsored by the York Environmental Sustainability Institute and by Friends of the Earth May 22, 2012, Biology Department, University of York
• Richard Leegood (USFD) gave a presentation on the 3to4 project at a workshop: ‘Sharing knowledge from EU FP7 projects in plant systems biology’, CRAG, Barcelona, Spain, 15 April, 2013.
• Richard Leegood (USFD) attended gave a presentation on the 3to4 project at a workshop: ‘Plant Breeding in the Context of Research Cooperation between the EU and China’ Brussels 5 December 2014.
• Ferit Kocacinar (KSU) presented a talk about 3to4 at a meeting held at KSU organised by TUBİTAK (Scientific and Technological Research Council of Turkey). March 2015.
• Richard Leegood (USFD) gave a presentation on the 3to4 project at the KEP Bioeconomy Summit – Bioeconomy Ecosystem: Policy, Industry, Science and Technology, 1st Knowledge Exchange Platform Peer-to-Peer event, Helsinki-Uusimaa, 2-3 June 2016.

Section B. Exploitable foreground and provide the plans for exploitation.
The owner of the foreground is identified after the work package title. No patents have so far been applied for.

WP1 Pursue forward genetics with Arabidopsis thaliana as the model species. UDUS.
The route for dissemination is publication, conference presentations, and public lectures.

WP2 Verify gene function in C3 dicot model (Arabidopsis) and crop plants (rapeseed) by synthetic biology. UDUS.
The route for dissemination is publication, conference presentations, and public lectures.

WP3 Transcriptomes of C3 and C4 pairs. UDUS/Nebion.
The transcriptomics data themselves are not exploitable commercially. They will published in individual papers by the experimenters. Nebion is committed to maintain the special instance of Genevestigator containing the 3to4 transcriptomics data and open it to the public free of charge as soon as the experimenters have all published their analyses. The processing pipeline could be reused for similar types of analyses and thus be exploitable for analysis services. Besides the processing of the transcriptomics data, a significant effort was put into making the generated data and results easily accessible for the consortium members (and later the public). Raw data, normalized read count, and initial analysis were deposited into a wiki (https://3to4.genevestigator.com/twiki/bin/view/NGSdata) and a special instance of the Genevestigator platform tailored for the 3to4 transcriptomics data was created (https://3to4data.genevestigator.com/local/; manuscript in preparation) which allows detailed and powerful analysis in a simple graphical analysis software. In addition to the 3to4 data, the public gene expression data was enriched with experiments that are particularly interesting to the 3to4 consortium. All data were collated in a joint database and visualized in the Genevestigator framework. The standardized methods and joint visualization enable ongoing and future analysis. (https://3to4.genevestigator.com/twiki/bin/view/NGSdata; https://3to4data.genevestigator.com/local/).

WP4 Identify candidate genes by comparative transcriptome profiling of mesophyll and bundle sheath UDUS/Nebion.
The route for dissemination is publication and then making the data publicly available through the Nebion interface (see above).

WP5 Identification and analysis of transcription factors controlling C4-specific gene expression. LUH/iBET.
The route for dissemination is publication, conference presentations, and public lectures.

WP6 Post-translational regulation of major C4 genes in maize and Setaria. iBET/UDUS/USFD.
The route for dissemination is publication, conference presentations, and public lectures.

WP7. Development of a generic model of C4 photosynthesis. SIBS/MPG.
The route for dissemination is publication, conference presentations, and public lectures. The information is also being introduced into the Bill and Melinda Gates Funded C4 Rice Phase III project.

WP8 Understanding Model Systems. CAM/LUH/KSU/CEA/CNR/EMU/UDUS/Nebion.
The route for dissemination is publication, conference presentations, and public lectures.

WP9 Build the C4 pathway in rice, both in terms of the metabolic pathway and in generating the Kranz anatomy. IRRI/UOXF/CAM.
Lines developed at IRRI in which components of the C4 cycle were introduced into rice at IRRI are being analysed and the data being prepared for publication. The information and lines are also being introduced into the Bill and Melinda Gates Funded C4 Rice Phase III project as background IP owned by IRRI. Lines of rice developed by Cambridge and IRRI that contain ribosomes labelled with a FLAG tag are now being used as part of the BMGF funded C4 Rice project as background IP owned by the University of Cambridge and IRRI. Any potential IP that is derived from these lines will be assessed as the data are interpreted.
For the Kranz regulators (transcription factors), most of the Kitaake rice lines generated at UOXF during testing for candidate regulators of Kranz anatomy have been submitted for publication and will not be worked on further. One remaining line, ZjPCKpro:ZmSHR1 is being further investigated by Dr Mara Schuler as part of the CEPLAS programme at UDUS (and a manuscript is also being prepared for publication). Setaria RNAi lines generated for Kranz candidate regulators are currently being propagated and will be introduced into the Bill & Melinda Gates Funded C4 Rice Phase III project as background IP owned by the University of Oxford.

WP12 Dissemination and Impact.
Presentations from Work Package Reports were also recorded though the duration of the project. All have been posted on the private pages of website (http://www.3to4.org/). Once sensitive material has been published these will be transferred to the public pages of the website.

List of Websites:
http://www.3to4.org/
This website will be maintained for at least 3 years after the conclusion of the project.

Final Report Summary - 3TO4 (3to4: Converting C3 to C4 photosynthesis for sustainable agriculture)

Share this page

Download