Upregulation of chloroplastic pyruvate dehydrogenase genes in rice leaf would potentially drive the in planta photorespiratory bypass for higher biomass

At ambient temperature (25–30 o C) and the prevailing atmospheric CO 2 levels (380 ppm), installing the C 4 photosynthetic machinery in a C 3 plant would potentially drive away the photorespiratory process through a carbon concentrating mechanism (CCM), thereby preventing oxygenation reaction of Rubisco. Development of C 4 rice is a global research priority, for enhanced water use eciency (WUE) and yield. At optimal environment, the difference in the solar energy to biomass conversion between C 3 and C 4 plants is mainly due to photorespiration. So, photorespiratory bypasses are the potential alternatives than conversion to C 4 . Genetically transformed C 3 model plants with photorespiratory bypass had demonstrated higher biomass (under same environmental conditions) than its wild type. Using a transcriptome approach, we report here the differential expression pattern for photorespiratory genes and chloroplastic pyruvate dehydrogenase (plPdc) gene between the leaves, peduncle, and the developing grain tissues in three rice genotypes. In addition to pyruvate, glycolate and glyoxylate also are the substrates for the plPdc gene product and hence a suitable candidate for photorespiratory bypass.


Introduction
At optimal environmental conditions, solar energy to biomass conversion e ciency is roughly 25% higher for C 4 plants than the C 3 ones, with key differences in photorespiration, while loss through respiration in light is unavoidable [1][2][3] . The C 4 trait, a carbon concentrating mechanism (CCM) that drives away photorespiration through preventing oxygenation reaction of Rubisco, is reported to have convergently evolved multiple times, nevertheless, same gene lineages were recruited in the C 4 trait evolution 4,5 .
Converting a C 3 crop plant with a C 4 pathway to enrich the solar energy conversion e ciency for higher biomass is one of the research priorities to improve yield. Installing photorespiratory bypasses is also a viable alternative, demonstrated to improve e ciency with higher biomass through reduced photorespiration in model C 3 plants 6-8 . Also, it is important for plants to metabolize 2-phosphoglycolate (2-PG, formed through the oxygenation process of rubisco) and glyoxylate (key intermediate of photorespiration), to overcome the metabolite toxicity that inhibits photosynthesis and starch biosynthesis 9,10 . Installing a C 4 machinery or a photorespiratory bypass in C 3 plants would help enhance the assimilation rate under optimal environmental conditions 7,8,11 .
In addition to pyruvate, glycolate and glyoxylate -the intermediates of photorespiration -also acts as a substrate for the chloroplastic pyruvate dehydrogenase complex (plPDC) in plants 12 . The PDC constitutes three components, E1 (pyruvate dehydrogenase in heterotetramer state-a2b2), E2 (dihydrolipoyl acetyltransferase, homodimer) and E3 (dihydrolipoyl dehydrogenase, monomer) with copy numbers of these components in the complex is variable 13,14 . Recent study highlights the E2 component's RNA binding activity with psbA mRNA coding for the D1 protein of the PSII reaction center 15 . To understand the expression pattern for genes of photorespiration and the C 4 pathway in leaf and non-leaf (photosynthetic) tissues in rice, transcriptome analysis was performed in the three rice genotypes, with two biological replicates.

Materials And Methods
Three rice (Oryza sativa ssp. indica) genotypes -Apo (EC734333), BAM4234 (EC497171), and Crossa (IC575838) -were grown under eld conditions during Kharif season 2018 in triplicate at the research farm of the Division of Plant Physiology at IARI (New Delhi). Flag leaf, peduncle and developing grains (3-5 days-post-anthesis, dpa) were collected in two replicates, snap frozen using liquid nitrogen (-196 o C) and stored at -80 o C for transcriptome studies. The experiment was planned with two replicates since the study involves three genotypes. The expression levels between genotypes for most of the genes studied were insigni cant (Supplemental File_S1, hence genotypes could equally be considered as biological replicates, totaling to six (two replicates x three genotypes). So, technically, the expression pattern reported in the study for each tissue is supported with an equivalent of six biological replicates. All methods pertaining to this study were performed in accordance with the relevant guidelines / regulations / legislation as applicable.
Total RNA from the samples (80-100mg) was extracted using a RNeasy plant mini kit (Qiagen, USA) following the manufacturer's protocol. The quality and quantity of the RNA was assessed using a Bioanalyzer 2100 (Agilent technologies, USA) and spectrophotometer ND-8000 (Thermo Scienti c, USA).
The RIN values for the 18 samples (3 genotypes and 3 tissues, repeated twice) ranged from 7.0 to 9.5. RNA-seq libraries were sequenced on an Illumina platform (2x150bp paired-end reads). A total of 521 million pairs of reads were obtained. Adapter trimmed reads were quality checked using FastQCv0.11.8 16 . These pre-processed reads were mapped against the indica rice (ASM465v1) genome sequence 17 . Mapping and alignment against the reference were done using Tophatv2.1 18 . Summary statistics on the number and percent reads mapped were provided in the Supplemental File_S2.
Cu inksv2.2.1 was used to assemble the individual transcripts for expression quanti cation 19 . The assembled transcripts were merged for the differential expression studies between each tissue (leaf vs peduncle, leaf vs grain, and peduncle vs grain) in every genotype and vice-versa (to con rm no signi cant differential expression for the genes / transcripts studied, between genotypes for the same tissue) using Cuffmerge 20 . The expression values (in RPKM) were tested for statistical signi cance at FDR 0.01 cutoff value using Cuffdiffv2.2.1 19 .
Including the key eight genes coding for the gene products being involved in photorespiration 21 , totally, 42 transcripts (with gene ids) were identi ed for the 11 genes involved in photorespiration viz., phosphoglycolate phosphatase-chloroplastic (cpPGLP), glycolate oxidase-peroxisomal (pGOX), glutamate:glyoxylate aminotransferase-peroxisomal (pGGT), serine hydroxymethyltransferasemitochondrial (mSHMT), glycine decarboxylase-mitochondrial (mGDC), glycerate kinase-chloroplastic (cpGLYK), glutamine synthetase-chloroplastic (cpGS2), glutamate synthase (cpGOGAT), serine:glyoxylate aminotransferase-peroxisomal (pSGT), and hydroxypyruvate reductase-1 and − 2 (HPR-1 & -2). The corresponding transcript ids were identi ed and extracted from the plants ensembl database 22 . This is done since few genes were not functionally annotated. Expression pro les (in RPKM -reads per kilobase of transcript per million mapped reads) including statistical signi cance and log fold change details for the genes of interest were extracted from the transcriptome analysis (Supplemental File_S1) and studied for its biological signi cance. Based on the results obtained, the expression pro les for transcript ids annotated with chloroplastic pyruvate dehydrogenase complex gene (pdc) were also studied from the transcriptome dataset and results are tabulated and given in Supplemental File_S1. In addition, expression levels for Rubisco small subunit (rbcS) transcripts were also compared between the three tissues for all the three genotypes. Since the expression values of the rbcS transcripts are also signi cantly downregulated in developing grains (Supplemental File_S1), when compared to leaves, ratio for expression level between leaf and developing grain' in each genotype has been worked out (excel sheet 'Ratio' in Supplemental File_S1). For those transcripts with expression values signi cantly higher in leaves are greater than one. The rbcS transcript with highest expression in both leaf and developing grain is identi ed, and its ratio is used as the threshold ratio to identify the set of photorespiratory genes that are signi cantly downregulated, and simultaneously above the threshold ratio (cells highlighted in green, in excel sheet 'Ratio' in Supplemental File_S1).

Result And Discussion
The eight genes of photorespiratory enzymes viz., phosphoglycolate phosphatase-chloroplastic (cpPGLP), glycolate oxidase-peroxisomal (pGOX), glutamate:glyoxylate aminotransferase-peroxisomal (pGGT), serine hydroxymethyltransferase-mitochondrial (mSHMT), glycine decarboxylase-mitochondrial (mGDC), glycerate kinase-chloroplastic (cpGLYK), serine:glyoxylate aminotransferase-peroxisomal (pSGT), and hydroxypyruvate reductase (HPR); two genes encoding for glutamine synthetasechloroplastic (cpGS2), glutamate synthase (cpGOGAT), were studied and found to be signi cantly downregulated in the developing grains (ca. 3-5 days post-anthesis) than the leaves, in all the three rice genotypes (Supplemental File S1). Although it can be argued that the downregulated expression pattern in developing grains is an expected one when compared to leaves, to identify the biological signi cance, expression pattern for rbcS gene transcript was also studied (Supplemental File_S1, 'ratio' worksheet). Those genes for which the expression pattern ratio between leaf and developing grains are greater than the ratio of rbcS gene, they were identi ed to play proportionately equal or higher role as in the leaves. However, downregulation of photorespiratory genes might lead to cell toxicity due to the accumulation of 2-PG and glyoxylate, notably when the plant is under abiotic stress 9,10 . Conversion of these two metabolites into non-toxic compounds is primarily important to overcome the cellular toxicity, as well as to sustain the availability of ADP and NADP for accepting light energy 23 . Diversion of the 2-PG to bypass the photorespiratory process is reported to improve the plant biomass 6-8 . Alternatively, chloroplastic pyruvate dehydrogenase complex (plPDC) is reported to detoxify glyoxylate, producing CO 2 in chloroplast 12 , potential for a natural photorespiratory bypass to enrich the CO 2 for rubisco carboxylation process. In addition to glyoxylate, glycolate also acts as a substrate for plPDC, and CO 2 production from these metabolites are competitively inhibited in the presence of pyruvate 12 .
So, to understand the expression pattern at transcriptional level, we compared the expression levels of plPdc (Supplemental File_S1) and found that the transcript levels of plPdc were signi cantly higher in the developing grains as compared to the leaves, in all the three genotypes. It gives an insight on the possible use of plPDC to establish a photorespiratory bypass (Fig. 1). This would potentially aid in developing an e cient photorespiratory bypass, in planta, through enhanced plPdc gene expression levels targeting for higher biomass or yield. Whether the upregulated plPdc driverts the photorespiratory process or viceversa, is unknown yet. Plants accomplishing C 2 photosynthesis have evolved for the preferential downregulation of mGDC (glycine decarboxylase, mitochondrial) or anatomical rearrangements (with more chloroplasts at the periphery) in mesophyll cells either reduce CO 2 release or provides high resistance to CO 2 e ux 7,24 . This is commonly called 'C 2 shuttle' and helps increase the plant productivity with high biomass or yield. With initiation of photorespiration through oxygenation reaction of Rubisco, conversion of the toxic metabolites 2-PG (through glycolate) and glyoxylate in chloroplast itself through plPDC to release CO2 will enrich the carbon ux for Rubisco's carboxylation process would simulate a natural photorespiratory bypass. Present study gives an insight for the probable existence of certain group of plants that have evolved to recapture the CO 2 released in the process of converting glycolate / glyoxylate to Acetyl-CoA through plPDC in chloroplast itself, possibly having the shortest C 2 shuttle that also help enhance the plant productivity. Expression levels of Acetyl-CoA carboxylase (ACCase), the key enzyme that channelizes the carbon ux for fatty acid (FA) biosynthesis, is insigni cant between leaf and grain tissues studied and suggesting for the expression of plPDC is not associated with FA biosynthesis. Alternatively, these Acetyl-CoA pools formed through the action of plPDC might possibly utilized for Nterminal acetylation through the action of plastidic N-terminal acetyltransferases (plNATs). This is in line with the reports suggesting ca. 30% of the plastid proteins are subjected to the action of NATs, especially the chlorophyll binding proteins and other enzymes of photosynthetic apparatus [25][26][27][28][29] .
Overall, our results show that, the signi cant downregulation of photorespiratory genes in the developing grains of rice as compared with leaves exhibit biological signi cance; with the ratio (leaf to developing grain tissues) for photorespiratory genes being greater than rbcS gene (Supplemental File_S1). The signi cant upregulation of the chloroplastic pdc gene speci cally in the developing grains, might convert glyoxylate / glycolate, to CO 2 in chloroplasts for carbon xation (Fig. 1), thereby preventing carbon loss 12,30 . This nding provides an insight for possible development of an in planta photorespiratory bypass in the leaves of C 3 plants to envision for higher biomass or yield. Figure 1