CIL and CIA2 show functional redundancy
The cia2 mutants showed pale green phenotype and reduced chlorophyll a and b content;
however, the leaf shape and size of the mutant plants were similar to those of wildtype plants (Col) (Figure 1). A previously study has shown that CIA2, driven by the 35S and CIA2 promoters, increased chlorophyll content of cia2 mutants to recover the phenotype to those of wildtype, indicating that CIA2 expression could rescue the cia2 mutant phenotype (Sun et al., 2009).
Therefore, we used similar strategies to confirm the functional redundancy of CIL and CIA2.
Vectors containing FL CIL CDS driven by the 35S or CIL promoter (-1504 ~ -1) were introduced into cia2 mutants via floral dipping, and four independent strains each of CILpro:HA-CIL/cia2 and 35Spro:HA-CIL/cia2 homozygous transgenic plants were obtained. The phenotype and chlorophyll content of the four CILpro:HA-CIL/cia2 strains were similar (Supplemental Figure S1); therefore, the transgenic strain CILpro:HA-CIL/cia2#1 was used as a representative. Similarly, the transgenic strain 35Spro:HA-CIL/cia2#1 was used as a representative (Figure 1). The phenotype and
chlorophyll content of plants of the CILpro:HA-CIL/cia2#1 strain were similar to those of cia2 mutants; however, the phenotype and chlorophyll content of plants of the 35Spro:HA-CIL/cia2#1 strain were similar to those of Col plants, suggesting that adequate CIL expression could restore the cia2 phenotype to wildtype.
CIL and CIA2 show similar expression patterns
CIA2 is expressed in leaves and young flower buds (Sun et al., 2001), but the expression pattern of CIL remains unknown. To understand the expression pattern of CIL in various tissues and at different developmental stages, we used RT-PCR and qRT-PCR. CIL was not expressed in roots, and its expression was restricted to aboveground tissues, including leaves and young flower buds, indicating that the expression patterns of CIL and CIA2 are similar (Figure 2). Therefore, based on
15
the similarity in expression patterns between CIL and CIA2 and the fact that adequate CIL expression restored the cia2 mutant phenotype, CIL is the homologous protein of CIA2 in Arabidopsis and CIL and CIA2 are complementary genes.
CIA2 and CIL co-regulate chloroplast development
CIA2 has previously been reported to regulate the expression of chloroplast development-related genes (Sun et al., 2001; 2009). Owing to the similar expression patterns of CIA2 and CIL and their complementary functions (Figure 2), CIL may also be involved in the regulation of chloroplast developmental. To determine the function of CIL, cil mutants were used in this study. The
phenotype of cil mutants did not differ from that of Col plants (Figure 3A). PCR was performed to confirm T-DNA insertion and verify the homozygous knockout of CIL in cil mutants. T-DNA (~5 kilobases, kb) was inserted into CIL between 328 and 725 basepair (bp) positions after the start codon to form an incomplete CIL transcript (Figure 3B and C). Furthermore, although the FL CIL CDS could not be amplified using PCR from the genomic DNA of cil mutants, the intersection fragment of CIL and T-DNA could be amplified, indicating that the generated cil mutants were homozygous (Figure 3D). Furthermore, we crossed the cia2 and cil mutants to obtain homozygous cia2/cil double mutants. The leaf size of cia2/cil double mutant was smaller than that of cia2 single mutant, and cia2/cil mutant showed a more severe pale-green phenotype than that of cia2 mutant (Figure 3A). Therefore, the functional defects caused by cil mutant could be observed once CIA2 is absent.
CIA2 enhances the expression of several nuclear genes, including Toc75, Toc33, CPN10-II, and cpRPs (RPL11, RPL15, RPL18, RPL28, RPL29, and RPS6), to regulate chloroplast development (Sun et al., 2009). CIL expression pattern is similar to CIA2 expression pattern; therefore, qRT-PCR was performed to assess whether the genes described above are regulated by CIL as well. Toc75 and Toc33 expression showed no significant difference between cil mutants and Col plants; Toc75 expression was decreased to similar levels in cia2 and cia2/cil mutants; and the decrease extant of Toc33 expression was in a greater extent in cia2/cil double mutants than that in cia2 single mutants
16
(Figure 4A). The expression levels of other Toc genes showed no significant difference between Col and mutant plants, with the exception of Toc34 expression, which was slightly increased in cia2/cil mutants (Figure 4A), suggesting that the Toc34 up-regulation somewhat compensated Toc33 down-regulation. CPN10-II expression was decreased to a greater extent in cia2/cil double mutants than in cia2 single mutants, but the expression level was similar between cil mutants and Col plants.
In contrast, CPN10-III expression showed no significant difference across the three mutants and wildtype plants (Figure 4B). Regarding the cpRPs, there was no difference in RPS6, RPL11, and RPL29 expression between cil mutants and Col plants. In cia2 and cia2/cil mutants, RPL15 expression was slightly increased but RPS6, RPL11, RPL15, and RLP29 expression was decreased (Figure 4C). These results indicate that CIL might regulate the expression of Toc75and cpRPs through CIA2 function and the expression of Toc33 and CPN10-II synergistically with CIA2;
however, CIA2 and CIL are not involved in the regulation of CPN10-III. Therefore, CIL functions similar to CIA2 in the regulation of chloroplast development.
To further confirm that CIA2 and CIL co-regulate chloroplast functions, chloroplast
microstructures were observed in leaves of 10-day-old Col plants as well as cia2, cil, and cia2/cil mutants. The results indicated that the chloroplasts of cia2/cil double mutants showed fewer granal thylakoids and greater development abnormalities (Figure 5J-L). Thus, CIA2 and CIL co-regulate chloroplast development.
CIA2 and CIL co-regulate downstream genes
The cia2/cil double mutants with smaller leaf size and paler green phenotype than the cia2 single mutants revealed functional redundancy of CIA2 and CIL (Figure 3A). To identify the genes
downstream that are co-regulated by CIA2 and CIL, microarray analysis of two independent experiments was performed and genes that showed significant foldchange in expression of ≥ 1.5-fold compared to Col (P ≤ 0.05) were determined. A total of 103, 26, and 279 genes were down-regulated and 70, 20, and 277 genes were up-regulated in cia2, cil, and cia2/cil mutants, respectively (Figure 6A and B; Supplemental Tables S2 and S3). These results suggest that only
17
cia2/cil double mutants affected the genes co-regulated by CIA2 and CIL, further confirming their functional redundancy. In cia2/cil double mutants, the chloroplast development-related and nuclear transcription-related genes constituted 55.6% (155/279) and 30.0% (84/279) of the downregulated genes, respectively (Figure 6C; Supplemental Table S2). Plastid translation-related genes, such as cpRPs, photosystem genes, and chlorophyll biosynthetic genes constituted 23.2% (36/155), 16.1%
(25/155), and 9% (12/155) of the chloroplast development-related genes (Figure 6D; Supplemental Table S2). The expression of these genes was much lower in cia2/cil double mutants than in cia2 single mutants and showed no obvious change in cil single mutants, indicating that these genes are primarily regulated by CIA2 and supplemented by CIL.
During chloroplast development, the plastid genome of land plants expresses
photosynthesis-related genes. This plastid gene expression involves two distinct types of RNA polymerases, namely NUCLEUS-ENCODED RNA POLYMERASE (NEP) and
PLASTID-ENCODED RNA POLYMERASE (PEP) (Hajdukiewicz et al., 1997; Yu et al., 2014).
PEPs form a complex involving interaction with at least 12 PEP-ASSOCIATED PROTEINs (PAPs) (Steiner et al., 2011; Pfalz and Pfannschmidt, 2013) and PLASTID RNA POLYMERASE
SIGMA-SUBUNITs (SIGs) that are involved in PEP regulation (Schweer et al., 2010). All PAPs in the proteomes of a transcriptionally active chromosome of the chloroplast nucleoid have recently been identified (Majeran et al., 2012; Melonek et al., 2016). SIGF expression in cia2/cil double mutants is much lower than that in cia2 single mutants but slightly lower than that in cil single mutants (Table 1; Supplemental Table S2). In addition, the expression of
PLASTID-TRANSCRIPTIONALLY ACTIVE 7 (pTAC7/PAP12), pTAC13, pTAC14/PAP7, and PLASTIDIAL THIOREDOXIN Z (TrxZ/PAP10) is much lower in cia2/cil double mutants than in cia2 single mutants but slightly down-regulation of that in cil single mutants. These results indicate loss of CIA2 and CIL functions affect plastid transcriptional genes, and these genes are primarily regulated by CIA2 and supplemented by CIL.
Conversely, mitochondrial-related, nuclear transcription-related, and chloroplast
development-related genes constituted 48.4% (134/277), 26.0% (72), and 17.3% (48/277) of the
18
up-regulated genes in cia2/cil double mutants, respectively (Figure 6E and F; Supplemental Table S3). In cia2/cil double mutants, the mitochondrial respiration- and oxidative stress-related genes constituted 14.9% (20/134) and 10.4% (14/134) of the mitochondria-related genes, respectively (Figure 6F; Table 2 and 3; Supplemental Table S3). The products of mitochondria-related genes, including TRANSLOCASE AT MITOCHONDRIALE INNER MEMBRANE (Tim) genes (e.g., Tim8, Tim13, and Tim23-3), mitochondrial respiratory chain complex I-related genes [e.g., PROHIBITIN 1 (PHB1), PHB2, and PHB4], and mitochondrial respiratory chain complex II-related genes (e.g., SUCCINATE DEHYDROGENASE 2), are encoded by the nucleus and transported to the
mitochondria. Therefore, in the presence of chloroplast developmental defect and increased cellular oxidative stress, the nuclear-encoded mitochondria-related genes may be up-regulated to increase mitochondrial function for maintaining cell redox hemostasis and generating energy.
On the other hand, functional chlorophyll biosynthesis involves the tetrapyrrole and
methylerythritol 4-phosphate (MEP) pathways (Lange and Ghassemian, 2003). GLKs, including GLK1 and GLK2, are nuclear transcription factors that up-regulate the expression of
photosynthesis-related and chlorophyll biosynthetic genes to promote chloroplast development (Waters et al., 2009). The chlorophyll biosynthetic genes regulated by GLKs are mainly involved in the tetrapyrrole pathway (Lange and Ghassemian, 2003; Waters al., 2009). In cia2/cil double mutants, the expression of GLK1 and its downstream regulated genes, such as GENOMES UNCOUPLED 4 (GUN4), PROTOCHLOROPHYLLIDE OXIDOREDUCTASE A (PORA), and PORB, was significantly decreased, but it remained unchanged in cia2 or cil single mutants (Figure 7; Supplemental Tables S2 and S4). These results indicate that CIA2 and CIL co-regulate the expression of GLK1 and its regulated chlorophyll biosynthetic genes. In addition, CIA2 and CIL regulate other chlorophyll biosynthetic genes, including GLUTAMATE-1-SEMIALDEHYDE AMINOTRANSFERASE 2 (GSA2), COPROPORPHYRINOGEN-III OXIDASE HEMB1 (HEMB1), HEMF1, FLOURESCENT IN BLUE LIGHT (FLU), and 1-DEOXY-D-XYLULOSE 5-PHOSPHATE SYNTHASE (DXS) (Supplemental Tables S2 and S4). GSA2, HEMB1, and HEMF1 are tetrapyrrole pathway enzyme genes, and DXS is an MEP pathway enzyme gene (Lange and Ghassemian, 2003);
19
therefore, CIA2 and CIL likely co-regulate genes involved in all chlorophyll biosynthesis pathways and key regulators in the tetrapyrrole pathway (Figure 7; Supplemental Tables S2 and S4).
Chloroplast developmental defects in cia2/cil double mutants lead to ROS overproduction
In plants, photosynthetic processes usually generate ROS; therefore, the reaction centers of photosystem I (PSI) and PSII in chloroplast thylakoids are the major ROS generation sites. In PSI, photoreduction of oxygen to superoxide radicals (O2
-) and their subsequent disproportionation produces hydrogen peroxide (H2O2) and oxygen is catalyzed by SUPEROXIDE DISMUTASE (SOD). Thereafter, H2O2 is reduced to water, catalyzed by ASCORBATE PEROXIDASE (APX) (Mehler et al, 1951; Asada et al., 1974; Miyake and Asada, 1994; Asada, 2006). In PSII, singlet oxygen (1O2) generates in the reaction center of chlorophyll (Telfer et al., 1994; Hideg et al., 1998).
ROS production levels in plant cells are low under normal conditions, but ROS production is enhanced in the presence of chloroplast dysfunction, and SOD expression increases (Myouga et al., 2008).
Arabidopsis expresses three COPPER/ZINC-SODs (CSD1-3), three FE-SODs (FSD1-3), and one MANGANESE-SOD (MSD1). FSD1 is localized in the stroma of chloroplasts, plasma membrane, and mitochondrial membrane (Kliebenstein et al., 1998; Brugie`re et al., 2004; Marmagne et al., 2004; Peltier et al., 2006). FSD2, FSD3, and CSD2 are attached to the thylakoid membranes, whereas FSD2 and FSD3, also known receptively as PAP9 and PAP4, are evenly colocalized in chloroplast nucleoids (Ogawa et al., 1995; Myouga et al., 2008; Pfalz and Pfannschmidt, 2013).
CSD1, CSD3, and MSD1 are respectively localized in the cytoplasm, peroxisomes, and
mitochondria (Peck, 2005; Baginsky and Gruissem, 2006). FSD3/PAP4 was down-regulated by nearly 0.60-fold in cia2 and cia2/cil mutants, and CSD2 was up-regulated by 1.66-fold in cia2 mutants and by 2.2-fold in cia2/cil double mutants (Table 2; Supplemental Table S2). MSD1 was only up-regulated by approximately 1.74-fold in cia2/cil double mutants (Table 2; Supplemental Table S3). There was no significant difference in the expression levels of other SOD genes between
20
Col and mutant plants. Additionally, APXs showed no change in expression between Col and mutant plants. These results of microarray analysis suggest increased oxidative stress downregulates FSD3 and up-regulates CSD2, MSD1, and mitochondria-related genes.
Histochemical staining was used to detect ROS levels in plants. cia2/cil double mutants, but not cia2 and cil single mutants, showed high O2- accumulation in leaves (Figure 8A). No H2O2
accumulation was detected in both Col and mutant plants (Figure 8B). Therefore, we hypothesize that chloroplast dysfunction increases cellular oxidative stress, resulting in the phenotype of cia2/cil double mutants, was proven.
CIA2 and CIL affect the flowering time of Arabidopsis
The cil and cia2/cil mutants showed obvious delay in flowering under LD (Table 4). In addition, CIA2 and CIL interact with CO (Results in Part II), and CO is involved in the regulation of flowering time (Putterill et al., 1995). In this study, microarray analysis found changed in the expression of several genes related to flowering time regulation in cia2, cil, and cia2/cil mutants, including the positive regulators of flowering such as APETAL 1 (AP1), CAULIFLOWER (CAL), AGAMOUS-LIKE6(AGL6), and REPRESSOR OF UV-B PHOTOMORPHOGENESIS
2/EARILYFLOWERING BY OVEREXPRESSION, among others (Irish and Sussex, 1990; Ma et al., 1991; Kempin et al., 1995; Wang et al., 2011), as well as negative regulators of flowering such as BBX30, BBX31, and BBX32 (Graeff et al., 2016; Tripathi et al., 2017) (Table5; Supplemental Table S2 and S3). In addition, the expression of 13 genes that regulate flowering time was reduced in cia2, cil, or cia2/cil mutants, including positive regulators such as COL5 and LONG VEGETATIVE
PHASE 1, among others (Hassidim et al., 2009; Xu et al., 2012), as well as negative regulators such as FLOWERING LOCUS C and MADS AFFECTING FLOWERING 4, among others (Rouse et al., 2002; Xu et al., 2013). Among these, the expression of AP1, CAL, and AGL6 was increased in cia2 and cia2/cil mutants but decreased in cil mutants. FLC expressing did not show significant change in cia2 single mutants, increased in cil single mutants, and decreased in cia2/cil double mutants, suggesting that CIA2 and CIL are involved in the regulation of flowering time; however, the
21
underlying mechanism remains unclear. Therefore, to confirm whether CIA2 and CIL affect flowering time, Col and mutant (cia2, cil, cia2/cil, and ppi1) strains were used. CIA2 transgenic strains (35Spro:CIA2/cia2 and CIA2pro:CIA2/cia2) were also used to observe flowering time under different daylight conditions. As CIA2 regulates Toc33 expression and affects chloroplast
development, we explored whether chloroplast dysplasia affected flowering time using a ppi1 mutant strain.
Under LD condition, the flowering time of cil and cia2/cil mutants was delayed by 2 days compared with that of other plants, and the average rosette leaf number of cil mutants was greater (by 3) than that of Col plants (Table 4). Under MD condition, cia2, cil, cia2/cil, and ppi1 mutants flowered 7 to 15 days later but 35Spro:CIA2/cia2 and CIA2pro:CIA2/cia2 mutants flowered 10 to 13 days earlier than Col plants (Table 4). Under SD condition, cia2, cil, cia2/cil, and ppi1 mutants flowered 4 to 44 days later but 35Spro:CIA2/cia2 and CIA2pro:CIA2/cia2 mutants flowered 7 to 13 days earlier than Col plants (Table 4). Based on these results, under MD and SD conditions, cia2, cil, and cia2/cil mutants showed delayed flowering, whereas 35Spro:CIA2/cia2 and CIA2pro:
CIA2/cia2 mutants showed early flowering. Although the flowering time of 35Spro:CIL/cil plants were not determined, CIA2 and CIL might be involved in the regulation of flowering time given that these two genes are complementary. This would be further verified by expressing the CIL protein in transgenic plants. In addition, the Y2H experiments showed that CIA2 and CO proteins interaction with each other; therefore, CO interaction likely regulates the effects of CIA2 and CIL on flowering time. This would be further verified by crosses of cia2 and cil mutants with co mutants to establish double and triple mutants.