Disruption of the Homogentisate Solanesyltransferase Gene Results in Albino and Dwarf Phenotypes and Root, Trichome and Stomata Defects in Arabidopsis thaliana

Homogentisate solanesyltransferase (HST) plays an important role in plastoquinone (PQ) biosynthesis and acts as the electron acceptor in the carotenoids and abscisic acid (ABA) biosynthesis pathways. We isolated and identified a T-DNA insertion mutant of the HST gene that displayed the albino and dwarf phenotypes. PCR analyses and functional complementation also confirmed that the mutant phenotypes were caused by disruption of the HST gene. The mutants also had some developmental defects, including trichome development and stomata closure defects. Chloroplast development was also arrested and chlorophyll (Chl) was almost absent. Developmental defects in the chloroplasts were consistent with the SDS-PAGE result and the RNAi transgenic phenotype. Exogenous gibberellin (GA) could partially rescue the dwarf phenotype and the root development defects and exogenous ABA could rescue the stomata closure defects. Further analysis showed that ABA and GA levels were both very low in the pds2-1 mutants, which suggested that biosynthesis inhibition by GAs and ABA contributed to the pds2-1 mutants' phenotypes. An early flowering phenotype was found in pds2-1 mutants, which showed that disruption of the HST gene promoted flowering by partially regulating plant hormones. RNA-sequencing showed that disruption of the HST gene resulted in expression changes to many of the genes involved in flowering time regulation and in the biosynthesis of PQ, Chl, GAs, ABA and carotenoids. These results suggest that HST is essential for chloroplast development, hormone biosynthesis, pigment accumulation and plant development.


Introduction
HST is an important enzyme that catalyzes the condensation of the tyrosine-derived aromatic compound, homogentisate (HGA), with the isoprenoid, all-trans-nonaphenyl diphosphate solanesyl diphosphate, to form 2-Me-6-solanesyl-1,4-benzoquinol (also known as 2-demethylplastoquinol-9) [1]. This branch-point compound directly leads to the biosynthesis of either vitamin E or the photosystem II (PSII) mobile electron transport co-factor, PQ [2,3]. Indirectly, it also links a number of other diverse, important metabolic pathways, including carotenoid biosynthesis, which uses PQ as a co-factor for phytoene desaturase, and ABA biosynthesis, which is derived from the breakdown of carotenoids [4][5][6]. Consequently, HST is likely to affect plant growth and development. While the homogentisate and solanesyl pathways are critical for forming prenylated electron carrier molecules, such as PQ in chloroplasts and ubiquinone in mitochondria (Figure 1, modified from Motohashi et al.) [7], the C 20 diterpene geranylgeranyldiphospate (GGPP) stands at the crux of a number of other biosynthetic pathways important to plant survival and adaptation. GGPP is also used as a substrate in the formation of dolichol (required for protein glycosylation) and phytol side chains during Chl biosynthesis. Changes in fluxes through many of these different pathways could influence other pathways through feedback regulation, changes to higher order regulatory genes and changes to the balance between substrate ''draw'', which depends on tissue-specific and organ-specific gene expressions, and expression timing. The feedback regulation balances between substrate ''draw'' and expression timing are known for GA biosynthesis, the 2-C-methyl-derythritol-4-phosphate (MEP) pathway and for carotenoid biosynthesis [8,9]. PQ transports electrons from PSII to cytochrome b 6 f in the light harvesting reactions of photosynthesis [1,10] and is the immediate electron acceptor (co-factor) in the -carotene to lycopene stage during carotenoid and ABA biosynthesis and in the desaturation of phytoene and f-carotene ( Figure 1). A certain quinone/hydroquinone balance is necessary for optimal phytoene desaturation. PQ is enriched in chloroplasts and is also found in Golgi membranes and in minor quantities in microsomes [11]. PQ is not detectable in mitochondria; instead the mobile quinone electron carrier, ubiquinone, dominates in mitochondria. The presence of PQ in the cytosol has also been reported, which probably reflects its transport from the site of synthesis to its final location [11,12].
In higher plants, GAs can regulate seed germination, stem elongation, leaf expansion, trichome development and plant fertility (stamen elongation, pollen release and germination and pollen tube growth) [8,[13][14][15], GAs are mainly synthesized by the MEP/isoprenoid pathway and this biosynthesis can be divided into three stages [16][17][18]. The GA1 and GA2 genes play an important role in the conversion of GGPP to ent-kaurene and the Arabidopsis ga1 mutant is a GA-responsive male-sterile dwarf due to the disruption of GA biosynthesis. Accumulation of the 1st stage enzyme, cyclase ent-copalyl diphosphate synthase (CPS, also known as GA1), in Escherichia coli, the conversion of GGPP to copalyl diphosphate (CDP) in GA1 + /GGPP synthase + transgenic E. coli and the transgenic complementation of Arabidopsis ga1-3 mutants, confirmed that the GA1 gene locus, U11034, encodes CPS [19]. The Arabidopsis ga2 mutant is also a GA-deficient dwarf because it contains mutated ent-kaurene synthase (KS) and has impaired CDP conversion to ent-kaurene [20].
Genes encoding HST or its homologs have been isolated and identified in Arabidopsis and many other plants [2,4,[21][22][23]. The HST gene, VTE2, was first isolated from Glycine max and the constitutive expression of VTE2-2 in transgenic Arabidopsis resulted in a significant 13% tocopherol increase compared to the transgenic vector control plants [22]. Previous studies have explored the function of the Arabidopsis HST by utilizing pds2 mutants and transgenic technology [2,4,21,22]. Enzyme assay results for cell expression of the HST gene in E. coli suggested that HST catalyzes the first step in PQ biosynthesis [2]. Overexpression of the HST gene improved prenyl lipid, PQ and tocopherol levels in transgenic Arabidopsis [2,22]. Disruption of the HST gene produced an albino phenotype and caused a deficiency in the synthesis of PQ and tocopherol by affecting the prenyl/phytyl transferase enzyme [4]. The in-frame 6 bp deletion in the coding region of the HST gene caused an albino phenotype and the expression of the HST gene in this pds2 mutant turned albino plants into green plants [21].
Recently, we screened and found a new recessive albino knockdown mutant called pds2-1. TAIL-PCR and DNA sequencing showed that the mutant phenotype was caused by a T-DNA insertion in the HST gene. To date, phenotypic analysis of HST has been restricted to an albino phenotype, which results from disruption to chloroplast development and photosynthetic pigment biosynthesis. However, in addition to pigment biosynthesis, HST indirectly regulates carotenoid and ABA biosynthesis, but analyses of the developmental and the physiological changes due to HST gene blockage have not been reported. Detailed analysis of pds2-1 indicated that HST knockdown impaired carotenoid and ABA biosynthesis, as well as GA biosynthesis and auxin content, which resulted in severe developmental defects. However, the application of exogenous GA 3 and ABA to pds2-1 partially restored the wildtype phenotype. Our new mutant line and analytical data confirm that HST is essential for carotenoid, ABA, and GA biosynthesis and that it plays a critical role in plant growth and development.

Plant Materials and Growth Conditions
Arabidopsis thaliana (ecotype Columbia-0) plants were grown in soil or on K strength Murashige and Skoog (MS) plates (pH 5.7) containing 0.8% agar and 3% sucrose at 22uC and 65% humidity under a 16 h light (100 mmol m 22 s 21 )/8 h dark cycle. Insertion mutants were obtained from a transgenic experiment via the transformation of Arabidopsis by the floral dip method [24] using Agrobacterium tumefaciens GV3101 that contained pART27 [25]. Self pollinated seed recovered from 16 independent transgenic events was screened on K strength MS plates containing 50 mg L 21 of kanamycin. Then the surviving seedlings were transferred to soil to generate seeds at 22uC under long-day conditions. For PCR, RT-PCR, Enzyme-linked immunosorbent assays (ELISA), RNAsequencing and microscope analyses, the seedlings were grown on KMS plates containing 3% sucrose.

TAIL-PCR
The CTAB method was used to isolate genomic DNA from WT seedlings and heterozygous Arabidopsis. Genomic DNA was used as a template for the TAIL-PCR, which was carried out using a Genome Walking Kit (TAKARA) according to the manufacturer's instructions. The three gene-specific primers used in TAIL-PCR were: R1 (59-GTGCTGCAAGGCGATTAAGTTGGGTAA-39), R2 (59-ATTGCGTTGCGCTCACTGCCCGCTTTC-39) and R3 . The PCR products were sequenced using R3 as the sequencing primer. The sequencing results were searched against the Arabidopsis genome sequence database (GenBank) using BLAST to localize the position of the T-DNA insertion.

Co-segregation Analysis
Three primers were designed and used to analyze the segregation pattern of the T-DNA insertion into the HST gene in seedlings that showed a mutant phenotype. The analysis was performed on DNA extracted from plates of WT and heterozygous mutant seedlings and the primer sequences were: hst-g1 (59-CGAAAGTAAGCAGAGCAAAGAGT-39), hst-g2 (59-CATTC-CCACAAATAAGAGCAAGA-39) and R3 (59-GTGGCTCCT-TCAATCGTTGCGGTTCTG-39). The PCR products were identified using 1% agar gel electrophoresis.

Cloning of the HST Gene and Binary Vector Construction
Based on the HST cDNA (Accession No.: NM_001161137.1) sequence, a pair of primers: hst-1 (59-AAACGAAAGTAAGCA-GAGCAAAG-39) and hst-2 (59-CTAGAGGA-AGGGGAATAA-CAGATAC-39), were designed to obtain the DNA sequence that included the coding region of the HST gene. Total RNA isolated from 3-week-old Arabidopsis was used for the reverse transcription-PCR. The PCR product was cloned into the T vector (TAKARA) and confirmed by DNA sequencing. The vector containing the HST full-length coding region was used as a template for binary vector construction, together with a pair of primers: HST-f (59-cccgggccATGGAGCTCTCGATCTCA-CAATC-39; where ''cccggg'' represents a Sma I restriction site and ''ccATGG'' represents a Nco I site) and HST-r (59-tcta-gaactagtTTAGAGGAAGGGGAATAACAGATACT-39; where ''tctaga'' represents a Spe I site and ''actagt'' represents a Xba I site). The PCR product was digested with Nco I and Spe I and cloned into the Nco I-Spe I site of the binary pCAMBIA1302 vector, creating 35S pro ::HST, in which the HST gene was expressed from the 35S promoter. The HST antisense plasmid, Anti-HST, was created by placing the Xba I-Sma I fragment of the PCR product between the 35S promoter and the NOS terminator of the binary pBI121 vector in an antisense orientation. The primers used for subcloning the HST gene promoter included Pro-HST-f  and Pro-HST-r (59-tctagaTTTGTGTCCAATCCTCTTTCC-GG-39). The PCR products were subcloned into the Cla I-Xba I site of pBI121, creating HST pro ::GUS, in which the GUS gene is expressed driven by the HST promoter.

RT-PCR and Quantitative Real-Time RT-PCR (qRT-PCR)
For RT-PCR, WT and albino Arabidopsis plant leaves were harvested after 3 weeks growth on agar medium containing 3% sucrose. Total RNA was extracted using Trizol Reagent (Biomed). The cDNA was synthesized with the reverse transcriptase (TaKaRa) and Oligo d(T) primer, and PCR was performed using the hst-1 and hst-2 primers. UBQ10 (Access No.: NM_116771.5) was used as the inner control gene along with two primers: UBQ-f (59-GCCGGAAAACAATTGGAGGATGGT-39) and UBQ-r (59-CATTAGAAAGAAAGAGATAACAGG-39). RT-PCR images were captured using a Gel Image Analysis System.
qRT-PCR analysis was used for the expression analyses and was performed on 7-week-old WT plants at different developmental stages and on a variety of tissues. It was also performed on 7-weekold WT or pds2-1 mutant plants for gene expression analyses. The system used was an ABI 7500 system with SYBR green detection. For analyses of HST expression in different tissues and developmental stages of WT Arabidopsis, UBQ10 was used as an internal control. For analyses of gene expression in pds2-1 mutants, TUB2 was used as an internal control. The two-step thermal cycling profile used was 15 s at 95uC and 1 min at 68uC. The qRT-PCR reactions were performed in biological triplicates using total RNA samples extracted from three independent plant materials grown under identical growth conditions. The comparative DDthreshold cycle method [27] was used to evaluate the relative quantities of each amplified product in the samples. All primers used in the qRT-PCR analyses are reported in Supplementary Table S1.

RNA Sequencing
The whole plants of 3-week-old pds2-1 and WT Arabidopsis were harvested and used for RNA extraction. The experiment was then repeated following the same collection scheme, thus providing two distinct biological replicates. Total RNA was extracted using TRIzol reagent following the manufacturer's instructions and was treated with RNase-free DNase I (NEB) to remove contaminated genomic DNA. The mRNA was isolated from the total RNA using Dynabeads oligo(dT) (Invitrogen). An illumina library was constructed using the NEB mRNA library prep kit instruction manual. The cDNA library was sequenced for paired ends on the Illumina Hiseq2000 platform at the Beijing Center for Physical and Chemical Analysis (Beijing, China). The raw reads were filtered by removing adaptor sequences, empty reads and low quality reads containing more than 50% bases with Q,30 using FASTX-Toolkit with methods described previously [28][29][30]. The resulting high-quality reads were mapped onto the Arabidopsis thaliana reference genome (TAIR 10) using Tophat (v2.0.5) [31]. Gene expression levels were measured as FPKM (fragments per kilobase of exon model per million mapped reads). EdgeR outputs the T-statistic and the p-values for each gene. Differential expression was estimated and tested using the edgeR software package (R version: 2.14, edgeR version: 2.3.52) [32]. We then calculated the FDR, estimated fold change (FC) and the FC log2 values. Transcripts that had an FDR#0.05 and an estimated absolute log2 (FC)$1 were determined to be significantly differentially expressed. Gene Ontology (GO), SEA and PAGE tools, found at the agriGO website (http://bioinfo.cau.edu.cn/ agriGO/), were used to analyze the differential expression genes during the functional annotation and enrichment analysis.

Chl and Carotenoid Assays
Chl and carotenoids were extracted from the aboveground parts of 40-day-old Arabidopsis, as described previously [33]. Absorbance (A) of the final solution at 663, 647 and 470 nm was measured and the concentrations of Chl a (Ca), Chl b (Cb) and carotenoids (Cc) were calculated as follows [34]:

ELISA Analyses
Initially, four plant hormones: ABA, GA 3 , IAA and ZR, were measured on the aerial parts of 3-week-old WT and pds2-1 plants by a previously described ELISA method [35].

b-Glucuronidase Staining
Histochemical b-Glucuronidase staining in HST pro ::GUS transgenic plants was performed as described previously [36].

Transmission Electron Microscope (TEM) and scanning electron microscope (SEM) Assays
Leaves harvested from 3-week-old WT and pds2-1 plants were soaked in fixation buffer (4% glutaraldehyde in 100 mM cacodylate buffer, pH 7.2) and post-fixed for 6 h in secondary fixation buffer (1% OsO 4 in 100 mM cacodylate buffer, pH 7.4) at 4uC. The fixed samples were dehydrated, embedded in Spurr resin and cut into ultrathin sections. The specimens were subsequently stained with uranyl acetate and observed with a TEM (Hitachi).
For SEM analysis, leaves of 3-week-old WT and pds2-1 plants were harvested and the analysis was performed as described previously [37]. The specimens were examined by a SEM (Hitachi).

Phenotype Analysis of the Albino Mutant
Arabidopsis thaliana was transformed by the binary vector, pART27, so that a small number (16) of T-DNA insertion lines could be generated. Seedlings from self-pollinated plants, selected on K-strength Murashige and Skoog plates containing kanamycin and sucrose, were mainly green, but one plate contained several albino plants. Progeny from heterozygous albino plants were mainly green, but progeny derived from self-pollinated heterozygotes segregated 3:1 into green and albino colored siliques ( Figure  S1). The cotyledons of all the young homozygous progeny from the albino type plants were purple on the selection medium (Figure 2 A), which suggested that Chl may be low or absent and that anthocyanin pigments may have accumulated. Most homozygous albino mutants gradually faded to white with continued growth on the selection medium. A small number of plants also showed the purple coloration on the rest of their plant parts (Figure 2 A-C). All albino seedlings died when they were grown on soil or K strength MS medium without sucrose, but in K MS medium with 3% sucrose, albino seedlings produced albino leaves (and even translucent stem tissues) and only one main shoot over their whole life cycle (Figure 2 B,E,F). Low resolution microscopy indicated that the albino plants had shorter roots, fewer root hairs, fewer and smaller leaves with shorter petioles and a reduced trichome density (Figure 2 B-F; Figure S2; Figure S3). Microscope or SEM was used to investigate the leaf surface morphology (trichomes and stomata) of pds2-1 and WT plants in greater detail. The results showed that pds2-1 trichomes were shorter in length and most (,80%) trichomes had two branches instead of three (Figure 3 A,B; Figure S3; Figure S4). Their stomata had larger openings with unusual swollen structures surrounding the opening (Figure 3 C-F), which suggested that pds2-1 stomata may close abnormally.

Albino and Dwarf Phenotypes are Caused by a T-DNA Insertion into HST
We hypothesized that the albino pds2-1 phenotype was caused by a T-DNA insertion because the albino plants were obtained from transgenic Arabidopsis T-DNA insertion lines and the green and albino plants of T 1 , T 2 and T 3 segregated in a 3:1 ratio on nonselective medium. To determine the insertion site within the Arabidopsis genome, DNA was isolated from T 3 heterozygous plants and non-transformed WT plants (as a negative control) for the TAIL-PCR ( Figure S5). The TAIL-PCR and sequencing showed that the insertion was located at Chr3:3782298, which is within the second intron of the Arabidopsis HST gene (At3g11945) (Figure 2 H).
To check whether the T-DNA insertion was responsible for the albino phenotype, DNA was isolated from WT, heterozygous (green) and homozygous (white) Arabidopsis plants for PCR analysis. Primers, hst-g1 and hst-g2 (Table S1) amplified a fulllength 969 bp fragment of HST gDNA from the WT and heterozygous plants. In contrast, primers R3 and hst-g2 amplified a 466 bp fragment from the heterozygous and homozygous pds2-1 plants. These results indicated that the T-DNA insertion in the HST gene co-segregated with the albino phenotype ( Figure 2J). To analyze the transcription of HST in the heterozygous and homozygous pds2-1 plants, RT-PCR was conducted using gene specific primers: hst-1 and hst-2. The results showed that HST expression was eliminated in the homozygous plants (Figure 2 J), which suggested that albino plants were HST null mutants and that the mutant phenotype was caused by a T-DNA insertion into the HST gene.

The HST Gene is Highly Expressed in Green Tissues
To understand the breadth of the role played by the HST gene, we re-examined the expression pattern of HST in WT Arabidopsis using qRT-PCR. The qRT-PCR assays revealed that HST was expressed at high levels in stems and leaves, but at relatively lower levels in the flowers and roots of adult plants (Figure 4 A). Furthermore, different HST expression levels were detected in the leaves at different developmental stages. The HST transcript levels were highest in non-senescent leaves, but they gradually decreased as the leaves began to senesce (Figure 4 B).
To further examine the HST expression patterns in different plant tissues, a 813 bp HST promoter sequence, upstream of the transcription start site, was fused to a GUS-coding sequence and this led to the appearance of HST pro ::GUS transgenic Arabidopsis plants. In 40-day-old seedlings, GUS activity was highly detected in green tissues, such as leaves and stems (Figure 4 C-J). Notably, GUS activity was detected at a high level in adult green leaves, but at decreased levels in senescent leaves (Figure 4 J). These results suggested that HST was highly expressed in green tissues and had an important role to play in their development. Total Chl and carotenoid (Car) contents were examined in pds2-1 and WT plants. The calculated results showed that Chl and Car were almost absent in the homozygous pds2-1 mutant plants. Chl a decreased to 0.55% of the WT level, Chl b to 1.86% and total Chl to 0.93% (Table 1), while carotenoids were reduced to 0.24% of the WT level. The Chl a/b ratio also decreased dramatically in pds2-1 mutants and was only 26.90% of the WT level. As a result,  the Chl/Car ratio increased to 20.79 in the pds2-1 mutants (Table 1). These results showed that disruption of the HST gene caused significant reductions in Chl and carotenoid levels and that carotenoids were affected more seriously than Chl in pds2-1 mutants.
To assess the effect of pds2-1 mutations on chloroplast development, plastids from the first leaves of 30-day-old seedlings were examined by TEM. The data showed that the chloroplasts in pds2-1 were smaller than those in the wild type and lacked starch granules, oil drops and thylakoids, but contained many densely stained globule structures that resembled plastoglobules (Figure 3 G-J). These results suggested that proplastids in the mesophyll cells of the pds2-1 mutant have lost the ability to develop into mature chloroplasts. The RuBisCO large subunit levels were also highly reduced in pds2-1 when total protein was analyzed by SDS-PAGE ( Figure S6) and this was consistent with the albino phenotype and the chloroplast development defects in pds2-1. These data suggested that HST played an important role in proplastid growth and thylakoid membrane formation.
An anti-sense HST-RNAi vector was used to transform Arabidopsis WT plants and six transgenic Arabidopsis lines were analyzed for phenotype and chloroplast defects (Figure S7  The HST Gene Rescues the pds2-1 Mutant Phenotype HST cDNA, driven by the 35S promoter, was inserted into the genome of the pds2-1 heterozygous mutant. Transgenic seeds were screened on KMS plates containing 3% sucrose and 20 mg L 21 hygromycin and six independent lines were selected. Homozygous pds2-1 plants with 35S pro ::HST were confirmed by PCR ( Figure 5 A-C). T 2 progeny of pds2-1 that were homozygous for 35S pro ::HST displayed a phenotype that was similar to the WT plants ( Figure 5  D). These results showed that the HST gene driven by the 35S promoter could completely rescue the mutant phenotype.
Severe Declines in GA and ABA Content Were Seen in pds2-1 Leaves ELISA were performed to determine the concentrations of ABA, GA 3 , zeatin riboside (ZR) and IAA in the aerial parts of 3week-old WT and pds2-1 plants. A substantial decrease in ABA (43.1%), GA 3 (61.5%) and ZR (69.8%) concentrations occurred in pds2-1 plants compared to WT plants (Table 2) but the ratio of GA 3 :ABA was increased from 0.0621 in WT to 0.0923 in pds2-1 ( Table 2). This disruption of the HST gene also led to a substantial increase (by 41.6%) in the IAA content of pds2-1.
Gene Transcriptions That Specified Carotenoids, GA, ABA, Trichomes and Roots Were Depressed in pds2-1 Previous reports indicated that albino mutants blocked the MEP pathway and led to a decrease in carotenoid, GA and ABA biosynthesis [33]. Therefore, we investigated whether disruption of the HST gene changed the expression of genes involved in these pathways using qRT-PCR. DXS, DXR, LYC, IM, PSY and ZDS are involved in the MEP pathway and in the biosynthesis of carotenoids [38][39][40]. GGPS encodes an enzyme involved in isoprenoid biosynthesis [41] and GGRS encodes a protein with geranylgeranyl reductase activity [42]. The PDS1 gene (encoding the p-hydroxyphenylpyruvate) plays an important role in the synthesis of both plastoquinone and tocopherols in plants [43], GA1, GA2 and GA3 are involved in GA biosynthesis [19,20,44]; ABA1 plays an important role in ABA accumulation [45] and GL2, WAVE1 and WAVE2 are involved in trichome and root hair initiation [46]. The results showed that transcription of all these selected genes was significantly lower in pds2-1 plants compared to WT plants and the key gene in ABA biosynthesis, ABA1, was almost absent in the pds2-1 plants ( Figure 6).

A Mutation in HST Promotes Early Flowering in pds2-1
Although the growth of homozygous mutants was severely retarded, the pds2-1 albino plants could bolt when grown on KMS containing 3% sucrose (Figure 7 A). All the homozygous plants had flower-like structures that never matured into normal flowers (Figure 7 A,B) and the flowers and stems were usually glabrous (Figure 2 D). Therefore, psd2-1 had to be propagated in the heterozygous state, which suggested that the gene was essential for plant development. Moreover, flowering time was disrupted in pds2-1 plants and they flowered earlier than the WT plants (Figure 7 A,C). This conclusion is supported by the light greencolored anti-sense HST-RNAi line, anti-L2, which also displayed an early flowering phenotype (Figure 7 D). Transcripts for two flowering genes, FLC and GI, were also significantly depressed in pds2-1 plants, but CO, SOC1 and FT were unaffected by the mutation (Figure 7 E).

HST Disruption Affects the Expression of Many Photosynthetic Genes, Cellular Components and Biological Processes
RNA sequencing was carried out to further explore the effects of the HST gene on metabolic pathways. A total of 1254 unique genes exhibited two fold or greater differential expression in pds2-1 plants compared to WT plants, with 351 genes up-regulated and 903 genes down-regulated. A total of 222 differentially expressed genes were then identified with a false discovery rate (FDR) of less than 0.05 (Table S2). Gene Ontology (GO) was used to analyze the differentially expressed genes by functional annotation and enrichment analysis, which required the use of singular enrichment analysis (SEA) and the Parametric Analysis of Gene set Enrichment tool (PAGE) [47]. In total, 212 of the 222 differentially expressed genes were assigned to at least one term in GO under the different biological process categories (molecular function, subcellular structure/components and biological process) and the number of significant GO terms were 138. In pds2-1 mutants, the expression levels of many genes related to cellular structure/components, including thylakoids, plastids, chloroplasts, photosynthetic membrane, plastoquinone, plastoglobules, lightharvesting complexes, organelles, the cytoplasm and the apoplast, were dramatically reduced (Table S3). Moreover, the expression levels of molecular components involved in photosynthesis, the light reaction, pigment metabolic processes and processes such as the regulation of catalytic activity, response to stimuli, and response to stress, had also decreased significantly. When differentially expressed genes involved in cellular components were submitted to the PAGE tool after RNA sequencing, 22 GO terms (including thylakoid, chloroplast, photosynthetic membrane, etc.) were significantly over-represented in the hierarchical tree ( Figure 8; Table S4).

GA 3 and ABA partially rescued the pds2-1 mutant phenotype
The 14-day-old pds2-1 seedlings were transferred onto KMS medium containing 3% sucrose and 10 mM GA 3 . Two weeks later, the leaf petioles and roots of all the tested pds2-1 plants were longer than those grown without GA 3 and root hair density had increased ( Figure 9 A-D, G). The pistils of pds2-1 mutants grew longer with GA 3 supplementation, but flower buds did not develop into mature flowers (Figure 9 E). Stem trichomes were obvious on pds2-1 mutant plants when treated with GA 3 (Figure 9 F). The pds2-1 plants treated with GA 3 also had extended petioles, were larger and grew faster than those without GA 3 and they produced more branches and inflorescences (Figure 9 G,H).
ABA plays an important role in regulating stomata closure in plants. To find out whether the stomata structural defect in pds2-1 plants was caused by ABA absence, 14-day-old pds2-1 seedlings were transferred onto K MS medium supplemented with ABA. Two days later, the leaves of the mutant plants were harvested and the stomata structure observed using light microscopy. The stomata closed normally in pds2-1 mutants with exogenous ABA (Figure 9 J-K), which suggested that the stomata closure defect was caused by an ABA deficiency in pds2-1 mutants.

Discussion
In this study, we identified an albino mutant, pds2-1, and showed that disruption to the HST gene resulted in an albino phenotype and developmental defects. Homozygous pds2-1 mutants all died when they were grown on soil or KMS medium without sucrose, but produced albino leaves or white inflorescences when 3% sucrose was included. Other Arabidopsis mutants exhibited a similar albino phenotype when DXR, PDS or the IspH homologue of an isoprene biosynthetic gene were disrupted [9,33,48,49].
RT-PCR and GUS staining showed that HST was expressed in almost all the plant tissues and at higher levels in green tissues. HST protein contains a putative 69 aa chloroplast transit peptide at its N-terminus. Subcellular localization analysis of HST in a  previous report showed that HST was targeted at chloroplasts [21]. In this study, TEM analysis demonstrated that pds2-1 chloroplasts were smaller and developmentally malformed. Our study confirmed previous reports, which showed that HST disruption resulted in an albino phenotype in Arabidopsis [4,21]. We measured Chl levels to confirm the negative effect of a mutated HST gene on the photosystem and found that the pds2-1 mutant had significantly reduced Chl levels, which was consistent with the reduced expression we measured for key genes involved in chloroplast biosynthesis. Moreover, SDS-PAGE analysis of the total proteins in pds2-1 plants also confirmed that the photosystem was destroyed and that RuBisCo was not produced after HST disruption. Transgenic mutants transformed with 35S pro ::HST restored the normal green phenotype, but RNAi lines had yellowish-green leaves. Together, these results confirmed that HST was essential for normal plant growth and chloroplast development.
Carotenoids play important roles in many physiological processes and in chloroplast biogenesis. They are also precursors of ABA. Biochemical analysis of the pds2-1 mutant plants revealed that carotenoids, ABA and PQ levels were dramatically reduced in plant leaves. HST is a critical enzyme in the PQ biosynthesis pathway. PQ is both a PSII electron carrier and a critical component that links carotenoid and ABA biosynthesis [4][5][6] with PS I, PS II and ATP synthesis [50]. Hence, the absence of PQ in pds2-1 caused disruption to electron transport, which may have contributed indirectly to the developmental defects in chloroplasts via negative feedback during phytoene desaturation, carotenoid synthesis and chloroplast biogenesis. Inhibition of PQ biosynthesis, leading to phytoene accumulation, has also been reported in previously identified pds2 mutants [4]. Our pds2-1 mutant plants also displayed stomata closure defects and an application of exogenous ABA restored the wild type plant stomata. These results suggested that HST also had an indirect impact on stomata closure by affecting carotenoid and ABA biosynthesis through the regulation of PQ levels.
The pds2-1 mutants showed a pronounced dwarf phenotype and trichome and root hair development defects. qRT-PCR analyses showed that three positive regulators (GL2, WAVE1 and WAVE4) of trichome or root hair development were down-regulated in pds2-1 mutants. ELISA analyses and qRT-PCR revealed that GA levels and the expression of GA biosynthesis genes, such as GA1, GA2 and GA3, declined dramatically in pds2-1 mutants. When exogenous of GA 3 was added to the KMS plates, the pds2-1 plants grew larger and had more root hairs and branches. GAs play important roles in plant height, trichome initiation and root development [51][52][53][54][55][56]. In plastids, GGPP is a precursor of carotenoids, ABA, tocopherols, the phytol tail of Chl, plastoquinones and GA. Qin et al. (2007) showed that disruption of the phytoene desaturase gene, PDS3, in carotenoid biosynthesis impeded the biosynthesis of GAs [9] and a negative feedback mechanism controlled upstream genes in the pds3 carotenoid mutant, which resulted in a decrease in GAs. Our novel GA findings, together with supporting literature and the other information, strongly suggest that HST disruption in pds2-1 affects many pathways through direct and indirect mechanisms.
GAs play an important role in the positive regulation of flowering time [57,58] and reduced gibberellin synthesis usually promotes late flowering phenotypes in Arabidopsis [59,60]. Hence, pds2-1 mutants should, in principle, exhibit late flowering because of the decrease in GA 3 . Instead, an unusually early flowering phenotype occurred in pds2-1 and in the anti-HST plant. This early flowering phenotype was consistent with the decrease in the expression of a key flowering repressor, FLC, and an increase in the expression of the flowering regulator, GI, in pds2-1 mutants. One interpretation of this data arises from the fact that ABA and Figure 6. qRT-PCR analyses of genes involved in the carotenoid, GA and ABA biosynthetic pathways and in the regulation of trichome and root hair development. Significant differences of the means (6 SD) between WT plants and pds2-1 plants (n = 3) are indicated by * (P,0.05) and ** (P,0.01). The Arabidopsis TUB2 gene was used as an internal control. doi:10.1371/journal.pone.0094031.g006 GA levels decreased in pds2-1 mutants. ABA is believed to act as an antagonist to GAs during plant growth and development [61,62]. ABA biosynthesis mutants have an early flowering phenotype [61,63] and ABA can delay flowering onset by upregulating the expression of FLC, as occurs in Arabidopsis [64]. In fact, although the levels of both hormones fell in pds2-1 seedlings, the overall ratio of GAs:ABA increased by 48.6% (Table 2). Genetic analyses of the interaction between GAs and ABA shows that GAs play a major rate-limiting role in floral promotion [61]. Our findings, together with an earlier report, suggested that the increased ratio of GA:ABA in pds2-1 mutants had maintained the predominant role played by GA and caused the triggering of the early transition from vegetative to reproductive growth. A second interpretation is that HST disruption affected other hormone levels. Earlier reports showed that a number of hormones, including GAs, ABA and auxins, played important roles in regulating flowering time [65]. For example, exogenous IAA was able to induce flowering in long day plants [66]. Our ELISA results showed that auxin levels (IAA) increased by 41.6% in the pds2-1 mutant compared to the WT plants. Hence, the early flowering phenotype displayed by pds2-1 mutant plants could also be due to their increased IAA levels. In summary, disruption of the HST gene promoted flowering by affecting the levels of several plant hormones.
Screening of a small Arabidopsis T-DNA population uncovered a novel albino mutant pds2-1 with a T-DNA insertion in the HST gene. This gene is known to play an important role in PQ biosynthesis. The pds2-1 mutation resulted in a typical carotenoiddeficient phenotype, with reduced PQ, abnormal chloroplast development, reduced photo-protection and reduced PS II Figure 8. Hierarchical tree of over-represented GO terms (subcellular structure category) generated by parametric analysis of gene set enrichment. Boxes in the graph represent GO terms as labeled by their GO ID, term definition and statistical information. Significantly overrepresented terms (adjusted P,0.05) are marked with color, while non-significant terms are shown as white boxes. A box's degree of color saturation is correlated to the representation level of the term. Boxes connected by dotted, dashed and solid lines represent zero, one and two down-regulation terms at the ends connected by the line, respectively. doi:10.1371/journal.pone.0094031.g008 activity. These findings confirmed that the HST gene plays an important role in pro-plastid growth and thylakoid membrane formation. However, the mutation also produced novel IAAenhanced, ABA-deficient and GA-deficient phenotypes that had not been reported previously. Furthermore, GA 3 and ABA application partially rescued the mutant phenotype. Other novel phenotypes not previously reported for this gene include short roots and petioles, fewer leaf numbers, root hairs and trichomes, more swollen stomata, an early flowering date and high expression of HST gene in green tissues. Moreover, qRT-PCR and RNA sequencing confirmed that the mutant had a blocked MEP pathway and the genes that played a role in GA biosynthesis did not function. This detailed analysis of the pds2-1 mutant paves the way for a comprehensive study of the physiological regulation of carotenoids, chloroplast components and other physiological processes by major plant hormones through the use of a combination of genetic, biochemical and molecular approaches. HST will probably lie at the heart of these processes as it appears to impact directly and indirectly on so many hormones and biological processes.