The Reduced Plastid-Encoded Polymerase-Dependent Plastid Gene Expression Leads to the Delayed Greening of the Arabidopsis fln2 Mutant

In Arabidopsis leaf coloration mutants, the delayed greening phenomenon is common. Nonetheless, the mechanism remains largely elusive. Here, a delayed greening mutant fln2–4 of FLN2 (Fructokinase-Like Protein2) was studied. FLN2 is one component of Transcriptionally Active Chromosome (TAC) complex which is thought to contain the complete plastid-encoded polymerase (PEP). fln2–4 displayed albino phenotype on medium without sucrose. The PEP-dependent plastid gene expression and chloroplast development were inhibited in fln2–4. Besides interacting with thioredoxin z (TRX z), we identified that FLN2 interacted with another two members of TAC complex in yeast including its homologous protein FLN1 (Fructokinase-Like Protein1) and pTAC5. This indicates that FLN2 functions in regulation of PEP activity associated with these TAC components. fln2–4 exhibited delayed greening on sucrose-containing medium. Comparison of the PEP-dependent gene expression among two complete albino mutants (trx z and ptac14), two yellow mutants (ecb2–2 and ys1) and the fln2–4 showed that fln2–4 remains partial PEP activity. FLN2 and FLN1 are the target proteins of TRX z involved in affecting the PEP activity. Together with the data that FLN1 could interact with itself in yeast, FLN1 may form a homodimer to replace FLN1–FLN2 as the TRX z target in redox pathway for maintaining partial PEP activity in fln2–4. We proposed the partial PEP activity in the fln2 mutant allowed plastids to develop into fully functional chloroplasts when exogenous sucrose was supplied, and finally the mutants exhibited green phenotype.


Introduction
The first step in the life cycle of a plant is seedling establishment after seed germination. When seedlings are exposed to light, they synthesize chlorophyll; the cotyledons expand and turn green [1]. During this process, the etioplasts of cotyledons develop into photosynthetically functional chloroplasts, which allow seedlings to become photoautotrophy and no longer depend on energy stored in the seed [2,3]. Disruption of chloroplast function often results in severe phenotype, such as embryo lethality, albinism, or palegreen plants [4]. Chloroplast development is regulated by both nuclear and plastid genomes, and recent investigations revealed that many factors are involved in the process [5,6].
The transcription of plastid genes in higher plants is dependent on two RNA polymerases, the nuclear-encoded plastid RNA polymerase (NEP) and plastid-encoded plastid RNA polymerase (PEP) [7]. NEP preferentially transcribes housekeeping genes, while PEP is responsible for transcribing genes involved in photosynthesis [8,9]. PEP and its associated proteins/DNA can be purified by different biochemical purification procedures [10,11]. This protein/DNA-complex, termed ''transcriptionally active chromosomes'' (TAC), can be separated from the plastid components by gel filtration [12][13][14]. In Arabidopsis and mustard (Sinapis alba), a total of 35 components have been identified in the TAC complex including the core subunits of PEP encoded by plastome-located rpo genes and nuclear encoded subunits [15].
In recent years, a group of Arabidopsis leaf coloration mutants defective in the TAC components have been isolated. The Fesuperoxide dismutase (FSD) double mutant fsd2-1 fsd3-1 exhibited low tolerance to oxidative stress and accumulated decreased amounts of mRNA for PEP-dependent plastid genes [16]. Studies on the knockout mutants of thioredoxin z (TRX z) and fructokinase-like protein (FLN) showed that TRX z interacts with FLN1 and FLN2 to regulate the PEP-dependent transcription and chloroplast development in redox signaling pathway [17][18][19]. Mutations in the ptac3, ptac7 and ptac14 genes also resulted in the expression defects of PEP-dependent genes [20][21][22]. Based on the decreased expression of PEP-dependent genes, the PEP activity in these leaf coloration mutants was severely impaired which led to complete albino phenotype.
However, there exists another group of delayed greening mutants including ptac2 [15] and some non-ptac mutants, such as wco [23], dg1 [24] and pisp1 [25]. These mutants initially showed an albino phenotype on medium without sucrose, but they could turn green when exogenous sucrose was provided. In addition, a group of yellow mutants were reported such as ys1 [26] and ecb2-2 [27]. These mutants shared a similar phenotype that their cotyledons and true leaves were initially very yellow, but then turned green gradually without supplementation with exogenous carbon. YS1 encodes a chloroplast-localized pentatricopeptide repeat (PPR) protein, which is required for the editing of rpoB transcripts encoding the beta subunit of PEP in Arabidopsis [26]. AtECB2 also encodes a PPR protein with a C-terminal DYW domain. It regulates the editing of the plastid genes accD and ndhF [27]. The PEP activity has not been analyzed in pisp1 [25]. In all other delayed greening mutants and yellow mutants, PEP activity was defective [15,23,24,26,27]. Nonetheless, the mechanism for the greening of these PEP-related leaf coloration mutants is largely unknown.
During the seedling greening process, one of the most important events is the formation of chlorophyll, which allows plants to absorb energy from light [28]. The chlorophyll biosynthesis is strictly regulated by environmental and endogenous cues such as light signals [29], hormone signals [30], and plastid retrograde signals [31]. Previous studies revealed that several signalingresponsive transcription factors are required for chlorophyll biosynthesis. The transposase-derived transcription factors FHY3/FAR1 are responsive to light signals and directly activate the expression of the key gene HEMB1 in chlorophyll biosynthetic pathway [32]. The ethylene-stabilized transcription factors EIN3/ EIL1 were found to promote chlorophyll synthesis in the ethyleneinduced signaling pathway [33]. In addition, recent studies have demonstrated that the regulation by light, auxin/cytokinin and plastid-derived retrograde signals is dependent on the golden2-like transcription factors (GLKs), which is required for the expression of several chlorophyll biosynthesis genes [34,35]. Apart from the known chlorophyll biosynthesis, photosynthesis gene expression is crucial for the greening process by affecting the assembly of the photosynthetic apparatus [35]. PEP is the major machinery in regulating photosynthesis-related plastid gene expression. Thus, it is necessary for us to elucidate the relationship between PEP and seedling greening.
Here, we report the characterization of a delayed greening mutant fln2-4 in Arabidopsis, which displays albino phenotype but can develop greenish true leaves on sucrose-containing medium. The PEP-dependent plastid gene expression and chloroplast development were inhibited in fln2-4. Comparison of the PEPdependent gene expression among five leaf coloration mutants (trx z, ptac14, fln2-4, ecb2-2 and ys1) indicates that the PEP activity is critical for the leaf color phenotypes. The different degrees of PEP activity often give rise to the different leaf colors. Based on our yeast two-hybrid assay, FLN1 may form a homodimer instead of the FLN1-FLN2 heterodimer to function in regulation of PEP activity, which supports that the PEP activity in fln2-4 is higher than that in the complete albino mutants (trx z and ptac14). The relatively high PEP activity in fln2-4 allows the slow accumulation of the PEP-dependent gene transcripts for chloroplast development when supplemented with sucrose. With the formation of functional chloroplasts, the fln2-4 mutant shows green phenotype. All of these investigations should prove to be helpful to understand the mechanism of the greening phenomenon in many PEP-related Arabidopsis leaf coloration mutants.

Identification and Characterization of the fln2 Mutant
To analyze the functional roles of FLN2 gene during plant growth and development, we obtained two T-DNA insertion lines, SALK_005734 and CS811853, from the Arabidopsis Biological Resource Center (ABRC, http://abrc.osu.edu/). In these two lines, the T-DNA was inserted in the 3 rd exon and the 5 th exon of the FLN2 gene, respectively ( Figure 1A). Reverse transcriptionpolymerase chain reaction (RT-PCR) analysis showed that the FLN2 transcript was absent in both mutants ( Figure 1B). Due to the SALK_005734 line named as fln2-3 [19], the CS811853 line in this work was termed as fln2-4. Both fln2-3 and fln2-4 displayed albino cotyledons and were seedling lethal on MS medium without sucrose ( Figure 1C). The phenotypes of the two allelic mutants were similar; thereby the fln2-4 mutant was chosen for further analysis. Transmission electron microscopy (TEM) observations revealed that the chloroplasts in the 7-day-old fln2-4 mutants had a visible change in ultrastructural organization with irregular morphology and lacked internal membrane structures ( Figure 1D). To confirm that the knockout of FLN2 was responsible for the defects in the fln2-4 phenotype, a construct containing the genomic sequence of the FLN2 gene, as well as 1517-base pair (bp) upstream and the FLAG sequence was introduced into the heterozygous plant (FLN2/fln2-4). A total of 53 transgenic plants were obtained. Six of them were identified to be homozygous for the T-DNA insertion, and exhibited normal morphology as the wild type (WT) ( Figure 1E). These results demonstrate that the FLN2 gene is responsible for the defective phenotype in fln2-4 mutant, and FLN2 is important for chloroplast development and seedling growth.

Loss of FLN2 Down-Regulates the Expression of PEP-Dependent Plastid Genes
FLN2 is one component of TAC, and many TAC members were reported to affect the plastid gene expression [15][16][17][20][21][22]. To investigate the effect of the knockout of the FLN2 gene on the plastid gene expression, we used Northern hybridization to examine the transcriptional levels of plastid genes in 7-day-old fln2-4 mutants grown on MS medium without sucrose. The plastid genes are categorized into three classes based on whether they are transcribed by PEP and/or NEP [36]. The class I and class III genes are exclusively transcribed by PEP and NEP, respectively. While the class II genes are both PEP-and NEP-dependent. Our results showed that the expression of the class I genes (psbA, psbB and rbcL) were strongly reduced in the fln2-4 mutant compared with that of the WT. By contrast, the transcript levels of the class-II genes (rrn16 and clpP) and the class-III genes (accD and rpoA) did not have any significant variation between fln2-4 mutant and WT ( Figure 2). These results confirm that loss of FLN2 mainly affects the expression of PEP-dependent plastid genes.

FLN2 Interacts with FLN1 and pTAC5 in Yeast
FLN2 belongs to the components of TAC complex in Arabidopsis [15]. To establish the relationship between FLN2 and the other components of TAC, we performed a yeast two-hybrid screen. FLN2 was fused to the GAL4 DNA-binding domain (BD) as a bait to screen a pool composed of the thirty-five reported TAC components fused to the GAL4 activation domain (AD). After yeast transformation, we randomly selected clones with blue appearance in selective dropout (SD) medium lacking trptophan (Trp), leucine (Leu), histidine (His) and adenine hemisulfate salt (Ade) with X-a-gal for PCR amplification. Sequencing the PCR products identified three genes including FLN1, pTAC5 and TRX z. Yeast two-hybrid experiments verified the interactions between FLN2 and the three proteins ( Figure 3A). We subsequently performed pull-down assay to further confirm the interactions between FLN2 and FLN1 or pTAC5. Results showed that both the recombinant proteins GST-FLN1 and GST-pTAC5 were able to pull down FLN2-His ( Figure 3B and C). FLN2 and FLN1 belong to the pfkB family, and share high peptide similarity [37]. To examine whether the FLN1 and FLN2 proteins form homodimers, we analyzed the interactions between them by yeast two-hybrid assay. Results showed that FLN1 can interact with itself, but FLN2 can not ( Figure 3D). RpoA, pTAC12 and pTAC14 are the essential subunits in the PEP complex [18]. Yeast two-hybrid assay showed that FLN2 did not interact with these essential subunits in yeast ( Figure 3E). Based on previous investigations [21,22] and data in this study, a working model for partial TAC components is proposed ( Figure 3F). In this model, pTAC7 interacts with FLN1, pTAC10, pTAC12 and pTAC14 while FLN2 interacts with FLN1, TRX z and pTAC5.  The delayed greening fln2-4 mutant exhibits sugar-dependent for the survival of seedlings and growth of green true leaves. When supplemented with sucrose, the 7-day-old fln2-4 seedlings displayed yellow-tinted cotyledons, and then developed greenish true leaves in two weeks after germination ( Figure 4A). After transplanting them into soil, all the mutant seedlings with greenish true leaves could flower and produce fertile seeds without sucrose supplementation ( Figure 4B). To unveil the ultrastructural basis of fln2-4 seedlings during the delayed greening process, we examined dynamic changes in chloroplasts using TEM. In the leaves of 7day-old WT grown with sucrose, chloroplasts were crescentshaped and contained well-formed internal membrane structures including stroma thylakoids and stacked grana thylakoids ( Figure 4C). In contrast, at the same stage, the chloroplasts in the 7-day-old fln2-4 mutants were still highly vacuolated, but the lamellar membrane structures appeared ( Figure 4C). In the 14day-old fln2-4 mutants grown under the same conditions, chloroplasts with well-organized thylakoid membrane could be observed ( Figure 4C). This suggests the fln2-4 chloroplast can be gradually formed although its development is slower than that in WT. In addition, we measured the levels of chlorophyll a and b in both the mutant and WT seedlings during the greening process. The total chlorophyll content of fln2-4 mutant (93.5862.45 mg g 21 fresh weight for the fln2-4 mutant type) was about 16.9% of WT (554.94615.51 mg g 21 fresh weight for the wild type) grown with sucrose for 7 days, and the chlorophyll a/b ratio of fln2-4 was about half of WT ( Figure 4D; Table 1). In 14-day-old mutant seedlings grown under the same conditions, the chlorophyll content (292.16624.91 mg g 21 fresh weight for the fln2-4 mutant type) was about 42% compared to that of the WT (695.19624.14 mg g 21 fresh weight for the wild type), and the chlorophyll a/b ratio was close to that of WT ( Figure 4D; Table 1). These observations indicated that the chlorophyll biosynthesis in fln2-4 was partially recovered. Furthermore, we investigated the plastid ultrastructure development of the WT and fln2-4 grown on sucrose-containing medium during de-etiolation. In the WT and fln2-4 seedlings grown in darkness for 5 days, the etioplasts contained a large prolamellar body ( Figure 5, left panels). When etiolated seedlings were exposed to light for 6 hours, the prolamellar in WT developed into stromal lamellae ( Figure 5, middle panels). In contrast, plastids in fln2-4 contained less stromal lamellae ( Figure 5, middle panels). After de-etiolation for 24 hours, plastids in WT contained well-developed thylakoid membrane system and starch granules, while loose thylakoid lamellae and grana thylakoids were detected in fln2-4 ( Figure 5, right panels). When grown on MS medium without sucrose in darkness for 5 days, the etioplasts of fln2-4 were similar with that of the WT. After de-etiolation for 24 hours, only a few thylakoid lamellae could be observed in fln2-4 ( Figure 5). These results revealed that plastid development in fln2-4 grown on sucrosecontaining medium proceeds slowly, but it can gradually form well-structured chloroplast during the de-etiolation process.
PEP-Dependent Plastid Gene Expression in fln2-4 Is Higher than that in the Complete Albino Mutants, but Lower than that in the Yellow Mutants The most distinct characteristic of fln2-4 mutant is its delayed greening phenotype when grown on sucrose-containing medium. We assayed the expression of PEP-dependent plastid genes in fln2-4 during its greening process. Results showed that the transcript levels of the three selected genes (rbcL, psbA, psbB) in 7-day-old fln2-4 mutant were still significantly lower than that in WT. However, after 14 days of growth on sucrose-containing medium, the rbcL mRNA accumulated to similar levels in WT, while the psbB transcription was slightly enhanced (Figure 6). These results indicated that the transcripts of several PEP-dependent genes were able to gradually accumulate during the greening of fln2-4. To disclose whether these differences in plastid gene expression were tightly linked with leaf phenotype, we chose four leaf coloration mutants for further research, which including two complete albino mutants trx z [17] and ptac14 [22], and two yellow mutants ecb2-2 [27] and ys1 [26]. When germinated on MS medium, trx z and ptac14 exhibited albino cotyledons and died before developing true leaves. While grown on MS medium containing 2% sucrose, they could produce pale yellow cotyledons and true leaves ( Figure 7A). Nevertheless, these mutants subsequently died. In contrast, the ecb2-2 and ys1 seedlings initially exhibited yellow cotyledons and were able to turn green without exogenous sucrose ( Figure 7A). However, the mature ecb2-2 and ys1 plants were slightly weaker than WT (data not shown). We compared the PEP activity in these four leaf coloration mutants (trx z, ptac14, ecb2-2 and ys1) and the fln2-4 through detecting the transcript abundance of four PEPdependent chloroplast genes (psaB, psbA, psbB and petD). Quantitative Real-Time PCR (qRT-PCR) assay showed that the expression of these genes in the two complete albino mutants (trx z and ptac14) were lower than that in the fln2-4, ecb2-2 and ys1, respectively. In addition, the expression of these genes in the delayed greening fln2-4 mutant was lower than that in the yellow mutant ecb2-2 and ys1 ( Figure 7B). Northern blot analysis using probes for psbA and psbB in these mutants exhibited the similar result ( Figure 7C). All of these results suggest that the PEP activity in fln2-4 is higher than that in the complete albino mutants but lower than that in the yellow mutants.

FLN2 is One Component of TAC to Regulate PEP-Dependent Plastid Gene Expression
The knockouts of many TAC members in Arabidopsis have been reported to affect plastid gene expression and chloroplast development including ptac6, ptac12 [15], fsd2, fsd3 [16], trx z [17], fln1 [17,18], ptac14 [18,22], ptac3 [20] and ptac7 [21]. FLN2 is one component of TAC and is involved in the redox pathway mediated by TRX z [15,17]. The knockout lines fln2-1 and fln2-2, exhibited pale green cotyledons and slow greening on growth medium and soil [19]. Here, we isolated another two knockout lines fln2-3 and fln2-4. These two lines displayed albino phenotype when grown on MS medium ( Figure 1C), but they could turn green when grown on sucrose-containing medium ( Figure 4A and data not shown), which was slightly different from the reported phenotypes of fln2-1 and fln2-2 [19]. This difference was probably due to the different growth conditions. Nevertheless, the PEP-dependent plastid gene expression was down-regulated in all the four lines ( Figure 2) [19], as well as in the Arabidopsis RNAi plants [17]. Therefore, FLN2 is similar with other functional TAC components in regulating the PEP-dependent plastid gene expression and chloroplast development.
TRX z interacts with FLN1 and FLN2 [17], and recently pTAC7 has been reported to interact with FLN1, pTAC10, pTAC12 and pTAC14 [21]. Besides interacting with TRX z, we found that FLN2 also interacts with FLN1 and pTAC5 in yeast ( Figure 3A). These interactions revealed the preliminary relationship of these TAC components in the complex ( Figure 3F). The TAC components appear to perform basic functions or play roles as regulators for the transcription regulation and environmental adaptation in chloroplast. Steiner et al. identified 10 essential nonrpo subunits of the PEP complex named PEP-associated proteins (PAPs) [18]. These PAPs might be the core subunits of the TAC complex and play the basic function in regulation of plastid gene expression. FLN1 and TRX z belong to PAPs, while the FLN2 and pTAC5 were not included. The fln1 [18,19] and trx z ( Figure 7A) displayed seedling lethality, which were more severe   than that of fln2. In addition, silencing of FLN2 results in weaker effect on PEP-dependent plastid gene expression than does silencing of FLN1 [17]. Furthermore, the ptac5 mutant showed similar phenotype with WT under normal condition, but was sensitive to heat stress ( Figure S1). Overall, FLN2 and pTAC5 may represent peripheral component rather than core subunit to function as a regulator for environmental adaptation in the TAC complex.

PEP Activity is Critical for the Early Development of Chloroplast and Leaf Greening
Leaf coloration mutants are frequently observed in plant. In Arabidopsis, seedling leaf coloration mutants can be classified into several groups. For the yellow mutants such as ys1 [26] and ecb2-2 [27], they initially exhibited yellow cotyledons and true leaves, and subsequently turned green and grew photo-autotrophically under normal condition. For the complete albino mutants such as trx z [17] and ptac14 [22], they displayed albino cotyledons and could not turn green even on sucrose-containing medium ( Figure 7A). These lines are seedling lethal and can not complete their life cycles. Apart from the complete albino mutants and the yellow mutants, there exists another group of delayed greening mutants such as ptac2 [15], wco [23], dg1 [24] and pisp1 [25]. These mutants showed albino phenotype initially, however, they could turn green on the sucrose-containing medium. The fln2-4 mutant is similar with the delayed greening mutants ( Figure 4A). These different groups of leaf coloration mutants often showed different degrees of defects in chloroplast development. During seed germination, the photosynthetically functional chloroplasts are quickly formed and seedlings turn green before the energy in the seed is exhausted. The rapid formation of mature chloroplast during seed germination is very important for seedling greening. Thus, it seems that the different degrees of chloroplast developmental status give rise to the different leaf colors of mutant seedlings.
The PEP-dependent plastid gene expression analysis in this work indicates the PEP activity in the delayed greening mutant is higher than that in the complete albino mutants, but lower than that in the yellow mutants ( Figure 7B and C). The PEP activity in these leaf coloration mutants is in agreement with their phenotypes. The more seriously reduced PEP activity results in the more severe phenotype. Recently, Zhelyazkova et al. have reported that PEP is the dominating RNA polymerase in chloroplast maturation [38]. Based on their work together with the data in this paper, we hypothesize that a threshold of PEP activity is critical for the chloroplast development and leaf greening. In complete albino mutants, the PEP activity is belowthreshold and the chloroplast development in these mutants may be completely blocked. In the fln2-4 mutant, relatively high PEP activity allows the slow chloroplast development. However, the energy in seed is not enough to support the formation of mature chloroplast and therefore seedlings show albino phenotype. The exogenous sucrose in the medium supports the mutant plants to overcome energy deficiency and gradually accumulate enough photosynthesis gene transcripts for functional chloroplast formation. The development of fully functional chloroplast allows the mutant plants to show green phenotype.
A Homodimer Formed by FLN1 may Partially Compensate the Loss of FLN2-FLN1 Heterodimer to Maintain some PEP Activity The cellular redox environment has a fundamental role in regulating most plastid processes, including secondary metabolism, gene transcription, and protein synthesis as well as import [39,40]. In Arabidopsis, Arsova et al. identified that a novel thioredoxin TRX z which may be involved in the redox pathway to regulate PEP-dependent plastid gene expression and chloroplast development through interacting with FLN1 and FLN2 [17]. Here, our data showed that there exist an interaction between FLN1 and FLN2, which indicated that the FLNs may form a heterodimer as the TRX z target in the TAC complex. The homologs of FLN1 and TRX z are present in moss, pteridophyta and angiosperm, while FLN2 homologs are only present in angiosperm, based on the available genomic data. This suggests that FLN1 might be an ancient protein, while FLN2 is a novel protein evolved from FLN1 after the emergence of the pteridophyta during the course of evolution. In this work, FLN1 was found to interact with itself in yeast ( Figure 3D). This suggests that FLN1 might form a homodimer in the TAC complex to regulate the PEP activity and chloroplast development in moss. With the presence of FLN2, the FLN1 may interact with FLN2 rather than itself. However, as the essential protein in the redox pathway, FLN1 remains the ability to interact with itself. In Arabidopsis, both the TRX z and FLN1-FLN2 heterodimer are the essential components of redox pathway in regulation of PEP activity. The PEP activity in the knockout of TRX z was severely impaired and therefore the trx z mutant showed complete albino phenotype. In the fln1 mutant, the redox pathway may be blocked due to the absence of the FLN1-FLN2 heterodimer, which leads to albinism phenotype. While in the fln2-4 mutant, FLN1 may form a homodimer to replace the FLN1-FLN2 heterodimer, and thus the mutant remains partial signaling of the redox pathway in regulation of PEP activity. This supports the slow accumulation of the PEP-dependent plastid gene transcripts for chloroplast development and the delayed greening of this mutant.

Plant Materials and Growth Conditions
WT and mutant Arabidopsis thaliana (Columbia) seeds were used in this study. The T-DNA insertion lines SALK_005734 (fln2-3), CS811853 (fln2-4), SALK_0028162 (trx z), SALK_005814 (ptac14), SALK_049133 (ptac5) were obtained from the ABRC. The ecb2-2 mutant was screened using an ethylmethane sulfonate mutagenesis strategy as described in our recent paper [27]. Plants were grown on Murashige and Skoog (MS) agar plates supplemented with or without 2% (w/v) sucrose. All plate-grown seedlings were vernalized under 4uC condition for 2 days and then were grown in a chamber at 22uC with a 16-h-light (120 mmol photonsm -2 s -1 )/8-h-dark cycle. The soil-grown plants were grown with a photon flux density of 120 mmol m -2 s -1 at 22uC.

Chlorophyll Determination and Chloroplast Ultrastructure
Total chlorophyll content was measured according to the record as described earlier [41]. Arabidopsis leaf segments were from the cotyledons of 7-day-old fln2-4 plants grown with or without sucrose and the true leaves of 14-day-old mutants grown on sucrose-containing medium. Transmission electron micrographs were obtained exactly as described earlier [42]. The specimens were examined Hitachi H7650 transmission electron microscopy (http://www.hitachi.com).

RNA Isolation, cDNA Synthesis, RT-PCR Analysis and Quantitative Real-Time RT-PCR
Procedures for the purification of total RNAs for cDNA synthesis, RT-PCR, and qRT-PCR were performed as previously described [41]. The specific primers used to quantify the expression of FLN2 were as follows: sense primer, 59-ATGGCTGCTGGTAGGAGAAAG-39; antisense primer, 59-TCATAAACTACCATCTTCAAA-39. The b-tublin was used as the internal standard for qRT-PCR analysis.

Northern Hybridization
Approximately 10mg total RNAs of each sample were separated on a 1.5% formaldehyde-agarose gel, transferred onto a nylon membrane (Pall, Mexica) and hybridized with the specific probes. The probes were synthesized with a PCR amplification-DIGlabeling kit. Sequence data for the PCR primers performed in this study can be found in previous report [22]. Chemiluminescent detection was carried out according to the Roche manual (Roche, http://www.roche.com).