Transgenic tobacco lines expressing ectopically CPO (coproporphyrinogen oxidase) antisense RNA suffer from photodynamic cell death in leaf tissue [23]. CPO catalyzes the oxidative decarboxylation of two propionate side chains to vinyl groups of coproporphyrinogen leading to protoporphyrinogen. The necrotic cell death phenotype of CPO-antisense lines (Figure 1A) is explained by accumulation of the photoreactive CPO substrate coproporphyrinogen and the oxidized coproporphyrin. In continuation to the initial characterization of CPO-deficient tobacco plants we were interested to identify early inducible genes in response to coproporphyrin-induced cell death. We made use of a light-dose dependent phenotypical change of coproporphyrin-accumulating lines for the search of early inducible genes upon photooxidative stress. The leaf phenotype of wild-type and the CPO-antisense line #41 grown under short day condition (6 h light/18 h dark) at 200 μmol photon m⁻²·s⁻¹ differed from that of plants exposed to long day condition (16 h light/8 h dark) at the same light intensity of 200 μmol photon m⁻²·s⁻¹. Although, the plants already accumulate coproporphyrin under short day condition, leaf necrosis is not detectable. The shift from low light to a higher light dosage under the same light intensity led to visible cell death symptoms within 24–48 h (Figure 1A) [23,24].

To identify genes induced early under photooxidative stress conditions in the porphyrin-accumulating CPO-antisense tobacco plants, eight-week-old control and CPO-antisense plants (line 1/41) were grown under short day condition before the light exposure was extended for six additional hours and leaf samples were harvested for RNA extractions. Total RNA of control and wild-type plants was subjected to subtractive suppression hybridization. The approach resulted in the identification of, in total, 184 cDNA sequences with differential expression pattern between CPO antisense line 1/41 and wild-type plants, which encode 70 different proteins (Table S1). Sixty-one percent of the cDNA sequences were confirmed by reverse northern blotting to be expressed with elevated levels in the CPO antisense RNA expressing line.

Among the genes with increased expression in the transgenic line, NtHAP5b was identified as encoding the tobacco transcription factor subunit NF-YC. For consistency with the designation of the homologous Arabidopsis genes for the NF-Y subunits, we named this gene NtNF-YC. The induced NF-YC expression upon porphyrin-derived photooxidative stress was confirmed with quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) analysis of tobacco wild-type and CPO antisense plants before and after transition to high light doses. The three-week-old plants were grown first under short day and low light condition (6 h light/18 h dark). Then control and transgenic plants were exposed six hours to low and high light before they were harvested for mRNA quantification (Figure 1B). NtNF-YC expression was induced during high light and the mRNA accumulated 4-fold compared to low light condition of control and CPO antisense seedlings as well as to high light-exposed wild-type seedlings. In conclusion, this screen enabled identification of a putative candidate gene encoding the C-subunit of the transcription factor NF-Y, which is inducible upon photoreactivity of accumulating tetrapyrrole intermediates.

Taking advantage of the previous exploitation of the entire Arabidopsis genome we searched for NtNF-YC homologs among the members of the Arabidopsis NF-YC gene family. To benefit from the complete collection of the NF-Y gene families for the three subunits we aimed at the identification of the candidate gene among the NF-YC family that is also induced upon porphyric stress. The highest similarity to NtNF-YC exists among a subgroup of the NF-YC family consisting of NF-YC1, NF-YC2, NF-YC3, NF-YC4 and NF-YC9 (Figure 2C). These investigations for the identification of the NtNF-YC-homologous gene were embedded in a comprehensive analysis of the transcriptome of the complete set of all NF-Y genes in response to different adverse growth conditions including low temperature, heat, low and high light intensities and treatments of photodynamic herbicides, e.g., acifluorfen, norflurazon (inhibitor of carotenoid synthesis), DCMU (inhibitor of the linear photosynthetic electron transport chain), as well as oxidants, e.g., hydrogen peroxide. We quantified the transcript levels of all 36 NF-Y genes encoding the NF-YA, NF-YB and NF-YC subunits, which are currently known in Arabidopsis with qRT-PCR.

Table 1 surveys the results of the expression analyses and highlights those NF-Y genes which show a rapid and stronger change in expression (<0.5 or >2.0) (or >5.0 for drought stress) after the exposure to different stresses and effectors. The different stress conditions and applications are described in Material and Methods. Among the NF-Y genes only a few representatives of the NF-YA, NF-YB and NF-YC subfamilies reveal pronounced changes of their transcript levels upon application of several inducible stress conditions in comparison to normal growth condition. Drought stress and heat treatments lead to induction of more NF-Y genes, most likely because of the intensity of the stress (heat) and the extended stress period (drought). Among those genes with enhanced accumulation of mRNA content in response to stress, NF-YA1 and NF-YC2 responded more extensively with increased mRNA levels than other NF-Y genes.

In default of a coproporphyrin-accumulating Arabidopsis cpo mutant we applied acifluorfen to wild-type seedlings. Acifluorfen is a peroxidising herbicide that inhibits protoporphyrinogen oxidase, an enzyme of tetrapyrrole biosynthesis [27]. Figure 2A displays the transcript profile of all NF-Y genes 3 h after application of 50 μM acifluorfen and in a different experiment after 3 h of high light stress. Among these genes ultimately only NF-YC2 accumulates six-fold higher RNA levels in response to acifluorfen treatment and high light stress (Figures 2A,B). Figure 2B depicts the transcript level of NF-YC2 after multiple stress treatments indicating the strong oxidative stress inducibility. In summary, elevated NF-YC2 transcript levels often correlate with oxidative stress (H₂O₂, heat, drought) or photooxidative stress (acifluorfen, norflurazon, high light), whereas NF-YC2 mRNA amounts are slightly reduced under conditions which predominantly possess a reduced potential for oxidative stress (low temperature, low light). Among the NF-YC subfamily (Figure 2C) NF-YC2 seems to have the strongest responsiveness against oxidative stress conditions. Enhanced NF-YC2 expression upon porphyrins-induced photooxidative stress resembles the expression of NtNF-YC in CPO-antisense plants.

Among the potential T-DNA-insertion mutant lines for NF-YC2 (SALK_11422; GABI_369E03, GABI_669A05) we identified only one mutant line with a T-DNA-insertion in the coding region (GABI_669A05). This line was designated nf-yc2. The exact T-DNA insertion site of nf-yc2 could be confirmed 90 base pairs upstream of the translation stop codon (Figure 3A,B), while the inserted T-DNA of SALK_11422 and GABI_369E03 interrupted NF-YC2 in the 5′ and 3′ untranslated region, respectively. The transcript levels of all T-DNA mutant lines were similar to wild-type seedlings (data not shown). But qRT-PCR revealed an increased content of the 3′ part of the cDNA downstream of the T-DNA-insertion site which can be explained with enhanced transcriptional activity due to promoter activity on the T-DNA (Figure 3C). It was expected that homozygous nf-yc2 synthesizes a truncated protein. But the size, the amount and the stability of NF-YC2 in control plants and the homozygous nf-yc2 mutant could not be determined. An antibody raised against a small non-conserved peptide sequence of NF-YC2 did not specifically recognize a protein of the expected size in the nuclear extract or total extracts of Arabidopsis wild-type seedlings. Efforts to enrich the specific antibody by purification attempts failed.

Seedlings of nf-yc2 and wild type exposed to different abiotic stress conditions did not show differences of the macroscopic phenotype. This holds true, regardless, whether mutant and control plants were either treated with acifluorfen or Rose Bengal as well as exposed to high light or heat stress. When toxic amounts of herbicides were applied to the plants, formation of necrotic tissue upon acifluorfen treatment or bleaching in response to applied norflurazon were quantitatively similar in leaves of mutant and wild-type seedlings. We derived from these experiments that an obvious phenotypic diversification between nf-yc2 and control seedlings could not be observed.

Additionally, more physiological properties of nf-yc2 were examined in comparison to wild-type seedlings. Analyses of chlorophyll fluorescence parameters of mutant and wild-type seedlings revealed no significant differences in photosynthetic properties (Figure S1). Additionally, levels of tetrapyrrole intermediates were determined after a 36 h acifluorfen treatment. But the seedlings did not display significant differences in the content of tetrapyrrole intermediates (Figure S2). These results indicate that nf-yc2 was not more sensitive towards photooxidative stress than control seedlings.

In order to examine the effect of ectopic AtNF-YC2 overexpression on photooxidative stress response, the full-length cDNA sequence of NF-YC2 was inserted downstream of a CaMV 35S promoter in the pCAMBIA 3301 vector and introduced into the genome of Arabidopsis thaliana (Figure 3D). Selection of transformed lines with high expression of the transgene revealed lines #21, #24 and #33 with highly elevated NF-YC2 mRNA levels (Figure 3F). Using the anti-NF-YC2 antibody, the overproduced protein was immunologically detectable in total cell extracts (Figure 3E), while the content of the endogenous NF-YC2 remained below the detection level in total wild-type extracts. The selected NF-YC2 overexpressor lines were sprayed with 100 μM acifluorfen. However, no phenotypic differences were detectable on leaves of transgenic lines in comparison to wild-type leaves. Upon application of different concentrations of acifluorfen transgenic and wild-type seedlings showed always a similar formation of leaf necrosis after herbicide treatments. The viability of transgenic lines was also not elevated upon acifluorfen treatments in comparison to control seedlings after herbicide application (data not shown). Also the expression of the stress related APX genes (APX1, APX2 and APX4), which are involved in the detoxification of reactive oxygen species (ROS) [28] and are induced by oxidative stress [29], were neither in nf-yc2 nor in 35SCaMV::NF-YC2 significantly increased in relation to wild-type plants independent if acifluorfen were applied or not (data not shown).

In addition, the impact of photooxidative stress was also tested in other Arabidopsis T-DNA insertion mutants of the Arabidopsis NF-YC family with high sequence similarity to NF-YC2 (Figure 3). T-DNA insertion mutants for the family members AtNF-YC1 (GABI_004H11, SALK_86334), AtNF-YC3 (GABI_51E10), AtNF-YC4 (SALK_32163) and AtNF-YC9 (SALK_058903) were isolated and homozygous lines were generated. With the exception of nf-yc9, homozygous nf-yc1, nf-yc3 and nf-yc4 mutant plants could always be identified with the T-DNA insertion in the coding region. In these T-DNA insertion lines no transcript of the respective AtNF-YC gene was detected (data not shown). nf-yc3 (GABI_51E10) and nf-yc4 (SALK_32163) as well as nf-yc1-1 (GABI_004H11) and nf-yc1-2 (SALK_86334) T-DNA insertion mutants neither showed phenotypically changes in response to treatments with acifluorfen nor to the exposure to high light. The growth of these mutants resembled that of wild-type and nf-yc2 seedlings.

Apart from the transcript analysis and the assessment of phenotypic differences between mutant and control seedlings upon abiotic stress, NF-YC2 overexpressor lines and the T-DNA insertion mutant nf-yc2 were examined for induction of flowering under short and long day conditions. In parallel, nf-yc1-2 as well as AtNF-YC1 overexpressor lines were also tested as representatives of the NF-YC subfamily with high homology to NF-YC2. Genotyping of nf-yc1-2 revealed the T-DNA insertion in the exon region (Figure S3A). Homozygous nf-yc1-2 does not contain detectable amounts of NF-YC1 transcripts (Figure S3B). 35SCaMV::NF-YC1 expression in transgenic Arabidopsis lines yielded a strong accumulation of the specific mRNA. The lines #41, #42 and #57 represent transformants with over-expression of NF-YC1 (Figure S3C,D).

Thirty-eight-day-old seedlings of the transgenic lines 35SCaMV::NF-YC1 #41, #42, #57 and 35SCaMV::NF-YC2 #21, #24 and #33 showed a significant early flowering phenotype under long day condition compared to wild-type plants and nf-yc2 and nf-yc1-2. Figure 4A shows always one representative plant of each set of overexpressor lines. The overexpressor lines flowered at least 16 days earlier than wild type (Figure 4B). Under short day condition all mutants and overexpressor lines did not significantly modulate the flowering time (Figure 4C). Flowering was set when petals started to be seen during flower development. As a control, co-1, a mutant of CONSTANS (co-1) showed a significantly delayed flowering than wild-type plants under long day condition (Figure 4B). In consistencies with a previous report co-1 flowered significantly earlier under short day condition (Figure 4C) [30].

In 21-day-old mutant seedlings with modified expression of NF-YC1 and NF-YC2 we explored transcript levels of FLOWERING LOCUS (FT) and SUPRESSOR OF OVEREXPRESSION OF CO1 (SOC1) [31], two genes in the photoperiodic control of flowering by qRT-PCR. Growth under long day condition (14 h light/10 h dark) leads to a 20-fold accumulation of FT transcripts in the earlier flowering periods of the 35SCaMV::NF-YC1 and 35SCaMV::NF-YC2 lines (Figure 5A). The two mutants nf-yc1-2 and nf-yc2 also showed elevated FT RNA levels, which do not correlate with a significant variation in flowering time in comparison to wild-type properties. As control co-1 contained lower FT mRNA levels under long day condition, which correlate with the late flowering phenotype (Figure 5A). Short day exposure (10 h light/14 h dark) of the NF-YC1 and NF-YC2 overexpressor lines did not result in modified expression of FT (Figure 5B). The SOC1 transcript levels were not modulated under long or short day condition in overexpressor plants and T-DNA insertion mutants in comparison to wild-type plants (Figure 5A,B).

Arabidopsis thaliana plants grew for 14, 21 or 34 days under light/dark conditions (120 μM photons m⁻²·s⁻¹, 12 h light/ 12 h dark at 23 °C) before they were exposed to high light (HL—500 μM photons m⁻²·s⁻¹), low light (LL—50 μM photons m⁻²·s⁻¹), increased (38 °C) or low temperature (4 °C) respectively. Alternatively, plants were treated with 50 μM acifluorfen, 10 mM DCMU, 10 mM H₂O₂ or 20 μM norflurazon by spraying leaves with these agents. Leaf material was harvested after exposure time of 3 h. Drought stress was induced when plants were not watered for 8 days.

For RNA quantification Nicotiana tabacum var. Samsun NN (SNN) and CPO-antisense plants [23] were initially cultured for 14 days under low light conditions (40 μM photons m⁻²·s⁻¹, 12 h light/ 12 h dark at 27 °C) before exposure to high light (400 μM photons m⁻²·s⁻¹, 12 h light/ 12 h dark at 27 °C). Leaf material was harvested before and 6 h after exposure to high light. For suppression subtractive hybridization experiments tobacco plants were cultivated for 8 weeks with a low light dose (200 μM photons m⁻²·s⁻¹, 6 h light/ 18 h dark at 25 °C) and were subsequently exposed to a high light dose (400 μM photons m⁻²·s⁻¹) for 12 h before leaf material was harvested.

For qRT-PCR total RNA of 100 mg leaf material was extracted with RNeasy mini kit (Qiagen, Germany) in accordance to the supplier’s manual. The amount of 1.250 ng of total RNA were treated with 2U TURBO DNA-free (Ambion, USA) for removing genomic DNA contaminations before first strand cDNA synthesis was performed using SuperScript III reverse transcriptase (Invitrogen, USA) in the presence of 50 ng oligo dT₁₈-primer (Sigma, Germany). As a template for qRT-PCR analysis ten times diluted cDNA were used in 20 μL reactions (1 μL cDNA, 500 nM each primer, 0.4 μL 50 × SYBR Green and 10 μL ImmoMix TM (Bioline, Germany) performed with the Rotor Gene 2000 (Corbert-Research, Australia) or 5 μL PCR reactions (0.5 μL cDNA, 200 nM each primer and 2.5 μL SYBR Green PCR Master Mix (Applied Biosystems, USA) performed in 384 wells plates using the ABI PRISM 7900HT sequence detection system (Applied Biosystems, USA). The amplification program consisted of an initial hot-start activation at 94 °C for 10 min, 45 cycles including 15 s denaturing at 94 °C, 20 s annealing at 60 °C and 20 s elongation phase at 72 °C. Primers of the Arabidopsis thaliana NF-Y gene families were kindly provided by the Max-Planck Institute of Molecular Plant Physiology Golm. Oligonucleotide sequences are listed in the Table S2. Transcript levels of each gene were normalized to an Arabidopsis member of the Arabidopsis SAND gene family (At2g28390) or Nicotiana tabacum ACTIN gene respectively. Primer efficiency was determined by linear regression on the Log(fluorescence) of each PCR reaction using the LinRegPCR software [40] and PCR efficiency was defined as mean value for each primer pair. For semi-quantitative RT-PCR, total RNA were isolated with TRIsure (Bioline, Germany) and treated with 1U DNAseI (Fermentas, Germany). First strand cDNA synthesis was performed by RevertAid M-MuLV reverse transcriptase (Fermentas, Germany) in accordance to the manufacturer instructions.

Full length cDNA constructs of AtNF-YC1 and AtNF-YC2 were amplified by PCR using specific primers (AtNF-YC1: 5′-ACCATGGATACCAACAACCAGC-3′ and 5′-TGACGTCCACCTTGGCCG TCGAGA-3′; AtNF-YC2: 5′-ACCATGGAGCAGTCAGAAGAGG-3′ and 5′-TGACGTCCAGACTC ATCAGGGTGTTG-3′) and ligated into NcoI and ActII restriction sites of the multiple cloning site of the modified pCAMBIA3301 plasmid. The correct sequence of the 35SCaMV::AtNF-YC1 and 35SCaMV:: AtNF-YC2 gene constructs was confirmed by sequencing and used for A. tumefaciens-mediated stable transformation of Arabidopsis (ectoype Col-0) plants by means of the floral-dip transformation method described previously [41]. Transformed plants were selected by successive treatments of two, three and four-week-old seedlings with 0.1% (w:v) BASTA (Bayer CropScience AG, Germany). T-DNA insertion mutants of the SALK and GABI-Kat collection were obtained from the European Arabidopsis Stock Centre (Nottingham, UK) or Bielefeld University, Germany. Homozygous T-DNA insertions were validated by PCR using specific oligo primer for either the wild-type allele or the T-DNA left boarder and for indication of different genes (primer sequences in Table S2). Transcript levels of the affected genes in mutant plants were analyzed as described above. AtNF-YC2 protein amounts of pCAMBIA3301::AtNF-YC2 plants were determined by immunoblot analysis in accordance to standard methods [42].

In vivo chlorophyll fluorescence was measured with attached leaves of dark-adapted plants as previously described [43]. The following fluorescence parameters were assessed: the maximum photochemical efficiency of PS II in the dark-adapted state F_v/F_m = (F_m – F₀)/F_m, effective quantum yield of photochemical energy conversion in PSII (ΦP_SII = (F_m′ – F_t) F_m′⁻¹) and non-photochemical quenching (NPQ = (F_m – F_m′) F_m′⁻¹ ) in accordance to [44]. Photon flux densities were measured using a quantum sensor (LI-189A, Li-Cor, Lincoln, NE).

Extraction and analysis of accumulated tetrapyrroles in tobacco leaves were essentially performed as described [23]. Modifications regard the extraction of liquid nitrogen-ground leaf material with 50 mM potassium phosphate buffer (pH 7.8), methano1:0.l M NH₄OH (9:1, v/v) and acetone: 0.l M NH₄OH (9:1, v/v). The porphyrins were separated by HPLC (Agilent, Germany) on a RP 18 column (Novapak C18, 4 μm particle size, 4.6 × 250 mm). Column eluent was monitored by fluorescence (λ_ex 405 nm, λ_em 625 nm) and porphyrins were identified and quantified by authentic standards.

Two μg polyA⁺ RNA of the sixth tobacco leaf (counting from the top of the plant) from wild-type (driver) and CPO antisense # 41 (tester) plants was extracted and applied to the forward subtracted and the reverse subtracted approach according to the protocol of the supplier (Clontech, Germany). After E. coli transformation with pCRII vector (Invitrogen, USA) containing forward subtracted cDNA fragments 1993 colonies were obtained, which were subjected to a colony hybridization using the labeled probes from forward and reverse subtracted cDNA. Finally, 234 bacterial colonies were confirmed and their cDNA sequenced.

All data shown in graphs are mean values and corresponding standard deviations are displayed as error bars calculated using the Excel 2007 software (Microsoft Corp., USA). For relative expression analysis three biological replicates were used, whereas fluorescence parameter, tetrapyrrole intermediates and days to visible petals were determined by evaluation of 12 biological replicates (different plants). Differences between WT and mutant lines were tested using a two-tailed Student’s t test and were regarded as significant for p < 0.05.