Proteome and metabolome profiling of cytokinin action in Arabidopsis identifying both distinct and similar responses to cytokinin down- and up-regulation

In plants, numerous developmental processes are controlled by cytokinin (CK) levels and their ratios to levels of other hormones. While molecular mechanisms underlying the regulatory roles of CKs have been intensely researched, proteomic and metabolomic responses to CK deficiency are unknown. Transgenic Arabidopsis seedlings carrying inducible barley cytokinin oxidase/dehydrogenase (CaMV35S>GR>HvCKX2) and agrobacterial isopentenyl transferase (CaMV35S>GR>ipt) constructs were profiled to elucidate proteome- and metabolome-wide responses to down- and up-regulation of CK levels, respectively. Proteome profiling identified >1100 proteins, 155 of which responded to HvCKX2 and/or ipt activation, mostly involved in growth, development, and/or hormone and light signalling. The metabolome profiling covered 79 metabolites, 33 of which responded to HvCKX2 and/or ipt activation, mostly amino acids, carbohydrates, and organic acids. Comparison of the data sets obtained from activated CaMV35S>GR>HvCKX2 and CaMV35S>GR>ipt plants revealed unexpectedly extensive overlaps. Integration of the proteomic and metabolomic data sets revealed: (i) novel components of molecular circuits involved in CK action (e.g. ribosomal proteins); (ii) previously unrecognized links to redox regulation and stress hormone signalling networks; and (iii) CK content markers. The striking overlaps in profiles observed in CK-deficient and CK-overproducing seedlings might explain surprising previously reported similarities between plants with down- and up-regulated CK levels.


(2) Measurements
The resulting peptides were desalted (SPEC plate C18, Agilent), dried and dissolved in 0.5% (v/v) formic acid in 5% (v/v) acetonitrile, and analyzed online by nanoflow C18 reverse-phase liquid chromatography, loading 5 g of protein onto a 15 cm Ascentis Express Column (0.1 mm inner diameter; Sigma-Aldrich) and an Eksigent UPLC system (Eksigent) directly coupled to an ESI source and an LTQ-Orbitrap XL mass spectrometer (Thermo Scientific). Peptides were eluted with a linear 155-min 5% to 95% acetonitrile gradient. The mass spectrometer was operated in positive ion mode using data-dependent automatic switching between MS and MS/MS acquisition modes. Fourier-transformed full scan mass spectra were acquired at a target value of 9E05 ions with resolution r = 30 000 at m/z 400 and an m/z range of 300-2000. The seven most intense ions were selected for collision-induced dissociation with a target value of 5000 ions in the LTQ. After MS analysis, raw files were searched against the TAIR10 Arabidopsis database using the Sequest algorithm. For identification and spectral count-based data matrix generation Proteome Discoverer (v 1.3, Thermo Scientific) was used. Database search criteria were as follows: TAIR10 database; enzyme -trypsin, two missed cleavages allowed; variable modification -acetylation (N-terminus), methionine oxidation, phosphorylation (S, T, Y), max. four modifications per peptide; peptide tolerance -7 ppm, MS/MS tolerance -0.4 Da; decoy database search -target FDR 0.01. Only high confidence peptides (false discovery rate < 1%) with better than 7 ppm precursor mass accuracy and at least one distinct peptide per protein met identification criteria.
(3) Data comparison Quantitative differences in protein abundance between DEX-and mock-treated samples were determined by spectral counting (Neilson et al., 2011) and further manually validated by comparison of respective peptide peak areas (Qual Browser 2.0.7, Thermo Scientific). Quantitative differences were deemed significant if there was an absolute DEX/mock ratio ≥1.5, with t-test p-values <0.05.

Metabolome analysis
(1) Extraction Briefly, two biological replicates, each consisting of approximately 100 Arabidopsis seedlings cultivated as described above, were pooled and analyzed in three technical replicates. Frozen seedlings (150 mg, approximately 200 seedlings) were homogenized in an MM 400 mill (Retsch). Metabolites were extracted with 1 ml methanol/chloroform/distilled water buffer (2.5:1:0.5 [v/v/v]), 4 °C, 8 min; and clarified by centrifugation. The resulting polar phase was separated by adding 0.5 ml of distilled water, divided into two equal parts, dried in a speedvac concentrator (Thermo-Scientific), and stored at -80 °C until GC-MS analysis.

(2) Derivatisation and GC-MS measurement
Samples were dissolved in 20 l methoxyamine hydrochloride in dry pyridine (40 mg/ml) and incubated for 90 min at 30 °C with rigorous shaking. They were then treated with 80 l N-methyl-N-trimethylsilyltrifluoroacetamide spiked with retention time index markers (alkanes C10-C40, 60 l/1 ml; Sigma-Aldrich), and incubated for an additional 30 min at 37 °C. They were then analyzed by GC-TOF-MS using an Agilent 6890 gas chromatograph (Agilent, Böblingen, Germany) coupled to a Pegasus IV TOF mass spectrometer (LECO Corp Inc., St. Joseph, MI, USA). One microliter of sample was injected in split (1:10) mode. For GC separation a HP5-MS capillary column (30 m, 0.25 mm I.D., 25 μm film; Agilent) was used with a 40 min temperature gradient (70 °C for 1 min followed by 9 °C per min gradient to 350 °C). The MS acquisition rate was 20 scans/s in the mass range m/z = 40-600.

(3) Data analysis
The acquired data were analyzed using ChromaTOF software (LECO), which enables automated data processing. Spectra were calibrated to the retention time index markers, the annotations of selected spectra of mock-and DEX-treated samples were manually checked and a reference peak table with ion traces specific for each analyte to be quantified was created. All chromatograms were compared to the reference and the peak areas were calculated. Quantitative differences in metabolite abundance were deemed significant if there was an absolute DEX/mock ratio ≥1.5, with t-test p-values <0.05.
(1) RNA isolation and reverse transcription RNA was prepared from 50 mg seedlings that had been frozen in liquid nitrogen, using TRIzol reagent (Invitrogen) according to the manufacturer's instructions. To remove contaminating DNA, 1 U of DNase I (Top-Bio, Czech Republic) per 1 g RNA was used. The resulting samples were incubated for 45 min at 37 ºC, then denatured at 65 ºC for 15 min. First-strand cDNA was prepared using SuperScript II reverse transcriptase (Invitrogen) and oligo(dT) primer according to the manufacturer's instructions. cDNA synthesis was performed with 4 g of total RNA.
(2) Quantitative PCR qPCR with specific UPL probes (Roche) and primers designed by ProbeFinder Software was performed using a LightCycler® 480 Instrument and LightCycler® 480 Probes Master (both Roche). Three independent biological replicates and three technical replicates were included for each PCR amplification. Expression levels were normalized to four reference genes and fold-changes in transcript levels were calculated using the ∆∆CT Method with verification of similar efficiencies (Livak and Schmittgen, 2001). Presented results are means obtained for biological replicates and corresponding standard deviations. The significance of differences between activated and non-activated plants was evaluated by Student's t-test (p<0.05). For details about primer sequences see supplementary tables.