Whole maize cobs with immature seeds (930 g), approximately 10 days after pollination, were collected in the field and homogenized using a food processor with 1 L of 50 mM Tris/HCl, pH 8.0, at 4°C for 1 h. After a centrifugation at 4,800g, the pellet was re-extracted with the same volume of fresh buffer and centrifuged again. Ballast proteins were precipitated with additions of protamine sulfate (1 g per 10 g of proteins), manganese(II) chloride (final concentration 7.5 mM) and ammonium sulfate (20% saturation, 114 g L^-1) according to published protocol (Šebela et al., 2000). After the addition of each precipitant, the solution was stirred for 10 min and then centrifuged at 4,800 g for 10 min. The supernatant was loaded onto a DEAE-Sepharose (GE Life Sciences) 5.0 cm i.d. × 30 cm column equilibrated with 50 mM Tris/HCl, pH 8.0, containing ammonium sulfate at 20% saturation. The flow-through fraction was desalted using a MiniKros^® hollow fiber tangential flow filtration system with a 10-kDa polysulfone module (Spectrum Laboratories) and loaded on a High Q (Bio-Rad) 1.5 cm i.d. × 20 cm column equilibrated with 50 mM Tris/HCl, pH 8.0. The bound proteins were eluted with a gradient of KCl from 0 to 1.0 M. Fractions showing the zeatin cis–trans isomerase activity were pooled and transferred to 20 mM Tris/HCl buffer, pH 7.4, containing 0.5 M NaCl and 5 mM of each MgCl₂, MnCl₂, and CaCl₂ using Amicon^® Ultra centrifugal filters (10,000 NMWL; Merck Millipore). The sample was then loaded on a Concanavalin A-Sepharose (Sigma–Aldrich) 1.0 cm i.d. × 10 cm column equilibrated with the same buffer and eluted with the buffer containing 1 M methyl α-D-mannopyranoside. The sample was then transferred to 50 mM Tris/HCl, pH 8.0 using the above centrifugal filters.

The enzyme was further purified using an FPLC BioLogic DuoFlow 10 system (Bio-Rad). The sample was first loaded onto a Resource Q 6 mL column (GE Life Sciences) equilibrated with 50 mM Tris/HCl, pH 8.0, and eluted with a linear gradient up to 0.5 M KCl. Active fractions were pooled, transferred to 5 mM K-phosphate, pH 7.0, containing 0.1 M NaCl and 0.5 mM CaCl₂. The recovered proteins were loaded onto a Bio-Scale CHT5-I column (Bio-Rad) and eluted with 0.75 M K-phosphate buffer, pH 7.0. Active fractions were pooled and transferred to 50 mM Tris/HCl, pH 8.0. Concentrated enzyme solution was loaded onto a HiTrap Blue HP column (GE Life Sciences) and eluted with 2 M KCl.

Finally, the purified enzyme was applied to Novex^® isoelectric focusing electrophoresis gels (ThermoFisher Scientific) with pH 3-10 buffers. The gels were run at a constant voltage of 100 V for 1 h, then at 200 V for 1 h and finally at 400 V for an additional hour. The margin lines were excised and silver-stained. Protein bands were excised from the rest of the gel slab in accordance to the staining and crushed by pressing through a syringe several times. Afterward it was extracted overnight with 100 μL of McIlvaine buffer pH 7.5 at 4°C.

After SDS-PAGE followed by Coomassie staining, protein bands of interest were excised from the gel slab and processed for in-gel digestion by a modified trypsin (Šebela et al., 2006). Peptides from the digests were desalted using ZipTip C18 pipette tips (Merck Millipore Ltd.) according to manufacturer’s instructions and finally reconstituted in 10 μl of 0.1% (v/v) trifluoroacetic acid (TFA). Prior to MALDI mass spectrometry (MS) and tandem mass spectrometry (MS/MS), α-cyano-4-hydroxycinnamic acid (a matrix compound) was dissolved to 0.7 mg ml^-1 in a solvent mixture containing 85% (v/v) acetonitrile, 15% (v/v) water, 0.1% (v/v) TFA and 1mM NH₄H₂PO₄. An aliquot (0.5 μl) of the sample solution corresponding to an initial protein amount of around 200 ng was spotted onto the target (MTP AnchorChipTM 384 BC; Bruker Daltonik), immediately mixed with 0.5 μl of the matrix solution and left to dry at laboratory temperature.

MS and MS/MS analyses were performed on an ultrafleXtreme MALDI-TOF-TOF instrument equipped with a LIFT cell and Smartbeam-II laser operating at a repetition rate up to 2 kHz (Bruker Daltonik). All mass spectra were obtained in the reflectron positive ion mode. The mass spectrometer was controlled by flexControl 3.3 software for acquisition and flexAnalysis 3.3 for spectra processing. The accelerating voltages in the ion source for MS and MS/MS analyses were 25 and 7.5 kV, respectively. In MS/MS mode (no collision gas was used), an accelerating potential of 19 kV was applied to fragments (coming from a timed ion gate) in the LIFT cell. The instrument was calibrated externally using peptide standards supplied by the manufacturer. Manual mass spectra acquisitions were done from 2,000 laser shots in the MS mode and 5,000–10,000 shots in the MS/MS mode. The following settings were applied for MS/MS in an AutoXecute method: primary choice mass range of precursors: 750–3000; number of precursor masses: 15; peak intensity: >800; peak quality factor: >30; signal/noise: >7; FAST minimal fragment mass: 250; LIFT: measure fragments only.

Combined MS/MS datasets were processed by flexAnalysis 3.3, uploaded to ProteinScape 3.0 (Bruker Daltonik) and searched against the NCBI non-redundant database with the Mascot 2.2 search engine (Matrix Science, London, United Kingdom). The search parameters were as follows: Viridiplantae (green plants) were set as a taxonomy; trypsin was set as a protease with 1 missed cleavage allowed; carbamidomethylation of cysteine was set as a fixed modification and methionine oxidation as a variable modification; +1 was set as a peptide charge; monoisotopic masses were considered; other settings: instrument – MALDI-TOF-TOF, significance threshold – p < 0.05, peptide mass tolerance – 25 ppm, MS/MS fragment mass tolerance – 0.7 Da.

A maize nucleotide pyrophosphatase/phosphodiesterase (ZmNPP) gene was synthesized by GeneArt service of ThermoFisher Scientific. Minor alterations from the native sequence were implemented as follows: a CAT triplet was placed before the start codon to form an NdeI restriction site, the sequence stretch GCCTCC from the position 107 was changed to GCTAGC to form an NheI restriction site, which allows to clone ZmNPP fragment without predicted signal sequence, and the triplet CTT at the position 901 was changed to TTG to remove a HindIII restriction site. Neither of these changes led to an amino acid substitution. The TGA stop codon was changed to TAA to form a new HindIII restriction site and an adjacent XhoI site was added to yield a sequence TAAGCTTCTCGAG. All changes were made in order to clone the gene into plasmids for expression in E. coli. Gateway attB1 and attB2 sequences were added up- and down-stream of the coding sequence, respectively.

The ZmNPP gene was amplified from synthesized DNA with the following primers: P1_fw (5′-GGAATTCATGGCGTCTCCGCCCCACTC-3′) and P2_rev (5′-GCCGCTCGAGTTTTGTTCGGCAACAGAATCGTGCC-3′) with EcoRI and XhoI restriction sites, respectively, shown underlined in italics. After restrictions with the respective endonucleases, the gene was cloned into a vector pPICZα (ThermoFisher Scientific). The plasmid was then linearized with MssI. Chemically induced transformation to Pichia pastoris X-33 was performed in accordance with the manufacturer’s protocol.

Transformed yeasts were cultivated in a bioreactor Biostat^® Cplus (Sartorius). The vessel was filled with water up to 8.5 L containing 134 g of yeast nitrogen base without amino acids (ThermoFisher Scientific), 400 mL of glycerol and 2 mL of antifoam A (Sigma–Aldrich) and steam sterilized. After cooling down, 1 L of 1 M potassium phosphate, pH 7.0, and 50 mL of filter-sterilized PTM₁ Trace Salts (Pichia Fermentation Process Guidelines, ThermoFisher Scientific) were added and the conditions were set as follows: agitation 400 rpm, air flow 2 vvm, temperature 28°C, and pH 7.0. After 3 h, 0.5 L of an overnight preculture of Pichia pastoris harboring pPICZα::ZmNPP was added and agitation was set to keep the concentration of dissolved oxygen above 20% of saturation. After 24 h of fermentation, 50% glycerol containing 5% PTM₁ Trace Salts was fed continuously to the bioreactor at the rate of 2 mL min^-1. After another 24 h, an induction of protein expression with methanol started, increasing the rate up to 3 mL min^-1, while the feeding with glycerol was decreased linearly. The methanol was fed to the bioreactor for 3 more days.

Yeast cells were then collected by a centrifugation at 4,800g for 30 min. The medium was conditioned with ¹/₄₀ volume of 1 M Tris and 2 M NaCl, run through a DEAE-Sepharose 5.0 cm i.d. × 30 cm column, concentrated and conditioned to 20 mM Tris/HCl, pH 8.0, on a SartoJet pump with Sartocon^® Slice Ultrasart polyethersulfone 10-kDa NMWCO cassettes (Sartorius). The enzyme was further purified on the High Q 1.5 cm i.d. × 20 cm and HiTrap Blue HP columns as described above.

The interconversion of zeatin isomers was measured with 0.2 mM cis-zeatin in 100 mM McIlvaine buffer containing 20 mM MgCl₂, 0.1 mM FAD and 0.8 to 2.0 mM dithiothreitol. Upon mixing, the samples were incubated at 37°C under white fluorescent light (500 μE m^-2 s^-1). The reaction was stopped by the addition of two volumes of methanol after 1 h. For each assay, a control reaction was set-up, where the enzyme sample was boiled for 5 min prior to addition of the other components and the enzymatic activity was calculated from the difference in the concentration of trans-zeatin between the two reactions. Zeatin content was analyzed on a Nexera UFLC (Shimadzu) equipped with a Zorbax RRHD Eclipse Plus C18 column, 2.1 mm i.d. × 50 mm, 1.8 μm (Agilent) thermostated at 40°C. Zeatin isomers were eluted with 15 mM formic acid, set up with ammonium hydroxide to pH 4.0 at the flow rate of 0.4 mL min^-1 with gradient of methanol as follows (min/%): 0.0/22; 3.0/22; 4.0/90; 5.5/90; 6.0/22; 8.0/22 and their content determined from peak areas at 268 nm using standard compounds. A typical chromatogram of the reaction is shown in Supplementary Figure S1A.

The assay of nucleotide pyrophosphatase/phosphodiesterase activity was done in identical manner, but without cis-zeatin and the reaction mixture was kept in dark. Moreover, dithiothreitol was omitted in the assays with purified recombinant protein as no longer needed. The samples were analyzed on the same UFLC system and conditions as above, but eluted with 20 mM phosphoric acid (set up to pH 6.5 with ammonium hydroxide) using the following methanol gradient (min/%): 0.0/15; 1.0/15; 4.0/22; 4.5/25; 5.0/95; 7.0/95; 7.5/15; 10.0/15 and monitored at 449 nm. A typical chromatogram of the reaction is shown in Supplementary Figure S1B. In plant samples, extracted flavins were analyzed by the same procedure but quantified on a FP-2020 Plus fluorescence detector (JASCO) with excitation at 265 nm and emission at 530 nm using standard compounds.

To determine the nucleotide pyrophosphatase/ phosphodiesterase activity with various substrates, the same reaction was set up with 0.5 mM substrate and with an appropriate amount of the purified recombinant enzyme hydrolyzing maximally 40% of the substrate. The reaction was then stopped by pipetting aliquots at time-points 0, 10, 20, and 30 min into two volumes of methanol. The samples were all centrifuged at 21,000 g for 10 min and mixed with 15 mM triethylamine set with phosphoric acid to pH 6.0 in such a ratio to bring methanol to 5% and filtered using 0.22 μm Costar^® Spin-X^® centrifuge tube filters (Corning Inc.) at 10,000g for 3 min. The samples were then analyzed by the UFLC equipped with a Polaris 180Å C18-A, 2.0 mm i.d. × 150 mm, 3 μm (Agilent) thermostated at 35°C using a methanol gradient in 15 mM triethylamine (set up to pH 6.0 with phosphoric acid) at the flow rate of 0.2 mL min^-1. The gradient was set up as follows (min/%): 0.0/10; 1.5/10; 4.0/40; 4.5/60; 7.0/60; 7.5/10; 10.0/10. The analytes were quantified using UV-Vis detector using standard compounds.

To determine enzyme specific activity, protein concentration was estimated using a Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad) with bovine serum albumin as a calibration standard with linearization (Ernst and Zor, 1996).

The gene was cloned into pGWB17 vectors (35S promoter, kanamycin and hygromycin resistance) using Gateway protocol (ThermoFisher Scientific) and electroporated into Agrobacterium rhizogenes strain 15834, which was used to transform tomato as described previously (Šmehilová et al., 2009). The main root was used for propagation; 2-cm long tips of lateral roots were used for the determination of gene expression and activity level and the rest was used for genotyping. Gene expression was determined as described previously (Šmehilová et al., 2009) with primers ZmNPP_fw (5′-CCCCAACCACTACTCCATCGT-3′) and ZmNPP_rev (5′-TCGTGGTTTTTCATGGTGAAGT-3′).

Arabidopsis plants expressing maize nucleotide pyrophosphatase/phosphodiesterase gene cloned into the pGWB17 vector were prepared from Arabidopsis thaliana ecotype Col-0 using a floral-dip method (Clough and Bent, 1998). Homozygous plants were selected by PCR detection of the inserted ZmNPP gene. A plant was considered homozygous when at least 100 offspring plants carried the insert and at the same time there was no negative offspring plant. The plants were grown for 4 weeks in an environmental chamber (16 h fluorescence light intensity of 150 μE m^-2 s^-1/8 h dark, 22°C, 55% relative humidity) in the soil. Third to sixth leaves were collected and pooled from 4 plants into each sample.

Rice (Oryza sativa japonica) T-DNA insertional line PFG_2B-60145.L with the corresponding wild type Hwayoung were purchased from POSTECH, South Korea. The line has an insertion in the promoter sequence of the gene LOC_Os01g10020, 300 bp upstream of the start codon, which contains the promoter sequence of OsTubA1 gene. Rice seeds were germinated in Agroperlit (Perlit); after 2–3 weeks the seedlings were planted into the soil and grown for 2 weeks in an environmental chamber (12 h fluorescence light intensity of 250 μE m^-2 s^-1, 28°C/12 h dark, 25°C, 65% relative humidity). Then shoots were collected from each plant separately with the exception of the albino plants that were pooled from two plants into one sample.

Enzymatic activity and protein content were assayed as above using samples prepared from 0.1 g of the leaves that were crushed in liquid nitrogen and extracted with 0.2 mL of 50 mM Tris/HCl, pH 8.0. Total flavins were first extracted with methanol/methylene chloride (9:10) from 5 mg of starting material as described previously (Hiltunen et al., 2012) then the sample was extracted again with 10 mM sodium phosphate containing 10% (v/v) acetonitrile. The whole procedure was conducted in a dim light. Finally, 10 μL of a mixture of both extracts (1:1) was mixed with 50 μL of 20 mM ammonium phosphate, pH 6.5, and analyzed using UFLC with fluorescence detector as above on the same type of column, but 150 mm long. Chlorophyll content was determined using published method (Porra et al., 1989). Cytokinins in Arabidopsis plants were determined using published method (Svačinová et al., 2012).

When we examined extracts from several plant species including maize, Arabidopsis, common bean, wheat and rice, the highest activity was found in maize, detectable in all developmental stages and organs (results not shown). Maize immature kernels were therefore chosen as the starting material for enzyme purification. As shown in Figure 1A, the rate of conversion of cis- to trans-zeatin was about 5-fold higher than the opposite reaction, both with a relatively high background after boiling that reached up to 25% of the rate with native extracts.

The whole purification procedure was quite tedious and required a series of six chromatographic steps to obtain 0.7 mg of a final enzyme preparation from almost one kilogram of maize cobs with kernels at the stage of liquid endosperm. The progress of individual purification steps with 10% activity yield resulted in 1600-fold purified enzyme with the specific activity of 65 nkat mg^-1 as summarized in Table 1. However, several protein bands were still detectable on an SDS-PAGE gel (Figure 1B). To identify a particular protein showing the isomerase activity, the active fraction from HiTrap Blue was further separated using isoelectric focusing (Figure 1C). The gel slab was then sliced, separated proteins extracted and tested for the enzymatic activity. The zeatin cis–trans isomerase activity was found in the fraction corresponding to a 50 kDa band on the original SDS-PAGE gel (Figure 1D).

MALDI-TOF/TOF MS and MS/MS allowed assigning the purified protein band with in vitro isomerase activity to the accession number NP_001146857 in the NCBInr database. The peptide mass fingerprinting data allowed assigning 19 peptides with m/z values in the range of 805–2816 Da providing a probability-based score of 131 and sequence coverage of 32.3% (Supplementary Figure S2). The MS/MS-based data provided much more convincing score value of 1345 (Supplementary Figure S3). This accession is annotated as a nucleotide pyrophosphatase/phosphodiesterase from Zea mays (ZmNPP; LOC100280465; Zm00001d039454_P001). The protein consists of 468 amino acids with a theoretical molecular mass of 51.2 kDa, which well corresponds to the results obtained by SDS-PAGE, and pI 6.36 determined by Compute pI/Mw (Bjellqvist et al., 1993).

A BLAST analysis against translated plant genomes using ZmNPP sequence as a query returned only a few results per plant. At the time of ZmNPP identification, the database of maize genome assembly contained an additional gene with a high homology in a stretch of 160 amino acids, while the rest of the gene encoded Sir2 domain that possess a NAD-hydrolysis-dependent deacetylase activity. However, at the time of completing this paper (July 2017), the current maize genome assembly (AGPv4) contains three annotated NPP genes with similar length and only minor alterations (indels and several amino acid substitutions, Supplementary Figure S4). Because we were not aware of the other two ZmNPPs, we have focused in our work only on the identified enzyme. For the same reason, we refer to the enzyme as ZmNPP, without the number. Nevertheless, the mass spectrometry analysis (Supplementary Figures S2, S3) unambiguously confirmed that a protein present in the purified fraction is ZmNPP1 (Zm00001d039454_P001) and not ZmNPP2 (Zm00001d047948_P001) or ZmNPP3 (Zm00001d044311_P001). A single gene LOC_Os01g10020 was found in rice and four in Arabidopsis, whose genes are located on chromosome 4, next to each other (At4g29680, At4g29690, At4g29700, and At4g29710).

Unlike human nucleotide pyrophosphatase/phosphodie-sterase genes that contain a large number of introns and undergo an alternative splicing, all plant genes, interestingly with the exception of the additional two ZmNPPs (ZmNPP2 and ZmNPP3), contain a single exon. As shown in Figure 2, plant NPP proteins cluster separately from animal enzymes in the phylogenetic tree. Mostly, when there is more than one encoding gene in the genome (e.g., in Arabidopsis and tomato), the paralogs are located sequentially on the same chromosome. This all suggests that before the evolutional divergence of animals and plants, there was only a single precursor gene. While there were multiple gene duplications in animals before the divergence of clades (Zimmermann et al., 2012), the genes in plants duplicated relatively recently, after the speciation.

The prediction of plant NPPs’ subcellular localization by TargetP (Emanuelsson et al., 2007) is mostly ambiguous. Two exceptions are ZmNPP2 and AtNPP3, both with high confidence to be secreted proteins. However, both of them have also predicted one trans-membrane helix by TMHMM server (Krogh et al., 2001). The other proteins’ localization is predicted with lower confidence with reliability class usually 4 or 5.

The plant NPPs in general contain all the amino acids that were shown to be crucial for the catalysis in animals (Kato et al., 2012) with two notable exceptions: Lys 277 from mouse NPP1 does not align to any specific amino acid in the plant NPPs and the Asp 308 that is crucial for substrate binding via water molecule is substituted with glutamate (Glu 191 in ZmNPP1) as in Xanthomonas enzyme. The plant enzymes are predicted to be glycosylated by NetNGlyc (Gupta and Brunak, 2002) at several asparagine residues. Without ambiguity, all enzymes shall be glycosylated at residues corresponding to Asn 108 and 208 of ZmNPP1 with the exception of AtNPP2 and AtNPP3 that have residue corresponding to Asn 108 substituted for Arg and Gly, respectively. The three Arabidopsis proteins that have asparagine at the position corresponding to ZmNPP1’s Lys 393 shall be glycosylated there. The last position undergoing glycosylation is the one corresponding to Asn 454 of ZmNPP1. At this position, about half of proteins are weakly predicted to be glycosylated, while the rest not to be glycosylated. Lastly, ZmNPP1, AtNPP1 and AtNPP2 shall be glycosylated at positions 35, 33 and 34, respectively. An alignment of all aforementioned plant NPP enzymes with indicated putative glycosylation sites, conserved amino acids and predicted signal peptides is shown in Supplementary Figure S4.

To study the catalytic reaction, heterologous expression was preferred to purification from maize kernel extract for obtaining sufficient amount of ZmNPP. First attempts using Escherichia coli did not lead to successful protein production. Alternatively, the expression was successful using a methanol inducible expression in the methylotrophic yeast Pichia pastoris with the expression vector pPICZα that fuses a signal peptide to the N-terminus of the recombinant protein thus driving its secretion into the growth medium. The recombinant protein was then concentrated and purified to homogeneity (Figure 1E) using three relatively simple chromatographic steps as described in Section “Materials and Methods.” The size of the recombinant protein was around 60 kDa due to the presence of His-tag and a linker (3.5 kDa) and probably due to a heavier glycosylation pattern in yeast.

The purified recombinant protein showed the zeatin cis–trans isomerase reaction in the presence of FAD and light. The enzyme does not significantly speed up cis–trans zeatin conversion in the presence of FMN; however, a non-enzymatic conversion in the presence of FMN is much faster than in the presence of FAD (Supplementary Figure S5). The obtained recombinant ZmNPP was also readily able to hydrolyze a range of typical nucleotide pyrophosphatase/phosphodiesterase substrates as shown in Figure 3 and none of these reactions required light. With the best substrate FAD, enzyme exhibited a specific activity of 496.8 nkat mg protein^-1 and K_m of 32.7 μM (Figure 3). In general, dinucleotides (FAD, diadenosine polyphosphates, NADP⁺, and UDP-Glc) were better substrates than mononucleotides. The hydrolytic reaction produced AMP from FAD, NADP⁺, ATP, and diadenosine tetra- and pentaphosphate. Further hydrolysis of produced adenosine polyphosphates was observed only after a prolonged incubation. Hydrolysis of other nucleoside triphosphates also led to monophosphate products.

In order to quickly verify ZmNPP activity in planta, the ZmNPP gene was first expressed in a tomato hairy root system (Collier et al., 2005) and several independent transgenic lines were propagated by main root excision. The ZmNPP expressing lines showed 2–10 times higher activity of nucleotide pyrophosphatase/phosphodiesterase with FAD (assayed as the production of FMN) compared to wild type and heat inactivated samples. Similarly, zeatin cis-to-trans isomerization rate measured in vitro in the extracts from various transgenic lines was higher as shown in Figure 4A. There was a strong correlation (R² = 0.9655) between zeatin isomerization and nucleotide pyrophosphatase/phosphodiesterase activity, which was, at the same time, independent of the presence of cis-zeatin (Figure 4B).

Further, Arabidopsis plants overexpressing the ZmNPP gene under 35S promoter were prepared. The specific activity of nucleotide pyrophosphatase/phosphodiesterase with FAD as the substrate increased in the extracts from leaves of 4-week-old Arabidopsis plants by an order of one magnitude compared to wild type (Table 2). Accordingly, the transgenic plants had decreased levels of endogenous FAD and increased levels of FMN and riboflavin (Figure 4C). However, no significant changes in the levels of endogenous zeatin isomers were observed. The levels of zeatins’ free bases in green and senescent leaves of Arabidopsis are depicted in Supplementary Figure S6. The plants also did not show any visible phenotypic changes or changes in the chlorophyll content (Table 2).

To assess the function of putative rice nucleotide pyrophosphatase/phosphodiesterase LOC_Os01g10020, 2-week-old rice plants in which the expression of encoding gene was enhanced by disrupting a native promoter of the gene with the promoter sequence of α-tubulin gene OsTubA1 were examined. Originally, this only available T-DNA insertion mutant of NPP gene in monocot species was purchased in order to study loss-of-function phenotype; however, it later turned out that the insertion of T-DNA to the promotor region led to its overexpression rather to disruption. Approximately one fourth of the obtained plants were albinos, in which chlorophyll content was below the detection limit and the nucleotide pyrophosphatase/phosphodiesterase activity increased about 50-times compared to the wild type, while green mutant plants showed only about 10-times activity increase. The total flavin content decreased about 3-times in albinos, but remained unchanged in green mutant plants (Table 2).