Arabidopsis Plastidial Folylpolyglutamate Synthetase Is Required for Seed Reserve Accumulation and Seedling Establishment in Darkness

Interactions among metabolic pathways are important in plant biology. At present, not much is known about how folate metabolism affects other metabolic pathways in plants. Here we report a T-DNA insertion mutant (atdfb-3) of the plastidial folylpolyglutamate synthetase gene (AtDFB) was defective in seed reserves and skotomorphogenesis. Lower carbon (C) and higher nitrogen (N) content in the mutant seeds than that of the wild type were indicative of an altered C and N partitioning capacity. Higher levels of organic acids and sugars were detected in the mutant seeds compared with the wild type. Further analysis revealed that atdfb-3 seeds contained less total amino acids and individual Asn and Glu as well as NO3 −. These results indicate significant changes in seed storage in the mutant. Defects in hypocotyl elongation were observed in atdfb-3 in darkness under sufficient NO3 − conditions, and further enhanced under NO3 − limited conditions. The strong expression of AtDFB in cotyledons and hypocotyl during early developmental stage was consistent with the mutant sensitivity to limited NO3 − during a narrow developmental window. Exogenous 5-formyl-tetrahydrofolate completely restored the hypocotyl length in atdfb-3 seedlings with NO3 − as the sole N source. Further study demonstrated that folate profiling and N metabolism were perturbed in atdfb-3 etiolated seedlings. The activity of enzymes involved in N reduction and assimilation was altered in atdfb-3. Taken together, these results indicate that AtDFB is required for seed reserves, hypocotyl elongation and N metabolism in darkness, providing novel insights into potential associations of folate metabolism with seed reserve accumulation, N metabolism and hypocotyl development in Arabidopsis.


Introduction
The role of seeds is to propagate offspring. In Arabidopsis thaliana, seed development can be divided into three stages: cell division or the pre-storage phase, maturation or the storage phase, and the desiccation phase [1,2]. Large quantities of carbon (C) and nitrogen (N) are stored in maturing seeds, mainly in the form of large insoluble compounds [3]. The major storage compounds that accumulate in mature seeds are triacylglycerols (TAGs) and seed storage proteins (SSPs), accounting for 30-45% of the seed dry weight. Small amounts of carbohydrate in the form of sucrose are stored within cotyledons [1,3,4,5,6]. SSPs, including soluble proteins and non-soluble proteins, include two predominant classes, namely, 12S globulin and 2S albumin [1,4,7]. Seed storage accumulation is regulated by many factors, such as hormones, sugars, master regulator genes and transcriptional factors [1]. These seed reserves are used to fuel germination and post-germinative seedling establishment until seedling photosynthesis autotrophy can be efficiently established [4].
Seed germination and post-germinative seedling establishment are metabolically distinct [8,9]. Germination initiates with release from dormancy and seed imbibition and is completed when the radicle emerges through the seed coat [10]. At the beginning of germination, seed reserves other than lipids (TAG) are rapidly converted to soluble metabolites (e.g. glycolysis products, organic acids, and amino acids) that can be transported throughout the seedling to support growth, while the breakdown of seed oil storage TAG is used for subsequent seedling establishment after the radicle has emerged [3,8,9,11,12,13]. Following germination, TAG is broken down to yield free fatty acids (FAs) and glycerol, both of which are ultimately converted to sugars required for postgerminative seedling development [6,12]. The sdp1 mutant, containing a mutation in sugardependent1 (SDP1), which encodes a patatin domain TAG lipase that initiates TAG breakdown in germinating seeds, displayed slightly delayed seed germination and a much slower post-germinative growth rate than the wild type [14]. Seedlings grown in darkness showed skotomorphogenesis, which is characterized by elongated weak hypocotyls, closed cotyledons, and shortened roots [15]. Seedling establishment and hypocotyl elongation are driven by the catabolism of TAG under dark conditions. Mutants (icl and pck1) defective in TAG mobilization show shortened hypocotyls in darkness, but hypocotyls could be rescued by providing alternative C sources, such as sucrose [16,17]. N metabolism is also essential for hypocotyl growth. In conifer plants grown in the dark, a portion of N mobilized from the megagametophyte is diverted toward the hypocotyl shortly after germination to produce high levels of Asn, which serves as a reservoir of N to meet subsequent specific developmental demands [18].
Tetrahydrofolate (THF) and its derivatives are collectively called folates. Most cellular folates carry a short poly-c-Glu tail, which is believed to affect their efficacy and stability. The tail can be removed by c-glutamyl hydrolase (GGH), a vacuolar enzyme who has an important influence on polyglutamyl tail length and hence on folate stability and cellular folate content [19]. Folylpolyglutamate derivatives are central cofactors for many folate-dependent enzymes [20,21,22,23,24,25]. During the germination process, de novo synthesis of THF occurs in pea (Pisum sativum) cotyledons, and the inhibition of THF de novo synthesis using folate analogs blocks seedling development [26,27,28]. The cotyledonary folate pool contains principally methylated derivatives [28]; the concentration of folylpoly-Glu derivatives increases gradually during germination [29], and the accumulation of folates peaks 3 days after sowing [27].
Plants with defective folate biosynthesis and metabolism showed various aberrant seed and seedling phenotypes. For example, the globular arrest1 (gla1) mutant, which contains a mutation in dihydrofolate synthetase folylpolyglutamate synthetase (DHFS-FPGS) homolog A (DFA), encoding a functional mitochondrial matrix-localized dihydrofolate (DHF) synthetase, exhibited defective embryonic development and did not undergo transition to the heart stage [23,30]. The double knockout (dKO) mutation of 10formyl-THF deformylase genes, At4g17360 and At5g47435, resulted in defective embryo development, with cells arresting between the heart and early bent cotyledon stages. Mature seeds of dKO were shriveled, accumulated low amounts of lipids, and failed to germinate [31]. A mutation in AtDFB, which encodes the plastidial folylpolyglutamate synthetase (FPGS) isoform, displayed short primary roots with a disorganized quiescent center [24,32]. A mutation in AtDFC, which encodes the mitochondrial FPGS, was characterized based on its altered N metabolism and enhanced phenotypes to low N stress, providing novel insights into folate biosynthesis and N utilization during early seedling development [33]. To date, the role of folate during skotomorphogenesis in plants remains poorly understood.
In this report, a mutant (atdfb-3) carrying a T-DNA insertion in the AtDFB gene was characterized for its altered seed reserves and defective seedling establishment with shortened hypocotyls under dark conditions. Early post-germinative growth (before 3 days) in atdfb-3 required external NO 3 2 sufficient conditions, and exogenous application of 5-formyl-tetrahydrofolate (5-F-THF) restored hypocotyl length in atdfb-3 when NO 3 2 was as the sole N source in the medium. The defective hypocotyl elongation could be due to altered seed storage, perturbed folate and N metabolism in atdfb-3. This report provides novel insights into a potential associations of folate metabolism with seed reserve accumulation, N metabolism and hypocotyl development elongation in darkness in Arabidopsis.

Results
Reduced seed size and altered C/N partition capacity in mature seeds of atdfb-3 A previous report demonstrated that the vegetative phenotype of atdfb (fpgs1, SALK_133817) did not differ visually from the wild type under light conditions [34]. In this report, SALK_015472 with a T-DNA insertion in the sixth intron of At5g05980 ( Figure S1A in File S1), which encodes the plastidial isoform of FPGS (AtDFB), was obtained from the Arabidopsis Biological Resource Center (The Ohio State University) and named atdfb-3, as described previously [32].
First, characteristics of seeds harvested from atdfb-3 and wildtype plants grown under light conditions, such as seed number per silique and 1000 seed weight, were examined. No significant difference was observed in seed number per silique between atdfb-3 and the wild type ( Figure 1A); however, the width and length of mature atdfb-3 seeds were slightly but significantly smaller than those of the wild type ( Figure 1B). The reduction in dimensions was somewhat reflected by the seed weight, with a significant decrease of 5% in atdfb-3 compared with the wild type ( Figure 1C).
Next, we explored seed reserves in mature atdfb-3 seeds. We found that C and N levels in atdfb-3 seeds were 94% and 122%, respectively, compared with wild-type levels (Figure 2A and B). In AtDFB complemented plants ( Figure S1B and C in File S1), these changes were restored to wild-type levels ( Figure 2A and B), indicating they were due to the loss of function of AtDFB. These results indicated the altered C and N partitioning capacity observed in atdfb-3 was due to the loss of function of AtDFB.

Altered C and N metabolites in atdfb-3 seeds
We analyzed metabolites in atdfb-3 and wild-type seeds using gas chromatography time-of-flight mass spectrometry (GC-TOF-MS). A total of nine metabolites, including two FAs (14:0 and 18:3), three organic acids (oxalic acid, pentanedioic acid, and phosphoric acid), two sugars (galactose and mannose), and two polyols (campesterol and phytol) were higher and three metabolites (20:1, benzoic acid, and lyxose) were lower in atdfb-3 than in the wild type ( Figure 3). In addition, the contents of other metabolites (mainly fatty acids) detected in atdfb-3 were similar to those in the wild type. These results suggested that the mutation in AtDFB altered C-rich metabolites accumulation in mature seeds.
The level of soluble protein was not significantly different from the wild type in atdfb-3 ( Figure 4A), while the total free amino acids were significantly less in atdfb-3 seeds than in the wild type ( Figure 4B). Many individual amino acids accumulated to lower levels in atdfb-3 than in the wild type, such as Asn, Glu, Asp, Cys, Gly, Pser, Pro, and His. Asn and Glu were both 50% less than the wild type, accounting for the main shortage of total amino acids in atdfb-3 ( Figure 4C). In contrast, some other amino acids accumulated more in atdfb-3 than in the wild type, such as Gln, Phe, Leu, Ile, Met, b-Aiba, b-Ala, Lys, and c-Aba. Among them, Leu in atdfb-3 accumulated to the highest level: 2.5-fold higher than that of the wild type ( Figure 4C). As a result, total amino acids in atdfb-3 seeds were 27% lower than in the wild type, and the Gln/Glu ratio in atdfb-3 (0.62) was higher than that of the wild type (0.19). Additionally, the NO 3 2 content in atdfb-3 seeds was only 21% of that in the wild type (19.6 mg g 21 in atdfb-3 vs 95.4 mg g 21 in the wild type; Figure 4D). These results indicated that the mutation in AtDFB reduced the accumulation of N-rich metabolites.
Sufficient N supply (but not C) was required for early post-germinative growth of atdfb-3 in darkness The post-germinative growth of the mutant was investigated in the dark under various N conditions. After growing on halfstrength MS medium (30 N) for 6 days, atdfb-3 had shortened hypocotyls and primary roots as well as expanded cotyledons and a larger apical hook curvature than the wild type ( Figure 5A and B). Similar results were obtained when ammonium (NH 4 + ) was omitted and 9.4 mM or 3 mM NO 3 2 (9.4 N or 3 N, respectively) was added to the medium ( Figure 5A). Interestingly, when the amount of NO 3 2 in the medium was decreased further (less than 3 mM), there were no significant changes in the lengths of hypocotyls of the wild-type seedlings, but the mutant displayed even shorter hypocotyls ( Figure 5A and B). When the medium was supplemented with 0.3 mM NO 3 2 (0.3 N) or 0 N, these hypocotyl and primary root phenotypes of atdfb-3 differed significantly from those of the wild type. The cotyledons of atdfb-3 were folded similarly to those of the wild type; however, the apical hook curvature in atdfb-3 appeared larger than that in the wild type ( Figure 5A). Next, we explored the hypocotyl phenotype further. Interestingly, when NH 4 + was used as the sole N source in the medium, it could not be used for hypocotyl development in atdfb-3 (unlike NO 3 2 ), whether at higher (9.4 and 3 mM) or lower (1 and 0.3 mM) concentrations ( Figure S2A and B in File S1). In addition, the mutant could not utilize organic nitrogen Asn or Gln in the medium under dark conditions ( Figure S2C in File S1). These results indicated that atdfb-3 was sensitive to external NO 3 2 concentrations during skotomorphogenesis, and the hypocotyl  Mutant and wild-type seedlings were also grown on media with other nutrient deficiencies. When grown on C-free medium, there was no obvious difference in hypocotyl length between atdfb-3 and the wild type ( Figure 5C). On phosphate (P)-free medium in darkness, the hypocotyl length of atdfb-3 was 80% compared with that of the wild type ( Figure 5C); the ratio was similar to that under N-sufficient conditions (3 mM or higher NO 3 2 concentration). These results indicated atdfb-3 had a specific response to the external NO 3 2 supply (but not C or P) during seedling establishment in darkness.
Since hypocotyl elongation in atdfb-3 was significantly inhibited by 0.3 N (0.3 mM NO 3 2 ), and the phenotype of the mutant on 9.4 N was similar to that on 1/2 MS, 0.3 N and 9.4 N were used as Nlimited and N-sufficient conditions, respectively, in subsequent experiments. Both N conditions were used in our previous report for N limitation analysis in Arabidopsis [33]. The epidermal cell length in atdfb-3, as measured using a field emission scanning electronic microscope (FE-SEM), was approximately 86% compared with the wild-type cell length on 9.4 N (396.8660.6 mm and 462.0635.8 mm, respectively) and 47% compared with the wildtype cell length on 0.3 N (226.2636.5 mm and 484.2635.6 mm, respectively) ( Figures 5D and E). These observations demonstrated that atdfb-3 was defective in hypocotyl cell elongation in darkness.
To explore why atdfb-3 was sensitive to external NO 3 2 concentrations, the stage at which N-sufficient conditions were required for atdfb-3 hypocotyl development was investigated by removing NO 3 2 from the medium at various time points after sowing. Seedlings first grown on 9.4 N for 0 to 6 days were transferred to 0.3 N for the remaining days, for a total growth time of 6 days. A significant difference in hypocotyl length between atdfb-3 and the wild type was observed when seedlings grown on 9.4 N for 2 days before transferring to 0.3 N ( Figure 5F). Hypocotyl length of atdfb-3 first grown on 9.4 N for 3 days or longer time and then transferred to 0.3 N was similar to that of the mutant grown on 9.4 N for 6 days ( Figure 5F). These results indicated that N-sufficient conditions were important for atdfb-3 during the first 3 days. In further time-course experiments, the hypocotyl length of seedlings grown under N-limited conditions and then transferred to N-sufficient conditions was shorter than those continuously grown on 9.4 N for 6 days and longer than those continuously grown on 0.3 N for 6 days ( Figure 5G). The less time atdfb-3 was grown on 0.3 N before transferring to Nsufficient conditions, the longer the hypocotyls ( Figure 5G). The hypocotyl in 6-day-old wild-type seedlings grown on 0.3 N for 2 to 5 days before transferring to 9.4 N was longer than those continuously grown on 0.3 N or 9.4 N for 6 days ( Figure 5G); however, this phenomenon was not observed when wild-type seedlings were transferred from N-sufficient to N-limited conditions ( Figure 5F). It is possible that the transferring from N-limited to N-sufficient conditions stimulated hypocotyl elongation in the wild type. These results indicated that N-sufficient conditions were required for early growth of atdfb-3, and that the response of atdfb-3 to low N stress occurred within a narrow developmental window (3 days or earlier).
The endosperm in atdfb-3 was removed under both N conditions to explore whether the storage in embryo or endosperm in atdfb-3 was altered, which would affect hypocotyl development under dark conditions. When the endosperm was removed, wildtype hypocotyls were slightly shorter than those with endosperm present under both 9.4 N and 0.3 N conditions ( Figure 5H). However, the hypocotyls in atdfb-3 without endosperm under 9.4 N were only 30% of those in atdfb-3 with the endosperm, and even shorter than those in atdfb-3 with endosperm under 0.3 N ( Figure 5H). Defects in hypocotyl elongation were more significant in atdfb-3 without endosperm than that with endosperm under both N conditions ( Figure 5H). These results indicated that, unlike the wild type, the embryo alone could not satisfy hypocotyl development in atdfb-3 under N-sufficient conditions, and that the endosperm was vital for atdfb-3 hypocotyl development. Under Nlimited condition, external N could not satisfy hypocotyl development of atdfb-3 even with the endosperm. The requirement of sufficient NO 3 2 during early hypocotyl development in atdfb-3 could be due to altered seed storage in embryo.

AtDFB was expressed in early developmental stage in Arabidopsis
Since AtDFB is important for early seedling establishment under dark conditions, we investigated the expression pattern of AtDFB during the early stage in etiolated seedlings to illustrate its importance in hypocotyl development. A plasmid containing an AtDFB promoter-driven GUS fragment was introduced into the wild-type plants ( Figure 6A). Histochemical GUS staining showed that AtDFB was strongly expressed in cotyledons and hypocotyls in 2-day-old germinating seeds, while it was strongly expressed only in cotyledons in 3-day-old etiolated seedlings ( Figure 6B). The expression pattern of AtDFB under light was similar to that in the dark ( Figure S3 in File S1). These results indicated that AtDFB was expressed in early seedling developmental stage.
Rescue of atdfb-3 seedling establishment by exogenous 5-F-THF depended on the nitrate supply 5-F-THF is able to rescue the atdfb defects in primary root development in light [27]. We added various concentrations of 5-F-THF or 5-methyl-tetrahydrofolate (5-M-THF) to 0.3 N medium to determine whether folate derivatives could rescue defects in hypocotyl elongation in atdfb-3 in dark. The difference in hypocotyl length between the mutant and wild-type seedlings was reduced when grown on 0.3 N with 0.5 mM or 5 mM 5-F-THF, while treatment with 50 or 500 mM 5-F-THF rescued hypocotyl elongation in atdfb-3 seedlings ( Figure S4A in File S1). 5-M-THF stimulated hypocotyl elongation in both atdfb-3 and wild-type seedlings. Interestingly, under N-limited conditions, disparities between the mutant and wild type decreased with increasing 5-M-THF concentrations, but the hypocotyl length in atdfb-3 was still only 67% of the wild type when grown with 500 mM 5-M-THF ( Figure S4B in File S1). These results indicated that 5-M-THF can only partially rescue the hypocotyl elongation  in atdfb-3. Thus, 50 mM 5-F-THF was chosen to rescue the hypocotyl elongation defect in atdfb-3 under dark conditions.
The hypocotyl length of 6-day-old atdfb-3 seedlings was restored to the wild-type level under both N-sufficient and Nlimited conditions with 50 mM 5-F-THF ( Figure 7A). Further analysis indicated that 5-F-THF could rescue the hypocotyl elongation defects in atdfb-3 seedlings under N-limited conditions at various experimental time points ( Figure 7B). We next examined the stages at which folate was vital for hypocotyl elongation. Both atdfb-3 and the wild type seeds were grown on Nlimited conditions with 5-F-THF for 0 to 6 days and then moved to conditions without 5-F-THF for the remaining time, for a total growth time of 6 days. We found that the mutant seedlings grown on medium with 5-F-THF for only 2 days and then transferred to conditions without 5-F-THF could adapt to low-N conditions, demonstrating the same hypocotyl length as the wild type at day 6 ( Figure 7C). However, atdfb-3 seedlings grown on medium with 5-F-THF for 1 day before transferring to conditions without 5-F-THF could not adapt to N-limited conditions, similar to those without 5-F-THF treatment under N-limited conditions ( Figure  S4C in File S1). Additionally, the mutant seedlings grown on medium without 5-F-THF for 2 or more days before transferring to medium with 5-F-THF showed shorter hypocotyls than did the wild-type seedlings; the longer time the mutant was grown on medium without 5-F-THF before transferring to medium with 5-F-THF, the shorter the hypocotyls ( Figure 7D). These results suggested that intact folate metabolism was necessary for early (2 days or earlier) developmental stages in Arabidopsis.
When no N was applied to the medium, 5-F-THF could not rescue hypocotyl defects in atdfb-3. Meanwhile, 5-F-THF could not restore hypocotyl defects in atdfb-3 when NH 4 + was the sole N source in the medium ( Figure S5A and B in File S1). These results indicated that the recovery of hypocotyl development in atdfb-3 by 5-F-THF depended on exogenous NO 3 2 supply.
Folate metabolism was altered in atdfb-3 germinating seeds and etiolated seedlings To increase our understanding of how folate metabolism was perturbed in the mutant, liquid chromatography/mass spectroscopy (LC-MS) was employed to profile various folate derivatives in early developmental stage of 2-day-old germinating seeds. We found that 5-M-THF was the major folate derivative, accounting for 70% of the total folates ( Figure S6 and Table S1 in File S1). Under N-sufficient conditions, the mutant contained less 5-F-THF, 5-M-THF, and total folates than the wild type (approximately 30%, 80%, and 75% of the wild type, respectively). Under N-limited conditions, the contents of most folate derivatives decreased in the wild type, e.g. 67% reduction for 5-F-THF and 35% for 5-M-THF, respectively, but remained unchanged in atdfb-3 ( Figure S6 and Table S1 in File S1).
To further determine how the AtDFB mutation interferes with folate metabolism in dark-grown seedlings, various folate derivatives and poly-glutamylated 5-M-THF and 5-F-THF were examined in 6-day-old etiolated seedlings (Figure 8). First, we found that there was a difference in folate derivative contents between the mutant and wild type. Under N-sufficient conditions, the mutant contained less 5-F-THF, 5-M-THF, and total folates than the wild type (approximately 70%, 36%, and 51% of the wild type, respectively). 5-M-THF, the major folate derivative, constituted the major deficiency in total folates in atdfb-3 seedlings. Under N-limited conditions, higher accumulation of folate derivatives including 5-F-THF and DHF was observed in atdfb-3 than that of the wild type, which was opposite to that under Nsufficient conditions; however, 5-M-THF remained less in atdfb-3 than in the wild type ( Figure 8A and Table S2 in File S1). Second, we found that the folate derivatives profiling of the mutant and wild type responded differentially to N limitation. For example, N limitation led to a 50% decrease in 5-F-THF in the wild type, but had no effects on atdfb-3, resulting in a 1.6-fold accumulation of 5-F-THF in atdfb-3 as compared to the wild type. N limitation led to no significant decrease in total folates in the wild type, but a 44% increase in atdfb-3 seedlings, resulting in a drastic reduction of the difference from 49% to 16% between the mutant and wild type ( Figure 8A and Table S2 in File S1).
The levels of polyglutamated folates with 5-, 6-, 7-, and 8-Glu tails were compared between the mutant and wild type based on relative peak areas due to a lack of standards. There was a significant difference between the two genotypes. Under Nsufficient conditions, most striking difference was observed for both 5-M-THF-Glu7 and 5-F-THF-Glu7, i.e. higher accumulation in atdfb-3 than the wild type. In addition, 5-M-THF-Glu6 was less and 5-F-THF-Glu8 was higher in atdfb-3 than that of the wild type, respectively ( Figure 8B). Under N-limited conditions, both 5-M-THF-Glu7 and 5-F-THF-Glu7 remained higher in the mutant than in the wild type as observed under N-sufficient conditions. 5-M-THF-Glu5 and 5-M-THF-Glu8 was around 19 folds less and 2.9 folds higher in atdfb-3 than in the wild type, respectively ( Figure 8B). Moreover, the folate derivatives with polyglutamates in the mutant and wild type differed in responding to N limitation. For example, N limitation led to a 3.1-fold increase of 5-M-THF-Glu5 in the wild type, but a 6.1-fold decrease in atdfb-3. The pattern of the folylpolyglutamation profile in the complemented transformants was similar to that in the wild type ( Figures 8B).
We also analyzed the expression of genes involved in folate biosynthesis and C1 metabolism ( Figure S7 in File S1). The expression of AtDFA and AtDFC were enhanced in 2-day-old atdfb-3 seedlings due to loss function of AtDFB especially under 0.3 N, but not in 6-day-old seedlings ( Figure S7A and B in File S1). Most of these genes, including AMINODEOXYCHORISMATE LYASE (ADCL), 10-FORMYL-THF DEFORMYLASE 2 (FDF2), 5-FORMYL-THF CYCLOLIGASE (5-FCL), 10-FORMYL-THF SYNTHETASE (THFS), c-GLUTAMYL HYDROLASE 1 (GGH1) and GGH2, showed higher expression in atdfb-3 than in the wild type under both N conditions. Low N stimulated the expression of ADCL, PDF2 and MTHFR2 in both genotypes, but only that of 5-FCL in the mutant and THFS in the wild type, respectively ( Figure S7C in File S1).

N metabolism was affected in atdfb-3 germinating seeds and etiolated seedlings
Under N-sufficient conditions, the C and N contents in atdfb-3 were unchanged compared with the wild type, while the N content in atdfb-3 increased by 9% under N-limited conditions ( Figure 9A and B). In addition, the mutant accumulated 23% more soluble protein under 9.4 N and 32% more under 0.3 N than did the wild type ( Figure 9C). These results indicated that N metabolism in germinating atdfb-3 seeds was altered under N-limited conditions. Because atdfb-3 showed a failure of seedling establishment when grown on low N, we sought to understand how N metabolism was affected by the AtDFB mutation by analyzing N-relating metabolites and enzyme activities. Under N-sufficient conditions, there was no significant difference in NO 3 2 and NO 2 2 contents between atdfb-3 and the wild type; however, under N-limited conditions lower level of NO 3 2 and higher level of NO 2 2 were detected in atdfb-3 than in the wild type ( Figure 10A and B), whereas no significant difference in NH 4 + contents was observed ( Figure S8 in File S1). Subsequently, activities of the enzymes involved in N metabolism, such as nitrite reductase (NiR), glutamine synthetase (GS), and glutamine 2-oxoglutarate amino transferase (GOGAT) were investigated under both N conditions. NiR activity was lower in atdfb-3 than in the wild type under both 9.4 N and 0.3 N, about 51% and 61% of the wild type, respectively ( Figure 10C). Under 9.4 N or 0.3 N, GS activity in atdfb-3 was 83% or 85% of the wild type, respectively ( Figure 10D). The GOGAT activity did not differ significantly between atdfb-3 and the wild type under 9.4 N, while that in atdfb-3 was only half of the wild type under 0.3 N ( Figure 10E). These results implied that activity of the enzymes involved in N reduction and assimilation was altered in atdfb-3 etiolated seedlings.
Next, transcripts of genes involved in NO 3 2 transport, NO 3 2 reduction, and N assimilation were also examined in atdfb-3. When N was sufficient, atdfb-3 had lower level of expression of NITRATE TRANSPORTER 1.1 (NRT1.1) than the wild type; however, transcripts of GS1:1 and GS2 were more abundant in atdfb-3 than in the wild type. Low-N stress increased the expression level of NRT1.1 and GS1:4 in both genotypes ( Figure 10F). Taken together, these results indicated that N metabolism was perturbed in atdfb-3.

Discussion
Under light conditions, mutation in the AtDFB resulted in a short primary root due to perturbed folate profile [32]. In this report, atdfb-3 was characterized for its defects in seed reserves and hypocotyl elongation in the dark due to loss of function of AtDFB, providing novel insights into a potential link among folate metabolism, seed reserves, and hypocotyl development in Arabidopsis.

AtDFB mutation altered seed storage
Seed storage compound synthesis and accumulation in maturing seeds of Arabidopsis are under the control of many factors, such as hormones, sugars, master regulator genes, and transcription factors [4]. Folates also play an important role in seed development, since gla1 and fpgs1fpgs2 exhibited defective embryo development [30,34]. The double knockout (dKO) mutation of 10-formyl-THF deformylase resulted in shriveled seeds and low amounts of lipids, such as 20:1 [31]. A slightly but significantly lower level of 20:1, one of the markers for storage oil in Arabidopsis [35], was also detected in atdfb-3 than in the wild type (Figure 3), probably indicative of a low oil storage in the mutant.
Arabidopsis mutants that have defective seed storage mobilization had shorter hypocotyls in the dark [16,17], and atdfb-3 also showed shortened hypocotyls ( Figure 5A). There is a possibility that high level of mannose in atdfb-3 seeds reduced storage mobilization rate, as exogenous mannose did greatly reduce the rate of storage lipid mobilization in germinating Arabidopsis seeds [36]. It was also reported that galactose that accumulated during seed maturation could provide easily available energy and also be an important component of the sugar signaling pathway during germination of pea seeds [9,37]. Thus, high content of galactose accumulated in atdfb-3 seeds might promote germination while mannose inhibits storage mobilization. Besides, mannose and galactose are intermediates of ascorbic acid biosynthesis [38]. Therefore, it needs further investigation that whether the accumuation of mannose and galactose is due to deficient ascorbic acid biosynthesis in atdfb-3.
It was also observed that oxalic acid was accumulated in atdfb-3 seeds (Figure 3). Oxalate could be produced by glycolate or glyoxylate during photorespiration, or by the breakdown of ascorbic acid [39,40]. Previously we reported that the mitochondrial AtDFC is involved in regulation of N metabolism in Arabidopsis by linking folate metabolism with photorespiration [33]. However, it seems unlikely that the oxalate accumulation is due to photorespiration alteration given the fact that Gly/Ser ratio, an indicator for photorespiration, was not changed in the mutant ( Figure 4C). Oxalate accumulation was observed both in atdfb-3 and the mutant of oxalyl-CoA synthetase, an enzyme that catalyzes the first step of oxalate catabolism [41]. However, it remains unclear that whether the oxalate accumulation in atdfb-3 seeds is attributable to decreased oxalate catabolism. Pentanedioic acid was dramatically increased in the atdfb-3 seeds (Figure 3), but it is unknown yet how folates affect its biosynthesis in plants to date. In addition, altered levels of many individual amino acids in atdfb-3 verified the role of folates in amino acids metabolism [31,32,33,42,43].
Folate biosynthesis and polyglutamylation were responsive to low N stress Low N stress enhanced the expression of folate synthesis and metabolism genes (such as AtDFA and THFS) in the mutant ( Figure S7A in File S1). As a result, similar contents of total folates in the mutant and wild type were achieved ( Figure S6 in File S1). This is different from our previous report that the total folates level in atdfc remained lower than in the wild type under N limitation [33]. Under N-sufficient conditions, hypocotyls of 6-day-old etiolated atdfb-3 seedlings were only slightly shorter than the wild type; however, under N-limited conditions, the mutant had significantly reduced hypocotyls ( Figure 5A and B). In association with this, 5-M-THF-Glu5 and 5-F-THF-Glu7 in atdfb-3 was 0.05and 6-fold of that in the wild type under N-limited conditions, respectively ( Figure 8B). Considering that 5-M-THF-Glu5 and 5-F-THF-Glu7 were the most changed folate derivative and responded in an opposite manner when the mutant was subjected to low N, we hypothesize that 5-M-THF-Glu5 and 5-F-THF-Glu7 may, at least in part, play an important but contrasting role in regulation of hypocotyl development. Meanwhile, we could also speculate that there is not a causal relationship between total folate level and shortened hypocotyl length in the mutant under N limitation because the mutant and wild type had a similar level of total folates (Figure 8A and S6). Exogenous nitrate was required for the rescue of atdfb-3 mutant by 5-F-THF application Seed reserves are mobilized to fuel seedlings until autotrophic growth. Given the massive reduction in nitrate level in atdfb-3 ( Figure 4D), it was reasonable that the mutant requires sufficient external N. 5-F-THF could rescue the defects in hypocotyl elongation in atdfb-3 only in the presence of NO 3 2 , but not NH 4 + in the medium ( Figure S5 in File S1). Owing to the observation of reduction of GS and GOGAT activity under N-limited conditions (Figure 10), NH 4 + assimilation is probably deficient in atdfb-3. Thus, the mutant could not utilize the sole N source NH 4 + to fuel seedling development. Additionally, exogenous 5-F-THF failed to rescue the mutant when NO 3 2 was absent ( Figure S5 in File S1), demonstrating an absolute necessity of NO 3 2 for seedling establishment in atdfb-3.
The exogenously supplied 5-F-THF was probably absorbed and converted into other active forms of folate by 5-FCL, MTHFR, or other enzymes that convert folate derivatives [44,45]. The recovery of hypocotyl elongation in atdfb-3 treated with 5-F-THF was probably due to the excess of folates in the monoglutamylated form, which might be able to complement the mutation by accomplishing the same functions as a small amount of polyglutamylated folates. Given the fact that microtubule cytoskeleton plays a crucial role during hypocotyl elongation [46], and abnormal actin cytoskeleton was observed in atdfb primary root [32], it's conceivable that 5-F-THF may promote hypocotyl cell elongation in atdfb-3 through regulating cytoskeleton stabilization. Taken together, these observations suggest that folate-regulated N metabolism is important for atdfb-3 seedling development in the dark.

Perturbation of N metabolism was caused by atdfb mutation
Defective hypocotyl elongation in atdfb-3 is accompanied by perturbed N metabolism due to loss function of AtDFB. Under Nsufficient conditions, activities of NiR and GS were both lower in atdfb-3 than those of the wild type, indicative of an impaired N reduction and assimilation, although the contents of NO 3 2 , NO 2 2 and NH 4 + in atdfb-3 were similar to those in the wild type ( Figure 10 and Figure S8 in File S1). Under N-limited conditions, lower content of NO 3 2 may reflect a decreased NO 3 2 uptake and/or reduction in atdfb-3 ( Figure 10A), and the lower NiR , each replicate consisted of 500 mg of pooled plant material. A significant difference at P,0.05 is indicated by *, and a highly significant difference at P,0.01 is indicated by ** (Student's t-test). doi:10.1371/journal.pone.0101905.g010 activity could partly explain why high content of NO 2 2 accumulated in atdfb-3 ( Figure 10B and C). Apart from this, low activities of GS and GOGAT were also observed under 0.3 N ( Figure 10D and E), suggesting a reduced N assimilation ability in atdfb-3 as compared to the wild type. Taken together, the ability of N reduction and assimilation in atdfb-3 was significantly lower than in the wild type under N limitation, resulting in defective hypocotyl development. Furthermore, it was reported that folates can be oxidized by NO 2 2 to several pterin products [47], therefore it is possible that insufficient folate derivatives could not effectively protect the etiolated seedling from toxicity of significantly accumulated NO 2 2 under N-limited conditions. Thus, the drastically shortened hypocotyl of atdfb-3 under 0.3 N could be partly due to the NO 2 2 toxicity. However, how the altered folates profiling or polyglutamylation affects N metabolism in atdfb-3 awaits further investigation. One of the possibilities is that polyglutamylated 5-M-THF or 5-F-THF may act as a regulator of the N metabolism enzymes.
Taking together, we provide genetic evidence that the plastidial isoform FPGS is required for normal seed reserve accumulation and hypocotyl elongation under dark conditions. The AtDFB mutation results in altered seed storage, perturbed folate profile, altered N metabolism and shorter hypocotyls in etiolated seedlings, and exogenous 5-F-THF recovered shortened hypocotyls of the mutant to the wild-type level when NO 3 2 was present in the growth conditions. However, the underlying mechanism through which folate regulates seed reserve accumulation and hypocotyl development during skotomorphogenesis as well as the relationship among folate metabolism, N metabolism and hypocotyl development require further investigation.

Plant materials and growth conditions
Arabidopsis wild-type (Arabidopsis thaliana, ecotype Columbia), the T-DNA insertion mutant of AtDFB (SALK_015472, called atdfb-3 in this report), and the AtDFB complemented line were grown in identical growth chambers under a 16-h photoperiod (photosynthetic photon flux density 60 mE m 22 s 21 ) and a day/ night temperature of 22/16uC before being harvested. For biochemical analysis, and metabolite measurement assays, seeds from various genetic backgrounds were harvested at the same time and were after-ripened for 3 months.
For the Petri dish-based N limitation experiments, when NO 3 2 was used as the sole N source, NH 4 + was removed from the halfstrength MS medium [48]. The K + level was balanced with KCl to maintain 9.4 mM K + . In this report, 9.4 mM NO 3 2 (9.4 N) was used as the N-sufficient condition and 0.3 mM NO 3 2 (0.3 N) as the N-limited condition. When NH 4 + was used as the sole N source, NO 3 2 was removed from the half-strength MS medium, and NH 4 Cl was then added to the desired N concentration. When Asn and Gln were used as the sole N sources, NO 3 2 and NH 4 + were removed from the half-strength MS medium, and the K + level was balanced with KCl to maintain 9.4 mM K + . For C-or Pdeficiency experiments, sucrose or KH 2 PO 4 was not added to halfstrength MS medium. Endosperm/seed coat tissues were removed from the embryo using a dissecting microscope, after allowing seeds to soften in water for 6 h at 4uC [17]. For all experiments mentioned above, the wild-type and atdfb-3 seeds were sterilized, grown on the same plate, treated at 4uC in the dark for 2 days, and then moved to a growth chamber at 22uC under continuous dark conditions. Digital photographs of hypocotyls at various stages of etiolated seedling growth were acquired using a Nikon 700 camera, and their lengths were measured using ImageJ.
For expression pattern analysis, 1,406 bp of the AtDFB promoter was amplified and cloned into the binary vector pKGWFS 7.0. The construct was introduced into wild-type plants using the floral dipping method. The homozygous AtDFB: GUStransformed seedling were stained according to Francisco [49], and observed under the stereoscope from Nikon DIGITAL CAMERA Dxm 1200F.
For the 5-F-THF and 5-M-THF [(6R, S)-5-formyl-5,6,7,8tetrahydrofolic acid and (6R, S)-5-methyl-5,6,7,8-tetrahydrofolic acid, calcium salt; Schircks Laboratories, Switzerland] supplementation experiments, a stock solution was added to the growth medium to achieve the desired working concentration. A stock solution of 5 mM 5-F-THF or 5-M-THF was prepared in deionized water. Seeds were planted directly on the medium with or without the abovementioned folate derivatives. The hypocotyl length assays were performed as described above.

Microscopic analysis
The hypocotyls of 6-day-old etiolated wild-type and atdfb-3 seedlings grown on 9.4 N or 0.3 N were observed according to the method by Cowling et al. under a Hitachi S1-4800 high-resolution FE-SEM [50]. The cells of the midportions of hypocotyls were observed using a ZEISS Imager M1 DIC microscope and a 106 objective lens [51].

Biochemical analysis
Biochemical analysis procedure was according to Jiang et al. [33]. C and N contents were analyzed using a Perkin Elmer 2400 Series II CHNS/O Elemental Analyzer (www. perkinelmer.com), and the value indicated the percentage of C or N in total dry weight (mg/100 mg DW). Free amino acids were analyzed using an ASI.KAUNER amino acid analyzer A200 (www.knauer.net). Soluble proteins were extracted from the frozen seedling powder using 100 mM HEPES-KOH (pH 7.5) and 0.1% Triton X-100 and assayed using a commercial protein assay kit (Bio-Rad). NO 3 2 and NO 2 2 were measured as described by Oliveira [39]. NH 4 + was measured according to Andrew et al. [52].

Seed metabolite profile analysis using GC-TOF-MS
Seed metabolite analysis using GC-TOF-MS was performed using a method modified from that described previously [9,53]. Seeds (approximately 20 mg) were homogenized using a precooled mortar and pestle with liquid nitrogen and extracted in 1.5 ml of a methanol: chloroform: water extraction solution (2.5:1:1, v/v/v). Internal standards (50 ml 1 mg ml 21 ribitol in water and 20 ml 2 mg ml 21 C 13 -nonadecanoic acid in chloroform) were subsequently added. The mixture was extracted for 2 h at 37uC with shaking at 1,500 rpm. After 10 min of centrifugation at 12,000 rpm, 400 ml water and chloroform were added to the supernatant, respectively. Following vortexing and a 5-min centrifugation at 12,000 rpm, 200 ml methanol-water phase was isolated and reduced to dryness in a vacuum. Meanwhile, 400 ml chloroform-lipid phase was obtained and concentrated to dryness using nitrogen gas. Residues were re-dissolved and derivatized for 2 h at 37uC (in 25 ml 20 mg ml 21 methoxyamine hydrochloride in pyridine) followed by a 30-min treatment with 50 ml N-methyl-N-(trimethylsilyl) trifluoroacetamide at 37uC. Each 1 ml aliquot of the derivatives was injected in a splitless mode using an autosampler into an Agilent 6890 GC system coupled to a LECO Pegasus IV time-of-flight mass spectrometer system (LECO Corporation, USA). A DB-5MS capillary column (30 m60.25 mm i.d., 0.25mm film thickness, Agilent J&W Scientific, USA) was used to separate the samples. The injector temperature was 280uC. The Helium gas flow rate through the column was 1.0 ml min 21 . The column temperature was held at 80uC for 1 min and then increased by 10uC min 21 to 310 uC and held there for 10 min. The column effluent was introduced into the ion source of a Pegasus IV TOF-MS. The transfer line and the ion source temperatures were 280 and 200uC, respectively. The electron energy was 70 eV, and mass data were collected in a full-scan mode (m/z 50-650). The detector voltage was set at 1,650 V. All samples were randomized, and five biological replicates were analyzed within 24 h of chemical derivatization. Raw data were processed using LECO ChromaTOF v3.32. Information, including the peak area and retention time, for each detected metabolite was obtained. According to the retention time, the name of each metabolite was obtained by searching the NIST MS Search 2.0 database. When the matching score was higher than 800, the result was considered credible, and the metabolite was further analyzed. The relative contents of the metabolites are shown, and those of the wild type were normalized to values of 1.

Folate profile analysis using LC-MS
The following folates were purchased from Schircks Laboratories: 5-M-THF, THF, 5-F-THF, 5,10-methenyltetrahydrofolate, and DHF. The 2-day-old etiolated seedlings grown on 9.4 N and 0.3 N medium plates were used for identification of folate profiles. The 6-day-old etiolated seedlings grown on 9.4 N and 0.3 N medium plates were used for identification of folate profiles and folylpolyglutamates of 5-M-THF-Glu n and 5-F-THF-Glu n (n = 5, 6, 7, and 8). The methods for sample preparation and metabolite measurement were described previously [33]. The experiments included five biological replicates.

NiR, GS and GOGAT enzyme activity analysis
The method for enzyme activity analysis was similar to our previous report [33]. The 200 mg powdered tissues of 6-day-old etiolated seedlings for NiR analysis were added to 0.6 ml of extraction buffer containing 50 mM potassium phosphate buffer (pH 7.5), 1 mM EDTA, 10 mM 2-mercaptoethanol, 100 mM phenylmethanesulfonyl fluoride, and 5 mg PVP and then homogenized. The homogenate was centrifuged, and the supernatant (crude enzyme solution) was used for the NiR activity analysis. A blank sample, in which sulfanilamide was added prior to the extract, was used for background reading. NiR activity was assayed following Takahashi et al. [54], with modifications, to measure the decrease of NO 2 2 in the assay mixture. A 45 ml sample of the crude enzyme solution was transferred to a 1.5 ml centrifuge tube, and 195 ml of the assay solution containing 50 mM potassium phosphate buffer (pH 7.5), 1 mM NaNO 2 , and 1 mM methyl viologen was added. The reaction was started by adding 60 ml of 57.4 mM Na 2 S 2 O 4 in 290 mM NaHCO 3 (final Na 2 S 2 O 4 concentration in the assay solution, 11.5 mM), and the reaction was run for 5 min at 30uC. A 0.3 ml aliquot was transferred to a new tube containing 0.7 ml water and mixed vigorously to stop the reaction, after which 1 ml 1% (w/v) sulfanilamide in 3 N HCl and 1 ml 0.02% (w/v) N-1naphthylethylenediameine dihydrochloride were added. The absorbance of this mixture at 520 nm was measured. NiR enzyme activity was expressed as nmole NO 2 2 used per min per mg protein.
For assessment of total GS activities, freshly harvested samples (500 mg) were ground on ice with 1 ml extraction buffer consisting of 100 mM Tris-HCl (pH 7.6), 1 mM MgCl 2, 1 mM EDTA, and 10 mM 2-mercaptoethanol. Semi-synthetase GS activity was assayed, with NH 2 OH used as an artificial substrate, by quantifying the formation of glutamic acid c-monohydroxamate.
The homogenates were centrifuged at 12,000 g for 30 min at 4uC, and the supernatant was analyzed for total GS activities. Total GS activity was measured in a preincubation assay buffer (30uC) consisting of 37.5 mM imidazole buffer (pH 7.0), 30 mM sodium glutamate, 25 mM MgSO 4 , 50 mM NH 2 OH, and 3 mM ATP. The reaction was terminated after 15 min at 30uC by addition of acidic FeCl 3 solution (88 mM FeCl 3 , 670 mM HCl, and 200 mM trichloroacetic acid). After allowing 10 min for the color development, the reaction mixture was centrifuged at 4,000 g at room temperature for 10 min, and 2 ml of supernatant was then transferred from each well into a new tube. The A 540 was measured in a spectrophotometer quantification reader [55]. GS enzyme activity was expressed as mmol Glu c-monohydroxamate formed per 15 min per mg protein.
For assessment of GOGAT activities, freshly harvested samples (200 mg) were ground on ice with 0.6 ml extraction buffer consisting of 100 mM potassium phosphate buffer (pH 7.4), 1.28 mM EDTA, and 10 mM 2-mercaptoethanol. GOGAT activity was assayed by quantifying the formation of Glu and using NADH used as the substrate. The reaction mixture consisted of 100 mM potassium phosphate buffer (pH 7.4), 10 mM Gln, 10 mM 2-oxoglutarate, 0.05 mM NADH, and extract. After a 5min pre-incubation at 30uC, the reaction was started by adding the reductant solution (1.68 mg Na 2 S 2 O 4 and 3.48 mg NaHCO 3 in 1 ml of reaction solution). After a 15 min of incubation at 30uC, the reaction was terminated by heating to 98uC for 5 min. The Glu concentration was then determined using the ninhydrin reaction [56]. GOGAT activity was expressed as mmol Glu formed per min per mg protein.

Accession numbers
Sequence data from this article can be found at the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers:

Supporting Information
File S1 Contains the following files: Figure S1. Identification of atdfb-3 and AtDFB complemented (COM) line.  Figure S6. Folate profiles in 2-dayold WT and atdfb-3 germinating seeds under 9.4 N or 0.3 N. Figure S7. Transcript levels of genes involved in folate biosynthesis and metabolism. Figure S8. Ammonium content in 6-day-old WT and atdfb-3 seedlings in the dark. Table S1. Profiles of total folates and various folate species in 2-day-old WT and atdfb-3 germinating seeds grown on 9.4 N or 0.3 N medium in the dark. Table S2. Profiles of total folates and various folate species in 6-day-old WT, atdfb-3, and AtDFB complemented (COM) etiolated seedlings grown on 9.4 N or 0.3 N medium. (DOC)