Proanthocyanidin oxidation of Arabidopsis seeds is altered in mutant of the high-affinity nitrate transporter NRT2.7

Summary The seed-specific nitrate transporter AtNRT2.7 is involved in flavonoid accumulation as evidenced by the higher proanthocyanidin content in nrt2.7-2 mutant seeds. As TT10 laccase activity is not modified, the link between NRT2.7 and proanthocyanidin accumulation has yet to be discovered.


Introduction
Seed development and maturation lead to accumulation of N and C compounds in embryo such as reserve proteins, lipids, and carbohydrates, which are then used as energy sources during germination. The N compounds accumulated in seeds originate from nitrate (NO 3 -), amino acids, and peptides transferred from vegetative organs and subsequent synthesis of storage proteins. NO 3 uptake by the roots and its translocation to the aerial part and to the seeds are achieved by transporters of high-affinity and low-affinity systems (reviewed in Dechorgnat et al., 2011). The high-affinity system is ensured by some members of the NRT2 family (seven members) and the low-affinity system by some members of the NRT1 family (or NPF according to the unified nomenclature proposed by Léran et al., 2013;54 members), which transport also dipeptides (reviewed in Tsay et al., 2007) and other compounds such as auxin, abscisic acid, and glucosinolates (Krouk et al., 2010;Kanno et al., 2012;Nour-Eldin et al., 2012). NO 3 uptake by roots is mediated mainly by NRT2.1 and NRT1.1 (AtNPF6.3), depending on the NO 3 concentration of the soil solution (below or above 1 mM, respectively). At extremely low NO 3 concentration (below 0.025 mM), NRT2.4 is also active for NO 3 uptake by roots (Kiba et al., 2012). Then, root xylem loading is due to NRT1.5 (AtNPF7.3)  and root phloem loading to NRT1.9 (AtNPF2.9) (Wang and Tsay, 2011). In shoots, xylem unloading is performed by NRT1.8 (AtNPF7.2) and NRT1.4 (AtNPF6.2) (Chiu et al., 2004;Li et al., 2010). In leaves, up to 50% NO 3 storage is achieved by an anion channel/transporter (chloride channel a, CLCa), which is a nitrate/proton antiporter localized in the tonoplast of foliar cells (Monachello et al., 2009). Regarding the travel of NO 3 through the plant, NRT1.7 (AtNPF2.13) has been suggested as an actor in the apoplastic loading of NO 3 into the phloem sap of older leaves (Fan et al., 2009). There, the delivery of NO 3 to the developing seeds is due to NRT1.6 (AtNPF2.12) located in the vascular tissue of the silique and funiculus (Almagro et al., 2008). NO 3 represents quantitatively a minor N compound in dry seeds and its accumulation is due to a high-affinity NO 3 transporter, NRT2.7, specifically expressed in seeds (Chopin et al., 2007). One-hour-imbibed seeds of transformants expressing a fusion between NRT2.7 promoter and β-glucuronidase (GUS) reporter gene have shown a GUS staining in the embryo and in the endosperm. Transgenic lines carrying the GFP reporter gene fused to NRT2.7 under the control of the 35S CaMV promoter have evidenced the tonoplastic localization of NRT2.7. NO 3 is not only an important N nutrient for plants but also a signalling molecule and the role of NO 3 in the physiology of the seed has been shown especially in breaking dormancy (Alboresi et al., 2005). Mutants deficient in NRT2.7 display lower NO 3 content in dry seeds but also a higher dormancy highlighting the signalling role of NO 3 in dormancy relief (Chopin et al., 2007).
Secondary metabolites such as flavonoids are synthesized during seed development and are accumulated in the seed coat and in the embryo . Flavonoids are polyphenolic compounds responsible for the brown seed colour. They have also important functions in various aspects of seed development and have health benefits when present in animal and human diet . Flavonoids are involved in protection of seeds against biotic and abiotic stresses, for instance against ultraviolet radiations, and in acting as scavengers of free radicals. The physiological functions of flavonoids in strengthening seed dormancy and viability have also been documented (Debeaujon et al., 2000). Proanthocyanidin (PA) oxidation generates quinones that behave as toxic compounds against pathogens. They also constitute an antinutritive barrier against herbivores and interfere with fungal enzymes necessary for plant cell invasion. Quinones can also act as antioxidants by scavenging reactive oxygen species (ROS) produced by UV radiation, for example (reviewed in Pourcel et al., 2007). Arabidopsis seeds contain flavonols (glycosylated aglycones derivatives) in the seed coat and embryo, and PAs or condensed tannins in the inner integument and chalaza zone (Pourcel et al., 2005;Routaboul et al., 2006). The biosynthesis pathway and regulations have been largely studied especially through transparent testa (tt) mutants , which are characterized by a lighter seed colour phenotype. The brown colour of Arabidopsis seeds occurring during desiccation is due to the oxidation of PAs and their epicatechin monomers by the laccase-like enzyme TT10/LAC15 (Pourcel et al., 2005). Moreover oxidized PAs cross-link with cell-wall components, thus becoming insoluble and as such difficult to extract . Seeds from the tt10 mutant deprived of TT10 laccase-like activity are yellow at harvest but slowly darken with storage time through chemical oxidation reactions. They exhibit more soluble (i.e. extractable) PAs than wild-type seeds but are not affected in PA biosynthesis per se. They also accumulate less biflavonols, which are dimers of the flavonol quercetin 3-O-rhamnoside and are also synthesized by TT10. Before oxidation, PA biosynthesis and polymerization involve transport and/or vesicle trafficking (Zhao et al., 2010). While the biosynthesis of PA precursors is believed to occur in the endoplasmic reticulum, transfer into the vacuole is performed by TT12 (a multidrug and toxic efflux transporter family) coupled to AHA10 a putative P-type H + -ATPase (Baxter et al., 2005;Marinova et al., 2007). However, the complete story of PA transport inside the cell has not yet been completely elucidated (reviewed in Zhao et al., 2010).
This work describes a new phenotype for the nrt2.7-2 mutant allele which exhibited seeds with more soluble PAs. Little is known about the mechanisms regulating the oxidation of tannins in seeds, and this study provides a new link between nitrogen signalling and PA metabolism. The role of NO 3 accumulated in seeds is discussed in relation to tannin oxidation, TT10 expression, and TT10 activity.

Growth conditions
Plants were grown in a growth chamber at 60% relative humidity with a 16/8 light/dark cycle at 21//17 °C and light intensity 150 μmol m -2 s -1 . Seeds were sown on sand in 5 × 5 cm pots and plants were subirrigated three times a week with a complete nutrient solution (10 mM NO 3 -) containing 5 mM KNO 3 , 2.5 mM Ca(NO 3 ) 2 , 0.25 mM MgSO 4 , 0.25 mM KH 2 PO 4 , 0.42 mM NaCl, 0.1 mM FeNa-EDTA, 30 μM H 3 BO 3 , 5 μM MnSO 4 , 1 μM ZnSO 4 , 1 μM CuSO 4 , and 0.1 μM (NH 4 ) 6 Mo 7 O 24 . For the experiments on dry seeds, plants were harvested at the end of the culture, whereas for the seed development experiments, flowers at the beginning of anthesis were tagged every 3 d after fertilization (DAF) on one stalk per plant and then 6-21-d-old siliques were harvested.
For the experiment with varying nitrogen nutrition, plants were subirrigated with 10 mM NO 3 from the sowing to the flowering stage and then with 0.2, 2, or 10 mM NO 3 -. In the 2 mM nutrient solution until harvest, KNO 3 and Ca(NO 3 ) 2 concentration was 1.75 mM and 0.125 mM, respectively. In the 0.2 mM nutrient solution, KNO 3 concentration was 0.2 mM, and Ca(NO 3 ) 2 was replaced with 0.25 mM CaCl 2 .
Nitrate content measurement Nitrate content of seeds was determined after extraction in water of 2 mg dry seeds or 1 mg developing seeds excised from siliques and silique tissues (siliques without seeds). The nitrate content was measured by a spectrophotometric method adapted from Miranda et al. (2001). The principle of this method is a reduction of nitrate by vanadium (III) combined with detection by the acidic Griess reaction.
C, N, total protein, amino acids, sugar, and fatty acid determination Total C and N determination were carried out on 1 mg seeds following the Dumas combustion method using a NA 1500 Serie 2 CN Fisons instrument analyser (Thermoquest) as described in Baud et al. (2010). Fatty acid analyses were performed on pools of 20 seeds by gas chromatography after extraction in methanol/sulphuric acid (100:2.5, v/v) as previously described (Li et al., 2006). Free amino acids and sucrose contents were determined after 80% (v/v) ethanolic extraction on batches of 20 seeds according to Baud et al. (2002). Free amino acid content was quantified by the ninhydrin colourimetric analysis according to Rosen (1957). Sucrose was determined enzymatically using a kit (Boehringer Mannheim). Starch was quantified from the pellet resulting of the ethanolic extraction. After hydrolysis of starch by amyloglucosidase and amylase (Baud et al., 2002), glucose was determined enzymatically using a kit (Boehringer Mannheim). Total protein content was determined on batches of 1.5 mg seeds by the ninhydrin colourimetric quantification of the amino acids released after 1 h hydrolysis of the seeds at 120 °C in 3M NaOH, as described in Baud et al. (2007).

Flavonoid composition analyses
Flavonoids were extracted from 15 mg dry seeds with acetonitrile/ water (75:25, v/v), as described in Routaboul et al. (2006). After centrifugation of the extracts, the supernatant was used for the analysis of flavonols and soluble PAs, while the pellet contained insoluble PAs. Analyses of soluble and insoluble PAs were further performed after acid-catalysed hydrolysis and absorbance measured at 550 nm using cyanidin as a standard molecule according to Routaboul et al. (2006). Epicatechin monomers and PA polymers were then analysed by LC-MS. Flavonol composition was also analysed by LC-MS using apigenin as an internal standard which was added at the time of extraction .

Results and discussion
Soluble PAs are more accumulated in seeds of the nrt2.7-2 mutant Because NO 3 is an important N nutrient for plants, the impact of the lack of NRT2.7 on the accumulation of other N and C reserve compounds was evaluated in nrt2.7 mutant seeds when grown on nonlimited supply of N (10 mM NO 3 -). Total N and free amino acid contents were not affected in the nrt2.7-2 mutant compared to the wild type (Table 1), while total protein content was slightly increased (Table 1) and NO 3 content was decreased (Fig. 1A). A decrease in NO 3 content was also observed in tt10-2 mutant seeds (Fig. 1A).
In contrast, total C, fatty acids, starch, and sugar contents were not changed in the nrt2.7-2 mutant (Table 1). Thus, the decrease in capacity to store NO 3 in nrt2.7-2 seed vacuole seemed to favour the accumulation of N-protein reserve compounds without affecting C content.
Interestingly the nrt2.7-2 mutant seeds displayed a slightly lighter colour compared to the wild-type Ws, resembling the phenotype of tt10 mutant seeds (Pourcel et al., 2005) (Fig. 2). This lighter colour phenotype was also observed in the double mutant nrt2.7-2 tt10-2 with a pale-brown seed coat colour and a dark-brown chalaza zone (Fig. 2). The analysis of flavonoids in mature seeds revealed that the soluble PA content was similarly increased in the nrt2.7-2 and tt10-2 mutants compared to Ws seeds (Fig. 1A) while insoluble PAs were not changed ( Supplementary Fig. 1, available at JXB online). LC-MS analyses showed that the soluble epicatechin monomers and oligomers were also increased in the nrt2.7-2 mutant (Fig. 1B) as well as in the tt10-2 mutant (see Pourcel et al., 2005). However, unlike the tt10-2 mutant that contains only very small amounts of biflavonols (dimer of quercetin 3-O-rhamnoside) and a slightly more quercetin 3-O-rhamnoside monomers (Fig. 1C), the flavonol composition of the nrt2.7-2 mutant was not modified (Fig. 1C). Thus, the nrt2.7-2 mutant was very peculiar as it exhibited a modification in flavonoid composition that was specific for PAs and perhaps, as in the tt10 mutant, linked to a defect in PA oxidation. It remains to be investigated whether the TT10 function is altered in nrt2.7-2 or whether another oxidative mechanism is involved.
To first confirm the link between the T-DNA insertion in At5g14570 and the phenotype of nrt2.7-2, a functional complementation of nrt2.7-2 was conducted with a construct, Pro35S:AtNRT2.7, which allows overexpression of a fulllength NRT2.7 under control of the CaMV 35S promoter. As a result, the soluble PAs and nitrate content phenotypes  were both restored in the two complemented lines nrt2.7-2 C12 and nrt2.7-2 C14 (Fig. 1A). However, the PA phenotype of the nrt2.7-2 mutant allele was specific for the Ws accession, since no difference in soluble PAs was observed in the nrt2.7-1 null mutant allele in Columbia background ( Supplementary Fig. S2 available at JXB online) whereas the nitrate phenotype was also encountered in nrt2.7-1 (Chopin et al., 2007). These data suggest that nitrate and soluble PA contents are not directly correlated. The specificity of the PA phenotype for the Ws accession was surprising but natural variability in PA accumulation has already been reported Routaboul et al., 2012), suggesting a variability in the regulation of PA oxidation. Besides, plant nitrate content varies also among accessions and, more precisely, Col accession displays a higher capacity to store nitrate than Ws accession in seeds (Chopin et al., 2007) and in foliar tissues (North et al., 2009), and consequently Col is more tolerant to N limitation. Control of PA oxidation originating from natural diversity of strategies for nitrate use and storage might explain the lack of PA phenotype for the Col accession. Thus, this work investigated further the relationship between nitrate accumulated in the seed and condensed PA accumulation. Since the PA phenotype of the nrt2.7-2 mutant was first observed when plants were grown on nonlimiting supply of N nutrition (10 mM) as described above, a more comprehensive range of NO 3 nutrition was also tested from 0.2 and 2 mM NO 3 as limited N levels to 10 mM NO 3 -. The NO 3 content of dry seeds was linked to the NO 3 nutrition in both genotypes and it was lower in the nrt2.7-2 mutant than in Ws on 10 mM NO 3 -, but not significantly affected on 2 and 0.2 mM NO 3 - (Fig. 3A). Epicatechin and soluble PAs (epicatechin oligomers) were more accumulated in the nrt2.7-2 mutant for all nutrition levels ( Fig. 3B and C), while still no change was observed for flavonols (Fig. 3D). The effect of nrt2.7-2 mutation on both NO 3 and soluble PA contents increased with the NO 3 nutrition level. Considering these results, subsequent experiments were performed at 10 mM NO 3 -, which allowed viewing of the most pronounced flavonoid and nitrate phenotypes.
The NO 3 content in seeds was dependent on supply of NO 3 nutrition (Fig. 3A) and thus may be relevant to the NO 3 availability for allocation to the seeds. This work speculated whether a limited capacity of NO 3 storage in leaves could also modulate NO 3 transfer to the seeds and could also influence the soluble PA level in seeds. Therefore, this work analysed the consequence of a knockout mutation in CLCa, encoding a nitrate/proton antiporter responsible for NO 3 accumulation in vacuolar compartment in leaves (Monachello et al., 2009). Interestingly, NO 3 content was decreased in clca1 and clca2 mutant seeds to the same extent as in the nrt2.7-2 mutants, while the soluble PA accumulation was not changed in clca1 and clca2 mutants (Ws background) ( Fig. 4A and B). This result suggested that the mechanism linking NO 3 accumulation and PA accumulation in seeds was specifically linked to NRT2.7 function in seeds rather than to global NO 3 accumulation.

Nitrate accumulation during seed development
It has already been described that PA oxidation in the testa starts with the desiccation of developing seeds (Pourcel et al., 2005). In order to better understand the link between NRT2.7 and PA oxidation/accumulation in seeds, the current work investigated more precisely the fluctuation of NO 3 content in seeds and in siliques tissues (siliques excluding seeds) during seed development. The NO 3 content was the highest in young seeds (9 DAF) and decreased abruptly (12 DAF) to the final low content in mature seeds (Fig. 5A). Conversely NO 3 content was the lowest in young siliques tissues (9 DAF) and increased regularly up to the senescing stage (21 DAF) (Fig. 5B). In the nrt2.7-2 mutant, the NO 3 contents were slightly lowered in seeds at 12 DAF and in mature seeds compared to those in Ws (Fig. 5A), concomitantly to the maxima of NRT2.7 expression in Ws (Fig. 5C). In contrast, NO 3 content was not affected in silique tissues of the nrt2.7-2 mutant (Fig. 5B). Thus, NRT2.7 was likely not the only actor responsible for NO 3 accumulation in these tissues. According to Almagro et al. (2008), the impact of the NRT1.6 (AtNPF2.12) mutation was strongly associated with a reduced NO 3 content in seeds and an increased seed abortion, but no colour phenotype of the nrt1.6 mutant seeds was reported. In the current study, no significant difference in NRT1.6 (AtNPF2.12) expression was measured in Ws and in nrt2.7-2 (data not shown). NRT1.6 (AtNPF2.12) was expressed in the vascular tissue of the silique and funiculus and was partially responsible for the delivery of NO 3 into the seed, but NRT1.6 (AtNPF2.12) was localized at the plasma membrane and, thus, may not be able to compensate the vacuolar nitrate storage in nrt2.7-2. Expression of the vacuolar anionic Colour phenotype of mature seeds of nrt2.7-2 and tt10-2 simple mutants, nrt2.7-2 tt10-2 double mutant, and wild-type (Ws) mature seeds. Mother plants were grown on 10 mM NO 3 and seeds were observed at harvest. Bar, 500 μm.
channel CLCa was detected in silique tissues (Fig. 5D) and, thus, could explain the partial compensation mechanism for the loss of NRT2.7 function in this organ, but no expression of CLCa was measured in excised seeds (Fig. 5C). Further study is required to find out if any other transporter is functional in these organs.
The PA phenotype of the nrt2.7-2 mutant is not due to a modulation of TT10 expression nrt2.7-2 mutant seeds accumulated less NO 3 and more soluble PAs and epicatechins compared to Ws partially resembling tt10 mutant phenotype. Thus, this phenotype was likely arising from a defect in PA oxidation leading to an accumulation of soluble forms of PAs during the development. According to Pourcel et al. (2005), TT10 expression in entire siliques begins to be detected at 4 DAF. Thus, the current work investigated TT10 and AtNRT2.7 expression in excised seeds and silique tissues excluding seeds of Ws and the nrt2.7-2 mutant during seed development. In Ws, the level of NRT2.7 mRNA was lower than TT10 but they were expressed in seeds and siliques ( Fig. 5C and 5D). The expression patterns of TT10 and NRT2.7 varied along seed development. TT10 expression was repressed in excised seeds from 9 DAF to 21 DAF (or mature seeds) (Fig. 5C). TT10 mRNA levels in silique tissues were measured 50% lower than those in excised seeds (Fig. 5D). In contrast, NRT2.7 expression showed two maxima in excised seeds, at 12 and 21 DAF (Fig. 5C) and increased slightly in silique tissues from 9 to 18 DAF (Fig. 5D). Furthermore, this work failed to observe a modified expression pattern of TT10 that was significantly reproducible in the nrt2.7-2 mutant compared to Ws (data not shown).
A role in signalling was previously suggested for NO 3 in relieving seed dormancy (Alboresi et al., 2005). However, considering that the maximum of TT10 expression preceded the first raise in NRT2.7 expression and the beginning of NO 3 content to decrease in nrt2.7-2, the current study excluded the hypothesis of a signalling role for NO 3 in downregulating the expression of TT10 and then lowering soluble PA oxidation. Is the PA phenotype of the nrt2.7-2 mutant due to a modulation of TT10 activity?
In order to find out a causal explanation for the PA phenotype of the nrt2.7-2 mutant, TT10 activity was considered. The enzymic activity of TT10 has never been successfully measured in vitro but an assay for in situ detection of browning in immature seed coat has been reported by Pourcel et al. (2005). In a first attempt, the current study looked into the in situ measurement of TT10 activity in young seeds (7-8 DAF) of Ws and the nrt2.7-2 mutant using the tt10-2 mutant as a negative control and the tt4-8 mutant as a positive control without endogenous supply of flavonoids (due to the lack of chalcone synthase). The browning intensity of the seeds incubated in presence of the epicatechin substrate revealed the PA oxidation activity of TT10. As expected, tt10-2 seeds stayed colourless and seeds of Ws and tt4-8 showed a brown colour, but nrt2.7-2 seeds became as brown as Ws (Table 2). These results suggested that the oxidative activity of TT10 was not altered in nrt2.7-2 seeds at this stage. However, this type of experiment is only feasible when the testa was still colourless in immature seeds. At this stage TT10 was highly expressed but these conditions were not favourable for a maximal NRT2-7 expression. Further investigation of TT10 activity by optimizing the in situ measurement at older stages is needed to understand the mechanism of higher soluble PA accumulation in nrt2.7-2 seeds.
Since the mechanisms for regulating the TT10 activity are largely unknown, the link between NRT2.7 and TT10 activity is difficult to assess. TT10 protein has been described as a putative laccase containing four His-rich copper-binding domains, corresponding to the putative catalytic sites of the multi-copper oxidase family (Pourcel et al., 2005). A phylogenetic analysis has revealed the highest homology of TT10 with four other dicotyledonous laccases (and for example with RvLAC2 from the sap of the Japanese lacquer tree Rhus vernicifera). Nitric oxide (NO) has been reported as a regulator of laccases, acting as a reducer of the R. vernicifera laccase RvLAC2 and also of fungal laccases (Torres and Wilson, 1999;Wilson and Torres, 2004). However, the consequences of the NO action on the enzymic activity of laccase are not completely understood (Torres et al., 2002). TT10 protein has recently been experimentally shown to be localized in vacuole (Pang et al., 2013), the same cellular compartment as NRT2.7, but a hypothetical link between NO, TT10 activity, and NRT2.7 remains uncertain.
What is a role for NRT2.7 in PA oxidation/ accumulation?
According to Pourcel et al. (2005), TT10 is expressed in the developing testa, firstly in the inner integument (PA-producing cells) and afterwards in the outer integument (location of flavonol synthesis). NRT2.7 expression has been previously localized in the endosperm and in embryo in imbibed seeds (Chopin et al., 2007) while PAs are synthesized and accumulated in the endothelium. Although the current work was able to measure NRT2.7 expression by qPCR in excised seeds, all attempts viewing the localization of NRT2.7 in the seed during its development by in situ hybridization or immunolocalization were unsuccessful. However, NRT2.7 expression is present in the seed coat according to the data available on the eFP browser web site (http://bbc.botany. utoronto.ca/efp_seedcoat/cgi-bin/efpWeb.cgi). NRT2.7 has already been described as a NO 3 transporter (Chopin et al., 2007), which is coherent with the lower NO 3 content in seeds of nrt2.7 mutant. Since the PA phenotype appeared more strictly correlated to the presence of NRT2.7 than to the vacuolar NO 3 content (Fig 3A and B), it was speculated whether the function of NRT2.7 in PA oxidation could be related to another function of NRT2.7 hitherto unknown. It has been demonstrated that NRT1 (NPF) proteins are able to transport molecules other than nitrate (Léran et al., 2013), although little is known about the NRT2 family. Further experiments are needed in order to ascertain such hypothesis. The transport of epicatechin into and out of the vacuolar compartment could have been disturbed in absence of NRT2.7. TT12 is a MATE transporter involved in the storage of PA precursor into the vacuole and its activity is coupled to AHA10, an H + -ATPase. Aha10 and tt12 mutants are affected in PA accumulation and also in the vacuolar biogenesis, supporting an endomembrane function for these transporters. There may be a direct or indirect link between these transport activities and NRT2.7 that involves pH stability, tonoplast stabilization, or other unknown mechanism. Table 2. In situ enzymic activity of the TT10 laccase in the wild type and mutants The analysis was performed according to the method described in Pourcel et al. (2005). The table describes seed coat colour with and without (control) the addition of epicatechin substrate to immature seeds (7-8 DAF). The browning colour intensity is positively correlated to TT10 activity. It is recorded by visual observation and noted as such: -, colourless; +++<++++, increasing browning colour. Approximately 50 seeds per sample were analysed.