Ascorbate-Deficient vtc2 Mutants in Arabidopsis Do Not Exhibit Decreased Growth

In higher plants the L-galactose pathway represents the major route for ascorbate biosynthesis. The first committed step of this pathway is catalyzed by the enzyme GDP-L-galactose phosphorylase and is encoded by two paralogs in Arabidopsis – VITAMIN C2 (VTC2) and VTC5. The first mutant of this enzyme, vtc2-1, isolated via an EMS mutagenesis screen, has approximately 20–30% of wildtype ascorbate levels and has been reported to have decreased growth under standard laboratory conditions. Here, we show that a T-DNA insertion into the VTC2 causes a similar reduction in ascorbate levels, but does not greatly affect plant growth. Subsequent segregation analysis revealed the growth defects of vtc2-1 mutants segregate independently of the vtc2-1 mutation. These observations suggest that it is the presence of an independent cryptic mutation that affects growth of vtc2-1 mutants, and not the 70–80% decrease in ascorbate levels that has been assumed in past studies.

Because vtc2-1 has detectable levels of full-length VTC2 transcript (Dowdle et al., 2007) and is thus not a complete lossof-function mutant, we obtained a T-DNA insertion mutant from the Arabidopsis stock center. We demonstrate that this mutant, here named vtc2-4, lacks wildtype VTC2 transcripts and, although it has a level of ascorbate similar to vtc2-1, exhibits a near wildtype growth phenotype under short-day and continuous light conditions. We suggest that the decreased growth phenotype of vtc2-1 is likely due to an independent cryptic mutation and is not due to ascorbate deficiency as widely assumed in the literature.

Plant Materials and Growth Conditions
Mutant alleles were all in the Col-0 background and were sourced from the ABRC stock center. For growth assays under short days, seeds were sown on soil and placed in a growth chamber and exposed to an 8 h light (150-250 µmol photons m −2 s −1 )/16 h dark cycle at 21 • C for 6 weeks. Plant weight was recorded and leaf surface area measurements taken by imagining freshly collected leaves and using ImageJ software to determine the surface area of leaf silhouettes. For growth under continuous light conditions, surface-sterilized seeds were germinated on agar media containing mineral salts (MM) with 2% sucrose (MMS) and 0.8% Bacto agar as described by Scholl et al. (1998). Seedlings were grown at 20 • C in a climate chamber with a 16 h light (150-250 µmol photons m −2 s −1 )/8 h dark cycle for 7 days before transfer to soil in a climate chamber with 24 h continuous light (150-250 µmol photons m −2 s −1 ).
For comparison of leaf area and chlorophyll fluorescence under short and long day conditions, plants were grown in pots (5 cm square and 5 cm deep) containing four parts Levington F2 compost (Scotts, Maryville, OH, USA) plus one part vermiculite. They were grown in short days (8 h) for 22 days after sowing in a controlled environment room at 23 • C and 65% relative humidity with a light intensity of 200 µmol photons m −2 s −1 . After 22 days, half the plants were transferred to long days (14 h) under otherwise identical conditions and the leaf area and chlorophyll fluorescence measured over 10 days. The plants were imaged with a CF Imager (Technologica Ltd., Colchester, UK). They were dark adapted for 30 min before measuring basal fluorescence (F 0 ). This was followed by measuring maximum dark adapted fluorescence (F m ) resulting from a saturating light flash (6, 349 µmol photons m −2 s −1 for 0.8 s). Darkadapted quantum efficiency of photosystem II was calculated (Baker, 2008). Leaf area of the imaged rosettes was extracted from the chlorophyll fluorescence images and the instrument was calibrated with leaf discs of known area.

Measurement of Leaf Ascorbate Content in Seedlings
Two-week old seedlings were harvested for ascorbate assay. Three seedlings from each control line (Col-0, single mutant lines) and 15-20 vtc2-4;vtc5 seedlings were pooled for each biological replicate. Total ascorbate (ascorbate and dehydroascorbate) was measured by the iron (III) reduction assay (Kampfenkel et al., 1995) in an 8x scaled-down protocol.

RNA Isolation and RT-PCR Analysis
Total RNA from 3-week old shoot tissues was isolated using the Qiagen RNeasy Plant Mini Kit 1 according to the manufacturer's protocol. The extracted RNA was further treated to remove any contaminating DNA using the DNA-free TM kit 2 according to the manufacturer's protocol. First strand cDNA was synthesized using the SuperScript R III First Strand Synthesis system 2 according to the manufacturer's protocol. The full length VTC2 coding DNA sequence was amplified with the following primers: forward (5 -CAAAAGAGTTCCGACCGTTG-3 ) and reverse (5 -ACTGAAGGACAAGGCACTCG-3 ). The transcript of the constitutively expressed ACTIN2 (ACT2) was used as an internal control with the following primers: forward (5 -GGTAACATTGTGCTCAGTGGTGG-3 ) and reverse (5 -CTCGGCCTTGCAGATCCACATC-3 ). The PCR was carried out using Promega GoTaq R Green Master Mix 3 with the following conditions: initial denaturation 5 min at 94 • C was followed by 30 cycles of 30 s denaturation step at 94 • C, 30 s annealing step at 60 • C and 60 s extension step at 72 • C. The final extension step was for 5 min at 72 • C.

Genotyping the vtc Mutants
Plants were genotyped by PCR using DNA extracted with the preparation method described in Edwards et al. (1991). The vtc2-1, vtc5-1, and vtc5-2 mutants were genotyped as described by Dowdle et al. (2007). The SAIL line of vtc2-4 was genotyped by performing PCR using a triplet of primers; two primers complementary to genomic DNA sequences situated on either side of the insertion site and a third primer complementary to the left border of the vector pDAP101. Primer sequences were as follows: forward (5 -TGATAATGGTTTCTGTAGCTTGGA-3 ), reverse (5 -AAAACCAAGCTCTCTGCACAA-3 ) and LB1 (5 -GCCTTTTCAGAAATGGATAAATAGCCTTGCTTCC -3 ). Accession Numbers vtc2-4 and vtc2-1W (line#4) have been deposited with ABRC and given the stock numbers CS69540 and CS69541, respectively.

Genetic Characterization of a vtc2 T-DNA Insertion Mutant
A line predicted to have a T-DNA insertion in VTC2 (SAIL_769_H05/CS876707) was obtained from the Arabidopsis stock center (ABRC). PCR genotyping and sequencing of genomic DNA adjacent to the T-DNA left border confirmed the presence of the T-DNA within the fourth exon of VTC2 at position +620 downstream of the start codon (+1). Lines homozygous for this insertion mutation, hereafter called vtc2-4, were backcrossed to wildtype (Col-0) and the resulting F2 progeny PCR genotyped for presence of the T-DNA allele. Numbers of VTC2/VTC2:vtc2-4/VTC2:vtc2-4/vtc2-4 plants were consistent with an expected 1:2:1 segregation ratio (48:112:56; χ 2 = 0.89; p > 0.6). Subsequent RT-PCR analysis failed to amplify a full-length VTC2 transcript from lines homozygous for vtc2-4 confirming presence of the T-DNA within this gene (Supplementary Figure S1). Measurement of total ascorbate in leaves of 4-week-old wildtype, vtc2-1 and vtc2-4 plants revealed that both vtc2-4 and vtc2-1 had approximately 30% the levels of ascorbate that are found in wildtype, showing that the T-DNA present in the vtc2-4 line conditions a similar decrease in ascorbate ( Figure 1A).
Previous studies have shown that the EMS-induced mutant vtc2-1 has a growth defect (Müller-Moulé et al., 2004;Pavet et al., 2005;Kerchev et al., 2011). However, when vtc2-4 mutants were propagated on soil under continuous light, they had a wildtype appearance ( Figure 1B) and fresh weight measurements differed little from wildtype ( Figure 1C). In contrast, vtc2-1 mutants were significantly smaller than vtc2-4 mutants, with approximately half the fresh weight of wildtype. To ensure that the observed growth difference between vtc2-1 and vtc2-4 mutants are a general feature of these mutants and not just induced by long-days, leaf surface area measurements were assessed over a 10-day growth period under both short and long-days conditions. Regardless of day length, vtc2-4 mutants had a marginally smaller leaf surface area when compared to wildtype, whereas vtc2-1 mutants were considerably smaller than Error bars indicate SE (n = 3). Significant differences from Col-0 wildtype were determined with Student's t-test ( * * p < 0.01). Significance values were adjusted for multiple comparisons using the Bonferroni correction. (B) Three-week-old Col-0 wildtype, vtc2-1, vtc2-4, vtc2-1W, VTC2-1S grown under continuous light. Scale bar = 10 mm. (C) Shoot fresh weight measurements of 3-week-old plants grown under continuous light conditions. Error bars indicate SE (n = 6). Significant differences from Col-0 wildtype were determined with Student's t-test ( * * p < 0.01). Significance values were adjusted for multiple comparisons using the Bonferroni correction.
wildtype and vtc2-4 mutants (Figures 2A,B). A multi-factor ANOVA test confirmed that differences in leaf size between vtc mutant lines and wildtype under either growth condition was significant (p < 0.0001; Table 1). To determine whether the observed difference in leaf surface area arises from altered rates of growth, relative leaf expansion rate of rosettes under long and short day conditions were calculated. As expected, the growth rate was largely influenced by day length as each line displayed significantly higher growth under long days ( Figure 2C). In contrast, there was no significant difference in the growth rate between wildtype and the vtc2 mutants under long days. While this was also true for wildtype and vtc2-4 FIGURE 2 | (A-C) Rosette and leaf area measurements of Arabidopsis thaliana vtc2 mutants. (A) Surface area measurements (mm 2 ) of leaves taken over a 10-day period under long-day (14 h) conditions. Values are mean and error bars indicate SD (n = 12). (B) Surface area measurements (mm 2 ) of leaves taken over a 10-day period under short-day (8 h) conditions. Values are mean and error bars indicate SD (n = 12). (C) Relative leaf expansion rate (RLER, day −1 ) was calculated as the slope of a linear fit to a graph of natural log leaf area (mm) versus time (days). RLER was calculated from data in Figures 2A,B. Values are mean and error bars indicate SD (n = 10-12). Significant differences were determined with a one-way ANOVA and Tukey HSD post hoc test ( * p = 0.005, * * p < 0.001).
(D) Silhouettes of leaves from 6-week-old plants grown under short-day conditions. Scale bar = 1 cm. (E) Surface area measurements (mm 2 ) of leaves taken from 6-week-old plants grown under short-day conditions. Values are mean and error bars indicate SD (n = 7-10). Differences between leaf areas of vtc2-1, vtc2-1W and VTC2-1S were assessed using a Student's t-test ( * p < 0.05). mutants under short days, vtc2-1 mutants had a significantly reduced growth rate when compared to wildtype (Figure 2C). Finding a smaller rosette leaf area for both vtc2-1 and vtc2-4 in long days (Figure 2A), but similar relative expansion rate to wildtype (Figure 2C), implies that the small size results from initially smaller seedlings rather than from differences in intrinsic leaf expansion rate. However, under short-day conditions, vtc2-1 relative leaf expansion rate is compromised as well. Measuring the surface area of individual leaves collected from 6-week old short-day grown plants revealed that the size differences between wildtype and vtc2-1 mutants is mostly confined to the fifth and subsequent leaves (Figures 2D,E). Taken together these data show that the growth defects of vtc2-4 mutants are not of the same magnitude as those seen in The leaf area data are shown in Figures 2A,B. vtc2-1 and hence may not be directly attributable to ascorbate deficiency.
The Small Size Phenotype of vtc2-1 Is Not Linked to Ascorbate Deficiency Since vtc2-1 and vtc2-4 have contrasting growth phenotypes, we undertook a genetic characterization of vtc2-1. From a cross between wildtype and vtc2-1 F1 individuals, heterozygous for vtc2-1, exhibited wildtype growth in soil under continuous light condition (data not shown). The presence of the vtc2-1 allele can be detected by CAPS (cleaved amplified polymorphic sequence) PCR (Dowdle et al., 2007). In a segregating F2 population grown in soil, genotyping confirmed that the vtc2-1 allele segregated in a Mendelian ratio ( Table 2; χ 2 = 2.23; p > 0.3). Four-week-old individual F2 plants were also visually scored for size as "Wildtype" or "Small." While wildtype and vtc2-1 mutants grown in parallel showed a clear difference in size as previously reported (Veljovic-Jovanovic et al., 2001;Müller-Moulé et al., 2004;Pavet et al., 2005;Kerchev et al., 2011), some individual wildtype plants exhibited a similar growth size to vtc2-1 mutant plants. However, an excess of small individuals over the expected number suggested it was likely that some wildtype individuals were misclassified as "Small." A number of individuals with a small growth phenotype were homozygous for the wildtype VTC2 allele and, conversely, some vtc2-1 homozygous individuals had a wildtype appearance ( Table 2).
To score the growth phenotype more accurately a number of F2 lines derived from the Col-0/vtc2-1 cross were allowed to self-fertilize to produce individual F3 families, which were then observed as populations after 4-week-growth in soil under continuous light conditions ( Table 3). Of seven F2 individuals wildtype at the VTC2 locus with an apparent small growth phenotype, five F3 families were uniformly small in size (called VTC2-1S; Figure 1B). Conversely, of 24 F2 individuals homozygous for vtc2-1 and apparently wildtype in  size, 18 F3 families were uniformly wildtype in size (called vtc2-1W; Figure 1B). Measuring fresh weight and total ascorbate levels under continuous light conditions revealed that vtc2-1W plants had a similar weight to wildtype, but with ascorbate levels decreased by approximately 70% compared to wildtype. Likewise, fresh weight of a representative VTC2-1S line was similar to vtc2-1, but with ascorbate approaching wildtype levels ( Figure 1C).
To better characterize the growth characteristics of these F3 lines, fresh weight and leaf surface area measurements of representative vtc2-1W and VTC2-1S plants grown under short-day were made (Figures 2D,E, Supplementary  Figure S2). In agreement with the measurements made under continuous light conditions, the fresh weight of the vtc2-1W line was statistically different from vtc2-1 mutant controls (Supplementary Figure S2). When compared to vtc2-1 mutants, late arising leaves of vtc2-1W plants were larger (leaf 7 to leaf 9; increase ranged from 18 to 41%). In contrast, the fresh weight of a representative VTC2-1S line was not significantly different from vtc2-1 mutants (Supplementary Figure S2). This is despite finding that a few VTC2-1S leaves were slightly smaller than those of vtc2-1 (leaf 5-6; decrease ranged from 13 to 23%). In summary, these analyses demonstrate that ascorbate-deficiency is genetically separable from growth defects associated with the vtc2-1 mutant line.

Leaf Senescence in vtc2 Mutants
Given that ascorbate deficiency in vtc2-4 mutants is not associated with a strong growth defect, we next considered whether this line displays other ascorbate-deficient characteristics. Previous studies have shown that reduced ascorbate is associated with faster leaf senescence (Barth et al., 2004). Senescence can be measured as a loss of leaf photosynthetic efficiency. Thus to assess the extent of leaf senescence in vtc2 mutants, we imaged dark-adapted quantum efficiency of photosystem II (F v /F m ). Examining the F v /F m images of wildtype, vtc2-4 and vtc2-1 grown under both short and long days suggested bigger decreases in older leaves of vtc2 mutants compared to wildtype (Figure 3A). To quantify the different rates of senescence in each line, the % of leaf area below a fixed F v /F m value (0.75) over days 7-10 was calculated. This revealed a slight elevation in the rate of senescence under short days compared to long days for all lines tested which was only significant for vtc2-4  Figures 2A,B). Error bars indicate SD (n = 6-12). ANOVA shows significant differences (p < 0.001) between Col-0 and the vtc2-1 mutant, whereas daylength only has a significant effect on vtc2-4 mutant plants (p < 0.05).
(p < 0.05) ( Figure 3B). However, under both growth conditions, vtc2-1 mutants displayed a noticeable and statistically significant (p < 0.001) increase in the proportion of F v /F m values below 0.75 when compared to wildtype, whereas vtc2-4 mutants only exhibit a marginal and statistically non-significant increase ( Figure 3B). These observations are consistent with vtc2-1 mutants displaying increased leaf senescence, which is not apparent in vtc2-4 under these growth conditions.

Arabidopsis Lacking Functional GDP-L-galactose Phosphorylase has Undetectable Ascorbate and Exhibits a Seedling-Lethal Phenotype
The importance of the L-ascorbate biosynthetic pathway for Arabidopsis growth and development was demonstrated by Dowdle et al. (2007) when they showed that vtc2-1;vtc5 double mutants displayed a growth arrest following germination; a phenotype that was also associated with cotyledon bleaching. Given that vtc2-1 is not a complete loss-of-function allele and also apparently harbors a cryptic mutation/s that affect growth, we reconstituted the vtc2;vtc5 double mutant line by combining the vtc2-4 mutant with two different vtc5 mutant alleles (vtc5-1 and vtc5-2; Dowdle et al., 2007). The resulting vtc2-4;vtc5-1 and vtc2-4;vtc5-2 double mutants also exhibited a seedling-lethal phenotype, and cotyledons were subsequently bleached upon germination ( Figure 4A). These were phenotypically indistinguishable from the vtc2-1;vtc5 double mutants previously described (Dowdle et al., 2007). Consistent with the previous study, both double mutants had undetectable levels of ascorbate ( Figure 4B). The double mutants could be rescued to complete their life cycles by supplementation with ascorbate ( Figure 4C). Thus these experiments show that the phenotype of the double mutants between vtc2-1 and vtc5 produced by Dowdle et al. (2007) is not caused by additional mutations in vtc2-1 and confirm the importance of the L-ascorbate biosynthetic pathway for seedling viability.

DISCUSSION
It has been widely reported that Arabidopsis mutants with severe ascorbate deficiency exhibit a small growth phenotype and that this decreased growth is thus casually linked to ascorbate deficiency (Jander et al., 2002;Müller-Moulé et al., 2004;Pavet et al., 2005;Dowdle et al., 2007;Giacomelli et al., 2007;Kerchev et al., 2011). Here we show that the decreased growth phenotype of the vtc2-1 mutant line is not linked to ascorbate deficiency using two approaches; (1) the vtc2-4 mutant, an apparent null mutant which has comparable levels of ascorbate to vtc2-1, shows a different growth profile when compared to vtc2-1 mutants; (2) after backcrossing vtc2-1 to wildtype plants we readily isolated vtc2-1 ascorbate-deficient lines with weight and leaf sizes that are closer to wildtype than vtc2-1 mutants, as well as lines with a small growth phenotype that were homozygous for VTC2. It appears that the decreased growth phenotype of the vtc2-1 line is caused by an additional mutation/s that have likely arisen during the original EMS mutagenesis process (Conklin et al., 1996). Because of the difficulty of accurately scoring the size of individual F2 plants it is not clear from our analysis whether the vtc2 mutation and those associated with the growth defect are linked. However, it is unlikely that they are closely linked because it was relatively easy to identify F2 individuals homozygous for the wildtype allele at one locus but homozygous mutant at the other.
The vtc1-1 mutant (Conklin et al., 1996) also exhibits a decreased growth phenotype (Veljovic-Jovanovic et al., 2001;Pavet et al., 2005;Kerchev et al., 2011;Talla et al., 2011). VTC1 produces GDP-mannose, which has important functions in cell wall carbohydrate biosynthesis and protein glycosylation in addition to ascorbate biosynthesis (Conklin et al., 1999;Lukowitz et al., 2001). It is therefore possible that deficiency of GDPmannose results in a number of different biochemical defects leading to a small growth phenotype that is not specifically due to ascorbate deficiency.
The vtc2-1 mutant line has been extensively analyzed (Müller-Moulé et al., 2003;Mukherjee et al., 2010;Gao et al., 2011;Zechmann, 2011;Botanga et al., 2012;Page et al., 2012) and has also been used to generate different mutant combinations in a number of studies. For example, interactions between redox metabolism and abscisic acid signaling were investigated by crossing vtc2-1 with an abscisic acid insensitive mutant (abi4). The vtc2-1;abi4 double mutant that was subsequently characterized, exhibited wildtype growth leading to the suggestion that abi4 suppressed the growth phenotype of vtc2-1 (Kerchev et al., 2011). It is plausible that from the cross, the vtc2-1;abi4 double mutant selected for analysis lacked the cryptic mutation(s) conferring the small growth phenotype. Furthermore, the use of whole genome microarray analysis to characterize differences between vtc2-1 and vtc2-1 abi4 mutant lines is likely to have been complicated by the possible absence of the cryptic mutation(s) in the double mutant background (Kerchev et al., 2011). The decreased growth phenotype of vtc2-1 has also been described in combination with mutations affecting chronic photo-oxidative stress (Müller-Moulé et al., 2004), chloroplastic ascorbate peroxidase (Giacomelli et al., 2007) and the autoimmune response (Zhu et al., 2013).
Our data indicate that the ascorbate deficiency of vtc2 mutants does not cause a large decrease in growth under laboratory conditions. While the vtc2-4 has a slightly smaller rosette area than wildtype, relative leaf expansion rate is identical. It is likely that vtc2-4 seedlings have smaller initial seedling size or vigor resulting in a magnifying size difference as the plants grow. Since the fully ascorbate deficient seedlings of the vtc2;vtc5 double mutant are not viable, it is apparent that a reduction greater than the ∼80% occurring in the vtc2-4 mutant is needed to impact growth. Ascorbate is amongst the most abundant of primary metabolites in Arabidopsis leaves (compare the widely reported values of 2-20 µmol g −1 fresh weight for ascorbate with the most abundant sugars, organic acids and amino acids reported in Szecowka et al., 2013). The presence of ascorbate concentrations in leaves that are much greater than needed to support growth in laboratory conditions emphasizes the importance of its protective functions for plants growing in stressful or fluctuating natural environments (Munné-Bosch et al., 2013).

CONCLUSION
In summary, we suggest that the interpretation of some studies of the commonly used vtc2-1 line or its derivatives may need to be re-evaluated. It is also important to take into consideration that the apparent cryptic mutation(s) affecting growth of vtc2-1 may have a significant impact on physiological status and gene expression unrelated to ascorbate deficiency, so although it was critical in identifying its biosynthetic pathway, it is not suitable for investigating the wider functions of ascorbate.

AUTHOR CONTRIBUTIONS
BL, NS, and JG designed and performed the experiments, analyzed the data and drafted the manuscript. CC conceived the study and its design and coordination, and assisted with revisions of the manuscript. All authors read and consented to the final version of the manuscript.

ACKNOWLEDGMENTS
This research is supported by a grant from the Australian Research Council (BL, CSC, JFG) and a Biotechnology and Biological Sciences Research Council grant (NS; BB/G021678/1).

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fpls.2016.01025 FIGURE S1 | Reverse transcription polymerase chain reaction analysis of the vtc2-4 mutant. Total RNA was extracted from 10-day old seedlings and RT-PCR was performed using VTC2 specific primers (top panel) that flank the full-length transcript. ACTIN2 (ACT2) specific primers were used as an internal control (bottom panel).
FIGURE S2 | Fresh weight measurements of short-day grown plants. Shoot fresh weights of indicated lines were assessed after 6-week growth under short-day conditions. Error bars indicate SE (n = 7-10). Significant differences were determined using a Student's t-test and statistical differences (p < 0.01) from Col-0 wildtype indicated with "a," whereas "b" denotes differences from vtc2-1 mutants.