Abnormal leaf development of rpt5a mutant under zinc deficiency reveals important role of DNA damage alleviation for normal leaf development

Leaf development in plants, including dorsoventral (adaxial–abaxial) patterning, is tightly regulated. The involvement of several subunits of the 26S proteasome in adaxial–abaxial polarity establishment has been reported. In the present study, we revealed that in Arabidopsis thaliana, a mutation in RPT5A, a subunit of 26S proteasome, causes abnormally narrow true leaves under zinc deficiency. mRNA accumulations of DNA damage marker genes in leaves were elevated by zinc deficiency. PARP2, a single-strand break (SSB) inducible gene, was more strongly induced by zinc deficiency in rpt5a mutants compared with the wild type. A comet assay indicated that SSB is enhanced in mutants grown under the zinc deficiency condition. These results suggest that SSB accumulation is accompanied by abnormal leaf development. To test if DNA damage is a sole cause of abnormal leaf development, we treated the wild type grown under normal zinc conditions with zeocin, a DNA damage-inducing reagent, and found that narrow leaves developed, suggesting that DNA damage is sufficient to induce the development of abnormally narrow leaves. Taken together with the observation of the abnormal leaf morphology of our mutant plant under zinc deficiency, we demonstrated that the alleviation of DNA damage is important for normal leaf development.

abaxial-promoting genes and results in very narrow blade leaves with a phloem-surrounding-xylem structure or radically symmetric needle-like leaves with no apparent vascular bundle. The disruption of other regulatory particles has also been reported to have a pleiotropic effect on leaf development. Mutants of RPT2A exhibit enlarged rosette leaves with enlarged cell size, which has been attributed to extended endoreduplication 10 . RPN10 is suggested to be involved in abscisic acid signaling, and its mutant exhibits a pleiotropic phenotype with a slower rate of leaf initiation and expansion and premature senescence 11 . RPN12 is considered to be involved in cytokinin regulation, and rpt12a-1 exhibits a slower rate of rosette leaf emergence 12 .
In the present work, we revealed that defects of a 26S proteasome subunit RPT5A result in abnormal, filamentous leaf formation under the zinc deficiency condition without as1 or as2 mutations. The 26S proteasome RPT5 subunit is encoded by two paralogous genes: RPT5A and RPT5B. In Col-0 accession, it is known that RPT5B compensates rpt5a mutation to some extent, but the functions of the two are not identical 13 . Previously, we reported that rpt5a-4 has reduced proteasome activity and increased accumulation of poly-ubiquitinated proteins 14 . It revealed that rpt5a-4 exhibits reduced shoot growth under zinc deficiency but not under normal conditions, suggesting that RPT5A, but not RPT5B, is involved in zinc deficiency tolerance, possibly through the alleviation of oxidative stress and/or processing of polyubiquitinated proteins. Zinc is known as a cofactor for Cu/Zn-superoxide dismutase (SOD), a major reactive oxygen species scavenger, and its deficiency decreases SOD activity in plants 15,16 . However, the relationship between the increased oxidative stress and leaf morphology remains unknown.
The possible involvement of oxidative stress in abnormal leaf development motivated us to examine the involvement of DNA damage, because generally, oxidative stress induces DNA damage 17 . The effects of the DNA damage response on leaf development have been reported. One effect is in cell compensation, where reduced cell division is compensated by increasing cell size to maintain the final leaf area (reviewed by 18 ). An analysis of the fas1 (fasciata1) mutant, which suffers from endogenous DNA damage, suggested that the ATAXIA TELANGIECTASIA MUTATED (ATM)-dependent DNA damage response is a trigger for compensated cell expansion 19 . Another effect of DNA damage on leaf development has been found in the Elongator complex, an evolutionarily conserved histone acetyltransferase complex that is an important regulator of the mitotic cell cycle to promote leaf patterning 20 . Loss of Elongator functions induced aberrant DNA replication and increased DNA damage, resulting in severe leaf polarity aberrance 20 .
Through this work, we dissect the mechanism of abnormal leaf development under zinc deficiency to shed light on the effect of DNA damage on leaf development. By analyzing the Arabidopsis thaliana rpt5a-4 mutant, we demonstrate that RPT5A is important for alleviating DNA damage under zinc deficiency and that DNA damage is sufficient to trigger the development of aberrant filamentous leaves.

Results
Zinc deficiency induced filamentous true leaves in rpt5a-4 and rpt2a-2 mutants. To investigate the involvement of RPT5A in leaf development under zinc deficiency, we observed first true leaves of the wild type and rpt5a-4 grown under control (1 µM zinc) or zinc deficiency (no zinc added) conditions. Grown under the control condition, rpt5a-4 true leaves exhibited morphological differences from the wild type (Fig. 1A). The wild type exhibited broader leaf blades and round leaf tips (Fig. 1B), while rpt5a-4 exhibited narrower leaf blades and edged leaf tips (Fig. 1C). This phenomenon has also been mentioned previously 9 . Under the zinc deficiency condition, no significant effect on leaf morphology was observed in the wild type. On the other hand, rpt5a-4 There is no significant difference in the ratio of ("round tip" + "edged tip"): ("narrow" + "filamentous") at p < 0.05 among groups sharing the same letter by Fisher's exact test with Holm correction.
leaf morphology was significantly affected by zinc deficiency. The majority of leaves exhibited reduced leaf blade width, with an ambiguous boundary between leaf blades and petioles (Fig. 1D,F). Approximately 10% of true leaves exhibited more severe morphological alteration: their leaf blades were undeveloped and indistinguishable from petioles (Fig. 1E).
To genetically establish that the loss of function in RPT5A is responsible for the abnormal leaf formation under zinc deficiency, we conducted a complementation test by introducing a genomic fragment of RPT5A cloned from the wild type into rpt5a-4. We obtained three independent homozygous transformant lines. Expression of the transgene in these transformants was confirmed by qRT-PCR ( Supplementary Fig. 1). All of the obtained transformant lines exhibited non-or less-frequent narrow or filamentous leaves compared with rpt5a-4 under zinc deficiency conditions. This result indicates that introducing a RPT5A genomic fragment into rpt5a-4 suppresses abnormal leaf formation in rpt5a-4 under zinc deficiency, establishing that RPT5A is responsible for circumventing abnormal leaf formation under zinc deficiency.
To examine whether the observed phenomenon is specific to RPT5A among proteasome components, we observed leaf morphology of a mutant of another proteasome subunit, RPT2A. As is the case in rpt5a-4, rpt2a-2 exhibited narrow leaf blades or filamentous leaves under zinc deficiency ( Supplementary Fig. 2), suggesting that involvement in leaf development under zinc deficiency is not specific to RPT5A, but at least common between the two proteasome components.
Adaxial identity was defective in rpt5a-4 leaves under zinc deficiency. A needle-like leaf structure is generally thought to be a result of defective leaf adaxial-abaxial polarity 9 . To investigate the alteration in adaxial-abaxial fate, we observed trichomes and vascular bundle patterns in the blade-petiole conjugation region in rpt5a-4 under zinc deficiency.
No trichome development was observed in severely filamentous leaves of rpt5a-4 under zinc deficiency, whereas rpt5a-4 under the control condition developed trichomes on the adaxial leaf surface ( Fig. 2A,B), suggesting that adaxial identity was not properly established in rpt5a-4 under zinc deficiency. Under zinc deficiency, rpt5a-4 infrequently developed a horn-like structure from the abaxial side of emerging leaves (Fig. 2C). The horn-like structure was not consecutive to the filamentous leaf development. Observation of the cross-sections revealed that filamentous rpt5a-4 leaves under zinc deficiency did not develop the expanded leaf blade but had the midrib alone, whereas the rpt5a-4 under control conditions and the wild type under either condition developed the leaf blade ( Fig. 2D-G). The most severely filamentous leaves did not develop the leaf blade at all (Fig. 2H).
In the wild-type vascular bundle, xylem is developed in the adaxial side, and phloem is developed in the abaxial side (Fig. 2I,J). The relative positioning of xylem and phloem was maintained in filamentous rpt5a-4 leaves ( Fig. 2K-M). However, the vascular bundle and tracheary element of xylem were thinner in rpt5a-4, which was more evident under zinc deficiency. In some filamentous leaves, two independent vascular bundles were observed (Fig. 2M). The most severely filamentous leaves did not develop the vascular bundle (Fig. 2H). These observations suggest that adaxial identity is defective in rpt5a-4 leaves developed under zinc deficiency.

expression of abaxial determinant genes was suppressed in rpt5a-4 leaves under zinc deficiency.
To investigate possible effects of zinc deficiency on leaf development, we examined the expression of adaxial and abaxial marker genes in the developing true leaves (Fig. 3). RNA was extracted from round-tip leaves from the wild type under the control or zinc deficiency conditions, edged-tip leaves from rpt5a-4 under the control condition, and narrow or filamentous leaves from rpt5a-4 under the zinc deficiency condition. qRT-PCR revealed that mRNA accumulation of the adaxial marker gene, AS2, was reduced, whereas PHV and PHB were not significantly affected by zinc deficiency or by rpt5a-4 mutation. On the other hand, abaxial marker genes, ETTIN (ETT) and YAB5, were significantly reduced in rpt5a-4 under zinc deficiency. These results suggest that RPT5A is required for the maintenance of adaxial-abaxial balance under zinc deficiency.
DNA damage was accumulated in rpt5a-4 leaves under zinc deficiency. To investigate the possible involvement of DNA damage in leaf development under zinc deficiency, we first measured mRNA accumulation of four DNA damage marker genes in shoots by qRT-PCR (Fig. 3). Rad51, Ku80, and BRCA1 tended to be induced by zinc deficiency, but no significant differences were detected between those of the wild type and rpt5a-4. Among the four genes, only PARP2 (At4g02390) exhibited a difference between the wild type and rpt5a-4 under zinc deficiency. In PARP2, induction by zinc deficiency was evident in rpt5a-4 but not significant in the wild type.
To assess DNA damage accumulation directly, we conducted a comet assay. We adopted two variations of the comet assay to distinguish a double-strand break (DSB) from a single-strand break (SSB): the neutral/neutral (N/N) method detects mainly DSB, while the alkali/neutral (A/N) method detects DSB and SSB 21 . Tail DNA percentage detected by the N/N method was increased by zinc deficiency in both genotypes, but no significant differences were detected between the wild type and rpt5a-4 (Fig. 4A). Tail DNA percentage detected by the A/N method was induced by zinc deficiency, and the levels were higher in rpt5a-4 compared with those of the wild type (Fig. 4B). These results suggest that zinc deficiency causes DNA damage in shoots and that RPT5A is involved in the alleviation of SSB in shoots.
DNA damage was sufficient to induce development of filamentous leaves. As DNA damage was highly accumulated in rpt5a-4 shoots under zinc deficiency, we hypothesized that the abnormal leaf morphology results from the accumulation of DNA damage. To test this hypothesis, we treated the seedlings with zeocin, a DNA damage-inducing reagent, to artificially induce DNA damage, and investigated whether it affected leaf morphology. In rpt5a-4, zeocin treatment with a concentration range of 1.3-3.3 µM produced filamentous or narrow leaves similar to those observed under zinc deficiency (Fig. 5A,C). With the severer treatment with 6.6 µM zeocin, rpt5a-4 was not viable, whereas the wild type was able to develop true leaves, suggesting that rpt5a-4 is www.nature.com/scientificreports www.nature.com/scientificreports/ more sensitive to DNA damage. The induction of filamentous or narrow leaves by zeocin was observed even in the wild type, although the sensitive dose range was higher than that of rpt5a-4, ranging from 3.3-6.6 µM (Fig. 5A,B).
To test whether the observed difference between the wild type and rpt5a-4 in sensitivity to zeocin treatment was attributable to a difference in DNA damage repair ability, we conducted a comet assay with rpt5a-4 treated with 3.3 µM zeocin (Fig. 5D-M). With A/N method, which detects SSB and DSB, higher damage accumulation was observed in rpt5a-4 treated with zeocin (Fig. 5D), whereas no significant difference was detected with N/N method, which detects DSB alone (Fig. 5E). This suggests that it is SSB rather than DSB that is highly accumulated in rpt5a-4 under zeocin treatment.
To test the generality of this DNA damage-induced abnormal leaf development in other types of abiotic stress, we observed the leaf morphology of rpt5a-4 under excess boron stress, which is known to induce DNA damage 22 . The wild type developed normal leaf morphology under toxic conditions with 3 mM boric acid ( Supplementary  Fig. 3). On the other hand, with 3 mM boric acid, most rpt5a-4 true leaves had reduced leaf blade width, with an ambiguous boundary between leaf blades and petioles ( Supplementary Fig. 3 B). Comet assay revealed that excess boron stress resulted in DNA damage accumulation pattern which is similar to zeocin treatment (Supplementary  Seedlings were grown for 14 days and total RNA was extracted from the first pair of true leaves with the representative shapes; for the control condition, the wild type and rpt5a-4 were represented by "round tip" and "edged tip", respectively, and for the zinc deficiency condition, the wild type was represented by "round tip" and "edged tip, " whereas rpt5a-4 was represented by "narrow" or "filamentous. " mRNA accumulations were normalized by the geometric mean of Actin8, PEX4, and SAND. Values represent mean ± standard deviation of four biological replicates. Groups sharing the same letter were not significantly different from each other at p < 0.05 by Tukey's multiple test.  Supplementary Fig. 3D), whereas no significant difference was detected with N/N method ( Supplementary  Fig. 3C), indicating that excess boron stress induces SSB in shoots. These results suggest that the mechanism of developing abnormal leaves where RPT5A plays a crucial role is not specific to zinc deficiency conditions but also occurs under abiotic stress conditions where DNA damage accumulation is involved, at least with excess boron stress.

Discussion
In this work, we demonstrated that proteasome components, at least RPT5A and RPT2A are essential for normal leaf development, especially under a specific nutritional environment. The combination of rpt5a-4 mutation and zinc deficiency or excess boron triggers severely aberrant leaves, but each factor alone does not (Figs 1 and S3). The phenotype of rpt5a-4 leaves under zinc deficiency or excess boron conditions is similar to that of the ae3-1 mutant: narrow or needle-like rosette leaves, which are attributed to a defective adaxial identity 9 . The ae3-1 mutation was isolated as an enhancer of as2-101 and identified as a mutation in RPN8a, a lid subunit of the 26S proteasome 9 . Huang et al. reported that abnormal rosette leaves were also observed in a selfed F 2 population of a cross between as2-101 and other proteasome subunit mutants: rpt2a, rpt4a, rpn1a, rpn8a, rpn9a, pad1, pbe1, as well as rpt5a. Based on this result, they suggested that the 26S holoenzyme plays a role in regulating leaf patterning. Along this line, our findings can be interpreted as indicating a zinc deficiency or excess boron stress effect on rpt5a-4 phenocopying aberrant leaf formation in double mutants of as2 and proteasome mutants, which suggests that these stress conditions affect the AS2 pathway in leaf adaxial-abaxial determination.
Morphological observation revealed that adaxial-abaxial determination failed in rpt5a-4 under zinc deficiency. The observations, including the absence of trichomes, absence of an expanding leaf blade, and reduction in AS2 expression, suggest that the adaxial identity was defective. In addition, abaxial determinant genes, ETT and YAB5, were less accumulated in rpt5a-4 than in the wild type. Given that the expression of both abaxial and adaxial marker genes is reduced in rpt5a-4 under zinc deficiency, it can be interpreted to indicate that the abaxial-adaxial determination fails and the tissue is undifferentiated.
Our results demonstrated that DNA damage-inducing treatment was sufficient to induce abnormal leaf development even in the wild-type plants. In this context, considering that DNA damage accumulated more in rpt5a-4 than in the wild type under zinc deficiency (Fig. 4), it is deduced that RPT5A contributes to normal leaf development through the alleviation of DNA damage induced by zinc deficiency. This is supported by the results that zeocin treatment induced abnormal leaf formation in both the wild type and rpt5a-4 and that the sensitivity was higher in rpt5a-4 than in the wild type. How does RPT5A alleviate DNA damage? Under zinc deficiency, the level of lipid peroxidation, an indicator of oxidative stress level, is higher in rpt5a-4 and rpt2a-2 compared with the wild type 14 . This suggests that these proteasome subunits are involved in alleviating oxidative stress. As oxidative stress induces DNA damage 17 , it is deduced that the reduced capacity for alleviating oxidative stress in rpt5a-4 resulted in DNA damage accumulation. We do not exclude the possibility that RPT5A is also involved in other pathways, namely DNA damage prevention and/or repair.
Zinc deficiency tended to induce the expression of the four DNA damage marker genes we examined (Fig. 3). However, a significant difference between the wild type and rpt5a-4 under zinc deficiency was detected only in PARP2. In mammalian cells, it is described that PARP1 and PARP2 contribute to SSB repair/base excision repair processes 23 . The direct measurements of DSB and SSB suggest that zinc deficiency induces both DSB and SSB and RPT5A is involved in the alleviation of SSB under these conditions (Fig. 4). Besides the DNA damage response, the induction of PARP in rpt5a-4 under zinc deficiency itself may be involved in leaf development. PARP is known to be involved in the stress response and growth control of plants 24,25 . Schulz et al. found that PARP activity is involved in leaf cell number by demonstrating that the inhibition of PARP activity by 3-Methoxybenzamide (3MB) treatment results in an increased cell number with more cell divisions, especially in young leaves. In light of this, the increased expression of PARP2 may result in reduced cell division in developing leaves, which could affect the leaf morphology.
Since rpt5a-4 mutation affected SSB, but not DSB, we speculate that it is SSB rather than DSB that induces abnormal leaf formation. From this point of view, the notion that a well-known DSB-causing reagent, zeocin, induced abnormal leaf development might sound controversial. However, although zeocin is generally regarded as a DSB-inducing reagent, it is reasonable to believe that it also induces SSB, because a structurally related DNA damage-inducing reagent, bleomycin, is expected to react with oxygen to generate free radicals, resulting in SSB as well as DSB [26][27][28] . Although zinc deficiency but not excess boron or zeocin treatment induced DNA damage which is detected by N/N method comet assay, with A/N method we observed high damage accumulation in rpt5a-4 consistently among these three treatments (Figs 4 and 5D-E, Supplementary Fig. 3C-D). These results also support the idea that SSB rather than DSB is responsible for the abnormal leaf formation.
Our finding is not the first instance of the involvement of DNA damage in leaf development in a broad sense. An analysis of fas1, a mutant of a large subunit of chromatin assembly factor-1 (CAF-1), revealed that DNA damage induces endoreduplication and increases leaf cell size 29 . However, in the present work, we demonstrated that DNA damage is involved in not only cell size determination but also developmental axis determination. We believe that our findings introduce a new perspective of leaf development. Intensive research studies have revealed comet assay with N/N method (F-I) or A/N method (J-M). Nuclei which give median Tail DNA [%] are shown with its value. (F,J) Wild type, control, (G,K) wild type, 3.3 µM zeocin, (H,L) rpt5a-4, control, (I,M) rpt5a-4, 3.3 µM zeocin. Bar, 50 µm. Contrast and brightness were adjusted for visualization purpose. Pseudo-color scale is shown in the bottom. DNA damage assessment by comet assay. For nuclear isolation, shoots of 5 to 20 seedlings were minced in PBS with a razor blade approximately 100 times, followed by filtration through a CellTrics 30-µm filter (Sysmex Partec, Görlitz, Germany) with a centrifuge at 800 g for 1 min. For the neutral/neutral (N/N) method 21 , 10 µL of the supernatant was taken from the bottom of the tube. For the alkali/neutral (A/N) method 21 , the chloroplast was removed by repeating the following steps: 1. Add 10% v/v of 10× suc. sln. (50% sucrose, 2% Triton-X100 in PBS) to the supernatant and mix it well by inversion. 2. Centrifuge 1 min at 8,000 g and remove the supernatant. 3. Resolve the pellet in 450 µL of PBS. Steps 1 to 3 were repeated to obtain a white pellet. In Step 3 of the final iteration, the pellet containing nuclei was suspended in 20 µL of PBS.
The comet assay was performed using a CometAssay Kit (25 × 2 well slides) (Trevigen, Gaithersburg, MD, USA) with the following modifications. First, 10 µL of the nuclear extract was combined with 100 µL molten LMAgarose held at 37 °C, and 50 µL of the mixture was immediately pipetted onto CometSlide. The slides were kept at 4 °C in the dark for 30 min. The slides were immersed in a lysis solution (2.5 M NaCl, 10 mM Tris-HCl, pH8.0, 100 mM EDTA (self-prepared, not the one supplied in the kit)) (for the N/N method) or alkaline unwinding solution (0.3 M NaOH, 5 mM EDTA) (for the A/N method) for 20 min at room temperature. The slides were washed three times with TBE buffer for 5 min on ice. The slides were subjected to electrophoresis on ice with TBE buffer 1 V/cm for 6 and 4 min for the N/N method and the A/N method, respectively. The slides were soaked in 1% Triton-X100 in PBS for 10 min at room temperature and then washed twice with 70% ethanol for 5 min and twice with 96% ethanol for 5 min. The slides were dried at 37 °C for 1 h. For staining the nuclei, 100 µL of SYBR Gold staining solution (SYBR Gold Nucleic Acid Gel Stain (10,000 × Concentrate in DMSO, Thermo Fisher Scientific) diluted 30,000 times with TE pH 8.0) was placed onto CometSlide and kept for 30 min at room temperature in the dark. The slices were rinsed in water and dried completely at 37 °C. The fluorescence was observed with fluorescent microscope BX50WI with filter unit U-MWIBA3 (OLYMPUS, Tokyo, Japan), and the images were captured with an equipped CCD camera (OLYMPUS).
Images were quantified with CASP 35 and with ImageJ macro OpenComet 36 for the N/N method and the A/N method, respectively.