A DNA2 Homolog Is Required for DNA Damage Repair, Cell Cycle Regulation, and Meristem Maintenance in Plants1[OPEN]

A DNA2 homolog plays an important role in DNA damage repair, cell cycle progression, and the development of the shoot apical meristem and root apical meristem in higher plants. Plant meristem cells divide and differentiate in a spatially and temporally regulated manner, ultimately giving rise to organs. In this study, we isolated the Arabidopsis jing he sheng 1 (jhs1) mutant, which exhibited retarded growth, an abnormal pattern of meristem cell division and differentiation, and morphological defects such as fasciation, an irregular arrangement of siliques, and short roots. We identified JHS1 as a homolog of human and yeast DNA Replication Helicase/Nuclease2, which is known to be involved in DNA replication and damage repair. JHS1 is strongly expressed in the meristem of Arabidopsis. The jhs1 mutant was sensitive to DNA damage stress and had an increased DNA damage response, including increased expression of genes involved in DNA damage repair and cell cycle regulation, and a higher frequency of homologous recombination. In the meristem of the mutant plants, cell cycle progression was delayed at the G2 or late S phase and genes essential for meristem maintenance were misregulated. These results suggest that JHS1 plays an important role in DNA replication and damage repair, meristem maintenance, and development in plants.

In higher plants, organ morphogenesis depends on the persistent activity of apical meristems during development (Donnelly et al., 1999;Fleming, 2006). Apical meristem development is sustained and driven by cell division in meristematic regions (Traas and Bohn-Courseau, 2005;Zhou et al., 2011). Plant meristems, which contain stem cells and have a powerful regenerative ability, enable sustained plant growth and give rise to new organs such as roots, stems, leaves, and flowers. Plant architecture and the final form of organs depend on spatially and temporally regulated cell proliferation in meristems.
Cells in the shoot apical meristem (SAM) function coordinately and are distributed in distinct regions, namely the central zone, which includes stem cells and the organizing center, the peripheral zone, and the rib zone (Aichinger et al., 2012). Meristem-specific genes, such as SHOOT MERISTEMLESS (STM), WUSCHEL (WUS), and CLAVATA3 (CLV3), are key regulators of SAM formation (Aichinger et al., 2012). STM encodes a transcription factor that promotes cytokinin biosynthesis and inhibits gibberellin biosynthesis, thus preventing cell differentiation in the SAM (Yanai et al., 2005). WUS, which also encodes a transcription factor, defines the organizing center (OC) of SAMs (Mayer et al., 1998;Lenhard and Laux, 2003). CLV3 encodes a small protein that can be processed into a small peptide, which interacts with and probably activates a putative receptor kinase, CLAVATA1 (CLV1), to regulate SAM size (Clark et al., 1997;Fletcher et al., 1999;Shinohara and Matsubayashi, 2015). The SAMs of clv1 and clv3 mutants are significantly enlarged relative to the wild type and produce more floral meristems than wild-type plants (Clark et al., 1993(Clark et al., , 1995. The wus mutant lacks a SAM and overexpression of WUS results in enlarged meristems, suggesting that WUS prevents the differentiation of stem cells and is essential for maintaining the structural and functional integrity of the SAM (Mayer et al., 1998;Lenhard and Laux, 2003). In the SAM, a feedback regulation loop exists between WUS and CLV3; WUS maintains CLV3 expression in the OC and CLV3 functions as an intercellular signal that represses WUS transcription (Schoof et al., 2000). The controlled expression of WUS in the distal region of the OC is important for maintaining the boundaries between stem cells and differentiating cells.
Quiescent center (QC) cells are the mitotically inactive cells in the center of the root apical meristem (RAM) that inhibit differentiation of the surrounding stem cells (van den Berg et al., 1997). Together with their adjacent cells, QC cells establish intercellular signal communication and regulate the balance between division and differentiation in tissues such as the steel, ground tissue, epidermis, lateral root caps, and starch-containing and gravity-sensing columella (Lenhard and Laux, 2003). WUSCHEL-RELATED HOMEOBOX5 (WOX5) is a homolog of WUS that is specifically expressed in QC cells (Haecker et al., 2004). WOX5 maintains stem cells in the RAM and represses the differentiation of columella stem cells (Sarkar et al., 2007). Several CLV3-related peptides, such as CLE40, antagonize WOX5 function and promote differentiation of stem cells (Stahl et al., 2009).
Several fasciated mutants display similar developmental defects as clv3 and clv1. However, the genes affected in most of those mutants seem not to be directly related to CLV3 and CLV1 function. For example, FAS1 and FAS2, which encode proteins identified as counterparts of human Chromatin assembly factor1, have also been shown to be involved in maintaining the cellular and functional organization of the SAM and RAM, and mutation of either FAS1 or FAS2 causes fasciation, abnormal phyllotaxy, and defective root growth (Kaya et al., 2001). Further analyses suggested that fas1 is sensitive to DNA damage agents and exhibits enhanced endoreplication (Endo et al., 2006;Kirik et al., 2006). In addition, fasciated mutants with defects in BRU1/MGO3/TSK, which is thought to be involved in chromatid separation during mitosis (Takeda et al., 2004;Suzuki et al., 2005), are hypersensitive to DNA damage agents, show a release of gene silencing, and are delayed in cell cycle progression. Furthermore, MGO1, which encodes a type-IB DNA topoisomerase, was shown to be involved in stem cell regulation Graf et al., 2010). DNA polymerase « was shown to influence the cell cycle, embryo patterning, epigenetic regulation, somatic homologous recombination, and ABA signaling (Jenik et al., 2005;Yin et al., 2009). teb, a mutant of DNA polymerase u, exhibited a constitutively activated DNA damage response and defects in cell cycle progression (Inagaki et al., 2006). Mutants of MRE11 and BRCA2, which function in DNA damage repair, were also shown to have a fasciated phenotype (Bundock and Hooykaas, 2002;Abe et al., 2009). These data together suggest that chromatin assembly, DNA replication, DNA damage repair, cell cycle regulation, and related events are important for meristem organization and plant development.
DNA2 is a conserved essential protein with a helicase and a nuclease domain, and is thought to be involved in Okazaki fragment synthesis and DNA damage repair (Budd et al., 2006). During maturation of the Okazaki fragment, the short RNA/DNA primer flap is removed by FEN1 (Zheng and Shen, 2011). However, long flaps may also occur, and these are removed by the sequential action of DNA2 and FEN1 (Kang et al., 2010). During DNA break repair, DNA2 mediates 59-end resection of DNA by cleaving the 59-single-stranded DNA (ssDNA), with the aid of Sgs1 and RPA Mimitou and Symington, 2009;Cejka et al., 2010). Sgs1 is a helicase, whose activity is stimulated by RPA, and which can unwind DNA and produce an ssDNA substrate for DNA2. Homologs of Sgs1 and RPA are also conserved in plants with important roles in DNA repair and other processes (Aklilu et al., 2014;Schröpfer et al., 2014). Yeast DNA2 has Pat4, Pat7, and bipartite nuclear localization signal (NLS) sequences and is localized to the nucleus (Zheng et al., 2008). Yeast dna2 mutants exhibit increased sensitivity to DNA-damaging agents, such as UV, MMS (methyl methane sulfonate), and x-ray irradiation, and yeast DNA2 participates in DNA repair by facilitating homologous recombination (Budd and Campbell, 2000). In mammals, DNA2 plays a role in DNA replication and repair in both mitochondria and nuclei (Zheng et al., 2008;Duxin et al., 2009). In humans, depletion of hDNA2 results in cell cycle delay and aberrant cell division (Duxin et al., 2009). DNA2 is widely believed to have important roles in DNA replication and repair and in mitochondrial and nuclear genomic DNA maintenance in metazoans and fungi (Kang et al., 2010). However, the role of DNA2 in plants has yet to be investigated.
In this study, we isolated an Arabidopsis mutant, jing he sheng1 (jhs1), which exhibits fasciation, abnormal phyllotaxy, and defective root growth, as observed in fas1, fas2, and numerous other mutants as mentioned above. Map-based cloning indicated that JHS1 encodes a DNA2 homolog. Our analysis suggests that this DNA2-like gene acts in the response to stalled DNA replication and is also involved in double-stranded DNA break repair. Further results indicated that a mutation in this gene causes an increased sensitivity to DNA damage stress, accumulation of DNA damage, and a delay of cell cycle progression in plants, which may interfere with the expression of key genes involved in meristem maintenance and normal SAM and RAM development.

Mutation of JHS1 Affects SAM and RAM Development and Organ Morphogenesis
We isolated jhs1 in a screen for EMS (ethyl methane sulfonate)-induced Arabidopsis mutants with abnormal shoot growth. The mutant had a fasciated stem, which is called "jing he sheng" in Chinese. Therefore, we named this mutant jhs1.
Compared with the wild type, the growth of true leaves was retarded in young jhs1 seedlings (Fig. 1,  A and B). Bolted jhs1 plants had a fasciated stem ( Fig. 1, C-E) and abnormalities in silique arrangement (Supplemental Fig. S1) and phyllotaxy (Fig. 1C). These phenotypes suggest that SAM development is abnormal in jhs1. Sectioning of the SAM of young jhs1 seedlings indicated that the SAM was enlarged and consisted of enlarged cells with an abnormal arrangement, especially in the L1 and L2 layers (Fig. 1B). In cross section, the stem of the jhs1 mutant was oval-shaped, whereas it was cylindrical in the wild type (Fig. 1E). We analyzed the size of pavement cells of the first true leaves of 15-d-old plants. The cell size of the jhs1 mutant was significantly larger than that of the wild type (Fig. 1F). . Shoot phenotypes of the jhs1 mutant. A, The phenotype of 8-d-old wild-type and jhs1 seedlings and jhs1 seedlings complemented with JHS1 (Com) grown on MS medium. Bars = 2 mm. B, Toluidine blue-stained SAM sections of 8-d-old wildtype, jhs1, and Com seedlings. The L1 and L2 layers are outlined in wild type (left) and Com (right). Individual cells in the L1 and L2 layers are indicated in jhs1 (middle). Bars = 20 mm. C, Floral organs of 40-d-old wild-type, jhs1, and Com plants. Bars = 1 cm. D, Floral organs of 50-d-old wild-type, jhs1, and Com plants. Bars = 1 cm. E, Toluidine blue-stained stem cross sections of 40-dold wild-type, jhs1, and Com plants. Bars = 100 mm. F, Comparison of the cell size of abaxial epidermal cells in the true leaves of 15-d-old wild-type, jhs1, and Com plants. The border of a cell is marked in white. Bars = 75 mm. G, Analysis of the cell area of abaxial epidermal cells in the true leaves of 15-d-old wild-type, jhs1, and Com plants. The average area of 50 cells is shown. Error bars represent SD. Student's t-test; for **, P , 0.01. WT, wild type.
The average cell size was 3101 mm 2 in the wild type, but 6502 mm 2 in the mutant (Fig. 1G). Thus, jhs1 has defects in cell division in the SAM and in SAM-derived organs, such as stems, leaves, and siliques.
In 10-d-old plate-grown plants, the root length of jhs1 mutants was approximately one-fifth that of the wild type (Fig. 2, A and B), suggesting that root growth is severely retarded in the mutant. Observation with a stereomicroscope revealed that the first root hair was much closer to the root tip in jhs than in the wild type, indicating that the RAM and elongation zone are much shorter in the mutant (Fig. 2C). Confocal microscopy analysis of roots stained with FM4-64 (a plasma membrane dye) further indicated that the cellular arrangement was disorganized in the RAMs of jhs1, that QC cells were irregular and difficult to distinguish, and that cells in the elongation zone were larger and not as organized as those in the wild type ( Fig. 2D and Supplemental Fig. S2). The RAM was also shorter in jhs than in the wild type, with fewer cells (Fig. 2D). To determine the location of the QC, we transformed jhs and wild-type plants with ProWOX5:GFP, which was previously shown to label the QC (González-García et al., 2011). In the jhs mutant, the QC was not clear and ProWOX5:GFP expression was altered in comparison with the wild type (Supplemental Fig. S2). We then stained the roots with Lugol's solution (a starch dye) to analyze the columella root cap. Whereas wild-type roots had four tiers of columella root cap cells with a regular arrangement (Fig. 2E, left panel), root cap columella cells were severely disordered in jhs1 (Fig. 2E, middle panel).
As shown in Figure 2A, the development of roots, including lateral roots, was severely affected in jhs1. We generated transgenic plants expressing DR5:GUS, a marker of auxin (Ulmasov et al., 1997), to further study the development of lateral roots in jhs1. b-glucuronidase (GUS) staining indicated sites of lateral root emergence (Supplemental Fig. S3A). In 9-d-old wild-type plants, lateral roots initiated and elongated normally (Supplemental Fig. S3A). However, in jhs1, the number of lateral roots was reduced at least 4-fold (Supplemental Fig. S3, A and B).

Identification of JHS1
We crossed the jhs1 mutant with wild-type plants of the Landsberg erecta (Ler) ecotype for map-based cloning of JHS1. All F1 plants had a wild-type phenotype. Among 692 F2 progeny, 182 exhibited a mutant phenotype and 510 had a wild-type phenotype, suggesting that jhs1 is a single recessive mutation (x 2 = 0.62 , x 2 0.05(1) = 3.84). Rough mapping indicated that JHS1 was located on chromosome 1, between the molecular markers LUG887 and CH1-4.3. Fine mapping indicated that the JHS1 locus was in a region of approximately 270 kb between molecular markers CH1-2.62 and CH1-2.89 (Supplemental Fig. S4). We sequenced this region and identified a G-to-A transition mutation in At1g08840 in the jhs1 mutant plant (Fig. 3A). This mutation is at the splicing site of the 11th intron, and cDNA sequencing indicated that the splicing site in the mutant was shifted 63 bp upstream, to include a piece of intron sequence, and that this caused premature termination of the protein encoded by At1g08840 (Fig. 3A).
A 10.259-kb DNA fragment containing the wild-type At1g08840 gene and the 1.323-kb upstream sequence spanning the promoter and 59 UTR was introduced into jhs1 plants by Agrobacterium-mediated transformation. The transgenic plants were verified by genotyping, DNA sequencing, and reverse transcription (RT)-PCR analysis (Supplemental Fig. S5). All of the mutant phenotypes described above, such as retarded and abnormal development of the shoot and root, were complemented in the transgenic plants ( Figs. 1 and 2). This finding confirmed that At1g08840 is JHS1 and that the mutation in jhs1 is responsible for the mutant phenotype. We analyzed the amino-acid sequence of JHS1 by PSORT II Prediction (http://psort.hgc.jp/form2.html) and found that JHS1 contained an NLS sequence (Fig.  3B). We studied the localization of JHS1-YFP in the leaves of Nicotiana benthamiana and Arabidopsis plants transgenically expressing Pro-35S:JHS1-YFP. In both cases, the YFP signal colocalized with the DAPI (49,6diamidino-2-phenylindole) signal (Supplemental Figs. S6 and S7), suggesting that JHS1 was localized to the nucleus. In control plants expressing YFP alone, YFP was localized to both the nucleus and cytoplasm (Supplemental Figs. S6 and S7).
To determine the expression pattern of JHS1, a fusion construct in which GUS expression is driven by the JHS1 promoter (ProJHS1:GUS) was transformed into wild-type plants. GUS staining indicated that JHS1 was strongly expressed in the RAM and SAM (Fig. 4, A and B). Moreover, GUS activity was detected in the vasculature at different developmental stages (Fig. 4, C and D) and in young floral tissues (Fig. 4E).
jhs1 Is Sensitive to DNA Damage Stress In yeast, dna2 mutants are sensitive to DNA damage stress induced by MMS and x-ray irradiation and DNA2 was shown to have an important role in DNA damage repair (Formosa and Nittis, 1999;Budd and Campbell, 2000). Therefore, it is likely that the jhs1 mutant is also sensitive to DNA-damaging agents. To test this hypothesis, we examined the effects of two types of DNA-damaging agents, HU (hydroxyurea) and zeocin, on jhs1 plants. HU, a DNA replication stress-inducing agent that functions as a ribonucleotide reductase inhibitor, reduces the dNTP pool and results in stalked replication forks with ssDNA, whereas zeocin, a radiomimetic drug of the bleomycin family, induces a double-strand break (DSB).
We evaluated the effect of HU on the growth of 15-dold plate-grown plants. Treatment with 1.0 mM and 2.0 mM HU severely limited the growth of the mutant, but not of the wild type. In the mutant, the fresh weight of treated plants was only 78% and 38% that of untreated plants, respectively. By contrast, the fresh weight of treated plants was 92% and 78% that of untreated plants in the wild type, respectively (Fig. 5, A and E). We also studied the effect of HU on root growth in the mutant. The root length of jhs1 plants treated with 1.0 mM HU was 28% that of the untreated jhs1 plants. However, on plates containing 2.0 mM HU, root growth was severely inhibited in both the wild type and jhs1, and no difference was observed between these lines (Fig. 5,C and F).
We then evaluated the effect of zeocin on the growth of 15-d-old plate-grown plants. Treatment with 5 mM and 10 mM zeocin severely limited the growth of the mutant, but not of the wild type. In the mutant, the fresh weight of treated plants was only 70% and 55% that of untreated plants, respectively. By contrast, the fresh weight of treated plants was 89% and 72% that of untreated plants in the wild type, respectively (Fig. 5, B and G). We also studied the effect of 5 mM and 10 mM zeocin on root growth in the mutant. The root length of the treated mutant was reduced to 40% and 32% that of untreated plants, respectively. However, the root length of treated wild-type plants was 81% and 60% that of untreated plants, respectively (Fig. 5,D and H).

JHS1 Is Localized to Nuclear Foci in Response to DNA Damage
The above results indicate that JHS1 has an important role in DNA damage repair in Arabidopsis. To further examine the role of DNA2 in the DNA damage response, we observed the subcellular localization of YFP-JHS1 in N. benthamiana and Arabidopsis plants treated or not with zeocin or HU (Fig. 6). JHS1-YFP was localized to subnuclear foci upon zeocin treatment in both plants. Because zeocin induces DSBs (Chankova et al., 2007), JHS1 may be localized to those foci to repair the breaks. In HU-treated plants, the localization of YFP-JHS1 was similar to that of the untreated control plants. This could be because no nuclear DNA replication and cell division occurred in these leaf cells after White lines indicate the transition zone between the meristem and the elongation differentiation zone. The QC cells are outlined in white. Bars = 100 mm. E, Lugol-stained root caps of 6-d-old wild type (left), jhs1 (middle), and Com (right) seedlings. Starch granules were stained in columella root caps. Bars = 50 mm. WT, wild type. the treatment. In the absence of severe DNA damage or stalled replication forks, no YFP-JHS1 foci formed.

jhs1 Has Increased DNA Damage and Damage Responses
To examine DNA damage in the jhs1 mutant, we used the terminal deoxynucleotide transferase-mediated dUTP nick-end labeling (TUNEL) method, which detects 39-OH DNA breaks. Signals were barely detected in the leaves and root tips of the wild type (Fig.  7, A and B), whereas they were observed in numerous TUNEL-stained nuclei in the jhs1 mutant (Fig. 7, A and B). These results suggest that more DNA damage occurred in the jhs1 mutant than in the wild type.
We next analyzed whether the jhs1 mutation affected the expression of several marker genes for the DNA damage response. Rad51 is a homolog of the bacterial RecA gene and plays a crucial role in DSB repair as the central component of the DNA recombinational repair system (Doutriaux et al., 1998;Li et al., 2004). BREAST CANCER SUSCEPTIBILITY1 (ATBRCA1) is a homolog of the human breast cancer susceptibility gene and has an important role in DNA damage repair as a key component of the BRCA1-associated genome surveillance complex (Wang et al., 2000;Lafarge and Montané, 2003). GAMMA RESPONSE1 and MRE11 are nucleases involved in DSB end-processing in the repair of crosslinked DNA (Bundock and Hooykaas, 2002;Uanschou et al., 2007;Roth et al., 2012). KU70 plays an important role in the maintenance of the telomeric region and DSB repair by modulating the nonhomologous end-joining pathway (Tamura et al., 2002). MSH2 and MSH6 play important roles in the DNA mismatch repair system (Li et al., 2006;Lario et al., 2011). We examined the expression of these genes by real-time, quantitative reverse transcription-PCR (qRT-PCR). As shown in Figure 7C, the expression levels of RAD51, BRCA1, and GR1 were all increased approximately 2-to 4-fold in jhs1 in comparison to the wild type. The expression of MRE11 and KU70 was a little higher in the jhs1 mutant than in the wild type. However, the expression of MSH2 and MSH6 was not changed in jhs1 mutants.
It has been shown that plant meristematic cells are hypersensitive to DNA damage (Fulcher and Sablowski, 2009). We stained the roots of young seedlings with PI (propidium iodide) to observe dead cells in the RAM of the wild type and jhs1 mutant (Fig. 7, D-G). The area of dead cells in PI-stained roots was significantly larger (50-fold) in the jhs1 mutant than in the wild type (Fig.  7G). We also analyzed the PI-stained roots of wild-type,  jhs1, and Com seedlings that had been treated with 1.0 mM HU or 5 mM zeocin for 36 h. After the treatment, the area of dead cells in the jhs1 mutant was significantly increased (Fig. 7, D-G). However, treatment did not affect the area of dead cells in the wild type (Fig. 7, D-G).

The Frequency of Homologous Recombination Is Increased in jhs1
In response to DNA damage, DNA repair mechanisms such as homologous recombination (HR) or nonhomologous end-joining are upregulated in the cell. To examine the frequency of HR in the jhs1 mutant, HR reporter lines 1415 and 1406, which contain two fragments of the GUS reporter gene in the direct and inverse orientation, respectively (Fig. 8A), were used. Line 1415 can be used to examine intramolecular HR, whereas line 1406 can be used to examine intermolecular HR. If HR events occurred between the homologous regions, i.e. the regions represented by "U" from "GU" and "US" in Figure 8A, DNA recombination between two U-regions of GUS would produce a complete and functional GUS gene. Cells that have undergone this kind of event appear blue upon GUS staining (Supplemental Fig. S8). We first crossed the jhs1 mutant with the 1406 and 1415 lines, respectively. We selected lines of the F 3 generation that were homozygous for both jhs1 and the GUS reporter gene for our analysis. GUS staining resulted in more blue spots in the jhs1 mutant than in the wild type (Fig. 8, B and C). For line 1415, the mutant plants had an average of 4.52 spots per plant, while the wild-type plants had only 1.18 blue spots (Fig. 8, B and C). For line 1406, the mutant plants had 3.42 spots per plant on average, while the wild-type plants had only 0.57 blue spots (Fig. 8, B and C). These results indicate that the jhs1 mutant has an increased level of HR.
After HU treatment, the average number of blue spots was increased to 2.22 in the wild type in line 1415, while the jhs1 mutant had 4.81 blue spots. The average number of blue spots was increased to 0.98 in the wild type in line 1406, while the jhs1 mutant had 4.1 blue spots (Fig. 8, B and C). After treatment with zeocin, the average number of blue spots increased to 7.34 in the wild type in line 1415, while the jhs1 mutant had 13.92 blue spots. The average number of blue spots was increased to 2.77 in the wild type in line 1406, while the jhs1 mutant had 6.26 blue spots (Fig. 8, B and C). Both HU and zeocin treatment significantly increased the HR events in the wild-type plants, while the latter had a much stronger effect. Furthermore, the HR events of both GUS reporters in the HU-treated jhs plants were DNA damage can affect cell cycle progression, at least partially through changes in the regulation of the expression of cell cycle-related genes. We thus compared the expression level of several cell cycle-related genes in the wild type and jhs1 mutant. Specifically, we evaluated the expression of CYCB1:1, which is induced at the G 2 phase and encodes a kinase protein with activity that peaks at the G 2 /M transition (Cools et al., 2010); CYCD4:1, which encodes a protein that is expressed from the G 2 to M phase and binds to and activates CDKB2;1 (Kono et al., 2003); CDKA;1, which is required for cell cycle regulation during pollen, embryo, root, and leaf development in Arabidopsis (Dissmeyer et al., 2009); CYCD3;3, which is expressed in the G 1 phase (Menges et al., 2005); histone H3.1 and histone H4, which are specifically expressed in the S-phase (Menges et al., 2003;Yin et al., 2009); CYCA3;1, whose transcripts are highly accumulated at the G 1 /S phase (Juraniec et al., 2016); CYCB1;2 and MAD2, which are markers for G 2 /M phase (Menges et al., 2005;Juraniec et al., 2016) and CYCA2;1, which is induced during the S phase and peaks at the end of the G 2 phase (Shaul et al., 1996), using qRT-PCR analysis (Fig. 9A). Besides HISTONE H2A PROTEIN9 (HTA9), PLASTID DIVISION2 was also used as a reference gene and similar results were observed (Supplemental Fig. S9). Expression of CYCD3;3, CYCA3;1, and MAD2 were the same in the mutant and the wild type. By contrast, expression of CYCB1:1, CYCB1:2, CDKB1:1, CYCD4:1, CDKA;1, H3.1, histone H4, and CYCA2;1 was greater in the jhs1 mutant than in the wild type. Most strikingly, the expression of CYCB1:1 was upregulated 4.3-fold compared with the wild type (Fig. 9A). The up-regulation of these genes suggests that the DNA damage response is activated in jhs1, which may affect cell cycle progression.  Data presented are averages of four replicates. HTA9 was used as a reference gene and error bars represent SD. Student's t-test; for **, P , 0.01. D, E, F, PI-stained root tips of 5-d-old wild-type, jhs1, and Com seedlings grown on MS medium and transferred to MS medium, without or with 1 mM HU or 5 mM zeocin for 3 d. Bars = 100 mm. G, Quantification of the area of dead cells indicated in (D), (E), and (F). Data represent the average of 10 biological replicates. Error bars represent SD. Student's t-test; for **, P , 0.01.
To analyze the G 2 checkpoint response in the jhs1 mutant, we crossed jhs1 mutant plants with a transgenic line carrying the CYCB1:1 promoter fused with the GUS reporter gene. CYCB1 and GUS are expressed upon entry into the G 2 phase (Doerner et al., 1996). Only a few cells in the wild-type SAM and RAM accumulated GUS protein, as revealed by GUS staining (Fig. 9, C and D). However, more cells were stained in the SAM and RAM of jhs1, which suggests that the DNA damage response was greater in these cells and that more cells underwent G 2 cell-cycle arrest in the jhs1 mutant. In addition, we treated wild-type and jhs1 mutant plants with 1.0 mM HU, which primarily induces G 1 arrest, but also induces G 2 arrest (Culligan et al., 2004); aphidicolin, which inhibits the replicative DNA polymerases d and « and induces G 2 arrest (Culligan et al., 2004); and zeocin, which induces G 2 arrest (Twentyman, 1983). HU treatment can induce the GUS level in the RAM of the wild type, but not in the RAM and SAM of jhs1.
Besides, aphidicolin and zeocin can induce GUS expression in the wild type and jhs1 mutant. The GUS level was higher in jhs1 plants subjected to these two treatments than in the wild-type plants. However, both the wild type and the mutant had similar responses to these two treatments.
To analyze cell cycle progression in jhs1 plants, we determined the endoreduplication level of 10-d-old seedlings by flow cytometry. As shown in Figure 9B and Supplemental Figure S10, the proportion of cells with a 2 C nuclear content was much lower in jhs1 mutant plants than in wild-type plants, and the proportion of cells with a 4 C, 8 C, and 16 C nuclear content was higher in jhs1 mutant plants. This result, in combination with the data shown above, suggests that the G 2 /M phase or late S/G 2 phase of the cell cycle is delayed in the jhs1 mutant.

Genes Important for SAM Development Are Misregulated in the jhs1 Mutant
To decipher the mechanism that controls the phenotype of the jhs1 mutant, we analyzed the expression level of SAM-specific genes, such as WUS, CLV3, and STM using real-time qRT-PCR. As shown in Figure  10A, the expression of STM and WUS was 7-fold and 3-fold greater, respectively, in the jhs1 mutant than in the wild type. However, the expression of CLV3 was slightly reduced.
To investigate the expression pattern of these three SAM-specific genes, we introduced the ProWUS:GUS, ProCLV3:GUS, and ProSTM:GUS reporter constructs into jhs1 plants by crossing, and analyzed GUS activity in the SAMs of the wild-type and jhs1 mutant. In 9-dold wild-type seedlings expressing ProWUS:GUS or ProSTM:GUS, GUS staining was localized to the organizing center of the SAM. However, in the jhs1 mutant of the same age, GUS staining extended to the outer cell layers (Fig. 10, B and D). ProWUS:GUS and ProSTM:GUS expression was also detected in the apex of the inflorescence (Supplemental Fig. S11). In the wild type, both of these constructs were expressed in the central region of the SAM. In the jhs1 mutant, GUS was detected in multiple regions arranged in a linear fashion. This arrangement is due to the multiple meristems present in the shoot apex of fasciated mutants. Similar to our findings in 9-d-old seedlings, these two GUS reporter genes were also expressed at a higher level and in a larger area in the mutant than in the wild type.
The expression level of CLV3, as determined by realtime qRT-PCR, was slightly reduced in the mutant (Fig.  10A). Our GUS staining results indicate that ProCLV3: GUS activity was low and limited to a small region of the SAM in young seedlings of the wild type (Fig. 10C). The expression area was smaller and the activity was lower in the jhs1 mutant (Fig. 10C). In the SAM that gives rise to flowers, ProCLV3:GUS expression was barely detectable (Supplemental Fig. S11). Arrows represent the process of HR. HR between two overlapping sequences produces a functional GUS gene. B, C, The average number of GUS spots per plant in the wild type and jhs1 background with or without 2.0 mM HU and 10 mM zeocin treatment, respectively. Error bars represent SD. "GU" and "US" are two GUS gene fragments.
Some of the fasciated mutants, such as fas1, bru1, mre11, and abo4, also have a defect in epigenetic regulation and a release of silencing of transcriptionally silent information (TSI) (Takeda et al., 2004;Kirik et al., 2006;Yin et al., 2009;Amiard et al., 2010), which may be related to the misregulation of WUS and STM expression in these mutants. However, in other mutants, such as teb, there is no release of TSI (Inagaki et al., 2006). To determine the TSI status in jhs1, we analyzed the expression of TSI repeats by RT-PCR (Supplemental Fig.  S12). We found that there was no difference between the mutant and wild type. This suggests that epigenetic regulation is not altered in jhs1, at least regarding TSI repeats.
To further elucidate the relationship between DNA damage and the fasciation phenotype, we treated 9-dold wild-type plants with HU and zeocin, and analyzed the expression level of WUS, CLV3, and STM using RT-PCR. We found that WUS and STM expression were induced by these DNA damaging agents and that CLV3 expression was not significantly changed (Fig. 10D). GUS staining also indicated that ProWUS:GUS and ProSTM:GUS was expressed at higher levels and over a larger region in plants treated with HU and zeocin (Fig.  10, B and D). Overall, WUS and STM expression was induced to a greater extent by 5 mM zeocin than by 1 mM HU, but it still did not reach the levels observed in the jhs1 mutant. These results further suggest that the DNA  . Cell cycle analysis of the wild-type and jhs1 plants. A, Real-time qRT-PCR analysis of the transcript levels of cell cycle-related genes in the wildtype and jhs1 mutant. The level of these genes was set to 1 in the wild type. Data represent the averages of four replicates. HTA9 was used as a reference gene and error bars represent SD. Student's t-test; for **, P , 0.01. B, Analysis of ploidy distribution in the leaves of 10-d-old wildtype and jhs1 seedlings. Data represent the average of three biological replicates. Error bars represent SD. C, D, GUS staining of the SAM and RAM. Fiveday-old wild-type and jhs1 seedlings transgenically expressing ProCYCB1;1: GUS were transferred to control MS plates or MS plates containing 1.0 mM HU, 12 mg/mL aphidicolin, or 2 mM zeocin, respectively, for 3 d. WT, wild type.
damage observed in the jhs1 mutant promotes the expression of WUS and STM, which may affect SAM development.

DISCUSSION
DNA2 is similar to its homologs in fungi and metazoans, suggesting its function may be conserved. Studies of DNA2 to date have mainly been performed in fungi and mammals (Kang et al., 2010). This protein removes the RNA primer during lagging-strand DNA replication and Okazaki fragment maturation (Budd and Campbell, 1997;Bae et al., 2001;Ayyagari et al., 2003) and is also involved in DNA resection during double-stranded DNA break repair Mimitou and Symington, 2009). In humans, DNA2 is targeted to both nuclei and mitochondria (Zheng et al., 2008;Duxin et al., 2009). The dna2D mutation is lethal in yeast, suggesting that the wild-type protein is essential for cell viability (Budd et al., 2011). In this study, we showed that JHS1 is targeted to the nucleus in Arabidopsis (Supplemental Figs. S6 and S7), as reported for DNA2 in yeast (Choe et al., 2002). The jhs1 mutation occurred at the splicing site of the 11th intron (Fig. 3), which resulted in mis-splicing of the intron and introduced a premature stop codon. This defect might severely affect or even abolish the function of the gene, as reflected by the many severe developmental and physiological defects of the jhs1 mutant shown in this study.
The yeast dna2 mutants are sensitive to DNAdamaging agents, such as MMS and x-ray and g-ray irradiation, suggesting that DNA2 is involved in DNA damage repair (Formosa and Nittis, 1999;Budd and Campbell, 2000). To test whether JHS1 is also involved in DNA damage repair, we evaluated the effect of two DNA-damaging agents, HU and zeocin. HU blocks DNA synthesis and arrests DNA replication by reducing dNTP levels and causing stalled replication forks. Zeocin is a DSB-inducing agent. The jhs1 mutant was more sensitive to all of these agents than was the wild type (Fig. 5). After Zeocin treatment, JHS1 was localized to foci in the nucleus (Fig. 6). Indeed, the TUNEL staining results indicated that the level of DNA damage was high in jhs1 (Fig. 7). Moreover, many dead cells were observed in the SAM of jhs1, which might be the result of severe DNA damage. All these data suggest that the DNA2 homolog in plants also has an important role in DNA damage repair.
To further examine the working mechanism of DNA2 homologs in plants, we analyzed the expression levels of key genes in DNA damage responses (Fig. 7C). The expression of BRCA1, which functions in genome surveillance and DNA damage repair (Wang et al., 2000;Lafarge and Montané, 2003), was increased more than 4-fold in jhs1, suggesting that much DNA damage occurs in this mutant. The GR1 and MRE11 nucleases, which are involved in DSB end processing (Bundock and Hooykaas, 2002;Uanschou et al., 2007;Roth et al., 2012), were also induced in the mutant. The expression level of RAD51, which mediates DNA repair directly by homologous recombination (Li et al., 2004), was increased more than 4-fold in jhs1. The increased expression of genes involved in DNA damage repair and HR likely reduces the amount of damaged DNA in jhs1 cells, but does not repair all damaged DNA.
We crossed two recombination lines, 1415 and 1406, with the jhs1 mutant to examine HR. We found that the frequency of HR events is several times higher in jhs1 than in the wild type (Fig. 8). Previous studies in Arabidopsis suggest that bru1 and abo4-1 plants have a greatly increased frequency of HR events (Takeda et al., 2004;Yin et al., 2009). In jhs1 and abo4, an increased amount of DNA damage was also observed. As stated above, genes involved in DNA recombination and damage repair, such as Rad51 and BRCA2, were also upregulated, probably in response to the DNA damage. The up-regulation of these genes would increase the frequency of HR, which has been shown by the two GUS reporter genes, and help the cell to repair damaged DNA through HR.
Our real-time qRT-PCR and GUS staining results showed that many cell cycle-related genes, such as CYCB1;1, CDKB1;1, CYCD4;1, CDKA;1, H3.1, and CYCA2;1, were upregulated in jhs1 (Fig. 9, A and C). In addition to being a response to DNA damage, this upregulation would affect the progression of the cell cycle in jhs1. The accumulation of DNA damage can activate the check-point response in the cell cycle, which can delay cell cycle progression and allow time to repair DNA damage. Otherwise, apoptosis and cell death may be triggered in response to severe DNA damage. The reduced growth of the SAM and RAM and a significant amount cell death in the RAM of jhs1 all suggest that cell cycle delay and apoptosis were triggered. On the other hand, cells with much DNA damage tend to stop dividing and undergo endoreplication (Kirik et al., 2007). Our flow cytometry analysis indicated that in the mutant, the nuclear DNA content was increased and the proportion of 2 C cells was significantly reduced (Fig. 9B). Moreover, the cells in the SAM, RAM, and leaf epidermis of the jhs1 mutant were significantly larger than those in the wild type (Figs. 1, B and F, and 2D), indicating that endoreplication occurred in these cells. Furthermore, the phenotypes of jhs1 plants were similar to those of other fasciated mutants, such as fas1, fas2, abo4-1, and teb-1 (Endo et al., 2006;Inagaki et al., 2006;Kirik et al., 2006;Yin et al., 2009). It has also been suggested that plant stem cells are hypersensitive to DNA damage (Fulcher and Sablowski, 2009). Thus, all these data support the notion that the DNA damage response may delay cell cycle progression and cause endoreplication in jhs1.
The ataxia telangiectasia and rad3-related (ATR) and ataxia telangiectasia-mutated (ATM) protein kinases are conserved among eukaryotes and involved in DNA damage sensing and the regulation of cell cycle checkpoints. ATR senses the ssDNA of stalled replication forks, whereas ATM is activated by DSBs (Sancar et al., 2004). HR events and the expression of ProCYCB1:GUS, which are triggered in response to DNA damage, were elevated in the jhs1 mutant (Figs. 8 and 9) and were further increased by zeocin treatment, but not by HU treatment. This is probably because jhs1 and HU cause the same type of DNA damage and activate the same signaling pathway, i.e. the ATR pathway. Zeocin treatment causes DSBs and activates the ATM pathway, further increasing the response to DNA damage.
Mutations in FAS1 and FAS2 caused defects in the maintenance of the expression state of WUS in the SAM and SCR in the RAM (Kaya et al., 2001). A mutation in the BRCA1-associated RING domain1 gene caused both a defect in DNA repair and in the maintenance of SAM organization by repressing WUS expression in plants (Reidt et al., 2006;Han et al., 2008). In a previous study, double-knockdown lines of the cyclin-dependent kinases CDKB2;1 and CDKB2;2 showed abnormal SAMs and phyllotaxis, suggesting a close link between cell cycle progression and the regulation of meristem organization in plants (Andersen et al., 2008). In jhs1, the expression of STM and WUS in the SAM and WOX5 in the RAM was misregulated, possibly due to DNA damage and defects in cell cycle progression, and likely contributed to the abnormal development of the SAM and the entire plant in the jhs1 mutant.
It has been shown that g-ray irradiation can induce fasciation in wild-type plants (Abe et al., 2009). In this study, we showed that DNA-damaging agents, such as HU and zeocin, induced WUS and STM expression (Fig. 10). It has been suggested that overexpression of WUS and STM may disorganize the SAM (Han et al., 2008) and result in fasciation. Thus, it is likely that DNA damage causes fasciation by promoting the overexpression of WUS and STM. In contrast to the almost 100% penetrance of the fasciation phenotype of the jhs1 mutant, only 20% to 30% of plants subjected to g-ray irradiation exhibit fasciation (Abe et al., 2009). In this study, the induction of WUS and STM by DNAdamaging agents did not reach the level observed in the jhs mutant. These differences could reflect differences between DNA damage induced by endogenous and exogenous factors.
Although many DNA repair and cell cycle mutants have been reported, only a subset of these has a fasciation phenotype. For the DNA repair mutants, it is possible that some of the mutant alleles were not strong enough to cause such a phenotype. For example, both the mre11-1 and mre11-2 mutants are sensitive to DNAdamaging treatments, but only the stronger allele, mre11-1, has a fasciation phenotype (Bundock and Hooykaas, 2002). On the other hand, if a mutant allele is too strong, it will cause an embryo-lethal phenotype. For example, strong alleles of the mutant of DNA polymerase « showed a sporophytic embryo-defective phenotype (Jenik et al., 2005;Ronceret et al., 2005), while fasciated stems were only found occasionally in the weak alleles (Yin et al., 2009). Similarly, disruption of DNA topoisomerase 1a in combination with reduction of DNA topoisomerase 1b caused seedling lethality, while disruption of DNA topoisomerase 1a alone only caused fasciation Graf et al., 2010). Thus, the phenotype of a mutant may largely depend on the severity and type of DNA damage. A similar situation may exist in cell cycle mutants. For example, only double knockdown of CDKB2;1 and CDKB2;2 caused an abnormal SAM, whereas the singleknockdown plants appeared normal (Andersen et al., 2008). Moreover, it is possible that only the mutants that affect specific steps of the cell cycle cause misexpression of WUS and STM and produce an abnormal SAM.
DNA2 has been shown to function in DNA replication and DNA damage repair in humans and yeast (Budd and Campbell, 2000;Zheng et al., 2008). In this study, we showed that, in plants, DNA2 is important for DNA damage repair during DNA replication in the stem cells of the SAM and RAM, and that mutation of this gene causes DNA damage, a delay in cell cycle progression, endoreplication, and severe developmental defects.

Plant Materials and Growth Conditions
The Arabidopsis plants used in this study were of the Col-0 ecotype unless otherwise stated. Seeds were surface-sterilized with 75% (v/v) ethanol for 5 min, washed five times with sterile water, and sown on Murashige and Skoog (MS) solid medium. The plates were placed at 4°C for 3 d and then transferred to a growth chamber under long-day conditions (16 h of light/8 h of darkness) at 22°C. Plants were grown in a perlite/vermiculite/peat soil (1:1:1, v/v/v) mixture under long-day conditions at 22°C and with a controlled humidity range of 40-60%.

Map-Base Cloning of jing he sheng1
The Arabidopsis jing he sheng 1 (jhs1) mutant plants were crossed with the Ler wild type to obtain F 1 seeds. The F 1 plants were self-crossed to obtain F 2 seeds. Approximately 700 plants from the F 2 population were used for the fine mapping of JHS1. PCR-based molecular markers were developed for the mapping and the primer sequences used are listed in Supplemental Table S1.

Plasmid Construction
For complementation analysis, a 10.259-kb DNA fragment of JHS1, including a 1.323-kb upstream sequence of the promoter and 59 UTR, and a 145-bp downstream region, was amplified from the genomic DNA of the wild type with the primers 59-CCT ACT AGT GCA CGA AGT TTA TCA GCT CCA TC-39 and 59-CCT ACG CGT GCA AGT GAT TCA TGC TCT CCT G-39. The PCR product was digested with MluI and inserted into a binary vector 3302Y3 digested with SmaI and MluI. The constructed plasmid was transformed into Agrobacterium tumefaciens C58C01 and the gene was introduced into the jhs1 mutant by floral dipping (Clough and Bent, 1998).
To construct the ProJHS1:GUS fusion gene, a 1368-bp promoter of JHS1 with the 59 UTR was amplified from the wild-type genomic DNA with the primers 59-GTC AAT AAG GAT AGT AGC TTT GTC AC-39 and 59-CCT GGA TCC TTC CAA TGG CTA ATA CTT GGC AC-39. The PCR product was digested with BamHI and inserted into the 3302GUS binary vector digested with SmaI and BglII to produce the fusion gene.
To construct the Pro-35S:JHS1-YFP fusion gene, a 8.788-kb fragment of JHS1 with the 59 UTR was amplified from wild-type genomic DNA with the primers 59-CGA GAC GTC ATG CCA AGA AAG CC-39 and 59-CCT ACG CGT TGG TTT TAA ATC TCC CGG CAA CAG-39. The PCR product was digested with AatII and MluI and inserted into the 3302Y3 binary vector digested with AatII and MluI to produce the fusion gene.

Microscopy Analysis
To study the cellular arrangement of shoot apical meristems (SAMs) and stems in the wild type, jhs1, and complementation lines, 8-d-old seedlings and the stems of 6-week-old plants were fixed in cold FAA solution (70% ethanol/ formalin/acetic acid, 18:1:1, v/v/v), dehydrated in a series of ethanol concentrations (50, 70, 80, 90, and 100%, v/v), and infiltrated and embedded in LR White resin (Electron Microscopy Sciences, Hatfield, PA). The resin was polymerized at 60°C for 24 h. Sections (1-mm thick) were cut with an Ultra Microtome (Leica Microsystems, Wetzlar, Germany), stained with 1% w/v toluidine blue, and viewed with a model no. CX21 microscope (Olympus America, Melville, NY).
To study the cellular arrangement of roots, 5-d-old seedlings of the wild-type, jhs1, and complemented mutant plants were stained with 5 mM FM4-64, a plasma membrane dye, for 5-10 min. FM4-64 fluorescence was observed using a confocal laser scanning microscope (model no. SP8; Leica Microsystems). Starch granules in root caps were stained with Lugol's solution as previously described in Takahashi et al. (2003) and viewed with a model no. CX21 microscope (Olympus America).
To analyze the root apical meristem, jhs1 mutant plants were crossed with transgenic plants carrying ProWOX5:GFP and ProDR5:GUS, which are specifically expressed in the quiescent center and roots, respectively. F 3 plants homozygous for the reporter gene and jhs1 were selected. Five-day-old seedlings were used to analyze GFP fluorescence, which was observed with a confocal laser scanning microscope (model no. SP8; Leica Microsystems). The number of lateral roots was determined in 9-d-old b-glucuronidase (GUS)-stained seedlings using a dissecting microscope (Optec, Chongqing, China).
To study the subcellular localization of JHS1, the Pro-35S:JHS1-YFP fusion gene was first transiently expressed in Nicotiana benthamiana leaf cells by Agrobacterium infiltration (Voinnet et al., 2003). Pro-35S:JHS1-YFP was also transformed into Arabidopsis to obtain stable transgenic lines. For colocalization studies, the nucleus was stained with a DNA dye, DAPI (49,6-diamidino-2phenylindole). Fluorescence images of YFP and DAPI were obtained with a confocal laser scanning microscope (model no. SP5; Leica Microsystems).
To study the subcellular localization of JHS1 after DNA damage agent treatment, Agrobacterium containing the Pro-35S:JHS1-YFP construct was mixed with mock, 10 mM zeocin, or 2.0 mM HU (hydroxyurea) and then infiltrated into N. benthamiana leaves. Five-day-old transgenic Arabidopsis seedlings harboring YFP-JHS1 were transferred to MS medium with mock, 10 mM zeocin, or 2.0 mM HU for analysis. The localization of JHS1-YFP in N. benthamiana and Arabidopsis was observed 48 h later with a confocal laser scanning microscope (model no. SP8; Leica Microsystems).
To study cell death in the roots of young seedlings, 5-d-old wild-type, jhs1, and complemented plants were transferred to MS solid medium without or with 1 mM HU or 5 mM zeocin for 36 h. Seedlings were stained for 2 min in a 1 mg/mL PI solution (Sigma-Aldrich, St. Louis, MO) and were observed with a confocal laser scanning microscope (model no. SP8; Leica Microsystems).
To detect the DNA damage in jhs1 mutants, a terminal deoxynucleotide transferase-mediated dUTP nick-end labeling (TUNEL) Kit (One Step TUNEL Apoptosis Assay Kit; Beyotime, Shanghai, China) was used. The manufacturer's instructions were modified. Briefly, seedlings were fixed in 4% (w/v) paraformaldehyde in PBS at 20°C for 1 h and washed three times with PBS buffer for 10 min. Then, samples were incubated in 0.1% (v/v) Triton X-100 with 0.1% (w:v) sodium citrate for 30 min (4°C). After three washes in PBS buffer, the roots were soaked in TUNEL reaction mixture and incubated at 37°C for 90 min in darkness. Finally, the roots were examined with a confocal laser scanning microscope (model no. SP8; Leica Microsystems).

DNA Damage Sensitivity Assay
For the images shown in Figure 5, A and B, the seeds were sown directly on MS solid medium containing HU or zeocin and the plants were grown for approximately 3 weeks before photographs were taken. For the images shown in Figure 5, C and D, the seeds were sown directly on MS solid medium containing different concentrations of HU or zeocin and the plants were grown on vertically oriented dishes for 8 d before photographs were taken and root lengths were measured.

Quantitative RT-PCR Analysis
Total RNA was extracted from 10-d-old Arabidopsis seedlings using an RNA Isolation Kit (Applygen, Beijing, China) and an equal amount of total RNA was reverse-transcribed into cDNA using the RevertAid Premium First Strand cDNA Synthesis Kit (Fermentas, Thermo Fisher Scientific, Rochester, NY). For real-time quantitative RT-PCR (qRT-PCR), the Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) and a model no. 7500 Fast RT-PCR system were used (Applied Biosystems). PCR reactions were run in quadruplicate for each sample and the gene expression level was normalized to that of the reference genes, HISTONE H2A PROTEIN9 or PLASTID DIVISION2. All analyses were repeated at least three times. The primer pairs used for qRT-PCR are listed in Supplemental Tables S2 and S3. Transcriptionally silent information analysis was done as before (Yin et al., 2009).

Homologous Recombination Assay
The jhs1 mutant plants were crossed with the homologous recombination (HR) reporter lines 1415 and 1406, which were in the Col background (Lucht et al., 2002). Plants homozygous for the GUS reporter gene in the jhs1 or wildtype background were selected. Fourteen-day-old seedlings of the F 3 generation were subjected to GUS staining and used for HR frequency analysis. For HU and zeocin treatment, seedlings were grown on MS medium containing 1 mM HU or 5 mM zeocin for 16 d. Then, the seedlings were subjected to GUS staining. The number of blue (GUS-positive) spots, each of which represents an HR event, was determined using a dissecting microscope (Optec, Chongqing, China). GUS staining was performed using an S102 GUS Staining Kit (Obiolab, Beijing, China). For each line, 100 individuals were analyzed.

Cell Cycle Assay
The jhs1 mutant was first crossed with Col wild-type plants transformed with a ProCYCB1;1:GUS reporter gene (Doerner et al., 1996). Plants homozygous for the ProCYCB1;1:GUS reporter gene and jhs1 in the F 3 generation were selected. Five-day-old seedlings were used to analyze GUS activity. For the HU, zeocin, and aphidicolin treatment, 5-d-old seedlings were transferred to MS medium containing 1 mM HU, 2 mM zeocin, or 12 mg/mL aphidicolin, respectively, and cultured for 3 d before being subjected to GUS staining. GUS activity in the SAMs and root apical meristems was observed using a dissecting microscope (Optec).

Flow Cytometry Analysis
To study the ploidy level of the jhs1 mutant and the wild type, 9-d-old seedlings were chopped with a blade in 1 mL cold extraction buffer (Dolezel et al., 2007). Crude samples were filtered through a nylon filter (30 mm) to remove cell debris and stained with 25 mL of 0.1 mg.mL 21 DAPI. The nuclei were analyzed using a MoFlo XDP Flow Cytometer (Beckman Coulter, Brea, CA) with CellQuestPro software (BD Biosciences, San Jose, CA). Four independent replicates were performed for each sample.

Expression Pattern Analysis of SAM-Specific Genes
The jhs1 mutant plants were crossed with wild-type transgenic plants carrying ProWUS:GUS, ProCLV3:GUS, or ProSTM:GUS (Angela et al., 2002;Meng et al., 2012;Song et al., 2013), respectively. F 3 plants homozygous for the reporter gene and jhs1 were selected. GUS expression was analyzed in 9-d-old seedlings treated without or with 1 mM HU or 5 mM zeocin and in the inflorescence of 45-d-old plants. After dehydration in 70% (v/v) ethanol, seedlings were cleared with HCG solution (30 mL H 2 O, 80 g chloral hydrate, and 10 mL 100% glycerol) at room temperature overnight and viewed with a model no. CX21 microscope (Olympus America).

Supplemental Data
The following supplemental materials are available.
Supplemental Figure S1. Silique arrangement in the wild-type, jhs, and complemented lines.
Supplemental Figure S2. Expression of ProWOX5:GFP in the roots of transgenic wild-type and jhs1 seedlings.
Supplemental Figure S3. Analysis of lateral root generation in wild-type and jhs1 plants.
Supplemental Figure S4. Map-based cloning and sequence analysis of JHS1.
Supplemental Figure S5. Genotyping and gene expression analysis of the complemented jhs1 plant.
Supplemental Figure S6. Subcellular localization of JHS1 in N. benthamiana leaves.
Supplemental Figure S7. Subcellular localization of JHS1 in Arabidopsis leaves.
Supplemental Figure S9. qRT-PCR analysis of cell cycle-related gene expression.
Supplemental Figure S10. Analysis of ploidy level distributions of wildtype and jhs1 plants.
Supplemental Figure S11. Expression analysis of SAM-specific genes in the floral tissues of 45-d-old wild-type and jhs1 plants.
Supplemental Table S1. Molecular markers and primers used for the fine mapping of JHS1.
Supplemental Table S2. Primers used for the genotyping and RT-PCR analysis in Supplemental Figure S5.
Supplemental Table S3. Primers used for the qRT-PCR analysis.