Introduction

In multicellular organisms, cell division plays a crucial role in development. The cell division cycle is well studied in single-cell systems, and molecular mechanisms that ensure the fidelity of DNA replication have been defined. The G2 to M (G2/M) phase transition is an important checkpoint in the regulation of the cell cycle, and is the point at which the fidelity of the replicated DNA is verified. Cells with defects in G2/M phase transition can enter the endoreduplication mode and undergo one or more rounds of DNA replication without intervening mitosis 1, 2. For instance, in transgenic Arabidopsis thaliana plants that have reduced activity of CDKB1;1, a regulator of G2/M transition, stomatal cells become arrested at G2/M and show enhanced endoreduplication 3. When the dominant-negative allele of CDKB1;1 was overexpressed in E2Fa-DPa-overproducing plants, the endoreduplication phenotype of these plants was stronger, but the extra mitotic activity caused by E2Fa-DPa was suppressed 4. In another case misexpression of ICK/KRP1, a possible inhibitor of CDKA;1, in Arabidopsis trichomes could block G1/S and G2/M transitions in a concentration-dependent manner 5. A high concentration of ICK/KRP1 can inhibit the G1/S transition and suppress endoreduplication 6, whereas low-level misexpression blocks the G2/M transition and induces endoreduplication 5.

In plants, shoot apical meristem (SAM) cells play critical roles in many aspects of development and the cell cycle of SAM cells is highly regulated 7. Two major signaling pathways that are involved in establishing and maintaining SAM cells have been characterized. These are the WUSCHEL (WUS)-CLAVATA (CLV) pathway and the SHOOT MERISTEMLESS (STM) pathway 8, 9, 10, 11, 12. WUS encodes a homeobox transcription factor that is necessary for the expression of CLV3. The protein encoded by CLV3 binds to its receptor CLV1 to trigger downstream signals, and this in turn represses the expression of WUS. Thus, a feedback loop is set between WUS and CLV3 13, 14. In the STM pathway, STM is a KNOTTED1-like homeobox (KNOX) gene 12, and its known downstream genes include AS1 and AS2 15, 16, which encode an Myb transcription factor and a leucine zipper transcription factor, respectively. Genetic studies have demonstrated that STM keeps SAM cells in an undifferentiated state by repressing the expression of AS1 and AS2. These two genes can promote the formation of primordia by suppressing KNOX genes, including KNAT1, KNAT2 and KNAT6 15, 17, 18.

Despite extensive knowledge about the genes that regulate the cell cycle and those that are involved in SAM functions, the link between these two categories of genes remains unclear 19. Myb-related CDC5 proteins are functionally and structurally conserved across the eukaryotic kingdoms 20. The CDC5 gene was first isolated from Schizosaccharomyces pombe and has been demonstrated to function at the G2/M phase transition 21. In yeast, mammals and Arabidopsis, the CDC5 proteins can bind to specific DNA sequences in vitro; therefore, CDC5 might serve as a transcription factor to regulate mitotic entry 21, 22, 23, 24, 25. The CDC5 protein has also been proved to be a crucial component of the spliceosome, and it is essential for mRNA splicing in fission yeast and humans 26, 27, 28, 29, 30. Further genetic studies in yeast show that the loss of CDC5/Cef1p function results in cell cycle defects by influencing the splicing of TUB1 mRNA 31, 32. The CDC5 homolog in Arabidopsis, AtCDC5, has been reported to be able to rescue the phenotype of the temperature-sensitive cdc5 mutant in S. pombe, and was found to bind DNA in vitro 24. Recent studies demonstrate that AtCDC5 is localized in the nucleus and that the C-terminus of this protein has transactivation activity in yeast. Moreover, silencing of this gene in Arabidopsis results in accelerated cell death 33.

In this study, we report the functional characterization of AtCDC5 in Arabidopsis to investigate its roles in plant development. We found that AtCDC5 is mainly expressed in proliferating tissues of all phases, and the AtCDC5 loss-of-function mutant is lethal at the zygote stage. In AtCDC5-RNAi plants, the G2/M phase transition is severely affected, SAM function is impaired, and the STM and WUS-CLV signaling pathways are suppressed. The results suggest that AtCDC5 is essential for the G2/M phase transition in Arabidopsis, and is critical for SAM function.

Material and Methods

Generating AtCDC5-RNAi transgenic plants

To create an RNAi construct, a 365-bp fragment (637-1001) of AtCDC5 cDNA was amplified by RT-PCR 33. It was then cloned, in the sense and antisense orientations, into the pQVF binary vector. These two fragments were separated by an 800-bp fragment of the β-glucuronidase (GUS) gene (Figure 1A) 34. The positive clones were sequence confirmed. This plasmid, designated pQVF-AtCDC5i, was introduced into Agrobacterium tumefaciens (GV3101). Arabidopsis plants (ecotype Columbia-0) were transformed using the floral-dip method 35. Transformed plants were screened against kanamycin and grown in the greenhouse under long-day conditions (16 h of light with 8 h of dark) after being transferred into soil. The transgenic plants were verified by PCR.

Figure 1
figure 1

Characterization of the T-DNA insertion mutant atcdc5-1. (A) Schematic representation of the T-DNA insertion sites in atcdc5-1. Exons and introns are represented by black boxes and lines, respectively. Gray boxes represent T-DNA inserts. (B) Siliques of wild type (WT) and atcdc5-1 (+/−) plants. Aborted seeds are indicated by arrows. (C) Wild type and atcdc5-1 (−/−) embryos at 36 hap. The wild-type embryo is at the 8-cell stage; however, the mutant embryo (indicated by an arrow) is arrested as a zygote in a seed containing a single endosperm nucleus (indicated by an asterisk). (D) Wild type and atcdc5-1 (−/−) embryos at 48 hap. The wild-type embryo is at the globular stage. The mutant embryo (indicated by an arrow) remains at the zygote stage but is enlarged. There is only a single endosperm nucleus at this time (indicated by an asterisk). (E) Wild type and atcdc5-1 (−/−) embryos at 72 hap. Note that the atcdc5 (−/−) embryo (indicated by an arrow) is still at the zygote stage, and no endosperm nucleus can be seen. Bars: (B) = 500 μm; (C-E) = 25 μm.

Full size image

Northern blot and qRT-PCR

To carry out northern blot analysis of STM, CLV1 or AtCDC5 expression, a part of the coding region was amplified by RT-PCR and used as a template for the synthesis of P35-labeled random probes. RNA extraction and northern blot analysis were carried out as previously reported 36, 37. cDNA synthesis and real-time quantitative RT-PCR were performed on a DNA Engine Opticon system (MJ Research, Cambridge, MA, USA) according to the protocols that have previously been described 38. The primers used for the northern blot and real-time qRT-PCR analyses are:

AtCDC5 5′-aag gcg gag gaa gaa gca-3′, 5′-gaa gct tcc atg gct atg gc-3′;

CDKB1;1 5′-ctt cat atc ttc agg ttg c-3′, 5′-tca gaa ctg aga ctt gtc aa-3′;

cyclinB1 5′-atg atg act tct cgt tcg att gtt-3′, 5′-atc ctc cac aag aag cgt ggg att-3′;

KRP1 5′-atg gtg aga aaa tat aga aaa g-3′, 5′-ccc gct aca aca aca atc taa cg-3′;

STM 5′-atg gag agt ggt tcc aac agc-3′, 5′-tca aag cat ggt gga gga gat gt-3′;

CLV1 5′-aag gac acg gtc tcc acg act g-3′, 5′-acc ggt gta gct gtt gta gt-3′;

WUS 5′-gtc tat gga tct atg gaa c-3′, 5′-cta gtt cag acg tag ctc aag aga-3′;

AS1 5′-gaa ccg tga cgc caa gtc ttg t-3′, 5′-ctc gtc aat agg ctc aac tct c-3′;

KNAT1 5′-acc gag aat tgc ttc cga tct g-3′, 5′-gcc gtg ccg ccg taa ttc tat c-3′;

Ubiquitin-10 5′-tcc gga tca gca gag gct ta-3′, 5′-tca gaa ctc tcc acc tca ag-3′.

Electron microscopy

For scanning electron microscopy, shoot apices were fixed in FAA (50% ethanol, 5% acetic acid and 3.7% formaldehyde), vacuumed and dehydrated in a series of ethanol. The 100% ethanol was then replaced by isoamyl acetate. The samples were dried with CO2 at critical point, sputter coated with gold and observed with an S-800 scanning electron microscope (Hitachi, Tokyo) at an accelerating voltage of 30 kV.

Flow cytometric analysis

Nuclei from seedlings and rosette leaves were prepared as already described 39. To eliminate the RNA, the nucleus was digested with RNase (with a final concentration of 4 μg/ml) at 37 °C for 20 min, and then 2 μl PI (5 mg/ml) was added. The samples were subjected to flow cytometric analyses with a FACSCalibur flow cytometer (Becton, Dickinson and Company, USA), equipped with an argon ion laser at 488 nm, and the fluorescence was examined at 580 nm. Histograms were produced using the CellQuest software (Becton, Dickinson and Company, USA). The number of nuclei present at each peak of the histogram (2C, 4C, 8C and 16C) was analyzed by measuring the peak area.

Histological and whole-mount tissue studies

In order to monitor embryogenesis, developing seeds from wild type and atcdc5-1 (+/−) plants were removed from the siliques and cleared as has been previously described 40 and then examined using Nomarski optics with an optical microscope (Olympus BX-51, Tokyo, Japan).

To carry out histological analysis, seedlings and inflorescences were fixed in FAA for 20 h at 4 °C, dehydrated in an ethanol/xylene series and embedded in paraffin. Embedded tissues were sectioned to an 8-μm thickness, placed onto glass slides, dried and stained with safranine and fast green.

GUS staining 41 was carried out as previously described.

In situ hybridization

In situ hybridizations were carried out as previously described 37. The templates used for synthesizing antisense transcripts were all cloned into pBS and sequence was conformed. The primers used for amplifying the templates are AtCDC5 5′-gcg ctt acg agc tct caa gt-3′, 5′-cca tct ctt ctt gcc tct tc-3′; cyclinB1 5′-atg atg act tct cgt tcg att gtt-3′; 5′-atc ctc cac aag aag cgt ggg att-3′; STM 5′-gcc gca gcc gct gcc gca tct a-3′; 5′-tca aag cat ggt gga gga gat gt-3′; Histone4 5′-atg tcg gg tcg tgg aaa ggg agg-3′; 5′-ctt cct cct agc gtg ctc ggt gt-3′.

Results

AtCDC5 loss-of-function mutant is lethal at the zygote stage

To investigate the function of AtCDC5, we obtained an AtCDC5 loss-of-function mutant, GABI-KAT line 278B09 (hereafter designated as atcdc5-1), in which two T-DNA insertions were found in the fourth exon of AtCDC5 (Figure 1A). Segregation was analyzed using the T3 population. Out of 191 T3 plants, 122 were heterozygous and no homozygous mutants were found. This suggests that atcdc5-1 is a recessive, lethal mutant of AtCDC5. Phenotypically, heterozygous atcdc5-1 (+/−) mutants were largely identical to wild-type plants. However, when the siliques were examined about one-quarter of the seeds had been aborted (Figure 1B), indicating that seed development for the homozygous atcdc5-1 (−/−) mutants terminates prematurely. We examined embryonic development in the ovules of atcdc5-1 (+/−) plants. At 36 h after pollination (hap), wild-type embryos were at the 8-cell stage, but atcdc5-1 (−/−) embryos were at the zygote stage (Figure 1C). At 48 hap, when the wild-type embryos were at the globular stage, the atcdc5-1 (−/−) embryos, still at the zygote stage, were enlarged abnormally (Figure 1D). At 72 hap, the wild-type embryos were at the heart stage; however, the seeds containing atcdc5-1 (−/−) embryos were aborted, and remains of a large cell, probably the defective zygote, were found inside (Figure 1E). Besides the defective zygote, the endosperm nucleus failed to divide in these atcdc5-1 (−/−) seeds (Figure 1C and 1D, indicated by asterisks). These results suggest that AtCDC5 is essential for the cell division of the embryo and endosperm, and that AtCDC5 loss-of-function results in lethality at the zygote stage.

AtCDC5 is predominantly expressed in proliferating cells

It has been reported that AtCDC5 encodes a putative cell cycle regulator that is mainly expressed in proliferating tissues 24. To further clarify the connection between the AtCDC5 expression pattern and development, in situ hybridization was conducted using wild-type embryos and flowers. At the one-cell stage, AtCDC5 transcripts could be detected in both the embryonic cells and the endosperm nucleus (Figure 2A, indicated by arrowhead). At the 4-cell stage and globular stage, AtCDC5 was expressed in all the embryonic and endosperm cells (Figure 2B and 2C). At the heart stage and torpedo stage, AtCDC5 was expressed throughout the embryo, but it was more abundant in the cotyledon and radicle primordia (Figure 2D and 2E). The transcript was detected throughout the entire mature embryo and it was particularly enriched in some cotyledon cells (Figure 2F, indicated by arrows). In the inflorescences, the AtCDC5 transcript was more abundant in inflorescence meristems (m) and developing flowers (Figure 2G and 2H). In flowers at different developmental stages, AtCDC5 was especially enriched in those tissues containing rapidly proliferating cells, such as the primordia of sepals (Figure 2G, stage 5), young stamens and gynoeciums (Figure 2G and 2H, stages 6-8), and ovule primordia (Figure 2H, stage 10, indicated by asterisks). No signal was detected in the negative control, which was hybridized with the sense probe (Figure 2I). Therefore, these results suggest that AtCDC5 is mainly expressed and enriched in proliferating cells.

Figure 2
figure 2

Expression pattern of AtCDC5. Labeled antisense transcripts were hybridized with sections of wild-type tissue. (A) One-cell embryo. (B) 8-cell embryo. SC: seed coat. (C) Globular-stage embryo. (D) Heart-stage embryo. AtCDC5 is expressed throughout, but particularly abundantly in the cotyledon primordia (arrow). (E) Torpedo-stage embryo. AtCDC5 is highly expressed in cotyledon primordia and radicle primordia cells. (F) Mature embryo. Expression of AtCDC5 is enriched in some cotyledon cells (arrows). (G) Inflorescence. AtCDC5 is highly expressed in inflorescence meristem (m) and developing flowers. The development stages of the flowers are indicated by numbers. S: Sepal primordia; St: young stamen; Gy: gynoecium. (H) Flower at stage 7 and stage 10. Ovule primordia are indicated by asterisks. (I) Sense control. No signal can be detected. Bars = 100 μm.

Full size image

AtCDC5-RNAi plants display pleiotropic phenotypes

Since homozygous loss-of-function plants are embryonic lethal, we adopted the RNAi strategy to investigate AtCDC5 function in post-embryonic development (Figure 3A). We generated 35 transgenic lines, of which three lines appeared to be wild type, whereas 32 partially or completely lacked the SAM. Out of the 32 RNAi mutants, only eight were fertile. Genetic segregation analysis determined that four out of the eight lines contained a single copy of the RNAi construct (Supplementary information, Table S1).

Figure 3
figure 3

Phenotype of AtCDC5-RNAi plants. (A) A schematic representation of the AtCDC5-RNAi construct. Green boxes indicate the AtCDC5 fragments. (B) Phenotypes of 12-day-old WT and AtCDC5-RNAi (+/+) plants. Bar = 4 mm. (C) Kinematic analysis of the root length of WT and AtCDC5-RNAi (+/+) plants. Bar = 8 mm. (D) Phenotypes of 35-day-old WT and AtCDC5-RNAi (+/−)-I plants. (E) Phenotypes of 60-day-old WT and AtCDC5-RNAi (+/−)-I plants. (F) Phenotypes of 40-day-old WT and AtCDC5-RNAi (+/−)-II plants. (G) Basal part of the AtCDC5-RNAi (+/−)-II plant shown in (F). Lesions are indicated by arrows.

Full size image

Phenotypic and co-segregation analysis showed that homozygous AtCDC5-RNAi plants (designated AtCDC5-RNAi (+/+) hereafter) had severe phenotypes (Supplementary information, Table S1). Death occurred at the cotyledon stage before mature true leaves emerged (Figure 3B), and their root growth was severely inhibited (Figure 3C). However, heterozygous AtCDC5-RNAi plants (designated AtCDC5-RNAi (+/−) hereafter) could generate true leaves, although at a slower pace, and showed pleiotropic phenotypes. Among the T3 generation of AtCDC5-RNAi (+/−) plants, about 13% died with only four or fewer true leaves (Figure 3D). Approximately 39% showed retarded, slow growth with abnormal rosette phyllotaxy and lacked primary inflorescence shoots (Figure 3E). The remaining 48% did not have obvious growth defects, but had a dwarf phenotype (Figure 3F), and paler lesions were found on young rosette leaves (Figure 3G) and siliques (Figure 3F). We divided AtCDC5-RNAi (+/) plants into two classes. The class I plants could not generate primary inflorescences, while the class II plants could. Supplementary information, Table S2 summarizes the phenotypes of the T3 generation of AtCDC5-RNAi (+/−) plants.

Phenotypic severity in AtCDC5-RNAi plants corresponds to AtCDC5 expression level

Northern blot analyses were conducted to investigate whether AtCDC5 was silenced in the AtCDC5-RNAi plants. The results showed that expression of AtCDC5 was reduced in four independent AtCDC5-RNAi (+/−)-II lines (Figure 4A), and in AtCDC5-RNAi (+/+), AtCDC5-RNAi (+/−)-I and AtCDC5-RNAi (+/−)-II plants (Figure 4B). The expression levels of AtCDC5 in the most severe AtCDC5-RNAi (+/+) plants, in the moderately severe AtCDC5-RNAi (+/−)-I plants and in the least severe AtCDC5-RNAi (+/−)-II plants were about 20%, 35% and 50% of the wild-type level, in corresponding order. This shows a positive correlation between the reduction of AtCDC5 expression level and the severity of the phenotypes (Figure 4B). The expression of RAN1, which had 18 identical nucleotides to the RNAi fragment of AtCDC5, was not affected (data not shown), thus excluding the possibility of nonspecific RNA silencing. Therefore, the variation of the phenotypes observed among the AtCDC5-RNAi lines likely reflects dosage-dependent defects that are due to loss of AtCDC5 function.

Figure 4
figure 4

Dosage effect of AtCDC5 on the phenotype of AtCDC5-RNAi plants. (A) Northern blot analysis of AtCDC5 in WT and AtCDC5-RNAi plants. Four independent lines with single insertions are shown. (B) RNA gel blot analysis of AtCDC5 in WT, AtCDC5-RNAi (+/+), AtCDC5-RNAi (+/−)-I and AtCDC5-RNAi (+/−)-II plants.

Full size image

G2/M phase transition is severely affected in AtCDC5-RNAi plants

In order to investigate whether AtCDC5 plays a similar role in the G2/M phase transition to the CDC5 proteins in yeast and animals 21, 22, we needed to clarify whether cells from AtCDC5-RNAi plants display defects during the G2/M phase transition. Such defects can include being arrested at the G2 phase 21, 23, 42 or entering the endocycle 4, 7. To address this problem, we used flow cytometry to measure the DNA content of cells in AtCDC5-RNAi (+/+) plants. At 7 days after sowing (DAS), nuclei from both wild type and AtCDC5-RNAi (+/+) seedlings displayed four main peaks that correspond to ploidy forms of 2C, 4C, 8C and 16C. The ploidy level distribution of these plants was similar (Figure 5A). At 8 DAS, AtCDC5-RNAi (+/+) plants had an increased proportion of 8C cells as well as other polyploidy cells, but a decreased proportion of 4C cells and 2C cells (Figure 5B), suggesting that the mitotic cell cycle was suppressed. At 10 DAS, the proportion of 16C cells was largely increased in the AtCDC5-RNAi (+/+) plants and 32C cells emerged, whereas, in contrast, the proportion of 4C and 2C cells was further decreased (Figure 5C). This suggests that more cells escaped from the mitotic cell cycle and endoreduplication was triggered. At 11 DAS, the ploidy level distribution of the AtCDC5-RNAi (+/+) seedlings was similar to that of the 10 DAS plants (Figure 5D), suggesting that the ploidy level reached a steady state. Given that endoreduplication is always accompanied by cell enlargement 7, we examined the pavement epidermal cells of the fifth leaves in 35-day-old wild type and AtCDC5-RNAi (+/−)-I plants to see whether endoreduplication is indeed increased in AtCDC5-RNAi plants. The results showed that the cells of the AtCDC5-RNAi (+/−)-I plants are larger than those of the wide type (Figure 5E). Therefore, these results suggest that the G2/M phase transition is blocked in AtCDC5-RNAi plants, and that the defective cells escape mitosis and enter the endocycle.

Figure 5
figure 5

Cell cycle analysis of AtCDC5-RNAi plants. (A to D) Kinetic analysis of ploidy level distribution in wild type (left) and AtCDC5-RNAi (+/+) (right) plants as measured by flow cytometry: (A) 7 DAS, (B) 8 DAS, (C) 10 DAS and (D) 11 DAS. Histograms represent average data from three independent measurements. (E) SEM image of adaxial epidermal cells of the fifth leaves at 5 weeks after sowing. One cell is outlined in red in each plant. Bars = 100 μm. The histogram represents the average size of 100 independent cells. (F) Expression analysis of KRP1, cyclinB1 and CDKB1;1. The level of KRP1, cyclinB1 and CDKB1;1 was normalized to that of Ubiqutin-10 mRNA and the relative mRNA level was calculated in reference to the WT level. Experiments were repeated three times and the standard deviation (SD) is shown.

Full size image

To further clarify the regulation mechanism for the G2/M phase transition, we examined the expression of CDKB1;1 and KRP1, two genes that are important during the G2/M phase transition 3, 4. We also examined the expression of cyclinB1 in AtCDC5-RNAi (+/+) plants. Quantitative RT-PCR analysis showed that CDKB1;1 expression was severely reduced in the AtCDC5-RNAi (+/+) plants, while both cyclinB1 and KRP1 were slightly reduced (Figure 5F). This result suggests that the downregulation of CDKB1;1 might be the main cause for the defects in the G2/M phase transition.

Severe AtCDC5-RNAi plants have impaired SAM function

To better understand the phenotype at the shoot apex, we examined SAMs of the AtCDC5-RNAi (+/+) and AtCDC5-RNAi (+/−)-I plants by scanning electron microscope and histological analysis. At 10 DAS, wild-type seedlings possessed a hemisphere-shaped SAM (Figure 6A), but in AtCDC5-RNAi (+/+) seedlings a cavity was observed where a SAM should be present (Figure 6C). Histological analysis found no SAM structure (Figure 6D compared with WT in Figure 6B). In the 35-day-old wild-type plants that bolted (data not shown), the primary SAMs were transformed into inflorescence meristems and flower primordia were observed (Figure 6E). However, the AtCDC5-RNAi (+/−)-I plants that were at the same developmental stage did not bolt, and no inflorescence meristems and flower primordia were observed (Figure 6G). Instead, the primary SAMs of the AtCDC5-RNAi (+/−)-I plants were tightly covered by leaf primordia, and clustered leaf primordia were observed between rosette leaves (Figure 6G). This suggests that the development of the primary SAMs was disturbed. Histological analysis confirmed that while inflorescence meristems had been produced in the wild-type plants as early as 30 DAS (Figure 6F), no inflorescence meristems were found in the AtCDC5-RNAi (+/−)-I plants even at 35 DAS (Figure 6H). This result suggests that the function rather than the establishment of SAMs was impaired. In addition, AtCDC5-RNAi (+/−)-I plants developed some axillary meristems at this stage (Figure 6H), implying that the impaired SAM development led to the loss of apical dominance. At approximately 60 DAS, these axillary meristems, but not the primary SAMs, developed into inflorescences (Figure 3E). These results, together with the dosage effect of AtCDC5 on the phenotype of AtCDC5-RNAi plants, suggest that AtCDC5 is essential for SAM function.

Figure 6
figure 6

SEM and histological analysis of AtCDC5-RNAi (+/+) and (+/−)-I shoot apices. (A) Shoot apex of a 10-day-old WT seedling. (B) Section of the shoot apex of a 10-day-old WT seedling. (C) Shoot apex of a 10-day-old AtCDC5-RNAi (+/+) seedling. (D) Section of the shoot apex of a 10-day-old AtCDC5-RNAi (+/+) plant. The SAM is absent. (E) Inflorescence meristem (m) of a 35-day-old WT plant. Floral meristems (fl) are initiated in a spiral arrangement. (F) Section of the shoot apex of a 30-day-old WT plant. (G) Shoot apices of a 35-day-old AtCDC5-RNAi (+/−)-I plant. The primary meristem leaf primordium (pl) and axillary meristem leaf primordium (al) are indicated by arrows. (H) Section of the shoot apex of a 35-day-old AtCDC5-RNAi (+/−)-I plant. The primary meristem (pm) and axillary meristem (am) are indicated by arrows. Bars: (A-D) = 100 μm; (E-H) = 50 μm.

Full size image

Both the WUS-CLV and STM pathways are impaired in AtCDC5-RNAi plants

Since no SAM was found in AtCDC5-RNAi (+/+) plants (Figure 6C) and SAMs with defective function were found in AtCDC5-RNAi (+/−)-I plants, it is possible that these plants have defects in the establishment and/or maintenance of SAM structures. To test this, we examined whether the three main component genes of the WUS-CLV signaling pathway, CLV3, WUS and CLV1, were affected by AtCDC5 deficiency. We crossed AtCDC5-RNAi plants with CLV3::GUS plants to examine the expression of CLV3 in the F2 generation 41. The results showed that CLV3::GUS was constitutively expressed in the wild type (Figure 7A, 7C and 7E). However, in AtCDC5-RNAi (+/+) plants CLV3 expression was detected at 6 DAS (Figure 7B), but was significantly downregulated at 8 DAS (Figure 7D), and was undetectable at 10 DAS (Figure 7F), suggesting that the SAM was established at an early stage but was not maintained. CLV3 was also suppressed in AtCDC5-RNAi (+/−)-I plants (Figure 7G and 7H). Northern blot analysis showed that the expression of CLV1 was undetectable in AtCDC5-RNAi (+/+) plants but not affected in AtCDC5-RNAi (+/−)-I plants (Figure 7I), suggesting that there is no SAM in AtCDC5-RNAi (+/+) plants. Because we could not detect the expression of WUS in either wild type or AtCDC5-RNAi plants by northern blot analysis, we used real-time qRT-PCR. Our results showed that WUS was significantly reduced in AtCDC5-RNAi (+/−)-I plants (Figure 7I) and could not be detected in AtCDC5-RNAi (+/+) plants (data not shown). Together, these data suggest that the WUS-CLV signaling pathway becomes suppressed in AtCDC5-RNAi plants, and defective AtCDC5 function interferes with the maintenance of SAM, rather than its establishment.

Figure 7
figure 7

Expression of meristem development genes in AtCDC5-RNAi lines. (A, C and E) GUS-stained CLV3::GUS seedlings at 6 DAS, 8 DAS and 10 DAS, respectively. (B, D and F) GUS-stained CLV3::GUS/AtCDC5-RNAi (+/+) plants at 6 DAS, 8 DAS and 10 DAS, respectively. (G and H) CLV3::GUS and CLV3::GUS/AtCDC5-RNAi (+/−)-I plants at 25 DAS. (I) Northern blot analysis showing CLV1 expression. CLV1 expression was significantly reduced in AtCDC5-RNAi (+/+) plants but not in AtCDC5-RNAi (+/−)-I plants. Real-time qRT-PCR showed that WUS was also suppressed in AtCDC5-RNAi (+/−)-I plants. (J) Expression analysis of STM, AS1 and KNAT1 in WT and AtCDC5-RNAi (+/−)-I plants. Equal loading of the gel was confirmed by EtBr staining (bottom panel). The level of WUS, AS1 and KNAT1 mRNA was normalized to that of Ubiqutin-10 mRNA and the relative mRNA level was calculated in reference to the WT level. Experiments were repeated three times and SD is shown. Bars = 50 μm.

Full size image

Since the STM signaling pathway is independent of the WUS-CLV pathway, we also examined the three main components of this pathway, i.e., STM, AS1 and KNAT1. Northern blot result showed that the expression of STM was suppressed in AtCDC5-RNAi (+/−)-I plants (Figure 7J). Quantitative RT-PCR analysis showed that the expression of AS1 was upregulated, while KNAT1 was downregulated (Figure 7J), suggesting that the STM signaling pathway was indeed affected in AtCDC5-RNAi plants.

STM expression is reduced in the shoot apex region of AtCDC5-RNAi plants

We wanted to clarify whether the SAM defects in AtCDC5-RNAi plants result from an impairment in the G2/M phase transition and a subsequent decrease in cell division, or from downregulation of both WUS and STM by AtCDC5. To do this, we examined the expression of two cell cycle marker genes, cyclinB1 and Histone4, and that of STM in the shoot apex region of AtCDC5-RNAi plants.

AtCDC5 was highly expressed in the SAM and leaf primordia in the wild-type plants, but was restricted to some cells in the shoot apex of the AtCDC5-RNAi plants at 7.7 DAS. The expression was further attenuated at 8.1 and 8.7 DAS (Figure 8A). These data suggest that AtCDC5 expression is indeed silenced in these plants.

Figure 8
figure 8

Expression patterns of AtCDC5, cyclinB1, Histone4 and STM in AtCDC5-RNAi (+/+) plants. The expression patterns of AtCDC5, cyclinB1, Histone4 and STM in the AtCDC5-RNAi (+/+) plants were determined by in situ hybridization. AtCDC5-RNAi (+/+) plants at 7.7 DAS (7 DAS + 16.8 h), 8.1 DAS (8 DAS + 2.4 h), 8.7 DAS (8 DAS + 16.8 h) and wild-type plants at 8 DAS were used in this assay. Wild-type plants that were hybridized with antisense probes and sense probes were used as positive and negative controls, respectively. (A) AtCDC5. V: vascular tissue. (B) cyclinB1. Arrowheads indicate cells expressing cyclineB1. (C) Histone4. Arrowheads indicate cells expressing Histone4. (D) STM. Arrowheads indicate cells expressing STM. Bars = 50 μm.

Full size image

In wild-type plants, cyclinB1, specifically expressed in the G2 or M phase 43, 44, was not detected in the top two cell layers of the SAM (Figure 8B). However, cyclinB1 was found to be abundant in the apical cells of the 7.7 DAS AtCDC5-RNAi plants (Figure 8B, 7.7 DAS, arrowhead). At 8.1 DAS, fewer cells were found to express cyclinB1; these were mainly restricted to the peripheral zone of the shoot apex (Figure 8B, 8.1 DAS, arrowheads). At 8.7 DAS, cyclinB1 expression could still be detected in some cells in the peripheral zone of the shoot apex (Figure 8B, 8.7 DAS, arrowhead). Histone4, which is S-phase-specific 45, was highly expressed in some leaf primordia cells in the wild-type plants, but could not be detected in the central zone of the SAM (Figure 8C). In the AtCDC5-RNAi plants, Histone4 could be detected in some cells localized in the central zone of the shoot apex at 7.7 DAS, and in more central zone cells at 8.1 DAS (Figure 8C). At 8.7 DAS, Histone4 expression had declined, and was mainly detected in the topside cell layer and peripheral zone of the shoot apex (Figure 8C, 8.7 DAS, arrowheads). The expression patterns of these two cell cycle marker genes in AtCDC5-RNAi plants suggest that the cell cycle is disturbed in the shoot apex region of AtCDC5-RNAi (+/+) plants.

STM transcripts were detected in shoot apex cells of the 7.7-day-old and 8.1-day-old AtCDC5-RNAi (+/+) plants. However, STM expression was reduced and restricted to one or two cell layers compared with the wild-type plants (Figure 8D), suggesting that the ability of STM to maintain SAM cells in an undifferentiated state was diminished 12. At 8.7 DAS, the STM transcript could not be detected (Figure 8D, 8.7 DAS), implying that STM regulation is completely lost in the shoot apex cells of AtCDC5-RNAi (+/+) plants.

STM overexpression can partially rescue defective AtCDC5 function

To test the hypothesis mentioned above, we overexpressed STM in AtCDC5-RNAi plants and obtained five transformants – two were sterile and the other three had inflorescences that terminated prematurely (data not shown). The T3 progeny of the three fertile lines were selected for an AtCDC5-RNAi (+/+) background and named 35S::STM/AtCDC5-RNAi (+/+)-1, 2 and 5. In these three lines, AtCDC5 expression was silenced and STM was overexpressed (Figure 9A).

Figure 9
figure 9

Partial rescue of the AtCDC5-RNAi (+/+) phenotype by 35S::STM. (A) RNA gel blot analysis of AtCDC5 and STM in 35S::STM/AtCDC5-RNAi (+/+) plants. Three independent lines were used in this assay. (B) Phenotypes of 13-day-old WT and 35S::STM/AtCDC5-RNAi (+/+) seedlings. (C) Phenotypes of 50-day-old WT and 35S::STM/AtCDC5-RNAi (+/+) plants. (D) Inflorescences of WT and 35S::STM/AtCDC5-RNAi (+/+) plants. The 35S::STM/AtCDC5-RNAi (+/+) inflorescences terminated prematurely and their flowers lacked carpels. Bars: (B and D) 0.5 cm.

Full size image

Unlike the AtCDC5-RNAi (+/+) plants, the 35S::STM/AtCDC5-RNAi (+/+) seedlings did not show any lesions. Leaves were formed, although at a slower rate, and at 13 DAS, when the wild-type counterpart had four leaves, the 35S::STM/AtCDC5-RNAi (+/+) plant generated one true leaf (Figure 9B). Throughout development, rosette leaves and shoots randomly emerged (Figure 9C), and no lesion-like pale spots were observed. Inflorescences of 35S::STM/AtCDC5-RNAi (+/+) plants showed a “stop and go” growth pattern, i.e. inflorescences prematurely terminated and, at the same time, new inflorescences emerged from the axils of cauline leaves (Figure 9C), resembling wus mutants 11. Additionally, most 35S::STM/AtCDC5-RNAi (+/+) flowers were missing carpels and two or more stamens (Figure 9D). Although similar to wus plants, their phenotype was weaker, and they were partially fertile. These results support our hypothesis that the SAM defects of the RNAi lines are partially mediated through the downregulation of STM.

These data, together with the expression pattern of CLV3, suggest that the defects of the AtCDC5-RNAi plants in maintaining the SAM might be caused by the reduced expression of STM or WUS, which in turn results in the complete consumption of the SAM cells.

Discussion

To characterize the function of AtCDC5, we examined plants that had reduced activity of this gene. Homozygotic lines of a T-DNA insertion mutant atcdc5-1 were not obtained, and analysis of seeds developing inside the heterozygote individuals showed that individuals that are homozygous for the insertion halted at the zygote stage. Therefore, the homozygous AtCDC5 loss-of-function mutant is embryonic lethal (Figure 1). The RNAi lines with partially reduced AtCDC5 activity exhibited pleiotropic defects, indicating continuous and widespread functions of the gene (Figure 3). This is supported by the expression pattern of AtCDC5, which was expressed throughout embryogenesis and in most post-embryonic tissues, such as inflorescence meristems, SAMs, and flower and leaf primordia (Figures 2 and 8). Publicly available AtCDC5 expression data from various microarray analyses, which we surveyed using the Genevestigator database 46, agree with our in situ hybridization results.

In this study, we found that downregulation of AtCDC5 reduced, rather than increased, the proportion of G2/M phase cells, and increased that of the polyploidy cells (Figure 5). This phenotype is different from that seen in cultured cells of yeast and mammals, in which CDC5 defects result in cell cycle arrest at the G2 phase 21, 23. This difference is possibly associated with the mechanism of organ size control that exists in multicellular organisms 47. For instance, in Drosophila, inactivation of the Cdc2 kinase blocked the G2/M phase transition and increased endoreduplication. Therefore, because the cell size increases, the final size and shape of the pupal wing was not affected 48. In Arabidopsis, it was also reported that cells often compensate for a shortage in number of cells by increasing cell size during leaf development 49. Our data showed that the AtCDC5-RNAi plants had both reduced expression of CDKB1;1 and increased cell size (Figure 5), supporting this hypothesis. Thus, we conclude that, instead of being arrested at the G2 phase, endoreduplication was used by the AtCDC5-defective cells as an efficient way to maintain the final size of organs 7.

The most striking developmental phenotype of the AtCDC5-RNAi plants was the loss of SAM (Figure 6). It is well known that mutants with affected cell cycles also exhibit meristem termination phenotypes (for instance, double mutants of ribonucleotide reductase (tso2-1 rnr2a-1) that have defects in S-phase progression 45). Cell cycle progression is vital for the maintenance of SAM and failure in cell cycle progression will generally destroy the balance in SAM cell numbers and result in SAM termination. In this study, we found that AtCDC5-RNAi plants had comparable CLV3 expression with wild-type plants at 6 DAS, but had much lower expression at 8 DAS, and even no expression at 10 DAS (Figure 7A-F). This suggests that stem cells exist before 6 DAS but they are lost after 8 DAS. Since both the G2/M phase transition and SAM maintenance pathways were suppressed in AtCDC5-RNAi plants, it is difficult to decide which one corresponds to the loss of stem cells. However, the fact that cyclinB1 and Histone4 were expressed in the shoot apical cells of the AtCDC5-RNAi plants at 8.1 DAS and 8.7 DAS suggests that these cells can still enter mitosis at this time. This observation argues against the possibility that the loss of stem cells was caused by the failure of the cell cycle. From the evidence that the reduction of STM was tightly correlated with the loss of stem cells (Figure 8D), together with the fact that the STM and WUS-CLV pathways are both required for the maintenance of the SAM 50, we propose that these two pathways are suppressed in the AtCDC5-RNAi plants. The suppression consequently causes the loss of stem cells, and then the absence of stem cells eventually contributes to the loss of SAM in the AtCDC5-RNAi plants.

Taking into consideration that AtCDC5 was expressed at the beginning of embryogenesis, and that WUS and STM were downregulated in AtCDC5-RNAi plants, a straightforward interpretation is that AtCDC5 is involved in regulating the expression of these two genes. This hypothesis is further supported by our data, which show that overexpression of STM can partially rescue the phenotype caused by defective AtCDC5 function (Figure 9). However, many questions remain to be answered. For example, are WUS and STM the direct targets of AtCDC5? Does AtCDC5 function as a transcription factor or as an mRNA splicing factor? All these issues need to be elucidated in the future.

In conclusion, our data in this study demonstrate that AtCDC5 is a cell cycle regulator that is important for the G2/M phase transition. It also plays an essential role in SAM maintenance, possibly through regulating WUS and STM expression.

(Supplementary information is linked to the online version of the paper on the Cell Research website.)