|Home | About | Journals | Submit | Contact Us | Français|
Cell production is an essential facilitator of fruit growth and development. Cell production during carpel/floral-tube growth, fruit set, and fruit growth, and its regulation by cell cycle genes were investigated in apple (Malus×domestica Borkh.). Cell production was inhibited during late carpel/floral-tube development, resulting in growth arrest before bloom. Fruit set re-activated cell production between 8 d and 11 d after full bloom (DAFB) and triggered fruit growth. The early phase of fruit growth involved rapid cell production followed by exit from cell proliferation at ~24 DAFB. Seventy-one cell cycle genes were identified, and expression of 59 genes was investigated using quantitative RT-PCR. Changes in expression of 19 genes were consistently associated with transitions in cell production during carpel/floral-tube growth, fruit set, and fruit growth. Fourteen genes, including B-type cyclin-dependent kinases (CDKs) and A2-, B1-, and B2-type cyclins, were positively associated with cell production, suggesting that availability of G2/M phase regulators of the cell cycle is limiting for cell proliferation. Enhanced expression of five genes including that of the putative CDK inhibitors, MdKRP4 and MdKRP5, was associated with reduced cell production. Exit from cell proliferation at G0/G1 during fruit growth was facilitated by multiple mechanisms including down-regulation of putative regulators of G1/S and G2/M phase progression and up-regulation of KRP genes. Interestingly, two CDKA genes and several CDK-activating factors were up-regulated during this period, suggesting functions for these genes in mediating exit from cell proliferation at G0/G1. Together, the data indicate that cell cycle genes are important facilitators of cell production during apple fruit development.
Fruit development in apple (Malus×domestica Borkh.) involves multiple phases of growth: carpel/floral-tube growth, fruit set, and early and later stages of fruit growth, similar to that in other fruits (Gillaspy et al., 1993). Transitions in cell production are closely associated with the above phases of growth. Growth of the carpel/floral-tube following bud break is probably mediated by changes in cell production. Pollination and fertilization, which typically occur several days after bloom in apple, result in seed set and generate signals that initiate fruit growth. Stimulation of growth during fruit set is facilitated by re-activation of cell production (Malladi and Hirst, 2010). Fruit set is followed by an early phase of fruit growth which is largely driven by multiple rounds of cell production (Denne, 1960; Malladi and Hirst, 2010). Transition from cell production-mediated to cell expansion-mediated fruit growth occurs at ~3–8 weeks after bloom, and is associated with exit from mitotic cell production (Denne, 1960; Harada et al., 2005; Malladi and Hirst, 2010). Key questions regarding the regulation of cell production during these growth phases remain unanswered. For example: How is cell production in the carpel/floral-tube prior to bloom regulated? What are the factors involved in the re-activation of cell production during fruit set? How is regulation of the rate and duration of cell production during early fruit growth achieved? Additionally, in apple, exit from the mitotic cell cycle during fruit growth occurs at G0/G1 and does not involve the onset of endoreduplication, an alternative form of the cell cycle commonly seen in plants such as Arabidopsis and tomato (Melaragno et al., 1993; Joubes et al., 2000; Harada et al., 2005; Malladi and Hirst, 2010). Mechanisms associated with exit from the mitotic cell cycle at G0/G1 are not well understood. To address these questions, a clear understanding of mechanisms regulating cell production during fruit development is essential.
Cell production is regulated by the sequential progression of a cell through different phases of the cell cycle. The plant mitotic cell cycle consists of G1, S, G2, and M phases, where cell growth, DNA replication, DNA repair, and mitosis occur, respectively (Inze and De Veylder, 2006). Regulation of the plant cell cycle is facilitated by a core group of cell cycle genes such as cyclin-dependent kinases (CDKs), cyclins, CDK inhibitors [Kip-related proteins (KRPs)], CDK subunits, E2F transcription factors, retinoblastoma-related genes (RBRs), WEE kinases, and CDC25 phosphatases. More than 80 genes constitute the core cell cycle machinery in Arabidopsis (Vandepoele et al., 2002; Menges et al., 2005). Similarly, core cell cycle genes have been identified in maize, tomato, and rice (Chevalier, 2007; Guo et al., 2007; Rymen et al., 2007). While A-type CDKs are constitutively expressed during the cell cycle, B-type CDKs are plant specific and are involved in regulating G2/M phase progression (Menges et al., 2005). CDKs facilitate progression through the cell cycle in association with specific cyclin partners. A-type cyclins typically facilitate S and/or G2/M phase progression, and B-type cyclins regulate progression of cells through the G2/M phase. D-type cyclins facilitate phosphorylation of RBRs, thereby releasing their inhibitory binding to E2F transcription factors, subsequently leading to the activation of S phase progression (Inze and De Veylder, 2006). Regulation of CDK activity is facilitated by their association with CDK-activating kinases, such as CDKF, and with CDK subunits. KRPs are CDK inhibitors that bind to and inactivate CDKs, thereby facilitating negative regulation of the cell cycle (De Veylder et al., 2001; Verkest et al., 2005b). An additional level of regulation of CDK activity is facilitated by WEE kinase and a putative CDC25 phosphatase, although the role of CDC25 in regulating CDK activity may be limited (De Schutter et al., 2007; Dissmeyer et al., 2009).
Changes in expression of cell cycle genes are often closely associated with transitions in cell production during organ growth. A set of 131 cell proliferation-related genes, including many cell cycle genes, was identified in Arabidopsis based on their expression during leaf and root growth (Beemster et al., 2005). Additionally, activation of cell production in emerging roots during seed germination is dependent on changes in the expression of A- and D-type cyclins, as well as other cell cycle genes (Masubelele et al., 2005). Expression of several cell cycle genes is associated with cell production during fruit growth in tomato (Joubes et al., 1999, 2000; Baldet et al., 2006; Chevalier, 2007). Additionally, ovary growth and fruit set in tomato involve extensive alterations in the cell cycle transcriptome (Vriezen et al., 2008; Pascual et al., 2009; Wang et al., 2009). It may be hypothesized that transitions in cell production during carpel/floral-tube growth, fruit set, and fruit growth in apple are similarly facilitated by extensive changes in the expression of cell cycle genes.
Cell cycle genes also play important roles in facilitating exit from the mitotic cell cycle (Buttitta and Edgar, 2007; Lee et al., 2009). The role of cell cycle genes in regulating mitotic exit through the onset of endoreduplication is well characterized in plants (Cebolla et al., 1999; Verkest et al., 2005b; Lammens et al., 2008; Boudolf et al., 2009). However, regulation of exit from the mitotic cell cycle at G0/G1 by the cell cycle genes has not been sufficiently explored. Absence of endoreduplication during fruit growth in apple allows investigation of the role of cell cycle genes in mediating cell cycle exit at G0/G1 in the context of organ development.
Several studies have previously addressed genome-wide changes in gene expression during fruit development in apple (Park et al., 2006; Lee et al., 2007; Janssen et al., 2008). However, association of quantitative changes in gene expression with changes in cell production and growth was not achieved in these studies. The primary goals of this study were: to identify key potential regulators (among cell cycle genes) of cell production during different phases of fruit growth and development in apple; to understand how these regulators facilitate transitions in cell production during fruit growth and development; and to explore the role of these genes in regulating exit from the mitotic cell cycle at G0/G1 during fruit growth. Kinematic analysis of cell production and growth, and quantitative analysis of the expression of 59 cell cycle genes were performed to achieve these objectives.
Mature ‘Gala’ apple (Malus×domestica Borkh.) trees growing on M.7 rootstocks at the Georgia Mountain Research and Experiment Station in Blairsville, GA, USA, were used for analysis of carpel/floral-tube growth prior to bloom (n=4). Carpel/floral-tube diameter at the widest part was measured on 10 king flowers per tree. King flowers were collected at 26, 17, 7, and 0 days before full bloom (DBB), stages roughly corresponding to ‘tight cluster’, ‘pink’, ‘open cluster’, and ‘bloom’ stages of apple flower development, respectively. The carpel/floral-tube region was excised from the flowers and used for histology and gene expression analyses. For histology, tissues were fixed in Histochoice (Amresco Inc., Solon, OH, USA). For gene expression analyses, tissues were frozen in liquid N2 and stored at –80 °C.
For analysis of fruit set, mature ‘Gala’ trees on M.7 rootstocks at the above location were used (n=4). At 6 DBB (‘open cluster’ stage of flower development), flower clusters were manually thinned to one king flower per cluster and randomly assigned to ‘Pollinated’ or ‘Unpollinated’ treatments. To prevent pollination and fertilization in the ‘Unpollinated’ treatment, flowers were opened by separating the petals and the styles were clipped with scissors at 6 DBB. For the ‘Pollinated’ treatment, pollen collected from various apple varieties grown at the Horticulture Research Farm in Athens, GA, USA, was applied to the stigmatic surface of open flowers after bloom. Additionally, flowers in the ‘Pollinated’ treatment were allowed to be open-pollinated. The carpel/floral-tube diameter was measured at regular intervals and samples were collected for histology and gene expression analyses at 0, 8, and 11 days after full bloom (DAFB). For histology, tissues were fixed in CRAF III fixative (Berlyn and Miksche, 1976). For gene expression analyses, tissues were frozen as described above.
Mature ‘Gala’ trees on M.7 rootstocks at the above location were used for fruit growth analyses (n=5). Trees were maintained according to commercial production practices except that no chemical thinning agents were applied. Clusters were manually thinned to one king fruit per cluster at 11 DAFB. Fruit diameter was measured at the widest part of the fruit (10 fruits per tree). Fruits from each tree were sampled at different stages during development for histology (four fruits per tree) or for gene expression analyses (4–20 fruits per tree), as described above for the fruit set study.
Carpel/floral-tube and fruit tissues were sectioned using a vibratome. To facilitate efficient sectioning, carpel/floral-tube tissues were embedded in 6% agarose prior to sectioning. Cell number was determined by counting the number of cell layers between the petal vascular trace and the peel. The relative cell production rate (RCPR) was determined from the cell number data as: [Ln(C2)–Ln(C1)]/T2–T1], where C1 and C2 are the cell numbers at time points T1 and T2, respectively. Cell area was determined by counting the number of cells within a defined area, at three locations between the epidermis and the petal vascular trace.
Apple expressed sequence tags (ESTs) coding for cell cycle genes were identified from publicly available EST databases (NCBI and the Apple EST Project: http://titan.biotec.uiuc.edu/apple/). Core cell cycle genes from Arabidopsis (Menges et al., 2005) were initially used for similarity searches in the apple databases, which resulted in the identification of several cell cycle genes. Additionally, queries for previously annotated cell cycle genes (cyclins, CDKs, KRPs, E2F transcription factors, RBR genes, WEE kinase, and CDC25) in the Apple EST Project database yielded genes with homology to known plant cell cycle genes. Sequences obtained by the above methods were further used to mine the NCBI apple unigene database. Sequences within the unigene sets were analysed to identify putative homeologues. Typically, the longest EST within the unigene was used for blastx analysis and a putative function was assigned to the gene based on its similarity to the Arabidopsis homologue (e-value ≤1e-05).
RNA was extracted from carpel/floral-tube and fruit tissues for the fruit set and fruit growth studies as described in Malladi and Hirst (2010). For extraction of RNA from developing carpel/floral-tube tissues, the EZ-RNA extraction kit (Omega Bio-Tek, Norcross, GA, USA) was used following the manufacturer's instructions. Total RNA (1 μg) was treated with DNase (Promega Corporation, Madison, WI, USA) prior to cDNA synthesis. ImPromII reverse transcriptase and oligo(dT) (Promega Corporation, Madison, WI, USA) were used for cDNA synthesis following the manufacturer's instructions (20 μl reaction volume). Samples without the reverse transcriptase were used to test for genomic DNA contamination. The cDNA was diluted 5-fold for the fruit growth study and 8-fold for carpel/floral-tube growth and fruit set studies.
Gene-specific primers for quantitative RT-PCR were designed after multiple alignment of related genes using Clustal W. In the case of closely related genes, primers were designed specifically in non-conserved regions. All primers were validated using a cDNA dilution series. A list of the validated primers used in this study is presented in Supplementary Table S1 available at JXB online. Primer efficiency ranged from 1.85 to 2.0, and ~80% of the primer pairs had an efficiency ≥95%. The specificity of the primer pairs was verified by melting curve analysis at the end of the RT-PCR. All primer pairs used in this study displayed a distinct, single peak in the melting curve analysis. Quantitative RT-PCR analyses were performed on the Stratagene Mx3005P real-time PCR system. Reaction volumes were 14/15 μl using the 2X SYBR-Green Master Mix (Applied Biosystems Inc., Foster City, CA, USA). Reaction parameters were: 95 °C for 10 min; 95 °C for 30 s followed by 60 °C for 1 min (40 cycles); and melting curve analysis. Relative expression of cell cycle genes was calculated according to Pfaffl (2001). Two reference genes were used for normalization of gene expression data for the carpel/floral-tube growth and fruit growth studies: MdACTIN (accession no. EB127077) and MdGAPDH (accession no. EB146750). The geometric mean of expression of the two reference genes (normalization factor) was used for normalization of expression of cell cycle genes. For the fruit set study, only MdGAPDH was used as a reference gene, as expression of MdACTIN decreased considerably at later stages in ‘Unpollinated’ flowers. Representative changes in the expression of the normalization factor during carpel/floral-tube growth and fruit growth, and of MdGAPDH during fruit set are shown in Supplementary Fig. S1 at JXB online. Expression of the normalization factor (or MdGAPDH) was generally altered by <2-fold. Hence, a cut-off value of 2-fold change in expression of cell cycle genes was used for analysing all quantitative gene expression data. Expression of a gene relative to its expression at bloom is presented for the fruit set and fruit growth experiments. For the carpel/floral-tube growth study, expression of a gene relative to its expression at 26 DBB is presented. Three CKLs, five P-type cyclins, and two T-type cyclins were excluded from analysis of gene expression as little is known regarding their function in plants. Additionally, expression of MdCYCD1 and MdCYCD3;3 could not be determined as primer pairs did not result in the unequivocal amplification of a single product.
Statistical analyses were performed using Minitab-15 (Minitab Inc., State College, PA, USA) and SigmaPlot 11 (Systat Software Inc., San Jose, CA, USA). All statistical analyses of gene expression were performed on log2-transformed data. Analysis of variance (ANOVA) was used to identify genes that exhibited significantly different expression during carpel/floral-tube growth (P ≤0.05; n=4). Two-way ANOVA with the general linear model: time, treatment, and time×treatment (three levels of time, 0, 8, and 11 DAFB; and two levels of treatment, ‘Pollinated’ and ‘Unpollinated’) was used for analysing gene expression data for the fruit set study (n=4), and was followed by Tukey's multiple comparison test for mean separation (α=0.05). Carpel/floral-tube diameter, cell layers, and cell area data were compared between the ‘Pollinated’ and ‘Unpollinated’ treatments using a paired t-test (n=4). ANOVA was used to determine if expression of a given cell cycle gene was significantly different across various stages of fruit development (P ≤0.05; n=5). Pearson product moment correlation analysis was used to analyse the association between RCPR and expression of cell cycle genes (P ≤0.05). RCPR data were normalized to the value at bloom prior to correlation analyses. Clustering of cell cycle genes based on expression profiles was performed using Cluster 3.0 (Eisen et al., 1998), and visualized in Java TreeView (http://jtreeview.sourceforge.net/manual.html). Hierarchical clustering using centred, Pearson correlation with complete linkage was adopted for all clustering analyses.
Seventy-one genes with homology to known core cell cycle genes in plants were identified in apple (Table 1). Core cell cycle-related genes from apple were also compared with predicted genes from Vitis vinifera (Supplementary Table S2 at JXB online). Overall, 14 CDKs representing seven classes (A, B, C, D, E, F, and G) were identified. The apple transcriptome contains two CDKA genes with the ‘PSTAIRE’ motif. Two B1- and two B2-type CDKs were identified, indicating the presence of multiple plant-specific, G2/M phase-regulating CDKs in apple. Two members each of C-, E-, and G-type CDKs, a D-type CDK, and a CDK-activating kinase (MdCDKF1;1) were also identified. Additionally, six genes with homology to Arabidopsis CDK-like (CKL) genes are present in the apple transcriptome.
Cyclins constituted the largest class of cell cycle genes in apple (34 genes; A, B, C, D, H, L, P, Q, and T classes). All three A-type cyclins identified here were classified within the A2 subclass. Seven B-type cyclins constituted the G2/M phase-specific group of cyclins. Ten genes representing putative D-type cyclins, of which five belonged to the D3 subclass, were identified. The apple transcriptome also contains six P-type and three T-type cyclins, the functions of which in plants are not yet clear.
Four genes with limited homology to Arabidopsis KRPs and one with similarity to a KRP from Poplar were identified in apple. The apple KRP genes were named in the order of increasing numbers as limited sequence homology exists among KRPs. Two ESTs with homology to the Arabidopsis CDK subunit gene (CKS) shared >90% identity between them and were grouped within the same unigene. However, sequence analysis of ESTs within the unigene allowed their classification into two putative homeologous genes (MdCKS1;1 and MdCKS1;2). Three members belonging to the E2F-type and related transcription factors (MdE2F, MdDEL1, and MdDPb) and two genes with similarity to the retinoblastoma genes (MdRBR1 and MdRBR2) were identified in the apple transcriptome. A single EST with similarity to the WEE kinase of Arabidopsis and a putative CDC25 homologue were also identified.
Carpel/floral-tube diameter increased by 32% between 17 DBB and 7 DBB, but was not significantly altered between 7 DBB and 0 DBB, indicating that growth of the carpel/floral-tube ceased around a week before bloom (Fig. 1). Cell number in the floral-tube increased between 26 DBB and 7 DBB by 56% (Fig. 1). However, cell number was not significantly different between 7 DBB and 0 DBB, indicating cessation of cell production concomitant with the arrest in growth. Between 26 DBB and 0 DBB, cell area increased by only 13% (Fig. 1). Hence, the majority of carpel/floral-tube growth prior to bloom is facilitated by cell production. These data also indicate that arrest in carpel/floral-tube growth prior to bloom is primarily facilitated by a reduction in cell production.
Expression of 59 cell cycle genes was analysed in this study. Among them, 41 genes showed >2-fold change in expression in the carpel/floral-tube tissue between 26 DBB and 0 DBB. Hierarchical cluster analysis allowed grouping of these genes into four major clusters (Fig. 2A). Genes within cluster 1 (21 genes) typically displayed high levels of expression during early stages of carpel/floral-tube growth (26 DBB), a period of higher cell production activity. These genes declined in expression between 26 DBB and 0 DBB by as much as 5.7-fold. A decline in expression was observed over two phases: between 26 DBB and 17 DBB, and between 7 DBB and 0 DBB (Fig. 2B). Genes associated with regulation of G2/M phase progression, such as B-type CDKs, A2-type cyclins, and B1- and B2-type cyclins, were grouped within this cluster. These data suggest that arrest in cell production prior to bloom is associated with decreased availability of G2/M phase regulators. In addition, MdDEL1, MdWEE1, and four D-type cyclins also displayed a similar decline in expression before bloom.
MdKRP1 and MdKRP2 constituted the second cluster. These genes decreased in expression from 26 DBB to 7 DBB and increased in expression between 7 DBB and 0 DBB (Fig. 2B). Cluster 3 included nine genes which increased in expression from 26 DBB to 7 DBB, and either remained constant or slightly declined thereafter (Fig. 2B). Two putative CDK inhibitors, MdKRP4 and MdKRP5, were included within this cluster. Interestingly, expression of MdKRP4 increased by >25-fold between 26 DBB and 7 DBB (Supplementary Fig. S2 at JXB online). Nine genes were included in cluster 4 which typically displayed a slight decrease in expression between 26 DBB and 17 DBB, followed by a sharp increase in expression between 17 DBB and 7 DBB. Expression of these genes either remained constant or decreased slightly between 7 DBB and 0 DBB (Fig. 2B). MdCDKF1;1, MdRBR2, and several cyclins exhibited this pattern of expression.
Growth kinematics during the period of fruit set were studied in ‘Pollinated’ and ‘Unpollinated’ flowers. In ‘Pollinated’ flowers, little change in carpel/floral-tube diameter was observed between 4 DBB and 3 DAFB, but it increased by >60% between 3 DAFB and 10 DAFB (Fig. 3). ‘Unpollinated’ flowers showed little increase in carpel/floral-tube diameter between 4 DBB and 10 DAFB. At 10 DAFB, carpel/floral-tube diameter in ‘Pollinated’ flowers was 60% higher than that in ‘Unpollinated’ flowers (P <0.001). In ‘Pollinated’ flowers, cell number increased only by ~1 cell layer between 6 DBB and bloom, and by ~7 cell layers between bloom and 8 DAFB (Fig. 3). Cell production in ‘Pollinated’ flowers increased sharply between 8 DAFB and 11 DAFB, resulting in a 50% increase in the number of cell layers. In ‘Unpollinated’ flowers, no stimulation of cell production was observed after bloom. At 11 DAFB, cell number in ‘Unpollinated’ flowers was 1.9-fold lower than that in ‘Pollinated’ flowers (P=0.008). By 14 DAFB, the majority of ‘Unpollinated’ flowers abscised and cell number within the remaining few flowers was highly variable. Cell size in ‘Unpollinated’ flowers was slightly higher than that in ‘Pollinated’ flowers prior to and at bloom, but was not different at 3 DAFB and 8 DAFB (Fig. 3). At 11 DAFB, cell size in ‘Pollinated’ flowers was ~26% higher than that in ‘Unpollinated’ flowers (P <0.001). These data demonstrate that pollination and fertilization are essential for re-activating and sustaining cell production, cell expansion, and fruit growth.
Expression of cell cycle genes during fruit set was evaluated at 0, 8, and 11 DAFB, in ‘Pollinated’ and ‘Unpollinated’ flowers. In ‘Pollinated’ flowers, 16 genes were differentially expressed by >2-fold between 0 DAFB and 8 DAFB, and 18 genes were differentially expressed between 0 DAFB and 11 DAFB (Supplementary Tables S3, S4 at JXB online). The majority of these genes decreased in expression during these time periods (10 and 16 genes, respectively). Also, four genes were down-regulated between 8 DAFB and 11 DAFB in ‘Pollinated’ flowers (Supplementary Table S4 at JXB online). In ‘Unpollinated’ flowers, 13, 35, and 31 genes were differentially expressed by >2-fold between 0 DAFB and 8 DAFB, 0 DAFB and 11 DAFB, and 8 DAFB and 11 DAFB, respectively (Supplementary Tables S3, S4). Over 75% of genes differentially expressed in ‘Unpollinated’ flowers between 0 DAFB and 8 DAFB displayed a similar pattern of change in expression in ‘Pollinated’ flowers. However, <10% of genes differentially expressed between 8 DAFB and 11 DAFB in ‘Unpollinated’ flowers displayed a similar pattern in ‘Pollinated’ flowers. These data indicate that the majority of differences in cell cycle gene expression between ‘Pollinated’ and ‘Unpollinated’ flowers emerged between 8 DAFB and 11 DAFB.
Expression of cell cycle genes was further compared between ‘Pollinated’ and ‘Unpollinated’ flowers at each stage. Across the three stages analysed, 38 cell cycle genes were differentially expressed by >2-fold between ‘Pollinated’ and ‘Unpollinated’ flowers (Fig. 4). At 8 DAFB, none of the genes exhibited higher expression in ‘Pollinated’ compared with ‘Unpollinated’ flowers by >2-fold, although several genes were significantly different. By 11 DAFB, 19 genes displayed higher expression in ‘Pollinated’ flowers compared with that in ‘Unpollinated’ flowers (Fig. 4A; all genes except MdCYCD3;4). Changes in expression ranged from 2.4-fold (MdCDKB1;1) to >93-fold (MdCYCB2;2; Fig. 4A). The majority of these genes were putative G2/M phase regulators (13 genes), suggesting that these genes are essential for re-activating and sustaining cell production during fruit set. Additionally, several D-type cyclins also displayed higher expression in ‘Pollinated’ flowers during fruit set.
At 8 DAFB, five genes (MdCDKF1;1, MdCKL3, MdCYCB3;1, MdKRP4, and MdKRP5) exhibited higher expression (>2-fold) in ‘Unpollinated’ flowers in comparison with that in ‘Pollinated’ flowers (Fig. 4B). By 11 DAFB, 18 genes showed higher expression in ‘Unpollinated’ flowers in comparison with that in ‘Pollinated’ flowers, with differences in expression ranging from 2.1-fold (MdCYCC1;2) to >21-fold (MdCYCB3;1; Fig. 4B). Interestingly, expression of several KRP genes was higher by 2- to 4-fold in ‘Unpollinated’ flowers at 11 DAFB. The data suggest that these genes facilitate inhibition of cell production in the absence of pollination.
Carpel/floral-tube diameter changed little between 4 DBB and 3 DAFB. The increase in fruit diameter was initiated between 3 DAFB and 10 DAFB (Fig. 5A). Fruit diameter increased almost exponentially during the early stages (until 32 DAFB) and linearly during later stages of fruit development (Fig. 5A). While cell number within the floral-tube increased by <5 cell layers between 0 DAFB and 8 DAFB, a rapid increase in cell number by >50% was observed between 8 DAFB and 11 DAFB, coincident with initiation of fruit growth (Fig. 5B). Cell number continued to increase rapidly until ~24 DAFB. This was followed by exit from mitotic cell production in the majority of fruit cortex cells (Fig. 5B). An increase in cell number between 8 DAFB and 11 DAFB was associated with a sharp increase in the RCPR (Fig. 5C). RCPR subsequently decreased from ~14 DAFB and reached basal levels by 32 DAFB. Although cell number within the fruit cortex continued to increase during subsequent stages of fruit development, this occurred at a lower rate than during early fruit growth (Fig. 5B, ,C).C). Cell area increased by >3-fold from bloom until 21 DAFB (Fig. 5D). However, the majority of increase in cell size (>24-fold) occurred during later stages of fruit growth (24–123 DAFB), following exit from cell production. The above data indicate that early fruit growth was closely associated with cell production while later stages of fruit growth were primarily driven by cell expansion.
Expression of cell cycle genes was evaluated at 13 stages of fruit development. All core cell cycle genes analysed here showed a significant change in expression by at least 2-fold among different stages of development while 58 of them were altered by >3-fold. Hierarchical clustering was used to identify related groups of cell cycle genes based on their expression profiles during fruit development (Fig. 6A). Representative genes from major clusters are shown in Fig. 6B. Genes within the first cluster (17 genes) exhibited peak expression at bloom or immediately after bloom, and displayed a decrease in expression during the period of cell production. Expression of some of these genes subsequently increased slightly at ~56 DAFB (Fig. 6B, and Supplementary Figs. S3, S4, S5 at JXB online). Two RBRs, E2F, and three D-type cyclins were included in this cluster.
Genes within the second cluster (15 genes) were characterized by enhanced levels of expression during the cell production period (Fig. 6B). Many of these genes exhibited an increase in expression by 2- to 4-fold at ~8–11 DAFB, coinciding with a rapid increase in cell production (Fig. 6B; Supplementary Fig. S3–S5). The decline in expression of these genes started at ~14 DAFB coincident with a decrease in RCPR, and subsequently reached levels lower than that at bloom by 24–32 DAFB, coincident with exit from cell production. These genes continued to display low levels of expression after exit from cell production (after 24–32 DAFB; Fig. 6B). Genes in this cluster included B-type CDKs, A2-type cyclins, and B1- and B2-type cyclins. Additionally, MdDEL1 and MdWEE1 were included within this cluster.
Several CDK inhibitors (MdKRP1, MdKRP2, and MdKRP3) and cyclins (MdCYCD3;2 and MdCYCD3;4) were grouped together in a third cluster (seven genes). These genes increased in expression during the cell production period but also displayed elevated expression at later stages of fruit growth (56–123 DAFB; Fig. 6B, and Supplementary Figs. S4, S5 at JXB online). The fourth cluster contained MdCYCB1;3 and MdCYCP4;1 which exhibited atypical expression patterns. MdCYCB1;3 expression increased by >2-fold during the cell production period and subsequently decreased during exit from cell production, similar to that observed in genes within the second cluster (Fig. 6B). However, its expression was enhanced at later stages of fruit development (79 DAFB and 123 DAFB). MdCYCP4;1 expression was variable during the period of cell production but remained lower at later stages of fruit development (Supplementary Fig. S4).
Genes within the fifth cluster (10 genes) were typically characterized by low expression during the period of cell production followed by elevated expression during and after exit from cell production (after 24 DAFB). Interestingly, MdCDKA1 and MdCDKA2 exhibited this pattern of expression. A CDK subunit, MdCKS1;2, a CDK-activating kinase, MdCDKF1;1, and a CDK-activating phosphatase, MdCDC25, showed patterns of expression similar to that of the A-type CDKs. MdKRP4 and MdKRP5 were also included in this cluster. MdKRP4 expression decreased by 6- to 8-fold during the cell production period and subsequently increased by >14-fold during later stages of fruit development (Supplementary Fig. S5 at JXB online). Similarly, MdKRP5 expression was lowest at the onset of cell production and subsequently increased by 11- to 19-fold during later stages of fruit development (56–123 DAFB; Fig. 6B). Genes within cluster 6 (seven genes) displayed high expression around bloom, a decline in expression during the cell production period, and expression levels similar to that around bloom at later stages of fruit development. This group included several putative cyclins and an E2F dimerization partner (MdDPb).
Analysis of correlation between cell cycle gene expression and RCPR during fruit growth revealed that 13 genes were positively correlated with RCPR (Table 2). All these genes belonged to the second cluster (Fig. 6A) and included B-type CDKs, and A2-, B1-, and B2-type cyclins. Interestingly, expression of MdDEL1 and MdWEE1 was also positively correlated with RCPR. Four genes were negatively correlated with RCPR during apple fruit growth: MdCDKA2, MdCDKF1;1, MdCKL1, and MdCKS1;2 (Table 2). All these genes were grouped together in cluster 5 in the hierarchical cluster analysis (Fig. 6A). These data suggest that the rate of cell production during apple fruit growth is regulated by cell cycle gene expression.
Cell cycle gene expression data from carpel/floral-tube growth, fruit set, and fruit growth analyses were used to identify genes consistently associated with alterations in cell production. Genes positively associated with cell production were identified by: high expression during the initial stages of carpel/floral-tube growth followed by low expression during later stages (cluster 1 of Fig. 2A); higher expression in ‘Pollinated’ flowers during fruit set (Fig. 4A and Supplementary Table S3 at JXB online); and high expression during the cell production phase of fruit growth followed by low expression at later stages (cluster 2 of Fig. 6A and genes positively correlated with RCPR). Fourteen genes were positively associated with cell production across all three data sets (Table 3). Putative G2/M phase-specific regulators of the cell cycle were over-represented within this group (12 genes). Genes negatively associated with changes in cell production were identified by: low expression during early carpel/floral-tube growth followed by peak expression around bloom (clusters 3 and 4 of Fig. 2A); lower expression in ‘Pollinated’ flowers during fruit set (Fig. 4B, and Supplementary Table S4 at JXB online); and low expression during the cell production phase followed by enhanced expression at later stages of fruit growth (clusters 5 and 6 of Fig. 6A, and genes negatively correlated with RCPR). Five genes, MdCDKF1;1, MdCYCC1;2, MdCYCD2;3, MdKRP4, and MdKRP5, were negatively associated with cell production across all data sets (Table 3). Together, the data allowed the identification of key potential regulators of cell production during fruit development.
All cell cycle genes analysed in this study exhibited a >2-fold change in expression during at least one growth phase of carpel/floral-tube or fruit development, suggesting important roles for the cell cycle gene family in facilitating cell production and growth in apple. A previous microarray study indicated that only three cell cycle genes were significantly altered in expression during apple fruit development (Janssen et al., 2008). However, in that study fewer cell cycle genes and only two stages within the cell production period of fruit growth were analysed. In the current study, a quantitative method for analysis of gene expression along with higher temporal resolution, particularly during the cell production phase of early fruit growth, enabled better comprehension of the roles of these genes in regulating fruit growth. In addition to temporal regulation, spatial regulation of gene expression may contribute significantly to fruit development (Lemaire-Chamley et al., 2005; Mounet et al., 2009). Also, regulation of cell production by the cell cycle genes is facilitated through post-transcriptional mechanisms. Analysis of spatial regulation of cell production in fruits and functional analysis of key cell production regulators are required to define further the roles of these genes in apple fruit growth control.
Genes positively associated with cell production were largely constituted by B-type CDKs, and A2-, B1-, and B2-type cyclins, suggesting that cell proliferation during different growth phases of fruit development is limited by the availability of putative G2/M phase regulators of the cell cycle. A majority of genes associated with cell proliferation during leaf and root growth in Arabidopsis were M phase-specific factors (Beemster et al., 2005). G2/M phase-regulating cell cycle genes were also associated with high mitotic activity during seed germination in Arabidopsis and during fruit growth in tomato (Joubes et al., 2000; Masubelele et al., 2005). Additionally, altered expression of these genes in Arabidopsis is often associated with changes in cell production and, in certain cases, with organ growth (Doerner et al., 1996; Lee et al., 2003; Li et al., 2005; Boudolf et al., 2009). Promotion of cell proliferation by G2/M phase-specific regulators may therefore constitute a conserved mechanism for facilitating organ growth in plants.
MdDEL1 and MdWEE1 were also positively associated with cell production in apple. In Arabidopsis, DEL1 is expressed mostly in actively dividing cell types and is a negative regulator of exit from the mitotic cell cycle and onset of endoreduplication (Vlieghe et al., 2005; Lammens et al., 2008). MdDEL1 may similarly prevent exit from cell proliferation as a decline in its expression was consistently associated with a reduction in cell production. In addition, positive correlation of MdDEL1 with RPCR suggests that it may directly promote cell production during apple fruit growth. WEE1 is expressed in tissues undergoing rapid cell proliferation during growth in tomato fruits and Arabidopsis leaves, although its expression peaks during the onset of endoreduplication (Gonzalez et al., 2004, 2007; Beemster et al., 2005). In Arabidopsis, WEE1 is essential for checkpoint control during G2/M to allow repair of DNA damage (De Schutter et al., 2007). Similarly, in apple, MdWEE1 may trigger checkpoint control mechanisms in replicating cells in the event of DNA damage, a role supported by its expression specifically in dividing cells. Alternatively, MdWEE1 may directly promote cell production in apple through as yet undefined mechanisms. Such a function is supported by the positive correlation between MdWEE1 and RCPR during fruit growth.
KRPs function as negative regulators of the mitotic cell cycle (Schnittger et al., 2003; Verkest et al., 2005a, b; Weinl et al., 2005). In apple, MdKRP4 and MdKRP5 were negatively associated with cell production, suggesting that inhibition of CDK activity by these KRP genes limits cell production. In Arabidopsis, CDKF;1 is a positive regulator of cell proliferation as loss of its function reduces cell production and endoreduplication (Takatsuka et al., 2009). Expression of at least one Arabidopsis C-type cyclin is associated with cell proliferation during leaf growth and seed germination (Beemster et al., 2005; Masubelele et al., 2005). D2-type cyclins enhance cell production in response to mitogenic signals (Riou-Khamlichi et al., 2000; Qi and John, 2007). Negative association of the above three genes with cell production in apple is inconsistent with their putative roles in Arabidopsis, and suggests novel functions for these genes in regulating apple carpel/floral-tube and fruit development.
Growth arrest in the carpel/floral-tube tissue from ~7 DBB until at least 3 DAFB was primarily due to reduced cell production. Reduction in cell production at this stage appears to be facilitated by a block in G1 progression as at least some cells retain competence for re-activation of cell production after fruit set. This indicates an arrest in cell production or quiescence prior to bloom. An arrest in ovary growth is also observed prior to anthesis in tomato (Gillaspy et al., 1993; Pascual et al., 2009), indicating that cell production arrest prior to anthesis/bloom is a conserved mechanism to prevent fruit growth in the absence of seed set. Down-regulation of several D3-type cyclins and enhanced expression of MdRBR2 during late carpel/floral-tube growth suggest a block in G1/S phase progression, while down-regulation of all putative G2/M phase regulators during this period suggests a block in G2/M progression, thereby facilitating the arrest in cell production. Similar to data reported here, expression of several cell cycle genes was found to decrease during pre-anthesis in tomato (Wang et al., 2009). Conversely, their expression is enhanced around anthesis in parthenocarpic tomato fruits which do not exhibit an arrest in ovary growth (Mazzucato et al., 1998; Pascual et al., 2009). Arrest in cell production was also associated with enhanced expression of MdKRP4 and MdKRP5. Enhanced KRP activity associated with these genes may result in the inactivation of G1/S phase-specific CDK–cyclin complexes as the cell cycle block occurs at G1 during this period. Together, these data indicate coordinated down-regulation of cell cycle activity during late carpel/floral-tube growth.
Several upstream inhibitory mechanisms may constrain cell production and growth prior to bloom/anthesis. Decreased auxin levels, and/or enhanced activity of negative regulators of auxin signalling, such as IAA9 and ARF7, may constitute one such mechanism (Wang et al., 2005, 2009; de Jong et al., 2009a, b). Alternatively, changes in gibberellic acid (GA) signalling through enhanced DELLA activity may facilitate the reduction in cell production. Prominent roles of GAs in regulating late ovary development and fruit set, and the recent discovery of the role of GAs in regulating cell proliferation support this hypothesis (Marti et al., 2007; Achard et al., 2009; de Jong et al., 2009a). Analyses of changes in the above hormone levels and the expression/activity of genes associated with their signalling are essential to better define these mechanisms.
Pollination and fertilization released the arrest of cell production and initiated fruit growth. Impaired progression of the above events resulted in a sharp decrease and enhancement of expression of genes positively and negatively ssociated with cell production, respectively. These data clearly indicate that pollination/fertilization-derived cues, possibly originating from the seeds, are essential for sustaining changes in cell cycle gene expression and for re-activating cell production during fruit set. Some of the early changes in expression (0–8 DAFB) in ‘Unpollinated’ flowers were similar to those in ‘Pollinated’ flowers, suggesting that these changes are at least initially driven by developmental signals. Alteration in hormone levels/signalling, either in a developmental context or as a result of pollination and fertilization, may facilitate re-activation of cell cycle activity and growth during fruit set. Interestingly, fruit set in tomato is also associated with extensive changes in the transcriptome, particularly that of cell cycle genes and hormone biosynthesis- and signalling-related genes (Vriezen et al., 2008; Wang et al., 2009). Analysis of genome-wide changes in the transcriptome during fruit set in apple may provide insights into the molecular nature of developmental and pollination/fertilization-derived signals that trigger cell production.
Early fruit growth in apple was characterized by a rapid but short period of cell production (8–24 DAFB). Several D-type cyclins displayed enhanced expression during this period, while the expression of RBR genes was reduced, suggesting the activation of mechanisms required for progression through the G1/S phase of the cell cycle. During this period, many putative G2/M phase regulators of the cell cycle exhibited expression patterns correlated with RCPR, suggesting that progression through the G2/M phase may determine the cell cycle duration and thereby the rate of cell proliferation in apple fruit cells. Negative correlation of the expression of MdCDKA2, MdCKL1, and two putative CDK-activating genes with RCPR during this period is surprising as reduction in CDKA activity in Arabidopsis is usually associated with decreased cell production (Dissmeyer et al., 2009). Analysis of CDKA activity is required to understand the significance of this relationship. Coordinated changes in cell cycle gene expression may also influence the duration of the cell production phase during fruit growth as the expression of all genes positively associated with cell production declined, while that of genes within clusters 5 and 6 (Fig. 6A) increased prior to the exit from cell production. Upstream factors controlling final cell number may facilitate such coordination of cell cycle gene expression.
Transition from cell production and subsequent differentiation during apple fruit growth may be facilitated by cell cycle exit at G0 or an arrest in G1, as the majority of fruit cells exhibit a 2C DNA content during later stages of fruit development (Harada et al., 2005; Malladi and Hirst, 2010). Multiple mechanisms appear to facilitate exit from cell proliferation at G0/G1. A dramatic reduction in expression of many putative G2/M phase-specific regulators to levels lower than that during the quiescent stage at bloom was observed during exit from the cell production phase of fruit growth. Down-regulation of mitosis-promoting factors such as the G2/M phase regulators is also required in endoreduplicating plants to facilitate exit from the mitotic cell cycle and to trigger the onset of endoreduplication (Edgar and Orr-Weaver, 2001; Larkins et al., 2001; Lee et al., 2009). Hence, down-regulation of mitosis-promoting activity may constitute an essential component of all mechanisms mediating exit from the mitotic cell cycle.
In addition to reduced mitosis-promoting activity, other mechanisms such as down-regulation of G1/S phase-promoting factors also govern exit from cell proliferation at G0/G1 (Buttitta and Edgar, 2007). E2F transcription factors are essential for entry and progression through the S phase during the mitotic cell cycle (De Veylder et al., 2002). Reduction in MdE2F, and possibly MdDEL1, expression at later stages of fruit growth in apple may prevent transcription of genes required for S phase entry subsequently aiding in exit from cell proliferation at G0/G1. S phase entry and progression are also negatively regulated by RBR genes as loss of RBR function in tobacco leaves enhances cell production in mitotically competent cells and promotes endoreduplication in differentiated cell types (Park et al., 2005). Expression of MdRBR1 and MdRBR2 was slightly enhanced after exit from cell production in apple fruits. Together these data suggest that S phase entry and progression may be down-regulated in apple fruit cells during exit from cell proliferation at G0/G1.
CDK inhibitors are important regulators of cell cycle exit (Buttitta and Edgar, 2007). While moderate overexpression of KRP genes induces precocious entry into endoreduplication, strong overexpression leads to complete exit from cell cycle activity in Arabidopsis (Verkest et al., 2005a, b; Weinl et al., 2005). Expression of MdKRP4 and MdKRP5 was greatly enhanced during exit from the cell proliferation phase of fruit growth, suggesting that these genes facilitate exit from the mitotic cell cycle in apple. However, enhanced expression of these genes during late carpel/floral-tube growth was associated with only a quiescent state. Hence, an increase in expression of these KRP genes may primarily aid in blocking progression of the cell cycle through G1, thereby facilitating exit from cell proliferation in apple fruit cells. Alternatively, a threshold level of KRPs may be essential to trigger exit at G0/G1. High KRP expression at later stages of fruit growth may aid in overcoming such a threshold, thereby triggering exit from cell proliferation.
Interestingly in apple, both A-type CDKs displayed enhanced expression specifically during and following exit from the cell production phase of fruit growth, suggesting an increase in CDKA levels and activity during this period. This is in contrast to mechanisms in endoreduplicating plant systems where mitotic cell cycle exit, and onset and progression of endoreduplication are often associated with a reduction in CDKA activity (Joubes et al., 1999; Verkest et al., 2005b; Inze and De Veylder, 2006; Vlieghe et al., 2007). During mitotic cell cycles, lower CDK activity during the G1 phase is essential to allow assembly of pre-replicative complexes (pre-RCs) required for DNA replication (Edgar and Orr Weaver, 2001; Arias and Walter, 2007; Lee et al., 2009). Even within endoreduplication cycles, oscillation in CDK activity is required to license multiple rounds of DNA replication (Edgar and Orr Weaver, 2001; Larkins et al., 2001). Exit from the mitotic cell cycle at G0/G1 may presumably require the inactivation of mechanisms involved in DNA replication licensing. Enhanced expression/activity of CDKA during later stages of apple fruit growth may reflect a mechanism which prevents pre-RC assembly thereby blocking DNA replication licensing and facilitating exit at G0/G1. Analysis of CDK activity and identification of its partners during exit from cell proliferation at G0/G1 may help in better defining such mechanisms.
Supplementary data are available at JXB online.
Table S1. List of primer sequences used for quantitative RT-PCR analyses.
Table S2. List of genes with similarity to apple cell cycle genes in Vitis vinifera.
Table S3. List of up-regulated genes in ‘Pollinated’ and ‘Unpollinated’ flowers across three stages during fruit set.
Table S4. List of down-regulated genes in ‘Pollinated’ and ‘Unpollinated’ flowers across three stages during fruit set.
Figure S1. Representative expression patterns of reference genes/normalization factors used for normalization of cell cycle gene expression.
Figure S2. Expression of cell cycle genes in the carpel/floral-tube.
Figure S3. Expression of cyclin-dependent kinases (CDKs) and CDK-like genes during fruit development in ‘Gala’ apple.
Figure S4. Expression of cyclins during fruit development in ‘Gala’ apple.
Figure S5. Expression of other cell cycle genes during fruit development in ‘Gala’ apple.
The authors thank the staff of the Georgia Mountain Research and Experiment Station, Blairsville, GA, for their help with field experiments and tree maintenance.