|Home | About | Journals | Submit | Contact Us | Français|
Aside from those on Arabidopsis, very few studies have focused on spatial expression of cyclin-dependent kinases (CDKs) in root apical meristems (RAMs), and, indeed, none has been undertaken for open meristems. The extent of interfacing between cell cycle genes and plant growth regulators is also an increasingly important issue in plant cell cycle studies. Here spatial expression/localization of an A-type and B-type CDK, auxin and cytokinins are reported in relation to the hitherto unexplored anatomy of RAMs of Cucurbita maxima.
Median longitudinal sections were cut from 1-cm-long primary root tips of C. maxima. Full-length A-type CDKs and a B-type CDK were cloned from C. maxima using degenerate primers, probes of which were localized on sections of RAMs using in situ hybridization. Isopentenyladenine (iPA), trans-zeatin (t-Z) and indole-3yl-acetic acid (IAA) were identified on sections by immunolocalization.
The C. cucurbita RAM conformed to an open transverse (OT) meristem typified by an absence of a clear boundary between the eumeristem and root cap columella, but with a distinctive longitudinally thickened epidermis. Cucma;CDKA;1 expression was detected strongly in the longitudinally thickened epidermis, a tissue with mitotic competence that contributes cells radially to the root cap of OT meristems. Cucma;CDKB2 was expressed mainly in proliferative regions of the RAM and in lateral root primordia. iPA and t-Z were mainly distributed in differentiated cells whilst IAA was distributed more uniformly in all tissues of the RAM.
Cucma;CDKA;1 was expressed most strongly in cells that have proliferative competence whereas Cucma;CDKB2 was confined mainly to mitotic cells. iPA and t-Z marked differentiated cells in the RAM, consistent with the known effect of cytokinins in promoting differentiation in root systems. iPA/t-Z were distributed in a converse pattern to Cucma;CDKB2 expression whereas IAA was detected in most cells in the RAM regardless of their proliferative potential.
Hormone systems coordinate plant growth and development through a complex network of signal interactions (Davies, 1995). Significant progress in understanding the molecular basis of hormonal action in plants has been achieved by using the model plant Arabidopsis thaliana (Estelle and Klee, 1994; Kieber, 2002; Schmülling, 2002; Benkovà et al., 2003). However, how interactions between the different hormone classes are coordinated with cell division during plant development remains a major challenge (Friml, 2003; Nordstrom et al., 2004; Ljung et al., 2005; Normanly et al., 2005; Dello Ioio et al., 2008a, b; Moubayadin et al., 2009). That said, there are numerous examples that link these classes of plant growth regulators to growth, development and cell division.
There is firm evidence that indole-3yl-acetic acid (IAA) promotes, but that cytokinins repress, the size of the root apical meristem (Werner et al., 2003). Auxin is also required for root elongation (Katekar and Geissler, 1980) and represents an important promotive signal for lateral root morphogenesis (Blakely et al., 1982; Reed et al., 1998; Bhalerao et al., 2002). In contrast, cytokinins inhibit both primary root growth and lateral root formation (Hewelt et al., 1994; Werner et al., 2001, 2003). For example, transgenic tobacco plants over-expressing the cytokinin biosynthesis gene ISOPENTENYL TRANSFERASE (IPT) exhibited reduced root growth (Hewelt et al., 1994) and exogenous application of the cytokinin 6-benzylaminopurine suppressed the formation of lateral roots in Lactuca sativa seedlings (Zhang and Hasenstein, 1999). Conversely, in Arabidopsis, over-expression of CYTOKININ OXIDASE led to a reduction in cytokinin levels in roots and an increase in the number of lateral roots (Werner et al., 2001). Hence, the trend is that root growth will increase when auxin levels are raised and when cytokinins are reduced.
Auxin and cytokinins also regulate the cell cycle. Cyclin-dependent kinases (CDKs) and cyclins are recipients of cytokinin signal transduction chains. For example, D-type cyclins are upregulated in response to an exogenous cytokinin signal (Riou-Khamlichi et al., 1999) and this initiates events that drive cells into S-phase. Also, dephosphorylation of CDKs at G2/M occurs in response to a cytokinin signal (Zhang et al., 1996, 2005; Orchard et al., 2005) followed by entry of cells into mitosis. Likewise, in auxin-treated roots, two positive G1/S regulators, D-type cyclins and E2F transcription factors, were upregulated whereas two negative regulators, KRP1 and 2, were downregulated (Himanen et al., 2002; Vanneste et al., 2007). Moreover, the latter authors suggested that an A-type cyclin (Arath;CYCA2;4) was a primary recipient of an auxin signal. This work, principally in Arabidopsis, has revealed a positive role for auxins on the cell cycle, root growth and development, but that whereas cytokinins are positive regulators of the G1/S and G2/M transition they repress root growth and development (see above). This apparent anomaly has yet to be explained but might be partly due to a differential sensitivity of different regulatory events to different concentrations/moieties of cytokinin. Aloni et al. (2006) proposed that basipetal polar movement of auxin in the pericycle is halted by a localized increase in ethylene biosynthesis. Auxin accumulates and lateral roots are initiated. In this model, cytokinins are highest in the root cap and meristem, they antagonize IAA and they repress lateral root formation in the near vicinity of the root apical meristem (RAM). Disruption in this balance leads to either more or fewer lateral roots, consistent with the above-mentioned examples of stimulation and repression of lateral roots by auxins and cytokinins, respectively.
Quite how genes that drive cells into divisions are interfaced with plant hormones in such a model is not known. During the past 10 years or so, 152 CDKs have been cloned from 41 species; there are six classes (A–G) together with an additional class of CDK-like kinases (Dudits et al., 2007). Arguably, most is known about CDKA and CDKB types. In Arabidopsis there is only one A-type, CDKA1;1, that can complement the temperature-sensitive cdc2– mutant of fission yeast (Ferreira et al., 1991). The CDKB class comprises CDKB1;1, 1:2 and CDKB 2;1 and 2;2. The first of these plant-specific B-types to be cloned was originally named cdc2bAt (Segers et al., 1996). Unlike CDKA;1, B class CDKs cannot complement the cdc2– mutant of fission yeast and these genes diverge from the A class in that the conserved PSTAIRE domain found in Arath;CDKA;1 is altered to either PPTALRE or PPTLRE (Dudits et al., 2007). Whereas CDKA;1 is required for both the G1/S and the G2/M transitions, the B-types function only at G2/M (Ferreira et al., 1991; Mironov et al., 1999). In Arabidopis, CDKA;1 expression is linked to meristematic competence in the pericycle (Hemerly et al., 1993) whilst a B-type CDK (CDKB1;1) is most strongly expressed in pericycle cells that initiate lateral root primordia (Beeckman et al., 2001; Vanneste et al., 2007). The RAM of Arabidopsis conforms to a three-tiered closed meristem in which there is a distinct boundary between the epidermis and root cap. As noted by Heimsch and Seago (2008), closed meristems are less common than open meristems (which lack that distinct boundary) even within eudicots. Hence, knowledge of auxin and cytokinin distribution and cell cycle gene activity in Arabidopsis may not be the basis for a universal model applicable to meristems in eudicots or, indeed, angiosperms as a whole. For example, polar movement and spatial distribution of auxin might be quite different in open than in closed meristems.
Cucurbita maxima, a popular cultivated species in southern Italy, as a member of the curcurbitales, should have a root apical meristem that conforms structurally to an open-transverse type (Heimsch and Seago, 2008). That is, it should lack a clear boundary between the pole of the meristem and the columella but should possess a very distinctive epidemis more basipetally, separating the cap in this region from the cortex. As such, a cucurbitales-type RAM should be structurally disparate from that of Arabidopsis and could serve as an interesting example to examine the extent to which CDKA and CDKB spatial expression varies in open compared with closed meristems. This was one aim of the work presented here. A further aim was to determine the distribution of auxin and cytokinins in RAMs of C. maxima in relation to the spatial expression of A-type and B-type CDKs in open meristems.
Neither in Arabidopsis nor in crop plants is there a complete understanding of cell cycle genes in relation to plant hormones. Indeed, except for Arabidopsis there are very few plant species in which cell cycle genes have been cloned; known variants of CDKB encompass only 18 angiosperms (Dudits et al., 2007). Data are presented here for the first time on the anatomy of root apical meristems of C. maxima and this structural analysis is used as a template to interpret spatial expression of an A- and B-type CDK. The extent to which the spatial expression of these genes is correlated with the distribution of endogenous auxin and cytokinins in the RAM was also examined. A distinctive feature of the data was spatial expression of Cucma;CDKA in a domain comprising the lateral root cap–epidermal initials, a highly distinctive domain in RAMs of the curcurbitales.
Seeds of Cucurbita maxima Wall. were surface-sterilized with ethanol and then sodium hypochlorite. After washing, seeds were germinated in Petri dishes on filter paper moistened with either distilled water or different incubation solutions in a growth chamber (PIARDI, Brescia, Italy) in the dark, at 23 °C and 65 % humidity.
When primary roots were 1 cm long (n >10), root tips were excised and fixed in 3 % (w/v) paraformaldehyde and 0·5 % (v/v) glutaraldehyde in PBS buffer (135 mm NaCl, 2·7 mm KCl, 1·5 mm KH2PO4, 8 mm K2HPO4, pH 7·3) for 3 h at 4 °C. After washing in the same buffer, samples were dehydrated and embedded in Tecknovitt 8100 resin. Semi-thin sections (3 µm) were obtained using an Ultracut microtome (Leica RM2155) and stained with 0·5 % (w/v) safranin O for histological observations.
Total RNA was isolated from roots of C. maxima. Tissues, frozen with liquid nitrogen (100 mg), were processed with the RNeasy Plant Mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions followed by DNase treatment (Qiagen). cDNA was synthesized with SuperScript III Reverse Trascriptase and oligo dT(22), according to the manufacturer's instructions (Invitrogen, Milan, Italy).
CDKA was PCR-amplified using degenerate primers: FW (5′-TCTCTSTTGAARGARATGCARCA-3′) and BW (5′-GGAGTCTCCAGGRAABTRDGG-3′). These were designed based on the class A CDK conserved amino acid stretches SLLKEMQH and PLFPGDS, respectively, in Arabidopsis and other eudicot species for which sequences were available (see Fig. 1). The 467-bp fragment was cloned, sequenced and confirmed to share high identity with higher plant CDKAs. The C. maxima CDKA full-length cDNA was obtained in two steps: (1) the 3′ region (PCR product of 1101 bp) was cloned by 3′ RACE (Invitrogen), using the anchor primers provided by the manufacturer and the FW primer; and (2) the 5′ region was isolated by using the degenerate primer FWdeg (5′-CCGGATGGCTTCCATTGCTCYCAA-30), based on conserved amino acids between Arabidopsis and tobacco, and the BW1 primer 50-TGGCGTTGTAATCCGTGTAA-30.
For cloning of C. maxima CDKB, the cDNA from root RNA was PCR-amplified using degenerate primers: FW (5′-GGAGAAGGAACNTAYGGVAA-3′)and BW (5′-AGAATCACCAGGAAANARWGC-3′), based on class B CDK conserved amino acid stretches and ALFPGDS from Arabidopsis and other eudicots for which this sequence was available. The PCR reaction yielded a 660-bp product, which was homologous to B-type CDKs. The 3′ end of the putative CDKB was amplified by RACE (Invitrogen). PCR was performed with the FW0 (5′-CAAATCCACAGGTTCAC CCTA-3′) and kit anchor primer. The 5′ CDKB end was amplified by using the degenerate primer FWdeg (5′-CCGGATGGCTTCCATTGCTCYCAA-3′), based on conserved amino acids between Arabidopsis and tobacco, and primer ZCDKBBW (5′-CAAAATCTCCAGGAAAGAGC-3′) based on the C. maxima CDKB sequence already obtained.
C. maxima CDKA and CDKB sequences were aligned with other CDKA and CDKB proteins using ClustalW (http://www.ebi.ac.uk/clustalw) and optimized by visual inspection (PILEUP program). Full-length C. maxima clones containing the open reading frame were obtained using primers ZCDKBFWall (5′-GCATTCGAGAAGCTGGAGAA-3′); ZCDKBBWall (5′-TTCGCTGAAATACGCTTTGA-3′) for CDKB and CDKAfFW1 (5′-GATGGAGCAGTATGAGAAAGTTGAGAAG-3′) and CDKAfR1 (5′-CGCAATTAA AACCCATAA AATGCTACAC-3′) for CDKA and sequenced in both directions by the Genelab ENEA service (Rome, Italy) or the in-house sequencing facility at the Cardiff School of Biosciences.
Phylogenetic trees were constructed in MegaBlast3 (based on the minimum evolution criterion), using bootstrap values performed on 1000 replicates and the 50 % value was accepted as indicative of a well-supported branch.
For in situ expression analysis, the specific probe for Cucma;CDKA spanned the 1134–1411 stretch (Fig. 1A) and was cloned after PCR amplification using the primers FW2/BW2 (5′-TCTTCTAAAACAATTAATTT-30). The Cucma;CDKB-specific probe spanned the 1134–1411stretch (Fig. 1A) and was cloned after PCR amplification using the primers FW2/BW2 (50-TCTTCTAAAACAATTAATTT-3′). Both probes were linearized using SpeI and NcoI endonucleases and the resulting products were used as a template to synthesize digoxigenin-labelled RNA sense and anti-sense probes, by T7 and SP6 polymerase-driven in vitro transcription, respectively. The DIG-RNA labelling Kit protocol (Roche Diagnostic GmbH, Mannheim, Germany) was used according to the manufacturer's instructions. Excised roots were fixed, dehydrated and embedded in paraplast (Sigma-Aldrich, Milan, Italy). Samples were cut with an RM 2125 RT microtome (Leica, Milan, Italy) into 8-μm sections that were transferred to charged slides and hybridized at 50 °C to digoxigenin-labelled antisense RNA probes as described by Cañas et al. (1994). For immunological detection, the slides were incubated as described by Chiappetta et al. (2001). Transcript accumulation was visualized as a violet/brown staining. After stopping the reaction with TE (50 mm Tris-HCl, pH 8·0, 5 mm EDTA), sections were mounted with 50 % glycerol in TE and analysed by bright-field illumination (Leica DMRB). Images were taken with a digital camera (Leica DFC 320).
For cytokinin immunolocalization, 1-cm-long primary roots (n >15) were excised and fixed in a 3 % (v/v) paraformaldehyde and 0·5 % (v/v) glutaraldehyde mixture in PBS buffer (135 mm NaCl, 2·7 mm KCl, 1·5 mm KH2PO4, 8 mm K2HPO4, pH 7·3) for 3 h at 4 °C. The cytokinin pre-embedding immunolocalization was performed as described by Dewitte et al. (1999). Auxin immunolocalization was performed as described in Moctezuma (1999) and Thomas et al. (2002) as modified by Chiappetta et al. (2009).
Excised roots (1 cm long; n = 15) were immediately pre-fixed in freshly prepared 4 % aqueous 1-ethyl-3-(dimethyl-aminopropyl)-carbodiimide hydrochloride (EDAC; Sigma) for 1 h at 4 °C. Subsequently, roots were post-fixed for 3 h in 0·5 % (v/v) glutaraldehyde and 3 % (w/v) paraformaldehyde mixture in 10 mm phosphate buffer, pH 7 (PBS). Thick sections (25–30 µm) were cut using a vibratome (Leica VT1000E, Germany) and incubated overnight at 4 °C with anti-IAA monoclonal primary antibody (Sigma-Aldrich) diluted 1 : 250 in PBS/BSA solution (10 mm phosphate solution, 0·8 % bovine serum albumin). Subsequently, sections were incubated with the secondary antibody (anti-mouse IgG alkaline phosphatase conjugate; Promega Italia, Milan, Italy) diluted 1 : 100 in PBS/BSA solution for 4 h. After washing, the sections were developed with NBT (nitro blue tetrazolium) and BCIP (5-bromo-4-chloro-3-indolylphosphate) mixture for 5 min, then rinsed with stop buffer (100 mm Tris-HCl, pH 8·0; 1 mm EDTA), mounted on slides and immediately observed and photographed. To verify the effectiveness of the immunolocalization technique, sections were processed with the omission of the primary IAA antibody and with a competition assay control as described by Kerk and Feldman (1995).
We identified a 885-bp cDNA with an open reading frame encoding 294 amino acid residues (molecular weight of 58 kDa). The encoded protein exhibited the highest degree of amino acid sequence identity with known CDKA of Prunus dulcis (92 %) and contained the PSTAIRE motif in the cyclin-binding domain, unique to the plant A-type CDKs (Fig. 1A, B). This cDNA was therefore designated as Cucma;CDKA (Cucurbita maxima CYCLIN DEPENDENT KINASE), in accordance with conventional plant CDK terminology (Renaudin et al., 1996).
A C. maxima CDKB was also isolated, comprising a 909-bp cDNA with an open reading frame encoding 302 amino acid residues. The encoded protein exhibited a high degree of amino acid sequence identity with known CDKB homologues of a variety of plants, the highest being 95 % with Glycine max CDKB2. The amino acid sequence of the newly isolated CDKB-related protein kinase cDNA displayed the PPTTLRE motif in the cyclin-binding domain, which characterizes the Arabidopsis B2-type CDKs (Fig. 1). We therefore designated this cDNA as Cucma;CDKB2. Most probably, the CDKB2 probe will be detecting both CDKB1 and CDKB2 expression, and from hereon these in situ patterns are mostly referred to as emanating from CDKB expression.
As illustrated in the longitudinal section in Fig. 2, the root apical meristem of C. maxima conforms to an open transverse (OT) meristem and as such it lacks a clear boundary between the epidermis and the root cap columella (Fig. 2A, B). The RAM is covered by a well-developed cap whose outermost layer can be traced basipetally up to a distance of 2400 µm away from the tip of the root cap (Fig. 2B–D). Two layers inside comprise a file of cells that shows substantial longitudinal thickening extending all the way down to the tip of the meristem proper (eumeristem); we take this to be the epidermis. This type of thickening is evident in a longitudinal section of the OT RAM of Julbernadia fabiflora (see Heimsch and Seago, 2008, fig. 27).
At the pole of the RAM is the predicted site of the quiescent centre (QC) but it is difficult to gauge its absolute size given the lack of a distinctive boundary between the pole of the RAM and root cap junction (Fig. 2A, B). The convergence of stelar and cortical domains can be also observed (Fig. 2A, B), but it is difficult to identify the exact end of these domains as the acropetal limit of the cortex is obscured by the open nature of the RAM. Consistent with the features of a eumeristem, isodiametric proliferative cells were observed at the RAM pole. However, and unusually, early vacuolization of the outermost cells of cortex extended all the way down to the eumeristematic pole (Figs 2A, B and 3A). As such, the eumeristem extends no longer than 400 µm and is a very small component of the RAM (Fig. 2A).
Basipetal to the eumeristem and extending over a distance of 1200 µm, vacuolarization underlying cell elongation and differentiation was clearly evident in both the outer cortex and the inner stele regions. Conversely, meristematic cells frequently dividing in either transverse or longitudinal planes were scattered in the inner cortex/outer stele (Figs 2B, C and and3C).3C). However, behind this zone and extending over a distance of 1200 µm, the meristematic state was maintained in very few layers of inner cortex/outer stele.
To try and define the transition point between the meristematic and non-meristematic state in Cucurbita roots, cell length was measured in each of the longitudinal files present in the RAM over a distance spanning a distance of 1200 µm from the pole (Fig. 2A). This might have indicated the cut-off point for the meristem on a file-by-file basis. However, for this RAM, we were unable to observe strict transition points in any of the files of cells that would have been highlighted by a sudden transition from small isodiametric to highly elongate cells (data not shown). This lack of a clear transition point between the meristematic and elongated/differentiated states is well illustrated in high-power images of median longitudinal sections that continue to show isodiametric cells frequently dividing in either transverse or longitudinal planes in the inner cortex/outer stele (Fig. 3A, C). Hence, alongside differentiated cells in some files, meristematic cells continued to be detected in the outer cortex and inner stele for at least a distance of 1200 µm from the tip of the eumeristem.
Another conspicuous feature of Cucurbita root tips, highlighted in Fig. 2, is precocious lateral root primordia which occur within the meristematic zone, the most apical of which was observed at a distance of approx. 750 µm from the root tip (Fig. 2A, C).
Overall, the anatomy of the root apex of C. maxima is consistent with an OT structure but the present analyses indicate some unusual features: (1) vacuolated and differentiated cells of the outer cortex and inner stele extending all the way down to the pole of the eumeristem; (2) proliferative cells of the inner cortex and outer stele extending basipetally way beyond the eumeristem; and (3) precocious formation of lateral root primordia close to the eumeristem.
Spatial expression patterns of CDKs in root apical meristems using in situ hybridization have been limited to Arabidopsis (Hemerly et al., 1993). Hence, and to our knowledge, this is the first report of spatial expression of cell cycle genes in open meristems (Fig. 4).
The heaviest in situ staining pattern for Cucma;CDKA;1 extended along the length of the eumeristem in the epidermis and in 2–3 longitudinal files of cells immediately inside of it (Fig. 4A, B). However, this signal is disrupted at the pole of this open meristem. A more diffuse signal was also detected in the outermost layers of the stele, including the pericycle, and relatively more strongly in lateral root primordia (Fig. 4A). Hence, the striking feature of this spatial expression pattern is the uniform signal along both the epidermis (heaviest) and the pericycle (lightest) (Fig. 4A, B) with an intermediate level of signal in the young lateral root primordia (Fig. 4C), which form comparatively close to the pole of the RAM. In contrast, Cucma;CDKB was extensive in the eumeristem but also along the meristematic regions of inner cortex and outer stele, as well as root primordia, but was notably absent from the epidermis (Fig. 4D–F). In the region tentatively identified as the QC, there was a much weaker CDKB expression pattern (Fig. 4E). Moreover, CDKB expression was not detected in all cells of the RAM, but instead in groups of cells scattered along the outer stele and inner cortex (Fig. 4E) and in lateral root primordia (Fig. 4F). Notably, signal was missing in the outer cortex, which was identified above as comprising vacuolated and differentiated cells that extended down to the tip of the eumeristem (Fig. 4D, F). Signal was missing in control sections processed with sense-probe (see Supplementary Data Fig. S1, available online).
The spatial expression of these cell cycle genes was consistent with CDKA marking tissues with a competence for proliferation such as the epidermis and pericycle whereas CDKB marks proliferating cells per se (see Discussion).
As mentioned in the Introduction, there is evidence that links the expression of cell cycle regulatory genes to auxin and cytokinin action. Our next step was to examine the extent to which spatial expression of Cucma;CDKA;1 and Cucma;CDKB matched the distribution of the endogenous cytokinins isopentenyladenine (iPA) and trans-zeatin (t-Z) and the auxin IAA (Figs 5 and and6).6). Immunolocalization of these plant growth regulators was performed for this purpose.
IPA was strongly localized in the root cap columella and cortical tissues of the primary root (Fig. 5A). However, there was a distinct absence of signal in the stelar and epidermal regions. It was also particularly noticeable that lateral root primordia were completely lacking in iPA signal except for the peripheral basal and apical regions of these secondary meristems (Fig. 5B). The distribution of t-Z was again strongest in the columella and cortical tissues but was absent from the eumeristem (Fig. 5C). Unlike iPA, t-Z was also concentrated in the inner stele that was basal to the eumeristem. However, as with the iPA distribution pattern, t-Z was absent in young primordia with the exception of the regions where a vascular connection at the base and a cap at the apex are first established (Fig. 5D). Taken together, the spatial distribution of these cytokinins was largely restricted to differentiated tissues in which cell proliferation would either be low or completely absent.
Unlike the spatially distinct distribution of iPA and t-Z, IAA was detected in a much more uniform pattern throughout the root (Fig. 6A). In particular, IAA signal was strong in the eumeristem and lateral root primordia, whereas a scattered distribution was detected in the root cap (Fig. 6A, B). Hence, IAA signal was a strong feature of the main body of cells in the primary root but unlike the cytokinins that we measured it was very prominent in meristematic regions such as the primordia and eumeristem although largely absent in the putative QC (Fig. 6C). For analysis of both cytokinins and auxin, an immunoreaction was not detected in control sections processed without secondary antibodies (Supplementary Data, Fig. S1).
We report, for the first time, on the cloning of pivotal cell cycle genes and their spatial expression, together with immunolocalization of cytokinins and IAA in the OT RAM of Curcubita maxima. The A-type CDK that we cloned has the conserved PSTAIRE sequence, placing it firmly alongside the solitary A-type CDK in Arabidopsis and, using the Renaudin et al. (1996) nomenclature, we classify it as Cucma;CDKA;1. The B-type CDK reported has PPTLRE instead of PSTAIRE in the conserved region and aligns most closely with an CDKB2-type rather than B1-type CDKs in which the conserved motif is PPTALRE (see Dudits et al., 2007) and hence we assign our B-type as Cucma;CDKB2. Clearly the B-type C. maxima gene is placed firmly in the family of B-type CDKs and is most closely related to that of Glycine max.
To our knowledge, the RAM of C. maxima has never been fully characterized. The root apical meristem of C. maxima is a typical OT meristem of cucurbitales type detailed by Heimsch and Seago (2008). For example, there is a prominent longitudinally thickened epidermis that typifies such OT RAMs. Moreover, vacuolated and differentiated cells were detected in the innermost stele and outer cortex that extend all the way down to the pole of the RAM. Apart from exhibiting an anatomy that is distinctly different from the well-documented closed meristems of A. thaliana (e.g. Jiang and Feldman, 2006) and Zea mays (Clowes, 1958; Barlow, 1974), the RAM of C. maxima also diverges quite markedly from other well-documented open meristems such as those of Vicia faba and Pisum sativum (MacLeod, 1991). In our view, the RAM of C. maxima comprises a eumeristem positioned centrally at the pole of the various tissue systems but the meristem extends basipetally only along the outer stele and inner cortical cell files. Hence, we conclude that this RAM is unlike anything reported previously.
Although Mallory et al. (1970) noted that lateral root primordia are initiated close to the root tip, they also commented on the difficulty in observing the initiation of primordia in this species. We detected lateral root primordia at approx. 750 µm from the tip of the eumeristem.
Cucma;CDKA;1 spatial expression is most prominent along the epidermis, a domain that in OT meristems contributes cells to the lateral root cap (Heimsch and Seago, 2008). In order to do so, these cells must be able to divide both periclinally and anticlinally; the latter is the normal plane of division for epidermal cells. Thus, the prominent expression of Cucma;CDKA;1 in the epidermis is a further example of the characteristic feature of an A-type CDK to mark mitotically competent cells. In other words, these cells are not permanently meristematic but have the potential to divide, with their descendants able to contribute to the root cap of a new secondary root.
In Arabidopsis, CDKA;1 is required for the G2/M transition of the cell cycle but unlike other CDKs is also expressed along the pericycle, a tissue that comprises competent meristematic cells, but not all of which divide (Hemerly et al., 1993); we also detect Cucma;CDKA;1 spatial expression in this tissue. A root system typically comprises clusters of lateral root primordia separated in time and space from other clusters. The clusters tend to conform to a basipetal sequence of younger (nearer the tip) to older (further from the tip) (Charlton, 1983). Proof of the meristematic competence of pericycle cells that normally do not divide is through the classic action of exogenous auxin in stimulating an over-production of lateral roots (Blakely et al., 1982; Vanneste et al., 2007). Hence, we suggest that Cucma;CDKA;1 spatial expression marks cells of tissues that have competence for mitosis but are not necessarily dividing, a conclusion not dissimilar from that of Hemerly et al. (1993) from their study of CDKA;1 expression in root apical meristems of Arabidopsis.
Spatial expression of Cucma;CDKB2 is rather different in that we observe alternating intense and diffuse expression across the various tissues of the RAM, which visually provides a scattered staining pattern. Notably, Cucma;CDKB2 is expressed in meristematic cells of the eumeristem, pericycle and lateral root primordia. Indeed, the restricted expression profile of Cucma;CDKB2 to meristematic cells conforms to the archetypal plant CDKB that is required to drive cells into mitosis (Sorrell et al., 2001; Orchard et al., 2005). The rather spotty pattern of CDK spatial expression is not dissimilar from that provided by cdc2c and cdc2d in Antirrhinum inflorescences (Doonan, 1998). These genes are expressed at G2/M, and are now reclassified as Antma;CDKB1;1 and Antma;CDKB2;1 (Dudits et al., 2007). Doonan (1998) concluded that the spotty transcription pattern of these genes indicated that their transcripts are present for only a short while during the cell cycle, which was also true for a tobacco B-type CDK in tobacco BY-2 cells (Sorrell et al., 2001; Orchard et al., 2005). This could well be the case for Cucma;CDKB.
Immunolocalization of the cytokinins iPA and t-Z is consistent with a major differential location of these plant growth regulators in differentiated cells of the cortex, innermost stele and root cap. Also, and remarkably, these cytokinins are completely absent from the lateral root primordia although they were detected in differentiated cells at the base of the primordia where typically a connection of differentiated vascular cells will serve to deliver nutrient to the developing primordium (MacLeod and Francis, 1976). In other words, quiescent cells formerly at the basipetal margin of the RAM become proliferative and contribute to the RAM. Conversely, manipulating roots to have increased amounts of cytokinins can repress root growth and lateral root initiation (Li et al., 1992; Dello Ioio et al., 2008a, b). Hence, what emerges and accords with our findings for the RAM of C. maxima, where cytokinins localize in differentiated cells, is an inverse relationship between cytokinins and proliferative domains. This is a somewhat unusual interpretation of a plant growth regulator that was defined as one that can stimulate cell division (at least in vitro) (Skoog and Miller, 1957) and is necessary for the G1/S and G2/M transitions of the cell cycle (Riou-Khamlichi et al., 1999; Zhang et al., 2005). Trewavas (1981, 1982) argued that different cells develop different sensitivities to plant growth regulators at different stages of development. Regarding cytokinins, it is also possible that different cytokinins act at different optimal concentrations at the cell, tissue and organ levels and that such subtle interactions cannot be unmasked solely by immunolocalization studies.
In the C. maxima RAM, IAA is distributed uniformly among the various tissue systems except the periphery of the root cap where cells are probably in the grip of programmed cell death. Hence, in the RAM, IAA is a ubiquitous marker of root cells regardless of whether they are meristematic or differentiated. That IAA delineates root cells in this way should not be surprising given the very early alignment of PIN proteins during early embryogenesis that facilitates IAA transport into the embryonic radicle (Jenik et al., 2007). Note also the classic dogma that a high auxin to cytokinin ratio favours root development (Skoog and Miller, 1957). Moreover, auxin promotes cell division by suppressing cytokinin signalling, whereas cytokinins promote cell differentiation by modulating auxin distribution (Růžička et al., 2009). Again, the differential sensitivity of dividing and differentiating cells might well be regulating the differential spatial expression of the A- and B-type CDKs in the root apex of C. maxima.
In conclusion, the spatial expression of Cucma;CDKA;1 is in accord with this gene being expressed in cells that have meristematic competence, but that of Cucma;CDKB2 is more localized to proliferative cells in meristematic tissues of the RAM. Clear links between the spatial expression of these cell cycle genes and the general distribution of iPA, t-Z and IAA are not immediately evident although we can conclude that cytokinin localization and the expression pattern of CucmaCDKA;1 appear to be mutually exclusive in these roots.
Supplementary data are available online at www.aob.oxfordjournals.org and illustrate that immunoreaction was not detected in control sections processed without secondary antibodies for cytokinins (iPA), t-Z and IAAs.
This work was supported by grants from Università della Calabria (MIUR-ex60 %) to M.B.B. and by Cardiff University.