Positive regulation of minichromosome maintenance gene expression, DNA replication, and cell transformation by a plant retinoblastoma gene

Abstract

Retinoblastoma-related (RBR) genes inhibit the cell cycle primarily by repressing adenovirus E2 promoter binding factor (E2F) transcription factors, which drive the expression of numerous genes required for DNA synthesis and cell cycle progression. The RBR-E2F pathway is conserved in plants, but cereals such as maize are characterized by having a complex RBR gene family with at least 2 functionally distinct members, RBR1 and RBR3. Although RBR1 has a clear cell cycle inhibitory function, it is not known whether RBR3 has a positive or negative role. By uncoupling RBR3 from the negative regulation of RBR1 in cultured maize embryos through a combination of approaches, we demonstrate that RBR3 has a positive and critical role in the expression of E2F targets required for the initiation of DNA synthesis, DNA replication, and the efficiency with which transformed plants can be obtained. Titration of endogenous RBR3 activity through expression of a dominant-negative allele with a compromised pocket domain suggests that these RBR3 functions require an activity distinct from its pocket domain. Our results indicate a cell cycle pathway in maize, in which 2 RBR genes have specific and opposing functions. Thus, the paradigm that RBR genes are negative cell cycle regulators cannot be considered universal.

Retinoblastoma-related (RBR) proteins are negative cell cycle regulators conserved in animals and plants (1–5). They are known as pocket proteins by virtue of a conserved region located in their C-terminal half (i.e., the pocket domain), which is largely responsible for protein–protein interactions and biological function. It comprises 2 highly conserved A and B domains separated by a nonconserved spacer of varying length, as well as a less-conserved C-terminal region. In mammals, these proteins are well characterized and comprise RB, p107 and p130. RBR proteins repress cell cycle progression primarily through the inhibition of E2 promoter binding factor (E2F)-dependent transcription, which is required to express many genes involved in S-phase and cell cycle progression, including the minichromosome maintenance 2–7(MCM2–7) family of DNA replication factors (4, 6–11). Thus, repression or expression of E2F targets, such as MCM2–7 genes, is a good indicator of the activity of RBR and E2F proteins.

In overexpression studies, all 3 mammalian pocket proteins inhibit E2F-dependent gene expression, recruit chromatin remodeling complexes, actively repress transcription, and arrest cell growth (3, 12, 13). However, gene knockouts in mice have shown that, in addition to having overlapping functions, individual pocket proteins have unique properties (14). In particular, although only RB is considered a bona fide tumor suppressor, both p107 and p130 can functionally compensate for RB inactivation or loss in certain contexts and restrain cell proliferation (15).

Inactivation of RBRs can occur by various mechanisms. For example, it is well documented that several oncoviral proteins, such as SV40 large T antigen (Tag), adenovirus E1A, human papillomavirus E7, and certain plant viral proteins, such as wheat dwarf virus RepA (16–19), interact with the A/B pocket domain through a conserved LxCxE motif (20); thus, disrupting the ability of RBRs to bind and repress E2F transcription factors. Also, mutation of key residues essential for pocket conformation can disrupt RBR activity (21). One such mutation involves the cysteine at position 706 in human RB (22–25). In maize, the equivalent cysteine residue is required for interaction of RBR1 and RBR3 with viral oncoproteins and RepA (18, 19).

The plant RBR-E2F pathway appears similar to that in animals (4, 5, 26); however, although many plants appear to possess only 1 RBR gene, maize and related cereals have at least 2 distinct types, RBR1 and RBR3 (19, 27, 28). Maize RBR1 represses expression of RBR3 through a mechanism that most likely involves inhibition of E2F transcription factors (19), and can be circumvented by expression of RepA (19, 29), which inhibits RBR1 through a pocket-dependent mechanism (17, 18, 30). Ectopic expression of RepA provides an effective means to down-regulate RBR1 and release E2F activity in an inducible manner, and considerably stimulates cell transformation and callus growth in maize embryos (29). Inhibition of RBR3 expression by RBR1 suggests 2 possible alternative functions for RBR3 in cell cycle control: RBR3 could have a novel positive role or, similar to p107 in mammals, it could provide a fail-safe mechanism by restraining the cell cycle after loss or inactivation of RBR1 (15, 19, 27, 31).

The ability of plant cells to undergo transformation and regeneration is associated with cell cycle activity (29, 32, 33), and increasing evidence supports the idea of a critical role played by the RBR/E2F pathway (29, 33, 34). A compensatory mechanism between RBR1 and RBR3 could explain, at least in part, the well-known recalcitrance of maize and related grasses to genetic transformation (19, 27, 35). However, the role of RBR3 in E2F-dependent gene expression, cell cycle progression, and cell transformation is unknown. A negative cell cycle role for RBR3 would be consistent with other RBR genes, but formal evidence is lacking. Also, the existence of an RBR1/RBR3 compensatory mechanism has not been demonstrated in vivo.

Using a combination of approaches, involving the up- or down-regulation of RBR3 (either alone or in combination with RBR1 down-regulation) in cultured maize embryos, we demonstrate that RBR3 has a necessary and positive role in the expression of the MCM gene family, DNA replication, and cell transformation. Also, evidence is presented indicating that an activity attributable to one or more domains distinct from the pocket domain is required for RBR3 function. These results rule out a compensatory mechanism between RBR1 and RBR3 in maize, and shed light on important differences concerning the function of the RBR-E2F pathway between cereals and other dicotyledonous plants.

Results

To understand the function of RBR3, we uncoupled its expression from the inhibitory control of RBR1, and analyzed how this condition impacts maize cell transformation, DNA replication, and the expression of selected E2F targets. We first investigated the consequences of RBR3 down-regulation on cell transformation. Ectopic expression of RepA is an effective means to inhibit RBR function in plants by a mechanism that specifically involves the release of E2F transcription factors, which increases cell transformation in maize (29, 36). However, RepA-dependent inhibition of RBR1 in maize resulted in increased expression of RBR3 (19), making it difficult to dissect their individual roles. Therefore, we uncoupled RBR3 from RBR1 by specifically down-regulating RBR3 using RNAi and RNA antisense approaches. A description of the constructs used in these experiments is given in Table 1. On expression of RepA with the nopaline synthase (Nos) promoter, transformation efficiency of embryos increased from 10.9% in control experiments to 27.5% (Table 2), consistent with previous results (29). However, expression of RBR3 RNAi (RBR3DSH) or antisense (RBR3NASx2) constructs, each driven by the ubiquitin 1 (UBI) promoter, resulted in only 3.2 and 6.1% transformation efficiency, respectively. This result indicates that RBR3 has a positive role in maize transformation, and that its down-regulation negatively impacts this process. Coexpression of RepA and RBR3DSH or RBR3NASx2 constructs resulted in transformation efficiencies close to control values (8.3 and 9.9%, respectively), consistent with both RepA and RBR3 having positive roles in the pathway controlling cell transformation.

Table 1.

Constructs used in transformation experiments

Plasmid no.	Plasmid name	Promoter	Coding sequence	3′ sequence	Reference or source
PHP9058	UBI::PAT∼GFP	UBI	moPAT fused to moGFP	pinII	ref. 29
PHP21829	UBI::PAT∼YFP	UBI	moPAT fused to ZsYFP-N1	pinII	Clontech Laboratories
PHP3953	UBI::GUS	UBI	β-glucuronidase from E. coli	pinII	ref. 54
PHP17008	Nos::RepA	Nos	WDV RepAm	pinII	ref. 29
PHP25473	UBI::RBR3NASx2	UBI	N/A	None	This work
PHP25475	UBI::RBRDSH	UBI	N/A	None	This work
PHP30001	UBI::RBR3HA	UBI	RBR3HA	pinII	This work
PHP30003	UBI::RBR3HA-C788G	UBI	RBR3HA-C788G	pinII	This work
PHP30004	35S::RBR3HA-C788G	35S	RBR3HA-C788G	pinII	This work

Maize-optimized moPAT and moGFP indicate maize codon-optimized PAT and GFP sequences. RepA_m indicates an ″intronless″ version of RepA. N/A, not applicable.

Table 2.

RBR3 is required for maize cell transformation

Construct	Embryos, transformed/bombarded	Transformation rate, %
RBR3 down-regulation
Control	38/350	10.9
Nos::RepA	103/375	27.5**
UBI::RBR3DSH	13/400	3.2**
Nos::RepA + UBI::RBR3DSH	31/375	8.3
UBI::RBR3NASx2	26/425	6.1*
Nos::RepA + UBI::RBR3NASx2	42/425	9.9
RBR3 overexpression
UBI::GUS, control	27/260	10.4
UBI::RBR3HA	28/270	10.4
UBI::RBR3HA-C788G	23/283	8.1
35S::RBR3HA-C788G	16/281	5.7*

* and **, significance levels (*, 0.05 probability; **, 0.01 probability) relative to control in each of the 2 independent experiments (RBR3 down-regulation and RBR3 overexpression).

Gene expression analysis by RT-PCR showed that on expression of RepA, RBR3 was up-regulated 3.8-fold relative to control calli (Fig. 1), which confirmed previous findings indicating that RBR3 is negatively controlled by RBR1, most likely through inhibition of E2F activity (19). In fact, we demonstrated earlier that up-regulation of RBR3 depends on the presence of the LxCxE motif in RepA that mediates interaction with and inhibition of RBR1, and that E2F and their partner DP proteins can interact with the RBR3 promoter (19). Accumulation of RBR3 transcript was reduced ≈50% in calli transformed with either the RBR3DSH or RBR3NASx2 construct, and coexpression of RepA and RBR3DSH or RBR3NASx2 resulted in RBR3 RNA expression levels that were 2.1- and 1.8-fold higher, respectively, than the control (Fig. 1). No major change in RBR1 or RBR2 RNA expression was detected, indicating that the transgenic experiments involving RepA expression and RBR3 down-regulation impinged specifically on the expression of RBR3 (Fig. 1).

Fig. 1.

Down-regulation of RBR3 expression reduces levels of all 6 MCM2–7 transcripts. Maize embryos were transformed with different constructs by particle bombardment, the resulting calli genotyped, and RNA was extracted, pooled, and assayed for gene expression by RT-PCR (19). See Table 1 for construct description. Calli transformed with constructs for RBR3 down-regulation by RNAi (RBR3DSH) and antisense (RBR3NASx2) displayed reduced expression of RBR3 RNA as well as transcripts of the MCM2–7 family. Expression of 3 non-E2F target RNAs (i.e., RBR1, RBR2, and actin) was not affected by down-regulation of RBR3. A representative experiment is shown. See Fig. S2 for comprehensive quantitative analysis.

We next examined the effect of down-regulating RBR3 on the expression of MCM2–7 genes. Three MCM genes from maize were previously characterized, MCM3 (37), MCM6 (38), and MCM7 (39). Through database mining, we identified the 3 remaining members of the MCM2–7 family [supporting information (SI) Fig. S1]. On expression of RepA, all 6 MCM genes were up-regulated 2- (i.e., MCM3) to 7.4-fold (i.e., MCM5) (Fig. 1 and Fig. S2), indicating maize MCM2–7 genes are negatively regulated by RBR1 and are most likely targets of E2F activity, consistent with what is known about MCM gene expression in higher eukaryotes (4, 9–11). However, down-regulation of RBR3 resulted in decreased expression of all members of the MCM2–7 gene family, which ranged from 10 to 80% of the control (Fig. 1 and Fig. S2). Thus, expression of RBR3 is required for transcription of at least a subset of known E2F targets (i.e., MCM2–7 genes), although individual MCM genes appeared to differ in their sensitivity to RBR3 down-regulation. These data support the transformation results, and indicate the positive role of RBR3 in cell transformation could be mediated by its regulation of DNA synthesis through controlling the expression of at least some essential DNA replication factors.

We next asked whether RBR3 is sufficient to increase cell transformation and MCM gene expression. RBR3 was engineered with a hemagglutinin (HA) epitope tag (RBR3HA) and overexpressed under the control of the UBI promoter (29). As expected, RBR3 RNA was expressed at high levels (21.7-fold over control values), and, with the exception of MCM2, all of the MCM genes were up-regulated 1.6- to 2.2-fold (Fig. 2 and Fig. S3). Expression of RBR3HA had no noticeable effect on actin and RBR1 gene expression. However, no significant change in transformation efficiency was observed (Table 2). One explanation for these results relates to the inability of RBR3 to stimulate expression of the full complement of genes required to achieve cell proliferation (see below). Because the spectrum of genes regulated by RBR3 is unknown, understanding of the reasons why RBR3 is not a limiting factor in maize transformation will probably require genome-wide profiling of gene expression. Thus, we conclude that RBR3, although required, is not limiting for transformation.

Fig. 2.

Overexpression of RBR3 induced increased expression of MCM3–7 genes, whereas expression of the C788G mutant resulted in the down-regulation of all 6 MCM2–7 genes. RBR3 was overexpressed under the control of the maize ubiquitin1 (UBI) promoter (RBR3HA). Also, an RBR3HA-C788G mutant was expressed with the UBI and CaMV 35S promoters. Expression analysis was carried out as in Fig. 1. Expression of RBR1 and actin RNAs was largely unaffected. A representative experiment is shown. See Fig. S3 for comprehensive quantitative analysis.

Having established that RBR3 has a positive role in both regulating MCM gene expression and cell transformation, we asked whether these effects solely depend on its pocket region or involved a pocket-independent mechanism. Although detailed information is available regarding the role of the C-terminal pocket domain of RBR proteins, relatively little is known about activity that is pocket-independent. We previously showed that RBR3 has a conventional pocket protein activity that requires a conserved cysteine in the B domain at position 788, similar to other RBR proteins (19). In vitro interaction studies indicated that mutating this residue to glycine (C788G) greatly reduced the ability of RBR3 to interact with RepA, adenovirus E1A, and human papillomavirus E7 proteins (19). Therefore, we reasoned that ectopically expressing the C788G-mutagenized form of RBR3 (RBR3HA-C788G) in callus would create the means, in a dominant-negative manner, to establish whether or not RBR3 also functions through a pocket-independent mechanism (Fig. S4). Expression of RBR3HA-C788G driven by the strong UBI promoter increased RBR3 transcript 16.7-fold over the control. However, it not only failed to stimulate expression of MCM genes, but also resulted in their collective down-regulation (Fig. 2 and Fig. S3). Thus, high expression of the dominant-negative RBRHA-C788G allele interfered with endogenous RBR3 activity. Essentially similar results were obtained on expression of the C788G mutant under control of the weaker cauliflower mosaic virus (CaMV) 35S promoter (Fig. 2 and Fig. S3), indicating that high expression levels of RBR3HA-C788G are not necessary for the inhibition of endogenous RBR3, and that expression of UBI::RBR3HA-C788G probably saturated the system. Consistent with the inhibition of MCM gene expression, a decrease in transformation efficiency was also observed on expression of the RBR3-C788G mutant allele (Table 2). This result indicates that expression of the C788G allele competed with endogenous RBR3 for some essential protein–protein interaction, which appears to be mediated by a domain(s) distinct from the pocket region (Fig. S4 and Discussion). Thus, we conclude that RBR3 controls MCM gene expression and cell transformation through both the pocket domain and a pocket-independent activity.

Because RBR3 controls expression of key genes involved in the regulation of DNA synthesis, we investigated whether its overexpression or down-regulation would be sufficient to stimulate or repress DNA replication, respectively. Control and transgenic calli were pulse-labeled with the thymidine analogue BrdU and analyzed for their rate of DNA synthesis by using a Southwestern method with anti-BrdU antibodies (Fig. 3). Expression of UBI::RBR3HA in 2 independent experiments resulted in 1.6- and 2-fold greater rates of DNA synthesis, compared with control, approaching the stimulation obtained by RepA expression (2.6-fold). In contrast, DNA replication was severely inhibited on expression of the RBR3HA-C788G dominant negative mutant allele (i.e., 0.2-fold relative to control). These results indicate that overexpression of RBR3 is necessary and sufficient to stimulate DNA replication in vivo, and are consistent with the positive roles of RBR3 in regulating MCM gene expression and cell transformation.

Fig. 3.

RBR3 is necessary and sufficient for DNA replication. Two independent experiments involving in vivo incorporation of BrdU into newly synthesized DNA in different transgenic maize calli are shown. Calli were pulse-labeled with BrdU, and their DNA samples were pooled and probed with an anti-BrdU antibody. Pools ranged from 5 to 30 independent samples, as indicated. DNA loadings were confirmed by hybridization to ³²P-labeled total maize DNA. Expression of either Nos::RepA and UBI::RBR3HA increases callus DNA replication, whereas expression of the dominant negative allele UBI::RBR3HA-C788G impairs it. Pairwise differences in BrdU incorporation levels for each treatment relative to the control were significant as determined by t test (P < 0.05).

To understand whether RBR3 also promotes the G2/M-phase transition, we compared the relative frequencies of G1 and G2 nuclei in calli expressing RBR3HA (UBI::RBR3HA) with control calli (Fig. 4). Representative flow-cytometric profiles are shown in Fig. 4A, which indicate an increase in the fraction of nuclei in G2 phase on expression of RBR3HA. This analysis revealed that cells tend to stabilize in the G2 phase on overexpression of RBR3, as shown by the ≈50% reduction in the G1/G2 nuclei ratio in UBI::RBR3HA-transformed nuclei, compared with control (0.56 versus 1.0) (Fig. 4B). This result suggests that although RBR3 overexpression stimulates DNA replication, it is not equally effective at stimulating the transition of G2 cells to M-phase. This apparent block in G2 would compensate for the stimulatory role of RBR3 in DNA replication, and curtail a potential net gain in the rate of cell proliferation. This interpretation is in agreement with the observation that RBR3 fails to stimulate cell transformation when overexpressed (Table 2). Therefore, we conclude that RBR3 does not promote the G2/M-phase transition as effectively as it does the G1/S-phase transition, which would explain the failure of RBR3 overexpression to stimulate cell transformation.

Fig. 4.

Overexpression of RBR3 results in a block in the G₂ phase of the cell cycle. Control and RBR3-overexpressing calli (UBI::RBR3HA) were analyzed by flow cytometry. (A) Representative flow-cytometric profiles. (B) Histogram showing an ≈50% decrease in the G1/G2 nuclei ratio in the calli transformed with UBI::RBR3HA compared with control, suggesting an inhibition of the G2/M-phase transition in the cell cycle. Differences in G1/G2 ratios were significant as determined by t test (P < 0.01).

Discussion

We demonstrated that a maize member of the RB gene family, RBR3, has an essential and positive role in regulating MCM gene expression, DNA replication, and cell transformation. These properties appear to be unique among pocket proteins. In fact, it is well established that all RB genes in animals have negative roles in DNA replication and cell cycle progression (3, 40). Similar findings were reported for plant RBRs, and a number of studies have shown that cell proliferation is stimulated on down-regulation or loss of RBR activity in tobacco, Arabidopsis, and maize (29, 36, 41–44). Interaction with and inhibition of plant RBR by geminivirus RepA proteins has been established (17–19, 45), and shown to cause up-regulation of E2F-dependent gene expression (19, 36), which is in agreement with the current model in which RBR inhibits the G1/S-phase transition during the cell cycle by repressing E2F transcription factors (6, 7, 46, 47). The requirement for RBR3 for the expression of MCM2–7 genes is novel and appears unique within the RBR family. For example, several previous studies profiled gene expression in animal systems on down-regulation or loss of pocket protein function (9, 10, 48–51); however, in no instance were MCM genes down-regulated. Also, our data indicate that RBR3 is limiting for the expression of all but one of the MCM2–7 genes. Thus, we have identified a previously undescribed function of a plant RBR gene: maize RBR3 appears to have a key, positive regulatory role between RBR1 and MCM2–7 genes in the RBR-E2F pathway leading to DNA replication (Figs. 1–3 and 5). Because MCM2–7 are conserved E2F targets in all systems studied so far, it is likely that RBR3 impinges directly or indirectly on at least a subset of E2F transcription factors. A detailed interaction analysis between RBR3 and prospective E2F protein binding partners will be required to define the specificities of RBR3/E2F binding.

Fig. 5.

Model showing RBR3 role in the RBR/E2F pathway controlling the expression of MCM2–7 genes, DNA replication, and cell transformation. RepA inhibits RBR1; thus, stimulating the pathway leading to S-phase gene expression, DNA synthesis, and cell transformation through up-regulation of RBR3. The transgenic approaches to down- or up-regulate RBR3 are indicated in italics. The dotted line illustrates a potential inhibitory effect of RepA on RBR3 ruled out by this work.

The outcomes of expressing the C788G allele of RBR3 suggest that the function(s) of RBR3 in cell cycle regulation and transformation requires an activity that is distinct from the canonical role of the pocket domain. The C788G mutation strongly diminishes the pocket binding activity of RBR3 (19). Ectopic expression of this allele would be expected to have no consequence if RBR3 function was exerted solely through its pocket domain, because an intact complement of RBR3 pocket activity would be retained due to the presence of endogenous RBR3. However, a marked reduction in MCM2–7 gene expression, DNA synthesis, and cell transformation was observed on expressing the C788G mutant allele. We interpret these outcomes as an indication of a requirement of RBR3 for an activity that is independent from its pocket domain. In the model shown in Fig. S4, we represent this unspecified activity as a protein or protein complex binding to the N-terminal region of RBR3. Overexpression of RBR3 with a functional pocket domain would increase the pool of RBR3 bound to the unknown factor, effectively enhancing the normal role of RBR3 in controlling MCM gene expression and DNA replication (Fig. S4B). According to this interpretation, the RBR3 expression level, rather than the concentration of the unknown binding factor, is limiting for RBR3 activity. Expression of the C788G allele would out-compete endogenous RBR3 for binding to the unknown factor(s), but, because of its compromised pocket domain, it could not function, causing the negative effects described above (Fig. S4C). Thus, both the unspecified activity independent of the RBR3 pocket and its pocket domain would be required for RBR3 function. More research will be needed to elucidate the identity of the unknown factor required for RBR3 activity. Investigation in various systems has primarily focused on describing the function of the pocket domain of RBRs, and there is little information on pocket-independent activities. For example, the N-terminal domains of RBRs show a good degree of sequence conservation, at least within RBR subfamilies; however, they have not been thoroughly investigated (52). We previously identified an amino acid sequence in proximity of the RBR3 N terminus spanning residues 78–124, which is similar to a conserved domain in p107 and p130 that is required for growth suppression and inhibition of CDK/CycA/E complexes (19, 53). However, it is unknown whether this domain has a role in the RBR3 functions we have described.

Several nonmutually exclusive pathways downstream of RBR3 could be envisioned that eventually stimulate E2F-dependent expression of MCM genes and subsequent DNA replication and transformation. They include the stimulation of activator E2Fs, the inhibition of repressor E2Fs, and an inhibition of RBR1 in a dominant-negative fashion. All 3 scenarios, in their most direct interpretation, would involve RBR3 binding to E2Fs, which appears likely as mentioned above. Although our data do not rule out the existence of partial overlap between RBR1 down-regulation and RBR3 overexpression, they suggest that the pathways controlled by RBR1 and RBR3 are not identical, and that RBR3 does not stimulate downstream processes simply by inhibiting RBR1 activity. If RBR3 solely down-regulated RBR1 activity, then RBR3 overexpression would essentially recapitulate the full spectrum of phenotypes resulting from inhibition of RBR1, such as those obtained by expressing RepA. However, the fact that ectopic expression of RepA is sufficient to trigger a substantial increase in transformation, whereas RBR3 overexpression is not, indicates that this scenario is not the case. Clearly, understanding the precise mechanism of action of RBR3 represents a challenge for future research. Profiling global gene expression changes on up- or down-regulation of RBR3, and investigating its interactions with members of the E2F family could help address this question.

Because RepA expression results in up-regulation of MCM genes, increased DNA replication, and transformation, all of which require RBR3 function, our findings appear to rule out a role for RepA in inhibiting RBR3 in vivo, although the 2 proteins clearly interact in vitro in a pocket- and LxCxE-motif-dependent fashion (Fig. 5) (19). At least 2 alternative explanations fit with these results. RepA and RBR3 may not interact in vivo, or, even if such an interaction does occur, it does not influence RBR3 activity. Alternately, because binding of RepA (or other oncoviral protein partners) to pocket proteins through their LxCxE-motifs results in a conformational change of the pocket domain resulting in the displacement of E2F, and because RBR3 pocket seems necessary for its activity, we favor the view that RepA and RBR3 do not interact in vivo in this context. Thus, our results rule out a compensatory mechanism between RBR1 and RBR3 in maize embryogenic callus that results in inhibition of the cell cycle and transformation (19, 27).

The organization of the RBR gene family in maize appears to be shared by and unique to cereals, but the biological implications are unknown. Because RBR genes have been universally viewed as repressors of cell cycle activity, the 3 maize RBR genes were expected to play largely redundant roles in controlling the cell cycle. Consequently, it was thought that stimulation of the pathways they control could be achieved by their simultaneous down-regulation. Our findings indicate that at least 1 type of RBR gene in maize (and possibly its orthologs in related grass species as well) has a positive cell cycle role. Therefore, strategies aimed at the manipulation of the RBR/E2F pathway to stimulate cell proliferation and increase crop yield in cereals will need to be revised. Although certain RBR genes may be down-regulated, others may be up-regulated, possibly simultaneously, to stimulate the cell cycle. Clearly, understanding the RBR/E2F pathway in Arabidopsis and other dicots provides only limited insight relative to cereals. Improvement of cereals through the manipulation of the RBR/E2F pathway will have to consider the different and opposing roles of distinct RBR genes.

Materials and Methods

For details, see SI Materials and Methods.

Plasmids.

The promoters and terminators used in this study were described earlier (29), and include maize UBI, Nos, and CaMV 35S promoters; the potato proteinase inhibitor (pin) II terminator was used as the flanking 3′ end sequence in most constructs (Table 1). The UBI::PAT∼GFP and Nos::RepA constructs were previously described (29). In some cases, YFP (Clontech Laboratories) replaced GFP in the UBI::PAT∼GFP plasmid, or the Escherichia coli uidA gene (encoding β-glucuronidase) replaced the PAT∼GFP fusion to give the UBI::GUS plasmid (54).

Cell Transformation Experiments.

Immature maize (Zea mays L.) high-type II embryos were transformed by particle bombardment essentially as described earlier (29). Each transformation experiment included the selectable-visible marker UBI::moPAT∼moGFP (or YFP)::PinII, which is a fusion between maize-optimized (mo) phosphinothricin acetyl transferase (PAT) that confers resistance to the herbicide BASTA, and GFP or YFP. Transgenic calli were selected and scored as described (29). Transformation efficiency is defined as the percentage of embryos resulting in transgenic calli relative to the total number of embryos targeted for transformation by particle bombardment. Significance was determined by using a 2 × 2 (pairwise fashion) contingency table to determine Chi² values for each treatment relative to control with a 0.05 probability (*) or a 0.01 probability (**). Significance was determined separately for the 2 different experiments (RBR3 down-regulation and RBR3 overexpression).

RT-PCR Analyses.

Individual calli originating from independent transformation events were screened for the presence/absence of the transgene(s) by PCR and, where appropriate, the expression of transgenic RNA by RT-PCR. RNA samples were pooled before RT-PCR analysis. Description of RT-PCR targets, primers used and RNA pools are given in Tables S1–S3. Each RT-PCR was repeated at least 3 times, and average expression and SD values are given in Fig. S2 and Fig. S3.

BrdU Incorporation Analysis.

Calli were incubated in an Eppendorf tube with Murashige and Skoog Basal Salt Mixture containing 1 mM BrdU for 5 h at 28 °C on a rotating incubator in the dark. Callus DNA was extracted from a number of independent transformants ranging from 5 to 30, depending on the treatment, and genotyped according to standard procedures (19). Pooled DNA samples (250 ng) were denatured, neutralized, and blotted onto a nitrocellulose membrane. BrdU was detected by using a primary anti-BrdU antibody and a secondary antibody conjugated to horse radish peroxidase. Three exposures were taken at different times and averaged. X-ray films were scanned and digitally quantified. Blots were then probed with ³²P-labeled total genomic B73 DNA, exposed to Phosphoimager screen, and digitally quantified. BrdU incorporation was quantified relative to the amount of DNA in each sample.

Flow Cytometry.

Ten control calli, transformed with the UBI::PAT∼GFP::pinII construct and 10 calli transformed in addition with the UBI::RBR3HA::pinII construct were analyzed by flow-cytometry to investigate the effect of overexpressing RBR3 on the relative frequency of unreplicated (G₁ phase) and replicated (G₂ phase) nuclei. Nuclei were released from callus tissue and analyzed with a Partec CAA-II flow analyzer as previously described (55, 56).