The retinoblastoma tumour suppressor, Rb, has two major functions. First, it represses genes whose products are required for S-phase entry and progression thus stabilizing cells in G1. Second, Rb interacts with factors that induce cell-cycle exit and terminal differentiation. Dictyostelium lacks a G1 phase in its cell cycle but it has a retinoblastoma orthologue, rblA.
Using microarray analysis and mRNA-Seq transcriptional profiling, we show that RblA strongly represses genes whose products are involved in S phase and mitosis. Both S-phase and mitotic genes are upregulated at a single point in late G2 and again in mid-development, near the time when cell cycling is reactivated. RblA also activates a set of genes unique to slime moulds that function in terminal differentiation.
Like its mammalian counterpart Dictyostelium, RblA plays a dual role, regulating cell-cycle progression and transcriptional events leading to terminal differentiation. In the absence of a G1 phase, however, RblA functions in late G2 controlling the expression of both S-phase and mitotic genes.
Citation: Strasser K, Bloomfield G, MacWilliams A, Ceccarelli A, MacWilliams H, Tsang A (2012) A Retinoblastoma Orthologue Is a Major Regulator of S-Phase, Mitotic, and Developmental Gene Expression in Dictyostelium. PLoS ONE 7(6): e39914. doi:10.1371/journal.pone.0039914
Editor: Juan Mata, University of Cambridge, United Kingdom
Received: June 8, 2011; Accepted: June 4, 2012; Published: June 29, 2012
Copyright: © 2012 Strasser et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by a Discovery grant from the Natural Sciences and Engineering Research Council of Canada, http://NSERC.ca. GB was supported by MRC core funding. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Dictyostelium is not a traditional model system for cell-cycle studies. Most of the work done on the cell cycle of this organism was motivated by other questions, particularly the desire to understand Dictyostelium development. In the latter process, amoebae differentiate into either stalk cells or spores. Pathway choice is determined by the cell-cycle position when development begins . Regulation of the eukaryotic cell cycle is fairly well understood, but this has brought few insights about developmental-pathway choice in social amoebae largely because the Dictyostelium cell cycle is atypical. In Dictyostelium the G1 phase is very short or non-existent . The G1/S transition is the major point at which the cell cycle is controlled in most eukaryotes. Since Dictyostelium appears to bypass this point, it has been unclear how the Dictyostelium cell cycle is organized.
Despite this unusual configuration Dictyostelium has a relatively normal-appearing set of genes regulating the cell cycle. There is a single cyclin-dependent protein kinase (cdk1) of the cell-cycle clade, three cell-cycle cyclins, a clearly recognizable anaphase promoting complex, a polo kinase, an aurora kinase , several nimA-related molecules, three wee kinases, and one cdc25 . Dictyostelium even possesses genes that are associated with the G1/S transition in higher cells. In particular, there is an orthologue of the retinoblastoma susceptibility protein, Rb, as well as the activator E2F (gene ID DDB_G0284129), the transcription factor with which Rb characteristically interacts, and the E2F-dimerization partner DP (gene ID DDB_G0276799) .
Rb has two well-known roles in animal and plant cells, one in regulating transcription at the G1/S transition, the other in facilitating terminal differentiation . The Dictyostelium Rb orthologue, rblA, is upregulated 200-fold in differentiating spores. Vegetative rblA-deficient cells are smaller than wild-type cells and they enter development prematurely. Although rblA-nulls develop normal fruiting bodies, they show a preference for the stalk pathway when mixed with wild-type cells . These results suggest that RblA plays a role during growth and late development of Dictyostelium.
In the present study, we used comparative transcriptomics to further investigate the role of RblA in Dictyostelium. Using DNA microarrays ,  to analyse the transcriptomes of wild-type and rblA disruptant cells, we found that the most significant differences occurred in S-phase specific genes. Using massively parallel mRNA sequencing (mRNA-Seq), we identified a larger group of transcripts, representing about 4% of Dictyostelium genes, that are overexpressed 2- to 80-fold in the absence of RblA function. The collection includes virtually all Dictyostelium genes with known or suspected roles in S phase or mitosis. A further experiment using synchronized wild-type cells showed that most RblA-repressed genes are maximally expressed at a single point in G2, one to two hours before mitosis. We also used an existing mRNA-Seq data set  to characterize the expression of RblA-regulated genes in the Dictyostelium multicellular stage. Virtually all RblA-repressed cell-cycle genes are upregulated strongly in mid-development; this is consistent with classical and recent reports of cell-cycle activity in Dictyostelium’s migrating slugs , , , , , . During terminal differentiation, most cell-cycle genes are turned off again while a different class of RblA targets, the RblA-activated genes, are upregulated. These results suggest that RblA is important for both the cell cycle and differentiation of Dictyostelium. Unlike the role of Rb in the G1/S phase transition of the animal cell cycle, however, Dictyostelium RblA exerts its effect in late G2 and is marked by its transcriptional regulation of both mitotic and S-phase specific genes.
Congruence of Microarray and mRNA-Seq Data
We prepared mRNA samples from growing and developing cells of a rblA-disruptant  and the parental strain, then analysed the samples using both microarrays  and mRNA-Seq. For all the conditions and strains, a total of three biological and two technical replicates were performed using DNA microarrays and two biological replicates were analysed by RNA-Seq. The number of genes showing significant differences in expression was higher in the mRNA-Seq data set than in the microarrays (367 versus 251 transcripts with p-values of <0.05). Many of the transcripts whose differential expression was judged significant in the microarrays were also strongly regulated in the mRNA-Seq data set (Figure S1). Expression changes were similar for the most abundant transcripts. Low abundance transcripts behaved heterogeneously. Expression changes were higher on the average in the mRNA-Seq data set but we also saw some genes, judged significant in the microarrays, which showed little or no regulation in the mRNA-Seq data. These genes included several whose transcripts were not detectable by mRNA-Seq. Such events are presumably microarray false-positives resulting from cross-hybridization and deficiencies in array construction. We also performed qRT-PCR to support this cross-platform validation on selected genes (Table S1).
Overall Structure of the mRNA-Seq Data
To visualize the mRNA-Seq data set as a whole, we plotted, gene by gene, the regulation factor in the rblA disruptant versus the regulation in late development (Figure 1). Most of the points fell near the origin in this plot but three separate clusters of genes were found to depart from this pattern. The largest group consists of developmentally upregulated genes not controlled by RblA (green dots). These genes have been extensively studied in Dictyostelium and are not discussed further here. A second group of transcripts (red dots) is upregulated in the disruptant, more so in developing than in growing cells. This group is referred to as the RblA-repressed genes. A third grouping (blue dots) contains genes that are upregulated in development but downregulated in the rblA mutant. We call these RblA-activated genes. The genes repressed or activated by RblA are described below.
. Scatter plot of RblA-specific regulation during development (x-axis; scale is log2) compared to developmental regulation (y-axis; scale is log2).Each point represents a gene. The red dots are genes repressed by RblA in both growth and development. The green dots are genes induced during development but not significantly regulated by RblA. The blue dots are genes whose activation depends wholly or partly on RblA.
Table 1 gives an overview of the genes that are overexpressed in the rblA disruptant and thus presumably normally repressed, directly or indirectly, by RblA. All of the RblA-regulated genes with regulation factors of 2 or greater in growing and/or developing cells are listed in Table S2. Macros are included to assist the user in graphing expression profiles and gene IDs are linked to DictyBase .
A group of genes found to be upregulated in the rblA disruptant code for proteins involved in the overall control of cell-cycle functions. These include cdk1, cks1, and aurora kinase. At least three functional groups of S-phase genes are overexpressed in the rblA disruptant. Genes whose products are directly associated with replication initiation or replication fork progression - including the ATP-binding Cdc6, members of the origin of replication (orc) complex, and all six subunits of the mini-chromosome maintenance (mcm) helicase - are upregulated in the mutant. Genes associated with the biosynthesis of deoxyribonucleotides such as the large and small subunits of ribonucleotide reductase and those involved in the disassembly/reassembly of chromatin are also upregulated in rblA-deficient cells.
Genes associated with DNA repair and rearrangement repressed by RblA include components of the DNA mismatch repair system. Genes whose products are involved in rescuing stalled replication forks by homologous recombination are also upregulated in the rblA disruptant, including the PCNA-targeting E3 ubiquitin ligase RAD18 and the recombination enzymes RAD51 and RAD52. There is however dichotomy in the role of RblA on the non-homologous end joining (NHEJ) pathway. The genes encoding members of the “X4L4” complex are RblA-repressed while several others (dnapkcs, ku70, ku80, polB, and wrn) are not regulated by RblA.
Mitotic genes are also repressed by RblA. For example, most Dictyostelium condensins are upregulated in the rblA disruptant. The annotation of this group of genes, consisting of complete sets of both type I and type II condensins, was based initially on the results of this study. Other RblA-repressed M-phase genes include those whose products are important in spindle assembly or spindle mechanics. Notably, the seven overexpressed kinesins correspond exactly to those thought to be involved in mitotic processes , ; none of the kinesins with putative functions in vesicle transport are regulated by RblA. RblA-repressed genes with probable roles in cytokinesis include the inner centromere protein INCENP and two genes coding for dynamin-like proteins . In addition, five of the ten identifiable subunits of the anaphase promoting complex are repressed by RblA.
Our data show RblA regulation of several highly conserved genes whose functions in the cell cycle are poorly understood. These include shaker, and ube2s whose expression correlates with normal or abnormal proliferation and two (wdr92 and SIVA) that are considered to be apoptotic genes in metazoans , , , , , . Dictyostelium may be useful in clarifying the functions of these genes.
Among the genes most strongly upregulated in the rblA disruptant are open reading frames (ORFs) of five transposable elements. A collective of 19 Tdd-4 elements is upregulated 25-fold. The Dictyostelium database lists two occurrences of the Tdd-5 transposon. One of these (DDB_G0290955) is upregulated in rblA-deficient cells 85-fold during growth (p-value <0.01) and 40-fold during development (p-value <0.01). The retrotransposable element Skipper and two divergent Skipper GAG fragments are also upregulated in the rblA disruptant.
Expression of rblA-repressed Genes during the Cell Cycle
Are rblA-repressed genes with putative cell-cycle functions regulated during the Dictyostelium cell cycle? To answer this question, we synchronized wild-type cells using the temperature block-and-release method . In parallel, we cultured an aliquot of the synchronized cells in the presence of bromodeoxyuridine to monitor progression through S phase (Figure 2). The number of bromodeoxyuridine-positive cells peaked 3 hours after release from the cell-cycle block. The duration of the Dictyostelium cell cycle is approximately 8 hours with a 10-minute M phase followed immediately by a 30-minute S phase . The levels of most RblA-repressed genes fluctuate during the cell cycle, reaching a maximum one hour after the cells have resumed growth. Since there is no G1 phase ,  this places their accumulation during the cell cycle in late G2 shortly before mitosis. These results agree with those obtained in RNA-blot analysis and reporter-gene studies for rnrB .
Cells were synchronized using cold arrest and released from the block at 0 h. RNA samples were collected at hourly intervals over a 12 hour period for mRNA sequencing. Cells were counted every 30 minutes (grey dashed line) and pulse-labelled every hour with bromodeoxyuridine to monitor passage through S-phase (grey line). Panel (A) shows the expression pattern of selected genes: cdk1, kif2, polA, and rblA. In panel (B), the expression profiles of genes which act at specific phases of the cell cycle are grouped. Genes with RblA-repression factors of 2 or greater whose products are known to play roles in the cell cycle, including 5 general cell cycle controllers, 41 M-phase genes, and 57 S-phase genes were averaged and graphed. Note that because of the incomplete synchronization, modulation of greater than threefold is not expected even if a gene is expressed exclusively at the G2/M transition. The time in hours (x-axis) and normalized regulation level (y-axis) are shown.
In this experiment, the rblA transcript itself behaved very much like the genes that it represses, increasing shortly before S phase. It is therefore probable that the RblA protein is inactivated at this point although this remains to be shown directly. In animal and plant cells, cyclin D initiates Rb inactivation at the G1/S transition , . Our cold-synchronization data suggest that it may play a parallel role in Dictyostelium; cycD mRNA levels increased 7.5-fold during the cold shock (p<0.01) and preceded by one hour the maximum expression of RblA-repressed genes.
Further examination of the cell-cycle data revealed a number of genes that show strong cell-cycle regulation but change little in the rblA disruptant. Some of these such as cdc25 and cycB are putative cell-cycle regulators. Others, like coronin B or the folliculin-interacting protein orthologue DDB_G0289243, code for proteins not previously recognized as being important in the cell cycle. It thus seems likely that a RblA-independent pathway for cell-cycle regulation exists.
Expression of RblA-regulated Genes in Development
Cell-cycle activity during Dictyostelium multicellular development has been reported , , , , , . We examined the developmental expression of Dictyostelium cell-cycle genes using previously published mRNA-Seq data . In this study, developing cells were harvested for mRNA-Seq analysis at 4 hour-intervals over a 24 hour period until terminal differentiation. To ensure strict comparability between datasets the raw reads were obtained (courtesy of G. Shaulsky and A. Parikh) and reanalysed using the same method used to analyse the cell-cycle dataset. Figure 3 shows the developmental profiles of cdk1, kif2, polA1, cycD, and rblA as well as the averaged profiles from 5 general cell-cycle control genes, 41 mitotic genes, 57 S-phase genes, and 19 DNA-repair genes. Most cell-cycle genes are maximally expressed in mid development at the tipped aggregate and slug stages. Like the cell-cycle experiment, the transcriptional profile of rblA is comparable to those of the genes that it represses with a peak between 8 to 16 hours of development . Taken together this suggests that RblA is inactive at the protein level in most cells during this period. Interestingly, the cycD transcript peaks shortly but significantly (p<10−44, by chi-squared comparison of profiles) before the maximal expression of RblA-repressed genes, again suggesting a role for cyclin D in RblA inactivation.
(A) Developmental regulation of the selected cell-cycle genes cdk1, kif2, polA1, cycD, and rblA. (B) Averaged transcriptional profiles during development of groups of cell-cycle genes including 5 general cell cycle controllers, 41 mitotic genes, 57 S-phase genes, and 22 DNA repair genes. These profiles were generated by reanalysing the raw data from Parikh et al . The developmental time in hours (x-axis) and normalized mRNA levels (y-axis) are shown.
In metazoans, Rb interacts with other activators of terminal-differentiation genes to facilitate differentiation and stabilize the differentiated state , . We identified a group of genes that was upregulated in development but downregulated in the rblA disruptant. These genes are listed in TableS2. Four genes activated by RblA during development code for transmembrane proteins containing an immunoglobulin E-set domain. Eight others are related to hssA (high copy suppressor of STATa). Two of these (sigN173 and sigN175) belong to a family of prestalk-specific genes transcriptionally regulated by the serum-response related factor SrfA . In contrast to the RblA-repressed genes, most of the RblA-activated genes are novel with no recognizable homologues outside the cellular slime moulds. This is not surprising since they presumably code for proteins that are responsible for the genesis of slime mould specific structures or natural products.
Figure 4 shows the average developmental expression profiles the RblA-activated genes. In wild-type cells, eight genes are expressed during terminal differentiation and 45 genes are upregulated during mid-development. In most cases the vegetative mRNA levels are too low to give usable cell-cycle profiles. An exception is the putative GATA-family transcription factor coded by comH, which is 100-fold downregulated in the rblA disruptant during growth and over 30-fold downregulated in the rblA disruptant during development. The comH transcript has an expression peak in mid-development and shows strong cell-cycle regulation, but it is expressed in early G2 in a phase that overlaps only slightly with the expression of RblA-repressed genes (Table S2).
Cells were developed for the amount of time indicated and analysed using mRNA-sequencing. The averaged-developmental profiles of 8 terminal-differentiation genes and 45 genes regulated during development by RblA, as well as the profiles for rblA, and srfA. Raw data from Parikh et al reanalysed . The developmental time in hours (x-axis) and normalized mRNA levels (y-axis) are shown.
A Dictyostelium Cell-cycle Transcriptional Network
To our knowledge, a comprehensive picture of cell-cycle transcriptional control exists only for yeast . Unfortunately most transcriptional regulators of the yeast cell cycle have no orthologues in higher eukaryotic cells. Our data suggests that the parallel between metazoans and Dictyostelium may be closer. In amoebae, we find orthologues of several histone modifiers and other regulators suspected to control the expression of metazoan cell-cycle genes (Table 1). The genes encoding these transcriptional proteins are repressed by RblA, and they are expressed during development in a fashion typical of cell-cycle genes. Prominent among them is a relative of the conserved histone lysine methyltransferase known as the Set and MYnd Domain-containing protein, SMYD3. This protein accumulates in the nucleus of S-phase and G2/M cells, and is upregulated in the majority of colorectal and hepatic cancers .
RbbD (an orthologue of RpAp48), lin9, and lin54 are putative members of the LINC complex implicated in the regulation of G2/M genes in mammals . These three genes are upregulated in the rblA disruptant. DDB_G0280079 codes for a protein similar to ADA2, a subunit of the trimeric histone-acetyltransferase complex SAGA. Levels of the ada2 transcript are 2.4-fold higher in the rblA disruptant in both growing and developing cells. Finally, DDB_G0274491, a gene similar to the budding yeast PCC1, is repressed by RblA and is cell-cycle and developmentally regulated. In yeast PCC1 recruits SAGA to specific promoters including those of genes involved in cell-cycle progression . The human orthologue of PCC1 is the cancer/testis antigen (L-antigen), a gene overexpressed in a wide variety of cancers .
The control of cell-cycle genes in Dictyostelium must involve a variety of pathways. This is certainly true in higher eukaryotes but the overall structure of the system has not been deciphered. In amoebae, where the cell-cycle transcriptional network appears to have RblA at its apex, the situation may be more approachable (Figure 5). Using genetic manipulations and mRNA-Seq technology, it may be possible to clarify the downstream network interconnections.
We propose a model based partly on our gene expression data in which the retinoblastoma homologue RblA (with E2F and DP) represses most of the important S phase and G2/M-phase genes at the transcriptional level. This repression acts directly by down-regulating activators such as cdk1, and indirectly by targeting factors that control the expression of genes required for cell-cycle progression. Among the factors repressed by RblA are the SAGA complex and the histone methyl transferase SMYD3, which are predicted to bind to the promoters of many cell-cycle genes. The LINC complex, also RblA-repressed, is implicated in the regulation of G2/M genes in mammals (see citations in the text). RblA-independent regulators, such as Cdc25, CycB, and CycD, acting with Cdk1 help to complete the network.
Previous studies on animal and plant cells over- or under-expressing Rb have consistently identified four classes of genes as targets of the Rb/E2F axis. Not surprisingly, the list includes many genes whose products are involved in DNA replication. Genes coding for chromatin-modifying proteins form a second group and DNA-repair genes a third. Finally, Rb and/or E2F have been shown to regulate a number of mitotic genes.
Most of the RblA-repressed genes that we have found in Dictyostelium fall into these four categories. However, mRNA-Seq technology allowed us to see these groups more clearly than microarray-based studies. We have recovered essentially all genes with important roles in DNA replication as well as most of the genes known or suspected to be involved in the Dictyostelium M phase. We also capture more detail in the groups of chromatin-modifying and DNA-repair molecules. The DNA-repair genes are highly heterogeneous in their response to disruption of rblA, and it is primarily the repair pathways linked to DNA replication (“methyl-directed” repair, homologous recombination, and replication fork rescue) that are overexpressed in the absence of RblA function.
In addition to these classes of RblA targets, two DNA transposons and three retrotransposons showed particularly strong overexpression in the rblA disruptant. DNA transposons increase in number primarily by transposition in S phase  while retrotransposons create double-strand breaks that must be repaired by the cellular machinery . Since transposable elements of both types rely on the DNA-replication machinery for dispersal in the genome, it makes sense that their expression should correlate with that of S-phase genes. Transposition is a major modality of mutagenesis and since Rb is inactivated, directly or indirectly, in most tumours  Rb-regulated transposon transcription may contribute to genomic instability in cancer.
RblA also appears to control cell-cycle genes during the Dictyostelium multicellular stage. Although initially interrupted, cell cycling resumes at the slug stage , , , , , . We found that cell-cycle genes are almost universally activated during an 8-hour period preceding slug formation. The upregulation levels are comparable to the overexpression levels in the rblA-null vegetative cells, suggesting that RblA is largely inactivated in most cells during mid-development.
Cell-cycle genes are, for the most part, repressed by RblA. A completely different group of genes is activated by RblA during late development. The rblA transcript itself is downregulated during this period so transcriptional regulation cannot be responsible for the increased expression of these genes. This implies that Dictyostelium RblA is activated at the protein level during late development. The set of rblA-activated genes includes two that are transcriptionally regulated by SfrA, a serum-response factor related protein . srfA is itself induced in late development but not to the same extent as its known targets. This observation suggests that the genes may be coregulated by SrfA and RblA, perhaps as the result of a direct protein-protein interaction, although a complex of this sort has not been described in vertebrate cells. Alternatively, a further component may be involved. Two Dictyostelium GATA family proteins, comH (also known as gtaB) and gtaT are activated by RblA. An interaction between Rb and a GATA-family transcription factor has been reported in mice .
The cell cycles of most eukaryotes include a prolonged G1 phase, but Amoebozoans appear to depart from this pattern. Classical studies suggest that a substantial G1 phase is lacking in both Physarum , ,  and Acanthaomeba , . Dictyostelium amoebae enter S phase directly after mitosis and often before the completion of cytokinesis. When cells are allowed to attach to a substrate and S-phase cells labelled by a short pulse of [3H]thymidine or bromodeoxyuridine, marked cells are frequently found in pairs (HKM, unpublished observations) . This is expected if cells that divide after attachment immediately begin DNA synthesis. One also occasionally sees late-telophase cells, sometimes with lagging chromosomes, in which DNA synthesis appears to have already begun , . Although one report, based largely on flow cytometry of whole cells, suggests that Dictyostelium prespore cells and spores are arrested in G1 , this was not confirmed by a study of the DNA contents of individual nuclei , .
The basis for the lack of G1 in Amoebozoans is obscure. Intriguingly, the replication-licensing factor cdt1 found in both animals and plants is missing and apparently not necessary in Dictyostelium. Dictyostelium also lacks identifiable homologues of Cdk inhibitors (CKIs) which play an important role in the G1/S transition of both fungi and metazoans . Another protein missing in Dictyostelium is p53. In vertebrates it plays an important role in cell-cycle arrest following DNA damage, which proceeds via p21 to Cdk inhibition ultimately leading to Rb repression of cell-cycle genes , . Both p21 and p53 are lacking in Dictyostelium but a DNA-damage induced cell cycle checkpoint has been described .
In some aspects, regulation of the Dictyostelium cell cycle more closely resembles regulation in plants than in metazoans. In metazoan cells, Rb inactivation at the G1/S transition is driven by a positive feedback loop involving cyclin E. Cyclin E binds to Cdk2 and the complex, in turn, phosphorylates Rb. This event frees the transcription factor E2F to activate the cycE gene. Plants do not have cyclin E. Rather the switch involves transcriptional upregulation of E2F and its dimerization partner, DP, at the G1/S transition and cdkB at G2/M , . In Dictyostelium, cdk1 and DP (tfdp2) are strongly upregulated in the rblA disruptant and both are regulated as cell-cycle genes.
In metazoans, mitotic entry is accompanied by a prominent peak in the transcription of cyclin A. In plants, both A- and D-type cyclins peak at the G2/M transition and in some cases D-type cyclins appear to be rate-limiting . In Dictyostelium both cyclins peak shortly before M phase; cycD was induced before most cell-cycle genes including cycA in our cold-synchronization experiment. In the developmental time course prepared by Parikh and colleagues , cycA and cycD show sharp peaks slightly before the maximum expression of RblA-repressed genes.
In animals, Rb is best known as a stabilizer of G1, blocking S-phase entry and supporting G1 cell-cycle exit to differentiation. There is, however, evidence for an involvement of Rb/E2F at the G2/M transition. Three studies of E2F-overexpressing mammalian cells , ,  each identified half-a-dozen mitotic genes as E2F targets, as did a survey of genes whose promoters are precipitable with antibodies to E2F1 or E2F4 . Markey et al.  found 20 mitotic genes whose activity decreased upon Rb overexpression; Jackson et al.  found overexpression of 14 mitotic genes in cells lacking the Rb-related proteins p107 and p130. In Drosophila, Dimova and colleagues  found 8 mitotic genes among Rb/E2F targets. More recently, Date et al.  showed that the mitotic regulator Borealin is a Rb target in mammalian cells. In green plants, most of the cell-cycle genes regulated by the Rb/E2F axis are associated with the G1/S transition  but the G2/M-specific cdkB is also a known E2F target .
Here we show that RblA represses virtually all S-phase genes in addition to most genes required for entry to, progression through, and exit from mitosis in Dictyostelium. We saw 57 S-phase genes and 41 mitotic genes that are strongly regulated by RblA. The degrees of repression are similar and both groups of genes are expressed at a single point in the cell cycle about two hours before entry to M.
The overall picture is one of a regulatory system similar to higher eukaryotic cells but missing a number of components known in metazoans, and reminiscent in some aspects of plants. The Dictyostelium cell cycle is thus configured as one might expect in a common ancestor of animals and plants. A recent phylogeny, incorporating new information from sponges and placozoans, places the Amoebozoan branch before the divergence of animals and plants . If correct, this raises the possibility that the G1-less cell cycle of Dictyostelium, in which RblA directs the major cell-cycle events, resembles that of a much earlier basal eukaryotic form.
Materials and Methods
Preparation of Cells
Dictyostelium discoideum strain Ax2 was used for all experiments described here. The rblA disruptant has been described previously . Vegetative cells were grown in HL5 and harvested at a density of 2×106 cells ml−1. For development cells were dispersed on Whatman 50 filters overlaid on Millipore absorbent pads moistened with LPS . Aggregates were harvested at the early culminant stage just after the spore mass lifted off the substrate and the basal disk became visible; this occurred at about 19 hours in the wild type and 16 hours in the rblA disruptant.
Cold synchronization was performed as described . Briefly, cultures between 0.5–1×106 cells ml−1 were incubated for 16 hours at 9.5°C then rapidly warmed to 22°C in a 40°C water bath. Changes in culture temperature were followed using an ethanol-washed thermometer. Cell proliferation was monitored using a Beckman-Coulter Z2 particle analyser; BrdU labelling was conducted as described , . BrdU-incorporated cells were counted automatically using a custom-written ImageJ macro.
Isolation of RNA, Microarray Analysis, Real-time PCR Validation, and Sequence Acquisition
Cells were snap-frozen in liquid nitrogen (mutant/parent comparison) or fixed in RNALater® (Ambion) at harvesting. RNA was extracted using Bio-Rad Aurum total RNA spin columns following the manufacturer’s protocol. To minimize column clogging, lysates were drawn-through a 23-gauge needle prior to loading. Samples were labelled for microarray analysis, hybridized to arrays, and data processed as described previously . Three biological and two technical replicates were conducted. The MIAME compliant array data have been deposited in ArrayExpress with accession number E-TABM-1087. Two S-phase genes found to be upregulated in the rblA disruptant and two genes that showed no difference in expression between the mutant and parental strain were selected for real-time PCR analysis (see Table S1). Primers were designed using the Primer Express Software (Applied Biosystems) and cross-checked with the Finnzymes Multiple Primer Analyzer program. cDNAs were generated from 100 nanograms of the same developmental RNA employed in the microarrays experiments using a RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas). qPCR reactions were done in triplicate on biological duplicates using SYBR Green PCR mix (Applied Biosystems) and 0.3 micromolars of each oligonucleotide. Efficiencies were determined using a pool of the mutant and wild type cDNAs on five 5-log dilutions. Dissociation curves showed the absence of non-specific products (not shown). The mRNA-Seq protocol of the Illumina/Solexa platform was used for whole transcriptome sequencing at the DNA Core Facility of the University of Missouri in Columbia or at the McGill University/Genome Quebec Innovation Centre in Montreal. Reads of 42 bp (wild type/mutant comparison) or 36 bp (cell-cycle experiment) were obtained. Two biological replicates were conducted for the wild type and rblA-null mutant. Three samples from the cell-cycle synchronization experiment (0 hr, 1 hr, and 3 hr) were sequenced twice and served as technical replicates.
Processing of Sequence Data
A transcriptome consisting of all predicted protein coding sequences was obtained from www.Dictybase.org; sequences for rRNA intersegments were added. The sequence of DDB_G0270824 which contains homologies to both tdd-4 and thugS transposable elements was split to prevent fusion of the corresponding “gene collectives” (see below).
mRNA-Seq reads were mapped to transcriptional units using software written specifically for this project. The software incorporates two features that are not yet widely used in RNA-Seq read mapping. (1) It eliminates all targets from the mapping process that are identical to other potential targets in all but one base. This obviates mismapping due to single-base sequencing errors. (2) Groups of near-identical genes are recognized automatically and organized into “collectives”. Reads that could be mapped to the collective but not to an individual gene were assigned to the “collective gene”. All reads that could be assigned uniquely to a gene within the collective continued to be so. Among other things, this procedure allowed us to monitor transposon-coded genes, some of which turned out to be very strongly regulated by RblA.
Our software uses a straightforward implementation of open hash tables for mapping. It is implemented in Java and designed so that all data structures for the Dictyostelium transcriptome fit in about 1 GB of main memory, allowing fast execution (on the order of minutes) on standard PCs.
The software executes in five phases:
- Index preparation: As a first step, the transcriptome is logically segmented into 42-base or 36-base potential mapping targets (“snippets”). A status array is constructed with one entry for each target. Snippets are represented as an integer which is their index in the transcriptome: snippet 0 starts at the beginning of the first gene, etc. In phases II-V, these indexes are placed into different hash tables. An additional array is constructed, mapping snippet indexes back to their gene numbers.
- Intra-gene uniqueness tests: within each gene, all snippets that are identical to other snippets within the same gene are marked as non-mappable within the status array. This uses a small hash table for each gene. This test eliminates the simplest sequences, such as extended trinucleotide repeats, from further processing.
- [LOOSER]Collective building: all remaining snippets are compared with each other by entering them all into a large hash table. When entering a snippet into the hash table, the algorithm detects whether an identical snippet has already been put there. After the processing of phase II, such a match necessarily must come from a different gene. The number of matches for each pair of genes is tallied in a sparse matrix. If a pair of genes contains more than 50% identical snippets and the two gene lengths differ by no more than a factor of 2, a “collective” is created for this pair of genes. The status and index arrays from phase I are extended to incorporate the new collective genes. 297 collective genes were created by the algorithm, each containing between 2 to 53 genes.
- Robust uniqueness tests: All snippets that are still marked as mappable are now compared with each other by entering them into a hash table. A snippet that occurs within two or more genes (that are not part of the same collective) is marked as non-mappable in the status array. This process is run 43 times, using 43 slightly different hash and comparison functions. First, only identical snippets are marked as non-mappable. Second, snippets that differ only in the first base are marked, then, those that differ only in the second, and so forth. The remaining snippets are considered to be “robustly unique” and can be used for mapping.
- Read mapping: all reads are mapped to the remaining snippets in the transcriptome, creating a number of matches per gene and a profile of the match over each gene’s length. The matching process uses the large hash table again. Only exact matches are counted.
The source code is available from the authors on request.
The mRNA-Seq data had been deposited in Gene Expression Omnibus  and is available through GEO series accession number GSE30368 http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE30368.
Selection of rblA Repressed Genes for Detailed Discussion
For all genes, we first calculated normalized hits in each experiment by dividing the number of mappable reads for a particular gene by the total number of fitted reads for the sample. Mutant-specific regulation factors were defined as normalized hits in the rblA disruptant divided by the normalized hits to the wild type; these were calculated separately for growing and developing cells using averages of the two biological replicates. The logarithms (base 2) of the growth phase and development-specific regulation factors were then plotted against one another. Two thirds of the points lay within a distance of 0.5 of the origin, corresponding to upregulation or downregulation of 40%. The diagram also revealed a clearly visible subgrouping of points in the northeast quadrant, corresponding to genes that are upregulated in the rblA disruptant in both growth and development. Depending on the exact definition of its boundaries, this cluster contains 3–5% of Dictyostelium transcript models. Transcripts or collectives with linear induction factors of 2 or more were selected for closer examination. For each of these, a statistical test using edgeR  was performed of the null hypothesis that the observed differences could occur by chance. Genes with adjusted p-values of <0.05 were considered to be RblA-regulated.
Regulation factors presented in this manuscript are expressed as reads per kilobase of exon model per million mappable reads . These were normalized to give a total value of 1 by adding the reads for a particular gene over the entire series (cell-cycle synchronization or developmental time course) and dividing the reads from a given time point by this sum to produce the regulation factor.
Genes selected as above were first analysed by using the gene-ontology tool GOAT. This procedure revealed little interesting substructure, probably because the cell cycle genes of Dictyostelium have not yet been systematically annotated. We then conducted BLAST and literature searches for most of the genes. Many previously unrecognized Dictyostelium cell-cycle genes were uncovered in this process (http://wiki.dictybase.org/dictywiki/index.php/MacWilliams_contributions).
Fundamental agreement between microarray and mRNA-Seq data. Compared are the data for developing cells, where the microarray signals were clearest. Normalized fold-change ratios (rblA disruptant mutant/AX2 parental strain) were compared and visualized as scatterplots. The abundance of a particular transcript measured by microarray (x-axis) compared to the abundance of the same transcript measured by mRNA-Seq (y-axis). The linear regression line is in orange and Pearson’s correlation coefficient (r) is included in the upper, left-hand corner of each graph. (A) Comparison of the 20 most abundant transcripts. (B) Comparison of 247 genes considered to be strongly differentially expressed between samples in the microarray analysis. (C) Correlation of 175 genes scored as significant (p<0.05) in the mRNA-Seq data with the corresponding microarray measurements.
qPCR validation on developmental RNA
Microsoft Excel file containing data on rblA-repressed genes as well as a number of additional genes included for comparison. This table contains raw data on regulation by RblA, developmental regulation, and regulation in the cold-synchronization experiment. Macros are included to assist the user in graphing the expression profiles and calculating error bars (the confidence limits are user-settable). Gene IDs are linked to DictyBase.
We thank Charles David, John Bonner, Richard Kessin, and John Stiller for helpful comments on the manuscript, and are especially grateful to the DictyBase team for support in all aspects of this project. HKM thanks his colleagues Angelika Boettger and Heinrich Leonhardt for encouragement and Dietfried Molter for IT expertise; he is particularly indebted to Barbara Goldstein-Müller whose intervention at a crucial juncture made it possible for him to complete this work. KS thanks Vineet Dua for preparing RNA from the cell-cycle experiment, Nadeeza Ismael for coordinating sample shipment, Marc Champagne for sharing his knowledge on qPCR, and Ian Reid for handling the sequence submission. GB wishes to thank Rob Kay and Al Ivens for their support, and Jason Skelton and the rest of the Sanger Pathogen Microarrays group for help with the array experiment.
Conceived and designed the experiments: AC AT HM. Performed the experiments: KS GB AC. Analyzed the data: GB HM KS AT. Contributed reagents/materials/analysis tools: AT HM GB AC. Wrote the paper: HM KS GB AT. Designed the software used in analysis: AM.
- 1. MacWilliams H, Gaudet P, Deichsel H, Bonfils C, Tsang A (2001) Biphasic expression of rnrB in Dictyostelium discoideum suggests a direct relationship between cell cycle control and cell differentiation. Differentiation 67: 12–24.
- 2. Weijer CJ, Duschl G, David CN (1984) A revision of the Dictyostelium discoideum cell cycle. J Cell Sci 70: 111–131.
- 3. Li H, Chen Q, Kaller M, Nellen W, Graf R, et al. (2008) Dictyostelium Aurora kinase has properties of both Aurora A and Aurora B kinases. Eukaryot Cell 7: 894–905.
- 4. MacWilliams H, Doquang K, Pedrola R, Dollman G, Grassi D, et al. (2006) A retinoblastoma ortholog controls stalk/spore preference in Dictyostelium. Development 133: 1287–1297.
- 5. Ciemerych MA, Sicinski P (2005) Cell cycle in mouse development. Oncogene 24: 2877–2898.
- 6. Urushihara H, Morio T, Saito T, Kohara Y, Koriki E, et al. (2004) Analyses of cDNAs from growth and slug stages of Dictyostelium discoideum. Nucleic Acids Res 32: 1647–1653.
- 7. Bloomfield G, Tanaka Y, Skelton J, Ivens A, Kay RR (2008) Widespread duplications in the genomes of laboratory stocks of Dictyostelium discoideum. Genome Biol 9: R75.
- 8. Parikh A, Miranda ER, Katoh-Kurasawa M, Fuller D, Rot G, et al. (2010) Conserved developmental transcriptomes in evolutionarily divergent species. Genome Biol 11: R35.
- 9. Muramoto T, Chubb JR (2008) Live imaging of the Dictyostelium cell cycle reveals widespread S phase during development, a G2 bias in spore differentiation and a premitotic checkpoint. Development 135: 1647–1657.
- 10. Bonner JT, Frascella EB (1952) Mitotic activity in relation to differentiation in the slime mold Dictyostelium discoideum. Journal of Experimental Zoology. pp. 561–572.
- 11. Zada-Hames IM, Ashworth JM (1978) The cell cycle and its relationship to development in Dictyostelium discoideum. Dev Biol 63: 307–320.
- 12. Durston AJ, Vork F (1978) The spatial pattern of DNA synthesis in Dictyostelium discoideum slugs. Exp Cell Res 115: 454–457.
- 13. Zimmerman W, Weijer CJ (1993) Analysis of cell cycle progression during the development of Dictyostelium and its relationship to differentiation. Dev Biol 160: 178–185.
- 14. Gaudet P, Fey P, Basu S, Bushmanova YA, Dodson R, et al. (2011) dictyBase update 2011: web 2.0 functionality and the initial steps towards a genome portal for the Amoebozoa. Nucleic Acids Res 39: D620–624.
- 15. Nag DK, Tikhonenko I, Soga I, Koonce MP (2008) Disruption of four kinesin genes in dictyostelium. BMC Cell Biol 9: 21.
- 16. Kollmar M, Glockner G (2003) Identification and phylogenetic analysis of Dictyostelium discoideum kinesin proteins. BMC Genomics 4: 47.
- 17. Miyagishima SY, Kuwayama H, Urushihara H, Nakanishi H (2008) Evolutionary linkage between eukaryotic cytokinesis and chloroplast division by dynamin proteins. Proc Natl Acad Sci U S A 105: 15202–15207.
- 18. White RJ (2004) RNA polymerase III transcription and cancer. Oncogene 23: 3208–3216.
- 19. Tavtigian SV, Simard J, Teng DH, Abtin V, Baumgard M, et al. (2001) A candidate prostate cancer susceptibility gene at chromosome 17p. Nat Genet 27: 172–180.
- 20. Hallows JL, Tempel BL (1998) Expression of Kv1.1, a Shaker-like potassium channel, is temporally regulated in embryonic neurons and glia. J Neurosci 18: 5682–5691.
- 21. Fortin A, MacLaurin JG, Arbour N, Cregan SP, Kushwaha N, et al. (2004) The proapoptotic gene SIVA is a direct transcriptional target for the tumor suppressors p53 and E2F1. J Biol Chem 279: 28706–28714.
- 22. Saeki M, Irie Y, Ni L, Yoshida M, Itsuki Y, et al. (2006) Monad, a WD40 repeat protein, promotes apoptosis induced by TNF-alpha. Biochem Biophys Res Commun 342: 568–572.
- 23. Jung CR, Hwang KS, Yoo J, Cho WK, Kim JM, et al. (2006) E2-EPF UCP targets pVHL for degradation and associates with tumor growth and metastasis. Nat Med 12: 809–816.
- 24. Maeda Y (1986) A new method of inducing synchronous growth of Dictyostelium discoideum cells using temperature shifts. J Gen Microbiol 1189–1196.
- 25. Lundberg AS, Weinberg RA (1998) Functional inactivation of the retinoblastoma protein requires sequential modification by at least two distinct cyclin-cdk complexes. Mol Cell Biol 18: 753–761.
- 26. de Jager SM, Scofield S, Huntley RP, Robinson AS, den Boer BG, et al. (2009) Dissecting regulatory pathways of G1/S control in Arabidopsis: common and distinct targets of CYCD3;1, E2Fa and E2Fc. Plant Mol Biol 71: 345–365.
- 27. Korenjak M, Taylor-Harding B, Binne UK, Satterlee JS, Stevaux O, et al. (2004) Native E2F/RBF complexes contain Myb-interacting proteins and repress transcription of developmentally controlled E2F target genes. Cell 119: 181–193.
- 28. Khidr L, Chen PL (2006) RB, the conductor that orchestrates life, death and differentiation. Oncogene 25: 5210–5219.
- 29. Vicente JJ, Galardi-Castilla M, Escalante R, Sastre L (2008) Structural and functional studies of a family of Dictyostelium discoideum developmentally regulated, prestalk genes coding for small proteins. BMC Microbiol 8: 1.
- 30. Simon I, Barnett J, Hannett N, Harbison CT, Rinaldi NJ, et al. (2001) Serial regulation of transcriptional regulators in the yeast cell cycle. Cell 106: 697–708.
- 31. Hamamoto R, Furukawa Y, Morita M, Iimura Y, Silva FP, et al. (2004) SMYD3 encodes a histone methyltransferase involved in the proliferation of cancer cells. Nat Cell Biol 6: 731–740.
- 32. Schmit F, Korenjak M, Mannefeld M, Schmitt K, Franke C, et al. (2007) LINC, a human complex that is related to pRB-containing complexes in invertebrates regulates the expression of G2/M genes. Cell Cycle 6: 1903–1913.
- 33. Kisseleva-Romanova E, Lopreiato R, Baudin-Baillieu A, Rousselle JC, Ilan L, et al. (2006) Yeast homolog of a cancer-testis antigen defines a new transcription complex. Embo J 25: 3576–3585.
- 34. Scanlan MJ, Gure AO, Jungbluth AA, Old LJ, Chen YT (2002) Cancer/testis antigens: an expanding family of targets for cancer immunotherapy. Immunol Rev 188: 22–32.
- 35. Feschotte C, Pritham EJ (2007) DNA transposons and the evolution of eukaryotic genomes. Annu Rev Genet 41: 331–368.
- 36. Gasior SL, Wakeman TP, Xu B, Deininger PL (2006) The human LINE-1 retrotransposon creates DNA double-strand breaks. J Mol Biol 357: 1383–1393.
- 37. Sherr CJ (1996) Cancer cell cycles. Science 274: 1672–1677.
- 38. Escalante R, Iranfar N, Sastre L, Loomis WF (2004) Identification of genes dependent on the MADS box transcription factor SrfA in Dictyostelium discoideum development. Eukaryot Cell 3: 564–566.
- 39. Kadri Z, Shimizu R, Ohneda O, Maouche-Chretien L, Gisselbrecht S, et al. (2009) Direct binding of pRb/E2F-2 to GATA-1 regulates maturation and terminal cell division during erythropoiesis. PLoS Biol 7: e1000123.
- 40. Nygaard OF, Guttes S, Rusch HP (1960) Nucleic acid metabolism in a slime mold with synchronous mitosis. Biochim Biophys Acta 38: 298–306.
- 41. Braun R, Mittermayer C, Rusch HP (1965) Sequential temporal replication of DNA in Physarum polycephalum. Proc Natl Acad Sci U S A 53: 924–931.
- 42. Kessler D (1967) Nucleic acid synthesis during and after mitosis in the slime mold, Physarum polycephalum. Exp Cell Res 45: 676–680.
- 43. Band RN, Mohrlok S (1973) The cell cycle and induced amitosis in Acanthamoeba. J Protozool 20: 654–657.
- 44. Jantzen H, Schulze I, Stohr M (1988) Relationship between the timing of DNA replication and the developmental competence in Acanthamoeba castellanii. J Cell Sci 91 (Pt 3): 389–399.
- 45. Zada-Hames IM, Ashworth JM (1978) The cell cycle during the vegetative stage of Dictyostelium discoideum and its response to temperature change. J Cell Sci 32: 1–20.
- 46. Chen G, Shaulsky G, Kuspa A (2004) Tissue-specific G1-phase cell-cycle arrest prior to terminal differentiation in Dictyostelium. Development 131: 2619–2630.
- 47. Sherr CJ, Roberts JM (1999) CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev 13: 1501–1512.
- 48. He G, Siddik ZH, Huang Z, Wang R, Koomen J, et al. (2005) Induction of p21 by p53 following DNA damage inhibits both Cdk4 and Cdk2 activities. Oncogene 24: 2929–2943.
- 49. Jackson MW, Agarwal MK, Yang J, Bruss P, Uchiumi T, et al. (2005) p130/p107/p105Rb-dependent transcriptional repression during DNA-damage-induced cell-cycle exit at G2. J Cell Sci 118: 1821–1832.
- 50. Magyar Z, Atanassova A, De Veylder L, Rombauts S, Inze D (2000) Characterization of two distinct DP-related genes from Arabidopsis thaliana. FEBS Lett 486: 79–87.
- 51. Inze D, De Veylder L (2006) Cell cycle regulation in plant development. Annu Rev Genet 40: 77–105.
- 52. Muller H, Bracken AP, Vernell R, Moroni MC, Christians F, et al. (2001) E2Fs regulate the expression of genes involved in differentiation, development, proliferation, and apoptosis. Genes Dev 15: 267–285.
- 53. Ishida S, Huang E, Zuzan H, Spang R, Leone G, et al. (2001) Role for E2F in control of both DNA replication and mitotic functions as revealed from DNA microarray analysis. Mol Cell Biol 21: 4684–4699.
- 54. Polager S, Kalma Y, Berkovich E, Ginsberg D (2002) E2Fs up-regulate expression of genes involved in DNA replication, DNA repair and mitosis. Oncogene 21: 437–446.
- 55. Ren B, Cam H, Takahashi Y, Volkert T, Terragni J, et al. (2002) E2F integrates cell cycle progression with DNA repair, replication, and G(2)/M checkpoints. Genes Dev 16: 245–256.
- 56. Markey MP, Angus SP, Strobeck MW, Williams SL, Gunawardena RW, et al. (2002) Unbiased analysis of RB-mediated transcriptional repression identifies novel targets and distinctions from E2F action. Cancer Res 62: 6587–6597.
- 57. Dimova DK, Stevaux O, Frolov MV, Dyson NJ (2003) Cell cycle-dependent and cell cycle-independent control of transcription by the Drosophila E2F/RB pathway. Genes Dev 17: 2308–2320.
- 58. Date DA, Jacob CJ, Bekier ME, Stiff AC, Jackson MW, et al. (2007) Borealin is repressed in response to p53/Rb signaling. Cell Biol Int 31: 1470–1481.
- 59. Ramirez-Parra E, Frundt C, Gutierrez C (2003) A genome-wide identification of E2F-regulated genes in Arabidopsis. Plant J 33: 801–811.
- 60. Srivastava M, Simakov O, Chapman J, Fahey B, Gauthier ME, et al. (2010) The Amphimedon queenslandica genome and the evolution of animal complexity. Nature 466: 720–726.
- 61. Sussman M (1987) Cultivation and synchronous morphogenesis of Dictyostelium under controlled experimental conditions. Methods Cell Biol 28: 9–29.
- 62. Edgar R, Domrachev M, Lash AE (2002) Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 30: 207–210.
- 63. Robinson MD, McCarthy DJ, Smyth GK (2010) edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26: 139–140.
- 64. Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B (2008) Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods 5: 621–628.