Runx-dependent expression of PKC is critical for cell survival in the sea urchin embryo

Background Runx transcription factors play critical roles in the developmental control of cell fate and contribute variously as oncoproteins and tumor suppressors to leukemia and other cancers. To discover fundamental Runx functions in the cell biology of animal development, we have employed morpholino antisense-mediated knockdown of the sea urchin Runx protein SpRunt-1. Previously we showed that embryos depleted of SpRunt-1 arrest development at early gastrula stage and underexpress the conventional protein kinase C SpPKC1. Results We report here that SpRunt-1 deficiency leads to ectopic cell proliferation and extensive apoptosis. Suppression of the apoptosis by pharmacological inhibition of caspase-3 prevents the ectopic proliferation and rescues gastrulation, indicating that many of the overt defects obtained by knockdown of SpRunt-1 are secondary to the apoptosis. Inhibition or knockdown of SpPKC1 also causes apoptosis, while cell survival is rescued in SpRunt-1 morphant embryos coinjected with SpPKC1 mRNA, suggesting that the apoptosis associated with SpRunt-1 deficiency is caused by the deficit in SpPKC1 expression. Chromatin immunoprecipitation indicates that SpRunt-1 interacts physically with SpPKC1 in vivo, and cis-regulatory analysis shows that this interaction activates SpPKC1 transcription. Conclusions Our results show that Runx-dependent activation of SpPKC1 is essential for maintaining protein kinase C activity at levels conducive to cell survival during embryogenesis.


Background
The Runt domain (Runx) is a highly conserved 128 amino acid sequence that defines a metazoan family of sequencespecific DNA binding proteins important for the developmental control of cell fate [1]. Vertebrates have three Runx genes that are essential for the development of specific tissues (blood, bone, and proprioceptive neurons of the dor-sal root ganglia, respectively, for Runx1, Runx2, and Runx3), each of which is associated with leukemogenic or carcinogenic mutations [1][2][3][4]. Some of these mutations cause loss of Runx function while others are gain-of-function; hence, mammalian Runx genes display attributes of both tumor suppressors and proto-oncogenes [4]. Runx proteins have been shown to regulate cell cycle transit via their function as context-dependent transcriptional regulators. For example, Runx1 stimulates the G1 to S phase transition in cultured mammalian cells, which may in part reflect its transcriptional activation of cyclin D3 [5][6][7]. Runx1 is also an transactivator of the p14 ARF tumor suppressor [8], while both Runx1 and Runx2 are transcriptional repressors of the p21 Cip1/Waf1 cyclin-dependent kinase inhibitor [9,10]. These findings indicate that Runx is a critical node in the gene regulatory network that controls cell proliferation and/or survival during animal development. Since many of the core linkages in this network are likely to be conserved throughout animal evolution, identification and cis-regulatory analysis of relevant Runx target genes in experimentally tractable animal models should help illuminate how Runx genes contribute to human development and health.
The sea urchin embryo is optically accessible and highly amenable to experimental manipulations such as gene transfer and morpholino antisense mediated knockdown, making it a useful system for discovering fundamental roles played by Runx proteins in the cell biology of animal development. Whole genome sequence data [11] as well as genomic Southern blot analysis [12] suggest that, contrary to previous reports [13][14][15], the sea urchin Strongylocentrotus purpuratus has two Runx genes (A. Poustka, personal communication; AJR and JAC, unpublished results). Only one of these (SpRunt-1, previously referred to as SpRunt) is expressed in the embryo [12,14]. SpRunt-1 is required for embryogenesis beyond blastula stage [16] and contributes to transcriptional activation of the CyIIIa actin gene, a marker of aboral ectoderm differentiation [12]. SpRunt-1 mRNA is expressed globally in the early embryo, whereas in the larva it accumulates predominantly in oral ectoderm and endomesoderm, tissues associated with continued growth and cell proliferation [14]. Embryos depleted of SpRunt-1 by morpholino antisense oligonucleotides (MASOs) that sequence-specifically block either translation or pre-mRNA splicing fail to gastrulate normally, display global defects in cell differentiation and proliferation and underexpress a number of genes including that encoding the conventional protein kinase C SpPKC1 [16].
Here we provide evidence that the gastrulation defect and abnormal proliferation associated with SpRunt-1 deficiency are secondary effects of extensive apoptosis that occurs in the post-blastula stage embryo. We further demonstrate that apoptosis is caused by a deficit in SpPKC1 activity, and that cell survival in SpRunt-1 morphant embryos can be rescued by exogenous SpPKC1 mRNA. Finally, we show that the SpPKC1 promoter region contains SpRunt-1 target sequences that are required for its transcriptional activation.

SpRunt-1 deficiency leads to apoptosis and secondary proliferation
To characterize further the abnormal cell proliferation associated with SpRunt-1 deficiency [16], gastrula stage embryos injected with either a nonspecific control MASO or the splice-blocking SpRunt-1 MASO were pulse-labelled with bromodeoxyuridine (BrdU). The embryos were fixed, and an anti-BrdU antibody was used to detect BrdU incorporation, which is indicative of DNA synthesis during the labelling period. Normal late gastrula stage embryos ( Figure 1A) contained very few cells in S phase, as shown by an absence of nuclear BrdU signal in the controls ( Figure 1C). In contrast, SpRunt-1 morphant embryos of the same age ( Figure 1B) displayed extensive nuclear BrdU incorporation throughout the embryo (Figure 1D; 11/14 imaged embryos with phenotypes similar to that shown), indicating an abnormally high number of proliferating cells, in agreement with our previous cytometric DNA analysis of cell cycle distribution in these embryos [16].
Paradoxically, SpRunt-1 morphant embryos had about half as much DNA per embryo as gastrula stage controls of the same age [16]. Since deregulated cell proliferation is often associated with programmed cell death [17][18][19], we asked whether cells in SpRunt-1 morphants undergo apoptosis, which could account for these embryos displaying lower than normal DNA content despite having ectopic cell cycles. Terminal transferase-mediated dUTP nick end labeling (TUNEL) showed that, in contrast to controls (Fig. 1E), SpRunt-1 morphants displayed extensive DNA fragmentation (Figures 1F, 2B; 20/23 imaged embryos with phenotypes similar to that shown), a hallmark of apoptosis. Apoptosis was further indicated by the observation that SpRunt-1 morphants contained elevated levels of activated caspase-3 ( Figure 1H; compare with control in Figure 1G) and caspase-6 (data not shown). The apoptosis was specifically caused by SpRunt-1 deficiency, as cell survival is rescued by coinjection of full-length SpRunt-1 mRNA with the splice-blocking SpRunt-1 MASO ( Figure 2C). Interestingly, RNAi-mediated knockdown of the Drosophila Runx gene CG15455 was recently shown to lead to the apoptosis of in vitro cultured cells [20]; combined with our data this suggests an evolutionarily fundamental role for Runx genes in the control of metazoan cell survival.
Ectopic cell cycles in the Drosophila eye disc lead to compensatory apoptosis via Jun kinase (JNK) signaling [19]. We hypothesized that apoptosis in SpRunt-1 morphant embryos might similarly be a secondary response to unregulated cell proliferation, in which case inhibiting apoptosis might allow unbridled cell proliferation leading to a tumor-like overgrowth phenotype. To test this we Morpholino antisense-mediated knockdown of SpRunt-1 leads to ectopic proliferation and apoptosis in the gastrula stage (48 h) embryo . Surprisingly, this drug rescued gastrulation (but not other aspects of development such as skeletogenesis) in about half of the SpRunt-1 morphants, allowing development of more normal looking embryos with fully invaginated guts ( Figure 2D and Table 1). The rescued embryos displayed negligible apoptosis when assayed by TUNEL ( Figure 2E), confirming the drug's efficacy in inhibiting caspase activity and suggesting that apoptosis is a principal cause of the gastrulation defect associated with SpRunt-1 deficiency. Apoptosis is also likely to be the underlying cause of the ectopic proliferation observed in these embryos (rather than the converse), as SpRunt-1 morphants rescued with the caspase-3 inhibitor displayed negligible BrdU incorporation at late gastrula stage, whereas SpRunt-1 morphants treated with the DNA polymerase inhibitor aphidicolin (which blocks cell proliferation while allowing for normal differentiation and morphogenesis from the mesenchyme blastula stage on [21]) were not rescued from apoptosis [see Additional file 1]. It is likely that the ectopic cell cycles associated with SpRunt-1 deficiency are a regulative response to a loss of embryo integrity, as occurs when cells from dissociated gastrula stage embryos re-enter the cell cycle [22].

Apoptosis associated with SpRunt-1 deficiency is caused by loss of PKC expression
Protein kinase C (PKC) is a key component of numerous developmental signalling pathways, and the conventional PKC SpPKC1 was previously shown to be one of a number of genes that are significantly underexpressed in SpRunt-1 morphant embryos [16]. Moreover, mammalian PKCα promotes cell survival [23]. To ask whether the apoptosis associated with SpRunt-1 deficiency might be caused by the loss of PKC expression, we made use of chelerythrine, a compound that selectively and completely inhibits mammalian PKC activity at a concentration of ~10 µM [24]. Embryos treated with 2-4 µM chelerythrine from the mesenchyme blastula stage on (see Methods) displayed gastrulation defects ( Figure 3B) as well as extensive apoptosis as indicated by TUNEL ( Figure 3G) and elevated lev-els of activated caspase-3 ( Figure 3L). A similar gastrulation-defective phenotype ( Figure 3D; 80/101 embryos) and cell survival defect ( Figure 3I, N) was obtained by injecting embryos with a MASO that targeted the translation initiation codon of SpPKC1 mRNA. In both cases cell survival (but not fully normal development) was rescued by treatment with the caspase-3 inhibitor ( Figure 3C, H, E and 3J) despite the continued presence of active caspase-3 ( Figure 3M, O; note that the inhibitor blocks the downstream activity of the activated caspase).
These data strongly suggest that the deficit in SpPKC1 expression in SpRunt-1 morphants might be the underlying cause of their apoptotic phenotype, in which case restoring SpPKC1 mRNA in these embryos should rescue cell survival. To test this, we coinjected zygotes with the SpRunt-1 MASO along with synthetic capped mRNA encoding SpPKC1. Unlike embryos injected with the SpRunt-1 MASO alone, most of the morphants coinjected with solutions containing 400-800 ng/µl of SpPKC1 mRNA gastrulated ( Figure 2F and Table 2) and displayed negligible TUNEL signal ( Figure 2G), indicating that underexpression of PKC is indeed the principal underlying cause of the apoptosis and resulting gastrulation defects associated with SpRunt-1 deficiency.

SpRunt-1 target sequences in the SpPKC1 promoter region are required for transcriptional activation
Our findings raised the question of whether the requirement for SpRunt-1 in support of PKC expression is direct or indirect. Examination of the 5' flanking sequence of the SpPKC1 gene revealed the presence of two sequences that match the Runx target consensus within 1.7 kb of the first exon ( Figure 4A and [see Additional file 2]). Fluorescently-labeled oligonucleotide probes representing each of these sites bound sequence-specifically to recombinant SpRunt-1 in an electrophoretic mobility shift assay (EMSA; Figure 4B). Furthermore, sequences from the SpPKC1 promoter region were specifically recovered by chromatin immunoprecipitation (ChIP) using an anti-  Figure 2A), rescued gastrulation (with guts but no skeletons, a phenotype similar to that depicted in Figure 2D), or normal (indistinguishable from uninjected control embryos with skeletons). The latter group probably represents embryos that were under injected; note that this particular experiment had an unusually high proportion of such embryos (which more typically are in the range of 10-15%).   IgG αSpRunt-1 ChIP:

Relative Expression
SpRunt-1 antibody [12], indicating that SpRunt-1 is bound in vivo to these sequences at gastrula stage ( Figure  4C). To test the functionality of the Runx target sequences in this region we constructed a luciferase reporter gene linked to ~1.9 kb 5' flanking sequence (including the endogenous promoter) of SpPKC1 ( Figure 4A and [see Additional file 2]). When introduced into embryos, this sequence activated transcription of the reporter by approximately threefold and in a SpRunt-1 dependent fashion, as reporter expression was reduced to near baseline levels in SpRunt-1 morphant embryos ( Figure 4D). Moreover, base substitutions that eliminate either the proximal or the distal Runx target sequences in the SpPKC1 cis-regulatory domain similarly reduced activity of the reporter, as did combined substitutions that eliminated both target sequences ( Figure 4D), suggesting that the two Runx sites function synergistically rather than additively to activate transcription.
The relatively weak activation of the reporter conferred by the 1.9 kb SpPKC1 promoter fragment, together with our previous results showing that endogenous SpPKC1 expression is reduced approximately seven fold in SpRunt-1 morphant embryos [16], suggests that there may be additional enhancers outside of the 1.9 kb fragment that collaborate with SpRunt-1 in activating transcription of the endogenous gene. Testing this possibility would require cis-regulatory analysis using a larger genomic fragment that contains the SpPKC1 locus. Nevertheless, the data presented here support the proposition that the direct interaction of SpRunt-1 with its target sequences in the SpPKC1 cis-regulatory domain is required for normal levels of SpPKC1 transcription in the post-blastula stage embryo, which in turn is required to maintain PKC activity at levels sufficient to antagonize caspase activity and suppress apoptosis ( Figure 4E).

Conclusions
Our results support the idea that, in animal development, cells do not survive by default but rather by actively suppressing apoptosis in response to permissive cell signalling [25,26]. Furthermore, we find that deregulated apoptosis can lead to secondary cell proliferation, similar to the compensatory cell proliferation described by others [27,28]. In contrast to the latter phenomenon however, we do not observe overproliferation caused by apoptotic cell signalling in the caspase-3 inhibited embryos. While the specific role of SpPKC1 in suppressing apoptosis is not known, there are several plausible pathways based on well-known functions of mammalian PKCα, most notably its role in phosphorylating and thereby activating the anti-apoptotic protein Bcl-2. [29][30][31]. Identification of the relevant substrate(s) and pathway(s) in sea urchins awaits further analysis, as does the basis for the secondary proliferation observed in SpRunt-1 morphants.
In addition to the conventional PKCs (cPKCs), the PKC family also includes non-conventional and atypical PKCs (nPKCs and aPKCs, respectively) [32]. We do not yet know how many PKC family members are in S. purpuratus, although available expressed sequence tag (EST) and genomic sequence data suggest that SpPKC1 is the only cPKC gene expressed in the embryo, and possibly the only cPKC gene encoded in the genome (JAC, unpublished results). We note in closing that mammals have three cPKC genes, PKCα, PKCβ, and PKCγ. PKCβ is actually proapoptotic in myeloid cells, and its gene was recently shown to be directly activated by a human Runx homolog, RUNX1 [33]. This suggests that Runx-mediated control of cPKC gene expression may be a primitive conserved character (i.e., plesiotypic) in animals, whereas the functional output of this interaction (cell death or survival) varies with context, something that would bear on whether a given Runx gene is an oncogene or tumor suppressor. It will therefore be interesting to determine whether the other mammalian cPKC genes are also direct Runx regulatory targets, and whether the Runx-dependency of cell survival in Drosophila [20] is a function of PKC.

Animals and gametes
Adult Strongylocentrotus purpuratus were purchased from either Charles Hollahan (Santa Barbara, California, USA) or Pat Leahy (Corona del Mar, California, USA). Gametes were obtained by vigorous shaking of animals or by intracoelomic injection of 0.5 M KCl.

Microinjection and embryo culture
Microinjection of zygotes was carried out by the method of McMahon et al. [34], essentially as described in Kirchhamer et al. [35]. MASO injection solutions contained 500 µM MASO and 120 mM KCl. Embryos were cultured in artificial seawater at 15°C.

BrdU labeling and antibody staining
Late gastrula stage embryos (44 h) were incubated in 300 µg/ml BrdU (Sigma-Aldrich; St. Louis, Missouri, USA) for four hours in artificial seawater (ASW) at 15°C then fixed in 4% formaldehyde in ASW for 90 min on ice. After the embryos were washed three times in ASW, they were treated with 2 M HCl on ice for 30 min, neutralized in 100 mM Tris, pH 9.5, 100 mM NaCl for 2 min on ice, and washed three times in phosphate buffered saline containing 0.2% Tween-20 (PBST). The embryos were post-fixed in 4% formaldehyde in PBST for 20 min on ice, and washed three times in ice-cold PBST. BrdU was detected by staining with an AlexaFluor 647-conjugated anti-BrdU antibody (Molecular Probes, Eugene, Oregon, USA) diluted 1:20 in PBST and incubated overnight at 4°C. Digital images were collected using a Leica TCS SP2 confocal microscope.
Antibodies directed against activated caspases 3 and 6 were obtained from Calbiochem (San Diego, California, USA). Embryos were stained with 5 µg/ml antibody in PBST with 5% bovine serum albumin (BSA) overnight at 4°C and then washed three times in PBST. Antibody labeling was detected by staining with AlexaFluor 488 goat anti-mouse IgG from Molecular Probes in PBST with 5% BSA for 1 h at room temperature. Following three washes in PBST and nucleic acid staining with 4',6-diamidino-2phenylindole (DAPI), the digital images were obtained with the Leica TCS SP2 confocal microscope.

Terminal transferase-mediated dUTP nick end labeling (TUNEL) assay
Late gastrula stage embryos (48 h) were fixed for 1 h in 4% paraformaldehyde at 4°C and washed for 1 h in PBST. Staining was performed using the In Situ Fluorescein Detection of Cell Death Kit from Roche (Indianapolis, Indiana, USA). Stained embryos were washed three times in PBST and labeled once in DAPI. Digital images were collected using a Leica TCS SP2 confocal microscope.

DNA quantitation by fluorometry
The DNA content of isolated nuclei was quantitated by fluorometry as described previously [16].

Drug treatments
The caspase-3 selective inhibitor Ac-DEVD-CHO (BD Biosciences; San Jose, California, USA) was added to SpRunt-1 morphant cultures at a final concentration of 50 µM at the 1-2 cell stage.
The PKC inhibitor chelerythrine (Sigma-Aldrich) was added to embryo cultures at concentrations ranging from 2-4 µM from the mesenchyme blastula stage on. Higher concentrations tended to be embryonic lethal whereas lower ones gave no phenotype at all. Since the efficacy of the drug in producing the gastrulation-defective phenotype varied from culture to culture, the optimal concentration for producing the gastrulation-defective phenotype at a frequency close to 100% was determined empirically for each experiment. For example, in one experiment, 85% of embryos treated with 3.5 µM chelerythrine were viable but gastrulation defective (the remaining 15% appearing normal) whereas 100% of embryos treated with 4 µM chelerythrine were viable but gastrulation defective. For other cultures the effective dose for a similar effect was somewhat lower.

Synthetic mRNAs
Full-length SpRunt-1 mRNA for rescue experiments was synthesized as previously described [16]. SpPKC1 mRNA for rescue experiments was obtained as follows: an EST (National Center of Biotechnology Information (NCBI) accession number CD290393) containing the SpPKC1 start codon was obtained by BLAST using the Genbank entry of Lytechinus pictus PKC1 (Genbank:U02967; [36]) as a query sequence. The SpPKC1 coding sequence was then assembled from the S. purpuratus genome traces using the LpPKC1 sequence as a scaffold. The assembled SpPKC1 sequence was used to design the following primers: GTGGATAATTTTTGCTGGATTTGCGCCTTGA, (forward; ends 13 bp upstream of start codon) and GTTCTACAGACAGACTGGCACTAGCT (reverse; reverse complement of stop codon underlined). These primers were used in a reverse transcription (RT)-PCR reaction using the Platinum Quantitative RT-PCR ThermoScript One-Step System from Invitrogen (Carlsbad, California, USA) and 200 ng total RNA (purified using the RNeasy Mini Kit from Qiagen; Valencia, California, USA) from 60 h embryos. An amplicon of the expected size was gel purified and inserted to pGEM-T Easy (Promega; Madison, Wisconsin, USA). Sequencing was performed to determine the insert orientation and also to confirm the identity of the PKC1 sequence. An Apa1-Spe1 restriction fragment containing the full length SpPKC1 sequence with some flanking sequence was then transferred into a pBluescript (Stratagene; LaJolla, California, USA) expression construct containing the Xenopus BetaGlobin 3' UTR and polyA sequence (pGloBlu) [16]. Pst1 was used to linearize the resulting construct for in vitro transcription of SpPKC1 using the T7 mMessage Machine from Ambion (Austin, Texas, USA).

Assembly of the SpPKC1 5' flanking sequence
A contiguous sequence of over 2,500 bases extending in the upstream direction relative to the start codon of SpPKC1 was assembled from trace sequences obtained from the Sea Urchin Genome Project [11] using the Con-tigExpress program in the InforMax Vector NTI suite (Invitrogen). Genotrace [37], which supports the reassembly of short spans of genomic DNA sequence from trace files, was used to generate independently a contig for comparison with the manually assembled one in ContigExpress. assembled the SpPKC1 promoter and obtained sequence from a cloned PCR amplicon from sea urchin genome data and assisted with ChIP. DER performed ChIP and EMSA with assistance from JJM, who developed the ChIP protocol. JAC directed the research, assisted with some of the microinjections, assisted with microscopy and data analysis and drafted the manuscript and figures.