Previous studies summarized above point to the c-Src protein-tyrosine kinase as an important regulator of the earliest stages of mES cell differentiation. These findings led to the question of the signaling pathways controlled by c-Src that account for its role in ES cell fate. To address this question, c-Src target protein capture experiments were performed using an immobilized, recombinant c-Src SH3 domain fusion protein and soluble protein extracts from both self-renewing mES cells as well as differentiated EBs. SH3 domains contribute not only to SFK regulation (see Introduction
) but also to substrate recruitment by binding to target proteins containing polyproline type II helices 
. Unique SH3-interacting proteins were captured by the c-Src SH3 domain but not by an inactive mutant control domain, indicative of specific binding (). Three prominent bands were excised from the gel, digested with trypsin, and identified by MALDI-TOF MS and MS/MS sequencing: 1) Dynamin II, a GTPase involved in vesicular trafficking 
; 2) hnRNPK, which regulates transcription, pre-mRNA processing, mRNA transport and translation 
; and 3) a 54 kDa N-terminal fragment of Tim. Both Dynamin II and hnRNPK have been identified previously as c-Src SH3-binding proteins 
, validating our experimental approach. In contrast, the SH3-dependent association of Tim with c-Src or other SFKs has not been reported, suggestive of a novel interaction. To confirm that Tim is an SH3-binding partner for c-Src, SH3 capture experiments were repeated using lysates from ES cells and EBs, followed by immunoblotting with an antibody to the Tim protein. Full-length Tim was captured in each case, as well as a prominent cleavage product that corresponds in size to the fragment originally identified by tryptic fingerprinting (). In contrast, no binding was detected with the inactive mutant of the c-Src SH3 domain or with GST alone, indicative of a specific SH3-mediated binding event.
Identification of Src SH3-binding proteins in ES cells and embryoid bodies (EBs).
Tim is a c-Src SH3 domain binding protein and substrate.
We next investigated whether Tim is a substrate for c-Src. For these experiments, full-length Tim was co-expressed with wild-type c-Src or a kinase-defective mutant in the human epithelial cell line, 293T. Tim was then immunoprecipitated from transfected cell lysates, and analyzed for tyrosine phosphorylation by immunoblotting with anti-phosphotyrosine antibodies. As shown in , Tim was strongly phosphorylated in the presence of active c-Src but not with a kinase-dead mutant, demonstrating that Tim is a Src substrate. To investigate a possible relationship between Tim tyrosine phosphorylation and ubiquitylation, as observed previously in Drosophila 
, aliquots of the same Tim immunoprecipitates were also immunoblotted with ubiquitin antibodies (). Co-expression with c-Src strongly enhanced Tim ubiquitylation, suggesting that Src-mediated Tim phosphorylation enhances this process. Interestingly, co-expression of Tim with kinase-inactive c-Src suppressed Tim ubiquitylation below control levels observed in the absence of c-Src expression, suggestive of a dominant negative effect ().
To investigate the role of Tim in mES cells, we first over-expressed the protein and observed a reduction in cell viability and enhanced apoptosis compared to untransfected control cells (data not shown). Because of the negative impact of Timeless expression on ES cell viability, we turned to the complementary approach of gene silencing. In these experiments, Tim expression was knocked down by lentiviral transduction of shRNAs targeting two distinct regions of the Tim transcript. Both lentiviral vectors yielded ES cell populations with substantial reductions in endogenous Tim protein levels (data not shown). Six Tim knockdown cell lines were subsequently cloned from each shRNA-transduced mES cell population, and screened for the extent of Tim knockdown by immunoblot analysis (). The two cell lines exhibiting the greatest extent of full-length Tim knockdown without changes in undifferentiated colony morphology were selected for further analysis. These lines were designated as lenti:87-22 and lenti:89-18.
Generation of Tim knockdown ES cell lines.
Morphologically, self-renewing cultures of both Tim-knockdown ES cell lines exhibited less spontaneous differentiation when compared to wild-type ES cells ( & ). To investigate whether the loss of Tim influenced the expression of self-renewal markers, we examined the levels of Oct4, Sox2, Nanog, and KLF4 by quantitative immunoblotting. As shown in , both Tim-knockdown lines exhibited modest increases in the expression of Oct4 and Sox2, while levels of Nanog and KLF4 were essentially unchanged. Conversely, expression of the differentiation marker AFP was decreased by more than 60% relative to control ES cells. These changes most likely reflect loss of spontaneous differentiation as a consequence of Tim knockdown, although an influence on the ES cell self-renewal program cannot be ruled out.
Knockdown of Tim suppresses spontaneous differentiation of mES cells.
To determine whether the presence of Tim is essential for early development, we next tested the ability of the Tim knockdown ES cell lines to form EBs. In the EB assay, ES cells are plated in suspension in the absence of LIF, the cytokine required for the maintenance of mES cell pluripotency. Under these conditions, the cells differentiate into organized cysts that recapitulate the initial stages of pre-implantation development, including formation of an endodermal surface layer, differentiation of columnar epithelium, and hollowing out of a central cavity via apoptosis 
. While wild-type ES cells differentiated into a typical heterogenous EB population with respect to size, both Tim knockdown lines produced smaller EBs of more uniform size (). Under normal conditions, cystic EBs undergo expansion as they differentiate. Thus the consistent formation of smaller EBs from the Tim knockdown ES cells suggested a failure to expand and a possible differentiation defect. To address this issue, we harvested control and Tim knockdown EBs after six days of development and imaged them for cavitation by confocal microscopy. As shown in , the number of EBs exhibiting cavity formation was substantially reduced in both Tim knockdown ES cell lines. This result may help to explain the early embryonic lethality previously observed in Tim knockout mice. At ED 7.5, homozygous Tim knockout embryos lack cellular organization, with necrotic cell debris filling the amniotic cavity 
. As cavitation is a prerequisite for gastrulation, the failure to cavitate resulting from loss of Tim may prevent subsequent development as well.
Tim knockdown EBs are smaller and of more uniform size.
Tim knockdown prevents EB cavitation.
To determine if loss of cavity formation was linked to an apoptotic defect, Caspase 3/7 activity was determined in wild-type and Tim knockdown EBs following induction with staurosporine (). EBs generated from both of the Tim knockdown mES cell lines exhibited a markedly blunted response to staurosporine in terms of caspase activity as well as production of the cleaved, active form of Caspase 3 (). Interestingly, Caspase activity has been linked to the cleavage of Nanog, one of the core transcription factors known to regulate self-renewal 
. ES cells lacking the Casp3
gene are defective for differentiation, which can be rescued by expression of a cleavage-resistant Nanog variant.
Timeless-knockdown EBs are resistant to apoptosis.
We next investigated whether the cells present in the failed cavity retained characteristics of pluripotent mES cells. Wild-type and Tim knockdown EBs were fixed and immunostained for the pluripotency marker Oct4. In addition, the EBs were immunostained for the zeta isoform of protein kinase C, which is expressed in the tight junctions of the outer visceral endoderm layer and thus defines the outer edge of the EB 
. Cells remaining in the centers of the Tim knockdown EBs exhibited strong staining for Oct4 after six days, suggesting that the cavity remains filled with undifferentiated cells (). After 12 days, EBs derived from control ES cells showed little Oct4 staining, and 2D projections of the confocal Images revealed substantial cavitation (). In contrast, 12 day EBs from both Tim knockdown ES cell lines retained very thick walls of Oct4-positive cells, with little to no cavitation evident.
EBs formed from Tim knockdown cells retain pluripotent cells.
To test whether the Oct4-positive cells present in the Tim knockdown EBs remain pluripotent, we assayed for secondary EB formation. This assay allows for a quantitative comparison of the number of pluripotent ES cells remaining in each culture of primary EBs, as only these cells can give rise to secondary EBs 
. Six-day EBs from control and Tim knockdown ES cell lines were trypsinized to single cells, plated in methylcellulose and the number of secondary EBs counted 10 days later. EBs from Tim knockdown cells produced substantially more secondary EBs compared to control EBs (), providing strong evidence that the cells present in the failed cavity remain undifferentiated. These data support a model in which Tim is required for cavitation and removal of pluripotent cells from the developing EB; without Tim, cavitation and subsequent development are arrested.
In a final series of experiments, we investigated whether Tim protein levels varied as a function of EB formation. Lysates were prepared from self-renewing ES cells as well as 3, 6 and 12 day EBs, and aliquots were analyzed for Tim protein levels by immunoblotting. Overall Tim levels began to decrease after 6 days of EB formation, and were dramatically reduced after 12 days (). The timing of Tim protein loss correlates with cavity formation, which is clearly evident after 6 days and complete after 12 days ( and ) 
. We also observed that tyrosine-phosphorylated Tim is present in ES cells as well as 3 day EBs, and diminishes as Tim protein levels decrease (). Treatment of ES cells with two inhibitors previously shown to block all Src-family kinase activity in ES cells (PP2 and SKI-1) 
substantially reduced endogenous Tim tyrosine phosphorylation, strongly implicating c-Src or another member of the Src-kinase family as the kinase responsible for Tim phosphorylation (). This result is consistent with the tyrosine phosphorylation of Tim following co-expression with active c-Src in 293T cells as shown in .
Changes in Tim protein levels and tyrosine phosphorylation during EB formation.
Data presented here provide the first evidence that the circadian rhythm protein Tim also has a central role in one of the first morphogenic changes that accompanies early development. Knockdown of Tim results in ES cells that are capable of forming EB-like structures, but these EBs fail to expand or cavitate and remain filled with undifferentiated cells. While the specific mechanism by which Tim regulates cavitation and subsequent development will require further investigation, our results clearly implicate Tim in the first wave of Caspase 3-dependent apoptosis essential for EB cavitation and morphogenic progression.
In addition to early embryogenesis, other studies suggest that Tim may regulate apoptosis required for organ formation later in development. In the mouse embryonic kidney, for example, strong Tim expression has been observed in regions of active ureteric bud branching. Conditional knockdown of Tim with antisense oligonucleotides in both kidney rudiments and isolated ureteric bud cells profoundly inhibits embryonic kidney growth and ureteric bud morphogenesis 
. Interestingly, blocking Caspase activity also prevents ureteric bud formation, suggesting that Tim may also regulate apoptotic signals involved in organogenesis 
Because Tim was identified as a c-Src SH3 domain-binding protein in ES cells and EBs, Src kinase activity may regulate Tim protein levels and activity during embryogenesis. In support of this hypothesis, we show that full-length Tim is a substrate for c-Src following co-expression in 293T cells, and that Src-mediated phosphorylation correlates with enhanced Tim ubiquitylation. This finding suggests that tyrosine phosphorylation controls Tim protein levels in a manner analogous to circadian control of Tim in Drosophila 
. Furthermore, we observed that endogenous Tim is tyrosine-phosphorylated in self-renewing ES cells and early stage EBs, and that tyrosine phosphorylation precedes a dramatic decrease in Tim protein levels after EB cavitation has been completed. These observations are consistent with the idea that Src-family kinase-mediated phosphorylation regulates Tim ubiquitylation and proteasomal degradation during differentiation of ES cells to EBs. Interestingly, our previous studies have shown that Src-family kinase inhibitors reversibly delay EB formation 
. Here we show that these same inhibitors block endogenous Tim phosphorylation in ES cells, consistent with a role for a Src-Tim connection in the ES cell to EB transition.
How Tim is linked to the apoptotic machinery and the relationship of that connection to ES cell differentiation is not clear. Clues to a possible mechanism may come from cellular responses to environmental insults, such as DNA damage, which can also trigger apoptosis. Under these conditions, activation of checkpoint pathways initially induces cell cycle arrest, allowing time for DNA repair to occur. However, if DNA damage is severe, apoptosis is triggered instead (reviewed in 
). In addition to its circadian role, Tim also functions in S-phase checkpoint control and can induce cell cycle arrest by localizing to the stalled replication fork via interaction with Tipin 
. This observation suggests the possibility that Tim may cause cell cycle arrest in ES cells via a similar mechanism, thus indirectly inducing apoptosis as required for cavity formation.
In summary, our data provide the first evidence that a protein previously implicated in the control of circadian rhythms also has a unique role to play in mES cell differentiation to EBs, two processes that require tight temporal regulation. Remarkably little is known about the presence or role of circadian rhythms in ES cells or in the developing embryo. One very recent study suggests that circadian clock genes are regulated differently in adult versus embryonic cells 
. These authors showed that although genes associated circadian rhythms are expressed in the developing embryo in vivo, they were not expressed in a synchronized fashion characteristic of circadian cycling. However, when embryonic heart, liver and kidney tissues were placed into culture, rhythmic expression of one of these clock proteins (PER2) was observed. These observations suggest that ex vivo culture of embryonic tissues, as well as in vitro differentiation of ES cells to EBs as described in our study, may require coordination by circadian rhythm proteins. The more general idea that circadian rhythm proteins may coordinate cell lineage and tissue development should be of significant interest to those exploring the directed differentiation of ES cells in vitro. This process often requires EB formation to allow for coordinated differentiation of the three primitive germ layers 
. Support for this concept comes from recent work by Yagita, et al
., who showed that circadian feedback loops fail to oscillate in self-renewing cultures of mES cells, but become activated upon differentiation 
. Interestingly, oscillation was lost again upon genetic reprogramming of the cells back to the pluripotent state, providing a strong connection between circadian oscillations of clock protein promoter activity and differentiation.