The 14-3-3 family members are intracellular dimeric adapter proteins that play a central role in many cellular processes, including signal transduction, nutrient-sensing, checkpoint control, and apoptotic pathways (Fu et al.,2000; Muslin and Xing,2000). The 14-3-3 proteins are present in all eukaryotes. In particular, budding yeast have two 14-3-3 proteins, BMH1p and BMH2p, whereas mammals—and possibly all vertebrates—have seven family members encoded by separate genes: 14-3-3 beta (β), eta (η), epsilon (ϵ), gamma (γ), sigma (σ), tau (τ), and zeta (ζ; Fu et al.,2000). The main biochemical function of 14-3-3 proteins is to bind to phosphoserine/threonine motifs present on partner proteins. The 14-3-3 proteins bind to phosphoserine/threonine chiefly in the context of two amino acid motifs: R-S-x-pS/T-x-P or R-x-Y/F-x-pS/T-x-P, where x is any amino acid and pS/T is phosphoserine/threonine (Muslin et al.,1996; Yaffe et al.,1997). The crystal structure of a mammalian 14-3-3 ζ dimer complexed to a phosphoserine-containing partner protein, serotonin N-acetyltransferase, has been solved and this structure revealed that each dimer contains two phosphoserine/threonine binding pockets that bracket a large central channel (Obsil et al.,2001). The amino acids that line the phosphoserine/threonine binding pockets are highly conserved among all seven mammalian 14-3-3 proteins, and most differences are detected in portions of the protein that are not involved in phosphoserine/threonine binding or in dimerization.
The location of 14-3-3 binding motifs appears to be nonrandomly located in partner proteins. For example, 14-3-3 binding motifs are located in close proximity to nuclear location signals in several transcription factors, such as FOX3A (FKHRL1), and cell cycle proteins, such as cdc25c (Muslin and Xing,2000). When a 14-3-3 protein dimer binds to these proteins after they are serine/threonine phosphorylated, the nuclear localization signal is obscured and the partner proteins localize to the cytosol. The 14-3-3 binding sites can also be placed in close proximity to other information-containing domains, so that reversible phosphorylation events can lead to 14-3-3 protein binding and silencing of the information contained within those domains.
One important signaling molecule that binds to 14-3-3 proteins is the protein kinase Raf-1, a participant in the ras-ERK MAPK (extracellular regulated kinase mitogen-activated protein kinase) cascade that is activated by fibroblast growth factor (FGF) and other growth factors in embryonic development (Fantl et al.,1994; Freed et al.,1994; Fu et al.,1994; Irie et al.,1994). The 14-3-3 binding sites are located in both the amino-terminal regulatory domain, at serine-259 (and possibly at other sites), and at the carboxy-terminus, at serine-621, close to the kinase domain (Muslin et al.,1996; Rommel et al.,1996). In one model, 14-3-3 protein binding to both sites on Raf-1, maintains Raf-1 in an inactive, but “activatable” conformation so that ras-GTP can properly bind and trigger kinase activity. In this complex manner, 14-3-3 protein binding is required for Raf-1 activation. Genetic studies in Drosophila melanogaster support the role of 14-3-3 ϵ as a positive regulator of ras-mediated activation of the ERK MAPK cascade (Chang and Rubin,1997).
Previously, we examined the role of 14-3-3 proteins in early Xenopus development by use of a nonphosphorylated peptide inhibitor, GST-R18, which prevents 14-3-3 protein binding to phosphoserine/threonine-containing binding partners (Wu and Muslin,2002). GST-R18 inhibits the ability of all 14-3-3 family members to bind to phosphoserine/threonine-containing motifs (Wang et al.,1999). Microinjection of RNA encoding GST-R18 caused embryos to develop prominent gastrulation and axial patterning defects. These phenotypic abnormalities were similar to those observed in embryos injected with dominant-negative forms of the FGF receptor, Raf-1, H-ras, or MEK (Amaya et al.,1991; Whitman and Melton,1992; MacNicol et al.,1993; Umbhauer et al.,1995). Indeed, embryos injected with GST-R18 had reduced gastrula-stage expression of the general mesodermal marker genes Xbra and Xwnt8 but exhibited normal expression of the Spemann Organizer marker genes chordin and goosecoid (Wu and Muslin,2002). Animal cap experiments showed that GST-R18 blocked FGF-stimulated cap elongation and mesodermal gene expression, and this defect was rescued by injection with activated MEK.
Although experiments with GST-R18 established a general requirement for 14-3-3 protein function in FGF-mediated mesodermal specification, they did not identify the specific 14-3-3 protein(s) required for FGF signaling. Also, the severe phenotype observed in GST-R18–injected embryos may have obscured a role for 14-3-3 proteins in later developmental processes. Other investigators have demonstrated that at least three 14-3-3 genes, 14-3-3 ζ, τ, and ϵ, are expressed in early Xenopus embryonic development (Kousteni et al.,1997; Kumagai et al.,1998; Wu and Muslin,2002; Bunney et al.,2003). In one study, global gene expression analysis revealed that 14-3-3 ζ mRNA was present throughout embryonic development, but that higher levels accumulated after gastrulation (Kousteni et al.,1997). Whole-mount in situ hybridization on tail bud stage embryos showed robust 14-3-3 ζ gene expression in the head, optic vesicles, spinal cord, and branchial arches, but weaker expression in the somites (Kousteni et al.,1997). Recent work by Bunney et al. (2003) showed that 14-3-3 ϵ is expressed asymmetrically in stage 2 and 3 embryos and that 14-3-3 ϵ regulates left-right patterning. Although the expression of three Xenopus 14-3-3 genes have been described in oocytes and embryos, it is likely that several additional family members play a role in embryonic development.
Analysis of the Temporal and Spatial Expression Patterns of 14-3-3 Genes in Normal Embryos
The cDNAs for six 14-3-3 genes were identified by key word search as well as sequence blast search on the TIGR Xenopus database (http://www.tigr.org/tdb/tgi/xgi/). Mus musculus 14-3-3 cDNA sequences were used as bait. Xenopus cDNAs identified by this method included 14-3-3 β, ϵ, η, γ, τ, and ζ. No homologue of 14-3-3 σ was identified by sequence blast search. Figure 1 shows the comparisons of derived amino acid sequences of the six 14-3-3 family members in X. laevis. The accession numbers are as follows: x14-3-3 τ = TC204326, TC204327; x14-3-3 η = TC202286, TC202289; x14-3-3 γ = TC 13154, TC 195846; x14-3-3 ζ = TC 204594, TC 204596; x14-3-3 β = TC188860; x14-3-3 ϵ = TC186676. Identical amino acids are in yellow, with blue representing highly conversed amino acids and green representing moderately conserved amino acids as determined by Vector NTI Suite 8.0 (Infomax, Inc., Bethesda, MD). The sequences are 60–76% conserved among the six 14-3-3 family members at the DNA level, and 60–85% conserved at the amino acid level. The level of conservation of the specific 14-3-3 family members to their homologs in rat ranges from 68–86% at the DNA level, and 79–99% at the amino acid level.
To determine the temporal expression patterns of specific 14-3-3 genes in Xenopus developing embryos, reverse transcriptase-polymerase chain reaction (RT-PCR) experiments were performed by use of 14-3-3 family member-specific primers to detect their relative expression levels at various embryonic stages (Fig. 2). Two distinct sets of primers were designed for each 14-3-3 family member and the two primer sets yielded similar mRNA expression patterns for each 14-3-3 family member. In addition, all amplified PCR products produced from both primer sets matched their predicted size between the forward and reverse primers, suggesting that the primer designs were valid for 14-3-3 mRNA detection.
Expression studies showed that 14-3-3 β, τ, and ϵ were the most abundantly expressed isoforms throughout early Xenopus embryonic development (Fig. 2). The expression of three other 14-3-3 genes, η, γ, and ζ, was limited to low levels until later embryonic stages (Fig. 2). The 14-3-3-β gene expression levels were constant between stages 2 and 38. The 14-3-3 ϵ and τ shared similar gene expression patterns: mRNA levels slowly decreased during blastula and gastrula stages (stage 2 through 11.5) but increased after neurula stages (after stage 14). The 14-3-3 η and ζ mRNA levels were both very low between stages 2 and 26, and transcription was induced after stage 26. The 14-3-3 γ expression levels were variable throughout embryogenesis. Therefore, 14-3-3 family members have differing patterns of gene expression in Xenopus embryogenesis.
To determine whether the different 14-3-3 family members have different spatial expression patterns, whole-mount in situ hybridization experiments were performed by use of 14-3-3 family member-specific RNA probes. Normal stage 25 (Fig. 3A) and 40 (Fig. 3B) embryos were incubated with digoxigenin-labeled RNA probes that were antisense to individual 14-3-3 family members. Digoxigenin-labeled probes were recognized by anti-digoxigenin antibody and visualized by anti-diaminobenzidine alkaline phosphate staining. At stage 25, in situ hybridization showed that 14-3-3 γ, ϵ, and τ were abundantly expressed on the body surface (Fig. 3Aa–c), but 14-3-3 ζ, η, and β were not (Fig. 3Ad–f). By stage 40, 14-3-3 γ was more abundant in the cranium and central nervous system than other 14-3-3 family members (Fig. 3Ba). The 14-3-3 ϵ and τ were abundantly expressed in the skeletal myotomes, the posterior trunk, and the tail fin regions (Fig. 3Ab–c, 3Bb–c). The 14-3-3 ζ was expressed peripherally in the translucent tail fin but not in the trunk of the embryo (Fig. 3Ad, 3Bd), and 14-3-3 η and β were preferentially expressed in the trunk but not in the tail fins (Fig. 3Be,f). Figure 3C shows the magnified views of the expressions of 14-3-3 mRNAs in the cranium in stage 40 embryos. Taken together, 14-3-3 family members have differing temporal and spatial patterns of gene expression in Xenopus embryogenesis.
Injection of Antisense Morpholinos Reduced Specific 14-3-3 Protein Levels
To determine whether 14-3-3 family member-specific morpholino oligos could effectively and specifically reduce the translation of individual 14-3-3 proteins, Xenopus oocytes were injected with varying doses (20 to 40 pmol) of family member-specific morpholinos and injected oocytes were incubated for 1 to 5 days at 19°C. Cytosolic lysates were obtained from injected oocytes, proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and bands were analyzed by anti–14-3-3 family member-specific immunoblotting. In preliminary morpholino injection experiments, we used an oligo targeted against Xenopus 14-3-3 τ because of its abundance. Injection of oocytes with a 14-3-3 τ-specific morpholino effectively reduced τ protein levels (Fig. 4A). The same 14-3-3 τ protein level reduction was observed when we used a second 14-3-3 τ-specific morpholino that targets the coding frame beginning one base pair upstream of the original morpholino (data not shown). In oocytes injected with control morpholinos, 14-3-3 τ protein levels were unchanged. In addition, ERK MAPK and 14-3-3 ζ protein levels were unaffected by injection of oocytes with a 14-3-3 τ-specific morpholino (Fig. 4A). Similar protein knockdown was observed for 14-3-3 ϵ in 14-3-3 ϵ morpholino-injected oocytes (Fig. 4B).
We next examined the ability of other 14-3-3 family member-specific morpholinos to specifically reduce the protein levels of their intended targets. Xenopus oocytes were injected with 20 or 40 pmol of morpholinos targeted against 14-3-3 β, ϵ, γ, or ζ. Injected oocytes were incubated for 72 hr before they were lysed, proteins were separated by SDS-PAGE, and bands were analyzed by family member-specific immunoblotting. In each case, injection of morpholinos resulted in decreased protein levels for the corresponding 14-3-3 family member (Fig. 4C). Furthermore, the specific morpholinos were effective in blocking the protein production of exogenously expressed 14-3-3s in Xenopus oocytes (Fig. 4D). In addition, we demonstrated that 14-3-3 η morpholino could specifically knock down the endogenous level of 14-3-3 η protein (Supplementary Figure S2b, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). Thus, all six 14-3-3 morpholinos were specific and effective.
To confirm the utility of 14-3-3 morpholinos in developing Xenopus embryos, we injected approximately 100 pmol of 14-3-3 τ or β morpholinos into both blastomeres of two-cell stage embryos. Cytosolic lysates were harvested 24 hr after injection and separated by SDS-PAGE, and bands were analyzed by immunoblotting. As shown in Figure 4E, the 14-3-3 family member-specific morpholinos reduced the expression of the targeted 14-3-3 protein. Lower doses of morpholinos were also effective at 14-3-3 protein knock down in embryos (Supplementary Figure S2). Taken together, these data demonstrate that antisense morpholino injection specifically reduced individual 14-3-3 protein levels in Xenopus oocytes and embryos.
Phenotypic Abnormalities in Xenopus Embryos Injected With 14-3-3 Family Member-Specific Morpholinos
To determine whether these differences in gene expression reflect divergent biological functions carried by specific 14-3-3 proteins, morpholinos that target specific 14-3-3 mRNA were injected into Xenopus developing embryos. Each blastomere of two-cell stage embryos were injected with 5 pmol of family member-specific morpholinos at the animal poles. The embryos were observed until the tadpole stage. The effects of morpholino injections on Xenopus developing embryos are summarized in Table 1. A low dose of 14-3-3 family member-specific morpholinos was used, because injection of higher dosages of control morpholino (20 pmol or above) caused the embryos to develop nonspecific abnormalities that included trunk shortening and delayed maturation, as previously described (Heasman,2002).
Table 1. Effect of 14-3-3 Family Member-Specific Morpholinos in Xenopus Embryos
Survival to tadpole
Exogastrulation similar to τ morpholino-injected embryos but less extensive
Small or absent eyes
Exogastrulation; most die by stage 17
Figure 5 shows representative images of embryos injected with specific 14-3-3 morpholinos. Uninjected embryos or embryos injected into both cells of two-cell embryos with 10 pmol of nonspecific control morpholino developed normally through the tadpole stage. Embryos injected with 14-3-3 β morpholino displayed slight ventralization defects, with a score of approximately 4.0/10 on the dorsoanterior index (Kao and Elinson,1988), but were otherwise normal (Fig. 5Ab). This lack of obvious phenotypic abnormalities was surprising given the high level of 14-3-3 β expression in developing embryos. Embryos injected with ϵ-specific morpholino (Fig. 5Ac) developed an exogastrulation phenotype similar to that observed in embryos injected with GST-R18 mRNA, a general inhibitor of 14-3-3 function (Wu and Muslin,2002). Of interest, γ-specific morpholino-injected embryos displayed a small eye or no eye phenotype, similar to those observed in Pax-6–deficient embryos (Hirsch and Harris,1997; Fig. 5Ad). Embryos injected with a 14-3-3 η or ζ morpholino did not develop any obvious phenotypic defects (Fig. 5Ae,f).
Role of 14-3-3 τ and ϵ in Mesoderm Induction
Initial morpholino injection experiments clearly demonstrated that the most severe phenotypic defects were observed in 14-3-3 τ morpholino-injected embryos. These embryos formed a blastopore lip normally (Fig. 5Ba), but the majority (87%) failed to complete gastrulation (Table 1). At the mid-to-late gastrula stage, the closure of the lateral and ventral regions of the blastopore was suspended and, to some extent, underwent evagination instead of invagination, resulting in an open blastopore in later neurula stages (Fig. 5Bb,c). In contrast, only 7% of the embryos injected with the control morpholino underwent exogastrulation. Most of the two-cell embryos injected bilaterally with 14-3-3 τ-specific morpholinos embryos died before stage 17. To further confirm that these exogastrulation defects were caused by specific knockdown of 14-3-3 τ proteins, we designed antisense DNA oligos that specifically targets 14-3-3 τ mRNA. Embryos injected with 14-3-3 τ-specific antisense DNA oligos displayed the same phenotype as morpholino-injected embryos, including exogastrulation defects and decreased survival through gastrula stages (data not shown). We therefore conclude that the phenotypic abnormalities associated with morpholino injection were specifically caused by the reduction of individual 14-3-3 proteins.
When embryos were injected in both cells of a two-cell embryo with a lower dose (5 pmol) of 14-3-3 τ-specific morpholino, some embryos survived past stage 17 but they displayed posterior axial defects similar to that observed in embryos overexpressing dominant-negative FGF receptor (Amaya et al.,1991), dominant-negative H-ras (Whitman and Melton,1992), dominant-negative Raf-1 (MacNicol et al.,1993), or the GST-R18 mRNA (Wu and Muslin,2002). Embryos injected unilaterally at the two-cell stage embryo (Fig. 5Bf,g) or into the two opposite quadrants of the four-cell embryos (Fig. 5Bh,i) had higher survival through gastrula and neurula stages than those injected bilaterally with the same dosage of 14-3-3 τ morpholino. For unilateral morpholino injection (Fig. 5Bf,g), embryos with apparent dorsal–ventral pigment demarcations were selected and τ-specific morpholinos were injected only to the left blastomere of the two-cell embryos. Trunk structures were diminished on the side of injection (left side), producing a bent phenotype. These bent tadpoles were similar to those observed by unilateral injection of GST-R18 mRNA (Wu and Muslin,2002). Anterior structures were relatively intact. When the embryos were injected in opposite quadrants at the 4-cell stage, the posterior trunk truncation defects became more severe, and the tadpoles were unable to swim or extend anterior-posteriorly into their normal configuration. To confirm our findings with the 14-3-3 τ morpholino injection, a second 14-3-3 τ-specific morpholino was generated for embryo injection (see the Experimental Procedures section) and identical exogastrulation phenotypes were observed.
Embryos injected with 14-3-3 ϵ morpholino displayed phenotypic abnormalities that were similar to, but less severe than, 14-3-3 τ morpholino-injected embryos (Table 1, Figure not shown). A second morpholino that specifically targeted 14-3-3 ϵ produced the same phenotypic defects.
Analysis of Mesodermal Marker Gene Expression in Embryos Injected With 14-3-3 τ and ϵ Morpholinos
Previous work in our lab established the role of 14-3-3 proteins in FGF signaling during early Xenopus development (Wu and Muslin,2002). To determine whether the phenotypic abnormalities associated with the loss of 14-3-3 τ and ϵ were due to defective mesodermal induction, stage 10.5 embryos were subjected to whole-mount in situ hybridization by use of the mesodermal marker genes Xbra and Chordin. Embryos subjected to in situ hybridization with the Xbra probe were injected either bilaterally, unilaterally, or at opposite end blastomeres with 10 pmol of 14-3-3 τ- or ϵ-specific morpholinos. Figure 6a and 6e show that two-cell embryos injected bilaterally with 14-3-3 τ and ϵ morpholinos had very low level of Xbra expression, compared with control-injected embryos (Fig. 6d), where Xbra was expressed throughout the marginal zone. When one of the two blastomeres of a two-cell stage embryo was injected with 10 pmol of 14-3-3 τ- or ϵ-specific morpholinos, Xbra expression was reduced on the injected side (Fig. 6b,f). When two opposite blastomeres of four-cell stage embryos were injected with 10 pmol of 14-3-3 τ- or ϵ-specific morpholino, Xbra expression was only reduced in these two quadrants (Fig. 6c,g). Chordin expression, on the other hand, was not affected by 14-3-3 τ or ϵ morpholino injection, indicating that the observed phenotypic defects were not due to abnormal Spemann Organizer function (Fig. 6h–j).
To determine whether 14-3-3 τ and ϵ morpholinos interfered with ras-ERK MAPK signaling, animal cap rescue experiments were performed with a constitutively active form of MEK (CA-MEK). Plasmid DNA containing the CA-MEK cDNA downstream of the Xenopus borealis cytoskeletal actin (CSKA) promoter was used in these studies (Harland and Misher,1988; Wu and Muslin,2002). The 14-3-3 τ- and ϵ-specific morpholinos were first injected into both blastomeres of two-cell embryos, followed by injection of CA-MEK plasmid DNA. Animal caps were dissected from double-injected stage 9 embryos and incubated with basic FGF (bFGF; 50 ng/ml) or vehicle for 12 hr. FGF-stimulated mesoderm induction was scored by elongation of the animal caps as described previously (MacNicol et al.,1993). Examples of elongated versus nonelongated animal caps are shown in Supplementary Figure S3. Animal caps isolated from uninjected embryos underwent elongation normally (100%) when stimulated with FGF and did not undergo elongation in the absence of FGF (0%). Animal caps isolated from 14-3-3 τ or ϵ morpholino-injected embryos failed to undergo cap elongation in the presence of FGF stimuli (both at 4% elongation; Table 2). Coinjection of CSKA-CA-MEK DNA with 14-3-3 τ or ϵ morpholinos partially rescued the defect of FGF-mediated mesoderm induction (62% and 67% elongation, respectively).
Table 2. Animal Cap Assay Revealed Partial Rescue of 14-3-3 τ or ϵ Morpholino-Induced Mesodermal Induction Failure
τ morpholino 20 pmol
τ morpholino 20 pmol
τ morpholino 20 pmol
τ morpholino 20 pmol
ϵ morpholino 20 pmol
ϵ morpholino 20 pmol
ϵ morpholino 20 pmol
ϵ morpholino 20 pmol
Microinjection of 14-3-3 τ mRNA Rescued Animal Cap Elongation and Exogastrulation Defects in 14-3-3 τ Morpholino-Injected Embryos
To confirm that the failure of cap elongation was specifically caused by reduced 14-3-3 τ protein levels, we co-injected myc-tagged X. laevis 14-3-3 τ mRNA with 14-3-3 τ-specific morpholino into two-cell embryos to determine whether supplementation of exogenous 14-3-3 τ mRNA could reverse the cap elongation defects. The N-terminal myc tag prevents the binding of the 14-3-3 τ morpholino to the translation initiation site.
Two-cell stage embryos were uninjected, injected with 5 pmol of 14-3-3 τ-specific morpholino, co-injected with 5 pmol of 14-3-3 τ-specific morpholino and 50 ng of myc-14-3-3 τ mRNA, or co-injected with 5 pmol of 14-3-3 τ morpholino and 50 ng of myc-14-3-3 η mRNA (Fig. 7A). Animal caps were dissected from stage 9 embryos and incubated with bFGF (50 ng/ml) or vehicle for 12 hr and then scored for elongation. All animal caps from uninjected embryos elongated in response to FGF (100% elongation), but 14-3-3 τ morpholino-injected caps were resistant to FGF-stimulated elongation (3 ± 0.5% elongation; Fig. 7A). Co-injection of 14-3-3 τ mRNA with the 14-3-3 τ morpholino treatment partially rescued (69 ± 9.9%) FGF-stimulated elongation. Co-injection with a different 14-3-3 mRNA, 14-3-3 η, was less effective (36 ± 7.1%) at rescuing FGF-stimulated elongation caused by the 14-3-3 τ morpholino (Fig. 7A).
To confirm that exogastrulation defects were specifically caused by 14-3-3 τ knockdown, we injected myc-tagged 14-3-3 τ mRNA with 14-3-3 τ morpholino. Co-injection partially rescued the exogastrulation defects caused by injection of 14-3-3 τ morpholino alone (Fig. 7B). Co-injection with a different 14-3-3 mRNA, 14-3-3 ϵ, also rescued the exogastrulation phenotype caused by 14-3-3 τ morpholino to a similar extent (Fig. 7A). This finding suggests that 14-3-3 τ and ϵ have overlapping functions and can compensate for the loss of each other during early embryogenesis.
Developmental Defects in 14-3-3 τ- or ϵ-Specific Morpholino-Injected Embryos Were Associated With Increased Apoptosis
The 14-3-3 proteins are key negative regulators of important pro-apoptotic proteins, including BAD and FKHRL-1 (Rosenquist,2003). To determine whether the phenotypic abnormalities found in 14-3-3 τ or ϵ morpholino-injected embryos were associated with an increased rate of apoptosis, two-cell stage embryos were injected with 14-3-3 τ- or ϵ-specific morpholinos and frozen at several early embryonic stages. Apoptosis was scored with a commercial enzyme-linked immunosorbent assay (ELISA) -based colorimetric assay that detects mono- and oligonucleosomes in the cytoplasmic fraction of embryonic lysates (Roche, Cell Death Detection ELISA plus). Embryos injected with 14-3-3 τ- or ϵ-specific morpholinos displayed a much higher rate of apoptosis than uninjected embryos at the same developmental stage (Fig. 8A).
Co-injection of 14-3-3 τ mRNA partially rescued animal cap elongation defects caused by the 14-3-3 τ-specific morpholino (Fig. 7). Embryos co-injected with 14-3-3 τ morpholino and 14-3-3 τ mRNA were evaluated for apoptosis. Lysates were obtained from uninjected, 14-3-3 τ mRNA-injected, 14-3-3 τ morpholino-injected, or morpholino/mRNA–co-injected stage 12 embryos (Fig. 8B). Apoptosis was increased in 14-3-3 τ morpholino-injected embryos when compared with uninjected embryos (Fig. 8B, P < 0.05). Injection of 14-3-3 τ mRNA by itself (without the morpholino) did not reduce apoptosis compared with control embryos. Co-injection of 14-3-3 τ mRNA with 14-3-3 τ morpholino reduced the rate of apoptosis to basal levels (Fig. 8B). Co-injection with a different 14-3-3 mRNA construct, 14-3-3 ζ, also reduced the rate of apoptosis (Fig. 8B).
Role of 14-3-3 γ in Eye Development
Initial morpholino injection experiments demonstrated that reduction in 14-3-3 γ protein caused head and eye defects in Xenopus embryos. To confirm that the effect of 14-3-3 γ morpholino on eye development was due to reduction in 14-3-3 protein activity, injection with a pan-14-3-3 inhibitor GST-R18 was performed. To target 14-3-3 inhibition specifically to anterior neural tissue and to avoid axial patterning defects, eight-cell embryos were injected in one or two dorsal animal pole blastomeres with 10 ng of GST-R18 mRNA. When two dorsal animal blastomeres of eight-cell embryos were injected with RNA encoding GST or GST-R18, 54% of GST-R18–injected embryos developed obvious eye/head defects (data not shown), but only 25% of GST-injected embryos developed similar defects. No embryos developed gastrulation or axial patterning defects with this injection method. RNA encoding GST-R18 was next injected into a single dorsal animal pole blastomere of eight-cell embryos in order that eye development on one side of the embryo could serve as a control for the other side. By this technique, embryos displayed markedly reduced (or absent) eyes on the side injected with GST-R18 RNA (Fig. 9A). Injection of GST RNA did not affect eye development on the injected side (Fig. 9Ab). To demonstrate that the inhibition of eye formation was restricted to cells that received the injected RNA, GST-R18 mRNA was co-injected with a cell lineage tracer, β-gal RNA, into single dorsal animal blastomeres of eight-cell embryos. When the embryos were stained with β-gal substrate, the accuracy of RNA targeting to the eye/head region was ∼80% (n = 79). Fifty-three percent (n = 58) of GST-R18–injected embryos with RNA distributed to the eye/head region had diminished eye/head size on the injected side of the embryo (Data not shown). Only 10% (n = 82) of GST-injected embryos with RNA distributed to the eye/head region had diminished eye/head size on the injected side of the embryo.
At the neural plate stage, the homeobox genes Rx, Six3, and Pax6 are expressed in an area of the embryo that demarcates the retina morphogenetic field and these genes are also expressed at a later time in the differentiated retina (Mathers et al.,1997). To analyze whether genetic markers of eye development were inhibited by GST-R18 RNA injection, whole-mount in situ hybridization experiments were performed. In stage 13 embryos that had been injected with GST-R18 RNA into a single dorsal animal pole blastomere at the eight-cell stage, Xenopus Rx (Xrx) expression was reduced by half (Fig. 9B). A single morphogenetic field, referred to as the retina field, gives rise to two separate, laterally located retina forming groups of cells (Li et al.,1997). The reduction in the size of the Xrx field of expression that we observed in response to GST-R18 injection demonstrates that eye development was disrupted at an early stage (stage 13) of organogenesis.
We next injected 14-3-3 γ-specific morpholinos into Xenopus developing embryos to observe the effect of reducing 14-3-3 γ protein level on eye development (Fig. 9Ac,d). When both blastomeres of two-cell stage embryos were injected with 14-3-3 γ morpholino, the size of both eyes in stage 35 embryos were reduced compared with wild-type (30/56, 54%). When only a single blastomere of two-cell stage embryos was injected with the 14-3-3 γ-specific morpholino, the injected side of the embryos had reduced or absent eyes (57/85, 67%) but the uninjected side developed normal head and eye structures. The two-cell embryos injected with scrambled morpholino did not have reduced eye size (2/26, 7.7%). To demonstrate that the effect on eye formation was specific to 14-3-3 γ, 14-3-3 η-specific morpholino was injected unilaterally into two-cell stage embryos and no reduction of eye size was found in these embryos (3/28, 11%; Fig. 9Ac,d).
To confirm that 14-3-3 γ is necessary for eye formation, in situ hybridization was performed to evaluate Pax6 marker gene expression. The two-cell stage embryos were injected bilaterally with 10 pmol of 14-3-3 τ-specific morpholinos. The developing embryos were fixed at stage 35 and subjected to in situ hybridization with the Pax6 probe (Fig. 9C). The two-cell stage embryos injected bilaterally with the 14-3-3 γ morpholino had reduced Pax6 staining compared with embryos injected with control morpholino oligo. When one of the two blastomeres of a two-cell stage embryo was injected with 10 pmol of 14-3-3 γ-specific morpholino, Pax6 expression was reduced on the injected side (data not shown).
We performed whole-mount immunostaining to examine the localization of 14-3-3 γ protein in developing embryos. The anti–14-3-3-γ antibody used in immunostaining is specific in recognizing this 14-3-3 isoform (Fig. 9C). We found that 14-3-3 γ protein was preferentially located in the head and eye regions, especially around the optic groove area (Fig. 9D). By contrast, incubation with the secondary antibody alone gave no specific staining patterns. This protein localization is consistent with our previous findings that reduction of 14-3-3 γ caused eye-specific defects (Fig. 5A). To further confirm that 14-3-3 γ is preferentially expressed in the head and eye regions, we performed RT-PCR experiments on different tissue sections of normal stage 26 embryos. We found that 14-3-3 γ was highly expressed in the head and eye region, but minimally expressed in the trunk and tail regions (Fig. 9E). Therefore, it is likely that the eye-specific function of 14-3-3 γ is due to its specific localization to the head and eye regions.
The current study is intended for a surface overview of individual 14-3-3 family member functions during X. laevis embryogenesis, and it provides the foundation for more in-depth analyses of the specific cellular pathways and mechanisms that regulate family-member specific functions. The 14-3-3 proteins play a vital role in the regulation of many eukaryotic cellular processes, but the unique functions of individual family members remain largely unexplored. The 14-3-3 protein dimers act by changing the conformation of binding partners in some cases and by obscuring nearby information-containing domains in other cases, such as nuclear localization signals, nuclear export signals, and membrane localization domains (Fu et al.,2000; Muslin and Xing,2000). All 14-3-3 family members bind to the same phosphoserine/threonine motifs in vitro, and this finding raises the question of why there are seven unique 14-3-3 genes in mammals. In this work, we have attempted to address this issue in the context of Xenopus embryonic development.
The experimental approach in this work took advantage of the ease of protein knockdown experiments in Xenopus by use of antisense morpholino oligo injections. Database searching yielded the cDNAs for six Xenopus 14-3-3 family members: β, η, ϵ, γ, τ, and ζ. The validity of this morpholino injection approach was confirmed by Western blotting (Fig. 4A–E) and specific rescue experiments using tagged 14-3-3 expression constructs (Figs. 7, 8) and activated cMEK microinjection (Table 2). Each of the morpholinos caused comparable levels of target protein reduction (approximately 50%, Fig. 4A–E). The lack of complete protein knockdown seen on Western blotting might be due to several factors, such as the persistence of maternally expressed 14-3-3 proteins, antibody cross-reactivity, or the low dosage of morpholinos injected.
Expression analysis revealed that 14-3-3 β, ϵ, γ, and τ were the most abundant family members during the early stages of embryogenesis. In situ hybridization analysis revealed differential expression patterns of 14-3-3 family members in early embryogenesis. Morpholino injection experiments revealed marked differences in the function of the various 14-3-3 family members. Knockdown of 14-3-3 ζ did not cause phenotypic abnormalities. In contrast, knockdown of 14-3-3 τ, and to a lesser extent ϵ, caused prominent gastrulation defects, reduced Xbra gene expression, and normal Chordin gene expression. Injection of 14-3-3 γ morpholino did not cause gastrulation defects but, instead, inhibited eye development in neurula and tail bud stages. Finally, despite its high level of expression, knockdown of 14-3-3 β protein levels did not result in significant phenotypic abnormalities. The observation that knockdown of 14-3-3β did not cause gastrulation abnormalities or mesodermal defects is significant, because 14-3-3β is highly and globally expressed in early development, and this result suggests that there are important binding partners for 14-3-3τ and 14-3-3ϵ that do not bind to or are not affected by 14-3-3β. Future studies will focus on the identification of partner proteins found in early Xenopus embryos that specifically bind to 14-3-3τ but not to 14-3-3β.
Our findings that the six X. laevis 14-3-3s contributed differently to early embryogenesis are mirrored by other studies. In barley, three 14-3-3 family members, 14-3-3A, B, and C, are differentially expressed and processed in microspore embryogenesis (Maraschin et al.,2003). Previous studies have shown that family member-specific expression patterns, ligand interactions, and transcriptional regulation are all important factors for the different functions of individual 14-3-3s (Roberts and de Bruxelles,2002). Our expression analyses showed that the transcriptions of individual 14-3-3s were differentially regulated both in time and space (Figs. 2, 3). It is important in the future to analyze the ligand specificity and promoter regulation of 14-3-3 ϵ, τ, and γ to determine how these factors may affect 14-3-3 family member specificity during embryogenesis. Other studies have shown that the C-terminus of 14-3-3 proteins mediate family member-specific ligand interactions (Benzinger et al.,2005; Bornke,2005). Our alignment of the X. laevis 14-3-3 protein sequences showed little amino acid variability in the C-terminus, with the exception of 14-3-3 ϵ, which is unique among the six family members in having an additional sequence of negatively charged amino acid residues (Fig. 1, position 271–277 for 14-3-3 ϵ). It is unclear how this may affect the functional specificity for 14-3-3 ϵ during early Xenopus embryogenesis.
The phenotype of 14-3-3 τ and ϵ morpholinos-injected embryos was similar to that observed in our previous work with GST-R18, a general peptide inhibitor of 14-3-3 action (Wu and Muslin,2002). In work with GST-R18, 14-3-3 protein action was required for FGF-mediated mesodermal specification and axial patterning. Animal cap experiments revealed that activated MEK could rescue FGF-stimulated mesoderm induction that was inhibited by either GST-R18 or 14-3-3 τ and ϵ morpholinos (Table 2). Furthermore, morpholino-treated embryos were found to have an increased rate of cellular apoptosis (Fig. 8). Taken together, these results suggest that 14-3-3 τ and ϵ are the key family members in FGF-mediated signaling in early Xenopus embryonic development and that reduced FGF signaling may be linked to increased programmed cell death during gastrulation.
In mammals, 14-3-3 γ is abundantly expressed in the brain, nervous tissues, and cerebral spinal fluid (Wiltfang et al.,1999; Chen et al.,2005). Of the different 14-3-3 family members, 14-3-3 γ alone is specifically induced in the nervous tissue when subjected to oxidative stress or metabolic stimuli (Traina et al.,2004; Chen et al.,2005). Furthermore, knockdown of 14-3-3 γ proteins by antisense RNA resulted in enhanced apoptosis in astrocytes subjected to ischemia (Chen et al.,2005). In one study, overexpression of 14-3-3 γ was found to specifically repress the transcription of several synaptic reporter genes in cultured myotubes and to alter the morphology of the neuromuscular junction (Strochlic et al.,2004). Our findings that reduction of 14-3-3 γ proteins inhibited head and eye development is consistent with the findings of others that 14-3-3 γ plays a specific role in the nervous system. Whole-mount in situ hybridization and immunostaining revealed preferential expression of 14-3-3 γ in the head and eye regions (Figs. 3, 9D), giving further proof that 14-3-3 γ is unique among different 14-3-3 family members to mediate neural development.
Given the large number of 14-3-3 protein binding partners and the number of potential pathways that take part in eye formation, it is not apparent which pathway is disrupted in eye development by reduction in 14-3-3 γ protein. Recently, other investigators successfully used an affinity isolation approach to identify more than two hundred 14-3-3 substrates that recognize 14-3-3s using the canonical binding motifs (Rubio et al.,2004). There is evidence that two of the identified substrates, FOXK1 and FOXJ1, may play an important role in Xenopus eye development. FOXK1 and FOXJ1 are 14-3-3–interacting transcription factors (Pohl and Knochel,2004). The expression of FOXK1 is initially restricted to neuroectoderm but is expressed later in other tissues including brain, eye, and otic vesicles.
It is unclear whether the effects of 14-3-3 γ morpholino injection on eye development are related to FGF receptor signaling. Previous work demonstrated that the FGF receptor 1 is not required for early Xenopus eye induction or patterning (Kroll and Amaya,1996), and inactivation of FGF by use of a dominant-negative FGF receptor 1 did not prevent animal blastomeres from contributing to the retina (Moore and Moody,1999). In contrast, a more recent study in transgenic frogs showed that FGF receptor 4a is required for the correct specification of retinal cell types (Zhang et al.,2003). Also, another group demonstrated that an interaction between FGF and ephrinB1 is required for morphogenetic movements that underlie eye field formation (Moore et al.,2004). EphrinB1 is a transmembrane ligand of the Eph receptor tyrosine kinase family and is highly expressed in the embryonic retina. Coincidentally, sequence analysis showed that EphrinB1 possesses a putative 14-3-3-interacting domain. It will be interesting to evaluate the role of EphrinB1, FOXJ1 or FOXK1 in 14-3-3 γ-mediated eye formation, and to determine whether these targets bind selectively to 14-3-3γ and not to other 14-3-3 family members.
In addition to the mesodermal induction and eye formation defects elaborated in this study, individual 14-3-3 family member proteins were found by another group to be involved in left–right patterning (Bunney et al.,2003). In our work, embryos injected with specific 14-3-3 morpholinos did not display abnormalities in left–right patterning. However, we injected embryos at the two-cell or four-cell stages, and 14-3-3 inhibitors injected after the one-cell stage were shown to be ineffective at inducing heterotaxia (Bunney et al.,2003). In addition, the phosphopeptide that Bunney et al. (2003) used to inhibit 14-3-3 function has a much shorter half-life than morpholino oligos that persist in embryos for days. Therefore, whereas Bunney et al. examined the transient effect of 14-3-3 inhibition, our morpholino approach evaluates the effect of prolonged loss of individual 14-3-3 proteins.
Morpholinos, Antisense Oligos, and Plasmid Constructs
The 14-3-3 family member-specific antisense morpholino oligos were purchased from GeneTools, LLC (Philomath, OR). Morpholinos were designed to be antisense to the initiation AUG and subsequent 22 ribonucleotide residues in the mRNAs encoding Xenopus 14-3-3 proteins. For 14-3-3 τ and ϵ, a second morpholino one base pair upstream or downstream of the original morpholino sequence was designed to validate the original experimental findings.
14-3-3 β morpholino, 5′-TCTGTACCAGTTCACTCTTGTCCAT-3′; 14-3-3 η morpholino, 5′-GCAACTGCTGCTCCCGATCAGCCAT-3′; 14-3-3 ϵ morpholinos, 5′-ACACTAAATCCTCTCGCTCTTCCAT-3′, 5′TACACTAAATCCTCTCGCTCTTCCA-3′; 14-3-3 γ morpholino, 5′-GCACCAGCTGCTCGCGGTCCACCAT-3′; 14-3-3 τ morpholinos, 5′-TCTGGATTTGTGCGGTCCTGTCCAT-3′, 5′-GTCTGGATTTGTGCGGTCCTGTCCA-3′; and 14-3-3 ζ morpholino, 5′- TCTGGACCAGTTCATTTTTATCCAT-3′. The control morpholino was a scrambled version of the experimental morpholinos commercially available from GeneTools, LLC.
Antisense DNA oligos were purchased from Integrated DNA Technology (Coralville, IA) and have identical sequences to the morpholinos that target the specific 14-3-3 family members. The three terminal linkages on both ends of the antisense oligos have a phosphorothioate-modified backbone to provide resistance to nucleases.
The CSKA-CA-MEK plasmid DNA was described previously (Wu and Muslin,2002). The 14-3-3 τ, ϵ, η, and ζ mRNAs used for microinjections were transcribed in vitro using Ambion mMachine SP6 mRNA preparation kit (Ambion, Inc., TX), following manufacturer's instructions.
Oocyte and Embryo Injections and Animal Cap Assays
Stage 6 oocytes were surgically isolated from mature females as described (Sive et al.,2000). In vitro fertilization of Xenopus laevis eggs was performed as previously described (MacNicol et al.,1993). Embryos were staged by the method of Nieuwkoop and Faber (1967). Ventralization was scored by the methods of Kao and Elinson (1988). Animal cap assays were performed as previously described (Wu and Muslin,2002).
Apoptosis was measured with the Roche Cell Death Detection ELISA Plus Kit (Roche Applied Science, Indianapolis, IN). Ten embryos per treatment were homogenized in 1,000 μl of lysis buffer provided by the kit and lysed for 30 min before centrifugation at 2,500 rpm for 5 min to collect cell membrane and debris to the bottom of the tubes. Twenty microliters of the supernatant containing the cytosolic fraction was used for the colorimetric assay according to the manufacturer's instructions.
Xenopus oocyte and embryo lysates were prepared as previously described (Wu and Muslin,2002). Western blots were prepared using standard molecular biology protocols. The membranes were incubated with 14-3-3 family member-specific antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and rabbit polyclonal anti-ERK p44/p42 antibody (Cell Signaling Technology, Inc., Danvers, MA) and visualized with horseradish peroxidase–coupled secondary antibody and ECL chemiluminescence-developing agent (Amersham Biosciences, Buckinghamshire, England). Selected bands on scanned immunoblots were quantified by the densitometry computer program Image J (NIH, Bethesda, MD). The target band densities were normalized to the amount of protein loaded in each lane (i.e., total ERK).
RNA Purification and RT-PCR
Total RNA was purified from Xenopus embryos at various stages by use of RNAzol (Tel-Test, Inc., Friendswood, TX). RT-PCR analyses were performed as previously described (Cui et al.,1995). Primers corresponding to individual 14-3-3 family members were designed with specific sequences located in the coding region of each family member. Two independent sets of RT-PCR primers were used for each 14-3-3 family member: · 14-3-3 β primer set #1: forward, 5′-AAGAGTGAACTGGTACAG-3′; reverse, 5′-CTAATTTCAAAGGCTTCC-3′. Set #2: forward, 5′-ACAAGAGTGAACTGGTACAGA-3′; reverse, 5′-TCTGGTGGGGTTGCGTTAGGT-3′ · 14-3-3 ϵ primer set #1: forward, 5′-TGATAAGGGAATATCGGC-3′; reverse, 5′-GTTCTCTGCAGCCTCCTT-3′. Set #2: forward, 5′-AAGAGCGAGAGGATCTAGTGT-3′; reverse, 5′-TCTCCACTGCTTGCAGCTGGA-3′ · 14-3-3 η primer set #1: forward, 5′-ATGGCTGATCGGGAGCAG-3′; reverse, 5′-GGTTCAGATGTTCAGTCAC-3′. Set #2: forward, 5′-CTGATCGGGAGCAGCAGTTG-3′; reverse, 5′-CGAATTGGAAGTCATTGCAGTTC-3′ · 14-3-3 γ primer set #1: forward, 5′-GACCTTCCGGTTCAGAGC-3′; reverse, 5′-ACGTGAGCGTTGCTCAGC-3′. Set #2: forward, 5′-AGCAGCTGGTGCAGAAAG-3′; reverse, 5′-TCGTATTGGGTTTCACTGCA-3′ · 14-3-3 τ primer set #1: forward, 5′-ATTGCCAAAGAGTACAAA-3′; reverse, 5′-AATCTTCTATTGTTTTAG-3′. Set #2: forward, 5′-ACAGGACCGCACAAATCCAGA-3′; reverse, 5′-TCAGTCGAAGTTGAAGAGCTT-3′ · 14-3-3 ζ primer set #1: forward, 5′-ATGGGCCCTATACTAGGT-3′; reverse, 5′-TCTCTTCATACAAGCAGC-3′. Set #2: forward, 5′-ATGAACTGGTCCAGAAGGCCA-3′; reverse, 5′-TCTGGTTGCGTGGCATTGGCA-3′
Primers for ornithine decarboxylase were as follows: forward, 5′-AATGGATTTCAGAGACCA-3′; reverse, 5′-AGAGTGGTGTGTGGAATC-3′ (Bassez et al.,1990).
Whole-Mount In Situ Hybridization and Immunostaining
Whole-mount in situ hybridization was performed by the method of Sive et al. (2000) and was described previously (Wu and Muslin,2002). Whole-mount immunostaining was performed by the method of Dent et al. (1989).
All data are reported as mean ± SEM. Statistical analysis was performed by two-tailed Student t-test or analysis of variance where applicable. A value of P < 0.05 was considered statistically significant.
The authors thank Kristen Kroll for technical advice and for helpful suggestions about the manuscript. A.J.M. was funded by an NIH grant and by a grant from the Burroughs Wellcome Fund.