Tob genes in development and homeostasis

Authors

  • Shunji Jia,

    Corresponding author
    1. Protein Science Laboratory of the Ministry of Education, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing, China
    • Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing 100084, China
    Search for more papers by this author
  • Anming Meng

    Corresponding author
    1. Protein Science Laboratory of the Ministry of Education, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing, China
    • Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing 100084, China
    Search for more papers by this author

Abstract

Members of the Btg/Tob protein family share a conserved N-terminal region that confers the activity to inhibit cell proliferation. Tob1 and Tob2 proteins, which constitute a Tob subfamily, have a longer C-terminal region than BTG proteins. Apparently, genomes of invertebrates and teleost species contain only a single Tob locus, whereas genomes of mammalian, avian, and amphibian species contain two Tob loci (Tob1 and Tob2). Tob genes are expressed in oocytes, sperm, early embryos, and various adult tissues, depending on the species. Recent reports indicate that Tob proteins play important roles in spermatogenesis, embryonic dorsoventral patterning, osteogenesis, T-cell activation, and learning and memory. Accumulating evidence supports the hypothesis that Tob proteins act primarily as transcriptional repressors in several signaling pathways. Developmental Dynamics 236:913–921, 2007. © 2007 Wiley-Liss, Inc.

INTRODUCTION

The regulation of cell proliferation involves dozens of extracellular signals and intracellular factors of various types. In the early 1990s, several novel intracellular factors sharing a high degree of sequence homology were found to play important roles in regulating cell proliferation in mammalian cells. Nerve growth factor induces expression of one of these factors, PC3, during neuronal differentiation of rat PC12 cells (Bradbury et al.,1991). TIS21 was found to be induced by the tumor promoter tetradecanoyl phorbol acetate in Swiss murine 3T3 cells (Fletcher et al.,1991), and BTG1 was cloned from near the breakpoint of a t(8;12) chromosomal translocation in a B-cell chronic lymphocytic leukemia (Rouault et al.,1992). Based on sequence homology and the antiproliferative effects of PC3 and BTG1, Rouault et al. (1992) proposed that these factors are members of a novel family of antiproliferative proteins. Since then, a few more members of this family, including TOB/TOB1 (transducer of ErbB2; Matsuda et al.,1996), BTG3/ANA (Yoshida et al.,1998), TOB2 (Ikematsu et al.,1999), and BTG4 (Buanne et al.,2000), have been identified in the human genome. This family of related proteins is now referred to as the BTG/Tob or APRO family (Matsuda et al.,2001). Orthologs of human BTG/TOB genes have been identified in several vertebrate and invertebrate genomes, suggesting a high degree of evolutionary conservation. Tob proteins (generally >300 residues) are significantly larger than BTG proteins (158–252 residues), however, Tob and BTG proteins have similar N-terminal sequences and antiproliferative activities. We suggest that Tob proteins be classified as members of a subfamily. In this review, we focus on the properties of this Tob protein subfamily, with special emphasis on their functions during development.

STRUCTURE AND BIOCHEMICAL PROPERTIES OF TOB PROTEINS

The Tob subfamily member, TOB/TOB1, was identified in a human cell line, where it interacts with the ERBB2 gene product p180; it was thus considered to be a possible transducer of ERBB2 (TOB; Matsuda et al.,1996). In an attempt to isolate new members of the BTG/Tob family, Ikematsu et al. (1999) cloned the human TOB2 gene (Ikematsu et al.,1999). To date, orthologs of the human TOB1 and TOB2 genes have been identified in mouse (Yoshida et al.,1997; Ajima et al.,2000), chicken (Dragon et al.,2002; Caldwell et al.,2005), and frog (GenBank accession no. AAH60329; Yoshida et al.,2000). However, only a single Tob-related gene has been identified in the invertebrate species Caenorhabditis elegans (Chen et al.,2000), C. briggsae (Chen et al.,2001), C. remanei (Chen et al.,2001), Drosophila melanogaster (Bourbon et al.,2002), and Branchiostoma floridae (Holland et al.,1997). In zebrafish, a teleost species, two orthologs of human TOB1, tob1a and tob1b, have been identified (Shi et al.,2004; Xiong et al.,2006). We have failed to identify a Tob2 ortholog in zebrafish or in Takifugu rubripes, another fish species, in an expressed sequence tag (EST) database searches. Thus, we postulate that the two Tob genes found in mammals, chicken, and frog were produced by a duplication event that occurred after the divergence of amphibian and teleost lineages. However, further analyses are needed to reveal whether two Tob genes in amphibian species were produced by duplications of a single gene, chromosomal segments, or entire genomes.

Tob1 and Tob2 proteins have a highly conserved N-terminal region. In vertebrates, the sequence identity for the Tob N-terminal 117 amino acid region, which includes box A and box B domains that are conserved in all BTG/Tob family proteins (see below), ranges from 70.1 to 100% (Fig. 1A,B). In this region, amphioxus and Drosophila Tob proteins share a sequence identity of greater than 57.3% with vertebrate Tob1 and Tob2. However, Fog-3, a Tob homolog first identified in C. elegans (Chen et al.,2000), shares a sequence identity with other Tob subfamily members of only 22.2–29.9% in the N-terminal region (Fig. 1B), and its sequence identity with BTG proteins in this region is even lower (data not shown), suggesting that of the known BTG/Tob proteins it may have diverged earliest from the common ancestor.

Figure 1.

Sequence similarity of N-terminal 117 amino acid region of Tob proteins. A: Sequence alignment of Tob subfamily members. Identical residues are shaded in black. The nuclear localization signal (NLS), as well as boxes A and B, is indicated. Sequence sources in GenBank: human TOB1 (hTOB1), P50616; human TOB2 (hTOB2), NP_057356; mouse Tob1 (mTob1), Q61471; mouse Tob2 (mTob2), NP_065253; chicken Tob1 (cTob1), NP_001001467; chicken Tob2 (cTob2), NP_001012552; Xenopus Tob1 (xTob1), AAH60329; Xenopus Tob2 (xTob2), NP_001016213; zebrafish Tob1a (zTob1a), NM_198363; zebrafish Tob1b (zTob1b), NM_212974; amphioxus Tob (aTob), AAB53747; Drosophila Tob isoform A (fTob), NP_573104; Caenorhabditis elegans Fog-3 (nFog3), NP_492646. B: Percentages of sequence identity among the various Tob subfamily members.

The box A and box B domains of the N-terminal region are the most conserved domains among BTG/Tob family proteins (Guehenneux et al.,1997). Among vertebrate proteins of the Tob subfamily, box A has a consensus sequence of KYEGHWYP(E/D)KP(Y/L)KGSG(F/Y)RC(I/V), and box B has a consensus sequence of (L/V)P(Q/E)(D/E)LSVWIDPFEVSYQIGE; these two domains are separated by a less conserved spacer sequence of 27 residues (Fig. 1A). The box A and box B domains of amphioxus and Drosophila Tob proteins also have a high homology with those in vertebrate Tob proteins; however, the equivalent nematode Fog-3 domains are less conserved. The N-terminal region of Tob/BTG proteins is important for forming complexes with target proteins, and for exerting biological effects (Rouault et al.,1998; Prevot et al.,2001; Yoshida et al.,2003b; Kawate et al.,2005; Xiong et al.,2006); the box B domain plays a particularly critical role in forming protein complexes and mediating biological activity (Ikematsu et al.,1999; Xiong et al.,2006).

Tob proteins are distributed throughout the cytoplasm and nucleus, but at higher levels in the nucleus, of various cell types. Kawamura-Tsuzuku et al. (2004) demonstrated that, during the cell cycle, subcellular localization of TOB1 is dynamic with higher levels found in the cytoplasm during late S phase than during other phases of the cell cycle. The nuclear localization of TOB1 is mediated by a bipartite nuclear localization signal (NLS), conserved in all vertebrate proteins of the Tob/BTG family, having a consensus sequence of RRR-X12-KKK at amino acid positions 22–39 (Fig. 1A). In contrast, human TOB1 may contain up to four nuclear export signals (NES; Kawamura-Tsuzuku et al.,2004; Maekawa et al.,2004). Kawamura-Tsuzuku et al. demonstrated that mutations in the TOB1 NLS result in preferential distribution of the mutant protein to the cytoplasm, and significant reduction of its antiproliferative activity in NIH3T3 cells (Kawamura-Tsuzuku et al.,2004). However, Maekawa et al. found that the addition of the potent SV40 NLS to the N-terminus of human TOB1 leads to exclusive nuclear accumulation of TOB1, but also results in reduced antiproliferative activity for this protein (Maekawa et al.,2004). Perhaps the additional NLS placed at the N-terminus of TOB1 causes a change in the conformation of TOB1, thus affecting its proper functioning. Amphioxus and Drosophila Tob proteins contain the first conserved cluster (RRR) of the bipartite vertebrate Tob NLS, but lack the second conserved cluster; nematode Fog-3 proteins does not contain any of the conserved vertebrate Tob NLS sequences. It is not known whether Tob proteins are present and functional in the nuclei of invertebrates.

Levels of Tob/BTG proteins change dynamically during the cell cycle (Fig. 2). For instance, in NIH3T3 cells, Tob1 protein is detectable during the resting phase (G0); its level declines dramatically during the G1 through S phase; and its expression resumes during the G2 phase (Suzuki et al.,2002). The Tob1 protein must be degraded rapidly during early G1 to ensure progression through late G1 phase. Treatment with proteasome inhibitors prevents the degradation of endogenous TOB1 protein in HeLa cells and inhibits the degradation of transiently expressed TOB1, TOB2, BTG1, and BTG2 in HEK293 cells (Sasajima et al.,2002), indicating that TOB/BTG proteins are degraded through the proteasomal pathway; the instability of TOB1, TOB2, BTG1, and BTG2 may be conferred by the C-terminal 60 amino acid region (Sasajima et al.,2002).

Figure 2.

Illustration of dynamic changes in Tob1 protein levels during the cell cycle in NIH3T3 cells. Tob1 levels are indicated by a gray gradient. Serum depletion allows the cell to withdraw from the cell cycle, and enter resting phase (G0). The cell cycle resumes after the serum supply is replenished. Tob1 levels are high during cell division. During resting phase (G0), Tob1 is detectable; however, Tob1 levels decline dramatically during G1 through S phase; Tob1 expression resumes during G2 phase. This illustration is based on the findings of Suzuki et al. (2002).

TOB1 is subject to phosphorylation at multiple sites. Suzuki et al. demonstrated that TOB1 is phosphorylated on serine/threonine residues, but not on tyrosines, and that receptor tyrosine kinase-activated p90rsk1 interacts with, and phosphorylates, TOB1 (Suzuki et al.,2001). Subsequently, they found that Ser152, Ser154, and Ser164 of TOB1 are rapidly phosphorylated by ERK1 and ERK2 upon growth-factor stimulation (Suzuki et al.,2002). Maekawa et al. also showed that ERK2 and JNK2 can bind to and phosphorylate TOB1 (Maekawa et al.,2002). Significantly, both groups reported that TOB1 phosphorylation negatively regulates the antiproliferative activity of TOB1 (Suzuki et al.,2001,2002; Maekawa et al.,2002). Indeed, Ser152, Ser154, and Ser164 and neighboring residues of TOB1 are conserved in other vertebrate Tob1 and Tob2 orthologs as well as in amphioxus Tob, suggesting the importance of these phosphorylation sites for repressing Tob antiproliferative activity.

Expression of Tob Genes

Expression patterns of Tob genes during early embryonic development have been studied in several species. Transcripts of zebrafish tob1a (Xiong et al.,2006) and tob1b (Shi et al.,2004) and Xenopus xTob2 (Yoshida et al.,2003b) are ubiquitously distributed during blastula and gastrula stages of development, suggesting a role in early embryogenesis in these species. However, in amphioxus, Tob transcripts were not detected by whole-mount in situ hybridization before late gastrulation (Holland et al.,1997). This finding may be due to low sensitivity of the detection method used, or may reflect a species divergence. Expression patterns of Tob genes during blastulation and gastrulation have not been reported for other species.

At later stages of embryonic development, the expression of Tob genes occurs in distinct domains, such as the notochord, hatching gland, blood islands, and gut, depending on the species. Of interest, during segmentation, Tob genes are expressed in somites, which ultimately give rise to axial skeleton, skeletal muscle, and dermis (Pourquie,2001), in amphioxus (Tob; Holland et al.,1997), zebrafish (tob1b; Shi et al.,2004), Xenopus (xTob2; Yoshida et al.,2003b), and mouse (Tob2; Ajima et al.,2000). In adults, mouse Tob1 (Yoshida et al.,1997) and Tob2 (Ajima et al.,2000) and human TOB1 (Matsuda et al.,1996) and TOB2 (Ikematsu et al.,1999) are expressed in skeletal muscle. In developing limbs of embryonic day (E) 18.5 mouse embryos, Tob1 transcript are also present in osteoblasts at high levels, in hypertrophic chondrocytes at moderate levels, and in osteoclasts at low levels (Yoshida et al.,2000). Furthermore, Park et al. showed that the TOB1 protein is stably expressed in the proliferating basal layer of the epidermis and in primary human keratinocytes (Park et al.,2006). Taken together, these expression data support the hypothesis that Tob genes play roles in development of the somites and their derivatives.

Some Tob genes are also expressed in the central nervous system (CNS). For example, Tob is expressed in the embryonic CNS of Drosophila (Bourbon et al.,2002), and amphioxus Tob is expressed in the nerve cord during the late neurula and larva stages (Holland et al.,1997). However, putative functions for Tob genes in neurogenesis remain unknown. Tob transcripts are also present in adult mouse (Yoshida et al.,1997; Ajima et al.,2000), rat (Jin et al.,2005), and human (Matsuda et al.,1996; Ikematsu et al.,1999) brain tissues.

Apparently, Tob genes are commonly expressed in the reproductive system and germ cells. Zebrafish tob1a (Xiong et al.,2006) and tob1b (Shi et al.,2004) transcripts and Xenopus xTob2 transcripts are present in one-cell stage embryos, suggesting maternal expression of these transcripts. TOB1 and TOB2 are expressed in human testis and ovary (Matsuda et al.,1996; Ikematsu et al.,1999) and, in mice, Tob1 (Yoshida et al.,1997) and Tob2 (Ajima et al.,2000) are expressed in the ovary and testis, respectively. Expression of the nematode fog-3 gene in germ cells has also been well studied. In C. elegans hermaphrodites, fog-3 is highly expressed in the germ cells of larvae at the L3 and L4 stages, when sperm are produced; fog-3 expression declines in young adult C. elegans as they begin to produce eggs (Chen and Ellis,2000; Chen et al.,2000). The same expression pattern is observed in hermaphrodites of C. briggsae (Chen et al.,2001). In C. remanei, which are not hermaphrodites, larval and adult males express high levels of fog-3, whereas females do not express detectable levels of fog-3 mRNA at the equivalent stages (Chen et al.,2001). Thus, nematode fog-3 expression is well correlated with spermatogenesis.

Currently available data also indicate that Tob genes are expressed in many other adult tissues and organs. Because the expression patterns of Tob genes are not studied in similar depth in each species, it is difficult at present to draw any further conclusions concerning the evolutionary conservation of Tob gene expression.

Tob Genes in Germ Cell Differentiation

The role played by Tob proteins in germ cell differentiation and gamete production has been best studied in nematodes. Germ cells of the hermaphrodite gonads of nematodes develop into sperm during larval stages; however, the remaining undifferentiated germ cells develop into oocytes in the adult nematode (Kimble and White,1981). In C. elegans, mutations in fog-3 cause XX hermaphrodites to produce oocytes rather than sperm during larval stages; XO males carrying fog-3 mutations also produce oocytes rather than sperm (Ellis and Kimble,1995; Chen et al.,2000). Similarly in C. briggsae, RNAi inhibition of fog-3 expression transforms both XX and XO animals into females that produce only eggs (Chen et al.,2001). On the other hand, it appears that loss of function of fog-3 has no effect on the sexual identity of other tissues in these animals. It seems that fog-3 is required, specifically and absolutely, for initiating spermatogenesis in the germ cells of nematodes. However, the mechanisms underlying Fog-3 regulation of the fate decision of germ cells in nematodes remain unclear. In a mutagenesis screen, 6 of 10 fog-3 mutant alleles were found to contain a missense mutation in the N-terminal conserved domain, indicating that this domain is important for the sex determination function of this protein.

Although Tob genes in vertebrate species are also expressed in testis and ovary, the role these genes play in the proliferation and differentiation of vertebrate germ cells remains unknown.

Tob Genes in Embryonic Development

The important role played by Tob genes in the development of invertebrate embryos clearly is apparent in Drosophila. A transposon insertion in Tob results in Drosophila embryos that are recessive lethals (Bourbon et al.,2002). However, no specific developmental defects were reported in this Tob-deficient line.

We have investigated the functions of Tob genes during embryogenesis in detail in zebrafish (Xiong et al.,2006). Overexpression of tob1a in zebrafish embryos can cause embryonic ventralization, including loss of the head and the notochord, whereas translational inhibition of endogenous tob1a mRNA by an antisense morpholino results in dorsalized phenotypes, such as loss of the ventral fin and a shorter twisted tail (Xiong et al.,2006). These results indicate that tob1a inhibits dorsal development and promotes ventral development in zebrafish embryos. Biochemical analyses reveal that Tob1a can inhibit transcriptional stimulation by β-catenin, a factor that is essential for the dorsal development of amphibian and fish embryos (Schneider et al.,1996; De Robertis and Kuroda,2004); Tob1a competes with the Lef1/Tcf cofactors for binding to β-catenin, effectively blocking formation of a Lef1/Tcf–β-catenin protein complex that can stimulate the transcription of several genes. Knockdown of tob1a is unable to rescue the developmental defects caused by antisense morpholino knockdown of the expression of β-catenin-2, a maternal gene that is essential for the dorsal development of zebrafish embryos (Bellipanni et al.,2006), indicating that tob1a acts upstream of maternal β-catenin. In addition, Tob1a is able to inhibit transcriptional activity of Smad3, an effector of transforming growth factor-beta (TGFβ)/Nodal signals (Tian and Meng,2006), by physically interacting with Smad3 and preventing it from binding to one of its cofactors, p300. Morpholino knockdown of tob1a in zebrafish embryos enhances expression of mesodermal and endodermal markers, suggesting that tob1a may antagonize the mesendodermal induction activity of Nodal signals. We have also demonstrated that zebrafish tob1b, as well as mouse Tob1 and Tob2, ventralize zebrafish embryos, supporting the idea that the functions of Tob genes are evolutionarily conserved.

Yoshida et al. demonstrated that Xenopus Tob2 (xTob2) physically interacts with Smad6 and Smad7 (Yoshida et al.,2003b). Smad6 and Smad7 can each inhibit bone morphogenetic protein (BMP) signaling in Xenopus embryos (Casellas and Brivanlou,1998; Hata et al.,1998; Nakayama et al.,1998). Although injection of xTob2 mRNA into ventral blastomeres does not induce secondary axes, coinjection with Smad6 mRNA results in a significantly higher ratio of embryos having a secondary axis, and a dorsalized phenotype, than does injecting Smad6 mRNA alone (Yoshida et al.,2003b), suggesting that xTob2 can enhance the ability of Smad6 to inhibit BMP signaling on the ventral side of embryos, consequently inhibiting ventral development. However, the inhibition of xTob2 expression by antisense morpholinos failed to induce significant abnormalities in early Xenopus embryos; this finding may be due functional overlap with other Tob proteins expressed in Xenopus embryos (Yoshida et al.,2003b).

Because Tob proteins have the ability of antagonizing both Bmp and β-catenin signaling, the difference in effects of overexpression of Tob genes in zebrafish and Xenopus might be due to the positions at which the mRNAs were injected into the embryos. In Xenopus, ventral injection of xTob2 mRNA should inhibit Bmp signaling without affecting β-catenin signaling on the dorsal side of the embryo, resulting in dorsalized phenotypes. In zebrafish, however, the inhibition of tob1a expression results in dorsalized phenotypes, whereas injections of other tob gene mRNAs into one-cell zebrafish embryos lead to ventralized phenotypes (Xiong et al.,2006), indicating that Tob1a acts as a ventralizing factor in zebrafish embryos. It remains to be elucidated whether endogenous Tob proteins in Xenopus embryos primarily inhibit dorsal or ventral development. Information on the distribution of Tob proteins along the dorsoventral axis during development would greatly clarify the roles of these proteins in dorsoventral patterning.

Tob1 in Bone Formation

The roles played by Tob genes in bone formation and resorption have been well studied in mice. Although Tob1 knockout (Tob1−/−) mice are born without apparent phenotypic abnormalities (Yoshida et al.,2000), Tob1-deficient adult mice have a higher amount of bone mass than do wild-type mice; this finding is due to increased numbers of osteoblasts and an accelerated rate of bone formation in the Tob1-deficient mice. Because BMP2-induced osteoblast proliferation and differentiation is enhanced in the calvariae of Tob1−/− mice, and Tob1 expression is induced by BMP2, Tob1 normally acts as a negative feedback regulator of BMP2 signal during bone formation (Yoshida et al.,2000; Usui et al.,2002). Mechanistically, Tob1 binds to the downstream effectors of BMP signaling, including Smad1, 5, 8, and 4, and promotes their localization to nuclear compartments, which may result in the suppression of the activity of these Smad transcription factors.

Sex steroid hormones, such as estrogen and androgen, are important signals for bone metabolism (Manolagas et al.,2002; Seeman,2004). It has been reported that the mineral apposition rate and the bone formation rate in the ovariectomized Tob1-deficient mice are significantly higher than those in the ovariectomized wild-type mice, suggesting that bone loss due to estrogen depletion can be compensated for by enhancement of osteoblastic activity in Tob1−/− background (Usui et al.,2004). Recently, Kawate et al. (2005) showed that overexpression of TOB1 in MC3T3-E1 osteoblast cells inhibits the ligand-dependent transactivation of androgen receptor (AR) or estrogen receptor, and that Tob1 binds to AR, preventing androgen/AR complexes from forming subnuclear foci, suggesting that Tob1 controls bone formation, not only by suppressing BMP signaling, but also by inhibiting sex hormone signaling in osteoblast cells.

TOB1 in T-Cell Activation

Human TOB1 is highly expressed in anergic and quiescent T cells, whereas its expression in primary unstimulated T lymphocytes is repressed during activation by the T-cell receptor signaling (Tzachanis et al.,2001). Induced expression of the VP22-TOB1 fusion protein inhibits T-cell activation caused by mitogenic stimulation and blocks the cell cycle at the G1 phase due to decreased expression of the positive cell cycle regulators cyclin E and Cdk2 as well as increased expression of the negative cell cycle regulator p27kip1 (Tzachanis et al.,2001). interleukin-2 (IL-2), a cytokine that plays a critical role in T-cell activation, has a promoter that contains a negative regulatory element to which Smad transcription factors bind. Therefore, TOB1 may interfere with T-cell activation, at least in part, by enhancing the inhibitory effect of Smad-mediated TGFβ signaling on IL-2 transcription (Tzachanis et al.,2001).

The inducible poly(A)-binding protein (iPABP/PABPC4), a homolog of PABP/BABPC4, binds to the poly(A) tail of eukaryotic mRNAs and promotes mRNA translation (Grange et al.,1987; Yang et al.,1995; Kahvejian et al.,2005). The expression level of iPABP is low in normal resting human T cells and is rapidly elevated upon T-cell activation (Yang et al.,1995). Interestingly, TOB1 binds to PABP and iPABP (Okochi et al.,2005), which allows it to inhibit the promotion of IL-2 mRNA translation by iPABP. In NIH3T3 cells that stably express exogenous iPABP, cotransfection of IL-2 with TOB1 results in the suppression of IL-2 production. These data suggest that the expression of IL-2 in anergic T cells is also suppressed at the translation step by TOB1 through its interaction with iPABP (Okochi et al.,2005). Significantly, this study revealed a novel function for Tob factors in the cytoplasm, attenuating target mRNA translation.

Tob1 in Tumor Development

Tob proteins have antiproliferative potential, suggesting that they play important roles in suppressing tumor development. Yoshida et al. found that, at 18 months of age, 77% of Tob1−/− mice had spontaneously developed a variety of tumors, including hemangiosarcomas, lung carcinomas, and hepatocellular adenomas, whereas only 16% of wild-type mice at the same age had developed tumors: primarily malignant lymphomas and lung adenomas (Yoshida et al.,2003a). After treatment with diethylnitrosamine, a liver-specific carcinogen, the rate of liver tumor formation was significantly higher in Tob1−/− mice than in wild-type mice. These results indicate that mice lacking Tob1 are much more susceptible to cancer than wild-type mice are.

In humans, TOB1 is expressed at lower levels in lung cancer tissue than in adjacent lung tissues (Iwanaga et al.,2003; Yoshida et al.,2003a). Iwanaga et al. also noted that the TOB1 present in lung cancer tissue is in the phosphorylated form, which is inactive, whereas this inactive form of TOB1 is rare in the adjacent lung tissues (Iwanaga et al.,2003). Similar results have been obtained for thyroid neoplasm: high levels of phosphorylated TOB1 occur in papillary carcinoma, and TOB1 protein is not detected in anaplastic carcinomas (Ito et al.,2005). In squamous cell carcinoma, an invasive carcinoma of the surface epidermis, TOB1 protein levels are lower than in normal epidermis (Park et al.,2006). The association of certain types of human cancers with decreased expression or inactivation of TOB1 suggests that it may be an important tumor suppressor.

How do Tob factors inhibit cell proliferation and suppress tumor development? TOB1 was first identified as a binding partner of the c-erbB-2 gene product p185erbB2; kinase-active p185erbB2 alleviates TOB1-mediated suppression of cell proliferation (Matsuda et al.,1996). The authors of this study suggested that TOB1 inhibits cell proliferation through intervention in other oncogenic pathways. Like some BTG proteins (Rouault et al.,1998; Prevot et al.,2001), TOB1 and TOB2 can also interact with human Caf1 (Ikematsu et al.,1999); the yeast Caf1 homolog (CCR4), with other cofactors, can form transcriptional complexes that activate or suppress target gene transcription (Liu et al.,1998). CAF1 directly interacts with cyclin-dependent kinases (CDKs) cyclin E, cyclin A, cyclin B, and cyclin D1 (Bogdan et al.,1998; Ikematsu et al.,1999). It is possible that Tob proteins regulate cell proliferation by modulating the activities of certain CDKs through their interaction with CAF1 (Ikematsu et al.,1999). Yoshida et al. found that Tob-1–deficient cells have increased levels of cyclin D1 expression. They also found that Tob1 inhibits cyclin D1 promoter activity by interacting with and recruiting histone deacetylase 1 (HDAC1; Yoshida et al.,2003a). We demonstrated that Tob proteins bind to β-catenin, and inhibit its ability to stimulate transcription. Because the Wnt/β-catenin signaling pathway plays important roles in cell proliferation, and increased activity of this pathway has been linked to tumorigenesis (Moon et al.,2004), β-catenin may be one of the major targets of Tob proteins for exerting their antiproliferative effect. Thus, proteins of Tob subfamily may inhibit cell proliferation by means of multiple mechanisms.

Tob1 in Learning and Memory

Rat Tob1 is expressed in various segments of the brain, including the cerebellum, cortex, hippocampus, and brainstem (Jin et al.,2005). Its expression in the hippocampus can be induced transiently by contextual fear conditioning, which reaches a peak well before memory consolidation. When the expression of Tob1 is inhibited in the hippocampus CA1 region by specific antisense oligonucleotides, long-term contextual fear memory, as well as the spatial learning and memory, which all depend on a properly functioning hippocampus, are impaired while auditory fear memory, visual ability, or motivation are not affected (Jin et al.,2005). The defects in learning and memory coincide with the suppression of the long-term potentiation. These results provided the first insights into the involvement of Tob1 in learning and memory in mammals.

Additionally, Tob1 may function in motor learning in rats (Wang et al.,2006). Training rats on a rotating rod can induce increased Tob1 expression in the cerebellum. When a Tob1 antisense oligonucleotide is infused into the fourth ventricle of the rat brain, the rat's ability to run on a rotating rod or walk across a beam is impaired, suggesting that Tob1 is required for motor skill learning.

PERSPECTIVES

Tob genes have been isolated from vertebrate as well as invertebrate species. Accumulating evidence indicates that Tob genes play diverse roles in development, and in other biological processes. It is noteworthy that Tob genes are expressed in a wide spectrum of tissues and organs, particularly in adult animals, and that the biological functions of these proteins in most of these tissues and organs remain unknown.

Tob proteins exert their biological functions by regulating a variety of signaling pathways by means of differing mechanisms (Fig. 3). Accumulating evidence indicates that Tob proteins usually bind to nuclear factors and negatively regulate the ability of these factors to stimulate transcriptional activity. For instance, zebrafish Tob1a and Tob1b, and mouse Tob1 and Tob2 all can bind to and repress the ability of β-catenin or Smad3 to stimulate transcription of target genes (Xiong et al.,2006); mouse Tob1 also binds to Smad1, Smad5, and Smad8 and inhibits their ability to stimulate gene expression (Yoshida et al.,2000); and human TOB1 can also bind to the AR, repressing its transactivation activity (Kawate et al.,2005). The interaction between TOB1 and Smad4 in primary peripheral T cells is an exception: TOB1 actually enhances the DNA binding affinity of Smad4 (Tzachanis et al.,2001). Although Tob proteins are present in both cytoplasm and the nucleus, little is known about the mechanisms that regulate the shuttling of Tob proteins between the cytoplasm and nucleus, because extracellular and intracellular signals that regulate the movement of Tob proteins between the cytoplasm and nucleus have not been identified.

Figure 3.

A summary of Tob functions and underlying mechanisms. A Tob protein binds to a factor that plays a role in a particular signaling pathway; this finding allows Tob to regulate particular biological process by modulating the corresponding signaling pathway. An arrow or a line with a bar indicates the positive or negative regulation of a particular process, respectively. A single line indicates that both positive and negative regulations are involved.

In vertebrates Tob genes are often coexpressed during development. Considering that Tob1 and Tob2 proteins share a high degree of sequence homology and biochemical properties, it is reasonable to expect that they may have redundant functions in tissues in which both are expressed, making it difficult to precisely determine their corresponding roles in particular biological processes in these tissues. For instance, although mouse Tob1 is expressed maternally and continuously throughout embryonic development (Yoshida et al.,1997), suggesting an important role in embryonic development, Tob1−/− embryos develop normally (Yoshida et al.,2000). One possible explanation for this finding is that Tob2, which is expressed ubiquitously in mouse embryos (Ajima et al.,2000), and possibly BTG proteins expressed during embryogenesis, can compensate for the Tob1 deficiency. If true, this finding suggests that the functions of Tob/BTG family members are so critical during embryogenesis that several of these genes have evolved similar, overlapping functions.

It appears that Tob proteins clearly function as important negative regulators of the cell cycle and have significant antiproliferative activities. A question of fundamental importance to biologists is how the regulation of cell proliferation by Tob proteins is coordinated with their roles in other biological processes, including cell differentiation and cell movements, during embryogenesis. On one hand, Tob proteins associate with positive regulators of the cell cycle to block the progression of the cell cycle. On the other hand, they also interact with, and inhibit, the activities of factors that regulate cell differentiation or cell movements or both. Future study is required to determine how the wide varieties of functions of Tob proteins are temporally and spatially coordinated.

Acknowledgements

We thank members of the Meng lab for helpful discussions of this manuscript. We also thank Dr. Michael Farrell for his comments.

Ancillary