| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Cell Biology (S.Y.T., M-J.T.) and Medicine (M-J.T.), Baylor College of Medicine, Houston, Texas 77030
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
In this review, we attempt to provide an overview of COUP-TFs with respect to their structural homology to members of the steroid receptor superfamily, their molecular and biochemical characteristics as putative negative regulators, their expression patterns during development, their regulation, and, most importantly, their possible physiological functions.
|
II. The COUP-TF Gene Family |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
). A 66-amino acid region in
COUP-TF, which contains two conserved Zn-finger motifs, signified the
presence of the DNA-binding domain (DBD) of a nuclear receptor. In
COUP-TFs, all 20 invariant amino acids were conserved and 11 of 12
conserved residues were identical except that a conserved lysine in the
second finger is replaced by a glutamine (10). Based on the P-box
sequence, COUP-TFs can be classified as members of estrogen receptor
(ER)/TR subfamily which bind to a Pu-GGTCA repeat. There is only one
amino acid difference within the DBD of COUP-TFI and II, and it is a
conservative change from Ser to Thr (10, 13, 14, 15). In addition to
the Zn-finger DNA-binding motif, there are two regions, II and III, in
the ligand-binding domain (LBD) that share significant homologies
between various members of the steroid receptor superfamily (10, 16).
These two regions have been shown to be important for the formation of
a ligand-binding pocket (18). COUP-TFI and II are 96% and 100%
identical in these two regions. COUP-TFs are most related to the
retinoid X receptor (RXR)/retinoid acid receptor (RAR) subfamily
members and share the highest homology with RXR
, having 67%, 48%,
and 43% in the DBD, region II and region III, respectively (Fig. 1).
The homology is slightly reduced in comparison to various members of
the RAR subfamily, but is greatly decreased when compared with members
of the steroid receptor family, such as mineralocorticoid receptor
(MR), glucocorticioid receptor (GR), androgen receptor (AR), and
progesterone receptor (PR) (Fig. 1).
|
. It is clear that
COUP-TFs are most closely related to EAR2 and members of the RXR
family. The groups with the second highest percentage of homology are
HNF4 and TLX. Further apart in sequential order are other orphan
receptors, TR2, TR4, SF1, LRH1, ERR1, and ERR2, with ER as the only
liganded receptor in these groupings.
|
Sequence comparison between various COUP-TF genes clearly shows that
they belong to a distinct subfamily within the steroid/TR superfamily.
The DBDs within different members of COUP-TF minimally share 94% amino
acid sequence identity, which is considerably greater than the
conservation between other subfamily members. Based on homology
alignments of the LBDs, vertebrate COUP-TFs can be subdivided into four
groups (28) (Fig. 3). The first group includes
hCOUP-TFI/EAR3, mCOUP-TFI, hamCOUP-TFI, rCOUP-TFI, xCOUP-TFI, and
zsvp[44]. The second group includes hCOUP-TFII/ARP-1, mCOUP-TFII,
cCOUP-TFII, xCOUP-TFII/xCOUP-TFB, and zsvp[40]. xCOUP-TFIII/xCOUP-TFA
is classified as the third group because it is equally homologous
(90%) to the subgroup of COUP-TFI and COUP-TFII. The
zCOUP-TFIV/svp[46] is considered as group 4 because it shows similar
homology (90%) to the other three groups. Invertebrate dsvp and
suCOUP-TF are slightly more distantly related to other vertebrate
COUP-TF members. Although EAR2 is the next closely related gene, its
homology is much reduced in comparison to members of the COUP-TF
subfamily (70% vs. >90%). In addition, the duplication
between COUP-TFs and EAR2 was completed after the division of these
early metazoans and the most modern ones such as arthropods and
vertebrates (Laudet, personal communication). Therefore, EAR2 is not
classified as a COUP-TF subfamily member and will not be discussed in
this review.
|
). The DBDs of COUP-TFI or II in
different species are identical, implying that they bind to a similar
or an identical response element (16). Most surprisingly, the putative
LBDs of COUP-TF I and II are 99.6% identical among vertebrates (16).
Even svp from Drosophila shares 90% identity. Such a high
degree of homology suggests that these domains are critical for the
biological function of COUP-TFs, even though a ligand or a specific
interacting protein unique for COUP-TFs has not been identified. In
contrast, the N-terminal domain of COUP-TFI and II are significantly
diverged, having only 45% identity, which may provide distinct
functions for the two different members (10, 13, 14, 15). There are
clusters of proline- and glutamine-rich regions that resemble the
activation domains of transcription factors such as CTF and SP1,
respectively. These domains also contain multiple consensus
phosphorylation sites for protein kinase A and proline-directed kinase,
consistent with the finding that COUP-TFs are phosphorylated proteins.
|
|
III. Biochemical Characteristics |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
To systematically analyze the effect of spacing and orientation of half-sites on the relative binding affinity of COUP-TFs, a series of oligonucleotides containing a direct repeat of the core half-site, GGTCA separated by a 0, 2, 5, 7, 9, or 12 nucleotide spacing, were synthesized and used as competitors in a gel mobility shift binding assay (31). The relative binding affinity for the direct repeats with different spacing is as follows: DR1, DR6, DR4, DR8, DR0, and DR11. COUP-TFs also bind to inverted and everted repeats of the consensus sequence (31). In general, COUP-TFs display higher binding affinity for direct repeats in comparison to the inverted palindromic sequence. Among the palindromic sequences, COUP-TFs bind to the consensus element with 0 spacing with the highest affinity, yet their binding is about 3-fold lower than the binding affinity of the DR1 element. As expected, homo- and heterodimers of COUP-TFI and II bind to the different repeats with similar relative affinity, consistent with the function of a nearly identical DBD of the two factors. Taken together, COUP-TFs display highest preference to bind DR1 elements (31).
In fact, the most common COUP-TF-binding site found in natural promoters is the DR1 consensus sequence, which is an AGGTCA direct repeat with one nucleotide spacer. COUP-TFs have been shown to bind and repress DR1 consensus-regulatory elements in the promoter of many genes, including rat and human apolipoprotein CIII (32, 33, 34), human Apolipoprotein AI (32, 33), chicken apolipoprotein VLDLII (35, 36), mouse lactoferrin (37, 38), mouse mammary tumor virus promoter (39), and mouse OCT4 promoter (40). In addition to the DR1 element, COUP-TFs can also bind to the DR0 sequence in oxytocin (41) and hemopexin promoters (42), the DR2 element in the sea urchin actin III B gene (27), the DR6 RIPE-1 element of the rat insulin 2 promoter (30), the DR7 of arrestin gene (43), the DR9 of HIV-LTR (44), and the everted repeats of eight- and 14-nucleotide spacings of the acyl-coA dehydrogenase gene promoter (45).
Biochemical studies indicate that COUP-TFs exist in solution as dimers (9, 12, 46) and have been shown to bind to the consensus response elements as dimers in gel mobility shift assays (9, 46). The ability of the dimeric COUP-TFs to bind to diverse structural motifs, such as direct repeats, inverted repeats, and everted repeats with a variety of spacings, suggests that COUP-TFs must be able to assume different conformations to accommodate the structural and spatial changes in the recognition sequences (31, 47). Indeed, the COUP-TF dimer bound to DR1 or DR6 elements possesses distinct conformations, as shown by its different susceptibilities to protease digestion (31). The binding versatility of COUP-TFs for direct repeats has tremendous biological implications because many of these binding sites are also response elements for the retinoid (DR2 and 5), thyroid hormone (DR4), and vitamin D (DR3) receptors (46, 48). Retinoids, thyroid hormones, and vitamin D are well known hormones and morphogens for vertebrate development and differentiation. Therefore, it is of great interest to delineate the molecular mechanism by which COUP-TFs modulate the processes of cellular development and differentiation.
B. Molecular mechanism of COUP-TFs action
Vitamin D receptor (VDR), TR, and RAR have been demonstrated to
activate target genes containing DR3, DR4, and DR5 response elements,
respectively. By virtue of their promiscuous DNA binding, COUP-TFs are
expected to down-regulate the hormonal induction of target genes by
VDR, TR, and RAR. Cotransfection of COUP-TFI or II expression vectors
inhibited the hormonal induction of VDR-, TR-, and RAR-dependent
activation of reporter activity (31, 46, 47, 49). The inhibition of
transcriptional activity by COUP-TFs is dose-dependent, as the reporter
activity is progressively inhibited by increasing concentrations of the
transfected COUP-TF expression vector. COUP-TFs not only inhibit the
hormone response of reporters containing synthetic AGGTCA repeats with
various spacings, they also inhibit the expression of reporters
containing natural vitamin D response element (VDRE), thyroid response
element (TRE), and retinoic acid response element (RARE) sequences as
in the respective cases of the osteocalcin, myosin heavy chain, and
ßRAR promoters (46). In addition, COUP-TFs have been shown to
antagonize the HNF4-dependent transcriptional activation of many
liver-specific genes (33, 50, 51) and to suppress OCT3/4 expression
during retinoid-induced differentiation of P19 embryonic carcinoma
cells (52). How do COUP-TFs inhibit the transactivation of other
members of the steroid receptor superfamily? Four mechanisms are
proposed, which are the subject of discussion here (Fig. 5).
|
2. Competition for RXR. It is well documented that RXR is a universal heterodimeric partner of RAR, TR, VDR, PPAR, and other orphan receptors (57, 58, 59, 60, 61). Homodimers of RAR, TR, VDR, and PPAR either bind poorly or not at all to their cognate response elements (3). Through association with RXR, the heterodimeric receptors can then bind to the cognate response elements with high affinity and, thus, enhance the transactivation potential of this group of receptors (3). Because the direct repeat recognition sequence is asymmetric, it has been shown that RXR occupies the 5' half-site while the other partner occupies the downstream 3' half-site, which confers the hormone responsiveness (62, 63). RXR can also bind to DR1 elements as a homodimer and as a heterodimer with RAR and PPAR (60, 61). The RXR homodimer is an activator that responds to 9-cis-retinoic acid. RXR/PPAR heterodimers respond to both 9-cis-retinoic acid- and PPAR-specific ligands (61). However, RAR/RXR heterodimers, in which RAR binds to the 5' half-site, are transactivationally inactive. It has been shown that RAR and TR bind to a corepressor [either silencing mediator for retinoid and thyroid hormone receptors (SMRT) or nuclear receptor corepressor (N-CoR)] in the absence of hormone (64, 65). Binding of these corepressors is necessary for receptors to silence the promoter activity. Binding of hormone then releases the corepressor and, thus, abolishes silencing activity of receptors. However, when RAR/RXR binds to DR1, the retinoic acid ligand is not able to release the corepressor from RAR; therefore, RAR/RXR heterodimer is not able to activate the DR1 reporter (63).
Although COUP-TFs exist in solution as homodimers and fail to form stable heterodimers with RXR in coimmunoprecipitation assays (12, 31), they do readily form DNA-binding heterodimers with RXR (31, 47, 49). Therefore, COUP-TFs are able to sequester the universal partner RXR in a functionally inactive complex and reduce the available concentrations of RXR (31, 46, 47, 49, 66). The loss of RXR indirectly decreases the DNA-binding affinity of TR, VDR, RAR, and PPAR and thereby interferes with the potential of this subgroup of receptors to transactivate their target genes (46, 47, 61). This notion is further verified by the relief of COUP-TF inhibition when RXR is overexpressed (46). In addition, it has been demonstrated that COUP-TFs form heterodimers with TR and RAR and disrupt their functions (67, 68, 69). Thus, the ability of COUP-TFs to form heterodimers with RXR, TR, and RAR may contribute significantly to the negative regulatory role of COUP-TFs in modulating hormone responsiveness of a large number of receptors of the TR and RAR subfamily (3).
Svp, the Drosophila homolog of COUP-TF, has also been shown to modulate the in vivo and in vitro function of Ultraspiracle (Usp), the Drosophila homolog of RXR (70, 71). Usp is the heterodimeric partner of ecdysone receptor. Like COUP-TFs, Svp binds to the consensus response element of ecdysone receptor and competes with the ecdysone/Usp heterodimer for the same binding site as well as forms heterodimers with Usp; thereby, it interferes with the signaling pathway of the ecdysone receptor in Drosophila (72). The induction of lethality during early metamorphosis by ectopic expression of Svp and the reversal of lethality by concomitant overexpression of Usp are consistent with the hypothesis that Svp negatively modulates the ecdysone signaling pathway in Drosophila in a manner similar to COUP-TFs’ modulation of thyroid hormone and retinoid function in mammalian systems.
3. Active repression. Like unliganded TR, COUP-TFs have been shown to repress basal transcriptional activity of a number of thymidine kinase reporters containing DR3, DR4, or DR5 hormone response elements (31, 46). This silencing of basal transcriptional activity is response element specific and is unlikely due to squelching of TFIIB, which interacts with COUP-TFs or other general transcription factors, since reporter genes lacking COUP-TF binding sites show little COUP-TF-mediated repression (69). Subsequently, it has been demonstrated that COUP-TFs, similar to TR and RAR, possess an active silencing domain within the C terminus of the putative LBD (69). This repressor domain can be transferred to a heterologous GAL4 DBD and can be shown to retain its ability to repress basal transcriptional activity (69).
In addition, we have recently shown that COUP-TFs can function as an active repressor to inhibit transactivation mediated by acidic (Gal4-RII), glutamine-rich (Gal4-ftzQ), proline-rich (Gal4-CTF1P), and Ser/Thr-rich (Gal4-ZenST) transactivators (69). The active repressor function of COUP-TFs is position independent, i.e. the binding sites of COUP-TFs can either be localized upstream of the activator binding site or downstream of the reporter gene without significantly affecting the active repression. The fact that COUP-TFs can repress such diverse groups of transactivators suggests that it is unlikely due to COUP-TFs directly quenching these transactivators or interfering with their interaction with their respective targets. Perhaps it is more likely that COUP-TFs interact with a common target, a putative corepressor that mediates the repression. Therefore, it is possible that COUP-TFs can interact with cellular repressors, such as SMRT and N-CoR, to silence basal and active transcription in a manner similar to RAR and TR (64, 65). Whether SMRT or N-CoR mediates the active silencing activity of COUP-TFs has yet to be defined.
4. Transrepression. As discussed earlier, COUP-TFs form heterodimers with RAR, TR, and RXR (46, 47, 49, 69). The dimerization is presumably mediated through interactions between the LBDs of COUP-TFs and these receptors. We have recently demonstrated that heterodimeric interactions can take place between DNA-bound wild type TR, RAR or RXR, and Gal4-COUP-TF (LBD). This dimeric interaction is sufficient for subsequent inhibition of the basal or activated promoter activity (69). Therefore, COUP-TFs can be tethered to DNA in the absence of their cognate response elements via LBD-LBD interactions with other receptors such as TR, RAR, and RXR to transrepress the ligand-dependent transactivation of the above nuclear receptors.
5. The role of COUP-TFs in transactivation. Although accumulated evidences indicate that COUP-TFs function as negative regulators in transient transfection assays, COUP-TF was initially found as an activator of chicken ovalbumin gene expression. COUP-TF has been shown to stimulate the transcription of transferrin promoter in Hep3B cells, but not in Sertoli cells (55). In addition, it has been shown that COUP-TF can activate fatty acid-binding protein (73), mouse mammary tumor virus (74), vHNF1 (75), and ornithine transcarbamylase (76) promoter activities. Recently, Hall and Sladek (77) demonstrated that COUP-TF and HNF-4 bind to the AF-1 element in the PEPCK gene and serve as accessory factors to augment glucocorticoid response in activation of the PEPCK gene expression. However, the physiological importance of COUP-TFs as activators is unclear, inasmuch as some of the activity has been observed only when a response element was analyzed out of the context of its promoter. In addition, the inducibility is low in general, and it is only observed in a few particular cell types.
|
IV. Expression Patterns of COUP-TFs During Development |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
Both COUP-TFI and II are widely expressed in the CNS and in many organs, with patterns that are overlapping yet distinct from each other (16, 17, 19, 20, 21). In general, COUP-TFI expression is higher in the CNS and lower in internal organs as compared with COUP-TFII. Because the lower vertebrates include many members of COUP-TFs that are not found in the higher vertebrates, we will describe the expression patterns of the two groups separately. To simplify the presentation, we will primarily describe results obtained from one species and discuss others in parallel so that the results can be presented in more meaningful and less complex terms.
A. Expression patterns in zebrafish and Xenopus
COUP-TFI/Svp[44] transcripts are first detected in 11- to 12-h
embryos of zebrafish (25). In the rostral brain of 13-h embryos, it is
expressed within the anterior half of the midbrain and the posterior
part of the diencephalon (25). In the presumptive hindbrain, it appears
in a segment-like stripe in the anterior region, resembling the
presumptive rhombomere units of the hindbrain (25). It is also detected
in the intermediate mesoderm posterior to the first somite. As
somitogenesis proceeds, its expression extends posteriorly and flanks
the 10 most anterior somites (25). The expression changes extensively
both in level and expansion of domains between 13 and 20 h. In the
rostral brain, its expression extends to include a major part of the
diencephalon and a caudal portion of the telencephalon. Within the
hindbrain, it is strongly expressed in the two most anterior
rhombomeres, and a lower but uniform expression is seen to extend
throughout rhombomere 7. In 28-h embryos, higher and more uniform
expression of COUP-TFI/Svp[44] is seen in both rostral and hindbrain
areas. Also, COUP-TFI/svp[44] is expressed in the retina of the eye.
Similar to COUP-TFI/svp[44], COUP-TFII/svp[40] is expressed as early as the 10-h stage (26). By 12 h, it is expressed in two domains of the brain: one at the border between diencephalon and hindbrain and the other located just posterior to the optic vesicles (26). In contrast to COUP-TFI/svp[44], there is no expression in the presumptive telencephalon. It is also expressed in the anterior part of somites 3–6 (26). In 16-h embryos, the expression in the hindbrain is higher and more sharply demarcated, corresponding to the primordia of the six most rostral rhombomeres (26). In 20-h embryos, COUP-TFII/svp[40] is expressed as a step gradient with specific levels in each of the first six rhombomeres and the expression domains corresponding directly to rhombomeres (26). The expression of COUP-TFII/svp[40] in the rhombomeres resembles that of xCOUP-TFB, having higher expression in r1, r2, r5, and r6 than in r3 and r4 in the hindbrain and very little expression in the major part of the midbrain (26).
The third COUP-TF member in zebrafish, COUP-TFIV/svp[46], is expressed in the primordia of the diencephalon, midbrain, hindbrain, and the anterior part of the spinal cord at the 11-h stage (25). In the hindbrain, it is expressed in the presumptive rhombomeres 1, 2, 4, and 5, but the expression is transient and is not detected by the 17-h stage (25). COUP-TFIV/svp[46] is also expressed in paraxial mesoderm in different developmental stages. In general, the expression is intense in the last somites being formed and in the anterior portion of each somite (25). Finally, COUP-TFIV/svp[46] is highly expressed in the optic stalk. The overall expression patterns of COUP-TFIV/svp[46] are very similar to xCOUP-TFIII/xCOUP-TFA, which is intensely expressed in rhombomeres 1, 2, 4, and 6 of the hindbrain and the eye anlagen (78). However, xCOUP-TFIII/xCOUP-TFA is also highly expressed in telencephalon (78).
B. Expression patterns in mouse and chick.
COUP-TFI and II from mouse and COUP-TFII from chick are expressed
in the developing CNS (19, 21). The expression patterns of COUP-TFI and
II are extremely similar to that observed with svp[44] and svp[40]
of the zebrafish, respectively (25, 26). Expression of mCOUP-TFI and II
was first detected at 7.5 days post coitus (p.c.), peaked at 14–15
days p.c., and declined sharply before birth (19). Whole-mount staining
of mCOUP-TFI at 8.5 days p.c. indicates that it is expressed in
specific regions of the rostral brain marked by sharp expression
domains and two intensively expressed stripes in the presumptive
hindbrain, correlating well with migratory neural crest cells (Y. Qiu,
F. A. Pereira, F. J. DeMayo, J. P. Lydon, S. Y. Tsai, and M.-J. Tsai,
unpublished results). It is also expressed in the anteriormost somites
(X. Qiu et al., unpublished results). In contrast, the
expression domains of COUP-TFII are much less defined in both the
rostral and hindbrain, and expression is seen in all somites (Y. Qiu,
F. Pereira, M.-J. Tsai, and S. Y. Tsai, unpublished result).
As development proceeds, mCOUP-TFI and II establish overlapping yet distinct patterns of expression (19). At 13.5 days p.c., COUP-TFI expression is seen throughout the pallium, while COUP-TFII is restricted to the caudal region. Interestingly, mCOUP-TFs are expressed in a segmental fashion in the diencephalic neuromeres. According to Figdor and Stern (80), the neuromeres are designated as D1 (ventral- and hypothalamus), D2 (dorsal thalamus), and D3 and D4 (pretectal region). Both mCOUP-TFI and II are expressed at equally high levels in D1 but at low levels in D3 and D4. However, high levels of mCOUP-TFI, but not mCOUP-TFII, are seen in D2 (19).
In the midbrain, mCOUP-TFI is distributed in the tectum with an anterioposterior gradient, with highest intensity at the rostral end, while the expression of mCOUP-TFII is limited to the rostral third of the tectum (19). mCOUP-TFII is also expressed at high levels in the ocular motor nucleus, whereas mCOUP-TFI transcripts are barely detectable. In the hindbrain, mCOUP-TFI is highly expressed in rhombomeres 1–5, and expression gradually declines within rhombomere 6 (Y. Qiu, F. A. Pereira, F. J. DeMayo, J. P. Lydon, S. Y. Tsai, and M.-J. Tsai, unpublished results). mCOUP-TFII is expressed highly in the anteriormost rhombomeres, and expression is significantly reduced in r4 and gradually increases to a higher level in r6 (Y. Qiu, F. Pereira, M.-J. Tsai, and S. Y. Tsai, unpublished observation). The hindbrain expression pattern is strikingly similar to its counterpart, svp[40] of zebrafish, suggesting the importance of COUP-TFs in hindbrain segmentation. In the spinal cord, mCOUP-TFI is expressed throughout the neural tube with slightly higher intensity in the motor neurons (19). In contrast, mCOUP-TFII expression is restricted to the motor neurons. In chick, it was also shown that cCOUP-TFII expression correlates with the differentiation, but not determination, of motor neurons (21).
In addition to the neuronal expression, mCOUP-TFs are also
differentially expressed in a restricted manner during organogenesis
(17, 20). mCOUP-TFI is expressed in the stroma of the nasal septum, in
the tongue, in the follicles of vibrissae, and in the cochlea (17).
mCOUP-TFII is expressed in the same regions, but at a considerably
lower level (Fig. 6) (17). In contrast, mCOUP-TFII is
highly expressed in the salivary gland, lung, esophagus, stomach,
pancreas primordium, mesonephros, prostate, and kidney and at lower
levels in testes, ovary, retina, limb bud, skin, and inner ear, while
mCOUP-TFI, with the exception of salivary gland, is expressed at much
lower levels (Fig. 6) (17). In general, in organs that require
mesenchymal and epithelial interactions, COUP-TFs are only expressed in
the mesenchymal cells of the developing organs, not in the terminally
differentiated epithelium. Thus, we hypothesize that COUP-TFs may be
important signals required for induction of epithelial differentiation.
|
|
V. Physiological Function of COUP-TFs |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
A. Ectopic expression of svp in Drosophila
It is known that svp in Drosophila is required to
specify photoreceptor subtype in the development of the compound eye,
preventing photoreceptors, R1/R3/R4/R6, from adopting R7 cell fate (24, 83). Ectopic expression of svp in cone cells converts the cone cells to
neuronal cells, and ectopic expression in other photoreceptor subtypes
maintains the neuronal characteristics but loses the specific subtype
identity (24). These results suggest distinct processes are required
for achieving neural and subtype identities and that svp plays a role
in the determination of both processes. Therefore, svp acts as a cell
fate switch, and the specific phenotype depends on the developmental
stage of the ommatidium at the time of svp expression (83). Molecular
and genetic studies suggest that ras signaling is required
for svp activity (84). However, the detailed mechanism has yet to be
defined.
B. Ectopic expression of COUP-TFI in Xenopus embryos
Misexpression of COUP-TFI dramatically affected early
Xenopus development (82). Overexpression of COUP-TFI in the
two-cell Xenopus embryos resulted in anterior truncation at
the tadpole stage. At four-cell embryos, overexpression of COUP-TFI in
the dorsal half, but not the ventral half, of the embryos led to
alterations in anterior development, including truncation of head
structure, loss of eyes, deletion of the cement gland, and malformation
of brain structure in a large percentage of tadpoles. The abnormal
early development might result from perturbation of transcription of
anterior early genes because expression of the anterior neural markers,
EN-2 and Krox-20, and the anterior neural crest markers, XDLL-1, XAP-2,
and Xtwi, were all greatly reduced when COUP-TFI was overexpressed
(82). Whether ectopic COUP-TFI expression will cause similar anterior
head developmental defects in higher vertebrates that mimic those
observed in Xenopus has yet to be defined.
C. Loss of function of COUP-TFI in mice
We have recently generated null mutants of COUP-TFI using
homologous recombination technology. The homozygous mutants died
perinatally between 8 and 36 h post birth (Y. Qiu, F. A. Pereira,
F. J. DeMayo, J. P. Lydon, S. Y. Tsai, and M.-J. Tsai, unpublished
results). The newborn mutant neonates appear highly dehydrated, lack
milk in their stomach, and contain air in the intestines. The apparent
symptoms resemble the phenotype of the knockout mice of neurotrophins
or their receptors, possibly resulting from an inability to feed (85, 86). Since COUP-TFI is highly expressed in the CNS, the nerve ganglia,
and the pharyngeal region, we examined the nerve ganglia of the null
mutants.
For this purpose, we used an antibody specific for the neurofilament heavy chain to probe for axonal pathways of cranial nerve ganglia. Among all the null mutants that have been examined, 95% of them have asymmetric fusion of the glossopharyngeal (IX) and vagus (X) cranial nerves. In addition, the arborization of axons, as seen by neurofilament staining, is severely affected in the cervical plexus regions (Y. Qiu et al., unpublished results). Therefore, lack of appropriate neuronal function due to the defect in these neurons may impair sensory and motor functions, which affect feeding behavior and result in perinatal death.
Retinoids are known to regulate the expression of Hox genes, which play
a major role in pattern formation and bone morphogenesis (87, 88, 89, 90).
Because COUP-TFI is hypothesized to antagonize retinoid function and
its expression is known to be regulated by retinoids, it is of interest
to assess whether loss of COUP-TFI function in the null mutants will
affect bone formation. We noted that the left or the right exoccipital
bone is fused to the basioccipital bone in 98% of the null mutants.
This result suggests that COUP-TFI plays a major role in the
development of these bones. Similar ossified fusions are also observed
in the mutant mice of the double knockout of the RAR1 and RARß
genes (91). Since COUP-TFI has been shown to be regulated by retinoids
during differentiation of embryonic carcinoma cells, it is considered
as one of the downstream targets of the retinoid-signaling pathways
(20, 92). Thus, it is not surprising that mutation of either COUP-TFI
or RARs will give rise to some common phenotypes. On the other hand,
many of the other defects seen in the RAR double knockouts, including
hyoid bone abnormality, renal hypoplasia, etc., are not seen in the
COUP-TFI null mutants (91). Whether this is due to functional
redundancy of two COUP-TFs or limited convergence of the signaling
pathways shared by COUP-TFI and retinoids has yet to be defined.
D. Loss of function of COUP-TFII in mice
Embryonic lethality is seen with COUP-TFII null mutants. Few
homozygous COUP-TFII mutants survive past 10 days p.c. The cause of the
death has yet to be defined. Since the mutants die very early, it is
difficult to assess the role of COUP-TFII during organogenesis or
during motor neuron differentiation. The function of COUP-TFII during
organogenesis must await a conditional, inducible knockout in the
future.
|
VI. Regulation of the Expression of COUP-TF Genes |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
B. Regulation of COUP-TF by sonic hedgehog (Shh)
COUP-TFs have been shown to be expressed at high levels in motor
neurons of the spinal cord (19, 21). We have previously demonstrated
that the expression of cCOUP-TFII correlates with the differentiation
pattern of motor neurons in the spinal cord (21). In addition, we have
shown that signals from the notochord, which induce motor neuron
differentiation, also induce cCOUP-TFII expression when the notochord
was ectopically transplanted to the dorsal regions of the spinal cord
(21). Since Shh is one of the best known morphogens produced by the
notochord, which is important for induction of differentiation of the
floor plate and motor neurons, we began to ask whether Shh will induce
COUP-TFII gene expression. We showed that Shh induced COUP-TFII mRNA
synthesis when P19 cells were cocultured with bacterially produced Shh
or with conditioned medium from COS-1 cells expressing sonic hedgehog
(G. Krishnan, S. Y. Tsai, and M.-J Tsai, unpublished observation). We
further delineated the regions in the promoter of mCOUP-TFII that are
responsible for conferring Shh induction. The Shh induction is
completely blocked by agents that activate cAMP signaling pathways,
consistent with the accumulated evidence that cAMP antagonizes Shh
function during limb development in vertebrates and eye development in
Drosophila (93, 94).
|
VII. Perspectives |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
The perinatal lethality caused by the COUP-TFI knockout in mice was not entirely anticipated, inasmuch as the expression patterns of COUP-TFI and COUP-TFII overlap in many different regions, particularly in the CNS during early development. For the same reasons, it is possible that the subtle defective phenotypes observed in COUP-TFI null mutants are contributed by the functional redundancy of the two factors and their overlapping expression patterns. Whether the double knockout of COUP-TFI and II will produce much more severe phenotypes must await future investigation.
The cranial nerve fusion defects detected in the COUP-TFI knockout mice are consistent with the fact that the expression of COUP-TFI mimics the migration patterns of the neural crest cells during early embryonic development. Although COUP-TFI is expressed highest in rhombomeres 2 and 4 of the hindbrain, little or no defects are seen with cranial nerves derived from these two rhombomeres. Perhaps the high expression of COUP-TFII in these regions can compensate for the loss of the COUP-TFI. It will be interesting to use organ cultures from the null mutants to assess whether the mutant cells will differentiate to the proper lineage or whether they will assume a different cell fate. Also, it will be interesting to examine whether other marker genes, which are expressed in these rhombomeres, are affected by the loss of COUP-TFI in the null mutants. These results should permit the identification of putative genes that are downstream of COUP-TFs. Also it is interesting to define what is COUP-TF’s role in the formation of ganglia.
The problems associated with arborization and varicosity in the axons of the COUP-TFI null mutants are particularly intriguing. It will be interesting to study whether the axon guidance cues are affected in the COUP-TFI knockout or whether the receptors for the guidance signals are missing in the cell body. It is possible that the loss of COUP-TFI interferes with the proper timing of the expression of the cues or the receptors, thereby leading to the inappropriate migration of neurons and/or projection of the nerve fibers.
The fusion of exoccipital and basioccipital bones in the COUP-TFI null
mutants is also seen in the double knockout mutants of the RAR and
RARß receptors. In addition, some of the cervical bones are ossified
inappropriately. These results reaffirm the concept that COUP-TFs are
involved in the retinoid-signaling pathways. Whether the same
phenotypes are consequences of COUP-TFI being regulated by retinoic
acid or COUP-TFs’ misexpression interfering with the
retinoid-signaling pathways has yet to be defined. Since retinoic acids
have been implicated in the regulation of many Hox genes, it will also
be of interest to determine whether the expression of some Hox genes
are altered, leading to the abnormal fusion and ossification in the
COUP-TFI knockout mice.
The early lethality of COUP-TFII null mutants does not permit us to examine the functional role of COUP-TFII in modulation of motor neuron differentiation in the spinal cord, nor does it allow us to properly investigate the possible role of COUP-TFII in organogenesis of tissues that require epithelial and mesenchymal interactions. Therefore, it will be necessary to generate an inducible knockout of the COUP-TFII gene so that we may begin to assess its role during organogenesis and during motor neuron differentiation and maturation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 This work was supported by NIH Grants DK-44988 (to S.Y.T.) and
DK-45641 and CA-58204 (to M-J.T.). 
|
References |
|---|
|
TopAbstractI. IntroductionII. The COUP-TF Gene...III. Biochemical CharacteristicsIV. Expression Patterns of...V. Physiological Function of...VI. Regulation of the...VII. PerspectivesReferences |
|---|
