May 1, 1997;
Cytoplasmically anchored plakoglobin induces a WNT-like phenotype in Xenopus.
is one of two vertebrate proteins closely related to the Drosophila segment polarity gene product armadillo. Overexpression of plakoglobin
induces neural axis duplication in Xenopus and the exogenous plakoglobin
is localized to nuclei (Karnovsky, A., and Klymkowsky, M. W., Proc. Natl. Acad. Sci. USA 92, 4255, 1995; Rubenstein, A., et al., Dev. Genet., 1997, in press). We have carried out a series of experiments to test whether the nuclear localization of plakoglobin
is required for its inductive effects. Prior to the midblastula transition exogenous plakoglobin
is cytoplasmic and concentrated in the cortical regions of blastomeres; after the midblastula transition exogenous plakoglobin
accumulates in embryonic nuclei. The addition of a "nuclear localization sequence" does not change the timing of plakoglobin''s nuclear localization, suggesting that it is anchored in the cytoplasm
prior to the midblastula transition. Next, we constructed two "membrane-anchored" forms of plakoglobin
. These are exclusively cytoplasmic; yet both were as effective at producing a "Wnt-like" axis duplication as were "free," unfettered forms of plakoglobin
. Moreover, expression of anchored plakoglobins had no apparent effect on the cytoplasmic or nuclear levels of beta-catenin. These data indicate that plakoglobin
can act cytoplasmically to generate a WNT-like phenotype. Taken together with the ventralizing effects of a mutant from of the XTcf-3 transcription factor, described by Molenaar et al. Cell 86, 391, 1996, we speculate that in the early Xenopus embryo
, activation of plakoglobin
(or beta-catenin) inhibits the activity of XTcf-3 or a XTcf-3-like factor.
[+] show captions
FIG. 1. A cartoon of the constructs used in this work. Xenopus plakoglobin (A) was tagged at its N-terminus with either a myc-epitope (myc) or a mutated (super) form of green fluorescent protein (GFP). Human plakoglobin (B) was tagged at its C-terminus with myc. Alternatively, a NLS was inserted between the plakoglobin and myc sequences to form HPkg-NLS-myc (C). Two types of mem-brane-anchored forms of HPkg were constructed. The transmembrane domain of rat connexin-32 was fused in-frame to the HPkg-myc sequence to form CnxHPkg-myc (D). A form of CnxHPkg in which
GFP replaces the myc-tag was also constructed (D). Alterna-tively, we fused the signal sequence and transmembrane domain of Xenopus C-cadherin to HPkg-myc to form the CAnchorHPkg-myc polypeptide (E). The CAnchor region lacks the bulk of C-cadherins extracellular domain and all of its cytoplasmic domain. As control, the HPkg sequence was omitted to form the Cnx-myc (F) and CAnchor-myc (G) polypeptides. In parts D–G the transmembrane
domains of the polypeptides are indicated.
FIG. 2. Axis duplication illustrated. Injection of RNA encoding HPkg-myc (A) leads to the appearance of duplicated and occasionally
triplicated neural axes in early-stage embryos; with further development, two distinct anterior–posterior axes can be discerned (B). Similar
axis duplication was produced by the injection of HPkg-NLS-myc (C, D), CnxHPkg-myc (E), and CAnchorHPkg-myc (F) RNAs. No axis
duplication effect was observed in embryos injected with Cnx-myc or CAnchor-myc RNA (not shown).
FIG. 3. The intercellular distribution of exogenous PKGs. Whole-mount immunocytochemistry reveals that prior to the midblastula
transition (a stage 7 embryo is illustrated in A), exogenous HPkgmyc
is found exclusively in the cytoplasm and concentrated at the
periphery of cells. After the midblastula transition (a stage 12–13
embryo is illustrated in B and C) exogenous PKG is found in nuclei
(punctate structures seen in B, arrow). Occasionally within the
same embryo (C), cells with no significant nuclear staining (arrow)
could also be seen. In these regions, the exogenous PKG was cytoplasmic
and localized largely to the cellular cortex. The distribution
of membrane-anchored forms of HPkg is quite different from that
of the free form. Whereas HPkg-myc is primarily nuclear (B), the
CnxHPkg-myc (D), and CancHPkg-myc (E) polypeptides are cytoplasmic.
In the case of CancHPkg-myc, staining is clearly concentrated
at the cellular periphery (arrow). The punctate staining seen
within the central region of some blastomeres corresponds to perinuclear
structures (see Figs. 6 and 7). We consistently observed that
the region of the embryo expressing membrane-anchored plakoglobins is significantly smaller in extent than that found to express the free forms (compare B/C, D, and F).
FIG. 4. Localization of plakoglobins in cultured cells. Intranuclear
injection of either pCS2mt-XPkg-GFP (A, B) or pCS2-CnxHPkg-
GFP (C, D) DNA into Xenopus A6 cells leads to the appearance of
a fluorescence signal within 2 to 3 hr. Living cells were photographed
at 20–24 hr after injection. The mycXPkg-GFP polypeptide
was both cytoplasmic, and localized within nuclei (A, phase; B,
fluorescein optics—nucleus marked N) and was excluded from
nucleoli (arrow marked nu). In contrast, the CnxHPkg-GFP polypeptide
was excluded from nuclei (C, phase; D, fluorescein optics);
it accumulated cytoplasmically as discrete aggregates.
FIG. 5. Localization of the HPkg-NLS-myc polypeptide. Whole-mount immunocytochemistry of a HPkg-NLS-myc RNA injected, postmidblastula
transition (stage 12) embryo (A and B) reveals the absence of cytoplasmic staining. In comparison, an HPkg-myc RNA-injected
embryo (C) at the same stage displays considerable cytoplasmic staining associated with the cell cortex (arrow). In HPkg-NLS-mycexpressing
embryos, the only cells with apparently cytoplasmic staining (A) lack discrete nuclei (arrow) and appear to be in mitosis. In
pre-midblastula transition, stage 7/8, embryos both HPkg-myc (D) and HPkg-NLS-myc (E and F) appear to be exclusively cytoplasmic.
FIG. 6. Expression of CnxHPkg-myc does not alter b-catenin distribution. Embryos, injected with CnxHPkg-myc RNA, were stained for
anti-b-catenin (A, C, E, G, and I) and for CnxHPkg-myc (B, C, D, F, and H) and examined using confocal microscopy. The images in A,
B, H, and I are presented in pseudocolor to facilitate comparison of staining intensities. At low magnification (A and B), the nuclear andcortical localization of endogenous b-catenin in a stage 10 embryo is readily apparent (A). The presence of CnxHPkg-myc (B) in subregions
of the embryo (pointed out by the arrows in B) does not alter the distribution of b-catenin (A). In C, CnxHPkg-myc (red) and b-catenin
(green) staining are superimposed; while b-catenin was localized within nuclei, CnxHPkg-myc was restricted to the nuclear periphery,
with no significant overlap (which would appear as yellow) between the distribution of the two polypeptides. The absence of an effect of
expressing CnxHPkg-myc on b-catenin levels or intracellular distribution can be seen more clearly at higher magnifications (D–I).
CnxHPkg-myc (D and F, shown in red; H shown in pseudocolor) and b-catenin (E–G, shown in green; I, shown in pseudocolor) staining
for three different regions are presented. The level and intracellular distribution of b-catenin staining in cells that express CnxHPkg-myc
is the same as that seen in cells that do not express CnxHPkg-myc. In D and E arrowheads mark cells that do not express CnxHPkg-myc;
in F and G, the two blastomeres that do express CnxHPkg-myc are marked by arrowheads. In H and I an arrow marks a cell that does
not express CnxHPkg-myc; in regions of high local concentration of CnxHPkg-myc (marked by arrowheads in H and I) there is little
FIG. 7. CAnchorHPkg-myc is cytoplasmic and does not alter endogenous b-catenin distribution. To facilitate comparison of staining
intensities, pseudocolor confocal images of embryos injected with CAnchorHPkg-myc RNA are presented. The CAnchorHPkg-myc polypeptide
was stained with 9E10 and Texas red-conjugated secondary antibody (A, C, and E), while endogenous b-catenin was stained with
a rabbit anti-b-catenin antiserum and fluorescein-conjugated secondary antibody (B, D, and F). CAnchorHPkg-myc (A) was seen associated
with the cellular periphery and in cytoplasmic aggregates (marked by arrowheads). Arrows in A and B point to a region that does not
express the CAnchorHPkg-myc polypeptide; the distribution of b-catenin appears similar in regions that do and do not express CnxHPkgmyc.
At higher magnification (parts C– F), cells that do not express CAnchorHPkg-myc and marked by red asterisks and are juxtaposed
to cells that do express CAnchorHPkg-myc. Two confocal planes are shown; no obvious change in b-catenin staining intensity or distribution
is observed between the two cell types. In contrast to the behavior of CnxHPkg-myc, intensely stained aggregates of CAnchorHPkgmyc
(arrows) are also strongly stained by the anti-b-catenin antibody.