QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online 21 April 2004
doi: 10.1242/dev.01115

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Related articles in Development Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Chen, Y. Articles by Mackem, S. Search for Related Content PubMed PubMed Citation Articles by Chen, Y. Articles by Mackem, S.

Direct interaction with Hoxd proteins reverses Gli3-repressor function to promote digit formation downstream of Shh

Yuting Chen^1,, Vladimir Knezevic^1,*,, Valerie Ervin¹, Richard Hutson^1,, Yvona Ward² and Susan Mackem^1,

¹ Laboratory of Pathology, Center for Cancer Research, NCI, NIH, Bethesda, MD 20892, USA
² Cell and Cancer Biology Branch, Center for Cancer Research, NCI, NIH, Bethesda, MD 20892, USA

View larger version (85K): [in a new window]   Fig. 1. Gli3 and Hoxd12 interact genetically during limb development. E17.5-18.5 limb skeletons (left and middle two columns), and E11.5-12.5 hindlimb bud Shh expression (right column) of (A) weak Tg-Hoxd12 line (identical to wild type, +/+), (B) Gli3+/-, (C) Tg-Hoxd12;Gli3+/-, (D) Gli3-/- and (E) Tg-Hoxd12;Gli3-/- embryos. Hindlimb long bones (fe, femur; ti, tibia; fi, fibula) and digits (I-V) are marked for Tg-Hoxd12. Extra digits (*) with distinct identities are marked for Gli3+/- and Tg-Hoxd12;Gli3+/-. Anterior is top, posterior bottom, for all panels except column 2 (anterior right, posterior left). Gli3+/- (B) have only an extra digit I (arrow), whereas Tg-Hoxd12;Gli3+/- (C) have more extensive polydactyly with posterior transformations and very distinct digit identities. By contrast, polydactyly in Tg-Hoxd12;Gli3-/- (E) is unchanged from Gli3-/- (D); both have 7-9 forelimb and 5-7 hindlimb digits that are all short and predominantly digit I-like (see also Fig. 2). Note that in some cases the posterior-most Gli3-/- digits show variable cartilage staining in an otherwise clear, amorphous region that is suggestive of a rudimentary third (middle) phalanx formation (e.g. D,E). In other instances (e.g. Fig. 2F), such rudiments are completely absent from all digits. Unlike digit phenotypes, long bone shortening worsens progressively, and is severest in Tg-Hoxd12;Gli3-/-. Normal Shh expression (E11.5-12, right column) in Tg-Hoxd12 (A) and Gli3+/- (B) is lost by E12.5, whereas some Tg-Hoxd12;Gli3+/- (C) have broad, deregulated Shh at E12. By contrast, Tg-Hoxd12;Gli3-/- (D) and Gli3-/- (E) both show only focal ectopic Shh (arrow) at E12.5.

View larger version (73K): [in a new window]   Fig. 2. Comparison of hindlimb digital morphologies in wild-type (+/+), Tg-Hoxd12;Gli3+/- and Gli3-/- embryos by Gdf5 expression at E14.5 (A-C), and skeletal staining at E18.5 (D-F). Anterior is top, posterior bottom for all panels. At E14.5, bands of Gdf5 expression in digits prefigure sites of future segmentation forming phalangeal joints (Storm and Kingsley, 1996). Both wild type (+/+; A) and Tg-Hoxd12;Gli3+/- (B) display one strong band of expression within digit I, and two bands in each of the more posterior digits (II-V). By contrast, Gli3-/- (C) has only one distinct band of expression and a second `abortive' zone, which never forms a complete band across the digit (evaluated at multiple stages, data not shown). Note that, by E14.5, the proximal-most Gdf5 expression band marking the phalangeal-metatarsal joint region in wild-type embryos has already declined. At E18.5, wild type (D) and Tg-Hoxd12;Gli3+/- (E) have distinguishable digits of varying size, with recognizable identities based on size, shape and number of phalanges. By contrast, Gli3-/- digits (F) are all short, similar in appearance, and have ill-defined phalanges with only a single ossification center. Arrows show ossification centers for digit I (single) compared with digit II-V (two centers), and brackets show phalangeal segments for digit I (2 segments) and digit II (3 segments) in wild type.

View larger version (76K): [in a new window]   Fig. 3. Gli3-Hoxd expression overlap and interaction between endogenous Gli3-Hoxd12 proteins. (A) Expression of Gli3, and of Hoxd10/11/12/13 RNA in nested posterior domains of E10 (left panel) and interdigits of E12 (right panel) forelimb buds (digit I-V, AP, indicated for Gli3). (B) Western blot comparing Gli3 protein in lysates from early chick (stage 22) anterior (A) or posterior (P) limb bud with late stage (27/28) distal digit arch region containing interdigit (ID) mesenchyme, either intact or separated into A and P parts. Lower panels show Hoxd12 and Hoxd13 proteins detected in the same lysates. Note these stages are comparable to mouse E10.5/E11 (early) and E12/E12.5 (late); chick and mouse RNA and protein expression profiles are generally similar (see Dolle et al., 1989; Nelson et al., 1996; Mo et al., 1997; Schweitzer et al., 2000; Wang et al., 2000; Litingtung et al., 2002). The ratio of the short-repressor form of Gli3 protein (83 kDa) relative to full length (190 kDa) in late interdigits is similar to the anterior early limb bud profile, consistent with lack of Shh expression at this stage. In posterior early limb buds, Shh activity results in a high ratio of full-length to repressor form of Gli3. p75kD is a Gli-related antigen of uncertain identity (Wang et al., 2000). (C) Co-immunoprecipitation (IP) of Gli3 and Hoxd12 from early (stage 22, upper panels) and late distal (stage 27/28, lower panel) chick limb bud lysates, using immobilized anti-Hoxd12 or control purified IgG for immunoprecipitation, and anti-Gli3 for detection of bound proteins. Endogenous Hoxd12 binds Gli3 from early limb bud, when both full-length (190kD) and repressor forms (83kD) of Gli3 are expressed, and from later interdigital zones, when mainly the repressor form of Gli3 protein is expressed.

View larger version (57K): [in a new window]   Fig. 4. Domains necessary for Gli3-Hoxd interaction. (A) Relevant Gli3 and Hoxd12 coding domains used in assays shown in B-D. All input lanes show 5-10% of the assay input (B D). (B) The N-terminal Gli3 domain interacts with the C-terminal homeodomain (HD) of Hoxd12 in normalized Gst pull-down assays. The HD-region in Hoxd12/Gst is essential for interaction with in vitro translated (IVT) full-length (FL; B, left panel) or truncated (TR; B, middle panel) Gli3. Hoxd12 mutated in DNA-binding function (mtHD) and N-terminal Gli3 lacking zinc fingers (N-ZnF) still interact (B, middle, right panels). (C) 5'Hox proteins interact with Gli3 preferentially over 3'Hox proteins. Hoxd13(HD)/Gst fusion also binds Gli3 TR (C, left panel). IVT tagged-Gli3 TR (precipitated with Anti-Xpress) also binds full-length IVT Hoxd11 (C, middle panel). Hoxd12 binds preferentially in assays challenged with IVT full-length Hoxa1 or Hoxb1 (C, right panel). (D) Hoxd12 and Gli3 co-immunoprecipitate (IP) from co-transfected cells. Full-length wild-type (wt) or mutant (mtHD) Hoxd12 binds co-transfected Gli3 TR, whereas homeodomain-deleted Hoxd12 (HD) does not (D, left panel). A representative input is shown; all inputs were similar and Hoxd12 recovery in IPs were equivalent (not shown). Hoxd12 binds Gli3 TR preferentially over co-transfected Hoxb1 (D, right panel). (E) Gli3 and Hoxd12 co-localize in transfected cell nuclei, as revealed in optical sections with FITC and Alexa Red antibodies. There are no differences in localization compared with the controls of cells transfected singly and expressing either Gli3 TR or Hoxd12 (full length) alone (data not shown).

View larger version (23K): [in a new window]   Fig. 5. Hoxd12 converts Gli3 repressor into an activator. (A,B) A 4 kb Ptc promoter/luciferase reporter (0.5 µg,) was co-transfected with varying amounts of Gli3 TR repressor (5-80 ng), with or without 25 ng of full-length Hoxd12 wild type (wt; A) or homeodomain mutant (mt; B). (C) Delta-crystallin basal promoter with eight Gli consensus elements (8xGli)/luciferase (1.5 µg) co-transfected with full-length mt Hoxd12 (25 ng) and varying amounts of Gli3 TR (2.5-20 ng). (D) The transfection results are most consistent with a model in which Gli3 recruits Hoxd12 to Gli3-target DNA-binding sites, and the bound Hoxd12 converts Gli3 TR into an activator of the Gli3-target promoters in a stoichiometric fashion that is independent of Hoxd12 DNA binding.

View larger version (15K): [in a new window]   Fig. 6. Hoxd and Gli3 limb expression suggest a quantitative model for modification of Gli3 repressor function by total Hoxd10/11/12/13 protein expression. Schematics show the expected gradient of Gli3:[total Hoxd] complexes across the limb bud AP axis at early (left) and late interdigit (right) stages, and possible Gli3-regulated processes that may be affected by the varying Gli3-Hoxd stoichiometry across the limb bud. This model is compatible with the known functional overlap, and the incremental additive effects of posterior Hox genes in regulating digit morphogenesis (e.g. Zakany et al., 1997).