| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,2
* Department of Experimental Medicine and Pathology, University "La Sapienza," Rome, Italy;
Department of Biology, University of Rome Tor Vergata, Rome, Italy;
Department of Experimental Medicine and Biochemical Sciences, University of Rome Tor Vergata, Rome, Italy;
Neuromed Institute, Pozzilli, Isernia, Italy;
|| Azienda Ospedaliera Sant'Andrea, Rome, Italy;
¶ Istituto Dermatologico Santa Maria e San Gallicano, IRCCS, Rome, Italy; and
Centro Ricerca Sperimentale, Istituto Regina Elena, Rome, Italy
2Correspondence: Department of Experimental Medicine and Pathology, University "La Sapienza," Viale Regina Elena 324 00161, Rome, Italy. E-mail: maurizio.alimandi{at}uniroma1.it
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: negative regulators • ubiquitin ligase • nuclear translocation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. On the other hand, this balance is believed to contribute to the specific cell response by recruiting different partner proteins through a variety of recognition signals that are specific for the different receptor types. In this context, timing and thresholds of signal intensity are regulated at different levels by the integrated activities of positive and negative regulators for the unique intent of maintaining finely tuned signals propagated by the receptors.
Alternative splicing of the erbB-4 transcripts adds further complexity to the ErbB network by generating four mRNA variants that differ in the sequence encoding the receptor juxta-membrane (JM-a and JM-b) and cytoplasmic domains (CYT-1 and CYT-2) (2
, 3)
. The ensuing structural diversity of the ErbB-4 isoforms has an impact on the activation mechanism and on coupling to specific signaling pathways. To this extent, assembly of signaling molecules at the ErbB-4 JM-b receptors results in spatial constraints at the plasma membrane, whereas only the kinase-active ErbB-4 intracellular domains (E4.ICDs, also referred to as s80 kDa ErbB-4) generated by regulated intramembrane proteolysis (RIP) of the ErbB-4 JM-a receptors can redirect signals in the cytoplasm or nuclei to where they relocate on ligand stimulation (4)
. Also unique to the ErbB receptor family is the ability of ErbB-4 to recruit proteins containing WW domains (5
6
7)
. WW are globular domains consisting of 40 amino acids, two of which are highly conserved tryptophans (WW) (8)
. Similar to SH3 domains, WW domains mediate interaction with proline-rich ligands. The two CYT isoforms of ErbB-4 differ by the presence in CYT-1 of a peptide fragment (PPPAYTPM) containing two overlapping protein recognition motifs. This peptide matches the motifs PPxY and YxxM (where x is any amino acid), typical docking sites for WW domains (6)
and for the phospho-tyrosine binding domain of PI-3K, respectively.
With the exceptions of kidney and heart, where expression of the JM-a and JM-b isoforms is specifically restricted, the presence of multiple isoforms in the same cellular context indicates the importance of utilizing the entire repertoire of signaling options offered by defined ratios of receptor isoforms. This may ensure adequate and balanced signal tuning necessary to control the execution of diverse functional programs. These considerations would also predict that deregulated stoichiometry of the ErbB-4 isoforms might be relevant in pathological conditions. This appears to be true in breast cancer or osteosarcoma, where the presence of nuclear forms of soluble proteins of s80 kDa of ErbB-4 is respectively associated with poor prognosis (9)
or aggressive behavior (10)
, and in medulloblastoma (11)
, where the CYT-1 isoforms are expressed to a higher degree than in normal cerebellum (12)
.
The physiological relevance of nuclear translocation of E4.ICDs has been demonstrated in lactating breast, where localization of E4.ICD/STAT5A complexes in the nuclei of secretory epithelial cells regulates the activity of the ß-casein promoter to induce milk gene expression (13)
. More recently it was shown that N-core/TAB2/E4.ICD complexes are recruited at the promoter regions of astrocytic genes to repress transcription and thus regulate cell fate in the developing brain (14)
. We previously demonstrated that nuclear relocated E4.ICDs may utilize YAP to implement transcription. Altogether, these findings indicate that the s80 kDa form of ErbB-4 works as a chaperone to facilitate the nuclear entry of transcription factors, including p63 or p73 and coactivators of transcription (5
, 6
, 15
, 16)
. In addition, the tumor suppressor gene WWOX encoding for a protein containing two WW domains that bind p73 (7)
is also able to interact with ErbB-4 (17)
. Thus, while the WWOX/p73 complexes limit the YAP-mediated transcriptional activity of p73, WWOX bound to ErbB-4 limits shuttling of the E4.ICDs from the cytoplasm to the nuclear compartment (17)
.
The emerging scenario implicates a number of WW domain proteins regulating propagation of the molecular signal initiated by the ErbB-4 receptor and the shuttling of E4.ICDs between cytosolic and nuclear compartments.
In search of further functional links mediated by WW domain proteins, we used the phage display technique to screen a human brain cDNA library using as bait short fragments of the ErbB-4 cytoplasmic domain containing WW recognition motifs.
We report here that all of the ErbB-4 receptor isoforms bind to and are ubiquitinated and down-regulated by AIP4/Itch (atrophin-1 interacting protein 4), a ubiquitin E3 ligase with HECT domain and four WW domains. AIP4/Itch, here referred to as Itch, is structurally related to the neural development of the down-regulated gene-4 (Nedd4) family of E3 proteins with ubiquitin-ligase functions (18)
.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. pGEX-2TK-MLVPQAFNIPPPIYTSRARIDS (aa 1023–1043), pGEX-2TK-MNRSEIGHSPPPAYTPMSGNQF (aa 1044–1064), and pGEX-2TK-MLSEFSLKPGTVLPPPPYRHRNTVV (aa 1285–1308) were used to screen a T7 phage-displayed cDNA library of human brain (Premade T7Select® cDNA display Libraries, Novagen, Madison, WI, USA). Aliquots of 400 µl of the phage library (
109 particles) were incubated overnight at 4°C with 50 µg of GST-fused ErbB-4 peptides in PBS 3% BSA. Unbound phages were removed by washes (PBS/Tween-20 0.05%) and saved for titration. Bound phages were recovered by incubating at 37°C with 2 ml of a culture of Escherichia coli BLT5615 (Novagen) induced with 1 mM IPTG for 1 h before phage addition. Phage lysates were used for three additional rounds of selection.
Plaque assay
Nitrocellulose membranes (Electran 43611–2B, BDH) were laid on bacterial lawns containing 200 affinity-selected T7 phage particles and maintained at 37°C for 1 h. Filters were then incubated overnight at 4°C with 1 mg/ml of purified GST fusion proteins. Washed membranes were incubated with anti-GST antibody and peroxidase-conjugated anti-goat IgG. Reactive plaques were visualized with enhanced chemiluminescence (SuperSignal, Pierce, Rockford, IL, USA).
Bacteriophage ELISA
Microtiter wells (Nunc, Rochester, NY, USA) coated with 0.2 mg of purified GST-ErbB-4 peptides (or control GST) were washed (PBS 0.05% Tween-20) and incubated with 108 T7 phage particles for 2 h. Washed plates were incubated with an anti-T7 MoAb, then with an anti-mouse IgG conjugated to alkaline phosphatase. Phage particles were revealed after addition of 100 µl/well of 1 mg/ml p-nitrophenyl phosphate in a 50 mM NaHCO3 pH 9.6, 2 mM MgCl2 solution.
Antibodies and plasmids
NRG1-ß3 was purchased from Upstate Inc. (Waltham, MA, USA). Immunoprecipitation was performed using Gammabind G beads (Pharmacia, Uppsala, Sweden). Anti-EGFR (
-C), anti-ErbB-2 (M6), anti-ErbB-3 (MK-4) (LCMB, NIH, Bethesda, MD, USA), and polyclonal ErbB-4 antibody (SC-283, Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used for Western blot. Anti-Flag M2 (Sigma, St. Louis, MO, USA), anti-myc (9E10), and anti-HA (12CA5), anti-Itch MoAb (611198, BD Transduction Laboratories, Lexington, KY, USA), or anti-Itch polyclonal antibody (SC-11890; Santa Cruz Biotechnology, CA, USA) were used for immunoprecipitation and Western blot. pFlag-CMV-2-Itch, pFlag-CMV-2-Itch-C830A, pjx40-HA-E4.ICD, p-myc-Itch, and p-myc-Nedd4 were used as indicated.
Biotinylated peptides and phospho-peptides for in vitro binding
Biotin-LC-IGHSPPPAYTPMSG, biotin-LC-AFNIPPPIYTSRAR, and biotin-LC-GTVLPPPPYRHRNT peptides were synthesized using either Tyr or a phosphorylated analog of Tyr (6)
. Biotinylated peptides spotted on avidin-conjugated nitrocellulose filters were exposed to the GST-WW domains of Itch, Nedd4, YAP, or GST alone for 3 h; washed filters were first incubated with anti-GST antibody, then with radio-labeled 125I -protein A; signals in dried filters were revealed by autoradiography.
Western blot, immunoprecipitation, ubiquitination, and pulse-chase assays
HEK-293 cells were transfected by Lipofectamine 2000 (Invitrogen). Two days after transfection, cells were lysed in HNTG buffer (50 mM HEPES pH 7.5; 150 mM NaCl; 10% glycerol; 5 mM EDTA; 5 mM EGTA; 1% Triton). Ubiquitination experiments were done in the presence of 100 µM chloroquine. Lysates were collected 36 h after transfection in HNTG buffer containing 5 mM N-ethylmaleimide (NEM, Sigma) and in HNTG buffer containing 1% SDS and 1% deoxycholate, boiled for 5 min, and subsequently diluted to 0.1% SDS before immunoprecipitation.
HEK-293 cells transiently transfected with either ErbB-4 or E4.ICD (6 µg) and Itch or Itch-C830A (2 µg) were starved for 45 min with DMEM Met-Cys-free (Sigma D0442) supplemented with 2% dialyzed FCS and pulsed with 100 µCi/ml of both 35S-methionine and cysteine (Expre35S35S Protein labeling mix; NEN, Boston, MA, USA) for 2 h. After a chase with 100x excess of cold methionine- and cysteine-containing medium, cells were lysed at different time points. Immunoprecipitated samples were resolved by SDS page and subjected to autoradiography.
Immunofluorescence analysis
NIH.E4 cells expressing the ErbB-4 JM-a/CYT-2 isoform and T47D were grown onto glass coverslips precoated with PBS 2% of gelatin (Sigma), washed, fixed (PBS 4% paraformaldehyde) for 30 min, and permeabilized with 0.1% Triton X-100 (NIH.E4) or with 0.1% saponin (T47D) for 5 min. Cells were stained using anti-ErbB-4 and anti-Itch Abs followed by FITC-conjugated anti-rabbit IgG or Texas Red goat anti-mouse IgG (Cappel Research Products, Durham, NC, USA). Fluorescent signals were analyzed using a Zeiss Confocal Microscope (Zeiss, Oberkochen, Germany). Lyso-Tracker internalization assay: starved T47D cells where treated with 100 ng/ml NRG for 1 h at 4°C, then for 2 h at 37°C before incubation with 100 nM Lyso-Tracker-Red (Molecular Probes, Eugene, OR, USA) for 30' at 37°C. Cells were stained with anti-ErbB-4 and anti-Itch, then visualized by goat anti-rabbit IgG-FITC or goat anti-mouse IgG-Alexa Fluor (Molecular Probes). Specimens were scanned in single sections of 0.5 µm obtained with an ApoTome system connected with an Axiovert 200 inverted Epifluorescence Microscope (Zeiss). Quantitative analysis of colocalized signals was calculated analyzing a minimum of 30 cells randomly taken for each experiment.
Depletion of endogenous Itch by siRNA
DAOY and TD47 cells were electroporated with either control siRNA duplex or a mix of two different Itch-targeting oligos (Ambion Inc., Austin, TX, USA) as described elsewhere (19)
.
Luciferase assays on SRE
HeLa were transfected with the SRE-Luciferase reporter and the indicated combinations of ErbB-4, Itch, Itch-C830A, and ß-Gal plasmids. Eight hours after transfection, cells were starved (16 h) and stimulated with 100 ng/ml of NRG1-ß3 (16 h). Luciferase activity was quantified with a TD-20E luminometer (Turner Biosystems, Sunnyvale, CA, USA).
Gal4 trans-activation assay
HeLa cells were transfected with Gal4-BD-E4.ICD (100 ng/well), Gal4-Luc (400 ng/well), ß-Gal-CMV (50 ng/well), and pcDNA3-YAP (50 ng/well) in the presence or absence of pFlag-Itch-C830A. Empty vectors were used for DNA content normalization and ß-Gal staining to normalize results for transfection efficiency. The experiments were done in triplicate.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). After three panning cycles, positive clones were identified by plaque assay. One of the eight clones isolated with the bait GST-MLVPQAFNIPPPIYTSRARIDS fusion protein (GST-PPIY: aa 1023–1043) displayed a fragment (aa 218–478) comprising all four WW domains of the AIP4/Itch protein (Fig. 1
A). The physical interaction between the proline-rich region of ErbB-4 and the WW domains of Itch was confirmed by ELISA assay (Fig. 1B
).
|
|
Protein-protein interaction studies
The first approach we used to study the interaction between ErbB-4 and Itch was aimed at verifying whether Sepharose-conjugated GST-Itch fusion proteins could capture ErbB-4 receptors in transiently overexpressed HEK-293 cells. Since the CYT-1 isoform contains an extra WW binding motif (Fig. 2
A), we tested the CYT-1 and CYT-2 isoforms of ErbB-4 in pull-down experiments. The results illustrated in Fig. 2B
indicate that both ErbB-4 CYT isoforms can serve as platforms for interaction with Itch.
|
To identify domains involved in the binding of Itch to the receptor, we used GST fusion proteins of the four WW domains of Itch to capture the ErbB-4 JM-b/CYT-1 expressed in HEK-293 cells. These experiments confirm that all the WW domains of Itch can mediate the interaction with ErbB-4 (Fig. 2C
). Given the conserved domain structure of the Nedd family of proteins, we wanted to test whether Nedd4 could also bind to ErbB-4 in pull-down experiments. GST-Nedd4 fusion proteins were also able to capture ErbB-4 (Fig. 2D
); however, when we looked at the binding ability of each of the four GST-WW domains of Nedd4 on ErbB-4, we found that only the WW2 and WW3 were able to bind the receptor (Fig. 2D
).
Next, to identify binding domains in the ErbB-4 sequence responsible for the interaction with Itch, we used a Far-Western approach. We probed a set of peptides containing the PPxY motifs of ErbB-4, either unphosphorylated or tyrosine phosphorylated, with the individual WW domains of Itch. As shown in Fig. 2E
, all four GST-Itch-WW domains (GST-I-WW) bind to the ErbB-4 peptides, although with quantitative and qualitative differences; most notably, the WW1 domain of Itch recognizes both phosphorylated and unphosphorylated ErbB-4 peptides whereas the WW2, WW3, and WW4 seem to interact with the unphosphorylated peptides only.
Finally, we wanted to define the ability of Itch and ErbB-4 to form a complex in vivo. We transiently transfected HEK-293 with constructs expressing ErbB-4, myc-Itch, and myc-Nedd4 in different combinations to evaluate the potential formation of Itch/ErbB-4 and Nedd4/ErbB-4 complexes after anti-myc immunoprecipitation. Despite the different stoichiometry of protein expression, we could detect ErbB-4 immune reactivity only from lysates coexpressing Itch and ErbB-4; conversely, no binding could be detected between Nedd4 and ErbB-4 in vivo in cells overexpressing both proteins (Fig. 2F
). Altogether, these results indicate that Itch, but not Nedd4, can form a complex with ErbB-4 in vivo by a WW/PPxY-mediated interaction.
Colocalization of ErbB-4 and Itch
To assess the subcellular compartments where binding occurs, we first analyzed ErbB-4 and Itch localization by confocal microscopy in NIH3T3 cells stably transfected with ErbB-4 (NIH.E4, with 106 receptors per cell) and in T47D breast cancer cells expressing endogenous ErbB-4. Staining of the endogenous Itch appeared in small dots localized mostly in a perinuclear Golgi area, but also at the cell periphery, resembling endosomes (Fig. 3
A). This pattern of staining agrees with previous studies performed in Itch-transfected cells (20
, 21)
. The ErbB-4 signal was distributed on the cell plasma membrane as well as in intracellular dots scattered all over the cytoplasm in T47D cells and more concentrated in the juxtanuclear area in NIH.E4 cells. Colocalization of the two signals (yellow in the merged images) was evident in punctate structures both centrally and peripherally located (Fig. 3A
). Treatment of T47D with NRG for 2 h did not affect the pattern of localization of Itch or (as expected; see ref. 4
) of the receptor (Fig. 3B
). To identify the endocytic structures where ErbB-4 and Itch colocalize, we performed a triple fluorescence in T47D cells using anti-ErbB-4, anti-Itch, and Lyso-Tracker, a specific marker of the lysosomal compartment. Details of the perinuclear area of cells analyzed by a epifluorescence microscope fitted with an ApoTome System showed that 20% of Itch colocalizes with the receptor in punctate structures; 60% of these dots were lysosomes, as demonstrated by staining with Lyso-Tracker (Fig. 3C
). This pattern was not significantly altered after 2 h of NRG treatment (Fig. 3C
).
|
Itch diverts ErbB-4 to the degradation pathways
The function of Itch is to coordinate sorting and ubiquitylation events that drive several proteins, including transmembrane receptors such as Notch and CXCR4, to proteasomes and/or lysosomes for degradation (22
, 23)
. To verify whether Itch could regulate ErbB-4 stability, we transfected HEK-293 with ErbB-4 and increasing amounts of Itch. As shown in Fig. 4
A, B, wild-type Itch, but not the C830A mutant, induces reduction of the CYT-1 and CYT-2 ErbB-4 levels. The different sensitivity of the two 180 kDa CYT isoforms to Itch might be related to their differential ability to stably complex Itch via their specific repertoire of PPxY motives. Notably, a progressive decrease of receptor levels is more evident for the s80 kDa than for the 180 kDa form of ErbB-4. Conversely, we could not detect decreased receptor levels in cells overexpressing ErbB-4 and Nedd4 (Fig. 2G
).
|
Furthermore, pulse-chase experiments performed to address the ability of Itch to affect the turnover of ErbB-4 and E4.ICD confirmed that Itch, but not Itch-C830A, increases the degradation of both the mature and cleaved form of ErbB-4 (Fig. 4C
) and of the E4.ICD (Fig. 4D
).
Finally, we tested the ability of Itch to induce degradation of EGFR, ErbB-2, and ErbB-3. No substantial degradation was observed when EGFR (Fig. 4E
), ErbB-2, or ErbB-3 (Fig. 4F
) were cotransfected with Itch, consistent with the observation that these receptors do not contain WW domain docking sites.
Opposing effects of Itch overexpression and interference on endogenous ErbB-4 levels
The experiments reported so far demonstrate that the interaction with Itch is instrumental to control ErbB-4 levels. To further confirm the role of Itch in controlling receptor density, we transfected T47D breast cancer cells with wild-type Itch and monitored endogenous ErbB-4 levels 48 h after transfection. The experiment in Fig. 5
A shows that upon Itch overexpression, both the 180 kDa membrane-bound and the s80 kDa form of the endogenous ErbB-4 are down-regulated, with a respective decrease of 55% and 60%, as revealed by densitometric analysis (Fig. 5B
).
|
Next, we electroporated siRNAs to reduce endogenous Itch expression in T47D breast cancer cells and in DAOY, a medulloblastoma cell line in which, despite expression at the mRNA level, ErbB-4 protein levels are below the detection limit of WB analysis (11)
. Twenty-four hours after electroporation, Itch-specific siRNAs, but not scrambled oligonucleotides, induce an increase of ErbB-4 basal levels in both cell lines (Fig. 5C
): the 180 kDa band becomes detectable by WB in DAOY cells and is up-regulated with a gain of 30% in T47D cells, as determined by densitometric analysis (Fig. 5D
). Altogether, these experiments substantiate the role of Itch in regulating endogenous ErbB-4 levels.
Itch ubiquitinates ErbB-4
The ubiquitination pathway regulates protein fate through covalent attachment of ubiquitin molecules. To determine whether the ErbB-4/Itch interaction induces ubiquitination of the receptor in vivo, we performed immunoprecipitation experiments in cells transfected with ErbB-4 and Itch or Itch-C830A in the presence or absence of a construct expressing ubiquitin tagged with the HA epitope (UB-HA). Unlike the C830A mutant, wild-type Itch was effective in increasing ubiquitination of both the JM-a and JM-b isoforms of ErbB-4; the reduction of ErbB-4 levels by Itch was also confirmed in this experimental setting (Fig. 6
A, B). The self-ubiquitination of wild-type Itch, but not of the catalytically inactive Itch-C830A, is consistent with the negative autoregulatory function proposed for the E3 domains. Ubiquitination experiments were further performed in immunoprecipitates from lysates denatured in 1% SDS buffer to exclude the possibility that anti-HA reactive bands could be due to ubiquitinated proteins associated with ErbB-4 (Fig. 6C
). These results indicate that Itch is an E3 ligase responsible for ErbB-4 ubiquitination.
|
Itch diverts ErbB-4 to lysosomes for degradation
A decision between lysosomal vs. proteasomal degradation generally depends on the sorting mechanisms distinguishing mono- vs. polyubiquitination (24)
. Covalent attachment of ubiquitin moieties can work as an endocytosis signal targeting receptor proteins to the lysosomal machinery (25
, 26)
. To unravel the relative contribution of the lysosomal and proteasomal pathways in the degradation of ErbB-4 induced by Itch, we monitored receptor levels in the presence of either the lysosomes inhibitor chloroquine or the proteasome inhibitor clasto-lactacystin-lactone. In accordance with previous reports (27)
, our results indicate that ErbB-4 is constitutively subjected to proteasomal degradation, as revealed by the increased protein levels obtained after lactacystin treatment, more evident for the s80 kDa than for the 180 kDa receptor form (Fig. 7
A–C). However, the effect of proteasome inhibitors on ErbB-4 levels is qualitatively and quantitatively similar in either the presence or absence of exogenous Itch or Itch-C830A. On the other hand, when ErbB-4 is expressed in the presence of Itch, the general decrease in the total amount of receptor is more pronounced for the s80 kDa than for the 180 kDa form and is accompanied by an ex novo susceptibility of the s80 kDa forms to a lysosome-dependent degradation, as revealed by the increased protein levels obtained after chloroquine treatment (Fig. 7A, B
). This effect is also dose dependent on Itch (Fig. 7A
). Furthermore, the effect exerted by the lysosome inhibitors on ErbB-4 levels requires the presence of a functionally active Itch, since it is not observed in the presence of the inactive Icth-C830A (Fig. 7C
). Notably, both Itch-independent proteasomal degradation and Itch-dependent lysosomal degradation of ErbB-4 seem to occur independent of NRG stimulation (Fig. 7B, C
).
|
Itch down-regulates ErbB-4-stimulated transcription
We next aimed at investigating the impact of Itch-dependent ubiquitination on ErbB-4 downstream signaling. To this end, we used a reporter system to measure the activation of the SRE element under neuregulin (NRG) stimulation in HeLa cells transiently transfected with ErbB-4 in the presence/absence of increasing amounts of Itch or Itch-C380A. The results obtained in this assay lead to two main conclusions: 1) Itch down-regulates receptor functions in both unstimulated and stimulated cells; and 2) Itch-C380A exerts a dominant negative effect, probably due to the ability to compete with endogenous Itch and titrate out its function (Fig. 8
). Parallel evaluation of protein levels in HeLa cell lysates subjected to the reporter assays confirmed the Itch-dependent progressive decrease of ErbB-4.
|
Itch and YAP compete for binding to ErbB-4
The ability of ErbB-4 to interact with a number of WW domain-containing proteins expands the possible ramifications of its signaling network. It is conceivable that partner selection would depend on competition between WW proteins with different binding affinities as well as on protein localization in specialized compartments. To evaluate the ability of Itch to compete with YAP for ErbB-4 binding, we transfected HEK-293 cells with E4.ICD fused to GAL4-BD, YAP, and Itch-C830A, unable to ubiquitylate or degrade ErbB-4 but still able to complex with the receptor (not shown). In the GAL4 assay, competitive interactions can be evaluated as variations of luminescence expressed by the transcriptional activation of a luciferase reporter construct. Since both endogenous Itch and transfected Itch are detectable in the cytosol (Fig. 3)
(20
21)
, the ability of Itch-C830A to sequester ErbB-4 is monitored through the decreased amount of E4.ICD/YAP complexes able to trans-activate the reporter gene by recruiting transcriptional elements at the nuclear level. The data reported in Fig. 9
indicate that increasing amounts of Itch-C830A compete with YAP for binding to ErbB-4 receptor, thus limiting formation of the transcriptionally active E4.ICD/YAP complexes; this indicates that Itch can sequester E4.ICDs in the cytoplasm and participate directly in the regulated entry of defined amounts of E4.ICDs in the nuclei.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
c-Cbl binds to ligand-stimulated EGFR and in turn becomes phosphorylated (28)
. Its activation results in mono-ubiquitination of the receptor, a phenomenon that promotes its internalization through endocytosis along multiple sorting steps, culminating in either recycling of the EGFR to the cell surface or its degradation by lysosomes or proteasome (29
, 30)
. The established paradigm linking c-Cbl/EGFR complexes to ubiquitination, sorting, endocytosis and receptor degradation is not directly applicable to the other members of the ErbB family. ErbB-2, ErbB-3, and ErbB-4 receptors are impaired in their ability to rapidly internalize from the plasma membrane (31)
; mechanisms controlling receptor density prior or after growth factor stimulation are still poorly defined.
Evidence has been produced linking LRIG1, the mammalian ortholog of Kekkon, to regulation of the ErbB network in virtue of its ability to down-regulate each member of the ErbB receptor family (32)
. Although the mechanistic details are unclear, at least in the case of the EGFR the recruitment of c-Cbl has been shown to induce the simultaneous ubiquitination and degradation of both EGFR and LRIG1 (32)
. Another protein that has been implicated in maintaining an adequate density of ErbB-3 and ErbB-4 is the Neuregulin receptor degradation protein 1 (Nrdp1). Nrdp1 is a RING finger E3 ubiquitin ligase that promotes redistribution of ErbB-3 from the cell surface to intracellular compartments and induces suppression of ErbB-3 and ErbB-4 receptor levels independent of receptor stimulation (33
34)
.
Our screening of a phage-displayed cDNA library, using peptide fragments of ErbB-4-containing WW recognition motifs as bait, has identified Itch as a candidate functional ligand. We show that a direct consequence of this interaction is the accelerated decay of the membrane-bound (180 kDa) and, to a greater extent, of the soluble-cleaved form (s80 kDa) of ErbB-4. Although each of the four WW domains of Itch can bind to each of the three ErbB-4 peptides containing a PPxY motif, the interaction mediated by the first WW domain is unique, since it appears to occur regardless of the phosphorylation status of the tyrosine within the peptide motif. Upon binding to ErbB-4, Itch induces ubiquitination and protein degradation by a mechanism that appears to involve sorting of the receptor to the lysosomal compartment rather than to the proteasome. Both ubiquitination and degradation require intact catalytic functions of the HECT domain since the Itch-C830A mutant does not degrade ErbB-4. Remarkably, although Nedd4 is able to form complexes with ErbB-4 in vitro through WW2 and WW3, it does not bind or induce degradation of ErbB-4 in vivo, suggesting that the interaction mediated by the two WW domains could be unstable and ineffective in activating the catalytic functions of the HECT domain. We did not test whether WWP1 or WWP2, closest homologous of Itch, or Nedd4–2, the closest homologue of Nedd4, are effective as well in the down-regulation of ErbB-4 receptor levels.
We demonstrate here that interference with the endogenous Itch induces an increase of endogenous ErbB-4 in breast carcinoma and medulloblastoma cells. These results indicate that Itch regulates ErbB-4 levels in physiological contexts as well as on overexpression of both proteins. However, we could not detect coimmunoprecipitation between Itch and ErbB-4 in T47D cells. Lack of coimmunoprecipitation might be explained by the short half-life of their complexes due to degradation of the Itch-bound receptor; alternatively, it is conceivable that their interactions might be regulated within spatial or temporal windows of the turnover of the ErbB-4 receptor, or by the JNK1-mediated activation of Itch (35)
.
The results presented in this work expand our understanding of the network of interactions linking ErbB-4 to partner proteins containing WW domains; YAP, Itch, and WWOX are part of this network. Their recruitment might serve to either implement or regulate the signals initiated by the activation of ErbB-4. In the case of the JM-a isoforms, the dual sequential mechanism of RIP generates membrane-free, kinase-active E4.ICD (36)
, which rapidly relocate to the nuclei where, in complex with either YAP, STAT-5, or TAB2/N-cor, they may modulate the activity of p73, induce transcription of ß-casein mRNA, or repress the expression of astrocyte genes. However, how nuclear translocation of the E4.ICDs is regulated is poorly understood.
We propose a model (Fig. 10
) according to which the network of interaction with WW domain proteins controls the ingress of stoichiometrically defined quotes of E4.ICDs migrating to the nuclear compartment. According to current information, the interplay of Itch, WWOX, and YAP with ErbB-4 per se might be sufficient to regulate the access of E4.ICDs to subcellular compartments where different functions take place. Indeed, binding of Itch, WWOX, and YAP to ErbB-4 occur via WW domain/PPxY interactions that are mutually exclusive. Partner selection and complex formation would be regulated in time and space by variations in reciprocal affinity and by the availability of the different partners in different cellular compartments. Thus, while binding of ErbB-4 to WWOX may negatively regulate nuclear access of E4.ICDs/YAP complexes by sequestering them in the cytosol, binding to Itch could in turn induce ubiquitination and consequent degradation of excessive E4.ICDs. The dynamic regulation of binding of WWOX, Itch, and YAP to ErbB-4 may define nodal mechanisms to control different receptor functions. The first level of control would be at the receptor level. In principle, the mechanism that regulates binding of WW domains to ErbB-4 is phosphorylation of the Tyr in position 1056 embedded in the PPxYxxM motif of the CYT-1 isoform (6
, 37)
. Tyr phosphorylation might influence the affinity of interaction of the receptor with Itch, WWOX, or YAP, thus determining protein composition, ratios, and stoichiometry of the different complexes.
|
The role of Itch in the ErbB network appears more complex than originally thought. In our hands, the overexpression of Itch did not alter receptor levels of unstimulated EGFR, ErbB-2, or ErbB-3. However, experimental evidence indicates a role for Itch in holding in check the EGFR kinase activity by regulating the ubiquitination of Cbl-C (38
, 39)
. It has also been reported that formation of the Cbl/CIN85/endophilin complexes that mediate ligand-induced down-regulation of the EGFR (29)
is sustained by an Itch/endophilin interaction (20)
. Altogether, these data seem to indicate that Itch may be a component of the regulatory pathways of endocytosis and trafficking of the EGFR. This role might be even more relevant on ligand-dependent triggering of EGFR/ErbB-4 dimers.
Our results represent the first demonstration that ErbB-4 receptors can be ubiquitinated and diverted to degradation by AIP4/Itch. This interaction can contribute to maintain adequate ErbB-4 receptor density at the cell surface and regulate nuclear access by degrading excessive quotes of kinase active E4.ICDs.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication December 13, 2006. Accepted for publication March 15, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. Zeng, J. Xu, and R. C. Harris Nedd4 mediates ErbB4 JM-a/CYT-1 ICD ubiquitination and degradation in MDCK II cells FASEB J, June 1, 2009; 23(6): 1935 - 1945. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Chan, R. Tian, Y.-F. Lee, S. T. Sit, L. Lim, and E. Manser Down-regulation of Active ACK1 Is Mediated by Association with the E3 Ubiquitin Ligase Nedd4-2 J. Biol. Chem., March 20, 2009; 284(12): 8185 - 8194. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-M. Feng, R. S. Muraoka-Cook, D. Hunter, M. A. Sandahl, L. S. Caskey, K. Miyazawa, A. Atfi, and H. S. Earp III The E3 Ubiquitin Ligase WWP1 Selectively Targets HER4 and Its Proteolytically Derived Signaling Isoforms for Degradation Mol. Cell. Biol., February 1, 2009; 29(3): 892 - 906. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sundvall, A. Korhonen, I. Paatero, E. Gaudio, G. Melino, C. M. Croce, R. I. Aqeilan, and K. Elenius Isoform-specific monoubiquitination, endocytosis, and degradation of alternatively spliced ErbB4 isoforms PNAS, March 18, 2008; 105(11): 4162 - 4167. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
