Human Molecular Genetics

Human Molecular Genetics Advance Access originally published online on July 21, 2005
Human Molecular Genetics 2005 14(17):2521-2531; doi:10.1093/hmg/ddi256

This Article Abstract FREE Full Text (PDF) All Versions of this Article: 14/17/2521    most recent ddi256v2 ddi256v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (38) Request Permissions Google Scholar Articles by Zawistowski, J. S. Articles by Marchuk, D. A. Search for Related Content PubMed PubMed Citation Articles by Zawistowski, J. S. Articles by Marchuk, D. A. Social Bookmarking          What's this?

© The Author 2005. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oupjournals.org

CCM1 and CCM2 protein interactions in cell signaling: implications for cerebral cavernous malformations pathogenesis

Jon S. Zawistowski^1,, Lisa Stalheim^2,, Mark T. Uhlik², Amy N. Abell², Brooke B. Ancrile¹, Gary L. Johnson² and Douglas A. Marchuk^1,*

¹Department of Molecular Genetics and Microbiology, Duke University Medical Center, Durham, NC 27710, USA and ²Department of Pharmacology and the Lineberger Comprehensive Cancer Center, University of North Carolina School of Medicine, Chapel Hill, NC 27599-7365, USA

^* To whom correspondence should be addressed. Tel:+1 9196843290; Fax:+1 9196819193; Email: march004{at}mc.duke.edu

Received April 28, 2005; Accepted July 14, 2005

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Cerebral cavernous malformations (CCMs) are sporadically acquired or inherited vascular lesions of the central nervous system consisting of clusters of dilated thin-walled blood vessels that predispose individuals to seizures and stroke. Familial CCM is caused by mutations in KRIT1 (CCM1) or in malcavernin (CCM2), the murine ortholog of which was concurrently characterized as osmosensing scaffold for MEKK3 (OSM). The roles of the CCM proteins in the pathogenesis of the disorder remain largely unknown. Here, we use co-immunoprecipitation, fluorescence resonance energy transfer and subcellular localization strategies to show that the CCM1 gene product, KRIT1, interacts with the CCM2 gene product, malcavernin/OSM. Analogous to the established interactions of CCM1 and ß₁ integrin with ICAP1, the CCM1/CCM2 association is dependent upon the phosphotyrosine binding (PTB) domain of CCM2. A familial CCM2 missense mutation abrogates the CCM1/CCM2 interaction, suggesting that loss of this interaction may be critical in CCM pathogenesis. CCM2 and ICAP1 bound to CCM1 via their respective PTB domains differentially influence the subcellular localization of CCM1. Furthermore, we expand upon the established involvement of CCM2 in the p38 mitogen-activated protein kinase signaling module by demonstrating that CCM1 associates with CCM2 and MEKK3 in a ternary complex. These data indicate that the genetic heterogeneity observed in familial CCM may reflect mutation of different molecular members of a coordinated signaling complex.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Cerebral cavernous malformations (CCMs) are vascular lesions of the central nervous system consisting of clusters of dilated thin-walled blood vessels or ‘caverns’. The cluster of vessels that defines the CCM lesion is surrounded by connective tissue while remaining distinct from the surrounding neural parenchyma (1). The cavernous malformation lesions have the capacity to hemorrhage, resulting in seizures, stroke and focal neurological deficits (2,3).

Three forms of autosomal dominant CCM have been mapped (4–7), and the disease gene products are known for all three of the mapped loci. CCM1 is caused by truncating mutations in KRIT1 (8,9), a protein containing ankyrin repeat motifs and a FERM domain. CCM2 results from mutations in MGC4607 (10,11), encoding the phosphotyrosine binding (PTB) domain protein malcavernin (10), the murine ortholog of which was characterized as a mitogen-activated protein kinase (MAPK) scaffold named osmosensing scaffold for MEKK3 (OSM) (12). CCM3 has recently been shown to result from mutations in PDCD10 (programmed cell death 10), a gene upregulated in the human myeloid cell line TF-1 upon induction of apoptosis (13). Germline mutations within CCM1 kindreds are primarily nonsense, frameshift or splice site mutations, predicted to be loss of function alleles (14). A similar class of mutations is found in CCM2 and CCM3 pedigrees (10,11,13); however, a missense mutation has been reported for a CCM2 kindred (11).

Although the functions of these novel proteins remain to be elucidated, initial insight into the function of the CCM1 protein came from its identification as a yeast two-hybrid binding partner for the Ras-like small GTPase Rap1A (15). It has since been shown that CCM1 binds the integrin binding protein ICAP1, suggesting a potential role of the CCM1 protein as a modulator of integrin-mediated cell–cell or cell–matrix adhesion (16,17). The interaction of ICAP1 with the ß₁ integrin cytoplasmic domain occurs via a canonical PTB domain/NPxY amino acid motif interaction (18–20). This interaction is well-established and is the prototypical means by which PTB domain scaffold proteins bind intracellular tails of transmembrane receptors (21). The ICAP1 interaction with CCM1 occurs in an analogous manner to that of the ß₁ integrin cytoplasmic domain, as mutagenesis of an NPxY motif within CCM1 abrogates the association (16,17).

Concurrent with the identification of MGC4607 as the CCM2 gene product, the murine ortholog of MGC4607 was identified as a yeast two-hybrid binding partner for the MAP kinase kinase kinase MEKK3. Murine CCM2 (OSM) was placed in the mammalian counterpart to the osmoregulation high-osmolarity glycerol MAPK pathway in yeast (12). CCM2 was shown to bind the MAPK members MEKK3 and MKK3 as well as Rac1. The study provided evidence that upon osmotic shock, the CCM2 protein serves as a scaffold for a Rac1/MEKK3/MKK3 signaling complex regulating p38 MAP kinase activation.

Here, we show that CCM1 and CCM2 are in a complex via a CCM1/CCM2 PTB domain interaction, analogous to the CCM1/ICAP1 PTB interaction. We provide evidence by co-immunoprecipitation, fluorescence resonance energy transfer (FRET) and subcellular localization to support the CCM1/CCM2 interaction. We show that the genetically linked CCM proteins 1 and 2 share a common molecular link, as well as invoke an expanded role for MAPK signaling in CCMs pathogenesis.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
CCM2^+/– cells have perturbed MAPK activation in response to osmotic shock
CCM2 (OSM) is required as a scaffold for MEKK3-mediated p38 MAPK phosphorylation during osmotic shock by sorbitol (12). We therefore assessed the role of the CCM2 protein in the p38 MAPK phospho-relay module by isolating mouse embryonic fibroblasts (MEFs) heterozygous for a gene trap insertion in the mouse Ccm2 gene. This gene trap cell line (Bay Genomics RRG051) has a LacZ insertion in exon 6 of Ccm2, which is predicted to be a loss of function mutation (N.W. Plummer et al., submitted for publication). We chose to analyze heterozygous MEFs, because they are the genotypic equivalent of CCM patients. We first verified that the CCM2 protein was reduced by the knockout allele in heterozygous MEFs by western blot with CCM2 antibody (Fig. 1A). Cells heterozygous for the Ccm2 mutation had a marked reduction of activated p38 MAPK in response to a sorbitol time course as assessed by Thr180/Tyr182 phosphorylation (Fig. 1A and B). Reduced p38 activity was restored by add-back of FLAG-CCM2 (Fig. 1B), demonstrating that the effect was specific to the reduction of CCM2. Perturbed p38 activation in response to hyperosmotic stress in CCM2^+/– cells is consistent with our previous study defining CCM2 as a scaffold for MEKK3-dependent p38 activation (12).

View larger version (35K): [in this window] [in a new window]   Figure 1. MEFs heterozygous for a CCM2 gene trap allele have impaired p38 activation upon hyperosmotic stress. (A) Western blotting with anti-CCM2 and anti-ERK1/2 as a loading control confirms reduction of CCM2 protein in CCM2+/– MEFs (top panel). Wild-type or CCM2+/– MEFs were serum-starved and treated for the indicated times with 0.2 M sorbitol or 15 min with 10 mg ml–1 anisomycin as a positive control for p38 activation (+). Lysates were subjected to western blotting with anti-p38 to reveal total p38 levels and with anti-phospho (Thr180/Tyr182) p38 to reveal activated p38 levels (bottom panel). (B) Quantification of impaired p38 activation of CCM2+/– MEFs in (A) and rescue by add-back of CCM2. Fold p38 activation was calculated (see Materials and Methods) for wild-type and heterozygous CCM2 MEFs. Add-back of FLAG-CCM2 by transient transfection rescues the impaired p38 activation in CCM2+/– MEFs, demonstrating that the phenotype is CCM2-dependent. Results are representative of three independent experiments. Error bars represent standard error of the mean.

 
CCM1 associates with the MEKK3/CCM2 scaffold complex
As CCM2 (OSM) scaffolds MEKK3 during p38 MAPK activation (12), we sought to determine whether CCM1 is a member of the CCM2/MEKK3 complex. Following transfection of epitope-tagged CCM1, CCM2 and MEKK3, all three proteins are detected in the immunoprecipitates using polyclonal CCM2 antibody (Fig. 2A).

View larger version (49K): [in this window] [in a new window]   Figure 2. CCM1 associates with CCM2 in the p38 MAPK module. (A) CCM1 is identified in a ternary complex with CCM2 and MEKK3. 293T cells were transfected with the indicated constructs, immunoprecipitated with anti-CCM2 and subjected to western blotting with the indicated antibodies. In addition to both HA-MEKK3 and FLAG-CCM1 singly co-immunoprecipitating with FLAG-CCM2, a FLAG-CCM1/FLAG-CCM2/HA-MEKK3 ternary complex is detected. Approximate molecular weights are FLAG-CCM1: 85 kDa and FLAG-CCM2: 60 kDa. (B) CCM1/CCM2 interact in live cells and this interaction is enhanced upon hyperosmotic stress. COS-7 cells transfected with CFP-CCM2 and YFP-CCM1 were visualized at room temperature and analyzed by micro-FRET. CFP fluorescence, YFP fluorescence and corrected FRET (FRETC) were examined for non-treated cells (top) and cells were treated with 0.2 M sorbitol for 5 min (bottom). Non-treated cells yield a diffuse FRETC signal throughout the cytoplasm. Cells treated with sorbitol display a re-distribution of CFP-CCM2 and YFP-CCM1 to the cell periphery (arrows) and an enhanced FRETC interaction signal in the cytoplasm and at the cell periphery.

 
We hypothesized that the presence of CCM1 in the CCM2/MEKK3 signaling complex may represent an interaction between CCM1 and CCM2. As an initial means of assessing the potential for the CCM proteins 1 and 2 to associate, we performed CYFRET analysis in live COS-7 cells. Cells co-expressing yellow fluorescent protein (YFP)-CCM1 and cyan fluorescent protein (CFP)-CCM2 yield a FRETN^C value of 7.43x10^–5, whereas control cells co-expressing YFP and CFP-CCM2 yield a FRETN^C value of 0.35x10^–5 (Table 1). Visualization of the corrected FRET value revealed a diffuse interaction signal throughout the cytoplasm (Fig. 2B).

View this table: [in this window] [in a new window]   Table 1. FRET values for CCM1/2 interactions

 
We next determined whether osmotic shock with sorbitol would affect CCM1/CCM2 association, as CCM2 has been demonstrated to be a scaffold for MEKK3-dependent MAPK activation upon sorbitol stimulation (12). Treatment of cells co-expressing YFP-CCM1 and CFP-CCM2 with sorbitol produced a 2-fold increase in the FRETN^C value relative to non-treated cells, yielding a FRETN^C value of 15.77x10^–5 (Table 1). Localization of both YFP-CCM1 and CFP-CCM2 shifted from the cytoplasm to the cell periphery as a consequence of sorbitol stimulation (Fig. 2B). In addition to the sorbitol-induced re-localization of CCM1 and CCM2 to the cell periphery, the proteins interact at this site as revealed by visualization of the corrected FRET signal (Fig. 2B).

Endogenous CCM1 and CCM2 are in a complex
To confirm the interaction of the CCM1 and CCM2 proteins, we performed co-immunoprecipitation experiments of the endogenous proteins. Protein complexes from lysates of MEFs were immunoprecipitated with a polyclonal CCM2 antibody and analyzed for the presence of endogenous CCM1 by western blot. Endogenous CCM1 efficiently co-immunoprecipitated with CCM2 (Fig. 3A).

View larger version (30K): [in this window] [in a new window]   Figure 3. The CCM1/2 interaction is PTB domain-dependent. (A) Endogenous CCM1 co-immunoprecipitates with endogenous CCM2. Following immunoprecipitation of a MEF lysate with anti-CCM2 and western blotting with anti-CCM1, CCM1 is detected as a species migrating near the predicted molecular weight of 83 kDa. About 100 µg of 293T cell lysate transfected with FLAG-CCM1 was utilized as a size control for endogenous CCM1. (B) Engineered and CCM2 patient PTB domain missense mutations abrogate the CCM1/2 interaction. FLAG-CCM1 and FLAG-CCM2 with the indicated mutations were expressed in 293T cells, and complexes were immunoprecipitated with anti-CCM2 followed by western blotting with anti-FLAG. The engineered CCM2 PTB domain mutation F217A and a CCM2 patient PTB domain mutation L189R (11) completely prevent FLAG-CCM1 from being co-immunoprecipitated with FLAG-CCM2, whereas NPxY mutation N192A in FLAG-CCM1 does not perturb the interaction. Mutant FLAG-CCM2 F217A and FLAG-CCM2 L198R proteins were produced in cells as efficiently as wild-type FLAG-CCM2 and immunoprecipitated as efficiently as wild-type FLAG-CCM2 with anti-CCM2.

 
The presence of a PTB domain in CCM2 and an NPxY motif in CCM1 suggested a mechanism for CCM1/CCM2 interaction. The function of the CCM2 PTB domain on FLAG-CCM1/FLAG-CCM2 co-immunoprecipitation was tested by mutagenesis of a residue conserved among Dab/Shc-like PTB domains, which resides in the PTB pocket critical for binding to the NPxY peptide (21). Point mutation of the corresponding residue in CCM2 (F217A) completely abrogated the ability of FLAG-CCM1 to co-immunoprecipitate with FLAG-CCM2 (Fig. 3B), indicating that the CCM2 PTB domain is critical for CCM1/2 interaction.

We also used FRET to analyze the effect of the F217A CCM2 mutation on the CCM1/2 interactions in living cells. COS-7 cells co-expressing YFP-CCM1 and CFP-CCM2 F217A had markedly reduced FRET^C (FRETN^C=1.355x10^–5) when compared with wild-type CCM1/2 proteins (FRETN^C=7.43x10^–5) (Table 1). A similar reduction of FRET^C was observed with sorbitol stimulation (CCM2-F217A FRETN^C=2.29x10^–5 and wild-type FRETN^C=15.77x10^–5) (Table 1). Thus, both loss of FRET^C in cells and co-immunoprecipitation from lysates demonstrate the CCM2 PTB domain requirement for CCM1/2 interactions.

Germline mutations in the CCM1 and CCM2 genes have been primarily nonsense, frameshift and splicing mutations predicted to prematurely truncate the message. This CCM mutation profile has been devoid of missense mutations, useful for gaining insight into structurally and functionally critical CCM residues or domains directly relevant to CCM pathogenesis. However, a single missense mutation, L198R, has recently been identified in a CCM2 kindred (11). Similar to the engineered F217A mutation, L198R resides in the C-terminal region of the CCM2 PTB domain. When co-expressed with CCM1 protein, this patient mutation in CCM2 also abolished co-immunoprecipitation of CCM1 (Fig. 3B). The findings suggest that the L198R patient mutation is sufficient to inhibit CCM1/2 interaction.

Both CCM2 and ICAP1 interact with CCM1 via their PTB domains. CCM1 encodes three NPxY motifs, suggesting that CCM2 and ICAP1 could bind to the same or different sites in the CCM1 protein. We tested this hypothesis by co-immunoprecipitation and found that mutagenesis of the critical asparagine (N192A) residue in the CCM1 NPxY motif critical for ICAP1 binding (16,17) did not disrupt the CCM1/2 association (Fig. 3B). FRET interaction values for CCM1-N192A and CCM2 were diminished compared with wild-type CCM1 and CCM2, but the interaction was not completely eliminated (Table 1). The FRET data indicate that the N192A CCM1 mutation has a partial effect on CCM1/2 binding, whereas in the context of co-immunoprecipitation, it does not perturb the interaction. Taken together, these data suggest that CCM2 can bind to additional motifs in CCM1.

CCM1 localization is influenced by association with CCM2
We hypothesized that association with CCM2 would influence CCM1 subcellular localization. In COS-7 cells, CCM1 exhibits whole-cell localization when expressed as a YFP fusion, whereas CCM2 exhibits exclusively cytoplasmic localization when expressed as a CFP fusion. In cells co-expressing YFP-CCM1 and CFP-CCM2, YFP-CCM1 nuclear localization is lost and becomes exclusively cytoplasmic (Fig. 4A). This indicates that CCM2 is capable of interacting with and sequestering CCM1, directly or indirectly, in the cytoplasm. To see whether the cytoplasmic sequestration of CCM1 by CCM2 was dependent on the PTB domain of CCM2, we analyzed the effect of the CCM2 missense mutation F217A. In COS-7 cells co-expressing YFP-CCM1 and PTB-mutant CFP-CCM2 F217A, CCM2 fails to interact with and sequester CCM1, thus nuclear localization of YFP-CCM1 is restored (Fig. 4A). These data are concordant with the co-immunoprecipitation results and provide support that the CCM gene products interact in a PTB domain-dependent manner. These data also suggest that CCM1 has nuclear-cytoplasmic shuttling capability.

View larger version (38K): [in this window] [in a new window]   Figure 4. CCM1 subcellular localization is influenced by association with binding partner CCM2 and by a CCM1 NLS. (A) The F217A PTB mutation in CCM2 prevents interaction and sequestration of CCM1 in the cytoplasm. COS-7 cells were transfected with CFP-CCM2 or PTB-mutant CFP-CCM2 F217A with YFP-CCM1 and fluorescence was examined in live cells at room temperature. YFP-CCM1 fluorescence is whole-cell, whereas CFP-CCM2 fluorescence is predominantly cytoplasmic. In cells co-expressing YFP-CCM1 and CFP-CCM2 (arrows), CCM2 is capable of interacting with and sequestering CCM1 in the cytoplasm (top panels). This sequestration phenotype is lost in cells co-expressing PTB-mutant CFP-CCM2 F217A and YFP-CCM1, as YFP-CCM1 exhibits wild-type, whole-cell fluorescence (bottom panels). (B) CCM1 possesses a functional NLS. Wild-type FLAG-CCM1 in COS-7 cells localizes to both the nucleus and the cytoplasm, whereas FLAG-CCM1 NLS mutant 2 (KKRKKKKAAAAK) is predominantly cytoplasmic in the majority of cells as revealed by anti-FLAG immunofluorescence. Magnification: 400x. (C) Percentage of cells exhibiting FLAG-CCM1 nuclear localization following NLS mutagenesis. FLAG-CCM1 NLS mutant 1 (KKRKKKKKAAKK) reduces the number of cells exhibiting nuclear localization to 40%, whereas the more severe mutation NLS mutant 2 reduces the number of cells exhibiting nuclear localization to 10% (n=150 cells).

 
CCM1 harbors a functional nuclear localization sequence
In addition to the contribution of CCM2 association to CCM1 localization, we reasoned that YFP-CCM1 nuclear localization may be dependent on a nuclear localization sequence (NLS). To identify a putative NLS in CCM1, we submitted the amino acid sequence to the online PSORT II program (EXPASY). A ‘pat4’ four-residue pattern, similar to the NLS of the SV40 large T antigen, was predicted in the unique N-terminal region of CCM1 (corresponding to 46-KKKRKK-51). In addition, a ‘pat4’ sequence was detected at residue 569 within the FERM domain of CCM1 (corresponding to KKHK).

The putative N-terminal NLS in CCM1 was altered by site-directed mutagenesis from the wild-type KKKRKK to KKAAKK (NLS mutant 1) and also to a more severe mutation of KAAAAK (NLS mutant 2). The resulting constructs were transfected into COS-7 cells and assessed for nuclear localization (Fig. 4B). The percentage of cells displaying nuclear localization for FLAG-CCM1 NLS mutant 1 was reduced to 40% from the 90% of wild-type FLAG-CCM1 cells displaying nuclear localization, whereas nuclear localization for FLAG-CCM1 NLS mutant 2 was reduced to 10% of the cells (Fig. 4C). These data suggest that the putative N-terminal NLS in CCM1 is functional and are also consistent with a nuclear-cytoplasmic shuttling role for CCM1.

CCM1/CCM2 interaction is not dependent on the same CCM1 NPxY sequence critical for ICAP1 interaction
Demonstration of a PTB domain-dependent CCM1/CCM2 interaction raised the possibility of competition between ICAP1 and CCM2 for CCM1 binding via the N192-Y195 NPxY sequence in CCM1. Figure 3B shows that mutation of the CCM1 NPxY sequence critical for ICAP1 binding does not abrogate CCM2 binding. This finding suggests that CCM2 may bind elsewhere on the CCM1 protein, potentially allowing concurrent ICAP1 and CCM2 binding with CCM1 in a ternary complex. To test this hypothesis, we performed a CCM1/2 co-immunoprecipitation assay in the presence of increasing amounts of ICAP1. ICAP1 overexpression did not prevent CCM1 from being efficiently co-immunoprecipitated with CCM2, instead, all the three molecules were present in the complex precipitated with the CCM2 antibody (Fig. 5A). The findings suggest that a CCM1/CCM2/ICAP1 ternary complex may exist in cells.

View larger version (35K): [in this window] [in a new window]   Figure 5. ICAP1 complexes with the CCM1 and CCM2 proteins and sequesters CCM1 in the nucleus. (A) ICAP1 does not compete with CCM2 for binding to CCM1, but is instead a member of a ternary complex with CCM1 and CCM2. 293T cells were transiently transfected with the indicated constructs. Cell extracts were analyzed by immunoprecipitation using a polyclonal antibody to CCM2 followed by western blotting with anti-FLAG. (B) ICAP1 sequesters CCM1 in the nucleus. COS-7 cells were transfected with YFP-CCM1 alone and co-transfected with CFP-ICAP1. Live COS-7 cells transfected with YFP-CCM1 alone exhibit whole-cell fluorescence, whereas YFP-CCM1 is sequestered in the nucleus when co-expressed with CFP-ICAP1. (C) Molecular model for CCM protein function. The CCM1/2 complex may modulate integrin-mediated cell adhesion via association with ICAP1 in the cytoplasm, as well as organize a p38 MAPK signaling complex. CCM1/ICAP1 complex may function in the nucleus and may be recruited by CCM2 to function in the cytoplasm.

 
As mutation of the N192-Y195 NPxY motif in CCM1 did not abrogate association with CCM2, alternative CCM1 NPxY motifs may be the sites of interaction with CCM2. We tested this possibility by mutation of the N231-F234 and N250-F253 NPxY-like sequences within CCM1. Single asparagine to alanine (N231A or N250A) or single phenylalanine to alanine (F234A or F253A) substitutions in FLAG-CCM1 did not prevent association with FLAG-CCM2 in co-immunoprecipitation assays (data not shown), suggesting that these NPxF sites are not strong sites of CCM2 interaction.

We have demonstrated that the expression of CCM2 influences the localization of CCM1; namely, CCM2 is capable of sequestering CCM1 in the cytoplasm. ICAP1 may similarly influence CCM1 localization. CFP-ICAP1 exhibits predominantly nuclear localization upon overexpression. When co-expressed with CFP-ICAP1, YFP-CCM1 localization shifts from whole-cell to predominantly nuclear (Fig. 5B). Thus, in this system, CCM2 or ICAP1 bound to CCM1 via their respective PTB domains determines the subcellular location of CCM1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
The identification of MGC4607 as the CCM2 gene (10,11) and concurrent biochemical studies with the murine ortholog OSM (12) have provided the opportunity to integrate the current model of CCM1 and ICAP1 with new molecular pathways, thus providing a more comprehensive picture of the pathogenesis of CCM. Here, we define a CCM protein complex by demonstrating an interaction between the CCM1 and the CCM2 gene products. This interaction is abrogated by a CCM2 patient missense mutation, L198R (11), suggesting that loss of the CCM1/2 association contributes to the development of the abnormal vessel architecture in a CCM lesion. Our hypothesis is that the CCM complex has distinct functions for integrin signaling via ICAP1 and for the control of the p38 pathway.

The work demonstrating OSM/CCM2 as a scaffold for Rac/MEKK3/MKK3 in the p38 MAPK module (12) has provided a new signal transduction branch point to examine in the context of CCMs pathogenesis. RNAi knockdown studies indicated that OSM and MEKK3 were required for activation of p38 in response to hyperosmotic stress (12). Our results presented here are concordant, as CCM2 heterozygous MEFs have significantly decreased levels of active, phosphorylated p38. We also demonstrated that CCM1 can be detected with MEKK3 as a multi-protein complex involving CCM1 and CCM2 proteins. Thus, a function of the CCM1/2 complex may be to organize and regulate p38 MAPK signaling.

CCM1 nuclear localization has not been described previously. CCM1 localization in the nucleus and cytoplasm suggests a dual role for the CCM1 protein, with nuclear functions and functions at sites of integrin localization in membranes. The relative abundance of CCM2 and ICAP1 or their binding status with other proteins may dictate CCM1 subcellular location and/or function at a given time. We propose that ICAP1 may tether CCM1 in the nucleus until the presence of a stimulus, whereby CCM2 serves as a cytoplasmic anchor and recruits CCM1 to a signaling complex for integrin and p38 MAPK-associated functions. CCM2 sequestration of CCM1 in the cytoplasm would therefore effectively coordinate p38 signaling with CCM1 function. The dynamics of CCM1 nuclear-cytoplasmic localization suggest that the formation of the CCM1/2 complex is tightly regulated. Loss of either CCM1 or CCM2 in patients would perturb the coordinated assembly of a CCM1/2 complex. Failure of the CCM1/2 complex to properly assemble may in turn result in dysregulated p38 MAPK function and loss of regulation of integrin-mediated cell adhesion via ICAP1.

It remains unclear what role CCM1 and ICAP1 may play in a transient nuclear location. It is possible that the nuclear location may be a means to regulate cytoplasmic CCM protein function by compartmentalization. Alternatively, CCM1 and ICAP1 may possess distinct nuclear functions. A recent report has implicated ICAP1 as a shuttling protein that binds to integrins and also translocates to the nucleus, affecting cell proliferation in a cell adhesion-dependent manner (22). In this study, ICAP1 was shown to regulate transcription of the c-myc promoter in concert with the protein Nm23-H2. Interestingly, ICAP1 subcellular localization was influenced by its binding partner ß₁ integrin—overexpression of ß₁ integrin shifted the nuclear pool of ICAP1 toward the cytoplasm. Therefore, ICAP1 function is dual in nature, where ß₁ integrin serves as a cytoplasmic anchor for ICAP1, although ICAP1 is also capable of regulating transcription in the nucleus with its binding partner Nm23-H2. This molecular interplay may be analogous to that described for CCM1 in this report. The CCM1/ICAP1 complex may function in the nucleus, whereby CCM2 recruits and anchors CCM1 or the CCM1/ICAP1 complex in the cytoplasm as part of a signaling complex for MAPK or integrin-mediated signal transduction events (Fig. 5C).

The nuclear localization for CCM1 described in this report and the experiments demonstrating nuclear-cytoplasmic shuttling contrast with results from a previous study describing subcellular localization of endogenous CCM1. Immunofluorescence of endogenous CCM1 protein using a peptide antibody to CCM1 protein yielded a microtubule staining pattern (23). This discrepancy in CCM1 subcellular localization may reflect differences between endogenous and overexpressed CCM1 protein. However, this discrepancy must also be explained in the light of a related discrepancy involving the molecular weight of the authentic CCM1 protein. The antiserum that exhibits a microtubule staining pattern detects a 58 kDa species for the endogenous CCM1 protein, which is significantly smaller than the predicted molecular weight of 83 kDa (23,24). In contrast, our CCM1 antibody detects an endogenous species of the predicted molecular weight following immunoprecipitation with CCM2 antibody. Clearly, additional antibody reagents will be required to resolve these issues. Despite these caveats, nuclear localization of CCM1 is consistent with the previously described nuclear localization of its binding partner ICAP1. Nuclear-cytoplasmic shuttling is also supported by the presence of a functional NLS in the protein.

What may be the consequences of perturbed p38 activation in the context of CCMs? It is well-established that p38 signaling plays an important role during angiogenesis. Targeted disruption of p38 (Mapk14) gene in mice results in embryonic lethality owing to placental defects, specifically lack of vascularization and increased apoptosis, as well as abnormal angiogenesis in the embryo (25). Furthermore, embryos null for Mekk3 die from impaired blood vessel development (26). Reduced p38 activation in response to cellular stress in cells deficient for CCM2 may influence downstream p38-specific transcriptional activation critical for the organization of new vessels and for the maintenance of the existing vessel architecture and may ultimately contribute to the formation of the closely packed, malformed vessels of the cavernous malformation.

In addition to the p38 MAPK signaling pathway, integrin signaling has been suggested to play a role in CCM pathogenesis on the basis of the CCM1/ICAP1 interaction (16,17). To study the role of the CCM proteins in integrin-mediated cell adhesion, we view the sorbitol stimulation experiments as a general tool for altering cell morphology. Upon sorbitol treatment, cells exhibit rapid shrinkage and cell shape changes as a consequence of an efflux of water from the cell. During this process, the formation and dissolution of focal contacts and focal adhesions are dynamic and may be a model to study the role of the CCM proteins in integrin signaling. YFP-CCM1 and CFP-CCM2 migration to the cell periphery during osmotic shock, in addition to reflecting a CCM2 scaffold harnessing Rac1/MEKK3/MKK3 (12), may reflect roles of the CCM1 and CCM2 proteins at sites of integrin engagement. Perturbation of CCM1, CCM2 or CCM1/2 protein complex integrin signaling functions in patients may, in addition to impaired p38 signaling, contribute to the formation of the CCM lesions.

Here, we demonstrate that the phenotypically linked CCM proteins 1 and 2 are in a complex likely involving both MAPK and integrin signaling, although each protein may have distinct individual roles. It remains to be determined whether CCM3 will be a member of the CCM1/2 protein complex defined so far. The nuclear-cytoplasmic shuttling data for CCM1 and stimulus-dependent re-localization observed for both CCM1 and CCM2 argue that the CCM complex is dynamic, responding to different signaling cues. Assaying the different cellular functions of the CCM complex as its protein composition changes will provide further insight into the molecular mechanisms of CCM pathogenesis.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Plasmids and cell culture
Full-length cDNAs for human KRIT1 and ICAP1 were isolated by RT–PCR from HUVEC RNA, N-terminally FLAG-tagged by PCR and subsequently cloned into the expression vector pRK5 (27). Full-length cDNA for malcavernin was N-terminally FLAG-tagged by PCR using human IMAGE Consortium clone 3501896 was used as template and subsequently cloned into pRK5. For FRET/subcellular localization constructs, full-length human KRIT1 was fused in-frame to the coding sequence of EYFP of pEYFPC1 (Clontech) and full-length human ICAP1 was fused in-frame to the coding sequence of ECFP of pECFPC1 (Clontech). CFP-OSM and hemagglutinin (HA)-MEKK3 have been described previously (12). Missense mutations were made using QuickChange^TM (Stratagene) site-directed mutagenesis kit and confirmed by sequencing.

293T cells and COS-7 cells were maintained in Dulbecco's modified Eagle medium (DMEM) (Invitrogen)/10% fetal bovine serum (FBS) and supplemented with amphotericin B (GensiaSicor Pharmaceuticals) and penicillin/streptomycin (Invitrogen) at the manufacturers' recommended concentrations.

Immunoprecipitation and western blotting
For co-immunoprecipitation assays, 293T cells were transfected with the constructs indicated in the figure legends using FuGene6^TM reagent (Roche). Transfected 293T cells or MEFs were lysed in a buffer containing 0.6% Triton X-100, 137 mM NaCl, 10% glycerol and 20 mM Tris pH 7.4 supplemented with Complete^TM protease inhibitor tablets (Roche). About 500 µg of cell lysate was incubated with a rabbit polyclonal CCM2 (OSM) antibody (12) with gentle rocking at 4°C followed by incubation with protein G-sepharose 4B beads (Amersham). The beads were washed six times with lysis buffer, boiled for 5 min in SDS-loading buffer and subjected to SDS–PAGE and western blotting with anti-FLAG (M2, Sigma), anti-HA (Sigma) or affinity-purified anti-CCM1 (raised against the peptide H38-L53, C*HEVPIEGQKKKRKKVL, Bethyl Laboratories, Inc.).

Fluorescence resonance energy transfer
For live-cell fluorescence microscopy and CYFRET experiments, COS-7 cells were grown on 25 mm round glass coverslips. Cells were subsequently transfected using Lipofectamine Plus^TM (Invitrogen). After 24 h, cells were placed in a media-filled imaging chamber (Molecular Probes) at room temperature. Cells were treated with 0.2 M sorbitol for the indicated times. CYFRET image acquisition and analysis were performed by the three-filter micro-FRET image subtraction method as described previously (28). Briefly, images for YFP excitation/YFP emission, CFP excitation/CFP emission and CFP excitation/YFP emission (raw, uncorrected CYFRET) were obtained (29). Background-subtracted YFP and CFP images were then fractionally subtracted from raw CYFRET images on the basis of measurements for CFP bleed-through (CFP-OSM=0.55 and CFP-OSM F217A=0.52) and YFP cross-excitation (YFP-CCM1=0.021 and YFP-CCM1 N192A=0.024). This fractional subtraction generated corrected FRET^C images that function as visual representations of interacting proteins.

MEF isolation
E10.5 primary MEF cultures were established from embryos heterozygous for a LacZ gene trap insertion in the mouse Ccm2 gene (cell line RRG051, Bay Genomics, allele designation Ccm2^{Gt(pGt0Lxf)1Dmar}). Cells were maintained in Iscove's modified Dulbecco's medium (Invitrogen)/10% FBS and supplemented with penicillin/streptomycin (Invitrogen) and amphotericin B (GensiaSicor Pharmaceuticals) at the manufacturers' recommended concentrations. Established CCM2 MEF lines were genotyped by PCR using the following primers: 5'-CATCCCTGTCTGGGAACCTA-3' (intron 6, reverse), 5'-TCTAGGACAAGAGGGCGAGA-3' (gene trap, reverse) and 5'-GAAGAGTTGTGCTCCCTGCT-3' (exon 6, forward).

CCM2 MEF lysates were subjected to western blotting with anti-CCM2 (OSM) (12) and anti-ERK1/2 (K-23, Santa Cruz) as a loading control to confirm reduction of CCM2 protein in CCM2^+/– cells.

MAPK activation assay
Spontaneously immortalized (greater than passage 15) CCM2^+/– MEFs were serum-starved in OptiMEM^TM (Invitrogen) overnight. Cells were treated with 0.2 M sorbitol or 10 µg ml^–1 anisomycin for the indicated times and harvested in lysis buffer containing 0.6% Triton X-100, 137 mM NaCl, 10% glycerol and 20 mM Tris pH 7.4 supplemented with Complete^TM protease inhibitor tablets (Roche). About 50 µg of each cell lysate was subjected to SDS–PAGE and western blotting with anti-p38 MAP kinase (Santa Cruz Biotechnology no. sc-7149/H-147) and anti-phospho-p38 MAP kinase (Thr180/Tyr182) (Cell Signaling Technology no. 9211). Thr180/Tyr182 phosphorylated p38 was quantified using densitometry and normalized to levels of total p38 present to yield an activation index. Fold induction was calculated as the activation index of the sample divided by the activation index of the sample not treated with anisomycin or sorbitol. For add-back experiments, CCM2^+/– MEFs were transfected with pRK5-FLAG-CCM2 using Nucleofector^TM electroporation technology (Amaxa Biosystems). After 24 h, cells were serum-starved and treated with sorbitol, lysed and subjected to SDS–PAGE/western blotting as described earlier.

Immunofluorescence
Immunofluorescence microscopy was performed on COS-7 cells grown on human fibronectin coverslips (Becton Dickinson). Cells were washed with phosphate-buffered saline (PBS), fixed for 14 min in 3% paraformaldehyde/sucrose/PBS, washed three times with PBS and permeabilized for 6 min in 0.1% Triton X-100/PBS. Coverslips were washed an additional three times and blocked for 30 min with DMEM containing 10% goat serum. Coverslips were incubated for 2 h at room temperature with anti-FLAG M2 (Sigma) diluted in blocking solution. After washing with PBS, coverslips were incubated for 1 h at room temperature with goat anti-mouse IgG-AlexaFluor^TM 594 (1:1000, Molecular Probes^TM, Invitrogen). After final washes with PBS, coverslips were mounted with ProLong Gold^TM antifade reagent (Molecular Probes^TM, Invitrogen).

	ACKNOWLEDGEMENTS

 
The authors thank Cheryl Bock of the Duke University Medical Center Transgenic Mouse Facility for the generation of the Ccm2^{Gt(pGt0Lxf)1Dmar} gene trap mice and Nicholas Plummer for critical reading of the manuscript. This work was supported by an American Heart Association Bugher Foundation award and NIH grant NS43543 (D.A.M.), NIH grants GM68820 and DK37871 (G.L.J.) and by an American Heart Association predoctoral fellowship award (J.S.Z.).

Conflict of Interest statement. None declared.

	FOOTNOTES

 
The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 

1. Gil-Nagel, A., Wilcox, K.J., Stewart, J.M., Anderson, V.E., Leppik, I.E. and Rich, S.S. (1995) Familial cerebral cavernous angioma: clinical analysis of a family and phenotypic classification. Epilepsy Res., 21, 27–36.[CrossRef][ISI][Medline]

2. Zabramski, J.M., Wascher, T.M., Spetzler, R.F., Johnson, B., Golfinos, J., Drayer, B.P., Brown, B., Rigamonti, D. and Brown, G. (1994) The natural history of familial cavernous malformations: results of an ongoing study. J. Neurosurg., 80, 422–432.[ISI][Medline]

3. Kondziolka, D., Lunsford, L.D. and Kestle, J.R. (1995) The natural history of cerebral cavernous malformations. J. Neurosurg., 83, 820–824.[ISI][Medline]

4. Dubovsky, J., Zabramski, J.M., Kurth, J., Spetzler, R.F., Rich, S.S., Orr, H.T. and Weber, J.L. (1995) A gene responsible for cavernous malformations of the brain maps to chromosome 7q. Hum. Mol. Genet., 4, 453–458.[Abstract/Free Full Text]

5. Gunel, M., Awad, I.A., Anson, J. and Lifton, R.P. (1995) Mapping a gene causing cerebral cavernous malformation to 7q11.2–q21. Proc. Natl Acad. Sci. USA, 92, 6620–6624.[Abstract/Free Full Text]

6. Marchuk, D.A., Gallione, C.J., Morrison, L.A., Clericuzio, C.L., Hart, B.L., Kosofsky, B.E., Louis, D.N., Gusella, J.F., Davis, L.E. and Prenger, V.L. (1995) A locus for cerebral cavernous malformations maps to chromosome 7q in two families. Genomics, 28, 311–314.[CrossRef][ISI][Medline]

7. Craig, H.D., Gunel, M., Cepeda, O., Johnson, E.W., Ptacek, L., Steinberg, G.K., Ogilvy, C.S., Berg, M.J., Crawford, S.C., Scott, R.M. et al. (1998) Multilocus linkage identifies two new loci for a Mendelian form of stroke, cerebral cavernous malformation, at 7p15–13 and 3q25.2–27. Hum. Mol. Genet., 7, 1851–1858.[Abstract/Free Full Text]

8. Laberge-le Couteulx, X.S., Jung, H.H., Labauge, P., Houtteville, J.P., Lescoat, C., Celillon, M., Marechal, E., Joutel, A., Bach, J.F. and Tournier-Lasserve, E. (1999) Truncating mutations in CCM1, encoding KRIT1, cause hereditary cavernous angiomas. Nat. Genet., 23, 189–193.[CrossRef][ISI][Medline]

9. Sahoo, T., Johnson, E.W., Thomas, J.W., Kuehl, P.M., Jones, T.L., Dokken, C.G., Touchman, J.W., Gallione, C.J., Lee-Lin, S.Q., Kosofsky, B. et al. (1999) Mutations in the gene encoding KRIT1, a Krev-1/rap1a binding protein, cause cerebral cavernous malformations. Hum. Mol. Genet., 8, 2325–2333.[Abstract/Free Full Text]

10. Liquori, C.L., Berg, M.J., Siegel, A.M., Huang, E., Zawistowski, J.S., Stoffer, T., Verlaan, D., Balogun, F., Hughes, L., Leedom, T.P. et al. (2003) Mutations in a gene encoding a novel protein containing a phosphotyrosine-binding domain cause type 2 cerebral cavernous malformations. Am. J. Hum. Genet., 73, 1459–1464.[CrossRef][ISI][Medline]

11. Denier, C., Goutagny, S., Labauge, P., Krivosic, V., Arnoult, M., Cousin, A., Benabid, A.L., Comoy, J., Frerebeau, P., Gilbert, B. et al. (2004) Mutations within the MGC4607 gene cause cerebral cavernous malformations. Am. J. Hum. Genet., 74, 326–337.[CrossRef][ISI][Medline]

12. Uhlik, M.T., Abell, A.N., Johnson, N.L., Sun, W., Cuevas, B.D., Lobel-Rice, K.E., Horne, E.A., Dell'Acqua, M.L. and Johnson, G.L. (2003) Rac-MEKK3-MKK3 scaffolding for p38 MAPK activation during hyperosmotic shock. Nat. Cell Biol., 5, 1104–1110.[CrossRef][ISI][Medline]

13. Bergametti, F., Denier, C., Labauge, P., Arnoult, M., Boetto, S., Clanet, M., Coubes, P., Echenne, B., Ibrahim, R., Irthum, B. et al. (2005) Mutations within the programmed cell death 10 gene cause cerebral cavernous malformations. Am. J. Hum. Genet., 76, 42–51.[CrossRef][ISI][Medline]

14. Laurans, M.S., DiLuna, M.L., Shin, D., Niazi, F., Voorhees, J.R., Nelson-Williams, C., Johnson, E.W., Siegel, A.M., Steinberg, G.K., Berg, M.J. et al. (2003) Mutational analysis of 206 families with cavernous malformations. J. Neurosurg., 99, 38–43.[ISI][Medline]

15. Serebriiskii, I., Estojak, J., Sonoda, G., Testa, J.R. and Golemis, E.A. (1997) Association of Krev-1/rap1a with Krit1, a novel ankyrin repeat-containing protein encoded by a gene mapping to 7q21–22. Oncogene, 15, 1043–1049.[CrossRef][ISI][Medline]

16. Zawistowski, J.S., Serebriiskii, I.G., Lee, M.F., Golemis, E.A. and Marchuk, D.A. (2002) KRIT1 association with the integrin-binding protein ICAP-1: a new direction in the elucidation of cerebral cavernous malformations (CCM1) pathogenesis. Hum. Mol. Genet., 11, 389–396.[Abstract/Free Full Text]

17. Zhang, J., Clatterbuck, R.E., Rigamonti, D., Chang, D.D. and Dietz, H.C. (2001) Interaction between krit1 and icap1alpha infers perturbation of integrin beta1-mediated angiogenesis in the pathogenesis of cerebral cavernous malformation. Hum. Mol. Genet., 10, 2953–2960.[Abstract/Free Full Text]

18. Chang, D.D., Wong, C., Smith, H. and Liu, J. (1997) ICAP-1, a novel beta1 integrin cytoplasmic domain-associated protein, binds to a conserved and functionally important NPXY sequence motif of beta1 integrin. J. Cell Biol., 138, 1149–1157.[Abstract/Free Full Text]

19. Zhang, X.A. and Hemler, M.E. (1999) Interaction of the integrin beta1 cytoplasmic domain with ICAP-1 protein. J. Biol. Chem., 274, 11–19.[Abstract/Free Full Text]

20. Chang, D.D., Hoang, B.Q., Liu, J. and Springer, T.A. (2002) Molecular basis for interaction between Icap1 alpha PTB domain and beta 1 integrin. J. Biol. Chem., 277, 8140–8145.[Abstract/Free Full Text]

21. Uhlik, M.T., Temple, B., Bencharit, S., Kimple, A.J., Siderovski, D.P. and Johnson, G.L. (2005) Structural and evolutionary division of phosphotyrosine binding (PTB) domains. J. Mol. Biol., 345, 1–20.[CrossRef][ISI][Medline]

22. Fournier, H.N., Dupe-Manet, S., Bouvard, D., Luton, F., Degani, S., Block, M.R., Retta, S.F. and Albiges-Rizo, C. (2005) Nuclear translocation of integrin cytoplasmic domain-associated protein 1 stimulates cellular proliferation. Mol. Biol. Cell, 16, 1859–1871.[Abstract/Free Full Text]

23. Gunel, M., Laurans, M.S., Shin, D., DiLuna, M.L., Voorhees, J., Choate, K., Nelson-Williams, C. and Lifton, R.P. (2002) KRIT1, a gene mutated in cerebral cavernous malformation, encodes a microtubule-associated protein. Proc. Natl Acad. Sci. USA, 99, 10677–10682.[Abstract/Free Full Text]

24. Guzeloglu-Kayisli, O., Amankulor, N.M., Voorhees, J., Luleci, G., Lifton, R.P. and Gunel, M. (2004) KRIT1/cerebral cavernous malformation 1 protein localizes to vascular endothelium, astrocytes, and pyramidal cells of the adult human cerebral cortex. Neurosurgery, 54, 943–949; discussion 949.[ISI][Medline]

25. Mudgett, J.S., Ding, J., Guh-Siesel, L., Chartrain, N.A., Yang, L., Gopal, S. and Shen, M.M. (2000) Essential role for p38alpha mitogen-activated protein kinase in placental angiogenesis. Proc. Natl Acad. Sci. USA, 97, 10454–10459.[Abstract/Free Full Text]

26. Yang, J., Boerm, M., McCarty, M., Bucana, C., Fidler, I.J., Zhuang, Y. and Su, B. (2000) Mekk3 is essential for early embryonic cardiovascular development. Nat. Genet., 24, 309–313.[CrossRef][ISI][Medline]

27. Naga Prasad, S.V., Laporte, S.A., Chamberlain, D., Caron, M.G., Barak, L. and Rockman, H.A. (2002) Phosphoinositide 3-kinase regulates beta2-adrenergic receptor endocytosis by AP-2 recruitment to the receptor/beta-arrestin complex. J. Cell Biol., 158, 563–575.[Abstract/Free Full Text]

28. Sorkin, A., McClure, M., Huang, F. and Carter, R. (2000) Interaction of EGF receptor and grb2 in living cells visualized by fluorescence resonance energy transfer (FRET) microscopy. Curr. Biol., 10, 1395–1398.[CrossRef][ISI][Medline]

29. Oliveria, S.F., Gomez, L.L. and Dell'Acqua, M.L. (2003) Imaging kinase—AKAP79—phosphatase scaffold complexes at the plasma membrane in living cells using FRET microscopy. J. Cell Biol., 160, 101–112.[Abstract/Free Full Text]

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				A. Glading, J. Han, R. A. Stockton, and M. H. Ginsberg KRIT-1/CCM1 is a Rap1 effector that regulates endothelial cell cell junctions J. Cell Biol., October 22, 2007; 179(2): 247 - 254. [Abstract] [Full Text] [PDF]

				P. Brouillard and M. Vikkula Genetic causes of vascular malformations Hum. Mol. Genet., October 15, 2007; 16(R2): R140 - R149. [Abstract] [Full Text] [PDF]

				D. A. Buchner, F. Su, J. S. Yamaoka, M. Kamei, J. A. Shavit, L. K. Barthel, B. McGee, J. D. Amigo, S. Kim, A. W. Hanosh, et al. From the Cover: pak2a mutations cause cerebral hemorrhage in redhead zebrafish PNAS, August 28, 2007; 104(35): 13996 - 14001. [Abstract] [Full Text] [PDF]

				S. Battistini, R. Rocchi, A. Cerase, A. Citterio, L. Tassi, G. Lando, M. C. Patrosso, R. Galli, P. Brunori, D. L. Sgro, et al. Clinical, Magnetic Resonance Imaging, and Genetic Study of 5 Italian Families With Cerebral Cavernous Malformation Arch Neurol, June 1, 2007; 64(6): 843 - 848. [Abstract] [Full Text] [PDF]

				X. Ma, H. Zhao, J. Shan, F. Long, Y. Chen, Y. Chen, Y. Zhang, X. Han, and D. Ma PDCD10 Interacts with Ste20-related Kinase MST4 to Promote Cell Growth and Transformation via Modulation of the ERK Pathway Mol. Biol. Cell, June 1, 2007; 18(6): 1965 - 1978. [Abstract] [Full Text] [PDF]

				N Revencu and M Vikkula Cerebral cavernous malformation: new molecular and clinical insights J. Med. Genet., September 1, 2006; 43(9): 716 - 721. [Abstract] [Full Text] [PDF]

				J. D. Mably, L. P. Chuang, F. C. Serluca, M.-A. P. K. Mohideen, J.-N. Chen, and M. C. Fishman santa and valentine pattern concentric growth of cardiac myocardium in the zebrafish Development, August 15, 2006; 133(16): 3139 - 3146. [Abstract] [Full Text] [PDF]

				A. Seker, K. L. Pricola, B. Guclu, A. K. Ozturk, A. Louvi, and M. Gunel CCM2 Expression Parallels That of CCM1 Stroke, February 1, 2006; 37(2): 518 - 523. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) All Versions of this Article: 14/17/2521    most recent ddi256v2 ddi256v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (38) Request Permissions Google Scholar