The Wiskott-Aldrich syndrome: The actin cytoskeleton and immune cell function

Disease Markers 29 (2010) 157–175 157 DOI 10.3233/DMA-2010-0735 IOS Press
The Wiskott-Aldrich syndrome: The actin cytoskeleton and immune cell function
Michael P. Blundella, Austen Wortha,b, Gerben Boumaa and Adrian J. Thrashera,b,∗
aMolecular Immunology Unit, UCL Institute of Child Health, London, UK bDepartment of Immunology, Great Ormond Street Hospital NHSTrust, Great Ormond Street, London, UK
Abstract. Wiskott-Aldrich syndrome (WAS) is a rare X-linked recessive primary immunodeficiency characterised by immune dysregulation, microthrombocytopaenia, eczema and lymphoid malignancies. Mutations in the WAS gene can lead to distinct syndrome variations which largely, although not exclusively, depend upon the mutation. Premature termination and deletions abrogate Wiskott-Aldrich syndrome protein (WASp) expression and lead to severe disease (WAS). Missense mutations usually result in reduced protein expression and the phenotypically milder X-linked thrombocytopenia (XLT) or attenuated WAS[1–3]. More recently however novel activating mutations have beendescribed that give rise to X-linked neutropenia (XLN), a third syndrome defined by neutropenia with variable myelodysplasia [4–6]. WASP is key in transducing signals from the cell surface to the actin cytoskeleton, and a lack of WASp results in cytoskeletal defects that compromise multiple aspects of normalcellular activity including proliferation, phagocytosis, immune synapse formation, adhesion and directed migration.
Keywords: Wiskott-Aldrich syndrome, actin polymerization, lymphocytes, dendritic cells, migration, cell activation, immune cell function
Abbreviations
AGM: Aorto-gonad-mesonephros ARPC: aActin related protein complex component CRIB: Cdc42/Rac interactive binding DC: Dendritic cells EVH1: Ena-VASP homology domain 1 FRET: Fluorescence Resonance Energy Transfer GBD: GTPase binding domain GEF: Guanine exchange factor IS: Immune synapse MZB: Marginal zone B cells N-WASp: Neuronal WASp PH: Pleckstrin homology PIP2: Phosphatidylinositol (4,5)-bisphosphate PLCγ: Phospholipase C gamma PSTPIP: Proline, serine threonine-rich phosphatase interacting
protein
∗Corresponding author: Adrian J. Thrasher, UCL Institute of Child Health, Molecular Immunology Unit, 30 Guilford Street, Lon- don, WC1N 1EH, UK. Tel.: +44 (0)20 7905 2289; Fax: +44 (0)20 7905 2810; E-mail: [email protected].
SNX9: Sorting nexin 9 TCR: T cell receptor Toca-1: Transducer of Cdc42-dependent actin assembly-1 Treg: Regulatory T cell VCA: Verprolin homology, central and acidic regions WAS: Wiskott-Aldrich syndrome WAS KO: WAS knockout WASH: WASp and SCAR homolog WASp: WAS protein WAVE: WASp family verprolin-homologous protein WHAMM: WASP homolog associated with actin, membranes,
and microtubules WIP: WASp interacting protein XLN: X-linked neutropenia XLT: X- linked thrombocytopenia
1. Introduction
Wiskott-Aldrich syndrome (WAS) is a rare X- linked recessive primary immunodeficiency charac- terised by immune dysregulation, microthrombocy- topaenia, eczema and lymphoid malignancies. Mu- tations in the WAS gene can lead to distinct syn- drome variations which largely, although not exclu- sively, depend upon the mutation. Premature termina-
ISSN 0278-0240/10/$27.50 2010 – IOS Press and the authors. All rights reserved
158 M.P. Blundell et al. / The Wiskott-Aldrich syndrome: The actin cytoskeleton and immune cell function
tion and deletions abrogate Wiskott-Aldrich syndrome protein (WASp) expression and lead to severe disease (WAS). Missense mutations usually result in reduced protein expression and the phenotypically milder X- linked thrombocytopenia (XLT) or attenuated WAS [1– 3]. More recently however novel activating mutations have been described that give rise to X-linked neutrope- nia (XLN), a third syndrome defined by neutropenia with variable myelodysplasia [4–6]. WASP is key in transducing signals from the cell surface to the actin cytoskeleton, and a lack of WASp results in cytoskele- tal defects that compromise multiple aspects of normal cellular activity including proliferation, phagocytosis, immune synapse formation, adhesion and directed mi- gration.
2. Molecular structure
WASp is a multidomain protein that lacks intrin- sic catalytic activity. It acts as an adaptor to bring together downstream mediators that facilitate Arp2/3- mediated actin polymerization. WASp consists of an N-terminal Ena-VASP homology domain 1 (EVH1), a basic domain, a GTPase binding domain (GBD), polyproline domain and the C-terminal domain com- prising of a cluster of verprolin homology (V), cen- tral (C) and acidic regions (A) (the VCA domain) (Fig. 1). WASp was the first member of a family of proteins that shares the ability to regulate Arp2/3 activ- ity through their VCA domains. The other mammalian members (Fig. 1) that share expression of the VCA domain include neural WASp (N-WASp), WASp fami- ly verprolin-homologous protein (WAVE) 1–3 and the more recently discovered Wiskott-Aldrich syndrome protein and SCAR homolog (WASH) and WASP ho- molog associated with actin, membranes, and micro- tubules (WHAMM) [7–9]. Many studies have investi- gated isolated protein domains to gain insight into the function of WASp family members and have extrap- olated findings of one member to predict function of another. In particular, studies of N-WASp have been used to predict function of WASp. The overall pro- tein sequence identity between WASp and N-WASp is around 50%, although sequence of the functional do- main is much higher. Importantly, however, expression of WASp is restricted to haematopoietic cells while N-WASp is ubiquitously expressed [10,11].
Fig. 1. Domain structure of WASp family protein members. Schematic representation of the protein domains of individual mem- bers of the WASp family. The C terminus, which is critical for regulation of actin polymerization, is highly conserved between the members, while the N terminus is divergent. EVH1, Ena Vasp ho- mology domain; B, basic domain; GBD, GTPase binding domain; PPP, polyproline domain; V, verprolin homology domain; C, cen- tral domain; A, acidic domain; SH, Scar homology domain; N, N-terminal domain with insignificant homology to known protein; CC, coiled-coils domain; WHD1/2, WASH homology domain 1 and 2.
3. EVH1 and WIP
The EVH1 domain is a proline-rich region that has been implicated in actin remodelling. This domain is found in actin scaffold proteins, such as Mena, VASP and Evl. The best characterised EVH1 domain binding partners are members of the verprolin family of pro- teins. These include WASp-interacting protein (WIP), CR16 and WICH (WIP and CR16 homologous protein, also referred to as WIRE) [12]. Of these, WIP is prob- ably the best characterized. It is a 503-amino acid long proline-richprotein, which is expressed in many tissues although most abundantly in haematopoietic cells [13]. While CR16 is known to be able to bind the WASp EVH1 domain, it is not expressed in haematopoietic cells; therefore, CR16/WASp interactions are not likely to be physiologically relevant [14]. WICH shows some ability to bind WASp, but this binding is much weaker compared to N-WASp binding. WICH is expressed at much lower levels in haematopoietic tissues compared to brain and gastrointestinal tissues, suggesting that N- WASp is the preferred binding partner to WICH [15].
Amino acids 416–488 of WIP were initially iden- tified as being required for WASp binding [16]. The crystal structure of WIP bound to the EVH1 domain of N-WASp shows that a 25 amino acid fragment (aa 461– 485) is minimally required for high affinity N-WASp binding [17]. Later the same group showed that for optimal binding, a larger sequence of WIP (aa 451– 485) wraps around the EVH1 domain of N-WASp; they also identified three WIP epitopes that are involved in WIP/N-WASp binding [18]. Many studies have used N-WASp to study the interaction of WASp with WIP. It is likely that these interactions are similar, but that
M.P. Blundell et al. / The Wiskott-Aldrich syndrome: The actin cytoskeleton and immune cell function 159
the WIP/WASp interaction is more relevantin vivogiv- en the similar tissue expression profiles. WIP inhibits N-WASp-mediated activation of Arp2/3 by direct in- teraction with N-WASp [19], and WIP maintains the autoinhibitory conformation of WASPin vivo [20]. A similar inhibitory mechanism exists for N-WASp, in which the WIP/N-WASp complex cannot initiate actin polymerization whereas free N-WASp can induce actin polymerization [21]. Toca-1 (transducer of Cdc42- dependent actin assembly) could release WIP binding, enabling Cdc42 to activate the protein and initiate actin polymerization [21].
Compelling evidence that WIP binding is important for WASp expression was provided by several groups who have shown that when WIP is absent, no or very low levels of WASp expression can be detected [22– 24]. Furthermore, reduced levels of WASp expression in lymphocytes occur in WAS patients who carry mu- tations that abrogate WIP binding [23]. Interestingly, the majority of missense WASp mutations that give rise to clinical disease are located in the EVH1 domain. Several of these mutations (e.g. T45M, V75M, R86H, Y107C and A134T) result in reduced binding affinity for WIP [17,25,26] and are associated with reduced WASp expression, but normal mRNA levels. These observations suggest that WASp levels are low due to proteolysis as a consequence of altered WIP binding [1, 3,27,28]. WASp expression in lymphocytes of these patients could be partially restored by proteasome in- hibition [23]. Together this evidence strongly sug- gests a role for WIP in protecting WASp from degra- dation. While N-WASp degradation appears mediat- ed by ubiquitination [29], WASp degradation seems a complex mechanism, which may be dependent on phosphorylation or activation state. WASp degrada- tion can be inhibited either by blocking the proteasome in unstimulated human and murine T cells [23] and murine splenocytes [30], or by blocking calpain pro- teases in stimulated human platelets [31,32], stimulat- ed human T cells [23] and GM-CSF/TGF-β-cultured splenic murine dendritic cells (DC) [22]. An attrac- tive hypothesis is that the proteasome is responsible for background turnover of WASp degradation and that cal- pain proteases are responsible for WASp degradation in response to activation and cell surface signalling, but further mechanistic investigation is required.
In addition, WIP binding mediates recruitment of WASp to podosomes [22,33], the phagocytic cup [34] and to the T cell receptor signalling complex after T cell stimulation [23,35,36]. Similarly, WIP binding to N- WASp is required for the actin-based intracellular mo-
bility of vaccinia virus [37]. These observations sug- gest that by binding to WASp, WIP prevents degrada- tion of WASp, keeps it in an inactive state and helps to achieve correct subcellular localization, thus providing an important spatial and temporal regulation of WASp activity.
4. Basic domain and PIP2
The basic domain is found in all members of the WASp family, although the specific sequence is poor- ly conserved. In WASp, the basic domain consists of 10 amino acids, of which six are lysine residues. The basic domain of N-WASp binds phosphatidylinositol (4,5)-bisphosphate (PIP2) in vitro [38,39]. WASp can also bind PIP2, but the specific binding site remains unclear. Higgs and Pollard reported that a region en- compassing the basic and GBD domains did not bind PIP2, in contrast to full length WASp [40]. Others have suggested that an N-terminal pleckstrin homology (PH) domain in WASp or N-WASp is responsible for PIP2 binding [11,41]. However, the sequence of such a PH domain shows weak alignment with the consensus PH domain sequence and overlaps with the EVH1 do- main, raising the question whether this domain should be called a PH domain [42]. Despite low sequence ho- mology, PIP2 binding and membrane localization were demonstrated for both WASp and N-WASp [11,41,43]. When the clinical C43W mutation was introduced in- to a truncated GST-WASp construct that contains the proposed PH domain, PIP2 binding was diminished. This observation suggests functionality of the PH do- main [41], although it should be noted that the C43W mutation likely affects binding of WIP as well [44]. Binding to WIP is important for N-WASp sub cellu- lar localization, and the corresponding C35W mutation in N-WASp abolished WIP-dependent recruitment of N-WASp to intracellular vaccinia virus [37].
Deletion of the basic domain of N-WASp diminished the ability of N-WASp to induce actin-driven motil- ity of coated beads and actin polymerization in cell lysates. By contrast, activity in a purified protein sys- tem containing Arp2/3 complex and monomeric actin was increased, suggesting that additional binding part- ners mediate the basic domain regulation of N-WASp activity [45]. One of these binding partners is likely to be Cdc42. Positively charged residues of the basic do- main contribute to Cdc42 binding, through an electro- static steering mechanism that mediates active Cdc42 to bind the GBD domain [46]. In addition, the ba-
160 M.P. Blundell et al. / The Wiskott-Aldrich syndrome: The actin cytoskeleton and immune cell function
sic domain of N-WASp can bind the Arp2/3 complex, probably by contributing to the closed, autoinhibito- ry conformation of cytosolic N-WASp. PIP2 binding could release this binding, enabling full activation by Cdc42 [38]. Whether this is also the case for WASp remains to be investigated.
5. GBD and activation by Cdc42 and phosphorylation
The GTPase binding domain (GBD), which spans from residue 235 to 288 in WASp, can bind Cdc42 and to a lesser degree Rac1 [12,47–49]. Contained within the GBD is the Cdc42/Rac interactive binding (CRIB) domain, a highly conserved sequence of 14 amino acid residues found in Rho type GTPase binding proteins [50,51]. The affinity of GTP-bound Cdc42 to WASp is at least 500-fold higher compared to the GDP-bound form [48] and its binding is important for WASp activation.
A second important binding partner of the GBD is the intrinsic WASp VCA domain. In its native, inactive state, residues 482–492 within the VCA domain bind to residues 242–310 at the carboxyl end of the GBD, as well as the adjacent sequence between GBD and the polyproline domain [52,53]. This intramolecular bind- ing, which has also been observed for N-WASp [39], inhibits the binding of Arp2/3 to the VCA domain, thereby keeping the protein in an inactive, autoinhibit- ed conformation that prevents actin polymerization.In vivo studies confirmed the autoinhibited conformation of WASp and N-WASp using Fluorescence Resonance Energy Transfer (FRET) techniques [20,54] and are fa- cilitated by a recently developed antibody that specifi- cally recognizes the open conformation [55,56]. Tyro- sine residue 291, within the GBD (aa 256 in N-WASp), is thought to be important in stabilising the open molec- ular conformation. Y291 can be phosphorylated by a variety of kinases from the Src and Tec kinase fami- lies. Although the ability to induce Arp2/3-mediated actin polymerization increases upon phosphorylation, more potent actin polymerization is observed when ei- ther GTP-Cdc42 or SH2 and SH3-domain-containing proteins are bound. Full potential activity is achieved when Cdc42 and SH2/SH3 binding act synergistically on phosphorylated WASp [57]. A model has been pro- posed in which GTP-Cdc42 disrupts the autoinhibited conformation of WASp, allowing Y291 to be phospho- rylated; this stabilises the opened molecular confirma- tion and enables maximum actin polymerization [52,
53,57–59]. Other studies, however, have shown that partial deletion of the GBD (but not Y291) and con- sequently loss of GTP-Cdc42 binding, does not affect WASp activity, suggesting phosphorylation as an alter- native Cdc42-independent mechanism of WASp acti- vation [60,61].
Phosphorylation of WASp is important for many cell processes including TCR signalling, filopodia and po- dosome formation, migration and phagocytosis, and appears to be a key regulatory mechanism of WASp activation. Phosphorylation of N-WASp is thought to regulate protein activation in a similar fashion. Sever- al studies have examined the importance of phospho- rylation by mutating the WASp Y291 to a ‘phospho- dead’ phenylalanine (Y291F) or to a phospho-mimetic glutamic acid (Y291E). While variability exists de- pending on the experimental system, expression of WASp Y291F reduces the ability of WASp to induce actin polymerization and formation of phagocytic cups, filopodia or podosomes [34,56,60,62]. FRET showed that the autoinhibited molecular conformation of WASp Y291F is disrupted by Cdc42 activation, suggesting phosphorylation is not required for WASp activation, although the actin polymerization ability was not in- vestigated [56]. Recently, knock-in mice were gen- erated to express the murine equivalent (Y293F) of the human Y291F WASp mutant. These mice show a strikingly similar cellular phenotype to WASp defi- cient mice including reduced actin polymerization, po- dosome formation, migration and phagocytosis [30]. In contrast, introducing the phospho-mimetic Y291E showed normal cellular and protein function in some experimental systems, but enhanced actin polymeriza- tion and formation of filopodia and podosome struc- tures in other systems [34,56,62]. Interestingly, mice expressing the phospho-mimetic Y293E phenotypical- ly resemble WASp deficient animals due to decreased protein expression levels, which could be partly re- stored by inhibition of the proteasome [30]. Phospho- rylation of Y256 by kinases such as Arg [63] is pre- dicted to stabilise the ‘opened’ molecular conformation of N-WASp. Experiments using Y256 N-WASp mu- tants indeed show enhanced actin polymerization with a phospho-mimetic Y256E and reduced actin polymer- ization with a ‘phospho-dead’ Y256F mutant [64].
6. Polyproline domain
The polyproline domain is the largest of the func- tional domains of WASp. It contains multiple binding
M.P. Blundell et al. / The Wiskott-Aldrich syndrome: The actin cytoskeleton and immune cell function 161
Table 1 WASp polyproline domain binding partners
Partner Known function References
Tyrosine kinases Btk Binds and phosphorylates WASp. [62,67,70–72] Hck Binds and phosphorylates WASp. [62,74] Fyn Binds and phosphorylates WASp. [60,68,69,73] Lyn Binds and phosphorylates WASp. [72] c-Src Binding to WASp demonstrated. [69,80,211] c-Fgr Binding to WASp demonstrated. [69,73,80] Tec Binding to WASp demonstrated. [70,212] Itk Binding to WASp demonstrated. [70,212]
Sh2-SH3 adaptor molecules Nck Binding and recruiting WASp to signalling complexes. [73,75] Grb2 Binding and recruiting WASp to signalling complexes. [32,69–71,75] p85α Direct and indirect binding to WASp, recruitment to signalling complexes. [69,80,81] CrkL Binding to WASp demonstrated. [79]
Actin binding molecules profilin Binds WASp and enhances Arp2/3-mediated actin polymerization. [86] cortactin Binds WASp and enhances Arp2/3-mediated actin polymerization. [213] VASP Binds WASp and enhances Arp2/3-mediated actin polymerization. [65]
Other PLCγ Binding to WASp demonstrated. [69–71,75,80] PSTPIP Binds WASp and allows dephosphorylation by recruiting phosphatase. [60,79,82–84,211] SNX9 Binding and recruiting WASp to signalling complexes. [81]
Btk, bruton’s tyrosine kinase; Grb2, growth factor receptor-bound protein; Hck, haematopoietic cell kinase; Itk, IL-2-inducible T cell kinase; Nck, non-catalytic region of tyrosine kinase; PLCγ, phospholipase C-gamma; PSTPIP, proline-serine-threonine phosphatase-interacting protein; SNX9, sorting nexin 9; Tec, tyrosine kinase expressed in hepatocellular carcinoma; VASP, vasodilator-stimulated phosphoprotein.
motifs for SH3 domain containing proteins (PXXP). Deletion of the polyproline domain abolishes the abil- ity of WASp and N-WASp to induce actin polymeriza- tion [45,65,66]. Many molecules can bind WASp via their SH3 domains including non-receptor tyrosine ki- nases, actin binding molecules and adaptor molecules (see Table 1). Members of the Src and Tec family of kinase families can bind to the polyproline domain of WASp through their SH3 domains (see Table 1). Of these, Btk, Hck, Fyn and Lyn can all bind and phospho- rylate WASp [60,62,67–74]. Many other kinases have the ability to bind WASp, but further investigations are required to elucidate their function andin vivo rele- vance (see Table 1). Whilst many kinases also contain SH2 domain binding motifs, no interaction with WASp through binding of SH2 domains with phosphorylat- ed tyrosine residues have been reported. Phosphoryla- tion of Y291 of WASp creates a potential SH2 binding motif, although this differs from the Src family SH2 consensus sequence (pYDFI instead of pYEEI).
A second important group of binding partners for WASp are adaptor molecules that contain SH3 do- mains (Table 1). Nck and Grb2 are SH2-SH3 adap- tor molecules, which were among the first identified to bind to WASp. They bind to the polyproline domain of WASp via their SH3 domains [73,75] and can re-
cruit other molecules through their SH2 domains. An example of this is the recruitment of WASp via Grb2 binding to the epidermal growth factor receptor after its stimulation [75]. Binding of Nck and Grb2 contributes to N-WASp activation and release of its autoinhibited conformation [76–78] and may function similarly for WASp. Two other SH2-SH3 adaptor molecules that can bind WASp are CrkL and the regulatory subunit p85α
of PI-3K. In platelets,CrkL binds WASp via its SH3 do- main and recruits the kinase Syk to the complex, which has the potential to tyrosine phosphorylate WASp [79]. Direct binding of p85α to WASp was shown in the lym- phoblastoid Namalwa cell line and in monocytic U937 cells [69,80]. In T lymphocytes, p85α binds CD28 and sorting nexin 9 (SNX9), with the latter binding WASp via its SH3 domain [81]. WASp function in platelets and the RAW macrophage cell line appears dependent on PI-3K activity, as inhibition of PI-3K by…