And Subcellular Targeting

Recent structural and biochemical studies of a wide variety of eukaryotic peripheral proteins participating in signal transduction and membrane-trafficking have shown that their molecular activity involves the reversible binding to specific lipid ligands in cell membranes (Hurley and Misra 2000). Importantly, the membrane-binding ability of these proteins is conferred by a reduced number of protein domains, the best known being PKC CI (Hurley et al. 1997), PKC C2 (Rizo and Sudhof 1998), FYVE domains (Misra and Hurley 1999), and pleckstrin homology domains (Rebecchi and Scarlata 1998). It must be remarked that the mechanism of membrane association of the different peripheral proteins involves a combination between both specific and non-specific interactions. Thus, in addition to specific binding pockets for signal molecules, basic and aromatic residues play important roles in providing a non-specific driving force for membrane association (Cho 2001).

Figure 3.8 shows 3D structures for different domains showing the main structural features of each family. The CI domains are known to occur not only in PKCs but also in more than 200 non-redundant proteins (Hurley and Misra 2000). This domain was first identified as the binding site for diacylglycerol (DAG) and phorbol ester in PKCs (Nishizuka 1988), although many of the known CI domains do not bind such molecules. Proteins bearing CI domains not competent for DAG or phorbol ester binding are called atypical. CI domains are compact motifs composed of -50 residues. They contain five P-strands, arranged into two sheets, a short a-helix and two 3-Cys-l-His-2-Zn2+ clusters (Hommel et al. 1994; Zhang et al. 1995). Briefly, the current model explaining the interactions of CI domains with membranes can be summarized as follows: DAG or phorbol ester bind to a specific binding pocket rich in aliphatic and aromatic residues, which is surrounded by basic residues which in turn are involved in electrostatic interactions with acidic phospholipids (Bittova et al. 2001). Whereas these last residues enhance the association rate with the membrane, the specific interaction with the second messenger decreases the dissociation rate, thus favouring the formation of a tight association of the domain with the membrane. Remarkably, there is a strong synergism between DAG or phorbol ester binding and membrane binding (Mosior and Newton 1996). It is evident that the presence of an exposed hydrophobic surface in CI domains demands a tightly controlled triggering mechanism of membrane targeting (Cho 2001). In fact, the scenario is more complex as

PH-domain FYVE-domain

Fig. 3.8. Three-dimensional structures of representative members of peripheral proteins involved in signal transduction and membrane-trafficking showing the main structural details of the membrane-interacting domains, (a) Cyan spheres in C1-domain correspond to Zn2+ ions (PDB code: 1KBE). (b) and (c) Green segments in the C2-domain correspond to those loops directly involved in Ca2+-binding (PDB code: 1DSY). Ca2+ ions are indicated as small blue spheres; phosphatidylserine molecule is indicated as PS. (d) The characteristic C-terminal a-helix above the p-structure core of the pleckstrin-homology domain (PDB code:1FB8) appears in green, (e) As above, cyan spheres in the FYVE-domain (PDB code: 1VFY) correspond to Zn2+ ions. Figures were prepared with the program PyMOL

PH-domain FYVE-domain

Fig. 3.8. Three-dimensional structures of representative members of peripheral proteins involved in signal transduction and membrane-trafficking showing the main structural details of the membrane-interacting domains, (a) Cyan spheres in C1-domain correspond to Zn2+ ions (PDB code: 1KBE). (b) and (c) Green segments in the C2-domain correspond to those loops directly involved in Ca2+-binding (PDB code: 1DSY). Ca2+ ions are indicated as small blue spheres; phosphatidylserine molecule is indicated as PS. (d) The characteristic C-terminal a-helix above the p-structure core of the pleckstrin-homology domain (PDB code:1FB8) appears in green, (e) As above, cyan spheres in the FYVE-domain (PDB code: 1VFY) correspond to Zn2+ ions. Figures were prepared with the program PyMOL

the majority of peripheral proteins have more than one membrane-targeting CI domain which may interact between them as well as with additional parts of the protein to modulate the interaction with the membrane.

C2 domains were originally discovered as the Ca2+ binding site of conventional PKCs (Xu et al. 1997). Proteins with C2 domains (>400 different proteins) are not only involved in signal transduction roles, but also in other cellular processes, which is consistent with the degree of variability of primary structures (Rizo and Sudhof 1998). Thus, many C2 domains bind membranes in a Ca2+-dependent manner, others do not bind Ca2+, and finally others do not bind membranes at all but other proteins both in a calcium-dependent or independent manner (Hurley and Misra 2000). The structure of the C2 domain is composed of an antiparallel P-sandwich made up of two four-stranded P-sheets, with the Ca2+-binding site located at one tip of the structure being formed by three connecting loops (Fig. 3.8c). It has been shown that Ca2+ ions may play two fundamental roles: they serve as a bridge between the C2 domain and anionic phospholipids (Verdaguer et al. 1999), and as inductor of intra- or interdomain conformational changes which may further affect the interaction with the membrane (Cho 2001). A model for membrane interaction of C2 domains has been proposed from the crystal structure of the phospholipase C5l (Grobler et al. 1996); in this case, the loops involved in Ca2+-binding are directed towards the membrane interface, where the acidic lipid headgroups are situated with the concave face of the domain facing the membrane plane.

Although most Ca2+-dependent C2 domains bind acidic phospholipids, others, like cPLA2-C2, seem to be phosphatidylcholine specific (Nalefski et al. 1998). Interestingly, whereas in the first group of domains there is a predominance of basic residues in the surroundings of the Ca2+-binding loops, in the second group hydrophobic residues can be found in equivalent positions. In addition, this lipid specificity correlates with the respective subcellular localization of the corresponding domains: those with preference for phosphatidylserine translocate to plasma membrane, and those with preference for phosphatidylcholine translocate to the PC-rich nuclear envelope and endoplasmic reticulum (Perisic et al. 1999).

FYVE domains constitute a recently described family of membrane-binding domains present in -60 proteins (Hurley and Misra 2000). They are involved in the endosomal translocation of proteins implicated in membrane-trafficking (Stenmark et al. 1996; Wurmser et al. 1999). FYVE domains specifically bind phosphatidylinositol-3-phosphate (PI3P), i.e. they are lipid-directed targeting domains. The crystal structure of the FYVE domain from Vps27p (Misra and Hurley 1999) reveals that it is formed by two antiparallel P-sheets and an a-helix stabilized by two Zn2+ ions, resembling CI domains (Fig. 3.8e). A basic region located in the surface of the domain, contributed by the sequence RKHHCR, has been shown by mutagenesis analyses to participate in PI3P-binding. In fact, this sequence motif, (R/K)(R/K)HHCR, is a defining feature of the FYVE domain. Analysis of the solvent accessible hydrophobic and basic residues suggests a nonspecific mechanism of membrane interaction (Misra and Hurley 1999), in which the tip of the N-terminal loop, containing two contiguous Leu residues, would penetrate into the PI3P-containing bilayer. Yet, experimental evidence points to the specific binding of PI3P as the major driving force for interaction with the membrane, which is consistent with the structural data.

Pleckstrin homology (PH) domains have been found in more than 500 proteins, most of them involved in phosphoinositides (Pis) binding, exhibiting a broad range of specificities (Rebecchi and Scarlata 1998). In addition, similarly to the functional variability observed with C2 domains, there also exist PH domains which do not bind Pis, but are involved in protein-protein interactions (Lemmon et al. 1997; Rebecchi and Scarlata 1998). The consensus structure of the PH domains as derived from the known three-dimensional structures can be described as follows: the domain consists of two nearly orthogonal and antiparallel P-sheets composed of three and four strands, respectively. The core of the P-structure is stabilized by a C-terminal a-helix which packs one of its edges (Fig. 3.8d). A distinct feature of PH domains is the clearly asymmetric distribution of charged residues in the protein surface: whereas basic residues predominate in one side of the domain (the membrane-binding face) and also partially in the loops, acidic residues are concentrated in the opposite side (Rebecchi and Scarla-ta 1998). The structures of complexes of PLC51-PH with Ins(l,4,5)P3 (Ferguson et al. 1995) and Btk-PH with Ins(l,3,4,5)P4 (Baraldi et al. 1999) define four different phosphate-binding subsites that participate in high-affinity specific PI-binding. Finally, membrane binding of PH domains is driven by non-specific electrostatic interactions with acidic phospholipids, together with the specific recognition of phosphoinositides. This agrees with the finding that some mutations increasing the number of basic residues in PH domains result in constitutive activation of the proteins (Baraldi et al. 1999).

Was this article helpful?

0 0
Healthy Chemistry For Optimal Health

Healthy Chemistry For Optimal Health

Thousands Have Used Chemicals To Improve Their Medical Condition. This Book Is one Of The Most Valuable Resources In The World When It Comes To Chemicals. Not All Chemicals Are Harmful For Your Body – Find Out Those That Helps To Maintain Your Health.

Get My Free Ebook


Post a comment