Perspective

Dancing with Multiple Partners

See allHide authors and affiliations

Science's STKE  19 Mar 2002:
Vol. 2002, Issue 124, pp. pe14
DOI: 10.1126/stke.2002.124.pe14

Abstract

Transmembrane proteins, such as G protein-coupled receptors (GPCRs) and integrins, activate intracellular signaling pathways through interactions with downstream binding partners. Woodside discusses two examples in which GPCRs and integrins interact in a noncompeting manner with more than one partner. The specific GPCR described is the thrombin receptor, in experiments where G protein peptides selectively block signaling through a particular G protein that does not appear to inhibit coupling of the receptor to other G proteins. The second system described is the αIIbβ3 integrin and its activation of the nonreceptor tyrosine kinase Syk. Syk appeared capable of interacting with both the integrin and intracellular domains of immune response receptors, because binding of Syk to the integrin was not inhibited by peptides based on the Syk binding site in immune response receptors. Thus, multiple, noncompeting binding partners add to the complexity of signal transduction outputs from a single receptor complex.

G protein-coupled receptors (GPCRs) and integrins are two classes of transmembrane receptors whose signals are largely defined by the intracellular molecules with which they interact. The cytoplasmic domains of these signaling receptors can have multiple, potentially noncompeting, binding partners. Gilchrist et al. (1) suggest that the specificity of GPCR coupling to heterotrimeric guanosine triphosphate-binding proteins (G proteins) is provided through the 11 COOH-terminal residues of the Gα subunits binding to nonoverlapping regions of the GPCR cytoplasmic domain. In a second example of noncompeting partners, the regions in integrin cytoplasmic domains responsible for regulation of the nonreceptor tyrosine kinases Syk and focal adhesion kinase (FAK) also appear to be distinct (2). Furthermore, the integrin-binding site in Syk involves its tandem Src homology type 2 (SH2) domains, yet appears to be independent of phosphotyrosine binding. The idea of protein domains having multiple, potentially noncompeting, binding partners adds another level of complexity to the signaling mechanisms linked to GPCRs and integrins.

GPCRs are heptahelical transmembrane receptors. Distinct GPCRs number in the thousands (3), and their diverse ligands include chemokines, neurotransmitters, hormones, light, and sometimes cryptic regions of the receptor that are themselves in the form of tethered ligands (4). Individual GPCRs can couple to multiple different intracellular G proteins composed of an α and a βγ subunit. There are four Gα subunit families (αs, αq, αi, and α12), each of which contains many members. Multiple different β and γ subunits are also described, culminating in what has been viewed by many as likely the most diverse receptor signaling system in eukaryotes (3).

Upon agonist binding, GPCRs undergo conformational changes that are propagated to their intracellular domains. These intracellular domains associate with G proteins, and Gα becomes activated by nucleotide exchange of guanosine diphosphate (GDP) for GTP. Following this sequence of interactions, active Gα and βγ subunits dissociate from each other and the GPCR to regulate downstream effectors. This apparently simple interaction between G proteins and GPCRs is precisely where the complexity of receptor signaling begins.

The intracellular loops of GPCRs contact the G proteins. In the model GPCR rhodopsin, the second and third intracellular loop, and the α helix 8, make contacts with the G protein α and βγ subunits (Fig. 1) (5-8). Although multiple points of contact occur, the COOH-terminal residues of Gα appear to play a pivotal role in receptor binding and selectivity (9-13). Gilchrist et al. (14) previously used the COOH-terminal 11 amino acid residues of Gαi expressed as a "minigene" to specifically inhibit muscarinic M2 GPCR signaling. This same strategy has since been used to demonstrate the specificity with which individual G proteins initiate different signal transduction pathways upon thrombin receptor activation and is highlighted in this perspective (1).

Fig. 1.

The putative interacting areas of the rhodopsin receptor and Gt. The regions of the two proteins are shown apart (A) and together (B). Highlighted are potential areas of charge complementarity in rhodopsin (K141, R147, K248, K341, in blue) and the α subunit of Gt (E212, D311, in red). [Adapted from (26)]

Thrombin signaling in endothelial cells can regulate intracellular 3′,5′-adenosine monophosphate (cAMP) (15) and Ca2+ concentration (16), actin stress fiber formation (17), and mitogen-activated protein kinase (MAPK) activation (18). These pleiotropic effects of thrombin are initiated by cleavage of the NH2-terminus of protease-activated receptors (PARs), which exposes a cryptic ligand that remains tethered to the receptor (4). PARs, like other GPCRs, can bind many different G proteins. Peptides based on the 11 COOH-terminal amino acid residues of Gαi, Gαq, Gα12, and Gα13 expressed as minigenes were used to determine G-protein specificity in thrombin receptor signaling in human dermal microvascular endothelial cells (HMEC-1) (Fig. 2A). Expression of Gαi prevented thrombin repression of isoproterenol-stimulated cAMP accumulation, whereas a randomized Gαi mini-gene had no effect. Isoproterenol stimulates cAMP accumulation through β-adrenergic receptor coupling to Gαs. This also demonstrates the fidelity of the β-adrenergic receptor toward Gαs, because cAMP accumulation appeared unchanged in the presence of the Gαi minigene. Expression of the Gαq minigene inhibited thrombin-generated increases in intracellular Ca2+ concentration and inositol phosphate accumulation. Gαi minigenes and sequence-scrambled Gαi had no effect, again demonstrating specificity. Thrombin induces stress fiber formation through activation of Gα12 (19), leading to activation of the small guanosine triphosphatase Rho (20). When Gα12 or Gα13 minigenes were expressed in endothelial cells, thrombin-induced stress fiber formation was abrogated. Unfortunately, Gαi and Gαq minigenes were not tested. Finally, all G protein minigenes (αs, αq, α12, and α13) inhibited thrombin-induced phosphorylation of extracellular signal-regulated protein kinase 1 (ERK1), which may be a reflection of Gβγ involvement in MAPK activation (21). For example, if Gβγ subunits are limiting, minigene-dependent increases in free Gα subunits could prevent Gβγ function (22, 23). Together, these results demonstrate that specific thrombin-mediated events in endothelial cells are dependent on unique G protein-GPCR coupling (Fig. 2A).

Fig. 2.

Multiple interactions between transmembrane receptors and their signaling components. (A) PAR1 signaling through multiple G proteins. PAR signaling in endothelial cells regulates intracellular cAMP and Ca2+ concentrations, cell morphology through rearrangement of the actin cytoskeleton, and MAPK activation (see text). The specificity of PAR1 coupling to each Gα subunit depends on the 11 COOH-terminal residues of Gα binding to nonoverlapping sites within the PAR1 cytoplasmic domain (1). (B) Integrin cytoplasmic domain interactions with signaling enzymes. Syk interaction with integrin β cytoplasmic domains appears to be independent of the sites required to regulate FAK activity. Likewise, integrin β cytoplasmic domains bind Syk's tandem SH2 domains through a mechanism that is independent of phosphotyrosine binding.

Assuming that thrombin is only signaling through PAR1 in HMEC-1 cells, what is the mechanism of action of these G protein minigenes? The simplest explanation is that they are binding to the PAR1 cytoplasmic domain and competing for endogenous G protein association. But how then could Gαi minigenes prevent some thrombin-dependent events and not others? Perhaps Gα minigenes can bind PAR1 through nonoverlapping, noncompeting binding sites. The minigenes expressed are only 11 amino acid residues in length. They presumably inhibit agonist signaling by preventing the binding of their parent G protein to the GPCR (1, 14, 24). However, they are likely not large enough to sterically inhibit other G proteins from interacting with the receptor, if in fact the PAR1 binding site(s) for the COOH-terminus of other Gα subunits are distinct and function primarily in receptor selectivity (25). One prediction of this hypothetical model would be that expression of one Gα minigene may alter the repertoire of other G proteins coupling to PAR1, potentially increasing the sensitivity of thrombin signaling in certain cellular responses. Such mechanisms could easily be tested and would have important therapeutic implications.

Alternatively, do the above results suggest that one GPCR can be simultaneously coupled to more than one G protein? Although not addressed, this possibility seems unlikely. In rhodopsin, for example, the cytoplasmic face is membrane fixed and small (26) in relation to the crystal structure of Gt (Fig. 1). The contact sites between Gt and rhodopsin appear broad and extensive, and it would be difficult to accommodate another G protein (26). However, if dimerization or higher order oligomers of PAR1 are the active signaling units (as has been suggested for other GPCRs [reviewed in (27)], then it is conceivable that more than one G protein could be associated with this active signaling complex.

The effects of the Gα minigenes might alternatively be due to the presence of another GPCR that could be regulated by thrombin. PAR2 is expressed on endothelial cells (28). It is activated by trypsin and by tryptase, but not by thrombin (29). However, it can act in trans as a surrogate signaling partner for PAR1 by heterodimerizing with PAR1 and "borrowing" the PAR1-tethered ligand (30). PAR1 coupling to Gαi and PAR2 coupling to Gαq (31) would explain why Gαi minigenes would have no effect on thrombin-induced increases in intracellular Ca2+ concentration, while still preventing thrombin-dependent decreases in cAMP levels. Although the signaling events in this scenario would still involve initial proteolysis of PAR1, some signaling events would be due to PAR2-G protein coupling. Dimerization (homo- and hetero-) of many different subfamilies of GPCRs occurs, and compelling computational studies involving evolutionary trace methods support models in which the functional unit of an active GPCR may be a dimer (27). The role played by PAR3 as a cofactor in thrombin cleavage and activation of PAR4 signaling (32) also suggests that PAR receptors can closely associate on the cell surface. It would be interesting to determine if function-blocking antibodies specific for the NH2-terminus of PAR2 could prevent thrombin-induced inositol phosphate production in the presence of Gαi minigenes. Ultimately, answers to many of the questions involving GPCR coupling to G proteins await high-resolution structural data of an active GPCR-G protein conjugate.

Another function of GPCR signaling is to regulate a distinct class of cell surface receptors, the integrins (31, 33, 34). Although structurally and functionally distinct, GPCRs and integrins share common mechanisms of action. Like GPCRs, integrins bind ligand and undergo putative conformational changes that are transmitted to the interior of the cell [reviewed in (33)]. The binding of intracellular molecules to the cytoplasmic domains of integrins can alter ligand-binding affinity (35), similar to the induction of a high-affinity ligand-binding state through GPCR-G protein interaction (36). And like GPCRs, integrins rely on their cytoplasmic domains for binding to intracellular molecules to transmit their signals.

Integrins are involved in the regulation of a number of cellular processes, including cell growth, differentiation, and apoptosis, as well as cell adhesion and migration [reviewed in (37)]. They are composed of αβ heterodimers of type I transmembrane proteins. A single-pass transmembrane domain connects the large ligand-binding extracellular regions to their cytoplasmic domains. Integrin cytoplasmic domains typically are 13 to 70 amino acid residues in length and have no known intrinsic catalytic activity, yet are required for integrin signaling. The associations of integrin cytoplasmic domains with intracellular proteins convey integrin-dependent signals to the cell. The list of intracellular molecules that have been shown to bind integrin cytoplasmic domains is long and includes cytoplasmic adaptor molecules, cytoskeletal components, and signaling enzymes [for a recent review, see (38)]. Like the binding of intracellular domains of GPCRs to Gα minigenes, integrin intracellular domains can interact with multiple binding partners in an apparently noncompetitive manner.

Engagement of platelet integrin αIIbβ3 activates two different nonreceptor tyrosine kinases, FAK (39) and the spleen tyrosine kinase (Syk) (40). Integrin-dependent activation of FAK may involve FAK binding to the membrane-proximal segment of integrin cytoplasmic domains (41) or, more likely (42), through an indirect association with the integrin-binding proteins talin (43) or paxillin (44, 45). Woodside et al. have recently shown that Syk (and its paralog Zap-70) can directly bind to integrin cytoplasmic domains and that this interaction is responsible for integrin αIIbβ3-dependent Syk activation (2). Removal of the last four COOH-terminal cytoplasmic amino acid residues of integrin β3 prevented Syk association with β3 both in vitro and in vivo and abrogated adhesion-dependent activation of Syk. However, αIIbβ3 integrins with this truncation still form integrin- and talin-containing focal adhesions, and integrin-dependent FAK phosphorylation is comparable to that of wild-type αIIbβ3. Furthermore, overexpression of the integrin-binding region in Syk prevents adhesion-dependent activation of Syk, but does not affect adhesion-dependent activation of FAK (2). Thus, it appears that Syk binding and FAK binding (either directly or indirectly) to the integrin β cytoplasmic domain occurs through unique, nonoverlapping regions of the integrin cytoplasmic domain (Fig. 2). It remains to be determined whether Syk binding and FAK binding proteins, such as talin or paxillin, can noncompetitively interact with integrin cytoplasmic domains. However, the bifurcation in the integrin-dependent regulation of these two kinase families suggests that integrin signaling could be targeted to inhibit some, but not all, integrin-dependent signaling events.

Further examination of the interaction between integrins and Syk demonstrated yet another example of a protein domain that has multiple, specific, noncompeting binding partners. Syk has a unique domain structure [reviewed in (46)] with two tandem SH2 domains in the NH2-terminus, followed by the interdomain B region, which contains multiple sites for tyrosine phosphorylation that can serve to regulate Syk function. The COOH-terminus of Syk is composed of a large kinase domain (46). Activation of this kinase occurs through binding of the tandem SH2 domains to dually phosphorylated tyrosines in immune receptor tyrosine-based activation motifs (ITAMs) (47, 48). These ITAMs contain the consensus binding motif YxxI/Lx(6-8)YxxI/L (49). The tandem SH2 domains of the Syk family interact with phosphorylated ITAMs (pITAMs) with high affinity (50) in a reverse colinear manner, such that the COOH-terminal SH2 domain of Syk binds the NH2-terminal phosphotyrosine in the ITAM sequence (51, 52). Mutation of Arg195 in the Syk COOH-terminal SH2 domain prevents its association with pITAMs (53, 54), but integrin-dependent activation of Syk still occurs (55). In direct binding assays, Syk with nonfunctional SH2 domains could still interact with the integrin cytoplasmic domain. Indeed, the Syk tandem SH2 domains bound integrin cytoplasmic domains in the presence of excess phosphorylated ITAM peptides (2). Together, these findings suggest that the SH2 domains of Syk interact with the integrin β cytoplasmic domains through a novel mechanism that is independent of pITAM binding and again provide an example of protein domains that have more than one specific, noncompeting binding partner.

This unique association between integrin cytoplasmic domains and the tandem SH2 domains of Syk raises a number of interesting questions. Because pITAM peptides based on the FcϵRIγ chain activate Syk (47), do the cytoplasmic domains of integrin β chains directly regulate Syk kinase activity? Although this could directly explain integrin activation of Syk, it would likely not be the complete story, because maximal adhesion-dependent Syk activation requires the presence of a Src kinase family member (55). Can Syk function as an adaptor between integrins and immune response receptors? For example, can integrin cytoplasmic domains and pITAMs from intact immune response receptors simultaneously bind Syk family members (Fig. 2B)? If so, do the integrin cytoplasmic domains potentiate pITAM activation of this kinase family? This could have important implications in integrin-dependent lymphocyte costimulation (56-58).

The two papers highlighted in this Perspective deal with vastly different receptor systems, namely, GPCRs and integrins. However, they both provide examples of protein domains having multiple specific noncompeting binding partners. Clearly, it will be important to define the contact points between integrins and the tandem SH2 domains of Syk, and those between activated GPCRs and G proteins. This will likely require high-resolution structural analysis of these two receptor systems, coupled with novel methodologies for tertiary structure analysis [reviewed in (26)]. Information gleaned from such studies will be extremely valuable in the design of the next generation of therapeutics that, instead of antagonizing receptor-ligand interactions, specifically antagonize receptor signaling.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
View Abstract

Stay Connected to Science Signaling

Navigate This Article