The Many Faces of SAM

See allHide authors and affiliations

Science's STKE  31 May 2005:
Vol. 2005, Issue 286, pp. re7
DOI: 10.1126/stke.2862005re7


Protein-protein interactions are essential for the assembly, regulation, and localization of functional protein complexes in the cell. SAM domains are among the most abundant protein-protein interaction motifs in organisms from yeast to humans. Although SAM domains adopt similar folds, they are remarkably versatile in their binding properties. Some identical SAM domains can interact with each other to form homodimers or polymers. In other cases, SAM domains can bind to other related SAM domains, to non–SAM domain–containing proteins, and even to RNA. Such versatility earns them functional roles in myriad biological processes, from signal transduction to transcriptional and translational regulation. In this review, we describe the structural basis of SAM domain interactions and highlight their roles in the scaffolding of protein complexes in normal and pathological processes.


SAM domains [also known as Pointed, SPM, SEP (yeast sterility, Ets-related, PcG proteins), NCR (N-terminal conserved region), and HLH (helix-loop-helix) domains] were initially identified about a decade ago by Ponting on the basis of the conservation of an ~70–amino acid domain in 14 eukaryotic proteins (1). The domain was christened the sterile alpha motif or SAM domain because four of the identified proteins (Byr2, Ste11, Ste4, and Ste50) are crucial in yeast sexual differentiation and secondary-structure prediction suggested a high content of α helix (1, 2). The SMART database now identifies more than 1300 SAM-containing proteins in all genomes. This number is comparable to that of the more widely known SH2 (Src homology 2) domain (which occurs in about 1600 proteins) (3, 4). The number of SAM domains in an organism roughly correlates with its complexity (Table 1). SAM domains are usually found in the context of larger multidomain proteins and may be found in all cellular compartments (Fig. 1), implying roles in complex and wide-ranging cellular processes.

Fig. 1.

Modular architecture of some SAM domain–containing proteins. SAM, sterile alpha motif; TM, transmembrane region; FN III, fibronectin repeats type III; PH, pleckstrin homology domain; ZBD, zinc binding domain; LZ, leucine zipper; ANK, ankyrin domain; SH3, Src homology 3 domain; PDZ, PDZ domain, PHAT, pseudo-HEAT repeat analogous topology; ETS, ETS DNA binding domain; MBT, malignant brain tumor repeats; Q, glutamine-rich domain; DG kinase, diacylglycerol kinase.

SAM domains have diverse functions. Thus, unlike some protein modules, such as SH2 and SH3 domains, for which the mere presence immediately suggests a likely function (5, 6), SAM domains are not so easily categorized. Indeed, even close homologs can have different functions (7). Thus, extensive experimental characterization of many SAM domains will be required before any sort of reliable bioinformatics approaches to classification is possible. This review summarizes our current understanding of SAM domain structure and biological function in several well-studied subfamilies and further discusses their possible roles in other important biological systems.

SAM Domains as Kinase Docking Sites

SAM domains occur in about 40% of ETS (initially named for erythroblastosis virus, E26—E twenty-six) family members [usually called the pointed (PNT) domain in this family], which are transcription factors defined by conserved DNA binding domains that recognize the core DNA sequence 5′-GGA(A/T)-3′ (8). ETS domain proteins function in a range of important biological processes, including cell growth control, differentiation, and embryonic development (9). One striking feature of this family of proteins is that they can be both transcriptional activators, as in the case of ETS-1 and ETS-2 (10, 11), and transcriptional repressors, as in the case of TEL and Yan (12, 13).

The structure of the SAM domain of ETS-1 has been solved (Fig. 2A) (14). ETS-1 is a transcriptional activator that is regulated by phosphorylation through the mitogen activated protein (MAP) kinase ERK2 (9). The nuclear magnetic resonance (NMR) solution structure of this monomeric SAM domain consists of a core formed by a bundle of four helices interfaced with an additional N-terminal helix (14). An adjacent MAP kinase phosphorylation site is located in a disordered region preceding the N-terminal helix. The SAM domain of ETS-1 provides a docking site for the ERK-2 MAP kinase. The site is composed of a cluster of hydrophobic residues, LXLXXXF (where L is leucine, F is phenylalanine, and X can be any amino acid) on the surface of the ETS-1 SAM domain (Fig. 2, B and C) (15). Mutations of docking-site residues inhibit phosphorylation in vitro and impair Ras-MAPK pathway–mediated enhancement of transactivation by ETS-1 in cell culture assays. A docking site of the same pattern also exists in ETS-2 and may be an interaction site with the Cdc2 family kinase Cdk10 (cyclin-dependent kinase 10). Phosphorylation by Cdk10 leads to the inhibition of ETS-2 transactivation in mammalian cells (16).

Fig. 2.

The structure of ETS-1 SAM domain and its role as an ERK-2 docking site. (A) Ribbon diagram of the NMR structure of ETS-1 SAM domain. The five α helices in the domain are labeled H1 through H5. The N terminus of the domain is disordered, and only a single conformation is shown. (B) Surface representation of the ETS-1 SAM domain. Three residues highlighted in green are kinase docking site residues (15). Phosphorylation site is in an N-terminal flexible region. The positions of MAP kinase phosphorylation site (T38) and kinase docking site (L114, L116, and F120) are indicated. (C) Schematic representation of the ETS-1 SAM domain. Sequences alignments of ETS-1, ETS-2, and Pnt-P2 in these two regions suggest that the MAP kinase phosphorylation site (Thr, colored in red) and the kinase docking site (in green) are conserved among them.

SAM domains may act in cis to facilitate phosphorylation (that is, on the same molecule that contains the kinase docking site), but they may also act in trans (that is, to bring a kinase into proximity of another substrate that has no kinase docking site). Phosphorylation of the Drosophila transcription repressor Yan is dependent on its interaction with another protein called Mae through their SAM domains (17). Mae may recruit the kinase, Rolled, to phosphorylate Yan. Another potential mechanism for Mae enhancement of Yan phosphorylation has also been suggested (18). The SAM domain of Mae acts to depolymerize polymers composed of Yan-SAM domains, which could thereby expose a phosphorylation site on Yan that is sterically inaccessible in the polymer.

SAM Domains as Polymers

The TEL transcriptional repressor

In addition to its presence in ETS family transcriptional activators, the SAM domain is also conserved in ETS family transcription repressors, such as TEL (also known as ETV6) (19) and its Drosophila ortholog Yan (13). TEL was originally identified at the breakpoint of the t(5;12) translocation in patients with chronic myelomonocytic leukemia (Fig. 3A) (20). This translocation fuses the N terminus of TEL to the tyrosine kinase domain of the platelet-derived growth factor receptor β. Some leukemia patients also harbor translocations that fuse the N terminus of TEL to the catalytic domains of other tyrosine kinases such as ABL (21) and JAK2 (22), or to transcription factors, such as AML1 (acute myeloid leukemia 1) (23) and ARNT (aryl hydrocarbon receptor nuclear translocator) (24). In each of these cases, the SAM domain of TEL is always present in the fusion protein, where it functions as a self-association motif necessary for either the activation of the kinase catalytic domain or transcriptional repression (Fig. 3A) (25, 26).

Fig. 3.

TEL and the polymeric structure of its SAM domain. (A) Chromosomal translocations of the SAM domain–containing region of the TEL gene generate many different TEL–Tyr kinase fusion genes. Self-association of the TEL-SAM domain can lead to constitutive activation of the fused kinases, causing cell transformation. (B) (Top) Polymeric structure of TEL-SAM. Nine subunits of the polymer viewed with the helix axis in the plane of the paper, pointing up. The ML and EH surfaces are shown in green and red, respectively. Two key residues in the interaction interface, Ala93 and Val112, are shown in stick representation. (Bottom) Schematic illustration of the TEL-SAM polymer with two key residues, Ala93 and Val112, in the ML and EH surface, respectively, highlighted. (C) Structure of TEL-SAM, showing the ML (left) and EH (right) binding surface. The hydrophobic residues that form the ML binding surface of TEL-SAM are colored green. The hydrophobic residues that make up the core of the EH surface are in red.

Isolated TEL-SAM domains form a head-to-tail polymer with six SAM domain monomers per turn, and a repeat distance of 53 Å (Fig. 3B) (27). The polymer interface is made from two different surfaces on the protein: the mid-loop (ML) surface, consisting of residues in the middle of the protein; and the end-helix (EH) surface, located around the C-terminal helix (Fig. 3, B and C) (28). Both the N and C termini point outward from the polymeric helix, so this polymer structure could be accommodated at any point in a protein sequence. The intersubunit interactions in the TEL-SAM polymer are very stable with a dissociation constant Kd ~2 nM, implying that the interaction may be of biological relevance (27). Moreover, polymer-blocking mutations (at interface residues Ala93 and Val112) (Fig. 3B) block TEL repression in cell culture, strongly suggesting that the polymeric structure is essential for the transcriptional-repression function of TEL (29). SAM polymerization may provide a mechanism for spreading of a domain of repression along the chromatin, although there is still no direct demonstration of this (27).

Tyrosine kinases can be activated by oligomerization, so it is not hard to see how the attachment of a polymerization module to a kinase domain, as found in the oncogenic TEL-SAM fusions, could lead to constitutive kinase activation (30). The story is not so simple, however. When the role of polymerization in the transformation activity of the TEL-NTRK3 chimera (also called ETV6-NTRK3 or EN) (31) is examined, both SAM domain binding interfaces are found to be essential for full EN transformation activity. EN mutants containing mutations that change hydrophobic amino acid residues to hydrophilic ones within the polymerization interface (A93D and V112E or V112R) fail to self-associate, do not become tyrosine phosphorylated, and do not transform NIH 3T3 cells. Moreover, gel-filtration chromatography and electron microscopy experiments suggest that, although the EN chimera polymerizes to form higher order complexes both in vitro and in vivo, EN variants with polymer-blocking mutations within their SAM domains are largely monomeric in solution. Dimerization alone is not sufficient to transform NIH 3T3 cells. When the SAM domain in the chimera was replaced by an FKBP-binding domain, a chemically inducible dimerization motif, the chimeric protein became tyrosine phosphorylated upon dimerization, but it did not induce NIH 3T3 cell transformation. Only transient MEK1 activation, cyclin D1 up-regulation, and AKT activation were observed in these cells after the addition of the dimerization inducer, which suggests that unlike EN, chemically induced FKBP-NTRK3 dimers lack the ability to constitutively activate the Ras-MAPK and phosphatidylinositol 3-kinase (PI3K)–AKT cascades. Although the stable coexpression of the V112E and A93D EN-SAM mutants in NIH 3T3 cells led to heterodimerization, tyrosine kinase activation was not observed and the cells were not transformed. These findings imply that SAM domain polymerization in these oncogenic fusions must do more than simply create the close proximity of tyrosine kinase domains.

The key role of polymerization in cell transformation by TEL-SAM oncogenic fusions suggests that one possible therapeutic strategy would be to find reagents that could block TEL-SAM polymerization. Isolated SAM domains should be able to bind to the EN chimeras and could thereby block cell transformation (31, 32). Indeed, when either wild-type or poylmer-blocking mutant SAM domains were expressed in NIH 3T3 cells, EN-mediated morphological transformation and soft agar colony formation were reduced. Cells expressing both the isolated SAM and EN chimeras showed a reduced ability to form tumors in nude mice compared with cells expressing the EN chimera only. This indicates that the SAM domain may act as a dominant negative regulator of polymerization-mediated kinase activation and cell transformation by the EN chimera. Thus, controlling cell transformation by targeting the polymerization interface might serve as a potential therapeutic strategy. However, polymer disruptors could also affect the transcriptional repression activity of wild-type TEL, which acts as a tumor suppressor.

The Yan, Mae, and Pointed-P2 proteins

The ETS family SAM domain–containing proteins Yan, Mae, and Pointed-P2 (Pnt-P2) are also involved in transcriptional regulation during the development of Drosophila eyes (Fig. 4A) (33). Yan is a transcriptional repressor (13, 34) and Pointed-P2 is a transcriptional activator (35, 36), and both are subject to regulation by the MAP kinase Rolled. Upon stimulation of the Sevenless or epidermal growth factor (EGF) receptors and activation of the Ras-MAP kinase pathway, Yan becomes inactive and Pnt-P2 becomes active, stimulating the transcription of previously repressed genes.

Fig. 4.

Mae regulation of Yan and Pnt-P2 through SAM domain interactions. (A) Upon the activation of the receptor tyrosine kinase (RTK) pathway, through the guanosine triphosphatase RAS and the mitogen-activated protein kinase (MAPK) cascade, MAP kinase Rolled becomes dual-phosphorylated and translocates to the nucleus. In the nucleus, activated Rolled phosphorylates two ETS family transcription factors: Yan, a transcription repressor; and Pnt-P2, a transcription activator. Phosphorylation of Yan results in abrogation of its repressor activity and its translocation to the cytoplasm. In contrast, phosphorylated Pnt-P2 becomes an active transcription activator, which then stimulates the expression of differentiation genes. Moreover, the phosphorylation of Yan and Pnt-P2 is regulated by Mae through interactions mediated by the SAM domains of Mae, Yan, and Pnt-P2 (B) Overall structure of Yan- and Mae-SAM domains compared with the Yan-SAM polymer model. Yan-SAM and Mae-SAM are shown in blue and violet, respectively. ML surface residues are colored green, and EH surface residues are colored red. (C) The formation of polymeric Yan on DNA by its SAM domain interaction makes its key phosphorylation site, Ser127, inaccessible to the MAP kinase Rolled. Depolymerization of Yan by Mae exposes this phosphorylation site to Rolled.

Yan was identified through a genetic screen as a negative regulator of R7 photoreceptor development acting antagonistically to the Sevenless–Ras–MAP kinase pathway (13). It functions as a repressor of ETS-responsive genes. Phosphorylation of Yan by the MAP kinase Rolled leads to inactivation and export from the nucleus (34). The most important phosphorylation site, Ser127, is very close to the C terminus of the Yan SAM domain. Although Yan becomes inactive upon phosphorylation, the transcriptional activator Pnt-P2 is activated when phosphorylated by Rolled. Pnt-P2 contains an N-terminal SAM domain and C-terminal ETS DNA binding domain (35, 37).

Another component of this signal transduction pathway is called Mae (modulator of the activity of ETS) (17). Mae is essential for the normal development and viability of Drosophila and is required in vivo for normal signaling of the EGF receptor. Mae is also required for Rolled phosphorylation of Yan. Mae contains a SAM domain that binds to the SAM domains of both Yan and Pnt-P2. Rolled also binds to Mae, suggesting that Mae recruits the kinase to the site of action, possibly in concert with some sort of conformational change enforced on Yan by Mae. Mae also inhibits DNA binding by Yan, although the mechanism is unclear (17).

Additional insight into this complex regulatory system came from structure-based studies of Yan-SAM. Like its human ortholog, the TEL-SAM domain, the isolated SAM domain of Yan forms a head-to-tail polymer (Fig. 4B), although the interaction between each subunit (Kd ~ 11 μM) is much weaker than that of the TEL-SAM domain. Yan variants bearing mutations that block polymerization do not repress transcription as efficiently as does wild-type Yan, indicating that the ability to polymerize is essential for Yan activity (18).

Mae causes depolymerization of Yan, revealing a new mechanism for controlling transcriptional repression by Yan. The crystal structure of interacting Mae and Yan-SAM domains explains in detail how Mae inhibits polymerization of Yan. The Mae-SAM domain interacts through its ML surface with the EH surface of Yan-SAM, making the EH surface unavailable for polymer formation. The binding mode of Mae almost perfectly mimics the binding mode in polymers of Yan-SAM domains. However, the affinity of the Mae-SAM domain for binding to the Yan-SAM domain is ~1000 times as high as that of Yan-SAM for binding to itself, so Mae is able to compete effectively with Yan polymer formation.

That Mae inhibits polymerization of Yan also suggests an alternative mechanism by which Mae can facilitate Yan phosphorylation. The key phosphorylation site on Yan, Ser127, is very close to the SAM domain and would therefore most likely be inaccessible to the kinase in the polymer. Depolymerization may therefore enhance access of the kinase to the phosphorylation site (Fig. 4C). Thus, by favoring depolymerization of Yan, Mae could have a multifaceted role in controlling Yan function. Moreover, Yan represses the transcription of Mae (38), so that inhibition of Yan would lead to increased production of Mae. Thus, the system is apparently designed for synergistic down-regulation of Yan and up-regulation of Mae. Thus, depolymerization of Yan by Mae represents a mechanism of transcriptional control that sensitizes Yan for regulation by receptor activation (18).

Polycomb group

Polycomb group (PcG) proteins are general transcriptional repressors essential for maintaining the transcription pattern of key regulatory genes throughout development in many organisms (39). They are best known as repressors that restrict Hox gene expression along the anterior-posterior animal body axis. They are generally classified not by their conserved domains or structural motifs, but by their common function. PcG proteins, working together in large complexes, organize chromatin of the target gene to maintain repressed states for long periods of time and through many cell divisions (40). Two major PcG complexes have been characterized in Drosophila, called the ESC-E(Z) complex (4143) and Polycomb repressive complex 1 (PRC1) (4446). The core subunits of both complexes are also conserved in mammals (45). Both purified fly embryo PRC1, and recombinant, reconstituted PRC1 complex can block nucleosome remodeling by SWI/SNF chromatin remodeling complex in vitro, possibly by generating a higher order chromatin structure refractory to gene transcription (44, 46).

SAM domains are found in several Drosophila Polycomb group proteins and their mammalian relatives. Information about the structure and function of PcG SAM domains comes largely from studies on Polyhomeotic (Ph) and Sex comb on midleg (Scm), which both contain C-terminal SAM domains (47) (Fig. 1).

Both Ph and Scm are found in PRC1, although Scm is found in substoichiometric amounts and is not considered to be one of the core components of PRC1 (44). Nevertheless, yeast two-hybrid and in vitro protein-protein interaction assays show that SAM domains of Ph and Scm interact both with themselves and with each other. Moreover, Ph and Scm both localize together at Drosophila polytene chromosome sites, implying that they might function together to repress transcription (47).

The Ph-SAM domain forms a polymer, and the polymer crystal structure has been solved. The polymer architecture is very similar to that of TEL-SAM, that is, a left-handed, head-to-tail helical polymer (48). The polymer subunits interact through similar ML and EH surfaces. The intersubunit interactions are strong (like those of TEL-SAM) with a Kd of 190 nM. The Scm-SAM domain forms a polymer that has a similar architecture to that of polymers formed for TEL-SAM or Ph-SAM (49). That TEL, Ph, and Scm show minimal sequence similarity and share no other common domains, yet generate very similar polymer structures, lends further weight to the argument that these polymers are biologically relevant and play important roles in transcriptional repression.

In vivo studies of the role of Scm-SAM in PcG repression further support a biological function for the Scm-SAM polymer and the interaction of the Scm- and Ph-SAM domains (50). In genetic rescue assays, a collection of Scm variants containing random SAM domain mutations were introduced to an Scm-null strain. Only those mutants that retained the ability to self-associate [as verified by two-hybrid and GST (glutathione S-transferase) pull-down assays] sustained formation of fertile and phenotypically normal flies (50). The mutants that did not self-associate mapped to the ML and EH surfaces of the Scm-SAM domain and likely block polymerization. This was further substantiated by PcG loss-of-function phenotypes caused by overexpression of an isolated Scm-SAM domain in vivo. Overexpression of the Scm-SAM domain presumably interferes with functional polymer formation by the full-length protein. PcG repression can occur over long distances. Although the mechanism of long-range repression is not known, it seems possible that polymerization of PcG along the chromosome could have a role (48).

The corresponding human homologs of several fly PcG proteins have been identified. Humans have multiple versions of each PcG gene. For example, three human homologs of Drosophila Ph have been identified so far: hPh1, hPh2 (51), and hPh3 (45). They function in repression complexes that inhibit SWI/SNF remodeling of nucleosomal arrays and may be important for transcriptional repression in vivo. The SAM domains of hPh1 and hPh2 mediate self-association, perhaps by polymerization (51). Despite the domain divergence of these human Ph homologs, they all contain SAM domains, indicating the importance of SAM domains to Ph function.

The first Polycomb group protein discovered in Caenorhabditis elegans, SOP-2, was recently isolated in a genetic screen for suppressors of the homeobox gene pal-1 (52). Sequence comparisons to other PcG proteins revealed that the only conserved domain in SOP-2 is the SAM domain, which is closely related to the SAM domains of Drosophila PcG proteins. Despite the lack of additional sequence similarity, the presence of a SAM domain in SOP-2 suggests that this C. elegans factor may also self-associate, allowing it to potentially play the same role as its counterparts in the fly.

A linkage between the SAM domain of the PcG group and that of ETS proteins was established by the finding that the human PcG protein homolog, L(3)MBT [lethal(3)malignant brain tumor protein], and TEL interact through their SAM domains (53). Moreover, expression of the h- L(3)MBT SAM domain enhances the efficiency of transcriptional repression mediated by TEL, indicating that the interaction is functional. These results suggest that general transcriptional repression machinery can be recruited, by SAM domain association, to individual target genes through the DNA binding specificity endowed by the ETS family protein TEL.

SAM Domain Binding to RNA

The traditional belief that SAM domains are protein-protein interaction motifs was recently turned on its head by work on Smaug and its relatives (54, 55). Smaug is required for the establishment of a morphogen gradient in the developing embryo through repression of Nanos mRNA translation in the bulk cytoplasm. Smaug binds to an RNA stem-loop consisting of a 9–base pair helix capped by a five-ribonucleotide loop within the translation control element of the nanos mRNA 3′-untranslated region. This interaction appears to be necessary for the proper functioning of Smaug because mutations in the loop abolish the repression of Nanos translation (5658). The RNA binding region of Smaug corresponds to its SAM domain (54, 55). The crystal structure of a Smaug RNA binding region, containing both a SAM and a PHAT (pseudo-HEAT repeat analogous topology) domain, shows that Smaug-SAM conforms to the consensus SAM domain architecture. One distinct feature of Smaug-SAM is the appearance of a cluster of positively charged residues on its surface (Fig. 5). A database search revealed a family of Smaug homologs from yeast to human, all of which have essentially the same RNA binding specificity as fly Smaug, even though their sequence similarity is restricted to the SAM domain. Thus, both structural and sequence conservation analysis suggest that the SAM domain of Smaug is the RNA binding motif. A yeast three-hybrid assay was used to select Smaug mutants that can still bind to RNA. This screen isolated variants with surface substitutions almost everywhere except the electropositive face of the SAM domain. Engineered substitutions in the conserved basic residues of the electropositive face reduced RNA binding affinity (54). The isolated SAM domain of Vts1, a yeast Smaug homolog, binds to non–stem-loop RNA with high affinity, comparable to that of the full-length Vts1 protein. SAM-mediated RNA binding, identified through a combination of structural, biochemical, and genetic studies on Smaug and its relatives, represents a new mode of SAM domain action. Perhaps SAM domains may continue to have other new faces exposed (7, 59).

Fig. 5.

Crystal structure of Smaug RNA binding region and the potential RNA binding surface on Smaug-SAM. The SAM domain is colored pink, and the PHAT domain is colored white. Residues that are essential for RNA binding are highlighted in blue.

Table 1.

The number of SAM domains in an organism roughly correlates with its complexity.

Nonpolymeric SAM Associations

Not all SAM self-associations result in polymer formation. Several SAM domain proteins involved in yeast MAP kinase cascades also seem to form closed oligomers.

In the fission yeast Sschizosaccaromyces pombe, pheromone-induced sexual differentiation is controlled via a MAPK pathway that includes the scaffold proteins Ste4 and Byr2 (60, 61). Byr2 is a MAPK kinase kinase that is activated by interactions with both Ras1 and Ste4. The SAM domain of Byr2 has been shown to bind to the N-terminal 160 amino acids of Ste4, a region containing a SAM domain followed immediately by a leucine zipper (Ste4-LZ) domain. This interaction is essential for proper signaling (62). The isolated SAM domains of Byr2 and Ste4 are both monomeric and bind to each other in a 1:1 stoichiometry with Kd ~ 56 μM (63). However, when the leucine zipper domain is included with Ste4-SAM, it forms a 3:1 Ste4-LZ-SAM:Byr2-SAM complex with much higher affinity (Kd ~ 19 nM) (63). Systematic mutagenesis of the surface residues on both Byr2- and Ste4-SAM mapped the binding interface between them to the ML surface of Byr2-SAM and the EH surface of Ste4-SAM, as did similar analysis of the interaction between Mae-SAM and Yan-SAM (64). This suggests that SAM domains can use similar surfaces for the formation of different types of oligomeric states. How this SAM interaction regulates activity of the Byr2 protein kinase remains unclear.

Structural and biochemical studies have also been done with Ste11 and Ste50 in Saccharomyces cerevisiae, which are orthologs of Byr2 and Ste4, respectively (65, 66). NMR solution structures of the Ste11 and Ste50 SAM domains show structures similar to those of other SAM domains (67, 68). Ste11-SAM forms a weak dimer in solution with Kd ~ 0.5 mM (68). Again, the interface was mapped to EH and ML surfaces by comparing the chemical-shift changes between concentrated and diluted solutions of Ste11-SAM. A combination of chemical cross-linking and surface plasmon resonance experiments demonstrated that SAM domains of Ste11 and Ste50 form a high-affinity heterodimer (Kd ~ 71 nM) (67). Preliminary mutagenesis studies showed that one of the Ste11-SAM ML surface mutants fails to interact with Ste50-SAM, implying that Ste11-SAM might use its ML surface to bind to the EH surface of Ste50-SAM (67, 68).

Polymers or Not Polymers?

Members of the Eph receptor tyrosine kinase family of proteins together with their activating ligands, termed ephrins, are involved in contact-mediated axon guidance, axon fasciculation, vascular network assembly, capillary morphogenesis, and angiogenesis. Interaction of Eph receptor proteins with ephrins can activate bidirectional signaling pathways, frequently resulting in repulsive cell-cell signaling (69).

SAM domains are found at the C terminus of all the known Eph proteins (3). Several crystallographic and NMR studies have yielded canonical SAM domain structures (7073). Crystal structures of the EphA4-SAM domain and the EphB2-SAM domain revealed potential oligomeric forms. In the case of the EphA4-SAM, a large interface was identified that could represent a dimer structure (73). In the case of EphB2-SAM, two large interfaces may lead to a polymeric structure (72). Although in both cases there are reasons to think that these interfaces are more than crystal contacts, self-association in solution is invariably quite feeble with these SAM domains (Kd > 1 mM). However, even weak interactions can lead to productive associations when proteins are tethered to a membrane, or when they are clustered by ligand interactions. It is therefore possible that the SAM domains play an ancillary role to the clustering process necessary for signaling. Nevertheless, it is very difficult to confirm or deny the importance of such weak interactions, so these oligomers remain hypothetical. Surprisingly, removal of EphA4-SAM domain does not disrupt Eph signaling, measured either as kinase activity of EphA4 kinase or clustering and ligand activation (74). A Phe-to-Tyr mutation at position 928 within the SAM domain, however, caused an increase in ectopic induction of posterior protrusions when compared with wild-type EphA4, suggesting that Tyr928 may negatively regulate EphA4 activity (75).

SAM domains are also present in Shank family proteins, which are highly enriched in the postsynaptic density (PSD), acting as scaffolds to organize assembly of postsynaptic proteins (76). The SAM domain of Shank1 self-associates and may have a role organizing the PSD (77).

SAM Domain Binding to Other Proteins

In addition to interactions with the kinases discussed previously, SAM domains also interact with other proteins that lack SAM domains. For example, the SAM domain of BAR (bifunctional apoptosis regulator), which can regulate both extrinsic and intrinsic apoptosis pathways, binds the cell death suppressors Bcl-2 and Bcl-XL (78). Moreover, BAR suppresses BAX-induced apoptosis though a SAM domain–dependent mechanism in both yeast and mammalian cells. However, the molecular details of these activities remain unclear.

The SAM domain of ETS-2 is both necessary and sufficient for the binding of ETS-2 to the C-terminal domain of CREB-binding protein [CBP-SID (steroid receptor coactivator interaction domain)]. Moreover, this binding region on ETS-2–SAM was narrowed down to its fifth helix. Cell culture assays showed that this interaction is important for ETS-2–mediated transactivation (79).

Posttranslational Modification of SAM Domains

As discussed above, RTK-Ras-MAPK pathway–mediated phosphorylation of Ser127 on Yan, a residue just 10 amino acids away from the C terminus of the Yan-SAM domain, abrogates Yan’s transcriptional repressor activity and facilitates Crm1-adapted translocation out of the nucleus (34, 80). Conversely, phosphorylation of a residue very close to the Pnt-P2 SAM domain, Thr151, by the same pathway increases transcriptional activation by PNT-P2 (35, 37). Thus, this common pathway of regulation can have distinct downstream consequences.

Two observations link SUMO (small unbiquitin-like modifier) directly to SAM domains. The ETS family transcription repressor TEL interacts with UBC9 (81), a member of the SUMO conjugating enzymes, through its SAM domain. This interaction leads to the SUMOylation of Lys99 in TEL-SAM (82), thus decreasing the repression activity of TEL. Mutation of Lys99 caused decreased nuclear export. Unexpectedly, SAM domain polymerization seems to be important for the SUMOylation reaction (29). The reason is still a mystery. Interestingly, SUMOylation has completely opposite effects on C. elegans PcG protein SOP-2, in which case, SUMOylation of SOP-2 through the interaction between UBC9 and the SOP-2 SAM domain is required for its localization to nuclear bodies in vivo and for its physiological repression of Hox genes (83).


In the decade since Ponting’s identification of SAM domains from 14 proteins, many more proteins have been identified with SAM domains. Ponting’s prediction that SAM domains might be involved in protein-protein interactions proved prescient. SAM domains can interact with themselves, bind to other SAM domains, bind to non–SAM domain proteins such as kinases or SUMO conjugating enzyme, and can even bind to RNA. We now have a detailed understanding of SAM domain function in a variety of systems. High-resolution structures of homo- and heterotypic SAM domain interactions have greatly facilitated our attempts to elucidate the molecular mechanism of several signal transduction pathways and gene regulation circuits that include SAM domain–containing proteins. Nevertheless, we have probably only scratched the surface of SAM domain functions and interaction modes, and there are likely roles for SAM domains that remain to be discovered. For example, p73, a homolog of the tumor suppressor p53, was recently found to interact with lipids by in vitro biochemical assays, although the biological significance of this finding still remains unclear (84). At present, it is hard to see how these alternative SAM domain functions can be deciphered without doing the hard work of further biological, biochemical, and biophysical characterization of individual systems.


  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.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
View Abstract

Stay Connected to Science Signaling

Navigate This Article