PMC:5026484 / 37582-47143 JSONTXT

{"project":"2_test","denotations":[{"id":"27552051-18940603-26907304","span":{"begin":2099,"end":2103},"obj":"18940603"},{"id":"27552051-18940604-26907305","span":{"begin":2156,"end":2160},"obj":"18940604"},{"id":"27552051-10025402-26907306","span":{"begin":2209,"end":2213},"obj":"10025402"},{"id":"27552051-16554019-26907307","span":{"begin":2729,"end":2733},"obj":"16554019"},{"id":"27552051-9524117-26907308","span":{"begin":3458,"end":3462},"obj":"9524117"},{"id":"27552051-17725553-26907309","span":{"begin":3523,"end":3527},"obj":"17725553"},{"id":"27552051-18946081-26907310","span":{"begin":3681,"end":3685},"obj":"18946081"},{"id":"27552051-23639309-26907311","span":{"begin":4484,"end":4488},"obj":"23639309"},{"id":"27552051-23639309-26907312","span":{"begin":4761,"end":4765},"obj":"23639309"},{"id":"27552051-17725553-26907313","span":{"begin":5429,"end":5433},"obj":"17725553"},{"id":"27552051-21951725-26907314","span":{"begin":5830,"end":5834},"obj":"21951725"},{"id":"27552051-23060965-26907315","span":{"begin":5855,"end":5859},"obj":"23060965"},{"id":"27552051-21041651-26907316","span":{"begin":6023,"end":6027},"obj":"21041651"},{"id":"27552051-26538022-26907317","span":{"begin":6029,"end":6033},"obj":"26538022"},{"id":"27552051-25086062-26907318","span":{"begin":6213,"end":6217},"obj":"25086062"},{"id":"27552051-26538022-26907319","span":{"begin":7624,"end":7628},"obj":"26538022"},{"id":"27552051-20008558-26907320","span":{"begin":7645,"end":7649},"obj":"20008558"},{"id":"27552051-18305261-26907321","span":{"begin":7667,"end":7671},"obj":"18305261"},{"id":"27552051-12938173-26907322","span":{"begin":8506,"end":8510},"obj":"12938173"},{"id":"27552051-23931141-26907323","span":{"begin":8668,"end":8672},"obj":"23931141"},{"id":"27552051-16034420-26907324","span":{"begin":8691,"end":8695},"obj":"16034420"},{"id":"27552051-17356578-26907325","span":{"begin":8877,"end":8881},"obj":"17356578"},{"id":"27552051-9297972-26907326","span":{"begin":8899,"end":8903},"obj":"9297972"},{"id":"27552051-21951725-26907327","span":{"begin":9518,"end":9522},"obj":"21951725"}],"text":"Discussion\nIn this publication, we present a thorough biochemical and structural analysis of a Rab effector domain termed bivalent Mical/EHBP Rab binding (bMERB) domain. The results show that the domains probably constitute a Rab8-effector family involved in endosomal trafficking, and the Rab-binding specificity can be well explained from the 3-dimensional structures of complexes determined in this work. Furthermore, we show that at least some of these domains contain two separate binding sites for Rab-proteins, suggesting previously unknown functions, as discussed below. The strong similarity between the 2 binding sites within one effector domain strongly suggests an evolutionary development via gene duplication.\nThe high specificity of the effector domains towards Rab8 family members can be well explained from the structural analysis of Rab:bMERB complexes. Specificity-determining interactions were seen between the effector domains and the RabSF1 and RabSF2 motifs. However, additional interactions were required to increase the specificity even further, thus allowing the proteins to distinguish Rab1- and Rab8-family members. In this regard, we showed that the N-terminal residues preceding the RabSF1 motif contribute to this specificity, an observation that has previously not been made in other Rab:effector interactions. However, as alluded to in the introduction, the presence of multiple isoforms (e.g. Mical-1, Mical-L1 etc.) of the proteins, as well as the demonstrated presence of two separate binding sites, might also point towards a broader and more diverse Rab-binding spectrum and is the subject of ongoing research in our work.\nThe 3-dimensional structures of the effector domains solved in this work showed a solely α-helical fold common to many other Rab effector proteins (Oesterlin et al., 2014). Further comparison with other known Rab:effector structures showed that the main interacting helix in Rab:bMERB complexes (e.g. α-helix 3 in Mical-cL) adopts a similar position to that of the main interacting helix in the structures of Rab27:Slp2-a (Chavas et al., 2008), Rab27:Slac2a/melanophilin (Kukimoto-Niino et al., 2008) and Rab3:Rabphilin-3a (Ostermeier and Brunger, 1999) (see Figure 5—figure supplement 2 for a comparison). Interestingly, in all three examples, the Arg/Lys contacting Asp45 in our Mical:Rab structures is also conserved in these effector proteins, and the Asp/Asn following this basic residue and contacting Gln61 in Rabs is conserved in both Slp2-a and Slac2a. Intriguingly, these effector domains also display similar binding affinities as bMERB domains towards their cognate Rabs (KD = 13.4 nM for Rab27:Slp2-a, KD = 112 nM for Rab27:Slac2-a/melanophilin) (Fukuda, 2006), and these are amongst the highest affinities observed for Rab:effector interactions.\nBoth the biochemical as well as the structural analysis identified a second binding site in Mical-1, Mical-3 and EHBP1L1, thus allowing these effectors to bind Rab proteins in a 1:2 stoichiometry. In contrast to the bMERB domain proteins, Rab:effector complexes that were previously characterized display either a 1:1 or 2:2 stoichiometry, where the 2:2 complexes are usually formed by a central effector dimer with symmetrical binding interfaces on both sites (Oesterlin et al., 2014). On the other hand, multivalent Rab effector proteins have been described previously (examples are Rab4 and Rab5 binding to Rabaptin-5 (Vitale et al., 1998), Rab6 and Rab11 binding to Rab6IP1 (Miserey-Lenkei et al., 2007) or the extreme case of Gcc185 with five sites binding to Rab1a/b, Rab2a/b, Rab6a/b, Rab9a/b, Rab15, Rab27, Rab30, Rab33, Rab35 and Rab36 (Hayes et al., 2009). However, all of these effectors contain separate Rab-binding domains, each in turn only binding one Rab protein. The work presented constitutes the first description of two Rab proteins binding a single effector domain.\nThe separate binding sites within one domain not only represent a novel finding for Rab effector molecules, but also suggest intriguing and hitherto unknown functions of these proteins. Such functions could include linking Rab-decorated vesicles to a target membrane or other vesicles via a central bivalent effector. On the other hand, concerted Rab cascades and feedback loops have been observed with effector domains fused to GEFs or GAPs of one Rab acting upstream or downstream of a second Rab, helping to recruit or remove Rab proteins from a certain membrane (Pfeffer, 2013). Bivalent effectors could act in a similar manner in a positive feedback loop, initially being recruited by activated Rab proteins and subsequently helping in the recruitment and stabilization of further Rabs at this site to establish Rab membrane microdomains (Pfeffer, 2013). In fact, the presence of one high affinity and one low affinity Rab binding site as observed in some bMERB effectors could further enhance the formation of Rab microdomains: Whereas the Rab bound to the high affinity site would essentially stay bound within physiologically relevant timescales, the additional Rab protein recruited by the low affinity site could dissociate again and recruit another effector molecule via the high affinity site, thus helping to concentrate Rabs within small areas on the membrane.\nAdditionally and similar to the suggested function of Rab6IP linking Rab6 and Rab11 mediated vesicular trafficking events (Miserey-Lenkei et al., 2007), bMERB domain containing effectors might fulfill analogous functions in vesicular trafficking and act as effector Rab hubs. The possible importance of such concerted membrane recruitment cascades of Rabs and other proteins involved in membrane trafficking has been previously highlighted for Mical-L1 connecting Rab35 and Rab8, and this was aptly referred to as a membrane hub (Rahajeng et al., 2012; Giridharan et al., 2012). Furthermore, recent studies on Mical-L2 dependent GLUT4 translocation showed that trafficking was dependent on a concerted action of Rab8 and Rab13 (Sun et al., 2010, 2016). In another study it was shown that Mical-L1 is recruited to recycling endosomes by Rab35 and subsequently recruits other Rab proteins (Rab8, Rab13 and Rab36) (Kobayashi et al., 2014). In this work, the authors concluded that dimerization of Mical-L1 allows a concerted recruitment and binding of two separate Rabs to an effector dimer. However, our data now show how the 2 separate binding sites presumably also present in Mical-L1 and Mical-L2 (see Figure 6—figure supplement 1) could help in establishing this concerted action of two Rabs by connecting them via one bivalent effector protein in an (intermediate) 1:1:1 complex, thus explaining for the first time the structural and biochemical basis of the Rab hub function. The strong sequence homology of different bMERB domains including both binding sites and the fact that all Rab8 family proteins reported to interact with Micals/EHBPs are implicated in different steps of endocytic trafficking (Wandinger-Ness and Zerial, 2014) as well as previously published data thus points towards an important function of Micals/EHBPs in sorting of endocytic cargo with different destinations in the cell.\nAnother possible functional implication of the separate binding sites follows from the observation of an auto-inhibited state within Micals. Previous studies have shown that the bMERB domains in Micals can bind to their CH/LIM domains, forming an auto-inhibited intramolecular interaction that can be released by competitive binding of Rab proteins, thus allowing for the interaction with actin only after binding of Rabs (Sun et al., 2016; Sakane et al., 2010; Schmidt et al., 2008). The structural basis as well as the functional significance of this competitive binding will be an interesting topic for future research, especially regarding the binding region at the RBD responsible for auto-inhibition. The two binding sites might therefore also have separate functions, serving as a membrane recruitment site by Rabs via one site (presumably the high-affinity binding site) and release of the auto-inhibition due to competitive binding of the CH/LIM domains and Rabs at the second binding site.\nIn the last part of the study, we have shown that the two binding sites share a strong similarity and bind Rab proteins via similar residues within the interaction surfaces, thus indicating that the RBD have evolved via duplication of a common ancestor supersecondary structure element (Figure 6d) (Söding and Lupas, 2003). The underlying single α-hairpin motifs making up the separate Rab-binding sites are known from other Rab-effectors such as Rabenosyn-5 (Khan and Ménétrey, 2013; Eathiraj et al., 2005), and the resulting fused α-hairpins observed in Micals and EHBPs strongly resemble the architecture of spectrin repeats, each being connected via one continuous helix (Han et al., 2007; Pascual et al., 1997). Because of the strong similarity between the repeats of the different isoforms of Micals and EHBPs, but less similarity between both repeats within the bMERB domains (Figure 6—figure supplement 1b), we suggest that the duplication must have occurred early in evolution in one common ancestor bMERB domain. The two repeats have since diverged in terms of overall sequence, but the ability to bind Rab proteins has remained in at least some cases, as shown in this study. Exploration of the binding specificity of both binding sites of the bMERB proteins towards Rabs (and possibly other GTPases [Rahajeng et al., 2012]) will therefore be of great interest."}

{"project":"MyTest","denotations":[{"id":"27552051-18940603-26907304","span":{"begin":2099,"end":2103},"obj":"18940603"},{"id":"27552051-18940604-26907305","span":{"begin":2156,"end":2160},"obj":"18940604"},{"id":"27552051-10025402-26907306","span":{"begin":2209,"end":2213},"obj":"10025402"},{"id":"27552051-16554019-26907307","span":{"begin":2729,"end":2733},"obj":"16554019"},{"id":"27552051-9524117-26907308","span":{"begin":3458,"end":3462},"obj":"9524117"},{"id":"27552051-17725553-26907309","span":{"begin":3523,"end":3527},"obj":"17725553"},{"id":"27552051-18946081-26907310","span":{"begin":3681,"end":3685},"obj":"18946081"},{"id":"27552051-23639309-26907311","span":{"begin":4484,"end":4488},"obj":"23639309"},{"id":"27552051-23639309-26907312","span":{"begin":4761,"end":4765},"obj":"23639309"},{"id":"27552051-17725553-26907313","span":{"begin":5429,"end":5433},"obj":"17725553"},{"id":"27552051-21951725-26907314","span":{"begin":5830,"end":5834},"obj":"21951725"},{"id":"27552051-23060965-26907315","span":{"begin":5855,"end":5859},"obj":"23060965"},{"id":"27552051-21041651-26907316","span":{"begin":6023,"end":6027},"obj":"21041651"},{"id":"27552051-26538022-26907317","span":{"begin":6029,"end":6033},"obj":"26538022"},{"id":"27552051-25086062-26907318","span":{"begin":6213,"end":6217},"obj":"25086062"},{"id":"27552051-26538022-26907319","span":{"begin":7624,"end":7628},"obj":"26538022"},{"id":"27552051-20008558-26907320","span":{"begin":7645,"end":7649},"obj":"20008558"},{"id":"27552051-18305261-26907321","span":{"begin":7667,"end":7671},"obj":"18305261"},{"id":"27552051-12938173-26907322","span":{"begin":8506,"end":8510},"obj":"12938173"},{"id":"27552051-23931141-26907323","span":{"begin":8668,"end":8672},"obj":"23931141"},{"id":"27552051-16034420-26907324","span":{"begin":8691,"end":8695},"obj":"16034420"},{"id":"27552051-17356578-26907325","span":{"begin":8877,"end":8881},"obj":"17356578"},{"id":"27552051-9297972-26907326","span":{"begin":8899,"end":8903},"obj":"9297972"},{"id":"27552051-21951725-26907327","span":{"begin":9518,"end":9522},"obj":"21951725"}],"namespaces":[{"prefix":"_base","uri":"https://www.uniprot.org/uniprot/testbase"},{"prefix":"UniProtKB","uri":"https://www.uniprot.org/uniprot/"},{"prefix":"uniprot","uri":"https://www.uniprot.org/uniprotkb/"}],"text":"Discussion\nIn this publication, we present a thorough biochemical and structural analysis of a Rab effector domain termed bivalent Mical/EHBP Rab binding (bMERB) domain. The results show that the domains probably constitute a Rab8-effector family involved in endosomal trafficking, and the Rab-binding specificity can be well explained from the 3-dimensional structures of complexes determined in this work. Furthermore, we show that at least some of these domains contain two separate binding sites for Rab-proteins, suggesting previously unknown functions, as discussed below. The strong similarity between the 2 binding sites within one effector domain strongly suggests an evolutionary development via gene duplication.\nThe high specificity of the effector domains towards Rab8 family members can be well explained from the structural analysis of Rab:bMERB complexes. Specificity-determining interactions were seen between the effector domains and the RabSF1 and RabSF2 motifs. However, additional interactions were required to increase the specificity even further, thus allowing the proteins to distinguish Rab1- and Rab8-family members. In this regard, we showed that the N-terminal residues preceding the RabSF1 motif contribute to this specificity, an observation that has previously not been made in other Rab:effector interactions. However, as alluded to in the introduction, the presence of multiple isoforms (e.g. Mical-1, Mical-L1 etc.) of the proteins, as well as the demonstrated presence of two separate binding sites, might also point towards a broader and more diverse Rab-binding spectrum and is the subject of ongoing research in our work.\nThe 3-dimensional structures of the effector domains solved in this work showed a solely α-helical fold common to many other Rab effector proteins (Oesterlin et al., 2014). Further comparison with other known Rab:effector structures showed that the main interacting helix in Rab:bMERB complexes (e.g. α-helix 3 in Mical-cL) adopts a similar position to that of the main interacting helix in the structures of Rab27:Slp2-a (Chavas et al., 2008), Rab27:Slac2a/melanophilin (Kukimoto-Niino et al., 2008) and Rab3:Rabphilin-3a (Ostermeier and Brunger, 1999) (see Figure 5—figure supplement 2 for a comparison). Interestingly, in all three examples, the Arg/Lys contacting Asp45 in our Mical:Rab structures is also conserved in these effector proteins, and the Asp/Asn following this basic residue and contacting Gln61 in Rabs is conserved in both Slp2-a and Slac2a. Intriguingly, these effector domains also display similar binding affinities as bMERB domains towards their cognate Rabs (KD = 13.4 nM for Rab27:Slp2-a, KD = 112 nM for Rab27:Slac2-a/melanophilin) (Fukuda, 2006), and these are amongst the highest affinities observed for Rab:effector interactions.\nBoth the biochemical as well as the structural analysis identified a second binding site in Mical-1, Mical-3 and EHBP1L1, thus allowing these effectors to bind Rab proteins in a 1:2 stoichiometry. In contrast to the bMERB domain proteins, Rab:effector complexes that were previously characterized display either a 1:1 or 2:2 stoichiometry, where the 2:2 complexes are usually formed by a central effector dimer with symmetrical binding interfaces on both sites (Oesterlin et al., 2014). On the other hand, multivalent Rab effector proteins have been described previously (examples are Rab4 and Rab5 binding to Rabaptin-5 (Vitale et al., 1998), Rab6 and Rab11 binding to Rab6IP1 (Miserey-Lenkei et al., 2007) or the extreme case of Gcc185 with five sites binding to Rab1a/b, Rab2a/b, Rab6a/b, Rab9a/b, Rab15, Rab27, Rab30, Rab33, Rab35 and Rab36 (Hayes et al., 2009). However, all of these effectors contain separate Rab-binding domains, each in turn only binding one Rab protein. The work presented constitutes the first description of two Rab proteins binding a single effector domain.\nThe separate binding sites within one domain not only represent a novel finding for Rab effector molecules, but also suggest intriguing and hitherto unknown functions of these proteins. Such functions could include linking Rab-decorated vesicles to a target membrane or other vesicles via a central bivalent effector. On the other hand, concerted Rab cascades and feedback loops have been observed with effector domains fused to GEFs or GAPs of one Rab acting upstream or downstream of a second Rab, helping to recruit or remove Rab proteins from a certain membrane (Pfeffer, 2013). Bivalent effectors could act in a similar manner in a positive feedback loop, initially being recruited by activated Rab proteins and subsequently helping in the recruitment and stabilization of further Rabs at this site to establish Rab membrane microdomains (Pfeffer, 2013). In fact, the presence of one high affinity and one low affinity Rab binding site as observed in some bMERB effectors could further enhance the formation of Rab microdomains: Whereas the Rab bound to the high affinity site would essentially stay bound within physiologically relevant timescales, the additional Rab protein recruited by the low affinity site could dissociate again and recruit another effector molecule via the high affinity site, thus helping to concentrate Rabs within small areas on the membrane.\nAdditionally and similar to the suggested function of Rab6IP linking Rab6 and Rab11 mediated vesicular trafficking events (Miserey-Lenkei et al., 2007), bMERB domain containing effectors might fulfill analogous functions in vesicular trafficking and act as effector Rab hubs. The possible importance of such concerted membrane recruitment cascades of Rabs and other proteins involved in membrane trafficking has been previously highlighted for Mical-L1 connecting Rab35 and Rab8, and this was aptly referred to as a membrane hub (Rahajeng et al., 2012; Giridharan et al., 2012). Furthermore, recent studies on Mical-L2 dependent GLUT4 translocation showed that trafficking was dependent on a concerted action of Rab8 and Rab13 (Sun et al., 2010, 2016). In another study it was shown that Mical-L1 is recruited to recycling endosomes by Rab35 and subsequently recruits other Rab proteins (Rab8, Rab13 and Rab36) (Kobayashi et al., 2014). In this work, the authors concluded that dimerization of Mical-L1 allows a concerted recruitment and binding of two separate Rabs to an effector dimer. However, our data now show how the 2 separate binding sites presumably also present in Mical-L1 and Mical-L2 (see Figure 6—figure supplement 1) could help in establishing this concerted action of two Rabs by connecting them via one bivalent effector protein in an (intermediate) 1:1:1 complex, thus explaining for the first time the structural and biochemical basis of the Rab hub function. The strong sequence homology of different bMERB domains including both binding sites and the fact that all Rab8 family proteins reported to interact with Micals/EHBPs are implicated in different steps of endocytic trafficking (Wandinger-Ness and Zerial, 2014) as well as previously published data thus points towards an important function of Micals/EHBPs in sorting of endocytic cargo with different destinations in the cell.\nAnother possible functional implication of the separate binding sites follows from the observation of an auto-inhibited state within Micals. Previous studies have shown that the bMERB domains in Micals can bind to their CH/LIM domains, forming an auto-inhibited intramolecular interaction that can be released by competitive binding of Rab proteins, thus allowing for the interaction with actin only after binding of Rabs (Sun et al., 2016; Sakane et al., 2010; Schmidt et al., 2008). The structural basis as well as the functional significance of this competitive binding will be an interesting topic for future research, especially regarding the binding region at the RBD responsible for auto-inhibition. The two binding sites might therefore also have separate functions, serving as a membrane recruitment site by Rabs via one site (presumably the high-affinity binding site) and release of the auto-inhibition due to competitive binding of the CH/LIM domains and Rabs at the second binding site.\nIn the last part of the study, we have shown that the two binding sites share a strong similarity and bind Rab proteins via similar residues within the interaction surfaces, thus indicating that the RBD have evolved via duplication of a common ancestor supersecondary structure element (Figure 6d) (Söding and Lupas, 2003). The underlying single α-hairpin motifs making up the separate Rab-binding sites are known from other Rab-effectors such as Rabenosyn-5 (Khan and Ménétrey, 2013; Eathiraj et al., 2005), and the resulting fused α-hairpins observed in Micals and EHBPs strongly resemble the architecture of spectrin repeats, each being connected via one continuous helix (Han et al., 2007; Pascual et al., 1997). Because of the strong similarity between the repeats of the different isoforms of Micals and EHBPs, but less similarity between both repeats within the bMERB domains (Figure 6—figure supplement 1b), we suggest that the duplication must have occurred early in evolution in one common ancestor bMERB domain. The two repeats have since diverged in terms of overall sequence, but the ability to bind Rab proteins has remained in at least some cases, as shown in this study. Exploration of the binding specificity of both binding sites of the bMERB proteins towards Rabs (and possibly other GTPases [Rahajeng et al., 2012]) will therefore be of great interest."}