FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2114578747> ?p ?o ?g. }

• W2114578747 endingPage "1477" @default.
• W2114578747 startingPage "1469" @default.
• W2114578747 abstract "Article9 April 2013free access Structure of the Legionella effector AnkX reveals the mechanism of phosphocholine transfer by the FIC domain Valérie Campanacci Valérie Campanacci Laboratoire d'Enzymologie et Biochimie Structurales, Centre de Recherche de Gif, CNRS, Gif-sur-Yvette, France Search for more papers by this author Shaeri Mukherjee Shaeri Mukherjee Department of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT, USA Search for more papers by this author Craig R Roy Craig R Roy Department of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT, USA Search for more papers by this author Jacqueline Cherfils Corresponding Author Jacqueline Cherfils Laboratoire d'Enzymologie et Biochimie Structurales, Centre de Recherche de Gif, CNRS, Gif-sur-Yvette, France Search for more papers by this author Valérie Campanacci Valérie Campanacci Laboratoire d'Enzymologie et Biochimie Structurales, Centre de Recherche de Gif, CNRS, Gif-sur-Yvette, France Search for more papers by this author Shaeri Mukherjee Shaeri Mukherjee Department of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT, USA Search for more papers by this author Craig R Roy Craig R Roy Department of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT, USA Search for more papers by this author Jacqueline Cherfils Corresponding Author Jacqueline Cherfils Laboratoire d'Enzymologie et Biochimie Structurales, Centre de Recherche de Gif, CNRS, Gif-sur-Yvette, France Search for more papers by this author Author Information Valérie Campanacci1, Shaeri Mukherjee2, Craig R Roy2 and Jacqueline Cherfils 1 1Laboratoire d'Enzymologie et Biochimie Structurales, Centre de Recherche de Gif, CNRS, Gif-sur-Yvette, France 2Department of Microbial Pathogenesis, Yale University School of Medicine, New Haven, CT, USA *Corresponding author. Laboratoire d'Enzymologie et Biochimie Structurales, CNRS, 1, Avenue de la Terrasse, Gif-sur-Yvette 91198, France. Tel.:+33 169823492; Fax:+33 169823129; E-mail: [email protected] The EMBO Journal (2013)32:1469-1477https://doi.org/10.1038/emboj.2013.82 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The FIC motif and the eukaryotic-like ankyrin repeats are found in many bacterial type IV effectors, yet little is known about how these domains enable bacteria to modulate host cell functions. Bacterial FIC domains typically bind ATP and transfer adenosine monophosphate moiety onto target proteins. The ankyrin repeat-containing protein AnkX encoded by the intracellular pathogen Legionella pneumophila is unique in that its FIC domain binds to CDP-choline and transfers a phosphocholine residue onto proteins in the Rab1 GTPase family. By determining the structures of unbound AnkX and AnkX with bound CDP-choline, CMP/phosphocholine and CMP, we demonstrate that the orientation of substrate binding in relation to the catalytic FIC motif enables this protein to function as a phosphocholinating enzyme rather than a nucleotidyl transferase. Additionally, the structure reveals that the ankyrin repeats mediate scaffolding interactions that resemble those found in protein–protein interactions, but are unprecedented in intramolecular interactions. Together with phosphocholination experiments, our structures unify a general phosphoryl transferase mechanism common to all FIC enzymes that should be conserved from bacteria to human. Introduction Many intracellular pathogens use specialized secretion systems to deliver proteins into host cells called effectors, which can have biochemical activities that modulate the cytoskeletal and membrane transport pathways of the infected host cell (reviewed in Cossart and Roy, 2010). The FIC domain (filamentation induced by cyclic adenosine monophosphate moiety (AMP)) and the eukaryotic-like ankyrin repeat domain are modules contained in many different bacterial effectors (reviewed in Roy and Mukherjee, 2009; Voth, 2011). Ankyrin repeats consist of strings of helix-helix-turn units, and in eukaryotes are among the most common protein–protein interaction motifs (reviewed in Li et al, 2006). It is thought that ankyrin repeats in bacterial effectors will serve as interaction modules involved in the targeting of host factors, but there is no clear evidence demonstrating such a role, and the function of most bacterial Ank proteins remains unknown. FIC domains represent another homology region found in multiple bacterial effectors. Work on the Vibrio parahaemolyticus type III effector VopS showed that the FIC domain in this protein had an enzymatic activity that resulted in the transfer of an AMP onto host Rho GTPases, resulting in the disruption of the actin cytoskeleton (Yarbrough et al, 2009). This adenylylation reaction mediated by VopS was termed as AMPylation. The protein IbpA from the respiratory pathogen Histophilus somni was also shown to have an activity that mediated adenylylation of Rho family GTPases to disrupt their function (Kinch et al, 2009; Worby et al, 2009). Thus, the FIC domain, which is conserved outside the prokaryotic kingdom (Kinch et al, 2009; Worby et al, 2009), was thought to have a conserved function in mediating the post-translational addition of AMP onto hydroxyl-containing residues of target proteins (Kinch et al, 2009; Roy and Mukherjee, 2009). Mechanisms for ATP cleavage and AMP transfer have been proposed recently based on structural and biochemical analysis of representative bacterial FIC domain-containing enzymes, which include VopS (Luong et al, 2010), IbpA bound to AMPylated Cdc42 (Xiao et al, 2010), Bartonella BepA with bound pyrophosphate (PPi) (Palanivelu et al, 2011) and Neisseria NmFic bound to AMP and ATP (Engel et al, 2012). These studies highlighted the conserved fold of the FIC domain and the catalytic role of its HPFx[D/E]GN[G/K]R signature in cleaving the Pα-Oαβ bond of ATP (Luong et al, 2010; Xiao et al, 2010; Palanivelu et al, 2011), as well as the contribution of adjacent elements in binding the adenine (Xiao et al, 2010; Engel et al, 2012) and possibly in regulating the AMPylation activity (Engel et al, 2012). It should be noted that although all characterized FIC enzymes seem to have a narrow protein target specificity, most undergo auto-modification (Kinch et al, 2009; Xiao et al, 2010; Engel et al, 2012; Feng et al, 2012; Goody et al, 2012). Although the significance of auto-modification remains unknown, it has been proposed to represent an intermediate in the phosphoryl transfer reaction mediated by the FIC domain or possibly be important in regulating the activity of the enzyme (Luong et al, 2010; Goody et al, 2012). The paradigm that all FIC domain-containing effectors function as AMPylating enzymes was challenged by the recent discovery that Legionella type IV effector protein AnkX uses CDP-choline as a substrate to transfer phosphocholine onto serine or threonine in conserved positions in the switch II region of Rab1 family GTPases (Mukherjee et al, 2011). Additionally, it has been shown that Xanthomonas AvrAC can use UTP to transfer UMP onto plant receptor-like kinases (Feng et al, 2012). Thus, FIC domains are able to catalyse chemical reactions others than AMPylation. All of the FIC domain-containing effectors that have been characterized mediate a phosphoryl transfer reaction requiring a nucleotide-based substrate. What makes the reaction mediated by the FIC domain in AnkX unique is that the transferred moiety is not the nucleotide monophosphate (CMP) but is instead the phosphocholine moiety. AnkX is also unique in that it carries multiple ankyrin repeats in addition to the FIC domain on a single polypeptide. When expressed in mammalian cells, AnkX modifies Rab1 and Rab35 proteins to disrupt membrane transport processes mediates by these GTPases, and this activity requires a conserved histidine residue found in the catalytic site of the FIC motif and several predicted ankyrin repeat domains located nearby in the amino-acid sequence (Pan et al, 2008; Mukherjee et al, 2011). The defects in membrane transport are likely due to impairments in the interaction of phosphocholinated Rab1 proteins with RabGDIs, RabGEFs, and cellular effectors (Mukherjee et al, 2011; Goody et al, 2012). Here, we report the crystallographic structures of apo-AnkX and of AnkX bound to its substrate, CDP-choline, and to its products, CMP and phosphocholine. These structural snapshots, together with further analysis of the phosphocholination reaction, provide important insight into how FIC and ankyrin domains have been co-opted by bacteria, and suggest a general model for the substrate recognition and phosphoryl transferase mechanism common to FIC enzymes from bacteria to human. Results The crystal structure of AnkX reveals that the ankyrin repeats scaffold constrains the FIC and substrate-binding domains Limited proteolysis studies indicated that the smallest protease-resistant domain that retained Rab1 phosphocholination activity contained residues 1–484 in AnkX, so this region was selected for protein crystallization studies. We solved the crystal structures of L. pneumophila AnkX1–484 in the apo form, in complex with its substrate, CDP-choline (using the inactive AnkXH229A mutant), and in complexes with CMP and phosphocholine and with CMP alone (Figure 1A; Table I). AnkX has an atypical FIC domain comprised of two sub-domains, of which the second carries a canonical FIC motif (Supplementary Figure 1A and B). The FIC sub-domains are separated by a 70-residue domain, which inserts within a β-hairpin that is conserved in all known FIC domain structures (Supplementary Figure 1A and C). Unlike the β-hairpin of the FIC domain of IbpA, which forms the binding site for AMPylated Cdc42 (Xiao et al, 2010), the β-hairpin of AnkX is covered by the insert domain and is unavailable for binding Rab1, suggesting that Rab1 has a different binding modality (Supplementary Figure 2). Next to the FIC motif, AnkX features a unique domain comprised of the N-terminal 50 residues and of 40 residues located downstream of the FIC domain, which binds the CMP moiety of CDP-choline (Supplementary Figure 1A and D). Residues beyond Pro351 form four consecutive ankyrin repeats (Figure 1A and B), two of which were not predicted using standard methods (Pan et al, 2008). All domains establish multiple interactions with each other, yielding a close-packed structure (Figure 1C). Remarkably, the ankyrin repeats are involved in a large interface with the CMP binding, FIC, and insert domain (about 2200 Å2 buried surface area, intramolecular contacts in Supplementary Figure 3), which resembles the interfaces found in protein–protein interactions but is seen for the first time in an intramolecular interaction. Consistent with this unprecedented role of the ankyrin repeats in constraining the structure of AnkX, a recombinant AnkX construct lacking the ankyrin repeats did not yield soluble expression in E. coli. Thus, our structures reveal that ankyrin repeats are not only protein–protein interaction domains, but can also contribute to intramolecular interactions. Figure 1.Structure of L. pneumophila AnkX. (A) Structure of AnkX with bound CDP-choline (in green). The sequences of the CMP binding, FIC, insert and ankyrin repeat domains are given in Supplementary Figure 1A. (B) Overlay of the four ankyrin repeats of AnkX1–484 (in magenta) to an archetypal ankyrin repeats protein (AnkyrinR, in orange Michaely et al, 2002). The longer loop in the first ankyrin motif is shown by an arrow. (C) The ankyrin repeats (in magenta) form extensive intramolecular contacts with the CMP-binding domain (in orange), the FIC domain (in cyan), and the insert domain (in red). Intramolecular contacts are shown in Supplementary Figure 3. Download figure Download PowerPoint Table 1. Data collection and refinement statistics AnkX1–484 AnkX1–484 AnkX/CMP AnkX/CMP/phosphocholine AnkXH229A/CDP-choline Data collection Form I Form II Form III Form IV Phasing method MR-SAD SAD MR MR MR Space group P212121 P21 P2221 P21212 P21 Cell constants (Å) 55.3, 91.6, 239.1 59.1, 198.4, 110.0β=98.5° 58.3, 103.6, 224.1 84.3, 122.3, 55.0 66.1, 122.1, 75.7β=107.3° Molecule/asu 2 4 2 1 2 Wavelength (Å) 0.9792 0.9790 0.9801 0.9801 0.9801 Resolution rangea (Å) 44.0–3.14(3.33–3.14) 48.63–3.60(3.69–3.60) 46.28–2.60(2.76–2.60) 43.10–2.54(2.68–2.54) 43.89–2.55(2.70–2.55) Rsyma,b, a,b (%) 9.2 (52.3) 7.6 (68.3) 10.5 (101.1) 10.1 (74.7) 9.1 (83.7) I/σIa 11.33 (2.59) 11.06 (1.99) 16.16 (2.30) 13.42 (2.50) 12.02 (1.75) Completenessa (%) 98.9 (96.9) 99.2 (92.1) 99.6 (98.8) 99.7 (98.8) 98.5 (97.4) Redundancya 3.8 (3.8) 4.9 (3.4) 7.3 (7.4) 5.8 (5.7) 4.7 (4.7) Chain A 5–482 1–328, 331–477 6–85, 96–484 5–484 Chain B 5–86, 96–477 1–85, 90–479 6–484 No. of reflections 20 682 42 690 19 367 36 935 Rwork/Rfree (%) 21.18/25.64 19.58/23.26 19.43/25.71 19.23/24.32 No. of atoms/waters 7434 7582/73 3738/54 7643/114 Mean B-factor (Å2) Protein 43.1 65.8 55.9 69.5 Ligands — 59.7 (CMP:A) 59.7 (CMP) 59.3 (CDC:A) — 62.5 (CMP:B) 99.4 (PC) 52.31 (CDC:B) r.m.s.d. Bond lengths (Å) 0.011 0.010 0.010 0.010 Bond angle (deg) 1.58 1.18 1.21 1.24 a Values in parentheses refer to the highest resolution shell. b Rsym=∑h∑i|Ihi–(Ih)|/∑h∑I(Ih), where Ii is the ith observation of reflection h and (Ih) is the weighted average intensity for all observations l of reflection h. The orientation of CDP-choline binding to AnkX provides a mechanism for transfer of phosphocholine by the FIC motif It was unknown why AnkX mediates the transfer of phosphocholine rather than CMP given that it carries the canonical catalytic FIC motif that functions in nucleotide transfer in AMPylating toxins. Our structures show that the AnkX FIC motif has the same conformation as in known AMPylating FIC enzymes (Supplementary Figure 4A). CDP-choline shares a nucleoside diphosphate group with ATP, but its binding to AnkX with the orientation of ATP as seen in AMPylating FIC domains would result in a head-to-tail chemical reaction with inversion of the transferred and leaving groups. Unprocessed CDP-choline was captured in complex with the inactive AnkXH229A mutant (Figure 2A), revealing that choline, rather than cytidine, overlays the adenosine moiety of ATP (Figure 2B). The positive charge of the trimethylammonium group of the choline moiety is recognized by the electron-rich aromatic ring of Phe107 in the β-hairpin and the carboxyls of two acidic residues, Asp265 and Glu226 in the FIC motif-containing sub-domain (Figure 2C). Phe107, Ile109, and Asp265 are located in the vicinity of the choline moiety and should exclude the binding of the adenine base through steric hindrance (Figure 2B), which would explain why AnkX cannot function as an AMPylating enzyme. The other unique feature of the active site pocket of AnkX is that it snugly fits the cytidine group, as shown by its well-defined structure in the complexes containing AnkXH229A/CDP-choline (Figure 2A), AnkX/CMP, and AnkX/CMP/phosphocholine. Specific recognition is mediated by the CMP domain unique to AnkX, which forms numerous interactions with the base (Tyr 41, Arg 44, and Cys 48) and the sugar (Asp28 and Arg30) (Figure 2C). Tyr41, notably, stacks its aromatic ring on the cytosine ring of CDP-choline. Mutations in the predicted choline-binding residues (Phe107Gly, Glu226Ala, and Asp265Ala) and of a cytosine-binding residue (Tyr41Ala), all of which are located outside the FIC motif, impaired Rab1 phosphocholination in vitro (Figure 3A; Supplementary Figure 5A), consistent with these residues playing an important role in enzyme function. The binding site of CDP-choline to AnkX shares a common use of aromatic and negatively charged residues to recognize the choline moiety with CTP-phosphocholine cytidylyltransferases from bacteria (Kwak et al, 2002) and mammals (Lee et al, 2009), but is otherwise completely different from that of these enzymes (Supplementary Figure 6). Thus, these structural snapshots provide molecular details on how AnkX uses its CMP domain and FIC domain to uniquely recognize CDP-choline as a substrate and exclude ATP, which makes AnkX a phosphocholine transferase exclusively. Figure 2.Structure of the CDP-choline binding site of AnkX. (A) Fo-Fc electron density omit map of CDP-choline bound to AnkXH229A contoured at 2.5σ. (B) Superposition of CDP-choline bound to AnkX (in orange) and AMP bound to cdc42 in the IbpA-cdc42 complex (in magenta, PDB 3N3V). Note that the AnkX active site (yellow surface) cannot accommodate the base of AMP. (C) Interactions of the CMP domain (in orange) and FIC domain (in cyan) residues with the cytidine and choline moieties of CDP-choline. Hydrogen bonds are shown in dotted lines. Download figure Download PowerPoint Figure 3.Analysis of Rab1 phosphocholination and AnkX auto-phosphocholination. (A) AnkX constructs carrying mutations in the CDP-choline binding site have impaired Rab1 phosphocholination. Rab1 phosphocholination is expressed as the relative percentage of Rab1 phosphocholination by wild-type AnkX. Data are from immunoblots from in vitro reactions that contained full-length His-tagged AnkX constructs and Rab1A in the presence of phosphocholination buffer. Blots were probed with an anti-PC antibody. Representative blots are shown in Supplementary Figure 5A. Errors bars represent the standard deviation based on three independent experiments. (B) AnkX constructs carrying mutations in the FIC motif have impaired Rab1 phosphocholination. Experimental conditions are as in Figure 3A. Representative blots are shown in Supplementary Figure 5B. Errors bars represent the standard deviation based on three independent experiments. (C) Mg2+ is necessary for phosphocholination. Full-length His-tagged AnkX was dialysed against a metal-free buffer prior to the experiment, and subsequently assessed for auto-phosphocholination without (left lane) or with (right lane) addition of 1 mM Mg2+ in the auto-phosphocholination buffer. (D) Auto-phosphocholinated residues are not located in the FIC domain. The AnkX1–484 construct (lane 1) phosphocholinates Rab1 but is not auto-phosphocholinated. The AnkX1–484 construct carrying the H229A mutation is shown as a control that Rab1 phosphocholination requires the FIC motif histidine (lane 2). Molecular weight markers (in kDa) are shown on the right. Download figure Download PowerPoint AnkX catalyses the same phosphoryl transfer reaction as AMPylating FIC domains The structures of substrate- and product-bound AnkX (Figure 4A and B) suggest that the catalytic mechanism of AnkX phosphoryl transfer would be similar to that of AMPylating FIC enzymes. FIC motif-based overlay of AnkX/CDP-choline to IbpA/AMPylated Cdc42 shows that the β-phosphate of CDP-choline superimposes with the α-phosphate of AMP (Figure 4C), indicating that the scissile bond in CDP-choline is the Oαβ-Pβ bond, rather than the Pα-Oαβ of ATP that is cleaved by AMPylating enzymes. Furthermore, the Oαβ-Pβ bond of CDP-choline is exactly co-linear with the phosphotyrosine bond of AMPylated Cdc42 with inversion of configuration of the phosphate (Figure 4C), as would be expected if both enzymes would have the same catalytic mechanisms. Consistently, mutations of conserved residues in the FIC motif in AnkX (His229Ala, Asp233Ala, and Arg237Glu) all severely impaired both auto-phosphocholination and Rab1 phosphocholination in vitro (Figure 3B; Supplementary Figure 5B). Thus, a major consequence arising from CDP-choline binding head-to-tail with respect to ATP is that AMPylation and phosphocholination can be achieved through the same chemical reaction, with conservation of the positions of the transferred and leaving groups relative to the FIC motif and with the same use of catalytic residues. Figure 4.Structural basis for the phosphocholination reaction. (A) Interactions of the phosphates of CDP-choline with AnkXH229A. His229 in the FIC motif is overlaid from wild-type AnkX. (B) Interactions of the phosphates of CMP and phosphocholine with wild-type AnkX. (C) The Oαβ-Pβ bond of CDP-choline aligns with the phosphotyrosine bond of Cdc42 in complex with IbpA. (D) Proposed interactions of Mg2+ with CDP-choline modelled from Bartonella BepA/PPi-Mg2+ (PDB 2JK8). PPi bound to BepA is overlaid (in light blue). (E) Model for the transition state of phosphoryl transfer by the conserved FIC motif. Hydrogen bonds are in dotted lines, partial bonds are in dashed lines. Superpositions in Figure 4A–D are based on the FIC motif. Download figure Download PowerPoint The catalytic mechanism of CDP-choline cleavage by AnkX Our structures reveal a diphosphate-containing substrate both before and after its cleavage by an FIC domain, allowing for unprecedented insight into the catalytic mechanism. In these structures, phosphoryl transfer to the hydroxyl-containing serine residue in Rab1 was mimicked by a water molecule. We observe that the CDP-choline-binding site has the same conformation in apo-AnkX, AnkXH229A/CDP-choline, and AnkX/CMP/phosphocholine (Supplementary Figure 4B), from which it can be safely assumed that the reaction does not require a disorder-to-order transition or an induced fit. This suggests that the transition state of CDP-choline cleavage is structurally close to the ground states observed in the crystals and can be analysed from these structures. An important contribution to the efficiency of biological phosphoryl transfer reactions is the stabilization of partial negative charges that develop at the transition state, which is provided by positive charges from enzyme active sites (reviewed in Lassila et al, 2011). AnkX has only one positively charged residue that interacts with the α-phosphate of both the substrate and leaving group, Arg237 from the FIC motif, and none that interacts with the β-phosphate (Figure 4A and B). Additional stabilization could stem from a cationic metal, one candidate being a Mg2+ cation. As shown in Figure 3C, auto-phosphocholination was impaired when no cationic metal was present in the mixture, and was recovered by the addition of Mg2+, indicating that the reaction requires a cationic ion. A Mg2+ ion has been observed in the structure of the BepA/pyrophosphate complex, where it bridges the phosphates of the ligand (Palanivelu et al, 2011). Superposition of this structure with CDP-choline-bound AnkX suggests that Mg2+ is coordinated by the conserved FIC motif aspartate (Figure 4D), which would be consistent with the critical role of this residue despite the fact that it does not make direct contacts with CDP-choline. An intriguing aspect of FIC enzymes, whose mechanism is currently debated, is their robust auto-modifications, which depend on the catalytic histidine and occurs regardless of whether they catalyse AMPylation, UMPylation, or phosphocholination reactions. One explanation that has been put forward is that the auto-modified species represents the accumulation of a phosphohistidine intermediate, which would indicate a ping-pong mechanism of phosphoryl transfer (Goody et al, 2012). Comparison of substrate-bound AnkX and product-bound IbpA identifies a conserved geometry for in-line nucleophilic attack of the transferred phosphorus (Figure 4C) that could not be structurally accommodated in a ping-pong mechanism, which would require that at least one of the nucleophilic attack is occurring from the opposite side of the phosphorus. We found that the AnkX1–484 protein used in our crystallization studies was not auto-phosphocholinated in vitro, even though it retained the ability to phophocholinate Rab1 by a mechanism that required the catalytic histidine (Figure 3D; Supplementary Figure 5C). These data suggest that the auto-phosphocholinated AnkX species does not represent an intermediate in the phosphoryl transfer reaction and that the auto-phosphocholinated residues must be located outside of the enzymatic domain in AnkX1–484. Thus, AnkX uses a concerted mechanism of phosphoryl transfer, in which in-line nucleophilic attack of the β-phosphate of CDP-choline mediated by the hydroxyl group of Ser79 in Rab1 is mimicked by a water molecule, according to the transition state summarized in Figure 4E. Discussion A large number of bacterial effector proteins contain ankyrin repeats and FIC domains, but understanding how these domains have been adapted for diverse functions remains a critical question. Our structures of AnkX provide new insight into how these domains function. These data show that the ankyrin repeats in AnkX have been co-opted by bacteria to constrain intramolecular interactions, in contrast with most eukaryotic ankyrin repeats where the exposed loops and flexible structure of the repeat functions as a versatile protein–protein interaction module (reviewed in Li et al, 2006). Sequence analysis predicts that AnkX may have up to 12 ankyrin repeats (Supplementary Figure 1A), in which the loops have more variable lengths than in eukaryotic ankyrin repeats, as exemplified by the loop of the first motif (Figure 1B). These repeats can be readily modelled by extending the first four repeats without clashes with the rest of the structure (Figure 1B), leaving open the possibility that they form additional interactions, possibly with Rab1 or other partners. FIC domains have recently been classified as an evolutionary conserved family present in genomes from bacteria, archae, fungi, and metazoan (Kinch et al, 2009; Worby et al, 2009). Characterizing what makes their air de famille and what defines their individual specificities is important to understand their roles in bacterial infections and eukaryotic regulatory functions. Our analysis of substrate- and product-bound AnkX, in perspective with previous analysis of AMPylating FIC proteins, shows that the FIC motif has a general role in catalysing the cleavage of diphosphate-containing substrates and transfer of the resulting phosphoryl-containing moiety onto an entering hydroxyl group, regardless of the specific nature of the substrate. Catalysis would be best explained by in-line nucleophilic attack by the entering hydroxyl group which is activated by the invariant histidine of the HPFx[D/E]GN[G/K]R signature, with stabilization of the developing negative charges by the invariant arginine and by a metal cation bound to the conserved acidic residue, according to the transition state depicted in Figure 4E. We propose that auto-modification, a side reaction reported in all characterized FIC enzymes, occurs using the same catalytic mechanism, but likely target residues located in poorly folded sequences outside of the catalytic region of the enzyme that can readily make it to the active site. This could explain why auto-AMPylation of the C-terminal helix of NmFIC increases upon mutations that render this helix flexible (Engel et al, 2012). When the structure of CDP-choline bound to AnkX is compared with the structure of AMP bound to IbpA (Xiao et al, 2010) and ATP bound to NmFIC (Engel et al, 2012), the only AMPylating toxins where these ligands have been observed, it becomes clear that the transferred and leaving groups of diphosphate-containing substrates are specifically recognized by variable regions located outside the FIC motif at both ends of the active site. The AMP and phosphocholine transferred groups bind near the β-hairpin of the FIC domain (Figure 5A), in which specific residues exclude illegitimate ligands (Figure 2B). The structures of AnkX in complex with CDP-choline and CMP also show for the first time how the leaving group is recognized by a native FIC domain, since ATP was observed only in an NmFIC mutant in which an extra C-terminal α-helix had been truncated (Xiao et al, 2010), whereas PPi bound to BepA (Palanivelu et al, 2011) likely mimicked the Pα-Pβ moiety of ATP rather than the leaving Pβ-Pγ group (see Figure 4D). The CMP domain in AnkX has highly variable counterparts in FIC proteins of known structures (Figure 5A and B). We propose that these extra structural elements determine the nature of the physiological substrate and/or exclude illegitimate ligands. FIC enzymes whose binding of the γ-phosphate of ATP is hindered by an extra α-helix, such as NmFIC, may thus have a physiological substrate other than ATP, a plausible alternative to their proposed auto-inhibition by this helix (Engel et al, 2012). According to this model, the fact that Pseudomonas AvrB, which is remotely related to FIC proteins (Kinch et al, 2009), binds ADP with the same orientation as the CDP moiety of CDP-choline to AnkX (Figure 5C; Supplementary Figure 7), would suggest that it could use ATP to transfer a phosphate group, consistent with its reported protein kinase activity (Desveaux et al, 2007). Figure 5.The active sites of FIC domains assemble from the conserved catalytic FIC motif and variable substrate-binding regions. (A) Superposition of active sites of phosphocholinating (AnkX, orange), AMPylating (IbpA, magenta) and unknown function (NmFic, blue and BepA, cyan) FIC enzymes. The FIC motif is shown in yellow and is in the same orientation in all views. Structural elements that recognize the leaving group are on the left, and the regions that recognize the transferred phosphoryl group are on the right. The latter regions are probably also involved in bind" @default.
• W2114578747 created "2016-06-24" @default.
• W2114578747 creator A5041693932 @default.
• W2114578747 creator A5045032777 @default.
• W2114578747 creator A5080222404 @default.
• W2114578747 creator A5084806589 @default.
• W2114578747 date "2013-04-09" @default.
• W2114578747 modified "2023-10-10" @default.
• W2114578747 title "Structure of the Legionella effector AnkX reveals the mechanism of phosphocholine transfer by the FIC domain" @default.
• W2114578747 cites W1726597626 @default.
• W2114578747 cites W1790797992 @default.
• W2114578747 cites W1984185074 @default.
• W2114578747 cites W1994475681 @default.
• W2114578747 cites W1995972219 @default.
• W2114578747 cites W2002756887 @default.
• W2114578747 cites W2011747741 @default.
• W2114578747 cites W2030979689 @default.
• W2114578747 cites W2032432541 @default.
• W2114578747 cites W2054275665 @default.
• W2114578747 cites W2056979628 @default.
• W2114578747 cites W2081435829 @default.
• W2114578747 cites W2083786522 @default.
• W2114578747 cites W2089390902 @default.
• W2114578747 cites W2090188358 @default.
• W2114578747 cites W2121950916 @default.
• W2114578747 cites W2122339645 @default.
• W2114578747 cites W2124026197 @default.
• W2114578747 cites W2124677069 @default.
• W2114578747 cites W2129196764 @default.
• W2114578747 cites W2142823843 @default.
• W2114578747 cites W2155682217 @default.
• W2114578747 cites W2159211495 @default.
• W2114578747 cites W2161936455 @default.
• W2114578747 cites W2162244749 @default.
• W2114578747 cites W2163306024 @default.
• W2114578747 cites W2163341755 @default.
• W2114578747 cites W2172191429 @default.
• W2114578747 cites W2180229411 @default.
• W2114578747 cites W4248872320 @default.
• W2114578747 doi "https://doi.org/10.1038/emboj.2013.82" @default.
• W2114578747 hasPubMedCentralId "https://www.ncbi.nlm.nih.gov/pmc/articles/3655470" @default.
• W2114578747 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/23572077" @default.
• W2114578747 hasPublicationYear "2013" @default.
• W2114578747 type Work @default.
• W2114578747 sameAs 2114578747 @default.
• W2114578747 citedByCount "64" @default.
• W2114578747 countsByYear W21145787472013 @default.
• W2114578747 countsByYear W21145787472014 @default.
• W2114578747 countsByYear W21145787472015 @default.
• W2114578747 countsByYear W21145787472016 @default.
• W2114578747 countsByYear W21145787472017 @default.
• W2114578747 countsByYear W21145787472018 @default.
• W2114578747 countsByYear W21145787472019 @default.
• W2114578747 countsByYear W21145787472020 @default.
• W2114578747 countsByYear W21145787472021 @default.
• W2114578747 countsByYear W21145787472022 @default.
• W2114578747 countsByYear W21145787472023 @default.
• W2114578747 crossrefType "journal-article" @default.
• W2114578747 hasAuthorship W2114578747A5041693932 @default.
• W2114578747 hasAuthorship W2114578747A5045032777 @default.
• W2114578747 hasAuthorship W2114578747A5080222404 @default.
• W2114578747 hasAuthorship W2114578747A5084806589 @default.
• W2114578747 hasBestOaLocation W21145787472 @default.
• W2114578747 hasConcept C121332964 @default.
• W2114578747 hasConcept C134306372 @default.
• W2114578747 hasConcept C2776330855 @default.
• W2114578747 hasConcept C2776923060 @default.
• W2114578747 hasConcept C2778918659 @default.
• W2114578747 hasConcept C2779164008 @default.
• W2114578747 hasConcept C33923547 @default.
• W2114578747 hasConcept C36503486 @default.
• W2114578747 hasConcept C41625074 @default.
• W2114578747 hasConcept C51785407 @default.
• W2114578747 hasConcept C523546767 @default.
• W2114578747 hasConcept C54355233 @default.
• W2114578747 hasConcept C62520636 @default.
• W2114578747 hasConcept C70721500 @default.
• W2114578747 hasConcept C86803240 @default.
• W2114578747 hasConcept C89611455 @default.
• W2114578747 hasConcept C95444343 @default.
• W2114578747 hasConceptScore W2114578747C121332964 @default.
• W2114578747 hasConceptScore W2114578747C134306372 @default.
• W2114578747 hasConceptScore W2114578747C2776330855 @default.
• W2114578747 hasConceptScore W2114578747C2776923060 @default.
• W2114578747 hasConceptScore W2114578747C2778918659 @default.
• W2114578747 hasConceptScore W2114578747C2779164008 @default.
• W2114578747 hasConceptScore W2114578747C33923547 @default.
• W2114578747 hasConceptScore W2114578747C36503486 @default.
• W2114578747 hasConceptScore W2114578747C41625074 @default.
• W2114578747 hasConceptScore W2114578747C51785407 @default.
• W2114578747 hasConceptScore W2114578747C523546767 @default.
• W2114578747 hasConceptScore W2114578747C54355233 @default.
• W2114578747 hasConceptScore W2114578747C62520636 @default.
• W2114578747 hasConceptScore W2114578747C70721500 @default.
• W2114578747 hasConceptScore W2114578747C86803240 @default.
• W2114578747 hasConceptScore W2114578747C89611455 @default.
• W2114578747 hasConceptScore W2114578747C95444343 @default.
• W2114578747 hasIssue "10" @default.

• next