Open conformation of tetraspanins shapes interaction partner networks on cell membranes

Article16 August 2020free access Open conformation of tetraspanins shapes interaction partner networks on cell membranes Yihu Yang Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Xiaoran Roger Liu orcid.org/0000-0003-3198-1588 Department of Chemistry, Washington University, St. Louis, MO, USA Search for more papers by this author Zev J Greenberg Division of Hematology and Oncology, Department of Pediatrics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Fengbo Zhou Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Peng He Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Lingling Fan Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Shixuan Liu orcid.org/0000-0003-3886-144X Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Guomin Shen orcid.org/0000-0002-5006-2790 Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Takeshi Egawa Department of Pediatrics Pathology and Immunology, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Michael L Gross Department of Chemistry, Washington University, St. Louis, MO, USA Search for more papers by this author Laura G Schuettpelz Division of Hematology and Oncology, Department of Pediatrics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Weikai Li Corresponding Author [email protected] orcid.org/0000-0002-8711-1904 Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Yihu Yang Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Xiaoran Roger Liu orcid.org/0000-0003-3198-1588 Department of Chemistry, Washington University, St. Louis, MO, USA Search for more papers by this author Zev J Greenberg Division of Hematology and Oncology, Department of Pediatrics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Fengbo Zhou Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Peng He Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Lingling Fan Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Shixuan Liu orcid.org/0000-0003-3886-144X Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Guomin Shen orcid.org/0000-0002-5006-2790 Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Takeshi Egawa Department of Pediatrics Pathology and Immunology, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Michael L Gross Department of Chemistry, Washington University, St. Louis, MO, USA Search for more papers by this author Laura G Schuettpelz Division of Hematology and Oncology, Department of Pediatrics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Weikai Li Corresponding Author [email protected] orcid.org/0000-0002-8711-1904 Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA Search for more papers by this author Author Information Yihu Yang1, Xiaoran Roger Liu2, Zev J Greenberg3, Fengbo Zhou1, Peng He1, Lingling Fan1, Shixuan Liu1, Guomin Shen1, Takeshi Egawa4, Michael L Gross2, Laura G Schuettpelz3 and Weikai Li *,1 1Department of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, USA 2Department of Chemistry, Washington University, St. Louis, MO, USA 3Division of Hematology and Oncology, Department of Pediatrics, Washington University School of Medicine, St. Louis, MO, USA 4Department of Pediatrics Pathology and Immunology, Washington University School of Medicine, St. Louis, MO, USA *Corresponding author. Tel: +1 314 362 8687; E-mail: [email protected] EMBO J (2020)39:e105246https://doi.org/10.15252/embj.2020105246 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 Abstract Tetraspanins, including CD53 and CD81, regulate a multitude of cellular processes through organizing an interaction network on cell membranes. Here, we report the crystal structure of CD53 in an open conformation poised for partner interaction. The large extracellular domain (EC2) of CD53 protrudes away from the membrane surface and exposes a variable region, which is identified by hydrogen–deuterium exchange as the common interface for CD53 and CD81 to bind partners. The EC2 orientation in CD53 is supported by an extracellular loop (EC1). At the closed conformation of CD81, however, EC2 disengages from EC1 and rotates toward the membrane, thereby preventing partner interaction. Structural simulation shows that EC1-EC2 interaction also supports the open conformation of CD81. Disrupting this interaction in CD81 impairs the accurate glycosylation of its CD19 partner, the target for leukemia immunotherapies. Moreover, EC1 mutations in CD53 prevent the chemotaxis of pre-B cells toward a chemokine that supports B-cell trafficking and homing within the bone marrow, a major CD53 function identified here. Overall, an open conformation is required for tetraspanin–partner interactions to support myriad cellular processes. Synopsis Crystal structure of CD53 captures an open conformation for partner interaction and reveals the structural regulation mechanism for tetraspanins to organize interaction networks on cell membranes and support myriad cellular processes. Crystal structure of CD53 captures an open conformation poised for partner interaction. This open conformation is supported by a stable EC1-EC2 interaction. Hydrogen-deuterium exchange experiments show that CD53 and CD81 use their C-D variable region in the EC2 domain to bind partners. This variable region approaches the membrane plane in the closed conformation, and through EC2 rotation in the open conformation, the above-membrane height of this region is raised to permit partner interaction. Partner specificity of tetraspanins is achieved by the various structures and sequences of the C-D region, by its above-membrane height that matches with interacting partners, and by interactions through the transmembrane domains. CD53 assists the migration of pre-B cells towards the CXCL12 chemokine, whereas disrupting the EC1-EC2 interaction abolishes this migration. Similar EC1 mutations in CD81 impairs the accurate glycosylation of CD19. Introduction Tetraspanins participate in a wide array of cellular processes including cell proliferation, migration, adhesion, fusion, and signaling (Termini & Gillette, 2017; Yeung et al, 2018). There are 33 tetraspanin members identified in the human genome. The well-studied CD81 is required for sperm–egg fusion during the fertilization process (Sutovsky, 2009), for cell composition regulation in the central nervous system (Geisert et al, 2002; Charrin et al, 2009), and for maturation of CD19 (Maecker & Levy, 1997; Shoham et al, 2003, 2006; Bagashev et al, 2018), a major immunotherapy target against lymphoma and leukemia (Park et al, 2016; Braig et al, 2017; Velasquez & Gottschalk, 2017). Most tetraspanins, including CD81, are widely distributed in cells and tissues. As an exception, CD53 is expressed exclusively in leukocytes. CD53 is involved in B-cell differentiation and modulation of inflammatory responses (Beckwith et al, 2015; Zuidscherwoude et al, 2017; Greenberg et al, 2020). Familial deficiency of CD53 expression manifests as recurrent infections (Mollinedo et al, 1997), whereas increased CD53 expression correlates with the active status of tuberculosis patients (Omae et al, 2017). Overall, these examples highlight the essential functions of tetraspanins under physiological and pathological conditions. Tetraspanins facilitate these cellular processes through organizing molecules at the plasma membrane, thereby generating a hierarchical interaction network (Charrin et al, 2014). Tetraspanins interact with each other and with a large repertoire of partners, many of which contain immunoglobulin-like (Ig) domains. The partner interactions are primarily through a large extracellular domain (EC2), which contains a variable region that may control the partner-interaction specificity such as CD81 with CD19 (Shoham et al, 2006) and CD53 with CD2 (Bell et al, 1992). Owing to this functional importance, the structural patterns of EC2 have been used to classify tetraspanins (Seigneuret et al, 2001). A small extracellular loop (EC1) in tetraspanins is also highly variable, but its functional role remains elusive (Van Deventer et al, 2017). Among the proteins with four transmembrane helices (TM), tetraspanins are distinguished by the unequal size of EC1 and EC2 and by the cysteine pattern of EC2, which contains two to four cysteine pairs including a CCG motif (Huang et al, 2005). The transmembrane region of tetraspanins is also involved in partner interactions; for example, TM1 of CD81 supports the exit of CD19 from the endoplasmic reticulum (Shoham et al, 2006). The two termini of tetraspanins are located at the cytosolic side, providing additional interaction sites for signaling molecules (Van Deventer et al, 2017). As an example, the N-terminal of CD53 interacts with PKC-β to regulate B-cell signaling (Zuidscherwoude et al, 2017). Taken together, multiple structural regions of tetraspanins are required to organize their partner interactions (Van Deventer et al, 2017). The molecular nature of these organizations and their regulation of cellular responses, however, remain largely unknown. In the entire tetraspanin superfamily, the full-length structure is known only for human CD81 and CD9 (reported during the submission of this manuscript), both captured in a closed conformation incapable of partner interactions (Zimmerman et al, 2016; Umeda et al, 2020). The four TMs of CD81 and CD9 form a cone-shaped structure that splits into two TM pairs at the extracellular side. The EC2 domain of CD81 adopts a mushroom-like structure with five helices (named A to E); the A and E helices form the mushroom “stalk”, and B and variable C–D region form the “head”. The isolated EC2 domain of CD81 shows essentially the same structure (Kitadokoro et al, 2001), and the mushroom fold is observed also in the CD9 structure and in the NMR structure of the EC2 domain of another tetraspanin, TSP-2, from Schistosoma mansoni (Jia et al, 2014). The EC1 loop, the loop connecting TM2 and TM3 (L2–3), and part of the N- and C-termini, however, are disordered in the full-length CD81 structure, the resolution of which is compromised by anisotropic crystal diffraction (5.5 and 2.95 Å). Nevertheless, the electron densities of this structure suggest that a cholesterol molecule is bound at the transmembrane region, and molecular dynamics simulation implies that cholesterol dissociation changes the structure of the transmembrane domain to induce the open conformation. Despite such modeling attempts (Zimmerman et al, 2016; Umeda et al, 2020), a tetraspanin structure at the partner-interactive state is required to understand the molecular mechanism regulating the tetraspanin network. Here, we report the 2.9 Å structure of CD53 in the open conformation. Unexpectedly, the transmembrane domain of CD53 without bound cholesterol adopts essentially the same structure as the closed conformation of CD81. EC2 of CD53, however, rotates to a different angle owing to supporting interactions from EC1. After the rotation, the EC2 variable region is exposed for partner binding at the open conformation. In contrast, this region faces the membrane surface and is partially buried in the closed conformation of CD81. Hydrogen–deuterium exchange mass spectrometry (HDX-MS) shows that CD53 and CD81 use a similar part of their variable region to interact with extracellular domains of their partners, CD2 and CD19, respectively. Mutations that disrupt the EC1-EC2 interface of CD81 interfere with its ability to promote accurate glycosylation of CD19. Similar mutations in CD53 interfere with B-cell migration, which we identify to be a major cellular function of CD53. Thus, the open and closed conformational changes may regulate partner interactions of tetraspanins in a wide range of cellular processes. Results CD53 structure Human CD53 was crystallized in lipid cubic phase (LCP), and the crystals diffracted isotopically to 2.9 Å (Table 1) with a construct carrying an Asn148Ala mutation to eliminate a predicted glycosylation site, and Cys80Ala and Cys208Ala mutations to prevent palmitoylation. The electron density map reveals nearly the entire structure of CD53 including the loop and termini regions (Fig EV1A). The CD53 molecules pack artificially into an antiparallel dimer in each asymmetric unit of the crystals (Fig EV1B); these two molecules face different packing interactions and yet their structures, including the termini tails, are nearly identical (Fig EV1C). Table 1. Data collection, phasing, and refinement statistics Protein CD53 Data collection aa The dataset was derived from a single crystal. Space group P21 Solvent content (%) 67.0 Cell dimensions a, b, c (Å) 49.4, 210.5, 73.4 α, β, γ (°) 90, 100.2, 90 Resolution (Å) 50–2.9 (2.95–2.90)bb Values in parentheses are for the highest-resolution shell. CC 1/2 cc CC1/2 = Pearson correlation coefficient between random half-datasets. 0.998 (0.772) R sym 0.128 (0.738) I/σI 8.4 (1.4) Completeness (%) 93.3 (97.2) Redundancy 2.8 (2.8) Refinement Resolution (Å) 50–2.9 No. reflections 24,301 Rwork/Rfreedd Rwork = ∑h||Fo(h)|-|Fc(h)||/∑h|Fo(h)||, where |Fo| and |Fc| are the observed and calculated structure-factor amplitudes, respectively. Rfree was calculated with 5% of the data excluded from the refinement. 22.3/26.5 No. atoms Protein 6,915 Ligand/ion 151 Water 19 B-factors (Å2) Protein 60.9 Ligand/ion 55.4 Water 61.1 Ramachandran plot Favored (%) 97.4 Allowed (%) 2.6 Outliers (%) 0.0 R.m.s deviations Bond lengths (Å) 0.006 Bond angles (°) 0.952 a The dataset was derived from a single crystal. b Values in parentheses are for the highest-resolution shell. c CC1/2 = Pearson correlation coefficient between random half-datasets. d Rwork = ∑h||Fo(h)|-|Fc(h)||/∑h|Fo(h)||, where |Fo| and |Fc| are the observed and calculated structure-factor amplitudes, respectively. Rfree was calculated with 5% of the data excluded from the refinement. Click here to expand this figure. Figure EV1. Electron density map and crystal packing Electron density map contoured at 1 σ. This density modification map is calculated with the partial phases from the molecular replacement model of fused GFP, and therefore, the CD53 part of the map is unbiased by the model. Inset, zoomed view of monoolein in the density map. CD53 molecules form an antiparallel dimer in the asymmetric unit. Superimposition of the two molecules in the antiparallel dimer. Download figure Download PowerPoint The four TMs of CD53 adopt a cone-shaped conformation (Fig 1A and B). At the cytoplasmic side, the TMs pack tightly into a bundle at an angle to each other. The packing primarily uses knob-and-hole interactions from the side chains in the TMs (Fig 1C), the region that human tetraspanins share high sequence similarity (Fig EV2). About halfway in the membrane, the TM1/TM2 and TM3/TM4 pairs start to split, creating a tunnel between TM1 and TM4 that is delineated by several hydrophobic residues (Fig 1B). Elongated electron densities were observed in this tunnel (Fig EV1A) and assigned to monoolein, a lipid used for the LCP crystallization. Further splitting of the TM pairs at the extracellular side leaves a large opening (Fig 1B), which affords a potential site designated for binding the transmembrane domain of interacting partners. Figure 1. Structure of CD53 in the open conformation Overall structure (side view). EC1 and secondary structure elements in EC2 are shown in different colors, and the four TMs in blue. Electrostatic surface representation of the structure (back view from A). The bundle region of TMs forms a tunnel that binds monoolein, and their splitting generates a lateral opening inside the membrane. Van der Waals surface (top view) showing interactions between the TM1/2 and TM3/4 pairs near their splitting site and residues forming the lipid-binding tunnel. Structure of the CD53 EC2 domain showing the mushroom fold with a head and stalk. The stalk helices connect to TM3 and TM4 through two hinges. EC1 loop binds in the EC2 groove (electrostatic surface) formed by helices A and B. Interactions stabilizing the EC1-EC2 binding. Download figure Download PowerPoint Click here to expand this figure. Figure EV2. Sequence conservation between human tetraspanins Conservation plot of 33 human tetraspanins. The similarity plot is generated by Plotcon, which is based on multiple sequence alignment from Clustal Omega. The secondary structures are indicated above. Heat presentation of the sequence similarity mapped onto the CD53 structure. Red color indicates high sequence conservation. Sequence alignment of CD53 and CD81. Identical residues are shaded in dark gray and similar residues in light gray. Secondary structures are indicated above. The dashed boxes show that CD53 lacks the C and D helices found in CD81. The red box indicates the small region of hydrophobic residues in EC1 that are involved in EC2 interaction. Download figure Download PowerPoint The EC2 domain of CD53 adopts a canonical mushroom fold (Fig 1D). The “stalk” helices A and E are connected to TM3 and TM4, respectively. Both connections are through short hinges and with sharp bends. The two hinges are associated with each other through hydrophobic residues (Fig 2D). The interactions continue between the interface of helices A and E, involving stacking between hydrophobic residues and a hydrogen bonding network formed between polar residues. These extensive interactions stabilize the relative positions of the two stalk helices. CD53 lacks the helices C and D found in CD81 (Kitadokoro et al, 2001). This C–D region instead folds into two antiparallel loops, whose conformation is stabilized by the two disulfide bonds characteristic of tetraspanins. The EC2 head region, in particular the C–D region, is well exposed for interactions in the open conformation (Fig 1A). Figure 2. Structural comparison of CD53 in open conformation and CD81 and CD9 in closed conformation Structures of CD53 (colored by structure elements) and CD81 (gray) superimposed by their transmembrane domains. The curved arrow indicates EC2 rotation. Letters with a prime sign (A′–E′) indicate structural components of CD81. Comparison of individual structures of CD53 in open conformation (left), and CD81 (middle) and CD9 in closed conformation (right). The dashed curve indicates disordered EC1 in CD81, and the angles indicate bending of TM3 in CD81 and CD9. The bound cholesterol (Chol; orange) in CD81 and bound monoolein (Mono; blue) in CD53 and CD9 are indicated. CD9 structure contains palmitoylation sites (Palm, green) and another monoolein bound between the TM bundle. Since this monoolein molecule is not modeled in PDB 6K4J, we can only show surrounding residues here. Among these, Leu14, Asn18, and Phe21 are conserved between CD9, CD81, and CD53. EC2 domain structure of CD53 superimposed with those of CD81, CD9, and TSP2. The comparisons show the same position of two characteristic disulfide bonds and show differences in the C–D conformation and A-helix orientation. Interactions within the EC2 domains of CD53, CD81, CD9, and TSP-2. Relative orientation of the stalk helices and bending at the hinges are stabilized by extensive side-chain interactions. Folding of the head region is maintained by the two characteristic disulfide bonds. Download figure Download PowerPoint This exposed orientation of EC2 is supported by its interaction with EC1. The central part of EC1 (residues 43–47) binds in a small groove of EC2 that is formed between helices A and B (Fig 1E). The side chains of Leu43 and Phe44 insert into the hydrophobic part of groove, and other interactions involve Phe40, His45, and backbone carbonyls of EC1 (Fig 1F). Owing to these stable binding interactions, EC1 is well ordered in the CD53 structure. With the support from EC1, EC2 is disengaged from the transmembrane domain (Fig 1E). Consequently, the head region of EC2 moves away from the membrane and sticks out into the extracellular space (Fig 1A). Thus, the head region is nearly free from steric hindrance and physically poised for partner interactions. At the cytoplasmic side of the CD53 molecule, an interaction “cage” is formed between the loop connecting TM2 and TM3 (L2–3) and the N- and C-terminal tails (Fig EV3). Glu77 in L2–3 forms salt bridges with Lys7 and Lys10 at the N-terminus. These highly conserved Lys residues have been identified as ubiquitination sites in CD81, CD151, and TSPAN6 (Termini & Gillette, 2017), and their structural organization may have regulative roles. Click here to expand this figure. Figure EV3. Interactions at the intracellular side of CD53Residues from the four-helix bundle and connecting loops form an interaction cluster. Download figure Download PowerPoint Open and closed conformations Structural comparison of CD53 and CD81 shows a large rotation between their EC2 and transmembrane domains (Fig 2A). With support by interactions from EC1, the EC2 head region of CD53 is twisted away from the membrane and exposed for interactions. In contrast, the EC1 loop of CD81 is disordered in the structure, and its EC2 instead interacts with TM1 and TM2 (Fig 2B). Similarly, EC2 interacts with the TM regions in the CD9 structure; although the EC1 loop is ordered in this structure, it flips to another side without forming interaction with EC2 (Fig 2B). Without the EC1 support, EC2 rotates to a different angle with its head region oriented toward the membrane, thereby generating a closed conformation in the CD9 and CD81 structures. The relative rotation between EC2 and transmembrane domains can be regarded as a rigid-body movement because each individual domain adopts a similar conformation. In particular, the transmembrane domains of CD53 and CD81 are nearly superimposable (Fig 2A), with a 1.2 Å RMSD between their Cα atoms. The same cone-shaped conformation suggests that this is a canonical fold shared by all tetraspanins in their transmembrane domain. Despite the overall structural similarity, the TM2 of CD81 is one turn longer than that of CD53. Moreover, the TM3 of CD81 is bent by approximately 18° at the extracellular end, whereas the TM3 of CD53 is straight (Fig 2B). The CD9 structure shows a similar bending in TM3. The different conformations of TM3 are probably associated with the relative movement of EC2. The transmembrane domains of these tetraspanins, at essentially the same conformation, form different lipid-binding sites (Fig 2B). In CD53, monoolein is bound in a tunnel that is formed by hydrophobic residues at the bottom part of the TM splitting. In CD9, a monoolein molecule is bound at a similar position that is surrounded by residues conserved between CD9, CD53 and CD81. In contrast, cholesterol is bound at an open pocket in CD81 and occupies a much larger area. At the bottom part of the TM splitting, the cholesterol hydroxyl group form hydrogen bonds with Asn18 in TM1 and Glu219 in TM4, and the cholesterol ring interacts with hydrophobic residues from all four TMs of CD81. After the TMs are split, however, the other part of the cholesterol molecule interacts only along TM2. Importantly, the cholesterol binding after the splitting does not involve TM3 and TM4. Given that EC2 connects only to TM3 and TM4, it is unlikely that binding or dissociation of cholesterol induces EC2 rotation. Instead, the role of the lipids is probably to stabilize folding of the transmembrane domain. The EC2 domains of different tetraspanins share the overall mushroom fold but with large structural deviations (Fig 2C). The C–D region of CD81 forms two helices in crystal structures, and the molecular shape of the head region differs from that of CD53 head. In CD9, the C region maintains a helical conformation, but the D region becomes a loop. On the other hand, the C–D regions of CD53 and TSP-2 form two loops that swing to different angles. These structural variations in the C–D region, along with sequence differences, may afford the specificity of partner recognition for different tetraspanins. Folding of these variable regions is reinforced by two disulfide bonds conserved in the tetraspanin family. Remarkably, these disulfides are nearly superimposable in all four structures (Fig 2C and D). In addition, the E helix and at least part of the B helix can be closely superimposed. Helix A adopts different angles in these structures, however, owing to its different but extensive side-chain interactions with B and E (Fig 2D). Given that the sequence of helix A is poorly conserved (Fig EV2), this helix may change angles in different tetraspanins and alter their EC2 orientation in both open and closed conformations. Mapping of tetraspanin–partner binding interfaces To understand how the open and closed conformations affect the partner interactions of tetraspanins, we mapped the partner-binding interfaces of CD53 and CD81 by using HDX-MS. CD53 physically associates with CD2 (Bell et al, 1992), a membrane-anchored NK- and T-cell marker that contains two extracellular Ig domains. The association is likely through the EC2 domain of CD53 and two Ig domains of CD2, similar to other tetraspanin–partner interactions (Van Deventer et al, 2017). To characterize the binding interface, we purified the CD53 EC2 domain and CD2 Ig domains and conducted HDX-MS for the unbound and bound proteins. Comparison of the HDX kinetic curves shows that CD53 EC2 domain and CD2 Ig domains each have a single region that, upon binding, exhibits significant decreases in deuterium uptake with cumulative differences greater than three times of propagated error (99.7% confidence that differences are significant) (Figs 3A and B, and EV4A and B). In CD53 EC2, the protection seen in HDX corresponds to residues 153WTSGPPASCPSD164 in the variable C–D region (Fig 3A). The protection in CD2 Ig domains is observed for two peptides (Fig EV4B), and their overlap indicates a binding region of 124LKIQE128 that is located at the joint between the two Ig domains (Fig 3C). The protection is observed to be nearly constant throughout the entire time of HDX (30 s to 4 h), suggesting a slow off rate for EC2 and Ig domains. Overall, HDX-MS reveals that the C–D region of CD53 binds to the Ig joint region of CD2 (Fig 3C). Figure 3. CD53-CD2 and CD81-CD19 partner interactions identified by HDX-MS and modeled onto open and closed conformations The partner-binding regions of CD53 and CD81. Their EC2 sequences are aligned. Identical residues are shaded in dark gray, similar residues in light gray, and partner-binding regions identified by HDX-MS in orange. Secondary structures are indicated above. Representative HDX-MS curves showing the protected regions in CD53 and CD2. Other curves at these regions are shown in Fig EV4A and B, and complete HDX-MS results in Fig 3 Source Data. The HDX experiments were performed (manually or with a LEAP automation system) in triplicate. Error bars show the standard deviation. Left, CD53-CD2 docking model. The HDX-MS identified regions are shown as a red rod in CD53 and as a blue arrow in CD2. The black box indicates the CD2 transmembrane region, and the dashed line indicates the linker to Ig domains. Right, CD53-CD2 docking model with CD53 modeled in the closed conformation. The modeling was based on the CD81 structure, in which EC2 reorients to interact with the TM domain. Docking of CD19 Ig to CD53 EC2 is in the same manner as Left. Representative HDX-MS curves showing the protected regions in CD81 and CD19. Other curves at these regions are