(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Heterophilic and homophilic cadherin interactions in intestinal intermicrovillar links are species dependent

['Michelle E. Gray', 'Ohio State Biochemistry Program', 'The Ohio State University', 'Columbus', 'Ohio', 'United States Of America', 'Department Of Chemistry', 'Biochemistry', 'Zachary R. Johnson', 'Debadrita Modak']

Date: 2022-01 

Enterocytes are specialized epithelial cells lining the luminal surface of the small intestine that build densely packed arrays of microvilli known as brush borders. These microvilli drive nutrient absorption and are arranged in a hexagonal pattern maintained by intermicrovillar links formed by 2 nonclassical members of the cadherin superfamily of calcium-dependent cell adhesion proteins: protocadherin-24 (PCDH24, also known as CDHR2) and the mucin-like protocadherin (CDHR5). The extracellular domains of these proteins are involved in heterophilic and homophilic interactions important for intermicrovillar function, yet the structural determinants of these interactions remain unresolved. Here, we present X-ray crystal structures of the PCDH24 and CDHR5 extracellular tips and analyze their species-specific features relevant for adhesive interactions. In parallel, we use binding assays to identify the PCDH24 and CDHR5 domains involved in both heterophilic and homophilic adhesion for human and mouse proteins. Our results suggest that homophilic and heterophilic interactions involving PCDH24 and CDHR5 are species dependent with unique and distinct minimal adhesive units.

To better understand how PCDH24 and CDHR5 interact to form intermicrovillar links, we used a combination of structural and bead aggregation assays to characterize the adhesive properties and mechanisms of these cadherin family members. We present the X-ray crystallographic structures of Homo sapiens (hs) PCDH24 EC1-2 in 2 distinct forms, as well as of Mus musculus (mm) PCDH24 EC1-3 and hs CDHR5 EC1-2. The structures give insight into possible binding mechanisms, which are further probed by bead aggregation assays revealing that the human and mouse intermicrovillar cadherins do not engage in the same homophilic and heterophilic interactions. Based on these results, we suggest models for the PCDH24 homophilic complex and the PCDH24/CDHR5 heterophilic interaction utilized to form intermicrovillar links.

PCDH24 and CDHR5 are members of a large superfamily of proteins with extracellular domains that often mediate adhesion and that have contiguous and similar, but not identical, extracellular cadherin (EC) repeats. The linker regions between the EC repeats typically bind 3 calcium ions essential for adhesive function [ 22 – 25 ]. PCDH24 belongs to the Cr-2 subfamily and has 9 EC repeats, a membrane adjacent domain (MAD10), a transmembrane domain, and a C-terminal cytoplasmic domain [ 21 , 26 – 30 ]. CDHR5 belongs to the Cr-3 subfamily and has 4 EC repeats, a mucin-like domain (MLD), a transmembrane domain, and a C-terminal cytoplasmic domain [ 31 ]. Some human CDHR5 isoforms lack the MLD, which is not essential for the interaction with human PCDH24 [ 9 ]. Sequence alignments of the N-terminal repeats (EC1-3) suggest that PCDH24 and CDHR5 are similar to another pair of heterophilic interacting cadherins: Cadherin-23 (CDH23) and protocadherin-15 (PCDH15), the cadherins responsible for the formation of inner-ear tip links [ 32 – 36 ]. PCDH24 is most similar to CDH23, having the residues and elongated N-terminal β-strand that are predicted to favor the formation of an atypical calcium-binding site at its tip [ 33 , 34 ]. CDHR5 is most similar to PCDH15, and features cysteine residues predicted to form a disulfide bond at its tip [ 32 , 36 ]. In addition, a mutation in CDHR5, p.R84G (residue numbering throughout the text corresponds to processed proteins, see Methods ), which mimics a deafness-related mutation in PCDH15, interferes with the intermicrovillar links formed by PCDH24 and CDHR5 [ 9 ], suggesting that these proteins interact in a similar fashion to the heterophilic tip-link “handshake” formed by CDH23 and PCDH15 [ 32 , 36 , 37 ].

The extracellular parts of the IMAC stem from a pair of cellular adhesion proteins that belong to the cadherin superfamily, protocadherin-24 (PCDH24, also known as cadherin-related family member 2 [CDHR2]) and mucin-like protocadherin (CDHR5) [ 9 , 21 ]. These 2 nonclassical protocadherins form intermicrovillar links essential for brush border morphogenesis and function [ 9 , 10 ], stabilizing the microvillar hexagonal patterns as shown by freeze-etch electron microscopy and PCDH24 immunolabeling combined with transmission electron microscopy [ 9 ]. Mutations that impair the PCDH24 and CDHR5 interaction, studied in knockdowns of PCDH24 and CDHR5, cause a remarkable reduction in microvillar clustering ex vivo [ 9 ]. A PCDH24 knockout mouse model was found to be viable, but body weight of mutant mice was lower than wild type, and there were defects in the packing of the microvilli in the brush border [ 8 ]. Despite the important physiological role played by PCDH24 and CDHR5, little is known about the molecular details of how these proteins interact to form the intermicrovillar links.

Enterocytes are specialized epithelial cells lining the luminal surface of the small intestine and are fundamental players in nutrient absorption with an additional role in host defense [ 1 – 4 ]. Key to the aforementioned processes is the brush border, so named because the apical surface of the enterocytes is coated by thousands of microvilli of similar length and size organized in a hexagonal arrangement [ 5 ]. The organization and structure of the microvilli is maintained by a network of cytoplasmic and transmembrane proteins known as the intermicrovillar adhesion complex (IMAC), which includes proteins with extracellular adhesive domains and cytoplasmic parts tethered to the actin cytoskeleton that forms the interior of the microvilli [ 6 – 11 ]. Perturbations of brush border function are often associated with disease [ 12 – 15 ], and, not surprisingly, IMAC components are associated to foodborne diarrhea, chronic atrophic gastritis, and cholangiocarcinoma, while disruption of the IMAC complex results in intestinal dysfunction [ 9 , 16 – 20 ].

Results

PCDH24 and CDHR5 tip sequences are poorly conserved across species Intermicrovillar links formed by PCDH24 and CDHR5 in the enterocyte brush border are similar to inner-ear tip links formed by CDH23 and PCDH15. Both types of links are essential for the development, assembly, and function of actin-based structures (microvilli in the gut and stereocilia in the inner ear), and both are made of long nonclassical cadherin proteins that have similar cytoplasmic partners involved in interactions with the cytoskeleton and in signaling [10,38]. Sequence analyses have revealed similarities between the tips of PCDH24 and CDH23 and between the tips of CDHR5 and PCDH15 [32]. Given that CDH23 and PCDH15 engage in a heterophilic “handshake” complex involving their EC1-2 tips, it has been proposed that PCDH24 and CDHR5 might use a similar binding mechanism [9,32]. To further explore this hypothesis, we performed sequence analyses comparing the tips of these cadherins to each other, to classical cadherins, and across species. Interestingly, alignments of EC repeat sequences across 13 species for intermicrovillar and tip-link cadherins reveal poor conservation for PCDH24 and CDHR5 when compared to CDH23 and PCDH15. Average percent identity computed from these alignments is 50.7% for CDH23 EC repeats and 49.5% for PCDH15 repeats, with the N- and C-terminal ends being more conserved than the middle EC repeats [39,40]. In contrast, average percent identity is 16.6% for PCDH24 and 10.3% for CDHR5, considerably lower when compared to values for CDH23 and PCDH15. Moreover, the N- and C-terminal ends of PCDH24 and CDHR5 tend to be less conserved than the middle region of these proteins (Fig 1A, S1 and S2 Tables, and S1 and S2 Figs). Nevertheless, multiple sequence alignments of PCDH24 and CDH23 EC1 repeats confirm that PCDH24 has an elongated N-terminus that should facilitate the formation of calcium-binding site 0 as observed in structures of CDH23 (S3A Fig). Binding of calcium at this site in CDH23 is mediated by several acidic residues, which are also present and conserved in PCDH24. Similarly, sequence alignments of CDHR5 and PCDH15 EC1 repeats confirm that CDHR5 has the conserved cysteine residues that form a stabilizing disulfide bond at the tip of PCDH15 EC1 (S3B Fig). None of these proteins feature the tryptophan residues that mediate homophilic binding in classical cadherins [41–43], further supporting the hypothesis that PCDH24 and CDHR5 might form a tip-link-like handshake complex mediating adhesion. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Comparison of sequence conservation across species. (A) Percent identity across the extracellular domains of PCDH24 (EC1-MAD10), CDH23 (EC1-MAD28), CDHR5 (EC1-4), and PCDH15 (EC1-MAD12) plotted against EC repeat/MAD number. Overall, the cadherins of the inner-ear tip link, CDH23 and PCDH15, have higher identity across a variety of species, while the intermicrovillar-link cadherins, PCDH24 and CDHR5, have poor sequence conservation across species (S1 and S2 Tables and S1 and S2 Figs). N- and C-termini are more conserved in CDH23 and PCDH15 than the middle region of these proteins. An opposite trend is observed for PCDH24 and CDHR5. (B) Percent identity of the first 3 EC repeats of PCDH24, CDHR5, CDH23, PCDH15, CDH1, and CDH2 demonstrate that PCDH24 and CDHR5 sequences are not highly conserved across several species (Homo sapiens, hs; Mus musculus, mm; Gallus gallus, gg; Anolis carolinesis, ac; Danio rerio, dr). However, the N-terminal repeats of CDH23, PCDH15, CDH1, and CDH2 are highly conserved. Sequences were obtained from NCBI (S3 Table and Methods). CDH1, Cadherin-1; CDH2, Cadherin-2; CDH23, Cadherin-23; CDHR2, cadherin-related family member 2; CDHR5, cadherin-related family member 5; EC, extracellular cadherin; MAD, membrane adjacent domain; PCDH15, protocadherin-15; PCDH24, protocadherin-24. https://doi.org/10.1371/journal.pbio.3001463.g001 Pairwise sequence identity computed across 5 species for the N-terminal repeats (EC1-3) of PCDH24, CDHR5, CDH23, PCDH15, and the classical cadherins E-cadherin (CDH1) and N-cadherin (CDH2), also show different trends of sequence conservation (Fig 1B, S3 Table). Average percent identity in pairwise comparisons of EC1-3 sequences is high for CDH2 (89%), CDH23 (86%), and PCDH15 (82%); moderate for CDH1 (61%); and lowest for PCDH24 (49%) and CDHR5 (38%). This is evident when comparing sequences from the most evolutionary distant species (human and fish). For instance, hs and Danio rerio (dr) PCDH24 EC1-3 sequences are 39% identical, while hs and dr CDH1 EC1-3 sequences are 54% identical. Sequence differences are still large for the human and mouse PCDH24 and CDHR5 protein tips, with 75% identity for the hs and mm PCDH24 EC1-3 sequences and 66% identity for the hs and mm CDHR5 EC1-3 sequences (compared to CDH1, CDH2, CDH23, and PCDH15 EC1-3 sequences with hs and mm pairwise identities between 83% and 98%) (Fig 1B). Proteins with sequence identity as low as 30% can share similar folds [44], yet the low PCDH24 and CDHR5 sequence conservation in multiple sequence alignments and pairwise comparisons suggests that details of their structures and function differ across species.

Crystal contacts in human and mouse PCDH24 structures suggest distinct adhesive interfaces Crystallographic contacts observed in cadherin structures have revealed multiple physiologically relevant interfaces [36,40,45,46,48,49,57]. The crystallographic packings observed for the hs and mm PCDH24 structures show various interfaces that further highlight differences across species. The asymmetric unit of the hs PCDH24 EC1-2 I structure shows 4 molecules with 2 similar antiparallel trans dimers (S8A Fig) arranged perpendicular to one another to form a dimer of dimers (S8B Fig). Each antiparallel dimer positions EC1 from 1 monomer in front of EC2 from the next monomer, slightly shifted with respect to each other and with the extended FG β-hairpins near the EC1-2 linker regions mediating the interaction head-to-head. The antiparallel dimers have interface areas of 979 Å2 and 939 Å2 for monomers A:D and B:C, respectively. These values are both larger than an empirical threshold (856 Å2) that distinguishes biologically relevant interactions from crystallographic packing artifacts [58]. Two other interfaces are small and unlikely to be biologically relevant (S8C and S8D Fig). Previous data have shown that hs PCDH24 mediates homophilic trans adhesion when hs CDHR5 is not present [9]. It is possible that the antiparallel EC1-2 interface seen in the hs PCDH24 EC1-2 I structure facilitates homophilic adhesion mediated by hs PCDH24, yet our SMD simulations suggest that this interface is weak (S9A Fig, S9 Data). Interestingly, the hs PCDH24 EC1-2 II structure also has 4 molecules in the asymmetric unit, but these arrange differently and form a distinct set of interfaces. Remarkably, there are 2 sets of potential antiparallel trans dimers (S10 Fig). In the first one, the EC1 repeat from 1 monomer is in front of EC2 from the next monomer, slightly shifted and with glycosylation sugars at p.N9 residues and the extended FG β-hairpins near the EC1-2 linker regions all pointing away from the interface (S10A Fig). These antiparallel dimers have interface areas of 1,023 Å2 and 1,019 Å2 for monomers A:B and C:D, respectively, and feature an overlap of aromatic rings contributed by p.Y67 and p.Y71 residues that might be critical for binding. In the second set of antiparallel trans dimers, the EC1 repeat from 1 monomer is also positioned in front of EC2 from the next monomer, but at an angle that reduces their contacts, gives space for sugars stemming from p.N9, and that favors overlap of the extended FG β-hairpins near the EC1-2 linker regions (S10B Fig). These antiparallel dimers have interface areas of 418 Å2 and 436 Å2 for monomers A:C and B:D, respectively. The 2 sets of antiparallel trans dimers in the hs PCDH24 EC1-2 II structure come together to form an asymmetric unit in which various other smaller interfaces are observed (S10C–S10G Fig). It is possible that the largest antiparallel EC1-2 interface seen in the hs PCDH24 EC1-2 II structure (A:B and C:D) facilitates homophilic adhesion mediated by hs PCDH24. Predictions from SMD simulations show that this interface is stronger than the antiparallel interface observed in hs PCDH24 EC1-2 I when probed under the same conditions (S9C Fig, S10 Data), thus supporting the possible physiological relevance of the hs PCDH24 EC1-2 II arrangement. In contrast to the crystal contacts observed in the hs PCDH24 EC1-2 I and II structures, the crystal packing of mm PCDH24 EC1-3 (space group P321) results in multiple large interfaces generated from rotations of a single molecule in the asymmetric unit. A parallel trimer is present around the axis of symmetry, with the EC1 N-termini from each protomer protruding away and inserting into pockets of adjacent EC1 repeats in a clockwise fashion, thereby interlocking the EC1 domains (S11A and S11B Fig). This is possible because of the open conformation observed for the mouse EC1 N-terminus without a bound calcium ion at site 0, and it is reminiscent of how the EC1 N-terminus of classical cadherins is exchanged resulting in the insertion of W2 into an hydrophobic binding pocket of the partner molecule to form a trans dimer (S7C and S7D Fig) [41,42]. Interestingly, a glycosylation site is predicted at p.N9, and sugars protruding from this site, as observed in the hs PCDH24 EC1-2 II structure, may interfere with or regulate the formation of interlocking EC1 repeats, as has been observed for other cadherins [57]. The interlocking of mouse EC1 repeats in the trimer facilitates the formation of 2 parallel cis interfaces between EC1-3 protomers involving the FG β-hairpin interacting with the neighboring EC2-3 linker. These cis interfaces are large with an interface area of 1,352.6 Å2 between 2 monomers (S11A Fig). In addition, a fully overlapped antiparallel trans interface is observed between EC1-3 protomers with an area of 1,221.2 Å2 (S11C Fig). Sugars at a second predicted glycosylation site located in the center of the cis trimer (p.N161) and at a third predicted glycosylation site near the trans interface between EC1 and EC3 (p.N280) may interfere with or regulate the formation of these interfaces. Together, the cis trimer and trans dimer could form a glycosylation-modulated interlocking arrangement of PCDH24 molecules that is different from the trans interactions seen in the hs PCDH24 EC1-2 I and II structures. Sequence conservation mapped to the mm PCDH24 EC1-3 structure reveals that only core residues are highly conserved, highlighting poor conservation for surface exposed residues potentially involved in adhesive interactions (S12 Fig). Motivated by the sequence and structural differences observed for the human and mouse PCDH24 tips, we used bead aggregation assays to determine the minimum adhesive unit of PCDH24 homophilic binding in both species.

[END]

[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001463

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/