In most immunology textbooks, TCR- and BCR complexes are depicted as symmetric structures with a signaling heterodimer placed on each side of the ligand-binding parts. Based on the fact that the mIgM molecule is a symmetric homodimer, we initially also proposed a symmetric model for the IgM-class BCR (IgM-BCR) complex, in which the mIgM molecule binds two CD79a/CD79b heterodimers (13). However, biochemical analysis has shown that the mIgM molecule and the CD79a/CD79b heterodimer form a 1:1 complex rather than a 1:2 complex (14). The now resolved cryo-EM structures confirm the 1:1 model of the IgM-BCR complex and show an asymmetric assembly between the ligand-binding and signaling part of these antigen receptors (15–17). Specifically, in the murine IgM-BCR structure the CD79a/CD79b heterodimer predominantly interacts with only one of the two heavy chains (μHC and μHC’) of the mIgM molecule (17). The monomeric αβTCR also shows an asymmetric organization with the CD3 complex occupying only one side of the αβTCR (18). Interestingly, the cryo-EM structures of the pMHC-bound or free αβTCR do not reveal major alterations upon ligand binding (19). Coupling between the antigen receptor components occurs via three distinct contact sites, namely the membrane proximal Ig domains, the connecting peptides and the TMDs, with the latter two playing a dominant role in complex formation. The TMDs of all antigen receptor components are single-spanning alpha-helixes, covering a space within the membrane that has been previously underestimated and drawn much too small in many schematic representations of antigen receptor organization. Thus, only the recent cryo-EM structures highlight the importance of the TMDs for the formation and stability of antigen receptor complexes.

A major question arising from the known cryo-EM structures is why the basic structure of the antigen receptor complexes is asymmetric. A possible answer to this question is that the asymmetric organization allows these receptors to interact on the lymphocyte membrane either with themselves (forming dimers and oligomers) or with coreceptor modules that regulate or amplify the signal transduction of the antigen receptors. Following this idea, I took a closer look at the conserved TMD AAs that are either engaged in the formation of the 4-alpha-helical bundle of the IgM-BCR TMDs or exposed to the lipid bilayer. The TMDs of the CD79a and CD79b signaling component have different positions in the IgM-BCR structure. The CD79a TMD binds on one side of the alpha helix to the CD79b TMD and on the other side to the TMDs of the μHC and μHC’ ( Figure 1A ). Indeed, it is predominantly CD79a that couples the CD79a/CD79b heterodimer to the mIgM molecule. The CD79b TMD is only engaged on one side in the binding to the CD79a TMD and, to a lesser extent, to the μHC’ ( Figure 1 ). The exposed, free side of the CD79b TMD carries either a leucine or isoleucine residue at the TMD position 8, 11, 15, 18, 22, which are part of a leucine zipper (LZ) heptad motif involved in the interaction of different TMD alpha-helixes (20). Within a heptad (abcdefg) motif, the a- and d-position residues form the hydrophobic core of the interface between the interacting alpha helices.

From our previous studies of the BCR conformation and interaction on the B cell membrane using the Fab-based proximity ligation assays (Fab-PLA), we learned that the IgM-BCR established a close contact with the CD19 coreceptor upon B cell activation (6). A look at the AAs composition of the human CD19 TMD revealed a series of 5 leucine residues located on one side of the CD19 TMD alpha helix at the same position as the LZ components of the CD79b TMD ( Figure 1A ). It is thus likely that the LZ structure within the CD19 TMD is involved in the heterodimerization of the IgM-BCR with the CD19 coreceptor. Interestingly, in our Fab-PLA studies, we detected the IgM-BCR/CD19 interaction only on activated but not on resting B cells. The reason for this restriction may be the finding that CD19 is only present on the B cell surface in association with the CD81 protein (21). Based on a recent cryo-EM study, a model of the CD19/CD81 complex can be drawn. Although the present CD19/CD81 cryo-EM is of low resolution, it shows that the LZ structure within the CD19 TMD is covered by one of the TMD alpha-helixes of CD81 and one can predict that the LZ residues are not available for binding to the CD79b TMD (22). However, B cell activation appears to be accompanied by the dissociation of the CD19/CD81 complex, releasing CD19 for the potential TMD interaction with the IgM-BCR complex (23). As explained above, the association between the ligand-binding and the signaling subunits of the IgM-BCR involves the extracellular Ig domains and the connecting peptides in addition to the TMDs. Thus, it is feasible that not only the LZ structure within the TMDs but also additional sites take part in the CD79b/CD19 complex formation. One of these sites seems to be the juxtamembrane region which contains a series of complementary positively (H23,Q25,R26) and negatively (D23,D25,D26) charged AAs in the CD19 and CD79b AA sequence, respectively ( Figure 1A ). An alignment of the human TMD sequences of CD79b and CD19 according to the found LZ structure shows that these two sequences carry 9 identical or similar AAs ( Table 1 ). Thus, I could define an LZ-containing A-motif: TLxx(LI)(LI)xx(LI)xx(LI)xx(LI)L, which is shared by both proteins and that may play an important role in the organization of antigen receptors and their coupling to lateral interactors. For this and related LZ motifs (see below), I propose the name “immunoreceptor organization and coupling motifs” (ICOMs).

A prosite motif search of the TMD + 5 AA flanking sequences of 2417 single-pass transmembrane proteins extracted from the human uniprot database (TMD5) for the defined A-motif yielded only 6 hits ( Table 1 ). In addition to CD79b and CD19, the A-motif is found in the TMD sequence of the erythropoietin receptor (EpoR), the insulin growth factor-like receptor 1 (IGFLR1), the macrophage stimulating 1 receptor (MST1R), also known as CD136 or RON, and the transmembrane proteins TMEM191 of unknown function. The result of the A-motif search was verified by a TMD sequence comparison according to the 5 positions of the heptad LZ motif ( Table 2 ). A problem that arises with such a comparison is that the start of a TMD sequence is not always well defined and that the length of this sequence can differ between 19-24 AAs. In my TMD sequence comparison, I aim to place the first heptad repeat AA at (or near) the position 8 and the last at position 22, which should be preferentially followed in the next 1-4 positions by a charged or polar AA. With such an alignment the 5 positions of the heptad repeat are occupying a central space within the alpha-helical barrel representation of a TMD sequence ( Figure 1B ). Of the 6 hits from the A-motif search, only the MST1R TMD caused a problem and allowed alternative ways of LZ motif sequence alignment. One should keep in mind that an LZ sequence is a rather flexible dimerization structure that is not restricted to leucine or isoleucine residues. Indeed, 1 or 2 of the 5 heptad repeat positions of an LZ structure could carry other AAs than leucine or isoleucine, such as hydrophobic or polar, without compromising the dimerization function (24–26). Cells appear to exploit this binding flexibility to either increase or decrease the affinity of an ICOM : ICOM interaction (see below).

Next, I conducted a prosite search for a leucine-only LZ motif (B motif): Lx(2)Lx(3)Lx(2)Lx(3)L found in the CD19 TMD. This search yielded 32 hits, including several adhesion proteins such as cadherins (CDH, CDH1, CDH3), desmogleins (DSG, DSG2, DSG4) and the integrin ITGB4. It is thus feasible that an ICOM-based interaction between adhesion and antigen receptors is involved in the activation of lymphocytes in contact with cell-bound antigens.

A prosite search for ICOM members with a mixed leucine/isoleucine LZ motif (C-motif): [IL]x(2)[IL]x(3)[IL]x(2)[IL]x(3)[IL] identified 131 membrane proteins. This search included the 32 MPs identified in the B-motif search and extended the group of adhesion proteins carrying an ICOM within their TMD. Interestingly, the identified ICOM-positive members expressed on B cells are CD40, CD180 also known as RP105 (see below). Furthermore, I identified the scavenger receptor family member CD6, which fine-tunes TCR signaling, as an ICOM-positive protein.

The finding that the TMD of EpoR contains a LZ structure similar to that of CD79b and CD19 is of particular interest since this receptor has been extensively studied over the last 20 years. Like the BCR, the EpoR was once thought to be activated either by dimerization of two monomers or by an EPO-dependent conformational change that brings two TMDs into close proximity to each other (27). However, this idea has been challenged by studies showing that the resting EpoR already forms a dimer and that the LZ motif within its TMD is involved in receptor dimerization (28, 29). The EpoR can be activated not only by the binding of its ligand EPO but also by coexpression of the envelope-related glycoprotein gp55 of the Friend spleen focus forming virus (F-SFFV), which causes erythroblastosis in an EpoR-dependent manner (30). Interestingly, the TMD of gp55 carries an ICOMs-related LZ motif ( Table 2 ). Newer data show that the TMD of gp55 either forms homodimers or binds to the TMD of EpoR (31). Thus, dissociation of the TMDs of the EpoR dimer either by Epo or by gp55 has been discussed as the activation mechanism of this receptor. Recent studies suggest that in addition to EpoR and gp55 also a third player, namely the tyrosine kinase sf-Stk, is required for fibroblast transformation by F-SFFV by forming a tripartite EpoR/p55/sf-Stk complex as an oncogenic driver (32, 33). Interestingly, sf-Stk is an N-terminal truncated form of murine MST1R/RON, which is one of the six hits of the A-motif search of the human membrane proteome. Thus, three different ICOM-positive membrane proteins are implicated in oncogenic complex formation and cellular transformation. In summary, genetic and biochemical studies support the notion that ICOM-carrying TMDs are involved in the homo- and/or heterodimerization of membrane proteins and regulate important biological processes such as cellular transformation.

The radio-protective 105 (RP105) antigen was discovered in a screen for monoclonal antibodies (mAb) that protect B cells from X-ray irradiation induced apoptosis. Surprisingly, B cells exposed to the RP105 mAb underwent a massive mitogenic B cell expansion and differentiation (34). The cloning and sequencing of RP105, now known as CD180, identified this protein as a new member of the toll-like receptor family (35). CD180 is an orphan receptor that forms a dimer and can interact with toll-like receptor 4 (TLR4). Similar to TLR4 which requires binding to MD-2, CD180 reaches the cell surface only in association with the adaptor MD-1 (36). It remained a puzzle that the mitogenic signaling via CD180 did not involve the classical Toll-like/Myd88 signaling pathway but rather depended on the components of the BCR signaling pathway such as BTK and Lyn (37, 38). My discovery that the TMD of CD180 carries an ICOM provides a solution to this puzzle ( Figure 2 ). I suggest that on resting B cells, CD180 is part of the inhibitory IgD-type nanocluster. Upon exposure of B cells to either cognate antigen or to anti-RP105 mAb the inhibitory complex is resolved and CD180 becomes part of the ICOM interactome on activated B cells that promotes CD19/IgM-BCR complex formation and ITAM/PI-3K signaling. The finding that the extracellular part of the CD180/MD-1 complex can bind to the TLR4/MD-2 complex (39) and my suggestion that the TMD of CD180 is associated with the ICOM interactome suggest a direct link between the innate and adaptive immune system. This suggestion also provides an explanation for the cooperation between TLR4 and BCR in B cell activation (40, 41).

When I discovered the dual tyrosine motif (now known as ITAM) in the cytoplasmic tail of the signaling components of the BCR and TCR, as well as Fc receptors, I noticed that the glycoprotein gp30 of the bovine leukemia virus (BLV) also carried such a motif (3). Subsequently, other ITAM-bearing viral proteins were described, demonstrating that viruses employ the ITAM-dependent signaling modules as part of their propagation strategy (42).

Given the importance of the ICOM interactions for lymphocyte regulation and activation, it is likely that viruses and tumor cells employ the ICOM coupling and organization function. Indeed, I found that BLVgp30 carries not only an ITAM in its cytoplasmic tail but also an ICOM within the TMD sequence ( Table 3 ). BLVgp30 may thus interact with CD19 in an ICOM-dependent manner to establish an ITAM/PI-3K signaling axis and this might be one of the drivers of bovine leukemia. Whether or not BLVgp30 or the env proteins of related viruses can also directly interact with the antigen receptors in the way that gp55 activates the EpoR remains to be investigated. Clearly, a deregulated ICOM interaction is involved not only in viral infection but also in oncogenic transformation. Another example for this is the platelet-derived growth factor receptor beta (PDGFRb), whose TMD can form ligand-independent homodimers (43) and interact with the bovine papillomavirus BPVE5R transmembrane protein (44, 45). The TMD of both proteins carries an ICOM-related LZ structure, which may promote their interaction and tumor development ( Table 3 ).

In our nanoscale Fab-PLA studies, we found that the transmembrane phosphatase CD45 (PTPRC) is associated with the IgM-BCR complex on resting but not on active B cells. By dephosphorylating the two ITAM tyrosines, CD45 has a negative signaling role as a gatekeeper of IgM-BCR activation (8). However, upon B cell activation, CD45 dephosphorylates the negative regulatory pY593 of Lyn and has a positive role in IgM-BCR signal amplification. Thus, like many signaling proteins, CD45 can alter its signaling functions depending on its location and association. To learn more about the molecular association of CD45 with the resting IgM-BCR, I had a look at the evolutionary highly conserved TMD sequence of CD45 and found that one side of the CD45 TMD carries an ICOM-related sequence in which the 2 outer AAs of the 5 LZ positions are occupied by a phenylalanine and tyrosine instead of a leucine ( Figure 1B ). When evaluating the CD45 ICOM, it should be kept in mind that within the lipid bilayer, a phenylalanine or tyrosine can form tight complexes with either a leucine or isoleucine residue (46). This notion is supported by the described AAs interactions between the TMDs in the BCR complex (17). A CD45 ICOM-related sequence has also been found in the TMD of human CD101, also known as immunoglobulin superfamily member 2 (IGSF2) and cadherin 26. A related sequence is also present in the TMD of two further human PTPRs namely PTPRE and PTPRA ( Table 3 ).

It has been suggested that CD45 and the related transmembrane phosphatase PTPRA can form dimers and that dimerization is mediated by the extracellular domain and the TMDs of these phosphatases. It is therefore possible that the ICOM of these phosphatases is involved in their dimerization. The CD45 dimer is less active and seems to be the autoinhibited form of this tyrosine phosphatase (47). Thus, dissociation and reorganization of the CD45 dimer appears to be associated with an increased enzymatic activity of CD45. Interestingly this process is regulated by the CD45-associated protein (CD45-AP or PTPRC-AP), a small membrane protein that binds to CD45 via its leucine-rich TMD (48). Closer examination of the CD45-AP TMD sequence reveals that it carries an ICOM-related LZ structure ( Table 3 ). T cell studies have shown that the binding of CD45-AP to CD45 not only leads to the dissociation of the CD45 dimer, but also to an increased phosphatase activity and CD45/TCR interaction (49). In this respect, the ICOM-mediated CD45/CD45-AP interaction is similar to that of the EpoR/gp55 pair. Thus, on resting B cells, CD45-AP may promote CD45:CD45 dissociation and the binding of CD45 to CD79b. In this way, an active tyrosine phosphatase is coupled to the IgM-BCR and this association may help to maintain the resting state of B lymphocytes.

The discovery of conserved ICOMs within the TMD of one of the two BCR signaling components and several BCR co-receptors suggests that B cell maintenance and activation is a very dynamic process involving different ICOM interactions. Based on this new information, one could imagine the following scenario for the regulation and activation of B cells via specific TMD : TMD interactions ( Figure 2 ). On the surface of resting B cells, the IgM-BCR may form several inhibitory complexes that prevent deregulated signaling. One complex is an autoinhibitory IgM-BCR oligomer that is maintained by an ICOM-dependent CD79b/CD79b and a HC:HC homodimerization (50). Within the closed IgM-BCR oligomer, the CD79b ICOMs are not accessible for interaction with CD19 or other positive signaling components. A second inhibitory complex could be formed via the ICOM-dependent interaction of the monomeric IgM-BCR with the CD45/CD45-AP complex as discussed above. In contrast to that of the IgM-BCR, the IgD-BCR ICOM interactome is located in a raft lipid domain and may be organized by a CD20/CD81/CD19 tetraspanin web (model not shown). Once an ICOM : ICOM-negative regulatory complex is formed and delivered to the plasma membrane, it could be further stabilized by extracellular cross-linkers. A likely candidate for this function is galectin-9 (Gal9), a dimeric adaptor protein carrying two carbohydrate recognition domains separated by a flexible linker. Exposure of B cells to Gal9 suppresses IgM-BCR mobility and signaling (51). Gal9-deficient B cells are hyperactive, thus demonstrating the negative regulatory role of this lectin. Interestingly, several of the identified binding targets of Gal9 such as CD45, CD79b and CD180 also carry an ICOM with in their TMD. Upon B cell activation these ICOM complexes may be remodeled as outlined in Figure 2 . Furthermore, an ICOM-mediated interaction could also be involved in the formation of an CD19:CD79b/CD79b:CD19 complex that is found on the B cell surface after BCR internalization and can emit a pro-survival signal (52). In summary, the discovery of conserved ICOM sequences in the TMD of CD79b and several coreceptor modules suggests that the IgM-BCR regulation and activation is a dynamic process involving stable or transient ICOM-based homo- and hetero-dimerization of the IgM-BCR with itself and with CD45 and CD19.

Does this new insight into the ICOM-dependent regulation and activation of B cells process also have implications for the regulation of T cells? Given the prominent role of CD79b within the IgM-BCR complex, I asked whether the αβTCR has a similarly exposed ICOM-carrying signaling component. The cryo-EM structure of the αβTCR shows an 8-helix instead of the 4-helix TMD bundle of the BCR (18). The signaling subunits are organized as two heterodimers (CD3γ/CD3ε and CD3δ/CD3ε) and one TCR-ζ/TCR-ζ homodimer. Interestingly, the TMD AAs of the CD3γ and CD3δ chains establish different contacts within the asymmetric 8-helix bundle. All sides of the CD3γ TMD are involved in binding to either the CD3ε, the TCR-α or TCR-ζ TMD whereas the CD3δ TMD has a free side with an ICOM-related LZ motif ( Table 3 ). It is thus likely that CD3δ either mediates TCR : TCR oligomerization (53, 54) or couples the αβTCR to lateral interactors such as the CD45/CD45-AP complex (8). Indeed, the ICOM-carrying side of the CD3δ TMD shares six identical or similar AAs with the CD45 TMD. Similar to the variant LZ motif of CD45, the human CD3δ LZ motif carries an isoleucine or leucine only at the three core positions, while the two outer positions are occupied by a threonine and a cysteine.

CD45 has been well studied as a gatekeeper of the resting state of T cells and the ICOM-based inhibitory CD3δ/CD45 complex may be a part of this important regulation. Another negative regulator of T cell activation is the transmembrane adaptor “phosphoprotein associated with glycosphingolipid-enriched microdomains 1 (PAG-1), also known as CSK binding protein (Cbp). PAG-1 is constitutively tyrosine phosphorylated in resting T cells and is associated with the C-terminal Src kinase CSK, which inhibits Src-family kinase (SFK) activity (55). It is only upon T cell activation that PAG-1 is dephosphorylated, allowing for the release of CSK and SFK inhibition. T cell activation is also under the control of the immune checkpoint inhibitor PD-1, which is targeted by anti-PD1 antibodies in several successful anti-tumor therapies. A recent study showed that the inhibitory function of PD-1 requires PAG-1 and that the two proteins are found in close proximity on the T cell membrane (56, 57). Genetic studies have also identified the scavenger receptor family member CD5 as a negative regulator of T cell selection and activation (58). Interestingly, all three of these proteins carry an ICOM-related LZ motif with a cysteine at the 5^th position ( Table 3 ), and the CD3δ ICOM may help to couple these negative regulators to the αβTCR. T cell activation is accompanied by the recruitment of the “linker for activation of T cells” (LAT) to the TCR complex. The TMD of LAT also carries an ICOM-related LZ motif with a cysteine at the 5^th position ( Table 3 ). Thus, it is possible that on activated T cells, LAT replaces PAG/PD-1/CD5 in its interaction with the CD3δ ICOM. Interestingly, it has been found that not only PAG-1 or LAT but also CD3δ can be palmitoylated on the cysteine at the 5^th position of the CD3δ ICOM (59). Such a palmitoylation prevents a S-S dimer formation but it does not prevent the Van der Waals interactions of the sulfur atom and thus may stabilize rather than compromise an ICOM-based interaction (46). The CD3δ ICOM may also recruit the adhesion protein CD44 and the T cell signaling modulator CD6 to the αβTCR. As described for the ICOM interactome of B cells, the ligand-dependent activation of T lymphocytes may involve a dynamic reorganization of ICOM-positive membrane proteins on the T cell surface ( Figure 2 ). The suggestion that an ICOM interactome of TCR homodimers and heterodimers regulates the resting state of T lymphocytes and their transition into activation may solve a long-lasting controversy about the oligomeric or monomeric structure of the TCR on the surface of living T lymphocytes. Specifically, the ICOM discovery suggests that the experimentally found TCR monomers (60, 61) may actually be TCR : CD45, TCR : PAG or TCR : CD5 heterodimers and that TCR : TCR homodimers coexist with these heterodimers on different topographical regions of the same T cell surface.