Vanderbilt Institute of Chemical Biology

 

 

Discovery at the VICB

 

 

 

 

 

 

How Integrins Sense the Extracellular Space


By: Carol A. Rouzer, VICB Communications
Published:  December 15, 2016

 

New structural and functional studies reveal the complexity of integrin activation and interaction with the extracellular matrix.

 

Integrins are integral membrane proteins found on nearly every cell in multicellular animals. They provide the means by which cells interact with components of the extracellular matrix (ECM). In humans, integrins are heterodimers composed of one of 18 alpha subunits and one of 8 beta subunits, with 24 combinations found in cells. All subunits with the exception of β4 comprise a large extracellular domain that binds to target ECM components, a single transmembrane domain (TM), and a short C-terminal cytosolic domain (CT) that interacts with the cytoskeleton and various signaling molecules. A key role of integrins is to respond to changes in the ECM by triggering appropriate intracellular signals or to respond to intracellular signals by altering ECM interactions. Thus, it is important that integrins be able to modulate ECM component binding affinity and signaling responses as needed to meet the demands of a changing environment. Much of what is known about integrin modulation comes from extensive studies of the αIIbβ3 protein from platelets, but it is not clear to what degree that integrin is representative of the entire class. To address this issue, Vanderbilt Institute of Chemical Biology member Chuck Sanders and his collaborator Roy Zent (Department of Medicine) investigated the structure and function of the β3 subunit in comparison to β1, the major beta subunit found in integrins in solid organs. They report striking differences between the two subunits and their mechanism of activation [Z. Lu, S. Mathew, J. Chen, et al., eLife, 2016:10.7554/eLife.18633].

 

Under normal conditions, the platelet αIIbβ3 is inactive, exhibiting low affinity for its primary ECM binding partner, fibrinogen. The structure of the dimer is notable for the 25o tilt of the β3 TM that brings it in close proximity to the αIIb subunit near the extracellular leaflet of the membrane bilayer. Here, the two TMs interact to form an outer membrane clasp (OMC, Figure 1). A second, inner membrane clasp (IMC) joins the two TMs near the cytosolic leaflet of the bilayer. A key structural component of this clasp is a salt bridge between D723 and R995 (Figure 1). Platelet activation leads to binding of the protein talin to the CT, resulting in destabilization of the clasps and an increase in binding affinity for fibrinogen. Thus, for αIIbβ3, interaction with the ECM requires prior activation in response to a cell-specific stimulus.

 

 

FIGURE 1. Structure of the transmembrane domains of the αIIbβ3 integrin structure in the inactive conformation. The regions, including the outer membrane clasp (OMC) and inner membrane clasp (IMC) and key amino acids involved in clasp formation, are highlighted. Note that the beta chain (red) is  tilted about 25o relative to the alpha chain (blue). Also shown is K716, the "snorkeling" lysine residue that is believed to form a salt bridge with negatively charged oxyanions in membrane phospholipid head groups. Figure reprinted by permission from Macmillan Publishers Ltd from C. Kim, et al. (2012) Nature, 481, 209. Copyright 2012.

 

 

The high degree of sequence homology between β3 and β1 led the investigators to hypothesize that β1-containing integrins might be activated by a mechanism similar to that of αIIbβ3. To test this hypothesis, they focused on K752, a residue of β1 that is homologous to K716 of β3. In β3, this residue, which is located within the TM, is believed to project its side chain toward the cytosolic leaflet of the membrane (a process referred to as "snorkeling") in order to form a salt bridge with oxyanions in the head groups of membrane phospholipids. Prior studies suggested that K716 snorkeling is required to maintain the 25o tilt of the β subunit in the inactive state of αIIbβ3. A charge reversal mutation of this residue disrupts the interaction with membrane phospholipids and results in permanent activation of αIIbβ3. Thus, the researchers began their studies by investigating the effects of various K752 mutations on β1-containing integrin function.

 

Two key β1-containing integrins are α1β1 and α2β1, the primary integrins involved in binding to collagen. The investigators first expressed wild-type β1, β1 containing a charge preserving K752R mutation, or β1 containing a charge reversing K752E mutation in β1 null renal collecting duct epithelial (CD) cells that naturally expressed α1 and α2. They then tested the ability of the cells to attach to culture dishes coated with collagens I or IV, the primary targets of α2β1 and α1β1, respectively. They found that cells expressing wild-type β1 could attach to and spread on both kinds of collagen, and that the K752R mutation had no effect on these properties. In contrast, the K752E mutation markedly reduced collagen adherence and mildly reduced spreading. These findings were highly unexpected as they suggested that the K752E mutation decreased rather than increased the activation state of α1β1 and α2β1 integrins. Further evidence for this conclusion came from experiments using an antibody that detects activated β1-containing integrins. They showed reduced levels of activated β1 in cells expressing the K752E mutation when compared to cells expressing wild-type β1 (Figure 2). An antibody that detects all β1 confirmed equal expression of the wild-type and mutant proteins. Treatment of cells with Mn2+, a nonspecific integrin activator, served as a positive control for both proteins. Together, the results suggest that α1β1 and α2β1 integrins may exist normally in an activated state (able to bind collagen) and that the K752E mutation serves an inactivating rather than an activating function.

 

 

FIGURE 2. The K752E mutation blocks collagen-binding integrin activation. CD cells adhering to collagen-coated plates expressed equal quantities of wild-type (WT) or K753E integrin, as indicated by fluorescence microscopy of cells treated with the AIIB2 antibody that detects all β1 chains. In contrast, the 12G10 antibody that detects only activated β1 reveals a markedly lower amount of activated integrin in cells expressing the K752E mutation as opposed to those expressing wild-type β1. Treatment of the cells with Mn2+ results in equal levels of integrin activation in both cell population. Figure reproduced under a Creative Commons Attribution 4.0 International License from Z. Lu, S. Mathew, J. Chen, et al., eLife, 2016:10.7554/eLife.18633.

 

 

To develop a better understanding of this striking difference between the platelet αIIbβ3 integrin and the β1 integrins studied here, the investigators performed extensive structural studies of the β1 and β3 TM/CT domains using NMR. For this work, they incorporated the proteins into bicelles, and multiple bicelle and buffer compositions were explored to find conditions that provided optimal data. The results (Figure 3) delineated the extent of the TM and CT domains of both proteins, noting some important differences. For example, although the alpha helix of both proteins extends beyond the TM into the cytosol, the length of this extension is much greater for β3 (16 amino acids) than β1 (8 amino acids). For both proteins, the TM helix is fairly rigid, but the cytosolic extensions are prone to fraying. The CT beyond the helical extensions is unstructured in both β1 and β3.

 

 

FIGURE 3. Diagrammatic representation of the structures of the β1 and β3 TM/CT as determined by NMR of the proteins incorporated into bicelles. The TMs are in blue, and the CTs in red. The positions of key residues at the start of the TM helix, the end of the CT helix, and the TM/CT junction are noted. Also shown is the position of the snorkeling lysine residue. Figure reproduced under a Creative Commons Attribution 4.0 International License from Z. Lu, S. Mathew, J. Chen, et al., eLife, 2016:10.7554/eLife.18633

 

 

Further structural studies employed a water-soluble and a lipophilic paramagnetic probe to identify regions in each protein that were bicelle-embedded versus exposed to solution. The results of these studies showed that the K752R and K753E mutations had no significant effect on β1 topology. This was surprising because prior experiments had shown that the K716E mutation of β3 caused a significant structural change, notable for the loss of the 25o tilt within the membrane. This led the researchers to carry out additional experiments to investigate the structure of β3. Their results demonstrated that the prior conclusions regarding the structural effects of the K716E mutation were the result of an artifact associated with the paramagnetic probe that had been used. When this artifact was eliminated, the data indicated that, as in the case of K752E β1, the K716E mutation of β3 had no major structural effects on the protein. These findings ruled out the hypothesis that the snorkeling lysine residue was responsible for the 25o tilt of either β3 or β1 and opened new questions about the role of K716 in αIIbβ3 activation.

 

Prior work had indicated that the K752E mutation led to permanent activation of the α5β1 integrin, which binds to fibronectin. As activation of αIIbβ3 is associated with loss of the clasps binding the α and β subunits together, the investigators used NMR and fluorescence anisotropy to evaluate the effect of the K752E and K716E mutations on the subunit binding affinities of α5β1 and αIIbβ3, respectively. Both approaches showed strong affinity between the alpha and wild-type beta subunits in each case, and the mutation led to an increase in KD (7-fold for α5β1 and 4-fold for αIIbβ3), confirming the reduction in affinity that would be expected from destabilization of the clasps.  Similar studies of subunit affinity for wild-type and K752E α1β1 and α2β1 revealed relatively poor affinity in the case of the wild-type β1 subunit and the mutant. In all four cases, an affinity constant could not be determined due to failure to reach saturated binding. These results are consistent with the hypothesis that α1β1 and α2β1 normally exist in an activated state lacking clasp-dependent association.

 

The finding that the K752E mutation reduces binding affinity between the subunits in α5β1 was consistent with previous reports that the mutation leads to permanent activation of this integrin. These studies were carried out in CHO cells expressing both the mutant and the wild-type β1 protein. In contrast, when the researchers explored the effects of the K752E mutation using their CD null cells transfected to express only the wild-type or the mutant subunit, they found that the K752E mutation reduced the activation level of α5β1 as indicated by binding to and spreading on fibronectin-coated culture dishes and by antibody-binding to total and activated β1 (Figure 4). Again, these results contrasted with those found previously for the β3 subunit.

 

 

FIGURE 4. The K752E mutation blocks fibronectin-binding integrin activation. CD cells adhering to fibronectin-coated plates expressed equal quantities of wild-type (WT) or K753E integrin, as indicated by fluorescence microscopy of cells treated with the AIIB2 antibody that detects all β1 chains. In contrast, the 12G10 antibody that detects only activated β1 reveals a markedly lower amount of activated integrin in cells expressing the K752E mutation as opposed to those expressing wild-type β1. Treatment of the cells with Mn2+ results in equal levels of integrin activation in both cell populations. Figure reproduced under a Creative Commons Attribution 4.0 International License from Z. Lu, S. Mathew, J. Chen, et al., eLife, 2016:10.7554/eLife.18633.

 

 

Activation of the platelet integrin αIIbβ3 occurs upon binding of talin to the CT. Direct interaction of talin with integrins occurs through its phosphotyrosine-binding F3 domain and the NPxY motif that is found on both β1 and β3. Studies of talin F3 domain binding to the TM/CT of wild-type and mutant beta chains showed that the K752E mutation had little effect on the binding affinity of β1 for the F3 domain of talin. In contrast, the K716E mutation increased the KD for talin F3 binding to β3 by six-fold. This substantial loss in affinity was not expected in light of the activating nature of the K716E mutation in β3.

 

Together, the findings reveal new insights and raise important questions about integrin activation. They clearly show that β1- and β3-containing integrins are structurally distinct and are likely activated by different mechanisms. In fact, they suggest that at least some β1 integrins are found in a default active state, rather than requiring activation as in the case of αIIbβ3. They also call into question the structural role of the snorkeling lysine present in the TM of both β1 and β3. Although it is clear that this amino acid plays a critical role in integrin function, that role appears to be opposite in the two integrins, and the mechanism of its effects is now uncertain. Clearly more work is required to understand exactly how these important proteins regulate the interactions between cells and their immediate environment.

 

 

View Elife article: Implications of the differing roles of the β1 and β3 transmembrane and cytoplasmic domains for integrin function

 

 

TwitterYouTube

 

 

The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center

  • Areas of Interest