Author response: Von Willebrand factor A1 domain stability and affinity for GPIbα are differentially regulated by its O-glycosylated N- and C-linker
Year: 2022
Type: peer-review
Abstract: Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Hemostasis in the arterial circulation is mediated by binding of the A1 domain of the ultralong protein von Willebrand factor (VWF) to GPIbα on platelets to form a platelet plug. A1 is activated by tensile force on VWF concatemers imparted by hydrodynamic drag force. The A1 core is protected from force-induced unfolding by a long-range disulfide that links cysteines near its N- and C-termini. The O-glycosylated linkers between A1 and its neighboring domains, which transmit tensile force to A1, are reported to regulate A1 activation for binding to GPIb, but the mechanism is controversial and incompletely defined. Here, we study how these linkers, and their polypeptide and O-glycan moieties, regulate A1 affinity by measuring affinity, kinetics, thermodynamics, hydrogen deuterium exchange (HDX), and unfolding by temperature and urea. The N-linker lowers A1 affinity 40-fold with a stronger contribution from its O-glycan than polypeptide moiety. The N-linker also decreases HDX in specific regions of A1 and increases thermal stability and the energy gap between its native state and an intermediate state, which is observed in urea-induced unfolding. The C-linker also decreases affinity of A1 for GPIbα, but in contrast to the N-linker, has no significant effect on HDX or A1 stability. Among different models for A1 activation, our data are consistent with the model that the intermediate state has high affinity for GPIbα, which is induced by tensile force physiologically and regulated allosterically by the N-linker. Editor's evaluation Through a series of elegant experiments, the authors concluded that the N-terminal linker together with its O-linked glycans stabilize the A1 in an inactive, native conformation. Removal of the N-linker or its O-linked glycans destabilizes the A1 structure and may activate the A1 into an intermediate, higher affinity state for GPIbα binding. The authors also concluded that the C-terminal linker and its O-linked glycans do not contribute to stabilizing the A1 structure, though they prevent GPIbα binding through steric hindrance and electrostatic repulsion. The conclusions are consistently supported by the data from multiple experimental techniques. The reported results are highly interesting for the hematology and biophysics communities. https://doi.org/10.7554/eLife.75760.sa0 Decision letter eLife's review process Introduction The ultra-long length of the blood plasma protein von Willebrand factor (VWF) enables its activation by hydrodynamic forces at sites of hemorrhage (Springer, 2014; Figure 1A, B). VWF, a long concatemer of monomers linked head-to-head and tail-to-tail, undergoes a change of its multimeric superstructure from a random coil form (Parker and Lollar, 2021) to a thread-like, extended form when exposed to elongational flow (Bonazza et al., 2015; Fu et al., 2017). Hydrodynamic flow activates VWF to bind to glycoprotein Ib (GPIb) on platelets, the essential interaction required for hemostasis and thrombosis in the arteriolar circulation. Fluorescent imaging of single VWF concatemers tethered in a flow cell showed that binding of GPIbα not only requires elongation of VWF but also a half-maximal tensile force of ~20 pN transmitted through the backbone of VWF concatemers (Fu et al., 2017). Activation could be well fit to a model of transition from a low-affinity state 1 to a high-affinity state 2 (Figure 1B). Measuring A1–GPIbα binding and unbinding with laser tweezers also revealed two states, with a force dependent switch between two states with different on- and off-rates (Kim et al., 2015; Kim et al., 2010). In addition to such flex-bond measurements, A1 and GPIbα are also reported to form catch bonds, although these measurements lack single molecule fiduciary markers or fits to models from which on- and off-rates could be extracted (Ju et al., 2013). Despite multiple crystal structures of the A1 domain bound to GPIb, distinct conformational states of A1 that explain force-dependent regulation of A1 affinity for GPIbα have yet to emerge (Blenner et al., 2014; Dumas et al., 2004). Figure 1 Download asset Open asset von Willebrand factor (VWF), hydrodynamic flow, A1 activation, and A1 linkers. (A) N- and C-terminal linkage and domain organization of a VWF monomer in a VWF concatemer. Monomers are connected head-to-head and tail-to-tail. (B) When exposed to elongational flow and tethered (black triangle) on a vessel wall, VWF concatemers extend to a linear shape. At higher elongational flow, which exerts higher mechanical tension along the length of the concatemer (highest at the tether point and zero at the downstream end), A1 undergoes transition from a low- to a high-affinity state as a consequence of tensile forces transmitted to it through its linkers. The disulfide bond between Cys1272 and Cys1458 is indicated with red SS. (C) A1 protein constructs studied here that differ in length of linkers and were expressed in E. coli (not glycosylated) or in mammalian cells (O-glycosylated) (O-glycans shown as lollipops). The long-range disulfide bond is schematized in red. (D) A1 constructs were subjected to sodium dodecyl–sulfate polyacrylamide gel electrophoresis (SDS–PAGE) and stained with Coomassie Blue. Amount of protein added per lane differed between protein constructs but was identical for each construct on reducing and nonreducing PAGE. The force-responsive A1 domain is unusual among protein domains in being linked to neighboring domains through O-glycosylated, mucin-like segments (Figure 1A, blue hashed lines). Tensile force exerted on A1 is transmitted through these linkers. It has been suggested that A1-flanking sequences (Auton et al., 2012; Interlandi et al., 2017; Ju et al., 2013; Miyata and Ruggeri, 1999; Sugimoto et al., 1993) or domains (Aponte-Santamaría et al., 2015; Ulrichts et al., 2006) may shield the GPIbα-binding epitope and thereby inhibit otherwise tight binding to GPIb. It was also proposed that the N- and C-linkers interact with each other and form an autoinhibitory module that masks A1 (Deng et al., 2017). A force-dependent signature for breakage of this module has been reported that correlates with the combined expected extension of the N- and C-linkers (Arce et al., 2021). An alternative mechanism has also been proposed by which the N-linker could regulate A1 affinity, that is, by regulating the relative stability of the A1 native and intermediate states. A1 has been found to have three states, native (N), intermediate (I), and denatured (D), while the homologous A2 and A3 domains in VWF have only two states, native and denatured (Auton et al., 2007). Lower stability of the A1 native state correlates with higher affinity (Auton et al., 2012; Auton et al., 2010a). This anticorrelation between A1 affinity and stability is also seen with von Willebrand disease type 2B gain-of-function mutations, which increase A1 affinity and decrease stability (Auton et al., 2010b; Tischer et al., 2017). Conversely, the G1342S decrease-in-function type 2M von Willebrand disease mutation decreases A1 affinity for GpIbα and increases A1 stability (Auton et al., 2009). Mucin-like regions in proteins have unique characteristics distinct from both intrinsically disordered polypeptide segments and folded domains. The bulky O-linked glycans which include sialic acid that are attached to threonine and serine residues in mucins are highly solvated, repel one another, and make mucins extended, with an average length of ~2 Å/residue (Clemetson, 1983; Woollett et al., 1985). Electron microscopy showed that D3 and A1 in VWF, which are connected by the N-linker, have many possible orientations, and an interdomain distance as long as ~7 nm (Zhou et al., 2011). A1 is mechanically stabilized by a long-range disulfide bond between Cys1272 and Cys1458 (Figure 1B, C). Under tension, the core of A1 between these two cysteines, residues 1273–1457, is protected from complete unfolding. Only the portions of A1 external to this disulfide bond, residues 1238–1272 and 1458–1489, are directly subjected to the mechanical force transmitted through the spine of the VWF concatemer. However, multiple internal residues are noncovalently associated with folded external residues, and thus the stability of internal residues is expected to be affected by elongation of external residues by tensile force. Here, to understand how binding of the VWF A1 domain to platelet GPIbα is regulated by its N- and C-terminal mucin-like linkers, we have measured the effects of both the glycan and polypeptide moieties of these linkers on stability, thermodynamics, and ligand-binding affinity and kinetics of A1. These linkers include the multiple serine and threonine residues that were found to be O-glycosylated during protein sequencing of native plasma VWF, including Thr-1248, Thr-1255, Thr-1256, and Ser-1263 in the N-linker (Titani et al., 1986). Mass spectrometry verified these assignments and showed a total of eight sialic acids over the four sites (Solecka et al., 2016). We show that both the N- and C-linkers affect affinity, that both their polypeptide and O-glycan moieties are important, and that the N-linker affects stability and hydrogen deuterium exchange (HDX) of A1 including in specific regions internal to its disulfide. Results Proteins We compared the effect of the linkers N- and C-terminal to A1, which separate A1 from D3 and A2, respectively, and which contain O-glycosylation sites that have been chemically identified in native VWF (Solecka et al., 2016; Titani et al., 1986; Figure 1). Because the linkers are O-glycosylated in mammalian Expi293 cells and not in E. coli, we expressed A1 protein constructs in both cell types in order to test the effect in some assays of both polypeptide and O-glycan components of the linkers (Figure 1C). The N- and C-terminal boundaries of the constructs were based on allowing comparability to previous studies; references to the previous literature and a structural rationale for defining the minimal size of A1 as residues 1270–1463 and the maximal size of A1 plus linkers as residues 1240–1494 are described in Methods. To obtain A1 disulfide bond formation in bacteria, thioredoxin–A1 fusion proteins were expressed in E. coli SHuffle cells. The C-terminal boundary of the A1+N+C bacterial construct at Ser-1486 that was already on hand was extended by three residues in the mammalian A1+OGly+N+C construct to increase the likelihood of O-glycosylation of residue Ser-1486 and to include O-glycosylated residue Thr-1487 (Solecka et al., 2016; Titani et al., 1986). Constructs were purified using His tag affinity and by Superdex S200 size exclusion, and for the bacterial constructs, also with an intervening heparin affinity chromatography step. As expected based on O-glycosylation, the mammalian constructs migrated more slowly and showed increased heterogeneity in sodium dodecyl–sulfate polyacrylamide gel electrophoresis (SDS–PAGE) compared to the corresponding bacterial A1 constructs (Figure 1D). Furthermore, faster migration of all seven A1 constructs in nonreducing than reducing SDS–PAGE confirmed that all contained the long-range disulfide bond. A1–GPIbα-binding kinetics and thermodynamics Using biolayer interferometry (BLI), we measured A1 association to and dissociation from GPIbα immobilized through a biotinylated C-terminal avitag to BLI sensor tips. A1 preparations were subjected to S200 gel filtration the day of each measurement to remove aggregates. Global fits at all analyte concentrations to a single on- and off-rate for each A1 construct were good (Figure 2). In contrast, when gel filtration was omitted, data could not be fit to a single on- and off-rate. Differences among all seven A1 constructs in 150 mM NaCl showed that the linkers were of great importance in regulating A1 affinity for GPIbα (Figure 2). Comparisons to A1 short showed that nonglycosylated linkers in A1+N+C lowered affinity by 10-fold while the presence of glycosylated linkers in A1+OGly+N+C lowered affinity by 50-fold. The N-linker in A1+N lowered affinity by 4.3-fold while the C-linker in A1+C lowered affinity by 2.5-fold. Glycosylation of the N-linker was also more important than the C-linker; the affinity of A1+OGly+N was 10-fold lower than A1+N while the affinity of A1+OGly+C was only 2-fold lower than A1+C. O-Glycosylated linkers consistently lowered on-rates (Figure 2B). Overall, the results showed that both the polypeptide linker moiety and the O-glycan moiety of A1 linkers contributed to lowering A1 affinity for GPIbα, that the N-linker was more important than the C-linker in lowering affinity, and that the combined effect was very large at 50-fold. Figure 2 Download asset Open asset Effect of A1 linkers, their glycosylation, and ionic strength on GPIbα-binding kinetics. (A) Biolayer interferometry (BLI) traces in gray are shown for the indicated concentrations of the indicated A1 construct at the indicated NaCl concentration. In each panel, all A1 concentrations were fit globally to a one-site-binding model to obtain a single kon and koff. Thin black lines show fits. Fitting errors ranged from 1% to 5% of the fit values. Vertical dashed lines mark the beginning of the dissociation phase. (B) Kinetic constants and derived KD. Errors show difference from mean of n = 2 independent experiments. To test the electrostatic contribution to binding affinity and on-rate, binding of A1 short and the three glycosylated constructs was additionally measured in 30 and 10 mM NaCl (Figure 2). Lowering NaCl concentration from 150 to 10 mM increased on-rates by an average of 17-fold for the four constructs. The effect of NaCl was predominantly on on-rate, since on- and off-rate together resulted in only a somewhat greater average increase in affinity of 23-fold (Figure 2B). Our results demonstrate that electrostatic complementarity between the binding interfaces in the basic A1 (pI 9.2) and acidic GPIbα (pI 5.5) domains (Huizinga et al., 2002) gives rise to electrostatic steering that speeds association kinetics. To independently test affinity, and obtain binding thermodynamics, we used isothermal titration calorimetry (ITC). Although the GPIbα protein used in ITC lacked the avitag and its biotinylation, KD measurements in BLI and ITC were on average within 1.3-fold of one another and showed the same trends (Figure 3). A1 short showed a 39-fold increase in affinity compared to A1+OGly+N+C, the N-linker was more important than the C-linker, and the O-glycan moiety was more important than the polypeptide moiety in regulating affinity. Binding of all A1 protein constructs to GPIbα absorbed heats that showed the reaction was endothermic, that is, entropically driven (Figure 3B). The thermodynamics of A1–GPIbα binding showed enthalpy–entropy compensation (Liu and Guo, 2001; Starikov and Nordén, 2007): the higher the enthalpic contribution to binding, the higher the entropic cost, with all seven constructs following the same rank order for the enthalpic and entropic terms. Thus, the linkers affected both the enthalpic and entropic components of GPIbα binding. Figure 3 Download asset Open asset Effect of A1 linkers and their glycosylation on GPIbα-binding thermodynamics. (A) Panels show ITC traces of heat absorption after each injection (above) and fit of enthalpy to the binding isotherm (below) in 150 mM NaCl at 22°C. (B) Table summarizing the reaction Gibbs free energies (∆G) and the enthalpic (∆H) and entropic (T∆S) contributions. Errors show difference from mean of n = 2 independent experiments, except for A1+N+C which shows the fitting error from one experiment. A1 stability and the thermodynamics of the intermediate state The relative stabilities of A1 short and the three O-glycosylated A1 constructs to denaturation by heat or urea were measured by tryptophan fluorescence, whose emission maximum shifts to higher wavelength upon exposure to solvent during unfolding. The fluorescence intensity ratio (FIR), I350/I330, showed that constructs had three states, that is, exhibited two unfolding transitions (Figure 4A, B, D). Constructs containing the N-linker, that is A1+OGly+N+C and A1+OGly+N, were more stable to heat (Tm of 53.2 and 53.3°C for the first transition, respectively) than constructs lacking the N-linker, that is A1+OGly+C and A1 short (Tm of 49.9 and 50.7°C for the first transition, respectively) (Figure 4A, B). These results suggest that the N-linker stabilizes the native state of A1 relative to the intermediate state. Furthermore, all four constructs showed a second inflection point between 60° and 65° (second peak or shoulder of the derivative curve, Figure 4B). These results suggested two-step unfolding behavior and thus that A1 has three states, native (N), intermediate (I), and D (denatured) as previously observed (Auton et al., 2007). Figure 4 Download asset Open asset Stability of A1 constructs to heat and urea and free energy difference between the native and intermediate A1 states. (A–C) Effect of heating from 30 to 85°C on A1 stability and aggregation. (A) Unfolding was measured by excitation of tryptophan at 275 nm and measuring the ratio of the fluorescence intensity at 350 and 330 nm (fluorescence intensity ratio [FIR] = I350/I330). (B) The derivative of data in A (dFIR/dT) shows two unfolding transitions for each A1 construct and provides evidence for three states of A1: N, I, and U as described in Results. (C) Backscattered light intensity as a measure of A1 aggregation. (D) Unfolding of A1 by urea measured at 20°C by FIR as in panel A. Lines show fits to the three-state model of A1 unfolding as described in Methods and Results (see also Source Data). (E) Tabulation of the melting temperature of the native state (Tm) from panels A and B and the free energy difference between the native and intermediate states of A1 (ΔGI–N), the concentration of urea at which the I and N states are equally populated (Curea I–N), and m, from the fits in panel D. Figure 4—source data 1 A1 short. https://cdn.elifesciences.org/articles/75760/elife-75760-fig4-data1-v1.txt Download elife-75760-fig4-data1-v1.txt Figure 4—source data 2 A1 short + C. https://cdn.elifesciences.org/articles/75760/elife-75760-fig4-data2-v1.txt Download elife-75760-fig4-data2-v1.txt Figure 4—source data 3 A1 short + N. https://cdn.elifesciences.org/articles/75760/elife-75760-fig4-data3-v1.txt Download elife-75760-fig4-data3-v1.txt Figure 4—source data 4 A1 short + NC. Download A1 by heat was measured by light the inflection point for the intermediate state, A1 short to (Figure compared to Figure 4B). required higher temperature for which to after the inflection point for the denatured state (Figure compared to Figure 4B). the O-glycosylated N-linker protected A1+OGly+N and A1+OGly+N+C from at to in urea confirmed the of an intermediate state and was used to the free energy of the intermediate state relative to the native state, that is, The A1 constructs showed two separate unfolding one between 1 and 4 and another urea (Figure constructs showed a in or in the of a but with a decrease in of a three-state unfolding were previously seen for a construct to A1+N+C using (Auton et al., 2007). We fit the unfolding data to a three-state We that all four constructs had the same that the of and that each state had a unique FIR as a of urea concentration that was among the four constructs We thus the difference in energy between the I and N states (ΔGI–N), the concentration of urea at which the I and N states are equally populated (Curea I–N), and m, a that is to the size of folded domains and and Figure The results showed that the O-glycosylated N-linker increased the stability of the native state and thus the gap in energy between the native and intermediate state, with higher and transitions at higher urea concentrations for A1+OGly+N and A1+OGly+N+C than for A1 short and A1+OGly+C (Figure in HDX among A1 constructs with affinity and stability HDX solvent and and on the and stability of proteins and The O-glycosylated A1 constructs and A1 short were to exchange their backbone for deuterium in for of with and the kinetics of deuterium exchange were measured for A1 Figure and Figure HDX as a of structure and for all four constructs studied for A1 short in Figure and for all in Figure 1 and 2). Figure shows exchange for all at all from 10 to 4 the A1 from N- to the exchange was seen for the the the and the and the which all had at one with than at 4 These slowly structural one another and the GPIbα-binding on A1 (Figure Figure with all Download asset Open asset The N-linker decreases A1 measured by hydrogen deuterium exchange (A) deuterium exchange at all for A1 short as of the backbone in each to the Figure 1 and 2 show data for all constructs. (B) in HDX at all of A1 short HDX for the other three constructs as In (A) and residues with type 2B are shown as red (C) HDX kinetics for Figure show data for all (D) of A1 bound to GPIbα A1 is from N- to with type 2B are shown as residues are shown in GPIbα is shown in from the to the in the of HDX were found in specific regions among the four A1 constructs (Figure and Figure The A1+OGly+N+C and A1+OGly+N constructs showed deuterium exchange at all in that were on the and the These near to GPIbα (Figure We also observed exchange in the but the difference was only seen at the 4 The is the long-range disulfide and in a with multiple gain-of-function type 2B (Figure The is near to GPIbα and has four basic residues in Figure but they are not to directly interact with acidic residues in the GPIbα The HDX data showed no significant HDX between A1 short and the of HDX between A1 short and both A1+OGly+N+C and A1+OGly+N was very However, while A1+OGly+N+C and A1+OGly+N were to one another in they both deuterium than A1 short and that both A1 short and were more than A1+OGly+N+C and These HDX the measurements of the stability of the native states relative to the intermediate states (Figure which showed that A1+OGly+N+C and A1+OGly+N were stable and more stable than A1 short and which were also in stability. Discussion Among domains in the A1 domain of VWF is highly unusual in being from neighboring domains by mucin-like linkers and being activated by tensile force transmitted through linkers that from hydrodynamic force on VWF concatemers and bound Here, we have the effects of these linkers on A1 thermodynamics and in binding to its have the effect of the N-terminal linker or both linkers on types of or binding assays and thermodynamics and found effects on GPIbα binding from the N-linker or both linkers to A1 (Arce et al., Auton et al., 2012; et al., et al., 2017; Interlandi et al., 2017; Ju et al., 2013; Miyata and Ruggeri, 1999; et al., Tischer et al., 2017; Tischer et al., We found that both the N- and C-linker A1 affinity for GPIbα and that for both the N- and C-linker, O-glycosylated linkers affinity more than polypeptide linkers. the C-linker, both the O-glycan and polypeptide moieties were in decreases in affinity while for the N-linker, the O-glycan was more than the polypeptide O-glycosylated linkers on the N- and C-terminal of A1 lowered affinity and the presence of both O-glycosylated linkers lowered binding compared to the presence of the N-linker The separate of the polypeptide and glycan moieties of the A1 N- and C-linkers had not previously been O-glycosylation of both linkers together was reported to decrease affinity from to et al., these reported differ from by to they in of the affinity difference by Kinetic measurements showed that the effect of both the polypeptide and O-glycan components of the linkers was Overall, among the seven constructs, on-rates by to short while off-rates to 10-fold binding is important for proteins that in proteins for binding they one another, as previously using the ionic strength of binding of GPIbα to VWF tethered to the of a flow (Fu et al., 2017). Our measurements in 10 mM compared to 150 mM NaCl showed a increase in affinity for A1+OGly+N+C driven by a increase in while VWF tethered to a flow showed a increase in affinity driven by an increase in In 150 mM the high-affinity state of mechanically VWF had an affinity of for GPIbα, compared to the of for A1 short and for A1+OGly+N+C measured Introduction of two into A1 was shown to increase affinity for GPIbα by 10-fold (Blenner et al., However, though such a be expected to increase the affinity of the A1 short construct to a higher than found for mechanically A1 in VWF, a crystal structure of A1 short with the two bound to GPIbα revealed only from a crystal structure (Blenner et al., tension on the one hand and of in the of tension and of the N- and C-linkers on the other hand may have different effects on the structure of the A1 consistent with the that A1 short with two bound to GPIbα not an A1 high-affinity state, an affinity for GPIbα to that seen with mechanically VWF (Blenner et al., We BLI affinity measurements here with ITC measurements of affinity and binding affinity measurements in the literature on identical A1 differ to from reported here (Huizinga et al., affinity measurements here on seven different A1 constructs by BLI and ITC were all within to of one We found it was to remove from A1 preparations by gel filtration to BLI measurements to kinetics that required two on- and off-rates to which in some of the literature on A1 binding to GPIbα (Arce et al., Tischer et al., 2017). In measurements in the are more than in such as and a of an binding kinetics. We are of previous measurements of A1 and GPIbα binding. ITC results show that the interaction of GPIbα with all seven A1 constructs was with a large entropic and that ranged from to at 22°C. Among all seven A1 constructs, the enthalpy and were highly
Source: eLife
Language: English
Subfield: Hematology
Field: Medicine
Domain: Health Sciences
Open Access status: gold
Estimated APC paid: $2,500
OpenAlex ID: W4229365068