The Ca2+-sensor synaptotagmin-1 that triggers neuronal exocytosis binds to negatively charged membrane lipids (mainly phosphatidylserine (PtdSer) and phosphoinositides (PtdIns)) but the molecular details of this process are not fully understood. Using quantitative thermodynamic, kinetic and structural methods, we show that synaptotagmin-1 (from Rattus norvegicus and expressed in Escherichia coli) binds to PtdIns(4,5)P2 via a polybasic lysine patch in the C2B domain, which may promote the priming or docking of synaptic vesicles. Ca2+ neutralizes the negative charges of the Ca2+-binding sites, resulting in the penetration of synaptotagmin-1 into the membrane, via binding of PtdSer, and an increase in the affinity of the polybasic lysine patch to phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P2). These Ca2+-induced events decrease the dissociation rate of synaptotagmin-1 membrane binding while the association rate remains unchanged. We conclude that both membrane penetration and the increased residence time of synaptotagmin-1 at the plasma membrane are crucial for triggering exocytotic membrane fusion.
Tsg101 is a highly conserved protein best known as an early-functioning component of cellular ESCRT machinery participating in recognition, sorting, and trafficking of cellular cargo to various intracellular destinations. It shares sequence and structural homology to canonical ubiquitin-conjugating (E2) enzymes and is linked to diverse events regulated by Ub signaling. How it might fulfill these roles is unclear. Here, we show that Tsg101 E2 mimicry permits interactions with diverse ubiquitin ligating (E3) enzymes and underlies its multifunctional capabilities. Coexpression of Tsg101 with the E3 ligase NNedd4-2s protected the enzyme from degradation and, remarkably, other widely divergent ligases as well. Structural alignment with UbcH5, a canonical E2 enzyme, revealed that recognition at the E2-E3 interface, a region broadly conserved despite sequence and structural differences in both E2 and E3 enzymes, was critical for protection. Nevertheless, UbcH5 failed to protect NNedd4-2s, indicating that the UEV chaperone function is unique to the variant. Studies using Cy5-Ub-VME showed that Tsg101-mediated protection reduced accessibility to Cys residues in the ligase. Access to Tsg101 Ub-binding sites was critical: Rabeprazole, which interferes with Tsg101 Ub-binding, diminished E3 ligase protection. Thus, E2 mimicry permitting control of E3 ligase ubiquitin signaling underlies Tsg101’s broad ability to participate in multiple cellular functions. The study provides mechanistic insight into how Tsg101, by partnering with diverse E3 ligases, can contribute to a broad range of cellular activities.
Tsg101, a component of the endosomal sorting complex required for transport (ESCRT), is responsible for recognition of events requiring the machinery, as signaled by cargo tagging with ubiquitin (Ub), and for recruitment of downstream acting subunits to the site. Although much is known about the latter function, little is known about its role in the earlier event. The N-terminal domain of Tsg101 is a structural homologue of Ub conjugases (E2 enzymes) and the protein associates with Ub ligases (E3 enzymes) that regulate several cellular processes including virus budding. A pocket in the domain recognizes a motif, PT/SAP, that permits its recruitment. PT/SAP disruption makes budding dependent on Nedd4L E3 ligases. Using HIV-1 encoding a PT/SAP mutation that makes budding Nedd4L-dependent, we identified as critical for rescue the residues in the catalytic (HECT) domain of the E3 enzyme that lie in proximity to sites in Tsg101 that bind Ub non-covalently. Mutation of these residues impaired rescue by Nedd4L but the same mutations had no apparent effect in the context of a Nedd4 isomer, Nedd4-2s, whose N-terminal (C2) domain is naturally truncated, precluding C2-HECT auto-inhibition. Surprisingly, like small molecules that disrupt Tsg101 Ub-binding, small molecules that interfered with Nedd4 substrate recognition arrested budding at an early stage, supporting the conclusion that Tsg101-Ub-Nedd4 interaction promotes enzyme activation and regulates Nedd4 signaling for viral egress. Tsg101 regulation of E3 ligases may underlie its broad ability to function as an effector in various cellular activities, including viral particle assembly and budding.
Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Outer membrane TonB-dependent transporters facilitate the uptake of trace nutrients and carbohydrates in Gram-negative bacteria and are essential for pathogenic bacteria and the health of the microbiome. Despite this, their mechanism of transport is still unknown. Here, pulse electron paramagnetic resonance (EPR) measurements were made in intact cells on the Escherichia coli vitamin B12 transporter, BtuB. Substrate binding was found to alter the C-terminal region of the core and shift an extracellular substrate binding loop 2 nm toward the periplasm; moreover, this structural transition is regulated by an ionic lock that is broken upon binding of the inner membrane protein TonB. Significantly, this structural transition is not observed when BtuB is reconstituted into phospholipid bilayers. These measurements suggest an alternative to existing models of transport, and they demonstrate the importance of studying outer membrane proteins in their native environment. eLife digest Bacteria must obtain nutrients from their surrounding environment in order to survive. In Gram-negative bacteria, proteins in the outer membrane surrounding the cell actively transport carbohydrates and trace nutrients like iron into the cell’s interior. Although the structures of many of these transport proteins have been determined, the mechanism they use to move molecules across the membrane is poorly understood. To better understand this process, Nilaweera, Nyenhuis and Cafiso examined the structure of BtuB, a transport protein found in the outer membrane of Escherichia coli that is responsible for absorbing vitamin B12. Previous experiments analyzing the structure of BtuB, and other similar transporters, have been carried out on purified proteins that were extracted from the outer membrane. However, these isolated proteins fail to replicate the transport activity observed in bacterial cells. Nilaweera, Nyenhuis and Cafiso therefore wanted to see how the structure of BtuB changes when it is still enclosed in the membrane of E. coli. This revealed that BtuB undergoes large structural changes when it binds to vitamin B12, suggesting that this is an important part of the transport process. However, when purified BtuB was placed into an artificial membrane, these structural changes did not occur. This indicates that the cellular environment in the bacteria is needed for BtuB to carry out its transport role, and explains why previous experiments using purified proteins struggled to see this structural shift. This work highlights the importance of studying bacterial membrane proteins in their native cell environment. BtuB and similar transporters represent a large family of proteins unique to Gram-negative bacteria that have an impact on human health. Since these proteins are structurally alike, the results of this study may help resolve the transport mechanisms of other proteins, ultimately leading to new ways to control bacterial growth. Introduction The passive permeation of low molecular weight solutes across the outer membrane (OM) of Gram-negative bacteria is typically facilitated by porins. However, many higher molecular weight solutes and trace nutrients, including carbohydrates, iron siderophores, cobalamin, copper, and nickel, are bound and transported across the OM by a family of active transporters that are TonB-dependent (Bolam and van den Berg, 2018; Noinaj et al., 2010). These TonB-dependent transporters (TBDTs) derive energy from the bacterial inner membrane by interacting with TonB, a transperiplasmic protein that interacts with the inner membrane proteins ExbB and ExbD (Celia et al., 2016; Maki-Yonekura et al., 2018). This family of transporters has two distinct domains: a β-barrel formed from 22 anti-parallel β-strands, and a core or hatch domain that fills the interior of the barrel. The β-strands on the extracellular surface of the protein barrel are often connected by long loops, while short turns join the strands on the periplasmic interface (see Figure 1B). TonB interacts with the transporter through a conserved motif on the N-terminal side of the core termed the Ton box (Pawelek et al., 2006; Shultis et al., 2006). Figure 1 Download asset Open asset Electron paramagnetic resonance (EPR) spectra obtained in vivo from spin labeled core sites on extracellular face of the BtuB. The spin labeled side chain (A) R1 was attached to sites on the extracelluar core of BtuB shown in (B). BtuB is shown in both extracellular (top) and side views (PDB ID: 1NQH), with the core in yellow and barrel in light blue. The Cα atom on site 188 on the second extracellular loop is shown, which is used as a reference point for measurements to the core. Labeled Cα sites in the core are shown in magenta, along with substrate binding loops SB1, SB2, and SB3. In (C) EPR spectra are shown in the absence (apo state in blue) and presence of substrate (vitamin B12 bound state in red). No change in the spectrum is observed at sites 65 and 90 upon the addition of substrate. The spectra are characterized by well-defined hyperfine extrema ((i) and location of arrow), where the difference between these points is approximately 69 Gauss. This value is close to 2Azz, twice the value of the hyperfine tensor, and indicates that the label is immobilized on the ns time scale. Spectra are the sum of 10–100 Gauss scans. The mechanism of transport in TBDTs is presently poorly understood; however, the large size of most substrates and the absence of any obvious pathway for substrate permeation (Faraldo-Gómez et al., 2003) have led to proposals that transport is mediated by a significant conformational event that involves a partial rearrangement or full removal of the core domain from the surrounding barrel (Chimento et al., 2005). Although many high-resolution structures are available for TBDTs, there is no direct evidence for a major structural change within the core of TBDTs that might indicate a transport mechanism. In the Escherichia coli vitamin B12 (cobalamin) transporter, BtuB, EPR spectroscopy shows that substrate binding unfolds the Ton box at the N-terminus and extends it into the periplasm, an allosteric event that may facilitate the binding of TonB to BtuB (Kim et al., 2007; Xu et al., 2006); however, no other significant structural changes have been observed in the core. High-resolution crystal structures have been obtained for a C-terminal fragment of TonB in complex with BtuB, the ferrichrome transporter FhuA, and the ferrioxamine B transporter FoxA (Pawelek et al., 2006; Shultis et al., 2006; Josts et al., 2019). When TonB binds, the Ton box extends from the core and interacts with the β-sheets of TonB in an edge-to-edge manner. Except for the Ton box, the remainder of the core remains folded and is essentially unchanged. Because TonB binding does not alter the core of BtuB in the BtuB-TonB structure, it has been proposed that TonB alters the core by exerting a mechanical force on the transporter, and current models for transport favor a mechanism where TonB acts by pulling the Ton box thereby unfolding the core (Gumbart et al., 2007; Hickman et al., 2017; Sverzhinsky et al., 2015). Models involving a rotation of TonB have also been proposed (Klebba, 2016); however, in FhuA there are four to five unstructured residues between the Ton box and core when TonB is bound, making the transfer of torque from TonB to the core unlikely (Sarver et al., 2018). Pulling models have been explored using steered molecular dynamics (MD) (Gumbart et al., 2007) as well as single-molecule AFM (atomic force microscopy) pulling experiments (Hickman et al., 2017), and these studies indicate that an N-terminal region of the core (up to residue 73) is preferentially unfolded to permit the movement of vitamin B12 into the periplasm. This work concludes that the C-terminal region of the core is static and does not unfold during transport, a result that is consistent with denaturation experiments on BtuB (Flores Jiménez and Cafiso, 2012). An important caveat to almost all the structural work on BtuB is that it has been carried out on purified or partially purified protein where the native OM environment is no longer present. Since transport in this family of transporters has never been reconstituted, it has never been established that the isolated, purified, and membrane reconstituted BtuB is capable of transport. Recently, we developed an approach to attach spin labels to either extracellular or periplasmic sites on BtuB in intact cells, thereby permitting EPR measurements to be made under conditions where the protein is known to be functional (Joseph et al., 2019; Nilaweera et al., 2019). Preliminary measurements made on BtuB indicate that it behaves differently in the intact cell than it does in a purified reconstituted phospholipid system. For example, a substrate-dependent change in the core domain of BtuB involving substrate binding loop 3 (SB3) is observed in situ but is not seen in a detergent-treated OM preparation (Nilaweera et al., 2019). Moreover, the extracellular loops of BtuB are also highly constrained in the intact cell, and substrate-induced structural changes and structural heterogeneity that is observed for BtuB in proteoliposomes are not observed for BtuB in situ (Nyenhuis et al., 2020a). In the present work, we perform double electron-electron resonance (DEER) on BtuB in intact cells to determine whether structural changes take place in the core domain that are associated with substrate binding and transport. Upon substrate binding, a movement of the core is observed involving sites 90 and 93 in SB3. No other movements in the core are detected. When the ionic lock is broken between site R14 on the C-terminal side of the Ton box and site D316 in the barrel, long-distance components appear upon substrate binding, indicating that SB3 can assume a state where it has moved into the BtuB barrel toward the periplasmic side of the protein. Under these same conditions, other sites on the N-terminal side of the core remain static. Since this ionic interaction would normally be broken upon TonB binding, this structural transition is likely to take place during transport. This result suggests that transport may involve allosteric changes in the C-terminal side of the core upon TonB binding. Remarkably, these substrate-induced structural changes are not observed for purified, membrane reconstituted BtuB, which may be due to the absence of lipopolysaccharide (LPS) or other periplasmic components in the reconstituted system. The importance of the native OM environment provides an explanation for why this structural transition has not been previously observed. Results Multiple sites in the core region of BtuB may be spin labeled in intact E. coli To investigate movements that might occur in the BtuB core region in situ, pairs of spin labels were placed into the extracellular region of BtuB, with one label located at an outer loop site that is known to be relatively fixed in the cell environment and a second label located at a site in the BtuB core. Previous work has demonstrated that several sites on the extracellular surface of BtuB may be spin labeled in vivo using site-directed cysteines and a standard methanethiosulfonate reagent to produce the side chain R1 (Figure 1A). These included multiple sites on the extracellular loops of BtuB (Nilaweera et al., 2019; Nyenhuis et al., 2020a; Joseph et al., 2016) as well as two sites in the core region (Nilaweera et al., 2019). Recent work has also shown that the efficient incorporation of pairs of spin labels to make distance measurements using DEER required the use of a strain deficient in the disulfide bond formation (Dsb) chaperone system (Nilaweera et al., 2019). We tested several additional single cysteine mutants in BtuB using a DsbA- strain to determine whether spin labeling of additional sites in the core was possible. Shown in Figure 1c are spectra from site 90, which was previously labeled, as wells as four additional sites in the core region. Sites 63 and 65 lie in substrate binding loop 1 (SB1), site 72 lies in substrate binding loop 2 (SB2), and sites 90 and 93 are positioned in SB3. These spectra arise from label having more than one motional component but are dominated by a broad feature that is characteristic of a population of label with hindered motion on the ns time scale, consistent with the confined environment in the extracellular region of the core. For sites 63, 72, and 93, the addition of vitamin B12 alters the spectra and increases the population of the immobile component indicating that incorporation of the label at these sites has not prevented the binding of substrate. No significant changes with substrate are seen for sites 65 and 90. At site 90, substrate does bind (see below) and the lack of a change in the EPR spectrum may reflect the fact that in the apo state the label is already highly immobile. Site 65 is also highly immobile, but we cannot exclude the possibly that incorporation of R1 at this site has blocked the binding of vitamin B12. The apex of the SB3 loop in the BtuB core undergoes a substrate-dependent conformational change For distance measurements using pulse EPR, each set of spin pairs included a label at position 188 on the 3/4 extracellular loop (the second loop connecting β-strands 3 and 4). This site was chosen as a reference point because previous work in whole cells demonstrated that this loop assumed a well-defined position and exhibited minimal or no movement upon substrate addition (Nyenhuis et al., 2020a). All data were analyzed using LongDistances. Positions of both components were held constant throughout and were set to the average of an initial round of fits where distance was varied. Width and area for the two components were allowed to vary freely. Error ranges were taken from the output of the fitting routine. Shown in Figure 2 are the results for measurements on the V90R1-T188R1 spin pair in cells. Preliminary results from this pair were presented in a previous study demonstrating the use of disulfide chaperone mutants to achieve double labeling of BtuB in whole cells (Nilaweera et al., 2019). The background corrected DEER data and resulting distance distributions are shown in Figure 2b,c. Both the apo (blue) and vitamin B12 bound (red) distributions yield two main intramolecular peaks at 2.4 and 3.2 nm, with a substrate-dependent shift observed toward the shorter component. Predicted distance distributions were generated from the apo and vitamin B12 bound in surfo crystal structures of BtuB (PDB IDs: 1NQG and 1NQH) using the program MMM (Jeschke, 2018). The distributions generated from these structures also show a shift toward a shorter distance in the substrate bound state, which is due to the unfolding of a helical turn in the SB3 loop (Chimento et al., 2003). However, the magnitude of the predicted shift is smaller than that observed by DEER. Because the position of site 188 in the 3/4 extracellular loop of BtuB is not altered with substrate addition in situ (Nyenhuis et al., 2020a), this structural change must involve a movement of the SB3 apex or a change in rotamers assumed by R1 at position 90. Figure 2 with 1 supplement see all Download asset Open asset Substrate-dependent shifts are detected at the apex of substrate binding loop 3 (SB3) in whole cells. A side view of BtuB (A) with the locations of site 90 in SB3 of the core domain (yellow) and site 188 near the apex of the 3/4 extracellular loop (light blue). (B) Background corrected double electron-electron resonance (DEER) data for the apo (blue) and substrate bound states (red) of the V90R1-T188R1 spin pair, where the red traces represent the fits to the data. The resulting distance distributions are shown in (C) where the histograms represent predicted distance ranges obtained from the in surfo crystal structures 1NQG (blue) and 1NQH (red) using the software package MMM (Jeschke, 2018). (D) This structural change titrates between two states (labeled 1 and 2) with the addition of vitamin B12. The conversion between states saturates at concentrations above 60 µM. For the distributions shown in (C), data were analyzed using LongDistances v932 using the model-free fitting mode, whereas for distributions in (D), data were fit to a two-Gaussian model where position, width, and amplitude were variables in the fit. Errors in the fit to the distributions in (D) are shown in Supplementary file 1. We also titrated this structural change by measuring the distance distribution with increasing concentrations of the vitamin B12 substrate, where the result is shown in Figure 2D. For this analysis data were processed using a model-based approach with two-Gaussian components where the position, width, and amplitude were varied. As seen in Figure 2D, there is a strong progressive response to increases in substrate until saturation is reached in the range of 30–60 μM vitamin B12. This titration likely reflects substrate loading. Because substrate concentrations greatly exceed the affinity of vitamin B12 to BtuB (Bradbeer et al., 1986), the saturation point likely reflects the concentration of BtuB in our sample and is roughly consistent with the spin concentrations expected from the EPR signal intensity. Substrate-dependent shifts in the extracellular face of the hatch domain are localized to the SB3 loop To determine whether the structural change observed in the apex of SB3 is limited to this site or part of a broader conformational change across the core domain, we tested additional spin pairs on the extracellular face of the protein using the core sites shown in Figure 1B. The spin pairs examined are shown in Figure 3A,B and include site 93, which also lies in SB3, as well as sites 63 and 65 in SB1 and site 72 in SB2. The distance distributions that result from these spin pairs are shown in Figure 3C, along with the V90R1-T188R1 spin pair. It should be noted that the distance distributions in Figure 3, as well as subsequent figures, have been truncated at 5 nm. As seen in the raw data (Figure 2—figure supplement 1 and Figure 3—figure supplement 1), a longer distance component is apparent in some of these dipolar evolutions. This is due to the presence of BtuB-BtuB interactions in the OM that leads to a 6.5 nm distance component (Nyenhuis et al., 2020b). Figure 3 with 4 supplements see all Download asset Open asset Substrate-dependent conformational shifts are limited to substrate binding loop 3 (SB3) and are altered by mutation of the R14-D316 ionic lock. (A) Top view of BtuB (PDB ID: 1NQH) showing the locations of the hatch sites relative to the reference site, 188, in the 3/4 extracellular loop. In (B) the location of the R14-D316 ionic interaction between the core and the barrel is also shown along with the Ton box (red). (C) Distance distributions obtained for hatch: barrel pairs in the apo state (blue) and with substrate (red). (D) Distance distributions obtained for hatch: barrel pairs in the apo state (blue) and with substrate (red) in the presence of the R14A mutation. Both V90R1-T188R1 and S93R1-T188R1 spin pairs show additional distances at 3.8 and 4.5 nm in the presence of R14A. Data were analyzed using LongDistances v932 and the model-free fitting regime. The shaded error bands in (C) and (D) represent variation due to background noise, start time, dimensionality, and regularization. Histograms are predicted distances generated from the in surfo crystal structures PDB ID: 1NQG (light blue) and PDB ID: 1NQH (light red) using the software package MMM (Jeschke, 2018). As seen in Figure 3C, distance distributions for sites located outside SB3 show little evidence for any structural change upon the addition of substrate, with spin pairs involving sites 63, 65, and 72 having nearly identical distributions for both apo and vitamin B12 bound conditions. This lack of a shift in these spin pairs is consistent with work showing that site 188 does not show a substrate-dependent shift in its position in situ (Nyenhuis et al., 2020a). In the SB3 loop, however, site 93 at the edge of the loop shows a substrate-dependent change in position along with site 90 at the loop apex, although the change is much larger for site 90 (about 8 Å) than for site 93 (about 4 Å) and in an opposite direction. The structural changes measured by DEER are qualitatively consistent with the predictions from the crystal structures, which show a loss in helical structure and change in position of the loop. For the sites examined here, substrate-dependent changes in the core appear to be confined to the region of the SB3 loop. Interestingly, as we demonstrate below, this substrate-dependent change does not occur when the protein is removed from its native environment. Breaking an ionic lock between the BtuB core and barrel triggers a large substrate-dependent change in SB3 As indicated above, BtuB is expected to bind to both substrate and TonB during transport, which will break the ionic lock between R14 in the core and D316 in the barrel (Shultis et al., 2006). However, under the conditions of our experiment, BtuB is in large excess relative to TonB, perhaps by a factor of 10–20 or more. As a result, only a small portion of the BtuB would be bound to TonB at any time during our distance measurement. To determine whether there might be a connection between the structure of the core and this internal ionic lock, we examined the effect of disrupting the R14-D316 interaction on the core by introducing the R14A mutation into the existing pairs of labels between site 188 in the 3/4 extracellular loop and the core. Distance measurements for the apo and vitamin B12 bound states made in the presence of this mutation are shown in Figure 3D. For distance distributions involving sites 63, 65, and 72 in SB1 and SB2, the core remains largely unchanged in response to substrate and unchanged by the R14A mutation. However, distance distributions involving sites 90 and 93 in SB3 are altered by breaking the D316/R14 ionic lock. For distances measured to sites 90 and 93, the R14A mutation has two main effects. First, for the V90R1-T188R1 pair, the substrate-dependent conversion between the 2.4 and 3.2 nm distance components is absent, and the shorter distance now dominates in both apo and vitamin B12 bound conditions. Second, for both the V90R1-T188R1 and S93R1-T188R1 pairs, additional distance components are observed in the presence of substrate centered at 3.8 and 4.5 nm. These long components result in a distribution that is substantially broader than that predicted by the crystal structures, and they indicate the formation of a novel conformation of the SB3 loop and an altered substrate binding mode. The longer distance components that appear for the V90R1-T188R1 and S93R1-T188R1 spin pairs represent a substantial movement of the SB3 loop toward the periplasmic side of BtuB. A movement of SB3 toward the extracellular surface is highly unlikely, in part because movement in this direction would require a major unfolding of the core that we do not observe. For measurements to SB3, there were relatively few positions that were both accessible and within the range of the pulse EPR measurement, but we made measurements to site 90 from site 237, which is located near the apex of the 5/6 extracellular loop. In the apo and vitamin B12 bound states, the predicted distances from this site are shorter than 2 nm and are not within a range that can accurately be measured by DEER. The results are shown in Figure 3—figure supplement 2. In the absence of the R14A mutation, no clear substrate-dependent shift in the position of SB3 is observed, which is likely due to the short distance involved. However, in the presence of the R14A mutation, a new distance component appears with substrate addition around 3 nm that is beyond the distance range predicted by the crystal structure. This is shorter than the 4.5 nm observed from position 188, which likely reflects differences in the side chain direction and the relative positions of the 3/4 and 5/6 loops. These ionic lock mutations generate structural states that are not seen in the wild-type protein. However, disrupting the ionic lock does not appear to abolish transport. The BtuB R14A mutant is functional in transport as determined by a growth assay (Figure 3—figure supplement 3). This is consistent with earlier work where a mutant containing a dipeptide insertion into one of the barrel strands in BtuB, which should have disrupted the D316-R14A ionic lock, was shown to support transport of vitamin B12 (Lathrop et al., 1995). It should be noted that we examined the stability of the substrate-induced conformation of SB3 for times as long as 60 min before freezing and preparing the cells for DEER. The data are shown in Figure 3—figure supplement 4, and indicate that the conformations are stable over time, indicating that the label is stable and not being reduced, and that conversion of the transporter back to the apo state does not occur under the conditions of this experiment. Substrate-dependent changes in the SB3 loop require a native environment Two conformations are observed by crystallography for SB3. In the apo structure (PDB ID: 1NQG), SB3 has a single helical turn and shorter conformation that we speculate may be associated with the distance observed by EPR at 3.2 nm for the V90R1-T188R1 spin pair (labeled 1 in Figure 2D). In the vitamin B12 bound structure (PDB ID: 1NQH), SB3 is more extended, and this state may be associated with the distance at 2.4 nm (labeled 2 in Figure 2D). Computational work suggests that the extended state of SB3 requires the interaction of BtuB with LPS, whereas the shorter helical state of SB3 occurs in the presence of phospholipid (Balusek and Gumbart, 2016), suggesting that environment, specifically LPS, may be important in controlling the configuration of SB3. To test for an environmental effect on SB3, we reconstituted four spin pairs of BtuB into 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine (POPC) proteoliposomes. These included the V90R1-T188R1 and S93R1-T188R1 spin pairs both in the absence and presence of the R14A mutation. The dipolar evolution data and distance distributions from DEER measurements on these reconstituted BtuB samples are shown in Figure 4 (the raw time domain data are presented in Figure 4—figure supplement 1). In the absence of the R14A mutation, the substrate-dependent conversion to the shorter distance that was seen for the V90R1-T188R1 spin pair in whole cells (Figure 2C and Figure 3C) is now much more limited, with only a minor shoulder appearing around 2.5 nm. For the S93R1-T188R1 spin pair, the change seen in Figure 3C with substrate is largely absent. The 0.4 nm substrate-dependent shift to a longer distance is absent, but the small 2.2 nm distance component is still present. Figure 4 with 1 supplement see all Download asset Open asset The substrate-induced changes in substrate binding loop 3 (SB3) are altered or absent in proteoliposomes. Background corrected double electron-electron resonance (DEER) signals (left) and distance distributions (right) for the V90R1-T188R1 and S93R1-T188R1 pairs involving SB3 in the absence and presence of the R14A mutation, where the labeled BtuB have been reconstituted into 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine (POPC) vesicles. Data were analyzed using LongDistances v932 and the model-free fitting mode. The shaded error bands represent variation due to background noise, start time, dimensionality, and regularization. Histograms are predicted distances generated from the in surfo crystal structures 1NQG (blue) and 1NQH (pink) using the software package MMM (Jeschke, 2018). The behavior of SB3 in the presence of the R14A mutation is also altered in the phospholipid reconstituted system. Rather than increasing the population of the shorter distance component, adding the R14A mutation to the V90R1-T188R1 pair in the reconstituted system results in a single short distance, where the resulting distribution aligns almost perfectly with the predicted distribution from the 1NQH structure. For the S93R1-T188R1 pair, R14A causes a small increase in the short distance component at 2.2 nm. But significantly, for neither spin pair are the longer substrate-induced shifts that were seen in Figure 3D observed in the reconstituted system. Thus, when removed from the native OM environment, the structure of SB3 is altered and the large substrate-induced movement of SB3 toward the periplasmic surface in the presence of the R14A mutation is no longer seen. It should be noted that in our initial work on the V90R1-T188R1 spin pair, we failed to observe a substrate-dependent conformational change in SB3 using an isolated OM preparation where the preparation includes a sarkosyl treatment (Nilaweera et al., 2019). This suggests that this detergent treatment of the OM to remove inner membrane components is sufficient to alter the behavior of BtuB. These observations provide an explanation for why these changes in the conformation of SB3 have not been previously observed. Mutating either R14, D316, or both alters SB3 conformations and populates a state where SB3 is moved toward the periplasmic surface In earlier work, we demonstrated that breaking the ionic lock between D316 and R14 altered a conformational equilibrium in the Ton box and promoted its unfolded state (Lukasik et al., 2007). In this work, the effect of the R14A mutation on the Ton box equilibrium was comparable to that of a D3
