Conformational and dynamics changes induced by bile acids binding to chicken liver bile acid binding protein

The correlation between protein motions and function is a central problem in protein science. Several studies have demonstrated that ligand binding and protein dynamics are strongly correlated in intracellular lipid binding proteins (iLBPs), in which the high degree of flexibility, principally occurring at the level of helix-II, CD, and EF loops (the so-called portal area), is significantly reduced upon ligand binding. We have recently investigated by NMR the dynamic properties of a member of the iLBP family, chicken liver bile acid binding protein (cL-BABP), in its apo and holo form, as a complex with two bile salts molecules. Binding was found to be regulated by a dynamic process and a conformational rearrangement was associated with this event. We report here the results of molecular dynamics (MD) simulations performed on apo and holo cL-BABP with the aim of further characterizing the protein regions involved in motion propagation and of evaluating the main molecular interactions stabilizing bound ligands. Upon binding, the root mean square fluctuation values substantially decrease for CD and EF loops while increase for the helix-loop-helix region, thus indicating that the portal area is the region mostly affected by complex formation. These results nicely correlate with backbone dynamics data derived from NMR experiments. Essential dynamics analysis of the MD trajectories indicates that the major concerted motions involve the three contiguous structural elements of the portal area, which however are dynamically coupled in different ways whether in the presence or in the absence of the ligands. Motions of the EF loop and of the helical region are part of the essential space of both apo and holo-BABP and sample a much wider conformational space in the apo form. Together with NMR results, these data support the view that, in the apo protein, the flexible EF loop visits many conformational states including those typical of the holo state and that the ligand acts stabilizing one of these pre-existing conformations. The present results, in agreement with data reported for other iLBPs, sharpen our knowledge on the binding mechanism for this protein family.     

Calcium coordination studies of the metastatic Mts1 protein

Mts1, also known as S100A4, is an 11 kDa calcium-binding protein strongly linked to metastasis. As a member of the S100 protein family, Mts1 is predicted to contain four alpha-helices and two calcium-binding loops, the second of which forms a canonical EF hand, while the first is a pseudo-EF hand, using two extra residues and principally backbone carbonyls rather than side chain oxygens to coordinate calcium. Here we follow chemical shift changes which occur in Mts1 upon titration of calcium. The results are consistent with calcium coordination by the EF hands described above. Filling of the first (pseudo) EF hand occurs at a lower calcium concentration than does filling of the second (canonical) EF hand. Concurrent with filling of site I, resonances from much of helix 4 vanish while the chemical shifts of a possibly nascent helical segment immediately C-terminal to helix 4 increase in helical character. Other smaller changes are seen, including a change in the linker joining helix 2 and helix 3. Since binding of effector molecules to S100 proteins has been shown to involve the C-terminus and linker regions, these calcium-induced changes have implications for the role of Mts1 in metastasis.     

Circular permutation and deletion studies of myoglobin indicate that the correct position of its N-terminus is required for native stability and solubility but not for native-like heme binding and folding

We studied the effect of deleted and circularly permuted mutations in sperm whale myoglobin and present here results on three classes of mutants: (i) a deletion mutant, Mb(1)(-)(99), in which the C-terminal helices, G and H, were removed; (ii) two circular permutations, Mb-B_GHA, in which helix B is N-terminal and helix A is C-terminal, and Mb-C_GHAB, in which helix C is N-terminal and helices A and B are C-terminal; and (iii) a deleted circular permutation, Mb-HAB_F, in which helix H is N-terminal, helix F is C-terminal, and helix G is deleted. The conformational characteristics of the apo and holo forms of these mutants were determined at neutral pH, by spectroscopic and hydrodynamic methods. The apo form of the deleted and permuted mutants exhibited a stronger tendency to aggregate and had lower ellipticity than the wild type. The mutants retained the ability to bind heme, but only the circularly permuted holoproteins had native-like heme binding and folding. These results agree with the theory that myoglobin has a central core that is able to bind heme, but also indicate that the presence of N- and C-terminal helices is necessary for native-like heme pocket formation. Because the holopermuteins were less stable than the wild-type protein and aggregated, we propose that the native position of the N-terminus is important for the precise structural architecture of myoglobin.     

Close approximation of putative alpha -helices II, IV, VII, X, and XI in the translocation pathway of the lactose transport protein of Streptococcus thermophilus

The lactose transport protein (LacS) of Streptococcus thermophilus belongs to a family of transporters in which putative alpha-helices II and IV have been implicated in cation binding and the coupled transport of the substrate and the cation. Here, the analysis of site-directed mutants shows that a positive and negative charge at positions 64 and 71 in helix II are essential for transport, but not for lactose binding. The conservation of charge/side-chain properties is less critical for Glu-67 and Ile-70 in helix II, and Asp-133 and Lys-139 in helix IV, but these residues are important for the coupled transport of lactose together with a proton. The analysis of second-site suppressor mutants indicates an ion pair exists between helices II and IV, and thus a close approximation of these helices can be made. The second-site suppressor analysis also suggests ion pairing between helix II and the intracellular loops 6-7 and 10-11. Because the C-terminal region of the transmembrane domain, especially helix XI and loop 10-11, is important for substrate binding in this family of proteins, we propose that sugar and proton binding and translocation are performed by the joint action of these regions in the protein. Indeed, substrate protection of maleimide labeling of single cysteine mutants confirms that alpha-helices II and IV are directly interacting or at least conformationally involved in sugar binding and/or translocation. On the basis of new and published data, we reason that the helices II, IV, VII, X, and XI and the intracellular loops 6-7 and 10-11 are in close proximity and form the binding sites and/or the translocation pathway in the transporters of the galactosides-pentosides-hexuronides family.     

Binding of retinol induces changes in rat cellular retinol-binding protein II conformation and backbone dynamics

The structure and backbone dynamics of rat holo cellular retinol-binding protein II (holo-CRBP II) in solution has been determined by multidimensional NMR. The final structure ensemble was based on 3980 distance and 30 dihedral angle restraints, and was calculated using metric matrix distance geometry with pairwise Gaussian metrization followed by simulated annealing. The average RMS deviation of the backbone atoms for the final 25 structures relative to their mean coordinates is 0.85(+/-0.09) A. Comparison of the solution structure of holo-CRBP II with apo-CRBP II indicates that the protein undergoes conformational changes not previously observed in crystalline CRBP II, affecting residues 28-35 of the helix-turn-helix, residues 37-38 of the subsequent linker, as well as the beta-hairpin C-D, E-F and G-H loops. The bound retinol is completely buried inside the binding cavity and oriented as in the crystal structure. The order parameters derived from the (15)N T(1), T(2) and steady-state NOE parameters show that the backbone dynamics of holo-CRBP II is restricted throughout the polypeptide. The T(2) derived apparent backbone exchange rate and amide (1)H exchange rate both indicate that the microsecond to second timescale conformational exchange occurring in the portal region of the apo form has been suppressed in the holo form.     

High resolution solution structure of apo calcyclin and structural variations in the S100 family of calcium-binding proteins

The three-dimensional solution structure of apo rabbit lung calcyclin has been refined to high resolution through the use of heteronuclear NMR spectroscopy and 13C,15N- enriched protein. Upon completing the assignment of virtually all of the 15N, 13C and 1H NMR resonances, the solution structure was determined from a combination of 2814 NOE- derived distance constraints, and 272 torsion angle constraints derived from scalar couplings. A large number of critical inter- subunit NOEs (386) were identified from 13C- select,13C-filtered NOESY experiments, providing a highly accurate dimer interface. The combination of distance geometry and restrained molecular dynamics calculations yielded structures with excellent agreement with the experimental data and high precision (rmsd from the mean for the backbone atoms in the eight helices: 0.33 Å). Calcyclin exhibits a symmetric dimeric fold of two identical 90 amino acid subunits, characteristic of the S100 subfamily of EF-hand Ca2+-binding proteins. The structure reveals a readily identified pair of putative sites for binding of Zn2+. In order to accurately determine the structural features that differentiate the various S100 proteins, distance difference matrices and contact maps were calculated for the NMR structural ensembles of apo calcyclin and rat and bovine S100B. These data show that the most significant variations among the structures are in the positioning of helix III and in loops, the regions with least sequence similarity. Inter-helical angles and distance differences for the proteins show that the positioning of helix III of calcyclin is most similar to that of bovine S100B, but that the helix interfaces are more closely packed in calcyclin than in either S100B structure. Surprisingly large differences were found in the positioning of helix III in the two S100B structures, despite there being only four non-identical residues, suggesting that one or both of the S100B structures requires further refinement.     

Two homologous rat cellular retinol-binding proteins differ in local conformational flexibility

Cellular retinol-binding protein I (CRBP I) and cellular retinol-binding protein II (CRBP II) are closely homologous proteins that play distinct roles in the maintenance of vitamin A homeostasis. The solution structure and dynamics of CRBP I and CRBP II were compared by multidimensional NMR techniques. These studies indicated that differences in the mean backbone structures of CRBP I and CRBP II were localized primarily to the alphaII helix. Intraligand NOE cross-peaks were detected for the hydroxyl proton in the NOESY spectrum of CRBP I-bound retinol, but not for CRBP II-bound retinol, indicating that the conformational dynamics of retinol binding are different for these two proteins. As determined by Lipari-Szabo formalism, both the apo and holo forms of CRBP I and CRBP II are conformationally rigid on the pico- to nanosecond timescale. transverse relaxation optimized spectroscopy-Carr-Purcell-Meiboom-Gill -based 15N relaxation dispersion experiments at both 500 MHz and 600 MHz magnetic fields revealed that 84 and 62 residues for apo-CRBP I and II, respectively, showed detectable conformational exchange on a micro- to millisecond timescale, in contrast to three and seven residues for holo-CRBP I and II, respectively. Thus binding of retinol markedly reduced conformational flexibility in both CRBP I and CRBP II on the micro- to millisecond timescale. The 15N relaxation dispersion curves of apo-CRBP I and II were fit to a two-state conformational exchange model by a global iterative fitting process and by an individual (residue) fitting process. In the process of carrying out the global fit, more than half of the residue sites were eliminated. The individual chemical exchange rates k(ex), and chemical shift differences, Deltadelta, were increased in the putative portal region (alphaII helix and betaC-betaD turn) of apo-CRBP II compared to apo-CRBP I. These differences in conformational flexibility likely contribute to differences in how CRBP I and CRBP II interact with ligands, membranes and retinoid metabolizing enzymes.     

Differential backbone dynamics of companion helices in the extended helical coiled‐coil domain of a bacterial chemoreceptor

Cytoplasmic domains of transmembrane bacterial chemoreceptors are largely extended four‐helix coiled coils. Previous observations suggested the domain was structurally dynamic. We probed directly backbone dynamics of this domain of the transmembrane chemoreceptor Tar from Escherichia coli using site‐directed spin labeling and electron paramagnetic resonance (EPR) spectroscopy. Spin labels were positioned on solvent‐exposed helical faces because EPR spectra for such positions reflect primarily polypeptide backbone movements. We acquired spectra for spin‐labeled, intact receptor homodimers solubilized in detergent or inserted into native E. coli lipid bilayers in Nanodiscs, characterizing 16 positions distributed throughout the cytoplasmic domain and on both helices of its helical hairpins, one amino terminal to the membrane‐distal tight turn (N‐helix), and the other carboxyl terminal (C‐helix). Detergent solubilization increased backbone dynamics for much of the domain, suggesting that loss of receptor activities upon solubilization reflects wide‐spread destabilization. For receptors in either condition, we observed an unanticipated difference between the N‐ and C‐helices. For bilayer‐inserted receptors, EPR spectra from sites in the membrane‐distal protein‐interaction region and throughout the C‐helix were typical of well‐structured helices. In contrast, for approximately two‐thirds of the N‐helix, from its origin as the AS‐2 helix of the membrane‐proximal HAMP domain to the beginning of the membrane‐distal protein‐interaction region, spectra had a significantly mobile component, estimated by spectral deconvolution to average approximately 15%. Differential helical dynamics suggests a four‐helix bundle organization with a pair of core scaffold helices and two more dynamic partner helices. This newly observed feature of chemoreceptor structure could be involved in receptor function.     

Crosslinking of transcription factor TFIIIA to ribosomal 5S RNA from X. laevis by trans-diamminedichloroplatinum (II)

Trans-diamminnedichloroplatinum (II) was used to induce reversible crosslinks between 5S rRNA and TFIIIA within the 7S RNP particle from X. laevis immature oocyte. The crosslinked fragments have been unambiguously identified. These fragments exclusively arise from three RNA regions centered around the hinge region at the junction of the three helical domains. Major crosslinking sites are located in region 9-21 (comprising loops A and helix II) and region 54-71 (comprising loop B, helices II and V). A minor site is also found in the 3' part of helix I and helix V (region 100-120). Our results point to the crucial role of the junction region and of the three-dimensional folding of the RNA in the recognition of the 5S rRNA by TFIIIA.     

High-resolution prediction of protein helix positions and orientations

We have developed a new method for predicting helix positions in globular proteins that is intended primarily for comparative modeling and other applications where high precision is required. Unlike helix packing algorithms designed for ab initio folding, we assume that knowledge is available about the qualitative placement of all helices. However, even among homologous proteins, the corresponding helices can demonstrate substantial differences in positions and orientations, and for this reason, improperly positioned helices can contribute significantly to the overall backbone root-mean-square deviation (RMSD) of comparative models. A helix packing algorithm for use in comparative modeling must obtain high precision to be useful, and for this reason we utilize an all-atom protein force field (OPLS) and a Generalized Born continuum solvent model. To reduce the computational expense associated with using a detailed, physics-based energy function, we have developed new hierarchical and multiscale algorithms for sampling the helices and flanking loops. We validate the method using a test suite of 33 cases, which are drawn from a diverse set of high-resolution crystal structures. The helix positions are reproduced with an average backbone RMSD of 0.6 A, while the average backbone RMSD of the complete loop-helix-loop region (i.e., the helix with the surrounding loops, which are also repredicted) is 1.3 A.     

Helix lap-joints as ion-binding sites: DNA-binding motifs and Ca-binding "EF hands" are related by charge and sequence reversal

The DNA-binding helix pairs in gene repressor and activator proteins were compared with other approximately perpendicular pairs of adjacent helices in the known protein structures. Two other examples of closely matching conformations were found in cytochrome c peroxidase (residues 153-174) and in ribosomal L7/L12 protein (residues 68-89). Another group of such offset "lap-joints" are the Ca-binding "EF hand" structures, which bind a positive rather than a negative ligand. The EF hands turn out to match the DNA-binding motifs quite well (outside of the loop) if their sequence direction is reversed. This conformation is thus not as unusual as had been thought, but may have a more generalized role in ion binding and occasionally occur in a purely structural role.     

Molecular Basis of Calcium-Sensitizing and Desensitizing Mutations of the Human Cardiac Troponin C Regulatory Domain: A Multi-Scale Simulation Study

Troponin C (TnC) is implicated in the initiation of myocyte contraction via binding of cytosolic and subsequent recognition of the Troponin I switch peptide. Mutations of the cardiac TnC N-terminal regulatory domain have been shown to alter both calcium binding and myofilament force generation. We have performed molecular dynamics simulations of engineered TnC variants that increase or decrease sensitivity, in order to understand the structural basis of their impact on TnC function. We will use the distinction for mutants that are associated with increased affinity and for those mutants with reduced affinity. Our studies demonstrate that for GOF mutants V44Q and L48Q, the structure of the physiologically-active site II binding site in the -free (apo) state closely resembled the -bound (holo) state. In contrast, site II is very labile for LOF mutants E40A and V79Q in the apo form and bears little resemblance with the holo conformation. We hypothesize that these phenomena contribute to the increased association rate, , for the GOF mutants relative to LOF. Furthermore, we observe significant positive and negative positional correlations between helices in the GOF holo mutants that are not found in the LOF mutants. We anticipate these correlations may contribute either directly to affinity or indirectly through TnI association. Our observations based on the structure and dynamics of mutant TnC provide rationale for binding trends observed in GOF and LOF mutants and will guide the development of inotropic drugs that target TnC.     

Internal dynamics of the tryptophan repressor (TrpR) and two functionally distinct TrpR variants, L75F-TrpR and A77V-TrpR, in their l-Trp-bound forms

Backbone amide dynamics of the Escherichia coli tryptophan repressor protein (WT-TrpR) and two functionally distinct variants, L75F-TrpR and A77V-TrpR, in their holo (l-tryptophan corepressor-bound) form have been characterized using (15)N nuclear magnetic resonance (NMR) relaxation. The three proteins possess very similar structures, ruling out major conformational differences as the source of their functional differences, and suggest that changes in protein flexibility are at the origin of their distinct functional properties. Comparison of site specific (15)N-T(1), (15)N-T(2), (15)N-{(1)H} nuclear Overhauser effect, reduced spectral density, and generalized order (S(2)) parameters indicates that backbone dynamics in the three holo-repressors are overall very similar with a few notable and significant exceptions for backbone atoms residing within the proteins' DNA-binding domain. We find that flexibility is highly restricted for amides in core α-helices (i.e., helices A-C and F), and a comparable "stiffening" is observed for residues in the DNA recognition helix (helix E) of the helix D-turn-helix E (HTH) DNA-binding domain of the three holo-repressors. Unexpectedly, amides located in helix D and in adjacent turn regions remain flexible. These data support the concept that residual flexibility in TrpR is essential for repressor function, DNA binding, and molecular recognition of target operators. Comparison of the (15)N NMR relaxation parameters of the holo-TrpRs with those of the apo-TrpRs indicates that the single-point amino acid substitutions, L75F and A77V, perturb the flexibility of backbone amides of TrpR in very different ways and are most pronounced in the apo forms of the three repressors. Finally, we present these findings in the context of other DNA-binding proteins and the role of protein flexibility in molecular recognition.     

Raman optical activity characterization of native and molten globule states of equine lysozyme: comparison with hen lysozyme and bovine alpha-lactalbumin

Vibrational Raman optical activity (ROA) spectra of the calcium-binding lysozyme from equine milk in native and nonnative states are measured and compared with those of the homologous proteins hen egg white lysozyme and bovine alpha-lactalbumin. The ROA spectrum of holo equine lysozyme at pH 4.6 and 22 degrees C closely resembles that of hen lysozyme in regions sensitive to backbone and side chain conformations, indicating similarity of the overall secondary and tertiary structures. However, the intensity of a strong positive ROA band at approximately 1340 cm(-1), which is assigned to a hydrated form of alpha helix, is more similar to that in the ROA spectrum of bovine alpha-lactalbumin than hen lysozyme and may be associated with the greater flexibility and calcium-binding ability of equine lysozyme and bovine alpha-lactalbumin compared with hen lysozyme. In place of a strong sharp positive ROA band at approximately 1300 cm(-1) in hen lysozyme that is assigned to an alpha helix in a more hydrophobic environment, equine lysozyme shows a broader band centered at approximately 1305 cm(-1), which may reflect greater heterogeneity in some alpha-helical sequences. The ROA spectrum of apo equine lysozyme at pH 4.6 and 22 degrees C is almost identical to that of the holo protein, which indicates that loss of calcium has little influence on the backbone and side chain conformations, including the calcium-binding loop. From the similarity of their ROA spectra, the A state at pH 1.9 and both 2 and 22 degrees C and the apo form at pH 4.5 and 48 degrees C, which are partially folded denatured (molten globule or state A) forms of equine lysozyme, have similar structures that the ROA suggests contain much hydrated alpha helix. The A state of equine lysozyme is shown by these results to be more highly ordered than that of bovine alpha-lactalbumin, the ROA spectrum of which has more features characteristic of disordered states. A positive tryptophan ROA band at approximately 1551 cm(-1) in the native holo protein disappears in the A state, which is probably due to the presence of nonnative conformations of the tryptophans associated with a previously identified cluster of hydrophobic residues.     

Conformational and sequence signatures in beta helix proteins

beta helix proteins are characterized by a repetitive fold, in which the repeating unit is a beta-helical coil formed by three strand segments linked by three loop segments. Using a data set of left- and right-handed beta helix proteins, we have examined conformational features at equivalent positions in successive coils. This has provided insights into the conformational rules that the proteins employ to fold into beta helices. Left-handed beta helices attain their equilateral prism fold by incorporating "corners" with the conformational sequence P(II)-P(II)-alpha(L)-P(II), which imposes sequence restrictions, resulting in the first and third P(II) residues often being G and a small, uncharged residue (V, A, S, T, C), respectively. Right-handed beta helices feature mid-sized loops (4, 5, or 6 residues) of conserved conformation, but not of conserved sequence; they also display an alpha-helical residue at the C-terminal end of L2 loops. Backbone conformational parameters (phi,psi) that permit the formation of continuous, loopless beta helices (Perutz nanotubes) have also been investigated.     

Characterization of the deoxynucleotide triphosphate triphosphohydrolase (dNTPase) activity of the EF1143 protein from Enterococcus faecalis and crystal structure of the activator-substrate complex

The EF1143 protein from Enterococcus faecalis is a distant homolog of deoxynucleotide triphosphate triphosphohydrolases (dNTPases) from Escherichia coli and Thermus thermophilus. These dNTPases are important components in the regulation of the dNTP pool in bacteria. Biochemical assays of the EF1143 dNTPase activity demonstrated nonspecific hydrolysis of all canonical dNTPs in the presence of Mn(2+). In contrast, with Mg(2+) hydrolysis required the presence of dGTP as an effector, activating the degradation of dATP and dCTP with dGTP also being consumed in the reaction with dATP. The crystal structure of EF1143 and dynamic light scattering measurements in solution revealed a tetrameric oligomer as the most probable biologically active unit. The tetramer contains four dGTP specific allosteric regulatory sites and four active sites. Examination of the active site with the dATP substrate suggests an in-line nucleophilic attack on the α-phosphate center as a possible mechanism of the hydrolysis and two highly conserved residues, His-129 and Glu-122, as an acid-base catalytic dyad. Structural differences between EF1143 apo and holo forms revealed mobility of the α3 helix that can regulate the size of the active site binding pocket and could be stabilized in the open conformation upon formation of the tetramer and dGTP effector binding.     

Kinetics for exchange of imino protons in the d(C-G-C-G-A-A-T-T-C-G-C-G) double helix and in two similar helices that contain a G . T base pair, d(C-G-T-G-A-A-T-T-C-G-C-G), and an extra adenine, d(C-G-C-A-G-A-A-T-T-C-G-C-G)

The relaxation lifetimes of imino protons from individual base pairs were measured in (I) a perfect helix, d(C-G-C-G-A-A-T-T-C-G-C-G), (II) this helix with a G . C base pair replaced with a G . T base pair, d(C-G-T-G-A-A-T-T-C-G-C-G), and (III) the perfect helix with an extra adenine base in a mismatch, d(C-G-C-A-G-A-A-T-T-C-G-C-G). The lifetimes were measured by saturation recovery proton nuclear magnetic resonance experiments performed on the imino protons of these duplexes. The measured lifetimes of the imino protons were shown to correspond to chemical exchange lifetimes at higher temperatures and spin-lattice relaxation times at lower temperatures. Comparison of the lifetimes in these duplexes showed that the destabilizing effect of the G . T base pair in II affected the opening rate of only the nearest-neighbor base pairs. For helix III, the extra adenine affected the opening rates of all the base pairs in the helix and thus was a larger perturbation for opening of the base pairs than the G . T base pair. The temperature dependence of the exchange rates of the imino proton in the perfect helix gives values of 14-15 kcal/mol for activation energies of A . T imino protons. These relaxation rates were shown to correspond to exchange involving individual base pair opening in this helix, which means that one base-paired imino proton can exchange independent of the others. For the other two helices that contain perturbations, much larger activation energies for exchange of the imino protons were found, indicating that a cooperative transition involving exchange of at least several base pairs was the exchange mechanism of the imino protons. The effects of a perturbation in a helix on the exchange rates and the mechanisms for exchange of imino protons from oligonucleotide helices are discussed.     

The proximity between helix I and helix XI in the melibiose carrier of Escherichia coli as determined by cross-linking

The melibiose carrier of Escherichia coli is a transmembrane protein that comprises 12 transmembrane helices connected by periplasmic and cytoplasmic loops, with both the N- and C-termini located on the cytoplasmic side. Our previous studies of second-site revertants suggested proximity between several helices, including helices XI and I. In this study, we constructed six double cysteine mutants, each having one cysteine in helix I and the other in helix XI: three mutants, K18C/S380C, D19C/S380C, and F20C/S380C, have their cysteine pairs near the cytoplasmic side of the carrier, and the other three, T34C/G395C, D35C/G395C, and V36C/G395C, have their cysteine pairs near the periplasmic side. In the absence of substrate, disulfide formations catalyzed by iodine and copper-(1,10-phenanthroline)(3) indicate that helix I and helix XI are in immediate proximity to each other on the periplasmic side but not on the cytoplasmic side, as shown by protease cleavage analyses. We infer that the two helices are tilted with respect to each other, with the periplasmic sides in close proximity.