Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2023 Feb 20;63(5):1592–1601. doi: 10.1021/acs.jcim.3c00072

Modeling and Simulation of Bacterial Outer Membranes with Lipopolysaccharides and Capsular Polysaccharides

Ya Gao †,‡, Göran Widmalm §,*, Wonpil Im ‡,*
PMCID: PMC10249353  PMID: 36802606

Abstract

graphic file with name ci3c00072_0009.jpg

Capsule is one of the common virulence factors in Gram-negative bacteria protecting pathogens from host defenses and consists of long-chain capsular polysaccharides (CPS) anchored in the outer membrane (OM). Elucidating structural properties of CPS is important to understand its biological functions as well as the OM properties. However, the outer leaflet of the OM in current simulation studies is represented exclusively by LPS due to the complexity and diversity of CPS. In this work, representative Escherichia coli CPS, KLPS (a lipid A-linked form) and KPG (a phosphatidylglycerol-linked form), are modeled and incorporated into various symmetric bilayers with co-existing LPS in different ratios. All-atom molecular dynamics simulations of these systems have been conducted to characterize various bilayer properties. Incorporation of KLPS makes the acyl chains of LPS more rigid and ordered, while incorporation of KPG makes them less ordered and flexible. These results are consistent with the calculated area per lipid (APL) of LPS, in which the APL of LPS becomes smaller when KLPS is incorporated, whereas it gets larger when KPG is included. Torsional analysis reveals that the influence of the CPS presence on the conformational distributions of the glycosidic linkages of LPS is small, and minor differences are also detected for the inner and outer regions of the CPS. Combined with previously modeled enterobacterial common antigens (ECAs) in the form of mixed bilayers, this work provides more realistic OM models as well as the basis for characterization of interactions between the OM and OM proteins.

Introduction

The outer membrane (OM) of Gram-negative bacteria is a unique and asymmetric lipid bilayer that has phospholipids in the inner leaflet and mostly lipopolysaccharides (LPS) in the outer leaflet.1,2 Capsule is one of the common virulence factors in Gram-negative bacteria and consists of long-chain capsular polysaccharides (CPS) anchored in the OM.3,4 CPS play a vital role in evading host defenses by preventing phagocytosis and complement-mediated killing, and thus represent an attractive therapeutic target.5,6 In Gram-negative bacteria, CPS are synthesized and translocated via one of two widespread assembly systems: Wzy-dependent or the ATP-binding cassette (ABC) transporter-dependent pathway.7,8 Both mechanisms have been well studied in Escherichia coli (E. coli), and there are more than 80 different CPS types identified so far, referred to as the K-antigens.9 The E. coli capsule systems have been categorized into four groups based on the assembly system, organization of key genes, and regulatory features.3 The Wzy-dependent pathway is used for groups 1 and 4 capsules attached to lipid A (KLPS), which are often found in isolates causing gastrointestinal disease. Capsules attached to a phosphatidylglycerol (PG) via alternating β-linked 2,4- and 2,7-Kdo (2-keto-3-deoxyoctulosonate) residues belong to groups 2 and 3 that use an ABC transporter-dependent assembly system, and these bacteria usually cause a variety of diseases in humans, including septicemia, meningitis, and urinary tract infections.10,11 Therefore, modeling and simulation of CPS in Gram-negative bacteria at the atomic and molecular level are critical to understand its biological functions as well as OM properties.

In most simulation studies, the outer leaflet of the OM is represented exclusively by LPS,12−23 and the OMs including CPS are rarely modeled and simulated due to the complexity and diversity of CPS. In our previous work, ECAs in E. coli were modeled and incorporated into the OM for all-atom molecular dynamics (MD) simulations.24 In this work, for groups 1 and 4, E. coli K-antigens K30 and K43 that are co-expressed with O-antigen LPS (O9:K3025,26 and O8:K4327,28) in the same serological group are selected as the representative examples. K30 CPS consists of →2-α-Manp-(1→3)-β-Galp-(1→ chains carrying a β-GlcpA-(1→3)-α-Galp-(1→ branch at position 3 of the mannose, and K43 CPS is composed of →4)-α-GlcpA-(1→3)-β-Manp-(1→4)-β-Manp-(1→3)-β-GlcpNAc-(1→ with a β-GlcpNAc-(1→ branch at position 2 of the second mannosyl residue whose position 3 is linked to GlcpA (Figure 1). The E. coli K129 and K530 glycolipids belong to group 2, and CPS in this group can be co-expressed with many different types of LPS, herein chosen to co-exist with the O164 antigen.31 K129 and K530 have been the prototypes for studying CPS assembly via ABC transporter-dependent pathways.3 K1 CPS was the first full polysaccharide gene cluster cloned,32 and it consists of a homopolymer of α-(2→8)-linked N-acetylneuraminic acid (polysialic acid; PSA), and K5 is composed of a heparosan-like glycan containing glucuronic acid (GlcA) and N-acetylglucosamine (GlcNAc) (Figure 1). A lyso-PG (lyso-PG) moiety with an alternating β-linked 2,4- and 2,7-Kdo linker was originally considered to attach to the reducing terminus of K1 and K5 CPS.33 In 2019, Whitfield et al. proposed that species containing lyso-PG are likely an artifact introduced during the isolation and characterization, and the diacyl form is the correct acceptor.34 Moreover, K1 and K5 differ by possessing either odd (K1) or even (K5) numbers of β-Kdo residues, meaning that K1 has a terminal (2→4)-linked β-Kdo, whereas K5 is built on a terminal (2→7)-linked residue.

Figure 1.

Figure 1

Open in a new tab

(A) Schematic structures of E. coli O164, O8, and O9 LPS having an R1 core (residues 2 to 10), KLPS (K43 and K30), and KPG (K1 and K5). The chemical structures of lipid A and PVPG are depicted in Figure S1. (B) Initial molecular system structure of 50EcO950K30. Lipid A in LPS and KLPS is colored by green and magenta, and O9 and K30 antigens are colored by cyan and red, respectively. Only the upper leaflet is shown for clarity.

In this work, both KLPS (K30 and K43) and KPG (K1 and K5) are modeled and incorporated into various symmetric bilayers with LPS in different ratios. All-atom MD simulations of these systems are conducted to characterize various bilayer properties, such as hydrophobic thickness, chain order parameters, area per lipid (APL), and the influence of CPS glycoconjugates on the pure LPS-containing bilayer structure, and to get a better estimate for these various properties that could be used in the future modeling of asymmetric OM bilayers. This work is significant in complementing the characterization of interactions between LPS, ECA, CPS, and OM proteins in more realistic OMs.

Methods

The sequences of each LPS, KLPS (K30 and K43), and KPG (K1 and K5) used in this study are summarized in Figure 1. The numbers of repeating units (RUs) are 5 (LPS), 30 (K30), 15 (K43), 50 (K1), and 25 (K5) to ensure that CPS are much longer than LPS. E. coli lipid A was used to model KLPS and LPS with an R1 core. For O8 and O9 antigens, a primer-adaptor of the tetrasaccharide (α-Manp-(1→3)-α-Manp-(1→3)-α-Manp-(1→3)-β-GlcpNAc) attached to the R1 core was added to form a conserved reducing terminal structure.35 For KPG, acyl chains with 16:0 in the C1 tail and 18:1 cis-11 in the C2 tail (palmitoylvacenoyl PG: PVPG) were used. The chemical structures of E. coli lipid A and PVPG are shown in Figure S1.

To investigate the properties of CPS-containing bilayers as well as the conformational properties of each CPS and LPS, symmetric O164, O8, and O9 LPS bilayer systems mixed with 0, 25, and 75% of the associated CPS type were modeled and simulated. Detailed system information, including the number of lipids in each leaflet, the initial size, and the total atom number of each system, is summarized in Table S1. Figure 1 shows an initial structure of 50EcO950K30 (n.b., the superscripts indicate the molar ratio of each LPS and CPS), in which K30 is extended outward to ∼360 Å and is much longer than O9 LPS. The pure LPS systems were generated using Membrane Builder36−40 in CHARMM-GUI (http://www.charmm-gui.org),41 while mixed CPS-containing systems were built following the protocol of Membrane Builder using the pre-modeled structures of KLPS and KPG. Ca2+ ions were added to the LPS lipid A and core sugar residues to neutralize each system, and 150 mM KCl was also added to the bulk region to mimic the bulk ion solution.

The CHARMM36 force field for LPS,42,43 lipids,44 and carbohydrates45−48 was used to describe the system’s energetics. For each system, equilibrations were first conducted by following the Membrane Builder equilibration protocol. Briefly, NVT (constant particle number, volume, and temperature) dynamics were first used and subsequently followed by NPT (constant particle number, pressure, and temperature) simulations. During the equilibration, various restraints were applied to the lipids and water molecules and gradually decreased over six steps to assure the gradual equilibration of the assembled system. After equilibration, 1.5 μs production NPT simulation with 4 fs time-step was conducted for each system. All bonds to hydrogen atoms were fixed using the SHAKE algorithm.49 The van der Waals interactions were smoothly switched off at 10–12 Å by a force-switching function,50 and the long-range electrostatic interactions were calculated using the particle-mesh Ewald method.51 Langevin dynamics was used for the temperature coupling, with a collision frequency of 1 ps–1. A semi-isotropic Monte Carlo barostat method with a pressure coupling frequency of 100 steps was used to maintain the pressure.52,53 The temperature was maintained at 303.15 K, and the pressure was set to 1.0 bar. MD simulations were performed using a hydrogen mass repartitioning scheme, which is available in CHARMM-GUI and was proven to be reliable and comparable with standard 2 fs MD simulations.54 Three independent replicas with different random seed numbers were generated for each system to improve sampling and to check the simulation convergence. All simulations were conducted utilizing OpenMM.55

Results and Discussion

Density Distributions of LPS and CPS

Different from previously studied LPS-only or ECA-containing systems, the Z-dimensions of CPS-containing systems are much larger due to the distinctly long CPS polysaccharides. Table S1 summarizes the initial size of each system, showing that the Z-dimensions of CPS-containing systems are about as twice large as those of pure LPS-containing systems. The variation of X or Y dimension length as a function of simulation time is shown in Figure S2. For each system, although divergences are detected among three replicas due to the slow relaxation of LPS/CPS, each replica gradually becomes equilibrated after 1200 ns simulation. In this work, the last 300 ns trajectories for each system were used for analysis.

Figure 2 shows the density distribution of each component along the membrane normal (i.e., the Z-axis) in a representative 50EcO16450K1 system. Due to the long polysaccharide of K1, the K1 CPS distribution along the Z-axis extends to 250 Å compared to ∼120 Å of O8 polysaccharide, which is consistent with the work of Phanphak et al.56 For 50EcO16450K5, 50EcO850K43, and 50EcO950K30 systems, the density distribution is shown in Figures S3–S5. Figure S6 shows the density distribution of Ca2+ along the Z-axis, and Ca2+ ions are dominantly occupied in the lipid headgroup and core regions, maintaining the integrity of the bilayer system. Some Ca2+ ions are detected in the O-antigen region, indicated by a minor distribution beyond 50 Å.

Figure 2.

Figure 2

Open in a new tab

(A) Density profiles along the Z-axis for different components of each LPS, K1 CPS, and water molecules in system 50EcO16450K1. In the distributions, only the Z > 0 membrane portion up to Z = 280 Å is shown after symmetrization. (B) Equilibrated molecular system structure of 50EcO16450K1. Colors for LPS and KPG are the same as those in (A).

Pairwise root mean square deviations (rmsd) for lipid and sugar portions of each LPS, KLPS, and KPG in each system were calculated to explore the structural variations in various systems. Comparison of pairwise rmsd between 701–1000 and 1201–1500 ns is shown in Figure S7, which indicates the convergence of the sampled ensemble. As shown in Figure 3, for LPS lipid A in each system, the acyl chains of lipid A in KPG-containing systems are more dynamic, with the rmsd peak located around 7 Å. In addition, LPS lipid A in 50% KPG-containing systems is more flexible than that in 25% KPG-containing systems, indicating that more unsaturated tails (more KPG included) make the membrane more fluid. For KLPS lipid A, similar ensembles of structures are sampled, although K43-containing systems are a little more dynamic than K30-containing systems. However, lipid A in 50% KLPS-containing systems is more rigid than that in 25% KLPS-containing systems, as indicated by the slightly left-shifted rmsd peaks. For PVPG of KPG, similar rmsd distributions are observed for KPG-containing systems. The rmsd peaks are located at smaller values compared to lipid A in LPS or KLPS due to its small size. For sugar portions, higher flexibility and dynamics are observed compared to a lipid A or PVPG portion, and broader rmsd distributions are sampled in KLPS- or KPG-containing systems due to their long polysaccharides. In addition, the sugars of KLPS and KPG in 50% CPS-containing systems are more rigid than those in 25% CPS-containing systems, indicating that longer polysaccharides make the membrane more rigid.

Figure 3.

Figure 3

Open in a new tab

Pairwise rmsd distributions for (A) residue 1 (i.e., lipid A or PVPG lipid) and (B) all sugars (i.e., except for residue 1) in each LPS, KLPS, and KPG.

Structural Properties of CPS-Containing Membrane Bilayers

Hydrophobic thickness is an important measure to characterize the acyl chain structures in a membrane bilayer. The calculated average hydrophobic thickness of each system is shown in Figure 4. The carbon atoms of both leaflets (shown in the red circles in Figure S1) were used for the thickness calculation. For pure LPS-containing systems, the averaged hydrophobic thickness is 20.4 ± 0.3 (100EcO164), 21.0 ± 0.4 (100EcO8), and 21.6 ± 0.3 Å (100EcO9), which are comparable with the previously studied 100EcO159 system (20.1 ± 0.2 Å).24,57 When KLPS is incorporated, the membrane bilayer gets thicker, and the averaged thickness becomes 22.0 ± 0.1 (75EcO825K43), 22.5 ± 0.2 (75EcO925K30), 22.7 ± 0.1 (50EcO850K43), and 23.7 ± 0.4 Å (50EcO950K30), indicating that the acyl chains of lipid A in LPS or KLPS are more rigid and ordered in KLPS-containing systems. In contrast, when KPG is included, the membrane bilayer becomes thinner, and the averaged thickness drops to 19.6 ± 0.2 (75EcO16425K1), 20.3 ± 0.1 (75EcO16425K5), 16.6 ± 0.1 (50EcO16450K1), and 17.9 ± 0.1 Å (50Ec16450K5). The trend of decreased hydrophobic thickness along with the increased population of KPG indicates that KPG glycolipids change the structural and dynamic properties of the lipid portion of each LPS and KPG and make LPS or KPG more dynamic and flexible.

Figure 4.

Figure 4

Open in a new tab

Averaged hydrophobic thickness with the standard error from three independent simulations.

The order parameters (SCD) of lipid acyl chains are used to provide information regarding the overall order of the membrane: SCDInline graphic, where θCH is the angle between a C–H bond vector and the Z-axis, and the bracket represents the time and ensemble average. Figure 5 shows the calculated SCD of acyl chain 6 for lipid A and acyl chain 2 for PVPG (see Figure S1 for the chain number). For LPS lipid A, the SCD of each carbon in KLPS-containing systems exhibits higher values compared to other systems, indicating more rigid conformations in KLPS-containing systems. In addition, acyl chains in 50% KLPS-containing systems are more rigid than those in 25% KLPS-containing systems. It is consistent with the hydrophobic thickness result, in which the membrane bilayer gets thicker when more KLPS is incorporated. For lipid A of LPS in KPG-containing systems, the lower SCD values compared to those in pure LPS-containing systems indicate that acyl chains are more dynamic and flexible. In addition, acyl chains in 50% KPG-containing systems are more dynamic than those in 25% KPG-containing systems, which is also observed for acyl chains in PVPG.

Figure 5.

Figure 5

Open in a new tab

Calculated order parameters of acyl chain 6 of lipid A and acyl chain 2 of PVPG (see Figure S1 for the chain number).

The APL provides important information about lipid packing in a bilayer resulting from attractions among head groups and repulsions among non-polar hydrocarbon tails. In this work, the APL of each glycolipid was calculated using the Voronoi tessellation method58 and averaged over the lipids. Carbons in the red circles of lipid A and those in the blue circles of PVPG in Figure S1 were used in the Voronoi tessellation calculation for LPS, KLPS, and KPG. Table 1 summarizes the APL of each LPS, KLPS, and KPG. In pure LPS-containing systems, LPS in 100EcO9 are more rigid and packed tightly with a smaller APL compared to 100EcO164 and 100EcO8. When KLPS is incorporated, the average APL of LPS in 75EcO825K43 and 75EcO925K30 decreases to 177.0 and 176.9 Å2. It is further decreased when more KLPS are mixed into the membrane (171.9 in 50EcO850K43 and 172.1 Å2 in 50EcO950K30), indicating that the acyl chains of lipid A in LPS in KLPS-containing systems become more rigid and ordered compared to those in pure LPS-containing systems. Correspondingly, the average APL of KLPS (K43/K30) also decreases from 173.6/165.3 to 168.3/157.2 Å2 when the population of KLPS increases from 25 to 50%. When 25% KPG is present, the average APL of LPS increases to 201.9/197.8 Å2 in 75EcO16425K1 and 75EcO16425K5 compared to a 100% O164-containing system (190.8 Å2). It is further increased to 238.4 and 225.0 Å2 when 50% KPG is incorporated. Therefore, in contrast to KLPS, the incorporation of KPG enhances disorder, affects structural dynamics, and makes glycolipids in membrane bilayers more flexible and dynamic. In addition, the p-values for APLs were calculated for each glycolipid of each system, and all p-values were less than 0.05, indicating the statistical significance.

Table 1. Calculated Average APL with the Standard Error for Each LPS and CPS.

  APL (Å2)
simulation systems LPS K43 K30 K1 K5
100EcO164 190.8 ± 2.9        
100EcO8 184.9 ± 3.9        
100EcO9 180.0 ± 2.8        
75EcO825K43 177.0 ± 0.5 173.6 ± 1.5      
75EcO925K30 176.9 ± 2.1   165.3 ± 1.3    
50EcO850K43 171.9 ± 2.2 168.3 ± 1.5      
50EcO950K30 172.1 ± 2.6   157.2 ± 4.5    
75EcO16425K1 201.9 ± 1.1     76.6 ± 3.1  
75EcO16425K5 197.8 ± 1.5       73.7 ± 3.0
50EcO16450K1 238.4 ± 3.1     96.0 ± 2.7  
50EcO16450K5 225.0 ± 1.6       87.1 ± 1.0
Open in a new tab

The increasing trend in APL of LPS and KPG along with the increased population of KPG is consistent with our previous study for ECAs,24 in which incorporating ECAPG (ECA linked to PVPG/PSPG/DPPG) into LPS-containing systems also increases the APL of LPS. It is possibly due to stronger electrostatic repulsive interactions between LPS and KPG polysaccharides compared to those between LPS and LPS. In EcO164 LPS, the charges of lipid A, core region, and O164 RUs are −4e, −5e, and 0, respectively. However, in KPG (K1/K5), the charges of PVPG, poly-Kdo, and RUs are −1e, −7e, −50e/–25e, respectively. Stronger repulsive interactions cause more space for the acyl chains of LPS and KPG, resulting in loosely packed membrane bilayers. Similarly, in our previous work, when more ECALPS was incorporated, stronger repulsive interactions between LPS and ECALPS also increased the average APL of both LPS and ECALPS. In this work, incorporating KLPS decreases the APL of LPS and KLPS and makes the acyl chains of lipid A more ordered. Presumably, the repulsive interaction between LPS and LPS is stronger than that between LPS and KLPS since there is no core region in KLPS, resulting in a tighter membrane bilayer.

Conformational Distributions of Glycosidic Linkages in LPS and CPS

Given that the probability distributions of the LPS and CPS as well as membrane properties differ depending on components and mixtures thereof (vide infra), further analysis gives insight into these membrane-anchored systems. In particular, the conformational preference and flexibility at the glycosidic linkages of the polysaccharides pose questions about the effects of the environment on preferred conformations and dynamics. For the O-antigen polysaccharides of E. coli LPS of serogroups O8, O9, and O164 and the KLPS with CPS serotypes K30 and K43, the hexose sugar residues have the d absolute configuration, which is also the case for the KPG CPS K5 structure. Thus, the conformational preference at the ϕ torsion angle is anticipated to be governed by the exo-anomeric effect with α-d-hexoses having ϕ ≈ −50° and β-d-hexoses having ϕ ≈ +50°,59,60 whereas the ψ torsion angles, originating from substitution at secondary hydroxyl groups, are populated in the vicinity of 0°, but with a quite large span possible; antiperiplanar conformations may also be populated to some extent.61 For β-(1→6)-linkages in O164 LPS and α-(2→8)-linkages in K1 KPG, additional aspects need to be considered due to the positions of substitution.

The O-antigen polysaccharide of E. coli O9 contains only α-linked sugars in which the 2- and 3-substituted residues are arranged in blocks within the repeating unit (RU) of the polymer (Figures 6 and S8). The α-(1→2)-linkages populate two conformational states for the ψ torsion angle in a syn arrangement, bisected by the eclipsed conformation at 0°, as a result of a bistable potential similar to its “mirror image” α-l-rhap-(1→2)-α-l-rhap-OMe,62 but essentially devoid of the non-exo-anomeric conformation of the ϕ torsion angle.63 This is contrasted by a quite skewed population preference for the α-(1→3)-linkages in which the major conformational state has the ψ torsion angle in a synperiplanar orientation with ψ ≈ −20° and the non-exo-anomeric conformation of the ϕ torsion angle populated, but only to a minor extent. The conformational behavior is fully consistent with an α-(1→3)-linked oligosaccharide (denoted cnX) representing a glucuronoxylomannan CPS of a pathogenic encapsulated Cryptococcus neoformans fungus,64 in particular as the CHARMM36 additive force field for carbohydrates was used for the MD simulations. However, the influence of the K30 CPS on the conformational distributions of the glycosidic linkages in the EcO9 LPS is small, with merely minor shifts in populations of accessible conformational states (Figure S8). The O-polysaccharide of E. coli O8 is also a homopolymer of d-mannosyl residues, though one of the three sugars in the RU contains a β-(1→2)-linkage that, like mannosyl α-(1→2)-linkages, exhibits a two-state conformational equilibrium divided at ψ ≈ 0° but also a third distinct conformational state at ϕ ≈ 60° and ψ ≈ 90° populated to a small extent (Figure S9). Like for the E. coli O9 O-antigen, the presence of E. coli K43 CPS influences the conformational distributions just marginally.

Figure 6.

Figure 6

Open in a new tab

Two-dimensional distributions of glycosidic torsion angles (ϕ, ψ) for each linkage in polysaccharides of selected LPS and CPS [(left) LPS of 50EcO950K30, (middle) LPS of 50EcO16450K5, and (right) CPS of 50EcO850K43] together with schematic structures in SNFG format (see Figures S8–S14 for all other combinations).65 The density was rescaled by the maximum value: white for 0, blue for 0.1, green for 0.3, yellow for 0.7, and red for 1.

A branched pentasaccharide structure is present in the RU of the E. coli O164 O-antigen with many different conformational sampling patterns in glycosidic linkages (Figure S10). Characteristics of the RU structure are a vicinally disubstituted galactopyranosyl residue, a β-(1→3)-linked galactofuranosyl residue, which in turn is substituted at O6, and in the side-chain a β-(1→6)-linked glucopyranosyl residue substituting a glucopyranosyl residue. The 3,4-disubstituted galactopyranosyl residue is substituted by β- and α-linked residues, respectively. Whereas the conformational space of the former is restricted mainly to a single syn conformational region, although both anti-ϕ and anti-ψ conformational states are populated transiently, the α-(1→4)-linkage displays a major but also an extended conformational syn region that is populated. For the β-(1→3)-linked galactofuranosyl residue, a non-exo conformation with ϕ < 0° is sampled besides the exo-anomeric conformation being the major one. Two β-(1→6)-linkages are present in the RU. In the side-chain, a glucopyranosyl residue is substituted and the major conformational state has the ψ torsion angle (C1′–O6–C6–C5) in an antiperiplanar orientation,66,67 and limited excursions to clinal states. In contrast, the β-(1→6)-linkage in the main-chain involving the galactofuranosyl residue shows not only the antiperiplanar orientation as the major one but also both of the clinal orientations with ψ ≈ +90° and ψ ≈ −90° in distinct conformational states. Neither the CPS of K1 nor that of K5 influence the conformational dynamics of the E. coli O164 O-antigen to any great extent (Figure S10).

The molecular systems with CPS as either KLPS (K30 and K43) or KPG (K1 and K5) have polysaccharides extending beyond those of the LPS. Therefore, for the subsequent analysis, they have been sub-divided into “inner” and “outer” regions, where the former corresponds to the region spanned by O-antigens of LPS (cf. Figure 1B). The CPS from K30 contains four sugars in the RU with a vicinally 2,3-disubstituted branch point sugar residue in which the α-(1→3)-linked galactosyl residue of the side-chain substituting the mannosyl residue has a quite restricted conformational space (Figure S11). Notably, the β-(1→2)-linked main-chain residue substituting the other glycosyloxylated carbon of the branch-point residue exhibits flexibility differences for residues in the “inner” versus the “outer” region of the CPS with those at the “inner” region populating a somewhat larger conformational space. In a similar manner, the populations of anti-ψ conformers are higher in the “inner” region for the β-(1→4)-linked main-chain residues of K43 (Figures 6 and S12). Likewise, for the K5 CPS of the KPG type consisting of a disaccharide RU with (1→4)-linkages with alternating α- and β-anomeric configured sugar residues having the gluco-configuration, anti-ψ conformers are more predominant in the “inner” section, in particular for the sugar having the α-anomeric configuration (Figure S13). In the K1 CPS, being a homopolymer of α-(2→8)-linked sialic acids, only minor differences are observed in the population preferences of the two regions (Figure S14).

The number of conformational states populated at the glycosidic linkage of the α-(2→8)-linked Neu5Ac CPS is quite large, and six conformational regions are possible to identify (Figure 7A), viz., ϕ ≈ −90° and ψ ≈ +90°, ϕ ≈ −60° and ψ ≈ +120°, ϕ ≈ −60° and ψ ≈ +180°, ϕ ≈ +60° and ψ ≈ +100°, ϕ ≈ +80° and ψ ≈ +170°, as well as ϕ ≈ +180° and ψ ≈ −100°, where the torsion angles are defined by ϕ = C1′-C2′-O8-C8 and ψ = C2′-O8-C8-C7. A comparison to an MD simulation using the Glycam-06h force field of an oligosaccharide containing α-(2→8)-linked N-acetylneuraminic acids68 reveals a highly flexible glycosidic linkage with populations at, in particular, ϕ ≈ −60° and ψ ≈ +120°, ϕ ≈ +180° and ψ ≈ +120°, and still significantly populated at ϕ ≈ −90° and ψ ≈ −90°. For the α-(2→8)-linked dimer of Neu5Ac, all three staggered conformations were populated in MD simulations thereof using the force field 53A6GLYC, with ψ ≈ +120° and additional excursions to −anticlinal orientations.69

Figure 7.

Figure 7

Open in a new tab

(A) Two-dimensional distributions of glycosidic torsion angles (ϕ, ψ) for the linkage in the inner portion of the K1 CPS in system 50EcO16450K1 with the schematic structure of the CPS RU in SNFG format. The inner CPS correspond to the region spanned by LPS O-antigens and the outer ones are above the O-antigens. The density was rescaled by the maximum value: white for 0, blue for 0.1, green for 0.3, yellow for 0.7, and red for 1. (B) Rotamer distribution of the ω8 torsion angle (O8–C8–C7–O7) in the sialic acid residue of the CPS from K1 in combination with the polysaccharide of the LPS of EcO164 (50% K1, blue) and (25% K1, red) for the CPS inner portion (top) and the CPS outer portion (bottom). (C) The corresponding plots for the ω7 torsion angle (O7–C7–C6–O6).

In addition to the glycosidic torsion angles in α-(2→8)-linked sialic acids,70 the exocyclic portion of the sialic acid (glycerol structural element) was analyzed with respect to the rotamer distribution of the torsion angles, ω8 = O8–C8–C7–O7 and ω7 = O7–C7–C6–O6, which revealed different sampling of conformational space. The former, involving the glycosidically substituted O8 atom, shows a preference for the antiperiplanar orientation, but other conformations are sampled in a more continuous fashion, although some of the eclipsed ones are not populated as anticipated (Figure 7B). The latter torsion angle containing O6 of the pyranose ring shows almost exclusively the +synclinal rotamer (cf. gauche effect) and a small antiperiplanar population with ω7 ≈ −160° (Figure 7C). The conformational behavior for the exocyclic torsions is consistent with that of the parent α-(2→8)-linked disaccharide in which ω8 populates all three staggered conformations whereas the ω7 only samples two of them.71 Notably, the recently performed MD simulations employing the 53A6GLYC force field showed that all three rotamers were significantly sampled for the ω8 torsion angle, whereas for ω7 only the +synclinal rotamer contributed.69 Interestingly, by extending the degree of polymerization to a decamer of α-(2→8)-linked sialic acid, the results of the MD simulation concluded that the polymer is flexible, which in particular gives credence to the current MD simulation with 50 sialic acids in the polysaccharide K1 where not only the glycosidic linkage populates several torsional states but also the ω8 contains close to a continuous distribution of conformationally accessible states (Figure 7).

Conclusions

Bacterial capsules are made of long-chain polysaccharides anchored to the OM of Gram-negative bacteria and represent major virulence factors that protect pathogens from host defenses. There are more than 80 different types of CPS in E. coli categorized into four groups. Groups 1 and 4 capsules are attached to lipid A (KLPS), and groups 2 and 3 capsules are attached to a PG (KPG) via alternating β-linked 2,4- and 2,7-Kdo residues. Modeling and simulation of CPS in Gram-negative bacteria at the atomic and molecular level are critical to understand its biological functions. However, the outer leaflet of the OM is represented exclusively by LPS in most simulation studies, and the OMs including CPS are rarely modeled and simulated due to the complexity and diversity of CPS.

In this work, representative E. coli CPS, KLPS (K43 and K30) and KPG (K1 and K5), were modeled and incorporated into various symmetric bilayers with co-existing LPS in different ratios. All-atom MD simulations of these systems were conducted to characterize various bilayer properties. Pairwise RMSD shows that dynamic and flexible conformations are sampled for each LPS and CPS, with broad rmsd distributions in each system. However, incorporating KLPS or KPG has different effects on the structural properties of each glycolipid in the membrane systems. On the one hand, the trend of decreased hydrophobic thickness along with the increased population of KPG and, on the other hand, the increased hydrophobic thickness as a result of the increased population of KLPS indicate that glycolipids change the structural and dynamic properties of the lipid portion of each LPS and CPS. The acyl chains of lipid A in LPS or KLPS become rigid and ordered when more KLPS glycolipids are incorporated, while when KPG are included, the acyl chains become dynamic and flexible, which is also observed in the order parameters analysis. The calculated APL of each glycolipid indicates that the averaged APL of LPS in KLPS-containing systems decreased compared to pure LPS-containing systems, while trends of increased APL for LPS correlate with the increase of the population of KPG. Conformational distributions of glycosidic linkages in LPS and CPS were also analyzed and showed that the influence on the conformational distributions of the glycosidic linkages of LPS from the presence of the CPS is small, with minor differences detected for the inner and outer regions of the CPS (where the inner CPS correspond to the region spanned by LPS O-antigens and the outer ones are above the O-antigens). The work provides a good estimate and starting point for building more realistic OMs with different ratios of LPS/CPS/ECA and characterizing the interactions between the OM and OM proteins.

Data and Software Availability

Input files for generating symmetric lipid bilayers containing CPS molecules and the structure file of separate K1, K5, K43, and K30 molecules can be obtained from https://github.com/graceYaGao/CPS.

Acknowledgments

This work was supported in part by grants from the NSF MCB-211172 and MCB-1810695 (to WI), and the Swedish Research Council 2017-03703 (to GW).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jcim.3c00072.

  • System information for MD simulation; chemical structures of E. coli lipid A and PVPG; time-series of X or Y system length; distributions of ions and charged moieties of LPS along the Z-axis in system 50EcO16450K1; two-dimensional distributions of glycosidic torsion angles of conformations for each linkage in the polysaccharide of the O9/O8/O164 LPS from O9/O8/O164-containing systems; two-dimensional distributions of glycosidic torsion angles of conformations for each linkage in the polysaccharide of the K30/K43/K5/K1 CPS from different K30/K43/K5/K1-containing systems (PDF)

The authors declare no competing financial interest.

Supplementary Material

ci3c00072_si_001.pdf (16.9MB, pdf)

References

  1. Kamio Y.; Nikaido H. Outer Membrane of Salmonella Typhimurium: Accessibility of Phospholipid Head Groups to Phospholipase C and Cyanogen Bromide Activated Dextran in the External Medium. Biochemistry 1976, 15, 2561–2570. 10.1021/bi00657a012. [DOI] [PubMed] [Google Scholar]
  2. Smit J.; Kamio Y.; Nikaido H. Outer Membrane of Salmonella Typhimurium: Chemical Analysis and Freeze-Fracture Studies with Lipopolysaccharide Mutants. J. Bacteriol. 1975, 124, 942–958. 10.1128/jb.124.2.942-958.1975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Whitfield C. Biosynthesis and Assembly of Capsular Polysaccharides in Escherichia coli. Annu. Rev. Biochem. 2006, 75, 39–68. 10.1146/annurev.biochem.75.103004.142545. [DOI] [PubMed] [Google Scholar]
  4. Willis L. M.; Whitfield C. Structure, Biosynthesis, and Function of Bacterial Capsular Polysaccharides Synthesized by ABC Transporter-Dependent Pathways. Carbohydr. Res. 2013, 378, 35–44. 10.1016/j.carres.2013.05.007. [DOI] [PubMed] [Google Scholar]
  5. Goller C. C.; Seed P. C. High-Throughput Identification of Chemical Inhibitors of E. coli Group 2 Capsule Biogenesis as Anti-Virulence Agents. PLoS One 2010, 5, e11642. 10.1371/journal.pone.0011642. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Corbett D.; Hudson T.; Roberts I.. Prokaryotic Cell Wall Compounds, 2010, pp 111–132. [Google Scholar]
  7. Cuthbertson L.; Mainprize I. L.; Naismith J. H.; Whitfield C. Pivotal Roles of the Outer Membrane Polysaccharide Export and Polysaccharide Copolymerase Protein Families in Export of Extracellular Polysaccharides in Gram-Negative Bacteria. Microbiol. Mol. Biol. Rev. 2009, 73, 155–177. 10.1128/mmbr.00024-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cuthbertson L.; Kos V.; Whitfield C. Abc Transporters Involved in Export of Cell Surface Glycoconjugates. Microbiol. Mol. Biol. Rev. 2010, 74, 341–362. 10.1128/mmbr.00009-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Whitfield C. Structure and Assembly of Escherichia coli Capsules. EcoSal Plus 2009, 3, PMID 26443756. 10.1128/ecosalplus.4.7.3. [DOI] [PubMed] [Google Scholar]
  10. Agrawal A.; Murphy T. F. Haemophilus Influenzae Infections in the H. Influenzae Type B Conjugate Vaccine Era. J. Clin. Microbiol. 2011, 49, 3728–3732. 10.1128/jcm.05476-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Wilkie I. W.; Harper M.; Boyce J. D.; Adler B. Pasteurella multocida: Diseases and Pathogenesis. Curr. Top. Microbiol. Immunol. 2012, 361, 1–22. 10.1007/82_2012_216. [DOI] [PubMed] [Google Scholar]
  12. Wu E. L.; Engström O.; Jo S.; Stuhlsatz D.; Yeom M. S.; Klauda J. B.; Widmalm G.; Im W. Molecular Dynamics and NMR Spectroscopy Studies of E. coli Lipopolysaccharide Structure and Dynamics. Biophys. J. 2013, 105, 1444–1455. 10.1016/j.bpj.2013.08.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ma H. L.; Irudayanathan F. J.; Jiang W. J.; Nangia S. Simulating Gram-Negative Bacterial Outer Membrane: A Coarse Grain Model. J. Phys. Chem. B 2015, 119, 14668–14682. 10.1021/acs.jpcb.5b07122. [DOI] [PubMed] [Google Scholar]
  14. Blasco P.; Patel D. S.; Engström O.; Im W.; Widmalm G. Conformational Dynamics of the Lipopolysaccharide from Escherichia coli O91 Revealed by Nuclear Magnetic Resonance Spectroscopy and Molecular Simulations. Biochemistry 2017, 56, 3826–3839. 10.1021/acs.biochem.7b00106. [DOI] [PubMed] [Google Scholar]
  15. Li A.; Schertzer J. W.; Yong X. Molecular Dynamics Modeling of Pseudomonas Aeruginosa Outer Membranes. Phys. Chem. Chem. Phys. 2018, 20, 23635–23648. 10.1039/c8cp04278k. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Piggot T. J.; Holdbrook D. A.; Khalid S. Electroporation of the E. coli and S. aureus Membranes: Molecular Dynamics Simulations of Complex Bacterial Membranes. J. Phys. Chem. B 2011, 115, 13381–13388. 10.1021/jp207013v. [DOI] [PubMed] [Google Scholar]
  17. Lee J.; Pothula K. R.; Kleinekathöfer U.; Im W. Simulation Study of Occk5 Functional Properties in Pseudomonas aeruginosa Outer Membranes. J. Phys. Chem. B 2018, 122, 8185–8192. 10.1021/acs.jpcb.8b07109. [DOI] [PubMed] [Google Scholar]
  18. Wu E. L.; Fleming P. J.; Yeom M. S.; Widmalm G.; Klauda J. B.; Fleming K. G.; Im W. E. coli Outer Membrane and Interactions with OmpLA. Biophys. J. 2014, 106, 2493–2502. 10.1016/j.bpj.2014.04.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Shearer J.; Jefferies D.; Khalid S. Outer Membrane Proteins Ompa, Fhua, Ompf, Esta, Btub, and Ompx Have Unique Lipo-polysaccharide Fingerprints. J. Chem. Theory Comput. 2019, 15, 2608–2619. 10.1021/acs.jctc.8b01059. [DOI] [PubMed] [Google Scholar]
  20. Im W.; Khalid S. Molecular Simulations of Gram-Negative Bacterial Membranes Come of Age. Annu. Rev. Phys. Chem. 2020, 71, 171. 10.1146/annurev-physchem-103019-033434. [DOI] [PubMed] [Google Scholar]
  21. Patel D. S.; Qi Y. F.; Im W. Modeling and Simulation of Bacterial Outer Membranes and Interactions with Membrane Proteins. Curr. Opin. Struct. Biol. 2017, 43, 131–140. 10.1016/j.sbi.2017.01.003. [DOI] [PubMed] [Google Scholar]
  22. Hwang H.; Paracini N.; Parks J. M.; Lakey J. H.; Gumbart J. C. Distribution of Mechanical Stress in the Escherichia coli Cell Envelope. Biochim. Biophys. Acta, Biomembr. 2018, 1860, 2566–2575. 10.1016/j.bbamem.2018.09.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Patel D. S.; Park S.; Wu E. L.; Yeom M. S.; Widmalm G.; Klauda J. B.; Im W. Influence of Ganglioside Gm1 Concentration on Lipid Clustering and Membrane Properties and Curvature. Biophys. J. 2016, 111, 1987–1999. 10.1016/j.bpj.2016.09.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Gao Y.; Lee J.; Widmalm G.; Im W. Modeling and Simulation of Bacterial Outer Membranes with Lipopolysaccharides and Enterobacterial Common Antigen. J. Phys. Chem. B 2020, 124, 5948–5956. 10.1021/acs.jpcb.0c03353. [DOI] [PubMed] [Google Scholar]
  25. Prehm P.; Jann B.; Jann K. The O9 Antigen of Escherichia coli. Structure of the Polysaccharide Chain. Eur. J. Biochem. 1976, 67, 53–56. 10.1111/j.1432-1033.1976.tb10631.x. [DOI] [PubMed] [Google Scholar]
  26. Chakraborty A. K.; Friebolin H.; Stirm S. Primary Structure of the Escherichia coli Serotype K30 Capsular Polysaccharide. J. Bacteriol. 1980, 141, 971–972. 10.1128/jb.141.2.971-972.1980. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jansson P. E.; Lönngren J.; Widmalm G.; Leontein K.; Slettengren K.; Svenson S. B.; Wrangsell G.; Dell A.; Tiller P. R. Structural Studies of the O-Antigen Polysaccharides of Klebsiella O5 and Escherichia coli O8. Carbohydr. Res. 1985, 145, 59–66. 10.1016/s0008-6215(00)90412-9. [DOI] [PubMed] [Google Scholar]
  28. Choy Y. M.; Dutton G. G.; Ruth Leslie M. R.; Parolis H.; Parolis L. A. The Structure of the Capsular Polysaccharide of Escherichia coli O8:K43:H11. Carbohydr. Res. 1995, 269, 295–302. 10.1016/0008-6215(94)00366-n. [DOI] [PubMed] [Google Scholar]
  29. McGuire E. J.; Binkley S. B. The Structure and Chemistry of Colominic Acid. Biochemistry 1964, 3, 247–251. 10.1021/bi00890a017. [DOI] [PubMed] [Google Scholar]
  30. Vann W. F.; Schmidt M. A.; Jann B.; Jann K. The Structure of the Capsular Polysaccharide (K5 Antigenn) of Urinary-Tract-Infective Escherichia coli 010:K5:H4. Eur. J. Biochem. 1981, 116, 359–364. 10.1111/j.1432-1033.1981.tb05343.x. [DOI] [PubMed] [Google Scholar]
  31. Linnerborg M.; Weintraub A.; Widmalm G. Structural Studies of the O-Antigen Polysaccharide from the Enteroinvasive Escherichia coli O164 Cross-Reacting with Shigella dysenteriae Type 3. Eur. J. Biochem. 1999, 266, 460–466. 10.1046/j.1432-1327.1999.00878.x. [DOI] [PubMed] [Google Scholar]
  32. Silver R. P.; Finn C. W.; Vann W. F.; Aaronson W.; Schneerson R.; Kretschmer P. J.; Garon C. F. Molecular cloning of the K1 capsular polysaccharide genes of E. coli. Nature 1981, 289, 696–698. 10.1038/289696b0. [DOI] [PubMed] [Google Scholar]
  33. Willis L. M.; Stupak J.; Richards M. R.; Lowary T. L.; Li J.; Whitfield C. Conserved Glycolipid Termini in Capsular Polysaccharides Synthesized by ATP-Binding Cassette Transporter-Dependent Pathways in Gram-Negative Pathogens. Proc. Natl. Acad. Sci. 2013, 110, 7868–7873. 10.1073/pnas.1222317110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Doyle L.; Ovchinnikova O. G.; Myler K.; Mallette E.; Huang B.-S.; Lowary T. L.; Kimber M. S.; Whitfield C. Biosynthesis of a Conserved Glycolipid Anchor for Gram-Negative Bacterial Capsules. Nat. Chem. Biol. 2019, 15, 632–640. 10.1038/s41589-019-0276-8. [DOI] [PubMed] [Google Scholar]
  35. Greenfield L. K.; Richards M. R.; Li J.; Wakarchuk W. W.; Lowary T. L.; Whitfield C. Biosynthesis of the Polymannose Lipopolysaccharide O-Antigens from Escherichia coli Serotypes O8 and O9a Requires a Unique Combination of Single- and Multiple-Active Site Mannosyltransferases. J. Biol. Chem. 2012, 287, 35078–35091. 10.1074/jbc.m112.401000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Jo S.; Kim T.; Im W. Automated Builder and Database of Protein/Membrane Complexes for Molecular Dynamics Simulations. PLoS One 2007, 2, e880. 10.1371/journal.pone.0000880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Jo S.; Lim J. B.; Klauda J. B.; Im W. CHARMM-GUI Membrane Builder for Mixed Bilayers and Its Application to Yeast Membranes. Biophys. J. 2009, 97, 50–58. 10.1016/j.bpj.2009.04.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Wu E. L.; Cheng X.; Jo S.; Rui H.; Song K. C.; Dávila-Contreras E. M.; Qi Y. F.; Lee J. M.; Monje-Galvan V.; Venable R. M.; Klauda J. B.; Im W. CHARMM-GUI Membrane Builder toward realistic biological membrane simulations. J. Comput. Chem. 2014, 35, 1997–2004. 10.1002/jcc.23702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Lee J.; Cheng X.; Swails J. M.; Yeom M. S.; Eastman P. K.; Lemkul J. A.; Wei S.; Buckner J.; Jeong J. C.; Qi Y. F.; Jo S.; Pande V. S.; Case D. A.; Brooks C. L.; MacKerell A. D.; Klauda J. B.; Im W. CHARMM-GUI Input Generator for NAMD, GROMACS, AMBER, OpenMM, and CHARMM/OpenMM Simulations Using the CHARMM36 Additive Force Field. J. Chem. Theory Comput. 2016, 12, 405–413. 10.1021/acs.jctc.5b00935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Lee J.; Patel D. S.; Ståhle J.; Park S. J.; Kern N. R.; Kim S.; Lee J.; Cheng X.; Valvano M. A.; Holst O.; Knirel Y. A.; Qi Y. F.; Jo S.; Klauda J. B.; Widmalm G.; Im W. CHARMM-GUI Membrane Builder for Complex Biological Membrane Simulations with Glycolipids and Lipoglycans. J. Chem. Theory Comput. 2019, 15, 775–786. 10.1021/acs.jctc.8b01066. [DOI] [PubMed] [Google Scholar]
  41. Jo S.; Kim T.; Iyer V. G.; Im W. CHARMM-GUI: A web-based graphical user interface for CHARMM. J. Comput. Chem. 2008, 29, 1859–1865. 10.1002/jcc.20945. [DOI] [PubMed] [Google Scholar]
  42. Kim S.; Patel D. S.; Park S.; Slusky J.; Klauda J. B.; Widmalm G.; Im W. Bilayer Properties of Lipid a from Various Gram-Negative Bacteria. Biophys. J. 2016, 111, 1750–1760. 10.1016/j.bpj.2016.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Jo S.; Wu E. L.; Stuhlsatz D.; Klauda J. B.; MacKerell A. D.; Widmalm G.; Im W.. Lipopolysaccharide Membrane Building and Simulation. In Glycoinformatics, Lütteke T., Frank M., Eds.; Springer New York: New York, NY, 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Klauda J. B.; Venable R. M.; Freites J. A.; O’Connor J. W.; Tobias D. J.; Mondragon-Ramirez C.; Vorobyov I.; MacKerell A. D.; Pastor R. W. Update of the CHARMM All-Atom Additive Force Field for Lipids: Validation on Six Lipid Types. J. Phys. Chem. B 2010, 114, 7830–7843. 10.1021/jp101759q. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Guvench O.; Greene S. N.; Kamath G.; Brady J. W.; Venable R. M.; Pastor R. W.; Mackerell A. D. Additive Empirical Force Field for Hexopyranose Monosaccharides. J. Comput. Chem. 2008, 29, 2543–2564. 10.1002/jcc.21004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Guvench O.; Hatcher E.; Venable R. M.; Pastor R. W.; MacKerell A. D. CHARMM Additive All-Atom Force Field for Glycosidic Linkages between Hexopyranoses. J. Chem. Theory Comput. 2009, 5, 2353–2370. 10.1021/ct900242e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Hatcher E.; Guvench O.; MacKerell A. D. CHARMM Additive All-Atom Force Field for Aldopentofuranoses, Methyl-Aldopentofuranosides, and Fructofuranose. J. Phys. Chem. B 2009, 113, 12466–12476. 10.1021/jp905496e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Guvench O.; Mallajosyula S. S.; Raman E. P.; Hatcher E.; Vanommeslaeghe K.; Foster T. J.; Jamison F. W.; MacKerell A. D. CHARMM Additive All-Atom Force Field for Carbohydrate Derivatives and Its Utility in Polysaccharide and Carbohydrate-Protein Modeling. J. Chem. Theory Comput. 2011, 7, 3162–3180. 10.1021/ct200328p. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Ryckaert J. P.; Ciccotti G.; Berendsen H. J. C. Numerical-Integration of Cartesian Equations of Motion of a System with Constraints - Molecular-Dynamics of N-Alkanes. J. Comput. Phys. 1977, 23, 327–341. 10.1016/0021-9991(77)90098-5. [DOI] [Google Scholar]
  50. Steinbach P. J.; Brooks B. R. New Spherical-Cutoff Methods for Long-Range Forces in Macromolecular Simulation. J. Comput. Chem. 1994, 15, 667–683. 10.1002/jcc.540150702. [DOI] [Google Scholar]
  51. Essmann U.; Perera L.; Berkowitz M. L.; Darden T.; Lee H.; Pedersen L. G. A Smooth Particle Mesh Ewald Method. J. Chem. Phys. 1995, 103, 8577–8593. 10.1063/1.470117. [DOI] [Google Scholar]
  52. Chow K. H.; Ferguson D. M. Isothermal Isobaric Molecular-Dynamics Simulations with Monte-Carlo Volume Sampling. Comput. Phys. Commun. 1995, 91, 283–289. 10.1016/0010-4655(95)00059-o. [DOI] [Google Scholar]
  53. Åqvist J.; Wennerström P.; Nervall M.; Bjelic S.; Brandsdal B. O. Molecular Dynamics Simulations of Water and Biomolecules with a Monte Carlo Constant Pressure Algorithm. Chem. Phys. Lett. 2004, 384, 288–294. 10.1016/j.cplett.2003.12.039. [DOI] [Google Scholar]
  54. Gao Y.; Lee J.; Smith I. P. S.; Lee H.; Kim S.; Qi Y.; Klauda J. B.; Widmalm G.; Khalid S.; Im W. CHARMM-GUI Supports Hydrogen Mass Repartitioning and Different Protonation States of Phosphates in Lipopolysaccharides. J. Chem. Inf. Model. 2021, 61, 831–839. 10.1021/acs.jcim.0c01360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Eastman P.; Friedrichs M. S.; Chodera J. D.; Radmer R. J.; Bruns C. M.; Ku J. P.; Beauchamp K. A.; Lane T. J.; Wang L. P.; Shukla D.; Tye T.; Houston M.; Stich T.; Klein C.; Shirts M. R.; Pande V. S. OpenMM 4: A Reusable, Extensible, Hardware Independent Library for High Performance Molecular Simulation. J. Chem. Theory Comput. 2013, 9, 461–469. 10.1021/ct300857j. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Phanphak S.; Georgiades P.; Li R.; King J.; Roberts I. S.; Waigh T. A. Super-Resolution Fluorescence Microscopy Study of the Production of K1 Capsules by Escherichia coli: Evidence for the Differential Distribution of the Capsule at the Poles and the Equator of the Cell. Langmuir 2019, 35, 5635–5646. 10.1021/acs.langmuir.8b04122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Gao Y.; Lee J.; Widmalm G.; Im W. Correction to “Modeling and Simulation of Bacterial Outer Membranes with Lipopolysaccharides and Enterobacterial Common Antigen”. J. Phys. Chem. B 2021, 125, 9384–9385. 10.1021/acs.jpcb.1c06516. [DOI] [PubMed] [Google Scholar]
  58. Shinoda W.; Okazaki S. A Voronoi Analysis of Lipid Area Fluctuation in a Bilayer. J. Chem. Phys. 1998, 109, 1517–1521. 10.1063/1.476702. [DOI] [Google Scholar]
  59. Lemieux R. U.; Koto S. The Conformational Properties of Glycosidic Linkages. Tetrahedron 1974, 30, 1933–1944. 10.1016/s0040-4020(01)97324-7. [DOI] [Google Scholar]
  60. Söderman P.; Jansson P.-E.; Widmalm G. Synthesis, NMR Spectroscopy and Conformational Studies of the Four Anomeric Methyl Glycosides of the Trisaccharide D-Glcp-(1→3)-[D-Glcp-(1→4)]-α-D-Glcp. J. Chem. Soc., Perkin Trans. 1998, 2, 639–648. 10.1039/a707346a. [DOI] [Google Scholar]
  61. Yang M.; Angles d’Ortoli T.; Säwén E.; Jana M.; Widmalm G.; MacKerell A. D. Delineating the Conformational Flexibility of Trisaccharides from NMR Spectroscopy Experiments and Computer Simulations. Phys. Chem. Chem. Phys. 2016, 18, 18776–18794. 10.1039/c6cp02970a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Zerbetto M.; Polimeno A.; Widmalm G. Glycosidic Linkage Flexibility: The Ψ Torsion Angle Has a Bimodal Distribution in α-L-Rhap-(1→2)-α-L-Rhap-OMe as Deduced from 13C NMR Spin Relaxation. J. Chem. Phys. 2020, 152, 035103. 10.1063/1.5134531. [DOI] [PubMed] [Google Scholar]
  63. Säwén E.; Massad T.; Landersjö C.; Damberg P.; Widmalm G. Population Distribution of Flexible Molecules from Maximum Entropy Analysis Using Different Priors as Background Information: Application to the Φ, Ψ-Conformational Space of the α-(1→2)-Linked Mannose Disaccharide Present in N- and O-Linked Glycoproteins. Org. Biomol. Chem. 2010, 8, 3684–3695. 10.1039/c003958f. [DOI] [PubMed] [Google Scholar]
  64. Kuttel M. M.; Casadevall A.; Oscarson S. Cryptococcus neoformans Capsular GXM ConformationEpitope Presentation: A Molecular Modelling Study. Molecules 2020, 25, 2651. 10.3390/molecules25112651. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Neelamegham S.; Aoki-Kinoshita K.; Bolton E.; Frank M.; Lisacek F.; Lütteke T.; O’Boyle N.; Packer N. H.; Stanley P.; Toukach P.; Varki A.; Woods R. J.; Darvill A.; Dell A.; Henrissat B.; Bertozzi C.; Hart G.; Narimatsu H.; Freeze H.; Yamada I.; Paulson J.; Prestegard J.; Marth J.; Vliegenthart J. F. G.; Etzler M.; Aebi M.; Kanehisa M.; Taniguchi N.; Edwards N.; Rudd P.; Seeberger P.; Mazumder R.; Ranzinger R.; Cummings R.; Schnaar R.; Perez S.; Kornfeld S.; Kinoshita T.; York W.; Knirel Y. Updates to the Symbol Nomenclature for Glycans Guidelines. Glycobiology 2019, 29, 620–624. 10.1093/glycob/cwz045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Pendrill R.; Säwén E.; Widmalm G. Conformation and Dynamics at a Flexible Glycosidic Linkage Revealed by NMR Spectroscopy and Molecular Dynamics Simulations: Analysis of β-L-Fucp-(1→6)-α-D-Glcp-OMe in Water Solution. J. Phys. Chem. B 2013, 117, 14709–14722. 10.1021/jp409985h. [DOI] [PubMed] [Google Scholar]
  67. Patel D. S.; Pendrill R.; Mallajosyula S. S.; Widmalm G.; MacKerell A. D. Conformational Properties of α- or β-(1→6)-Linked Oligosaccharides: Hamiltonian Replica Exchange MD Simulations and NMR Experiments. J. Phys. Chem. B 2014, 118, 2851–2871. 10.1021/jp412051v. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Novakovic M.; Battistel M. D.; Azurmendi H. F.; Concilio M.-G.; Freedberg D. I.; Frydman L. The Incorporation of Labile Protons into Multidimensional NMR Analyses: Glycan Structures Revisited. J. Am. Chem. Soc. 2021, 143, 8935–8948. 10.1021/jacs.1c04512. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Turupcu A.; Blaukopf M.; Kosma P.; Oostenbrink C. Molecular Conformations of Di-, Tri-, and Tetra-α-(2→8)-Linked Sialic Acid from NMR Spectroscopy and MD Simulations. Int. J. Mol. Sci. 2019, 21, 30. 10.3390/ijms21010030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Soares C. O.; Grosso A. S.; Ereño-Orbea J.; Coelho H.; Marcelo F. Molecular Recognition Insights of Sialic Acid Glycans by Distinct Receptors Unveiled by NMR and Molecular Modeling. Front. Mol. Biosci. 2021, 8, 727847. 10.3389/fmolb.2021.727847. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Vasudevan S. V.; Balaji P. V. Conformational Analysis and Implications for Binding to Proteins Molecular Dynamics Simulations of α2→8-Linked Disialoside. Biopolymers 2002, 63, 168–180. 10.1002/bip.10019. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ci3c00072_si_001.pdf (16.9MB, pdf)

Articles from Journal of Chemical Information and Modeling are provided here courtesy of American Chemical Society

RESOURCES