Modeling Diversity in Structures of Bacterial Outer Membrane Lipids

Jan 12, 2017 - The primary immunostimulatory center of the LPS molecule is lipid A, ... Considering the broad diversity in bacterial species, there ar...
0 downloads 0 Views 5MB Size
Subscriber access provided by University of Newcastle, Australia

Article

Modeling diversity in structures of bacterial outer membrane lipids Huilin Ma, Daniel D. Cummins, Natalie Brooke Edelstein, Jerry Gomez, Aliza Khan, Masud Dikita Llewellyn, Tara Picudella, Sarah Rose Willsey, and Shikha Nangia J. Chem. Theory Comput., Just Accepted Manuscript • DOI: 10.1021/acs.jctc.6b00856 • Publication Date (Web): 12 Jan 2017 Downloaded from http://pubs.acs.org on January 13, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Chemical Theory and Computation is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Modeling diversity in structures of bacterial outer membrane lipids Huilin Ma, Daniel D. Cummins,† Natalie Brooke Edelstein,† Jerry Gomez,† Aliza Khan,† Masud Dikita Llewellyn,† Tara Picudella,† Sarah Rose Willsey,† and Shikha Nangia* Department of Biomedical and Chemical Engineering, Syracuse University, Syracuse NY 13244 †

Denotes equal contribution from these co-authors

*

Corresponding author email: [email protected]

ABSTRACT Lipopolysaccharides are vital components of the outer membrane of gram-negative bacteria, and they act as extremely strong stimulators of innate immunity in diverse eukaryotic species. The primary immunostimulatory center of the LPS molecule is lipid A—a disaccharide-bound lipophilic domain. Considering the broad diversity in bacterial species, there are variations in the lipid A structures and their immunogenic potency. In this work, we model the lipid A structures of eight commensal or human pathogenic bacterial species: Helicobacter pylori, Porphyromonas gingivalis, Bacteroides fragilis, Bordetella pertussis, Chlamydia trachomatis, Campylobacter jejuni, Neisseria meningitidis, and Salmonella minnesota. The membrane properties of these bacterial species were characterized and compared using molecular simulations. The structureproperty relationships that emerge from this lipid A molecular library highlight the roles of acyl chain lengths, number of chains, phosphorylation state, membrane composition and charge of the counterions in regulating the phase transition temperature of the membrane, diffusion coefficient of the lipids, and membrane thickness. Molecular and structural insights provided reveal the

1 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 47

diversity in bacterial outer membrane lipids and their contribution to human disease and immunity. 1. INTRODUCTION Gram-negative bacteria have evolved to protect themselves from hostile environments by developing a double membrane envelop surrounding their cellular contents. The outermost membrane is highly asymmetric and consists of an inner leaflet of glycerophospholipids and an outer leaflet that is rich in lipopolysaccharides (LPS) macromolecules that provide a negatively charged envelope around the bacterial cell.1-6 The LPS has three distinct domains—lipid A, a core oligosaccharide, and an O-antigen polysaccharide. Although all three domains play integral roles in the outer membrane, the amphiphilic lipid A domain anchors LPS to the outer membrane via hydrophobic interactions.2, 7 Lipid A also is a well-established endotoxin that stimulates innate immune response in diverse eukaryotic species.2, 8 Highly conserved among bacterial species, lipid A’s distinctive molecular structure is recognized as a pathogen-associated molecule by Toll-like receptor 4/myeloid differentiation factor 2 (TLR4/MD2) present on host immune cells.9 In response to lipid A exposure, host cells secrete pro-inflammatory cytokines to neutralize the bacteria and their endotoxic effects. Structurally, all lipid A molecules consist of a hydrophilic 1,4′bisphosphorylated disaccharide head and variable numbers of saturated fatty acid tails (Figure 1). The head unit, a β(1→6)-linked D-glucosamine disaccharide, is linked to acyl carbon chains at positions 2 and 3 as well as 2′ and 3′ via amide or ester linkages.10 The proximal reducing glucosamine residue (GlcN I) of the disaccharide head group has an α-phosphate at position 1 and an ester-bound phosphate at position 4′ of the distal nonreducing glucosamine residue (GlcN

2 ACS Paragon Plus Environment

Page 3 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

II). In some bacterial species, the primary acyl chains at positions 2′ and/or 3′ can be further esterified to support additional secondary carbon chains.10 The primary hydroxyl at position 6′ acts as the binding site of the core LPS oligosaccharide domain. Furthermore, to evade detection by the host immune system, bacteria undergo subtle modifications to alter their quintessential primary lipid A structure. For example, to evade detection by the host immune system, some bacteria undergo subtle modifications to alter the lipid A structural template at the glucosamine head group, or by the degree of phosphorylation or presence of phosphate substituents; other stress-induced changes may include the nature, number, location, and length of acyl chains.6, 11, 12 Often, these structural modifications are employed as an active response to changing environmental chemical stresses.13, 14 These structural modifications directly affect pathogenesis by changing outer membrane permeability and promoting resistance to antimicrobial peptides. There is, therefore, a need to understand structure-property relationships between the lipid A structures and the properties they confer to the outer membrane of a bacterial species. The experimental characterization of LPS remains challenging, due to the complexity and heterogeneity of the bacterial membrane. The isolation of an LPS macromolecule is non-trivial because the amphiphilic nature of lipid A causes micellization. Determination of high-resolution LPS structure requires iterative extractions followed by refinement and fragmentation. Such advances in extraction methods coupled with improved characterization techniques such as mass spectrometry (MS),15, 16 matrix-assisted laser desorption/ionization (MALDI)17, 18 and electrospray ionization (ESI)19, 20 have been invaluable. To expand beyond experimentally determined, static structural properties of these lipids, complementary computational approaches

3 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 47

are now being employed to assess the dynamical and thermodynamical properties of the bacterial membranes.21-28 Molecular simulations have become an indispensable tools to understand both the dynamic and nanoscale organization of bacterial membrane structures. While multiple computational techniques are available and have been used to investigate these membranes, coarse-grain (CG) representation provides an equitable balance between (1) the complexity and chemical specificity of membrane lipids and (2) the length and timescales required to characterize these systems.25, 2834

In our previous work, we adopted a multiscale approach to bridge atomistic and CG

representations by developing force field parameters for Pseudomonas aeruginosa LPS macromolecule.25 In this work, we extend the MARTINI force field parameters29 to a library of eight commensal or human pathogenic gram-negative bacteria species: Helicobacter pylori, Porphyromonas gingivalis, Bacteroides fragilis, Bordetella pertussis, Chlamydia trachomatis, Campylobacter jejuni, Neisseria meningitidis, and Salmonella minnesota. This representative set of gramnegative bacteria have lipid A domains that differ in the degree of phosphorylation, presence of phosphate substituents, glucosamine head group, as well as the nature, number, location, and length of acyl chains. After 80 independent simulations and close to 150 µs of total simulation time, this library of representative bacterial lipids provides an excellent example of the structureproperty relationships between lipid A structural modifications and bacterial outer membrane properties. The results highlight the role of acyl chain length, number of chains, and phosphorylation state in regulating the phase transition temperature of the membrane, and the role of membrane composition and charge of the counterions on membrane permeability. Prior to

4 ACS Paragon Plus Environment

Page 5 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

presenting results, background information on the eight bacterial species, their differences in preferred habitat, and lipid A structure (Table 1) are briefly discussed. 2. BACKGROUND Helicobacter pylori H. pylori is a spiral, rod-shaped bacteria that grow in the upper gastrointestinal tract. It is associated with a variety of gastrointestinal diseases such as peptic ulcers, gastric adenocarcinoma and can lead to stomach cancer.35 The H. pylori infection is found in over 50 percent of the world's population, especially among the young, and is transmitted through direct human contact. H. pylori can often be a lifelong infection in many of its hosts. The outer cell membrane of H. pylori is similar to that of other gram-negative bacteria. The temperature range supporting H. pylori’s growth is 307 K to 313 K, with an optimum temperature of 310 K, which is the average temperature of the human body. The chemical structure of its lipid A has glucosamine β-(1-6) disaccharide with phosphate at position 1 and four acyl chains (Figure S1). The acyl groups are (R)-3-hydroxyoctadecanoic acid, (R)-3-hydroxyhexadecanoic acid, and (R)3-(octadecanoyloxy)octadecanoic acid at the 2-, 3- and 2′-positions, respectively.36 Porphyromonas gingivalis P. gingivalis is a non-motile, gram-negative, endotoxic, anaerobic bacillus of the phylum Bacteriodetes found in gingival tissue and in atheromatous plaque and thrives best at 310 K. It is a suspected periodontal pathogen because it produces collagenase; however, about 25% of people without periodontitis test positive for P. gingilvalis, while 21% of patients with periodontitis test negative for P. gingivalis.37 The chemical structure of its lipid A comprises a

5 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 47

hydrophilic β-(1,6)-linked D-glucosamine disaccharide head that is monophosphorylated at position 1, and hydrophobic N- and/or O-acylation at positions 2, 3, 2′, and 3′ (Figure S2).38 Bacteroides fragilis The anaerobic bacteria B. fragilis is part of the normal microflora of the human large intestine. It is the most frequent cause of abdominal and wound infection in post-surgical procedures of the gastrointestinal or urogenital tract. B. fragilis is an enterobacteria with low endotoxicity, primarily attributed to its monophosphorylated lipid A that has five acyl residues, which are relatively long chains, each with 15-17 carbon atoms (Figure S3). The (R)-3hydroxyhexadecanoic acid and (R)-3-hydroxypentadecanoic acid residues are present at the positions 3' and 3 of the distal GlcN and reducing GlcN groups, respectively. The amino group at 3' position carries (R)-3-(13-methyltetradecanoyloxy)-15-methylhexadecanoic acid and the one at position 3 carries (R)-3-hydroxyhexadecanoic acid.39 Bordetella pertussis B. pertussis causes pertussis, a highly contagious respiratory infection commonly known as whooping cough, because of the characteristic sound patients make when they inhale. Transmission between people typically occurs by coughing or sneezing. Its lipid A structure contains a common bisphosphorylated disaccharide head group with hydroxytetradecanoic acid in the amide as well at the 3′ position (Figure S4).40 Its shorter acyl chains enable the bacteria to escape the receptor signaling system. Chlamydia trachomatis C. trachomatis is the most common cause of sexually transmitted bacterial infection, with more than 90 million new cases annually worldwide.41 C. trachomatis also is a cause of pelvic 6 ACS Paragon Plus Environment

Page 7 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

inflammatory disease in women and is a significant cause of blindness in the developing world, where treatment is largely absent. Members of the Chlamydiae genus are obligate intracellular parasites, and C. trachomatis is specifically reliant on human cells to carry out its life cycle. The physiological effects of C. trachomatis, like all gram-negative bacteria, are invoked by the bacteria’s lipid A component. Mass spectrometry shows that the LPS of C. trachomatis is composed mainly of a glucosamine disaccharide with five-fatty acid chains and two phosphates (Figure S5). The long fatty acid chains of C. trachomatis (up to 21 carbons) are anomalous to most gram-negative bacteria lipid A components, and is thought to bring about its relatively low toxicity.4, 42 Campylobacter jejuni C. jejuni is a microaerobic strain of proteobacteria with a helical shape. It is primarily responsible for food-borne bacterial gastroenteritis.43 C. jejuni is often found in animal feces and is transmitted easily between animals and humans. Its capacity to form a biofilm increases the survival of C. jejuni under detrimental conditions; when in a biofilm, the bacteria is onethousand times more resistant to disinfectants.44 Unlike the other lipid A structures, in C. jejuni one of the glucosamine residues of the lipid A head group is replaced by a GlcN3N monosaccharide, a phosphorylated 2,3 diamino-2,3-dideoxy-D-glucose (GlcN3N) disaccharide (Figure S6). Neisseria meningitidis N. meningitidis is a leading cause of bacterial meningitis and sepsis worldwide.45 The meningococcal LPS has a bisphosphorylated disaccharide head group with 12:0(3-OH) acyl chains bound to each of the two hydroxyl groups at positions 3 and 3', and 14:0(3- OH) acyl chains linked to the amino groups at positions 2 and 2', and the hydroxyl groups of the amide7 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 47

linked chains acylated by 12:0 carbon tails.46 Additionally, O-phosphorylethanolamine residues cap the phosphates at positions 1 and 4' (Figure S7). Salmonella minnesota Typically, S. minnesota, the second leading cause of intestinal infections, is transmitted through ingestion of contaminated food. S. minnesota infection commonly occurs in the intestinal tract and is associated with bloody diarrhea, abdominal cramps, and other related symptoms. Most Salmonella serotypes are able to grow and thrive in environments whose temperatures falls between 280 K and 321 K. The S. minnesota lipid A has a typical 1,4′-bisphosphorylated disaccharide head group with seven acyl chains that are 12–14 carbons in length. Position 2 and 3 have (R)-3-hydroxy fatty acids and 2' and 3' have (R)-3-acyloxyacyl residues. Additionally, hexadecanoic acid and dodecanoic acid residues are on the (R)-3-hydroxytetradecanoic acid at positions 2 and 2', respectively (Figure S8).47

3.

METHODS

3.1 Parameterization The CG parameterization of the library of eight lipid A molecules is developed on the MARTINI many-to-one mapping approach;29 and in most cases, four heavy atoms are mapped onto one bead. The structural similarities in the disaccharide head groups and the dissimilarities in the phosphorylation state and acyl chain patterns have been incorporated into the parameterization (see Table 1). The proximal reducing (GlcN I) and non-reducing (GlcN II) glucosamine residues were mapped individually to four beads with bead types ranging from P1–P4, based on the number of hydroxyl groups (Figure 2). The phosphates at positions 1 and 4′ were assigned a Qa

8 ACS Paragon Plus Environment

Page 9 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

bead type with a unit negative charge. The beads linking the acyl carbon chains via amide or ester linkages at positions 2, 3, 2′ and 3′ were assigned Na bead type. For N. meningitidis lipid A, the additional NH3+ groups linked to the phosphates were assigned Qd bead type. The acyl chain beads were assigned C1 bead type (Figure 2). The GROMACS compatible topology file with equilibrium bonded distances, bond angle parameters and their corresponding force constants will be distributed freely to the interested researchers upon request. Monovalent (Na+) or divalent (Ca2+) counterions were used to make the systems electrically neutral. The ion parameterization accounts for the first hydration shell around the ion, and both ions were assigned the Qd bead type. As in our previous work,25 no additional parameterization of Ca2+ ions was performed, and the only difference in Na+ and Ca2+ was their net integral charge.

3.2 Simulation and analysis details Eight sets of simulations, which include variations in membrane composition, membrane size, solvent, counterions, and temperature were performed for each of the eight membrane systems (Tables 2 and 3). The simulations were performed using the molecular dynamics engine GROMACS, version 5.1.2.48-50 The workflow of the simulations involved the initial construction of the membrane, energy minimization, short isothermal-isochoric (NVT) and isothermal-isobaric (NPT) equilibration runs, and long-production NPT runs. For each bacterial species, the membrane was built using a python script, which is a locally modified enhanced version of the insane script, a versatile membrane-building tool routinely used in constructing CG membranes.51 The library of eight bacterial lipids was coded in the freely distributed insane script programmed in python. The workflow of the insane script was not changed from the published version.51 9 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 47

The outer leaflet of the membrane is a mixture of lipid A and 1,2-dihexadecanoyl-sn-glycero-3phosphoethanolamine (DPPE) in 9:1 ratio for all eight sets. For the inner leaflet either a pure DPPE (Sets I-V) or a mixture of DPPE, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE), and Cardiolipin (CDL2) in the ratio 7:2:1 (Set VI-VIII) was used.52 POPE and CDL2 lipids have −1 and −2 charge, respectively. In generating the membranes, the total numbers of acyl chains in the inner and outer leaflet were kept the same to avoid unphysical undulations of the membrane. The membranes were solvated with either standard water (W) or polarizable (PW) MARTINI water53 as specified (Tables 2 and 3). All systems were made charge neutral by adding Na+ or Ca2+ counterions. In all sets, except Set III, a 10×10 nm2 membrane patch was generated spanning the xy-plane. In Set III, a larger 25 × 25 nm2 membrane cross-section was built to enable the comparison of membrane properties of smaller systems. Periodic boundary conditions were applied in all three dimensions. Energy minimization was performed using the steepest-decent algorithm with a 20 fs time-step until the maximum force on any bead was below the tolerance parameter of 10 kJmol−1nm−1. The NVT and NPT simulation runs were performed for 0.2 µs. The production simulations were run for at least 2 µs and up to 10 µs in some cases with a 20 fs time-step. Semi-isotropic pressure coupling was used, and systems were maintained at 1 bar using the Berendsen barostat with time constant, τp = 4.0 ps. Temperature was maintained at 310 K by independently coupling the lipids and solvent to an external velocity rescaling thermostat with τT = 1.0 ps. The heating scans were performed for a wider temperature range, varying from 275–360 K (Table 2 and 3). The neighbor list was updated every 25 steps using 1.4 and 1.2 nm for short-range van der Waals and electrostatic cutoffs, respectively. For simulations with polarizable water, particle mesh ewald

10 ACS Paragon Plus Environment

Page 11 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

(PME) algorithm was used for the long-range electrostatics, with an electrostatic screening constant of εr = 2.5. The structural and dynamic properties of the membranes were compared by computing area per lipid, membrane thickness, density profiles, order parameters, phase transition temperatures, and diffusion coefficients. The membrane microstructure was quantified by the average area per lipid (AL) and membrane thickness (DM), and hydrophobic thickness (DH). For bacterial membranes, the AL values were computed by dividing the cross-sectional area of the membrane by the number of lipid A molecules in the leaflet. Standard utilities available in the GROMACS software suite were employed for quantitative analysis. To determine Tm, the characteristic phase-transition temperature values for the model systems, we performed annealing simulations starting from well-equilibrated configurations to mimic phase transition conditions. The heating scans were performed over the 275–360 K temperature range with intermediate temperatures of 292, 309, 326, and 343 K and 2 µs of simulation time.

4.

RESULTS

4.1.

Bonded parameters

The analysis of bond distances and bond angles was performed for all eight bacterial lipid A membranes using the same protocol. For the ease of comparison, the analysis was sub-divided into—the head group and acyl chains. Since the proximal reducing (GlcN I) and non-reducing (GlcN II) glucosamine residues in the lipid A head group are the same in seven of the eight species investigated (the exception being C. jejuni), the average bond distance frequency distribution observed is similar (Figure 3). A unimodal frequency distribution of the bonded pairs

11 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 47

centered at 0.30 ± 0.01 nm shows that the bonded pairs in a saccharide head group range between 0.29-0.39 nm both in atomistic and CG simulations. A similar frequency distribution of the average internal angles also shows a unimodal distribution centered at 75.2°±1.2° for all eight membranes (Figure 4). These results are consistent with analysis reported earlier for P. aeruginosa CG parameterization.25 For the acyl chains in all the systems, the average bond distance distribution shows unimodal distribution centered at 0.442 ± 0.012 nm (Figure 5). The average bond distance of the acyl tails is +0.142 nm larger than the head group bond distance because head group beads are smaller and, unlike tails, do not always follow the 4-to-1 mapping prescription. The acyl chain bond angle distribution is unimodal for all of the lipid A membranes, but the location of the peak depends on the specific bacterial lipid A structure (Figure 6). For example, lipid A structures with 17-21 carbon acyls chains (H. pylori, P. gingivalis, B. fragilis, and C. trachomatis) have peaks centered at 158 ± 2°, while structures with shorter 14-16 carbon acyl chains (C. jejuni, N. meningitidis, and S. minnesota) have peaks at 151 ± 1°, and B. pertussis with shortest (10-14) carbon acyl chains has peak at 145°. Despite having exactly same bond angle parameters for the acyl beads, the variation in the average angle with the acyl chain length has significant implications for membrane properties such as membrane thickness, area per lipid, and phase transition temperature. These findings not only demonstrate that the CG parameterization is able to capture the molecular differences in these lipid A structures, but also validate the reliability of membrane properties predicted by the force field. The dihedral angles for the disaccharide head group or the acyl chain beads were not included because reports in the literature indicated that including dihedral angle parameters requires the use of time steps that are an order of magnitude smaller than those optimal for CG simulations.54, 12 ACS Paragon Plus Environment

Page 13 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

55

Despite the absence of explicit dihedral angle parameters, the average dihedral angle was

computed for 2 µs trajectory. In all eight membranes (Set II) the acyl chains are linear with average dihedral angle of 180° ± 11° or (0°± 11) through the trajectory (Figure S9). 4.2. Area per lipid (AL) and phase-transition temperature (Tm) In general, the AL for a lipid increases with increases in temperature as it acquires higher thermal energy, and if varied over a long-enough temperature range, the lipids undergo phase transition marked by a sharp increase in the AL versus T plot. In this work, the variation in AL for all eight membranes was computed over a broad 275-360 K temperature range (Figure 7) to determine the phase transition melting temperature (Tm). To determine the Tm more precisely, the change in AL (∆AL) as a function of temperature was computed as a function of T, where the peak in the curve reflects a sharp change in the area per lipid over a small change in temperature for individual membrane (Figure 8). As with the AL values of the lipids, characteristic changes in lipid tails were observed for the membranes below and above their Tm values. The tetra-acylated H. pylori lipid A has the smallest AL value compared to the penta-, hexa- and hepta-acylated lipid As. The AL values of the penta-acylated lipid A (P. gingivalis, B. fragilis, C. trachomatis, and B. pertussis) are 1.2-1.3 nm2 in the ordered phase below their phase transition temperature. In the disordered phase, about 10 K above the Tm, the AL values increase to 1.5–1.6 nm2. With six tails, the AL values increase for both C. jejuni and N. meningitidis membranes. The computed AL values are in the range 1.45–1.48 nm2 and 1.85–1.9 nm2, 10 K below and above their Tm, respectively. The heptaacylated S. minnesota, has the highest AL of 1.65 nm2 and 2.0–2.1 nm2 10 K below and above the Tm. Snapshots of the membranes below the Tm show ordered and fully extended lipid tails and disordered and compacted lipid tails above the Tm (Figure 9). 13 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 47

Changing the lower leaflet composition (Sets VI and VII) to include negatively charge POPG and cardiolipin resulted in lipid A AL values that were 0.2 nm2 larger than those in Set II. The slight increase in lipid A AL is the direct consequence of the presence of charge in the lower leaflet, which causes the increase in the bilayer cross-sectional area. 4.3. Membrane thickness (DM) The DM values were computed by measuring the perpendicular distance between the planes formed by the phosphate head groups of the top leaflet and the lipid heads from the bottom leaflet. As expected, DM is larger at temperatures below Tm and smaller above the Tm. To capture this change in membrane structure, thickness was computed as a function of temperature for all eight membranes, and the results are shown in Figure 10. All membranes, except B. pertussis, show ~0.51 nm decrease in DM after phase transition, which is consistent with the change in thickness observed experimentally in S. minnesota over a 30 K variation in temperature.56 For B. pertussis, the decrease is only about 0.24 nm because of its short 10-12 carbon acyl chains relative to others (Table 4). The simulated S. minnesota DM was found 0.441 and 0.393 nm 10 K below and above the Tm (Set II), which is in good agreement with the electron density profile for rough mutant lipopolysaccharides Re (LPS Re) of S. Minnesota (strain R595).56 The experimentally reported upper leaflet head-group to lower leaflet head-group distance of the bilayer is 0.429 nm at 293 K and 0.387 nm at 323 K.56 4.4. Density profiles The distribution of individual components within the lamellar asymmetrical bilayers was computed for all the membranes 10 K below their Tm. Because the lipids are in a thermal

14 ACS Paragon Plus Environment

Page 15 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

equilibrium, they adopt a highly variable instantaneous molecular orientation; therefore, density profiles of all membrane components were calculated over 1 µs of the simulation trajectory to account for ensemble averaging. The density profiles were computed for Set VI to determine the key features of the membranes for comparison (Figure S10). At each membrane interface, the Ca2+ ions interact with the lipid head groups and do not penetrate the hydrophobic tails of outer and inner leaflets. The Ca2+ ion density is more pronounced in lipid A head groups of N. meningitidis due to the presence of additional phosphorylated residues that cap the phosphates at positions 1 and 4'. This also explains the higher density of water surrounding the lipid A head groups than phospholipids in the inner leaflet. The counterion peaks in the density profile were used as a measure of the membrane thickness. Additionally, the density profile of C1 beads (representing the acyl chains in both leaflets) was plotted as a function of the membrane normal (z-coordinate), as a measure of the hydrophobic thickness (DH). The hydrophobic thickness lies in the 2.5-3.3 nm range depending on the number of carbons in the acyl chains. The C. trachomatis membrane with an average of 17 carbon acyl chains has the highest hydrophobic thickness of 3.3 nm, which can be an important factor in determining the nature of the transmembrane porin proteins that can span the relatively thick outer membrane. Additionally, the high hydrophobic thickness in C. trachomatis also prevents the penetration of water deeper into the lipid A head groups. The role of acyl chain length on the overall thickness of the membrane is discussed in Section 5.2.

15 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

4.5.

Page 16 of 47

Radial distribution function

The radial distribution functions (RDF) of Na+ (Set IV) and Ca2+ (Set V) ions interacting with the negatively charged phosphate and carboxyl groups of lipid A were calculated for all eight membranes (Figure 11) in polarizable water model. The change in water model from polarizable to non-polariazable did not alter the RDF functions. The curves for all membranes show similar trends, but notable is the peak for Ca2+-carboxylate, which occurs at a longer distance (~1.3 nm) than the Na+-carboxylate peak (at 0.51 nm). The peak positions imply that Na+ is able to penetrate deeper into the membrane and interact with carboxylate groups that lie below negatively charged phosphates. The discussion of interaction of ions with phosphorylated lipid head groups is discussed in Sec. 5.4.

5.

DISCUSSION

The complexity of gram-negative bacterial outer membranes has been a limiting factor in the computational modeling and characterization of these membranes. Until recently, the inherently asymmetric outer membranes (with LPS/phospholipids leaflets) were simplified57-65 as symmetric phospholipids in molecular simulations due to the lack of atomistic and coarsegrained force field parameterization. The development of LPS models is in its infancy with models available for one or two bacterial species in atomistic21-24, 26 and coarse-grain representation.25, 55 There is, therefore, limited molecular level understanding of the role of the number of acyl chains, the length of acyl chains, and phosphorylation of a lipid on membrane properties. The library of eight coarse-grained bacterial lipids models studied here will provide a systematic evaluation of the factors contributing to membrane properties.

16 ACS Paragon Plus Environment

Page 17 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

5.1. Effect of number of acyl chains on membrane properties Bacterial species adopt various acylation patterns to promote their survival by evading detection by the host innate immune system. Others have shown that penta-, tetra,- and tri-acylated lipid A analogs stimulate a smaller immune response and lower cytokines levels compared to hexaacylated lipid A.10, 13 Some bacteria actively modify lipid A in response to changes in temperature of the host. For example, Yersinia pestis produces hexa-acylated lipid A under ambient conditions but shifts to a tetra-acylated form at temperatures close to mammalian body temperature.10, 66 In S. typhimurium and P. aeruginosa, acylation patterns are modified by enzyme activity in response to hostile chemical stimuli, such as depletion of cationic counterions, changes in pH, and presence of antimicrobial peptides, among others.10 The variability in acyl chains permits up and down regulation of outer membrane permeability, lipid diffusivity, structural integrity, and the Tm. The eight bacterial lipids studied here represent a range in lipid A structural diversity with 4–7 acyl chains. Analyzing the AL data (Table 4) 10 K below the phase transition (Tm −10 K) shows that on average, each lipid tail contributes ~0.24 nm2 to the area occupied by a lipid molecule. While this rule-of-thumb holds well for the membranes below their phase transition temperature, the contribution of the lipid tails increases to ~0.31 nm2 above the Tm in the disordered phase. Additionally, increasing the number of chains decreases the diffusivity of the lipid. Diffusion coefficient data from simulations of the eight lipid, under similar physiochemical conditions show that hepta-acylated lipid A is an order of magnitude lower than the tetra-acylated analogs, and values of hexa- and penta-acylated lipid A range between the two extremes (Figure S11).

17 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 47

The results show that S. minnesota has the lowest diffusion coefficient, while H. pylori has about an order of magnitude higher diffusion coefficient (Table 4). Although there is a difference in the phosphorylation state of these two lipids, we attribute the differences in D values primarily to the number of lipids and differences in molecular weight. The hexa-acylated C. jejuni and N. meningitidis have D values in the same order of magnitude. A trend in the D values for the penta-acylated lipid A membranes (P. gingivalis, B. fragilis, B. pertussis, and C. trachomatis) was less apparent, but B. pertussis with the shortest acyl chain length lipid A has the highest diffusion coefficient (Figure S11). As evident from the AL and D data, acyl chain addition or deletion has significant effect on the membrane properties. The acyl chain variability is an excellent example of structural-property relationships of how bacteria can employ this attribute to adapt to their habitat.

5.2. Effect of acyl chain length on membrane properties

Membrane microstructure is dependent on the lipid-lipid interactions between adjacent molecules and is intimately tied to the length of the acyl chains and the average hydrophobic thickness. A membrane with a larger hydrophobic thickness experiences increased van der Waals attractions between neighboring lipids, resulting in lower area per lipid and a higher phase transition temperature.67 The results from the eight membranes studied here reflect this expected trend. Considering B. pertussis and C. trachomatis membranes, which have the shortest and longest acyl chains of the group, respectively, C. trachomatis (with at least 2–6 additional carbons in the acyl chains) has a higher hydrophobic thickness (∆DH = +0.6 nm), lower area per lipid (∆AL = −0.07 nm2), and a higher phase transition temperature (∆Tm = + 27 K) than does B. pertussis. Density profiles of all membranes (Figure 13) show that the difference in the total

18 ACS Paragon Plus Environment

Page 19 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

membrane thickness (∆DM = +0.6 nm) arises due to the hydrophobic thickness alone and not due to the disaccharide head groups. In addition, comparing density profiles (Figure S10) of P. gingivalis and B. fragilis membranes, whose lipid A structures are similar in terms of the number of acyl chains and phosphorylation state, and differ only by one carbon in two of their acyl chains, have similar values for AL, DM, DH, and Tm (Table 4). These results highlight once again, the importance of the acyl chain structure, particularly the sensitivity of the chain length to the properties of the membrane.

5.3. Effect of phosphorylation on membrane properties The phosphorylation state of the disaccharide head group influences lipid A-mediated endotoxicity. Bacterial species with monophosphorylated lipids are resistant to antimicrobial peptides and are less active than the diphosphorylated lipids.68 For example, H. pylori consists of a tetra-acylated lipid A that lacks the 4ʹ-phosphate group to evade detection by TLR4 and resists action by antimicrobial peptides.10 In Salmonella typhimurium, neutralizing the phosphates results in increased antimicrobial resistance and decreased immunogenic response.68 Beside the biological activity, change in the phosphorylation pattern of the lipid head group affects the ordered to fluid phase transition in the membranes because of the change in the surface charge density. This trend was observed in lipid A modifications that were performed on a pair of representative systems, specifically, C. jejuni and P. gingivalis. Deletion of 4ʹphosphate in C. jejuni resulted in a +7 K change in the Tm of the modified lipid A membrane compared to its native membrane, whereas addition of 4ʹ-phosphate in P. gingivalis lipid A resulted in −11 K change in the Tm compared to its native counterpart (Figure 13). In case of C. jejuni modification, deletion of the phosphate group, decreases the ionic interactions among the lipid A head groups, reduces the surface charge density, enhances the lipid-lipid packing, 19 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 47

resulting in higher Tm. The converse of this explanation is valid for modified P. gingivalis. Both these examples of lipid A modifications, illustrate the sensitivity of the lipid A structure and the relative ease with which bacterial species can mask the phosphoryl groups in response to environmental factors such as changes in the pH, presence of antimicrobial peptides, or nature of counter ions.

5.4. Role of counterions in the surrounding medium Ionic environment and the charge distribution on bacterial outer membrane can induce gross alterations in membrane microenvironment. The counterions in the surrounding solution play an important role in lipid A head group organization by screening the strong electrostatic repulsions that exist between the charged phosphates groups. Structural changes of the outer membrane in response to different salt conditions have been reported.69

The negatively charged phosphates act as coordination sites for divalent ions to chelate adjacent LPS molecules. Comparison of phosphate-Ca2+ radial distribution functions, g P-Ca 2+ ( r ) in H. pylori, C jejuni, and N. meningitidis, all show a predominant peak at r = 0.5 nm, irrespective of the number of phosphates on lipid A head groups (Figure 14). The separation distance of 0.5 nm is particularly important because it is the signature of the closest non-bonded distance between two CG beads. Unlike g P-Ca 2+ ( r ), phosphate-phosphate (P-P) radial distribution functions gP-P (r ) clearly show differences in the phosphorylation states among these lipid A structures. H. pylori, with one phosphate (at position 1 of the disaccharide head group), shows a low intensity P-P peak at 0.5 nm, mediated by the Ca2+ ions, but the majority of the phosphates are less organized, illustrated by the broader peak centered at 0.8 nm (Figure 14A). On the other hand, C. jejuni,

20 ACS Paragon Plus Environment

Page 21 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

with two phosphate groups (at positions 1 and 4'), shows well-defined peaks at 0.6 and 0.9 nm that correspond to the head-on intermolecular and the intramolecular P-P interactions, respectively (Figure 14B). Finally, N. meningitidis, with four phosphates (bonded pair at positions 1 and 4'), shows a bonded P-P peak at 0.3 nm, and broader peak centered at 0.9 nm. The differences in the ionic charge density at the lipid A-water interface is shown in the inset snapshots in Figure 14. As expected, H. pylori shows the lowest charge density, which may contribute to its higher resistance to cationic antimicrobial peptides.

5.5.

Aiding knowledge-based antimicrobial peptide design

The investigation of the eight bacterial outer membranes reveals that the microstructure and interfacial properties of the membrane vary among bacterial species. The structure-property relationship between the lipid A molecules and the membrane properties should be factored into consideration in designing synthetic antimicrobial peptides that can target a desired pathogenic bacterial species. Designing efficient antimicrobial peptides, however, is arduous because of the variability in peptide length (20-50 amino acids), charge (number of positively charged amino acids), secondary structure (helical, β-stranded, loop, or unstructured), and amphipathicity (percent hydrophilic/phobic residues). To mitigate the challenge, one approach can be to utilize the ensemble-averaged membrane properties (membrane thickness, hydrophobic thickness, membrane lipid density, and charge density) from the simulations as a reference for knowledgebased design. Based on the bacteria specific data set, antimicrobial peptides can be tailored for optimal length and amphipathicity that will allow them to partition into the bacterial outer membrane successfully. Additionally, the membrane charge distribution can be utilized as a reference for selection amount of charge and location of charged residues so that the

21 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 47

antimicrobial peptides associate with lipid A head groups. In the next phase of this work, current library of bacterial membrane properties will be employed to design knowledge-based antimicrobial peptides.

6. CONCLUSIONS This work provides thermodynamic and dynamical properties of a diverse set of eight bacterial membranes of commensal or human pathogenic gram-negative bacteria species: Helicobacter pylori, Porphyromonas gingivalis, Bacteroides fragilis, Bordetella pertussis, Chlamydia trachomatis, Campylobacter jejuni, Neisseria meningitidis, and Salmonella minnesota. This representative set of gram-negative bacteria has lipid A domains that differ in the degree of phosphorylation, presence of phosphate substituents, glucosamine head group, as well as the nature, number, location, and length of acyl chains. After multiple independent simulations for all membranes, several key characteristics emerge. First, we find that on average each lipid tail contributes ~0.24 nm2 to the total area of the lipid; therefore AL values of hepta-acylated S. minneosta and tetra-acylated H. pylori lipid A are in 7:4 ratio. Second, membranes composed of longer acyl chain lipid A have smaller AL and a higher phase transition temperature compared to their shorter acyl chain counterparts. Third, the membrane composition and charge of the inner leaflet can influence the phase transition temperature of the membrane by 20-30 K. Fourth, monovalent ions bury themselves deeper into the membrane head groups, whereas divalent ions are at the surface and act as chelating agents, binding to the phosphates on adjacent lipid A molecules. The insights from this work, coupled with the development of library of lipid A coarse-grained topology, will facilitate advances in knowledge-based design of antimicrobial agents. 22 ACS Paragon Plus Environment

Page 23 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

7. ACKNOWLEDGMENTS The authors thank XSEDE supercomputing facility and Information and Technology Services at Syracuse University (Eric Sedore, Larne Pekowsky, and Michael R. Brady) for providing computational resources for part of the simulations presented here. The authors also thank various funding sources that funded one or more undergraduate students for this project. We thank National Science Foundation (NSF) EFRI-MIKS-1137186 grant for providing financial support for SRW, TP, and MDL and NSF DMR-1460784 (REU) for DDC; The Student Association, the Vice Chancellor and Provost for Academic Affairs, the Associate Provost for Academic Programs, and the Vice President for Research, and the New York State Collegiate Science and Technology Entry Program at Syracuse University for JG; Syracuse University’s College of Engineering and Computer Science Dean's Leadership Fund for NBE; and Syracuse University for AK and HM.

Supporting Information: The supporting document includes-details of Set VIII; atomistic structure of lipid A molecules; dihedral angle fluctuations as a function of simulation time; density profile for Set VI membranes; snapshots of thermal phase transition of Set VII membranes; diffusion coefficient of lipid A molecules in Set I. This information is available free of charge via the Internet at http://pubs.acs.org

23 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 47

REFERENCES 1.

2.

3. 4. 5. 6. 7. 8. 9. 10. 11.

12.

13. 14.

15.

16.

Knirel, Y. A.; Kochetkov, N. K., The structure of lipopolysaccharides of gram-negative bacteria. The structure of O-antigens - a review. Biochemistry-Moscow 1994, 59 (12), 1325-1383. Zahringer, U.; Lindner, B.; Rietschel, E. T., Molecular-structure of lipid-A, The endotoxic center of bacterial lipopolysaccharides. Adv. Carbohydr. Chem. Biochem. 1994, 50, 211-276. Alexander, C.; Rietschel, E. T., Bacterial lipopolysaccharides and innate immunity. J. Endotoxin Res. 2001, 7 (3), 167-202. Erridge, C.; Bennett-Guerrero, E.; Poxton, I. R., Structure and function of lipopolysaccharides. Microbes Infect. 2002, 4 (8), 837-851. Caroff, M.; Karibian, D., Structure of bacterial lipopolysaccharides. Carbohydr. Res. 2003, 338 (23), 2431-2447. Dixon, D. R.; Darveau, R. P., Lipopolysaccharide heterogeneity: Innate host responses to bacterial modification of lipid A structure. J. Dent. Res. 2005, 84 (7), 584-595. Needham, B. D.; Trent, M. S., Fortifying the barrier: The impact of lipid a remodelling on bacterial pathogenesis. Nat. Rev. Microbiol. 2013, 11 (7), 467-481. Raetz, C. R. H.; Whitfield, C., Lipopolysaccharide endotoxins. Annu. Rev. Biochem 2002, 71, 635-700. Maeshima, N.; Fernandez, R. C., Recognition of lipid A variants by the TLR4-MD-2 receptor complex. Front.Cell.Infect. Microbiol. 2013, 3. doi: 10.3389/fcimb.2013.00003 Needham, B. D.; Trent, M. S., Fortifying the barrier: The impact of lipid a remodelling on bacterial pathogenesis. Nat. Rev. Micro. 2013, 11 (7), 467-481. Park, B. S.; Song, D. H.; Kim, H. M.; Choi, B. S.; Lee, H.; Lee, J. O., The structural basis of lipopolysaccharide recognition by the TLR4-MD-2 complex. Nature 2009, 458 (7242), 1191-U1130. Brandenburg, K.; Andrä, J.; Müller, M.; Koch, M. H. J.; Garidel, P., Physicochemical properties of bacterial glycopolymers in relation to bioactivity. Carbohydr. Res. 2003, 338 (23), 2477-2489. Munford, R. S., Sensing gram-negative bacterial lipopolysaccharides: A human disease determinant? Infect. Immun. 2008, 76 (2), 454-465. Maldonado, R. F.; Sa-Correia, I.; Valvano, M. A., Lipopolysaccharide modification in gram-negative bacteria during chronic infection. FEMS Microbiol. Rev. 2016, 40 (4), 480-493. Seydel, U.; Lindner, B.; Wollenweber, H. W.; Rietschel, E. T., Structural studies on the lipid-a component of enterobacterial lipopolysaccharides by laser desorption massspectrometry - location of acyl-groups at the lipid-a backbone. Eur. J. Biochem. 1984, 145 (3), 505-509. Gibson, B. W.; Melaugh, W.; Phillips, N. J.; Apicella, M. A.; Campagnari, A. A.; Griffiss, J. M., Investigation of the structural heterogeneity of lipooligosaccharides from pathogenic Haemophilus and Neisseria species and of R-type lipopolysaccharides from Salmonella typhimurium by electrospray mass-spectrometry. J. Bacteriol. 1993, 175 (9), 2702-2712.

24 ACS Paragon Plus Environment

Page 25 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Molinaro, A.; Silipo, A.; Lanzetta, R.; Parrilli, M.; Malvagna, P.; Evidente, A.; Surico, G., Determination of the structure of the lipid a fraction from the lipopolysaccharide of DI-TOF mass spectrometry. Eur. J. Org. Chem. 2002, (18), 3119-3125. Nichols, F. C.; Bajrami, B.; Clark, R. B.; Housley, W.; Yao, X. D., Free lipid A isolated from Porphyromonas gingivalis lipopolysaccharide is contaminated with phosphorylated dihydroceramide lipids: Recovery in diseased dental samples. Infect. Immun. 2012, 80 (2), 860-874. Knirel, Y. A.; Kondakova, A. N.; Bystrova, O. V.; Lindner, B.; Shaikhutdinova, R. Z.; Dentovskaya, S. V.; Anisimov, A. P., New features of Yersinia lipopolysaccharide structures as revealed by high-resolution electrospray ionization mass spectrometry. Adv. Sci. Letts. 2008, 1 (2), 192-198. Man-Kupisinska, A.; Bobko, E.; Gozdziewicz, T. K.; Maciejewska, A.; Jachymek, W.; Lugowski, C.; Lukasiewicz, J., Fractionation and analysis of lipopolysaccharide-derived oligosaccharides by zwitterionic-type hydrophilic interaction liquid chromatography coupled with electrospray ionisation mass spectrometry. Carbohydr. Res. 2016, 427, 2937. 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 (45), 13381-13388. Kirschner, K. N.; Lins, R. D.; Maass, A.; Soares, T. A., A GLYCAM-based force field for simulations of lipopolysaccharide membranes: Parametrization and validation. J. Chem. Theory Comput. 2012, 8 (11), 4719-4731. Wu, E. L.; Engstrom, 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 (6), 1444-1455. 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 (11), 2493-2502. 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 (46), 14668-14682. Patel, D. S.; Re, S.; Wu, E. L.; Qi, Y. F.; Klebba, P. E.; Widmalm, G.; Yeom, M. S.; Sugita, Y.; Im, W., Dynamics and interactions of OmpF and LPS: Influence on pore accessibility and ion permeability. Biophys. J. 2016, 110 (4), 930-938. Pavlova, A.; Hwang, H.; Lundquist, K.; Balusek, C.; Gumbart, J. C., Living on the edge: Simulations of bacterial outer-membrane proteins. Biochim. Biophys. Acta, Biomembr. 2016, 1858 (7), 1753-1759. DeSalvo, S. C.; Liu, Y.; Choudhary, G. S.; Ren, D.; Nangia, S.; Sureshkumar, R., Signaling factor interactions with polysaccharide aggregates of bacterial biofilms. Langmuir 2015, 31 (6), 1958-1966. Marrink, S. J.; Risselada, H. J.; Yefimov, S.; Tieleman, D. P.; de Vries, A. H., The MARTINI force field: Coarse grained model for biomolecular simulations. J. Phys. Chem. B 2007, 111 (27), 7812-7824. Yesylevskyy, S. O.; Schafer, L. V.; Sengupta, D.; Marrink, S. J., Polarizable water model for the coarse-grained MARTINI force field. PLoS Comput. Biol. 2010, 6 (6) 1-17.

25 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

31. 32.

33.

34.

35. 36.

37.

38.

39.

40.

41. 42.

43.

44.

45.

Page 26 of 47

Nangia, S.; Sureshkumar, R., Effects of nanoparticle charge and shape anisotropy on translocation through cell membranes. Langmuir 2012, 28 (51), 17666-17671. Jiang, W.; Luo, J.; Nangia, S., Multiscale approach to investigate self-assembly of telodendrimer based nanocarriers for anticancer drug delivery. Langmuir 2015, 31 (14), 4270-4280. Irudayanathan, F. J.; Trasatti, J. P.; Karande, P.; Nangia, S., Molecular architecture of the blood brain barrier tight junction proteins–a synergistic computational and in vitro approach. J. Phys. Chem. B 2016, 120 (1), 77-88. Reem, E.; Manuela, R.; Jörg, O.; Yon, R.; Wenjuan, J.; Shikha, N.; Cerasela Zoica, D., Combinatorial approaches to evaluate nanodiamond uptake and induced cellular fate. Nanotechnology 2016, 27 (8), 085107. Kusters, J. G.; van Vliet, A. H. M.; Kuipers, E. J., Pathogenesis of Helicobacter pylori infection. Clin. Microbiol. Rev. 2006, 19 (3), 449-490. Moran, A. P.; Lindner, B.; Walsh, E. J., Structural characterization of the lipid A component of Helicobacter pylori rough- and smooth-form lipopolysaccharides. J. Bacteriol. 1997, 179 (20), 6453-6463. Griffen, A. L.; Becker, M. R.; Lyons, S. R.; Moeschberger, M. L.; Leys, E. J., Prevalence of Porphyromonas gingivalis and periodontal health status. J. Clin. Microbiol. 1998, 36 (11), 3239-3242. Kumada, H.; Haishima, Y.; Umemoto, T.; Tanamoto, K. I., Structural study on the free lipid-A isolated from lipopolysaccharide of Porphyromonas gingivalis. J. Bacteriol. 1995, 177 (8), 2098-2106. Weintraub, A.; ZÄHringer, U.; Wollenweber, H.-W.; Seydel, U.; Rietschel, E. T., Structural characterization of the lipid a component of Bacteroides fragilis strain nctc 9343 lipopolysaccharide. Eur. J. Biochem. 1989, 183 (2), 425-431. Basheer, S. M.; Bouchez, V.; Novikov, A.; Augusto, L. A.; Guiso, N.; Caroff, M., Structure activity characterization of Bordetella petrii lipid A, from environment to human isolates. Biochimie. 2016, 120, 87-95. Brunham, R. C.; Rey-Ladino, J., Immunology of Chlamydia infection: Implications for a Chlamydia trachomatis vaccine. Nat. Rev. Immunol. 2005, 5 (2), 149-161. Qureshi, N.; Kaltashov, I.; Walker, K.; Doroshenko, V.; Cotter, R. J.; Takayama, K.; Sievert, T. R.; Rice, P. A.; Lin, J. S.; Golenbock, D. T., Structure of the monophosphoryl lipid A moiety obtained from the lipopolysaccharide of Chlamydia trachomatis. J. Biol. Chem. 1997, 272 (16), 10594-10600. Haddad, N.; Burns, C. M.; Bolla, J. M.; Prévost, H.; Fédérighi, M.; Drider, D.; Cappelier, J. M., Long-term survival of Campylobacter jejuni at low temperatures is dependent on polynucleotide phosphorylase activity. Appl. Environ. Microbiol. 2009, 75 (23), 73107318. Reuter, M.; Mallett, A.; Pearson, B. M.; van Vliet, A. H. M., Biofilm formation by Campylobacter jejuni is increased under aerobic conditions. Appl. Environ. Microbiol. 2010, 76 (7), 2122-2128. Reinhardt, A.; Yang, Y.; Claus, H.; Pereira, C. L.; Cox, A. D.; Vogel, U.; Anish, C.; Seeberger, P. H., Antigenic potential of a highly conserved Neisseria meningitidis lipopolysaccharide inner core structure defined by chemical synthesis. Chem. Bio. 2015, 22 (1), 38-49.

26 ACS Paragon Plus Environment

Page 27 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

46.

47.

48. 49.

50.

51.

52.

53. 54. 55. 56.

57.

58.

59.

60.

61.

Kulshin, V. A.; Zahringer, U.; Lindner, B.; Frasch, C. E.; Tsai, C. M.; Dmitriev, B. A.; Rietschel, E. T., Structural characterization of the lipid-A component of pathogenic Neisseria meningitidis. J. Bacteriol. 1992, 174 (6), 1793-1800. Janusch, H.; Brecker, L.; Lindner, B.; Alexander, C.; Gronow, S.; Heine, H.; Ulmer, A. J.; Rietschel, E. T.; Zahringer, U., Structural and biological characterization of highly purified hepta-acyl lipid A present in the lipopolysaccharide of the Salmonella enterica sv. Minnesota re deep rough mutant strain r595. J. Endotoxin Res. 2002, 8 (5), 343-356. Van Der Spoel, D.; Lindahl, E.; Hess, B.; Groenhof, G.; Mark, A. E.; Berendsen, H. J. C., Gromacs: Fast, flexible, and free. J. Comput. Chem. 2005, 26 (16), 1701-1718. Hess, B.; Kutzner, C.; van der Spoel, D.; Lindahl, E., Gromacs 4:  Algorithms for highly efficient, load-balanced, and scalable molecular simulation. J. Chem. Theory Comput. 2008, 4 (3), 435-447. Abraham, M. J.; Murtola, T.; Schulz, R.; Páll, S.; Smith, J. C.; Hess, B.; Lindahl, E., Gromacs: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 2015, 1–2, 19-25. Wassenaar, T. A.; Ingólfsson, H. I.; Böckmann, R. A.; Tieleman, D. P.; Marrink, S. J., Computational lipidomics with insane: A versatile tool for generating custom membranes for molecular simulations. J. Chem. Theory Comput. 2015, 11 (5), 2144-2155. Wydro, P., The influence of cardiolipin on phosphatidylglycerol/phosphatidylethanolamine monolayers--studies on ternary films imitating bacterial membranes. Colloids Surf., B. 2013, 106, 217-223. Yesylevskyy, S. O.; Schäfer, L. V.; Sengupta, D.; Marrink, S. J., Polarizable water model for the coarse-grained MARTINI force field. PLoS Comput. Biol. 2010, 6 (6), e1000810. López, C. A.; Sovova, Z.; van Eerden, F. J.; de Vries, A. H.; Marrink, S. J., MARTINI force field parameters for glycolipids. J. Chem. Theory Comput. 2013, 9 (3), 1694-1708. Van Oosten, B.; Harroun, T. A., A MARTINI extension for Pseudomonas aeruginosa pao1 lipopolysaccharide. J. Mol. Graphics Modell. 2016, 63, 125-133. Garidel, P.; Rappolt, M.; Schromm, A. B.; Howe, J.; Lohner, K.; Andra, J.; Koch, M. H.; Brandenburg, K., Divalent cations affect chain mobility and aggregate structure of lipopolysaccharide from Salmonella minnesota reflected in a decrease of its biological activity. Biochim. Biophys. Acta 2005, 1715 (2), 122-131. Li, J. G.; Lakshminarayanan, R.; Bai, Y.; Liu, S. P.; Zhou, L.; Pervushin, K.; Verma, C.; Beuerman, R. W., Molecular dynamics simulations of a new branched antimicrobial peptide: A comparison of force fields. J. Chem. Phys. 2012, 137 (21). Mihajlovic, M.; Lazaridis, T., Charge distribution and imperfect amphipathicity affect pore formation by antimicrobial peptides. Biochim. Biophys. Acta, Biomembr. 2012, 1818 (5), 1274-1283. Parton, D. L.; Akhmatskaya, E. V.; Sansom, M. S. P., Multiscale simulations of the antimicrobial peptide maculatin 1.1: Water permeation through disordered aggregates. J. Phys. Chem. B 2012, 116 (29), 8485-8493. Wang, Y.; Schlamadinger, D. E.; Kim, J. E.; McCammon, J. A., Comparative molecular dynamics simulations of the antimicrobial peptide CM15 in model lipid bilayers. Biochim. Biophys. Acta, Biomembr. 2012, 1818 (5), 1402-1409. Yuan, T. H.; Zhang, X.; Hu, Z. H.; Wang, F.; Lei, M., Molecular dynamics studies of the antimicrobial peptides Piscidin 1 and its mutants with a DOPC lipid bilayer. Biopolymers 2012, 97 (12), 998-1009. 27 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

62.

63.

64.

65. 66.

67.

68.

69.

Page 28 of 47

Zhao, J.; Zhao, C.; Liang, G. Z.; Zhang, M. Z.; Zheng, J., Engineering antimicrobial peptides with improved antimicrobial and hemolytic activities. J. Chem. Inf. Model. 2013, 53 (12), 3280-3296. Khatami, M. H.; Bromberek, M.; Saika-Voivod, I.; Booth, V., Molecular dynamics simulations of histidine-containing Cod antimicrobial peptide paralogs in self-assembled bilayers. Biochim. Biophys. Acta, Biomembr. 2014, 1838 (11), 2778-2787. Bennett, W. F. D.; Hong, C. K.; Wang, Y.; Tieleman, D. P., Antimicrobial peptide simulations and the influence of force field on the free energy for pore formation in lipid bilayers. J. Chem. Theory Comput. 2016, 12 (9), 4524-4533. Pino-Angeles, A.; Leveritt, J. M.; Lazaridis, T., Pore structure and synergy in antimicrobial peptides of the magainin family. PLoS Comput. Biol. 2016, 12 (1) 1-17. Rebeil, R.; Ernst, R. K.; Jarrett, C. O.; Adams, K. N.; Miller, S. I.; Hinnebusch, B. J., Characterization of late acyltransferase genes of Yersinia pestis and their role in temperature-dependent lipid a variation. J. Bacteriol. 2006, 188 (4), 1381-1388. Petrache, H. I.; Dodd, S. W.; Brown, M. F., Area per lipid and acyl length distributions in fluid phosphatidylcholines determined by (2)H NMR spectroscopy. Biophys. J. 2000, 79 (6), 3172-3192. Kong, Q.; Six, D. A.; Liu, Q.; Gu, L.; Wang, S.; Alamuri, P.; Raetz, C. R.; Curtiss, R., 3rd, Phosphate groups of lipid a are essential for Salmonella enterica serovar typhimurium virulence and affect innate and adaptive immunity. Infect. Immun. 2012, 80 (9), 3215-3224. Kucerka, N.; Papp-Szabo, E.; Nieh, M. P.; Harroun, T. A.; Schooling, S. R.; Pencer, J.; Nicholson, E. A.; Beveridge, T. J.; Katsaras, J., Effect of cations on the structure of bilayers formed by lipopolysaccharides isolated from Pseudomonas aeruginosa Pao1. J. Phys. Chem. B 2008, 112 (27), 8057-8062.

28 ACS Paragon Plus Environment

Page 29 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Table 1. Summary of chemical structure of lipid A in various species of gram-negative bacteria. Labels A, A', B, B', C, C', D, and D' correspond to acyl chains shown in Figure 1. Label P denotes the total number of phosphates. Species H. pylori P. gingivalis B. fragilis B. pertussis C. trachomatis C. jejuni N. meningitidis S. minnesota

A 18 17 16 14 20 14 14 14

A' − − − − − − 12 16

Saturated acyl chain lengths B B' C C' D 16 − 18 18 − 16 − 17 16 15 15 − 17 15 16 10 − 14 14 14 14-15 − 20 18-21 14-16 14 − 14 16 14 12 − 14 12 12 14 − 14 12 14

D' − − − − − 16 − 14

Total chains 4 5 5 5 5 6 6 7

P 1 1 1 2 2 2 4 2

29 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 47

Table 2. System details of the membrane simulations involving DPPE in the inner leaftet.a Set

Species

b

H. pylori c P. gingivalis B. fragilis B. pertussis C. trachomatis d C. jejuni N. meningitidis S. minnesota H. pylori P. gingivalis B. fragilis B. pertussis C. trachomatis C. jejuni N. meningitidis S. minnesota H. pylori P. gingivalis B. fragilis B. pertussis C. trachomatis C. jejuni N. meningitidis S. minnesota H. pylori P. gingivalis B. fragilis B. pertussis C. trachomatis C. jejuni N. meningitidis S. minnesota H. pylori P. gingivalis B. fragilis B. pertussis C. trachomatis C. jejuni N. meningitidis S. minnesota

Set I

Set II

Set III

Set IV

Set V

inner DPPE 224 188 187 176 188 224 178 260 224 188 187 176 188 224 178 260 224 188 187 176 188 224 178 260 152 188 264 188 188 224 177 260 152 188 264 188 188 224 177 177

outer Lipid A DPPE 72 8 72 8 72 7 64 16 72 8 72 8 57 6 72 8 72 8 72 8 72 7 64 16 72 8 72 8 57 6 72 8 72 8 72 8 72 7 64 16 72 8 72 8 57 6 72 8 72 8 72 8 96 24 72 8 72 8 72 8 57 6 72 8 72 8 72 8 96 24 72 8 72 8 72 8 57 6 57 6

Water

Ion

W W W W W W W W W W W W W W W W W W W W W W W W PW PW PW PW PW PW PW PW PW PW PW PW PW PW PW PW

Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+

Number of Water ions 4420 72 4419 72 4535 72 4484 128 4341 144 4341 144 4546 114 4344 144 4420 72 4419 72 4535 72 4484 128 4341 144 4341 144 4546 114 4344 144 4420 36 4419 36 4535 36 4484 64 4341 72 4341 72 4546 57 4344 72 4429 72 4407 72 8937 113 4450 144 4339 144 4340 144 4544 114 4343 144 4429 36 4407 36 8937 56 4450 72 4339 72 4340 72 4544 57 4544 57

T (K) 310 310 310 310 310 310 310 310 315→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 275→360 310 310 310 310 310 310 310 310 310 310 310 310 310 310 310 310

a

Simulations were performed in duplicates Set I was simulated for 10 µs c Simulation repeated after addition of 4' phosphate and adjusting the counterions d Simulation repeated after deletion of 4' phosphate and adjusting the counterions b

30 ACS Paragon Plus Environment

Page 31 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Table 3. System details of the membrane simulations involving a complex composition of the inner leaftet.a Set b

Set VI

Set VII

a b

inner DPPE POPG H. pylori 118 33 P. gingivalis 118 33 B. fragilis 117 33 B. pertussis 118 33 C. trachomatis 118 33 C. jejuni 137 39 N. meningitidis 137 39 S. minnesota 137 39 H. pylori 118 33 P. gingivalis 118 33 B. fragilis 117 33 B. pertussis 118 33 C. trachomatis 118 33 C. jejuni 137 39 N. meningitidis 137 39 S. minnesota 137 39 Species

outer CDL2 Lipid A DPPE 16 87 9 16 70 8 16 68 12 70 8 16 16 70 8 19 69 7 19 68 10 19 60 4 87 9 16 16 70 8 16 68 12 70 8 16 16 70 8 19 69 7 19 68 10 19 60 4

Water Ion PW PW PW PW PW PW PW PW PW PW PW PW PW PW PW PW

Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+ Ca2+

Number of Water ions 4318 76 4378 67 4704 68 4627 102 4283 103 4261 110 4619 106 4276 110 4318 76 4378 67 4704 68 4627 102 4283 103 4261 110 4619 106 4276 110

T (K) 310 310 310 310 310 310 310 310 275→330 275→330 275→330 250→330 250→330 275→330 250→330 275→330

Simulations were performed in duplicates Set VI was simulated for 10 µs

31 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 32 of 47

Table 4. Key properties of the membrane and comparison phase transition temperature. The Tm values of Set II and Set VII are compared to the available experimental data. AL (nm2) Tm−10 K Tm+10 K H. pylori 1.02 1.27 P. gingivalis 1.27 1.54 B. fragilis 1.25 1.55 B. pertussis 1.33 1.52 C. trachomatis 1.26 1.51 C. jejuni 1.46 1.83 N. meningitidis 1.49 1.80 S. minnesota 1.71 2.08 a At T=275 K b Reference 12 c Reference 53 Bacteria

DM (nm) Tm−10 K Tm+10 K 4.63 4.14 4.64 4.11 4.59 4.09 3.96 3.74 4.55 4.05 4.54 4.01 4.55 4.12 4.31 3.93

DH (nm) Tm−10 K 2.8 2.9 3.0 2.6 3.3 2.7 2.5 2.8

Set II 341 ± 4 338 ± 4 349 ± 6 309 ± 2 330 ± 3 336 ± 4 324 ± 3 337 ± 1

Tm (K) Set VII Expt. 299 ± 1 − 305 ± 2 − 309 ± 1 − 279 ± 2 − 289 ± 2 − 296 ± 6 325b 284 ± 2 318c 299 ± 8 310, 321c

D (cm2s−1) Set Ia 1.9 × 10−7 3.0 × 10−8 1.3 × 10−7 1.6 × 10−7 6.5 × 10−8 4.4 × 10−8 7.3 × 10−8 1.4 × 10−8

32 ACS Paragon Plus Environment

Page 33 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Figure 1. Lipid A template structure with primary (A, B, C, and D) and secondary acyl chains (A', B', C', and D').

33 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 34 of 47

Figure 2. Coarse grain mapping scheme for lipid A tails of (A) N. meningitidis, (B) H. pylori, (C) P. gingivalis, (D) B. fragilis, (E) B. pertussis, (F) C. trachomatis, (G) C. jejuni, and (H) S. minnesota. The bead type is shown in bold.

34 ACS Paragon Plus Environment

Page 35 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Figure 3. Average disaccharide head group bond distance frequency distribution for (A) H. pylori (B) P. gingivalis, (C) B. fragilis, (D) B. pertussis, (E) C. trachomatis, (F) C. jejuni, (G) N. meningitidis, and (H) S. minnesota.

35 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 36 of 47

Figure 4. Average disaccharide head group angle frequency distribution for (A) H. pylori (B) P. gingivalis, (C) B. fragilis, (D) B. pertussis, (E) C. trachomatis, (F) C. jejuni, (G) N. meningitidis, and (H) S. minnesota.

36 ACS Paragon Plus Environment

Page 37 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Figure 5. Average acyl chain bond distance frequency distribution for (A) H. pylori (B) P. gingivalis, (C) B. fragilis, (D) B. pertussis, (E) C. trachomatis, (F) C. jejuni, (G) N. meningitidis, and (H) S. minnesota.

37 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 38 of 47

Figure 6. Average acyl angle frequency distribution for (A) H. pylori (B) P. gingivalis, (C) B. fragilis, (D) B. pertussis, (E) C. trachomatis, (F) C. jejuni, (G) N. meningitidis, and (H) S. minnesota.

38 ACS Paragon Plus Environment

Page 39 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Figure 7. Area per lipid (AL) of Lipid A as a function of temperature for (A) Set II and (B) Set VII bacterial outer membrane. Color scheme: H. pylori (brown), P. gingivalis (red), B. fragilis (purple), B. pertussis (gray), C. trachomatis (black), C. jejuni (orange), N. meningitidis (yellow), and S. minnesota (green).

39 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 40 of 47

Figure 8. Phase transition temperature (Tm) for Set II membranes determined by the change in the AL (nm2) versus T (K). The Tm values (K) are labeled for each curve. Color scheme for the lines and labels: H. pylori (brown), P. gingivalis (red), B. fragilis (purple), B. pertussis (gray), C. trachomatis (black), C. jejuni (orange), N. meningitidis (yellow), and S. minnesota (green).

40 ACS Paragon Plus Environment

Page 41 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Figure 9. Snapshots of thermal phase transition of (A) H. pylori, (B) P. gingivalis, (C) B. fragilis, (D) B. pertussis, (E) C. trachomatis, (F) C. jejuni, (G) N. meningitidis, and (H) S. minnesota in Set II. The panels show ordered phase (283 K, left) and disordered phase (350 K, right) membrane structure. Color scheme: Lipid A head groups (orange); Lipid A acyl chains (cyan); DPPE head group (blue); DPPE carbon tails (magenta).

A

B

C

D

E

F

G

H

41 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 42 of 47

Figure 10. Membrane thickness (nm) as a function of T (K) for (A) H. pylori, (B) P. gingivalis, (C) B. fragilis, (D) B. pertussis, (E) C. trachomatis, (F) C. jejuni, (G) N. meningitidis, and (H) S. minnesota in Set II.

42 ACS Paragon Plus Environment

Page 43 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Figure 11. RDF for Set II (Na+, dashed) and Set VI (Ca2+, solid) for phosphate (black) and carboxylate (red) for (A) H. pylori, (B) P. gingivalis, (C) B. fragilis, (D) B. pertussis, (E) C. trachomatis, (F) C. jejuni, (G) N. meningitidis, and (H) S. Minnesota.

43 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 44 of 47

Figure 12. Density profile of key components of (A) B. pertussis and (B) C. trachomatis membranes (Set VI). Color scheme: Water (blue); Lipid A phosphates (purple); DPPE head groups (red); inner leaflet carbon tails (black, dotted); and outer leaflet carbon tails (orange, dotted). The Ca2+counterion density (green) is shown on the secondary y-axis.

44 ACS Paragon Plus Environment

Page 45 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Figure 13. Shift in the Tm values due to lipid A modification in (A) C. jejuni and (B) P. gingivalis. The peaks in the ∆AL versus T plot denote the Tm values. In panel A, native C. jejuni (blue) shows increase in Tm after deletion of 4' phosphate (red); and in panel B, native P. gingivalis (red) shows decrease in Tm after addition of 4' phosphate (blue).

45 ACS Paragon Plus Environment

Journal of Chemical Theory and Computation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Page 46 of 47

Figure 14. Radial distribution functions for lipid A head group phosphate-phosphate ( gP-P (r) ; solid line) and phosphate-Ca2+ counterions ( gP-Ca2+ (r ) ; dashed line) for (A) H. pylori, (B) C. jejuni, and (C) N. meningitidis. The inset images show the top-view of the lipid A head group phosphates (green) and Ca2+ (orange) counterions from the surrounding medium.

46 ACS Paragon Plus Environment

Page 47 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

Journal of Chemical Theory and Computation

Table of content graphic

47 ACS Paragon Plus Environment