Book of Abstracts: Albany 2003

category image Albany 2003
Conversation 13
Abstract Book
June 17-21 2003

Structural Analysis of β-barrel Membrane Proteins

Structures of integral membrane proteins have proven difficult to obtain. Structural analysis with existing known structures of these classes of proteins can help in modeling the structure of a new protein from the protein sequence. Classification and associated grouped analysis allow comparisons within a class to infer shared properties of other members (1). Moreover, comparisons between classes can yield information of specific-function relationships. In order to understand the β-barrel membrane protein structure for their oligomerization state, pore dimension and loop conformation, a combined program to analyze all the structural details of β-barrel membrane proteins has been developed in visual basic (Structural Comparison of Outer Membrane β-barrel Proteins SCOMPS). Classified as sixteen and eighteen stranded b-barrel membrane proteins were analyzed for their structural details.

Since membrane permiability barriers are among the factors contributing to the intersinic resistance of bacteria to antibiotics. OmpF porin from E. coli is the the pricipal path way for the β-lactam antibiotics. It has been showed that ampicilin exhibits a strong interaction with the residues of the constriction zone of OmpF channel (2). The structure based sequence alignment shows that except Y 32 the constricting zone residues of OmpF porin involved in ampicilin binding (R 42, R 82, R 132, E 117, Y 32 and Y 22) are structurally aligned with the PhoE porin and OmpK36. This suggest that Y32 may not be essential for ampicilin the binding. The alignment also suggest that the R. capsulatus, R. balastics may not posses this ampicilin binding property as none of the residues are conserved as in the case of E. coli porins. These predictions need to be experimentally verified.

The core regions of these porins have well conserved and structurally positioned positive and negatively charged residues on the opposite walls. R. capsulatus porin and R. blastica porin in comparison with OmpF have no structurally conserved charged residues. In all the sixteen-stranded porins except R. blastica, the positively charged residues are aligned on one side of the wall (between 0°-180°) and the negatively charged residues are on the opposite side of the wall (between 180°-360°) along the pore axis (Figure 1). All these charged residues along the wall are structurally conserved. This arrangement can cause a strong transversal electric field across the pore, which could be of functional importance.

Figure 1: The charged residues that are lining the opposite walls of the pore for sixteen-stranded porins are shown. The molecules are identified by their PDB ID. Negatively charged residues are colored in red and the positively charged residues are colored in blue.

Molecular basis for the polyamine-OmpF porin interactions, an inhibitor and mutant studies (3) showed three residues (D 113, D 121 and Y 294) as putative sites of interaction between the channel and spermine. These three residues that are protruding into the pore are conserved among the E. coli porins and OmpK36 but not in the R. blastica and R. capsulatus porins, which differ in their pore architecture from the E. coli porins.

In the case of specific porins most of the channel-lining residues are conserved between LamB and ScrY porins. Maltoporin from E. coli and from Salmonella typhimurium have similar positional preference of amino acids and also a conserved set of positive and negative charges on opposite sides across the channel. The distribution of positively and negatively charged residues are almost equal on both sides of the wall in all the specific porins. The lining of the maltodextrin-specific maltoporin (LamB) channel exibits a string of aromatic residues forming the greasy slide shown by the crystallography study (4). Among the aromatic residues involved in substrate binding, W 74, Y 41, Y 6, W 420 are structurally conserved among the three specific porins. The other residues W 358, F 227 and Y 118 are structurally conserved only within maltoporins.

Modeling channels as cylinders with a biconical profile (5) yields more acceptable estimation of channel conductance. In microscopic models of ion conductance through channels, ion permeation can be expressed in terms of known physical quantities: ion properties (mass, radius and hydration number) and channel properties (pore length, radius, location and type of specific reacting residues). Pores with a biconical profile are found to have the highest conductance for a given diameter at their narrowest point. Similarly the functional models of channel proteins had identified a set of possible structural folds (6). Knowledge of the geometry of the pore, surface charges and interaction of permeant ions with the channel are essential for the theoretical channel conductance analysis through different models. In porin family of membrane proteins diffusion through the channel is determined by the structure of the eyelet (7), which is formed by the long loop (L3) between β-strands β5 and β6. This internal loop L3 constricts the pore to different sizes. The ion translocation is mainly governed by pore geometry and charges in OmpF porin (8). Based on the above said studies on conductance in ion channels and the correlation of the channel geometry with the biconical shape defined in the microscopic model, for the pore forming β-barrel membrane proteins, the pore radius at every plane is calculated and compared with other similar porins ( Figure 2). For the sixteen-stranded porins the pore dimension varies from 6.5Å to 9.6Å at the constriction zone. Constriction zone diameter of the eighteen-stranded porins varies between 4Å to 4.6Å. As deduced from the crystal structures of E. coli porins; the PhoE pore is more constricted than that of OmpF. Sixteen-stranded porins have a wider pore than the eighteen-stranded porins.

Figure 2: a. The inner and outer dimension (calculated using SCOMPS) at an interval of 1Å along the pore axis for sixteen-stranded porins. b. The inner and outer dimension of the pore along the pore axis for eighteen-stranded porins.

All the sixteen and eighteen stranded porins are trimers and have conserved sub unit and trimer interfaces. The dimer contacts are closely packed than the trimer interface. Surface accessibility at the interfaces shows the buried surface is more at the trimer than the ?dimer? interface in all the specific and general porins.

Porins also tend to act as receptors for molecules such as phage proteins, lactoferrin etc. The interaction is primarily with the loops as the beta barrel is embedded in the membrane. However, the loop conformation does not show any common pattern as seen for example in the case of the Complementarity Determining Regions (CDRs) of the Fab structures (9). The loops are made of mostly polar residues. The loop geometry analysis showed that the strands are anti parallel with the packing angle of the connecting loops to be almost 180° in all the cases. The hoist and meridian angles of all the loops of these porins are different. The hoist angle for loop3 (L3) for all porins is less than 90° indicating its inward fold into the barrel. The over all distribution of charged residues along the pore contribute to the formation of specific electric field, giving each porin its unique properties.

Vasanthi G1
Krishnaswamy S2

1Department of Physics
Lady Doak College
2Bioinformatics Centre
School of Biotechnology
Madurai Kamaraj University Madurai
625021, India

Reference and Footnotes
  1. Chemistry & Biology 3, 779-783 (1996)
  2. Proc. Natl. Acad. Sci. USA 99, 9789-9794 (2002)
  3. J. Mol. Biol. 297, 933-945(2000)
  4. J. Mol. Biol. 272, 56-63 (1997)
  5. Biophys. Chem. 52, 15-24 (1994)
  6. Curr. Opin. Struct. Biol. 6, 99-510 (1996)
  7. Mol. Microbiol. 37, 239-253 (2000)
  8. Biochemistry 40, 6319-6325 (2001)
  9. Nature 342, 887-883 (1989)