Modified November 1998
Department of Chemistry and Biochemistry Home Page
University of Guelph Home Page
View of the trimeric membrane protein maltoporin looking down on the membrane plane. The view has been restricted to amino acids 1-360 so that the pore is clearly visible. Beta strands are coloured yellow, turns and loops blue and white, helix is shown in red.
|
Porins are members of an extended family of transmembrane proteins found in the outer membrane of gram negative bacteria such as E. coli, or in the outer membranes of mitochondria in eukaryotes. The "general" porins provide a simple diffusion pathway across the membrane for molecules less than 1 kDa with little substrate selectivity. This allows most soluble nutrients to enter while excluding dangerous toxins or proteases that might by released by adversary organisms. A sub-class of substrate-specific porins also exists, for example maltoporin, which shows selectivity for linear oligosaccharides. An unrelated class of proteins called aquaporins is found in eukaryotes, and regulates passage of H2O across the membrane. Aquaporins are fairly conventional alpha helix-based transmembrane proteins. |
![]() |
|
The true porins share a common structure, based on a polypeptide of 300-400 amino acids forming a 16 to18 stranded antiparallel beta barrel. The barrel traverses the membrane with its axis normal to the plane of the bilayer, and is arranged as a trimer of three identical subunits. Although the porin structure is shared by a wide variety of outer membrane channels, there is surprisingly little sequence homology. A similar structural organization of a 14-stranded antiparallel beta barrel is found in Staph. aureus alpha haemolysin, although functionally this is somewhat different. References S.W Cowan et al., Nature, 358: 727-733 (1992) M.S. Weiss and G.E Schulz, J. Mol. Biol. 227: 493-509 (1992). |
|
|
The porin beta barrel has a non polar exterior embedded in the bilayer and a polar interior to the channel, an organization which is inside out relative to normal soluble globular proteins. Overall hydropathy is more polar than a typical soluble protein, and because the beta strand organization alternates polar and non polar amino acids in the transmembrane region, porins fail the usual tests which purport to detect transmembrane proteins. However the alternating polar/nonpolar arrangement is readily detected if one is aware of this possibility. |
|
The beta barrel is roughly cylindrical, diameter 40 angstroms, with a wall varying from 30 to 50 angstroms high around the perimeter. In the trimer, the low point of the wall faces the centre, where the subunits join; the high point of the wall faces out. The lower edge of the barrel is fairly uniform and smooth, and adjacent strands are linked by fairly uniform tight turns; this faces the periplasm. The extracellular surface of porin is rough, due to the beta strands being connected by loops of irregular length and form. The extracellular surface is also very polar and rich in charged amino acids. Many bacterial cell surface antigens derive from the rough end of porins. |
|
The organization of the cylindical wall is a continuous antiparallel beta sheet. In some porins, e.g. E. coli OmpF, the N-and C-terminals link via a salt bridge to form a pseudo-cyclic peptide. |
![]() |
Each beta strand is inclined about 45o to the barrel axis. The low point of the wall occurs partly because strands 1-6 + 16 are shorter (7-10 amino acid) and partly because they are at a shallower angle. The high spot on the wall is formed from strands 8-14 which are 12-14 amino acids long and slightly steeper. When amino acids are selectively colored, red = non-polar, white = aromatic, green = polar, a belt of nonpolar amino acids about 25 angstoms high is seen to surround the barrel. This corresponds to the fatty acid zone of the bilayer. The non-polar belt is bordered by aromatic amino acids which associate with the lipid head group region. A narrow polar region on the periplasmic side, and broader region on the extracellular end protrude from the bilayer. |
|
| ![]() |
Switching to a ribbon view, where the inward facing amino acids are not masked, one sees the alternating polar non-polar arrangement of aminoacids in the beta strand in the bilayer belt. | ![]() |
|
The view into the barrel core reveals a largely polar interior. The interior is partly blocked by the extended irregular loop structure of loop 3, which restricts passage to a small eyelet of 10 angstroms instead of the full 20-25 angstroms of the barrel bore. The loop is held in position partly by hydrophobic interactions, and contacts between loop 3 and the barrel have inward facing non polar amino acids, so the alternation of polar/non polar is disrupted at these points. |
![]() |
| In the maltoporins, the bore is even more obstructed, and amino acids lining the eyelet bind with high affinity to linear oligosaccharide. Presence of bound sugar blocks passage of other molecules, but another entering sugar molecule can displace the current occupant inwards, since the new entrant replaces the binding energy for the current occupant with interactions of equal strength. |
|
Porin has a normal signal sequence, and is translocated across the bacterial inner membrane as a secreted protein. The mechanism of translocation in bacteria involves sequestration of a signal containing unfolded nascent polypeptide by the chaperone protein Sec B. This complex then binds Sec A ATPase, which delivers the polypeptide to the SecE-SecY translocation channel in the inner membrane. |
|
|
On entry to the periplasm, the polypeptide begins folding. It is believed that folding commences by trimerization as a soluble protein, since the region of subunit contact has a conventional hydrophobic core. The partly assembled barrel at this stage exposes its polar concave face to the periplasmic solvent. |
|
A lipid binding region of this hydrophobic core could pull the semi-folded trimer into the membrane, where beta sheet formation would be driven by the alternating polar-nonpolar side chains. |
|
![]() |
If the method requires building test datasets e.g. as used by neural network based systems, web-based software automatically keeps its test datasets up to date, whereas downloaded software runs the risk of relying on outdated local test datasets.
Tools fall into the following classes:
Sequence searches:
These days this is usually some variant of BLAST (Basic Local Alignment Search Tool)
The National Center for Biotechnology Information (NCBI) hosts BLAST servers as well as GenBank, the major sequence database.
http://www.ncbi.nlm.nih.gov/BLAST
BLAST and its variants match sequences based on localized similarity,
| ALSQ |||| ALSQ | identity | SITN .... ALSQ | similiarity | |||
| AKMQ | | ALSQ | partial identity | A--Q | | ALSQ | gapped identity |
One submits a sample sequence (or accession number or identifier of an existing GenBank entry), and BLAST will retrieve all similar sequences from a sequence or structure database that you specify. A parameter E sets the limit for degree of mismatch (defaults to 1, but values greater than 0.1 can retrieve very tenuous matches.) Results return to your browser after a brief wait.
The test sequence is submitted in FASTA format,
>comment line e.g. to identify protein
QAQITGRPEWIWLALGTALMGLGTLYFLVKGMGVSDPDAKKFYAITTLVP
AIAFTMYLSMLLGYGLTMVPFGGEQNPIYWARYADWLFTTPLLLLDLALL
etc
The sequence can be prepared in a text editor and pasted into the query window.
Simple structure predictions:
Hydropathy plots:
Weizmann Institute has a simple server for hydropathy plots
http://bioinformatics.weizmann.ac.il/hydroph/plot_hydroph.html
Again a plain text sequence is pasted into the query window. One can choose the size of the "window", and choose between Kyte-Doolittle scales or Hoppe-Woods scales for hydrophobicity. The Kyte-Doolittle method should find tentative transmembrane seqments, whereas Hoppe-Woods searches for surface exposed loops on a globular protein that are candidates for antigenic determinants. Results return directly to your web browser.
However, the Cadillac webservere for Protein Structure Prediction must be PHD, located at EMBL in Heidelberg:
http://www.embl-heidelberg.de/predictprotein/predictprotein.html
This server includes a variety of tools for sequence alignment, motif searches, transmembrane segment searches, and secondary structure prediction, as well as some tertiary structure matching methods.
The structure analysis methods are based on advanced neural network analysis, and use pattern matching to sequences of homologous proteins to produce highly reliable predictions. The output includes extensive commentary on the results as well as the graphical presntation. Because this output is more extensive than some of the simpler servers, results do not return directly to your browser, but may be e-mailed or picked from the PHD ftp server. Typically, a wait of a few minutes may be expected before results are returned.