Практикум 6. Transmembrane proteins

Task 1

Omp33 porin

Type:

Transmembrane (3 classes)

Class: Beta-barrel transmembrane (36 superfamilies)

Superfamily: OmpG-like porins (n=14,S=16) (3 families) PF09381

Family: Omp33 porin (1 proteins) 1.B.1 (TCDB) PF16956 IPR031593 PDBsum

Species: Acinetobacter baumannii (37 proteins)

Localization: Bacterial Gram-negative outer membrane (421 proteins)

pdb: 6gie

Uniprot: A0A081GU02_ACIBA

Function: Omp33's main function is to serve as a channel for the passive diffusion of small molecules, such as ions and nutrients, across the bacterial outer membrane. Porins like Omp33 play a critical role in the uptake of essential nutrients and in the defense mechanisms of the bacterial cell.

Img.1. Image of the Omp33 porin structure location in the cell membrane, this picture was taken form OPM site.

Coordinates of the transmembrane regions of the protein from OPM:

1( 3- 10)

2( 22- 30)

3( 51- 58)

4( 74- 83)

5( 90- 97)

6( 115- 121)

7( 127- 134)

8( 169- 177)

9( 185- 194)

10( 200- 206)

11( 214- 219)

12( 232- 239)

13( 246- 251)

14( 265- 272)

The protein sequence was downloaded from the PDB record in fasta format and uploaded to DeepTMHMM.

Img.2. The DeepTMHMM output

On the horizontal axes, the positions of amino acids of the protein are plotted. On the vertical axes - the probability of the predicted existence of the amino acid in a certain position. Different colors denote different positions: Red - in the membrane, green - in the periplasm,blue - outside the cell, orange - signal sequence.

The output file in .gff3

results:

1(21-27)

2(45-51)

3(74-80)

4(96-103)

5(113-119)

6(136-143)

7(150-156)

8(191-197)

9(206-213)

10(221-228)

11(234-242)

12(253-260)

13(267-274)

14(286-294)

Chain Number of predicted TMRs: 14,which corresponds to the experimentally obtained model, but the boundaries of the transmembrane fragments in the experimental model and in the predicted one do not coincide at all. This could be arised from protein conformational flexibility or post-translational modifications, or even the presents of other molecules could influence localization of regions.

Task 2.Comparison of predictions of transmembrane regions in the alpha-helical protein

1. My protein which currently has no close homolog with an experimentally obtained three-dimensional structure (terra incognita) is Leucine efflux protein (LEUE_ECO57)

Name: Leucine efflux protein/Белок выхода лейцина

Swiss-prot AC: Q8XDS6

Organism: Escherichia coli O157:H7

Function:Exporter of leucine

2.Run DeepTMHMM with Leucine efflux protein:

Img.3.Arrangement of helices by the DeepTMHMM algorithm for Leucine efflux protein

The upper graph shows the probable structure, while the lower one displays the probability distribution for each position to be in a specific region. The X-axis represents the protein sequence. The colors indicate the following: red-represents transmembrane segments, blue- represents extracellular regions, and pink- represents intracellular regions.

It was found that there are 6 transmembrane alpha helices

The output file in .gff3

Coordinates of the transmembrane regions of the protein from OPM:

1(11-31)

2(45-69)

3(74-92)

4(130-139)

5(153-175)

12(195-208)

3.Number of Membranes:1

Type of membrane:gram-negative bacteria outter membrane

Allow curvature: yes

Topology (N-ter): out or in does not make differences according to DeepTMHMM prediction

Coordinates of the transmembrane regions of the protein from PPM:

1( 12- 22)

2( 43- 68)

3( 74- 99)

4( 120- 141)

5( 153- 175)

The number of findings does not coincide with DeepTMHMM. However, the transmembrane regions intersect according to the results of two predictions, despite the shifts. If we look at the AlphaFold prediction, it can be seen that the boundaries of the transmembrane domains have lower prediction accuracy. This may be explained by the fact that DeepTMHMM and AlphaFold use different approaches and training datasets for predicting protein structures, which should influence how the algorithms determine the start and end of transmembrane domains. These differences could also be explained by protein conformational flexibility, as each model may represent a specific conformation. These differences in prediction accuracy can explain the variations in the length of the transmembrane domains predicted by the two methods.