Data on diverse roles of helix perturbations in membrane proteins

The various structural variations observed in TM helices of membrane proteins have been deconstructed into 9 distinct types of helix perturbations. These perturbations are defined by the deviation of TM helices from the predominantly observed linear α-helical conformation, to form 310- and π-helices, as well as adopting curved and kinked geometries. The data presented here supplements the article ‘Helix perturbations in Membrane Proteins Assist in Inter-helical Interactions and Optimal Helix Positioning in the Bilayer’ (A. Shelar, M. Bansal, 2016) [1]. This data provides strong evidence for the role of various helix perturbations in influencing backbone torsion angles of helices, mediating inter-helical interactions, oligomer formation and accommodation of hydrophobic residues within the bilayer. The methodology used for creation of various datasets of membrane protein families (Sodium/Calcium exchanger and Heme Copper Oxidase) has also been mentioned.


Tables and figures How data was acquired
Data was retrieved from public databases

Data format Analyzed data Experimental factors
Protein structures were retrieved from OPM database and analyzed. Sequence and structural alignments of proteins were performed using Clustal Ω and MAPSCI respectively Experimental features This work uses X-ray crystal structure data of membrane proteins that has been deposited in the Protein Data Bank (PDB) Data source location

Bangalore, India
Data accessibility Data is within this article. Membrane protein structures aligned along the Z-axis can be readily retrieved from the OPM database (http://opm.phar.umich.edu/ download.php).

Value of the data
The data on different types of helices shows that, apart from the commonly observed α-helices, 3 10 and π-helices are also present within the bilayer and have varying lengths as well as distinct sequence signatures. This data provides experimentalists with options to model new 3 10 -and πhelices in the bilayer and reorient the locations of active sites in TM helices.
The data on backbone torsion angle variation in perturbed helices indicates that in these regions the disrupted hydrogen bonds lead to free NH-and C¼ O groups that mediate inter-helical interactions. This information can be used by the scientific community to engineer the desired inter-helical interactions at appropriate locations in TM helices.
The data showing conservation of a kink in proteins from the Sodium/Calcium exchanger family highlight its crucial functional role in this family. This data can be used for homology modeling of proteins within this family by computational biologists.

Data
The data used in this analysis has been generated after a detailed structural examination of membrane proteins. This structural data provides solid evidence for the utility and various roles of perturbed helices in membrane proteins. See Figs. 1-17 and Tables 1-5.

Experimental design, materials and methods
Structural analysis of membrane protein structures was performed after they were downloaded from the Orientation of Proteins in Membrane (OPM) database [9]. The identification of secondary structures was carried out using Assignment of Secondary Structures in Proteins (ASSP) [10] and nonbonded interactions were identified using MolBridge [11]. Next, we identified geometries of helical fragments using Helanal-Plus [2] and computed the backbone torsion angles (φ-ψ). Multiple sequence alignment of protein sequences was carried out using ClustalΩ [12].
We prepared datasets of proteins belonging to Sodium Calcium family of transporters as mentioned in [1] to examine conservation of kinks in functionally important helices. A dataset of proteins belonging to Heme Copper Oxidase (HCO) superfamily was created to gain insights about the presence of the π-helix in each protein (Table 3). To understand the variation if any in the π-helix within different types of HCOs, we analyzed two crystal structures from the A-type, one from B-type and         Table 3). The presence of the unusually long π-helix in Cytochrome-c-oxidase (PDB ID: 1v55) defined by ASSP was reconfirmed by its identification using DSSPa program based on hydrogen bond energetics for secondary structure identification (http://www.cmbi.ru.nl/dssp.html)   [3,4] and several studies have elucidated their importance in other bio molecules as well [5][6][7][8].   Table 2). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article) .    15. Ramachandran Map for π-helical (64-82) region in mitochondrial COX and multiple sequence alignment of the heme copper oxidase (HCO) superfamily members. a) Pro84 is not a part of the π-helix but the φ-ψ for it has been represented to show that it has similar torsion angles outside the helix perturbation as well. b) Multiple sequence alignment for the helical region analogous to TM2 of the reference protein containing the 19 residue π-helix for all HCO superfamily members (see Table 3).   17. Long π-helices allow accommodation of more amino acids in the membrane. Helical regions have been represented as ribbons with C α atoms highlighted as spheres. The αand π-helical regions of the reference protein (Mitochondrial COX-1v55) have been represented in blue and red ribbons respectively. The corresponding α-helical regions of 3mk7, 3o0r and 3ayf have been shown in green, orange and grey colours. a) The bacterial COX has a small interspersed π-helix that accommodates a Phenylalanine within the helical region as observed in the reference protein. b and c) The long π-helix accommodates two extra residues (Phe67 and Gly76) in the helical region as compared to α-helices observed in NORs. The entry and exit points of the helix in the membrane have been represented as a '' and '*' respectively. Kinked-Pro-P2 Kinked-Non-Pro  Table 3 Proteins from the Heme-Copper Oxidase (HCO) superfamily considered for the analysis of the π-helical region. A total of 8 proteins (at least one member of a particular HCO subtype) have been selected for analysis. The 'Mitochondrial COX (1v55:A)' belongs to the initial dataset of 90 proteins used for analysis and contains the interspersed 19 residue long π-helix. The 'Helical region' (fifth column) represents the entire TM segment considered for analysis. The 'Helix assignment' (sixth column) includes the helix boundaries for α and π-helices defined by ASSP (see methods).   Table 4 Tabulated output files of ASSP and DSSP defining the long π-helical region in mitochondrial COX. ASSP defines a π-helix from (64 V-82 L) based on twist, rise per residue and helical radius whereas DSSP defines a π-helix from (64 V-79 G) denoted by the symbol 'I' based on backbone hydrogen bond energetics.

Protein
( Table 4). Pair-wise crossing angles between TM helices were determined by calculating the cross products of direction cosines (l, m, n) as computed by Helanal-Plus.

Acknowledgements
AS acknowledges IISc for a fellowship. AS acknowledges Anasuya Dighe for suggestions and critical reading of the manuscript. MB is a recipient of J.C. Bose National Fellowship of Department of Science and Technology, Government of India. MB acknowledges DBT-IISc partnership program.