Use of Bioinformatics to Enhance Biophysical Exploration of Ion ...

dasypygalstockingsΒιοτεχνολογία

2 Οκτ 2013 (πριν από 4 χρόνια και 1 μήνα)

90 εμφανίσεις

Use of Bioinformatics to
Enhance Biophysical
Exploration of Ion Channels

Main points of talk
:


Goals of project


Introduction to ion channels & methods of study


Introduction to K channels (functional & topological
classifications, KcsA structure)


Bioinformatics & sequence
-
function prediction methods


construction of a comprehensive alignment of the K
channel family permeation pathway


Sequence
-
function predictions: Ca binding site in BK & K
channel selectivity analysis


Strategy for exploration of Na, Ca channels


Ion channel database proposal

Goal of project
: Discover sequence
-
function relationships in
ion channels and other membrane proteins by identifying K
channel residues responsible for ion selectivity, conductance,
and toxin affinities. This will be done by:


Gathering available experimental data (ex. from
mutation experiments) regarding function of individual
residues within K channels


Using bioinformatics (identifying structural features,
sequence searches in databases, comparing sequence
profiles, etc.) to check whether these functionalities can
be projected within the channel class, throughout the K
channel family, and across different ion channel
families

Ion channels


Ion channels are excitable
protein molecules in the lipid
bilayer.


Passage of ions across cell
membrane is essential for
excitation and electrical
signaling


Many ion channels are
characterized by high
conductance and ion
selectivity


Na
+

& Ca
2+

channels have 4
homologous repeats of 6 TMS
each. K
+

channels have 4
identical subunits, of 2, 6, or
10 TMS each.

Methods of studying ion channels


Patch
-
clamp techniques help determine single channel activities to
analyze channel electrophysiology


Studying protein sequences of ion channels and evaluation of data
from mutation experiments


Using FRET (Fluorescence Resonance Energy Transfer)
experiments to measure gating movements; determining contiguity
of channel segments from NMR spectra; estimating channel
topology from cysteine
-
scanning experiments


Using modeling programs to visualize channel activities (toxin
binding, ion passage through pore, etc.)


Structural determination of ion channels (such as the gramicidin
channel, MscL, certain porins, & KcsA K
+

channel)

K
+

channels


Characterized by wide variety
in structure & function


Exhibit similar ion
permeabilities: selectivity
sequence is
K
+
=Rb
+
>Cs
+
>Na
+
>Li
+


Can be blocked by TEA


a

subunits have 4 identical
subunits arranged around
central pore. Each contains 2
(M1, M2), 6 (S1
-
S6), or 10
(S1
-
S10) TMS.



K
+

Channel classifications
:

Functional characteristics of K channels
:


The S4 sequence in voltage
-
gated channels contains positively
charged basic residues at every 3
rd

position and serves as a
voltage
-
sensor


In certain channels, an extended N
-
terminal segment occludes the
pore in a ball
-
and
-
chain mechanism causing inactivation


In IRK’s, an Asp residue in M2 affects channel blocking by Mg
2+
,
which influences inward rectification and K selectivity


A C
-
terminal Ca
2+
-
sensing domain is responsible for Ca
2+
-

dependent activation in BK channels


In each subunit, S5 & S6 (in 6 or 10 TM channels), M1 & M2 (in 2
TM channels), and the linker joining them form part of the
permeation pathway. Residues responsible for ion selectivity, toxin
sensitivities, and channel conductance are suspected to be located
here. Linkers contain K channel “signature” G[Y/F]G sequence


Structure of the KcsA K
+

channel
:

The KcsA channel structure was discovered and refined to 3.2 A accuracy
by Doyle et al (1998).


The pore is constructed like an
inverse teepee, with selectivity filter
sequence at wide end


Selectivity filter is narrow (3 A) and
12 A long. Rest of pore is wider
with inert hydrophobic lining (thus
minimizing distance of strong ion
-
channel interactions and favoring
high K
+
throughput)


A large water
-
filled cavity near
channel center combined with helix
dipoles overcomes the electrostatic
destabilization of a K
+

ion in the
pore


The K
+

selectivity filter is lined with main chain oxygen
rings, which provide multiple closely
-
spaced binding
sites. As a dehydrated ion enters the filter, the carbonyl
oxygen atoms substitute for the water oxygen atoms.
The filter is constrained at an optimal radius to
coordinate a dehydrated K
+

ion, but the Na
+

ion is too
small for the carbonyl oxygens to provide similar
compensation.


Two K
+

ions 7.3 A apart in the selectivity filter provide a
force of repulsion which overcomes the strong ion
-
protein interaction to allow rapid conduction in an
environment of high selectivity

Sequence
-
function prediction methods


Identification of and filtering for structural features: mask
low
-
complexity regions (to prevent spurious hits); identify
transmembrane regions, internal repeats; predict
secondary structure


Identification of homologs: identify annotated domains
from dedicated databases prior to a BLAST search (to
reduce search space for remaining parts of protein);
search complete sequence databases individually using
subsequences separated by known domains; perform
exhaustive, iterative database searches; combine search
for sequence similarity with profile, motif, and pattern
searches


Prediction of protein function: consider domain
organization and distinct functions of individual domains
(keeping in mind that many proteins are multifunctional),
analyze database annotation if inconsistencies between
different homologues are detected; perform cluster
analysis of homologs to determine level of precision for
functional prediction; identify similarities to proteins with
known 3D structure to aid in model construction, which
may provide further functional insights

Bioinformatics


Bioinformatics can be loosely defined as the use of
computer technology to manage biological information.
One of its goals is to construct and utilize tools in order
to extract useful information pertaining to protein function
from available sequence data


Bioinformatics will be used in this project to expedite the
study of sequence
-
function relationships in K channels
(primarily to identify residues responsible for ion
selectivity, conductance, and toxin sensitivities) and to
project these relationships across different types of ion
channels.

Construction of a comprehensive alignment of K channel
families
:

The following procedure was used for construction:


All K channel classes and isoforms to be included in the alignment
were selected (K channel classes: DRK & A
-
type, Ca
2+
-
activated,
inward rectifiers, KCNG, EAG & ERG, plant K channels, 2
domains/subunit, and the KcsA K channel)


DRK/A was identified as the channel class with highest sequence
similarity to KcsA and S5 & S6 TMS in these channels were located


Profiles of the DRK S5 & S6 TMS were constructed


Corresponding TMS in all other K channel classes (called TM1 and
TM2) were located


Linker segments in each channel type were identified as the
difference between TM1 and TM2 and aligned


Procedure for K channel alignment construction


Selection of isoforms: A BLAST
search was performed on the
NCBI non
-
redundant database
using representatives from each of
the channel classes. Channels
were selected on the basis of
completeness in database records
and published literature


Identification of class closest to
KcsA: The DRK/A class was
chosen because it produced
ungapped alignments with KcsA.
S5 & S6 TMS in DRK were
identified based on alignments
with KcsA. All available DRK/A
channel sequences were found
and profiles for TM1 and TM2
were constructed (Figures 4, 5, &
6).

Procedure for K channel alignment construction (contd.)


Location of TM1 & TM2 in other K
channels: The S5 & S6 profiles
were used as probes to locate
TM1 & TM2 in other K channels.
A binary scoring scheme was
implemented.


Identification of linker segments:
Linker sequences joining TM1 &
TM2 were extracted from each
channel, arranged in order of
length, and aligned.

Observations from K channel alignment


Region of highest conservation is pore helix
-
selectivity
filter. KcsA structure here contains a hydrophobic
cluster (interactions serve to stabilize twisting backbone)
and a hydrophilic cluster (formation of intersubunit H
-
bonds creates a cuff
-
like sheet around selectivity filter,
allowing passage to dehydrated K
+

ions)


Region of maximum sequence and length variation is in
the turret segment. This part constitutes the
extracellular entryway of channel and interacts with
different toxins depending on channel class

Sequence
-
function predictions
:

a) Identification of putative Ca
2+
-
binding region of BK


Sequence analysis showed this
channel class contained a long
C
-
terminal domain absent in
other K channels


BLAST searches using the C
-
terminal segment yielded other
Ca
2+
-
binding proteins such as
troponin & calmodulin


Alignment of this segment with
troponin showed sequence
similarity in troponin’s known
Ca
2+
-
binding region


Hydropathy plots showed this
segment was likely to be present
in the cytoplasm

Sequence
-
function predictions
:

b) K channel selectivity analysis


2 sets of sequence profiles of
the pore helix
-
selectivity filter
region were constructed: one
from 94 highly
-
selective, one
from 8 weakly
-
selective
channels.


Profile comparison showed
Trp68 & Thr72 were present in
highly sel. Channels whereas
Lys68 & His72 were at
corresponding positions in
weakly sel. Channels

Sequence
-
function predictions
:

b) K channel selectivity analysis (contd.)


In KcsA, Trp68 & Thr72 form
H
-
bonds with Tyr78 (of GYG).
This bond, which might be a
determinant of K channel
selectivity, is preserved only in
highly K
-
sel. Channels


Molecular dynamics
calculations show that ease of
movement through selectivity
filter is modulated by tightness
of restraint in the pore helix
region.


Homology modeling & Fragment cluster
analysis


A homology model of the Shaker channel was built using
MODELLER and evaluated using PROCHECK by R.
Shealy. Models of other K channels can be similarly
constructed and might provide insights into the
functioning of K channels


A fragment clustering analysis technique was applied on
the A chain of KcsA by G. Hunter. The KcsA permeation
pathway sequence showed a reasonable fit within the
bounds of the available fragment cluster set. An
investigation into which proteins from the PDB are
similar in sequence and structure to the KcsA
permeation pathway still remains to be made.

Na
+

& Ca
2+

channel analysis

The strategy for exploring properties of these channels is
similar to what is being done with K channels:


Select all Na & Ca channel isoforms to be included in
study


Identify channel class from each family with highest
sequence homology to KcsA and identify S5 & S6 in
these (each of the 4 repeats in Na and Ca channels
have to be analyzed separately)


Construct profiles of S5 and S6 segments


Use these profiles to locate corresponding TMS in all
other channel classes in each family


Identify the linkers from each domain, arrange in order of
length, align, and study observable characteristics

Proposal for construction of an ion channel
database


This will be a queryable HTML database, initially
containing voltage
-
gated K channels (other types of
channels will be added once a working format has been
constructed)


Existing ion channel
-
related databases on the web
contain limited information on selected channel types
and are not as comprehensive as the one we are
constructing

Proposal for construction of an ion channel
database (contd.)


We have created a web query form and an indexing
program called Ndjinn. Ndjinn can be used to search,
index, and retrieve records from the database
-

it works
by indexing the entire text of files contained in multiple
database entries


Upon its completion, the database will be incorporated
into the Biology Workbench. This will enable users to
access DNA and protein sequences, and obtain
mutational, electrophysiological, and functional data of
ion channels from one source.