The Influence of Alternative Splicing in Protein Structure

1
The Influence of Alternative Splicing in Protein
Structure

• Durham, E. H. A. B.1, Garratt, R. C.2, de Souza,
  S. J.3
• 1 Bioinformatics PhD student - University of
  São Paulo - Brazil
• 2 Physics Insitute (São Carlos) - University of
  São Paulo - Brazil
• 3 - Ludwig Institute for Cancer Research São
  Paulo Branch - Brazil
• e-mailelza_at_compbio.ludwig.org.br

R E S U L T S
I N T R O D U C T I O N
The fact that gene number is not significantly
different between mammals and some invertebrates
suggests that other mechanisms are being used to
generate diversity, such as alternative splicing
(AS) and post-translational modifications. AS
could be understood as a single gene originating
different mRNA sequences which can occur by the
use of alternative splice sites. The major types
of AS are intron retention (IR), alternative
splice sites usage (AU), exon skipping (ES) and
mutually exclusive exons. It is know that some
AS variants are tissue-specific and/or associated
with several diseases in humans, as cancer.
However, AS can create thousand of mRNA sequences
and their functional viability has been
questioned. Some studies indicate that variants
with a frame shift and/or premature stop codons
will be degraded. Some suggested that a high
number of ESTs/mRNAs supporting a variant
correlates with its functionality while others
use the comparison between human and other
organisms (mouse, rat) to exclude not functional
sequences. Many computational tools have been
used to find and compare alternative splicing
variants. Generally, cDNA, mRNA, ESTs and protein
sequences that are public available are aligned
against each other or against the genome to
identify splicing isoforms. Most of this
information is usually deposited in relational
databases with open access. This can be used to
join all sequence information related to variants
as size, frame shift, insertions, deletions,
repetitive elements and domains. Some previous
studies correlated the effect of alternative
splicing in protein structures. Of them, some are
focused on protein families while others do not
cover all possible protein modifications caused
by alternative splice sites. So, there still
exists a lack of information about the protein
structure modifications as a consequence of
alternative splicing.
Aminoacids composition of alternative and
constitutive boundaries
A
B
(A) Conditional probability for a pairs of
aminoacids in human PDB data set (B) Comditional
probability for alternative boundaries
Frequency of human protein aminoacids (black),
constitutive (white) and alternative (grey)
boundaries
Spatial distances of alternative boundaries
deleted in protein structures
A
B
Distance (Angstron)
1 10 20 30 40 Size (aminoacids)
A B O U T T H I S W O R K
(A) Distance of human protein regions with
different sizes
(B) Distance of deleted regions with alternative
splice boundaries
In this study we intend to identify and
distinguish human protein structures modified by
alternative splicing. In order to do it, mRNAs
and EST sequences from UCSC were mapped to the
human genome using BLAT and SIM4. All mapped
sequences were deposited in a local database and
the splicing boundaries from all sequences from a
gene were compared to identify splicing variants.
Those variants were assigned as IR, AU or ES
events. We constructed a pipeline where TBLASTN
was performed between those variants (829.212
mRNA and EST sequences) and a set of 3.196
non-redundant PDB human sequences. .Some BLAST
parameters were carefully adjusted to allow gap
opening and extension and identity was
recalculated considering the gap size. Terminal
regions without alignment were resubmitted to
TBLASTN and the correct splice boundaries were
assigned. Sequences with identity greater than
70 were included in our analysis, except for
those containing stop codons. Initially, the
non-redundant PDB structures were related to
1.364 Unigene clusters allowing a directly
association between the genes with alternative
splicing sequences and their structural effects
on proteins. Events in proteins were separated in
insertion and deletion, depending of the splicing
sequence alignment. Proteins with deletions
presented 7.427 donor and acceptor splice
boundaries mapped into 1.662 structures (716
Unigene clusters) while insertion had 5.673 cases
were related to 1.314 structures (585 Unigene
clusters). Other structural features were
analyzed, as motility (measured through
experimental B-factor values from PDB files)
which is one feature expected to vary in
determined regions of proteins was measured to
deletion and insertion boundaries as to deleted
regions. Spatial distance restraints between CA
atoms which can be used by alternative splicing
sequences to restraint the energy needed to fold
a new protein, was measured in deleted regions
and compare between the prototype and variant
structures. Association between interaction
regions (intra-protein and inter-protein) and
diseases are also in course.
Exposition and interaction of deleted alternative
boundaries
Contact Structural Units (interaction between
chains)
ProfBval (exposition and flexibility)
Alternative boundaries
Random Draw
RandomDraw
Alternative boundaries
Exposed and rigid
Total
Total
Exposed and rigid
Interacting
Interacting
Total
Total
D I S C U S S I O N
This work brings an authomatical method to find
proteins structures related to mRNA sequences
modified by alternative splicing. The pipeline
used here gives the precise position of the
splice site in protein structure and assign it as
an alternative or constitutive boundary. The
assignment of the boundaries is a specially
difficult task once, most of times, we did not
find the alternative boundaries from genomic data
in proteins structures. This study still intend
to highligth the structural modifications caused
by alternative splicing and each type of AS
events. Besides it will be a detailed description
of structures related to alternative splicing,
including their dynamic behavior (Molecular
Modeling and Dynamics in course) and one
experimental structure (X-ray) in future studies.
T H A N K S
Alan Durham for Perl lessons and support, Pedro
Galante for initial set of alternative splice
cases, Joao Muniz for usefull Modeller tips PhD
Bioinformatics program and CAPES for financial
support