Today it is simple to determine protein sequence, but

1
Bioinformatics and Genomics - Spring 2006 -
Protein Structure
Today it is simple to determine protein sequence,
but difficult to determine protein structure and
function. Tools to find structural and
functional info on proteins membrane bound or
soluble? secondary structure amphipathic
nature post-translational modifications
sub-cellular location any motifs or domains
Experimental Annotation
2
Bioinformatics and Genomics - Spring 2006 -
Analysis of Protein Sequences
Steps in process Run protein sequence through
BLASTp. Will reveal identity of subject
protein or similar proteins. Or will reveal that
you are analyzing a unique protein.
Predict hydrophobicity. Is it a soluble or
membrane protein? Predict secondary
structure. Predict sequence motifs and
patterns. Predict sub-cellular location.
3
Bioinformatics and Genomics - Spring 2006 -
Analysis of Protein Sequences
Steps in process Run protein sequence through
BLASTp. Will reveal identity of subject
protein or similar proteins. Or will reveal that
you are analyzing a unique protein.
Predict hydrophobicity. Is it a soluble or
membrane protein? Predict secondary
structure. Predict sequence motifs and
patterns. Predict sub-cellular location.
Assignment Pick a protein that relates to
your thesis work (or rotation). Analyze the
protein sequence for transmembrane domains,
secondary structure, motifs and patterns,
sub-cellular location, and functional
class. Present a predicted topology or
structure for the test protein, indicating where
key motifs and localization signals are found.
4
Proteomics Servers and Tools
5
Proteomics Servers and Tools
6
Transmembrane Domains
7
Transmembrane Domains
Hydropathy scale - aa assigned value reflecting
relative phob/phil, based on Partition
coefficient of aa between water and phase that
resembles protein interior (ethanol).
Determine free energy of transfer of aa from
ethanol to water. Distribution of each aa in
12 globular proteins whose atomic coordinates
were known. Strong correlation between aa
buried and PC.
8
Transmembrane Domains
Simple computer program to systematically scan
amino acid sequence, determine average hydropathy
of moving segment as it advances through the
sequence.
Window too short noisy, too wide miss
features. MRVGKHDEKGHEDRGLMVEAVIFGSVGRILMGFWIPD
KEHRHKR VIFSHDEKHEDRLMVGARVIFSVFGGILEMGFPDKEHRHVKR
HG DEKHEKVEHRHK
http//fasta.bioch.virginia.edu/fasta_www/grease.h
tm
9
Transmembrane Domains
Human glycophorin
10
Transmembrane Domains
Transmembrane sequence is thought to be in what
conformation? a-helix, ß-sheet or random
coil An a-helix of amino acids will
span the membrane. A ß-sheet of amino
acids will span the membrane. Will the KD
algorithm predict an amphipathic a-helix? What
type of protein will have an amphipathic
transmembrane domain?
11
Transmembrane Domains
VSV G protein
12
Transmembrane Domains
13
Transmembrane Domains
14
Transmembrane Domains
Test protein Niemann-Pick disease type C 1
(NPC1) protein MTARGLALGL LLLLLCPAQV FSQSCVWYGE
CGIAYGDKRY NCEYSGPPKP LPKDGYDLVQ ELCPGFFFGN
VSLCCDVRQL QTLKDNLQLP LQFLSRCPSC FYNLLNLFCE
LTCSPRQSQF LNVTATEDYV DPVTNQTKTN
VKELQYYVGQ SFANAMYNAC RDVEAPSSND KALGLLCGKD
ADACNATNWI EYMFNKDNGQ APFTITPVFS DFPVHGMEPM
NNATKGCDES VDEVTAPCSC QDCSIVCGPK PQPPPPPAPW
TILGLDAMYV IMWITYMAFL LVFFGAFFAV WCYRKRYFVS
EYTPIDSNIA FSVNASDKGE ASCCDPVSAA FEGCLRRLFT
RWGSFCVRNP GCVIFFSLVF ITACSSGLVF VRVTTNPVDL
WSAPSSQARL EKEYFDQHFG PFFRTEQLII RAPLTDKHIY
QPYPSGADVP FGPPLDIQIL HQVLDLQIAI ENITASYDNE
TVTLQDICLA PLSPYNTNCT ILSVLNYFQN SHSVLDHKKG
DDFFVYADYH THFLYCVRAP ASLNDTSLLH DPCLGTFGGP
VFPWLVLGGY DDQNYNNATA LVITFPVNNY YNDTEKLQRA
QAWEKEFINF VKNYKNPNLT ISFTAERSIE DELNRESDSD
VFTVVISYAI MFLYISLALG HIKSCRRLLV DSKVSLGIAG
ILIVLSSVAC SLGVFSYIGL PLTLIVIEVI PFLVLAVGVD
NIFILVQAYQ RDERLQGETL DQQLGRVLGE VAPSMFLSSF
SETVAFFLGA LSVMPAVHTF SLFAGLAVFI DFLLQITCFV
SLLGLDIKRQ EKNRLDIFCC VRGAEDGTSV QASESCLFRF
FKNSYSPLLL KDWMRPIVIA IFVGVLSFSI AVLNKVDIGL
DQSLSMPDDS YMVDYFKSIS QYLHAGPPVY FVLEEGHDYT
SSKGQNMVCG GMGCNNDSLV QQIFNAAQLD NYTRIGFAPS
SWIDDYFDWV KPQSSCCRVD NITDQFCNAS VVDPACVRCR
PLTPEGKQRP QGGDFMRFLP MFLSDNPNPK CGKGGHAAYS
SAVNILLGHG TRVGATYFMT YHTVLQTSAD FIDALKKARL
IASNVTETMG INGSAYRVFP YSVFYVFYEQ YLTIIDDTIF
NLGVSLGAIF LVTMVLLGCE LWSAVIMCAT IAMVLVNMFG
VMWLWGISLN AVSLVNLVMS CGISVEFCSH ITRAFTVSMK
GSRVERAEEA LAHMGSSVFS GITLTKFGGI VVLAFAKSQI
FQIFYFRMYL AMVLLGATHG LIFLPVLLSY
IGPSVNKAKS CATEERYKGT ERERLLNF
15
Transmembrane Domains
16
Transmembrane Domains
Human Niemann-Pick C 1 (NPC1)
17
Transmembrane Domains
18
Transmembrane Domains
TopPred of NPC1
Analyzes hydrophobicity and topology (positive
residues stay on cytoplasmic side of ER membrane).
19
Transmembrane Domains
SOSUI of NPC1
hydropathy plus helical wheel diagrams
20
Transmembrane Domains
SOSUI of NPC1
21
Transmembrane Domains
Test protein Niemann-Pick disease type C (NPC2)
protein MRFLAATFLL LALSTAAQAE PVQFKDCGSV
DGVIKEVNVS PCPTQPCQLS KGQSYSVNVT FTSNIQSKSS
KAVVHGILMG VPVPFPIPEP DGCKSGINCP IQKDKTYSYL
NKLPVKSEYP SIKLVVEWQL QDDKNQSLFC WEIPVQIVSH L
22
Secondary Structure
Three dimensional structure of protein is a
topological organization of small local
structures, i.e. secondary structures.
helices strands other (loops) Different
amino acids prefer to be in different secondary
structures. PHDsec based on algorithms that
learn amino acids probability of adopting
particular structure from a large set of
examples. Server searches for homologous
proteins that are assumed to have similar
structures as the query. PSIPRED and SAM-T99
similar. Metaservers submit query to multiple
prediction servers and generate consensus
prediction. Solvent accessibility predicts
regions with potential to interact with other
proteins, metal atoms, or ions. buried
exposed
23
Secondary Structure
PHDsec results for NPC1
Helix
Strand
Neither
24
Secondary Structure
PHDsec results for NPC2
25
Tertiary Structure
Quick PDB of NPC2
Helix / Extended / Turns
26
Predicting Function
Predict function of new protein from known
information on homologous proteins. Identify
motifs, i.e. sequence elements that are
evolutionarily conserved and have functional
importance. Identify patterns or signatures,
i.e. common characteristics of a protein
family. PROSITE Pfam BLOCKS Predict
sub-cellular location based on known signal
peptides. PSORT TargetP Predict functional
class based on cellular role, enzyme class and
Gene Ontology category. ProtFun
27
Motifs and Patterns
28
Motifs and Patterns
29
Motifs and Patterns
30
Motifs and Patterns
31
Sub-cellular Location
32
Sub-cellular Location
PSORT II of NPC1
33
Sub-cellular Location
PSORT Users Manual In this version of PSORT,
the prediction accuracy for lysosomal/vacuolar
proteins is disastrous.
34
PSORT II of NPC2
35
Sub-cellular Location
SP secretory pathway
36
Functional Class
37
Functional Class
38
Functional Class
Prob Probability that protein belongs to class
in question. Influenced by prior probability
of that class. Odds Odds that sequence belongs
to the class independent of prior probability.
39
Assignment
Pick a protein that relates to your thesis
work (or rotation). Analyze the protein
sequence for transmembrane domains, secondary
structure, motifs and patterns, sub-cellular
location, and functional class. Present a
predicted topology or structure for the test
protein, indicating where key motifs and
localization signals are found.