Sandra Orchard (V2 15/09/09)
UniProt – the protein sequence database
www.uniprot.org
UniProt (Universal Protein Resource) is the world's most comprehensive
catalogue of information on proteins. It is a central repository of protein
sequence and function. The UniProt Knowledgebase (UniProtKB) is the
central access point for extensive curated protein information, including
function, classification, and cross-reference. The database is divided into two
section UniProtKB/SwissProt which is manually curated and UniProtKB/TrEMBL
which is automatically maintained. During this course you will concentrate on
UniProtKB/SwissProt and learn how to access the entries in the database,
extract the maximum amount of information from them.
BLAST (Basic Local Alignment Search Tool), finds regions of sequence
similarity and gives functional and evolutionary clues about the structure and
function of your novel sequence.
For this exercise, we will use the BLAST program on the UniProt website,
which allows you to restrict your search by taxonomic group.
This work is licensed under the Creative Commons Attribution-Share Alike 3.0 License. To view a copy
of this license, visit http://creativecommons.org/licenses/by-sa/3.0/ or send a letter to Creative Commons, 543
Howard Street, 5th Floor, San Francisco, California, 94105, USA.
1
Exercise 1 – Exploring UniProtKB
Go to the BLAST tool on the UniProt.org website. Paste in the Mystery protein sequence
below. Restrict the search to Vertebrates.
Mystery Protein sequence
MSRESDVEAQQSHGSSACSQPHGSVTQSQGSSSQSQGISSSSTSTMPNSSQSSHSSSGTL
SSLETVSTQELYSIPEDQEPEDQEPEEPTPAPWARLWALQDGFANLECVNDNYWFGRDKS
CEYCFDEPLLKRTDKYRTYSKKHFRIFREVGPKNSYIAYIEDHSGNGTFVNTELVGKGKR
RPLNNNSEIALSLSRNKVFVFFDLTVDDQSVYPKALRDEYIMSKTLGSGACGEVKLAFER
KTCKKVAIKIISKRKFAIGSAREADPALNVETEIEILKKLNHPCIIKIKNFFDAEDYYIV
LELMEGGELFDKVVGNKRLKEATCKLYFYQMLLAVQYLHENGIIHRDLKPENVLLSSQEE
DCLIKITDFGHSKILGETSLMRTLCGTPTYLAPEVLVSVGTAGYNRAVDCWSLGVILFIC
LSGYPPFSEHRTQVSLKDQITSGKYNFIPEVWAEVSEKALDLVKKLLVVDPKARFTTEEA
LRHPWLQDEDMKRKFQDLLSEENESTALPQVLAQPSTSRKRPREGEAEGAETTKRPAVCA
AVL
The
resulting
scores
should
suggest
a
100%
match
to
CHK2_HUMAN (O96017). You should also note a match to the isoforms of
CHK2 – each isoform is treated as a separate transcript, information which is
often lost when using other protein sequence databases.
Click on the top green bar in the Local Alignment column to display the aligned sequences.
The UniProt Consortium is comprised of the European Bioinformatics Institute,
the Swiss Institute of Bioinformatics and the Protein Information Resource.
The UniProt consortium aims to support biological research by maintaining a
high quality database that serves as a stable, comprehensive, fully classified,
richly and accurately annotated protein sequence knowledgebase, with
extensive cross-references and querying interfaces freely accessible to the
scientific community.
All data stored in UniProt can be downloaded from the Download Centre at
http://www.ebi.uniprot.org/database/download.shtml
Return to the search page and open O96017/CHK2_HUMAN.
Click on the EC number link under protein names.
Q 1. What is the reaction catalysed by this enzyme?
Return to the UniProtKB/Swiss-Prot entry and click on the taxonomic identifier.
2
The NEWT database is a compilation of the information within the NCBI
Taxonomy database together with proteins found in the Swiss-Prot and
TrEMBL section of UniProtKB. It is maintained by the Swiss-Prot group in
Switzerland. For each species, NEWT displays the following taxonomy data:
Swiss-Prot scientific name, Swiss-Prot common name and Swiss-Prot
synonym, lineage, number of protein sequence entries in Swiss-Prot and
TrEMBL.
The NEWT data is available from the European Bioinformatics Institute.
Q 2. In UniProtKB how many proteins are there from Homo sapiens
neanderthalensis?
(Hint – Use the taxonomic hierarchy.)
Return to the entry and scroll down.
Q 3. How many other species also have proteins which are members of the
CHK2 protein kinase subfamily?
Return to the entry and scroll to Binary Interation.
Q 4. How many evidences are there for the interaction of CHK2 with PLK1.
Click on the hyperlink to IntAct to see which publication these evidences come
from.
Alternative Products in UniProtKB
Return to the entry and go to the Alternative Products section.
Chek2 is a protein for which multiple (12) isoforms have been identified. All of
these are mapped within UniProtKB, and given stable identifiers.
Press the button “Align” to see how all the isoforms differ.
Return to the Alternative Products section.
Q 5. What functional characteristic has the alternative splicing has
conferred to isoform 12 (and other isoforms)?
3
To see the location and length of the missing region in isoform 12 go the UniProt entry
(click O96017) and look in the Sequence annotation section under Natural variations.
Q 6. Using the information in the table, explain why isoform 12 is
catalytically inactive?
(Hint – where’s the active site.)
Q 7. Where in the cell is isoform 12 to be found?
(Hint – look in Comments –> Subcellular location)
Q 8. What is the length of isoform 12?
Adding information using the InterPro Database
At the bottom of the entry, are all the database cross-references for this
protein. Look for those contained within the InterPro database. InterPro is an
integrated documentation resource for protein families, domains and sites.
InterPro combines a number of databases (referred to as member databases)
that use different methodologies and a varying degree of biological
information on well-characterised proteins to derive protein signatures. By
uniting the member databases, InterPro capitalizes on their individual
strengths, producing a powerful integrated diagnostic tool
Signatures describing the same protein family, domain repeat or site are
grouped into unique InterPro entries. Each combined InterPro entry has a
unique accession number, an abstract describing the features of proteins
associated with the entry and literature references and has links to the
relevant member database(s). All UniProtKB protein sequences that have
matches to a particular InterPro entry are listed in the Match Table associated
with that entry. There are also links to the InterPro graphical views. The
graphical views, which can be sorted by UniProtKB accession number,
structure or taxonomy, show the position of the signatures on the protein,
mousing over the signature brings up a pop-box, giving the accession, name
and position.
InterPro graphically represents the location of a protein domain and
information pertaining to the origin of that domain and the proteins that
contain it. Families are also defined and may contain several InterPro domains
which are often, but not always, in the same order.
InterPro and InterProScan are accessible for interactive use over the EBI web
server (www/ebi.ac.uk/interpro), they are distributed as stand-alone copies
by anonymous ftp.
InterPro entries can be linked to one another through PARENT/CHILD and
CONTAINS/FOUND IN relationships. PARENT/CHILD relationships indicate
4
superfamily/family/subfamily relationships, as well as domain hierarchies,
where sequences can be subdivided into more specific sub-sets.
CONTAINS/FOUND IN relationships apply to domains, repeats and sites within
families, and are used to describe the composition of protein sequences.
InterPro entries that belong to a UniProtKB entry can be found in Crossreferences -> Family and domain databases
 Moving back and forwards between the entry and the
InterPro domains and active sites which it contains,
answer the following questions
Exercise 2 – Searching InterPro using a UniProt identifier
Open the development version of the InterPro homepage (InterPro beta) in a web browser
(http://wwwdev.ebi.ac.uk/interpro/).
Using the “Text” Search box mid-way down the page, type in the UniProtKB accession
‘O15075’ (without the quotes. That’s a letter O at the start and a zero in the middle). Click on the
purple “Search” button.
You should now have a page describing the signature matches for this protein. (the protein view):
Question 1: Looking at the InterPro protein view for O15075, how many
InterPro entries (not individual signatures) match the query protein
sequence?
Question 2: How many domains is the protein divided up into?
On the protein page for O15075, click on “Detailed results” in the left hand
side menu. This will show you all of the protein signatures within InterPro that
were used to create the information in the preceding page.
Question 3: How many member database signatures contribute to InterPro
entry IPR003533?
Hint: You can see the contributing member database signatures to InterPro
entry IPR003533 in the “Detailed results” view or alternatively click on the link to
IPR003533, which will take you to the entry page for that domain.
Go back to the “Overview” and scroll down to the “Structural features” just below the view
of unintegrated signatures.
Under the “Structural Features” heading you will find the PDB structure. Its length indicates the
region of the protein for which the structure is known. You will also see bars representing a CATH
database match and a SCOP database match, both of which are structural classification databases
that break down the PDB structures for the protein into their constituent domains.
5
Question 4: What region is covered by the PDB structure (ie which domain)?
Hint: Compare it to entry IPR003533.
Not all of the protein has been structurally characterised, shown by the fact that only a small
region of this protein is covered by the PDB match. To help address this problem, there are
homology models from both ModBase and Swiss-Model found under the “Structural Predictions”
section. These are models based on aligning our protein with its closest homologue whose
structure has been determined. (Note: these are predictive models that provide a ‘best guess’ at the
remaining structure).
Question 5: Why does IPR003533 have two domain hits compared to the single domain the
PDB structure?
Note the structural view at the top right of the page and click on the ‘Go’ purple bottom. It
will bring you to the PDB page, where you can visualise the 3-D structure by clicking on Jmol (it
will pop up in a new window).
Exercise 3 – Exploring InterPro entries
General annotation
Still on the protein page for O15075, look at the match to the entry IPR000719. Notice that there
are other domain entries that cover approximately the same sequence position.
Click on the hyperlink to IPR000719
Question 1: What is the name of this domain?
Now look at the “Contributing signatures” section.
Question 2: Which signatures make up this entry?
This section lists the signatures in an entry, the database they come from, and the number of
proteins they match.
Relationships
Question 3: What “Child” entries is IPR000719 subdivided into?
InterPro links related signatures through Parent/Child relationships
which indicate domain/family hierarchies. Child entries subdivide
IPR000719 into more closely related subgroups.
Question 4: What is the name of the “Parent” of IPR000719?
In this case, the parent entry represents domains with a structural fold homologous to that of the
protein kinase domain (even if they have no enzyme activity), whereas IPR000719 represents a
more specific form of the domain that has catalytic protein kinase activity.
GO (Gene Ontology) terms
Scroll down to the “GO terms annotation” section.
6
Question 5: What GO terms are provided for this entry?
InterPro provides its own mappings to GO terms based on the curated
UniProt/Swiss-Prot proteins matching an entry. These are useful for
the annotation of TrEMBL proteins that do not otherwise have GO
terms associated with them. (Note: Next to the Go terms you can find
a link to other entries in InterPro that share that Go term. This is of
value if you are interested in searching for InterPro entries that match
proteins with a specific function or those involved in a specific
process).
Now look at the information in the side menu for entry
Question 6: How many proteins are matched by IPR000719?
Structures
Now click on the "Structures" link on the left hand side menu.
InterPro provides a list of all the PDB entries associated with an entry. There are also structural
links to SCOP and CATH at the bottom of the page, which provide structural classifications of the
proteins that match this entry.
Scroll to the bottom of the page and follow the “SCOP d.144.1.7”
link to the SCOP database to find out the structural classification of
this domain.
Question 7: What type of structure does the protein kinase-like
fold consists of?
Hint: look at the information under “Fold” in the “Linage” section.
Taxonomy
Click on the browser back button till you are in the InterPro page again, then
click on the "Species" link on the left hand side of the InterPro entry page for
IPR000719. You may explore the taxonomic spread by expanding the table.
InterPro divides all the protein hits in an entry by their taxonomy.
Question 8: How wide a taxonomic coverage do proteins containing a protein kinase domain
have?
7
Searching UniProtKB
 Return to the search engine at the top of the entry page
and click on Fields
 Use the fields to search on IPR000719 AND
Organism=human to see how many protein kinases you
can find.
Q 11. How many of those are in Swiss-Prot (i.e. reviewed
records)?
Q 12. Furthermore, how many of those have experimental
evidence of being associated with the cell membrane?
(As well the subcellular location fields in the Comments section other fields
such keywords and gene ontology also have subcellular location information.
However, in the UniProt search facility only the subcellular location fields
allows the evidence to taken in to account.)
8
Further Info
Sequence Searching
For more sequence searching tools visit http://www.ebi.ac.uk/Tools/similarity.html.
There are two main programs that implement BLAST searches; WU-BLAST 2.0
and NCBI BLAST2. They are distinctly different software packages, although they have a
common lineage for some portions of their code, so the two packages do their work
differently and obtain different results and offer different features. You can also check for
vector contamination with Blast2 EVEC.
Fasta can be very specific when identifying long regions of low similarity
especially for highly diverged sequences. You can also conduct sequence similarity
searching against complete proteome or genome databases using the Fasta
programs.
MPsrch – Smith and Waterman algorithm, capable of identifying hits in cases
where Blast and Fasta fail and also reports fewer false-positive hits.
Further Reading

The UniProt Consortium “The Universal Protein Resource (UniProt).”
Nucleic Acids Res. (2008) 35:D193-197

Leinonen R, Nardone F, Zhu W, Apweiler R “UniSave: the UniProtKB
sequence/annotation version database.” Bioinformatics (2006) 22:12841285

Suzek BE, Huang H, McGarvey P, Mazumder R, Wu CH “UniRef:
comprehensive and non-redundant UniProt reference clusters.”
Bioinformatics (2007) 23:1282-8

Mulder NJ et al “New developments in the InterPro database.” Nucleic
Acids Res. (2007) 35:D224-8

Kerrien S et al “IntAct--open source resource for molecular interaction
data.” Nucleic Acids Res. (2007)35:D561-565
Exercise 1 – UniProt
9
Q 1. What is the reaction catalysed by this enzyme?
ATP + a protein = ADP + a phosphoprotein
Q 2. In UniProtKB how many proteins are there from Homo sapiens neanderthalensis?
23
Q 3. How many other species also have proteins which are members of the CHK2 protein
kinase subfamily?
13 proteins in 6 species
Q4 How many evidences are there for the interaction of CHK2 with PLK1. Click on the
hyperlink to IntAct to see which publication these evidences come from.
6 (Jan 2012) – PMID: 12493754
Q 5. What functional characteristic has the alternative splicing has conferred to isoform 12
(and other isoforms)?
Inactive
Q 6. Using the information in the table, explain why isoform 12 is catalytically inactive?
Alternative splicing removes the active site
Q 7. Where in the cell is isoform 12 to be found?
Nucleus
Q 8 What is the length of isoform 12?
514
Exercise 2 - Searching InterPro using a UniProt identifier.
Question 1: Looking at the InterPro protein view for O15075, how many InterPro entries
(not individual signatures) match the query protein sequence?
Answer 1: Seven (six domain/sites entries plus the family membership entry)
Question 2: How many domains is the protein divided up into?
Answer 2: Three (one of them is found repeated in the protein)
Question 3: How many member database signatures contribute to InterPro entry
IPR003533?
10
Answer 3: Five
Question 4: What region is covered by the PDB structure (ie which domain)?
Answer 4: The first doublecortin domain of the two represented by InterPro entry
IPR003533.
Question 5: Why does IPR003533 have two domain hits compared to the single domain the
PDB structure?
Answer 5: IPR003533 predicts the presence of two doublecortin domains, but only the area
corresponding to the first one has been structurally characterised and therefore appears in
the PBD structure.
Exercise 3 - Exploring InterPro entries
Question 1: What is the name of this domain?
Answer 1: Protein kinase, catalytic domain
Question 2: Which signatures make up this entry?
Answer 2: Two signatures make up this entry (PF00069 and PS50011)
Question 3: What “Child” entries are IPR000719 subdivided into?
Answer 3: Two: Serine-threonine/tyrosine protein kinase catalytic domain and Serinethreonine/dual-specificity protein kinase catalytic domain
Question 4: What is the name of the “Parent” of IPR000719?
Answer 4: Protein kinase-like domain
Question 5: What GO terms are provided for this entry?
Answer 5: GO:0006468 (protein phosphorylation), GO:0004672 (protein kinase activity),
GO:0005524 (ATP binding)
Question 6: How many proteins are matched by entry IPR000719?
Answer 6: 111054 proteins are matched
Question 7: What type of structure does the protein kinase-like fold consists of?
11
Answer 7: Consists of two alpha+beta domains, and the C-terminal domain is mostly alpha
helical.
Question 8: How wide a taxonomic coverage do proteins containing a protein kinase domain
have?
Answer 8: It is widely spread, being found in eukaryota, bacteria, archaea and viruses.
Exercise 3- Analysing sequences using InterProScan
Question 1: Based on the InterPro matches, what domains is the protein predicted to
possess? What family is it predicted to belong to?
Answer 1: A six-hairpin glycosidase-like domain (IPR008928) and belongs to the
Lanthionine synthetase C-like family (IPR007822), subfamily LanC-like eukaryotic
(IPR20464).
Question 2: What do your BLAST results suggest your protein to be?
Answer 2: G protein coupled receptor
Question 3: Are the results consistent with those returned by InterProScan? Is there anything
in the InterPro annotation that might explain any discrepancies?
Answer 3: InterPro entry IPR007822 explains that due to its structure some authors
considered LanC-1 and other members of the family to be novel G protein-coupled
receptors, but this claim has since being refuted. Nevertheless, many sequences in the
protein databases were submitted as G protein-coupled receptors and still carry this
description. This creates some confusion when doing a BLAST search.
12