1. Recerca de selenoproteïnes en el genoma d’organimes eucariotes
Bioinformàtica
Didac Santesmasses
PhD student
Bioinformatics and genomics programme
Roderic Guigó's group
Centre for Genomic Regulation, Barcelona

2. Structure of a selenoprotein gene

3. Selenoprotein families include selenoproteins and cysteine homologues (= orthologues or paralogues)

4. Selenoproteins are generally misannotated
…protein DBs have the same limitations
Percentages are computed by comparison Selenoprofiles-Ensembl annotations – see Mariotti and Guigó, Bioinformatics.

5. Bioinformatics methods for selenoproteins
De novo: Selenogeneid (Castellano et al. 2001)
Homology based approaches:
UGA / Sec or UGA / Cys alignments
(e.g. Kryukov et al. 2003)
- Selenoprofiles (Mariotti and Guigó 2010)
- Seblastian (Mariotti et al. 2013)
SECIS prediction:
- SECISearch (Kryukov et al. 2003)
- SECISearch3 (Mariotti et al. 2013) – explained later

6. Why selenocysteine? What is cysteine used for?
mainly for: Disulfide bonds
Inter or intramolecular, important for the folding of proteins, their stability, and in many cases necessary for catalysis.
Thioredoxins are a large class of proteins that perform redox reactions using thiol/disulfide switches (catalytic cysteines).

7. Why selenocysteine?
The term “Thioredoxin” generally designates small oxidoreductase proteins, constituing a pool of enzymes for anti-oxidant defense. The thioredoxin system includes these proteins and those operating on them, both using them as electron donors, or maintaining them reduced.
Plenty of other proteins possess a thioredoxin-like fold, relying on the same local structure that includes a thiol/disulfide switch, and performing redox functions.
Selenocysteine is found almost always replacing a catalytic cysteine in redox proteins. Many selenoproteins possess a thioredoxin-like fold. Some are involved in the thioredoxin system. Consistently, selenocysteine is more reactive than cysteine, and it is a better reductant.
Anyway, some selenoproteins possess totally unrelated functions
(e.g. SelJ = eye crystallin, SelP = selenium storage and transport)

8. Thioredoxin Reductases (TR)
Examples:
Thioredoxin Reductases (TR) S
+NADPH+H+ TR SH
Trx
Trx
+NADP- S SH
3 paralogous genes are present for this family in human, all of which are selenoproteins. Sec is present as penultimate residue, and it is catalytic.
TR are the only responsible for the reduction of thioredoxins in cell.
TR3: mitochondrial
Found in all vertebrates
TGR (TR2): can reduce glutathione disulfide. Contains a glutaredoxin (Grx) domain
Found only in tetrapodes
TR1: cytosolic.

9. Glutathione Peroxidases (GPx)
R-OOH + 2 GSH R-OH + H2O + GSSG
GPx
3 groups:
GPx1/GPx2,
GPx3/GPx5/GPx6,
GPx4/GPx7/GPx8
Reduces superoxides at expenses of glutathione (GSH).
8 paralogues in human, 5 with Sec
GPx1: cytosolic, abundant in liver and red blood cells
GPx2: cytosolic, found in liver and gastrointestinal system
GPx3: secreted, found in plasma and intestine
GPx4: cytolic / mytochondrial, abundant in testis.
Can reduce phospholipid hydroperoxides
GPx5: secreted/membrane bound.
Found only in epididymis
GPx6: secreted (?).
Found in olfactory epithelium and embryonic tissues
GPx7: secreted (?)
GPx8: membrane bound (?)
GPx6 converted to Cys independently in 3 species

11. Selenoproteins in mammals / vertebrates
20 duplications
9 gene losses
13 Sec  Cys 28

12. This is very different than the situation in insects:
Among vertebrates, selenoproteins are quite conserved: most of them are found in all species, few transformations occurred.
This is very different than the situation in insects:
(Chapple and Guigó, 2008)

13. Selenoproteins as test case
Selenoproteins have the peculiar characteristic of possessing a UGA codon, recoded because of the presence of the SECIS element.
If you learn how to predict selenoproteins, you are able to do the same with any “standard” protein family.
Bioinformatics project:
find all selenoproteins in a given genome

14. UPF Biologia. Curs
2007/08 – 2008/09: find all selenoproteins in a given protist genome
2009/10 – 2011/12: find a given selenoprotein family in all protist genomes
2012/13 – 2016/17: find all selenoproteins in a given vertebrate genome

16. selenoproteins in vertebrates
Project
selenoproteins in vertebrates
Objective: find all selenoproteins in a given vertebrate genome

17. selenoproteins in vertebrates
Project
selenoproteins in vertebrates
viquipedia: species description
selenoDB: insert your selenoprotein genes predictions into a real world database. Available to the scientific community.

18. Insert genes form

19. Useful resources Sequences of your assigned species: Ensembl
Collection of genomes (and annotations)
Your assigned genomes will be available in the UPF computers when you will start the project

20. Useful resources Sequences of your assigned species: NCBI nucleotide
Collection of all sequences (genomes, ESTs, etc)
If you do not find something that you expect to be there, look for other sources of sequences. Most genomes today are low/medium-quality

21. Useful resources Selenoprotein sequences: SelenoDB 2.0 (and 1.0)
Database containing manual annotations for human, and automated annotations (selenoprofiles) for other vertebrate species.
NCBI protein
NCBI hosts the sets of all known proteins. Noisy, but comprehensive. You can find here more selenoprotein sequences searching by homology (blast) or by keywords

22. Useful resources Tools: Blast - typically tblastn
Exonerate - protein2genome mode
Genewise
T-coffee
Webserver with SECISearch3 and Seblastian:
S14. Anotació de genomes (I)
S15. Anotació de genomes (II)

23. SECISearch 3 Mariotti et al. 2013
Based on a manually curated 2ndary structure alignment
Combines up to 3 methods to ensure maximum sensitivity
Filter and grading procedure based on manual inspection of hundreds of SECIS elements
Mariotti et al. 2013

24. annotated homologue(s) (Sec/Cys) in the reference protein database
Seblastian
Assumptions:
the presence of a detectable SECIS within acceptable genomic distance from the Sec-UGA
annotated homologue(s) (Sec/Cys) in the reference protein database
Concept: you look for SECISes, then upstream for selenoprotein coding sequences
We’re running a large scale project, we are confident to identify new selenoproteins in basal eukaryotes
Mariotti et al. 2013

26. Notes for the project
Results must be presented in a web page with the structure of a scientific paper
aminoacid sequences; SECIS sequences
genes in gff format (insert selenoDB form)
All genes should be as complete as possible: starting with a AUG, ending with a stop codon, and with an identified SECIS element downstream
Ignore alternative isoforms (if any), just choose one

27. Notes for the project
Report also the genes of selenoprotein machinery: SecS, eEFsec, pstk, SBP2, SPS1, (SPS2).
Ignore tRNAsec, for technical reasons

28. Common pitfalls Zero, one or many genes?
Careful with superfamilies, and gene duplications
Know what to expect
Genome assemblies are not perfect!

29. Evaluation - results - discussion - methods
The projects will be evaluated based on:
- results
you are expected to find all selenoproteins in your assembly
- discussion
interpret your results logically
- methods
scripting is encouraged (but not compulsory)
- presentation
the web page should present the work as clearly as possible