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

Structure of a selenoprotein gene

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

Selenoproteins are generally misannotated …protein DBs have the same limitations Percentages are computed by comparison Selenoprofiles-Ensembl annotations – see Mariotti and Guigó, 2010 - Bioinformatics.

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

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).

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)

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.

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

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

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)

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

UPF Biologia. Curs 2007-17 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

selenoproteins in vertebrates Project 2016-17 selenoproteins in vertebrates Objective: find all selenoproteins in a given vertebrate genome http://bioinformatica.upf.edu/

selenoproteins in vertebrates Project 2016-17 selenoproteins in vertebrates viquipedia: species description selenoDB: insert your selenoprotein genes predictions into a real world database. Available to the scientific community.

Insert genes form

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

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

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

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

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

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

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

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

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

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