Identification of Novel Mammalian Caspases Reveals an Important Role of GeneLoss in Shaping the Human Caspase Repertoire

Leopold Eckhart,* Claudia Ballaun,* Marcela Hermann,� John L. VandeBerg,� Wolfgang Sipos,§Aumaid Uthman,*k Heinz Fischer,* and Erwin Tschachler**Department of Dermatology, Medical University of Vienna, Vienna, Austria; �Max F. Perutz Laboratories, Medical University ofVienna, Vienna, Austria; �Department of Genetics, Southwest Foundation for Biomedical Research, San Antonio, TX; §Departmentfor Farm Animals and Herd Management, University of Veterinary Medicine Vienna, Vienna, Austria; and kCollege of VeterinaryMedicine, University of Sulaimani, Sulaimani, Iraq

Proteases of the caspase family play central roles in apoptosis and inflammation. Recently, we have described a new geneencoding caspase-15 that has been inactivated independently in different mammalian lineages. To determine thedynamics of gene duplication and loss in the entire caspase gene family, we performed a comprehensive evolutionaryanalysis of mammalian caspases. By comparative genomics and reverse transcriptase–polymerase chain reactionanalyses, we identified 3 novel mammalian caspase genes, which we tentatively named caspases-16 through -18.Caspase-16, which is most similar in sequence to caspase-14, has been conserved in marsupials and placental mammals,including humans. Caspase-17, which is most similar to caspase-3, has been conserved among fish, frog, chicken, lizard,and the platypus but is absent from marsupials and placental mammals. Caspase-18, which is most similar to caspase-8,has been conserved among chicken, platypus, and opossum but is absent from placental mammals. These genedistribution patterns suggest that, in the evolutionary lineage leading to humans, caspase-17 was lost after the split ofprotherian and therian mammals and caspase-18 was lost after the split of marsupials and placental mammals. In thecanine genome, the number of caspases has been reduced by the fusion of the neighboring genes caspases-1 and -4,resulting in a single coding region. Further lineage-specific gene inactivations were found for caspase-10 in murinerodents and caspase-12 in humans, rabbit, and cow. Lineage-specific gene duplications were found for caspases-1, -3,and -12 in opossum and caspase-4 in primates. Other caspases were generally conserved in all mammalian speciesinvestigated. Using the positions of introns as stable characters during recent vertebrate evolution, we define3 phylogenetic clades of caspase genes: caspases-1/-2/-4/-5/-9/-12/-14/-15/-16 (clade I), caspases-3/-6/-7/-17 (clade II),and caspases-8/-10/-18/CFLAR (clade III). We conclude that gene inactivations have occurred in each of the 3 caspaseclades and that gene loss has been as critical as gene duplication in the evolution of the human repertoire of caspases.

Introduction

Caspases are cysteine-dependent aspartate-directed pro-teases that have central roles in apoptosis, proinflammatorysignaling, immune cell proliferation, epidermal barrier for-mation, and various other physiological processes (Degterevet al. 2003; Denecker et al. 2007; Lamkanfi et al. 2007).Because of their critical functions and their potential astherapeutic targets (Reed 2002; Hotchkiss and Nicholson2006), caspases have attracted much interest and have beenwell characterized with respect to gene expression, struc-ture, activation, and catalytic activity (Earnshaw et al.1999; Degterev et al. 2003).

All caspases contain a peptidase C14 domain (caspasedomain, pfam00656), which is approximately 230 aminoacids long and comprises both alpha-helical segmentsand a central beta sheet that facilitates homodimerization(Fuentes-Prior and Salvesen 2004). Activation of procas-pase molecules involves proteolytic processing, therebygenerating 2 subunits from each caspase domain and re-moving the N-terminal prodomain. In humans, 14 proteinswith a peptidase C14 domain have been described: cas-pases-1 (originally named interleukin-1 beta-converting en-zyme), -2, -3, -4, -5, -6, -7, -8, -9, -10, -12, and -14; CLFAR(CASP8 and FADD-like apoptosis regulator, also known asc-FLIP); and the product of the mucosa-associated lym-phoid tissue lymphoma translocation gene 1. The latter

is not a true caspase but a paracaspase that deviates signif-icantly from the canonical caspase structure (Uren et al.2000). Proteolytic activity has been reported for all caspasesexcept for caspase-12, which is inactivated by deleteriousmutations in the majority of the human population (Fischeret al. 2002; Wang et al. 2006; Xue et al. 2006) and functionsas an inhibitor of inflammation in the mouse (Saleh et al.2006). Likewise, CFLAR inhibits caspase-8–dependent ap-optosis (Irmler et al. 1997). Inconsistent with the rules ofcaspase nomenclature (Alnemri et al. 1996), caspases-11and -13 represent the murine and bovine orthologs of hu-man caspase-4, respectively (Koenig et al. 2001). We haverecently cloned and characterized a new caspase, caspase-15, in a number of mammalian species (Eckhart et al. 2005;Eckhart et al. 2006).

Mammalian caspases have been grouped according todifferent criteria. Based on their functions, caspases-2, -8, -9, and -10 are apoptosis initiators, caspases-3, -6, and -7 areapoptosis executors, caspases-1, -4, and -5 are involved ininflammation, and caspase-14 has a role in terminal differ-entiation of epidermal keratinocytes (Fuentes-Prior andSalvesen 2004). Based on the length and structural organi-zation of the N-terminal prodomain, caspases-3, -6, -7, and-14 are short-prodomain proteases, whereas all others havea long prodomain folding into a caspase activation andrecruitment domain (CARD) (caspases-1, -2, -4, -5, -9,and -12), a tandem of 2 death effector domains (DEDs)(caspases-8 and -10 and CFLAR), or a pyrin domain(caspase-15). On the basis of different optimal cleavagesubstrate motifs, 3 caspase groups have been proposed:group I (caspases-1, -4, and -5), II (caspases-2, -3, and -7),and III (caspases-6, -8, and -9) (Thornberry et al. 1997),whereas a phylogenetic analysis based on amino acid

Key words: caspase, gene loss, gene duplication, comparativegenomics, mammals.

E-mail: leopold.eckhart@meduniwien.ac.at.

Mol. Biol. Evol. 25(5):831–841. 2008doi:10.1093/molbev/msn012Advance Access publication February 14, 2008

� The Author 2008. Published by Oxford University Press on behalf ofthe Society for Molecular Biology and Evolution. All rights reserved.For permissions, please e-mail: journals.permissions@oxfordjournals.org

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

sequence comparison suggested the existence of 3 ‘‘clus-ters’’: cluster I (caspases-1, -2, -4, -5, -12, and -14), II (cas-pases-3, -6, and -7), and III (caspases-8, -9, and -10)(Lamkanfi et al. 2002).

Comparative genomics has revealed that genes similarto caspases are present in alpha-proteobacteria, which arerelatives of the free-living ancestors of mitochondria (Kooninand Aravind 2002). This indicates that the ancestral caspasegenes of eukaryotes have been derived from the mitochon-drial endosymbiont. ‘‘Protocaspase’’ genes evolved intometacaspases in plants and fungi, paracaspases in slimemolds and animals, and genuine caspases in animals.The number of caspase genes per genome differs amonganimal lineages with 4, 5, 7, and 31 genes encoding caspasedomains being present in the nematode Caenorhabditiselegans, the cnidarian Nematostella vectensis, the fruitfly Drosophila melanogaster, and sea urchin Strongylocen-trotus purpuratus (Robertson et al. 2006), respectively. Asdescribed above, mammals contain more than 10 caspasegenes, most of which were cloned in the late 1990s. How-ever, our recent identification of caspase-15, a novel ancientmammalian caspase that was lost in primates and other evo-lutionary lineages (Eckhart et al. 2006), indicated that thecomplexity of caspase evolution in mammals had previ-ously been underestimated.

Here, we extended our comparative and evolutionaryanalysis to all mammalian caspase genes and to a largenumber of sequenced genomes. We report on the evolution-ary history of novel mammalian caspases and proposea new classification of caspases into 3 phylogenetic clades.Our study demonstrates that evolutionary dynamics of thecaspase gene content was strongly influenced by gene lossin mammals.

Materials and MethodsSequence Queries, Alignments, and PhylogeneticAnalyses

Blast searches for caspase homologs were performedusing query sequences such as the human caspases,CFLAR, human caspase pseudogenes (Rocher et al.1997; Centola et al. 1998), and caspases predicted by au-tomated algorithms of the ENSEMBL and GenBank data-bases in the genome sequences of amniotes. Sequenceswere retrieved from the GenBank database and the EN-SEMBL genome browser (http://www.ncbi.nlm.nih.gov/BLAST/; http://www.ensembl.org/Multi/blastview). Addi-tional nucleotide sequence alignments were made with theUniversity of California Santa Cruz genome browser(http://www.genome.ucsc.edu/) and with the programLALIGN (http://www.ch.embnet.org/software/LALIGN_form.html).

Gene orthology was evaluated according to the fol-lowing criteria: 1) Reciprocal best hits (also known as sym-metrical best hits) in Blast searches (Koonin 2005); 2)Features of gene organization that are stable in recent ver-tebrate evolution, namely the localization and phase of in-trons (Roy et al. 2003); 3) Syntenic position in the genome,that is, similarity in the arrangement of neighboring genes(Zheng et al. 2005).

Phylogenetic analyses were performed at the Phyle-mon Web server (Tarraga et al. 2007). Amino acid sequen-ces of caspase catalytic domains were aligned usingClustalW. Phylogenetic trees were built with PHYLIP usingthe Jones–Taylor–Thornton matrix as distance algorithm andthe Neighbor-Joining method for clustering. Trees were dis-played using the TreeView program (Page 1996).

DNA Preparation, Polymerase Chain Reactions andSequence Analysis

Genomic DNA was prepared from animal tissues ac-cording to a standard protocol (Ausubel 1998). TissueRNAs were extracted with the Trizol reagent (Invitrogen,Vienna, Austria) according to the manufacturer’s instruc-tions and reverse transcribed according to standard proto-cols (Eckhart et al. 1999). Polymerase chain reactions(PCRs) were performed according to a published protocolwith a primer annealing temperature of 55 �C (Eckhartet al. 1999). PCR products were either purified and se-quenced or cloned into the pCRII Topo vector (Invitrogen)according to the manufacturer’s instructions. Primersequences are shown in supplementary table S1 (Supple-mentary Material online).

ResultsIsolation of Caspase Genes from Databases andNomenclature of Novel Mammalian Caspase Genes

Genome sequences of diverse mammals, mostly usedas model species in biomedical research, as well as chickenand lizard (table 1) were searched for caspase homologs. Inthe course of these screenings, we identified novel caspasegenes for which we propose the following taxonomy. 1)Caspases that evolved by gene duplication in evolutionarylineages not leading to humans are termed caspase-x–like y,where x is the number of the orthologous human caspaseand y is a consecutive number ranging from 1 to the max-imum number of homologs arisen by duplication (fig. 1A).2) Caspases that are either still present in the human ge-nome or which have been present in the genome of a mam-malian ancestor of humans are assigned a new numberbecause they are paralogous to all (other) human caspases(fig. 1B). We consider the criterion of presence in a mam-malian ancestor of humans met if an ortholog of such a cas-pase is present in a mammalian sister group of humans andin a (mammalian or nonmammalian) sister group of thecommon ancestor of humans and the first, evolutionarilymore close sister group (fig. 1). An example for this taxon-omy is caspase-15, which is absent in humans but whoseformer presence in an ancestor of humans can be inferredfrom the detection of caspase-15 in pig (representing sistergroup 1) and in marsupial opossum (representing sistergroup 2) (Eckhart et al. 2006). The proposed caspase tax-onomy extends the original guidelines of caspase nomen-clature (Alnemri et al. 1996) by considering nonhumancaspases while ensuring phylogenetic relevance for humanphysiology.

832 Eckhart et al.

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

Identification of Caspase-16, an EvolutionarilyConserved Mammalian Caspase Gene Most Similar toCaspase-14

Comparative genomics of all caspase-like genes in thehuman genome showed that a transcribed gene previouslyconsidered to be a pseudogene (Centola et al. 1998) wasconserved among mammals. The chromosomal locus ofthis gene was syntenic in human (chromosome 16p13.3),opossum (fig. 2A), and other mammals (data not shown).Because the gene encoded a caspase domain and the cys-teine and histidine residues forming the catalytic center ofcaspases were conserved (fig. 2B), it appeared to representa novel mammalian member of the caspase family. In ac-cordance with the proposed guidelines of caspase nomen-clature, we refer to the novel gene as caspase-16.

Caspase-16 orthologs were identified in all mamma-lian genome sequences (including platypus and opossum)available in September 2007, whereas no orthologs werefound in nonmammalian genomes including the closest out-group to mammals, that is, sauropsids, represented bychicken (Gallus gallus) and the lizard Anolis carolinensis.This suggested an evolutionary origin of caspase-16 afterthe divergence of the sauropsids (including reptiles andbirds) and synapsids (including mammals) and prior tothe split of protherians (monotremes) and therians (marsu-pials and placental mammals) approximately 220 MYA(van Rheede et al. 2006). Caspase-16 corresponds toa so-called caspase-14–like gene identified in the rat ge-nome (Puente and Lopez-Otin 2004). Homology of human,chimpanzee, dog, mouse, and rat caspase-16 genes has alsobeen detected by the automated system that feeds the Na-tional Center for Biotechnology Information HomoloGenedatabase (Wheeler et al. 2007), and caspase-16 gene–related information is available at the HomoloGene data-base entry 72522 (http://www.ncbi.nlm.nih.gov/sites/entrez/query.fcgi?db5homologene). Analyses of the genomesequences listed in table 1 showed that the open reading frameof the caspase domain of caspase-16 was intact in all speciesexcept guinea pig, which contained in-frame stop codons.

Caspase-16 contains a unique insertion of 25 aminoacids at a position corresponding to the L3 loop of thecaspase domain (Fuentes-Prior and Salvesen 2004). Ex-cluding this segment from the alignment, the caspase-16peptidase C14 domain displays the highest amino acid se-quence similarity with caspase-14 (45% identity) and cas-pase-15 (29% identity). The prodomain of caspase-16 lackssignificant similarity to other proteins. However, the struc-ture of the N-terminus of the human caspase-16 protein re-quires further characterization because of differences in thecorresponding region of the mRNA predicted by the GNO-MON algorithm of the GenBank and the caspase-16cDNAs reported by Centola et al. (1998).

We analyzed the expression of caspase-16 by reversetranscriptase–polymerase chain reaction (RT-PCR) andnested PCR and found that caspase-16 was expressed in hu-man spleen (supplementaryfig. S1A, SupplementaryMaterialonline), confirming Northern blot data published by Centolaet al. (1998). Unlike its closest genetic relative, that is, cas-pase-14 (Eckhart et al. 2000), caspase-16 was not expressedin differentiated epidermal keratinocytes (supplementary fig.S1A, Supplementary Material online). In a separate experi-ment, expression of caspase-16 was also detected in the liverof the opossum (GenBank accession number: EF397305)(supplementary fig. S1B, Supplementary Material online).Further studies of caspase-16 expression at the mRNA andprotein levels as well as the investigation of its catalytic po-tential as a protease have been initiated in our laboratories.

Identification of Caspase-17, a Novel Caspase Similar toCaspase-3, in the Platypus and NonmammalianVertebrates

The genome sequence of the platypus, a protherianmammal, contained sequences homologous to 2 exons ofan uncharacterized caspase gene present in the genomeof chicken (supplementary fig. S2, Supplementary Materialonline). Expression of the novel gene, which we tentativelynamed caspase-17, was detected by RT-PCR in chickenoviduct and liver (supplementary fig. S2, Supplementary

Table 1Species and Databases Used for Comparative Genomic Analyses

Clade Species Database Genome assembly

Teleostei Pufferfish (Tetraodon nigroviridis) http://www.ensembl.org/Tetraodon_nigroviridis/index.html TETRAODON 7Amphibia Frog (Xenopus tropicalis) http://www.ensembl.org/Xenopus_tropicalis/index.html JGI 4.1Aves Chicken (Gallus gallus) http://www.ensembl.org/Gallus_gallus/index.html WASHUC2Lepidosauria Green anole (Anolis carolinensis) http://www.genome.ucsc.edu/cgi-bin/hgGateway?hgsid5

104371369&clade5vertebrate&org5Lizard&db50anoCar1

Monotremata Platypus (Ornithorhynchus anatinus) http://www.ensembl.org/Ornithorhynchus_anatinus/index.html Oana-5.0Marsupialia Opossum (Monodelphis domestica) http://www.ensembl.org/Monodelphis_domestica/index.html monDom5Afrotheria Elephant (Loxodonta africana) http://www.ensembl.org/Loxodonta_africana/index.html BROAD E1Laurasiatheria Cow (Bos taurus) http://www.ensembl.org/Bos_taurus/index.html Btau_3.1

Dog (Canis familiaris) http://www.ensembl.org/Canis_familiaris/index.html CanFam 2.0Euarchontoglires Mouse (Mus musculus) http://www.ncbi.nlm.nih.gov/projects/genome/guide/mouse/ Build 37.1

Rat (Rattus norvegicus) http://www.ncbi.nlm.nih.gov/projects/genome/guide/rat/ RGSC 3.4Guinea pig (Cavia porcellus) http://www.ensembl.org/Cavia_porcellus/index.html cavPor2Rabbit (Oryctolagus cuniculus) http://www.ensembl.org/Oryctolagus_cuniculus/index.html RABBITRhesus monkey (Macaca mulatta) http://www.ensembl.org/Macaca_mulatta/index.html MMUL 1.0Chimpanzee (Pan troglodytes) http://www.ensembl.org/Pan_troglodytes/index.html PanTro 2.1Humans (Homo sapiens) http://www.ncbi.nlm.nih.gov/projects/genome/guide/human/ Build 36.2

Evolution of Mammalian Caspases 833

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

Material online) (GenBank accession numbers EU165361and EU165362). Orthologs of caspase-17 were also foundin the genome of the lizard A. carolinensis, the Westernclawed frog Xenopus tropicalis, and the fish species Daniorerio and Tetraodon nigroviridis (fig. 3) but not in opossumand placental mammals. Expressed sequence tags corre-sponding to caspase-17 of fugu (Takifugu rubripes), stick-leback (Gasterosteus aculeatus), medaka (Oryzias latipes),and the clawed frogs X. tropicalis and Xenopus laevis wereidentified in the GenBank (supplementary fig. S2, Supple-mentary Material online). This species distribution patternsuggested that caspase-17 originated prior to the divergenceof the lineages leading to bony fish and tetrapods, respec-tively (450 MYA) (Hedges 2002), and was retained in allvertebrate lineages investigated except for therian mam-mals (marsupials and placental mammals). We concludethat, in the evolutionary lineage leading to humans, the cas-pase-17 gene was inactivated between 220 MYA (diver-

gence of protherian and therian mammals) and 190MYA (divergence of marsupials and placental mammals)(van Rheede et al. 2006).

Phylogenetic analysis supported the notion that cas-pase-17 is a distinct caspase, nonorthologous to the geneswith which it shows highest sequence similarity, that is,caspases-3, -6, and -7 (fig. 3). Both the chicken and the frogcaspase-17 genes were located between the genes RETSAT(retinol saturase) and TGOLN2 (trans-golgi network pro-tein 2), suggesting that this is the ancestral locus of cas-pase-17 (fig. 3). The corresponding chromosomal locusof the dog and the human genome (human chromosome2p11.2) did not contain sequences similar to caspase-17(fig. 3).

Caspase-17 contained the histidine and cysteine resi-dues forming the catalytic site of caspases (fig. 3B, blackshading), indicating that it is enzymatically active. How-ever, all caspase-17 orthologs identified in this study lackeda tyrosine residue (corresponding to caspase-6 Y37, fig. 3B)that is conserved among other caspases (supplementary fig.S4, Supplementary Material online). In addition, an argi-nine residue (corresponding to caspase-6 R220, fig. 3B),implicated in substrate binding of caspases (Fuentes-Priorand Salvesen 2004), was conserved in caspase-17 of theplatypus but not in other caspase-17 orthologs. Together,these features indicate that the substrate specificity and/or other enzymatic properties of caspase-17 differ from pre-viously characterized caspases.

Identification of Caspase-18, a Caspase-8–Like Gene inChicken, Platypus, and Opossum

Caspases-8 and -10 and CFLAR are arranged in closeproximity at a chromosomal locus (human chromosome2q33–q34), henceforth referred to as the caspase-8 subfam-ily locus. Comparison of this locus in the genomes of hu-man, mouse, opossum, and chicken showed that theoutermost genes, CFLAR and CASP8, were conservedand always represented the borders of this locus. In contrastto the other 3 species, the mouse lacked a caspase-10 gene,as has been reported previously (Janicke et al. 2006). Sur-prisingly, a previously undescribed gene was found in bothopossum and chicken between the CASP10 and CASP8genes (fig. 4A). This novel gene encoded a protein withthe same structure as that of caspase-8. It comprised 2N-terminal DEDs and a caspase domain in which all aminoacid residues critical for catalytic activity were conserved(fig. 4B). We have therefore tentatively termed the novelgene caspase-18. In a phylogenetic analysis, chicken andopossum caspase-18 genes clustered together to the exclu-sion of CFLAR and caspases-8 and -10 of chicken, opos-sum, and humans (fig. 4C). An uncharacterized cDNAcorresponding to chicken caspase-18, but originally termed‘‘initiator caspase-8/-10’’ (gi 67097571; submitted byKazuhiro Sakamaki and Masami Nozaki, Kyoto University,Kyoto, Japan) was identified in the GenBank, whichshowed that this gene is expressed. We confirmed expres-sion of caspase-18 in the liver of opossum and chicken, re-spectively, by RT-PCR (GenBank accession numbersEU165363–EU165365) (supplementary fig. S3, Supple-mentary Material online).

FIG. 1.—Nomenclature of mammalian caspases. (A) We propose touse the term caspase-x–like genes for caspase genes that have arisen bygene duplications in a lineage not leading to humans. The combinedspecies (thick-lined frames) and gene (thin lines) tree shows an examplein which the caspase-x gene has been duplicated in sister group 1 ofhumans. Both copies are orthologs of human caspase-x. (B) By contrast,caspase genes that have been present in evolutionary ancestors of humansand have been retained in at least one mammalian sister group of humansshould be assigned a new name, that is, caspase-n þ 1, where n is thehighest number of a caspase identified previously. This is irrespective ofwhether the caspase-n þ 1 gene has been maintained or inactivated(indicated by a strike symbol) in the human lineage.

834 Eckhart et al.

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

To determine the evolutionary history of caspase-18,we analyzed, in addition to the species mentioned above,Western clawed frog (X. tropicalis), platypus, armadillo,elephant, dog, and cow. A putative ortholog of caspase-18 was identified in the genome of the platypus but notin the genomes of the other species. This distribution sug-gested that caspase-18 originated, presumably by duplica-tion of caspase-8 or -10, after the split of the amphibianlineage from amniotes, was further conserved in birdsand protherian as well as metatherian mammals, but wasdeleted in a common ancestor of all placental (eutherian)mammals.

Exon–Intron Organization Defines 3 PhylogeneticClades of Caspases

We reevaluated the phylogenetic relationships ofmammalian caspases by defining phylogenetic clades ofcaspase genes using a character that is considered stable

over the evolutionary period investigated, namely, theposition of introns within the coding sequence (Royet al. 2003). Amino acid sequence alignment and mappingof intron positions within the coding regions of mammaliancaspases showed that there are 3 main, distinct patterns ofexon–intron structure (fig. 5; supplementary fig. S4, Sup-plementary Material online). The genes encoding cas-pases-1, -2, -4, -5, -9, -12, -14, -15, and -16 (clade I)share 4 common intron positions within the region codingfor the caspase domain, whereas caspases-3, -6, -7, and -17(clade II) as well as caspases-8, -10, and -18 and CFLAR(clade III) share 3 common intron positions, respectively(fig. 5). One intron position is shared between clade Iand II caspases. In addition, caspases-1, -4, -5, and -12share homologous introns flanking the CARD sequencein the prodomain (not shown), at the 5# end of the caspasedomain region and in the 3# untranslated region. This com-monality supports the notion that these caspases, which arealso known as caspase-1 subfamily, form a subclade ofclade I. Caspases-2 and -9 share a homologous intron in

FIG. 2.—Caspase-16 is a phylogenetically ancient gene and has been conserved among mammals. (A) Schematic depiction of the genomicenvironment of the caspase-16 gene (CASP16) in humans and opossum. ZNF213, zinc finger protein 213; MEFV, Mediterranean fever gene. (B) Aminoacid sequence alignment of the caspase domain of caspase-16 (C16) orthologs; hum, human; mou, mouse; and opo, opossum. The correspondingsequence of human caspase-14 (C14) is shown for comparison. Black shading indicates amino acid residues forming the catalytic site. Sequencesimilarity is indicated by the following symbols below the alignment using the symbols ‘‘asterisk,’’ ‘‘colon,’’ and ‘‘dot’’ for identical, highly similar, andweakly similar residues, respectively.

Evolution of Mammalian Caspases 835

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

the region encoding the CARD motif of the prodomain(data not shown), supporting their close relationship, whichis also suggested by common roles in the initiation phase ofapoptosis.

The 3 phylogenetic clades of caspases, defined byshared intron positions within the caspase domain regionsof the genes, have not been predicted by the phylogeneticanalysis of amino acid sequences (Lamkanfi et al. 2002) butcorrelate well, though not perfectly, with the grouping ofcaspases according to their prodomain structure. All cladeIII caspases contain a tandem of 2 DEDs in their prodomain,and all clade II caspases have a short (caspases-3, -6, and -7)or no prodomain (caspase-17). The predominant prodomainmotif among clade I caspases is CARD. Absence ofa CARD in caspases-14, -15, and -16, which also clusterin phylogenetic analyses using the amino acid sequenceof the caspase domain only, has probably been causedby loss of an ancestral CARD. It exceeded the scope of this

study to assess which of the 3 caspase clades is phyloge-netically closest to the primordial caspase of animals. How-ever, a preliminary comparative genomics screening ofmore primitive metazoan species revealed that caspaseswith prodomains containing a CARD or 2 DEDs as wellascaspaseswithshort prodomains areencoded in thegenomeof the sea anemone N. vectensis (Eckhart L, unpublisheddata), which represents an evolutionary lineage that divergedfrom the bilaterian lineage approximately 600 MYA.

Evolutionary History of Caspase Genes in Mammals

Next we assessed the evolutionary history of each in-dividual caspase in multiple mammalian lineages. Becauseof limitations in the availability of assembled genome se-quences, which are required for an in-depth comparativegene analysis, and in order to facilitate follow-up studiesin accessible animal models, we focused on a series of

FIG. 3.—Caspase-17 is a phylogenetically ancient gene that has been lost in an ancestor of therian mammals. (A) Schematic depiction of thegenomic environment of the caspase-17 gene (CASP17) in frog and chicken and of the homologous locus of dog and humans. RETSAT, retinolsaturase; TGOLN2, trans-golgi network protein 2. The human gene locus contains 2 pseudogenes, LOC647305, which is similar to RETSAT, andLOC647307, which is similar to phosphatidylethanolamine-binding protein. (B) Amino acid sequence alignment of caspase-17 (C17) orthologs. Thealigned sequences of human caspases-3 and -6 (C3 and C6) are shown for comparison. Black shading indicates amino acid residues forming thecatalytic site. The platypus caspase-17 sequence is incomplete, and unknown amino acid residues are substituted by X; dan, Danio rerio; tet, Tetraodonnigroviridis; xen, Xenopus tropicalis; ano, Anolis carolinensis; chi, chicken; pla, platypus; and hum, human. (C) A phylogenetic tree of caspases-3, -6, -7,and -17 was built using the Neighbor-Joining method. The branch lengths in the tree are proportional to the number of substitutions per site (scale 0.1substitution per site).

836 Eckhart et al.

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

mammalian model species (opossum, cow, dog, rabbit,guinea pig, rat, mouse, macaque, and chimpanzee). Blastsearches in the respective databases and RT-PCR analyseson mRNAs from various animals were performed. Cas-pases-2, -6, -7, -8, -9, and -14 and CFLAR were conservedwithout deletion or duplication events in all mammalianspecies investigated, whereas interspecies differences werefound for the other caspases. Besides the gene losses of cas-pases-17 and -18 described above, caspase-10 was lost inmurine rodents but not in guinea pig and rabbit (GenBankaccession number EF397306) and caspase-3 was duplicatedin the opossum (supplementary fig. S5, SupplementaryMaterial online).

The most striking interspecies differences were foundat the locus of the so-called caspase-1 subfamily genes (cas-pases-1, -4, -5, and -12), also known as the subfamily ofinflammatory caspases (Martinon and Tschopp 2004),which are clustered on human chromosome 11q23. Com-parison of this chromosomal locus among mammals sup-ported previous studies that have proposed a presumably

primate-specific gene duplication leading to caspases-4and -5 (Lamkanfi et al. 2002) and gene inactivation of cas-pase-12 in humans (Fischer et al. 2002). In addition, a geneduplication giving rise to a caspase-4–like transcribed pseu-dogene was identified in humans and other primates (fig. 6A)and inactivation of caspase-12 was also detected in rabbitand cow (supplementary fig. S6, Supplementary Materialonline). The opossum genome contains lineage-specificgene duplicates of caspases-1 and -12 (fig. 6A). By contrast,the dog genome contains only 2 genes at the caspase-1 fam-ily locus, namely, caspase-12 and a gene with maximumsequence similarity to human caspases-1 and -4. The cas-pase-1– and caspase-4–like gene, which we now name ‘‘hy-brid caspase-1/caspase-4’’ or short, ‘‘hybrid caspase-1/-4,’’is most similar to caspase-1 in its first 3 exons and mostsimilar to caspase-4 in its remaining exons (fig. 6B). RT-PCR analysis of dog peripheral blood mononuclear cellsdemonstrated the existence of 2 alternatively spliced tran-scripts, dog hybrid caspase-1/-4/a and hybrid caspase-1/-4/b (GenBank accession number EU183118), which differed

FIG. 4.—Caspase-18 is a gene of the caspase-8 gene cluster of chicken, opossum, and platypus. (A) Schematic depiction of the genomicenvironment of the caspase-18 gene (CASP18) in chicken, opossum, human, and mouse. Note that CASP18 has been lost in humans and mouse andthat CASP10 has been lost in mouse. (B) Amino acid sequence alignment of caspase-18 (C18) orthologs. The aligned sequences of human caspases-8and -10 (C8 and C10) are shown for comparison. Black shading indicates amino acid residues forming the catalytic site. The position of DEDs isindicated by thin lines, and the caspase domain is indicated by a thick line above the sequence. (C) A phylogenetic tree of caspases-8 (C8), -10 (C10),and -18 (C18) and CFLAR of the species chicken, opossum, and human was built using the Neighbor-Joining method. The branch lengths in the treeare proportional to the number of substitutions per site (scale: 0.1 substitutions per site); chi, chicken; opo, opossum; and hum, human.

Evolution of Mammalian Caspases 837

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

by the facultative inclusion of exon 4 (fig. 6B). The shortervariant was identical to a cDNA, which prior to our reportthat caspase-13 is the bovine ortholog of human caspase-4(Koenig et al. 2001) had been named ‘‘hybrid caspase-1/caspase-13’’ (Taylor et al. 2000). From the presence of cas-pases-1 and -4 in humans, mouse, and cow, it can be in-ferred that both genes were present in the genome of thelast common ancestor of Boreoeutheria (Murphy et al.2001), a clade that includes these 3 species as well asthe dog. Therefore, we propose that the gene structure ofthe dog hybrid caspase-1/-4 gene has emerged by fusionof the ancestral caspases-1 and -4 genes via deletion ofthe chromosomal segment containing exons 4–11 of cas-pase-1 and exon 1 of caspase-4. This rearrangement musthave occurred early in the evolution of carnivores becausea hybrid caspase-1/-4 gene is also present in the cat genome(GenBank accession number EU183117). Notably, the pre-dicted translation product of caspase-1/-4/a represents thefirst caspase with 2 consecutive CARDs in the prodomain.

By utilizing the new assembly of the anole genome,we could extend our recent investigation of the evolutionof caspase-15. Regions homologous to mammalian cas-pase-15 exons 3–7 were identified in the anole, whereasexons 1–2 (encoding the prodomain) escaped detectionbecause of a gap in the genome sequence. In the anolis ge-nome assembly of the University of California Santa Cruzgenome browser, the caspase-15–like gene is in chromo-somal vicinity to an ortholog of FLJ20551, a gene thatflanks the mammalian caspase-15 gene (Eckhart et al.2006). Based on the new data, we propose that the evolu-tionary origin of caspase-15 preceded the divergence of

sauropsids (including reptiles and birds) and synapsids (in-cluding mammals) and that this gene was not only lost inseveral mammalian lineages (Eckhart et al. 2006) but alsoin chicken. Together with the absence of caspases-14 and-16 in nonmammalian species, this pattern indicates thatcaspase-15 originated first among these 3 caspases and thatduplication of caspase-15, followed by gene copy–specificsequence modification, gave rise to caspases-14 and -16 inthe evolutionary lineage leading to mammals.

Having identified commonalities and differences inthe caspase gene complement in various mammals as wellas chicken and anole, we could map the presence and ab-sence of individual caspases onto the known phylogenetictree of vertebrates (fig. 7). This allowed us to infer caspasegene birth and death events during mammalian evolution.The results are shown in the upper panel of figure 7.

Discussion

The results of our comparative genomic analysis implythat the number of caspase gene deletions (caspases-17, -18,-15, and -12, ranked in the chronological order of inactiva-tion) almost equaled the number of caspase gene innova-tions by duplication (caspases-12, -14, -16, -4, and -5) inthe evolutionary lineage leading to humans. Furthermore,we found novel examples of caspase gene deletions that oc-curred in parallel in different lineages: loss of caspase-15 inchicken (this study) and some mammals (Eckhart et al.2006) and deletions of caspase-12 in cow and rabbit (thisstudy) and humans (Fischer et al. 2002) as well as caspase-18 in anole and the last common ancestor of placental mam-mals (this study). We conclude that gene deletion shouldnot generally be considered a rare event during evolution(Rokas and Holland 2000), and great care should be takenwhen phylogenetic relationships among species are inferredfrom common presence or absence of individual genes.However, the use of near-complete genome data from manyspecies allows concomittant analyses of multiple genes thatare likely to give phylogenetic results with high confidence.

The differences in caspase gene content in mammalsmay reflect adaptations to species-specific constraints or re-dundant functions of caspases. In the recent evolution ofhumans, caspase-12 (Fischer et al. 2002) has been underpositive selection for mutational inactivation (Wanget al. 2006; Xue et al. 2006). Whether positive selectionhas also driven inactivation of other caspase genes, suchas caspases-15, -17, and -18, or duplication of caspasegenes, such as caspase-3 in the opossum and caspase-4in primates, remains to be determined. In our view, anyhypotheses regarding the evolutionary roles of these genereorganization events require the understanding of the bi-ological functions of the respective caspase in at least onespecies. All murine caspase genes except for caspase-16 havebeen investigated by targeted knockout. Mice deficient incaspases-1, -2, -6, -7, -11, -12, and -14 have relatively mildphenotypes under standard laboratory conditions, whereasinactivation of caspases-3, 8, or, 9 is lethal (Kuida et al.1996; Hakem et al. 1998; Varfolomeev et al. 1998). Thisindicates that most caspases have functions not critical fordevelopment and survival under nonstressed conditions

FIG. 5.—Exon–intron structure defines phylogenetic relationship ofcaspase genes. Schematic overview of the caspase domains ofmammalian caspase genes. The position of introns is shown by triangles.Vertical lines indicate homologous intron positions. An intron, markedwith an asterisk, of caspase-16 is interrupted by a gene-specific exon.Caspase-5 is not included in this scheme because it represents a primate-specific duplication product of the caspase-4 gene and because the exonsencoding its caspase domain have the same organization as the caspase-4exons. The exon–intron structure of caspase-4 is also identical to those ofcaspases-11 and -13, which are actually the murine and bovine orthologsof human caspase-4. An alignment of caspase sequences with exactpositions and phases of exon borders is shown in the SupplementaryMaterial online. CARD, caspase activation and recruitment domain;DED2, death effector domains in tandem arrangement; and PYD, pyrindomain.

838 Eckhart et al.

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

but, most likely, are essential in distinct states of infection (Liet al. 1995) or exposure to noxious environmental challengessuch as UV irradiation (Denecker et al. 2007). Becausecaspases-10, -15, -17, and -18 have been lost in the evolu-tionary lineage leading to the mouse, gene knockout orknockdown approaches must be established in alternativemodel species or cultured cells derived from species express-ing these caspases.

Notably, we found the highest interspecies variabilityamong members of the caspase-1 subfamily, which is in-volved in inflammation and host defense against pathogens(Martinon and Tschopp 2004). The observed variability in-cluded gene duplications, gene inactivations, and gene fu-sions. The fusion of caspases-1 and -4 in carnivores is ofparticular interest as it has led to the loss of the catalytic do-main of caspase-1. This suggests that the catalytic reactionsmediated by caspase-1 in humans are catalyzed by either thecaspase-4–derived caspase domain of hybrid caspase-1/-4 orby caspase-12 in the dog. A more comprehensive evaluationof the interdependency of caspase gene functions and pat-terns of gene conservation/amplification should include thecomparison with other gene families at a multigenome level.

The results of the present study suggest that one of thecentral gene families involved in apoptosis, the caspases,

has undergone extensive changes during mammalian evo-lution with approximately one-third of genes being deletedand a similar fraction of genes being duplicated. Our datashow that the apoptotic machinery has not constantly in-creased in complexity during evolution in the vertebrate lin-eage as it was extrapolated from a comparative genomicanalysis of model organisms such as nematode, fly, andmouse (Aravind et al. 2001) but, at least in mammals,has undergone both expansions and reductions in genenumbers during recent evolution.

Supplementary Material

Supplementary table S1 and figures S1–S6 are avail-able at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).

Acknowledgments

We thank Ulrich Koenig and Wolfgang J. Schneiderfor helpful discussions and Maria Buchberger, SonjaGawlas, Christian Kittel (Medical University of Vienna,Austria), and Jeannie Chan (Southwest Foundation for

FIG. 6.—Comparison of the caspase-1 gene subfamily locus in humans and nonhuman mammals. (A) Schematic overview of the caspase-1subfamily gene locus in mammals. Genes are represented by arrows. Note that cDNAs of the caspase-1 subfamily genes have not been cloned in theopossum and that orthology relationships of these genes should be considered preliminary. Asterisks indicate the presence of mutations that prevent theexpression of a catalytically active caspase. (B) The caspases-1 and -4 genes have fused in the dog. Exons are represented by boxes in which the exonnumber is shown.

Evolution of Mammalian Caspases 839

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

Biomedical Research, San Antonio, TX) for excellent tech-nical support. The authors acknowledge all providers ofpublic genome sequences, in particular the Broad Institute,Boston, MA, and the Genome Sequencing Center at theWashington University School of Medicine, St Louis,MO. We thank Heidemarie Rossiter for critically readingthe manuscript.

Literature Cited

Alnemri ES, Livingston DJ, Nicholson DW, Salvesen G,Thornberry NA, Wong WW, Yuan J. 1996. Human ICE/CED-3 protease nomenclature. Cell. 87:171.

Aravind L, Dixit VM, Koonin EV. 2001. Apoptotic molec-ular machinery: vastly increased complexity in vertebratesrevealed by genome comparisons. Science. 291:1279–1284.

Centola M, Chen X, Sood R, et al. (29 co-authors). 1998.Construction of an approximately 700-kb transcript maparound the familial Mediterranean fever locus on humanchromosome 16p13.3. Genome Res. 8:1172–1191.

Degterev A, Boyce M, Yuan J. 2003. A decade of caspases.Oncogene. 22:8543–8567.

Denecker G, Hoste E, Gilbert B, et al. (18 co-authors). 2007.Caspase-14 protects against epidermal UVB photodamageand water loss. Nat Cell Biol. 9:666–674.

Earnshaw WC, Martins LM, Kaufmann SH. 1999. Mammaliancaspases: structure, activation, substrates, and functionsduring apoptosis. Annu Rev Biochem. 68:383–424.

Eckhart L, Ballaun C, Uthman A, Kittel C, Stichenwirth M,Buchberger M, Fischer H, Sipos W, Tschachler E. 2005.Identification and characterization of a novel mammaliancaspase with proapoptotic activity. J Biol Chem. 280:35077–35080.

Eckhart L, Ban J, Ballaun C, Weninger W, Tschachler E. 1999.Reverse transcription-polymerase chain reaction products ofalternatively spliced mRNAs form DNA heteroduplexes andheteroduplex complexes. J Biol Chem. 274:2613–2615.

Eckhart L, Declercq W, Ban J, Rendl M, Lengauer B, Mayer C,Lippens S, Vandenabeele P, Tschachler E. 2000. Terminaldifferentiation of human keratinocytes and stratum corneumformation is associated with caspase-14 activation. J InvestDermatol. 115:1148–1151.

Eckhart L, Uthman A, Sipos W, Tschachler E. 2006. Genomesequence comparison reveals independent inactivation of thecaspase-15 gene in different evolutionary lineages ofmammals. Mol Biol Evol. 23:2081–2089.

FIG. 7.—Distribution of caspase genes in mammalian species. The presence or absence of caspase genes in mammalian species and, forcomparison, in chicken and anole is indicated by ‘‘plus’’ and ‘‘minus’’ symbols. ‘‘Double plus’’ and ‘‘triple plus’’ indicate lineage-specific geneduplications. Plus/minus f indicates partial deletion and fusion of the Casp-1 and -4 genes in the dog. Above the table, the phylogenetic tree of thespecies is shown. Mutations affecting the repertoire of caspases are mapped onto the tree. On the left side of the table, the grouping of caspases in cladesI, II, and III is indicated; dupl, duplication; del, deletion; mut, point mutations; C, caspase; fus, fusion; platy, platypus; opo, opossum; rabb, rabbit; guin,guinea pig; mou, mouse; mac, rhesus macaque; chim, chimpanzee; and hum, humans.

840 Eckhart et al.

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018

Fischer H, Koenig U, Eckhart L, Tschachler E. 2002. Humancaspase 12 has acquired deleterious mutations. BiochemBiophys Res Commun. 293:722–726.

Fuentes-Prior P, Salvesen GS. 2004. The protein structures thatshape caspase activity, specificity, activation and inhibition.Biochem J. 384:201–232.

Hakem R, Hakem A, Duncan GS, et al. (16 co-authors). 1998.Differential requirement for caspase 9 in apoptotic pathwaysin vivo. Cell. 94:339–352.

Hedges SB. 2002. The origin and evolution of model organisms.Nat Rev Genet. 3:838–849.

Hotchkiss RS, Nicholson DW. 2006. Apoptosis and caspasesregulate death and inflammation in sepsis. Nat Rev Immunol.6:813–822.

Irmler M, Thome M, Hahne M, et al. (13 co-authors). 1997.Inhibition of death receptor signals by cellular FLIP. Nature.388:190–195.

Janicke RU, Sohn D, Totzke G, Schulze-Osthoff K. 2006.Caspase-10 in mouse or not? Science. 312:1874.

Koenig U, Eckhart L, Tschachler E. 2001. Evidence that caspase-13 is not a human but a bovine gene. Biochem Biophys ResCommun. 285:1150–1154.

Koonin EV. 2005. Orthologs, paralogs, and evolutionarygenomics. Annu Rev Genet. 39:309–338.

Koonin EV, Aravind L. 2002. Origin and evolution of eukaryoticapoptosis: the bacterial connection. Cell Death Differ. 9:394–404.

Kuida K, Zheng TS, Na S, Kuan C, Yang D, Karasuyama H,Rakic P, Flavell RA. 1996. Decreased apoptosis in the brainand premature lethality in CPP32-deficient mice. Nature.384:368–372.

Lamkanfi M, Declercq W, Kalai M, Saelens X, Vandenabeele P.2002. Alice in caspase land. A phylogenetic analysis ofcaspases from worm to man. Cell Death Differ. 9:358–361.

Lamkanfi M, Festjens N, Declercq W, Vanden Berghe T,Vandenabeele P. 2007. Caspases in cell survival, proliferationand differentiation. Cell Death Differ. 14:44–55.

Li P, Allen H, Banerjee S, et al. (17 co-authors). 1995. Micedeficient in IL-1 beta-converting enzyme are defective inproduction of mature IL-1 beta and resistant to endotoxicshock. Cell. 80:401–411.

Martinon F, Tschopp J. 2004. Inflammatory caspases: linking anintracellular innate immune system to autoinflammatorydiseases. Cell. 117:561–574.

Murphy WJ, Eizirik E, O’Brien SJ, et al. (11 co-authors). 2001.Resolution of the early placental mammal radiation usingBayesian phylogenetics. Science. 294:2348–2351.

Page RD. 1996. TreeView: an application to display phylogenetictrees on personal computers. Comput Appl Biosci.12:357–358.

Puente XS, Lopez-Otin C. 2004. A genomic analysis of ratproteases and protease inhibitors. Genome Res. 14:609–622.

Reed JC. 2002. Apoptosis-based therapies. Nat Rev DrugDiscov. 1:111–121.

Robertson AJ, Croce J, Carbonneau S, Voronina E, Miranda E,McClay DR, Coffman JA. 2006. The genomic underpinningsof apoptosis in Strongylocentrotus purpuratus. Dev Biol.300:321–334.

Rocher C, Faucheu C, Blanchet AM, Claudon M, Herve F,Durand L, Harnois M, Diu-Hercend A, Lalanne JL. 1997.Identification of five new genes, closely related to the

interleukin-1beta converting enzyme gene, that do not encodefunctional proteases. Eur J Biochem. 246:394–400.

Rokas A, Holland PW. 2000. Rare genomic changes as a tool forphylogenetics. Trends Ecol Evol. 15:454–459.

Roy SW, Fedorov A, Gilbert W. 2003. Large-scale comparisonof intron positions in mammalian genes shows intron loss butno gain. Proc Natl Acad Sci USA. 100:7158–7162.

Saleh M, Mathison JC, Wolinski MK, Bensinger SJ, Fitzgerald P,Droin N, Ulevitch RJ, Green DR, Nicholson DW. 2006.Enhanced bacterial clearance and sepsis resistance in caspase-12-deficient mice. Nature. 440:1064–1068.

Strauss WM. 1998. Preparation of genomic DNA frommammalian tissue. In: Ausubel FM, Brent R, Kingston RE,Moore DD, Seidman JG, Smith JA, Struhl K, editors. Currentprotocols in molecular biology. New York: John Wiley &Sons. p. 2.

Taylor S, Hanlon L, McGillivray C, Gault EA, Argyle DJ,Onions DE, Nicolson L. 2000. Cloning and sequencing offeline and canine ice-related cDNAs encoding hybrid caspase-1/caspase-13-like propeptides. DNA Seq. 10:387–394.

Tarraga J, Medina I, Arbiza L, Huerta-Cepas J, Gabaldon T,Dopazo J, Dopazo H. 2007. Phylemon: a suite of web toolsfor molecular evolution, phylogenetics and phylogenomics.Nucleic Acids Res. 35:W38–W42.

Thornberry NA, Rano TA, Peterson EP, et al. (12 co-authors).1997. Combinatorial approach defines specificities of mem-bers of the caspase family and granzyme B. Functionalrelationships established for key mediators of apoptosis.J Biol Chem. 272:17907–17911.

Uren AG, O’Rourke K, Aravind LA, Pisabarro MT, Seshagiri S,Koonin EV, Dixit VM. 2000. Identification of paracaspasesand metacaspases: two ancient families of caspase-likeproteins, one of which plays a key role in MALT lymphoma.Mol Cell. 6:961–967.

van Rheede T, Bastiaans T, Boone DN, Hedges SB, deJong WW, Madsen O. 2006. The platypus is in its place:nuclear genes and indels confirm the sister group relation ofmonotremes and therians. Mol Biol Evol. 23:587–597.

Varfolomeev EE, Schuchmann M, Luria V, et al. (18 co-authors).1998. Targeted disruption of the mouse caspase 8 gene ablatescell death induction by the TNF receptors, Fas/Apo1, andDR3 and is lethal prenatally. Immunity. 9:267–276.

Wang X, Grus WE, Zhang J. 2006. Gene losses during humanorigins. PLoS Biol. 4:e52.

Wheeler DL, Barrett T, Benson DA, et al. (30 co-authors). 2007.Database resources of the National Center for BiotechnologyInformation. Nucleic Acids Res. 35:D5–D12.

Xue Y, Daly A, Yngvadottir B, et al. (14 co-authors). 2006.Spread of an inactive form of caspase-12 in humans isdue to recent positive selection. Am J Hum Genet. 78:659–670.

Zheng XH, Lu F, Wang ZY, Zhong F, Hoover J, Mural R. 2005.Using shared genomic synteny and shared protein functions toenhance the identification of orthologous gene pairs. Bio-informatics. 21:703–710.

Arndt von Haeseler, Associate Editor

Accepted January 8, 2008

Evolution of Mammalian Caspases 841

Downloaded from https://academic.oup.com/mbe/article-abstract/25/5/831/1195498by gueston 18 February 2018