Phylogenetic analysis of cnidarian TIR sequences in comparison to a selection of TIR domains from other species. The maximum likelihood (ML) tree shown is the result of analysis of an HMM-based alignment of TIR domains. A number of the TIR sequences identified and discussed in the text are incomplete due to the presence of introns of unknown size, and hence were not included in the phylogenetic analyses. Three clades are resolved by these analyses, corresponding to the TIR domains characteristic of the 'MyD88-type', 'Toll/TLR-type' and 'IL-1R-type'. In addition to the TIR domain, the first of these types contains a death domain and the second contains multiple LRRs. Like the mammalian receptors for interleukin 1, the three Nematostella proteins falling into the third clade each also contain multiple immunoglobulin domains. Note that HyTRR1 does not contain such domains and that it is not yet clear whether either of the Acropora proteins does. The Acropora sequences included in the analysis were predicted from A. palmata genomic clones (ApGenomic) and from an A. millepora cDNA clone (AmTIR-1). Hydra lacks a canonical Toll/TLR, having only two MyD88 genes and the two sequences known as TRR-1 and TRR-2; H. magnipapillata and N. vectensis sequences are indicated by the prefixes Hy and Nv, respectively. Reference sequences: HsMyD88, human MyD88 (SwissProt:Q99836); DmMyD88, fly MyD88 (GenBank:AAL56570); SdMyD88, Suberites MyD88 (EMBL:CAI68016); Dmtoll, fly Toll (SwissProt:P08953); HsTLR4, human TLR4 (EMBL:CAD99157); Arabidopsis (GenBank:AAN28912).