Identification of rice TUBBY-like genes and their evolution Qingpo Liu School of Agriculture and Food Science, Zhejiang Forestry University, Hangzhou, Lin’an, China TUBBY-like proteins (TULPs) are present in eukary- otes from single-celled to multicellular organisms. The first TUBBY gene was identified in obese mice [1]. A typical feature of TUBBY proteins is the approxi- mately 270-amino acid tubby domain at their C-termi- nal region. Accordingly, a number of such genes have been successively isolated from other organisms, such as Homo, Gallus, Xenopus, Zea and Arabidopsis [2], based on sequence characteristics. Interestingly, these genes form a small family in mammals, which consists of TUBBY and four TULPs [1,3]. By contrast, plants appear to harbour a large number of TULPs [2,4]. Moreover, compared with the high divergence of the N-terminal sequence of animal TULPs, most plant TULPs contain a conserved F-box domain at the cor- responding region [5]. Although the high conservation of the tubby domain across species suggests that TULPs may carry out cer- tain fundamental biological functions in common [4], little is known about their target genes or precise action mechanisms. In animals, TULPs are important for normal neuronal development and function [4]. It has been shown that TULPs play crucial roles in vesic- ular trafficking [6], mediation of insulin signalling [7] and gene transcription [8]. In Arabidopsis, Lai et al. [2] have demonstrated that TULP9 (AtTLP9) interacts with Skp1-like 1 (ASK1). They speculated that, as well as acting as transcription regulators, F-box domain- containing plant TULPs should have cellular function activities of F-box proteins in signal transduction. Sup- pression and overexpression analyses of Arabidopsis TULPs(AtTLPs) have shown that at least one AtTLP may function in the abscisic acid-regulated pathway [2]. Rice is one of the most important crops for human consumption. The final completion of the Oryza sativa genome [9] has made it possible to identify all the TUBBY-like family members in this plant species at the genome-wide level. In order to investigate the evolution and divergence of TULPs in eukaryotes, a phylogenetic, evolutionary and functional divergence analysis of the TUBBY-like gene family has been Keywords evolution; functional divergence; phylogeny; rice; TUBBY-like Correspondence Q. Liu, School of Agriculture and Food Science, Zhejiang Forestry University, Hangzhou, Lin’an, China Fax: +86 571 86971117 Tel: +86 571 86971611 E-mail: liuqp@genomics.org.cn (Received 28 September 2007, revised 7 November 2007, accepted 12 November 2007) doi:10.1111/j.1742-4658.2007.06186.x The identification of TUBBY-like genes in organisms ranging from single- celled to multicellular eukaryotes has allowed the phylogenetic history of this gene family to be traced back to the early evolutionary stages of eukaryote development. Rice TUBBY-like genes were located on chromo- somes 1, 2, 3, 4, 5, 7, 8, 11 and 12 without any obvious clustering. On a genomic scale, it was revealed that the rice TUBBY-like gene family proba- bly evolved mainly through segmental duplication produced by polyploidy. The altered selective constraints (or site-specific rate changes), related to functional divergence during protein evolution between plant and animal TUBBY-like genes, were statistically significant. Based on posterior proba- bility analysis, five amino acid sites (103, 312, 315, 317 and 319) are thought to be responsible for functional divergence. Abbreviations EST, expressed sequence tag; GEO, gene expression omnibus; NCBI, National Center for Biotechnology Information; RED, Rice Expression Database; RGP, Rice Genome Project; TIGR, Institute for Genomic Research; TULP, TUBBY-like protein. FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS 163 performed. This first description of the whole rice TUBBY-like gene family will aid in our understanding of the function of TULPs in plants. Results and Discussion Identification and sequence analysis of rice TULPs After carefully surveying the rice genome, 14 genes were defined as rice TULPs(OsTLPs; Table 1). Domain analysis showed that, with one exception (OsTLP13), a conserved F-box domain (PF00646) was found at the N-terminal region of OsTLPs. In addition to the tubby (PF01167) and F-box domains, most OsTLPs had two PROSITE signature patterns, termed TUB1 (PS01200) and TUB2 (PS01201), at their C-ter- minal region, evidence that strongly supports their reliability as members of the rice TUBBY-like family. The results of expressed sequence tag (EST) and cDNA blast searches showed that 13 of the 14 OsTLPs matched at least one significant EST sequence, and 11 of the 14 OsTLPs exactly matched a corre- sponding full-length cDNA sequence in the GenBank or KOME database (Table 1). These results indicated that most TULPs are expressed in the rice genome. The comparison of gene structure showed a con- served exon number pattern in OsTLPs, although the length of introns was different (Table 1). With three exceptions (OsTLP7, OsTLP13 and OsTLP14), all of the OsTLPs had four exons. Further analysis found that the F-box domain, except for OsTLP13, was encoded by the first exon, whereas the tubby domain was encoded by the following three exons. Phylogenetic analysis blast search against the GenBank database showed that both single-celled (Ostreococcus tauri, Tetrahy- mena thermophila, etc.) and multicellular (mammals, plants, etc.) organisms possessed tubby domain-con- taining proteins, indicating their functional importance for eukaryotes. In this study, in order to avoid biased analysis, the C-terminal tubby domain, rather than the highly divergent N-terminus, was used to perform phy- logenetic analysis. Supplementary Fig. S1 shows the multiple sequence alignment of tubby domains. The phylogenetic trees reconstructed using the neighbour- joining and minimal evolution methods in mega v3.1 (trees not shown) and the neighbour-joining method in phylip (Fig. 1) revealed similar topologies. At first glance, three clades (plant, animal and mixed clades) were evident in the tree (Fig. 1). In animals, the T. thermophila TULP (TtTLP) was evolutionarily distant from the other animal TULPs, suggesting long periods of divergence from this metazoan. Further analysis showed that invertebrate (insects, nematodes and Echinodermata) and vertebrate metazoans were not clustered into distinct clades, indicating that the divergence of animal TULPs probably occurred before the invertebrate–vertebrate split. In addition, mamma- lian TULPs were clustered into four subclades, three of which were evolutionarily clustered together with Actinopterygii TULPs; this indicates that divergence may have predated the Mammalia–Actinopterygii split. With five exceptions (AtTLP8, OsTLP13, PtaTLP10, CrTLP1 and CrTLP2), plant TULPs were tightly clus- tered together, supported by highly significant boot- strap values (996; Fig. 1). In the plant-specific clade, Table 1. List of rice TUBBY-like family members. OsTLP Accession number Chromosome Amino acid length Gene structure cDNA OsTLP1 AAT07611 5 445 AK103583 OsTLP2 BAD33172 8 451 AK120706 OsTLP3 CAE01783 4 462 AK104333 OsTLP4 ABA95864 12 445 AK102298 OsTLP5 BAC01219 1 448 AK102221 OsTLP6 AAX95106 11 440 OsTLP7 BAC20077 7 406 AK071159 OsTLP8 BAD08037 2 428 AK064855 OsTLP9 BAD73373 1 356 AK070401 OsTLP10 AAU03104 5 372 AK061747 OsTLP11 AAU10642 5 352 OsTLP12 BAB90233 1 368 AK106853 OsTLP13 BAD28008 2 427 OsTLP14 ABF95936 3 403 AK102370 TUBBY-like genes in rice Q. Liu 164 FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS monocotyledon and dicotyledon TULPs were not found in distinct groups, but, instead, were inter- spersed, suggesting that the significant expansion of plant TULPs should be no younger than the diver- gence time between monocots and dicots (approxi- mately 200 million years ago [10]). The mixed clade is very special, and consists of one fungus, one euglenozoan and five plant TUBBY-like proteins. In addition, one tubby domain-containing protein was identified in both Plasmodium yoelii yoelii and Plasmodium falciparum. These two proteins were excluded from phylogenetic analysis because of their failure in the chi-squared test for homogeneity of amino acid composition. However, if these sequences were arbitrarily included in the phylogenetic analysis, they were clearly classified into this clade (bootstrap value 920; tree not shown). Although tubby domain- containing proteins were found in several protozoans, this type of protein was only identified in one fungus after an exhaustive database search. It should be noted that O. tauri belongs to the Prasinophyceae, that diverged at the base of the monophyletic green lineage, which includes green algae and land plants [11]. Thus, the identification of one TULP in O. tauri and three TULPs in Chlamydomonas reinhardtii (CrTLPs) allowed the origin of this gene family to be traced back to the early evolutionary stages of eukaryote development. Genomic organization of OsTLPs Gene families can arise through tandem amplification, resulting in a clustered occurrence, or segmental dupli- cations of chromosomal regions, resulting in a scat- tered occurrence of family members [12]. It was observed that OsTLPs were located on chromosomes 1, 2, 3, 4, 5, 7, 8, 11 and 12. By contrast with only one OsTLP found on chromosomes 3, 4, 7, 8, 11 and 12, three (OsTLP5, OsTLP9 and OsTLP12), two (OsTLP8 and OsTLP13) and three (OsTLP1, OsTLP10 and OsTLP11) genes were located on chromosomes 1, 2 and 5, respectively (Fig. 2A). Although the distribution of OsTLPs on rice chromosomes was obviously uneven, no two OsTLPs were found to be located Fig. 1. Phylogenetic tree of eukaryote TUBBY domains. The neighbour-joining method wrapped in PHYLIP [28] was used to reconstruct the phylogenetic tree based on the multiple sequence alignment of tubby domains (supplementary Fig. S1). The numbers beside the branches represent bootstrap values (‡ 600) based on 1000 replications. To identify the species of origin for each tubby domain, a species acronym is included before the protein name: Aa, Aedes aegypti; Ag, Anopheles gambiae; Am, Apis mellifera; At, Arabidopsis thaliana; Bt, Bos taurus; Ca, Cicer arietinum; Cb, Caenorhabditis briggsae; Ce, Caenorhabditis elegans; Cf, Canis familiaris; Cr, Chlamydomonas reinhardtii; Dm, Drosophila mela- nogaster; Dp, Drosophila pseudoobscura; Dr, Danio rerio; Ec, Encephalitozoon cuniculi; Gg, Gallus gallus; Hs, Homo sapiens; Lm, Leishmania major; Lp, Lemna paucicostata; Mam, Macaca mulatta; Mm, Mus musculus; Mt, Medicago truncatula; Os, Oryza sativa; Ot, Ostreococcus tauri; Pc, Pyrus communis; Pt, Pan trog- lodytes; Pta, Populus trichocarpa; Pxa, Platanus · acerifolia; Rn, Rattus norvegicus; Sp, Strongylocentrotus purpuratus; Tc, Triboli- um castaneum; Tn, Tetraodon nigroviridis; Tt, Tetrahymena ther- mophila; Xl, Xenopus laevis; Xt, Xenopus tropicalis. Q. Liu TUBBY-like genes in rice FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS 165 close to each other, for example on the same scaffolds or bacterial artificial chromosomes. Thus, segmental duplications probably contributed to the expansion of the rice TUBBY-like gene family. To test this hypothe- sis, the method of Schauser et al. [12] was performed to investigate the evolutionary relationships between duplicated segments. In this way, five pairs of dupli- cated segments, including OsTLP1 ⁄ 5, OsTLP4 ⁄ 6, OsTLP7 ⁄ 14, OsTLP9 ⁄ 11 and OsTLP10 ⁄ 12, were iden- tified (Fig. 2B). Further examination of the rice dupli- cation blocks identified by Yu et al. [13] and Wang et al. [14] revealed that the five pairs of OsTLPs con- stituted three duplication blocks corresponding to part of the long arm of chromosome 1 (OsTLP9, OsTLP12 and OsTLP5) and part of the long arm of chromo- some 5 (OsTLP11, OsTLP10 and OsTLP1), part of the short arm of chromosome 3 (OsTLP14) and part of the long arm of chromosome 7 (OsTLP7), and part Fig. 2. Genomic organization of OsTLPs. (A) Localization of OsTLPs on rice chromosomes. Black boxes indicate the three duplication blocks between chromosomes 1 and 5, 3 and 7, and 11 and 12. The relative sizes of chromosomes are derived from RGP. (B) Detection of seg- mental duplications in regions of the rice genome encompassing OsTLPs. The sequences of 10 proteins surrounding each OsTLP (five on each side) were concatenated to form one block. A vertical black bar indicates the concatenation of two protein sequences. This was per- formed for all 14 OsTLPs, resulting in 14 blocks, which were then searched against each other using a reciprocal best-hit BLAST strategy. The five pairs of OsTLPs identified resulting from segmental duplications are shown here. TUBBY-like genes in rice Q. Liu 166 FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS of the short arm of chromosome 11 (OsTLP6) and part of the short arm of chromosome 12 (OsTLP4). The three pairs (OsTLP9 ⁄ 11, OsTLP12 ⁄ 10 and OsTLP5 ⁄ 1) located on chromosomes 1 and 5 may originate from an ancient whole genome duplication, followed by a segmental inversion. Another two pairs (OsTLP6 ⁄ 4 and OsTLP7 ⁄ 14) may be the result of recent segmental duplication of chromosomes 11 and 12 and chromosomes 3 and 7, respectively [13,14]. With regard to OsTLP8 and OsTLP13, they may have arisen as a result of duplication events, and lost their counterparts over the long period of evolution, because there was only one copy of each in the duplicated regions on chromosome 2. This explanation is reason- able, as it has been shown that the rice genome was probably generated through two rounds of ancient polyploidy events that were followed by massive gene losses and numerous chromosome rearrangements [15]. Functional divergence (altered functional constraint) analysis A maximum likelihood test of functional divergence was performed on the basis of the Gu method [16], using the program diverge [17], which evaluates the shifted evolutionary rate after gene duplication or spe- ciation [18]. The advantage of the Gu method [16] is that it is not sensitive to saturation of synonymous sites. The estimation was based on the phylip neigh- bour-joining tubby domain tree (Fig. 1). The result showed that the coefficient of type I functional diver- gence between the plant and animal clades was statisti- cally significant (h = 0.387 ± 0.149; likelihood ratio test statistic, 6.669; P < 0.05), indicating that signifi- cantly different site-specific shifts of evolutionary rate may take place at certain amino acid sites [18] between plant and animal tubby domains. In order to identify these variant amino acid sites, the posterior probability of divergence was determined for each site. The results showed that the functional divergence between plant and animal TUBBY-like proteins could be partially attributed to variation on at least five amino acid sites (103, 312, 315, 317 and 319, counting from PtaTLP10; supplementary Fig. S1 and Fig. 3). Boggon et al. [8] performed X-ray crystallographic analysis of the tubby domain of mouse TUBBY and found a unique protein structure: a 12-stranded b-barrel conformation filled with a central hydrophobic core that traversed the entire barrel. It was observed that these five amino acids fall within the fifth (E5, site 103) and ninth (E9A, site 312; E9B, site 319) b-strands, and the loop between E9A and E9B (sites 315 and 317) (supplemen- tary Fig. S1). Santagata et al. [19] demonstrated experimentally that the amino acids that interact with l-a-glycero- phospho-d-myo-inositol 4,5-bisphosphate (GPMI-P 2 ) are mostly in b-strands 4, 5 and 6 and helix 6A. More importantly, three positively charged amino acid resi- dues, R332, R363 and K330 (corresponding to sites 103, 182 and 101 in this study), were found to be crucial for the tubby domain of mouse TUBBY protein to bind phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P 2 ] [19]. Figure 3 shows that site 103 (Q k = 0.877) is invariant arginine in animal tubby domains, a result supporting its importance for animal TULPs specifically binding a number of phosphory- lated phosphoinositides [20]. By contrast, the same position in plant tubby domains contains several amino acids (arginine, serine or lysine) with different chemical properties, such as the uncharged polar amino acid serine (S). This variation may be explained as a relaxed selective constraint at site 103 in plant tubby domains. In addition, it was observed that site 182 was not an arginine (R) and showed a plant-spe- cific deletion (supplementary Fig. S1). These results suggest that at least some of the plants may have lost their ability to bind phosphatidylinositol phosphates. Boggon et al. [8] observed a groove of highly posi- tive charge that was bordered at the top by helix H8 and at the bottom by the large 7–8 loop and the three- stranded ‘extra’ 9ABC sheet. The result showed that the amino acid sites 312, 315, 317 and 319 were func- tional divergence related, in which one site was located at E9A (site 312) and E9B (site 319) respectively, and two sites (315 and 317) were positioned in the E9A– E9B loop (supplementary Fig. S1). It was observed that the four amino acid sites were significantly con- served in plants (Fig. 3), indicating that strong func- tional constraints were imposed on these sites. In comparison with plants, great variation was observed in these sites in animal tubby domains. Of these sites, site 317 was predicted to be highly functional diver- gence related (Q k = 0.954). It was observed that the chemical properties of the amino acid at site 317 were significantly different between plant and animal tubby domains. In plants, the amino acid at site 317 was the invariant nonpolar leucine (L), whereas it was changed to uncharged polar serine (S), threonine (T) or aspara- gine (N) in animals (Fig. 3). Molecular structural and genetic analyses have sug- gested a common function for animal tubby domains that can bind to double-stranded DNA and phospha- tidylinositol phosphates [4,8]. With regard to the N-terminus, although this region shows a lack of con- servation in animals, it is able to activate transcription [8]. Unlike animals, the N-terminal region of most Q. Liu TUBBY-like genes in rice FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS 167 plant TUBBY-like family members often contains a well-conserved F-box domain. Experimental evidence has shown that AtTLP1, AtTLP2, AtTLP3, AtTLP6, AtTLP7, AtTLP9, AtTLP10 and AtTLP11 are expressed in all tested organs, whereas AtTLP5 and AtTLP8 are tissue specifically expressed in Arabidopsis [2]. After querying the rice dbEST database at the National Center for Biotechnology Information (NCBI), it was found that, with three exceptions (OsTLP11, OsTLP12 and OsTLP13), OsTLP genes showed a tissue-specific expression pattern, although most were expressed in nearly all the examined tissues (Table 2). However, the expression of several phylo- genetically closely related OsTLP genes showed similar or overlapping tissue specificity, for example OsTLP1 and OsTLP5 (Table 2). Interestingly, TULPs can be expressed with cell-type specificity and can be regu- lated by their subcellular localization [4,21]. He et al. [21] found that, in hypothalamic neurones, TUB was localized in the cytoplasm and nucleus, whereas, in photoreceptor cells, it appeared to be found only in the cytoplasm. The reason for the localization of TULPs in different cell populations still remains unknown. In addition, Lai et al. [2] demonstrated that AtTLP9 might participate in the abscisic acid signal- ling pathway. The Rice Expression Database (RED) [22] and gene expression omnibus (GEO) in NCBI were also queried; it was found that OsTLP4, OsTLP5, OsTLP7, OsTLP9, OsTLP10 and OsTLP12 were prob- ably involved in the abscisic acid and gibberellin sig- nalling processes. Nevertheless, more in-depth studies are needed to establish their distinctive activities and biological roles. Experimental procedures Collection of rice TULPs The consensus sequence of the tubby domain (PF01167) was obtained from the Pfam database. The Arabidopsis TUBBY-like proteins (accession numbers: AtTLP1, AF487267; AtTLP2, AY045773; AtTLP3, AY045774; AtTLP5, AY046921; AtTLP6, AF487268; AtTLP7, Fig. 3. Functional divergence related amino acid site candidates (Q k > 0.6). A site-specific profile based on the posterior probability (Q k ) was used to identify critical amino acid sites that were respon- sible for functional divergence [18] between the animal and plant tubby domains. According to the definition, a large Q k value indi- cates a high possibility that the functional constraint (or the evolu- tionary rate) of a site is different between two clusters. (A) Animals; (B) Plants; and (C) Posterior probability values (Q k ) of five amino acid sites. TUBBY-like genes in rice Q. Liu 168 FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS AY092403; AtTLP8–AtTLP10, AF487269–AF487271; AtT- LP11, AY046922) were downloaded from the GenBank database. In an attempt to obtain all the TULP members, the rice protein sequences collected in the Rice Genome Project (RGP) [9], Institute for Genomic Research (TIGR) and NCBI were downloaded to construct a local rice pro- tein database. With the tubby domain consensus and the Arabidopsis TUBBY-like proteins (AtTLPs) as queries, psi- blast was seeded to search the local and Oryza sativa pro- tein database in NCBI with an e-value of 10. Moreover, a psi-blast search against the nonredundant GenBank data- base was performed to collect tubby domain-containing proteins in other species using the tubby domain consensus sequence as query. The collected Arabidopsis and other published TULPs were used to construct a hidden Markov model (HMM) profile; this was followed by an hmmer (version 2.3.2) [23] search of the rice proteome. In addition, a tblastn search against the rice genomic sequences depos- ited in RGP, NCBI and TIGR was also conducted. Signifi- cant hits were collected and redundant hits were removed by manual inspection. The domain architecture of eukary- ote TUBBY-like proteins was analysed using the domain analysis program interproscan [24] with the default parameters. The accession numbers of the collected TULPs are listed in supplementary Table S1. Analysis of rice TULP evolution TULPs were found to show a scattered distribution pattern on rice chromosomes. Consequently, segmental duplication was assumed to have contributed to the expansion of this gene family. Schauser et al. [12] demonstrated that the effective way to detect this type of duplication event was to identify additional paralogous protein pairs in the neigh- bourhood of each of the family members. Accordingly, the present study focused on 10 proteins encoded by genes flanking each of the 14 rice TULPs (five on each side). Multiple sequence alignment and construction of the phylogenetic tree Alignment of the tubby domains was performed using the clustalw program [25] with the default parameters. The multiple aligned sequences were initially subjected to a chi-squared analysis for homogeneity of amino acid com- position, implemented in tree-puzzle v5.2 [26]. Sequences that failed in this test were excluded. To investigate the evolutionary relationships amongst tubby proteins, a phy- logenetic tree was reconstructed by employing the neigh- bour-joining method and the minimal evolution method wrapped in mega v3.1 [27]. For both methods, the param- eters p-distance model and pairwise deletion of gaps ⁄ miss- ing data were selected. In addition, phylip [28] was employed to reconstruct a neighbour-joining tree from the same data. A bootstrap test of phylogeny was performed with 1000 replications for each method. The programs njplot and mega v3.1 [27] were used to display the phy- logenetic trees. Estimation of functional divergence diverge, a program developed by Gu and Velden [17], was used to detect functional divergence between members of a protein family [29]. In the TUBBY-like gene family, two gene clusters of interest, including plant and animal TULPs, were selected. The coefficient of type I functional divergence h and the likelihood ratio test statistic between Table 2. EST-derived expression profile for rice TUBBY-like genes. OsTLP Tissue Root Leaf Shoot Panicle Flower Callus OsTLP1 ++ + +++ +++ + ++++++++ OsTLP2 ++++++ +++ +++++++++ +++ + + OsTLP3 ++ +++ ++++++++++++ ++++++++ ++++ +++++++++++++ OsTLP4 ++ ++++++++++ ++++++++ ++++++++ + +++++++ OsTLP5 ++++++ + ++++++ ++++++ ++++ ++++++++ OsTLP6 ++ +++ ++++ ++ ++ OsTLP7 ++ ++ +++ ++ +++ OsTLP8 ++++ +++ ++++ +++++++++++++++++++ ++ OsTLP9 + +++++++ ++++++++ ++++ +++ ++ OsTLP10 +++++++++++++++ ++++++++++++ ++++++ +++++ OsTLP11 a OsTLP12 + OsTLP13 + + ++ OsTLP14 ++ ++++ ++ +++ ++ +, Presence of gene sequences in EST collection derived from the indicated tissues. a No significant EST hit was found for OsTLP11 in the present EST database, indicating that this gene might be weakly expressed in rice. Q. Liu TUBBY-like genes in rice FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS 169 the two clusters were quickly calculated. A h value signifi- cantly greater than zero indicates altered selective con- straints of amino acid sites after gene duplication [18]. Acknowledgement This project was supported by the China Postdoctoral Science Foundation (No. 20060390348). References 1 Kleyn PW, Fan W, Kovats SG, Lee JJ, Pulido JC, Wu Y, Berkemeier LR, Misumi DJ, Holmgren L, Charlat O et al. (1996) Identification and characterization of the mouse obesity gene tubby: a member of a novel gene family. Cell 85, 281–290. 2 Lai CP, Lee CL, Chen PH, Wu SH, Yang CC & Shaw JF (2004) Molecular analysis of the Arabidopsis TUBBY-like protein gene family. Plant Physiol 134 , 1586–1597. 3 North MA, Naggert JK, Yan Y, Noben-Trauth K & Nishina PM (1997) Molecular characterization of TUB, TULP1, and TULP2, members of the novel tubby gene family and their possible relation to ocular diseases. Proc Natl Acad Sci USA 94, 3128–3133. 4 Ikeda A, Nishina PM & Naggert JK (2002) The tubby- like proteins, a family with roles in neuronal develop- ment and function. J Cell Sci 115, 9–14. 5 Gagne JM, Downes BP, Shiu SH, Durski A & Vierstra RD (2002) The F-box subunit of the SCF E3 complex is encoded by a diverse superfamily of genes in Arabidopsis. Proc Natl Acad Sci USA 99, 11519– 11524. 6 Hagstrom SA, Adamian M, Scimeca M, Pawlyk BS, Yue G & Li T (2001) A role for the tubby-like protein 1 in rhodopsin transport. Invest Ophthalmol Vis Sci 42, 1955–1962. 7 Kapeller R, Moriarty A, Strauss A, Stubdal H, The- riault K, Siebert E, Chickering T, Morgenstern JP, Tartaglia LA & Lillie J (1999) Tyrosine phosphoryla- tion of tub and its association with Src homology 2 domain-containing proteins implicate tub in intracellu- lar signaling by insulin. J Biol Chem 274, 24980– 24986. 8 Boggon TJ, Shan WS, Santagata S, Myers SC & Shap- iro L (1999) Implication of tubby proteins as transcrip- tion factors by structure-based functional analysis. Science 286, 2119–2125. 9 International Rice Genome Sequencing Project (2005) The map-based sequence of the rice genome. Nature 436, 793–800. 10 Wikstrom N, Savolainen V & Chase MW (2001) Evolu- tion of the angiosperms: calibrating the family tree. Proc R Soc Lond B Biol Sci 268, 2211–2220. 11 Bhattacharya D & Medlin L (1998) Algal phylogeny and the origin of land plants. Plant Physiol 116, 9–15. 12 Schauser L, Wieloch W & Stougaard J (2005) Evolution of NIN-like proteins in Arabidopsis, rice and Lotus japonicus. J Mol Evol 60, 229–237. 13 Yu J, Wang J, Lin W, Li S, Li H, Zhou J, Ni P, Dong W, Hu S, Zeng C et al. (2005) The genomes of Oryza sativa: a history of duplications. PLOS Biol 3, 266–281. 14 Wang X, Shi X, Hao B, Ge S & Luo J (2005) Duplica- tion and DNA segmental loss in the rice genome: impli- cations for diploidization. New Phytol 165, 937–946. 15 Guyot R & Keller B (2004) Ancestral genome duplica- tion in rice. Genome 47, 610–614. 16 Gu X (1999) Statistical methods for testing functional divergence after gene duplication. Mol Biol Evol 16, 1664–1674. 17 Gu X & Velden KV (2002) DIVERGE: phylogeny- based analysis for functional–structural divergence of a protein family. Bioinformatics 18, 500–501. 18 Gu X (2003) Functional divergence in protein (family) sequence evolution. Genetica 118, 133–141. 19 Santagata S, Boggon TJ, Baird CL, Gomez CA, Zhao J, Shan WS, Myszka DG & Shapiro L (2001) G-protein signaling through tubby proteins. Science 292, 2041– 2050. 20 Xi Q, Pauer GJT, Marmorstein AD, Crabb JW & Hagstrom SA (2005) Tubby-like protein 1 (TULP1) interacts with F-actin in photoreceptor cells. Invest Ophthalmol Vis Sci 46, 4754–4761. 21 He W, Ikeda S, Bronson RT, Yan G, Nishina PM, North MA & Naggert JK (2000) GFP-tagged expres- sion and immunohistochemical studies to determine the subcellular localization of the tubby gene family mem- bers. Brain Res Mol Brain Res 81, 109–117. 22 Yazaki J, Kishimoto N, Ishikawa M & Kikuchi S (2002) The Rice Expression Database (RED): gateway to rice functional genomics. Trends Plant Sci 7, 563– 564. 23 Eddy SR (1998) Profile hidden Markov models. Bioinformatics 14, 755–763. 24 Quevillon E, Silventoinen V, Pillai S, Harte N, Mul- der N, Apweiler R & Lopez R (2005) InterProScan: protein domains identifier. Nucleic Acids Res 33, W116–W120. 25 Thompson JD, Higgins DG & Gibson TJ (1994) CLU- STALW: improving the sensitivity of progressive multi- ple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673–4680. 26 Schmidt HA, Strimmer K, Vingron M & von Haeseler A (2002) tree-puzzle: maximum likelihood phyloge- netic analysis using quartet and parallel computing. Bioinformatics 18, 502–504. TUBBY-like genes in rice Q. Liu 170 FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS 27 Kumar S, Tamura K & Nei M (2004) MEGA3: inte- grated software for molecular evolutionary genetics analysis and sequence alignment. Brief Bioinform 5, 150–163. 28 Retief JD (2000) Phylogenetic analysis using phylip. Methods Mol Biol 132, 243–258. 29 Gu J, Wang Y & Gu X (2002) Evolutionary analysis for functional divergence of Jak protein kinase domains and tissue-specific genes. J Mol Evol 54, 725–733. Supplementary material The following supplementary material is available online: Fig. S1. Multiple sequence alignment of eukaryote tubby domains. Table S1. Accession numbers of collected TULPs in 33 species. This material is available as part of the online article from http://www.blackwell-synergy.com Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corre- sponding author for the article. Q. Liu TUBBY-like genes in rice FEBS Journal 275 (2008) 163–171 ª 2007 The Author Journal compilation ª 2007 FEBS 171 . Identification of rice TUBBY-like genes and their evolution Qingpo Liu School of Agriculture and Food Science, Zhejiang Forestry. description of the whole rice TUBBY-like gene family will aid in our understanding of the function of TULPs in plants. Results and Discussion Identification and