Insect Molecular Biology (2003) 12 (4), 345–351© 2003 The Royal Entomological Society 345 Blackwell Publishing Ltd. The evolution of HOM-C homeoboxes in the Dipteran family Drosophilidae R. DeSalle*, M. A. Branham*, P. O’Grady* and J. Gatesy† * Division of Invertebrates, American Museum of Natural History, 79th Street at Central Park West, New York, NY 10024, USA; † Department of Biology, Spieth Hall, University of California, Riverside, CA 92521, USA AbstractForty-five new Homeotic Complex (HOM-C) homeoboxsequences from six species of Drosophilidae ( Dro-sophila heteroneura , D. adiastola , Zaprionus vittiger , Chymomyza amoena , Scaptodrosophila pattersoni and Hirtodrosophila pictiventris ) were obtained usinga PCR-cloning method. These new homeoboxes arefrom the labial , proboscipedia , Deformed , Sex combsreduced , fushi tarazu , Antennapedia , Ultrabithorax , abdominal-A and Abdominal-B genes. Phylogeneticsignal in the homeobox sequences was assessed andseveral aspects of sequence evolution were examined.In particular, codon bias was examined and found toexist between the drosophilid species examined hereand Anopheles gambiae outgroup sequences. In addi-tion, different patterns of codon bias were detected inhomeoboxes interrupted with introns when comparedto homeoboxes that are uninterrupted.Keywords: Drosophila , homeobox, codon bias, phylogeny,evolution.Introduction While homeobox genes have been studied in a wide varietyof organisms (de Rosa et al ., 1999; Banerjee-Basu et al .,2000; Ferrier & Holland, 2001) very few studies havefocused on homeobox evolution in a group of relativelyclosely related organisms (Zardoya et al ., 1996). Here wecompare homeobox sequences from seven taxa in thefamily Drosophilidae selected because their phylogeneticdistribution covers a wide range of the diversity in this fam-ily. Drosophila heteroneura and D. adiastola are species inthe Hawaiian Drosophila clade and members of the genus Drosophila (Remsen & O’Grady, 2002). Both of these spe-cies are endemic to the Hawaiian archipelago. The remain-der of flies examined are from genera representative of themajor branching events in the family. Scaptodrosophila pat-tersoni and Chymomyza amoena are considered basalflies in the family (Grimaldi, 1990; Remsen & O’Grady,2002). Hirtodrosophila pictiventris and Zaprionus vittiger are considered more derived genera closely related to thesubgenera Drosophila and Sophophora , the latter of whichcontains the intensively studied species D. melanogaster (for which all HOM-C homeobox sequences are also avail-able). Phylogenetic relationships of the drosophilid taxa inthis study are well known and robust (Tatarenkov et al .,2001; Remsen & O’Grady, 2002), as are divergence timesfor these taxa, which range between 10 and 60 millionyears ago (Grimaldi, 1990; Russo et al ., 1995).The Homeotic Complex (HOM-C) of Drosophila is com-posed of two separate units, the Antennapedia (ANTC) andBithorax (BXC) complexes. Although both these complexesare located on the right arm of chromosome 3 (3R) in Drosophila melanogaster , they are considered geneticallyunlinked because they are separated by 7.5 Mb. We obtainedhomeobox sequences from both the ANTC and the BXCcomplexes for several drosophilid species by amplifyinggenomic DNA with a single pair of ‘HOM-C specific’primers. Resultant PCR products were cloned and 150 clonesfrom each species were sequenced. Nine homeoboxes( labial , proboscipedia , Deformed , Sex combs reduced , fushitarazu , Antennapedia , Ultrabithorax , abdominal-A and Abdominal-B ) were recovered from these species usingthis approach. Our taxonomic sampling strategy allows usto examine phylogenetic signal in HOM-C as well as test forpatterns of codon usage bias in homeoboxes across taxain the family Drosophilidae. Because several differenthomeoboxes were obtained for all the species in this study,we also examined codon usage bias as a function of genestructure by comparing those homeoboxes with introns( lab , pb and AbdB ) to those that lack these interruptions. Received 2 July 2002; accepted after revision 3 March 2003. Correspond-ence: Dr R. DeSalle, Division of Invertebrates, American Museum of NaturalHistory, 79th Street at Central Park West, New York, NY 10024, USA. E-mail:[email protected] R. DeSalle et al. © 2003 The Royal Entomological Society, Insect Molecular Biology , 12 , 345–351Figure 1. Nucleic acid sequences of 45 new partial Drosophila HOM-C homoeboxes. Top sequence for each gene is the D. melanogaster sequence from Genbank. Dots in the blocks indicate nucleotide identity to the D. melanogaster sequence. Abbreviations: Ag = Anopheles gambiae, Dm = Drosophila melanogaster, Hp = Hirtodrosophila pictivientris; Dh = D. heteroneura; Da = D. adiastola; Zv = Zaprionus; vittiger; Sp = Scaptodrosophila pattersoni; Ca = Chymomyza amoena. Sequences for A. gambiae and D. melanogaster were obtained from Genbank as detailed in the experimental procedures. All sequences have been deposited in Genbank under accession numbers AY194803 to AY194847.Evolution of homeoboxes in Drosophilidae 347 © 2003 The Royal Entomological Society, Insect Molecular Biology , 12 , 345–351 Results and discussion Results from PCR and cloning Of the 900 sequenced clones, about three-quarters con-tained inserts of the target length 105 bases (or slightlylonger to accommodate homeobox products with introns).The majority of the clones were either Ubx , abdA , Scr , Antp or ftz homologues. Since the amino acid sequences of Scr and Antp are identical over the region we amplified, diag-nosis of clones using amino acids was not possible.Instead, putative orthology was assigned to Scr and Antp based on nucleotide sequence similarity. Figure 1 showsthe nucleotide alignments of the 45 new Drosophila home-obox genes and the orthologs from Anopheles gambiae and D. melanogaster . Amino acid translations of thesequences for these genes were highly invariant except forthe ftz , pb , AbdB and lab genes (Fig. 2). These sequences,while generally lacking variation at the amino acid level,displayed a great deal of variation in their nucleotidesequences, primarily in third codon positions. Although welooked for polymorphism by sequencing multiple clones ofthe same homeobox gene in each species, none was dis-covered. This result is not surprising given that the genomicDNAs used to do