Proc. Null. Acud. Sci. USA Vol. 86, pp. 7716-7720, October 1989 Biochemistry kmphila z&homeobox genesJ (NK-1, NK-2, NK-3, and NK-4 DNA clones/chromosome locations of genes) YONGSOK KIM AND MARSHALL NIRENBERG Laboratory of Biochemical Genetics, National Heart, Lung and Blood Institute, National Institutes of Health, Building 36. Room lC-06. Bethesda, MD 20892 Contributed by Marshall Nirenberg, July 6, 1989 ABSTRACT i 6&r Drosophila melanoEaster homeohox genes were found by screening a genomic fiNA library with oligodeoxynucleotides that correspond to a conserved amino acid sequence that is part of the putative site of homeobox proteins that recognizes nucleotide sequences in DNA. The amino acid sequences of NK-2, NK-3, and NK-4 homeoboxes are more closely related to one another (59-66% homology) than they are to other Drosophila homeoboxes (28-54% ho- mology), whereas the homeobox of NK-1 is most closely related, in order of decreasing homology, to muscle segment homeobox, zerkniillt-1, NK-3, and distal-less homeoboxes. Three of the genes, NK-1, NK-3, and NK-4, comprise a cluster of homeobox genes located in the 9331-5 region of the right arm of the third chromosome, whereas the fourth homeobox gene, NK-2, is located in the lCl-5 region of the X chromosomd Homeobox genes encode DNA binding proteins that regulate gene expression during development or in the adult (l-4). In most cases, the similarity between different kinds of ho- meobox proteins extends only over a segment of the protein that consists of 60-61 amino acid residues, the homeo- domain, which is thought to be the portion of the protein that recognizes nucleotide sequences in DNA. Homeobox genes are particularly well expressed in nervous system, and the homeobox family of genes encodes the largest set of proteins that regulate gene expression in the nervous system that has been identified thus far (5-9). In this report, we describe four newly discovered, related Drosophila homeobox genes that were detected with oligo- nucleotide probes corresponding to an amino acid sequence that is thought to be part of the nucleotide sequence recog- nition site of homeobox proteins. METHODS AND MATERIALS Oligodeoxynucleotide. An Applied Biosystems DNA syn- thesizer 380B was used to synthesize oligodeoxynucleotides. Oligonucleotides with the trityl groups attached were purified by OPC column chromatography and trityl groups then were iemoved as described by Applied Biosystems. [y"P]ATP with a specific activity of 6000 Ci/mmol (1 Ci = 37 GBq) (New England Nuclear) was used for phosphorylation of oligode- oxynucleotides catalyzed by T4 polynucleotide kinase (10). Detection and Cloning of Homeobox Genes. A Drosophila melanogas~er genomic DNA library in Charon 4A (11) was obtained from the American Type Culture Collection. Recom- binant phage [48,000 plaque-forming units (pfu)] and 2 x 109 Escherichia coli KH802 cells were plated in Petri dishes (150 mm) at a concentration of 12,000 pfu per dish. Four nitrocel- lulose replica filter plaque lifts were obtained from each Petri dish, and each filter was hybridized with a different [32P]- oligodeoxynucleotide preparation [16-64 oligodeoxynucle- otide species per preparation; 1.5 x lo6 cpm/ml; 120-150 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "adverfisement" in accordance with 18 U.S.C. 61734 solely to indicate this fact. fmol/ml (the sum of all species of oligodeoxynucleotides)] at 37oC overnight and washed with a solution containing tetra- methylammonium chloride at 53oC or 50oC for 30 min for 17- mers or 16-mers, respectively, as described by Wood ef a/. (12). DNA Sequencing.* Cloned genomic DNA fragments cleaved by restriction enzymes were subcloned into Bluescript pKS+. Both strands of the homeobox regions of the following DNA fragments were sequenced by the dideoxynucleotide chain- termination method (13) using Ml3 universal primers or spe- cific oligodeoxynucleotide primers and Sequenase 2 (United States Biochemical): NK-1, 1.4-kilobase (kb) EcoRl/Pst I DNA fragment; NK-2, 1.2-kb EcoRl/Psf I DNA fragment; NK-3, 0.7-kb Pst I DNA fragment; NK-4, 0.4-kb and 2.3-kb upstream Hind111 DNA fragments. dlTP was used to reduce compression of DNA bands. Locations of Genes on Chromosomes. Salivary gland poly- tene chromosomes were hybridized with EcoRl-cleaved ge- nomic DNA fragments that contained the appropriate ho- meobox region and had incorporated biotin-16 dUMP in place of some dTMP residues as described (14). Detek l-HRP kits (Enzo Biochemicals) and the protocol supplied by the man- ufacturer were used. RESULTS AND DISCUSSION Detection of Homeobox Genes. The Drosophila genomic DNA library of Maniatis et al. (11) in Charon 4A was screened for recombinants corresponding to homeobox genes with five [32P]oligodeoxynucleotide probe preparations designed to hybridize to highly conserved homeobox nucle- otide sequences. The oligonucleotide preparations were 16 OJ 17 nucleotides long and each consisted of multiple species of oligodeoxynucleotides (described in the legend to Fig. 2). Replica filters were prepared and each filter was hybridized to a different [32P]oligodeoxynucleotide preparation. The filters were washed under high-stringency conditions with a solution that contained tetramethylammonium chloride, which selectively binds to A*T base pairs and raises the melting temperature (fm) of A-T base pairs to that of CC base pairs (12). The t, of each [32P]oligodeoxynucleotide-DNA duplex then was dependent on the nuniber of contiguous base pairs formed but was not affected by the proportion of GC vs. A*T base pairs (12). Consequently, all species of 17-mer oligodeoxynucleotides hybridized to DNA were washed ~11 the same temperature (53oC) for the stringent wash, and ;`I 16-mers were washed at 50oC. Of the 48,000 phage plaques that were screened, =X1() clones were obtained that exhibited a positive autoradio- graphic signal with one of the five [32P]oligonucleotide probe preparations, and 7 recombinant clones were obtained that gave positive signals with two or more probe preparations. Many of the 200 clones that were detected with only one *The sequences reported in this paper for NK-1 to NK-4 have been deposited in the GenBank data base (accession nos. M273sy. M27290, M27291, M27292, respectively). 7716 Biochemistry: Kim and Nirenberg Proc. Natl. Acad. Sci. USA 86 (1989) 7717 probe were cloned, but they were not studied further. The 7 clones that were detected with two or more 32P-labeled probes were characterized by restriction site analysis and the nucleotide sequences of the homeobox regions of some of the clones were determined by using unlabeled probes as se- quencing primers. Five of the 7 clones were found to be previously unknown homeobox genes. The 2 remaining ,lones correspond to known homeobox genes; 1 clone con- ..iins zerkntillt-1 (zen-1) and Zen-2 DNA (15), and the other clone corresponds to either en or inv (16) (data not shown). Characterization of Homeobox Genes. In Fig. 1 are shown partial restriction site maps of the homeobox genes NK-1, NK-2, NK-3, and NK-4, the locations of homeobox regions within the DNA inserts, and the direction of transcription. The approximate chain lengths of the cloned NK-1 and NK-2 genomic DNA fragments are 15.0 and 14.1 kb, respectively. Three of the 7 clones detected with two or more probes correspond to NK-3. Clone 6 is a 14.7-kb DNA fragment that L-ontains the NK-3 homeobox sequence. Clones 3 and 9 contain similar or identical DNA inserts (14.6 kb) that overlap clone 6 and contain both NK-3 and NK-4 homeobox se- quences separated by 27.8 kb. The restriction site map of NK-3 and NK-4 shown is derived from data obtained from clones 6, 3, and 9. Subcloned EcoRI DNA fragments from clone 3 were used to determine NK-3 and NK-4 nucleotide ATGGCACCAACATGTGCCGAMAATTCCAATTRATCGAAC -29; CCGTGGTGATTGATTTCCGTTTTCCAATCCCCCAGGACATT~CA~TGTCTGTGAT~ -234 ATGGCCCTAGCCTGTTGACTTATGCAARAAGAGAGACACCC~~CTTATCGT~C~ -175 Characterization of NK-2. The nucleotide sequence of the homeobox region of the NK-2 gene is shown in Fig. 3. The deduced amino acid sequence before the homeobox contains repetitive asparagine residues and a highly acidic region consisting of 14 aspartyl or glutamyl residues in a 31-amino acid segment (45% acidic amino acid residues). Twenty-five percent of the amino acid residues before the homeobox are glycine plus alanine. The carboxyl-terminal 30 amino acid residues of NK-2 are rich in histidine (20%), proline (17%), and glycine (17%). A 168-nucleotide 3'-untranslated region also is shown. Characterization of NK-3 and NK-4. The nucleotide se- quence and deduced amino acid sequence of part of the NK-3 homeobox gene are shown in Fig. 4. The initial part of the sequence consists of part of an exon that encodes 54 amino acids (17% alanine, 19% serine and threonine, and 9% asparagine), which is followed by a short, 11Pnucleotide intron within the 26th codon before the homeobox. The intron-exon structure of the NK-3 gene was confirmed by sequencing NK-3 cDNA clones (K. Webber, Y.K., and M.N., unpublished data). The initial portion of the second sequences. Partial Nucleotide Sequence of NK-1. The sequence of an 811~nucleotide portion of the NK-1 gene is shown in Fig. 2. The first 198 nucleotides correspond to the 3' portion of an i Itron. Another intron, 217 nucleotides long, was found \dllI: P. Pst 1; R, I3oRI; (R), EcoRl cloning \rte created by ligation of an EcoRl linker to genomic DNA. v ATCTCCTCTTCTTTTTTTTGTCTTGCAG CC CAG GAT TTG AAT GAC ATG GAT - Gln m Leu Asn m Met iAspl CAG GAC GAT ATG TGT GAC GAT GGC AGC GAT ATC GAC GAT CCC AGC Gln(Aspd Met Cys RspAsplGly Ser m Ile P,d Pro Ser AGC GAG ACG GAC TCC AAA AAG GGA GGC AGT CGT AAT GGG GAT GGA rq Thr Ala,Phe Thr Tyr Glu Gin Leu Val Ser Leu Glu As,, Lys he Lys Thr Thr Arq Tyr Leu Ser Val cys Glu Arq Leu Asn Leu ,124 -42 -79 -2, -34 -12 12 4 5: 19 102 34 I CC CTC AGC TTG AGC CTG ACA GAG ACG CAG GTGAGCAATGATATATACT 151 Leu Ser Leu Ser Leu Thr Glu Thr Gin 44 TATTGTTAAAGATTAAAATCCAGAGAAGTTATGTATATTTTGCAAQAGTTGGTATAA 214 26; 328 --"----------- ATTTCTCATCATCCTTTTACAG GTT AAA ATT TGG TTC CAG AAC CGC 0.X 377 Val Lys Ile Trp Phe Gin As" Arq Arg 53 I CC MG TGG AAG AAG CAG AAC CCC GGC ATG GAT GTC AAC TCC CCC 421 Thr Lys Trp Lys Lys Gin As" Pro Gly Met Asp Val Asn Ser Pro 68 ACC ATC CCC CCG CCC GGC GGC GGC TCC TTC GGA CCG GG 460 Thr Ile Pro Pro Pro Gly Gly Gly Sex Phe Gly Pro Gly 81 FIG. 2. Nucleotide sequence and deduced amino acid sequence of the homeobox region of the NK-1 gene. Deoxynucleotide and amino acid residues are numbered on the right; 1 corresponds to the first deoxynucleotide or amino acid residue in the homeobox, which is enclosed in a large box. The acidic amino acid region is indicated by boxed Asp or Glu residues. Repetitive Gly residues before the homeobox also are enclosed in a box. The 1st and 2nd boxed nucleotide sequences in intron 2 are possible sites for binding of zeste protein to DNA (17). The 3rd site, ANNNNCATTA, is an Antp protein binding site (18). Arrowheads represent intron-exon junc- tions. Nucleotide 12 in an NK-1 genomic DNA clone was C. whereas the corresponding nucleotide residue found in an NK-1 cDNA clone was T. All oligodeoxynucleotide probes are complementary to the DNA strand shown: probe sequences, starting from the 5'-terminal nucleotide residues are as follows: -, probe 121, 24 species of 17-mers, (-)TNTGRAACCA(T/A/G)TARAA; - -, probe 125, 48 species of l7-mers, (-)AACCA(T/G/A)ATYTTNACYTG; --. probe 126,64 species of 17-mers, (-)AAYTCYTTYTCNAGYTC; - -, probe 127. 64 species of 17-mers, (-)-C(T/G)RTlYTCRTTRAAYTC; . . . , probe 130, 16 species of 16-mers, (-)A(C/G)T(C/G,(C/G)TIT/ G)CTCCAGCTC. Y, pyrimidine; R, purine. 7718 Biochemistry: Kim and Nirenberg Proc. Natl. Acad. Sci. USA 86 (1989) exon before the homeobox consists of 35% serine and 19% proline residues. The 54 amino acid residues after the ho- meobox are rich in alanyl and glycyl residues (22%) as well as leucyl residues (11%). In Fig. 5 is shown the nucleotide sequence of part of the NK-4 gene. The initial part of the sequence consists of part of an intron, which is followed by an exon that contains the homeobox domain. The carboxyl-terminal region of the de- duced NK-4 protein contains repetitive glutamine residues (M or opa repeats and a CAX repeat in the corresponding DNA). Locations of Genes on Chromosomes. The cytological lo- cations of the NK-1, NK-2, NK-3, and NK-4 genes in Drosophila third-instar larvae salivary gland polytene chro- mosomes are shown in Fig. 6. Unexpectedly, NK-1, NK-3, and NK-4 genes were found to reside in neighboring chro- mosomal bands in the right arm of chromosome 3. The NK-3 and NK-4 genes reside at 93El-3, and the NK-1 gene resides at 93E3-5. When two probes, one for NK-1 and one for NK-3, were added to the same in situ hybridization reaction mixture, two labeled chromosomal bands were obtained at 93El-5 that were separated only slightly. However, the NK-2 gene resides in the lCl-5 region of the X chromosome. In Fig. 6B, the relative positions of the NK-3/NK-4 and NK-1 genes are shown correlated with the chromosomal bands in the 93E region of chromosome 3 in Bridges' revised map of chromosomal bands (24). These results show that NK-1, NK-3, and NK-4 comprise a cluster of homeobox genes. Either NK-1, NK-3, or NK-4 genes may be the same as torso-like, a maternal effect gene that resides at 93E and is one of the ensemble of genes that determine the anterior- posterior pattern of the embryo (2). The torso-like gene and four other genes are required for the formation of both the anterior and posterior terminal, unsegmented portions of the embryo (the acron and telson) (2). Another candidate is paired gene 9, which is thought to contain repetitive alternating codons for histidine and proline termed a paired repeat [also found in paired (25) and bicoid KG GCC CRT KC CTIA CAC AAC ARC RAT AAT AAT ACG RCA AAC AAC -160 Thr Ala HIS Ala Leu "1sl4s.n Rsn Asr hsn ASK Thr Thr Asn Asn -54 RAT AAC CAC AK CTG RAG GCC GAG GGG ATC AAC GGA GCA GGC AGT -115 m HIS Ser Leu Lys hla Glu Giy Ile Asn Gly Ala Gly Ser -39 GGT CAC GAC GRT AK CTC AAC GAA GAT GGC ATC GAG GAG GAT ATC -70 Gly HIS iAsp Ser Le'x ~sn m Gly Ile klu Glu Asd Ile -24 GAC GAC GTG GAC GRC GCC GAC GSC XT GGC GGC GGG GAT GCA RAT As s 7 "al 1-d Ala m i;y Ser Gly Gly G1y vispl Ala As" BamH -1 +1 GGA TCC GAC GGT CTG CCR ART RRG AM CGG ARG CGR CGA GTC CTG Gly Serm Gly Le" Fro hsn Lys Lys Arg iys Arg Arq Val Leu TTC ACC AAG GCG CR.4 ACR TAT GA; CTG GM CGT CGG TTT CGA CEA Phe Thr Lys Ala Gin Thr Tyr Glu Le.2 Slii Rrg Rrg Fhe Arg Gln CAA CGT TRC TTG AGT GCC 'CCG GAA CGC GAG CAC CTG GCC AGT TTG Gin Arg Tyr Leu Ser Ala Frc Glu Arq Gli; HIS Leu Ala Ser Le" ATC CGC CTG ACG CCG ACC CAG GIG AAG ATC TGG TTT CAA AAC CAT Ile Arq Leu Thr Pr3 Tnr Gin Val Lys Ile Trp Fhe Gin As" tils CGC TAC AAG ACG AA: CG:: GCG CAA AAC GAG AAG GGC TAC GAG GGT &rg Tyr Lys Ttr Lys AIM A.la Gln As" Glu Lys Gly Tyr Glu Gly CAT CCT GGT CTA CTG CAC GGC CAT GCC ACC CAT CCG CAT CAC Ccc 246 IHis Pro] Giy L~'J Leu a Gly a Ala Thr kis Pro His His Prd 82 AGT GCC CTG XA ?CG CCC G-C GGG TAG CCGYTCCRGTTCTGGTGAGGAAC Ser Ala Leu m Ser m 'Jai Gly '** GGAAAGeCCTGCTTGGGCGATAGT;CCF.ARCTGGGAGCCGACTGCG?CTCCGTGTCATC AGCCACCGCCACCGCCATGCAGAATGeCSC`CCGC~CATCACTTGGTTGCCCTAAATGGAG Pst I CGGCCGCCTATCAACATGCCGUUZ -25 -9 21 7 6 6. 22 111 37 156 52 201 67 291 90 341 400 440 FIG. 3. Nucleotide sequence and deduced amino acid sequence of the homeobox region of NK-2 genomic DNA. The homeobox domain is enclosed in a large box. Repetitive Asn residues are enclosed in a box. The acidic amino acids enclosed in boxes before the homeodomain comprise an acidic region of NK-2 protein. (25) genes] and a homeobox, which resides at 93El-2 and was cloned by Frigerio et al. (25). Elsewhere, we will show that NK-1 contains a paired repeat (unpublished data): however, it is not known whether NK-1 is the same as paired gene 9. It should be noted that binding sites for polycomb protein have been detected at 93El-4 (27). One of several candidates for the NK-2 gene is twisted, discovered by Demerec et al. (28), which is located between lC-5 and 2C-10. The abdo- mens of adult TV' mutants, viewed from behind. are rotated -30oC clockwise. However, further work is needed for the identification of NK-1, NK-2, NK-3, and NK-4 genes. Homeobox Homology. The deduced amino acid sequences of the homeobox domains of NK-1, NK-2, NK-3, and NK-4 are shown in Fig. 7 and are compared with the 23 Drosophila homeobox sequences that have been reported thus far. NK-2, NK-3, and NK-4 homeoboxes are more closely related to one another (59-66% homology) than they are to other Droso- phila homeoboxes. The maximum homology to a previously reported homeobox is to muscle segment homeobox (msh) (29) (54% homology). In order of decreasing homology, the homeobox of NK-1 is most closely related to msh, Zen-l, NK-3 and Dll homeoboxes. NK-1, lab, Dll, and Abd-B genes may have originated from a common precursor because each gene contains an intron between the codons for the 44th and 45th homeobox amino acid residues. It has been suggested that homeobox proteins bind to DNA via a helix-turn-helix motif in the homeodomain and that amino acid residues 42, 43, and 47 in the third a-helix of the homeodomain interact with nucleotide residues in the major groove of DNA and determine, at least in part, the nucleotide sequence recog- nized (29-31). Since NK-1, lab, Abd-B, and Dll genes each contain an intron between the codons for the 44th and 45th homeobox amino acid residues, the part of each gene that is thought to encode the DNA recognition site of the corre- Pst I 'C;G CAG TRT TAT GCG GCG Xc; ATG GA: AAC AAT RAC CAC CAT CAC -31r LCJ Gin Tyr Tyr Ala Ala A;a ME: Asp Asn Asr. As" H15 H2s His -hi CA; GCA ACG GGC ACA TCS AAC XC AGT GCC KC GAC TAC ATG CAC -z7: Gin Ala Thr Gly Thr Ser ARC Ser SPI A13 Ala Asp Tyr Met Gin -5: BornHI CGC A.&A TTG KC TAT TX GGA TCC AC: CT: GCT XT ZCT TTG GAC -::; Arq Lys Le" Ala Tyr Phe G1j Ser Thr Le" Ala Ala Pril re" Asp AT; AGA cGC TGC AC:: AGZ AK SAT TCC v G GTA.AGTAACXCA:GAAAT`-A -I'- Yet Arq Arq Cys Thr Ser Asr As? Ser A -2t ACGCCATTCAGGcTcTAATIZGACTCTG~~AAGAAC^CCCT?TTGT -::f v ATRGGA_GTA~GCTAACTTT_GGTAAT~~~C'~~T~~*'~A~ AC TGC SAC TCA 'C:A -^' sp Cys Asi; Ser Fro -: CCG CCA TTG AC, AGT 'ICC CC. TCC GAG TCG CC: STA TCZ CAC GAC -!" Prc Fro Leu Se: Ser Ser Pro Se: G1> Ser Fro Leu Ser 41s Asp -- CGG TCG CGT GCC GCC TTC AGC ^~ - Arq Ser Arq Ala A:a Fhe Ser, ' AC GCC CRS G-C TTC GAG TT; GAC, CGC CG: 'XT XC CM CAG ' 7: ':GCl Slu Lerl Glu Arg Arg Phe Ala Sin Gin Arg ;i AC TTG TCC GGT CCG GAA ZGC AX GAG RTG GCC AK RGC CTG CGC '3; !.eu Ser Gly Frs GIL Arg Ser Glu Me: Ala Lys SPL Leti Arq ' AAG ATS TGG TTC ZM AAC CGZ CGZ TAC lL- Thr Glu Thr Gin b'ril Lys ile :rp Phe Sin Asn Arq Arq Tyr c AA; AC': AAG CA: CAC GAG XC XC CT TTG ; Gin 91s G11: Ala Ala Leu ie" GG: GCC AGC AAG AGG GTT CCC GTC CAA GTC TT; GTG CGA GAG GAT 2:: Gly Ala Ser Lys Arq "al Prc '/a! Gin "al Leu "al Arg Glu ASP GSC AX ACC ACC TAC GCT CAC ATG GCT GCT CCC GGT GCT GGA CAC ';I G;y Ser Thr Thr Tyr Ala H15 Yet Aid Ala Pro Gly Ala Gly ,416 " Pst ! SGC CTC GAT CCC GCC CTG ATC AAC ATC TPC CGC CAT i`AG CTG LAG ;;.; Giy Le" Asp Pro Ala Leu Ile As" Ile Tyr Arq His GA" 'eu Gin FIG. 4. Nucleotide sequence and deduced amino acid sequence of the homeobox region of the NK-3 gene. The homeobox is enclosed in a box. Arrowheads represent exon-intron junctions. I - Biochemistry: Kim and Nirenberg Proc. Natl. Acad. Sci. USA 86 (1989) 7719 spending protein is interrupted by an intron. Further work is needed to determine whether the specificity of DNA recog- nition by NK-1 protein is altered by alternative splicing. Amino acid replacements that alter the 42nd or 43rd amino acid residues in the homeobox are of special interest since they may determine part of the nucleotide sequence that is recog- nized by the homeobox protein. The 42nd homeobox amino :Lcid residues of NK-2 and NK-4 are proline and alanine, I espectively. The unspliced form of Saccharomyces cerevi- :,,ue mating-type factor a-l has proline at this site (29); how- ever, neither proline nor alanine has been found at this site in any metazoan homeobox protein, A proline residue would not be expected to be part of an a-helix, unlike alanine or glutamic acid residues, which promote a-helix formation. The 43rd homeobox amino acid residue of NK-1, NK-2, NK-3, and NK-4 is threonine; however, the only other homeobox pro- teins that contain threonine at this site are msh, Dll, lab, and ro in Drosophila and Hox 1.6 and Hox 7.1 in the mouse. The amino acid sequences of most or all of these homeobox domains share other unusual features [for example, see alanine ( ilth amino acid residue), lysine or arginine (19th residue), glutamic acid (30th residue), tyrosine (54th residue), and serine or threonine (56th residue)]. The presence of the same or similar unusual amino acid replacements in most or all ofthese homeodomains provides additional evidence that the newly discovered homeobox genes are related to one another. The combined use of probes 121 and 125, which corre- spond to the overlapping hexapeptides shown at the top of Fig. 7, should detect only those homeobox genes that encode the amino acid sequence Gln-Val-Lys-Ile-Trp-Phe-Gln-Asn `!s residues 44-51 of the third a-helix of the homeodomain, vhich is part of the putative nucleotide sequence recognition site of homeobox proteins. Only Zen-l, zen-2, lab, and cad, in addition to the homeobox genes described in this report would be expected to give positive autoradiographic signals with both 121 and 125 probes. It is uncertain whether msh, D/l. and Abd-B genes would give positive signals with 121 and 125 probes because both Dll and Abd-B genes contain introns UK-3 4-k UK-1 UK-b (93E3-S) (93El-3) FIG. 6. (A) In situ hybridization of genomic DNA probes for NK-1 (first panel), NK-4 (second panel), NK-3 (third panel). NK-1 and NK-3 (fourth panel), and NK-2 (fifth panel) to Drosophila polytene chromosomes. A-F and vertical markers represent chro- mosomal band subdivisions in the 93 A-F region of the right aim of the third chromosome. The DNA probes hybridize to the following locations: NK-1, 93E3-5; NK-4, 93El-3; NK-3, 93El-3; NK-1 and NK-3,93El-5; NK-2, lCl-5. Arrowheads indicate labeled chromo- somal bands. (B) The approximate locations of the NK-3, NK-4, and NK-1 genes are indicated on Bridges' revised map of chromosomal bands (24). of unknown sequence between codons for the 44th and 45th homeobox amino acid residues, and it is not known whether the msh gene contains an intron at this site. Probe 125 hybridizes to the NK-1 gene because the 3'-terminal nucle- otide sequences of both exon 1 and intron 2 are CAG. Both Zen-f and Zen-2 genomic DNA were detected and cloned with probes 121 and 125 in addition to NK-1, NK-2, NK-3. and NK-4, but lab and cad genes were not detected. 1 `:~n';,:i~C?nnG;nTTCrRC,T:TCT.~. rTTRTTS;.~GT'TT'^nF,,~~~~~~*~~,C,,~GT~ .j : : Y" ;xT~,AI;;:,P,AcI;~. ~CnT~GTTCThGTTC_TT;T-? :