Structure and Regulation of the Salivary Gland Secretion Protein Gene Sgs-1 of Drosophila melanogaster

Corresponding author: Günther E. Roth, Institut für Genetik, Freie Universität Berlin, Arnimallee 7, 14195 Berlin, Germany., groth{at}genetik.biologie.fu-berlin.de (E-mail)

Communicating editor: S. HENIKOFF

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The Drosophila melanogaster gene Sgs-1 belongs to the secretion protein genes, which are coordinately expressed in salivary glands of third instar larvae. Earlier analysis had implied that Sgs-1 is located at the 25B2-3 puff. We cloned Sgs-1 from a YAC covering 25B2-3. Despite using a variety of vectors and Escherichia coli strains, subcloning from the YAC led to deletions within the Sgs-1 coding region. Analysis of clonable and unclonable sequences revealed that Sgs-1 mainly consists of 48-bp tandem repeats encoding a threonine-rich protein. The Sgs-1 inserts from single clones are heterogeneous in length, indicating that repeats are eliminated. By analyzing the expression of Sgs-1/lacZ fusions in transgenic flies, cis-regulatory elements of Sgs-1 were mapped to lie within 1 kb upstream of the transcriptional start site. Band shift assays revealed binding sites for the transcription factor fork head (FKH) and the factor secretion enhancer binding protein 3 (SEBP3) at positions that are functionally relevant. FKH and SEBP3 have been shown previously to be involved in the regulation of Sgs-3 and Sgs-4. Comparison of the levels of steady state RNA and of the transcription rates for Sgs-1 and Sgs-1/lacZ reporter genes indicates that Sgs-1 RNA is 100-fold more stable than Sgs-1/lacZ RNA. This has implications for the model of how Sgs transcripts accumulate in late third instar larvae.

IN Drosophila melanogaster the salivary glands of third instar larvae produce a sticky secretion by which the larvae attach themselves to a solid surface prior to puparium formation. The secretion consists of glycoproteins (BECKENDORF and KAFATOS 1976 ; KORGE 1977 ) that electrophoretically separate into at least eight bands. According to the mobility of the proteins their corresponding genes have been named Sgs-1–8 (salivary gland secretion genes 1–8; KORGE 1975 , KORGE 1981 ). The electrophoretic mobility of the secretion proteins varies between different stocks of D. melanogaster. As this variation is due to different alleles, it has been used to genetically map the genes (KORGE 1975 ). The Sgs genes are coordinately activated, which leads to the formation of puffs that regress when the titre of the steroid hormone 20-hydroxyecdysone increases at the end of the third larval instar (BECKER 1959 ; ASHBURNER 1972 ). All of the five Sgs genes cloned to date are heavily transcribed during the third larval instar in salivary glands only (reviewed in LEHMANN 1996 ). Thus, this group of genes provides an excellent opportunity to analyze the mechanisms by which tissue-specific and temporarily restricted gene expression is controlled. Presumably, the expression of all Sgs genes is controlled by the same trans-acting factors, such as possibly by the ecdysone receptor (LEHMANN and KORGE 1995 ). However, when the 5' upstream regions of Sgs genes 3, 4, 5, 7, and 8 are compared to each other, it is difficult to detect common sequence motifs, which could serve as binding sites for common factors. To analyze the regulation of Sgs-1 and to have an additional member of this gene family available for comparison, we started out to clone the Sgs-1 gene, the last uncloned member of the most prominent Sgs genes.

Sgs-3, -4, -5, -7, and -8 have been isolated from a salivary gland cDNA library, using as probes total poly(A) RNA isolated from salivary glands of different developmental stages (MUSKAVITCH and HOGNESS 1980 ; MEYEROWITZ and HOGNESS 1982 ; CROWLEY et al. 1983 ; GUILD 1984 ). In this screen the Sgs-1 gene that had been inferred to be located at the 25B2-3 puff (VELISSARIOU and ASHBURNER 1980 ) had not been isolated, although ¹²⁵iodine-labeled poly(A) RNA from salivary glands gave a strong signal at the 25B puff in in situ hybridizations (KORGE 1980 ). Different attempts were made in our laboratory to acquire the Sgs-1 gene, including microcloning, microsequencing of the Sgs-1 protein, and in situ hybridization screening of a cDNA library, specific for genes active in salivary glands of third instar larvae (BORNSCHEIN 1994 ). None of these approaches was successful. We then cloned nearly the complete Sgs-1 gene by combining yeast artificial chromosome (YAC) and PCR technologies. Like other Sgs genes, Sgs-1 mainly consists of repetitive elements. Despite the use of different cloning strategies, parts of the transcribed region, very likely repetitive elements, could not be cloned on bacterial plasmids or bacteriophage vectors. However, the gene can be maintained on a YAC. The cause for the instability in E. coli is unknown, but very likely it was the reason for the difficulties in cloning this gene.

Analysis of the transcriptional regulation of Sgs-1 by means of P-element transformations and mobility shift assays shows that Sgs-1 is a target gene of the transcription factor FKH, encoded by the fork head (fkh) gene, which previously has been shown to control the tissue-specific expression of Sgs-3 and Sgs-4 (LEHMANN and KORGE 1996 ; MACH et al. 1996 ; LEHMANN et al. 1997 ). In addition, we identified a site that binds the factor SEBP3 (secretion enhancer binding protein 3) that is required for full transcriptional activity of Sgs-4 (LEHMANN and KORGE 1995 ). We also present evidence that Sgs-1 transcripts are 100-fold more stable than Sgs-1/lacZ reporter transcripts. This observation has an impact on the conception of how Sgs transcripts accumulate at the end of the third larval instar.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila strains:
D. melanogaster wild-type strains Hikone, Sevelen, Oregon-N, Falsterbo, Bakup, Karsnäs, Kochi-R, Canton-S, and the transformant line E5 (HOFMANN et al. 1987 ) were used.

Protein analysis:
Secretion proteins were analyzed as described (KORGE 1975 ) except that 3% acrylamide gels were used.

DNA methods:
A cosmid library of strain E5 was constructed in pWE15 (Stratagene, La Jolla, CA). Overlapping fragments from Cos1 and Cos2 were sequenced by using the Sequenase 2.0 kit (Amersham, Buckinghamshire, UK). Both strands of the Sgs-1 upstream sequence up to nucleotide (nt) -1022 were sequenced. For details of the sequencing strategy see WATTLER 1995 .

Isolation of YAC E02-38:
The Saccharomyces cerevisiae strain containing YAC E02-38 was provided by D. Hartl (AJIOKA et al. 1991 ). Cells were incubated for 2 hr at 30° in 2 ml SCE (1 M sorbitol, 100 mM sodium citrate, 60 mM EDTA pH 7.0), 50 µl ß-mercaptoethanol, and 0.1 mg Zymolyase-100T (Seikagaku Kogyo). Spheroplasts were mixed with an equal volume of 1% low-melting-point agarose in SCE and plugs were incubated in 0.5 M EDTA pH 9.0, 1% Sarkosyl, 10 mM Tris-HCl pH 8.0, and 0.05 mg/ml Proteinase K for 24 hr at 55° and stored in 0.5 M EDTA at 4°. DNA was separated by pulsed-field gel electrophoresis and YAC DNA was isolated by means of Gelase (Epicenter).

Cloning of Sgs-1 fragments in E. coli:
YAC DNA was digested by restriction enzymes and fractionated on agarose gels, and DNA was eluted from the appropriate region. Southern hybridization confirmed that the eluted DNA contained sequences homologous to cDNA-fr2 (see RESULTS). Restriction fragments, vectors, and E. coli strains used for subcloning are given in Table 1. In Experiments 1–3 (Table 1) 5000 transformants were screened using cDNA-fr2 as a probe. DNA from 20 clones was analyzed. Most inserts had the size expected; none of them hybridized to cDNA-fr2. Experiment 7 yielded pYS-38 (Fig 2). pLEX5R is a derivative of vectors described in DIEDERICH et al. 1994 with a temperature-dependent R1 origin with a plasmid copy number of 1 at 30°. In Experiments 4, 5, and 6 (Table 1) 1000 transformants were screened by pYS-38 insert as a probe. At 30°, pLEX5R yielded two positive clones, whose inserts had the same size as that of pYS-38 and were unstable upon propagation. In Experiment 8 a partial genomic library was constructed by using data from the analysis of Cos1 and Cos2 and of genomic fly DNA. Plaques were screened by pYS-38 insert DNA. Five positive clones contained large deletions, including most of the Sgs-1 gene. Experiment 9 yielded Cos1 and Cos2 (Fig 2).

View larger version (12K): In this window In a new window Download PPT slide   Figure 1. Map of the coding region of Sgs-1 of strain Canton-S compared to that of YAC E02-38, pYS-38, and termini of cosmids Cos1 and Cos2, and position of cDNA clones pFEG4.2 and pZ321. R, EcoRI; RV, EcoRV; H, HindIII; S, SstI; X, XbaI; R*, EcoRI sites of the cosmid linker. Initiation and stop codons and the 66-bp intron are indicated. Dotted lines designate stretches that were sequenced. Vertical lines indicate positions of 48-bp repeats detected by sequence analysis. During subcloning of pYS-38 from YAC, 2.5 kb from the EcoRI-EcoRV fragment that contains most of the Sgs-1 coding region were lost (indicated by the dashed lines). Terminal fragments of Cos1 and Cos2 contain 5' and 3' ends of Sgs-1, respectively. Alignment of cosmids and YAC according to sequence and restriction sites leads to a gap of 1 kb between the cosmids. Since the cosmids come from line E5, which has longer Sgs-1 transcripts than Canton-S (Fig 2), the real gap between Cos1 and Cos2 is ~2 kb. The sequence of the left end of pYS-38 is identical to the Cos1 sequence located downstream of the second EcoRI site. The simplest assumption is that the deletion of pYS-38 occurred within the unsequenced part of the 1.2-kb EcoRI-EcoRV fragment. The leftmost EcoRI site is located at position -44 relative to the start site of transcription. Additional 1.8 kb of upstream DNA were sequenced.

View larger version (121K): In this window In a new window Download PPT slide   Figure 2. Comparison of the SGS-1 protein band pattern from different D. melanogaster strains with RNA and DNA hybridization signals demonstrates that pYS-38 contains Sgs-1 gene sequences. (a) Secretion proteins from strains Hikone H, Sevelen S, Oregon O, Falsterbo F, transformant E5 E, Bakup B, Karsnäs K, Kochi Ko, and Canton-S C separated on an acrylamide gel. Coomassie staining. (b) Northern analysis of third larval instar total RNA (phenol/cresol method). (c) Southern analysis of genomic DNA, digested by EcoRI, from the same strains. The probe was pYS-38.

RNA analysis:
Lithium-salt method: Third instar larvae were dissected in ice-cold "Tübingen" Ringer (ASHBURNER 1989 ) and tissue collected in 100 µl 1% LETS buffer [10 mM Tris pH 7.4, 10 mM lithium-EDTA, 100 mM lithium chloride, 1% lithium lauryl sulfate (LiDS)] and frozen in liquid nitrogen. A total of 100 µl 1% LETS buffer and 100 µl phenol were added and the tissue was homogenized. After centrifugation the interphase was extracted with 0.2% LETS (LETS buffer containing 0.2% LiDS). The RNA was reextracted with phenol and precipitated with 3 volumes of ethanol and 0.1 volume of 0.25 M LiCl. The guanidine-salt method and the phenol/cresol method were used as described in KRUMM et al. 1985 and CHOMCZYNSKI and SACCHI 1987 .

TRIzol reagent method: TRIzol reagent (GIBCO BRL, Gaithersburg, MD) was used following the instructions of the manufacturer. Poly(A) RNA was purified from RNA isolated by the guanidine-salt method followed by centrifugation through cesiumtrifluoroacetate. By two cycles of oligo(dT) cellulose chromatography poly(A) RNA was enriched. For Northern blots total RNA was electrophoretically separated according to FOURNEY et al. 1988 , transferred to Nytran (Schleicher and Schuell, Keene, NH), and hybridized with ³²P-labeled DNA synthesized by random priming.

RT-PCR:
Poly(A) RNA was reverse transcribed by M-MLV RT (GIBCO BRL) and the oligonucleotide bo17 (5' ATGAGTCGAGACTCCATAAGCGGCCGCTTTACG-d(T)₁₇ 3') as a 3' adapter primer. A d(A) anchor was added by terminal transferase as an annealing site for the bo17 primer-adapter in the second strand synthesis and amplification. RNA was digested with RNaseH, RNaseA, and RNaseT1. The amplification reaction started from 1/20,000th salivary gland template cDNA with 5 pmol bo17 and 100 pmol bo16 primer (5' ATGAGTCGAGACTCCATAAGCGGCCGCTTTACG 3'). PCR products were fractionated on low-melting-point agarose and reamplified.

Isolation of cDNA clones:
cDNA clone pZ321 was isolated from an oligo(dT)-primed cDNA library, constructed from salivary gland RNA of third instar larvae of strain Karsnäs. The 5' rapid amplification of cDNA ends (RACE) system (GIBCO BRL) was used to amplify the 5' end of Sgs-1 cDNA. RNA from salivary glands of strain Oregon-N was isolated by the lithium method. The Sgs-1-specific primer Sg-5 (5'-AGAGCATGTGCACTCCG-3', position 500-484) was chosen for the first strand synthesis, and the nested primer Sgn5 (5'-GGTAGTCCTCTTGGTGGC-3', position 407-390) for amplification. After gel purification of PCR products the 341-bp fragment was subcloned into EcoRV-restricted dT-tailed pBluescript to yield pFEG 4.2 (Fig 1).

Construction of transformation plasmids and P-element transformation:
The starter plasmid for all transformation constructs consisted of Sgs-1 sequences from the EcoRI site at nt -44 to the MseI site at nt +56, linked by a BamHI linker to a 3.4-kb BamHI-HindIII lacZ fragment from E. coli, derived from pPASL (HOFMANN and LEHMANN 1998 ). The fusion contains 44 bp of Sgs-1 upstream sequences, 32 bp of untranslated leader, the first eight amino acids of Sgs-1, an arginine residue encoded by the linker, and lacZ beginning at codon eight. This fragment was ligated with its EcoRI site to different Sgs-1 upstream fragments (using restriction sites given in Fig 5) and cloned into pBluescript. To have SalI sites on both ends of the fusion fragment a second cloning step into pGem-3 was performed. SalI fragments were then cloned into the P-element vector pC20. In all injected DNAs the orientation of the fusion gene was opposed to that of the rosy⁺ gene of pC20. Thereby poly(A) addition sites are provided by rosy sequences. Base changes in C-1021-mu were introduced by PCR mutagenesis and controlled by sequencing. Germline transformations were performed by standard methods (ASHBURNER 1989 ). Host strain was Ko; S4-94; ry, which is isogenic for Sgs-1 and carries the Sgs-4-underproducer allele from strain Kochi-R. Transformants were made homozygous using strain T(2;3)Ata/CyO;TM3ry^RK. Sites of integrations were localized by in situ hybridization using biotin-labeled probes.

View larger version (78K): In this window In a new window Download PPT slide   Figure 3. The DNA elimination process acts on the 48-bp Sgs-1 repeat unit. YAC DNA was digested by EcoRI, the 6-kb region was gel purified and cloned into Zap, and clones were screened by pYS-38 insert. DNA from positive clones ( 1–7) was digested with EcoRI + EcoRV, run on a 1.2% agarose gel, and hybridized with pYS-38 insert DNA. All DNAs show the 2.5-kb EcoRI-EcoRV fragment that carries the 3' end of Sgs-1 (see Fig 1). In addition, some clones contain a dominant fragment similar in size to the 1.25-kb EcoRI-EcoRV fragment of pYS-38 ( 3, 4, and 6). The inserts of the other clones are heterogeneous in lengths, appearing as a ladder. The difference in length between adjacent bands of the ladders was determined to be close to 48 bp (46, 47, and 48 bp in independent determinations). The original autoradiography shows faint ladder bands also in 3, 4, and 6.

View larger version (59K): In this window In a new window Download PPT slide   Figure 4. Sgs-1 RNA is detected only in salivary glands of third instar larvae. Northern analysis of different stages of development and from different tissues of strain Oregon. (a) 1, 0- to 20-hr embryos; 2, first instar larvae; 3, second instar larvae; 4, third instar larvae; 5, white prepupae; 6, brown prepupae; 7, pupae; 8, adult female flies; 9, adult male flies. (b) 1, third instar larvae; 2, salivary glands; 3, third instar larvae with glands removed. RNA was isolated by TRIzol and run-on formaldehyde gels. Filters were hybridized with pYS-38 insert DNA and then with the tubulin (Tub) probe pTU-56. Comparing the signals to those in Fig 2B shows less RNA degradation. Besides the main band slower migrating RNAs are detected that disappear on glyoxal gels (data not shown). They probably contain Sgs-1 RNA with secondary structures not removed by formaldehyde.

View larger version (32K): In this window In a new window Download PPT slide   Figure 5. Sgs-1-lacZ fusion genes used for germline transformations and mean of the ß-galactosidase activity in the transformed lines in percent (Table 2). The activity of C-1806 lines was set as 100%. All transgenes contain 32 bp of untranslated leader and 24-bp coding sequence of Sgs-1. Restriction sites used for cloning and their position relative to the start site of transcription are given at the top (B, BamHI; D, DraI; Hi, HindII; R, EcoRI; Rs, RsaI; S, SalI; Sa, Sau3AI; X, XmnI). Numbers specifying the transgenes give the length of the Sgs-1 upstream sequence. C-1021-mu was mutated at positions indicated by * (Fig 6). Arrowheads indicate positions of putative FKH binding sites tested by bandshifts; only the site indicated in C-1021-mu was positive (Fig 8). In C-1021-del the region between nt -271 and nt -431 bp was deleted.

LacZ expression assays:
Histochemical staining assays on dissected larvae were performed as described (RAGHAVAN et al. 1986 ). ß-Galactosidase activity was measured fluorometrically as described in HOFMANN et al. 1991 .

Run-on analysis:
A total of 100 pairs of salivary glands was incubated in 40 µl of 35 mM Tris pH 8.0, 17 mM MgCl₂, 270 mM KCl, 18 µl of 10% N-lauryl sarcosine, 210 µl of Ringer solution, 25 µl of ³²P-UTP (6000 Ci/mmol), and 0.75 µl of each ATP, CTP, and GTP (100 mM) at 30° for 30 min. The mixture was incubated with 5 units of RNase-free DNase I for 10 min at 37°. After addition of 36 µl of 10% SDS, 50 mM EDTA, 100 mM Tris, pH 7.4, and incubation with Proteinase K (10 µl of 20 mg/ml) the mixture was extracted twice with phenol/chloroform, ethanol precipitated, resuspended in 100 µl TE, and spun through a Sephadex G50 column. The eluted RNA was added to 500 µl of 10 mM TES (Sigma, St. Louis), 1% SDS, 10 mM EDTA, 300 mM NaCl, 1x Denhardt's solution, 0.25% milk powder, 200 µg/ml E. coli RNA, and hybridized with filter-bound DNA at 65° for 18 hr. Filters were washed at room temperature in 2x SSC and incubated at 37° for 30 min with 10 mg/ml RNase A in 2x SSC. Filters were rinsed in 1x SSC/0.1% SDS, washed at 65° in 0.1x SSC/1% SDS, and exposed at -80°. Signals were analyzed densitometrically. A total of 1 µg of the following DNAs was dot blotted: 3.3-kb HindIII-BamHI fragment of pPASL (lacZ); 2.4-kb SalI fragment of paDm 2023 (Sgs-3); 2.0-kb EcoRI fragment of pOR6A (Sgs-4); 2.0-kb NotI-HindIII fragment of Cos1 (Sgs-1); linearized pBluescript.

Mobility shift DNA-binding assay:
Mobility shift assays were carried out as described (LEHMANN and KORGE 1995 ). Double-stranded oligonucleotide probes (Fig 8) with 5'-GGGG overhangs were labeled with [-³²P]dCTP by a fill-in reaction. The anti-FKH antiserum was a gift from P. Carrera and H. Jäckle. Anti-FKH antibodies and antibodies from a normal serum were purified prior to use (LEHMANN and KORGE 1996 ). Anti-USP antibody AB11 was a gift from D. L. King and F. C. Kafatos. Competitor oligonucleotides (LEHMANN and KORGE 1995 , LEHMANN and KORGE 1996 ) were added at a 10-fold molar excess over the probe. The nucleotide sequences are available at GenBank under accession nos. AF156227 and AF156228.

View larger version (12K): In this window In a new window Download PPT slide   Figure 6. Comparison of the upstream regions of Sgs-1 and Sgs-3. Sequences underlined once are important for expression of Sgs-3 and bind salivary gland nuclear protein (GEORGEL et al. 1991 , GEORGEL et al. 1993 ). Double-underlined sequences show similarity to sequences important for expression of Sgs-3, contained in its proximal enhancer at position -93 and -62, respectively (TODO et al. 1990 ). Sequences in dark or light gray are identical between Sgs-1 and Sgs-3. The sequence between the arrowheads was used in the gel shift experiment (Fig 8). Asterisks indicate core motive important for FKH protein binding. GC-rich sequences specified below the Sgs-1 sequence replace the three AT-rich elements in the construct C-1021-mu (Fig 5).

View larger version (44K): In this window In a new window Download PPT slide   Figure 7. An Sgs-1-lacZ transgene of type C-1806 is transcribed at a rate similar to that of the endogenous Sgs-1 gene, but its transcripts are unstable. (a) RNA dot blot. Different relative amounts of total salivary gland RNA were loaded on the filters. (b) Northern analysis. Third larval instar poly(A)+ RNA was first hybridized with a lacZ probe (exposure time 24 hr) and subsequently with an Sgs-1 probe (exposure time 30 min). There was no lacZ signal in the fraction of poly(A)- RNA. (c) Run-on analysis. DNA fragments encoding different genes and vector DNA as a control were loaded on the filters in duplicates and hybridized with in vivo-labeled nascent RNA. The Sgs-4 signal is weak due to the Sgs-4-underproducer allele of strain Kochi in the transformed lines. The Sgs-1 signal is almost as strong as that of Sgs-3; the lacZ signal is ~50% of that of Sgs-1. The steady-state level of lacZ RNA is at least 100-fold less than that of Sgs-1 RNA; nascent lacZ RNA, however, is only about 2-fold less.

View larger version (51K): In this window In a new window Download PPT slide   Figure 8. The Sgs-1 regulatory region contains binding sites for FKH and SEBP3. (a) Probe sequences and positions with respect to the Sgs-1 transcription start site. Core motifs important for FKH domain binding are underlined and the E-box in probe II is printed in bold. (b) Binding of nuclear extract proteins from third larval instar salivary glands to probes I and II was analyzed by a mobility shift assay in the absence or presence of the indicated antibodies or competitor oligonucleotides. Anti-FKH antibodies (FKH; LEHMANN and KORGE 1996 ) interfere with the formation of the fast-migrating complex formed by probe I, while antibodies from normal serum (NS) or an anti-USP antibody (AB11) do not react with this complex. An oligonucleotide containing the FKH binding site of the Krüppel gene (KR; KAUFMANN et al. 1994 ) competes for formation of the complex, and two point mutations of the core motif abolish the ability to compete (KRM). Likewise, neither an oligonucleotide containing a strong SEBP3 binding site of Sgs-4 (O5) nor a mutant form of this oligonucleotide carrying two point mutations that interfere with SEBP3 binding (O6; LEHMANN and KORGE 1995 ) competes with probe I. Probe II forms a complex that migrates slightly faster than the slowly migrating complex (SMC) that is also formed with probe I and is due to nonspecific protein binding (LEHMANN and KORGE 1996 ). Formation of this complex is disturbed only by O5. The mutation of the SEBP3 binding site in O6 abolishes the ability to compete, indicating that this complex is formed by SEBP3.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

YAC E02-38 contains sequences complementary to salivary gland cDNA:
By means of variants of the SGS-1 protein the Sgs-1 gene had been mapped to the second chromosome between 25A3 and 25D2 (VELISSARIOU and ASHBURNER 1980 ). Since the bands at 25B2-3 form a puff during the secretion synthesis phase, it was concluded that this is the locus of Sgs-1. Because the yeast artificial chromosome YAC E02-38 contains 200 kb from the 25B1-2 to 25B11 region (AJIOKA et al. 1991 ) it was a possible source for the Sgs-1 gene. Since ¹²⁵-iodine-labeled poly(A) RNA isolated from salivary glands hybridizes to the loci of some of the Sgs genes, including the 25B region (KORGE 1980 ), we used an RT-PCR approach to detect Sgs-1 sequences in YAC DNA. Poly(A) RNA from salivary glands was used for cDNA synthesis and amplification. The cDNA was ³H-labeled and hybridized to polytene chromosomes. Most of the Sgs gene loci were labeled, including the presumed Sgs-1 locus (data not shown). The cDNAs were size fractionated and again hybridized to polytene chromosomes. Since the fraction >2 kb (cDNA-fr2) showed the strongest signal at 25B2-3 it was hybridized against YAC DNA: single hybridization bands were obtained after digestion by EcoRI (6.3 kb) or HindIII (4.6 kb) or by EcoRI + HindIII (4.3 kb; data not shown).

YAC DNA fragments are not stable in E. coli:
We soon realized that subcloning of these fragments was not trivial. Different plasmid vectors and E. coli hosts were used without success (see MATERIALS AND METHODS and Table 1). When we used Zap II to clone DNA from the 6.3-kb region of EcoRI-digested YAC DNA, positive plaques were obtained by screening with cDNA-fr2. Insert DNA from 12 clones was analyzed, both from plasmid DNA isolated by in vivo excision and infection of XL-1 cells, as well as from DNA, isolated directly. Surprisingly, in all clones the size of the EcoRI insert was ~3.8 kb. The insert of one clone, pYS-38, was hybridized against restricted YAC DNA: it detects a 6.3-kb EcoRI fragment, a 4.6-kb HindIII fragment, and a 4.3-kb EcoRI-HindIII fragment (data not shown, but see Fig 1), as does the cDNA-fr2 probe. In addition, fragments of 2.1 kb and 1.9 kb appeared in the HindIII and the HindIII + EcoRI digest, respectively. Thus, in the course of the subcloning, 2.5 kb of the 6.3-kb EcoRI fragment was deleted (Fig 1).

YAC subclone pYS-38 contains Sgs-1 sequences:
The SGS-1 protein polymorphism was used to analyze whether pYS-38 contains Sgs-1 sequences. Nine D. melanogaster stocks were selected, showing either three SGS-1 protein bands (Hikone), two bands (E5, Karsnäs), one band (Samarkand, Oregon, Falsterbo, Kochi-R, Canton-S), or no band at all (Bakup) in the gel (Fig 2A). Third larval instar RNA and genomic DNA were hybridized with pYS-38 (Fig 2B and Fig C). The hybridization patterns of both Northern and Southern analyses are strikingly similar to the SGS-1 protein pattern of the different stocks. The patterns correspond to each other in six of the nine stocks (Hikone, H; E5, E; Karsnäs, K; Samarkand, S; Oregon, O; Canton-S, C). In the stocks Falsterbo (F) and Kochi-R (Ko), RNA and DNA analyses show a major band that in its mobility corresponds to that of the protein and an additional minor band. In the case of Falsterbo (F) it is very likely that this allele was underrepresented in those animals from which the secretion was prepared. In the case of Kochi-R (Ko) the band is detected only in the DNA analysis, running close to the major band, and possibly was unresolved by both protein and RNA analysis. Remarkably, the stock Bakup (B) that apparently produces no SGS-1 protein does show a band in the Northern analysis; it is the longest RNA (6.5 kb) of the stocks analyzed. Despite these inconsistencies, from the overall picture it is obvious that pYS-38 contains sequences of the Sgs-1 gene. The variation in the electrophoretic mobility of the proteins is caused by Sgs-1 alleles of different lengths, which is reflected by EcoRI fragments of different sizes.

Genomic organization and partial sequence of Sgs-1:
By the use of subfragments of pYS-38 and of Cos1 and Cos2 as probes in Southern analyses of genomic and YAC DNA and in Northern experiments we found that Sgs-1 coding sequences are all represented in the left-hand 1.8-kb EcoRI-HindIII fragment of pYS-38. This fragment derives from a 4.3-kb EcoRI-HindIII fragment of the YAC, parts of which were deleted during the cloning of pYS-38 (Fig 1). Screening of a salivary gland cDNA library with pYS-38 resulted in the isolation of cDNA clone pZ321. The sequence of pZ321 positions the 3' end of Sgs-1, as shown in Fig 1. Since we had not obtained clones longer than pZ321 we selected gene-specific primers from the 3.4-kb EcoRI fragment of Cos1 and amplified the 5' terminal region of the Sgs-1 RNA by means of a 5'-RACE system. This led to cDNA clone pFEG 4.2 (Fig 1). Sequence analysis of cosmid and cDNA clones revealed the structure of Sgs-1: the start site of transcription lies within the 139-bp EcoRI fragment. The transcript contains an untranslated leader sequence of 32 bp followed by the first coding region of 30 bp and an intron of 66 bp. Reading 5' to 3', the deduced amino acid sequence of the second exon shows an increasing frequency of threonine residues. The terminal sequences of both Cos1 and Cos2 and also the cDNA clone pZ321 contain a series of a 48-bp repetitive unit that codes for 16 amino acids with a threonine content of 56%. The amino acid sequence of the repetitive units is not strictly conserved. A representative example has the following sequence: T S T S R P T T T T P R S T T T. The array of repeats ends about 450 bp 5' to the stop codon. A total of 23 repeats have been found by sequence analysis.

Between the different D. melanogaster strains the transcript size varies from 4 kb to 6.5 kb (Fig 2). It is very likely that this variation is caused by different numbers of repeats. This assumption is supported by the observation that the size difference of the alleles lies within the EcoRI-EcoRV fragment (Fig 1) that contains the repeats (data not shown).

In situ hybridization to polytene chromosomes using pYS-38 as a probe shows that Sgs-1 is located at 25B1-3, the site which shows puffing during the secretion synthesis phase. Sequence comparison of the transcribed region of Sgs-1 with that of the other secretion genes reveals a relationship to Sgs-3, -7, and -8 but not to Sgs-4 (see DISCUSSION).

The uncloned sequences consist of repetitive elements similar to the 48-bp repeats:
Sequence analysis of pYS-38 and of Cos1 DNA had shown that most of the 48-bp repeat units contain Sau3A cleavage sites. By using the size of the EcoRI-EcoRV fragment of pYS-38, and the Sau3A restriction pattern as well as the pattern of hybridization when using a 96-bp dimer fragment as a probe, we determined the number of 48-bp repeats in pYS-38 to be eight (data not shown). This was used as a measure to estimate the number of repeats in YAC DNA. YAC and pYS-38 DNA were digested by EcoRI + EcoRV and hybridized with the 96-bp dimer fragment. The amount of loaded DNAs was normalized by means of the 1.5-kb EcoRI-XbaI fragment of pYS-38 as a probe. Densitometric analysis showed that the 3.6-kb EcoRI-EcoRV fragment of the YAC contains about 10 times more sequences hybridizing to the 96-bp fragment than pYS-38 (data not shown), which is equivalent to 3840 bp (48 x 8 x 10). This obviously is an overestimation; however, the result supports the hypothesis that the uncloned sequences consist of elements similar to the 48-bp repeat unit.

This interpretation is supported by a second independent observation. To analyze more clones containing YAC DNA, we repeated the experiments that had led to the isolation of pYS-38 and analyzed seven new clones. The size of the EcoRI-EcoRV fragment that carries the 48-bp repeats was determined directly from lysates without subcloning. Fig 3 shows that the insert DNA of all clones is heterogeneous in length, appearing as a ladder on an agarose gel. The difference in size between the bands of the ladder equals the length of one repeat unit (48 bp). This strongly suggests that the process that leads to the elimination of DNA acts on the Sgs-1 repeat unit.

Sgs-1 RNA is detected only in the secretion cells of salivary glands of third instar larvae:
The Sgs-1 gene is expressed exclusively in salivary glands of third instar larvae. In Northern analysis of total RNA isolated from various developmental stages and from adult flies, using pYS-38 as a probe, a strong signal is obtained only with RNA from third instar larvae and a weak signal is detectable in white prepupae (Fig 4A). The signal is restricted to salivary glands; it is not detectable in RNA from larvae after removal of the glands (Fig 4B). This result is confirmed by in situ hybridization with tissues of third instar larvae. Sgs-1 expression is found only in the posterior—not in the anterior—cells of the salivary gland (data not shown).

A total of 1 kb of Sgs-1 upstream sequence is sufficient for proper expression of a reporter gene:
To localize cis-regulatory sequences of Sgs-1, upstream fragments were fused with the E. coli ß-galactosidase gene and stably transformed into D. melanogaster (Fig 5). Gene expression in transformants was analyzed by ß-galactosidase staining of third instar larvae and by measuring ß-galactosidase activity (Table 2). Besides construct C-44, which does not show any staining or measurable enzyme activity, all constructs lead to correct expression, i.e., staining exclusively in salivary glands of third instar larvae. Table 2 shows that there are considerable differences in Sgs-1 expression between lines transformed with the same construct. This is likely due to position effects. Comparing the ß-galactosidase activity of the lines carrying shortened upstream sequences, three groups with clearly differing activities are observed: (i) lines that carry C-1806 and C-1021 show the highest activity. The expression of C-1806 with the longest upstream fragment was taken as 100%. (ii) Shortening of the upstream sequences to nt -623, nt -369, and nt -271 leads to a 4- to 10-fold reduction of expression. (iii) Shortening to nt -182 leads to a further 10-fold decrease; only a few salivary gland cells are stained in these lines.

By comparing the Sgs-1 upstream sequence with that of the other Sgs genes, we detected a region similar to the distal promoter region of Sgs-3 (Fig 6). Deletion or replacement of those sequences led to a 95% reduction of Sgs-3 expression (GEORGEL et al. 1991 , GEORGEL et al. 1993 ). By mobility shift assays two sequence elements that interact with nuclear proteins from salivary glands have been identified. Both these elements can be found within the Sgs-1 upstream sequence. In addition, the Sgs-1 sequence contains two elements similar to two other elements of importance for Sgs-3 transcription, located in the proximal promoter region of Sgs-3. We therefore deleted the Sgs-1 sequence between -431 bp and -271 bp in construct C-1021-del (Fig 5). This reduces ß-galactosidase expression to 13% (Table 2). Replacing these three AT-rich elements by GC elements in construct C-1021-mu leads to a similar decrease (25%). This shows that at least one of the AT-rich elements is required for full activation of the Sgs-1 promoter.

Taken together, the data indicate that several sequence elements required for full expression are distributed within 1 kb of upstream sequence. Elements sufficient for weak but specific expression are still contained in the construct C-182.

Compared to Sgs-1 RNA the Sgs-1-lacZ fusion RNA is transcribed at "normal" rates, but is unstable:
The enzyme activity of the Sgs-1-lacZ fusion containing 1.8 kb of Sgs-1 upstream sequences was taken as 100% (C-1806, Fig 5). To compare the transcriptional activity of these transgenes with that of the endogenous Sgs-1 gene we analyzed the levels of steady state mRNAs transcribed from both genes in line C-1806-25B (Table 2). Surprisingly, in dot blot analyses on total salivary gland RNA we found that lacZ RNA is about 100-fold less abundant than Sgs-1 RNA. Northern analyses of poly(A)⁺ RNA confirm this ratio and, in addition, indicate that lacZ RNA is polyadenylated as expected (see MATERIALS AND METHODS). To measure promoter strength more directly we performed run-on analyses. In two independent assays we found that nascent lacZ RNA is only ~2-fold less than nascent Sgs-1 RNA (Fig 7). Considering position effects we conclude that 1.8 kb of upstream sequences are sufficient for full expression of the Sgs-1 gene. The comparison of the results of the two types of experiments indicates that the endogenous Sgs-1 RNA is considerably more stable than the Sgs-1-lacZ fusion RNA.

The Sgs-1 regulatory region contains binding sites for known regulators of Sgs gene expression:
To test if the transcription factor FKH, which has previously been shown to control the tissue-specific expression of Sgs-3 and -4 (LEHMANN and KORGE 1996 ; MACH et al. 1996 ; LEHMANN et al. 1997 ), also binds to regions important for Sgs-1 expression, we scanned the Sgs-1 upstream region for sequences showing similarity to known FKH binding sites. From 21 candidate sequences in the nt -1 to nt -1200 region containing the core motif TNNGTNA/T, we selected four sequences for binding studies. Only one of these sequences, which is deleted in the C-1021-del lines and mutated by base exchanges in the C-1021-mu lines, proved to be an in vitro FKH target site (Fig 8, and data not shown). Surprisingly, a second putative FKH target sequence, located between positions nt -980 and nt -1000, was not recognized by FKH but was efficiently bound by SEBP3 (Fig 8). SEBP3 is a factor that binds to two sites in the upstream region of the Sgs-4 gene that are required for full transcriptional activation of this gene (LEHMANN and KORGE 1995 ). The SEBP3 binding site of Sgs-1 identified in this study shows a high similarity to the strong SEBP3 binding site of Sgs-4 located around position nt -422. In particular, both binding sites contain an E-box, suggesting that SEBP3 belongs to the basic helix-loop-helix family of transcription factors. The detection of a SEBP3 binding site in the regulatory region of another Sgs gene beside Sgs-4 suggests that SEBP3 might generally be required for the activation of Sgs genes.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The genes Sgs-3, -4, -5, -7, and -8 have been isolated from the same library of cDNAs, synthesized from poly(A) RNA obtained from larval salivary glands of the intermolt puff stage, and inserted into the plasmid vector pSC105 (WOLFNER 1980 ). In situ hybridization mapping led to the identification of these genes, but this strategy did not lead to the isolation of Sgs-1, although, as we show here, the Sgs-1 transcripts appear with the same temporal profile and are as abundant as those of Sgs-3. Different fruitless approaches have been undertaken in our laboratory to clone Sgs-1 (see Introduction). From the data presented here the main reason for the failure of those experiments seems to be the instability of Sgs-1 DNA in E. coli.

The complete Sgs-1 gene can be maintained on a YAC in yeast cells. However, all our approaches to subclone Sgs-1 gene fragments from the YAC, or directly from genomic DNA, onto plasmid or vectors in E. coli led to deletions of DNA sequences, despite using low-copy-number vectors and E. coli host strains that lack homologous recombination (XL-1) or, in addition, are defective in DNA repair systems (Sure). The latter strain has been reported to allow cloning of inverted repeats (GREENER 1990 ). Could the cloning problems be due to the synthesis of a lacZ/Sgs-1-encoded fusion protein, toxic to E. coli? This is unlikely in view of the negative results obtained also for pBR322, pBR328, and dash II, which should not produce fusion proteins.

It seems that those sequences that we cannot subclone from the YAC are related to the repeats that can be cloned, as they cross-hybridize with each other and consist, at least in part, of repeats with 48 bp in length (Fig 3). The clonable sequences do not contain inverted repeats, yet it is possible that the unclonable repeats contain such sequences. A mechanism known to lead to the elimination of direct repeats is "replication slippage." In an in vitro system it has been shown that DNA polymerase holoenzyme III of E. coli can slip between repeats, provided that a hairpin structure can form between the repeats (CANCEILL and EHRLICH 1996 ). Further progress of DNA sequencing technology, which would allow direct sequencing of Sgs-1 from the YAC, might reveal the reasons for the DNA elimination.

The structure of the SGS-1 protein is similar to that of the majority of the other secretion proteins, including those of D. virilis, LGP-1 and LGP-3 (LANIO et al. 1994 ). They all carry at their N termini a signal peptide of ~20 amino acids. SGS-1, SGS-3, SGS-4, LGP-1, and LGP-3 are highly glycosylated (BECKENDORF and KAFATOS 1976 ; KORGE 1977 ) and contain repeated amino acid segments, rich in threonine residues that are probably the sites of oligosaccharide linkage. In contrast, SGS-5 is only lightly glycosylated (BECKENDORF and KAFATOS 1976 ) and SGS-7 and SGS-8 are not glycosylated (CROWLEY et al. 1983 ). They do not contain amino acid repeats.

For two reasons it seems that Sgs-1, -3, -7, and -8 of D. melanogaster and Lgp-1 and Lgp-3 of D. virilis belong to a family of homologous genes. First, all these genes contain a single intron of ~60 bp at identical positions, following the first nucleotide of the 10th codon. Second, a comparison of their amino acid sequences using Clustal-W (THOMPSON et al. 1994 ) reveals that all six proteins contain, within a region extending 24–29 amino acids from their C terminus, identical residues at five positions and chemically similar residues at five additional sites. Sgs-4 and Sgs-5 are different from the other secretion genes: Sgs-5 does show some similarity at its C terminus, but has two introns at positions different from that of the other secretion genes, whereas Sgs-4 contains no introns, and its C-terminal amino acid sequence cannot be aligned with the other C termini. Nevertheless, it seems that the transcriptional activity of Sgs-1 and Sgs-3 and of Sgs-4 is controlled by the same set of transcription factors. Sgs-3 and Sgs-4 have previously been shown to be controlled by FKH (LEHMANN and KORGE 1996 ; MACH et al. 1996 ; LEHMANN et al. 1997 ) and the binding site for SEBP3 has been demonstrated to be essential for full transcriptional activation of Sgs-4 (LEHMANN and KORGE 1995 ). We here show the existence of binding sites for both FKH and SEBP3 within the Sgs-1 regulatory region and demonstrate their relevance for the transcriptional activation of Sgs-1-lacZ constructs. Interestingly, only one site of the four putative FKH sites tested could be shown to bind FKH in the band shift assay. However, the positions of the other sites correlate well with the functional data obtained for the transgenic flies (Fig 5). It is tempting to assume that the progressive shortening of the upstream sequence removes putative FKH binding sites. The shortest construct that is able to drive the expression of the lacZ reporter still carries a single putative FKH site at nt -51. Its removal in C-44 (Fig 5) could lead to the nonfunctional promoter. The significance of this correlation between gene function and the presence of these putative FKH sites remains to be explored further.

Most surprising to us is the observation that the steady-state level of Sgs-1-lacZ RNA is about 100-fold less than that of the endogenous Sgs-1 RNA, whereas their rates of transcription are similar. Steady-state levels were determined by Northern and dot hybridizations, transcription rates by run-on analyses. Both transcripts are polyadenylated and there is no indication for inefficient polyadenylation of Sgs-1-lacZ RNA as we detected only faint signals in the flow through obtained by the fractionation of RNA (data not shown). We therefore conclude that the Sgs-1 enhancer is contained within 1.0 kb of upstream sequence and that the difference in the amount of steady-state RNA is due to a difference in the stability of the two types of poly(A) RNA. One plausible explanation for this is that Sgs-1-lacZ RNA is unstable as it contains E. coli sequences. An alternative view is that Sgs-1 transcripts are specifically stabilized. This is supported by the observation that a rosy⁺ transgene controlled by the Sgs-4 enhancer forms strong puffs and produces xanthine dehydrogenase in salivary glands, but leads to only very low levels of steady-state rosy⁺ RNA (G. E. ROTH and A. KRUMM, unpublished results). In the extreme consequence this could mean that the dramatic increase in Sgs RNA between early and late third instar (BARNETT et al. 1990 ) is not due to an increase in the rate of transcription controlled by an enhancer element, but in fact is due to a protection of Sgs RNA against degradation, possibly by RNA secondary structure or by (a) specific factor(s). This model would be in concordance with the binary model of transcriptional enhancer action proposed recently (WEINTRAUB 1988 ; WALTERS et al. 1995 , WALTERS et al. 1996 ) which functions independently of an effect on the rate of transcription. The validity of this "stabilizer model" for the Sgs genes can be tested.

	FOOTNOTES

¹ These authors contributed equally to this work.
² Present address: Lexicon Genetics Incorporated, The Woodlands, Texas 77381.

	ACKNOWLEDGMENTS

We thank Ruth Brockmann and Madeleine Brünner for technical assistance and Anton Krumm for advice with the run-on experiments. We also thank D. Hartl for providing YAC E02-38, W. Messer for various plasmids, and P. Carrera, H. Jäckle, D. King, and F. Kafatos for antibodies. This work has been supported by the Freie Universität Berlin through an FPS grant to G.E.R., by the Graduiertenförderung through a grant to S.W., and by the Deutsche Forschungsgemeinschaft (SFB 344) through a grant to G.K.

Manuscript received February 8, 1999; Accepted for publication June 4, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AJIOKA, J. W., D. A. SMOLLER, R. W. JONES, J. P. CARULLI, and A. E. VELLEK et al., 1991  Drosophila genome project: one-hit coverage in yeast artificial chromosomes. Chromosoma 100:495-509[Medline].

ASHBURNER, M., 1972 Puffing patterns in Drosophila melanogaster and related species, pp. 101–151 in Developmental Studies on Giant Chromosomes, edited by W. BEERMANN. Springer-Verlag, Berlin.

ASHBURNER, M., 1989 Drosophila: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

BARNETT, S. W., K. FLYNN, M. K. WEBSTER, and S. K. BECKENDORF, 1990  Noncoordinate expression of Drosophila glue genes: Sgs-4 is expressed at many stages and in two different tissues. Dev. Biol. 140:362-373[Medline].

BECKENDORF, S. K. and F. C. KAFATOS, 1976  Differentiation in the salivary glands of Drosophila melanogaster: characterization of the glue proteins and their developmental appearance. Cell 9:365-373[Medline].

BECKER, H. J., 1959  Die Puffs der Speicheldrüsenchromosomen von Drosophila melanogaster.. Chromosoma 10:654-678.

BORNSCHEIN, H., 1994 Klonierung und Charakterisierung des Pre-intermolt-Gens Pig-2 und PCR-Amplifikation einer cDNA-Fraktion zur Klonierung des intermolt-Gens Sgs-1 von Drosophila melanogaster. Thesis, Fachbereich Biologie. Freie Universität Berlin, Berlin.

CANCEILL, D. and S. D. EHRLICH, 1996  Copy-choice recombination mediated by DNA polymerase III holoenzyme from Escherichia coli.. Proc. Natl. Acad. Sci. USA 93:6647-6652[Abstract/Free Full Text].

CHOMCZYNSKI, P. and N. SACCHI, 1987  Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156-159[Medline].

CROWLEY, T. E., M. W. BOND, and E. M. MEYEROWITZ, 1983  The structural genes for three Drosophila glue proteins reside at a single polytene chromosome puff locus. Mol. Cell. Biol. 3:623-634[Abstract/Free Full Text].

DIEDERICH, L., A. ROTH, and W. MESSER, 1994  A versatile plasmid vector system for the regulated expression of genes in Escherichia coli.. Biotechniques 16:916-923[Medline].

FOURNEY, M. A., I. MIYAKOWSK, R. S. DAY and M. C. PATERSON, 1988 Northern blotting: efficient RNA staining and transfer, pp. 5–7 in Focus, Vol. 10, edited by N. SASAVAGE. Life Technologies Inc., Gaithersburg, MD.

GEORGEL, P., P. RAMAIN, A. GIANGRANDE, G. DRETZEN, and G. RICHARDS et al., 1991  Sgs-3 chromatin structure and trans-activators: developmental and ecdysone induction of a glue enhancer-binding factor, GEBF-I, in Drosophila larvae. Mol. Cell. Biol. 11:523-532[Abstract/Free Full Text].

GEORGEL, P., G. DRETZEN, K. JAGLA, F. BELLARD, and E. DUBROVSKY et al., 1993  GEBF-I activates the Drosophila Sgs-3 gene enhancer by altering a positioned nucleosomal core particle. J. Mol. Biol. 234:319-330[Medline].

GREENER, A., 1990 E. coli SURE TM strain: clone unclonable DNA, pp. 5–6 in Strategies in Molecular Biology, Vol. 3, edited by A. POLLARD. Stratagene, La Jolla, CA.

GUILD, G. M., 1984  Molecular analysis of a developmentally regulated gene which is expressed in the larval salivary gland of Drosophila.. Dev. Biol. 102:462-470[Medline].

HOFMANN, A. and M. LEHMANN, 1998  The transcriptional switch between the Drosophila genes Pig-1 and Sgs-4 depends on a SEBP1 binding site within a shared enhancer region. Mol. Gen. Genet. 259:656-663[Medline].

HOFMANN, A., A. KEINHORST, A. KRUMM, and G. KORGE, 1987  Regulatory sequences of the Sgs-4 gene of Drosophila melanogaster analysed by P element-mediated transformation. Chromosoma 96:8-17[Medline].

HOFMANN, A., M. D. GARFINKEL, and E. M. MEYEROWITZ, 1991  Cis-acting sequences required for expression of the divergently transcribed Drosophila melanogaster Sgs-7 and Sgs-8 glue protein genes. Mol. Cell. Biol. 11:2971-2979[Abstract/Free Full Text].

KAUFMANN, E., M. HOCH, and H. JÄCKLE, 1994  The interaction of DNA with the DNA-binding domain encoded by the Drosophila gene fork head. Eur. J. Biochem. 223:329-337[Medline].

KORGE, G., 1975  Chromosome puff activity and protein synthesis in larval salivary glands of Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 72:4550-4554[Abstract/Free Full Text].

KORGE, G., 1977  Larval saliva in Drosophila melanogaster: production, composition, and relationship to chromosome puffs. Dev. Biol. 58:339-355[Medline].

KORGE, G., 1980 Gene activities in larval salivary glands of insects. Verh. Dtsch. Zool. Ges., 94–110.

KORGE, G., 1981  Genetic analysis of the larval secretion gene Sgs-4 and its regulatory chromosome sites in Drosophila melanogaster.. Chromosoma 84:373-390[Medline].

KRUMM, A., G. E. ROTH, and G. KORGE, 1985  Transformation of salivary gland secretion protein gene Sgs-4 in Drosophila: stage- and tissue-specific regulation, dosage compensation, and position effect. Proc. Natl. Acad. Sci. USA 82:5055-5059[Abstract/Free Full Text].

LANIO, W., U. SWIDA, and H. KRESS, 1994  Molecular cloning of the Drosophila virilis larval glue protein gene Lgp-3 and its comparative analysis with other Drosophila glue protein genes. Biochim. Biophys. Acta 1219:576-580[Medline].

LEHMANN, M., 1996  Drosophila Sgs genes: stage and tissue specificity of hormone responsiveness. Bioessays 18:47-54[Medline].

LEHMANN, M. and G. KORGE, 1995  Ecdysone regulation of the Drosophila Sgs-4 gene is mediated by the synergistic action of ecdysone receptor and SEBP 3. EMBO J. 14:716-726[Medline].

LEHMANN, M. and G. KORGE, 1996  The fork head product directly specifies the tissue-specific hormone responsiveness of the Drosophila Sgs-4 gene. EMBO J. 15:4825-4834[Medline].

LEHMANN, M., F. WATTLER, and G. KORGE, 1997  Two new regulatory elements controlling the Drosophila Sgs-3 gene are potential ecdysone receptor and fork head binding sites. Mech. Dev. 62:15-27[Medline].

MACH, V., K. OHNO, H. KOKUBO, and Y. SUZUKI, 1996  The Drosophila fork head factor directly controls larval salivary gland-specific expression of the glue protein gene Sgs-3.. Nucleic Acids Res. 24:2387-2394[Abstract/Free Full Text].

MEYEROWITZ, E. M. and D. S. HOGNESS, 1982  Molecular organization of a Drosophila puff site that responds to ecdysone. Cell 28:165-176[Medline].

MUSKAVITCH, M. A. and D. S. HOGNESS, 1980  Molecular analysis of a gene in a developmentally regulated puff of Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 77:7362-7366[Abstract/Free Full Text].

RAGHAVAN, K. V., M. A. CROSBY, P. H. MATHERS, and E. M. MEYEROWITZ, 1986  Sequences sufficient for correct regulation of Sgs-3 lie close to or within the gene. EMBO J. 5:3321-3326[Medline].

THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON, 1994  CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680[Abstract/Free Full Text].

TODO, T., M. ROARK, K. V. RAGHAVAN, C. MAYEDA, and E. MEYEROWITZ, 1990  Fine-structure mutational analysis of a stage- and tissue-specific promoter element of the Drosophila glue gene Sgs-3.. Mol. Cell. Biol. 10:5991-6002[Abstract/Free Full Text].

VELISSARIOU, V. and M. ASHBURNER, 1980  The secretory proteins of the larval salivary gland of Drosophila melanogaster: cytogenetic correlation of a protein and a puff. Chromosoma 77:13-27[Medline].

WALTERS, M. C., S. FIERING, J. EIDEMILLER, W. MAGIS, and M. GROUDINE et al., 1995  Enhancers increase the probability but not the level of gene expression. Proc. Natl. Acad. Sci. USA 92:7125-7129[Abstract/Free Full Text].

WALTERS, M. C., W. MAGIS, S. FIERING, J. EIDEMILLER, and D. SCALZO et al., 1996  Transcriptional enhancers act in cis to suppress position-effect variegation. Genes Dev. 10:185-195[Abstract/Free Full Text].

WATTLER, S., 1995 Isolierung des larvalen Sekretproteingens Sgs-1 von Drosophila melanogaster und Analyse der Regulation seiner Expression. Thesis, Fachbereich Biologie. Freie Universität Berlin, Berlin.

WEINTRAUB, H., 1988  Formation of stable transcription complexes as assayed by analysis of individual templates. Proc. Natl. Acad. Sci. USA 85:5819-5823[Abstract/Free Full Text].

WOLFNER, M. F., 1980 Ecdysone-responsive genes of the salivary glands of Drosophila melanogaster. Ph.D. Thesis, Stanford University, Stanford, CA.

This article has been cited by other articles:

				C. Cao, Y. Liu, and M. Lehmann Fork head controls the timing and tissue selectivity of steroid-induced developmental cell death J. Cell Biol., March 12, 2007; 176(6): 843 - 852. [Abstract] [Full Text] [PDF]