The non-dosage compensatedLSPI -α gene ofDrosophila melanogaster lies immediately downstream of the dosage compensatedL12 gene

The X-linked geneLSPI-α ofDrosophila melanogaster, expressed in the third larval instar, does not exhibit dosage compensation at its normal locus but does compensate when it is relocated to ectopic sites on the X chromosome. A transcription unit designatedL12, which is active in the second larval instar and capable of encoding a putative protein of 28.5 kDa, lies immediately downstream fromLSPI-α. We have determined thatL12 is dosage compensated by measuring the steady-state level of its transcript in male and female larvae. The difference in response of these two adjacent genes should be taken into consideration when models of the mechanism of dosage compensation are formulated.


Introduction
The difference in the dosage of X-linked genes present in males and females of Drosophila melanogaster is compensated by an increase in the transcriptional activity of most of these genes in males resulting in comparable levels of gene product in the two sexes (reviewed in Lucchesi and Manning 1987). X-linked genes transposed or transduced to autosomal sites often remain fully or partially compensated while autosomal genes relocated to the X chromosome often gain some level of compensation. These observations suggest that dosage compensation is mediated by compensatory elements distributed along the X chromosome.
Although the X-linked gene LSPI~c~ does not exhibit dosage compensation (Roberts and Evans-Roberts 1979;Brock and Roberts 1982), we have shown that it is capable of compensation when it is relocated to ectopic sites on the X chromosome (Ghosh et al. 1989). This observa-Correspondence to: J.C. Lucchesi tion suggests that LSPI-~ normally resides in a chromosomal domain where compensatory elements are not present or where, for some reason, they have no influence on its transcriptional activity. Therefore it seemed of interest to determine whether the genes immediately adjacent to LSPI-c< exhibit dosage compensation. Here we report the existence of a gene, L12, located immediately upstream of LSPI-~. Measurements of steady-state RNA in males and females indicate that, in contrast to LSPI-~, L12 is fully compensated.

Materials and methods
Collection of timed developmental stages. Adults from an Oregon-R strain were maintained at room temperature in population cages under a 10 h light/14 h dark cycle. Eggs were collected on trays of standard cornmealmolasses-agar medium overlaid with a yeast paste. Embryos from 0-24 h of age, in 3 h increments, first (22-46 h), second (40-70 h) and third (70 120 h) instar larvae, pupae (120-220 h) and adults were obtained by allowing collected eggs to develop at 25 ° C. To mass isolate second instar larvae of one particular sex, the genetic scheme of Lyttle (1989) and Walker et al. (1989) was employed as described in Polito et al. (1990).
Analytical procedures. DNA was extracted from adult flies as described in Manning et al. (1975). Plasmid and cosmid DNA was purified as described in Maniatis et al. (1982). Agarose gel electrophoresis of DNA, Southern transfer and filter hybridization were performed according to Fouts et al. (1981) and Beard et al. (1988). DNA restriction fragments were radiolabelled with (a-32p)dNTP using a nick-translation kit from Bethesda Research Laboratories (Gaithersburg, Md). Synthetic oligonucleotides were endlabelled with (y-32p)ATP using T4 polynucleotide kinase Boehringer-Mannheim, Indianapolis, Ind) according to Maniatis et al. (1982). All restriction enzymes were purchased from Boehringer-Mannheim and used as recommended. Total RNA was extracted by the hot phenol/ chloroform method (Jowett 1988) and poly(A)+RNA was isolated as described in Levy and Manning (1981). Northern blot analysis was performed by fractionation of the RNA on denaturing 1.1% agarose gels containing formaldehyde followed by transfer to Hybond N membranes (Amersham, Arlington Heights, Ill) and hybridization with radiolabelled DNA probes (Maniatis et al. 1982). For quantitation, appropriate regions of the membrane were excised using an exposed film as template and the radioactivity determined by liquid scintillation counting (Ghosh etal. 1989).
A cDNA library was constructed in phage Lgtl0, using poly(A +)RNA from second instar larvae, by methods described in Maniatis et al. (1982). The library was screened with the 1.4 kb HindIII-PstI restriction fragment from the pELBA plasmid of Ghosh et al. (1989). The inserts present in positive phages were subcloned into plasmid BSSK (Stratagene, LaJolla, Calif), and partial restriction maps were established for each insert.
A cosmid library was constructed in the Cosper transformation vector of Pirrotta (1988). The library was screened with a HindIII fragment containing the LSPI-a gene (Ghosh et al. 1989).
Sequencing of cDNA fragments that encode L12, inserted into the BSSK vector, was performed by the dideoxynucleotide chain termination method of Sanger et al. (1977). Oligonucleotides were synthesized in the Department of Molecular Biology and Biochemistry, UC-Irvine. The cDNA sequence was edited and analyzed using the IBI Pustell program.

Results and discussion
Based on the restriction endonuclease cleavage pattern, one cosmid clone (LSP 1-~Cos3) was found to contain the entire LSPI-~ gene (Fig. 1). Hybridization of this cosmid to Northern blots of RNAs extracted from different developmental stages detected a 2.8 kb transcript in third instar larvae which, on the basis of size and developmental specificity, is likcly to be the LSPI-~ mRNA. In addition, a 1.4 kb transcript, prevalent in second instar larvae, was detected (Fig. 2). Probing of duplicate blots with a 2.0 kb EcoRI cDNA fragment containing a portion of the Drosophila glucose-6-phosphate dehydrogenase gene (Zw ÷) indicated to us that transcripts synthesized in relatively low abundance would have been detected by our approach in all of the developmental stages tested (Ganguly et al. 1985).
The transcription unit encoding the 1.4 kb RNA was mapped by probing Southern blots of restriction endonuclease digests of LSPl-eCos3 with 32p-labelled poly(A+)RNA isolated from second instar larvae. This transcription unit, designated L12, lies within the 1.4 kb HindlII-PstI and 1.6 PstI-BamHI fragments immediately 5' to the LSPI-e gene (Fig. 1). Using strand-specific transcripts of the 1.4 kb HindlII-PstI fragment, the orientation of transcription of L12 was found to be the same as that of the LSPI-e gene. Since no transcription unit was detected downstream of LSPI-c~ in the LSPI-aCos3 insert, a new cosmid (1.9Cosl) was selected by screening the library with the 1.9 kb HindlII fragment from the distal end of the insert. No sequences from the 1.9Cosl insert were found to be represented in poly(A)+RNA from the various developmental stages tested. This observation suggests that there are no transcriptional units within approximately 44 kb of DNA 3' to LSPI-cc Clearly, we cannot rule out the existence of genes encoding very rare RNAs such as those found in low abundance in specific tissues.
A cDNA library from second instar larvae was screened with the 1.4 kb HindIII-PstI genomic fragment containing the 5' end of L12 and three positive clones were identified. The complete nucleotide sequence of the o G6PD LI2 ADH Fig. 3. Determination of the relative steady-state level of L12 RNA in male and female second instar larvae of Drosophila. A northern blot containing either male or female poly(A) + RNA (7 gg per lane) isolated from second instar larvae was hybridized with 32p-labeled L12 DNA, ADH DNA and G6PD DNA. The areas of the blot containing the hybridizing bands were excised and the radioactivity determined. The results are shown in Areas showing hybridization in the Northern blot in Fig. 3 were excised and their radioactivities were determined. An area of membrane of equivalent size was excised from a region in which no hybridization was observed, and assayed to determine background radioactivity, azp-labeted probes used were: L12, 1.4 kb HindIII-PstI fragment of LSPl-aCos 3; ADH, Sacl fragment; G6PD, EcoRI fragment from )~DmC20. Three independent pairs of male and female RNA samples were measured. The means of these measurements and their standard deviations are presented longest c D N A insert (1265 bp) was determined (data to be published elsewhere). This sequence, representing approximately 90% of the sequences present in the 1.4 kb L12 transcript, reveals a single long open reading frame with a codon usage in excellent agreement with that of several Drosophila genes (O'Connell and Rosbash 1984;Marsh et al. 1986;Fouts et al. 1988). The consensus polyadenylation sequence in the L12 c D N A is located 871 bp 5' to the transcriptional start site of the L S P I~ gene (Jowett 1985).

51
To determine whether L12 exhibits dosage compensation, Northern blots of poly(A+)RNA isolated from male or female second instar larvae were probed with a 3Zp-labelled 1.4 kb HindIII-PstI fragment from the LSPI-(~Cos3 genomic insert to detect the L12 transcript, as well as with an 11.8 kb Sacl fragment of the Adh + gene encoding alcohol dehydrogenase (Goldberg 1980), and a 2.0 kb EcoRI fragment from ZDmC20 containing a portion of the Z w + gene insert to detect the G6PD message (Ganguly et al. 1985). The latter two transcripts are expected to occur in equal amounts in males and females since Adh + is an autosomal gene and Z w + is X-linked and is known to be fully dosage compensated. These transcripts serve as internal controls to allow for correction of the differences in the relative amount of m R N A loaded onto the gels or transferred to the Northern blots (Fig. 3). Male to female ratios of the L12 transcript, normalized with either the alcohol dehydrogenase or the G 6 P D transcripts, are approximately equal to 1.0 (Table 1). Therefore, it is reasonable to conclude that the L12 gene is dosage compensated.
A number of rationalizations can be provided to explain why two adjacent genes may differ with respect to dosage compensation. For example, the elements which are responsible for the compensation of the L12 gene m a y not influence L S P I < t because their effect is narrowly localized and does not extend even to closely adjacent genes. Alternatively, these regulatory elements may exhibit developmental specificity and function only when the gene is transcribed (in the case of L12 and L S P I -e maximal gene activity occurs at distinctly different stages of larval development). A third possibility is that sequences are present which shield the LSPI-c~ gene from the long-ranging effects of the regulatory elements. While additional experiments are needed to test these possibilities, it is important to recognize that closely adjacent genes can differ with respect to dosage compensation.