A New Pathway Expressed during a Distinct Stage of Drosophila Development for the Removal of dUMP Residues in DNA*

In view of removing lesions in DNA produced by the deamination of cytosine to uracil, uracil-DNA glyco-sylases were anticipated to be ubiquitous. However, an analogous activity in Drosophila melanogaster was not detected. Instead, a nuclease was identified that acts specifically upon DNA containing uracil. The cleavage of uracil-containing DNA by the nuclease generates acid-soluble oligonucleotides in a reaction which can be inhibited by pretreatment of the DNA with Esche- richia coli uracil-DNA glycosylase. Uracil-containing DNA with either A:U base pairs or G:U base pairs were susceptible to cleavage by the nuclease, whereas other damaged DNA substrates were not. The nuclease activity is transient and appears only in third instar larvae, with other developmental stages of Drosophila lacking significant levels of the nuclease. Uracil latter

In view of removing lesions in DNA produced by the deamination of cytosine to uracil, uracil-DNA glycosylases were anticipated to be ubiquitous. However, an analogous activity in Drosophila melanogaster was not detected. Instead, a nuclease was identified that acts specifically upon DNA containing uracil. The cleavage of uracil-containing DNA by the nuclease generates acid-soluble oligonucleotides in a reaction which can be inhibited by pretreatment of the DNA with Escherichia coli uracil-DNA glycosylase. Uracil-containing DNA with either A:U base pairs or G:U base pairs were susceptible to cleavage by the nuclease, whereas other damaged DNA substrates were not. The nuclease activity is transient and appears only in third instar larvae, with other developmental stages of Drosophila lacking significant levels of the nuclease.
Uracil represents a nonconventional base in DNA that can arise either through the accidental misincorporation of dUMP residues instead of dTMP or by the spontaneous deamination of cytosine to uracil in DNA. The latter event is considered to be more hannful since it can lead to transition mutations. To counteract the presence of uracil in DNA, most organisms contain a uracil-DNA glycosylase, an activity which removes uracil in DNA by hydrolyzing the sugar-base glycosylic bond, forming an apyrimidinic site susceptible to base excision repair (1).
In the search for a uracil-DNA glycosylase activity in Drosophila melanogaster, we found instead a nuclease that specifically acts upon DNA containing uracil to produce acidsoluble oligonucleotides in a reaction similar to that identified for Escherichia coli endonuclease V (2). Additionally, we were able to detect the nuclease activity only in third instar larvae, as other developmental stages of Drosophila lacked significant levels of the nuclease.

EXPERIMENTAL PROCEDURES
Preparation of Third Instar Larval Extracts-Drosophila melanogaster (Oregon-R) embryos were collected 4 h subsequent to establishing a population cage. Embryos were washed with 0.01% Triton * This work was supported by the Louisiana Agricultural Experiment Station and National Institutes of Health Grant GM 27358. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. t: To whom reprint requests should be addressed. X-100,0.7% NaCl, then repeatedly with H20. The embryos were then transferred to a 15-ml polystyrene centrifuge tube using 70% ethanol.
All subsequent operations were carried out in a sterile environmental hood. After repeated washings in 70% ethanol, the embryos were collected on a 140-pm Nitex monofilament screen (Tetko, Inc.), divided, and introduced into sterile bottles containing sterile, dead yeast-sucrose medium (3). Incubations were at 25 "C in a Forma diurnal growth chamber. Individual developmental stages were collected, identified (4), and washed as above. After the final 70% ethanol wash, the ethanol was removed and replaced by 50 m M potassium phosphate, pH 7.5, 1 m M EDTA, and 25 pl of the protease inhibitor Aprotinin (Sigma) per ml of buffer. After removal of excess buffer, the individual developmental stages were stored at -20 "C. Crude extracts of the various developmental stages were prepared as described previously (5).
Enzyme Assay-To monitor the release of acid-soluble material, assay mixtures (0.1 ml) contained 25 m~ Tris-HC1, pH 7.5,100 pg/ml of acetylated bovine serum albumin, 3 nmol of T5 [3H]DNA (containing 7% substitution of thymine by uracil, Ref. 6; specific activity of 19,800 cpm/nmol of DNA nucleotide), and protein. Incubations were for 30 min at 37 "C, at which time incubation tubes were placed on ice and 0.1 ml of 5 mg/ml of bovine serum albumin and 0.25 ml of 7% trichloroacetic acid were added. After 5 min, tubes were centrifuged for 5 min at 10, OOO X g and 0.3 ml of the supernatant was subsequently removed, aqueous scintillation fluor was added, and the presence of radioactivity was determined.
Other Methods-Uracil-DNA glycosylase was prepared from E. coli strains deficient for apurinic/apyrimidinic endonuclease VI and was essentially similar to that described by Lindahl et al. (8) and Ljungquist (9), in which hydroxyapatite fractions were concentrated with ammonium sulfate (IO) and then dialyzed versus 10 m M Tris-HCl, pH 7.5, 1 m M EDTA, 1 m M dithiothreitol, and 5% glycerol.

RESULTS
We initially attempted, without success, to identify from crude extracts of Drosophila embryos an activity that would liberate acid-soluble material from uracil-containing T5 E3H] DNA. We then resorted to screening other developmental stages of Drosophila. Among the different developmental stages examined, the liberation of acid-soluble material from T5 [3H]DNA was detected only for crude extracts of third instar larvae. We subsequently made a more detailed analysis of different developmental stages of Drosophila to determine when the activity first appeared. We were additionally interested in determining whether the activity remained or alternatively became inactive at later developmental stages (Table   I). Third instar larvae represent the first stage in the development of Drosophila in which the liberation of acid-soluble material from uracil-containing DNA can be detected. Furthermore, the activity is transient, as late pupal stages reflect little or no activity. Note that activity due to bacterial contamination would not be predicted to behave in this fashion. Also, we were able to detect activity in the absence of any divalent cation. As a result, our interpretation of those developmental stages that were indeed active was not compromised by nonspecific DNases that require divalent cations, such as Mg2+, for activity.
We have additionally examined the germ-line tissue of    Table I. Incubations were for 30 min at 37 "C. The complete reaction showed that 0.14 nmol of DNA nucleotide was liberated as acid-soluble material by E . coli uracil-DNA glycosylase; 0.26 nmol of DNA nucleotide was liberated as acid-soluble material in complete reactions for third instar extracts. Twenty pl of the total reaction mixture were spotted on polyethyleneimine-impregnated cellulose thin layer plates and developed as described by Gates and Linn (2).
strate such an activity in crude extracts of either unfertilized embryos or adult testis. Contrary to our expectations, the activity observed in third instar larvae was not attributable to a uracil-DNA glycosylase, for while a partially purified fraction of E. coli uracil-DNA glucosylase liberated free uracil from uracil-containing T 5 [3H]DNA, crude extracts of third instar larvae did not liberate free uracil (Fig. 1). Reactions with extract from all stages were additionally checked, without success, for an activity that would liberate free uracil. T h e nuclease from third instar extracts toward uracil-containing DNA also did not degrade the DNA substrate to deoxymononucleotides, but instead acts in a manner similar to that observed for the E. coli endonuclease V enzyme isolated by Gates and Linn (2). T o determine whether the nuclease present in third instar extracts was dependent upon DNA containing uracil, PM2 ["]DNA was treated with sodium bisulfite according to the Uracil-containing T5, heat-denatured 0.24 a Five pl of an ammonium sulfate concentrated fraction (capable of liberating approximately 2 nmol of DNA-uracil) was added to the bisulfite-treated ?M2 DNA (specific activity of 2700 cpm/nmol of DNA nucleotide) and incubated for 30 min at 37 "C prior to the addition of the third instar extract. method of Lindahl et al. (8). T o induce the deamination of cytosine to uracil, conditions were chosen to form roughly 1 dUMP residue/genome. This DNA substrate was indeed rendered partially acid-soluble1 (Table 11). That the susceptibility of this DNA substrate was dependent upon the presence of uracil is shown by the absence of nuclease activity on bisulfitetreated DNA fist reacted with E. coli uracil-DNA glycosylase prior to exposure to third instar crude extracts.
Consistent with the absence of activity on bisulfite-treated DNA in which uracil is removed by E . coli uracil-DNA glycosylase is the apparent lack of activity toward other damaged DNA substrates (Table   11). It appears that the activity analyzed by the liberation of acid-soluble material using either uracil-containing T 5 or PM2 DNA is relatively specific for the presence of uracil. Thus, the Drosophila nuclease is capable of acting at both A:U base pairs, as demonstrated by the uracil-containing T 5 DNA, and at G:U base pairs, as seen for bisulfite-treated PM2 DNA. Also, denatured T5 DNA was consistently somewhat less susceptible than the duplex form to the release of acid-soluble material (Table 11).

DISCUSSION
The apparent absence in Drosophila of a uracil-DNA glycosylase is to our knowledge unique. It is possible, however, that although the uracil-DNA glycosylase activity is easily demonstrated in both bacteria and mammalian cells, this may not be the case in Drosophila. Since Drosophila have a highly active nuclease to act on uracil-containing DNA, cleavage at uracil sites in the DNA may inhibit the DNA glycosylase from acting in vitro. Similar inhibition has been observed for purine base insertase acting on depurinated DNA subsequent to the exposure of the DNA to apurinic/apyrimidinic endonuclease (11). This does not explain the apparent absence of a uracil-DNA glycosylase activity in other developmental stages, however, where the nuclease appears to be inactive. E . coli endonuclease V, which also shows a strong preference in vitro for uracil-containing DNA (2), provides a prec-' The amount of degradation observed for bisulfite-treated DNA appears to be excessive in light of the amount of uracil present. This could be due to contaminating exonucleases present in crude extracts of third instar larvae or the fact that uracil acts as an entry point for a processive enzyme. The amount of degradation, however is dependent solely upon the prior action of the uracil nuclease, as this degradation is not observed for other damaged DNA substrates or on bisulfite-treated DNA that is reacted with E. coli uracil-DNA glycosylase prior to incubations with extracts of third instar larvae. edence with regard to an endonuclease degrading uracil-containing DNA to acid-soluble oligonucleotides. Endonuclease V does act on other damaged DNA substrates ( e g . apurinic/ apyrimidinic sites in duplex DNA) that were not susceptible to the Drosophila activity. However, the preference of endonuclease V for uracil-containing DNA does not appear to be a major pathway in vivo (12).
Our results thus far indicate that uracil-containing DNA remains the only substrate acted on by the Drosophila nuclease and, furthermore, the nuclease acts regardless of how the dUMP was produced in the DNA. It is conceivable that dUMP residues in DNA arising from cytosine deamination are in part responsible for the presence of the nuclease activity observed in third instar larvae. One would predict that an activity directed toward this type of occurrence would remain active, particularly in adult germ-line, so that potentially mutagenic sites in DNA would not be passed on to the next generation. Alternatively, the nuclease may be produced in response to misincorporated uracil. With this in mind, it will be of interest to learn in future studies if Drosophila do incorporate a fairly s i g d c a n t level of dUMP into their DNA. If this is found to be the case, it may be that the incorporation of uracil into DNA is not a biological accident, but instead part of some cellular design (13) utilized by Drosophila that is facilitated by the cleavage at uracil sites in the DNA molecule.