A novel fruitfly protein under developmental control degrades uracil-DNA

Biochemical and Biophysical Research Communications 355 (2007) 643–648 www.elsevier.com/locate/ybbrc

A novel fruitfly protein under developmental control degrades uracil-DNA Ange´la Be´ke´si a, Ma´ria Puka´ncsik a, Vill} o Muha a, Imre Zagyva a, Ibolya Leveles a, b b E´va Hunyadi-Gulya´s , E´va Klement , Katalin F. Medzihradszky b, Zolta´n Kele c, Anna Erdei d, Ferenc Felfo¨ldi e, Emese Ko´nya a, Bea´ta G. Ve´rtessy a,* b

d

a Institute of Enzymology, Karolina u´t 29. H-1113, Budapest, Hungary Proteomics Research Group, Biological Research Center, Szeged, Hungary c Department of Medical Chemistry, University of Szeged, Hungary Department of Immunology, Eo¨tvo¨s Lora´nd University, Budapest, Hungary e MolCat Ltd., Budapest, Hungary

Received 24 January 2007 Available online 12 February 2007

Abstract Uracil in DNA may arise by cytosine deamination or thymine replacement and is removed during DNA repair. Fruitfly larvae lack two repair enzymes, the major uracil-DNA glycosylase and dUTPase, and may accumulate uracil-DNA. We asked if larval tissues contain proteins that specifically recognize uracil-DNA. We show that the best hit of pull-down on uracil-DNA is the protein product of the Drosophila melanogaster gene CG18410. This protein binds to both uracil-DNA and normal DNA but degrades only uracil-DNA; it is termed Uracil-DNA Degrading Factor (UDE). The protein has detectable homology only to a group of sequences present in genomes of pupating insects. It is under detection level in the embryo, most of the larval stages and in the imago, but is strongly upregulated right before pupation. In Schneider 2 cells, UDE mRNA is upregulated by ecdysone. UDE represents a new class of proteins that process uracil-DNA with potential involvement in metamorphosis.  2007 Elsevier Inc. All rights reserved. Keywords: Uracil-DNA; Nuclease; Development; Metamorphosis; Drosophila melanogaster; Pupating insect

Uracil is considered to be a mistake in DNA that may arise by cytosine deamination or thymine replacement. Spontaneous cytosine deamination occurs frequently, generating mutagenic U:G mismatches. Thymine-replacing uracil incorporation is not mutagenic and occurs if the cellular dUTP/dTTP ratio becomes much elevated since DNA polymerases will incorporate either U or T against A. Two enzymes have primary roles in safe-guarding against uracilDNA: dUTPases remove dUTP from the DNA polymerase pathway [1], while uracil-DNA glycosylases (UDGs) excise uracil from DNA, either produced by cytosine deamination or thymine-replacing incorporation [2]. To ensure uracilfree DNA, representatives of these two enzymes are encoded in all free-living organisms and even in diverse viruses. *

Corresponding author. Fax: +361 4665 465. E-mail address: [email protected] (B.G. Ve´rtessy).

0006-291X/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2007.01.196

In Drosophila, however, the finding that the ung gene coding for the major UDG enzyme is absent from the genome [3] presented an intriguing problem: how can a developed multicellular organism exist without a robust uracil-DNA glycosylase? The UDG superfamily members present in Drosophila are characterized with decreased catalytic rate constants and restricted substrate range [4]. In addition, dUTPase protein is also absent from larval tissues [5], i.e., Drosophila larvae lack both enzymes acting against uracil in DNA. These data suggested potential accumulation of uracil-DNA in fruitfly larvae (cf. also [6]), as shown for cells deficient in UDG or both in dUTPase and UDG activities [7,8]. Here we asked if uracil-DNA in larval tissues might trigger any physiological response and set out to identify uracil-DNA recognizing proteins from Drosophila extracts. We show that the most abundant hit of uracil-DNA

644

A. Be´ke´si et al. / Biochemical and Biophysical Research Communications 355 (2007) 643–648

pull-down experiments is the protein product of the CG18410 gene. The protein, termed uracil-DNA degrading factor (UDE), does not exhibit glycosylase activity but it is capable of specifically degrading uracil-DNA. It has no detectable homology to nucleases/glycosylases or any other proteins except a group of sequences present in pupating insects. UDE protein in the fruitfly is under strict developmental control: it is absent in most stages but becomes strongly upregulated right before pupation. Data imply the existence of a novel protein family that specifically recognizes uracil-DNA. UDE may have a role in metamorphosis and may be used as a molecular biology tool. Materials and methods Uracil-DNA affinity chromatography. Uracil-DNA was prepared by amplification of plasmid DNA (pSUPERIOR-puro (Invitrogene)) in dut-/ ung- CJ236 Escherichia coli [7]. Control plasmid was prepared in XL1Blue E. coli. NotI-linearized plasmid was immobilized on cyanogen bromide activated Sepharose (Amersham). Late third larval extract (cf. [5]) in 50 mM Pipes, pH 7.9 buffer, also containing 50 mM KCl, 4 mM EDTA, 1 mM dithiothreitol and 1 mM phenylmethanesulfonyl fluoride (starting buffer) was prefractionated on heparin column and loaded onto either uracil-DNA or normal DNA column. Bound proteins were eluted by adding 0.4 M NaCl to the starting buffer and analyzed on SDS–PAGE. Five protein bands were detected as exclusive to uracil-DNA pull-down mixture. The most intense band was subjected to mass spectrometry investigations. Protein identification by mass spectrometry. Protein identification by mass spectrometry was performed as in [5] using a Bruker Daltonics Reflex III MALDI-TOF MS instrument in 2,5-dihydroxybenzoic acid matrix without further fractionation and analysis was done by ProteinProspector against the NCBInr 05.10.2004 database. Identification was confirmed by MS/MS (post source decay) spectra. DNA cloning and recombinant protein expression. The cDNA for gene CG18410 of Drosophila melanogaster (Open Biosystems) was cloned into pET19b (Novagene), using NdeI and XhoI sites. The recombinant construct included a 6· His tag and a linker segment at the N-terminus so that the amino acid sequence at the N-terminus was (M)GSSHHHHHHSSGLVPRGS to be followed by the sequence as shown in Fig. 1A, lacking the first four residues. This plasmid was termed as pET-HisUDE. The same plasmid was also used for cloning the gene of human dUTPase. Protein production was induced in Bl21(DE3)pLysS and BL21(DE3)ung-151 or BL21(DE3)ung-151pLysS E. coli by addition of 1 mM isopropyl-beta-D-thiogalactopyranoside at logarithmic growth and shifting temperature to 25 C for four more hours. Ni–NTA-purified proteins were stored in 20 mM Hepes, pH 7.5, buffer containing 150 mM KCl and 1 mM dithiothreitol. Plasmid DNA processing assays. Linearized uracil-containing plasmid DNA or control plasmid DNA at 20 lg/ml was incubated with 10 lg/ml UDE protein or 2 · 103 U/ml UNG protein (Sigma) at 37 C in 25 mM Tris/HCl, pH 7.5, buffer, also containing 0.1 mg/ml albumin (assay buffer). At given reaction times, reaction mixtures were incubated at 75 C for 10 min, followed by agarose gel electrophoresis. UNG reaction assay was performed in the same assay buffer and the reaction mixtures were incubated at room temperature in the presence of 60 mM NaOH for 10 min (to allow cleavage of AP sites). Products were detected by the standard ethidium-bromide staining after agarose gel electrophoresis on 0.75% (w/w) agarose gels. Determination of uracil by HPLC/mass spectrometry. Following isocratic HPLC (Applied Biosystems 140 C instrument, methanol:water 10:90; flow rate: 250 ll/min; sample volume: 20 ll; column YMC J‘sphere ˚ , 50 · 2.1 mm and Phenomenex Luna 5 lm, 100 A ˚, H80 4 lm, 80 A 50 · 2.0 mm), samples were analyzed on a Finnigan TSQ_7000 triple quadrupole mass spectrometer (Finnigan_MAT, San Jose, CA) equipped with a Finnigan Atmospheric Pressure Chemical Ionization (APCI) source

in positive ion mode using selective reaction monitoring (SRM) [9–11]. In this mode, the mass spectrometer examines only the previously set precursor(s) (113 in the case of uracil) which can produce previously set fragment ions (96 for NH3 loss and 70 for HNCO loss for uracil). For tuning optimal fragmentation parameters of uracil, tandem mass spectra were acquired. Further analysis of uracil content was based on the transition 113–96 with better signal to noise ratio. Uracil standard calibration curve was determined by injecting 20 ll aliquots of 50–500 ng/ml uracil in water. Reaction mixtures contained 1 mg/ml uracil-containing plasmid DNA in 10 mM ammonium bicarbonate buffer, pH 8.0, and either 50 lg/ml UDE protein or 102 U/ml UNG protein (Sigma). Reaction mixtures were incubated for 60 min at 37 C, stopped by 1.2 M HCl, and diluted 10-fold in water or in watercontaining uracil at 50–500 ng/ml before injection. Aldehyde reactive probe (ARP) assay. Aldehyde reactive probe (ARP) assay was performed using ARP reagent (Dojindo Molecular Technologies). Reaction mixtures contained 1 mg/ml uracil-containing plasmid DNA in plasmid assay buffer and either 50 lg/ml UDE protein or 102 U/ ml UNG protein (Sigma). Reaction mixtures were incubated for 60 min at 37 C, spotted on nitrocellulose membrane and analyzed as in [12]. Biotintagged AP sites were detected using streptavidin-conjugated horseradish peroxidase (Sigma). Drosophila Schneider 2 (S2) cell line. Drosophila Schneider 2 (S2) cell line (Invitrogen) was cultured at 27 C in Drosophila SFM medium (cf. [5]). At 60% confluency, ecdysone was added to the medium to a final concentration of 2.4 lg/ml, and cells were further grown for the intervals indicated. Semi-quantitative RT-PCR. Semi-quantitative RT-PCR was performed as described for dUTPase [5] using forward and reverse primers: 5 0 GTCCTGACAATGACCAAGGCGG-3 0 , and 5 0 -CCTCGCCATCG GAATCCTGGC-3 0 , or 5 0 -CAACCCGATCCACAGCCAGAACCAGA CAC-3 0 and 5 0 -GCTGTTGTCCGAGCCGCGCAGAC-3 0 for UDE or Broad-Complex transcripts, respectively. Preparation of polyclonal IgG and Western blotting. Preparation of polyclonal IgG and Western blotting was performed as in [5]. Protein A-purified serum IgG was used at 1:180,000 dilution. Extracts from developmental stages were prepared with protease inhibitor cocktail (Sigma). The same amount of total protein from each extract was loaded on SDS–PAGE gels. This was confirmed by running the SDS–PAGE gels in duplicate and using one gel for Coomassie protein quantification by laser densitometry, while using the other gel for blotting. Densitometry data indicated that protein content in the lanes differed by less than 4%. Blot results were detected by enhanced chemiluminescence.

Results and discussion Identification, cloning, and characterization of UDE The most abundant hit of uracil-DNA pull-down experiments from late third stage larval extract was identified by mass spectrometry as the product of gene CG18410 (with unknown function in [3]) (Supplementary Fig. 1). A few homologous sequences, all present in genomes of other pupating insects (e.g., silk worm, honeybee, and mosquito) were found (Fig. 1A). The estimated pI of the protein is rather high (9.41), arguing for its potential to bind to DNA. The protein was cloned and expressed as a His-tagged construct in E. coli (Fig. 1B). Purified recombinant protein was checked for DNA-binding ability in electrophoretic mobility shift assays (Fig. 2A) that indicated significant shift in the DNA position in a dose-response characteristic manner. If the uracil-substituted DNA was replaced by the normal DNA, the extent of gel-shift was less pronounced.

A. Be´ke´si et al. / Biochemical and Biophysical Research Communications 355 (2007) 643–648

A

645

B

Fig. 1. Identification and expression of the CG18410 gene product protein. (A) Sequence alignment for CG18410 homologues. Tblastp and NCBI genomial blast were used for thorough similarity searches. Gray boxes indicate conserved residues (stars: absolutely conserved, dots: similar residues) revealing extended conserved motifs. (B) Expression and purification of recombinant His-tagged CG18410 gene product protein. (Upper) Expression in BL21(DE3)pLysS strain, lanes: extract before induction, extract after induction, flow-through on Ni–NTA column, empty lane, purified protein. Markers in kDa. (Lower) Expression of dUTPase and CG18410 gene product (UDE) in BL21(DE3)ung-151 and BL21(DE3)ung-151pLysS strains. Note that dUTPase is expressed in both these strains, with strong basal expression before induction if cells lack the control plasmid pLysS, while UDE is expressed only in the BL21(DE3)ung-151pLysS strain. Arrows indicate expected positions for His-dUTPase and His-UDE at the left and right side, respectively. Mid-labels: markers in kDa.

Gel-shift results did not depend on the presence or absence of available divalent metal ions, sequestrable by EDTA, arguing that the protein binds to DNA under both conditions. Incubation experiments of the recombinant protein with uracil-DNA indicated efficient DNA degradation (Fig. 2B– D). DNA degradation is apparently strictly dependent on the presence of dUMP moieties in the DNA. Normal DNA is not degraded under the same experimental conditions. DNA degradation by the CG18410 gene product protein could be visualized on the agarose gels only after heat-induced separation of the DNA strands, indicating single-stranded DNA cleavage by the protein. As expected [13], the control UNG reaction could only be visualized after alkaline cleavage of abasic sites. These results argue that the UDE recombinant protein is capable of inducing specific degradation of uracil-DNA. We therefore termed this protein uracil-DNA degrading factor (abbreviated as UDE). Expression of UDE in ung-E. coli requires strong control The recombinant UDE protein used in the above experiments was produced in E. coli cells that usually show appreciable activities of endogenous UNG. Co-purification of E. coli UNG with UDE may result in contaminations interfering with the independent assessment of UDE function. We therefore expressed UDE in BL21(DE3)ung-151 E. coli strain lacking detectable UNG activity [14]. Fig. 1B shows that UDE was produced in these cells only if the strong control plasmid pLysS was also included indicating that UDE may possess an activity negatively inter-

fering with BL21(DE3)ung-151 cell. The ung- cell line facilitates uracil-DNA accumulation and expression of UDE in these cells may lead to fragmentation of E. coli genomic DNA, a toxic situation for the bacteria. The strong control of the pLysS plasmid confines the expression of UDE strictly to the period after induction thereby allowing controlled protein production. These expression patterns provided an independent assessment for the role of UDE within the cellular milieu. The uracil-DNA processing activity of the protein expressed in the BL21(DE3)ung-151pLysS cells was indistinguishable from the activity of UDE produced in BL21(DE3)pLysS cells (data not shown), arguing for the conclusion that the observable function is not due to E. coli UNG contamination. The presently identified UDE is therefore capable of uracil-DNA recognition and processing. UDE does not possess DNA-glycosylase activity To address the possibility of UDE catalyzing glycosidic bond cleavage, two independent methods were used. First, uracil release was addressed by analyzing reaction products generated in mixtures of UDE or UNG and uracil-DNA plasmid by HPLC/MS. Uracil eluted at 1.4 min and the calibration curve indicated a detection limit of 96 h; 13: emerging young fly; 14: adult fly (>5 min). Fifty micrograms of total protein was loaded. (C) Ecdysone induction. RT-PCR was performed for UDE and Broad-Complex transcripts using equal amounts of cDNA prepared from S2 cells grown in the absence of ecdysone (C: control) or after induction with ecdysone for the time intervals indicated.

decreased intensity, arguing that although the matrix may decrease the signal, uracil detection is still possible with high sensitivity (cf. calibration curves on Fig. 2E, top right). These results show that UDE does not generate any detectable uracil (