1846-1854 Nucleic Acids Research, 1994, Vol. 22, No. 10

Identification of the rat xanthine dehydrogenase/oxidasepromoter

Chi-Wing Chow1, Melissa Clark2, Jean Rinaldo1'2 and Roger Chalkley'l*1Department of Molecular Physiology and Biophysics and 2Department of Medicine,School of Medicine, Vanderbilt University, Nashville, TN 37232, USA

Received February 2, 1994; Revised and Accepted March 31, 1994

ABSTRACTInflammation and ischemia - reperfusion tissue injuryare important pathophysiologic processes with a widespectrum of clinical presentations; the enzymexanthine dehydrogenase/oxidase (XDH/XO) Is thoughtto play a key role in ischemia - reperfusion injury.Recent studies have shown the transcriptionalregulation of XDH/XO by cytokines (Dupont etal., 1992,J. Clin. Invest. 89, 197 - 202). In the present study, the5' structure of the XDH/XO gene and characterizationof its promoter are undertaken providing an initial stepto further elucidate the regulatory mechanism(s) of thisenzyme. XDH/XO cDNA from rat bone marrowmacrophage has been isolated and used to screen arat genomic library in order to identify and characterizethe promoter of the XDH/XO gene. By Southernanalysis, XDH/XO was found to be a single copy genein the rat genome. Primer extension, RNase protection,and anchor-PCR studies indicate the presence ofmultiple start sites within a 65 bp window located some20- 85 bp upstream of the translation initiator (ATG).Functional studies of the sequences up to 116 ntupstream of the translational start site, which encom-passes the several transcriptional start sites, indicatethat this region is sufficient to drive the expression ofa luciferase reporter gene and is presumed to representthe promoter. Neither a TATA box nor a GC-rich regionare present in close proximity to any of the transcript-ional start sites; however, sequences with homologyto known initiator elements are found within this 116bp fragment. Several possible regulatory elements,including a NF-IL6 motif, are also located upstream ofthe transcriptional start site. This study represents thefirst description of the XDH/XO promoter from avertebrate system.

INTRODUCTIONXanthine oxidase (XO) and xanthine dehydrogenase (XDH) haveimportant physiologic and pathophysiologic actions (1). XDHcatalyzes the oxidation of hypoxanthine and xanthine to uric acid,accompanied by reduction of NAD to NADH. XO is thought

EMBL accession nos U08120- U08123 (incl.)

to be derived from the XDH molecule, either by proteolyticcleavage or by oxidation of sulfhydryl groups (1). XO catalyzesthe oxidation of xanthine and hypoxanthine by reduction of 02to superoxide radical (02-), which undergoes furtherdismutation to H202. The hydroxyl radical (-OH), which ishighly reactive and can cause cell injury, can then be produceddirectly (2) or indirectly (3). The reactive oxygen metabolitesgenerated from these reactions may contribute to ischemia-reperfusion tissue injury. Apparently, during ischemia, the rateof conversion ofXDH to XO is accelerated while the availabilityof purine substrates is also increased due to ATP depletion. Theseevents are thought to lead to increased generation of 02- by XOwhen molecular oxygen becomes available as an electron acceptorduring reperfusion (4,5).

In addition to playing a key role in ischemia-reperfusioninjury, XO may also be important in inflammation, and XDH/XOactivity has been found in macrophages (6-10). A wide spectrumof models of inflammation are characterized by an increase inXO activity, including thermal bums, complement activation, theadult respiratory distress syndrome, and influenza pneumonia(11-14). While it had been proposed that higher activity ofXOduring inflammation resulted from a higher rate of conversionof XDH to XO, some studies have also suggested that the totalamount of XDH is increased (16,17).

Since XO is derived by proteolysis or oxidation of XDH, theXDH gene is presumed to encode XO activity; however, littleis known about the mechanism(s) of regulation ofXDHIXO geneexpression. The sequences of rat and mouse liver XDH/XOcDNA have been reported by Amaya et al. (15) and Terao etal. (16), respectively. Recendy, Dupont et al. (17) reported thatinterferon gamma (IFN) is a potent inducer ofXDH/XO in termsof activity, mRNA levels, and transcriptional rate in ratendothelial cells. Similar studies have also been performed inmouse liver (18). These studies suggested that the expression ofXDH/XO gene could be up-regulated by IFN duringinflammation.

Little is known about the genomic structure and regulatoryelements that govern transcription of XDH/XO. While thesequences for XDH from Drosophila melanogaster (19) andCalliphora vicini (20) have been reported, their homology to themammalian genes seemed to be limited to the coding region. To

*To whom correspondence should be addressed

\.) 1994 Oxford University Press

Nucleic Acids Research, 1994, Vol. 22, No. 10 1847

our knowledge, no mammalian XDH/XO genomic clone has beenreported.To explore the possibility of modulation of XDH/XO gene

activity in inflammatory cells during inflammation or duringischemia-reperfusion injury, we have (1) cloned an XDH/XOcDNA from rat bone marrow derived macrophages, (2) clonedthe 5' end of the XDH/XO gene from rat DNA and (3) identifiedthe promoter region of this gene.

MATERIALS AND METHODSMaterialsRestriction enzymes and other DNA modification enzymes werepurchased from Promega and/or New England Biolabs. TheSequenase 2.0 sequencing kit was obtained from United StatesBiochemical Corporation. Radioactive chemicals were purchasedfrom NEN. Oligonucleotides were synthesized on a Milligen/Biosearch 7500 DNA synthesizer and the sequences are shownbelow.

Isolation of rat bone marrow macrophageRat bone marrow macrophages (RBMM) were isolated andcultured in alpha-minimal essential medium containing 10% fetalcalf serum and 10% L-cell conditioned media (LCM) (21). Underthe influence of LCM, bone marrow macrophage precursorsdifferentiate. After 5-7 days, the cells exhibit macrophagemorphology, plastic adherence, and expression of macrophage-specific mannose receptor activity (22,23).

Isolation of cDNA clones of XDH/XO from RBMMA RBMM cDNA library was constructed using total RNA frombone marrow-derived macrophages after 7 days in culture. Toscreen the library, we relied upon known homologies betweenthe rat liver and Drosophila XDH/XO (15). We synthesizedprimers from two regions of 100% base homology between therat liver and Drosophila cDNAs. The 5' sense primer(oligonucleotide 885) and the 3' antisense primer (oligonucleotide886) were used in a PCR reaction with the rat BMM cDNAlibrary as template. A single PCR product of approximately 440base pairs was generated. 1 x 105 plaques were screened by filterhybridization using E. coli C600 HFL as host, and using the 440bp PCR product as probe. Nine positive plaques were identified.One purified clone, GTO-XO-1, was amplified in liquid cultureand the phage DNA purified by the CsCl2 method. Two EcoRIsubfragments were isolated from the GTIO-XO-l clone andsubcloned into Bluescript KS II plasmid (Stratagene) byblue-white selection using X-gal and IPTG. Plasmid DNA wasscreened to identify clones with a 900 bp insert (clone # 20) andsome with a 600 bp insert (clone # 10). cDNA clones # 10 and#20 were amplified in liquid culture and sequenced by thedideoxy method.

Oligonucleotide Position737 (antisense) - 19 to -53885 (sense) +16 to +39825 (antisense) +42 to +22694 (antisense) +84 to +581314 (antisense) + 107 to + 871292 (antisense) +141 to 121886 (antisense) +450 to +421296 (sense) +436 to +456

Anchor oligonucleotideComplementary anchor oligo

Isolation of a genomic clone of rat XDH/XOAbout 5 - 10 x 105 plaques of a Sall partial digest rat genomiclibrary were screened by plaque hybridization. The 900 bpfragment from the XDH/XO cDNA (cDNA clone #20, seeabove) was labeled for the initial screening. The purified clonewas subjected to restriction enzyme digestion and subcloned intoBluescript KS II plasmid for sequencing. The nucleotidesequence(s) reported in this paper has been submitted to theGenbank/EMBL Data Bank with accession numbers U08120,U08121, U08122, and U08123.

Southern analysisRat liver genomic DNA (20-50 ug) was digested with restrictionenzymes and subjected to electrophoresis in a 1 % agarose gel.The gel was transferred to Zeta-Probe (Bio-Rad), hybridized tocDNA probes from +5 to + 863 (clone #20) and from + 858to + 1518 (clone # 10), and analyzed by autoradiography.

Primer extension of XDH/XO mRNATotal RNA was extracted from rat bone marrow macrophagesusing Tri-reagent (Molecular Research Center Inc.). About 20,zg total RNA was used for primer extension together with3 x 105 cpm of oligonucleotide primer (oligonucleotides 1292and 694). Primer and RNA were annealed in buffer containing20 mM Tris-HCI (pH 7.5), 250 mM NaCl, and 1 mM EDTA,heated at 85°C for 10 min, and transferred to a 500 or 60°Cwater bath for another 60 min. Reverse transcriptase buffer, 500/%M dNTPs, and 5 U of avian myeloblastosis virus reversetranscriptase were added for elongation. The extended productswere separated by electrophoresis in a 5% polyacrylamide gelunder denaturing condition and subsequently autoradiographed.

Identification of the 5' untranslated region by anchor-PCRAnchor-PCR was performed as described by Troutt et al (24).In brief, primer extension was performed as mentioned aboveexcept the primer oligonucleotide (oligonucleotide 694) was notphosphorylated and the elongation step was performed at 50°C.Primer extended products (cDNA) were purified from excessprimer oligonucleotides by using a Sephadex G-50 spin column.The anchor oligonucleotide (27) was blocked at the 3'-end withddATP using terminal deoxynucleotidyltransferase and 5' endlabeled with ATP using T4 polynucleotide kinase. Ligation ofcDNA to the anchor oligonucleotide was carried out at roomtemperature for at least 24 h with 5 pmol of the 3' blocked and5' phosphorylated anchor oligonucleotide, T4 RNA ligasereaction mix (50 mM Tris-HCI, pH 8.0, 10 mM MgCl2, 1mM hexamine cobalt chloride, 20 OM ATP, 25% PEG 8000),and 5 U T4 RNA ligase (New England Biolab) in a total volumeof 10 1d. Ligation reaction was terminated and amplified asdescribed (24). The first PCR amplification was carried out asfollows using 50 pmol of both internal oligonucleotide

Sequences5'-TGAGCTCTTGAGAAGTAGTTGTAGATATCACTGCC-3'5'-TTGGTCTTCTTTGTGAATGGCAAA-3'5'-CTTTTTGCCATTCACAAAGAA-3'5'-GACCAGAAGTGTTGTTTCAGGGTCCGC-3'5'-AGCCCCAACTTTCTTCTCAGG-3'5'-GCCACCTTCTCCACAGCCAAG-3'5'-TGTACAGCGACAGAGGTTTCCTTGGAAGGC-3'5'-CTCTGTCGCTGTACAGGCTAC-3'5'-CAGCTGTTGGCTGCAATTGCGCCACCGCCACAG-3'5'-CTGCTGTGGCGGTGGCGCAATTGCAGCCAACAG-3'

1848 Nucleic Acids Research, 1994, Vol. 22, No. 10

(oligonucleotide 825) and complementary anchor oligonucleotide:25 cycles of PCR with annealing temperature at 560C for 1 min,elongation temperature at 720C for 1 min, and denaturingtemperature at 92°C for 1 min. The second PCR amplificationwas carried out with a more upstream internal oligonucleotide(oligonucleotide 737), in a similar fashion except nocomplementary anchor oligonucleotide was added, and theannealing step and elongation step was carried out at the sametemperature (720C) for 2 min. PCR amplified products were gelpurified and subcloned into Bluescript KS II plasmid, which hasbeen digested with EcoRV and terminally transferred with dTTPin the presence of Taq DNA polymerase. Positive clones weresequenced with Sequenase 2.0 from USB.

RNase protectionAn EcoRI-FokI fragment (-282 to +26) and a reversetranscriptase-PCR product from oligonucleotide 486 and 1314(-53 to + 107) were each subcloned into Bluescript KS II.Uniformly labeled [ct-32P]CTP antisense RNA (cRNA) wasgenerated from HindmI-linearized plasmids with T7 RNApolymerase. A total of 4xi05 cpm antisense RNA was co-precipitated with 30 yIg RBMM total RNA, and resuspended in20,l hybridization buffer (40 mM piperazine-N,N'-bis[2-ethanesulphonic acid] (PIPES) (pH 6.4), 1 mM EDTA (pH 8.0), 0.4M NaCl, and 80% formamide). After hybridization at 540C for16 h, 300 Al of RNase digestion mix (300 mM NaCl, 10 mMTris-HCI, pH 7.5, 5 mM EDTA, 100 U/ml RNase TI, and 5,tg/ml RNase A) was added and incubated at 370C for 1 h.Proteinase K and SDS was added to a final concentration of 300tg/4l and 1%, respectively, and further incubated at 37°C for30 min after the digestion. Protected fragments were resolvedin a 6% denaturing polyacrylamide gel electrophoresis andautoradiography.

Transcriptional activity of the putative XDH/XO promoterBased on the restriction map in Fig. 3, upstream restriction siteswere utilized for the construction of XDH/XO -luciferasereporter plasmids. For the XO-256 reporter plasmid, aSacI-BanI (-25 to -256) fragment was excised from a genomicsubclone, and subcloned into a promoterless luciferase construct(pSVOARPL2L-AA5'). In a similar fashion, fragments from theSacI site to an upstream KpnI site (at about -6000), Nsil site(at about -3000), PstI site (at -555), DdeI site (at -231), BstNI

Rat TCCTGGAAT G TGCAGAGCGT GTVT2CCGGC TTGGCAGQ&.h=CMacrophage SerTrpAsn GluGluLeuL euGinSerVa lCysAlaGly LeuAlaGluG luLeuHis

Rat Liver TCCTGGAAT ---------- ---------- ---------- ------GAGG AGCTGCAGSerTrpAsn ---------------------------------GluG luLeuGln

Mouse Liver TCCTGGAAT GAGGAGTTCC TGCAGGATGT GTGTGCTGGC TTGGCAGAGG AGCTGCACSerTrpAsn GluGluLeuL euGlnAspVa lCysAlaGly LeuAlaGluG luLeuHis

Drosophila GAGTGGAGC CACCAGCTCG TGGAGCGCGT GGCGGAGAGC TTGTGCACGG AGCTGCCTmelanogaster GluTrpSer HisGlnLeuV alGluArgVa lAlaGluSer LeuCysTyrG luLeuPro

Drosophila CCTTTGGAT CATCACCTGG TGGAGCGCGT GGCGGAAAGC CTGTGCGGAG AGCTTCCApseudobscura ProLeuAsp HisHisLeuV alGluArgVa lAlaGluSer LeuCysGlyG luLeuPro

Calliphora CAATGGAAT AAGGTGCTTA TGGAGCGTGT GGTGGAAAAT CTGTWTWCAG AGTTGCCTvicina GlnTrpAsn LysValLeum etGluArgVa lValGluAsn LeuCysAlaG luLeuPro

Figure 1. Comparison of the 36 base pair insert in the rat bone mafrow macrophage(RBMM) cDNA to other known XDH/XO cDNAs. Italics represent the extra36 nucleotides found in the RBMM cDNA. Boldface letters represent the singlebase change-derived amino acid substitution. Underlined nucleotides representthe 10 base pair repeats in this region.

site (at - 124), and Hpall site (at - 116) were used to constructXO-6000, XO-3000, and XO-500, XO-231, XO-124, andXO-1 16, respectively. For XO-73, a pair of complementaryoligonucleotides (from -7 to -73) were synthesized andsubcloned into the reporter plasmid. Transient transfections wereperformed with 10 ,tg of XO-Luc or pSVOARPL2L-AA5'plasmid DNA, and 5 yg of pRSV-3Gal plasmid DNA. Thecalcium phosphate precipitation method was employed to transfectplasmid DNA into HeLa cells. After about 16 h in the presenceof calcium phosphate, cells were washed once with PBS andincubated in medium for another 24 h before harvesting cellularextract.

RESULTSIsolation of a cDNA clone for XDH/XO from rat bone marrowmacrophagesOur overall goal was to study the regulation of the activity ofXDH/XO in macrophages because of the important role that thisenzyme may play in tissue injury during inflammation (1,4- 15).Thus, we aimed to investigate if transcription, processing, andregulation of the gene might be altered in any way in these cells.Therefore, we elected to isolate an XDH/XO cDNA from ratbone marrow macrophages (RBMM) to provide us with a reagentthat could be used for screening a rat genomic library.To screen a RBMM cDNA library, we relied upon known

homologies between rat liver and Drosophila XDH/XO (15). Wesynthesized primers from two regions which show 100%nucleotide homology between the rat and Drosophila cDNAs.Since the transcriptional start site in the rat liver gene had notbeen identified, the adenine of the translational start site wasassigned as + 1. A single PCR product of approximately 440base pairs was generated from oligonucleotide 885 andoligonucleotide 886 using the RBMM cDNA library as template.This fragment was used to screen the library and nine clones wereobtained. One purified clone, GT10-XO-1, was digested withEcoRI to yield two fragments of approximately 900 and 600 bp,consistent with known restriction sites present in the rat liverXDH/XO cDNA, which contains internal EcoRI sites at +884and + 1502, respectively. The 900 and 600 bp EcoRI fragmentswere further subcloned as described in Materials and Methodsand are identified as cDNA subclone # 20 and # 10, respectively.The sequence of the isolated RBMM cDNA compares closely

with that of the rat liver cDNA isolated by Amaya et al. Thesequence of the macrophage cDNA is 100% homologous to theliver cDNA until + 1435. There are an additional 36 base pairsat this point in the rat macrophage clone. This sequence has a10 base repeat (GAGGAGCTGC) at the beginning, + 1435 to+ 1444, and the end, + 1471 to + 1480, of the 36 bp region.In addition, the second nucleotide following the second repeat(i.e. at position 1482) is changed, which results in an amino acidchange from glutamine in the rat liver cDNA to histidine in themacrophage sequence. Comparison to other known XDH/XOgenes (Fig. 1) indicates that the extra 12 amino acids in the ratmacrophage clone is the norm. Although it is not possible toexclude expression of different but very similar genes, thesimplest explanation for this observation resides either indifferential splicing (25) in which the single base change hasoccurred in alternative exons, or in a cloning deletion in the ratliver cDNA which provides, thus far, the only example of anXDH/XO cDNA missing these 36 nucleotides (see discussionbelow). Certainly, the protein sequence deduced from the recently

Nucleic Acids Research, 1994, Vol. 22, No. 10 1849

published mouse liver XDH/XO cDNA (16) differs by only oneamino acid from that described by us for the RBMM gene(Fig. 1).Complexity of the XDH/XO gene system in the genomeWe have probed various restriction enzyme digests of rat genomicDNA with fragments from the XDH/XO cDNA. The results ofsuch a Southern analysis are presented in Fig. 2A. Probing anEcoRI digest with the cDNA clone #10 (+858 to + 1518)generated a single band ofMW ca. 8 kb. A single band is alsoobserved following a BamHI digest (data not shown). However,if the same blot is re-probed with cDNA clone #20 (+5 to+863), we detect a positive signal in three bands in an EcoRIdigest, ofMW ca. 4.7, 4.2 and 3.9 kb, as well as a like numberof bands ofMW ca. 8.0, 1.8 and 0.7 kb from a BamHI digest.There are several possible interpretations of these results. Theremight be a single gene for XDH/XO, which consists of codingsequences that are extensively interrupted toward the 5' end withfairly large introns which contain restriction sites leading tomultiple fragments. On the other hand, there could be at leastthree different copies of XDH/XO-related sequences. Also, theexistence of pseudogenes cannot be excluded at this point.To distinguish between these possibilities, further Southern

hybridization was done using short oligonucleotides whichrecognize localized sequences at different positions along thecDNA between +5 and + 863 bp and on average are less likely

to span multiple exons. The Southern analysis indicated that eacholigonucleotide was able to recognize a specific, individual bandwhich together make up the complement of bands recognizedby the 5' portion of the cDNA (Fig. 2B). This indicates that the5' heterogeneity can most simply be explained by introniccomplexity in the 5' terminus of XDH/XO. Thus, we concludethat we are most likely dealing with a single copy gene whichis extensively interrupted by introns, a conclusion which isdirectly supported by the cloning studies described below.

Isolation of rat XDH/XO genomic clonesWe utilized the isolated cDNA clone # 20 as a probe to screena rat genomic DNA library. Two positive plaques were identifiedand subsequently purified to homogeneity. At the present timeonly one of these clones (XO genomic clone II) has beencharacterized. This clone contains approximately 23 kb of ratDNA. It hybridized with the 5' portion of the XDH/XO cDNA.In contrast, no hybridization was noted to the 3' portion of thecDNA (+858 to + 1518). Accordingly, clone II was analyzedin detail as we were most interested in obtaining information aboutthe promoter of this gene.The restriction map of a 17 kb SalI-Sall fragment, obtained

from a Sall digest of the 23 kb insert in clone II, is shown inFig. 3. The positions of various oligonucleotides used in itsanalysis are also indicated in the figure. Comparison of thesequence information from the genomic subclones and the cDNA

A ~~~cDNA probe +858 to+4 to +863 40 w +1 51 8.::i.;: : ..kb .q....9

7: < 7.2 >

;: 3

..1 -4

885 1 314 1 292E P E P E P

..........E

886 296 Probe

E P E p RestrictionEnzyme

_- 6500

4Q 3000

*0-- 2300Q 2000

40l4- 1200

Figure 2. (A) Southern blot analysis of the XDH/XO gene. About 20 Ag of rat liver genomic DNA were digested with restriction enzymes as indicated, subjectedto 1% agarose gel electrophoresis and transferred to a nylon membrane. Two different cDNA clones (number as indicated) were randomly labeled with [a-32P]dATPand used for hybridization. Arrow markers indicate apparent molecular size in kilo-base pairs. (B) Southern blot of the XDH/XO gene. EcoRI (E) or PstI (P) digestedrat liver genomic DNA were separated in a 1% agarose gel and transferred as described in Materials and Methods. Probes used for hybridization were generatedfrom specific oligonucleotide primer (as shown) labeling on different genomic subclones. Arrow markers indicate apparent molecular size in base pairs.

B

n0m

:;.

3n0

1850 Nucleic Acids Research, 1994, Vol. 22, No. 10

clone revealed that the 5' portion of the XDH/XO gene has indeedbeen isolated, and that it is highly interrupted by introns. In factthe 17 kb SaII-Sall fragment contains only 191 bp of the ratliver cDNA sequences. The 191 bp exonic sequences are dividedamong three exons with sizes of 68 bp (-26 to +42, exon 1),58 bp (+43 to + 100, exon 2), and 65 bp (+ 101 to + 165, exon3), respectively. The estimated sizes for introns 1 and 2 are about8 kb and 3 kb, respectively. The 3' limit of the genomic insertis in all probability at a SalI site, which is located just insidea third intron as indicated in Fig. 3. Although we know fromSouthern analysis that oligonucleotides for coding regions justdownstream of the 191 bp exon position do not detectcomplementary sequences in the genomic clone TI, we cannotexclude the possibility that the genomic insert in clone II containsan additional intron 3' to the SalT site prior to terminating at yetanother, more downstream, Sall site. In fact, from the ratgenomic DNA Southern analysis (Fig. 2B), we notedoligonucleotide 1292 and oligonucleotide 886 hybridized todifferent bands in the PstI digest. Since there is no PstI site locatedbetween oligonucleotide 1292 and oligonucleotide 886 in thecDNA sequence, these differences indicate the presence of anintron (intron 3) which lies between oligonucleotide 1292 andoligonucleotide 886, and evidently this intron contains a PstI site.Preliminary study indicates that the size of intron 3 is about 3kb (unpublished data).Sequence information on the genomic organization of the 5'

end of the XDH/XO gene is presented in Fig. 4A. Most of thesequencing activity was concentrated on the exons and the regionsimmediately surrounding them. Analysis of the intron/exonborders is presented in Fig. 4B. Both acceptor sequences are fullycanonical (AG) and two of the donor sequences are likewise (GT).In contrast, the sequence immediately following the third exon

(CGTTG) is unusual and appears to diverge from the establishedrules. This raises the possibility that the gene we have clonedis in fact not spliced at this point.

Determination of the XDH/XO transcriptional start siteThe sequence upstream of the 5' cDNA terminus in the rat livergenomic clone is shown in Fig. 4A. Inspection of the sequence

c694

&*1314 1292 d!.0 I

shows potential splice acceptor sites at -24 and -59 with respectto the adenine of the translation initiator codon. Since theorganization of untranslated parts of the XDH/XO message isnot known, the presence of potential splice acceptor sites raisedthe possibility that there might be another intron in the 5'untranslated region (UTR). The size of the XDH/XO mRNAobserved in this study and two others (16,17) is approximatelyS kb, which is larger than the rat liver cDNA reported by Amayaet al. (15). Further, none of the rat cDNA sequences include

A A1-2

TMGCA TG CCAAOGCOCO GCCSCCCTA ATCTAAGCCT-202AGTAGACCAC OTTCCC CCGAGGT TAQ&TCOGA

GTGCACTTTO ACTAGGAGG CAGAOTAGO& SSOCCO-162CACGTGAAAC TGTCCTCCC GTCTCLTCCY AAACOGCCA

Eta-1 WXL-6

GGCTOTCCG TCCTCCTO ATgCAA CCTGTOACTC-122CCGACLGGOC AGGAAGGCC TAACACOTTT GOACACTOAG

TTOCCAAGAA CCOTCCATOC CTcCOO T 102AACGGTTCTT GOCAGOTACG OACCTCGOCC ACAAT

t

cCA&T l3 > tTTGOGTAACT T STCATT OCOCTATCTTT-62AACCCATTGL AA&&G?AA AACQACCCTC COCATAGAAA

>t > > t **t * *

CAAGTTGCAG GGCAGTGATA TCTACAACTA CTTCTCAAGA-22GTTCAACGTC CCOTCACTAT AOA$0TTOAT QLAAGTTCT

* t>t metQCTCAGTGAC TCCAOCAGCC ACGATQACTG CGATGAGTT+17CQaGTCACTO AGGTCOTCGG TGCTACTGAC OCCTACTCAA

+42GGTCTTCTTT OTOAATOOCA AAAAGgtaag cagg .....

CCAGAAGAAA CACTTACCGT TTTTCcattC gtcc .....

.43Intron 1 (- 8 kb) ... ttttcttggt LgOTOGTGGA..................... aaaagaacca tcCACCACCT

GAAAA&TGCG GACCCTOAAA CAACACTTSCT GQTCTACCTOCTwTTACOC CTOGGACTTT OTTOnGALG CCAGATGGAC

+100AGAAGAAAOT gtatcctga. Intron 2 (- 3 kb). aaTCTTCTTTCA cataggact .................. tt

+101gtggttgttc agTOGOOCTA TGTSGLCCA AOCTTGOCTGcacceaceag tcACCCCOAT ACACCCTGOT TCGIAACCAC

.165

TOGAGAAOGT GOCTGTG_G CATOCACCOT QATGATCcgtACCTCTTCCA CCGACACCCC GTACOTGGCA CTACTAGgca

tgacctgcac tcg .... Intron 3 ....

actggacgtg agc .................

Bconsonsus Zzon .... gtragt ..... Intron .. (Y)Yag... Etonsequence

(Zion 1) --CAAAAatagacgg Intron1 ttcttggtagGlGGTGGAGA (axon 2)

(ZIon 2) G&AQAA&GTgtatcctga Intron 2 ggttgttcagTGGGGCTATG (Zxon 3)

(Zzon 3) ATGATGaTCcgttgacct Intron 3

= intron

-~ - exon= promoter/s nhancer region

= oligo primer

Figure 3. Structure of the 5' end of the XDH/XO gene. A 17 kb Sail-Sallfragment derived from the genomic clone II was digested with different restiction

enzymes and subcloned into Bluescript II KS plasmid (Stratagene) for sequencingand further studies. Filled boxes represent the exons; opened boxes representthe intron; and the hatched box represents the promoter/enhancer region ofXDH/XO. Arrows represent polarity of oligonucleotides used for sequencing andprimer extension.

Figure 4. (A) Nucleotide sequences of the 5' end ofXDH/XO gene. Uppercaseletters indicate the promoter/enhancer or the coding sequences while lowercaseletters represent the intervening sequences. Putative transcriptional factor bindingsites are scored on the top of the sequence with the abbreviated names. Trnslatonalstart site (ATG) is represented in italics and the adenine is numbered as + 1.Arrow heads represent the transcriptional start sites identified by using anchor-PCR method. The cross marks indicate transcriptional start sites identified byusing primer extension (see Fig. SA). The asterisk indicates the transcriptionalstart sites identified by RNase protection assay (see Fig. SB). Abbreviations used:ISRE, interferon-a stimulatory response element. (B) Exon/intron boundary ofthe XDH/XO gene. Junction between exons (uppercase letters) and introns(lowercase letters) are shown. Except for the intron donor site at the end of exon3, all the other intron donor/acceptor sites are highly conserved compared to knowncanonical splice donor/acceptor sequences.

I Icc cc

K12 kb 885

cc - CCo

2o-boL. 250 bp 885

251

i %I I

cc

33 R'.. I 9 AI 11 I I I I I I

65 bp 500 bi

f]I

Nucleic Acids Research, 1994, Vol. 22, No. 10 1851

any indication of the location of the polyadenylation site at the3' end of the XDH/XO mRNA. Comparison of the recentlypublished mouse liver XDH/XO cDNA to the rat liver XDH/XOcDNA indicates the that sizes of the coding regions of these twogenes are very similar; 4005 bp in the mouse liver XDH/XO,and 3957 bp in the rat liver XDH/XO (not including the extra36 bp found in this report and the extra 9 bp found in the 3'terminus of the mouse XDH/XO). In the mouse gene, however,the size of the 3' UTR has been identified (531 bp). Taking intoaccount 250 bp for the average size of a poly A tail in mammalianmRNA, the total length from the translational start site to theend of the poly A tail in mouse XDH/XO is about 4800 bp. Thus,the estimated size for the 5' UTR in mouse liver XDH/XO, andby interpolation for the rat liver XDH/XO, is no more than 200bp, and possibly somewhat less. Nonetheless, this is sufflcientto extend beyond the splice acceptor sites discussed above, andcertainly requires consideration of the hypothesis that there isan additional intron in the 5' UTR.

Accordingly, we have attempted to identify the transcriptionalstart site of the rat XDH/XO gene using several strategies.Initially, we employed primer extension analysis witholigonucleotide 1292. Instead of a defined, extended product asshown in the positive control from the phosphoenolpyruvatecarboxykinase (PEPCK) gene, multiple extended products wereobserved (Fig. SA). Two possibilities may account for thisobservation. First, since the 5' terminus of XDH/XO message

contains somewhat repetitive sequences, these could lead to astrong secondary structure and interfere with primer extension.Second, there might be multiple transcriptional start sites in theXDH/XO gene.To test the first possibility, another oligonucleotide primer

(oligonucleotide 694), which is longer in length, was used to allowus to adopt a higher elongation temperature in primer extension.It was anticipated that the higher temperature would weaken anysecondary structure in the 5' terminus, and allow more efficientprocessivity of the reverse transcriptase. As shown in Fig. 5A,multiple extended products were still observed. The mostprominent extended products were about 20 and 40 bp beyondthe ATG site. However, longer products, from 55-83 bp inlength, could also be observed, though with a much weakersignal. This observation raises the distinct possibility that multiplestart sites exist in the XDH/XO gene.The second possibility was tested by RNase protection. Two

cRNA probes were generated; one probe contained sequencesfrom -53 to + 107 and the other from -282 to +26 bp. Theprobes were annealed to mRNA isolated from RBMM. After theRNase digestion, several protected fragments of differing sizeswere obtained (Fig. SB). The -53 to + 107 probe gave fragmentsof size 56, 88, 125, and 147 bp. The two longest fragments areconsistent with transcripts containing sequence information from-19 and -41 bp upstream of the translation initiator. The twoshorter fragments (i.e. mapping to sites downstream of the ATG)

A 1 2 3 M 4 5 6

338-

201 , :

iw '4

t.t-

t.

I

B ElongationTemp.OligoPrimer

370C 500C

1292 PEPCK Size Std 694

4- 147

4 125

.4- 88

75 _6 g9 n}b

6

e4

49 v_

.: 44 o

20_00 .1 i6 _-O --.-

*.

:.- 83

.4- 70

.4- 61

*41

_- 21,*- 1 9. - 56

Figure 5. (A) Primer extension of the XDH/XO gene. About 20 jig rat bone marrow macrophage (RBMM) total RNA was used for the study. Rat liver total RNA,with PEPCK oligonucleotide primer, were used as a positive control. Numbers on the side represent the position of extended products with respect to the XDH/XOtranslational start site. (B) RNase protection of the XDH/XO gene. Antisense RNA were synthesized as mentioned in Materials and Methods. Lane 1, -282 to+26 cRNA probe was generated from a genomic subclone that contains the 5' terminus of XDH/XO; lane 2, 30 t&g RBMM total RNA digested with 100 U/mlRNase TI and 5 itg/ml RNase A; lane 3, cRNA alone; lane 4, -53 to + 107 cRNA probe was generated from a RT-PCR product subclone; lanes 5-6 weresimilar to lanes 2-3 except for the differences of cRNA probe; lane M, size marker. Arrows indicated the sizes of different protected products.

1852 Nucleic Acids Research, 1994, Vol. 22, No. 10

may well reflect inability of the cRNA probe to hybridizeextensively because of secondary structure in the XDH mRNA.The -282 to +26 probe yielded protected fragments of size 49,52, 54, and 69 bp which correspond to RNA molecules extendingfrom 23 to 43 bp upstream of the ATG codon.The results from the primer extension and RNase protection

are difficult to interpret. In all likelihood, strong secondarystructure in the 5' portion of the XDH/XO mRNA is generatingpremature termination in primer extension as well as decreasedhybridization in the RNase protection experiments, and we areunable to resolve whether we are detecting several differenttranscription initiation points or strong secondary structure. Eventhe observation that the start sites mapped by both techniquesare the same is subject to the same concern.Our fumdamental goal in attempting to define the transcriptional

start site lay in our intent to identify the promoter for this gene.Since the region immediately upstream of the ATG contains twopossible splice acceptor sites (at -24 and -59), if we couldobtain evidence that the primer extended products extended intoknown 5' sequences beyond -59, then we could concludedefinitively that transcription initiation must have occurred at anearby site. Hence, the possibility that the promoter and an

A-6000 L rmerte

-3000

-500

-256

_2ZzZZZZZZ722z

Promoterle"rEM o

B-256 -2 5 Lucife reorter

-230 -25_ /E~000277777-1 23

-1 1 6

Promoter activity in HeLa cells

72 ± 27

55 ± 19

30 ± 6.2

100

2.8 ± 1.3

Prornoter activity in HeLa cells

100

118 ± 38

-25

unidentified first exon were located in an undefined upstreamposition could be eliminated. Accordingly, we have utilizedanchor-PCR to isolate those primer extended products which aregreater than 59 nts. These PCR products were sequenced andthe identified transcriptional start sites are summarized in Fig.3B, along with the other transcriptional start sites identified byprimer extension and RNase protection. The results indicate thatonly immediately 5' sequences are found in the mRNA and thereis no evidence of additional splicing. Thus, we conclude that,at least in RBMM, the XDH/XO promoter is located 20-85 bpupstream of the ATG. Also, we would argue that there is tentativeevidence for the existence of a number of separate start sites inthis region as summarized in Fig. 4A.

Identification of the XDH/XO promoterInspection of the sequences upstream of the 65 bp domain oftranscription initiation indicates the presence of a considerableconcentration of cis-acting binding sites typically found ineucaryotic promoters (Fig. 4A). These include the sites for theubiquitous factors AP-2, CCAAT factor, and NF-IL6. In additionto the ubiquitous sites, there is a 10/11 match to the a-interferonstimulatory response element (ISRE, 31) as well as a perfectmatch to an Ets-1 binding site (26). There is, however, nocanonical TATA binding site or GC-rich sequence (Spltranscription factor binding site) in close proximity to thetranscription initiation regions.To test whether the region upstream of the translational start

site has functional promoter activity, a fragment of DNAupstream of the ATG (from -25 to -256 bp) was subclonedin front of a luciferase reporter gene (XO-256), and assayed bytransient transfection. Plasmid pRSV-,BGal was co-transfectedwith the luciferase reporter constructs as an internal control. Asshown in Fig. 6A, the XO-256 luciferase construct provided anapproximate 50-fold increase in activity compared to thepromoterless construct (the parent plasmid used for initialsubcloning). Further, insertion of additional XDH/XO upstreamfragments (XO-500, XO-3000, and XO-6000) 5' of the luciferasereporter gene, in general, leads to a decrease in promoter activity(Fig. 6A). This indicates the presence of possible negativeregulator(s) upstream of the 250 fragment.

43 ± 20-25

66 ± 34

.73 -7

.1~~~~~~~~~~~3~~~-1 17 -73

-7r3 -1 17

_ zzZ////

Promot erless r7,,,Z

2 ± 0.6

2.5 ± 0.7

0.8 ± 0.2

2 ± 0.7

Figure 6. (A) Deletion analysis of the rat XDH/XO promoter. Constructscontaining various XDH/XO upstream fragments were subcloned in front ofluciferase reporter gene, and transiently transfected into HeLa cells. (-CGalactosidasereporter gene was co-transfected as an internal control. Luciferase activity wasnormaized to (3-galactosidase and used to represent the transcriptional activityof XDH/XO promoter. (B) Deletion analysis within the 256 bp rat XDH/XOpromoter. Various restriction enzymes were used to generate a series of XDH/XOreporter constructs and then transiently transfected into HeLa cells as described.Promoterless construct repsents the parent plasmid (pSVOARPL-AA5'). At leastthree independent experiments were done and all observations were performedin triplicate. The results are all normalized to a value of 100 for the XO-256construct.

Initiator sequence

Human PBGD -2 TOAGCG +10

Rat XO/XDH -22 TCAGTGACTCCA -10t >t

Mouse DEFR -13 GATI'IGCGCCAAA rCGGCA +12

Rat XO/XDH -73 GCGTATC1I17AAGT'PCAGGGCA -50t >t > > t

Mouse rpS16 C¶CCT1 N

Rat XO/XDH -91 GIcAl'TTGc -80t

Figure 7. Comparison of upstream nucleotide sequences of XDH/XO to knowninitiator sequences. Arrow indicates the transcription start site of known initiators.Cross mark, asterisk, and arrow head represent start sites found in rat XDH/XOby primer extension, RNase protection and anchor-PCR, respectively. PBGD,porphobilinogen deaminase; DHFR, dihydrofolate reductase; rpS16, smallribosomal protein 16.

7A6

Nucleic Acids Research, 1994, Vol. 22, No. 10 1853

Further deletion analysis within the 256 bp fragment wassubsequently performed (Fig. 6B). A smaller fragment, -116bp upstream of the ATG, was identified as containing basalXDH/XO promoter function. This -116 bp fragment leads toa 30-fold increase in promoter activity compared to thepromoterless construct. Further deletion to position -73abolished the promoter activity. The -25 to -116 fragment wassub-divided into two fragments, from -73 to - 116 and from-73 to -25. Neither subfragment was capable of directingtranscription. This indicates that either, (1) the two subfragmentscontain binding activities which need to communicate in someway or (2) a critical binding domain around -73 was destroyedwhen the basal promoter was dissected into two halves.

DISCUSSIONXanthine oxidase is a molecule of considerable clinical interestbecause of its potential role in generating reactive oxygen speciesin ischemia-reperfusion tissue injury and in inflammation (1).XO is thought to be derived from a proteolysis/oxidationconversion of the protein product of XDH. Recent studies havesuggested that in addition to cytokine-mediated XDH/XOconversion, inflammatory cytokines such as interferon mightinduce transcription of the XDH/XO gene (17,18). However,no XDH/XO promoter has been identified and it has not beenpossible to study the regulation of the gene in detail.

In this report, we describe the isolation of a XDH/XO cDNAfrom a rat bone marrow macrophage library. Using this cDNAclone as a probe, we have isolated the 5'-most three exons ofthe XDH/XO gene from a rat liver genomic library. Further,we have identified the transcriptional start sites and the promoterwithin this genomic clone. In addition, several observations inthis report provide insight into the basal regulation of theXDH/XO gene.We have presented evidence based on Southern analysis that

there appears to be only one gene for XDH/XO. However, thecDNA isolated from RBMM while very similar to that isolatedby Amaya et al. (15) from rat liver, is nonetheless clearlydifferent. The macrophage cDNA contains an additional 36nucleotides (which are also seen in cDNAs from other species).In addition, there is a single nucleotide substitution downstreamof the 36 nucleotide insertion. Given the observation that thereis no indication of multiple genes, as well as the finding that theXDH/XO gene is extensively interrupted with introns, we testedfor the existence of differentially spliced messages. Using areverse transcriptase-PCR approach, we have found that thereis no evidence of differential splicing. In tissues that express highXDH/XO activity (small intestine, liver, and RBMM), weobserved no differences in terms of their sizes of PCR products,or the restriction digests of the amplified products (unpublisheddata). However, the expression of a very minor form of XDH/XOthat is missing the 12 amino acids could not be excluded. Otherpossible explanations for the missing 12 amino acids are eitherallelic variation between different rat species or a cloning variationdue to the presence of 10 bp repeats flanking the 36 bp insertthat led to a cloning artifact.Primer extension of the XDH/XO mRNA indicates multiple

transcriptional start sites which lie within a 65 nucleotide windowupstream of the translational start site. However, in primerextension experiments of this type, if the message has a regionof strong secondary structure, premature termination of reversetranscriptase may occur (28). Based on an energy minimization

algorithm (29), there is an extensive secondary structure at the5' end of the XDH/XO message. If this is the case, then onewould expect that the shortest extended cDNA will have thehighest yield and vice versa. Indeed this was observed as reportedin Fig. 5A. In addition, sequence analysis of primer extendedproducts amplified by PCR indicates that at least two of the startsites contain uninterrupted sequences from a position immediatelyupstream of the splice acceptor sites in the 5' UTR providingan indication that splicing to these sites is not occuring.Observations from the RNase protection were consistent with theprimer extension and anchor-PCR. Similar indications of strongsecondary structure at the 5' terminus are seen in the RNaseprotection assay.There is no canonical TATA box or GC-rich region within

normally expected distances from any of these transcriptional startsites in the XDH/XO gene. Comparison of sequences around thetranscription initiation sites with other genes lacking TATA boxesreveals a considerable degree ofhomology (Fig.7). These initiatorsequences have been known to direct transcription start in aTATA-less gene. Also, cooperation between such an initiatorregion and an upstream regulator greatly enhances the promoteractivity (30). This notion was substantiated in the XDH/XO genewhen the putative non-TATA initiation sequence (-73 to -25)which most resembles the canonical initiator was coupled to aluciferase reporter gene, and shown to direct a significant level(30-fold induction) of transcriptional activity in the presence ofupstream regulators (-116 to -73). In fact, a putative CCAATtranscriptional factor binding site has been identified within this-116 bp fragment. Deletion of the putative CCAAT binding siteabolished the promoter activity of the XDH/XO gene; howeverthe CCAAT sequence by itself could not initiate transcription.We conclude that the bulk of the initiation of the XDH/XO

mRNA occurs over a definable 65 bp region of the gene. Thesequences immediately upstream of what we are calling the basalpromoter were then examined for ability to modulate the basaltranscriptional activity. We find that sequences out to around-250 bp elevate transcriptional activity to 50-fold above the basalpromoter. Extension of this analysis to around -6000 bp revealeda modest diminution of transcriptional ability, suggesting thepresence of additional modulating elements. Recent observations(unpublished data) have provided evidence that substantial up-regulation in response to PMA occurs in this latter domain.

ACKNOWLEDGEMENTSWe thank Josef Ozer for his generous gift of the Southern blotshown in Fig. 2A. We also thank V.Shepherd and the membersof her laboratory for input and advice, and for the generous giftof the macrophage cDNA library. Joseph Parinello and HaijingLi provided invaluable technical support, and members of theRC laboratory provided knowhow and encouragement. We aregrateful to Drs Linda Sealy and Mark Magnuson for usefulreagents. This work was supported by the Department ofVeterans' Affairs (JR) and the National Institutes of Health (RC).

REFERENCES1. Parks D., and Granger D.N. (1986) Acta. Physiol. Scand. 548, 87-99.2. Kuppusamy, P., and Zweier, J. (1989) J. Bio. Chem. 264, 9880-9884.3. Haber, F., and Weiss, J. (1934) Proc. R. Soc. Lond. (A) 147, 332.4. Granger, D.N., Rutili G., and McCord J.M. (1981) Gastroenterology 81,

22-29.5. McCord, J.M. (1985) N. Engl. J. Med 312, 159-1263.

1854 Nucleic Acids Research, 1994, Vol. 22, No. 10

6. Rowe G.T., Eaton L.R., and Hess M.L. (1984) J. Mol. Cell. Cardiol. 16,1075-1079.

7. Mullane K.M., Read N., Salmon J.A., and Moncada S. (1984) J. Phannacol.Ezp. Ther. 228, 510-522.

8. Hess M.L., Rowe G.T., Caplan M., Romson J., and Lucchesi B. (1985)Adv. Myocardiol. 5, 159-175.

9. McCord, J.M. (1987) Fed. Proc. 46, 2402-2406.10. Grum C., Gross T., Mody C.H., and Sitrin R.G. (1990) J. Lab. Clin. Med.

116, 211-218.11. Akaike T., Ando M., Oda T., Doi T., Ijiri S., Araki S.,and Maeda H. (1990)

J. Clin. Invest. 85, 739-745.12. Friedl H.P., Till G.O., Trentz O., and Ward P.A. (1989) Am. J. Pathol.

135, 203-217.13. Till G.O., Friedl H.P., and Ward P.A. (1991) Free Radical Biol. Med. 19,

379-386.14. Grum C.M., Ragsdale R.A., Ketai L.H., and Simon R.H. (1987) J. Crit.

Care 2, 22-26.15. Amaya Y., Yamazaki K., Sato M., Noda K., Nishino T., and Nishino, T.

(1990) J. Biol. Chem. 265, 14170-14175.16. Terao M., Cazziniga G., Ghezzi P., Bianchi M., Falciani F., Perani P.,

and Garatdni E. (1992) Biochem. J. 283, 863-870.17. Dupont G.P., Huecksteadt T.P., Marshall B.C., Ryan U.S., Michael J.R.,

and Hoidal J.R. (1992) J. Clin. Invest. 89, 197-202.18. Falciani, F., Ghezzi, P., Terao, M., Cazzaniga, G., and Garattini, E. (1992)

Biochem. J. 285, 1001-1008.19. Lee C.S., Curtis D., McCarron M., Love C., Gray M., Bender W., and

Chovnick A. (1987) Genetics 116, 55-66.20. Houde M., Tiveron M., and Bregegere F. (1989) Gene 85, 391-402.21. Christman J.W., Abdohrasulnia R., Shepherd V.L., Rinaldo J.E. (1991)

Prostaglandins 41, 252-262.22. Chung K.N., Shepherd V.L., Stahl P.D. (1984) J. Biol. Chem. 259,

14637-14641.23. Schreiber S., Blum J.S., Stenson W.F., McDermott R.P., Stahl P.D.,

Teitelbaum S.L., Perkins S.L. (1991) Proc. Natl. Acad. Sci. USA 88,1616-1620.

24. Troutt, A., McHeyzer-Williams, M.G., Pulendran, B., and Nossal, G.J.V.(1992) Proc. Natl. Acad. Sci. USA 89, 9823-9825.

25. Breitbart R.E., Andreadis, A and Nadal-Girard, B. (1987) Ann. Rev. Biochem56, 467-496.

26. Karim, F. D., Umess, L.D., Thummel, C. S., Klemsz, M. J., McKercher,S. R., Celada, A., Van Beveren, C., Maki, R. A., Gunther, C. V., Nye,J. A., and Graves, B. J. (1990) Genes Dev. 4, 1451-1453.

27. Karnitz, L., Poon, D., Weil, P.A., and Chaildey, R. (1989) Mol. Cell. Biol.9, 1929-1939.

28. Pallanwh L., Beswick, H., Talian, J., and Zelenka, P. (1989) Anal. Biochem.185, 57-62.

29. Jaeger, J.A., Turmer, D.H., and Zuker, M. (1989) Proc. Natl. Acad. Sci.USA 86, 7706-7710.

30. Kollmar, R., and Farmham, P.J. (1993) Proc. Soc. Exp. Biol. Med. 203,127-139.

31. Hug, H., Costas, M., Steheli, P., Aebi, M., and Weissman, C. (1988) Mol.Cell. Biol. 8, 3065-3079.