Mammalian PASKIN, a PAS-Serine/Threonine Kinase Related to Bacterial Oxygen Sensors

MR

TI*S

R

mooghmndcwaearPtdPs

ts

atcmtbu

i

dL1

Biochemical and Biophysical Research Communications 288, 757–764 (2001)

doi:10.1006/bbrc.2001.5840, available online at http://www.idealibrary.com on

ammalian PASKIN, a PAS-Serine/Threonine Kinaseelated to Bacterial Oxygen Sensors

homas Hofer,*,†,1 Patrick Spielmann,*,1 Petra Stengel,‡ Bettina Stier,‡ Dorthe M. Katschinski,‡sabelle Desbaillets,*,† Max Gassmann,*,† and Roland H. Wenger*,‡,2

Institute of Physiology and †Institute of Veterinary Physiology, University of Zurich, CH-8057 Zurich,witzerland; and ‡Institute of Physiology, Medical University of Lubeck, D-23538 Lubeck, Germany

eceived September 24, 2001

slight reductions in normal oxygen concentrations(iiilHpardsosvk(r

pactmpefsambsotsds

si

The PAS domain is a versatile protein fold found inany archaeal, bacterial, and plant proteins capable

f sensing environmental changes in light intensity,xygen concentration, and redox potentials. The oxy-en sensor FixL from Rhizobium species contains aeme-bearing PAS domain and a histidine kinase do-ain that couples sensing to signaling. We identified aovel mammalian PAS protein (PASKIN) containing aomain architecture resembling FixL. PASKIN is en-oded by an evolutionarily conserved single-copy genehich is ubiquitously expressed. The human PASKINnd mouse Paskin genes show a conserved intron–xon structure and share their promoter regions withnother ubiquitously expressed gene that encodes aegulator of protein phosphatase-1. The 144-kDaASKIN protein contains a PAS region homologous to

he FixL PAS domain and a serine/threonine kinaseomain which might be involved in signaling. Thus,ASKIN is likely to function as a mammalian PASensor protein. © 2001 Academic Press

Key Words: gene expression; hypoxia-inducible fac-or; iron; LOV-domain; nitrogen fixation; oxygen sen-or; PAS-domain; protein phosphatase.

Adaptation to environmental changes is essential toll organisms for survival and evolution. This adapta-ion requires specific sensors coupled to signaling cas-ades. While the sensing principles for some environ-ental parameters (such as light, sound, tension, or

emperature) are relatively well known, others are justeing detected. Until recently, for instance, it has beennclear how mammalian cells sense oxygen. Even

The mouse Paskin genomic nucleotide sequence has been depositedn the EMBL/GenBank/DDBJ databases (Accession No. AJ318757).

1 The first two authors contributed equally to this work.2 To whom correspondence and reprint requests should be ad-

ressed at Institut fur Physiologie, Medizinische, Universitat zuubeck, Ratzeburger Allee 160, D-23538 Lubeck, Germany. Fax:49 451 500 4171. E-mail: [email protected].

757

hypoxia) can cause the induction of specific genesnvolved in mammalian oxygen homeostasis (reviewedn Refs. 1, 2). Under hypoxic conditions, hypoxia-nducible factors (HIFs) serve as transcriptional regu-ators of these genes. Under normoxic conditions,IF-a subunits are bound by the von Hippel-Lindaurotein (pVHL) and subjected to rapid ubiquitinylationnd proteasomal degradation (3). Binding of pVHLequires oxygen- and iron-dependent prolyl-hy-roxylation of a conserved proline residue of HIF-a,uggesting that a prolyl-4-hydroxylase serves as thexygen sensor that regulates the stability of HIF-aubunits (4, 5). However, because both the transacti-ation activity of HIF-a as well as the activity of manyinases are also regulated by oxygen concentrations6–10), we postulate additional oxygen sensor(s) thategulate hypoxia-inducible phosphorylation of HIF-a.In the nitrogen-fixing bacteria Bradyrhizobium ja-

onicum and Rhizobium melilotti, the oxygen sensor ishistidine kinase termed FixL which under anaerobic

onditions phosphorylates and thereby activates theranscription factor FixJ (11). The oxygen sensing do-ain of FixL is a so-called PAS domain, a versatile

rotein fold resembling to a left-handed glove thatncloses a heme cofactor (12). When oxygen is bound toerrous heme, this domain is in the flattened “off”tate, resulting in a change of the protein conformationnd inactivation of the kinase domain. The PAS do-ain is very often found in sensor proteins of archaea,

acteria and plant species, where the correspondingensor proteins measure light, oxygen and reduction–xidation (redox)-potentials. These PAS sensor pro-eins often contain kinase domains, allowing to linkensing with signaling. While the structure of the PASomain is highly conserved, there is only little primaryequence conservation (reviewed in Refs. 13–16).In mammalian proteins, the PAS domains identified

o far seem to serve primarily as heterodimerizationnterfaces of transcription factors involved in the xeno-

0006-291X/01 $35.00Copyright © 2001 by Academic PressAll rights of reproduction in any form reserved.

biotic response, circadian rhythm and hypoxic adapta-tNeancid

M

DfcNe(mPCePt

(welsfhSplPswd(

uHgbB(PNcwwp

pspbSpaioe

In vitro translation, stable transfection, and immunoblotting. TheoAtP(fe

iEi(eNm

pltidhdmm

R

C

lBcmsfbp(sctcTPpHtpt

onsR(btsi

Vol. 288, No. 4, 2001 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ion. In fact, both HIF-a and HIF-b subunits contain-terminal PAS domains that are essential for het-

rodimerization. Because the PAS domain turned upgain in signaling pathways related to sensor mecha-isms, we wondered whether it might be evolutionarilyonserved also in the sensing mechanism itself. Wedentified a single mammalian gene containing a PASomain and a serine/threonine kinase domain.

ATERIALS AND METHODS

Cloning and sequencing of PASKIN. The cDNA cloneKFZp434O1522 encoding full-length human PASKIN was obtained

rom the Resource Center/Primary Database (Berlin, Germany). TheDNA clone HA0/203 encoding KIAA0135 was a kind gift from T.agase (Chiba, Japan). The IMAGE cDNA clone IMAGp998H022323

ncoding part of mouse PASKIN was obtained from Research GeneticsInvitrogen) and the insert was completely sequenced. Part of the

ouse PASKIN cDNA was amplified by RT-PCR using the primersaskin5.1 (59-CCACCTTCCCTCTCAGTTTG-39) and Paskin3.1 (59-AGCTCCAACTGAGCTTCCT-39). Total RNA derived from mouserythroleukemia (MEL) cells served as template for the RT-PCR. TheCR product was subcloned into the SmaI site of pUC19 resulting inhe plasmid pmPaskin-PCR1.

The mouse Paskin gene was cloned from a genomic l phage librarykind gift of U. Muller, Zurich, Switzerland) prepared from DNAhich has been isolated from the mouse strain 129Sv(ev)-derivedmbryonic stem cell line AB-1, partially digested with Sau3AI andigated into the vector LambdaGEM-11 (Promega). This library wascreened twice by plaque hybridization to gel-isolated probes derivedrom the mouse (resulting in the l phage clones lP1 to lP21) or theuman (lP22 to lP36) cDNA. The mouse probe was the 1386 bpalI–NotI fragment from mouse IMAGp998H022323 and the humanrobe was the 1060 bp XhoI–BamHI fragment from human HA0/203,abeled by random-primed labeling with [a-32P]dCTP (Hartmann).ositive l clones were plaque purified, and the XhoI fragments wereubcloned into pBluescript vectors (Stratagene). The Paskin geneas sequenced on both strands using automated sequencing proce-ures according to the instructions provided by the manufacturerApplied Biosystems).

DNA blot analysis. Genomic DNA was isolated and analyzedsing standard techniques (17). Bacterial DNA was a kind gift of.-M. Fischer (Zurich, Switzerland). If not otherwise indicated,

enomic DNA (10 mg) was digested overnight with EcoRI, separatedy electrophoresis through 0.7% agarose gels, blotted to unchargediodyne A membranes (Pall) and cross-linked by UV irradiation

Stratalinker, Stratagene). The blots were hybridized to the humanASKIN probe in 63 SSC, 103 Denhardt’s, 0.1% SDS, 1.1 mMa4P2O7, 17 mM Na2HPO4/NaH2PO4 (pH 7.7), and 200 mg/ml soni-

ated salmon sperm DNA (Sigma) for 15 h at 63°C. The blots wereashed for 1 h in 0.13 SSC, 0.2% SDS at 45°C, exposed, and re-ashed for 1 h at 65°C. Radioactive signals were recorded by phos-hoimaging (Molecular Dynamics).

RNA blot analysis. Total RNA isolation and Northern blottingrocedures were performed as described previously (18). Mouse tis-ue RNA was subsequently hybridized to the mouse PASKIN cDNArobe and to a ribosomal protein L28 cDNA probe (18). The mem-ranes were washed to a final stringency of 50°C in 0.13 SSC, 0.2%DS. A multiple human tissue dot blot containing 115 to 659 ngoly(A)1 RNA per dot was hybridized to the human PASKIN probeccording to the manufacturer’s instruction (Clontech). Final wash-ng stringency was 1 h at 65°C in 0.13 SSC, 0.2% SDS. The amountf poly(A)1 RNA was normalized to the mRNA expression levels ofight different housekeeping genes.

758

ligonucleotide 59-CATGGTGAGAGGATCGCATCACCATCACCATC-CGC-39, encoding the peptide sequence MVRGS(H)6A, was annealed

o its antisense strand and inserted into the NcoI site of the humanASKIN cDNA clone DKFZp434O1522. The resulting plasmid

pSportHis2PASKIN) was linearized with NotI and used as a templateor coupled in vitro transcription/translation (IVTT) in wheat germxtracts according to the manufacturer’s instructions (Promega).The expression vector pcDNA3His2PASKIN was constructed by clon-

ng the blunted MluI fragment from pSportHis2PASKIN into thecoRV site of pcDNA3 (Invitrogen). pcDNA3His2PASKIN was linear-

zed with PvuI and stably transfected into Chinese hamster ovaryCHO) cells as described previously (19). Total cellular protein wasxtracted with 10 mM Tris/HCl (pH 8.0), 1 mM EDTA (pH 8.0), 400 mMaCl, 0.1% NP-40, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 1M Na3VO4, and 1 mg/ml each aprotinin/leupeptin/pepstatin A.IVTT and CHO protein extracts were fractionated by 6% SDS–

olyacrylamide gel electrophoresis and transferred to a nitrocellu-ose membrane (Amersham–Pharmacia). The membrane wasreated with blocking solution (4% defatted milk powder in PBS) andncubated for 1 h with a monoclonal anti-RGS-His antibody (Qiagen),iluted 1:1000 in blocking solution, followed by a rabbit anti-mouseorseradish peroxidase-coupled secondary antibody (Santa Cruz),iluted 1:1000, which was detected by chemiluminescence with 100M Tris–HCl, pH 8.5, 2.65 mM H2O2, 0.45 mM luminol, and 0.625M coumaric acid for 1 min, and exposure to X-ray films (Fuji).

ESULTS

loning of Human PASKIN

With the aim of identifying mammalian proteins re-ated to FixL, a BLAST search was performed using the. japonicum FixL sequence (EMBL/GenBank/DDBJ Ac-ession No. P23222) as query. A single mammalian ho-ologue was identified which we named PASKIN. The

equence of the human PASKIN cDNA was assembledrom overlapping cDNA sequences deposited in the data-ases as a result of several large-scale sequencingrojects. The cDNA clone HA0/203 encoding KIAA0135Accession No. D50925) contained most of the PASKINequence and was completed using the DKFZp434O1522DNA (Accession No. AL043004) which contained addi-ional 309 bp (88 amino acids) at the 59 end, including aonsensus TCCCATGG translation initiation codon (20).his assembled cDNA sequence, designated humanASKIN, was 4306 bp long and encoded a predictedrotein of 1323 amino acids (Fig. 1A). In addition, theSU79240 cDNA clone (Accession No. U79240) was ex-

ended at the non-translated 39 terminus by 193 bp, mostrobably because of the use of alternative polyadenyla-ion sites.

The human PASKIN gene was located to a previ-usly reported 175-kb fragment of the human ge-ome (Accession No. AC005237). Sequence compari-on confirmed the assembled cDNA sequence.ecently, a novel PASKIN cDNA has been reported

Accession No. AF387103) which was extended by 29p at its 59 end. This fragment represents the un-ranslated first exon which was located 6.4 kb up-tream of the second exon containing the translationnitiation codon. Thus, PASKIN consists of 18 exons

seeps

C

wfimullPc(mst

t3rpoot

atsp

taiwmtbNthui

P

Pocssddw

rmtN


pread over 43.3 kb (Fig. 1B). Interestingly, the firstxon of the PPP1R7 gene (Accession No. AF067130),ncoding a putative mitotic regulator of proteinhosphatase-1 (21), was located only 1.0 kb up-tream of the first exon of PASKIN (Fig. 1B).

loning of Mouse PASKIN

The IMAGE clone 934321 (Accession No. AA544838)as identified by BLAST analysis and was obtained

rom an IMAGE clone depository. Sequencing of thensert revealed that it encodes 1.5 kb of the 39 end of

ouse PASKIN beginning at exon 14. The insert wassed as a probe to screen a genomic mouse l phage

ibrary and 21 phage clones (lP1 to lP21) were iso-ated. This library was then re-screened using a humanASKIN 59 probe derived from the DKFZp434O1522DNA clone, resulting in 15 additional phage cloneslP22 to lP36). The location of the phage clones wasapped (Fig. 1C) and the mouse Paskin gene was

equenced on both strands by a combination of restric-ion fragment subcloning and primer walking.

The intron–exon structure is well conserved betweenhe human and mouse genes (Fig. 1B). The intron 59 and9 splice sites of both genes conformed all with the GT/AGule (22). As deduced by sequencing of the plasmidmPaskin-PCR1, containing part of the mouse cDNA, thenly difference in the splice site locations is the presencef additional 5 bp at the end of mouse exon 9 and addi-ional 208 bp at the beginning of exon 10 (numbering

FIG. 1. Primary structure of PASKIN. (A) Human PASKIN cespectively, as well as the PAS repeats and the serine/threonine kiouse Paskin. Exons are indicated by black boxes numbered accord

entative and could not be deduced by homology to the human firstote the different scale bars.

759

ccording to the human sequence). This results in addi-ional 71 amino acids in the predicted mouse PASKINequence, and the assembled mouse cDNA encodes for aredicted protein of 1390 amino acids.An alternative splice acceptor 21 bp downstream of

he start of intron 14 was identified in both the humannd mouse genes, leading to a 7-amino-acid insertionn the serine/threonine kinase domain of KIAA0135hich is not present in other database entries of hu-an PASKIN cDNAs. By sequence comparison with

he tentative mouse genome sequence (Celera data-ase), we found that the mouse Ppp1r7 gene (Accessiono. AF067129) is located at a similar position relative

o Paskin gene as it is found in the correspondinguman locus. Therefore, we predict the existence of anntranslated, short mouse first exon at a position sim-

lar to the human first exon (Fig. 1B).

ASKIN Is Encoded by an Evolutionarily ConservedSingle-Copy Gene

As shown by Southern blot analysis, the humanASKIN PAS domain probe hybridized to a single bandf the expected size with human genomic DNA, andross-hybridized also with genomic DNA derived fromeveral rodent species (Fig. 2A), making it suitable forcreening the mouse l phage library (see above). Aetailed restriction mapping of mouse genomic DNAerived from 129 or C57Bl/6 strains and hybridizationith a mouse probe resulted in bands of the expected

A. The ATG translation initiation and TAA termination codons,e domain are indicated. (B) Gene structure of human PASKIN and

to the human sequence. The mouse first exon (in parentheses) isn. (C) Map of the cloned l phages encoding the mouse Paskin gene.

DNnasing

exo

ss(denhbpgre

M

Ptcasmttwmn

P

t

bFrstdpb(oaisdhwilcmitc

P

e1t

h(


ize without polymorphisms between the two strains,uggesting that mouse Paskin is a single copy geneFig. 2B). These data could be confirmed by recentatabase analyses of the human and mouse genomentries, which did not result in additional matches. Ofote, the human PAS domain probe also cross-ybridized with genomic DNA derived from distinctacterial species, including Escherichia coli and B. ja-onicum (Fig. 2C), under the same hybridization strin-ency used for the human and rodent blots. Theseesults indicate evolutionary conservation of the genencoding PASKIN.

ammalian PASKIN Is Related to FixL

Using ProfileScan analysis, two N-terminal consensusAS repeats, a PAS associated sequence in-betweenhem, and a C-terminal serine/threonine kinase domainould be identified (Fig. 3). Human and mouse PASKINre highly similar within these domains, but share lessimilarity outside of them. An alignment of human andouse PASKIN to the B. japonicum FixL sequence using

he Clustal W program revealed sequence conservation ofhe PAS repeats (Fig. 3). The FixL histidine kinase regioneakly aligns to the long exon 10 of both human andouse PASKIN, rather than to the serine/threonine ki-ase regions (Fig. 3).

ASKIN Is Ubiquitously Expressed

Northern blot analyses were performed to determinehe mRNA size and the expression pattern of PASKIN in

FIG. 2. PASKIN is encoded by an evolutionarily conserved singluman and rodent species (A), mouse (B), and several bacterial specieA, C) or to a mouse PASKIN probe (B).

760

oth adult and embryonic mouse tissues. As shown inig. 4, the mouse PASKIN hybridization probe revealed aelatively weak band of approximately 4.5 kb in all tis-ues examined. To obtain a more complete overview onhe tissue expression pattern, a human multiple tissueot blot was hybridized with the human PAS domainrobe under high stringency conditions that allowed hy-ridization with human but not mouse PASKIN DNAdata not shown). The dot blot contains normalized levelsf poly(A)1 mRNA from a total of 61 different humandult tissues, 7 fetal tissues and 8 cell lines. As depictedn Fig. 5, the PASKIN probe hybridized with all of theseamples, yielding relatively uniform signal intensities. Asetermined by phosphoimaging analysis, 2- to 4-foldigher expression levels compared to the mean valueere found in caudate nucleus and putamen of the brain,

n prostate and in testis; and a 5-fold reduced expressionevel was found in placenta. The PASKIN probe did notross-hybridize with yeast RNA, human poly r(A) or hu-an C0t-1 DNA, confirming the specificity of the hybrid-

zation. In contrast, as shown by genomic Southern blot-ing (Fig. 2), the PASKIN probe cross-hybridized with E.oli DNA.

ASKIN Is a Soluble 144-kDa Protein withoutExtensive Posttranslational Modifications

The open reading frame of the human PASKIN cDNAncodes for a protein of a predicted molecular weight of44 kDa. To examine whether PASKIN is posttransla-ionally modified, a RGSH6 tag was added to the

py gene. Southern blot analyses of the genomic DNA isolated from). The blots were hybridized to a human PASKIN PAS domain probe

e-cos (C

NdR

was synthesized in vivo and in vitro, by stable transfec-tbsbb(bdlqlt

D

aplmmspftftohPptedt

ilapu

(akhasa


-terminus of the human PASKIN cDNA allowing theetection of the fusion protein with a monoclonal anti-GSH6 tag antibody. RGSH6-tagged PASKIN protein

FIG. 3. Alignment of human and mouse PASKIN. PAS domainunderlined) and PAS-associated domains (dashed line) of PASKINligned along with Bradyrhizobium japonicum FixL. The histidineinase domain of FixL and the serine/threonine kinase domains ofuman and mouse PASKIN are indicated by dotted lines. Identicalmino acids are shown in bold (indicated by asterisks). Colon, con-erved amino acid; period, semiconserved amino acid. Pairs of aminocids shown in italics indicate the position of introns.

761

ion of CHO cells (resulting in the cell line PKCHO) andy coupled in vitro transcription/translation (IVTT), re-pectively. Recombinant PASKIN protein was analyzedy immunoblotting as shown in Fig. 6. The detection of aand with a relative molecular weight of about 146 kDaincluding the RGSH6 tag) following protein synthesisoth in vitro and in vivo, but not in maternal CHO cells,emonstrated that PASKIN is not extensively posttrans-ationally modified. As expected from the primary se-uence data, PASKIN remained in the supernatant fol-owing centrifugation at 20,000g for 30 min, suggestinghat PASKIN is soluble (data not shown).

ISCUSSION

The PASKIN protein reported in this work is anttractive candidate for a mammalian cellular sensorrotein because first, it represents the first mamma-ian PAS protein containing a kinase domain that

ight link sensing to signaling; second, its PAS do-ain is more similar to the PAS domain of the oxygen

ensor FixL than to any other known mammalian PASrotein; and third, it is ubiquitously expressed. Apartrom this circumstantial evidence, we also tried to ob-ain experimental evidence for this hypothetical sensorunction. The PAS domain of FixL contains heme ashe oxygen-binding co-factor. Thus, the identificationf a prosthetic group within the PAS domain wouldave given a further hint as to what sensing function ofASKIN might be. However, while bacterially ex-ressed FixL showed the expected absorption spec-rum, we did not observe a chromophore in bacteriallyxpressed fragments of PASKIN containing the PASomain (data not shown). Whether this is an artifact ofhe bacterial expression system or whether PASKIN

FIG. 4. PASKIN expression in mouse tissues. Total RNA wassolated from several adult and embryonic mouse tissues and ana-yzed by Northern blotting. The blot was subsequently hybridized to

mouse PASKIN probe and a ribosomal protein L28 probe. Theositions of ribosomal 28S (4712 nt) and 18S (1869 nt) RNA weresed as size markers.

dcAPtpbe

dRbttmv

i


oes indeed not contain a chromophore (or even anyofactor at all) is currently a matter of investigation.lthough there are many different co-factors known forAS domains of archaeal and bacterial species (13–16),he only ligand known so far for a mammalian PASrotein is dioxin, or dioxin-related xenobiotics, whichind the aryl hydrocarbon receptor, another het-rodimerization partner of the HIF-1b subunit.

FIG. 5. Ubiquitous PASKIN expression in human tissues. Dot bndicated tissues. The blot was hybridized to a human PASKIN pro

762

As shown by a multiple alignment, the PAS sensoromains have been structurally conserved between thehizobium FixL and mammalian PASKIN proteins,ut the histidine kinase domain (weakly) matched tohe intervening domain rather than to the serine/hreonine kinase domain of PASKIN. Interestingly,ouse Paskin exon 10 to 12, which compose the inter-

ening domain, are part of one long open reading frame

analysis of normalized amounts of poly(A)1 RNA derived from thender high stringency conditions.

lotbe u

(rspiddiTsswc

tsPmSmtcmpdeabdAattmcm

1bPgr

anism of an environmental sensor requires de-pFprmpi

atndhtrbotPdweietP

A

DFCttcU(

R

tCi


810 amino acids) that is interrupted by single stopegions within small introns. Thus, it is tempting topeculate that this region is the remnant of an ancestorrotein where the histidine kinase-like domain becamenactivated during evolution and now spaces the PASomain from the newly gained serine/threonine kinaseomains in the mammalian PASKIN proteins, suggest-ng that PASKIN and FixL share a common ancestor.he idea of this “evolutionary exon erosion” is furtherupported by the fact, that the human gene contains ahortened exon 10 when compared to the mouse gene,hereas the positions of all other splice sites has been

onserved.Intriguingly, the analysis of the genomic structure of

he human PASKIN gene revealed the presence of ahort (1 kb) promoter region that is shared by thePP1R7 gene (21). Analogously, we also found theouse Ppp1r7 gene in the 59 region of the Paskin gene.o far, the location of the putative, very short firstouse exon could not be deduced by comparison with

he human sequence. However, regarding the overallonservation of the structure of these four genes, theouse Paskin exon 1 is likely to exist at a similar

osition as in the human PASKIN gene, although itsefinitive allocation awaits the identification of the 59nd of the cDNA. A comparison of the human PPP1R7nd mouse Ppp1r7 promoter regions has been reportedy Ceulemans and colleagues (21). This promoterrives ubiquitous expression of the PPP1R7 gene (21).ssuming a bicistronic promoter activity, this is ingreement with our finding that PASKIN is ubiqui-ously expressed. The PPP1R7 gene had been mappedo human chromosome 2q37.3 and the Ppp1r7 gene toouse chromosome 1 (21), hence representing also the

hromosomal localizations of the human PASKIN andouse Paskin genes.The PPP1R7 gene encodes Sds22, a regulator of typeserine/threonine protein phosphatases (PP1). It will

e interesting to elucidate whether coexpression of theASKIN and PPP1R7 genes from a single conservedenomic locus means that they are also functionallyelated. Phosphorylation as part of the signaling mech-

FIG. 6. PASKIN protein analysis. Immunoblotting of His-agged, recombinant human PASKIN produced by stably transfectedHO cells (CHO, maternal cell line; PKCHO, transfected cells) or by

n vitro transcription/translation (IVTT).

763

hosphorylation to turn off the signal. Intriguingly, theixL protein is known to be regulated by autophos-horylation. Because Sds22 mediates target proteinecognition of the PP1 protein phosphatase (23), itight be involved in dephosphorylation of autophos-

horylated PASKIN, thereby resetting the sensor to itsnactive state.

While this article was in preparation, S. L. McKnightnd co-workers reported the cloning of a human cDNAhat is identical to PASKIN (24). This protein has beenamed PASK for PAS kinase. However, in order toistinguish PASKIN from an unrelated kinase whichas been named PASK previously (25–27), we preferhe designation “PASKIN.” McKnight and colleagueseported that the PASKIN kinase activity is activatedy autophosphorylation and inhibited by the presencef the N-terminal PAS domain, whether in cis or inrans. Therefore, by assaying the kinase activity ofASKIN in an in vitro assay, it should be possible toefine the nature of the environmental parameters forhich PASKIN might serve as a sensor protein. Nev-rtheless, to unequivocally demonstrate the physiolog-cal relevance of PASKIN as a sensor protein, the gen-ration of knock-out mice will be required. Based onhe data reported here, we are currently targeting theaskin gene in mouse embryonic stem cells.

CKNOWLEDGMENTS

The authors thank M. Sporri, G. Tomio, and D. Zimmermann forNA sequencing work; F. Parpan for technical assistance; G.letschinger for artwork; K. F. Wagner for helpful discussions; and. Bauer and W. Jelkmann for support. This work was supported by

he Swiss National Science Foundation (31-56743.99) to M.G. andhe “Sondermassnahmen des Bundes zur Forderung des akademis-hen Nachwuchses,” the FSP Oncology (Project 22) of the Medicalniversity of Lubeck and the Deutsche Forschungsgemeinschaft

We2672/1-1) to R.H.W.

EFERENCES

1. Semenza, G. L. (2001) HIF-1 and mechanisms of hypoxia sens-ing. Curr. Opin. Cell Biol. 13, 167–171.

2. Wenger, R. H. (2000) Mammalian oxygen sensing, signaling andgene regulation. J. Exp. Biol. 203, 1253–1263.

3. Maxwell, P. H., Wiesener, M. S., Chang, G. W., Clifford, S. C.,Vaux, E. C., Cockman, M. E., Wykoff, C. C., Pugh, C. W., Maher,E. R., and Ratcliffe, P. J. (1999) The tumour suppressor proteinVHL targets hypoxia-inducible factors for oxygen-dependent pro-teolysis. Nature 399, 271–275.

4. Ivan, M., Kondo, K., Yang, H., Kim, W., Valiando, J., Ohh, M.,Salic, A., Asara, J. M., Lane, W. S., and Kaelin, W. G., Jr. (2001)HIFa targeted for VHL-mediated destruction by proline hy-droxylation: Implications for O2 sensing. Science 292, 464–468.

5. Jaakkola, P., Mole, D. R., Tian, Y. M., Wilson, M. I., Gielbert, J.,Gaskell, S. J., von Kriegsheim, A., Hebestreit, H. F., Mukherji,M., Schofield, C. J., Maxwell, P. H., Pugh, C. W., and Ratcliffe,P. J. (2001) Targeting of HIF-a to the von Hippel-Lindau ubiq-uitylation complex by O2-regulated prolyl hydroxylation. Science292, 468–472.

6. Conrad, P. W., Rust, R. T., Han, J., Millhorn, D. E., and Beitner-

1

1

1

1

1

1

1

superfamily: Sensors of environmental and developmental sig-

1

1

1

2

2

2

2

2

2

2

2


Johnson, D. (1999) Selective activation of p38a and p38g byhypoxia. Role in regulation of cyclin D1 by hypoxia in PC12 cells.J. Biol. Chem. 274, 23570–23576.

7. Minet, E., Arnould, T., Michel, G., Roland, I., Mottet, D., Raes,M., Remacle, J., and Michiels, C. (2000) ERK activation uponhypoxia: Involvement in HIF-1 activation. FEBS Lett. 468, 53–58.

8. Hur, E., Chang, K. Y., Lee, E., Lee, S. K., and Park, H. (2001)Mitogen-activated protein kinase kinase inhibitor PD98059blocks the trans-activation but not the stabilization or DNAbinding ability of hypoxia-inducible factor-1a. Mol. Pharmacol.59, 1216–1224.

9. Chen, E. Y., Mazure, N. M., Cooper, J. A., and Giaccia, A. J.(2001) Hypoxia activates a platelet-derived growth factorreceptor/phosphatidylinositol 3-kinase/Akt pathway that resultsin glycogen synthase kinase-3 inactivation. Cancer Res. 61,2429–2433.

0. Aragones, J., Jones, D. R., Martın, S., San Juan, M. A., Alfranca,A., Vidal, F., Vara, A., Merida, I., and Landazuri, M. O. (2001)Evidence for the involvement of diacylglycerol kinase in theactivation of hypoxia-inducible transcription factor 1 by low ox-ygen tension. J. Biol. Chem. 276, 10548–10555.

1. Gilles-Gonzalez, M. A., Ditta, G. S., and Helinski, D. R. (1991) Ahaemoprotein with kinase activity encoded by the oxygen sensorof Rhizobium meliloti. Nature 350, 170–172.

2. Gong, W., Hao, B., Mansy, S. S., Gonzalez, G., Gilles-Gonzalez,M. A., and Chan, M. K. (1998) Structure of a biological oxygensensor: A new mechanism for heme-driven signal transduction.Proc. Natl. Acad. Sci. USA 95, 15177–15182.

3. Ponting, C. P., and Aravind, L. (1997) PAS: A multifunctionaldomain family comes to light. Curr. Biol. 7, R674–R677.

4. Zhulin, I. B., and Taylor, B. L. (1997) PAS domain S-boxes inarchaea, bacteria and sensors for oxygen and redox. TrendsBiochem. Sci. 22, 331–333.

5. Taylor, B. L., and Zhulin, I. B. (1999) PAS domains: Internalsensors of oxygen, redox potential, and light. Microbiol. Mol.Biol. Rev. 63, 479–506.

6. Gu, Y. Z., Hogenesch, J. B., and Bradfield, C. A. (2000) The PAS

764

nals. Annu. Rev. Pharmacol. Toxicol. 40, 519–561.7. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular

Cloning: A Laboratory Manual, Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, NY.

8. Rolfs, A., Kvietikova, I., Gassmann, M., and Wenger, R. H.(1997) Oxygen-regulated transferrin expression is mediated byhypoxia-inducible factor-1. J. Biol. Chem. 272, 20055–20062.

9. Wanner, R. M., Spielmann, P., Stroka, D. M., Camenisch, G.,Camenisch, I., Scheid, A., Houck, D. R., Bauer, C., Gassmann,M., and Wenger, R. H. (2000) Epolones induce erythropoietinexpression via hypoxia-inducible factor-1a activation. Blood 96,1558–1565.

0. Kozak, M. (1991) An analysis of vertebrate mRNA sequences:Intimations of translational control. J. Cell Biol. 115, 887–903.

1. Ceulemans, H., Van Eynde, A., Perez-Callejon, E., Beullens, M.,Stalmans, W., and Bollen, M. (1999) Structure and splice prod-ucts of the human gene encoding sds22, a putative mitotic reg-ulator of protein phosphatase-1. Eur. J. Biochem. 262, 36–42.

2. Horowitz, D. S., and Krainer, A. R. (1994) Mechanisms for se-lecting 59 splice sites in mammalian pre-mRNA splicing. TrendsGenet. 10, 100–106.

3. Bollen, M. (2001) Combinatorial control of protein phosphatase-1.Trends Biochem. Sci. 26, 426–431.

4. Rutter, J., Michnoff, C. H., Harper, S. M., Gardner, K. H., andMcKnight, S. L. (2001) PAS kinase: An evolutionarily conservedPAS domain-regulated serine/threonine kinase. Proc. Natl.Acad. Sci. USA 98, 8991–8996.

5. Ushiro, H., Tsutsumi, T., Suzuki, K., Kayahara, T., and Nakano,K. (1998) Molecular cloning and characterization of a novelSte20-related protein kinase enriched in neurons and transport-ing epithelia. Arch. Biochem. Biophys. 355, 233–240.

6. Miao, N., Fung, B., Sanchez, R., Lydon, J., Barker, D., and Pang,K. (2000) Isolation and expression of PASK, a serine/threoninekinase, during rat embryonic development, with special empha-sis on the pancreas. J. Histochem. Cytochem. 48, 1391–1400.

7. Leiserson, W. M., Harkins, E. W., and Keshishian, H. (2000)Fray, a Drosophila serine/threonine kinase homologous to mam-malian PASK, is required for axonal ensheathment. Neuron 28,793–806.

Documents

Mammalian PASKIN, a PAS-Serine/Threonine Kinase Related to Bacterial Oxygen Sensors