CBAC-5986; No.of Pages11 ARTICLE IN PRESS · +Model ARTICLE IN PRESS CBAC-5986; No.of Pages11 2 G. Pampalakis et al. / Computational Biology and Chemistry xxx (2008) xxx–xxx In

C

A

bpsevpn(Ccm©

KK

1

(ni

6TK

ern

1d

ARTICLE IN PRESS+ModelBAC-5986; No. of Pages 11

Available online at www.sciencedirect.com

Computational Biology and Chemistry xxx (2008) xxx–xxx

Identification and analysis of mammalian KLK6 orthologue genes forprediction of physiological substrates�

Georgios Pampalakis a, Maria Arampatzidou a, Grigoris Amoutzias b,Sofia Kossida c, Georgia Sotiropoulou a,∗

a Department of Pharmacy, School of Health Sciences, University of Patras, Rion-Patras 26500, Greeceb Department of Ecology and Evolution, Biophore, University of Lausanne, 1015 Lausanne, Switzerland

c Bioinformatics Group, Foundation for Biomedical Research of the Academy of Athens, Athens 11527, Greece

Received 1 June 2007; received in revised form 25 October 2007; accepted 17 November 2007

bstract

Human kallikrein-related peptidase 6 (KLK6) is a novel serine protease that is aberrantly expressed in human cancers and represents a serumiomarker for the molecular diagnosis and monitoring of ovarian cancer. Here, we report the cloning and analysis of human kallikrein-relatedeptidase 6 gene (KLK6) orthologues in model organisms and farm animals. The corresponding full-length cDNAs were assembled from partialequences retrieved from EST and genomic databases. Alignment of inferred protein sequences indicated a high degree of conservation of thencoded enzyme. We found that, similarly to (HUMAN)KLK6, monkey, cattle, mouse and rat orthologue genes encode for multiple transcriptariants. This strengthens our previously published data showing that (HUMAN)KLK6 transcription is coordinately regulated by alternativeromoters. Analysis of the KLK6 upstream genomic region led to the identification of multiple conserved regulatory regions with motifs foruclear receptor transcription factors. Interestingly, we found that specific CpG dinucleotides in the proximal promoter, that were shown to regulateHUMAN)KLK6 gene expression via DNA methylation, are conserved in orthologue genes, indicating epigenetic regulation of the KLK6 gene.
onstruction of a protein–protein interaction network indicated that KLK6 likely acts on the TGF-b1 signal transduction pathway to regulateertain cytoskeletal proteins, such as vimentin and keratin 8, thus, KLK6 may control cell shape that, in turn, regulates cell migration andotility.2007 Elsevier Ltd. All rights reserved.
M/zy

caf

eywords: Human kallikrein-related peptidase 6 (tissue kallikrein 6/proteaseLK6 interactome

. Introduction

The gene encoding human kallikrein-related peptidase 6

Please cite this article in press as: Pampalakis, G., et al., Identification anphysiological substrates, Comput Biol Chem (2008), doi:10.1016/j.compb

tissue kallikrein 6 or protease M/zyme/neurosin) was origi-ally identified and cloned by differential display based on itsnactivated expression in metastatic breast cancer cells when

Abbreviations: KLK6, gene encoding human kallikrein-related peptidase; KLK6, human kallikrein-related peptidase 6; EST, expressed sequence tag;F, transcription factor; UTR, untranslated region; ORF, open reading frame;RT8, keratin 8; VIM, vimentin; CALR, calreticulin.

� This work was supported by a research grant sponsored by the Greek Gen-ral Secretariat of Research and Technology and by the European Union throughesearch project PENED2003 (03E�430) of operational program competitive-ess.∗ Corresponding author. Tel.: +30 2610 969939/40; fax: +30 2610 969940.

E-mail address: [email protected] (G. Sotiropoulou).

mctgiii(a(cIba

476-9271/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.oi:10.1016/j.compbiolchem.2007.11.002

me/neurosin); Orthologues; Splice variants; Conserved regulatory sequences;

ompared to corresponding primary tumor cells (Anisowicz etl., 1996). KLK6 is a member of the human tissue kallikreinamily, a cluster of 15 genes tandemly localized on human chro-osome 19q13.4 that are predicted to encode for trypsin- or

hymotrypsin-like serine proteases. Tissue kallikreins consti-ute the largest serine protease family within the entire humanenome (Borgono and Diamandis, 2004). The enzymatic activ-ty of KLK6 has been implicated in neurodegenerative andnflammatory diseases of the central nervous system, includ-ng multiple sclerosis, Alzheimer’s and Parkinson’s diseaseBorgono and Diamandis, 2004; Iwata et al., 2003; Bernett etl., 2002). The three-dimensional structures of mature KLK6active) and pro-KLK6 (zymogen form) were resolved by X-ray

d analysis of mammalian KLK6 orthologue genes for prediction ofiolchem.2007.11.002

rystallography (Bernett et al., 2002; Gomis-Ruth et al., 2002).t was shown that the enzymatic activity of KLK6 is regulatedy an autoactivation/autolytic inactivation mechanism (Bayes etl., 2004; Blaber et al., 2007).

dx.doi.org/10.1016/j.compbiolchem.2007.11.002

mailto:[email protected]


IN+ModelC

2 Biolo

tgoss(tPrdmrnwaiomteSterlopcKovc

bmcmiTtm

2

2

shwwbHiMn

2

themeM

2

we(Q

2

wr

2

fi(ife

3

3o

ooimulotgjjfiel

ARTICLEBAC-5986; No. of Pages 11

G. Pampalakis et al. / Computational

In this study, we applied biocomputational approacheso analyze partial DNA sequences retrieved from EST andenomic databases, in order to identify orthologue genesf (HUMAN)KLK6 and assemble their full-length cDNAequences (virtual transcripts). Recently, we suggested that tran-cription of KLK6 is controlled by three alternative promotersP1, P2 and P3) that all drive the synthesis of wild-type pro-ein and display tissue-specific activity (Pampalakis et al., 2004;ampalakis and Sotiropoulou, 2006). Interestingly, differentialegulation of P1 and P2 promoters has been associated with CNSemyelinating diseases (Christophi et al., 2004). Production ofultiple transcript variants is corroborated by the observation

eported here that orthologue genes are also regulated by alter-ative promoters that exploit intronic sequences. In addition,e have analyzed the KLK6 promoter sequences in mammals

nd identified multiple conserved consensus sequences for bind-ng of transcription factors. In addition, conserved positionsf CpG dinucleotides were identified within the KLK6 pro-oter sequences from different species. Previously, we showed

hat cytosine methylation at these sites underlies the aberrantxpression of KLK6 in human tumor cells (Pampalakis andotiropoulou, 2006). Consistently with the previous identifica-

ion of human KLK6 isoforms produced by splicing out codingxons (splice variants), mining of EST orthologue sequencesevealed novel splice variants in cattle. It is likely that accumu-ation of genome sequencing data will lead to the identificationf additional KLK6 isoforms in other species. In order to identifyotential substrates and biological functions of KLK6, we haveonstructed a protein–protein interaction network for KLK6 (theLK6 interactome) based on human and rat data. Expressionf KLK6 up-regulates keratin 8 (KRT8) and down-regulatesimentin (VIM). These changes are known to control theytoskeletal structure and reduce cellular motility and migration.

In conclusion, we showed that KLK6 activity is regulatedoth at the level of transcription and post-translationally. Theechanisms that control the expression of KLK6 are likely

onserved among species and involve the usage of multiple pro-oters and production of splice variants. The encoded protein

s highly conserved indicating a conserved biological function.he KLK6 interactome predicted that this secreted serine pro-

ease may also act intracellularly to regulate cell migration andotility.

. Materials and methods

.1. Nomenclature

The nomenclature that was recently proposed for the tis-ue kallikrein family by Lundwall et al. (2006) was followedere. Briefly, gene names are given in italics and protein namesith standard font. Animal symbols used in the present study,ere obtained from SWISS-PROT (http://www.expasy.ch/cgi-in/speclist). Under this system, the prefix (HUMAN) denotes


omo sapiens, (BOVIN) Bos taurus, (CANFA) Canis famil-aris, (PANTR) Pan troglodytes, (PIG) Sus scrofa, (MACMU)

acaca mulatta, (MOUSE) Mus musculus and (RAT) Rattusorvegicus.

Knr(

PRESSgy and Chemistry xxx (2008) xxx–xxx

.2. Nucleotide and protein databases

Expressed sequence tag sequences were retrieved fromhe National Center for Biotechnology Information (NCBI,ttp://www.ncbi.nlm.nih.gov) and ENSEMBL (http://www.nsembl.org). Search of EST databases was performed byegaBLAST, BLASTN and TBLASTN algorithms (Altschul

t al., 1997), while the KLK6 gene structure was retrieved fromapViewer (http://www.ncbi.nlm.nih.gov/mapview).

.3. Nucleotide and protein tools

Translation of nucleotide sequences and motif predictionere performed with tools available in Expasy (http://www.

xpasy.org). Sequence alignments were performed by Bl2seqBLAST two sequences) or DNAMAN (Lynnon Corporation,uebec, Canada) for multiple alignments.

.4. Identification of regulatory elements

Identification of regulatory elements upstream of KLK6 geneas conducted with the program ECR (evolutionary conserved

egions, http://www.dcode.org).

.5. Protein–protein interaction network

Protein–protein interactions of human KLK6 were identi-ed and displayed by the ingenuity pathway analysis (IPA)Ingenuity Systems, Inc., Redwood city, California, http://www.ngenuity.com/). The pathway was constructed with a directionrom KLK6 to the nodes, which are upstream of the differentiallyxpressed genes and regulate their expression.

. Results

.1. Identification of full-length cDNA sequences of KLK6rthologue genes

Fig. 1 shows a schematic representation of full-lengthrthologue cDNA sequences of KLK6. Accession numbersf representative EST clones used in this study are givenn Table 1. EST databases containing sequences from ani-

al genomes were searched with the KLK6 protein sequencesing the TBLASTN algorithm (Altschul et al., 1997). Full-ength cDNAs (virtual transcripts) were derived as contigsf partially overlapping EST sequences. The derived virtualranscripts encoding KLK6 orthologues were searched againstenomic animal databases, in order to identify intron–exonunctions and the corresponding gene structures. All spliceunctions satisfied the rule GT (donor)–AG (acceptor) thatlters out transcripts possibly resulting from spliceosomalrrors (Modrek and Lee, 2002). For all identified ortho-ogues, translational start sites lied within a conserved


ozak sequence (Kozak, 1987) and a polyadenylation sig-al was identified in the 3′ untranslated region. The mouse,at and dog orthologues were identified by Homologenehttp://www.ncbi.nlm.nih.gov/). Full-length transcripts were


http://www.expasy.ch/cgi-bin/speclist

http://www.ncbi.nlm.nih.gov/

http://www.ensembl.org/

http://www.ensembl.org/

http://www.ncbi.nlm.nih.gov/mapview

http://www.expasy.org/

http://www.expasy.org/

http://www.dcode.org/

http://www.ingenuity.com/



ARTICLE IN PRESS+ModelCBAC-5986; No. of Pages 11

G. Pampalakis et al. / Computational Biology and Chemistry xxx (2008) xxx–xxx 3

Fig. 1. Schematic representation of (HUMAN)KLK6 orthologue genomic sequences. The 5′ and 3′ untranslated ends were obtained from multiple contigs of partialE l transr s. Nui m alt

oietd

fcsactirpeii

sK

las

cfig

Aio

ST clones. Introns were identified by alignment of orthologue cDNA (virtuaegions whereas gray boxes indicate protein-coding regions. Lines depict intronncomplete genomic sequencing. P1, P2 and P3 represent transcripts derived fro

btained by EST database mining. KLK6 orthologues displaydentical gene structure and consist of five coding exons withqual or similar number of nucleotides. Sequences of all virtualranscripts identified here were included as a Supplementaryata.

P. troglodytes. Because no ESTs encoding for KLK6 wereound, probably due to the low number of available ESTlones, we searched the chimpanzee high throughput genomeequencing (HTGS) database at NCBI with the TBLASTNlgorithm using the human KLK6 amino acid sequence. Thelone CH251-355A20 (GenbankTM: AC130782) was identifiedhat represents a fragment of chromosome 19 encompass-ng the chimpanzee kallikrein locus. This sequence wasetrieved and used for further analysis. Aligning of chim-


anzee genomic sequence against (HUMAN)KLK6 sequencenabled prediction of the (PANTR)KLK6 gene structure shownn Fig. 1, that however, contains a number of gaps due toncomplete genomic sequencing. The inferred partial protein

fc

m

cripts) against genomic animal sequences. White boxes indicate untranslatedmbers represent the length of exons–introns in base pairs. An asterisk indicatesernative promoters.

tructure of chimpanzee KLK6 is 100% identical to the humanLK6.M. mulatta. For identification of the (HUMAN)KLK6 ortho-

ogue in Rhesus monkey, data were retrieved from ENSEMBLnd NCBI. (MACMU)KLK6 is localized on monkey chromo-ome 19.

B. taurus. In cow genome, kallikrein family of genes map onhromosome 18. Multiple overlapping EST clones were identi-ed that corresponded to the complete ORF of (BOVIN)KLK6ene including 5′ and 3′ untranslated sequences.

S. scrofa. In pig genome, a sequence (GenbankTM:C149292) was identified that contained the kallikrein locus,

ncluding the (PIG)KLK6 gene, as described for chimpanzeerthologues but its chromosomal localization is unknown. A


ull-length cDNA transcript for pig KLK6 was assembled as aontig of partially overlapping ESTs.

C. familiaris. In dog genome, kallikrein gene family wasapped on chromosome 1.



4 G. Pampalakis et al. / Computational Biology and Chemistry xxx (2008) xxx–xxx

Table 1GenbankTM accession numbers of representative EST clones encoding for KLK6 orthologues and tissue expression pattern

Animal GenbankTM accession number KLK6 gene region Expression

Pig BF441391 5′ end MixedBX671525 3′ end MixedBX671526 ORF Mixed

Cattle DV924321 P2 SkinDV871627 P1 (without un-translated Exon 2) HippocampusDV886953 P1 (with smaller Exon 2) ThalamusDV467443 Extended 3′ end MixedCO882996 ORF Brain

Chimpanzee AY410890 Sequence based on alignment

Dog CN497238 Extended 3′ end BrainDN394359 5′ end Brain

Rhesus monkey BQ807957 P1 HypothalamusENSG00000167755 (ENSEMBL) P2

Rat BF550821 P2 MixedAI145671 Extended 3′ end Mixed

Mouse BY727409 P1 Corpora quadrigeminaBB571443 P2 Skin

k7it

wHlitKK

nKomipnpfD

c(i(dsrr

sK

3p

hmasllitacfeaesmKrBrPP

BB180902

M. musculus and R. norvegicus. In mouse and rat, theallikrein gene family was previously mapped on chromosomeand 1, respectively (Yousef and Diamandis, 2003). Here, we

dentified full-length KLK6 cDNAs by assembling EST cloneshat extended further 5′ upstream or 3′ downstream.

Xenopus tropicalis. A transcript homologous to KLK6as identified in X. tropicalis (GenbankTM: BC087753).owever, it is unlikely that this represents the Xenopus ortho-

ogue of (HUMAN)KLK6, since a TBLASTN search of thenferred protein against human nucleotide sequences showedhat trypsin-like proteases, such as trypsin 1/PRSS1, KLK1,LK13, KLK11 and KLK8 display higher homology thanLK6.Drosophila melanogaster. Using Homologene (http://www.

cbi.nlm.nih.gov/) as another tool to identify the (HUMAN)LK6 orthologues proteins, we have identified a putative KLK6rthologue in D. melanogaster (GenbankTM: NM 166114). ThisRNA encodes for a protein with 33% identity and 52% pos-

tivity to human KLK6 that has a conserved trypsin-like serinerotease structure. Search of this sequence against the humanucleotide database, showed that the (HUMAN)KLK6 gene dis-layed the highest similarity. The reciprocal best BLAST hiturther supports the idea that this protein may represent therosophila KLK6 orthologue protein.Other mammalian species. (HUMAN)KLK6 orthologues

ould not be identified in cat (Felis catus) neither in sheepOvis aries) that is likely due to incomplete genome sequenc-ng and the low number of available ESTs. In elephantLoxodonta africanis) genomic DNA and ab initio RNA


atabase available by ENSEMBL, small fragments with highimilarity (>60%) to (HUMAN)KLK6 were identified, buteciprocal BLAST hit showed that these sequences rep-esented orthologues of other kallikrein-related peptidases,

rtdo

Extended 3′ end Hypothalamus

uch as (HUMAN)KLK12, (HUMAN)KLK11 and (HUMAN)LK8.

.2. KLK6 transcription is regulated by alternativeromoters and alternative splicing

Based on experimental data, we recently suggested thatuman KLK6 is coordinately regulated by three alternative pro-oters (P1, P2 and P3) (Pampalakis et al., 2004; Christophi et

l., 2004; Pampalakis and Sotiropoulou, 2006) as schematicallyhown in Fig. 1. The P1 transcript is derived from P1 promoterocated in the 5′ upstream region and contains two 5′ untrans-ated exons, the P2 transcript is derived from P2 promoter locatednside intron I and contains one 5′ untranslated exon, and the P3ranscript is derived from P3 promoter located inside intron IInd contains no untranslated exons. As shown in Fig. 1, ESTlones derived from either P1 or P2 promoters were detectedor most animals. More specifically, products of both promot-rs were predicted for Rhesus monkey, B. taurus, M. musculusnd R. norvegicus. P1- or P2-derived transcripts are predicted toncode for full-length KLK6 protein. In B. taurus, three distinctplice variants were detected that were derived from P1 pro-oter (Fig. 1). All splice variants encode for wild-type bovineLK6, since variation in sequence occurs in the 5′ untranslated

egion. For the other species, no splice variants were detected. In. taurus, of 10 EST clones that extended to the 5′ end, 3 clones

epresented the longest P1 transcript, 4 clones represented the1 transcript with a smaller second exon, 1 clone represented the1 product that lacks the second untranslated exon and 2 clones


epresented the P2 transcript, providing a rough estimation ofhe relative abundance of these transcripts. Taken together, ourata suggest that conserved mechanisms regulate the expressionf KLK6 in mammalian species.






Fig. 2. KLK6 orthologue protein sequences. Alignment of KLK6 orthologue protein sequences. The catalytic triad amino acid residues are separated by verticalw sites ge sit

3

otbciKta

aRiowss

hite lines and are marked by an asterisk. A circle indicates an N-glycosylationites are not conserved in rodents. The arrow indicates the signal peptide cleava

.3. Comparison of KLK6 orthologue proteins

Based on the protein alignment shown in Fig. 2, all KLK6rthologue proteins have six fully conserved cysteine residueshat are probably engaged in the formation of conserved disulfideonds. The amino acid residues of the catalytic triad are also fullyonserved (His, coding exon 2; Asp, coding exon 3; Ser, cod-


ng exon 5). Recently, we showed that the enzymatic activity ofLK6 is self-regulated by an autocatalytic mechanism of activa-

ion and subsequent inactivation as a result of autolytic cleavaget internal R76 (Bayes et al., 2004) that is located in an exposed

taei

and inverted triangle an autoinactivation site (R76 in human KLK6). Autolysise, which was predicted by SignalP 3.0 (Bendtsen et al., 2004).

utolysis loop (Bernett et al., 2002; Gomis-Ruth et al., 2002).76 is conserved in Rhesus monkey, dog, pig and bovine but not

n mouse and rat. Characterization of the enzymatic propertiesf rat KLK6 revealed two different autoinactivation sites usedith similar preference (Blaber et al., 2002). An N-glycosylation

ite at N132 of human KLK6 was experimentally confirmed byite-directed mutagenesis (Gomis-Ruth et al., 2002), however,


he sugar moiety does not significantly affect the enzymaticctivity due to its distal location from the active-site (Bernettt al., 2002; Gomis-Ruth et al., 2002). This site is conservedn most species examined here, except for rodents. Indeed, rat


ARTICLE IN+ModelCBAC-5986; No. of Pages 11

6 G. Pampalakis et al. / Computational Biolo

Table 2Identities of KLK6 orthologues among animal species

Human Monkey Pig Cattle Dog Mouse Rat

Human 100Monkey 97 100Pig 84 86 100Cattle 80 80 86 100Dog 86 87 91 86 100Mouse 68 69 70 69 69 100R

%

K32lsi

3

ssruwr

dbt(klitoertHtEm(LtHtg(a

FgmrvaopKSat

at 66 67 68 68 67 88 100

identities were obtained by pair wise alignment using blast2sequences.

LK6 is a deglycosylated protein (Blaber et al., 2002). SignalP.0 (http://www.cbs.dtu.dk/services/SignalP) (Bendtsen et al.,004) identified a signal peptide cleavage site with high simi-arity in all proteins that are predicted to be secreted. Table 2hows the identities between KLK6 orthologue proteins in var-ous animals.

.4. Promoter analysis

Analysis of KLK6 upstream sequences predicted con-erved transcription factor binding sites and CpG dinucleotideequences, indicating common mechanisms of transcriptional


egulation. The KLK6-KLK7 intergenic region that is located 5′pstream of KLK6 and includes the putative promoter sequenceas analyzed in several species in order to identify conserved

egulatory regions. Multiple conserved regulatory regions were

Hao2

ig. 3. Mechanisms of regulation of KLK6 gene expression are conserved among anenes and visualization of the sequence between KLK7 and KLK6 genes using ECRouse, rat and dog orthologue genes. Compared genomes were plotted as horizontal

epresents synteny. Each layer contains a pip-type plot that consists of multiple shorertical height of each line describes the nucleotide identity underlying the alignmend UTRs in yellow. Arrow lines indicate 5′ → 3′ orientation of gene sequences. Thf ECRs and is used to flag underlying regions of alignment. A conserved alignmenink if it overlaps with an intron and red if it overlaps with an intergenic region. GreLK6 P1 promoter region. Asterisks indicate positions of cytosines experimentally footiropoulou, 2006). The transcriptional start site is indicated with an arrow for (HUnd pig, sites for initiation of transcription are located further downstream (not showno the web version of the article.)


etected, as shown in Fig. 3A. Putative transcription factorinding sites (TFBS) were identified using the ECR (evolu-ionary conserved regions) browser at http://www.dcode.orgOvcharenko et al., 2004). We aligned human, Rhesus mon-ey, mouse, rat and dog sequences, using as a cutoff detectionevel of ECRs a minimum length of 100 bp and a minimumdentity of 70%. The identified ECRs were scanned for puta-ive TF consensus sequences. The upstream non-coding regionf the KLK6 gene is conserved within the mammalian lin-age and contains several consensus sequences for nucleareceptors or basic helix-loop-helix family of transcription fac-ors. More specifically, binding sites for HEB, E2A, PU.1,NF4 (hepatocyte nuclear factor 4) and LXR (liver X recep-

or) were conserved in the proximal KLK6 promoter. HEB,2A and PU.1 are factors involved in lymphocyte develop-ent (Murre, 2005; Hagman and Lukin, 2006), while HNF4

Lemaigre and Zaret, 2004) is mainly a liver-specific factor.XR is involved in regulation of genes involved in choles-

erol metabolism and lipid biosynthesis (Makishima, 2005).EB functions synergistically with myogenin, also predicted

o bind to the KLK6 promoter, in order to activate the myo-enin expression, thus myogenin regulates its own expressionParker et al., 2006). Therefore, HEB and myogenin may inter-ct for binding to the KLK6 proximal promoter. In addition,


EB forms heterodimers with E2A that bind to E-boxes toctivate gene expression, an interaction involved in regulationf T-cell development (Barndt et al., 2000; Takeuchi et al.,001).

imal species. (A) Predicted regulatory motifs in mammalian KLK6 orthologueBrowser. Conservation profiles between human KLK6 and Rhesus monkey,

layers of conservation diagrams. The light blue bar at the bottom of each layert horizontal black lines. Each line represents an ungapped alignment, and thent. Gene exons are depicted in blue and yellow, protein-coding exons in bluee dark red bar on top of every layer provides an overview of the distributiont is blue if it overlaps with a coding exon, yellow if it overlaps with an UTR,en bars at the bottom indicate repetitive elements. (B) Sequence alignment ofund to be methylated in human cells that do not express KLK6 (Pampalakis andMAN)KLK6 and a gray circle for chimpanzee, dog, cattle (bovine). In monkey). (For interpretation of the references to color in the text, the reader is referred


http://www.cbs.dtu.dk/services/SignalP

http://www.dcode.org/



ontin

uccnsoKcshcsmntnfo

iatc

3

anpcoIb

Fig. 3 (C

We have previously shown that DNA hypermethylationnderlies the observed silencing of KLK6 expression in humanancer cells (Pampalakis and Sotiropoulou, 2006). More specifi-ally, we found that the extent of cytosine methylation in a smallon-typical-CpG island region (−72, +15) that spans the tran-criptional start site of P1 promoter correlates well with levelsf expression of the (HUMAN)KLK6 gene. Alignment of P1LK6 sequences of human, chimpanzee, monkey, pig, dog andattle indicated that the genomic regions that span the P1 tran-criptional start site are conserved, as shown in Fig. 3B. Theighest degree of conservation is observed between the human,himpanzee and monkey sequences. Also, conserved are theeven CpG dinucleotides (marked by asterisks) that were foundethylated in human KLK6 non-expressing cells. In addition,

ovel CpG sites are present in KLK6 upstream sequences in


he dog, cattle and pig orthologue genes indicating that epige-etic mechanisms may be an alternative conserved mechanismor regulation of KLK6 gene expression in evolutionary higherrganisms. High degree of similarity of nucleotides surround-

waee

ued ).

ng the transcriptional start site of (HUMAN)KLK6 (depicted byrrow in Fig. 3B) indicates that all other species probably exploithis sequence as a transcriptional initiation signal. However, ESTlones that extend to this region were not found.

.5. The KLK6 interactome

Fig. 4 depicts a simple flowchart for the IPA function, as wells a protein–protein interaction network that involves KLK6,amed the KLK6 interactome, as designed by IPA. The inputroteins (KLK6, vimentin, calreticulin, and keratin 8) used toonstruct the KLK6 interactome were retrieved from previ-us proteomic analysis in our laboratory (unpublished data).PA transforms a list of genes into a set of relevant networksased on extensive records maintained in the ingenuity path-


ays knowledge base (IPKB). This tool can help biologistsnd bioinformaticians to understand how their genes of inter-st are biologically related either directly (by interacting withach other) or indirectly (by interacting with neighboring genes).



8 G. Pampalakis et al. / Computational Biology and Chemistry xxx (2008) xxx–xxx

F alysiI prote

Mobr

adafbamsfbn

dtoIhaeoIo

o(

tsfimstdcpuss2cvb1dte

ig. 4. KLK6 interactome. (A) A simple flowchart for the ingenuity pathway anPA. Although KLK6 is an extracellular protease, it can control the cytoskeletal

oreover, the users can obtain a great amount of informationn a single screen such as the kind of the biological relationsetween their genes of interest, other genes that are closelyelated with the user’s ones, etc.

In order to succeed that, the developers of IPA have designedprocedure that creates quickly relevant networks. The proce-ure consists of six steps and the final result is a graph that iss highly connected as possible. It uses simple methods derivedrom graph algorithmic techniques such as triangle connectivity,est “neighbor”, etc. that have low computational cost. There isnetwork size limitation (by default 35) so the results can beore useful for the researchers. In many cases, the selection of

ome genes that are included in the final network has been per-ormed in a random fashion. Perhaps, additional criteria shoulde taken into consideration in order to produce better qualityetworks.

The data used in these functions come from the IPKBatabase, which contains a large volume of information abouthe genes. So, the success of the IPA tool is mainly basedn the quality of the IPKB data and on the structure of thePKB. In order to increase the quality, the IPA developersave included data manually obtained from literature researchnd only a small percentage of them (<10%) is automatically


xtracted and modeled. Furthermore, it uses a well-organizedntology that consists of a hierarchical data structure and givesPA the opportunity to be fast. Details about the developmentf the algorithm can be obtained from Calvano et al. (2005)

mKli

s (IPA) function. (B) The KLK6 interactome was constructed with the usage ofins VIM and KRT8.

r can be downloaded from the help menu of the IPA programhttp://www.ingenuity.com/).

The intermediate nodes and their role(s) in signal transduc-ion were uncovered, as following: rat KLK6 (myelencephalon-pecific protease or MSP) protein rapidly degrades rat plasmabronectin (FN) to a polypeptide with an apparent molecularass of 200 kDa, which is subsequently degraded to numerous

maller fragments (Blaber et al., 2002). FN is a glycopro-ein present in a soluble dimeric form in plasma, and in aimeric or multimeric form at the cell surface and in extra-ellular matrix. FN is involved in cell adhesion and migrationrocesses including embryogenesis, wound healing, blood coag-lation, host defence and metastasis. FN treatment of vascularmooth muscle cells significantly increases the mRNA expres-ion of transforming growth factor-beta 1 (TGF-b1) (Hu et al.,000). TGF-b1 is an extracellular growth factor, involved inell transformation. TGF-b1 induces the expression of FN inarious cell types (Yoshida et al., 1992; Pricci et al., 1996) itinds (Mooradian et al., 1989) and activates it (Grunt et al.,993). TGF-b1 induces TGF-b1 mRNA levels in a time- andose-dependent manner (Baldwin and Korc, 1993). TGF-b1reatment of pancreatic cancer cells results in reduced cytok-ratin 8 (KRT8) expression but increases the expression of the


esenchymal marker vimentin (VIM) (Ellenrieder et al., 2001).RT8, a simple epithelial keratin is a component of the intracel-

ular cytoskeleton in the cells of the single-layered sheet tissuesnside the body. It is a member of the intermediate filament



IN+ModelC

Biolog

fwTmeipTlClrr2

4

olisfaciteiwirtfo(

lelteicm(a(aoemAs(rc

pK

cpihmstnem2(ie

iiKt((doasrirDa(

Ktbecscmccddwas

A



amily of proteins. VIM is an intermediate filament protein,hich moves bi-directionally along microtubules in the cell.he level of the transcriptional regulator factor E2F1, is down-odulated in TGF-b1/vitamin D3-treated HL-60 cells (Peiretti

t al., 2001; DeGregori et al., 1995). E2F1 plays a crucial rolen the control of cell cycle and action of tumor suppressorroteins. E2F1 can also modulate the transcriptional level ofGF-b1 (Thatikunta et al., 1997). The gene encoding for car-

eticulin (CALR) is up-regulated by E2F1 (Hlaing et al., 2004).ALR, a Ca2+ binding protein, can act as an important modu-

ator of the regulation of gene transcription by nuclear hormoneeceptors. Finally, it is interesting to note that it has been foundecently that human KLK6 cleaves FN as well (Ghosh et al.,004).

. Discussion

Mapping and sequencing of genomes from a large numberf species enabled us to identify and clone in silico ortho-ogue genes of the (HUMAN)KLK6 gene. We have identifiedn total 13 mammalian KLK6 orthologues including novel tran-cript and splice variants and the putative KLK6 orthologuerom D. melanogaster. The longest KLK6 cDNA transcript wasssembled for every orthologue. Similarly to humans, KLK6o-localized with all other kallikreins in a chromosomal locusn all mammalian genomes examined. All EST clones used inhis analysis were identified in tissues known in humans toxpress (HUMAN)KLK6. Further, all genes contained five cod-ng exons and encoded for putative secreted serine proteasesith high identity to human KLK6 (>66% in rodents and >80%

n other mammals). The codons that encode for the amino acidesidues of the catalytic triad were found in conserved posi-ions. Furthermore, all intron phases were fully conserved asor (HUMAN)KLK6 (I–II–I–0). The above data confirmed ourriginal hypothesis that the identified genes are orthologues ofHUMAN)KLK6.

Based on the high similarity observed between KLK6 ortho-ogues, it is very likely that mechanisms regulating theirxpression are well conserved among species. Indeed, simi-arly to humans, we have demonstrated presence of multipleranscript variants in monkey, cattle, mouse and rat that allncode for the same full-length protein, since they differ onlyn the 5′ UTR. Splice variants resulting from splicing out ofoding exons were found in cattle. Specific splice variantsay have differential expression in disease, as observed for

HUMAN)KLK13 in testicular cancers (Chang et al., 2001)nd (HUMAN)KLK5 and (HUMAN)KLK7 in ovarian cancerDong et al., 2003). In addition, differential regulation of P1nd P2 (HUMAN)KLK6 promoters in inflammatory conditionsf the central nervous system was demonstrated (Christophit al., 2004). Alignment of genomic sequences from variousammals showed the presence of conserved regulatory motifs.lso, differential DNA methylation likely underlies the tissue-


pecific expression of KLK6, as it was demonstrated for maspinFutscher et al., 2002). Here, we have found that positions ofegulatory CpGs (Pampalakis and Sotiropoulou, 2006) are wellonserved in the other animals, suggesting that methylation may

f2

PRESSy and Chemistry xxx (2008) xxx–xxx 9

lay a significant role in determining the expression profile ofLK6.

The high similarity in the structure of KLK6 orthologues indi-ates a conserved function at the protein level. It was reportedreviously that MSP, the rat orthologue of human KLK6, ismplicated in multiple sclerosis (Blaber et al., 2004). Recently,uman KLK6 was implicated in multiple sclerosis and inflam-ations of the central nervous system (Iwata et al., 2003). In

pinal cord injury, MSP and KLK6 are up-regulated at sites ofrauma (Scarisbrick et al., 2006) and both can degrade compo-ents of the extracellular matrix (Blaber et al., 2002; Magklarat al., 2003). MSP is a trypsin-like enzyme, able to cleaveyelin and extracellular matrix components (Blaber et al.,

002). Human KLK6 was similarly shown to degrade myelinBernett et al., 2002), while inhibition of its enzymatic activitymproved the course of experimental multiple sclerosis (Blabert al., 2004).

To predict potential function(s) of KLK6, a protein–proteinnteraction network of KLK6 was constructed. Although KLK6s a secreted protein, here, we demonstrate for the first time thatLK6 is able to regulate the levels of two intracellular cytoskele-

al proteins, VIM and KRT8 via the TGF-b1 signaling pathwayFig. 4B). KLK6 rapidly and specifically degrades both lamininLN) and fibronectin (FN). Decreased levels of FN result inecrease of TGF-b1. This results in an even further decreasef the expression of FN and, subsequently, TGF-b1 throughloop between the two genes. As TGF-b1 plays a suppres-

ive role in the expression of KRT8 and E2F1 and an inducingole in the expression of VIM, decreased TGF-b1 allows for anncrease of KRT8 and E2F1 expression and a decrease of VIM,espectively. Increased E2F1 results in up-regulation of CALR.ifferential proteomic analysis showed that KLK6 expression

ffects the concentration of these protein substrates as predictedour unpublished data).

According to the KLK6 protein–protein interaction network,LK6 reduces the expression of FN, which in turn reduces

he expression of TGF-b1. Thus, levels of KLK6 and TGF-1 are inversely correlated. Further, TGF-b1 up-regulates thexpression of VIM, which is found in aggressive forms of breastancer in contrast to KLK6 that is down-regulated in the aggres-ive forms. Therefore, we expect that the treatment of breastancer cells with TGF-b1 will inhibit the expression of KLK6RNA. In order to test whether the KLK6 interactome that we

onstructed here operates in vivo, we treated the T47D breastancer cells with TGF-b1. We found that TGF-b1 caused a time-ependent down-regulation of KLK6 mRNA (our unpublishedata). In conclusion, this study for the first time predicts a path-ay for KLK6 that should be tested in cellular and animal models

nd may potentially be exploited for the design of therapeutictrategies.

ppendix A. Supplementary data


Supplementary data associated with this article can beound, in the online version, at doi:10.1016/j.compbiolchem.007.11.002.


http://dx.doi.org/10.1016/j.compbiolchem.2007.11.002

http://dx.doi.org/10.1016/j.compbiolchem.2007.11.002

IN+ModelC

1 Biolo

R

A

A

B

B

B

B

B

B

B

B

B

C

C

C

D

D

E

F

G

G

G

H

H

H

I

K

L

L

M

M

M

M

M

O

P

P

P

P

P


0 G. Pampalakis et al. / Computational

eferences

ltschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W.,Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: a new generation ofprotein database search programs. Nucleic Acids Res. 25, 3389–3402.

nisowicz, A., Sotiropoulou, G., Stenman, G., Mok, S.C., Sager, R., 1996. Anovel protease homolog differentially expressed in breast and ovarian cancer.Mol. Med. 2, 624–636.

aldwin, R.L., Korc, M., 1993. Growth inhibition of human pancreatic car-cinoma cells by transforming growth factor beta-1. Growth Factors 8,23–34.

arndt, R.J., Dai, M., Zhuang, Y., 2000. Functions of E2A-HEB heterodimersin T-cell development revealed by a dominant negative mutation of HEB.Mol. Cell Biol. 20, 6677–6685.

ayes, A., Tsetsenis, T., Ventura, S., Vendrell, J., Aviles, F.X., Sotiropoulou,G., 2004. Human kallikrein 6 activity is regulated via an autoproteolyticmechanism of activation/inactivation. Biol. Chem. 385, 517–524.

endtsen, J.D., Nielsen, H., von Heijne, G., Brunak, S., 2004. Improved predic-tion of signal peptides: SignalP 3.0. J. Mol. Biol. 340, 783–795.

ernett, M.J., Blaber, S.I., Scarisbrick, I.A., Dhanarajan, P., Thompson, S.M.,Blaber, M., 2002. Crystal structure and biochemical characterization ofhuman kallikrein 6 reveals that a trypsin-like kallikrein is expressed in thecentral nervous system. J. Biol. Chem. 277, 24562–24570.

laber, S.I., Scarisbrick, I.A., Bernett, M.J., Dhanarajan, P., Seavy, M.A., Jin,Y., Schwartz, M.A., Rodriguez, M., Blaber, M., 2002. Enzymatic propertiesof rat myelencephalon-specific protease. Biochemistry 41, 1165–1173.

laber, S.I., Ciric, B., Christophi, G.P., Bernett, M.J., Blaber, M., Rodriguez,M., Scarisbrick, I.A., 2004. Targeting kallikrein 6 proteolysis attenuates CNSinflammatory disease. FASEB J. 18, 920–922.

laber, S.I., Yoon, H., Scarisbrick, I.A., Juliano, M.A., Blaber, M., 2007. Theautolytic regulation of human kallikrein-related peptidase 6. Biochemistry46, 5209–5217.

orgono, C.A., Diamandis, E.P., 2004. The emerging roles of human tissuekallikrein gene family in cancer. Nat. Rev. Cancer 4, 876–890.

alvano, S.E., Xiao, W., Richards, D.R., Felciano, R.M., Baker, H.V., Cho, R.J.,Chen, R.O., Brownstein, B.H., Cobb, J.P., Tschoeke, S.K., et al., 2005. Anetwork-based analysis on systemic inflammation in humans. Nature 473,1032–1037.

hang, A., Jung, K., Rajpert-DeMeyts, E., Yousef, G.M., Diamandis, E.P., 2001.Identification and molecular characterization of five novel kallikrein gene13 (KLK13; KLK-L4) splice variants: differential expression in the humantestis and testicular cancer. Anticancer Res. 21, 2152–3147.

hristophi, G.P., Isackson, P.J., Blaber, S., Blaber, M., Rodriguez, M., Scaris-brick, I.A., 2004. Distinct promoters regulate tissue-specific and differentialexpression of kallikrein 6 in CNS demyelinating disease. J. Neurochem. 91,1439–1449.

eGregori, J., Kowalik, T., Nevins, J.R., 1995. Cellular targets for activation bythe E2F1 transcription factor include DNA synthesis- and G1/S-regulatorygenes. Mol. Cell Biol. 15, 4215–4224.

ong, Y., Kaushal, A., Brattsand, M., Nicklin, J., Clements, J.A., 2003. Dif-ferential splicing of KLK5 and KLK7 in epithelial ovarian cancer producesnovel variants with potential as cancer biomarkers. Clin. Cancer Res. 9,1710–1720.

llenrieder, V., Hendler, S.F., Boeck, W., Seufferlein, T., Menke, A., Ruhland,C., Adler, G., Gress, T.M., 2001. Transforming growth factor beta1 treatmentleads to an epithelial–mesenchymal transdifferentiation of pancreatic cancercells requiring extracellular signal-regulated kinase 2 activation. Cancer Res.61, 4222–4228.

utscher, B.W., Oshiro, M.M., Wozniak, R.J., Holtan, N., Hanigan, C.L., Duan,H., Domann, F.E., 2002. Role for DNA methylation in the control of celltype specific maspin expression. Nat. Genet. 31, 175–179.

hosh, M.C., Grass, L., Soosaipilli, A., Sotiropoulou, G., Diamandis, E.P., 2004.Human kallikrein 6 degrades extracellular matrix proteins and may enhance


the metastatic potential of tumour cells. Tumour Biol. 25, 193–199.omis-Ruth, F.X., Bayes, A., Sotiropoulou, G., Pampalakis, G., Tsetsenis,

T., Villegas, V., Aviles, F.X., Coll, M., 2002. The structure of humanprokallikrein 6 reveals a novel activation mechanism for the kallikrein family.J. Biol. Chem. 277, 27273–27281.

S


runt, T.W., Oeller, H., Somay, C., Dittrich, E., Fazeny, B., Mannhalter, C.,Dittrich, C., 1993. Modulation of the immunophenotype of ovarian cancercells by N,N-dimethylformamide and transforming growth factor-beta 1. J.Cell. Physiol. 156, 358–366.

agman, J., Lukin, K., 2006. Transcription factors drive B cell development.Curr. Opin. Immunol. 18, 127–134.

laing, M., Spitz, P., Padmanabhan, K., Cabezas, B., Barker, C.S., Bern-stein, H.S., 2004. E2F-1 regulates the expression of a subset of targetgenes during skeletal myoblast hypertrophy. J. Biol. Chem. 279, 43625–43633.

u, W.Y., Fukuda, N., Satoh, C., Jian, T., Kubo, A., Nakayama, M., Kishioka,H., Kanmatsuse, K., 2000. Phenotypic modulation by fibronectin enhancesthe angiotensin II-generating system in cultured vascular smooth musclecells. Arterioscler. Thromb. Vasc. Biol. 20, 1500–1505.

wata, A., Maruyama, M., Akagi, T., Hashikawa, T., Kanazawa, I., Tsuji, S.,Nukina, N., 2003. Alpha-synuclein degradation by serine protease neurosin:implication for pathogenesis of synucleinopathies. Hum. Mol. Genet. 12,2625–2635.

ozak, M., 1987. An analysis of 5′-noncoding sequences from 699 vertebratemessenger RNAs. Nucleic Acids Res. 15, 8125–8148.

emaigre, F., Zaret, K.S., 2004. Liver development update: new embryo mod-els, cell lineage control, and morphogenesis. Curr. Opin. Genet. Dev. 14,582–590.

undwall, A., Band, V., Blaber, M., Clements, J.A., Courty, Y., Diamandis, E.P.,Fritz, H., Lilja, H., Malm, J., Maltais, L.J., Olsson, A.Y., Petraki, C., Scorilas,A., Sotiropoulou, G., Stenman, U.H., Stephan, C., Talieri, M., Yousef, G.M.,2006. A comprehensive nomenclature for serine proteases with homologyto tissue kallikreins. Biol. Chem. 387, 637–641.

agklara, A., Mellati, A.A., Wasney, G.A., Little, S.P., Sotiropoulou, G., Becker,G.W., Diamandis, E.P., 2003. Characterization of the enzymatic activity ofhuman kallikrein 6: autoactivation, substrate specificity, and regulation byinhibitors. Biochem. Biophys. Res. Commun. 8, 948–955.

akishima, M., 2005. Nuclear receptors as targets for drug development: reg-ulation of cholesterol and bile acid metabolism by nuclear receptors. J.Pharmacol. Sci. 97, 177–183.

odrek, B., Lee, C., 2002. A genomic view of alternative splicing. Nat. Genet.30, 13–19.

ooradian, D.L., Lucas, R.C., Weatherbee, J.A., Furcht, L.T., 1989. Transform-ing growth factor-beta 1 binds to immobilized fibronectin. J. Cell. Biochem.41, 189–200.

urre, C., 2005. Helix-loop-helix proteins and lymphocyte development. Nat.Immunol. 6, 1079–1086.

vcharenko, I., Nobrega, M.A., Loots, G.G., Stubbs, L., 2004. ECR Browser:a tool for visualizing and accessing data from comparisons of multiplevertebrate genomes. Nucleic Acids Res. 32, W280–W286.

ampalakis, G., Sotiropoulou, G., 2006. Multiple mechanisms underlie the aber-rant expression of the human kallikrein 6 gene in breast cancer. Biol. Chem.387, 773–782.

ampalakis, G., Kurlender, L., Diamandis, E.P., Sotiropoulou, G., 2004. Cloningand characterization of novel isoforms of the human kallikrein 6 gene.Biochem. Biophys. Res. Commun. 320, 54–61.

arker, M.H., Perry, R.L., Fauteux, M.C., Berkes, C.A., Rudnicki, M.A., 2006.MyoD synergizes with the E-protein HEB beta to induce myogenic differ-entiation. Mol. Cell Biol. 26, 5771–5783.

eiretti, F., Lopez, S., Deprez-Beauclair, P., Bonardo, B., Juhan-Vague, I., Nal-bone, G., 2001. Inhibition of p70(S6) kinase during transforming growthfactor-beta 1/vitamin D(3)-induced monocyte differentiation of HL-60cells allows tumor necrosis factor-alpha to stimulate plasminogen activatorinhibitor-1 synthesis. J. Biol. Chem. 276, 32214–32219.

ricci, F., Pugliese, G., Romano, G., Romeo, G., Locuratolo, N., Pugliese, F.,Mene, P., Galli, G., Casini, A., Rotella, C.M., Di Mario, U., 1996. Insulin-likegrowth factors I and II stimulate extracellular matrix production in humanglomerular mesangial cells. Comparison with transforming growth factor-


beta. Endocrinology 137, 879–885.carisbrick, I.A., Sabharwal, P., Cruz, H., Larsen, N., Vandell, A.G., Blaber,

S.I., Ameenuddin, S., Papke, L.M., Fehlings, M.G., Reeves, R.K., Blaber,M., Windebank, A.J., Rodriguez, M., 2006. Dynamic role of kallikrein 6 intraumatic spinal cord injury. Eur. J. Neurosci. 24, 1457–1467.


IN+ModelC

Biolog

T

T

Y



akeuchi, A., Yamasaki, S., Takase, K., Nakatsu, F., Arase, H., Onodera,M., Saito, T., 2001. E2A and HEB activate the pre-TCR� pro-


moter during immature T cell development. J. Immunol. 167, 2157–2163.

hatikunta, P., Raj, G.V., Kundu, M., Khalili, K., Amini, S., 1997. The tran-scription factor E2F-1 modulates TGF-beta1 RNA expression in glial cells.Oncogene 14, 2959–2969.

Y

PRESSy and Chemistry xxx (2008) xxx–xxx 11

oshida, M., Romberger, D.J., Illig, M.G., Takizawa, H., Sacco, O., Spurzem,J.R., Sisson, J.H., Rennard, S.I., Beckmann, J.D., 1992. Transforming growth


factor-beta stimulates the expression of desmosomal proteins in bronchialepithelial cells. Am. J. Respir. Cell Mol. Biol. 6, 439–445.

ousef, G.M., Diamandis, E.P., 2003. An overview of the kallikrein gene fami-lies in humans and other species: emerging candidate tumor markers. Clin.Biochem. 36, 443–452.


Documents

CBAC-5986; No.of Pages11 ARTICLE IN PRESS · +Model ARTICLE IN PRESS CBAC-5986; No.of Pages11 2 G. Pampalakis et al. / Computational Biology and Chemistry xxx (2008) xxx–xxx In