Oncogene (1999) 18, 6667 ± 6676 ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $15.00 http://www.stockton-press.co.uk/onc Growth factor ± regulated expression of enzymes involved in nucleotide biosynthesis: a novel mechanism of growth factor action

Marcus G Gassmann1, Andreas Stanzel1 and Sabine Werner*,1,2

1Max-Planck-Institute of Biochemistry, Am Klopferspitz 18a, D-82152 Martinsried, Germany; 2Institute of Cell Biology, ETH HoÈnggerberg, CH-8093 ZuÈrich, Switzerland

Keratinocyte growth factor (KGF) is a potent and ®broblast-derived mitogen which stimulates the prolif- speci®c mitogen for epithelial cells, including the eration of mouse keratinocytes (Rubin et al., 1989). In keratinocytes of the skin. We investigated the mechan- contrast to most FGFs which are broad-spectrum isms of action of KGF by searching for genes which are mitogens, responsiveness to KGF is a speci®c feature regulated by this growth factor in cultured human of epithelial cells. Thus KGF has been shown to keratinocytes. Using the di€erential display RT ± PCR stimulate proliferation of various types of epithelial technology we identi®ed the gene encoding adenylosucci- cells in vitro and in vivo (for review see Rubin et al., 1995). nate lyase [EC 4.3.2.2] as a novel KGF-regulated gene. In addition, it a€ects migration and di€erentiation Adenylosuccinate lyase plays an important role in purine processes of these cells and it also acts as a protective de novo synthesis. To gain further insight into the factor (reviewed by Werner, 1998). Therefore, KGF has potential role of nucleotide biosynthesis in the mitogenic the properties expected of a wound cytokine. Indeed, our e€ect of KGF, we cloned cDNA fragments of the key laboratory and others demonstrated a strong upregula- regulatory enzymes involved in purine and pyrimidine tion of KGF expression after injury to murine and metabolism (adenylosuccinate synthetase [EC 6.3.4.4], human skin (Werner et al., 1992; Marchese et al., 1995) phosphoribosyl pyrophosphate synthetase [EC 2.7.6.1], and also to other tissues (reviewed by Werner, 1998). amidophosphoribosyl transferase [EC 2.4.2.14], hypox- Furthermore, expression of KGF was shown to be anthine guanine phosphoribosyl transferase [EC 2.4.2.8] increased in human in¯ammatory diseases, such as and the multifunctional protein CAD which includes the psoriasis (Finch et al., 1997) and in¯ammatory bowel enzymatic activities of carbamoyl-phosphate syntheta- disease (Brauchle et al., 1996; Finch et al., 1996). These se II [EC 6.3.5.59], aspartate transcarbamylase results suggest that KGF stimulates proliferation and/or [EC 2.1.3.2] and dihydroorotase [EC 3.5.2.3]). Expres- migration of epithelial cells in injured and in¯amed sion of all of these enzymes was upregulated after tissues, a hypothesis which was supported by inhibition treatment with KGF and also with epidermal growth of wound re-epithelialization in transgenic mice expres- factor (EGF), indicating that these mitogens stimulate sing a dominant-negative KGF receptor in the epidermis nucleotide production by induction of these enzymes. To (Werner et al., 1994). determine a possible in vivo correlation between the However, mice lacking KGF had no obvious wound expression of KGF, EGF and the enzymes mentioned healing phenotype, suggesting the presence of other above, we analysed the expression of the enzymes during KGF receptor ligands which can compensate for the cutaneous wound repair, where high levels of these lack of KGF in the knockout mice (Guo et al., 1996). mitogens are present. Indeed, we found a strong mRNA Thereby the recently identi®ed ®broblast growth expression of all of these enzymes in the EGF- and factor 10 (FGF-10) is the most likely candidate, since KGF-responsive keratinocytes of the hyperproliferative it binds to the KGF receptor (a splice variant of FGF- epithelium at the wound edge, indicating that their receptor 2) with high anity (Igarashi et al., 1998). expression might also be regulated by growth factors Furthermore, both KGF and FGF-10 have a similar during wound healing. expression pattern in adult mouse tissues (Beer et al., 1997). Finally, preliminary studies from our laboratory Keywords: keratinocyte growth factor; keratinocytes; demonstrated similar e€ects of KGF and FGF-10 on purine; pyrimidine; nucleotide biosynthesis; salvage cultured keratinocytes. pathway Although a series of studies have been performed to elucidate the biological functions of KGF and its receptor, little is known about the molecular mechan- isms of KGF action. In an attempt to answer this Introduction question, we used the di€erential display RT ± PCR technology to identify genes which are regulated by this Keratinocyte growth factor (KGF, FGF-7) is a member growth factor in cultured keratinocytes. This approach of the rapidly growing ®broblast growth factor (FGF) led to the identi®cation of novel KGF-regulated genes family of mitogens. It was originally isolated as a lung encoding a homologue of the yeast protein CHL-1 which in yeast is responsible for chromosome segregation and cell cycle progression (Frank and Werner, 1996) and a non-selenium glutathione perox- idase (Frank et al., 1997). In this study we demonstrate *Correspondence: S Werner, Institute of Cell Biology, ETH HoÈ nggerberg, CH-8093 ZuÈ rich, Switzerland a potent e€ect of KGF on the expression of enzymes Received 14 April 1999; revised 22 July 1999; accepted 5 August 1999 which are involved in nucleotide biosynthesis. Using Growth factors and nucleotide biosynthesis MG Gassmann et al 6668 the di€erential display RT ± PCR strategy we identi®ed 10 ng/ml recombinant human KGF. Total cellular the adenylosuccinate lyase [EC 4.3.2.2] gene as a novel RNA was harvested after 1.5, 5 and 8 h and reverse

KGF-regulated gene in HaCaT keratinocytes. This transcribed using 5'-d(T12CA)-3' as a primer. DDRT ± enzyme catalyzes two steps in the biosynthesis of PCR was carried out using the same 3' primer and 5'- purine nucleotides, an early step required for the TGGATTGGTC-3' as 5' primer. A 320 bp fragment synthesis of all purines and a later step which is was found to be ampli®ed predominantly after KGF speci®c for the generation of adenosine monophos- treatment. It was isolated from the gel, reampli®ed by phate (for review see Zalkin and Dixon, 1992). To gain PCR and cloned into the transcription vector further insight into the potential role of nucleotide pBluescript KS II(+). Sequence comparison revealed biosynthesis for the mitogenic e€ect of KGF we cloned its identity with nt 1083 ± 1580 of the human fragments of the key regulatory enzymes involved in adenylosuccinate lyase (ADL) cDNA (Fon et al., purine and pyrimidine metabolism (shown in Figure 1). 1993; accession number: X65876). However, 178 bp Expression of all of these enzymes was induced after (nt 1190 ± 1367) are missing in our sequence, suggest- treatment of keratinocytes with KGF and also with ing that the transcript that we detected is the result of epidermal growth factor (EGF), indicating that these alternative splicing. To con®rm the KGF-regulated mitogens stimulate production of nucleotides by expression of ADL we analysed the expression of this increasing the expression of enzymes involved in the gene by RNase protection assay. As shown in Figure de novo synthesis and in the salvage pathway. 2, upper panel and in Table 2, a fourfold increase in Furthermore, we show a strong mRNA expression of ADL mRNA expression was observed within 5 h after four of these enzymes in the hyperproliferative epithelium of murine cutaneous 5d wounds where high levels of these growth factors are present. These results indicate a possible growth factor regulation of Table 1 List of primers used for PT-PCR enzymes involved in nucleotide biosynthesis during wound healing. mADL 1 5'-TGCTGAAAGTTACCGTTCTCCGCTG-3' mADL 2 5'-CCCAGATGAATAATGCCCGCAGCTT-3' ADS 1 5'-CTGGCGCAGGACGCCGACATCGTGTGCCGC-3' ADS 2 5'-TGTTCCTGGATACCATCAGCTGCTTGATGA-3' APRT 1 5'-CACTCTGGGACTCGTGGGGCTGCAGCACCG-3' Results APRT 2 5'-CGCCAGTAACTGGGTAATCATTTCACTATC-3' CAD 3 5'-GTTGGTGGGCTCTGGGCGCTGCTTTCTGAC-3' Adenylosuccinate lyase mRNA expression is upregulated CAD 4 5'-GAGTGGGAATCCCAGGGGGACAGAGGCGCT-3' by KGF and EGF HPRT 1 5'-GTGATAGATCCATTCCTATGACTGTAGATT-3' HPRT 2 5'-GAAGTATTCATTATAGTCAAGGGCATATCC-3' To identify novel KGF-regulated genes, we stimulated PPPS 1 5'-TGACCGCCTGGGCCTGGAGCTAGG-3' quiescent human keratinocytes (HaCaT cell line) with PPPS 3 5'-TTTAGGACAGCCGGCTCTGCATAC-3'

Figure 1 Overview of nucleotide biosynthesis. Substrates and products of nucleotide biosynthesis are shown in boxes. One step reactions are indicated by solid arrows. Multistep reactions are indicated by dotted arrows. Key regulatory enzymes involved in the di€erent reactions are indicated above or beside the arrows Growth factors and nucleotide biosynthesis MG Gassmann et al 6669 Table 2 Regulation of the expression of key regulatory enzymes involved in nucleotide biosynthesis by KGF, EGF, TNF-a, and TGF-b KGF EGF TNF-a TGF-b ADL 4 5.5 2 N.R. ADS 2 3 2 1.5 APRT 4.5 6 1.5 2 CAD 5.5 6.5 N.R. 2.5 HPRT 4.5 5.5 3.5 2.5 PRPS 3 3 2.5 3 The maximal degree of induction of the di€erent mRNAs compared to the basal level is indicated. N.R.: no regulation. Data represent the mean of two experiments

expression of this gene (Figure 2, lower panel and Table 2).

Cloning of cDNA fragments of the key regulatory enzymes in purine and pyrimidine metabolism To determine whether the KGF inducibility observed for adenylosuccinate lyase is a general feature of enzymes involved in purine and pyrimidine biosynth- esis, we cloned cDNA fragments of ®ve key regulatory enzymes involved in nucleotide metabolism: human adenylosuccinate synthetase (ADS), the enyzme responsible for converting inosine monophosphate into adenylosuccinate, phosphoribosyl pyrophosphate synthetase (PRPS) which generates phosphoribosyl pyrophosphate, a molecule required for de novo production of both purine and pyrimidine nucleotides and for the purine salvage pathway, amidophosphor- ibosyl-transferase (APRT), a key regulatory enzyme in purine biosynthesis, hypoxanthine guanine phosphor- ibosyl-transferase (HGPRT, HPRT), an important Figure 2 KGF-regulated expression of adenylosuccinate lyase. enzyme of the salvage pathway, and of the multi- HaCaT keratinocytes were rendered quiescent by serum functional protein CAD (carbamoylphosphate synthe- starvation as described in Materials and methods. Cells were tase II, aspartate transcarbamylase dihydroorotase), subsequently treated with di€erent growth factors and cytokines which catalyzes the initial steps in pyrimidine de for 1.5 ± 8 h as indicated. Twenty mg total cellular RNA was novo synthesis (Figure 1). We obtained a 293 bp isolated before and at di€erent time points after growth factor or cytokine addition and analysed by RNase protection assay for cDNA fragment of the human ADS. This sequence is the presence of ADL mRNA. Twenty mg tRNA were used as a 100% identical to the sequence (nt 219 ± 511) pub- negative control. Thousand c.p.m. of the hybridization probe lished by Powell et al. (1992). Furthermore, we cloned were loaded in the lane labeled `probe' and used as a size fragments encoding parts of PRPS subunit one marker. One mg of each RNA sample was loaded on a 1% agarose gel and stained with ethidium bromide. Pictures of the (398 bp; nt 183 ± 580; Sonoda et al., 1991), APRT RNA gels are shown below the RNase protection assays (labeled (368 bp; nt 220 ± 587; Iwahana et al., 1993a), HPRT RNA) (326 bp; nt 355 ± 680; Jolly et al., 1983), and of the multifunctional protein CAD (310 bp; nt 871 ± 1180; Iwahana et al., 1996). In addition to the human cDNAs we ampli®ed murine cDNA fragments corresponding to ADL, KGF treatment. The induction was reversible and APRT, HPRT and the multifunctional protein CAD. expression returned to almost basal levels 3 h later. A For ADL and HPRT we obtained cDNAs 100% similar regulation was found with the KGF homo- identical to the published sequences (ADL: 301 bp; logue FGF-10 (data not shown). In addition to the nt 60 ± 360, (Wong and O'Brien, 1995); HPRT: KGF receptor ligands, EGF also stimulated the 336 bp; nt 335 ± 670, (Konecki et al., 1982)). For the expression of the ADL gene but with a di€erent APRT we obtained a 369 bp fragment which was kinetic. The induction of ADL expression by EGF 91.4% identical to nt 167 ± 535 of the rat cDNA was delayed compared to the regulation by KGF. (Iwahana et al., 1993b) and 82.7% identical to However, it was stronger and long-lasting with nt 220 ± 588 of the human cDNA (Iwahana et al., maximal mRNA levels seen 8 h after addition of the 1993a). For the CAD protein we cloned a 314 bp mitogen (Figure 2, upper panel and Table 2). In cDNA fragment which was 89% identical to nt 869 ± contrast to the epithelial mitogens EGF and KGF, 1182 of the human sequence (Iwahana et al., 1996) TGF-b1 and TNF-a which inhibit keratinocyte and 88% identical to nt 108 ± 421 of the hamster proliferation had no or only a minor e€ect on the sequence (Simmer et al., 1990). Growth factors and nucleotide biosynthesis MG Gassmann et al 6670

Figure 3 Growth factor regulation of key regulatory enzymes involved in nucleotide biosynthesis. Quiescent HaCaT keratinoytes were incubated with di€erent growth factors and cytokines as described in the legend of Figure 2. Twenty mg total cellular RNA were analysed for the expression of ADS, PRPS, HPRT, CAD and APRT. Twenty mg tRNA were used as a negative control in all protection assays. Thousand c.p.m. of the di€erent hybridization probes were loaded in the lanes labeled `probe' and used as size markers. The same batch of RNAs was used for all protection assays shown in Figures 2 and 3. One mg of each RNA sample was loaded on a 1% agarose gel and stained with ethidium bromide. Pictures of the RNA gels are shown below the RNase protection assays (labeled RNA)

the presence of EGF than in the presence of KGF. The mRNA expression of the key regulatory enzymes In addition to the keratinocyte mitogens EGF and involved in purine and pyrimidine metabolism is KGF, the cytokine TNF-a also stimulated the upregulated by KGF and EGF expression of most of the enzymes, although to a As shown in Figure 3 and in Table 2, the mRNA lesser extent. TGF-b1 had only a minor e€ect on expression of all the enzymes described above was the expression of the HPRT and PRPS genes, stimulated by KGF and EGF treatment. The whereas the other genes were una€ected by this weakest increase was observed for ADS (twofold factor. These results were reproduced with RNAs after KGF and threefold after EGF treatment) while from two independent tissue culture experiments. the strongest increase was detected for the CAD Taken together, these ®ndings demonstrate a similar mRNA (5 ± 6-fold after KGF and 6 ± 7-fold after regulation of various key regulatory enzymes in EGF treatment). As already demonstrated for ADL, nucleotide biosynthesis by growth factors and EGF was a stronger inducer of all enzymes cytokines, indicating that each factor coordinately compared to KGF. In addition, upregulation of regulates multiple steps in the production of purines ADS, APRT and CAD expression occurred later in and pyrimidines. Growth factors and nucleotide biosynthesis MG Gassmann et al 6671 Expression of adenylosuccinate lyase, amidophosphoribosyltransferase, hypoxanthine phosphoribosyltransferase and the multifunctional protein CAD in normal and wounded skin To determine a possible regulation of the enzymes by KGF in vivo we analysed their expression during cutaneous wound repair in mice. This model was chosen since high levels of KGF and of other growth factors are present in the healing wound. As shown in Figure 4, ADL and APRT mRNA levels increased twofold within 1 day after wounding. Expression of these enzymes subsequently declined. HPRT mRNA levels increased continuously after injury and maximal levels were seen between day 7 and 14 after wounding. No signi®cant changes in CAD expression were observed during wound healing. These results were reproduced with a di€erent set of RNAs from an independent wound healing experiment.

Enzymes involved in nucleotide biosynthesis are predominantly expressed in the hyperproliferative epithelium at the wound edge In order to localize the expression of ADL, APRT, HPRT and CAD in the wound we performed in situ hybridizations with 5 day wounds. At this stage after injury, partial re-epithelialization of the wound has occurred, and a thick hyperproliferative epithelium as Figure 4 Expression of ADL, APRT, CAD and HPRT during well as an extensive granulation tissue is present in the cutaneous wound repair in mice. Mice were wounded as described wound. As shown in Figures 5 and 6b,f, weak signals in Materials and methods and sacri®ced at di€erent time points were observed throughout the entire wound area, after injury. Twenty mg total cellular RNA from normal and wounded skin were analysed by RNase protection assay for the re¯ecting the ubiquitous expression of these enzymes. expression of the di€erent enzymes. Twenty mg tRNA were used A particularly strong and speci®c signal was detected as a negative control. Thousand c.p.m. of the hybridization in the hyperproliferative epidermis within the wound probes were loaded in the lanes labeled `probe' and used as size (Figures 5 and 6a,b,d,f,g,h) but not in the epidermis markers. The same set of RNAs were used for all protection adjacent to the wound (data not shown). No speci®c assays shown in this ®gure. One mg of each RNA sample was loaded on a 1% agarose gel and stained with ethidium bromide. signals were obtained with the antisense riboprobes A picture of the RNA gel is shown below the RNase protection (Figures 5 and 6c,g). The enzymes involved in assays (labeled RNA) nucleotide de novo synthesis (ADL, APRT, CAD) showed a strong expression only in the suprabasal layers of the hyperproliferative epithelium at the wound edge (Figures 5 and 6a ± d), whereas no signals were seen in the basal and lower suprabasal of the regulatory enzymes has also been demonstrated cells as well as in the migrating keratinocytes of the (for review see Carrey, 1995). By contrast, little is most extended epithelial tongue. By contrast, expres- known about a possible long-term regulation of sion of HPRT, an enzyme of the purine salvage enzymes involved in nucleotide biosynthesis at the pathway, was distributed over the entire hyperproli- transcriptional level. A cell cycle dependent expression ferative epithelium. In addition, the migrating of the CAD and APRT genes was demonstrated in keratinocytes at the wound edge expressed this di€erent cell types (Liao et al., 1986; Rao and enzyme (Figure 6e ± h). Davidson, 1988; Rao and Church, 1988; Iwahana et al., 1995), indicating that the expression of these enzymes might be regulated by growth factors. Discussion In this study we identi®ed the ADL gene as a novel KGF-regulated gene. ADL catalyzes two similar When quiescent cells reenter the cell cycle there is the reactions in the purine de novo biosynthetic pathway: need for increasing the cellular nucleotide pool. the conversion of succinylaminoimidazole carboxa- Nucleotides are required for many cellular processes mide ribotide to aminoimidazole carboxamide and including DNA and RNA synthesis, energy storage the reaction from adenylosuccinate to adenosine and metabolism, activation of phospholipids and monophosphate (Zalkin and Dixon, 1992). In addi- carbohydrates and as general metabolic regulators tion to ADL, we found a positive regulation of (Voet and Voet, 1995). The extent of nucleotide various key regulatory enzymes involved in purine and synthesis is predominantly regulated by allosteric or pyrimidine de novo synthesis and in the salvage competitive inhibition or activation of key regulatory pathway by EGF and KGF in cultured keratino- enzymes by various intermediates or end products of cytes. Since both factors are strongly mitogenic for the metabolic pathways. In addition, phosphorylation primary keratinocytes and for HaCaT cells (Rhein- Growth factors and nucleotide biosynthesis MG Gassmann et al 6672 wald and Green, 1977; Rubin et al., 1989; Game et al., 1986; Detmar and Orfanos, 1990), these factors al., 1992; H Smola, personal communication), our caused only a minor or no induction of the enzymes results indicate that stimulation of nucleotide bio- involved in nucleotide biosynthesis. synthesis via induction of enzymes with regulatory A remarkable ®nding was the similarity in the function could be a novel mechanism of growth factor regulation of all enzymes. Thus EGF was the stronger action. Consistent with the inhibitory e€ect of TGF-b inducer compared to KGF, but KGF caused an earlier and TNF-a on keratinocyte proliferation (Shipley et induction. These data suggest that growth factors

Figure 5 Expression of ADL, and APRT in 5 day full-thickness excisional wounds. Wounds were generated and processed as described in Materials and methods. Frozen sections from the middle of the wound were hybridized with 35S-labeled sense (c,g)or antisense (a,b,d,e,f,h) riboprobes complementary to the murine ADL, and APRT mRNAs. Overviews of half of the wound area are shown in a,b,c,e,f and g (magni®cation 1006). 4006 magni®cations of the area indicated by the rectangles in (a and e) are shown in (d and h). Signals appear as black dots in the bright ®eld surveys and as white dots in the dark ®eld surveys. ES: Eschar, HE: hyperproliferative epithelium, G: granulation tissue Growth factors and nucleotide biosynthesis MG Gassmann et al 6673 coordinately regulate a complete metabolic pathway mice. We previously demonstrated a strong induction such as nucleotide biosynthesis. of the expression of KGF (Werner et al., 1992), To determine a possible regulation of nucleotide TGF-b (Frank et al., 1996) and TNF-a (HuÈ bner et biosynthesis by growth factors in vivo, we analysed al., 1996) in the same wound healing model. the expression of four important enzymes involved Furthermore, a strong expression of epidermal in purine and pyrimidine biosynthesis during the growth factor receptor ligands in the healing wound healing process of full-thickness excisional wounds in has also been demonstrated (Todd et al., 1991;

Figure 6 Expression of CAD and HPRT in 5 day full-thickness excisional wounds. Wounds were generated and processed as described in Materials and methods. Frozen sections from the middle of the wound were hybridized with 35S-labeled sense (c,g)or antisense (a,b,d,e,f,h) riboprobes complementary to the murine, CAD and HPRT mRNAs. Overviews of half of the wound area are shown in a,b,c,e,f and g (magni®cation 1006). 4006 magni®cations of the area indicated by the rectangles in (a and e) are shown in (d and h). Signals appear as black dots in the bright ®eld surveys and as white dots in the dark ®eld surveys. ES: Eschar, HE: hyperproliferative epithelium, G: granulation tissue Growth factors and nucleotide biosynthesis MG Gassmann et al 6674 Marikowsky et al., 1993), indicating that these 20 h and subsequently treated with fresh DMEM containing factors might stimulate expression of enzymes 10 ng/ml recombinant human keratinocyte growth factor involved in nucleotide biosynthesis during wound (KGF), 20 ng/ml epidermal growth factor (EGF), 1 ng/ml repair. Indeed, we found increased mRNA levels of transforming growth factor-b1 (TGF-b1) or 300 U/ml some of these enzymes after skin injury. Most tumour necrosis factor-a (TNF-a), respectively. Cells were harvested at di€erent time points after growth factor interestingly, all enzymes were expressed at particu- stimulation and used for RNA isolation. Each experiment larly high levels in the keratinocytes of the was repeated at least twice. DMEM and FCS were purchased hyperproliferative epithelium at the wound edge. from Gibco/BRL; growth factors were from Roche Since keratinocytes are targets for all the factors Biochemicals (Germany). analysed in this study in vitro and in vivo (Rheinwald and Green, 1977; Trefzer et al., 1991; Wenczak et al., 1992; Werner et al., 1994; Marchese RNA isolation and RNase protection assay et al., 1995; Gold et al., 1997; Ashcroft et al., 1997; Isolation of total cellular RNA was performed as described Schmid et al., 1998; this study), it seems likely that by Chomczynski and Sacchi (1987). RNase protection assays the expression of the enzymes involved in nucleotide were carried out according to Werner et al. (1992). biosynthesis might also be regulated by growth Autoradiograms were scanned and analysed with the Scion factors in the healing wound. Image software (Scion Corporation, Maryland, USA). The broadest distribution within the hyperproli- ferative epithelium was detected for HPRT, an DDRT ± PCR enzyme of the purine salvage pathway. The mRNA encoding this enzyme was found in the proliferating DDRT ± PCR was carried out according to Bauer et al. (1994). Brie¯y, 2 mg of total cellular RNA were reverse basal and lower suprabasal cells and also in the transcribed using 5'-d(T12CA)-3' (A) as a primer. PCR was redi€erentiating cells of the upper suprabasal layers. performed in a total reaction volume of 20 ml using 5 mCi The expression of this enzyme in the proliferating [35S]dATP, 25 mM primer A, 5 mM primer B (5'- keratinocytes at the wound edge is consistent with TGGATTGGTC-3') and 1 U Taq polymerase (Perkin- previous ®ndings demonstrating high activities of the Elmer, Norwalk, CT, USA). Thirty-®ve cycles were salvage enzymes in hyperkeratotic epidermis (Ogura performed (30 s 948C, 60 s 428C, 30 s, 728C). and Kumano, 1983). It has been suggested that the increased nucleotide pool plays a role as a source of Cloning of cDNA fragments cellular energy and as a precursor of nucleic acids for epidermal proliferation (Ogura and Kumano, 1983). The cDNA fragments which we used as templates for the Whereas the HPRT in the proliferating basal and generation of riboprobes were all generated by RT ± PCR. lower suprabasal keratinocytes of the hyperthickened For the ampli®cation of mouse probes RNAs from 1 day wound epidermis is likely to be important for the wounds and lung tissue were used. For the human probes RNA from exponentially growing HaCaT keratinocytes was production of nucleotides required for DNA synth- used. RNA was reverse transcribed and the resulting cDNA esis, the expression of this enzyme in the differ- was ampli®ed by PCR and cloned into the transcription entiated suprabasal cells suggests an additional role vector pBluescript KSII(+) (Stratagene, La Jolla, CA, USA). of APRT in the recycling of free purine bases which The primers that were used are listed in Table 1. PCR was are generated upon degradation of the cellular DNA performed at annealing temperatures ranging from 55 ± 688C, during the process of terminal di€erentiation. depending on the primers. In contrast to HPRT, high levels of CAD, ADL and APRT were only found in a speci®c population of Animals keratinocytes in the higher suprabasal layers which have already undergone redi€erentiation and which BALB/c mice were obtained from the animal care facility of have lost the proliferative capacity. Therefore, the the Max-Planck-Institute of Biochemistry. They were housed strong expression of enzymes involved in purine and and fed according to Federal guidelines, and all procedures pyrimidine de novo synthesis in these cells is unlikely to were approved by the local Government of Bavaria. play a role in the regulation of cell proliferation. These enzymes might thus be required for the production of Wounding and preparation of wound tissue nucleotides which serve as energy donors in various Two independent wound-healing experiments were per- cellular processes occurring in the course of keratino- formed. For each experiment, 24 BALB/c mice (8 ± 12 weeks cyte di€erentiation. of age) were anaesthesized with a single intraperitoneal In summary, our data suggest that induction of key injection of avertin. The hair on the animals' back was cut, regulatory enzymes involved in nucleotide biosynthesis and the skin was wiped with 70% ethanol. Six full-thickness could be a novel mechanism of growth factor action in excisional wounds (6 mm diameter, 3 ± 4 mm apart) were vitro and possibly also in vivo. generated on the back of each animal by excising skin and panniculus carnosus. The wounds were allowed to dry to form a scab. At di€erent time points after injury (1 ± 14 days), animals were sacri®ced, and the complete wounds including Materials and methods 2 mm of the wound margins were isolated. At each time point, the tissue from four animals was combined, immediately frozen in liquid nitrogen, and used for isolation Cell culture and growth factor treatment of keratinocytes of total cellular RNA. A similar amount of skin from HaCaT keratinocytes (Boukamp et al., 1988) were grown to nonwounded animals served as control. con¯uency in Dulbecco's modi®ed Eagle's medium (DMEM) For in situ hybridization, complete wounds were isolated containing 1% Pen/Strep and 10% fetal calf serum (FCS). at day 5 after injury, bisected and ®xed overnight at 48Cin They were rendered quiescent by serum starvation for 16 ± 4% paraformaldehyde in PBS. They were subsequently Growth factors and nucleotide biosynthesis MG Gassmann et al 6675 incubated in 15% sucrose in PBS for 4 h at 48C and frozen in Acknowledgements tissue freezing medium (Jung, Nussloch, Germany). We thank Dr Hans Smola, University of Cologne, for providing unpublished data, and Helga Riesemann, Max- Planck-Institute of Biochemistry, Martinsried, for excellent In situ hybridization technical assistance. This work was supported by grants Sense and antisense riboprobes were generated using T3 or from the Deutsche Forschungsgemeinschaft (We 1983/1-3), T7 RNA polymerases and [a-35S]UTP. Frozen sections (6 mm the Human Frontier Science Program and the Hermann- from the middle of 5d wounds were hybridized as described and-Lilly Schilling Stiftung (to S Werner). by Wilkinson et al. (1987). After hybridization, sections were coated with NTB2 nuclear emulsion (Kodak) and exposed in the dark for 3 ± 4 weeks. After development sections were counterstained with hematoxylin/eosin and mounted.

References

Ashcroft GS, Horan MA and Ferguson MW. (1997). J. Jolly DJ, Okayama H, Berg P, Esty AC, Filpula D, Bohlen P, Anat., 190, 351 ± 365. Johnson GG, Shively JE, Hunkappillar T and Friedman Bauer D, Warthoe P, Rohde M and Strauss M. (1994). T. (1983). Proc. Natl. Acad. Sci. USA, 80, 477 ± 481. Genome Res., 4, 97 ± 108. Konecki DS, Brennand J, Fuscoe JC, Caskey CT and Beer HD, Florence C, Dammeier J, McGuire L, Werner S Chinault AC. (1982). Nucl. Acid. Res., 10, 6763 ± 6775. and Duan DR. (1997). Oncogene, 15, 2211 ± 2218. Liao WS, Heller R, Green P and Stark GR. (1986). J. Biol. Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Chem., 261, 15577 ± 15581. Markham A and Fusenig NE. (1998). J. Cell Biol., 106, Marchese C, Chedid M, Dirsch OR, Csaky KG, Santanelli F, 761 ± 772. Latini C, Larochelle WJ, Torrisi MR and Aaronson SA. Brauchle M, Madlener M, Wagner AD, Angermeyer K, (1995). J. Exp. Med., 182, 1369 ± 1376. Lauer U, Hofschneider PH, Gregor M and Werner S. Marikowsky M, Breuing K, Liu PY, Eriksson E, Higashiya- (1996). Am.J.Path.,149, 521 ± 529. ma S, Farber P, Abraham J and Klagsbrun M. (1993). Carrey EA. (1995). Bioc. Soc. Transact., 23, 899 ± 902. Proc. Natl. Acad. Sci. USA, 90, 3889 ± 3893. Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162, Ogura R and Kumano S. (1983). Curr. Probl. Dermatol., 11, 156 ± 159. 145 ± 158. Detmar M and Orfanos CE. (1990). Arch. Dermatol. Res., Powell S, Zalkin H and Dixon JE. (1992). FEBS Lett., 303, 282, 238 ± 245. 4±10. FinchPW,PricoloV,WuAandFinkelsteinSD.(1996). Rao GN and Church RL. (1988). Exp. Cell Res., 178, 449 ± Gastroenterol., 110, 441 ± 451. 456. Finch PW, Murphy F, Cardinale I and Krueger JG. (1997). Rao GN and Davidson JN. (1988). DNA, 7, 423 ± 432. Am.J.Pathol.,151, 1619 ± 1628. Rheinwald JG and Green H. (1977). Nature, 265, 421 ± 424. FonEA,DemczukS,DelattreO,ThomasGandRouleau RubinJS,BottaroDP,ChedidM,MikiT,RonD,Cheon GA. (1993). Cytogenetics Cell Genetics, 64, 201 ± 203. HG, Taylor WG, Fortney E, Sakata H, Finch PW and Frank S, Madlener M and Werner S. (1996). J. Biol. Chem., Larochelle WJ. (1995). Cell Biol. Intern., 19, 399 ± 411. 271, 10188 ± 10193. Rubin JS, Osada H, Finch PW, Taylor WG, Rudiko€ S and Frank S and Werner S. (1996). J. Biol. Chem., 271, 24337 ± Aaronson SA. (1989). Proc.Natl.Acad.Sci.USA,86, 24340. 802 ± 806. Frank S, Munz B and Werner S. (1997). Oncogene, 14, 915 ± Schmid P, Itin P, Cherry G, Bi C and Cox DA. (1998). Am. J. 921. Pathol., 152, 485 ± 493. Game SM, Huelsen A, Patel V, Donnelly M, Yeudall WA, Shipley GD, Pittelkow MR, Wille JJJ, Scott RE and Moses Stone A, Fusenig NE and Prime SS. (1992). Int. J. Cancer, HL. (1986). Cancer Res., 46, 2068 ± 2071. 52, 461 ± 470. Simmer JP, Kelly RE, Rinker Jr AG, Scully JL and Evans Gold LI, Sung JJ, Siebert JW and Longaker MT. (1997). Am. DR. (1990). J. Biol. Chem., 265, 10395 ± 10402. J. Pathol., 150, 209 ± 222. Sonoda T, Taira M, Ishijima S, Ishizuka T, Lizasa T and GuoL,DegensteinLandFuchsE.(1996).Genes Dev., 10, Tatibana M. (1991). J. Biochem., 109, 361 ± 364. 165 ± 175. Todd R, Dono€ BR, Chiang T, Chou MY, Elovic A, HuÈ bner G, Brauchle M, Smola H, Madlener M, FaÈ ssler R Gallagher GT and Wong DT. (1991). Am.J.Pathol.,138, and Werner S. (1996). Cytokine, 8, 548 ± 556. 1307 ± 1313. Igarashi M, Finch PW and Aaronson SA. (1998). J. Biol. Trefzer U, Brockhaus M, Loetscher H, Parlow F, Kapp A, Chem., 273, 13230 ± 13235. Schopf E and Krutmann J. (1991). J. Invest. Dermatol., 97, Iwahana H, Oka J, Mizusawa N, Kudo E, Ii S, Yoshimoto K, 911 ± 916. Holmes EW and Itakura M. (1993a). Biochem. Biophys. Voet D and Voet JG. (1990). Biochemistry, 2nd edn. Wiley: Res. Commun., 190, 192 ± 200. New York. Iwahana H, Yamaoka T, Mizutani M, Mizusawa N, Ii S, Wenczak BA, Lynch JB and Nanney LB. (1992). J. Clin. Yoshimoto K and Itakura M. (1993b). J. Biol. Chem., 268, Invest., 90, 2392 ± 2401. 7225 ± 7237. Werner S. (1998). Cytokine Growth Factor Rev., 9, 153 ± 165. Iwahana H, Honda S, Tsujisawa T, Takahashi Y, Adzuma K, Werner S, Peters KG, Longaker MT, Fuller-Pace F, Banda Katashima R, Yamaoka T, Moritani M, Yoshimoto K MJ and Williams LT. (1992). Proc. Natl. Acad. Sci. USA, and Itakura M. (1995). Biochim. Biophys. Acta, 1261, 369 ± 89, 6896 ± 6900. 380. Werner S, Smola H, Liao X, Longaker MT, Krieg T, IwahanaH,FujimuraM,IiS,KondoM,MoritaniM, Hofschneider PH and Williams LT. (1994). Science, 266, Takahashi Y, Yamaoka T, Yoshimoto K and Itakura M. 819 ± 822. (1996). Biochem. Biophys. Res. Commun., 219, 249 ± 255. Growth factors and nucleotide biosynthesis MG Gassmann et al 6676 Wilkinson D, Bailes J, Champion J and McMahon A. (1987). Zalkin H and Dixon JE. (1992). Progr. Nucl. Acid Res. Mol. Development, 99, 493 ± 500. Biol., 42, 259 ± 287. Wong LJ and O'Brien WE. (1995). Genomics, 28, 341 ± 343.