139

Molecular and Cellular Biochemistry 221: 139–145, 2001.© 2001 Kluwer Academic Publishers. Printed in the Netherlands.

Estrogen modulates the expression of L-arginine:glycine amidinotransferase in chick liver

Yunfeng Zhu1 and Marilyn I. Evans2

Departments of 1Anatomy; 2Biochemistry, West Virginia University School of Medicine, Morgantown, WV, USA

Received 17 January 2001; accepted 19 March 2001

Abstract

Identification of estrogen-responsive genes is important to understand the molecular mechanisms of estrogen action. Suppres-sion subtractive hybridization was employed to screen estrogen-responsive genes in chick liver. A single injection of estrogeninto 6-week-old chick induced up-regulation of several known genes encoded for yolk proteins, such as Vitellogenin I and IIand very low density lipoprotein II (apo-VLDL II). One novel sequence displayed a dramatic change (3-fold increase) in re-sponse to estrogen treatment. This cDNA fragment was extended and the resultant sequence was analyzed. Translated aminoacid sequence was 90, 88, 83 and 87% identical to the L-arginine:glycine amidinotransferase of pig, rat, frog and human, re-spectively. The sequence has a conservative catalytic site of L-arginine:glycine amidinotransferase. The expression pattern ofthis gene in organs is consistent with previous reports of L-arginine:glycine amidinotransferase in chick. Thus, this clone rep-resented the chicken L-arginine:glycine amidinotransferase. It appeared that estrogen-induced alteration of arginine:glycineamidinotransferase was not dependent on protein synthesis, because concurrent administration of cycloheximide did not af-fect the estrogen-mediated expression pattern. This is the first study demonstrating that L-arginine:glycine amidinotransferaseis a target of the estrogen receptor. (Mol Cell Biochem 221: 139–145, 2001)

Key words: cycloheximide, subtractive hybridization, Southern blot, Northern blot, transcriptional regulation

Introduction

Sex hormones, such as estrogen, have profound effects onbiological systems. Estrogen, acting via its receptor, regulatesthe development and the function of various human organsystems including reproductive, skeletal, and the nervous sys-tems. Estrogen is also directly involved in numerous patho-logical processes, such as carcinogenesis, mental disorder,osteoporosis, cardiovascular disease [1–6]. Identification ofmolecular targets of estrogen is a critical step to elucidate themechanisms of estrogen action.

The present study uses a chick model to screen estrogen-re-sponse genes (ERGs). The hormone-naïve liver from 6-week-old chicks provides an excellent model system for studyingestrogen action [7]. In contrast to the oviduct, which requiresestrogen for differentiation, the liver of chick is differentiatedand fully functioning in the absence of estrogen. This unique

feature of chick liver allows the examination of gene regulationprior to, during, and following the administration of estrogen.

Suppression subtractive hybridization was employed toidentify the ERGs. This technique offers two major advan-tages, i.e. the inclusion of all of the sequences in the cDNApool, and the equalization of rare and abundant messagesfollowing the initial hybridization [8]. To isolate both estrogen-induced and suppressed genes, two patterns of subtractivehybridization were designed. First, to isolate estrogen-inducedgenes, inductive subtractive hybridization was performedwith control birds (no estrogen treated) as a driver andestrogen-treated birds as a tester. In contrast, suppressivesubtractive hybridization with control birds as a tester andestrogen-treated birds as a driver was carried out to screenestrogen-suppressed genes. With this technique, we demon-strate that L-arginine:glycine amidinotransferase is a noveltarget of estrogen action.

Address for offprints: Y. Zhu, Department of Anatomy, West Virginia University School of Medicine, Robert C. Byrd Health Science Center, Morgantown,WV 26506-9128, USA

140

Materials and methods

Animals: hormone and cycloheximide treatment

Six week old White Leghorn chickens (Truslow Farms, MD,USA) received a single intramuscular injection of 17β-hy-droxyl estradiol (20 mg/kg body wt) dissolved in propyleneglycol. Control birds were injected with propylene glycolonly 2 h prior to sacrifice. For the cycloheximide (CHX)experiments, the birds received a concurrent treatment ofestradiol at 20 mg/kg body wt and CHX at 3 mg/kg body wt.Our results showed that CHX injection at this concentrationblocked protein synthesis and inhibited estrogen-mediatedexpression of egg yolk precursor proteins (data not shown).All birds were sacrificed by cervical dislocation at the timesindicated in the figures following either estradiol only orCHX plus estradiol treatment. The liver was quickly minced,frozen in liquid nitrogen, and stored at –70°C until use. Allexperiments were conducted in accordance with the NIBGuide for the Care and Use of Laboratory Animals.

Preparation of mRNA

Total RNA was isolated from frozen liver tissue by the so-dium dodecyl sulfate (SDS)-phenol-LiCl procedure [9].mRNA was purified using Poly A tract (Promega, Madison,WI, USA). The concentration was calculated by absorbancevalue at 260 nm and the ratio of 260/280 nm.

Subtractive hybridization cloning

To reduce the effects of variation among individuals, poolsof equal amounts poly A+ RNA mixture from four femalebirds were used for each time point. The PCR-Select cDNASubtraction Kit (Clontech Laboratories, Palo Alto, CA, USA)[10] was used to isolate differences between untreated and48 h estrogen-treated chicken liver. The subtracted fragmentswere cloned directly into the TA vector pCR 2.1 (Invitrogen,Carlsbad, CA, USA), then transfected into INVaF’ cells(Invitrogen, Carlsbad, CA, USA). The sequence of the DNAinserts was determined with a Pharmacia ALF DNA sequencerusing the ALF express Auto Cycle Sequencing Kit (AmershamPharmacia Biotech, Piscataway, NJ, USA).

Northern blot hybridization

The cDNA probes were derived from subcloned DNA. Theprobes were labeled with [32P]dCTP by PCR. The reactionmixture contained the following components: [32P]dCTP (3000Ci/mM, ICN), templates (plasmids harbored subtractive DNA

fragments, 20 ng/ml), 50 × dNTP (10 mM of dATP, dGTP, anddTTP, and 100 µM of dCTP), Taq DNA polymerase (5 U/µl),and primers (same as subtractive hybridization PCR primers).Amplification was performed at 95°C for 3 min for denatur-ing followed by 30 cycles of additional reaction (94°C for30 sec, 55°C for 30 sec, and 72°C for 2 min). Labeled probeswere then purified by G-50 column (Promega, Madison, WI,USA).

Ten microgram of total RNA per lane was separated byelectrophoresis on a 1.0% agarose/fomaldehyde gel. TheRNA was transferred to Nytran filters by capillary action andwas ultraviolet-cross-linked. The hybridization was perfomedwith the following procedure: Pre-hybridization occurred byincubating filters with a hybridization buffer (ULTRAhybTMBuffer, Austin, TX, USA) at 42°C for 30 min. After pre-hy-bridization, the filters were hybridized with cDNA probes atfinal concentration of 1 × 106 cpm/ml in the same hybridiza-tion buffer (42°C, overnight). The filters were washed threetimes (5 min each) at 42°C with 2 × SSC , 0.1% SDS solutionfollowed by three additional washes (15 min each) at 42°C in0.1 × SSC , 0.1% SDS solution. The signals were processedwith a PhosphorImager and analyzed with Image QuaNT soft-ware (Molecular Dynamics, Sunnyvale, CA, USA).

Southern blot hybridization

The chicken genomic DNA was extracted and digested withrestrictive enzymes EcoRI, BamHI, HindIII and Bgl II re-spectively and separated on a 0.8% agarose gel. The DNAwas transferred to Nytran filter by capillary action after de-naturing with 1.5 M NaCl, 0.5 N NaOH, and was cross-linkedby ultraviolet. The procedures of cDNA probe preparationand hybridization were the same as the Northern blot.

5′ Rapid Amplification of cDNA Ends (5′ RACE)

The extension of isolated cDNA fragments was performedusing a commercial kit (Rapid Amplification of cDNA Ends,version 2.0, Life Technologies, Grand Island, NY, USA).Three Gene Specific Primers (GSP) were designed based ontarget gene sequence. In brief, mRNA was reverse transcribedusing GSP1. Then, the cDNA was tailed with poly-dC by ter-minal deoxynucleotidyl transferase and amplified by PCRwith a bridged anchor primer (provided by kit) and GSP2.The extended fragment was inserted into the TA cloning vec-tor pCR2.1 (Invitrogen, Carlsbad, CA, USA). The positiveclones were screened by PCR using a bridged universal am-plification primer (provided by kit) and GSP3. The extendedfragment was sequenced by the Research Services Molecu-lar Genetics Facility (University of Georgia, Athens, GA,USA).

141

Statistical analysis

Differences among treatment groups were tested using a one-way analysis of variance (ANOVOA). Differences in whichp < 0.05 were considered statistically significant. In caseswhere significant differences were detected, specific post-hoccomparisons between treatment groups were examined withStudent-Newman-Keuls tests.

Results

Isolation of estrogen responsive sequences

Chicken liver Poly A+ RNA from control (no estrogen treated)and estrogen-treated (48 h of exposure) were prepared asdescribed in the Materials and methods section and subjectedto inductive and suppressive subtractive hybridization respec-tively. The subtractive gene fragments were isolated andsequenced. Total of 12 estrogen-induced clones and 26estrogen-suppressed clones were obtained. All estrogen-induced clones were known egg yolk protein cDNA, i.e.vitellogenin and apoVLDLII. In order to characterize theestrogen-mediated alteration, a study of time sequence ofgene expression with Northern blot was performed. One450 bp cDNA fragment cloned as an estrogen-suppressedgene, tentatively named 4A3, displayed an initial 3-fold in-crease at 1 h after an injection of estradiol followed by a de-cline below baseline at 1.5 h for its transcription (Fig. 1). Thenucleotide sequence of 4A3 did not match any known geneavailable in the GenBank/EMBL. Since this novel gene ismost affected by estrogen exposure, further characterizationof this gene was performed.

Characterization of estrogen-responsive gene

Using a specific probe derived from 4A3, further analy-sis with Northern blot indicated that the mRNA was around2.4 kb (data not shown). Southern blot showed a single bandfollowing digestion by four different restrictive enzymes (Fig.2). This result suggested that this fragment belong to a sin-gle copy gene. To extend this gene fragment, the 5′ RACEwas employed as description in Materials and methods. Anapproximately 1.7 kb additional fragment was obtained andits sequence was analyzed. The total length of the sequenceis approximately 200 bp shorter than the size estimated byNorthern-blot. The sequence obtained consists of a 1266 bpopen reading frame with an 884 bp 3′ untranslated region.The 3′ end of this sequence is not polyadenylated nor does itcontain an AATAAA cleavage signal, but rather ends in a RsaI half site (Fig. 3). Rsa I is the four base cutter used to frag-ment the cDNA prior to ligation of the adapters during thecloning procedure, so it is obvious that the 3′ end of the

mRNA is missing. However, the flank sequence of the firstATG complies with the rule of defining starting codon [11],and with a suggestion that the code sequence we obtained wasfull length. By searching the GenBank/EMBL database, thisgene display a high identification with L-arginine:glycineamidinotransferase (AT) gene in different species and thededuced amino acid sequence was aligned with that of hu-man, pig, rat and frog (Fig. 4).

Data base analysis revealed that the amino acid sequencefrom the cleavage site for the signal peptide to the end of theopen reading frame (amino acids 38–422) is 87% identicalto the AT sequence of human (accession number NM-001482),88% identical to the rat (accession number U07971), 90%identical to the pig [12], and 83% identical to the frog se-quence (accession number AF187863). The catalytic site ofAT consists of cysteine 406, histidine 302, and aspartate 253.The arginine substrate is fixed to the active site through hy-drogen bonds with aspartate residues 169 and 304 [13]. Allof these amino acids are conserved in the chicken sequence.

A

B

Fig. 1. Analysis of time course of expression of an estrogen-responsive gene(4A3) with Northern blot. Chicks were injected with estradiol. (A) mRNAwas isolated from the liver at various time points (0–72 h) post estradiolinjection. The expression of the 4A3 gene was determined by Northern blotas described in Materials and methods. After hybridization with a probe for4A3, the same blot was stripped and re-hybridized with a cDNA probe forpreproalbumin. (B) Expression of 4A3 (data in panel A) was quantified witha PhosphorImager and the relative amount of expression was normalizedwith the level of preproalbumin. The result was the mean ± S.E.M. of 4replicates.

142

This gene is expressed primarily in the liver and kidney,and a very low level is detected in the testis (Fig. 5), whichis consistent with previous reports of AT expression in thechicken [14]. Based on these data, we conclude that this clonerepresents the chicken L-arginine:glycine amidinotransferasegene (EC 2.1.4.1). The sequence has been submitted to Gen-Bank and assigned accession number AF237950.

L-arginine:glycine amidinotransferase is a direct responseto estrogen

To determine whether the observed increase in RNA is a di-rect effect of estrogen, protein synthesis was inhibited byCHX during the stimulation by estradiol. The chickens weresacrificed and poly A+ RNA was isolated as described inMaterials and methods. Northern blot analysis revealedthat CHX did not affect expression pattern of AT follow-ing estrogen exposure (comparing Fig. 1 and Fig. 6), demon-strating that protein synthesis is not required for this increasein AT mRNA.

Discussion

The present study uses a subtractive hybridization methodto identify estrogen-responsive genes in chicken liver. Ourresults show that (1) only the yolk protein genes are isolatedfollowing estrogen exposure by inductive subtractive hy-bridization; (2) estrogen induces a transient up-regulationof L-arginine:glycine amidinotransferase (AT). Its expression

Fig. 2. Southern blot analysis of 4A3. Chicken genomic DNA was iso-lated and digested with different restrictive enzymes. Lane 1: BglII; lane2: HindIII; lane 3: BamHI; lane 4: EcoRI. The position of the molecularweight marker (1 kb ladder) was shown on the right.

is increased 3-fold at 1 h after an injection of estradiol andfollowed by a decline below baseline at 1.5 h.

The regulation of yolk protein genes by estrogen has beenpreviously reported [15]. It is interesting to note that no otherestrogen-induced genes except yolk protein genes were iso-lated following 48 h of estrogen exposure. This result is prob-ably due to the significant up-regulation of yolk protein genefollowing estrogen exposure. After estrogen treatment, themRNAs for these yolk proteins are dramatically increasedand constitute as much as half of the total mRNA in the hepa-tocyte. The proportion of non-yolk protein genes is smallerin estrogen-treated samples than in control samples. There-fore, the genes with modest up-regulation by estrogen willnot be identified by inductive subtractive hybridization, but

Fig. 3. The sequence of the chicken AT gene. The nucleotide sequence isnumbered on the left. The deduced amino acid sequence is shown belowthe nucleotide sequence and numbered on the right. The asterisk indicatesthe stop codon.

atcI

143

Fig. 4. The alignment of amino acid sequences of chicken AT with other species (human, pig, rat and frog).

rather be removed during subtraction process, because therelative amount of non-yolk genes in the Driver (control sam-ples) is much more than in Tester (estrogen-treated samples).So, only those genes with dramatic up-regulation will be re-tained.

Our results show that AT is dramatically affected by es-trogen. The effect of estrogen is bimodal. There is an initialup-regulation followed by a down-regulation after long-termestrogen exposure. The estrogen-mediated AT transcriptionis a fast process (the increase peaks in 1 h) and the expres-

144

sion pattern is not dependent on protein synthesis. This re-sult suggests that AT is a direct target of the estrogen receptor.

AT is the enzyme that catalyzes the transfer of an amidinogroup from L-arginine to glycine. The resultant guanidoaceticacid is the immediate precursor of creatine [16]. Creatine andits phosphorylated form play an important role in the energymetabolism and act as a dynamic reservoir of high-energyphosphate, which buffers the rapid fluctuation of the ATP/ADP ratio [13, 17]. In mammals, creatine is synthesized pri-marily in the liver, pancreas, and kidney, transported bybloodstream and up by Na+-dependent creatine transportersinto muscle and neural tissues [18, 19]. AT exhibits broad sub-strate specificity. In addition to the physiological substrates,AT can utilize a wide variety of amidine donors and accep-tors for the transamidination reaction [17]. Currently, twoforms of AT are presented, dependent on their localization;one is located in the cytoplasm and another is in the inter-membrane space of the mitochondria [13].

AT is regulated in a variety of ways, including productinhibition of its catalytic activity by ornithine [20], end prod-uct repression of its synthesis at a pretranslational level bycreatine, induction by growth hormone and thyroxine [21],and repression during embryogenic development [22, 23]. Aprevious study shows that sex hormones can regulate theactivity of AT [24]. In that study, treatment of male rats withtestosterone propionate increases AT activity. In contrast,estrogen treatment decreases AT activity and induces weightloss. So, this is the first report showing that estrogen canregulate the expression of AT transcript. However, it is cur-rently unclear whether the changes in the level of AT tran-script results from altered mRNA stability or enhancedtranscriptional rate. If estrogen-mediated alteration resultsfrom transcriptional regulation, the site of estrogen action isyet to be determined. It is noted that the short upstream se-quence of chick AT gene isolated here fails to display thetypical Estrogen Response Element (ERE, 5′-GGTCAGACT-GACC-3′). On the other hand, upstream sequence of humanAT gene contains a perfect palindromic sequence (5′-GGC-TCGAGCC-3′) at –242 bp that is similar to ERE. It has beenknown that some estrogen-responsive genes may be regulatedby imperfect ERE [25]. Therefore, further investigation ofupstream sequence of chick AT gene is necessary to identifythe potential site of estrogen action. We conclude that AT asan important enzyme in regulation of energy metabolism andgrowth was first identified that can be regulated by estrogentransiently and also suggested that some effects of estrogencan be mediated by this important enzyme.

Acknowledgements

This work was supported by NIH grant HD 26339 and anaward from the Office of Research for Women’s Health.

Fig. 5. Tissue specific expression of the AT gene by Northern Blot. mRNAwas isolated from gizzard, intestine, kidney, oviduct, skeletal muscle andliver of an adult hen, testis from an adult rooster and hybridized with theAT gene fragment as described in Materials and methods . Lane 1: gizzard;lane 2: intestine; lane 3: kidney; lane 4: oviduct; lane 5: skeletal muscle;lane 6: testis; lane 7: liver (no estrogen treatment); lane 8: liver treated withestrogen for 15 h.

A

B

Fig. 6. The effect of CHX on AT gene expression. Liver mRNA was iso-lated from chicken treated with 17beta-estradiol and CHX for various times(0–4 h). The expression of AT gene was analyzed by Northern blot hybridi-zation. (A) Results from a Northern blot. After hybridization with a cDNAprobe for AT, the blot was stripped and re-hybridized with a probe forpreproalbumin. (B) Expression of AT (data in panel A) was quantified witha PhosphorImager and the relative amount of expression was normalizedwith the level of preproalbumin. The result was the mean ± S.E.M. of 3replicates.

145

References

1. Fink G, Sumner BE, Rosie R, Grace O, Quinn JP: Estrogen control ofcentral neurotransmission: Effect on mood, mental state, and memory.Cell Mol Neurobiol 16: 325– 344, 1996

2. Pelzer T, Shamim A, Neyses L: Estrogen effects in the heart. CellBiochem 161: 307–313, 1996

3. Yager JD, Liehr JG: Molecular mechanisms of estrogen carcinogen-esis. Annu Rev Pharmacol Toxicol 36: 203–232, 1996

4. Frost HM: On the estrogen-bone relationship and postmenopausal boneloss: A new model. J Bone Miner Res 14: 1473–1477, 1999

5. Scallet AC: Estrogens: Neuroprotective or neurotoxic? Ann NY AcadSci 890: 121–132, 1999

6. Muramatsu M, Inoue S: Estrogen receptors: How do they control re-productive and nonreproductive functions? Biochem Biophys ResCommun 270: 1–10, 2000

7. Evans MI, Silva R, Burch JB: Isolation of chicken vitellogenin I andIII cDNAs and the developmental regulation of five estrogen-respon-sive genes in the embryonic liver. Genes Dev 2: 116–124, 1988

8. Duguid JR, Dinauer MC: Library subtraction of in vitro cDNA librar-ies to identify differentially expressioned genes in scrapie infection.Nucleic Acids Res 18: 2786–2792, 1990

9. Sambrook IFE, Maniatis T: Molecular Cloning. A Laboratory Manual, 2ndedn. Cold Spring Harbor Laboratory Press, New York, 1989, pp 7–16

10. Diatchenko L, Lau YF, Campbell AP, Chenchik A, Moqadam F, HuangB, Lukyanov S, Lukyanov K, Gurskaya N, Sverdlov ED, Siebert PD:Suppression subtractive hybridization: A method for generating differ-entially regulated or tissue-specific cDNA probes and libraries. ProcNatl Acad Sci USA 93: 6025–6030, 1996

11. Kozak M: Recognition of AUG and alternative initiator codons isaugmented by G in position +4 but is not generally affected by thenucleotides in positions +5 and +6. EMBO J 16: 2482–2492, 1997

12. Humm A, Huber R, Mann K: The amino acid sequences of human andpig L-arginine:glycine amidinotransferase. FEBS Lett 339: 101–107, 1994

13. Humm A, Fritsche E, Steinbacher S, Huber R: Crystal structure andmechanism of human L-arginine:glycine amidinotransferase: A mito-chondrial enzyme involved in creatine biosynthesis. EMBO J 16: 3373–3385, 1997

14. Grazi E, Magri E, Balboni G: On the control of arginine metabolismin chicken kidney and liver. Eur J Biochem 60: 431–436, 1975

15. Berkowitz EA, Evans MI: Functional analysis of regulatory regionsupstream and in the first intron of the estrogen-responsive chickenvery low density apolipoprotein II gene. J Biol Chem 267: 7134–7138, 1992

16. Walker JB, Skorvaga M: Streptomycin biosynthesis and metabolismphosphate transfer from dihydrostreptomycin 6-phosphate to inosam-ines, streptamine, and 2-deoxystreptamine. J Biol Chem 248: 2441–2446, 1973

17. Walker JB: Creatine: Biosynthesis, regulation, and function. AdvEnzymol Relat Areas Mol Biol 50: 177–242, 1979

18. Loike JD, Zalutsky DL, Kaback E, Miranda AF, Silverstein SC: Ex-tracellular creatine regulates creatine transport in rat and human mus-cle cells. Proc Natl Acad Sci USA 85: 807–811, 1988

19. Guimbal C, Kilimann MW: A Na(+)-dependent creatine transporter inrabbit brain, muscle, heart, and kidney. cDNA cloning and functionalexpression. J Biol Chem 268: 8418–84 21, 1993

20. Sipila I: Inhibition of arginine-glycine amidinotransferase by ornithine.A possible mechanism for the muscular and chorioretinal atrophiesin gyrate atrophy of the choroid and retina with hyperornithinemia.Biochim Biophys Acta 613: 79–84, 1980

21. McGuire DM, Tormanen CD, Seagall S, Van Pilsum JF: The effect ofgrowth hormone and thyroxine on the amount of L-arginine:glycineamidinotransferase in kidneys of hypophysectomized rats. Purificationand some properties of rat kidney transamidinase. J Biol Chem 255:1152–1159, 1980

22. Tormanen CD, Sutter BE: Changes in kidney transaminase activityduring development in male and female rats. Biosci Rep 5: 309–314,1985

23. Van Pilsum JF, Ungar F: Effect of castration and steroid sex hormoneson rat kidney transamidinase. Arch Biochem Biophys 124: 372–379,1968

24. Krisko I, Walker JB: Influence of sex hormones on amidinotransferaselevels. Metabolic control of creatine biosynthesis. Acta Endocrinol(Copenh) 53: 655–662, 1966

25. Jan W, Jacques NJP, Geert AB: Tissue-specific and steroid-depend-ent interaction of transcription factors with the oestrogen-inducibleapoVLDL II promoter in vivo . EMBO J 7: 2757–2763, 1988

146