factor (IRF) in Saccharomyces cerevisiae - BioMedSearch

. 1993 Oxford University Press

5316-5322 Nucleic Acids Research, 1993, Vol. 21, No. 23

Translational repression by the human iron-regulatory factor (IRF) in Saccharomyces cerevisiae Carla C.Oliveira, Britta Goossen1, Nilson l.T.Zanchin, John E.G.McCarthy*, Matthias W.Hentze1 and Renata Stripecke1 Department of Gene Expression, Gesellschaft fOr Biotechnologische Forschung mbH (GBF), Mascheroder Weg 1, 38124 Braunschweig and 'European Molecular Biology Laboratory (EMBL), Gene Expression Programme, Meyerhofstrasse 1, 69117 Heidelberg, Germany Received September 23, 1993; Revised and Accepted October 25, 1993

ABSTRACT The regulation of the synthesis of ferritin and erythroid 5-aminolevulinate synthase in mammalian cells is mediated by the interaction of the Iron regulatory factor (IRF) with a specific recognition site, the iron responsive element (IRE), In the 5' untransiated regions (UTRs) of the respective mRNAs. A new modular expression system was designed to allow reconstruction of this regulatory system in Saccharomyces cerevlslae. This comprised two components: a constitutively expressed reporter gene (luc; encoding luciferase) preceded by a 5' UTR including an IRE sequence, and an inducibly expressed cDNA encoding human IRF. Induction of the latter led to the in vivo synthesis of IRF, which in turn showed IRE-binding activity and also repressed translation of the luc mRNA bearing an IRE-containing 5' UTR. The upper stem-loop region of an IRE, with no further IRE-specific flanking sequences, sufficed for recognition and repression by IRF. Translational regulation of IRE-bearing mRNAs could also be demonstrated In cell-free yeast extracts. This work defines a minimal system for IRF/IRE translational regulation in yeast that requires no additional mammalian-specific components, thus providing direct proof that IRF functions as a translational repressor In vivo. It should be a useful tool as the basis for more detailed studies of eukaryotic translational regulation. INTRODUCTION mRNA is an important site for the regulation of gene expression in prokaryotic and eukaryotic organisms (1, 2). In particular, studies of Escherichia coli have revealed the existence of numerous cases of regulation involving protein-mRNA interactions, whereby the initiation of the translation process is generally the step acted upon (2). It is therefore all the more remarkable that only one regulatory system involving a specific translational repressor protein has as yet been well characterized *

To whom correspondence should be addressed

in eukaryotes. This system is responsible for the regulation of the uptake, intracellular storage and utilization of iron in mammalian cells (3-5). The focal point of regulation is the recognition of a cis-acting, palindromic mRNA motif, the ironresponsive element (IRE), by a trans-acting cytoplasmic binding protein, the iron-regulatory factor [IRF; alternatively IRE-binding protein (IRE-BP)]. Single IREs are present in the 5' untranslated regions (UTRs) of the mRNAs encoding ferritin and erythroid 5-aminolevulinic acid synimase (eALAS) in a range of organisms, and the binding of IRF to them blocks translation (6-9). IRF also binds to multiple IREs in the 3' UTR of the mRNA encoding transferrin receptor (TfR), thus apparently protting this mRNA against degradation (10, 11). Thus the consequences of the binding of IRF to IRE sites are on the one hand the repression of ferritin translation, and on the other, the stabilization of TfR mRNA. A number of properties of IRF distinguish it from other known RNA-binding proteins. It does not contain any of the previously identified RNA-binding motifs. It does however, show striking sequence homology to the aconitases of mitochondria and of E. coli (12,13), and purified human IRF has been shown itself to possess aconitase activity (14). Indeed, IRF may belong to a family of (4Fe-4S) cluster proteins. The affinity of IRF for IREs is high (Kd = 10-10 to 10-11 M) at low iron concentrations, but is reduced 50- to 100-fold upon the addition of iron salts. There is a reciprocal relationship between the activation of the IRE binding site and of the aconitase activity. It has been suggested that the iron-dependent regulation of the equilibrium between the two states (IRE affinity high/aconitase activity low -~ IRE affinity low/aconitase activity high) is linked to switching between the [4Fe-4S] and apoprotein forms of IRF (4, 5). The molecular basis of the mode of action and regulation of IRF remains unknown. An important strategy for the investigation of this as yet unparalleled regulatory system is to isolate individual components from their normal environment and study their behaviour in a heterologous system where no equivalent mechanism is normally active. We have chosen to use the yeast

Nucleic Acids Research, 1993, Vol. 21, No. 23 5317 A H/Xb

XIS

A

H

H

Ii.;,0-,. .. ,00-i:ll .I............. GALPGK FUSION PROMOTER

CEN4 ARS1 URA3

PGK1 TERM

IRF

B

7N

B

C

Xb

H

CEN4 ARS1 TRP1 A

TEF 1 PROMOTER

PGKI TERM

LUC

C Bg/B

Xtb

N

Nsi

Im LEADER

LUC/CAT

POLY - A

D

STABIUTY

KCaVMd 3s AATTATCTACTTAAGAACACAAAACTCGAGAACAT -

FL:

IREWT:

-s58

b.

.4

AATTATCTACTTAAGCTTCAACAGTGCTTGAACTTAAGAACACAAAACTCGAGAACAT ATA7 .A

I REMUT:

AATTATCTACTTAAGCTTCAAAGTGCTTGAACTTAAGAACACAAAACTCGAGAACAT

(8)

-0 .9

(Aj) (@i

-7 .3 -7

.5

Figure 1. Physical map of the plasmids used. (A) YCpSUP-IRFl. The human IRF cDNA was cloned using the XJzoI and Xbal sites of the plasmid YCpSUPEXl (17) by fusing these restriction sites with SmaI and HindIII (respectively), present in the pMS-56-hIRF fragment (20). The IRF gene was under control of the GAL-PGK fusion promoter. URA3 was the selection marker for this plasmid. (B) YCp22FLl plasmid. The modified TEFI promoter (18) was cloned into YCplac22 (19) using the EcoRI and BamHI sites. The SUPEXI leader sequence, the luciferase gene and the PGK terminator sequence were excised from YCpLUCEXl (17) and cloned using BamHI and HindIII sites. Oligonucleotides IREWT and IREMUT were cloned into an AflII site present in the leader sequence, as indicated by the stem-loop structure. TRP1 was the marker for this plasmid. (C) Physical map of the basic constructs for synthesis of mRNA in vitro. It contains the T7 RNA polymerase promoter, the leader sequence, the reporter gene and a 30 nucleotides long poly-A tail. (D) Sequences of the mRNA leaders upstream of the luc gene (compare B and C). The stabilities of the secondary structures are indicated on the right-hand-side. Leader lengths are also indicated, where A of the ATG start codon is nucleotide + 1. In the in vitro system the mRNA leaders were six bases longer than those of the in vivo mRNAs because transcription starts at the first G of the BamHI/BglII fusion. A = AflII, B = BamHI, Bg = BglII, C = Ciad, E = EcoRI, H = HindIII, N = NdeI, Nsi = NsiI, S = SmaI, X = XhoI, Xb = XbaI.

Saccharomzyces cerevisiae as the host for investigations of the interactions between human IRF and IREs inserted into the 5' UTR of a reporter gene. In contrast to all mammalian cells analyzed so far, it has been shown previously that S. cerevisiae does not normally possess an IRE-binding activity (15,16). Using a newly conceived two-plasmid in vivo expression system we have been able to reproduce faithfully the translational repression function of IRF in this genetically versatile lower eukaryote. MATERIALS AND METHODS Bacterial and yeast strains Saccharomzyces cerevisiae: RS453a (ade2-1 trpl-l leu2-3 his3-1 1 ura3-52 lys+ canl-100) [kindly provided by Dr. Ramon Serrano, Valencia]

SKQ2n (17) Escherichia coli: TG2 (SupE hsdA5 thiA(lac-proAB)A(srlrecA)306: :TnlO(tetr) F'(traD36 proAB + lacIq lacZ M15) Yeast was cultured and transformed according to methods described previously (18).

Plasmid construction YCpSUP-IRFl: The human IRF cDNA (19) was cloned as a 2.9 kb Smal-HindIII fragment from pMS-56-hIRF (20) into YCpSUPEX1 (21) digested withXoI andXbaI. The DNAs were cleaved using the respecti-ve enzymes and treated with Klenow prior to ligation. YCp22FL1: The TEF1 modified promoter (22) was cloned into YCplac22 (23) as an EcoRI-BamHI fragment. After digestion with EcoRI, the DNA of both the vector and the TEFl fragment was treated with the Klenow fragment of DNA polymerase I, and then digested with BamHI. After the ligation, the YCplac22-TEF1 construct had no EcoR! site. This plasmid was digested with BamHI and HindIll and a fragment bearing the SUPEXI leader sequence, the LUC gene, and PGK terminator [excised from YCpLUCEX1 (21)], was cloned into it, generating YCp22FL1. The IREWT and IREMUT oligonucleotides were inserted into YCp22FL1, after digestion with AflII, generating YCp22FL-IREWT and YCp22FLIREMUT. The IREMUT oligonucleotide differs from the wild type by a deletion of one C in the loop of the secondary structure formed in the mRNA leader sequence (6, 24). The in vitro transcription vector was pHST7 [derived from pHSTO (25)]. A 30 nucleotide long poly-A sequence was inserted

5318 Nucleic Acids Research, 1993, Vol. 21, No. 23 ~. XL L.

IREWT

iL

c:

-

" :"'

qj., ,s::;s;s

%:.

4:4,

-t

.

.*

*mustsAA

FIgure 3. Gel retardtion assay. Cell-free extracts were prepared from cells bearing the plasid YCpSUP-IRFl and either YCp22FL-IREWT or YCp22FL-IREMUT and were incubated with a radiolabelled IRE RNA probe. Extracts were prepared from cells growing in lactate (LACT) or galactose (GAL) medium. Two clones of each transformation were used. B.P. = bound protein = IRE/IRF complex; F.P. = free IRE probe. Experiments were perforned either in the presence (+) or the absence (-) of 2% 2-mercaptoethanol.

HnE zuciteiase? t'qF-

Flgure 2. Primer extension analysis of total RNA. The 5' ends of the luc mRNAs were mapped using MMLV reverse transcriptase. YCp22FL-IREWT DNA was sequenced and is shown as a reference which can be used to determine the positions of the primer extension terminations. The full extensions are equivalent to the expected lengths of the respective leaders. These are double bands which correspond to the two As downstrem of the Bamr!1 site (see Fig. IB,D). Weaker, slighdy shorter bands correspond to the additional transcriptional start sites of the TEF) promoter (21). The results shown here were obtained using extracts from cells after induction with galactose (as in Fig.7). The same results were obtained using RNA from cells that had not been induced (data not shown).

between

the XbaI and EcoRl sites of this vector. This polyA cassette contained an NsiI site immediately downstream of the

polyA

sequence. The luciferase gene together with the leader sequences was removed from YCp22FL1 and its derivatives

YCp22FL-IREWT and IREMUT respectively, and inserted into the BglII and XbaI sites upstream of the poly-A sequence to obtain the constructs LUCEX, LUCIREWT, and LUCIREMUT. Then the LUC gene of these three constructs was substituted by the CAT gene, using the NdeI and XbaI restriction sites to obtain

CATEX, CATIREWT, and CATIREMUT. The constructs containing a truncated LUC gene (LUC') were generated by digesting YCp22FLl with BamHI and EcoRI and inserting these fragments into the BglII and EcoRl sites (in the luciferase gene, the EcoRl is located at 599 nt from the AUG start codon) of the vector upstream of the poly-A. The ORF of the truncated constructs ends at a stop codon present upstream of the poly-A sequence. All the in vitro expression constructs are under the control of the T7 RNA polymerase promoter. In vitr transcription and translation Templates for in vitro transcription were prepared by digesting the constructs with NsiI. In vitro transcription and cotrnscriptional capping were performed as described by Nielsen & Shapiro (26). Yeast cell-free extracts were isolated and

HE-iLuciterase/hr"frw

4-*

ijb I V:

I

--

Figure 4. Time course of RNA binding activity of IRF. This figure depicts the ratio of luciferase activity measured at the different time points from cells growing in galactose (+GAL) to the activity detected at corresponding time points using lactate (-GAL) medium. The same cell-free extracts were used for gel retdation assays of the formation of the IRE/IRF complex (see top of panels, and compare Fig. 3). The two panels compare the results obtained using the IREWT and IREMUT structures, respectively, in the luc leader (Fig. IB).

translation assays were performed as described previously (17, 27). Recombinant IREF was prepared as described previously (20). Briefly, the in vitro synthesized mRNAs (500 ng per reaction) were incubated with IRF in a volume of 6 gl for 10 minutes in an ice bath and then 19 $l of yeast translational extract were added, after which the reaction mixes were incubated at 20°C

Nucleic Acids Research, 1993, Vol. 21, No. 23 5319

A

IREGIREMUT YCp L L -G L -G L -G G~

YCp FL Rd LucA 120110 1009080-

Mslum * LAT * GAL

LUC

708050 -

403020100-

B Leader

LUC Activity

Figure 6. Western blots of cell extracts subjected to SDS-PAGE. Cell extracts were prepared from cells bearing YCpSUP-IRF1 together with one of the plasmids YCp22FLI, YCp22FL-IREWT or YCp22FL-IREMUT. The outer lanes show the results obtained with an extract from cells carrying only YCplac22 (YCp; as a negative control). Cells were grown either in lactate medium (L) or for 9 hours in the induced state (G). Western blotting was performed using rabbit antiluciferase serum.

(GAL/LACr ± SD)

FL

2.11 ± 0.32

IREWT

0.20 ± 0.05

IREMUT

1.48 ± 0.18

Figure 5. A. Graphical presentation of the inhibitory effects of expression of the IRF cDNA on the activity of luciferase in double transformants. The values represent the luciferase activity measured in extracts of cells in lactate medium (LACT) and after 9 hours of induction of IRF gene expression by growing the cells in galactose medium (GAL). The activities corresponding to FL-IREWT and FL-IREMUT were normaized to that of ofthe FL construc (set to the arbitrary value of 100). B. Averages of, for each leader, six independent measurements of luciferase activity of the respective constructs 9 hours after induction of IRF synthesis using galactose (GAL), expressed as ratios to the activities of the corresponding samples grown widtout induction (LACT). The standard deviations are given on the right-hand-side.

for 60 minutes and analysed by SDS-PAGE and fluorography tested for luciferase or CAT activity.

for preparing cell extracts. Total proteins were quantified by the BCA method (Sigma) and 1 ,ug total cellular protein was used for each enzyme assay. The CAT assay (21) was performed using samples removed from the in vitro translation reactions. Western blotting For the purpose of Western blotting (performed as in 21) 10 jig of total protein were loaded on an SDS-polyacrylamide gel, submitted to electrophoresis, and blotted onto an Immobilon (PVDF) membrane. Rabbit anti-luciferase antibody was kindly provided by Dr. Hansjorg Hauser, GBF. RNA gel retardation assay Cell-free extracts of yeast transformants bearing the IRF and luciferase constructs were incubated together with a radiolabelled IRE RNA probe (31) at 25°C for 30 min. Reactions were then submitted to electrophoresis on non-denaturing polyacrylamide gels (see 31). Where indicated, extracts were pretreated with 2% 2-mercaptoethanol prior to the addition of the RNA probe.

or

RNA isolation, Northern blotting and primer extension Total yeast RNA was isolated (28) from 50 ml of each yeast culture used for luciferase assays. 10 ,tg of total RNA were glyoxylated and subjected to electrophoresis through a 1.3% agarose gel. The RNA was blotted onto nylon membranes (Hybond-N, Amersham) and cross-linked by UV irradiation. Hybridization (29) was performed using DNA fragments derived from the luciferase and actin (as an internal control) genes labelled by random-priming, using the Prime-a-genetm' labelling system (Promega). Mapping of the 5'-ends of mRNAs was carried out by means of primer extension (30) using an oligodeoxyribonucleotide complementary to the 5'-end of the LUC coding region (position +46 to +69, relative to the A of the start codon). Luciferase and CAT assays The luciferase assay was performed as described elsewhere (21). Fractions corresponding to 2 ml of yeast cell culture were used

RESULTS AND DISCUSSION A modular expression system for studying translational regulation in yeast We constructed a new modular expression system to enable us to study translational regulation of mRNAs bearing IRE stemloops in their leaders. Two types of expression plasmid were constructed: (1) YCpSUP-IRFl was constructed by insertion of the IRF cDNA sequence (19) into YCpSUPEX1 (21) (Fig. IA). (2) The reporter gene (encoding luciferase) was inserted into the plasmid YCplac22-TEFl, yielding YCp22FL1 (Fig. 1B). Synthetic DNA fragments bearing either the wild-type (IREWT) or a mutant form (IREMUT) of a (consensus) IRE sequence (6,8,24) were inserted into the leader region of YCp22FL1 at the AflIl site (Fig. 1D). The synthetic DNA sequences bore only the upper stem-loop region of the IRE elements observed in a range of eukaryotes (32,33). They were predicted to be capable of forming stem-loop structures of stability less than -8 kcal

5320 Nucleic Acids Research, 1993, Vol. 21, No. 23

,

Alik

0

w W

...W.

6

::

::::.

t

40

Am

4Ia ab 40

4m do qw".

B

Figure 7. Total RNA analysis. (A) Northern blot showing the relative abundance of luc mRNAs in the transformed strains in the absence of IRF induction (LACT) and after 9 hours induction of IRF gene expression in galactose medium (GAL). 32P-labelled DNA fragments specific to the luciferase (LUC) and actin (ACT) genes were used as probes. (B) Table showing the concentration of luc mRNA relative to actin mRNA in LACT- or GAL-medium. After hybridization, the bands were excised and their radioactivity determined by means of scintillation counting.

mol-I and thus to exert in themselves only a limited effect on translational initiation efficiency (compare 21,22). The deletion of a cytosine residue in the loop of IREWT (yielding here IREMUT) is known to reduce drastically the binding affinity of IRF. The first C in the IRE consensus sequence was found to lie 15 nucleotides from the 5' end of the mRNA encoded by the IREWT construct (compare Figs. 1 and 2). This is similar to the distances observed in a range of mRNAs encoding ferritin or erythroid 5-aminolevulinic acid synthase in other organisms (32,33), and was expected to allow effective binding of IRF to the mRNA leader. Cotransformation of yeast cells with YCpSUPIRFI together with either the control construct YCp22FLl or one of its derivatives (IREWT/IREMUT) yielded the required in vivo test system.

Synthesis of active IRF in yeast Transformants carrying the IRF expression plasmid YCpSUPIRF 1 were found capable of synthesizing active IRF upon induction with galactose. Gel retardation studies revealed that a protein activity that binds a small in vitro-synthesized RNA fragment bearing an IRE sequence (compare 31) was present in extracts from yeast cells in which IRF synthesis had been induced (Fig. 3). It has been reported previously (31,34,35) that the RNA (IRE) binding activity of IRF in extracts from higher cells is maximized by the addition of 2-mercaptoethanol. This effect was not observed in the experiments described here, possibly because IRF was largely expressed as an apoprotein in the yeast cells, at least under the conditions described here. However, further experimental work will be necessary to clarify this issue.

Figure 8. In vitro synthesis of CAT using yeast cell-free translation extracts in the presence and absence of IRF. The amount of IRF used in each reaction is given below each lane in the fluorograph (A) or on the right-hand-side of the graphical representation of the CAT enzymic activity data (B). Panel A shows a fluorograph prepared after SDS-PAGE of samples from the in vitro translation reactions. Samples were taken from these reactions for analysis of CAT activity (B). The relative activities are expressed in terms of the acetylation of 14Cchloramphenicol.

In double transformants carrying both YCpSUP-IRF 1 and YCp22FL-IREWT, the luciferase activity encoded by the luc gene-IREWT combination was progressively reduced with time subsequent to induction of IRF gene expression (Fig. 4). The maximum loss of luciferase activity was observed approximately 9 hours after induction, which was also the time-point where the maximal IRE-binding activity could be measured. In parallel experiments using the IREMUT derivative of YCp22FLl (Fig. 4), induction of IRF synthesis resulted in the same development of IRE-binding activity, but this was not accompanied by a progressive loss of luciferase activity in the cells. Fig. 5 shows a direct comparison of the effects of IRF synthesis (after 9 hours of induction) upon the luciferase activities encoded by mRNAs bearing the IREWT sequence or either of the two other control leaders. The observed changes in luciferase activity were also reflected in the amounts of detectable protein reacting with a polyclonal antibody preparation specific for luciferase (Fig. 6).

Nucleic Acids Research, 1993, Vol. 21, No. 23 5321 This indicates that the accumulation of luciferase directed by the IREWT mRNA in yeast cells is inhibited by the presence of de novo synthesized IRF. Inhibition of luciferase activity involves translational repression In order to be able to draw conclusions about changes in the translational activity of the reporter gene mRNAs, we determined their relative amounts as a function of the conditions of growth of the yeast strains. Northern blot analysis of yeast cell extracts revealed that the amounts of the luc mRNAs transcribed from the TEFI promoter increased under conditions of galactose induction (Fig. 7). Thus while luciferase activity was reduced approximately five times nine hours after the transition from lactate to galactose in a double transformant bearing YCpSUPIRFI/YCp22FL-IREWT (Fig. 5), the amount of luciferasespecific mRNA (normalized to the amount of actin mRNA) increased by a factor three (Fig. 7). Given that the reduction in luciferase activity correlates with a reduction in the amount of luciferase protein in the cell (Fig. 6), we interpret these data to mean that the effective level of synthesis of luciferase directed by the (IREWT) luc mRNA was attenuated by more than 90% relative to its maximum level. That IRF inhibits translation of the IREWT luc mRNA could be directly confirmed by examining the effects of adding purified recombinant IRF to yeast cell-free extracts translating in vitro synthesized mRNAs (Fig. 8). The translation of a cat mRNA bearing the IREWT leader was strongly inhibited (by at least 90%) in the presence of IRF, whereas the translation of the corresponding control mRNAs (IREMUT and CATEX) was insignificantly affected under the same conditions. Similar results were obtained using a reporter mRNA bearing either the complete, or a truncated form of the luc gene downstream of the IREWT leader (data not shown). Overall, the described experimental data lead us to the conclusion that the regulation of reporter gene activity induced by IRF in yeast is primarily attributable to translational repression. The minimal IRE sequence used here allows IRF binding of sufficiently high affinity to ensure effective inhibition of translation. At the same time, the IREWT and IREMUT stemloop structures do not in themselves strongly inhibit translation of the reporter mRNAs (Fig. 5). We have not as yet investigated other consequences of the IRF/IRE interaction. For example, comparison of the relative amounts of reporter mRNAs in the various double transformants (Fig. 7) indicates that IRF binding may increase mRNA abundance, perhaps by stabilizing it against degradative processes.

CONCLUSIONS The present work describes the reconstruction of a translational regulatory circuit derived from mammalian cells in the yeast S. cerevisiae. The expression system allows constitutive synthesis of the 'target' IRE-bearing mRNA and inducible expression of the IRF cDNA within one cell using transcriptional and translational signals tailored for use in yeast. We have shown that the IRF synthesized in the yeast cell binds an IRE target and reduces the translation of a reporter mRNA whose reading frame is preceded by a leader bearing an IRE sequence. The inhibitory effects are specific to interactions between IRF and a bona fide IRE sequence, since a one-base deletion in the loop region of the latter eliminates regulation imposed by the repressor protein. Overall, our data indicate that IRF can fulfil the same

regulatory function in yeast as it does in the mammalian cell. The finding that IRF is the only additional component required in the yeast cell to allow repression of an mRNA bearing an IRE in its 5' UTR constitutes direct proof of its function as a translational repressor in vivo. IRF/IRE interactions have previously been studied in higher eukaryotic systems either by cloning the ferritin leader upstream of a reporter gene (7, 36, 37) or by introducing synthetic IRE sequences (6, 24, 32, 38). However, these studies cannot exclude the role of additional endogenous components potentially involved in the iron regulatory system. In contrast, the present work reduces the system to a minimal set of defined components. In fact, our results show that the translational repressor activity of IRF cannot be absolutely dependent on any specific additional factors that are present in mammalian (but not in yeast) cells. The minimal IRE sequence used here in yeast suffices to allow strong translational regulation despite the fact that it is lacking flanking sequences proposed elsewhere to play a role in translational regulation by IRF (33). At the present stage we can only speculate as to the mechanism of the translational repression caused by IRF binding. However, it is likely that, as in a number of prokaryotic systems (2), the binding of the repressor blocks translational initiation. In eukaryotes, initiation at the start codon is believed to be preceded by binding of the 43S initiation complex at or near the 5' cap structure, followed by 'scanning' of this complex along the leader sequence (39). Any of the component steps of this pathway might theoretically be blocked by the IRF-IRE interaction (5). Finally, the IRF/IRE regulatory system established here in yeast promises to be a useful tool in future studies of the mechanism of action of IRF.

ACKNOWLEDGEMENTS We thank Dr. Hansjorg Hauser, GBF (Braunschweig) for the gift of rabbit anti-luciferase serum, Dr. Lukas Kuhn (Epalinges, Switzerland) for a full-length human IRF cDNA, and Yves Henry for helpful technical advice. The Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPq) of Brazil provided grants to C.C.O., N.I.T.Z. and R.S.

REFERENCES 1. Lindahl, L. and Hinnebusch, A. (1992) Curr. Opinion Genet. Devel. 2, 720-726. 2. McCarthy, J.E.G. and Gualerzi, C. (1990) Trends Genet. 6, 78-85. 3. Leibold, E.A. and Guo, B. (1992) Annu. Rev. Nutr. 12, 345-368. 4. Klausner, R.D., Rouault, T.A. and Harford, J.B. (1993) Cell 72, 19-28. 5. Melefors, 0. and Hentze, M.W. (1993) BioEssays 15, 85-90. 6. Hentze, M.W., Caughman, W., Rouault, T.A., Barriocanal, J.G., Dancis, A., Harford, J.B. and Klausner, R.D. (1987) Science 238, 1570-1573. 7. Aziz, N. and Munro, H.N. (1987) Proc. Natl. Acad. Sci. USA 84, 8478-8482. 8. Dandekar, T., Stripecke, R., Gray, N.K. Goossen, B., Constable, A., Johansson, H.E. and Hentze, M.W. (1991) EMBO J. 10, 1903-1909. 9. Melefors, O., Goossen, B., Johansson, H.E., Stripecke, R., Gray, N.K. and Hentze, M.W. (1993) J. Biol. Chem. 268, 5974-5978. 10. Milliner, E.W. and Kuhn, L.C. (1988) Cell 53, 815-825. 11. Casey, J.L., Koeller, D.M., Raniin, V.C., Klausner, R.D. and Harford, J.B. (1989) EMBO J. 8, 3693-3699. 12. Rouault, T.A., Stout, C.D., Kaptain, S., Harford, J.B. and Klausner, R.D. (1991) Cell 64, 881-883. 13. Hentze, M.W. and Argos, P. (1991) Nucl. Acids Res. 19, 1739-1740. 14. Kennedy, M.C., Mende-Mueller, L., Blondin, G.A. and Beinert, H. (1992) Proc. Natl. Acad. Sci. USA 89, 11730-11739. 15. Paino, M.M. and Walden, W.E. (1992) J. Biol. Chem. 267, 19011-19016.

5322 Nucleic Acids Research, 1993, Vol. 21, No. 23 16. Rothenberger, S., Muillner, E.W. and Kuhn, L.C. (1990) Nucl. Acids Res. 18, 1175-1179. 17. Tuite, M.F. and Plesset, J. (1986) Yeast 2, 35-52. 18. Guthrie, C. and Fink, G.R. (eds.) (1991) Methods in Enzymology 194. 19. Hirling, H., Emery-Goodman, A., Thompson, N., Neupert, B., Seiser, C. and Kuihn, L.C. (1992) Nucl. Acids Res. 20, 33-39. 20. Gray, N.K., Quick, S., Goossen, B., Constable, A., Hirling, H., Kuihn, L.C. and Hentze, M.W. (1993) Eur. J. Biochem., in press 21. Oliveira, C.C., van den Heuvel, J.J. and McCarthy, J.E.G. (1993) Molec. Microbiol. 9, 521-532. 22. Vega Laso, M.R., Zhu, D., Saglioco, F., Brown, A.J.P., Tuite, M.F. and McCarthy, J.E.G. (1993) J. Biol. Chem. 268, 6453-6462. 23. Gietz, R.D. and Sugino, A. (1988) Gene 74, 52k7-534. 24. Rouault, T.A., Hentze, M.W., Caughman, S.W., Harford, J.B. and Klausner, R.D. (1988) Science 241, 1207-1210. 25. Jobling, S.A., Cuthbert, C.M., Rogers, S.G., Fraley, R.T. and Gehrke, L. (1988) Nucl. Acids Res. 16, 4483-4498. 26. Nielsen, D.A. and Shapiro, D.J. (1986) Nucl. Acids Res. 14, 5936. 27. Gerstel, B., Tuite, M.F. and McCarthy, J.E.G. (1992) Molec. Microbiol. 6, 2339-2348. 28. K6hrer, K. and Domdey, H. (1991) Methods Enzymol. 194, 398-401. 29. Sambrook, J., Maniatis, T. and Fritsch, E.F. (1989) Molecular Cloning: A laboratory manual. 2nd Edition Cold Spring Harbour, New York: Cold Spring Harbour Press. 30. Leer, R.J., van Raamsdonk-Duin, M.M.C., Hagendoorn, M.J.M., Mager, W.H. and Planta, R.J. (1984) Nucl. Acids Res. 12, 6685-6700. 31. Weiss, G., Goossen, B., Doppler, W., Fuchs, D., Pantopoulos, K., WernerFelmayer, G., Wachter, H. and Hentze, M.W. (1993) EMBO J. 12, 3651-3657. 32. Goossen, B. and Hentze, M.W. (1992) Mol. Cell. Biol. 12, 1959-1966. 33. Dix, D.J., Lin, P.-N., McKenzie, A.R., Walden, W.E. and Theil, E.C. (1993) J. Mol. Biol. 231, 230-240. 34. Hentze, M.W., Rouault, T.A., Harford, J.B. and Klausner, R.D. (1989) Science 244, 357-359. 35. Miillner, E.W., Rothenberger, S., Miller, A.M. and Kuhn, L.C. (1992) Eur. J. Biochem. 208, 597-605. 36. Brown, P.H., Daniels-McQueen, S., Walden, W.E., Patino, M.M., Gaffield, L., Bielser, D. and Thach, R.E. (1989) J. Biol. Chem. 264, 13383-13386. 37. Coccia, E.M., Profita, V., Fiorucci, G., Romeo, G., Affabris, E., Testa, U., Hentze, M.W. and Battistini, A. (1992) Mol. Cell Biol. 12, 3015-3022. 38. Goossen, B., Caughman, S.W., Harford, J.B., Klausner, R.D. and Hentze, M.W. (1990) EMBO J. 9, 4127-4133. 39. Kozak, M. (1989) J.Cell.Biol. 108, 229-241.