ArgRII, a Component of the ArgR-Mcm1 Complex Involved in the Control of Arginine Metabolism in Saccharomyces cerevisiae, Is the Sensor of Arginine

Abstract

Repression of arginine anabolic genes and induction of arginine catabolic genes are mediated by a three-component protein complex, interacting with specific DNA sequences in the presence of arginine. Although ArgRI and Mcm1, two MADS-box proteins, and ArgRII, a zinc cluster protein, contain putative DNA binding domains, alone they are unable to bind the arginine boxes in vitro. Using purified glutathione S-transferase fusion proteins, we demonstrate that ArgRI and ArgRII1-180 or Mcm1 and ArgRII1-180 are able to reconstitute an arginine-dependent binding activity in mobility shift analysis. Binding efficiency is enhanced when the three recombinant proteins are present simultaneously. At physiological concentration, the full-length ArgRII is required to fulfill its functions; however, when ArgRII is overexpressed, the first 180 amino acids are sufficient to interact with ArgRI, Mcm1, and arginine, leading to the formation of an ArgR-Mcm1-DNA complex. Several lines of evidence indicate that ArgRII is the sensor of the effector arginine and that the binding site of arginine would be the region downstream from the zinc cluster, sharing some identity with the arginine binding domain of bacterial arginine repressors.

Yeast ArgRII (Arg81) is one of the four proteins which coordinate the expression of arginine anabolic and catabolic genes in response to arginine. ArgRII is 880 amino acids (aa) long and belongs to the Zn₂Cys₆ binuclear cluster proteins (21). Unlike Gal4 and Ppr1, which bind as dimers to DNA sequences with the palindromic CGG separated by 11 bp for Gal4 and 6 bp for Ppr1, ArgRII does not bind by itself to the arginine boxes. It requires the presence of two other proteins, Mcm1 and ArgRI, belonging to the MADS-box family of transcription factors (6, 19). The target site of these three proteins (called the ArgR-Mcm1 complex) consists of a large DNA region of about 40 to 60 nucleotides containing two arginine boxes homologous to the binding site of Mcm1 (PBox) (1, 7, 20).

Pairwise interactions between ArgRII, ArgRI, and Mcm1 were identified using the two-hybrid system. ArgRI and Mcm1 interact also with ArgRIII, a pleiotropic regulatory factor required for the stability of these two proteins (9). Binding of the ArgR-Mcm1 proteins to DNA requires the presence of arginine, whereas the interactions between ArgRII, ArgRI, and Mcm1 occur in the absence of the effector. Arginine is thus required for the interaction of the complex with the arginine boxes and not for the modulation of the activation ability, as in most of the other systems identified in yeast. Since Mcm1 is a pleiotropic regulator, one of the two specific regulatory proteins of the system, ArgRI or ArgRII, could contain the arginine binding site. Comparison of the amino acids sequence of ArgRII with those of the arginine repressors of Escherichia coli (ArgR) and Bacillus subtilis (AhcR) (18, 26) revealed that two regions of ArgRII, located between aa 89 and 114 and aa 563 and 587, share some identity with the C-terminal domain of the two bacterial repressors (Fig. 1). Different studies showed that this domain of the hexameric E. coli ArgR repressor contains an arginine binding pocket defined in part by two aspartic acid residues and is responsible for oligomerization (4, 25, 29, 30).

FIG. 1.

Amino acid alignment between yeast ArgRII and E. coli ArgR and B. subtilis AhrC. Identical amino acids are shaded, and amino acids contacting arginine in E. coli ArgR are underlined. The first and last residue numbers are indicated.

The present study aimed at defining in the ArgRII protein the regions interacting with DNA, with ArgRI and Mcm1, and also with arginine, the effector. Although the mode of action of many regulatory proteins in response to physiological signals has been extensively studied, the site of action of the small molecule effector in signaling environmental changes has been defined in very few cases. It was shown only recently that the C-terminal end of Ppr1, the activator of the pyrimidine biosynthetic pathway, contains the dihydroorotic acid-responsive domain, which colocalizes with the activation domain. The binding of the effector converts DNA bound Ppr1 from a transcriptionally inactive state to an active one (11).

In this report, we provide evidence that the N-terminal end of ArgRII is sufficient for the formation of the DNA-ArgR-Mcm1 complex in response to arginine.

MATERIALS AND METHODS

Strains and media.

Saccharomyces cerevisiae HY (diploid strain obtained by crossing strains HF7c (MATa ura3-52 his3-200 lys2-801 ade2-101 trp1-901 leu2-3,112 gal4-542 gal80-538 LYS2::GAL1-HIS3 URA3::(GAL4_(17-mers)3-CYC1-lacZ) and Y187 (MATα ura3-52 his3-200 ade2-101 trp1-901 leu2-3,112 met⁻ gal4Δ gal80Δ URA3::GAL1-lacZ; Clontech) (15) was the recipient strain for experiments using the two-hybrid system. These strains lack both GAL4 and GAL80 genes and contain an integrated GAL1-lacZ reporter gene activated by the GAL upstream activation sequence (URA3::GAL1-lacZ) and an integrated GAL1-HIS3 reporter gene.

Strain 02463dII (MATa leu2 ura3 argRII::KanMX4) (9) was used as recipient strain for transformation with plasmids expressing wild-type or mutated ARGRII genes.

All yeast strains were grown on minimal medium containing 3% glucose or 1% galactose, vitamins, and mineral traces (17) (M.glucose or M.galactose, respectively). Nitrogen source was 0.02 M ammonium sulfate. For the two-hybrid experiments, yeast cells were grown in synthetic medium containing 0.7% yeast nitrogen base without amino acids. This medium was supplemented with 2% glucose and all amino acids except those whose omission was required for plasmid selection.

The lithium acetate procedure was used to transform the recipient yeast strains (13).

E. coli XL1B and JM109 (Stratagene) were used for plasmid amplification and in vitro mutagenesis.

Construction of plasmids expressing different N-terminal portions of ArgRII under GAL10 or its own promoter.

To overexpress the first 128 and 180 aa of ArgRII, plasmid pME52 containing the wild-type ARGRII gene was used to synthesize by PCR BamHI-NotI DNA fragments using oligonucleotides RII38 (BamHI)-RII39 (NotI) and RII38-RII58 (NotI), respectively (Table 1). These fragments were inserted in the BamHI and NotI sites of vector pYeF2 (pUC19, 2μm URA3 GAL10 promoter) (5), yielding plasmids pNA44 (GAL10-ArgRII1-128) and pNA53 (GAL10-ArgRII1-180). To fuse the activation domain of Gal4 (GAD) to the first 180 aa of ArgRII, we synthesized by PCR a NotI DNA fragment encoding aa 768 to 881 of Gal4 from plasmid pCL1 (10), using oligonucleotides OAD1 and OAD2. This fragment was inserted into the NotI site of plasmid pNA53, generating in-frame ArgRII1-180-GAD fusion protein (pNA54). All genes were sequenced to ensure that no mutation was introduced during the PCR procedure.

TABLE 1.

Oligonucleotides used for in vitro mutagenesis in ARGRII and CAR1 genes

Oligomer	Length (nucleotides)	Sequence	Amino acid change(s)
RII38	27	CCCGGGGATCCAGATATAATGGGAATT	Creation of a BamHI site at position −7
RII39	30	CCCAAATTTCTTTATGAGGATCCACTAGTT	Creation of a NotI site at position +410
RII58	30	CCGGCGGCCGCTAGTTGAAGAAGATGGTAA	Creation of a NotI site at position +542
RII103	37	GCCGGATCCTCTTCTCTGTACCTCTTTAATG	Creation of a BamHI site at position +901
RII78	21	AAAGTTAAGTTAGATCTTCGG	C31L
RII40	30	CATCCCCACTTACAACGATTAGAAAAGTCT	C38L, C41L
RII61	27	GATCTTCGGCATTTACACTGCCAACGA	P36L
RII79	21	GATGAACCAGCATACCAACGA	Q89A
RII80	21	CGGAACATCGCTTTTGTGCGC	D96A
RII81	27	GTGCGCTATGCTGCAGCATACGTGTAT	D101A, E102A, E103A
RII82	24	GTGTATCATGCAGCTATGGATGAT	E108A, D109A
RII83	24	GAAGATATGGCTGCTGAGCTAACA	D111A, D112A
RII52	30	AAGTTGCAGATAGGATCCGAATTTTCAGCA	Creation of a BamHI site at position +1
RII09	30	GCCGGGATCCTAAGTTGAAGAAGATGGTAA	Creation of a BamHI site at position +542
RII16	28	GCCGCCGGGATCCTTAATCATTGGCACTGGC	Creation of a BamHI site at position +901
RII64	27	GCCGGGATCCCTATGCAATTTGACCCG	Creation of a BamHI site at position +179
RII08	27	GCCGGGATCCAACGACGGAACATCGAT	Creation of a BamHI site at position +272
RII10	30	GCCGGGATCCATATTATACCCAAAACAG	Creation of a BamHI site at position +1139
RII11	33	GCCGGGATCCTATTTTTGTAGTTTGTAGTTCCT	Creation of a BamHI site at position +1412
RII93	27	GCCGGGATCCACAAACTACAAAAATAC	Creation of a BamHI site at position +1397
RII94	27	GCCGGGATCCTAAAATTTTCGGGGTAC	Creation of a BamHI site at position +2146
RII14	27	GCCGGGATCCCTGTACCCCGAAAATTT	Creation of a BamHI site at position +2129
RII15	30	GCCGGGATCCTTATGATAGCATCAGATT	Creation of a BamHI site at position +2641
CAR1-99	39	CAAGCACCGTGTTTCCTAATTAGGAAATCAACAGCGC	Replacement of arginine box B by PPAL
CAR1-100	39	GCGCTGTTGATTTCCTAATTAGGAAACACGGTGCTTTG	Replacement of arginine box B by PPAL
CAR1-DE 08	30	CCGCTCGAGCGCCGCGAAAATATCGGCTAG	Primer to generate CAR1 fragment
CAR1-DE 09	30	CCGCTCGAGTGATAGAAAGTGGGCCGCAAG	Primer to generate CAR1 fragment
OAD1	30	CCGTCTGCTTTGGGCGGCCGCGCCAATTTT	Primer to generate GAD fragment in GAL4
OAD2	33	GGGGGCGGCCGCTTACTCTTTTTTTGGGTTTGG	Primer to generate GAD fragment in GAL4

Creation of mutations in the full-length ARGRII gene.

Different oligonucleotides (Table 1) were used to create substitutions by in vitro mutagenesis on single-stranded DNA prepared from pME52 (pGem7-ARGRII) or pNA84 (pALTER1-ARGRII) containing the full-length ARGRII gene expressed from its own promoter. The resulting plasmids were pNA31 (pGem7-argRII_C31L), pNA56 (pGem7-argRII_C38L,C41L), pWS1 (pALTER1-argRII_P36L), pNA88 (pALTER1-argRII_{D101A,E102A,E103A}), pNA89 (pALTER1-argRII_E108A,D109A), pNA119 (pALTER1-argRII_{Q89A,D96A,D111A,D112A}), pNA120 (pALTER1- argRII_{D96A,D111A,D112A}), pNA123 (pALTER1-argRII_{D101A,E102A,E103A,E108A,D109A}), pNA124 (pALTER1-argRII_{D101A,E102A,E103A,D111A,D112A}), pNA131 (pALTER1-argRII_{Q89A,E108A,D109A}), and pNA132 (pALTER1-argRII_{D96A,E108A,D109A}). The wild-type and different mutated 3.4-kb BamHI-BamHI DNA fragments were inserted into the centromeric vector pFL38 (pUC19, CEN6 ARS URA3), leading to plasmids pNA109 (wild type), pNA36, pNA110, pWS4, pNA114, pNA115, pNA125, pNA126, pNA129, pNA130, pNA139, and pNA140, respectively. We also recreated in vitro three mutations localized in the N-terminal end of ArgRII corresponding to mutations isolated by in vivo selection, leading to the following changes: D32N, G50D, and R99P. After mutagenesis, the 3.4-kb BamHI-BamHI DNA fragments were inserted into pFL38 vector, leading to plasmids pBJ250, pBJ202, and pBJ211.

Creation of mutations in the N-terminal end of ArgRII (aa 1 to 180).

To overexpress the wild-type and mutated N-terminal ends of ArgRII, we used plasmids pNA84, pNA31, pNA56, pBJ250, pWS1, pBJ202, pBJ211, pNA123, and pNA124 (described above) to synthesize BamHI-NotI DNA fragments by PCR, using oligonucleotides RII38 and RII58, containing a BamHI and a NotI restriction site, respectively (Table 1). After amplification by PCR, these different fragments were digested by BamHI and NotI and inserted into the BamHI and NotI sites of vector pYeF2 (pUC19, 2μm URA3 GAL10 promoter) (5), yielding plasmids pNA53 (GAL10-ArgRII1-180), pNA63 (GAL10-argRII1-180_C31L), pNA64 (GAL10-argRII1-180_C38L,C41L), pNA78 (GAL10-argRII1-180_D32N), pNA79 (GAL10-argRII1-180_P36L), pNA77 (GAL10-argRII1-180_G50D), pNA76 (GAL10-argRII1-180_R99P), pNA137 (GAL10-argRII1-180_{D101A,E102A,E103A,E108A,D109A}), pNA138 (GAL10-argRII1-180_{D101A,E102A,E103A,D111A,D112A}). All genes were sequenced to ensure that no additional mutation was introduced during the PCR procedure.

Construction of GAD-ARGRII fusions.

GBD refers to the DNA binding domain of the Gal4 activator, Gal4(1-147), and GAD refers to its activation domain, Gal4(768-881). GBD and GAD will refer to the DNA sequences encoding these domains. GBD-ARGRI and GBD-MCM1 fusions were constructed as described elsewhere (9). GAD-ARGRII fusions were constructed in vector pACTII (8); transformants harboring the vector or a derivative thereof were selected by omitting leucine from the growth medium.

(i) GAD-ARGRII fusion.

To construct the GAD-ARGRII gene fusion, we used oligonucleotide-directed in vitro mutagenesis to create a BamHI restriction site at the initiator codon of the ARGRII gene in plasmid pME52, bearing the ARGRII gene on a 3.4-kb DNA fragment (using oligonucleotide RII52), yielding plasmid pME8. The 3.2-kb BamHI-BamHI DNA fragment from plasmid pME8 was inserted in the BamHI site of the vector pACTII, leading to plasmid pME9 (GAD-ARGRII). In this GAD-ARGRII fusion, we determined the nucleotide sequence of the junction between the GAD-encoding region and the ARGRII gene to ensure that the fusions were in frame.

(ii) GAD-argRII fusions containing deletions in the ARGRII gene.

To fuse different portions of ARGRII to GAD, we amplified by PCR different DNA fragments, using as template the ARGRII gene present on pME52 and as primers synthetic oligonucleotides containing a BamHI restriction site. Oligonucleotides RII52-RII09, RII52-RII16, RII52-RII39, RII64-RII09, RII08-RII09, RII08-RII16, RII10-RII11, RII93-RII94, and RII14-RII15 (Table 1) allowed amplification of the regions from aa 2 to 180 (534 bp), 2 to 302 (900 bp), 2 to 128 (378 bp), 60 to 180 (360 bp), 91 to 180 (267 bp), 91 to 302 (633 bp), 381 to 470 (267 bp), 467 to 715 (744 bp), and 710 to 881 (513 bp), respectively. The different BamHI-BamHI DNA fragments were inserted in the BamHI site of pACTII vector, leading to plasmids pNA47 (GAD-ArgRII2-180), pNA58 (GAD-ArgRII2-302), pNA43 (GAD-ArgRII2-128), pNA48 (GAD-ArgRII60-180), pNA23 (GAD-ArgRII91-180), pNA27 (GAD-ArgRII91-302), pNA46 (GAD-ArgRII381-470), pNA150 (GAD-ArgRII470-710), and pNA25 (GAD-ArgRII710-880) in vector pACTII. All genes were sequenced to ensure that the fusions were in frame and that no mutation had been introduced during the PCR procedure.

(iii) GAD-argRII2-180 fusions containing different mutations in the ARGRII gene.

Plasmids pNA31, pNA56, pBJ250, pWS1, pBJ202, pNA86, pBJ211, pNA87, pNA88, pNA89, pNA90, pNA123, and pNA124 (described above) were used to synthesize 540-bp BamHI-BamHI DNA fragments by PCR, using oligonucleotides RII52 and RII09 (Table 1). These fragments were inserted in the BamHI restriction site of plasmid pACTII (GAD), yielding plasmids pNA65 (GAD-argRII2-180_C31L), pNA66 (GAD-argRII2-180_C38L,C41L), pNA74 (GAD-argRII2-180_D32N), pNA75 (GAD-argRII2-180_P36L), pNA73 (GAD-ArgRII2-180_G50D), pNA72 (GAD-argRII2-180_R99P), pNA146 (GAD- argRII2-180_{D101A,E102A,E103A,E108A,D109A}), and pNA161 (GAD-argRII2- 180_{D101A,E102A,E103A,D111A,D112A}). All genes were sequenced to ensure that the fusions were in frame and that no additional mutation had been introduced during the PCR procedure.

Construction and purification of GST fusion proteins.

To produce purified ArgRI, ArgRII, and Mcm1 proteins, we expressed them as glutathione S-transferase (GST) fusions in E. coli. To construct GST-ArgRI, we inserted a 1.7-kb BamHI fragment from plasmid pYM3 (9) containing the ARGRI gene in which a BamHI restriction site was introduced in the initiator codon, allowing in-frame fusion with GST. This fragment was inserted into the BamHI site of plasmid pGEX-5X-3 (Pharmacia), yielding plasmid pME53. To construct GST-Mcm1, we inserted a 0.9-kb BamHI fragment from plasmid pME15 (9) containing the MCM1 gene in which a BamHI restriction site was introduced in the initiator codon, allowing in-frame fusion with GST. This fragment was inserted into the BamHI site of plasmid pGEX-5X-3, yielding plasmid pME58. To construct fusions between GST and the wild-type and mutated N-terminal end (180 aa) of ArGRII, we synthesized by PCR BamHI-NotI fragments using oligonucleotide pair RII52-RII58 and plasmids pME52 (ArgRII wild type) and pNA123 (argRII_{D101A,E102A,E103A,E108A,D109A}). These fragments were inserted into the BamHI-NotI sites of plasmid pGEX-5X-3, yielding plasmids pME106 and pNA162, respectively, allowing in-frame fusions of the different proteins with GST.

Following transfection of E. coli XL1-B by plasmids expressing the different GST fusion proteins, induction of the fusion genes was achieved by addition of 500 μM isopropyl-β-d-thiogalactopyranoside for 3 h at 37°C. Bacterial pellets were resuspended in phosphate-buffered saline (PBS; 140 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄) containing a mixture of protease inhibitors and sonicated on ice for 3 min. After spinning at 12,000 rpm for 15 min, the supernatants were collected and sieved through a column containing glutathione-Sepharose 4B beads (Pharmacia) at 4°C. After extensive washes with PBS buffer, the GST fusions proteins were eluted by 0.1 M glutathione–50 mM Tris-HCl buffer (pH 8).

Replacement of the arginine boxes of the CAR1 promoter by a perfect PPAL sequence.

Nucleotides from −211 to −199 were replaced by the sequence 5′TTTCCTAATTAGGAAA3′. In vitro mutagenesis was performed as described by Stratagene, using plasmid pCV7 (pFL38-CAR1) (7) and oligonucleotides CAR1-99 and CAR1-100 (Table 1), yielding plasmid pFV72.

Enzyme assay.

β-Galactosidase activity was assayed as described by Miller (24). Protein contents were determined by the Folin method. Ornithine carbamoyltransferase (OTCase) and arginase activities were assayed as described previously (22).

DNA manipulation and DNA sequencing.

Restriction reactions were performed as recommended by the enzyme supplier. DNA fragments were isolated from agarose gels by Geneclean. Plasmid DNA was prepared by the alkaline lysis method (2) or rapid boiling lysis (12).

Denatured double-stranded DNA was used as template for DNA sequencing. Double-stranded DNA was prepared using Qiagen columns. DNA was sequenced by the dideoxynucleotide chain termination method of Sanger et al. (28), oligonucleotides being used as primers. Site-directed in vitro mutagenesis was performed with a Sculptor in vitro mutagenesis system (Amersham), an Altered Sites II in vitro mutagenesis system (Promega), or a QuikChange site-directed mutagenesis kit (Stratagene). Preparation of single-stranded DNA templates for in vitro mutagenesis is described by Messing (23).

Gel retardation assays.

Extract preparation and the binding assays were performed as described in reference 6. For binding studies, a 160-bp AluI-AluI fragment containing the control region of the ARG5,6 gene was used. The different CAR1 DNA fragments were synthesized by PCR from plasmids pCV7 and pFV72 as templates, using oligonucleotides CAR1-DE08 and CAR1-DE09, generating 170-bp DNA fragments. These fragments were end labeled with [γ-³²P]ATP (Amersham) by using polynucleotide kinase by the standard method (16). For the experiments described in Fig. 5, the bands were scanned with a Sharp JX330 scanner and quantified using Macintosh computer image analysis software (BioImage IQ version 2.1.1).

FIG. 5.

Effects of different mutations in the argRII1-180 protein on the formation of ArgR-Mcm1 complexes with ARG5,6 DNA as a function of different arginine concentration. (A) Gel retardation assays. The end-labeled 160-bp AluI-AluI ARG5,6 DNA fragment (about 1 ng) was incubated with 10 μg of yeast extracts prepared from strain 02463dII (ura3 leu2 argRII::KanmX4) (lanes 1 to 4) and strain 02463dII transformed with plasmids pNA53 (pGAL10-argRII1-180 URA3 2μm; lanes 5 to 8), pNA76 (R99P; lanes 9 to 12), pNA137 (D101A, E102A, E103A, E108A, D109A; lanes 13 to 16) and pNA138 (D101A, E102A, E103A, D111A, D112A; lanes 17 to 20). All strains were grown on M.ammonia-galactose and 50 μg of l-leucine; 25 μg uracil was also added in the culture of strain 02463dII. The different amounts of l-arginine added in the in vitro binding assays are indicated. (B) Western blot. Proteins were extracted from aliquots from the cultures described above, separated on an SDS–10% polyacrylamide gel, and electrotransferred to a Hybond membrane. Each lane contains about 50 μg of proteins. ArgRII proteins were visualized using anti-GST-ArgRII2-180 antibodies as described in Materials and Methods. (C) Gel retardation assays. The end-labeled 160-bp AluI-AluI ARG5,6 DNA fragment (about 1 ng) was incubated with 10 μg of yeast extracts prepared from strain 02463dII (ura3 leu2 argRII::KanMX4) transformed with plasmid pNA53 (pGAL10-argRII1-180 URA3 2μm; lanes 1 to 7), pNA138 (D101A, E102A, E103A, D111A, D112A; lanes 8 to 14), or pNA76 (R99P; lanes 15 to 21). All strains were grown on M.ammonia-galactose and 50 μg of l-leucine. The different amounts of l-arginine added in the in vitro binding assays are indicated. (D) Quantification of the protein-DNA binding activities presented in Fig. 6A. The images were captured using a Sharp scanner and quantified by Macintosh computer image analysis software. The saturation value for the wild-type ArgRII1-180 protein with 100 mM l-arginine was taken as 100%, and the values for the wild type (□) and mutants (argRII1-180_{D101A,E102A,E103A,D111A,D112A} [■] and argRII1-180_R99P [○]) were plotted accordingly.

Western blot analysis.

For ArgRII detection, 25 ml of exponentially growing cells was harvested by centrifugation, and the proteins were extracted by the trichloroacetic acid method described by Clontech. About 50 μg of total proteins was separated on a 10% polyacrylamide gel containing sodium dodecyl sulfate (SDS) as described by Laemmli (14). After electrotransfer of proteins to Hybond membranes, specific proteins were detected with polyclonal antibodies raised against GST-ArgRII2-180 obtained by injection of this purified protein in mice. Antibodies were a gift from Paul Jacobs. After incubation with anti-mouse immunoglobulin G-specific antibody conjugate to horseradish peroxidase, peroxidase activity was revealed with an enhanced chemiluminescence kit as specified by the supplier (Boehringer).

In vitro protein-protein interaction.

GST-ArgRII2-180 and the GST moiety were prepared as described above and independently immobilized on glutathione-Sepharose 4B beads. After being washed, beads were split into several portions for subsequent binding experiments. Semipurified yeast extracts (100 μg) (6) overexpressing ArgRI or Mcm1 were incubated overnight at 4°C with GST-ArgRII2-180 fusion protein immobilized on glutathione-Sepharose 4B beads. After being extensively washed with cold PBS, beads were boiled in SDS loading buffer, samples were separated on SDS–10% polyacrylamide gels, and proteins were transferred to Hybond membranes and detected by Western analysis with polyclonal antibodies raised against GST-Mcm1, GST-ArgRI, and a C-terminal ArgRI peptide (6), and peroxidase activity was revealed with an enhanced chemiluminescence kit as specified by the supplier (Boehringer).

RESULTS

The first 180 aa of ArgRII are sufficient for the formation of an arginine-dependent regulatory complex at the arginine boxes.

We previously showed that the formation of a protein-DNA complex with the arginine boxes in vitro required the presence of arginine and the integrity of ArgRI, Mcm1, and ArgRII. The use of antibodies in gel shift assays demonstrated that ArgRI and Mcm1 were part of the DNA-protein complex. Overexpression of ArgRII strongly enhanced the formation of this complex, indicating its participation (6, 19).

Previous experiments have also shown that most of the deletions created along the ARGRII gene affected ArgRII functions in vivo, but only the deletions of aa 1 to 60, containing the zinc cluster, and aa from 96 to 165, containing the first putative arginine binding domain, impaired the binding of the ArgR-Mcm1 complex to DNA in vitro (27). It is noteworthy that the deletion of the region from aa 533 to 625, comprising the second putative arginine binding domain (aa 563 to 587), impaired the ArgRII function but not its capacity to participate in the formation of the ArgR-Mcm1-DNA complex (27). These data suggested that about the first 200 aa could be sufficient to ensure an arginine-dependent binding of the complex to DNA.

To determine the minimal domain of ArgRII required for the formation of the arginine-dependent DNA-protein complex, we expressed the 128 and 180 N-terminal aa under the control of the GAL10 promoter from plasmids pNA44 and pNA53 (see Materials and Methods). After growth on M.galactose or M.glucose of strain 02463dII (ura3 leu2 argRII::KanMX4) transformed with pNA44 or pNA53, proteins were extracted and semipurified on heparin-Sepharose as described by Dubois and Messenguy (6). Gel shift assays performed with the ARG5,6 promoter showed that the first 180 aa, but not the first 128 aa of ArgRII, were sufficient for the formation of ArgR-Mcm1 complexes (Fig. 2, lanes 2 and 4). This binding was strongly enhanced with the extract from the strain overexpressing ArgRII1-180 (lanes 3 and 4) and was arginine dependent (lanes 5 and 6). To determine whether this truncated protein is able to fulfill the ArgRII functions in vivo (repression of anabolism and induction of catabolism), we measured the activities of the anabolic enzyme OTCase and the catabolic enzyme arginase, in the presence and absence of arginine in the growth medium. Therefore, the argRII deletion strain 02463dII (ura3 leu2 argRII::KanMX4) was transformed with plasmids expressing different portions of ArgRII under the control of the GAL10 promoter: pYeFRII1-880 (pME50), pYeFRII1-128 (pNA44), pYeFRII1-180 (pNA53), and pYeFRII1-180-AD_Gal4 (pNA54). The transformed strains in which ARGRII expression is dependent on the GAL10 promoter were grown on galactose as the carbon source, with or without l-arginine (1 mg/ml). Measurements of OTCase and arginase specific activities showed that the first 180 aa of ArgRII repressed partially the synthesis of OTCase but did not induce the synthesis of arginase (Table 2). The recovery of induction of arginase required the addition of the Gal4 activation domain at the C end of the 180 aa of ArgRII (Table 2). This hybrid protein had two effects on expression of the ARG3 gene, encoding OTCase. The basal enzyme level on M.ammonia was enhanced, probably resulting from the presence of the Gal4 activation domain, but when arginine was added to the growth medium, a twofold repression was observed.

FIG. 2.

Arginine-dependent binding of ArgRII1-180, ArgRI, and Mcm1 to ARG5,6 DNA. The end-labeled 160-bp AluI-AluI ARG5,6 DNA fragment (about 1 ng) was incubated with 10 μg of yeast extracts prepared from strain 02463dII (ura3 leu2 argRII::KanMX4) transformed with plasmid pNA44 (pGAL10-argRII1-128 2μm, URA3) (lanes 1 and 2) and pNA53 (pGAL10-argRII1-180 2μm URA3) (lanes 3 to 6) grown on M.ammonia-glucose and 50 μg of l-leucine per ml (lanes 1 and 3) or M.ammonia-galactose and 50 μg of l-leucine per ml (lanes 2, 4, 5, and 6). In all in vitro assays, 5 mM l-arginine was added except in lane 5.

TABLE 2.

Capacity of the N-terminal end of ArgRII1-180 to repress the expression of arginine anabolic genes and induce the expression of arginine catabolic genes^a

Strain 02463dII plus:	Sp act
	Arginase (μmol of urea formed/h/mg of protein)		OTCase (μmol of citrulline formed/h/mg of protein
	M.ammonia	M.ammonia + arginine	M.ammonia	M.ammonia + arginine
No plasmid	4	5	30	28
pME50 (URA3 2μm GAL10-ArgRII1-880)	41	153	13	5.6
pNA44 (URA3 2μm GAL10-ArgRII1-128)	8	9	22	21
pNA53 (URA3 2μm GAL10-ArgRII1-180)	7	7	23	11
pNA54 (URA3 2μm GAL10-ArgRII1-180-ADGal4)	14	55	51	27

^aArginase and OTCase activities were measured in strain 02463dII (ura3 leu2 argRII::KanMX4) transformed with various plasmids after growth on 2% galactose as the carbon source. Specific activities are the means of three independent measurements that did not differ by more than 15%; 50 μg of l-leucine per ml was added in the growth media; 25 μg of uracil was also added when no plasmid was present in the strain. 

All of these data suggest that the first 180 aa of ArgRII are able to bind the arginine boxes in vitro as well as in vivo. The N-terminal end of ArgRII contains thus a region contacting DNA (presumably the Zn₂C₆ cluster), a domain interacting with ArgRI and Mcm1, and a sequence binding the effector, arginine.

To determine the amino acids responsible for these different interactions, we tested the interaction of the N-terminal end of ArgRII with ArgRI and Mcm1 using the two-hybrid system and created a series of mutations in the Zn₂C₆ cluster and in the putative arginine binding domain.

The first 180 aa of ArgRII interact with ArgRI and Mcm1 in vivo and in vitro.

Using the two-hybrid system, we have shown that the full-length ArgRII interacts with ArgRI and Mcm1 (9). To determine the domains of ArgRII interacting with the two MADS-box proteins, we fused different portions of ArgRII to GAD (see Materials and Methods) and determined their interaction with GBD-ArgRI and GBD-Mcm1. Strain HY was transformed with plasmid pME46 (pGBD-ARGRI TRP1 2μm) and the different GAD-ARGRII plasmids (pGAD-ARGRII LEU2 2μm) (Fig. 3). Expression of the lacZ reporter gene was monitored by β-galactosidase activity assays. The first 180 aa of ArgRII interacted with ArgRI with about the same efficiency as the full-length ArgRII (Fig. 3, lines 1 and 2). In contrast when the first 128 aa or aa 60 to 180 of ArgRII were used, only 20% of the interaction capacity was retained (Fig. 3, lines 2, 4, and 5). All other regions showed only a poor interaction, since the β-galactosidase levels were comparable to the level obtained in a strain expressing only GBD-ArgRI (lines 8 to 10 compared to line 11). A strain transformed with only GAD-ArgRII exhibited no detectable β-galactosidase activity.

FIG. 3.

Determination of the region of ArgRII interacting with ArgRI and Mcm1 in vivo. The full-length ArgRII and different portions of ArgRII were fused in frame with the activation domain of Gal4 (GAD-ArgRII), and full-length ArgRI and Mcm1 were fused in frame with the DNA binding domain of Gal4 (GBD). Transcription activation of the lacZ reporter gene was determined by β-galactosidase activity assays, performed at 30°C on extracts of at least three independent transformants containing both plasmids. The standard error was 15% of the mean. Specific activity is expressed in nanomoles of o-nitrophenyl-β-d-galactopyranoside hydrolyzed per minute per milligram of protein. Hatched boxes correspond to the GAD, black boxes represent portions of ArgRII, and grey boxes represent the Zn₂C₆ zinc cluster (aa 21 to 48).

Similarly, the interactions between different portions of ArgRII and Mcm1 were measured by assaying β-galactosidase activity in HY strains transformed with plasmid pNA51 (pGBD-MCM1 TRP1 2 μm) and the different GAD-ARGRII plasmids (pGAD-ARGRII LEU2 2μm). As for ArgRI, the first 180 aa of ArgRII proved to be sufficient to interact with Mcm1, but the region between aa 181 and 302 could also contribute to increase the efficiency of the interaction (Fig. 3, lines 2 and 3 compared to lines 6 and 7). All other regions showed only a poor interaction, since the β-galactosidase levels were comparable to the level obtained in a strain expressing only GBD-Mcm1 (lines 8 to 10 compared to line 11).

ArgRII thus interacts with ArgRI and Mcm1 through its first 180 aa. However, we cannot exclude that another region of ArgRII could contact the MADS-box proteins, since we have no proof of the stability of the GAD-ArgRII hybrid proteins; none of the GAD-ArgRII proteins could be detected by Western blotting using antibodies raised against GAD or GST-ArgRII2-180.

To provide biochemical evidence for interaction between the N-terminal end of ArgRII and the two MADS-box proteins, we performed GST pull-down experiments using bacterially expressed GST-ArgRII2-180 and semipurified yeast extracts overexpressing ArgRI or Mcm1 (see Materials and Methods). Equivalent amounts of GST-ArgRII2-180 and GST proteins immobilized on glutathione-Sepharose 4B beads were incubated with yeast extracts containing ArgRI or Mcm1. After extensive washes, bound proteins were visualized by Western blotting with antibodies raised against GST-Mcm1, allowing us to detect GST, GST fusion proteins, and Mcm1. GST alone did not retain Mcm1 (Fig. 4A, lane 2), whereas Mcm1 bound to immobilized GST-ArgRII2-180 (lane 4). Figures 4B and C show the results obtained for ArgRI. Since GST and ArgRI migrate at the same position, we did not use antibodies raised against GST-ArgRI but instead used two antibodies, one raised against GST to identify GST and GST fusion proteins and a second raised against an ArgRI C-terminal peptide to detect ArgRI. Therefore, two identical gels were transferred to Hybond membranes and hybridized with each antibody. GST alone did not interact with ArgRI (Fig. 4B, lane 2). In contrast, a significant amount of ArgRI bound to GST-ArgRII2-180 (Fig. 4C, lane 4).

FIG. 4.

In vitro association of the first 180 aa of ArgRII with ArgRI and Mcm1. (A) In vitro interaction between ArgRII2-180 and Mcm1. Purified GST (lanes 1 and 2) and GST-ArgRII2-180 (lanes 3 and 4) were immobilized on glutathione-Sepharose 4B beads. About 100 μg of semipurified proteins (6) from strain 02463d (ura3 leu2) transformed with plasmid pED40 (overexpressing the MCM1 gene) was allowed to bind to the beads (lanes 2 and 4). Lane 5 contains 10 μg of the yeast semipurified extract. After extensive washing, bound proteins were separated on an SDS–10% polyacrylamide gel and detected by Western blotting using polyclonal antibodies against GST-Mcm1. Size standards are indicated on the left. (B and C) In vitro interaction between ArgRII2-180 and ArgRI. Purified GST (lanes 1 and 2) and GST-ArgRII2-180 (lanes 3 and 4) were immobilized on glutathione-Sepharose 4B beads. About 100 μg of semipurified proteins (6) from strain 02463d (ura3 leu2) transformed with plasmid pME51 (overexpressing the ARGRI gene on galactose) was allowed to bind to the beads (lanes 2 and 4). Lane 5 contains 10 μg of the yeast semipurified extract. After extensive washing, bound proteins were separated on an SDS–10% polyacrylamide gel and detected by Western blotting using polyclonal antibodies against GST (B) or against a C-terminal peptide of ArgRI (C). Size standards are indicated on the left.

Analysis of mutations created in the N-terminal end of ArgRII (aa 1 to 180).

To further analyze the region of the protein sufficient to form a complex with ArgRI and Mcm1 able to interact with DNA in an arginine-dependent fashion, we created a series of mutations in the Zn₂C₆ cluster and in the putative arginine binding domain. The mutations were analyzed for their effect on binding of the ArgR-Mcm1 complex to DNA in vitro as a function of the arginine concentration, for the ability to interact with ArgRI and Mcm1, and for the capacity to repress the expression of anabolic genes and to induce the catabolic genes.

In the Zn₂C₆ cluster, we mutated by in vitro mutagenesis (see Materials and Methods) some residues among the Zn₂C₆ regulatory proteins, thus creating the substitutions C31L, C38L, C41L, D32N, P36L, and G50D. The D32N and G50D substitutions, which corresponded to mutations selected in vivo, were recreated by in vitro mutagenesis. Even at high arginine concentrations, all of the mutations impaired the binding of the ArgR-Mcm1 complex to DNA, as expected for this type of protein (data not shown).

In the putative arginine binding domain, from aa 89 to 114, we replaced by alanine a series of acidic residues and some of the amino acids that were shown to contact arginine in the E. coli ArgR repressor. Most of the single or combined alanine replacements of residues Q89, D96, D101, E102, E103, E108, D109, D111, and D112 had no significant effect in vitro, although some changes partially impaired the ArgRII function in vivo, since the repression of OTCase and induction of arginase were reduced (see Table 4 and comments in the next section). In contrast, the multiple substitutions D101A,E102A,E103A,E108A,D109A introduced in ArgRII1-180 (pNA137) abolished formation of the protein-DNA complexes, whereas the combined substitutions D101A,E102A,E103A,D111A,D112A (pNA138) and the substitution R99P (pNA76), which corresponds to a mutation isolated in vivo, reduced the binding efficiency. As shown by Western blot, the various ArgRII1-180 mutated proteins were present in all extracts (Fig. 5A and B).

TABLE 4.

Effect of mutations in different regions of the full-length ArgRII protein on expression of arginine anabolic and catabolic genes^a

Strain 02463dII plus:	Arginase (CAR1) sp act (μmol of urea formed/ h/mg of protein)		Induction factorb	OTCase (ARG3) sp act (μmol of citrulline formed/h/mg of protein)		Repression factorc
	M.ammonia	M.ammonia + arginine		M.ammonia	M.ammonia + arginine
No plasmid	3	3	1	43	42	1
pNA109 (ArgRII1-880)	23	90	30	25	11	3.9
pBJ211 (argRII1-880R99P)	6.5	32	11	47	33	1.3
pNA114 (argRII1-880D101A,E102A,E103A)	14	40	13	31	21.5	2
pNA115 (argRII1-880E108A,D109A)	10	40	13	32	19	2.3
pNA125 (argRII1-880Q89A,D96A,D111A,D112A)	13	43	13	38	22	1.9
pNA126 (argRII1-880D96A,D111A,D112A)	18	60	20	23	15	2.8
pNA129 (argRII1-880D101A,E102A,E103A,E108A,D109A)	5	7	2	64	56	1
pNA130 (argRII1-880D101A,E102A,E103A,D111A,D112A)	10.5	28.5	9	41.5	33	1.3
pNA139 (argRII1-880Q89A,E108A,D109A)	9	36	12	50	28	1.5
pNA140 (argRII1-880D96A,E108A,D109A)	12	29	10	47	26	1.7

^aArginase and OTCase activities were measured in strain 02463dII (ura3 leu2 argRII::KanMX4) transformed with various plasmids (in all experiments, the low-copy-number plasmid pFL38 was used; see Materials and Methods). Specific activities are the means of three independent measurements that did not differ by more than 15%; 50 μg of l-leucine per ml was added in the growth media; 25 μg of uracil was also added when no plasmid was present in the strain. 

^bRatio between the level on M.ammonia plus arginine and the basal level of the mutant argRII strain. 

^cRatio between the level on M.ammonia plus arginine and the derepressed level of the mutated argRII strain. 

It is worth noting that the mutated ArgRII1-180 protein produced from plasmid pNA138 (D101A, E102A, E103A, D111A, D112A) required at least a 10- to 20-fold-higher arginine concentration than the wild-type protein to obtain 50% binding efficiency of the ArgR-Mcm1 complex to DNA, as shown by densitometric analysis of autoradiographs of the different DNA-protein complexes obtained in gel shift experiments performed with different arginine concentrations (Fig. 5C and D). The R99P substitution also impaired the DNA binding capacity and to a lesser extent the requirement for arginine. In contrast to mutations in the zinc cluster which cannot be rescued by arginine, some mutations in the region of ArgRII showing sequence identity with bacterial repressors led to an apparent reduced affinity for arginine, suggesting that this region could contain an arginine binding site.

To determine whether the lack of formation of an ArgR-Mcm1-DNA complex observed in the argRII mutants described above resulted from the loss of interaction of ArgRII with ArgRI or Mcm1, we fused the coding sequence of the first 180 aa containing the different mutations to the GAD (see Materials and Methods). Strain HY was transformed with a pGBD-ARGRI (pME46) or pGBD-MCM1 (pNA51) plasmid and plasmids pGAD-ArgRII2-180 (pNA47), pGAD-argRII2- 180_C31L (pNA65), pGAD-argRII2-180_C38L,C41L (pNA66), pGAD-argRII2-180_D32N (pNA74), pGAD-argRII2-180_P36L (pNA75), pGAD-argRII2-180_G50D (pNA73), pGAD-argRII2-180_R99D (pNA72), pGAD-argRII2-180_{D101A,E102A,E103A,E108A,D109A} (pNA146), and pGAD-argRII2-180_{D101A,E102A,E103A,D111A,D112A} (pNA161). As shown in Table 3, β-galactosidase assays from the different transformed strains revealed that none of these mutations led to the loss of interaction between the mutated argRII proteins and the two MADS-box proteins, although the efficiency of interaction was reduced in some mutants. Thus, these mutations do not affect primarily the formation of the ArgRI-ArgRII-Mcm1 complex but rather its interaction with the arginine boxes or with the effector, arginine.

TABLE 3.

Ability of mutated argRII2-180 proteins to interact with ArgRI and Mcm1

Hybrid	β-Galactosidase sp act (nmol of o-nitrophenyl-β-d-galactopyranoside hydrolyzed/min/mg of protein)
	GBD-ArgRI (pME46)	GBD-McmI (pNA51)
None	<1	4
GAD-ArgRII2-180 (pNA47)	95	56
GAD-argRII2-180C31L (pNA65)	66.5	40
GAD-argRII2-180C38L,C41L (pNA66)	45	23
GAD-argRII2-180D32N (pNA74)	46	38
GAD-argRII2-180P36L (pNA75)	56	22
GAD-argRII2-180G50D (pNA73)	69	18
GAD-argRII2-180R99P (pNA72)	43	30.5
GAD-argRII2-180D101A,E102A,E103A,E108A,D109A (pNA146)	46	27
GAD-argRII2-180D101A,E102A,E103A,D111A,D112A (pNA161)	37	23

^aTranscription activation of the lacZ gene was estimated by determination of β-galactosidase activity in strain HY cotransformed with plasmids expressing GBD-ArgRI (pME46) or GBD-Mcm1 (pNA51) and the different wild-type or mutated GAD-argRII2-180 fusions. Specific activities are the means of three independent measurements with variation was less than 15%. 

Effects of mutations created in the full-length ArgRII protein on its ability to regulate the expression of arginine genes.

The effects of mutations analyzed above were studied using the overexpressed N-terminal portion of ArgRII protein. To further investigate the involvement of some amino acids located in the zinc cluster or in the putative arginine binding site, we created a series of nucleotide substitutions in the full-length ARGRII gene, expressed from its own promoter and thus in physiological conditions (Table 4).

All mutations created in the zinc cluster impaired, as expected, both the induction of arginase (CAR1 gene product) and the repression of OTCase (ARG3 gene product) by arginine (data not shown). In the putative arginine binding domain, only the combined alanine substitutions of five amino acids (D101,E102,E103,E108,D109) led to a complete loss of ArgRII function (Table 4) and impaired binding to the arginine boxes even at high arginine concentrations (data not shown), as observed with the N-terminal end of ArgRII containing the same substitutions (Fig. 5A, lanes 13 to 16). Some other combinations and the substitution R99P reduced significantly induction of arginase and repression of OTCase (Table 4). Most of the mutations affected more readily the repression of anabolic genes than the induction of catabolic genes.

Reconstitution of the arginine-dependent binding activity to arginine boxes from recombinant GST-ArgRI, GST-ArgRII2-180 and GST-Mcm1.

Although previous results demonstrated that ArgRI, Mcm1, and ArgRII were required for the assembly of a heteromeric complex at the arginine boxes, they did not prove that they were the only proteins required for assembly and arginine-dependent DNA binding. To address this point, we performed mobility shift studies using various combinations of purified recombinant GST-ArgRI, GST-Mcm1, and GST-ArgRII2-180 proteins (see Materials and Methods and Fig. 6). The individual proteins were unable to bind to the arginine boxes, even in the presence of arginine (Fig. 6A, lanes 1 to 3). In contrast, the combination of GST-ArgRII2-180 with GST-ArgRI or GST-Mcm1 could reconstitute an arginine-dependent binding activity, whereas GST-ArgRI and GST-Mcm1 were not able to bind to the arginine boxes (lanes 4 to 9). Interestingly, there was no complex formation when GST-Mcm1 was combined with the mutant GST-argRII2-180_{D101A,E102A,E103A,E108A,D109A} (lane 13). The GST-ArgRII2-180 and GST-ArgRI combinations led consistently to a weaker binding activity, with the formation of one complex with faster mobility. The combination of the three recombinant proteins increased significantly the amount of DNA-protein complex formed (lanes 10 and 11). When suboptimal concentrations of GST-ArgRII2-180 and GST-Mcm1 were used in this assay, DNA binding activity was significantly reduced, and addition of GST-ArgRI restored the formation of an ArgR-Mcm1-DNA complex showing cooperativity between ArgRI and Mcm1 (Fig. 6B, lanes 3 and 4).

FIG. 6.

Reconstitution of the arginine-dependent binding activity to the arginine boxes from recombinants GST-ArgRI, GST-ArgRII2-180, and GST-Mcm1. DNA mobility shift assays were performed with a radiolabeled AluI-AluI ARG5,6 fragment, incubated with about 3 μg of purified GST-ArgRI, wild-type GST-ArgRII2-180, mutated GST-argRII2-180_{D101A,E102A,E103A,D111A,D112A}), and GST-Mcm1 recombinant proteins in the various combinations indicated. In lanes 1 to 5 of panel B, various concentrations of the different GST-recombinant proteins were used, as indicated; 5 mM l-arginine was added in the binding assay where indicated.

Taken together, these data demonstrate that in vitro, the arginine box binding activity requires at least arginine, the N-terminal end of ArgRII, and one of the two MADS-box proteins when used at nonphysiological concentrations. In vivo, the two MADS-box proteins could cooperate to recruit ArgRII. The binding of arginine to the first 180 aa of ArgRII would stabilize the interaction of the ArgR-Mcm1 complex with DNA.

Reconstitution of the arginine-dependent binding activity to CAR1 promoter containing PPAL sequence replacing arginine box B.

Mcm1 and ArgRI are able to interact with PPAL DNA, although the affinity of ArgRI for this sequence is much weaker (data not shown and reference 31). However, as shown above, the two MADS-box proteins did not bind alone or in combination with the arginine boxes, sharing homology with the PBox sequence (20). Their binding to DNA required the N-terminal end of ArgRII and arginine. To determine if the requirement for ArgRII could be bypassed by recruiting more efficiently Mcm1 to a promoter containing the arginine boxes, we replaced arginine box B of the CAR1 promoter by the perfectly palindromic PPAL (Fig. 7B) (7). It is worth noting that this box B is absolutely required for CAR1 induction by arginine. In Fig. 7A we show the binding of GST-Mcm1, alone or in combination with wild-type or mutated GST-ArgRII2-180, to the wild-type arginine boxes (DNA fragment from pCV7; lanes 1 to 6) and to a DNA fragment in which the PPAL sequence replaces box B (upstream PPAL [UPPAL] fragment from pFV72; lanes 7 to 15) (see Materials and Methods for construction of modified CAR1 promoter). As for the ARG5,6 promoter, the formation of a complex with the wild-type CAR1 promoter required the presence of GST-Mcm1, GST-ArgRII2-180, and arginine (lanes 4 and 5). No complex was obtained with the mutated GST-argRII2- 180_{D101A,E102A,E103A,E108A,D109A} (lane 6), although this protein still interacts with Mcm1 in vivo (Table 2). In contrast, GST-Mcm1 was able to bind the DNA fragment from pFV72 independently of arginine (lanes 7 and 8), but in the presence of GST-ArgRII2-180 we observed the formation of a complex of slower mobility only with arginine (lanes 9 and 10); this latter complex was absent when the mutated GST-argRII2- 180_{D101A,E102A,E103A,E108A,D109A} protein was used (lane 15). In contrast, GST-ArgRI interacted very poorly with the modified CAR1 DNA sequences (data not shown) and did not change the complex formed with GST-Mcm1 (lanes 11 and 12). All of these data suggest that even when Mcm1 is artificially recruited to the CAR1 promoter by insertion of a perfect palindromic P sequence, ArgRII, but not ArgRI, is still required to obtain an arginine-dependent response in vitro.

FIG. 7.

Reconstitution of the arginine-dependent binding activity to the CAR1 promoter with modified arginine boxes. (A1 and A2) Gel shift assays. DNA mobility shift assays were performed with about 1 ng radiolabeled wild- type CAR1 probe (A1) or 0.05 ng of mutated CAR1 probe (A2). These probes were incubated with about 3 μg (A1) or 1 μg (A2) of purified wild-type GST-ArgRII2-180 or mutated GST-argRII2-180_{D101A,E102A,E103A,D111A,D112A}), GST-ArgRI, and GST-Mcm1 recombinant proteins in the various combinations indicated; 5 mM l-arginine was added in the binding assay where indicated. Positions of the different protein-DNA complexes are indicated by arrows. (B) Different CAR1 probes were synthesized by PCR as described in Materials and Methods from plasmids pCV7, containing the CAR1 wild-type promoter (6), or pFV72, containing the CAR1 promoter in which box B is replaced by the PPAL sequence (UPPAL means upstream location of PPAL).

DISCUSSION

The first 180 aa of ArgRII are sufficient for conveying the formation of an arginine-dependent ArgR-Mcm1-DNA complex.

Previous results had shown that at physiological concentration, almost the entire ArgRII protein was required to ensure its function in anabolic repression and in catabolic activation. However, the only mutations that affected the formation of an arginine-dependent protein-DNA complex were located in the N-terminal end of ArgRII. The loss of function did not result from a reduced stability of the mutated argRII proteins (27). Here we show that qualitatively the first 180 aa of ArgRII are sufficient for conveying interaction with its two partners ArgRI and Mcm1 to form a regulatory complex required for binding to DNA in an arginine-dependent manner. However, at physiological levels the C-terminal end of ArgRII is essential for its function since most of the deletions created along the protein impaired its ability to repress the anabolic genes and to induce the catabolic genes. This part of the protein may be involved in conformation changes or in intramolecular or ArgRII-ArgRII interactions.

The N-terminal end of ArgRII contains two domains.

Amino acid alignment of the N-terminal end of ArgRII with yeast and bacterial regulatory proteins reveals the presence of a Zn₂C₆ zinc cluster between amino acids 21 to 48, similar to findings for the other members of this family such as Gal4, Put3, and Ppr1, and a region between aa 89 to 114 similar to the arginine binding domains of E. coli ArgR and B. subtilis AhrC repressors.

Modification of several conserved amino acids in the zinc cluster led to total or partial loss of binding of the ArgR-Mcm1 complex to DNA, although these changes did not impair significantly the interaction of ArgRII with ArgRI and Mcm1. This region seems thus to be involved in DNA recognition as it is in other regulatory proteins of this family.

The region between aa 89 to 114, which presents some identity with the two bacterial arginine repressors, seems to be the binding site of arginine. The combination of different amino acid changes in this region impairs ArgRII function, especially the combined replacement by alanine of D101,E102,E103 with E108,D109 or with D111,D112. The first combination of five changes leads to a total argR phenotype in vivo (no growth on arginine or ornithine as a nitrogen source) as well as a total loss of DNA binding activity in vitro, whereas the second combination shows a strong arginine-dependent phenotype in vivo and the requirement for high arginine concentration in vitro. The growth of such a mutant is strongly affected when ornithine but not when arginine serves as the sole nitrogen source (data not shown), and binding of the ArgR-Mcm1 complex to DNA requires at least 10-fold more l-arginine than with the wild-type ArgRII protein, indicating a strong decrease in the apparent affinity of this modified protein for arginine. Our analysis shows the importance of acidic residues in this region, possibly because they interact with arginine. It is noteworthy that D111 and D112 correspond to the two aspartic residues 128 and 129 of the E. coli ArgR repressor. The double-mutant protein D128N and D129V shows an arginine-independent binding of the E. coli repressor to DNA (4). The R99P substitution, which corresponds to a mutation obtained by in vivo selection, also modifies the affinity of the protein for arginine. This replacement could modify the conformation of that region rather than impair directly the interaction of that residue with arginine, whereas the acidic residues could bind arginine. When these multiple changes of acidic residues to alanine are created in the N-terminal end of the full-length ArgRII, similar results are obtained, confirming that the second region of similarity with the bacterial arginine repressors (between aa 563 and 587 in ArgRII [Fig. 1]) does not compensate for the loss of function of the proximal region. Further biochemical experiments as well as structural studies will be required to determine if the N-terminal end of ArgRII directly binds arginine.

ArgRI, ArgRII, and Mcm1 are sufficient to interact with the arginine boxes in a cooperative fashion.

Four proteins, ArgRI, ArgRII, Mcm1, and ArgRIII, are necessary for the regulation of arginine metabolism. Recent results showed that the role of ArgRIII was to stabilize the two MADS-box proteins ArgRI and Mcm1 (9). These two proteins and ArgRII are required for an arginine-dependent DNA binding activity, and the reconstitution of this activity from recombinants GST-ArgRI, GST-ArgRII2-180, and GST-Mcm1 demonstrates that these proteins are sufficient for the assembly of the heteromeric complex at the arginine boxes when arginine is present. None of these purified recombinant proteins interact alone with DNA, whereas the combination of GST-ArgRII2-180 with GST-Mcm1 allows the formation of a protein-DNA complex, which is enhanced when GST-ArgRI is added. GST-ArgRII2-180 and GST-ArgRI bind very weakly to DNA to form a complex of faster mobility. All of these interactions require arginine and the N-terminal end of ArgRII, supporting the idea that ArgRII is the sensor of arginine. This is confirmed by the experiments in which Mcm1 is artificially recruited to the CAR1 promoter. Mcm1 bound to the PPAL sequence replacing box B (UPPAL) forms a complex of slower mobility only with wild-type GST-ArgRII2-180 and when arginine is present.

Since these in vitro experiments were performed with large amounts of purified recombinant proteins compared to the physiological concentrations, the in vivo situation could be quite different. ArgRI could facilitate the interaction of Mcm1-ArgRII-arginine with the arginine boxes, which contain imperfect PBox sequences. It is worth noting that a strain in which the ARGRI gene is deleted is unable to grow on ornithine as the sole nitrogen source but is only moderately impaired when arginine is the sole nitrogen source. Similarly, MCM1 gene mutants are partially impaired in their growth on arginine. The absence of growth on this medium is observed only in a strain lacking the ARGRI gene and mutated in the MCM1 gene (unpublished result), suggesting that each MADS-box protein can fulfill the function of the other protein, to a certain extent, when the arginine intracellular pool is high. In contrast, an argRII deletion strain is unable to grow on either arginine or ornithine, in full agreement with the key role of ArgRII in mediating the cellular response to arginine.