Copyright 0 1982 by the Genetics Society of America A CIS-DOMINANT MUTATION IN ASPERGILLUS NIDULANS AFFECTING THE EXPRESSION OF THE amdS GENE IN THE PRESENCE O F MUTATIONS IN THE UNLINKED GENE, amdA MICHAEL J. HYNES’ Department of Genetics, La Trobe University, Bundooro, Victoria, Australia Manuscript received February 8, 1982 Revised copy accepted June 1,1982 ABSTRACT A mutant producing very high levels of the acetamidase enzyme encoded by the amdS gene has been isolated in a strain containing the amdA7 mutation, which itself causes high levels of this enzyme. Genetic analysis has shown that this mutation, designated amd166, is adjacent to the amdS gene and is cisdominant in its effect. The amd166 mutation has little effect on amdS expression when present in strains not containing the amdA7 mutation. Two other omdA mutations investigated also interact with the amd166 mutation to result in high acetamidase levels. No interaction between amd166 and any of the other putative regulatory genes affecting amdS expression has been observed. The amd166 mutation has been located by fine structure mapping at the extreme end of the controlling region, which has previously been defined by genetic mapping (HYNES1979). Analysis of this region has been extended by mapping new mutations resulting in loss of amdS expression. One of these defines the most extreme site capable of mutation to loss of gene function found so far. N lower eukaryotes many apparent regulatory genes have been defined that Igenes. appear to act in trans to regulate the expression of one more structural Genetic methods can provide evidence for a direct regulatory action of or a proposed regulatory gene if cis-acting mutations, adjacent to the controlled structural gene, affect regulation of gene expression and show specific interactions with mutations in regulatory genes. Such specific mutations are particularly valuable because, in a number of cases in fungi, apparent cis-acting mutations have been found to arise by fusions to new controlling regions (promoters) resulting from chromosomal rearrangements (ARST, RAND and BAILEY1979; MCKNIGHT,CARDILLO and SHERMAN 1981; SHERMAN and HELMS 1978) or insertions (ERREDE et al. 1980). A new cis-dominant mutation affecting expression of the amdS gene coding for the acetamidase enzyme of Aspergillus nidulans is described here. This mutation interacts with mutations in the unlinked amdA gene to cause high levels of amdS expression. The complex control of the amdS gene has been reviewed previously (HYNES 1978a). Briefly, the acetamidase enzyme is inducible by sources of acetylcoenzyme A such as acetate, L-threonine and acetamide by a mechanism sharing at least some control elements with the regulation of acetate metabolism via the glyoxylate by-pass. Benzoate also results in enzyme induction but no genes Present address and address for correspondence Department of Genetics, Melbourne Unlversity, Parkville, Victoria 3052, Australia Genetics 102 139-147 October. 1982 140 M. I. HYNES specifically involved in this control are known. The acetamidase is highly inducible by w-amino acids (e.g., p-alanine), and the amdR gene, unlinked to the amdS gene, appears to act as a positive regulatory gene in this process. Nitrogen metabolites such as ammonium and L-glutamine result in strong repression of the acetamidase. The areA gene has been implicated in this control mechanism, which also affects most nitrogen catabolic pathways (ARSTand COVE1973). The amdA gene on linkage group VI1 is of prime interest in this work. Mutations (e.g., amdA7) in this gene have been found to cause elevated acetamidase levels compared with wild type (HYNES1978a). Mutations in the amdA gene do not interact with other regulatory mutations, and induction by all inducers is observed in amdA mutants. Thus the effect of amdA mutations is to elevate the overall level of amdS expression. No mutations in amdA have been observed to result in lowered levels of amdS expression. Mutations in the amdA gene have no observed effect on growth on a variety of carbon and nitrogen sources. Thus the effect of the amdA gene has been thought to be specific for amdS expression. However, recent observations of labeled polypeptides separated on polyacrylamide gels show that amdA mutants synthesize high levels of a polypeptide of unknown function (M. J. HYNESand J. A. KING,unpublished). A cis-dominant mutation, amdI66, adjacent to amdS, is described here. This mutation interacts with amdA lesions resulting in very high levels of acetamidase in both the presence and absence of inducers. By itself amdI66 has only a small effect. Mapping of mutations in the amdS region has resulted in the definition of an adjacent controlling region defined by cis-acting mutations (HYNES1979). Analysis of this region has been extended by the mapping of the amdI66 mutation and derived secondary mutations resulting in loss of amdS expression. One of these new mutations defines the most extreme site on the genetic map capable of mutating to loss of gene function found so far. MATERIALS AND METHODS Strains and genetic techniques: All strains were derived from those previously described (HYNES 1978a, 1978b, 1979, 1980). Where necessary, the genotypes of double mutant strains were checked by appropriate crosses. Genetic techniques were standard for A. nidulans (CLUTTERBUCK 1974). Methods used for the isolation of mutants and for fine structure mapping of the amdS region were those previously described (HYNES1979). Media and growth conditions: The standard minimal medium used was that of COVE(1966) with 1%glucose, 1%sucrose or no carbon source added. Ammonium was added as ammonium tartrate. Growth tests were performed with solid medium (1%agar) and incubation at 37' for 2-3 days. Mycelium for enzyme determinations was grown at 30' in a Gallenkamp orbital incubator and harvested as described (HYNES1972). Preparation of crude extracts and enzyme ussays: Crude extracts of freshly harvested mycelium were made as described (HYNES1972) using 100 mM sodium orthophosphate buffer, pH 7.2. Acetamidase (E.C. 3.5.1.4) was assayed as described (HYNES1972). All enzyme activities are expressed as milli-units per milligram of soluble protein where one unit is the amount of enzyme that catalyzes the production of 1pmol of ammonium/min. Results given are the average of two or more independent determinations. RESULTS Only strains producing high acetamidase levels can grow strongly on acrylamide as the sole nitrogen source (HYNESand PATEMAN 1970a). Strains of 141 CONTROL-REGION MUTATIONS IN A . NIDULANS genotype umdR4b; amdA7 are unable to use acrylamide as a nitrogen source (HYNES1978a). The amdI66 mutation was found in a N-methyl-N'-nitro-nitrosoguanidine induced mutant by selecting for acrylamide utilization in a strain of genotype amdM4; amdA7. Initial outcrosses showed that the mutant retained both amdR4h and amdA7 lesions. Crosses to amdS-; amdA7 strains showed that the amdI66 mutation responsible for allowing acrylamide utilization segregated in a 1:l ratio with the amdS mutation. A cross between a strain of genotype amdI66; amdA7 to an amdSl2 mutant gave one-quarter progeny capable of strong growth on acrylamide medium (genotype, amdI66; amdA7), one-quarter progeny with a similar amide utilization phenotype to wild type (genotype, amdI66; amdA +) and one-half of the progeny showed the acetamide nonutilization phenotype of amdS12. These results indicated that the amdI66 mutation was close to the amdS gene and increased acetamidase levels only in the presence of the amdA7 mutation. Growth properties of double mutants: A wide variety of double and triple mutants containing amdI66; amdA7 and other mutations affecting umdS expression were constructed. When necessary these were outcrossed to appropriate strains to check for the presence of the amdI66 mutation. Results of growth tests on amide media for these strains are shown in Table 1. Provided both amdI66 and amdA7 mutations were present, growth on amide media was very strong. By itself the amdI66 mutation had no detectable effect compared with the wild-type strains. No interactions between the amdI66 mutation and mutations in the regulatory genes amdR and areA were observed. That the effect of the amdI66 lesion was not specific for the amdA7 allele TABLE 1 Effects of the amdI66 mutation on growth on amide media" Growth medium 1%glucose Relevant genotype amd/R44; amdA7 amdRh; amdI66; amdA7 amdA7 amdI66; amdA7 Wild type amdI66 amdRh amdR44; amdI66 areA102 amdI66 areA102 amdI66 areA102; amdA7 areA217 amdI66 areA217 amdI66 areA217; amdA7 oreA217 amdA7 amdA301 amd166; amdA301 amdA4 amd166: amdA4 ~~ - mM + 10 acrylamide - +++ + +++ - - ++ ++ +++ - - ++ + ++ + ++ 1%sucrose + 10 mM acetamide ++ ++++ +++ ++++ + + * * ++++ ++++ ++++ - +++ ++ +++ ++++ +++ ++++ 50 mM acetamide + ++++ +++ ++++ + + ++- + + ++++ + + ++++ ++ ++ +++ ++ +++ ~ "Growth represented by symbols in order of increasing growth: -, f,+, Symbols on different media are not equivalent. ++, +++, ++++. 142 M. J. HYNES was shown by the increased amide utilization of the double mutants amdI66; amdA3OZ and amdI66; amdA4 compared to the amdI66, amdA301 or amdA4 single mutants. Acetamidase activities: As expected from their growth properties, mutants containing both amdI66 and amdA7 mutations were found to have very high acetamidase activities in media lacking added inducers (Table 2). Responses of the high constitutive enzyme level to repression by ammonium and induction by p-alanine, benzoate or acetate were still observed, compatible with the independence of amdA-mediated regulation from other regulatory mechanisms (see HYNES1978a). By itself the amdI66 mutation led to only slightly higher enzyme activities (particularly in glucose-containing media) compared to the wild-type strain (Table 2). This effect was clearly too small to result in observable growth differences in the growth tests. When the amdl66 mutation was combined with either the amdA302 or amdA4 mutations in double mutants high acetamidase activities were observed, indicating that the interaction was not specific for the amdA7 mutation. However, the effect of these amdA alleles was not as great as for amdA7, indicating allele specific interaction with amdI66. Cis-dominance of the amdI66 mutation: In order to construct diploids with the amdI66 mutation in cis to amdS mutations, it was necessary to isolate amdS mutations in strains of genotype amdI66; amdA7. Such mutants were isolated by selecting for fluoroacetamide resistance (HYNESand PATEMAN, 1970b) and were shown to contain amdS mutations by fine-structure mapping and complementation tests (see below), with known amdS alleles. One such strain of genotype amdS401 amdI66; amdA7 was used to construct various heterozygous diploids that were growth tested on amide media and assayed (Table 3). As previously described (HYNES1978a) the amdA7 allele was semi-dominant to wild type for its effects on amide utilization. The amdI66; amdA7 combination was found to also be semi-dominant to the amdl' and amdA+ alleles, but only when amdI66 was in the cis position relative to a wild-type amdS+ allele. Similar results were found using amdS alleles other than amdS402 (data not TABLE 2 Acetamidase activities of amdI66-containing strains Acetamidase specific activity with Relevant genotype amdRz4; amdA7 amdRT4; amd166; amdA7 amdA7 amd166; amdA7 Wild type amd166 amdI66; amdA301 amd166; amdA4 1% glucoseD+ 1% glucasea 10mM 20mM alanine NH4+ 24 341 41 295 13 21 133 187 + 3 82 8 102 0 11 48 42 10 mM SOdium dium acebenzoateb tate* 10 mM SO- No addition* 10 mMP- 54 44 355 439 1002 186 127 ND ND 364 75 448 25 35 ND' ND alanine* 127 587 210 619 96 82 ND ND 95 672 115 725 60 89 ND ND "Mycelium grown for 16-18 hr at 30" on this medium before harvesting, extraction and assay. Mycelium grown for 16 hr at 30° in 1%glucose + 20 mM NH,+ medium and then transferred to minimal medium with the indicated additions for 4 hr before harvesting, extraction and assay. ND = not determined. ' 143 CONTROL-REGION MUTATIONS IN A . NIDULANS TABLE 3 Cis-dominance of amdI66 Growth on" acetamide Acetamidase specific activity with 1% glucose + 10 mM L-alanine* + ++ ND" amd166; amdA7 amdI+; amdA' ++ +++ ND amdI66amdS'; amdA7 amdl+amdS17; amdA+ ++ +++ 173 amd166amdS401; amdA7 amdl+amdS+amdA' + + 12 amd166amdS401; amdA7 amdl+omdS+;amdA7 + ++ 34 amdl66amdS+; amdA7 amdPamdS223; amdA7 +++ ++++ Relevant genotype of diploid amdA7 amdA7 1%glucose + 10 mM acrylamide 1%sucrose + 10 mM 223 Growth symbols as for Table 2. * Mycelium grown for 16 hr before harvesting, extraction and assay. e ND = not determined. shown). Therefore, the dominance properties of the amdZ66 mutation were compatible with it defining a cis-acting site affecting amdS gene regulation. Fluoroacetamide resistant mutants of amdI66; amdA7 strains: Seven spontaneous revertants showing reduced amide utilization were isolated from the amdRq4; amdI66 amdA7 strain by selection for fluoroacetamide resistance. Two distinct classes were observed. Three of the revertants grew as well as amdR44 and amdR44; amdI66 strains on amide media and did not segregate the amdA7 mutation in outcrosses to wild-type strains. They retained the amdI66 mutation because in crosses to amdS; amdA7 strains they segregated one-quarter progeny capable of strong growth on acrylamide medium. Apparently these mutants arose as a result of reversion events at the amdA locus. For two of these mutants scoring of large numbers of progeny from outcrosses to a wild-type strain showed a low proportion of amdA7 segregants. This indicated that mutation at a second site within the amdA locus could reverse the amdA7 phenotype. Further revertants of amdA7 have been obtained in subsequent experiments. So far no amdA7 revertant isolated in this way has been shown to have an observable effect on wild-type amide utilization or to have any other phenotypic effect. The remaining four mutants showed a complete loss of ability to utilize amides. Segregation of low frequencies of progeny capable of strong growth on acrylamide medium in outcrosses to wild-type or amdA7 strains showed that they retained the amdI66 and amdA7 mutations. Complementation tests with amdS alleles and fine structure mapping (see below) showed that these mutants contain mutations in the amdS region leading to loss of enzyme activity. 144 M . J. HYNES Fine structure mapping of amdI66 in relation to the amdS gene and its controlling region: Using the methods and strains described previously (HYNES 1979) three of the new amdS region mutants isolated in the amdI66; amdA7 background were investigated in detail in order to map the amdI66 mutation relative to amdS. Figure 1 gives the orientation of the genetic map. Mapping experiments with defined amdS deletion mutants showed that one of the mutants had a lesion (designated amdS401) in the left hand region of the amdS gene, and clearly mapped to the left of amdI66 (cross 1 in Table 4). The other two strains were found to have lesions in the right hand end of the amdS region. Because it was not clear whether these mutations were inside or outside the amdS coding region they were designated amd-406 and amd-407 in accordance with previous practice (HYNES1979). The marker mutation cbxA17, which is on the left-hand centromere proximal side of the amdS gene, was used to order the mutations relative to defined point mutations and to amdI66 (Table 4). The amdS224 mutation is the most rightward confirmed structural gene mutation (HYNES1979). Both amd-406 and amd-407 were mapped to the right of this site (crosses 6 and 7 in Table 4 ). The amd-406 mutation yielded a low frequency of amdS+ recombinants with the amd-205 mutation (an apparent promoter-down mutation-see HYNES1979) but has not been observed to recombine with amd407. Recent physical mapping studies employing Southern blot analysis with amdS DNA probes have shown that amd-406 and amd-407 are small deletions (M. J. HYNESand C. M. CORRICK, unpublished). amd-407 clearly was well to the right of the amd-205 site (crosses 4 and 5 in Table 4). The amd-407 mutation therefore has defined the most rightward site capable of mutation to loss of amdS function so far isolated. The amd166 mutation was found to be located to the right of amd-407 and amd-406 (crosses 2 and 3 in Table 4). Subsequent mapping of further amdS region mutations has so far failed to show any new mutations mapping to the right of either the amd-407 or amdI66 mutations (M. J. HYNES,unpublished data). Frequencies of amdS+ recombinants in crosses between several of the mutants were accurately calculated by determining total numbers of ascospores plated. Assuming equal frequencies of undetected double mutant recombinants the following map distances were obtained: amd-406-andS224 = 0.03 map units; amd-407-amdS224 = 0.15 map units and amd-407-amdS222 = 0.65 map units. These results were comparable with those obtained previously (HYNES 1979). The amdI93 mutation, which specifically affects a-amino acid induction s222 cbxA I I Y sun 205 h 5224 1 l!B I18 407 166 I 1 406 --- FIGURE1.-Map order of mutations at the right hand end of the amdS gene. No recombination has been observed between amd-205 and amd-I18 (HYNES1979). The amd193 mutation is to the right of amdll8, but its position with respect to amd-407 and amdI66 is unknown. The amd-406 mutation appears to be a small deletion. The amdS401 mutation has not been ordered with respect to amdS222. 145 CONTROL-REGION MUTATIONS IN A. NIDULANS TABLE 4 Fine structure mapping of amdI66 No. of recombinantsa cross Relevant genotypes in cross cbxA17 cbxA+ 1 2 amdS401; amd166; amdA7 x cbxA17 omd-407 omd166; amdA7 X cbxA17; amdA7 amd-406 omdI66; amdA7 X cbxA17; amdA7 cbxA17amd-407 amdI66; amdA7 X amd-205; amdA7 amd-407 amdI66; amdA7 X cbxA17omd-205; amdA7 cbxA17amd-407 amdI66; amdA7 x amdS224 amdlld; omdA7 cbxA17amd-406 amd166 X amdS224 amd118; amdA7 amdI18 amd193; omdA7 x cbxA17amd-205; amdA7 22 13 14 21 0 45 23 1 2 0 3 3 4 5 6 7 8 8 6 5 3 38 Recombinants were selected an either glucose-acrylamide medium (to select the genotype amdS+amdI66; amdA7 in crosses 1, 2 and 3 or amdS+amdI18; amdA7 in cross 8) or sucroseacetamide medium (to select amdS+ recombinants in crosses 4, 5, 6 and 7). Recombinants were purified and tested for the presence of the cbxA17 allele by determining carboxin resistance. The principles and details of amdS mapping were as in HYNES1979. mediated by the amdR gene product, has been described (HYNES1980). This mutation has been observed to recombine at a low frequency with amd-407. A cross with a strain of genotype cbxAZ7 amd-207 allowed further mapping of amdI93 within the controlling region (cross 8 in Table 4). It is clear that the amdI93 mutation was to the right of amdI28 and amd-205. Because of their different phenotypes it has not been possible to map amdI93 with respect to amdI66. The map order of mutations at the right-hand end of the amdS gene determined in this study and the previous one (HYNES1979) is shown in Figure 1. DISCUSSION The amdI66 mutation described here represents an especially interesting addition to the variety of cis-acting control mutations adjacent to the amdS structural gene. This mutation only exerts a strong effect on gene expression if a mutation in the unlinked amdA gene is present. No obvious interactions with other proposed trans-acting regulatory genes affecting amdS expression have been observed. Therefore, it appears that the amdI66 mutation alters a receptor site for the control signal generated or modified by amdA mutations without detectably affecting responses to the regulation mediated by other regulatory genes. The role of the amdA gene in regulating amdS expression is not clear. There is no evidence for the wild-type amdA gene product affecting amdS because no amdA mutations resulting in reduced expression have been isolated. This may result from amdA effects being weak in comparison to other control signals. Alternatively, loss of function mutations in amdA could be lethal. The phenotype of deletion mutations in amdA would be of great interest. It is possible that the wild type amdA gene normally does not play a role in amdS regulation, but is involved in regulation of other genes. Evidence for an unknown polypeptide under amdA control has been obtained recently (see Introduction). The amdA mutations that lead to increased acetamidase synthesis could alter the 146 M. J. HYNES specificity of the amdA gene product so that it now recognizes a DNA sequence adjacent to the amdS gene and stimulates gene expression. On this hypothesis amdI66 would alter this sequence so that the amdA gene product has a greater effect. Until more information is available on the role of the amdA gene it is not possible to distinguish between these possibilities. The interaction between the amdI66 mutation and mutations in the amdA gene is the third case of specificity between controlling region mutations and one of the control mechanisms affecting amdS expression. The amdI93 mutation specifically affects a-amino acid induction and interacts with mutations in the amdR gene (HYNES 1980), and the amdI9 mutation causes increased inducibility by acetate that is probably mediated by the facB gene product (HYNES 1977; M. J. HYNESand J. M. KELLY,unpublished data). These mutations provide strong evidence for independent control mechanisms operating by a n interaction between the product of a regulatory gene and a specific site adjacent to the structural gene. No data regarding the molecular level at which gene expression is affected by the amdI66-amdA interaction is presented in this paper. It is formally possible that the amdI66 mutation affects the primary structure of the amdS-coded polypeptide and the altered product of mutant amdA genes interacts with the altered amdS product to increase enzyme activity. However, data based on separation of labeled polypeptides on polyacrylamide gels, immunological titration of the amdS product and determinations of translatable messenger RNA levels indicate that regulation is at the level of gene activity (M. J. HYNES, unpublished data). amdS DNA has been recently cloned, using recombinant DNA techniques, and is being used to directly measure mRNA levels and in physical mapping of the controlling region (M. J. HYNESand C. M. CORRICK, unpublished data). The results of these studies on the molecular nature of regulation of the amdS gene will be the subject of future communications. This research was supported by the Australian Research Grants Committee LITERATURE CITED ARST,H. N. and D. J. COVE,1973 Nitrogen metabolic repression in Aspergillus nidufans. Mol. Gen. Genet. 126 111-141. ARST,H. N., K. N. RANDand C. R. BAILEY,1979 Do the tightly linked structural genes for nitrate and nitrite reductases form an operon? Evidence from an insertional translocation which separates them. Mol. Gen. Genet. 174: 89-100. CLUTTERBUCK, A. I., 1974 Aspergillus nidulans genetics. pp. 447-510. In: Handbook of Genetics, Vol. 1,Edited by R. C. KING.Plenum Press, New York. COVE,D. J., 1966 The induction and repression of nitrate reductase in the fungus Aspergillus nidulans. Biochem. Biophys. Acta 113: 51-56. ERREDE, B., T. S. CARDILLO, F. SHERMAN, E. DUBOIS, J. DESCHAMPS and J.-M. WIAME,1980 Mating signals control expression of mutations resulting from insertion of a transposable repetitive element adjacent to diverse yeast genes. Cell 22: 427-436. HYNES,M. J., 1972 Mutants with altered glucose repression of amidasc enzymes in Aspergillus nidulans. J. Bacteriol. 111:717-722. HYNES,M. J., 1977 Induction of the acetamidase of Aspergillus nidulans by acetate metabolism. J. Bacteriol. 131: 770-775. HYNES,M. J., 1978a Multiple independent control mechanisms affecting thc acetamidasc of Aspergillus nidulans. Mol. Gcn. Genet. 161: 59-65. CONTROL-REGION MUTATIONS IN A. NIDULANS 147 HYNES, M. J., 1978b An “up-promoter’’mutation affecting the acetamidase of Aspergillus nidulans. Mol. Gen. Genet. 166 31-36. HYNES,M. J . , 1979 Fine structure mapping of the acetamidase structural gene and its controlling region in Aspergillus nidulans. Genetics 91: 381-392. HYNES,M. J., 1980 A mutation, adjacent to gene amdS, defining the site of action of positive control gene amdR in Aspergillus nidulans. J. Bacteriol. 142 400-406. 1970a The genetic analysis of regulation of amidase synthesis in HYNES,M.J.and J. A. PATEMAN, Aspergillus nidulans. I. Mutants able to use acrylamide. Mol. Gen. Genet. 108: 95-106. HYNES,M. J. and J. A. PATEMAN, 1970b The genetic analysis of regulation of amidase synthesis in Aspergillus nidulans. 11. Mutants resistant to fluoroacetamide. Mol. Gen. Genet. 108: 107-116. MCKNIGHT, G. L.,T. S. CARDILLO and F. SHERMAN, 1981 An extensive deletion causing overproduction of yeast iso-2-cytochrome C. Cell 25: 409-419. F. and C. HELMS,1978 A chromosomal translocation causing over-production of iso-2SHERMAN, cytochrome C in yeast. Genetics 88: 689-707. Corresponding editor: R. C. ULLRICH