The expression of the structural genes of the proline utilization cluster of Aspergillus nidulans is repressed efficiently only when both repressing carbon and nitrogen sources are present. Two hypotheses can account for this fact. One is a direct or indirect competition mechanism between the positive‐acting AreA GATA factor, mediating nitrogen metabolite repression, and the negative‐acting CreA protein, mediating carbon catabolite repression. The second is to propose that CreA prevents the binding or activity of another, as yet unidentified, positive‐acting factor, here called ADA. We show the second possibility to be the correct one, and we localize the new positive cis‐acting element within 290 bp of the prnD–prnB divergent promoter.
Proline catabolism is a useful model system for studying the regulation of gene expression in Aspergillus nidulans. In this organism, and differently from Saccharomyces cerevisiae, proline can serve as both a carbon and a nitrogen source. All the genes required to convert external l‐proline to glutamate are clustered in 13 kb on chromosome VII (Arst and MacDonald, 1975; Arst et al., 1981; Jones et al., 1981; Hull et al., 1989; see Figure 1). The cluster comprises the prnA, prnB, prnC, prnD and prnX genes. prnA codes for a transcription factor mediating pathway‐specific induction (by proline) which is unconditionally required for the expression of the prn structural genes. These structural genes are prnD, encoding proline oxidase, prnB, encoding the specific proline permease, and prnC, encoding l‐Δ1‐pyrroline‐5‐carboxylate dehydrogenase. A fifth gene of unknown function, prnX, is also present in the prn cluster (Gavrias, 1992).
In A.nidulans, the utilization of carbon and nitrogen sources is under the control of, respectively, CreA, a negative‐acting zinc finger protein (Bailey and Arst, 1975; Dowzer and Kelly, 1991), and AreA, a positive‐acting transcription factor belonging to the GATA family (Arst and Cove, 1973; Kudla et al., 1990). Expression of genes involved in the utilization of less favoured carbon sources is repressed by glucose, irrespective of the nitrogen nutrient present in the medium or of the areA allele present in the strain. Good examples of glucose‐repressible genes are the genes of the ethanol regulon (Mathieu and Felenbok, 1994). Conversely, expression of genes involved in the utilization of compounds which serve only as nitrogen sources is repressed by ammonium or glutamine irrespective of the carbon source or of the presence of a creAd (carbon catabolite‐derepressed) mutation (Arst and Cove, 1973; Bailey and Arst, 1975). Examples of such genes are those involved in the utilization of nitrate (Cove, 1966; Arst and Cove, 1973) and of purines (Gorfinkiel et al., 1993; Oestreicher and Scazzocchio, 1993).
The expression of the prn cluster is repressible by the simultaneous presence of ammonium and glucose. The prime target of glucose/ammonium repression is the permease gene prnB. Although synthesis of the prnD and prnC products is also repressed, the limiting step for proline utilization under repression conditions is proline uptake (Arst et al., 1980). prnd mutations (derepressed, see below), map in the prnD–prnB intergenic region and result in derepression of prnB and also of prnD and prnC (Sophianopoulou et al., 1993; see Discussion).
Glucose or ammonium on their own reduce the expression of prnB, prnD and prnC, but full repression requires the simultaneous presence of repressing carbon and nitrogen sources (Gavrias, 1992). Strains carrying areA loss‐of‐function mutations (areA−) are unable to utilize proline as a source of nitrogen if a repressing carbon source (e.g. glucose) is present, but grow on proline in the presence of non‐repressing carbon sources (Arst and Cove, 1973). This implies that the AreA transcription factor is only required for the expression of the prn structural genes in the presence of an active CreA protein. The pattern of transcriptional regulation of the prn cluster makes teleonomic sense. It allows the use of proline as a nitrogen source even if a rich carbon source is present and as a carbon source even if a rich nitrogen source is present. The utilization of some other metabolites, which can serve as both carbon an nitrogen sources (e.g. acetamide, γ‐aminobutiric acid), is subject to the same pattern of regulation as that of proline (Arst and Bailey, 1977).
A model where AreA is strictly necessary, and CreA acts by preventing its binding or action, is incompatible with both the genetic and physiological evidence. If this were the case, either ammonium or glucose alone would repress prn expression, and areA− strains would be unable to use proline, irrespective of the carbon source. Two models could account for the pattern of repression of the prn genes. The first is that AreA acts by preventing the binding or repressing activity of CreA. We shall call this model the ‘AreA/CreA competition model’, without pre‐judging whether the effect of AreA on CreA binding or activity is direct or indirect. An analogous mechanism, whereby GATA‐1 prevents the binding of a negative‐acting factor has been demonstrated for the glycophorin B gene (Rahuel et al., 1992). In the second model, the transcription of the prnB gene can be elicited by either AreA or, alternatively and only somewhat additively, by an unidentified positive‐acting element whose binding or activity is prevented by the CreA protein in conditions of carbon catabolite repression. This model may also be valid for the prnD and prnC genes (Scazzocchio, 1992; Sophianopoulou et al., 1993). In fact, while in the first model AreA is preventing the action of CreA, in the second model it is by‐passing it. In spite of their formal similarity, these two models make strictly different predictions. In the AreA/CreA competition model, the only function of AreA is to prevent, directly or indirectly, the repressing activity of CreA. Thus, a deletion removing the physiologically relevant CreA sites should be derepressed, in every case, except in those in which other essential elements are removed, in which case no transcription will be seen. In the second model, a deletion not removing other essential elements will have two different phenotypes, according to whether the hitherto unidentified positive‐acting element is removed or not. If the positive‐acting element is not removed, the phenotype will be derepressed, exactly as in strains carrying point mutations of the CreA‐binding sites (Sophianopoulou et al., 1993; Cubero and Scazzocchio, 1994). If it is removed, and the physiological AreA‐binding sites are not, the promoter will become repressible by ammonium on its own, and strictly dependent on AreA, irrespective of whether the organism is grown under carbon repressing or derepressing conditions or whether the strain carries a creA+ or a creAd allele. The possibility of generating an AreA strictly dependent phenotype is only compatible with the second model.
Here we establish that an AreA/CreA competition model is untenable and that CreA acts by preventing another, hitherto unidentified, positive element from eliciting prnB transcription. We call this element ADA (absolute dependence on AreA). This hypothesis, but not the experimental evidence, has been presented previously (Scazzocchio, 1992).
Structure of the prnD–prnB intergenic region
In Figure 1, we show schematically the 1672 bp intergenic region, between the initiation codons of genes prnB (Sophianopoulou and Scazzocchio, 1989) and prnD (Gavrias, 1992). This sequence has been submitted to the DDBJ/EMBL/GenBank database under accession No. U74465. The direction of transcription of prnD has been determined by hybridization of wild‐type RNA with single‐stranded probes (data not shown). At variance with a previous report (Hull et al., 1989), prnD is transcribed from the prn intergenic region divergently from prnB. The initiation codon shown for prnD in Figure 1 is the most 5′ in the transcribed sequences (see below) and also the one which maximizes the protein similarity of prnD to PUT1, the isofunctional protein of S.cerevisiae (Wang and Brandriss, 1987). Although no sequences corresponding to a TATA box are present in the prnD gene promoter, the sequence TCATTAAT at position −75 is similar to the sequence CATTTAAT, which can act as a TATA box for the yeast transcription factor TFIID (Singer et al., 1990). The sequence, transcription initiation point and possible promoter structure of prnB have been published (Sophianopoulou and Scazzocchio, 1989). The region contains seven actual CreA‐binding sites (shown in Figure 1), of which two have been shown to be necessary for repression of the gene cluster (see figure legend) (Sophianopoulou et al., 1993; Cubero and Scazzocchio, 1994), and 17 GATA sites of which 14 follow the HGATAR consensus established for the AreA protein (Ravagnani et al., 1997). There are also 21 GATT and 16 GATG sites which may bind AreA, mainly if another, canonical, strongly binding site is adjacent (D.Gomez, C.Scazzocchio and B.Cubero, unpublished). Only true GATA sites are shown in Figure 1. The binding ability and role of the different putative AreA‐binding sites will be published elsewhere.
Deletions prn301 and prn304 and construction of prn+ strains carrying a deleted prnD–prnB intergenic region
Two previously characterized deletions have proven useful to generate strains suitable for distinguishing the two possible models discussed in the Introduction. Deletion prn304 is a prnD–prnB–prnC deletion by complementation and recombination criteria (Arst et al., 1981; Hull et al., 1989). The relevant portion of the prn cluster of a strain carrying this mutation was sequenced. It is a 3479 bp deletion, from position −284 of prnD to 373 bp downstream of the last codon of prnB, 310 bp before the ATG of prnC (Figure 2). Northern blots show that deletion prn304 results in almost non‐detectable transcription of prnC and reduced, but clearly inducible and ammonium‐repressible, transcription of prnD (not shown, see below). This deletion was used to generate two types of strains altered in the prnB promoter: (i) strains carrying the prnB and prnC genes in an ectopic position; and (ii) similar strains generated by homologous recombination at the prnC gene. A prn304pabaA1riboB2 strain was co‐transformed with plasmids pAN225 and pPL5 (see Materials and methods for the structure of these plasmids). Transformants were selected for complementation of the riboB2 mutation and tested for growth on proline as sole nitrogen source. A number of transformants were able to utilize proline, albeit at a reduced rate when compared with a wild‐type strain. The fact that such transformants were obtained at all indicates that the low level of prnD transcript is sufficient to allow some proline utilization (see below). Several transformants were characterized by Southern blot analyses, and two of them, P1 and PR4, which contain single‐copy integrated plasmid pAN225, were selected for transcript analysis (see below). Transformant P1 contains the prnB–prnC EcoRI fragment inserted homologously in the prn gene cluster, while transformant PR4 contains the same construction inserted at an unknown chromosomal location (Figure 2). The riboB+‐containing plasmid pPL5 has integrated independently from the pAN225 plasmid in both P1 and PR4. Both transformants have an intact prnB–prnC intergenic region and have the same sequences upstream of prnD as the original deletion strain. The physiological CreA‐binding sites are not present. These transformants carry two copies of the prnC gene, the copy present in the recipient strain, which should be virtually non‐transcribed, and a second one under the control of the intact prnC promoter contained in the transforming plasmid.
Deletion prn301 has been shown by both genetic and physical mapping to terminate within the prnD and prnB open reading frames (ORFs) (Arst et al., 1981; Hull et al., 1989). Sequencing has shown that it is a 3058 bp deletion with boundaries at positions +459 of prnD and +946 of prnB (from the ATGs of each gene, Figure 2). This deletion is a null mutation for both prnB and prnD, and thus results in an inability to use proline as nitrogen and/or carbon source under any conditions. This deletion has been used to generate strains with altered prnD–prnB intergenic regions. A strain carrying this deletion has been transformed with DNA fragments containing prnD and prnB genes truncated at their 3′ ends, so that transformants selected for growth on proline can only be generated by gene conversion (or double cross‐over). This is a way to introduce any modification in the intergenic region, at the original chromosomal location, with the only proviso that prnD–prnB intergenic regions not restoring the growth on proline under at least some conditions cannot be obtained by this method. Strains carrying deletions prn942 and prn943 have been constructed using this system. The missing sequences in the prnD–prnB intergenic region in these two strains are almost identical to those in transformants P1 and PR4. The left boundaries of these two deletions were chosen so as to bracket that of deletion prn304 and, consequently, that of transformants P1 and PR4 (Figure 2). The second non‐transcribed copy of prnC and the heterologous plasmid sequences are absent in strains carrying deletions prn942 and prn943. No expression of the prnD, prnB and prnC genes is seen in any of these four transformant strains in the absence of proline induction, exactly as for the wild‐type strain (results not shown).
The prnB 5′ truncated promoters extant in transformants P1, PR4, prn942 and prn943 lack ADA: genetic evidence
A diagnostic test for the presence of the ADA element is as follows. A strain carrying a prnB wild‐type promoter (ADA+) and an areA− null allele will use proline as sole carbon and nitrogen sources or proline as nitrogen source in the presence of non‐repressing carbon sources (such as ethanol). However, such a strain will be unable to use proline as nitrogen source when a repressing carbon source (glucose) is present. A strain carrying the wild‐type promoter will use proline under all conditions if, together with an areA null mutation, a creAd mutation is also present in the strain. In the presence of an areA null allele, an ADA− strain will be unable to use proline under any conditions, regardless of whether a creAd mutation is also present in the strain. We have constructed P1areA600, PR4areA600, prn942areA600 and prn943areA600 double mutants. In each case, the introduction of an areA null mutation resulted in inability to use proline under any conditions. The results are illustrated in Figure 3 and summarized in Figure 4. We then crossed each of the areA600‐containing strains to a creAd1 strain. While the strain containing the wild‐type promoter in an areA600creAd1 context is able, as expected, to use proline as nitrogen source in the presence of glucose, no such ability is present in the deleted strains (data not shown). We can deduce that all four transformant strains have lost ADA.
Fine mapping of the ADA element
Using deletion prn301 and the methodology described above, we have constructed strains carrying a series of deletions in the prnD–prnB intergenic region. Each of the deletion strains was crossed with areA600. The results of the whole deletion analysis are shown in Figure 4. We include results with strains where the two relevant CreA‐binding sites were eliminated by point mutation (prnd20prnd22), with a small deletion encompassing both physiologically important CreA‐binding sites (prn927) and with the large deletions discussed above. Figure 3 illustrates the phenotypes associated with three significant deletions. prn943 has the characteristics expected for an ADA− mutation as described above. prn968 is the largest of a series of deletions which remove the CreA‐binding sites, and whose phenotype is similar to that of point mutations in the CreA sites (prnd20prnd22). ADA activity is not affected in these strains. We can conclude that ADA maps upstream of the prnD proximal border of prn968. prn945 (not shown in Figure 3) is a small deletion, adjacent to prn943 (Figure 4), which has no effect on proline utilization. It brackets the left boundary of the ADA element. Deletion prn970 has the expected ADA− phenotype, locating the ADA element within a 290 bp DNA fragment. Thus the ADA element is located between nucleotides 369 and 657 in the sequence GenBank U74465.
Deletion of the ADA element: transcriptional evidence
Figure 5 confirms, using mRNA steady‐state levels, earlier data obtained measuring proline uptake and proline oxidase activity (Arst and MacDonald, 1975; Arst et al., 1980). These data indicate that, in the wild‐type strain, full repression necessitates the simultaneous presence of repressing carbon and nitrogen sources. A further prediction of the ‘ADA’ model is that prnB promoters lacking ADA will be repressed by ammonium, irrespective of the carbon source. This prediction is borne out for a number of the deletion strains described above (Figure 5). We have tested transformants P1 and PR4 and strains carrying deletions prn943, prn942 and prn970. In these transformant strains, prnB and prnD transcripts are inducible. As in the wild‐type strain, no transcript is detectable in the absence of proline induction (results not shown). For all transformants, the steady‐state levels of prnB and prnD were determined under non‐repressing conditions, under carbon‐repressing conditions, under nitrogen‐repressing conditions and under carbon‐ and nitrogen‐repressing conditions. Figure 5 shows the results for PR4 (lower panel) and for transformants prn943 and prn970 (upper panel). Results for P1 and prn942 (data not shown) are equivalent to those for PR4 and prn943 respectively.
The transcription of prnD is, under inducing, non‐repressing conditions, as low as it is in the original prn304 deletion in all the deletion strains tested in Northern blots (see below for discussion of prn970). Surprisingly, an increase of transcription of prnB in the presence of glucose is seen for transformants P1 and PR4. No such increase is seen in strains carrying deletions prn942 and prn943. In every one of the deleted strains, including prn970 carrying the 290 bp deletion, ammonium is by itself a repressor of prnB and prnD independently of the carbon source, as predicted for mutants lacking the ADA element.
We then compared repression by glucose in the absence of a repressing nitrogen source in strains which lacked the ADA element in an areA+ and an areA600 background. We compared the prnB and prnD mRNA levels in strains carrying deletions prn970 and prn943 with those in strains carrying an intact prnD–prnB intergenic region. This experiment is the ‘mRNA equivalent’ of the growth tests shown in Figure 3 and summarized in Figure 4. The results (Figure 6) are strikingly in accord with the predictions. An areA+ strain shows marginal repression by glucose (as shown before, Figure 5) and an areA600 strain shows repression by glucose; in both deletion strains, transcript levels are not repressed by glucose. In an areA600 background, both deletion strains show, irrespective of whether a repressing or non‐repressing carbon source is included in the culture medium, levels of transcription which are equivalent or lower than those observed for an areA600 single mutant in the presence of glucose. This is what was predicted; repression by glucose in an areA600 background is equivalent to repression by ammonium and glucose in an areA+ background; the ADA element is dispensable in the presence of areA+ in the absence of a repressing nitrogen source, but indispensable in an areA600 background. We note, however, that while the mRNA steady‐state levels of prnB and prnD are clearly dependent on the presence of an active areA gene product for both deletions tested, the absolute levels are higher for the longer deletion. This was already observed in Figure 5 (see Discussion).
It was to be expected that the induced, non‐repressed level of expression of prnD in transformants P1, PR4 and in those carrying deletions prn942 and prn943 would be as low as those found in the original prn304 deletion, as the same (for P1 and PR4) or almost the same sequences are missing from the promoter. Deletion prn970 shows that within the 290 bp which define the ADA element, there are sequences which are important to set the maximal level of prnD. What is quite surprising is that the induced non‐repressed level of expression of prnB in the small deletion prn970 is significantly lower than in the wild‐type and the large deletions prn943 and prn942 (the latter not shown).
We have proposed two different models to explain the repression pattern of the prn cluster in A.nidulans. These models differ in the role of the AreA protein. Strains carrying our larger deletions are sensitive to ammonium repression and necessitate an areA+ allele for prnB (and prnD) expression. This implies that the few AreA potential binding sites left in deletions prn304 (and its P1 and PR4 derivatives), prn943 and prn942 are sufficient for the promoter(s) to respond to AreA transcriptional activation. Note that deletion prn946, which results in an in vivo phenotype identical to the above‐mentioned deletions, leaves only five canonical AreA‐binding sites in the prnD–prnB intergenic region (see Figure 1). Work which will be published elsewhere shows that these sites bind both to a GST–AreA fusion protein and, for those which have been tested, to native A.nidulans AreA protein (D.Gomez, C.Scazzocchio and B.Cubero, unpublished).
The deletion analysis carried out in this study shows conclusively that besides AreA, a second positive‐acting element is operative in the prnD–prnB intergenic region, and that CreA works by preventing its action. Every single prediction of the ADA model has been confirmed by this deletion analysis. Deletion of ADA results, as predicted, in (i) unconditional dependence on AreA and (ii) repression by ammonium whether the carbon source be repressing or derepressing. The promoter of prnB (and prnD, but see below) behaves, when ADA is deleted, as do the promoters of, for example, nitrate and nitrite reductase or urate permease. These are only repressible by ammonium and are indifferent to carbon source (Cove, 1966; Arst and Cove, 1973; Gorfinkiel et al., 1993; Oestreicher and Scazzocchio, 1993).
We have used a dual approach to define ADA. We have analysed the same ADA− deletion in three different genomic locations and carried out nested deletions in an otherwise unmodified prn cluster. The agreement of both approaches is satisfying and precludes artefacts due to chromosomal position effects. The comparison of the strains carrying deletions prn942, prn943 and transformants P1 and PR4 is illustrative. All these constructions contain a prnB promoter truncated at the EcoRI site (Figure 2). prn942 and prn943 strains have a deletion and no other modification of the prn gene cluster, except that the distances between other remaining cis‐acting elements (e.g. AreA‐ and PrnA‐binding sites) are altered. P1 contains the same prnB truncated promoter in the same chromosomal location, but separated from the truncated prnD promoter by ∼8 kb, including a non‐transcribed prnC gene and the plasmidic sequences. PR4 has a prnB gene with a truncated promoter at an unknown location in the genome. Besides the unexplained glucose dependence found in P1 and PR4, which may be due to plasmidic sequences absent in prn942 and prn943, the response of the three topographically different constructions is the same: the three behave as predicted by a deletion of ADA. The nested deletion approach has finally allowed mapping of the ADA element within the limits of 290 bp of the prnD–prnB intergenic region. Finally, we have shown the equivalence of the genetic and transcriptional analyses by testing the mRNA levels of critical deletions defining the ADA element, prn970 and the large deletion prn943 in an areA600 background. As expected, these double mutants, which cannot utilize proline under any conditions, also have very low levels of prnB and prnD mRNA irrespective of the carbon source.
Within the 290 bp deletion prn970, there is an element or elements that set the maximum induced level of both prnB and prnD. This is not a PrnA‐binding site for prnB, as all larger overlapping deletions tested are normally inducible for this gene (see below). It cannot contain an essential PrnA‐binding element for prnD, as this and larger deletions are still inducible, albeit at a lower level than in the wild‐type. This element is conceptually different from ADA, but may be overlapping with it. It is surprising that prn970 has an effect on the induced, non‐repressed level of prnB and that this effect is not seen in larger deletions, completely overlapping with prn970. This may indicate the presence of still other, unknown negative‐acting factors in this complex promoter, which must be counteracted by sequences in the 290 bp fragment defined by prn970. What is relevant here is that the pattern of repression is identical in all these deletions and this fact is not altered by this apparent anomaly.
The pattern of repression of prnD is identical to the pattern of repression of prnB. Two mechanisms may account for this fact. The whole intergenic region may be acting bi‐directionally on both genes. Alternatively, the repression of prnD may be due to inducer exclusion mediated through the repression of prnB. The latter mechanism is also sufficient to account for the repression pattern of prnC. These two mechanisms are not exclusive. It is relevant here that cis–trans tests using prnd mutations have shown that the expression of prnB is the limiting factor affecting growth on proline under conditions of repression (Arst et al., 1980).
The model proposed here might be quite general; other metabolites (acetamide, γ‐aminobutiric acid) are utilized as the nitrogen source in an areA− background if a non‐repressing carbon source is present (Arst and Cove, 1973; Arst and Bailey, 1977). Recent work has identified areA and creA homologues in a number of filamentous fungi (Fu and Marzluf, 1990; Drysdale et al., 1993; Haas et al., 1995; Strauss et al., 1995), and it is thus likely that the integration mechanism proposed here is not limited to A.nidulans. An apparently similar pattern of repression has been described for the gene encoding the NAD‐dependent glutamate dehydrogenase of S.cerevisiae (Coschigano et al., 1991). However, the mechanism proposed to account for this pattern is completely different from the one proposed here (Miller and Magasanik, 1991).
We purposely have left the nature of ADA undefined. The most obvious possibility is that the 290 bp region contains the binding site for an unidentified transcriptional activator, and indeed preliminary experiments show that A.nidulans nuclear extracts contain proteins which bind the 290 bp fragment. Binding of CreA to both sites shown in Figure 1 (see legend) is necessary for repression (Scazzocchio, 1992; Sophianopoulou et al., 1993; and this study, Figures 3 and 4). Thus, CreA must prevent the binding or activity of an ‘ADA factor’ at a distance of at least 500 bp. However, a less obvious possibility should be considered. This is that there is no trans‐acting factor, but rather that ADA is a sequence which maintains the promoter in a transcriptionally competent ‘open’ configuration, which is ‘closed’ at a distance by CreA, and that binding of AreA to sites near the transcription start points of prnB and prnD can by‐pass this effect. Further work, including the analysis of chromatin structure in the prnD–prnB intergenic region, may distinguish these formally equivalent but mechanistically different possibilities.
Materials and methods
Strains, plasmids and genetic methods
biA1pantoB100 and pabaA1 strains were used as wild‐types. prn304 is a prnD–prnB–prnC deletion (Arst et al., 1981). prn301 is a deletion removing all the prn cis‐acting region and most of the coding regions of prnD and prnB (Arst et al., 1981). creAd1 is a derepressed, partial loss‐of‐function mutation (null mutations in creA are lethal) and areA600 is an early chain termination, null mutation (Bailey and Arst, 1975; Kudla et al., 1990). biA1, pabaA1, riboB2 and pantoB100 are standard auxotrophic markers (Clutterbuck, 1993). pPL5 is a pUC derivative containing a riboB+ gene (Oakley et al., 1987). pAN225 contains a 6.6 kb EcoRI–EcoRI fragment, comprising genes prnB and prnC, in vector pBR322 (Hull et al., 1989). bAN926 contains a 3 kb XhoI–Asp718 fragment, comprising a 907 bp fragment of the prnD ORF, the prnD–prnB intergenic region and 1512 bp of the prnB ORF, in vector Bluescript KS+. Other plasmids were constructed by site‐directed mutagenesis (prnd20prnd22), using standard methods (Kunkel et al., 1991), or by restriction and ligation of bAN926, in order to obtain the corresponding deletions of the intergenic region (see Results). Nucleic acid manipulations followed standard procedures (Sambrook et al., 1989). Genetic methods for A.nidulans were as described by Pontecorvo et al. (1953).
Media and growth conditions
Media and supplements were described by Cove (1966). Inducing and repressing conditions are described in the figure legends.
The sequence of the prnB‐proximal half of the prnD–prnB intergenic region has been published (Sophianopoulou and Scazzocchio, 1989). The prnD‐proximal half was sequenced from subclones of plasmid pAN224 (Hull et al., 1989). Deletions prn301 and prn304 were sequenced from PCR fragments amplified from DNA of the corresponding mutant strains. All DNA fragments were sequenced by the method of Sanger et al. (1977).
This followed Tilburn et al. (1990). A strain prn304pabaA1riboB2 was co‐transformed with plasmid pAN225 and plasmid pPL5. riboB+ transformants were selected and tested for co‐transformation in medium containing 5 mM proline as the sole nitrogen source. To construct the strains used in the mapping of the ADA element, prn301pabaA1 was transformed with the insert of the corresponding bAN926 derivative plasmids. Transformants were selected on medium containing 20 mM proline as the sole source of nitrogen. Transformants were checked by Southern blots, using the 6.3 kb EcoRI–EcoRI fragment of plasmid pAN225 and a 2.2 kb PstI–PstI fragment of plasmid pPL5 (for prn304 transformants) or a 2.2 kb XhoI–SmaI fragment of plasmid bAN942 (for prn301 transformants) as probes.
RNA was isolated from A.nidulans mycelia as described (Chomczynski and Sacchi, 1987; Lockington et al., 1987), glyoxal treated and separated on a 1% agarose gel (Sambrook et al., 1989). Northern blots were carried out by standard techniques, using a PCR‐amplified whole prnD gene, a 1.8 kb PstI–PstI fragment of plasmid pAN225, containing most of the ORF of prnB, or a 2 kb KpnI–SmaI fragment of plasmid pSF5, containing the A.nidulans actin gene (Fidel et al., 1988), as a probe.
We thank George Diallinas and Tania Said for helpful discussions and H.N.Arst,Jr for deletions prn304 and prn301. This work was supported by CEE grants SCI‐CT92‐0815 and BIO2‐CT93‐0147. R.G. has been the recipient of, successively, EMBO fellowship ALTF457‐1993 and CE fellowship BIO4‐CT96‐5010. D.G. is the recipient of a doctoral fellowship from the Ministère de l'Enseignement Supérieur et de la Recherche of the French Government.
- Copyright © 1997 European Molecular Biology Organization