Polyamine Regulation of Ornithine Decarboxylase Synthesis in Neurospora crassa

Abstract

Ornithine decarboxylase (ODC) of the fungus Neurospora crassa, encoded by the spe-1 gene, catalyzes an initial and rate-limiting step in polyamine biosynthesis and is highly regulated by polyamines. In N. crassa, polyamines repress the synthesis and increase the degradation of ODC protein. Changes in the rate of ODC synthesis correlate with similar changes in the abundance of spe-1 mRNA. We identify two sequence elements, one in each of the 5′ and 3′ regions of the spe-1 gene of N. crassa, required for this polyamine-mediated regulation. A 5′ polyamine-responsive region (5′ PRR) comprises DNA sequences both in the upstream untranscribed region and in the long 5′ untranslated region (5′-UTR) of the gene. The 5′ PRR is sufficient to confer polyamine regulation to a downstream, heterologous coding region. Use of the β-tubulin promoter to drive the expression of various portions of the spe-1 transcribed region revealed a 3′ polyamine-responsive region (3′ PRR) downstream of the coding region. Neither changes in cellular polyamine status nor deletion of sequences in the 5′-UTR alters the half-life of spe-1 mRNA. Sequences in the spe-1 5′-UTR also impede the translation of a heterologous coding region, and polyamine starvation partially relieves this impediment. The results show that N. crassa uses a unique combination of polyamine-mediated transcriptional and translational control mechanisms to regulate ODC synthesis.

Ornithine decarboxylase (ODC) catalyzes an initial, rate-limiting reaction in the biosynthesis of the polyamines, the conversion of the amino acid ornithine to the divalent polyamine putrescine (1,4-diaminobutane). Putrescine undergoes two subsequent aminopropyl transfer reactions, in which it is converted first to the trivalent spermidine and then to the tetravalent spermine. Polyamines are essential for the growth of normal, and particularly neoplastic, cell types (25, 38), but excess spermidine and spermine are toxic (8, 28).

ODC is a one of the most highly regulated enzymes of eukaryotic cells, its activity varying over a 100-fold range. ODC activity responds to extracellular signals such as mitogens and growth factors and to changes in the intracellular concentrations of the polyamines themselves. Polyamine regulation of ODC activity is unusual in that the end products do not act as allosteric effectors of this initial enzyme but control only the synthesis and degradation of the ODC protein (reviewed in references 6 and 41).

In the filamentous fungus Neurospora crassa, ODC is encoded by the spe-1 gene (3, 9). The regulation of ODC in this fungus resembles that in other eukaryotic organisms. Polyamines reduce the rate of synthesis and increase the rate of degradation of ODC protein (1). However, unlike all other eukaryotes in which polyamine-mediated regulation has been studied, changes in the rate of synthesis ODC in N. crassa are correlated with similar changes (ca. 10-fold) in the abundance of spe-1 mRNA (43). Previously, we identified two regions of the spe-1 gene that affect its expression (27). An upstream activation region (UAR) was required for normal expression. Its elimination reduced spe-1 mRNA abundance approximately fivefold, with some loss of regulation by polyamines. The second region was a 473-bp AflIII-NruI segment of the region encoding the long 5-untranslated region (5′-UTR) of the mRNA. Deletion of this segment resulted in high levels of spe-1 expression and greatly diminished polyamine regulation.

Here we report a general analysis of sequences required for polyamine regulation of spe-1 mRNA and ODC activity in N. crassa. We show that regulation of spe-1 mRNA involves interactions between sequence elements in the DNA encoding the 5′-UTR of the mRNA and regions upstream of or overlapping the start of transcription, as well as an independently acting 3′ activator element. In addition, we identify an element in the spe-1 5′-UTR that modestly regulates the translation of the ODC coding region in response to polyamines.

MATERIALS AND METHODS

N. crassa strains and growth conditions.

The N. crassa arginaseless strain IC3 (aga) was used as the wild-type control in this study. Plasmids were transformed into strains IC2747-22a (his-3 aga), IC54 (spe-1::hph his-3 aga), and IC2794-5 (spe-1 inl his-3 aga) as indicated below. The specific alleles of the genes carried by these strains were aga(UM906), his-3(Y155M261), inl(89601), and spe-1(JP209). The isolation of the strain carrying a disrupted spe-1 gene (spe-1::hph) is described below.

All strains carry the aga mutation, which eliminates arginase activity and renders the strains unable to catabolize arginine as a source of ornithine (5). The addition of 1 mM arginine to these strains leads to feedback inhibition of ornithine synthesis from glutamate and, because arginine catabolism is blocked, results in ornithine and polyamine starvation. Under these circumstances, ODC activity is highly derepressed and polyamine synthesis is confined to the formation of small amounts of the polyamine analogs cadaverine (1,4-diaminopentane) and aminopropylcadaverine (24). These analogs, formed by the inefficient decarboxylation of lysine by derepressed ODC activity (24) and subsequent aminopropylation, allow indefinite, slow growth of arginine-treated aga cultures (4).

N. crassa strains were grown and maintained by standard methods (2). Cultures were provided with 1 mM spermidine trihydrochloride to support the growth of spe-1 mutant strains and in some cases to test the effect of polyamine supplementation. In other cases, 5 mM ornithine was added to restore cellular polyamines to arginine-grown cultures. Mutants deficient in inositol synthase (inl) were supplemented with 100 μg inositol per ml of culture. Cultures of his-3 mutants were supplemented with 100 μg of histidine per ml.

Plasmid constructions and N. crassa transformation. (i) Nested deletions of the spe-1 UAR.

Plasmid pPHL2 (27) consists of a BglII-HindIII cassette containing the entire spe-1 gene (Fig. 1) in a pDE1-based transformation vector. This plasmid was used as starting material for 5′-to-3′ deletions from the spe-1 upstream PstI site using an exonuclease III/mung bean nuclease deletion kit (Stratagene) as specified by the manufacturer. The BglII site in the multiple-cloning site of the pDE1 vector was used to linearize the plasmid and then filled in with α-thiophosphate nucleoside triphosphates. Deletions were initiated from an adjacent EcoRI site (also in the pDE1 multiple-cloning site) immediately upstream of the spe-1 PstI site. The blunt, deleted plasmid ends resulting from the exonuclease III/mung bean nuclease reactions were ligated with BglII linkers, digested with BglII, recircularlized, and used to transform Escherichia coli DH1 cells. A set of plasmids (see Fig. 5) carrying nested deletions of the spe-1 5′ region were identified by restriction analysis, and their 5′ ends were determined by manual sequencing with T7 primers using Sequenase version 2.0 (U.S. Biochemical Corp.) as specified by the manufacturer. These plasmids were used to transform strain IC2794-5.

FIG. 1.

Restriction map of the spe-1 gene and flanking sequences found in plasmids pPHL1 and pPH1. The boxed area represents the transcribed region, beginning with the right-pointing arrow. The coding sequence, interrupted by one intron, is shown in black. Abbreviations: MCS, multiple-cloning site; B, BglII; V, EcoV; C, ClaI; R, EcoRI; P, PstI; S, SacI; A, AflIII; N, NruI; K, KpnI; Sa, SalI; H, HindIII.

FIG. 5.

Effects of 5′-to-3′ deletions of the spe-1 upstream region on ODC activity and derepression of spe-1 mRNA. (A) Schematic representation of the wild-type (P2) and deleted spe-1 genes integrated at the his-3 locus of strain IC2794-5. Distances from the major transcription start site, indicated by the arrow, are given in base pairs. The relative positions of the PstI (−1000), SacI (−167), and AflIII (+97) sites are also shown. ODC activity (in units per milligram of protein) of the transformants grown with 1 mM spermidine (SPD) or 1 mM arginine (ARG) are given to the right. (B) Northern blots of total RNA (10 μg) from repressed (left) and derepressed (right) cultures of these transformants were probed with spe-1 cDNA (spe-1) or a fragment of the β-tubulin gene (tub), the latter as a loading control. Lanes: 1, P2; 2, PΔ1; 3, PΔ2; 4, PΔ3; 5, PΔ4; 6, PΔ5; 7, S8; 8, PΔ6; 9, PΔ7. (C) Approximately 25 μg of total RNA from the derepressed transformants was analyzed by primer extension reactions with the MH12 primer to determine the 5′ ends of their spe-1 transcripts. The molecular size marker on the right is given in nucleotides. Lanes are labeled as in panel B.

(ii) spe-1 5′-UTR deletions.

Deletions within the long (622-bp) spe-1 5′-UTR were constructed using convenient restriction sites: AflIII (at position +101 from the transcription start site), Bsu36I (+266), StuI (+403), and NruI (+574). The starting material for the deletions was plasmid pSS1 (27), containing the 1.3-kb spe-1 SacI-SalI region in a pSP72 vector (Promega), in which the AflIII site of the vector had been obliterated. The pSS1 plasmid was digested with two of the appropriate restriction endonucleases, and where necessary (AflIII and Bsu36I digests), overhanging ends were filled in using Klenow DNA polymerase. The doubly digested plasmid was recircularized, and the SacI-SalI fragment was cloned into spe-1 sequences in pPH1 (27). The spe-1 genes containing 5′-UTR deletions were cloned into pDE1 transformation vectors, creating plasmids pMH11 (AflIII-Bsu36I deleted), pMH12 (Bsu36I-StuI deleted), pMH13 (StuI-NruI deleted), and pMH15 (AflIII-StuI deleted). Construction of plasmid pDPH1 (AflIII-NruI deleted) was described previously (27). These plasmids were used to transform strain IC54.

(iii) spe-1::qa-2 constructs.

For construction of spe-1::qa-2 chimeric genes, a 0.7-kb spe-1 SacI-NruI fragment (containing most of the spe-1 5′-UTR) was joined to a 0.7-kb Asp718-BamHI qa-2 fragment from plasmid pMSK338 (32). The qa-2 Asp718 site lies 13 bp upstream of the initiation codon of the coding region of N. crassa catabolic dehydroquinase. The NruI end of the spe-1 fragment was blunt-end ligated to the Asp718 end of the qa-2 fragment that had been filled by treatment with Klenow polymerase, and the spe-1::qa-2 fragment was inserted into a SacI-BamHI-cut pSP72 vector to create plasmid pMH34. Similarly, a 262-bp SacI-AflIII spe-1 fragment (lacking most of the spe-1 5′-UTR) was blunt-end ligated to the Asp718-BamHI qa-2 fragment, after the AflIII and Asp718 ends had been filled in, and inserted into a SacI-BamHI-cut pSP72 vector to create plasmid pMH35. The upstream region of the spe-1 gene was reconstituted in the spe-1::qa-2 vectors by insertion of a 0.8-kb BglII-SacI spe-1 cassette (containing the UAR) from pPH1 into the BglII-SacI-cut pMH34 and pMH35 by ligation, generating plasmids pMH36 and pMH37, respectively. The entire spe-1 5′-UTR was joined to the qa-2 coding region at the translation initiation codon using gene splicing by overlap extension (15). The spe-1 sequences were amplified as a 211-bp fragment from plasmid pSS1 using the sense spe-1 primer MH1 (5′-CGTACCGACACCGACCCCCC-3′) and the antisense splice primer SQ2 (5′-ATGTGACGGGGGGACGCCATATCCCAAGATTTGACTG-3′). A 0.4-kb qa-2 fragment was amplified from plasmid pMSK338 using the sense splice primer SQ1 (5′-CAGTCAAATCTTGGGATATGGCGTCCCCCCGTCACAT-3′) and the antisense qa-2 primer Q1 (5′-CACATGAACCTCCACAAACG-3′). The PCR products from these reactions were then combined and used as templates for splice overlap extension PCR with primers MH1 and Q1. The product was attached to upstream spe-1 sequences in plasmid pMH79. Transformation vectors were generated by ligation of BglII-HindIII cassettes containing the entire spe-1::qa-2 genes from pMH36, pMH37, and pMH79 into BglII-HindIII-cut pDE1 vectors, generating plasmids pMH40, pMH41, and pMH82, respectively. These plasmids were used to transform strain IC2747-22a.

(iv) tub::spe-1 constructs.

The basic constructs used as starting material for construction of chimeric tub::spe-1 genes (see Fig. 7 and 8 and Results) were generated by introducing a BamHI linker 10 bp upstream of the spe-1 initiation codon and 25 bp downstream of the 5′ end of the tub transcribed region (21), and an EcoRI linker immediately following the spe-1 and tub termination codons. The spe-1 coding region was amplified from plasmid pPH1 with primers MH8 (5′-TTATCACCCAGGATCCTCTTGGGATATGGTTATGCCGAC-3′) and MH9 (5′-AAGCTGCACGGAATTCTTACAATCCCAAGAGCGCCATAGC-3′), which introduced 5′ BamHI and 3′ EcoRI sites (underlined), respectively. The tub 5′- and 3′-UTRs and vector sequences of plasmid pβT6, which carries the β-tubulin gene (20), were amplified with primers MH11 (5′-ACGCATCTTGGGATCCTGGTGATGACGAACACGGGTCTAT-3′) and MH10 (5′-CCTTGAGGGCGAATTCTAAATCATTCCACTCAACATTCAG-3′), which introduced 5′ BamHI and 3′ EcoRI sites, respectively. The PCR products were digested with BamHI and EcoRI, and the spe-1 coding region was inserted between the tub untranslated sequences by ligation to create plasmid pMH25. Several subcloning steps followed to place a BglII site upstream of the SalI site at the 5′ end of the tub promoter sequences to create plasmid pMH30. This plasmid consists of a BglII-HindIII cassette containing the spe-1 coding region flanked by tub 5′- and 3′-UTRs in a pSP72-derived vector. The BglII-HindIII cassette from plasmid pMH30 was subcloned into the BglII-HindIII-cut pDE1 vector, generating the transformation vector pMH33. A tub::spe-1 chimeric gene containing the spe-1 coding and 3′ regions attached to the tub promoter and 5′-UTR was created by ligating a spe-1 NotI-HindIII fragment from pPH1 into the similarly cut pMH33 to create plasmid pMH42.

FIG. 7.

Expression of various spe-1 and chimeric transcripts driven by the β-tubulin (tub) promoter of N. crassa. (A) Schematic diagram of tub::spe-1 genes introduced into strain IC54, with functional regions of each gene listed across the top. The spe-1 sequences are represented by open boxes; and tub sequences are represented by shaded boxes. The positions of the AflIII and NruI sites in the spe-1 5′-UTR are shown. (B) Northern blots of 10 μg of total RNA from repressed (SPD) or derepressed (ARG) cultures, probed with spe-1 cDNA or the coding region of tub DNA. The relative abundance of tub::spe-1 mRNA in each transformant, normalized to tub mRNA and relative to that in DMH43/SPD, is given below the panel.

FIG. 8.

Role of the UAR on the regulation of spe-1 genes lacking the AflIII-NruI segment of the 5′-UTR. (A) Schematic of spe-1 genes lacking 5′-UTR sequences in which the spe-1 UAR is either present (DMH3) or absent (DMH4). The positions of the AflIII and NruI sites of the 5′-UTR are indicated. (B) Northern blot analysis of repressed (SPD) and derepressed (ARG) cultures of the transformants. Northern blots were hybridized with probes derived from spe-1 cDNA or the coding region of the tub gene.

To create chimeric genes with the tub promoter attached to different lengths of the spe-1 gene, BamHI sites were introduced as linkers into the spe-1 5′-UTR at the AflIII or NruI site. Plasmid pSS1 was digested with AflIII, and pSP3 (containing a 3.2-kb spe-1 SacI-HindIII cassette) was digested with NruI. The linearized plasmids were treated with Klenow polymerase, BamHI linkers (Stratagene) were ligated onto the filled in ends, and the plasmids were digested with BamHI and recircularized. The resulting plasmids were digested with BamHI and DsaI, and the resulting fragments, containing portions of the spe-1 5′-UTR and coding region, were used to replace the BamHI-DsaI spe-1 sequences in pMH42. This resulted in plasmids containing the tub promoter fused to the spe-1 5′-UTR at the former positions of AflIII (in pMH45) and NruI (in pMH47) sites, now replaced by BamHI linkers. A BamHI site (underlined) was introduced at the 5′ end of the spe-1 transcribed region by PCR amplification with primer MH37 (5′-CAAGTCCAACCTACCTCTTGGATCCTTTCTCACCCTTCT-3′) and a downstream antisense spe-1 primer GB9 (5′-TGCGGAGGAAAAGCTCGGCG-3′). The PCR product was digested with BamHI and DsaI, and the 0.7-kb spe-1 fragment was used to replace BamHI-DsaI spe-1 sequences in pMH42. The resulting plasmid, pMH86, contained the tub promoter fused to the major transcription start site (+1) of the spe-1 gene. In plasmid pMH107, the spe-1 NotI-HindIII 3′ region of pMH86 was replaced with that of pMH30, described above, in which the tub 3′ region follows the spe-1 coding region. These plasmids were used to transform strain IC54.

PCR method.

PCRs used in cloning procedures were carried out using the Expand high-fidelity PCR system (Boehringer Mannheim). Each reaction mixture contained 100 pmol of each primer, 0.1 μg of template DNA, 200 nM each deoxyribonucleoside triphosphate, 2.5 U of Expand enzyme mix, and 1× PCR buffer (10 mM Tris-HCl [pH 8.3], 1.5 mM MgCl₂, 50 mM KCl) in a 100-μl volume. The reactions were performed in a PTC-100 thermal controller (MJ Research) for 15 cycles (denaturing for 2 min at 94°C [4 min for the first cycle only], annealing for 2 min at 55°C, and extension for 2 min at 68°C). Products of the appropriate size were purified by agarose gel electrophoresis.

Transformation.

Plasmid DNA (5 μg) was used for transformation of N. crassa spheroplasts as previously described (20). The pDE1-derived vectors, containing various spe-1 constructs, were used to target the entire transforming plasmid DNA to the his-3 locus. N. crassa strains carry the Y155M261 allele of his-3, which has a point mutation in the 3′ region of the coding region; the transforming plasmid lacks a 5′ region of the gene. A His⁺ phenotype results only when the truncated his-3 gene of the plasmid recombines homologously with the mutant allele in the recipient (see Fig. 3 in reference 27). Transformants were selected and screened as previously described (27) for those carrying single copies of the spe-1 plasmids integrated at the his-3 locus. This allowed spe-1 expression of different transformants to be determined with the constructs in a common chromosomal context. N. crassa transformants were made using recipients carrying either the wild-type spe-1⁺ allele or the disrupted spe-1 gene described below.

Construction of a strain with a spe-1 gene disruption (IC54, spe-1::hph).

Plasmid pGS1 carries the 5.0-kb HindIII fragment that includes the entire spe-1 gene (43). We replaced the 1.3-kb SacI-SalI segment, containing the spe-1 promoter and transcription and translation starts with the bacterial hygromycin phosphotransferase gene (hph) in reverse orientation (Fig. 2A). The source of the hph gene was plasmid pCSN43 (36), in which the hph gene is transcribed from the Aspergillus nidulans trpC promoter. The 5.0-kb spe-1::hph insert of the resulting plasmid (pMB1) was used for homologous replacement of the endogenous spe-1 gene of strain IC2747-22a (his-3 aga) by transformation and selection of hygromycin-resistant (Hyg^r) colonies followed by screening for those that were Spe⁻. One transformant identified by Southern analysis that met the criteria for homologous replacement by the disrupted spe-1 fragment was backcrossed to a his-3 aga strain of the opposite mating type, and a Spe⁻ Hyg^r isolate was selected and designated strain IC54. This strain behaves as expected with respect to diagnostic probes of its DNA (Fig. 2B), and it lacks detectable spe-1 mRNA under repressing and derepressing conditions (results not shown).

FIG. 2.

Construction of a spe-1::hph deletion-insertion mutant. (A) Restriction maps of the wild-type N. crassa spe-1 genomic region (top) and the spe-1::hph deletion-insertion allele (bottom). Boxes indicate the spe-I (open) and trpC::hph (shaded) transcribed regions. Arrows indicate the direction of transcription. Restriction fragments used to probe Southern and Northern blots are indicated. Distances from the 5′ HindIII site are given in base pairs. Abbreviations: B, BamHI; H, HindIII; P, PstI; S, SalI. (B) Southern blot of N. crassa genomic DNA from strains IC3 (aga) and IC54 (spe-1::hph his-3 aga) digested with BamHI and HindIII and probed with the probes indicated in panel A.

Inhibition of RNA synthesis by thiolutin.

Thiolutin, an inhibitor of all three RNA polymerases, has been used previously for measurements of mRNA half-life in Saccharomyces cerevisiae (14). Thiolutin was kindly provided to us by Pfizer, Inc., and a 1.5-mg/ml stock solution was prepared in dimethyl sulfoxide. The effects of thiolutin on the transcription of total cellular RNA were measured by the addition of 4 × 10⁻² μCi of [³H]uridine and 50 nmol of cold uridine per ml of an aga culture (strain IC3, inoculated with 10⁶ conidia per ml) during exponential growth in Vogel's minimal medium. The cells were labeled for 10 min before the addition of 0, 1.5, or 3 μg of thiolutin per ml of culture. Culture samples (5 ml) were added to an equal volume of ice-cold 10% trichloroacetic acid (TCA), and the cells were then collected and washed with cold 5% TCA on 5-μm-pore-size membrane filters (Micron Separations Inc.). The filters were boiled for 10 min in 5% TCA, and the acid-soluble radioactivity was determined using a scintillation counter. The effects of thiolutin on translation in strain IC3 were measured in similar cultures growing in the presence of 1.25 × 10⁻² μCi of [³⁵S]methionine and 50 nmol of cold methionine per ml of culture. Cells were labeled for 10 min before thiolutin was added, as in the transcription experiments. At given times after thiolutin addition, 5-ml culture samples were collected on 5-μm membrane filters. The filters were washed with 5% cold TCA and counted.

N. crassa RNA analyses.

N. crassa total RNA was prepared as previously described (43), with minor modifications. Wet mycelial pads were collected by filtration, frozen at −80°C, lyophilized overnight, and powdered by vortexing in a 7- by 100-mm polypropylene tube with a spatula. Solubilization of RNA in later steps was routinely followed by a 2-min centrifugation step (10,000 × g) to remove insoluble material. RNA was resuspended and stored in deionized formamide.

Total RNA was denatured in 1× MOPS buffer–37% formaldehyde (2:15 [vol/vol]; 1× MOPS buffer is 200 mM morpholinepropanesulfonic acid [MOPS], 50 mM sodium acetate, and 10 mM EDTA [pH 7.0]) by heating to 65°C for 3 min and separated by electrophoresis on 6% formaldehyde–1% agarose gels made in 1× MOPS buffer. RNA was transferred to Nytran nylon membranes (Schleicher & Schuell) by capillary transfer using 10× SSC (1× SSC is 0.15 M NaCl plus 0.015 sodium citrate) (31). The membranes were UV cross-linked as specified by the manufacturer and prehybridized in 50% deionized formamide–5× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7])–5× Denhardt's reagent–0.1% sodium dodecyl sulfate (SDS)–0.1 mg of sheared salmon testis DNA per ml (31). Hybridization was carried out in the same buffer overnight at 42°C using random-primed ³²P-labeled DNA probes. A 0.5-kb EcoRI-KpnI fragment from pCS1 (to probe for spe-1 mRNA [43]), a 1.2-kb SstI fragment from pβT6 (for β-tubulin mRNA [20]), or a 0.7-kb Asp718-BamHI fragment from plasmid pMSK338 (for qa-2 mRNA [32]) were used as probes. Hybridized membranes were washed three times at room temperature for 5 min with 2× SSPE–0.1% SDS and three times at 65°C for 30 min with 0.1 × SSPE–0.1% SDS. Signals from hybridized membranes were visualized by autoradiography, and scanned autoradiographs were quantified by densitometry with Adobe Photoshop software. A standard series of autoradiographic signals was also scanned and quantified to determine the linear range of values obtainable by this method. Multiple autoradiographic exposures were quantified for each experiment, and only values within the linear range were used for mRNA determinations.

Primer extension analysis was carried out using the avian myeloblastosis virus reverse transcriptase primer extension system (Promega). Between 10 and 25 μg of total RNA in deionized formamide was precipitated with 4 volumes of ethanol and resuspended in a total volume of 11 μl of 1× primer extension (PE) buffer (50 mM Tris-HCl [pH 7.5], 50 mM KCl, 10 mM MgCl₂, 10 mM dithiothreitol, 0.5 mM spermidine, 1 mM each dATP, dCTP, dGTP, and dTTP) with 100 fmol of the end-³²P-labeled spe-1 primer (5′-CCGTGTTAGATACGGTTGCC-3′). The primer and RNA were heated to 83°C for 5 min and annealed at 55°C for 20 min. The annealed primers were extended by the addition of 9 μl of the extension mix (final concentration, 1× PE buffer, 2.8 mM sodium pyrophosphate, and 1 U of avian myeloblastosis virus reverse transcriptase) and given a 30-min incubation at 42°C. Extension products were resolved on a denaturing 6% polyacrylamide gel and visualized by autoradiography.

Polysome analyses.

Methods for polysome analysis were adapted from those previously described (44). Standard exponential cultures were treated with cycloheximide (final concentration, 50 μg/ml) 5 min prior to harvesting. Mycelia were homogenized in buffer (10 mM KCl, 10 mM MgCl₂, 30 mM triethanolamine, and 1 mM EGTA, treated with 0.01% diethylpyrocarbonate, and sterilized by autoclaving) in a 30-ml Bead Beater chamber (Biospec) with acid-washed 0.5-mm-diameter glass beads. Nuclei and mitochondria were removed from the homogenate by centrifugation at 12,000 × g for 10 min at 4°C. A volume of the supernatant containing 4 absorbance at 260 nm (A₂₆₀) units of the supernatant was layered on a 11.6-ml linear gradient (10 to 40% sucrose in 50 mM Tris-HCl [pH 7.6]–25 mM KCl–3 mM MgCl₂ treated with 0.01% diethylpyrocarbonate). Polysomal RNA was sedimented through the gradients by centrifugation in a Beckman SW41 Ti rotor at 35,000 rpm for 90 min at 4°C. The A₂₅₄ profile was determined by pumping the gradient through a Gilson 111B UV detector. Approximately 0.5-ml fractions were collected from the bottom of the gradient into microcentrifuge tubes containing 0.5 ml of isopropanol and precipitated in isopropanol overnight at −20°C.

Polysomal RNA was extracted from the precipitated fractions by the sequential addition of 100 μl of solution D (4 M guanidinium thiocyanate, 25 mM sodium citrate [pH 7.0], 0.5% sarcosyl, 0.1 M β-mercaptoethanol), 10 μl 2 M sodium acetate, 100 μl of water-saturated phenol, and 20 μl of chloroform-isoamyl alcohol (48:1). Samples were mixed, and the phases were separated by centrifugation (3,000 × g for 10 min at 4°C). The aqueous phase was precipitated with 220 μl of isopropanol at −20°C. Precipitated RNA was resuspended in 10 μl of deionized formamide and stored at −80°C until analyzed by Northern blotting.

Enzyme assays.

ODC activity was measured in duplicate 5-ml permeabilized cell samples by previously described methods (4, 43).

Catabolic 5-dehydroquinase activity was assayed in extracts of lyophilized mycelial powders as described by Hautala et al. (13). The lyophilized powders, in extraction buffer (0.1 M K⁺PO₄ [pH 7.5], 0.4 mM dithiothreitol, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride) at a ratio of 1 g of powder per 20 ml of buffer, were resuspended on a vortex mixer and mixed further on an orbital shaker for 30 min at room temperature. The extract was cleared by centrifugation, and protamine sulfate was added to a final concentration of 0.14% and mixed for 15 min on an orbital shaker. The precipitated nucleic acids were removed by centrifugation, and the extract was then heated to 71°C for 10 min to inactivate biosynthetic dehydroquinase. The precipitated proteins were removed by centrifugation, and the supernatant was dialyzed against extraction buffer overnight at 4°C. Enzyme activity was measured in reaction mixtures containing 100 mM Tris-HCl (pH 7.4), 0.2 mM 5-dehydroquinic acid (kindly provided by M. E. Case), 0.2 mM EDTA, and 1/12 volume of dialyzed cell extract. Activity was measured as the rate of appearance of the product, 5-dehydroshikimic acid, by monitoring the change in A₂₄₀. Protein was determined by the method of Lowry et al. (18), using 100 μl of the cell extract.

RESULTS

As noted in the introduction, previous studies revealed that deletion of two regions of the spe-1 gene, the UAR and a segment encoding the 5′-UTR of the mRNA, affect its expression. The results suggested that the UAR includes an enhancer-like element and that sequences in the 5′-UTR encode a negatively acting, polyamine-responsive regulatory element. Two questions might be asked regarding the mechanism of polyamine regulation of spe-1 mRNA. (i) Because the polyamine-responsive element appears to reside within the transcribed region of the gene, is regulation exerted through changes in the stability of the mRNA? (ii) Are the sequences in the 5′-UTR able to impart polyamine regulation to other genes? We show here that although sequences in the 5′-UTR negatively affect expression, they are insufficient to impart polyamine regulation to reporter constructs in the absence of upstream spe-1 sequences. Furthermore, we demonstrate that in the absence of any upstream untranscribed and 5′-UTR spe-1 sequences (hereafter collectively referred to as the 5′ polyamine-responsive region [5′ PRR]), a 3′ polyamine-responsive region (3′ PRR) remains. The 3′ PRR was identified using a heterologous promoter driving the expression of spe-1 coding and 3′ regions. The effects of the 3′ PRR are obscured by high levels of expression in spe-1 genes lacking regulatory elements in the 5′-UTR.

Effects of polyamines and 5′-UTR sequences on the stability of spe-1 mRNA.

Because we had previously identified elements affecting the expression of spe-1 mRNA within the spe-1 transcribed region, we asked whether these elements, or polyamines themselves, altered the stability of spe-1 mRNA. To determine the half-life of cytoplasmic spe-1 mRNA, we inhibited transcription with the drug thiolutin at 1.5 μg/ml in exponential-phase cultures. This concentration inhibited total-RNA synthesis by 75% and inhibited translation by 20%. Higher concentrations of thiolutin were more effective in inhibiting transcription. However, they also inhibited translation to an extent that might seriously compromise the degradation of short-lived mRNAs (30). The residual transcription at the lower concentration of thiolutin should not affect our conclusions, since our interest is in differences in the rate of mRNA decay, rather than the true half-life.

The inhibition of transcription in a polyamine-starved (arginine-grown) culture of the aga strain (IC3) resulted in the disappearance of spe-1 mRNA with a half-life of approximately 15 min (Fig. 3, top panel). To determine the effects of polyamines, ornithine was added to an arginine-grown culture following the inhibition of transcription in these experiments. Normally, supplementation of the growth medium from inoculation with spermidine is sufficient to repress ODC activity (27). However, cellular polyamine pools are restored only slowly by the direct addition of spermidine to the growth medium of starved cultures, owing to poor uptake (4, 7). The addition of ornithine leads to its rapid conversion to excess putrescine and spermidine by the derepressed ODC in these cells (23). Restoration of polyamine synthesis by this method had no effect on the turnover of spe-1 mRNA compared to that in the polyamine-starved culture (Fig. 3, middle panel). Similar experiments were carried out with cultures of a transformant (DMH3) in which the AflIII-NruI region of the spe-1 5′-UTR had been deleted and in which spe-1 expression was highly derepressed even in the presence of excess ornithine (27). The stability of the mutant mRNA was similar to that of the normal transcript in both the presence and absence of polyamines despite the deletion of the 5′-UTR sequences (Fig. 3, bottom panel). We conclude that polyamine regulation does not involve changes in the stability of cytoplasmic spe-1 mRNA, since neither excess polyamine levels nor removal of the negatively acting 5′-UTR sequence altered the rate of spe-1 mRNA turnover.

FIG. 3.

Effects of polyamines on the stability of spe-I mRNA. Cultures of strain IC3 (aga) or transformant DMH3 (Δ5′-UTR) were starved for polyamines by growth on 1 mM arginine. Transcription was inhibited by the addition of thiolutin (final concentration, 1.5 μg/ml) at time zero. Incubation was continued in arginine-containing medium (ARG) or in cultures supplemented with ornithine (5 mM) to restore polyamine synthesis (ORN). Total RNA was extracted, and 10 μg from each time point was subjected to Northern blotting, using a 0.5-kb NruI-SalI fragment of a spe-1 cDNA as a probe. rRNA was visualized on the gels prior to transfer with ethidium bromide to confirm equivalent sample loading.

Determination of spe-1 mRNA transcription start sites.

Previous S1 endonuclease protection experiments (43) indicated that spe-1 mRNA transcription was initiated at three adjacent sites, 535, 534, and 533 nucleotides upstream of the first ATG in the transcribed region. We carried out primer extension analysis to confirm this result, using an antisense spe-1 primer with a 5′ end located 87 nucleotides downstream from the previously determined spe-1 transcription start site. Primer extension failed to detect the 5′ end determined by S1 nuclease protection in both the control IC3 strain (Fig. 4A, lanes 2 and 3) and a transformant (DMH1) carrying a normal copy of the spe-1 gene at the his-3 locus (lanes 4 and 5). Instead of the expected ca. 87-nucleotide transcript extension product, transcripts with 5′ ends ranging in size from about 140 to 200 nucleotides were found. The major product was 173 nucleotides. We calculated that this product belonged to a transcript whose 5′ end was 622 nucleotides upstream from the first ATG in the spe-1 transcribed region (Fig. 4B). We arbitrarily designate the 5′ end of this transcript the actual start site (+1) of the spe-1 transcript. Other, less abundant transcripts that initiated 5′ and 3′ to the major transcript were also seen. The conclusion that the primer extension analysis detected bona fide spe-1 transcripts was based on the following observations. (i) In both the IC3 and DMH1 strains, the abundance of the extension products was derepressed by polyamine starvation (Fig. 4A, compare lane 2 with lane 3 and compare lane 4 with lane 5). (ii) No extension products were detectable in the IC54 strain, in which the corresponding spe-1 sequences were been deleted (lanes 6 and 7). (iii) Finally, we detected sequences corresponding to the new 5′ end of the spe-1 transcripts independently by reverse transcription-PCR analysis (results not shown).

FIG. 4.

Determination of the 5′ ends of spe-1 transcripts by primer extension analysis. (A) Total RNA was isolated from cultures of strains IC3 (aga), IC54 (spe-1::hph his-3 aga), and transformant DMH1, supplemented as indicated below with 1 mM spermidine (SPD), 1 mM arginine (ARG), or 100 μg of histidine per ml (HIS). Total RNA (20 μg) was analyzed by primer extension analysis using the MH12 primer. Lanes: 1, φX174/HincII molecular weight marker; 2, IC3/SPD; 3, IC3/ARG; 4, DMH1/SPD; 5, DMH1/ARG; 6, IC54/SPD+HIS; 7, IC54/HIS; 8, no RNA. (B) The 5′ ends of the extension products from total RNA of the IC3 strain, grown in the presence of arginine, were determined by comparing primer extension products (lane 5) to dideoxynucleotide chain termination sequencing reactions (lanes 1 to 4) with a spe-1 DNA template using the same primer. The sequence of the sense DNA strand is given to the right, with the major extension product indicated by the arrow.

Effects of upstream sequences on spe-1 expression and polyamine regulation.

A HindIII genomic fragment (GenBank accession no. L16920) includes a PstI-HindIII region sufficient for normal expression of the spe-1 gene (27). Deletion of sequences between the upstream PstI site, 1,000 bp 5′ of the transcription start site, and a SacI site 167 bp 5′ of +1 (a region defined as the UAR) leads to a three- to fivefold reduction in mRNA abundance without eliminating its regulation by polyamines (27). To better define upstream sequences affecting expression, a series of transformants carrying 5′-to-3′ nested deletions extending from the upstream PstI site to a site downstream of the spe-1 transcription start site was constructed (Fig. 5A). The spe-1 constructs carrying these deletions were targeted by transformation to the his-3 locus of strain IC2794-5, which lacked detectable ODC activity, in earlier studies. No spe-1 mRNA or ODC activity was detectable in the untransformed recipient strain (a chromosomal aberration mutant), even in polyamine-starved cultures. The P2 transformant in this series, carrying the full-length UAR, had somewhat (but reproducibly) lower ODC activity and spe-1 mRNA than did comparable transformants carrying the same plasmid in the IC54 ODC-null strain. However, the effect of deleting the entire PstI-SacI sequence of the spe-1 gene on ODC activity and mRNA abundance (approximately fivefold decrease in polyamine-starved cultures) in the two types of transformant was similar.

In the deletion series, both ODC activity and spe-1 mRNA abundance diminished to similar extents in transformants as more of the upstream region of the spe-1 gene was removed (Fig. 5A and B), confirming previous evidence (27). The deletion of sequences upstream of −778 (in transformant PΔ2 [Fig. 5B, lane 3]) led to a ca. 50% decrease in spe-1 expression in the polyamine-starved cultures; further deletion to −347 (transformant PΔ5 [lane 6]) reduced expression to 25% of the control. Similar decreases in the ODC activity and spe-1 mRNA of these transformants occurred in spermidine-supplemented cultures. The deletion that overlapped the spe-1 transcription start site (transformant PΔ7 [lane 9]) almost completely eliminated expression under both growth conditions. In all transformants carrying upstream deletions, there was some derepression of spe-1 mRNA abundance in cells starved for polyamines. However, the deletion of the UAR clearly reduced the amplitude of the response to polyamines (compare the effect of polyamines on transformants P2 [lane 1] and S8 [lane 7]).

Primer extension was used to determine the 5′ ends of spe-1 transcripts in transformants with these upstream deletions. Removal of upstream sequences had no effect on transcription start site selection until sequences between −543 and −347 were eliminated (transformant PΔ5 [Fig. 5C, lane 6]). The deletion of this region greatly reduced initiation from the normal start site and led to the appearance of new transcripts with 5′ ends 40 to 60 nucleotides downstream (lanes 6 to 8). This change in the site of transcript initiation occurred although the original start site remained intact. The deletion that overlapped +1 (in transformant PΔ7 [lane 9]) eliminated all transcripts detectable with the primer.

The spe-1 UAR and 5′-UTR impart polyamine regulation to a heterologous coding and 3′ region.

We next asked whether the UAR and 5′-UTR of the spe-1 gene would confer polyamine regulation to a heterologous downstream reporter gene. We attached various portions of the spe-1 5′ region to the coding and 3′ region of the N. crassa qa-2 gene (Fig. 6A) and determined the effects of polyamines on expression of the chimeric gene. The qa-2 gene encodes catabolic dehydroquinase, an enzyme involved in quinic acid utilization, and qa-2 mRNA is undetectable in wild-type strains in the absence of quinic acid supplementation (22). In quinate-induced cultures, manipulation of cellular polyamine levels had no effect on qa-2 expression (data not shown).

FIG. 6.

Regulation of spe-1::qa-2 chimeric genes by polyamines. (A) Schematic diagram of the spe-1::qa-2 genes used to transform strain IC2747-22a. Functional regions of each gene are listed across the top. Open boxes represent spe-1 sequences, while shaded boxes represent qa-2 sequences. The relative positions of the AflIII (+101) and NruI (+574) sites within the spe-1 5′-UTR are indicated. The fold regulation of spe-1::qa-2 mRNA in each transformant, derived from the Northern blot analysis in panel B, shown on the right, is the ratio of qa-2 mRNA abundance in derepressed over repressed cultures after normalization with tub mRNA. (B) Northern blot analysis of 10 μg of total RNA from repressed (SPD) or derepressed (ARG) cultures of the transformants. Northern blots were probed with qa-2- or tub-specific probes. The amount of spe-1::qa-2 mRNA relative to that in SPD-grown SMH12 is given below the qa-2 panels. The order of transformants on the Northern blots has been rearranged for ease of presentation, preserving the relative positions of the autoradiographic signals. (C) The turnover of spe-1 and spe-1::qa-2 mRNA in the SMH41 transformant was determined after inhibition of transcription as described in the legend of Fig. 3. The spe-1 AflIII-NruI sequence was used as a probe for Northern blots.

In transformant SMH13 (Fig. 6A), spe-1 sequences upstream of the AflIII site (at +101) in the 5′-UTR were fused to the qa-2 coding and 3′ region. Polyamines had little or no effect on expression of the spe-1::qa-2 mRNA detected by a qa-2 probe, in this transformant (Fig. 6B). If the chimeric gene also included the spe-1 AflIII-NruI region (in transformant SMH12), expression of the resulting spe-1::qa-2 mRNA was reduced in spermidine-supplemented cultures but was derepressed threefold upon polyamine starvation. The effects of polyamines on spe-1::qa-2 mRNA in this transformant were modest but repeatable and were always greater than in the transformant lacking the AflIII-NruI sequences. The further inclusion of sequences between the NruI site and the spe-1 translation initiation codon (at +623) in transformant SMH41 resulted in slight increases of spe-1::qa-2 mRNA abundance in both polyamine-supplemented and starved cultures compared to the SMH12 transformant and in a slight increase in the regulatory amplitude.

Transformants carrying the spe-1::qa-2 chimeric gene integrated at the his-3 locus also contained a functional spe-1 gene at its normal chromosomal location. This allowed us to compare the expression of the spe-1 and spe-1::qa-2 genes in transformant SMH41 by Northern analysis with probes of the 5′ spe-1 transcribed sequences. In a polyamine-starved culture, expression of the spe-1::qa-2 mRNA is much lower than that of the endogenous spe-1 gene (Fig. 6C, 0 min). We measured the turnover of spe-1::qa-2 mRNA using thiolutin and compared it to that of spe-1 mRNA in the same transformant (Fig. 6C). Both mRNAs disappeared with similar half-lives, approximately 15 min, indicating that changes in stability cannot account for the differences in expression.

These results indicate that spe-1 sequences upstream of the AflIII site in the 5′-UTR, including sequences in the UAR, are incapable of imparting significant polyamine regulation in the absence of downstream spe-1 sequences. The inclusion of sequences in the AflIII-NruI region in addition to these upstream sequences results in polyamine regulation of the chimeric gene. However, polyamines did not regulate the spe-1::qa-2 genes to same extent as they regulated the normal spe-1 gene, and its expression was lower. This raised the possibility that spe-1 sequences downstream of the 5′-UTR were required for normal expression (but not stability) and/or regulation of spe-1 mRNA. We next asked whether the AflIII-NruI region of the 5′-UTR, necessary for polyamine regulation, was sufficient in itself to impart regulation to a heterologous upstream promoter and what role, if any, downstream sequences play.

Downstream sequences confer promoter-independent polyamine regulation of spe-1 mRNA.

To test whether the polyamine regulation imparted by the spe-1 5′-UTR in the spe-1::qa-2 transformants was independent of upstream untranscribed sequences, we placed all transcribed spe-1 sequences and its 3′ flank downstream of a heterologous promoter, that of the N. crassa β-tubulin (tub) gene (Fig. 7A). (This promoter can force constitutive expression of the normally inducible N. crassa cys-3 gene [21].) The abundance of the full-length spe-1 mRNA in transformant DMH43, carrying such a construct, responded to changes in the cellular polyamine status (Fig. 7B). However, the ca. 4-fold regulation of spe-1 mRNA in this transformant was somewhat attenuated compared to the 10-fold regulation in the DMH1 transformant, carrying a normal spe-1 gene (27, 43). The result shows that at least some of the regulatory amplitude is imparted by spe-1 DNA sequences within the transcribed region, independently of the spe-1 promoter.

We next asked whether sequences in the 5′-UTR required for repression of the normal spe-1 gene were responsible for the regulated expression of the tub::spe-1 chimeric gene. Various 5′ deletions were made in the spe-1 5′-UTR, and the remainder of the gene was attached to the tub promoter (Fig. 7A). In transformant DMH26, all spe-1 5′UTR sequences downstream of the AflIII site (at +101) were attached to the tub promoter. The expression and polyamine regulation of tub::spe-1 mRNA in DMH26 (Fig. 7B) were similar to those in DMH43, demonstrating that sequences between +1 and AflIII are dispensable for regulation of the chimeric gene.

Transformant DMH27 carries the portion of the spe-1 gene downstream of the NruI site at +574 attached to the tub promoter, eliminating most of the spe-1 5′-UTR. This transformant has greatly increased mRNA accumulation and diminished regulation compared to transformants in which most of the spe-1 5′-UTR is present (DMH43 or DMH26 [Fig. 7B]). This behavior of DMH27, lacking sequences upstream of NruI, resembles the effect of deleting the AflIII-NruI segment in the intact spe-1 gene (27) (see below). Thus, the AflIII-NruI segment of the 5′-UTR negatively affects expression from the tub promoter, as it does in the spe-1::qa-2 constructs and the wild-type spe-1 gene.

We then tested a chimeric gene, in transformant DMH11, in which the tub promoter is joined to spe-1 sequences 9 bp upstream of the translation initiation codon (at +613). The absence of most of the sequences from the NruI site to the start codon in this transformant had two surprising effects. First, it caused a lower abundance of tub::spe-1 mRNA compared to that in transformant DMH27 (tub promoter fused at the NruI site of spe-1), suggesting that this 40 bp region contains a positively acting sequence element. Deletion of this sequence in an otherwise intact spe-1 gene reduced expression by 50% but had no effect on polyamine regulation (results not shown). We designated this region the downstream activation region (DAR). The second surprising effect of extending the deletion of the 5′-UTR by the additional 40 nucleotides was the restoration of polyamine regulation to the level observed in DMH43 (tub promoter joined to spe-1 at +1) and DMH26 (joined at AflIII). Thus, the AflIII-NruI segment, part of a major regulatory element in transformants carrying the spe-1 gene and qa-2 chimeras, appears wholly dispensable in those carrying the tub::spe-1 chimeras. Therefore, DMH27 and DMH11, both lacking the AflIII-NruI segment, may differ not in whether they have a remaining regulatory element but, rather, in their ability to express it. The high, constitutive expression of the tub::spe-1 chimera in DMH27 may limit further derepression upon polyamine starvation, while DMH11, lacking the DAR, may express the gene at a low enough level to display its full regulatory range.

If our assumption that the high level of expression limits the regulation of genes lacking the spe-1 AflIII-NruI segment is correct, lowering expression by removal of sequences other than the DAR should have similar effects. To that end, we compared polyamine regulation of transformants carrying spe-1 genes lacking either the 5′-UTR AflIII-NruI region (DMH3) or both the 5′-UTR and positively acting sequences in the UAR (DMH4) (Fig. 8A). Reducing expression by removing the UAR might once again allow the full regulatory amplitude of constructs lacking the AflIII-NruI segment to become apparent. Indeed, the reduced expression of spe-1 mRNA lacking both the positively acting UAR and the negatively acting AflIII-NruI segment in transformant DMH4 resulted in greater polyamine regulation of this mRNA compared to that in transformant DMH3, in which the UAR was still present (Fig. 8B). This result supports the finding with the tub::spe-1 transformants that the high level of expression in the absence of the AflIII-NruI sequences limits the amplitude of regulation.

We make the important additional inference that the regulation of spe-1 mRNA expression in DMH4 and tub::spe-1 mRNA expression in DMH11, DMH43, and DMH26 (the last two including the AflIII-NruI segment) must be governed by sequences within their coding or 3′ regions. In this connection, we can also infer that both the UAR and the 5′-UTR are essential for the action of the 5′ PRR: the removal of the UAR from an otherwise normal spe-1 gene causes a substantial reduction in the regulatory amplitude (compare lanes 1 and 7 in Fig. 5) to a level comparable to that imparted by the 3′ region in DMH4, DMH11, DMH43, and DMH26.

The spe-1 3′ region contains a positively acting, polyamine-responsive element.

The question of the location of the remaining regulatory element was addressed by replacing sequences downstream of the spe-1 coding region in transformant DMH11 (lacking 5′-UTR sequences) with the corresponding region of the tub gene. This yielded a tub::spe-1 chimeric gene (in transformant DMH9) in which the only spe-1 region that remained was the ODC coding sequence. The tub::spe-1 mRNA of DMH9 was unresponsive to polyamine starvation, although its expression in spermidine-supplemented cultures was equivalent to that in DMH11 (Fig. 7B). These results imply that the spe-1 3′ region contains sequences required for polyamine-mediated derepression. In-frame deletions within the ODC coding region had no effect on polyamine regulation or the expression of the resulting spe-1 mRNAs (results not shown). We refer to the spe-1 sequences downstream of the ODC termination codon as the 3′ PRR. This segment contains not only the region encoding the 3′ UTR of spe-1 mRNA but also the untranscribed region extending to the HindIII site some 552 nucleotides downstream.

The amplitude of regulation of tub::spe-1 mRNA expressed from the tub promoter in the DMH43, DMH26, and DMH11 transformants is about fourfold in all cases (Fig. 7B). Because the chimeric tub::spe-1 gene in each of these transformants includes the spe-1 3′ PRR, we sought to confirm the inference above that the sequences in the 3′ PRR are solely responsible for the regulation seen in these transformants. The spe-1 3′ sequence in the DMH43 tub::spe-1 gene was replaced with the 3′ end of the tub gene, leaving the spe-1 5′-UTR sequences and coding region flanked by the tub promoter and the tub 3′ region (transformant DMH52) (Fig. 9A). The expression of tub::spe-1 mRNA in this transformant showed little response to changes in cellular polyamine levels and resembled the behavior of tub::spe-1 mRNA in transformant DMH9 (containing only the ODC coding region), even though the AflIII-NruI segment was present in the tub::spe-1 chimeric gene (Fig. 9B). We conclude that the 3′ PRR is wholly responsible for the polyamine regulation observed in transformant DMH43, in which the β-tubulin promoter is attached to the entire spe-1 transcribed region and its downstream flank. In addition, these results indicate again that the AflIII-NruI segment cannot confer polyamine regulation in the absence of upstream, untranscribed spe-1 sequences. We also infer that the presence of the 3′ PRR in the spe-1 transformant lacking both the UAR and 5′-UTR (DMH4) accounts for its polyamine responsiveness.

FIG. 9.

Effects of replacing the spe-1 3′ region with that of the tub gene on spe-1 and tub::spe-1 gene expression. (A) Schematic of tub::spe-1 constructs in transformants DMH1 (wild-type spe-1 gene), DMH52, and DMH10. (B) Northern blot analysis of repressed (SPD) and derepressed (ARG) cultures of transformant DMH52. (C) Northern blot analysis of similar cultures of transformants DMHI and DHMI0. Blots were probed with spe-1 cDNA or the coding region of the tub gene.

We also tested the regulatory effect of the 3′ PRR by replacing the 3′ region of a normal spe-1 gene with that of the tub gene in transformant DMH10 (Fig. 9A). The replacement of spe-1 3′ sequences with those of tub had little effect on spe-1::tub mRNA (Fig. 9C) or ODC activity (results not shown) in spermidine-supplemented cultures compared to those in a transformant carrying the normal spe-1 gene (DMH1). However, neither the spe-1::tub mRNA nor ODC activity in the DMH10 transformant was derepressed in polyamine-starved cultures to the same extent as in the DMH1 transformant. This result indicates that the 3′ PRR is required for part, but not all, of the observed normal depression of spe-1 mRNA in response to polyamine starvation. This supports the conclusions already reached by examination of polyamine regulation in the spe-1::qa-2 transformants and suggests that the 5′ and 3′ regulatory elements act independently.

Regulation of ODC translation by polyamines and the spe-1 5′-UTR.

Change in the rate of ODC translation has been proposed as a regulatory feature in mammalian cells (25). Because sequences in the 5′-UTR might affect the translation of spe-1 mRNA as well as its abundance, we tested the effects of three segmental deletions within the 5′-UTR on ODC translation, and in the process, on mRNA expression (Fig. 10). The abundance of spe-1 mRNA increased upon deletion of the proximal AflIII-Bsu36I region (+101 to +266) or the middle Bsu36I-StuI region (+266 to +403), but deletion of both was required to duplicate the effect of removing most of the 5′-UTR (AflIII-NruI). Deletion of the StuI-NruI region had no effect on mRNA abundance. Therefore, either a negatively acting sequence overlaps the Bsu36I site or there are at least two such elements, one in each of the first two segments.

FIG. 10.

Effects of deletions within the 5′-UTR on spe-1 mRNA abundance in repressed cultures. (A) Schematic of spe-1 5′-UTR deletions. The positions of restriction sites used to make the deletions and the ODC start codon are indicated, along with their distances from the start of transcription. The ODC activity (mean milliunits per milligram of protein ± standard deviation) and activity relative to the normal repressed level (1×) in spermidine-grown cultures of each transformant is given on the right. (B) Northern blot analysis of the transformants grown under repressing conditions. The numbers given across the top correspond to the transformant numbers in panel A. Northern blots were probed with fragments of spe-1 cDNA, and loading was normalized by hybridization with a tub gene probe. The relative abundance of spe-1 mRNA in each transformant, normalized to tub mRNA (data not shown) and relative to that in DMH1, is given below the panel.

We assessed effects on translation by comparing the ratio of ODC activity (and thus of protein) to spe-1 mRNA abundance in repressed cultures of transformants carrying the deletions above. The comparison revealed that the increases of ODC activity were three- to fourfold greater than the increases of spe-1 mRNA abundance in transformants lacking the Bsu36I-StuI region (shown qualitatively in Fig. 10 [compare DMH1, DMH5, and DMH6]). This result suggests that removal of the Bsu36I-StuI region eliminates an impediment to translation of the spe-1 coding region and that this effect is distinct from or overlaps effects on mRNA abundance. The sequence of the spe-1 5′-UTR in this (or any other) region contains no upstream open reading frame or obvious secondary structure that might impede scanning ribosomes.

Polyamines increase the turnover of ODC protein (1), making ODC activity a poor reporter of translational regulation of ODC by polyamines. We therefore used transformants carrying spe-1::qa-2 genes to determine the effects of the spe-1 5′-UTR and polyamines upon translation of the dehydroquinase coding region. Dehydroquinase activity and spe-1::qa-2 mRNA abundance were compared in transformants carrying spe-1::qa-2 genes with (SMH12) and without (SMH13) the spe-1 5′-UTR AflIII-NruI region (Table 1). In repressed cultures, deletion of the spe-1 5′-UTR led to 3.4-fold-greater spe-1::qa-2 mRNA abundance and 10-fold-higher enzyme activity. We interpret these results to indicate that the translational efficiency of the mRNA rises by about threefold when the AflIII-NruI region is removed. We conclude that the presence of the spe-1 5′-UTR impedes the translation of the coding region in the spe-1::qa-2 gene, as it does in the spe-1 gene itself.

TABLE 1.

Effects of spe-1 5′-UTR sequences and polyamine starvation on translation of the qa-2 coding region

Transformant	spe-1::qa-2 mRNAa in:		DHQ activityb in:		DHQ/mRNAc
	SPD	ARG	SPD	ARG	SPD	ARG	ARG/SPDd
SMH12	1.0	1.8	101	347	101	192	1.9
SMH13	3.4	3.3	1,017	993	299	301	1.0

^aRelative abundance of spe-1::qa-2 mRNA in spermidine- (SPD) or arginine (ARG)-supplemented cultures, as determined by Northern blot analysis and normalied to tub mRNA. 

^bDehydroquinase (DHQ) specific activity in nanomoles of dehydroshikimate per milligram of protein at 37°C. 

^cTranslational efficiency of spe-1::qa-2 mRNA, as measured by DHQ specific activities divided by the relative abundances of spe-1::qa-2 mRNA. 

^dThe effect of polyamine starvation is measured as the translation of spe-1::qa-2 mRNA in ARG cultures divided by that in SPD cultures. 

In derepressed cultures, deletion of the 5′UTR sequence led to an approximately 1.8-fold increase in spe-1::qa-2 mRNA and a 2.8-fold increase in enzyme activity, nominally about a 1.5-fold increase in translational efficiency. Indeed, in the transformant with deletion of the 5′UTR, polyamine starvation had little effect on either parameter. We conclude that polyamine starvation partially relieves the translational impediment imposed by 5′-UTR. We have previously inferred that polyamine starvation inhibits translation generally (43), and it is of interest that the relief of the translational block in the spe-1 mRNA mitigates that effect for this particular mRNA.

A reduction in the efficiency of ODC translation could be due to a reduction in either the frequency of ribosomes initiating at the ODC start codon or the elongation rate of translating ribosomes. Reduced initiation should correlate with fewer ribosomes associated with spe-1 mRNA, while reduced elongation rates should correlate with greater ribosomal loading of spe-1 mRNA. Polysomal profiles of various cultures, after probing with spe-1 probes, revealed that spe-1 mRNA could be found in both monosomal and polysomal fractions (Fig. 11). In repressed cultures, all the spe-1 mRNA was associated with ribosomes, indicating that repression was not due to a failure to recruit ribosomes. The removal of the 5′UTR (transformant DMH3) led to a modest increase in the average polysome size. Polyamine starvation led to greater ribosomal loading of both wild-type spe-1 mRNAs and those in which the 5′-UTR had been removed. The results of polysome analysis are consistent with greater ribosomal loading or their impeded elongation or both under conditions of polyamine starvation.

FIG. 11.

Analysis of polysomal mRNA. Data from spermidine- and arginine-grown cultures are shown on the left and right, respectively, and the tops of the gradients are to the right. (A) A₂₅₄ profile of 10 to 40% sucrose gradients. (B) Northern blot analysis of spe-1 mRNA from the aga strain (IC3). (C) Northern blot analysis of spe-1 mRNA from transformant DMH3, lacking the 5′-UTR sequences between the AflIII and NruI sites. (D) Northern blot analysis of tub mRNA of strain IC3 shown in panel B (tub mRNA from DMH3 was similar). Polysomal fractions are aligned below their approximate positions in the gradient.

DISCUSSION

We have shown that expression of spe-1 mRNA depends upon both negatively and positively acting elements of the gene and that the full range of polyamine regulation requires the independent action of both upstream (5′ PRR) and downstream (3′ PRR) PRRs. The 5′ PRR comprises a negatively acting element in the region encoding the 5′-UTR and a collaborating, enhancer-like element in the 5′ UAR of the gene. The positively acting spe-1 3′ PRR lies downstream of the coding region, but its exact location is not known. Its effects are obscured in strains lacking the 5′-UTR, owing to the high level of expression caused by such deletions. In addition, modest translational regulation is imparted by an element in the 5′-UTR of spe-1 mRNA that impedes translation of the downstream coding region. The 5′ and 3′ PRRs, each exerting 3- to 4-fold regulation, can account for the full 10- to 12-fold regulatory amplitude of the mRNA of the intact spe-1 gene.

Polyamine regulation by the 5′ PRR in N. crassa cells is a nuclear phenomenon, most probably transcription, that reduces the abundance of spe-1 mRNA. This conclusion is based on the findings that (i) the stability of spe-1 mRNA is not altered by manipulations of cellular polyamine pools or by the removal of the 5′-UTR and (ii) regulation is governed in part by a sequence upstream of the start of transcription. At present, we cannot distinguish between regulatory mechanisms acting at the initiation of the spe-1 transcript and those that might act on its continued elongation or termination, nor can we exclude potential effects of the 5′-UTR on nuclear export of the spe-1 mRNA. Because the 5′ PRR and 3′ PRR act independently, it is possible that more than one of these steps in spe-1 expression are targets of regulation.

The sequences governing spe-1 expression are distributed over an unusual length of DNA, both upstream and downstream of the coding region. In particular, it is rare that sequences more than 300 bp upstream of the normal transcription start site determine the position of the latter. It is also unusual for sequences almost 600 bp 3′ of the transcription start (the DAR, just 5′ to the coding region) or 3′ to the coding region to influence mRNA abundance without having any effect on its stability. Recently, Pollard et al. (27a) demonstrated that polyamines act to repress the transcription of a subset of genes in the yeast Saccharomyces cerevisiae, potentially by stabilizing condensed chromatin states. In addition, they showed the activity of the GCN5 histone acetyltransferase antagonizes these transcriptional effects of polyamines. We suggest that the sequences affecting spe-1 mRNA abundance may be part of a larger chromatin domain that may be remodeled or modified as cells respond to polyamines.

The 5′ PRR is particularly complex. It includes sequences in both the upstream, untranscribed region and the 5′-UTR. The −1000 to −778 region (within the UAR) contains several poly(dA-dT) sequences characteristic of upstream transcriptional activators in other N. crassa and S. cerevisiae genes (11, 37). Curiously, three copies of an 11-bp conserved sequence element found in the UAR [CCCTCC(A/T)CCAC] do not appear to play much of a role in regulation, as judged from the deletion analysis. Clearly, the role of the UAR will have to be examined in the absence of the 3′ PRR. Because deletion of sequences upstream of and overlapping the transcription start site (−148 to +22, in transformant PΔ7) impairs expression and affects initiation site selection, we cannot exclude the possibility that polyamine-responsive regulatory sequences also reside in this region. Furthermore, although sequences in the 5′-UTR are required for regulation (as demonstrated in the spe-1::qa-2 transformants) and have negative effects on expression (in all contexts in which they were tested), more extensive analysis of this region is required to determine whether these effects are attributable to a single sequence element.

In yeast (10) and mammalian (34) cells, ODC mRNA abundance is unresponsive to fluctuations in the levels of cellular polyamines. In fact, the response of ODC mRNA abundance to polyamine status in N. crassa is unique among eukaryotes. However, regulation of ODC in N. crassa and mammalian cells has some common features. In both, polyamines reduce ODC activity by accelerating the degradation of ODC protein (1, 33). Second, ODC mRNA from most organisms has an extremely long 5′-UTR. In mammalian cells, the ca. 200-nucleotide 5′-UTR is GC rich and may form stable secondary structures (42). The removal of these sequences increases the rate of translation of the ODC coding region (16, 19, 29). However, polyamines do not appear to act through these sequences to regulate ODC translation (39–41), but some debate on this issue remains (26, 35). Our results show that the long 5′-UTR of the N. crassa spe-1 mRNA impedes translation of the ODC coding region, although it lacks the obvious secondary structures found in mammalian ODC mRNA (43). The 5′-UTR of mammalian ODC mRNA also contains a short upstream open reading frame, absent in the N. crassa spe-1 5′-UTR, but this open reading frame has little effect. Our results indicate that polyamine starvation partially relieves the impediment to translation imposed by the spe-1 5′-UTR of N. crassa. While the spe-1::qa-2 chimeric genes used in our studies of translation lacked spe-1 3′-UTR sequences, further investigations are required to determine the role, if any, of spe-1 3′-UTR sequences in ODC translation in N. crassa, as they appear to have in the mammalian system (12, 17).

The regulation of spe-1 mRNA expression by polyamines distinguishes N. crassa polyamine regulation from that in all other eukaryotic organisms studied to date. Why does N. crassa go to such extraordinary lengths to regulate ODC activity, employing mechanisms acting at almost every level of gene expression? The constitutive derepression of ODC in transformants lacking the 5′-UTR does not adversely affect their growth in laboratory culture, and other mechanisms appear to prevent spermidine accumulation to toxic levels (27). The information so far indicates that most organisms do not require precise control of cellular polyamine levels (6). To accomplish this, an allosteric enzyme regulatory mechanism would be required, and no organism known displays a polyamine-responsive ODC enzyme. The mechanisms governing ODC activity in N. crassa appear simply adequate to maintain the small amount of polyamine synthesis needed to sustain growth while preventing wasteful use of nitrogen over the long term (6). While the inefficient and redundant modes of regulation appear to be individually rather insignificant, together they may contribute substantially in nature to the fitness of the organism.