TEXT

Top Abstract Text References

Silencing of gene expression by transcription of an artificial antisense construct has been successfully employed with fungi for a number of practical applications (e.g., see references 10 and 16). This approach is particularly useful when the target gene is represented by multiple copies in the genome (5) or when complete disruption of the gene function is lethal or gives an undesirable phenotype. In this study, the antisense technique was successfully employed to suppress the broad-domain carbon catabolite repression in Aspergillus nidulans. The key component of this repression is the creA gene. Most commercially important promoters in aspergilli are subject to the CREA-mediated repression. Loss-of-function creA mutants usually exhibit impaired growth parameters (15) and are inapplicable for industrial protein production. In this study, partial alleviation of glucose repression was achieved without affecting growth parameters of the fungus. Production of intracellular and secreted glucose-repressible enzymes (heterologous secreted -amylase, endogenous intracellular alcohol dehydrogenase, and secreted endoarabinase) increased substantially.

Strain construction. The A. nidulans strain G 861 (yA2 argB2 trpC801) was employed as the recipient for transformation. For solid medium and shake-flask cultivation, standard growth media and culturing techniques were used (7). The Aspergillus oryzae -amylase gene on pTAKA-17 (courtesy of Novo Nordisk A/S) (6) was introduced into the strain G 861 by cotransformation with pTA11 (12). The transformants were analyzed by Southern blotting (Hybond N+ membranes; suppliers' protocol), and a single-copy transformant (T1) was selected. The creA antisense expression vector contained a BamHI/XhoI fragment of pTA11 with the terminator of the trpC gene (12), a 700-bp BamHI/NcoI fragment from pCD5 containing a fragment of creA cDNA, and an EcoRI/NcoI fragment from pEC7 containing PgpdA (provided by B. Felenbok, Institut de Génetique et Microbiologie, Université Paris-Sud, Paris, France) assembled in the plasmid pILJ16 with the argB gene as a marker (9). The construct, termed pMH-C, is represented in Fig. 1A. This plasmid was introduced into the original transformant by transformation with selection for the Arg⁺ phenotype. For a control, a similar plasmid without the insert of creA cDNA was introduced into the same original transformant. Arg⁺ transformants obtained with pMH-C were indistinguishable from one another and from the control on complete and minimal media with or without glucose. Screening of the transformants for the derepressed phenotype was performed by measuring clearing zones on starch plates in the presence of different glucose concentrations (1 to 5%). Most transformants formed larger clearing zones than the control. Three transformants were selected, with the clearing zones ranging from 0.2 (T59) to 0.5 cm (T24 and T62). Southern blot hybridization suggested integrations into different genomic sites, probably in more than one copy in the case of T24.

View larger version (61K): [in this window] [in a new window]   FIG. 1.   (A) The antisense expression plasmid pMH-C with an exploded view of the expression cassette. Base pairs are numbered from the transcription start of the antisense RNA, which lies within the gdhA promoter-containing fragment. The putative antisense transcript and the creA-encoding mRNA (8) are shown as broken lines, with the potential pairing region indicated. The CREA polypeptide is shown as a solid arrow underneath. Positions of oligonucleotide probe complementary to the sense (SCA) and antisense (ACA) creA RNA are indicated by small arrows pointing towards the 3' ends of the oligonucleotides. The sequences of the probes ACA and SCA were 5'-CACATGCCGCAACCAGGATCGTCAGTGG-3' and 5'-TAAGAATGAGAGCCGGAATGGAGGATGG-3', respectively. (B) Northern blot hybridization of RNA from A. nidulans strain G051 (wt), the control transformant (C), and antisense transformants (as indicated). The probes (indicated underneath the corresponding panels) were either radioactively labeled DNA fragments of creA cDNA, the -amylase gene from A. oryzae ("amy"), the alcA and actin gene from A. nidulans, or the 3'-end-labeled oligonucleotide probes ACA and SCA. The position of the endogenous creA DNA fragment and its transcript on panel B is indicated by the solid arrow. The panel hybridized with the creA cDNA was produced from a separate blot with poly(A)+-RNA and can be used to estimate relative amounts of the sense and antisense transcripts.

Growth, productivity, and gene expression patterns during batch cultivation. Batch cultures were carried out in an in-house 5-liter glass fermentor as described previously (3, 11);under these conditions, -amylase synthesis is strongly induced. -Amylase activity was determined using an Amyl kit (Boehringer Mannheim GmbH, Mannheim, Germany) in a Cobas Mira analyzer (F. Hoffman-La Roche Ltd., Basel, Switzerland). For endoarabinase activity measurements, commercial Arabinazyme tablets (Megazyme Int. Ireland Ltd.) were used according to the instructions given by the manufacturer. Glucose in the fermentation medium was determined with the Unimate 7 Gluc GDH kit (F-Hoffman-La Roche Ltd.) using a Cobas Mira analyzer. DNA and RNA were isolated from fermenter biomass samples taken at the early exponential phase for RNA or at the end of the experiment for DNA (in order to make sure that the inserts of recombinant DNA were not rearranged during the cultivation). DNA was isolated by phenol extraction (2). RNA was isolated using the Trizol reagent (Gibco BRL) (supplier's protocol). All auxiliary DNA and RNA techniques were performed as described by Sambrook et al. (14).

1

creA 4

Chemostat experiments. One of the antisense transformants (T24) and the control strain were further characterized in glucose-limited chemostat experiments. These experiments were used to study the effects of the glucose concentration and the specific growth rate on the productivity of two repressible enzymes, the recombinant -amylase and endogenous endoarabinase. Chemostat experiments were performed with a 2-liter Biostat M fermentor (B. Braun Melsungen AG, Melsungen, Germany). The experiments began as batch cultures, and the continuous feed started when the glucose concentration in the medium dropped below 0.5 g/liter. The residual glucose concentration in the medium was low (<10 mg/liter) and increased with the specific growth rate. The glucose uptake rate (r_s) increased linearly with the dilution rate (which is equal to the specific growth rate), and from the linear plot the true yield coefficient (Y_sx^true) and the maintenance coefficient (m_s) were estimated using the following equation:

(1)

The values of Y_sx^true obtained for both strains (Table 1) are very similar and agree well with results reported for the wild-type A. nidulans (4), indicating that the basic growth parameters have not been significantly affected for the antisense transformant. The values of Y_sx^true are similar to the yield coefficients estimated from the batch cultivations.

1

View this table: [in this window] [in a new window]   TABLE 2.   Specific productivities of -amylase and endoarabinase at different dilution rates in glucose-limited chemostats

Conclusions. The experimental settings of this study were deliberately chosen to emulate a typical industrial process rather than to investigate the mechanism of antisense silencing in any detail. The strategy was typical for antisense techniques in that we used a relatively short transcription template covering the 5' moiety of the target. Although there is still no consensus on the mechanisms of antisense silencing (13), it seems likely that in our study the effect was due to impaired capping-ribosome binding-translation initiation reactions without physical degradation of the target mRNA. The data indicate that the antisense strategy is promising for creation of strains of filamentous fungi for enzyme production. In particular, it is possible to achieve carbon catabolite derepression in Aspergillus at a substantial level without affecting the growth rate and morphology. Phenotypic abnormalities have been previously reported for creA mutants with pronounced derepression (e.g., see references 8 and 15). Our data suggest either that the morphological manifestations require a decrease in protein activity much larger than that achieved by this technique or that these manifestations are less dependent on the de novo CREA synthesis and hence on the mRNA status.