Inducible RNA Interference of brlAβ in Aspergillus nidulans

Abstract

An inducible RNA interference (RNAi) construct composed of inverted repeating alcA promoters flanking the developmental regulatory gene brlAβ was tested in Aspergillus nidulans. On inducing medium, the RNAi strains failed to sporulate and lacked brlAα and brlAβ expression. RNAi was specific for brlAβ, but not brlAα, silencing, indicating brlAα regulation by brlAβ.

Aspergillus nidulans has been a filamentous fungal model for RNAi using inverted repeat transgenes (10, 11) and short interfering RNAs (siRNAs) (15). A. nidulans reproduces asexually by developing radial hyphae and aerial conidiophores whose tips differentiate into vesicles, primary and secondary sterigmata (metullae and phialides, respectively), and uninucleate conidia (3, 6, 18). bristle (brlA) mutants maintain normal hyphal extension, but conidiophores are elongated and fail to develop viable conidia (9). brlA null mutations result in the complete obliteration of conidiophores, and leaky mutations may form vesicles and sterigmata but not viable conidia (3, 9). The leaky mutants are thought to have partial brlA expression, and variability between phenotypes may be a result of a dose effect (3, 23). brlA has two overlapping transcripts, brlAα and brlAβ (Fig. 1A). The brlAβ transcript begins at −851 and has an intron between −99 and +293 and extends to +2085. The brlAα transcript begins within the brlAβ intron at +1 and extends to +2085 (23).

FIG. 1.

brlA locus and RNAi construct. (A) The brlA locus consists of two overlapping transcriptional units, brlAα and brlAβ. The top line represents the brlA genomic DNA (gDNA). The portion of the locus flanked by BamHI sites (−2902 to −404) used in the RNAi construct included only brlAβ. The line of the transcripts (mRNA) represents untranslated RNA, and the box represents translated mRNA. brlAβ has one intron (−99 to +293) where brlAα transcription begins. (B) The RNAi construct consists of inverted repeats of inducible alcA promoters [alcA_(p)] flanking brlAβ in a unique BamHI site. The RNAi construct also contains an argB marker for the nutritional selection of transformants.

Our RNAi construct (Fig. 1B) flanks brlAβ with inverted repeats of an inducible alcohol dehydrogenase promoter that is repressed on glucose but strongly induced on threonine (1). Double-stranded RNA is transcribed, triggering the RNAi mechanism to silence brlAβ. To engineer the construct, alcA promoters were cloned as inverted repeats into EcoRI and NotI sites of a plasmid containing the brlAβ sequence (−2902 to −404) cloned into the BamHI restriction site on pBluescript KS(+). An argB nutritional selection gene was cloned into the SacI site (14), and the final construct was sequenced for verification.

A. nidulans strains RMS11 (pabaA1 yA2 argB::trpC veA1 trpC801) and LR191 (pyrG pabaA1 yA2 argB pyroA veA1) were transformed with the construct as previously described (26), except that germlings were digested with Driselase (Sigma) instead of Novozyme234. Eight primary transformants were obtained, and two of these RNAi strains, JB1 and JB3, were chosen for further analysis. The integration of the construct into the genomes was verified with PCR and Southern blot analysis (data not shown) of genomic DNA. Southern blots also verified that the native brlA gene was not disrupted by homologous recombination.

In submerged culture, A. nidulans conidiophore development is suppressed, but spores develop vegetatively to produce mature mycelia. Conidiophore development begins in a synchronized manner when the mycelia are plated and exposed to air (5, 6). The wild-type, JB1, and JB3 spores were inoculated in minimal medium (22) broth and shaken for 24 h at 250 rpm. Mycelia were harvested by filtration, washed thoroughly with sterile water, and plated on glucose and threonine (Fig. 2A). After 24 h of developmental induction, the wild type developed yellow-pigmented conidia on both glucose and threonine, but the RNAi strains developed conidia only on glucose (Fig. 2A). On threonine, JB1 displayed a brlA null phenotype, but JB3 displayed a brlA leaky phenotype that develops rudimentary vesicles and sterigmata (Fig. 2B and C). This indicated that brlA is silenced to various degrees when the RNAi strains develop on threonine, but it is not silenced on glucose.

FIG. 2.

Phenotypes of RNAi strains. (A) Plates that contained either glucose (G) or threonine (T) as the sole carbon source were scanned with a Microtek ScanMaker 4700. (B) Plates were photographed with a Canon Powershot A620 under a Nikon SMZ-U stereomicroscope at ×60 magnification. (C) Tape mounts were taken from the plates and photographed under a Nikon TMS microscope at ×1,000 magnification. These images show the wild type (WT) under both conditions and the RNAi strains (JB1 and JB3) on glucose with mature conidiophore development and viable conidia. The RNAi mutants on threonine show brlA phenotypes.

In brlA expression analysis, RNA was isolated with Tri reagent (Sigma-Aldrich) from mycelia 24 h after being plated. Northern blots and reverse transcription-PCR (RT-PCR) showed a drastic reduction of both brlAα and brlAβ on threonine compared to that on glucose (data not shown). In order to quantify the level of change (n-fold) of the brlAα and brlAβ transcripts, real-time RT-PCR on an Applied Biosystems 7500 real-time PCR system was employed (Fig. 3A). For analysis, the threshold cycle (C_T) value was set at 1.5, and ΔΔC_T calculations were performed relative to values for actin and the mean ΔC_T. The changes (n-fold) in gene expression were derived from 10^{(−ΔΔCT/ln10/ln2)}, where the primer efficiency (ln10/ln2) was calculated experimentally for each primer set. In the wild type, brlAα and brlAβ expression was not reduced on threonine. In JB1, brlA expression on glucose is approximately equal to that of the wild type on glucose, but there is only a slightly detectable abundance of brlAα and brlAβ on threonine. JB3 had a large amount of brlAα and brlAβ on glucose, nearly fivefold more than the wild type, but brlAα and brlAβ expression on threonine was three- to fourfold less abundant than that on glucose. These experimental results imply that the expression of brlA is silenced in the RNAi strains on threonine. The silencing in JB1 is more drastic than the silencing in JB3, which corresponds to the respective brlA null and leaky phenotypes.

FIG. 3.

brlA expression and siRNA analysis. (A) Primers used in real-time RT-PCR were specific for brlAα, brlAβ, and the overlapping portion of brlAα and brlAβ (brlAαβ). The wild type (WT) had normal expression of brlA on both glucose (G) and threonine (T). JB1 had normal brlA expression on glucose but almost undetectable expression on threonine. JB3 had high levels of brlA expression on glucose but reduced expression on threonine. (B) siRNAs specific for brlAβ in the RNAi strains on threonine were detected. No siRNAs were detected in the wild-type or RNAi strains on glucose. Untreated probes (+) or probes treated with RNaseA/T1 (−) were used as controls. (C) No siRNAs were detected using a probe specific for the overlapping portion of brlAα and brlAβ. This indicates that the RNAi mechanism is specifically targeting for brlAβ, not brlAα. Low-molecular-weight (MW) RNA is shown to verify its integrity.

With a few rare exceptions, RNAi is sequence specific but not locus specific in fungi (20). A. nidulans contains two RNA-dependent RNA polymerases that could amplify the RNAi signal through transitive RNAi, which has been shown to occur in Caenorhabditis elegans (4), but transitive RNAi has been absent in previous A. nidulans silencing experiments (11). Furthermore, if the RNAi signal were amplified, all RNAi strains would have null brlA phenotypes. To investigate RNAi specificity, Northern blot analyses were performed with ³²P-UTP-labeled RNA probes constructed with a Maxiscript T7 kit (Ambion). In antisense Northern blots (data not shown), a band specific for brlAβ was present in the RNAi strains on threonine, but antisense RNA specific for brlAα was undetectable. For siRNA analysis (Fig. 3B and C), sodium carbonate-treated probes (7) were used with a mirVana microRNA detection kit (Ambion). siRNAs specific for brlAβ (Fig. 3B) were present in the RNAi strains on threonine but not in the wild type or on glucose. No siRNAs were detected for the overlapping portion of brlAα and brlAβ (Fig. 3C). siRNAs were more abundant in JB1 than JB3, corresponding to the respective brlA null and leaky phenotypes. These blots verified that RNAi mechanisms are the direct cause of brlAβ silencing and that brlAβ, not brlAα, is the target of RNAi.

brlA regulates developmental genes such as abaA (2, 19), wetA (16), rodA (8, 24), and stuA (17). The most current models suggest that brlAβ initiates asexual development, since it is present in small quantities in vegetative cells, and brlAα continues the development through a feedback mechanism with abaA (3, 12, 13). Furthermore, in a brlAβ knockout mutant, brlAα expression was not detected after 12 h of developmental induction (23), and the overexpression of brlAβ induces the expression of brlAα even when abaA is not present (12). This evidence and our results strongly suggest that brlAβ plays a key role in regulating the expression of brlAα.

The novel RNAi construct serves as a powerful genetic tool, because RNAi is easily induced and any gene of interest may be cloned into the single BamHI site in either orientation. Problems associated with lethal knockouts of essential genes is circumvented, because these strains are isolated on glucose and then characterized by RNAi induction on threonine. Additionally, the one-step cloning of a gene in either orientation is advantageous over the currently popular use of inverted repeat transgenes, where the gene must be cloned twice in specific orientations to produce double-stranded hairpin RNA (10, 11, 21, 25).