Essential gene identification and drug target prioritization in Aspergillus fumigatus.

Abstract

Aspergillus fumigatus is the most prevalent airborne filamentous fungal pathogen in humans, causing severe and often fatal invasive infections in immunocompromised patients. Currently available antifungal drugs to treat invasive aspergillosis have limited modes of action, and few are safe and effective. To identify and prioritize antifungal drug targets, we have developed a conditional promoter replacement (CPR) strategy using the nitrogen-regulated A. fumigatus NiiA promoter (pNiiA). The gene essentiality for 35 A. fumigatus genes was directly demonstrated by this pNiiA-CPR strategy from a set of 54 genes representing broad biological functions whose orthologs are confirmed to be essential for growth in Candida albicans and Saccharomyces cerevisiae. Extending this approach, we show that the ERG11 gene family (ERG11A and ERG11B) is essential in A. fumigatus despite neither member being essential individually. In addition, we demonstrate the pNiiA-CPR strategy is suitable for in vivo phenotypic analyses, as a number of conditional mutants, including an ERG11 double mutant (erg11BDelta, pNiiA-ERG11A), failed to establish a terminal infection in an immunocompromised mouse model of systemic aspergillosis. Collectively, the pNiiA-CPR strategy enables a rapid and reliable means to directly identify, phenotypically characterize, and facilitate target-based whole cell assays to screen A. fumigatus essential genes for cognate antifungal inhibitors.

Full text

Page 1

Essential Gene Identification and Drug Target
Prioritization in Aspergillus fumigatus
Wenqi Hu1, Susan Sillaots1, Sebastien Lemieux1¤, John Davison1, Sarah Kauffman2, Anouk Breton1, Annie Linteau1,
Chunlin Xin1, Joel Bowman3, Jeff Becker2, Bo Jiang1, Terry Roemer1*
1 Merck Frosst Center of Fungal Genetics, Montreal, Quebec, Canada, 2 Department of Microbiology, University of Tennessee, Knoxville, Tennessee, United States of America,
3 Infectious Diseases, Merck Research Laboratories, Rahway, New Jersey, United States of America
Aspergillus fumigatus is the most prevalent airborne filamentous fungal pathogen in humans, causing severe and often
fatal invasive infections in immunocompromised patients. Currently available antifungal drugs to treat invasive
aspergillosis have limited modes of action, and few are safe and effective. To identify and prioritize antifungal drug
targets, we have developed a conditional promoter replacement (CPR) strategy using the nitrogen-regulated A.
fumigatus NiiA promoter (pNiiA). The gene essentiality for 35 A. fumigatus genes was directly demonstrated by this
pNiiA-CPR strategy from a set of 54 genes representing broad biological functions whose orthologs are confirmed to be
essential for growth in Candida albicans and Saccharomyces cerevisiae. Extending this approach, we show that the
ERG11 gene family (ERG11A and ERG11B) is essential in A. fumigatus despite neither member being essential
individually. In addition, we demonstrate the pNiiA-CPR strategy is suitable for in vivo phenotypic analyses, as a
number of conditional mutants, including an ERG11 double mutant (erg11BD, pNiiA-ERG11A), failed to establish a
terminal infection in an immunocompromised mouse model of systemic aspergillosis. Collectively, the pNiiA-CPR
strategy enables a rapid and reliable means to directly identify, phenotypically characterize, and facilitate target-based
whole cell assays to screen A. fumigatus essential genes for cognate antifungal inhibitors.
Citation: Hu W, Sillaots S, Lemieux S, Davison J, Kauffman S, et al. (2007) Essential gene identification and drug target prioritization in Aspergillus fumigatus. PLoS Pathog 3(3):
e24. doi:10.1371/journal.ppat.0030024
Introduction
Aspergillus fumigatus is a ubiquitous soil-dwelling sapro-
phytic fungus that propagates through the prolific produc-
tion of air-borne conidia [1]. Large-scale genome
comparisons have shown that no gene sets are shared
exclusively by both Aspergillus fumigatus and any other human
pathogenic fungi sequenced to date, such as Candida or
Cryptococcus species [2]. Thus, it has been recently suggested
that A. fumigatus pathogenesis is based on its saprophytic
lifestyle in combination with the immunosuppressed state of
the host, rather than from genuine fungal virulence factors
[2]. Although A. fumigatus conidia are constantly inhaled and
seldom cause serious medical conditions in healthy individ-
uals, immunocompromised patients (e.g., those with HIV
infection or AIDS, solid organ and bone marrow transplant
recipients, and those receiving chemotherapy) are highly
susceptible to invasive aspergillosis, a fatal systemic infection
[1,3,4]. Current treatment options for invasive aspergillosis
are limited to three classes of antifungal therapeutics:
polyenes (amphotericin B and various liposomal formula-
tions), azoles (e.g., fluconazole, voriconazole, itraconazole),
and, more recently, semisynthetic echinocandins (e.g., caspo-
fungin and anidulafungin) [4,5]. Despite current therapeutic
options, mortality associated with invasive aspergillosis
remains high (ranging from 60% to 90%) and more effica-
cious antifungal drugs with novel mechanisms of action are
needed [4,5].
The identification of conserved essential genes required for
the growth of fungal pathogens offers an ideal strategy for
elucidating novel antifungal drug targets. A comprehensive
determination of all the essential genes has been achieved in
the nonpathogenic yeast Saccharomyces cerevisiae [6,7]. Extend-
ing similar genetic approaches to fungal pathogens has
proved difficult due to limited available molecular technol-
ogies and to the asexual nature of most medically relevant
fungi, preventing the use of classical genetics. Nonetheless,
large-scale essential gene identification in Candida albicans has
begun through a number of alternative approaches, including
antisense-based gene inactivation [8], transposon-based het-
erozygote screens for hypomorphs [9], homozygote null
mutants [10], and a promoter replacement strategy to
construct conditional mutants [11].
Large-scale functional analysis and essential gene identifica-
tion in A. fumigatus have proved more difficult. Although gene
disruption methodologies have been adapted to A. fumigatus,
they are limited due to the organism’s poor efficiency of
homologous recombination as well as the inherent inability to
study essential genes by such means. A. fumigatus essential genes
have been defined using parasexual genetics in which gene
Editor: Brendan P. Cormack, Johns Hopkins University School of Medicine, United
States of America
Received May 9, 2006; Accepted January 8, 2007; Published March 9, 2007
Copyright: ? 2007 Hu et al. This is an open-access article distributed under the
terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author
and source are credited.
Abbreviations: ACM, Aspergillus complete medium; AMM, Aspergillus minimal
medium; CFU, colony-forming units; CPR, conditional promoter replacement; MIC,
minimum inhibitory concentration; pNiiA, A. fumigatus NiiA nitrogen-regulatable
promoter; RT-PCR, reverse transcription PCR
* To whom correspondence should be addressed. E-mail: terry_roemer@merck.
com
¤ Current address: Institute of Research in Immunology and Cancer, Universite ´ de
Montre ´al, Montreal, Quebec, Canada
PLoS Pathogens | www.plospathogens.orgMarch 2007 | Volume 3 | Issue 3 | e240001

Page 2

essentiality is inferred from the failure to recover haploid
segregants carrying a gene knock-out [12,13]. Nevertheless,
such an approach is likely unsuitable to systematically identify
all possible essential genes due to irregularities in the para-
sexual cycle of A. fumigatus. Direct demonstration of A. fumigatus
gene essentiality and phenotypic analyses, however, may be
achieved using molecular genetics strategies including RNA
interference or promoter replacement strategies [14,15].
Mouyna et al. (2004) have used RNA interference to produce
the predicted phenotypes associated with both a nonessential
gene involved in melanin biosynthesis (alb1) and an essential
glucan synthase, FKS1 [14]. (Note: In this report, we maintain
gene nomenclature for A. fumigatus genes as previously
described [e.g., alb1 or FKS1]; in cases where previously
uncharacterized A. fumigatus genes are described, standard S.
cerevisiae gene nomenclature and provisional gene names are
adoptedaccordingtotheiryeastortholog.)Todate,conditional
promoter replacement (CPR) strategies applied to A. fumigatus
have been restricted to heterologous promoters including the
Aspergillus nidulans alcA promoter and Escherichia coli tetracy-
cline-regulated promoter [15,16].
Here we report a CPR strategy to identify A. fumigatus
essential genes and to prioritize potential antifungal drug
targets. This strategy uses the A. fumigatus NiiA nitrogen-
regulatable promoter (pNiiA) to delete and replace the
endogenous promoter of selected genes. By applying this
strategy to 54 A. fumigatus genes of diverse biological
functions whose orthologs are known to be essential for
growth in S. cerevisiae and C. albicans, we have identified 35
genes essential for mycelial growth, with many displaying a
cidal terminal phenotype. We also demonstrate that the
ERG11 gene family is essential in A. fumigatus and that the
resulting pNiiA-CPR strains may be used as cell-based whole
cell assays to examine target-specific chemical hypersensitiv-
ity. Finally, we show that a number of pNiiA-CPR mutants of
essential genes fail to establish a terminal aspergillosis
infection in a mouse model system. Therefore, both in vitro
and in vivo phenotypic analysis of essential genes may be
performed. This initial gene set comprises an experimentally
validated drug target set that is broadly conserved within
fungal pathogens and therefore of significant relevance to
antifungal drug discovery.
Results
Construction of A. fumigatus Conditional Mutant by
Promoter Replacement
A. fumigatus conditional mutants were constructed by CPR
using pNiiA. pNiiA was chosen based on previous work in
both A. fumigatus and A. nidulans demonstrating its tight
nitrogen-dependent transcriptional regulation [17,18]. Two
operationally independent signals control pNiiA expression:
an inducer (e.g., nitrate or other secondary nitrogen source)
and a repressor (e.g., ammonium or other primary nitrogen
source). Expression is achieved solely in the absence of
ammonium and the presence of nitrate; however, repression
is achieved by the presence of ammonium regardless of the
presence of other nitrogen [17,18]. pNiiA-CPR mutants were
constructed by homologous recombination-mediated dele-
tion of the endogenous promoter (approximately 250 base
pairs of promoter sequence immediately preceding the start
codon of the gene) and replacement with an NiiA promoter
cassette marked with the pyrG selectable marker (Figure 1A
and 1B; see Materials and Methods for details). To test the
pNiiA-CPR strategy, TUB1 and MET2 were first selected.
TUB1 encodes the broadly conserved and essential protein a-
tubulin, and MET2 encodes a homoserine O-acetyltransferase
involved in methionine biosynthesis, mutations of which
cause methionine auxotrophy. TUB1 and MET2 pNiiA-CPR
mutants displayed wild-type mycelial growth on inducing
medium (Aspergillus minimal medium [AMM] plus nitrate),
but neither mutant was able to grow on repressing medium
(AMM plus ammonium, Figure 1C). Furthermore, the A.
fumigatus MET2 conditional mutant displayed a tight methio-
nine auxotrophy that was fully suppressed by methionine
added to the medium (Figure 2A). pNiiA-CPR conditional
mutants of GFA1 and ALR1, whose orthologs are essential in
both S. cerevisiae and C. albicans, were also examined (Figure
1C). GFA1, a glutamine-fructose-6-phosphate aminotransfer-
ase catalyzing the first step in the chitin biosynthesis pathway,
was shown to be essential in A. fumigatus. However, CPR
mutants of ALR1 (a metal cation transporter belonging to a
gene family in A. fumigatus) failed to display any growth defect
under repressing conditions, revealing that ALR1 is non-
essential in A. fumigatus (see below).
According to the severity of the growth phenotype
displayed by CPR mutants in repressing conditions, qualita-
tive scores were assigned to each of the CPR mutants to
classify the terminal phenotype (Figure 1D). A 4þ shutoff
phenotype was assigned to those strains (e.g., GFA1) that
completely fail to grow on repressing medium after 48 h at 30
8C. Similarly, a 3.5þor 3þshutoff phenotype was used to score
those strains that showed a very severe (e.g., PFS2) or a severe
(e.g., ALG7) growth phenotype, respectively. Other growth
defects were scored as 2þ(mild), 1þ(minor; e.g., CDC24), or 0þ
(no defect; e.g., ALR1). All essential genes described here
showed either a complete absence of growth (4þ)or a dramatic
growth defect (3.5þ and 3þ) under repressing conditions.
Evaluating pNiiA-CPR Reliability by Genetic Approaches
To test its reliability, the pNiiA-CPR strategy was further
applied to multiple A. fumigatus genes whose null phenotype
PLoS Pathogens | www.plospathogens.orgMarch 2007 | Volume 3 | Issue 3 | e24 0002
A. fumigatus Essential Gene
Author Summary
Aspergillus fumigatus is an opportunistic filamentous fungal
pathogen of emerging clinical significance. Although virulence
factors are seen as potential drug targets, neither genetic analyses
nor genomic comparisons have identified genuine virulence factors
in A. fumigatus. Essential genes required for fungal growth and
viability also serve as potential drug targets, yet few have been
described in this pathogen. To begin to catalog essential genes in A.
fumigatus, we devised a genetic strategy for creating conditional
mutants by promoter replacement of target genes using a nitrogen-
regulated promoter. Applying this genetic approach to A. fumigatus
genes orthologous to known essential genes of the nonpathogenic
yeast, Saccharomyces cerevisiae and Candida albicans, we demon-
strate a robust enrichment for identifying essential genes conserved
within this pathogen. We show that A. fumigatus conditional
mutants can be evaluated according to their terminal phenotypes
(e.g., conidial germination, growth, morphology, and cidal versus
static consequences) and pathogenesis in a murine model of
systemic aspergillosis to prioritize essential genes as novel drug
targets suitable for developing broad-spectrum antifungal agents.

Page 3

was either known or independently verified by other means.
First, pNiiA-CPR mutants were constructed for five genes
(HIS3, TRP5, MET16, LYS9, and LYS4) involved in amino acid
biosynthesis. Growth was completely impaired for all strains
under repressing conditions lacking amino acids but fully
restored when the cognate amino acid was provided (Figure
2A). Second, pNiiA-CPR mutants were constructed for four
genes (FKS1, GUS1, SPE2, and HEM1) whose essentiality in A.
fumigatus has been demonstrated by a parasexual genetic
strategy [12,13]. In each instance, the conditional mutant
displayed an essential growth phenotype (Figure 2B). Third,
we constructed pNiiA-CPR mutants of two nonessential genes
(ayg1 and arp2) involved in conidial pigmentation whose null
phenotypes result in the production of yellow-green and
pink-red conidiospores, respectively [19]. Hence, conidial
pigmentation phenotypes of ayg1 and arp2 mutants serve as
whole cell reporters to monitor the level of repression
achieved by the pNiiA-CPR system. Indeed, pNiiA-ayg1 and
pNiiA-arp2 conidia color shifted from dark-green (wild-type
conidia color) under inducing conditions to yellow-green
(ayg1) or red-pink (arp2) under repressing conditions (Figure
2C). Finally, we applied this pNiiA-CPR strategy to A. fumigatus
nudC, a gene shown to be essential by promoter replacement
with A. nidulans alcA promoter [15]. Consistent with a previous
report [15], the pNiiA-nudC mutant displayed (1) an essential
growth phenotype (Figure 2D), (2) a germination defect of
conidia with approximately 70% of conidia failing to
germinate and approximately 30% of conidia forming short
growth-arrested germ tubes (Figure S1), and (3) an nudC
Figure 2. Genetic Evaluation of the pNiiA-CPR Strategy
(A) Phenotypes of pNiiA-CPR mutants for HIS3, TRP5, MET16, MET2, LYS9,
and LYS4. Under standard repressing conditions (Re, AMM plus
ammonium), all strains lacked detectable growth (4þ phenotype).
Growth was unimpaired under inducing conditions (In; AMM plus
nitrate). Growth phenotypes of pNiiA-CPR mutants under repressing
conditions were specifically suppressed if the cognate amino acid was
provided to the growth media (His, histidine; Trp, tryptophan; Met,
methionine; or Lys, lysine).
(B) Growth phenotypes of pNiiA-CPR mutants for the previously reported
A. fumigatus essential genes, FKS1, GUS1, SPE2, and HEM15 [12,13].
Reproducible 4þ essential growth phenotypes are observed for each
pNiiA-CPR mutant under repressing conditions with the exception of
FKS1, which produced a 3.5þ growth phenotype.
(C) arp2 and ayg1 conidial color phenotypes by the pNiiA-CPR strategy.
Wild-type (WT) strain CEA10, pNiiA-arp2, and pNiiA-ayg1 mutants display
normal dark-green conidia color under inducing conditions. Gene-
specific conidia color phenotypes characteristic of their known null
phenotype [19] are specifically detected under repressing conditions.
(D) nudC growth phenotype by the pNiiA-CPR strategy. Highly
reproducible growth and morphological phenotypes associated with
nudC mutants are observed (see Figure S1) as similarly determined using
an alcA heterologous conditional promoter [15].
doi:10.1371/journal.ppat.0030024.g002
Figure 1. Outline of the A. fumigatus pNiiA-CPR Strategy
(A) A cloning vector, pPyrG-pNiiA, was created specifically for the
construction of CPR cassettes. This plasmid contains an A. niger pyrG
(encoding orotidine-59-phosphate decarboxylase and required for
uridine-uracil prototrophy) as a selection marker [41], as well as an
Ampþselection marker and the A. fumigatus pNiiA conditional promoter.
Unique restriction sites have been engineered at either side of the pyrG-
pNiiA cassette to facilitate subcloning of flanking sequences.
(B) Schematic overview of the pNiiA-CPR strategy and strain confirmation
by PCR genotyping. Promoter replacement cassettes were constructed
by inserting approximately 1.5 kb of homologous flanking sequences of
the target gene (L-arm and R-arm) into the Not1/Mlu1 and Asc1/Pac1
restriction sites, respectively. After Not1 and Pac1 double digestion, the
linearized pNiiA-CPR cassette was introduced into CEA17 strain (pyrG?)
by protoplast transformation. As indicated, three sets of primers were
used to perform genotypic PCRs to map the expected promoter
replacement junctions (L1/L2 for left-arm junction, R1/R2 for right-arm
junction) and to confirm the deletion of the native promoter (L1/P).
(C) Phenotype of transformants obtained by the pNiiA-CPR strategy for
MET2, TUB1, GFA1, and ALR1. CPR mutants were examined under pNiiA-
inducing (AMM plus nitrate) and repressing (AMM plus ammonium)
conditions after 36 h at 30 8C.
(D) Growth phenotypes of GFA1, PFS2, ALG7, CDC24, and ALR1 pNiiA-CPR
mutants under these standard repressing conditions are scored
qualitatively as the following: 4þ, essential for cell viability, no growth
under repressing conditions; 3.5þ or 3þ, showing very strong or strong
growth defect; 2þor 1þ, mild to minor growth defect; and 0þ, no growth
phenotype observed.
doi:10.1371/journal.ppat.0030024.g001
PLoS Pathogens | www.plospathogens.orgMarch 2007 | Volume 3 | Issue 3 | e240003
A. fumigatus Essential Gene

Page 4

nuclear distribution terminal phenotype characterized by
multinucleate conidia (Figure S1).
An advantage to the use of large homologous flanking
sequences to pNiiA-CPR cassettes is that multiple independ-
ent promoter replacement mutants could be recovered for a
given gene. Thus, gene essentiality can be independently
validated from multiple pNiiA-CPR mutants. Indeed, multiple
independent pNiiA-CPR mutants (n . 2) were recovered for
18 of the 19 genes mentioned above, and in all cases, their
independent pNiiA-CPR mutants displayed reproducible
growth phenotypes under repressing conditions. Although
the essential growth phenotype of GUS1 was based on the
single pNiiA-GUS1 mutant recovered, its essentiality is
consistent with previous work and with its predicted function
as a glutamate-tRNA sythetase [13].
Direct Comparison between pNiiA-CPR and Gene
Disruption Terminal Phenotypes
Reliability of the pNiiA-CPR system was also assessed bypyrG-
based gene disruption of multiple genes shown to be essential
by the pNiiA-CPR approach (Table 1). In fact, viable null
mutants of ERG11A, ERG11B, ERG27, ROM2, and MET2 (when
exogenous methionine was provided) were recovered, consis-
tentwiththeirnonessentialorconditionalessentialphenotypes
determined by pNiiA-CPR mutants. Moreover, no viable null
mutants were recovered for either TOM40 or LUC7 by gene
disruption (two essential genes identified by the pNiiA-CPR
approach) despite a significantly large number of transform-
ants examined (84 for TOM40 and 107 for LUC7, respectively)
(Table1). Thus,theessentialityofTOM40andLUC7isshown by
both the pNiiA-CPR strategy and gene deletion analysis.
Collectively, these data demonstrate that robust genetic
conclusions spanning an extensive set of ‘‘phenotypic reporter
genes’’ are reliably reproduced by the pNiiA-CPR strategy.
Quantitative Analysis of pNiiA-CPR Transcriptional
Regulation
Reverse transcription (RT)-PCR was first performed to
monitor pNiiA-CPR gene expression levels for all previously
described genes in Table 1. As pNiiA-MET2 mutants grow
normally in repressing media supplemented with methionine,
MET2 mRNA expression could be assessed in both inducing
and pNiiA-repressing conditions by RT-PCR. MET2 mRNA
was detected in both the pNiiA-MET2 mutant and wild-type
CEA10 on inducing medium; however, no MET2 transcript
was detected in the pNiiA-MET2 mutant on repressing
medium supplemented with methionine (Figure 3A). Sim-
ilarly, mRNA transcripts were detected by RT-PCR in CEA10
and pNiiA-CPR mutants for ERG11A, ERG11B, ERG27, ALR1,
and ROM2 under inducing conditions but not from pNiiA-
CPR mutants maintained under repressing conditions (Figure
S2). In addition, Northern blot analysis also confirmed a
significant depletion of ALR1 mRNA levels under repressing
Table 1. Comparison of Growth Phenotypes Observed by pNiiA-
CPR Methods versus Gene Knockout
A. fumigatus
Gene
Growth
Phenotype
of pNiiA-CPR
Mutanta
Number of Viable
Null Mutants
Recoveredb
Reproducible
Growth
Phenotype
ERG11A
ERG11B
ERG27
ROM2
LUC7
TOM40
MET2
0þ
0þ
0þ
0þ
4þ
4þ
4þ
4/23c
2/23
1/12
1/15
0/107
0/84
3/24d
Yes, viable
Yes, viable
Yes, viable
Yes, viable
Yes, essential
Yes, essential
Yes, auxotroph
aPhenotypes were examined under standard pNiiA-repressing conditions, and qualitative
shutoff phenotype is shown.
bDemonstrated viability or inferred essentiality of null phenotype is shown.
cNumber of viable null mutants identified versus the number of transformants screened
by PCR genotyping (e.g., 4/23 reflects four confirmed, viable null mutants obtained
among 23 transformants examined).
dViable deletion mutants identified by plating onto methionine-supplemented medium.
doi:10.1371/journal.ppat.0030024.t001
Figure 3. Expressional Level Analysis of pNiiA-CPR Mutants
(A) RT-PCR of pNiiA-MET2: (a) pNiiA-MET2 mutant under inducing
conditions, (b) pNiiA-MET2 mutant under repressing conditions plus
100 lg/ml methionine, (c) pNiiA-MET2 mutant under inducing condition
plus 100 lg/ml methionine, and (d) wild-type A. fumigatus strain (CEA10)
under repressing conditions. To monitor and ensure even sample
loading, RT-PCRs for the ACT1 transcript were also performed using
identical samples. In addition, standard PCR was performed to confirm
that there is no detectable genomic-DNA contamination.
(B) Northern blot analysis of pNiiA-ALR1 mutant expression levels.
Northern blot was performed with RNA samples prepared from the
nonessential ALR1 mutant and wild-type cells growing in standard
inducing (In) or repressing (Re) conditions. ACT1 transcript levels served
as sample loading controls.
(C) Real-time RT-PCR analysis of expression level of pNiiA-ALR1 and pNiiA-
MET2. pNiiA-CPR mutants and wild-type strain (CEA10) were grown in
inducing (In) or repressing (Re) medium at 37 8C for 20 h, and total RNA
was extracted from identical time points. The relative expression level
normalized to total input RNA [43] is displayed on the y-axis. Error bars
represent SD. Compared to wild-type, the relative expression level for
ALR1 and MET2 is 0.013 and 0.061 under repressing conditions and 2.49
and 23.58 under inducing conditions, respectively.
(D) pNiiA-TUB1 expression level under inducing (In) and repressing (Re)
conditions versus wild-type level is displayed on the y-axis. Error bar
represents SD. The relative expression level of pNiiA-TUB1 versus wild-
type is 0.06 under repressing conditions and 6.35 under inducing
conditions, respectively. Note: Since real-time RT-PCR was performed
using primers detecting both pNiiA-TUB1 and wild-type allele, data
shown were calculated by subtracting wild-type level from total inducing
(In) and total repressing (Re) level, respectively.
doi:10.1371/journal.ppat.0030024.g003
PLoS Pathogens | www.plospathogens.orgMarch 2007 | Volume 3 | Issue 3 | e24 0004
A. fumigatus Essential Gene

Page 5

conditions (despite the lack of a clear growth phenotype)
versus its endogenous level of expression in wild-type CEA10
(Figure 3B). Although the ALR1 mRNA levels of the ALR1
conditional mutant were clearly elevated compared to the
wild-type, no deleterious growth phenotype was detected
(Figure 1C).
Real-time RT-PCR was performed to further evaluate the
achievable range of expression levels for three pNiiA-CPR
mutants shown to yield either a nonessential (ALR1), a
conditional essential (MET2), or an essential (TUB1) terminal
phenotype (Figure 3C and 3D). Compared to wild-type, ALR1
and MET2 expression levels in CPR mutants were dramati-
cally reduced by 75.9-fold and 16.3-fold under repressing
conditions and elevated by 2.8-fold and 23.6-fold under
inducing conditions, respectively (Figure 3C). To overcome
the technical difficulty of performing real-time RT-PCR on
an essential gene, a TUB1 tandem duplication mutant was
constructed in which the pNiiA-TUB1 allele is balanced by a
wild-type TUB1 copy under the control of its native
promoter. Thus, expression of the pNiiA-regulated TUB1
allele could be monitored under repressing conditions (see
Materials and Methods). Real-time RT-PCR performed with
this strain revealed expression of the pNiiA-TUB1 allele to be
approximately 0.06-fold when repressed and 6.3-fold under
inducing conditions relative to normal TUB1 expression
levels (Figure 3D). Collectively, these data demonstrate that
the pNiiA-CPR strategy typically achieves tight regulatable
expression and suggest that it could be reliably applied to
large-scale analysis of gene essentiality in A. fumigatus.
Large-Scale Identification of Essential Genes
The genomic sequence of A. fumigatus strain AF293 [20], as
well as a second clinical isolate used in this study, CEA10,
where approximately 103 coverage has been obtained (N.
Fedorova, V. Joardar, J. Crabtree, M. Anderson, R. Maiti, et al.,
unpublished data), provides an opportunity to systematically
identify essential genes. Although global annotation of the A.
fumigatus genome is ongoing, homologs to genes of interest
were identified through search alignment tools and corre-
sponding promoter replacement cassettes designed accord-
ingly. In total, 54 A. fumigatus genes were selected for pNiiA-
CPR analysis, of which 49 are known to be essential (or
conditionally essential) for growth in S. cerevisiae with all but
three of these being essential for normal growth in C. albicans
(Table 2). In addition, genes were selected so as to represent a
broad diversity of gene functions and cellular processes.
Biasing this A. fumigatus gene set toward those essential in
both S. cerevisiae and C. albicans builds on previous work
demonstrating that at least 60% of C. albicans genes
homologous with essential S. cerevisiae genes were also
essential in C. albicans [11]. Therefore, we reasoned that a
set of genes experimentally demonstrated as essential in two
distinct hemiascomycetes would be highly enriched for genes
sharing conserved essential functions within the euascomy-
cetes.
Analysis of this A. fumigatus gene set revealed 35 genes
displaying a 3þ or greater (3.5þ or 4þ) essential growth
phenotype (Table 2). Excluding genes involved in amino acid
biosynthesis, no significant differences in terminal growth
phenotypes were observed between minimum medium
(AMM) and rich medium (Aspergillus complete medium
[ACM]). Essential genes comprise a spectrum of biological
functions including lipid, ergosterol, cell wall, amino acid,
protein, and heme biosynthesis, as well as glycosylation,
secretion, RNA processing, and novel genes of unknown
functions (Table 2). Moreover, among those genes displaying
an essential phenotype in both S. cerevisiae and C. albicans and
predicted to have only a single ortholog in A. fumigatus (n ¼
44), 32 genes (73%) are experimentally demonstrated as
sharing a conserved growth phenotype of 3þ or greater in A.
fumigatus.
Essential genes identified in this manner constitute a
collection of possible antifungal drug targets, and micro-
scopic examination of their terminal growth phenotypes may
assist in their prioritization (Figure 4). For example,
repression of GFA1 produced enlarged and highly disrupted
nongerminating conidia that failed to undergo any polarized
growth. As GFA1 is predicted to encode the sole glutamine-
fructose-6-phosphate aminotransferase activity responsible
for the first and rate-limiting step in chitin synthesis, this
terminal phenotype likely reflects the normal interdepend-
ence between cell wall biosynthesis and polarized growth.
Conidia of SEC31 and SLY1 incubated under repressing
conditions were unable to germinate, suggesting that these
genes are required for the initiation of germ tubes, whereas
terminal phenotypes of TUB1, ERG10, and, most promi-
nently, PRI1, resulted in growth-arrested germ tubes. Inter-
esting, HEM15 displayed a more complex terminal phenotype
of dramatically swollen conidia with single or multiple germ
tubes. Repression of pNiiA-FKS1 yielded stubby and highly
branched micromycelial germlings that morphologically
resemble the chemotype observed when A. fumigatus is treated
with Fks1p-specific inhibitors such as caspofungin and
related echinocandins [21]. Thus, distinct terminal pheno-
types are identified among pNiiA-CPR mutants, and in some
instances, they appear more severe than that caused by either
genetic or chemical inactivation of FKS1.
Essential genes with a cidal terminal phenotype are
commonly preferred since genetic evidence predicts that
chemical inhibitors of such targets may display similar
chemotypes. An assay to evaluate cidal or static terminal
phenotypes was applied to essential genes displaying a 3.5þor
4þ growth phenotype (Table 2) by incubating conidia in
pNiiA-repressing medium for various durations of time
before washing and plating conidia onto pNiiA-inducing
medium where viability was scored by counting colony-
forming units (CFU) (see Materials and Methods and Figure
5). Of 28 genes tested, six genes (approximately 21%),
including GFA1 (Figure 5), displayed a cidal terminal
phenotype, as scored by a significant reduction (greater than
90%) of CFU after 24-h incubation in repressing medium.
Additional genes demonstrating a cidal terminal phenotype
include KRR1, SEC31, SLY1, TUB1, and GUS1 (Table 2).
Additional genes displayed a slow cidal phenotype, as scored
by significant reduction (greater than 90%) in CFU after 48-
or 72-h incubation in repressing medium, or a static terminal
phenotype after 48-h incubation in repressing conditions as
shown with TRR1 (Figure 5, Table 2).
The ERG11 Gene Family Is Essential in A. fumigatus
Identifying gene families that are essential for the growth
and viability of A. fumigatus can be missed by our approach
and thus require additional considerations. For example,
duplicate ERG11 genes (encoding 14a-demethylase, the
PLoS Pathogens | www.plospathogens.orgMarch 2007 | Volume 3 | Issue 3 | e240005
A. fumigatus Essential Gene