Content uploaded by Raymond St leger
Author content
All content in this area was uploaded by Raymond St leger
Content may be subject to copyright.
Variation in gene expression patterns as the
insect pathogen Metarhizium anisopliae adapts
to different host cuticles or nutrient deprivation
in vitro
Florian M. Freimoser,3 Gang Hu and Raymond J. St Leger
Correspondence
Raymond J. St Leger
rl106@umail.umd.edu
Department of Entomology, University of Maryland, 4112 Plant Sciences Building, College
Park, MD 20742, USA
Received 16 August 2004
Revised 25 October 2004
Accepted 28 October 2004
Metarhizium anisopliae infects a broad range of insects by direct penetration of the host cuticle.
To explore the molecular basis of this process, its gene expression responses to diverse insect
cuticles were surveyed, using cDNA microarrays constructed from an expressed sequence tag
(EST) clone collection of 837 genes. During growth in culture containing caterpillar cuticle
(Manduca sexta), M. anisopliae upregulated 273 genes, representing a broad spectrum of
biological functions, including cuticle-degradation (e.g. proteases), amino acid/peptide transport
and transcription regulation. There were also many genes of unknown function. The 287
down-regulated genes were also distinctive, and included a large set of ribosomal protein genes.
The response to nutrient deprivation partially overlapped with the response to Man. sexta
cuticle, but unique expression patterns in response to cuticles from another caterpillar
(Lymantria dispar), a cockroach (Blaberus giganteus) and a beetle (Popilla japonica) indicate
that the pathogen can respond in a precise and specialized way to specific conditions. The
subtilisins provided an example of a large gene family in which differences in regulation could
potentially allow virulence determinants to target different hosts and stages of infection.
Comparisons between M. anisopliae and published data on Trichoderma reesei and
Saccharomyces cerevisiae identified differences in the regulation of glycolysis-related genes
and citric acid cycle/oxidative phosphorylation functions. In particular, M. anisopliae has multiple
forms of several catabolic enzymes that are differentially regulated in response to sugar levels.
These may increase the flexibility of M. anisopliae as it responds to nutritional changes in its
environment.
INTRODUCTION
Current molecular and genomic methods are being applied
to Metarhizium anisopliae, the causative agent of green
muscardine disease, because of its importance for the bio-
logical control of insect pests. It is a very versatile fungus,
being able to infect a broad range of insects, 200 species
from over 50 insect families (Samuels et al., 1989), and is
also adapted to life in the root rhizosphere (Hu & St Leger,
2002). Consi stent with its promiscuous nature, an array of
expressed sequence tags (ESTs) from M. anisopliae strain
2575 identified large numbers of genes dedicated to host
interaction and countering insect defences, as well as regu-
lators for coordinating their implementation (Freimoser
et al., 2003). Sequence comparisons and conserved motifs
suggest that about 60 % of the ESTs of strain 2575 expressed
during growth on cuticle encode secreted enzymes and
toxins. Acting collectively, the number and diversity of these
effectors may be the key to this pathogen’s ability to infect a
wide variety of insects. In contrast, ESTs from the specialized
locust pathogen M. anisopliae sf. acridum strain 324 revealed
very few toxins (Freimoser et al., 2003). This relates to life-
styles. Strain 2575 kills hosts quickly via toxins, and grows
saprophytically in the cadaver. In contrast, strain 324 causes
a systemic infection of host tissues before the host dies. This
shows that by utilizing ESTs, multiple virulence factors and
pathways can be viewed simultaneously, and the different
lifestyles that exist in insect–fungus interactions can be
understood from a broader perspective.
3Present address: Institute of Plant Sciences, ETH Zurich,
Universita¨tsstr. 2, CH-8092 Zurich, Switzerland.
The expression ratios for Metarhizium anisopliae ESTs in different
cuticle-containing media are shown in Supplementary Table S1 with
the online version of this paper at http://mic.sgmjournals.org.
Abbreviations: BC, beetle (Popilla japonica) cuticle; CC, cockroach
(Blaberus giganteus) cuticle; ESTs, expressed sequence tags; GC, gypsy
moth (Lymantria dispar) cuticle; HL, haemolymph; MC, Manduca sexta
cuticle; RT-PCR, reverse-transcription PCR.
0002-7560
G
2005 SGM Printed in Great Britain 361
Microbiology (2005), 151, 361–371 DOI 10.1099/mic.0.27560-0
In this report, we use cDNA microarrays for high-throughput
expression profiling of how M. anisopliae strain 2575
responds over a 24 h period to cuticle from tobacco horn-
worm caterpillars (Manduca sexta). As a control, we also
define the response of M. anisopliae to nutrient deprivation.
In addition we obtained snapshots of gene expression at
24 h to compare and contrast the responses of M. anisopliae
to gypsy moth caterpillar cuticle (Lymantria dispar), and
hard (sclerotized) cuticles from a beetle (Popilla japonica)
and a cockroach (Blaberus giganteus). Each of these insects is
a susceptible host for M. anisopliae.
These studies demonstrated that M . anisopliae can rapidly
adjust its genomic expression patterns to adapt to insect
cuticle, and identified specific responses to different cuticles.
Genes specifically induced by cuticle incl uded a plethora of
cuticle-degrading enzymes, transporters for cuticle degra-
dation products and a subset of transcription factors.
METHODS
Strains and culture conditions. To measure the variation in the
expression of genes during starvation conditions or during adapta-
tion to growth on different insect cuticles, we transferred cultures to
minimal medium (MM) or cuticle-containing media after a period
of unrestricted growth on a nutrient-rich medium. This is an effec-
tive and reproducible procedure for obtaining proteins that require
release from catabolite repression and/or specific induction by a
cuticular component (St Leger et al., 1994). M. anisopliae sf. aniso-
pliae (strain ARSEF 2575) was routinely grown at 27 uC, either in
liquid (SDB) or on solid (SDA) Sabouraud dextrose medium sup-
plemented with 0? 5 % yeast extract. For RNA extraction, the fungus
was grown for 48 h in 50 ml liquid SDB broth. The cultures were
then washed with sterile distilled water and 2 g wet weight of the
fungal biomass was transferred for up to 24 h to 10 ml MM con-
taining 0?1% KH
2
PO
4
,0?05 % MgSO
4
and 50 % tap water, supple-
mented with 1 % of the following additives: tobacco hornworm
caterpillar (Man. sexta cuticle, MC); cockroach cuticle (CC, B.
giganteus); beetle cuticle (BC, P. japonica); gypsy moth cuticle (GC,
L. dispar ). Cuticles were prepared as described previously (St Leger
et al., 1986b). Alternatively, M. anisopliae sf. anisopliae was trans-
ferred to 10 ml of Man. sexta haemolymph (HL) obtained and trea-
ted as described by Grundschober et al. (1998).
cDNA microarray experiments. All unique ESTs with significant
BLAST matches (Freimoser et al., 2003) were amplified using T3
and T7 primers and standard PCR protocols. It should be noted
that, as with most other bioinformatic studies, gene identities are
based on computer-predicted homologies, and in very few cases (e.g.
serine proteases and hydrophobins) have the protein products of
these genes been demonstrated experimentally. Genes found among
the EST sequences of M. anisopliae sf. acridum (ARSEF 324), such
as chitinases and chitosanase (Freimoser et al., 2003), which were
absent from the M. anisopliae sf. anisopliae (ARSEF 2575) EST col-
lection, were amplified from M. anisopliae sf. anisopliae genomic
DNA with specific primers and included on the array. This
resulted in 837 clones, which were precipitated and resuspended
in 36 SSC (16 SSC=0?15 M sodium chloride, 0?015 M sodium
citrate, pH 7?0) to give a final DNA concentration between 100 and
300 ng ml
21
.
Printing, hybridization and scanning of slides were performed with
an Affymetrix 417 Arrayer and 418 Scanner (see http://www.
umbi.umd.edu/~cab/macore/macorestart.htm for detailed protocols)
at the University of Maryland Biotechnology Institute’s Microarray
Core Facility located at the Center for Biosystems Research. PCR
products were spotted in triplicate on poly-lysine-coated glass slides,
with a mean spot diameter of 100
mm and a spot spacing of 375 mm.
Following printing and cross-linking, slides were washed with 1 %
SDS to remove background, treated with blocking solution (0?2M
succinic anhydride, 0?05 M sodium borate, prepared in 1-methyl-
2-pyrrolidinone) and washed with 95 uC water and 95 % ethanol. After
drying, slides were kept in the dark at room temperature.
RNA was extracted as previously described for M. anisopliae (Joshi &
St Leger, 1999). For experiments comparing different media, RNA
from a culture transferred to SDB was used as the reference sample.
For time-course experiments, mycelium was collected after 1, 2, 4, 8,
12, 18 and 24 h from MM or from medium containing MC. RNA
from the 0 h time-point was used as a reference. Hybridizations
were done with Cy3- and Cy5-labelled probes derived from 50–80
mg
of total RNA. All hybridizations were repeated at least three times
with RNA from independent experiments and with switched labelling
for the reference and test RNA samples.
Analysis of microarray data. The images of the scanned slides
were analysed with Scanalyse (available from Eisen Lab: http://rana.
lbl.gov/) and the data obtained from each scanned slide were nor-
malized using global normalization, as performed by J-Express
(Dysvik & Jonassen, 2001). All data were log
2
-transformed, and for
further analysis the mean (Em) and the standard deviation (
SD)of
the log-transformed expression ratios of the replicates were calcu-
lated for all genes. A gene was defined as differently regulated if the
expression varied by at least a factor of two (1<Em<21). Expres-
sion ratios not fulfilling this requirement (21<Em<1) were defined
as zero, and the same was done for cases where the interval
Em±1?966
SD (95 % confidence interval around the mean value for
the three replicate spots) included the value 0. Further analysis of the
processed data was performed using J-Express (Dysvik & Jonassen,
2001),
EPCLUST (http://ep.ebi.ac.uk/EP/EPCLUST/) and Excel.
Validation of differentially expressed clones through real-
time PCR. A total of 16 clones predicted to be differentially
expressed by microarray analysis were tested by quantitative reverse-
transcription PCR (RT-PCR) by using an Applied Biosystems
GeneAMP 5700 sequence detection system and an Applied
Biosystems TaqMan RT kit. Transcript abundance was calculated
by using the comparative
DCt method relative to the amount of
the tubulin alpha chain transcript AJ273998 or 18S rRNA in the
sample, with primers and conditions as described by Parsley et al.
(2002). Differential expression based on RT-PCR measurements
was defined as a change in transcript abundance accumulation of
twofold or more.
RESULTS
Overall patterns of cuticle-induced gene
expression
The libraries we employed to obtain ESTs were made
from fungi growin g on Man. sexta cuticle (MC) for 24 h
(Freimoser et al., 2003). The complete list of ESTs class-
ified into functional groups is available at http://
mic.sgmjournals.org/cgi/content/full/149/1/239/DC1. These
ESTs were hybridized with labelled RNA probes isolated
from mycelium harvested up to 24 h after the transfer from
a nutrient-rich medium (SDB) to media containing an
insect cuticle or HL. As a control for time-cou rse studies
with MC, mycelium was challenged by transfer to MM,
362 Microbiology 151
F. M. Freimoser, G. Hu and R. J. St Leger
revealing the response to nutrient deprivation after growth
in SDB. The responses to MC and MM were studied in
parallel time-course experiments, each with seven time-
points (1–24 h) that, together with the redundant sequence
representation in the microarrays, ensured the robustness of
the expression profiles. The expression ratios for ESTs in the
different media are shown in Supplementary Table S1 with
the online version of this paper at http://mic.sgmjournals.
org. In addition, the ESTs are characterized in Table S1 with
their accession numbers, E values and a description of the
best
BLAST hit.
An overview of the microarray results is presented in Fig. 1.
They illustrate the rapid changes in expression of some genes
in response to MC. Overall, these changes increased in
magnitude with time. Thus at 4 and 18 h post-inoculation
in MC medium, 88 and 154 genes were upregulated, res-
pectively. Similarly, 66 genes were down-regulated at 8 h,
and 143 genes were down-regulated at 18 h. During the
first hour, there was no overlap between genes upregulated
in response to MC and those upregulated in resp onse to
starvation conditions (MM). However, by 18 and 24 h, 30 %
of the genes were concomitantly upregulated in MC and
MM, indicating that catabolite repression is involved in
regulating at least some cuticle-induced genes. A cluster of
41 gen es was rapidly activated (<2 h) by MM, but down-
regulated in response to cuticle (Fig. 1). Only eight of
these genes had homologues with known biological activity
in databases, and these included the subtilisin Pr1G and
ribosomal proteins. At least with respect to the regulation of
these 41 genes, nutrient deprivation may be perceived as
having an effect distinct from and even opposite to that of
induction by cuticle.
In contrast, the magnitude of expression of most genes
upregulated by nutrient deprivation, including the majority
of the secreted proteases, was sharply increased by the pre-
sence of cuticle (Supplementary Table S1). In addition, a
large subset of diverse genes was upregulated by MC, and
not by MM, during the first 2 h, suggesting that these genes
are specifically involved in adaptation to growth on cuticle
(Fig. 2). A broad view of the nature of the adaptations
made by M. anisopliae following transfer from nutrient-rich
(SDB) medium to MC was obtain ed by grouping func-
tionally related genes (Fig. 3). Changes involving upregula-
tion, measured on the microarray for each functional
category during the 24 h of growth on MC, were either
gradual following the first hour, as for secreted proteases,
or abrupt during the first hour followed by a slow decline, as
for genes for amino acid/peptide uptake. Down-regulated
genes included many for protein synthesis machinery,
excluding RNA synthesis and processing. Genes encoding
ribosomal proteins and translational machinery were coor-
dinately regulated, showing an initial decrease, followed
by an increase and a decrease, resulting in an 8- to 13-fold
down-regulation. The repression of ribosomal protein
genes has been reported in yeasts during multiple stress
responses, including glucose deprivation (Warner, 1999),
and may therefore be a general feature of fungi transferred
to a low-nutrient medium. Overall, ho usekeeping genes for
cell metabolism, including endocellular proteases, showed
stable expression.
Fig. 1. Gene expression patterns of M. anisopliae in response
to starvation conditions (MM), haemolymph (HL) from Man.
sexta, or cuticles from a beetle (BC), a cockroach (CC) and
caterpillars (Man. sexta, MC, and Lymantria dispar, GC).
Mycelia growing on MC or in starvation conditions were assayed
in time-course experiments. The 837 cDNA clone set was ana-
lysed by hierarchical clustering based on their expression pat-
terns. Genes showing at least twofold regulation, compared
with a reference probe from mycelia grown on SDB, are shown
in red (upregulated) and green (down-regulated). Colour inten-
sity is directly relative to magnitude of differential expression
ratios. Experiments were carried out in triplicate, and represen-
tative clusters are shown.
http://mic.sgmjournals.org 363
Microarray analysis of Metarhizium anisopliae
Based on the time-course experiments under starvation
conditions or in MM supplemented with MC, we chose the
24 h time-point to obtain a snapshot of gene expression
during growth on other insect cuticles. The large-scale
features of the expression patterns illustrate shared features
between the responses to different cuticles and to starva-
tion conditions, indicative of a stereotyped programme
of gene expression. However, no two expression patterns
were identical in terms of the genes affected and the
magnitude of expression alteration (Fig. 1). Of the 136
genes upregulated on MC at 24 h, 87 (64 %) were similarly
regulated on CC, 96 (71 %) were similarly regulated on
BC and 95 (70 %) were similarly regulated on GC. Among
these commonly regulated genes, 64 were upregulated on
all four cuticles at 24 h. The balance of genes demonstrated
specific responses to different cuticles, including up- or
down-regulation of genes not observed at any time-point
on MC (Fig. 4). This implies that the pathogen can pre-
cisely respond to different conditions. In some cases, genes
coordinately upregulated on a particular cuticle were func-
tionally related. Thus, several sequences upregulated at
24 h on GC, but not MC, CC or BC at this time-point,
have homologues in yeast that are involved in integrating
nutrient and growth signals with morphogenesis. These
include LAS1, a nuclear protein required for cell-surface
growth and bud formation (Doseff & Arndt, 1995); SLA2,
required for morphogenesis and polarization of the
membrane cytoskeleton (Holtzman et al., 1993); Ecm15p,
involved in yeast cell-wall biogenesis (Goffeau et al., 1996);
and PIG-L, essential in the synthes is of glycosylpho-
sphatidylinositol, used as a membrane anchor by cell-
surface proteins (Watanabe et al., 1999).
Fig. 2. The cluster of M. anisopliae genes upregulated within 2 h in medium containing Man. sexta cuticle, and which were
not upregulated in minimal medium (starvation conditions). Samples were reordered from Fig. 1 according to the time-scale
shown across the top, and genes were hierarchically clustered. Gene names and accession numbers are shown to the right of
the figure.
364 Microbiology 151
F. M. Freimoser, G. Hu and R. J. St Leger
Identification of genes regulated by nutrient
deprivation and by insect cuticles
Energy metabolism. While the overall expression of sev-
eral functional categories, including cell metabolism, was
largely unaltered by transfe r to cuticle (Fig. 3), individual
ESTs among categories were altered in regulation, possibly
indicative of pivotal enzymes involved in metabo lic repro-
gramming (Supplementary Table S1). Although pathways
are incomplete in the M. anisopliae array, infere nces can
be made from the differential expression of representative
genes, as this likely reflects alterations in the pathways
in which these genes are involved. We thus compared our
results with experiments performed with Saccharomyces
cerevisiae (DeRisi et al., 1997) and Trichoderma reesei
(Chambergo et al., 2002), using very similar nutrient-rich
and nutrient-poor media. In all three species, the regula-
tion of many genes that participate in key metabolic pro-
cesses is not affected by being in sugar-rich media, such
as SDB (Supplementary Table S1). However, in T. reesei,
expression of genes encoding tricarboxylic acid (TCA)
cycle components and mitochondrial proteins favours
the oxidation of pyruvate via the TCA cycle, rather than
its reduction to ethanol by fermentation. In contrast,
S. cerevisiae preferentially ferments glucose, even in the
presence of oxygen. Only when glucose is exhausted do
yeast cells use the ethanol as a carbon and energy source
for aerobic respiration (the ‘diauxic shift’). M. anisopliae
resembled T. reesei in that the abundance of transcripts
encoding enzymes of the glycolytic pathway and TCA
cycle (e.g. isocitrate dehydrogenase, AJ272972) was mostly
unaffected upon transfer from a sugar-rich (SDB) to a
sugar-deficient medium (MM). In yeast, these genes are
strongly repressed in sugar-rich media. Yeast mitochon-
drial genes are also subject to strong repression by glu-
cose. However, levels of M. anisopliae transcripts encoded
by the mitochondrial genome (e.g. NADH ubiquinone
dehydrogenase, AJ273010) and nuclear genes encoding
mitochondrial proteins (e.g. cytochrome c oxidase chain
V, AJ272726) were the same or higher in sugar-rich media
than in MM. These results indicate that, like T. reesei, but
unlike yeast, M. anisopliae will respire in the presence of
sugar.
However, M. anisopliae appears to differ from T. reesei in
the extent to which aerobic respiration prevails. As in
yeast and T. reesei,aM. anisopliae pyruvate decarboxy-
lase (AJ274332) is upregulated in the presence of sugar.
Fig. 3. Regulation of functionally related
genes. The curves represent the average
induction or repression ratios for all the
genes in each indicated group. The total
number of genes in each group was as fol-
lows: cell metabolism, 71; cofactors/vitamins,
6; energy metabolism, 27; ribosomal pro-
teins, 25; translation, 15; tRNA synthesis, 4;
secreted protease, 23; intracellular pro-
teases, 12; transport proteins, 14; amino
acid/peptide transporters, 6; cell wall struc-
ture/formation, 26; stress response, 26;
RNA metabolism, 28.
http://mic.sgmjournals.org 365
Microarray analysis of Metarhizium anisopliae
However, in contrast to these fungi, M. anisopliae has
an additional pyruvate decarboxylase (AJ274298) that is
repressed in nutrient-rich medium but upregulated within
1 h on MC (Fig. 2). In S. cerevisiae, the acetaldehyde
formed from pyruvate decarboxylase is reduced to ethanol
by alcohol dehydrogenase, and is not converted to acetate,
due to repression of aldehyde dehydrogenase by glucose.
Two paralogous genes for aldehyde dehydrogenase hav e
been identified in T. reesei, only one of which is repressed
by nutrient-rich conditions. In contrast , both the aldehyde
dehydrogenases (AJ272833 and AJ273869) in M. anisopliae
are down-regulated in SDB compared to cuticle-containing
media, suggesting that readily utilized nutrients repress
acetate production. It is of interest that AJ272833 is
upregulated at an earlier time-point on MC than in
MM (Fig. 2). We also identified two paralo gues of
acetyl coenzyme A synthetase: the AJ273955 transcript is
upregulated early during growth on cuticle and late
during growth in MM (Fig. 2), while regulation of
AJ274191 is not affected. If both enzymes have compara-
ble specificity, production of acetyl coenzyme A in glucose-
poor media such as cuticle will increase the entry of
acetate, produced via the pyruvate bypass route, into the
TCA cycle. Interestingly, M. anisopliae also has two
paralogous genes for alcohol dehydrogenase. AJ273792 is
regulated in a similar fashion to pyruvate decarboxylase
AJ274332 (upregulated in SDB), whereas AJ273547, like
pyruvate decarboxylase AJ274298, is repressed in SDB.
Thus, M. anisopliae has multiple gene families of cata-
bolic enzymes, some of which include isoforms that are
differentially regulated by sugars. These alternative forms
may give M. anisopliae the flexibility to shunt any avail-
able pyruvate into fermentation or the TCA cycle,
irrespective of sugar levels.
(a)
(b)
(c)
Fig. 4. Subclusters of genes specifically upregulated on only one of the cuticles (BC, beetle; CC, cockroach; MC, Man.
sexta; GC, Lymantria dispar). Samples were reordered from Fig. 1 according to the time-scale shown across the top, and
genes were hierarchically clustered. Gene names and accession numbers are shown to the right of the figure.
366 Microbiology 151
F. M. Freimoser, G. Hu and R. J. St Leger
Amino acid, carbohydrate and lipid metabolism. Genes
with homologues involved in amino acid catabolism and
which were upregulated on cuticle included glutami nase
A (AJ273512) and NADH-specific glutamate dehydrogen-
ase (AJ274362). Glutamate is the preferred amino acid
substrate for M. anisopliae (St Leger et al., 1986a).
Otherwise, diverse genes involved in amino acid synthesis
were commonly down-regulated in MM and on cuticle,
consistent with the reduced availability of raw materials
for biosynthesis. Insect cuticle also contains diverse lipids,
and seven of 13 genes for lipid metabolism were upre gu-
lated on at least one cuticle. Only a cytochrome P450
monooxygenase (AJ274003) wa s also upregulated during
growth in MM. Lipases are the last class of depolymerases
to be secreted in insect cuticle (St Leger et al., 1986 b),
consistent with which, lipase AJ274124 was upregulated
in late cuticle-containing cultures (24 h) only. Enzyme
assays have also detected a secreted DNase activity during
growth on cuticle (St Leger et al., 1986b), and in this
study DNase (AJ273950) was upregulated in cuticle-
containing media.
Protein a side, the major component of insect cuticle is
chitin, and predictably therefore chitinases were upregu-
lated on cuticle. Chitinase AJ274366 was exp ressed
within 1 h on MC, but was not expressed in MM (Fig. 2).
Chitosanase was only produced on GC (Fig. 4). As this
coincides with the GC-specific expression of genes involved
in morphogenesis, it is possible that the chitosanase may
be involved in modifying cell wall components. However,
five additional enzymes involved in metabolizing carbo-
hydrates not known to occur in cuticle were also upregu-
lated in one or more of the cuticle media: formate
dehydrogenase (AJ274347), usually involved in detoxifica-
tion reactions; 1,2-
a-D-mannosidase (AJ273630); b-D-
galactosidase (AJ273808);
L-sorbosone dehydrogenase
(AJ273834); and b-glucosidase (AJ273623). These could
be involved in digesting glycoproteins, but were also
more weakly upregulated in starvation conditions, consis-
tent with catabolite repression in SDB . Only one of
the genes for carbohydrate metabolism (AJ272928)
was upregulated in response to HL, while seven genes
were down-regulated (Supplementary Table S1). Seven
carbohydrate-metabolizing enzymes were down-regulated
on cuticle-containing media, including a transketolase
(AJ274194) and fructose-bisphosphate aldolase (AJ273952).
RNA synthesis. Elements required for mRNA synthesis,
such as RNA polymerase (AJ272996) and RNA polymer-
ase transcription factor (AJ274125) were upregulated in
cuticle-containing media, but not in MM or HL. This
presumably adapts the fungus for the rapid synthesis of
cuticle-degrading enzymes.
Transport proteins. The ESTs included homologues
of two distinct peptide transport systems, one for di-/
tripeptides (PTR transporter AJ273551 and PTR-2 trans-
porter AJ272830) and another for tetra-pentapeptides
(OPT transporter AJ273568), as well as diverse amino
acid transporters (e.g. the INDA1 homologue AJ272773).
These all required induction by cuticle, and most were
upregulated 8- to 12-fold within 1 h on MC (Figs 2 and
3). In contrast, the PTR transporter in T. reesei is upregu-
lated by glucose exhaustion alone (Chambergo et al., 2002),
consistent with the M. anisopliae transporters having
acquired more specialized functions in pathogenicity.
Only the M. anisopliae oligopeptide transporter OPT2
(AJ273118) was not upregulated in cuticle-containing
media. Regulation of peptide/amino acid transporters was
not altered in HL compared to growth on SDB.
Proteolytic enzymes. It had been shown previ ously
that total subtilisin activity is produced in response to
nutrient deprivation, but that production is enhanced by
the addition of cuticle to media (Paterson et al., 1994).
Consistent with this, subtilisins Pr1A and Pr1B were upre-
gulated on MM, and to a greater extent on insect cuticles
(Fig. 5). Increased induction by cuticle compared with
nutrient deprivation alone suggests that subtilisin produc-
tion is controlled by multiple regulatory systems evoked
under different environmental conditions. In contrast,
Pr1C, Pr1D, Pr1E, Pr1F, Pr1I and Pr1J were down-
regulated at most time-p oints in MM. Of these, Pr1C and
Pr1D were rapidly upregulated (Pr1C within 1 h of trans-
fer to MC; Fig. 2), while upregulation of Pr1E and Pr1K
in MC was delayed by 4 and 8 h, respectively. Pr1J was
upregulated on all the cuticles, except BC. Pr1G was shar-
ply down-regulated in CC. Pr1F and Pr1I were upregu-
lated on MC and on GC. Expression of Pr1H was slightly
upregulated by transfer to MC and MM.
The exo-acting carboxypeptidase AJ274343 was upregulated
after 18 h in MM, but showed earlier and much stronger
upregulation in all cuticle-containing media. Most other
categories of exopeptidases (e.g. aminopeptidases AJ273806
and AJ274061) and endopeptidases, including trypsin
(AJ272743), chymotrypsin (AJ273663), metalloprotease
(AJ273481) and aspartyl protease (pepsinogen)
(AJ274168) were only upregulated in the presence of cuticle.
Transcription factors and signal transduction. Of the
17 arrayed ESTs enc oding homologues of proteins known
to be involved in transcription in other organisms, ten
(AJ272823, AJ272967, AJ273078, AJ273134, AJ273171,
AJ273219, AJ273260, AJ273589, AJ273694 and AJ274235)
were upregulated on at least one cuticle. The positive
sulfur transcription regulator homologue (AJ273134)
was down-regulated in MM and BC, suggestive of parti-
cularly low sulfur levels in these media (sulfite reductase,
AJ273620, but not sulfite oxidase, AJ272866, was upregu-
lated on cuticle, within 1 h in MC, but not in MM and
HL). In contrast, the pH signalling transcription factor
PacC (AJ273219) was upregulated on cuticle, but not in
MM or HL . AJ272977 was unique in being upregulated in
HL. In contrast, AJ273694 was very strongly down-regulated
in HL and strongly upregulated on the lepidopteran
http://mic.sgmjournals.org 367
Microarray analysis of Metarhizium anisopliae
cuticles GC and MC. Among gene products involved
in signalling (category 4f), adenylate cyclase (AJ251971)
(the enzyme that produces cAMP) and protein kinase A
(AF116597) (PKA: the major effector of cAMP responses)
were not upregulated on cuticle-containing media, while a
downstream activity, MAP kinase kinase 2 (AJ273356)
was upregulated in GC- and BC-containing media, and at
two time-points in MC.
Cell wall proteins. Of 30 genes encoding proteins
involved in cell structure and function, 18 were upregu-
lated in at least one cuticle-containing medium. The
hydrophobins are differenti ally regulated. Thus, AJ273847
was upregulated in HL and MM, and down-regulated in
cuticle-containing media, while AJ274156 was upregulated
in MM and on sclerotized cuticles (CC and BC), un-
altered on lepidopteran cuticles (GC and MC) and down-
regulated in HL. The other cell wall proteins upregulated
on cuticle were AJ273845, a homologue to an antigenic
cell wall protein from the human pathogen Aspergillus
fumigatus, and AJ274019, which is very similar to the
antifungal glucan 1,3-
b-glucosidase fro m Trichoderma
atroviride (Donzelli et al., 2001). Clearly, besides cell wall
biosynthesis and structure, these proteins may have addi-
tional functions in pathogenicity or in protecting scarce
resources from competitors.
Stress response. Several arrayed M. anisopliae ESTs
are similar to peptide synthases, reductases and other
enzymes that take part in the synthesis of fungal toxins,
such as destruxins, trichothecene and enniatin (Freimoser
et al., 2003). This is in agreement with the observation
that M. anisopliae strain 2575 rapi dly kills its host after
infection through the action of toxins, and subsequently
colonizes the insect host by saprob ic growth (Samuels
et al., 1989). Genes upregulated in at least one cuticle-
containing medium included those encoding a peptide
synthase (AJ272930, in BC and GC), a protein involved
in sterigmatocystin biosynthesis (AJ273515, in GC), versi-
colorin B synthase (AJ272697, in CC, GC, MC and HL)
and a bacteriolytic enzyme (AJ272917, in CC, BC, GC,
early in MC and after 12 h in MM).
Validation of microarray results. An external quality
control check on the lists of differentially expressed clones
generated through microarray profiling was provided by
analysing a subset of 16 clones by quantitative RT-PCR.
Each clone tested by RT-PCR was measured in triplicate
for each of two independent RNA isolations. All 16 clones
were confirmed by RT-P CR, indicating a very high success
rate for predicting differentially expressed clones. How-
ever, expression ratios were consistently underestimated at
least 10-fold by cDNA microarrays, compared to PCR-
based methods.
DISCUSSION
The construction of a M. anisopliae cDNA microarray
provides many advantages over previous labour-intensive
techniques to monitor transcriptional responses to host
tissues. For this study, it provided a powerful tool with
which to examine the influence of culture conditions on the
magnitude and spectrum of cuticle-induced gene expres-
sion. The analysis presented here has expanded the number
of identified M. anisopliae genes that respond to cuticle
from about 20 (Joshi & St Leger, 1999; Joshi et al., 1997) to
more than 200. Expression patterns of known pathogeni-
city genes, incl uding Pr1A subtilisin, hydrophobin, trypsins,
chymotrypsin and carboxypeptidase, matched previously
published data (Screen & St Leger, 2000; St Leger et al.,
1986b, 1987, 1992, 1996). This provides a high level of
confidence that the arrays accurately identify differentially
expressed clones. For selec ted genes, including subtilisins
(Pr1A, Pr1H and Pr1K), a trypsin (Try1) and tubulin, the
Fig. 5. Subtilisin gene expression during growth on cuticle from Man. sexta compared with expression during nutrient
deprivation. The curves show average expression ratios for different subtilisins in minimal medium (MM) and on Man. sexta
cuticle (MC): &, Pr1A; %, Pr1B; #, Pr1I;
N
, Pr1E; +, Pr1J; n, Pr1D; m, Pr1C; x, Pr1K.
368 Microbiology 151
F. M. Freimoser, G. Hu and R. J. St Leger
expression patterns were also verified by quantitative real-
time RT-PCR. Normally (e.g. Yuen et al., 2002), expression
ratios are greatly underestimated by cDNA microarrays,
compared to PCR-based methods. Together with our strict
analysis of the replicates within and between different arrays,
this suggests that our estimates of the magnitude of changes
in expression are conservative.
The demonstration of the differential regulation of genes
encoding cuticle-degrading enzymes, cell w all proteins,
toxins and toxin-producing enzymes on the different
cuticles, and in HL and MM suggests that M. anisopliae
may have the ability to target the production of these
proteins to different hosts. Like other ascomycete patho-
gens, M. anisopliae secretes a great variety of proteases (Hu
and St Leger, 2004), some of which have been associated
with virulence, because they allow rapid physical ingress,
nutrient solubilization and the disabling of antimicrobial
peptides (St Leger et al., 1996). The subtilisin cluster pro-
vides a good example of de novo protein synthesis req uired
for adaptation to growth on cuticle (Figs 1 and 5, Supple-
mentary Table S1), particularly as the differences in regula-
tion of subtilisins imply differences in their function. This
supports homology-modelling studies based on seque nces
that predict differences between the Pr1s in their secondary
specificities, adsorption properties to cuticle and alkaline
stability (Bagga et al., 2004). It is likely that these differences
in regulation and structure–function allow M. anisopliae to
respond flexibly, producing proteases that are appropriate
to the composition of the environment, consistent with its
opportunistic lifestyle. Thus the proteases such as Pr1A
produced as part of a general response to nutrient depriva-
tion could also function outside of pathogenesis by scaveng-
ing for nutrients during saprophytic existence. During early
infection processes, they could also function in concert with
the exopeptidases to provide host degradation products.
These may include specialized signals that allow the fungus
to ‘sample’ the cuticle and then respond with the secretion of
a plethora of cuticle-induced proteins. This will include the
proteases that require cuticle for induction, as they
presumably have specialized roles in breaching host barriers.
The very early induction of peptide/amino acid transport
systems (Figs 2 and 3) would enhance the ability of the
fungus to rapidly and precise ly monitor host degradation
products.
Hydrophobins provide another example besides subtilisins
where members of a family are differentially regulated,
consistent w ith different functions. Thus, AJ273847 was
upregulated in HL, while AJ274156 was down-regulated
in HL. This suggests that adaptation to HL may include
alterations in cell wall composition.
Of key importance to understanding the mechanisms
behind adaptation to cuticles is the identification of com-
ponents of signal transduction that will allow M. anisopliae
to screen its surroundings to regulate protein synthesis and
secretion. PacC-mediated pH signalling is crucial to the
pathogenicity of the human pathogen Candida albicans and
the plant pathogen Fusarium oxysporum (Caracuel et al.,
2003; Davis et al., 2000). Consistent with a crucial role for
PacC in M. anisopliae, extrac ellular pH rises during cuticle
degradation and acts as a key signal for the production of
alkaline-active enzymes such as subtilisins (St Leger et al.,
1998). Significant for their absence of response were
adenylate cyclase and protein kinase A, as transcriptional
regulation in response to the cAMP signalling pathways
seems central to infection-related development in M. aniso-
pliae (St Leger, 1993). Constitutive expression may be a
feature of some primary initiators of physiological pro-
cesses, so their importance will not be detected in micro-
array analyses. A downstream activity, MAP kinase kinase 2
(AJ273356) was upregulated in GC- and BC-containing
media. This enzyme, and other transcription factors, may
constitute downstream ‘ground-level’ components which
are immediately concerned with recognizing and respond-
ing to specific host features, and which do not control fungal
metabolism as a whole. As such, they may be useful for strain
improvement purposes.
Microarray technology has made it possible to decipher
the transcriptional programmes of organisms by studying
gene expression en masse while assessing individual gene
function in a detailed manner (Brown & Botstein, 1999).
Thus, knowing when and where a gene is expressed ofte n
provides a strong clue as to its function (DeRisi et al.,
1997). Almost 50 % of the arrayed ESTs upregulated in
cuticle-containing media have undiscovered biological acti-
vities (Table 1 and Supplementary Table S1), and 25 % of
these are not upregulated in MM or HL. These genes have
never been recognized to have a role in pathogenicity, but
are now implicated by co-regulation with known virulence
factors. They thus pro vide an additional rich resource for
future research.
Evolutionary theory has long held that the process of
adaptation is driven by competition for limited resources.
Among heterotrophic micro-organisms, the availability of
carbon limits the ability of these organisms to multiply. As a
result, the machinery of central metabolism is tuned to
exploit reduced carbon resources in natural environments,
where they vary greatly in both form and abundance (Ferea
et al., 1999 ). Comparisons between M. anisopliae, T. reesei
and S. cerevisiae suggest that the three fungi will respond
differently to environmental changes, presumably reflecting
their adaptation to predictable differences in the composi-
tion of thes e environments. The similarities between M.
anisopliae and T. reesei may reflect their close relationship
as clavicipitaceous pyrenomycetes. However, the alterna-
tively regulated forms of catabolic enzymes in M. anisopliae
and T. reesei suggest they will differ in how they coordinate
the regulation of key parts of metabolism, such as fermen-
tation at different levels of glucose. This could affect the
extent to which aerobic respiration prevails in glucose-rich
media.
Evidently, gene-duplication events and altered patterns
of regulation could provide mechanisms for evolution to
http://mic.sgmjournals.org 369
Microarray analysis of Metarhizium anisopliae
fine-tune ATP-producing pathways, allowing these organ-
isms to adapt to their different environments and nutri-
tional requirements. It is tempting to speculate that
fermentation may play a more pivotal role in the life of
M. aniso pliae, compared to that of T. reesei, to enable it to
exploit sugars in the anaerobic environment of the dead
host. However, complicating interpretation of these results,
ATP-producing pathways can be co-opted to other func-
tions. Thus, some fungal acetyl coenzyme A synthetases are
involved in the biosynthesis of secondary metabolites such
as penicillin, as well as in primary metabolism (Martinez-
Blanco et al., 1993). It is axiomatic that as more is learned
about the function of each gene, comparative studies on
transcriptomes will become an increasingly powerful tool
allowing predictive insights into the behavioural plasticity of
each saprophyte or pathogen.
ACKNOWLEDGEMENTS
This work was supported by the USDA (grant 2003-35302-13588).
F. M. F. acknowledges support from the Swiss National Science Founda-
tion. We would like to thank A. Wilson and S. Grundschober for
computer support.
REFERENCES
Bagga, S., Hu, G., Screen, S. E. & St Leger, R. J. (2004).
Reconstructing the diversification of subtilisins in the pathogenic
fungus Metarhizium anisopliae. Gene 324, 159–169.
Brown, P. O. & Botstein, D. (1999). Exploring the new world of the
genome with DNA microarrays. Nat Genet 21, 33–37.
Caracuel, Z., Roncero, M. I., Espeso, E. A., Gonzalez-Verdejo, C. I.,
Garcia-Maceira, F. I. & Di Pietro, A. (2003). The pH signalling
transcription factor PacC controls virulence in the plant pathogen
Fusarium oxysporum. Mol Microbiol 48, 765–779.
Chambergo, F. S., Bonaccorsi, E. D., Ferreira, A. J., Ramos, A. S.,
Ferreira, J. R. J. J. R., Abrahao-Neto, J., Farah, J. P. & El-Dorry, H.
(2002). Elucidation of the metabolic fate of glucose in the
filamentous fungus Trichoderma reesei using expressed sequence
tag (EST) analysis and cDNA microarrays. J Biol Chem 277,
13983–13988.
Davis, D., Wilson, R. B. & Mitchell, A. P. (2000). RIM101-dependent
and-independent pathways govern pH responses in Candida albicans.
Mol Cell Biol 20, 971–978.
DeRisi, J. L., Iyer, V. R. & Brown, P. O. (1997). Exploring the
metabolic and genetic control of gene expression on a genomic scale.
Science 278, 680–686.
Donzelli, B. G., Lorito, M., Scala, F. & Harman, G. E. (2001). Cloning,
sequence and structure of a gene encoding an antifungal glucan
1,3-
b-glucosidase from Trichoderma atroviride (T. harzianum). Gene
277, 199–208.
Doseff, A. I. & Arndt, K. T. (1995). LAS1 is an essential nuclear
protein involved in cell morphogenesis and cell surface growth.
Genetics 141, 857–871.
Dysvik, B. & Jonassen, I. (2001). J-Express: exploring gene
expression data using Java. Bioinformatics 17, 369–370.
Ferea, T. L., Botstein, D., Brown, P. O. & Rosenzweig, R. F. (1999).
Systematic changes in gene expression pattern following adaptive
evolution in yeast. Proc Natl Acad Sci U S A 96, 9721–9726.
Freimoser, F. M., Screen, S., Bagga, S., Hu, G. & St Leger, R. J.
(2003).
Expressed sequence tag (EST) analysis of two subspecies of
Metarhizium anisopliae reveals a plethora of secreted proteins with
potential activity in insect hosts. Microbiology 149, 239–247.
Goffeau, A., Barrell, B. G., Bussey, H. & 13 other authors (1996).
Life with 6000 genes. Science 274, 563–567.
Grundschober, A., Tuor, U. & Aebi, M. (1998). In vitro cultivation
and sporulation of Neozygites parvispora (Zygomycetes: Entomo-
phthorales). Syst Appl Microbiol 21, 461–469.
Holtzman, D. A., Yang, S. & Drubin, D. G. (1993). Synthetic–lethal
interactions identify two novel genes, SLA1 and SLA2, that control
membrane cytoskeleton assembly in Saccharomyces cerevisiae. J Cell
Biol 122, 635–644.
Hu, G. & St Leger, R. J. (2002). Field studies using a recombinant
mycoinsecticide (Metarhizium anisopliae) reveal that it is rhizosphere
competent. Appl Environ Microbiol 68, 6383–6387.
Hu, G. & St Leger, R. J. (2004). A phylogenomic approach to
reconstructing the diversification of serine proteases in fungi.
J Evolution Biol 17, 1204–1214.
Joshi, L. & St Leger, R. J. (1999). Cloning, expression, and substrate
specificity of MeCPA, a zinc carboxypeptidase that is secreted
into infected tissues by the fungal entomopathogen Metarhizium
anisopliae. J Biol Chem 274, 9803–9811.
Joshi, L., St Leger, R. J. & Roberts, D. W. (1997). Isolation of a
cDNA encoding a novel subtilisin-like protease (Pr1B) from the
entomopathogenic fungus, Metarhizium anisopliae using differential
display-RT-PCR. Gene 197, 1–8.
Martinez-Blanco, H., Orejas, M., Reglero, A., Luengo, J. M. &
Penalva, M. A. (1993).
Characterisation of the gene encoding acetyl-
CoA synthetase in Penicillium chrysogenum: conservation of intron
position in plectomycetes. Gene 130, 265–270.
Parsley, T. B., Chen, B., Geletka, M. & Nuss, D. L. (2002).
Differential modulation of cellular signaling pathways by mild and
severe hypovirus strains. Eukaryot Cell 1, 401–413.
Paterson, I. C., Charnley, A. K., Cooper, R. M. & Clarkson, J. M.
(1994).
Partial characterization of specific inducers of a cuticle-
degrading protease from the insect pathogenic fungus Metarhizium
anisopliae. Microbiology 140, 3153–3159.
Samuels, K. D. Z., Pinnock, D. E. & Allsopp, P. G. (1989).
The potential of Metarhizium anisopliae (Metschnikoff) Sorokin
(Deutermycotina, Hyphomycetes) as a biological control-agent of
Inopus rubriceps (Macquart) (Diptera, Stratiomyidae). J Aust Entomol
Soc 28, 69–74.
Screen, S. E. & St Leger, R. J. (2000). Cloning, expression, and
substrate specificity of a fungal chymotrypsin. Evidence for lateral
gene transfer from an actinomycete bacterium. J Biol Chem 275,
6689–6694.
St Leger, R. J. (1993). Biology and mechanisms of insect cuticle
invasion by deuteromycete fungal pathogens. In Parasites and
Pathogens of Insects, pp. 211–229. Edited by N. C. Beckage, S. N.
Thompson & B. A. Federici. San Diego, CA: Academic Press.
St Leger, R. J., Charnley, A. K. & Cooper, R. M. (1986a). Cuticle-
degrading enzymes of entomopathogenic fungi: mechanisms of
interaction between pathogen enzymes and insect cuticle. J Invertebr
Pathol 47, 295–302.
St Leger, R. J., Charnley, A. K. & Cooper, R. M. (1986b). Cuticle-
degrading enzymes of entomopathogenic fungi: synthesis in culture
on cuticle. J Invertebr Pathol 48, 85–95.
St Leger, R. J., Charnley, A. K. & Cooper, R. M. (1987).
Characterization of cuticle-degrading proteases produced by the
entomopathogen Metarhizium anisopliae. Arch Biochem Biophys 253,
221–232.
370 Microbiology 151
F. M. Freimoser, G. Hu and R. J. St Leger
St Leger, R. J., May, B., Allee, L. L., Frank, D. C., Staples, R. C. &
Roberts, D. W. (1992).
Genetic differences in allozymes and in
formation of infection structures among isolates of the entomo-
pathogenic fungus Metarhizium anisopliae. J Invertebr Pathol 60,
89–101.
St Leger, R. J., Bidochka, M. J. & Roberts, D. W. (1994). Isoforms of
the cuticle-degrading Pr1 proteinase and production of a metallo-
proteinase by Metarhizium anisopliae. Arch Biochem Biophys 313,
1–7.
St Leger, R. J., Joshi, L., Bidochka, M. J., Rizzo, N. W. & Roberts,
D. W. (1996).
Biochemical characterization and ultrastructural locali-
zation of two extracellular trypsins produced by Metarhizium aniso-
pliae in infected insect cuticles. Appl Environ Microbiol 62, 1257–1264.
St Leger, R. J., Joshi, L. & Roberts, D. (1998). Ambient pH is a major
determinant in the expression of cuticle-degrading enzymes and
hydrophobin by Metarhizium anisopliae. Appl Environ Microbiol 64,
709–713.
Warner, J. R. (1999). The economics of ribosome biosynthesis in
yeast. Trends Biochem Sci 24, 437–440.
Watanabe, R., Ohishi, K., Maeda, Y., Nakamura, N. & Kinoshita, T.
(1999). Mammalian PIG-L and its yeast homologue Gpi12p are
N-acetylglucosaminylphosphatidylinositol de-N-acetylases essential
in glycosylphosphatidylinositol biosynthesis. Biochem J 339, 185–192.
Yuen, T., Wurmbach, E., Pfeffer, R. L., Ebersole, B. J. & Sealfon,
S. C. (2002). Accuracy and calibration of commercial oligonucleotide
and custom cDNA microarrays. Nucleic Acids Res 30, 48.
http://mic.sgmjournals.org 371
Microarray analysis of Metarhizium anisopliae
