1 Cryptic Aspergillus nidulans antimicrobials 1 2 Steve S. Giles1*, Alexandra A. Soukup2*, Carrie

30 Pages·2011·0.68 MB·English

AEM Accepts, published online ahead of print on 8 April 2011
Appl. Environ. Microbiol. doi:10.1128/AEM.02000-10
Copyright © 2011, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.
1 Cryptic Aspergillus nidulans antimicrobials
2
3 Steve S. Giles1*, Alexandra A. Soukup2*, Carrie Lauer1, Mona Shaaban1,2, Alexander
4 Lin4, Berl R. Oakley3, Clay C. C. Wang4,6, Nancy P. Keller1,7,**
5
6 1Department of Medical Microbiology and Immunology, University of Wisconsin-
D
o
7 Madison, Madison, WI
w
n
8 2Department of Genetics, University of Wisconsin-Madison, Madison, WI lo
a
9 3Microbiology and Immunology, Faculty of Pharmacy, Mansoura University, Egypt. d
e
10 5 Department of Molecular Biosciences, University of Kansas, 1200 Sunnyside Ave., d
f
r
11 Lawrence, KS 66045, USA om
12 4 Department of Pharmacology and Pharmaceutical Sciences, 6Department of Chemistry, h
t
t
p
13 University of Southern California, 1985 Zonal Avenue, Los Angeles, California, USA :
/
/
a
14 90033 e
m
15 7Department of Bacteriology, University of Wisconsin-Madison, Madison, WI .
a
s
16 * equal contribution m
.
o
17 r
g
/
18 o
n
19 J
a
n
20 **Corresponding author:
u
a
21 Phone: [phone]; fax: [phone] r
y

4
22 E-mail: [email protected] ,

2
23 0
1
9

b
24 Running Title: silent secondary metabolites y

g
u
25
e
26 s
t
27
28
29
30
31
32
1
33 Summary
34 Secondary metabolite (SM) production by fungi is hypothesized to provide some
35 fitness attribute for the producing organisms. However, most SM clusters are ‘silent’
36 when fungi are grown in traditional laboratory settings, and it is difficult to ascertain any
37 function or activity of these SM cluster products. Recently, the creation of a chromatin
D
o
w
38 remodeling mutant in Aspergillus nidulans induced activation of several cryptic SM gene
n
lo
a
39 clusters. Systematic testing of nine purified metabolites from this mutant identified an
d
e
d
40 emodin derivate with efficacy against both human fungal pathogens (inhibiting both
f
r
o
m
41 spore germination and hyphal growth) and several bacteria. The ability of catalase to

h
t
t
42 diminish this antimicrobial activity implicates reactive oxygen species generation, p
:
/
/
a
43 specifically the generation of hydrogen peroxide, as the mechanism of emodin hydroxyl e
m
.
a
44 activity. s
m
.
o
r
g
/

o
n

J
a
n
u
a
r
y

4
,

2
0
1
9

b
y

g
u
e
s
t
2
45 Introduction
46 For decades, natural products produced by bacteria and fungi have been mined for use as
47 antimicrobials. The easily accessible compounds (e.g. β-lactams, macrolides) were the
48 first to be mined and patented for drug development. Recently, advances in genome
49 sequencing have revealed an arsenal of untapped secondary metabolite (SM) gene D
o
w
50 clusters, in both Kingdoms (Bentley et al., 2002; Nierman et al., 2005; Omura et al., n
lo
a
51 2001). In many of these organisms the number of SM gene clusters (20-50) far exceeds d
e
d

52 the number of known metabolites; hence the concept of ‘silent’ SM gene clusters fr
o
m
53 (Brakhage et al., 2008; Chiang et al., 2008; Cichewicz, 2010). The inaccessibility to
h
t
t
p
54 these compounds in laboratory settings is hypothesized to be due to inability to replicate :
/
/
a
e
55 the conditions that induce SM activation in nature. m
.
a
56 In their ecological habitat, microbes are constantly being challenged, often by s
m
.
o
57 other microbes. These microbial interactions activate SM clusters that produce bioactive r
g
/

o
58 compounds that offer protection for the organism, niche security, and a myriad of other n

J
a
59 benefits (Rolhfs et al., 2007). A case in point is the A. fumigatus gliotoxin gene cluster. n
u
a
60 Although most widely referred to in its role as a toxin in invasive aspergillosis (Dagenais ry

4
,
61 and Keller, 2009), gliotoxin is a potent antifungal and likely helps secure A. fumigatus 2
0
1
62 dominance in soil environments (Aliaa, 2008; Brian and Hemming, 1945; Howell and 9

b
y
63 Stipanovic, 1995; Larkin and Fravel, 1998; Lorito et al., 1994). A. fumigatus self protects g
u
e
64 from endogenous gliotoxin by expressing a gliotoxin reductase, GliT- also known to s
t
65 exhibit oxidase activity, encoded in the cluster (Schrettl et al., 2010; Scharf et al. 2010).
66 Because the gliotoxin gene cluster is expressed under most laboratory conditions, this
67 metabolite was identified decades ago (Weindling and Emerson, 1936). However, most
3
68 antimicrobial gene clusters are silent in laboratory environments presumably because
69 there is little to no competition, and therefore no need for protection. Until recently, the
70 hurdle of silent gene cluster activation in laboratories has limited the mining resources for
71 drug discovery and hindered any understanding of the ecological role of these SM. With
72 the advent of genome sequencing and recent advances in epigenetic modification, D
o
w
73 formerly silent gene clusters can be activated resulting in the production of novel n
lo
a
74 compounds. d
e
d
75 Epigenetic modification is best illustrated in the model organism, Aspergillus f
r
o
m
76 nidulans. The secondary metabolome of this fungus contains at least 25 SM gene
h
t
t
p
77 clusters, many of which are characteristically silent in the laboratory environment, and :
/
/
a
e
78 therefore considerable numbers of its natural products are unknown (Brakhage et al., m
.
a
79 2008; Chiang et al., 2008; Cichewicz, 2010). However, through chromosomal s
m
.
o
80 remodeling of A. nidulans, an array of novel bioactive compounds can be retrieved. For r
g
/

o
81 example, deletion of cclA, which encodes a protein (CclA) that is involved in histone 3 n

J
a
82 lysine 4 (H3K4) methylation, results in the activation of the several gene clusters yielding n
u
a
83 emodin derivatives including monodictyophenone and the anti-osteoporosis polyketides, r
y

4
,
84 F9775A and F9775B derived from orsellinic acid (Bok et al., 2009; Sanchez et al., 2010). 2
0
1
85 At the same time loss of CclA was found to induce these gene clusters, a co- 9

b
y
86 culture of A. nidulans with Streptomyces hygroscopicus resulted in the induction of g
u
e
87 several cryptic SMs including the F9775 metabolites (Schroeckh et al., 2009). We s
t
88 hypothesized that the cclA deletant unveiled microbial SM induction pathways utilized by
89 A. nidulans during interaction with other microbes. The purpose of this present study was
90 to test this hypothesis by examining potential antimicrobial activity of compounds
4
91 released by activated pathways against possible competitor microbes (Table 1). Here we
92 find that a derivative of emodin, 2-hydroxyemodin (2-OH), effectively inhibits the
93 growth of A. fumigatus, A. flavus, Candida albicans, Bacillus cereus, and Micrococcus
94 luteus. We suggest the process of SM cluster activation due to competition in nature can
95 be mimicked through epigenetic modifications, resulting in an arsenal of novel D
o
w
96 compounds that can be characterized for their ecological roles and mined for new n
lo
a
97 therapeutics. d
e
d
98 f
r
o
m
99 Results
h
t
t
p
100 Competition growth assays :
/
/
a
e
101 Our laboratory has previously shown that the loss of CclA, which is involved in m
.
a
102 H3K4 methylation, turns on cryptic secondary metabolite gene clusters in A. nidulans. s
m
.
o
103 This resulted in the production of several silent secondary metabolites (SM) including r
g
/

o
104 monodictyophenone, emodin and emodin derivatives, and the anti-osteoporosis n

J
a
105 polyketides, F9775A and F9775B (Bok et al., 2009). The expansive armamentarium of n
u
a
106 SM produced by fungi is thought to provide an adaptive advantage, allowing them to r
y

4
,
107 compete with other microbiota and colonize diverse ecological niches. We hypothesized 2
0
1
108 that the cryptic SM gene clusters (e.g. the monodictyophenone cluster emodin derivates) 9

b
y
109 turned on in the A. nidulans cclA∆ strain might produce SM with antimicrobial activity g
u
e
110 against other fungi and bacteria. s
t
111 To investigate this hypothesis we co-cultured wild type and three A. nidulans
112 mutants (cclA∆ upregulated in monodictyophenone and other SM clusters, alcA(p):mdpE
113 upregulated in just the monodictyophenone cluster, and cclA∆ mdpG∆ unable to express
5
114 metabolites in the monodictyophenone cluster but upregulated in other SM) with wild
115 type A. fumigatus on GMM agar medium (WT, cclA∆, and cclA∆ mdpG∆) or LMM agar
116 + cyclopentanone (alcA(p):mdpE). Interestingly, we observed that the radial growth of A.
117 fumigatus mycelia would not extend into regions of the plate containing strains where the
118 monodictyophenone cluster was expressed (cclA∆ and alcA(p):mdpE) (Fig. 1). In D
o
w
119 contrast, the mycelia of wild type A. nidulans and A. fumigatus strains overlapped with n
lo
a
120 no signs of inhibitory activity. Inhibition by the cclA∆ strain was abrogated in a strain d
e
d
121 lacking the backbone PKS for this cluster (cclA∆ mdpG∆). Similar results were observed f
r
o
m
122 with the mutant series against wild type A. nidulans (data not shown). These results
h
t
t
p
123 indicated that upregulation of the monodictyophenone gene cluster produced SM with :
/
/
a
e
124 inhibitory activity against wild type strains of A. fumigatus and A. nidulans. m
.
a
125 Next, crude extracts were prepared from A. nidulans cclA∆ and wild type strains s
m
.
o
126 grown on GMM agar at 37˚C. The aqueous and organic fractions of the crude extracts r
g
/

o
127 were then assessed for antifungal activity against A. fumigatus wild type AF293 using a n

J
128 96-well microtiter plate based assay. A. fumigatus spores (1x103 CFU/well) were added an
u
a
129 to wells containing various concentrations of the crude extracts and germination was r
y

4
,
130 assessed microscopically over time to evaluate inhibitory activity. Although aqueous 2
0
1
131 extracts had no effect on fungal growth (data not shown), we did observe that organic 9

b
y
132 extracts from A. nidulans cclA∆, but not A. nidulans wild type, possessed potent g
u
e
133 antifungal activity and largely inhibited the germination of A. fumigatus spores (Fig. 2A s
t
134 and C). This result was consistent with our hypothesis that SM(s) with antifungal activity
135 were produced as a result of the loss of CclA.
136
6
137 Antimicrobial activity of purified compounds from A. nidulans cclA∆
138 We previously identified several organically soluble SM produced by the A.
139 nidulans cclA∆ mutant strain, including desacetylaustin, dehydroaustinol, citreorosein,
140 emodic acid, 2 hydroxyemodin (2-OH), 2 aminoemodin, emodin, and the polyketides,
141 F9775A and F9775B (Bok et al., 2009). We speculated that one or more of these SM D
o
w
142 might mediate the inhibitory activity observed with the crude extracts. First, each of the n
lo
a
143 compounds were purified and assessed for antifungal activity against A. fumigatus spores d
e
d
144 using the 96-well microtiter plate assay at concentrations ranging from 9 µM to 567 µM. f
r
o
m
145 We observed that of the nine compounds tested only 2-OH exhibited significant
h
t
t
p
146 antifungal activity against A. fumigatus, and that 2-OH activity reconstitutes the :
/
/
a
e
147 inhibition of germination effects seen in treatment with cclA∆ extracts (Fig. 2B and C). m
.
a
148 To further investigate whether the antifungal activity exhibited by 2-OH was s
m
.
o
149 specific for A. fumigatus, or if 2-OH exhibited broader antimicrobial activity, we assessed r
g
/

o
150 2-OH activity against three fungi (A. fumigatus, A. flavus, and C. albicans) and three n

J
a
151 bacteria (P. aeruginosa, B. cereus, and M. luteus). We observed that 2-OH exhibited n
u
a
152 robust antimicrobial activity against both Aspergilli, C. albicans, and two of the three r
y

4
,
153 bacterial species tested (Table 2). A. flavus (MIC100 142 to 284 µM) was more resistant 2
0
1
154 than A. fumigatus (MIC 35 to 71 µM), which was more resistant than C. albicans 9
100
b
y
155 (MIC100 9 to 18 µM). 2-OH was also quite effective towards B. cereus (MIC100 ≤ 9 µM) g
u
e
156 and M. luteus (MIC ≤ 9µM), while P. aeruginosa showed significantly higher s
100 t
157 resistance to 2-OH (MIC 284 to 567 µM) (Table 2).
158
159 2-OH mechanism of action
7
160 To begin the investigation of potential mechanisms of action, we assessed the
161 inhibitory activity of 2-OH against a collection of A. fumigatus strains with known
162 antifungal drug resistance to Voriconazole, Itraconazole, or Posaconazole. The rationale
163 for this approach was that if a strain with a known phenotype (and corresponding
164 genotype) exhibited resistance to 2-OH then we could gain insights into the mechanism D
o
w
165 by which 2-OH kills A. fumigatus. All of the A. fumigatus drug resistant strains assessed n
lo
a
166 exhibited susceptibility to 2-OH that was comparable to wild type A. fumigatus (data not d
e
d
167 shown). This suggested that the mechanism of activity was novel and independent of f
r
o
m
168 defined pathways known to be involved in resistance to azole drugs.
h
t
t
p
169 Emodin (1, 3, 8-trihydrozy-6-methylanthraquinone) is a free radical generator that :
/
/
a
e
170 when used as a therapeutic can cause immunosuppression, vasorelaxation, and m
.
a
171 hypolipidemia (Wu et al., 2007). The production of free radicals, specifically hydrogen s
m
.
o
172 peroxide, is thought to contribute to the immunosuppressive activities of emodin (Huang r
g
/

o
173 et al., 1992). 2-OH has also been shown to specifically produce hydrogen peroxide and n

J
a
174 not superoxide (Kodama et al., 1987). Therefore, we hypothesized that hydrogen n
u
a
175 peroxide might account for the fungicidal activity of 2-OH against A. fumigatus. To test r
y

4
,
176 this hypothesis, we assessed whether catalase, which catalyzes the reduction of hydrogen 2
0
1
177 peroxide to water and oxygen, could diminish the inhibitory activity of 2-OH against A. 9

b
y
178 fumigatus. In agreement with our hypothesis, we observed that catalase did effectively g
u
e
179 abrogate the inhibitory activity of 2-OH against A. fumigatus (Table 3). The complete s
t
180 loss of antifungal activity in the presence of catalase suggested that the inhibitory activity
181 of 2-OH was dependent on the production of hydrogen peroxide. Interestingly,
8
182 challenge of A. fumigatus with 2-OH did not result in increased endogenous catalase
183 activity (data not shown), in accordance with the sensitivity to this metabolite.
184
185 Lack of self resistance activity.
186 Frequently fungal SM gene clusters include a gene encoding a protein that imparts D
o
w
187 a resistance mechanism for the fungus against the respective SM produced by the cluster. n
lo
a
188 For instance, gliT in the gliotoxin gene cluster encodes a reductase that detoxifies d
e
d
189 gliotoxin in order to protect A. fumigatus against endogenous gliotoxin (Scharf et al., f
r
o
m
190 2010; Schrettl et al., 2010). We were interested to see if a cclA∆ strain (over-producing
h
t
t
p
191 2-OH), a cclA∆, mdpG∆ (unable to produce 2-OH or other metabolites in this pathway :/
/
a
e
192 due to deletion of mpdG encoding the monodictyophenone cluster polyketide synthase, m
.
a
s
193 Chiang et al., 2010), and an alcA(p):mdpE strain (overproducing 2-OH due to m
.
o
194 overexpression of mdpE encoding the monodictyophenone pathway specific transcription rg
/

o
195 factor, Chiang et al., 2010) showed variances in susceptibility to 2-OH. Table 4 shows n

J
a
196 that there was little difference in 2-OH sensitivity between these three A. nidulans strains n
u
a
r
197 and wild type. The slight increase in sensitivity of cclA∆ may be suggestive of the y

4
,

198 endogenous overproduction of 2-OH in this strain, although this does not appear to be the 2
0
1
9
199 case for the alcA(p):mdpE strain. Regardless, this suggests that the monodictyophenone
b
y

200 gene cluster does not include a gene that encodes resistance to 2-OH mediated growth g
u
e
s
201 inhibition. t
202 We also speculated that the reduced rate of growth and/or altered morphology of
203 all three A. nidulans mutants might give insight into SM function in A. nidulans.
204 Macroscopic and microscopic examination of all three strains as compared to wild type
9
205 revealed a dominant impact of cclA∆ on diameter growth that was not alleviated in the
206 double cclA∆ mpdG∆ strain (Fig. 3). However, loss of monodictyophenone cluster
207 expression in cclA∆ did result in an increase in altered colony morphology. Because
208 induction of the alcA(p):mdpE allele led to decreased/delayed conidial production, it is
209 possible that one or more of the monodictyophenone cluster generated metabolites has a D
o
w
210 negative impact on conidiophore development. n
lo
a
211 d
e
d
212 Discussion f
r
o
m
213 Historically, the vast majority of antimicrobial compounds have either been
h
t
t
p
214 natural products or derived from natural products. These include a diverse spectrum of :
/
/
a
e
215 drugs including antibiotics, antifungals, antiparasitics, lipid control agents, m
.
a
216 immunosuppressants, and anticancer drugs (Li and Vederas, 2009). In the last two s
m
.
o
217 decades there has been a shift in pharmaceutical research that has emphasized high r
g
/

o
218 throughput screening of large libraries of synthetically produced compounds. This shift in n

J
a
219 focus away from natural products occurred in part because of the difficulty of finding n
u
a
220 new novel natural products. For example, members of the Aspergilli are known to be rich r
y

4
,
221 sources of bioactive metabolites but there remains a discontinuity between the vast 2
0
1
222 number of SM gene clusters and the relatively few compounds described for the genus - 9

b
y
223 leaving the true SM potential of these fungi unknown (Brakhage et al., 2008; Chiang et g
u
e
224 al., 2008; Cichewicz, 2010). s
t
225 The disproportionate cluster to SM ratio of Aspergillus species is due largely to
226 the inability to promote the expression of SM gene clusters in the laboratory. However,
227 recent genomic mining of A. nidulans has uncovered several techniques to stimulate
10