Core oxidative stress response in Aspergillus nidulans

Oxidative and salt stress induced genome-wide transcriptional changes in A. nidulans, which were highly depended on the type and strength of the stress (Tables 3 and 4, Fig. 2). The observed global stress responses were similar to those found by other researchers
earlier. For example, cell division-related processes, which are decisively important
in the maintenance of vegetative growth and which influence replication, transcription,
translation and cytoskeleton functions as well as sterol metabolism were inhibited
under severe stress conditions in this study (Additional file 6: Table S6) as well as in other previous studies and in other fungi 7], 8], 36], 67], 68]. Up-regulation of catalase and peroxidase genes, furthermore genes coding for elements
of glutathione, thioredoxin and trehalose metabolisms as well as for heat shock proteins
and parts of DNA repair (Additional file 1: Table S1) have also been observed by other researchers in various fungal species
7], 8], 36], 69], 70].

On the other hand, only weak correlation was found between the current gene expression
data sets and the data coming from our earlier transcriptome 36] and proteome 62] studies. Besides of the variations in the culture and stress conditions applied in
these works, the weak correlation can be clearly explained by several other important
differences between the stress conditions employed. In the recent study, we used a
whole-genome-based DNA chip with 60-mer oligonucleotide probes designed to reduce
cross-hybridizations between probes. In previous experiments by Pócsi et al. 36], an expressed sequence tag based DNA was applied, and in this case, potential cross-hybridization
between paralogue genes can be a serious problem 71]. It is also worth mentioning that the FGSC26 strain used in our previous study, in
contrast to strains used recently, harbored biA1 (biotin auxotrophy) and veA1 mutations 36]. Recent studies demonstrated that VeA is important in the regulation of oxidative
stress response in several fungi including A. flavus72], Cochliobolus heterostrophus73] and Botrytis cinerea74]. Moreover, nutritional supplements, e.g. riboflavin, paba, pyridoxine, can also have an impact on the observed stress sensitivity
of the strains tested 48]. Although biotin did not affect significantly the growth of A. nidulans48] the presence of this supplement in the medium may have influenced the stress response
at the level of transcriptome. Poor correlation is a common problem when proteome
and transcriptome data are compared, and the majority of the differences can be explained
with divergent post-transcriptional and post-translational regulations 75]. Considering all these variations in the experimental arrangements, it is remarkable
that certain genes, notably AN2846 (gpxA) encoding a putative glutathione peroxidase and AN3581 (trxR) coding for thioredoxin reductase 76], still showed steady up-regulation independently of the applied strains and methods
(Additional files 2 and 3: Tables S2 and S3). Up-regulation of these genes seems to be crucial during MSB induced
stress, which demonstrates the paramount importance of both the glutathione-based
and the thioredoxin-based elements of the oxidative stress defense system in A. nidulans36], 76], 77].

Environmental Stress Response (ESR) and COSR

ESR was first defined in the budding yeast Saccharomyces cerevisiae as a sum of stereotypical changes observable in the transcription of more than 900
genes in response to very different types of stress 15], 78]. In their study, Gasch et al. 15] found approximately 300 genes up-regulated and 600 genes down-regulated in more than
20 different stresses in baker’s yeast. Later, Chen et al. 16] studied global gene expression changes under five stress conditions (heat, H
₂
O
₂
, Cd
²⁺
, sorbitol and methylmethane sulfonate stress) in the fission yeast Schizosaccharomyces pombe, and demonstrated that approximately 140 genes showed more than two-fold increases
in their transcription in at least four stresses and approximately 100 genes showed
more than two-fold decreases in their expression in at least three stresses 16]. Importantly, the number of stress specific genes induced by only one stressor was
less than 100 in each case 16]. In the present study, the number of co-regulated genes (merely 7 genes co-induced
and 6 genes co-repressed under all the five oxidative stress conditions as well as
under NaCl exposures, and when l-H
₂
O
₂
treatments were omitted from analyses 51?+?65 genes showed co-regulation (Table 3). Meanwhile the number of co-regulated genes was small the number of genes regulated
exclusively by one certain type of stress was well above 1000 (Table 3). These observations together with the sharply decreasing number of co-regulated
genes as a function of the number of stress initiating agents studied (Table 3) does not support the existence of a S. cerevisiae-type ESR in A. nidulans.

We assume that the observed co-regulations were most likely consequences of the overlapping
physiological effects of the stressors especially in the case of severe stress treatments
and not of the existence of a general ESR. Severe stress causes aspecific damages
in versatile biomolecules like proteins, nucleic acids and lipids, decreases the ATP/AMP
ratio or influences the redox balance and ion homeostasis independently of the way
of the initiation of stress. Such non-specific physiological changes may be reflected,
at least to some extent, in the stress-initiated alterations in the transcriptome
profiles.

On the other hand, comparing stress treatments similar in type and strength can be
a useful and beneficial strategy to identify a group of genes co-regulated by the
same stress sensing and signaling pathways. COSR gene groups were constructed by identifying
and collecting co-regulated genes through mapping the global transcriptional changes
recorded under three “severe” oxidative stress conditions elicited by MSB, tBOOH and diamide treatments (Additional file 5: Table S5). Based on this experimental arrangement, COSR genes were found in great
number (873 genes) and, similarly to the ESR genes in S. cerevisiae15], 78], the function of the up-regulated genes was very diverse with no significant shared
GO term identified meanwhile the majority of the down-regulated genes was related
to the maintenance of vegetative growth, e.g. replication, cytoskeleton functions as well as nuclear and cell divisions. Further
studies are needed to identify the stress signaling and regulatory pathways governing
the expression of the COSR genes. It is noteworthy that two bZIP-type transcription
factors, NapA and RsmA, are transcriptionally regulated within the frame of COSR in
A. nidulans (Additional files 1 and 5: Tables S1 and S5), which suggests the importance of both the maintenance of the
redox homeostasis of the cells and the production of secondary metabolites 50], 63] as an inseparable part of the oxidative stress defense.

The characteristics of ESRs observed in various fungi are summarized in Table 7. Unfortunately, the experimental design (e.g. the type, the strength and the number of the tested stresses as well as the criteria
used to define stereotypical changes) was different in these experiments, which is
a limitation when we compare the data. However, Table 7 suggests that ESR may be limited only to the budding yeast S. cerevisiae and to its close relatives like C. glabrata, where Msn2/4 transcription factors evolved to regulate stress responses under a
wide spectrum of environmental stress. It is noteworthy that Msn2/4 regulate numerous,
but not all, genes up-regulated in ESR, and these transcription factors are probably
not involved in down-regulations 15]. Roetzer et al. 18] found only limited overlap (268 genes) between the ESRs of S. cerevisiae and C. glabrata, and this overlap is even smaller when other species are considered 16], 17]. These data question the existence of a universal stress-response set of genes in
fungi, the induction of which were equally beneficial in all fungal species and in
all ecological niches they occupy. Fungi seem to choose one of two options, evolving
a set of ‘unique’ stress responses or, instead, a ‘general’ stress response. A set
of stress-specific “unique” stress responses can provide the fungus with an appropriate
adaptation to a wide array of stress but need numerous genes and a complex and robust
signaling network to regulate them like that described in the a Aspergilli 9], 48], 79], 80]. On the other hand, a general stress response can be operated well even with less
genes and with a less complex signaling network and can provide the fungi like saccharomycetous
yeasts with a perhaps less sophisticated but instantaneous stress response even to
cope with impending stress 12], 13]. Importantly, the number of S. cerevisiae genes is approximately half of that of the Aspergilli, which indicates that the type
of stress response (“unique” vs. “general”) is likely also dependent on the size of the fungal genome (Table 7).

Table 7. Properties of ESR in different fungi

A wide spectrum of genetic evidence demonstrates that overexpression of even a single
gene can increase the stress tolerance 81], 82] and, therefore, if this gene is part of ESR its up-regulation by one stress can cause
adaptation to another stress 16], 78]. This explanation is commonly used to explain cross-stress adaptation phenomena and
the physiological significance of ESR. Cross-stress adaptation was also observed in
both the control and the ?atfA strains in our experiments (Fig. 3). The most interesting cross-stress adaptation was developed with H
₂
O
₂
when employed at 5 mM concentration, which alone caused only small transcriptional
changes in A. nidulans and these alterations in gene expressions were quite different from those caused
by MSB (Fig. 2; only 81 up-regulated and 24 down-regulated genes overlapped). Importantly, l-H
₂
O
₂
exposures did not elicit even two-fold increases in the transcriptions of genes encoding
basically important elements of MSB stress response, including FeS cluster assembly
and DNA repair proteins, trehalose, glutathione or thioredoxin metabolic and antioxidant
enzymes as well as heat shock proteins and metallo-chaperones (data not shown). However,
using 5 mM H
₂
O
₂
in stress pre-treatments resulted in clear-cut adaptation to severe MSB stress (Fig. 3). In accordance with these observations, Berry et al. 13] were able to induce H
₂
O
₂
tolerance in S. cerevisiae by pre-exposing baker’s yeast cultures to mild NaCl, dithiotreitol or heat stress
although there were only little overlaps in the lists of genes induced by different
pre-treatments. In another study, Guan et al. 83] found that Ctt1 catalase produced under NaCl pre-treatment was distributed to daughter
cells during subsequent divisions and was responsible for the elevated H
₂
O
₂
tolerance of S. cerevisiae cells. They also demonstrated that stress pre-treatments caused a faster response
in gene expression during subsequent high-dose stress treatments, which required the
nuclear pore component protein Nup42 83]. Furthermore, several studies have demonstrated the overlapping nature of stress
signaling pathways with numerous interplays, co-operations and even cross-talks between
them 8], 79], 80]. The regulation of this complex and robust network is based on protein-protein interactions
and/or modifications rather than on transcriptional changes alone. A possible explanation
for cross-stress adaptation therefore is that various pre-treatments can activate
the signaling network, which increases subsequently the efficiency of sensing of and/or
responding to versatile types of environmental stress. It is possible that changes
in the expressions of stress response genes during pre-treatments contribute to the
adaptation to impending, more severe environmental stress. However, we suggest that
transcriptional up-regulations of stress response genes under pre-treatments are not
essential to reach cross-stress adaptations in A. nidulans.

Involvement of AtfA in the regulation of stress response in A. nidulans

Atf1 is a bZIP-type transcription factor regulated by the Sty1 MAPK pathway in S. pombe and is responsible for regulation of genes involved in various stress responses including
heat, oxidative, reductive, osmotic and starvation stress 84]. Atf1 can form heterodimer with another bZIP-type transcription factor, Pcr1 and
some of the target genes are regulated by this hetrodimer 85]. In A. nidulans, the Atf1 orthologue AtfA is regulated by the HogA/SakA MAPK pathway 49]. The phenotypes of the ?atfA gene deletion strains demonstrate that AtfA is necessary for normal vegetative growth
and sporulation as well as for oxidative and heat stress tolerance in A. nidulans46]–49] (Fig. 1, Table 1). In the present study, the deletion of atfA affected the transcription of an unexpectedly high number of genes under MSB stress
(Fig. 2, Table 5). In contrast, the transcriptome profiles of the ?atfA mutant and the control strains were more similar during H
₂
O
₂
, tBOOH, NaCl and especially under diamide treatments (Fig. 2, Table 5). Moreover, the lack of atfA also affected the transcription of several genes under unstressed conditions. Deletion
of atf1 in S. pombe also caused inductions and repressions of several genes even in unstressed cultures
and also prevented the induction of numerous genes under stress treatments 16].

The stress-dose dependent activation of Atf1 is also well described in S. pombe. Atf1 regulates the oxidative stress response in high dose H
₂
O
₂
treatments while its importance is less significant when fission yeast was exposed
to low H
₂
O
₂
concentrations when the Pap1 transcription factor played a key role 86]. In A. nidulans, the ratio of AtfA-dependent genes was much higher in l-H
₂
O
₂
(5 mM) elicited stress than in h-H
₂
O
₂
(75 mM) triggered stress (Table 5). However the overlap between the two stress responses was significantly less in
the ?atfA mutant than in the control strain (Table 4). These observations suggest that the majority of AtfA-dependent genes in l-H
₂
O
₂
stress were also part of the h-H
₂
O
₂
stress response. In fact, 77 out of the 98 AtfA-dependent genes recorded in l-H
₂
O
₂
stress also showed stress responsive regulation under h-H
₂
O
₂
exposures. Therefore, the low ratio of the AtfA-dependent genes during h-H
₂
O
₂
elicited stress was not the consequence of the decreased number of AtfA-dependent
genes but could be attributed to the increased number of AtfA-independent genes instead.

One of the main differences between the regulations of the oxidative stress responses
in A. nidulans and S. pombe is that the number of genes likely under AtfA control was more stress-type-dependent
in A. nidulans than in S. pombe (Table 5). Several genes showing AtfA-dependent regulation in one stress treatment did not
show any AtfA-dependency under another stress condition in our experiments. It is
remarkable that even in the group of the COSR genes merely two showed an AtfA-dependent
regulatory pattern (Additional file 5: Table S5).

Several GO terms related to stress signaling and regulation (“phosphorelay signal
transduction system”, “regulation of protein phosphorylation”, “calcium ion transmembrane
transport” as well as “response to stimulus”) were typical of the group of down-regulated
genes in the ?atfA mutant under unstressed conditions (Additional file 7: Table S7). According to this, we hypothesize that AtfA coordinates the up-regulation
of certain regulatory genes (e.g. members of the phosphorelay signal transduction
system; Additional file 7: Table S7) under environmental stress and also determined their basal transcription
levels under unstressed conditions. Decreases in the expressions of these genes in
the gene deletion mutant grown in unstressed cultures disturbed the homeostasis of
the strain, resulted in alterations in the transcription patterns of a large number
of other genes (Fig. 2, Table 5, Additional file 7: Table S7).

Due to the networking nature of signaling pathway, the missing AtfA was compensated
by other regulatory proteins under H
₂
O
₂
, tBOOH, diamide, NaCl, which resulted in global transcriptional profiles very similar
to those recorded for the control strain (Fig. 2, Table 5). Considering a most recently published study of Bok et al. 50] the transcription factor NapA (orthologue of S. pombe Pap1 86]), another bZIP-type oxidative stress response regulator, can be a candidate which
may take over AtfA functions under hydrogen peroxide induced oxidative stress. NapA,
which is under RsrA control, seems to be the master regulator of the specific response
to peroxide stress 50].

When cells were exposed to MSB the signaling network was unable to substitute AtfA
satisfactorily, which resulted in serious disturbances in the cell homeostasis and
concurrently altered the transcription levels of a large group of genes, which were
therefore described as potential AtfA targets (Fig. 2, Table 5). Further research is needed to identify which genes among the AtfA target genes
responsible for the efficient stress response under MSB stress treatment.

Secondary metabolism and stress response

Emerging data demonstrate that there are interplays between the regulations of oxidative
stress response and secondary metabolism in fungi. Induction of secondary metabolite
production by oxidative stress and its inhibition by antioxidants have been observed
in several species, and even transcription factors affecting both secondary metabolism
and oxidative stress response have been identified 42]–44], 63], 66], 87]–90]. Our transcriptome data also support the importance of stress in the regulation of
secondary metabolism because all stressors including NaCl affected significantly the
transcription of secondary metabolite biosynthesis genes (Table 6, Additional file 8: Table S8) under experimental conditions like culturing at 37 °C in glucose containing
minimal medium, which are generally not beneficial for secondary metabolite production
in this species. Moreover, the ?atfA strain, in addition to being more sensitive to oxidative stress (Fig. 1), also had a more altered expression pattern of secondary metabolism genes under
various stress treatments when compared to the control strain (Table 6, Additional files 8 and 9: Tables S8 and S9).

Both rsmA (AN4562) and napA (AN7513) was part of the COSR in the control strain (Addtional file 5: Table S5) which is in good accordance with increased number of up-regulated secondary
metabolit genes a secondary metabolit key genes under MSB, tBOOH and diamide stress in comparison to the other stress applied (Table 6). It supports the view that RsmA and NapA can be a link between the regulation of
stress response and secondary metabolite production in A. nidulans43], 63], 89].

Cryptic secondary metabolite gene clusters are in the center of industrial investigations
since they may be exploitable in the production of novel secondary metabolites. In
addition to the overexpression of the complete gene cluster in a suitable organism
91], the overexpression of a cluster-specific transcription factor or, alternatively,
a global regulator of secondary metabolism, e.g. LaeA, in the host organism 92], 93] are frequently used techniques to identify the products of cryptic gene clusters.
Our study demonstrate that deletion of an oxidative stress response regulator gene
in combination with mild oxidative stress can also be applicable to overproduce the
products of certain gene clusters, which may lead to identification of new secondary
metabolites (Additional file 9: Table S9).

It is worth noting that many known secondary metabolite gene clusters of A. nidulans were not stress responsive in our experiments (Additional files 9 and 10: Tables S9 and S10). Since we studied only the early global transcriptional changes
further studies will address the question if the regulation of these clusters and
genes are independent of environmental stress or the progression of the transcriptional
changes will need more time. More importantly, environmental stress cannot only induce
some secondary metabolite gene clusters but can also repress others (Table 6, Additional file 9: Table S9). Strategies based on the application of certain antioxidants to prevent
the accumulation of ROS and, consequently, the formation of mycotoxins may efficiently
inhibit the production of certain mycotoxins but, concomitantly, may also induce the
formation of other unwanted secondary metabolites.

The ecological and/or physiological value of the redox regulation of secondary metabolite
production is unclear. Reverberi et al. 42] suggested that production at least some the secondary metabolites (e.g. aflatoxins) contained several oxidative steps therefore their biosynthesis helped
maintaining the redox status of the cells under oxidative stress. On the other hand,
beside well-known and well-characterized ROS productions observable during pathogen
– host interactions, competing micro-organisms can also elicit oxidative stress either
through the extracellular formation of H
₂
O
₂94], 95] or via do novo synthesis of secondary metabolites, which generate oxidative stress in sensitive
organisms 96]–98]. Such ecological role can be attributed for example to aflatoxin B1 produced by certain
aspergilli 99]. Similarly, hyperosmotic stress is also frequently induced e.g. by ethanol producing microbes 100]. The “artificial” stressors used in research laboratories may imitate the attack
of a competitive species or a host organism, which may explain the stress dependent
regulation of secondary metabolite clusters. Other explanations, like the use of secondary
metabolite spectra to inform other cells of the same species about the physiological
status of a given cell cannot be ruled out either. Connections between secondary metabolite
production and development as well as the importance of secondary metabolites or secondary
metabolite-like compounds in the regulation and coordination of sporulation, germination
or sexual development of a colony have been demonstrated by several researchers 101]–103].