While ustiloxin B is a nonribosomal peptidelike compound, no predicted NRPS genes could
be linked to its biosynthesis. The MIDDAS-M
analysis of the transcriptome data from A. flavus grown under different conditions showed
three differentially expressed clusters. Finally,
the ustiloxin B cluster was identified and verified by gene deletion.
C. Proteomics
To complement the genomic and metabolomic
strategies, proteomic studies offer a snapshot of
an organism’s secondary metabolism from the
perspective of active biosynthetic enzymes.
This approach is, naturally, based on an
assumption that the changes in enzyme types
and quantities are projected in modifications of
the metabolome. Indeed, the expression levels
of biosynthetic proteins are often directly correlated to the levels of NP produced (Vo ¨disch
et al. 2011; Gubbens et al. 2014).
Furthermore, the comprehensive nature of
a proteome study allows for the detection of
posttranslational modifications that may
occur to identify active enzymes. Du and van
Wezel (2018) provide an in-depth review of
activity-based protein profiling (ABPP), a
cost-efficient set of methods providing a snapshot of active biosynthetic potential in a complex biological sample. Combining ABPP with
probes for conserved features of biosynthetic
enzymes, such as carrier protein (CP) and
thioesterase domains, and coupling to mass
spectrometry was used to explore and quantify
PKS and NRPS enzymes [the OASIS (Orthogonal Active Site Identification System) by Meier
et al. (2009)]. It is important to note that these
methods are limited in their range by the availability of target domains in the sample, potentiating bias in the final results. Thus, future
developments of novel probes for modular
enzymes could enable a wide implementation
of probe-based methods in secondary metabolism research in fungi, as well as other phyla
(Meier and Burkart 2011).
Beyond discovery of biosynthetic enzymes,
proteomics is also useful for characterizing
peptidic NPs and determining the mode of
action of bioactive NPs. Comparing the whole
proteome of an NP-treated organism to an
untreated control provides a global overview
of changed pathways. Alternatively, a metabolite can be immobilized and used as bait to
identify proteins with a sufficient binding affinity, thus identifying potential target proteins
that can be further characterized by proteomics
(Du and van Wezel 2018).
The effects of gliotoxin on A. fumigatus were elucidated
in this manner: the toxin was found to trigger oxidative
stress by induction of a superoxide dismutase (Carberry et al. 2012), but also attenuated the effects of
H 2 O 2 -induced stress on the fungus (Owens et al. 2014).
Even a modified version of genome mining
is possible from a proteomics starting point:
the platform PrISM [Proteomic Investigation of Secondary Metabolites (Bumpus
et al. 2009)] allows for targeted detection of
NRPS- and PKS-derived peptides as well as
the corresponding BGC in unsequenced genomes. The method makes use of the labile
phosphodiester linkage of the phosphopantetheinyl cofactor to the thiolation domain of
NRPS and PKS enzymes.
This bond easily breaks during mass spectrometry,
releasing a cofactor fragment, thus creating a characteristic mass difference that allows the identification of
peptides from the thiolation domain.
Peptide sequences can then be used as templates to create primers and amplify the
corresponding gene from the genome without
any previous sequence information. Sequencing of the gene, domain predictions, and alignment to known genes can then be used to
postulate a hypothesis regarding the enzyme
product, leading to its targeted detection.
The complementation of proteomic methods with other approaches, e.g., proteogenomics (an integrated methodology combining
genomics, transcriptomics, and proteomics)
(Albright et al. 2014), along with the advancing
possibilities of exploiting metaproteomics from
microbial communities (Wang et al. 2016b),
promises deep insight into, and large expansion
of our knowledge of fungal biosynthetic pathways.
284
M. Flak et al.
be linked to its biosynthesis. The MIDDAS-M
analysis of the transcriptome data from A. flavus grown under different conditions showed
three differentially expressed clusters. Finally,
the ustiloxin B cluster was identified and verified by gene deletion.
C. Proteomics
To complement the genomic and metabolomic
strategies, proteomic studies offer a snapshot of
an organism’s secondary metabolism from the
perspective of active biosynthetic enzymes.
This approach is, naturally, based on an
assumption that the changes in enzyme types
and quantities are projected in modifications of
the metabolome. Indeed, the expression levels
of biosynthetic proteins are often directly correlated to the levels of NP produced (Vo ¨disch
et al. 2011; Gubbens et al. 2014).
Furthermore, the comprehensive nature of
a proteome study allows for the detection of
posttranslational modifications that may
occur to identify active enzymes. Du and van
Wezel (2018) provide an in-depth review of
activity-based protein profiling (ABPP), a
cost-efficient set of methods providing a snapshot of active biosynthetic potential in a complex biological sample. Combining ABPP with
probes for conserved features of biosynthetic
enzymes, such as carrier protein (CP) and
thioesterase domains, and coupling to mass
spectrometry was used to explore and quantify
PKS and NRPS enzymes [the OASIS (Orthogonal Active Site Identification System) by Meier
et al. (2009)]. It is important to note that these
methods are limited in their range by the availability of target domains in the sample, potentiating bias in the final results. Thus, future
developments of novel probes for modular
enzymes could enable a wide implementation
of probe-based methods in secondary metabolism research in fungi, as well as other phyla
(Meier and Burkart 2011).
Beyond discovery of biosynthetic enzymes,
proteomics is also useful for characterizing
peptidic NPs and determining the mode of
action of bioactive NPs. Comparing the whole
proteome of an NP-treated organism to an
untreated control provides a global overview
of changed pathways. Alternatively, a metabolite can be immobilized and used as bait to
identify proteins with a sufficient binding affinity, thus identifying potential target proteins
that can be further characterized by proteomics
(Du and van Wezel 2018).
The effects of gliotoxin on A. fumigatus were elucidated
in this manner: the toxin was found to trigger oxidative
stress by induction of a superoxide dismutase (Carberry et al. 2012), but also attenuated the effects of
H 2 O 2 -induced stress on the fungus (Owens et al. 2014).
Even a modified version of genome mining
is possible from a proteomics starting point:
the platform PrISM [Proteomic Investigation of Secondary Metabolites (Bumpus
et al. 2009)] allows for targeted detection of
NRPS- and PKS-derived peptides as well as
the corresponding BGC in unsequenced genomes. The method makes use of the labile
phosphodiester linkage of the phosphopantetheinyl cofactor to the thiolation domain of
NRPS and PKS enzymes.
This bond easily breaks during mass spectrometry,
releasing a cofactor fragment, thus creating a characteristic mass difference that allows the identification of
peptides from the thiolation domain.
Peptide sequences can then be used as templates to create primers and amplify the
corresponding gene from the genome without
any previous sequence information. Sequencing of the gene, domain predictions, and alignment to known genes can then be used to
postulate a hypothesis regarding the enzyme
product, leading to its targeted detection.
The complementation of proteomic methods with other approaches, e.g., proteogenomics (an integrated methodology combining
genomics, transcriptomics, and proteomics)
(Albright et al. 2014), along with the advancing
possibilities of exploiting metaproteomics from
microbial communities (Wang et al. 2016b),
promises deep insight into, and large expansion
of our knowledge of fungal biosynthetic pathways.
284
M. Flak et al.