Iterative l-Tryptophan Methylation in Psilocybe Evolved by Subdomain Duplication

Article abstract of DOI:10.1002/cbic.201800336

Psilocybe mushrooms are best known for their l-tryptophan-derived psychotropic alkaloid psilocybin. Dimethylation of norbaeocystin, the precursor of psilocybin, by the enzyme PsiM is a critical step during the biosynthesis of psilocybin. However, the “magic” mushroom Psilocybe serbica also mono- and dimethylates l-tryptophan, which is incompatible with the specificity of PsiM. Here, a second methyltransferase, TrpM, was identified and functionally characterized. Mono- and dimethylation activity on l-tryptophan was reconstituted in vitro, whereas tryptamine was rejected as a substrate. Therefore, we describe a second l-tryptophan-dependent pathway in Psilocybe that is not part of the biosynthesis of psilocybin. TrpM is unrelated to PsiM but originates from a retained ancient duplication event of a portion of the egtDB gene that encodes an ergothioneine biosynthesis enzyme. During mushroom evolution, this duplicated gene was widely lost but re-evolved sporadically and independently in various genera. We propose a new secondary metabolism evolvability mechanism, in which weakly selected genes are retained through preservation in a widely distributed, conserved pathway.

Full text of DOI:10.1002/cbic.201800336

Accepted Article
Title: Iterative L-tryptophan methylation in Psilocybe evolved by sub-
domain duplication
Authors: Felix Blei, Janis Fricke, Jonas Wick, Jason Slot, and Dirk
Hoffmeister
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201800336
Link to VoR: http://dx.doi.org/10.1002/cbic.201800336
A Journal of
www.chembiochem.org

10.1002/cbic.201800336
ChemBioChem
COMMUNICATION
Iterative -tryptophan methylation in Psilocybe evolved by sub-
L
domain duplication
Felix Blei,[a] Janis Fricke,[a] Jonas Wick,[a] Jason C. Slot,*[b] and Dirk Hoffmeister*[a]
PsiH, the kinase PsiK, and the N-methyltransferase PsiM
(Scheme 1). Importantly, The preference of PsiM for
norbaeocystin (Figure 1), i.e., substrate with 4-
phosphoryloxy group, places this enzyme at the end of the 1
biosynthetic cascade.[4a,4c] However, during mass spectrometric
analyses of P. serbica extracts, we surprisingly found signals
a
a
Abstract: Psilocybe mushrooms are best known for their
L-
tryptophan-derived psychotropic alkaloid psilocybin. The
dimethylation of its precursor norbaeocystin by the enzyme PsiM is a
critical step during its biosynthesis. However, the “magic” mushroom
that were consistent with the masses of
L-abrine (=N-α-methyl-L-
Psilocybe serbica also mono- and dimethylates
is incompatible with the specificity of PsiM. Here,
methyltransferase, TrpM, was identified and functionally
characterized. Mono- and dimethylation activity on -tryptophan was
reconstituted in vitro, while tryptamine was rejected as substrate.
Therefore, we describe a second -tryptophan-dependent pathway
L
-tryptophan, which
tryptophan, 4) and N,N-α-dimethyl- -tryptophan (5), respectively
L
a
second
(Figure 2), which are known natural products, e.g., from the
jequirity bean (Abrus precatorius).[5] Since PsiM does not accept
3 as substrate, occurrence of 4 and 5 in P. serbica indicates a
PsiM-independent methylation by an as yet unknown enzyme.
L
L
in Psilocybe which is not part of the psilocybin biosynthesis. TrpM is
R1=O-PO3H2, R2=N(CH3)2
R1=OH, R2=N(CH3)2
R1=O-PO3H2, R2=NH2
R1=H, R2=NH2
1
psilocybin ( ):
unrelated to PsiM, but originates from a retained ancient duplication
R1
2
psilocin ( ):
event of
a portion of the egtDB gene which encodes an
R2
norbaeocystin:
tryptamine:
ergothioneine biosynthesis enzyme. During mushroom evolution,
this duplicated gene was widely lost, but re-evolved sporadically and
independently in various genera. We propose a new secondary
metabolism evolvability mechanism, in which weakly selected genes
are retained through preservation in a widely distributed, conserved
pathway.
N
H
N,N-dimethyl-
R1=H, R2=N(CH3)2
R1=OH, R2=NH2
tryptamine:
4-hydroxy-
tryptamine:
R1=H, R2=NH2
-tryptophan (3):
L
L
COOH
R2
R1=H, R2=NHCH3
R1
4
-abrine ( ):
N,N-dimethyl-
-tryptophan ( ):
The fungal genus Psilocybe comprises numerous mushroom
species that biosynthesize tryptamine alkaloids, primarily
psilocybin (4-phosphoryloxy-N,N-dimethyltryptamine, 1, Figure
1).[1] It represents a prodrug to its dephosphorylated congener
psilocin (2), which agonistically targets serotonin receptors in the
central nervous system.[2] The colloquial term “magic
mushrooms” for these fungi alludes to their psychotropic effects
and to their use as recreational drug. However, clinical trials
have recognized 1 as a valuable candidate to be developed into
a medication against cancer-related anxiety and treatment-
resistant depression.[3]
R1=H, R2=N(CH3)2
R1=OH, R2=NH2
5
L
4-hydroxy-
L
N
H
-tryptophan:
-hypaphorine ( ):
1
2
+
6
L
R =H, R =N (CH3)3
COOH
7
ergothioneine ( )
CH3
CH3
N
HN
CH3
N
H
S
Figure 1. Structures of indole alkaloids isolated from Psilocybe and other
fungal species, products of the methyltransferase TrpM, and of ergothioneine.
Relying on Psilocybe cubensis, P. cyanescens, and other
species, evolutionary, ecological, and biochemical aspects of 1
biosynthesis have been investigated by our laboratories.[4]
Evolutionarily, 1 production has been advantageous to alter the
behaviour of mycophagous and wood-inhabiting insects that
We here report on the functional and phylogenetic
characterization of TrpM, an S-adenosyl-L-methionine (SAM)-
dependent methyltransferase of P. serbica. This enzyme does
not take part in 1 biosynthesis, but catalyzes a second pathway
that originates from 3 and includes formation of a tertiary amine
by iterative N-methylation.
share their habitat with the fungus. For its biosynthesis,
L-
tryptophan (3) is fed into a sequential four-step pathway,
catalyzed by the decarboxylase PsiD, the P450 monooxygenase
PsiM-independent
3 methylation in Psilocybe mushrooms
[a]
[b]
F. Blei, J. Fricke, J. Wick, Prof. Dr. D. Hoffmeister
Department Pharmaceutical Microbiology at the Hans Knöll Institute
Friedrich-Schiller-Universität
Beutenbergstrasse 11a, 07745 Jena (Germany)
E-mail: dirk.hoffmeister@leibniz-hki.de
Prof. Jason C. Slot, Ph.D.
appeared plausible, given the report on EgtD, a mycobacterial
enzyme that features a MT_33 methyltransferase superfamily
domain. EgtD is required for ergothioneine (7) biosynthesis and
processively trimethylates L
-histidine.[6] 7 occurs in various fungi
Department of Plant Pathology
Ohio State University
2021 Coffey Road
Columbus, OH 43210 (USA)
(and bacteria and plants alike) and is hypothesized to participate
in cellular redox processes.[6a] Specifically, crystallographic work
and subsequent targeted active site engineering of EgtD
(EgtDE282A,M252V) to emulate a fungal active site converted this
enzyme into a 3 methyltransferase. Combined with surveys on
fungal genomic data, this finding suggested that mushrooms use
E-mail: slot.1@osu.edu
Supporting information for this article is given via a link at the end of
the document.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800336
ChemBioChem
COMMUNICATION
EgtD-like methyltransferases to produce
L
-hypaphorine (N,N,N-
encoded within the 1 biosynthesis gene cluster. Also, TrpM and
PsiM are phylogenetically unrelated, as the former is a member
of the methyltransferase family 33 whereas the latter belongs to
family 10. The trpM cDNA was sequenced to confirm the
predicted introns, while a synthetic codon-optimized gene was
used to create expression plasmid pFB06. N-terminally tagged
hexahistidine TrpM fusion protein was produced in E. coli KRX
and purified by immobilized metal chelate affinity
chromatography (Figure S1).
α-trimethyl-L
-tryptophan, 6, Figure 1).[6a] This fungal product is
ecologically relevant as an agonist of the phytohormone indole-
3-acetic acid, which is e.g., secreted by the tree partner in
ectomycorrhizal symbioses.[7]
Figure 2. Mass spectrometric analysis of P. serbica mycelial extracts. Left
panel: extracted ion chromatogram for m/z 219-220. At tR=3.6 min, the mass
of monomethylated 3, e.g., L
-abrine (4, calculated 219.1129 [M+H]+, found
219.1124) was detected. Right panel: extracted ion chromatogram for m/z
233-234. Only the signal at tR=3.9 min (m/z 233.1285 [M+H]+) is consistent
with the exact mass of dimethylated 3, e.g., N,N-α-dimethyl-L-tryptophan (5,
calculated m/z 233.1284 [M+H]+). Ion chromatograms of 4 and 5 standards are
shown below the sample traces.
Gel permeation chromatography confirmed that TrpM is a
monomer under native conditions (Figure S2), which is
consistent with previous findings by Seebeck and colleagues on
EgtD.[8] Subsequently, TrpM was tested for activity in vitro.
Hypothesizing that it catalyzes 3 methylation, we first used this
substrate to test the activity of the recombinantly produced
enzyme. Assays were run in TRIS-buffer (pH=8.0) for 15 min.
For a time course of the reaction, samples were taken in 30 s
intervals over 150 s, and subsequently analyzed by LC/MS
(Figure 3, Figure S3). The analysis showed simultaneous
formation of mono- and dimethylated 3 across this time course.
One product was identical to an authentic standard of 4
regarding its retention time (tR=16.6 min, Figure 3), UV/Vis
spectrum, and molecular mass (m/z 219.1134 [M+H]+;
calculated 219.1129 [M+H]+). The identity of the expected
second product (5, tR=15.6 min) was confirmed after purification
Scheme 1. Methyl transfer steps during biosynthetic pathways to 5 and 1 in
Psilocybe serbica. The parentheses and the dashed arrow indicate low
turnover to 6 by TrpM.
The genomic sequences of a North American P. cyanescens
isolate is published, as well as a Central European isolate of P.
serbica (which falls into P. cyanescens species complex in the
wide sense).[4] This sequence data was browsed for further
genes encoding small molecule methyltransferases. In the
European isolate (P. serbica FSU12416), two egtD-like genes
were identified: one was fused to an egtB homolog, which is the
expected gene for sulfoxide synthase required for putative 7
assembly. Fused egtDB genes for 7 biosynthesis are common in
fungi. The second gene encoded a non-fusion methyltransferase,
which we considered a candidate enzyme for 3 methylation.
Standalone egtD-like genes rarely exist in basidiomycetes.
Examples include HyoA,[6a] a putative 6 synthase of the
mushroom Dichomitus squalens which shares 37% identical
amino acids, and a hypothetical protein of Galerina marginata
(protein ID: KDR66884.1, 70% identical aa).
1
by H NMR spectroscopy (Figure S4). Km values for 3 and 4
were 3.7 µ
M and 1.0 µM, respectively, as determined with a
luminescence assay. The kcat was 3  10-2 s-1 and 2.9  10-2 s-1.
Low Km and kcat values are well documented for other small
molecule
EgtDE282A,M252V
methyltransferases,
e.g.,
for
EgtD
and
L-
[6a-c]
,
or the C-methyltransferase TylC3 for
mycarose biosynthesis.[9] The kcat values for the first and second
TrpM-catalyzed methylation virtually do not diverge, which points
to increased affinity of TrpM to the monomethylated intermediate,
as indicated by the respective Km values, which is also
consistent with findings for EgtD.
In P. serbica FSU12416, the 1237 bp long candidate gene trpM
is disrupted by three introns, as predicted by Augustus software.
The 1077 bp reading frame thus encodes a 358 aa protein with
a calculated pI=5.1 and a mass of 39.6 kDa. Notably, trpM is not
Under standard conditions, 6 formation was not observed. In
separate assays, S-adenosylhomocysteine (SAH) nucleosidase
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800336
ChemBioChem
COMMUNICATION
and adenine deaminase were added to irreversibly remove SAH
from the reaction. After prolonged incubation for 4 h, 6 was
detectable as minor product as well (tR=13.3 min), besides 4 and
(Figure S5). In mycelial extracts of P. serbica, mono- and
dimethylated
S6) while not even traces of methyl derivatives of
-tyrosine were detected.
L-phenylalanine was identified by LC/MS (Figure
L-histidine and
5 (Figure 3). 4-Hydroxy-L-tryptophan and
D-tryptophan did not
L
serve as substrate (Figure 3).
Seebeck et al. characterized the mycobacterial wild type EgtD
(that is specific for -histidine), the engineered variant
EgtDE282A,M252V (specific for 3), and Ybs, an Aspergillus
methyltransferase, specific for
-tyrosine.[6] Our in vitro results
indicate that TrpM is more flexible for acceptor substrates than
the above methyltransferases. Yet, our data also support the
notion that wild type EgtD is outstanding in that it efficiently
trimethylates, while EgtDE282A,M252V and TrpM virtually stop at the
dimethyl stage. Of note, the residues of EgtD known to interact
Psilocybe mushrooms control the PsiM-catalyzed methyl
transfer steps during 1 biosynthesis by requiring substrates that
have a 4-phosphoryloxy group. Consequently, tryptamine is not
accepted. This important property minimizes, or prevents
altogether, that the reactive and instable 2 is made. Otherwise, 2
could be formed from N,N-dimethyltryptamine which may
undergo subsequent 4-hydroxylation by the rather unspecific
monooxygenase PsiH. Here, we tested whether tryptamine was
a substrate for TrpM. Under the applied conditions, turnover was
not observed by chromatographic analysis, and not even traces
of N,N-dimethyltryptamine or of its N-monomethyl precursor
were detected by mass spectrometry (Figure 3, traces g and h).
4-Hydroxytryptamine was not turned over either. An important
feature of 1 biosynthesis is that it 2 formation is avoided. TrpM
does not participate in 1 biosynthesis, but its substrate specificity
prevents 2 formation as well, as the products of TrpM, 4, 5, and
6, are not intermediates or shunt products of the 1 biosynthetic
pathway. Both 1 and 5 biosynthesis originates from 3 (Scheme
1). However, cellular localization of the respective enzymes or
mutually exclusive gene expression, e.g., dependent on the
developmental stage, may prevent competition for 3 as principal
building block.
L
L
with the (methylated) α-amino group of
Asn166, are conserved in TrpM (Gly192 and Asn197). Asn 166
in EgtD is essential to make N,N-dimethyl- -histidine a 100-fold
stronger ligand than
-histidine.[6b] Crystallographic work would
L-histidine, Gly161 and
L
L
be warranted to elucidate the respective roles of active site
residues for TrpM.
PsiM and TrpM share a functional relationship. Therefore, we
hypothesized that the evolution of their respective pathways is
correlated. To address this question, and to understand the
evolutionary relationship between EgtD and 7 metabolism in
fungi, we conducted a phylogenetic analysis of homologs of
TrpM-coding genes in Agaricomycetes. A maximum likelihood
tree places trpM in a clade with 12 orthologs from diverse
agaricomycete species that is nested in a larger group of
enzymes from a larger diversity of Agaricomycetes (Figure 4A,
Figure S7). Homologs of trpM in this phylogeny are variable in
structure; while all orthologs of trpM code for
a single
MT_33/EasF (EgtD) domain, the rest of the tree is composed
mainly of EgtDB bifunctional proteins, with a C-terminally linked
YfmG-superfamily sulfoxide synthase (EgtB) domain. Based on
the topological congruence between the trpM clade and that of
its mostly EgtDB sister clade (Figure 4A), which includes the
trpM taxa, we infer that trpM originated by partial duplication of
egtDB at an intermediate stage of agaricomycete diversification.
The phylogeny also reveals at least ten recent (mostly species-
specific) duplications that resulted in taxonomically diverse EgtD
proteins. Since its origin by an ancient duplication, trpM has
been periodically lost and is no longer present in most
Agaricomycetes, and it is the only long-retained duplication in
the phylogeny, suggesting EgtD-only paralogs have been
instable during Agaricomycetes diversification. This instability is
exemplified by the recent loss of EgtD in P. cubensis and
pseudogenization in P. cyanescens, where it is truncated and
contains an in-frame stop codon (Figure S8).
Figure 3. LC/MS analyses of TrpM in vitro activity tests. HR-ESIMS spectra
were recorded in positive mode. Experimentally determined masses are
indicated. Top trace: overlaid separate chromatograms of standard
compounds. Trace a: TrpM reaction with 3. Trace b: TrpM reaction with
D-
tryptophan. Trace c: control with heat-treated TrpM. Trace d: prolonged TrpM
reaction with 3 in the presence of SAH nucleosidase and adenine deaminase.
Traces e and f: TrpM reaction with 4-Hydroxy-L-tryptophan and negative
control, respectively. Traces g and h: TrpM reaction with tryptamine and
negative control. Traces i and j: TrpM reaction with 4-´hydroxytryptamine and
negative control. The signal at tR=12.9 min corresponds to m/z 298.0974
[M+H]+, which is compatible with the mass of 5´-methylthioadenosine, a
breakdown product of SAM. Tram: tryptamine, trp: tryptophan.
The functions of convergently evolved EgtD are only partly
understood in fungi. The only other functionally characterized
enzyme in the phylogeny (Figure 4A) is the D. squalens EgtD
previously predicted to produce 6,[6a] that is the product of a
recent partial duplication of EgtDB. Pisolithus tinctorius egtD
genes, which may also be involved in the production of 6, are
distantly related homologs, analyzed separately (Figure S9A),
while genes for EgtDB from P. tinctorius and other Boletales are
present, along with four recent egtD paralogs from Scleroderma
citrina. The gene for EasF is another distant egtD homolog in the
Given the activity of EgtD,[6] we further tested if
converted by TrpM in vitro as well. ESI-MS showed mono- and
dimethylation whereas trimethylation was not found. -tyrosine
and -phenylalanine were also mono- and dimethylated in vitro
L-histidine was
L
L
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800336
ChemBioChem
COMMUNICATION
Claviceps purpurea (ergot fungus, Ascomycota) ergotamine
biosynthesis cluster, which is more similar to Ascomycota egtDB
sequences than to egtDB sequences presented here (Figure
S9B).
The evolutionary instability of EgtD is consistent with a transition
from a broadly selected function in 7 biosynthesis[12] to one in
secondary metabolism, which tends to be conditionally retained
through diversifying ecological selection. In keeping with
functional diversification of EgtD in Agaricomycetes, the
substrate specificity-determining amino acid position 282
(relative to Mycobacterium smegmatis) is highly conserved as
glutamic acid for histidine binding in 97 EgtDB sequences, but is
variable as mainly Met, Thr, Glu, Ala, and Ser in 25 EgtD
sequences (Figure 4B). Similarly, position 252 is highly
conserved as alanine in EgtDB, but is variable as mainly Ala,
Met, Gly, and Ser in EgtD.
The phylogenetic distribution of gene families indicates
ecological functions.[10] EgtD is important for mycorrhizal
formation aided by 6 in P. tinctorius,[7a] which has four paralogs
of a very distantly related egtD, all of which are co-orthologous
with just one EgtD encoded in Scleroderma citrinum (Figure
S9A). If the parallel recent expansion of egtD in S. citrinum
(Figure 4A) represents convergent ectomycorrhiza-related
metabolism, this is in contrast to the broader distribution of egtD
homologs. The majority of Agaricomycetes with orthologs of
trpM and/or recent egtD paralogs shown here are involved in
lignocellulose decay, often of exposed substrates, which
suggests their products are involved in competition with
invertebrates or bacteria. It is also interesting to note that EgtD
in the Saprolegniales (oomycetes) appears to have originated by
partial horizontal transfer of an EgtDB from aquatic true fungi
(Figure S9C).
It is also plausible that fitness costs to the organism account in
part for the instability of recently evolved EgtD, by detrimentally
interfering with 7 synthesis due to the irreversibility of histidine
N-methylation; this hypothesis should ultimately be tested in a
genetically tractable model fungus. Gene duplication has long
been identified as a source of genetic novelty,[13] and has
strongly influenced the metabolic diversity of fungi through the
sub- and neofunctionalization of paralogs.[14] Further, gene
fissions are important sources of novel functions in fungi and
oomycetes.[15] We have characterized the function of TrpM, a
methyltransferase generated by duplication of the first domain of
a presumed 7 synthesis bifunctional enzyme, EgtDB. We further
demonstrated several convergent origins of EgtD in
Agaricomycetes by the same mechanism. Filamentous
Ascomycota maintain their EgtD enzymes encoded in diverse
secondary metabolism gene clusters, which can be specialized
and retained in large pan-genomes, and acquired by horizontal
gene transfer (HGT).[16] However, our data indicate that
Basidiomycota benefit more from de novo neo-functionalization
of vertically-inherited enzymes because of lower rates of
clustering and HGT.
Our results highlight
mechanism, in which weakly or sporadically selected functions
remain accessible through their preservation in widely
a secondary metabolism evolvability
a
distributed, structurally conserved bifunctional protein. Selection
on within-protein interactions and broadly conserved functions
possibly maintain these domains as a source of secondary
metabolic diversity, where they otherwise are lost in transiently
adaptive roles. In short, genes for these highly conserved
bifunctional proteins may function as genetic templates from
which less conserved functions can repeatedly evolve.
Traditionally, the fungal genus Psilocybe impacted natural
product chemistry due to its capacity to biosynthesize 1. Our
results demonstrate that the relevance of this genus reaches
further and help address more fundamental aspects of the
evolution of fungal small molecule-processing enzymes.
Figure 4. Evolutionary origin of TrpM and related enzymes. TrpM originated
by partial duplication of a putative ergothioneine (7) bifunctional enzyme
(EgtDB). A) Maximum Likelihood phylogeny of MT_33 Superfamily domains
extracted from nearest PsiM homologs in Agaricomycetes (phylogeny with
accession numbers in Figure S7A). EgtD enzymes (green branches) emerged
once anciently (TrpM), and at least 10 times recently by partial duplication of
bifunctional putative
7 synthesis enzyme EgtDB. There are few TrpM
monofunctional proteins surviving since the ancient duplication in
Agaricomycetes, and it has been lost in P. cubensis and pseudogenized in P.
cyanescens. The P. cyanescens pseudogene (ψ) is shown for illustration, but
was not included in phylogenetic analysis. A reversed phylogeny of EgtDB
enzymes from genomes containing TrpM (see also Figure S7B) demonstrates
topological congruence between the alternate paralogs. Species phylogeny in
Experimental Section
the key follows Hibbett et al.[11]
. B) Motifs associated with specificity-
General and microbiological procedures: P. serbica FSU12416 was
grown at 25°C in the dark on malt extract peptone (MEP) agar plates or
in MEP liquid medium shaken at 140 rpm for 20 d at room temperature.
Plasmid isolation, DNA restriction and ligation followed the instructions of
the manufacturers of kits and enzymes (NEB, Promega, Thermo, Zymo).
determining residues (*) 252 (left) and 282 (right) of Agaricomycetes EgtD
domains/proteins. Position number is relative to M. smegmatis EgtD
(A0R5M8.1).
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800336
ChemBioChem
COMMUNICATION
Chemicals, media ingredients and solvents were purchased from Roth,
Sigma-Aldrich, VWR, and Deutero.
µm particle size and the following conditions were used: initial hold at 5%
B for 1 min, linear gradient to 100% B within 15 min at a flow rate of 1 mL
min-1. For other analyses, a Betasil C18 column (150  2.1 mm, 3 µm
particle size) and these conditions were used: initial hold at 5% B for 1
min, a linear gradient of 5-98% B within 15 min at a flow of 0.2 mL min-1.
Cloning of trpM cDNA and construction of expression plasmid:
mRNA was isolated from P. serbica mycelium using Promega’s SV Total
RNA isolation kit. First strand cDNA synthesis was primed with oligo-
d(T)-primers and RevertAid reverse transcriptase. The first strand
Chemical analysis of P. serbica extracts: To prepare mycelium for
LC/HRESIMS-analyses, it was collected from liquid cultures by filtration,
washed with distilled water, blotted dry, snap-frozen in liquid nitrogen,
homogenized with mortar and pestle, and shaken in 200 mL MeOH for
18 h at 240 rpm, before the biomass was removed by filtration and dried
over sodium sulfate. Residual solvent was evaporated under reduced
pressure. The dry residue was dissolved in MeOH, filtered, and subjected
to LC/HRESIMS measurements (above).
reaction served as template in a subsequent PCR with primers (0.2 μ
each) oFB51 (5´-ATGCCGCGAATCCAGGTT-3´) and oFB52 (5´-
TTAGCTTCGTCCGGTTACTTCG-3´). The reaction contained 0.2 m
(each) deoxynucleoside triphosphate, 2 m MgSO4, in HF buffer that was
M
M
M
supplied with the enzyme (Phusion DNA polymerase, 1 U). Amplified
DNA was purified by agarose gel electrophoresis and subsequent
extraction from the gel. The PCR amplicon was ligated to pJET1.2 and
several clones were sequenced to confirm predicted exon/intron
junctions. For heterologous production of TrpM, E. coli KRX (Promega)
and a codon-optimized gene (Invitrogen) were used that was inserted
between the BamHI and XhoI sites of pET28a, to create expression
plasmid pFB06. The codon-optimized sequence is provided in Figure
S10, the genomic DNA sequence of trpM is deposited at GenBank under
accession number MH423322.
Purification and NMR spectroscopy: From in vitro assays, 5 was
purified in two steps on the above instrument. The first step was
accomplished with an Eclipse XDB-C18 column (250 × 9.4 mm, 5 µm
particle size). H2O + 0.1% TFA was solvent A and MeOH solvent B. A
linear gradient (2 mL min-1) was applied, beginning with 5% B, increased
to 15% in 5 min, to 37% B within 18 min, and to 100% B within 1 min.
The second purification step included a Hypercarb column (150 × 4.6 mm,
5 µm), 0.1% TFA in H2O as solvent A and ACN as solvent B. A linear
gradient (1 mL min-1) was applied, beginning with 21% B and increased
Protein purification: Cultivation of E. coli and gene expression were
performed as described.[17] TrpM was purified by metal chelate affinity
chromatography on Protino Ni2+-NTA agarose (Macherey-Nagel).
Purification was verified by polyacrylamide gels (12% Laemmli gel), and
protein concentrations were determined by Bradford’s method.[18] TrpM
was desalted on a PD-10 column (GE Healthcare), equilibrated with
1
to 100% B in 9 min. The H NMR spectrum of 5 was recorded at 300 K
on a Bruker Avance III spectrometer at 500 MHz in D2O. Chemical shifts
were referenced to residual non-deuterated solvent (δH 4.79 ppm).
Bioinformatic analysis: Initially, Blast[19] and Augustus (v.3.3.1)[20]
software was used to browse the P. serbica genome.[4a] Homologs of
trpM were obtained from a local database of 540 fungal and oomycete
proteomes using usearch (v. 8.0.1517)[21] with TrpM (KDR66884.1) and
EgtDB (KDR82860.1) from G. marginata as queries, an e-value of 1e-3,
and protein identity of 30%. Domain structure of proteins was determined
using rpsblast (NCBI BLAST v. 2.6.0+)[19] to search the NCBI Conserved
Domain Database (v. 3.16)[22] with an e-value of 0.001. Absence of YfmG
superfamily domains in proteins was confirmed by tblastn against the
respective genome assemblies using the nearest EgtDB sequence from
phylogenies (see below) as a query. MT_33 superfamily domains were
extracted from protein sequences using hmmer (v. 3.1b2).[23] MT_33
phylogenies were constructed by first aligning with mafft (v. 7.221),[24]
removing ambiguously aligned characters with Trimal (v. 1.4)[25] using the
-automated1 method. Exploratory phylogenetic analyses were performed
using fasttree (v. 2.1.7).[26] A well-supported clade consisting exclusively
of Agaricomycetes was selected for subsequent analyses. For analyses
of distant relatives of TrpM, hmmer was used to search for MT_33
domains (952 total) across the proteome database. Figure S9A (P.
tinctorius EgtD) combined the same four additional Agaricomycetes
rooting sequences along with near and distant clades of Boletales
EgtDB/EgtD. Figure S9B (EasF) combined two clades of Pezizomycotina
MT_33 homologs with two orthologs each encoding Agaricomycetes
TrpM and EgtDB for rooting. Figure S9C (oomycetes EgtD) combines a
clade of early diverging fungi that includes oomycete EgtD sequences
and the Agaricomycetes root. Datasets were re-aligned and curated as
above, followed by maximum likelihood analysis using RAxML (v.
8.2.9)[27] with the model (LG) selected automatically according to
Bayesian information criterion. Protein domains identified by rpsblast
were mapped to the Agaricomycetes PsiM2/EgtDB phylogeny using the
ETE Toolkit (v. 3.1.1).[28]
buffer (50 m
M TRIS-HCl, pH 8.0) for enzyme assays. Production of E.
coli S-adenosylhomocysteine nucleosidase and adenine deaminase[6]
using pET28-based expression plasmids followed the same protocol,
except that the E. coli culture to express the adenine deaminase gene
was supplemented with 50 µ
exclusion chromatography, TrpM was eluted in phosphate buffer (10 m
sodium dihydrogen phosphate, 140 m NaCl, pH 7.4). Size-exclusion
chromatography was performed by FPLC (Äkta Pure 25, GE Healthcare)
and a Superdex 200 increase 10/300 GL column with 24 mL bed volume.
Signals were referenced to the GE Healthcare high molecular weight
standard.
M 2,2’-bipyridyl and 1 mM MnCl2. For size-
M
M
Product formation by TrpM: In vitro TrpM assays were carried out in
triplicates in a volume of 200 µL, buffered in 50 m
the enzyme concentration was 200 n . The methyl acceptor (3 or others)
was added at 500 µ final concentration, SAM at 1.5 m . The reactions
were incubated at 25°C for 15 min. The time course with 3 as acceptor
substrate included additional sampling in 30 s intervals from 0-150 s. The
reaction with SAH-nucleosidase and adenine deaminase was incubated
for 4 h. Reactions were stopped in liquid nitrogen and then lyophilized,
dissolved in MeOH and centrifuged. The supernatant was collected for
chemical analysis. TrpM kinetics were recorded using the MTase-Glo
luminescence assay (Promega).
M TRIS-HCl, pH=8.0,
M
M
M
Chemical analysis of TrpM assays: LC/ESIMS-analyses of TrpM in
vitro assays were carried out on an Agilent Infinity 1260 instrument with a
Zorbax Eclipse XDB-C18 column (150 × 4.6 mm, 5 µm particle size, flow:
1 mL min-1), coupled to an Agilent 6130 Single Quadrupole mass
detector. Diode array detection was between λ=200–400 nm.
Chromatograms were extracted at λ=280 nm. To analyze indolic
compounds, the following linear gradient was applied with 0.1%
trifluoroacetic acid (TFA) in H2O (solvent A) and methanol (MeOH,
solvent B): initially 5% B, increased to 15% within 5 min, to 37% B within
further 13 min, and to 100% B within 1 min. LC/HRESIMS was performed
on a Thermo Accela liquid chromatograph coupled to an Exactive
Orbitrap spectrometer operated in positive and negative mode and
equipped with a C18 column. Solvent A was 0.1% (v/v) formic acid in
H2O, solvent B was acetonitrile (ACN).To analyze mycelial extracts for
Acknowledgments
We thank A. Perner H. Heinecke, F. Baldeweg, and H. Schoeler
(Hans Knöll Institute Jena) for recording high-resolution mass
methylated L-phenylalanine, a Grom-Sil 100 ODS-0 AB, 250  4.6 mm, 3
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800336
ChemBioChem
COMMUNICATION
Beliaeva W. Blankenfeldt, F.P. Seebeck, ACS Chem. Biol. 2018, 13,
1333-1342; c) F.P. Seebeck, J. Am. Chem. Soc. 2010, 132, 6632–
6633.
spectra, for recording and interpreting NMR spectra, and for
luminescence measurements, respectively. M. Gressler
(Friedrich Schiller University Jena) is acknowledged for helpful
comments on the manuscript. We also thank F.J. Ruiz-Dueñas,
Ph.D. (Consejo Superior de Investigaciones Científicas, Madrid,
Spain), J.M. Barrasa, Ph.D. (Universidad de Alcalá, Alcalá de
Henares, Spain), D.S. Hibbett, Ph.D. (Clark University,
Worcester, MA, USA), D. Lindner, Ph.D. (US Dept. of Agriculture,
Forest Service, Madison, WI, USA), and L. Nagy, Ph.D.
(Hungarian Academy of Sciences, Szeged, Hungary), for access
to unpublished genome sequences. These data were produced
by the US Department of Energy Joint Genome Institute in
collaboration with the user community. Evolutionary analyses
were performed using the resources of the Ohio Supercomputer
Center. This work was supported by the National Science
Foundation grant (grant DEB-1638999 to JCS) and by the
Deutsche Forschungsgemeinschaft (DFG, grant HO2515/7-1 to
DH). Work in DH’s laboratory is also funded by the Collaborative
Research Center ChemBioSys (SFB1127).
[7]
a) T. Kawano, Plant Cell Rep. 2003, 21, 829-837; b) A. Jambois, A.
Dauphin, T. Kawano, F.A. Ditengou, F. Bouteau, V. Legue, F. Lapeyrie,
Physiol. Plantarum 2005, 123, 120-129.
[8]
[9]
A. Vit, L. Misson, W. Blankenfeldt, F.P. Seebeck, Acta Crystallogr. F
2014, 70, 676-680.
H. Chen, Z. Zhao, T.M. Hallis, Z. Guo, H.-w. Liu, J. Am. Chem. Soc.
2001, 40, 607-610.
[10] D. Floudas, M. Binder, R. Riley, K. Barry, R.A. Blanchette, B. Henrissat,
et al. Science, 2012, 336, 1715-1719.
[11] D.S. Hibbett, M. Binder, J.F. Bischoff, M. Blackwell, P.F. Cannon, O.E.
Eriksson, et al., Mycol. Res. 2007, 111, 509-547.
[12] G.W. Jones, S. Doyle, D.A. Fitzpatrick, Gene, 2014, 549, 161-170.
[13] S. Ohno, Evolution by gene duplication, 1st ed., Springer-Verlag,
Heidelberg, 1970.
[14] a) I. Wapinski, A. Pfeffer, N. Friedman, A. Regev, Nature, 2007, 449,
54-61; b) J.H. Wisecaver, J.C. Slot, A. Rokas, PLoS Genetics, 2014, 10,
e1004816.
[15] G. Leonard, T.A. Richards, Proc. Natl. Acad. Sci., 2012, 109, 21402-
21407.
[16] M. Marcet-Houben, T. Gabaldón, Fungal Genet. Biol., 2016, 86, 71-80.
[17] D. Kalb, T. Heinekamp, S. Schieferdecker, M. Nett, A. A. Brakhage, D.
Hoffmeister, ChemBioChem 2016, 17, 1813-1817.
Keywords: biosynthesis • enzymes • methyltransferase •
psilocybin • tryptophan
[18] M. M. Bradford, Anal. Biochem. 1976, 72, 248-254.
[19] S. F. Altschul, T.L. Madden, A.A. Schäffer, J. Zhang, Z. Zhang, W.
Miller, D.J. Lipman, Nucleic Acids Res. 1997, 25, 3389-3402.
[20] M. Stanke, R. Steinkamp, S. Waack, B. Morgenstern, Nucleic Acids
Res. 2004, 32, W309-312.
[1]
[2]
[3]
a) A. Hofmann, R. Heim, A. Brack, H. Kobel, A. Frey, H. Ott, T.
Petrzilka, F. Troxler, Helv. Chim. Acta 1959, 42, 1557-1572; b) A.Y.
Leung, A.H. Smith, A.G. Paul, J. Pharm. Sci. 1965, 54, 1576-1579.
a) D.E. Nichols, Pharmacol. Ther. 2004, 101, 131-181; b) F. Hasler, U.
Grimberg, M.A. Benz, T. Huber, F.X. Vollenweider, Psychopharmacol.
2004, 172, 145-156.
[21] R.C. Edgar, Bioinformatics, 2010, 26, 2460-2461.
[22] A. Marchler-Bauer, M.K. Derbyshire, M.R. Gonzales, S. Lu, F. Chitsaz,
L.Y. Geer, et al., Nucleic Acids Res., 2015, 43, D222-226.
[23] S.R. Eddy, Bioinformatics, 1998, 14, 755-763.
[24] K. Katoh, D.M. Standley, Mol. Biol. Evol. 2013, 30, 772-780.
[25] S. Capella-Gutierrez, J.M. Silla-Martinez, T. Gabaldón, Bioinformatics,
25, 2009, 1972-1973.
a) C.S. Grob, A.L. Danforth, G.S. Chopra, M. Hagerty, C.R. McKay, A.L.
Halberstadt, G.R. Greer, Arch. Gen. Psychiat. 2011, 68, 71-78; b) M.W.
Johnson, A. Garcia-Romeu, M.P. Cosimano, R.R. Griffiths, J.
Psychopharmacol. 2014, 28, 983-992; c) A.M. Sherwood, T.E.
Prisinzano, Exp. Rev. Clin. Pharmacol. 2018, 11, 1-3.
[26] M.N. Price, P.S. Dehal, A.P. Arkin, PLoS One, 2010, 5, e9490.
[27] A. Stamatakis, Bioinformatics, 2014, 30, 1312-1313.
[28] J.F. Huerta-Cepas, Serra, P. Bork, Mol. Biol. Evol., 2016, 33, 1635-
1638.
[4]
a) J. Fricke, F. Blei, D. Hoffmeister, Angew. Chem. Intl. Ed. 2017, 56,
12352-12355; b) H.T. Reynolds, V. Vijayakumar, E. Gluck-Thaler, H.B.
Korotkin, P.B. Matheny, J.C. Slot, Evol. Lett. (doi: 10.1002/evl3.42); c)
F. Blei, F. Baldeweg, J. Fricke, D. Hoffmeister, Chem. Eur. J. 2018, 24,
10028-10031; d) J. Fricke, F. Blei, D. Hoffmeister, Angew. Chem. 2017,
129, 12524-12527.
[5]
[6]
S. Ghosal, S.K. Dutta, Phytochemistry 1971, 10, 195-198.
a) A. Vit, L. Misson, W. Blankenfeldt, F. P. Seebeck, ChemBioChem
2015, 16, 119-125; b) L. Misson, R. Burn, A. Vit, J. Hildesheim, M.A.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800336
ChemBioChem
COMMUNICATION
Entry for the Table of Contents
COMMUNICATION
Magic evolution - The
Felix Blei, Janis Fricke, Jonas
Wick, Jason C. Slot*, and Dirk
Hoffmeister*
methyltransferase TrpM of the
“magic” mushroom and psilocybin
producer Psilocybe serbica was
characterized. TrpM catalyzes
Page No. – Page No.
iterative L-tryptophan N-
methylation, its specificity profile
prevents interference with the
psilocybin pathway. The trpM
gene evolved from an
Iterative L-tryptophan
methylation in Psilocybe
evolved by sub-domain
duplication
ergothioneine biosynthetic gene.
Its phylogenetic history provides
insight how mushrooms evolve
natural product biosyntheses.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.