Structure and Catalytic Mechanism of a Bacterial Friedel–Crafts Acylase

Article abstract of DOI:10.1002/cbic.201800462

C?C bond-forming reactions are key transformations for setting up the carbon frameworks of organic compounds. In this context, Friedel–Crafts acylation is commonly used for the synthesis of aryl ketones, which are common motifs in many fine chemicals and natural products. A bacterial multicomponent acyltransferase from Pseudomonas protegens (PpATase) catalyzes such Friedel–Crafts C-acylation of phenolic substrates in aqueous solution, reaching up to >99 % conversion without the need for CoA-activated reagents. We determined X-ray crystal structures of the native and ligand-bound complexes. This multimeric enzyme consists of three subunits: PhlA, PhlB, and PhlC, arranged in a Phl(A₂C₂)₂B₄ composition. The structure of a reaction intermediate obtained from crystals soaked with the natural substrate 1-(2,4,6-trihydroxyphenyl)ethanone together with site-directed mutagenesis studies revealed that only residues from the PhlC subunits are involved in the acyl transfer reaction, with Cys88 very likely playing a significant role during catalysis. These structural and mechanistic insights form the basis of further enzyme engineering efforts directed towards enhancing the substrate scope of this enzyme.

Full text of DOI:10.1002/cbic.201800462

Accepted Article
Title: Structure and catalytic mechanism of a bacterial Friedel-Crafts
acylase
Authors: Tea Pavkov-Keller, Nina G. Schmidt, Anna Żądło-
Dobrowolska, Wolfgang Kroutil, and Karl Gruber
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.201800462
Link to VoR: http://dx.doi.org/10.1002/cbic.201800462
A Journal of
www.chembiochem.org

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
Structure and catalytic mechanism of a bacterial Friedel-Crafts
acylase
Tea Pavkov-Keller,[a,b] Nina G. Schmidt,[a,c] Anna Żądło-Dobrowolska,[c] Wolfgang Kroutil,*[a,c,d] and Karl
Gruber*[a,b,d]
Abstract: C–C bond forming reactions are key transformations to
set up the carbon framework of organic molecules. In this context,
the Friedel-Crafts acylation is commonly used for the synthesis of
aryl ketones, which are common motifs in many fine chemicals and
natural products. A bacterial multi-component acyltransferase from
Pseudomonas protegens (PpATase) catalyzes such a Friedel-Crafts
C-acylation of phenolic substrates in aqueous solution reaching up
to >99% conversion without the need of CoA-activated reagents. We
determined x-ray crystal structures of the native and ligand-bound
complex. This multimeric enzyme consists of three subunits – PhlA,
PhlB and PhlC which are arranged in a Phl(A2C2)2B4 composition.
The structure of a reaction intermediate obtained from crystals
soaked with the natural substrate monoacetylphloroglucinol together
with site-directed mutagenesis studies revealed that only residues
from the PhlC subunits are involved in the acyl transfer reaction, with
Cys88 very likely playing a significant role during catalysis. These
structural and mechanistic insights form the basis of further enzyme
engineering efforts towards enhancing the substrate scope of this
enzyme.
numerous biosynthetic pathways, especially in the biosynthesis
of membrane phospholipids[9], wax esters and triacylglycerols[10]
and polyketides[11] and also play role in lysozyme resistance[12]
.
In contrast to the O- and N- acylation, natural C-acylation
reactions are scarce. So far, only a couple of acyltransferases,
e.g.
from
Pseudomonas
protegens
(PpATase)
and
Pseudomonas fluorescens have been reported to perform C–C
bond formation in the biosynthesis of the antibiotically active
polyketide
2,4-diacetylphloroglucinol
(DAPG)[13]
.
The
biosynthesis of DAPG is regulated by the phlACBDEFGHI gene
cluster, which is divided into regulatory genes phlEFGHI and the
biosynthetic operon phlACBD[14]. For many years, biocatalytic
applications were limited to gene phlD encoding a type-III
polyketide synthase which was employed for the in vivo
production of the DAPG-precursor PG in either Pseudomonas
sp. or E. coli under controlled conditions in bioreactors[15]
.
Very recently, the scope of applications was extended to phlACB,
which encodes the cofactor-independent acyltransferase
PpATase mentioned above. It catalyzes the reversible
disproportionation of monoacetylphloroglucinol (MAPG) into
phloroglucinol (PG) and DAPG in a divergent reaction (Scheme
1). PpATase is a multi-component enzyme and is catalytically
active without addition of cofactors like CoA or ATP. A functional
enzyme is only obtained upon expression of the entire phlABC
operon, whereas mixing and incubation of the individually
expressed proteins – PhlA, PhlB, and PhlC – did not lead to the
Introduction
Acyltransferases belong to the EC 2.3 groups of enzymes. This
subclass contains enzymes that transfer acyl groups from a
donor molecule to the hydroxyl, amino or thiol group of an
acceptor molecule, forming either esters, amides or thioesters.
Different acyltransferase classes have been identified among
living organisms that utilize for instance 1-O-acylglucosides[1],
acylated-acyl carrier proteins[2] or quinic acid ester[3] as high
energy activated acyl donors. The most abundant group of
acyltransferases requires mostly acyl-CoA derivative as a donor
substrate catalyzing various reactions involved in the primary
and secondary metabolism[4]. In bacteria, CoA-dependent O-
and N-acylation plays a key role in detoxification of antibiotics,
such as chloramphenicol[5], aminoglycosides[6], streptothricin[7]
and phosphinothricin[8]. Acyltransferases are also involved in
disproportionation of MAPG[13a, 16]
.
Scheme 1. Natural reaction catalyzed by the Pseudomonas protegens
acyltransferase (PpATase) involved in the biosynthesis of 2,4-
diacetylphloroglucinol (DAPG).
It was shown that PpATase accepts various C- or O-acyl donors
such as isopropenyl acetate and transfers an acetyl moiety to a
[a]
Dr. T. Pavkov-Keller, Dr. N. G. Schmidt, Prof. Dr. W. Kroutil, Prof.
Dr. K. Gruber
Austrian Centre of Industrial Biotechnology (ACIB)
Petersgasse 14, 8010 Graz, Austria
phenolic acceptor in a Friedel-Crafts-type acetylation reaction[16b,
17]
.
The enzyme complex also shows chemical reaction
promiscuity and accepts aniline derivatives as substrates for
amide formation[18]
[b]
[c]
Dr. T. Pavkov-Keller, Prof. Dr. K. Gruber
.
Institute of Molecular Biosciences, University of Graz
Humboldtstrasse 50, A-8010 Graz, Austria
Dr. N. G. Schmidt, Dr. A. Żądło-Dobrowolska, Prof. Dr. W. Kroutil
Department of Chemistry, Organic & Bioorganic Chemistry,
University of Graz
Heinrichstrasse 28/2, A-8010 Graz, Austria
Prof. Dr. W. Kroutil, Prof. Dr. K. Gruber
BioTechMed-Graz, Mozartgasse 12/II, 8010 Graz Austria
Here, we report on the crystal structure determination of
PpATase in its substrate-free form as well as after soaking the
crystals with MAPG. The structural results indicate that only one
of the three subunits (PhlC) is involved in catalyzing the
acylation reaction. Site-directed mutagenesis studies and activity
measurements complement the structural data and enable the
proposal of a catalytic mechanism.
[d]
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.201800462
ChemBioChem
FULL PAPER
A cavity analysis identified four, large, contiguous cavities per
dodecamer which are lined by residues from one particular copy
of PhlA, PhlB and PhlC. In this hetero-trimeric arrangement the
proposed active sites of PhlA and PhlC (see below) are adjacent
to each other on one side of the cavity, whereas PhlB (especially
through a long loop formed by residues Glu74 to Val87) closes
the cavity on the opposite side (Figure 1c).
Results and Discussion
Overall structure
We determined x-ray crystal structures of a CoA-independent
acyltransferase produced by Pseudomonas protegens DSM
19095 (PpATase) using diffraction data from two different crystal
forms at 2.8 and 3.4 Å resolution, respectively (Table 2 and
Supporting Information Figure 1). The structures were solved by
molecular replacement involving extensive density modification
as well as automated and manual rebuilding (see Methods). It
had been known before that PpATase is a multimeric enzyme
consisting of three subunits – PhlA, PhlB and PhlC[36]. The final
structures showed the hexagonal crystal form (space group
P6122) to contain two copies of each of those subunits and the
orthorhombic crystal form (space group P212121) to contain eight
copies of each subunit in the asymmetric unit, respectively.
In both structures, an analysis of protein-protein interactions
within the crystal using the EBI-Pisa server[19] yielded a hetero-
dodecamer with four copies of each subunit as the most likely
biologically active oligomer of PpATase. Thus, the hexagonal
crystal form contains half a dodecamer in the asymmetric unit
with the other half being generated by a crystallographic diad.
The orthogonal crystals, on the other hand, contain two copies
of the full dodecamer in the asymmetric unit. Closer inspection
of the inter-chain contacts indicated that PhlA and PhlC each
form strongly interacting homodimers in the crystal (Supporting
Information Figure 2). The composition of the multimeric enzyme
complex is thus best described as Phl(A2C2)2B4, in which the
four copies of PhlB mediate the binding of two PhlA and two
PhlC dimers (Figure 1b) to form the complete oligomer (Figure
1a and 1b).
Comparison with other structures
We used the Dali server[20] as well as PDBeFold[21] to identify
structurally similar proteins in the PDB. According to these
analyses, the closest structural homologs to PhlA are
hydroxymethylglutaryl-CoA synthases (e.g. the enzyme MvaS
from Myxococcus xanthus, PDB-entry: 5hwq)[22] and two β-
oxoacyl-(acyl-carrier-protein) synthases: from Aquifex aeolicus
(PDB-entry: 2ebd, RIKEN Structural Genomics/Proteomics
Initiative) and from E. coli (PDB-entry: 1hnh)[23]. The sequence
identity of PhlA with these homologs is below 24% with root-
mean-square-deviations (rmsd) of approximately 2 Å.
Previous sequence analyses had suggested that PhlA could be
involved in the first step of diacetyl-phloroglucinol (DAPG)
biosynthesis, i.e. the formation of acetoacetyl-CoA from acetyl-
CoA[24]. Inspections of the cavity observed in our structures in
the vicinity of PhlA suggests that there is indeed enough space
for acetyl-CoA binding. However, key residues necessary to
catalyze this Claisen-type condensation (a cysteine and a
glutamic acid) are not present in PhlA. Compared to
homologous enzymes, the crucial cysteine corresponds to
Gly115 in PhlA and the active site glutamate aligns with Cys83
in PhlA. We aimed at obtaining a complex structure by soaking
PpATase with acetyl-CoA, but the soaked crystal did not diffract
sufficiently well for structure determination.
Structural analysis of PhlB suggested a close similarity to a
protein of unknown function from Sulfolobus solfataricus (PDB-
entry: 3irb, rmsd 2.6 Å, seq-id: 18%)[25]. This protein from the
DUF35 family (Pfam PF01796) exhibits
a
two-domain
architecture comprising an N-terminal, rubredoxin-like zinc
ribbon and a C-terminal oligonucleotide/oligosaccharide-binding
(OB) fold domain[26] with an additional N-terminal helical
segment possibly involved in protein-protein interactions[25]
(Supporting information Figure 3). For members of this protein
family, a general role in fatty acid and polyketide biosynthesis as
acyl-CoA binding proteins has been predicted[25]
.
Although a similar domain organization is observed in PhlB and
the zinc ribbon with its Cys4 metal ion binding site is conserved,
severe differences are evident in the OB-domain and at the N-
terminus. Instead of two N-terminal helices, PhlB exhibits an
elongated N-terminal tail and a short, kinked a-helix. In the
dodecameric arrangement of PpATase, this region of PhlB (up
to Arg28) is not accessible to the solvent but buried within the
enzyme complex structure, where it is involved in tight
interactions with adjacent PhlA and PhlC molecules as well as
with the N-terminal tail of a neighboring PhlB molecule (Figure
1b). The same is true for the two long loop regions within the OB
domain (residue 73 to 89 and residues 127-136).
Figure 1. Structural analysis of PpATase. (a) Structure of the hetero-
dodecamer Phl(A2C2)2B4 (PhlA = blue and cyan, PhlB = magenta, PhlC = grey
and wheat). (b) PhlB mediates the binding of two PhlA and two PhlC dimers.
Two N-terminal tails of PhlB (up to Arg28) is involved in tight interactions with
adjacent PhlA and PhlC molecules as well as with the N-terminal tail of a
neighboring PhlB molecule. PhlB is shown in ribbon representation. (c) The
continuous cavity between PhlA, PhlB and PhlC subunit trimer. The active site
of PhlC (in violet, residues shown in stick representation) is situated adjacent
to the cavity of PhlA. The long loop in PhlB (Glu74-Val87, magenta) is closing
the cavity from the other side. The cavity is shown in surface representation
(red-hydrophobic to blue-hydrophilic). See Methods for detailed descriptions.
Finally, the sequence of PhlC shows that it belongs to the
thiolase superfamily[27]
,
although sequence identities to
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
members of this enzyme family are below 45%. The structure of
this subunit exhibits an α/β-hydrolase-type fold and is most
similar to structures of the thiolase-like protein ST0096 from
Sulfolobus tokodaii (PDB-entry: 4yzo, rmsd 1.6 Å, seq-id: 26%)
and the SCP2 thiolase from Trypanosoma brucei (PDB-entry:
4bi9, rmsd 1.7 Å, seq-id: 24%)[28]. In thiolases, a cysteine and a
histidine residue (corresponding to Cys88 and His347 in PhlC)
are highly conserved and were found to be important for the
enzymatic reaction. Beside those two residues, the active site
cavity of PhlC is lined by amino acid residues His56, Asn87,
His144, Trp211, Tyr298 and Ser349, which may play a role in
substrate binding or catalysis (Supporting Information Figure 7).
Even in the untreated (hexagonal) crystal, residual density was
observed at the sidechain of Cys88 indicating that this residue is
at least partially acetylated.
was solved at 3.4 Å resolution. Residual density was observed
in the vicinity of residue Cys88 in four out of the eight
crystallographically independent copies of the PhlC subunit
(Supporting Information Figure 10). This density was compatible
with an acetylated Cys88 and with
a
(deacetylated)
phloroglucinol molecule bound at this site (Figure 2). The
resulting structure revealed that only residues from PhlC interact
with the bound phloroglucinol. Specifically, the sidechains of
His56 as well as Tyr124 and His347 form hydrogen bonds with
the hydroxyl groups at positions 1 and 3 of the aromatic ring,
respectively (Figure 2). In contrast to that, the hydroxyl group at
position 5 is apparently not involved in any direct interactions
with PhlC. Instead, it points towards a small cavity lined by
mostly hydrophobic residues (such as Phe148, Leu209 and
Leu300).
Recently, the structure of an archaeal acetoacetyl-CoA
The carbonyl oxygen of the Cys88-bound acetyl group is
positioned in the “oxyanion hole” of the α/β-hydrolase-type fold
forming hydrogen bonds with the mainchain amide groups of
Cys88 itself and of Gly385. Additional polar stabilization may be
provided by the helix dipole oriented favorably towards Cys88
(the “nucleophile elbow” present in this type of fold[30]). The
acetyl group is oriented more or less parallel to the aromatic ring
and its carbonyl carbon atom is appropriately positioned for an
electrophilic attack on the C-6 atom of the substrate with
distances between the two atoms ranging from 3 to 3.5 Å
(Figure 2). Its methyl group points towards Phe148.
thiolase/HMG-CoA
synthase
(HMGCS)
complex
from
Methanothermococcus thermolithotrophicus was reported[29]
.
The arrangement of the individual protein subunits within this
complex is very similar to that in PpATase: the HMGCS subunit
corresponds to PhlA in PpATase (sequence identity 38.5%),
whereas the thiolase subunit corresponds to PhlC (sequence
identity 28.9%). The complex also contains a protein from the
DUF35 family that is related to PhlB in PpATase (sequence
identity 26.7%). Sequence alignments of the three separate
subunits are shown in Supporting Information Figures 4-6. In
contrast to the PpATase, however, this complex utilizes acetyl-
CoA as an acyl donor and is involved in the mevalonate pathway.
The HMGCS subunit is involved in the exergonic condensation
of
acetoacetyl-CoA
and acetyl-CoA
to
3-hydroxy-3-
methylglutaryl-CoA. The key Cys- and Glu-residues necessary
to catalyze this Claisen-type condensation are indeed present in
HMGCS (Cys114 and Glu82), whereas they are missing in PhlA
as mentioned above (Supporting Information Figure 8). The
presence of the active Cys in the thiolase subunit is preserved in
both complexes (Cys88 for PhlC and Cys85 in the thiolase
subunit). Soaking of the thiolase/HMGCS complex with acetyl-
CoA (PDB-entry: 6esq) revealed the binding of CoA at the
[29]
subunit interface comprising residues of all three proteins
.
This region is significantly different when compared with the
structure of PpATase (Supporting Information Figure 9) showing
altered relative positions of individual secondary structure
elements, the presence of additional residues in the PhlB region
of PpATase, two additional short b-strands in HMGCS as well as
a lack of CoA-binding residues in PpATase. The overall fold of
PhlB is very similar to the protein from the DUF35 family in the
thiolase/HMGCS complex (Supporting Information Figure 3).
The elongated N-terminus of PhlB that is buried within the
PpATase complex structure, however, is completely missing in
the other structure and notable conformational changes are
observed in the opposite loop region.
Figure 2. Active site structure of PhlC after soaking with MAPG. This depicts
the intermediate formed upon acetyl transfer from the substrate to Cys88 of
the enzyme (the transforming co-product PG and the transferred acetyl group
are depicted in cyan). Trp211 can act as a lid covering the entrance to the
active site (closed: grey; open: magenta). Hydrogen bonds are shown as
yellow dashed lines, the contact between the carbonyl carbon of the acetyl
group and C-6 of the substrate aromatic ring is indicated as a green dashed
line. The carbonyl oxygen of the Cys88-bound acetyl group is positioned in the
“oxyanion hole” of the α/β-hydrolase-type fold forming hydrogen bonds
(indicated in the light blue dashed lines) with the main chain amide groups.
Structure of the complex with MAPG
In order to identify the subunit responsible for the observed
transferase activity, we performed crystal-soaking experiments.
Orthorhombic crystals of PpATase were soaked with the native
acetyl donor/acceptor MAPG and the structure of the complex
Whereas the conformation of most residues forming the active
site of PhlC is similar in both structures, especially Trp211
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
stands out as it adopts an open conformation in the untreated
crystal and closes the cavity upon substrate binding in the
soaked crystal (Figure 2).
(entry #2), while the non-enzymatic reaction (entry #1) yields
only the O-acetylation product
3 under these conditions.
Replacing Cys88 by alanine or serine completely abolishes
ATase activity (entries #6 and #7). The same is true for
replacements of His56 (entries #3 and #4), whereas exchanges
of Asn87 (entry #5), His144 (entries #8 and #9), Tyr298 (entries
#12-#14), His347 (entry #15), Ser349 (entry #16) and Asp352
(entry #17) yielded variants which retained minute activities in
some cases. As anticipated, the exchange of Trp211 by alanine
or phenylalanine did not abolish the enzymatic activity of
PpATase (entries #10 and #11). The variant W211A, however,
significantly shifted the product spectrum in favor of the C-
acetylation product.
Mutagenesis of PhlC active site residues
The analysis of the PhlC active site identified Cys88, His144,
Asn87, His56, Ser349, Tyr124 Tyr298, Asp352, His347 and
Trp211 as a group of mostly polar or charged residues lining the
cavity (Supporting Information Figure 7). Those residues were
selected for site-directed mutagenesis and we analyzed the
activity of the corresponding enzyme variants in the acetylation
of resorcinol 1 using isopropenyl acetate (IPEA) as acetyl donor
and imidazole (see Methods). The results are shown in Table 1
as relative amounts of the educt resorcinol 1, the C-acetylation
product 4-acetyl-resorcinol 2 and the O-acetylation product 3-
hydroxyphenyl acetate 3.
Proposed mechanism of acyl-transfer
PpATase
catalyzes
the
reversible
acetylation
of
monoacetylphloroglucinol (MAPG) into phloroglucinol (PG) and
2,4-diacetylphloroglucinol (DAPG)[13]. It has been shown that
only a multi-component complex consisting of PhlA, PhlB and
PhlC subunits performs a disproportionation of MAPG[13a, 16]. The
crystal structures presented here revealed that only residues
from PhlC interact with the bound PG, strongly suggesting that
PhlA and PhlB are not directly involved in the actual acyl transfer
step. Instead, these subunits are very likely required for the
formation of a properly folded and functional dodecameric
enzyme complex. This finding also supports previous
suggestions regarding their possible involvement in preceding
Table 1. Activity of PpATase variants for the Friedel-Crafts acetylation of
benzene-1,3-diol (resorcinol, 1) and specific activity for the natural reaction
Acetylation
using
steps of DAPG biosynthesis[24]
.
IPEA/Im[a]
Entry PpATase
1
[%]
2
[%]
3
[%]
Specific
activity
Combining the structural results with activity data obtained for
enzyme variants (Table 1), a plausible catalytic mechanism can
be proposed (Figure 3). The observation of an acetylated Cys88
in both crystal structures and the lack of activity measured for
the cysteine to alanine variant indicates that Cys88 very likely
plays a significant role during catalysis. We suggest that thiol
group of Cys88 attacks the acyl moiety of the donor and
subsequently forms a covalent intermediate similar to acyl-
enzyme intermediates formed by serine or cysteine
hydrolases[46]. As in these hydrolases, the nucleophilic attack is
facilitated by stabilizing the ensuing negative charge at the acyl-
oxygen atom by polar interactions within the “oxyanion hole”.
There is no clear indication of a base in the vicinity of Cys88,
which could activate the thiol by deprotonation. It is reasonable,
however, that the sidechain is already deprotonated to some
extent at neutral or slightly basic pH-values. Additional
stabilization of the thiolate is possible through polar interactions
with the OH-group of Tyr298 and the imidazole moiety of His347,
although the distances to both groups are larger (>3.6 Å) than
observed for typical hydrogen bonds. The proposed formation of
an acyl-enzyme intermediate is also consistent with our previous
finding that conversions with DAPG/MAPG as acetyl donor did
not yield any phenyl acetate derivatives and that the enzyme
variant
[mU/mg][b]
1
w/o
68
n.d.
32
n.d.
enzyme
wt enzyme
H56A
2
3
41
81
81
84
74
69
79
81
31
48
79
82
82
77
73
73
44
n.d.
n.d.
6
n.d.
n.d.
5
n.d.
66
40
3
n.d.
3
15
19
19
10
26
31
16
19
3
12
18
18
15
22
27
26
44
n.d.
n.d.
4
n.d.
n.d.
n.d.
n.d.
16
24
1
n.d.
5
<1
4
H56S
5
N87A
6
C88A
7
C88S
8
9
H144A
H144S
W211A
W211F
Y298A
Y298V
Y298F
H347F
S349A
D352V
10
11
12
13
14
15
16
17
1
n.d.
1
n.d.
<1
[a]Conditions: cell-free extract of PpATase or variant (30 mg protein), resorcinol
(1, 10 mM), potassium phosphate buffer (50 mM, pH 7.5, total volume 1 mL),
IPEA (100 mM) and imidazole (100 mM added from a 1 M stock solution
prepared in the reaction buffer), 35 °C, 1.5 h and 750 rpm. The relative
amounts of 1-3 were determined by HPLC according to standard curves with
authentic samples. n.d. not detected.[b] Specific activity was determined
spectrophotometrically following the disproportionation of MAPG into DAPG
and PG
also catalyzes an intermolecular Fries-rearrangement[16b]
.
The second step of the reaction involves the transfer of the acyl
moiety from Cys88 to the aromatic ring of an acceptor molecule.
As mentioned above, the carbonyl carbon of the cysteine bound
As already shown previously[16b], the reaction using the wildtype
enzyme produces predominantly the C-acetylation product 2
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
acetyl group is appropriately positioned to attack the C-6 atom of
the bound phloroglucinol in the complex structure (Figure 3).
Electronic activation of the aromatic ring most likely involves
deprotonation of the phenolic OH group either at C-1 (by His56)
and/or at C-3 (by the diad His347-Asp352). We have previously
shown that – in addition to phloroglucinol – resorcinol but not
phenol can act as an acyl acceptor[16b] indicating that at least two
OH-groups (at positions 1 and 3) are necessary for substrate
binding and/or activity. In the structure, the OH-group at C-5 of
phloroglucinol does not form any polar interactions with the
enzyme and it can indeed be replaced by other (alkyl)
substituents in PpATase substrates[16b]. The rearomatization
step by an intramolecular proton transfer does very likely not
involve any enzyme intervention.
PhlC of PpATase differs from other, more common thiolases by
the presence of the tryptophan residue Trp211. This residue
appears to have
a lid function, because its conformation
significantly changes upon binding of the substrate (Figure 3).
This conformational flexibility could very well influence the
activity and above all the substrate specificity of the enzyme.
Regarding substitutions at the aromatic ring, the active site
cavity appears to provide more space for attachments at C-4 as
compared to C-5. This is in good agreement with the observed
preference for substituents at C-4 of the resorcinol core structure
over the C-5 position[16b]
.
Figure 3. Proposed mechanism of the acyl-transfer.
employing the sitting drop vapor-diffusion method in 96-well plates using
Index HT (Hampton Research), JCSG+ and Morpheus (Molecular
Dimensions) screens. A stock solution of PpATase (12 mg mL-1, in 50mM
potassium phosphate buffer pH 7.5) was used for all crystallization
experiments. Both in initial screens and subsequent optimizations, drops
of 1 µL were set with a 1:1 ratio of protein and precipitant solution. The
crystallization plates were incubated at 289 K.
Conclusions
The presented structural characterization of the multi-component
acytransferase PpATase reveals a close interaction of all three
enzymes PhlA, PhlB and PhlC. However, the C-C bond
formation without the utilization of CoA is performed only by the
PhlC subunit. This structural information provides a basis for
developing libraries of catalysts tailored for specific chemical
substrates. The extension of the substrate scope will be
achieved using established protein engineering techniques.
Crystal clusters of PpATase were readily obtained in several conditions.
Microseed-matrix-screening experiments[31] were set up using initial
crystals obtained in conditions Index #2 and #81 as seeding stock
(1:1000 dilution, 0.1 µL added to the drop). Using this technique, single
crystals were obtained in conditions Index #91 (0.15 M DL-malic acid pH
7.0, 20% w/v polyethylene glycol 3,350) and #55 (0.05 M magnesium
chloride hexahydrate, 0.1 M HEPES pH 7.5, 30% v/v polyethylene glycol
monomethyl ether 550) after about one week. Soaking experiments were
performed by adding small amounts of pure monoacetylphloroglucinol
(MAPG) dissolved in DMSO directly to the crystallization drop with a
cryo-loop. Soaking times varied between 20 to 60 sec. Untreated and
soaked crystals were flash-cooled in liquid nitrogen using 15% v/v
glycerol as cryoprotectant. Numerous crystals from different
crystallization conditions were tested for diffraction.
Experimental Section
Expression and purification
The CoA-independent acyltransferase from Pseudomonas protegens
DSM 19095 (PpATase) was expressed in E. coli as described
previously[16b] from a plasmid containing codon optimized ORFs coding
for the three subunits of the enzyme: PhlA, PhlB and PhlC. The
nucleotide sequence of the expression plasmid is available from
GenBank using the accession number KY173355. Purification of the
enzyme was achieved using size-exclusion-chromatography as
Data collection, processing, structure determination and analysis
Data were collected at 100 K on synchrotron beamlines ID23-1 and
ID30B (ESRF, Grenoble, France)[32] from an untreated crystal (hexagonal,
space group P6122) and a crystal soaked with MAPG (orthorhombic,
space group P212121) to crystallographic resolutions of approximately 2.8
and 3.4 Å, respectively. Diffraction data were processed and scaled
using the XDS package[33]. Initial automated molecular replacement
attempts using Balbes[34] and the better resolved hexagonal dataset
indicated the dimer of 3-hydroxy-3-methylglutaryl-coenzyme A synthase
described previously[16b]
.
Crystallization and soaking
Screening for crystallization conditions was performed with an Oryx8
crystallization robot (Douglas Instruments). Initial trials were set up
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
from Staphylococcus aureus (PDB-entry: 1tvz, 23% sequence identity)[35]
as a suitable search template for PhlA. A homology model of PhlC was
generated using the Phyre2 server[36] with the structure of SCP2 thiolase
from Leishmania mexicana (PDB-entry: 3zbg, 27% sequence identity)[28]
as template. Molecular replacement was continued within the CCP4
suite[37] using the program Phaser[38]. Fixing the previously positioned
1tvz-dimer (template for PhlA), two copies of the PhlC homology model
could be placed in the asymmetric unit. Density modification based on
phases from this partial model using the program Resolve[39] yielded well-
defined electron density for the whole PpATase-complex, including
density for two missing PhlB molecules. Manual rebuilding of PhlA and
PhlC was continued using the program Coot[40] and the improved model
was then submitted to automated rebuilding using the program
Buccaneer[41]. The resulting model containing two copies of PhlA, PhlB
and PhlC each was completed manually in Coot and refined using the
CC(work)
CC(free)
0.96 (0.83)
0.94 (0.74)
0.95 (0.90)
0.91 (0.82)
54563
Number of non-H 14084
atoms
Macromolecules
Ligands
13594
20
54398
116
Protein residues
RMS(bonds)
RMS(angles)
1800
0.011
0.91
96
7196
0.005
0.72
97
Ramachandran
favored (%)
Ramachandran
allowed (%)
3.7
3.1
PHENIX software suite[42]
.
Ramachandran
outliers (%)
0.11
0.18
The structure of the PpATase soaked with MAPG was solved by
molecular replacement using part of the previously determined
a
Rotamer outliers (%)
Clashscore
3.2
1.5
hexagonal structure (one copy of PhlA, PhlB and PhlC each) as search
template. Structure solution resulted in eight copies of this trimeric
arrangement in the asymmetric unit of the orthorhombic unit cell.
Structure refinement continued in the same manner as described above
using the programs Coot and PHENIX. Clear difference electron density
was observed in all eight chains of PhlC in the vicinity of residue Cys88.
In four of those chains we interpreted this density as a molecule of
phloroglucinol. Additional density around Cys88 was interpreted as an
acetyl group covalently attached to the Sg-atom of this amino acid.
4.29
10.12
43.64
43.63
52.78
30.28
5MG5
Average B-factor
Macromolecules
Ligands
21.32
21.24
25.53
23.51
5M3K
Solvent
PDB code
[a]Statistics for the highest-resolution shell are shown in parentheses.
Table 2. Crystal structure of PpATase, data collection and refinement
statistics
For both structures, validation was performed using the program
MolProbity[43]. Data collection and refinement statistics are summarized
in Table 2.
PpATase
hexagonal[a]
PpATase orthorhombic
(soaked)[a]
Wavelength
0.8726
0.95
All structure-related figures
were generated using PyMOL
Resolution range
48.74 - 2.83 (2.94 - 49.72 - 3.44 (3.56 -
2.83)
(http://www.pymol.org). Structures were superimposed using the SSM
Superposition tool[21] as implemented in Coot. Cavities were identified in
the final structures using the LIGSITE algorithm[44] as implemented in the
CaSoX plugin for PyMOL. The analysis of the hydrophobicity of these
3.44)
Space group
P 61 2 2
P 21 21 21
Unit cell dimensions
222.63
237.10
222.63 104.79 229.78 311.13
90 90 90
cavities utilized the corresponding function of the program VASCo[45]
.
Total reflections
Unique reflections
Multiplicity
695209 (67477)
82036 (7914)
8.5 (8.5)
602513 (51680)
99724 (9200)
6.0 (5.6)
Site directed mutagenesis and activity measurements
Variants of PpATase were prepared in order to investigate the role of
selected amino acid residues in the enzymatic reaction. Gene mutations
were introduced with the QuickChange II site-directed mutagenesis kit
(Agilent Genomics) according to the standard procedure provided by the
supplier without modifications. All primer sequences and plasmids used
in this study are collated in Supporting Information Table 1. The following
Completeness (%)
Mean I/s(I)
Wilson B-factor
R-merge
1.00 (0.98)
7.87 (2.23)
36.41
0.99 (0.93)
8.10 (2.57)
53.96
variants carrying
a single amino acid exchange within phlC were
generated: H56A, H56S, N87A, C88A, C88S, H144A, H144S, W211A,
W211F, Y298A, Y298V, Y298F, H347F, S349A and D352V. All variants
were expressed as described for the wildtype enzyme[16b] and were used
as cell free extracts in the activity measurements (Table 1).
0.255 (0.891)
0.271 (0.948)
0.98 (0.67)
0.99 (0.90)
0.268 (0.788)
0.294 (0.868)
0.97 (0.81)
0.99 (0.95)
99695 (9198)
R-meas
CC1/2
CC*
The activity of wildtype PpATase and of its variants was tested using the
acetylation of benzene-1,3-diol (resorcinol) under the following
conditions: Resorcinol (10 mM final concentration in the reaction mixture)
was dissolved in potassium phosphate buffer (50 mM, pH 7.5) and
preheated to 35 °C for 10 minutes. Cell-free extracts of the recombinant
enzymes (30 mg protein) were subsequently added to the preheated
mixture. The bioacetylation was started by addition of imidazole (100 mM,
added from a 1 M stock solution prepared in the reaction buffer causing
the pH to increase to 8.30) followed by the addition of isopropenyl
Reflections used in 82033 (7914)
refinement
Reflections used for 4102 (395)
R-free
4986 (460)
R-work
R-free
0.163 (0.251)
0.204 (0.316)
0.166 (0.219)
0.219 (0.292)
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
acetate (IPEA, 100 mM). To ensure proper suspension of the donor in
the mixture, the vessel was manually shaken thoroughly right after
starting the reaction. The reaction mixture (1 mL total volume) was
horizontally shaken for 1.5 h at 35 °C and 750 rpm in an orbital shaker.
Reactions were aborted by addition of acetonitrile (1 mL). The
precipitated protein was removed by centrifugation (10 min, 14,000 rpm)
and the supernatant (900 µL) was transferred to an Eppendorf tube and
left standing for another 40 minutes. Any residual precipitated protein
was once again removed by centrifugation and the supernatant was
directly subjected to HPLC for determination of conversions. The relative
amounts of resorcinol, the C-acetylation product 4-acetyl-resorcinol and
the O-acetylation product 3-hydroxyphenyl acetate were determined by
HPLC according to standard curves with authentic samples. Each
reaction was performed as a duplicate.
[9]
a) L. Paoletti, Y. J. Lu, G. E. Schujman, D. De Mendoza, C. O. Rock, J.
Bacteriol. 2007, 189, 5816-5824; b) Y. M. Zhang, C. O. Rock, J. Lipid
Res. 2008, 49, 1867-1874.
[10] R. Kalscheuer, A. Steinbüchel, J. Biol. Chem. 2003, 278, 8075-8082.
[11] Y. Q. Cheng, G. L. Tang, B. Shen, Proc. Natl. Acad. Sci. USA 2003,
100, 3149-3154.
[12] A. Bera, S. Herbert, A. Jakob, W. Vollmer, F. Götz, Mol. Microbiol. 2005,
55, 778-787.
[13] a) A. Hayashi, H. Saitou, T. Mori, I. Matano, H. Sugisaki, K. Maruyama,
Biosci. Biotechnol. Biochem. 2012, 76, 559-566; b) F. Yang, Y. Cao,
Appl. Microbiol. Biotechnol. 2012, 93, 487-495.
[14] a) C. Keel, D. M. Weller, A. Natsch, G. Defago, R. J. Cook, L. S.
Thomashow, Appl. Environ. Microbiol. 1996, 62, 552-563; b) J. A.
Moynihan, J. P. Morrissey, E. R. Coppoolse, W. J. Stiekema, F. O'Gara,
E. F. Boyd, Appl. Environ. Microbiol. 2009, 75, 2122-2131.
[15] a) Y. Cao, X. Jiang, R. Zhang, M. Xian, Appl. Microbiol. Biotechnol.
2011, 91, 1545-1552; b) Y. Cao, M. Xian, Biotechnol. Lett 2011, 33,
1853-1858; c) G. Rao, J. K. Lee, H. Zhao, Appl. Microbiol. Biotechnol.
2013, 97, 5861-5867; d) F. Yang, Y. Cao, Appl. Microbiol. Biotechnol.
2012, 93, 487-495; e) W. Zha, S. B. Rubin-Pitel, H. Zhao, J. Biol. Chem.
2006, 281, 32036-32047.
Specific activities were measured on a Thermo Scientific Genesys 10 UV
Scanning UV/Vis spectrophotometer according to a modified procedure
from literature[13a]. When following the disproportionation of MAPG into
DAPG and PG spectrophotometrically, an increase of absorption is
recorded due to the formation of DAPG (ε = 20 mM-1 cm-1, λ = 370 nm).
One unit of activity was defined as the µmol of product formed by an
enzyme in 1 min per 1 milligram of protein under the following conditions:
potassium phosphate buffer (960 µL, 100 mM, pH 7.5) and MAPG (1.2
µmol, 30 µL of a 40 mM stock solution prepared in DMSO) were added to
a cuvette and preheated to 35 °C. The reaction (1 mL total volume, 3
vol% DMSO) was started by the addition of the enzyme-containing cell-
free extract (10 µL). The reaction was followed for 1 minute. All reactions
were performed as a duplicate. The protein concentration was measured
using Bradford reagent [(ε = 0.083 mL mg-1 cm-1, λ = 595 nm)] and
specific activities were determined as units per mg protein.
[16] a) J. Achkar, M. Xian, H. Zhao, J. W. Frost, J. Am. Chem. Soc. 2005,
127, 5332-5333; b) N. G. Schmidt, T. Pavkov-Keller, N. Richter, B.
Wiltschi, K. Gruber, W. Kroutil, Angew. Chem. Int. Ed. 2017, 56, 7615-
7619.
[17] N. G. Schmidt, W. Kroutil, Eur. J. Org. Chem. 2017, 2017, 5865-5871.
[18] A. Zadło-Dobrowolska, N. G. Schmidt, W. Kroutil, Chem. Commun.
2018, 54, 3387-3390.
[19] E. Krissinel, K. Henrick, J. Mol. Biol. 2007, 372, 774-797.
[20] L. Holm, S. Kaariainen, P. Rosenstrom, A. Schenkel, Bioinformatics
2008, 24, 2780-2781.
[21] E. Krissinel, K. Henrick, Acta Cryst. D 2004, 60, 2256-2268.
[22] T. Bock, J. Kasten, R. Müller, W. Blankenfeldt, ChemBioChem 2016,
1257-1262.
Acknowledgements
[23] X. Y. Qiu, C. A. Janson, W. W. Smith, M. Head, J. Lonsdale, A. K.
Konstantinidis, J. Mol. Biol. 2001, 307, 341-356.
This study was financed by the Austrian FFG, BMWFJ, BMVIT,
SFG, Standortagentur Tirol and ZIT through the Austrian FFG-
COMET- Funding Program. COST Action CM1303 “Systems
Biocatalysis” is acknowledged. Additional support was provided
by the Austrian Science Fund (FWF) through a Lise-Meitner
fellowship (M2172-B21) to AZD and through project P29432 to
TPK. We acknowledge the ESRF (ID23-1 and ID30B, Grenoble,
France) and DESY (P11, PETRAIII, Hamburg, Germany) for
provision of synchrotron-radiation facilities and support during
data collection.
[24] M. G. Bangera, L. S. Thomashow, J. Bacteriol. 1999, 181, 3155-3163.
[25] S. S. Krishna, L. Aravind, C. Bakolitsa, J. Caruthers, D. Carlton, M. D.
Miller, P. Abdubek, T. Astakhova, H. L. Axelrod, H. J. Chiu, T. Clayton,
M. C. Deller, L. A. Duan, J. Feuerhelm, J. C. Grant, G. W. Han, L.
Jaroszewski, K. K. Jin, H. E. Klock, M. W. Knuth, A. Kumar, D.
Marciano, D. McMullan, A. T. Morse, E. Nigoghossian, L. Okach, R.
Reyes, C. L. Rife, H. van den Bedem, D. Weekes, Q. P. Xu, K. O.
Hodgson, J. Wooley, M. A. Elsliger, A. M. Deacon, A. Godzik, S. A.
Lesley, I. A. Wilson, Acta Cryst. F 2010, 66, 1160-1166.
[26] R. L. Flynn, L. Zou, Crit. Rev. Biochem. Mol. Biol. 2010, 45, 266-275.
[27] A. M. Haapalainen, G. Meriläinen, R. K. Wierenga, Trends Biochem.
Sci 2006, 31, 64-71.
Keywords: acyltransferase • crystal structure • Friedel-Crafts
[28] R. K. Harijan, T. R. Kiema, M. P. Karjalainen, N. Janardan, M. R.
Murthy, M. S. Weiss, P. A. Michels, R. K. Wierenga, Biochem. J. 2013,
455, 119-130.
acylation • multi-component enzyme
[29] B. Vögeli, S. Engilberge, E. Girard, F. Riobé, O. Maury, T. J. Erb, S.
Shima, T. Wagner, Proc. Natl. Acad. Sci. USA 2018, 115, 3380-3385.
[30] L. David, E. Cheah, M. Cygler, B. Dijkstra, F. Frolow, M. Sybille, M.
Harel, S. James Remington, I. Silman, J. Schrag, L. Joel, H. G. V. Koen,
A. Goldman, Protein Eng. Des. Sel. 1992, 5, 197-211.
[1]
a) Y. Liu, L. Gao, L. Liu, Q. Yang, Z. Lu, Z. Nie, Y. Wang, T. Xia, J. Biol.
Chem. 2012, 287, 44406-44417; b) R. Niemetz, G. G. Gross,
Phytochemistry 2001, 58, 657-661.
[2]
[3]
D. M. Byers, H. Gong, Biochem. Cell Biol. 2007, 85, 649-662.
L. Hoffmann, S. Maury, F. Martz, P. Geoffroy, M. Legrand, J. Biol.
Chem. 2003, 278, 95-103.
[31] A. D'Arcy, F. Villard, M. Marsh, Acta Cryst. D 2007, 63, 550-554.
[32] P. Theveneau, R. Baker, R. Barrett, A. Beteva, M. W. Bowler, P.
Carpentier, H. Caserotto, D. de Sanctis, F. Dobias, D. Flot, M. Guijarro,
T. Giraud, M. Lentini, G. A. Leonard, M. Mattenet, A. A. McCarthy, S. M.
McSweeney, C. Morawe, M. Nanao, D. Nurizzo, S. Ohlsson, P. Pernot,
A. N. Popov, A. Round, A. Royant, W. Schmid, A. Snigirev, J. Surr, C.
Mueller-Dieckmann, 11th International Conference on Synchrotron
Radiation Instrumentation (Sri 2012) 2013, 425.
[4]
M. C. Hunt, M. I. Siponen, S. E. H. Alexson, BBA - Mol. Bas. Dis. 2012,
1822, 1397-1410.
[5]
[6]
W. Zhang, J. Huffman, S. Li, Y. Shen, L. Du, BMC Biotechnol. 2017, 17.
L. Favrot, J. S. Blanchard, O. Vergnolle, Biochemistry 2016, 55, 989-
1002.
[7]
[8]
S. R. Partridge, R. M. Hall, J. Clin. Microbiol. 2005, 43, 4298-4300.
D. M. Stalker, K. E. McBride, L. D. Malyj, Science 1988, 242, 419-423.
[33] W. Kabsch, Acta Cryst. D 2010, 66, 125-132.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
[34] F. Long, A. A. Vagin, P. Young, G. N. Murshudov, Acta Cryst. D 2008,
64, 125-132.
[41] K. Cowtan, Acta Cryst. D 2006, 62, 1002-1011.
[42] P. D. Adams, P. V. Afonine, G. Bunkoczi, V. B. Chen, I. W. Davis, N.
Echols, J. J. Headd, L.-W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve,
A. J. McCoy, N. W. Moriarty, R. Oeffner, R. J. Read, D. C. Richardson,
J. S. Richardson, T. C. Terwilliger, P. H. Zwart, Acta Cryst. D 2010, 66,
213-221.
[35] N. Campobasso, M. Patel, I. E. Wilding, H. Kallender, M. Rosenberg, M.
N. Gwynn, J. Biol. Chem. 2004, 279, 44883-44888.
[36] L. A. Kelley, S. Mezulis, C. M. Yates, M. N. Wass, M. J. Sternberg, Nat.
Protoc. 2015, 10, 845-858.
[37] M. D. Winn, C. C. Ballard, K. D. Cowtan, E. J. Dodson, P. Emsley, P. R.
Evans, R. M. Keegan, E. B. Krissinel, A. G. Leslie, A. McCoy, S. J.
McNicholas, G. N. Murshudov, N. S. Pannu, E. A. Potterton, H. R.
Powell, R. J. Read, A. Vagin, K. S. Wilson, Acta Cryst. D 2011, 67, 235-
242.
[43] V. B. Chen, W. B. Arendall, J. J. Headd, D. A. Keedy, R. M. Immormino,
G. J. Kapral, L. W. Murray, J. S. Richardson, D. C. Richardson, Acta
Cryst. D 2010, 66, 12-21.
[44] M. Hendlich, F. Rippmann, G. Barnickel, J. Mol. Graph. Model. 1997,
15, 359-363.
[38] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C.
Storoni, R. J. Read, J. Appl. Crystallogr. 2007, 40, 658-674.
[39] T. C. Terwilliger, Acta Cryst. D 2000, 56, 965-972.
[40] P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan, Acta Cryst. D 2010,
66, 486-501.
[45] G. Steinkellner, R. Rader, G. G. Thallinger, C. Kratky, K. Gruber, BMC
Bioinformatics 2009, 10, 1-11.
[46] L. Hedstrom, Chem. Rev. 2002, 102, 4501-4523.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800462
ChemBioChem
FULL PAPER
Entry for the Table of Contents
Layout 1:
FULL PAPER
Friedel-Crafts acylase: This
multimeric enzyme consists of
three subunits – PhlA, PhlB and
PhlC which are arranged in a
Phl(A2C2)2B4 composition. The
subunit PhlC catalyzes a Friedel-
Crafts C-acylation of phenolic
substrates in aqueous solution
without the need of CoA-activated
reagents.
Tea Pavkov-Keller, [a,b] Nina G.
Schmidt, [a,c] Anna Żądło-Dobrowolska,
[c] Wolfgang Kroutil, *[a,c,d] and Karl
Gruber *[a,b,d]
Page No. – Page No.
Structure and catalytic mechanism
of a bacterial Friedel-Crafts acylase
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.