DptE was proven to preferentially activate branched mid- to long-chain fatty acids under ATP consumption, and these fatty acids are subsequently transferred onto DptF, the cognate acyl c
Trang 1of daptomycin
Melanie Wittmann, Uwe Linne, Verena Pohlmann and Mohamed A Marahiel
Department of Chemistry ⁄ Biochemistry, Philipps-University Marburg, Germany
Daptomycin is a clinically important semi-synthetic
derivative of the A21978C branched cyclic lipopeptide
antibiotics produced by Streptomyces roseosporus [1]
Acidic lipopeptide antibiotics present a new class of
therapeutic agents that includes compounds such as
calcium dependent antibiotic (CDA) [2], A54145 [3,4]
and friulimicin [5,6] with a unique mechanism of
action Daptomycin binds to Gram-positive cell
membranes via its lipid moiety, followed by
calcium-dependent insertion and oligomerization Subsequently,
oligomers form ion channels that disrupt the bacterial
membrane potential, leading to rapid cell death [7,8]
Daptomycin comprises a 13-amino acid peptide core
coupled to a fatty acid moiety (Fig 1) The peptide
core is assembled nonribosomally by dptA and dptBC
The thioesterase DptD of the daptomycin biosynthetic
gene cluster catalyses the cyclization reaction between
the hydroxyl group of Thr4 and the C-terminal
Kyn13, resulting in a ten-membered ring [8]
More-over, several ORFs localized within the gene cluster are associated with the biosynthesis of non-proteino-genic amino acids and incorporation of the fatty acid moiety [1]
All acidic lipopeptides (except CDA) produced
in vivo show some flexibility with respect to the length and branching of their N-terminally attached fatty acid groups (Fig 1) The activity of lipopeptide antibiotics
as well as the toxicity towards eukaryotic cells strongly depends on the nature of the acyl moiety [9,10] The fine tuning between these two features is of consider-able importance for the development of selective potent drugs
The biosynthesis of the peptide core of these acidic lipopeptides via nonribosomal peptide synthetases (NRPSs) is well understood, but little is known about the incorporation of the acyl residue into the final product [11,12] As revealed by sequence comparison, the initiation modules of such NRPSs contain unique
Keywords
acidic lipopeptide antibiotics; AMP ligase;
daptomycin; lipidation reaction;
nonribosomal peptide synthetases
Correspondence
M A Marahiel, Department of
Chemistry ⁄ Biochemistry, Philipps-University
Marburg, Hans-Meerwein-Strasse, D-35043
Marburg, Germany
Fax: +49 6421 2822191
Tel: +49 6241 2825722
E-mail: marahiel@chemie.uni-marburg.de
(Received 4 June 2008, revised 29 August
2008, accepted 1 September 2008)
doi:10.1111/j.1742-4658.2008.06664.x
Daptomycin and A21987C antibiotics are branched, cyclic, nonribosomally assembled acidic lipodepsipeptides produced by Streptomyces roseosporus The antibacterial activity of daptomycin against Gram-positive bacteria strongly depends on the nature of the N-terminal fatty acid moiety Two genes, dptE and dptF, localized upstream of the daptomycin nonribosomal peptide synthetase genes, are thought to be involved in the lipidation of daptomycin Here we describe the cloning, heterologous expression, purifi-cation and biochemical characterization of the enzymes encoded by these genes DptE was proven to preferentially activate branched mid- to long-chain fatty acids under ATP consumption, and these fatty acids are subsequently transferred onto DptF, the cognate acyl carrier protein Addi-tionally, we demonstrate that lipidation of DptF by DptE in trans is based
on specific protein–protein interactions, as DptF is favored over other acyl carrier proteins Study of DptE and DptF may provide useful insights into the lipidation mechanism, and these enzymes may be used to generate novel daptomycin derivatives with altered fatty acids
Abbreviations
CDA, calcium dependent antibiotic; PKS, polyketide synthase.
Trang 2condensation (CIII) domains that are thought to
cata-lyse N-acylation of the first amino acid in the peptide
chain [13] However, the fatty acid moiety must be
activated prior to being incorporated into the
product Two classes of enzymes are known to
catalyse such reactions One class, acyl CoASH
synthetases, recognize and activate fatty acids as acyl
adenylates (acyl AMPs), and subsequently couple
them to coenzyme A (CoASH) The second class,
fatty acyl ACP ligases, activate and transfer fatty
acids from acyl AMP to cognate acyl carrier proteins
(ACPs) [14,15]
The genes dptE and dptF are localized immediately
upstream of the NRPSs of A21987C The resulting
proteins DptE and DptF were predicted to be
involved in the lipidation reaction based on sequence
similarity [1] DptE is similar to other
adenylate-forming enzymes such as acyl CoASH synthetases,
and DptF is a putative ACP Both proteins are
thought to be important for the initiation of
dapto-mycin biosynthesis [1,16]
In this study, we describe the biochemical character-ization of DptE as an acyl ACP ligase, and demon-strate transfer of various fatty acids onto the ACP encoded by dptF (Fig 2) This biochemical character-ization of the lipidation mechanism during acidic lipopeptide biosynthesis may facilitate engineering of new derivatives with altered activities
Results
Initial biochemical characterization of DptE and DptF
DptE shares approximately 20% sequence identity with several members of the acyl AMP⁄ CoASH ligase super-family [17] These enzymes catalyse the formation of fatty acyl AMP⁄ CoASH from a fatty acid substrate, ATP and CoASH in a Mg2+-dependent two-step reaction [17–19]
In general, a fatty acyl adenylate intermediate is formed
in the first step, followed by conversion of the fatty acyl adenylate to fatty acyl CoA with release of AMP
Fig 1 Chemical structures of the lipopeptide antibiotics daptomycin, A54145 and CDA, and their natural fatty acid moieties Daptomycin and A54145 are naturally produced with various fatty acid side chains For daptomycin, the major fatty acids are shown CDA is produced with the epoxidized hexanoyl moiety exclusively.
Trang 3DptE was cloned into the pBAD102⁄ D-TOPO
vector and overexpressed in Escherichia coli
BL21(DE3) The C-terminally His6-tagged and
N-ter-minally thioredoxin-fused protein was purified, yielding
4.4 mgÆL)1 of culture The identity of the protein was
confirmed by SDS–PAGE (Fig 3) and mass
spectro-metry (Table 1) An initial fatty acid-dependent
ATP⁄ PPi exchange assay according to functionally
related adenylation domains of NRPSs showed no
activity (data not shown) To determine whether
CoASH is the physiological substrate of DptE and
required for enzyme activity, we determined the
activity of DptE with ATP, MgCl2 and CoASH under
various conditions However, no acyl CoA was detect-able by HPLC-MS (data not shown)
As we were not able to detect any in vitro activity of DptE using ATP⁄ PPiexchange assays, and no lipidation
of CoASH was observed in the presence of fatty acids,
we next focused on the transcriptionally coupled dptF, which encodes a stand alone putative ACP [20] ACPs contain the modestly conserved motif GxDS(I⁄ L), in which the serine residue is post-translationally modified
by covalent attachment of a 4¢-phosphopantethein group [21,22] The motif present in DptF is GLDSV, indicating that this putative ACP domain is one of the few ACPs in which valine replaces isoleucine (I) or leucine (L) in the conserved sequence To determine whether DptF is the putative partner of DptE, we expressed dptF using the pQTev vector in E coli and purified the resulting ACP as an N-terminal His7 fusion protein (Fig 3) with a yield of 9.5 mgÆL)1 of culture The identity of the protein was proven by SDS–PAGE
+
decanoic acid
DptE
+ATP
PPi
O
O–
7
SH
holo-ACP DptF
DptA DptBC DptD
daptomycin
DptF
S
O
decanoyl-S-ACP
7
AMP DptE
O
O
Fig 2 Proposed mechanism for the
lipidation of daptomycin by DptE and DptF.
Decanoic acid is activated by the putative
adenylating enzyme DptE under ATP
con-sumption The fatty acid is then transferred
onto the acyl carrier protein DptF The C
domain of DptA is predicted to catalyse the
condensation reaction between the fatty
acid and tryptophan.
80
25
20
30
40
50
60
100
150
Fig 3 Coomassie blue-stained SDS–PAGE gel of purified apo-DptF
(aDptF, 13.5 kDa), holo-DptF (hDptF, 13.8 kDa), apo-LipD (aLipD,
11.1 kDa), holo-LipD (11.5 kDa), hAcpK (B subtilis, holo form;
10.9 kDa) and DptE (80.5 kDa) SDS–PAGE was performed using a
NuPAGE 4-12% Bis-Tris gel (Invitrogen) The protein ladder was
from New England Biolabs (P7703, 10-250 kDa).
Table 1 [M+H] + mass values for the proteins, substrates and products.
[M+H] + (Da)
Trang 4and tryptic digestion followed by mass spectrometry.
Subsequently, DptF was incubated with the
promiscu-ous 4¢-phosphopantetheinyl transferase Sfp from
Bacil-lus subtilis and fluoresceinyl CoA [23] The successful
4¢-phosphopantetheinylation of DptF was monitored by
the in-gel fluorescence of the reaction mixture (Fig 4)
For subsequent acylation studies, holo-DptF was
produced in the sfp-containing E coli strain HM0079
[24] The in vivo modification of DptF by Sfp resulted
in 100% conversion of apo-DptF to holo-DptF as
estimated by tandem fourier transform ion cyclotron
resonance-MS (Fig 5 and Table 1)
Lipidation of DptF by DptE
Initially, 50 lm holo-DptF was incubated with 500 lm
decanoic acid, 10 mm MgCl2, 1 mm ATP and 1 lm
DptE (Fig 5) The reaction mixture was quenched with
10% formic acid after 10 min and subjected to
HPLC-ESI-MS analysis (Table 1) DptF was quantitatively
acylated with decanoic acid Subsequently, we
deter-mined the pH and temperature for maximum
forma-tion of decanoyl-S-ACP catalysed by DptE Suitable
reaction conditions were determined to be pH 7.0 and
37C, in agreement with those reported for other acyl
AMP⁄ ACP ⁄ CoASH ligases [15,25] Omitting DptE or
ATP abolished the acylation of DptF completely These results indicate that decanoic acid is activated as
a fatty acyl AMP and subsequently transferred onto holo-DptF by the acyl ACP synthetase DptE To detect the adenylate intermediate, we repeated the reaction with apo-DptF rather than holo-DptF, which should lead to accumulation of the acyl adenylate intermedi-ate The reaction was stopped using 10% formic acid and subjected to LC-MS to detect decanoic AMP (data not shown) However, we were not able to detect the adenylate intermediate using this approach Next, we performed an ATP⁄ PPi exchange assay with apo-DptF
in the presence of phosphate buffer Control reactions were performed without radioactively labelled PPi, DptE, apo-DptF, MgCl2, or ATP In the presence of apo-DptF, we observed an approximately 100-fold higher activity of DptE compared to the control reac-tions (Fig 6) The above-mentioned condireac-tions were used for determination of steady-state kinetic para-meters The KMand kcatvalues of DptE for holo-DptF (with concentrations between 2.5 and 250 lm) were 29.4 lm and 7.4 min)1 under decanoic acid satura-tion (500 lm), resulting in a catalytic efficiency of 0.25 min)1Ælm)1 Addition of CoASH to the reaction
10
15
20
30
SDS-PAGE UV-irradiation at 312 nm
Fig 4 In vitro phosphopantetheinylation of DptF and
apo-AcpK Coomassie blue-stained SDS–PAGE gel (left) and in-gel
fluo-rescence (right) of the fluoresceinyl-ACP (+) indicates the reaction
with Sfp; ( )) indicates the reaction without Sfp.
m/z
13491.6 apo-DptF
+ Na
+ Ka
13 200 13 300 13 400 13 500 13 600 13 700 13 800 13 900 13 500 13 600 13 700 13 800 13 900 14 000 14 100
13831.7 holo-DptF
+ Na + Ka
m/z
13986.9 decanoyl-DptF
+ Ka + Na
13 500 13 600 13 700 13 800 13 900 14 000 14 100
Sfp / CoASH/ Mg 2+
-5'-3'-ADP
DptE -AMP + PPi +ATP
m/z
Fig 5 Fourier transform MS spectra of apo-DptF (left), holo-DptF (middle) and decanoic acid-loaded DptF (right).
0
5000
10 000
15 000
20 000
25 000
30 000
Fig 6 ATP ⁄ PP i exchange assay of DptE in the presence of apo-DptF, and control reactions without radioactive labeled PPi(PPi*), DptE, ATP or MgCl2.
Trang 5mixture did not affect the product formation activity
(data not shown) Therefore, the results clearly
demon-strate that holo-DptF is the cognate acceptor subdemon-strate
of DptE (Table 2)
Fatty acid specificity of the AMP ligase DptE
Having proven a functional interaction of DptE and
DptF utilizing decanoic acid as a standard substrate,
we addressed the important question of DptE
specific-ity We systematically utilized a range of linear and branched chain fatty acids as well as hydroxy-fatty acids with various chain lengths and varied the concen-trations between 2.5 and 500 lm Kinetic constants were determined by Michaelis–Menten fitting of the data sets The summarized kinetic data, which were obtained under ATP and holo-DptF saturation, are presented in Table 2 As expected, the fatty acids (branched and linear, between 10 and 12 carbon units) that are known to be present in naturally produced A21987C lipopeptides and in the drug Cubicin (dap-tomycin formulated for injection) were observed to be excellent substrates, with KMvalues ranging from 8 to
20 lm and kcat values between 3.4 and 18.3 min)1 Catalytic efficiencies were 0.29–0.95 min)1Ælm)1 These values are in good agreement with those observed for other systems in which a fatty acyl ACP synthetase lipidates a cognate holo-ACP in trans [26] Octanoic acid, tetradecanoic acid and the 3-hydroxy fatty acid, which have not been reported as occurring in the natural compound, were relatively poor substrates, with KM values 2–13-fold higher than those for fatty acids naturally found in A21987C Hexanoic acid, palmitic acid and 15-methylhexadecanoic acid were not accepted by DptE
In summary, DptE is capable of transferring a variety of fatty acids to the cognate ACP DptF in vitro The kinetic data presented in this study indicate that DptE has a general preference for linear fatty acids with chain lengths between 8 and 14 carbon units, particularly iso⁄ anteiso-branched chain fatty acids and
Table 2 Kinetic parameters for steady-state analysis of the DptE-catalysed lipidation of DptF determined at varying concentrations of fatty acids or DptF, LipD and AcpK.
Substrate KM(l M ) kcat(min)1) kcat⁄ K M (min)1Æl M )1) Linear a
C12 10.9 ± 0.3 3.1 ± 0.1 0.29 ± 0.01 C14 26.6 ± 0.9 1.3 ± 0.2 0.05 ± 0.01 Branched b
iso-C10 19.3 ± 0.5 17.9 ± 0.5 0.93 ± 0.05 iso-C12 19.2 ± 0.4 18.3 ± 0.5 0 95 ± 0.05 iso-C13 16.1 ± 0.6 15.3 ± 0.7 0.95 ± 0.08 anteiso-C12 14.1 ± 0.8 13.1 ± 0.2 0.93 ± 0.08 Hydroxylated
3OH-C12 114.2 ± 4.2 5.1 ± 0.4 0.04 ± 0.01 ACPs
holo-DptF 29.4 ± 0.4 7.4 ± 0.2 0.25 ± 0.01 holo-LipD 135.0 ± 0.5 6.3 ± 0.3 0.05 ± 0.03
a Substrates C6 and C16 were not activated b Substrate anteiso-C16 was not activated.
+ Na
10901.6 holo-AcpK
10 600
10 500 10 700 10 800 10 900 11 000
m/z
10901.6 holo-AcpK
+ Na
10561.5
apo-AcpK
m/z
10 600 10 700 10 800 10 900 11 000 11 100 11 200 11 300
-5'-3'-ADP Sfp / CoASH/Mg2+
Fig 7 AcpK expressed in its active holo form in M15 ⁄ pRep4-gsp
cells (HM404) Only approximately 40% of AcpK is expressed in
the holo form (upper) Phosphopantetheinylation of apo-AcpK with
Sfp after expressing in M15 ⁄ pRep4-gsp cells (HM404) (lower).
Trang 6decanoic acid, while long chain fatty acids such as
palmitic acid or 15-methylhexadecanoic acid are not
recognized at all Hydroxylated fatty acids are
accepted, but with lower efficiencies
ACP specificity of the acyl ACP ligase DptE
The results presented above confirm that DptE
acti-vates various fatty acids and transfers them onto
DptF To address the question of the specificity of
DptE towards ACPs, we utilized as alternative ACPs
LipD, an ACP that is involved in friulimicin
biosyn-thesis and shows approximately 31% sequence identity
with DptF, and holo-AcpK (Fig 7) from B subtilis,
which shares approximately 13% sequence identity
with DptF Mass spectrometry analysis of the assayed
holo-AcpK showed no product formation In a
reaction mixture containing both DptF and AcpK,
acylation of DptF exclusively was observed (data not
shown) LipD was only partially acylated in presence
or absence of DptF For better comparison of
the reaction velocities obtained with DptF and LipD,
we performed kinetic studies To determine kinetic
data for LipD, this protein was expressed in vivo in its
active holo form (see Experimental procedures) The
reaction mixtures contained 1 mm ATP, 10 mm
MgCl2, 2–250 lm holo-LipD, 1% dimethylsulfoxide
and 500 lm decanoic acid Michaelis–Menten fitting of
the experimental data set resulted in a KMof 135 lm
and a kcat of 6.3 min)1 The catalytic efficiency of the
transfer reaction to LipD (0.047 min)1Ælm)1) was
approximately five times lower than that for DptF
(0.25 min)1Ælm)1) (Table 2) In conclusion, these
results suggest that there is specific recognition
between DptE and DptF
Discussion
Daptomycin is a prominent member of the
pharmaco-logically important class of antimicrobial acidic
lipo-peptides It has been commercialized as Cubicin
(Cubist Pharmaceuticals Inc., Lexington, PA, USA)
for the treatment of serious infections caused by
Gram-positive bacteria [27] Recently, it has been
shown that the activity of these acidic lipopeptides is
significantly influenced by the length and structure of
their fatty acid moieties [9,10] In the fermentation of
these natural products, some flexibility with respect to
the length and branching of the lipid side chain has
been observed [10] Complete biochemical
characteri-zation of the lipidation reaction may allow the
engineering of lipopeptides with modified fatty acid
moieties, which could lead to new antibiotics active
against a wide range of bacteria, preventing damage to eukaryotic cells For incorporation of the fatty acid moiety into nonribosomal peptides, condensation of the fatty acid with the N-terminal tryptophan of the nonribosomally synthesized peptide is necessary Here, we report the results of a steady-state kinetic analysis of DptE The aim of this kinetic study was to determine the specificity of DptE for various fatty acids and noncognate ACPs The Michaelis–Menten kinetic values indicate catalysis of the two-step reaction with one substrate (fatty acid or ACP) under saturating or non-saturating conditions The kinetic data for the various fatty acids transferred onto DptF
by DptE reported here indicate the preference of DptE for those found in the naturally produced daptomycin derivatives Additionally, it was observed that long-chain (16 carbon units or more) and short-long-chain fatty acids (six carbon units or fewer) are not accepted by DptE The observation that DptE is able to activate and transfer a broad range of fatty acids fits well with results for other fatty acid CoASH synthetases such as Faa1p from Saccaromyces cerevisiae [28] or
CpPKS1-AL from Cryptosporidium parvum [26] Faa1p func-tions in the vectorial acylation of exogenous long-chain fatty acids, and has a preference for fatty acid sub-strates with 10–18 carbons The KM value of Faa1p for oleate is 71.1 lm The CpPKS1-AL domain has been proposed to be involved in the biosynthesis of a yet undetermined polyketide This domain also shows broad substrate acceptance but with a preference for long-chain fatty acids, particularly arachidic acid The actual substrates for the fatty acid CoASH⁄ ACP synthetases will be limited by the availability of fatty acids in the host organism
Interestingly, comparison of the kcat⁄ KM values for DptE revealed that it is five times more active with the physiologically relevant ACP DptF than with to LipD (Table 2), and is inactive with AcpK Therefore, the
in trans lipidation of DptF appears to be the result of specific protein–protein communication [29]
Faa1p, which functions by a common ‘ping pong BI-BI’ mechanism [30–32], showed a KM of 18.3 lm for its cognate ACP In the case of CpPKS1-AL, the
KM for the lipidation of ACP was 3.53 lm These findings are in good agreement with those for DptE, which has a KMof 29.4 lm for its cognate ACP
In microorganisms, various strategies exist for the activation of fatty acids Gokhale et al [14,33,34] found several enzymes for fatty acid activation in Mycobacterium tuberculosis These putative enzymes were cloned and expressed in E coli, and two distinct classes were found, namely fatty acyl AMP ligases and fatty acyl CoASH ligases The AMP ligases activate
Trang 7metabolic fatty acids as acyl adenylates, which are
sub-sequently transferred to a cognate holo-ACP domain
In contrast, the acyl CoASH ligases catalyse transfer
onto CoASH, forming an acyl thioester, which
subsequently undergoes transthiolation with the
HS-phosphopantetheine group of an ACP [14,33,34]
The loading module of the polyketide synthase (PKS)⁄
NRPS hybrid mycosubtilin was recently characterized
[35,36] It was shown that priming of ACP1 with a
fatty acid occurs via an acyl AMP ligase domain in cis
The latter type of fatty acid activation and loading
was exclusively reported for PKS systems [14]
Recently, lipidation of the acidic lipopeptide CDA
was investigated in vivo and in vitro [16] However,
CDA is an exception within the acidic lipopeptides, as
only 2,3-epoxy-hexanoic acid is incorporated into the
final product, and two specific enzymes encoded by
fabH3 and fabH4 are thought to synthesize hexanoic
acid directly on an ACP Two additional proteins
encoded in the CDA fab operon, HxcO and HcmO,
are responsible for the subsequent epoxidation of
hexa-noyl S-ACP [37]
Interestingly, in the case of the lipopeptide surfactin,
neither an acyl CoASH ligase-like domain nor an ACP
could be identified within the biosynthetic gene cluster
using bioinformatic tools [38] Previously, an unknown
40 kDa protein was thought to be the candidate for
lipidation However, it has been suggested that the
activated 3-hydroxymyristoyl CoA substrate is
bio-synthesized by the primary metabolism Recently, it
was reported that the acyl CoA substrate is transferred
to the initiation module SrfA-A1 This transfer is
stimulated by the surfactin thioesterase II SrfD [38]
However, the reaction also took place in the absence
of the thioesterase, but with reduced turnover To
date, no additional enzyme such as an acyltransferase
or an acyl CoASH ligase has been reported to be
involved in the surfactin initiation process
Another possibility for lipidation of secondary
metabolites could be the interaction of fatty acid
synthase-like enzymes or substrates from the primary
metabolism with NRPSs or PKSs, as shown for
afla-toxin produced by the fungi Asparagillus parasiticus
and A flavus [39,40] In this example, the fatty acid
synthase-like enzymes HexA and HexB synthesize
hexanoic acid from acetyl CoA and two units of
malonyl CoA This hexanoic acid serves as a precursor
for initiation of the PKS of aflatoxin biosynthesis
As shown here, the acyl ACP ligase DptE of the
daptomycin biosynthetic gene cluster appears to
directly select and activate cytosolic fatty acids from
primary metabolism as fatty acyl adenylates in a
mech-anism analogous to the adenylation domains of
NRPSs [41] Subsequently, the fatty acids are trans-ferred in trans onto holo-DptF to generate fatty acyl S-ACP No lipidation was observed without ATP, confirming our conclusion that the fatty acid has to be activated as an adenylate prior to esterfication by the cognate ACP
Interestingly, we detected a 100-fold higher activity over background in the ATP⁄ PPi exchange assay with DptE when it was performed in the presence of nonre-active apo-DptF (approximately 26 500 c.p.m.) In the absence of DptE we found only a marginal activity (approximately 250 c.p.m.) This leads to the conclu-sion that DptE requires DptF for its activity We sug-gest that the reason that a fatty acid adenylate intermediate was not detected using apo-DptF in an LC-MS approach is that the back reaction was too fast or the amount of product was below the detection limit
Summarizing, the present study focuses on the bio-chemical characterization of DptE and DptF To date, we cannot rule out the possibility that similar fatty acid CoA derivatives will also be recognized by the C domain of the initiation module of dapto-mycin That DptE and DptF are involved in the lipidation process of daptomycin was first shown by Miao et al [1] In their work, the daptomycin gene cluster was heterologously expressed in Streptomyces lividans Only authentic daptomycin derivatives were found and no derivatives with common fatty acids
of the S lividans organism Studies utilizing deletion mutants or biochemical studies involving the initia-tion module of daptomycin synthetase are required
to prove whether DptE and DptF are essential for lipidation or whether there are additionally alterna-tive pathways
In conclusion, DptE was observed to recognize a variety of fatty acid moieties After activation of the fatty acids under ATP consumption, most likely as fatty acyl AMPs, DptE subsequently catalyses specific transfer onto the 4¢-phosphopantethein group of DptF The observed substrate tolerance for loading a variety
of fatty acids onto the ACP will facilitate future pro-jects on the manipulation and combinatorial biosyn-thesis of acidic lipopeptides Hopefully, the recognition and efficient transfer of new building blocks can be achieved using DptE and DptF This is important, as the fatty acid moiety has been proven to have a high impact on the bioactivity and bioselectivity of these antibiotics [9,10] It remains to be clarified whether all
of the fatty acids activated by DptE can be incorpo-rated into the final product or whether there is an interfering specificity of the CIII domain of the initia-tion module
Trang 8Experimental procedures
Materials
Electrocompetent Top10 and BL21 (DE3) E coli cells were
purchased from Invitrogen (Carlsbad, CA, USA) All
restriction endonucleases and T4 DNA ligase were obtained
Germany) Oligonucleotides were purchased from Operon
Plasmid DNA isolation was performed using a Qiagen spin
miniprep kit (Qiagen GmbH, Hilden, Germany) DNA
was purchased from Invitrogen The pQTev vector, which
is a derivative of pQE60, was purchased from Qiagen
Fatty acids were purchased from Larodan (LARODAN
Fine Chemicals AB, Malmoe, Sweden) All other materials
were purchased from Sigma-Aldrich (Sigma Aldrich Chemie
GmbH, Munich, Germany)
DNA isolation
was isolated using a DNeasy Blood and Tissue kit (Qiagen)
Cloning and expression of DptF
The 270 bp dptF gene was amplified by PCR from S
DNA polymerase (Finnzymes, Espoo, Finland) and primers
dptF-for (5¢-TATGGATCCAACCCGCCCGAAGCGGTC-3¢)
and dptF-rev (5¢-ATAGCGGCCGCGGTGCGGTCGGCC
AACTG-3¢) (underlining indicates artificial BamHI and NotI
restriction sites) The amplified product was purified on a
1.2% agarose gel using a PCR gel extraction kit (Qiagen),
digested with BamHI and NotI, and ligated into the same
sites of the pQTev vector to yield the plasmid pQTev-dptF
The integrity of the plasmid was confirmed by sequencing
The resulting plasmid was used to transform E coli BL21
(DE3) or E coli HM0079 [24] for gene expression The
cul-tures were grown in LB medium supplemented with
attenuance at 600 nm of 0.5, and then the temperature was
addi-tion of 0.1 mm isopropyl thio-b-d-galactoside (IPTG, final
concentration) Cultures were grown for an additional 4 h
Cloning and expression of DptE
The 1795 bp dptE gene was amplified from
high-fidelity Phusion DNA polymerase (Finnzymes) and primers
G-3¢; underlining indicates the sequence overhang for the TOPO cloning) and dptE-rev (5¢-CGCGGGGTGCGGA TGTGGAG-3¢) The amplified product was purified from a 0.8% agarose gel using a PCR gel extraction kit (Qiagen),
accord-ing to manufacturer’s instructions to yield the plasmid
confirmed by sequencing The resulting plasmid was used
to transform E coli BL21 (DE3) for gene expression The
expression was induced by addition of 0.1 mm IPTG (final concentration) Cultures were grown for an additional 4 h
Purification of recombinant expressed proteins DptE and DptF
For purification of DptE and DptF, cell pellets from 1 litre
of culture were resuspended in 10 mL of buffer A (50 mm phosphate buffer, 300 mm NaCl, pH 7.0) and disrupted
G Heinemann Labortechnik, Schwaebisch Gemuend, Ger-many) Insoluble cell debris was removed by centrifugation
was performed on an FPLC system (Amersham Pharmacia Biotechnology, Amersham, UK) according to manufac-turer’s standard protocol Briefly, fractions containing the recombinant proteins were monitored by SDS–PAGE, pooled, and dialysed against phosphate buffer with 100 mm NaCl using HiTrap desalting columns (GE Healthcare Eur-ope GmbH, Freiburg, Germany) The recombinant proteins
Ultra-15 concentrators (Millipore GmbH, Schwalbach, Germany) with a molecular mass cut-off of 10 kDa (DptF) and 50 kDa (DptE) Protein concentrations were determined
Biotechnologie GmbH, Erlangen, Germany) measurements
In vitro 4¢-phosphopantetheinylation of apo-DptF
A reaction mixture containing 200 lm fluoresceinyl CoA or
recom-binant Bacillus subtilis 4¢-phosphopantetheine transferase Sfp in assay buffer (50 mm phosphate buffer, 100 mm
analysed on an SDS–PAGE gel by measuring the in-gel
was generated as previously described [23] The CoASH
Trang 9modification of DptF was verified by ESI-MS using
an LTQ-FT mass spectrometer (Thermo Fisher Scientific,
Bremen, Germany)
ATP-pyrophosphate exchange assay
activity and substrate specificity of DptE For all assays,
the enzyme concentration varied from 300 nm to 1 mm,
and the ATP concentration was at a saturating level of
reactions were performed for 30 s to 1 min Reaction
mixtures contained 50 mm Hepes, pH 8.0, 100 mm NaCl,
DptE (in a final volume of 100 lL) The reaction was
initiated by addition of ATP, 50 lm tetrasodium
Waltham, MA, USA) The reactions were quenched by
acti-vated charcoal, 0.1 m tetrasodium pyrophosphate and
0.35 m perchloric acid Subsequently, the charcoal was
twice with 1 mL water (vortexed for 30 s), and once with
0.5 mL water After addition of 0.5 mL water and 3.5 mL
of liquid scintillation fluid (Rotiscint Eco Plus, CarlRoth
GmbH and Co KG, Karlsruhe, Germany), the
charcoal-bound radioactivity was determined by liquid scintillation
counting using a 1900CA Tri-carb liquid scintillation
analyser (Packard Instruments, Meriden, CT, USA)
Activity assay of DptE with CoASH
thioesterification of a fatty acid with CoASH In this study,
we showed that DptE was not able to react with CoASH
as a substrate However, a typical reaction mixture
(100 lL) was composed of 50 mm phosphate buffer,
500 lm decanoic acid, 1% dimethylsulfoxide and 1 mm
stopped with 10 lL formic acid The product formation
was measured by HPLC-MS Separation of the reaction
Germany) by applying the following gradient at a flow rate
5 min linear gradient up to 95% buffer B in 37 min, and
then holding 100% buffer B for 5 min The product was
identified by UV detection at 215 nm and by on-line
ESI-MS analysis with an Agilent 1100 MSD (Agilent
Technologies Deutschland GmbH, Boeblingen, Germany)
in the negative single ion monitoring (SIM) mode
DptE-mediated transfer of long-chain fatty acids
to holo-DptF
For this reaction, we used holo-DptF that was heterolo-gously expressed in sfp-containing E coli HM0079 cells A
dimethyl-sulfoxide and 1 lm DptE in a total volume of 25 lL
50 mm phosphate buffer (pH 7.0) with 100 mm NaCl After
quenched by addition of 7.5 lL formic acid and directly analysed by mass spectrometry using an LTQ-FT instru-ment (Thermo Fisher Scientific), with desalting using an
solvents used were water with 0.05% formic acid and
95% buffer A for 3 min, followed by a linear gradient down to 60% buffer A in 15 min, followed by a linear gradient down to 5% buffer A in 2 min 5% buffer A was held for an additional 2 min and followed by a linear gradient up to 95% buffer A in 6 min
Determination of the acyl adenylate intermediate LC-MS approach
To identify decanoic AMP by LC-MS, reactions (100 lL)
(10 mm), apo-DptF (50 lm), 1% dimethylsulfoxide and phosphate buffer (pH 7.0, 50 mm) were performed at
and stopped after 1 h by addition of 30 lL formic acid Samples were analysed by HPLC-MS as described above
ATP/PPi-exchange approach
For activity measurements, DptE (1 lm) was rapidily mixed
stopped after 60 min by addition of 500 lL stop mix Sam-ples were washed and analysed as described above
Determination of the kinetic parameters of DptF lipidation by DptE
To determine the kinetic parameters for holo-DtpF lipida-tion by DptE, we performed the reaclipida-tions under ACP satu-ration and varied the fatty acid concentsatu-rations Depending
on the substrate, we varied the reaction time between 30 s
Trang 10and 4 min The reactions were carried out at 37C in a
total volume of 25 lL Unless otherwise indicated, the
reaction mixtures contained 50 mm phosphate buffer,
DptE, 50 lm holo-DptF, 1% dimethylsulfoxide and various
concentrations of fatty acids (10–250 lm) The reactions
were stopped by the addition of 7.5 lL acetic acid The
conversion rate of holo-DptF into fatty acyl S-DptF was
analysed by LC-ESI-MS as described above The
were determined using nonlinear regression with sigmaplot
8.0 (Systat Software GmbH, Erkrath, Germany) to fit the
data to the Michaelis–Menten equation
Determination of DptE specificity towards other
ACPs
with pQE60-acpK) [43] were a gift from H D Mootz
(Fachbereich Chemische Biologie, Technische Universita¨t
Dortmund, Germany) The expression The E coli HM404
kanamy-cin, induced, harvested and disrupted, and the crude cell
extract was centrifuged as described above for DptF
Pro-tein purification was performed as described previously
culture (Fig 7)
The lipD gene was amplified from genomic DNA using
Phusion DNA polymerase (Finnzymes) and the synthetic
oligonucleotide primers 5¢-AAAAAAGAATTCATGTCA
GACCTCAGCACCGC-3¢ and 5¢-AAAAAAAGCTTTCA
GGCGGAACGCAGCTC-3¢ (EcoRI and HindIII
restric-tion sites are underlined) The resulting 291 bp PCR
frag-ment was purified, digested with EcoRI and HindIII, and
ligated into a pET28a(+) derivative (Novagen, Merck
KGaA, Darmstadt, Germany), digested with the same
enzymes The identity of the resulting plasmid
DNA sequencing
The plasmid was used to transform E coli strain
BL21(DE3) (Novagen) and the enzyme was overproduced
was induced by the addition of IPTG to a final
concentra-tion of 0.1 mm The cultures were incubated for a further
18 h After harvesting by centrifugation (6500 g, 15 min,
NaCl, purification of the recombinant protein was
per-formed as previously described [44] Fractions containing
LipD (11.1 kDa) were identified by 15% SDS–PAGE
using HiTrap desalting columns (Amersham Pharmacia
coefficient at 280 nm The yield of purified protein was
apo-LipD was performed as described above for apo-DptF
car-ried out in BL21(DE3)-pRep4-gsp The transfer assays to holo-LipD and holo-AcpK and determination of the kinetic parameters for the DptE-mediated transfer to holo-LipD were performed as described above (see Determination of the kinetic parameters of DptF lipidation by DptE)
Acknowledgements
We thank Dr Georg Scho¨nafinger, Dr Christoph Mahl-ert and Thomas Knappe (Department of Chemistry⁄ Biochemistry, Philipps-University Marburg, Germany) for helpful discussions and critical comments on the manuscript Dr Henning D Mootz provided the HM0079 and HM404 strains This work was supported
by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie
References
1 Miao V, Coeffet-Legal MF, Brian P, Brost R, Penn J, Whiting A, Martin S, Ford R, Parr I, Bouchard M et al (2005) Daptomycin biosynthesis in Streptomyces roseosp-orus: cloning and analysis of the gene cluster and revision
of peptide stereochemistry Microbiology 151, 1507–1523
2 Hojati Z, Milne C, Harvey B, Gordon L, Borg M, Flett
F, Wilkinson B, Sidebottom PJ, Rudd BA, Hayes MA
engi-neered biosynthesis of calcium-dependent antibiotics from Streptomyces coelicolor Chem Biol 9, 1175–1187
3 Fukuda DS, Debono M, Molloy RM & Mynderse JS (1990) A54145, a new lipopeptide antibiotic complex: microbial and chemical modification J Antibiot 43, 601–606
4 Miao V, Brost R, Chapple J, She K, Gal MF & Baltz RH (2006) The lipopeptide antibiotic A54145 biosynthetic gene cluster from Streptomyces fradiae
J Ind Microbiol Biotechnol 33, 129–140
5 Heinzelmann E, Berger S, Muller C, Hartner T, Poralla
K, Wohlleben W & Schwartz D (2005) An acyl-CoA dehydrogenase is involved in the formation of the Delta cis3 double bond in the acyl residue of the lipopeptide antibiotic friulimicin in Actinoplanes friuliensis Microbi-ology 151, 1963–1974
6 Vertesy L, Ehlers E, Kogler H, Kurz M, Meiwes J, Sei-bert G, Vogel M & Hammann P (2000) Friulimicins: novel lipopeptide antibiotics with peptidoglycan synthe-sis inhibiting activity from Actinoplanes friuliensynthe-sis sp
Anti-biot 53, 816–827