PqsL uses reduced flavin to produce 2-hydroxylaminobenzoylacetate, a preferred PqsBC substrate in alkyl quinolone biosynthesis in Pseudomonas aeruginosa

JBC Papers in Press. Published on April 18, 2018 as Manuscript RA117.000789

The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.RA117.000789

PqsL uses reduced flavin to produce 2-hydroxylaminobenzoylacetate, a preferred

PqsBC substrate in alkyl quinolone biosynthesis in Pseudomonas aeruginosa

Steffen Lorenz Drees^1*, Simon Ernst¹, Benny Danilo Belviso², Nina Jagmann¹, Ulrich

Hennecke³and Susanne Fetzner^1*

From the ¹Institute for Molecular Microbiology and Biotechnology, University of Münster, D-48149

Münster, Germany, ²Institute of Crystallography, Consiglio Nazionale delle Ricerche, 70126 Bari, Italy,

³Organic Chemistry Institute, University of Münster, D-48149 Münster, Germany

Running title: Function of PqsL in AQNO biosynthesis

*To whom correspondence may be addressed: Susanne Fetzner: Institute for Molecular Microbiology and

Biotechnology, University of Münster, D-48149 Münster, Germany; fetzner@uni-muenster.de; Tel.: +49

251 8339824; Fax: +49 251 8338388.

Steffen Lorenz Drees: Institute for Molecular Microbiology and Biotechnology, University of Münster,

D-48149 Münster, Germany; s.drees@uni-muenster.de; Tel.: +49 251 8339835; Fax: +49 251 8338388.

Keywords: Pseudomonas aeruginosa, biosynthesis, secondary metabolism, quorum sensing, crystal struc-

ture, 2-alkyl-4-hydoxyquinoline-N-oxides, p-hydroxybenzoate 3-hydroxylase, monooxygenase, flavin.

downstream

enzyme

PqsBC.

In

vitro

ABSTRACT

Alkyl

hydroxyquinoline

N-oxides

reconstitution of the PqsL/PqsBC reaction was

feasible by using the FAD reductase HpaC, and

we noted that the AQ:AQNO ratio is increased in

an hpaC-deletion mutant of P. aeruginosa PAO1

compared with the ratio in the wild-type strain. A

structural comparison with pHBH, the model

enzyme of class A flavoprotein monooxygenases,

revealed that structural features associated with

NAD(P)H binding are missing in PqsL. Our study

completes the AQNO biosynthetic pathway in P.

aeruginosa, indicating that PqsL produces the

unstable product 2-hydroxylaminobenzoylacetate

from 2-ABA and depends on free reduced flavin

as electron donor instead of NAD(P)H.

(AQNOs) are antibiotic compounds produced by

the opportunistic bacterial pathogen Pseudomonas

aeruginosa. They are products of the alkyl

quinolone (AQ) biosynthetic pathway, which also

generates the quorum sensing molecules 2-heptyl-

4(1H)-quinolone (HHQ) and 2-heptyl-3-hydroxy-

4(1H)-quinolone (PQS). Although the enzymatic

synthesis of HHQ and PQS had been elucidated,

the route by which AQNOs are synthesized

remained elusive. Here, we report on PqsL, the

key enzyme for AQNO production, which

structurally resembles class

A

flavoprotein

monooxygenases such as p-hydroxybenzoate 3-

hydroxylase (pHBH) and 3-hydroxybenzoate 6-

hydroxylase. However, we found that unlike

related enzymes, PqsL hydroxylates a primary

aromatic amine group, and it does not use

NAD(P)H as cosubstrate, but unexpectedly

required reduced flavin as electron donor. We also

observed that PqsL is active toward 2-

aminobenzoylacetate (2-ABA), the central

intermediate of the AQ pathway, and forms the

The opportunistic pathogen Pseudomonas

aeruginosa produces numerous 2-alkyl-4-hy-

droxyquinoline-type secondary metabolites (AQ),

which differ in the length and degree of saturation

of the alkyl side chain, the substituent (–H or

–OH) at the C3 position, and the oxidation state of

the quinoline nitrogen (1). AQs exert various and

unstable

hydroxylaminobenzoylacetate,

preferred over 2-ABA as substrate of the

compound

2-

was

often multiple

biological

activities

and

which

significantly contribute to virulence as well as

competitiveness within bacterial communities (2).

1

Function of PqsL in AQNO biosynthesis

While

2-heptyl-3-hydroxy-4(1H)-quinolone

family PS01304, 17). Structure analyses suggested

(Pseudomonas quinolone signal, PQS) and its

immediate biosynthetic precursor 2-heptyl-4(1H)-

quinolone (HHQ) mainly serve as quorum sensing

signals, contributing to the control of virulence

similarity

to

class

A

flavoprotein

monooxygenases. The prototype enzyme of this

class is p-hydroxybenzoate-3-monooxygenase

(pHBH, 18). Class A enzymes depend on NADPH

or NADH, contain a tightly bound FAD cofactor,

are encoded by a single gene, and feature a gluta-

thione reductase fold comprising one dinucleotide

binding domain (Rossman fold). The oxygenating

flavin species, the C(4a)-hydroperoxyflavin, per-

forms an electrophilic attack on the substrate,

which for the majority of the known members of

this class is an aromatic compound containing a

hydroxyl or amino group. Many members of this

flavoprotein monooxygenase class are involved in

the microbial degradation of aromatic compounds.

However, several halogenases taking part in

biosynthetic pathways of antibiotics also show

structural resemblance to pHBH (19). They

require reduced FAD, produced by a reductase, as

well as molecular oxygen for catalysis, and have

been classified within the class F flavoprotein

monooxygenases (18). Class F monoxygenases

were suggested to have a class A-like ancestor

(20).

gene

expression

(3,

4),

2-heptyl-4-

hydroxyquinoline-N-oxide (HQNO) acts as toxin

against both, microorganisms and eukaryotic cells.

HQNO binds to the quinone reduction (Q_i) site of

the respiratory cytochrome bc₁complex (5), and

also inhibits the quinol oxidase activity of cyto-

chrome bo₃and bd (6, 7) and quinone reduction by

type II NADH-quinone oxidoreductase (8). Expo-

sure of Staphylococcus aureus to HQNO selects

for small-colony variants, which are known for

long persistence during chronic infections. HQNO,

together with siderophores, plays a key role in

reduction of S. aureus viability by P. aeruginosa

(9). However, HQNO also causes self-poisoning

of P. aeruginosa, finally leading to autolysis,

which potentially is advantageous to the bacterial

population, promoting biofilm formation by the

surviving cells (10). HQNO and its nonyl con-

gener are among the most abundant AQ

compounds produced by P. aeruginosa (1, 11, 12).

2-Alkyl-4-hydroxyquinoline-N-oxide

In this study, we complete the HQNO bio-

(AQNO) biosynthesis involves the enzymes en-

coded by the pqsABCDE operon (PA0996-

PA1000), which mediate 2-alkyl-4-hydroxyquino-

line biosynthesis from anthranilic acid and acti-

vated fatty acids (Figure 1), and additionally re-

synthetic pathway by characterizing the PqsL reac-

tion, which produces the instable product 2-

hydroxylaminobenzoylacetate (2-HABA) from 2-

ABA, and unexpectedly depends on free reduced

flavin as electron donor in lieu of NAD(P)H. By

re-analyzing the available crystallographic data

and comparing them with the pHBH structure, we

found indication that key residues involved in

NAD(P)H recognition are lacking in PqsL, which

supports our biochemical data. We also identified

PqsBC as the enzyme which catalyzes the conden-

sation of 2-HABA with activated octanoate to pro-

duce HQNO. We propose that the preference of

PqsBC for 2-HABA over 2-ABA prevents

accumulation of the highly reactive N-aryl

hydroxylamine compound.

quires PqsL,

monooxygenase encoded by

a

putative flavin-dependent

presumably

a

monocistronic gene (PA4190) (1, 13). Because P.

aeruginosa supplemented with deuterated HHQ

failed to produce labeled HQNO, the N-oxides

most probably are not synthesized by

hydroxylation of 2-alkyl-4-hydroxyquinolines

(14). However, feeding of 2-aminobenzoylacetate

(2-ABA; Figure 1) to a pqsA⁻/pqsH⁻mutant

(which is unable to produce HHQ and to convert

exogenously supplied HHQ to PQS) resulted in

formation of HQNO along with HHQ (15),

suggesting that AQNO biosynthesis branches off

from 2-ABA or an immediate downstream

intermediate.

Results and Discussion

Purification of PqsL

The crystal structure of the PqsL protein

PqsL, the key enzyme for biosynthesis of

(PDB code: 2X3N) was determined by Oke et al.

(16) within the frame of a high-throughput crystal-

lization screen, but the functional properties of the

protein have not been investigated. PqsL is a

member of the UbiH family (Prosite/Expasy

HQNO, catalyzes the N-oxidation of

a

biosynthetic precursor molecule other than HHQ

(14). Despite long-standing evidence for an

alternative biosynthetic route, the substrate of

PqsL was unknown and its reaction had not been

2

Function of PqsL in AQNO biosynthesis

characterized so far. We purified PqsL as fusion

protein with an N-terminal His- or Strep-tag as

well as a C-terminal His-tag, in order to identify

its most active form. Cofactor content analysis by

regression of UV/Vis spectra showed that only

between 75-82 % of the PqsL protein carried an

FAD cofactor, irrespective of the type and position

of the affinity tag. However, FAD occupancy

could be improved to over 95 % by incubating the

protein with equimolar concentrations of FAD and

subsequent size exclusion chromatography. The

protein was stable for 24 h at 4°C. The purified

recombinant proteins did not show any differences

in activity or substrate selectivity. Because the N-

terminally His-tagged protein was easiest to pro-

duce at high yields (>20 mg l^-1of culture), it was

used for the experiments described in the

following sections.

produced by a single PqsL turnover, was observed

besides HHQ. Without pre-reduction of the flavin

cofactor of PqsL, HQNO was not detected. We

hence concluded that, once FAD is reduced, 2-

ABA or a hypothetical intermediate formed by

PqsBC is the substrate oxidized by PqsL.

Remarkably, addition of anoxically re-

duced free FAD (assay concentration between 20

and 200 µM, approx. 90% total FADH₂content) to

a mixture of 10 µM PqsL, 1 µM PqsBC and 50

µM of substrates 2-ABA and octanoyl-CoA,

instantaneously mixed in an air-saturated buffer,

also led to formation of HQNO (< 10 µM),

indicating that PqsL is able to receive electrons

from reduced FAD. Alternatively, PqsL might

exchange its internal FAD cofactor for FADH₂,

however, since the reaction could be supported by

reduced FMN as well (albeit with lower

efficiency), we believe that a cofactor exchange is

rather unlikely.

In vitro activity of PqsL toward 2-ABA with

chemically reduced FAD

Initially, we tested NADH or NADPH oxi-

Redox potential of PqsL is sufficiently positive

for direct reduction by free flavin

dation by PqsL in the presence of 2-ABA and

other available intermediates of the AQ pathway

as possible substrates (see Figure 1). Despite

trying an array of reaction conditions, substrate-

stimulated NAD(P)H oxidation was not detected.

Interestingly, in contrast to homologous group A

flavoprotein monooxygenases such as pHBH,

flavin reduction by NADH or NADPH under

anoxic conditions was not observed for PqsL,

irrespective of the presence of potential substrates.

In pHBH, absence of the substrate pHB was

reported to diminish the FAD reduction rate by a

factor of 10⁵, but nevertheless FAD reduction was

measurable within minutes (21). Thus, even if

PqsL used another organic substrate (such as the

hypothetical 1-(2-aminophenyl)decane-1,3-dione,

see Figure 1), FAD reduction by the reducing co-

substrate would have been expected to occur.

Therefore, we attempted to identify the organic

substrate of PqsL in a co-substrate-independent

approach. First, the PqsL-bound FAD was pre-

reduced under anoxic conditions with sodium

dithionite until the yellow color was almost

bleached out. Then, we performed an assay com-

bining PqsL with the PqsBC reaction in which the

expected reaction products HHQ and/or HQNO

were quantified by HPLC. When dithionite-

reduced PqsL was mixed with an excess of 2-ABA

and added to air-saturated buffer containing 1 µM

PqsBC and an excess of octanoyl-CoA, HQNO,

To investigate the electrochemical proper-

ties of the PqsL-internal flavin, we determined its

two-electron midpoint potential. The potential of

free FAD (-210 mV, 22, 23, 24) can be severely

affected by protein binding. Based on available

data, it can be generalized that this potential seems

to shift to slightly more positive values of

about -180 to -200 mV in class

monooxygenases (22, 25,

A

26).

Thermodynamically, this small gap in potential

would suffice to facilitate electron transfer from

free FADH₂/FADH^-to protein-bound FAD. In

light of the peculiarities of PqsL regarding co-sub-

strate utilization, the protein may have adapted to

utilization of reduced flavin as electron donor.

For determining the two-electron midpoint

potential of the PqsL-internal FAD, we used an as-

say based on the method published by Massey

(22). In brief, xanthine oxidase is used to reduce

PqsL as well as a reference dye with known redox

potential. From ratios between oxidized and

reduced species, the redox potential of the PqsL-

internal flavin can be derived. Of three reference

dyes tested, only indigo carmine (E_m7=-125 mV)

was suitable for the competitive reduction

experiment whereas phenosafranine (E_m7=-252

mV) and anthraquinone (E_m7=-184 mV) potentials

were too negative. Exemplary spectra including

the respective least squares fits are shown in

3

Function of PqsL in AQNO biosynthesis

Figure 2 A. Spectra of oxidized and reduces

species of PqsL and indigo carmine used for the fit

are shown in Fig. S1. The kinetics of the reaction

is displayed in panel B. Based on the linear fit

within the Nernst plot shown in Figure 2 C a mid-

point potential of -140 ± 4 mV was calculated for

PqsL (not taking in account slight deviations due

to pH). Hence, PqsL features a rather positive

midpoint potential compared to other class A

monooxygenases. While it is debatable whether

this could be interpreted as an adaptation to an

alternative electron donor, it can certainly be

concluded that, electrochemically, efficient

reduction of PqsL by reduced flavin is feasible.

and releases the FADH¯ product into solution

(31). Remarkably, the activity of HpaC of P.

aeruginosa is not influenced by the monooxy-

genase substrate pHPA, (31), as this is the case in

the HpaAC homologs of P. putida, Acinetobacter

baumannii or E. coli (32–34). This property makes

HpaC an ideal supporting enzyme for in vitro

characterization of the PqsL reaction.

HpaC was purified as a pale yellow

protein with approx. 1% of protein molecules

carrying

a

flavin cofactor according to

spectroscopic analysis. Its reductase activity,

monitored by following NADH oxidation in a

spectrophotometric assay, was consistent with

dpreviously reported data (31). We first

investigated whether HpaC was capable of

reducing the PqsL-bound FAD. To this end,

anoxic solutions of PqsL and HpaC (10,000:1

molar ratio) were mixed and the flavin bleaching

upon addition of NADH was monitored at 448 nm.

The instantaneous reduction of PqsL observed

suggests that either HpaC is able to directly reduce

the PqsL-bound FAD, or residual free reduced

FAD serves as catalyst for a very efficient

diffusion-based reduction.

The FAD reductase HpaC supports PqsL activity

in the coupled PqsL/PqsBC assay

Due to its sensitivity to oxidation, chemi-

cally prepared FADH₂as electron donor for PqsL

is highly inefficient and inconvenient for use in

enzyme assays. We therefore searched for other

ways to promote the PqsL reductive half-reaction.

It turned out that in an assay where PqsL, PqsBC,

2-ABA, octanoyl-CoA and NADH were provided,

addition of either cell-free extracts, or cell-free ex-

tracts depleted of dissolved low molecular weight

substances, of either P. aeruginosa PAO1 or E.

coli DH5α led to the formation of small amounts

of HQNO (< 5 µM). This suggested that a protein

rather than an additional low molecular compound

supported PqsL activity, either by transferring

electrons from NADH to PqsL, or – more likely –

by providing free reduced FAD. Such a reaction

scheme has previously been reported for flavin-de-

pendent halogenases which have structural

In accordance with the observed reduction

of PqsL, small quantities of HpaC supported

HQNO formation in the coupled PqsL/PqsBC

reaction (Figure 3). While the reaction showed a

clear dependence on the amounts of each

individual enzyme used, catalytic amounts of FAD

had a positive effect on HQNO yields. In contrast,

excess FAD, particularly in conjunction with

excess HpaC, rendered the reaction inefficient.

This can be explained by oxygen depletion due to

hydrogen peroxide formation out of FADH₂

autoxidation, a notion which is supported by the

observation that catalase, which produces 1 O₂

from 2 H₂O₂, rescued the reaction to some extent

(3 E). Because HQNO yields decreased

considerably when NADH concentrations were re-

duced in the assay, it can be assumed that the reac-

tion is rather inefficient in vitro. This may be due

to the short life time of reduced flavin in the

presence of oxygen (35), and the possibility that

PqsL, similar to other group A monooxygenases,

is incapable of sustaining a reactive flavin in-

termediate in the absence of substrate. In contrast

to, e.g., the oxygenase subunits of two-component

monooxygenases, in which the C(4a)-hydroperox-

yflavin is preserved to some extent, group A

monooxygenases regulate FAD reduction depend-

similarity

to

group

A

flavoprotein

monooxygenases (both share a glutathione reduc-

tase fold) and require the supply of reduced flavin

for enzyme activity (27). In several cases, these

proteins are not associated with a cognate flavin

reductase (19, 28), as observed for two-

component monooxygenases like StyAB or

HpaAC (28–30).

The genome of P. aeruginosa PAO1 con-

tains several predicted flavin reductase genes, but

FAD reductase activity has only been confirmed

for HpaC (PA4092). HpaC is the 18 kDa reductase

protein of the two-component monooxygenase

system HpaAC, which catalyzes the hydroxylation

of p-hydroxyphenylacetate (pHPA), and is

conserved in both, E. coli and P. aeruginosa (29).

HpaC of P. aeruginosa reduces FAD with NADH

4

Function of PqsL in AQNO biosynthesis

ing on the presence of substrate (21, 33). The data

presented in Figure 33 of course raise the question

of how a PqsL-reductase system can function effi-

ciently in vivo. While we cannot offer any

explanation so far, we hypothesize that reduced

flavin may behave differently in the cellular

environment. Moreover, the benefit of a flexible

enzyme system, with a monooxygenase that can

accept reduced FAD as well as FMN from

different reductases, might outweigh the disad-

vantage of increased formation of reactive oxygen

species from free reduced flavins. Low

efficiencies in NADH utilization and high rates of

autoxidation have also been reported for in vitro

type strain and the ΔhpaC mutant, grown in the

presence or absence of pHPA, the substrate of the

HpaAC two-component monooxygenase. As

shown in Fig. S2, supplementation with pHPA in-

duced reductase activity several fold in the wild-

type strain, probably by upregulation of hpaAC

gene expression, but not in the deletion mutant. It

was therefore not surprising that addition of 5 mM

pHPA to the growth medium led to a slight in-

crease in relative AQNO levels in the PAO1

culture (Figure 4). In contrast, the ΔhpaC strain

showed a further decrease, possibly due to the

competition of PqsL and HpaA, the

monooxygenase component of HpaAC, which pre-

sumably is also upregulated in presence of pHPA,

for the remaining free flavin. It has to be

mentioned that the observed differences were not

present when cultures were insufficiently aerated

or when media were not supplemented with the

AQ precursor anthranilic acid, which boosts AQ

productions (41). Hence, while not being evidence

for physiologically relevant dependence of PqsL

on HpaC, the data clearly support the hypothesis

that free reduced FAD is the electron donor to

PqsL in vivo.

assays

with

designated

two-component

monooxygenases and halogenases (29, 32, 33, 36,

37).

Deletion of hpaC in P. aeruginosa PAO1 influ-

ences the HQNO to HHQ ratio

The pqsL gene is not clustered with any

flavin reductase gene. This is not utterly

uncommon for enzymes utilizing reduced flavin as

electron donor, and in the respective cases, the

cognate reductase is unknown (38). Because the in

vitro activity of PqsL was supported by HpaC, we

were interested in whether the hpaC gene of P.

aeruginosa can be linked to AQNO levels in vivo.

Analyzing available transcriptomic data, e.g., from

the NCBI GEO database, no connection between

hpaC and pqsL expression levels was found.

However, because no clear induction of pqsL

under any conditions has been observed, it is

possible that its activity is not primarily regulated

on a genetic level, but rather by substrate/co-

substrate availability.

PqsL releases 2-hydroxylaminobenzoylacetate (2-

HABA), another substrate of PqsBC

The coupled enzyme assay involving PqsL

and PqsBC, described above, did not indicate how

and in which order both enzymes act in forming

HQNO. In order to identify a reaction product re-

leased by one of the enzymes and converted by the

other, PqsL and PqsBC reactions were separated

by a dialysis membrane. Pretests showed that,

despite the 3 kDa MWCO specification, besides

molecular oxygen, only 2-ABA was able to

efficiently pass the membrane (90 % equilibrium

after 15 minutes), whereas NADH, acyl-CoA,

HHQ and protein diffusion was not detected

(diffusion monitored by UV spectroscopy, data not

shown).

Reduced flavins are produced from an

array of enzymes involved in various processes

(reviewed in 42). P. aeruginosa features at least

one additional flavin reducing enzyme whose

function has been proven experimentally, namely,

MsuE (PA3446). However, use of MsuE was pre-

cluded for the in vitro experiments of this study,

because it possesses nitroreductase side activity

(43), which might interfere with the hydrox-

ylamine product of PqsL.

Despite the separation of PqsBC and

PqsL, an average yield of 62 ± 9 % (SEM, 3

experiments) HQNO was observed, suggesting

that a diffusible compound is exchanged between

PqsL and PqsBC. Because this compound is stable

enough to survive transfer through the membrane

for being converted by the downstream enzyme,

the hypothetical PqsBC condensation product 1-

(2-aminophenyl)decane-1,3-dione, whose half-life

was below detection limit (UV spectroscopy,

When growing P. aeruginosa PAO1

ΔhpaC in LB medium supplemented with 2 mM

anthranilic acid, a decrease of AQNOs relative to

AQs was observable compared to the PAO1 wild

type (Figure 4). We then assayed the total FAD

reductase activity of cell free extracts of the wild-

5

Function of PqsL in AQNO biosynthesis

steady state kinetics (42) and unpublished MS and

spectroscopic data), can most likely be excluded.

This leaves the hydroxylated form of 2-ABA as

the only plausible intermediate.

synthesized 2-HABA had the same retention time

as the compound released by PqsL, moreover, UV

spectra were identical (not shown). However, we

were not able to unambiguously identify a mass

corresponding to 2-HABA by HPLC-MS. Using

repeated HPLC injections, it was possible to

estimate its stability (Figure 5 C). Assuming a

first-order reaction, a half-life of 16.2 min was

estimated for both, synthetic and enzyme-

produced 2-HABA. We observed minor amounts

of 2-NBA and 2-ABA (approx. 10 % peak area) as

potential decomposition products, as well as a

number of additional unidentified compounds.

An additional outcome of the dialysis ex-

periment was that, despite compartmentalization,

relatively high yields of HQNO were achieved

(the one-compartment reactions reached close to

60 % only when enzyme and cofactor balance

were optimal, Figure 3 E). The contrary would

have been expected given the fact that osmotic

permeation is a slow process and besides the

hypothetical 2-HABA, 2-ABA would also diffuse

out of the PqsL compartment. A conceivable, but

admittedly speculative explanation would be that

PqsL is inhibited by its own product. With the

hydroxylation product diffusing out of the

compartment, feedback inhibition would be lower

than in a single compartment setup and therefore

lead to higher production of HQNO (which is also

promoted by 2-HABA being preferentially con-

verted by PqsBC, see below). However, given the

complexity of the reaction and the instability of

cofactors and substrates involved, performing

kinetic analyses of the system to assess product

inhibition of PqsL is not feasible.

Because

hydroxyarylamines

are

highly

nucleophilic and can undergo variety of reactions

(45), we suggest that 2-HABA forms various

products, depending on which reaction partners

are available.

For further verification of 2-HABA

authenticity, we employed

a

derivatization

intended to stabilize the hydroxylamine group. It

was demonstrated before that amines and

hydroxylamines can react with acetic anhydride at

0°C and strongly acidic pH to N- and O-acetylated

products, respectively (46). We adopted the

procedure for derivatization of 2-ABA and 2-

HABA, using an alkaline buffer, which acidified

during the reaction due to anhydride

decomposition. This modification was required to

maintain stability of the starting materials. The

reaction and potential reaction products are shown

in Figure 5 D. Although the efficiency of the

reaction was greatly reduced under the conditions

used, we were able to identify compounds with

molecular masses corresponding to those of the

postulated acetylated products. Extracted ion

chromatograms of the positive ion mode MS are

shown in Figure 5 E. The data verified the

presence of the hydroxylamino group in 2-HABA,

however, due to the low efficiency of the derivati-

zation, the procedure is not suitable for

quantitative measurements, such as enzyme

kinetics.

To verify 2-HABA formation, we at-

tempted to track down the direct product of the

PqsL reaction by HPLC. By injecting a sample

containing PqsL, HpaC, NADH and 2-ABA to

HPLC immediately after starting the reaction, an

additional compound could indeed be observed

(Figure 5 A). From the retention time and UV ab-

sorption characteristics (Figure 5 B) it seemed

likely that the compound was a derivative of 2-

ABA. When this compound was isolated from the

HPLC eluate and directly converted by PqsBC

with octanoyl-CoA, HQNO was detected as sole

reaction product, indicating that the compound

must be 2-HABA. For further verification, 2-

HABA was produced by chemical synthesis. To

this end, 2-nitrobenzoylacetate (2-NBA) was

partially reduced using zinc dust under argon at-

mosphere. This approach was described previously

for production of hydroxyarylamines (43, 44).

Reactions were conducted in an aqueous buffer

because the amino-derivative, 2-ABA, rapidly

decomposes in all solvents except for water (42).

Typical yields of 2-HABA were around 30%

according to HPLC analysis (estimated from peak

area at 360 nm), with minor amounts of 2-ABA

formed. As shown in Figure 5 A, the chemically

2-HABA is preferred over 2-ABA by PqsBC

The observation that PqsBC catalyzes the

condensation of octanoyl-CoA with either 2-ABA

or 2-HABA raises the question of which substrate

is preferred by the enzyme. The preference of

PqsBC is a critical factor for the balance between

AQ and AQNO, which are produced in compara-

ble quantities by P. aeruginosa (1, 14). Since it

6

Function of PqsL in AQNO biosynthesis

was impossible to isolate and handle 2-HABA

without significant decomposition, determining

the specificity constant k_catK_mof PqsBC for the

Pseudomonas fluorescens (pdb code 1PBE), the

model enzyme of the group A monooxygenase

subfamily (27). In the first step, we used the PqsL

structure (pdb code 2X3N), which was solved to a

resolution of 1.75 Å by Naismith and coworkers

(16), and remodeled a loop region, which aligns

with a loop reported to be involved in NADPH

and FAD binding in pHBH (Figure 7 A). As

confirmed by the 2Fo-Fc map calculated by using

the published structure factors of 2X3N, the 40-44

loop region of the PqsL was not interpreted in the

first place due to poor electron density (Fig. S3,

A). By using the electron density from the OMIT

map density modification strategy (Fig. S3, B), we

were able to complete the loop (Fig. S3, C),

obtaining structural information on the cofactor

binding mode although some residues, such as

A42, showed incomplete electron densities.

-1

two substrates by steady-state kinetics was not

feasible. An alternative approach was therefore

employed using the combined PqsL-PqsBC assay

in which 2-ABA and 2-HABA (produced by PqsL

in situ) were offered in excess whereas octanoyl-

CoA, the second substrate of PqsBC, was limited.

Whichever substrate is preferred by PqsBC will be

converted until all octanoyl-CoA is used up (reac-

tion scheme displayed in Figure 6 A). Varying the

octanoyl-CoA concentration can therefore provide

an estimate of which compound is the preferred

substrate to PqsBC. Comparable assays were used

for analyzing the conversion of competing

substrates and simulations were undertaken to

understand the kinetics underlying the enzymatic

conversion of competing substrates (47–50).

We set up a system in which ap-

proximately equal initial concentrations of both

substrates were generated in the assay. This was

indicated by the observation that equal amounts of

HHQ and HQNO are formed when a huge excess

of octanoyl-CoA and PqsBC is added to the

reaction. The pattern displayed in Figure 6 B

clearly suggests that conversion of 2-HABA is far

more effective than conversion of 2-ABA. How-

ever, the experimental setup does not allow us to

conclude whether this is due to a higher affinity or

a more efficient turnover.

Interestingly, the PqsL-bound FAD shows

a slight deviation from the planar geometry of the

isoalloxazine in the 2X3N model, which can be

justified by the electron density observed. Such a

distortion could result from X-ray radiation

damage, so that the protein-bound flavin is

reduced in situ to some extent. However, density

modification improved the resolution around the

flavin, so a planar geometry of the isoalloxazine is

conceivable as both, the crystallographic reality

and the native cofactor conformation. The

refinement has been deposited in the protein

database as 6FHO.

With respect to the physiological context,

pHBH is thoroughly studied with respect

the higher efficiency of 2-HABA conversion

would also make sense, because 2-HABA as

to its reaction mechanism, as well as substrate and

co-substrate binding. Despite limited sequence

similarity, its structure is very similar to PqsL

(rmsd 2.4 Å with 394 of 399 residues aligned, 19

% sequence identity). Both proteins share a

glutathione reductase fold with one Rossman fold

for binding FAD. In contrast to other NAD(P)H

utilizing proteins, pHBH-like enzymes do not

feature a designated NAD(P)H binding domain,

but instead use a loop substructure and several

residues on both sides of the FAD binding ridge

for interaction with the co-substrate (18, 56–59).

Both enzymes, PqsL and pHBH, comprise the

fingerprint motif (G₁₂xG₁₄xxG₁₇in PqsL) for

binding the FAD adenine moiety and the highly

conserved DG motif (D₁₅₉G₁₆₀) that was

demonstrated to interact with both, NADPH and

FAD in pHBH (59). PqsL also shares a glutamic

acid (E₃₅), conserved in several pHBH homologs

within the Rossman fold, that interacts with the

primary

aromatic

hydroxylamine

(N-aryl

hydroxylamine) is highly reactive toward DNA,

nucleotides and presumably other biomolecules

(51). N-Aryl hydroxylamines or their metabolic

successors have been demonstrated to possess

genotoxicity in several studies (52–55). The

efficient downstream conversion of 2-HABA may

therefore be important for protecting P.

aeruginosa from cellular damage.

Analysis of the PqsL crystal structure supports a

dysfunctional NAD(P)H binding site

To resolve why PqsL, despite being

closely related to NAD(P)H dependent pHBH-

type monooxygenases, would accept neither co-

substrate, we performed an in-depth analysis of the

crystal structure of PqsL and compared its putative

co-substrate binding region to that of pHBH of

7

Function of PqsL in AQNO biosynthesis

FAD ribose. In contrast, a conserved arginine in

proximity of the adenine, involved in NADPH

binding in pHBH (R₃₃, 57), is altered for a

glutamine (Q₃₆). Additionally, NADPH transiently

interacts with the loop of pHBH that aligns with

the remodeled PqsL loop (58). This substructure is

shorter in PqsL and a helix present within this loop

in pHBH (helix H2, residues 33-40 in pHBH) is

broken up. To compare the key residues for

NAD(P)H binding in pHBH with those in PqsL

structure, the two structures are superposed in

Figure 7 C. A table with the key residues

structurally aligned between the two proteins is

shown in Figure 7 A.

changes, allowing interaction with what appears to

be the phosphoribose of NADPH. Strikingly, R₄₄

of pHBH seems to be substituted with a glycine in

PqsL (G₄₅, Figure 7 C, #5). G₄₅interacts by means

of a peptide bond-aromatic interaction with the

isoalloxazine moiety of FAD (distance between

NH and the center of the carbocyclic aromatic ring

is in agreement with calculations by Bendová et

al., 62). Although this interaction is rather weak, it

is potentially crucial for orienting the flavin at its

position because in pHBH, the FAD open confor-

mation is retained by van-der-Waals interaction of

the flavin with R₄₄(63)_.The PqsL flavin is further-

more stabilized by hydrophobic interactions with

I₂₇₂and P₂₇₃, as well as by hydrogen bonds of the

P₂₉₉and G₃₀₄backbone amide groups to N3 and

the C2 carbonyl, respectively. Notably, the flavin

conformation found in PqsL cannot be classified

in analogy to the open and closed orientations

found in pHBH. Structural alignments suggest that

the conformation found in PqsL rather resembles

an ‘open’ state, which is rather unusual for a group

A monooxygenase without substrate bound (See

Fig. S4). However, there is no apparent space for a

conformational change within the structure for the

flavin to move into a ‘closed’ conformation, as

this is the case in pHBH (63, 64). Moreover, due

to the above-mentioned differences, the FAD

binding pocket is widened and the flavin is more

exposed than in related enzymes. Lacking

coordination, the FAD possesses a higher degree

of conformational freedom, possibly for enhancing

electron uptake from reduced flavin.

Two residues in the loop were demon-

strated to be essential for the reductive half-

reaction in pHBH as well as for NADPH

specificity, R₄₂and Y₃₈(57, 60). While R₄₂

interacts with the ribosyl alcohols of the FAD

adenosine nucleotide and supposedly is mandatory

for positioning the FAD, Y₃₈may directly or

indirectly interact with NADPH (58, 61). It is

therefore not surprising that the ribosyl-

coordinating R₄₂is functionally conserved in PqsL

(R₄₁), whereas Y₃₈is not. Notably, R₄₁of PqsL has

a different conformation than R₄₂of pHBH and

does not align with R₄₂, but with Y₃₈of pHBH

(Figure 7 C, #1 and #2). It is therefore ques-

tionable whether R₄₁of PqsL, despite its different

position within the loop, is capable of supporting

NAD(P)H binding via 2’-phosphate recognition in

a similar way as it has been reported for pHBH

(57). Figure 7 C illustrates the conformational

differences in both structures that could be

resolved by reevaluation of the 2X3N crystal

structure. The remodeling of the loop with the

respective electron density maps is shown in Fig.

S3.

Taken together, the available structural in-

formation on PqsL supports the hypothesis that the

enzyme is not capable of binding NAD(P)H in the

alleged mode because several of the crucial posi-

tions are replaced by dysfunctional substitutions.

However, provided that PqsL indeed binds

reduced FAD in lieu of NAD(P)H, instead of

exchanging its cofactor, it remains to be elucidated

how PqsL facilitates electron transfer and whether

different flavin conformations are involved in

controlling the catalytic reaction.

Located on the opposite side of the FAD

binding cleft, R₂₆₉and H₁₆₂of pHBH were proven

to be essential for NADPH oxidation while not af-

fecting hydroxylation. These two residues align

with A₁₆₄and G₂₇₅in PqsL (Figure 7 C, #3 and 4),

which are unable to support the same interactions

(Figure 7 C, #3 and 4).

Unfortunately, pHBH is the only group A

Crystallographic studies of the pHBH-

monooxygenase that has been extensively studied

with respect to co-substrate recognition. The se-

quence diversity among even closely related en-

zymes of the subclass and the rather few determi-

nants for co-substrate binding calls for additional

studies employing other enzymes in order to

recognize recurring binding patterns. There might

NADPH complex moreover had illustrated the role

of R₄₄for both cofactor and NADPH binding (58).

Without substrates bound (closed conformation), it

forms hydrogen bonds with the diphosphate and

with ribityl hydroxyls of the flavin nucleotide. In

the co-substrate-bound form its conformation

8

Function of PqsL in AQNO biosynthesis

even be flavin monooxygenases with promiscuous

co-substrate utilization, utilizing both reduced fla-

vin and NAD(P)H. It is known that flavin

dependent monooxygenases can be very efficient

catalysts, e.g., > 50 turnovers are reached by

pHBH (21, 56). While this is desirable in degrada-

tion processes, in case of PqsL the balance

between AQ and AQNO pathway products would

be disturbed by excess PqsL activity. Whereas

intracellular pools of NADH and NADPH are

readily accessible and millimolar cellular

concentrations were determined for E. coli (65),

reduced flavin, due to its potential to generate

reactive oxygen species, is not relied on as a

default reducing agent in the cell and probably

much less available. Little is known about the reg-

ulation and total concentrations of intracellular re-

duced flavin (39), still, it is conceivable that its use

as an alternative electron donor is more

widespread than expected so far.

20 mM Tris-Cl, pH 8, 150 mM NaCl and 0.1%

NP40, and sonicated for 4 × 2 min. Lysates were

centrifuged (40 min, 25,000 × g, 4°C) and filtered

through a 0.45 µm PVDF syringe filter (Carl Roth,

Karlsruhe, Germany). PqsL was purified by means

of either a hexahistidine or a StrepII tag fused to

the N-terminus. Immobilized metal affinity

chromatography was carried out with 1 ml

HisTrap FF columns and Äkta Prime plus au-

tomated chromatography units for gradient elution

(both GE Life sciences). Purification of StrepII-

tagged PqsL was performed using a 1 ml Strep-

Tactin Sepharose column (IBA life sciences,

Göttingen, Germany) with a washing buffer

containing 20 mM Tris-Cl, pH 8 and 150 mM

NaCl. Elution was conducted using 2.5 mM d-

desthiobiotin in the same buffer. Subsequent to

affinity chromatography, the prepared proteins

were buffer-exchanged into 20 mM HEPES, pH 8

and 50 mM NaCl using GE HiTrap desalt

columns, and concentrated to more than 10 mg ml^-

1

with 10 kDa MWCO centrifugal concentrators

Experimental Procedures

(Sartorius, Göttingen, Germany). HpaC was

purified with a C-terminal hexahistidine tag

according to (31); PqsD, PqsE and PqsBC were

purified as previously reported (42, 66). For all

proteins, 10% glycerol was added as

cryoprotectant before flash-freezing protein ali-

quots in liquid nitrogen for later use.

Preparation of cell extracts and protein purifica-

tion

Cell extracts were produced from E. coli

DH5α, P. aeruginosa PAO1 or P. aeruginosa

PAO1 ΔhpaC by sonicating 1 g cells ml^-1for 4 × 2

min in 50 mM Tris-Cl, pH 7.5. Lysates were

cleared by centrifugation (25,000 × g, 4°C, 40

min), filtered through 0.22 µm filters (VWR,

Radnor, PA, USA), and flash-frozen in liquid

nitrogen until further use. Alternatively, for

removal of low molecular weight compounds, cell

extracts were run through 5 ml size exclusion col-

umns (HiTrap desalt, containing Sephadex G-25

Superfine, GE, Pittsburgh, PA, USA) equilibrated

with 50 mM Tris-Cl, pH 7.5, prior to freezing in

liquid nitrogen.

Cultivation of P. aeruginosa PAO1 and

PAO1 ΔhpaC for FAD reductase assays

P. aeruginosa PAO1 wild-type and the hpaC dele-

tion strain were pre-grown in LB medium before

1:1000 dilution into M9 minimal medium with 10

mM succinate as sole carbon source. 50 ml of

overnight

cultures

were

harvested

and

subsequently resuspended in 50 ml of fresh M9

medium supplemented with either 10 mM

succinate or 0.5 mM succinate and 10 mM pHPA.

Cells were then grown for another 3 hours, har-

vested by centrifugation (6,000 × g, 10 min) and

stored at -80°C until further use.

Proteins

were

heterologously

overproduced in E. coli Rosetta2 pLysS (Merck

Millipore, Billerica, MA, USA), transformed with

pET28b or pET23a expression vectors containing

the respective genes. For production of PqsL and

HpaC, the respective recombinant strains were

grown at 37°C in LB medium to an OD₆₀₀of 0.5,

then cooled down to 20°C for 20 min before

inducing overexpression with 0.5 mM IPTG. After

overnight incubation, cells were harvested by

centrifugation (15 min, 6,000 × g) and frozen at -

80°C until further use. For preparation of cell free

protein extracts, pellets were thawed, suspended in

Synthesis of anthraniloyl-CoA, 2-ABA and 2-

HABA

Anthraniloyl-CoA and 2-ABA were pre-

pared as described previously (42, 66). For prepa-

ration of 2-(hydroxylamino)benzoylacetic acid (2-

HABA), 5 mg of 2-nitrobenzoylacetic acid (2-

NBA), prepared according to (42, 67–69), were

dissolved in 1 ml ice cold 20 mM Tris base

9

Function of PqsL in AQNO biosynthesis

solution (pH 9.5) (Carl Roth). 10 mg of zinc

powder and a magnetic stirrer were added and the

suspension was vigorously stirred for 30 min at

0°C under an argon atmosphere (1 bar

overpressure) for reduction of the nitro group (43,

44). Longer incubation times did not lead to in-

creased yields of 2-HABA. The suspension was

passed through a glass frit and a 0.45 µm filter,

and immediately subjected to preparative HPLC.

Louis, MO, USA and Santa Cruz, Dallas, TX,

USA).

HPLC-MS analysis of 2-ABA, 2-HABA

and their derivatives was performed using a proce-

dure described before for HHQ using a Thermo

Scientific Ultimate 3000/ Bruker amazon speed in-

strument (66). In order to increase the sensitivity

for detecting labile compounds, ESI ionization

parameters were varied according to the manufac-

turer’s recommendations.

Derivatization of 2-HABA and 2-ABA

2-ABA and unstable 2-HABA produced

UV spectroscopic enzyme assays

by PqsL were derivatized according to a procedure

used by Lee and coworkers (46), in which

aromatic amines and hydroxylamines are reacted

with acetic anhydride. The reaction was performed

in 50 mM HEPES, pH 9.5 on ice to slow down

decomposition of the compounds. After addition

of acetic anhydride to a final concentration of 1%

(v/v), the mixtures were kept on ice for another 30

minutes and centrifuged to remove precipitated

protein. After filtration through a 0.45 µm filter,

samples were analyzed with LC-MS.

For

the

detection

of

substrate,

NAD(P)H:oxygen oxidoreductase activity of

PqsL, the oxidation of either NADH or NADPH

was monitored spectrophotometrically at 340 nm

as a function of the potential substrates anthranilic

acid (Sigma-Aldrich), anthraniloyl-CoA, and 2-

ABA. Assays were conducted in 20 mM HEPES

or Tris-Cl, pH 8.0 and contained 0.1 to 20 µM

PqsL, 100 µM of substrate and 100 µM of either

NADH or NADPH (Sigma-Aldrich). Additives

were PqsD, PqsE or PqsBC (up to 10 µM), 100

µM malonyl-CoA, and 100 µM octanoyl-CoA

(both Sigma-Aldrich). Absorbance changes were

measured for up to 30 min (V-550 or V-750

spectrophotometers, Jasco, Tokyo, Japan or

HPLC and HPLC-MS analysis of reaction prod-

ucts

All HPLC experiments were performed

with an Agilent 1100 Series instrument (Agilent,

Santa Clara, CA, USA) equipped with

autosampler, binary pump and diode array

detector. Knauer Eurospher II 4x150 mm C18 and

C18-H columns (Berlin, Germany) were used for

all experiments. For preparation of chemically

synthesized 2-HABA or for analysis of PqsL

reaction products, separation of 2-ABA, 2-HABA,

2-NBA and DHQ was performed using a 20 mM

ammonium carbonate buffer, pH 8.7 as solvent A

and acetonitrile (both Carl Roth) as solvent B.

Elutions were started with 1 minute of isocratic

flow (100% A) before linearly increasing to 20%

B over 12 min. B was then increased to 80% over

5 min. The flow rate was kept constant at 0.8 ml

min^-1.

Evolution

201

UV-Vis-spectrophotometer,

ThermoFisher Scientific, MA, USA ). For anoxic

measurements, assays were prepared in an

anaerobic chamber in a gastight microcuvette

(Hellma Analytics, Mülheim, Germany). PqsL was

chemically reduced by titration with a solution

containing 50 mM Tris-Cl, pH 8, and 1 mM so-

dium dithionite (Sigma Aldrich). Single turnover

experiments were conducted by injecting air-satu-

rated solutions containing the substrate combina-

tions denoted above. Flavin reductase assays with

HpaC were conducted in 20 mM HEPES or Tris-

Cl, pH 8.0, with 50 µM FAD or FMN (Sigma-Al-

drich) as electron acceptor and between 5 and 100

µM NADH. Enzyme concentrations were varied

between 0.1 nM and 1 µM.

HPLC analyses of AQ reaction products

were conducted as described in (66). In brief, sol-

vents were 6 mM aqueous citric acid (A) and

methanol (B) with a gradient from between 10 and

40% B to 100 % B. HHQ and HQNO were

quantified from calibration with the commercially

available authentic compounds, detected at 315

nm and 326 nm, respectively (Sigma Aldrich, St.

Determination of the redox potential of PqsL-

bound FAD

The redox potential of enzyme-bound

FAD was determined employing xanthine oxidase

as described in (22). Xanthine oxidase oxidizes

xanthine to uric acid and in the absence of oxygen

provides electrons to redox dyes with a potential

more positive than -410 mV. In the presence of

10

Function of PqsL in AQNO biosynthesis

two or more redox dyes, the more positive ac-

ceptor is favored. Combining PqsL and a dye with

known potential, the redox potential of PqsL can

be calculated according to the Nernst relation.

Initial experiments indicated that PqsL had very

limited stability in the assay. We therefore

accelerated the reaction by increasing the

concentration of xanthine oxidase in the assay,

and employed multivariate data analysis using

component spectra to quantify reduced and oxi-

dized states of PqsL and the respective dye. As-

says contained 20 µM PqsL, different redox dyes

(Phenosafranine, E_m7= -252 mV, Anthraquinone-

2,6-disulphonate, E_m7= -184 mV and Indigo car-

mine, E_m7= -125 mV) in varying concentrations to

give an absorbance between 0.1-0.5, 5 mM

glucose, 10 µg ml^-1glucose oxidase, 5 µg ml^-1

catalase, 200 µM xanthine, 2 µM benzyl viologen

and 200 nM xanthine oxidase (all Sigma-Aldrich,

St. Louis, MO, USA) in 25 mM Tris-Cl, pH 8.0.

Samples were set up in a sealable cuvette in an an-

aerobic chamber and incubated for 10 min at room

temperature before starting the reaction by adding

3 µL pre-diluted xanthine oxidase. Spectra were

recorded every 15 s over 10 minutes at a scan rate

of 6000 nm min^-1. Due to overlapping absorption

peaks, concentrations of oxidized and reduced dye

and PqsL were determined by least squares fitting.

Pre-recorded spectra of PqsL and the respective

dyes in fully oxidized and reduced states served as

composites for the fit. Calculations were

performed with Matlab R2016B (The Mathworks,

Natick, MA, USA)

FAD. To this end, protein concentrations were de-

termined using a modified Bradford assay (70) and

extracts were diluted to the same concentration

with a buffer containing 30 mM Tris, pH 8, and

150 NaCl and stored on ice until used.

To minimize reoxidation of the reduced

FAD, reactions were carried out in air-tight

cuvettes and nitrogen-purged buffers and solution.

Cell extracts were diluted into a buffer containing

30 mM Tris, pH 8, 150 mM NaCl, 1% (w/v)

glucose, 25 U ml^-1glucose oxidase and 250 U ml^-1

catalase. 100 µM of FAD was added and the

mixture was incubated at room temperature for 5

minutes. Reactions were started by addition of 1

mM NADH. FAD reduction was measured over 2

minutes at room temperature with a Jasco J-750

spectrophotometer. Initial rates were converted to

specific activities using the extinction coefficient

of FAD_ox-red(ε_{450 nm}= 10,600 M^-1cm^-1, (71).

Construction of the hpaC deletion mutant of P.

aeruginosa PAO1

DNA manipulations were performed ac-

cording to standard protocols. Genomic DNA of

P. aeruginosa PAO1 was purified with the

Puregene Tissue Core Kit B (Qiagen, Hilden, Ger-

many). Two PCR products spanning parts of the

up- and downstream regions of hpaC were

obtained from genomic DNA of strain PAO1 with

the primer pairs hpaC Up fw (5’-AAATCTAGA-

CAACCTGTTCTGGTCGCTGA-3’)/hpaC Up rev

(5’-GGGGATAACCTCCGAGGC-3’) and hpaC

Dn fw (5’-CGCCTCGGAGGTTATCCCCAC-

GGCGCCTCCAGGAATCC-3’)/hpaC Dn rev (5’-

ATTAAGCTTCTATCGCCACGGCTTCTCCA-

3’) and used as templates for a second overlapping

PCR (SOE-PCR; 72) with the primers hpaC Up fw

and hpaC Dn rev. The resulting fragment was di-

gested with XbaI/HindIII and ligated with the cor-

responding sites of the suicide vector pEX18Ap.

The resulting plasmid was mobilized into strain

PAO1 by triparental mating with E. coli strain

DH5α as donor and strain HB101 pRK2013 as

helper as described previously (73). Mutants were

selected on Pseudomonas Isolation Agar plates

containing 350 µg ml^-1carbenicillin and

transferred onto LB plates containing 7% sucrose

for selecting for vector excision by a second cross-

over. Single colonies were repeatedly streaked out

and checked for gene deletion by PCR using the

same primer pair as for the SOE-PCR. Finally,

Multi-enzyme assays

Enzyme assays combining the PqsL and

PqsBC reactions were performed using the stable

AQ reaction products HHQ and HQNO as

readouts, quantified by HPLC. PqsBC and PqsL

concentrations were varied between 1 nM and 10

µM, concentrations of substrates 2-ABA and

octanoyl-CoA (Sigma Aldrich), putative substrates

NADH, NADPH, additives FAD, FMN, and

additives catalase, PqsE, PqsD and HpaC were

varied. All assays were conducted in either 30 mM

MOPS, pH 7.3 or 30 mM HEPES, pH 8.2 at 30°C

and 1,000 rpm.

FAD reductase activity of cell extracts

Cell-free extracts were analyzed on FAD

reductase activity with a spectrophotometric assay

measuring the NADH-dependent reduction of

11

Function of PqsL in AQNO biosynthesis

genomic DNA was purified from single mutant

colonies and checked for gene deletion by PCR.

crystallographic suite and the residues thereby

added again according to the Fo-Fc Fourier map,

slightly improving R and R_freevalues (75, 76). The

omit map of the 40-44 loop, obtained by the

Autobuild tool of Phenix crystallographic suite,

was used as map to fit the residues of the loop by

means of the Phenix Fit Loop tool (77, 78).

Finally, Refmac5 was run again to perform the

final refinement of the whole model. Pymol

(Schrödinger, LLC, NY, USA) was used for

alignments as well as analysis and visualization of

the interactions between the protein and the FAD

cofactor.

Refinement and analysis of the PqsL structure

The structure of the PqsL protein

deposited in the Protein Data Bank (PDB code

2X3N) was used to analyze the interaction with its

FAD cofactor. Because the structure lacks the

residues between Glu40 and Asn44 that

presumably play an important role in co-substrate

recognition, the 2X3N protein model and structure

factors were reprocessed. First, the 2Fo-Fc

electron density map was inspected and residues

showing critical mismatches were removed from

the model using Coot software (74). The structure

was then refined using Refmac5 of the CCP4

Acknowledgments: This work was supported by the Deutsche Forschungsgemeinschaft (grant no. FE

383/23-1 and FE 383/23-2 to SF, and HE 6020/3-1 to UH). We thank Prof. B. Philipp (University of

Münster) for access to the electrospray ionization ion trap mass spectrometer (DFG grant INST 211/646-

1).

Conflict of interest: The authors declare that they have no conflicts of interest with the contents of this

article.

Footnotes:

⁴The abbreviations used are:

2-ABA

2-HABA

2-NBA

AQ

2-Aminobenzoylacetate

2-(Hydroxylamino)benzoylacetate

2-Nitrobenzoylacetate

2-Alkyl-hydroxyquinolines

AQNO

DAD

2-Alkyl-4-hydroxyquinoline-N-oxides

Diode array detector

DHQ

2,4-Dihydroxyquinoline

E_m7

Midpoint potential at pH 7

HHQ

2-Heptyl-4(1H)quinolone

HQNO

MWCO

NAD(P)H

pHBH

pHPA

PQS

2-Heptyl-4-hydroxyquinoline-N-oxide

Molecular weight cutoff

NADH or NADPH

p-Hydroxybenzoate-3-monooxygenase

p-Hydroxyphenylacetate

Pseudomonas quinolone signal, 2-heptyl-3-hydroxy-4(1H)-quinolone

Root mean square deviation

rmsd

12

Function of PqsL in AQNO biosynthesis

References

1.

2.

Lépine, F., Milot, S., Déziel, E., He, J., and Rahme, L. G. (2004) Electrospray/mass spectrometric

identification and analysis of 4-hydroxy-2-alkylquinolines (HAQs) produced by Pseudomonas

aeruginosa. J. Am. Soc. Mass Spectrom. 15, 862–869

Cao, H., Krishnan, G., Goumnerov, B., Tsongalis, J., Tompkins, R., and Rahme, L. G. (2001) A

quorum sensing-associated virulence gene of Pseudomonas aeruginosa encodes a LysR-like

transcription regulator with a unique self-regulatory mechanism. Proc. Natl. Acad. Sci. U. S. A. 98,

14613–14618

3.

4.

Pesci, E. C., Milbank, J. B., Pearson, J. P., McKnight, S., Kende, A. S., Greenberg, E. P., and

Iglewski, B. H. (1999) Quinolone signaling in the cell-to-cell communication system of

Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. U. S. A. 96, 11229–112234

Rampioni, G., Falcone, M., Heeb, S., Frangipani, E., Fletcher, M. P., Dubern, J.-F., Visca, P.,

Leoni, L., Cámara, M., and Williams, P. (2016) Unravelling the genome-wide contributions of

specific 2-alkyl-4-quinolones and PqsE to quorum sensing in Pseudomonas aeruginosa. PLoS

Pathog. 12, e1006029

5.

6.

7.

8.

Cooley, J. W., Ohnishi, T., and Daldal, F. (2005) Binding dynamics at the quinone reduction (Qi)

site influence the equilibrium interactions of the iron sulfur protein and hydroquinone oxidation

(Qo) site of the cytochrome bc₁complex. Biochemistry. 44, 10520–10532

Kita, K., Konishi, K., and Anraku, Y. (1984) Terminal oxidases of Escherichia coli aerobic

respiratory chain. I. Purification and properties of cytochrome b₅₆₂-o complex from cells in the

early exponential phase of aerobic growth. J. Biol. Chem. 259, 3368–3374

Yap, L. L., Lin, M. T., Ouyang, H., Samoilova, R. I., Dikanov, S. A., and Gennis, R. B. (2010)

The quinone-binding sites of the cytochrome bo3 ubiquinol oxidase from Escherichia coli.

Biochim. Biophys. Acta. 1797, 1924–1932

Sena, F. V., Batista, A. P., Catarino, T., Brito, J. A., Archer, M., Viertler, M., Madl, T., Cabrita, E.

J., and Pereira, M. M. (2015) Type-II NADH:quinone oxidoreductase from Staphylococcus aureus

has two distinct binding sites and is rate limited by quinone reduction. Mol. Microbiol. 98, 272–

288

9.

Filkins, L. M., Graber, J. A., Olson, D. G., Dolben, E. L., Lynd, L. R., Bhuju, S., and O’Toole, G.

A. (2015) Coculture of Staphylococcus aureus with Pseudomonas aeruginosa drives S. aureus

towards fermentative metabolism and reduced viability in a cystic fibrosis model. J. Bacteriol.

197, 2252–2264

10.

Hazan, R., Que, Y. A., Maura, D., Strobel, B., Majcherczyk, P. A., Hopper, L. R., Wilbur, D. J.,

Hreha, T. N., Barquera, B., and Rahme, L. G. (2016) Auto poisoning of the respiratory chain by a

quorum-sensing-regulated molecule favors biofilm formation and antibiotic tolerance. Curr. Biol.

26, 195–206

11.

12.

Lépine, F., Déziel, E., Milot, S., and Rahme, L. G. (2003) A stable isotope dilution assay for the

quantification of the Pseudomonas quinolone signal in Pseudomonas aeruginosa cultures.

Biochim. Biophys. Acta. 1622, 36–41

Ortori, C. A., Dubern, J.-F., Chhabra, S. R., Cámara, M., Hardie, K., Williams, P., and Barrett, D.

A. (2011) Simultaneous quantitative profiling of N-acyl-L-homoserine lactone and 2-alkyl-4(1H)-

quinolone families of quorum-sensing signaling molecules using LC-MS/MS. Anal. Bioanal.

Chem. 399, 839–850

13.

14.

Stover, C. K., Pham, X. Q., Erwin, A. L., Mizoguchi, S. D., Warrener, P., Hickey, M. J.,

Brinkman, F. S., Hufnagle, W. O., Kowalik, D. J., Lagrou, M., Garber, R. L., Goltry, L.,

Tolentino, E., Westbrock-Wadman, S., Yuan, Y., Brody, L. L., Coulter, S. N., Folger, K. R., Kas,

A., Larbig, K., Lim, R., Smith, K., Spencer, D., Wong, G. K., Wu, Z., Paulsen, I. T., Reizer, J.,

Saier, M. H., Hancock, R. E., Lory, S., and Olson, M. V (2000) Complete genome sequence of

Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature. 406, 959–964

Déziel, E., Lépine, F., Milot, S., He, J., Mindrinos, M. N., Tompkins, R. G., and Rahme, L. G.

(2004) Analysis of Pseudomonas aeruginosa 4-hydroxy-2-alkylquinolines (HAQs) reveals a role

13

Function of PqsL in AQNO biosynthesis

for 4-hydroxy-2-heptylquinoline in cell-to-cell communication. Proc. Natl. Acad. Sci. U. S. A. 101,

1339–1344

15.

16.

Dulcey, C. E., Dekimpe, V., Fauvelle, D. A., Milot, S., Groleau, M. C., Doucet, N., Rahme, L. G.,

Lépine, F., and Déziel, E. (2013) The end of an old hypothesis: The Pseudomonas signaling

molecules 4-hydroxy-2-alkylquinolines derive from fatty acids, not 3-ketofatty acids. Chem. Biol.

20, 1481–1491

Oke, M., Carter, L. G., Johnson, K. a, Liu, H., McMahon, S. a, Yan, X., Kerou, M., Weikart, N.

D., Kadi, N., Sheikh, M. A., Schmelz, S., Dorward, M., Zawadzki, M., Cozens, C., Falconer, H.,

Powers, H., Overton, I. M., van Niekerk, C. a J., Peng, X., Patel, P., Garrett, R. a, Prangishvili, D.,

Botting, C. H., Coote, P. J., Dryden, D. T. F., Barton, G. J., Schwarz-Linek, U., Challis, G. L.,

Taylor, G. L., White, M. F., and Naismith, J. H. (2010) The Scottish Structural Proteomics

Facility: targets, methods and outputs. J. Struct. Funct. Genomics. 11, 167–180

D’Argenio, D. A., Calfee, M. W., Rainey, P. B., and Pesci, E. C. (2002) Autolysis and

autoaggregation in Pseudomonas aeruginosa colony morphology mutants. J. Bacteriol. 184,

6481–6489

17.

18.

19.

20.

van Berkel, W. J. H., Kamerbeek, N. M., and Fraaije, M. W. (2006) Flavoprotein

monooxygenases, a diverse class of oxidative biocatalysts. J. Biotechnol. 124, 670–689

van Pée, K. H., and Patallo, E. P. (2006) Flavin-dependent halogenases involved in secondary

metabolism in bacteria. Appl. Microbiol. Biotechnol. 70, 631–641

Mascotti, M. L., Juri Ayub, M., Furnham, N., Thornton, J. M., and Laskowski, R. A. (2016)

chopping and changing: the evolution of the flavin-dependent monooxygenases. J. Mol. Biol. 428,

3131–3146

21.

22.

23.

24.

Husain, M., and Massey, V. (1979) Kinetic studies on the reaction of p-hydroxybenzoate

hydroxylase. Agreement of steady state and rapid reaction data. J. Biol. Chem. 254, 6657–6666

Massey, V. (1991) A simple method for determination of redox potentials, Flavins and

Flavoproteins (B. Curti, S. Ronchi, G. Z. ed), Walter de Gruyter, Berlin

Crozier-Reabe, K., and Moran, G. R. (2012) Form follows function: structural and catalytic

variation in the class a flavoprotein monooxygenases. Int. J. Mol. Sci. 13, 15601–15639

Palfey, B. A., Ballou, D. P., and Massey, V. (1995) Oxygen activation by flavins and pterins. in

Active Oxygen in Biochemistry, pp. 37–83, Springer Netherlands, Dordrecht, 10.1007/978-94-011-

0609-2_2

25.

26.

Einarsdottir, G. H., Stankovich, M. T., Powlowski, J., Ballou, D. P., and Massey, V. (1989)

Regulation of oxidation-reduction potentials of anthranilate hydroxylase from Trichosporon

cutaneum by substrate and effector binding. Biochemistry. 28, 4161–4168

Williamson, G., Edmondson, D. E., and Müller, F. (1988) Oxidation-reduction potential studies on

p-hydroxybenzoate hydroxylase from Pseudomonas fluorescens. Biochim. Biophys. Acta. 953,

258–262

27.

28.

Huijbers, M. M. E., Montersino, S., Westphal, A. H., Tischler, D., and van Berkel, W. J. H. (2014)

Flavin dependent monooxygenases. Arch. Biochem. Biophys. 544, 2–17

Hammer, P. E., Hill, D. S., Lam, S. T., van Pée, K. H., and Ligon, J. M. (1997) Four genes from

Pseudomonas fluorescens that encode the biosynthesis of pyrrolnitrin. Appl. Environ. Microbiol.

63, 2147–2154

29.

30.

31.

Galán, B., Díaz, E., Prieto, M. A., and García, J. L. (2000) Functional analysis of the small

component of the 4-hydroxyphenylacetate 3-monooxygenase of Escherichia coli W: A prototype

of a new flavin:NAD(P)H reductase subfamily. J. Bacteriol. 182, 627–636

Otto, K., Hofstetter, K., Röthlisberger, M., Witholt, B., and Schmid, A. (2004) Biochemical

characterization of StyAB from Pseudomonas sp. strain VLB120 as a two-component flavin-

diffusible monooxygenase. J. Bacteriol. 186, 5292–5302

Chakraborty, S., Ortiz-Maldonado, M., Entsch, B., and Ballou, D. P. (2010) Studies on the

mechanism of p-hydroxyphenylacetate 3-hydroxylase from Pseudomonas aeruginosa: A system

composed of a small flavin reductase and a large flavin-dependent oxygenase. Biochemistry. 49,

372–385

14

Function of PqsL in AQNO biosynthesis

32.

33.

34.

Sucharitakul, J., Tinikul, R., and Chaiyen, P. (2014) Mechanisms of reduced flavin transfer in the

two-component flavin-dependent monooxygenases. Arch. Biochem. Biophys. 555–556, 33–46

Arunachalam, U., Massey, V., and Miller, S. M. (1994) Mechanism of p-hydroxyphenylacetate-3-

hydroxylase. A two-protein enzyme. J. Biol. Chem. 269, 150–155

Louie, T. M., Xie, X. S., and Xun, L. (2003) Coordinated production and utilization of FADH₂by

NAD(P)H - Flavin oxidoreductase and 4-hydroxyphenylacetate 3-monooxygenase. Biochemistry.

42, 7509–7517

35.

36.

Gibson, Q. H., and Hastings, J. W. (1962) The oxidation of reduced flavin mononucleotide by

molecular oxygen. Biochem. J. 83, 368–377

Xun, L., and Sandvik, E. R. (2000) Characterization of 4-hydroxyphenylacetate 3-hydroxylase

(HpaB) of Escherichia coli as a reduced flavin adenine dinucleotide-utilizing monooxygenase.

Appl. Environ. Microbiol. 66, 481–486

37.

38.

Unversucht, S., Hollmann, F., Schmid, A., and van Pée, K. H. (2005) FADH₂-dependence of

tryptophan 7-halogenase. Adv. Synth. Catal. 347, 1163–1167

Buedenbender, S., Rachid, S., Müller, R., and Schulz, G. E. (2009) Structure and action of the

myxobacterial chondrochloren halogenase CndH: A new variant of FAD-dependent halogenases.

J. Mol. Biol. 385, 520–530

39.

40.

Tu, S. C. (2001) Reduced flavin: donor and acceptor enzymes and mechanisms of channeling.

Antioxid. Redox Signal. 3, 881–897

Green, L. K., Storey, M. A., Williams, E. M., Patterson, A. V, Smaill, J. B., Copp, J. N., and

Ackerley, D. F. (2013) The flavin reductase MsuE is a novel nitroreductase that can efficiently

activate two promising next-generation prodrugs for gene-directed enzyme prodrug therapy.

Cancers (Basel). 5, 985–997

41.

42.

Palmer, G. C., Jorth, P. A., and Whiteley, M. (2013) The role of two Pseudomonas aeruginosa

anthranilate synthases in tryptophan and quorum signal production. Microbiol. (United Kingdom).

159, 959–969

Drees, S. L., Li, C., Prasetya, F., Saleem, M., Dreveny, I., Williams, P., Hennecke, U., Emsley, J.,

and Fetzner, S. (2016) PqsBC, a Condensing Enzyme in the Biosynthesis of the Pseudomonas

aeruginosa Quinolone Signal: Crystal structure, inhibition, and Reaction Mechanism. J. Biol.

Chem. 291, 6610–6624

43.

44.

45.

Huntress, E. H., Lesslie, T. E., and Hearon, W. M. (1956) A new route to 1-aryl pyrroles. J. Am.

Chem. Soc. 78, 419–423

Kamm, O. (2003) β-Phenylhydroxylamine. in Organic Syntheses, pp. 57–57, John Wiley & Sons,

Inc., Hoboken, NJ, USA, 10.1002/0471264180.os004.21

Millen, M. H., and Waters, W. A. (1971) Reactions between aromatic nitro-compounds and

hydroxylamines in alkaline solution. Part II. Product studies. J. Chem. Soc. B Phys. Org.

10.1039/j29710002398

46.

Kim, Y. H., Song, W. S., Go, H., Cha, C. J., Lee, C., Yu, M. H., Lau, P. C. K., and Lee, K. (2013)

2-nitrobenzoate 2-nitroreductase (NbaA) switches its substrate specificity from 2-nitrobenzoic acid

to 2,4-dinitrobenzoic acid under oxidizing conditions. J. Bacteriol. 195, 180–192

Schnell, S., and Mendoza, C. (2000) Enzyme kinetics of multiple alternative substrates. J. Math.

Chem. 27, 155–170

Schäuble, S., Stavrum, A. K., Puntervoll, P., Schuster, S., and Heiland, I. (2013) Effect of

substrate competition in kinetic models of metabolic networks. FEBS Lett. 587, 2818–2824

Cárdenas, M. L. (2001) The Competition Plot: A kinetic method to assess whether an enzyme that

catalyzes multiple reactions does so at a unique site. Methods. 24, 175–180

Pocklington, T., and Jeffery, J. (1969) Competition of two substrates for a single enzyme. A

simple kinetic theorem exemplified by a hydroxy steroid dehydrogenase reaction. Biochem. J. 112,

331–4

47.

48.

49.

50.

51.

52.

Beland, F. A., and Kadlubar, F. F. (1985) Formation and persistence of arylamine DNA adducts in

vivo. Environ. Health Perspect. 62, 19–30

Humphreys, W. G., Kadlubar, F. F., and Guengerich, F. P. (1992) Mechanism of C8 alkylation of

15

Function of PqsL in AQNO biosynthesis

guanine residues by activated arylamines: evidence for initial adduct formation at the N7 position.

Proc. Natl. Acad. Sci. U. S. A. 89, 8278–8282

53.

54.

Kriek, E. (1974) Carcinogenesis by aromatic amines. Biochim. Biophys. Acta - Rev. Cancer. 355,

177–203

Verna, L., Whysner, J., and Williams, G. M. (1996) 2-Acetylaminofluorene mechanistic data and

risk assessment: DNA reactivity, enhanced cell proliferation and tumor initiation. Pharmacol.

Ther. 71, 83–105

55.

56.

Borenfreund, E., Krim, M., and Bendich, A. (1964) Chromosomal Aberrations induced by

hyponitrite and hydroxylamine derivatives. J. Natl. Cancer Inst. 32, 667–679

Eppink, M. H. M., Schreuder, H. A., and van Berkel, W. J. H. (1998) Interdomain binding of

NADPH in p-Hydroxybenzoate Hydroxylase as Suggested by Kinetic, Crystallographic and

Modeling Studies of Histidine 162 and Arginine 269 Variants. J. Biol. Chem. 273, 21031–21039

Eppink, M. H. ., Overkamp, K. M., Schreuder, H. A., and Van Berkel, W. J. H. (1999) Switch of

coenzyme specificity of p-hydroxybenzoate hydroxylase. J. Mol. Biol. 292, 87–96

Wang, J., Ortiz-Maldonado, M., Entsch, B., Massey, V., Ballou, D., and Gatti, D. L. (2002)

Protein and ligand dynamics in 4-hydroxybenzoate hydroxylase. Proc. Natl. Acad. Sci. U. S. A. 99,

608–613

57.

58.

59.

60.

Eppink, M. H. M., Bunthof, C., Schreuder, H. A., and van Berkel, W. J. H. (1999) Phe ¹⁶¹and Arg

¹⁶⁶variants of p -hydroxybenzoate hydroxylase. FEBS Lett. 443, 251–255

Eppink, M. H., Schreuder, H. A., and van Berkel, W. J. (1998) Lys42 and Ser42 variants of p-

hydroxybenzoate hydroxylase from Pseudomonas fluorescens reveal that Arg42 is essential for

NADPH binding. Eur. J. Biochem. 253, 194–201

61.

van Berkel, W. J., Muller, F., Jekel, P. A., Weijer, W. J., Schreuder, H. A., and Wierenga, R. K.

(1988) Chemical modification of tyrosine-38 in p-hydroxybenzoate hydroxylase from

Pseudomonas fluorescens by 5’-p-fluorosulfonylbenzoyladenosine: a probe for the elucidation of

the NADPH binding site? Involvement in catalysis, assignment in sequence and fitting . Eur. J.

Biochem. 176, 449–459

62.

63.

Bendová, L., Jurecka, P., Hobza, P., and Vondrásek, J. (2007) Model of peptide bond-aromatic

ring interaction: correlated ab initio quantum chemical study. J. Phys. Chem. B. 111, 9975–9979

Schreuder, H. A., Mattevi, A., Obmolova, G., Kalk, K. H., Hol, W. G. J., van der Bolt, F. J. T.,

and van Berkel, W. J. H. (1994) Crystal structures of wild-type p-hydroxybenzoate hydroxylase

complexed with 4-aminobenzoate, 2,4-dihydroxybenzoate, and 2-hydroxy-4-aminobenzoate and

of the Tyr222Ala mutant complexed with 2-hydroxy-4-aminobenzoate. evidence for a proton

channel and a new binding mode of the flavin. Biochemistry. 33, 10161–10170

Palfey, B. A., Ballou, D. P., and Massey, V. (1994) Flavin conformational changes in the catalytic

cycle of p-hydroxybenzoate hydroxylase substituted with 6-azido-and 6-aminoflavin adenine

dinucleotide. Sci. Biochem. 266, 110–114

64.

65.

Wimpenny, J. W., and Firth, A. (1972) Levels of nicotinamide adenine dinucleotide and reduced

nicotinamide adenine dinucleotide in facultative bacteria and the effect of oxygen. J. Bacteriol.

111, 24–32

66.

67.

68.

Drees, S. L., and Fetzner, S. (2015) PqsE of Pseudomonas aeruginosa acts as pathway-specific

thioesterase in the biosynthesis of alkylquinolone signaling molecules. Chem. Biol. 22, 611–618

Needham, E. R., and Perkin, W. H. (1904) XVI.—o-Nitrobenzoylacetic acid. J. Chem. Soc.,

Trans. 85, 148–155

Schöpf, C., Koepke, G., Kowald, B., Schülde, F., and Wunderlich, D. (1956) Darstellung und

Eigenschaften der o-Formamino- und der o-Amino-benzoylessigsäure, zweier zellmöglicher

Abbauprodukte des Tryptophans und Heteroauxins. Chem. Ber. 89, 2877–2886

Sicker, D., and Mann, G. (1988) Synthesis of ethyl ortho-substituted benzoylacetates and

investigation of the influence of ortho-substituents on keto-enol tautomerism and MS

fragmentation behaviour. Collect. Czechoslov. Chem. Commun. 53, 839–850

Zor, T., and Selinger, Z. (1996) Linearization of the Bradford protein assay increases its

sensitivity: theoretical and experimental studies. Anal. Biochem. 236, 302–308

69.

70.

16

Function of PqsL in AQNO biosynthesis

71.

72.

73.

Whitby, L. G. (1953) A new method for preparing flavin-adenine dinucleotide. Biochem. J. 54,

437–442

Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K., and Pease, L. R. (1989) Site-directed

mutagenesis by overlap extension using the polymerase chain reaction. Gene. 77, 51–59

Jagmann, N., Brachvogel, H.-P., and Philipp, B. (2010) Parasitic growth of Pseudomonas

aeruginosa in co-culture with the chitinolytic bacterium Aeromonas hydrophila. Environ.

Microbiol. 12, 1787–1802

74.

75.

Emsley, P., Lohkamp, B., Scott, W. G., and Cowtan, K. (2010) Features and development of Coot.

Acta Crystallogr. D. Biol. Crystallogr. 66, 486–501

Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P., Evans, P. R., Keegan, R.

M., Krissinel, E. B., Leslie, A. G. W., McCoy, A., McNicholas, S. J., Murshudov, G. N., Pannu,

N. S., Potterton, E. A., Powell, H. R., Read, R. J., Vagin, A., and Wilson, K. S. (2011) Overview

of the CCP4 suite and current developments. Acta Crystallogr. D. Biol. Crystallogr. 67, 235–242

Murshudov, G. N., Vagin, A. A., Dodson, E. J., and IUCr (1997) Refinement of macromolecular

structures by the maximum-likelihood method. Acta Crystallogr. Sect. D Biol. Crystallogr. 53,

240–255

Adams, P. D., Afonine, P. V., Bunkóczi, G., Chen, V. B., Davis, I. W., Echols, N., Headd, J. J.,

Hung, L.-W., Kapral, G. J., Grosse-Kunstleve, R. W., McCoy, A. J., Moriarty, N. W., Oeffner, R.,

Read, R. J., Richardson, D. C., Richardson, J. S., Terwilliger, T. C., and Zwart, P. H. (2010)

PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta

Crystallogr. D. Biol. Crystallogr. 66, 213–221

76.

77.

78.

Terwilliger, T. C., Grosse-Kunstleve, R. W., Afonine, P. V., Moriarty, N. W., Zwart, P. H., Hung,

L. W., Read, R. J., and Adams, P. D. (2008) Iterative model building, structure refinement and

density modification with the PHENIX AutoBuild wizard. Acta Crystallogr. D. Biol. Crystallogr.

64, 61–69

79.

80.

Coleman, J. P., Hudson, L. L., McKnight, S. L., Farrow, J. M., Calfee, M. W., Lindsey, C. A., and

Pesci, E. C. (2008) Pseudomonas aeruginosa PqsA is an anthranilate-coenzyme A ligase. J.

Bacteriol. 190, 1247–1255

Zhang, Y.-M., Frank, M. W., Zhu, K., Mayasundari, A., and Rock, C. O. (2008) PqsD is

responsible for the synthesis of 2,4-dihydroxyquinoline, an extracellular metabolite produced by

Pseudomonas aeruginosa. J. Biol. Chem. 283, 28788–28794

17

Function of PqsL in AQNO biosynthesis

Figure 1. Biosynthesis of alkyl quinolones in P. aeruginosa (shaded box) (15, 66, 79, 80) and possible reactions for HQNO syn-

thesis. It was demonstrated before that HHQ (#5) is not the precursor of HQNO (#8) and that feeding 2-ABA (#3) to a 2-ABA

non-producing mutant yields HQNO (14, 15). Therefore, two possible substrates can be hypothesized for PqsL, 2-

aminobenzoylacetate (#3) or the hypothetical 1-(2-aminophenyl)decane-1,3-dione (#4, initial, short-lived product of the PqsBC-

catalyzed condensation reaction). Possible reaction routes are illustrated by dashed arrows. However, alternate or additional

reactions involving anthranilate/anthraniloyl-CoA (#1) or 2-aminobenzoylacetyl-CoA (#2, short-lived) as PqsL substrates cannot

be excluded entirely, as indicated by gray dashed arrows. malCoA – malonyl CoA, oCoA – octanoyl-CoA.

18

Function of PqsL in AQNO biosynthesis

Figure 2. Midpoint potential of PqsL-bound FAD. A Representative spectra of the reduction of a PqsL/indigo carmine solution

by xanthine oxidase. Black lines indicate real spectra, red lines correspond to composite spectra of least squares fitting using the

PqsL and indigo carmine spectra shown in Fig. S1, and of reduced/oxidized benzyl viologen. B Reduction of PqsL and indigo

carmine as calculated from the fitting (one representative experiment). Solid points reflect oxidized species, open points indicate

reduced species concentrations (PqsL – black, indigo carmine – blue). E Nernst Plot comprising data collected in three

independent experiments.

19

Function of PqsL in AQNO biosynthesis

Figure 3. HQNO formation in the coupled PqsL/PqsBC reaction as a function of individual components of the assay. Reactions

were run for 2 hours for complete conversion of PqsBC substrates before products HHQ (not shown) and HQNO were extracted

with ethyl acetate and analyzed by HPLC. A, 5 µM PqsL was incubated with varying concentrations of the FAD reductase HpaC,

0.5 µM PqsBC and 100 µM 2-ABA, 120 µM octanoyl-CoA and 5 mM of NADH, B PqsL was varied, concentrations like in A,

but with 0.2 µM HpaC, C variation of PqsBC shows that excess PqsBC depletes 2-ABA before conversion by PqsL, leading to a

decrease in HQNO (conditions like A, but with 0.05 µM HpaC and 0.2 µM FAD), D variation of NADH illustrates the

inefficiency of the reaction in vitro (conditions like in A, but with 0.2 µM HpaC), E supplementation with 0.2 µM FAD increases

reaction efficiency and requires less HpaC for efficient HQNO formation (NADH was 5 mM). Excess of FAD (2 µM series)

decreases the yields. To confirm that oxygen depletion by FADH₂autoxidation was the cause of reduced product formation, the

experiment was repeated with catalase supplementation (right side). Oxygen recovery out of H₂O₂by catalase leads to an

increase of reaction efficiency. Error bars represent standard deviations.

20

Function of PqsL in AQNO biosynthesis

Figure 4. Impact of the deletion of hpaC on the production of AQ and AQNO. Pseudomonas aeruginosa strains PAO1 and

PAO1 ΔhpaC were cultivated in LB medium with 2 mM anthranilic acid at 37°C. AQ and AQNO were quantified by HPLC 12

hours and 24 hours after inoculation, and ratios between AQs and AQNOs were calculated. Supplementation of media with 5

mM p-hydroxyphenylacetic acid (pHPA, indicated) led to different effects in both strains. Error bars represent standard

deviations.

21

Function of PqsL in AQNO biosynthesis

Figure 5. Identification of the PqsL reaction product 2-HABA. A The diagram shows chromatograms (stacked) of enzyme assays

(top) and chemically synthesized (chem. syn.) 2-HABA (middle) from reverse phase HPLC. Control experiments (e.g. using

PqsBC, bottom trace) and spectra comparisons (not shown) allow the assignment of peaks to 1 – NAD, 2 – NADH, 3 – 2-HABA,

4 – 2-ABA, 5 – 2-NBA, 6 – 2-AA. For confirmation of 2-HABA authenticity, peak 3 was prepared and converted with PqsBC to

yield exclusively HQNO (not shown). B Comparison of normalized UV absorption spectra of 2-HABA (gray) and 2-ABA

(black) as recorded with the HPLC diode array detector. C Decomposition of 2-HABA determined with HPLC. Error bars reflect

standard error of mean (three representative replicates of synthetic and enzymatically prepared 2-HABA). Data were fitted with

one exponential function (gray line) revealing a half-life of 16.2 ± 0.1 min. D Reaction schemes of derivatizations of 2-ABA and

2-HABA with acetic anhydride. E LC-MS identification of the products of derivatization (according to D) of 2-ABA and of the

of PqsL reaction products. Upper trace shows UV absorbance as measured with the LC-MS DAD detector, middle and lower

traces show extracted ion chromatograms of the denoted masses of the positive MS.

22

Function of PqsL in AQNO biosynthesis

Figure 6. Substrate preference of PqsBC. A The reaction scheme on the left illustrates the two substrates 2-ABA (1) and 2-

HABA (2); the latter being generated in situ by PqsL in the coupled PqsL-PqsBC reaction (additionally required substrates

NADH, O₂and the supporting enzyme HpaC were left out for clarity). Depending on its substrate preference, PqsBC converts

either compound to the respective product (3 – HQNO, 4 – HHQ) until octanoyl-CoA is depleted. B The dependence on

octanoyl-CoA reveals which substrate, 2-ABA or 2-HABA, is the preferred and/or more efficient one. Initial 2-ABA

concentration was 250 µM, HQNO – grey, HHQ – black. Error bars reflect SEM out of three independent replicates.

23

Function of PqsL in AQNO biosynthesis

Figure 7 Structural investigation of cofactor and co-substrate binding in PqsL. A Survey of key residues for NADPH interaction

in pHBH and corresponding residues of PqsL, as deduced from the structural alignment of PqsL with pHBH. B View of the

6FHO model (green) with the remodeled loop (cyan). Positions involved in NADPH recognition in pHBH that likely are dys-

functional in PqsL are highlighted in red: R₄₁(1), which aligns with Y₃₈of pHBH and adopted the function of R₄₂(2) of pHBH

(see C) is shown in stick representation. C Close up of the flavin binding site of PqsL (green) aligned with pHBH (1PBE,

yellow).

24

PqsL uses reduced flavin to produce 2-hydroxylaminobenzoylacetate, a preferred

PqsBC substrate in alkyl quinolone biosynthesis in Pseudomonas aeruginosa

Steffen Lorenz Drees, Simon Ernst, Benny Danilo Belviso, Nina Jagmann, Ulrich Hennecke

and Susanne Fetzner

J. Biol. Chem. published online April 18, 2018

Access the most updated version of this article at doi: 10.1074/jbc.RA117.000789

Alerts:

• When this article is cited

• When a correction for this article is posted

Click here to choose from all of JBC's e-mail alerts

Article Doi

PqsL uses reduced flavin to produce 2-hydroxylaminobenzoylacetate, a preferred PqsBC substrate in alkyl quinolone biosynthesis in Pseudomonas aeruginosa

DOI: 10.1074/jbc.RA117.000789

Source and publish data:

Authors:

Article abstract of DOI:10.1074/jbc.RA117.000789

Full text of DOI:10.1074/jbc.RA117.000789

Products guided by the article

R&D Labs maybe for 341-88-8

Relevant to this article

Hot Product