Variation in pantothenate kinase type determines the pantothenamide mode of action and impacts on coenzyme A salvage biosynthesis

Variation in pantothenate kinase type determines the

pantothenamide mode of action and impacts on

coenzyme A salvage biosynthesis

ꢀ

Marianne de Villiers*, Leanne Barnard*, Lizbe Koekemoer, Jacky L. Snoep and Erick Strauss

Department of Biochemistry, Stellenbosch University, South Africa

Keywords

N-substituted pantothenamides are analogues of pantothenic acid, the vita-

min precursor of CoA, and constitute a class of well-studied bacterial

coenzyme A; growth inhibition; pantetheine;

pantothenamide; pantothenate kinase

growth inhibitors that show potential as new antibacterial agents. Previous

studies have highlighted the importance of pantothenate kinase (PanK;

EC 2.7.1.33) (the ﬁrst enzyme of CoA biosynthesis) in mediating pantothen-

Correspondence

E. Strauss, Department of Biochemistry,

amide-induced growth inhibition by one of two proposed mechanisms:

ﬁrst, by acting on the pantothenamides as alternate substrates (allowing

their conversion into CoA antimetabolites, with subsequent effects on

CoA- and acyl carrier protein-dependent processes) or, second, by being

directly inhibited by them (causing a reduction in CoA biosynthesis). In

the present study we used structurally modiﬁed pantothenamides to probe

whether PanKs interact with these compounds in the same manner. We

show that the three distinct types of eubacterial PanKs that are known to

exist (PanK_I, PanK_IIand PanK_III) respond very differently and, conse-

quently, are responsible for determining the pantothenamide mode of

action in each case: although the promiscuous PanK_Ienzymes accept them

as substrates, the highly selective PanK_IIIs are resistant to their inhibitory

effects. Most unexpectedly, Staphylococcus aureus PanK (the only known

example of a bacterial PanK_II) experiences uncompetitive inhibition in a

manner that is described for the ﬁrst time. In addition, we show that pan-

tetheine, a CoA degradation product that closely resembles the pantothen-

amides, causes the same effect. This suggests that, in S. aureus,

pantothenamides may act by usurping a previously unknown role of pante-

theine in the regulation of CoA biosynthesis, and validates its PanK as a

target for the development of new antistaphylococcal agents.

Stellenbosch University, Stellenbosch 7600,

South Africa

Fax: +27 21 808 5863

Tel: +27 21 808 5866

E-mail: estrauss@sun.ac.za

*These authors contributed equally to this

work.

(Received 30 May 2014, revised 18 July

2014, accepted 18 August 2014)

doi:10.1111/febs.13013

Introduction

The N-substituted pantothenamides are a class of pan-

tothenic acid (Pan, vitamin B₅) analogues that were

ﬁrst described in 1970 as growth inhibitors of selected

lactic acid bacteria and Escherichia coli [1]. Because

Pan is the biosynthetic precursor of the universal acyl

group carrier CoA (Fig. 1A), which serves to activate

Abbreviations

ACP, acyl carrier protein; AcpH, [ACP]hydrolase; dN5, ^DL-4⁰-deoxy-N-pentylpantothenamide; DPCK, dephospho-CoA kinase; EcPanK_I,

Escherichia coli type I pantothenate kinase; HoPanAm, homopantothenamide; HoPan, homopantothenic acid; HRMS, high resolution mass

spectrometry; LDH, lactate dehydrogenase; MIC, minimal inhibitory concentration; N354-Pan, N-[2-(1,3-benzodioxol-5-yl)ethyl]

pantothenamide; N5-Pan, N-pentyl pantothenamide; N7-Pan, N-heptyl pantothenamide; n-PanAm, n-pantothenamide; PanK, pantothenate

kinase; Pan, pantothenic acid; PantSH, pantetheine; PK, pyruvate kinase; P-Pan, 4⁰-phosphopantothenic acid; P-PantSH, 4⁰-

phosphopantetheine; PPAT, phosphopantetheine adenylyltransferase; SaPanK_II, Staphylococcus aureus type II pantothenate kinase;

a-PanAm, a-pantothenamide.

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4731

PanK type affects pantothenamide mode of action

M. de Villiers et al.

PanK_Ior

PanK_IIor

PanK_III

A

CoA biosynthesis

O

N

H

OH

N

H

OH

-

OH OH

O

OH

P

O

-

ATP ADP

Pan

P-Pan

O O

SH

N

H

N

H

O

OH

O

PPCS

P

NH₂₂

N

--

O

PPCDC

N

CoA salvage

PPAT

--

O

PanK_Ior

N

DPCK

N

O

PanK_II

SH

O

N

H

N

H

N

H

N

H

OH

--

O

P

OH OH

O

OH

PPAT

DPCK

O

--

P

O

--

ATP ADP

PantSH

P-PantSH

CoA

O O

B

O

N

H

N

H

OH OH

O

N5-Pan

N

H

N

H

O

OH

O

P

NH₂

N

^-O

O

PPAT

DPCK

PanK_Ior

N

CoA-

dependent

enzymes

^-O

PanK_II

O

N

N5-Pan

PanK

N

P

O

P

OH

^-O

O

CoA biosynthesis

Ethyldethiacarba-CoA

Pan

N5-Pan

3',5'-ADP

P

OH

AcpS

AcpH

PPCS

PPCDC

apo-ACP

crypto-ACP

N5-Pan

P

SH

CoA

P

Pan

P

Pant

SH

PPAT

DPCK

AcpS

3',5'-ADP

AcpH

SH

P

Pant

Fatty acid

biosynthesis

holo-ACP

Fig. 1. Biosynthesis of CoA and the potential targets of the pantothenamide mode of action. (A) Biosynthesis of CoA from Pan in the ﬁve-

step pathway catalyzed by panothenate kinase (PanK), phosphopantothenoylcysteine synthetase (PPCS), phosphopantothenoylcysteine

decarboxylase (PPCDC), phosphopantetheine adenylyltransferase (PPAT) and dephospho-CoA kinase (DPCK), or from PantSH in the

three-step salvage pathway consisting of PanK, PPAT and DPCK. (B) The three major biological targets of N5-Pan: (i) PanK, the ﬁrst CoA

biosynthetic enzyme; (ii) CoA-dependent enzymes, after its transformation into the antimetabolite ethyldethiacarba-CoA; and (iii) fatty acid

biosynthesis, when ethyldethiacarba-CoA serves as substrate for AcpS to form a crypto-ACP instead of the catalytically active holo-ACP.

Holo- and crypto-ACP are recycled back to apo-ACP by AcpH.

4732

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

these groups for both acyl transfer reactions and

biological Claisen condensation reactions [2], pantothen-

amides have raised much interest as antimetabolite-

based lead compounds for the development of new

selective antimicrobial agents [3]. Although recent

studies have again highlighted the potential of these

compounds [4–6], attempts at improving their potency

are hampered by the fact that their mode of action still

has not been unambiguously deﬁned and appears to

show variation between different organisms.

In combination, these factors were considered to result

in the complete inhibition of fatty acid biosynthesis;

this has subsequently been hailed as the major tar-

get for pantothenamide-induced growth inhibition in

bacteria.

Several subsequent studies have drawn this asser-

tion into question. First, studies of the puriﬁed AcpH

and E. coli strains in which the acpH gene was

either knocked out or overexpressed indicated that

the accumulation of ethyldethiacarba-ACP following

N5-Pan treatment is not a result of reduced ACP

turnover because this crypto-ACP was found to be

readily hydrolyzed by AcpH [12]. Instead, N5-Pan

treatment caused a signiﬁcant reduction in the CoA

pool (in contrast to the previous ﬁndings), indicating

N5-Pan-mediated inhibition of CoA biosynthesis.

Attempts at identifying the speciﬁc target by individ-

ual overexpression of the CoA biosynthetic genes

failed because these strains instead showed an

increase in N5-Pan sensitivity. This again pointed to

the increased ﬂux of the antimetabolite through the

pathway as an important additional determinant of

pantothenamide-meditated growth inhibition. Second,

an investigation of pantothenamide-mediated inhibi-

tion of Staphylococcus aureus identiﬁed PanK as its

point of action [13] because only pantothenamides

that caused inhibition of the S. aureus PanK enzyme

in vitro (using a standard kinase assay that couples

ADP formation to NADH reduction) showed cell

The ﬁrst mode of action study, which was conducted

in E. coli using N-pentyl pantothenamide (N5-Pan),

the prototypical example of the class, showed that it is

transformed into the CoA antimetabolite ethyldethia-

carba-CoA by three of the ﬁve CoA biosynthetic

enzymes (Fig. 1B) [7]. These enzymes are pantothenate

kinase (PanK; EC 2.7.1.33), the ﬁrst enzyme of the

pathway that usually catalyzes the ATP-dependent

phosphorylation of Pan to give 4⁰-phosphopantothe-

nic acid (P-Pan), and the fourth and ﬁfth enzymes

phosphopantetheine adenylyltransferase (PPAT) and

dephospho-CoA kinase (DPCK), which transform

4⁰-phosphopantetheine (P-PantSH) into CoA through

the addition of adenylyl and phosphate groups, respec-

tively (Fig. 1A). Moreover, the formation of ethylde-

thiacarba-CoA took place faster than the ﬁve-step

conversion of Pan into CoA when conducted under

competitive conditions in vitro. This suggested that

N5-Pan exerted its inhibitory effect by reducing the

rate of CoA synthesis, and/or through its biosynthetic

product inhibiting enzymes and processes dependent

on CoA.

growth inhibition. However,

a subsequent study

revisited these results and used [c-³²P]ATP to demon-

strate that S. aureus PanK phosphorylates both N5-

Pan and its heptyl analogue N7-Pan, and that this

(as in the case of E. coli) also leads to the formation

of modiﬁed crypto-ACPs and the inhibition of fatty

acid synthesis [14]. However, because no studies have

investigated the relative importance of the two inhibi-

tion modes, the role of PanK in the pantothenamide-

mediated growth inhibition of S. aureus remains

unresolved.

A second study focused on fatty acid biosynthesis as

the target for pantothenamide-induced bacterial

growth inhibition, based on the essential requirement

of the holo-acyl carrier protein (holo-ACP) in type II

fatty acid synthase systems (as found in E. coli) [8].

Although holo-ACP is formed when [ACP]synthase

transfers the P-PantSH group from CoA to apo-ACP

[9], it is also possible for CoA analogues such as

ethyldethiacarba-CoA to serve as the substrate instead;

this leads to the formation of crypto-ACPs that do not

have the requisite thiol group and are therefore unable

to act as acyl carriers (Fig. 1B). N5-Pan treatment

led to the formation of the N5-Pan-containing crypto-

ACP (ethyldethiacarba-ACP) as predicted, and to a

reduction in normal holo-ACP levels. CoA levels

apparently remained unaffected. Ethyldethiacarba-

ACP was also shown to accumulate and persist in N5-

Pan-treated cells even after N5-Pan was removed from

the culture, suggesting that ethyldethiacarba-ACP is a

poor substrate for the [ACP]hydrolase (AcpH) that is

responsible for ACP prosthetic group turnover [10,11].

Although the exact mode of action of the pantothen-

amides still remains a matter of debate, these studies

have all clearly highlighted the importance of PanK in

mediating their growth inhibitory effects by one of

two mechanisms: (a) by acting as the cellular target of

inhibition, leading to the suppression of CoA bio-

synthesis or (b) by accepting pantothenamides as sub-

strates, thereby allowing them to be converted into

CoA analogues that exert their effects on CoA- and

ACP-dependent processes. Unfortunately, structure–

activity relationship studies that set out to establish a

correlation between the growth inhibitory potency of a

a pantothenamide and its potential to act as either an

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4733

PanK type affects pantothenamide mode of action

M. de Villiers et al.

inhibitor or substrate of the target organism’s PanK

have not been able to provide more insight into the

enzyme’s apparent dual role in pantothenamide-based

inhibition [15,16].

of chemical motifs, yielding three sets of pantothen-

amides, referred to as a-pantothenamides (a-PanAm),

n-pantothenamides (n-PanAm, where

n

signiﬁes

‘normal’) and homopantothenamides (HoPanAm),

respectively (Table 1).

Importantly, the distinction between a pantothen-

amide acting as either a PanK substrate or inhibitor

may be linked to other aspects of CoA metabolism.

For example, only organisms with PanKs that show

poor substrate selectivity are able to form CoA from

pantetheine (PantSH) (a CoA-derived degradation

product structurally related to the pantothenamides)

by means of a salvage pathway made up of PanK,

PPAT and DPCK (Fig. 1A) [2]. PanKs occur as one

of three distinct forms, known as type I, type II and

type III PanK enzymes, respectively [17–20], with only

PanK_Iand PanK_II(the subscript denotes the type)

showing activity towards both Pan and PantSH [21].

Organisms such as Pseudomonas aeruginosa that have

PanK_IIIenzymes cannot salvage CoA from pan-

tetheine [22] and are also resistant to pantothenamide-

induced growth inhibition [17,20]. This suggests that

the inhibition mode of the pantothenamides (as struc-

tural mimics of PantSH) may be linked to the targeted

organism’s PanK type and its ability to salvage CoA.

In the present study, we explored the existence of

such links through a combination of bacterial cell

growth inhibition and kinetic characterization studies.

We ﬁnd that PanK type is the major determinant of

the pantothenamide mode of action, and particularly

so in S. aureus, which has a unique CoA metabolism

supported by a PanK_IIwith several exceptional fea-

tures [14,20]. These results not only provide signiﬁcant

new insights for antimetabolite-based antimicrobial

drug design, but also add to our understanding of the

implications of PanK diversity for CoA metabolism

and regulation.

Although a-pantothenic acid has not previously been

studied in the context of PanK and/or CoA inhibitors,

HoPan was shown to act as a competitive inhibitor of a

murine PanK_II(MmPanK1a) and to negatively affect

CoA levels in mice in vivo [24]. The same result was

found in insect cells [25]. By contrast, E. coli type I

PanK (EcPanK_I), which is known to accept a range of

pantothenic acid analogues [21], does not accept HoPan

as a substrate, nor is it signiﬁcantly inhibited by it

[15,24,26]. This diverse response indicated that such

modiﬁcations in the Pan structure could successfully be

used to probe PanK activity and inhibition.

Organisms with different PanK types exhibit

different pantothenamide-induced growth

inhibition proﬁles

Three bacterial species, representing all three PanK

types, were selected to evaluate the potency of the

pantothenamide library members: E. coli, a Gram-

negative bacterium with

a

typical PanK_I, the

Gram-positive S. aureus, the only bacterium known to

have an active (albeit atypical) PanK_II(normally only

eukaryotes have PanK_IIenzymes) and P. aeruginosa, a

Gram-negative PanK_III-containing bacterium [2].

Although P. aeruginosa has been shown to be resistant

to N5-Pan inhibition as a result of the selectivity of its

PanK_IIIenzyme [20], it was included in the present

study because we considered that the smaller a-PanAm

series could potentially be accommodated in its active

site. As in several previous studies [16,27], growth inhi-

bition assays were conducted in 1% tryptone (pancre-

atic digest of casein), a medium that contains low

amounts (< 1 l^M) of Pan [28]. Such Pan concentrations

are not strictly physiologically relevant to humans with

bacterial infections because the concentration of total

Pan (i.e. free Pan and Pan bound up in CoA and other

Pan-derived metabolites) in human whole blood ranges

between 1 and 3 l^M[29,30]. However, it allows suscep-

tibility data to be obtained under conditions where Pan

is not sufﬁciently abundant to counteract any potential

inhibition, and therefore serves the point of investigat-

ing the mode of action of these compounds.

Results

Pantothenamide library and study design

The link between PanK type and pantothenamide-

induced bacterial cell growth inhibition was investi-

gated using a library of N-substituted pantothenamides

that was previously prepared in our laboratory and

recently used to identify antiplasmodial pantothen-

amides that are resistant to pantetheinase-mediated

degradation [5,23]. The library was constructed from

Pan and two Pan analogues: a-pantothenic acid, which

has the b-alanine of Pan replaced with glycine, and ho-

mopantothenic acid (HoPan) in which it is exchanged

for c-aminobutyric acid. The acids were subsequently

coupled to 47 different amines representing a variety

Susceptibility tests were performed by obtaining

minimal inhibitory concentration (MIC) values for

those compounds that showed inhibition in initial

screens performed at 200 and 50 l^M, respectively. The

structures and MIC values for the pantothenamides

4734

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

Table 1. MIC values for the inhibition of E. coli or S. aureus grown for 20 h in 1% tryptone medium in the presence of the indicated

pantothenamides at different concentrations. The reported values represent the mean of two or more independent experiments; the errors

indicate the range/2. References to compounds tested in previous studies are provided in the main text.

O

R_Am

N

H

N

H

R_Am

N

H

OH OH

O

OH OH

O

α-PanAm

n-PanAm

HoPanAm

E. coli MIC (µ^M)

S. aureus MIC (µM)

PanAm entry

1

R_Amgroup

a-PanAm

50–200

n-PanAm

HoPanAm

a-PanAm

> 200

n-PanAm

HoPanAm

> 200

30.8 ꢀ 4.3

50–200

> 200

N

H

2

3

50–200

> 200

N

H

64.4 ꢀ 8.2

50–200

> 200

18.0 ꢀ 0.6

3.14 ꢀ 1.20

0.77 ꢀ 0.04

0.74 ꢀ 0.17

50–200

N

H

4

5

6

7

50–200

> 200

N

H

N

H

> 200

N

H

> 200

N

H

8

> 200

50–200

> 200

N

H

9

101 ꢀ 3

> 200

50–200

> 200

34.7 ꢀ 4.2

> 200

N

H

10

N

H

11

12

13

50–200

55.3 ꢀ 0.8

> 200

50–200

> 200

N

H

22.3 ꢀ 1.8

50–200

N

H

N

H

14

136 ꢀ 26

> 200

24.7 ꢀ 0.3

> 200

N

H

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4735

PanK type affects pantothenamide mode of action

M. de Villiers et al.

Table 1. (Continued).

E. coli MIC (µ^M)

S. aureus MIC (µM)

a-PanAm

PanAm entry

15

R_Amgroup

a-PanAm

> 200

n-PanAm

HoPanAm

n-PanAm

HoPanAm

> 200

50–200

> 200

N

H

16

> 200

N

H

O

17

18

19

> 200

50–200

> 200

50–200

> 200

O

N

H

S

N

H

N

H

O

20

21

50–200

> 200

50–200

> 200

50–200

> 200

N

H

S

N

H

O

22

23

> 200

H

O

N

H

50–200

H

O

N

H

24

25

> 200

50–200

> 200

N

H

O

N

H

O

H

O

26

27

> 200

50–200

> 200

N

H

N₃

NHAc

N

H

4736

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

Table 1. (Continued).

E. coli MIC (µ^M)

S. aureus MIC (µM)

PanAm entry

28

R_Amgroup

a-PanAm

> 200

n-PanAm

HoPanAm

a-PanAm

> 200

n-PanAm

> 200

HoPanAm

> 200

H

N

H

Boc

29

> 200

47.7 ꢀ 7.9

> 200

N

H

30

31

> 200

N

H

N

H

Me

O

32

33

> 200

55.0 ꢀ 10.2

> 200

N

H

> 200

Me

O

N

H

O

Me

O

34

35

> 200

50–200

> 200

O

N

H

O

N

H

OCF₃

36

> 200

N

H

F

C

3

37

38

> 200

F

C

3

N

H

O

S

N

H

N(CH₃)₂

N

H

39

> 200

N

O

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4737

PanK type affects pantothenamide mode of action

M. de Villiers et al.

Table 1. (Continued).

E. coli MIC (µ^M)

S. aureus MIC (µM)

a-PanAm

PanAm entry

40

R_Amgroup

a-PanAm

n-PanAm

HoPanAm

n-PanAm

HoPanAm

> 200

N

S

41

42

> 200

N

H

> 200

O

N

H

43

> 200

N

H

N

O

44

45

> 200

N

H

N

H

46

47

> 200

N

H

O

tested are given in Table 1. Importantly, the data

obtained for the n-PanAm set correlate well with the

results of previous pantothenamide structure–activity

relationship studies, with the MIC values differing by

no more than 10-fold (some differences are expected

due to variation in the amount of Pan present in tryp-

tone) [8,14,15].

because TolC-defective strains do show sensitivity

[8,15]. By contrast, S. aureus was only inhibited by

n-PanAm compounds. These showed MIC values that

varied by almost two orders of magnitude, with the best

inhibitors (n-PanAm-5/N7-Pan and n-PanAm-6) having

MIC values of ~ 0.7 l^M. As expected, P. aeruginosa

showed no inhibition by any of the pantothenamides

tested, including the smaller sized a-PanAm series.

Comparative analysis of the results indicated that

E. coli was only inhibited by a small number of

a-PanAm and n-PanAm series members, most of which

had MIC values in the 50–100 l^Mrange. Previous stud-

ies have demonstrated that, at least in some cases

(such as with n-PanAm-5, otherwise known as N-heptyl

pantothenamide or N7-Pan), this relatively poor inhibi-

tion proﬁle can be ascribed to TolC-dependent efﬂux

Pantothenamide-mediated growth inhibition

correlates with PanK activity in E. coli but not

S. aureus

Next, we investigated whether the differences in the

E. coli and S. aureus growth inhibition proﬁles could

4738

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

be correlated with the activity and/or selectivity of

their respective PanKs. The pantothenamides that

showed the most potent growth inhibition were

tested as substrates of the puriﬁed overexpressed

enzymes at a ﬁxed concentration (100 l^M) using an

established pyruvate kinase/lactate dehydrogenase

(PK/LDH)-based coupled enzyme assay that links

ADP production to the consumption of NADH [7].

Although this assay measures the enzymes’ activity

indirectly, it has been shown previously that the

expected monophosphorylated ester is produced and

that the PK/LDH-assay is a viable alternative for

measuring PanK activity [7,13,21,31]. The correspond-

ing compounds that have the same amide substituent

but different substitutions of the b-alanine moiety

were included for comparison. The data indicate that,

for EcPanK_I, compounds that show growth inhibition

belong to the a- and n-PanAm series and have spe-

ciﬁc activities very similar to that measured for Pan,

whereas the HoPanAms, which do not inhibit E. coli

growth, show poor PanK activity (Fig. 2A). However,

PanK activity is not necessarily a predictor of growth

inhibition because even a- and n-PanAm compounds

that do not show growth inhibition still show similar

speciﬁc activity levels. Unexpectedly, the situation is

reversed in S. aureus, with only the non-inhibitory

HoPanAm members showing S. aureus type II PanK

(SaPanK_II) activity approaching that seen for Pan

(Fig. 2B).

n-PanAm-5, n-PanAm-9, n-PanAm-12 and n-PanAm-

29 that combines a low K^a_m^ppwith low turnover

(Fig. 2F). Among these compounds, n-PanAm-5 (N7-

Pan) showed the lowest turnover; importantly, it is

also the pantothenamide that exhibits the most

potent growth inhibition of S. aureus identiﬁed to

date [13,14]. These results strongly suggest that in

E. coli PanK activity is a necessary (but not sufﬁ-

cient) requirement for pantothenamides to show

growth inhibition. By contrast, pantothenamides that

show growth inhibition of S. aureus bind its PanK

enzyme with high apparent afﬁnity, yet show poor

turnover.

Quantifying the variations in PanK kinetics:

constructing a kinetic model

Several previous studies have used the Michaelis–

Menten equation to obtain kinetic parameters for

both EcPanK_Iand SaPanK_IIacting on a range of

substrates [7,13,14,16,32–34]. In the present study we

took special care to obtain all measurements dur-

ing the initial phase of the reaction (< 10% substrate

consumed), and found that EcPanK_Ihas a hyper-

bolic activity proﬁle only for Pan, whereas it

shows sigmoidal saturation curves for the pantothen-

amides (Fig. 3A). SaPanK_IIhas sigmoidal satura-

tion curves for both Pan and HoPanAm-3 (the low

K^a_m^ppfor N5-Pan limited our ability to determine if

this is also true in its case). These ﬁndings indicate

To further investigate these differences and their rel-

evance to the pantothenamide mode of action, kinetic

proﬁles were obtained for both enzymes with Pan and

the three pantothenamides with N-pentyl substituents

[i.e. a-PanAm-3, n-PanAm-3 (referred to as N5-Pan

from here on) and HoPanAm-3] (Fig. 2C). These con-

ﬁrm that the pantothenamides that acted as E. coli

inhibitors (N5-Pan and a-PanAm-3) have EcPanK_I

activity proﬁles that closely resemble that of Pan.

HoPanAm-3, which did not act as an E. coli growth

inhibitor, is clearly a poor EcPanK_Isubstrate. By con-

trast, the S. aureus growth inhibitor N5-Pan has a very

different activity proﬁle that shows both a lower

apparent K_m(K^a_m^pp) and signiﬁcantly reduced turnover

compared to Pan (Fig. 2D). The two pantothenamides

that did not show inhibition of S. aureus, a-PanAm-3

and HoPanAm-3, are apparently poor and excellent

substrates, respectively.

that some substrates have

effect on PanK catalysis, a phenomenon that has

a positive cooperative

previously only been described for certain PanK_III

[35].

s

The sigmoidal saturation data were best described

by simple Hill-type equation (Eqn 1) derived

according to mechanism based on dimeric

a

enzyme with one active site per subunit; the avail-

able structural data shows that both EcPanK_Iand

SaPanK_IIconform to this description [20,26,36]. This

mechanism, shown schematically for SaPanK_IIacting

on Pan in Fig. 3B, is based on the assumption that

only enzyme states with both of the active sites

occupied (e.g. E_Pan.Pan) have to be considered and

lead to product formation (as is usually the case for

cooperative enzymes). The differential binding of the

substrate to the two active sites is described by the

parameter a that modiﬁes the dissociation constant

for the enzyme form with substrate bound in both

active sites. In Eqn (1), k_fdenotes the rate of trans-

formation (phosphorylation) of Pan, E_Tis the total

enzyme concentration and K_Panis the dissociation

To conﬁrm these trends, we extended the studies

to other selected growth inhibitory pantothenamides.

We found that EcPanK_Ialso accepts a-PanAm-12

and a-PanAm-14 as substrates (Fig. 2E), whereas

SaPanK_IIshows the same unusual kinetic proﬁle for

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4739

PanK type affects pantothenamide mode of action

M. de Villiers et al.

SaPanK_II

6.4

5.6

4.8

4.0

3.2

2.4

1.6

0.8

0.0

3.2

A

EcPanK_I

B

-PanAm-x

2.8

2.4

2.0

1.6

1.2

0.8

0.4

0.0

n-PanAm-x

HoPanAm-x

1

3

4

5

6

9

12 14 29 32

1

3

4

5

6

9

12 14 29 32

PanAm Entry

2.0

1.6

1.2

0.8

0.4

0.0

8.0

7.2

6.4

5.6

4.8

4.0

3.2

2.4

1.6

0.8

0.0

EcPanK_I

Pan

SaPanK_II

Pan

C

D

-PanAm-3

N5-Pan

HoPanAm-3

-PanAm-3

N5-Pan

HoPanAm-3

0

50

100

150

200

250

0

50

100

150

200

250

[Substrate] (M)

2.0

1.6

1.2

0.8

0.4

0.0

8.0

6.0

4.0

E

F

EcPanK_I

SaPanK_II

Pan

N5-Pan

-PanAm-12

-PanAm-14

n-PanAm-5

n-PanAm-9

n-PanAm-12

n-PanAm-29

1.6

1.2

0.8

0.4

0.0

0

50

100 150 200 250 300

0

10 20 30 40 50 60 70 80 90 100

[Substrate] (M)

Fig. 2. Activity of EcPanK_Iand SaPanK_IItowards selected pantothenamides. (A) The speciﬁc activities for EcPanK_Itowards Pan (dashed line)

and the pantothenamides (bars) at 100 l^M. The data were obtained from a single experiment performed in triplicate; the error bars denote

the SD. (B) The data for SaPanK_IIagainst Pan (dashed line) and the same panel of pantothenamides (bars). (C) Activity proﬁles for EcPanK_I

with Pan, a-PanAm-3, N5-Pan and HoPanAm-3 as substrates. For information on statistical analysis and equations used to ﬁt the data, see

Table 2. (D) As for (C) but with SaPanK_II. (E) EcPanK_Iactivity proﬁles with the E. coli growth inhibitory pantothenamides N5-Pan, a-PanAm-

12 and a-PanAm-14 as substrates, with the Pan proﬁle shown for comparison. Description and data analysis is as indicated for (C). (F)

SaPanK_IIactivity proﬁles of the S. aureus growth inhibitory pantothenamides N5-Pan, n-PanAm-5 (N7-Pan), n-PanAm-9, n-PanAm-12 and

n-PanAm-29, with the Pan proﬁle shown for comparison. Note the y-axis break. Description and data analysis is as indicated for (C).

constant of Pan; V_maxis then equal to 2ꢁk_fꢁE_T, and

0:5

Note that, from these conditions, it follows that

qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

Pan

K

(the K_0.5value for Pan) is numerically equal to

2

Pan

0:5

K_Pan

¼

a ꢁ ðK Þ .

the concentration of Pan where v_PanK= 0.5ꢁV_max

.

4740

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

ꢀ

K_Pan

ꢁ

2

both enzymes acting on mixtures of Pan and N5-Pan;

this pantothenamide was chosen for this particular

experiment because it inhibits both E. coli and

S. aureus growth and therefore allows for a compara-

tive analysis. For each experiment, the total substrate

concentration was increased at the same time as

maintaining a constant ratio of Pan:N5-Pan; in subse-

quent experiments, the amount of Pan in the ratio

was increased in a stepwise manner, from 0% (pure

N5-Pan) to 100% (pure Pan). The combined turnover

of Pan and N5-Pan was measured using the PK/

LDH-based assay that responds to the formation of

ADP.

½Panꢂ

2 ꢁ k_fꢁ E_Tꢁ a

K_Pan

v_PanK¼2 ꢁ k_fꢁ ½E_PanꢁPanꢂ ¼

ꢀ

ꢁ

2

½Panꢂ

1 þ a ꢁ

ð1Þ

ꢀ

ꢁ

2

½Panꢂ

0:5

ꢁ

V_max

ꢁ

Pan

K

¼

ꢀ

2

½Panꢂ

0:5

1 þ

Pan

K

Fitting Eqn (1) to the saturation data showed a good

ﬁt in all cases, as exempliﬁed by the curve of SaPanK_II

with Pan (Fig. 3C), and allowed determination of the

relevant kinetic parameters. These values, summarized

in Table 2, indicate that although for E. coli there is

general agreement between a pantothenamide’s growth

inhibitory potency and the speciﬁc activity that EcPanK_I

shows towards it, there is no direct correlation, as would

be expected if the enzyme merely served as a gateway for

the metabolic activation of these compounds, which

subsequently have their inhibitory effect elsewhere. For

S. aureus, potency correlates with compounds that have

the unusual combination of very low (< 5 l^M) values of

K_0.5and low turnover rates, and, consequently, very

high apparent speciﬁcity constants (k_cat/K_0.5).

The results demonstrate that, for EcPanK_I, the

resulting activity proﬁles show very small differences,

and that there is a gradual progression from the proﬁle

for Pan to that for N5-Pan (Fig. 4A). However, for

SaPanK_II, even small amounts (i.e. 5%) of N5-Pan

exert both an inhibitory effect (at mixture concentra-

tions of 50 l^Mand higher) and an apparent stimula-

tory effect (at mixture concentrations of ~ 12.5 l^Mand

below) on activity (Fig. 4B). Overall, the impact of

N5-Pan on SaPanK_IIactivity is unexpectedly complex.

In comparison, mixtures of Pan and HoPanAm-3,

which acts as a substrate of SaPanK_IIbut does not show

S. aureus growth inhibition, yielded essentially the same

kinetic proﬁle regardless of the composition of the sub-

strate mixture (Fig. 4C). These results conﬁrm that

pantothenamides that show growth inhibition in

N5-Pan has an inhibitory binding interaction with

SaPanK_IIbut not with EcPanK_I

To further investigate the importance of the activity

proﬁle differences, we obtained kinetic proﬁles for

A

C

1.6

8.0

EcPanK_I

Pan

α-PanAm-3

N5-Pan

SaPanK_II

7.2

6.4

5.6

4.8

4.0

3.2

2.4

1.6

0.8

0.0

1.2

0.8

HoPanAm-3

0.4

Fig. 3. Constructing a kinetic model for

0.0

SaPanK_IIactivity. (A) Activity proﬁles for

SaPanK_II

3.2

2.4

1.6

0.8

0.0

EcPanK_Iand SaPanK_IIshown in Fig. 2C,D

at low substrate concentration to highlight

the sigmoidal nature of the curves. (B)

Schematic representation of the SaPanK_II

kinetic model, based on an enzyme with

two subunits, from which Eqn (1) is

derived. K_Panis the dissociation constant

of Pan, and a denotes the change in

dissociation constant in the enzyme form

with Pan bound to both subunits. This is

the active form of the enzyme (shown in

green), which catalyzes the

Pan

0

50

100 150 200 250

0

10

20

30

40

50

[Substrate] (μM)

[Pan] (μM)

B

Pan

K_Pan

Pan

K_Pan/α

Pan

SaPanK_II

phosphorylation reaction as denoted by

the arrow with the open square. (C) Fitting

Eqn (1) to the data points for the activity

of SaPanK_IItowards Pan, with the line

indicating the best ﬁt and the shaded area

showing the 90% conﬁdence interval.

Parameter values and SEs are given in

Table 2.

+

2

K_0.5

=

(K_Pan

)

α

/

Pan

P

2

Pan

SaPanK_II

Legend

Active sites

K_Pan/α

Pan

K_Pan

Pan

Catalytically

inactive

subunit

Catalytically

active

subunit

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4741

PanK type affects pantothenamide mode of action

M. de Villiers et al.

S. aureus have multifaceted interactions with its PanK

enzyme, and that they do not simply act as alternative

substrates.

Expanding the kinetic model to account for

complex interaction of N5-Pan with SaPanK_II

We next used the dN5 data to expand the kinetic

model for SaPanK_IIto include a mechanism that

could account for the dual stimulatory/inhibitory effect

exerted by both dN5 and N5-Pan. Because the N5-Pan

saturation kinetics did not show substrate inhibition at

high concentrations, we started with an uncompetitive

inhibition mechanism for dN5 [i.e. one in which it only

binds to the SaPanK_II–Pan ([E_Pan]) complex]. In addi-

tion, we also included a competitive inhibition mecha-

nism in which dN5 can bind to any unoccupied active

site. This led to the mechanistic scheme shown in

Fig. 7A, which can be translated into a rate equation

that assumes that only the [E_Pan.Pan], [E_Pan.dN5] and

[E_dN5.Pan] complexes have catalytic activity, and in

which no enzyme form that has only one active site

occupied is considered (Eqn 2). The parameters in this

equation are deﬁned as those used in Eqn (1), with the

addition of K_dN5(dissociation constant for the com-

DL-4⁰-Deoxy-N5-Pan acts as an inhibitor of

SaPanK_IIturnover and S. aureus growth

To simplify the analysis of the interaction of

N5-Pan with SaPanK_II, we prepared its structural

analogue DL-4⁰-deoxy-N-pentylpantothenamide (dN5)

by NaBH₄-mediated reduction of a ketoamide precur-

sor, to give the product as the racemate (Fig. 5A).

This compound is expected to exclusively act as an

inhibitor because it has the 4⁰-OH group of N5-Pan

removed, thereby preventing it from being a PanK sub-

strate (although all other binding interactions are

retained).

Experiments were performed on both EcPanK_I

and SaPanK_IIusing reaction mixtures that contained

25 l^MPan and increasing concentrations of either

N5-Pan or dN5. The addition of N5-Pan increased

the total observed activity for EcPanK_Ias expected

for an alternate substrate, whereas dN5 acted as a

very poor inhibitor of the enzyme (IC₅₀~ 500 lM)

(Fig. 5B). Full kinetic analysis subsequently indicated

0

petitive inhibition by dN5) and K_dN5(dissociation

constant for the uncompetitive inhibition by dN5):

Equation (2) was able to describe the data very well

(Fig. 5D, solid lines) and accounted for both the stim-

ulatory effect of the inhibitor at low substrate concen-

app

max

that this is the result of a minor reduction of the

(data not shown), conﬁrming that the 4⁰-OH

trations and the reduction in V at high substrate

concentrations. These effects are based on the dissocia-

app

max

V

group is an important determinant for ligand bind-

ing in EcPanK_I[26].

tion constants for the binding of dN5 in a competitive

0

(K_dN5= 3.81 ꢀ 0.45 lM) and uncompetitive (K_dN5

=

By contrast, both N5-Pan and dN5 showed the

same apparent inhibitory effect on SaPanK_IIwith IC₅₀

values of 4.8 ꢀ 1.2 and 7.3 ꢀ 0.9 lM, respectively

(Fig. 5C). For this enzyme, full kinetic analysis again

41.0 ꢀ 7.1 l^M) binding mode, respectively, which

clearly show that, for dN5, the former is the stronger

interaction (Table 3). Additionally, the dissociation

constant for Pan (K_Pan= 9.31 ꢀ 0.70 lM) and the

value for a (0.111 ꢀ 0.017 l^M) were also calculated in

this manner. The K_0.5value of 27.9 lM calculated from

these values is almost identical to the value of 27.8 l^M

obtained by ﬁtting Eqn 1 to the Pan activity proﬁle

(Fig. 3C and Table 2).

showed a dual effect (Fig. 5D), with dN5 causing a

app

small but signiﬁcant reduction in the K

for Pan and

0:5

a large reduction in the V^a_m^p_a^p_x. Additionally, growth

inhibition tests performed in minimal medium showed

that dN5 has a MIC of ~ 50 l^Mfor S. aureus but does

not inhibit E. coli (Fig. 6). By comparison, N5-Pan

still has a much lower MIC (approximately 1.5 l^M)

for S. aureus under these conditions. Nonetheless, this

result unambiguously shows that even pantothenamide

analogues unable to act as PanK substrates have a

negative impact on S. aureus growth.

The newly expanded mechanism can easily account

for the data obtained from SaPanK_IIacting on mix-

tures of Pan and N5-Pan (Fig. 4B) by modifying the

competitive inhibition component exerted by dN5 to

that of a competitive substrate with catalytic turnover.

This confers activity on the two boxed complexes in

ꢀ

ꢁ

ꢁ ꢀ

ꢁ

½Panꢂ

½dN5ꢂ

V_max

ꢀ

ꢁ

ꢁ a ꢁ ^½Panꢂþ

K_Pan

ꢃ

K_Pan

K_dN5

ꢁ ꢀ

ꢂ

v_PanK¼2 ꢁ k_fꢁ ½E_PanꢁPanꢂ þ 2 ꢁ k_fꢁ ½E_PanꢁdN5ꢂ ¼

ꢀ

ꢁ

ꢀ

ꢁ

ꢀ

ꢁ

2

1 þ a ꢁ

²ꢁ 1 þ

þ 2 ꢁ

ꢁ 1 þ

þ

½Panꢂ

½dN5ꢂ

½Panꢂꢁ½dN5ꢂ

½dN5ꢂ

K_Pan

K_dN5

0

K_PanꢁK_dN5

K_dN5

0

K_dN5

ð2Þ

4742

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

Table 2. PanK kinetic parameters with Pan and various pantothenamides. Kinetic parameters were determined by keeping the ATP

concentration constant at 1.5 m^Min all cases. All reported parameters are the mean of those obtained by ﬁtting the given equation to the

data obtained for each individual experiments; the error values represent the range/2 (for parameters obtained from two independent

experiments) or SEM (for parameters obtained from three or more independent experiments). References to compounds tested in previous

studies are provided in the main text. ND, not determined.

EcPanK_I

SaPanK_II

Equation

ﬁtted

a

ꢃ1

Compound

Pan

K_0.5(lM)

k_cat(s^ꢃ1

)

k_cat/K_0.5^a(mM ꢁs^ꢃ1

ꢃ1

)

N^b

R²value^c

0.9955

K_0.5(lM)

k_cat(s^ꢃ1

)

k_cat/K_0.5(mM ꢁs^ꢃ1

)

N^b

Equation ﬁtted

Eqn (1)

R²value^c

0.9928

21.3 ꢀ 1.4

14.6 ꢀ 1.1

15.6 ꢀ 1.2

1.16 ꢀ 0.14

0.95 ꢀ 0.13

1.06 ꢀ 0.03

54.5 ꢀ 3.1

64.3 ꢀ 4.3

68.6 ꢀ 3.4

2

Michaelis–

Menten

Modiﬁed

Eqn (1)^d

Eqn (1)

27.8 ꢀ 3.3

3.37 ꢀ 0.24

4.14 ꢀ 0.14

0.44 ꢀ 0.03

2.85 ꢀ 0.11

124 ꢀ 7

5.29 ꢀ 0.62

> 290

4

a-PanAm-3

2

0.9679

0.9988

783 ꢀ 64

< 1.5^f

1^e

4

Michaelis–Menten

Eqn (1)

0.8719

0.8606

0.9740

n-PanAm-3

[N5-Pan]

HoPanAm-3

a-PanAm-12

a-PanAm-14

n-PanAm-5

[N7-Pan]

720 ꢀ 95

63 ꢀ 11

151 ꢀ 56

ND

0.62 ꢀ 0.04

1.05 ꢀ 0.12

1.37 ꢀ 0.15

0.87 ꢀ 0.17

18.2 ꢀ 4.3

13.8 ꢀ 6.6

1^e

3

Eqn (1)

0.8866

0.9881

0.9911

27.8 ꢀ 2.3

ND

103 ꢀ 6

3

Eqn (1)

3

ND

2.04 ꢀ 0.54

0.22 ꢀ 0.02

118 ꢀ 41

2

Eqn (1)

0.8677

n-PanAm-9

n-PanAm-12

n-PanAm-29

PantSH

ND

4.94 ꢀ 0.94

4.47 ꢀ 0.11

3.49 ꢀ 0.50

< 1.5^f

0.33 ꢀ 0.12

0.32 ꢀ 0.04

0.75 ꢀ 0.06

0.39 ꢀ 0.09

65.4 ꢀ 11.8

72.7 ꢀ 10.0

232 ꢀ 59

> 260

2

3

4

Eqn (1)

0.9711

0.9326

0.8714

ND

23.6 ꢀ 2.1

1.00 ꢀ 0.09

42.4 ꢀ 0.3

3

Michaelis–

0.9949

Menten

a

In those cases where the Michaelis–Menten equation was used to ﬁt the data, K_0.5is K_m

.

b

c

d

Number of independent experiments.

R²value for the indicated equation ﬁtted to the data averaged from all the experiments.

a-PanAm-3 showed inhibition of EcPanK_Iat high substrate concentrations; using

a modiﬁed version of Eqn (1) that accounts for the uncompetitive substrate inhibition, a K_iof

720 ꢀ 120 l^Mwas determined.

As a result of the high K_0.5value, kinetic parameters were determined by ﬁtting the equation to the data from a single experiment performed in triplicate; errors indicate the SE of the ﬁt.

e

f

app

0:5

The low K

values of these compounds cannot be reported with accuracy because they fall below the sensitivity limit of the assay.

Fig. 7A as shown in Fig. 7B, and gives the corre-

sponding rate equation Eqn (3) that was ﬁtted to the

Pan/N5-Pan experimental data, constraining the values

The equation provided a very good description of the

experimental data set (Fig. 8A) and gave the parameter

values listed in Table 3. These indicate that, although

the dissociation constant of N5-Pan for binding at the

active site (i.e. as a substrate) is similar to that of dN5

(K_N5-Pan= 2.22 ꢀ 0.48 lM), its dissociation constant

for binding as an uncompetitive inhibitor (K_N5-

Pan

for V

(the V_maxfor Pan), K_Panand a to those

max

obtained before. In this equation, k^P_f^anand k_f^N5-Pan

denote the rates of phosphorylation of Pan and N5-

Pan

max

Pan respectively, giving V

N5-Pan

¼ 2 ꢁ k^Panꢁ E_Tfor Pan

f

¼ 2 ꢁ k^N_f^5-Panꢁ E_Tfor N5-Pan. Addition-

= 4.89 ꢀ 0.66 lM) is much smaller. Because N5-

Pan⁰

and V

ally, K_N5-Panis the dissociation constant of N5-Pan

max

Pan also has a much reduced MIC compared to dN5,

this provides additional evidence for the uncompetitive

inhibition of SaPanK_IIas an important contributor to

the inhibition of S. aureus growth by N5-Pan.

0

acting as a substrate, whereas K_N5-Panis the dissocia-

tion constant for the uncompetitive inhibition compo-

nent that it exerts:

v_PanK¼ 2 ꢁ k_f^Panꢁ ½E_PanꢁPanꢂ þ 2 ꢁ k^P_f^anꢁ ½E_PanꢁN5-Panꢂ þ 2 ꢁ k^N_f^5-Panꢁ ½E_{N5-PanꢁN5-Pan}ꢂ þ 2 ꢁ k_f^N5-Panꢁ ½E_N5-PanꢁPan

ꢂ

N5-Pan

þ 2 ꢁ k^N_f^5-Panꢁ ½E

ꢂ

^N5-PanꢁPan

ꢀ

ꢁ ꢀ

K_Pan

ꢁ

ꢀ

ꢁ ꢀ

ꢀ

ꢁꢁ

½Panꢂ

½N5-Panꢂ

K_N5

Pan

max

N5-Pan

ð3Þ

V

ꢁ

ꢁ a ꢁ ^½Panꢂþ

þ V

ꢁ

^½N5-Panꢂþ ^½Panꢂþ 1 þ

max

K_Pan

ꢂ

K_N5

Pan

K_Pan

0

-

Pan

-

Pan

-

ꢃ

¼

ꢀ

ꢁ

ꢀ

ꢁ

ꢀ

ꢁ ꢀ

ꢁ

ꢀ

ꢁ

2

½Panꢂ

½N5-Panꢂ

½Panꢂꢁ½N5-Panꢂ

½N5-Panꢂ

1 þ a ꢁ

²ꢁ 1 þ

þ 2 ꢁ

ꢁ 1 þ

þ

K_N5

K_PanꢁK_N5

K_N5

0

Pan

K_N5

Pan

0

-

Pan

-

Pan

-

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4743

PanK type affects pantothenamide mode of action

M. de Villiers et al.

SaPanK_II

8.0

6.4

4.8

3.2

1.6

0.0

A

2.0

1.6

1.2

0.8

0.4

0.0

B

8.0

6.4

4.8

3.2

1.6

0.0

C

100% Pan

95% Pan

75% Pan

50% Pan

25% Pan

N5-Pan

EcPanK_I

SaPanK_II

100% Pan

75% Pan

50% Pan

25% Pan

N5-Pan

100% Pan

75% Pan

50% Pan

25% Pan

HoPanAm-3

0

5

10 15 20 25 50 100 150 200

0

50

100

150

200

250

0

50

100

150

200

[Pan+N5–Pan] (μM)

[Pan+HoPanAm–3] (μM)

[Pan+N5–Pan] (μM)

Fig. 4. Analyzing the interaction of Pan and N5-Pan with EcPanK_Iand SaPanK_IIunder competitive conditions. (A) Kinetic proﬁles of mixtures

of Pan and N5-Pan determined with EcPanK_I. The x-axis indicates the total substrate concentration (Pan and N5-Pan combined); for each

data set, the ratio of Pan:N5-Pan was kept at a constant value. The data points represent the mean of two independent experiments, each

performed in triplicate; the error bars denote the range/2. Because the curves progress from a hyperbolic to a sigmoidal shape, the solid

lines represent the best ﬁt of the data to the Hill equation with the value for h (Hill coefﬁcient) constrained to between 1 and 2. (B) Kinetic

proﬁles of mixtures of Pan and N5-Pan determined with SaPanK_IIas described for (A). The solid line represents the best ﬁt of Eqn (1) to the

100% Pan data. (C) Kinetic proﬁles of mixtures of Pan and HoPanAm-3 determined for SaPanK_IIas described for (A).

In the absence of Pan, Eqn (3) reduces to a much

simpler form (Eqn 4), which describes the saturation

kinetics of inhibitory pantothenamides such as N5-

Pan.

tained 25 l^MPan and increasing concentrations of

PantSH. The results are strikingly similar to those

obtained for N5-Pan, with PantSH increasing the total

observed activity for EcPanK_Ibut inhibiting SaPanK_II

activity (Fig. 8B). Similarly, when constant ratio mix-

tures of Pan and PantSH were analyzed as conducted

for N5-Pan, the proﬁles clearly demonstrate that

PantSH has the same complex interaction with SaPan-

K_II(Fig. 8C), showing both a stimulatory (at low con-

centrations) and an inhibitory (at high concentrations)

effect on turnover that is even more pronounced than

that caused by N5-Pan. Based on the similarity between

N5-Pan and PantSH, Eqn (3) was used to describe the

Pan/PantSH experimental data by substituting PantSH

for N5-Pan, and again constraining the V^a_m^p_a^p_x, K_Panand

a values to those obtained previously. The equation

was able to describe the experimental data set well

(Fig. 8C), and gave values for K_PantSH(binding at

v_PanK¼ 2 ꢁ k_f^N5-Panꢁ ½E_{N5ꢃPanꢁN5-Pan}

ꢂ

ꢀ

K_N5

ꢁ

2

½N5-Panꢂ

N5ꢃPan

V

ꢁ

ð4Þ

max

K_N5

-

Pan

¼

ꢀ

ꢁ

2

½N5-Panꢂ

1 þ

-

Pan

This equation is equivalent to Eqn (1), which was

used to obtain the parameters shown in Table 2, if

K_0.5= K_Pan(i.e. if a = 1). This would indicate that

there is no differentiation in the binding of N5-Pan to

the enzyme’s two active sites.

SaPanK_IIatypical kinetic mechanism follows

from its role in CoA salvage biosynthesis

0

the active site, i.e. as substrate) and K_PantSH(binding in

The question arises as to why there is such a selective

binding site in SaPanK_IIthat it accepts N5-Pan at the

same time as excluding HoPanAm-3 (note that the lat-

ter did not show any uncompetitive inhibition mode;

Fig. 4C). Clearly, the selectivity is for compounds that

contain the natural b-alanine moiety, suggesting that

the site is predisposed to bind compounds containing

the native pantothenoyl group. Based on its close

structural similarity to n-PanAm series compounds

(such as N5-Pan), we considered PantSH (the substrate

of the CoA salvage pathway; Fig. 1A) as the most

likely natural ligand for this binding site.

an uncompetitive mode) of 0.46 ꢀ 0.08 l^Mand

6.03 ꢀ 1.27 l^M, respectively (Table 3). In comparison

with the values obtained for N5-Pan, PantSH appears

to be the better substrate at the same time as showing

similar uncompetitive inhibition to N5-Pan. Conse-

quently, PantSH is more effective at activating SaPanK_II

at low substrate concentrations (Fig. 8A,C, insets).

Interestingly, when the activity of SaPanK_IIactivity

against constant ratio mixtures of Pan and n-PanAm-5

(N7-Pan, the most potent S. aureus growth inhibitor) was

measured and analyzed in a similar manner, the obtained

kinetic parameters were almost identical to those found

for PantSH (data not shown). Taken together, these

results strongly suggest that pantothenamides showing

To test such a hypothesis, we measured the activity

of EcPanK_Iand SaPanK_IIagainst mixtures that con-

4744

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

O

A

B

a

b

H₂N

N

H

N

H

N

H

N

H

N

H

OH

+

O

OH

O

dN5

C

D

200

175

150

125

100

75

10.0

120

100

80

60

40

20

0

0.12

0.06

0

EcPanK_I

SaPanK_II

8.0

6.0

4.0

2.0

0.0

0

50

1.875

3.75

7.5

15

30

50

N5-Pan

dN5

N5-Pan

dN5

25

0

100 200 300 400 500 600

0

10

20

30

40

50

0

50

100

[Pan] (μM)

150

200

[N5-Pan or dN5] (μM)

Fig. 5. dN5 as a PanK inhibitor. (A) Synthetic scheme for the preparation of dN5. (a) 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide

hydrochloride (EDC), hydroxybenzotriazole (HOBt), N,N-diisopropylethylamine (DIPEA), CH₂Cl₂; (b) NaBH₄, methanol. (B) Activity/inhibition

proﬁles for EcPanK_Iwhen treated with N5-Pan and dN5 at a ﬁxed concentration of 25 lM Pan. The data points represent the mean of two

independent experiments, each performed in triplicate; the error bars denote the range/2. The solid lines represent the best ﬁts to the

Michaelis–Menten equation (for N5-Pan) or to Eqn (5) (for dN5). (C) Inhibition proﬁles for SaPanK_IIas described for EcPanK_Iin (A). (D)

Kinetic proﬁles for SaPanK_IIwith Pan in the presence of increasing concentrations of dN5 (indicated on the right of each curve). The data

points represent the mean of two independent experiments, each performed in triplicate; the error bars denote the range/2. The solid lines

represent the best ﬁt curves obtained by ﬁtting Eqn (2) to all the data; the shaded areas indicate the 90% conﬁdence intervals. The insert

shows the stimulatory effect of the inhibitor at low substrate concentrations which results from the change Pan’s saturation kinetics proﬁle

from sigmoidal for Pan only (black) to hyperbolic for Pan in the presence of 3.75 l^MdN5 (purple).

120

100

80

60

40

20

0

120

100

80

60

40

20

0

A

B

E. coli

S. aureus

Fig. 6. dN5 as a bacterial cell growth

inhibitor. (A) Growth inhibition proﬁles for

E. coli grown in minimal medium in the

presence of increasing concentrations of

N5-Pan and ^DL-4⁰-deoxy-N5-Pan (dN5). The

data points represent the mean of two

independent experiments, each performed

in triplicate; the error bars denote the

range/2. The solid lines represent the best

ﬁts to Eqn (5). (B) Growth inhibition

N5-Pan

dN5

N5-Pan

dN5

0.1

1

10

100

0.1

1

10

100

proﬁles for S. aureus grown in minimal

medium as described for E. coli in (A).

[N5-Pan or dN5] (μM)

inhibition of S. aureus growth mimic the interactions of

PantSH with SaPanK_II.

establish a clear link between these two features

[1,6–8,12,14–16,26,32,37]. In the present study, we

performed the ﬁrst detailed comparative kinetic

analysis of pantothenamides that cause bacterial

growth inhibition, showing that PanK type deter-

mines their mode of action. For organisms with

PanK_Ienzymes, our data indicate that pantothen-

amides only show growth inhibition if they are

accepted as alternate PanK substrates. This corre-

sponds to these compounds having a mode of action

Discussion

Although several previous studies have investigated

the growth inhibitory potency of pantothenamides

against various bacteria, and/or have shown that

these compounds act as substrates of the target

organisms’ PanK enzymes, none have been able to

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4745

PanK type affects pantothenamide mode of action

M. de Villiers et al.

A

Pan

Legend

Pan

P

2

+

SaPanK_II

Pan

K_Pan/α

2

Pan

Competitively

Uncompetitively

inhibited subunit

K_dN5'

inhibited subunit

2

Pan

K_Pan

Pan

K_dN5

*

Pan

P

+

SaPanK_II

Putative

allosteric

binding site

Pan

SaPanK_II

K_dN5'

K_Pan/α

Pan

K_dN5

Pan

‡

K_dN5

N5-Pan

dN5

B

Replacing

with

:

*

‡

Pan

N5-Pan

P

2

Pan

P

+

SaPanK_II

+

N5-Pan

SaPanK_II

K_N5-Pan'

Pan

N5-Pan

P

+

SaPanK_II

Fig. 7. Kinetic model for the interaction of dN5 and N5-Pan with SaPanK_II. (A) Schematic representation of the SaPanK_IIkinetic model

adapted to account for the dual stimulatory/inhibitory effect shown by d5N. K_Panis the dissociation constant of Pan and a denotes the

change in dissociation constant in the enzyme form with Pan bound to both subunits. K_dN5denotes the dissociation constant for

0

competitive inhibition, whereas K_dN5is the dissociation constant for uncompetitive inhibition. Note that, for the latter, it is assumed that

inhibition takes place through binding at an allosteric site as shown, although no evidence for such a site has yet been found. For enzyme

complexes shown in square brackets, the symmetrical alternative complex is not shown but implied. All active enzyme subunits are shown

in green, whereas the competitively and uncompetitively inhibited complexes are shown in shades of grey as indicated. (B) The model

shown in (A) can be modiﬁed to account for the interaction of the N5-Pan with SaPanK_IIby converting the two competitively inhibited

complexes [boxed and identiﬁed with an asterisk (*) and double dagger (‡), respectively] to catalytically active complexes that

phosphorylates N5-Pan (and Pan) as indicated. K_N5-Pan’is the dissociation constant for the uncompetitive inhibition shown by N5-Pan.

4746

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

Table 3. Kinetic parameters for SaPanK_IIdetermined from ﬁtting various data sets to the indicated equations.

Data set

Equation

Eqn (2)

Parameter

Value^a

Unit

dN5 inhibition data (Fig. 5D)

V_max

6.47 ꢀ 0.15

9.31 ꢀ 0.70

0.111 ꢀ 0.017

3.81 ꢀ 0.45

41.0 ꢀ 7.1

lmolꢁmin^ꢃ1ꢁmg

lM

K_Pan

a

K_dN5(competitive inhibition)

lM

0

K_dN5(uncompetitive inhibition)

lM

Pan:N5-Pan mix data (Fig. 4B; Fig. 8A)

Pan:PantSH mix data (Fig. 8C)

Eqn (3)

V

0.755 ꢀ 0.100

2.22 ꢀ 0.48

4.89 ꢀ 0.66

0.656 ꢀ 0.061

0.46 ꢀ 0.08

6.03 ꢀ 1.27

lmolꢁmin^ꢃ1ꢁmg

N5-Pan

max

K_N5-Pan(substrate)

lM

K_N5-Pan’(uncompetitive inhibition)

PantSH

max

lM

Eqn (3)^b

V

lmolꢁmin^ꢃ1ꢁmg

K_PantSH(substrate)

lM

K_PantSH’(uncompetitive inhibition)

^aErrors indicate the SE.

^bPantSH replaces N5-Pan in Eqn (3) as written.

10.0

A

B

C

0.12

0.06

0

SaPanK_II

200

175

150

125

100

75

E. coli

0.06

0

Pan

N5-Pan

Pan

PantSH

8.0

6.0

4.0

2.0

0.0

8.0

6.0

4.0

2.0

0.0

0

50

0

50

95% Pan

90% Pan

75% Pan

50% Pan

25% Pan

95% Pan

90% Pan

75% Pan

50% Pan

50

25

S. aureus

0

50

100

150

200

0

50

100

150

200

250

0

50

100

150

200

[Pan+N5-Pan] (μM)

[PantSH] (μM)

[Pan+PantSH] (μM)

Fig. 8. The interaction of SaPanK_IIwith PantSH mimics that of N5-Pan. (A) Fitting Eqn (3) to the data obtained from SaPanK_IIacting on

mixtures of Pan and N5-Pan (Fig. 3B); the solid lines represent the best ﬁt curves, whereas the shaded areas indicate the 90% conﬁdence

intervals. The insert shows the stimulatory effect caused by the presence of 5% N5-Pan (purple) compared to the 100% Pan curve (black).

(B) Inhibition proﬁles for EcPanK_Iand SaPanK_IItreated with PantSH at a ﬁxed concentration of 25 lM Pan. The data points represent the

mean of two independent experiments, each performed in triplicate; the error bars denote the range/2. The solid lines represent the best

ﬁts to the Michaelis–Menten equation (for EcPanK_I) or Eqn (S5) (for SaPanK_II). (C) Kinetic proﬁles of mixtures of Pan and PantSH

determined with SaPanK_IIperformed and analyzed as for (A).

that is dependent on their metabolic activation to

form inhibitory CoA antimetabolites and/or inactive

crypto-ACPs, and that these act as inhibitors of tar-

get(s) downstream of CoA biosynthesis. Such an

analysis is in agreement with the conclusions of pre-

vious studies [8,12] and is supported by recent crys-

tallographic data showing that growth inhibitory

pantothenamides bind to PanK_Ienzymes in a man-

ner similar to Pan [38,39].

low concentrations but turning into uncompetitive

inhibitors as their concentrations gradually increase.

Additionally, we show that a pantothenamide ana-

logue that cannot act as PanK substrates can still inhi-

bit S. aureus growth. Taken together, these results

clearly show that, in this organism, growth inhibition

is likely the result of several factors working in combi-

nation: ﬁrst, by being converted into CoA antimetabo-

lites and crypto-ACPs (and having concomitant

negative effects on fatty acid biosynthesis as was

found in a previous study) [14] and, second, by reduc-

ing CoA levels by inhibiting the ﬁrst step of its bio-

synthesis. This multifaceted mode of action is most

likely one of the reasons why the pantothenamides

(such as N7-Pan) show much higher potency against

For S. aureus, which is the only bacterium known

to have an active type II PanK, the situation is very

different. Speciﬁcally, we demonstrate that growth

inhibitory pantothenamides have a complex interac-

tion with this organism’s PanK enzyme, acting as

substrates that stimulate its activity when present at

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4747

PanK type affects pantothenamide mode of action

M. de Villiers et al.

S. aureus than any other bacterium that has been

tested to date.

concentrations but inhibits it at high concentrations.

Such an analysis is supported by previous ﬁndings

showing that P-PantSH does not accumulate in either

intra- or extracellular compartments, conﬁrming that

the regulation of CoA biosynthesis in S. aureus appar-

ently occurs at the level of SaPanK_II[14].

The kinetic model that was constructed for

SaPanK_IIprovides an accurate description of the vari-

ous data sets obtained for the enzyme acting on its

native substrate (Pan), on various pantothenamides

and on combinations of the two. Speciﬁcally, our

model indicates that SaPanK_IIis inhibited by the panto-

thenamides (and by PantSH) via an uncompetitive

mechanism, which would imply the existence of an

allosteric site in which n-PanAm series compounds

bind. However, only two SaPanK_IIstructures have

been deposited in the Protein Data Bank to date, nei-

ther of which has the natural substrate Pan bound.

These structures provide no indication of the likely

location of such an allosteric site. Moreover, the more

recently published structure [i.e. that of a ternary com-

plex of the enzyme bound to a phosphorylated pant-

othenamide (N354-Pan) and ADP (Protein Data Bank

code: 4NB4)] [39] indicates that the enzyme has dis-

tinct open and closed conformations, with the latter

preventing the product from being released; this dis-

covery complicates the structural investigation of a

potential uncompetitive inhibition mode through allo-

steric inhibition even further. Consequently, we cannot

exclude the possibility that alternative kinetic models

(including those that do not need to invoke the exis-

tence of an allosteric site) could also provide accurate

descriptions of our data. Nonetheless, our proposed

mechanism is the simplest one giving an accurate

description of the total data set at the same time as

taking all the information that is currently available

on the enzyme into account.

Clearly, such a conclusion has important implica-

tions for S. aureus physiology that will have to be

explored further, especially in light of the fact that the

CoA and ACP degradation pathways responsible for

PantSH formation remain very poorly characterized in

all organisms [2]. Nonetheless, the evidence reported in

the present study suggests that the speciﬁc interaction

of SaPanK_IIwith PantSH presents a unique opportu-

nity for the development of new antistaphylococcal

agents, one which the pantothenamides have already

started to exploit.

Materials and methods

General materials and methods

All the pantothenamides and precursors were prepared and

their purity conﬁrmed by ¹H NMR analysis, as described

previously [23]. The pantothenamides were dissolved in

50% acetonitrile–water solution to yield stock solutions at

a concentration of 50 m^Mand assays were performed with

the ﬁnal acetonitrile concentration never exceeding 3% (v/

v). Pantetheine was obtained by the reduction of the disul-

ﬁde pantethine (Sigma-Aldrich, St Louis, MO, USA) in the

presence of 1.5 equivalents of dithiothreitol or tris(2-carb-

oxyethyl)phosphine. General chemicals, reagents and media

were purchased from Sigma-Aldrich, Merck Chemicals

(Darmstadt, Germany) or Acros Organics (Thermoﬁsher,

Fair Lawn, NJ, USA) and were of the highest purity. Sol-

vents used for reactions were Chromasolv HPLC grade sol-

vents (Sigma-Aldrich) and the hexanes, dichloromethane

and ethyl acetate used for puriﬁcation were purchased from

Merck Chemicals. Dry N,N-dimethylformamide was pre-

pared by shaking up over potassium hydroxide, distilled

under reduced pressure and a nitrogen atmosphere, and

Regardless of the mechanism or structural basis of

the interaction between the growth inhibitory pantothen-

amides and SaPanK_II, our ﬁnding that PantSH (the

only pantothenamide known to occur naturally) [40,41]

mirrors this complex and speciﬁc interaction suggests

that it is a native ligand of the enzyme. Such a character-

istic could be a compensatory regulatory mechanism for

SaPanK_IIbecause, unlike other type I and type II

PanKs, this enzyme does not experience any feedback

inhibition by CoA or its thioesters [14]. This is most

likely as result of S. aureus maintaining higher levels of

intracellular CoA as part of its unique redox biology

because it does not contain any glutathione but instead

uses CoA (and most likely the recently discovered

low molecular weight thiol bacillithiol), as well as a

highly speciﬁc CoA disulﬁde reductase, to maintain its

intracellular redox potential [42,43]. Our ﬁndings sug-

gest that CoA biosynthesis in S. aureus may be regu-

ꢁ

ﬁnally stored over 4-A molecular sieves in the dark. Dry

dichloromethane was distilled from CaH₂under a nitrogen

atmosphere.

The E. coli K12 strain was available in our laboratory,

whereas S. aureus RN4220 and P. aeruginosa ATCC 27853

where kind gifts from L. M. T. Dicks at the Department of

Microbiology (Stellenbosch University). EcPanK_Iand

SaPanK_IIwere overexpressed and puriﬁed as described pre-

viously [7,14,34]. PK and LDH used in the kinetic assays

were obtained from Roche (Basel, Switzerland).

1

All H and ¹³C NMR spectra were obtained using a 300-

MHz Varian VNMRS (75 MHz for ¹³C) or 400-MHz

Varian Unity Inova (100 MHz for ¹³C) instruments (Varian

lated by a unique mechanism in which PantSH

stimulates SaPanK_IIactivity when it is present at low

4748

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

Inc., Palo Alto, CA, USA) at the Central Analytical Facil-

ity (CAF) of the University of Stellenbosch. All chemical

shifts (d) were recorded using the residual solvent peak

and reported in p.p.m. All high resolution mass spectrome-

try (HRMS) was performed on a Waters API Q-TOF

Ultima spectrometer (Waters, Milford, MA, USA) at the

MS unit of CAF. All OD₆₀₀measurements and kinetic

studies were performed using a Thermo Varioskan spectro-

photometer (Thermo Scientiﬁc, Bremen, Germany). Inhibi-

tion studies were performed in Greiner Bio-One Cellstar

ﬂat-bottomed 96-well suspension plates (Greiner Bio-One

GmbH, Frickenhausen, Germany). Kinetic studies were

performed in Greiner Bio-One polystyrene ﬂat-bottomed

96-well plates. The MICs were determined by plotting the

percentage (relative to a control containing no inhibitor)

bacterial cell growth against the logarithm of increasing

compound concentration, followed by ﬁtting Eqn (5) to

the data (with the values of y₀and a generally ﬁxed at 0

and 100, respectively).

128.1, 128.5, 136.5, 156.8 and 171.3. HRMS: m/z [M+H]⁺

293.1866 (calculated [C₁₆H₂₅N₂O₃]⁺= 293.1860).

3-Amino-N-pentylpropanamide

Benzyl-3-oxo-3(pentylamino)propylcarbamate

(1.50 g,

5.13 mmol) was dissolved in methanol (80 mL) at room

temperature followed by the addition of palladium on acti-

vated carbon (Pd/C) (80.0 mg, 0.752 mmol). The reaction

mixture was stirred overnight at room temperature under a

H₂gas balloon. Additional Pd/C (100 mg, 0.940 mmol)

was added and the reaction mixture was stirred for a fur-

ther 5 h under H₂. The reaction mixture was ﬁltered and

concentrated in vacuo to give the desired amide (800 mg,

99%) as a white solid. R_f= 0.05 (10% methanol/dichlo-

romethane). d_H(300 MHz; CDCl₃) 0.87 (3H, t, J = 7.1 Hz,

–CH₃), 1.29–1.33 (4H, m, –CH₂–), 1.50 (2H, m, –CH₂–),

2.28 (2H, t, J = 5.6 Hz, –CH₂–), 2.98 (2H, t, J = 6.1 Hz, –

CH₂–), 3.20 (2H, q, J = 7.3, 12.9 Hz, –CH₂–) and 6.85

(1H, br s, –NH–). d_C(300 MHz; CDCl₃; 25 °C) 14.4, 22.7,

29.5, 29.6, 38.0 38.3, 39.7 and 172.4. HRMS: m/z [M+H]⁺

159.1499 (calculated [C₈H₁₉N₂O]⁺= 159.1492).

a

y ¼ y₀þ

ð5Þ

ꢀ

ꢁ

b

x

IC₅₀

1 þ

Kinetic data were ﬁtted according to the derived equa-

tions. Analysis of all concentration–response curves

and kinetic data was carried out using ^SIGMAPLOT, ver-

sion 12 (Systat Software Inc., Chicago, IL, USA) or

MATHEMATICA (Wolfram Research, Champaign, IL,

USA).

3,3-Dimethyl-2-oxo-butyric acid

A solution of pinacolone (10 mL, 83.4 mmol) in water

(160 mL) was added to a solution of NaOH (8.00 g, 200

mmol) and potassium permanganate (12.2 g, 77.2 mmol) in

water (196 mL) at 0 °C. The reaction mixture was stirred for

1 h at 0 °C, allowed to warm up to room temperature and

stirred for an additional 2 h. The reaction mixture was

ﬁltered through celite, acidiﬁed to pH 2 with concentrated

sulphuric acid, and the aqueous layer was extracted with

diethyl ether (3 9 40 mL). The combined organic extracts

were dried (Na₂SO₄), ﬁltered and concentrated in vacuo.

Bulb-to-bulb distillation (110 °C, 27.5 mmHg) gave the acid

(3.99 g, 77%) as a clear oil. d_H(300 MHz; CDCl₃) 1.23–1.33

(9H, m, –CH₃). ¹H NMR data are consistent with those

reported previously [44].

Synthesis of dN5

Benzyl-3-oxo-3(pentylamino)propylcarbamate

Pentylamine (803 lL, 6.96 mmol) and diethyl phosphoryl

cyanide (1.05 mL, 7.25 mmol) were added to a solution of

CBz-b-alanine (1.41 g, 6.32 mmol) in dry N,Ndimethylform-

amide (8 mL) at room temperature. The reaction mixture

was cooled to 0 °C before triethylamine (1.85 mL,

13.3 mmol) was added. The reaction mixture was stirred

for 2 h at 0 °C and left to stir overnight at room tempera-

ture. Ethyl acetate (50 mL) was added and the organic

layer was washed sequentially with 5% citric acid

(3 9 10 mL), 1 ^Maqueous NaHCO₃(2 9 10 mL) and sat-

urated sodium chloride (1 9 10 mL). The organic layer

was dried (Na₂SO₄), ﬁltered and concentrated in vacuo

before puriﬁcation by ﬂash column chromatography (3:2 to

4:1 ethyl acetate:hexanes) afforded the carbamate (1.50 g,

81%) as a white solid. R_f= 0.18. d_H(300 MHz; CDCl₃)

0.87 (3H, t, J = 6.9 Hz, –CH₃), 1.26–1.33 (4H, m, –CH₂–),

1.54 (2H, m, –CH₂–), 2.38 (2H, t, J = 6.0 Hz, –CH₂–), 3.19

(2H, q, J = 7.2, 12.9 Hz, –CH₂–), 3.45 (2H, q, J = 6.2,

11.8 Hz, –CH₂–), 5.09 (2H, s, –CH₂–), 5.34 (2H, br s, –

NH–) and 7.32–7.36 (5H, m, –CH–). d_C(400 MHz; CDCl₃;

25 °C) 14.0, 22.3, 29.0, 29.2, 36.1, 37.1, 39.5, 66.6, 128.0,

3,3-Dimethyl-2-oxo-N-[3-oxo-3-(pentylamino)propyl]

butanamide

N,N-Diisopropylethylamine (1.23 mL, 7.06 mmol) was

added dropwise over 5 min to a solution of 3-amino-N-

pentylpropanamide (800 mg, 5.06 mmol) in dichlorome-

thane (40 mL) at 0 °C. Hydroxybenzotriazole (172.2 mg,

1.27 mmol), 3,3-dimethyl-2-oxo-butyric acid (820 mg,

6.30 mmol) and 1-ethyl-3-(3-dimethylaminopropyl) carbo-

diimide hydrochloride (1.34 g, 6.99 mmol) were then added

consecutively and the reaction mixture was stirred over-

night at room temperature. The reaction was quenched by

the addition of 3 ^MHCl (50 mL) and the organic layer was

washed with 3 ^MHCl (1 9 50 mL) and saturated NaHCO₃

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4749

PanK type affects pantothenamide mode of action

M. de Villiers et al.

(1 9 50 mL). The organic layer was dried (Na₂SO₄), ﬁl-

tered and concentrated in vacuo before puriﬁcation by ﬂash

column chromatography (0.75 : 1 ethyl acetate : hexanes)

afforded the desired amide (700 mg, 57%) as a white solid.

R_f= 0.27. d_H(300 MHz; CDCl₃) 0.75 (3H, t, J = 7.0 Hz, –

CH₃), 1.14–1.19 (13H, m, –CH₃–, –CH₂–), 1.33–1.40 (2H,

m, –CH₂–), 2.34 (2H, t, J = 6.1 Hz, –CH₂–), 3.07 (2H, q,

J = 7.0, 13.1 Hz, –CH₂–), 3.42 (2H, q, J = 6.1, 12.6 Hz, –

CH₂–), 5.57 (1H, br s, –NH–) and 7.47 (1H, br s, –NH–).

d_C(400 MHz; CDCl₃; 25 °C) 13.9, 22.3, 26.2, 29.0, 29.2,

35.2, 35.4, 39.6, 42.9, 160.4 and 170.6, 203.0. HRMS: m/z

[M+H]⁺271.2014 (calculated [C₁₄H₂₇N₂O₃]⁺= 271.2016).

(diluted from 50 m^Mstock solutions prepared in 50% aque-

ous acetonitrile to aid solubility). The compounds that

exhibited positive results in initial screens at 200 l^Mwere

subsequently tested at 50 l^M; those that still showed inhibi-

tion at this concentration were submitted to concentration–

response analysis. Final concentrations of compounds were

in the range 0.039–200 l^Mdepending on the potency of the

pantothenamide. The plates were incubated at 37 °C for

20 h before the cell densities were measured (OD₆₀₀). The

extent of growth in each well was determined by normaliz-

ing the OD₆₀₀values relative to those of the negative control

(containing 3% acetonitrile instead of pantothenamide),

which was taken as 100% bacterial cell growth. Each com-

pound was tested in triplicate and all experiments were

repeated at least once after the initial experiment.

DL-4⁰-Deoxy-N-pentylpantothenamide

3,3-Dimethyl-2-oxo-N-[3-oxo-3-(pentylamino)propyl]bu-

tanamide (700 mg, 2.89 mmol) was dissolved in methanol

(28 mL) at 0 °C under an inert atmosphere followed by the

addition of sodium borohydride (NaBH₄, 163.44 mg,

4.32 mmol) in portions. The reaction mixture was stirred for

1 h at 0 °C and left to stir overnight at room temperature.

Additional NaBH₄(163.44 mg, 4.32 mmol) was added and

the reaction mixture was stirred for another 3 h at room tem-

perature. The reaction was quenched by the addition of satu-

rated ammonium chloride (28 mL) and methanol was

removed in vacuo. The aqueous solution was extracted with

ethyl acetate (3 9 30 mL) and the combined organic extracts

were dried (Na₂SO₄), ﬁltered and concentrated in vacuo to

yield dN5 (760 mg, 96%) as a colourless oil. R_f= 0.39 (10%

methanol/dichloromethane). d_H(300 MHz; CDCl₃): 0.83

(3H, t, J = 7.1 Hz, –CH₃), 0.93–0.95 (9H, m, –CH₃–), 1.24–

1.27 (4H, m, –CH₂–), 1.43–1.48 (2H, m, –CH₂–), 2.37 (2H, t,

J = 6.2 Hz, –CH₂–), 3.13 (2H, q, J = 7.0, 13.6 Hz, –CH₂–),

3.47 (2H, q, J = 6.6, 12.7 Hz, –CH₂–), 3.66 (1H, s, –CH–),

6.38 (1H, br s, –NH–) and 7.17 (1H, br s, –NH–); d_C

(400 MHz; CDCl₃; 25 °C): 13.9, 22.3, 25.9, 29.0, 29.2, 35.0,

35.2, 35.6, 39.6, 79.4, 171.1 and 172.7. HRMS: m/z [M+H]⁺

273.2172 (calculated [C₁₄H₂₉N₂O₃]⁺= 273.2173).

Bacterial growth inhibition studies of N5-Pan and

dN5 in minimal media

The inhibition of E. coli K12 and S. aureus RN4220 by N5-

Pan and dN5 was tested in minimal media appropriate for

each bacterium. For E. coli, this medium consisted of

0.8 m^MMgSO₄, 10 mM citric acid, 60 mM K₂HPO₄, 20 mM

NaNH₄HPO₄and 0.5% glucose; for S. aureus, it contained

40 m^MKCl, 160 mM NaCl, 5.3 mM MgSO₄, 30 mM

(NH₄)₂SO₄, 0.11 mM CaCl₂, 1.0 mM KH₂PO₄, 0.02 mM

FeSO₄, 0.04 mM MnSO₄, 0.03 mM citric acid, 100 mM Tris,

28 m^Mglucose, 0.8 mM L-arginine, 1.0 mM L-proline,

1.9 m^{M L}-glutamic acid, 1.5 mM L-valine, 1.5 mM L-threo-

nine, 1.0 m^ML-phenylalanine, 1.3 mM L-leucine, 0.78 mM

^L-cysteine, 0.4 lM biotin, 6.6 lM thiamine and 16 lM

nicotinic acid [45]. The inhibition was determined by con-

centration–response analysis as described above with minor

modiﬁcations. Brieﬂy, starter cultures of E. coli K12 and

S. aureus RN4220 in 1% tryptone were inoculated with four

separate colonies grown on LB agar plates. The starter cul-

ture was grown to exponential phase and then diluted 10-

fold into the applicable minimal media. A 10-lL aliquot of

the diluted cell suspension was used to inoculate each well of

a 96-well ﬂat-bottomed plate containing 100 lL of minimal

medium. Subsequently, the media was supplemented with

either N5-Pan or dN5. For E. coli and S. aureus, the ﬁnal

concentrations of N5-Pan varied in the range 0.781–100 l^M

and 0.391–50 lM, respectively; for both organisms, dN5 was

varied in the range 1.56–200 l^M. The plates were incubated

at 37 °C for 24 h before the cell densities were measured

(OD₆₀₀) and analyzed further as described above. Each com-

pound was tested in either two (E. coli) or three (S. aureus)

independent experiments, each performed triplicate.

Bacterial growth inhibition tests of

pantothenamide library

The MICs of the library of N-substituted pantothenamides

(Table 1, entries 1–47) against E. coli K12, S. aureus

RN4220 or P. aeruginosa ATCC 27853 were determined by

microbroth dilution in 96-well microtiter ﬂat-bottomed

plates and turbidometric analysis at OD₆₀₀. Starter cultures

of either E. coli K12, S. aureus RN4220 or P. aeruginosa

ATCC 27853 in 1% tryptone were prepared by inoculation

with four separate colonies grown on LB agar plates. The

starter culture was grown to exponential phase and then

diluted 10 000-fold in the same medium. A 10-lL aliquot

of the diluted cell suspension was used to inoculate each

well of a 96-well plate containing 100 lL of 1% tryptone

broth supplemented with the speciﬁc compound of interest

PanK assays

PanK activity was determined using a continuous spectro-

photometric assay that coupled the production of ADP to

the consumption of NADH and was monitored by the

4750

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

decrease in A₃₄₀, as described previously [7]. An extinction

ꢃ1

analysis. ES conceived the project, contributed to the

data analysis and wrote the paper with contributions

from all of the authors.

coefﬁcient of 6220 ^Mꢁcm^ꢃ1was used for NADH. Each

300-lL reaction mixture contained 50 m^MTris-HCl

(pH 7.6), 10 m^MMgCl₂, 20 mM KCl, 1.5 mM ATP, 0.5 mM

NADH, 0.5 mM phosphoenolpyruvate, 3 units of PK,

3 units of LDH and either 3.0 lg of EcPanK_Ior 1.5 lg of

SaPanK_II. When PantSH was tested as substrate, 0.5 mM of

tris(2-carboxyethyl)phosphine was also added to the reaction

mixture. The concentration ranges (and ratios of substrate

used in the mixture experiments) are indicated as appropri-

ate. The reaction was initiated by the addition of substrate

(or mixtures of substrates, or mixtures of substrates and

inhibitors) and was monitored for 5 min at 25 °C.

References

1 Clifton G, Bryant SR & Skinner CG (1970) N¹-

(substituted) pantothenamides, antimetabolites of

pantothenic acid. Arch Biochem Biophys 137, 523–528.

2 Strauss E (2010) Coenzyme A biosynthesis and

enzymology. In Comprehensive Natural Products II

Chemistry and Biology (Mander L & Liu H-W, eds),

pp. 351–410. Elsevier, Oxford.

3 Spry C, Kirk K & Saliba KJ (2008) Coenzyme A

biosynthesis: an antimicrobial drug target. FEMS

Microbiol Rev 32, 56–106.

Statistical analysis

Using the raw kinetic data, initial velocities were calculated

for each substrate concentration (or substrate mixture, or

substrate/inhibitor mixture) by linear regression of the read-

ings made in the 50s period after the initial 10 s (i.e. the

period from 10 to 60 s after the reaction was started). For

each experiment, the three readings made for each data

point were averaged and plotted with the SD to give the

respective kinetic proﬁle. Kinetic parameters reported in

Table 2 were determined for each experiment by ﬁtting the

appropriate equation to the data; the reported values are

the mean values of the parameters determined from all the

individual independent experiments, and are given with

errors that indicate the range/2 or SEM as appropriate. The

parameters (Table 3) were determined in a global ﬁt by

minimizing the squared difference between experimental

data points and model simulation of the appropriate equa-

tion, and are given with errors that indicate the SE.

4 Spry C, Macuamule C, Lin Z, Virga KG, Lee RE,

Strauss E & Saliba KJ (2013) Pantothenamides are

potent, on-target inhibitors of Plasmodium falciparum

growth when serum pantetheinase is inactivated. PLoS

One 8, e54974.

5 de Villiers M, Macuamule C, Spry C, Hyun Y-M,

Strauss E & Saliba KJ (2013) Structural modiﬁcation of

pantothenamides counteracts degradation by

pantetheinase and improves antiplasmodial activity.

ACS Med Chem Lett 4, 784–789.

6 Jansen PAM, Hermkens PHH, Zeeuwen PLJM,

Botman PNM, Blaauw RH, Burghout P, van Galen

PM, Mouton JW, Rutjes FPJT & Schalkwijk J (2013)

Combination of pantothenamides with vanin

inhibitors as a novel antibiotic strategy against Gram-

positive bacteria. Antimicrob Agents Chemother 57,

4794–4800.

7 Strauss E & Begley TP (2002) The antibiotic activity of

N-pentylpantothenamide results from its conversion to

ethyldethia-coenzyme A, a coenzyme A antimetabolite.

J Biol Chem 277, 48205–48209.

Acknowledgements

We thank Professor Willem van Otterlo for providing

advice and guidance regarding the synthesis of dN5.

This project was funded by grants from the National

Research Foundation (NRF) of South Africa

(CPR20110726000022216) and Medical Research

Council of South Africa to ES. JLS is supported by

the South African Research Chairs Initiative of the

Department of Science and Technology and the NRF

of South Africa. MdV, LB and LK were supported by

bursaries from the NRF.

8 Zhang Y-M, Frank MW, Virga KG, Lee RE, Rock CO

& Jackowski S (2004) Acyl carrier protein is a cellular

target for the antibacterial action of the

pantothenamide class of pantothenate antimetabolites.

J Biol Chem 279, 50969–50975.

9 Mercer AC & Burkart MD (2007) The ubiquitous

carrier protein – a window to metabolite biosynthesis.

Nat Prod Rep 24, 750–773.

10 Jackowski S & Rock CO (1984) Turnover of the 4⁰-

phosphopantetheine prosthetic group of acyl carrier

protein. J Biol Chem 259, 1891–1895.

11 Thomas J & Cronan JE (2005) The enigmatic acyl

carrier protein phosphodiesterase of Escherichia coli.

J Biol Chem 280, 34675–34683.

Author contributions

MdV and LB contributed equally to the syntheses,

growth inhibition assays and enzyme kinetic analyses,

as well as to the data analysis. LK performed enzyme

kinetic assays and data analysis. JLS constructed the

kinetic model and performed the associated data

12 Thomas J & Cronan JE (2010) Antibacterial activity of

N-pentylpantothenamide is due to inhibition of

coenzyme A synthesis. Antimicrob Agents Chemother

54, 1374–1377.

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4751

PanK type affects pantothenamide mode of action

M. de Villiers et al.

13 Choudhry AE, Mandichak TL, Broskey JP, Egolf

RW, Kinsland C, Begley TP, Seefeld MA, Ku TW,

Brown JR, Zalacain M et al. (2003) Inhibitors of

pantothenate kinase: novel antibiotics for

RE, Rock CO et al. (2007) Chemical knockout of

pantothenate kinase reveals the metabolic and genetic

program responsible for hepatic coenzyme A

homeostasis. Chem Biol 14, 291–302.

staphylococcal infections. Antimicrob Agents

Chemother 47, 2051–2055.

25 Siudeja K, Srinivasan B, Xu L, Rana A, de Jong J,

Nollen EAA, Jackowski S, Sanford L, Hayﬂick S &

Sibon OCM (2011) Impaired coenzyme A metabolism

affects histone and tubulin acetylation in Drosophila

and human cell models of pantothenate kinase

associated neurodegeneration. EMBO Mol Med 3, 755–

766.

14 Leonardi R, Chohnan S, Zhang Y-M, Virga KG, Lee

RE, Rock CO & Jackowski S (2005) A pantothenate

kinase from Staphylococcus aureus refractory to

feedback regulation by coenzyme A. J Biol Chem 280,

3314–3322.

15 Virga KG, Zhang Y-M, Leonardi R, Ivey RA, Hevener

K, Park H-W, Jackowski S, Rock CO & Lee RE (2006)

Structure–activity relationships and enzyme inhibition

of pantothenamide-type pantothenate kinase inhibitors.

Bioorg Med Chem 14, 1007–1020.

26 Ivey RA, Zhang Y-M, Virga KG, Hevener K, Lee RE,

Rock CO, Jackowski S & Park H-W (2004) The

structure of the pantothenate kinase.ADP.pantothenate

ternary complex reveals the relationship between the

binding sites for substrate, allosteric regulator, and

antimetabolites. J Biol Chem 279, 35622–35629.

27 van der Westhuyzen R, Hammons JC, Meier JL,

Dahesh S, Moolman WJA, Pelly SC, Nizet V, Burkart

MD & Strauss E (2012) The antibiotic CJ-15,801 is an

antimetabolite that hijacks and then inhibits CoA

biosynthesis. Chem Biol 19, 559–571.

16 Mercer AC, Meier JL, Hur GH, Smith AR & Burkart

MD (2008) Antibiotic evaluation and in vivo analysis of

alkynyl coenzyme A antimetabolites in Escherichia coli.

Bioorg Med Chem Lett 18, 5991–5994.

17 Brand LA & Strauss E (2005) Characterization of a

new pantothenate kinase isoform from Helicobacter

pylori. J Biol Chem 280, 20185–20188.

28 Wittwer CT, Beck S, Peterson M, Davidson R, Wilson

DE & Hansen RG (1990) Mild pantothenate deﬁciency

in rats elevates serum triglyceride and free fatty acid

levels. J Nutr 120, 719–725.

18 Nicely NI, Parsonage D, Paige C, Newton GL, Fahey

RC, Leonardi R, Jackowski S, Mallett TC & Claiborne

A (2007) Structure of the type III pantothenate kinase

ꢁ

from Bacillus anthracis at 2.0 A resolution: implications

29 Wittwer CT, Schweitzer C, Pearson J, Song WO,

Windham CT, Wyse BW & Hansen RG (1989)

Enzymes for liberation of pantothenic acid in blood:

use of plasma pantetheinase. Am J Clin Nutr 50, 1072–

1078.

for coenzyme A-dependent redox biology. Biochemistry

46, 3234–3245.

19 Yang K, Eyobo Y, Brand LA, Martynowski D,

Tomchick D, Strauss E & Zhang H (2006) Crystal

structure of a type III pantothenate kinase: insight into

the mechanism of an essential coenzyme A biosynthetic

enzyme universally distributed in bacteria. J Bacteriol

188, 5532–5540.

30 Shibata K, Fukuwatari T, Ohta M, Okamoto H,

Watanabe T, Fukui T, Nishimuta M, Totani M,

Kimura M, Ohishi N et al. (2005) Values of water-

soluble vitamins in blood and urine of Japanese young

men and women consuming a semi-puriﬁed diet based

on the Japanese dietary reference intakes. J Nutr Sci

Vitaminol 51, 319–328.

20 Hong BS, Yun MK, Zhang Y-M, Chohnan S, Rock

CO, White SW, Jackowski S, Park H-W & Leonardi R

(2006) Prokaryotic type II and type III pantothenate

kinases: the same monomer fold creates dimers with

distinct catalytic properties. Structure 14, 1251–1261.

21 Strauss E, de Villiers M & Rootman I (2010)

Biocatalytic production of coenzyme A analogues.

ChemCatChem 2, 929–937.

31 Spry C, Saliba KJ & Strauss E (2014) A miniaturized

assay for measuring small molecule phosphorylation in

the presence of complex matrices. Anal Biochem 451,

76–78.

32 Meier JL, Mercer AC, Rivera H Jr & Burkart MD

(2006) Synthesis and evaluation of bioorthogonal

pantetheine analogues for in vivo protein modiﬁcation.

J Am Chem Soc 128, 12174–12184.

22 Balibar CJ, Hollis-Symynkywicz MF & Tao J (2011)

Pantethine rescues phosphopantothenoylcysteine

synthetase and phosphopantothenoylcysteine

decarboxylase deﬁciency in Escherichia coli but not in

Pseudomonas aeruginosa. J Bacteriol 193, 3304–3312.

23 van Wyk M & Strauss E (2008) Development of a

method for the parallel synthesis and puriﬁcation of

N-substituted pantothenamides, known inhibitors of

coenzyme A biosynthesis and utilization. Org Biomol

Chem 6, 4348–4355.

33 Worthington AS & Burkart MD (2006) One-pot

chemo-enzymatic synthesis of reporter-modiﬁed

proteins. Org Biomol Chem 4, 44–46.

34 van Wyk M & Strauss E (2007) One-pot preparation of

coenzyme A analogues via an improved chemo-

enzymatic synthesis of pre-CoA thioester synthons.

Chem Commun, 398–400.

24 Zhang Y-M, Chohnan S, Virga KG, Stevens RD,

Ilkayeva OR, Wenner BR, Bain JR, Newgard CB, Lee

35 Yang K, Strauss E, Huerta C & Zhang H (2008)

Structural basis for substrate biding and the catalytic

4752

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

M. de Villiers et al.

PanK type affects pantothenamide mode of action

mechanism of type III pantothenate kinase.

Biochemistry 47, 1369–1380.

40 Peters VJ, Brown GM, Williams WL & Snell EE (1953)

Isolation of the Lactobacillus bulgaricus factor from

natural sources. J Am Chem Soc 75, 1688–1691.

41 Jackowski S & Rock CO (1984) Metabolism of 4⁰-

phosphopantetheine in Escherichia coli. J Bacteriol 158,

115–120.

36 Yun M, Park C-G, Kim J-Y, Rock CO, Jackowski S &

Park H-W (2000) Structural basis for the feedback

regulation of Escherichia coli pantothenate kinase by

coenzyme A. J Biol Chem 275, 28093–28099.

37 Akinnusi TO, Vong K & Auclair K (2011) Geminal

dialkyl derivatives of N-substituted pantothenamides:

synthesis and antibacterial activity. Bioorg Med Chem

19, 2696–2706.

42 Van Laer K, Hamilton CJ & Messens J (2013) Low-

molecular-weight thiols in thiol-disulﬁde exchange.

Antioxid Redox Signal 18, 1642–1653.

43 Gaupp R, Ledala N & Somerville GA (2012)

Staphylococcal response to oxidative stress. Front Cell

Info Microbiol 2, 33.

38 Li B, Tempel W, Smil D, Bolshan Y, Schapira M &

Park H-W (2013) Crystal structures of Klebsiella

pneumoniae pantothenate kinase in complex with N-

substituted pantothenamides. Proteins 81, 1466–1472.

39 Hughes SJ, Antoshchenko T, Kim KP, Smil D & Park

H-W (2014) Structural characterization of a new N-

substituted pantothenamide bound to pantothenate

kinases from Klebsiella pneumoniae and Staphylococcus

aureus. Proteins 82, 1542–1548.

44 Tuck KL, Saldanha SA, Birch LM, Smith AG & Abell

C (2006) The design and synthesis of inhibitors of

pantothenate synthetase. Org Biomol Chem 4, 3598–

3610.

€

45 Rudin L, Sjostrom J-E, Lindberg M & Philipson L

(1974) Factors affecting competence for transformation

in Staphylococcus aureus. J Bacteriol 118, 155–164.

FEBS Journal 281 (2014) 4731–4753 ª 2014 FEBS

4753

Article Doi

Variation in pantothenate kinase type determines the pantothenamide mode of action and impacts on coenzyme A salvage biosynthesis

DOI: 10.1111/febs.13013

Source and publish data:

Authors:

Article abstract of DOI:10.1111/febs.13013

Full text of DOI:10.1111/febs.13013

Products guided by the article

R&D Labs maybe for 815-17-8

Relevant to this article

Hot Product