Allicin induces thiol stress in bacteria through S-allylmercapto modification of protein cysteines

JBC Papers in Press. Published on March 23, 2016 as Manuscript M115.702308

The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M115.702308

Allicin Induces Thiol Stress in Bacteria through S-allylmercapto Modification of Protein Cysteines*

1

‡

1‡

2

3

4

Alexandra Müller , Jakob Eller , Frank Albrecht , Pascal Prochnow , Katja Kuhlmann , Julia

3

2

1

Elisabeth Bandow , Alan John Slusarenko , Lars Ingo Ole Leichert

1

Institute of Biochemistry and Pathobiochemistry – Microbial Biochemistry, Ruhr-University Bochum,

4780 Bochum, Germany

Dept. of Plant Physiology, RWTH Aachen University, 52056 Aachen, Germany

4

2

3

Angewandte Mikrobiologie, Ruhr University Bochum, 44780 Bochum, Germany

4

Medizinisches Proteom-Center, Ruhr University Bochum, 44780 Bochum, Germany

authors contributed equally

Running Title: Allicin-modified proteins in E. coli

‡

*

To whom correspondence should be addressed: Lars I. Leichert, Ruhr-University Bochum, Institute of

Biochemistry and Pathobiochemistry, Microbial Biochemistry, Universitätsstr. 150, 44780 Bochum,

Germany, Phone: +49 234 32 24585, Fax: +49 234 32 14332, Email: lars.leichert@ruhr-uni-bochum.de

Keywords: Allicin, antibiotic action, bacteria, disulfide, Escherichia coli (E. coli), garlic (Allium

sativum), thiol, S-allylmercapto modification, phytoanticipin, redox proteomics

ABSTRACT

Allicin (diallyl thiosulfinate) from garlic

highly potent natural antimicrobial

of glutathione levels, unfolding stress, and

inactivation of crucial metabolic enzymes

through S-allylmercapto modification of

cysteines.

is

a

substance. It inhibits growth of a variety of

microorganisms, among them antibiotic

resistant strains. However, the precise mode of

action of allicin is unknown. Here, we show that

growth inhibition of Escherichia coli during

allicin exposure coincides with a depletion of

the glutathione pool and S-allylmercapto

modification of proteins, resulting in overall

decreased total sulfhydryl levels. This is

accompanied by the induction of the oxidative

and heat stress response. We identified and

quantified the allicin-induced modification S-

allylmercapto cysteine for a set of cytoplasmic

proteins by using a combination of label-free

mass spectrometry and differential OxICAT

labeling. Activity of isocitrate lyase AceA, an S-

allylmercapto-modified candidate protein, is

largely inhibited by allicin treatment in vivo.

Allicin-induced protein modifications trigger

protein aggregation, which largely stabilizes

RpoH and thereby induces the heat stress

response. At sublethal concentrations, the heat

stress response is crucial to overcome allicin

stress. Our results indicate that the mode of

action of allicin is a combination of a decrease

The treatment of infectious diseases is,

and has been, one of the main challenges faced by

medicine. Prior to the discovery of the first

antibiotics, treatment of infections mainly relied

on herbal medication. A plant that frequently was

considered as effective against a variety of

infectious diseases is garlic (Allium sativum). It

has traditionally been used in many cultures to

treat a variety of human medical conditions, such

as lung diseases, arthritis, and the common cold

(1).

In 1944, diallyl thiosulfinate, the main

antibacterially active ingredient of garlic, was

isolated and its chemical structure was described

(

2, 3). Due to its origin, diallyl thiosulfinate was

named allicin. Allicin is a phytoanticipin released

from the inactive odourless precursor alliin when

garlic tissues are damaged by pathogens or pests.

In intact allium plant tissue, the inactive precursor

alliin is present in the cytoplasm, separate from the

C-S lyase alliinase, which is stored in the vacuoles

(

4). Upon mechanical damage by herbivores or

due to infections by microorganisms, the vacuoles

are disrupted and alliin is converted to allyl

sulfenic acid and dihydroalanine by alliinase (5,

). Two molecules of allyl sulfenic acid

spontaneously condense to allicin (Fig. 1).

The inhibitory effect of allicin against a

variety of microorganisms, such as methicillin-

allicin, synthesized from diallyl disulfide and

hydrogen peroxide, inhibits growth of a broad

range of pathogenic microorganisms, including

Gram-positive and Gram-negative bacteria as well

as fungi. Treatment of E. coli with a sublethal

amount of allicin results in a significant reduction

of total free sulfhydryl levels. This is, at least in

part, due to a decrease of reduced glutathione. On

a proteomic level, allicin treatment induces the

heat stress and oxidative stress responses. Mass

spectrometry revealed that allicin causes

substantial S-allylmercapto cystein modification of

proteins in vivo. We were able to quantify the

extent of modification for a subset of S-

allylmercapto-modified proteins. Activity of

isocitrate lyase, a major S-allylmercapto-modified

candidate protein, was largely inhibited by allicin

exposure. On a global scale, allicin-induced

protein modifications resulted in enhanced protein

aggregation, which in turn causes induction of the

heat stress response through stabilization of the

6

resistant

Staphylococcus

aureus

(MRSA),

Streptococcus spp., Escherichia coli, and

Salmonella typhimurium as well as certain fungi

and parasites has been studied extensively (2, 3, 7-

1

2). Based on the results of these studies, allicin

can be considered a broad-range antimicrobial

compound. The exact mode of antimicrobial

action of allicin is not yet fully understood. As

allicin is a hydrophobic substance that can easily

pass membranes, it is likely that it can directly

react with microbial molecules (13). In vitro

studies using pure allicin demonstrated the

inhibition of several enzymes, including

thioredoxin reductase, papain, and alcohol

dehydrogenase (14). As this inhibition could be

rescued by the addition of reductants, such as DTT

or 2-mercaptoethanol, the inhibition is likely

caused by allicin-induced modification of cysteine

thiols. Upon reaction of allicin with a sulfhydryl

group, a disulfide is formed, resulting in an S-

allylmercapto modification (15). This type of

modification has been detected in vitro after

incubation of allicin with proteins, as well as with

free cysteine and other thiol-containing molecules

such as glutathione (14-16). S-allylmercapto-

modified proteinogenic cysteines could, in

principle, react further with free thiols present in

the vicinity via a thiol-disulfide exchange reaction.

This would result in the formation of intra- or

intermolecular disulfide bonds between proteins

and/or low molecular weight thiols such as

glutathione (17). Until now, detailed analyses of

the global cellular consequences of allicin on the

metabolism of microorganism as well as a global

identification of allicin's targets have not been

performed. Feldberg et al. observed an immediate

inhibition of RNA synthesis in Salmonella

typhimurium after treatment with allicin, while

protein and DNA synthesis was only delayed and

partially inhibited (9). In yeast, allicin caused an

increase in the cellular redox potential by

oxidation of the glutathione pool, ultimately

resulting in apoptosis (18).

3

2

alternative heat shock sigma factor σ (RpoH).

Our results show that the antimicrobial effect of

allicin coincides with a decrease in reduced

glutathione levels that is not accompanied by a

corresponding increase of glutathione disulfide, S-

allylmercapto modification of proteins, and

inhibition of metabolic enzymes.

EXPERIMENTAL PROCEDURES

Synthesis of allicin − Allicin was obtained

by oxidation of diallyldisulfide with H O₂as

2

previously

reported

with

the

following

modifications (18). 2 g of diallyldisulfide and

3 mL 30 % H O were dissolved in 5 mL glacial

2

acetic acid prior to shaking at room temperature

for 4 h. The reaction was stopped by addition of

25 mL Aqua dest. (Aqua destillata). Extraction

was performed two times with 30 mL of

diethylether prior to evaporation in vacuo. The

remaining volume was resuspended in 25 mL

methanol

(90 % v/v)

and

remaining

diallyldisulfide was extracted several times with

10 mL n-hexane until no diallydisulfide was

detectable by HPLC. The methanolic phase was

removed by evaporation. The remaining volume

was mixed with Aqua dest. and extracted against

diethylether. The ether phase was then dried with

dry copper sulfate until no further discoloration

from white to blue was observable. The ether was

separated from the copper sulfate by filtration

In the present study, we provide a global analysis

of allicin-induced cellular effects in the model

organism Escherichia coli. We show that pure

ꢀ

2

prior to evaporation. The resulting allicin had a

purity of 95 % as analyzed by HPLC.

Determination of sulfhydryl contents by

DTNB assay − An E. coli MG1655 overnight

culture was used to inoculate MOPS Minimal

Medium (1:100). The culture was grown

aerobically until an OD600 of 0.4 was reached. The

culture was split into three 15 mL cultures for

stress treatment. An untreated culture served as

MIC determination

−

E. coli K12

MG1655 was grown overnight in MOPS Minimal

Medium (Teknova MOPS Minimal Media Kit; 1x

MOPS mixture, 2 % (w/v) Glucose, 1.32 mM

-

1

K

2

HPO

4

, 1 µg mL thiamine) (19). 5 mL MOPS

5

-1

medium were inoculated with 10 cells of the

overnight culture. Allicin (10 mg mL stock) was

added to each tube at different end concentrations

0-35 µg mL ). Cultures were grown overnight at

7 °C on a roller drum. A non-inoculated culture

negative control. 0.79 mM allicin (128 µg mL ) or

-

1

1 mM diamide were added to one of the remaining

two cultures each. Cultures were incubated for

15 min. 5 mL of each culture were harvested by

centrifugation (8,525 g, 4 °C, 10 min). Cells were

washed twice with 1 mL PBS (137 mM NaCl,

-

1

(

3

-

1

with allicin (110 µg mL ) and a culture inoculated

with E. coli, but without allicin served as controls.

MIC determination of allicin against Escherichia

coli DSM 30083, Acinetobacter baumannii DSM

2 4 2 4

2.7 mM KCl, 10 mM Na HPO , 2 mM KH PO ,

pH 7.4, stored anaerobically prior to use) and

centrifuged (13,000 g, 4 °C, 10 min). Cells were

resuspended in lysis buffer (PBS with 6 mM

guanidinium HCl pH 7.4) prior to disruption at

3

0007, Pseudomonas aeruginosa DSM 50071,

Staphylococcus aureus DSM 20231,

Staphylococcus aureus ATCC 43300 (MRSA),

and Candida albicans DSM 1386 was performed

in a microtiter plate assay according to CLSI

guidelines, as described previously (20). E. coli,

A. baumannii, and S. aureus, were grown in

Mueller Hinton broth, P. aeruginosa in cation-

adjusted Mueller Hinton II, and C. albicans in

Sabourand medium. Allicin was dissolved in

4 °C

by

ultrasonication

(VialTweeter

ultrasonicator, Hielscher GmbH, Germany)

(3 x 1 min). Cell debris was pelleted by

centrifugation (13,000 g, 4 °C, 15 min). The

supernatant was transferred to a 3.5 mL QS-macro

cuvette (10 mm) with a magnetic stir bar and

mixed with 1 mL lysis buffer. Extinction of the

samples was monitored at 412 nm with a JASCO

V-650 spectrophotometer equipped with the

temperature-controlled cell holder PSC-718

(JASCO, Darmstadt, Germany) at room

temperature. 100 µL of a 3 mM DNTB (dithiobis-

(2-nitrobenzoic acid)) solution were added.

Extinction was monitored until it reached

saturation. Calculation of thiol concentration was

performed using _-₁the extinction coefficient

-

1

DMSO as 1.25, 2.5, 5, or 10 mg mL stock

solution. Serial dilutions in culture media were

prepared with the Tecan Freedom Evo 75 liquid

handling

workstation

(Tecan,

Männedorf,

Switzerland). Dilutions, starting from

a

-

1

0 mg mL stock solution, typically covered a

-

1

range from 512 to 0.5 µg mL . Compound

5

dilutions were inoculated with 5 x 10 bacteria per

-

1

mL from late exponential cultures grown in the

same media. Assay volumes were 200 µL per

well. Cells were incubated for 16-18 hours at

ε

412 = 13,700 M cm for TNB. Cellular thiol

concentrations were calculated based on a volume

-

15

of E. coli cells of 6.7 x 10 liters and a cell

density of (equivalent to

1 x 10 cells mL culture) (21).

3

7 °C. The lowest allicin concentration inhibiting

OD600 = 0.5

8

-1

visible bacterial growth was recorded as MIC.

Stress treatment of E. coli during

In vitro glutathione determination − GSH

and GSSG levels were determined as described

previously with only minor changes (22). To test

reactivity of allicin towards glutathione in vitro,

13 µM of reduced and oxidized glutathione in

KPE buffer (0.1 M potassium phosphate, pH 7.5,

5 mM EDTA) were incubated with 6.5 or 13 µM

allicin in a total volume of 300 µL KPE buffer for

15 or 60 min at room temperature. For the

determination of total glutathione, 100 µL reaction

mixture were added to a 1.5 mL QS-macro cuvette

exponential growth − E. coli K12 strains

MG1655, MC4100, and MC4100_ΔrpoH were

grown overnight in MOPS Minimal Medium or

LB medium. 60 mL of MOPS medium or LB

medium were inoculated to OD600 of 0.05 and

strains were grown at 25 °C or 37 °C with shaking

until OD600 of 0.5 was reached. Cultures were split

into three 20 mL cultures and treated with

0

.79 mM allicin, 1 mM diamide or DMSO

(

vehicle control). Growth of cultures was recorded

-

1

for at least 700 min.

(10 mm) containing 45 µg mL

DTNB and

ꢀ

3

-

1

0

.7 U mL

glutathione

reductase

(from

10 mL

(non-radioactive

gels)

of

an

Saccharomyces cerevisiae; Sigma-Aldrich, St.

Louis, USA) in a total volume of 880 µL KPE

buffer. After the addition of 45 µg mL β-NADPH

E. coli MG1655 culture in MOPS minimal

medium were exposed to 0.79 mM of allicin in the

early exponential growth phase. After 10 min,

-

1

3

5

the extinction of TNB was measured at 412 nm for

cells were pulse-labeled with [ S]-L-methionine

for 5 min. Precursor incorporation was stopped by

inhibition of protein biosynthesis by addition of

9

00 s in a JASCO V-650 spectrophotometer

equipped with the temperature-controlled cell

holder PSC-718 (JASCO, Darmstadt, Germany) at

room temperature. To block residual GSH, GSSG

reaction mixtures were treated with 0.4 % (v/v)

-

1

1 mg mL chloramphenicol and an excess of non-

radioactive methionine. Cells were harvested by

centrifugation, washed three times with TE buffer

(100 mM Tris, 1 mM EDTA, pH 7.5), and

2

-vinylpyridine for 60 min at room temperature

prior to addition of 0.4 % (v/v) triethanolamine.

10 µL of these reaction mixtures were used for

glutathione determination. The initial slopes in

disrupted

by

ultrasonication.

Methionine

1

incorporation rates per protein content in the

lysates were determined in duplicate for every

replicate using a Perkin Elmer Liquid Scintillation

Analyzer Tri-Carb 2800 TR. Proteins from 1 µL

sample were TCA precipitated on a Whatman

filter disc by 20 % (w/v) TCA for 20 min on ice.

Filter discs were washed twice with 10 % (w/v)

TCA, then twice with 96 % ethanol for 10 min

each. Filter discs were dried and placed into a

scintillation vial. 5 mL of scintillation solution

(Perkin Elmer Ultima Gold) were added prior to

measurement of scintillation counts. Protein

(55 µg for analytical, 300 µg for preparative gels)

was loaded onto 24 cm immobilized pH gradient

strips (pH 4-7, GE Healthcare, Uppsala, Sweden)

by passive rehydration for 18 hours. Proteins were

separated in a first dimension by isoelectric

focusing and in a second dimension by SDS-

PAGE using 12.5 % acrylamide gels. Gels were

stained with 0.003 % ruthenium(II) tris(4,7-

diphenyl-1,10-phenantrolin disulfonate) (RuBPs)

and imaged as described earlier (24). Analytical

(radioactive) gels were dried and exposed to

storage phosphor screens prior to scanning (24).

Analytical gels of controls and allicin-treated

samples were merged using Decodon Delta 2D

image analysis software (Decodon, Greifswald,

Germany). The 25 protein spots with the highest

expression level on gels after allicin treatment

were selected for protein identification. 21 of them

were found on non-radioactive gels and were

excised for mass spectrometric identification.

Protein spots were identified by nanoUPLC-ESI-

MS/MS as previously reported using a Synapt

G2-S HDMS mass spectrometer equipped with a

lock spray source for electrospray ionization and a

ToF detector (23). Mass spectrometry proteomics

data for identification of spots from 2D gels have

been deposited to the ProteomeXchange

-

1

ΔE min , (ΔE, change of absorbance at 412 nm)

of GSH and GSSG standards with a concentration

range between 26 and 0.1 µM in KPE were used to

calculate

standard

curves.

Glutathione

concentrations in the untreated control samples

were set to 100 %.

In vivo glutathione determination −

E. coli MG1655 was grown in MOPS Minimal

Medium in a total volume of 200 mL until OD600

of 0.5 was reached. The culture was split into

5

0 mL cultures for stress treatment. After 15 min

of incubation with 0.79 mM allicin, 1 mM diamide

or DMSO (control), cells were harvested at

4

,000 g at 4 °C for 10 min. Cells were washed

twice with KPE buffer prior to resuspension of

pellets in 700 µL KPE. For deproteination, 300 µL

1

0 % (w/v) sulfosalicylic acid were added prior to

disruption of cells by ultrasonication (3 x 1 min;

VialTweeter ultrasonicator, Hielscher GmbH,

Germany). Supernatants were collected after

centrifugation

(30 min,

13,000 g,

4 °C).

Sulfosalicylic acid concentrations were decreased

to 1 % by the addition of 3 volumes of KPE.

Measurements of total glutathione and GSSG was

performed as described above. Cellular glutathione

concentrations were calculated based on a volume

-

15

of E. coli cells of 6.7 x 10 liters and a cell

density of (equivalent to

x 10 cells mL culture). GSH concentrations

OD600 = 0.5

8

-1

1

were calculated by subtraction of 2 [GSSG] from

total glutathione.

2

D-PAGE analysis − Radioactive labeling

of newly synthesized proteins and subsequent

separation of the cytosolic proteome by two-

dimensional polyacrylamide gel electrophoresis

(

2D-PAGE) was performed as previously

described (23). In short, 5 mL (radioactive gels) or

ꢀ

4

Consortium (25) via the PRIDE partner repository

with the dataset identifier PXD003158.

For peer review the data can be accessed at

http://www.ebi.ac.uk/pride/archive/login

Username (“Email”): reviewer38975@ebi.ac.uk

Password: rMGfBnbK

Label-free mass spectrometric analysis of

the E. coli proteome after allicin exposure –

Culture growth, stress treatment, and cell

disruption was performed as described for the

DTNB assay. The supernatant was precipitated

with five volumes of acetone over night at -20 °C.

After centrifugation (13,000 g, 30 min, 4 °C) and

washing with five volumes of acetone, the pellet

was resuspended in ammonium bicarbonate buffer

and digested with trypsin over night (37 °C). Two

individual replicates were analyzed for allicin

treatment and one sample was analyzed for control

conditions and after diamide exposure. The LC-

MS/MS consisted of an Ultimate 3000 RSLCnano

system coupled online to an Orbitrap Elite Hybrid

Ion Trap-Orbitrap mass spectrometer (Thermo

Scientific). A total of 300 ng (in 15 µL) was

preconcentrated on a trap column (Acclaim

PepMap 100, 100 µm x 2 cm nanoViper, 5 µm,

Alkaline phosphatase activity assay − The

determination of PhoAΔ2-22 was performed as

described previously (26). In brief, E. coli DHB4

harboring plasmid pAID135 were grown in MOPS

minimal medium until an OD600 of 0.2 was

reached. IPTG was added to a final concentration

of 5 mM to induce expression of leaderless

alkaline phosphatase. At an OD600 of 0.4, allicin

-

1

(

0.79 mM; 128 µg mL ) or diamide (1 mM) was

added to induce oxidative stress. After 15 min,

00 µL cells were added to 100 µL 1 M

9

iodoacetamide prior to incubation on ice for 20

min. Cells were harvested and washed twice in

-

1

mL pre-chilled washing buffer (50 mM NaCl,

0 mM NH Cl, 10 mM MgCl 10 mM

30 µL min ) prior to separation on an analytical

column (Acclaim PepMap RSLC, C18, 2 µm,

100 Å, 5 µm i.d. x 50 cm, nanoViper) using a

gradient from 5-40 % solvent B within 98 min

(solvent A: 0.1 % (v/v) formic acid; solvent B:

0.1 % (v/v) formic acid, 84 % (v/v) acetonitrile).

4

2

,

iodoacetamide, 40 mM MOPS–KOH, pH 7.3)

prior to resuspension in 100 µL lysis buffer

-

1

(

10 mM EDTA, 2 mg mL lysozyme, 10 mM

iodoacetamide, 20 mM Tris-HCl, pH 8.0) and

0 min incubation on ice. After lysing cells by

three freeze-thaw cycles, 900 µl resuspension

buffer (10 mM MgCl , 10 mM ZnCl , 1 M Tris-

-

1

3

The flow rate was set to 400 nL min with a

column oven temperature of 60 °C. Full scan data

were acquired in the data-dependent mode.

MS/MS analysis of the 20 most intense peaks was

performed by collision-induced dissociation using

the linear ion trap.

2

HCl, pH 8.0) were added. Cell debris was cleared

by centrifugation (11,000 g, 10 min, 4 °C) prior to

transfer of 900 µL of the supernatant to a quartz

cuvette. Suspensions were stirred in a JASCO V-

Data analysis of S-allylmercapto-modified

6

50 spectrophotometer equipped with the

temperature-controlled cell holder PSC-718 for

0 min at 37 °C. The baseline extinction at 405 nm

was recorded for 30 sec prior to addition of

00 µL para-nitro-phenyl-phosphate (pNPP)

peptides − Data analysis was performed with

MaxQuant Software Version 1.5.3.12 (27). For the

identification, the S-allylmercapto modification of

cysteines was added via the Andromeda

configuration tab to the MaxQuant modification

1

solution (0.4 % (v/v) end concentration). Product

formation was monitored for 2 h. Alkaline

phosphatase activity in µM min was calculated

3 4

list (C H S; addition of 72.00337 Da) (27). Raw

data files of both replicates were loaded into

MaxQuant in parallel. Oxidation (M) and S-

allylmercapto were set as variable modifications.

Label-free quantification was selected. After data

processing, the table evidence.txt was used for

further analysis. Peptides that were found

modified by S-allylmercapto (“S-allylmercapto”,

"Oxidation (M), S-allylmercapto”, "2 S-

allylmercapto”) were copied to a new excel sheet.

The peptide list was further edited in order to

display only those peptides that were identified in

both replicates with a posterior error probability

-

1

based on the initial velocity according to the

following equation (ε405 p-nitrophenol

=

-

1

-1

1

4

8,000 M cm ; ΔE, change of absorbance at

05 nm min ):

-

1

∆

"

A = _#ꢀ∗ 10⁺

$%&

The relative activity of the untreated control was

set to 1.

(

PEP) below 0.01. In case a peptide appeared in

ꢀ

5

various modification states, the peptide with the

lowest PEP was selected.

modification. After data processing, the table

peptides.txt was used for further data analysis.

Data comparison of label-free and

OxICAT experiments − The peptides from label-

free analysis were matched to peptides from the

OxICAT analysis. Peptides identified in both types

of experiments were selected for OxICAT-based

quantification. For this purpose, the degree of

oxidation of a particular peptide in each individual

experiment and replicate was calculated. ꢀ

Subsequently, the percentage change between an

untreated control and its allicin-treated counterpart

was calculated.

Peptides were sorted by the average percentage

change from all three replicates. Peptides with a

percentage change above 10 % in each of the

replicates were considered consistent and reliable.

OxICAT labeling, tryptic digest, peptide

enrichment − MOPS minimal medium was

inoculated with 500 µL of an E. coli MG1655

overnight culture and cultivated until OD600 of 0.4

was reached. The main culture was split and

stressed as described above. After 15 min of stress,

1

.7 mL (corresponding to 100 µg protein) of each

culture were harvested and centrifuged (13.000 g,

°C, 10 min). The supernatant was removed

4

carefully and pellets washed twice with 1 mL TE

buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.5).

Pellets were resuspended in 80 µL of DAB (6 M

Urea,

200 mM

Tris

(

tris(hydroxymethyl)aminomethane)

pH 8.5,

1

0

0 mM EDTA (Ethylenediamine-tetraacetic acid),

.5 % (w/v) sodium dodecyl sulfate (SDS), freshly

Mass spectrometry proteomics data

for

prepared and stored anaerobically prior to use).

ICAT labeling was performed as described

previously (28). Briefly, each sample was mixed

with the content of one vial of light ICAT.

Samples were incubated, protected from light, for

identification of S-allylmercapto modified

cysteines (label-free) and quantification (OxICAT)

have been deposited to the ProteomeXchange

Consortium (25) via the PRIDE partner repository

with the dataset identifier PXD003063.

2

h at 37 °C with 300 rpm on a shaker. Samples

For peer review the data can be accessed at

were precipitated with 5 volumes of chilled

acetone for 4 h or overnight. After washing and

drying, the pellets were resuspended in 75 µL

DAB supplemented with 5 µL 0.5 M TCEP and

incubated for 10 min. Subsequently, samples were

mixed with the content of one vial heavy ICAT

and incubated for 2 h at 37 °C with 300 rpm on a

shaker. For each two ICAT samples, one vial of

lyophilized trypsin from the ICAT kit was restored

in 200 µL Aqua dest. Samples were resuspended

in 80 µL denaturing buffer from the ICAT kit and

http://www.ebi.ac.uk/pride/archive/login

Username (“Email”): reviewer10712@ebi.ac.uk

Password: t84FSJpc

Isocitrate lyase activity assay − Isocitrate lyase

activity assays were performed as described earlier

(29). In brief, E. coli MG1655 was grown as

described for the DTNB assay. From each

condition (untreated, allicin-treated, diamide-

treated), 4 mL of culture were harvested by

centrifugation (13,000 g, 10 min, 4 °C). Pellets

were washed twice with imidazole buffer (50 mM

imidazole, pH 6.8) prior to resuspension with

200 µL imidazole buffer supplemented with

2

0 µL acetonitrile. 100 µL of the trypsin solution

was added to each sample. Digests were

performed for 16 h at 37 °C. Peptide purification

by cation exchange and enrichment by affinity

purification was carried out as described

previously (28). For each condition (untreated,

allicin-treated, and diamide-treated), three

biological replicates were analyzed. LC-MS/MS

analysis was performed as described above.

-

1

2 mg mL lysozyme. Suspensions were incubated

on ice for 30 min. Samples were disrupted by

freeze-thaw lysis (3 cycles of freezing in liquid

nitrogen and thawing at room temperature; 30 sec

each). Cell debris was removed by centrifugation

(13,000 g, 10 min, 4 °C). Four 2 mL quartz

cuvettes were filled with 750 µL of imidazole

buffer, 150 µL of 10 mM EDTA, 150 µL of

Data analysis of OxICAT samples − Raw

data files of allicin-treated and control samples

2

50 mM MgCl , 150 µL of 40 mM phenylhydrazine

(

three replicates each; parameter group 1 to 6)

HCl, and 150 µL 10 mM DL-isocitric acid. The

solutions were equilibrated to 30 °C using a

JASCO V-650 spectrophotometer equipped with

the temperature-controlled cell holder PSC-718.

Absorption was monitored at 324 nm until

were loaded to MaxQuant in parallel. ICAT-0 and

ICAT-9 labels were selected in the group-specific

parameters tab. Oxidation (M) was set as variable

ꢀ

6

constant. 150 µL of isocitrate lyase (blank) and

50 µL of supernatants (the three test conditions)

were added to the cuvettes. The increase in

absorption was measured for 1400 sec. Activity

was then calculated based on the extinction

Bis-Tris gels (Life Technologies, Darmstadt,

Germany) in MES running buffer. Separated

proteins were transferred to nitrocellulose

membranes using the iBlot system (Life

Technologies, Darmstadt, Germany) according to

the manufacturer's instructions. Detection of RpoH

was performed using polyclonal rabbit anti-RpoH

antibodies and IRDye 680 LT goat anti-rabbit IgG

(H + L; Li-cor Biosciences, Bad Homburg,

Germany). Band intensities were quantified using

ImageJ 1.49v (31).

1

coefficient

324 = 16,800 M cm and the velocity of the

reaction.

of

phenylydrazin

glyoxylate

-

1

-1

ε

Crude extract aggregation assay − E. coli

MC4100 was grown in 250 mL LB medium to an

OD600 of 0.6. Cells were harvested by

centrifugation at 4,000 g for 20 min at 4 °C. The

pellet was washed with 25 mL washing buffer

RESULTS

(

1

50 mM HEPES/KOH, pH 8.0, 175 mM NaCl,

mM MgCl ) and resuspended in 1.25 mL

Allicin inhibits bacterial growth − To

analyze the effect of allicin on the growth of

microorganisms, we determined the minimal

inhibitory concentration (MIC) according to CLSI

standards on a set of pathogenic bacteria and fungi

(Fig. 2A). The MIC of a substance represents the

lowest concentration fully inhibiting visible

growth. Allicin exhibited comparable MICs (32 to

2

washing buffer. 100 µL of protease inhibitor

cocktail cOmplete, EDTA-free (1 tablet per 2 mL

of Aqua dest.; Roche Diagnostics, Mannheim,

Germany) were added prior to cell disruption by

ultrasonication

(VialTweeter

ultrasonicator,

Hielscher GmbH, Germany; 3 x 1 min) at 4 °C.

Lysates were cleared by centrifugation at 16,100 g

for 30 min at 4 °C. 40 µL aliquots were treated

with allicin (1-100 mM) for 15 min at 30 °C.

Samples were centrifuged for 30 min at 16,100 g

at 4 °C. Supernatants were mixed with 10 µL

5

1

sample were separated by denaturing SDS-PAGE.

In vivo degradation of RpoH − Turnover of RpoH

was analyzed as described previously (26, 30). In

brief, E. coli MG1655 harboring plasmid pEC5217

for the expression of RpoH was grown in LB

medium at 30 °C until OD600 of 0.5 was reached.

The culture was split into three 20 mL cultures

prior to addition of 1 mM IPTG. Diamide was

added after 5 min to a final concentration of

-

1

64 µg mL )

towards

methicillin-resistant

Staphylococcus aureus ATCC 43300, the type

strain Staphylococcus aureus DSM 20231, E. coli

DSM 30083, Acinetobacter baumannii DSM

30007, and Candida albicans DSM 1386. Growth

of the type strain Pseudomonas aeruginosa DSM

50071 was not affected by allicin concentrations

x SDS sample buffer. Pellets were washed with

00 µL washing buffer prior to resuspension in

00 µL 1 x SDS sample buffer. 2.5 µL of each

-

1

lower than 512 µg mL . The Müller-Hinton

medium used in standard MIC assays is a rich

medium that contains thiols, which could

inactivate allicin. We therefore determined the

MIC of our lab strain E. coli MG1655 in a more

defined medium, which lacks the presence of

potentially inhibiting thiols (MOPS medium). A

-

1

concentration of 23 µg mL was sufficient to

completely inhibit growth of E. coli in this

medium (Fig. 2B).

1

mM, whereas allicin stress was started after

0 min by the addition of 0.79 mM allicin. After a

We then determined allicin-induced

inhibition of exponentially growing E. coli. Mid-

logarithmic cells showed a considerable growth

total of 25 min, protein synthesis was stopped by

the addition of 200 µg mL chloramphenicol.

-

1

-1

inhibition at 128 µg mL of allicin (0.79 mM),

1

mL samples were collected after 0, 1, 3, 5, 7, 9,

1, 15, and 25 min and immediately frozen in

which is approximately six times higher than the

concentration needed for growth inhibition of the

significantly smaller number of cells used in MIC

tests (Fig. 2C). The disulfide stress inducer

diamide served as positive control for growth

inhibition (26, 32, 33). Addition of both

compounds led to an immediate growth arrest.

After approximately 60 min of diamide treatment

and 90 min of allicin treatment, the cultures

liquid nitrogen. Samples were thawed on ice and

centrifuged for 5 min at 4 °C and 16,100 g. Pellets

were resuspended in 40 µL TE buffer (10 mM

Tris, 1 mM EDTA, pH 8.0) and 10 µL 5 x SDS

sample buffer. Samples were vortexed for 30 s and

incubated at 95 °C for 5 min. Separation was

performed by loading 10 µL samples on 4-12 %

ꢀ

7

resumed growth and reached the same final optical

densities as the control cultures. These results

show that, while allicin seems to be slightly more

inhibitory when compared to diamide, cells are

able to overcome the stress induced by both

compounds at the levels used. This experiment

provided a suitable basis for the in-depth analysis

of the cellular consequences of allicin exposure at

sublethal levels.

and at higher molar ratios (Fig. 4A). Oxidized

glutathione (GSSG) was not affected by allicin

(Fig. 4B). The assay we used is able to determine

the total glutathione concentration (i.e. [GSH] + 2

[GSSG]). Our results indicate that allicin is indeed

reacting with free GSH, but this does not result in

the formation of GSSG. Instead, the reaction

product cannot be reduced by glutathione

oxidoreductase. It has been previously reported

that the reaction of GSH with allicin leads to the

formation of S-allylmercapto glutathione (GSSA)

(16). The observed decrease of total glutathione in

the presence of allicin suggests that GSSA was not

effectively converted to GSSG by a reaction with

GSH and was not transformed to GSH and

allylmercaptan by the action of glutathione

oxidoreductase present in the assay.

In vivo analysis of glutathione levels in E. coli

cells during the exponential growth phase revealed

that GSH levels drop significantly under allicin

stress, similar to our in vitro observation (Fig. 4C).

This drop in GSH is not accompanied by a rise in

GSSG levels. Quite the contrary, GSSG levels in

allicin-stressed cells decreased when compared to

control cells (Fig. 4D). Thus, although the overall

glutathione level drops, the ratio of the

GSH/GSSG couple in allicin-stressed cultures is

significantly increased about 5 fold (Fig. 4E). It

seems, that once reacted with allicin, glutathione

does not reenter, or only slowly reenters the GSH

pool.

Allicin reduces cellular thiol levels − Four

main groups of thiol-containing molecules

contribute to the total cellular thiol content:

glutathione, free cysteine and other low molecular

weight thiols, and cysteine-containing proteins. As

allicin is considered a thiol-reactive compound, we

directly

measured

the

total

sulfhydryl

concentrations in crude extracts of cells treated

with allicin. The quantification of free thiol levels

was carried out using 5,5’-dithio-bis-(2-

nitrobenzoic acid) (DTNB; Ellman’s reagent).

DTNB reacts with free sulfhydryls and forms a

mixed disulfide as well as thio-2-nitrobenzoic acid

(

TNB). The resulting TNB concentration can be

determined spectrophotometrically at 412 nm and

is directly proportional to the thiol concentration

in the sample. The sulfhydryl concentrations of

cells treated with 0.79 mM of allicin or 1 mM of

diamide decreased by approximately 40 %, when

compared to the untreated control (Fig. 3). This

corresponded to a drop of free thiols by 12 mM

within the cell, thus allicin reacts with free thiols

with high efficiency.

Allicin induces heat shock and oxidative

Allicin reacts with glutathione in vitro and in vivo

stress response − To explore the cellular

consequences of the drop in free cellular thiols and

glutathione, we performed proteomic profiling

using 2D gel electrophoresis (Fig. 5). Cells were

treated with allicin, while newly synthesized

−

Glutathione is the most important low-molecular

weight antioxidant in the cell. With its thiol group,

reduced glutathione (GSH) can react with a variety

of reactive species. This often leads to the

formation of oxidized glutathione (GSSG).

Typically, the GSH to GSSG ratio in the cell is

kept high by the enzyme glutathione

oxidoreductase and the concentration of GSH

exceeds the concentration of GSSG by a factor of

3

5

proteins were labeled with S [L-] methionine.

3

5

The S methionine incorporation rates indicated a

moderate decrease of protein synthesis after allicin

exposure to 44 % of the synthesis rate of untreated

cells. Newly synthesized proteins could then be

quantified by densitometry of radioactive spots

after 2D-PAGE (Fig. 5). Allicin treatment led to a

drastic change in the protein synthesis profile (Fig.

5). We then identified the 25 protein spots with the

highest synthesis rates in allicin-treated cells. 21 of

these spots could be identified by mass

spectrometry from non-radioactive gels (Fig. 5,

Table 1, Supplemental Table 1). From 16 spots, a

single protein was identified, while the remaining

>

100 (34). We asked, if the observed decrease in

total thiol levels during allicin stress is due to

reaction of allicin with reduced glutathione (GSH)

(

Fig. 4). First, we tested if allicin reacts with

reduced GSH in vitro. decrease to

A

approximately 40 % of the initial total glutathione

level was observed at molar ratios of allicin to

GSH of 0.5 after 15 min of incubation. GSH levels

dropped more with prolonged incubation (60 min)

ꢀ

8

five spots represented two or more proteins.

Remarkably, many of the identified proteins, such

does not seem to be as common in allicin-treated

cells as it is in cells that experience disulfide stress

caused by diamide or knock out of thiol-disulfide

oxidoreductase systems.

Allicin exposure leads to S-allylmercapto

modification of proteins in vivo − Our in vitro and

in vivo data on the reaction of GSH with allicin

3

2

as IbpA and AhpC belong to the σ or the OxyR

3

2

regulon. The heat shock sigma factor σ together

with RNA polymerase regulates the expression of

genes crucial for the adaptation to heat stress,

while OxyR is an important transcriptional

regulator for the oxidative stress response in

E. coli (35, 36). We previously reported that

disulfide stress induces the heat stress response in

E. coli (26). In this earlier study, disulfide stress

was induced chemically by diamide, and

genetically by a deletion of trxB and gor, which

are the ultimate NADPH-dependent reductases of

the thioredoxin and glutathione/glutaredoxin

systems, respectively. The proteomic results

presented here are strikingly similar to our

findings in our previous study, we thus

hypothesized that allicin may induce disulfide

stress similarly to diamide.

suggested

that

S-allylmercapto-modified

glutathione is a major product of this reaction. It

thus seemed likely to us that proteins are also S-

allylmercapto-modified upon treatment of cells

with allicin. To analyze if the proposed S-

allylmercapto modification is indeed a stable end

product of protein modification by allicin in vivo,

lysates of allicin-treated E. coli cells were

prepared under non-reducing conditions. Lysates

of untreated cells served as negative control, and

chemically induced disulfide stress in diamide-

treated cells served as a further control. After a

tryptic digest, samples were directly analyzed by

mass spectrometry in a label-free approach. We

then searched for the S-allylmercapto modification

at cysteine residues. A total of 90 allylmercapto-

modified peptides were identified in our

experiments with high confidence. These 90

peptides belong to a total of 73 individual proteins

(Supplemental Table 2).

Allicin did not induce protein disulfide

bond formation in vivo − We thus tested if allicin

causes non-native protein disulfide bond formation

in the cytoplasm. For our experiment, we

measured the activity of leaderless alkaline

phosphatase

PhoAΔ2-22

(37).

Alkaline

phosphatase activity depends on the formation of

two disulfide bonds that contribute to the native

protein fold. Wild-type alkaline phosphatase is

secreted into the periplasm, where oxidative

conditions lead to disulfide bond formation (38).

Leaderless alkaline phosphatase, however, lacks

the signal sequence for the export into the

periplasm, and thus is inactive in the reducing

environment of the cytoplasm under non-stress

conditions. Oxidative stressors, such as diamide

and genetically-induced disulfide stress, however,

can activate PhoAΔ2-22 despite its localization in

the cytoplasm (26).

In contrast to allicin, no S-allylmercapto-modified

peptides were identified in the control and the

diamide-treated sample.

Individual cysteines were up to 90 % S-

allylmercapto-modified − To quantify the extent

of S-allylmercapto modification, we then

performed differential OxICAT labeling of

cysteines from the E. coli proteome after allicin

exposure (28). This differential labeling method is

based on the isotopically different but chemically

1

2

13

identical labels ICAT- C (light) and ICAT- C

(heavy). The light version of the probe was used to

label and thereby block all initially reduced

cysteines. After treatment with a strong reductant,

newly accessible cysteine thiols were labeled with

the heavy probe. After digestion, peptide

purification, and mass spectrometry, the ratio of

light and heavy fractions of a peptide could be

calculated. This ratio directly reflects the in vivo

ratio of thiol-modification of protein cysteines.

Of the cysteines, we initially identified as

However, lysates of cells treated with

allicin did not show an activity level of

PhoAΔ2-22 above the levels of untreated cells

(

Fig. 6). In contrast, 1 mM diamide, a potent

inducer of protein disulfide bond formation,

significantly activated PhoAΔ2-22. This result

shows that although allicin and diamide both

decrease total sulfhydryl levels of the cell as well

as levels of reduced and oxidized glutathione, and

induce the heat stress response, the underlying

mechanisms differ significantly. Formation of

cystine disulfide bonds in cytoplasmic proteins

S-allylmercapto-modified, 12 showed a rate of

modification consistently higher than 10 %,

reaching as high as 90 % (Table 2, Supplemental

ꢀ

9

Table 3). The respective proteins are high-

abundant proteins, mostly involved in the primary

metabolism of E. coli, suggesting that allicin

targets cytosolic proteins indiscriminately.

Interestingly, our MS data suggest that in some of

these proteins, which possess cysteines that are

important for their activity, allicin predominantly

modified cysteines that have not been implied in

catalysis or redox regulation. E.g. 25 % of cysteine

when compared to allicin treatment. This result

indicates that enzyme activity, at least of AceA, is

inhibited by allosteric modification in allicin-

treated cells.

Allicin causes protein aggregation − Our data

indicates that allicin induces S-allylmercapto

modification of proteins and that this modification

can inhibit protein activity. This observation could

be explained by unfolding stress caused by S-

allylmercapto modification. Additionally, protein

unfolding and aggregation would be a potent

inducer of the heat shock response, which could

explain the strong up-regulation of heat shock

proteins on our 2D gels. To test this hypothesis,

we performed aggregation assays with increasing

allicin concentrations using crude extracts of

E. coli MC4100. This strain was previously used

to analyze protein aggregation upon diamide stress

(26). Allicin caused protein aggregation in a

concentration-dependent manner, which was

already visible at 1 mM of allicin (Fig. 9). At

concentrations higher than 50 mM, the majority of

proteins aggregated and could be found in the

pellet fraction. Diamide applied at the same

concentrations did not cause aggregation in wild-

type E. coli crude extracts in our previous study

(26). This finding suggests that allicin-induced

modifications affect proper folding and structural

integrity of cellular proteins more than disulfide

stress caused by non-native cystine bond

formation.

2

89 of glyceraldehyde 3-phosphate dehydrogenase

(

GapA), an important glycolytic enzyme, were

modified. GapA is known to be subject to redox

regulation of the enzyme upon oxidative stress

(

28, 39). Typically, this regulation occurs directly

at the active site cysteine of this enzyme.

Strikingly, allicin, unlike other oxidants, such as

HOCl, did not seem to modify the catalytic

cysteine, but instead targets an unrelated cysteine.

Similarly, GrxC, a disulfide reductase of E. coli, is

modified exclusively at a single cysteine that is not

part of the catalytic CXXC motif (40, 41). In

isocitrate lyase (AceA) we only found Cys244 to

be S-allylmercapto-modified, while peptides

containing Cys195, an important residue for

catalytic activity (42), were not detected in all

three replicates and changed to a much lesser

extent.

Allicin reduces isocitrate lyase activity in

cell lysates of E. coli − To further explore, if these

modifications lead to an allosteric inactivation of

cellular enzymes, we looked directly at the activity

of isocitrate lyase AceA. Isocitrate lyase plays a

crucial role in the glyoxylate cycle which bypasses

the decarboxylation steps of the TCA cycle (43).

We tested AceA activity in lysates of E. coli cells

after allicin and diamide treatment. This candidate

was chosen for two reasons. First, in our OxICAT

analysis, we found that Cys244 of this enzyme, a

residue not involved in catalysis, was S-

allylmercapto-modified (Fig. 7A, Table 2). This

cysteine showed an average change in the

oxidation state of 51 %, which makes it the second

most oxidized candidate identified in our analysis

Allicin exposure leads to RpoH stabilization − In

E coli, the induction of the heat stress response

3

2

depends on the heat shock sigma factor σ (RpoH)

(36). Under non-stress conditions, RpoH is bound

by DnaK/J and rapidly targeted to degradation by

FtsH. Under heat stress, however, DnaK/J bind to

denaturing proteins to protect them from

aggregation or to target them to proteolysis. RpoH

thus is stabilized and induces the expression of

chaperones and other heat shock proteins in

concert with RNA-polymerase (36). Induction of

the heat stress response upon allicin stress may

thus be due to increased stability of RpoH, caused

by protein unfolding and aggregation. We,

therefore, tested RpoH stability in cells exposed to

allicin. RpoH levels remained stable during 25 min

of analysis (Fig. 10). In contrast, the RpoH half-

life was very short under control conditions. Thus,

like diamide (26), allicin exposure leads to a

(

Fig. 7B; Table 2). Second, activity assays for

isocitrate lyase are well-established and can be

performed with cell extracts. AceA activity was

significantly diminished after treatment with

allicin (Fig. 8). Diamide treatment, which was

performed as control, also had an inhibitory effect

when compared to control lysates. However,

activity was significantly higher and twice as high

ꢀ

10

strong stabilization of RpoH, which results in the

induction of the heat shock response.

target are thiols, which occur in high abundance

(29 mM) in the E. coli cell (Fig. 2A), the

substantially higher number of cells present in a

RpoH is crucial to overcome allicin stress − At

sublethal concentrations (0.79 mM), E. coli can

overcome allicin stress and resume growth after

approximately 90 min (Fig. 2C). One explanation

could be the increased expression of chaperones

and proteins belonging to the oxidative stress

response, which would ultimately remove allicin-

damaged proteins from the cell. We thus asked if

chaperones and other heat shock proteins

synthesized during allicin exposure are essential

for this recovery. We tested recovery of an E. coli

rpoH mutant from allicin stress. This mutant is not

able to induce the heat stress response. Growth of

this mutant strain is very slow and cells can only

grow at moderate temperatures (25 °C). While this

strain could recover from diamide stress after a

long phase of growth arrest, it did not resume

growth after allicin exposure (Fig. 11). It thus

seems that the induction of the heat stress response

accompanied by the expression of chaperones and

other heat stress proteins is essential upon allicin

stress.

8

-1

mid-logarithmic culture (~10 cells mL ) vs. the

small inoculum used for MIC tests

4

-1

(2 x 10 cells mL ) could account for this

difference.

After treatment of E. coli with a sublethal

amount of allicin, total cellular sulfhydryl levels

dropped significantly, which seems to be a

cumulative effect of the conversion of reduced

glutathione (GSH) to its S-allylmercapto adduct

(Fig. 4) and modification of protein cysteine thiols

(Fig. 7). Interestingly, a one-time sub-millimolar

dose of allicin causes the cellular thiol content to

drop by 11 mM, indicating that allicin permeates

the cell much faster than the cellular repair system

can remove the thiol-modifications caused by it.

GSH is one of the most important cellular

antioxidants and together with its oxidized

counterpart GSSG constitutes the main thiol redox

buffer. We could show that the loss in free thiols is

caused, at least in part, by modification of this

tripeptide to S-allylmercapto glutathione, a known

reaction-product of GSH with allicin in vitro(16).

Glutathione oxidoreductase seems then unable to

return S-allylmercapto-modified glutathione to the

cellular GSH pool.

DISCUSSION

Allicin from garlic is

a

potent

antimicrobial substance. In the present study, we

aimed at understanding the molecular mechanism

of its antibacterial effects.

Loss of reduced GSH in vivo could result in

oxidative stress. Such a drop in free GSH caused

by allicin was also shown for Saccharomyces

cerevisiae with the help of genetically encoded,

fluorescent roGFP redox probes (18). At the

concentrations used in S. cerevisiae this drop

coincided with induction of apoptosis in a

proportion of the cells. In contrast, in our

experiment, when E. coli cells were treated with

128 µg mL , they resumed growth after

approximately 90 min of growth arrest. The

cultures reached the same final optical densities

when compared to the control, indicating that

effective mechanisms to overcome allicin stress

exist in E. coli. Indeed, lowering of the GSSG

levels compensates for the loss in total GSH in

allicin-treated E. coli. This results in a slightly

lower GSH/GSSG redox potential (-270 mV) in

these cells when compared to wild type (-257 mV)

and could be part of the cell’s defense strategy.

2D-PAGE analysis of the protein

We first tested allicin for its ability to

inhibit growth of a set of pathogenic organisms.

Growth of the microorganisms was inhibited by

allicin concentrations ranging between 32 and

-

1

6

4 µg mL . A remarkable difference in the

effective concentration of allicin was found for the

Pseudomonas aeruginosa DSM50071 strain

tested. The unusually high concentration of

-

1

-

1

5

12 µg mL necessary to prevent growth has been

reported before for this organism (44) and

suggests that concentrations of allicin needed to be

effective against some strains of this organism are

therapeutically not achievable. In minimal medium

-

1

allicin exhibited a MIC of 23 µg mL against

E. coli MG1655. These results are in good

agreement with a previously reported LD50 of

-

1

5 µg mL of allicin against E. coli (45). To

achieve inhibition of growth in exponentially

growing E. coli, we needed to add substantially

higher concentrations. Given that allicin’s main

expression pattern during treatment with allicin

demonstrated that proteins belonging to the heat

ꢀ

11

shock and oxidative stress responses are up-

regulated significantly. A similar observation was

made previously in our lab when cells were

subjected to chemically or genetically induced

disulfide stress (26). Under these conditions,

OxyR is activated by thiol oxidation and non-

native disulfide bonds lead to protein unfolding.

The unfolded proteins are then bound by DnaK,

which is no longer available to target the heat

suggesting that in isocitrate lyase, too, an allosteric

cysteine is the major target of allicin.

AceA was tested for its activity in vivo.

Allicin exposure of cells significantly reduced

their isocitrate lyase activity. A number of

enzymes from both pro- and eukaryotic organisms

has been reported to be inhibited by allicin in vitro

(14, 46). Rabinkov et al reported that papain and

alcohol dehydrogenase were efficiently inactivated

by allicin as well as by GSSA in vitro (16). In

papain, Cys25 is part of the active site, whereas in

alcohol dehydrogenase, Cys203 is located close to

3

2

shock sigma factor σ to proteases for its

3

2

immediate degradation. σ is thus stabilized and

induces the heat shock response. While we have

shown that allicin, in contrast to diamide, does not

primarily induce protein disulfide bond formation

in the cytoplasm, allicin exposure resulted in

enhanced protein aggregation accompanied by

+

the NADP binding site. Activity of both enzymes

could be restored by DTT or 2-mercaptoethanol

and a released allyl mercaptan moiety was

detected. In the present study, we could

exemplarily demonstrate that enzyme inhibition in

vivo coincides with the S-allylmercapto

modification of cysteines in cells exposed to

allicin. Although only about half of Cys244 of

AceA was found to be modified in vivo, activity

dropped by significantly more than 50 %. This

points to an allosteric effect, where the

modification of one subunit might cause the

inactivation of the isocitrate lyase tetramer or an

additive effect of allosteric C244 and active site

C195 modification.

Taken together, it seems likely that accessible

cysteines, which are part of the active sites of

enzymes, can be reduced by the major reducing

systems in vivo and thus appear mostly reduced in

our approach. In contrast, cysteines that are buried

in the core of a protein might still be accessible to

the small hydrophobic molecule allicin, but

potentially cannot be reached by the repair

machinery. Indeed, most cysteines affected by S-

allylmercapto-modification in our study had an

accessible surface area well below 25%, which is a

threshold above which an amino acid could be

considered surface accessible (47) (Table 2).

Overall, the mechanism of action of allicin

3

2

increased stability of σ . This finding thus could

explain the observed induction of heat shock

proteins.

Many up-regulated proteins showed a shift

in pI on our 2D gels, indicating a direct

modification by allicin. Mass spectrometric

analysis of lysates in a gel-free approach showed

that, 15 min after allicin treatment, a substantial

set of proteins is S-allylmercapto-modified. To our

knowledge, a global study that analyzes the total

proteome of an organism for allicin-induced

modifications has not been reported yet. Currently,

a study of the allicin-yeast redoxome is in progress

in our laboratories and it will be interesting to

compare the response in yeast to that in E coli.

Recently, a Burkholderia cepacia peroxiredoxin,

purified after expression in E. coli was shown to

be S-allylmercapto-modified after treatment with

allicin in vitro, as determined by mass

spectrometry (15). The S-allylmercapto-modified

proteins identified in our study are mostly high-

abundant proteins of primary metabolism,

indicating that allicin modifies proteins

unspecifically. Interestingly, none of the peptides

detected for GapA, GrxC, or AceA, contained

catalytically relevant or already described redox-

sensitive cysteines. OxICAT analysis of these

described redox-active cysteines revealed no

significant change in their oxidation state in GapA

and GrxC. We were able to quantify the redox

state of cysteine 195 in AceA (isocitrate lyase), a

cysteine residue adjacent to the substrate binding

site (42) in two of our three biological replicates.

The oxidation state of Cys195 changed by only 18

percentage points as opposed to 51 for Cys244,

is thus likely a combination of the decrease of

reduced glutathione levels through the formation

of GSSA and inhibition of enzymes crucial for the

metabolism of the cell and the oxidative stress

defense through S-allylmercapto modification.

This modification could be formed directly by

reaction of cysteine residues with allicin or

through the reaction of GSSA with protein

cysteines (16).

ꢀ

12

At sublethal allicin concentrations, cells can

overcome stress triggered by allicin. The inability

of an rpoH-deficient E. coli strain to resume

growth after allicin exposure demonstrates that the

induction of the heat stress response is essential

for this recovery.

When compared to most clinically applied

antibiotics, allicin seems to have an advantage. It

does not target a specific protein, so that resistance

development by target mutation is therefore rather

unlikely. However, its high instability in the

gastric tract and therefore inability to reach the

desired target and its high toxicity when

subcutaneously injected, prevent its systemic

clinical use as antibiotic (48-50). Furthermore, the

high GSH concentration in mammalian tissues and

blood are likely to make it difficult to achieve

therapeutically relevant concentrations of allicin in

the body. Additionally, the targets we identified in

E. coli are not specific to bacteria, but essentially

occur in all living beings and thus will be attacked

in human cells, as well (51). Nevertheless, since

allicin is volatile, direct inhalation of allicin

vapour or aerosols may be able to target

pulmonary infections. This promising line of

investigation is being pursued in the authors’

laboratories.

Along the same lines, first steps have been

made for the site-specific application of allicin. In

2010, a monoclonal antibody against Aspergillus

fumigatus, was fused to alliinase (52). In vivo

application of this antibody together with the

enzyme's substrate alliin successfully cleared

intratracheal fungi in mice. These efforts together

with a deeper understanding of the mode of action

of allicin make this molecule a promising

candidate structure for the treatment of numerous

human diseases.

Acknowledgements: We thank Christina John for her help with mass spectrometry, and Lisa-Marie

Bittner and Franz Narberhaus for providing plasmid pEC5217 and anti-RpoH antibodies.

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

article.

Author contributions: A.M. performed glutathione determinations, crude extract aggregation assays,

RpoH stability assays, growth tests with the RpoH mutant, the analysis of the experimental data and

wrote the manuscript. J.E. performed growth tests, MIC determination of E. coli, assays for sulfhydryl

contents, alkaline phosphatase activity, and isocitrate lyase activity, and preparation of mass spectrometry

samples, and drafted parts of the manuscript. 2D-PAGE, protein identification, and MIC determination of

microorganisms was carried out by P.P. Analysis of 2D gels and analysis of data for protein identification

was done by J.E.B. Allicin was synthesized by F.A. F.A. and A.J.S. contributed to editing and discussion

of the manuscript. Mass spectrometry for S-allylmercapto-modified protein identification and OxICAT

experiments was performed by K.K. L.I.L. conceived the experiments and wrote part of the manuscript.

REFERENCES

1

2

.

Rivlin, R. S. (2001) Historical perspective on the use of garlic. J. Nutr. 131, 951S–4S

Cavallito, C. J., and Bailey, J. H. (1944) Allicin, the antibacterial principle of Allium sativum. I.

Isolation, physical properties and antibacterial action. J. Am. Chem. Soc. 66, 1950–1951

Cavallito, C. J., Buck, J. S., and Suter, C. M. (1944) Allicin, the antibacterial principle of Allium

sativum. II. Determination of the chemical structure. J. Am. Chem. Soc. 66, 1952–1954

Lancaster, J. E., and Collin, H. A. (1981) Presence of alliinase in isolated vacuoles and of alkyl

cysteine sulfoxides in the cytoplasm of bulbs of onion (Allium cepa). Plant Science Letters. 22,

3

4

.

1

69–176

5

6

.

Ellmore, G. S., and Feldberg, R. S. (1994) Allin lyase localization in bundle sheaths of the garlic

clove (Allium sativum). American Journal of Botany. 81, 89

Stoll, A., and Seebeck, E. (2006) Chemical Investigations on Alliin, the Specific Principle of

Garlic. in Advances in Enzymology and Related Areas of Molecular Biology (Nord, F. F. ed), pp.

ꢀ

13

3

77–400, Nord/Advances, John Wiley & Sons, Inc., Hoboken, NJ, USA

7

8

.

Cutler, R. R., and Wilson, P. (2004) Antibacterial activity of a new, stable, aqueous extract of

allicin against methicillin-resistant Staphylococcus aureus. Br. J. Biomed. Sci. 61, 71–74

Curtis, H., Noll, U., and Stormann, J. (2004) Broad-spectrum activity of the volatile extracts of

garlic (Allium sativum), against plant pathogenic bacteria, fungi and oomycetes, Physiological

and Molecular Plant Pathology

9

.

Feldberg, R. S., Chang, S. C., Kotik, A. N., Nadler, M., Neuwirth, Z., Sundstrom, D. C., and

Thompson, N. H. (1988) In vitro mechanism of inhibition of bacterial cell growth by allicin.

Antimicrob. Agents Chemother. 32, 1763–1768

1

0.

1.

2.

3.

Ghannoum, M. A. (1988) Studies on the anticandidal mode of action of Allium sativum (garlic). J.

Gen. Microbiol. 134, 2917–2924

Yamada, Y., and Azuma, K. (1977) Evaluation of the in vitro antifungal activity of allicin.

Antimicrob. Agents Chemother. 11, 743–749

Mirelman, D., Monheit, D., and Varon, S. (1987) Inhibition of growth of Entamoeba histolytica by

allicin, the active principle of garlic extract (Allium sativum). J. Infect. Dis. 156, 243–244

Miron, T., Rabinkov, A., Mirelman, D., Wilchek, M., and Weiner, L. (2000) The mode of action

of allicin: its ready permeability through phospholipid membranes may contribute to its biological

activity. Biochim. Biophys. Acta. 1463, 20–30

1

4.

5.

6.

Rabinkov, A., Miron, T., Konstantinovski, L., Wilchek, M., Mirelman, D., and Weiner, L. (1998)

The mode of action of allicin: trapping of radicals and interaction with thiol containing proteins.

Biochim. Biophys. Acta. 1379, 233–244

Wallock-Richards, D., Doherty, C. J., Doherty, L., Clarke, D. J., Place, M., Govan, J. R. W., and

Campopiano, D. J. (2014) Garlic revisited: antimicrobial activity of allicin-containing garlic

extracts against Burkholderia cepacia complex. PLoS ONE. 9, e112726

Rabinkov, A., Miron, T., Mirelman, D., Wilchek, M., Glozman, S., Yavin, E., and Weiner, L.

(

2000) S-Allylmercaptoglutathione: the reaction product of allicin with glutathione possesses SH-

modifying and antioxidant properties. Biochim. Biophys. Acta. 1499, 144–153

Borlinghaus, J., Albrecht, F., Gruhlke, M. C. H., Nwachukwu, I. D., and Slusarenko, A. J. (2014)

Allicin: chemistry and biological properties. Molecules. 19, 12591–12618

Gruhlke, M. C. H., Portz, D., Stitz, M., Anwar, A., Schneider, T., Jacob, C., Schlaich, N. L., and

Slusarenko, A. J. (2010) Allicin disrupts the cell's electrochemical potential and induces apoptosis

in yeast. Free Radic. Biol. Med. 49, 1916–1924

1

7.

8.

1

2

9.

0.

Neidhardt, F. C., Bloch, P. L., and Smith, D. F. (1974) Culture medium for enterobacteria. J.

Bacteriol. 119, 736–747

Albada, H. B., Chiriac, A.-I., Wenzel, M., Penkova, M., Bandow, J. E., Sahl, H.-G., and Metzler-

Nolte, N. (2012) Modulating the activity of short arginine-tryptophan containing antibacterial

peptides with N-terminal metallocenoyl groups. Beilstein Journal of Organic Chemistry. 8, 1753–

1

764

2

1.

2.

Ingraham, J. L., and Neidhardt, F. C. (1996) Escherichia coli and Salmonella: cellular and

molecular biology

Rahman, I., Kode, A., and Biswas, S. K. (2006) Assay for quantitative determination of

glutathione and glutathione disulfide levels using enzymatic recycling method. Nat Protoc. 1,

3

159–3165

2

3.

Wenzel, M., Patra, M., Senges, C. H. R., Ott, I., Stepanek, J. J., Pinto, A., Prochnow, P., Vuong,

C., Langklotz, S., Metzler-Nolte, N., and Bandow, J. E. (2013) Analysis of the mechanism of

action of potent antibacterial hetero-tri-organometallic compounds: a structurally new class of

antibiotics. ACS Chem. Biol. 8, 1442–1450

2

ꢀ

4.

5.

Wenzel, M., Patra, M., Albrecht, D., Chen, D. Y.-K., Nicolaou, K. C., Metzler-Nolte, N., and

Bandow, J. E. (2011) Proteomic signature of fatty acid biosynthesis inhibition available for in vivo

mechanism-of-action studies. Antimicrob. Agents Chemother. 55, 2590–2596

Vizcaíno, J. A., Deutsch, E. W., Wang, R., Csordas, A., Reisinger, F., Ríos, D., Dianes, J. A., Sun,

14

Z., Farrah, T., Bandeira, N., Binz, P.-A., Xenarios, I., Eisenacher, M., Mayer, G., Gatto, L.,

Campos, A., Chalkley, R. J., Kraus, H.-J., Albar, J. P., Martinez-Bartolomé, S., Apweiler, R.,

Omenn, G. S., Martens, L., Jones, A. R., and Hermjakob, H. (2014) ProteomeXchange provides

globally coordinated proteomics data submission and dissemination. Nat. Biotechnol. 32, 223–226

Müller, A., Hoffmann, J. H., Meyer, H. E., Narberhaus, F., Jakob, U., and Leichert, L. I. (2013)

Nonnative disulfide bond formation activates the σ -dependent heat shock response in

Escherichia coli. J. Bacteriol. 195, 2807–2816

Cox, J., and Mann, M. (2008) MaxQuant enables high peptide identification rates, individualized

p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. Biotechnol. 26, 1367–

2

6.

7.

8.

3

2

1

372

Leichert, L. I., Gehrke, F., Gudiseva, H. V., Blackwell, T., Ilbert, M., Walker, A. K., Strahler, J.

R., Andrews, P. C., and Jakob, U. (2008) Quantifying changes in the thiol redox proteome upon

oxidative stress in vivo. Proc. Natl. Acad. Sci. U.S.A. 105, 8197–8202

2

3

9.

0.

1.

2.

3.

4.

Chell, R. M., Sundaram, T. K., and Wilkinson, A. E. (1978) Isolation and characterization of

isocitrate lyase from a thermophilic Bacillus sp. Biochem. J. 173, 165–177

Obrist, M., and Narberhaus, F. (2005) Identification of a turnover element in region 2.1 of

3

2

Escherichia coli σ by a bacterial one-hybrid approach. J. Bacteriol. 187, 3807–3813

Schneider, C. A., Rasband, W. S., and Eliceiri, K. W. (2012) NIH Image to ImageJ: 25 years of

image analysis. Nat Meth. 9, 671–675

Kosower, N. S., and Kosower, E. M. (1995) Diamide: an oxidant probe for thiols. Meth. Enzymol.

2

51, 123–133

Leichert, L. I. O., Scharf, C., and Hecker, M. (2003) Global characterization of disulfide stress in

Bacillus subtilis. J. Bacteriol. 185, 1967–1975

Aslund, F., Zheng, M., Beckwith, J., and Storz, G. (1999) Regulation of the OxyR transcription

factor by hydrogen peroxide and the cellular thiol-disulfide status. Proc. Natl. Acad. Sci. U.S.A.

9

6, 6161–6165

3

5.

6.

7.

8.

9.

Storz, G., Tartaglia, L. A., and Ames, B. N. (1990) Transcriptional regulator of oxidative stress-

inducible genes: direct activation by oxidation. Science. 248, 189–194

Arsène, F., Tomoyasu, T., and Bukau, B. (2000) The heat shock response of Escherichia coli. Int.

J. Food Microbiol. 55, 3–9

Derman, A. I., Prinz, W. A., Belin, D., and Beckwith, J. (1993) Mutations that allow disulfide

bond formation in the cytoplasm of Escherichia coli. Science. 262, 1744–1747

Sone, M., Kishigami, S., Yoshihisa, T., and Ito, K. (1997) Roles of disulfide bonds in bacterial

alkaline phosphatase. J Biol Chem. 272, 6174–6178

Cotgreave, I. A., Gerdes, R., Schuppe-Koistinen, I., and Lind, C. (2002) S-glutathionylation of

glyceraldehyde-3-phosphate dehydrogenase: role of thiol oxidation and catalysis by glutaredoxin.

Meth. Enzymol. 348, 175–182

4

0.

Aslund, F., Ehn, B., Miranda-Vizuete, A., Pueyo, C., and Holmgren, A. (1994) Two additional

glutaredoxins exist in Escherichia coli: glutaredoxin 3 is a hydrogen donor for ribonucleotide

reductase in a thioredoxin/glutaredoxin 1 double mutant. Proc. Natl. Acad. Sci. U.S.A. 91, 9813–

9

817

4

1.

2.

Holmgren, A. (1989) Thioredoxin and glutaredoxin systems. J Biol Chem. 264, 13963–13966

Robertson, A., and Nimmo, H. G. (1995) Site-directed mutagenesis of cysteine-195 in isocitrate

lyase from Escherichia coli Ml308. Biochem. J. 305, 239–244

4

ꢀ

3.

4.

5.

Müller, A., Hoffmann, J. H., Meyer, H. E., Narberhaus, F., Jakob, U., and Leichert, L. I. (2013)

Non-native disulfide bond formation activates the σ32-dependent heat shock response in

Escherichia coli. J. Bacteriol. 10.1128/JB.00127-13

Cai, Y., Wang, R., Pei, F., and Liang, B.-B. (2007) Antibacterial activity of allicin alone and in

combination with β-lactams against Staphylococcus spp. and Pseudomonas aeruginosa. The

Journal of Antibiotics. 60, 335–338

Ankri, S., and Mirelman, D. (1999) Antimicrobial properties of allicin from garlic. Microbes and

15

Infection. 1, 125–129

4

6.

7.

8.

9.

Wills, E. D. (1956) Enzyme inhibition by allicin, the active principle of garlic. Biochem. J. 63,

5

14–520

Zhang, H., Zhang, T., Chen, K., Shen, S., Ruan, J., and Kurgan, L. (2009) On the relation between

residue flexibility and local solvent accessibility in proteins. Proteins. 76, 617–636

Rao, R. R., and Natarajan, S. (1949) Toxicity of pterygospermin and allicin. Proceedings of the

Indian Academy of Sciences - Section B. 29, 148–154

Freeman, F., and Kodera, Y. (1995) Garlic Chemistry - Stability of S-(2-Propenyl) 2-Propene-1-

Sulfinothioate (Allicin) in Blood, Solvents, and Simulated Physiological Fluids. J. Agric. Food

Chem. 43, 2332–2338

5

0.

1.

2.

Lawson, L. D., and Gardner, C. D. (2005) Composition, stability, and bioavailability of garlic

products used in a clinical trial. J. Agric. Food Chem. 53, 6254–6261

Banerjee, S. K., Mukherjee, P. K., and Maulik, S. K. (2003) Garlic as an antioxidant: the good, the

bad and the ugly. Phytother Res. 17, 97–106

Appel, E., Vallon-Eberhard, A., Rabinkov, A., Brenner, O., Shin, I., Sasson, K., Shadkchan, Y.,

Osherov, N., Jung, S., and Mirelman, D. (2010) Therapy of murine pulmonary aspergillosis with

antibody-alliinase conjugates and alliin. Antimicrob. Agents Chemother. 54, 898–906

Cavallo, L., Kleinjung, J., and Fraternali, F. (2003) POPS: A fast algorithm for solvent accessible

surface areas at atomic and residue level. Nucleic Acids Res. 31, 3364–3366

5

3.

ꢀ

16

FOOTNOTES

This work was supported by DFG Grant LE 2905/1-1 under the Priority Program SPP1710 of the German

Research Foundation and received additional funding from the European Research Council under the

European Union's Seventh Framework Program (FP7/2007-2013) / ERC Grant agreement n° 281384–

FuMe to L.I.L. F.A. and A.J.S. gratefully acknowledge financial support from RFwN and RWTH Aachen

University. J.E.B. acknowledges funding from the German Federal state of North Rhine-Westphalia

"

Translation innovativer Antibiotika aus NRW". This manuscript was written in part during a paper

PLUS

writing retreat sponsored by the Ruhr-University Research School

.

The abbreviations used are: 2D-PAGE, two-dimensional PAGE; Aqua dest., Aqua destillata; CLSI,

Clinical & Laboratory Standards Institute; DAB, denaturing alkylation buffer; DTNB, 5,5’-dithio-bis-(2-

nitrobenzoic acid); ICAT, isotope-coded affinity tag; IPTG, Isopropyl β-D-1-thiogalactopyranoside; LC-

MS/MS, liquid chromatography coupled with tandem mass spectrometry; MIC, minimal inhibitory

concentration; MRSA, Methicillin resistant Staphylococcus aureus; MS/MS, tandem mass spectrometry;

nanoUPLC-ESI-MS/MS, nano ultra-high-pressure liquid chromatography coupled with tandem mass

spectrometry; PEP, posterior error probability; pNPP, para-nitro-phenyl-phosphate; TCEP, tris(2-

carboxyethyl)phosphine; TNB, thio-2-nitrobenzoic acid; RuBPs, ruthenium(II) tris(4,7-diphenyl-1,10-

phenantrolin disulfonate).

ꢀ

17

FIGURE LEGENDS

FIGURE 1. Biosythesis of allicin from alliin (adapted from (17)). Upon damage of garlic clove cells,

alliin comes into contact with alliinase. Alliin is converted to allyl sulfenic acid and dihydroalanine. The

latter is unstable and rapidly degrades to pyruvate and ammonia. Two molecules of allyl sulfenic acid

spontaneously condense to allicin.

FIGURE 2. Allicin inhibits growth of E. coli and other microorganisms.

(

A) Minimal inhibitory concentrations (MIC) of allicin against pathogenic microorganisms, tested

according to CLSI standards in a microtiterplate-based assay using rich medium. (B) A more precise

determination of the minimal inhibitory concentration of allicin against E. coli MG1655 in MOPS

minimal medium. Overnight cultures did not show visible growth at allicin concentrations above

-

1

2

0

3 µg mL . Negative control: NC; not inoculated, Positive control: PC; without allicin. (C) Addition of

-

1

.79 mM of allicin (128 µg mL ) or 1 mM diamide to mid-logarithmic E. coli MG1655 (black arrow)

causes a growth arrest. Cultures resumed growth after 60 and 90 min, respectively.

FIGURE 3. Allicin induces a decline of sulfhydryl levels. Treatment of cells with 0.79 mM of allicin or

1

mM of diamide leads to a significant decrease of total cellular free thiols as measured with DTNB.

Values are the means and standard deviations from three biological replicates. Asterisks indicate a p-

value below 0.05 as determined by an unpaired student's t-test.

FIGURE 4. Allicin reacts with reduced glutathione in vitro and in vivo. Allicin was incubated with GSH

(

A) or GSSG (B) in a 0.5 or 1 fold molar ratio for 15 or 60 min. Subsequently, GSH and GSSG levels

were measured and compared to standard curves. Concentrations of the untreated controls were set to

00 %. Levels of reduced (C) and oxidized glutathione (D) were measured in E. coli cells after treatment

1

with allicin for 15 min. (E) The GSH:GSSG ratio is increased upon allicin stress. Values are the means

and standard deviations from three biological replicates. Asterisks indicate a p-value below 0.05 (*) or

0

.01 (**) as determined by an unpaired student's t-test.

FIGURE 5. Allicin treatment induces heat shock and oxidative stress response in E. coli. Newly

synthesized proteins of untreated (green) and allicin-treated (red) E. coli cultures were analyzed by

3

5

S methionine labeling prior to separation by 2D-PAGE and detection of spots using a phosphor screen.

S methionine incorporation was found to be 44 % of the control in the allicin-treated culture suggesting

a decreased protein synthesis upon allicin treatment. An overlay of gel images was performed to identify

proteins expressed in response to allicin treatment. The 25 spots with the highest synthesis rates in allicin-

treated E. coli were cut from non-radioactive gels, trypsin digested, and analyzed by mass spectrometry.

The numbers behind protein names represent spot numbers in Table 1. ALI indicates allicin-induced spots

that could not be identified by mass spectrometry. Shown are representative gels from three individual

biological replicates.

FIGURE 6. Allicin does not induce the formation of disulfide bonds in PhoAꢀ2-22. Measurement of the

activity of leaderless alkaline phosphatase shows that allicin in contrast to diamide does not readily

induce cystine bond formation in this protein. The pNPP conversion rate of the control was set to 1.

Values are the means and standard deviations from three biological replicates. Double asterisks indicate a

p-value below 0.01 as determined by an unpaired student's t-test.

ꢀ

18

FIGURE 7. Allicin leads to substantial S-allylmercapto modification of cellular proteins.

A) MS/MS annotation of peptide AceA232-256 identifies the S-allylmercapto modification of cysteine

44 induced by allicin treatment. The b- and y-ion series are marked in red. The total mass is consistent

(

2

with an S-allylmercapto modification of cysteine 244 and fragments b13, b14, and y16 provide direct

evidence. (B) OxICAT quantification of AceA232-263 containing the S-allylmercapto-modified cysteine

2

44. MS spectra of this peptide show increased labeling with the heavy ICAT label after allicin treatment

corresponding to a higher level of oxidation of the respective cysteine. Representative spectra from one of

the three biological replicates are displayed.

FIGURE 8. Isocitrate lyase activity is reduced in lysates of E. coli cells treated with allicin. While

diamide treatment (1 mM) had a minor effect on AceA activity, treatment with 0.79 mM allicin led to

significant enzyme inhibition. Values are the means and standard deviations from three biological

replicates. Asterisks indicate a p-value below 0.05 as determined by an unpaired student’s t-test.

FIGURE 9. Allicin induces protein aggregation. Increasing concentrations of allicin were applied to

crude extracts from E. coli for 15 min at room temperature. Proteins were separated into soluble (S) and

insoluble fraction (P) by centrifugation prior to SDS-PAGE. An increased amount of proteins in the pellet

fraction indicates aggregation.

FIGURE 10. RpoH is strongly stabilized upon allicin stress. After 15 min of allicin stress, protein

synthesis was blocked by the addition of chloramphenicol and samples were taken after time points

indicated. RpoH was visualized by SDS-PAGE and Western blot. Half-lives (t1/2) were calculated from

band intensities after analysis with ImageJ 1.49v (31).

FIGURE 11. E. coli lacking RpoH does not recover from allicin stress. E. coli strains (BB7222 and

BB7224; MC4100 and MC4100 ΔrpoH (26)) were grown in LB medium at 25 °C (○). At OD600=0.5,

cultures were split and treated with 0.79 mM allicin (◼), 1 mM diamide (▲), or DMSO (control) (ꢁ).

Growth was controlled for approximately 25 hours.

ꢀ

19

TABLES

TABLE 1. The 25 proteins with the highest synthesis rate in allicin-treated E. coli, identified

from 2D gels (see Fig. 3).

Spot

ID

Protein

theoretical theoretical

MW (Da) pI (pH)

NA NA

(EG Number) (47) Protein function (47)

1

2

ALI1 (NA)

Unidentified allicin-induced protein 1

RplL (EG10873)

GroS (EG10600)

IbpB (EG11535)

AhpC (EG11384)

ALI5 (NA)

50S ribosomal subunit protein L7/L12

Chaperonin Cpn10

12287

10380

16083

20748

NA

4.3989

4.9585

5.02

3

Chaperone, heat-inducible protein of HSP20 family

Alkyl hydroperoxide reductase, subunit C

Unidentified allicin-induced protein 5

D-ribose pyranase

4

5

4.8486

NA

RbsD (EG10817)

RplS (EG10880)

ALI7 (NA)

15282

13125

NA

5.9326

11.0625

NA

6

50S ribosomal subunit protein L19

Unidentified allicin-induced protein 7

Alkyl hydroperoxide reductase, subunit C

Flagellin, structural gene

7

8

9

0

1

2

3

4

AhpC (EG11384)

FliC (EG10321)

RplL (EG10873)

TrxC (EG11887)

NfuA (EG12935)

Hns (EG10457)

IbpA (EG11534)

Lpp (EG10544)

YbeD (EG11592)

GrpE (EG10416)

RpsM (EG10912)

IbpB (EG11535)

UspA (EG11390)

GroS (EG10600)

IbpB (EG11535)

ZapB (EG11878)

ALI20 (NA)

20748

51264

12287

15544

20984

15530

15764

8318

4.8486

4.3008

4.3989

4.8164

4.3271

5.2412

5.4785

9.7734

5.4053

4.4839

11.2163

5.02

1

50S ribosomal subunit protein L7/L12

Thioredoxin 2, zinc-binding

Fe/S biogenesis protein

DNA-binding global regulator H-NS

Chaperone, heat-inducible protein of HSP20 family

Murein lipoprotein

1

5

6

UPF0250 family protein

9821

Nucleotide exchange factor for the DnaKJ chaperone

30S ribosomal subunit protein S13

Chaperone, heat-inducible protein of HSP20 family

Global regulatory gene for stress response

Chaperonin Cpn10

21784

13091

16083

16056

10380

16083

9628

1

7

4.9219

4.9585

5.02

1

8

Chaperone, heat-inducible protein of HSP20 family

FtsZ stabilizer, coiled-coil protein

Unidentified allicin-induced protein 20

FtsZ stabilizer, coiled-coil protein

Flagellin, structural gene

1

2

9

0

1

2

3

4

5

4.4883

NA

ZapB (EG11878)

FliC (EG10321)

TrxA (EG11031)

NfuA (EG12935)

OmpA (EG10669)

9628

4.4883

4.3008

4.4751

4.3271

5.9678

51264

11799

20984

37177

Thioredoxin 1

Fe/S biogenesis protein

Outer membrane protein A

ꢀ

20

FIGURES

FIGURE 1

FIGURE 1. Biosythesis of allicin from alliin (adapted from (17)). Upon damage of garlic clove

cells, alliin comes into contact with alliinase. Alliin is converted to allyl sulfenic acid and

dihydroalanine. The latter is unstable and rapidly degrades to pyruvate and ammonia. Two

molecules of allyl sulfenic acid spontaneously condense to allicin.

ꢀ

22

FIGURE 2

FIGURE 2. Allicin inhibits growth of E. coli and other microorganisms.

(

A) Minimal inhibitory concentrations (MIC) of allicin against pathogenic microorganisms, tested

according to CLSI standards in a microtiterplate-based assay using rich medium. (B) A more

precise determination of the minimal inhibitory concentration of allicin against E. coli MG1655

in MOPS minimal medium. Overnight cultures did not show visible growth at allicin

-

1

concentrations above 23 µg mL . Negative control: NC; not inoculated, Positive control: PC;

-

1

without allicin. (C) Addition of 0.79 mM of allicin (128 µg mL ) or 1 mM diamide to mid-

logarithmic E. coli MG1655 (black arrow) causes a growth arrest. Cultures resumed growth after

6

0 and 90 min, respectively.

ꢀ

23

FIGURE 3

FIGURE 3. Allicin induces a decline of sulfhydryl levels. Treatment of cells with 0.79 mM of

allicin or 1 mM of diamide leads to a significant decrease of total cellular free thiols as measured

with DTNB. Values are the means and standard deviations from three biological replicates.

Asterisks indicate a p-value below 0.05 as determined by an unpaired student's t-test.

ꢀ

24

FIGURE 4

FIGURE 4. Allicin reacts with reduced glutathione in vitro and in vivo. Allicin was incubated

with GSH (A) or GSSG (B) in a 0.5 or 1 fold molar ratio for 15 or 60 min. Subsequently, GSH

and GSSG levels were measured and compared to standard curves. Concentrations of the

untreated controls were set to 100 %. Levels of reduced (C) and oxidized glutathione (D) were

measured in E. coli cells after treatment with allicin for 15 min. (E) The GSH:GSSG ratio is

increased upon allicin stress. Values are the means and standard deviations from three biological

replicates. Asterisks indicate a p-value below 0.05 (*) or 0.01 (**) as determined by an unpaired

student's t-test.

ꢀ

25

FIGURE 5

FIGURE 5. Allicin treatment induces heat shock and oxidative stress response in E. coli. Newly

synthesized proteins of untreated (green) and allicin-treated (red) E. coli cultures were analyzed

3

5

by S methionine labeling prior to separation by 2D-PAGE and detection of spots using a

3

5

phosphor screen. S methionine incorporation was found to be 44 % of the control in the allicin-

treated culture suggesting a decreased protein synthesis upon allicin treatment.. An overlay of gel

images was performed to identify proteins expressed in response to allicin treatment. The 25 spots

with the highest synthesis rates in allicin-treated E. coli were cut from non-radioactive gels,

trypsin digested, and analyzed by mass spectrometry. The numbers behind protein names

represent spot numbers in Table 1. ALI indicates allicin-induced spots that could not be identified

by mass spectrometry. Shown are representative gels from three individual biological replicates.

ꢀ

26

FIGURE 6

FIGURE 6. Allicin does not induce the formation of disulfide bonds in PhoA ꢀ 2-22.

Measurement of the activity of leaderless alkaline phosphatase shows that allicin in contrast to

diamide does not readily induce cystine bond formation in this protein. The pNPP conversion rate

of the control was set to 1. Values are the means and standard deviations from three biological

replicates. Double asterisks indicate a p-value below 0.01 as determined by an unpaired student's

t-test.

ꢀ

27

FIGURE 7

FIGURE 7. Allicin leads to substantial S-allylmercapto modification of cellular proteins.

(

A) MS/MS annotation of peptide AceA232-256 identifies the S-allylmercapto modification of

cysteine 244 induced by allicin treatment. The b- and y-ion series are marked in red. The total

mass is consistent with an S-allylmercapto modification of cysteine 244 and fragments b13, b14

,

and y16 provide direct evidence. (B) OxICAT quantification of AceA232-263 containing the S-

allylmercapto-modified cysteine 244. MS spectra of this peptide show increased labeling with the

heavy ICAT label after allicin treatment corresponding to a higher level of oxidation of the

respective cysteine. Representative spectra from one of the three biological replicates are

displayed.

ꢀ

28

FIGURE 8

FIGURE 8. Isocitrate lyase activity is reduced in lysates of E. coli cells treated with allicin.

While diamide treatment (1 mM) had a minor effect on AceA activity, treatment with 0.79 mM

allicin led to significant enzyme inhibition. Values are the means and standard deviations from

three biological replicates. Asterisks indicate a p-value below 0.05 as determined by an unpaired

student's t-test.

ꢀ

29

FIGURE 9

FIGURE 9. Allicin induces protein aggregation. Increasing concentrations of allicin were applied

to crude extracts from E. coli for 15 min at room temperature. Proteins were separated into

soluble (S) and insoluble fraction (P) by centrifugation prior to SDS-PAGE. An increased amount

of proteins in the pellet fraction indicates aggregation.

ꢀ

30

Article Doi

Allicin induces thiol stress in bacteria through S-allylmercapto modification of protein cysteines

DOI: 10.1074/jbc.M115.702308

Source and publish data:

Authors:

Article abstract of DOI:10.1074/jbc.M115.702308

Full text of DOI:10.1074/jbc.M115.702308

Products guided by the article

R&D Labs maybe for 539-86-6

Relevant to this article

Hot Product