Báo cáo khoa học: A novel ATP-binding cassette transporter is responsible for resistance to viologen herbicides in the cyanobacterium Synechocystis sp. PCC 6803

The Slr1174 protein of the DUF990 family is related to the permease subunit of an ABC-2-type transporter and its R115 mutation was found to be solely responsible for the observed methyl

for resistance to viologen herbicides in the

cyanobacterium Synechocystis sp PCC 6803

Jana Prosecka1,2, Artem V Orlov3, Yuri S Fantin3, Vladislav V Zinchenko3, Michael M Babykin4 and Martin Tichy1,2

1 Department of Autotrophic Microorganisms, Institute of Microbiology, Trebon, Czech Republic

2 Institute of Physical Biology, University of South Bohemia, Nove Hrady, Czech Republic

3 Department of Genetics, Moscow State University, Russia

4 International Biotechnological Centre, Moscow State University, Russia

Introduction

Molecular oxygen is essential for most organisms

However, during aerobic respiration or oxygenic

photosynthesis, reactive oxygen species, including the

superoxide anion radical (O2)), are formed In the photosynthetic electron transport chain, instead of NADP, O2may accept an electron from the reduction

Keywords

ABC-type transporter; cyanobacteria;

DUF990 family proteins; oxidative stress;

viologen herbicide resistance

Correspondence

M Tichy, Laboratory of Photosynthesis,

Institute of Microbiology, Opatovicky mlyn,

Trebon 379 81, Czech Republic

Fax: +400 384 340415

Tel: +400 384 340433

E-mail: tichym@alga.cz

(Received 23 January 2009, revised 18 May

2009, accepted 22 May 2009)

doi:10.1111/j.1742-4658.2009.07109.x

The charged quaternary ammonium compounds – methyl, ethyl and benzyl viologens – generate reactive oxygen species in photosynthetic cells Three independent methyl viologen-resistant spontaneous mutants of Synechocys-tis sp PCC 6803 were identified, in which the conserved R115 residue of the Slr1174 protein was replaced with G115, L115 and C115 The Slr1174 protein of the DUF990 family is related to the permease subunit of an ABC-2-type transporter and its R115 mutation was found to be solely responsible for the observed methyl viologen resistance Bioinformatic anal-ysis showed that in various bacterial genomes, two genes encoding another permease subunit and the ATPase component of an ATP-binding cassette transporter form putative operons with slr1174 orthologs, suggesting that the protein products of these genes may form functional transporters The corresponding genes in Synechocystis sp PCC 6803, i.e slr0610 for the per-mease and slr1901 for the ATPase, did not form such an operon However, insertional inactivation of any slr1174, slr0610 or slr1901 genes in both the wild-type and the R115-resistant mutant resulted in increased sensitivity to methyl, ethyl and benzyl viologens; moreover, single- and double-insertion mutants did not differ in their viologen sensitivity Our data suggest that Slr1901, Slr1174 and Slr0610 form a heteromeric ATP-binding cassette-type viologen exporter, in which each component is critical for viologen extru-sion Because the greatest increase in mutant sensitivity was observed in the case of ethyl viologen, the three proteins have been named EvrA (Slr1901), EvrB (Slr1174) and EvrC (Slr0610) This is the first report of a function for a DUF990 family protein

Abbreviations

ABC, ATP-binding cassette; BV, benzyl viologen; DQ, diquat; EV, ethyl viologen; MATE, multidrug and toxic compounds extrusion;

MDT, multidrug transporter; MFS, major facilitator superfamily; MV, methyl viologen; NBD, nucleotide binding domain; SMR, small multidrug resistance; TMD, transmembrane domain; TMH, transmembrane helices.

side of photosystem I to form O2) [1] This so-called

Mehler reaction [2] becomes a major electron-transfer

route in the presence of the redox cycling agent methyl

viologen (MV) This nonselective herbicide, also

known as paraquat (1,1-dimethyl-4,4¢-bipyridinium

dichloride), is a charged quaternary ammonium

compound that generates O2)under aerobic conditions

by cyclic univalent reduction and reoxidation [3] As

such, MV is commonly used to mimic and magnify the

oxidative stress that cells normally encounter during

photosynthesis and respiration

Organisms usually cope with MV by impairing MV

uptake [4,5] or enhancing MV efflux from the cell

[6–12], both of which lower the intracellular MV

con-centration Drug efflux pumps (also known as drug

efflux carriers, drug transporters or exporters) are

believed to play a major role in cell resistance to

vari-ous toxic agents in all domains of life Of the more

than 900 known families of transport systems, at least

35 include toxic ion and drug efflux carriers

More-over, in Gram-negative bacteria with two cellular

membranes, auxiliary transport proteins, such as

mem-brane fusion proteins and outer memmem-brane channels

(b-barrel porins; subclass 1.B), may be involved in the

extrusion of antibiotics and other toxic compounds

(Transporter Classification Database, TCDB: http://

www.tcdb.org/) [13,14] Although some efflux pumps

are selective for a given substrate, many transporters

are able to extrude a plethora of structurally unrelated

drugs, conferring the organism with a

multidrug-resistance phenotype [14,15]

On the basis of bioenergetic and structural criteria,

multidrug transporters (MDTs) can be divided into

two major classes: (a) primary ATP-binding cassette

(ABC)-type MDTs that use the free energy of ATP

hydrolysis to pump drugs out of the cell, and (b)

sec-ondary MDTs that utilize the transmembrane

electro-chemical gradient of protons or sodium ions to

extrude drugs from the cell [13,16–20] The majority of

MDTs identified and characterized to date are

ener-gized by the proton motive force (secondary

transport-ers) Depending on their size, similarity in primary

structure and topology, they can be divided into four

main transporter families or superfamilies [21]: major

facilitator superfamily (MFS) [22], small multidrug

resistance (SMR) [23], resistance-nodulation-cell

divi-sion [24], and multidrug and toxic compound extrudivi-sion

(MATE) [25] ABC transporters generally have two

similar halves, each of which contains two parts – a

hydrophobic transmembrane domain (TMD) and

a nucleotide-binding domain (NBD) Two TMDs form

a passageway for cargo, and two NBDs located in the

cytoplasm bind and hydrolyze ATP In most cases,

eukaryotic ABC exporters are expressed with all four domains in a single TMD⁄ NBD ⁄ TMD ⁄ NBD structure [26] In bacteria, one gene can encode a protein with a single combined TMD⁄ NBD to generate the functional homodimer, or two genes may separately encode NBD and TMD Moreover, the TMD⁄ TMD and NBD ⁄ NBD structures are described as components of full-size ABC transporters [27]

Recently, several bacterial genes have been isolated that encode transporters conferring MV resistance These include emrB of Escherichia coli, smvA of Sal-monella enterica [6], pqrB of Streptomyces coelicolor [7] and pqrA of Ochrobactrum anthropi [8,9], all of which encode MFS proteins Another example is the E coli emrE (or mvrC) gene which encodes a protein belong-ing to the SMR transporter family; this family is restricted to prokaryotic cells, and its members are the smallest multidrug efflux pumps with only four helices and no significant extramembrane domain [10] It has been shown that the SMR exporter YddG of S

enteri-ca extrudes MV in cooperation with OmpD porin, whereas the less abundant TolC porin is involved in

MV import into the cell [5] In general, more than half

of the known SMR proteins can export MV [27] Full genome analysis of the cyanobacterium Syn-echocystis sp PCC 6803 (hereafter Synechocystis sp.) has revealed several putative multidrug resistance exporters (http://www.membranetransport.org): five TMDs of ABC transporters; and five, three and two proteins from the MFS, resistance-nodulation-cell divi-sion and MATE families, respectively It is surprising that the complement of MFS transporters in Synecho-cystis sp is much fewer than the 70 found in E coli Moreover, no putative MV efflux pumps were found based on homology to known MV transporters A search for spontaneous MV-resistant Synechocystis sp mutants led to the identification of the MATE family PrqA protein, which functions as an MV exporter The prqAgene is cotranscribed with prqR, which encodes a TetR-like repressor protein A point mutation in prqR affects the putative DNA-binding domain of the PrqR repressor and results in derepression of the prqRA operon, leading to increased MV resistance [11,12] Our search for new MV-resistant mutants of Syn-echocystis sp with mutations that do not map within the prqR gene yielded three MV-resistant strains Each

of these carries a different substitution of a single amino acid residue in the Slr1174 permease protein that represented the TMD of an ABC transporter A bioinformatics search revealed two potential structural partners of this TMD, i.e Slr0610 and Slr1901 (another TMD and NBD), which could form a func-tional ABC-type viologen exporter The existence of

this heteromeric exporter in Synechocystis sp was

con-firmed by insertional mutagenesis in which inactivation

of each subunit in the MV-resistant mutant led to the

MV-sensitive phenotype In this study, we demonstrate

that even the native exporter is involved in the efflux

of MV and other quaternary ammonium herbicides,

and document how a single amino acid substitution

influences the substrate specificity of the transporter

Because the greatest increase in mutant sensitivity was

observed for ethyl viologen (EV), the three proteins

were named EvrA (Slr1901), EvrB (Slr1174) and EvrC

(Slr0610)

Results

Mapping of the MV-resistant mutants

In a previous study, the MV-resistant phenotype of a

Synechocystis sp Prq20 mutant was a result of

ele-vated expression of a MATE family drug exporter

encoded by the prqA gene, derepressed because of the

L17Q substitution in the PrqR repressor of the TetR

family [11,12] Also, the first MV-resistant mutant

iso-lated in this study mapped to the prqR gene, leading

to the Y52E substitution in PrqR The phenotype of

this new Prq mutant was identical to that of the

previ-ously described Prq20 mutant, suggesting that both the

L17Q and Y52E substitutions derepress the prqRA

operon

To identify new proteins providing protection

against MV, two strategies were employed to avoid the

characterization of other prqR mutants In the first

approach, the prqR gene from the MV-resistant strain

was amplified by PCR and examined for its ability to

transform wild-type cells into MV-resistant ones The

second approach was based on the generation of

MV-resistant mutants in a DprqA background Using these

approaches, we isolated three independent spontaneous

mutants with increased resistance to MV; two in the

wild-type background and one in the DprqA

back-ground Surprisingly, in all three MV-resistant

mutants, the mutation was localized in a single gene,

slr1174, which has been renamed evrB Moreover,

sequencing of evrB from all three mutants revealed

that the three different point mutations replaced the

same amino acid (R115) in the EvrB protein with three

different amino acids, i.e G115, L115 and C115,

respectively To transfer the C115 mutation from the

DprqA background into the wild-type background,

PCR-amplified evrB of the C115 strain was used to

transform wild-type cells The point evrB MV-resistant

mutants have been named EvrB–R115G, EvrB–R115L

and EvrB–R115C

All three substitutions in EvrB provided a similar level of protection to MV in three different sensitivity assays: doubling time in the presence of MV (Fig 1),

MV disk-diffusion assay (Fig 2) and threshold MV concentration (Table 1) (only results for EvrB–R115C are shown) This resistance was comparable with that

of the Prq20 mutant with the derepressed MATE-type exporter PrqA Wild-type cells grew poorly in the pres-ence of 0.6 lm MV, and no growth impairment was apparent in the EvrB–R115C and Prq20 mutants at the same MV concentration (Fig 1) In the disk-diffu-sion assays, at least 10-fold higher concentrations of

MV were needed to obtain the same level of growth inhibition (Fig 2) In EvrB–R115C, a 25-fold higher threshold concentration of MV was needed to inhibit growth on plates (Table 1)

The fact that EvrB–R115C and Prq20 can tolerate a much higher MV concentration than the wild-type without inducing the cell stress response was also

A

–1.4 –1.2 –1 –0.8 –0.6 –0.4 –0.2 0

0

Time (h)

0 μ M

0.3 μ M

0.6 μ M

1.2 μ M

B

–1.4 –1.2 –1 –0.8 –0.6 –0.4 –0.2

0

Prq20 EvrB-R115L WT ΔevrC ΔevrB EvrB-R115L/ ΔevrA

30 20

10

0

Time (h)

30 20

10

Fig 1 Growth curves of wild-type and mutant strains in the pres-ence of MV (A) Growth of the wild-type at different MV concentra-tions (B) Growth of the wild-type and mutants in the presence of 0.6 l M MV.

demonstrated by following expression of the slr1544

gene, which is transiently induced by various stresses

such as high light, low temperature, high salinity,

organic peroxide or MV [28–30] In the wild-type, we detected a 10–15-fold increase in the slr1544 transcript level after treatment with 5 lm MV for 20 min No such increase in the slr1544 transcript level was observed in the EvrB–R115C and Prq20 mutants after the same treatment, indicating that, unlike the wild-type, mutant cells were not stressed

EvrB protein The EvrB protein is annotated as a hypothetical protein

of the DUF990 family of functionally uncharacterized proteins that are related to the permease subunit of ABC-2-type transporters [31] This indicates that EvrB may be a component of an ABC transporter involved

in MV transport To determine whether the observed

MV resistance was caused by increased MV export or impaired import, the evrB gene was inactivated in the wild-type background (Fig S1) The resulting DevrB strain did not exhibit any obvious alteration in pheno-type with respect to its growth rate or pigment content However, it exhibited moderately increased sensitivity

to MV, which was reflected in a decreased MV thresh-old concentration in comparison with the wild-type (Table 1) evrB inactivation in the EvrB–R115C mutant background led to the same MV sensitivity as in the DevrB, indicating that the R115 substitution in EvrB is solely responsible for the elevated MV resistance in EvrB–R115C (Fig 1 and Table 1) This suggested that EvrB is involved in MV efflux in the wild-type and that this efflux is greatly enhanced in R115 point mutants

0.1 0.5

1 5

20 25 50

EvrB-R115C

WT

Prq20

10

Fig 2 The disk-diffusion assay was used to compare the

sensiti-vity of the wild-type, EvrB–R115C and Prq20 mutants to MV

Aliqu-ots (0.1 mL) of a particular exponential culture (adjusted to

D750= 0.5) were spread on BG11–agar plates, grown for one day,

treated with 5 lL of MV (at concentrations of 0.1, 0.5, 1, 5, 10, 15,

20, 25 and 50 l M) and incubated for 5 days.

Table 1 Threshold concentrations of methyl viologen (MV), ethyl viologen (EV) and benzyl viologen (BV) for wild-type, EvrB–R115C point mutant and mutants with insertion inactivation of the evrA, evrB and evrC genes Threshold concentrations were determined by serial dilutions as described in Experimental procedures.

Strain

Threshold concentration (l M)

The inhibitory effect of MV on photosynthetic

organisms involves accepting electrons from

photosys-tem I, followed by reduction of oxygen to the

super-oxide anion radical, leading to oxidative damage to

various cellular components To address the possible

role of EvrB in the adaptive response of

cyanobacte-rium to oxidative stress, we followed evrB expression

after treatment with 5 lm MV The treatment did not

influence the relative transcript abundance of evrB,

indicating the absence of a link between evrB

expres-sion and oxidative stress caused by MV

Putative EvrABC transporter

Because most ABC transporters consist of four

domains – two TMDs and two NBDs – the EvrB may

require protein partner(s) to form an active ABC-type

MV exporter The STRING search tool was used to

retrieve possible interacting genes⁄ proteins [32] This

tool predicted two potential functional partners of

EvrB, Slr0610 (renamed EvrC) and Slr1901 (renamed

EvrA), which had unknown functions Interestingly,

EvrC represents another putative permease of the

DUF990 family, whereas EvrA is a putative

ATP-binding protein of an ABC transporter

Advanced membrane topology algorithms based on

a combination of a cascade-neural network and hidden

Markov models using the evolutionary information

from multiple sequence alignments predicted the

pres-ence of six or seven a-helical transmembrane helices

(TMH) for EvrB (not shown) It is proposed that the

protein is embedded in the cytoplasmic membrane with

the N-terminus located inside the cell The six-TMH

structure was also predicted for EvrC Based on this,

there would be 12 TMH in the EvrB–EvrC

heteromer-ic transporter, in agreement with the canonheteromer-ical ABC

exporter topology

Functional analysis of the genes encoding the

heteromeric ABC-type exporter EvrABC

If all three components, EvrA, EvrB and EvrC, are

necessary for building the functional EvrABC

trans-porter, inactivation of any of the subunits should

result in a loss of drug efflux function and an increase

in mutant sensitivity Moreover, no increase in mutant

sensitivity should be observed after consecutive

inacti-vation of an additional subunit To confirm the

exis-tence of the hypothetical MV exporter and to elucidate

its role in Synechocystis sp., functional analysis of

genes encoding putative transporter components was

carried out Therefore, the evrC and evrA genes were

inactivated by insertions in the wild-type background

and in mutants with increased and decreased MV resis-tance – EvrB–R115C and DevrB The resulting DevrC and DevrA strains did not exhibit any obvious alter-ation in the phenotype

As mentioned above, a point mutation in the R115 evrB mutants resulted in increased cell resistance to

MV Remarkably, such mutations did not influence the resistance to other diquaternary ammonium herbi-cides that are structurally similar to MV such as EV and benzyl viologen (BV) (Table 1) However, inacti-vation of evrB in the wild-type background increased the relative sensitivity to these compounds by 160- and 50-fold, respectively This is a much more pronounced increase than that observed with MV (Table 1), indi-cating lower affinity of the wild-type EvrB for MV Inactivation of evrC and evrA, each of which encodes two putative EvrABC exporter subunits, decreased the resistance to all three viologens to similar levels in both the wild-type and EvrB–R115C backgrounds (Table 1) As in the case of the DevrB mutant, the slightly enhanced MV sensitivity of the DevrA and DevrC strains was strikingly different from their EV or

BV sensitivity, which again increased by 160- and 50-fold, respectively, relative to the wild-type strain Moreover, no significant differences in viologen sensitivity were detected among the single- and double-insertion mutants, i.e DevrA, DevrB, DevrC, EvrB–R115C⁄ DevrA, EvrB–R115C⁄ DevrC, DevrA⁄ -DevrB and -DevrB⁄ DevrC (Table 1) To prove that evrA and evrC inactivations are responsible for the observed sensitive phenotypes, the EvrB–R115C⁄ DevrA and EvrB–R115C⁄ DevrC mutants were complemented

by intact evrA and evrC from the recombinant plasmids, yielding EvrB–R115C⁄ evrA+ and EvrB– R115C⁄ evrC+ strains The complemented strains exhibited the same viologen resistance as the EvrB– R115C mutant (Table 1) The subunit inactivation data suggest that the EvrA (NBD), EvrB (TMD) and EvrC (another TMD) proteins form a heteromeric ABC-type viologen exporter in which each component

is necessary for viologen extrusion

Most of the known MV exporters belong to a hetero-geneous group of multidrug efflux pumps that confer resistance to a variety of structurally unrelated com-pounds [27] In order to determine the substrate specificity of the EvrABC exporter, the wild-type, EvrB–R115C strain and insertion mutants with inacti-vated exporter subunits were tested using the disk-diffu-sion assay for growth in the presence of an oxygen stressor (hydrogen peroxide), superoxide generators (pyrogallol, menadione and duroquinone) and quater-nary ammonium compounds (acriflavine, ethidium bro-mide and diquat) Diquat (DQ) was used because it

represents the 2,2¢-bipyridine derivative that is

structur-ally related to viologens, which are 4,4¢-bipyridines No

difference was observed in the sensitivity of the tested

strains to this set of toxic compounds, indicating that

the substrate specificity of the EvrABC exporter is

probably rather narrow, limited to particular viologens

Different substrate specificity was observed for the

MATE-type exporter PrqA derepressed in the Prq20

mutant Here, we observed 30-fold increase in the

resis-tance to both MV and DQ, but only twofold increase in

the resistance to BV in comparison with the wild-type

Discussion

The novel ABC transporter EvrABC

A novel ABC-type drug exporter that confers

resis-tance to the diquaternary ammonium herbicides MV,

EV and BV in Synechocystis sp was identified by the

insertional inactivation of genes encoding the putative

exporter components This insertion always abolished

the function of the transporter Two of the genes, evrB

and evrC, encode different permease subunits (TMDs)

from the DUF990 family of proteins of unknown

func-tion, and the third gene, evrA, encodes the

ATP-bind-ing subunit (NBD) of an ABC transporter Functional

association of the evrA, evrB and evrC genes was

pre-dicted by the STRING database based on the

observa-tion that in various bacterial genomes, the genes

encoding the orthologous counterparts of the EvrA,

EvrB and EvrC proteins are organized in putative

ope-rons The gene organization in such operons varied

considerably among organisms (Fig 3), suggesting

functional relevance of these gene clusters Also, the

tree view option of the pair-wise alignment in the

NCBI BLASTP database search (http://blast.ncbi.nlm

nih.gov/Blast.cgi) produced almost the same tree

topol-ogies for all three proteins (not shown), implying that

they coevolved as a functional complex Among

cyano-bacteria, the two proteins from the DUF990 family

were found in all strains with known genome

sequences, with the exception of

Thermosynechococ-cus elongatus It is interesting to note that the genes encoding the three transporter subunits, which form operons in most organisms, are dispersed all over the Synechocystis sp genome Apparently, in Synechocystis sp., most genes of the ancestral operons are scattered throughout the chromosome, possibly because of the presence of abundant repetitive elements that can lead

to genome rearrangements [33,34]

By selecting spontaneous MV-resistant mutants, one can generally expect two types of mutations In the first type, the ability to transport or metabolize MV is altered, whereas in the second type, the ability to deal with reactive oxygen species produced by MV is improved Surprisingly, only mutants affecting MV transport were found in numerous studies [5–10] In Synechocystis sp., two classes of mutants with increased resistance to MV were found The first class

of mutants mapped onto the prqR gene, encoding a regulator from the TetR family that controls expres-sion of the MATE-type exporter PrqA [11,12] The second class of mutants characterized in this study mapped onto the evrB gene, encoding a subunit of an ABC transporter Also, for evrB R115 mutants, enhanced MV export is the most likely explanation for the observed MV resistance We believe that the cir-cumstantial evidence that evr genes encode a true viologen exporter is rather convincing, although we cannot exclude some indirect effects of the EvrABC transporter on viologen resistance The observed violo-gen resistance was not caused by impaired violoviolo-gen import because inactivation of any of the evrA, evrB

or evrC genes resulted in a decrease and not an increase in viologen resistance Unlike the wild-type, evrB R115 mutants were not stressed in the presence

of MV, suggesting that there was a low intracellular concentration of superoxide anion radical in the mutant cells However, resistance to the superoxide anion radical itself did not change in any null Devr mutant, excluding a direct role of the transporter in superoxide resistance Finally, the different sensitivity

of the wild-type and the R115 evrB mutants for struc-turally and chemically similar viologens can be best

Fig 3 Gene organization in the evr cluster

in various organisms The position and size

of the genes encoding proteins that are most homologous to EvrA, EvrB and EvrC in different organisms are shown Note that the genes tend to cluster together despite changes in the gene order or in the gene orientation (in Thermus thermophilus) The disconnected genes are present at different locations of the genome.

explained by direct interaction of viologens with the

EvrABC transporter

Role of R115 in the substrate selectivity of

EvrB permease

Interestingly, the R115 residue replaced by G115, L115

or C115 in EvrB, is highly conserved in the

cyanobac-terial cluster despite the lower overall EvrB homology

(36% identity) There was only one conservative

replacement for K115 found in Trichodesmium

erythra-eum Moreover, a similar conserved R116 residue is

observed at the same location in the EvrC protein

This suggests a potential functional importance of this

amino acid Irrespective of the six- or seven-TMS EvrB

model, the conserved R115 is located on the

cytoplas-mic side of the third TMS It is probable that the

R115 substitutions in the EvrB of the MV-resistant

mutants influence the substrate specificity of the

trans-porter In all three reported R115 substitutions

(R115G, R115L and R115C), a positively charged

arginine residue was replaced, which resulted in the

same mutant phenotype, although the properties of the

three amino acids are quite different: glycine is tiny,

whereas leucine is large and hydrophobic and cysteine

is polar It seems that there is no strict requirement for

a particular amino acid at position 115 and that

removal of the positively charged arginine from the

cytoplasmic side of the EvrABC exporter is required to

increase the specificity or processivity for MV DQ, a

structural analog of viologens, did not seem to be

transported by the native EvrABC transporter This is

different from the Prq20 MV-resistant mutant, which

was also partially characterized in this study Here DQ

and MV, but not BV, were efficiently removed by

MATE-type exporter PrqA High selectivity and a

large difference in the transport efficiency of two

clo-sely related substrates is unusual among exporters of

quaternary ammonium compounds, with members of

the MFS family generally conferring a broader

resis-tance phenotype than members of the SMR family

[35] All of these results are related to secondary

MDTs that utilize the transmembrane electrochemical

gradient to drive the export There are no data on the

involvement of ABC transporters in the transport of

diquaternary ammonium compounds to date

DUF990 family proteins

The DUF990 family is formed by two groups of

pro-teins with the same predicted secondary structure, but

with weak mutual homology Interestingly, in most

organisms, there is a single protein from each group,

indicating that both groups may originate from an ancient duplication event For example, protein FN0879 from Fusobacterium nucleatum is paralogously related to FN0881 (E-value = 9e)10) in the same puta-tive operon and is also significantly homologous to both EvrB (E-value = 1e)14) and EvrC (E-value = 4e)7) Currently, DUF990 family proteins have been identified in 135 species across the bacterial kingdom, implying their potential functional importance [31] No function has yet been assigned to transporters contain-ing DUF990 proteins; however, based on the much higher conservancy observed among NBDs than TMDs, EvrA belongs to the ABC-2 subfamily of NBD components of bacterial ABC transport systems [36]

Of the proteins from this subfamily whose functions are known, the EvrA protein is most similar to NatA and DrrA NatA is an NBD component of the NatAB system catalyzing ATP-dependent Na+ extrusion [37] and DrrA is a part of the DrrAB bacterial exporter that confers resistance to the antibiotics daunorubicin and doxorubicin [38]

Regarding the function of the EvrABC transporter

in Synechocystis sp and other (cyano)bacteria, it is questionable whether its native function is to export viologens, because viologens are not normally present

in bacterial cells However, we believe that the infor-mation obtained from model viologen transport may help us to establish the true role of the EvrABC trans-porter and determine its natural substrate

Altogether, our data suggest the existence of a new heteromeric ABC-type drug exporter that, to our knowledge, is the first ABC transporter shown to be involved in the export of diquaternary ammonium viol-ogen herbicides This is also the first time that a func-tion has been associated with the widely distributed putative permease subunit from the DUF990 family

Experimental procedures

Growth conditions and generation of spontaneous mutants

Wild-type and mutants of Synechocystis sp were grown autotrophically in liquid BG11 medium supplemented with

agar Measurements of autotrophic growth rates were per-formed in microtitre plates (culture volume 0.25 mL) with intense shaking (900 rpm) Attenuance values at 750 nm were measured using a microplate reader (Tecan Sunrise, Vienna, Austria) The values plotted against time were used

to calculate the doubling time Spontaneous MV-resistant mutants were obtained by consecutive cultivation of the

wild-type or DprqA strain [12] in liquid media, with

increas-ing MV concentrations from 0.5 to 10 lm After platincreas-ing on

BG11–agar containing 20 lm MV, single colonies were

picked and grown on plates containing up to 60 lm MV All

MV-resistant mutants were maintained in media

supple-mented with 20 lm MV

Sensitivity to MV and other toxic compounds

The disk-diffusion assay used to estimate the sensitivity of

the wild-type and mutant strains to various toxic compounds

was a modification of standard methods [39,40] An 0.1-mL

aliquot of a particular exponential culture (adjusted to

1 day, treated with 5 lL of a solution containing an

appro-priate concentration of the test compound and incubated for

5 days The threshold concentrations of MV, BV and EV for

each strain were determined by serial dilution The

aliquot from each dilution was spotted on BG11–agar plates

containing different concentrations of the inhibitor The

low-est concentration of MV, BV or EV that inhibited cell growth

at all three dilutions was defined as the threshold

concentra-tion The values reported here are representative of those

obtained in three independent experiments

Mapping of MV-resistant mutants

described previously [41,42] Briefly, the isolated

chromo-somal DNA of a MV-resistant mutant was digested with a

set of restriction enzymes The restriction fragments were

fractionated by agarose gel electrophoresis and used to

trans-form the wild-type cells The transtrans-formed cells were plated

on BG11 plates covered with nitrocellulose filters (0.85 lm)

After 4 days, the filters were transferred to plates containing

20 lm MV The filter was subsequently transferred to plates

containing 40 and 60 lm on the eighth and tenth day,

respec-tively The sizes of the restriction fragments yielding

MV-resistant transformants were compared with an ordered

database of chromosomal restriction fragments and yielded

putative chromosome fragments carrying the resistance

mutation The genome region containing the resistance

mutation was determined as the intersection of these sets of

putative transforming fragments resulting from individual

restriction digests using the vhelper restriction analysis

program (http://orlovsergei.com/Progs/VH/VH.zip)

Mutant construction

The DNA fragments carrying the evrA, evrB and evrC

genes were amplified from the Synechocystis sp wild-type

genomic DNA by PCR using the following primers:

for evrA, 5¢-TTTGTCAGGTCAGTCGGGTGATG-3¢ and

5¢-GCCAACGGGAAGAAGCCAAGAC-3¢ and 5¢-TTGC-CGGATATCAAAGCCCAAG-3¢; and for evrC, 5¢-TGAT-CCTTTACCTGTGGCCCTGAC-3¢ and 5¢-GCCGCCCTT-GACTGAACTTTG-3¢ The evrA gene was cloned into pTZ18R (Pharmacia, Uppsala, Sweden) as a whole 1.7 kb PCR fragment, whereas evrB and evrC were cloned as 1.4 kb SmaI–SalI and 2.0 kb DraI–Ecl136II fragments, respectively The evrA gene was interrupted by insertion of a 1.3 kb kanamycin-resistance cassette For the evrB interruption, a portion of gene between nucleotides 210 and 584 was removed by inverse PCR (divergent primers 5¢-GAC

CAGC-3¢) and replaced with a 1.8 kb spectinomycin-resis-tance cassette The evrC gene was interrupted by insertion

of a 0.9 kb gentamycin-resistance cassette The resulting plasmids were transformed into Synechocystis wild-type cells to generate single Devr mutants To generate double mutants, the plasmids carrying DevrA and DevrC constructs were transformed into the EvrB–R115C point mutant and into the DevrB insertion mutant The transformants were selected and segregated on plates containing the

of the single- and double-insertion mutants were confirmed

by PCR analysis (data not shown) The relevant chromo-somal regions of the Devr mutants are shown in Fig S1

Gene expression, RNA isolation, reverse transcription and quantitative PCR Wild-type and mutants growing exponentially were treated

20 min, 1 h and 6 h A 5-mL aliquot of Synechocystis sp

The pellets were immediately frozen in liquid nitrogen, and

extracted using the modified hot phenol method [43] Briefly, frozen cells were thawed on ice and washed twice with the resuspension buffer, lysed in SDS lysis buffer, heated twice and extracted with hot acid phenol and chlo-roform Total RNA was precipitated by LiCl and ethanol Twenty nanograms of purified RNA was used for cDNA synthesis using random primers and SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA) Real time-quantitative PCR was performed on the Rotor-Gene 3000 system using the iQ SYBR Green Supermix (BioRad, Hercules, CA, USA) Each quantitative PCR experiment was performed in duplicate for two independent RNA isolations from the same culture 16S rRNA was used as the reference Its level was found to be proportional to that

of total RNA (estimated on an RNA agarose gel) under all

conditions The DDCt method was used to calculate the

gene expression levels

Secondary structure prediction of membrane

proteins

The web application server TRAMPLE, which is dedicated

to the detection and annotation of transmembrane protein

sequences, was used to predict the overall membrane

topol-ogy and the ability to form TMHs [44] TRAMPLE

provides different membrane topology algorithms based on

a combination of a cascade-neural network and hidden

Markov models The methods were tested on the recently

published structure of the S aureus Sav1866 protein, which

is a probable bacterial multidrug ABC transporter, and

good results were obtained [45]

Acknowledgements

The authors are grateful to Prof Teruo Ogawa for

providing the original evrA mutant and to Eva

Prach-ova for technical assistance This study was supported

by the Ministry of Education, Youth and Sports of the

Czech Republic (project nos MSM6007665808 and

ME 881) and by the Czech Academy of Sciences

(AV0Z50200510) It was also supported by a grant

from the Russian Foundation for Basic Research

07-04-00117 and the program ‘Leading Scientific

Schools’ (RI-112⁄ 001 ⁄ 211)

References

1 Asada K (2000) The water–water cycle as alternative

photon and electron sinks Phil Trans R Soc Lond B

Biol Sci 355, 1419–1431

2 Mehler AH (1951) Studies on reactions of illuminated

chloroplasts I Mechanism of the reduction of oxygen

and other Hill reagents Arch Biochem 33, 65–77

3 Hassan HM & Fridovich I (1978) Superoxide radical

and the oxygen enhancement of the toxicity of paraquat

in Escherichia coli J Biol Chem 253, 8143–8148

4 Kao SM & Hassan HM (1985) Biochemical

character-ization of a paraquat-tolerant mutant of

Escherichi-a coli J Biol Chem 260, 10478–10481

5 Santiviago CA, Fuentes JA, Bueno SM, Trombert AN,

Hildago AA, Socias LT, Youderian P & Mora GC

(2002) The Salmonella enterica sv Typhimurium smvA,

yddG and ompD (porin) genes are required for the

effi-cient efflux of methyl viologen Mol Microbiol 46, 687–

698

6 Hongo E, Morimyo M, Mita K, Machida I,

Hama-Inaba H, Tsuji H, Ichimura S & Noda Y (1994) The

methyl viologen-resistance-encoding gene smvA of

7 Cho YH, Kim EJ, Chung HJ, Choi JH, Chater KF, Ahn

BE, Shin JH & Roe JH (2003) The pqrAB operon is responsible for paraquat resistance in Streptomyces -coelicolor J Bacteriol 185, 6756–6763

8 Jo J, Won SH, Son D & Lee BH (2004) Paraquat resis-tance of transgenic tobacco plants over-expressing the

1391–1396

9 Won SH, Lee BH, Lee HS & Jo J (2001) An

heterologous host Escherichia coli Biochem Biophys Res Commun 285, 885–890

10 Lomovskaya O & Lewis K (1992) emr, an

USA 89, 8938–8942

11 Babykin MM, Sidoruk KV, Zinchenko VV, Nefedova

LN, Cerff R & Shestakov SV (2003) On the involve-ment of the regulatory gene prqR in the developinvolve-ment of resistance to methyl viologen in cyanobacterium

12 Nefedova LN, Fantin Iu S, Zinchenko VV & Babykin

MM (2003) The prqA and mvrA genes encoding carrier proteins control resistance to methyl viologen in the cyanobacterium Synechocystis sp PCC6803 Genetika

39, 336–340

13 Saier MH Jr (2000) A functional–phylogenetic classifi-cation system for transmembrane solute transporters Microbiol Mol Biol Rev 64, 354–411

14 Van Bambeke F, Glupczynski Y, Plesiat P, Pechere JC

& Tulkens PM (2003) Antibiotic efflux pumps in pro-karyotic cells: occurrence, impact on resistance and strategies for the future of antimicrobial therapy

J Antimicrob Chemother 51, 1055–1065

15 Poole K (2005) Efflux-mediated antimicrobial resis-tance J Antimicrob Chemother 56, 20–51

16 Baranova NN & Neyfakh AA (1997) Apparent involve-ment of a multidrug transporter in the fluoroquinolone resistance of Streptococcus pneumoniae Antimicrob Agents Chemother 41, 1396–1398

17 De Rossi E, Branzoni M, Cantoni R, Milano A, Riccardi G & Ciferri O (1998) mmr, a

cat-ionic dyes and inhibitors J Bacteriol 180, 6068–6071

18 Heir E, Sundheim G & Holck AL (1998) The

family encoding multidrug resistance FEMS Microbiol Lett 163, 49–56

19 Littlejohn TG, Paulsen IT, Gillespie MT, Tennent JM, Midgley M, Jones IG, Purewal AS & Skurray RA (1992) Substrate specificity and energetics of antiseptic and disinfectant resistance in Staphylococcus aureus FEMS Microbiol Lett 74, 259–265

20 Paulsen IT, Brown MH & Skurray RA (1996) Proton-dependent multidrug efflux systems Microbiol Rev 60, 575–608

21 Putman M, van Veen HW & Konings WN (2000)

Molecular properties of bacterial multidrug

transport-ers Microbiol Mol Biol Rev 64, 672–693

22 Marger MD & Saier MH Jr (1993) A major superfamily

of transmembrane facilitators that catalyse uniport,

symport and antiport Trends Biochem Sci 18, 13–20

23 Paulsen IT, Skurray RA, Tam R, Saier MH Jr, Turner

RJ, Weiner JH, Goldberg EB & Grinius LL (1996) The

SMR family: a novel family of multidrug efflux proteins

involved with the efflux of lipophilic drugs Mol

Micro-biol 19, 1167–1175

24 Saier MH Jr, Tam R, Reizer A & Reizer J (1994) Two

novel families of bacterial membrane proteins concerned

with nodulation, cell division and transport Mol

Micro-biol 11, 841–847

25 Brown MH, Paulsen IT & Skurray RA (1999) The

mul-tidrug efflux protein NorM is a prototype of a new

family of transporters Mol Microbiol 31, 394–395

26 Hollenstein K, Dawson RJ & Locher KP (2007)

Struc-ture and mechanism of ABC transporter proteins Curr

Opin Struct Biol 17, 412–418

27 Borges-Walmsley MI, McKeegan KS & Walmsley AR

(2003) Structure and function of efflux pumps that

con-fer resistance to drugs Biochem J 376, 313–338

28 Hihara Y, Sonoike K, Kanehisa M & Ikeuchi M (2003)

DNA microarray analysis of redox-responsive genes in

the genome of the cyanobacterium Synechocystis sp

strain PCC 6803 J Bacteriol 185, 1719–1725

29 Li H, Singh AK, McIntyre LM & Sherman LA (2004)

Differential gene expression in response to hydrogen

peroxide and the putative PerR regulon of

30 Murata N & Suzuki I (2006) Exploitation of genomic

sequences in a systematic analysis to access how

cyano-bacteria sense environmental stress J Exp Bot 57, 235–

247

31 Finn RD, Tate J, Mistry J, Coggill PC, Sammut SJ,

Hotz HR, Ceric G, Forslund K, Eddy SR,

Sonnham-mer EL et al (2008) The Pfam protein families

data-base Nucleic Acids Res 36, D281–D288

32 von Mering C, Jensen LJ, Kuhn M, Chaffron S,

Doerks T, Kruger B, Snel B & Bork P (2007)

STRING 7– recent developments in the integration

and prediction of protein interactions Nucleic Acids

Res 35, D358–D362

33 Fujibuchi W, Ogata H, Matsuda H & Kanehisa M

(2000) Automatic detection of conserved gene clusters

in multiple genomes by graph comparison and P-quasi

grouping Nucleic Acids Res 28, 4029–4036

34 Itoh T, Takemoto K, Mori H & Gojobori T (1999)

Evolutionary instability of operon structures disclosed

by sequence comparisons of complete microbial

genomes Mol Biol Evol 16, 332–346

35 Anthonisen IL, Sunde M, Steinum TM, Sidhu MS &

Sorum H (2002) Organization of the antiseptic

resis-tance gene qacA and Tn552-related beta-lactamase genes in multidrug-resistant Staphylococcus haemolyticus strains of animal and human origins Antimicrob Agents Chemother 46, 3606–3612

36 Reizer J, Reizer A & Saier MH Jr (1992) A new sub-family of bacterial ABC-type transport systems catalyz-ing export of drugs and carbohydrates Protein Sci 1, 1326–1332

37 Cheng J, Guffanti AA & Krulwich TA (1997) A

Mol Microbiol 23, 1107–1120

38 Guilfoile PG & Hutchinson CR (1991) A bacterial analog of the mdr gene of mammalian tumor cells is present in Streptomyces peucetius, the producer of daunorubicin and doxorubicin Proc Natl Acad Sci USA 88, 8553–8557

39 Bauer AW, Kirby WM, Sherris JC & Turck M (1966) Antibiotic susceptibility testing by a standardized single disk method Am J Clin Pathol 45, 493–496

40 Buchmeier N, Bossie S, Chen CY, Fang FC, Guiney

DG & Libby SJ (1997) SlyA, a transcriptional regulator

of Salmonella typhimurium, is required for resistance to oxidative stress and is expressed in the intracellular environment of macrophages Infect Immun 65, 3725– 3730

41 Ermakova-Gerdes S & Vermaas W (1999) Inactivation

of the open reading frame slr0399 in Synechocystis sp PCC 6803 functionally complements mutations near the Q(A) niche of photosystem II – a possible role of slr0399 as a chaperone for quinone binding J Biol Chem 274, 30540–30549

42 Tichy M & Vermaas W (2000) Combinatorial mutagen-esis and pseudorevertant analysis to characterize regions

in loop E of the CP47 protein in Synechocystis sp PCC

6803 Eur J Biochem 267, 6296–6301

43 Mohamed A & Jansson C (1989) Influence of light on accumulation of photosynthesis-specific transcripts in the cyanobacterium Synechocystis 6803 Plant Mol Biol

13, 693–700

44 Fariselli P, Finelli M, Rossi I, Amico M, Zauli A, Mar-telli PL & Casadio R (2005) TRAMPLE: the transmem-brane protein labelling environment Nucleic Acids Res

33, W198–W201

45 Dawson RJ & Locher KP (2006) Structure of a bacte-rial multidrug ABC transporter Nature 443, 180–185

Supporting information

The following supplementary material is available Fig S1 Schematic representation of the procedure used for the insertional inactivation of the evrB (A), evrA (B) and evrC (C) genes in the Synechocystis genome