EF0176 and EF0177 from Enterococcus faecalis V583 are substrate-binding lipoproteins involved in...

Downloaded from www.microbiologyresearch.org by

IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

EF0176 and EF0177 from Enterococcus faecalisV583 are substrate-binding lipoproteins involved inABC transporter mediated ribonucleoside uptake

Ingrid Lea Karlskas, Zhian Saleihan, Helge Holo, Geir Mathiesenand Vincent G. H. Eijsink

Correspondence

Vincent G. H. Eijsink

[email protected]

Received 26 September 2014

Accepted 26 January 2015

Department of Chemistry, Biotechnology and Food Science, Norwegian University of Life Sciences,P.O. Box 5003, N-1432 As, Norway

One of the ABC transporter systems in Enterococcus faecalis V583 is encoded by the ef0176–

ef0180 gene cluster, which differs from orthologous operons in related bacteria in that it contains

two genes putatively encoding substrate-binding proteins (SBPs). These SBPs, EF0176 and

EF0177, have previously been identified on the surface of E. faecalis. By phenotypic studies of

single and double knockout mutants, we show here that EF0176 and EF0177 are specific for

ribonucleosides and, by inference, that the EF0176–EF0180 ABC transporter plays a role in

nucleoside uptake. The specificity of the SBPs was mapped using growth experiments on a

medium, RPMI 1640, that only supports growth of E. faecalis when supplemented with purine

nucleosides or their corresponding bases. This analysis was complemented by studies with toxic

fluorinated pyrimidine ribonucleoside analogues and competition experiments. The data show that

EF0176 and EF0177 have broad and overlapping, but not identical, substrate specificities and

that they, together, are likely to bind and facilitate the transport of all common ribonucleosides.

Comparative sequence analysis and inspection of an available crystal structure of an orthologue,

PnrA from Treponema pallidum, showed that the strongest binding interactions between the

protein and the ligand involve the ribose moiety and that sequence variation in the binding site

primarily affects interactions with the base. This explains both the broad substrate specificity of

these binding proteins and the observed variations therein. The presence of two SBPs in this

nucleoside ABC transporter system in E. faecalis may improve the bacterium’s ability to scavenge

nucleosides.

INTRODUCTION

Enterococci are found in different environments such asfood, soil and the gastrointestinal tract of humans. Whileenterococci normally are harmless and even beneficial forhuman health through probiotic effects, they are also oppor-tunistic pathogens involved in nosocomial infections withhigh occurrence of antibiotic resistance (Lebreton et al.,2014). Thus, it is of interest to study how these bacteriainteract with their environment and to identify factors thatpotentially affect survival or could be targets for drug design.

Approximately 25 % of Enterococcus faecalis V583 proteinsassociated with the bacterial cell envelope are lipoproteins,and among these, many are putative components of ATPbinding cassette (ABC) transporters. ABC transporters mayhave a wide variety of functions, including uptake of

nutrients (e.g. sugars, peptides, polysaccharides, trace elements)and export of antibiotics or toxins (Biemans-Oldehinkelet al., 2006; Davidson et al., 2008; Garmory & Titball, 2004).A classical ABC transporter typically comprises an extra-cellular solute/substrate-binding protein (SBP), an intracel-lular ATPase, and two integral membrane proteins (permeases)(Saier, 2000), where the SBPs are thought to provide affinityand specificity. In Gram-negative bacteria, SBPs are solubleperiplasmic proteins whereas in Gram-positive bacteria,lacking the outer membrane, the SBPs are often anchored tothe cytoplasmic membrane by various mechanisms (Biemans-Oldehinkel et al., 2006). Most ABC-type transporters exhibita high specificity for one single substrate, or a group ofrelated substrates (Lubelski et al., 2007). E. faecalis ABCtransporters are known to be involved in antibiotic resis-tance (Manson et al., 2004; Singh et al., 2002), multidrugresistance (Davis et al., 2001; Lee et al., 2003), amino aciduptake (Fulyani et al., 2013) and secretion of toxins andvirulence factors (Gilmore et al., 1990; Low et al., 2003).

In a recent study on the surface proteome of E. faecalis(Bøhle et al., 2011) 17 lipoproteins (of a predicted total of

Abbreviations: ABC, ATP binding cassette; SBP, substrate-bindingprotein.

Three supplementary figures and one supplementary table are availablewith the online Supplementary Material.

Microbiology (2015), 161, 754–764 DOI 10.1099/mic.0.000045

754 000045 G 2015 The Authors Printed in Great Britain


IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

74) were detected, among which were EF0176 and EF0177that are putative SBPs of an ABC transporter encoded bythe ef0176–ef0180 gene cluster (Fig. 1). These genes areprobably co-transcribed, since homologues from otherbacteria comprise one operon (Deka et al., 2006; Webb &Hosie, 2006). EF0176 and EF0177 encode proteins of 357and 361 amino acids, respectively and share 72 % sequenceidentity. The proteins are homologous to PrnA from Trep-onema pallidum (Deka et al., 2006), RsnB from Streptococcusmutans (Webb & Hosie, 2006), BmpA from Lactococcuslactis (Martinussen et al., 2010), YufN from Bacillus subtilis(Belitsky & Sonenshein, 2011) and SP0845 from Strepto-coccus pneumoniae (Bidossi et al., 2012) for which a role innucleoside transport has been demonstrated or suggested(Fig. 1). Nucleosides can be exploited as C-source, N-source,or precursors for nucleotides or nucleic acids (Kilstrup et al.,2005). It is known that E. faecalis is prototrophic for purinesand pyrimidines (Kilstrup et al., 2005; Murray et al., 1993),but overall not much is known about nucleoside metabolismof E. faecalis.

Importantly, there is ample evidence for a connectionbetween the ability to synthesize or take up nucleosides andbacterial virulence. For example, a pyrimidine biosynthesisoperon of E. faecalis V583 (ef1721–ef1712) is upregulatedduring growth in the presence of chloramphenicol (Aakraet al., 2010) and differentially regulated in the presence oferythromycin (Aakra et al., 2005). It has been shown thatpurine biosynthesis genes of uropathogenic Escherichia coliare important for growth of the bacterium in urine and for

urovirulence in vivo, i.e. colonization of the bladder (Alteri& Mobley, 2012). Nucleotide biosynthesis pathways arecritical for growth of bacteria (E. coli, Salmonella typhimu-rium, Bacillus anthracis) in human blood, and the enzymesinvolved are considered as targets for development of newantibiotics (Samant et al., 2008). Extracellular nucleotidemetabolism is a well-known mediator of bacterial virulence(Firon et al., 2014). Over 50 % of FDA-approved antiviraland anticancer drugs are nucleoside or nucleobase analo-gues, and the potential use of these drugs as antibiotics isreceiving increasing attention (Sun & Wang, 2013).

We have studied the role of the predicted ABC transporterEF0176–EF0180 of E. faecalis V583 by analysing the phe-notypic consequences of deleting one or both of the puta-tive SBP encoding genes, ef0176 and ef0177. To map thefunctionality of these genes, the WT strain and the mutantswere characterized using the Biolog phenotypic microarray(Bochner et al., 2001), substrate utilization studies usingRPMI 1640, a medium that only supports exponentialgrowth of E. faecalis upon addition of purines, and sensi-tivity assays with toxic fluororibonucleosides.

METHODS

Bacterial strains and standard growth conditions. Bacterial

strains used in this study are listed in Table 1. E. faecalis V583 was

used as a parent strain for deletion mutant constructions, and as WT

control in all assays. E. faecalis strains were grown, unless otherwise

stated, at 37 uC without shaking in M17 medium (Oxoid)

cdd ef017 ef017 ef017 ef017 ef018deoC cdd ef0176 ef0177 ef0178 ef0179 ef0180deoC

cdd 10124 10125 10126 10127 10128deoC

cdd 10041 10042 10043 10044deoC

T. pallidum cdd hyp. Tp0321 Tp0322 Tp0323pnrAdeoC

S. mutans cdd rnsB rnsA rnsC rnsDdeoC

L. lactis MG1363 cdd bmpA nupA nupB nupCdeoC 60 kB

S. pneumoniae cdd SP0845 SP0846 SP0847 SP0848deoC

nupN nupO nupP nupQ

E. faecalis V583

E. faecalis OG1RF

E. faecium DO

B. subtilis

Fig. 1. Organization of gene clusters (putatively) encoding bacterial ABC transporters in several bacteria. The clusters typicallycomprise four genes encoding one extracellular solute/substrate binding domain (SBP; light grey), one intracellular ATPase(ATP binding protein; dark grey), and two integral membrane proteins (permeases; white). Other genes are represented byblack arrows and include deoC and cdd, which encode the ribonucleoside metabolic enzymes pyrimidine phosphorylase andcytidine deaminase, respectively, and a hypothetical protein (hyp).

Ribonucleoside uptake in Enterococcus faecalis

http://mic.sgmjournals.org 755


IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

supplemented with 0.4 % (w/v) glucose (GM17). E. coli GeneHogs

(Invitrogen) or TOP10 (Invitrogen) was cultured at 37 uC in Luria–

Bertani (LB) broth, with shaking. Solid media were prepared by addition

of 1.5 % (w/v) agar to the broth. When appropriate, antibiotics were

added at the following final concentrations: 100 mg ampicillin ml21 for

E. coli and 12.5 mg tetracycline ml21 for E. faecalis.

Construction of Def0177, Def0176 and Def0177/0176 knockout

mutants. Genomic DNA of E. faecalis V583 was prepared using a

E.Z.N.A. Bacterial DNA kit (Omega Bio-Tek) after cell disruption by

glass beads (106 mm; Sigma-Aldrich). Primers used in this study are

shown in Table 2. Mutants carrying deletions were constructed using

pAS222 as a cloning vector (Biswas et al., 1993; Jonsson et al., 2009).

To delete approximately 80 % of the gene encoding the 37 kDa

protein EF0177, a deletion fragment was constructed by splicing two

PCR fragments amplified from the E. faecalis V583 chromosome. The

two PCR fragments, about 1000 bp each and generated using primer

pairs 177LF / 177LR, and 177RF / 177RR, contained the ends of the

ef0177 gene as well as their upstream and downstream flanking

regions, respectively. The fragments were fused by using SOE (splicing

by overlapping extension) PCR (Horton et al., 1990) using primers

177LF and 177RR to create the final replacement fragment. The final

PCR product was treated with BamHI, and ligated into BamHI

digested pAS222 using an In fusion HD Cloning kit (Clontech

Laboratories). This procedure yielded the plasmid pAS222_DEF0177,

which was propagated in E. coli TOP10.

To delete approximately 70 % of the gene encoding the 35 kDa

protein EF0176, two fragments of about 1000 bp each were amplified

from the E. faecalis V583 genome using primer pairs 176LF / 176LR,

and 176RF / 176RR. The fragments were annealed and spliced by PCR

using primers 176LF and 176RR to create the final replacement

fragment. The Def0176 fragment was digested with BamHI and NotI,

and ligated into pAS222 pretreated with the same restriction enzymes,

yielding pAS222_DEF0176, which was propagated in E. coli GeneHogs.

To construct a double knockout construct with most of the EF0176

and EF0177-encoding genes deleted, two fragments were amplified

from the E. faecalis V583 genome using primer pairs 7677LF / 7677LR

and 7677RF / 7677RR. The two fragments thus obtained were spliced

by SOE–PCR using primers 7677LF / 7677RR to create the final

replacement fragment. The construction of this fragment was such

that it created a deletion with a total length of 2035 nt, comprising

992 downstream nucleotides of ef0176, the 53 nt intergenic region

and the 990 upstream nucleotides of ef0177. The Def0176/0177

fragment was phosphorylated using poly nucleotide kinase (New

England Biolabs) and ligated to pAS222 [previously cut with SnaBI

Table 1. Bacterial strains and plasmids used in this study

Strain or plasmid Characteristics Reference

Strains

E. coli GeneHogs Cloning host Invitrogen

E. coli TOP10 In-fusion cloning host Invitrogen

E. faecalis V583 Parental strain / reference strain Sahm et al. (1989)

E. faecalis D177 ef0177 deletion mutant This study

E faecalis D176 ef0176 deletion mutant This study

E. faecalis D176/177 ef0176 & ef0177 deletion mutant This study

Plasmids

pAS222 Cloning vector Jonsson et al. (2009)

pAS222_ DEF0177 Tetr, Ampr This study

pAS222_ DEF0176 Tetr, Ampr This study

pAS222_ DEF0176/EF0177 Tetr, Ampr This study

Table 2. Primers used in this study

Strain Primer Sequence (5§–3§)*

D177 177LF GCAGCCCGGGGGATCCGGACCTCAAGGGTACGCGTAT

D177 177LR GTCAACGCCGCCACCACCACATGCTGC

D177 177RF GCATGTGGTGGTGGCGGATGGTGGCGTTGACTTAACA

D177 177RR TAGAACTAGTGGATCCACCCACCATCATATCGGCTAA

D176 176LF GCAGCCCGGGGGATCCGAAGCAGCTGATGCGATTA

D176 176LR GGTTGGTCTGGGGCTTGGTCAATGTTTGATGTA

D176 176RF TTGACCAAGCCCCAGACCAACCTTAGAAACAAC

D176 176RR ACCGCGGTGGCGGCCGCCGACACCTTTAAGCGTTGAAGTTA

D176/177 7677LF GTGGCGCTTGTCGTCA

D176/177 7677LR TAAATAGCCGTCTGCTATTGCCGCCACA

D176/177 7677RF GCGGCAATAGCAGACGGCTATTTAAACGACAA

D176/177 7677RR ACCATCATATCGGCTAACTGT

*Underlining indicates the position of restriction sites (BamHI: GGATCC and NotI: GCGGCCGC). Italics indicate primer tails needed for SOE-

PCR.

I. L. Karlskas and others

756 Microbiology 161


IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

and treated with alkaline phosphatase (Promega)].The resultingplasmid was propagated in E. coli GeneHogs. The sequences of allpAS222 deletion constructs were verified by DNA sequencing.

Electrocompetent E. faecalis V583 cells were prepared with 4 % to 6 %glycine in the growth medium and transformed with the pAS222deletion plasmids as described by Holo & Nes (1989). Selection forcrossover events was done essentially as described by Biswas et al.

(1993), with some adaptations. E. faecalis transformants were obtainedby plating on GM17 with tetracycline at 28 uC, and then grown at37 uC (GM17, with tetracycline) to obtain integrants. Single crossoverintegration mutants obtained on plates with tetracycline at 37 uC were

then grown in liquid medium without the antibiotic for 50–100generations at 28 uC, to promote a second crossover event leading toplasmid excision and curing. Screening for tetracycline sensitivity wasdone by replica plating. Sensitive colonies were isolated and screenedfor the desired deletions by PCR. In the rest of this paper, the mutants

are referred to as D177, D176 and D176/177.

Biolog phenotype microarray (PM) assay. To determine substrate

utilization profiles, the E. faecalis V583 WT and the D176/177 doublemutant were plated on GM17 soft agar, and shipped to Biolog foranalysis. Growth measurements were performed using PM1-PM8plates (http://www.biolog.com/pdf/pm_lit/PM1-PM10.pdf) using thestandard procedures for E. faecalis (see www.biolog.com and e.g. Xue

et al., 2011). The best signal-to-noise ratio was obtained when usingpH 5.5 and Biolog Dye H, and these conditions were used throughout.Growth data were collected from cultures incubated at 37 uC for 20 h.

Growth assays. All growth assays were performed in duplicate ortriplicate, and the results presented are the mean values±SD. E.faecalis strains were grown in BHI at 37 uC overnight, and after beingharvested and washed, the cells were resuspended in BHI broth

(Oxoid) or RPMI 1640 medium with no added glutamine (R7509,Sigma-Aldrich), with a starting OD600 of 0.15–0.2. Cells were incubatedat 37 uC without agitation, and vortexed briefly before every OD600

measurement. For the analysis of substrate utilization, strains were

grown in RPMI 1640 supplemented with either 0.06 mM ribonucleo-sides (inosine, guanosine, adenosine, thymidine, cytidine and uridine;all from Sigma), 0.06 mM nucleobases (hypoxanthine, adenine, cytosine,uracil and thymine; all from Sigma) and/or 2 mM sugars (ribose andglucose; all from Sigma). For the sensitivity and competition assays,

strains were grown overnight as described above, and resuspended inBHI broth (to a starting OD600 of 0.15–0.2) containing 50 mg (i.e.approximately 0.2 mM) of the toxic ribonucleoside derivatives 5-fluorocytidine (5-FC) or 5-fluorouridine (5-FU) (both from Sigma)

ml21, in the absence or presence of 5 mM competing ribonucleosides.

RESULTS

Bioinformatic analysis

Fig. 1 shows the organization of the E. faecalis V583ef0176–ef0180 genes encoding a putative ABC transporterwith two SBPs. The figure also shows selected orthologousgene clusters identified by BLAST searches. Gene clusterscomprising four of the five genes in the ef0176–ef0180cluster are commonly found in Gram-positive bacteria,including members of the generae Streptococcaceae, Entero-coccaceae, Listeriaceae, Lactobacillaceae, Leuconostocaceaeand Bacillaceae (Martinussen et al., 2010). Unlike mosthomologous (putative) ABC transporters, the putative E.faecalis V583 ABC transporter has not one, but twopredicted SBPs, EF0176 and EF0177 (Fig. 1), sharing 72 %

sequence identity. Using BLAST searches with the completeef0176–ef0177 gene fragment, we identified this duplicationto be present in six E. faecalis strains, including strainsknown as pathogenic and strains known as probiotic [E.faecalis V583 (Fig. 1), E. faecalis OG1RF (Fig. 1), E. faecalis62, E. faecalis E32, E. faecalis Symbioflor and E. faecalis7L76]. The sequence identity, at the DNA level, was in theorder of 99 %. Other Enterococcus species, such as Entero-coccus faecium DO (Fig. 1), Enterococcus casseliflavus andEnterococcus hirae, and other Gram-positive bacteria onlyhad one SBP in this otherwise conserved ABC transporter.

Initial screening of phenotypes

To identify possible roles of EF0176 and EF0177 the E.faecalis V583 WT strain and the D176/177 double knockoutmutant strain were subjected to Biolog phenotypic micro-array screening. Eight standard bacterial plates (PM1-PM8)were used, including more than 700 carbon, nitrogen,phosphorous and sulfur sources. While results obtained withplates PM3–8 were difficult to interpret due to high signal-to-noise ratios (as is not uncommon for E. faecalis strains;Biolog, personal communication), the results of duplicatescreening with plates PM1 and PM2 (carbon sources,including four nucleosides, but no nucleotides or nucleo-bases) clearly showed that the D176/177 strain had a reducedability to utilize uridine, thymidine and inosine (Figs S1–S3;Table S1; available in the online Supplementary Material).The mutant also showed reduced growth on the fourthnucleoside in the screen, adenosine, but the difference wasbelow the applied threshold.

Substrate utilization studies

When grown in rich media (M17, BHI), the three E.faecalis mutants showed similar growth characteristics asthe WT (data not shown). We therefore turned to a RPMI1640 medium devoid of glutamine that is commonly usedfor culturing eukaryotic cells, but that does not sustaingrowth of E. faecalis very well (Fig. 2a). Glutamine is notessential for E. faecalis as long as the bacterium has accessto glutamate (Murray et al., 1993), which is present in RPMI1640. Since glutamine is a crucial precursor in purinebiosynthesis and a central compound in nitrogen metabol-ism, we figured that this medium might be used to detectdeficiencies in nucleoside transport, at least for purines.

Growth studies with RPMI 1640 showed that addition ofsingle purine nucleosides (adenosine, inosine or guano-sine) or the corresponding purine bases (adenine or hypo-xanthine; guanine not tested) stimulated growth of WT E.faecalis (Fig. 2b–d; Table 3). The bases adenine and hypo-xanthine also stimulated growth of the three mutants(Table 3), indicating that EF0176 and EF0177 are notinvolved in the uptake of these compounds. The singlemutants showed WT-like growth in media supplied withinosine, guanosine or adenosine, except for D176, whichconsistently showed impaired growth on adenosine. The




IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

double mutant, however, showed strongly impaired growthon adenosine and inosine, and also some growth retarda-tion on guanosine. Taken together, these results suggestthat EF0176 and EF0177 are involved in purine nucleosideutilization and that the two proteins have overlapping butnot identical functions.

Growth of E. faecalis strains on RPMI 1640 was notstimulated by pyrimidine bases (cytosine, uracil or thymine)nor by pyrimidine nucleosides (cytidine, uridine or thymi-dine). Thus, the experimental set-up described above couldnot be used to assess involvement of EF0176 and EF0177 intransport of pyrimidine nucleosides.

Uptake of toxic fluoro-pyrimidines

Possible involvement of EF0176 and/or EF0177 in uptakeof pyrimidine nucleosides was assessed by studying theeffects of pyrimidine analogues 5-fluorocytidine (5-FC)and 5-fluorouridine (5-FU) on growth of WT and mutantstrains on BHI. These nucleoside analogues become toxicto bacteria after uptake and subsequent conversion throughsalvage pathways, and are therefore useful for studying

pyrimidine metabolism (Martinussen et al., 1994). Fig. 3 (a–c) shows that growth of the WT strain is inhibited by 5-FCand, to a somewhat larger extent, 5-FU. This demonstratesthat these toxic pyrimidine analogues are taken up and meta-bolized. The double mutant strain was less susceptible to 5-FC(Fig. 3b) and 5-FU (Fig. 3c) than the WT strain, indicatingthat uptake of the pyrimidine nucleosides is impaired in theabsence of both EF0716 and EF0177. Both single mutants(D176 and D177) showed WT-like sensitivity to 5-FU,indicating that EF0176 and EF0177 facilitate uptake of FUin a similar manner (Fig. 3b). The D176 strain also showed aWT-like response to 5-FC (Fig. 3b), whereas the D177 straingrew somewhat better than the WT (Fig. 3c), indicatingreduced uptake of 5-FC, albeit not as strongly reduced as inthe double mutant. Taken together, the results suggest thatboth EF0176 and EF0177 are involved in uptake ofpyrimidine nucleosides. The data indicate that both EF0176and EF0177 interact with uridine and cytidine and that thetwo proteins may differ slightly when it comes to theirinteractions with cytidine.

To gain further insight into the specificity of EF0176 andEF0177 for pyrimidines, the effects of adding excess

(b)

0.9

1.0 No supplements(a)

0.5

0.6

0.7

0.8

WT

D177

0.2

0.1

0.3

0.4D176

D176/177

00 1 2 3 4 5 6 7 8

Time (h)

+ Guanosine(c) (d)

+ Inosine

+ Adenosine

OD

600

0.91.0

0.50.60.70.8

WT

D177

0.20.1

0.30.4

D176

D176/177

00 1 2 3 4 5 6 7 8 9

Time (h)

OD

600

0.9

1.0

0.5

0.6

0.7

0.8

WT

D177

0.2

0.1

0.3

0.4D176

D176/177

00 1 2 3 4 5 6 7 8

Time (h)

OD

600

0.9

1.0

0.5

0.6

0.7

0.8

WT

D177

0.2

0.1

0.3

0.4D176

D176/177

00 1 2 3 4 5 6 7 8

Time (h)

OD

600

Fig. 2. Growth of E. faecalis V583 and mutants on RPMI 1640 medium supplemented with 0.06 mM purine nucleosides. Cellswere grown in RPMI 1640 (a) or RPMI 1640 supplemented with inosine (b), guanosine (c), or adenosine (d). The OD600 wasdetermined every hour. Each curve corresponds to the mean of three independent experiments; the SD is indicated. An overviewover growth on these and other compounds is provided in Table 3.




IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

amounts of various ribonucleosides (5 mM) on the toxicityof 5-FC and 5-FU (50 mg ml21, i.e. approximately 0.2 mM)were assessed in competition experiments. If the protectiveeffect of a ribonucleoside on a mutant would differ fromthe effect on the WT, this would show that EF0176 orEF0177 interacts with that ribonucleoside. The results ofthe competition experiments are summarized in Table 4.For all strains, thymidine reduced the sensitivity to 5-FCand 5-FU, whereas the Biolog data (Figs S1–S3) show thatat least one of the two SBPs transport thymidine. Thus, theprotective effect of thymidine, which was only partial, islikely to be due to interactions with other transporters thattransport 5-FC and 5-FU. Excess uridine reduced sensitiv-ity in both single mutants, but not in the double mutant,showing that both EF0176 and EF0177 have specificity foruridine, as was to be expected since the sensitivity datashow that (only) the double mutant has reduced sensitivityto 5-FU. The third pyrimidine, cytidine, showed betterprotection of WT and the D176 mutant compared to theD177 mutant. While this difference was very small, it is inaccordance with the slight difference in affinity for cytidinesuggested by the sensitivity data presented in Fig. 3(c).

DISCUSSION

The EF0176 and EF0177 proteins caught our attentionbecause they were detected as part of the surface proteomeof E. faecalis V583 (Bøhle et al., 2011) and because theirfunctions were not known. We also noted that the struc-ture of the ef0176–ef0180 gene cluster differed from thestructure of potentially orthologous gene clusters (Fig. 1).Finally, we considered the potential role of these proteinsin transport of nucleosides of interest because the ability toeffectively scavenge such valuable compounds from theenvironment could provide the bacterium with a compet-itive advantage, e.g. in pathogen fitness (Jenkins et al.,2011; Samant et al., 2008). Rigottier-Gois et al. (2011) havepreviously performed a large-scale screening of insertionalmutants of E. faecalis V583 that included mutants withinsertions in ef0176 and ef0177. The mutant library wassubjected to several test conditions related to host adapta-tion, which showed no phenotypic effects for the ef0177mutant, whereas the ef0176 mutant showed decreasedgrowth under non-stress conditions, and increased sens-itivity to serum independent opsonophagocytosis (Rigottier-Gois et al., 2011). We created single and double deletionmutants of these two genes, and did not detect any growthdefects on regular rich medium. As part of ongoing studieson the interplay between enterococci and human cells in ourlab, we had been growing low density cultures of E. faecalison RPMI 1640, a medium meant for growing eukaryotic cellsand far from optimal for the bacteria. It turned out that E.faecalis V583 can grow well on RPMI 1640 if the medium issupplied with purine nucleosides or purine nucleobases, andthis allowed us to demonstrate that both EF0176 and EF0177are involved in the transport of purine nucleosides (Table 3).In the double mutant, growth on RPMI 1640 supplementedwith adenosine or inosine is only marginally above back-ground, suggesting that the EF0176–EF0180 ABC trans-porter system, with its two SBPs, EF0176 and EF0177, is theby far most dominant, if not the only, system dedicated toimport of these two nucleosides in E. faecalis V583. The dataindicate that alternative import routes exist for guanosine.

Since pyrimidines did not stimulate growth on RPMI 1640,the role of EF0176 and EF0177 in pyrimidine importneeded to be addressed in a more indirect manner, usingtoxic uridine and cytidine variants (5-FU and 5-FC) andcompetition experiments. This approach showed thatEF0176 and EF0177 also are capable of importing thepyrimidine nucleosides uridine and cytidine, whereas datafor thymidine were non-conclusive (note that the Biologdata indicate that at least one of the two SBPs interacts withthymidine). The data for the pyrimidine nucleosides do notallow conclusions as to whether the EF0176–EF0180 ABCtransporter system is the dominant or perhaps the onlysystem dedicated to import of these compounds in E.faecalis V583.

Taken together, the data show that EF0176 and EF0177both transport all common nucleosides. The data alsoindicate that the two proteins have minor differences in

Table 3. Schematic overview over growth phenotypesobserved in substrate utilization assays

E. faecalis WT and mutants were grown on RPMI 1640 medium

supplemented with 0.06 mM nucleosides, 0.06 mM nucleobases and/

or 2 mM sugars. Background growth, as on RPMI with no additions,

is scored as 22. Growth above background is scored as – (e.g. the

double mutants in Fig. 2b, d), ++ indicates ‘maximum’ growth

(e.g. the WT in Fig. 2b–d), and + indicates growth in between

– and ++ (as observed in Fig. 2c, d for D176/177 and D176,

respectively).

Strain: WT D176 D177 D176/177

Addition

None (RPMI) 22 22 22 22

Adenosine ++ + ++ 2

Inosine ++ ++ ++ 2

Guanosine ++ ++ ++ +

Thymidine 22 22 22 22

Uridine 22 22 22 22

Cytidine 22 22 22 22

Adenine ++ ++ ++ ++

Hypoxanthine ++ ++ ++ ++

Cytosine 22 22 22 22

Thymine 22 22 22 22

Uracil 22 22 22 22

Glucose 22 22 22 22

Ribose 22 22 22 22

Adenosine + glucose* ++ + ++ 2

Adenosine + ribose* ++ + ++ 2

*Same growth as on adenosine without any further additions for all

strains.




IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

functionality: EF0176 seems to be more important forimport of adenosine than EF0177 (Fig. 2d), whereas EF0177seems to interact more strongly with cytidine than EF0176.Similar broad substrate profiles have been described for the

few orthologues (Fig. 1) for which such substrate profileshave been tested, either by in vitro binding studies with therecombinantly produced protein (PnrA; Deka et al., 2006)or by studying transport directly (RnsB; Webb & Hosie,

2.0(a) (b)

1.0

1.5

0.5

5-FU WT

BHI WT

5-FU Δ177

5-FU Δ176

5-FU Δ176/177

5-FC WT

BHI WT

5-FC Δ177

5-FC Δ176

5-FC Δ176/177

0

Time (h)

(c)

OD

600

2.0

1.0

1.5

0.5

0

OD

600

0 1 2 3 4 5 6

2.0

1.0

1.5

0.5

0

Time (h)

OD

600

0 1 2 3 4 5 6

Time (h)

0 1 2 3 4 5 6

BHI Δ177

BHI Δ176

BHI Δ176/177

BHI WT

Fig. 3. Growth of E. faecalis WT and mutants in BHI broth supplemented with fluorinated nucleosides. (a) BHI broth, noadditions; (b) BHI broth plus 50 mg ml”1 (0.2 mM) 5-fluorocytidine; (c) BHI broth plus 50 mg ml”1 (0.2 mM) 5-fluorouridine.Each curve corresponds to the mean of two independent experiments; the SD is indicated.

Table 4. Sensitivity of E. faecalis strains to toxic pyrimidine ribonucleoside derivatives and protective effect of natural pyrimidineribonucleosides

Strain Sensitivity to 5-FU* Protecting pyrimidine

nucleosidesD

Sensitivity to 5-FC* Protecting pyrimidine

nucleosidesD

WT ++ U, T, C ++ U, T, C

D177 ++ U, T, (C)d +(+)d U, T, (C)d

D176 ++ U, T, C ++ U, T, C

D176/177 + T + T

*Growth curves are shown in Fig. 3. Scoring: ++, as sensitive as WT; +, less sensitive than WT.

DCompetition experiments; see text for details.

dWe observed a statistically non-significant difference between D176 and D177 in terms of the effect of 5-FC and the protective effect of C; see text

for details.




IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

2006) or by using strategies similar to the ones used here(BmpA; Martinussen et al., 2010).

The growth rate of bacteria often correlates with the rate atwhich they can generate GMP and AMP pools (Hedstrom,2009), hence purine availability may have a direct effect onbacterial growth. Studies with minimal media have shownthat purines and pyrimidines are not essential for growth ofE. faecalis (Murray et al., 1993), but recent data show thataddition of purines can stimulate growth of some strains(Khan et al., 2013). Maximizing the ability to scavengepurines from the environment could thus be an importantdeterminant of bacterial fitness and even pathogenicity. Itis conceivable that the presence of two SBPs in the purineimporting ABC transporter of E. faecalis V583 endows thisbacterium with a competitive advantage. Having two SBPscould be beneficial by broadening the substrate specificityof the system or by simply increasing the total number ofavailable nucleoside binding sites. Further work is neededto study the biological significance of the duplication of theSBPs in E. faecalis V583, e.g. in infection models.

ABC-type nucleoside transporter systems in bacteria wereonly discovered in 2006, when the structure and bindingproperties of PnrA from Treponema pallidum were

described (Deka et al., 2006; Webb & Hosie, 2006).Binding studies, using isothermal titration calorimetry,showed that PnrA binds to purines with higher affinitythan pyrimidines. The structure of PnrA provides insightinto the structural basis of the broad substrate specificitiesof SBPs. Fig. 4 shows interactions between PnrA and boundinosine whereas Fig. 5 shows an annotated sequencealignment of PnrA with EF0176 and EF0177 and otherorthologues that are included in Fig. 1. Interactions withthe ribose moiety include several strong hydrogen bondsinvolving fully conserved residues (Asp108, Asp238 andLys260 in PnrA). Also, residues sterically contributing toshaping the ribose binding pocket are fully (Phe 161 inPnrA) or functionally (Met158 in PnrA) conserved. Bindingof the base is mediated in part by aromatic stacking betweentwo fully conserved phenylalanines (Phe36 and Phe186 inPnrA) and the conjugated double bond systems in the base.This type of interactions is less specific than hydrogenbonds. In addition, the base has two direct hydrogen bondsinvolving a fully conserved (Asp27 in PnrA) and a func-tionally conserved (Asn37 in PnrA and all orthologuesexcept SP0845 which has Asp) residue. Other residuesinvolved in base binding, mainly via water-mediated hydro-gen bonds, are Asp33, Ser28, Ser187 and Gly212, the latter

Phe161

Lys260

Asp238

Gly212

Asp33

Asn37

Ser28

Ser187

Asp27

Phe186Asp108

Met158

Phe36

Fig. 4. Structure of PrnA in complex with inosine (pdb code 2fqw; Deka et al., 2006). The picture shows inosine and the sidechains of all residues that directly interact with the nucleoside, or are involved in water mediated hydrogen bonds to thenucleoside and/or seem crucial for shaping the binding site. Backbone atoms are only shown in cases where they are relevantfor interactions with the ligand (Gly212). Water molecules are indicated by red spheres and hydrogen bonds are drawn asdashed lines. Carbons appear in green (protein), pink (protein) and yellow (inosine); other colours: red, oxygen; blue, nitrogen;gold, sulphur. Residues with green carbons are primarily involved in interactions with the ribose moiety; residues with pinkcarbons are primarily involved in interactions with the base moiety. Slightly darker colours are used for the three residues thatare not fully conserved and for which structural inspection indicated that the variation could affect the affinity for certainnucleosides (Gly212, Ser28, Ser187; see alignment in Fig. 5 and text for more details).




IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

prnA 55575555565957576060

115

115115115119117117120120

171169168168168176174174177177

229228228228228235233233236236

284285288288288293293293296296

333337329329329338335335339339

116

NupNBmpaRnsBSP0845DO10041EF0176

EF0177OG10124

OG10125* * * *

* *

*

*

*

* * *

* * * * * * * * * * * * * * * * *

* * * * * * * *

* * * * * * * * * * * * * * *

* *

* * * * * *

prnANupNBmpaRnsBSP0845DO10041EF0176

EF0177OG10124

OG10125


EF0177OG10124

OG10125


EF0177OG10124

OG10125


EF0177OG10124

OG10125


EF0177OG10124

OG10125




IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

three of which show sequence variation that could affectbinding specificity (Fig. 5). Thus, the ribose moiety interactsvery tightly, using ‘precise interactions’ (i.e. direct hydrogenbonds) and involving fully conserved residues, whereas thebase interacts through less specific interactions (aromaticstacking) and more indirect hydrogen bonds involving lessconserved residues. A dominating role of interactions withribose may explain the broadness of the binding specificity.Interestingly, the O2 of the ribose is heavily involved inbinding interactions, interacting strongly with fully con-served Asp108 and Lys260. So, one might expect somedifferences in affinity between ribonucleosides and deoxy-ribonucleosides, as indeed has been suggested for PnrA(Deka et al., 2006).

EF0176 and EF0177 differ in that Gly212 (PnrA number-ing) in EF0177 is an alanine in EF0176. It remains to beseen whether this difference accounts for the difference insubstrate specificity. Generally, further studies to addressfunctional differences between EF0176 and EF0177 andtheir biological implications would be of interest.

ACKNOWLEDGEMENTS

This work was funded by the FUGE programme of the NorwegianResearch Council, grant number 183637. We thank Phil Pope for helpfuldiscussions and Zarah Forsberg for assistance in preparing Fig. 4.

REFERENCES

Aakra, A., Vebø, H., Snipen, L., Hirt, H., Aastveit, A., Kapur, V., Dunny,G., Murray, B. E. & Nes, I. F. (2005). Transcriptional response ofEnterococcus faecalis V583 to erythromycin. Antimicrob AgentsChemother 49, 2246–2259.

Aakra, A., Vebø, H., Indahl, U., Snipen, L., Gjerstad, O., Lunde, M. &Nes, I. F. (2010). The Response of Enterococcus faecalis V583 toChloramphenicol Treatment. Int J Microbiol 2010, 483048.

Alteri, C. J. & Mobley, H. L. (2012). Escherichia coli physiology andmetabolism dictates adaptation to diverse host microenvironments.Curr Opin Microbiol 15, 3–9.

Belitsky, B. R. & Sonenshein, A. L. (2011). CodY-mediated regulationof guanosine uptake in Bacillus subtilis. J Bacteriol 193, 6276–6287.

Bidossi, A., Mulas, L., Decorosi, F., Colomba, L., Ricci, S., Pozzi, G.,Deutscher, J., Viti, C. & Oggioni, M. R. (2012). A functional genomicsapproach to establish the complement of carbohydrate transporters inStreptococcus pneumoniae. PLoS ONE 7, e33320.

Biemans-Oldehinkel, E., Doeven, M. K. & Poolman, B. (2006). ABCtransporter architecture and regulatory roles of accessory domains.FEBS Lett 580, 1023–1035.

Biswas, I., Gruss, A., Ehrlich, S. D. & Maguin, E. (1993). High-efficiency gene inactivation and replacement system for gram-positivebacteria. J Bacteriol 175, 3628–3635.

Bochner, B. R., Gadzinski, P. & Panomitros, E. (2001). Phenotypemicroarrays for high-throughput phenotypic testing and assay of genefunction. Genome Res 11, 1246–1255.

Bøhle, L. A., Riaz, T., Egge-Jacobsen, W., Skaugen, M., Busk, O. L.,Eijsink, V. G. & Mathiesen, G. (2011). Identification of surfaceproteins in Enterococcus faecalis V583. BMC Genomics 12, 135.

Davidson, A. L., Dassa, E., Orelle, C. & Chen, J. (2008). Structure,function, and evolution of bacterial ATP-binding cassette systems.Microbiol Mol Biol Rev 72, 317–364.

Davis, D. R., McAlpine, J. B., Pazoles, C. J., Talbot, M. K., Alder, E. A.,White, C., Jonas, B. M., Murray, B. E., Weinstock, G. M. & Rogers, B. L.(2001). Enterococcus faecalis multi-drug resistance transporters: applica-tion for antibiotic discovery. J Mol Microbiol Biotechnol 3, 179–184.

Deka, R. K., Brautigam, C. A., Yang, X. F., Blevins, J. S., Machius, M.,Tomchick, D. R. & Norgard, M. V. (2006). The PnrA (Tp0319; TmpC)lipoprotein represents a new family of bacterial purine nucleosidereceptor encoded within an ATP-binding cassette (ABC)-like operonin Treponema pallidum. J Biol Chem 281, 8072–8081.

Firon, A., Dinis, M., Raynal, B., Poyart, C., Trieu-Cuot, P. & Kaminski,P. A. (2014). Extracellular nucleotide catabolism by the Group BStreptococcus ectonucleotidase NudP increases bacterial survival inblood. J Biol Chem 289, 5479–5489.

Fulyani, F., Schuurman-Wolters, G. K., Zagar, A. V., Guskov, A.,Slotboom, D. J. & Poolman, B. (2013). Functional diversity of tandemsubstrate-binding domains in ABC transporters from pathogenicbacteria. Structure 21, 1879–1888.

Garmory, H. S. & Titball, R. W. (2004). ATP-binding cassettetransporters are targets for the development of antibacterial vaccinesand therapies. Infect Immun 72, 6757–6763.

Gilmore, M. S., Segarra, R. A. & Booth, M. C. (1990). An HlyB-typefunction is required for expression of the Enterococcus faecalishemolysin/bacteriocin. Infect Immun 58, 3914–3923.

Hedstrom, L. (2009). IMP dehydrogenase: structure, mechanism, andinhibition. Chem Rev 109, 2903–2928.

Holo, H. & Nes, I. F. (1989). High-frequency transformation, byelectroporation, of Lactococcus lactis subsp. cremoris grown with glycinein osmotically stabilized media. Appl Environ Microbiol 55, 3119–3123.

Horton, R. M., Cai, Z. L., Ho, S. N. & Pease, L. R. (1990). Gene splicingby overlap extension: tailor-made genes using the polymerase chainreaction. Biotechniques 8, 528–535.

Jenkins, A., Cote, C., Twenhafel, N., Merkel, T., Bozue, J. & Welkos,S. (2011). Role of purine biosynthesis in Bacillus anthracispathogenesis and virulence. Infect Immun 79, 153–166.

Jonsson, M., Saleihan, Z., Nes, I. F. & Holo, H. (2009). Construction andcharacterization of three lactate dehydrogenase-negative Enterococcusfaecalis V583 mutants. Appl Environ Microbiol 75, 4901–4903.

Khan, H., Flint, S. H. & Yu, P. L. (2013). Development of a chemicallydefined medium for the production of enterolysin A fromEnterococcus faecalis B9510. J Appl Microbiol 114, 1092–1102.

Fig. 5. Sequence alignment of the nucleoside-binding proteins shown in Fig. 1. The alignment was generated using CLUSTAL

Omega (Sievers et al., 2011) using the sequences without their signal peptides. Fully, highly and less conserved residues areindicated by an asterisk, a colon and a dot placed under the sequences, respectively. Residues interacting with the nucleoside(as shown in Fig. 4) are coloured blue for residues interacting with the ribose moiety and red for residues interacting with thebase moiety. Gly212 has a potentially mixed role and is coloured purple. Interacting residues that vary in a manner that mostlikely affects ligand specificity are placed in a shaded box. Interacting residues that vary in a manner that most likely does notaffect interactions with the ligand are Asn37 and Met158 (in PnrA). See text for more details.




IP: 23.22.50.124

On: Mon, 18 Apr 2016 09:05:06

Kilstrup, M., Hammer, K., Ruhdal Jensen, P. & Martinussen, J.(2005). Nucleotide metabolism and its control in lactic acid bacteria.FEMS Microbiol Rev 29, 555–590.

Lebreton, F., Willems, R. J. L. & Gilmore, M. S. (2014). Enterococcusdiversity, origins in nature, and gut colonization. In Enterococci: FromCommensals to Leading Causes of Drug Resistant Infection. Edited byM. S. Gilmore, D. B. Clewell, Y. Ike & N. Shankar. Boston: MassachusettsEye and Ear Infirmary.

Lee, E. W., Huda, M. N., Kuroda, T., Mizushima, T. & Tsuchiya, T.(2003). EfrAB, an ABC multidrug efflux pump in Enterococcusfaecalis. Antimicrob Agents Chemother 47, 3733–3738.

Low, Y. L., Jakubovics, N. S., Flatman, J. C., Jenkinson, H. F. & Smith,A. W. (2003). Manganese-dependent regulation of the endocarditis-associated virulence factor EfaA of Enterococcus faecalis. J MedMicrobiol 52, 113–119.

Lubelski, J., Konings, W. N. & Driessen, A. J. (2007). Distribution andphysiology of ABC-type transporters contributing to multidrugresistance in bacteria. Microbiol Mol Biol Rev 71, 463–476.

Manson, J. M., Keis, S., Smith, J. M. & Cook, G. M. (2004). Acquiredbacitracin resistance in Enterococcus faecalis is mediated by an ABCtransporter and a novel regulatory protein, BcrR. Antimicrob AgentsChemother 48, 3743–3748.

Martinussen, J., Andersen, P. S. & Hammer, K. (1994). Nucleotidemetabolism in Lactococcus lactis: salvage pathways of exogenouspyrimidines. J Bacteriol 176, 1514–1516.

Martinussen, J., Sørensen, C., Jendresen, C. B. & Kilstrup, M. (2010).Two nucleoside transporters in Lactococcus lactis with differentsubstrate specificities. Microbiology 156, 3148–3157.

Murray, B. E., Singh, K. V., Ross, R. P., Heath, J. D., Dunny, G. M. &Weinstock, G. M. (1993). Generation of restriction map of Enterococcusfaecalis OG1 and investigation of growth requirements and regionsencoding biosynthetic function. J Bacteriol 175, 5216–5223.

Rigottier-Gois, L., Alberti, A., Houel, A., Taly, J. F., Palcy, P., Manson,J., Pinto, D., Matos, R. C., Carrilero, L. & other authors (2011).

Large-scale screening of a targeted Enterococcus faecalis mutant library

identifies envelope fitness factors. PLoS ONE 6, e29023.

Sahm, D. F., Kissinger, J., Gilmore, M. S., Murray, P. R., Mulder, R.,Solliday, J. & Clarke, B. (1989). In vitro susceptibility studies of

vancomycin-resistant Enterococcus faecalis. Antimicrob Agents Chemother

33, 1588–1591.

Saier, M. H., Jr (2000). A functional-phylogenetic classification system

for transmembrane solute transporters. Microbiol Mol Biol Rev 64,

354–411.

Samant, S., Lee, H., Ghassemi, M., Chen, J., Cook, J. L., Mankin, A. S.& Neyfakh, A. A. (2008). Nucleotide biosynthesis is critical for growth

of bacteria in human blood. PLoS Pathog 4, e37.

Sievers, F., Wilm, A., Dineen, D., Gibson, T. J., Karplus, K., Li, W.,Lopez, R., McWilliam, H., Remmert, M. & other authors (2011). Fast,

scalable generation of high-quality protein multiple sequence

alignments using CLUSTAL Omega. Mol Syst Biol 7, 539.

Singh, K. V., Weinstock, G. M. & Murray, B. E. (2002). An Enterococcus

faecalis ABC homologue (Lsa) is required for the resistance of this

species to clindamycin and quinupristin-dalfopristin. Antimicrob

Agents Chemother 46, 1845–1850.

Sun, R. & Wang, L. (2013). Inhibition of Mycoplasma pneumoniae

growth by FDA-approved anticancer and antiviral nucleoside and

nucleobase analogs. BMC Microbiol 13, 184.

Webb, A. J. & Hosie, A. H. (2006). A member of the second

carbohydrate uptake subfamily of ATP-binding cassette transporters

is responsible for ribonucleoside uptake in Streptococcus mutans.

J Bacteriol 188, 8005–8012.

Xue, X., Sztajer, H., Buddruhs, N., Petersen, J., Rohde, M., Talay,S. R. & Wagner-Dobler, I. (2011). Lack of the delta subunit of RNA

polymerase increases virulence related traits of Streptococcus mutans.

PLoS ONE 6, e20075.

Edited by: J. Kok



EF0176 and EF0177 from Enterococcus faecalis V583 are substrate-binding lipoproteins involved in...

Documents

Transcript of EF0176 and EF0177 from Enterococcus faecalis V583 are substrate-binding lipoproteins involved in...