Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Sandrini, M. P. B. Articles by Piskur, J. Search for Related Content PubMed PubMed Citation Articles by Sandrini, M. P. B. Articles by Piskur, J.

 Previous Article  |  Next Article 

Antimicrobial Agents and Chemotherapy, August 2007, p. 2726-2732, Vol. 51, No. 8
0066-4804/07/$08.00+0     doi:10.1128/AAC.00081-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Deoxyribonucleoside Kinases Activate Nucleoside Antibiotics in Severely Pathogenic Bacteria

Michael P. B. Sandrini,^1,2* Oonagh Shannon,³ Anders R. Clausen,² Lars Björck,³ and Jure Pikur²

BioCentrum-DTU, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark,¹ Cell and Organism Biology, Lund University, Sölvegatan 35, SE-22362 Lund, Sweden,² Section for Clinical and Experimental Infection Medicine, BMC, B14, Lund University, Tornavägen 10, SE-22184 Lund, Sweden³

Received 19 January 2007/ Returned for modification 6 March 2007/ Accepted 21 May 2007

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Common bacterial pathogens are becoming progressively more resistant to traditional antibiotics, representing a major public-health crisis. Therefore, there is a need for a variety of antibiotics with alternative modes of action. In our study, several nucleoside analogs were tested against pathogenic staphylococci and streptococci. We show that pyrimidine-based nucleoside analogs, like 3'-azido-3'-deoxythymidine (AZT) and 2',2'-difluoro-2'deoxycytidine (gemcitabine), are specifically activated by the endogenous bacterial deoxyribonucleoside kinases, leading to cell death. Deoxyribonucleoside kinase-deficient Escherichia coli strains become highly susceptible to nucleoside analogs when they express recombinant kinases from Staphylococcus aureus or Streptococcus pyogenes. We further demonstrate that recombinant S. aureus deoxyadenosine kinase efficiently phosphorylates the anticancer drug gemcitabine in vitro and is therefore the key enzyme in the activation pathway. When adult mice were infected intraperitoneally with a fatal dose of S. pyogenes strain AP1 and afterwards received gemcitabine, they failed to develop a systemic infection. Nucleoside analogs may therefore represent a promising alternative for combating pathogenic bacteria.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Antimicrobial resistance has been emerging as a major public-health crisis (28). Staphylococcus aureus virulence determinants together with antibiotic resistance contribute to its emergence as a leading pathogen in both nosocomial and community settings (16, 29). Among the streptococci, Streptococcus pneumoniae and Streptococcus pyogenes are the most significant pathogens. However, only a few classes of novel antibiotics have been introduced in the past 40 years. There is therefore a need for antibiotics with new modes of action to combat the so-called "super-bugs" (8).

Nucleoside analogs mimic the building blocks of genetic material. From the medical point of view, they comprise a major group of chemotherapeutic antiviral and anticancer prodrugs. They serve as potent inhibitors of viral replication (18) and are successful antiproliferatives in the battle against various cancers (22). The antiretroviral nucleoside analog AZT (Retrovir; 3'-azido-3'-deoxythymidine) is a prodrug used in anti-human immunodeficiency virus treatment. In its triphosphorylated form, AZT inhibits human immunodeficiency virus reverse transcriptase and blocks viral replication (18). Gemcitabine (dFdC; 2',2'-di-fluoro-2'-deoxycytidine) is a very successful anticancer prodrug. This drug causes both inhibition of ribonucleotide reductase and termination of DNA synthesis (22). In the cell, analogs of deoxyribonucleosides need to be activated by an endogenous deoxyribonucleoside kinase (dNK) to exert their effect.

The dNKs are the key enzymes in the salvage of deoxyribonucleosides (dNs), as deoxyadenosine (dAdo), deoxycytidine (dCyd), deoxyguanosine (dGuo), and thymidine (Thd). The first and committed step in the salvage of dNs is the phosphorylation of the dN by dNKs to the corresponding deoxyribonucleoside monophosphate. The monophosphates are further phosphorylated in two steps to deoxyribonucleoside triphosphates, the terminal precursors for DNA synthesis (11). Humans have four different dNKs, which phosphorylate the native dNs with overlapping specificities (for a review see references 3 and 11). However, the number of kinases and their substrate specificities vary in other eukaryotes (21). Not much is known about the diversity and properties of prokaryotic dNKs, and the few bacteria which have been characterized for their capacity to salvage dNs have been either model organisms, nonpathogenic bacteria, or mollicutes (2, 5, 6, 9, 13, 17, 27). So far, not one gram-positive bacterium has a complete set of well-characterized dNKs.

Herein we show that pathogenic staphylococci and streptococci are sensitive to pyrimidine-based dN analogs. Gemcitabine, in particular, is efficiently activated in vitro and also in vivo by a specific bacterial deoxyadenosine kinase (dAK). Thus, nucleoside analogs represent an efficient additional tool for combating staphylococcal and streptococcal infection.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Strains. Clinical isolates and type strains were from the collection of the Danish Institute for Food and Veterinary Research and are listed in Table 1. S. pyogenes AP1 is the 40/58 S. pyogenes strain from the WHO Collaborating Centre for References and Research on Streptococci, Institute of Hygiene and Epidemiology, Prague, Czech Republic. The S. aureus strain used for the viable count susceptibility assay was from the American Type Culture Collection (ATCC 29213). For heterologous expression, Escherichia coli BL21 was used, and for susceptibility testing of heterologous dNKs in E. coli, we used the dNK-deficient strain KY895 (F^– tdk-1 ilv) (12).

View this table: [in this window] [in a new window]

TABLE 1. Susceptibility of bacterial isolates and type strains towards nucleoside analogsa

MIC determination. Gemcitabine was from Thykn (India) International (Mumbai, India). The remaining nucleoside analogs used in the susceptibility tests were from Sigma.

Bacterial isolates were grown overnight on Columbia blood agar plates from SSI Diagnostika (Hilleroed, Denmark), harvested from the plates, diluted to McFarland 0.5 in 0.9% NaCl, diluted additionally 10 times, and spotted on Mueller-Hinton agar (Oxoid) containing various nucleoside analog concentrations. Plates were incubated for 20 h at 37°C, and the MIC was determined by visually inspecting the plates for growth. The MIC is defined as the lowest concentration of nucleoside analog which completely inhibits the growth of bacteria.

Viable count susceptibility assay. S. aureus ATCC 29213 and S. pyogenes AP1 were grown overnight in Todd-Hewitt broth in the presence of 5% CO₂ at 37°C. This represents cells in the stationary phase of growth. An aliquot of these cells was added to fresh medium and grown to the exponential phase (optical density at 620 nm [OD₆₂₀] = 0.4). The cells were pelleted, washed twice, and diluted 1,000-fold in wash buffer (10 mM Tris-HCl [pH 7.5], 5 mM glucose) supplemented with 10% Todd-Hewitt broth. A 10-µl portion of bacterial cells was incubated with 40 µl wash buffer containing gemcitabine (0 to 200 µg/ml) for 3 h at 37°C in 5% CO₂. The incubation was stopped by the addition of 450 µl ice-cold wash buffer, and duplicate samples were plated onto Todd-Hewitt agar plates for viable count determination.

Cloning of dNKs. Genomic DNA from S. aureus CCM 885 was isolated by using a Easy-DNA isolation kit (Invitrogen) according to the manufacturer's instructions. The genomic DNA from S. pyogenes AP1 was isolated by using a Fast Prep FP120 (Bio 101, Savant) according to the manufacturer's instructions.

Putative dNKs from S. aureus and S. pyogenes were identified using the Basic Local Sequence Alignment Tool (BLAST) (1). We searched the S. aureus Mu50 and S. pyogenes M1 group A streptococcus genomes using human TK1, E. coli TK, and B. subtilis dAK and dGK as the query. Open reading frames corresponding to the identified putative kinases were amplified from the genomic DNA by PCR using specific primers (DNA-Technology, Aarhus, Denmark) with overhangs for cloning into the pGEX-2T vector (GE Healthcare). The S. pyogenes TK start codon was TTG in the GenBank sequence, but we changed it to ATG to maintain the methionine in this position. All open reading frames were inserted in the BamHI/EcoRI sites of the pGEX-2T vector, except for SaTK, which was inserted into the BamHI site. All clonings were verified by sequencing using a commercial service (MWG-Biotech). The pGEX-2T vector produces an N-terminal glutathione S-transferase (GST) fusion.

Phylogenetic analysis. Multiple alignments of the amino acid sequences were done in ClustalX v. 1.81 using the default settings (25). Subsequently, the phylogeny was reconstructed with TreeCon v. 1.3b (26) by the neighbor-joining method, with Herpes simplex virus TK as the outgroup.

Susceptibility of E. coli harboring heterologous kinases. Overnight cultures of KY895 freshly transformed with the pGEX-2T vector constructs described above were diluted 200 times in 10% glycerol and spotted sequentially on M9 plates (14) supplemented with 100 µg/ml ampicillin, 0.2% glucose, 5 mg/liter thiamine, 1 g/liter Casamino Acids, and various nucleoside analog concentrations. Susceptibility was visually determined, and the MIC was determined as the concentration where growth was no longer visible.

Expression and purification of dNKs. All cultivations were in LB medium containing 100 µg/ml ampicillin. Overnight cultures of E. coli BL21 transformed with SaTK, SadAK, or SadGK in the pGEX-2T vector were used to start the expression cultures. Cultures were grown at 37°C to an OD₆₀₀ of 0.6 and induced with IPTG (isopropyl-ß-D-thiogalactopyranoside) to a final concentration of 100 µM. Upon induction, the temperature was immediately lowered to 25°C, and induction was maintained for 6 h before harvest of the cells by centrifugation. The pellet was lysed in 25 column volumes of 1x PBS (140 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄ [pH 7.3]), 5 mM EDTA, 5 mM dithiothreitol (DTT), 150 µg/ml lysozyme, 10% glycerol, 0.1% Triton X-100, 0.1 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine, 50 mM aminocaproic acid. Filtered cell homogenate was applied to a glutathione-Sepharose column pre-equilibrated with buffer A (1x PBS, 1 mM DTT, 10% glycerol, 0.1% Triton X-100, 0.1 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine, 50 mM aminocaproic acid). Unbound material was removed by washing with 50 column volumes of buffer A. Strongly bound contaminating proteins were removed by recirculating for 16 h at room temperature with 2 column volumes of buffer A containing 10 mM ATP and 10 mM MgCl₂. The column matrix was then equilibrated with 10 column volumes of 1x PBS. The expressed protein was cleaved from GST by recirculation for 16 h at room temperature with 2 column volumes of 1x PBS containing 500 U of thrombin and 0.1% Triton X-100. The eluate was collected, and the matrix was washed twice with 2 column volumes of 1x PBS containing 0.1% Triton X-100. Finally, the remaining GST and uncleaved fusion protein were eluted with 3 column volumes of 50 mM Tris-HCl (pH 7.5) containing 10 mM reduced glutathione. Enzyme-containing fractions were stored at –80°C, in the presence of 10% glycerol, 5 mM MgCl₂, and 5 mM DTT. Purification was verified by denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (15), and the protein concentration was determined by the Bradford method using a bovine serum albumin standard (7).

Kinetic measurements. dNK activities were determined by initial velocity measurements based on four time samples by the DE-81 (Whatman Inc.) filter paper assay using various tritium-labeled deoxyribonucleoside substrate concentrations (20). All radiolabeled substrates were from Moravek Biochemicals Inc. except thymidine, which was from GE Healthcare. Standard assay conditions were 50 mM Tris-HCl (pH 8.0), 10 mM DTT, 2.5 mM ATP, 2.5 mM MgCl₂, 3 mg/ml bovine serum albumin, 0.5 mM CHAPS (3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate).

One unit of dNK activity is defined as 1 µmol of the corresponding monophosphate product formed per min.

Analysis of kinetic data. Kinetic data were evaluated by nonlinear regression analysis using the Michaelis-Menten equation v = V_max x [S]/(K_m + [S]), where v is the reaction velocity and [S] is the substrate concentration; whenever substrate inhibition was evident, nonlinear regression was applied using the equation for substrate inhibition v = (V_max x [S])/(K_m + [S] + ([S]²/K_is)), where K_is is the substrate inhibition constant.

Infection models. Adult female BALB/c mice weighing approximately 20 g were housed under standard conditions of light and temperature and were fed standard laboratory chow and water ad libitum. The local ethical committee approved all procedures. S. pyogenes AP1 was grown in Todd-Hewitt broth at 37°C in the presence of 5% CO₂. Cells were harvested at the mid-exponential phase of growth (OD₆₂₀ = 0.5). The cells were washed once and resuspended to approximately 1 x 10⁷/ml in 1x PBS. The number of bacteria given to each animal per experiment was verified by colony counts on blood agar. The animals received an intraperitoneal injection of 2.5 x 10⁶ CFU per animal. Six hours later, one group received an intraperitoneal injection of 100 µl 1x PBS and the other group received 100 µg gemcitabine. Twenty-four hours after the initial injection of bacteria, all of the animals were sacrificed, and the spleens were harvested on ice. The spleens were homogenized, and a viable count was performed to evaluate bacterial spread.

Fatal bacterial infection was tested with a similar model, but a higher bacterial dose, 1 x 10⁷ CFU per animal, was used. Survival of the animals was monitored at regular intervals. Animals still alive after 5 days were sacrificed according to ethical requirements.

Nucleotide sequence accession numbers. Nucleotide sequences were deposited in GenBank with the following accession numbers: SaTK, DQ384605; SadAK, DQ384604; SadGK, DQ384606; SpTK, EF061224; SpdAK; EF061223.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Nucleoside analogs kill streptococci and staphylococci. The present investigation was initiated with a susceptibility experiment to determine whether streptococci and staphylococci are susceptible to commonly used antiviral and anticancer nucleoside analogs. Six clinical isolates and type strains representing staphylococci, streptococci, and E. coli were tested by spotting cell suspensions of the strains on Mueller-Hinton agar plates containing dilution series of the clinically important nucleoside analogs AZT, 5-fluoro-2'-deoxyuridine (FdUrd), 2-chloro-deoxyadenosine (CdA), 9-beta-D-arabinofuranosyl-2-fluoroadenine (F-AraA), D-arabinosyl adenine (AraA), D-arabinosyl cytosine (AraC), 5-ethyl-2'-deoxyuridine (EdUrd), and gemcitabine (dFdC). The susceptibility was reported as the lowest concentration of nucleoside analog that completely inhibited visible growth on the plates (Table 1). The arabinofuranosyl analogs AraA, F-AraA, and AraC, like EdUrd and CdA, did not have any noticeable effect on the tested strains, and AZT was efficacious only against E. coli. The MIC of AZT on E. coli was in the concentration range of 10 µM to 31.6 µM. The two fluorinated pyrimidine-based nucleoside analogs, FdUrd and gemcitabine, were efficacious against streptococci and S. aureus. The FdUrd susceptibility varied from 0.1 µM to 10 µM in streptococci and from 0.00316 µM to 0.01 µM in S. aureus. Gemcitabine did not have any notable effect on the tested E. coli control strain (ATCC 25922), even at 1,000 µM. However, the streptococci and the S. aureus strain were highly susceptible and were killed by gemcitabine concentrations in the ranges of 0.01 µM to 0.1 µM and 0.1 µM to 1.0 µM, respectively. For comparison and confirmation of the streptococcus results above, we also tested the susceptibility of S. pyogenes AP1 to gemcitabine. The MIC of gemcitabine for S. pyogenes AP1 was 0.1 µM (Table 1). Thus, gemcitabine in particular showed promising antimicrobial efficacy against S. aureus and S. pyogenes.

Growth is a prerequisite for the efficacy of gemcitabine. The effect of growth phase on susceptibility of both S. aureus and S. pyogenes was pronounced. Viable count assays of the two strains revealed that during the exponential growth phase the cells were highly susceptible to low gemcitabine concentrations (Fig. 1). Effective killing of exponentially growing S. aureus was observed at 0.002 µg/ml gemcitabine, whereas in S. pyogenes the same effective killing was seen at 0.2 µg/ml gemcitabine. When the cells were in stationary growth phase, however, the susceptibility was profoundly diminished (Fig. 1). The killing curves clearly illustrate the efficacy of gemcitabine on these two potential pathogens and show that growth of the cells is a prerequisite for optimal killing conditions.

View larger version (12K): [in this window] [in a new window]

FIG. 1. Gemcitabine is an efficient antibacterial compound. A viable count antibacterial assay was performed for S. aureus ATCC 29213 (A) and S. pyogenes AP1 (B) in the stationary or exponential phase of growth. Gemcitabine effectively kills both bacteria when they are in exponential phase.

Three dNKs in S. aureus and two in S. pyogenes. In humans, activation of gemcitabine by the endogenous deoxycytidine kinase is required as the first step in the salvage pathway leading to the inhibitory compounds gemcitabine di- and triphosphate (22). A database search for putative dNKs encoded by the S. aureus Mu50 and S. pyogenes M1 group A streptococcus genomes revealed three dNK genes in S. aureus and two in S. pyogenes. Subsequently, we cloned the five predicted dNK genes from genomic DNA of S. aureus CCM 885 and S. pyogenes AP1 into the pGEX-2T expression vector.

In both organisms, one of the encoded kinases was homologous to human TK1. The TKs of S. aureus and S. pyogenes were named SaTK and SpTK, respectively. Phylogenetic analysis revealed that the cloned SaTK was closer to the human TK1, whereas SpTK was more closely related to E. coli TK (Fig. 2B). The remaining three proteins were homologous to Bacillus subtilis dAK and dGK. S. aureus and S. pyogenes both contained a protein that was phylogenetically closer to B. subtilis dAK (Fig. 2B), and for this reason they were named SadAK and SpdAK, respectively. The third kinase from S. aureus was phylogenetically closer to B. subtilis dGK and consequently named SadGK (Fig. 2B).

View larger version (22K): [in this window] [in a new window]

FIG. 2. Phylogeny of the cloned dNKs. Three and two putative dNK genes were found in the S. aureus and S. pyogenes genomes, respectively. The encoded amino acid sequences were analyzed for their phylogenetic relationship with other dNKs. The phylogenetic relationships of TKs (A) and non-TKs (B) are shown. Both trees were reconstructed by the neighbor-joining method. The GenBank accession numbers are in brackets, and bootstrap values from 100 replications are given at the nodes. Apparently, S. aureus contains a gram-positive-type TK and two bacterial-type dAKs and dGKs, whereas S. pyogenes contains a gram-negative-type TK and one bacterial-type dAK.

E. coli is susceptible to gemcitabine when expressing dAKs. To determine whether any of the cloned kinases conferred the activation of either AZT or gemcitabine, the dNK-deficient strain E. coli KY895 was transformed with the expression vector harboring SaTK, SpTK, SadAK, SpdAK, or SadGK. Subsequently, we determined the MICs of AZT and gemcitabine for these E. coli strains (Fig. 3). AZT was efficacious against SaTK and SpTK, with MICs of 31.6 nM and 10 nM, respectively (Fig. 3A). The other dNKs did not sensitize the E. coli strain to AZT at the tested concentration range (0 to 1,000 nM AZT). Gemcitabine was efficacious against both SadAK and SpdAK, with MICs of 1 nM and 31.6 nM, respectively. The results clearly show that the TKs are responsible for the activation of AZT, whereas in S. aureus and S. pyogenes the dAKs confer the activation of gemcitabine.

View larger version (32K): [in this window] [in a new window]

FIG. 3. Susceptibility to AZT and gemcitabine of E. coli KY895 transformed with S. pyogenes and S. aureus dNKs. (A) When transformed with SaTK and SpTK, KY895 shows impaired growth at 3.16 nM AZT and complete inhibition at 10.0 nM. AZT had no visible effect on the cells transformed with SadAK, SadGK, or SpdAK. (B) When transformed with SadAK, KY895 shows impaired growth at 0.316 nM gemcitabine and complete inhibition at 1.0 nM. When transformed with SpdAK, the cells could not grow at 31.6 nM gemcitabine. When cells were transformed with SpTK, SaTK, or SadGK, no inhibition was observed.

S. aureus dAK phosphorylates gemcitabine. To study the activation kinetics on the molecular level, we purified and characterized the recombinant SaTK, SadAK, and SadGK proteins. SaTK efficiently phosphorylated Thd and AZT. The K_m for AZT was determined to be 11.6 µM, and the V_max was 4.0 U/mg (Table 2). However, SaTK showed pronounced substrate inhibition, with declining reaction velocities at AZT concentrations above 100 µM, and the K_is was determined to be 944 µM (Fig. 4). Likewise, substrate inhibition also complicated the determination of K_m and V_max for Thd and caused great variability within reaction velocities at Thd concentrations greater than 8 µM. The K_m of 2.8 µM for Thd was determined by combining the results from three experiments and fitting the data to the mathematical model for substrate inhibition.

View this table: [in this window] [in a new window]

TABLE 2. Kinetic parameters for purified SaTK, SadAK, and SadGKa

View larger version (5K): [in this window] [in a new window]

FIG. 4. Kinetics of S. aureus recombinant dNKs, showing concentration-dependent reaction velocities for SaTK with AZT (A) and SadAK with gemcitabine (B). AZT confers characteristic substrate inhibition, resulting in declining reaction velocities at concentrations above 100 µM. In comparison, SadAK showed normal Michaelis-Menten kinetics with gemcitabine as the substrate. The figure clearly illustrates that dAK phosphorylates gemcitabine.

SadAK phosphorylated dAdo and dCyd almost equally well. The K_ms for dAdo and dCyd were 29.2 µM and 34.4 µM, respectively, and the catalytic efficiency constants (k_cat/K_m) were 6.2 x 10⁴ s^–1 M^–1 and 1.6 x 10⁴ s^–1 M^–1, respectively. As expected from the recombinant E. coli susceptibility experiments, SadAK could also phosphorylate gemcitabine, albeit 16 times less efficiently (k_cat/K_m = 9.8 x 10² s^–1 M^–1) than it could dCyd. The K_m for gemcitabine was 192 µM. SadGK was extremely specific, and we could measure significant phosphorylation only of dGuo. K_m and k_cat/K_m for dGuo were 1.1 µM and 5.2 x 10⁵ s^–1 M^–1, respectively. These results clearly show that gemcitabine is phosphorylated by the dAK.

Gemcitabine protects against S. pyogenes infections in mice. A mouse model of S. pyogenes AP1 infection was developed to investigate the potential in vivo antibacterial efficacy of gemcitabine. S. pyogenes is a human-specific pathogen, and large doses of bacteria are required to mimic human infection in mice. Two groups of six mice received a large intraperitoneal dose of exponentially growing S. pyogenes AP1. After 6 h, one group was treated with 100 µg gemcitabine, whereas the control group received only PBS buffer. After 18 h, the spleens from the animals were harvested, and the bacterial load was determined. In all of the animals that received PBS, significant bacterial dissemination had occurred. These animals had bacterial loads ranging from 2 x 10³ to 3 x 10⁶ CFU/ml in the spleen, confirming the systemic spread of S. pyogenes during infection (Fig. 5). In profound contrast, bacterial dissemination failed to occur in the gemcitabine-treated animals, and none of these animals had any S. pyogenes organisms in the spleen.

View larger version (6K): [in this window] [in a new window]

FIG. 5. Gemcitabine protects animals from bacterial dissemination during systemic infection. BALB/c female mice were subjected to a systemic infection with S. pyogenes AP1. Six hours after initiation of infection, the animals were treated with either gemcitabine or PBS (control), and 24 h after initiation of infection, the bacterial load in the spleen was determined for each animal. Each dot represents an individual animal.

To further investigate the efficacy of gemcitabine during bacterial infection in vivo, a similar animal experiment was performed with an even higher dose of bacteria. Two groups of six animals received an intraperitoneal injection of a fatal dose of S. pyogenes AP1. Six hours later, the control group was treated with PBS and the test group received 100 µg gemcitabine. Within 30 h of initial infection, all six mice in the control group had died (Fig. 6). In contrast, 5 days after infection, 83% (5/6) of the mice treated with gemcitabine were still alive and healthy. The results clearly show that gemcitabine can prevent bacterial dissemination and death caused by S. pyogenes infection.

View larger version (9K): [in this window] [in a new window]

FIG. 6. Kaplan-Meier survival curve for BALB/c mice after intraperitoneal infection with S. pyogenes AP1. Female BALB/c mice were subjected to a systemic infection with S. pyogenes AP1. Six hours after initiation of infection, the animals were treated with either gemcitabine or PBS (control), and the survival within the group was monitored. All six (100%) animals in the untreated group died of streptococcal infection within 30 h of infection. Only one of six (17%) animals treated with gemcitabine died as a result of streptococcal infection.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Pathogenic bacteria are rapidly acquiring resistance to almost all antibiotics employed today. As a result, bacterial infections remain a significant problem in hospital and community environments. The situation with gram-positive bacteria, such as staphylococci and streptococci, is worrisome (23, 28). Although the scientific community has been engaged in the discovery of new antibiotics, there has been only limited success since the early 1960s (8, 28). In this study, we examined nucleoside analogs, used in the treatment of various cancers and viral infections, for their potential against staphylococci and streptococci.

Nucleoside analogs mimic the building blocks of the genetic material and can impair the metabolism and structure of nucleic acids. In order to be effective, they must first be successfully transported into the cell and then, upon activation, must interact with a target. Eight different nucleoside analogs were tested for bactericidal activity against an E. coli control strain and six gram-positive clinical isolates. The cytidine analog gemcitabine killed all six gram-positive isolates at very low concentrations, while the other seven analogs required higher concentrations or were not successful at all (Table 1).

Two clinically important gram-positive pathogens were then tested in a more detailed experiment for their growth in the presence of gemcitabine (Fig. 1). Our results clearly show that gemcitabine is very effective at killing S. aureus and S. pyogenes, at concentrations lower than 0.002 µg/ml and 0.2 µg/ml, respectively.

Gemcitabine is expected to first be activated in the cell and then act on its targets. We therefore analyzed the two bacteria for their potential to activate gemcitabine through phosphorylation. The two sequenced genomes were analyzed for putative dNK genes, and three positive hits in S. aureus and two in S. pyogenes were obtained. Alignment and phylogenetic analysis (Fig. 2) of the encoded proteins revealed that in addition to a TK, S. aureus and S. pyogenes also contain an enzyme that is closely related to B. subtilis dAK. The difference in the dNK coding potential between S. aureus and S. pyogenes is that in S. aureus, a third dNK, closely related to B. subtilis dGK, is also found. However, these enzymes may vary a great deal in their substrate specificities; therefore, the five genes were subcloned and tested in E. coli for their ability to activate nucleoside analogs in vivo (Fig. 3). When the SaTK and SpTK genes were expressed in a kinase-deficient strain, the susceptibility of the host to AZT increased dramatically. AZT is, however, known to have a greatly reduced antimicrobial efficacy against gram-positive bacteria (Table 1) (10). In contrast, when the SadAK gene was expressed in the kinase-deficient E. coli strain, the susceptibility of the host to gemcitabine increased tremendously. Relatively low concentrations of gemcitabine (approximately 1 nM) could kill the host bacterium. Also, the SpdAK gene increased the susceptibility, and the host was killed at 31.6 nM gemcitabine. These experiments suggest that dAK is the key enzyme in the activation of gemcitabine.

To elucidate the activation kinetics and demonstrate on a molecular level that dAK activates gemcitabine, the three kinases from S. aureus were expressed as recombinant proteins and studied for their substrate specificities and kinetic parameters. SaTK phosphorylated Thd and AZT, as was expected on the basis of the in vivo E. coli experiments. The phosphorylation of both Thd and AZT was prone to substrate inhibition (Fig. 4) at high substrate concentrations (K_is,Thd = 257 µM and K_is,AZT = 944 µM), and the K_ms for Thd and AZT were 2.8 µM and 11.6 µM, respectively. An earlier study of AZT activation by E. coli TK reported a K_m of 21 µM for AZT (10), and kinetic parameters more similar to those reported here for SaTK were recently reported for Ureaplasma parvum TK, which also showed substrate inhibition for both Thd and AZT (9).

The recombinant SadAK phosphorylated both dAdo and dCyd equivalently (Table 2). In addition, SadAK clearly had the potential to effectively phosphorylate gemcitabine into its monophosphate, with a k_cat/K_m of approximately 1 x 10³ s^–1 M^–1 (Table 2; Fig. 4), thus representing a reasonably efficient kinase -substrate pair (19). The recombinant SadGK had practically the same specificity as recombinant B. subtilis dGK. Both enzymes have extremely narrow substrate specificities, and only the phosphorylation of dGuo is measurable. The efficiencies of SadGK and B. subtilis dGK are 5.17 x 10⁵ s^–1 M^–1 and 2.4 x 10⁵ s^–1 M^–1 (2), respectively. The close resemblance of kinetic values and narrow substrate specificities for SadGK and B. subtilis dGK are consistent with their close phylogenetic relationship (Fig. 2). The results of our kinetic studies, in combination with the recombinant E. coli susceptibility data, confirm that the activation-dependent killing of S. aureus and S. pyogenes by gemcitabine is attributable to dAK.

Could nucleosides be used as antibiotics? To address the question of whether nucleosides could indeed be used as antibiotics, we used a murine model of S. pyogenes infection. It is important to point out that S. pyogenes is a human-specific pathogen and that large doses of bacteria are required to mimic infection in murine models. S. pyogenes AP1 is a particularly virulent strain and is the leading cause of invasive and severe S. pyogenes infection (24). Administration of the S. pyogenes AP1 strain to the mouse peritoneum has previously been shown to result in bacterial spread from the site of inoculation (4). This develops into a systemic infection with bacteremia and/or sepsis, which can be fatal. In our first infection model, we used a nonfatal dose of S. pyogenes in order to compare the survival and spread of the bacteria from the site of infection. Bacterial spread was compared between two groups by harvesting the spleens of the animals postmortem and determining the bacterial load. The animals which received gemcitabine after the bacterial infection had no bacteria in their spleens (Fig. 5). This indicates that a systemic infection failed to develop in these animals. Since the PBS-treated animals had significant bacterial loads in their spleens, we conclude that gemcitabine protected the treated animals from bacterial spread and systemic infection. In the second model, we used a higher dose of bacteria, which in pilot experiments resulted in 100% mortality in infected animals. Figure 6 clearly shows that treatment with gemcitabine 6 h after bacterial infection significantly improved survival rates within the group. Only one animal died in the group which received gemcitabine (Fig. 6), whereas all six animals in the untreated group died from the infection. Five of the six treated animals survived and completely recovered from infection with no adverse effects. Gemcitabine is therefore a potent antibacterial agent in vivo against S. pyogenes. It is likely that gemcitabine would be equally effective against S. aureus and other dAK-containing bacteria.

One cannot exclude the possibility that bacteria would develop resistance against nucleoside drugs. A mutation in the uptake system or the dNK gene would likely impair sensitivity. Our preliminary tests showed that overnight cultures of bacteria plated on selective medium containing gemcitabine gave resistant colonies at a frequency of 10^–7. However, such clones may have a severely impaired metabolism.

We have demonstrated in this study that the nucleoside analog gemcitabine is specifically activated by dAK from gram-positive organisms and is responsible for the bactericidal effects in the mouse model of S. pyogenes infection. Therefore, gemcitabine and other nucleoside analogs have a clear antimicrobial potential. Many nucleosides, especially anticancer drugs, have a side effect, i.e., cytotoxicity in the patient. However, our approach should provide a more precise way to look for new nucleoside compounds which are preferentially toxic for microbes and not mammalian cells. We should stress here that such selective drugs already exist among nucleosides; for example, acyclovir is now successful against herpesvirus and has virtually no noticeable side effects.

 
We thank F. M. Aarestrup and S. L. W. On for providing the S. aureus and Streptococcus isolates from the Danish Institute for Food and Veterinary Research collection.

The work presented herein was supported by a grant from the Danish Technical Research Foundation (STVF) and the Swedish Research Foundation (VR).

 
* Corresponding author. Present address: Department of Molecular Biology, University of Copenhagen, Universitetsparken 13, DK-2100 Copenhagen, Denmark. Phone: 45 3532 1734. Fax: 45 3532 1567. E-mail: msandrini{at}aki.ku.dk

Published ahead of print on 25 May 2007.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1
1. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403-410.[CrossRef][Medline]
2
2. Andersen, R. B., and J. Neuhard. 2001. Deoxynucleoside kinases encoded by the yaaG and yaaF genes of Bacillus subtilis. Substrate specificity and kinetic analysis of deoxyguanosine kinase with UTP as the preferred phosphate donor. J. Biol. Chem. 276:5518-5524.[Abstract/Free Full Text]
3
3. Arner, E. S. J., and S. Eriksson. 1995. Mammalian deoxyribonucleoside kinases. Pharmacol. Ther. 67:155-186.[CrossRef][Medline]
4
4. Bjorck, L., P. Akesson, M. Bohus, J. Trojnar, M. Abrahamson, I. Olafsson, and A. Grubb. 1989. Bacterial growth blocked by a synthetic peptide based on the structure of a human proteinase inhibitor. Nature 337:385-386.[CrossRef][Medline]
5
5. Black, M. E., and D. E. Hruby. 1991. Nucleotide sequence of the Escherichia coli thymidine kinase gene provides evidence for conservation of functional domains and quaternary structure. Mol. Microbiol. 5:373-379.[CrossRef][Medline]
6
6. Bockamp, E. O., R. Blasco, and E. Vinuela. 1991. Escherichia coli thymidine kinase: nucleotide sequence of the gene and relationships to other thymidine kinases. Gene 101:9-14.[CrossRef][Medline]
7
7. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254.[CrossRef][Medline]
8
8. Brown, E. D. 2006. Microbiology: antibiotic stops ‘ping-pong’ match. Nature 441:293-294.[CrossRef][Medline]
9
9. Carnrot, C., R. Wehelie, S. Eriksson, G. Bolske, and L. Wang. 2003. Molecular characterization of thymidine kinase from Ureaplasma urealyticum: nucleoside analogues as potent inhibitors of mycoplasma growth. Mol. Microbiol. 50:771-780.[CrossRef][Medline]
10
10. Elwell, L. P., R. Ferone, G. A. Freeman, J. A. Fyfe, J. A. Hill, P. H. Ray, C. A. Richards, S. C. Singer, V. B. Knick, and J. L. Rideout. 1987. Antibacterial activity and mechanism of action of 3'-azido-3'-deoxythymidine (BW A509U). Antimicrob. Agents Chemother. 31:274-280.[Abstract/Free Full Text]
11
11. Eriksson, S., B. Munch-Petersen, K. Johansson, and H. Eklund. 2002. Structure and function of cellular deoxyribonucleoside kinases. Cell Mol. Life Sci. 59:1327-1346.[CrossRef][Medline]
12
12. Igarashi, K., S. Hiraga, and T. Yura. 1967. A deoxythymidine kinase deficient mutant of Escherichia coli. II. Mapping and transduction studies with phage phi 80. Genetics 57:643-654.[Free Full Text]
13
13. Ives, D. H., and S. Ikeda. 1998. Life on the salvage path: the deoxynucleoside kinase of Lactobacillus acidophilus R-26. Prog. Nucleic Acid Res. Mol. Biol. 59:205-255.[Medline]
14
14. Knecht, W., B. Munch-Petersen, and J. Piskur. 2000. Identification of residues involved in the specificity and regulation of the highly efficient multisubstrate deoxyribonucleoside kinase from Drosophila melanogaster. J. Mol. Biol. 301:827-837.[CrossRef][Medline]
15
15. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685.[CrossRef][Medline]
16
16. Lowy, F. D. 2003. Antimicrobial resistance: the example of Staphylococcus aureus. J. Clin. Investig. 111:1265-1273.[CrossRef][Medline]
17
17. McNab, R. 1996. Cloning and sequence analysis of thymidine kinase from the oral bacterium Streptococcus gordonii. FEMS Microbiol. Lett. 135:103-110.[CrossRef][Medline]
18
18. Mitsuya, H., K. J. Weinhold, P. A. Furman, M. H. St Clair, S. N. Lehrman, R. C. Gallo, D. Bolognesi, D. W. Barry, and S. Broder. 1985. 3'-Azido-3'-deoxythymidine (BW A509U): an antiviral agent that inhibits the infectivity and cytopathic effect of human T-lymphotropic virus type III/lymphadenopathy-associated virus in vitro. Proc. Natl. Acad. Sci. USA 82:7096-7100.[Abstract/Free Full Text]
19
19. Munch-Petersen, B., and J. Piskur. 2006. Deoxynucleoside kinases and their potential role in deoxynucleoside cytotoxicity, p. 53-79. In G. J. Peters (ed.), Cancer drug discovery and development: deoxynucleoside analogs in cancer therapy. Humana Press Inc., Totowa, NJ.
20
20. Munch-Petersen, B., L. Cloos, G. Tyrsted, and S. Eriksson. 1991. Diverging substrate specificity of pure human thymidine kinases 1 and 2 against antiviral dideoxynucleosides. J. Biol. Chem. 266:9032-9038.[Abstract/Free Full Text]
21
21. Piskur, J., M. P. Sandrini, W. Knecht, and B. Munch-Petersen. 2004. Animal deoxyribonucleoside kinases: ‘forward’ and ‘retrograde’ evolution of their substrate specificity. FEBS Lett. 560:3-6.[CrossRef][Medline]
22
22. Plunkett, W., P. Huang, and V. Gandhi. 1995. Preclinical characteristics of gemcitabine. Anticancer Drugs 6(Suppl. 6):7-13.
23
23. Rice, L. B. 2006. Antimicrobial resistance in Gram-positive bacteria. Am. J. Med. 119:S11-S19.[Medline]
24
24. Stevens, D. L. 2001. Invasive streptococcal infections. J. Infect. Chemother. 7:69-80.[CrossRef][Medline]
25
25. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25:4876-4882.[Abstract/Free Full Text]
26
26. Van de Peer, Y., and R. De Wachter. 1994. TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Comput. Appl. Biosci. 10:569-570.[Free Full Text]
27
27. Wang, L., J. Westberg, G. Bolske, and S. Eriksson. 2001. Novel deoxynucleoside-phosphorylating enzymes in mycoplasmas: evidence for efficient utilization of deoxynucleosides. Mol. Microbiol. 42:1065-1073.[CrossRef][Medline]
28
28. Wise, R. 2004. The relentless rise of resistance? J. Antimicrob. Chemother. 54:306-310.[Abstract/Free Full Text]
29
29. Zetola, N., J. S. Francis, E. L. Nuermberger, and W. R. Bishai. 2005. Community-acquired meticillin-resistant Staphylococcus aureus: an emerging threat. Lancet Infect. Dis. 5:275-286.[CrossRef][Medline]

---

Antimicrobial Agents and Chemotherapy, August 2007, p. 2726-2732, Vol. 51, No. 8
0066-4804/07/$08.00+0     doi:10.1128/AAC.00081-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Sandrini, M. P. B. Articles by Piskur, J. Search for Related Content PubMed PubMed Citation Articles by Sandrini, M. P. B. Articles by Piskur, J.