JAC Advance Access originally published online on July 5, 2007
Journal of Antimicrobial Chemotherapy 2007 60(3):510-520; doi:10.1093/jac/dkm240
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Nucleoside analogues are activated by bacterial deoxyribonucleoside kinases in a species-specific manner
kur21 BioCentrum-DTU, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark 2 Cell and Organism Biology, Lund University, Sölvegatan 35, SE-22362 Lund, Sweden 3 Danish Institute for Food and Veterinary Research, Bülowsvej 27, DK-1790 Copenhagen, Denmark 4 Department of Science, Systems and Models, Roskilde University, DK-4000 Roskilde, Denmark
* Corresponding author. Tel: +45-3532-1734; E-mail: msandrini{at}aki.ku.dk
Received 26 March 2007; returned 27 April 2007; revised 5 June 2007; accepted 8 June 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Objectives: To investigate the bactericidal activity of antiviral and anticancer nucleoside analogues against a variety of pathogenic bacteria and characterize the activating enzymes, deoxyribonucleoside kinases (dNKs).
Methods: Several FDA-approved nucleoside analogue drugs were screened for their potential bactericidal activity against several clinical bacterial isolates and type strains. We identified and subcloned the genes coding for putative deoxyribonucleoside kinases in Escherichia coli, Pasteurella multocida, Salmonella enterica, Yersinia enterocolitica, Bacillus cereus, Clostridium perfringens and Listeria monocytogenes. These genes were tested for their ability to increase the susceptibility of a dNK-deficient E. coli strain to various analogues. We overexpressed, purified and characterized the substrate specificity and kinetic properties of the recombinant enzymes from S. enterica and B. cereus.
Results: The tested Gram-negative bacteria were susceptible to 3'-azido-3'-deoxythymidine (AZT) in the concentration range 0.032–31.6 µM except for a single E. coli isolate and two Pseudomonas aeruginosa isolates which were resistant to the tested AZT concentrations. Purified recombinant S. enterica thymidine kinase phosphorylated AZT efficiently with a Km of 73.3 µM and kcat/Km of 6.6 x 104 s–1 M–1 and is the activator of this drug in vivo. 2',2'-Difluoro-2'-deoxycytidine (gemcitabine) was a potent antibiotic against Gram-positive bacteria in the concentration range between 0.001 and 1.0 µM. The B. cereus deoxyadenosine kinase had a Km for gemcitabine of 33.5 µM and kcat/Km of 5.1 x 103 s–1 M–1 and activates gemcitabine in vivo. S. enterica and B. cereus are now amongst the first bacteria with a completely characterized set of dNK enzymes.
Conclusions: Bacterial dNKs efficiently activate nucleoside analogues in a species-specific manner. Therefore, nucleoside analogues have a potential to be employed as antibiotics in the fight against emerging multiresistant bacteria.
Keywords: antibacterial , multidrug-resistant , thymidine kinase , salvage
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Multiresistant pathogenic bacteria are emerging all over the world as an increasing threat to our health.1–3 Selected by the use of antibiotics for more than half a century, the resistant pathogenic bacteria thrive in different places including hospitals, animal husbandries, pig and poultry farms where they spread to humans and cause severe illness.4,5 Vancomycin, a glycopeptide antibiotic preventing peptidoglycan synthesis in Gram-positive bacteria, was for long thought of as a last resort antibiotic to treat Gram-positive bacterial infections. However, multiresistant bacteria have emerged during recent years, including glycopeptide- and this methicillin-resistant Staphylococcus aureus, multiresistants Pseudomonas aeruginosa and cephalosporin-resistant Gram-negative bacilli leaving the clinician with few therapeutic alternatives and this emphasizes the urgent need for new ways of combating these multiresistant pathogens.4,6–8 In particular, immunocompromised individuals suffer severely from infections by multiresistant pathogenic bacteria, often with fatal outcome since ordinary antibiotic treatment is inadequate.9 Therefore, there is a need to develop new antibiotics for the future.
Nucleoside analogues comprise a major group of chemotherapeutic prodrugs. In their activated form, they serve as potent inhibitors of viral replication10 and as anti-proliferatives in the treatment of cancer.11 For many years, 5-fluorouracil (FU) and 5-fluoro-deoxyuridine (FdUrd) have been the anticancer drugs of choice when treating colorectal cancers and simultaneously FU and FdUrd have served as invaluable antibacterial agents in the elucidation of bacterial DNA metabolism.12,13 The antimetabolite action of FdUrd is achieved mainly through inhibition of thymidylate synthase by fluoro-deoxyuridine monophosphate (FdUMP).14 2',2'-Difluoro-2'-deoxycytidine (gemcitabine) is also a very successful anticancer prodrug. Gemcitabine causes both inhibition of ribonucleotide reductase and termination of DNA synthesis.11 The antiretroviral nucleoside analogue 3'-azido-3'-deoxythymidine (AZT; zidovudine) is a prodrug used in the treatment of human immunodeficiency virus (HIV) infections. In its tri-phosphorylated form, AZT inhibits HIV reverse transcriptase and thereby blocks viral replication.10 Common for all three substances is that they are prodrugs and need activation by an endogenous deoxyribonucleoside kinase (dNK) to exert their cytostatic effect.
The dNKs are the key enzymes in the salvage of deoxyribonucleosides (dNs). The first and committed step in the salvage of dNs is the phosphorylation of the dN by dNKs to the corresponding deoxyribonucleoside monophosphate (dNMP). The monophosphates are further phosphorylated in two steps to deoxyribonucleoside triphosphates (dNTPs). Through the salvage of dNs, cells supplement the de novo synthesis of deoxyribonucleic acids precursors for DNA synthesis and repair. Humans have four different dNKs [thymidine kinase 1 (TK1), deoxycytidine kinase (dCK), deoxyguanosine kinase (dGK) and thymidine kinase 2 (TK2)], which phosphorylate the native dNs [thymidine (dThd), deoxycytidine (dCyd), deoxyguanosine (dGuo) and deoxyadenosine (dAdo)] with overlapping specificities.15–17 Recently, based on similarity and structural data, dNKs have been divided into two superfamilies. The TK1-like family is composed of homologues of the human TK1 and the non-TK1-like family is composed of homologues of the human dCK, dGK and TK2.17 The number of kinases and their substrate specificities vary in other eukaryotes. Insects have only one dNK (in Drosophila melanogaster Dm-dNK) with broad specificity and the ability to phosphorylate all four native substrates with high efficiency.18 In general, each organism has kinases with very specific substrate specificity.
However, not much is known about the diversity of dNKs in bacteria, but at least three different kinds of enzymes, TK, deoxyadenosine kinase (dAK) and dGK, have been reported.19–25
Here, we present a screening of several anticancer and antiviral deoxyribonucleoside analogues for their bactericidal effect on various pathogenic bacterial isolates. In addition, we present the identification and cloning of dNKs from Bacillus cereus, Salmonella enterica, Clostridium perfringens, Listeria monocytogenes, Yersinia enterocolitica and Pasteurella multocida and we show that the activation and lethal effect of AZT and gemcitabine in these bacteria is mediated in a species-specific manner by endogenously encoded dNKs. B. cereus is the first Gram-positive bacteria for which the whole repertoire of dNKs has been characterized.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Strains
Strains tested with various analogues were from the Royal Danish Veterinary Institute (DVI) collection. Strains and identifiers are listed in Table 1. For cloning and expression of the recombinant dNKs, the TK-negative Escherichia coli strain KY895 (F–, tdk-1, ilv) was used.26
|
Susceptibility of bacterial isolates
Nucleoside analogues used in the susceptibility tests were obtained from Sigma-Aldrich (Broendby, Denmark) except for gemcitabine, which was from Thykn (India) International (Mumbay, India).
Bacterial isolates were grown overnight on Columbia blood agar plates from SSI Diagnostika (Hilleroed, Denmark), harvested from the plates and diluted to a turbidity equivalent to that of a 0.5 McFarland standard in 0.9% NaCl, diluted additionally 1:10 and spotted on Mueller–Hinton agar (Oxoid, Greve, Denmark) containing 1:3.16 dilutions series of nucleoside analogues in the range 0.1 nM to at least 100 µM as shown in Table 1.27 Plates were incubated for 20 h at 37°C and the MIC was determined by visually inspecting the plates for growth. MIC is defined as the lowest concentration of nucleoside analogue that completely inhibits the growth of bacteria.
Cultures of different bacterial strains employed in this study were harvested and genomic DNA was purified using the Easy-DNA Isolation Kit (Invitrogen, Taastrup, Denmark) using the instructions provided by the manufacturer. Gram-positive bacteria were pretreated with either lyostaphin or lysozyme.
Genes encoding putative dNKs were identified in GenBank by searching all complete genomes available at the time for homologues of human TK, E. coli TK, BsdAK, BsdGK, human dCK, human dGK, L. acidophilus dAK and dGK and Dm-dNK. Open reading frames identified by homology to known dNKs were amplified from the genomic DNA by PCR and cloned into the BamHI and EcoRI site of the commercially available expression vector pGEX-2T (GE Healthcare, Hilleroed, Denmark) using standard molecular biology techniques. The resulting constructs have a N-terminal GST fusion tag with a thrombin protease cleavage site between the GST-tag and the protein of interest. One particular TK from S. enterica was also subcloned in pASK75-8His.28 This construct provides a C-terminal 8 x histidine tag. The cloned genes were sequenced using a commercial service (MWG-Biotech, Martinsried, Germany). Sequences of the cloned genes were deposited in GenBank under the accession numbers given in Table 2.
|
Susceptibility of E. coli harbouring recombinant kinases
Overnight cultures of KY895 freshly transformed with pGEX-2T vector constructs described earlier were diluted 1 in 200 in 10% glycerol and spotted sequentially on M9 plates supplemented with 100 mg/L of ampicillin, 0.2% glucose, 5 mg/L of thiamine, 1 g/L of casamino acids and various nucleoside analogue concentrations. Susceptibility was visually inspected and MIC was noted as the concentration where growth was no longer visible (Figure 2 and Table 2).
Expression of recombinant proteins
Expression of recombinant GST-fusion or histidine-tagged dNKs was done in E. coli KY895. Freshly transformed KY895 was grown to an OD600 = 0.5–0.7 at 37°C and induced for 4–6 h at 25°C with 100 µM IPTG for pGEX-2T constructs or 200 µg/L of anhydro-tetracycline for the pASK75-8His construct. Cells were harvested by centrifugation and the pellets frozen at –80°C.
Purification of recombinant deoxyribonucleoside kinases
Pellets from pGEX-2T-fusion expressions (BcTK, BcdGK and BcdAK) were resuspended in binding buffer (PBS, pH 7.3, 10% glycerol, 0.1% Triton X-100) with EDTA-free complete inhibitor cocktail (Roche diagnostics, Hvidovre, Denmark) and disrupted by running the suspension three times through a French Press. Cell debris was removed by centrifugation at 20 000 g for 30 min and subsequently the supernatant was filtered through a 0.45 µm cellulose acetate filter. Liquid chromatography was done on AKTAprimeTM and AKTAexplorerTM 100 systems (GE Healthcare). The crude extract was loaded on a glutathione sepharose FF column and washed with binding buffer. In the purification of BcTK, 10 mM ATP/MgCl2 in binding buffer was recycled over the column for 1 h at room temperature and washed away with binding buffer. The column was then either (i) incubated for 16 h with 1 column volume (CV) binding buffer containing thrombin (50 U/mL) from GE Healthcare, or (ii) 1 CV binding buffer containing thrombin (100 U/mL) from Biofac A/S (Kastrup, Denmark) was recycled for 16 h and pure protein was eluted. Uncleaved fusion protein and bound GST-tag were eluted with elution buffer (50 mM Tris/HCl, pH 8.0, 10% glycerol, 0.1% Triton X-100, 10 mM reduced L-glutathione).
Lysate from the pASK75-8His-based expression of S. enterica thymidine kinase (SeTK) was prepared as above by French press and centrifugation, but the binding buffer was Tris/HCl, pH 7.5, 500 mM NaCl, 0.1% Triton X-100 and 10% glycerol. Following centrifugation, 10% streptomycin sulphate was added slowly to the supernatant while stirring until a final concentration of 2% was reached. After another centrifugation (20 000 g for 30 min), the supernatant was subjected to gel filtration on a G25 Sepharose column. The protein-containing fractions were pooled and loaded on a Ni–NTA sepharose column. After washing with binding buffer + 10% elution buffer (Tris/HCl, pH 7.5, 500 mM NaCl, 0.1% Triton X-100, 10% glycerol, 500 mM imidazole), pure protein was eluted with a 20 CV gradient to 100% elution buffer. Peak fractions were pooled and buffer was exchanged by gel-filtration on G25 Sepharose to 20 mM Tris/HCl, pH 7.5, 20 mM NaCl, 10% glycerol and then reapplied to the Ni–NTA column and washed with buffer without Triton X-100 (20 mM Tris/HCl, pH 7.5, 20 mM NaCl, 10% glycerol) and eluted with 20 mM Tris/HCl, pH 7.5, 20 mM NaCl, 10% glycerol, 500 mM imidazole. Finally, the pure protein was subjected to gel-filtration in 20 mM Tris/HCl, pH 7.5, 20 mM NaCl, 10% glycerol and concentrated with an Amicon centriprep cartridge (Millipore, Copenhagen, Denmark).
The purifications were followed by monitoring the A280 online and visualized by denaturing SDS–PAGE and Coomassie staining.29 Protein concentrations were determined by Bradford assay using BSA as standards30 and by measuring A280 in a Nanodrop ND-1000TM spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA).
Deoxyribonucleoside kinase activities were determined by initial velocity measurements based on four-time samples (0, 4, 8 and 12 min) by the Whatman DE-81 (Whatman Int. Ltd) filter paper assay using varying tritium-labelled deoxyribonucleoside substrate concentrations.31 All radiolabelled substrates were from Moravek Biochemicals Inc. (Brea, CA, USA) and GE Healthcare. The 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 of BSA and 0.5 mM CHAPS.
One unit (U) of deoxyribonucleoside kinase activity is defined as 1 µmol of the corresponding monophosphate formed per minute.
Kinetic data were evaluated by non-linear regression analysis using the Michaelis–Menten equation v = Vmax x [S]/(Km + [S]) or whenever substrate inhibition was evident, non-linear regression was applied using the equation for substrate inhibition v = (Vmax x [S])/(Km + [S] + ([S]2/Kis)). Eadie–Hofstee plots (v/[S] against v) were also calculated to illustrate the substrate inhibition.
Amino acid sequences were aligned using ClustalX v. 1.81 with default settings,32 and phylogenetic analysis was done with TreeCon v. 1.3b.33
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Susceptibility of pathogenic bacteria
Twenty-five bacterial isolates of clinical origin and type strains (Table 1), covering 11 Gram-positive and Gram-negative genera, were tested for their susceptibility to different nucleoside analogues, including AZT, FdUrd, 2-chloro-deoxyadenosine (CdA), 9-ß-D-arabinofuranosyl-2-fluoroadenine (F-AraA), D-arabinosyl adenine (AraA), D-arabinosyl cytosine (AraC) and gemcitabine. Cell suspensions of the pathogenic bacteria were spotted on Mueller–Hinton agar plates containing a dilution series of the aforementioned nucleoside analogues. Of the analogues tested, CdA, F-AraA, AraA and AraC had no or very limited effect on the isolates. However, gemcitabine was very active against the Gram-positive bacteria (Table 1). Listeria, Bacillus, Enterococcus and Staphylococcus isolates were susceptible to gemcitabine concentrations between 0.001 and 1.0 µM (Table 1). Gram-negative strains did not show any significant susceptibility to gemcitabine. The Gram-negative strains were all, with the exception of P. aeruginosa and one E. coli isolate (07529-1), susceptible to AZT in concentrations between 0.032 and 31.6 µM (Table 1). Conversely, none of the Gram-positive bacteria was susceptible to AZT. FdUrd was active towards both Gram-positive and Gram-negative bacteria; however, the lowest minimal inhibitory concentrations (MICs) were for the Gram-positive strains in the range from 0.003 to 1 µM (Table 1). The Gram-negative strains responded to somewhat higher FdUrd concentrations having MICs between 1 and 10 µM, only the E. coli K12 strain was more susceptible and again P. aeruginosa and one E. coli isolate were unaffected (Table 1). Apparently, in several cases a nucleoside drug was successfully transported into the cell, likely phosphorylated and clearly had a detrimental effect on the cell metabolism. To understand the mechanism behind the specific activation of each successful drug, we investigated the possible role of endogenous dNKs in the observed susceptibilities. We analysed the genomes of the susceptible and two other (C. perfringens and P. multocida) pathogenic bacteria for putative dNK-encoding genes, cloned putative dNK genes and tested them in E. coli for their ability to activate the prodrugs in a heterologous surrounding.
Identification and cloning of bacterial dNKs
Bacterial genome sequences of six selected human or domesticated animal pathogens were searched for dNK homologues. Several known dNK amino acid sequences were employed in the search. Seven TKs and four non-TKs were identified and subsequently the corresponding tdk, dak and dgk genes were cloned from isolated genomic DNA into the pGEX-2T expression vector (Table 2). The TKs were from B. cereus (BcTK), L. monocytogenes (LmTK), C. perfringens (CpTK), P. multocida (PmTK), S. enterica (SeTK) and Y. enterocolitica (YeTK). Also the E. coli TK (EcTK) was cloned into the same vector. The overall amino acid identity to human TK1 was lowest for LmTK (14.9%) and highest for BcTK (29.9%). The amino acid identity of SeTK to BcTK was 21.1% and the observed TKs phylogenetic relationship showed that BcTK was more closely related to eukaryotic TK1-like TKs than to the Gram-negative type (Figure 1) consistent with our previous studies.34 Non-TK1-like kinases were found only in the Gram-positive bacteria fraction of the pathogenic bacteria searched. B. cereus dAK (BcdAK) and dGK (BcdGK); L. monocytogenes dAK (LmdAK) and C. perfringens dAK (CpdAK) encoding genes were cloned. The bacterial non-TK-like kinases are related but distinct from the eukaryotic non-TK1-like kinases such as human dCK, dGK and TK2. They comprise a separate group of kinases and it appears that CpdAK and LmdAK are similar to the L. acidophilus dAK/dGK kinases.22 Overall, CpdAK and LmdAK are 40% identical to the L. acidophilus dAK/dGK kinases, which seem to originate from a post-speciation duplication of the original gene (Figure 1). Furthermore, the dGK gene found in the Bacillus genus was absent in C. perfringens, L. monocytogenes and L. acidophilus. No dNK homologues were found in the genomes of P. aeruginosa or Helicobacter pylori. Likely, these two bacteria have lost their dNK-encoding genes.
|
Susceptibility of E. coli harbouring heterologous dNKs
In Table 2, we show the susceptibility of the TK-deficient E. coli strain KY895 transformed with the vectors harbouring the cloned kinase genes. Expression of the kinases relied on the low basal level expression. AZT was effective whenever the strain was transformed with a tdk gene regardless the origin, Gram-negative or Gram-positive, which is contradictory to the in vivo susceptibility test (Table 1) and earlier studies that showed that Gram-positives were not susceptible to AZT.35 AZT concentrations in the range 0.01–10 µM inhibited the growth of KY895 transformed with tdk. CpTK was the least effective recombinant TK, with an MIC in recombinant E. coli of 10 µM AZT, but gave at least a 10-fold increase in susceptibility compared with the control (empty pGEX-2T vector). More than 1000- to 10 000-fold increase in susceptibility was observed with SeTK, BcTK (Figure 2a) and YeTK (Table 2). KY895 did show significantly increased susceptibility towards gemcitabine when transformed with the BcdAK or CpdAK genes (Figure 2b). Both genes/kinases increased the susceptibility of KY895 to gemcitabine dramatically as they resulted in at least a 10 000-fold decrease in MIC compared with the control (Table 2). TKs, LmdAK and BcdGK did not cause such a significant change in MIC with gemcitabine. Only KY895 transformed with BcTK and PmTK showed increased susceptibility towards FdUrd, with an MIC of 0.01 µM, whereas all others, including the control, had an MIC of 0.316 µM. The result is not immediately surprising since other enzymes like the thymidine phosphorylase and to some extent also uridine phosphorylase rapidly degrade FdUrd.13 The results presented in Figure 2 and Table 2 demonstrate that dNKs are responsible for the susceptibilities observed in Table 1. To further investigate the substrate specificities of the most characteristic kinases, we measured the kinetic parameters of purified recombinant enzymes.
|
Purification of recombinant kinases
We attempted to fully characterize the whole dNK repertoire of two bacteria. Recombinant SeTK from the Gram-negative S. enterica and BcTK, BcdAK and BcdGK from the Gram-positive B. cereus were expressed in E. coli and purified using affinity chromatography. The N-terminal GST-fusion provided by the pGEX-2T vector was used as an affinity-tag for the purification of BcTK, BcdAK and BcdGK. Thrombin was used as a specific protease cleaving the GST-tag from the kinase of interest while still bound on the GSH-column leaving only two extra amino acids (glycine and serine) at the N terminus. Afterwards, pure recombinant protein was eluted from the GSH-column. The GST-fusion with SeTK was not a good substrate for thrombin yielding only trace amounts of the cleaved SeTK. Therefore, the tdk gene from S. enterica was subcloned into the pASK75-8His vector28 to obtain SeTK with a C-terminal histidine tag. This protein was easily purified on a Ni–NTA column. SDS–PAGE of the pure proteins is shown in Figure 3 and is in reasonable agreement with the theoretical molecular weights. The theoretical molecular weights for the purified recombinant BcdAK, BcdGK, BcTK and SeTK were 26.2, 24.8, 21.8 and 25.7 kDa, respectively, and include the modifications that remain after purification (for BcdAK, BcdGK and BcTK two amino acids on the N terminus originating from the thrombin cleavage site; for SeTK a nineteen amino acid long tag on the C terminus). Substrate specificity and kinetics of these enzymes were characterized using ATP as the phosphate donor.
|
Kinetics of thymidine kinases
SeTK showed typical Michaelis–Menten kinetics with Thd and dUrd, having Thd as the preferred substrate (Figure 4a and c, and Table 3). The Km for Thd was 34 µM, 
4 times lower than for dUrd. However, SeTK was almost equally specific for AZT as for Thd (Table 3) having kcat/Km values of 6.62 x 104 and 9.41 x 104 s–1 M–1, respectively, although SeTK showed characteristic substrate inhibition with declining reaction rates at AZT concentrations above 200 µM (Figure 4b). Using non-linear regression and the expression for substrate inhibition, the Km for AZT was determined to 73.3 µM and Kis 754 µM. BcTK also had Thd as the preferred substrate over dUrd, but in comparison to SeTK the BcTK enzyme was much more specific for Thd with a kcat/Km of 1.08 x 106 s–1 M–1, 10 times higher than for dUrd and 4 times higher than for AZT. Generally, the BcTK had much lower Km for the three substrates than did SeTK. We could not detect any ATP-dependent phosphorylation of dAdo, dGuo or dCyd by either of the BcTK or SeTK. BcTK, like SeTK, showed pronounced substrate inhibition with AZT and declining reaction rates were observable at AZT concentrations above 8 µM. The Km was determined to be 4.43 µM and Kis for the AZT substrate inhibition was 84.6 µM. The clear substrate inhibition is demonstrated from the Eadie–Hofstee plots in Figure 4(b and e). In short, both TK enzymes could efficiently phosphorylate AZT.
|
|
Kinetics of dAK and dGK kinases
BcdAK was almost equally specific for dAdo and dCyd, the only two natural substrates accepted by BcdAK. kcat/Km was 3.68 x 104 and 2.02 x 104 s–1 M–1 for dAdo and dCyd, respectively (Table 3). In addition, BcdAK also efficiently phosphorylated gemcitabine, albeit with lower turnover than for the natural substrates and kcat/Km was 5.05 x 103 s–1M–1. The reactions followed Michaelis–Menten kinetics for all three substrates and the Km values were in the same range, 33.2, 27.0 and 33.5 µM for dAdo, dCyd and gemcitabine, respectively. Purified BcdGK was extremely specific and phosphorylated only dGuo with a Km of 4.4 µM for dGuo, the only natural dN it accepted (Table 3). The efficiency constant kcat/Km was 2.09 x 105 s–1 M–1. These results clearly show that BcdAK efficiently phosphorylated gemcitabine.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Only a few kinases of bacterial origin have so far been studied for substrate specificity and kinetics and not even a single Gram-positive bacterial species has been thoroughly analysed for the whole set of dNKs.19–25 In Bacillus subtilis, two overlapping genes, yaaG and yaaF, encode the dGK and dAK/dCK kinases, respectively. The dGK phosphorylates dGuo, whereas dAK/dCK phosphorylates dAdo and dCyd.23 Thymidine kinase activity has been measured in extracts of B. subtilis,36–38 but the corresponding gene and enzyme has not been characterized. Mycoplasma species, despite the absence of a cell wall, are phylogenetically related to Gram-positive bacteria and like the Gram-positive bacteria described in this study, mycoplasmas also encode TK1 and non-TK1-like kinases.24,25 Unlike the previously described organisms, E. coli only has one deoxyribonucleoside kinase, TK, which only phosphorylates dThd and dUrd.39 In this study, B. cereus and S. enterica were fully characterized for all their dNKs. In addition, several dNKs from other bacteria were described.
It seems that the preservation of non-TK1-like kinases within Gram-negative bacteria is sporadic. For instance, we did not find any other dNK-encoding genes than TKs within the alpha-, gamma- or epsilon proteobacteria. However, several putative non-TK-encoding genes were identified within the group of beta proteobacteria species (data not shown). We identified and cloned seven novel TKs and four novel non-TK1-like dNKs from bacteria that cause mild to life-threatening diseases in humans. The phylogenetic analysis of the cloned kinases was consistent with earlier studies of bacterial putative TKs and grouped Gram-positive TKs together with eukaryote TK1-like kinases.34 The low Km value of BcTK with Thd is comparable to that of human TK1 and supports the obtained phylogenetic relationships.31 Furthermore, the phylogenetic analysis revealed that CpdAK and LmdAK are related to the L. acidophilus dAK/dGK kinases (LadAK/LadGK), which seemingly result from of a post-speciation duplication of the original gene (Figure 1).
Gemcitabine was quite active in vivo when tested on L. monocytogenes isolates. However, it is puzzling that the LmdAK did not cause any significant increase in susceptibility when tested in KY895 since this is the only dNK in L. monocytogenes that could activate gemcitabine. However, the similarity of LmdAK to LadAK and LadGK and their post-translational modification in L. acidophilus could explain the missing effect of gemcitabine in the heterologous environment of KY895.22 BcdAK had a much broader substrate preference than BcdGK. BcdAK accepted dAdo and dCyd with only slightly higher preference for the purine nucleoside. In addition, gemcitabine was also a substrate for BcdAK, albeit four times poorer than dCyd. Like B. subtilis dGK (BsdGK),23 BcdGK phosphorylated only dGuo. The two highly similar kinases BsdGK and BcdGK (50% identity) phosphorylated dGuo equally well with kcat/Km of 2.4 x 105 and 2.1 x 105 s–1 M–1, respectively. Thus, it seems that the non-TK1-like kinases in B. cereus behave like those in B. subtilis. The substrate specificities of the non-TK1-like kinases from bacteria that have been characterized to date vary to a great extent. BcdGK and BsdGK only phosphorylate one natural substrate, dGuo,23 whereas the dAK from Mycoplasma mymySC phosphorylates dAdo, dGuo and dCyd.24 By contrast, the dAKs in B. subtilis and B. cereus, as described here, only accept dAdo and dCyd. So far no kinase specific only for dAdo has been characterized. The wide variation in kinase substrate preferences and organization of the kinase-encoding genes suggest that each bacterial species could have very specific affinity towards different nucleoside analogues.
Apparently, kinase genes are differentially distributed among different bacteria and represent potential species-specific activators of nucleoside prodrugs. Indeed, we showed that several nucleoside analogues could be employed as species-specific antibiotics (Table 1). AZT was specific for Gram-negative Enterobacteriaceae, while it had no effect on P. aeruginosa or any of the Gram-positive bacteria employed. Our findings that no dNK-encoding genes are present in the P. aeruginosa genome are consistent with its susceptibility profile towards the nucleoside analogues presented here and in earlier studies, which conclude that no TK activity can be measured in the extracts from this organism.37 Some of the most prolific human pathogens of Gram-negative bacterial origin belong to the Salmonella and Yersinia genera. The results presented here show that members of both genera are highly susceptible to AZT and FdUrd. Earlier studies have concluded that both E. coli and Salmonella typhimurium were much more susceptible to AZT than Y. enterocolitica.35 In our investigation, all three species were susceptible to AZT in the range of 1–31.6 µM (Table 1). The reason for this discrepancy with earlier literature is not clear. By contrast, gemcitabine was specific for the Gram-positive bacteria and targeted Bacilli, Enterococci, Staphylococci and Listeria. The result is consistent with the presence of non-TK1-like kinases that eventually could activate gemcitabine in these organisms (Figure 1 and Table 2). Plasma concentrations above 2 µM are usually achieved during treatment of HIV patients with AZT and for shorter periods the dose could be increased to reach even higher plasma concentrations.40,41 Similarly, plasma concentrations of 2.5–20 µM are achievable during treatment with gemcitabine.42
The AZT susceptibility observed in this study clearly relates to TK activity and apparently the KY895 strain is a good model to study AZT activation (Figure 2a). Non-TK1-like dNKs did not render the KY895 susceptible to AZT (Table 2). Earlier studies have suggested the bactericidal effect of AZT to be exerted by the tri-phosphorylated form as it causes chain termination upon incorporation in DNA during synthesis. Furthermore, mutants lacking TK activity have been shown to be resistant towards AZT,35 and we show here that the recombinant TKs efficiently phosphorylate AZT in a concentration-dependent manner, where AZT concentrations well above Km inhibits the reaction (Figure 4). Thus, the first step in the activation of AZT is its phosphorylation by TK. The results presented here further establish that TKs from Gram-positive bacteria also efficiently phosphorylate AZT. Therefore, it is rather surprising that Gram-positive organisms themselves are not susceptible to AZT.
HIV-infected patients are known to have a high occurrence of Salmonella bacteraemia, which left without successful treatment could lead to sepsis and death. However, reports that AIDS patients receiving treatment with AZT (Retrovir®) do not get Salmonella infections or have a lower recurrence indicate a new strategy for antibacterial treatment.43 The high dose tolerability of AZT in man establishes AZT as a candidate for a last resort treatment of some bacterial infections like multiresistant salmonellae. To our surprise, one of the clinical E. coli isolates (Strain ID 07529-1) tested in this study was also resistant to AZT (Table 1). In theory, AZT resistance could easily evolve as a result of mutations affecting the activity of the nucleoside transporter, TK activity, thymidylate kinase specificity or even DNA-polymerase specificity. Indeed, a few bacterial isolates from AIDS patients receiving anti-HIV treatment with AZT have been shown to become AZT-resistant,44 and Elwell et al.35 showed that when liquid medium containing AZT was inoculated with E. coli or S. typhimurium, TK-deficient mutants arose. The antibacterial activity of AZT in Gram-negative Enterobacteriaceae is mainly due to the phosphorylation and thereby activation of the drug by an endogenous TK.35 Our results presented in this communication further establish AZT as a Gram-negative-specific antibiotic and elucidates the underlying activation kinetics.
In Gram-positive bacteria, AZT is believed to utilize the same membrane transporter as Thd, and it has been suggested that the resistance towards AZT was due to insufficient phosphorylation by TK.45 When we transformed E. coli with tdks from Gram-positive bacteria, E. coli became susceptible to AZT (Table 2). In addition, purified recombinant BcTK efficiently phosphorylated AZT with 16 times lower Km (4.43 µM) than SeTK (Table 3). Compared to BcTK, E. coli TK also has a relatively high Km for AZT, 21 µM.35 Therefore, we have to assume that AZT is readily phosphorylated also in Gram-positive bacteria to AZT-monophosphate. The AZT-monophosphate is most likely not a good substrate for the thymidylate kinases in Gram-positive bacteria as has already been described for human, yeast and Yersinia pestis TMPKs.46,47 The AZT insensitivity of Gram-positive bacteria could then result from a combination of several plausible mechanisms. (i) Insufficient TMPK activity with AZT-monophosphate. (ii) Insufficient AZT transport. (iii) AZT-triphosphate is not a good substrate for the Gram-positive DNA-polymerase. (iv) Rapid degradation of AZT. Which of these scenarios are responsible for the AZT resistance in Gram-positive bacteria is unclear. Nonetheless, the bactericidal effect of AZT seen in Gram-negative bacteria is not achieved in the Gram-positive bacteria employed in this study, ruling out AZT as a possible agent for antibacterial treatment of infections by multiresistant Gram-positive bacteria.
FdUrd caused very increased susceptibility in the Gram-positive strains; however the effect of FdUrd was not immediately visible or distinguishable when the KY895 was transformed with heterologous kinases. Only BcTK and PmTK caused increased susceptibility of KY895 to FdUrd. In a similar study, mycoplasmas were shown to be susceptible to FdUrd but the complete nucleoside metabolism in this group of mollicutes has not yet been elucidated.25 It is possible that the degradation of FdUrd to FU in KY895, which was derived from K12, is so rapid that the MIC observed in KY895 and K12 is caused by FU and its metabolites rather than FdUMP but it is worth noting that the MIC of the E. coli K12 strain (Table 1) with FdUrd is the same as the KY895 transformed with the empty vector and most of the other constructs. Hence, we cannot, based on our results, claim the dNKs as important players in the activation of FdUrd. For the purified BcdAK, gemcitabine was almost a 10 times poorer substrate than dAdo and dCyd. Regardless, the effect of gemcitabine on several of the Gram-positive bacteria was astonishing. Already a successful anticancer drug, gemcitabine and its analogues could also turn out to be an effective antibiotic in the fight against multiresistant pathogenic bacteria.
How easily could the resistance towards nucleoside analogues be developed? It could be that dNKs are dispensable for the cell metabolism under the optimal growth conditions since the salvage pathway is redundant to the de novo synthesis of dNs in most organisms. Therefore, resistance against nucleoside analogues could be developed by diminishing the dNK activity. Elevated activity of analogue-degrading enzymes, i.e. nucleoside phosphorylases, could also decrease the level of sensitivity. In addition, a reduced nucleoside uptake could also increase the resistance. However, these impairments may greatly reduce the general fitness and competitive ability of the mutated microbes.
In conclusion, nucleoside analogues used in treatment of viral infections and cancer also have a potential to be activated by dNKs in bacteria and thereby used as species-specific antibiotics for treatment of bacterial infections, especially those caused by multiresistant bacteria. Given the alarming rise in the incidence of multiresistant bacteria, this is a welcome discovery.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
The work presented in this paper was supported by the Danish Technical Research Council (STVF), Swedish Research Foundation (VR), Crafoord Foundation (Sweden) and Meyer's Foundation (Denmark).
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
None to declare.
|
Footnotes |
|---|
Present address. Department of Molecular Biology, Copenhagen University, Copenhagen, Denmark. 
Present address. Institute of Environmental Science and Research, Christchurch, New Zealand.
|
Acknowledgements |
|---|
We thank Wolfgang Knecht for his comments on the manuscript.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
1 WHO. Antimicrobial resistance. World Health Organization Fact Sheet (2002) 194. http://www.who.int/mediacentre/factsheets/fs194/en.
2 Jones RN, Pfaller MA. Bacterial resistance: a worldwide problem. Diagn Microbiol Infect Dis (1998) 31:379–88.[CrossRef][ISI][Medline]
3 Alekshun MN, Levy SB. Commensals upon us. Biochem Pharmacol (2006) 71:893–900.[CrossRef][ISI][Medline]
4 Silva Coimbra MV, Silva-Carvalho MC, Wisplinghoff H, et al. Clonal spread of methicillin-resistant Staphylococcus aureus in a large geographic area of the United States. J Hosp Infect (2003) 53:103–10.[CrossRef][ISI][Medline]
5 Aarestrup FM. Veterinary drug usage and antimicrobial resistance in bacteria of animal origin. Basic Clin Pharmacol Toxicol (2005) 96:271–81.[CrossRef][ISI][Medline]
6
Weigel LM, Clewell DB, Gill SR, et al. Genetic analysis of a high-level vancomycin-resistant isolate of Staphylococcus aureus. Science (2003) 302:1569–71.
7 Bratu S, Quale J, Cebular S, et al. Multidrug-resistant Pseudomonas aeruginosa in Brooklyn, New York: molecular epidemiology and in vitro activity of polymyxin B. Eur J Clin Microbiol Infect Dis (2005) 24:196–201.[CrossRef][ISI][Medline]
8 Rice LB. Antimicrobial resistance in Gram-positive bacteria. Am J Med (2006) 119:S11–19.[Medline]
9 Stein M, O'Sullivan P, Wachtel T, et al. Causes of death in persons with human immunodeficiency virus infection. Am J Med (1992) 93:387–90.[CrossRef][ISI][Medline]
10
Mitsuya H, Weinhold KJ, Furman PA, et al. 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 (1985) 82:7096–100.
11 Plunkett W, Huang P, Gandhi V. Preclinical characteristics of gemcitabine. Anticancer Drugs (1995) 6(Suppl 6):7–13.
12 Van Laar JA, Rustum YM, Ackland SP, et al. Comparison of 5-fluoro-2'-deoxyuridine with 5-fluorouracil and their role in the treatment of colorectal cancer. Eur J Cancer (1998) 34:296–306.[CrossRef][ISI][Medline]
13
Beck CF, Ingraham JL, Neuhard J, et al. Metabolism of pyrimidines and pyrimidine nucleosides by Salmonella typhimurium. J Bacteriol (1972) 110:219–28.
14
Cohen SS, Flaks JG, Barner HD, et al. The mode of action of 5-fluorouracil and its derivatives. Proc Natl Acad Sci USA (1958) 44:1004–12.
15 Arner ESJ, Eriksson S. Mammalian Deoxyribonucleoside Kinases. Pharmacol Ther (1995) 67:155–86.[CrossRef][ISI][Medline]
16 Eriksson S, Munch-Petersen B, Johansson K, et al. Structure and function of cellular deoxyribonucleoside kinases. Cell Mol Life Sci (2002) 59:1327–46.[CrossRef][ISI][Medline]
17 Sandrini MP, Piskur J. Deoxyribonucleoside kinases: two enzyme families catalyze the same reaction. Trends Biochem Sci (2005) 30:225–8.[CrossRef][ISI][Medline]
18
Munch-Petersen B, Piskur J, Sondergaard L. Four deoxynucleoside kinase activities from Drosophila melanogaster are contained within a single monomeric enzyme, a new multifunctional deoxynucleoside kinase. J Biol Chem (1998) 273:3926–31.
19 Black ME, Hruby DE. Nucleotide sequence of the Escherichia coli thymidine kinase gene provides evidence for conservation of functional domains and quaternary structure. Mol Microbiol (1991) 5:373–9.[CrossRef][ISI][Medline]
20 Bockamp EO, Blasco R, Vinuela E. Escherichia coli thymidine kinase: nucleotide sequence of the gene and relationships to other thymidine kinases. Gene (1991) 101:9–14.[CrossRef][ISI][Medline]
21 McNab R. Cloning and sequence analysis of thymidine kinase from the oral bacterium Streptococcus gordonii. FEMS Microbiol Lett (1996) 135:103–10.[CrossRef][ISI][Medline]
22 Ives DH, Ikeda S. Life on the salvage path: the deoxynucleoside kinase of Lactobacillus acidophilus R-26. Prog Nucleic Acid Res Mol Biol (1998) 59:205–55.[ISI][Medline]
23
Andersen RB, Neuhard J. 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 (2001) 276:5518–24.
24 Wang L, Westberg J, Bolske G, et al. Novel deoxynucleoside-phosphorylating enzymes in mycoplasmas: evidence for efficient utilization of deoxynucleosides. Mol Microbiol (2001) 42:1065–73.[CrossRef][ISI][Medline]
25 Carnrot C, Wehelie R, Eriksson S, et al. Molecular characterization of thymidine kinase from Ureaplasma urealyticum: nucleoside analogues as potent inhibitors of mycoplasma growth. Mol Microbiol (2003) 50:771–80.[CrossRef][ISI][Medline]
26
Igarashi K, Hiraga S, Yura T. A deoxythymidine kinase deficient mutant of Escherichia coli. II. Mapping and transduction studies with phage phi 80. Genetics (1967) 57:643–54.
27 Clinical and Laboratory Standards Institute. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically—Seventh Edition: Approved Standard M7-A7 (2006) Wayne, PA, USA: CLSI.
28 Knecht W, Petersen GE, Munch-Petersen B, et al. Deoxyribonucleoside kinases belonging to the thymidine kinase 2 (TK2)-like group vary significantly in substrate specificity, kinetics and feed-back regulation. J Mol Biol (2002) 315:529–40.[CrossRef][ISI][Medline]
29 Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (1970) 227:680–5.[CrossRef][Medline]
30 Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem (1976) 72:248–54.[CrossRef][ISI][Medline]
31
Munch-Petersen B, Cloos L, Tyrsted G, et al. Diverging substrate specificity of pure human thymidine kinases 1 and 2 against antiviral dideoxynucleosides. J Biol Chem (1991) 266:9032–8.
32
Thompson JD, Gibson TJ, Plewniak F, et al. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res (1997) 25:4876–82.
33
Van de Peer Y, De Wachter R. TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Comput Appl Biosci (1994) 10:569–70.
34 Sandrini MP, Clausen AR, Munch-Petersen B, et al. Thymidine kinase diversity in bacteria. Nucleosides Nucleotides Nucleic Acids (2006) 25:1153–8.[CrossRef][ISI][Medline]
35
Elwell LP, Ferone R, Freeman GA, et al. Antibacterial activity and mechanism of action of 3'-azido-3'-deoxythymidine (BW A509U). Antimicrob Agents Chemother (1987) 31:274–80.
36
Mollgaard H. Deoxyadenosine/deoxycytidine kinase from Bacillus subtilis. Purification, characterization, and physiological function. J Biol Chem (1980) 255:8216–20.
37 Saito H, Tomioka H. Thymidine kinase of bacteria: activity of the enzyme in actinomycetes and related organisms. J Gen Microbiol (1984) 130:1863–70.[Medline]
38 Saito H, Tomioka H, Ohkido S. Further studies on thymidine kinase: distribution pattern of the enzyme in bacteria. J Gen Microbiol (1985) 131:3091–8.[Medline]
39
Okazaki R, Kornberg A. Deoxythymidine Kinase of Escherichia coli. I. Purification and Some Properties of the Enzyme. J Biol Chem (1964) 239:269–74.
40 Barry MG, Khoo SH, Veal GJ, et al. The effect of zidovudine dose on the formation of intracellular phosphorylated metabolites. AIDS (1996) 10:1361–7.[ISI][Medline]
41 Singlas E, Pioger JC, Taburet AM, et al. Comparative pharmacokinetics of zidovudine (AZT) and its metabolite (G.AZT) in healthy subjects and HIV seropositive patients. Eur J Clin Pharmacol (1989) 36:639–40.[CrossRef][ISI][Medline]
42
Cattel L, Airoldi M, Delprino L, et al. Pharmacokinetic evaluation of gemcitabine and 2',2'-difluorodeoxycytidine-5'-triphosphate after prolonged infusion in patients affected by different solid tumors. Ann Oncol (2006) 17(Suppl 5):v142–7.
43 Casado JL, Valdezate S, Calderon C, et al. Zidovudine therapy protects against Salmonella bacteremia recurrence in human immunodeficiency virus-infected patients. J Infect Dis (1999) 179:1553–6.[CrossRef][ISI][Medline]
44 Lewin CS, Watt B, Paton R, et al. Isolation of zidovudine resistant Escherichia coli from AIDS patients. FEMS Microbiol Lett (1990) 58:141–3.[Medline]
45 Lewin CS, Allen RA, Amyes SG. Mechanisms of zidovudine resistance in bacteria. J Med Microbiol (1990) 33:235–8.[Abstract]
46 Lavie A, Konrad M, Brundiers R, et al. Crystal structure of yeast thymidylate kinase complexed with the bisubstrate inhibitor P1-(5'-adenosyl) P5-(5'-thymidyl) pentaphosphate (TP5A) at 2.0 A resolution: implications for catalysis and AZT activation. Biochemistry (1998) 37:3677–86.[CrossRef][Medline]
47 Chenal-Francisque V, Tourneux L, Carniel E, et al. The highly similar TMP kinases of Yersinia pestis and Escherichia coli differ markedly in their AZTMP phosphorylating activity. Eur J Biochem (1999) 265:112–9.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||