OUP user menu

Deoxyribonucleoside kinases in two aquatic bacteria with high specificity for thymidine and deoxyadenosine

Tinkara Tinta, Louise Slot Christiansen, Anke Konrad, David A. Liberles, Valentina Turk, Birgitte Munch-Petersen, Jure Piškur, Anders R. Clausen
DOI: http://dx.doi.org/10.1111/j.1574-6968.2012.02565.x 120-127 First published online: 1 June 2012


Deoxyribonucleoside kinases (dNKs) are essential in the mammalian cell but their ‘importance’ in bacteria, especially aquatic ones, is less clear. We studied two aquatic bacteria, Gram-negative Flavobacterium psychrophilum JIP02/86 and Polaribacter sp. MED152, for their ability to salvage deoxyribonucleosides (dNs). Both had a Gram-positive-type thymidine kinase (TK1), which could phosphorylate thymidine, and one non-TK1 dNK, which could efficiently phosphorylate deoxyadenosine and slightly also deoxycytosine. Surprisingly, the four tested dNKs could not phosphorylate deoxyguanosine, and apparently, these two bacteria are missing this activity. When tens of available aquatic bacteria genomes were examined for the presence of dNKs, a majority had at least a TK1-like gene, but several lacked any dNKs. Apparently, among aquatic bacteria, the role of the dN salvage varies.

  • aquatic bacteria
  • deoxyribonucleoside kinase
  • 3H-thymidine incorporation
  • nucleoside salvage


Deoxyribonucleotides are the building blocks for the synthesis or repair of the genetic material (Eriksson et al., 2002). In the animal cell, deoxyribonucleosides are provided through the de novo biosynthesis and salvage, and both pathways are essential. In the salvage pathway, the phosphorylation of deoxyribonucleosides (dNs) into dN monophosphates (dNMP) is the first step and considered as the bottle-neck. A phosphate group is transferred from a phosphate donor, usually a nucleoside triphosphate, like ATP, to the 5′-hydroxygroup of the dN substrate (Eriksson et al., 2002) by deoxyribonucleoside kinases (dNKs). Two superfamilies of dNKs exist, the thymidine kinase 1 (TK1-like) and the non-TK1-like family (Sandrini & Piškur, 2005). TK1s are specific only for thymidine (dT) and deoxyuridine (dU), while the dNKs of the non-TK1-like family are rather unspecific compared to the TK1s, typically phosphorylating one or several of the native dNs (Eriksson et al., 2002; Sandrini & Piškur, 2005). However, the level of amino acid identity to the already characterized dNKs is still not a sufficient parameter to predict the substrate specificity of new dNKs.

In mammals, four essential dNKs can be found, while in bacteria so far it has been thought that Gram-negative bacteria have only one dNK, TK1, while Gram-positive bacteria seem to have several dNKs (Sandrini et al., 2007a, b). There are only limited data available so far on the situation in aquatic bacteria. However, because DNA pool in aquatic environments is the largest pool of DNA and dNs on Earth, aquatic microorganisms might gain a fitness benefit from the ability to degrade DNA and re-use the building blocks (DeFlaun et al., 1987).

In this study, we examined the sequenced genomes from several aquatic bacteria for genes encoding dNKs. We focused on Polaribacter sp. MED 152, which serves as a model to study the cellular and molecular processes in bacteria that express proteorhodopsin, their adaptation to the oceanic environment, and their role in the C-cycling (González et al., 2008), and on Flavobacterium psychrophilum JIP02/86, which is a widely distributed fish pathogen, capable of surviving in different habitats (Duchaud et al., 2007).

Materials and methods

Bioinformatics and phylogenetic analysis

Database searches for putative dNK genes in the sequenced genomes from various aquatic bacteria were made using the genome basic local alignment search tool (blast) at the National Center for Biotechnology Information (NCBI). Details on the sequence used in the search can be found in the Supporting Information, Data S1.

The two newly identified TK1-like protein sequences [Polaribacter sp. MED 152 (PdTK1, ZP_01053169) and F. psychrophilum JIP02/86 (FpTK1, YP_001295968)], which were extracted from the genome sequences data but then resequenced in our laboratory, were aligned against the previously biochemically characterized TK1 sequences (see above) using MAFFT (Katoh & Kuma, 2002) with JTT 200 as the substitution matrix. A phylogenetic tree was then reconstructed via maximum likelihood using PhyML (Guindon & Gascuel, 2003) with the WAG+I+G+F model and rooted using the human TK1 as an outgroup.

Genomic DNA

Genomic DNA of F. psychrophilum JIP02/86 was provided by E. Duchaud, Unité de Virologie et Immunologie Moléculaires, INRA — Domaine de Vilvert (GeneBank database accession number NC_009613). Genomic DNA of Polaribacter sp. MED152 was provided by J. Pinhassi, Marine Microbiology, University of Kalmar, Sweden (GeneBank database accession number NZ_AANA00000000).

Cloning and expression

Open reading frames identified by homology to the known dNKs were amplified from the genomic DNA by PCR using primers with the restriction enzyme overhang for BamHI and EcoRI/MfeI (Tables S1 and Table S2). Amplified ORFs were digested with appropriate restriction enzymes and subcloned into the BamHI and EcoRI site of the commercially available expression vector pGEX-2T (Pharmacia Biotech) using standard molecular biology techniques. The resulting constructs expressed a hybrid protein with the N-terminal glutathione-S-transferase (GST) fusion tag, the thrombin protease cleavage site, and the dNK of interest. Expression and purification details can be found in the Data S1.

Steady-state kinetic measurements

Phosphorylating activities of purified dNKs were determined by initial velocity measurements based on four time samples (4, 8, 12, and 16 min) using the DE-81 filter paper (Whatman Inc.) assay and various radio-labeled dN substrate concentrations (Munch-Petersen et al., 1991). Standard assay conditions were 50 mM Tris–HCl pH 7.5, 10 mM DTT, 2.5 mM ATP, 2.5 mM MgCl2, 3 mg mL−1 BSA, 0.5 mM CHAPS, and the indicated concentration of radio-labeled dN substrate in a final volume of 50 µL. The radioactive dNs (3H-dT, 3H-dA, 3H-dG, and 3H-dC) used in the assay were obtained from Moravek or PerkinElmer. When determining the activities in crude bacterial extracts, NaF (6 mM) was added to the reaction mixture to inhibit phosphatase activities, and when dC was used as the substrate, also 0.5 mM tetrahydrouridine (THUR) was added to inhibit possible cytidine deaminase activity. The activities were measured at 37 °C, except for PdTK1 and FpTK1, which were measured at 21 °C. When necessary, the enzyme or crude extract was diluted in the enzyme dilution buffer (50 mM Tris–HCl pH 7.5, 1 mM CHAPS, 3 mg mL−1 BSA, and 5 mM DTT). One unit (u) of enzyme activity is defined as the amount of kinase that can phosphorylate 1 nmol of nucleoside per minute under standard assay conditions (Munch-Petersen et al., 1998). Kinetic data were evaluated by fitting the data to the Michaelis–Menten equation ν = Vmax*(S)/(Km + (S)) using nonlinear regression analysis using Graph prism software.

Effect of temperature on enzyme activities

In order to determine the effect of the temperature on the PdTK1 phosphorylating activity, the activity of enzyme was measured at 5, 10, 15, 21, 25, 30, and 37 °C. In this case, all radio-assays were performed with 500 µM 3H-dT as substrate and ATP as phosphate donor. When measured at 21 and 25 °C, activities were determined by initial velocity measurements based on the four time samples, retrieved after 3, 6, 9, and 12 min. In the assays performed at 5, 10, and 15 °C, the four time samples were taken after 5, 15, 30, and 45 min. In order to determine the activity at 30 and 37 °C, the assays also had to be performed with the pro-longed time series, with time samples taken after 2, 5, 10, 20, 30, and 40 min, owing to the low activities. In a separate experiment, thermostability at 0 and 37 °C was investigated by incubating the enzyme 1 h prior to the measurement of the activity at 21 °C. In this experiment, time samples were taken after 2, 5, 10, 20, 30, and 40 min. Also FpTK1 was initially found to exhibit the effect of temperature on the phosphorylation activity. Therefore, the assays were conducted at 21 °C.


Two aquatic bacteria have several dNKs

Several aquatic bacterial genome sequences were searched for genes homologous to the known, previously characterized bacterial and eukaryote dNKs. Two of the analyzed bacteria, F. psychrophilum JIP02/86 and Polaribacter sp. MED 152, both Gram-negative and both belonging to Bacteroidete class, served as model organisms in our studies. Putative genes encoding dNKs in the bacterial genomes of F. psychrophilum JIP02/86 (NC_009613) and Polaribacter sp. MED 125 (NZ_AANA00000000) are listed in Table S2. In each species, we identified one TK1-like kinase (FpTK1 and PdTK1, respectively; Table S2). The two identified TK1s amino acid sequences and previously characterized TK1s were aligned and analyzed for their phylogenetic relationship. Interestingly, both FpTK1 and PdTK1, although Gram-negative, grouped together with the usual Gram-positive TK1-like dNKs, rather than with the previously characterized Gram-negative ones (Fig. 1).

Phylogenetic tree of the TK1-like kinases from Gram-negative bacteria Polaribacter sp. MED 152 (PdTK1) and Flavobacterium psychrophilum JIP02/86 (FpTK1) together with previously biochemically characterized TK1s. The following bacterial, plant, and human TK1s were used in the analysis: BaTK1 — Bacillus anthracis TK1 (YP_031418.1), BcTK1 — Bacillus cereus TK1 (NP_834992.), EcTK1 — Escherichia coli TK1 (NP_415754.1), HsTK1 — Homo sapiens TK1 (AAH07986.1), MmTK1 — Mycoplasma mycoides TK1 (CAC85214.1), PmTK1 — Pasteurella multocida TK1 (NP_246173.1), RmTK1 — Rhodothermus marinus TK1 (ACY49479.1), SeTK1 — Salmonella enterica TK1 (NP_455750.1), SlTK1 — Solanum lycopersicum TK1 (AAQ08180), SaTK1 — Staphylococcus aureus TK1 (BD37699.1), SpTK1 — Streptococcus pyogenes TK1 (ABO07417.1), TmTK1 — Thermotoga maritima TK1 (Q9WYN2.1), UpTK1 — Ureaplasma parvum TK1 (NP_078433.1),and UuTK1 — Ureaplasma urealyticum TK1 (Q9PPP5.1). Note that TK1s from both Gram-negative bacteria (FpTK1 and PdTK1) group together with TK1s from Gram-positive bacteria, while some TK1s from Gram-negative bacteria, including EcTK1, PmTK1, SeTK1, SpTK1, and TmTK1, form a separate group. Bootstrap values for each node are shown.

Surprisingly for Gram-negative bacteria, in F. psychrophilum JIP02/86, we identified also one non-TK1 dNK (FpdNK), and in Polaribacter sp. MED 152, we found two non-TK1 dNKs, one of them representing a hybrid between non-TK1 and a sequence encoding HPPK (PdHPPK + dNK) (Table S2). The HPPK, 6-hydroxymethyl-7,8-dihydropterin pyrophosphokinase, catalyzes the attachment of pyrophosphate to 6-hydroxymethyl-7,8-dihydropterin to form 6-hydroxymethyl-7,8-dihydropteridine pyrophosphate. This is the first step in the three-step pathway that leads to 7,8-dihydrofolate. Similar hybrid genes were also found in several other bacteria belonging to Bacteroidetes class (Fig. S1).

All the identified dNKs genes were successfully amplified from genomic DNA (Table S1) and subcloned into the pGEX-2T expression vector. In addition, in order to test the significance of the HPPK domain for the phosphorylating activity of the Polaribacter sp. MED 152 PdHPPK + dNK hybrid, also a recombinant dNK without the HPPK domain was constructed (PdHPPKΔdNK) (Tables S1 and Table S2).

Activity and overexpression of recombinant dNKs

Initially, the substrate specificity of recombinant dNKs was tested in transformed TK1-negative Escherichia coli KY895 crude extracts (Table S2). dNKs phosphorylating activities were tested with all native dNs substrates: dT, deoxyadenosine (dA), deoxyguanosine (dG), and deoxycytidine (dC). All assays were performed at 37 °C, except for FpTK1 and PdTK1. For these two enzymes, it was determined that 21 °C was the optimal temperature to measure their activity. In short, the subcloned TK1s and non-TK1s indeed represented active dNKs; however, the PdHPPK + dNK hybrid showed poor activity with dG, and the shortened recombinant enzyme PdHPPKΔdNK (without HPPK) also showed very low activity with dA, dC, and dG (Table S2). The hybrid proteins were not characterized further, and the function of the hybrid gene is so far unclear.

All recombinant dNKs were expressed in E. coli BL21 and purified using affinity chromatography. The N-terminal GST fusion provided by the pGEX-2T vector was used as the affinity tag. Thrombin was used as a specific protease cleaving the GST tag from the kinase of interest, leaving only two extra amino acids (glycine and serine) at the N terminus. Afterward, pure recombinant protein was eluted from the GSH column. In the case of FpdNK and PddNK, we were not able to remove the GST tag from the dNK of interest; therefore, the whole GST fusion protein was eluted from the column and characterized. Purified dNKs were visualized by denaturing SDS-PAGE and Coomassie staining (Fig. 2). The size of the pure proteins was in reasonable agreement with theoretical molecular weights. The molecular weight for the pure recombinant proteins, FpTK1 and PdTK1, was around 20–25 kDa (Fig. 1a and c), while for the GST fusion proteins, FpdNK + GST and PddNK + GST, it was around 50 kDa, because of the GST part (Fig. 2b and d).

Coomassie-stained SDS-PAGE gels of purified recombinant dNKs. (a) Flavobacterium psychrophilum JIP02/86 TK1 (FpTK1). Lane 1, protein marker (BioLabs); lane 2, crude extract; lane 3, flow-through; lane 4, pure FpTK1. (b) F. psychrophilum JIP02/86 dNK (FpdNK). Lane 1, protein marker (BioLabs); lane 2, crude extract; lane 3, flow-through; lane 4, pure FpdNK + GST (uncleaved fusion protein with bound GST tag). (c) Polaribacter sp. MED152 TK1 (PdTK1). Lane 1, protein marker (BioLabs); lane 2, crude extract; lane 3,flow-through; lanes 4 and 5, pure PdTK1. (d) Polaribacter sp. MED152 dNK (PddNK). Lane 1, protein marker (BioLabs); lane 2, crude extract; lane 3, flow-through; lane 4, pure PddNK + GST (uncleaved fusion protein with bound GST tag).

Steady-state kinetics of recombinant dNKs

Kinetic parameters were determined for purified recombinant PdTK1, PddNK + GST, FpTK1, and FpdNK + GST (Table 1). The activity was measured at 37 °C, except for FpTK1 and PdTK1, which were tested at 21 °C. All enzyme reactions followed classical Michaelis–Menten kinetics. The FpTK1 specifically phosphorylated dT and dU, having dT as the preferred substrate, because the Km for dT was 2.2 µM, which is ∼ 64 times lower than for dU. The apparent maximal velocity (Vm-app) was 5.78 U mg−1 for dT and 4.64 U mg−1 for dU (Table 1, Fig. S2a and Fig. S2b). Similarly, also PdTK1 preferred dT as substrate over dU, with Km for dT being 32 µM, which is ∼ 27 times lower than for dU. The m-app for dT was 3.43 and 2.11 U mg−1 for dU (Table 1, Fig. S2e and f). The catalytic efficiency (Vm-app/Km) was manyfold higher for FpTK1 than for PdTK1, and the activity with dT was higher for both enzymes. This indicates that FpTK1 has higher specificity toward dT and dU (Table 1). We could not detect any significant phosphorylation of dA, dC, or dG by either FpTK1 or PdTK1.

View this table:
Table 1

Kinetic parameters of purified recombinant dNKs from Flavobacterium psychrophilum JIP02/86 and Polaribacter sp. MED 125. All enzyme reactions followed classical Michaelis–Menten kinetics. The measurements were taken twice, except for FpTK1, which was tested three times. The values are expressed as the average of two or three independent experiments, and the error is shown as the variation range. Radio-assay of dNKs activity was performed at 37 °C or at 21 °C. None of the tested recombinant dNKs showed any measurable activity on dG

OrganismPutative dNKSubstrateKm (µM)VmaxApp (U mg−1)VmaxApp/Km
F. psychrophilum JIP02/86FpTK1 (21 °C)dT2.2 ± 0.55.78 ± 1.412.63
dU141 ± 294.64 ± 0.470.033
FpNK (37 °C)dA51 ± 23 ± 0.290.059
dC228 ± 131.97 ± 0.180.0086
Polaribacter sp. MED 152PdTK1 (21 °C)dT32 ± 23.43 ± 0.040.11
dU855 ± 1082.11 ± 0.540.0025
PdNK (37 °C)dA58 ± 85.43 ± 0.280.094
dC174 ± 123.29 ± 0.260.019

The FpdNK was able to phosphorylate both dA and dC, but had dA as the preferred substrate, with the K m for dA being 4.5 times lower than for dC. The catalytic efficiency (Vm-app/Km) was manyfold higher for dA than for dC, indicating that dA was the preferred substrate (Table 1, Fig. S2c and Fig. S2d).

Also PddNK preferred dA as substrate over dC, the catalytic efficiency being almost 6-fold higher for dA than for dC (Table 1, Fig. S2g and Fig. S2h). In short, both non-TK1 kinases were much more specific for dA than for dC, and none of them was able to phosphorylate dG.

Effect of temperature on enzyme activities

Initially, attempts to measure the substrate specificity of pure recombinant PdTK1 and FpTK1 at 37 °C failed; therefore, we decided to determine PdTK1 phosphorylating activity as a function of temperature, by measuring the activity at 500 µM 3H-dT and 2.5 mM ATP, at different temperatures, and with prolonged sampling times. It turned out that the activity of PdTK1 increased with temperature, up to 21 °C, where the highest activity was detected (7.77 ± 1.56 U mg−1); thereafter, the activity decreased with increasing temperature (Fig. 3). Therefore, 21 °C was used for further investigation of PdTK1 and FpTK1. Upon pre-incubating the enzyme at 0 °C for one hour, the obtained activity at 21 °C was considerably lower (0.71 ± 0.04 U mg−1), and after pre-incubation at 37 °C for one hour, the enzyme was irreversibly denatured, because we could not detect any activity at all.

Polaribacter sp. MED152 TK1 (PdTK1) activity as a function of temperature. Radio-assays were performed at 5, 10, 15, 21, 25, 30, and 37 °C with 500 µM 3H-dT as substrate. All measurements were taken four times, except at 30 and 37 °C (three times). The shown activities represent the average ± standard deviation.


One of the approaches to estimate the aquatic bacteria biomass production is the incorporation of 3H-dT into newly synthesized DNA, and it is based on the assumption that all actively growing bacteria can incorporate external dT into DNA (Furhman & Azam, 1982). This also assumes that active dNKs are present in all bacterial cells. In our study, we have examined the genetic and biochemical potential of several aquatic bacteria to phosphorylate native dNs. In two Gram-negative bacteria, F. psychrophilum JIP02/86 and Polaribacter sp. MED 152, we identified TK1-like kinase (FpTK1 and PdTK1, respectively, Table S2), which group together with the already characterized Gram-positive TK1-like dNKs (Fig. 1). The corresponding enzymes followed classical Michaelis–Menten kinetics and were strictly specific for dT and dU (Table 1, Fig. S2a,Fig. S2b,Fig. S2e,Fig. S2f). The kinetics parameters for FpTK1 resemble those of the Gram-positive TK1s from Staphylococcus aureus TK1 (SaTK1) and Bacillus cereus TK1 (BcTK1), which also supports the obtained phylogenetic relationship (Sandrini et al., 2007a, b). On the other hand, PdTK1 is from the kinetic point of view similar to Gram-negative TK1 from Salmonella enterica (SeTK1) (Sandrini et al., 2007a). Surprisingly, while the previously characterized bacterial TK1s are relatively thermostable, both FpTK1 and PdTK1 were not active at 37 °C, and by measuring PdTK1 phosphorylating activity as a function of temperature, it turned out that the activity of PdTK1 increased with temperature up to 21 °C (Fig. 3). When measured at higher temperatures, 25, 30, and 37 °C, the activity decreased over time. Furthermore, when pre-incubating the enzyme at 0 °C for one hour, the measured activity at 21 °C was 10-fold lower, while pre-incubation at 37 °C for one hour resulted in irreversible denaturation. These data can be important for the interpretation of the results obtained in the past studies, when the indirect activity of TK1 from aquatic bacteria was measured at 37 °C (Jeffrey & Paul, 1990). In short, when measuring the activity of bacteria isolated from cold niches by the incorporation of 3H-dT into newly synthesized DNA, one should keep in mind that the activity should be measured at different temperatures.

So far it has been thought that Gram-negative bacteria have only one dNK, TK1, while Gram-positive bacteria seem to have several dNKs (Sandrini et al., 2007a). The FpdNK and PddNK kinases followed classical Michaelis–Menten kinetics and were able to phosphorylate dA and dC; however, both of them had dA as the preferred substrate (Table 1, Fig. S2ac, Fig. S2d,Fig. S2g,Fig. S2h). None of them was able to efficiently phosphorylate dG; therefore, these two enzymes seem to act like dAK from S. aureus dAK (SadAK) or B. cereus dAK (BcdAK) (Sandrini et al., 2007a, b), but with much lower specificity for dC. The substrate preferences could be partially explained by the genome composition. Both bacteria have AT content of approximately 70%; therefore during DNA replication, they need more A and T than G and C. However, it is puzzling why the activity on dG is so low in both species (Table 1 and Fig. S2).

We examined the sequenced genomes from several aquatic bacteria for the genes encoding dNKs, the key enzymes in the salvage of dNs. Database searches revealed that most of the examined aquatic bacteria genomes contained 1–3 genes encoding TK1-like and non-TK1-like dNKs, and even hybrids between dNKs and other open reading frames (Table 2). We show that these bacteria indeed have the potential to phosphorylate dNs. It seems that the occurrence of dNK genes in examined bacteria is sporadic, because large majority of analyzed Alpha- and Gamma-proteobacteria and Firmicutes contained only TK1-like genes; on the other hand, most of the examined Beta-proteobacteria had only genes encoding for non-TK1-like dNKs and some of them did not possess any dNKs genes at all (Table 2). Analyzed bacteria from Bacteroidete class contained both the TK1-like genes and non-TK1-like genes, and most of Bacteroidete also contained a hybrid between putative dNKs and hydroxymethyldihydropterin pyrophosphokinase (HPPK) (Table 2, Fig. S1). Several groups, like Cyanobacteria, Delta-, and Epsilon-proteobacteria, apparently do not have any dNK genes (Table 2).

View this table:
Table 2

Several available sequenced aquatic bacteria genomes (as available in 2011) were searched for genes with homology to the known dNKs using the genome blast service at NCBI. Analyzed bacterial classes and species, together with the dNKs accession numbers, are listed; n.p. stands for not present. Note a variability in the presence and absence of TK1s and non-TK1s

Putative dNKs
CyanobacteriaProchlorococcus marinusn.p.n.p.
Synechococcus sp. CC9311n.p.n.p.
AlphaproteobacteriaSulfitobacter NAS-14.1ZP_00948322.1n.p.
Silicibacter sp. TrichCH4BZP_05742484.1n.p.
Jannaschia sp. CCS1YP_511266.1n.p.
Erythrobacter litoralis HTCC2594YP_4575871n.p.
Candidatus Pelagibacter ubiquen.p.n.p.
Hirischia baltica ATCC49814YP_003058673.1n.p.
Hyphomonas neptunium ATCC51888YP_761824.1n.p.
Magnetospirillum magneticum AMB-1n.p.n.p.
Rhodospirillum centenum SWYP_002298152.1n.p.
Roseobacter denitrificans Och 114YP_683664.1n.p.
Rickettsia rickettsiin.p.n.p.
Rugeria pomeroyi DSS-3YP_166167.1n.p.
Sphingomonas wittichii RW1YP_001264297.1n.p.
BetaproteobacteriaNitrosospira multiformis ATCC 25196n.p.YP_411575.1
Nitrosomonas eutropha C91n.p.YP_748476.1
Methylotenera mobilis JLW8n.p.YP_003049315.1
Janthinobacterium sp. Marseillen.p.YP_001354567.1
Nitrosomonas europaea ATCC 19718n.p.NP_840172.1
Methylovorus sp. SIP3-4n.p.YP_003051935.1
Herminiimonas arsenicoxydansn.p.YP_001100883.1
Thauera sp. MZ1Tn.p.YP_002889742.1
Methylibium petroleiphilum PM1n.p.YP_001022205.1
Bordetella petrii DSM 12804n.p.n.p.
Acidovorax ebreus TPSYn.p.n.p.
Leptothrix cholodnii SP-6n.p.n.p.
Polaromonas naphtalenivorans CJ2n.p.n.p.
Thiomonas intermedia K12n.p.n.p.
Rhodoferax ferrireducens T118YP_516069.1n.p.
GammaproteobacteriaMarinomonas MWYL1YP_001341734.1n.p.
Pseudoalteromonas atlantica T6cYP_660209.1n.p.
Idiomarina loihiensis L2TRYP_154753.1n.p.
Acinetobacter baumanniin.p.n.p.
Alcanivorax borkumensis SK2n.p.YP_692058
Alteromonas macleodiiYP_004426609.1n.p.
Marinobacter auqaeoli VT8n.p.n.p.
Nitrosococcus watsonii C-113n.p.n.p.
Shewanella baltica OS185YP_001367058.1n.p.
Vibrio harveyi ATCC BAA-1116YP_001445038.1n.p.
DeltaproteobacteriaDesulfovibrio desulfuricansn.p.n.p.
Geobacter sp. M21n.p.n.p.
Pelobacter carbinolicusn.p.n.p.
EpsilonproteobacteriaSulforuspirillum deleyianum DSM 6946n.p.n.p.
BacteroidetesMicroscilla marina ATCC 23134ZP_01694268.1ZP_01694275.1
Rhodothermus marinus DSM 4252YP_003291867.1YP_003291613.1
Zunongwangia profunda SM-A87YP_003587106.1YP_003586157.1
Gramella forsetti KT0803YP_863362.1YP_860359.1
Croceibacter atlanticus HTCC2559YP_003715249.1YP_003715952.1
Robiginitalea biformata HTCC2501YP_003194488.1YP_003196852.1
Maribacter sp. HTCC2170YP_003862165.1YP_003860881.1
Chlorobaculum parvum NCIB 8327n.p.n.p.
ActinobacteriaSalinispora arenicola CNS-205YP_001537320.1n.p.
Acidothermus cellulolyticus 11Bn.p.n.p.
FirmicutesOceanobacillus iheyensis HTE831NP_693921.1NP_690937.1
Geobacillus kaustophilus HTA426YP_149233.1n.p.
Halothermothrix orenii H168YP_002509543.1n.p.
Desulfotomaculum ruminis DSM 2154n.p.n.p.
ThermotogaeThermotoga maritima MSB8NP_228211.1n.p.
Thermotoga petrophilla RKU-1YP_001244117.1n.p.
SpriochaetalesSpirochaeta sp. BuddyYP_004249028.1n.p.
PlanctomycetesPirellula staleyi DSM 6068n.p.n.p.
Planctomyces limnophilus DSM 3776n.p.n.p.
Rhodopirellula baltica SH 1NP_868343.1n.p.

In conclusion, we showed that several examined aquatic bacterial genomes possess genes encoding putative dNKs; therefore, these bacteria have a potential to salvage dNs, but the presence of genes is variable and some substrate specificities are missing. It also turned out that a majority of sequenced aquatic Beta-proteobacteria lack TK1-like genes, which means that a whole fraction of the aquatic bacterial community can be overlooked, when measuring bacterial biomass production by the incorporation of external 3H-dT into newly synthesized DNA.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1. Alignment of the sequences representing hybrids between dNK and hydroxymethyldihydropterin pyrophosphokinase (HPPK) found in the sequenced bacterial genomes of Bacteroidetes class.

Fig. S2. Kinetics of purified dNKs from F. psychrophilum JIP02/8 and from Polaribacter sp. MED152.

Tab S1. Primers for PCR amplification of dNKs from genome of F. psychrophilum JIP02/86 (NC_009613) and Polaribacter sp. MED 152 (NZ_AANA00000000).

Tab S2. Genes encoding dNKs in the bacterial genomes of F. psychrophilum JIP02/86 (NC_009613) and Polaribacter sp. MED 152 (NZ_AANA00000000), their abbreviations, accession numbers and plasmids used for overexpression (the Piskur collection in Lund), together with their phosphorylation capacity for the native substrates.

Data S1. Material and methods.


This work was supported by a grant from Swedish Research Council (VR) and a grant from Ministry of Higher Education, Science and Technology of the R Slovenia. The authors thank E. Duchaud and J. Pinhassi for the bacterial genomic DNA. T.T. acknowledges a travel grant from FEMS. A.K. and D.A.L. receive funding from NSF DBI-0743374.


View Abstract