Purine and Pyrimidine Metabolism by Estuarine Bacteria

AQUATIC MICROBIAL ECOLOGY Vol. 42: 215–226, 2006 Published March 29 Aquat Microb Ecol

Purine and pyrimidine metabolism by estuarine bacteria

Gry Mine Berg1,*, Niels O. G. Jørgensen2

1Department of Geophysics, Stanford University, Stanford, California 94305, USA 2Department of Ecology, Royal Veterinary and Agricultural University, Thorvaldsensvej 40, 1871 Frederiksberg, Denmark

ABSTRACT: Nucleotide bases are ubiquitous in living organisms, but the fate of these compounds in natural environments is poorly understood. Here we studied the metabolism of selected purines and pyrimidines in estuarine bacterial assemblages from the Øresund, Denmark. Depletion of nucleotide bases in the incubations was followed by the appearance of urea. The purine guanine and the catabolic intermediates hypoxanthine and xanthine were depleted 2 times faster from seawater incubations than the purine adenine, and 8 to 35 times faster than the pyrimidines thymine, cytosine, and uracil over the course of 141 h. After 48 h, 45 to 60% of guanine-N, hypoxanthine-N and xanthine-N was converted to urea-N while the conversion of adenine and pyrimidines to urea was slower, corresponding to 34% and 19 to 23%, respectively. After 96 h, urea concentrations declined in most of the incubations, indicating hydrolysis of urea by the bacterial populations. Bacterial metabolism of adenine in Øresund water was estimated to contribute up to 10% of the urea pool, but due to the efficient conversion of adenine, the ecosystem importance of adenine degradation to urea production was most likely greater. Growth of bacterial microcolonies on 0.2 µm pore-sized polycarbonate filters floating on natural seawater enriched with individual nucleotide bases varied significantly with substrate enrichment. Despite growth of only a small fraction of bacteria present in the natural assem- blage on the filters, variation in bacterial microcolony biomass explained most of the variation in substrate utilization and urea production among the treatments. Our results suggest that bacterial catabolism of particularly purines and their intermediates, and to lesser extent pyrimidines, is a major process by which the N moiety of natural, heterocyclic bases are converted into urea which is easily assimilated.

KEY WORDS: Purines · Pyrimidines · Urea production · Estuarine bacteria · Microcolony approach

Resale or republication not permitted without written consent of the publisher

INTRODUCTION archaea (Stuer-Lauridsen & Nygaard 1998), protozoa (Hassan & Coombs 1988) and phytoplankton (Shah & Purines and pyrimidines, the nitrogenous bases of Syrett 1982, 1984, Antia et al. 1991, Sarcina & Cassel- nucleotides, are integral components of DNA and ton 1995). RNA. The bases are also constituents of many coen- Given their prevalence in organisms, and as excre- zymes that are involved in energy-carrying reactions, tory products, a high input of purines and pyrimidines in transfer of organic molecules and in oxidation- to the environment is expected to occur. However, reduction reactions. In addition, purines are excreted ambient concentrations of purines and pyrimidines in as waste products by a range of terrestrial and marine aquatic environments are typically below 0.1 to 10 µM, animals (Nicol 1960, Vogels & van der Drift 1976). suggesting they are cycled quickly (Antia et al. 1991). Purines and pyrimidines are nitrogen-rich molecules Microbial purine oxidation involves up to 7 transfor- with 5 and 2 atoms of nitrogen per molecule, respec- mation steps before release of the end product urea tively (Fig. 1) and can serve both as nitrogen (N) and (Fig. 1). Initially, adenine is deaminated to hypoxan- carbon (C) sources for growth in a variety of microor- thine, which is subsequently oxidized to xanthine, uric ganisms, including yeast (LaRue & Spencer 1968), bac- acid, allantoin, and ultimately urea. For each molecule teria (DeMoll & Auffenberg 1993, Schultz et al. 2001), of purine metabolized, 1 molecule of ammonia and 2

*Email: [email protected] © Inter-Research 2006 · www.int-res.com 216 Aquat Microb Ecol 42: 215–226, 2006

molecules of urea are produced. Similarly, aerobic Urea can be an important N source to phytoplankton pyrimidine oxidation results in the production of 1 mol- in coastal environments (Remsen 1971, McCarthy ecule of ammonia and 1 molecule of urea per pyrimi- 1972, Kristiansen 1983, Berg et al. 1997). In addition, dine (Vogels & van der Drift 1976). Therefore, purine urea may sustain a significant portion of the N demand and pyrimidine degradation can serve as a major by estuarine bacteria when other N sources are scarce source urea in nature. (Jørgensen 2006, this issue). Whether a significant portion of the urea observed in coastal systems is produced from purine and pyrimidine degradation depends on Adenine Guanine the prevalence of purine- and pyrimi- NH2 O dine-degrading bacteria, and whether

N N HN N these bacteria hydrolyze the urea produced to ammonium. Addition of N N purines or nucleic acids to anoxic N H2N N H H marine sediments (Therkildsen et al. H O H2O Adenine 2 Guanine 1996) and to seawater and freshwater deaminase deaminase NH3 NH3 bacterial incubations (Berman et al. 1999) leads to an increase in urea con- Hypoxanthine Xanthine centrations, but the relationship be- O O tween purine and pyrimidine degrada- H2O 2H HN N HN N tion and urea production has not been examined, nor has the role of purine Xanthine O N N N N and pyrimidine hydrolysis versus urea dehydroginase H H H hydrolysis by marine bacterial com- H2O Xanthine munities. dehydroginase 2H We performed a series of purine and Uric acid pyrimidine additions to incubations of O natural estuarine bacteria in order to A Adenine Guanine HN NH determine the lability of purines and NH 2 O pyrimidines, and to investigate the po- N O O N N HN N N tential production and consumption of H H urea by the bacteria. In order to inves- O2+2H2O N N Uricase tigate differences in bacterial biomass N H2N N H H CO2+H2O2 with treatment, we used a microcolony epifluorescence technique (Binnerup

Thymine Cytosine Uracil Allantoin B O et al. 1993) where formation of micro- NH O NH2 O colonies on 0.2 µm polycarbonate O CH3 NH O filters, floating on natural seawater HN N H2N—C—HN N H enriched with purines and pyrim- O H2O Allantoinase O N O N N idines, was quantified. This technique H H H demonstrated good agreement be-

O Allantoic acid tween final gross bacterial biomass and the rate of purine and pyrimidine H2N—C—HN O CH—C—OH utilization in natural seawater incubations.

H2N—C—HN O H2O Allantoicase

Glyoxylic acid Urea Ureidoglycolic acid Urea MATERIALS AND METHODS

O O O O OH O O Sampling and preparation of bacte- H—C—C—OH H N—C—NH H N—C—NHCH—C—OH H N—C—NH + 2 2 2 + 2 2 rial cultures. Water for all experiments Ureidoglycolase was collected between April and Fig. 1. Microbial aerobic purine degradation pathway modified from Vogels & August 1998 along the Øresund coast van der Drift (1976). Molecular structure of the purines adenine and guanine, north of Copenhagen, Denmark. and the pyrimidines thymine, cytosine and uracil are shown in inserts A and B Salinity was 12 to 18 ppt and water Berg & Jørgensen: Purine and pyrimidine metabolism 217

temperature was 10 to 19°C. For all experiments, parti- Table 1. Solvent gradient used for determination of nucleotide cles above bacterial size were removed by filtration base concentrations. A = 100 mM ammonium acetate, B = through Whatman GF/F filters with an estimated nom- methanol. Gradient curves: 0 = no change, 6 = linear change, 11 = immediate change inal pore size of 0.7 µm. Uptake of purines, pyrimidines and purine degradation products. Bacterial uptake of the purines ade- Time (min) %A %B Curve nine and guanine, their degradation products hypo- 0 100 0 0 xanthine and xanthine, and the pyrimidines cytosine, 9.5 96 4 6 uracil and thymine were tested by the addition of the 13.5 96 4 6 substrates to duplicate samples of 20 ml GF/F-filtered 19.0 85 15 6 water. Final concentrations of nucleotide bases were 2 26.5 85 15 6 –1 27.0 50 50 6 to 2.5 µM (4 to 12 µg-atom N l ). The samples were 30.5 100 0 11 incubated at 15 to 20°C for 141 h in the dark on a shak- 46.0 100 0 11 ing table with gentle rotation. At 0, 19, 48, 98 and 141 h, samples for bacterial counts (2 ml) and for HPLC analysis of the concentration of bases (1 ml; filtered through 0.2 µm pore size filters and frozen until analy- acetonitrile according to the recommendations of the sis) were taken. Samples for bacterial abundance were manufacturer, and collected. After the acetonitrile was preserved with 0.2-µm-filtered and buffered formalde- evaporated, the samples were dissolved in 500 µl hyde (pH 8.0; adjusted with borate). The bacterial Milli-Q water. Subsamples were analyzed by HPLC as abundance was determined by acridine orange stain- above or were radioassayed. The average recovery of ing according to Hobbie et al. (1977). Bacteria were [14C]-adenine was determined to be 11% (range 10 to counted at 10 different sites on the filters and at each 13%; results not shown). The relatively low recovery of site between 30 and 100 bacteria were counted. adenine most likely indicates that other organic Concentrations of the nucleotide bases were deter- molecules also absorbed to the C-18 material. The mined by UV absorption at 254 nM after separation by measured recovery was used for calculation of in situ reversed phase HPLC according to the procedure by adenine concentrations. Waters Corporation. The chromatographic equipment Dissolved nitrogen determinations. Changes in con- – consisted of the following Waters equipment: two 510 centrations of urea, nitrate (NO3 ) and ammonium + pumps, a 717 WISP autoinjector, a 2687 UV/VIS detec- (NH4 ) were also determined in the incubations with tor and a 4.6 × 250 mm Waters Symmetry C-18 column additions of nucleotides. Urea concentration was (5 µm particle size). Millenium32 software was used for determined by the monoxime method according to – control of the system and for manipulation of the Price & Harrison (1987). Concentrations of NO3 and + chromatograms. Eluents were 100 mM ammonium NH4 were determined by standard autoanalyzer tech- acetate adjusted to pH 6.0 with (A) acetic acid and nique. (B) methanol. Column temperature was 30°C and the Bacterial adenine metabolism: uptake, respiration flow rate was 1 ml min–1. The solvent gradient was and conversion into urea. Bacterial utilization of ade- according to Table 1. The detection limit of the bases nine and its conversion to urea was investigated under varied from 50 to 100 nM depending on the elution ambient adenine concentration by trace addition of time (the fastest eluting peaks had the lowest detection [14C]-adenine to 20 ml GF/F-filtered water samples. limit). The sample injection volume was 100 µl. Each bottle received 0.045 µCi [U-14C]-adenine (spe- To investigate the production of urea from adenine cific activity of 287 mCi mmol–1, Amersham Bio- at naturally occurring adenine concentrations, in situ sciences), corresponding to a final concentration of adenine concentrations were measured by HPLC as 7.95 nM adenine. At 3, 24, 48, 96 and 168 h, (1) incor- above with the following modifications to improve poration of adenine into bacterial biomass (collection detection: adenine in Øresund water was pre-concen- of bacteria on 0.2 µm filters), (2) respiration of adenine 14 trated using extraction cartridges made of C-18 mater- (collection of [ C]-CO2), and (3) conversion of adenine ial (Waters Oasis HLB Plus). The cartridges were con- into urea (urease mediated hydrolysis of [14C]-urea to 14 ditioned with 5 ml Milli-Q water after which 40 ml [ C]-CO2) were determined in triplicate. Details of the sample was slowly loaded onto the cartridge with a different procedures follow below. peristaltic pump (1 ml min–1). A trace amount of [14C]- (1) Incorporation of adenine was determined at each adenine was added to the 40 ml water sample and was sampling time by addition of formaldehyde to a final used to determine recovery of adenine on the car- concentration of 2% to stop biological activity. The tridge. The cartridges were washed with 10 ml Milli-Q water was subsequently filtered through 0.2 µm mem- water after which the content was slowly eluted with brane filters and the filters were radioassayed. 218 Aquat Microb Ecol 42: 215–226, 2006

(2) Respiration of adenine was determined by from urea, we assumed a 22% recovery when calculat- mounting rubber membranes with CO2 traps (1 ml ing the total urea production from adenine. Eppendorf-type vial with an accordion-folded wick Bacterial growth on purines and pyrimidines inves- attached to the membrane with a 0.5 mm 20 fishing tigated through a microcolony approach. Uptake of nylon line) on the bottles. At each sampling time, 3 purines, their degradation products and pyrimidines replicate bottles received 500 µl 10% H3PO4 to release was also investigated through a microcolony technique 14 [ C]-CO2. The wicks in the CO2 traps were wetted according to Binnerup et al. (1993). The principle of with 100 µl phenylethylamine with a hypodermic nee- this technique is that bacteria are filtered onto polycar- dle and incubated for 1 h on a shaking table to adsorb bonate filters that float on top of a medium enriched the CO2 produced. The traps were subsequently trans- with the substrate of interest. Only bacteria capable of ferred to scintillation vials and radioassayed after addi- utilizing the substrate(s) in the medium will divide and 14 tion of scintillation cocktail. Linearity of the [ C]-CO2 form new cells. In addition to selection for substrate production from degradation of [14C]-adenine over utilization, the technique also separates bacterial time was tested in preliminary studies. The linearity growth from protozoan grazing if natural water and experiments demonstrated that there was a linear soil samples are studied. At the end of the incubation, 14 development in [ C]-CO2 production in samples col- the number of colonies, the number of bacteria within lected 0.5, 1, 2, 3 and 4 h after [14C]-adenine addition, the colonies, and the size of the bacteria can be enu- indicating that adenine taken up by the bacteria was merated. The bacterial biomass is expected to repre- immediately metabolized (results not shown). sent the quality of the substrate to the bacteria. (3) Conversion of adenine to urea was determined at In the present investigation, 500 µl GF/F-filtered each sampling time by addition of 50 µl concentrated Øresund water (ca. 2 × 105 bacterial cells) was filtered 14 HCl to stop biological activity and to release [ C]-CO2. onto 0.2 µm pore size and 25 mm diameter Nuclepore Preliminary tests demonstrated that a small percent- polycarbonate filters and transferred to 9 cm plastic 14 14 age of [ C]-CO2 from respiration of C-adenine Petri dishes with 50 ml of 0.2 µm filtered Øresund remained in the samples after acidification. Since all water. Each dish was enriched with 2.5 µM of one 14 [ C]-CO2 in the samples in the following urease treat- purine, degradation product, pyrimidine or urea as in ment would be assumed to originate from [14C]-urea, the initial incubation experiment. After 7 d, the filters 14 traces of [ C]-CO2 were removed from the samples by were gently blotted, transferred to a solution of 0.015% freeze-drying. The samples were re-dissolved in the acridine orange in a 2.5 cm Petri dish, and allowed to serum bottles in 20 ml 10 mM potassium phosphate float for 5 min. After 5 min, filters were transferred 2 buffer with 10 mM lithium chloride and 1 mM EDTA additional times to dishes with Milli-Q water to rinse (pH 8.2) prior to addition of 10 units of urease enzyme. out excess acridine orange. The filters were inspected This buffer was recommended by Sigma-Aldrich to by epifluorescence technique (Zeiss Axioplan micro- optimize the enzyme reaction. The serum bottles were scope) using low magnification to identify presence closed with membranes with attached CO2 traps. Dur- and numbers of microcolonies on the filters. Individual ing a 2.5 h incubation period, urease hydrolyzed [14C]- colonies were recorded by a Quantix KAF 1400-G2 14 labelled and unlabelled urea to [ C]-CO2 and CO2 CCD camera (Photometrics) at 100× magnification + (and NH4 ). Finally, 1.5 ml H3PO4 was added to each (Fig. 2A). All microcolonies on each filter were sample and the wicks were wetted with 100 µl recorded. Number and area of individual bacteria in phenylethylamine to absorb the CO2 produced. In situ the microcolonies were subsequently determined urea concentrations were determined as described using the Image-Pro Plus software (v. 4.5; Media above. Cybernetics). In some instances bacteria in the micro-

The efficiency of the applied CO2 traps for collection colonies were embedded in an organic matrix that was 14 14 of [ C]-CO2 from [ C]-urea after treatment with ure- slightly stained with acridine orange (Fig. 2B). Since ase was tested for different urea concentrations. Water this stained, organic matrix interfered with the digital samples of 20 ml received combinations of unlabelled registration of number and size of the bacteria, the urea at 0.5 to 4.0 µM urea and [14C]-urea at concentra- matrix was removed with Image-Pro Plus. With the tions of 0.01 to 0.5 µCi per sample (specific activity was software a background image without bacteria was 56 mCi mmol–1; Perkin Elmer). The recovery of [14C]- produced by erosion (11 × 11 pixels, 3 passes) and dila-

CO2 varied from 19 to 24% (mean of 22%; results not tion (11 × 11 pixels, 3 passes) of the original image. The 14 shown). When tested with [H2 CO3] (ampoules for background image was subtracted from the original [14C] primary measurements), the phenylethylamine image, leaving only the bacteria in the image, as can 14 traps had a recovery of 97% [ C]-CO2 after acidifica- be seen in Fig. 2C. Subsequently, the number of bacte- tion (results not shown). Since we were unable to ria in the microcolonies was quantified using Image- 14 determine the reason for the low recovery in [ C]-CO2 Pro Plus. Berg & Jørgensen: Purine and pyrimidine metabolism 219

RESULTS thine were depleted from the medium within the first 48 h of the experiment, whereas adenine was depleted Uptake of purines, pyrimidines and production within 96 h. Uracil appeared to be the most bioavail- of urea able pyrimidine and was depleted from the medium after 141 h. Thymine and cytosine concentrations Purines and their intermediates were taken up faster decreased 26 and 46%, respectively, over the course of than pyrimidines over the course of 141 h (Fig. 3A). the experiment (Fig. 3A). Concomitant with changes in Guanine and the intermediates hypoxanthine and xan- purine concentrations, changes in concentrations of + – NH4 and NO3 were monitored (data not + shown). During the initial 48 h, NH4 declined from 9 ± 1 to 3 ± 2 µM and remained at that level during the rest of the incubation. Initial and final concen- – trations of NO3 were similar, 28 ± 2 and 31 ± 2 µM, respectively (mean concentrations ± 1 SD). The additions of purines and pyrimidines most likely exceeded natural concentrations in Øresund water several-fold, as the concentration of adenine varied between 0.015 and 0.053 µM in April when the water was collected. Consistent with the pattern in purine utilization, more urea was produced in the incubations with guanine, xanthine and hypoxanthine compared with the other incubations (Fig. 3B). In the first 48 h of the time course, urea production corresponded to 60, 46 and 45% of xanthine, hypoxanthine and guanine additions, respectively (based on N content in the bases and in urea) (Fig. 3A,B). In contrast, urea production equalled 19 to 23% of the pyrimidine additions in the same period. The urea concentration in the adenine treatment corresponded to 34% of the adenine addition, intermediate between that in the treatments with the other purines and pyrimidines. It was assumed that 2 urea molecules were produced for each purine molecule, and that 1 urea molecule was produced for each pyrimidine molecule, when calculating the efficiency of purine or pyrimidine conversion to urea (Fig. 1). After 48 h, urea concentrations decreased or increased at a slower rate in 3 of the purine incubations (guanine, hypoxanthine and adenine) suggesting that urea was con- sumed simultaneously with its production (Fig. 3B). The highest urea concen- Fig. 2. (A) Example of an acridine orange stained microcolony recorded at trations of 5.2 and 4.0 µM occurred at × 100 magnification. (B) Example of bacterial cells embedded in an organic 96 h in the xanthine and guanine incuba- matrix in the microcolony. (C) Image manipulation with Image-Pro Plus software to remove the organic matrix background from the bacterial cells in (B). tions, respectively, and corresponded to The background image was subtracted from the original image, leaving only 84% (xanthine) and 80% (guanine) of the the bacteria in the image purine additions. In a parallel incubation 220 Aquat Microb Ecol 42: 215–226, 2006

with untreated Øresund water, the urea concentration was filtered seawater without nutrient addition. The was 0.35 ± 0.15 µM and no increase or decrease in con- urea treatment supported less bacterial growth than centration was observed (data not shown). the purine and pyrimidine treatments (Fig. 3C). By The addition of purines, pyrimidines and urea stim- 48 h, bacterial abundance in all the treatments was ulated bacterial growth relative to the control, which on average 52 ± 17% higher than in the control. By 96 h, the xanthine and urea treatments had 10 ± 3% lower bacterial densities than the control, and Adenine 3.0 Guanine thymine, hypoxanthine and guanine treatments had Hypoxanthine 35 ± 3% higher densities compared to the control A Xanthine 2.5 Cytosine (Fig. 3C). By 141 h, there were no further increases in Thymine Uracil bacterial abundance in the different treatments. At 2.0 this point, we observed the presence of nanoflagellates in some treatments. Their densities were not 1.5 quantified but could have been a factor in the increased variability among the treatments leading 1.0 up to 141 h. 0.5 Purine or pyrimidine (µM) pyrimidine or Purine

0 Bacterial adenine incorporation and conversion 6 to urea B 5 The [14C]-adenine incubations were made to obtain conservative estimates of the in situ rates of urea pro- 4 duction due to bacterial decomposition of purines. 3 Adenine was chosen because we could not detect concentrations of other purines or pyrimidines with the Urea (µM) Urea 2 present HPLC method. [14C]-labeled adenine was added at trace level to incubations with GF/F filtered 1 seawater (nominal pore size 0.7 µm) and followed for 169 h. In the first 49 h, approximately 50% of the 0 added [14C]-adenine isotope was recovered in the bac- Control terial biomass and only a fraction (<5%) of [14C] was Urea C 8 Adenine incorporated into urea (Fig. 4). By 96 h, the proportion Guanine of [14C] in the bacteria had declined to 29%, while 34% 7 Hypoxanthine Xanthine of the [14C] was recovered in urea. From 96 to 169 h, Cytosine 14 6 Thymine the incorporation of [ C] slightly increased while the Uracil 14 6-1 proportion of [ C] in urea declined to 14%. Respira- 5 tion of adenine increased throughout the time course 4 with 50% of the added adenine having been respired by 169 h. Recovery of the added [14C]-adenine isotope 3 in bacterial biomass, CO and urea was 75% at 25 and

Bacteria x 10 ml Bacteria x 2 49 h, and increased to 97 and 104% at 96 and 169 h, 2 respectively. The conversion efficiency of adenine-C to 1 urea-C in this experiment was 34%, corresponding to a conversion efficiency of adenine-N to urea-N of 68% 0 020406080100120140 (2 urea molecules are produced from each adenine molecule). In a similar, preliminary experiment, 28% Time (h) of the added [14C]-adenine-N was converted to urea-N Fig. 3. (A) Depletion of adenine, guanine, hypoxanthine, xanthine, urea, cytosine, thymine, and uracil in cultures of Øre- (data not shown). This was comparable to the N con- sund water, filtered through Whatman GF/F-filters (0.7 µm version efficiency of 34% in the purine uptake experi- pore size). The cultures were spiked with free bases to final ment (Fig. 3A,B). Relative to the ambient urea and ade- concentrations of ca. 2.5 µM. GF/F-filtered Øresund water nine concentrations of about 350 nM and 15 to 53 nM, without additions served as a control. (B) Production of urea in the cultures. (C) Acridine orange direct counts of bacteria respectively, in Øresund water, this means that 3 to in the cultures as described above. Error bars indicate 10% of the urea pool in the water might have origi- deviation from the mean (n = 2) nated from adenine alone. Berg & Jørgensen: Purine and pyrimidine metabolism 221

and in the control, the bacterial size was 120 Respiration of [ 14 C]-adenine to 140 pixels (difference in size among these 14 60 [ C]-adenine in bacterial biomass treatments not significant; t-test, p> 0.05). In 14 Produced [ C]-urea contrast, the bacteria in the adenine treat- 104% 74% 76% ment were 26% larger in size relative to the 50 control (significantly different from the con-

14 trol and cytosine; -test, p < 0.05). The differ- 97% t 40 ence in size suggested that bacteria metabolizing adenine tended to produce larger cells, whereas bacteria metabolizing gua- 30 nine and the purine intermediates divided faster and therefore produced smaller cells. As a result, total bacterial biomass (bacterial 20 area × bacterial abundance) demonstrated that growth on adenine was interme- 10 diate between the control-urea-pyrimidines group and the guanine-xanthine-hypoxan- 0.9% thine group (Fig. 5D). 0

Adenine transformation (% of added [ C]-adenine) Microcolony bacterial biomass agreed 3 25 49 96 169 well with the uptake of purines and pyrim- Time (h) idines (Fig. 3A) and with the production of urea (Fig. 3B) measured independently 14 Fig. 4. Transformation of [ C]-adenine added at trace level to Whatman (Fig. 6A,B). Bacterial biomass was slightly GF/F filtered seawater, as percent of the added adenine, into [14C]-urea, 14 14 better correlated with purine and pyrimi- bacterial biomass and [ C]-CO2. Recovery of the added [ C] isotope 2 in urea, biomass and CO2 is indicated over the columns. Mean ± 1 SD dine uptake (r = 0.92) than with urea pro- (n = 3) shown duction (r2 = 0.87).

Microcolony growth DISCUSSION

The number of bacterial microcolonies on each of the Although purine and pyrimidine bases are ubiqui- polycarbonate filters was 53 ± 5 in the control, adenine tous components of all living material, almost no quan- and guanine treatments (Fig. 5A). In the hypoxanthine titative data exist on concentrations and fate of the and xanthine treatments, 25 and 52% more colonies bases in marine environments. Similarly, the signifi- were found than in the control. In contrast, the number cance of purines and pyrimidines to the production of of microcolonies in the uracil, urea, cytosine and urea in natural environments has received little atten- thymine treatments was 34 to 71% lower than in the tion in the literature, even though urea is an important control. N source to phytoplankton compared with inorganic + – A more pronounced difference among the treatments nitrogen such as NH4 and NO3 (McCarthy 1972, was evident in the total number of bacteria (i.e. sum of Berman & Bronk 2003). Our results suggest significant bacteria in colonies). Here, the number increased 420, differences in the lability of purines and pyrimidines to 480 and 540% in the hypoxanthine, guanine, and xan- marine bacterial populations, and we further demon- thine treatments, respectively, relative to the control strate a direct link between purine and pyrimidine (Fig. 5B). In contrast, the urea, adenine and the pyrimi- decomposition and urea production. dine treatments demonstrated no difference in total bacterial number relative to the control. It appeared that hypoxanthine, guanine, and xanthine sustained a Degradation of purines higher growth rate than the other substrates and the control, which also influenced the size of the bacteria in Assuming that the rates of purine and pyrimidine the microcolonies. The bacteria in the fast-growing mi- uptake measured in the present investigation approxi- crocolonies (hypoxanthine, guanine, and xanthine) mate the true potential of natural marine bacterial pop- were on average 32 ± 5% smaller in size relative to the ulations, guanine is taken up 8 to 35 times faster than bacteria in the control (Fig. 5C). The bacteria in the the pyrimidines thymine, cytosine and uracil, and fast-growing colonies were similar in size (t-test, p > twice as fast as adenine. The reason for adenine being 0.05). In the cytosine, thymine and uracil enrichments, metabolized significantly slower compared to guanine 222 Aquat Microb Ecol 42: 215–226, 2006

100 7x103

A 6x103 B 80 5x103

60 4x103

3 40 3x10 2x103 20 1x103 Total number of bacteria number of Total Number of microcolonies 0 0

-1 180 3

-1 600x10 C D 150 500x103

120 400x103

90 300x103

60 200x103

30 100x103

0 filter ) biomass (pixels Total 0

Bacterial size (pixels bacteria ) size (pixels Bacterial l e ne ci hine Urea mineUracil ani thineUrea Ura Control Controldenine AdenineGuanin Xanthine CytosineThy A Gu Xan CytosineThymine Hypoxanthine Hypoxant

Fig. 5. Growth of microcolonies on polycarbonate filters floating on top of 0.2 µm filtered Øresund seawater spiked with 2.5 µM adenine, guanine, hypoxanthine, xanthine, urea, cytosine, thymine, uracil and a control with no addition. (A) Number of microcolonies per 25 mm filter. (B) Total number of bacteria in microcolonies per filter. (C) Size of bacterial cells (bacterial area in terms of recorded numbers of pixels per bacteria). (D) Total bacterial biomass (bacterial area × bacterial abundance) per filter by marine bacterial populations is intriguing and Ewing 1992). The prevalence of guanine and hypoxan- unclear. Both purines contain 5 atoms of nitrogen and thine excretion by ciliates and other zooplankton in can be deaminated to the intermediates hypoxanthine marine environments is not known, but the high labil- and xanthine via adenine deaminase and guanine ity of guanine and hypoxanthine to marine bacteria deaminase, respectively (Fig. 1). may well reflect its rate of production by higher trophic Many bacteria can degrade adenine, but some levels and its availability in the water column. Consis- eubacteria, among these actinomycetes, lack adenine tent with this, investigations of purine utilization by deaminase and have a low or no degradation of marine phytoplankton demonstrate that guanine and adenine (Watanabe & Ohe 1972, Vitols et al. 1974, hypoxanthine are preferentially assimilated to adenine Vogels & van der Drift 1976). Microorganisms without in several different species (Antia et al. 1975, Shah & adenine deaminase may produce hypoxanthine from Syrett 1982, 1984). adenine by synthesizing adenosine which is converted to inosine and hypoxanthine (LeChevalier et al. 1982). In other microorganisms, adenine can inhibit purine Degradation of pyrimidines utilization in general (Eschmann & Kaltwasser 1980). These metabolic differences may be related to the The pyrimidines thymine, uracil and cytosine were availability of the different purines in the habitat of the degraded even slower than adenine. In the literature, microorganism. For example, guanine and hypoxan- few studies have focused on bacterial uptake and uti- thine are produced as excretion products by a marine lization of pyrimidines. According to Vogels & van der ciliate (Soldo et al. 1978) and guanine is a major com- Drift (1976) some bacteria possess reductive pathways + ponent of fish skin and scales (Nicol 1960, Staley & for degradation of pyrimidines into NH4 and amino Berg & Jørgensen: Purine and pyrimidine metabolism 223

acid derivatives (β-alanine and β-aminoisobutyric by the bacteria. Nevertheless, hydrolysis rates were acid), but the reactions are energy-demanding and slow enough for urea to accumulate in all the treat- require NADPH. Alternatively, other bacteria utilize ments, suggesting that urea is less attractive to bac- oxidative pathways in which pyrimidines are degraded teria than nucleotide bases. Consistent with this, + to NH4 , urea and organic acids, but only in the microcolony growth on urea-enriched seawater did + absence of NH4 . The reason for the low consumption not differ from growth in the control. Cho et al. of pyrimidines in our experiments is not clear. Possibly, (1996) found rates of bacterial urea production to be + the ambient concentrations of NH4 (3 to 9 µM) and 2 orders of magnitude higher than the bacterial urea – NO3 (30 µM) lowered the activity of pyrimidine decomposition rates in coastal surface waters. These degrading enzymes as observed for cytosine deami- data suggest that nucleotide bases may turn over nase in Escherichia coli (Ban et al. 1972). faster than urea in marine environments, and that bacteria may be net producers of urea. Studies of N utilization by estuarine bacteria show that uptake of Production of urea urea usually is lower than uptake of free amino acids + and NH4 , except when these components are Urea production mirrored the uptake of purines and depleted (Jørgensen 2006). pyrimidines in that it occurred faster in the guanine, The slow utilization of urea by bacteria suggests that hypoxanthine and xanthine treatments, compared other trophic levels benefit from bacterial urea produc- with the cytosine, thymine, and uracil treatments. A tion. In the sea, phytoplankton uptake typically con- time lag in urea production was observed in the ade- trols the community urea consumption (Tamminen & nine-enriched treatments. In the [14C]-adenine incu- Irmisch 1996). During summer months, when sources bations, the production of [14C]-urea peaked at 96 h, of dissolved inorganic N are scarce in many marine after the bacterial density began declining (data not habitats, urea may supply upwards of 28% of the shown). A time lag in urea production coinciding with phytoplankton N demand (McCarthy 1972, Kristiansen late exponential bacterial growth phase was also 1983, Berg et al. 1997), and recent reports suggest that noted in Therkildsen et al. (1996) where no exoge- urea may be an especially important N source for nous purine was added. In that study, production of cyanobacteria, such as Synechococcus (Berman & urea was attributed to an increased degradation of Bronk 2003). intracellular RNA, occurring immediately after carbon Historical data on community urea demand for the exhaustion. This was also observed by Mason & Egli Baltic Sea region, including the Danish sea coast, (1993). In addition to the time lag, the efficiency of allowed us to evaluate if the urea production rates urea production was low (28 to 68%), suggesting that observed in our experiments could contribute signifi- adenine-N was used for purposes other than urea pro- cantly to the community N demand in this region. In duction. Urea production could be reduced by recy- summer, community urea uptake rates range from 50 cling adenine for the production of nucleotides or ATP to 350 nmol urea-N l–1 h–1 (Sörenson & Sahlsten 1987, (Karl & Winn 1984), and in salvage pathways, as Berg et al. 2001). Our enrichment cultures demon- observed in some Archaea (Stuer-Lauridsen & strated that the estuarine bacterial populations had the Nygaard 1998). The rate of adenine incorporation into potential to produce urea in this range (urea produc- nucleotides in marine environments is on the order of tion from degradation of both adenine and guanine 0.025 nmol l–1 h–1 (Karl & Bossard 1985). This is very was 150 nmol urea-N l–1 h–1; Table 2). We hypothesize low compared to the adenine uptake rate of 129 nmol that the low concentrations of purines in natural waters l–1 h–1 measured in the present investigation (Table 1), are caused by a fast uptake and conversion to urea by suggesting this was not a major process in our experi- bacterial assemblages. The fact that guanine and ment. Other processes, possibly urea uptake, could hypoxanthine were taken up more rapidly relative to have resulted in the incomplete conversion of ade- adenine, the only detectable purine in seawater, sug- nine-N into urea-N. gests they cycle quicker than adenine. It should be Incomplete conversion of purine to urea or con- mentioned, however, that we optimized the present sumption by bacteria was evident even in the enrich- approach (retention on C-18 cartridge) for detection of ments with the greatest urea production rates, xan- adenine in seawater and that modification of the tech- thine and guanine, where the conversion efficiency nique may have allowed detection of other purines was approximately 16 to 20% below the theoretical (and pyrimidines). Except for concentrations of free maximum efficiency (based on Fig. 1). Irregular adenine and uracil, 0.8 to 13 and 1.8 to 6 µM, respec- decreases in urea concentrations towards the latter tively measured in the Gulf of Mexico by a biological half of the incubation period in 5 out of 7 treatments assay (Litchfield & Hood 1966), we have not been able indicated that urea was hydrolyzed to some extent to find published data on purines or adenines in nat- 224 Aquat Microb Ecol 42: 215–226, 2006

ural waters. The concentration of adenine in Øresund assume that the apparent low ambient concentrations (range 0.015 to 0.053 µM) is lower than the range of purines and pyrimidines reflect a high microbial found by Litchfield & Hood (1966), though the 2 analyt- potential for degradation of these compounds. ical methods may not be directly comparable. We

Microcolony growth Table 2. Purine and pyrimidine uptake rates and urea production rates based on changes in concentration over time in Fig. 4 The formation of microcolonies on the filters floating on water enriched with urea, purines, pyrimidines and Uptake ratea Urea production rateb intermediates agreed well with the bacterial capacity (nmol N l–1 h–1) (nmol N l–1 h–1) for uptake of these compounds (Figs. 3, 5 & 6). Al- though the number of colonies on the filters explained Adenine 129 50 slightly less of the variation in substrate utilization (r2 = Guanine 264 100 0.87) compared with total bacterial biomass (r2 = 0.92, Hypoxanthine 210 130 Xanthine 207 90 Fig. 6), our microcolony observations agree well with Cytosine 24 20 other microcolony studies in which the number of Thymine 7 20 colonies was found to represent the substrate utiliza- Uracil 30 10 tion (Højberg et al. 1997). aAdenine uptake based on 96 h; guanine, hypoxanthine The present correlation between substrate utiliza- and xanthine uptake over 48 h; cytosine, thymine and tion and colony density and/or biomass, may have uracil uptake over 141 h bUrea production based on the first 48 h of incubation due been influenced by the relative number of micro- 5 to concurrent uptake later in the time course colonies produced. Initially ca. 2 × 10 bacteria were added to each filter, but only 17 to 88 of these bacteria produced microcolonies, assuming that each colony originated from one 300

-1 bacterium. The low number of viable cells on the filters may have been -1 250 caused by desiccation of the bacteria, 200 as the filters remained uncovered when floating in the petri-dishes. 150 To circumvent desiccation, silicone- embedding of bacteria on top of the fil- 100 ters has been applied in other microcolony studies (Højberg et al. 1997). 50 Regression (r 2 = 0.924) ) Further, the small number of viable

Purine uptake (nmol N l h ) h N l (nmol uptake Purine 0 cells may have been caused slow diffusion of nutrients through the filter pores, effectively resulting in nutrient 140 Adenine -1 Guanine limitation of the cells. Using a more -1 120 Hypoxanthine permeable filter that allows for greater Xanthine 100 diffusion of substrates may circumvent Cytosine this problem. Fry & Zia (1982) found Thymine 80 Uracil that 10 to 20% of freshwater bacteria 60 (based on epifluorescent microscopy counts) were able to form micro- 40 colonies on agar slides, and only when the medium was enriched with a mix- 20 Regression (r 2 = 0.866) ture of casein, peptone and starch. 0 Urea production (nmol N l h ) h N l (nmol production Urea This suggests that even under com- 100x103 200x103 300x103 400x103 500x103 600x103 plete nutrient sufficiency another fac- Bacterial biomass (pixels) tor limits growth. Despite the potential limitations of the microcolony ap- Fig. 6. (A) Purine and pyrimidine uptake as a function of total bacterial biomass in microcolonies (combination of Figs. 3A & 5D): y = 6 × 10–7 x – 0.0627; p < 0.01. proach, the agreement between the (B) Urea production as a function of total bacterial biomass in microcolonies microcolony biomass index and the (combination of Figs. 3B & 5D): y = 2 × 10–7 x – 0.0155; p < 0.01 uptake of purines and pyrimidines Berg & Jørgensen: Purine and pyrimidine metabolism 225

suggests that the microcolony formation mirrors a true Acknowledgements. We thank R. E. Jensen for skilful techni- metabolic potential among estuarine bacteria. cal assistance, M. Hansen for help with the image software A disadvantage of the present microcolony biomass in- and 2 referees for valuable comments on the manuscript. This study was supported by grants from the Danish Strategic dex is the time-consuming computer manipulation of the Environmental Program II and the Danish Natural Sciences individual microcolonies, especially when the micro- Research Council to N.O.G.J. colony staining includes extracellular products, as in our case. The present digital manipulation of the micro- LITERATURE CITED photographs to remove extracellular material could also have influenced the actual cell size in some cases. Appli- Antia NJ, Harrison PJ, Oliveira L (1991) The role of dissolved cation of more DNA-specific fluorochromes than acri- organic nitrogen in phytoplankton nutrition, cell biology dine orange will undoubtedly minimize the problem of and ecology. Phycologia 30:1–89 interference of extracellular material. Antia NJ, Berland BR, Bonin DJ, Maestrini SY (1975) Compar- ative evaluation of certain organic and inorganic sources Our observation of a grouping in bacterial cell size in of nitrogen for phototrophic growth of marine microalgae. microcolonies, ranging from small cells (<100 pixels; J Mar Biol Assoc UK 55:519–533 guanine, hypoxanthine and xanthine), intermediate Ban J, Vitale L, Kos E (1972) Thymine and uracil catabolism in cells (120 to 140 pixels; control, urea and pyrimidines), Escherichia coli. J Gen Microbiol 73:267–272 Berg GM, Glibert PM, Lomas MW, Burford MA (1997) to large cells (>160 pixels; adenine), on the different Organic nitrogen uptake and growth by the chrysophyte media may indicate that different metabolic types of Aureococcus anophagefferens during a brown tide event. bacteria were produced. The present coincidence of Mar Biol 129:377–387 small cells and a fast purine metabolism may reflect Berg GM, Glibert PM, Jørgensen NOG, Balode M, Purina I that enrichment with guanine, hypoxanthine and xan- (2001) Variability in inorganic and organic nitrogen uptake associated with riverine nutrient input in the Gulf thine led to selection of bacteria with a small cell size, of Riga, Baltic Sea. Estuaries 24:204–214 in contrast to bacteria with a larger size growing on the Berman T, Bechemin C, Maestrini SY (1999) Release of other media. Molecular analysis of bacteria in the ammonium and urea from dissolved organic nitrogen in microcolonies would have allowed a more precise aquatic ecosystems. Aquat Microb Ecol 16:295–302 Berman T, Bronk DA (2003) Dissolved organic nitrogen: a indication of genotypic relationships among bacteria dynamic participant in aquatic ecosystems. Aquat Microb growing on the different media. Ecol 31:279–305 An advantage of the microcolony technique, in addi- Binnerup SJ, Jensen DF, Thordal-Christensen H, Sørensen J tion to isolating bacteria that can grow on a specific (1993) Detection of viable but non-culturable Pseudo- substrate, is the ability to physically separate bacteria monas fluorescens DF57 in soil using a microcolony epifluorescence technique. FEMS Microbiol Ecol 12:97–105 growing on the filters from their potential grazers. This Cho BC, Park MG, Shim JH, Azam F (1996) Significance of may be the reason for the good agreement between bacteria in urea dynamics in coastal surface waters. Mar the microcolony biomass and rates of substrate uptake, Ecol Prog Ser 142:19–26 as compared to the poorer agreement between bacter- DeMoll E, Auffenberg T (1993) Purine metabolism in Methanococcus vannielii. J Bacteriol 175:5754–5761 ial abundance in the enriched incubation flasks and Eschmann K, Kaltwasser H (1980) Inhibition of purine utiliza- substrate uptake. Nanoflagellates were observed in tion by adenine in Alcaligenes eutrophus H16. Arch some treatments at 141 h, but it cannot be excluded Microbiol 125:29–34 that grazing protists occurred earlier in some of the Fry JC, Zia T (1982) Viability of heterotrophic bacteria in incubations and may have played a role in a bacterial fresh-water. J Gen Microbiol 128:2841–2850 Hassan HF, Coombs GH (1988) Purine and pyrimidine metab- abundance not matching the substrate depletion. As olism in parasitic protozoa. FEMS Microbiol Rev 4:47–83 such the microcolony technique may provide an Hobbie JE, Daley RJ, Jasper S (1977) Use of nuclepore filters advantage when working with natural aquatic com- for counting bacteria by fluorescence microscopy. Appl munities where separating bacteria from their grazers Environ Microbiol 33:1225–1228 Højberg O, Binnerup SJ, Sørensen J (1997) Growth of sili- is desirable. cone-immobilized bacteria on polycarbonate membrane Our data demonstrate that natural marine bacterial filters, a technique to study microcolony formation under populations have a high potential for cycling purines anaerobic conditions. Appl Environ Microbiol 63: and their intermediates, while pyrimidines are less 2920–2924 labile. The exception to this was the purine adenine Jørgensen NOG (2006) Uptake of urea by estuarine bacteria. Aquat Microb Ecol 42:227–242 (this issue) which was intermediate in lability between purines Karl DM, Bossard P (1985) Measurement of microbial nucleic 32 and pyrimidines. Bacterial catabolism of purines and acid synthesis and specific growth rate by PO4 and their intermediates appears to be an important process [3H]adenine: field comparison. Appl Environ Microbiol 50: in the production of urea in coastal marine systems. 706–709 Karl DM, Winn CD (1984) Adenine metabolism and nucleic Because coastal marine bacteria act as net producers of acid synthesis: applications to microbiological oceanogra- urea, they may serve as a source of urea-N for phyto- phy. In: Hobbie J, Williams P (eds) Heterotrophic activity plankton. in the sea. Plenum Press, New York, p 197–216 226 Aquat Microb Ecol 42: 215–226, 2006

Kristiansen S (1983) Urea as an important source of nitrogen Shah N, Syrett PJ (1984) The uptake of guanine and hypoxan- in the Oslofjord. Mar Biol 74:17–24 thine by marine microalgae. J Mar Biol Assoc UK 64: LaRue TA, Spencer JFT (1968) The utilization of purines and 545–556 pyrimidines by yeasts. Can J Microbiol 14:79–86 Soldo AT, Godoy GA, Larin F (1978) Purine excretory nature LeChevalier MP, Gerber NN, Umbreit TA (1982) Transforma- of refractile bodies in the marine ciliate Parauronema acu- tion of adenine to 8-hydroxyadenine by strains of Oer- tum. J Protozool 25:416–418 skovia xanthineolytica. Appl Environ Microbiol 43: Sörenson F, Sahlsten E (1987) Nitrogen dynamics of a 367–370 cyanobacteria bloom in the Baltic Sea—new versus Litchfield CD, Hood DW (1966) Microbiological assay for regenerated production. Mar Ecol Prog Ser 37:277–284 organic compounds in seawater. II. Distribution of adenine Staley KB, Ewing RD (1992) Purine levels in the skin of juve- uracil and threonine. Appl Microbiol 14:145–151 nile coho salmon (Oncorhynchus kisutch) during parr- Mason CA, Egli T (1993) Dynamics of microbial growth in the smolt transformation and adaptation to seawater. Comp decelerating and stationary phase of batch culture. In: Biochem Physiol B 101B:447–452 Kjelleberg S (ed) Starvation in bacteria. Plenum Press, Stuer-Lauridsen B, Nygaard P (1998) Purine salvage in two New York, p 81–102 halophilic archaea: characterization of salvage pathways McCarthy JJ (1972) The uptake of urea by natural popula- and isolation of mutants resistant to purine analogs. J Bac- tions of marine phytoplankton. Limnol Oceanogr 17: teriol 180:457–463 738–748 Tamminen T, Irmisch A (1996) Urea uptake kinetics of a mid- Nicol JAC (1960) The biology of marine animals, 1st edn. Sir summer planktonic community on the SW coast of Fin- Isaac Pitman & Sons, London land. Mar Ecol Prog Ser 130:201–211 Price NM, Harrison PJ (1987) A comparison of methods for the Therkildsen MS, Isaksen MF, Lomstein BA (1996) Urea pro- measurement of dissolved urea concentrations in sea- duction by the marine bacteria Delaya venusta and water. Mar Biol 92:307–319 Pseudomonas stutzeri grown in a minimal medium. Aquat Remsen CC (1971) The distribution of urea in coastal and Microb Ecol 13:213–217 oceanic waters. Limnol Oceanogr 16:732–740 Vitols M, Volsky V, Maurina H (1974) Utilization and cata- Sarcina M, Casselton PJ (1995) Degradation of adenine by bolic transformation of exogenous purine and pyrimidine Prototheca zopfii (Chlorophyta). J Phycol 31:575–576 bases, nucleosides and nucleotides by actinomycetes. In Schultz AC, Nygaard P, Saxild HH (2001) Functional analysis Brownell GH (ed) Proceedings of the first international of genes that constitute the purine catabolic pathway in conference on biology of Nocardia. Merida, Venezuela, p Bacillus subtilis and evidence for a novel regulon con- 64–65 trolled by the PucR transcription activator. J Bacteriol 183: Vogels GD, van der Drift C (1976) Degradation of purines and 3293–3302 pyrimidines by microorganisms. Bacteriol Rev 40:403–468 Shah N, Syrett PJ (1982) Uptake of guanine by the diatom Watanabe Y, Ohe T (1972) Oxidation of hypoxanthine to uric Phaeodactylum tricornutum. J Phycol 18:579–587 acid by Streptomyces. Agric Biol Chem 36:785–792

Editorial responsibility: Frede Thingstad, Submitted: August 25, 2005; Accepted: November 21, 2005 Bergen, Norway Proofs received from author(s): March 9, 2006