Purine Salvage in Two Halophilic Archaea: Characterization of Salvage Pathways and Isolation of Mutants Resistant to Purine Analogs

Abstract

In exponentially growing cultures of the extreme halophile Halobacterium halobium and the moderate halophile Haloferax volcanii, growth characteristics including intracellular protein levels, RNA content, and nucleotide pool sizes were analyzed. This is the first report on pool sizes of nucleoside triphosphates, NAD, and PRPP (5-phosphoribosyl-α-1-pyrophosphate) in archaea. The presence of a number of salvage and interconversion enzymes was determined by enzymatic assays. The levels varied significantly between the two organisms. The most significant difference was the absence of GMP reductase activity in H. halobium. The metabolism of exogenous purines was investigated in growing cultures. Both purine bases and nucleosides were readily taken up and were incorporated into nucleic acids. Growth of both organisms was affected by a number of inhibitors of nucleotide synthesis. H. volcanii was more sensitive than H. halobium, and purine base analogs were more toxic than nucleoside analogs. Growth of H. volcanii was inhibited by trimethoprim and sulfathiazole, while these compounds had no effect on the growth of H. halobium. Spontaneous mutants resistant to purine analogs were isolated. The most frequent cause of resistance was a defect in purine phosphoribosyltransferase activity coupled with reduced purine uptake. A single phosphoribosyltransferase seemed to convert guanine as well as hypoxanthine to nucleoside monophosphates, and another phosphoribosyltransferase had specificity towards adenine. The differences in the metabolism of purine bases and nucleosides and the sensitivity to purine analogs between the two halobacteria were reflected in differences in purine enzyme levels. Based on our results, we conclude that purine salvage and interconversion pathways differ just as much between the two archaeal species as among archaea, bacteria, and eukarya.

The halophilic archaea form part of the group Archaea, one of the three domains suggested by Woese et al. (38) to comprise all living organisms; the other domains are the Bacteria and the Eukarya. The halophilic archaea (family Halobacteriaceae), here called halobacteria, have been isolated from a variety of salt-enriched habitats ranging from natural and artificial salt lakes to hides and fish preserved by treatment with crude salt from solar evaporation ponds. All members of the halobacteria require high concentrations of salt for growth. Some are extremely halophilic, while others are only moderately halophilic (11, 17, 27, 32). The two species analyzed in the present study, Haloferax volcanii and Halobacterium halobium, are examples of moderate and extreme halobacteria, respectively. The physical and chemical conditions of the natural environments of these organisms pose intriguing questions regarding the nature of adaptation. The halobacteria are often found at the top level of the short food chain of hypersaline environments in which a gradual increase in salinity has taken place (11, 29). As the preceding microbial communities release all cellular constituents when decaying, purines are likely to be present in substantial amounts in the environment. Consequently, purine salvage most likely is important for halobacteria. All microorganisms so far examined contain a network of pathways designed to reutilize already formed purine bases, nucleosides, and nucleotides. The extent and composition of these metabolic reactions are highly variable between organisms at all taxonomic levels (10, 24, 25, 35–37). Purine metabolism has scarcely been investigated in the archaea. Most investigations of purine salvage metabolism within the archaea have dealt with members of the methanogens (2, 6–8, 39). The presence of a few purine salvage enzymes in the halophilic archaeon Halobacterium cutirubrum has been demonstrated (1, 9).

The aim of the present work was to establish and compare the purine salvage pathways of the halophilic archaea H. volcanii and H. halobium. An integrated approach which involved assessment of enzymatic activities in crude extracts, determination of uptake and metabolism of [¹⁴C]purine bases and nucleosides, and isolation and characterization of mutants resistant to toxic purine analogs was used.

MATERIALS AND METHODS

Organisms, media, and growth conditions.

H. halobium R₁, a gas vesicle-deficient strain isolated by Stoeckenius and Kunau, (34) was kindly donated by A. S. Mankin (19). The growth medium for H. halobium contained a basal salts mixture (5) including 4.3 M NaCl, 18 mM MgSO₄, 13 mM KCl, 10 mM trisodium citrate, 1.4 mM CaCl₂, and 50 mM Tris-HCl (pH 7.4) supplemented with 0.4% Casamino Acids. The Casamino Acid stock solution (20%) was boiled for 1 min in the presence of 1 g of activated charcoal per liter and cleared by filtration through Whatman I filter paper. This removes contaminant purines and pyrimidines. H. volcanii WFD11, cured for the endogenous plasmid pHV2, was kindly provided by W. F. Doolittle. The medium for H. volcanii consisted of a basal salts mixture (20) containing 2.14 M NaCl, 246 mM MgCl₂, 29 mM K₂SO₄, 1.4 mM CaCl₂, and 25 mM Tris-HCl (pH 7.4). The salts mixture was supplemented with 1 mM K₂HPO₄, 5 mM NH₄Cl, 0.45% sodium succinate, 0.05% glycerol, trace elements (Cu, Fe, Mn, and Zn), thiamine, and biotin (16). Cells of H. volcanii and H. halobium were cultured in Erlenmeyer flasks at 40°C. Growth was monitored by observing the increase in absorbance (optical density [OD]) at 436 nm. The initial OD₄₃₆ was 0.02 to 0.03. Solid media contained 2% agar. Plates of both species were incubated at 40°C in sealed plastic bags to minimize evaporation of water.

Protein and RNA analysis.

Typically, cells from 20-ml cell cultures, OD₄₃₆ of 0.8 to 1.2, were harvested by centrifugation for 15 min at 6,000 × g. For protein analysis, the pellet was resuspended in a solution containing 30 mM KH₂PO₄-K₂HPO₄ (pH 7), 1 mM EDTA, 1 mM dithiothreitol, and 0.2% Nonidet P-40 and was homogenized in an ultrasonic disintegrator (Measuring and Scientific Equipment, Ltd., London, United Kingdom). Protein levels were determined by the method of Lowry et al. (18) with bovine serum albumin as the standard. For RNA analysis, the harvested cells were resuspended in 0.6 N perchloric acid and RNA was precipitated at 0°C for 30 min. The precipitate was washed twice with 0.6 N perchloric acid and twice with 96% ethanol, followed by centrifugation. RNA in the precipitate was hydrolyzed to nucleotides by treatment with 0.3 N KOH for 18 h at 37°C, and the extract was adjusted to pH 2 with 6 N perchloric acid. The precipitate formed was removed by centrifugation. To determine the concentration of nucleotides in the supernatant, the absorbance at 260 nm was measured. Yeast RNA (Sigma Chemical Co., St. Louis, Mo.) treated the same way was used as the standard.

Determination of nucleotide pools.

Cells were grown in the presence of 1 mM [³²P]phosphate (0.4 MBq/ml) for several generations. A 200-μl culture (OD₄₃₆, 0.6 to 0.8) was harvested by filtration on membrane filters and extracted with 200 μl of 0.33 M HCOOH at 0°C. Nucleoside triphosphates and PRPP (5-phosphoribosyl-α-1-pyrophosphate) in the extract were separated by two-dimensional chromatography on polyethyleneimine-impregnated cellulose on plastic sheets (PEI plates) and quantitated as described previously (14). NAD was separated from other ³²P-labeled compounds in another PEI two-dimensional chromatography system (28).

Enzyme assays.

Enzyme activities of crude extracts were measured at 40°C. Extracts were made from cells grown to an OD₄₃₆ of 0.8 to 1.2. Cells in 40 ml of culture were harvested by centrifugation and resuspended in 0.5 ml of buffer containing 100 mM Tris-HCl (pH 7.5), 3.5 M KCl, 0.2% Nonidet P-40, and 15 units of DNase I (Boehringer, Mannheim, Germany). When no KCl was added to the buffer, the salt concentration of the crude extract was found by conductivity measurements to be equivalent to 0.25 M KCl. To disrupt the cells, the suspensions were incubated on ice for 30 min and subsequently the cell debris was pelleted. In the assay mixture (50 μl), the conversion of ¹⁴C-labeled substrate to product was followed by these compounds being separated chromatographically on PEI plates. Ten-microliter samples were applied to PEI plates in the start spot at different times ranging from 10 to 60 min. During this period the enzyme activities were proportional to the elapsed time. Appropriate markers were applied to the spots, and the plates were developed in water. The chromatograms were examined under UV light and were visualized by autoradiography on Agfa Curix film. The UV spots of interest were cut out of the PEI plate and counted in a liquid scintillation counter (Rackbeta 1209; LKB, Bromma, Sweden). Activities of guanine deaminase, AMP deaminase, adenine deaminase (26), and xanthine oxidase were measured in a mixture of 75 mM Tris-HCl (pH 7.5) and 2.5 M KCl containing 0.1 mM guanine (0.4 MBq/μmol) [γ-¹⁴C]guanine, [8-¹⁴C]AMP, [8-¹⁴C]adenine, or [8-¹⁴C]xanthine, respectively. For the purine nucleoside kinases, the assay mixture contained 150 mM Tris-HCl (pH 7.5), 2 M KCl, 50 mM MgCl₂, 10 mM ATP, and 0.2 mM [8-¹⁴C]purine nucleosides (0.1 MBq/μmol). Coformycin (5 μM) was added to the adenosine kinase assay mixture to inhibit adenosine deaminase activity (23). Activity of adenosine deaminase was determined under the same conditions as for the purine nucleoside kinases. Purine nucleoside phosphorylase activities were measured in a solution containing 75 mM KH₂PO₄-K₂HPO₄ (pH 7.1), 1.5 M KCl, and 0.5 mM [8-¹⁴C]purine nucleosides (0.2 MBq/μmol). The activities of adenine, guanine, xanthine, and hypoxanthine phosphoribosyltransferase were determined essentially as described previously (15). Adenine, hypoxanthine, and xanthine phosphoribosyltransferase were assayed in the presence of 3 M KCl, and guanine phosphoribosyltransferase was assayed in the presence of 1 M KCl. Adenylosuccinate synthetase and lyase activities were determined as described previously (31) with 1 M KCl in the assay mixture. Radiolabeled purine compounds were obtained from NEN Life Science Products.

Metabolism of purine bases and nucleosides.

[¹⁴C]purine bases or nucleosides were added to 2-ml cultures of exponentially growing cells. The concentration of the purine compound added to the medium was 100 μM (0.02 MBq/μmol) at the start of the experiment. Ten-microliter samples of the culture were withdrawn and applied to PEI plates at intervals until the culture reached an OD₄₃₆ of 1.0 to 1.5. The PEI plates were developed in water. This allows the separation of the purine bases and nucleosides from each other and from the label incorporated in the cells. Radioactivity remaining in the application spot corresponds to that incorporated by the cells into nucleotides and nucleic acids. The chromatograms were analyzed as described above.

Incorporation of ¹⁴C-labeled bases into RNA was determined by the perchloric acid procedure described above. Ten to 20 μl of the neutralized supernatants were applied to PEI plates, which were subsequently developed in 0.5 M ammonium formate (pH 3.4). Relevant spots (mononucleotides) were excised and counted in a liquid scintillation counter.

Uptake of purine bases and nucleosides.

Uptake of purine bases and nucleosides was determined essentially as described previously (30). Exponentially growing cells were harvested by centrifugation and resuspended in fresh medium. After 5 min of incubation at 40°C, 1-ml samples were transferred to vials containing 1 μM [¹⁴C]purine base (2 MBq/μmol). At 30, 60, 100, and 150 s, samples of 200 μl were withdrawn and filtered through a 0.45-μm-pore-size membrane filter. The filters were immediately washed with 2.5 ml of basal salts mixture, dried, and counted in a liquid scintillation counter. The rate of uptake of purine bases was proportional to the time elapsed during the uptake experiment.

Isolation of mutants resistant to purine analogs.

Cells growing exponentially in liquid medium were used. A 200-μl culture at an OD₄₃₆ of 0.5 to 0.7 was plated on agar medium containing different concentrations of a given analog. At the highest concentrations used, no mutations occurred; at the lowest concentration, a lawn of growth was seen. For H. volcanii, optimal concentrations were between 5 and 50 μM; for H. halobium, optimal concentrations were between 0.2 and 0.5 mM. Resistant colonies were picked and streaked on a new agar plate containing the same concentration of the analog. All mutants were tested for cross-resistance to other analogs in the indicated concentration range.

RESULTS

Growth characteristics and nucleotide pool analysis.

For each of the two halobacteria analyzed, the composition of the salts mixture reflected the composition of the natural habitat. The medium for H. halobium contained amino acids as nitrogen and carbon sources, while that of H. volcanii contained ammonium ions, succinate, and glycerol. Despite the richer medium, growth of H. halobium was slower than that of H. volcanii (Table 1). For both strains the exponential growth phase ended at an OD₄₃₆ of 1.2 to 1.5. Throughout, we observed a 5 to 10% stimulation of the growth rate when purine compounds were added to the growth medium. The nucleotide pool sizes were generally two to three times higher in H. volcanii than in H. halobium. The ratios between the content of RNA and protein were 0.132 for H. halobium and 0.176 for H. volcanii, compared with 0.38 for Escherichia coli growing with a doubling time of 1 h.

TABLE 1.

Intracellular amounts of nucleoside triphosphates, NAD, and PRPP in H. volcanii and H. halobium^a

Compound	Amt (nmol/ml) inb:
	H. volcanii (doubling time, 3.8 h)	H. halobium (doubling time, 8.5 h)
ATP	1.77 ± 0.07	0.82 ± 0.08
GTP	0.79 ± 0.08	0.28 ± 0.01
CTP	0.58 ± 0.14	0.12 ± 0.03
UTP	0.72 ± 0.05	0.29 ± 0.05
dATP	0.06 ± 0.01	0.02 ± 0.01
dGTP	0.04 ± 0.02	0.02 ± 0.01
dCTP	0.05 ± 0.01	0.02 ± 0.01
dTTP	0.06 ± 0.03	0.03 ± 0.02
PRPP	0.98 ± 0.06	0.38 ± 0.10
NAD	0.12 ± 0.04	0.12 ± 0.02

^aCells growing exponentially in liquid medium at 40°C were analyzed as described in Materials and Methods. Data are based on the following growth characteristics at an OD₄₃₆ of 1. H. volcanii: cell numbers, (5.3 ± 0.5) × 10⁹ per ml; protein content, 176 ± 13 μg/ml; RNA content, 31 ± 5 μg/ml. H. halobium: cell numbers, (5.4 ± 0.5) × 10⁹ per ml; protein content, 197 ± 17 μg/ml; RNA content, 26 ± 4 μg/ml. A Burker counting chamber and a light microscope with a magnification of ×400 were used to determine total cell numbers. 

^bValues are means of at least four experiments ± standard deviations. OD₄₃₆, 1. 

Purine salvage.

Enzyme activities known to catalyze anabolic, catabolic, and interconversion reactions of purine compounds were analyzed in cell extracts. Both strains possessed enzyme activities catalyzing the deamination of adenosine and guanine (Table 2). Adenosine, guanosine, and inosine can be cleaved phosphorolytically to the free base and ribose-1-phosphate. An alternative pathway for the metabolism of nucleosides is phosphorylation to the corresponding nucleoside monophosphate. Except for adenosine kinase, all of these reactions were found to occur in both strains (Table 2). The adenosine deaminase level was high while the level of adenosine phosphorylase was low in H. halobium. Exactly the opposite conditions were observed in H. volcanii. This suggested that adenosine is predominantly deaminated in H. halobium and phosphorolyzed in H. volcanii. The level of guanosine and inosine kinase was twofold higher in H. volcanii, which also possessed much higher levels of guanosine phosphorylase than did H. halobium. Both species expressed similar levels of purine phosphoribosyltransferase activities towards adenine, hypoxanthine, and guanine (Table 2), and these levels were not affected by free purine bases in the growth medium (data not given). Neither xanthine phosphoribosyltransferase activity nor xanthine oxidase activity could be measured in the presence of 3 M KCl, 1 M KCl, or no KCl in the assay mixture. Adenylosuccinate synthetase and adenylosuccinate lyase activities were detected in both species (Table 2).

TABLE 2.

Activities of purine salvage and interconversion enzymes and purine uptake in H. volcanii and H. halobium^a

Purine	Activity (nmol/min/mg of protein) in:
	H. volcanii	H. halobium
Salvage and interconversion enzymes
Adenine deaminase	<0.01	<0.01
Adenosine deaminase	0.12 ± 0.02	0.76 ± 0.19
AMP deaminase	<0.02	<0.02
Guanine deaminase	1.14 ± 0.27	0.53 ± 0.12
Adenosine phosphorylase	2.63 ± 0.52	0.10 ± 0.02
Guanosine phosphorylase	5.42 ± 0.51	0.37 ± 0.04
Inosine phosphorylase	3.74 ± 0.19	2.89 ± 0.14
Adenosine kinase	<0.01	<0.01
Guanosine kinase	0.12 ± 0.03	0.06 ± 0.02
Inosine kinase	0.12 ± 0.02	0.05 ± 0.01
Adenine phosphoribosyltransferase	1.43 ± 0.12	2.05 ± 0.17
Guanine phosphoribosyltransferase	2.45 ± 0.23	0.87 ± 0.17
Hypoxanthine phosphoribosyltransferase	2.13 ± 0.18	2.34 ± 0.21
Xanthine phosphoribosyltransferase	<0.01	<0.01
Xanthine oxidase	<0.01	<0.01
Adenylosuccinate synthetase	0.72 ± 0.04	0.56 ± 0.03
Adenylosuccinate lyase	0.02 ± 0.003	0.04 ± 0.004
Uptake
Adenine	0.31 ± 0.03	0.65 ± 0.06
Guanine	0.23 ± 0.04	0.19 ± 0.04
Hypoxanthine	0.25 ± 0.07	0.40 ± 0.04
Adenosine	0.13 ± 0.03	0.44 ± 0.03
Guanosine	0.09 ± 0.03	0.63 ± 0.04
Inosine	0.14 ± 0.02	0.48 ± 0.03

^aCells were grown in liquid medium and were analyzed as described in Materials and Methods. Values are means of at least two experiments ± standard deviations.  

Uptake and metabolism of [¹⁴C]purine bases and nucleosides.

Determination of the initial rates of purine base and nucleoside uptake at 1 μM concentrations showed that both strains possessed high-affinity transport systems (Table 2). The highest rates of uptake were found in H. halobium. The metabolism of purine bases and nucleosides was monitored by analyzing the disappearance of the added purine compound (100 μM) in exponentially growing cells. The disappearance of adenine, guanine, and hypoxanthine from the growth medium was paralleled by incorporation into nucleotides and nucleic acids (Fig. 1). The enzyme data (Table 2) indicated that adenine, guanine, and hypoxanthine are phosphoribosylated to AMP, GMP, and IMP, respectively. Guanine was also converted to xanthine, which was excreted into the growth medium, most dramatically in H. halobium. Xanthine was not metabolized by the two halobacteria. Generally, nucleosides were more rapidly taken up and metabolized than their corresponding purine bases. The catabolism of adenosine resulted in inosine and adenine. From the amount and composition of the excretion products, it was apparent that more adenine and less inosine was excreted by H. volcanii than by H. halobium. This actually reflects the differences in the levels of adenosine deaminase and adenosine phosphorylase in the two strains (Table 2). In both strains, catabolism of inosine resulted in the excretion of excess hypoxanthine. Catabolism of guanosine, on the other hand, did not result in the excretion of guanine. The guanine formed intracellularly from guanosine was to a significant extent deaminated to xanthine, which was excreted.

FIG. 1.

Metabolism of exogenous purine bases and nucleosides in H. volcanii and H. halobium. The uptake, incorporation, and excretion of [¹⁴C]purine bases and nucleosides in growing cells were monitored. Samples of the culture were directly subjected to ion-exchange chromatography. The system used allowed the separation of purine bases and nucleosides in the medium from nucleotides and nucleic acids in the cells. The distribution of the labeling was calculated as described in Materials and Methods. Ade, adenine; Ado, adenosine; Hyp, hypoxanthine; Ino, inosine; Gua, guanine; Guo, guanosine.

To analyze for possible purine interconversion pathways, cultures were grown in the presence of radiolabeled adenine, guanine, or hypoxanthine. The RNA was isolated and degraded, and the distribution of labeling in AMP and GMP was determined (Table 3). Exogenous adenine was incorporated almost to the same extent into AMP and GMP of both species. Guanine was incorporated into GMP of RNA in both species but into AMP only in H. volcanii. Hypoxanthine was incorporated into both AMP and GMP, with most of the label being channelled into GMP in both species. These incorporation studies complement the enzyme analysis by providing evidence for the conversion of IMP into both GMP and AMP (Fig. 2).

TABLE 3.

Ratio of incorporation of [¹⁴C]purine bases into AMP to that of incorporation into GMP of RNA^a

Purine added	AMP/GMP incorporation ratio
	H. volcanii	H. halobium
Adenine	1.23 ± 0.05	0.92 ± 0.05
Guanine	0.24 ± 0.01	<0.01
Hypoxanthine	0.67 ± 0.05	0.64 ± 0.06

^aCells were grown in liquid medium. Incorporation was determined as described in Materials and Methods. Values are means of three experiments ± standard deviations. 

FIG. 2.

Purine salvage and interconversion pathways in H. volcanii and H. halobium. The individual reactions are identified by numbers: 1, adenosine deaminase; 2, adenosine phosphorylase; 3, guanosine phosphorylase; 4, inosine phosphorylase; 5, inosine kinase; 6, guanosine kinase; 7, adenine phosphoribosyltransferase; 8, hypoxanthine(guanine) phosphoribosyltransferase; 9, guanine deaminase; 10, adenylosuccinate synthetase; 11, adenylosuccinate lyase; 12, IMP dehydrogenase; 13, GMP synthetase; 14, GMP reductase (found only in H. volcanii).

Isolation and characterization of mutants resistant to purine analogs.

Analysis of sensitivity and the development of resistance to analogs affecting nucleotide metabolism provides a tool by which nucleotide metabolism can be studied (21). As a first approach, both halobacteria were cultivated in liquid medium to which different analogs were added. Both species were found to be sensitive to a number of analogs (Table 4). Purine base analogs were more inhibitory overall than purine nucleoside analogs, and H. volcanii was more sensitive to the analogs than was H. halobium. Five analogs, 2-fluoroadenine, 6-thioguanine, 6-methylpurine, 8-azaguanine, and 8-azahypoxanthine, were very toxic to both strains. This is probably because all can be activated to toxic compounds as a result of phosphoribosylation. The two guanosine analogs 6-mercaptoguanosine and 8-mercaptoguanosine and 6-mercaptopurine riboside were much more toxic to H. volcanii than to H. halobium. This might be explained by the significantly higher levels of guanosine phosphorylase and guanine phosphoribosyltransferase in H. volcanii, which may catalyze the conversion of the analogs to the toxic nucleotide compounds. A similar argument could be used to explain the sensitivity to 2-fluoroadenosine.

TABLE 4.

Effects of analogs on the growth of H. volcanii and H. halobium in liquid medium

Analog added to medium	Concn (mM)	Growth (%)a
		H. volcanii	H. halobium
2-Fluoroadenine	1	<5	<5
2,6-Diaminopurine	1	67	71
6-Mercaptopurine	1	<5	33
6-Thioguanine	1	<5	<5
6-Methylpurine	1	<5	<5
8-Azaadenine	1	36	48
8-Azaguanine	1	<5	<5
8-Azaxanthine	1	100	100
8-Azahypoxanthine	1	<5	<5
2-Fluoroadenosine	0.5	16	34
6-Mercaptopurine riboside	0.5	15	83
6-Mercaptoguanosine	0.5	<5	67
8-Mercaptoguanosine	0.5	<5	83
Purine riboside	0.5	67	90
Psicofuranine	0.5	80	83
Trimethoprim	0.03	<2	100
Sulfathiazole	0.03	<10	100
Mycophenolic acid	0.3	100	100

^aExponentially growing cells were diluted to an OD₄₃₆ of 0.02 to 0.03 and were supplemented with the indicated analogs. Growth was monitored in control cultures grown in the absence of analogs until an OD₄₃₆ of 1.2 was reached (defined as 100% growth). The growth yields (OD₄₃₆) of the analog-supplemented cultures were determined and expressed as percentages of the control cultures. 

Spontaneous mutants of H. halobium resistant to either 0.5 mM 8-azaadenine or 0.2 mM 8-azaguanine were selected as described in Materials and Methods. The mutants appeared with a frequency of about five per 10⁶ cells. Mutants resistant to 0.2 mM 8-azahypoxanthine were also isolated. However, they grew extremely slowly irrespective of growth conditions. A number of mutants were tested for cross-resistance toward other purine analogs. The 8-azaadenine-resistant mutants were also resistant to 6-methylpurine and 2-fluoroadenine but were sensitive to 8-azaguanine. The mutants resistant to 8-azaguanine, on the other hand, were resistant to 8-azahypoxanthine but were sensitive to 8-azaadenine, 6-methylpurine, and 2-fluoroadenine. Several of the 8-azaadenine-resistant mutants were further characterized with respect to their purine-metabolizing capabilities, and data from two mutants are shown in Table 5. Both were defective in adenine phosphoribosyltransferase activity and adenine uptake and showed increased guanine-hypoxanthine phosphoribosyltransferase activity and uptake. Several mutants resistant to 8-azaguanine were isolated and classified into three classes. Data from a representative mutant strain from each class is shown in Table 5. Mutant strain Hh15 showed normal guanine and hypoxanthine phosphoribosyltransferase activities and uptake. Mutants of this class were not further investigated. Mutant strain Hh16 was defective in guanine and hypoxanthine phosphoribosyltransferase activities and in guanine and hypoxanthine uptake. Mutant strain Hh18, on the other hand, showed a moderate decrease in guanine phosphoribosyltransferase activity, normal hypoxanthine phosphoribosyltransferase activity, and reduced uptake of guanine and hypoxanthine. All three mutant strains showed increased adenine uptake.

TABLE 5.

Purine phosphoribosyltransferase activity and purine uptake in mutants of H. halobium and H. volcanii resistant to purine analogs^a

Organism	Phenotype of mutant	Strain	Phosphoribosyltransferase activity (nmol/min/mg of protein)			Uptake (nmol/min/mg of protein)
			Ade	Hyp	Gua	Ade	Hyp	Gua
H. halobium	Wild type	Hh1	2.05	2.34	0.87	0.65	0.40	0.19
	8-Azaadenine resistant	Hh11	<0.01	4.57	—	0.02	0.59	0.30
		Hh12	<0.01	4.14	—	0.04	0.52	0.26
	8-Azaguanine resistant	Hh15	—	3.30	0.78	0.81	0.31	0.25
		Hh16	—	0.05	<0.01	0.81	0.09	0.10
		Hh18	—	2.10	0.37	0.96	0.12	0.07
H. volcanii	Wild type	WFD11	1.43	2.13	2.45	0.31	0.25	0.23
	2-Fluoroadenine resistant	Hv166	<0.01	4.30	—	0.13	0.38	0.34
		Hv174	<0.01	5.70	—	0.01	0.13	0.21
	8-Azaguanine resistant	Hv105	—	<0.01	0.07	—	0.01	0.01
		Hv118	—	2.32	2.48	—	0.03	0.02

^aAde, adenine; Hyp, hypoxanthine; Gua, guanine; —, not determined. Values are averages of two independent experiments. 

H. volcanii appeared to be very sensitive to the analogs when tested on plates. To avoid killing of all cells, much lower concentrations of analogs were used to select for resistant mutants. Spontaneous H. volcanii mutants resistant to 2-fluoroadenine (50 μM), 8-azaguanine (5 μM), 8-azahypoxanthine (5 μM), or 2-fluoroadenosine (50 μM) were isolated. As with H. halobium, the mutants of H. volcanii resistant to 8-azahypoxanthine appeared to be very sick and were difficult to grow. Overall, the mutants of H. volcanii resembled those of H. halobium with respect to patterns of cross-resistance. The 2-fluoroadenine-resistant mutants appeared with a frequency of two per 10⁶ cells. Two mutants, Hv166 and Hv174, were defective in adenine phosphoribosyltransferase activity, but adenine uptake was not reduced to the same extent (Table 5). In both mutant strains, hypoxanthine and guanine utilization was affected. Attempts to isolate mutants resistant to 2-fluoroadenosine that were defective in adenosine phosphorylase failed. All resistant mutants obtained were defective in adenine phosphoribosyltransferase (data not given). Mutants resistant to 8-azaguanine appeared with a frequency of one per 10⁸ cells. Two classes were obtained, represented by mutant strains Hv105 and Hv118 in Table 5. Both strains showed reduced uptake of guanine and hypoxanthine. In addition, Hv105 was defective in guanine and hypoxanthine phosphoribosyltransferase activity.

Inhibitors of de novo purine synthesis.

The toxicities of known inhibitors of purine synthesis were tested in both halobacteria. Mycophenolic acid, known to inhibit IMP dehydrogenase (33), and psicofuranine, known to inhibit GMP synthetase (24), had no significant effects on growth in either species at concentrations of 0.3 mM. Trimethoprim, a structural analog of folic acid (40), and sulfathiazole, a structural analog of p-aminobenzoic acid (4), are inhibitors of one-carbon metabolism. Both compounds seriously affected the growth of H. volcanii at concentrations of 0.03 mM (Table 4) but were not toxic to H. halobium even at concentrations of 0.3 mM. When H. volcanii was grown with trimethoprim (0.03 mM) plus hypoxanthine (0.1 mM) and thymidine (0.04 mM), growth was restored to a normal rate. Single addition of hypoxanthine to trimethoprim-supplemented cultures did not restore growth. When thymidine was added together with trimethoprim, the growth yield was 60 to 80% of normal. The reduction in growth caused by sulfathiazole was restored to normal by the addition of p-aminobenzoic acid (0.05 mM) to the medium but not by the addition of hypoxanthine and thymidine. To study the effects of trimethoprim (0.03 mM) and sulfathiazole (0.3 mM) on nucleotide metabolism, nucleotide pool sizes were determined. After 90 min of incubation with trimethoprim, the most significant changes in pool sizes were a fourfold reduction in the ATP and dTTP pools and an eightfold increase in the PRPP pool. The sulfathiazole-treated cells showed a threefold increase in the PRPP pool (data not shown). This swelling of the PRPP pools is an indication of arrested purine biosynthesis, also seen in bacteria (14, 25).

DISCUSSION

Purine metabolism was investigated in the halobacteria H. volcanii and H. halobium. Nucleotide and PRPP pool sizes were determined for the first time in archaea. The pool sizes were four to eight times lower than those found in enterobacteria, but the relative concentrations were similar (21). The RNA/protein ratio was also lower than that of E. coli. The lowest pool size and RNA/protein ratio observed was in H. halobium. Most likely these figures reflect the lower growth rate of H. halobium than H. volcanii under the experimental conditions. Both halobacteria contain extensive and active salvage and interconversion pathways as well as high-affinity transport systems for purine bases and nucleosides. Enzyme studies and analyses of the metabolism of exogenous purine bases and nucleosides have revealed that the two halobacteria contain the same purine salvage enzymes. The major differences were that H. halobium excreted large amounts of xanthine when supplemented with guanine (Fig. 1) and that it was unable to convert guanine to AMP (Table 2). Guanine is initially converted to GMP in both species, and the only known way of conversion of GMP, and guanine compounds, to IMP is through the reductive deamination of GMP catalyzed by GMP reductase (21, 24). This pathway seems to operate only in H. volcanii. Both H. volcanii and H. halobium were capable of converting adenine compounds to GMP (Table 3). The enzyme activities present in both halobacteria suggest that this conversion implies formation of adenosine from adenine and ribose-1-phosphate catalyzed by adenosine phosphorylase, followed by deamination of adenosine to inosine catalyzed by adenosine deaminase. Subsequently, inosine is phosphorolyzed to hypoxanthine and ribose-1-phosphate. The latter is recycled, while hypoxanthine can be converted to IMP and hence GMP. Such a pathway has been identified in enterobacteria (21). The differences in the levels of guanosine phosphorylase, adenosine phosphorylase, and adenosine deaminase between the two species were reflected in the patterns seen of the metabolism of exogenous nucleosides (Fig. 1) and of the sensitivities to purine analogs (Table 4).

By selecting for resistance to purine analogs, strains defective in purine phosphoribosyltransferase activity and/or purine base uptake in both halobacteria were isolated. From our analysis (Tables 2 and 5), we conclude that both species contain an adenine phosphoribosyltransferase and a hypoxanthine(guanine) phosphoribosyltransferase. The cross-resistance pattern observed indicates that 8-azaadenine, 6-methylpurine, and 2-fluoroadenine react as adenine analogs and that 8-azaguanine, 8-azahypoxanthine, and 6-mercaptopurine are substrates for hypoxanthine(guanine) phosphoribosyltransferase. The enzyme and uptake data further indicate that phosphoribosylation is an important route by which exogenous purine bases and base analogs are metabolized. The observation that a defective adenine phosphoribosyltransferase results in reduced adenine uptake and causes an increase in hypoxanthine and guanine uptake (Table 5) is most likely explained by a stimulation of hypoxanthine and guanine metabolism (21). This stimulation may be explained by an increase in the size of the PRPP pool as a result of the loss of a purine phosphoribosyltransferase (13). A more complex picture was seen with the 8-azaguanine-resistant mutants. The H. halobium mutant strain Hh15 was not significantly affected in hypoxanthine and guanine salvage reactions. A similar phenotype in E. coli has been identified as a mutation in a regulatory gene that increase the expression of the purine biosynthetic genes (23). The result of this is an increased capacity to synthesize purine nucleotides. Whether a similar explanation is valid in the present case was not investigated. Mutant strains Hh16 of H. halobium and Hv105 of H. volcanii were defective in hypoxanthine(guanine) phosphoribosyltransferase activity and uptake, while mutant strains Hh18 and Hv118 were defective only in hypoxanthine and guanine uptake. On the basis of all of the results obtained, we propose the scheme shown in Fig. 2 for the purine salvage and interconversion pathways in the two halobacteria studied.

The most significant difference between the enzyme composition of bacteria and that of the halobacteria is the absence of xanthine phosphoribosyltransferase in both halobacterial species and the lack of GMP reductase activity in H. halobium. Inosine and guanosine kinase activities were detected in both halobacteria (Table 2). These activities have been identified in only a few species, including enterobacteria, plants, yeast, a parasite, and human mitochondria (12).

Two known inhibitors of purine biosynthesis, trimethoprim and sulfathiazole, severely affected the growth of H. volcanii, but they had no effect on H. halobium (Table 4). Inhibition caused by trimethoprim was primarily on the de novo synthesis of dTTP, although purine biosynthesis was also affected, as judged from PRPP and nucleotide pool size analyses. The inhibition caused by sulfathiazole was overcome by p-aminobenzoic acid, an intermediate compound in folic acid biosynthesis. These findings and the isolation of dihydrofolate reductase from H. volcanii (40) indicate that this organism uses folic acid in its one-carbon metabolism. Since none of the inhibitors affected H. halobium, it may be that this organism uses not folic acid but rather modified folates in its one-carbon metabolism, as reported for the archaea Methanosarcina thermophila and Sulfolobus solfataricus (22). Another explanation is that neither trimethoprim nor sulfathiazole are transported into H. halobium.

Purine salvage metabolism has been little investigated in archaea, and most studies have involved methanogens. However, two reports have been published on purine enzymes from H. cutirubrum, one regarding a search for deaminase activities in crude extracts by using guanine, guanosine, adenine, adenosine, and 2′-deoxyadenosine, as well as a number of purine nucleotides, as substrates. Only adenosine and 2′-deoxyadenosine deaminase activities were found (1). This agrees with the results of the present study, except for our demonstration of the presence of guanine deaminase activity. Adenine, hypoxanthine, and guanine phosphoribosyltransferase activities have been demonstrated previously, but it was not established how many enzymes were involved (9).

The most detailed studies of purine salvage in methanogens have been performed with Methanobacterium thermoautotrophicum, using both wild-type cells and mutants resistant to purine analogs (39). Purine salvage was assessed by determining the incorporation of purine bases and by testing resistance to purine analogs. Several enzymes were detected in cell extracts. The major differences between the data obtained and our data were the following. M. thermoautotrophicum possesses only a low level of adenine phosphoribosyltransferase activity and showed xanthine phosphoribosyltransferase activity. The levels of guanosine and inosine phosphorylase were low, while the levels of inosine and guanosine kinase were high. Both adenine and AMP deaminase activities were found in M. thermoautotrophicum, while this organism did not contain guanine deaminase activity. Adenosine kinase activity was observed, but it was not established whether the activity was a direct phosphorylation of adenosine or involved prior deamination to inosine followed by phosphorylation to IMP (39).

A few studies have dealt with Methanococcus voltae. This organism incorporates guanine, adenine, and hypoxanthine into nucleic acids in amounts which indicate that both adenine and guanine nucleotides can be derived from any of the three bases (2). Growth of M. voltae was also found to be inhibited by several purine analogs. Mutants resistant to analogs like 8-azaguanine, 8-azahypoxanthine, and 6-mercaptopurine that were defective in the incorporation of guanine and hypoxanthine have been isolated (2). In cell extracts of M. voltae, guanosine phosphorylase, hypoxanthine phosphoribosyltransferase, and guanine phosphoribosyltransferase activities were demonstrated. However, it was not determined whether the latter two activities were catalyzed by a single enzyme or by two enzymes (3).

Purine degradation has been observed in Methanococcus vannielii, which is capable of degrading purines to an extent that allows this archaeon to use guanine, xanthine, or hypoxanthine, but not adenine, as the sole nitrogen source (7, 8). Guanine can, at low concentrations, be salvaged and converted into GMP but not into AMP (6), indicating that M. vannielii does not posses GMP reductase activity.

Overall, the purine salvage pathways of H. volcanii and H. halobium resemble the investigated methanogens, but significant differences exist. The resemblance towards this phylogenically distinct group of the Archaea is not more marked than toward members of the Bacteria and the Eucarya.