Codenitrification and Denitrification Are Dual Metabolic Pathways through Which Dinitrogen Evolves from Nitrate in Streptomyces antibioticus

Abstract

We screened actinomycete strains for dinitrogen (N₂)-producing activity and discovered that Streptomyces antibioticus B-546 evolves N₂ and some nitrous oxide (N₂O) from nitrate (NO₃⁻). Most of the N₂ that evolved from the heavy isotope ([¹⁵N]NO₃⁻) was ¹⁵N¹⁴N, indicating that this nitrogen species consists of two atoms, one arising from NO₃⁻ and the other from different sources. This phenomenon is similar to codenitrification in fungi. The strain also evolved less, but significant, amounts of ¹⁵N¹⁵N from [¹⁵N]NO₃⁻ in addition to ¹⁵N¹⁵NO with concomitant cell growth. Prior to the production of N₂ and N₂O, NO₃⁻ was rapidly reduced to nitrite (NO₂⁻) accompanied by distinct cell growth, showing that the actinomycete strain is a facultative anaerobe that depends on denitrification and nitrate respiration for anoxic growth. The cell-free activities of denitrifying enzymes could be reconstituted, supporting the notion that the ¹⁵N¹⁵N and ¹⁵N¹⁵NO species are produced by denitrification from NO₃⁻ via NO₂⁻. We therefore demonstrated a unique system in an actinomycete that produces gaseous nitrogen (N₂ and N₂O) through both denitrification and codenitrification. The predominance of codenitrification over denitrification along with oxygen tolerance is the key feature of nitrate metabolism in this actinomycete.

The biological process of dinitrogen (N₂) gas formation from fixed nitrogen compounds such as nitrate (NO₃⁻) plays an important role in maintaining homeostasis of the global environment. Bacterial denitrification was long considered the sole biological reaction responsible for this condition, and its systems have been characterized in detail (3, 8, 20). Denitrification physiologically functions as anaerobic respiration in which NO₃⁻ is used as the terminal electron acceptor when oxygen (O₂) is unavailable. Known bacterial denitrifying systems consist of four steps that successively reduce NO₃⁻ to N₂ and involve nitrite (NO₂⁻), nitric oxide (NO), and nitrous oxide (N₂O) as intermediates. Production of the enzymes catalyzing each step is induced by nitrogen oxides and suppressed by O₂ (3, 20). During the past decade, denitrifiers have been discovered in a variety of taxa including filamentous fungi (9, 10), yeasts (15), and actinomycetes (1, 11). All of these novel denitrifiers produce N₂O as the major denitrification product. Both nitrogen atoms in the N₂O product are derived from nitrate (or nitrite) (1, 9-11, 15). The phenomenon is therefore defined as denitrification, since the process should include the formation of an N—N bond (20). This system in several fungi is localized at respiring mitochondria, where it acts as anaerobic respiration, as it does in bacterial systems (5, 13, 16). Another unique feature of the fungal system is the involvement of cytochrome P450 (P450nor) as NO reductase (Nor) (6).

Biological processes other than bacterial denitrification evolve N₂ from fixed nitrogen compounds, but their molecular mechanisms and physiological significance remain to be elucidated. We identified simultaneous fungal codenitrification and denitrification, in which a hybrid N₂ species is formed by combining two nitrogen atoms derived from NO₂⁻ and from other nitrogen compounds (10, 14). The denitrifying fungi Fusarium solani and Cylindrocarpon tonkinense evolve hybrid N₂ species (10). The fact that the denitrifying fungus Fusarium oxysporum evolves N₂O instead of N₂ by codenitrification only when a nitrogen compound in addition to NO₂⁻ such as azide, salicylhydroxamic acid, or ammonium (NH₄⁺) is available (14) suggests that the mechanisms of this process differ among fungal species. Anammox is a third N₂-generating metabolic process that has been identified in the strictly anaerobic chemolithotrophic Planctomycetales (12), in which NH₄⁺ is combined with NO₂⁻ to form N₂.

The actinomycetes form a unique taxon among gram-positive bacteria. Although actinomycetes naturally proliferate in soil and in aqueous environments, little is known about how they accomplish denitrification. Denitrifiers also occur among actinomycetes (1, 11), and the system of Streptomyces thioluteus has been characterized previously (11). All of the actinomycete denitrifiers found in these studies evolve N₂O from NO₃⁻ or NO₂⁻, and thus, no known actinomycete strains contain a complete denitrifying system that can thoroughly reduce NO₃⁻ to N₂.

The present study continues screening for denitrifying actinomycetes (11) by using a highly sensitive N₂ detection system equipped with an isotope mass spectrometer and NO₃⁻ labeled with a stable isotope ([¹⁵N]NO₃⁻). The results showed that Streptomyces antibioticus B-546 has N₂-producing activity and that most of the N₂ molecules are formed via intracellular codenitrification that is induced simultaneously with denitrification.

MATERIALS AND METHODS

Strains and media.

The strains used in this study originated from the JCM (Japan Collection of Microorganisms, RIKEN) or IFO (Institute for Fermentation, Osaka, Japan) type culture collections, except for strains A, B, C, and N. We isolated the latter four strains from soil, and S. antibioticus B-546 originated from a different culture collection in the Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo. All cells were seeded, precultured, or cultured in glycerol-peptone (GP) medium (3% glycerol, 0.2% peptone, 10 mM KH₂PO₄, 0.02% MgSO₄·7H₂O, and inorganic salts) (11) unless otherwise stated.

Batch cultures in flasks.

Actinomycetes were batch cultured in flasks as follows. Seed cultures (10 ml) in 50-ml tubes were inoculated into 300 ml of GP medium in 500-ml Erlenmeyer flasks and rotary shaken at 30°C for 4 days at 120 rpm (preculture). Finally, portions of the preculture (100 ml) were transferred to 500-ml flasks containing 100 ml of GP medium supplemented with 10 mM NaNO₃ and incubated as described above (for preculture). Denitrifying actinomycetes were screened under initially aerobic conditions that were attained by sealing the flask after inoculation with a rubber stopper without replacing headspace air. NaNO₃ was labeled with 10% heavy isotope nitrogen (¹⁵N). Time-dependent changes in nitrogen oxides were investigated under two aerating conditions. Initially aerobic cultures were incubated in the same manner as for screening. Anaerobic cultures were incubated in the same manner, but the headspace air was replaced with argon gas before the flask was sealed. Both cultures used [¹⁵N]NaNO₃ instead of [¹⁴N]NaNO₃. When necessary, acetylene gas was added to a ratio of 10% of the headspace volume.

Quantification of nitrogenous oxides, dinitrogen, and cell weight.

We determined NO₃⁻ and NO₂⁻ with an ion-pair, high-performance liquid chromatograph equipped with a TSK gel IC-Anion PW column (Tosoh, Tokyo, Japan) as described previously (9). Oxygen, N₂O, and N₂ concentrations in the gas phase were determined by gas chromatography (GC) as described previously (9). ¹⁵N-labeled N₂ gas was analyzed by isotope mass spectrometry (MS) with a Finnigan DELTA Plus isotope mass spectrometer as described previously (15). Collected cells were dried at 80°C for 2 h and then weighed (cell dry weight).

Preparation of crude extract and subcellular fractionation.

Denitrifying cells of S. antibioticus B-546 were harvested by centrifugation, resuspended in buffer (100 mM potassium phosphate [pH 7.2], 0.1 mM EDTA, 0.1 mM dithiothreitol, 10% glycerol, 0.25 μM N-tosyl-l-phenylalanyl chloromethyl ketone, and 0.25 μM phenylmethanesulfonyl fluoride), and sonicated (Sonifier 250; Branson) for 60 min with occasional cooling. The sonicate was separated by centrifugation at 10,000 × g for 20 min to obtain the supernatant (crude extract), which was further sedimented by centrifugation at 100,000 × g for 60 min. The resulting supernatant (soluble fraction) and pellet (membranes) were analyzed as follows.

Enzyme assay.

Nitrate reductase (Nar) was assayed as described previously (5) with methyl viologen-dithionite as the electron donor. Nitrite reductase (Nir) was assayed with NADH-phenazine methosulfate as the electron donor by determining the amount of NO produced by the P450nor-trap method (5). Nor activity was determined by measuring NADH-dependent N₂O formation by GC as described previously (6). Nitrous oxide reductase (N₂Or) was assayed as described previously (4) with the following modification. The reaction mixture (0.85 ml of 10 mM potassium phosphate [pH 7.2] and 0.05 ml of 10 mM methyl viologen) in 1-ml Zumberg-type cuvettes was flushed with argon and an appropriate volume of 1 mM sodium dithionite to give an A₆₀₀ of 1.0 ± 0.2. An aliquot of the enzyme was injected, and the A₆₀₀ was monitored for 1 min to measure the background oxidation rate before starting the reaction by injecting 0.08 ml of N₂O gas. The reaction was measured by monitoring the N₂O-dependent oxidation of reduced methyl viologen at 600 nm. The protein concentration was determined with a protein assay reagent (Bio-Rad Laboratories, Inc., Richmond, Calif.).

RESULTS

Screening actinomycetes that produce N₂.

We examined denitrification in the following strains: Saccharopolyspora gregorii IFO15045T, Microbispora grisealba IFO14840T, Nocardia asteroides IFO3384, Nocardia brasiliensis IFO14402T, Streptomyces prunicolor IFO13075T, Streptomyces roseolus IFO12816T, Streptomyces acrimycini IFO12736T, Streptomyces badius IFO12745T, Nocardia carnea IFO14403T, Streptomyces daghestanicus IFO12762T, Streptomyces umbrinus IFO13091T, Streptomyces variabilis IFO12825T, Micromonospora citrae IFO14025T, Streptomyces coelicolor JCM4375, Streptomyces lividans JCM4783, Nocardia salmonicida JCM4826, Streptomyces rubescens IAM0074, S. antibioticus B-546, strain B, strain C, strain E, and strain N. Each strain was initially cultured under aerobic conditions (see Materials and Methods) in medium containing NO₃⁻, 10% of which was labeled with a heavy isotope ([¹⁵N]NO₃⁻). The headspace gas of the culture flask was analyzed after the incubation by GC-MS for isotopically labeled N₂ gas. Among the strains tested, only S. antibioticus B-546 produced a significant amount (9.5 × 10¹ μmol) of N₂ species (¹⁴N¹⁵N) after a 4-day incubation, indicating N₂ production from NO₃⁻. We then screened four other S. antibioticus strains (JCM3117, JCM4690, IFO12652, and IFO13271) and found that strain JCM3117 evolved up to 1.9 × 10¹ μmol of N₂ species (¹⁴N¹⁵N) after a 4-day incubation. These results showed that strains of this actinomycete species evolve more N₂ than do the other species tested.

Codenitrification and denitrification by S. antibioticus.

We investigated N₂ production by S. antibioticus B-546 in more detail. The strain was initially incubated under the aerobic conditions described above but with the stable isotope species of NO₃⁻, in which 99% of N atoms are labeled with ¹⁵N. We changed the content of ¹⁵N with the expectation that the ¹⁴N¹⁵N-N₂ species observed in the previous experiment would be replaced by ¹⁵N¹⁵N-N₂ species. However, the major N₂ species produced remained ¹⁴N¹⁵N (Fig. 1) with a very small but distinct amount of ¹⁵N¹⁵N. Furthermore, the N₂O that was also formed was in fact a ¹⁵N¹⁵NO species (data not shown). Neither ¹⁵N¹⁵N nor ¹⁵N¹⁵NO was observed in the control experiment without cells (data not shown), indicating that ¹⁵N¹⁵N and ¹⁵N¹⁵NO were formed because of denitrification. Since almost all of the added NO₃⁻ was of the heavy isotope species, the ¹⁴N¹⁵N-N₂ species should be generated by a combination of two nitrogen atoms, one derived from NO₃⁻ and the other from other nitrogen sources (codenitrification). The hybrid N₂ species was not detected in the control experiment without the cells. The medium contained 10 mM phosphate buffer (pH 7.2), and the final pH of the medium after 5 days of incubation was above 6.0. These results rule out the possibility that the hybrid N₂ species was formed by chemical means (2) but indicate that N₂ arises from a physiological reaction like that in fungal denitrifiers (10, 14).

FIG. 1.

N₂ production by S. antibioticus B-546 under aerobic conditions. Flask headspace was filled with air to initially maintain aerobic conditions. (A) Amounts of NO₃⁻ (solid circles) and NO₂⁻ (open circles) in medium and O₂ in the gas phase (solid squares) were determined at indicated times. (B) Amounts of ¹⁴N¹⁵N (solid circles), ¹⁵N¹⁵N (open circles), and N₂O (solid squares) in the gas phase were determined at indicated times. (C) Cell growth. Results are representative values from three experiments.

Time-dependent changes in each component during the initial aerobic culture show that the culture was divided into two phases (Fig. 1). In the first phase, NO₃⁻ and O₂ were rapidly consumed with the concomitant production of NO₂⁻ (Fig. 1A) and ¹⁵N¹⁵N-N₂ species (Fig. 1B), as well as cell growth (Fig. 1C). Most (70%) nitrogen atoms of the consumed NO₃⁻ seemed to be recovered into NO₂⁻. However, these activities almost stopped or significantly declined within 24 h. During the second phase after the 1-day incubation, N₂O began to be evolved whereas the emission of ¹⁵N¹⁵N-N₂ species declined. By contrast, the hybrid ¹⁴N¹⁵N-N₂ species was emitted during the first and the second phases throughout the incubation. Some portion of the nitrogen atoms of consumed NO₃⁻ were recovered into ¹⁴N¹⁵N-N₂ (6%), ¹⁵N¹⁵N-N₂ (0.0001%), and N₂O (0.3%). In contrast, NO was not detected throughout the incubation. The consumption of O₂ stopped in the second phase, although O₂ still accounted for 2% of the headspace gas. These results showed that the strain can reduce ¹⁵NO₃⁻ to produce ¹⁵N¹⁴N, ¹⁵N¹⁵N, N₂O, and NO₂⁻ species under aerobic conditions, which is in sharp contrast to other bacterial denitrifiers.

Nitrate metabolism by S. antibioticus under anoxic conditions.

The aerating (initial aerobic) conditions above (Fig. 1) mimicked those required for fungal denitrification (18). Next we examined similar cultures under anoxic conditions. Figure 2 shows that NO₃⁻ was first converted to NO₂⁻ as in the aerobic culture (Fig. 1), that the conversion was almost stoichiometric, and that ¹⁵N¹⁴N, ¹⁵N¹⁵N, and N₂O gas species evolved. Cell growth was distinct even under anoxic conditions (Fig. 2C) although cell yield was much lower than that under aerobic conditions (Fig. 1C). Figure 3 shows the time-dependent accumulation of each gas species per milligram of cells (dry weight) during aerobic (Fig. 1) and anoxic (Fig. 2) cultures. More of all of the gas species accumulated (per milligram of cells) under anoxic conditions. In particular, the amounts of N₂O evolved were increased by 500- to 1,000-fold under anoxic, compared with aerobic, conditions.

FIG. 2.

N₂ production by S. antibioticus B-546 under anaerobic conditions. Flask headspace was filled with argon to maintain anaerobic conditions. Symbols are as described for Fig. 1. Results are representative values from three experiments.

FIG. 3.

Denitrification by S. antibioticus B-546. Time-dependent accumulation of each gas species per milligram of dry cells was calculated and plotted. Circles, aerobic culture (Fig. 1); rectangles, anoxic culture (Fig. 2).

Denitrifying enzyme activities.

The formation of ¹⁵N¹⁵N, ¹⁵N¹⁵NO, and NO₂⁻ from ¹⁵NO₃⁻ suggests that S. antibioticus B-546 contains a conventional denitrifying system common to bacteria. We therefore examined the reconstitution of cell-free activities of enzymes that should be involved in the system. Table 1 shows that nitrate reductase (Nar), nitrite reductase (Nir), and nitric oxide reductase (Nor) activities were detected in crude extract or subcellular fractions prepared from the denitrifying cells (on day 3 in Fig. 1). Significant Nar, Nir, and Nor activities were recovered in the membranes as well as in the soluble fraction. In contrast, we could not reconstitute nitrous oxide reductase (N₂Or) activity.

TABLE 1.

Denitrifying enzyme activities of S. antibioticus B-546^a

Denitrifying enzyme	Total activity (nmol min−1)
	Crude extract	Soluble fraction	Membranes
NO−3 reductase	12.5	6.7	2.7
NO−2 reductase	396	260	49
NO reductase	56	13	6.0

^aS. antibioticus B-546 was initially cultured under aerobic conditions for 3 days, homogenized, and then fractionated as described in Materials and Methods.

Effect of acetylene on denitrification of S. antibioticus.

Since cell-free N₂Or activity could not be reconstituted, we further examined this step with respect to the in vivo activity of the actinomycete strain. N₂Or activities both in vivo and in vitro are generally inhibited by acetylene; thus, in vivo incubation of the denitrifying system usually results in N₂O accumulation (17). When S. antibioticus B-546 was initially cultured under aerobic conditions, acetylene did not affect the in vivo production of ¹⁴N¹⁵N, ¹⁵N¹⁵N, and N₂O from ¹⁵NO₃⁻ (data not shown), showing that the production of N₂ by strain B-546 is insensitive to acetylene.

Dissimilar nitrate reduction by S. antibioticus.

The above results indicate that nitrate metabolism by S. antibioticus supports anoxic cell growth (Fig. 2C). This notion was further supported by the following findings. The cell-free specific activity of Nar was higher in the anoxic cells than in those grown under aerobic conditions (Fig. 4), which is consistent with the complete conversion of NO₃⁻ to NO₂⁻ (Fig. 2A). The addition of tungstate, an inhibitor of molybdenum enzymes, to the culture medium considerably inhibited the cell-free enzyme activity in both the aerobic and anoxic cells, indicating that Nar activity is dependent on a molybdenum enzyme (Fig. 4). These properties are similar to those of bacterial Nar, which is involved in nitrate respiration (20).

FIG. 4.

Nitrate reductase activity of S. antibioticus B-546 under anaerobic conditions. Flask headspace was filled with air (bars 1 and 2) or argon (bars 3 and 4) and incubated for 24 h (the first phase) in medium with (bars 2 and 4) or without (bars 1 and 3) 10 mM sodium tungstate. Results are typical of over three experiments.

NO₃⁻ reduction during the first phase accompanied cell growth even in the absence of O₂ (Fig. 2). When NO₃⁻ was omitted from medium that still contained another nitrogen source (peptone), little growth was attained by the anoxic culture in sharp contrast to the normal growth attained under (initial) aerobic conditions in the absence of NO₃⁻ (Fig. 5). The addition of tungstate to the culture inhibited the NO₃⁻-dependent cell growth under anoxic conditions (Fig. 5, bar 5), again contrasted with the absence of a tungstate effect on the aerobic growth (data not shown). These results unequivocally demonstrated that anoxic cell growth depended on the Nar-dependent, dissimilatory reduction of NO₃⁻, whereas the high cell yield by the aerobic culture was mostly supported by O₂ respiration.

FIG. 5.

Nitrate-dependent cell growth of S. antibioticus B-546 under anaerobic conditions. Flask headspace was filled with air (bars 1 and 2) or argon (bars 3 to 5) and incubated for 24 h (the first phase in Fig. 1 and 2). GP medium was supplemented with (bars 2, 4, and 5) or without (bars 1 and 3) 10 mM NaNO₃. Sodium tungstate (10 mM) was added (bar 5). Typical results of more than three experiments are shown.

The anoxic cell growth (Fig. 2C) consisted of two phases, consistent with the initial reduction of NO₃⁻ to NO₂⁻ and the following initiation of N₂O evolution (Fig. 2A and B). These results support the notion that this actinomycete strain prefers, as the anoxic energy-yielding process, the reduction step from NO₃⁻ to NO₂⁻ by Nar to the subsequent reduction steps. When NO₃⁻ became unavailable due to consumption, the strain began to utilize the accumulated NO₂⁻ for further reduction, which resulted in denitrification. The strain was selective about the denitrification product, preferring N₂O to N₂ (detected above as ¹⁵N¹⁵N). The Nir and Nor activity detected in the cell extracts (Table 1) should correspond to the further reduction of NO₂⁻ to N₂O.

DISCUSSION

We are the first to demonstrate that denitrifiers produce N₂ as the denitrification product in actinomycetes. We also showed that the denitrifying system of S. antibioticus B-546 is distinct from other known systems of bacteria in that most N₂ production depends on codenitrification, where N₂ is formed by combining two nitrogen atoms, one from NO₃⁻ and one from other nitrogen sources. The hybrid N₂ species appears to be formed from NO₂⁻ (or NO) as it is in fungal codenitrification (10) but not directly from NO₃⁻, since the formation continued long after NO₃⁻ was consumed (Fig. 1 and 2). Another N₂ species (¹⁵N¹⁵N from ¹⁵NO₃⁻) should be formed by reducing N₂O since both nitrogen atoms of N₂O were derived from NO₃⁻ (¹⁵N¹⁵NO when ¹⁵NO₃⁻ was applied), showing that both products (N₂O and homo-N₂ species) are formed by normal denitrification. Thus, N₂ is formed in S. antibioticus B-546 via dual pathways of nitrate metabolism, namely, denitrification and codenitrification. Less hybrid N₂ species than denitrification product (N₂O) is produced by known fungal codenitrification processes (10). Therefore, the unique feature of nitrate metabolism by the actinomycete strain is that codenitrification predominates over, or at least is comparable to, denitrification.

These results are also the first to show that an actinomycete strain can attain growth even when the O₂ supply is completely lost (Fig. 2) and that anoxic growth depends on denitrification (nitrate respiration) (Fig. 2 and 4). Another unique feature of this type of nitrate metabolism is its aerobic or O₂-resistant nature, which is in sharp contrast to the nitrate metabolic processes of other bacteria that are suppressed by atmospheric O₂ (3). NO₃⁻ was converted to NO₂⁻ and all the gas species were formed under both aerobic and anoxic conditions (Fig. 1 and 2), although the formation rate for each metabolite differed between the conditions. The formation rate per milligram of cell (Fig. 3) for the denitrification products (N₂O and ¹⁵N¹⁵N) was much higher under anoxic than under aerobic conditions, supporting the importance of denitrification for anoxic growth. By contrast, the rates for conversion of NO₃⁻ to NO₂⁻ or the formation of the hybrid N₂ gas species (¹⁴N¹⁵N) were comparable between the conditions. The aerobic nitrate metabolism does not apparently contribute to cell growth (Fig. 1C and 5). However, this does not rule out the possibility that nitrate metabolism is an energy-yielding process, since NO₃⁻ was converted to NO₂⁻ (possibly supported by Nar) simultaneously with O₂ uptake, and O₂ respiration should be much more effective than nitrate respiration for energy production. This means that, under conditions such as the first stage shown in Fig. 1, suppressing nitrate respiration would be energetically even more favorable because nitrate should compete for electrons with O₂. In the second stage of the aerobic culture, the cell mass remained almost constant (Fig. 1C). The number of cells grown under aerobic conditions should have been too large for the weak denitrifying activity to support more growth (compare Fig. 1C and 2C).

Codenitrification was expressed in S. antibioticus irrespective of the extent of aeration. Thus, codenitrification appears to be associated with the formation of NO₂⁻ from NO₃⁻. However, the present results cannot confirm whether codenitrification contributes to cell growth. Another nitrogen source for codenitrification could be the peptone that was added to the medium as well as nitrate. The O₂-resistant nature of Nar along with codenitrification in the actinomycete strain has provided new insight into the dissimilar metabolism of nitrate by microorganisms, whereas their physiological significance or mechanism(s) remains to be elucidated.

We recently showed that many soil fungi are not obligatory aerobes but facultative anaerobes and that facultative anaerobiosis is supported not only by denitrification but also by ammonia fermentation, which is a novel anoxic metabolism of nitrate by fungi (19). Most known actinomycetes reside in soils, and like fungi, they have been considered obligatory aerobes. Therefore, the present study has revealed the possibility that many soil actinomycetes are in fact facultative anaerobes.

Our characterization indicates that the denitrifying system of the actinomycetes is more similar to that of fungal systems than bacterial systems. Codenitrification (10, 14), aerobic qualities (9, 18), formation of N₂O as the main product of denitrification (9, 10, 15), and rapid excretion of NO₂⁻ into the medium after conversion from NO₃⁻ (18) are all features of fungal denitrification. On the other hand, the P450nor involved in the fungal systems (Nor) seems not to occur among actinomycetes, although P450nor is classified into the same phylogenetic branch as other actinomycete P450s within the P450 superfamily (7). The phylogenetic relationship among the denitrifying systems of fungi, actinomycetes, and other bacteria requires further clarification. The contributions of the novel denitrifying systems of actinomycetes and of fungal systems to the nitrogen cycle in nature remain to be assessed.