Steady-state Nitric Oxide Concentrations during Denitrification”

Three species of denitrifying bacteria, Paracoccus denitrificans, Pseudomonas stutzeri strain JM300, and Achromobacter cycloclastes, were allowed to reduce nitrate or nitrite in anaerobic, closed vials while the equilibration of gases between aqueous and gas phases was facilitated by vigorous stirring. The gas phase was sampled and analyzed for NO with use of a chemiluminescence detector calibrated against bottled NO standards or against NO produced by the nitrite-iodide reaction. [NOaq] was inferred from [NOg] and the solubility of NO. NO was detected only during denitrification in amounts that, once established, did not change with time, were independent of the initial concentration of nitrate or nitrite, and were largely independent of cell concentration, at least when nitrate was the oxidant. The usual level of NO was promptly re-established following the addition of exogenous NO or following the loss of NO by sparging. The aforementioned properties are expected for a steady-state intermediate in denitrification. Steady-state [NOaq] ranged between 1 and 65 nM depending on species and conditions. Similar results were also obtained in a related experiment in which P. stutzeri strain ZoBell respired nitrite under growth conditions. The very low steady-state [NOaq] observed during denitrification imply that the maximum activity of nitric oxide reductase in vivo, if it could be realized, would be large relative to that for nitrite reductase. This circumstance allows NO to be an intermediate without reaching toxic steady-state levels.

during the steady-state of denitrification. A value of 2.2 nM was inferred for extracellular NO,, from the kinetics of the complexation of NO by extracellular hemoglobin and 560 nM was estimated for intracellular NO,, from the rate of denitrification and the diffusive mean residence time of NO,, (1). The latter was a rough estimate.  (4)) during the reduction of nitrite (3). This extrapolated value is taken to be the steady-state value and would therefore apply also to the intracellular [NO,,].
It seemed desirable to expand the data on steady-state [NO,,]  of NO, injected into the intake gas of the NO detector were always so low (a40 ppm), the subsequent dilution so rapid, and the time between injection and detection so short (5 s) that we observed no difference between the use of N, or air as the intake gas. The ability to use air is a consequence of the fact that the NO-O, reaction is second order in NO,, with k = 8 x lo9 ml2 x mol-* X s-i (12), and becomes slow in air when NO, drops below 100 ppm.  (Fig. 1). The two separate detector channels of the NO detector had slightly different sensitivities as indicated by the slopes for the two curves of Fig. 1. NO standards were run with every experiment, and the sensitivity of the detector was found to be very stable over a period of some 3 months over which the data were recorded. denitrificans was observed to decrease some 5-lo-fold when the cell concentration was lowered from 1 to 0.01 mg of protein X ml-'.
The reason for this dependence is at present unknown. Incubation of Pa. denitrificans with 10 pM azide, which completely blocked nitrate uptake and brought [NO,,] to undetectable levels, had no effect on the rate of nitrite uptake or the [NO.,] generated from nitrite. This rules out the possibility that additional NO was being made from nitrite by a pathway involving the direct reduction of nitrite to NO by nitrate reductase, as can occur with enteric bacteria (16). The fact that YP and SP media gave very similar [NO,,] values in the case of Pa. denitrificans, reducing nitrite tends to rule out explanations based on selective utilization of certain reducing substrates relative to others.
The judgment that the [NO.,] values of Table I represented the steady-state was based in part on their constancy over time, rapid recovery from perturbations, and reproducibility between two or more separate episodes of denitrification with the same lot of cells. NO., was always undetectable prior to the initiation of denitrification, reached some value within 5-10 min after addition of nitrate or nitrite, remained at this value until the N-oxide was exhausted and then returned to undetectable levels within the equilibration time of 5-10 min. This sequence of events could be repeated closely by addition of a second amount of nitrate or nitrite. Recovery from perturbation can be illustrated with the following examples. A system containing Pa. denitrificans (1 mg of protein X ml-') and reducing 10 mM nitrite showed a steady-state [NO.,] of 36 nM. When the system was sparged with Nz to lower NO  (6) 7-17 (8) 17-21 (6)  "The designator, R = 8 or R = 10 following a range of values signifies that the ratio, R, of initial [nitrate] or [nitrite] (in millimolar) to the cell density (in milligrams of protein X ml-') was constant at 8 or 10. This situation allowed a running t,ime for denitrification of 23-46min.
levels, the apparent [NO.,] immediately after sparging was 3 nM but rose to 40 nM at the next sampling 10 min later. In an analogous experiment, exogenous NO,, was injected to give a concentration of 130 nM, but the system returned to 39 nM within 15 min and remained at about that level until nitrite was exhausted. During the short term experiments, little or no increase in cell density occurred.
The results of long term experiments with P. stutzeri ZoBell are given in Table II. The time courses for appearance and  disappearance of nitrite, NO, and N20 qualitatively resembled those reported previously (3). [NO,] rose initially to rather stable plateau values which were maintained only as long as nitrite was present.
These plateau values, expressed as [NO,], are reported in Table II. During the period over which data were taken, the cells grew with a doubling time of 3-4 h.' Growth depended on nitrite respiration and was not fermentative. Initial cell concentrations were about lo6 cells X ml-' (0.2 c(g of protein x ml-') and particulate carbon values lay in the range of 0.5-5 mg of C x liter-' (about 0.25-2.5 pg of protein x ml-') during the times that NO determinations were being made. These cell densities were 4-40 times lower than the lowest cell densities for which [NO.,] values are reported in Table I. The [NO,,] values of Table II showed no discernable dependence on medium or nitrite concentration.
Controls-A number of chemical and biological controls were used to verify that the source of NO., was bacterial denitrification.
In the short term experiments NO was not detected in the absence of denitrifying bacteria, in the presence of an equivalent amount of heat-inactivated denitrifying bacteria, and in the presence of denitrifying bacteria but absence of nitrate and nitrite. As indicated above, 10 fiM azide prevented detection of NO from nitrate by virtue of its powerful inhibition of nitrate reductase (17). Similarly, NO was not detected when denitrifying bacteria were replaced with equivalent concentrations of aerobically grown Pa. denitrificans or E. coli, or with W. succinogenes grown anaerobically on nitrate. The denitrification apparatus apparently remained ' T. T. Packard (1986), personal communication.  (18).
In experiments with uninoculated medium contained in the gas recirculation system, the progressive appearance of extremely low levels of NO (of the order of 1 nM for NO,, after some 8 h) was routinely observed. The chemistry leading to this production of NO is unknown. Scavenging of NO in the gas recirculation system used for longer term experiments was assessed by injecting NO, into the head space of the system charged with sterile medium and following its disappearance with time. The amount of NO, introduced provided an equilibrium [NO,,] of 2-10 nM. No uptake of NO was observed within the analytical error (f4%) over several hours.

Technical
Limitations-Steady-state [NO.,] in the short term experiments proved to be lower than expected from Table I or undetectable for systems containing 1 pg of protein X ml-'. We suggest that this breakdown in the experiment may occur when the rate of reaction of NO with components of the system or with Oz as the result of its slow diffusion into the system begins to approach the flux of nitrogen through the NO., pool due to denitrification.
This flux can be approximated as the rate of nitrite uptake in the steady-state. At the highest cell densities used (1 mg of protein x ml-') the flux was about 300 nmol of N x min-' X ml-' or about lo4 times greater per min than the size of the steady-state pool of NO,, which we shall take here to be 30 pmol X ml-'. But at a cell density of 1 fig of protein x ml-', the flux should decrease to 300 pmol of N X min-' X ml-l or only some 10 times greater per min than the pool size. At some point, the size of the NO,, pool will be diminished by competing reactions that scavenge NO. Although we have been unable satisfactorily to quantitate the instability of lo-' M levels of NO,, in the systems used for the short term experiments, it is clear that NO at low concentrations was not stable in these systems and decayed in a few minutes or tens of min if not replenished. On the other hand, NO appeared to be stable for at least 30 min in the KI/acetic acid mixtures used to generate the NO standards.
Because the scavenging of NO was negligible in the gas recirculation system used for the long term experiments, it was possible to use this system to obtain data at cell densities in the vicinity of 1 pg of protein x ml-'.

DISCUSSION
Methods based on the equilibration of NO between gas and aqueous phases allowed a determination of the extracellular [NO,,] during denitrification for several bacteria under a variety of conditions. The properties of extracellular NO,, inferred in these studies are, for the most part, those expected for a steady-state intermediate in denitrification. Because the systems used were at equilibrium, the extracellular [NO.,] determined can be assumed also to hold for the intracellular concentration. The results do not bear on whether the reduction of nitrite in denitrification proceeds by two pathways, one generating NO as an intermediate and the other not, because the methods used did not permit determination of the kinetics of approach to the steady-state or a decay from it. The frequency of sampling was limited by the equilibration time between gas and aqueous phases, which was 5-10 min in the short term experiments and about 10 min in the long term experiments. It has been shown, however, that at least 60-70% of the flux of N in certain denitrifying bacteria (1,3) and 35% in another (P. stutzeri JM300) (1) passes through a pool of extracellular NO.,. Diffusion theory suggested that much or all of the remaining flux of N may pass through an intracellular pool of NO., (1). The values for [NO.,] reported in Tables I and II are in general agreement with a value of 2.2 nM (Pa. denitrificam) calculated from the kinetics of trapping of NO,, by hemoglobin (1) and a value of about 50 nM (P. stutzeri ZoBell) based on the flow rate and stripping efficiency of a single-pass sparged system in which denitrification was supported by low concentrations of nitrite (3). It is particularly reassuring that the previous (3) and present determinations of steady-state NO,, for P. stutzeri ZoBell agree so well. The system used in the first study was a dynamic one in which much of the initial nitrogen was swept out, whereas the present system was an equilibrium, closed one.
Knowledge of the steady-state [NO.,] is useful, along with the maximum velocity of nitrite uptake, VNi, and the K,,, for NO uptake KmcNOj, in estimating VNO. Assuming Michaelis-Menten kinetics, VN~ = VNi([NOaq] + K,,,(N~,)/[NO.,] in the steady-state. The only direct experimental value for Km(~o) of which we are aware is 400 nM and it relates to dilute suspensions of P. stutzeri ZoBell (3). That system also yielded a steady-state [NO,] C= 50 nM. Accordingly, VNo/VNi = 9. The data also yielded V&VNi = 40, but this estimate was dependent on the kinetic model used. Experiments in which both VNO and VNi were measured, the former at rather high (lo-100 pM) [NO.,], gave VNo/VNi 2 2-3 (2). This value is smaller than the actual one because the levels of NO used partially inhibited NO uptake. The (NO,] of Tables I and II, when coupled with KmcNOj = 400 nM, predict VNo/VNi = 7-400. These results overall suggest that nitric oxide reductase of denitrifying bacteria is a comparatively very active enzyme, probably the most active of the N-oxide reductases in uiuo, and is capable of maintaining NO., at such low steady-state concentrations that the potential toxicity of NO (19,20) is not realized. This represents a simple but effective solution to a common evolutionary problem among denitrifiers. A second utility of a highly active enzyme for the reduction of NO is the interception of intracellular NO, before it can diffuse from the cell. Efficient interception and reduction would tend to maximize the energy yield of denitrification, at least in open systems.
As cited above, studies to date (1, 3) indicate that 35-65% of the flux of N in denitrification may be intercepted and unavailable to diffusive loss as NO.
It is likely that a bacterial enzyme must itself be diffusion controlled if it is to compete against diffusive loss of a compound which faces no barrier to diffusion. Purified nitric oxide reductase from P. stutzeri ZoBell is reported to have a VNO of 40 rmol of NO X min-' X mg-' of protein in an ascorbate-based reducing system and a M, = 55,000 (21). If Km(p,o) = 400 nM, k,,t/K,,,c~o, for this enzyme is 9 X lo7 W' X s-l. Therefore, the enzyme in uitro can operate at or near the ordinary diffusion controlled limit for proteins and enzymes (lo7-los M-' X s-l).
The lower values of [NO,] of Table I are within a factor of 10 or so of those obtained from the few observations extant on natural, anaerobic, aqueous environments where denitrification is thought to control [NO,] (22). [NO.,] values as high as 0.5 nM and turnover times as short as a few minutes have been observed in these environments.