Distribution of Nitrosomonas europaea and Paracoccus denitrificans Immobilized in Tubular Polymeric Gel for Nitrogen Removal

Abstract

To improve the cooperative removal of nitrogen by Nitrosomonas europaea and Paracoccus denitrificans, we controlled their distribution in a tubular gel. When ethanol was supplied inside the tubular gel as an electron donor, their distributions overlapped in the external region of the gel. By changing the electron donor from ethanol to gaseous hydrogen, the distribution of P. denitrificans shifted to the inside of the tube and was separated from that of N. europaea. The separation resulted in an increase of the oxidation rate of ammonia by 25%.

Cells immobilized by fixation into inorganic carriers or entrapment in polymeric gel beads can serve as biocatalysts within bioreactors and have been productively applied in the pharmaceutical (24) and food industries (20) as well as in the area of environmental protection. For example, bioreactors used for wastewater treatment include a biofilm containing bacterial cells fixed onto carriers (2, 9, 10, 27, 30) or gel beads (6, 12, 14, 18, 19). In such arrangements, only the surfaces of the carriers or beads have activity, because substrates and oxygen for the immobilized cells are supplied only from the surface (11, 16, 17, 23, 25, 31). Recently, however, improved methods were described whereby the oxygen-rich surface of a gel was used to mediate aerobic reactions, while the oxygen-poor interior was used for anaerobic reactions (7, 8, 13, 15, 26).

We have also reported on an immobilized cell bioreactor capable of both aerobic and anaerobic reactions (28, 29). Instead of beads, our bioreactor makes use of a tubular polymeric gel with an ammonia oxidizer, Nitrosomonas europaea, and a denitrifier, Paracoccus denitrificans, trapped within. The immobilized N. europaea at the external surface of the tube mediated ammonia oxidation aerobically, while the immobilized P. denitrificans mediated anaerobic reduction of nitrite to nitrogen on the inside of the tube. Ethanol serving as an electron donor for denitrification flowed through the lumen of the tube. This configuration was advantageous because it enabled our bioreactor to use the ethanol effectively without having to add any ethanol directly to wastewater. Unfortunately, since P. denitrificans is a facultative anaerobe with both aerobic and anaerobic respiratory enzymes (3, 32), it grew on both the inside and the external surface of the tube, and the distributions of P. denitrificans and N. europaea overlapped (29). P. denitrificans then competed with N. europaea for oxygen, which suppressed the ammonia oxidation rate by N. europaea. Here, we describe the use of molecular hydrogen instead of ethanol as an electron donor for denitrification. This substitution controlled the distribution of P. denitrificans and avoided the competition for oxygen.

Bacterial strains and their immobilization.

The ammonia oxidizer N. europaea IFO-14298 and the denitrifier P. denitrificans JCM-6892 were used in this study. As previously described (28), the bacteria were separately aerobically cultured at 30°C and then coimmobilized with photocross-linkable polymer PVA-SbQ (SPP-H-13; Toyo Gosei Kogyo Corp.). A glass tube was used as a mold to form the polymeric gel containing the bacteria into a tube (outside diameter, 12 mm; inside diameter, 5 mm; length, 125 mm).

Batch experiments for nitrogen removal.

Solutions of ammonia or nitrate (200 ml) were treated with the tubular gel at 30°C while stirring (100 rpm) (Fig. 1). The ammonia solution contained (per liter) 0.472 g of (NH₄)₂SO₄, 0.2 g of MgSO₄ · 7H₂O, 9 g of Na₂HPO₄, 1.5 g of KH₂PO₄, and 1 ml of solution containing trace elements (1). The nitrate solution contained the same components, except that the (NH₄)₂SO₄ was replaced with 0.722 g of KNO₃. Hydrogen gas (99.99%) serving as an electron donor for denitrification flowed through the lumen of the tube at a rate of 360 ml/h. To acclimate the tube, it was exposed to the ammonia solution for 96 h, after which the solution was renewed. This process was repeated four times.

FIG. 1.

Schematic diagram of a batch system with a tubular gel. Artificial wastewater (200 ml) containing 100 mg of N/liter (as ammonia or nitrate) was treated. The wastewater was agitated by stirring (100 rpm) at 30°C. Hydrogen gas serving as an electron donor for denitrification flowed through the lumen of the tube (360 ml/h).

Ammonia and nitrite in the solution were measured colorimetrically as previously described (5). Nitrate concentrations were determined using an ion-chromato analyzer (DX-AQ1120; Dionex Corp.) equipped with an IonPac AS12A column. Discharged gas from the tube was collected and analyzed using a gas-chromato analyzer (CP-2002; Chrompack Corp.) equipped with a PoraPLOT-Q column and a microthermal conductivity detector.

Fluorescent-antibody labeling.

Prior to and then after the five batch experiments, sections of gel were cut away from the tube, fixed, dehydrated, embedded in polyethylene glycol, and sliced using the procedure of Hunik et al. (11). The sliced sections were labeled for 45 min at 52°C in the dark with rabbit anti-N. europaea-fluorescein isothiocyanate and with rabbit anti-P. denitrificans-fluorescein isothiocyanate. The labeled sections were then examined under a fluorescent microscope as described previously (28). The distributions of fluorescence within the gels were analyzed in photomicrographs with an IPLab image analysis system (Spectrum Signal Analytics Corp.).

Nitrogen removal by the tubular gel bioreactor.

When the ammonia solution was aerobically treated with the tubular gel in a batch system, the ammonia concentration in the solution decreased linearly from 100 to 1.7 mg of N/liter within 72 h (Fig. 2A). The nitrite concentration in the solution increased to 10.0 mg of N/liter within the first 48 h and then decreased to 2.0 mg of N/liter during the next 48 h. Nitrate was not detected (<0.05 mg of N/liter) at any time during the experiment. When nitrate solution was treated in the same way, the concentration of nitrate decreased from 100 to 8.3 mg of N/liter within 72 h (Fig. 2B). A small amount of nitrite (∼3 mg of N/liter) but no ammonia (<0.05 mg of N/liter) was detected. Nitrogen gas production was thought to accumulate in the lumen of the tube, since nitrous oxide was detected within the lumen of the tube exposed to acetylene, which is the inhibitor of N₂O reductase (data not shown). Neither gas production nor leakage from the external surface of the tube was observed during the experiments. To confirm the utilization of hydrogen, the nitrate solution was treated with nitrogen (99.99%) flowing through the tube instead of hydrogen. Under these conditions, the nitrate concentration in the solution did not decrease, and neither ammonia nor nitrite was detected in the solution (data not shown).

FIG. 2.

Time-dependent changes of the nitrogen concentration in wastewater containing ammonia (A) or nitrate (B) during the first batch experiment. Data are expressed as means ± standard deviations (SD) (n = 4).

These results show that the tubular gel bioreactor can effectively remove ammonia, nitrite, and nitrate by aerobic nitrification and anaerobic denitrification with hydrogen as the electron donor and with the produced N₂ gas going out from the lumen of the tube. Our bioreactor did not depend on the denitrifying activity of N. europaea, which has been recently reported (4, 21, 22), since the N. europaea used in this study showed no denitrifying activity (28). N. europaea showed no growth in medium containing nitrite and nitrate with hydrogen (data not shown).

Acclimation of tubular gel.

To acclimate the tubular gels, batch experiments were repeated five times. In the second experiment, the ammonia concentration in the solution declined to <1 mg of N/liter within 48 h; during the same period, nitrite increased to 36.0 mg of N/liter and then decreased to <1 mg of N/liter over the following 24 h. In the third, fourth, and fifth batch experiments, ammonia and nitrite disappeared within 24 and 72 h, respectively. The rates of ammonia oxidation (ammonia to nitrite) and nitrogen removal (ammonia to nitrogen gas) by the tube were calculated on the basis of changes in the ammonia and nitrite concentrations during the initial 12 h of the batch experiments. During the first batch experiment, the rates were calculated, respectively, to be 1.623 ± 0.194 and 1.469 ± 0.441 g of N/day/m² of gel surface (Table 1). By the fifth experiment, these rates had increased to 6.886 ± 0.353 and 3.371 ± 0.239 g of N/day/m² of gel surface.

TABLE 1.

Changes in rates of ammonia oxidation and nitrogen removal in tubular gel by acclimation

Batch process	Ammonia oxidation (NH4+→NO2−)a	Nitrogen removal (NH4+→N2)a
1	1.623 ± 0.194	1.469 ± 0.441
2	4.860 ± 0.268	3.240 ± 0.759
3	5.391 ± 0.272	3.587 ± 0.244
4	6.120 ± 0.193	3.631 ± 0.007
5	6.886 ± 0.353	3.371 ± 0.239
CFb	5.518 ± 0.453	4.458 ± 0.887

^aRates were calculated as the number of grams of N per day per square meter of the external surface of the gel tube. Values are expressed as means ± SD (n = 3 or 4). 

^bThe conferred values (CF) were the rate of the oxidation or nitrogen removal in the tube using ethanol after 33-day acclimation, as previously published (29). 

The nitrogen removal rate by the tubular gel was half the ammonia oxidation rate (Table 1). This reflects, in part, the dimensions of the tube: the internal surface area (denitrification zone) was less than half that of the external surface area (nitrification zone). In experiments using ethanol as the electron donor where denitrification and nitrification zones were almost same size, the nitrogen removal rate was 19% lower than the ammonia oxidation rate (Table 2). In an experiment using a plate-shaped gel with external and internal surface areas of equal size, the nitrogen removal rate was 22% lower than the ammonia oxidation rate. The lower rate of nitrogen removal might be ascribed to the fact that denitrification requires prior nitrification and would therefore not exceed the ammonia oxidation rate in an ammonia-rich solution. On the other hand, we observed a nitrogen removal rate per unit area of internal surface that was higher than the ammonia oxidation rate (Table 2), indicating that when hydrogen is used, nitrogen removal rates can approach ammonia oxidation rates.

TABLE 2.

Relationship between rates of ammonia oxidation and nitrogen removal in tubular gel reactor

Electron donor for denitrification	Ammonia oxidation (NH4+→NO2−)a	Nitrogen removal (NH4+→N2)a	Relative rate (%)
Hydrogen	6.886 ± 0.353	8.090 ± 0.574b	117
Ethanol	5.518 ± 0.453	4.612 ± 0.918c	84

^aRates were calculated as the number of grams of N per day per square meter of gel surface. Values are expressed as means ± SD (n = 3 or 4). 

^bThe nitrogen removal rate is shown as the rate per square meter of internal surface area of the tube. 

^cThe surface area of the denitrification zone using ethanol was assumed to be 96.7% of the external surface area of the tube, since the denitrification zone existed at a 200-μm depth from the external surface. 

Bacterial distribution within the tubular gel bioreactor.

The distributions of N. europaea and P. denitrificans within the tubular gel were microscopically investigated with fluorescently-labeled antibodies. Before the batch experiments, small colonies of N. europaea and P. denitrificans were sparsely spread throughout the tube. After the fifth batch experiment, colonies of N. europaea were concentrated within a region extending from the external surface to a depth of 200 μm (Fig. 3A) and were not detected in the interior of the tube (Fig. 3B); colony sizes were largest in regions closest to the external surface and decreased in size as a function of distance from the external surface. Conversely, after the batch experiments, colonies of P. denitrificans were concentrated in a region extending from the internal surface to a depth of 100 μm (Fig. 3D), with colonies situated closer to the internal surface growing larger than those situated at a greater depth. The size and number of those colonies farthest from the internal surface did not increase during the course of the experiments (Fig. 3C). To compare the distributions of P. denitrificans and N. europaea more quantitatively, their respective distribution densities were assessed by analyzing the photomicrographs described above. Figure 4 shows that when hydrogen was used as the electron donor, the distribution of P. denitrificans was clearly separate from that of N. europaea.

FIG. 3.

Fluorescence photomicrographs of N. europaea (A and B) and P. denitrificans (C and D) in regions of a tubular gel cut in cross section. Panels A and C show the external surface of the tubular gel, whereas panels B and D show the internal surface. The corresponding areas within the tubular gel are illustrated in the center diagram.

FIG. 4.

Distributions of N. europaea and P. denitrificans within a tubular gel. The solid and dotted curves show the relative biomasses of N. europaea and P. denitrificans, respectively. These curves were determined by analysis of the images in Fig. 3A and D.

With hydrogen as the electron donor, ammonia was removed by the tubular bioreactor just as efficiently as with ethanol and without accumulating nitrite or nitrous oxide. Indeed, the ammonia oxidation rate was as much as 25% higher than that seen with ethanol (Table 2). What is more, when hydrogen served as the electron donor, the distributions of P. denitrificans and N. europaea were completely distinct from one another. Because hydrogen is only slightly soluble in water (0.9 mmol/liter at 0°C and 1 atm), its diffusion from the inner surface of the tube was limited, which restricted the distribution of P. denitrificans. This resultant clear separation of the two bacterial strains and the extinction of the competition for oxygen was most likely the reason for the 25% increase in the ammonia oxidation rate.

In summary, we showed that the distributions of P. denitrificans and N. europaea could be separated within a tubular gel bioreactor. Their separation effectively increased the cooperation between nitrification and denitrification within our bioreactor, and this approach should also be effective with a variety of other bioreactors requiring combinations of aerobic and anaerobic reactions.