Microbiological Basis of Phosphate Removal in the Activated Sludge Process for the Treatment of Wastewater G. W. FUHS AND MIN CHEN Abstract. Several strains resembling members of the Acbwtobacter-Moraxella-Mima group of bacteria were isolated from activated sludge-type sewage treatment plants designed for phosphate removal. The bacteria are obligate aerobes but utilize as carbon and energy sources low-molecular intermediates generated anaerobically, particularly acetate and ethanol. These bacteria can be shown to be responsible for the phosphate luxury uptake occurring in these treatment plants. The bacteria are physiologically unusual in that they perform luxury uptake of phosphates in a complete growth medium. Phosphate release occurs on addition of a carbon source to the carbon-starved bacteria, lowering pH or both. The bacteria persist in the system by virtue of their ability to form floc.

Removal of phosphates from wastewater is important in order to protect lakes and other natural waters from cultural eutrophication. Conventional biological treatment processes remove only 50% or less of the sewage phosphate, and substantial improvement is needed to achieve 90% or more removal to reach effluent concentrations of 0.5 to 1.0 mg phosphate per liter. This can be accomplished by chemical means either in physical-chemical treatment processes or as part of the activated sludge process of wastewater treatment [22]. In these processes, salts of iron, calcium, or aluminum are added to form sparingly soluble phosphates, which are then removed by settling. Observations at treatment plants in San Antonio, Texas [23] and Baltimore, Maryland [15] have indicated that activated sludge can accumulate substantial amounts of phosphorus without chemical additions. A controversy has arisen as to whether such accumulation is a biological phenomenon or a chemical effect. In the latter case the calcium from naturally hard water presumably forms an insoluble phosphate on the activated sludge floc during aeration, when carbon dioxide is expelled and the pH rises, and the precipitate is again dissolved in the settling tank, when respiratory carbon dioxide accumulates and the pH value decreases. This purely chemical mechanism, which has been advanced particularly by Menar and Jenkins [ ! 3], resembles calcium phosphate deposition on the surface of algal cells in high phosphate media under the effect of a light-dark cycle [3]. By contrast, Levin and Shapiro [10, 19, 20] advanced the theory that aeration promotes the uptake of excess phosphate by activated sludge bacteria, while lowering the pH value or anaerobiosis causes this phosphate to be released 119 MICROBIAL ECOLOGY, Vol. 2, 119-138 9 1975 by Springer Verlag New York Inc.

120

G . w . Fuhs and Min Chen

into the medium. Inhibition of phosphate accumulation by 2,4-dinitrophenol tends to support the idea of biological uptake. Such a mechanism would permit a "stripping" of phosphates from sewage during sludge aeration and a release into a much smaller volume during extended settling, when anaerobic conditions prevail. From the concentrated supernatant in the settling tank the phosphates can be removed chemically with greater efficiency than is possible with the wastewater. Such processes have in fact been developed on a purely empirical basis. An aerobic-anaerobic cycle has been incorporated into the Seneca Falls, New York wastewater treatment system ("Phostrip" Process by Biospherics, Rockville, Md.), while a conventional flow scheme is adhered to at the City of Baltimore Back River Wastewater Treatment Plant [ 15]. At the Baltimore plant, operators report that phosphate removal efficiencies have been quite variable. We believe, and our results confirm, that such failures are the result of the poor understanding of the underlying mechanisms and, as a consequence, poor process control. Prediction from biological theory would be that under the conditions of the activated sludge process, phosphate accumulation by bacteria in excess of immediate need ("luxury uptake") is a highly unlikely event. Luxury uptake typically occurs when growth is arrested by lack of a nutrient other than phosphate and of a source of carbon and energy. (An energy source is required for phosphate uptake. Also required but rarely ever limiting are potassium as a neutralizing cation and magnesium as a cofactor [7].) In the activated sludge process, by contrast, with domestic wastewater as the substrate, bacterial growth is limited by the supply of carbon and energy sources, while all other nutrients are present in excess. A phenomenon often confused with luxury uptake is the "phosphate overplus phenomenon," which also consists of uptake in excess of immediate need and which occurs when phosphate-starved microorganisms are transferred into a phosphate-rich growth medium [24]. Since phosphate starvation doest not occur at any step in the treatment of domestic wastewater, the overplus phenomenon is not likely to occur. The more recent literature gives conflicting evidence with regard to the biochemical nature of the phosphate stored in activated sludge. Some describe it as an acid-soluble granular polyphosphate [ 16] and others as an acid-soluble fraction [28]. In preliminary experiments with a bench-scale activated sludge plant which had been run under completely aerobic conditions with synthetic sewage [25] and which contained mainly zoogloea-type bacteria and a variety of Pseudomonas and Flavobacterium, we were able to reproduce the reversible chemical precipitation of phosphate on the floc as a function of water hardness. We could also demonstrate the capability of the bacteria for luxury uptake of phosphate and for the

Phosphate Removal in Activated Sludge Process

121

overplus phenomenon when unrealistic conditions of nutrient limitation were established. In this system, unmodified except for an aerobiosis-anaerobiosis cycle maintained over a period of 2 weeks, we were unable to reproduce reversible luxury uptake in a complete growth medium. The finding by Chen (unpublished results) that internally stored organic carbon in the form of poly-/3hydroxybutyrate can serve as an energy source for phosphate accumulation, thereby permitting separation of the phases of carbon and phosphate accumulation, offered little help, inasmuch as few bacteria in our mixture were capable of storing both poly-fl-hydroxybutyrate and polyphosphate. We concluded that differences in species composition between our laboratory system and the systems studied by the other authors were responsible for this discrepancy. To continue the study we obtained in January, 1973 a sample of activated sludge from the Baltimore Back River Wastewater Treatment Plant and a supply of primary effluent from the same source for the experiments described below. Confirmation of hypotheses developed from these experiments were obtained with material from the Seneca Falls wastewater treatment plant collected on N o v e m b e r 2, 1973. Methods and Results 1. Appearance of Phosphate-Removing Activated Sludge The activated sludge of the Baltimore and Seneca Falls plants differs from normal activated sludge in that it appears black. After acidification with mineral acids it releases hydrogen sulfide, indicating that the black color is due to ferrous and possibly other sulfides. Whether the sulfide is derived from proteinaceous sulfur alone or from the reduction of sulfate is not clear. The sludge contains colorless filamentous sulfur bacteria, which presumably are engaged in the oxidation of hydrogen sulfide in the aerobic phase of the process. The possible existence of a bacterial sulfur cycle in this sludge led us to suspect that cyclic phosphate binding may occur in a biologically mediated chemical process, with the formation of ferric phosphate in the aerobic phase and of ferrous sulfide with the liberation of phosphate if the sludge turns anaerobic. The sludge, however, retains its black color during the aerobic phase, and chemical oxidants and reductants failed to produce a significant binding or release of phosphate. If this mechanism exists, it is either too slow or otherwise not pronounced enough to be of consequence. The Baltimore sludge in our sample contained many round clumps of bacteria that could be classified as zoogloeal~ masses, but the characteristic colonies ofZoogloea sp. with their fingerlike protrusions were absent. Consistent with this observation, we found differences in the major metabolic routes of glucose utilization in the Baltimore sludge as opposed to the aerobic sludge grown previously in our laboratory.

2. Glucose Metabolic' Routes in Activated Sludge Glucose dissimilation of the laboratory-model sludge was studied by diluting 2 ml of activated sludge ( ~ 2000 mg/liter of suspended solids) to a volume of 10 ml by adding water and solutions containing glucose and 14C-labeled glucose. The final glucose concentrations were 50 mg]liter and 0,01 p, Ci/ml, respectively. Six different species of glucose were used: C-l; C-2: C-3; C-3,4; C-6, and uniformly labeled compounds (New England Nuclear Corp.). The time course of glucose uptake from the medium and the appearance of ~4CO~ were monitored until chemical

122

G . W . Fuhs and Min Chen

tests (anthrone) showed that the glucose had disappeared from the medium. At that point, oxidative assimilation ends and the respiration of substrates takes over. The radioactive CO2 was collected in ethanol-methyl-cellosolve (1:1). The incubation mixture was separated by 0.22-/z membrane filtration, and all fractions were processed for scintillation counting. The Baltimore sludge was studied in a similar manner by diluting settled sludge with 9 parts of water to give 1830 mg of suspended solids per liter. Three milliliter of the suspension was incubated

"a 100~ 0 "r'

~

1-C

~

4-C

i

6-C

0

i

1

2

3

4

i

5

Hours Fig. 1. Release of 14COo from specifically labeled C-glucose by aerobic activated sludge. C-2, not shown, virtually coincides with C-6. The C-4 pattern was determined as the difference between C-3 and C-3,4. The sum of CO._,released from carbon positions 3 and 4 equals the sum of CO~ released from positions 1 and 6. Chemical analysis showed that added glucose was consumed in approximately 2 hr.

100 r

? "D

.13

"6

5~

x

........

~176176

~ 9

...:..=

.....-,....,,. 8 .......

8

....z" 0

~ ,

1

~)

2

o-o

2--C 0 Metabolite3_c

~ o

o

3

4

8

5

1-C 0-4-c J

6

Hours Fig. 2. Disappearance of added 14C-glucose label from suspension of aerobic activated sludge. Uptake of C- I and C-4 corresponds to uptake of glucose. Apparent slow uptake of C-2, C-3, and C-6 indicates excretion of metabolic intermediate(s) from C-2,3 and C-5,6 of glucose (dotted curve).

Phosphate Removal in Activated Sludge Process

123

with unlabeled (50.1 rag/liter) and labeled (40 p.g/liter) g] ucose. The collection methods for ~4CO~and other details were as described by Wood and Chua [27]. This experinaent was kindly carried out by Dr. Lindsay W. Wood, In the laboratory-model sludge, the pattern of ~4CO.~release (Figs. 1 and 2) was primarily from the atomic positions I and 4 of glucose, indicating predominance of the Entner-Doudoroff pathway, which is common in aerobic organisms such as Pseudomonas and rnay be a property of Zoogloea also. There is some indication of recycling of the C-4-6 fragments, as it occurs in Pseudomonasfluorescens and Acetobacter xylinum [ I 1, 26]. By contrast, the activated sludge from Baltimore, tested 2 days after collection and in aerobiosis, showed a release of ~4CO2 predominantly from the C-3 and C-4 positions of glucose (Fig. 3). This indicates glucose degradation along the Embden-Meyerhof pathway, which is characteristic for many facultatively anaerobic bacteria, such as the hemofermentative and heterofermentative lactic acid bacteria. These result,; show a profound difference between the two types of activated sludge. 3. Chemical and Microscopic Observations on Phosphate Uptake in Baltimore Actiwlted Sludge

After the activated sludge from Baltimore had been transported to the laboratory (6 hr) and kept in the coldroom (4~ overnight, 300 ml of settled sludge were diluted with 1500 ml of the primary

5000

3,4--C 40OO

3-C

3000 U--C

2000 1-C 6-C 2--C 1000

20

40

60

80

100

120

Minutes

Fig. 3. ~4CO., release from specifically labeled glucose by Baltimore activated sludge. External ~zlucose was exhausted after approximately 60 rain. Additions of [3,4~4CJ glucose and ['6-~ 4C] glucose were similar within 7.53% with a maximum deviation of 16% (results of experiment by L. W. Wood.)

124

G.W. Fuhs and Min Chen

Fig. 4. Baltimore activated sludge, charged with phosphate. Polyphosphate stain, water mount. Zoogloea-like mass of specific microorganisms showing massive staining reaction (frame), embedded in large sludge floc. Slight pseudo-three-dimensional effect by interference contrast. Scale marker, I0/xm.

Fig. 5. Seneca Falls activated sludge, charged with phosphate. Same stain as in Fig. 4, but bright-field illumination without interference contrast. One larger aggregate and many smaller ones of polyphosphate-storing microorganisms embedded in large bacterial floc. Scale marker, 10/zm.

125

Phosphate Removal in Activated Sludge Process

effluent. This mixture was placed in a 3-1iter Erlenmeyer flask and aerated at 22~ with magnetic stirring by passing about 6 liter/rain of moist air through the mixture. After 4 hr of aeration, the stirrer and air supply were shut off, and the liquor was left standing for 20 hr, after which the cycle was repeated. From the first day on, orthophosphate in the supernatant was determined after Murphy and Riley [ 17], and the presence of intracellular accumulations of inorganic polyphosphates was determined by high-power light microscopy after applying the volutin stain I (see Ref. [ 14]). This procedure consists in staining the ethanol-fixed smear with methylene blue (1% aqueous solution, 5 min), followed by quick differentiation (5-15 sec) in 1% sulfuric acid. After thorough rinsing in water, the specimen was generally dried and observed in immersion oil. If morphological detail of the bacteria was to be preserved, the specimen was mounted in water with a coverslip and then examined under a Zeiss oil immersion microscope, either in bright field or with a Nomarski interference contrast system. The volutin stain may leave a faint blue tone with some cell detail. The polyphosphate granules, however, show a metachromatic effect and stand out in a dark purple tone. After the extended period of anaerobiosis in transit on the day of collection and during the following night, no polyphosphate granules were seen in the activated sludge. After feeding with primary effluent and 4 hr of aeration, massive deposits of polyphosphate were found in one morphological type of bacterium, while all others were devoid of such deposits (Figs. 4 and 5). At the same time, a marked removal of orthophosphate had occurred.

40

35 30 \ - x

2,4--Dinitrophenol added ~ ' ~ -

-

--

"(D-

-

-.Q

25

2O.

2O

JE O.

2

15

O 10

5

0

1

2

3

4

5

6

24

25

Hours Fig. 6. Orthophosphate uptake and release by Baltimore activated sludge during aeration (hours 0 - 5 ) and anaerobiosis (hours 5-24). A 300-ml sample o f settled sludge was fed with 1500 ml of primary effluent (3660 rag/liter suspended solids). Dinitrophenol addition was 2 mM/liter. Orthophosphate is reported as P.

126

G . W . Fuhs and Min Chen

Figure 6 shows the results of an aeration-anaerobiosis experiment with the Baltirfiore sludge and also demonstrates the inhibition of phosphate uptake by 2,4-dinitrophenol. After the sludge had been left standing overnight without aeration, a high concentration of phosphate was again found in the supernatant (Fig. 6), and microscopic examination revealed that the polyphosphate granules had decreased in number and size but had not disappeared completely (Fig. 7). Over the following several days, whenever phosphate uptake during a daily aeration period was poor, microscopic observation showed that this occurred not because the microorganism failed to accumulate phosphate during aeration but because it had failed to release phosphate during the preceding anaerobic phase. Even though the anaerobic phase was as long as 16 hr phosphate release was often incomplete. Staining with Sudan black B revealed that the microorganism is also capable of storing poly-/3-hydroxybutyrate.

4. Isolation of Phosphate-Accumulating Organisms Enrichment and isolation of the phosphate-storing organism was facilitated by two factors: (1) The organism was easily recognized by its cell size and shape. A coccobacillus of approximately l-/zm diameter, it was by far the largest bacterium in the assemblage, not counting the filamentous sulfur organisms (Fig. 8). (2) As the sludge was kept in the laboratory, the abundance of the microorganism increased. This phenomenon was subsequently explained by its preference for acetate and ethanol, intermediate products likely to occur in the fouling primary effluent, which was used for feeding. When dilutions of the sludge were plated on nutrient agar and colonies picked and examined microscopically, many isolates showed the desired characteristics, although minor differences in colonial morphology were observed. The isolates appeared to be identical in their ability to grow on filtered primary effluent or on activated sludge extract and to accumulate phosphate into microscopically identifiable granules on aeration. Finally, the identity of the isolates with the bacteria in the sludge was established by suspending in the activated sludge system a dialysis bag with an aeration tube containing the pure culture in sludge extract. The cytochemical reactions of the isolate and the wild form were parallel over the course of a 24-hr aeration cycle. Concurrently with the platings on nutrient agar, the bacteria were successfully isolated on an agar medium prepared from activated sludge extract. On the basis of subsequent studies, the following medium was developed for fast, efficient enrichment and growth of the microorganism: 5 g of Na-acetate; 2 g of (NH4)2 SO4; 0.5 g of MgSO~ 9 7 H20; 0.25 g of KH~PO4; 0.2 g of CaClz 9 2 H20; 200 ml of mixed liquor, autoclaved and filtered; 800 ml of distilled water; pH adjusted to 7.0.

5. Description of the Phosphate-Accumulating Bacteria Physiological and biochemical properties of the isolates were determined according to Conn et al. [2]. The oxidase test was performed according to Gordon and McLeod [5]. Utilization of organic compounds and amino acids was determined according to Baumann et al. [ 1] and by the auxanogram technique (see Ref. [12]). Presence or absence of gliding movement was determined by Halvorson's [6] procedure. The identification of the bacteria followed Skerman [21], Henriksen [8], and Baumann et al. [ 1]. The bacteria are aerobic gram-negative short rods, 0.8-1.2/xm wide, 1.0-1.5/zm long (when mounted in water). They are nonmotile (flagella absent), but twitching movements are observed on cytophaga agar (5 /zm/min). Other characteristics are as follows. Colony morphology on yeast-extract agar or nutrient agar: circular, glistening, smooth, moist, translucent, slightly gummy; margin entire, undulate; no pigment. Growth on yeast extract agar slant: filiform and otherwise as above, moderately abundant. Growth on potato slant: glistening colonies, moderately abundant.

Phosphate Removal in Activated Sludge Process

127

Fig. 7. Baltimore activated sludge after 16-hr anaerobic incubation. Aggregates of the specific microorganism; some polyphosphate granules remaining (frames). Scale marker, 10/zm.

Fig. 8. Aggregate of the phosphate-storing microorganism (frame) freshly mounted in water, unstained. Scale marker, 10/zm.

128

G . W . Fuhs and Min Chen

Although acetate was utilized by all isolates, a detailed study of carbon sources was performed only with one strain from the Baltimore plant, which accumulated phosphate most vigorously. Carbon sources are: ethanol, acetate, citrate, succinate, l-arginine, l-histidine, l-glutam[c acid, tyrosine. Dubious growth on n,L-serine, phthalic acid, phenol, pentane, 2-propanol, n-butynol, methanol, l-proline. No growth on glucose, lactose, arabinose, lactate, formate, glycerin, carboxylic acids from propionic through decanoic, benzoic acid, benzyl alcohol, ethyl benzene, tartrate, fl-hydroxybutyric acid, butynic alcohol, pyonoic acid, O,L-tryptophane, D,L-threonine, D,L-valine, D,t.-aspartic acid, L-leucine, D,L-methionine, O,L-alanine, L-isoleucine, L-lysine, L-cysteine, L-cystine. Acid produced from citrate. No acid produced from glucose, lactose, sucrose, mannitol, maltose, xylose, rhamnose, dulcitol. Nitrogen sources: requires combined nitrogen; utilizes ammonia and nitrate. Positive biochemicat reaction: catalase. Negative biochemical reactions: oxidase, urease, gelatinase. No indol or HzS formed; no nitrate reduction. Storage products: poly-/J-hydroxybutyrate, polyphosphate (metachromatic granules). Temperature range (generation times); 8~ 3.6 hr; 20~ 1.5-1.8 hr; 22"C, 1.0 hr; 24~ 1.0-1.5 hr; 28~ 2.0 hr; 37~ 45~ no growth. If activated sludge extract or yeast extract is omitted from the medium, growth appears to be slower. However, there is no absolute requirement for growth factors. The taxonomic position of this bacterium corresponds to that of Acinetobacter lwoffi in the sense of Baumann et al. (see Ref. [1]), except that it does accumulate poly-/3-hydroxybutyrate as a storage material and the range of carbon sources is even more restricted than in the strains ofA. lwoffi studied by these authors. We shall refer to it in the rest of this paper as an A. lwoffi isolate. For the other isolates, only the genus (Acinetobacter) has been established so far. 6. Phosphate Uptake by a Pure Culture of Acinetobacter Figure 9 shows an experiment with a pure culture of ourA. lwoffi isolate grown in the acetate medium described earlier. Nine-hundred milliliters of medium were dispensed into three acid-washed dialysis bags, which were then suspended in aerated liquor and inoculated. After 20 hr the culture was transferred into a l-liter cylinder and the surface sprayed with Dow-Corning silicon antifoam. The experiment was begun by aerating the culture for 4 hr with 1.5 liter of moistened sterile air per minute. During the experiment, the pH was adjusted to between 7.0 and 7.2. Anaerobiosis was produced by turning off the air supply and sealing the cylinder with plastic film. After 4 hr of aeration, the culture had removed 17 mg of phosphorus per liter of solution. After 20 hr of anaerobiosis, 24% of this amount was released into the supernatant, but release continued for ablaut 1 hr after aeration had resumed. During this second aeration period, phosphate uptake reached approximately the previous level. A second period of anaerobiosis, 20 hr long, failed to produce phosphate release. A third period of aeration lasting 6 hr caused release rather than uptake of phosphate. At this point, however, a new phase of phosphate uptake could be induced by adding a suitable carbon source, such as acetate. A similar experiment was conducted with another carbon source, ethanol, added at time 0, and 48.72, and 96 hr. While the starving culture failed to accumulate phosphate, ethanol, like acetate,

Phosphate Removal in Activated Sludge Process

1 29

caused rapid uptake. Release of phosphate during anaerobiosis was quite variable. Ethanol, unlike acetate, did not cause phosphate release in the first hour after addition.

7. Fractionation of lntracellular Phosphorus in Activated Sludge and in Acinetobacter Cultures Phosphorus fractions in activated sludge and bacterial samples were determined at various phases of the uptake experiments under four conditions: (a) without further treatment, (b) after treatment with 0.05 M ethylenediaminetetraacetate (EDTA) to release externally bound phosphate, (c) after gassing with CO., for 5 min, or (d) after setding and letting the culture stand overnight. Treatments (c) and (d) were intended to induce phosphate release by establishing anaerobiosis and to show which phosphate fractions were affected by the release. All samples were extracted for 30 rain and again for 15 rain with cold 5% trichloroacetic acid (TCA). After centrifugation at 12,000 g for 15 rain, the supernatant was divided into three equal portions. Two portions were used to determine orthophosphate and total phosphate. The third portion was treated with acid-washed charcoal to remove nucleotide phosphorus, and the filtrate was hydrolyzed with l N HCI for 7 rain at 100~ Orthophosphate found in the hydrolysate was determined (7-min phosphate). This fraction diminished by the orthophosphate value and by the 7-rain phosphate value was designated acid-soluble polyphosphate.

7.0 O-~ 6 ~

.

~

~

.

~

~

t

46~henoladded 42

38 E 34 O. t-

~ o

-4~-

3o

-4

o

26

20

0

Aeration

I

I

Aeration

I

I

2

4

24

26

28

48

Aeration 50

I 52

54

Hours Fig. 9. Uptake and release of external phosphate by a pure culture of Acinetobacter lwoffi. Upper curves: pH and dissolved oxygen. Lower curve: reactive (ortho)phosphate in medium. A YSI dissolved oxygen meter was used for the oxygen measurements. Phosphate is reported as P.

Table 1

7.0

Nucleic acid and protein P 40.8 2.0

7.6

2.8 7.0

6.3 4.9 12.2

39.4 0.7

8.7

1.8 5.5

8.1 5.2 I0.1

22.4

6.8

1.4 3.7

0 4.1 6.4

0

36.5

9.1

2.4 6.9

1.8 4.1 10.2

2.5

33.7

8.4

1.9 6.0

3.8 3.8 8.7

5

0.05 M EDTA (hr)

Treatment and time

28.8

7.0

1.6 4.3

4.7 4.9 6.3

0

38.0

8.0

3.2 7.1

5.1 3.7 10.9

2.5

7.8

1.6 8.1

5.5 4.5 8.6

5

36.1

CO 2 gas (hr)

aIn both the EDTA and CO 2 treatments, the pH decreased from 7.0 to 5.2. Suspended solids concentration was 2740 - 2800 rag/liter in all experiments.

31.8 32.0

1.6 4.0

Lipoid P Acid-insoluble polyphosphates

Sum Total P in supernatant (rag/liter)

4.6 4.4 10.2

Ortho P. Nucleotide P Acid-soluble polyphosphates

5

0

compound

2.5

No treatment (hr)

Phosphorus

Phosphorus Compounds (lag P/mg dry wt) in Mixed Liquor during Aeration of Baltimore Activated Sludgea

==

c~

O

Phosphate Removal in Activated Sludge Process

]31

The residue from the TCA extraction was treated with cold 95% ethanol and with ethanol-ether (3:1) for 1 hr each. The total phosphorus in the extract was lipid phosphorus. The residue was divided into two portions. In one portion, 7-min phosphate was determined and designated acid-insoluble polyphosphate. The other portion was used to determine total phosphorus, an d the difference was listed as protein and nucleic-acid phosphorus. The results with activated sludge (Table 1) suggest an accumulation of acid-insoluble polyphosphate, while the pure culture (Table 2) showed an increase of acid-soluble phosphates during the first phase of phosphorus uptake. In both cases some superficial (or chemical) fixation of phosphate took place, as indicated by the sol ubilization of phosphate by EDTA. Treatment with CO., gas resulted in a decrease in orthophosphate, or acid-soluble polyphosphate, or both. The acid-insoluble form of polyphosphate is generally thought to represent a mobile intermediate fraction, which exchanges phosphate both with the orthophosphate pool and with other fractions, e.g., nucleic acids, while the granular polyphosphates are the ultimate storage product (see Ref. [18]).The differentiation of acid-soluble and acid-insoluble polyphosphate, however, is sometimes difficult, as the acid-soluble (granular) form dissolves only slowly in 5% TCA and, unless completely dissolved, is bound to appear in the acid-insoluble fraction (see Ref. [4] for discussion). The phosphate uptake in these experiments caused the phosphorus content ofA. lwoffi to cycle between 3.6 and 5.0% (untreated) and between 2.9 and 4.0% (CO,, treated). This explains, within the limits of experimental accuracy, the drop of 17-19 mg of dis solved phosphorus per liter at a bacterial dry weight of 960 mg/liter.

8. Factors Affecting the Release of Phosphorus into the Medium The mechanism by which the Acinetobacter isolate releases phosphate is difficult to understand. Shapiro [ 19] and Shapiro et al. [20] think that anaerobiosis is a prerequisite for phosphate excretion or release. In activated sludge systems, however, anaerobiosis is accompanied by the accumulation of respiratory CO._,and, consequently, by a lowering of the pH. To separate these factors we exposed a culture of the A. hvoffi isolate which had accumulated phosphate to anaerobiosis (a) by gassing it with nitrogen, hydrogen, or helium gas and thereby expelling respiratory CO,,, while the pH remained at 7.0; (b) by gassing it with CO2, which lowered the pH value instantly to 5.5; and (c) by adding 0.01% acetic acid, which brought the pH to 6.0 and served as a carbon and energy source. The addition of CO2 or acetic acid caused instantaneous release of phosphate, while bubbling with the inert gases did not (Fig. 10). The results of this experiment can be reconciled with Levin and Shapiro's early observation of pH-induced phosphate release [ 1 0 ] . " A n o x i c release" [19, 20] may consist in the lowering of the pH in a microzone within the activated sludge floc, or it may be the expression of other chemical factors. Biospherics Inc. assert that in the recently installed plant at Seneca Falls, phosphate stripping (phosphate release) occurs in anaerobiosis, while the pH remains in the vicinity of 7 (G. V. Levin, personal communication). The possible involvement of unknown chemical factors in the release of intraeellular phosphate was investigated as follows. A 420-ml portion of activated sludge (2009 mg of suspended solids and 7.5 mg of orlhophosphate per liter in supernatant) was placed into each of four cylinders. To the first cylinder was added a cellophane bag containing 42 ml of an aerated and phosphate-charged A. lwoffi culture in acetate medium (2155 mg of suspended solids and 19.4 mg of orthophosphate per liter in the supematant). The seond cylinder received a cellophane bag with 42 ml of a 1:1 mixture of A. lwoffi culture and a solution of orthophosphate (19.4 rag/liter). The third cylinder received a bag with 42 ml of the orthophosphate solution alone. The fourth cylinder served as a control, and a portion of the bacterial suspension was kept separately as a control. All solutions and suspensions were adjusted to a pH value of 7.7, sealed from the air, and kept for 20 hr.

Table 2

14.0 14.2 13.6 13.6 14.5 13.6

12.1

36.0 49.9 42.3 39.8 47.5 48.1 44.1 7.0 7.2 7.1 7.0 7.1 7.1 7.0

Sum

2.5 6.7

4.5 7.9

3.1 7.9

6.2 3.5

5.7

3.0

2.4

4.7

5.5

5.6

5.5

5.6

35.7 29.2 40.7 35.1

13.0

2.2 4.9

2.5

6.1

6.7 2.1

12.9 10.8 13.5

4.0

4.3

5.8 5.6

48

6.2

5.2 2.1

6.2 4.1

28

6.7

24

4

CO 2 gas b (hr)

Treatment and time

14.6

11.7

5.5

5.5

30.3 39.5

7.5

3.2

5.8

3.7 4.7

~28

5.8

3.2

5.0

2.9 1.7

4

EDTA (hr)

7.2

49.1

14.8

7.5

3.2

8.2

11.2 4.2

54

Na-Acc (hr)

aDry wt of the bacterium, 960 mg/liter; external phosphate concn., 45 mg/l!ter; suspended solids, 925-980 mg/liter. bCO2 passed through to decrease pH. The 24- and 48-hr values were found after the sample had been kept anoxically overnight. CNa-acetate (0.2 g, pH 7.0) added to l liter of culture at 54 hr.

pH

3.2 5.8

5.7 6.3

9.8

Nucleic acid and protein P

2.8 6.2

6.0 7.9

9.2

2.7 6.6

5.0 6.0

6.2 10.8

2.6 5.1

Acid-soluble polyphosphates

5.0 5.4

8.7

Lipoid P Acid-insoluble potyphosphates

54

6.0 8.5

48

12.1

28

3.3 5.2

25

7.7

Ortho P Nucleotide P

24

4

0

Phosphorus compound

No treatment (hr)

Phosphorus Compounds (pg P/mg dry wt) in Acinetobacter lwoffi lsolate during Continuous Aerobic and Anaerobic Cycling of Baltimore Activated Sludge a

r=-

c-')

t~

133

Phosphate Removal in Activated Sludge Process

Orthophosphate was then determined in all components. The first cylinder (A. hvoffi) had 12.4 rag/liter in the sludge and 15.2 mg/liter in the culture. The second cylinder (A. Iwoffi plus orthophosphate) had 10.2 mg/liter in the sludge and 14.9 mg/liter in the diluted culture. The third cylinder (orthophosphate) had 8.5 mg/liter in the sludge and 12.7 mg/liter in the phosphate solution. The activated sludge control showed a slight increase to 9.5 rag/liter; the bacterial culture Control showed no change. Calculation of changes of total extracellular phosphate in all systems showed that the complete system had released 1881 /xg of phosphate-phosphorus. The sludge alone had released 840/xg. In contact with the higher concentrated orthophosphate solution which served as a control for the A. hvoffi suspension, however, only 138 /zg were released by the sludge. Since the A. /woffi culture which was kept separate from the sludge did not release any phosphate, the difference of 1881 /zg - 138 /~g = 1743 t.t.g wouid seem to represent bacterial release of phosphate stimulated by products from the anaerobic sludge. The experiment with a 1:1 mixture of A. lwoffi suspension and orthophosphate solution shows intermediate concentration values in all compartments and a release of 807 t~g or somewhat less than one-half of 1743 /~g.

40 ,~.'~ CO 2

35 Acetic acid I I

3O E E

25

I

/

If II iI if it iI

20 .E

I-t2 He

15 tt~

No treatment

10

I

I

I

|

1

2

3

4

Days in anaerobic state

Fig. 10. Phosphate release from culture o f A . hvoffi after anaerobiosis induced by CO,_, gas (pH 5.5), by adding 0.01% acetic acid (pH 6.0), or by bubbling with Na, H~, or helium (pH 7.0). At zero time the culture contained 4.55% phosphate after aeration in phosphate-acetate m e d i u m . Phosphate is reported as P.

134

G.W.

Fuhs and Min Chen

40 45~

4~ 37~ 10~

35

E "0

30

2c). 8 25 24~

20

20~ 1

2

3

4

Hours Fig. 11. Effecl of t e m p e r a t u r e on p h o s p h a t e uptake by A. /wojg'i' isolate in p h o s p h a t e - a c e t a t e m e d i u m . P h o s p h a t e is reported as P.

9. Temperature I~'[/t,cts on Pho,v~hale Ad.sorption

As reported above, the optimum growth temperature for the A. /wry}" isolate is 20~176 Similarly, phosphate uptake is nlaximunl at lhese lemperalures (Fig. I I). At 10~ and 37~ phosphale uptake is; significanlly inhibited. Theref~}re the activated sludge process with biological phosphate removal may not operate at winter temperatures in the higher latitudes or at midsummer temperatur,.:s in the lower latitudes, or it may require adaptation or strain selection. [0. Cotlfirmalot O, IErperiments with Aclivr

Slud&,e-/)'om the Seneca1 F~dls Wastewater Tre~llmettl Pl{ml

Sludge taken from the Seneca Falls plant on Novemt_x3r I, 1973 was transported to the Albany laboratory on the same day. Three portions were kept overnight, one without aeration, one with aeratkm, and the third with aeration and additions of settled primaff effluent. The aerobic portions showed heavy polyphosphate deposits in microcolonies of bacteria resembling theAcinetob~tcter from the Baltimore plant. The Acinetobacter in the portion which had not ~ e n aeraled showed small granules only. The remainder of the activated sludge bacteria showed no polyphosphate inclusions. The isolales have characteristics sinlilar Io those of the Baltimore Acittetoh~tcter strains, although the carbon sources utilized have not yet been determined in detail. The Seneca Fails sludge on microscopic

Phosphate Removal in Activated Sludge Process

135

examination also showed small numbers of a polyphosphate-accumulatingstreptococcus. This form has not yet been isolated. At Seneca Falls. New York, phosphate removal occurred in winter or below 10~ (G. V. Levin, personal communication).

Discussion Our observations confirm the opinion by Levin and Shapiro [10] that luxury uptake of phosphate can occur during microbiological treatment o f wastewater by the activated sludge process. This effect is essentially caused by a single microorganism or several very closely related forms which differ from other bacteria in their ability to store large amounts of polyphosphate in the presence of all nutrients required for growth. Phosphate storage in microscopically undetectable form by some other bacteria cannot be ruled out, and the decision in this matter must await electron microprobe or other analysis. The granular storage product which occupies a substantial portion of the cell volume of phosphate-charged Aeinetobacter cells has the properties of a highly packed mineral polyphosphate, (KPO3),,[4], and therefore must be a quantitatively significant storage form. In experiments not described above, addition o f a pure culture o f

Acinetobacter to our aerobic laboratory culture of activated sludge caused this system to accumulate phosphate without acclimatization.

Achtetobacter-type bacteria were present in both of the sewage treatment processes we have studied so far. In additon, Acinetobacter was found in an activated sludge plant in Newark, New York, which was sampled on or about November 1, 1973. The sludge in this plant had gone anaerobic for reasons not further investigated, and phosphate removal was never attempted. About 5 years ago, aerobically grown activated sludge had been used to seed our completely aerobic, laboratory activated-sludge plant which did not contain Ac'hletobacter and could not be induced to remove phosphate biologically. Baumann et al. [ 1] point out that AcbTetobacter is easily enriched from natural waters or soil using acetate, a compound which can be expected to form in anaerobic sewage treatment. Anaerobic conditions preceding aerobiosis in sewage treatment therefore could well be related to the appearance of Acinetobacter. In the activated sludge, Acinetobacter forms microcolonies which are held together by a capsular material of unknown composition (Fig. 7). It can bc assumed that this a g g r e g a t i o n - - o r the inclusion of the microcolonies in the activated sludge floc, as in Figs. 4, 5, and 8 - - l e a d s to the sedimentation of the bacteria. As a result they tend to remain in the cycling floc, while single bacteria are normally washed out.

136

G.W. Fuhs and Min Chen

The greatest discrepancy between our conclusions and the concept of Levin and Shapiro concerns the anaerobic phase of the treatment process. We found that the anaerobic phase is not itself necessary to induce release of phosphate from the bacteria into the supernatant. The CO2 accumulation and the lowering of the pH which result from anaerobiosis may be important factors, but they could be obtained by other measures, at least in theory. Even prolonged anaerobiosis (20 hr/day) does not always produce complete release of polyphosphate from the cells. We believe that this release level should b e - - a n d can b e - - i m p r o v e d upon. The principal function of the anaerobic treatment, in our present view, is to establish a facultatively anaerobic microflora, as indicated by the EmbdenMeyerhof fermentation pattern. During anaerobiosis this flora would tend to produce compounds, such as ethanol, acetate, and succinate, which serve as a carbon source forAcinetobacter. During the acclimatization phase, Acinetobacter would be enriched from the sewage. Without anaerobiosis, an obligately aerobic assemblage is likely to develop, and the intermediate products would not be formed. Acinetobacter, which cannot attack sugars or polysaccharides and are likely to be subject to heavy competition for the utilization of amino acids, would then not develop. In the Seneca Falls Plant, cycling of the activated sludge through an anaerobic phase is likely to have caused the enrichment of Acinetobacter during acclimatization, while in the Baltimore and Newark, New York plants, fermentation products originating from the primary settling tank or elsewhere in the system may have caused enrichment. This idea is supported by our observation that Acinetobacter sp. became increasingly abundant in the Baltimore activated sludge after it was fed with primary effluent which had been left standing, in the coldroom from several days to a week. While this sewage may have shown a decreasing biochemical oxygen demand (BOD), the concentration of fermentation products, such as acetate and ethanol, had certainly increased and made this medium most suitable for the enrichment of the phosphate-accumulating bacterium. In the presence of acetate, Acinetobacter can form poly-/3-hydroxybutyrate, a carbonaceous reserve material which can serve as a storage form for energy for the uptake of phosphorus. The accumulation of ethanol and presumably also lactate in the activated sludge during anaerobiosis has a possible side effect: The growth of sulfatereducing bacteria which, to the extent that sulfates are present, may compete with the Acinetobaeter for one of its substrates, ethanol. Detailed isotopic and chromatographic studies are needed to elucidate this aspect further. To the extent that the anaerobic phase in Levin's Phostrip process [9] serves to produce nutrients for theAcinetobacter and does little for the processes of BOD and phosphate removal, it may be a step either not necessary or not desirable for the most efficient treatment of sewage. Our findings suggest that other solutions

Phosphate Removal in Activated Sludge Process

137

are equally possible. One is the bleeding-in of fouled primary effluent or digestor supernatant to provide the necessary carbon sources for the control o f A c i n e t o b a c ter, while sewage gas, which may be stripped o f C02 in order to be more useful as a fuel, would provide the CO._, needed for mild acidification and phosphate release from the biomass. As another alternative, phosphate might be r em o v ed in a stable enrichment culture separate from the activated sludge process. There is little risk that the A c i n e t o b a c t e r will b e c o m e too abundant in any activated sludge system, since they rely on the action o f bacteria which produce low-molecular intermediates from carbohydrates, sugars, and other sewage components, which the A c i n e t o b a c t e r cannot utilize. Some questions remain with regard to the factors controlling the release o f orthophosphate from the cells. In our experiments, low pH, addition of a carbon source, and unidentified diffusible subtances from fermented sewage were more effective than anaerobiosis. It cannot be ruled out that, in addition, anaerobiosis may cause the death o f some A c i n e t o b a c t e r and consequently the release o f orthophosphate during autolysis.

References 1,

Baumann,P., M. Doudoroff, and R. Y. Stanier, 1968. A study of the Moraxella group. II. Oxidase-negative species (genusAcinetobacter). J. Bacteriol. 95: 1520-1541.

2.

Corm,H. J., M. W. Jennison, and O. B. Week. 1957. Routine tests for the identification of bacteria. In: "Manual of Microbiological Methods" [H. J. Corm, editor], pp. 140-168, McGraw-Hill, New York.

3.

Frank, E. 1962. Vergleichende Untersuchungen zum Calcium-, Kalium- und Phosphathaushalt von Gr/inalgen. I1. Calciummangel bei Hydrodictyon, Sphaeroplea und Chlorella. Flora 152: 157-167.

4.

Fuhs, G. W. 1969. Interference-microscopic observations on polyphosphate granules and gas vacuoles in cyanophyceae [ in German with English summary]. Osterr. Bot. Z. 116 (Geitler-Festschrifl): 411-422.

5.

Gordon, J., and J. W. McLeod. 1928. The practical application of the direct oxidase reaction in bacteriology. J. Pathol. Bacteriol. 31: 185-190.

6.

Halvorson, J. F. 1963. Gliding motility in the organisms Bacterium anitratum (B5W), Moraxella lwoffi and Alkaligenes haemolysans, as compared to Moraxella nonliquefaciens. Acta Path. Microbiol. Scand. 59: 200-204.

7.

Harold, F. M. 1966. Inorganic polyphosphates in biology: structure, metabolism, and function. Bact. Rev. 30: 772-784.

8.

Henriksen,S. D. 1963. Mimeae. The standing in nomenclature of the names of this tribus and of its genera and species, lntern. Bull. Bacteriol. Nomen. Taxon. 13: 51-57.

9.

Levin, G. V., inventor. U. S. Patent No. 236, 766 "'Sewage Treatment Process," issued February 22, 1966.

10.

Levin, G. V., and J. Shapiro. 1965. Metabolic uptake of phosphorus by wastewater organisms. J. Water Pollut. Control Fed. 37: 800-821.

138

G. W. Fuhs and Min Chen

11.

Lewis, K. F., H. J. Blumenthal, R. W. Weinrach, and S. Weinhouse. 1955. An isotope tracer study of glucose catabolism in Pseudomonas fluoreseens. J. Biol. Chem. 216: 273-286.

12.

Lodder, J., and N. J. W. Kreger-van Rij. 1952. "The Yeasts." North Holland, Amsterdam.

13.

Menar, A. B., and D. Jenkins. 1970. Fate of phosphorus in wastewater treatment processes: enhanced removal of phosphate by activated sludge. Environ. Sci. Teehnol. 4: 1115-1121.

14.

Meyer, A. 1904. Orientierende untersuchungen 'uber die verbreitung, morphologie und chemie des volutins. Bot. Ztg. 62: 113-152.

15.

Milbury, W. F., D. McCauley, and C. H. Hawthorne. 1971. Operation of conventional activated sludge for maximum phosphorus removal. J. Water Pollut. Control Fed. 43: 1890-1901.

16.

Moore, H. G., R. B. Higgins, and E. G. Fruh, 1969. Surplus phosphorus uptake by microorganisms. Batch tests with diluted activated sludge cultures. Center for Research in Water Resources Report No. 41, Environmental Health Engineering Program, The University of Texas, Austin.

17.

Murphy, J., and J. P. Riley. 1962. A modified single solution method for the determination of phosphate in natural waters. Anal. Chim. Acta 27: 31-36.

18.

Rowan, K. W. 1966. Phosphorus metabolism in plants. Intern Rev. Cytol. 19: 201-290.

19.

Shapiro, J. 1967. Induced release and uptake of phosphate by microorganisms. Science 155: 1269-1271.

20.

Shapiro, J., G. V. Levin, and H. G. Zea. 1967. Anoxically induced release of phosphate in wastewater treatment. J. Water Pollut. Control Fed. 39:18 I0-1818.

21.

Skerman, V. B. D. 1959. "A Guide to the Identification of the Genera of Bacteria." The Williams and Wilkins Co., Baltimore.

22.

Thomas, E. A. 1965. Phosphat-Elimination in der Belebtschlammanlage yon M~innedorf und Phosphate-Fixation in See und Kl~rschlarnm. Vierteljahresschr. Naturf. Ges. Ziirich 110: 419-434.

23.

Vacker, D., C. H. Connell, and W. N. Wells. 1967. Phosphate removal through municipal wastewater treatment at San Antonio, Texas. J. Water Pollut. Control Fed. 39: 750-771.

24.

Voelz, H., V. Voelz, and R. O. Ortigoza. 1966. The "polyphosphate overplus" phenomenon in Myxococcus xanthus and its influence on the architecture of the cell. Arch. Mikrobiol. 53: 371-388.

25.

Weinberger, L. W. 1949. Nitrogen metabolism in the activated sludge process. Ph.D. thesis, Massachusetts Institute of Technology.

26.

White, G. A., and C. H. Wang. 1964. The dissimilation of glucose and gluconate by Acetobacter xylinum, 1, 11. Biochem..I. 90: 408-430.

27.

Wood, L. W., and K. E. Chua. 1973. Glucose flux at the sediment-water interface of Toronto Harbour, Lake Ontario, with reference to pollution stress. Canad. J. Microbiol. 19: 413-420.

28.

Yall, 1., W. H. Boughton, R. C. Knudsen, and N. A. Sinclair. 1970. Biological uptake of phosphorus by activated sludge. Appl. Microbiol. 20: 145-150.