Glyphosate degradation by immobilized bacteria: laboratory studies showing feasibility for glyphosate removal from waste water MICHAEL A.
HEITKAMP' AND
WILLIAMJ.
ADAMS~
Environmental Sciences Center, Monsanto Company, 800 North Lindbergh, St. Louis, MO 63167, U.S.A. AND
LAURENCE E. HALLAS
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
Monsanto Agricultural Company, Monsanto Company, 800 North Lindbergh, St. Louis, MO 63167, U.S.A. Received December 3, 1991 Revision received February 19, 1992 Accepted February 21, 1992
HEITKAMP,M. A., ADAMS,W. J., and HALLAS,L. E. 1992. Glyphosate degradation by immobilized bacteria: laboratory studies showing feasibility for glyphosate removal from waste water. Can. J. Microbiol. 38: 921-928. T o evaluate immobilized bacteria technology for the removal of low levels of glyphosate (N-phosphonomethylglycine) from aqueous industrial effluents, microorganisms with glyphosate-degrading activity obtained from a fill and draw enrichment reactor inoculated with activated sludge were first exposed to glyphosate production wastes containing 500-2000 mg glyphosate/L. The microorganisms were then immobilized by adsorption onto a diatomaceous earth biocarrier contained in upflow Plexiglasmcolumns. The columns were aerated, maintained at p H 7.0-8.0, incubated at 25"C, supplemented with NH4N03(50 mg/L), and exposed to glyphosate process wastes pumped upflow through the biocarrier. Glyphosate degradation to aminomethylphosphonic acid was initially >96% for 21 days of operation at flows yielding hydraulic residence times (HRTs) as short as 42 min. Higher flow rate studies showed > 98% removal of 50 mg glyphosate/L from the waste stream could be achieved at a H R T of 23 min. Glyphosate removal of > 99% at a 37-min HRT was achieved under similar conditions with a column inoculated with a pure culture of Pseudomonas sp. strain LBr, a bacterium known to have high glyphosate-degrading activity. After acid shocking (pH 2.8 for 18 h) of a column of immobilized bacteria, glyphosate-degrading activity was regained within 4 days without reinoculation. Although microbial growth and glyphosate degradation were not maintained under low organic nutrient conditions in the laboratory, the low levels of degradable carbon (45-94 mg/L) in the industrial effluent were sufficient t o support prolonged glyphosate-degrading activity. The results demonstrated that immobilized bacteria technology is effective in removing low levels of glyphosate in high-volume liquid waste streams. Key words: glyphosate, degradation, immobilized bacteria technology. HEITKAMP,M. A., ADAMS, W. J., et HALLAS,L. E. 1992. Glyphosate degradation by immobilized bacteria: laboratory studies showing feasibility for glyphosate removal from waste water. Can. J. Microbiol. 38 : 921-928. Cette etude a eu pour but d'evaluer la technologie de bacteries immobiliskes pour l'enlevement de faibles niveaux de glyphosate (N-phosphonomethylglycine) a partir d'effluents aqueux industriels. Les microorganismes possedant une activite de degradation du glyphosate ont ete obtenus a partir d'un reacteur d'enrichissement de type remplir et puiser, inocule avec de la boue activee. 11s ont d'abord Cte exposes a des boues productrices de glyphosate contenant 500 a 2000 mg glyphosate/L. Les microorganismes ont ensuite ete immobilises par adsorption sur un support de terre de diatomees contenue dans des colonnes de PlexiglasB a courant ascendant. Les colonnes ont ete aerees, maintenues entre pH 7,O et 8,0, incubees a 25°C en presence de NH4N03(50 mg/L) et exposees aux eaux usees contenant du glyphosate par pompage ascendant a travers le support biologique. La degradation du glyphosate en acide aminomethylphosphonique fut initialement superieure a 96070, pendant 21 jours d'opkration a des debits permettant des temps de residence hydrauliques (HRTs) aussi courts que 42 min. Les etudes ont montre qu'a des debits superieurs, l'enlevement des 50 mg glyphosate/L dans le courant d'eaux usees a pu etre accompli a plus de 98% avec un HRT de 23 min. Plus de 99% du glyphosate a pu &re retire dans des conditions similaires, avec un HRT de 37 min, en utilisant une colonne inoculee avec une culture pure de la souche LBr de Pseudomonas sp.; cette bacterie est reconnue pour posseder une activite elevee de degradation du glyphosate. L'activite de degradation du glyphosate fut recuperee a nouveau sans reinoculation, moins de 4 jours suivant un traitement de choc a l'acide (pH 2,8 pendant 18 h) d'une colonne de bacteries immobilisees. Les faibles niveaux de carbone degradable (45 a 94 mg/L) dans l'effluent industriel ont ete suffisants pour supporter une activite prolongee de degradation du glyphosate, tandis qu'au laboratoire, la croissance microbienne et la degradation du glyphosate n'ont pu etre maintenues sous des conditions reduites en elements nutritifs organiques. Les resultats ont demontre que la technologie de bacteries immobilisees est efficace pour retirer de faibles niveaux de glyphosate dans des flots a grand volume d'eaux usees. Mots cles : glyphosate, degradation, technologie de bacteries immobilisees. [Traduit par la redaction]
Introduction Glyphosate (N-phosphonomethylglycine) is a broadspectrum postemergent herbicide that Monsanto Company produces at several manufacturing facilities. Glyphosate is
cleaved either directly to aminomethylphosphonic acid (AMPA) (Balthazor and Hallas 1986) or, by cleavage of the phosphate moiety, to sarcosine (Kishore and Jacob 1987). Glyphosate, AMPA, and sarcosine are readily degradable
' ~ u t h o rto whom all correspondence should be addressed. 2 ~ r e s e n address: t ABC Laboratories, Columbia, MO 65205, U.S.A. Printed in Canada / imprime au Canada
922
CAN. J. MICROBIOL. VOL. 38, 1992
TABLE1. Physical characteristics of the immobilized bacteria columns Height of packed bed, cm
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
Internal diameter, cm Packed bed dry weight, g Packed bed wet weight, g Packed bed volume, L Interstitial fluid volume, mL Interstitial air volume, cm3 Aeration rate, cm3/min
8.26 829 1398 1.6 580 70 1500- 1800
8.26 1520 2563 2.9 1050 70 1500-1800
in environmental samples (Balthazor and Hallas 1986; Cook et al. 1978; Kishore and Jacob 1987; Mueller et al. 1981; Rueppel et al. 1977), and several bacteria with glyphosatedegrading activity (GDA).in pure culture have been reported (Jacob et al. 1985, 1988; Kishore and Jacob 1987; Liu et al. 1991; McAuliffe et al. 1990; Pipke et al. 1987; Pipke and Amrhein 1988; Shinabarger et al. 1984; Wackett et al. 1987). Glyphosate production wastes have been successfully managed at manufacturing facilities by using aerobic activated sludge for secondary treatment (Balthazor and Hallas 1986; Murthy et al. 1989); a previous study has reported the enumeration, isolation, and characterization of activated sludge bacteria having GDA (Hallas et al. 1988). However, increased glyphosate production anticipated from higher worldwide demand and stricter environmental standards for glyphosate residues in discharged waters have spurred interest in new tertiary biotreatment technologies for the removal of low levels of glyphosate from high-volume effluents. Immobilized bacteria technology (IBT) uses highly selected chemical degrading bacteria attached to a biocarrier to remove organics from aqueous waste streams. The high chemical removal rates achievable result from the high density of active biomass in the IBT reactor and from the maintenance of optimal physical, chemical, and nutritional conditions for the immobilized bacteria (Heitkamp et al. 1990). Furthermore, the use of IBT is especially attractive for selectively removing specific chemicals present at low concentration in high-volume waste streams. These waste streams are low sludge producers and high levels of active biomass remain immobilized in IBT reactors at hydraulic flow rates not possible in conventional waste treatment systems. These characteristics have generated increased interest in IBT as a tertiary treatment or "polishing" technology for treated waste water, or as a stand-alone technology for removing chemicals from contaminated groundwaters in pump and treat remediation projects. In this laboratory study, IBT was evaluated for removal of glyphosate from a high-volume waste stream that had already undergone extensive biological treatment in a conventional activated-sludge system. The waste stream was very low in organic nutrients and IBT was evaluated as a tertiary polishing technology to remove low levels of glyphosate from the effluent prior to discharge. Glyphosate removal was compared using IBT columns inoculated either with mixed cultures obtained from activated sludge or with a pure culture known to have high GDA. The experiments included high flow rate studies to determine the capacity for glyphosate removal from aqueous wastes. The performance
DAYS YY
**
DAYS
FIG. 1. Glyphosate removal in a 4-L aerobic, activated-sludge bioreactor over a 92-day period. The bioreactor was fed CSW water enriched with technical-grade glyphosate and cycled as a fill and draw reactor on days 27,30,47, and 66. Inorganic nitrogen addition (50 mg NH,N03/L) was begun on day 55 (*). The feedstock was changed to a 1: 1 mixture of CSW water and clarifier overflow enriched with glyphosate for cycles on days 74 and 78 (**), and to 100% clarifier overflow enriched with glyphosate for cycles on days 82, 85, and 90 (***). of IBT was also evaluated using a synthetic waste stream almost completely devoid of organic nutrients, during a simulated system upset (pH shocking) and after repeated fluidization to remove excess biomass. Materials and methods Chemicals and waste streams Analytical-grade glyphosate was synthesized by Monsanto's Agricultural Technology Department in St. Louis, Mo., and was >99% pure as determined by high pressure liquid chromatography (HPLC) analyses (U.S. Environmental Protection Agency 1983). Centrifuge spent wash (CSW) water collected from a glyphosate production process was used as the high strength glyphosate waste stream. This waste stream normally flows into the activated sludge biotreatment facility and contains formic acid, formaldehyde, glycine, iminodiacetic acid, nitrilotriacetic acid, methyliminodiacetic acid, and several phosphonates, including 500-2000 mg glyphosate/L. CSW water was fed to columns during start-up and recovery phases to promote microbial growth throughout the biocarrier bed while maintaining selective pressure for microorganisms having GDA. Water collected from the clarifier overflow of the activated-sludge biotreatment facility was used as the low organic nutrient waste stream for the IBT columns. This clarified waste stream, which normally flows through a settling basin and goes directly to the monitored discharge point, has a moderate chemical oxygen demand (COD) with only trace levels of glyphosate. To evaluate IBT in this study, the clarified waste stream was amended by adding 10-50 mg technical-grade glyphosate/L. Such concentrations of glyphosate could result in clarifier overflow as a result of a severe upset of the activated-sludge secondary biotreatment system. To determine whether immobilized
HEITKAMP ET AL.
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
bacteria could maintain GDA in waste streams where glyphosate was the sole source of carbon and energy synthetic waste stream with extremely low organic nutrients (COD I 15 mg/L) was prepared by adding 50 mg technical-grade glyphosate/L to pristine well water collected from St. Peters, Mo. The pH of all waste streams was adjusted to 6.5-7.0 before it passed through the IBT columns. Waste streams were pumped with FMI model RP-GI50 metering pumps (Fluid Metering Inc., Oyster Bay, N.Y .). The pumps for all IBT columns were calibrated at each change in the pumping rate and were checked weekly. Immobilized bacteria columns A complete description of the construction of IBT columns used in this study has been previously reported (Heitkamp et al. 1990). The columns were packed with biocarrier to bed heights of 30.5 or 55.9 cm. The biocarrier used was Manville R635 diatomaceous earth beads (Manville Co., Denver, Colo.). Before inoculation, the beads were coated with chitosan by soaking overnight in an acidic chitosan solution (pH 4.0) to facilitate microbial attachment and growth. The biocarrier was rinsed with water for 24 h and pH adjusted to 7.0 before addition to the columns and subsequent microbial seeding. The physical characteristics of the laboratoryscale IBT columns are presented in Table 1. Microbial seed A microbial consortium enriched for GDA was developed as a mixed-culture inoculum for IBT columns A and C. An activatedsludge sample was collected from the manufacturing facility's central waste treatment unit. The activated sludge was used to establish a 4-L bioreactor in the laboratory. This bioreactor was continuously mixed, aerated, and maintained at pH 7.0-8.0. To enrich it with bacteria having GDA to be used to inoculate IBT columns (Fig. I), the bioreactor was fed batches of CSW water spiked with technical-grade glyphosate at concentrations ranging from 500 to 2000 mg/L. Glyphosate degradation in the bioreactor was monitored by HPLC. After complete glyphosate removal was achieved, the bioreactor was operated in a fill and draw mode in which solids were allowed to settle for 2 h before 80% of the liquid was decanted. The bioreactor was refilled by adding fresh CSW water, spiked with technical-grade glyphosate, over a 4-h interval, using a Cole-Parmer MasterFlex model 7520 peristaltic pump (Cole Parmer, Chicago, Ill.). This fill and draw cycle was repeated each time complete glyphosate removal was detected. Beginning with the third fill and draw cycle, the medium in the reactor was amended with NH4N03(50 mg/L). Once high GDA was achieved by the microbial consortium in the bioreactor fed CSW, 50% of the biomass was removed on day 74 and used to inoculate the first immobilized cell column, designated column A. Inoculation of column A was achieved by recycling a suspension of the biomass in fresh CSW water (1 100 mg glyphosate/L) through the column (hydraulic residence time of 1.75 h) for 72 h. Pseudomonas sp. strain LBr, a bacterium previously reported to have GDA (Jacob et al. 1988), was used to inoculate a 250-mL Erlenmeyer flask containing 100 mL of L-salts (Leadbetter and Foster 1958) lacking inorganic phosphate but supplemented with 200 mg technical-grade glyphosate/L. The culture was incubated on a rotary shaker (100 rpm) and sampled twice weekly for removal of glyphosate. After complete removal was achieved, the culture was harvested by centrifugation (8000 x g for 10 min) and transferred into duplicate 2-L Erlenmeyer flasks containing 1 L of L-salts supplemented with 200 mg technical-grade glyphosate/L and 0.1 % yeast extract. The cultures were incubated on a rotary shaker until they showed complete removal of glyphosate and high cell densities, then they were used to inoculate an immobilized bacteria column (designated column B) as described above for column A. The microbial biomass remaining in the 4-L fill and draw reactor was acclimated to clarifier overflow in a two-step process. The
DAYS
FIG. 2. Glyphosate degradation by IBT column A. The initial feedstock was CSW water containing 1200 mg glyphosate/L pumped at a rate of 2.0 mL/min. The column underwent a series of changes in chemical loading (designated A-H) described as follows: A, 50 mg NH4N03/L added; B, flow increased to 2.7 mL/min; C, glyphosate concentration dropped to 600 mg/L and flow increased to 5.0 mL/min; D, glyphosate concentration dropped to 400 mg/L and flow increased to 10.0 mL/min; E, flow increased to 15.5 mL/min; F, feed changed to extremely low organic nutrient well water enriched with 50 mg glyphosate/L; G, feed changed to CSW water containing 700-900 mg glyphosate/L and flow was dropped to 3.0 mL/min; H , flow increased to 7.0 mL/min. first step used a feedstock consisting of a 1: 1 mix of CSW water and clarifier overflow (550 mg glyphosate/L). After complete removal of glyphosate was achieved, a second feedstock consisting of 100% clarifier overflow amended with 500 mg technical-grade glyphosate/L was introduced into the bioreactor. When high GDA was again achieved, the microbial biomass was used to inoculate (as for column A) an immobilized bacteria column (designated as column C). Microbiological analyses Populations of microorganisms were measured in the IBT columns by removing several diatomaceous earth beads from a sample port 12.7 cm from the bottom of each column; the beads were placed in a sterile grinding tube containing 10 mL of sterile phosphate buffer (0.05 M, pH 7 . 9 , ground with a Tekmar grinder (Tekmar Company, Cincinnati, Ohio), and serially diluted (1: 10) in sterile phosphate buffer. Samples (0.1 mL) at l o p 2 to l o p 6 dilution were plated in duplicate on L-salts agar containing 200 mg glyphosate/L. Cultures were incubated in the dark at 28°C for 2-7 days. The residues in the original grinding tube were dried overnight and weighed. Populations of bacteria were expressed as total colony-forming units (cfu) per gram dry weight of biocarrier. Scanning electron microscopy (SEM) analyses were performed by placing 8-10 diatomaceous earth beads into a 10-mL glass vial containing cacodylate buffer (0.1 M, pH 7.4). The beads were cross-sectioned with a razor blade and affixed using methylcellulose onto an EMscope (EMscope Laboratories Inc., Kent, England) transfer mount. The affixed samples were then plunged into liquid nitrogen and transferred to a Jeol840 scanning electron microscope (Jeol Application Laboratories, Peabody, Mass.). The surface ice coating on the beads was removed by raising the temperature to - 60°C while examining in the Jeol SEM at 5 kV. Once surface ice was sublimed, the temperature was reduced to - 180°C and the sectioned beads were transferred back to the EMscope SP-2000 and sputter coated with gold palladium. The samples were then examined by SEM on the Jeol 840 scanning electron microscope at 15 kV. Chemical analyses Glyphosate, AMPA, and ~ 0 were ~ measured ~ - by HPLC, using a Waters model 510 HPLC pump and model 480 spectrophotometer (Waters Milford, Mass.), Varian model 8000 autosampler (Varian, Walnut Creek, Calif .), Rheodyne 700 1
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
924
CAN. J. MICROBIOL. VOL. 38, 1992
autoinjection valve (Rheodyne, Cotati, Calif.), Rainin Rabbit peristaltic pump (Rainin Instruments, Woburn, Mass.), and DuPont column oven (DuPont, Wilmington, Del.). Aqueous samples were filtered (0.45 pm Acrodisc, Gelman Sciences Inc., Ann Arbor, Mich.), adjusted to pH 2-3, and injected onto a Brownlee AX 300 ion-exchange column (P.J . Cobert, St. Louis, Mo.), using a mobile phase containing enough trifluoroacetic acid to titrate pH 2.1. The eluting compounds were oxidized with a 1 % NaOCr solution in a 7-cm coil of Teflon tubing (0.8 mm inside diameter, Atlantic Tubing Company, Paterson, N.J.). The resulting P O , ' was reacted with molybdate to form a blue complex, which was quantitated by absorbance change at 600 nm (U.S. Environmental Protection Agency 1983). TOC was determined with a Xertex/Dohrmann DC-180 analyzer (Santa Clara, Calif.) utilizing the boat mode for individual sample insertion. Aqueous samples were collected from column feedstocks and effluents and were stored at 4°C in 4-mL vials containing inert cap liners. A 40-pL aliquot of each sample was loaded into the sample boat and analyzed by the standard method described in the DC 180 operating systems manual (Dohrmann Rosemount Analytical Division 1988). '
Results Glyphosate degradation in an activated sludge bioreactor Figure 1 shows glyphosate degradation in a fill and draw aerobic activated-sludge bioreactor. The microbial consortium did not show GDA during the 1st week of incubation with CSW water containing 1100 mg glyphosate/L. A relatively low level of GDA was observed during the 2nd and 3rd weeks of incubation, resulting in only a 40% total reduction in glyphosate. However, GDA increased sharply beginning on day 24 and more than 800 mg glyphosate/L was degraded within 4 days. Throughout the bioreactor studies, levels of AMPA increased stoichiometrically to observed decreases in glyphosate levels, indicating that AMPA was the primary degradation product of glyphosate. In addition, a pH increase, characteristic of glyphosate degradation, was also observed in the bioreactor. The pH was monitored and routinely maintained at 7 .O-8.5 in the bioreactor . Although high GDA was observed in the bioreactor on days 24-30, the rate of GDA dropped during the third fill and draw cycle; 16 days were required for the degradation of 2100 mg glyphosate/L in CSW water. This slowing of GDA continued in the fourth fill and draw cycle, when only 5% of the glyphosate was degraded during days 47-55. This progressive reduction in GDA suggested that the microorganisms were nitrogen limited. Subsequently, when 50 mg NH4N03/L was added to the bioreactor on day 55, complete degradation of glyphosate occurred within 72 h. Inorganic nitrogen additions were continued in the bioreactor for the remainder of the study. Half of the microbial biomass in the bioreactor was used to inoculate column A on day 74 as described above, and the remaining microorganisms in the fill and draw bioreactor were successfully acclimated from CSW water to clarifier overflow on days 74-92. In all cases, 50 mg NH4N03/L was added to the feed, and high GDA was observed throughout the remainder of the study. Column C was inoculated on day 92 with microorganisms from ,the bioreactor acclimated to clarifier overflow.
IBT column A This column had a biocarrier bed depth of 55.9 cm and was inoculated with microorganisms from the fill and draw activated-sludge bioreactor before its acclimation to GDA
in the clarifier overflow. Glyphosate removal by column A is shown in Fig. 2. No GDA was observed in column A during the first 16 days of exposure to CSW water, which contained 1200 mg glyphosate/L at a pumping rate of 2 mL/min. This pumping rate resulted in a hydraulic residence time (HRT) of 525 min. However, GDA jumped to > 99070 removal over 3 days after 50 mg NH4N03/L was added to the CSW water feed on day 14. This increase in GDA on column A after adding inorganic nitrogen was similar to the increased GDA observed for the fill and draw activated-sludge bioreactor after adding inorganic nitrogen (Fig. I). The pumping rate was increased to 2.7 mL/min (HRT 389 min) on day 16, and GDA increased to > 99% removal over the following 4-5 days and remained at > 99%. The glyphosate concentration in the CSW water feedstock was dropped to 600 mg/L and the pumping rate was increased to 5.0 mL/min (HRT 210 min) on day 27. GDA temporarily dropped to 77070 removal in response to this change but returned to >99% removal within 72 h. Similarly, on day 36 the glyphosate concentration in the CSW water feedstock was lowered to 400 mg/L and the pumping rate was increased to 10 mL/min (HRT 105 min). In this case, no breakthrough of undegraded glyphosate was observed. On day 41, the pumping rate for column A was increased to 15.5 mL/rnin (HRT 67.7 min) and the glyphosate concentration in CSW was held at 400 mg/L. Glyphosate breakthrough of about 43% was observed after 24 h, but GDA on column A returned to >99070 within 4 days. In all cases, microbial biomass was clearly visible on the column and was denser near the bottom. Each time column A received an increase in chemical loading, an increase in biomass was observed on the column after GDA had returned to >99% removal of glyphosate. An experiment was begun on day 54 to determine whether the bacteria on column A could survive and degrade 50 mg glyphosate/L in a low organic nutrient synthetic waste stream at 15.5 mL/min (HRT 67.7 min). This synthetic waste stream had 40-50 mg TOC/L above that accounted for by the glyphosate addition. Although column A continued to degrade > 99% of the glyphosate under these conditions for 5 days ( > 106 HRTs), GDA dropped to below 10% for the next 14 days, and some of the biomass detached from the biocarrier. Although GDA was maintained on column A for > 106 HRTs, these results suggest that immobilized bacteria are not capable of long-term survival on glyphosate as a sole source of carbon and energy. The bacteria on column A did not recover high GDA until the feedstock was switched to CSW water containing 700-900 mg glyphosate/L and the flow rate was decreased to 3 mL/min (HRT 350 min) on day 74. It is noteworthy that GDA on column A then returned to >99070 removal of glyphosate within 48 h and high GDA was regained within 2 days after the flow rate was increased to 7 mL/min (HRT 150 min) on day 85.
IBT column B This column had a biocarrier bed depth of 55.9 cm and was inoculated with Pseudomonas sp. strain LBr (Jacob et al. 1988). Glyphosate removal by column B is presented in Fig. 3. Low levels of GDA (0-37%) were established during the first 3-4 days of exposure to CSW water containing 600 mg glyphosate/L at a pumping rate of 2 mL/ min (HRT 525 min). On day 3 the flow was increased to
HEITKAMP
925
ET AL.
70
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
DAYS
FIG. 3. Glyphosate degradation by IBT column B after inoculation with Pseudornonas sp. strain LBr. The initial feedstock was CSW water containing 600 mg glyphosate/L and 50 mg NH4N03/L pumped at a rate of 2.0 mL/min. The column underwent a series of changes in chemical loading (designated A-F) described as follows: A, flow increased to 3.0 mL/min; B, glyphosate concentration dropped to 400 mg/L; C, flow increased to 5.0 mL/min; D, flow increased to 10.0 mL/min; E, pH of feedstock was dropped to 2.8 for 18 h; F, feed changed to clarifier overfow enriched with 50 mg glyphosate/L and flow was increased to 15.5 mL/min.
3.0 mL/min (HRT 350 min), and on day 5 the concentration of glyphosate in CSW water feed was dropped to 400 mg/L. The flow rate was increased to 5.0 mL/min (HRT 210 min) on day 8. A GDA of > 99% was established on column B by day 10. Doubling the flow rate (10 mL/ min, HRT 105 min) on day 17 resulted in a 40% breakthrough of glyphosate on days 20-22, but GDA returned to > 90% removal by day 24 and >99% removal by day 28. A pH shock of column B was inadvertently conducted on day 34 when CSW water feedstock was not titrated to neutral pH. This resulted in a pH drop on column B to 2.8 for about 18 h. The feedstock was then titrated to pH 6.5, and GDA was monitored in the column. Total GDA ranged from 30 to 60% removal of glyphosate for 4 days after the pH upset, but it returned to > 99% by day 45, 10 days after the pH upset. Column B was transferred from St. Louis, Mo., to Monsanto's glyphosate production plant on day 55. The biocarrier bed depth was dropped from 55.9 cm to 30.5 cm upon reassembly and the pumping rate was increased to 15.5 mL/min. This decrease in biocarrier bed volume and higher pumping rate resulted in a HRT of 37.4 min. The removal of glyphosate dropped to 45%, but it returned to >99% within 2 days. Throughout all experiments with column B, the increase in GDA observed in response to increases in glyphosate loading was accompanied by an increase in visible biomass on the column.
IBT column C This column had a biocarrier depth of 55.9 cm and was inoculated with microorganisms from the fill and draw activated-sludge bioreactor after acclimation to high GDA in clarifier overflow. Glyphosate removal by column C is presented in Fig. 4. Initial feed for column C was CSW water containing 400 mg glyphosate/L at a pumping rate of 3 mL/min (HRT 350 min). In contrast to columns A and B, column C showed >99% GDA immediately. To rapidly increase the bacterial biomass on the column, the concentration of glyphosate in CSW was increased on day 9 to 1400 mg/L. The GDA of column C dropped to 85% removal on day 14 but returned to >99% removal by
DAYS
FIG. 4. Glyphosate degradation by IBT column C. The initial feedstock was CSW water containing 400 mg glyphosate/L and 50 mg NH4N03/L pumped at a rate of 3.0 mL/min. The column underwent a series of changes in chemical loading (designated A-C) described as follows: A, glyphosate concentration increased to 1400 mg/L; B, feed changed to clarifier overflow enriched with 50 mg glyphosate/L and 25 mg NH4N03/L, and flow was increased to 25 mL/min; C, column C was transferred to Monsanto's glyphosate production plant.
day 16, when a significant increase in bacterial biomass was observed on column C. Flow studies were initiated on column C on day 27 by increasing the pumping rate to 25 mL/min (HRT 42 min). The feedstock for the high flow rate studies consisted of 50 mg glyphosate/L and 25 mg NH4N03/L added to clarifier overflow. The GDA on column C averaged > 96% removal during 21 days of incubation in the laboratory. However, as a result of the high volumes of clarifier overflow needed for feed, column C was transferred on day 48 to Monsanto's glyphosate production plant. Column C was set up with a 55.9-cm biocarrier bed depth, and > 99% GDA was established within 4 days of transfer at a pumping rate of 25 mL/min (HRT 42 min) using a feedstock consisting of clarifier overflow that contained 50 mg glyphosate/L and 50 mg NH4Cl/L.
Microbial colonization of biocarrier Total microbial populations immobilized on the biocarrier in each IBT column were determined by plating studies at the conclusion of the cell-growth phases. This time point corresponded to the appearance of dense biomass on the columns resulting from bacterial growth on CSW water containing 500-2000 mg glyphosate/L. Microbial populations presented as total colony forming units per gram dry weight of biocarrier were as follows: column A, 9.5 x lo7; column B, 3.3 x lo7; column C, 1.1 x lo8. Although the bacterial strains were not identified in this study, plates from columns A and C routinely showed six to eight different colony types. Plates from column B showed predominantly one colony type, with two others occurring a low levels. Light microscopy showed the predominant colony type on column B was a short gram-negative rod, which presumably was Pseudomonas sp. strain LBr (Jacob et al. 1988). Microbial attachment and penetration into the biocarrier were determined by SEM analyses for column C on day 20. The SEM analyses of the outer 20% of a cross-sectioned R635 diatomaceous earth bead (Fig. 5A) showed significant, but not confluent, bacterial colonization. More confluent microbial growth probably occurred later in the study, when column C was operated at maximum chemical loading. The
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
926
CAN. J. MICROBIOL. VOL. 38, 1992
FIG. 5. Scanning electron micrographs taken near the surface of a cross-sectioned Manville R635 diatomaceous earth bead (A) showing microbial colonization occurring in the outer 20% of the bead and (B) taken near the center showing bacteria that had penetrated deep into the bead matrix.
SEM analyses also showed that some bacteria had penetrated deep into the center of the diatomaceous earth bead (Fig. 5B). TOC analyses were conducted on column influents and effluents at several time points throughout the study. Unspiked clarifier overflow had TOC levels that ranged from 230 to 280 mg/L. TOC removal on the columns receiving clarifier overflow spiked with technical-grade glyphosate ranged from 18 to 37%. Since microbial growth and GDA were not maintained under conditions of extremely low organic nutrients, as described earlier (Fig. 2 for column A), GDA by the immobilized cells is most likely a co-metabolic phenomenon that requires the presence of a sufficient level of degradable TOC for maintaining viable and active cells. Although the minimum TOC requirements were not determined in this study, the results indicate that levels of degradable TOC in the clarifier overflow and the polishing pond are sufficient to maintain GDA activity on immobilized cell columns. High-flow studies on ZBT column A The biocarrier bed depth of column A was reduced from 55.9 to 30.5 cm by removing 25.4 cm of the R635 biocarrier from the top of the column. The 30.5-cm immobilized cell column had a smaller fluid volume (Table I), which resulted in a shorter HRT than the 55.9-cm columns at comparable pumping rates. This shortened column was used to monitor GDA in a series of high flow rate studies presented in Fig. 6. Initially, feedstock for these experiments consisted of diluted CSW water (100 mg glyphosate/L) spiked with 400 mg technical-grade glyphosate/L. GDA in column A remained > 98% during the first 19 days of the experiment as the flow rate was increased from 3 (HRT 193 min) up to 20 mL/min (HRT 29 min). On day 20 the pumping rate was increased to 25 mL/min (HRT 23.2 min) and the CSW water concentration was lowered to 50 mg glyphosate/L; technical-grade glyphosate was no longer added to the feedstock. GDA remained >98% for the next 5 days, and on day 26, the flow rate was increased to 30 mL/min (HRT 19.3 min). GDA dropped steadily to 77070 removal over the next 6 days but then recovered to 82% during the follow-
10 0
I
0
,
10
I
,
20
,
I
30
I
40
,
50
I
,
60
70
DAYS
FIG. 6. Glyphosate degradation by IBT column A during high flow and fluidization studies. The initial feedstock was CSW water containing 500 mg glyphosate/L and 50 mg NH,NO,/L pumped at a rate of 3.0 mL/min. The column underwent a series of changes in chemical loading (designated A-G) described as follows: A, flow was increased to 5.0 mL/min; B, flow was increased to 10 mL/min; C, flow was increased to 20 mL/min; D, concentration of glyphosate was dropped to 50 mg/L, 100 mg K,HPO,/L was added, and flow was increased to 25.0 mL/min; E, flow was increased to 30 mL/min; F, biocarrier was completely fluidized for 5 min to remove excess biomass; G, biocarrier was completely fluidized for 5 min to remove excess biomass.
ing 3 days. These results indicate that the optimal performance of column A occurred at a HRT between 23.2 and 19.3 min. Throughout the remainder of this experiment, the pumping rate was maintained at 30 mL/min (HRT 19.3 min) to provide constant glyphosate breakthrough needed to examine the effects of fluidization on column performance. Since excessive biomass was observed on column A on day 36, the column was purged (fluidized) with a fivefold increase in the aeration rate for 2 min. Although the R635 biocarrier was not fluidized enough for complete mixing of the biocarrier to occur on the column, a large portion of the biomass was blown off the biocarrier and pumped out of the column. This fluidization resulted in a drop in GDA to 55070 removal on day 39. However, GDA recovered to 77% removal on day 44 and ranged from 71 to 83% removal
HEITKAMP ET AL.
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
during days 45-57. Although excessive biomass had not accumulated on the column by day 57, the column was fluidized again to observe the effects of another fluidization cycle on the loss and recovery of GDA. The second fluidization resulted in a smaller loss of GDA (65% after 3 days), a faster recovery (1 day), and a higher level of GDA after recovery (78-86%). Fluidization probably benefits GDA by removing biomass (increasing fluid volume) and enriching for bacteria that attach more securely or penetrate more deeply into the biocarrier. These results indicate that an immobilized cell operational strategy should include a periodic fluidization cycle to benefit GDA. Discussion This study demonstrates that IBT can achieve >98% removal of low levels of glyphosate ( 1 5 0 mg/L) from aqueous waste streams at HRTs as short as 23 min. The high glyphosate removal rates that IBT achieves in this study probably result from combining the high densities of bacterial cells on the columns, using highly selected microbial cultures having high GDA, and optimizing physical and nutritional conditions on the IBT columns. In addition, the degradation of glyphosate to AMPA, its primary metabolite, only requires a one-step, two-carbon cleavage, which can be expected to occur rapidly. In previous laboratory studies, IBT has been reported to degrade > 99% of pnitrophenol (PNP) in a synthetic waste stream containing 1200-1800 mg PNP/L at an HRT of 48 min (Heitkamp et al. 1990). However, IBT's capacity to remove low levels ( s50 mg/L) of PNP from high-volume waste streams was not determined in that previous study. The present study shows that IBT is promising as a tertiary treatment technology for removing low levels of glyphosate from high-volume aqueous wastes. The ability to withstand and recover quickly from pH and osmotic shocks may be a significant advantage for IBT over some conventional technologies. Scanning electron microscopy in this study showed that most of the immobilized bacteria occurred on the outer portions of the biocarrier, but a significant number of bacterial cells were detected throughout the interior of the diatomaceous earth beads. This penetration of bacteria into the R635 diatomaceous earth biocarrier may serve to maintain a stable bacterial inoculum in the reactor during system upsets. This hypothesis would suggest that bacteria inside .the beads may exist in a microenvironment that is protected from severe chemical or pH shocks, which may be lethal to unprotected bacteria on the bead's surface. Furthermore, the ability of these protected bacteria to serve as inoculum could explain the rapid recovery of GDA by the IBT columns in this study when favorable conditions were restored. The minimal loss of GDA after repeated fluidization of the IBT column in this study suggests that full-scale IBT reactors should be routinely fluidized to remove excess biomass. Fluidization serves to remove interstitial biomass from the biocarrier bed and maximize interstitial fluid volume. This larger fluid volume results in a longer hydraulic residence' which benefit GDA by increasing the exposure time of waste water to immobilized bacteria in the reactor. Routine fluidization may also benefit IBT performance by selecting bacteria that attach more securely to the porous biocarrier, resulting in high GDA at high flow rates with minimal washout of immobilized bacteria and mini-
927
ma1 channeling of liquid wastes through the reactor. The immobilized bacteria in this study were not able to maintain high GDA when fed glyphosate as a sole source of carbon and energy. This suggests that GDA is linked to some minimal level of additional carbon loading needed to maintain the viability and activity of the immobilized cells. However, this should not create a problem for using IBT for tertiary biotreatment of effluents, since this study did show that clarifier overflow exiting an activaged-sludge secondary-treatment facility contained sufficient low levels of degradable organic carbon to support viability and activity of immobilized microorganisms having GDA. Furthermore, since the immobilized bacteria were able to maintain >99% removal of glyphosate from an extremely low organic carbon waste stream for a total of 106 HRTs occurring over a 5-day period, small fluctuations or temporary decreases in levels of degradable organic carbon in effluent would not be expected to negatively impact IBT performance for removing low levels of glyphosate. The minimum level of secondary carbon needed to support GDA was not determined in this study but warrants further investigation. The immediate increase in GDA in the fill and draw reactor and the IBT columns after adding NH4N03 suggests that nitrogen supplementation should be used for IBT treatment of glyphosate in aqueous wastes. However, inorganic nitrogen should only be added at the minimum levels needed to support GDA, since excessive nitrogen in the bioreactors could select for high populations of nitrifying bacteria, which would require excessive amounts of oxygen. Since high rates of GDA were achieved by IBT in this laboratory-scale study, pilot-scale studies were conducted to confirm the applicability of IBT for glyphosate removal from high-volume waste effluents. The pilot-scale study was designed to provide the GDA rate data needed to engineer and price a full-scale system and confirm the optimum operating conditions (pH, aeration, temperature, nitrogen supplementation, fluidization frequency, etc.) for IBT. In addition, the pilot-scale study was designed to determine the response of immobilized bacteria to surge loading of glyphosate to simulate a worst-case biosystem upset as well as investigate the recovery of GDA after long periods of dormancy in which the immobilized bacteria are exposed to minimal concentrations of glyphosate. The results of pilot-scale testing of IBT for removal of glyphosate from high-volume liquid wastes are presented in a separate paper (Hallas et al. 1992). Acknowledgements The excellent technical assistance of Martina BianchiniAkbeg, Stephanie Lannert, Will Renaudette, Tom Reuter, and David Schlictman of the Environmental Sciences Center, and Eileen Hahn and Harold Crouch of the Monsanto Agricultural Company contributed greatly to the successful completion of this project. The authors thank Carol Pellegrine for performing scanning electron microscopy analyses. Balthazor , T.M., and Hallas, L.E. 1986. Glyphosate-degrading microorganisms from industrial activated sludge. Appl. Environ. Microbial. 51: 432-434. cook, A.M., ~ ~ c.G., and ~ Alexander, ~ M.h 1978. ~ Phos~honateutilization by bacteria. J. Bacteriol. 133: 85-90. ~ o h r m a n nRosemount ~ n a i ~ t i c Division. al 1988. Total organic carbon analyzer. 3rd ed. Dohrmann, Santa Clara, Calif.
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
928
CAN. J. MICROBIOL. VOL. 38, 1992
Hallas, L.E., Hahn, E.M., and Korndorfer, C. 1988. Characterization of microbial traits associated with glyphosate biodegradation in industrial activated sludge. J . Ind. Microbiol. 3: 377-385. Hallas, L.E., Adams, W.J., and Heitkamp, M.A. 1992. Glyphosate degradation by immobilized bacteria: field studies showing removal of glyphosate from industrial wastewater. Appl. Environ. Microbiol. 58: 1215-1219. Heitkamp, M.A., Camel, V., Reuter, T.J., and Adams, W.J. 1990. Biodegradation of p-nitrophenol in an aqueous waste stream by immobilized bacteria. Appl. Environ. Microbiol. 56: 2967-2973. Jacob, G.S., Schaefer, J., Stejskal, E.O., and McKay, R.A. 1985. Solid-state NMR determination of glyphosate metabolism in a Pseudomonas sp. J. Biol. Chem. 260: 5899-5905. Jacob, G.S., Garbow, J.R., Hallas, L.E., et al. 1988. Metabolism of glyphosate in Pseudomonas sp. strain LBr. Appl. Environ. Microbiol. 54: 2953-2958. Kishore, G.M., and Jacob, G.S. 1987. Degradation of glyphosate by Pseudomonas sp. PG2982 via a sarcosine intermediate. J. Biol. Chem. 262: 12 164 - 12 168. Leadbetter, E.R., and Foster, J.W. 1958. Studies on more methaneutilizing bacteria. Arch. Mikrobiol. 30: 91-1 18. Liu, C.M., McLean, P.A., Sookdeo, C.C., and Cannon, F.C. 1991. Degradation of the herbicide glyphosate by members of the family Rhizobiaceae. Appl. Environ. Microbiol. 57: 1799-1804. McAuliffe, K.S., Hallas, L.E., and Kulpa, C.F. 1990. Glyphosate degradation by Agrobacterium radiobacter isolated from activated sludge. J . Ind. Microbiol. 6: 219-221. Mueller, M.M., Rosenberg, C., Siltanen, H., and Wartiovaara, T. 1981. Fate of glyphosate and its influence on nitrogen cycling in two Finnish agricultural soils. Bull. Environ. Contam. Toxicol. 27: 724-730.
Murthy, D.V.S., Irvine, R.L., and Hallas, L'.E. 1989. Principles of organism selection for the degradation of glyphosate in a sequencing batch reactor. Proceedings of the 43rd Annual Purdue Industrial Waste Conference. Lewis Publishers, Chelsea, Mich. pp. 267-275. Pipke, R., and Amrhein, N. 1988. Isolation and characterization of a mutant of Arthrobacter sp. strain GLP-1 which utilizes the herbicide glyphosate as its sole source of phosphorus and nitrogen. Appl. Environ. Microbiol. 54: 2868-2870. Pipke, R., Schulz, A., and Amrhein, N. 1987a. Uptake of glyphosate by Arthrobacter sp. Appl. Environ. Microbiol. 53: 974-97 8. Pipke, R., Amrhein, N., Jacob, G.S., et al. 19876. Metabolism of glyphosate in Arthrobacter sp. GLP- 1. Eur . J. Biochem. 165: 267-273. Rueppel, M.L., Brightwell, B.B., Schafer, J., and Marvel, J.T. 1977. Metabolism and degradation of glyphosate in soil and water. J. Agric. Food chem. 25: 517-522. Shinabarger, D.L., Schmitt, E.K., Braymer, H.D., and Larson, A.D. 1984. Phosphonate utilization by the glyphosate-degrading Pseudomonas sp. strain PG2982. Appl. Environ. Microbiol. 48: 1049-1050. U.S. Environmental Protection Agency. 1983. Methods for nonconvenient pesticide analysis of industrial and municipal wastewater: method 127-determination of glyphosate in wastewater. U.S. EPA Publication No. 440/1-83/079-C, U.S. Environmental Protection Agency, Washington, D.C. pp. 1- 10. Wackett, L.P., Shames, S.L., Venditti, C.P., and Walsh, C.T. 1987. Bacterial carbon-phosphorus lyase: products, rates and regulation of p hosp honic and p hosp hinic acid metabolism. J. Bacteriol. 169: 7 10-717.
HEITKAMP' AND
WILLIAMJ.
ADAMS~
Environmental Sciences Center, Monsanto Company, 800 North Lindbergh, St. Louis, MO 63167, U.S.A. AND
LAURENCE E. HALLAS
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
Monsanto Agricultural Company, Monsanto Company, 800 North Lindbergh, St. Louis, MO 63167, U.S.A. Received December 3, 1991 Revision received February 19, 1992 Accepted February 21, 1992
HEITKAMP,M. A., ADAMS,W. J., and HALLAS,L. E. 1992. Glyphosate degradation by immobilized bacteria: laboratory studies showing feasibility for glyphosate removal from waste water. Can. J. Microbiol. 38: 921-928. T o evaluate immobilized bacteria technology for the removal of low levels of glyphosate (N-phosphonomethylglycine) from aqueous industrial effluents, microorganisms with glyphosate-degrading activity obtained from a fill and draw enrichment reactor inoculated with activated sludge were first exposed to glyphosate production wastes containing 500-2000 mg glyphosate/L. The microorganisms were then immobilized by adsorption onto a diatomaceous earth biocarrier contained in upflow Plexiglasmcolumns. The columns were aerated, maintained at p H 7.0-8.0, incubated at 25"C, supplemented with NH4N03(50 mg/L), and exposed to glyphosate process wastes pumped upflow through the biocarrier. Glyphosate degradation to aminomethylphosphonic acid was initially >96% for 21 days of operation at flows yielding hydraulic residence times (HRTs) as short as 42 min. Higher flow rate studies showed > 98% removal of 50 mg glyphosate/L from the waste stream could be achieved at a H R T of 23 min. Glyphosate removal of > 99% at a 37-min HRT was achieved under similar conditions with a column inoculated with a pure culture of Pseudomonas sp. strain LBr, a bacterium known to have high glyphosate-degrading activity. After acid shocking (pH 2.8 for 18 h) of a column of immobilized bacteria, glyphosate-degrading activity was regained within 4 days without reinoculation. Although microbial growth and glyphosate degradation were not maintained under low organic nutrient conditions in the laboratory, the low levels of degradable carbon (45-94 mg/L) in the industrial effluent were sufficient t o support prolonged glyphosate-degrading activity. The results demonstrated that immobilized bacteria technology is effective in removing low levels of glyphosate in high-volume liquid waste streams. Key words: glyphosate, degradation, immobilized bacteria technology. HEITKAMP,M. A., ADAMS, W. J., et HALLAS,L. E. 1992. Glyphosate degradation by immobilized bacteria: laboratory studies showing feasibility for glyphosate removal from waste water. Can. J. Microbiol. 38 : 921-928. Cette etude a eu pour but d'evaluer la technologie de bacteries immobiliskes pour l'enlevement de faibles niveaux de glyphosate (N-phosphonomethylglycine) a partir d'effluents aqueux industriels. Les microorganismes possedant une activite de degradation du glyphosate ont ete obtenus a partir d'un reacteur d'enrichissement de type remplir et puiser, inocule avec de la boue activee. 11s ont d'abord Cte exposes a des boues productrices de glyphosate contenant 500 a 2000 mg glyphosate/L. Les microorganismes ont ensuite ete immobilises par adsorption sur un support de terre de diatomees contenue dans des colonnes de PlexiglasB a courant ascendant. Les colonnes ont ete aerees, maintenues entre pH 7,O et 8,0, incubees a 25°C en presence de NH4N03(50 mg/L) et exposees aux eaux usees contenant du glyphosate par pompage ascendant a travers le support biologique. La degradation du glyphosate en acide aminomethylphosphonique fut initialement superieure a 96070, pendant 21 jours d'opkration a des debits permettant des temps de residence hydrauliques (HRTs) aussi courts que 42 min. Les etudes ont montre qu'a des debits superieurs, l'enlevement des 50 mg glyphosate/L dans le courant d'eaux usees a pu etre accompli a plus de 98% avec un HRT de 23 min. Plus de 99% du glyphosate a pu &re retire dans des conditions similaires, avec un HRT de 37 min, en utilisant une colonne inoculee avec une culture pure de la souche LBr de Pseudomonas sp.; cette bacterie est reconnue pour posseder une activite elevee de degradation du glyphosate. L'activite de degradation du glyphosate fut recuperee a nouveau sans reinoculation, moins de 4 jours suivant un traitement de choc a l'acide (pH 2,8 pendant 18 h) d'une colonne de bacteries immobilisees. Les faibles niveaux de carbone degradable (45 a 94 mg/L) dans l'effluent industriel ont ete suffisants pour supporter une activite prolongee de degradation du glyphosate, tandis qu'au laboratoire, la croissance microbienne et la degradation du glyphosate n'ont pu etre maintenues sous des conditions reduites en elements nutritifs organiques. Les resultats ont demontre que la technologie de bacteries immobilisees est efficace pour retirer de faibles niveaux de glyphosate dans des flots a grand volume d'eaux usees. Mots cles : glyphosate, degradation, technologie de bacteries immobilisees. [Traduit par la redaction]
Introduction Glyphosate (N-phosphonomethylglycine) is a broadspectrum postemergent herbicide that Monsanto Company produces at several manufacturing facilities. Glyphosate is
cleaved either directly to aminomethylphosphonic acid (AMPA) (Balthazor and Hallas 1986) or, by cleavage of the phosphate moiety, to sarcosine (Kishore and Jacob 1987). Glyphosate, AMPA, and sarcosine are readily degradable
' ~ u t h o rto whom all correspondence should be addressed. 2 ~ r e s e n address: t ABC Laboratories, Columbia, MO 65205, U.S.A. Printed in Canada / imprime au Canada
922
CAN. J. MICROBIOL. VOL. 38, 1992
TABLE1. Physical characteristics of the immobilized bacteria columns Height of packed bed, cm
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
Internal diameter, cm Packed bed dry weight, g Packed bed wet weight, g Packed bed volume, L Interstitial fluid volume, mL Interstitial air volume, cm3 Aeration rate, cm3/min
8.26 829 1398 1.6 580 70 1500- 1800
8.26 1520 2563 2.9 1050 70 1500-1800
in environmental samples (Balthazor and Hallas 1986; Cook et al. 1978; Kishore and Jacob 1987; Mueller et al. 1981; Rueppel et al. 1977), and several bacteria with glyphosatedegrading activity (GDA).in pure culture have been reported (Jacob et al. 1985, 1988; Kishore and Jacob 1987; Liu et al. 1991; McAuliffe et al. 1990; Pipke et al. 1987; Pipke and Amrhein 1988; Shinabarger et al. 1984; Wackett et al. 1987). Glyphosate production wastes have been successfully managed at manufacturing facilities by using aerobic activated sludge for secondary treatment (Balthazor and Hallas 1986; Murthy et al. 1989); a previous study has reported the enumeration, isolation, and characterization of activated sludge bacteria having GDA (Hallas et al. 1988). However, increased glyphosate production anticipated from higher worldwide demand and stricter environmental standards for glyphosate residues in discharged waters have spurred interest in new tertiary biotreatment technologies for the removal of low levels of glyphosate from high-volume effluents. Immobilized bacteria technology (IBT) uses highly selected chemical degrading bacteria attached to a biocarrier to remove organics from aqueous waste streams. The high chemical removal rates achievable result from the high density of active biomass in the IBT reactor and from the maintenance of optimal physical, chemical, and nutritional conditions for the immobilized bacteria (Heitkamp et al. 1990). Furthermore, the use of IBT is especially attractive for selectively removing specific chemicals present at low concentration in high-volume waste streams. These waste streams are low sludge producers and high levels of active biomass remain immobilized in IBT reactors at hydraulic flow rates not possible in conventional waste treatment systems. These characteristics have generated increased interest in IBT as a tertiary treatment or "polishing" technology for treated waste water, or as a stand-alone technology for removing chemicals from contaminated groundwaters in pump and treat remediation projects. In this laboratory study, IBT was evaluated for removal of glyphosate from a high-volume waste stream that had already undergone extensive biological treatment in a conventional activated-sludge system. The waste stream was very low in organic nutrients and IBT was evaluated as a tertiary polishing technology to remove low levels of glyphosate from the effluent prior to discharge. Glyphosate removal was compared using IBT columns inoculated either with mixed cultures obtained from activated sludge or with a pure culture known to have high GDA. The experiments included high flow rate studies to determine the capacity for glyphosate removal from aqueous wastes. The performance
DAYS YY
**
DAYS
FIG. 1. Glyphosate removal in a 4-L aerobic, activated-sludge bioreactor over a 92-day period. The bioreactor was fed CSW water enriched with technical-grade glyphosate and cycled as a fill and draw reactor on days 27,30,47, and 66. Inorganic nitrogen addition (50 mg NH,N03/L) was begun on day 55 (*). The feedstock was changed to a 1: 1 mixture of CSW water and clarifier overflow enriched with glyphosate for cycles on days 74 and 78 (**), and to 100% clarifier overflow enriched with glyphosate for cycles on days 82, 85, and 90 (***). of IBT was also evaluated using a synthetic waste stream almost completely devoid of organic nutrients, during a simulated system upset (pH shocking) and after repeated fluidization to remove excess biomass. Materials and methods Chemicals and waste streams Analytical-grade glyphosate was synthesized by Monsanto's Agricultural Technology Department in St. Louis, Mo., and was >99% pure as determined by high pressure liquid chromatography (HPLC) analyses (U.S. Environmental Protection Agency 1983). Centrifuge spent wash (CSW) water collected from a glyphosate production process was used as the high strength glyphosate waste stream. This waste stream normally flows into the activated sludge biotreatment facility and contains formic acid, formaldehyde, glycine, iminodiacetic acid, nitrilotriacetic acid, methyliminodiacetic acid, and several phosphonates, including 500-2000 mg glyphosate/L. CSW water was fed to columns during start-up and recovery phases to promote microbial growth throughout the biocarrier bed while maintaining selective pressure for microorganisms having GDA. Water collected from the clarifier overflow of the activated-sludge biotreatment facility was used as the low organic nutrient waste stream for the IBT columns. This clarified waste stream, which normally flows through a settling basin and goes directly to the monitored discharge point, has a moderate chemical oxygen demand (COD) with only trace levels of glyphosate. To evaluate IBT in this study, the clarified waste stream was amended by adding 10-50 mg technical-grade glyphosate/L. Such concentrations of glyphosate could result in clarifier overflow as a result of a severe upset of the activated-sludge secondary biotreatment system. To determine whether immobilized
HEITKAMP ET AL.
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
bacteria could maintain GDA in waste streams where glyphosate was the sole source of carbon and energy synthetic waste stream with extremely low organic nutrients (COD I 15 mg/L) was prepared by adding 50 mg technical-grade glyphosate/L to pristine well water collected from St. Peters, Mo. The pH of all waste streams was adjusted to 6.5-7.0 before it passed through the IBT columns. Waste streams were pumped with FMI model RP-GI50 metering pumps (Fluid Metering Inc., Oyster Bay, N.Y .). The pumps for all IBT columns were calibrated at each change in the pumping rate and were checked weekly. Immobilized bacteria columns A complete description of the construction of IBT columns used in this study has been previously reported (Heitkamp et al. 1990). The columns were packed with biocarrier to bed heights of 30.5 or 55.9 cm. The biocarrier used was Manville R635 diatomaceous earth beads (Manville Co., Denver, Colo.). Before inoculation, the beads were coated with chitosan by soaking overnight in an acidic chitosan solution (pH 4.0) to facilitate microbial attachment and growth. The biocarrier was rinsed with water for 24 h and pH adjusted to 7.0 before addition to the columns and subsequent microbial seeding. The physical characteristics of the laboratoryscale IBT columns are presented in Table 1. Microbial seed A microbial consortium enriched for GDA was developed as a mixed-culture inoculum for IBT columns A and C. An activatedsludge sample was collected from the manufacturing facility's central waste treatment unit. The activated sludge was used to establish a 4-L bioreactor in the laboratory. This bioreactor was continuously mixed, aerated, and maintained at pH 7.0-8.0. To enrich it with bacteria having GDA to be used to inoculate IBT columns (Fig. I), the bioreactor was fed batches of CSW water spiked with technical-grade glyphosate at concentrations ranging from 500 to 2000 mg/L. Glyphosate degradation in the bioreactor was monitored by HPLC. After complete glyphosate removal was achieved, the bioreactor was operated in a fill and draw mode in which solids were allowed to settle for 2 h before 80% of the liquid was decanted. The bioreactor was refilled by adding fresh CSW water, spiked with technical-grade glyphosate, over a 4-h interval, using a Cole-Parmer MasterFlex model 7520 peristaltic pump (Cole Parmer, Chicago, Ill.). This fill and draw cycle was repeated each time complete glyphosate removal was detected. Beginning with the third fill and draw cycle, the medium in the reactor was amended with NH4N03(50 mg/L). Once high GDA was achieved by the microbial consortium in the bioreactor fed CSW, 50% of the biomass was removed on day 74 and used to inoculate the first immobilized cell column, designated column A. Inoculation of column A was achieved by recycling a suspension of the biomass in fresh CSW water (1 100 mg glyphosate/L) through the column (hydraulic residence time of 1.75 h) for 72 h. Pseudomonas sp. strain LBr, a bacterium previously reported to have GDA (Jacob et al. 1988), was used to inoculate a 250-mL Erlenmeyer flask containing 100 mL of L-salts (Leadbetter and Foster 1958) lacking inorganic phosphate but supplemented with 200 mg technical-grade glyphosate/L. The culture was incubated on a rotary shaker (100 rpm) and sampled twice weekly for removal of glyphosate. After complete removal was achieved, the culture was harvested by centrifugation (8000 x g for 10 min) and transferred into duplicate 2-L Erlenmeyer flasks containing 1 L of L-salts supplemented with 200 mg technical-grade glyphosate/L and 0.1 % yeast extract. The cultures were incubated on a rotary shaker until they showed complete removal of glyphosate and high cell densities, then they were used to inoculate an immobilized bacteria column (designated column B) as described above for column A. The microbial biomass remaining in the 4-L fill and draw reactor was acclimated to clarifier overflow in a two-step process. The
DAYS
FIG. 2. Glyphosate degradation by IBT column A. The initial feedstock was CSW water containing 1200 mg glyphosate/L pumped at a rate of 2.0 mL/min. The column underwent a series of changes in chemical loading (designated A-H) described as follows: A, 50 mg NH4N03/L added; B, flow increased to 2.7 mL/min; C, glyphosate concentration dropped to 600 mg/L and flow increased to 5.0 mL/min; D, glyphosate concentration dropped to 400 mg/L and flow increased to 10.0 mL/min; E, flow increased to 15.5 mL/min; F, feed changed to extremely low organic nutrient well water enriched with 50 mg glyphosate/L; G, feed changed to CSW water containing 700-900 mg glyphosate/L and flow was dropped to 3.0 mL/min; H , flow increased to 7.0 mL/min. first step used a feedstock consisting of a 1: 1 mix of CSW water and clarifier overflow (550 mg glyphosate/L). After complete removal of glyphosate was achieved, a second feedstock consisting of 100% clarifier overflow amended with 500 mg technical-grade glyphosate/L was introduced into the bioreactor. When high GDA was again achieved, the microbial biomass was used to inoculate (as for column A) an immobilized bacteria column (designated as column C). Microbiological analyses Populations of microorganisms were measured in the IBT columns by removing several diatomaceous earth beads from a sample port 12.7 cm from the bottom of each column; the beads were placed in a sterile grinding tube containing 10 mL of sterile phosphate buffer (0.05 M, pH 7 . 9 , ground with a Tekmar grinder (Tekmar Company, Cincinnati, Ohio), and serially diluted (1: 10) in sterile phosphate buffer. Samples (0.1 mL) at l o p 2 to l o p 6 dilution were plated in duplicate on L-salts agar containing 200 mg glyphosate/L. Cultures were incubated in the dark at 28°C for 2-7 days. The residues in the original grinding tube were dried overnight and weighed. Populations of bacteria were expressed as total colony-forming units (cfu) per gram dry weight of biocarrier. Scanning electron microscopy (SEM) analyses were performed by placing 8-10 diatomaceous earth beads into a 10-mL glass vial containing cacodylate buffer (0.1 M, pH 7.4). The beads were cross-sectioned with a razor blade and affixed using methylcellulose onto an EMscope (EMscope Laboratories Inc., Kent, England) transfer mount. The affixed samples were then plunged into liquid nitrogen and transferred to a Jeol840 scanning electron microscope (Jeol Application Laboratories, Peabody, Mass.). The surface ice coating on the beads was removed by raising the temperature to - 60°C while examining in the Jeol SEM at 5 kV. Once surface ice was sublimed, the temperature was reduced to - 180°C and the sectioned beads were transferred back to the EMscope SP-2000 and sputter coated with gold palladium. The samples were then examined by SEM on the Jeol 840 scanning electron microscope at 15 kV. Chemical analyses Glyphosate, AMPA, and ~ 0 were ~ measured ~ - by HPLC, using a Waters model 510 HPLC pump and model 480 spectrophotometer (Waters Milford, Mass.), Varian model 8000 autosampler (Varian, Walnut Creek, Calif .), Rheodyne 700 1
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
924
CAN. J. MICROBIOL. VOL. 38, 1992
autoinjection valve (Rheodyne, Cotati, Calif.), Rainin Rabbit peristaltic pump (Rainin Instruments, Woburn, Mass.), and DuPont column oven (DuPont, Wilmington, Del.). Aqueous samples were filtered (0.45 pm Acrodisc, Gelman Sciences Inc., Ann Arbor, Mich.), adjusted to pH 2-3, and injected onto a Brownlee AX 300 ion-exchange column (P.J . Cobert, St. Louis, Mo.), using a mobile phase containing enough trifluoroacetic acid to titrate pH 2.1. The eluting compounds were oxidized with a 1 % NaOCr solution in a 7-cm coil of Teflon tubing (0.8 mm inside diameter, Atlantic Tubing Company, Paterson, N.J.). The resulting P O , ' was reacted with molybdate to form a blue complex, which was quantitated by absorbance change at 600 nm (U.S. Environmental Protection Agency 1983). TOC was determined with a Xertex/Dohrmann DC-180 analyzer (Santa Clara, Calif.) utilizing the boat mode for individual sample insertion. Aqueous samples were collected from column feedstocks and effluents and were stored at 4°C in 4-mL vials containing inert cap liners. A 40-pL aliquot of each sample was loaded into the sample boat and analyzed by the standard method described in the DC 180 operating systems manual (Dohrmann Rosemount Analytical Division 1988). '
Results Glyphosate degradation in an activated sludge bioreactor Figure 1 shows glyphosate degradation in a fill and draw aerobic activated-sludge bioreactor. The microbial consortium did not show GDA during the 1st week of incubation with CSW water containing 1100 mg glyphosate/L. A relatively low level of GDA was observed during the 2nd and 3rd weeks of incubation, resulting in only a 40% total reduction in glyphosate. However, GDA increased sharply beginning on day 24 and more than 800 mg glyphosate/L was degraded within 4 days. Throughout the bioreactor studies, levels of AMPA increased stoichiometrically to observed decreases in glyphosate levels, indicating that AMPA was the primary degradation product of glyphosate. In addition, a pH increase, characteristic of glyphosate degradation, was also observed in the bioreactor. The pH was monitored and routinely maintained at 7 .O-8.5 in the bioreactor . Although high GDA was observed in the bioreactor on days 24-30, the rate of GDA dropped during the third fill and draw cycle; 16 days were required for the degradation of 2100 mg glyphosate/L in CSW water. This slowing of GDA continued in the fourth fill and draw cycle, when only 5% of the glyphosate was degraded during days 47-55. This progressive reduction in GDA suggested that the microorganisms were nitrogen limited. Subsequently, when 50 mg NH4N03/L was added to the bioreactor on day 55, complete degradation of glyphosate occurred within 72 h. Inorganic nitrogen additions were continued in the bioreactor for the remainder of the study. Half of the microbial biomass in the bioreactor was used to inoculate column A on day 74 as described above, and the remaining microorganisms in the fill and draw bioreactor were successfully acclimated from CSW water to clarifier overflow on days 74-92. In all cases, 50 mg NH4N03/L was added to the feed, and high GDA was observed throughout the remainder of the study. Column C was inoculated on day 92 with microorganisms from ,the bioreactor acclimated to clarifier overflow.
IBT column A This column had a biocarrier bed depth of 55.9 cm and was inoculated with microorganisms from the fill and draw activated-sludge bioreactor before its acclimation to GDA
in the clarifier overflow. Glyphosate removal by column A is shown in Fig. 2. No GDA was observed in column A during the first 16 days of exposure to CSW water, which contained 1200 mg glyphosate/L at a pumping rate of 2 mL/min. This pumping rate resulted in a hydraulic residence time (HRT) of 525 min. However, GDA jumped to > 99070 removal over 3 days after 50 mg NH4N03/L was added to the CSW water feed on day 14. This increase in GDA on column A after adding inorganic nitrogen was similar to the increased GDA observed for the fill and draw activated-sludge bioreactor after adding inorganic nitrogen (Fig. I). The pumping rate was increased to 2.7 mL/min (HRT 389 min) on day 16, and GDA increased to > 99% removal over the following 4-5 days and remained at > 99%. The glyphosate concentration in the CSW water feedstock was dropped to 600 mg/L and the pumping rate was increased to 5.0 mL/min (HRT 210 min) on day 27. GDA temporarily dropped to 77070 removal in response to this change but returned to >99% removal within 72 h. Similarly, on day 36 the glyphosate concentration in the CSW water feedstock was lowered to 400 mg/L and the pumping rate was increased to 10 mL/min (HRT 105 min). In this case, no breakthrough of undegraded glyphosate was observed. On day 41, the pumping rate for column A was increased to 15.5 mL/rnin (HRT 67.7 min) and the glyphosate concentration in CSW was held at 400 mg/L. Glyphosate breakthrough of about 43% was observed after 24 h, but GDA on column A returned to >99070 within 4 days. In all cases, microbial biomass was clearly visible on the column and was denser near the bottom. Each time column A received an increase in chemical loading, an increase in biomass was observed on the column after GDA had returned to >99% removal of glyphosate. An experiment was begun on day 54 to determine whether the bacteria on column A could survive and degrade 50 mg glyphosate/L in a low organic nutrient synthetic waste stream at 15.5 mL/min (HRT 67.7 min). This synthetic waste stream had 40-50 mg TOC/L above that accounted for by the glyphosate addition. Although column A continued to degrade > 99% of the glyphosate under these conditions for 5 days ( > 106 HRTs), GDA dropped to below 10% for the next 14 days, and some of the biomass detached from the biocarrier. Although GDA was maintained on column A for > 106 HRTs, these results suggest that immobilized bacteria are not capable of long-term survival on glyphosate as a sole source of carbon and energy. The bacteria on column A did not recover high GDA until the feedstock was switched to CSW water containing 700-900 mg glyphosate/L and the flow rate was decreased to 3 mL/min (HRT 350 min) on day 74. It is noteworthy that GDA on column A then returned to >99070 removal of glyphosate within 48 h and high GDA was regained within 2 days after the flow rate was increased to 7 mL/min (HRT 150 min) on day 85.
IBT column B This column had a biocarrier bed depth of 55.9 cm and was inoculated with Pseudomonas sp. strain LBr (Jacob et al. 1988). Glyphosate removal by column B is presented in Fig. 3. Low levels of GDA (0-37%) were established during the first 3-4 days of exposure to CSW water containing 600 mg glyphosate/L at a pumping rate of 2 mL/ min (HRT 525 min). On day 3 the flow was increased to
HEITKAMP
925
ET AL.
70
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
DAYS
FIG. 3. Glyphosate degradation by IBT column B after inoculation with Pseudornonas sp. strain LBr. The initial feedstock was CSW water containing 600 mg glyphosate/L and 50 mg NH4N03/L pumped at a rate of 2.0 mL/min. The column underwent a series of changes in chemical loading (designated A-F) described as follows: A, flow increased to 3.0 mL/min; B, glyphosate concentration dropped to 400 mg/L; C, flow increased to 5.0 mL/min; D, flow increased to 10.0 mL/min; E, pH of feedstock was dropped to 2.8 for 18 h; F, feed changed to clarifier overfow enriched with 50 mg glyphosate/L and flow was increased to 15.5 mL/min.
3.0 mL/min (HRT 350 min), and on day 5 the concentration of glyphosate in CSW water feed was dropped to 400 mg/L. The flow rate was increased to 5.0 mL/min (HRT 210 min) on day 8. A GDA of > 99% was established on column B by day 10. Doubling the flow rate (10 mL/ min, HRT 105 min) on day 17 resulted in a 40% breakthrough of glyphosate on days 20-22, but GDA returned to > 90% removal by day 24 and >99% removal by day 28. A pH shock of column B was inadvertently conducted on day 34 when CSW water feedstock was not titrated to neutral pH. This resulted in a pH drop on column B to 2.8 for about 18 h. The feedstock was then titrated to pH 6.5, and GDA was monitored in the column. Total GDA ranged from 30 to 60% removal of glyphosate for 4 days after the pH upset, but it returned to > 99% by day 45, 10 days after the pH upset. Column B was transferred from St. Louis, Mo., to Monsanto's glyphosate production plant on day 55. The biocarrier bed depth was dropped from 55.9 cm to 30.5 cm upon reassembly and the pumping rate was increased to 15.5 mL/min. This decrease in biocarrier bed volume and higher pumping rate resulted in a HRT of 37.4 min. The removal of glyphosate dropped to 45%, but it returned to >99% within 2 days. Throughout all experiments with column B, the increase in GDA observed in response to increases in glyphosate loading was accompanied by an increase in visible biomass on the column.
IBT column C This column had a biocarrier depth of 55.9 cm and was inoculated with microorganisms from the fill and draw activated-sludge bioreactor after acclimation to high GDA in clarifier overflow. Glyphosate removal by column C is presented in Fig. 4. Initial feed for column C was CSW water containing 400 mg glyphosate/L at a pumping rate of 3 mL/min (HRT 350 min). In contrast to columns A and B, column C showed >99% GDA immediately. To rapidly increase the bacterial biomass on the column, the concentration of glyphosate in CSW was increased on day 9 to 1400 mg/L. The GDA of column C dropped to 85% removal on day 14 but returned to >99% removal by
DAYS
FIG. 4. Glyphosate degradation by IBT column C. The initial feedstock was CSW water containing 400 mg glyphosate/L and 50 mg NH4N03/L pumped at a rate of 3.0 mL/min. The column underwent a series of changes in chemical loading (designated A-C) described as follows: A, glyphosate concentration increased to 1400 mg/L; B, feed changed to clarifier overflow enriched with 50 mg glyphosate/L and 25 mg NH4N03/L, and flow was increased to 25 mL/min; C, column C was transferred to Monsanto's glyphosate production plant.
day 16, when a significant increase in bacterial biomass was observed on column C. Flow studies were initiated on column C on day 27 by increasing the pumping rate to 25 mL/min (HRT 42 min). The feedstock for the high flow rate studies consisted of 50 mg glyphosate/L and 25 mg NH4N03/L added to clarifier overflow. The GDA on column C averaged > 96% removal during 21 days of incubation in the laboratory. However, as a result of the high volumes of clarifier overflow needed for feed, column C was transferred on day 48 to Monsanto's glyphosate production plant. Column C was set up with a 55.9-cm biocarrier bed depth, and > 99% GDA was established within 4 days of transfer at a pumping rate of 25 mL/min (HRT 42 min) using a feedstock consisting of clarifier overflow that contained 50 mg glyphosate/L and 50 mg NH4Cl/L.
Microbial colonization of biocarrier Total microbial populations immobilized on the biocarrier in each IBT column were determined by plating studies at the conclusion of the cell-growth phases. This time point corresponded to the appearance of dense biomass on the columns resulting from bacterial growth on CSW water containing 500-2000 mg glyphosate/L. Microbial populations presented as total colony forming units per gram dry weight of biocarrier were as follows: column A, 9.5 x lo7; column B, 3.3 x lo7; column C, 1.1 x lo8. Although the bacterial strains were not identified in this study, plates from columns A and C routinely showed six to eight different colony types. Plates from column B showed predominantly one colony type, with two others occurring a low levels. Light microscopy showed the predominant colony type on column B was a short gram-negative rod, which presumably was Pseudomonas sp. strain LBr (Jacob et al. 1988). Microbial attachment and penetration into the biocarrier were determined by SEM analyses for column C on day 20. The SEM analyses of the outer 20% of a cross-sectioned R635 diatomaceous earth bead (Fig. 5A) showed significant, but not confluent, bacterial colonization. More confluent microbial growth probably occurred later in the study, when column C was operated at maximum chemical loading. The
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
926
CAN. J. MICROBIOL. VOL. 38, 1992
FIG. 5. Scanning electron micrographs taken near the surface of a cross-sectioned Manville R635 diatomaceous earth bead (A) showing microbial colonization occurring in the outer 20% of the bead and (B) taken near the center showing bacteria that had penetrated deep into the bead matrix.
SEM analyses also showed that some bacteria had penetrated deep into the center of the diatomaceous earth bead (Fig. 5B). TOC analyses were conducted on column influents and effluents at several time points throughout the study. Unspiked clarifier overflow had TOC levels that ranged from 230 to 280 mg/L. TOC removal on the columns receiving clarifier overflow spiked with technical-grade glyphosate ranged from 18 to 37%. Since microbial growth and GDA were not maintained under conditions of extremely low organic nutrients, as described earlier (Fig. 2 for column A), GDA by the immobilized cells is most likely a co-metabolic phenomenon that requires the presence of a sufficient level of degradable TOC for maintaining viable and active cells. Although the minimum TOC requirements were not determined in this study, the results indicate that levels of degradable TOC in the clarifier overflow and the polishing pond are sufficient to maintain GDA activity on immobilized cell columns. High-flow studies on ZBT column A The biocarrier bed depth of column A was reduced from 55.9 to 30.5 cm by removing 25.4 cm of the R635 biocarrier from the top of the column. The 30.5-cm immobilized cell column had a smaller fluid volume (Table I), which resulted in a shorter HRT than the 55.9-cm columns at comparable pumping rates. This shortened column was used to monitor GDA in a series of high flow rate studies presented in Fig. 6. Initially, feedstock for these experiments consisted of diluted CSW water (100 mg glyphosate/L) spiked with 400 mg technical-grade glyphosate/L. GDA in column A remained > 98% during the first 19 days of the experiment as the flow rate was increased from 3 (HRT 193 min) up to 20 mL/min (HRT 29 min). On day 20 the pumping rate was increased to 25 mL/min (HRT 23.2 min) and the CSW water concentration was lowered to 50 mg glyphosate/L; technical-grade glyphosate was no longer added to the feedstock. GDA remained >98% for the next 5 days, and on day 26, the flow rate was increased to 30 mL/min (HRT 19.3 min). GDA dropped steadily to 77070 removal over the next 6 days but then recovered to 82% during the follow-
10 0
I
0
,
10
I
,
20
,
I
30
I
40
,
50
I
,
60
70
DAYS
FIG. 6. Glyphosate degradation by IBT column A during high flow and fluidization studies. The initial feedstock was CSW water containing 500 mg glyphosate/L and 50 mg NH,NO,/L pumped at a rate of 3.0 mL/min. The column underwent a series of changes in chemical loading (designated A-G) described as follows: A, flow was increased to 5.0 mL/min; B, flow was increased to 10 mL/min; C, flow was increased to 20 mL/min; D, concentration of glyphosate was dropped to 50 mg/L, 100 mg K,HPO,/L was added, and flow was increased to 25.0 mL/min; E, flow was increased to 30 mL/min; F, biocarrier was completely fluidized for 5 min to remove excess biomass; G, biocarrier was completely fluidized for 5 min to remove excess biomass.
ing 3 days. These results indicate that the optimal performance of column A occurred at a HRT between 23.2 and 19.3 min. Throughout the remainder of this experiment, the pumping rate was maintained at 30 mL/min (HRT 19.3 min) to provide constant glyphosate breakthrough needed to examine the effects of fluidization on column performance. Since excessive biomass was observed on column A on day 36, the column was purged (fluidized) with a fivefold increase in the aeration rate for 2 min. Although the R635 biocarrier was not fluidized enough for complete mixing of the biocarrier to occur on the column, a large portion of the biomass was blown off the biocarrier and pumped out of the column. This fluidization resulted in a drop in GDA to 55070 removal on day 39. However, GDA recovered to 77% removal on day 44 and ranged from 71 to 83% removal
HEITKAMP ET AL.
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
during days 45-57. Although excessive biomass had not accumulated on the column by day 57, the column was fluidized again to observe the effects of another fluidization cycle on the loss and recovery of GDA. The second fluidization resulted in a smaller loss of GDA (65% after 3 days), a faster recovery (1 day), and a higher level of GDA after recovery (78-86%). Fluidization probably benefits GDA by removing biomass (increasing fluid volume) and enriching for bacteria that attach more securely or penetrate more deeply into the biocarrier. These results indicate that an immobilized cell operational strategy should include a periodic fluidization cycle to benefit GDA. Discussion This study demonstrates that IBT can achieve >98% removal of low levels of glyphosate ( 1 5 0 mg/L) from aqueous waste streams at HRTs as short as 23 min. The high glyphosate removal rates that IBT achieves in this study probably result from combining the high densities of bacterial cells on the columns, using highly selected microbial cultures having high GDA, and optimizing physical and nutritional conditions on the IBT columns. In addition, the degradation of glyphosate to AMPA, its primary metabolite, only requires a one-step, two-carbon cleavage, which can be expected to occur rapidly. In previous laboratory studies, IBT has been reported to degrade > 99% of pnitrophenol (PNP) in a synthetic waste stream containing 1200-1800 mg PNP/L at an HRT of 48 min (Heitkamp et al. 1990). However, IBT's capacity to remove low levels ( s50 mg/L) of PNP from high-volume waste streams was not determined in that previous study. The present study shows that IBT is promising as a tertiary treatment technology for removing low levels of glyphosate from high-volume aqueous wastes. The ability to withstand and recover quickly from pH and osmotic shocks may be a significant advantage for IBT over some conventional technologies. Scanning electron microscopy in this study showed that most of the immobilized bacteria occurred on the outer portions of the biocarrier, but a significant number of bacterial cells were detected throughout the interior of the diatomaceous earth beads. This penetration of bacteria into the R635 diatomaceous earth biocarrier may serve to maintain a stable bacterial inoculum in the reactor during system upsets. This hypothesis would suggest that bacteria inside .the beads may exist in a microenvironment that is protected from severe chemical or pH shocks, which may be lethal to unprotected bacteria on the bead's surface. Furthermore, the ability of these protected bacteria to serve as inoculum could explain the rapid recovery of GDA by the IBT columns in this study when favorable conditions were restored. The minimal loss of GDA after repeated fluidization of the IBT column in this study suggests that full-scale IBT reactors should be routinely fluidized to remove excess biomass. Fluidization serves to remove interstitial biomass from the biocarrier bed and maximize interstitial fluid volume. This larger fluid volume results in a longer hydraulic residence' which benefit GDA by increasing the exposure time of waste water to immobilized bacteria in the reactor. Routine fluidization may also benefit IBT performance by selecting bacteria that attach more securely to the porous biocarrier, resulting in high GDA at high flow rates with minimal washout of immobilized bacteria and mini-
927
ma1 channeling of liquid wastes through the reactor. The immobilized bacteria in this study were not able to maintain high GDA when fed glyphosate as a sole source of carbon and energy. This suggests that GDA is linked to some minimal level of additional carbon loading needed to maintain the viability and activity of the immobilized cells. However, this should not create a problem for using IBT for tertiary biotreatment of effluents, since this study did show that clarifier overflow exiting an activaged-sludge secondary-treatment facility contained sufficient low levels of degradable organic carbon to support viability and activity of immobilized microorganisms having GDA. Furthermore, since the immobilized bacteria were able to maintain >99% removal of glyphosate from an extremely low organic carbon waste stream for a total of 106 HRTs occurring over a 5-day period, small fluctuations or temporary decreases in levels of degradable organic carbon in effluent would not be expected to negatively impact IBT performance for removing low levels of glyphosate. The minimum level of secondary carbon needed to support GDA was not determined in this study but warrants further investigation. The immediate increase in GDA in the fill and draw reactor and the IBT columns after adding NH4N03 suggests that nitrogen supplementation should be used for IBT treatment of glyphosate in aqueous wastes. However, inorganic nitrogen should only be added at the minimum levels needed to support GDA, since excessive nitrogen in the bioreactors could select for high populations of nitrifying bacteria, which would require excessive amounts of oxygen. Since high rates of GDA were achieved by IBT in this laboratory-scale study, pilot-scale studies were conducted to confirm the applicability of IBT for glyphosate removal from high-volume waste effluents. The pilot-scale study was designed to provide the GDA rate data needed to engineer and price a full-scale system and confirm the optimum operating conditions (pH, aeration, temperature, nitrogen supplementation, fluidization frequency, etc.) for IBT. In addition, the pilot-scale study was designed to determine the response of immobilized bacteria to surge loading of glyphosate to simulate a worst-case biosystem upset as well as investigate the recovery of GDA after long periods of dormancy in which the immobilized bacteria are exposed to minimal concentrations of glyphosate. The results of pilot-scale testing of IBT for removal of glyphosate from high-volume liquid wastes are presented in a separate paper (Hallas et al. 1992). Acknowledgements The excellent technical assistance of Martina BianchiniAkbeg, Stephanie Lannert, Will Renaudette, Tom Reuter, and David Schlictman of the Environmental Sciences Center, and Eileen Hahn and Harold Crouch of the Monsanto Agricultural Company contributed greatly to the successful completion of this project. The authors thank Carol Pellegrine for performing scanning electron microscopy analyses. Balthazor , T.M., and Hallas, L.E. 1986. Glyphosate-degrading microorganisms from industrial activated sludge. Appl. Environ. Microbial. 51: 432-434. cook, A.M., ~ ~ c.G., and ~ Alexander, ~ M.h 1978. ~ Phos~honateutilization by bacteria. J. Bacteriol. 133: 85-90. ~ o h r m a n nRosemount ~ n a i ~ t i c Division. al 1988. Total organic carbon analyzer. 3rd ed. Dohrmann, Santa Clara, Calif.
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by McMaster University on 12/02/14 For personal use only.
928
CAN. J. MICROBIOL. VOL. 38, 1992
Hallas, L.E., Hahn, E.M., and Korndorfer, C. 1988. Characterization of microbial traits associated with glyphosate biodegradation in industrial activated sludge. J . Ind. Microbiol. 3: 377-385. Hallas, L.E., Adams, W.J., and Heitkamp, M.A. 1992. Glyphosate degradation by immobilized bacteria: field studies showing removal of glyphosate from industrial wastewater. Appl. Environ. Microbiol. 58: 1215-1219. Heitkamp, M.A., Camel, V., Reuter, T.J., and Adams, W.J. 1990. Biodegradation of p-nitrophenol in an aqueous waste stream by immobilized bacteria. Appl. Environ. Microbiol. 56: 2967-2973. Jacob, G.S., Schaefer, J., Stejskal, E.O., and McKay, R.A. 1985. Solid-state NMR determination of glyphosate metabolism in a Pseudomonas sp. J. Biol. Chem. 260: 5899-5905. Jacob, G.S., Garbow, J.R., Hallas, L.E., et al. 1988. Metabolism of glyphosate in Pseudomonas sp. strain LBr. Appl. Environ. Microbiol. 54: 2953-2958. Kishore, G.M., and Jacob, G.S. 1987. Degradation of glyphosate by Pseudomonas sp. PG2982 via a sarcosine intermediate. J. Biol. Chem. 262: 12 164 - 12 168. Leadbetter, E.R., and Foster, J.W. 1958. Studies on more methaneutilizing bacteria. Arch. Mikrobiol. 30: 91-1 18. Liu, C.M., McLean, P.A., Sookdeo, C.C., and Cannon, F.C. 1991. Degradation of the herbicide glyphosate by members of the family Rhizobiaceae. Appl. Environ. Microbiol. 57: 1799-1804. McAuliffe, K.S., Hallas, L.E., and Kulpa, C.F. 1990. Glyphosate degradation by Agrobacterium radiobacter isolated from activated sludge. J . Ind. Microbiol. 6: 219-221. Mueller, M.M., Rosenberg, C., Siltanen, H., and Wartiovaara, T. 1981. Fate of glyphosate and its influence on nitrogen cycling in two Finnish agricultural soils. Bull. Environ. Contam. Toxicol. 27: 724-730.
Murthy, D.V.S., Irvine, R.L., and Hallas, L'.E. 1989. Principles of organism selection for the degradation of glyphosate in a sequencing batch reactor. Proceedings of the 43rd Annual Purdue Industrial Waste Conference. Lewis Publishers, Chelsea, Mich. pp. 267-275. Pipke, R., and Amrhein, N. 1988. Isolation and characterization of a mutant of Arthrobacter sp. strain GLP-1 which utilizes the herbicide glyphosate as its sole source of phosphorus and nitrogen. Appl. Environ. Microbiol. 54: 2868-2870. Pipke, R., Schulz, A., and Amrhein, N. 1987a. Uptake of glyphosate by Arthrobacter sp. Appl. Environ. Microbiol. 53: 974-97 8. Pipke, R., Amrhein, N., Jacob, G.S., et al. 19876. Metabolism of glyphosate in Arthrobacter sp. GLP- 1. Eur . J. Biochem. 165: 267-273. Rueppel, M.L., Brightwell, B.B., Schafer, J., and Marvel, J.T. 1977. Metabolism and degradation of glyphosate in soil and water. J. Agric. Food chem. 25: 517-522. Shinabarger, D.L., Schmitt, E.K., Braymer, H.D., and Larson, A.D. 1984. Phosphonate utilization by the glyphosate-degrading Pseudomonas sp. strain PG2982. Appl. Environ. Microbiol. 48: 1049-1050. U.S. Environmental Protection Agency. 1983. Methods for nonconvenient pesticide analysis of industrial and municipal wastewater: method 127-determination of glyphosate in wastewater. U.S. EPA Publication No. 440/1-83/079-C, U.S. Environmental Protection Agency, Washington, D.C. pp. 1- 10. Wackett, L.P., Shames, S.L., Venditti, C.P., and Walsh, C.T. 1987. Bacterial carbon-phosphorus lyase: products, rates and regulation of p hosp honic and p hosp hinic acid metabolism. J. Bacteriol. 169: 7 10-717.
