Testing of Some Assumptions about Biodegradability in Soil as Measured by Carbon Dioxide Evolutiont

APPLID AD VIROMTAL MICROBIOLOGY, Apr. 1993, p. 121-125 99-224/93/4121-5$2./ Copyright 1993, Amerian Soiety for Mirobiology Vol. 59, o. 4 Testing of Some Assumptions about Biodegradability in Soil as Measured by Carbon Dioxide volutiont AGIM L-DI SHARABIt AD RICHARD BARTHA* Department of Biohemistry and Mirobiology, Cook College, Rutgers University, ew Brunswik, ew Jersey 893-231 Reeived 2 Otober 1992/Aepted 27 January 1993 Conversion to CO2 upon inubation in aerobi soil is one of the standard test proedures to assess biodegradability. It may be measured with unlabeled test ompounds in biometer flasks. In this ase, the bakground CO2 evolution by unamended soil is subtrated from the CO2 evolution by the amended soil and the resulting net CO2 evolution beomes the measure of biodegradation. Alternately, '4Co2 release from radioarbon substrates is measured to assess biodegradability. Both approahes measure ultimate (omplete) biodegradation and bypass the theoretial and tehnial limitations of residue analysis. This report examines the underlying assumptions that, exept for arbon ontent, onversion perentage to CO2 is relatively independent of hemial omposition, that CO2 prodution is proportional to the amount of added test ompound, and that the bakground CO2 evolution of the soil is not influened by the test substane. Work with unlabeled and radiolabeled substrates proved the first two assumptions to be essentially orret. However, more than half of net CO2 prodution may represent the mineralization of biomass and soil organi matter, some of it unrelated to the test ompound. The soil mirobial ommunity in its nongrowing steady state appears to onvert a muh lower perentage of a radioarbon substrate to "4Co2 than a growing soil ommunity that responds to a substantial substrate addition. These findings may help to improve test methods and may aid in the interpretation of test results. The Toxi Substanes Control At alls for a premanufature review of novel hemial substanes onerning their environmental fate and effets, inluding biodegradability (1). The methodology of testing is in ontinuous evolution, but first-tier tests for aerobi biodegradation in soil normally involve CO2 evolution studies from nonlabeled or from 14C-labeled materials. Test protools, as exemplified by the U.S. Food and Drug Administration (FDA) nvironmental Assessment Tehnial Assistane Handbook (11) evolved from tehniques used in biodegradation studies on pestiides and hlorobenzenes, respetively (4, 7). They were seleted as standard proedures beause of their relative simpliity and broad appliability and appear to be satisfatory in these respets. evertheless, they were originally developed for a more limited purpose and did not undergo the bakground investigations that broadly applied standard tehniques deserve. The work desribed here attempts to provide some of this missing yet neessary bakground information. The urrent FDA test protool assumes an organi ompound to be biodegradable if during the test period more than 5% of its arbon ontent is released as CO2. The work performed here attempts to answer the following speifi questions: (i) Is the perent onversion to CO2 of a biodegradable ompound influened by its hemial omposition other than arbon ontent? (ii) Is the perent onversion to CO2 of a test ompound influened by the onentration applied to soil? (iii) Does net CO2 evolution represent atual substrate mineralization as refleted by 14CO2 evolution? * Corresponding author. t ew Jersey Agriultural D-1512-2-92. xperiment Station publiation no. Permanent address: University of Damasus, Damasus, Syria. 121 MATRIALS AD MTHODS Test ompounds and soil. Test materials were hosen to be highly biodegradable, of low volatility, and to represent a variety of hemial strutures. Unless speified otherwise, gluose (arbohydrate), adipi aid (aliphati diaid), benzoi aid (aromati aid), and n-hexadeane (aliphati hydroarbon) were applied to soil at 1 mg/g of soil (dry weight basis). These hemials were obtained from Fisher Sientifi Co. (Springfield,.J.) and were of ertified reagent grade (over 99% pure). The aids were neutralized with aoh prior to use [U-'4C]gluose, speifi ativity of 1.6 GBq/ mmol and purity of 99.6%, determined by high-pressure liquid hromatography by the manufaturer, was obtained from Amersham Corp. (Arlington Heights, Ill.). The delivered solution was diluted either with water or with unlabeled gluose solution to obtain the desired dose per biometer flask. This dose was then added as part of the liquid used in adjusting the soil moisture level. Biometer flasks (4) ontained the equivalent of 25 g of dry soil eah. Freshly olleted ixon sandy loam (texture: sand, 5%; silt, 21%; lay, 29%; organi matter, 5%; ph 5.5 to 6.; water-holding apaity,.65 g/g of dry soil) (3) was used in all experiments. The soil samples were olleted between May and August, from the same loation. The plant over of the soil was landsape lawn. To optimize biodegradation onditions for the tests, we raised the original ph of the soil to 7.5 by addition of 1 mg of CaCO3 per g of soil. This was done at least 5 days prior to the biodegradation tests to avoid the release of neutralization-derived CO2 in the biometers. To avoid any limitation of the biodegradation proess by mineral nutrient defiieny, we added.6 ml of a 1% solution of (H4)2PO4 to every biometer flask, plus suffiient distilled water to bring the soil moisture level to 6% of holding apaity (8). All test substrates exept n-hexadeane were dissolved in this water. Hexadeane was applied to 1 g of air-dried soil, and this dry soil was

122 L-DI SHARABI AD BARTHA thoroughly mixed with fresh soil equivalent to 24 g by dry weight. For this type of sample appliation, a separate ontrol was prepared with 1 g of air-dried soil but no hexadeane. Otherwise, the ontrol was fresh soil that was treated in every respet like the test samples but reeived no test ompound. Inubation was at 27 C. Titrations for CO2 and aeration of the flasks through the Asarite filters were performed initially daily and at 2- to 3-day intervals later in the experiment. et CO2 evolution, defined as the CO2 evolution by substrate-amended soil samples minus CO2 evolution by idential but unamended soil, was alulated by subtrating, at eah measurement point, the average umulative CO2 evolution of the unamended soil from the average umulative CO2 evolution of the substrate-amended soil. To alulate standard deviation, the umulative average CO2 evolution values of the unamended soil samples were also subtrated from the umulative CO2 evolution values of the individual amended soil samples. The CO2 evolution by unamended soil is referred to, for brevity, as bakground CO2 evolution. Analytial and radiohemial proedures. Gluose in soil was determined by the proedure of Fales (6). Anthrone reagent (astman Kodak Co., Rohester,.Y.) was freshly prepared for the measurements. For a standard urve, 25-g soil samples were dosed with the appropriate onentrations of gluose. Immediately, 25 ml of water was added, and the sample was shaken for 2 min and entrifuged to obtain a lear supernatant. This was deanted, and the soil residue was washed in a similar manner with 15 ml of water. The extrats were ombined and adjusted to a volume of 5 ml, and 1 ml of this extrat was used for gluose determination. Inubated soil samples were extrated in an idential manner. A62 was measured with a Shimadzu model UV-265 spetrophotometer. Soil extrat with no added gluose plus anthrone reagent served as a blank. In experiments that measured both total CO2 and 14CO2 evolution, 1 ml of KOH was retrieved from the biometer side arm and plaed in a ounting vial with 15 ml of Oxosol (ational Diagnostis, Manville,.J.) ounting solution. Thereafter, the normal biometer titration proedure was performed on the residual alkali, but the results were adjusted for the volume derease due to radioativity ounting. Radioativity was measured in a liquid sintillation ounter (Beta-Tra 6895; TM Analyti, lk Grove Village, Ill.). Samples from biometer flasks without radioative gluose were used for bakground orretion. Counting effiieny was orreted by the external standard ratio method. All tests were performed with tripliate biometer flasks. Standard deviation (1 SD) is indiated in form of error bars. They were omitted when smaller than the symbol in the figures. RSULTS ffets of hemial omposition and mode of appliation. Figure 1 shows net CO2 evolution from four test ompounds of strongly differing hemial omposition during a 3-day inubation. All these ompounds were found to be highly biodegradable in preliminary tests (data not shown). As expeted, net CO2 evolution strongly depended on the arbon ontents of the test ompounds, but other features of hemial omposition had little effet on the perentage of the arbon mineralized. About 5% of the substrate arbon was released in an initial flush of CO2 within the first 1 days of the experiment. More-gradual mineralization followed a, K 4 8 12 16 2 24 28 32 FIG. 1. Cumulative net CO2 evolution (ACO2) during inubation in soil from the substrates n-hexadeane (H), benzoi aid (B), adipi aid (A), and gluose (G), 1 mg/g of soil eah, with 25 g of soil per biometer flask. The points represent the average of three repliates. The error bars represent 1 standard deviation (SD), omitted when smaller than the point symbol. The numbers at 9, 2, and 3 days represent the perent onversion of eah substrate's alulated arbon ontent to CO2. thereafter, but the rate did not return to soil bakground level within the 3-day inubation. This seond slope of CO2 release was more steep for gluose than for the other test ompounds, and net CO2 from gluose reahed 1% by day 3. This pattern was reproduible with gluose and losely related ompounds suh as starh (data not shown), but its ause needs further investigation. Small amounts of hydrophobi liquids may form droplets in wet soil. In our experiene, they are more evenly distributed with air-dry soil as a vehile as was done here with n-hexadeane. It is very important in suh a ase to treat the soil ontrol in an idential manner. Figure 2 ompares CO2 prodution by the two soil ontrols, one ontaining 1 g (4%) of air-dried soil. This treatment resulted in the release in 3 days of.4 mmol of exess CO2, as muh as the addition of 12 mg of gluose would produe. (n ac -5 ) i.6 83.7% - 1.4 7H9 1.2 / 76.5% 1. 51.7%-/ _. 67.1% A,j B 84./, ).6 <6i-}-3% / ).2 O.~. APPL. VIRO. MICROBIOL. 4 8 1 2 16 2 24 28 32 Ti me(days) FIG. 2. Cumulative total CO2 evolution by the normal soil ontrol (C1) of the experiment shown in Fig. 1. The seond soil ontrol (C2) is similar, exept 1 g of the 25-g soil samples was first air dried and then rewetted. Average of tripliates; the error bars represent 1 SD.

VOL. 59, 1993 BIODGRADABILITY TSTIG I SOIL 123 Un 4 mg/g -o2. l..2-- 78.6% I, 1../ '.8 /- ).6 1.6!_b_i-' 7.8%.4- n -5 )2,' _{-i 7.5% ).2 1.25 C.. 75.5% 4 8 12 16 2 24 FIG. 3. Cumulative net CO2 evolution from 25-g soil samples, amended with.25,.5, 1., and 2. mg of gluose per g of soil. Average of tripliates; error bars represent 1 SD. The numbers at day 22 represent the perent onversion of the alulated substrate arbon to CO2. A omparison of the net CO2 values in Fig. 1 with the bakground CO2 evolution of unamended soil in Fig. 2 gives an idea of the typial bakground noise to biodegradation measurements in soil when unlabeled substrates are used. Seleting, somewhat arbitrarily, day 2 for omparison, the umulative bakground CO2 evolution of unamended soil is 1.5 mmol, and 1.8 mmol when dry soil was added. This ompares to a net umulative CO2 evolution of.65 mmol from gluose and 1.35 mmol from hexadeane, respetively, when the ompounds were applied at the 1 mg/g of soil (1, ppm) level. This is a omfortable ratio, with the signal (net C2) being 43 and 75% of the bakground CO2 evolution. Obviously, at lower substrate appliation levels, the signalto-bakground ratio beomes less favorable. ffets of test substrate onentrations. When gluose (Fig. 3) and adipate (Fig. 4) were added at levels from.25 to 2. mg/g of soil, net CO2 evolution was essentially proportional. Considering that in this experiment the umulative bakground CO2 evolution of the ontrol soil on day 22 was 2.28 mmol, it is obvious that the.5- or 1.-mg/g substrate addition levels give more reliable results than lower additions. At less than.1-mg/g substrate addition level, the tehnique is not reliable (2), and the.2-mglg addition level reommended by the FDA (11) requires very preise tehnique for good results. Soures of net CO2 evolution. In Fig. 1, net CO2 release from gluose reahed 1%, yet this soil sample ontinued to evolve exess CO2 ompared with the untreated soil ontrol. This and similar observations not shown here suggested that stimulated mineralization of soil organi matter may ontribute to net CO2 evolution. Figure 5 shows the disappearane of gluose from soil, as measured hemially by the anthrone method, versus net CO2 evolution over a 21-day inubation. Gluose was not detetable after day 4 and was presumably exhausted by day 5. This oinided with a rate hange in CO2 evolution. The seondary lower rate of exess CO2 evolution ontinued linearly and reahed 78% by day 21. Between days 5 and 21, exess CO2 evolved from something other than gluose, this something other ontributing at least one half of the reorded net CO2. Figure 6 reveals a rather unexpeted relationship between 14C2 evolution from radiolabeled gluose and net CO2 evolution. When radiolabeled gluose alone was applied to ok Ti me (days) FIG. 4. Cumulative net CO2 evolution from 25-g soil samples, amended with.25,.5, 1., and 2. mg of adipi aid per g of soil. Average of tripliate flasks. The numbers at day 22 represent the perent onversion of the alulated substrate arbon to CO2. soil, CO2 evolution did not hange measurably ompared to the soil ontrol (urve not shown). This is understandable, sine the appliation of 1,25, dpm per flask resulted in the addition of only 35 ng of gluose. In the first 5 days of the experiment, a surprisingly low amount (17%) of the radioarbon was mineralized. Thereafter, mineralization of radioarbon beame very slow, amounting to.2% per day. When radiolabeled gluose was added in ombination with 1 mg of unlabeled gluose per g of soil, radioarbon mineralization in the first 5 days inreased to 26%. After that, the radioarbon mineralization rate delined to.4% per day. The same experiment was onduted also at 125, and 35, dpm of radiolabeled gluose added (35 and 95 ng of gluose, respetively). The results were idential in terms of perent radioarbon releases and are, therefore, not presented. et CO2 release (the differene between the 1 mg of gluose per g of soil and soil ontrol urves) at day 5 was.31 mmol, and it was.37 mmol at day 1, representing 37. < I.,..8 78.%. ).8.6 ).6 U1) A 42.%. ).4.2. ).2 CD o. 4 8 12 1 6 2 24 Time (days) FIG. 5. Cumulative net CO2 evolution (-) from 25-g soil samples, amended with 1. mg of gluose per g of soil. Average of tripliate samples; the error bars represent 1 SD. Residual gluose in soil () was analyzed in individual flasks daily by the anthrone method (6). The numbers at day 5 and 21 represent the perent onversion of the alulated substrate arbon to CO2.

124 L-DI SHARABI AD BARTHA n CD -5w C) Vo a U) -o UL) (3) l FIG. 6. Cumulative total CO2 evolution from 25 g of soil amended by 14 ng of radiolabeled gluose (5, dpm) per g of soil (). This CO2 evolution was idential to that of the soil ontrol (data not shown). The same amount of radiolabeled gluose was also applied in ombination with 1. mg of unlabeled gluose per g of soil (@). Perent onversion of the radiolabel to 1'4CO2 is also shown from the soil that reeived radiolabeled gluose only (O) and from soil that reeived, in addition, 1. mg of unlabeled gluose per g of soil (U). Average of tripliate measurements; the error bars (1 SD) are smaller than the symbols and are not shown. and 45.2% gluose mineralization, respetively. Compared with the 14C2 mineralization rates, net CO2 evolution rates after day 5 remained relatively high, amounting to 2% per day. Additional work will be neessary to address, in a systemati way, the variables assoiated with the nature of test soils. While no attempt was made here to aomplish this, when the experiment shown in Fig. 6 was repeated with ixon sandy loam from a loation with a higher organi matter ontent (7% instead of 5%), both 14Co2 and net CO2 evolution perentages were found to inrease (data not shown). At day 1, 4Co2 evolution from 1 mg of gluose per g of soil (563, dpm per flask) was 42.5% + 1.2%, inreasing to 49.4% + 1.3% at day 21. et CO2 evolution values in the same experiment were 82% + 3.9% and 129% 2.5% at days 1 and 21, respetively. DISCUSSIO Advantages of net CO2 and 14CO2 evolution biodegradability testing. In ontrast to residue analysis, net CO2 and 14C2 evolution measurements are simple, nondestrutive, and measure ultimate biodegradation (mineralization) rather than a possibly limited strutural hange. If appropriately 14Clabeled test material is available, the measurements and their interpretations are relatively straightforward. This is unfortunately balaned by the tehnial diffiulty and expense involved in obtaining labeled material. The liensing and the waste disposal problems onneted with radioative work are also a drawbak. A 1992 update of the test guidelines by the U.S. nvironmental Protetion Ageny (5) reommends the use of radiolabeled test ompounds in biometer flasks, yet the doument aknowledges that labeled test ompounds in some ases are impossible to obtain. Thus, some biodegradation tests with unlabeled ompounds are likely to ontinue. Fortunately, biometer measurements show a high degree of preision and onsisteny. In experiments 3 to 4 weeks long, the umulative standard deviation among tripliate flasks was a maximum of 5%, whether alulated for o APPL. VIRO. MICROBIOL. total or net CO2 prodution, and was usually onsiderably less than that. In shorter experiments of 1 days, the umulative standard deviations of total, net, or 14C2 evolutions in tripliate flasks were within 2% and ould not be represented by error bars due to their low values. This onsisteny is a definite advantage for a standard testing tehnique. There is some variation between the mirobial ativity of soil samples, even when freshly olleted from the same loation and handled in an appropriate manner (8). These differenes are quantitative rather than qualitative and result in higher or lower bakground CO2 evolution, along with faster or slower net CO2 evolution rates. During 25 years of experiene with the tehnique, we found extensive rate variations but never positive-negative ontraditions when we onduted biodegradation tests in soil samples olleted in various seasons and from various loations. This observation may not apply to soils that fail to support normal plant life or are mistreated during olletion or storage. An anomalous burst of CO2 evolution upon remoistening of air-dried soil is a well-known phenomenon and it was autioned against earlier in onnetion with biodegradation tests (8). If the homogeneous appliation of a hydrophobi sample neessitates the air drying of some of the soil, a orresponding portion of the ontrol soil needs to be treated identially to avoid serious errors. The burst of CO2 represents primarily mineralization of soil biomass killed by the drying proess. Physial hanges aused by drying and remoistening may ontribute to some desorption and greater availability of non-biomass soil organi matter. Chemial omposition other than arbon ontent appears to influene minimally the perentage onversion to CO2 (Fig. 1), although gluose, the presently reommended positive ontrol substane in biodegradation tests, appears to sustain unusually high CO2 evolution rates long after its disappearane (Fig. 5). Conversion perentage of test ompounds to CO2 was relatively independent of onentration within the tested.25- to 2.-mg/g of soil range. To ahieve easily measurable net CO2 evolution above soil bakground,.5- or 1.-mg levels of test ompound per g of soil appear to be preferable to the presently reommended.2 mg/g. Toxiity of a test ompound in soil is rarely a problem, beause reversible adsorption greatly redues the effetive onentrations to whih miroorganisms are exposed. Origin of arbon in net CO2 evolution. A somewhat disturbing aspet of our results is the fat that while CO2 evolution seemed proportional to the arbon ontent and the onentration of the test substane, at least one-half of the evolved net CO2 did not ome diretly from the test substane (Fig. 5). In some ases, the net CO2 evolution may exeed 1% of the arbon added in the form of the test substane. It has been reported that the addition of readily deomposable organi matter stimulates the mineralization of indigenous soil organi matter (1). The magnitude of this priming effet ould vary from soil to soil, depending on organi matter ontent and on seasonal or limatologial fators. However, in our tests, the responses appeared to be proportional to the added substrate arbon, and the inlusion of a positive ontrol, as reommended in the FDA test protool (11), should allow a orretion for the variable of soil quality. Biomass yield oeffiients in aerobi mirobial metabolism depend greatly on the energy ontent of a substrate (9), but onversion to CO2 in soil appeared to depend on arbon ontent only (Fig. 1). It has been assumed that in short-term soil biodegradation studies, roughly 5% of substrate arbon

VOL. 59, 1993 is onverted to CO2 (1, 11). In experiments of 1 month or longer, substrate arbon initially fixed in biomass is also mineralized in part. Together with substrate-stimulated mineralization of indigenous soil organi matter, this may raise the net CO2 evolution in response to a substrate above 1% of the added substrate arbon. It remains to be explored whether this phenomenon is restrited to highly degradable substrates like the ones used in this study, or whether it applies broadly to all substrate additions. The relationship between radioarbon evolution and net CO2 evolution from gluose in Fig. 6 was unexpeted. The low 16 to 17% radioarbon release from submirogram amounts of gluose may be asribed to unavailability due to absorption. Alternatively, the nongrowing steady-state mirobial ommunity may turn over gluose very differently from a ommunity that is growing in response to signifiant substrate addition. The differene between radioarbon mineralization and net CO2 evolution (.5 and 2% per day, respetively) after day 5 onfirms that soil organi matter not diretly related to gluose addition was part of the net CO2 evolution response. If only gluose breakdown produts and mirobial biomass synthesized in response to gluose addition were mineralized, one would expet the two rates to be similar. As a result of these investigations, we onlude that net CO2 and 14Co2 evolution measurements are useful as firsttier tests for assessing biodegradability in soil. onradiolabeled ompounds should be tested in the.5- to 1.-mg/g of soil range to ahieve net CO2 evolution rates that are easily distinguished from the soil bakground. Radiolabeled gluose and probably also other test ompounds, if applied below 1,ug/g of soil, are onverted to 1'4CO2 in muh lower perentages than test ompounds applied at 1. mg/g of soil. Radiolabeled CO2 originates stritly from the test substrate plus test substrate-derived biomass, but our results show that stimulated biodegradation of indigenous soil organi matter ontributes to net CO2 evolution. Consequently, for mineralization of the same amount of test substrate in soil, onversion perentages to 1'4CO2 are inherently lower than BIODGRADABILITY TSTIG I SOIL 125 net CO2 values, when the latter are expressed as perentage of the theoretial maximum. ACKOWLDGMTS During this work, the senior author was on a sabbatial leave supported by the University of Damasus, Syria. This work was also supported by ew Jersey State funds. RFRCS 1. Alexander, M. 1977. Introdution to soil mirobiology, 2nd ed., p. 128-147. John Wiley & Sons, In., ew York. 2. Bartha, R. 199. Isolation of miroorganisms that metabolize xenobioti ompounds, p. 283-37. In D. P. Labeda (ed.), The isolation and sreening of miroorganisms from nature. Mamillan, ew York. 3. Bartha, R., and L. Bordeleau. 1969. Cell-free peroxidases in soil. Soil Biol. Biohem. 1:139-143. 4. Bartha, R., and D. Pramer. 1965. Features of a flask and method for measuring the persistene and biologial effets of pestiides in soil. Soil Si. 1:68-7. 5. Code of Federal Regulations. 1992. Inherent biodegradability in soil. CFR 4 796.34, p. 224-229. U.S. Government Printing Offie, Washington, D.C. 6. Fales, F. W. 1951. The assimilation and degradation of arbohydrates by yeast ells. J. Biol. Chem. 193:113-124. 7. Marinui, A. C., and R. Bartha. 1979. Apparatus for monitoring the mineralization of volatile 14C-labeled ompounds. Appl. nviron. Mirobiol. 38:12-122. 8. Pramer, D., and R. Bartha. 1972. Preparation and proessing of soil samples for biodegradation studies. nviron. Lett. 2:217-224. 9. Reed, G. 1982. Mirobial biomass, single ell protein and other mirobial produts, p. 541-592. In G. Reed (ed.), Presott and Dunn's industrial mirobiology. Avi Publishing Co., Westport, Conn. 1. U.S. nvironmental Protetion Ageny. 1979. Toxi Substanes Control At Premanufature Testing of ew Chemial Substanes. Fed. Regist. 44:1624-16292. 11. U.S. Food and Drug Administration. 1987. nvironmental assessment tehnial assistane handbook PB87-175345. ational Tehnial Information Servie, Washington, D.C.