Sterols as bio-markers for waste impact and source characterization in stream sediment.

Introduction

Contamination from human and animal waste is a complex but important aspect of environmental concern. Human and animal fecal waste is a major pollutant of domestic wastewater, streams, and rivers in rural and agricultural areas.

The approximately 3 billion gallons of water consumed daily in Iowa are derived from both surface-water and groundwater systems. Streams contribute 75 percent, while underground aquifers account for 25 percent of the water consumed (Iowa Department of Natural Resources, 1997). Ordinarily, thermo-tolerant fecal coliforms and nitrate loads are monitored as point-in-time indicators of water pollution. These analytes cannot, however, provide information for identification of pollution source, and they lack specificity for human waste and various sources of agricultural runoff (Nichols, Leeming, Latham, & Rayner, 1996; Phillips, Venkatesan, & Bowen, 1996; Rayner, Nichols, Nivens, & White, 1996; Sherblom, Henry, & Kelly, 1996; Standley & Kaplan, 1996; Venkatesan, Ruth, & Kaplan, 1986; Vivian, 1986; Walker, Wun, & Litsky, 1982).

Determination of sterols in stream sediment may complement determination of fecal coliform and nitrate as means of indicating waste impact. Identification of sterol compounds as well as measurement of other pollutants may provide information regarding pollution source.

Chemical pollutants originating from human, animal, or industrial wastes and runoff of agricultural chemicals may be divided into two groups: 1) biological metabolites and 2) synthetic chemicals. Biological metabolites originate from human and animal feces (stanols, sterols, and urobilin); urine (uric acid and its metabolites, and urobilin); antioxidants; antibiotics; and vitamins/nutrients (Leeming, Ball, Ashbolt, & Nichols, 1996). Caffeine has also been used to indicate waste of human origin. Synthetic chemicals that originate from industrial and agricultural activity include various pesticides (Takada, Satosh, Bothner, Tripp, & Farrington, 1996). Analysis of stream sediment and identification of sterol compounds has been used to provide an anthropogenic history of fecal contamination and identification of animal source; however, in the past, analytical methods were complex and involved the use of highly undesirable solvents.

[FIGURE 1 OMITTED]

Work by Leeming and Nichols (1996) has demonstrated the potential of utilizing 5[beta]-cholestan-3[beta]-ol (coprostanol) as a key biochemical marker. Coprostanol is produced in the digestive tract of humans by the microbial hydrogenation of cholesterol (Rosenfield & Gallagher, 1964). It has been proposed as a measure of human fecal pollution by a large number of researchers (Murtaugh & Bunch, 1967; Dutka, Chau, & Coburn, 1974; Walker et al., 1982). Fecal wastes from humans; household pets (cats and dogs); livestock (cattle, ducks, poultry, swine and horses); and sea mammals (Takada et al., 1996) have been analyzed to obtain reference profiles.

Analysis of human and animal feces has provided data for identification and characterization of 32 different cholesterol derivatives and a quantitative estimation of individual sterols (Leeming et al., 1996). Concentration patterns of sterols vary in animals and humans; consequently, evaluation of sterol patterns can assist in possible identification of pollution origin. Species-specific animal sterols and metabolites generating a sterol fingerprint (Figure 1) are controlled by 1) the ability of the animals to biosynthesize the endogenous sterol, cholesterol; 2) presence of hydrogenating anaerobic bacteria in the digestive tract; and 3) dietary behavior and load. The usefulness of sterol pattern to indicate possible sources depends on environmental stability, pattern consistency in species of interest, and the very real possibility of "interference" from sterols contributed by other species. Major sterol compounds produced by human, farm animals, and domestic pets are summarized in Table 1.

Major sterol components found in human feces include coprostanol, 24-ethylcoprostanol, cholesterol, 24-ethylcholesterol, 5[beta]-stigmasterol, 5[beta]-epistigmasterol, campestanol, sitosterol, and sitostanol. Coprostanol and 24-ethylcoprostanol are most abundant in human feces. Coprostanol and 24-ethyl-coprostanol generally are present at a ratio of approximately 3:1, with individual concentrations greater than 1 [micro]g/g dry weight in feces. Total-sterol content in human feces ranges from 3 to 9 [micro]g/g dry weight. Treatment of fecal waste with anaerobic microorganisms in wastewater treatment plants alters the concentrations of individual sterols. The digestive biota of pigs and humans are similar, and the observed fingerprint pattern of sterols is also similar. Among the omnivores, pigs and humans lack the metabolic production of isofucosterol (28-homo-24-methylenecholesterol). Subtle differences in total concentration and abundance of individual sterols are present, however. Total and individual sterol compound concentrations are found in pigs and humans in ratios ranging from 1:4 to 1:10. Pigs lack the metabolic chemistry to produce 5[beta]-stigmasterol and 5[beta]-epistigmasterol; therefore, stigmasterol and stigmastanol are not found in pig feces, while these sterols are present in human feces. [C.sub.29]-cholesterols are found at greater concentrations in pigs than [C.sub.27]-cholesterols. These sterols are therefore very important biomarkers for differentiating human-source and animal-source contamination.

Cholesterol is the second most abundant sterol present in wastewater. In human and animal cell membranes, cholesterol is the major sterol constituent. Cholesterol has been shown to biologically hydrogenate into coprostanol and cholestanol (dihydrocholesterol) in anoxic sediments, and the conversion to coprostanol has been shown to be negligible. It has been suggested that ratios comparing coprostanol concentration to cholesterol-plus-cholestanol concentration could be used as a simplified indicator for sewage contamination in river sediment (Writer, Leener, Barber, Amy, & Chapra, 1995). When coprostanol is observed in environmental sediment, it is considered to be from fecal sources only, while summation of cholesterol and cholestanol represents sewage and non-sewage sources. Furthermore, concentration ratios are independent of particle size and the percentage of organic matter. In the investigation by Writer and co-authors (1995), most sites had sterol ratios of approximately 1:17, representing small-scale impact from sewage. Sites with ratios greater than 1:17 (>0.06) are indicative of large source pollution.

Sterols attach to particulates and tend to accumulate in stream sediment. Sterol concentrations in sediment can act as an integrating exposure index of animal impact, human impact, or both, because these compounds remain in the streambed after other soluble markers have been washed downstream (Venkatesan & Kaplan, 1990). Development and validation of a simplified method of determining sterol compounds in sediment was a primary objective of the study reported here. The study also investigated application of the simplified method for determination of sterols in actual environmental samples, especially from rural agricultural areas, natural wildlife ecosystems, and urban populations.

Materials and Methods

Materials

Analytical standards were used for the target sterol compounds coprostanol (5[beta]-choles-tan-3[beta]-ol), cholesterol, dihydrocholesterol, stigmasterol, sitosterol, stigmastanol, and ergosterol (Sigma, St. Louis, Missouri), and [d.sub.6]-cholesterol (Cambridge Isotope Laboratories, Inc., Andover, Massachusetts) was used as a surrogate standard. Perylene-[d.sub.12] (Restek Corporation, Bellefonte, Pennsylvania) was used as an internal standard. The derivatizing agent was N,O-bis (trimethylsilyl) trifluoroace tamide (BSTFA) with 1 percent trimethylchlorosilane (1 percent TMCS) as catalyzing agent (Pierce, Rockford, Illinois).

Sampling Sites

Sediment samples were collected from streams passing through both agricultural and wastewater treatment plants (upstream and downstream) in urban areas. Three Iowa sites representing streams affected by livestock and crop farming were sampled, as were urban impact sites near waste treatment discharge areas. The Maquoketa River at Backbone State Park, in Strawberry Point, Iowa, was sampled as a control site. A Laurel Run site at Yellow Creek State Park in Indiana, Pennsylvania, was included as a remote reference.

Analytical Methods

The authors adapted methods outlined by previous investigators (Bligh & Dyer, 1959; Marty, Quemeneur, Aminot, & Lecorre, 1996; Mayer, Bourbonniere, & Crowe, 1996; Nguyen, Bruchet, & Aprino, 1995; Protection of Environment, 1994, Appendix B; Writer et al., 1995) using ultrasonication, derivatization, and subsequent analysis by fused-silica capillary column gas chromatography/mass spectrometry (GC/MS).

Sterols were extracted from sediment samples with a modified version of U.S. Environmental Protection Agency (U.S. EPA) Method 3550A--Ultrasonic Extraction--from Test Methods for Evaluating Solid Waste (SW-386) (U.S. EPA, 1994). Sediments were initially saponified to free esters that occur naturally in the environment. Saponification resulted in few natural esters, however, and lower recovery under some conditions. This step was subsequently eliminated for most samples. The percentage of moisture content was determined for all sediment samples.

Sterol compounds were extracted from sediment by the addition of 100 mL of methylene chloride followed by sonication; an ultrasonic disruptor with a 3/4-inch horn-type device was used, at full power in pulse mode and at 50 percent duty cycle (Heat Systems, Ultrasonics, Inc.). The resulting extract was decanted into a centrifuge bottle and centrifuged for five minutes at 2,000 rpm to separate the solvent (and analytes of interest) from the sediment. The authors removed any residual water from the methylene chloride extract by passing the extract through dried anhydrous sodium sulfate. This process was repeated with two additional 100-mL portions of methylene chloride. The extract was concentrated with a TurboVap nitrogen evaporation device (ZymarkCorp, Hopkinton, Massachusetts), dried again with sodium sulfate, and evaporated to a final volume of 0.5 mL. To derivatize the sterols, the authors added 0.5 mL of BSTFA + 1 percent TMCS to the extract and maintained the extract at 70[degrees]C for one hour to form the trimethyl silylether (TMS) derivative, which was stored at 4[degrees]C until completion of analysis. The TMS derivative improved chromatography and sensitivity, Prior to extraction, [d.sub.6]-cholesterol was added to blanks, spikes, and samples as a surrogate standard for monitoring of precision and method accuracy. The analytical method, an adaptation of U.S. EPA Method 8270C--Semivolatile Organic Compounds by GC/MS--provides mass-spectral confirmation data for specific compounds and is capable of distinguishing sterol-based compounds from other late-eluting compounds in environmental samples. The latter capability permitted a rough total-sterol concentration to be determined, under the assumption of similar response factors among compounds, and it also permitted crude fingerprinting of sterols and closely related compounds.

The Hewlett-Packard GC/MS system had a ProSep 800 Plus precolumn separator with large-volume injection capability, and [d.sub.12]-perylene was used as the internal standard. The method detection limit was determined to range from 90 [micro]g/kg to 190 [micro]g/kg for individual sterols by preparation and analysis of seven replicates (Protection of Environment, 1994).

[FIGURE 2 OMITTED]

Results and Discussion

Coprostanol, cholesterol, dihydrocholesterol (cholestanol), stigmasterol, sitosterol, and stigmastanol were identified and quantified by GC/MS method. In many environmental and fecal samples, 24-ethyl coprostanol appears as a major peak (Figure 2). Other tentatively identified sterols were quantified with the total ion current to the internal standard.

Recovery of the surrogate standard ([d.sub.6]-cholesterol) was 73.6 [+ or -] 7.8 percent for water (n = 10) and 77.9 [+ or -] 16.3 percent for field sediment samples (n = 124). Duplicate sample collection was included in the study design for approximately 10 percent of all field samples, and of the total number of sediment samples (124) included in the study, 17 samples were collected in duplicate. In six duplicate sets, no sterols were detected in either sample. In an additional six duplicate sets, target sterols were detected at low levels (<4 mg/kg total sterols), with very low levels detected in one sample (<1 mg/kg total sterols) and no detection of sterols in the duplicate sample. Calculation of a pooled standard deviation for the remaining four low-level sediments resulted in a relative standard deviation (RSD) of 87 percent, generally consistent with the large uncertainty expected near the detection limit for any method. For the remaining five duplicates, total target sterol concentrations ranged from 4 to 24 mg/kg, total tentatively identified sterol concentrations ranged from 0.56 to 3.3 mg/kg, and total sterol concentrations ranged from 4.6 to 27 mg/kg; calculated pooled standard deviations were 24 percent, 34 percent, and 27 percent, respectively.

Ratios observed in the study were different in some respects from those found by Leeming and co-authors (1996). Analysis of lagoon samples from two different swine confinement operations shows a different pattern of sterols. Sitosterol and cholesterol are the predominant sterols in raw feces. Lagoon slurries contain coprostanol and dihydrocholesterol as the primary sterols, while sitosterol and cholesterol are relatively minor components. These data suggest that the sterol content depends on factors other than the species of animal. Factors may include type of animal feed and degradation of sterols during lagoon retention.

The concentration of target sterols observed in sediments ranged from undetectable (no sterol peak observed) to 74 mg/kg. The total concentration of tentatively identified sterols in the samples ranged from undetectable to 30 mg/kg, with a ratio to the total target compounds of 0.33 [+ or -] 0.21 for samples in which measurable concentrations in both categories were observed (n = 46). The sum of the two concentrations (total sterols) for each sample ranged from undetectable to 94 mg/kg. A relatively linear sterol increase up to a concentration of about 15 mg/kg of total target sterols (approximately 17 mg/kg total sterols) was observed, and there were several sporadically higher concentrations, in 9 of 124 samples (7.25 percent). Interestingly, three of the nine were from north-central Iowa, an area with a high concentration of swine-feeding operations; one was from a state park in Iowa; two were from a stream in a natural area in northeastern Iowa; two were from the background sites in Pennsylvania; and one was from a reservoir in southeastern Iowa. It was observed that sampling performed in the concentrated-animal-feeding-operations (CAFO) areas of Iowa did not necessarily find higher total sterol concentrations than sampling in non-CAFO areas. Thus, the data indicate that "natural" areas are likely to have sterol concentrations comparable to those in CAFO areas. Additional studies are therefore necessary to determine whether the "break" in concentrations observed (approximately 15-20 mg/kg) represents an indication of higher animal input from CAFO, wildlife, or other biological or environmental factors.

The authors propose that the ratio of coprostanol concentration to cholesterol-and-cholestanol (dihydrocholesterol) concentration in environmental samples may be a better tool than total sterols for measuring the impact of fecal waste. Since coprostanol is thought to originate from fecal sources, and cholesterol and cholestanol represent sewage and nonsewage sources, this ratio may be used as a more specific and simplified indicator of sewage contamination. In the study reported here, many sites had a sterol ratio significantly greater than 0.06, representing impact from large point sources. A sterol ratio of less than 0.06 indicates that a site has no fecal contamination or minimal impact from distant sources. The indicator ratios for samples from Yellow and Twolick creeks in Pennsylvania were similar to those for samples with the highest ratios, from the heavy animal-feeding (CAFO) areas in Iowa, indicating possible sewage contamination from rural treatment plants.

Acknowledgements: The authors would like to acknowledge Lazlo Toth, a Fogarty Scholar from Hungary, who assisted with methodology and extraction. Thanks also go to Mike Scheller, Earlene Erbe, Greg Jacobs, and Ebtsam Selim of the University of Iowa Hygienic Laboratory for analytical measurements, and to John Miller and the staff of Limnology in Des Moines for collection of samples. The research was a collaborative effort of the Environmental Assessment Facility of the Environmental Health Sciences Research Center, the University of Iowa Hygienic Laboratory, and Indiana University of Pennsylvania. The development of the methodology was funded by a grant from the Iowa State Water Resources Research Institute.

Corresponding Author: Amadu Ayebo, Associate Professor of Biology, Indiana University of Pennsylvania, Department of Biology, 124 Weyandt Hall, Indiana, PA 15705-1090. E-mail: [e-mail].

REFERENCES

Bligh, E.G., & Dyer, W.J. (1959). A rapid method of total lipid extraction and purification. Canadian Journal of Biochemistry and Physiology, 37, 911-917.

Dutka, B.J., Chau, A.S.Y., & Coburn, J. (1974). Relationship between bacterial indicators of water pollution and fecal steroids. Water Research, 18, 1047-1055.

Iowa Department of Natural Resources. (1997, January/February). Water: Iowa water, a decade of progress. Iowa Conservationist, 55(1), 1-16.

Leeming, R., Ball, A., Ashbolt, N., & Nichols, P.D. (1996). Using faecal sterols from humans and animals to distinguish faecal pollution in receiving waters. Water Research, 30, 2893-2900.

Leeming, R., & Nichols, P.D. (1996). Concentrations of coprostanol that correspond to existing bacterial indicator guideline limits. Water Research, 30, 2997-3006.

Marty, Y., Quemeneur, M., Aminot, A., & Lecorre, P. (1996). Laboratory study on degradation of fatty acids and sterols from urban wastes in sea water. Water Research, 30, 1127-1136.

Mayer, T., Bourbonniere, R.A., & Crowe, A. (1996). Use of coprostanol in tracking nutrients sources in Point Pelee Marsh. Division of Environmental Chemistry Preprints of Extended Abstracts, 36(2), 268-271.

Murtaugh, J., & Bunch, R.L. (1967). Sterols as a measure of fecal pollution. Research Journal of the Water Pollution Control Federation, 39, 404-409.

Nguyen, D-K., Bruchet, A., & Aprino, P. (1995). Determination of sterols in sewage sludge by combined in situ trimethylsilylation/supercritical fluid extraction and GCMS. Environmental Science & Technology, 29, 1686-1690.

Nichols, P.D., Leeming, R., Latham, V., & Rayner, M. (1996). Detection of sewage pollution in Australean inland and coastal waters. Division of Environmental Chemistry Preprints of Extended Abstracts, 36(2), 175-78.

Phillips, C.R., Venkatesan, M.I., & Bowen, R. (1996). Use of principal component analysis for evaluating the fate and distribution of organic markers in sediments from the San Pedro Shelf. Division of Environmental Chemistry Preprints of Extended Abstracts, 36(2), 168-69.

Protection of Environment (U.S. Environmental Protection Agency), 40 C.F.R. [section] 136, Appendix B (1994). Retrieved October 10, 2004, from http://www.setonresourcecenter.com/cfr/40CFR/P136_008.HTM.

Rayner, M., Nichols, P., Nivens, D., & White, D.C. (1996). Rapid method for extraction of the sterol coprostanol using critical fluid carbon dioxide. Division of Environmental Chemistry Preprints of Extended Abstracts, 36(2), 273-276.

Rosenfield, R.S., & Gallagher, T.F. (1964). Further studies of the bio-transformation of cholesterol to coprostanol. Steroids, 4, 515-520.

Sherblom, P.M., Henry, M.S., & Kelly D. (1996). Evaluation of the "domestic waste" markers coprostanol and epicoprostanol in coastal Florida. Division of Environmental Chemistry Preprints of Extended Abstracts, 36(2), 265-267.

Standley, L.J., & Kaplan L.A. (1996). Tracking sources of dissolved organic matter in streams using molecular markers such as fecal sterols and caffeine. Division of Environmental Chemistry Preprints of Extended Abstracts, 36(2), 191-193.

Takada, H., Satosh, F., Bothner, M., Tripp, B., & Farrington, J. (1996). Anthropogenic molecular markers: Tools to identify the sources and transport pathways of pollutants. Division of Environmental Chemistry Preprints of Extended Abstracts, 36(2), 158-161.

U.S. Environmental Protection Agency. (1994). Ultrasonic extraction. In Test methods for evaluating solid waste: Physical/chemical methods (5th ed.) (EPA SW-386, Method 3550A). Washington, DC: Author.

Venkatesan, M.I., & Kaplan, I.R. (1990). Sedimentary coprostanol as an index of sewage addition in Santa Monica Basin, Southern California. Environmental Science Technology, 24, 208-214.

Venkatesan, M.I., Ruth, E., & Kaplan, I.R. (1986). Coprostanols in Antarctic marine mammals and not human pollution. Marine Pollution Bulletin, 17, 554-557.

Vivian, C.M. (1986). Tracers of sewage sludge in the marine environment: A review. Science Total Environment, 53, 5-40.

Walker, R. W., Wun, C., & Litsky, W. (1982) Coprostanol as an indicator of faecal pollution. CRC Critical Reviews in Environmental Control, 12, 91-112.

Writer, J.H., Leener, J.A., Barber, L.B., Amy, G.L., & Chapra, S.C. (1995). Sewage contamination in the upper Mississippi river as measured by the fecal sterol coprostanol. Water Research, 29, 1427-1436.

Amadu Ayebo, Ph.D.

George M. Breuer, Ph.D.

Terry G. Cain

Michael D. Wichman, Ph.D.

Periyasamy Subramanian, Ph.D.

Stephen J. Reynolds, Ph.D.

TABLE 1 Concentration of Representative Fecal Sterols from Selected
Herbivores and Omnivores*

Sterol                                  Fecal Concentration (Decreasing)

5[beta]-Cholestan-3[beta]-ol            Human > pig > cat > cow > sheep
  (coprostanol)                         > horse > duck > hen > dog
Cholest-5-en-3[beta]-ol (cholesterol)   Dog > cat > duck > human > hen >
                                        cow > pig > sheep >horse
5[beta]-Cholest-3[beta]-ol              Dog > human > cow > sheep > pig
  (cholestanol, 24-dihydrocholesterol)  > cat > horse > duck > hen
24-Ethyl-5[beta]-cholestan-3[beta]-ol   Human > pig > sheep > cow > cat
  (24-ethylcoprostanol)                 > horse > hen > duck > dog
24-Ethylcholestan-5-en-3[beta]-ol       Duck > hen > dog > cat > horse >
  (sitosterol)                          sheep > cow > pig > human

*Adapted from Leeming and co-authors (1996).