Anal. Chem. 1996, 68, 72-78
On-Line Immunoaffinity Extraction-Coupled Column Capillary Liquid Chromatography/Tandem Mass Spectrometry: Trace Analysis of LSD Analogs and Metabolites in Human Urine Jianyi Cai and Jack Henion*
Analytical Toxicology, Diagnostic Laboratory, Cornell University, 927 Warren Drive, Ithaca, New York 14850
An on-line immunoaffinity extraction-coupled column capillary liquid chromatography/tandem mass spectrometry (IAE/LC/LC/MS/MS) method is described. The system involves three columns, a 2.1-mm-i.d. protein G immunoaffinity column with noncovalently immobilized antibody specific to the analytes of interest, a packed capillary trapping column, and a packed capillary analytical column. With use of a short packed capillary trapping column, the protein G column could be operated at flow rates of 2.5-4 mL/min while the packed capillary analytical column was maintained at a flow rate of 3.5 µL/min. Human urine diluted 1:1 with phosphate-buffered saline was pumped directly onto the immunoaffinity column without pretreatment and was analyzed by electrospray mass spectrometry following the column switching process. Sample handling and transfer procedures were eliminated. The system was optimized and evaluated for the determination of LSD, its analogs, and metabolites in spiked human urine at low part-per-trillion (ppt) levels using mass spectrometric detection. LSD-positive human urine specimens from LSD users were also analyzed. Concentrations as low as 2.5 ppt of LSD and several of its analogs were detected in spiked human urine using IAE/LC/LC/MS/MS. This is 20-fold below our previous limit of detection using solid phase extraction and LC/ MS/MS. Mass spectrometry (MS) has been increasingly used in pharmaceutical, clinical, and biochemical applications due to its high specificity, high sensitivity, and ability to provide structural information. In particular, mass spectrometry, coupled with a modern analytical separation such as high-pressure liquid chromatography (HPLC),1,2 gas chromatography (GC),3 or capillary electrophoresis (CE),4,5 provides tremendous analytical capability for real-word problem solving. The success of these techniques depends largely on proper sample pretreatment, an aspect which is frequently tedious and time-consuming for complex sample analysis. Considerable effort has been directed toward the development of sample preparation techniques which are less (1) Huang, E. C.; Wachs, T.; Conboy, J. J.; Henion, J. D. Anal. Chem. 1990, 62, 713A-725A. (2) Suter, M. J.-F.; Dague, B. B.; Moore, W. T.; Lin, S. N.; Caprioli, R. M. J. Chromatogr. 1991, 553, 101-116. (3) Nelson, C. C.; Foltz, R. L. Anal. Chem. 1992, 64, 1578-1585. (4) Smith, R. D.; Wahl, J. H.; Goodlett, D. R. Anal. Chem. 1993, 65, 574A584A. (5) Cai, J.; Henion, J. J. Chromatogr. 1995, 703, 667-692.
72 Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
laborious and amenable to automation. At present, most sample cleanup methods involve methodologies such as liquid-liquid extraction,6 solid phase extraction (SPE),7,8 and supercritical fluid extraction (SFE).9,10 These off-line extraction methods are often labor-intensive and involve possible losses of analytes during evaporation and reconstitution steps. With the introduction of online trace enrichment, it is possible to circumvent these drawbacks. A widely used on-line method is a dual-column approach using a precolumn with an appropriate sorbent for the compounds of interest11-15 or a membrane extraction disk16,17 and an analytical column. In applications such as trace analysis of chemically complex samples, where high sensitivity is required, a selective sample pretreatment technique is often preferred. Immunoaffinity extraction (IAE) using antibody-antigen interactions can provide high specificity for targeted analytes. This technique can provide selective sample enrichment, which can improve detection limits for trace analytes. Immunoaffinity extraction as a means of sample enrichment has been demonstrated, either under low pressure and off-line18,19 or under high pressure and on-line with another analytical separation such as HPLC20,21 and GC.22 A system for on-line immunoaffinity extraction and HPLC using continuousflow fast atom bombardment mass spectrometry detection has also (6) Lim, H. K.; Andrenyak, D.; Francom, P.; R. T., J.; Foltz, R. L. Anal. Chem. 1988, 60, 1420-1425. (7) Zief, M.; Kiser, R. Am. Lab. 1990, January, 70-83. (8) Horack, J.; Majors, R. E. LC-GC 1993, 11, 74-90. (9) Camel, V.; Tambute, A.; Caude, M. J. Chromatogr. 1993, 642, 263-281. (10) Chester, T. L.; Pinkston, J. D.; Raynle, D. E. Anal. Chem. 1994, 66, 106R130R. (11) Opper, C.; Wesemann, W.; Barka, G.; Haak, G. LC-GC 1994, 12, 684-698. (12) Campins-Falco, P.; Herraez-Hernandez, R.; Sevillano-Cabeza, A. Anal. Chim. Acta 1993, 284, 67-71. (13) van der Vlis, E.; Irth, H.; Tjaden, U. R.; van der Greef, J. Anal. Chim. Acta 1993, 271, 69-75. (14) Buick, A. R.; Sheung, C. T. C. F. J. Chromatogr. 1993, 617, 65-70. (15) Pichon, V.; Hennion, M.-C. J. Chromatogr. 1994, 665, 269-281. (16) Brouwer, E. R.; Lingeman, H.; Brinkman, U. A. T. Chromatographia 1990, 29, 415-418. (17) Chiron, S.; Alba, A. F.; Barcelo, D. Environ. Sci. Technol. 1993, 27, 23522359. (18) Bagnati, R.; Castelli, M. G.; Airoldi, L.; Oriundi, M. P.; Ubaldi, A.; Fanelli, R. J. Chromatogr. 1990, 527, 267-278. (19) Shahtaheri, S. J.; Katmeh, M. F.; Kwasowski, P.; Stevenson, D. J. Chromatogr. 1995, 697, 131-136. (20) Haasnoot, W.; Ploum, M. E.; Paulussen, R. J. A.; Schilt, R.; Huf, F. A. J. Chromatogr. 1990, 519, 323-335. (21) Moretti, V. M.; van de Water, C.; Haagsma, N. J. Chromatogr. 1992, 583, 77-82. (22) Farjam, A.; Vreuls, J. J.; Cuppen, W. J. G. M.; Brinkman, U. A. T.; de Jong, G. J. Anal. Chem. 1991, 63, 2481-2487. 0003-2700/96/0368-0072$12.00/0
© 1995 American Chemical Society
Figure 1. Schematic diagram of the column switching system.
been described.23 Recently, we implemented an on-line enrichment method using immunoaffinity extraction carried out under HPLC conditions and coupled column chromatography with mass spectrometric detection.24,25 An objective of this study was to evaluate the feasibility of coupling immunoaffinity extraction on-line with packed capillary HPLC and tandem mass spectrometry (IAE/LC/LC/MS/MS). The strategy for this combination is derived from the known benefits of small-bore HPLC columns for increasing the sensitivity for LC/MS under electrospray conditions.26 An immunoaffinity column operated at high flow rates (2.5-4 mL/min) was coupled to a packed capillary HPLC analytical column (flow rate of 3.5 µL/min) through a short packed capillary trapping column. The use of a packed capillary analytical column has several advantages, including improved mass sensitivity27,28 and facile coupling to mass spectrometry.29-31 The fact that the flow rates are only in the order of one-hundredth of those for conventional HPLC also makes capillary HPLC environmentally and economically attractive. Thus, the column switching system has the unique advantage for trace analysis of complex samples since it offers improved detection sensitivity while eliminating sample handling and manual sample transfer steps. On-line tandem mass spectrometric detection provides additional selectivity and a means for characterization and confirmatory information.32,33 This report demonstrates the performance of the IAE/LC/ LC/MS/MS system for the determination of LSD and its analogs (23) Davoli, E.; Fanelli, R.; Bagnati, R. Anal. Chem. 1993, 65, 2679-2685. (24) Rule, G. S.; Henion, J. D. J. Chromatogr. 1992, 582, 103-112. (25) Rule, G. S.; Mordehai, A. V.; Henion, J. Anal. Chem. 1994, 66, 230-235. (26) Hopfgartner, G.; Wachs, T.; Bean, K.; Henion, J. Anal. Chem. 1993, 65, 439-446. (27) de Jong, G. J.; Lammers, N.; Spruit, F. J.; Dewaele, C.; Verzele, M. Anal. Chem. 1987, 59, 1458-1461. (28) Moritz, R. L.; Simpson, R. J. J. Chromatogr. 1992, 599, 119-130. (29) Huang, E. C.; Henion, J. D. Anal. Chem. 1991, 63, 732-739. (30) Kassel, D. B.; Musselman, B. D.; Smith, J. A. Anal. Chem. 1991, 63, 10911097. (31) Wolf, S. M.; Vouros, P. Chem. Res. Toxicol. 1994, 7, 82-88. (32) Rule, G.; McLaughlin, L. G.; Henion, J. Anal. Chem. 1993, 65, 857A-863A.
or metabolites in spiked human urine at low part-per-trillion (ppt) levels. Human urine diluted with phosphate-buffered saline (PBS) was pumped directly onto the immunoaffinity column without pretreatment and was analyzed by electrospray mass spectrometric detection following a column switching sequence. Representative data from LSD-positive human urine specimens from LSD users are also presented. EXPERIMENTAL SECTION Materials. Lysergic acid diethylamide (LSD) and N-demethylLSD (nor-LSD) were purchased from Alltech Associates (Deerfield, IL) and Radian Corp. (Austin, TX), respectively. Iso-LSD and the LSD-positive human urine specimens were kindly provided by Drs. C. Nelson and R. Foltz of Northwest Toxicology (Salt Late City, UT). 6-Nor-6-allyllysergic acid diethylamide was purchased from Sigma Chemical Co. (St. Louis, MO). Deethyl-LSD was synthesized in this laboratory according to the method described by Johnson and co-workers.34 All chemicals used for the HPLC mobile phases were purchased from Fisher Scientific (Rochester, NY). The antibody against LSD was supplied by Dr. S. Salomone of Roche Diagnostic Systems (Belleville, NJ). Water was deionized in-house using a Barnstead NANOpure analytical type D4700 deionization system (Dubuque, IA). Apparatus. Mass Spectrometer. The mass spectrometer was a Sciex TAGA 6000E atmospheric pressure ionization (API) triple quadrupole mass spectrometer (Thornhill, Ontario, Canada) upgraded to an API-III. Electrospray interface was used for highsensitivity capillary LC/MS and was maintained at 4.5 kV. Both mass analyzers were operated at unit mass resolution. The declustering energy (the potential difference between orifice and the AC rods) for both MS and MS/MS experiments was set at 40 eV. Ultrapure argon was used as collision gas in the second (33) Burlingame, A. L.; Boyd, R. K.; Gaskell, S. J. Anal. Chem. 1994, 66, 634R683R. (34) Johnson, F. N.; Ary, I. E.; Teiger, D. G.; Kassel, R. J. J. Med. Chem. 1973, 16, 532-537.
Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
73
Table 1. Procedure for the Determination of LSD Analogs and Metabolites in Human Urine by IAE/LC/LC/MS and IAE/LC/LC/MS/MS step ASV position 1 2 3 4 5
inject inject inject inject load
6
load
7 8
inject inject
events equilibrate protein G column with PBS by pumping PBS through the protein G column to waste at 200 µL/min prime the protein G column by injecting PBS-diluted antibody solution load drugs onto the protein G column by pumping PBS-diluted urine through the column to waste at 2.5-4 mL/min remove unretained impurities from protein G column by pumping PBS through the column to waste at 2.5 mL/min for 1 min equilibrate both protein G and trapping columns with PBS by pumping PBS through both protein G and trapping columns at 200 µL/min for 2 min release the drugs and antibodies from the protein G column and collect them on the trapping column by pumping elution solvent through protein G and trapping column at 200 µL/min for 5 min back-flush trapped drugs from trapping column to the analytical column where HPLC separation takes place remove residual IgG from protein G column by pumping stripping solvent through protein G column at 2.5 mL/min for 1 min before reequilibrating with PBS
quadrupole. For MS/MS experiments, the collision gas thickness was 250 × 1012 atoms/cm2, and the collision energy was 45 eV. Coupled Column Chromatography. A schematic diagram of the column switching system for this study is shown in Figure 1. Pump 1 was a Waters Model 510 (Division of Millipore Corp., Milford, MA); pump 2 was a Brownlee Labs microgradient system (Applied Biosystems, Inc., Santa Clara, CA); pump 3 was an Applied Biosystems 140A solvent delivery system (Applied Biosystems, Inc., Foster City, CA). The immunoaffinity column was a HiPac protein G column (2.1 mm × 3.3 mm, 30 µm, ChromatoChem, Missoula, MT), operated at 2.5-4 mL/min flow rates. The analytical column was a C-18 packed capillary column (0.3 mm × 150 mm, 3-µm particles, LC Packings, Amsterdam, the Netherlands), operated at a flow rate of 3.5 µL/min. The two columns were coupled through a short C-18 packed capillary trapping column (0.5 mm × 15 mm, 5 µm, LC Packings). Other LC equipment included an Acurate microflow processor (Model AC-20-CAP, LC Packings), an Applied Biosystem 757 absorbance detector (ABI, Foster City, CA) equipped with a U View capillary flow cell (LC Packings), a Hewlett-Packard 3390A integrator (Palo Alto, CA), a Waters automated switching valve (Milford, MA), a Rheodyne Model 7010 switching valve (Cotati, CA), and a Rheodyne Model 7125 injector equipped with a 100-µL external loop. Adsorption and washing buffers were phosphate-buffered saline (PBS), which is 0.01 M sodium phosphate (pH 7.4) containing 0.15 M sodium chloride. The elution solution for the desorption of IgG from the protein G column was 2% acetic acid in deionized water. The stripping solution for removal of all the residual IgG from the protein G column after each analysis was 20% acetic acid in deionized water. The HPLC eluent (30% ACN/ 30% MeOH/0.1% HOAc/5 mM NH4OAc) was maintained at 3.5 µL/min through the analytical packed capillary column using a precolumn splitting device (Acurate microflow processor, Model AC-20-CAP, LC Packings). Procedure. The experiments with on-line IAE/LC/MS and IAE/LC/MS/MS for the determination of LSD and its metabolites in human urine were performed according to the sequence described in Table 1 using the hardware shown in Figure 1. The automated switching valve was initially set at the “inject” position. The protein G column was equilibrated with PBS at 200 µL/min. An aliquot (30 µL) of PBS-diluted antibody solution (10% antibody/ 90% PBS) was injected onto the protein G column. Next, human urine diluted with PBS (50% urine/50% PBS unless otherwise specified) was pumped through the protein G column at 2.5-4 mL/min. This was immediately followed by flushing with PBS (2.5 mL/min for 1 min) to remove unretained and weakly bound 74 Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
Figure 2. Structures of LSD and its analogs.
impurities. Meanwhile, both the trapping column and the analytical column were equilibrated with mobile phase for HPLC separation. The automated switching valve was turned to the “load” position to allow PBS to be pumped through both protein G and trapping columns. Bound analytes and antibodies were desorbed from the protein G column with the elution solution (2% HAc) and directed to the trapping column. The switching valve was then returned to the inject position to back-flush trapped analytes from the trapping column to the analytical column, where analytical separation takes place under isocratic conditions. At the end of each run, the protein G column was re-equilibrated with PBS for the next sample after being treated with the stripping solution (20% HAc) to remove any residual components from the previous sample. RESULTS AND DISCUSSION Determination of LSD and Its Analogs in Spiked Human Urine. Due to its low dose and extensive metabolism,11,12 the determination of LSD and its metabolites in human body fluids is a challenging analytical task. Sample cleanup methods for the characterization of LSD and its metabolites in urine include liquid-liquid3,6,35 and solid phase extraction.36 Studies by radio immunoassay (RIA) suggest that the actual concentration of LSD in body fluids is often considerably lower than indicated by the RIA.37 The difference is likely due to the immunocross-reactivity of LSD metabolites. Since the elimination half-lives of LSD (35) Henion, J.; Wachs, T.; Foltz, R. L. In Proceedings of the 39th ASMS Conference on Mass Spectrometry and Allied Topics, Nashville, TN, May 20-24, 1991; pp 1653-1654. (36) Cai, J.; Henion, J. J. Anal. Toxicol., in press. (37) Blum, L. M.; Carenzo, E. F.; Rieders, F. J. Anal. Toxicol. 1990, 14, 285287.
Figure 3. IAE/LC/LC/UV chromatograms obtained with urine samples. (A) control blank urine; (B) urine spiked with LSD analogs at 10 ppt level.
Figure 4. Chromatograms obtained with mass spectrometric detection using selected ion monitoring. (A) Control blank urine; (B) urine spiked with LSD analogs at 10 ppt level.
metabolites are longer than those of LSD itself,6 the presence of these metabolites may be detectable much longer after the level of LSD has dropped below the limit of detection. Thus, the detection of LSD metabolites may aid the confirmation of illicit LSD use. Focusing on the identified LSD metabolites,36,38 a group of LSD analogs were selected for this study; their structures are shown in Figure 2. Urine samples spiked with the LSD analogs at the 10 ppt level were analyzed to evaluate the described system for trace analysis from complex matrices. Figure 3 shows the IAE/LC/LC/UV chromatograms obtained for the analysis of blank and spiked urine samples using the column switching system and the procedures described in the (38) Nelson, C. C.; Foltz, R. L. J. Chromatogr. 1992, 580, 97-109.
Experimental Section. In both cases, an aliquot of 50 mL of PBSdiluted (1:1) human urine was pumped directly through the protein G column at a 2.5 mL/min flow rate. The chromatogram shown in (A) was obtained from a control blank urine, while the one shown in (B) is from a urine sample spiked with LSD and its analogs at 10 ppt each. The arrow denotes the expected retention time for LSD. Little difference between these two chromatograms is observed, indicating the complete occlusion of analyte signals by the significant interference from the coeluting endogenous material. These results suggest that UV detection is unable to provide the required selectivity or sensitivity for trace analysis from a complex matrix, even with the prior on-line immunoaffinity extraction. Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
75
Figure 5. Chromatograms obtained with mass spectrometric detection using selected reaction monitoring. (A) Control blank urine; (B) urine spiked with LSD analogs at 10 ppt level.
Figure 4 shows the corresponding IAE/LC/LC/MS chromatograms obtained with a blank urine (A) and urine spiked with the LSD analogs at the 10 ppt level (B) using selected ion monitoring (SIM) instead of UV detection. The mass spectrometer was set such that no “up-front” CID took place.26 The protonated molecules of the analytes spiked into the urine were monitored at the appropriate m/z values. The expected retention times of the spiked components are denoted by arrows. Only norallylLSD (m/z ) 350) was detected under these SIM/IAE/LC/LC/ MS conditions (Figure 4B). The coelution of interference components precluded the detection of the remaining analytes. Although SIM provided some degree of selectivity, it is still not sufficient for the determination of these trace levels of components in untreated human urine samples. The above experiments were repeated using selected reaction monitoring (SRM)39 to obtain improved selectivity. The selected reaction for each compound is determined on the basis of the fragmentation pathways of the LSD analogs studied previously.36 Figure 5 shows the SRM/IAE/LC/LC/MS chromatograms of the corresponding blank urine (A) and urine spiked with the LSD analogs at the 10 ppt level (B). The arrows denote the expected retention times for the spiked components. No peaks were observed in the control blank (A) as expected, while the chromatograms in (B) for the spiked urine clearly show the expected chromatographic profiles. These results dramatically demonstrate the feasibility of the SRM/IAE/LC/LC/MS system for the determination of components at low ppt levels from a complex matrix without additional sample pretreatment. The potential of the SRM/IAE/LC/LC/MS system for even lower levels of these analytes was demonstrated for the analysis (39) Watson, J. T. Introduction to Mass Spectrometry; Raven Press: New York, 1985; pp 321.
76
Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
Figure 6. Chromatogram of urine spiked with LSD analogs at 2.5 ppt level, obtained with mass spectrometric detection using selected reaction monitoring.
of a urine sample spiked at the 2.5 ppt level, as shown in Figure 6. This is 20-fold below our previous limit of detection using solid phase extraction and SRM/LC/MS for the determination of LSD and analogs from spiked human urine.36 Two hundred milliliters of PBS-diluted (1:1) human urine was pumped onto the immunoaffinity column at a 4 mL/min flow rate. Most of the spiked
Figure 7. SRM/LC/MS chromatogram obtained with (A) a control blank urine and (B) a LSD-positive urine specimen. The original measured concentration of LSD in the specimen determined by GC/MS is 0.9 ng/mL
analytes were detected at this low levels. The loss of the deethylLSD signal is probably due to the combination of the competitive binding of all the spiked LSD analogs with the immobilized antibody, the decrease in concentration of the analyte in the urine, and the increase in flow rate during loading. By optimizing the selectivity of the antibody and the trapping condition used in this study, it may be possible to further improve the limit of detection for all the spiked analytes. Determination of LSD and Metabolites in LSD-Positive Human Urine Specimens. Focusing on possible in vitro and in vivo human metabolites and their CID characteristics,36,38,40 several LSD-positive human urine specimens were analyzed using SRM/IAE/LC/LC/MS. One hundred milliliters of PBS-diluted human urine (10% urine/90% PBS) was pumped through the protein G column at 4 mL/min. Figure 7 shows SRM/IAE/LC/ LC/MS chromatograms obtained with a control blank urine (A) and an LSD-positive urine specimen (B) for the characterization of LSD and its metabolites. The original measured concentration of LSD in the specimen determined by GC/MS is 0.9 ng/mL.41 The labeled peaks shown in Figure 7B indicate the presence of LSD and its metabolites in the LSD-positive human urine extracts; these peaks were absent in the control blank urine (Figure 7A). The exact structures of the detected metabolites were not confirmed. However, based on their molecular weights and previous metabolism studies of the drug,36,38,40 it is reasonable to postulate the metabolic incorporation of oxygen atoms into the LSD molecule. Also, there is evidence indicating that LSD is metabolized into an aromatic amine compound in both rhesus monkey42 and man.40 (40) Foltz, R. L. In Proceedings of California Association of Toxicologists; Burlingame, CA, February 4, 1995; pp 20-29. (41) Foltz, R. L., personal communications.
This application suggests the potential for the on-line immunoaffinity extraction system in drug metabolism studies. The system can be easily adapted for the characterization of other drugs and their metabolites by immobilizing antibodies specific to the drugs of interest in the immunoaffinity column.25 Since most metabolites retain some substructural features of the parent drug, a bioselective extraction method such as immunoaffinity extraction might be most appropriate for isolation and concentration of these structurally related substances. Coupled with mass spectrometry, this technique may enable the selective enrichment of drug metabolites and obtain valuable information for their characterization. CONCLUSIONS This report describes for the first time the direct combination of on-line immunoaffinity extraction and coupled column capillary HPLC/tandem mass spectrometry. It also demonstrates the feasibility of the system for the determination of trace analytes in spiked as well as LSD-positive urine samples. The system has combined both automated sample pretreatment and improved detection sensitivity. Levels as low as 2.5 ppt of LSD and its analogs or metabolites were detected in spiked human urine. This is 20-fold below our previous limit of detection, obtained using solid phase extraction and SRM/LC/MS for the determination of LSD and its analogs from spiked human urine.36 The technique is shown to be useful for the selective enrichment and characterization of drug metabolites with similar substructures as the parent drug. Future work will be directed toward multiresidue (42) Siddik, Z. H.; Barnes, R. D.; Dring, L. G.; Smith, R. L.; Williams, R. T. Biochem. Pharmacol. 1979, 28, 3093-3101.
Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
77
analysis and quantitative determination of trace analytes in complex matrices. ACKNOWLEDGMENT Partial financial support for this project was provided by the Office of Naval Research Laboratories Grant No. N00014-93-K2000. We acknowledge LC Packings for providing us with several packed capillary HPLC columns and helpful advice. We thank Drs. C. Nelson and R. Foltz of Northwest Toxicology, Salt Lake City, UT, for many helpful suggestions and for providing us with an analytical standard of iso-LSD and LSD-positive human urine
78
Analytical Chemistry, Vol. 68, No. 1, January 1, 1996
specimens, Dr. R. Hammen of ChromatoChem for providing us with the protein G column, and Dr. S. Salomone of Roche Diagnostic Systems for supplying several LSD antibodies. Special thanks also go to Dr. T. Wachs for his invaluable technical support.
Received for review July 31, 1995. Accepted October 11, 1995.X AC950763I X
Abstract published in Advance ACS Abstracts, November 15, 1995.
