W _ m .4.” 4x. «I u 5.1;. . 1..) .C... 7. 12.. . r31! .. sulrn “sin-”l . {fv MEG mm mm m 5%. 9...... v5.51. .4 4ry.-‘-. ,olo . 1.. '3’: in. ;“""'g This is to certify that the thesis entitled ANALYSIS OF EGG STEROIDS presented by Frido Hans Harnarm has been accepted towards fulfillment of the requirements for - 31.13. degree in Food Science "N , ( ‘- 71/60 flair-ocean :1 Major professor Dateggiz. CZ @717/ 0-7639 ANALYSIS OF MG STEROIDS By Frido H. Hanann The cholesterol content of eggs can be lowered by adding certain sterols to the diet of chickens. Since no method could be found in the literature to evaluate the steroid canposition and its changes due to such feeding or other causes, this study was initiated to develop an analytical procedure with which modifications could be made evident. Since all. steroids in eggs, accept cholesterol, occur only at trace level concentration, several methods had to be deve10ped for this procedure. A method was cmceived to quantitatively extract the total lipids from an entire egg yolk and to saponify this lipid extract with minimal oaddative degradation, yielding the nonsaponifiable fraction of egg yolk. Three solvent systems were selected for thin-layer chromatograpm' (TIC) to maximize the preparative separation of groups of egg yolk non- saponifiable canpounds of corresponding polarity. A micro TIC-band recovery and extraction procedure was devised to reduce the couple-city of regular TIC-band extracts and for trace caupound analysis. Two columns were adopted for gas liquid chromatography (arc) showing 'I E" VSI’MrA vv-x g; screening of derivatizati . short-time ( eat reactio tim conte 501V'Bnt By. W am of maneii 9° peaks. m Frido H.1-Lamann separation of canpounds based on molecular size or configuration and functional substituents and operating conditions were optimized for screening of individual TIC-band extracts . Methyloodme -trimethylsilyl- derivatization was adopted for analysis of thermally labile canp‘ounds or specific separations by gas chromatography (GC) . Preparative GC was investigated to recover canpounds separating insufficiently or only on the column which could not be used for mass spectrometry. Repetitive, short-time digitonin reaction was developed, taking advantage of differ- ent reaction rates, to tentatively enhance trace canpounds within frac- ticms containing predominantly cholesterol. Two silver nitrate TIC solvent systems were selected for canpounds which could not be separated by any of the above methods. Mass spectranetry was used to investigate homogeneity and to obtain mass spectra for identification of individual 60 peaks. All these methods have been canbined to a procedure which is pro- posed for the evaluation of changes in egg steroid composition. To facilitate identification, a table was developed in which molec- ular fornmlae of steroids with 5 1+ double bonds, composed of carbon, hydrogen, and £ 5 oxygen atoms, were arranged by molecular weight. Mathematical formulae for the interconversion of GC relative retention times and methylene units were derived allowing the use of corresponding data from the literature. The use of the methods is demonstrated for 3 canpounds which were isolated fran eggs and identified by the described procedure: 48(9) ,lh- ergostadienol, cholest-h-en-B-one, and squalene. ANALYSIS 0? MG STEROIDS FridoHans Hamann A DISSERTATION Submitted to Michigan State University in partial mnillment of the requirements for the degree of DmTOR 01" 3111080151? Department of Food Science and Hanan Nutrition 19714 ©Copyright by FREDO HANS HAW 1971; To nw wife and m mother 11 m A“ ——q—_ — .— .-— _— _—.—_—'—__ __ The Dr. L. E. carry cu he Bergen, doctors; ACMGMENTS The author wishes to express sincere gratitude to his advisor, Dr. L. E. Dawson, for his various contributions which enabled me to carry out this research. He also would like to express his appreciation to Drs. W. G. Bergen, D. R. Romsos, C. M. Stine, and M. E. Zabik for serving on the doctoral committee. Further expressions of gratitude are extended to the Department of Food Science and Human Nutrition for financial assistance during 11w graduate studies at Michigan State University. Sincerest appreciation is extended to R. H. Thompson for his helpful suggestions and cooperation in common research interests. Finally, the author is especially grateful to his wife, Heide, who has provided moral support and help throughout his graduate program. 111 TABIEOFCONI'I'INTS DEICATION O O O O O O O O O O O O O O O 0 O O 0 O O O W O O O O O 0 O O O O O O O O O 0 O O O m .0? mm 0 O O 0 O O O O O O O O O O O O O O O m w ”W O O O O O O O O O O O O O 0 O O O O O mmm O O O O O O O O C O O O C O C O O O O Gmem O O O O O O O O C O O O O I O O O O O O Steroids in Egg Y01k O O O O O O O O O O O O O O BackgroundonMethodology............ DEVBIDBWTSOFWS ............... Wtim O O O O O O O O O O O O O O 0 mm 0 O O O O O O 0 O O O O O O O O 0 Discussion of Extraction Procedures . . . . GeneralObservations............ Explanation of the Extraction Procedure . . SEMficationeeeeeeeeoeeeeeoee medum O O O O 0 O O O O O O O O O O O 0 tion of the Saponification Procedure Digitonin Precipitation . . . . . . . . . . medm O O O I O O O O O O O O O 0 Review of Digitonin Precipitation . . Ebcplanation of the Digitonin Procedure Recovery of Digitonin Precipitated Material Procedure....o......... Review of Recovery Procedures . . . . Ebcplanation of the Recovery Procedure mn'mrmtommeeoeeeoeo memseeeeeeeeeeoeeoe Review of Separation on Thin-Layer Plates Preparation of TIC-Plates . . . . . . . . Application of Yolk Nonsaponifiable Materialtothe Plates . . . . . . .. ElutimOftheHateS eeeeoeeeeeo iv 0 O O O O O O O O O O O O O O O O O O O Page iii vii «now-a E d? $fl5555§§%$%%33$835§§ Visualization of Individual Bands . . . ....... 58 Recovery of Individual Bands from TLC-Plates . . . . 59 Preparation of Silver Nitrate TIC-Plates . . . . . 61 Use of Silver Nitrate TIC-Plates . . . . . . . . . . 63 Break-down and Adsorption on TIE-Plates . . . . . . . 65 Gas-LiquidChranatography................ 69 Preparation of Columns and Operation of GC . . . . . 69 Derivatization................... 72 mative GC 0 O O O O O O O O O I O O O O O O O 0 7h Relative Retention, Steroid Number and MethyleneUnits........ ....... 76 MassSpectrometry............ ....... 81 Miscellaneousnethods.......... ....... 85 Recrystallization of Cholesterol . . . . . . . . . 85 Recrystallization of Yolk Extract and of Nonsaponifiable Material . . . . . Liquid Column Chranatograptnr . . . . . Removal of Free Fatty Acids . . . . . . ROPOSIDOVERAILMUMTOEVAHIATECWESIN msmmcmmm ......OOOOOOOOOO. m WOOOOOOOOOOOOOOOOOOOOOO0.... ” CIWOfGJ-asm ......OOOOOOOOOOO. ” PurificationofSolvents................. 100 Drying of Nitrogen and Use of Plastics . . . . . . . . . . 101 WtionOme-natea 0000000000000... 102 PreparationongNO TIC-Plates ............. 103 Preparation of Mic TIC-Band Collectors . . . . . . . . . 10h Mflcatim Of Digitonin I O O O O O O O O O O O O O O 0 1w PackingandOptimizingGC-Colms............ 106 SpecificMaterialsanquuipnent............. 109 W m DISC‘JSSIM O O 0 O O O O O O O O O O O O O O 0 O O 1.12 mmmm O O O O O O O O O O O O O O O 0 O O O O O O O O 0 B9 all...» 1 WI GI... J: F . LISTOFTABIE Table Page 1. Solubilities of digitmin, cholesterol-digitonide, andcholesterolinvarimzssolvents......... 1+0 2. Molecular formulae of steroids arrangedby molecular weight for steroids canposed of carbon, oxygen, andtnrdrogen withé hdoublebmdsandi'jooqrgenatans......... 95 LIST OF FIGURES Figure Page 1. Arrangement for collection of GC-fractions in preparativegaschromatograpby........... 75 2. Correlation of relative retention times and memlelleunitseeeeoeeeeeeooeeooe sea 3. GummmzatimOfmflmoeeeooso... h. GC-trace of the total nonsaponifiable fraction OfmeggyOJ-keeeeeeeoeeeeeeeoeee m 5. PreparativeTIC,nonpolarelution............ 115 6. Preparative TIC, medium-polar elutim . . . . . . . . . . 117 7. PreparativeTIC,polarelution............. 1.18 8. GC-traceofamacroTIC-bandonfiiSP-ZHOJ. 1.1.9 9. GC-traceofamacroTIC-bandonfiSP-aloo 120 10. ism-graph of.aGC-peakofamacro.TI£-band. . . . . . 122 11. GC-trace of amicro TIC-band on 5% SP-2h01 . . . . . . . 123 12. GC-trace of the supernatant of 2 consecutive, short-time digitonin precipitations . . . . . . . . 126 13. MS-chromatogram of a repetitive scan with seleCtiveimsemhoeeeeeeeeoeeeooe 128 1"". mbar‘mmOfacmmmmeoeoeeeeeee 15. MSbar-graphofareasmablypurecaspound ....... l6. LBbar-graph of eholest-h-en-3-one ........... EQEQ l7.lear-gra1hofsqualene ................ vii INTRODIETION General The report of the Inter-Society Omission for Heart Disease Resources (1970) on Primary Prevention of the Atherosclerotic Diseases cited the habitual diet high in saturated fat-cholesterol-calories as one of the risk factors for coronary heart disease. They atmnarized the general finding that human populations consuming diets high in saturated fat and cholesterol have high mean senmi cholesterol levels and high incidence and mortality rates fran premature coronary heart disease, whereas the opposite is true for those on diets low in saturated fat and cholesterol. For these reasons they recannended for the general public an adjustment of the caloric intake to maintain optimal weight , a reducticn of the dietary cholesterol intake to less than 300 mg per day, and a substantial reduction in the dietary saturated fats. With respect to particular staples they stated that "the public should be encouraged to avoid egg yolk consumption and the food industry should be persuaded to minimize egg yolk content of cmercially prepared foods" since "the yolk is the single higiest source of cholesterol in the aver- age American diet, as well as a source of considerable ammmt of satu- rated fat." This ban on the consumption of egg yolk due to its high cholesterol and saturated fat content has led to the industrial developmnt of egg substitutes such as Viobin' s (Viobin Corp. , Monticello, Illinois 61856) 1 desiccated defatted egg powder with the egg oil replaced by corn oil, or Pleischmann's ”Egg Beaters" (Standard Brands Inc., New Ybrk, nay. 10022) which.contains only the egg white part of the egg with a.host of other nonpegg ingredients. Another approach which does not attempt to eliminate cholesterol fran yolk altogether but to reduce it considerably is by selection and genetic modification, i.e. by selecting hens within particular strains which consistently lay eggs with cholesterol content below average and then to try genetically to produce a new strain in which all hens would lay eggs with low'amounts of cholesterol, possibly even lower than the parent chicken. That egg,composition can be changed by using different feeds has long been recognized and was, again, confirmed by Couch and Salome (1973) who stated that although "the quantity of phospholipids, neutral fat, total lipids and cholesterol in egg yolk.cannot be influenced‘by the type or quality of fat in the diet of the hen", "the fatty acid content of neutral fat and phospholipids can be influenced significantly by the fatty acid composition of the fat in the diet of the hen." ‘mis indicates that by feeding a ration with more unsaturated fats it is possible to produce eggs with.an.increased ratio of unsaturated to saturated fats while keeping the total lipid content constant. unfor- tunately,‘weiss, et a1. (196M) reported a simultaneous increase in egg cholesterol content with increasing unsaturation of the yolk lipids. Turk and Barnett (1972) confirmed the observation of others that corn diets when compared with other isocaloric, isonitrogenous diets de- creased the cholesterol content of eggs. They reported that alfalfa added to corn rations was most effective in decreasing egg cholesterol N... .1 -..ln 1.: ., ... II: with least loss of egg size, feed efficiency, and egg production. There was, however, no indication.which.particular compound or sub- stances in either corn oil or alfalfa was responsible for this de- crease. The influence of'drugs on lipid metabolism.was reviewed.by' Britchevsky'(197h) who summarized that serum lipids can.be lowerd by inhibition of cholesterol synthesis, increased cholesterol catabolimm, inhibition of reabsorption of cholesterol or bile acids, inhibition of lipolysis, inhibition of fatty acid synthesis and interference with lipoprotein synthesis or release. Among the many compounds mentioned were D-thyroxine and nicotinic acid.which were fed to laying hens. No attempt is made to extrapolate results with drugs obtained with other animals, since the literature is full of emples demonstrating that equivalent results may be obtained.by switching animals, but that very often this is not the case and one should only rely on results obtained with the animal under consideration. Singh (1972) reported that with nicotinic acid changes in plasma cholesterol, egg choles- terol, egg production.and.egg weight were all statistically insignif- icant, whereas D-thyroxine significantly decreased plasma cholesterol but increased egg cholesterol. Jones (1969) reported the results ob- tained.when cholestyramine, which is known as hypocholesterolemic agent, was included in the chicken feed. He feund.significantly lowered serum.cholesterol levels, as would be expected, but no effect on the cholesterol content in the egg. This result is supported by the finding of Bartov, et a1. (1971) that neither a general reduction in plasma cholesterol nor its periodic fluctuations were reflected in egg yolk cholesterol concentrations. Other conpounds which have been investigated with a potential for lowering the cholesterol content in egg yolk are plant sterols and.some synthetic azasterols. Singh, et al. (1972) reported the effects on 'cholesterol deposition of 2 azasterols, s.c-12937 (63D. Searle Co.) and UA22593A (Upjohn 00.). When the drugs were administered.at levels at which egg production and egg weights remained practically unaffected, the cholesterol was mostly replaced.by desmosterol without marked changes in total egg sterol deposition. They pointed out that sterol composition of eggs could be altered without interference with repro- ductive perfOrmance but did.not elaborate on the desirability of des- mosterol substitution for cholesterol. Clarenburg, et al. (1971) obtained reductions up to 33% in egg cholesterol content by including 2% or more emulsified sitosterol in standard chicken diets. They chose this sterol because it interferes with the absorption of dietary as well as enterOhepatically circulating cholesterol and because no harmful effects of plant sterol administra- tion to humans had been observed. Besides the lowered cholesterol level in the eggs they believed that an additional benefit could be derived from.the approximately h2 mg sitosterol deposited in the yolk due to its interference with the absorption of cholesterol in the intes- tine. Their data indicated not only a reduction in cholesterol but also a reduction in total sterol content in the egg yolk. This was probably not the case but rather reflects the inadequateness of their techniques to analyze for sterols. The difficulty of assessing the sterol content of the nonsaponifiable fraction of egg yolk by colori- metric methods was reported by Singh, et al. (1972). Bartov, et al. (1971) added 1 and at levels of soy sterols to the diet for short periods of time and found no effect on the yolk choles- terol level and assmed that plant sterols do not interfere with the cholesterol metabolism in the laying hen. Considering that the administration of drugs to laying hens in an attmpt to modify the natural canposition of their eggs may have rather serious consequences, it seems Justified to request that eggs be suhnitted to rather extensive analyses of all their substituents and not simply be limited to the measurement of the reduction in choles- terol content. This is certainly to be expected before eggs produced by amt such drug treatment, on a research basis, would be released to the consumers. Since, up to this date, the only successful cholesterol lowering attanpts for eggs were those which used plant sterols or azasterols, further investigations with sterols are anticipated. It can be assumed that the most significant changes on feeding sterols to chickens to lower egg cholesterol would occur in the steroid composition of the egg yolk. Thus, a literature search was initiated to find the methods which are commonly used to determine the steroid profile of egg yolk and which could be followed conveniently. It became clear that no uniform, simple method was available which would provide reliable infor- mation with respect to the composition of egg steroids or alterations in their profile. For this reason the onective of this study was to provide such a method and to show how it should be used within the context outlined above . Steroids in Egg Yolk The most complete identification of egg steroids was reported by Tu, et al. (1970) who found, in addition to large amounts of cholesterol, cmsiderable amotmts of desmosterol, cholestanol, ergosterol, £3 -sitos- terol, A7-cholestenol, lanosterol, A7-methostenol, h,h-d.imethyl- A 7, 2h-cholestadien-3f3 ~01, ditsdmanosterol, and small quantities of AB-methostenol, h0< -methy1-A8,2h-cholestadiene-3fi -ol, and ho< methyl- 4 7,2h-cholestadiene-33 -ol. Pemoch, et al. (1962) found cholesta-3,5-diene-7-one in egg yolk after recrystallization of the tocopherol fraction frm light petroleum ether. ‘lhey stated that 871. of the nonsaponifiable fraction was sterols and recovered 0.3 g cholesterol per yolk. Ergosterol was de- termined as 1 part in 1,350 parts of total sterols. Egg yolk scans to contain a chicken growth promoting factor. Henge, et a1. (1957) showed that this factor is a fat soluble substance and is not identical with oleic acid, linoleic acid, or lecithin and that it is destroyed by hot saponification in 10% ethanolic K011. Other investigators did ascribe an estrogenic activity to the egg yolk but neither Hertelendy and Cannon (1965) nor Tu, et a1. (1970) could detect steroid estrogens in eggs. Bertelendy and Cannon checked their separation procedure with labeled estrone and recovered approximtely 80%. They did not exclude the possibility that the estrogens are pre- sent in amounts below the limit of detectability of their method or that they are of unstable nature. In an investigation using labeled hormones, Arcos (19(2) separated the hormones estrone and estradiol from egg yolk extract. Background on Methodology The first major difficulty in assessing the total steroid content of egg yolk lies in its extreme variability. A number of factors dras- tically influence the reported cholesterol content. The first is re- lated to the way of expressing it, either in absolute amount or as a relative percentage. Jones (1%9) reported that while the size of the eggs of an individual hen changed, the total amount of cholesterol deposited in the egg remained constant and that there are differences between strains of birds and birds of the same strain. Clarenburg, et al. (1971) and Bartov, et a1. (1971) confirmed the differences in yolk cholesterol levels of eggs produced by different hens but found that cholesterol levels in eggs laid by the same hen were remarkably constant. No relationship between weight and cholesterol levels was apparent. Bartov, et al. (1971) found an inverse relationship between the number of eggs laid per unit of time and the amount of cholesterol deposited within each egg. They stated that such marked natural varia- bilities seriously handicap the evaluation of the effects of various dietary factors on yolk cholesterol, especiallywhen the effects are not particularly pronounced. They recamnended using each bird as its own control. This indicates that a method for analyzing changes in steroid composition of eggs should be based on individual eggs rather than on a cross-sample of a series of eggs or even of eggs from different hens. Gems (1972) found that the steroid content of egg yolk is inversely related to its fat content. This seems to be a questionable correla- tion, especially in view of the variables already mentioned above. The sterol content he found averaged 1.55% for the liquid yolk, whereas we VII IriIN.lr, I Ii. I the fat averaged 32 . 011.. Another difficulty associated with the determination of steroids in egg yolk is caused by the extremely high percentage of lipids; nearly two-thirds of the approximately 50% yolk solids, (Parkinson, 1966). In this respect it is quite different from the well documented analysis for steroids of urine, feces, plants, etc. , and seeds rather similar to brain or sane fat tissues without having their organized tissue structure, or to fats and oils, but in coexistence with consid- erable amounts of proteins . With respect to elaborating the steroid profile of eggs an addi- tional complication ananates from the extreme predominance of choles- terol. Tu, et al. (1970) found a cholesterol content of 1.1+; for the wet yolk, whereas the next most abundant sterol, desmosterol, was indi- cated to occur at only 7.5 mg/100 g wet yolk or roughly 0.5% of the cholesterol content. It is, thus, not surprising that the older liter- ature acknowledges the occurrence of plant sterols in eggs but seldanly specified then and mostly reported only cholesterol values . Doorman and Fisher (1966) identified campesterol and fi-sitosterol in eggs fm hens given a diet supplemented with maize sterols. they mentioned that separation and determination of small ammmts of plytosterols in the presence of large amounts of cholesterol had only been achieved at that time by Kuksis and Huang (1962), who had analyzed the lymph of dogs which were fed plant sterol diets. Kritchevsky and Topper (1961) found approximately 20; of the choles- terol in eggs esterified, whereas the rest was present as the free sterol. It may be speculated that part of the other steroids which are present would possibly also be esterified. The exact interdependence Tell. (Eff I. III! E, 1 of steroids and proteins in egg yolk is not known. Parkinson (1966) reported on the relative distribution of cholesterol within various yolk protein fractions and that most of the lipids appeared to be bound in form of lipoproteins , but no specific interdependence was provided. These observations indicate that a method to survey the steroids in yolk should include a step to separate the lipids fran the proteins and that hydrolysis is needed if the esterified and free steroid frac- tions are to be combined. Hydrolysis and the subsequent extraction of the soap solution is also a convenient way to separate steroids from glycerides . To be able to isolate the steroids fran fatty acid con- taining material is especially important when working with eggs in view of their lipid canposition; 95$ glyceridic against only 14 steroidal material, (Parkinson, 1966). It must be mentioned, however, that some steroids are degraded during hydrolysis and also that some steroids are water soluble and, thus, are lost in the usual extraction procedure of the soap solution. These are, however, rather special groups of steroids for which specific separation procedures will be needed and, therefore, they are not considered in this study. ' To circumvent the possibility of breakdown during saponification the fat extract has been subjected to recrystallization procedures by redeli and Jacini (1971), thus achieving an enrichment of the steroidal fraction. 'me behavior of yolk extract in some organic solvents has been studied and is reported in the section on miscellaneous methods. Since this precaution is necessary only for a restricted group of can- pounds and makes the analysis for those steroids that do not degrade in the hydrolysis step more difficult due to the incomplete separation of glyceridic and steroidal material, it was thought that 10 recrystallizatim may be included in later investigations once the methodologr for sterols and the more resistant steroids has been worked out. Saponification can be carried out in either acidic or alkaline median. Enzymatic or acidic hydrolysis is applied to estrogen gluco- siduronates and sulfates, as reviewed by Adlercreutz and Imminen (1968). they showed, as did Eik-Nes (1968), that each method has advantages and disadvantages based on such things as instability of the canpound in slimline mdium or destruction of estrogens in presence of glucose during hot acid twdrolysis. No reference has been found which would indicate that sterols are prone to degradation in either acidic or slimline saponification and, thus, it was decided to use alkaline mdrolysis. Once the nonsaponifiable fraction containing most or all of the steroids is obtained, several methods of subdivision are imaginable. Conventional separation by liquid coltmm chromatograpmr was canpared to thin-layer chromatography (TIC) and it was found that with the latter, modifications could be made more rapidly and the results made visible quicker. This confirmed the finding of Avigan, et a1. (1963). High pressure liquid chromatography (EPIC) had to be discarded, also, since the only system available did not have a solvent programing capability and, thus, solvent gradients could not be run. Since the changeover from one fixed solvent system to the next with the available HPIC system required ample manipulations, such as solvent degassing, dis- mantling and reassembling of fittings, often exchange of the optical filter systmns, and rather lengthy equilibration with the new solvent system it was thought that with this type of instrument, no real 11 advantage over a thin-layer system could be expected. It must be added, though, that once a particular separation is worked out liquid column chrmatograpmr offers the advantage over TLC of innediate recovery of the fractions which are of particular interest. With respect to recrystallization of particular steroid fractions fran nonsaponifiable material the approximately 1000-fold predominance of cholesterol over any other steroid must be considered, and it was thought highly unlikely that by this means amr better separation of individual steroids could be achieved than with thin-layer chranato- graphy. The purity of recrystallized material was questioned by Wortman, et al. (1972). The behavior of yolk nonsaponifiable material in different organic solvents is described in the section on miscella- neous methods. Another method very often used to separate sterols from other non- saponifiable material, e.g. Dietschy and Siperstein (1967), is the precipitation of 3p -sterols with digitonin, as reported by Windaus (1%9). This method has been adapted to egg yolk nonsaponifiable material and is described in the section on digitonin precipitation. It became apparent that practically all material in the nonsaponifiable fraction can be precipitated with digitonin and , thus, no appreciable subfractionation is achieved. 'ihe cmercially available digitonin scans to introduce foreign material into the precipitated and the supernatant fractions and, thus, should be used. only after rigorous purification. A weak point of digitonin with respect to quantification is its varying affinity for different 3p -sterols , (Scthheimer and Dam, 1933, Emberg and Seher, 1972), such that sane of them may be in the precipitated as well as in the supernatant fraction and that even 12 some nonpsteroddal.compounds react with digitonin, (Anon., 1968, Honberg and Seher, 1972). For these reasons, extremely high percentage of’precipitable:material, requirement fer extensive purification of commercial.digitonin, uncertainty with respect to its specificity, the digitonin reaction can.not be recommended as direct workeup'procedure for egg yolk nonsaponifiable:materials Thin-layer chromatography of steroids has been reviewed‘by Lisboa (1%9) and Neher (1969). Based on an average madman nonsaponifiable ' fraction.of egg:yolk.of 300 mg, the most logical mode of separaohon seems to be the canpranise solution indicated by Neher, i.e. using plates with.thicker absorbent layers. For unstable compounds, care should.be taken with respect to exposure to air and light on.the relap tively large thinplayer surface.. Neher also pointed out that the sensitivity of detection is highest with.adsorption TLC and that for quantification this mode yields a.purer eluate than the partition chromatographic mode. Identification.of steroids, as shown in the two reviews above, can.be done on the plates directly, Based on the results of Tu, et al. (1970) it is apparent, however, that most of the steroids found in egg yolkLlie within a relatively'narrow'band close to cholesterol. ‘When the nonsaponifiables are applied at such.amounts on the thin-layer plates that minor constituents can.be visualized, the width of the cholesterol spot or band becanes so large that many of these components overlap with cholesterol and the application of specific color reactions is practically'impossible. For this reason, to identify egg yolk.non- saponifiable compounds which.mdgrate with or close to cholesterol on thinslayer plates, additional separation techniques which help to 13 further separate minor constituents from cholesterol are needed. mis scans to be done best by recovering the separated substances from the plates and to subject them to gas liquid chranatographic (Gm) analysis, (Gardiner, et al. 1%6). Spectroscopic methods were not considered to follow the thin-layer chranatographic separation since they usually require highly purified single component samples which could not be obtained by simple extrac- tion, saponification and thin-layer separation as demonstrated by GIG. Even the Iiebermann-Burckardt test should be applied with caution when more than 1 sterol is present since unsaturated sterols vary widely in their response to the reagents and saturated sterols do not react, (Bpary, 1%3). Since a gas chranatography-mass spectrometry (GO-LS) in- strument was available and it is being stressed that this combination is the most powerful tool available at this time for identification of steroidal. material, it was thought that thin-layer and gas chranatogramy should provide the separation necessary to be able to use mass spectros- etry as the final step for identification. This approach was used by Gardiner, et al. (1966) and was recamended by Homing (1968). Concluding this background on methodology it seems that a method to yield information with respect to egg steroid composition and its - possible variations should be based on the analysis of lipid extracts fran single eggs which are subjected to alkaline hydrolysis to remove glyuidic material. Separation of individual steroids from the bulk constituent of the nonsaponifiable fraction, cholesterol, and fran each other should be achieved by thin-layer and gas chromatographic techniques and final identification by means of a mass spectrometer, (Brooks, et al. 1968, 1973). DEVBIDPMENI‘ WW8 Extraction Procedure The following is the procedure which was developed for the lipid extraction of egg yolk: 1. 2. u. «‘5‘ L“) 5. fl 6. \” ¢ 3*”; 7' :8. \WL 9. 10. Tare weight of 200 :11 glass bottle wt. Tare weight of 500 ml flat-bottom boiling flask wt. 388 night wt- Crack egg shell, remove white with running water, cut off chalazae. Roll yolk on paper towel to remove adhering white. Transfer the whole yolk into the 200 ml glass bottle, weigh yolk wt. weigi the washed shell-halves after drying. wt. white wt. Add 60 ml 1:2 chloroform-methanol mixture , close bottle, shake vigorously, let stand 10 min. Add 20 ml chloroform, proceed as in 5. Add 1&0 ml chloroform, proceed as in 5. Add 30 ml methanol, close bottle, shake, prepare dtm of equal weight with 2nd glass bottle and centrifuge head. Centrifuge on I30, 10 min at 2,500 rm, bottles closed- Decant the solvents through c-fritted glassfilter filled 1/3 with anhydrous 11.2301, into the 500 m1 flat-bottm boiling flask. After solvents have passed, rinse wall of filter ‘ with small emmmt of 03013 . 1h 15 \r n?" 11. Add 50 ml methanol to the protein precipitate in the bottle, .109 close, shake vigorously for complete dispersion, add 50 ml. 011313, shake again, let sit 10 min. 12. Repeat 9. 13. Karat 10. 1h. Add its ml chloroform to the protein precipitate, shake vigorously, let sit 10 min, add 50 ml methanol, shake, repeat 9. and 10. 15. Repeat 1h. without waiting period. 16. Evaporate the solvents under slight vacumn on a rotary evaporator with intermediary condense-trap; water temp. its C, use dry Hz to break vacuum. 17. If extraction is followed by saponification, take the extract up in the appropriate amount of 95% ethanol; if continued by removal of non-lipid material, add appropriate amount of chloroform. With this method the total lipids from a single but entire egg yolk can be extracted as completely as possible by mixed-solvent tech- niques . Discussion of Extraction Procedures For those techniques of steroid analysis which did include lipid extraction it was mostly done based on either the method of Folch, et al. (1957) or the method of Bligh and Dyer (1959). The method of Folch, et al., developed for brain tissue, consisted of extracting the tissue by homogenizing it in a 20-fold 2:1 (v/v) chlo- roform-methanol dilution and filtering the homogenate. Mixing the ex- tract with two tenth its volume with water or salt solution yielded a biphasic systan without "fluff" (insoluble material accumulating at the interphase of the chloroform and water-methanol phases). The upper phase contained the non-lipid and the lower phase the lipid material. After several rinses of the interphase, they took the washed extract 16 to dryness without foaming. The method of Bligh and Dyer (1959) was developed for frozen fish, a tissue containing 80% water and 1% lipid. The method consisted essen- tially of an initial homogenization of the tissue in 3 volumes (v/w) of chloroform-methanol 1 :2, followed by homogenization in chloroform- methanol adjusted to 1:1, and a final hanogenization after addition of water to yield a chloroform-methanoldwater ratio of 2:2:l.8. Then the hanogenate was filtered and left for phase separation. A variation of their own procedure included filtration after the first homogenization, followed only by chloroform rinses and phase separation by adding water to the combined filtrate. In this case, however, the separation was reported to be rather slow but still satisfactory. The amount of non- lipid material in the chloroform layer was below the sensitivity of several other methods and the lipids in the methanol-water layer only about 1% of the total lipids. Entenmann (1961) indicated that often a more satisfactory sub- division is obtained when the tissue is first blended with methanol or chloroform-methanol 1:1, instead of chloroform-methanol 2:1 or chloro- form. He pointed out that the lipids tend to be protected by the solvent vapors during drying, but recommended to vent the flask with nitmgen to prevent oxidative changes in the lipids. Galanos and Kapoulas (1965) observed interfacial "fluff" of protec- lipids during the washing procedure when extracting lipid fractions from whole milk with a small volume of chloroform-methanol 2 :1, whereas "fluff" never formed when chloroform-methanol 1:1 or chloroform-ethanol 1:1 or 2:1 solutions were washed with one fifth volumes of saline solu- tim nor when chloroform-methanol 2 :1 solutions were washed with one 17 tenth volumes of saline solution. they reported that this "fluff" occurred only when the washing was carried out on a concentrated protec- lipdd solution in chloroform, whereas it did not occur when the washing produced lower phases with chloroform-alcohol ratios of approximately 2:1 or higher. lhey also stressed that all solvents used in the fat extraction should be neutral, especially alcohols . To remove nonlipid contaminants from chloroform-methanol solutions they recommended that chloroform be added to obtain a solution containing less than 15% metha- nol. They found much higher lipid losses with the aqueous washing pro- cedures recanmended by Eolch, et a1. (1957) than were previously re- ported. Johnson (1971) indicated that for lipid extraction in liquid samples such as yolk, it is not necessary to honogenise and that such mixtures need only to be shaken or left standing for some time. To adjust for the required 805 water content of the sample in the Bligh and Dyer method he added 214 sodim chloride solution. Considering that an egg yolk of 20 g contains approximately 10 grams of water and 8 g of fat it requires a dilution with saline to 50 ml to yield a sample containing 80% water or to 800 ml to yield a sample with 1% lipids, which would be equivalent to the starting prod- ucts of the Bligh and Dyer method. Following the Bligh and Dyer (1959) method either quantity should be shaken up with 3 volumes of 1:2 chloroform-methanol mixture to which additional voltnes of chloroform and water should be added. 'niese solvent quantities indicate that the phase separation cannot be aided by using any of the readin available centrifuge equipments . Using the Folch, et al. (1957) or the Bligh and Dyer (1959) methods, considerable amounts of interfacial "fluff" appeared 18 and did not disappear in a reasonable length of time. Saponification is the succeeding step in this procedure for analysis of egg steroids and it seems that the removal of non-lipid and polar material could be dale in a single step, i.e. after saponification. 6334M Observations When yolk is placed in a container, it has a tendency to stick to the walls and bottom, thus making extractions or transfers rather dif- ficult. For some time butylated hydroxytoluene (BET) was dissolved in the chlorofom used for extraction at 1 mg per 100 ml but this precaution was later dropped after it was realized that egg yolk contains sufficient natural antioxidants of its own, (Brooks and Taylor, 1955). The use of Waring blendors to homogenize yolk in solvents can not be recannended since it is difficult to transfer yolk quantitatively into the blender and to pour the homogenate into filters without loss of material. Blenders should be equipped with gaskets which withstand the action of chlorofom-methanol mixtures . When water and solvents were added slowly to the yolk during hanog- enization a fine dispersion of proteinaceous material was obtained. The finer the dispersion, however, the more difficult the filtration. Screw-caps of centrifuge bottles should be lined with Teflon. The use of aluminum foil or other linings may result in contamination or even degradation of the sample by these materials. When centrifuging the yolk-organic solvent suspension, the protein precipitate often accumulated as a pellet at an intermediary level and recovery of the bottom solution was rather inconvenient . With l9 chloroform-methanol ratios greater than 1 the protein precipitate usual- ly did not settle, whereas at ratios smaller than 1 it settled, but filtration of the supernatant is recon-handed to retain loose particles. Addition of water or 214 saline solution to the yolk, as recanmended by Johnson (1971), always resulted in a two-phasic system and, thus, made recovery of the lipids more difficult. It is not clear why water should already be added for the extraction and not only later for the phase separation of the lipid extracts. In the beginning Whatman No. 1 filter paper was used which showed a sufficient flow-through rate. It was, however, replaced by C-fritted glass-filters because it happened often that a small hole developed at the point of the cone and canplete recovery of lipid material hm filter paper is rather difficult, whereas a fritted glass-filter is easy to rinse. Lengtlnr investigation into the behavior of chloroform-methanol- water mixtures with respect to phase separation in separatory funnels showed that even with pure solvents there were many solvent canbinations which formed interfac ial "fluff" , and with either the chloroform or the methanol layer or both remaining turbid or cloudy. Thus, for the Bligh and Dyer (1959) cmbination of 2:2 :l.8 , chloroform-methanol-water, inter- facial "fluff" developed and both phases remained turbid. 'me ratio of the Folch, et al. (1957) method was given by them as 8:148, chloroform- methanol-water. After shaking this combination in a separatory funnel, the phases separated without visible "fluff" but renained turbid. In both cases, many chloroform droplets remained on the walls of the sepa- ratory funnel, above the liquid phase, and also on top of the water- methanol layer which would cause a loss of lipids when withdrawing and 2O continuing only with the chloroform layer. This latter problem could be overcome by centrifugation. Neutralizing the pH of the methanol and the water with either acetic acid or aqueous 1m, did not visibly influence the clearing of the solvent phases nor the appearance of interfacial "fluff". Acidity- ing the water and/ or the methanol seemed to improve the interphase some- times. Addition of methanol to the already separated phases usually resulted in some clearing of the top of the water-methanol layer, leav- ing a foggy lower water-methanol phase or an interfacial "fluff", where- as adding chloroform usually caused both phases to turn millw. Warming of the separatory funnel and careful swirling improved the clearing of the methanol and/ or chloroform phases and eliminated inter- facial "fluff". A faintly turbid methanol phase could often be cleared by tilting the separatory funnel and slight rocking. When sane of the solvent combinations which cleared up without interfacial "fluff" were used to extract egg yolk, they all showed interfacial "fluff" . W of the Extraction Procedure In the procedure develOped in this study the extraction of the yolk lipids was done in a single 200 ml centrifuge bottle with a Teflon lined screw-cap. To get the yolk canpletely into the centrifuge bottle, it was rolled to the edge of the paper towel, which was then. wrapped around it and thus allowed to hold the yolk directly above the opening of the bottle. The yolk sac was then slit quickly with a sharp edge such that the yolk liquid dropped freely into the bottle. After most of the liquid had drained, the yolk sac was pulled down from the paper and 21 dropped also into the bottle. With sme exercise, a cmnplete yolk could be transferred into the centrifuge bottle without spilling yolk on either the paper or the outside of the bottle or the slitting edge. The solvent sequence was selected to' result in a fine suspension of proteins, thus, improving the extractability. The total solvent anoint was close to the recommended 20-fold volume of the Folch, et al. (1957) method but the repeated extractions insured that no lipid residue remained in the protein precipitate. The last extraction step could possibly be left out since it did not seem to produce additional lipids. It is included in the procedure as a safety step, in case extremely large yolks should be investigated. In the latter case a two-phasic system sometimes developed in the first extraction. ihen, only the supernatant was removed as indicated and the procedure continued with the second extraction. Since practi- cally all the water was removed with the first supernatant, the second extract was monophasic and the procedure could be continued in the usual way. After the first centrifugation, the yolk sac floated quite often intact in the supernatant solvents. This could indicate that the chloroform-methanol solvent mixture is quite suitable to remove lipids while leaving the proteinaceous structure intact, which may be of inter- est when studying membrane proteins. The extract was poured into a 60 m coarse-frittad glass-filter to retain precipitate and for removal of water. lhe ,canpleteness of the removal of solvents on the rotary evaporator was improved in the following way: Shortly before frothing of the lipid extract, the solution became rather viscous and turned from a dark yellow _ 22 color to light yellow. At this point, a small flow of nitrogen was leaked into the system, the speed of rotation of the flask was in- creased, and the water-flow of the ejector increased to incite frothing. As soon as the frothing began, the water flow was adjusted to cause only mall bubbles to form in the extract. with the nitrogen, the rota- tion, and the water-flow thus adjusted, the solvent removal could be continued as long as desired. Its progress was followed by observing the dripping of solvent from the condenser coil. In this analysis, with subsequent saponification, one half to one hour of operation under these conditions was found to be desirable and sufficient. At this point, the extract still contained solvents but its total weight was reduced to approximately 50% of the yolk wet weight. A method to remove non-lipid material from the extract was also developed. It can be done in the following way: After reducing the extract to the volune when frothing starts, it should be transferred quantitatively with chloroform to another 200 ml screw-cap centrifuge bottle. Following the recommendations of Galanos and Kapoulas (1965), chlo- roform should be added until the estimated methanol content represents less than 15%. For a 20 g yolk this was achieved when diluting the transferred extract to 100 ml with chloroform. At this point, the washing procedure of Folch, et al. (1957), using 0.29% NaCl can be applied. It is also possible to follow the procedure of Folch, et a1. (1957) entirely, by using for the transfer and the dilution a. 2:1 chloroform- methanol mixture or to follow Bligh and Dyer (1959), using chloroform- nethanol 1:1. 23 After phase separation by centrifugation, the water-methanol phase should be removed with an aspirator, the interphase washed according to Folch, et al. (1957), and the chloroform layer passed through a C-fritted glass-filter with anhydrous sodium sulphate into a 500 ml flat-bottom, tared boiling flask while keeping the material of the interphase in the centrifuge bottle. This material should be resuSpended in 25 ml chloro- form whioh should then be poured onto the filter. After a few rinses of the bottle and the filter with chloroform, the lipid extract should be concentrated on the rotary evaporator. The lipid extraction procedure was checked for completeness by subjecting the protein residue to method 16.008 of the AOAC (1965) for the determination of fat in dried eggs and by exhaustive ether extrac- tion on a Goldfisch apparatus. In either case, less then 0.5% of the fresh yolk weight was recovered. The method was also checked for oxidative effects by applying the extraction procedure to 80 mg of pyrogallol in 1&0 ml methanol without an other material. 0n drying, the final step of the extraction proce- dure, the pyrogallol crystallized onto the walls of the flask without any indication of oxidation. The method described for the extraction of total lipids from eggs was deve10ped for individual whole eggs. By using 200 ml centri- fuge bottles the extraction can be performed without loss of yolk material. The solvent ratios were selected for efficient extraction and so that the protein can be pelletized at the bottom of the bottle by centrifugation. Thus, the extract can be easily recovered and trans- fers of yolk material are eliminated. Since all extractions can be carried out in the same bottle, loss of extract can be prevented and ah the method made quantitative. The sample material canes only in contact with glass or {reflux surfaces, thus contamination can be avoided. Flushed-off particulate matter is retained on a filter which serves todrytheextrnct. monothodtopreventemcessivefrothing when drying the extract was devised and the glassware was chosen such that the extraction can be followed imediately by saponification. Although not considered essential for the analysis of egg steroids, a convenient method for the removal of non-lipid material fran yolk extract was developed. The technique to canpletely dry yolk lipid extract is described in the section on recrystallization of yolk extract and yolk nonsaponifiable material. 2'5 Saponification Procedure The AOAC (1%5) lists, under 16.015, a method to separate nonsaponi- fiable matter fran eggs or under 26.071 to determine the nonsaponifiable residue from oils, fats, and waxes. Method 16.015 prescribes 3 hours of heating on a steam bath to disintegrate the lumps formed by the proteins with the possibility of thermal degradation due to the long heat treatment. To reduce thermal degradation as much as possible it was decided to precede the saponification by a yolk fat extraction and, therefore, method 26.071 was considered to be more appropriate for saponification of the yolk lipids. However, in either method the amount of sample is only 2 to 2.5 g. Therefore, method 26.071 was adapted to the total amount of fat extracted from an entire egg yolk. The following is the procedure which was adapted for the saponi- fication of yolk lipid extract from method 26.071 of the AOAC (1965). 1. Yolk weight (from extraction procedure) wt. 250 ml flat-bottan boiling flask precision tare wt. 2. Roughly estimate the fat content of yolk as 10% of the liquid yolk weight. Per g of fat add 5 ml 95% ethanol and 100 mg pyrogallol to the 500 ml boiling flask con- taining the lipids extract, and, in a separate beaker, dissolve 0.5 g KOH (0.59 g 85% KGB-pellets) in a minimum of distilled water. 3. Install the boiling flask with a short, water-cooled con- denser on a steam-bath and bring to reflux. Once a con- densatim ring is well visible in the condenser, the ROE-solution is added very slowly with a long Pasteur pipette along the walls of the condenser, beneath the condensation ring. h. After completing the addition of KOH, the top of the condenser is blanketed with a slow stream of nitrogen and refluxing continued for 30 min. (N0 browning should occur.) 5. 10. 13. 1h. 26 For dilution and transfer of the soap solution, prepare distilled water and ether h times the volume of the ethanol used in 1. Half of the quantities of water and other are already poured into a 500 ml separatory funnel A. A second 500 ml separatory funnel B is set up containing 50 ml distilled water. A small funnel is used on each separatory funnel. (The sizes of glassware are based on the quantity of lipid extract from an average yolk and should be adjusted when significantly larger or smaller fat quantities are saponified.) After the saponification has proceeded for 30 min, the nitrogen is st0ppered firmly onto the condenser, the boiling flask and condenser removed from the steam bath and dipped into an ice-bath. During the short cooling the flask is continuously swirled. (Browning of the soap solution will occur but should not be excessive.) Pour the lukewarm soap solution into A and use first the remaining water, then the ether for a quantitative trans- fer by rinsing repeatedhr the saponification flask and condenser very carefully. Swirl the separatory funnel, close it, shake vigorously, and let phases separate completely (no fluffiness around interface). Drain lower alcoholic layer into a 1400 ml beaker with a good pouring tip and pour ether phase through the top into B. Rinse pouring edge with ether into B. Transfer beaker content back into A, rinse with same quan- gity of ether as used for the first extraction and repeat . and 9. Repeat extractions with 50 m1 aliquots of ether until the ether layer is no more yellowish (when pyrogallol is used, it remains faintly brown, without it it becanes completely colorless). After all ether extracts are combined in B, close it, shake vigorously and let phases separate capletely. (file volume of the water layer should not increase more than approximately 205.) Drain aqueous layer. Wash twice more with additional portions of 50 ml distilled water, shaking vigorously each time. Wash ether solution It times with alternate 50 ml portions 0.51! aq. K03 and water, shaking vigorously each time. (In the first alkali wash an emulsion forms which usually breaks up within 1; an hour, drain only well-sepa- rated part of water-phase.) I' I! ll: Ill].lll|l Ill al.-III 1" III (I -5: 27 15. After the last KGB-treatment, wash ether solution suc- cessively with 50 ml aliquots of distilled water until washings are no longer alkaline to phenolphtalein. (It takes about 6 to 7 washes.) 16. Transfer ether layer quantitatively into the 250 ml precision-tared flat-bottom flask through a 60 mm C-fritted glass-filter filled 1/3 with anhydrous NagSOu. l7. Evaporate the ether under slight vacuum on a rotary evaporator with intermediary condense-trap; water temp. 35 C; use dry N2 to break the vacuum. 18. Dry the nonsaponifiable material completely in a vacuum oven at 145 C for % hour, break the vacuum with N2, and let the flask cool for exactly é- hour in a desiccator, then weigh. wt. nonsaponifiable material wt. 19. For storage add appropriate amount of chloroform, 1 ml per 25 mg. Store in freezer. With this method the total nonsaponifiable material of a single but entire egg yolk can be recovered without oxidative degradation. Elanation of the Saponification Procedure Direct saponification of an entire egg yolk is difficult since it was found virtually impossible to canpletely introduce a yolk into one of the ncrmally used boiling flasks due to their narrow necks. 0n start- ing to reflux the yolk in the alcoholic potassium hydrondde solution, the proteins lumped together and it was difficult to assess the efficiency of saponification under such conditions. This problem could be allevi- ated, sanewhat, by shaking immediately after the egg was added to the alcoholic KOH solution. Whenever direct saponification was done, the soap solution was filtered before being passed into the separatory funnel to eliminate the protein precipitate and, thereby, preventing the separatory tunnel from being clogged. It is recommended that a water-cooled condenser be used instead of I...“ {slid-11.5. "1" w 28 the reflux air condenser of the AOAC method, since the air condenser can become quite warm and, thus, alcohol may be lost during refluxing. 'me use of glass-beads during saponification can not be recommended since it was difficult to retain them in the boiling flask, and once they were in the separatory funnel they interfered with the withdrawal of phases. The irregularly shaped boiling chips are better, but they were not necessary. Since the AOAC saponification method uses 25 ml alcohol and 1.5 ml son solution (3+2) for 2.5 g fat it was thought that tripling these quantities would be correct for the lipid extract of a 20 g yolk based on approximately 37.5% y'olk lipids. when mixing 75 ml ethanol and 15.5 ml K011 solution (3+2) with 150 ml water and 150 ml ether, draining the water-ethanol layer, reextracting it two more times with 150 ml ether the following phase volumes were measured: lst separation water-phase 205 ml, ether-phase 163 m1 2nd separation water-phase 18h m1, ether-phase 167 m1 3rd separation water-phase 172 ml, ether-phase 160 ml These values show an absorption of ether into the water and vice-versa with the ether layers gaining regularly. When the procedure was contin- ued, however, by washing the combined ether extracts with 60 ml aliquots of water, the water phases successively measured 8h, 80 and 75 ml. This would probably be acceptable since during the succeeding alternate alkali and water washes the volume of the water-layer gets closer and closer to a limit volume of 61+ ml. However, when the procedure was applied to the soap solution resulting from yolk lipid extract, the first water phase resulting after washing the combined ether extracts measured 15h ml instead of the above 8h ml. This discrepancy and the frequent finding 29 that it was impossible to canpletely dry the nonsaponifiable material on the rotary evaporator prompted a search for the necessary modifica- tions of the Adm-method to obtain a proper saponification procedure for the em yolk lipid extract. First it was established that boiling approxintely 3 g of lipid extract for half an hour with 25 ml 11! ethanolic ms yielded a complete saponification, as would be expected, based on the Ame-method. For 6 g of lipid extract the same conditions were insufficient, but when 50 m1 1!! or 25 ml 21! ethanolic ROB-solution were used, the saponifica- tion was again complete. Further experimentation showed that 0.5 g mm per g of lipid yielded a complete saponification when refluxed for 15 to 30 minutes. When checking the saponification for oxidation by using pyrogallol, it was seen that the potassius hydroxide should be added to the reflux- ing solution, as indicated in the vitamin E assay, Bunnel (1967) and in form of an aqueous solution. When a m-pellet fell into a refluxing pyrogallol solution, some brown-black substance fell out which was derived frm the oxidation of pyrogallol. This way of adding the Km may also cause other canpounds to oxidize and, therefore, the concen- trated ROB-solution should be added very slowly with a long Pasteur pipette to the wall of the water-condenser, below the condensation ring. Adding K08 in this way is advantageous, since no air is dragged into the refluxing solution, and the ROB-solution, while running down the hot walls of the lower part of the condenser and the boiling flask, is diluted and degassed. For complete saponification it was essential to add enough ethanol and to prevent loss of ethanol during refluxing, thus avoiding KOH 30 crystallisation. 0n the other hand, the use of emcessive quantities of ethanol seems undesirable since this would require large ether quanti- ties. A quite acceptable cmpromise was established with the use of 5mlethsmolpergofyolkfat. ’ Since the question of oxidation dining saponification was raised in the literature quite often a means was sought to prevent oxidation. Antioxidants readily available were butylated hydroxyanisole (BHA), M, and pyrogallol and, therefore, their suitability was evaluated. BHA and Mranainedliquidchlringthedryingstepsontheroteryevaporatorand in the new even. This seemed undesirable because then the endpoint of the drying process is uncertain and no indicatim for the canplete- ness of the saponification procedure can be obtained. Pyrogallol, on the otherhand, semedtobequite suitableandhas alsobeenusedin the vitamin E assay, (Bunnel, 1%7). 'nleuse ofcmlleOmgofpyrogallolpergoffatwasadoptedsince itwas seenthatwhenusing5$ (w/v) pyrogallolinethanol, as inthe vitamin E assay, the refluxing solution was very viscous, such that pyrogallol my even interfere with the saponification itself. Brooks an! Tyler (1955) reported that sane natural antioxidants such as toco- pherols are found in the nonsaponifiable fraction of egg yolk, thus allowing for a corresponding reduction in pyrogallol. Succeeding ether extractions were easier to perform when less pyrogallol was used. Pyrogallol is an extremely valuable indicator of oxidaticm since non-oxidised it showed no color in the solvents, whereas oxidized it turned to brown-black at alkaline 11!, and to yellow-brown in acidic medi- lsn. Whenthe KDHsolutionwas addedtoo rapidlytothe refluxing fat so- lution, or when the saponification was performed improperly, the flask 31 content turned brown and indicated that oxidative conditions prevailed. 0n careful addition of the potassiua.hydroxide and preper saponification, the refluxing could be carried on for hours without any brown color development. After the refluxing was stopped and the soap solution was trans- ferred to the separatory funnel the water turned brown-black due to the oxidized pyrogallol. This could be expected since pyrogallol undergoes autoxidation in aqueous solution in presence of finely distributed metals, (Grimm, 1967). It can be seemed that this phenomenon is a peculiarity of pyrogallol and is not shared by steroids sensitive to oxidation and.which are protected by pyrogallol during saponification. Instead of glass stoppers and glass stop-cocks on separatory fun- nels, which leak quite often, Teflon stoppers and stopmcocks are recomr mended to avoid losses. Also in this context, it is a good practice to withdraw part of the water-phases first until the flow practically stops before opening the stopper on the separatory funnel. By preceding in this sequence, a slight vacuum is created within the funnel and when the stopper is removed, the liquid accumulated around the stopper is drawn into the funnel instead of being blown out. The extraction of the soap solution without lipid extract in which there is a 2-fold dilution with water and its extraction with 2 volumes of ether as in the.AOAC (1965) method resulted in acceptable phase vol- umes, whereas in the presence of yolk lipid extract the volume of the water wash of the combined ether extracts was excessively large. When using a h-fold volume of water fer the dilution of the soap solution and extracting with h volumes of ether, all phase volumes appeared.normal and these relationships were retained for the procedure. It can be 32 speculated that the water is necessary to dilute the alcdhol and that the abnormal behavior Observed with the yolk lipid extract may be due to methanol-carried over from the extraction procedure. In the earlier German literature on saponification, (Thaysen, 191h and Fax, 1920), the saponification flask, after transfer of the hot soap solution to the separatory funnel, was rinsed first with ether and the separatory funnel shaken vigorously without any water dilution. Only thereafter was water added. The newer sequence of the AOAC-method was used here since time did not allow further investigation of this point. Collander (l9h9) measured a distribution coefficient of 1.7 between ether and water fer pyrogallol, thus indicating that it provides protec- tion against oxidation in both solvents. Once the nonsaponifiable com- pounds are extracted into the ether, they are protected there by the reducing action of ether, as mentioned by Collander (1999). Therefore, no other lipid solvent was used for the extraction of the nonsaponifiable material from the soap solution. ‘Hhen the combined ether extracts were washed with alternate aliquots of 0.5N’XOH’and water in step 1h of the procedure, quite strong interfacial emulsions occurred initially, which decreased and finally disappeared with succeeding washes. This hetero- philic material is eliminated from.the nonsaponifiable material by these washings since no such emulsions occurred when saponified.material was resaponified. When drying on a Buchi rotary evaporator it was observed that part of the evaporated solvents condensed inside the steam duct and ran back into the sample flask and when the vacuum was broken, many of the con- densed droplets were blown into the flask. Since the steam ducts of the evaporators are usually not cleaned, this "reflux" of solvents may 33 entrain one material foreign to the sample. For this reason, a Kontes evaporator trap was always used, since it could be cleaned like any other glass-ware and the vacuua could be broken on the trap such that the solvents were not blown into the sample but rather into the con- densing part of the apparatus. To measure the weight of nonsaponifiables and, in fact, of any other subtraction precisely it was necessary to obtain the tare weight of the flask under conditions identical to those used to dry the sample, 1 .e. the acid-washed flask was rinsed with 2:1 chloroform-methanol and ether, the outside was wiped with a moist Kimwipe, then the flask was placed in the vacuun oven and dried for one half hour at usual drying temperature. Thereafter, it was transferred to the desiccator to cool to ambient temperature and the tare measured after exactly one half hour. After the sample was transferred to the flask and dried on either the rotary evaporator or under nitrogen, it was vacuum dried and weighed at the same temperatures and with the same time schedule. The procedure to saponify the lipid extract from an entire yolk was adapted fran method 26.071, (AOAC, 1965). It is based on a preliminary extraction to reduce refluxing times and to minimize thermal degradation. By substituting the air condenser with a water condenser, losses of ethanol and overconcentration of the m1 solution could be prevented. A proper way of adding K03 was devised and the use of pyrogallol and of a nitrogen blanket was adopted to minimize oxidation. The amount of pyrogallol was adjusted to allow for efficient saponification and a quantity of ethanol selected to prevent overconcentration of K03 and to keep ether volumes minimal. The dilution of the soap solution was adapted to yolk lipid extract to avoid losses of nonsaponifiable 3h material. The rotary evaporator was equipped with an evaporator trap to prevent contamination and the weighing procedure standardized to ensure precise weighing. Digitonin Precipitation Procedure 9. 10. The following is the method which was developed for the precipita- of 3-beta-hydroxysteroids with digitonin. Vial precision tare wt. = Dissolve the sample in 95% ethanol (150 ml/g sample) in centrifuge tube A, (1150 ml/g sample). Prepare a 0.5% digitonin in 90% ethanol solution (1000 mi/g sample) in a beaker with pointed tip. After warming of the contents of A and the beaker in 80 0 water, pour beaker content imediately into A. let cool slowly. After overnight standing in cooler, flush walls with sane 90% ethanol, centrifuge 10 min at 2,500 rpa. Transfer supernatant into 100 ml round-bottom flask. Reduce to small volume on flash evaporator with inter- mediary condense-trap; water bath at 60 C. Add 90% ethanol (”00 ml/g sample) to A, bring precipitate into suspension by swirling and sonication, flush walls with small amount of 90% ethanol, centrifuge 10 min at 2.500 rin- Repeat 5. Transfer the cmcentrated supernatant into a 2nd centrifuge tube, B dilute with at least 2 volumes of water (total volune i:00 nl/g sample). Repeat 6 three times with ether, transferring the supernatant each time to B for extraction of the water mase. Mix thoroughly on vortex mixer, centrifuge 5 min at 2,500 rpm. bansfer supernatant ether phase into 100 ml round-bottm flask through Routes F-fritted glass filter filled to 4; with anhydrous “3250!;- 35 ll. Concentrate the combined ether phases to small volume, under liglt vacuum, on flash evaporator with interme- diary condense-trap; water bath at 35 c; use dry N2 to break the vacuum 12. Dry digitonides in A very carefully with dry Na, then invacumnovenforfihourat 1&5 C. 13. Transfer the ether concentrate quantitatively into the small precision-taxed vial, dry as in 12. 1h. After cooling for % hour in desiccator, weigh the supernatant fraction. vial wt. == supernatant wt. 8 15. Take supernatant up in pyridine or 10% MeOH in benzene, l ml/25 mg. Dissolve the digitonides in pyridine (60 ml/g sample). With this method 3-beta-hydroxysteroids can be almost quantitatively precipitated from nonsaponifiable material or individual TIC-fractions therefrom. Since commercial digitonin released contaminants when used in above precipitation procedure, a method was developed to purify the digitonin and is described in the section on materials. Review of Digtonin Precipitation Precipitation of cholesterol with digitonin was first reported by Hindaus (1909) and has since been applied to separate sterol fractions from: nonsaponifiable material. The method consisted of pouring togeth- er the hot solutions of 1 g of digitonin in 100 ml 90% ethanol and 0.h g of cholesterol in 60 ml 95% ethanol, filtering after 1 hour of standing, washing with ethanol, and drying. lhe digitonides were recrystallized from 120 ml boiling methanol by careful addition of sane water. This basic method was modified by 'Ihaysen (19110 to yield a quanti- tative analysis by using a reaction time of 2 hours and an excess digitonin of 1.5 to 2.5 per mill over the theoretically required amount. 36 He showed that digitonin could be removed from the supernatant solution by adding water to the concentrated filtrate and.extracting it with ether or petroleum ether, thus leaving the digitonin in the water phase while extracting the lipids into the ether; This extraction step'was. also included in the procedure used by Boughton and wheetley (1959) who subjected.the supernatant material recovered from.the first reaction to a second precipitation. Fex (1920) extended the reaction time to 12 hours and increased the digitonin excess to 1.75 percent, and.Windaus (1925) reported the need.for purification of commercial digitonin and the procedure worked out by Schah in which ether was added in excess to a cold 5% aqueous digitonin solution. The precipitate was vacuum-dried and the procedure repeated once or twice. Windaus confirmed that the recrystallization of digitonin fran 85% ethanol, as reported by Kiliani (1918), yielded beautiful crystals. Damr(l928) found that cholesterol-digitonide was very'hygroscopic and could absorb up to 6% of'moisture from the surroundings, that alcdhol washes of the digitonides caused losses of up to 2%, whereas ether washes did.not cause such losses. He recommended the use of a much higher digitonin excess and, by using standard curves, to establish corrected cholesterol values. Applying this precipitation method to sterols such as c0prostenol, cholesterol and cholestanol, Schdnheimer and Dam.(l933) showed that digitonin precipitated individual sterols to varying degrees. Thus, they concluded that quantitation of unknown steroid mixtures based on the precipitate alone was rather unreliable. In another modification of the digitonin precipitation method 3? Butt, et al. (191+8) precipitated digitonin and digitonides from the alcohol solution with ether and recovered the steroids by centrifugation. They, too, washed the combined ether extracts with water. Since the Windaus digitonin method and its various modifications by other investigators was not quantitative, Sperry (1963) developed a method in which the sterols were dissolved in acetone-ethanol 1:1 and did not require heating or stirring. In his procedure the precipitate was washed 3 times each.with 80% ethanol and ether but he stressed that the method did.not attempt to remove all digitonin from the digitonides which, therefore, did not allow for gravimetric determination of the reacted sterols. The AOAC methods (1965) are inconsistent in the use of a particular procedure, since under 13.131 Sperry's (1963) solvents were used and under 26.062 those of the method of‘Windaus (1909). Homberg and Seher (1972) reported that quantitative isolation of sterols required h times the theoretical amount of digitonin and that the excess digitonin could not be removed completely from the digito- nides. For lanosterol alone they Observed no precipitate with digitonin, whereas in a 1:1 mixture with cholesterol, lanosterol was partly co- precipitated. Their procedure gave quantitative results only when a concentration of approximately 1 mg sterol per ml solution, a great excess of digitonin, and only small quantities of alcdhol for washing were used. 38 Wtion of the Digitonin Procedure Sch8nheimer and Dam (1933) suggested that all commercial digitonin preparations contained digitonides to varying degrees which are diffi- cult to separate and Homberg and Seher (1972) found considerable quan- tities of polar and nonpolar artefacts in the steroid fraction when isolated by the digitonin precipitation method. In this study contam- inants were found in the digitonin supernatant and, to a lesser degree, in the fraction recovered from the digitonides. Therefore, it is recom- mended that the commercial digitonin be purified before its use in the precipitation procedure. A method of purification was developed so that the digitonin, when used in the precipitation procedure, would not give off contaminants to either the supernatant or the later recov- ered precipitated material. It was not intended nor necessary to pro- duce pure digitonin. Therefore, the purification procedure followed very closely the actual digitonin precipitation and steroid recovery procedures without a real sample. The purification of digitonin by precipitation from aqueous solutions was found to require too much ether in relation to the amount of digitonin treated and was, therefore, not adopted. For purification the digitonin was precipitated from 90% ethanol with ether. Ether had to be added slowly to yield a precipitate Which could be centrifuged down easily. Digitonin is not easily dissolved in ethanol and, therefore, it usually had to be heated and/or sonicated. An attempt was also made to work on more than 100 mg at a time by using 200 ml centrifuge tubes. It was found, however, that with these flat-bottom centrifuge tubes some digitonin remained suspended and would be lost with the supernatant, whereas in the conical 100 ml centrifuge tubes the digitonin was 39 pelleted and remained in the cone. With respect to solubilities the infonmation in Table l was accu- mmlated from various sources. It proved to be useful for recrystalli- zation studies, in the attempt to eliminate digitonin from.supernatant material, and to find solvents in which the supernatant material was soluble. This was usually accomplished by either using pyridine or 10% methanol in benzene. The digitonin precipitation procedure was adapted from Windaus (1909) and takes into account some of the modifications proposed by later investigators. The acetone-ethanol solvent system of Sperry (1963) was not used since,for steroid analysis,the sample material is subjected to a far more rigid heat treatment during saponification than is required during the short warming period under step h of the preci- pitation procedure. Since Sperry (1963) reported the precipitation of cholestanol by his method, it was verified that it could also be preci- pitated in the solvents of the Windaus (1909) method and that neither cholestanol nor digitonin alone precipitated out of solution. Although chloroform and ethanol show similar solubility behavior toward digitonin, digitonides, and cholesterol, (Table 1), an attempt to perform.the digitonin precipitation in chloroform was unsuccessful. The quantity of digitonin used for precipitation is 1:5 (wzw) sample to digitonin.and.yielded a nearly quantitative precipitation of cholesterol. ‘When the digitonin precipitation procedure was applied directly to the nonsaponifiable material from egg yolk, more than 99% of the material was precipitated and yielded a supernatant which still contained predominantly cholesterol. This low purification - separation result was one of the reasons why the digitonin precipitation was not ho mefiHOm odosaom mamcwaomm mHooHOm hho> esoeaom easemaae oaobaow manodom oaooaom huo> oaoodom mandaom oaooaoe saohmaae oahsaoe aapnmaae nonmaOOmeo mopoaOOmmwe onoofloqu oHQoHOmoa oaoodomow oHooHOmoa oHodHOmoH oHnSHomoa oaooaoo_saoomaae aso> magaom passage mahoaomoa Honopmmaonu oeaeopamao-aohoomoaoeo oHpoHOm mHooHOmc« oaooaoefi oaooaomofi oHooHOmoa manoHOm oHooHOm bHfiwmo manoHOm hawowsoAm mandaom oaeoaom scameommdm bacon nasceaMHn overflow ooaoasao coaxooflsoaaoeeaao nonpo aboaoapom totem econoo< weapons Hampm oooucom auomoaOHno Hue mapmo.aoeoepm acetone see: .mhe .aooeepm Honorees hopm3.EHo3 ameo3_odoo pao>Hom .mpso>Hou msoaho> ma Hoaopmoaono one .onfioovwmaoudoaopmoaono .oHnOpfiwae mo mofipwdwooaom .H manna hl considered suitable for subfractionation of egg yolk nonsaponifiable material. The procedure should rather be applied to the material recovered fran sane specific thin-layer bends. As already mentioned, several investigators have found that quantitative precipitation of sterols depended on a large excess of digitonin and (:1 allowing the precipitation to proceed for a long time. With respect to the latter, it was found that overnight standing was more efficient in removing cholesterol frcn nonsaponifiable material than two consecutive short contact periods. Since quantitative removal of a particular canpound is not always necessary, the 2 parameters, relative mount of digitonin and reaction time, should be adapted to take advantage of differences in affinity andspeed of canplexing ofthe various canpounds inamixture. Itwas advantageous, in the presence of large amounts of cholesterol, to use repetitive precipitation with a digitonin amount less than theoretically required, suchas 1:2 (wzw), and, thereby, to simultaneously reduce the cholesterol content while retaining less reactive compounds in the supernatant which otherwise would have co-precipitated. The use of less than theoretically required amounts of digitonin has the additional. advantage of reducing the digitonin concentration in the supernatant. Adding 1 or 2 drops of water to the centrifuge tube usually indi- cated whether or not additional precipitate could be formed. Since the precipitated material, after centrifugation, did not represent a solid pellet, supernatants were best transferred by pipetting. Here, again, flat-bottcn centrifuge tubes were not well suited and it is recs-ended that the sample size be limited to the corresponding available conical centrifuge tubes. ’42 To circumvent the problem caused by centrifuging in flat-bottan tubes, the supernatant and washings of the digitonides were filtered. However, the material deposited in the transfer pipette and on the walls of the filter were extremely difficult to recover, and the digi- tonides tended to clogg the filter, thus filtration is not recommended. When the precipitation method was checked quantitatively, a loss of digitonin was traced to the residue in the beaker used in step 3. Depending on the configuration of this beaker the loss was up to 10% of the digitonin in solution. When the solution was pipetted a loss oc- curred also in form of a residue in the pipette. The elimination of digitonin fran the supernatant was also achieved by canpletely drying the concentrated supernatant and alcohol wash in centrifuge tube 3, of step 9. me nonprecipitated material was then reextracted with the ether-washes of the digitonides , sonicated and centrifuged, and these supernatants treated as indicated in the proce- dure. Benzene extraction of the material in B is also feasible, where- as chloroform is not recommended because of its high specific granty. A third possibility of removing digitonin from the supernatant is by purification on TIC. The supernatant should be spotted in 10% meth- anol in benzene since pyridine interfered with the elution. Digitonin stayed near the origin while the steroidal material was eluted. For this purpose either regular silica gel G or Agm3-inpregnated altmnnum oxide neutral, type ‘1' plates were found suitable. Rexnoval of digitonin fron the supernatant is necessary when analyzing this material by gas chromatography and mass spectrometry. Digitonin did not appear as a peak in either the free form or as trimethylsilyl ether derivative but contributed a very undesirable high background due to degradation. The “3 supernatant which resulted when egg yolk nonsaponifiable material was subjected to the digitonin precipitation had a yellow color and had an odor very different from that of the nonsaponifiable material, some- what like a sweet wax. Elbe digitonides had to be dried very carefully since they were very light and easily blown off with the nitrogen stream. 'me ether also had to be canpletely evaporated before drying in the vacuum oven. Losses due to the alcohol washes of the digitonides have been men- tioned in the literature but no preferential resolubilisations have been reported. If desirable to investigate this matter, it is recom- mended that the contents of centrifuge tube 3, after the last extraction be dried and. then to proceed as with the digitonides in A. ‘me procedure worked also well with small amounts of sample, as shown when 1.3 mg of a sample recovered from the origin of a thin-layer plate was subjected to the treatment and precipitate was detected in the solution. Later it was found, by mass spectrometry, that this origin zone contained residual cholesterol. The 111 etho d for precipitation of 3-beta-hydroxysteroids was developed for nonsaponifiable material or individual TIC-fractions therefrom, independent of amount of sample, which sealed only limited by the size of available conical centrifuge tubes. In these tubes the precipitable and nonprecipitable compounds could be separated and the supernatant easily recovered without loss of material in filtration. A large excess of digitonin and a long reaction time made the method nearly quantitative. By reducing the amount of digitonin and the reac- tion time, cholesterol could be removed preferentially and the digitonin content of the supernatant decreased while enhancing other canpounds. M4. Using thin-layer chromatography or partitioning between aqueous alcohol and ether, excess digitonin could be removed from the supernatant. This could also be achieved by completely drying the supernatant and extrac- ting the residue with ether or benzene. To avoid contamination of the sample by comercial digitonin, a purification procedure was developed which treats the digitonin similar to its use in the precipitation and recovery procedures, thus eliminating contaminants, which would other- wise pass into the sample. By using conical centrifuge tubes, the con- taminants could be easily removed and the loss of digitonin minimized. Recrystallization of digitonin was developed for easy handling of the purified digitonin. Recovery of Digitonin Precipitated Material Procedure Digitonin precipitation yielded 2 fractions, a supernatant which could be analyzed directly and the digitonides which had first to be split before the precipitated material could be analyzed further. The following method was developed to recover the material which was preci- pitated with digitonin. l. emu vial precision tare wt. = 2. Add pyridine (60 ml/g sample) to the thoroughly dried digitonides in the centrifuge tube and get them into solution by sonification for Q hour. 3. Concentrate the solution to approximately fi its volume under a stream of dry N2. 14. Add ether (600 ml/g sample) and hold digitonin in sus- pension by swirling during 15 min. 5. Let stand in ice-bath for 2 hours. 6. Centrifuge 10 min at 2,500 rpm. 1&5 7. Transfer supernatant into round-bottom flask (1,200 ml/g sample). 8. Wash the digitonin precipitate 3 times with ether (600 ml/g sample) transferring each time. 9. Evaporate the ether on a rotary flash evaporator with intermediary condense-trap; water bath at 1+0 C; break vacuum with dry N2. 10. Eliminate pyridine by leaving the flask over H2801, in vacuo overnight. 11. Using ether, and sonication if necessary, transfer the recovered material quantitatively through F-fritted glass-filter into the small precision- tared vial. 12. Dry with N , then in vacuum oven for % hour, at 1&5 C. Letecool in desiccator for % hour, weigh the recovered fraction vial wt. recovered fraction wt. 13. Dissolve material in ethyl acetate for GO analysis. With this method the material which was precipitated with digitonin can be recovered in the free fom. Review of Recovery Procedures The most widely used method for the recovery of steroids from digi- tonides was reported by Sch8nheimer and Dam (1933) in which the digito- nides were dissociated in pyridine. To this solution a tenfold volume of ether was added which caused digitonin to precipitate, whereas the steroids remained in solution. The digitonin was removed by filtration and the ether evaporated to yield the free steroids. For quantitation this dissociation in pyridine and precipitation with ether may have to be repeated. They pointed out that the method was quick, quantitative, and required no heat treatment. This was in contrast to the method used before, when the digitonides had to be extracted for days with boiling xylene. With their method they avoided oxidative and thermal. degradation of steroids. serum (191m) modified the Schahheiner and Dam (1933) method since it was rather inconvenient when applied to the splitting of larger manta of digitonides. Thus, he recovered the steroids by keeping the digitonides in pyridine at 70 - 100 C for about 1 hour, removing the pyridine by distillation in vacuo, treating the residue twice with an- twdrous ether, grinding, and then extracting it for 1 hour in a Soadilet apparatus. the canbined ether extracts contained the steroids . To eliminate the precipitated digitonin Butt, et al. (1908) used centrifugation instead of filtration and washed the pyridine-ether extracts repeatedly with all 112801; and water. A similar procedure was adopted by “erg and Seher (1972) who used 10% hydrochloric acid and water to wash the canbined ether extracts. Another cleavage medium was reported by Issidorides, et a1. (1962) in which steroidal digitonides were canpletely dissociated in dimetlwl- sulfoxide at steambath temperature and from which the sterols precipi- tated on cooling. After transfer of the dimetmrlsulfoxide solution to a separatory funnel the sterols were extracted with hexane. After following the Schdnheimer and Dam method for the dissociation in pyridine and precipitation in ether, Sperry (1963) included a step, after overnigit elimination of pyridine in vacuo over sulfuric acid, in which the sterols were redissolved in ether and the digitonin removed by filtration. Mon of the Recovery Procedure The recovery of precipitated material from digitonides was adapted frm Scthheimer and Dam (1933) and takes into account sane of the 1.7 modifications proposed by later investigators. The method of Bergman (1940) was not used since it included a rather vigorous heat treatment and, therefore, could cause thermal degradation. Kornberg and Seher (1972) observed that many artefacts were produced, especially when the method of Bergman was used. The method of Issidorides, et al. (1962) was not adOprted since it has found little acceptance and would have required to work with a rather uncamnon solvent . The digitonides had to be absolutely dry, especially when they were dissolved on a steam bath rather than by sonication. The standing time of 2 hours under step 5 of the procedure is very essential to give the digitonin sufficient time for crystallization. When filtration instead of centrifugation was used under step 6, a problem of clogged filter occurred, again, as already mentioned for a similar situation with the digitonin precipitate. Digitonin was eliminated from the supernatant by drying and filtra- tion rather than acid washing to consistently expose the steroids only to neutral or alkaline medium. When the pyridine-ether extracts were washed with water only, the digitonin accmflnted as interfacial "fluff" and was very difficult to eliminate. In case there is any doubt about the quantitative recovery of the material which was precipitated with digitonin the procedure should be repeated on the precipitate of step 8 of the recovery procedure. When the digitonin precipitation and steroid recovery methods were checked quantitatively, more than 95% of the starting material was recovered from the supernatants of the precipitation procedure and the recovery procedure. The method was adapted to concord with the digitonin precipitation 14.8 method and to recover digitonin precipitated material. It was made independent of sample size and it did not require heat treatment. Using centrifugatim, the filtration of large quantities of precipi- tated digitonin could be avoided. Efficient precipitation of digitonin was achieved by selecting a sufficiently long standing time. Since pyridine could be eliminated fran the recovered sample by standing over concentrated sulfuric acid in vacuo, contact of the steroids with acids could be avoided. By drying the supernatant, extracting the residue with ether, and filtration, digitonin could be eliminated fro: the re- covered sample, thuspreventingmdesirablebackgrotmdinGCandlB analysis. Thin-War (‘.’hrcmatograptnr Procedures The following is the procedure which was developed for the prepar- ative separation of egg yolk nonsaponifiable material on TIC-plates. l. Dissolve the nonsaponifiable material in chloroform, 1 ml/25 mg. . 2. Apply this solution to the plate, 2 cm frm the bottan, with a 1 ml all-glass Tuberculin syringe, drop by drop, fanning a streak between 1+ and 16 cm of the template to yield approximately 2.5 Ins/cm on a 0.5 In or 5 ll'B/cm on a 1mm silica gel G layer. 3. At positions 2 and 18 apply respectively 0.5 mg or 1.25 mg of the nonsaponifiable material. it. mute the plate to groove in silica layer with the chosen solvent system, in developing tank lined with chrontography paper and saturated with the solvent system. 1&9 5. Runove the plate, let dry, if necessary. Detect ‘ desired zones in daylight, or on illuminated plate. If this is not sufficient, cover central part of the plate with a glass-plate, spray side-bands *- with 03$ 12 in methanol or with 50:50 sulfuric-acetic acid, developing colors with a heat gun. 6. Subdivide the silica layer into desired zones with a needle. Solvent systems non-polar elution : chloroform (free of methanol) medium-polar elution : ether polar elution : chloroform-methanol-water, 90+10+l (v/v/v) The preparation and purification of thin-layer plates is described in the section on materials. To recover and extract individual bands separated by TLC, the following macro-procedure was used. 1. 31mll vial Precision tare wt. = 2. If the side-bands were sprayed for visualization, mark the individual zones of interest clearly, then remove l side-band canpletely over corresponding area, using tissue dampened with chloroform-methanol 2:1 to wipe the area absolutely clean, also that edge of the plate over the whole length. 3. Holding the plate over a large funnel which feeds into a Kontes F-iritted glass-filter, scrape individ- ual band of silica off the plate with a flat-ended stainless steel spatula. 1+. Using a Pasteur pipette and chloroform-methanol 2:1, rinse the area from which the band was removed, with- out soaking neighboring layer material, the silica on the funnel, and the tip of the funnel. 5. Add enough chloroform to the filter to produce a thin slurry on swirling. Let sit 15 min, force solvent through the filter into 50 m1 centrifuge tube by means of lie-pressure without drying the silica. 6. Repeat 5. with chloroform-methanol 2:1, chloroform- methanol 1:1, chloroform-methanol 1:1 saturated with water, methanol, and methanol-water 1:1. 50 7. Add 0.29% aq. NaCl to the centrifuge tube with intermediate vigorous swirling, until the chloro- form mase represents less than 140%. Centrifuge 5 min at 2,500 rpm. 8. Remove water-methanol phase with aspirator and repeat 7 three times with water. 9. Concentrate chloroform solution to small volume with dry N2, water-bath at ’45 C. 10. Transfer the concentrate quantitatively into the small precision-tared vial. 11. Dry extract with N2, then in vacuum oven for % hour at ’45 C. Break vacuum with N2, let cool in desiccator 5 hour, weigh vial wt. = extract wt. = 12. Dissolve extract in small, known quantity of ethyl acetate or 10% methanol in benzene. Store in freezer. For trace compound analysis and to subfractionate macro TIC-bands, the following micro TIC-band recovery and extraction procedure was developed. 1. Small vial precision tare wt. = 2. Remove zone with a micro TIC-band collector connected to an aspirator. Inverse and accumulate the silica on the plug by tapping the pipette slightly. 3. Remove collector fran aspirator connection, wipe collector bottom with tissue, dampened with chloroform-metha- nol 2:1. 14. Shorten tip of collector so that the needle of a syringe can reach into the larger part of the collector. 5. Install collector on 25 ml centrifuge tube, extract the silica consecutively with 2 m1 of chloroform, chloroform-methanol 2 :l, chloroform-methanol 1:1, chloroform-methanol l :l saturated with water, methanol, and methanol-water 1:1. 6. Continue with 7. of the macro procedure. 'nie preparation and cleaning of micro TLC-band collectors is described in the section on materials. 51 Egg yolk nonsaponifiable material could be subfractionated very efficiently with these procedures. Review of Separation on Thin-Layer Plates Adsorption TIC-separation of the nonsaponifiable material fran egg yolk is a convenient way to separate most of the compounds from the predominant cholesterol and from each other, and is the most suitable mode to recover a pure eluate. Partition or reversed-phase chromato- graptw is undesirable for preparative TIC because of contamination of the sample with the liquid phase, or the need for extensive heating to remove it, with the possibility of thermal degradation. In addition, impregnated plates may not allow as much sample material to be spotted as nonimpregnated plates of the same thickness, (Souza and Nes, 1969). Besides adsorption chromatography on regular TLC-plates, many investigators have utilized AgN03 thin-layer chromatography, taking advantage of the complexing characteristics of AgNO3 which subdivides compound groups on the basis of unsaturation. To separate steroids on thin-layer plates, the free compounds as well as a number of derivatives have been used, such as acetates by Copius-Peereboom and Beekes (1965), 3,5-dinitrobenzoates by Seher and Hanberg (1972), trimethylsilylethers by Brooks and Watson (1967). These derivatives were not preferred over the free compounds for any reason other than improved resolution under specific conditions, and this was usually demonstrated with standard material. Suggestions have been made that derivatives can undergo hydrolysis during their contact with the layer material, (Brooks and Watson, 1967 , Idndgren and Svahn, 1%6, Idler, et al., 1966). Since the nonsaponifiable material of egg yolk '52 is an extremely complex mixture, it was decided not to use derivatives with the possibility of increasing the complexity of the sample even more by the occurrence of Irydrolysis products. Therefore, all. the TIC-systems were worked out with either free standards or free sample mterial. This also simplified the procedure by avoiding derivatiza- tion before and hydrolysis after TIC with their problems of purifica- tion and sample work-up. Another reason for not using derivatives in this particular case is that derivatized steroids are more uniform with respect to polarity, thus eluting as a group which is unnecessary for yolk nonsaponifiable material, as already pointed out for the digitonin precipitation pro- cedure. Preparation of TLC-Plates The method for preparing the plates is described in the section on materials and follows standard procedures, (Kaufinann, 1970a). Predevel- apnent with a methanol-ether 80:20 (v/v) mixture, as suggested by Brown and Benjamin (1964) and Adlercreutz and Luukkainen (1968), was included in the preparation. In this study the plates were predeveloped in the intended direction of the subsequent sample elution since this helped in evaluating the elution behavior of the individual plates and allowed the discarding of irregular ones before any sample material. was wasted. Since TIC was used in this research as a preparative technique, the contamination of the sample by the layer material was thorou@ly inves- tigated. It was found that silica gel G was not pure; but even silica gel 60 HR or silica gel C after acid washing according to Kottke, et a1. (1966), yielded still some colored residue after chloroform-methanol. 53 extraction. Silica gel 60 HR and. acid-washed silica gel G layers were very fragile and could be used, at best, for analytical purposes for which this high purity is not required. This fragility was very incon- venient for the following reasons: during the drop-wise application of the sample the layer was usually disrupted, during handling or on placing the plate into the eluting solvents frequently part of the layer fell off, and during visualization by spraying the layer was easily blown off. On the other hand, using silica gel G plates predeveloped with Ethanol-ether, it was found by gas chromatogralmic analysis that some TIE bands did not show any peaks. his shows that predevelopnent and, possibly, the washing of the TLC-band extract eliminated contam- inants originating in the layer material. The cutting of a l m groove across the silica layer served to clearly separate the area used for sample separation fran the area into which contaminants were eluted during predeveloment . When the band recovery was not intended close to the solvent front, this precaution was not necessary and the layer was not cut, thus allowing the use of this additional length of layer for further separation. Sane ready-made plates were experimented with, such as ISM-Reagent plates precoated with 2.0 mm silica gel F-Zsh. They were abandoned because of their extremely long elution times. Eastman Chransgram sheets 6&1 featuring silica gel without fluorescent indicator were inconvenient since acids could not be used for visualization. 51+ Application of Yolk Nonsaponifiable Material to the Plates Chloroform was found to be the best solvent for dissolving the nonsaponifiable material from egg yolk and for application on thin- layer plates. It combines high solubility with good volatility. Hexane was less suitable in the former characteristic, whereas ether evaporated too quickly forming residues on the tip of the applicator. Benzene was acceptable, whereas pyridine influenced the elution and ethyl acetate appeared incompatible with the layer material, especially aluminum oxide. Hamilton syringes are recommended for spotting small quantities and all-glass Tuberculin syringes for larger quantities. The Haak.pipette had no advantage over a syringe and with pipettes in general, it seemed more difficult than with syringes to control the drop by drOp spotting. .A Kontes spot applicator (Kontes Glass Co., Vineland, N.J.) was evaluated for streaking by moving the plate correspondingly. Since the apparatus was equipped.with too fine needles it required dilution of the sample and application of pressure. Therefore, it was not used. In this study, when not working in an inert atmosphere, the syringe was fixed in an inclined position, while the plate was moved along a smooth edge during streaking. When using the TLC spotting box with nitrogen or carbon dioxide, the plate rested in the box while the syringe moved with the built-in slider. While the solubilizing power of chloroform is much higher, the concentration obtained on adding l.ml of chloroform per 25 mg of sample was found to be a good compromise between high concentration for quick application and.suitable behavior during application. 55 Fer preparative separation of compounds other than cholesterol, the thinrlayer plates needed to be loaded with as much material as possible to yield sufficient material for further analysis. A streak, in ferm.of a single row of spots, did not allow detection of all bands present and, therefore, repetitive streaking was necessary to apply enough.material. The limits for medium-polar and polar elutions were found to be approxi- 1mately'2.5 mg and.5 mg per cm for 0.5 and l.mm silica gel G plates, respectively; For the reference spots in the side-bands, 10 drops of the l.ml/25 mg solution and 1.25 mg were suitable. For nonspolar elu- tions the plates could be loaded to double these values. These load concentrations should not be exceeded since the performance of the eluting solyents depends very much on them.in preparative TLC. For reference material in the side-bands it is preferable to use the same solution which is being separated, since it was found that standards did not always elute to exactly the same position as their counterpart in the biological mixture. This phenomenon was also. reported by Wotiz and Clark (1969). Elution of the Plates The suitability of solvents which yield a good separation of egg yolk nonsaponifiable material on silica gel G plates was investigated. Elution by a single solvent system did not yield a satisfactory resolu- tion of compounds. The resolution was vastly improved by eluting the nonsaponifiable material in three different systems. Chloroform was selected fOr non-polar elution in which cholesterol remained very close to the origin, whereas less polar compounds were distributed between cholesterol and the solvent front. This s stem.is recommended 56 for analyzing nonpolar canpounds of yolk nonsaponifiable material. Ether was selected for medium-polar elution since it moved cholesterol to apprcncimately half height on the plate. This system was used and is recamnended to recover canpounds of similar polarity as cholesterol. For the polar system the mixture of chlorofom-methanol-water 90+10+l (v/v/v) was selected which placed cholesterol into the upper half of the plate and allowed recovery of bands of polar substances. Eluting cholesterol with various solvents, the following series of solvent polarities resulted: hexane < benzene < chloroform < ether < ethyl acetate < methanol; which confirmed the arrangement for pure sol- vents reported by Neher (1%9). Deviations from his values, which were based on the average Rf-values of 20 steroids, were found with respect to the elution of cholesterol by solvent mixtures . Ether and the mixture of chloroform-methanol 90+10 were more eluotropic than benzene -methanol 85+ls. Based on the elution of squalene the following series of solvent polarities resulted: hexane < chloroform < ether < benzene. This shows that chloroform eluted nonpolar compounds less and medium-polar further than benzene. With respect to the number of nonpolar compounds separa- ted, it was found advantageous to use the chloroform system with con- tinuum developnent. The progress of the ccntimous developnent could be controlled by watching the yellow squalene band either in daylight or under Ultra Violet (uv) light. A lamp placed behind the plate was very helpful in determining the exact position of the advancing solvent front. (bntinwls developnent with chloroform, accomplished by removing the cover from the developing tank when the solvent front reached the groove in the silica layer, (Lisboa, 1969), was not suitable for cholesterol 57 and similar canpounds since they were still in the lower half of the plate after hours of continuous developnent. For this reason, the ether system was selected for these compounds. The ether system also allowed continuous developnent which was controlled by observing the advancing yellow bands or the UV-yellow-fluorescing band immediately ahead of cholesterol. The polar system was controlled during continuous developnent by observing one of the strong visible yellow bands. The suitability of solvents for continuous developnent is largely a function of their volatility and, based on their vapor pressures at standard pressure and temperature, they can be arranged to the following series: ether > acetone > chloroform > methanol > hexane > ethyl ace- tate > benzene; thus, ether is the most and benzene the least suitable solvent in this respect. When mixing chloroform and ether some heat developed and, therefore, this mixture must be used cautiously. Based on the most polar, detectable band of egg yolk nonsaponifi— able material, the polarities of solvent mixtures could be arranged to the following series: ether < benzene - ethyl acetate 50450 < benzene- methanol 85 +15 < chloroform-methanol 90+10. The separation patterns were mostly governed by the polarity of the eluting solvents. For components of nonsaponifiable material only minor shifts relative to cholesterol were observed in different solvents. To increase the polarity of a solvent mixture in preparative TIC, the methanol or ethyl acetate part could not be increased deliberately without encountering a channeling effect. When very high Rf-values are desired, repetitive elutions or continuous developnent should be uti- lized. Although switching solvents sanetimes leads to a better isola- tion of a particular compound, the three recommended elution systems, 58 and using the micro-recovery procedure in canbination with preparative GO, should be sufficient for the analysis of egg steroids. Visualization of Individual Bands Sane bands were easily distinguished because of their color, while others were visible in daylight against a dark background, on an illu- minated plate, or under UV, direct or when placed behind the plate. Exposure to UV-light should, however, be as brief as possible, (Horlick and Avigsn, 1963). An attempt was made to use the Kontes Chranaflex 1! 149-5000 Densitaneter (Kontes Glass Company, Vinelsnd, NJ.) for non- destructive visualization of individual bands. This was not successful since the instrument scanning speed was not synchronized with the speed of the recorder paper and the pick-up head could not be lowered flat onto the plates. light spraying of the plates with methanol or water is sanetimes reconnnended but all of these methods require fairly concentra- ted bands or strongly reflecting compounds, such as cholesterol or squa- lene in the nonsaponifiable material of egg yolk. To obtain consistently good orientation on minor compounds, the side bands were sprayed with either 0.5% iodine in methanol or with 50:50 acetic-sulfuric acid and heated locally with a heat gun. The use of the 2 sprays gave the most reliable information on TIC-band position. A number of fluorescent sprays such as 2' ,7'-dichlorofluorescein, Rhodamine B, Rhodamine 66 and other sprays such as 50:50 acetic acid - sulfuric acid, phosphomolybdic acid, phosphotungstic acid, conc. phospho- ric acid, iodine, and bromothymol blue were screened to supply maximum information. The 0.5% iodine in methanol and the 0.01% aqueous moda- mine 60 were considered the best nondestructive sprays, whereas 50:50 59 acetic-sulfuric acid and 5% phosphanolybdic acid in 99% ethanol were the best destructive sprays for TIC-separated egg yolk nonsaponifiable material. Layer sprayed with Rhodamine 6C} and the 50:50 acetic-sulfu- ric acid sprays had the additional advantage that they could be evalu- ated in daylight and under UV. The acid mixture had to be stirred before use. To attempt the detection of a particular compound, an approach similar to that proposed by Mchgold (1961) is recamnended. The plate should first be inspected in daylight on a dark background, then on an illuminated plate and under LIV-light. Thereafter, 1 side-band of the plate should be sprayed with 0.5% iodine in methanol, the other with 0.01% aq. modamine 60, and inspected in daylight and under uv. After evaporation of the iodine, the corresponding side-band should be sprayed with 50:50 acetic-sulfuric acid, heated locally with a heat gun, and then inspected in daylight and under UV. Finally, it should be sprayed over with 5% phosphanoly‘bdic acid in 95% ethanol for improved contras- ting. After each of the above steps, the observations should be care- fully recorded and the bands on the plate marked with a needle. Recoveg of Individual Bajnds frcm TIC-Plates The macro-procedure followed normal procedures as reported by Kaufmann (1970b). The micro-procedure was devised after it was recog- nized that bands as small as 8 m were still far too canplicated when analyzed by gas chromatograplw. The preparation, cleaning, and use of the micro-band collectors is described in the section on materials. The collectors are, in princi- pal, similar to those used by Goldrick and Hirsch (1963), but have the 6O advantage of being quickly prepared in large quantities and easy to use. 'me micro-procedure has drastically reduced the couplexity of new of the recovered fractions and the Pasteur pipettes and 313.33.. wool can be discarded after use. The glass-wool plug should not be pushed further than the constricted area. Thus, collected layer material accimmlates in form of a column on the plug and does not form a ring around it which would allow the solvents to drain directly through the glass-wool instead of percolating through the layer material. With the micro-collectors the width of individual bands could be held very mall, even Just to the material loosened by a single stroke of a needle, since the loose layer material could be sucked up without the tip of the pipette touching the layer on the plate. Bands which were as wide as the tip of the pipette or wider could be completely ranoved by scraping the plate with the pipette tip itself. Smaller bands could be loosened with a needle or other suitable device and sucked up. For steadiness, the TIC-template was used like a ruler. To avoid oncidation during collection, the plate can be placed into a larger box flooded with inert gas. For best results it was essential that the elution of the plate was as equal as possible across the plate to minimize mixing of neighboring bands and the nunber of bands in which an individual canpound appeared. This also masdmized the concentration of an individual cmpound in the extract of its band. The solvent sequence given for the extraction of the layer material was found to elute compounds fran even the most polar fractions. Non- polar compounds could be eluted by simply using chloroform-methanol 2:1 or ether 0 61 The removal of water and methanol and the washing of the extract was devised on the basis of experiences made with the phase separation used by Folch, et al. (1957). Since the methanol content of the first phase separation was relatively high, the use of 0.29% aqueous sodium chloride for this phase separation is recommended to increase the polarity of the water-methanol phase, thus avoiding losses of polar material. Washing with water was repeated until the chloroform layer cleared without centrifugation and no more interfacial "fluff" was seen after centrifugation. In this work, no more than 3 washes were neces- sary. This washing procedure should remove all contaminants extracted from the layer material, including silicic acid which was demonstrated by Eder (1972) to be eluted in increasing amounts with increasing pola- rity of the extraction solvents. Soxhlet extraction of layer material was also used but the mixt- solvent extraction, described above, was simpler and quicker. Prgzgation of Silver Nitrate TLC-Plates Silver nitrate plate preparation is described in the section on materials. Basically, the same technique was used as for regular TIC- plates, except that, true the moment the AgN03 solution was poured into the aluminum oudde until the plates were fully activated, no light was admitted. Although important for nondestructive detection, this was not mentioned in the literature on silver nitrate thin-layer chromato- graphy such as Morris (1963), Kamereck (1967), Stahl (1969), Idler and Safe (1972). It was also essential that the plates received enough dry- ing time, such as overnight, followed imediately by activation. Plates prepared in this way were absolutely white and could be kept for weeks 62 without appreciable browning or loss of efficiency. Activation not only prevented browning but also improved resolution and control of the layer conditioning. This method also worked well for silica gel G using a 1:2 powder to water ratio. However, based on the sterol-pair, cholesterol and cholestanol, a much better separation was achieved on almiimnn oxide and this layer material was, therefore, preferred. A similar material, Neutral Almfina AG? (Labeled 2.1a. microns) containing 5% by weight of assoc, was used by Ksmereck, et al. (1967) but this is no longer avail- able and it was substituted by Ransos (1973) with the material used in this research. With layer materials such as silica gel 60 HR and acid- washed silica gel G the problem of fragility was encountered, as pre- viously mentioned. With respect to the AgNO3 concentration on thin-layer plates, sane investigators (Morris, 1966, Stahl, 1969), claim that beyond 31. AgN03 no additional separation can be obtained. The use of 5% is thus recan- mended. In this study a better separation of cholesterol and choles- tanol was obtained at 10%, and for consistently good results, the use of 15% was adopted. When Agm3-TIC was used in recent investigations such as Copius-Peereboan and Beekes (1965), Idndgren and Svahn (1966), and Idler and Safe (1972) , the concentration was usually appreciably higher than 10%. Contrary to regular TIC no solvent system was found suitable for predevelopnent. The polar solvents all dissolved AgN03 to sane degree and thus reduced the separation of saturated and unsaturated compounds. Increased Rf-values for unsaturated compounds were also observed in mul- tiple elutions with chlorofom-acetone 90+10. Ethyl acetate may be used it! 63 predevelopent. However, it is of relatively low polarity and causes thelayertoturnsmewhatyellowandthepredevelnpedplates toshow slightly reduced separation capability. To circumvent the possibility of contamination from the layer material, the method of Cubero and hagold (1%5) was evaluated with predeveloped AgllOB-free plates by developing the plates with various concentrations of aqueous silver nitrate solutions. Since erratic re- sults were obtained with these plates and the authors already stated that the AgID3 gradients were not constant nor reproducible, this approach was abandmed. As a corollary, predevelopent of silver nitrate plates may be used to produce reproducible gradients which may be useful in certain applications. AglO3-plates prepared as described and then predeveloped with acetone or any other suitable solvent should yield a plate with a continually increasing silver nitrate content up to its naninal value at the end or the plate. Samples could then be run with the gradient, against it, perpendicular to it, and even two-dimensionally. A layer thickness of 0.75 m was found necessary for streaking 2.5 mg sample/cm. Apparently, for equal layer thickness, the silica gel G plates absorb more sample material which may be due to the dif- ferent layer to water ratios used in the preparation. Use of Silver Nitrate TIC-Plates These plates were used essentially the same way as regular silica gel G plates. lhe sample was dissolved in chloroform, the load was 2.5 mg sample per cm for the 0.75 m thick AgNO3 plates and the reference sample material was spotted at positions 2 and 18 of the template. 51. The non-polar system of chlorofom-acetone 90+lO and the medium- polar system of ether-acetone 50-60 gave good elutions. Here, again, the polarity of the solvents could not be increased beyond limits with- out a channeling effect and the afore-mentioned washout of Agm3 by these solvents. When Imiltipleelutions were used, residue from former elutions could be seen and, therefore, munmus development with the ethwystein is recmnenm when higher lit-values are desired. M were usually carried out in the dark, although the need for this precaution was not investigatd. Contrary to silica gel G plates, the circular elution technique, (Mangold, 1961, Randerath, 1%6), did not work well on these silver nitrate plates. ihe effects of multiple developments could not be evaluated by this technique on an plate. The eluting system of Subbiah (1972) , chloroform-methanol-acetic acid 100+l+0.2, was canpared with chloroform-acetone 90+10 and was not superior for separating cholesterol and cholestanol. To reduce tailing, Lisboa (1%9) found that the addition of water and/or acetic acid was very useful. water is not recomended for the silver nitrate plates but acetic acid at 0.5 ml or less per 100 ml decreased tailing tranendously while not impairing separation of canpounds and may be considered for the ether-acetone mixture. In this case it is advisible to start with a completely dry tank, to saturate it with the maJor solvents, such as ether - acetone, and to recharge the tank with a fresh load of solvents containing the acetic acid Just before actually eluting the plate. In this way more reproducible results were obtained and the additive effects of acetic acid residues were avoided. Visualization during continuous runs is, again, based on hands visible in daylight or UV. Side-bands were visualized with 0.01% 65 modemine 6G or with acids and heat as for regular silica gel G plates. Since Ebodamine 66 was not eluted by the solvent systems used, it may be applied to the plates with the slurry during spreading, (Marigold, 1961, Avigan, et al., 1963). To collect individual bands and extract the compounds, the same methods were used as for the silica gel G plates. The washing procedure eliminated the extracted Agso3, (Morris, 1963 and Randerath, 1966). Break-down and Adsogtion on TIC-Plates A problaa canon to both plate systems is the danger of degradation of cmpounds. Oxidation can be largely prevented during spotting by using the inert-gas spotting box and during collection of bands by keeping the plate flooded with inert gas. While oeddation during elu- tiui can be minimized for silica gel G by adding EST to the eluting solvents, no antioxidant was found for use in connection with AgNOB TIC. To prevent oxidation, Kaufmann (1970c) advised that elution be done in the dark in tanks filled with 002 or N2. He did not indicate, however, how the 002 or N2 influenced the elution. These conditions were also used by Avigan, et a1. (1963). Because of the lack of a suitable antioxidant and since no pre- developing system was found, results obtained with the Agllo3 plates should be evaluated with caution. It should, however, be kept in mind that the overall procedure is proposed primarily for canparative and only secondarily for qualitative analysis. When the scheme of analysis developed here is used quantitatively, the break-down of cholesterol should be considered because of its pre- daninant presence in egg yolk nonsaponifiables. This is best done by 66 using half of each plate for the developent of cholesterol standard material and the other half for the nonsaponifiables. A canparison of equivalent bands from both plate halves by 60 should indicate which peaks of a trace originate frm cholesterol. Break-down of cholesterol was detected, especially on the AgMyplates and, therefore, this can- parative evaluation is necessary when making qualitative assessments. ‘nie break-down of steroids on thin-layer plates was also demonstrated by Idler, et al. (1966). Not all steroid losses, determined gravimetrically or otherwise on the extract of their respective bands, are due to degradation. Sate cholesterol on these preparative plates was strongly adsorped in the origin and was incaspletely desorbed during elution. Also, all bands between the origin and the actual cholesterol band showed traces of cholesterol, whereas those bands between the cholesterol band and the solvent fruit contained no cholesterol detectable by gas chrasatograplnr. An analysis of the origin fraction with the mass-spectrcneter showed that this cholesterol is not degraded. Seher and Hanberg (1972) observed the same phenasenon by recording radiochromatograms of their thin-layer elutions. They found a small radioactivity in the origin but did not elaborate m the cmdition of this residue. Their data also showed increased radioactivity between the origin and the proper band itself, but this fact was not mentioned. Avigan, ct al. (1963) detected radioactivity of the less polar lanosterol in the band of the non-labeled cholesterol on silica gel G and of dihydrolanosteryl acetate in lanosteryl acetate on Agm3 plates. The relative amounts of adsorbed or incompletely eluted steroids was not determined, but is very small. It could be demonstrated in this 67 case only because of the large quantity of cholesterol applied and by Seher and Homberg (1972) because of the extremely high sensitivity of the tracer method. It can be implied that other steroids would show a similar behavior and that quantitation by mere spectroscopic or gravi- metric measurements on the extracts of corresponding bands would yield low results. Thin-layer chromatography was adapted to yield preparative separa- tion of minor and trace level compounds from cholesterol, the major cmponent of yolk nonsaponifiable material. The adsorption mode was selected over the partition or reversed-phase mode to avoid contamina- tion and/or them]. degradation of sample material due to the liquid phase. Predevelopnent of the plates was adopted and performed in the direction of ensueing separations to allow the use of silica gel G layer material for preparative separations, to retrieve contaminants deriving fran it, find to evaluate the quality of the plates. Chloroform was selected for application of sample material because of its high solubili- zing capacity and the concentration of the sample in chloroform adjusted for quick application and suitable behavior during application. Fixed syringes were found most convenient for spot by spot streaking while either guiding the plate or the syringe. Eluting systems were selected allowing maximum loads on the plates to enhance minor and trace compounds and to minimize elution times while yielding even elutions. Thus, mixing of neighboring bands and umber of bands in which an individual cmpound appeared could be minimized while the concentration of a com- pound in its hand extract was maximized. Three solvent systems of different polarity were selected for elution to maximize resolution of non-polar, medium-polar, and polar canpounds, respectively, and to 68 allow continuous development. A schmae was adapted for efficient visua- lization without degradation or contamination. The micro TIC-band recovery and extraction method was M which drastically reduced the complexity of individual TIC-fractions using inexpensive collectors which are quickly prepared and easy to use. A solvent sequence was devised to recover compounds fran the most polar TIC-bands and a washing procedure developed to eliminate contaminants extracted from the layer material, including silicic acid. The technique to prevent browning of Agm3 TIC-plates was developed allowing nondestructive detection during and after elutim. (the opti- mum concentration of silver nitrate necessary for separation of choles- terol and cholestanol and the layer thickness to allow preparative separation of smnple material was determined. the solvent systems of different polarity were selected for good resolution, even elution, and continuous developnent . A visualization procedure for nondestructive and destructive detection was adopted. Precautionary measures to avoid or evaluate degradation in thin-layer chromatography were developed and the effects of adsorption investigated. 69 Gas -Liquid Chranatography on: was used to control the purity of standards, solvents,and reagents, to monitor TIE-separations and digitonin precipitations, to separate sample mixtures, and, in a later mase, for preparative work. mum of Colmnns and %ration of GC Column packing material was originally prepared according to the procedure of Homing (1968) who described how the support material was graded, acid-washed, freed of fines, silanized and coated. In later studies, colmn packings were purchased directly from suppliers. Sila- nization of the glass surfaces and the glass wool to reduce adsorption was accomplished using a 5% solution of dichlorodimethylsilsne in benzene, (Homing, 1968). Glass columns were used since metal columns tend to adsorb or react with trace compounds, (Rosanski, 1%6), where- as the use of Teflm is limited with respect to temperatures. Packing of columns is described in the section on materials and follows procedures of Vanden Heuvel and Homing (1962), Homing, et al. (1%3), and Johnson and Stocks (1971). Columns were conditioned by holdim for 1:8 hours at the madman admissible temperature of the coating material, while pusing a small stream of nitrogen through the calm. the carrier-gas flow-rate was optimized for all columns with cho- lesterol at isothermal temperature and the hydrogen according to McRair and Bonelli (1968). The flows were all measured at the detector with a bubble flow meter, since the built-in rotameters were erratic. lime oven temperatures were chosen such that the cholesterol retention time was between 7' and 12 min and the sample size so that peaks rose to 7O 60 to 90% of full scale. The flow-rates, optimized for isothermal operation, were also used for programed temperature runs. Since no differences were noted when helium was replaced by nitrogen as carrier gas, nitrogen was retained. A flame ionization detector was used. With respect to percentage of coating the following theoretical plate numbers were measured with SP-ZhOl coltmms: for no coating, at all, l+00, for 1% 1,200, for 3% 2,700, and for 5% 3,500 theoretical plates per 6 ft colulm at the respective temperatures of 200, 200, 210, and 250 C. This shows that, for the range covered, stronger coating yielded greater plate numbers but required hidier oven tallperatures. ESP-2100, an improved version of 83-30 column pachng material, was used to separate steroids on the basis of carbon number or configuration of molecules. To separate saturated and unsaturated canpounds addition- al packing materials were evaluated. When using OV-l, SP-2250, and Oil-225, cholesterol and cholestanol did not separate, whereas when using SP-2h01, a newer version of the QF-l reported by Vanden Heuvel, et al. (1961) and Copius-Peereboan (1965). cholesterol and cholestanol separated and resolution increased as coating percentage increased. The influence of SP-2h01 on cholesterol and. cholestanol separation is opposite to the effect of Agm3 in T18. Cholestanol is retained on SP-2h01 more strongly than is cholesterol. me sample was introduced into the 60 by on-colunm injection with a 10 ul Handlton syringe. A variety of solvents may be used to dissolve the free steroids, (Homing, 1%8). In this research, ethyl acetate was used whenever complete dis- solution could be achieved, otherwise 10% methanol in benzene was sub- stituted. Chloroform was not used because it strips the column of its liquid phase and causes detector fouling, (Kuksis, 1971b and Johnson “I ..P ,E.I‘.l. n A 5V3; 71 and Stocks, 1971). When the sample weight was known, 1 ul solvent per lOugsamplewasadded. Hhentheweightwasnotknown, the solution was hept concentratedandwas dilutedonthebasis of triach runs. To show trace calpomds within cholesterol samples by 00, the auxin loads without overloading were injected. The respective amounts forhoancolunswere 50and12.5ugbasedoncholesterol. Over- loading causad trailing compomds, not.separated fran cholesterol, to elute later, whereas separated compounds which were eluted later were not affected. Minor canpounds which elute close to cholesterol appear asshouldersand, thus, umIteamomtsofasuhstancemaynotshowon the (IO-trace when eluted in a high attenuation setting. A lowering of the oven temperature usually increases the resolution of trace commands from descending peaks. These were often seen with cholesterol m: fractions obtained frat egg yolk nonsaponifiable materi- al. For nomal screening pm-poses the temperature was progranned from 180t02750 atZC/min, holding20minatthemaximmtemperature levelforthe3$8P-2100endfran150toz750at20/minwithouthold- ing for the 5% BIN-21001 calm. Since the baseline of the GC-trace increased calm during temperature programsd runs, the dual-mode of the 60 was used to cm- pensatelcolmnegainstthe other. However, thismodewouldhave required 2 columns filled with the same packing material, as was sugges- ted by mksis (19(1a), and, thus, would have reduced the capability of theavailableGbehalf. Itwasthoughtmoreimportant tohave2dif- ferent columns available than to prevent base line deflection. The ac was usually operated at relatively high sensitivity to yield maximum possible information from trace substances while adjusting to strong 72 signals with an automatic attenuator. This operation, again, showed increased base line deflection, and an electronic base line empensator would have been helpful. When excessive tailing of compounds was observed, attempts were made to sharpen the elution with the built-in auxiliary gas feature. The use of auxiliary gas, however, upset the optimized carrier gas and hydrogen flows which could not be readJusted. Derivatisation When an elution pattern was poor due to degradation on the 00, the trace was poor on both coluans and the fraction had to be derivatized. Self-built Pasteur pipette vials were used for this purpose, (nu-ten, 1973). The following is the procedure which was adopted. from Thenot and Earning (1972) and used for nethomy-t rinethyhilylether-derivatisation. l. Fire-close tip of Pasteur pipettes to a blunt end, shorten large cylindrical part to half its length, fire polish the out border. 2. Clean these vials as all. regular glass-ware. 3. Transfer 0.2 ng steroid into the absolutetly dry vial, evaporate the solvent with a stream of I2. It. Add 50u12$mX, close with Teflon laninated septa andreactforlSIinatSOC. 5. lvaporate the pyridine with a stream of NZ an a water- bath at 60 C. 6. AddSOulTsnl, close, and react at 1000 far2hours. 7. After cooling, withdraw samples with a syringe fron the closed vial and inject directly on the ac. With this method free steroids are converted to methaty-trinethyl- 311:1 (tn-us) derivatives. 73 be 2% max reagent was substituted for the recommended 100 mg Wflde per m1 pyridine. It was not necessary nor desirable to derivatize mast TIC-fractions since many nonderivatized sallples chmtogrsphed well, as was also reported by Patterson (1971) for 92 sterol standards. In general, derivatized fractions eluted on 8P-2100 at still. higher temperatures than the free compounds, whereas on SP-2h01 they eluted at lower temperatures. Patterson (1971) noted that sterol acetate derivatives of compounds differing by one or two methyl groups at C-h or differing in configuration at the 3 or 5 posi- tions were less well separated than the free compounds. Another such ample is shown in Figure 11 in the section on results and discussion. The twin peak appeared as a single peak when derivatized. Trimethyl- silyl derivatives of cholesterol and cholestanal separated also less wellthanthefreecamomdsinthepresent study. Ellhus, inmany instances a nonderivatized fraction shows a better separation, but some stereoisaners can be resolved better after derivatization. The mass spectra of many free steroids are more meaningful than of derivatized steroids (Wyllie and Dderassi, 1%8, Budzikiewicz, 1972), and most non- derivatized steroids can be analyzed, (Budzifiewicz, 1972). Budsikiewicz also reported the extreme variability of so-called "diagnostic” ions for EBB-derivatives depending on the mode of operation or the instrument used, see also Diehman and DJerassi (1967). The derivatization proce- dure also calls for extensive heat treatment and, thus, degradation of some thermally labile canpounds may occur. Therefore, derivatization is desirable only in a few cases. Incanplete derivatization or partial hydrolysis may add to the com- plexity of the fractions analyzed. In addition, contaminants derived '41 IV. )el.‘ l t r! a, e 71+ fra the derivatization reagents may also increase the complexity. These reagents were often highly contaminated, especially the trimethyl- silylimidasole. Therefore, to evaluate the quality, 2 ul of the res-- gents should be inJected on the 60 directly out of the battle when received. When a series of derivatives were prepared for GO analysis, a reagent dum was prepared in parallel for comparison. mauve 00 Although preparative 60 was not successfully performed on the avail- able instrument due to degradation in the detector, its application is discussed here, since it is an important feature of the overall proce- dure. The equipaent layout, adapted iran Kobayashi (1971+), is shown in Figure 1 and cmsists of a 30 cm long standard 6 m glass-tube, filled to 10 cm with glass-wool to break aerosols, (Johnson and Stocks, 1971). These GC-collectors were prepared by tightly spinning glass- wool aromd the tip of a long Pasteur pipette and drilling the spun tip into the glass-tube. They were cleaned the same way as the micro TIC-band collectors. The stopcock served to evacuate the desiccator (vacuum sink) before the actual collection of a peak. When collection was started the twdrogen flow was stopped first, then the tube pushed over the detector tip, and the valve opened. The collection was re- peated a sufficient number of times and then the collected material recovered from the tube by chloroform-methanol extraction . The extract was quantitatively transferred into special micro-vials made from Pasteur pipettes as described in above section on derivatization; fire- closing, in this case, to a sharp tip. To assure that the desired fraction was collected, part of it was reinJected on the same column. 75 ago ter- aaarufor .Sl‘opcack r—o-e-r N \ Plastic hoses '\ r:\ \ I Ibcuum Sink D an [my . glass wool (desiccator) plug *r- Standard glass tube » 6mm x .300 mm Figure l. Arrangmaent for collection of (IO-fractims in preparative gas chromatography. 76 Another injection was made on the companion column to appreciate the homogeneity of the fraction and to obtain a GC-trace which should have served as the basis for the analysis on the (EC-$8 system. Relative Retention, Steroid Number, and Mefllene Units Several semi-mathematical cmcepts have been proposed to facilitate identification of compounds and canparison of inter-laboratory results obtained by ac and to reduce the need to procure a large number of standards for each laboratory. ‘me relative retention is the quotient of the retention time of an unknown to that of a reference substance. It has several disadvantages, especially that no uniform reference substance has been adapted and that it depends on isothermal data. Homing and Vanden Heuvel (196M pointed out that small changes in operating conditions significantly changed the relative retention times, thus this concept is not reliable. Vanden Heuvel and Homing (1%2) proposed the steroid number con- cept which fixed the elution times of steroids on a scale detemined by the 2 nonsubstituted steroids, androstane and cholestane, with their respective steroid mmbers of 19 and 27, corresponding to the number of carbon atom in the molecules. This system has 2 major disadvantages in the context of this research. First, androstane has very short retention times and, when used, temperature programs must be initiated at very low temperatures and require extremely long times since the temperature increase should not exceed 1 to 2 C/min. Secondly, most sterols elute later than cholestane, and sane even significantly later, making it necessary to extrapolate, with the inherent risk of signifi- cant error. 77 'memostmitable syetmnappearstobethatproposedbyxovats (1959) which is either expressed as retention index or in methylene mute (m) which correlate by the formla retention index - 1.00 x metInrlene units, (Horning, 1968 and. Dalgliesh, et al., 1966). ibis systaisbasedmtheeltrtionofaseries ofn-allsaneswiththe scale ' corresponding to the amber of carbon stuns or metlwlene units and is plotted either against the logarithm of retention time, in isothermal runs, or the logarithm of tmuperature, in temperature proposed runs. menthodhas the advantage that 2 n-alkanes can always be foundwhich elute close to an unknown cosmound of interest and whose elutim time ortemperature liesbetweenthose ofthe standards, thus, themore precise interpolation approach can be used. More recently, the group of Homing were using III-values to characterise steroids and steroid nuabers (SH) for calculations involving functional group effects, (Home. at al.. .1967. 1969). Since mam canpounds are tabulated in the literature in form of relative retention times, the need did arise to transform them into methylene units and vice versa. The conversion is based as the linear correlation of the logarithm of relative retention times and methylene units over short ranges of time or temperature, (Vanden Heuvel and Horning, 1962 and. Homing, et al., 1%8), presented in Figure 2. Based «1 relative retention times (am) of 2 compounds fran litera- tIn-edataandtheirm-veluesmeasurcdonone's owninstrunsntwiththe appropriate colunn packing, the following relation serves for conversion: 1% “,0. -. m REL! 8 “1,3 - m“ I“ m,‘ " M m,b M1,!» - MU,b 4.- RRT log Reloh‘ w Retention firm a- n I I I 1 s b ' a =MU Hethylene Units Figure 2. Correlation of relative retention times and methylene units. whichwas rearranged to: 111,: . 10,. - (um - 111,1») 1.518 m» ' 108 mm: (1) log mm: - log m,b ma: - antilog [log ma - (log me» - log smut) 311,9. '— "3“] (2) m,e - sun) with equation (1) literature values could 'be cmverted to an indi- vidual situation, whereas equation (2) served to convert measured m- values to the corresponding relative retmition tithes of other inves- tigations . 'me latter possibility shows that, with a relatively few standards, it is possible to use the information contained in other plblicationl such as Kovats (1958), Knights (1%6), Item, et al. (1968), Patterson (1971) . The following 2 examples serve as illustration. a) cholesterol and lanosterol were used as standards for which Patterson (1971) indicated the relative retention times of 1.00 and 1.66 on a 3% 33-30 column, whereas with a 3% SP-2100 column correspond- ing til-values of 30.33 and 32.36 were found. Using equation (1) the relative retention of fl-sitosterol of 1.63, (Patterson, 1971), is con- verted to an Ill-value of 32.29. This is exactly what was measured on the SP-2100 colum. b) Using cholesterol and stignasterol with respective man": of 1.00 and 1.102 and W's of 30.33 and 31.72 as steam, the III-value of campesterol of 31.143 transforms to a corresponding relative retention of 1.323 by the use of equation (2) which compares quite favorably with Patterson's (1971) value of 1.30. 80 lbs power of this interpretation technique is extremely remarkable considering the fact that the lax-values were all measured in 2 C/min temperature programmed runs, whereas the relative retention times re- ported by Patterson (1971) were measured isothermally. It should also be noted that this interconversion of data does not require that the consulted literature provide relative retention times of n-alkanes. As long as methylene mite are determined for reference caspmmde which are close to the mlmown(s) and whose relative retention times are given in the literature, a sufficiently precise correlation can be made. Since steroid where and metlwlene mite are based on the same principle, it is easy to convert methylene mite to steroid numbers. me III-values of androstane and cholestane should be measured on the column used for evaluation, using the correlation of Horning (1968) for transformation. The real advantage of steroid nmbers lies in their use for identification on the basis of contributions of specific fmc- tional groups. Mass spectrometry should, however, be a more revealing technique. Gas -liquid chranatography was used for the separation of sample mixtures. Glass columns were used and silanized to minimize adsorption and degradation of steroids in the columns. A procedure for packing and conditioning of columns and to adjust carrier gas and hydrogen flow rates was adopted to yield optimum separation conditions. Two GC-pack- ings were selected, SP-2100, separating according to size and. configura- tion of molecules, and SP-2h01, separating according to fmctional sub- stituents on molecules and also cholesterol and cholestanol. Ethyl acetate and 10% methanol in benzene were selected for introduction of 81 senile material into the GC because of suitable solubilizing capacity and relatively small solvent front . A convenient sample concentration was chosen. Overloading of GC-columns was investigated to yield maximum information from trace canpounds and the limits were established. The tanperature program and the coating percentage for each of the 2 colmms were adopted for convenient screening of. individual TIC-band extracts, efficient separation, and complete elution. A derivatization procedure was adopted for use with thermally labile canpounds and for specific separations. Preparative 00 was investigated and the mathematical formulae derived for easy interconversion of relative retention times and metllylene mite. Mass Spectrometry mes spectrometry was used to reveal the purity of canpomds separated by GC and to help identify the compounds. The operation of the mass spectrometric facility at Michigan State University was described by Sweeley, et al. (1972). Samples,prepared as for regular (EC-analysis, were submitted to the technicians responsible for the m 9000 60-16 low resolution and the Varian CH5 (BC-LB high resolution instruments. Only the SP-2100 type GC-packings could be used for steroid analy- sis on these combination instruments. Although quite sane information relative to 60 column packings is provided by suppliers in form of Rohrschneider constants , Rohrschneider (1%6) , and about the field of application of the colmun packings, no indications were foundwith respect to the suitability of column packings for cc-nn instruments. lhus , after determining the separation conditions for SP-2h01 on an 82 individual GC, this packing material could not be used on the GC-MS instrument because the technician detected long term residual contamina- tion in the m-part when using SP-2h01 packings. Unfortunately, no equivalent column with a different packing was available which was compatible with the local instruments. This is one of the reasons why preparative 60 was introduced into the overall procedure. A parameter of actual elution temperatures for GO packing material is also not provided by suppliers. As an example, cholesterol appeared in a 2 C/min program on 3% SP-2100 at 260 0, whereas on at sp-zhOl it appeared at 210 C. Thus, the indication in the catalogs of suppliers of a high operating temperature limit, which for these particular phases were 350 and 275 C respectively, is no guarantee for an easy GC-elution and the indication of bleed rates at equal temperatures is absolutely irrelevant. Fran an operational point of view, the SP-2h01 column with its significantly lower temperature requirement is certain- ly easier to handle. It seems that this information is not contained in the Rohrschneider constants, since for each of them the SP-2100 values lie much lower than for SP-2‘401. Analysis of trace compounds and of samples with overwhelming cholesterol content requires careful. attention to details in the GC-MS operation. The objective is not simply confirmation of known canpounds but the identification of mknowns in a complex sample which requires that the instrument combination be run under optimal conditions for which a series of precautionary measures are recomended. Their respec- tive reasons are given. After the usual calibration procedure with helium, hexane, and perfluorokerosene, the mass spectrometer should be focused on an ion 83 with an m/e-value close to the molecular ions expected in the samples. This important step is equivalent to optimizing the sensitivity of a Gc-instrmsent by adjusting the hydrogen flow. Since the same GC-colmns are used for investigation of various canpomds of widely differing nature, the flow-rates are average and not optimized for a particular need. To insure that a sufficient plate number is available during analysis, the separatim efficiency is checked by injecting a mixture such as cholesterol and A7-cholestenol or desmosterol when using SP-2100 or cholesterol and cholestanol when using SP-2h01. It is important to use a steroid pair which is diffi- cult to separate. This prerm should be conducted mder cmditions identical to those under which the samples will be evaluated. Every aspect of the GC-part of the instrument with respect to operating conditions becomes thus apparent and allows for necessary adjustments. 'lhe amount of standard material injected should be kept as low as or lower than the amount of the unknowns of succeeding runs to check for sufficient sensitivity which is essential for trace canpound analysis. Taking mass spectrometric scans of the trial steroid pair insures that the instrument is operating properly and that all interfacing de- vices and the computer are also operating correctly. Examination of the 2 spectra will reveal the condition of the column and the extent of interfering background, both of which are especially important in trace compound analysis . When no thermal degradation occurs, the 14-15 and M-18 peaks of cholesterol should be about equal in size, and with cortisol includedas athirdstandard, its K-60andm/e 163 ions shouldbe minimal with careful operation, (Mildewicz, 1972). 8h For the analysis of mknown GC-peaks, the repetitive scan mode, (Reimendal and devall, 1973), should be used over a reasonable m/e range, e.g. not higher than 500 for free steroids, to obtain maximum scanning frequency, and the various scans of a single peak should be examined for homogeneity. If there is doubt with respect to purity of a peak, chranatograms should be retrieved from the computer by the ion search method. The intensity of ions in the big: molecular weight range are plotted against scan nmbers and, for a pure compomd, they should all peak over the same scan umber, mereas for a mixture there will be several scan numbers over which various ions peak and, thus, reveal nonhanogeneity. me availability of several m-scans per peak permits an assess- ment of possible 'bias' in the spectra due to changing ion source con- centrations and pressures, (Brooks, et al., 1%8). For each unknown GC-peak a UV-trace should be taken. This trace provides a means for correction of computer misassignments . lhe trace reveals also metastable ions which are extrmsely valuable in determin- ing fragsentation patterns or in elucidating rearrangement pathways , (Beynon and Caprioli, 1972). When analysis by low resolution has shown that the peak(s) of interest is pure, a GC-high resolution mass spectrum should be obtained with a table for the elemental. composition of the molecular ion and of some of the characteristic ions of the spectrm. For mass spectrometric analysis no specific methods had to be de- veloped but control measures had to be applied to insure optimal opera- tion of the 6046 combination instrumnts. These measures served to insure sufficient separation efficiency, detection sensitivity, a 85 properly conditioned column with minimal interfering background, mini- mal thermal degradation and proper operation of all deVices. The repetitive scan mode was adopted to assess the purity of individual GC-peaks and it is recommended to obtain 18 [IV-traces of individual peaks to correct computer misassigmnmts and to obtain information about molecular fragmentations and rearrangements . Miscellaneous Methods Recmtallization of Cholesterol The cholesterol standard was recrystallized before use. The method of Fieser, reported by Cook (1958), was used in which the cholesterol was dissolved in hot glacial acetic acid and left standing overnight at h C. 10 mg cholesterol per ml acetic acid was a suitable concentration for easy hot filtration and efficient purification. At this concentra- tion long crystals developed, which were vacuun filtered on Whatmn No. he filter paper in a Bllchner funnel. The crystals required rather long dryirg times in a vacmm oven at 100 C to completely eliminate the odor of acetic acid. Cholesterol was also recrystallized in absolute ethanol, ( Anon. , 1965). The sterol was dissolved in ethanol at 10 mg/ml in a centri- fuge tube and distilled water was added dropwise with intermittent resolubilization of precipitated material by means of a sonicator and a hot water bath. It took about 50 drops of water frm a Pasteur pi- pette to a 7.5 ml cholesterol in absolute ethanol solution to achieve saturation at steam bath temperature. After overnight crystallization, centrifugation at 2,500 rpm, and rmnoval of the supernatant with an aspirator, the crystals were dried in vacuo using freeze-drying 86 equipnent. To avoid splash-out, the tubes were tilted to horizontal or near horizontal position. Recmtallizatig of Yolk Extract and of Nonsa fiable interial To prevent thermal degradation during saponification, recrystalli- zation behavior of the mixt-solvent extract and the nonsaponifiable material from egg yolk in various solvents was investigated. The extract was taken to near frothing on the rotary evaporator, benzene was added , and the solution reconcentrated. 'nle concentrate was diluted with benzene and convenient aliquots were transferred into centrifuge tubes in which they were frozen at ~18 C and then freeze-dried. Freeze-dry- ing, after transferring the extract into benzene, was the only method found to obtain a dry yolk extract. The dry nonsaponifiable material was dissolved in benzene and treated the same way as the extract. To each fraction of extract or nonsaponifiable material a specific volume of various solvents - chloroform, benzene, ethyl acetate, hexane, ether, acetone, ethanol, methanol - was added and the following observatims were made: chloroform is not suitable for recrystallization because it has an extremely high solubilizing power, and benzene because of high solubilizing power and its high freezing point (+5.5 C). Yolk extract did not dissolve canpletely at 10 mg/ml methanol or ethanol; there was sane oily residue. Acetone dissolved the extract caspletely, but a precipitate acctmulated which did not disappear by warming on the steam bath. 'nle precipitate could have been phospho- lipids according to Hertelendy and Cannon (1965). Ettnrl acetate, hexane, and ether dissolved the extract completely. After overnight standing of the yolk extract solutions in the freezer at ~18 C, the 87 amount of visible precipitate provided data to derive the following solubility sequence for yolk extract in the various solvents: hexane > ether > ethyl acetate > ethanol > acetone > methanol. This indicates that fat fractionation of yolk extract could be achieved by recrystal- lizing it first in hexane, transferring the supernatant into ether, and repeating the recrystallization. These recrystallizations of material in the supernatant would be repeated following the solvent sequence through methanol. Each crop of crystals recovered fron the solvents would represent a fat fraction. Fat fractionation could also be achieved by proceeding through the solvent sequence in the opposite direction, thus extracting the yolk fat first with methanol and then, after centrifugation and recovery of the supernatant, repeating the extraction of the residue or precipitate with acetone, ethanol, etc . , through hexane. In this case the material recovered from each super- natant would represent a fat fraction. Working with a concentration of 2 mg nonsaponifiable material per ml solvent the following solubility sequence was derived: ether > ethanol > ethyl acetate > acetone > hexane > methanol. Thus, to circmn- vent hot saponification the solvent pair, hexane and ethanol, appears most praising. It seems that nonsaponifiable material could be preci- pitated from the fat extract by recrystallization in humane. the super- natant could be transferred into ethanol from which the fats would be precipitated. ‘Ihus , 2 nonsaponifiable fractions, the hexane precipitate and the ethanol supernatant, would be obtained which should contain all nonsaponifiable material. This method was, however, not evaluated. Liquid Column mm Liquid column chrolatography of the fat extract or the nonsaponifi- able material with the equipnent and instruments avilable was highly unsatisfactory. It seems, however, that high pressure liquid chroma- tograptw of the most recent design will, in the future, replace at least the TLC separation in the kind of research described here. At this time the real weakness of EPIC-systems are their unsatisfactory detection devices for nomad. steroids. KPH: has several advantages: the sample is always in solvents and, therefore, not exposed to dry layer material nor to oxygen and other gases. The elutions are usually rather short, thus minimizing the contact time between layer material and sample which is highly desirable to avoid destruction of steroids by the layer material, (Lisboa, 1969). The recovery of individual fractions from a EPIC-system is done in a fraction of the time required for the recovery of material separated by TIC and is much purer. In this aspect it may even replace preparative CO with its possible thermal degradation. Removal of Free Fatty Acids The method of Capella, et al. (1960) for removal of free fatty acids was adapted to egg yolk nonsaponifiable material in the following way: Cu(II)-CO3 corresponding to Q- the weight of nonsaponifiables to be treated was placed into a centrifuge tube of 20-fold volume. Ample chlorofom was added, the copper carbonate was suspended, and then centrifuged down. The supernatant was removed and the washing repeated if it seanei advisable. The nonsaponifiable material in chloroform was added to the centrifuge tube, the tube vigorously swirled, and then centrifuged. The nonsaponifiable material was recoverd and the copper 89 carbonate washed with an appropriate amount of chloroform which was recovered after centrimgation. PROPOSED OVERALL mm TO EVAIUATE CHANGES IN EEG STEROID CNPOSITION The individual methods described above were evaluated and combined to result in a cohesive procedure. It is proposed that this procedure be used for the analysis of egg steroid canposition and its changes. The analysis is based, at least initially, on a comparative basis and, therefore, eggs laid under conditions which are considered 'normal' should be used as reference material. Control and experimental eggs should be extracted and saponified separately but in parallel as described in the sections on extraction and saponification. The first significant information may be obtained, at this point, by canparing the values of the percent weight (of non- saponifiable material based on the liquid yolk. A mall sample should be drawn from each batch of nonsaponifiable material in chloroform and transferred into ethyl acetate or 10% metha- nol in benzene. The 2 solutions should be injected on the SP-2100 and SP-ZhOl columns, using the temperature program described on page 71, and the GC-traces should be carefully compared. The amount of cholesterol in the liquid yolk of the control egg and the experimental egg should be calculated based on the size of the 2 GC-peaks. This provides the information if there was a major change of the egg steroid composition. A sizeable difference in arm of the ratios of individual peaks to the cholesterol peak between the 2 samples should be considered significant. When the peak(s) of different relative size is strong and sufficiently 90 91 separated fron cholesterol by the SP-ZlOO column, a direct mass spec- trometric analysis of this peak should be obtained. If the peak sep- arates only on SP-2h01, it should be collected by the preparative GC technique described and then submitted to SP-2100 6046 analysis. The MU-value of each peak submitted to m-analysis should be measured, as well as the til-values of androstane and cholestane for the columns used. This will allow the conversion of LII-values to Sit-values for functional group analysis, (Vanden Heuvel and Hornim, 1%2). Differences visualized on the total nonsaponifiable fraction of egg yolk can be considered to be of major significance. To visualize smaller variations , the nonsaponifiable material should be subfraction- ated by TIC. Applying the control nonsaponifiable material on one half of the silica gel G plates and the experimental nonsaponifiable material cm the other, the plates should be eluted with the 3 solvent systems and visualized as described in the section on thin-layer chrmatogram. The macro recovery procedure should be applied in parallel to the con- trol and the experimental material for the appropriate area on the plates. The GC-evaluations should be performed, as described above for the nonsaponifiable material. To establish peak ratios, any well- sized peak appearing in both traces may be used for reference. It should be established, however, whether the reference peak itself occurs at varying concentrations . Differences visualized by GC should be correlated with the TIC patterns. Before submitting am of the peaks to mass spectrometric analysis they should be well enough separated from other peaks and fran each other. Peaks separated on SP—2h01 only should be collected from the GC, others can be submitted directly to SP-2100 GC-MS analysis . At this point of the analysis, TIC-bend extracts which are complex should be established. Good criteria are: more than 5 unknown peaks per GC-trace or that corresponding peaks on the SP-2100 and SP-eltOl traces can not be correlated without applying preparative GO. A further subfractionation usually yields more information about trace substances and, to subfractionate coupled: bands, the particular elution should be repeated, using now the whole width for the sample material and the micro-recovery procedure with sufficient overlapping for the particular band under investigation. For trace camound analysis all 3 plates should be subdivided over the appropriate area and treated according to the micro recovery procedure. The GC-anslysis of the micro-band extracts should be performed and the decision on tB-analysis of individual peaks should be taken as described above for the nonsapo- nifiable material and the macro-band extracts. For fractions showing canpound degradation or excessive adsorption on the 60, an approximately correct quantity of the fraction should be derivatized according to the methoxy-trimethylsilylether-derivatization procedure. When mass-spectroletric analysis reveals superposed com- pounds in a specific peak, derivatization should be tried to separate these canpounds. For cholesterol fractions showing trace compounds which are not sufficiently separated frcm cholesterol, the repetitive, short time digitonin precipitation procedure should be used as described in the section on digitonin precipitation. When a specific compound can not be separated sufficiently by either derivatization or digitonin precipitation, separation by 93 AgNO3-T'IC, described in the section on thin-layer chranatography, should be attempted, using reference material in parallel or at least choles- terol standard material for cholesterol fractions. The mass-spectrum of a particular GC-peak should be searched for the molecular ion which may have to be deduced. The identification process of this peak should be started with this ion. Based on the molecular weight of the ion, corresponding steroid molecular formulae can be obtained fran Table 2 with an indication for possible occurrence. Since the table is incomplete, all the steroids should, however, be considered as possible candidates for the GC-peak. This first selection of compounds based on the molecular weight should be narrowed down by examining the ill-values of the particular peak. Using the ill-value measured on 8P-2100 the number of carbon atoms in the molecule should be determined by comparison with steroid stand- ards , whereas the III-value obtained with the SP-2h01 collmm should be interpreted with respect to the number of hydroxyl groups or the presence of a ketone. This may either be done by running corresponding standards on the CC or consulting relative retention data in the liter- ature as described in the section on gas-liquid chranatography. If a cc-m high resolution instrument is available, it may be used to obtain the elemental formula of the molecular ion and of specific ion frac- tions. The isoneric configuration of the compound will, however, not be known at this point of the identification process. After this first elimination process on the basis of CC retention data, the elution behavior of the remaining steroid possibilities should be compared with the relative position of the TIC-band fron which the peak under investigation was derived. This may either be done by eluting assoc cannon no assess . nave 850 «o hogs naps sump—ohm no nuns—E « and go no RUDE—.5 u .ANRH .583: one toned 38m 33 3368 one no .8me :83; Johanna one bong no response So can define 5 ocean afloat». 33s: 300.38 OOQMNN Hflama NNOMON... NHNMHN MNQNON MHOMHN :NwNON +~ HmNHN Odamom commas Hammom NHOMON MHQNON :Hmwom. oozmom momotmomotmowotmomosmomos m 1.. JMNHO £MNHO o m ommmde 8N emu... i .532 dong m w one nosom cannon : w 5.3 smog one .5906 .8930 no ooaoosoo schooner. no.“ unease .3983: an nonsense soaonosn so essence seasons: .N «Sea cntd. ka2. * C H 0 B * C H O B * C H 0 B * C H O B * C H O B * C H O B MN a3610 JMNHO NNNNN (DON-:70 Nmmmm 33383 ***** JMNHO 2h h2 O O JMNHO HHHHH 23% fi 23% 23% a :mme mmwmm °$$33 fififififi JMNHO mmmmm CD NONfiSW 38538 * JMNHO HHHI-ll-I $3332 JMNHO NNNNN "300003;!" #:fiafl': #:d—‘J’efl' NNNNN JMNHO NNNNN N \DCDO m mm: mmmmm NNNNN JMNHO mmmmm ON <0 mm gm fififififi NNNNN JMNI—‘IO mmmmm N—‘JQCOO mmmm: mmmmm NNNNN :MNH :znz °%i% fifififi :J’MN mmm (DON Nmm Had *** :JMNHO .d'MNHO O arsrsrsrxr U\U\U\U\U\ C) 8®O .3 \080 .3 3 8888 O80 8 :S'MNHO O JMNHO U\U\U\U\U\ C> Firiririri 8888 3 32$ 8 88888 8 8888 cntd. Table 2 . *CHOB*CHOB*CHOB*CHOB*CHOB *CHOB M4 JMNHO :MNHO :mNHo o JMN mmmmm :znnm mmmmm o HHH : mo : mON no mgm: mgmqn mmggg fi 333 mans: arms: ::::: o 000 NNNNN NNNNN NNNNN m mmm an a: a: a: a: :MNHO :MNHO o JMNHO :m 4.3.3.1.: mmuuntn O u-h-lu-lo-IH NN N mo N mmo N ON 8888: 88 a 888um 88 mmmmm mmmmm m mmmmm mm NNNNN NNNNN N NNNNN NN **** o :mNHo o :mNHo :mNHo 4 : mmmmm o HHHHH NNNNN m o 8 888 8 3§33° sessa s NNN m m m w 8 888 8 88888 88888 8 * ***** * HO 0 :MNHO :MNHO :MNHO mm o HHHHH NNNNN mmmmm no ON ON mm 3 23:33 .e.-:23??? #:jgg r up» NNN P N am N NNN NNNNN 5555 * ***** ** * ** :mNHo :MNHO :MNHO :MNHO HHHHH NNNNN mmmmm mane: QON 0N QON QON massif? fissfifg Mddjg «3:84:33 0 0 88888 888 88888 8888 * :MNHO :mNHo :MNHO smNHo NNNNN mmmmm ::::: mmmmm mON mmON QON ow N out: mmddfi mm: 00003.: Inmlnlnln mmmmm mmmmm U‘Mlnln NNNNN NNNNN NNNNN NNNNN cntd. Table 2. * C H 0 B * C .H 0 B * C H 0 B * C H. 0 B * C HI 0 B * C H O B MW HO Flt-i N in 00 mm Nl-IO NNN 00] run O\O\O\ NNN MNHO mmmm £838 ooaooo <(x\>lNNN JMNHO 23.3.3.3: 3§§83 85835 * ..‘JMNH NNNN 838% 000 mmm JMN comm $83 00m NNN 31560 0 .II'MNFIO thinlhlhlh 83§§8 88888 32 58 O O :MNHO POI-I O Fit-IF! N: mmm 33838 etc. 98 standards on corresponding thinrlayer plates or by'consulting the appropriate literature, (Lisbon, 1%9, Neher, 1%9). For a final identification, the mass spectrum.of the unknown.peak should be compared.withumass spectra in the literature. It may‘be necessary to make the corresponding derivative of the fraction and to obtain an additional mass spectrum of the derivatized unknown. It may happen that.mass spectra of the final choices can not be found in the literature, then appropriate computer libraries should‘be consulted, (liming and Foster, 1972). At this point of the identificaticn process, only 1 or 2 compounds remain as possible compounds. If available, samples should be obtained.and.sub3ected to:mass spectrometry'for'comr parison. The interpretation of the mass spectrum of the unknown.GC+peak may also be attempted on the basis of information given‘by Knights (1967), Budzikiewicz (1972), Brooks, et al. (1973), and Brooks and Middleditch (1973), and especially in the series of articles 'mass spectrometry in structural and stereochemical.problems' by the group of DJerassi. Cleaning of Glassware The glassware used for the analysis of steroids occurring at trace level concentrations should be absolutely clean. The following proce- dure was used in this study. All glassware went first through normal laboratory washing proce- dures and, after drying, underwent a special acid treatment. Sodiu- dichraaate-concentrated sulfuric acid solution was prepared, according to Anon. (1963). Large items such as separatory funnels, condensers, graduated cylinders were subsergsd in the acid solution in a large container which was sitting in a sink with hot, overflowing water. Usual residence time was at least 8 hours. Small items such as Q dram vials, special Pasteur pipette vials, stoppers were treated in appropri- ately sized beakers directly on a steam bath for at least 2 hours. After the acid trea‘hsent, the glassware was carefully drained and rinsed 3 times with distilled water. Then, after careful washing with 0.111 K01! they were rinsed with distilled water and dried. The acid treated glassware was always stored separate from non-treated items. Before use, all surfaces casing in contact with the sample were flushed with chloroform-methanol 2 :1, to insure that no dust or other air-borne material could cause contamination. 99 1% When sodium sulfate was used for drying of solutions, it was first flooded in the fritted glassfilter with either chloroform-methanol or ether. Purification of Solvents It was established by gas chrmtogram that reagent grade sol- vents such as benzene, ethyl acetate, etc., are not pure. Therefore, all solvents used in the analyses were subjected to the following purification treatments : Methanol, ethanol: similar to the method used by mm (1961), the solvents were refluxed 2 hours over ms and then distilled over. Chloroform: according to Goodspeed and Millson (1%7), the solvent was distilled over frcn anhydrws calcium chloride. Except for the chloroform used for m, it was stabilized by addition of approximately 1% methanol, (htenmann, 1%1) . Diethyl ether: first the method of Adlercreutz and Imminen (1%8) with ferrous sulfate was used for purification, later the method of mm (1969). Ethyl acetate: this solvent was refluxed for h hours over phosphorous pentoxide (P205) , then distilled over. Benzene: it was purified according to the method of Henick, et al. (1951+). ‘me refluxing period was, however, extended to 12 hours. Acetone: it was simply distilled over fran anhydrous calcium chloride. Hexane: according to 'I'hmpson (1911+), the hexane was treated with concentrated sulfuric acid and potassium permanganate, washed to neutrality with distilled water, dried over entwdrous sodim sulfate, and distilled over. 101 Pyridine: it was either treated according to Sperry (1963) or silyla- tion grade quality was purchased (Pierce Chemical Co.,Rockford, Illinois). For all these redistillations, Vigreux 1&7 cm distilling, fractiona- ting columns were used. The first 50 to 100 ml of the distillate as well as the last 50 to 100 ml of the residue in the distilling flask were discarded. The solvents were never left standing over the reagents for too long periods of time at laboratory temperature and were dis- carded when they appeared abnormal, such as seen for chloroform after extended periods over CaClQ or for methanol and ethanol over KDH. All these solvents were stored in a cooler in the dark when not in use. Aliquots, corresponding to volumes used in the overall proce- dure, were concentrated to small volume and evaluated by so for residues, as recommended by Homing, et al. (1%3). Drying of Nitrogen and Use of Plastics The technically prepurified nitrogen was used for all applications. As a precautionary measure, a drying tube filled with anm'drous calcium sulfate and some indicating Drierite (W.A. Hsmond Drierite Co. , Xenia, Ohio) was inserted in the delivery line. Whenever the sample had the slightest chance of getting in contact with plastic material, it was replaced by either equivalent glass or Teflon components. This restricted the use of plastic material to hoses for gas supplies and water aspirator connections. Screw-caps were all lined with Teflon. The'size G, 1% inch diam- eter, cap size 38, bottle cap liners of Teflon were purchased (Lab Apparatus Co. , Cleveland, Ohio) and fitted the extraction centrifuge 102 bottle caps. Smaller sizes were punched out with cork-borers by placing the discs on flat pieces of wood. Before placing the discs into the caps they were thoroughly cleaned in chloroform-asthma 2:1, methanol, and acetone, sonicating the batch each time for 5 min. hereafter, they were only touched with clean tweezers. Preparation of TIE-Plates nefollowingistheprocedurewhichwasadoptedtopreparethin- layer plates . 1. 2. 3. 5 plates, 2W3 m thick, are thoroughly cleaned with Ajax and carefully rinsed with distilled water (it mustrunoffasasheet, notbreakawayanywhereonthe plate, and all detergent residue must be wiped off). Airdry the plates. Align the plates on mounting board with small strips at start and at end. Just before spreading, wipe plates with tissue danpened with alcohol to remove any dust or fingerprints. Adjust applicator to deliver appropriate layer thickness, put in spreading position. Weigh proper ammmt of silica gel into glass-stoppered 250 m1 erlemeyer, stopper, shake vigorously for 30 sec. Pour slurry into applicator. Reverse handle and pull applicator with uniform speed beyond the end strip. Ietgel sitforQ-lhour (3 -hhrs for silicagel6OER), put away for overnight storage. Predevelop the plates with methanol-ether 80:20 (v/v) , twice, letting the solvent evaporate under the hood. Renove aécmwide band of silicaoneach side andcut almgrooveacross the silica, 2cmfromtheendof the layer. Activate the plates for 1 hr invacuum oven at 100 0. Cool and store the predeveloped plates in desiccator. 103 Quantities: silica gel G 0.25 m 30 g; 320 60 ml 0.5 m 60 g; 120 ml silica gel 60 HR 0.25 an 27 8; 63 ml 0.5 m 5‘: s; 126 :1 Since the 0.5 nun-quantities are the maxim the applicator can hold, the number of plates spread simltaneously is correspondingly lower for thicker layers . Preparation of Aglll3 TIC Plates The following is the procedure which was developed to prepare AgN03 thin-layer plates . 1. 3 plates, 200x200x3.9 m thick, are thoroughly cleaned with Ajax and carefully rinsed with distilled water (it must run off as a sheet, not break away anywhere on the plate, and all detergent residue must be wiped off). Airdry the plates . Align the plates on mounting board with small strips at start and at end. Just before spreading, wipe plates with tissue dampened with alcohol to remove any dust or fingerprints. Adjust applicator to deliver 0.75 m layer thickness, put in spreading position. Weigh 60 g aluminum oxide neutral, type T into 250 m1 glass-stoppered erlenmeyer, dissolve 10.6 g AgN03 in 9) ml water in a 250 ml beaker. Turn all lights off to an absolute minumum. Pour Adv? solution into erlenmeyer, stopper, shake vigorous- ly or 30 sec, pour slurry into applicator. Reverse handle and pull applicator with uniform speed beyond the end strip. Let plates sit in canplete dark- ness for 1 hr. Store away overniglt in the dark. The followingmorning remove aficmwidebandof alminum oxide on each side and activate the plates for 1 hr in vacuum oven at 100 0. Cool and store the plates in a completely dark desiccator. 1. 2. 101!- Preparation of Micro TIC-Band Collectors The following is the procedure by which micro TIC-band collectors prepared. Plug tightly with glass-wool the end of Pasteur pipettes, up to the constricted area but not further. Place pipettes in 100 ml or greater beaker, large, plugged end down, fill beaker with chloroform, sonicate 5 min. Place larger beaker over pipette tips, inverse the whole set swiftly such that the chloroform drains into the large beaker. Repeat 2. and 3. with chloroform-methanol 2:1, methanol, and methanol-water 1:1. Repeat 2 . with acetone and check each pipette for slow drainage by pulling the pipettes out of the acetone. Eliminate those from which the solvent drains too quickly. The others are ready for use and may be dried. They should be kept in a clean place, such as a desiccator, to avoid contamination. 105 Purification of Digitonin The followingisthemethodwhichwasdevelopedtopurifycamnercial digitonin. 1. 2. Weigh out 100 mg digitonin, transfer into precision-tared 100 ml conical centrifuge tube. precision tare wt. = Add 20 ml 90% ethanol, bring digitonin into solution by careful warming on steam bath and/or sonication. Cool the centrifuge tube with content, then slowly add 30 ml ether, swirl, let sit 10 min. Add additional 16 ml ether, swirl, centrifuge 5 min at 2:500 rm- Add additional 16 m1 ether, swirl to mix the liquids with- out excessive abuse on the pellet, let sit 10 min, centri- fuge 5 min at 2,500 rpn. Remove the liquid with an aspirator. Dry the pellet very carefully with dry N2; water bath at 1&5 C, theninvacuumovenforéhour. Dissolve digitonin in 6 m1 pyridine with a sonicator and slight warming. Concentrate the solution with dry R2 to approximately f; its volume . Add 60 m1 ether and hold digitonin in suspension for 15 min with the sonicator, then let stand for 2 hours in refrigerator. Centrifuge 10 min at 2,500 rpm, remove supernatant. Wash the digitonin 3 times with 60 ml ether as in 9. and 10. Repeat 7., let cool in desiccator for Q hour, weigh wt. = recovered digitmin wt. = Dissolve digitonin completely in 10 parts of 50% ethanol, place in freezer for 21: hours for recrystallization. Dry in freeze-dryer, with the centrifuge tube tilted. 106 Packing and Optimizing (QC-Columns nefollowingisthemethodwhichwasadoptedtopackGC-colms and to optimize operating conditions. 1. 2. 3. ’4. 5. 10. The clean glass-column, glass S-piece and glass wool are bahed overnight in the oven of the glass-blower shop. Soaktheglasswoolinabeakerandfillthecolmand the S-piece with 5“ dimetlwldichlorosilane in redistilled benzene (toluene, hexane) for & hour, drain. Rinse once with benzene, 3 times with methanol. Drywith nitrogen; the glass wool in vacuum oven. With the aid of tweezers plug detector and of colm with silanized glass wool (1%" deep, overbording). Heat colminovenatl500whilepreparingthestand for the column and funnel for filling. Remove colm frua oven, mount on stand with funnel ad- .justed onto injector end and aspirator connected to detector end. Pour packing material into colm while firmly tapping with rubber-covered pencil. When the colm is filled to within the last 2 inches, close the vacuum and let pressure equili- brate. Repeat heating and filling until no more packing material can be added. With tweezers plug injector and of calm with glass wool, push the short plug down onto the column bed. Install the column and S-piece with Teflon ferrules in the GC. Do not connect to detector, close if off with alumimm foil. Condition the column for MS hours at maximum admissible temperature (see catalog of supplier) and minimal flow of nitrogen carrier gas. After conditioning inspect colm for irregularities . If discontinuities appear in the calm packing, remove colm fran oven and injection glass wool plug from column, and fillsomemoreasin6. through9. 1h. 15. 17 . 19. 107 with an airflow of approndmat 300 ml/min, adjust initial nitrogen und hydrogen flaws to o ml/Idn. Make a few injections to adjust the oven temperature that the peak merges within 7 to 12 min and the sample size that the peak height is between 60 and 90% of mi]. scale deflection. While optimizing the carrier gas flow, keep the nitrogen- hydrogen ratio approacimtely 1:1. With the chart speed at 1 in/min, “a standard (cholesterol for sterol analysis) is repeatedly injected and t' and w measured, see Figure 3a. Calculate the theoretical plates, 1? - 16-(t' xv)? and plot them against the carrier gas flow rate. The flow rate with the highest number of theoretical plates is the optimum. With the optimun carrier gas flow and the recorder zeroed and balanced, set the baseline with the fine control to 50% at an appropriate attenuation. Optimize the turdrogen flow by starting with a fairly big: flow rate and decreasing it contunously, thus reproducing the diagram of Figure 3b. who indicated optimum is adopted for normal operation of the GC . . Air flows usually need not be adjusted unless the carrier gas flow is extremely low (<20 ml/min or high (>100ml/min). It is a matter of econanv to choose a reasonable air flow, see Figure Be. When the air flow is changed, the hydrogen flow should be optimized, again. v a ' x . ..a ’ 108 / Tangent! a} ’3 how 1" ' figure 3a. GC-fma (if-peak . with l" and w for theo- retical plate substation. rJ Solvent front 710" Jg‘ech'on figure 3b . .fcnsih'oiq a: {emotion of hydrogen flaw. mm:- deflection § 3 . “ figure 3:. “(My as flirtation of as“ flow. g o: lncroasing afi- flaw Figure 3. Gas flow optimization of (DC-calms. 109 Specific Materials or Equipnent For the concentration of solvent solutions a model Blichi, type KRv 65/15 rotary evaporator of the Glasapparaturenfabrik Flawil/Switzer- land was used. The evaporator traps were purchased from the Kontes Glass Co., in Evanstan, Illinois. To control the evaporation rate, a Magni Whirl, model W—llZOA-l constant temperature bath of the Blue M Electric Company in Blue Island, Illinois was used. A Thelco, Model 19, range to 200 C, vacuum oven of the Precision Scientific Corp., Chicago, Illinois served for final drying of samples or activation of TIC-plates. Weidiing was either performed on a type B 5, to max. 200 g, precision balance or a type P-lOOO, to max. 100 g, overhead balance, both from E. Mettler, Zflrich/Switzerland. Small solvent quantities were evaporated on The Meyer N-Evap, Model 1%, of the Organanation Assoc., Inc., in Shrewbury, Massachusetts. To freeze-dry samples a model lO-lOO freeze- dryer of the Virtis Research Equipnent Co. in Gardiner, New York, was used. All centrifugations were done with a model CS centrifuge of the International Equiment Co. in Needham Heights, Massachusetts. To discard supernatant solutions, liquid aspirators were assembled, consis- ting of a long-tip Pasteur pipette, a plastic hose, and a water aspira- tor. To suspend particles, dissolve samples, or clean certain devices, the containers were held into an ultrasonic cleaner, model ME h.6, of the Electronics Corp. in Anaheim, California. To mix solutions or keep particles in suspension a Vortex Genie mixer, model 8 8223, frcn Scientific Products, a Division of American Hospital Supply Corp. , 110 Evanston, Illinois was very helpful. For TIC mostly Desaga/Brinkmsnn material was used such as the mounting board, the spreadcrs, the plates, the plate storage-desiccator cabinet, the inert gas spotting box, and the spray guns. Aluminum Oxide neutral, type T, E. Merck (Darmstadt), was also purchased fran Brinkmann Instrmnents, Inc. in Des Plaines, Illinois, whereas silica gel G, E. Merck (Darmstadt), was obtained from Applied Science Labora- tories, Inc. in State College, Pennsylvania. Larger amounts of sample material were spotted with a 1 m1 all-glass luer tip Yale Tuberculin syringe frcm Becton-Dickinson, a Division of Becton Dickinson'and Co. in Rutherford, New Jersey, whereas Hamilton syringes from the Hamilton Co. in Reno, Nevada were used for smaller samples or for 60 injections. The syringes were cleaned with chlorofom-methanol 2 :1 in a Hamilton syringe cleaner, part no. 76610. Ultraviolet light absorbing or fluorescing bands were visualized with a Shannon model 50h, mm, 100 Watt uv-iamp of the Shannon Luminous Material Co. in Hollywood, California. After spraying, the side bands of the plates were developed with a heat gun, model no 201, mam1factured by Master Appliance Corp. in Racine, Wisconsin. Digitonin used in the digitonin precipitation was either certified reagent of the Fisher Scientific Co. in Fairlawn, New Jersey, or Digitonin Merck of the Merck 8: Co., Inc. in Rahway, New Jersey. For derivatizations the 2% solution methooqramine HCl in pyridine, 2% mx, was obtained from Pierce Chemical Co. in Rockford, Illinois, and the trimethylsilylimidazole fran Supelco Inc. in Bellefonte, Pennsylvania. To close the little self-built Pasteur pipette reaction vials, Teflon laminated septa, Microsep type F 1145, size HH from Canton Bianedical 111 Products, Inc. in Boulder, Colorado were used. GC-wark was performed in a model 810 R-15 Research Chranatagraph with dual flame ionization detector and an autanatic attenuator, model 50, of the F + M Scientific Corp. in Avondale, Pennsylvania. It in- cluded a built-in Honeywell Elektronik 15 Continuous Balance Potentiom- eter of the Brown Instruments Div. in Philadelphia, Pennsylvania with a model 201-8 Disc Chart Integrator of Disc Instruments, Inc. in Santa Ana, California. The coiled calm and Teflon ferrules were purchased from the Anspec Co. in Ann Arbor, Michigan, the dimethyldichlorosilane from Applied Science Laboratories, Inc. in State College, Pennsylvania, and the pacldngs from Supelco Inc., Supelco Park, Bellefonte, Pennsylvania. To measure gas flows , a 25 ml bubble flowmeter was used and a short piece of Teflon tubing pushed over the detector jet or into the castle (after removal of the anode and the ignition coil). Steroid standards were obtained fran the following sources: Applied Science Laboratories, State College, Pennsylvania Fisher Scientific Co. , Fair Lawn, New Jersey Mann Research Laboratories, Inc. , New York Serdary Research Laboratories, London, Ontario, Canada Sign Chanical Co., St. Louis, Missouri RESUIfl‘S AND DISCUSSION The purpose of this research was to develop an analytical procedure to follow changes in egg steroid canposition. The overall procedure which was developed in this study for use in such investigations con- sists of techniques which were adapted to the particular analytical problems encountered with eggs such as the similtaneous high-protein, high-fat content of egg yolk, the extrane predaninance of cholesterol in the yolk nonsaponifiable material, and that other steroids occur only at trace level concentrations. To allow for trace compound analysis the methods were conceived such that contaminants frcm extraneous sources can be excluded or eliminated. Observations on individual methods with respect to specific tech- niques, requirements, or decisions can be found in the corresponding sections. This part covers only those aspects which pertain to the overall procedure. A midst-solvent extraction procedure in a 200 ml centrifuge bottle was developed to free the egg yolk of protein. In this method a single, entire yolk could be transferred into the, bottle without loss of yolk material. The repetitive extractions were all perfomed in the same bottle. Thus, no loss of yolk material occurred. Only minimal amounts of solvents are used for extraction, yielding maxim amounts of extract obtainable by mixt-solvent extraction techniques. 'nle extraction 113 procedure is rapid and does not cause oxidation nor thermal degradation. Saponification was used to free the yolk extract of glyceridic material. The method was adapted from the AGAC (1965) saponification procedure for oils and fats to accanodate mixt-solvent extracts and to make it independent of sample size. The use of pyrogallol and of a nitrogen blanket was included in the method to prevent oxidation during saponification. By repeated saponification it was established that the method saponified the yolk extract completely. Free fatty acids were originally removed fran nonsaponifiable material by the method described in the section on removal of free fatty acids. Once the appropriate saponification conditions and extraction volumes were established, free fatty acids could not be detected in the nonsaponifiable material and, therefore, this step was anitted from the proposed overall procedure. By eliminating this additional treatment, the procedure was not only simplified but also avoided the uncertain effects of copper carbonate on some of the labile compounds present. The amount of nonsaponifiable material on a wet yolk basis was quite stable and represented 1.h to 1.5% 1y weight. A typical GC-trace of nonsaponifiable material is shown in Figure h. It was obtained on the 3% SP-2100 column after MO-TDB-derivatization. The peak with 128 x attenuation represents cholesterol and clearly demonstrates its predaninance. To obtain sufficient amounts of material of the mall peaks seen in this figure and of others which were not concentrated enough to even appear, preparative TIC for egg yolk nonsaponifiable material was de- veloped. Since a single elution system could not resolve the complex mixture, 3 systems were worked out, based on varying polarity. By non- polar elution, shown in Figure 5 , non-polar canpounds were further 11h .oouggmuoougnoz 30308 . as n x ..rH 0 cadmium u EH00 Haas 30 no mo sadness.“ odpsausoasanos Hana» onmmuo massive .: whom?“ Figure 5. Preparative TIC, nonpolar elution. 0.5 m silica gel G, 5 n3 nonsaponifiable material per cm, solvent chloroform, spray 0.9% I2 in methanol, in UV-light. 0 = origin; B = cholesterol band. 1.16 resolved. Figure 6 shows the pattern of medium-polar elution, which gave best resolution for canpounds with Rr-values similar to choles- terol, whereas the third system allowed to separate the polar canpounds and is shown in Figure 7. To recover canpaunds which were separated by TIC a solvent sequence was developedwhich extracted canpounds franeventhemastpalarthin- layer bands. A procedure to wash these extracts was developed and adapted to centrifuge tubes thus making it easy and rapid. This washing process removes water soluble contaminants which may be extracted from the thin-layer material and also eliminates silver nitrate fran AgN03- TIC-band extracts. nae techniques related to TIC are described in the section on thin-layer chromatography. A GC-trace obtained with the 9% sir-21m. colulm is shown in Figure 8. It represents the extract from a TIC-band of 9 in width, 2 on ahead of the cholesterol band after nondpalar elation. It clearly dmstrates the canplexity of yolk nonsaponifiable material once the predaninant cholesterol can be separated out. A similar canplexity of biological samples was found by Wyllie and Djerassi (1%8) who stated: "in our hands , most of the naturally occurring sterols of narine origin - repor- ted as pure compounds in the literature - were proved by mass spectra- metry to be mixtures of 2 to 10 canpanents." Canparison of Figure 8 and Figure 9, which shows the same extract run on 3% SP-2100, shows the difficulty of correlating corresponding peaks simply on the basis of peak height. When fractions of such a canplex nature were submitted to mass spec- tranetric analysis, the spectra showed trenendous overlapping of a number of canpounds , thus inking detection of molecular ions and 1.1.7 Figure 6. Preparative TIC, medium-polar elutian. 0.5 m silica gel G, 2.5 mg nonsaponifiable material per cm, solvent ether, spray 50:50 acetic-sulniric acid, left side sprayed additionally with 5% phosphomolybdic acid in ethanol, in regular light. 0 = origin; B - cholesterol band. Figure 7. Preparative TIC, polar elutian. 0.5 m silica gel G, 2.5 mg nonsaponifiable material per cm, solvents chloroform-methanoldwater 90+10+1. After spraying the left side with modamine 6G and the right side with 0.5% 12 in methanol, the plate was sprayed with 50:50 acetic- sulfuric acid, in UV-light. 0 = origin; B = cholesterol band. 119 E N a ..C w "EH00 4335 E so snaffle Boos s so tornado .m page , I h .0 . . . . 03. it! .0.-.‘.I. l...¢s!-..%..'ru n. a h . 7“» .x. .....Jv. 1.0 r . , .... i ‘ ‘Oumiico i: _ . V we T. i . .. i . .I . 3 . . I l 1.. r . .... . 1 . ,. h n 1 _ . . I i . ’.' I ..’.I~‘IID-IIII.-IDD.. 120 . as a x pm 0 «EH00 gamumm fim do meanings 0.59: a Mo ocean-0o .m 0.3m?“ 121 interpretation of the spectrum as a whole virtually impossible. Figure 10 represents a scan obtained on the same fraction shown in Figures 8 and 9. The molecular ions at m/e #28, 1:26, hlh, 1412, 398, and 38h may all be postulated and, thus, make the spectrum practically worthless. For this reason, the technique of micro TIC-band collection was devel- oped and is described in the section an thin-layer chromatography. The effectiveness of the micro-technique can be seen in Figure ll which shows the extract of a micro-band firm a different TIC-plate but from within the same area from which the extract of Figures 8 and 9 was obtained. Besides a drastically reduced complexity of individual TIC- band extracts, an advantage of the micro-technique is that canpounds are not confined to single fractions. This means that,on GC-traces of consecutive bands, individual peaks appear first small, develop to a maximum, and redisappear over several fractions. For a double peak, such as the one seen in Figure 11, this meant that the earlier eluting peak appeared practically pure on the GC-trace of an extract obtained from the plate, 14 fractions closer to the cholesterol band, whereas the later eluting peak was highly predominant l fiaction mrther. Thus, with the micro-technique, the position of the maximum concentration of a particular canpound can be exactly located an the TIC-plate and the fraction which is best suited for mass spectrometric analysis of a particular compotmd can be selected. Figure ll may also serve to explain the usefulness of preparative GC. The degree of resolution of the 5% SP-2h01 column in the individual GC could not be obtained on the GC-MS instrument and, therefore, mass spectra of a mixture were obtained which are difficult to interpret. The 2 canpounds could not be separated better by methoxy-trimethylsilyl .Eoouaa 983 o «o x398 a mo nonunion m2 .3 are {a one 8,. sum .PbPPPPPt—fbb? Phhrbt \‘ .- fi 'l 1 M. a so...” mmm o4 i V r o5 a2 on 8~x.mu.luqm mm ...:uhu muqzumm m. 219w vxuemuho Ion relative intensity J23 Figure ll. GC-trace of a micro Tic-bend on 5% 39-21401. column: 6 ft x 2 m 121+ ether derivatization. Assuming 2 compounds elute to exactly the same position on thin-layer plates , it would not be possible to separate than by the micro TIC-band recovery technique. The mode of preselec- tion of individual GC-peaks could not be applied successfully to TIC micro-fractions in which overlapping compounds of similar GC relative retention times occur. Enriched fractions of the front and tail peak could, however, be collected by preparative GC and, thus, mass spectra of reasonably pure canpounds could‘ be obtained. The technique of pre- parative 60, although it could not be applied successfully due to an unsuitable GC detector, is described in the section on gas-liquid chromatograpmr. GC-traces of micro-bands from within the cholesterol TIC-band looked alike. This may mean that the trace canpomds can not orient themselves because the predaninant cholesterol acts as a sort of liquid mase. It definitely does not mean that cholesterol breaks down in the procedure on either the TIC-plates or on the GC since this was checked with cholesterol standard material. It is believed that repetitive, short-time digitonin precipitation is a good approach to enrich trace canpounds of a cholesterol band. The sample size in the developed method can be up to 85 mg and is only limited by the size of available, conical centrifuge tubes. The method was first adapted to quantitatively isolate 3-beta sterols from egg yolk nonsaponifiable material. However, for the purpose of quantitative iso- lation the method was quite futile since more than 99% of the sample was precipitated. On the other hand, it was observed that when digitonin was used at subtheoretical amounts and with insufficient reaction time, cholesterol digitonide seemed to crystallize preferentially. AlthOugh 125 this method did not eliminate cholesterol completely, it reduced its concentration sufficiently such that the trace compounds appeared enhanced. Figure 12 represents the digitonin supernatant after only 2 successive reactions at a ratio of 1:2.5 nonsaponifiable material:digi- tonin and, when compared with Figure it, shows the relative increase in trace compounds. To eliminate the possibility of contamination of the sample material by cannercial digitonin, a procedure for digitonin purification was developed which treats the digitonin similar to the precipitation method and , thus, removes possible contaminants. The methods of precipitation, purification of camercial digitonin, and recovery are described in the sections on digitonin precipitation and on recovery of digitonin precipitated material. To overcane the . predominance of cholesterol, the canbined methods were applied to egg yolk nonsaponifiable material directly and on cholesterol zones recovered from silica gel G and A3103 thin-layer plates. 'nie GC-traces of the precipitated fractions looked all essentially the same as the nonsaponi- fiable minial due to the predominance of cholesterol and were, there- fore, of no particular interest. The recovery procedure is, however, included since the precipitated material may be of interest as soon as other than cholesterol bands are subjected to digitonin precipitation. When the digitonin precipitation and recovery procedures were checked quantitatively, more than 95% of the starting material could be recovered in the supernatants of the 2 procedures. To enhance epimeric effects, methooqr-trimethylsilylether derivati- zation may be applied. Ellie method of Thenot and Homing (1972) was adopted because the authors investigated the derivatization conditions thoroughly, and it can be applied to ketonic as well as hydroxylic + cholesteryl-ms Figure 12. GC-trace of the supernatant of 2 consecutive, short-time digitonin precipitations. column : 3% SP-2100, 6 ft x 15 m fraction: Bio-M derivatized steroids. ‘ Observations made while working with Agm3 TIC-plates seem to indicate that degradation took place on the plates. Since no appropriate antimcldant was found and these plates were not suitable for purifica- tion by predeveloment, it is recomnended that they be used only after repetitive digitonin precipitation or derivatizaticn have not brought about the desired result. A method of preparation was developed such that the plates were still white when used, which facilitates their evaluation during elution and before spraying. Aluminum oxide was adopted for AgNO3-TIB because cholesterol and cholestanol separated clearly in a single elation. The preparation of AgNO3-plates and their use are described in the section on thin-layer chranatograwy. The 2 GC-packings, 3% SIP-2100 and 5% sr-zlml, were selected because they showed significantly different separation characteristics . Both have sufficient plate numbers in a regular 6 ft. column; approximately 2500 and 3500, respectively. The coating percentages have been chosen such that the temperature program could be limited to a maximum tempera- ture of 275 C, thus allowing both colmms to be kept inthe 6C, simulta- neously. The elution of most of the TIE-band extracts gave nice, clean chranatograms on both columns, which was also true for the other band extracts after methouqr-trimetlnrlsilylether derivatization. Canpounds which appeared as single peaks on the 2 GC-columns were subjected to Its-analysis. The chranatogram in Figure 13 demonstrates the resolving power of the method of repetitive scanning canbined with ensueing selective ion search. In the figure, ions 126 and ’41]. peaked over scan 15, whereas ions l+12 and 1‘10 peaked over the maximum of the total ion current of this particular Gc-peak, at scan 16. Examination 128 .mohsou 3d $3033 and: neon 059330.» a «o gong—630:9 .MH shew—rm 0v 0' mm QM mm mm ma .3 m. 730w r r r P f P r 'P P -h ‘r ‘r m . _ _ a P mmv mm — "£9”. a r P 1“ r t l bl P r 1‘! n . u, a a. 9mm. u H... i l” k l? 3P! r u? r P tr a 3.. ”l r 0 a v Mm.— "300“ mxuvmana Ion relative intensity 129 of other ion-chranatograms and of bar-graphs of individual scans, such as no. 16 of Figure 11}, showed that this particular peak contained molecular ions at m/e 1428, has, hlh, tile, 1410, and possibly more. The peak is, therefore, canposed of several compounds and allowed only tentative identification on the basis of the molecular ions. The rest of the bar-graph, except for a few simple ion fractions such as M-lS, M-18, and 14-33, could not be used for identification since the contribu- tions of the various molecules to each individual bar were not known. The bar-graph of scan 5, Figure 15, shows a quite different situa- tion. Only insignificant contaminants were present and the molecular ion appeared clearly at m/e 398, with 14-15, M-18, and M-33 all present and typical for sterols. This bar-braph may serve to illustrate the identification process which is described in the section on proposed overall procedure to evaluate changes in egg steroid canposition. Based on the molecular weight of 3%, the steroid molecular formulae 02633803, cannoa, and 028344-60 with 1+, 3, and 2 double bonds, reapec- tively, the last being reported as occurring in nature, was obtained from Table 2. The MIL-value on SP-ZlOO was 31.5 and 314.0 on 813-21401. This latter value excluded dihydroxy or ketonic canpounds since they would have shown much higher ill-values on SP-zhOl. Using equation (2) and the Mil-values measured for cholesterol and lanosterol standards, the relative retention times corresponding to the conditions used by Patterson (1971) were obtained: For SP-2100 map: = antilog [log 1.66 - (log 1.66 - log 1.00) W a 1.3% a 32036 " 30033 130 .mthDH_uflERE86.mhfivnfiflnffldn“Ea .afliflumdfi {a one use 09.. \‘ r r r 50M. OWN 90m 2 i _ _ T r my r on“ 00— &m Qwux.mm.omux0¢m mm when Niuazunm w“ zquw mxmvmnko Ion relative intensity 686963 0.9m handguns.“ a mo magnum: m2 .3 953m 0\fi 8v 5.0m rbbthb-t LPh-L~bbt.LF \ J NILIJ V T 8» an.” new a r 1 .e.: l on" .... V r oa—x.mm.om.n.iuim m; mnpa music—m. ...u ZIuw MIuvmuko Ion relative intensity 332 For SP-ZhOl RRI',x - antilog[log 1.62 - (log 1.62 - log l.OO)35'29 " 3&0] = 1.20 35.29 - 33-21 On the basis of Patterson's table and taking into account that the canpound should have a molecular weight of 398 and 2 double bonds, 438(9) ,lh-ergostadienol with the respective relative retention times of 1.33 and 1.214 corresponded the closest and the next closer one was A8,2h(28)-ergostadienol with 1.33 and 1.29, respectively. By canparing relative retention times with literature values, other compounds of the same molecular weight and number of double bonds, but of different isomeric structure such as brassicasterol, 47,22-ergosta- dienol, 2h-metmrlene cholesterol and other ergostadienols were elimina- ted. A number of other ismers were not included in the list of Patterson (1971) such as Zh-methyldesmosterol and zit-methylene- A7-cho- lestenol. This problem of steroid ismerism could not be overcome by use of a high resolution mass-spectrometer, since when analyzing the molecular ion alone, all these compounds would reveal the same molecular formula of 02831.50. A clue with respect to the isaneric structure of the molecule was obtained by analysis of specific molecular ion frac- tions. ‘lhe particular TIE-band fran which this fraction was extracted was located imediately in front of the cholesterol band in medium-polar elution. Tu, et al. (1970) reported that methostenols, B -sitosterol, demosterol, ergosterol reference standards eluted to approximately this position. Table 2 in Lisboa's work (1969) indicated that less unsatura- tion seemed to slightly increase the elution of similar compounds on silica gel G, e.g. cholesterol versus cholestanol or campesterol versus J33 campestanol, which mar be applied to ergosterol and ergostadienols. Thus, the occurrence of the ergostadienols selected on the basis of relative retention times is higfly likely at this specific position on the TIC-plate. Since no mass-spectra of these 2 ergostadienols were found in the literature and no such compounds or their spectra were on file in the mass-spectrometer library, a further characterization of the possible caupound could be obtained only fran the bar-graph itself, on the basis of information contained in the literature. A quite characteristic feature of the bar-graph is the relative importance of the molecular ion itself and of 14-15 being greater than 11-18. This seems to be indicative of A7 or A8-sterols which do not possess a Zh-methylene or ethylidene group, (Knights, 1967). The aim- ilarity of A7— and A8-sterol mass spectra is well documented for compounds such as 47/4 8-methostenols, to: , 11w: .mmtml-sx -cholest- 47/48-en-3/3 -ols, or 2h,25-dihydro- 47/4 8-1anosterols, (Eneroth, et al., 1%9). A 42h(28)-double bond (methylene or ethylidene) usually causes the mass-spectrum to show a major peak at m/e 31h, (Wyllie and Djerassi, 1968) , which is not the case in the bar-gram. For these 2 reasons, the 48,2h(28)-ergostadienol could be eliminated. The presence of a double bond in ring 0, such as 48(9), suppresses the loss of the side chain with the resultant ion of m/e 269, (Wyllie and Djerassi, 1%8, Brooks, et al., 1968), whereas the presence of a Alh double bond prevents ring D cleavage and, therefore, the loss of the side chain is the most important feature, (Paul and Djerassi, 1970, Djerassi, 1970). Further interpretation of the origin of certain specific ions is based on general steroid fragmentations. Thus, it can be postulated 13h thattheions atm/e 287and285areduetothe loss ofac7-chainfran the side chain. The ion which is cmated by. the complete loss of the side chain is mind at m/e 269 and confirms the presence of 2 double bonds in the steroid nucleus. The ion at 213 is produced when the side chain plus 016 - 017 are lost fromLthe2molecule. The peak at m/e of 227 is obtained by loss of water from the ion frament at 2h5 for which it is difficult to provide a certain structure, unless a,gain of 2 hydrogens for the ion of m/e 2&3 is postulated. All these fragaentations are quite canon for sterols and are found in investigations such as by Knights (1967), Knights and Laurie (1967), Bmizikiewics, et al. (1967), T8kes, et al. (1968), Wyllie and Djerassi (1%8), Brooks, et al. (1973). Thus, based on the molecular ion of 398, 3 molecular fonnulae of possible steroids were obtained. Multihydroxy and ketonic canpounds were eliminated due to the elution‘behawior of’the compound on SPHZhOl, leaving only the 02381.60 monohydroaw canpounds with 2 double bonds . By comparison of relative retention times on SP-2100 and SPb2h01'with literature data, 2 ergostenols were retained as possible candidates. Both compounds would be expected to elute to the position on the TEC- plate from which they'were extracted and, therefore, none of them.could be eliminated based on Rf-values. By analysis of the characteristic features of the mass spectrum the ergostadienol with a A 2h (28) double bond was eliminated, whereas it could be shown that all significant ions visible in the bar-graph would be expected to occur as normal fragmen- tations induced by electron impact on A8(9),1l+-ergostadienol. There- fore, it is postulated that this cmpound is 48(9),lh-ergostadienol. Interpretation is much easier when a corresponding spectrum can be found in the literature, such as was the case for the spectra of 7|||III| J35 Figures 16 and 17. Figure 16 shows characteristic peaks at m/e 38h, 369, 3142, 327 , 299, 271, 229, and especially 12h, which match the spectrum shown in the work of Budsikiewicz (19(2) for cholest-h-en-3-one. Figure 17 shows the mass spectrun of the steroid precursor squalene which was isolated from egg yolk. Most ions correspond with the data of the Wiley 272 reference spectrmn, Anon. (1973), except that the major ion was measured at m/e 69 instead of m/e 70. It is postulated, however, that the spectrum represents truly squalene because Eneroth, et al. (1969) reported m/e 69 as prominent peak for authentic squalene without mentioning m/e 70 and the major ions such as m/e 395 (Fl-15), 368 (14-03%), 3&1 (M-C5Ii9), and 2143 (62131.1) are readily explained as normal fragments of squalene. In addition, the position of the TIC-band from which the compound was extracted corresponded with that of squalene standard reported by Tu, et al. (1970). The overall procedure and the methods to circumvent specific prob- lems encountered when analyzing egg yolk nonsaponifiable material can be used to show qualitative differences in steroid composition betvmen control and experimental eggs. The differences can be revealed even at trace levels. A systematic approach to identify individual GC-peaks is also described. The procedure could also be applied to the analysis of nonsaponifiable canpounds which are not steroidal in nature. For quan- titation the approach is practically the same, except that deuterated, tritiated or otherwise labeled species of the identified canpounds are added to the yolk (Thompson, 1971}, BJSrkhem, et al., 19714). The latter investigators were, thus, able to quantitate cholesterol at the level of only it ug with a standard deviation of approximately 1.3%. I'Allllllllllll l m w _ .ooonmuaolzuuaodoao mo iguana m: .9” g one one . smm {a and as" an 4? guxdduxoam huummuba waning—m m; zaum “Invwog Ion relative intensity J37 .saqosaannisessrsxsmz .aamzswn_ 8v own 03.. PH {a b‘B by PLI ’ b {P L LP 4 d d f j r amuse T oow on" man i 3. nm Q , m. « ”Ha m: "ISM. H .w ..... :1 New. NIuVNONm‘ Ion relative intensity m1I h Ilrllluil .J, .- o l I III I I. ‘Illl..ll|. ff 1.. 138 The proof that an identified capound really originates in the egg yolk, i.e. it is not an artefact created smewhere during the work- up of the sample, can only be brought about by analysis techniques which would be totally different from the one proposed, here, and such a method may not even exist at the present tile. It should be added that, as a by-product of this research, several procedures and techniques were developed which can also be useful in applications outside the realm of steroid analysis. These are especial- ly the fat extraction frat eggs, the saponification of unlimited amounts of mixt-solvent extracts, the micro TIC-band recovery and extraction, the repetitive, short-time digitonin precipitation and precipitate recovery, the preparation of silver nitrate TIC-plates, and the calcul- ations to correlate gas chranatographic data of different investigations. The concept of an easy reference table for a particular compound group, such as Table 2 prepared for steroids, should also be applicable to other canpound families . BIBI-ICBRAPHY 3311mm Adlercreutz, H. and T. Luukkainen, 1%8. Gas phase chromatographic methods for estrogens in biological fluids, ch. 2 in Gas Rinse Chromatography of Steroids, X.B. Eik-Nes and E.C. Homing, edts. , Springer-Verlag New York Inc. , New York. Anonymous, 1963. CR: Handbook of Chemistry and Physics, lI-hth ed., C.D. Hodgnan, ed., The Chanical Rubber Publishing Co., Cleveland, Ohio. Anonymous, 1%5. Hawk's mysiological Chemistry, lhth ed., B.L. Oser, ed., The Blakiston Div., McGraw-Hill Book Casp., New York. ' Anonymous, 1%8. The Merck Index, 8th edn., P.G. 8techer, ed., Merck and Co., Inc., Rahway, N.J.. Anonymous, 1%9. CK: Handbook of Chemistry and Raysics, 50th ed., R.C. Weast, ed., The Chanical Rubber Co., Cleveland, Ohio. Anonymous, 1973. CH: Atlas of Spectral Data and Physical Constants of Organic Canpounds, J.G. Grasselli, ed., cm Press, a Div. of The Chemical Rubber Co., Cleveland, (bio. AGAC, 1%5. Official Methods of Analysis of the Association of Official Agricultural Chemists, 10th ed., AGAC, Washington, D.C. Areas, 24., 1972. Steroids in egg yolk. Steroids 19:25-31». Avigan, J., D.S. Goodman, and D. Steinberg, 1%3. Thin-layer chromatografliy of sterols and steroids, J. Lip. Res. hleO-lOl. Bartov, I., S. Bernstein, and P. Budowski, 1971. Variability of cholesterol concentration in plasma and egg yolks of hens and evaluation of the effect of sane dietary oils. Poultry Sci. 50:1357-136h. Bermnn, W., 1940. The splitting of digitonides. J. Biol. Chem. 132:1:71-h72. J39 lho Beynm, J.H. and R.M. Caprioli, 1972. Metastable ions as an aid in the interpretation of mass spectra, ch. 5 in Biochemical Applications of Mass Spectrometry, G.R. Waller, ed., Wiley- Interscience, A Div. of John Wiley 8. Sons, Inc., New York. Bdflrkhem, 1., R. Blmhmstrand, and L. Svonssm, 197k. Serum choles- terol detemination by mass fragmentograxinr. Clin. Chim. Aeta 5,42185-193. Bligh, E.G. and W.J. Dyer, 1959. A rapid method of total lipid extrac- tion and purification. Can. J. Biochan. Emsiol. 37 :9ll-917. Boorman, K.N. and H. Fisher, 1966. Elbe absorption of plant sterols by the lee Brita Jo Rutre 20:689‘7010 Wm, B. and V.R. Wheatley, 1959. Studies of sebum Biochan. J. 73:1hh-1h9. Brooks, C.J.N. and B.S. Middleditch, 1973. Some aspects of mass spectranetry in steroid analysis, in Modern Methods of Steroid Analysis, E. Heftmann, ed., Academic Hess, New York. Brooks, C.J.W., W. Henderson, and G. Steel, 1973. The use of trimethylsilyl ethers in the characterization of natural sterols and steroid dials by gas chromatograpmbmass spectros- etry. Biochim. Bioptws. Acts. 2962181416. Brooks, C.J.w., E.C. Homing, and J.S. Young, 1%8. Characterization of sterols by gaschrmatograptw - mass spectrometry of the trimetlwlsilyl ethers. Lipids 3:391-M2. Brooks, C.J.H. and J. Watson, 1%7. Quantitative thin-layer chroma- togram of trimethylsilyl ethers of hydroxylic steroids. J. Chranatog. 31:3%-1+Oh. Brooks, J. and D.J. Taylor, 1955. Eggs and Egg Products. Dept. of Scientific and Industrial Research, London. Food Investigation. Special new NO. 60. Brown, T.L. and J. Benjamin, 1%!“ Useful thin layer chranatography techniques. Anal. Chem. 36:“16. Budzikiewicz, H. , 1972. Steroids, ch; 10 in Biochemical Applications of Mass Spectranetry, G.R. Waller, ed. , Niley-Interscience, a Div. of John Wiley 8. Sons, Inc., New York. Budzilciewicz, H., C. Djerassi, and D.H. Williams, 1%7. Mass Spectrum- etry of Organic Caupomds. Holden-Day, Inc., San Francisco, California. Bunnel, 11.11., 1967. Vitamin E assay by chanical methods, ch. 6 in he Vitamins, 2nd ed., vol. 6, P. GyBrgy and v.11. Pearson, edts., Academic Press, New York. lhl Butt, W.R., A.A. Henley, and C.J.O.R. Norris, 19MB. The raphic estimation of steroid hormones. Biochem. J. 1&2: 74452. Capella, P., G. de Zotti, G.S. Ricca, A.P. Valentini, and G. Jacini, 1%0. Chromatogram on silicic acid of the nonsaponifiable matter of fats. J. Am. Oil Chan. Soc. 37:56h-567. Clarenburg, R., I.A.K. Chung, and L.M. Uakefield, 1971. Reducing the egg cholesterol level by including emilsified sitosterol in standard chicken diet. J. Nutr. 101:289-297. Collander, R. , 199. Die Verteilung organischer Verbindungen zwischen Ether und Nasser. Acta Chem. Scand. 3:717-7’47. Cook, R.P., 1958. Cholesterol: Chemistry, Biochemistry, and Pathology. Academic Press Inc., Publishers, New York. Copius-Peereboan, J .W., 1%5. Gas chromatography of pmrtosterol mixtures. Part I: identification of phytosterols. JeGeCe 3:325-3290 Copius-Peereboan, J.R. and HM. Beekes, 1%5. The analysis of mixtures of animal and vegetable fats. 5. Separation of sterol acetates by thin-layer chromatography in reversed-phase systmns and on silica gel G-silver nitrate layers. J. Chrcmatog. 17:99-113. Couch, J.R. and A.E. Selma, 1973. Effect of diet on triglyceride structure and casposition of egg yolk lipids. Lipids 8: 385-3?- Cubero, J.M. and H.M. Mgold, 1%5. Chranatography on adsorbent layer;6 impregnated with silver nitrate. Microchem. J. 9: 227-2 0 Dalgliesh, 0.3., 3.0. Homing, M.G. Homing, ILL. Knox, and K. Yarger, 1%6. A gas-liquid chranatographic procedure for separating a wide range of metabolites occurring in urine or tissue extracts. Biochan. J. 101:792-810. Dam, H., 1%8. Einige Baserkungen in bezug auf die Bestimmg von Cholesterin. Biochem. z. 19+:177-1%. Diekman, J. and C. Djerassi, 1967 . Mass spectrometry in structural and stereochemical problems . 125. Mass spectrometry of some steroid trimethylsilyl ethers. J. Org. Chan. 32 :1005-1012. Dietschy, J.M. and M.D. Siperstein, 1%7. Effect of cholesterol feeding and fasting on sterol synthesis in seventeen tissues of the mt. Jo up. R88. 8:97-101!“ 11:2 Djerassi, C. , 1970. Application of mass spectrometry in the steroid fields Pure Appl. Chane 21, ”-2250 Eder, S.R. , 1972. Prllparative Schichtchrasatographie phosphatid- haltiger Lipide. Pette, Seifen, Anstrichm. 7h:519-52h. Eik-Nes, K.B., 1%8. Gas phase chromatography of androgens in biological samples, ch. ’4 in Gas Elase Chromatography of Steroids, K.B. Eik-Nes and E.C. Homing, edts., Springer Verlag New York Inc. , New York. Eneroth, P., J .-A. Gustafsson, and E. Nystrfin, 1969. Steroids in newborns and infants. Europ. J. Biochem. 11:’+56-h6h. Entenmann, C., 1961. The preparation of tissue lipid extracts. Jo Me 011 Ch“. 30¢. 38:53h'5380 Paul, W.H. and C. Djerassi, 1970. Mass spectrometry in structural and stereochemical problems - 19%: The mass spectrometric framentations of steroidal sapogenins. Org. Mass Spectrom. 3:1187-1213. Fedeli, s. and o. Jacini, 1971. Lipid composition of vegetable oils, ch. 7 in Advances in Lipid Research, vol. 9, R. Paoletti and D. Kritchevsky, edts. , Academic Press, New York. Fex, J. , 1920. Chanische und morphologische Studien fiber das Cholesterin und die Cholesterinester in nomalen und patho- logisch verflnderten Organen. Biochem. Z. 10h:82-17h. Folch, J ., M. Lees, and 6.1!. Sloanestanley, 1957. A simple method for the isolation and purification of total lipids fran animal tissue. J. Biol. Chem. 226:h97-509. Galanos, D.S. and v.24. Kapoulas, 1%5. Preparation and analysis of lipid extracts from milk and other tissues. Biochim. Biophys. Acta 98:278-292. Gardiner, W.L., C.J.W. Brooks, and E.C. Homing, 1966. Urinary steroid pattern of the human newbom infant. Biochim. Biophys. Acta 130:278-281. Germs, A.C., 1972. The total sterol (cholesterol) content of various weight-categories of eggs. Z. Lebensm.-Untersuch. Goldrick, B. and J. Hirsch, 1%3. A technique for quantitative l’Irgcouvgry of lipids from chrauatoplates. J. Lip. Res. h: 2 3. Goodspeed, F.E. and M.F. Millson, 1967. The purification of chloro- form for use as solvent in IR-spectroscopy. Chan. Ind. London , 1591+. 1&3 Grim, x., 1967. 1,2,3-trihydroucy-benzol, pyrogallol dingo? Beilstein's Handbuch der Organischen Chanie, It. uf1., 3. Erglnzungswerk, 6. Band, 8. Teil, 6260-626h. flaming, M.C. and N.G. Foster, 1972. Interpretation of Mass Spectra of Organic Caspounds. Acadmnic Press, New York. Harten, J.E., 1973. Personal camunication. Henick, A.S., M.F. Dance, and J.H. Mitchell, Jr., 1951*. Estimating carbonyl compounds in rancid fats and foods. J. Am. Oil Chm. 30¢. 31:88-91. Hertelendy, l". and R.H. Cannon, 1%5. A chromatographic investigation of egg yolk for the presence of steroid estrogens. Poultry Sci. l41-»:120‘5-4209. Hasberg, E. and A. Seher, 1972. Isolierung von Steringemischen aus Fetten und Olen. z. Lebenmn.-Untersuch. 1h9:129-J33. Horlick, L. and J. Avigan, 1%3. Sterols of skin in the normal and triparanol-treated rat. J. Lip. Res. l”160-165. Homing, E.C., 1%8. Gas phase analytical methods for the study of steroid hormones and their metabolites, ch. 1 in Gas mass Chromatography of Steroids, LB. Eik-Nes and E.C. Homing, edts., Springer-Verlag New York Inc., New York. Homing, E.C., C.J.W. Brooks, and N.J.A. Vanden Heuvel, 1%8. Gas phase analytical methods for the study of steroids, ch. 7 in Advances in Lipid Research, vol. 6, R. Paoletti and D. Kritchevsky, etds., Academic Press, New York. Homing, M.G., E.M. Chambaz, C.J. Brooks, A.M. Moss, E.A. Boucher, E.C. Homing, and R.M. Hill, 1969. Characterization and estimation of urinary steroids of the newborn humm by gas-phase analytical methods. Anal. Biochem. 31:512-531. Homing, E.C., M.G. Homing, E.M. Chambaz, P.I. Jaakonmaki, and C.J.N. Brooks, 1967. Studies of analytical separations of human steroids and steroid glucuronides. J.G.C. 5:283-289. Homing, 3.0. and W.J.A. Vanden Heuvel, 1961;. Gas chrasatogramy in lipid investigations. J. Am. Oil Chem. Soc. I+1:70"l--716. Homing, E.C., N.J.A. Vanden Heuvel, and 3.6. Creech, 1%3. Separation and determination of steroids by gas chranatograptw, ch. 2 in Methods of Biochemical Analysis, vol. 11. D. Glick, ed., Interscience Publication, Division of John Wiley 8. Sons, New York. 1M Idler, D.R., N.R. Kimball, and B. Truscott, 1%6. Destruction of micro quantities of steroids on silica gel as shown by repeated thin-layer chmtograxhy. Steroids 8: 865-876. Idler, D.R. and L.M. Safe, 1972. Separation of steryl acetates by silver ion chrmatograpmr. Steroids 19315-321» Idler, D.R. and P.M. Niseman, 1972. Sterols: Physical Properties, ch. Ii6 in Biology Data Book, 2nd. ed., vol. 1, P.L. Altman and D.S. Dittmer, edts., Federation of American Societies for Experimental. Biology, Bethesda, Maryland. Ikekua, N., R. Watanuki, H. Tsuda, and E. Sakai, 1%8. Correlation between molecular structure of sterols and retention time in gas chrusatogralily. Anal. Chem. 1+0:ll39-11h1. Inter-Society Camission for Heart Disease Resources, 1970. Primary prevention of the atherosclerotic diseases. Circulation ham 55-95. Issidorides, C.H., I. Eitagawa, and E. Mosettig, 1%2. Cleavage of steroidal digitonides in dimethyl sulfmclde . Jo Ora. Chas 27318693-1‘69‘. Johnson, A.R., 1971. Extraction and purification of lipids, ch. 6 in Biochemistry and Methodology of Lipids, A.R. Johnson and J.B. Davenport, edts., John Wiley and Sons, Inc., New York. Johnson, A.R. and R.B. Stocks, 1971. Gas-liquid chrmatography of lipids, ch. 10 in Biochmistry and Methodology of Lipids, A.R. Johnson and J.B. Davenport, edts., 1st ed., John Wiley 8: Sons, Inc., New York. Jones, D., 1%9. Variations in the cholesterol content of egg yolk. Nature 221:780. Kannareck, R., W.-H. Lee, A. Paliokas, and C.J. Schroepfer, 1%7. Thin-layer chranatography of sterols on neutral alumina impregnated with silver nitrate. J. Lip. Res. 8:282-28h. Naufmenn, H.P., 1970a. Fortschritte auf dem Gebiet der Lipoid- Analyse h: Diinnschicht-Chranatographie, Teil 1. Fette, Seifen, Anstrictn.‘ 72:811-825. Kauimann, H.P. , 1970b. Portschritte auf dem Gebiet der Lipoid- Analyse 5: Dllnnschicht-Chranatographie, Teil 2. Pette, Seifen, Anstrichm. 72:902-915. Wu, HeP. , 1970C. .Portschritte auf dem Gebiet der LipOid- Analyse 7: Dunnschicht-Chranatographie, Teil h. Fette, Seifen, Anstrichm. 72:1091-1103. 1145 Kiliani, H., 19113. User Digitalis-Stoffe. (38. Mitteilung) Ber. ”to CDC. GOO. 513163-16390 Knights, B.A., 1966. Application of retention indices to steroid gas chranatograplw. J.G.C. h:329—331. Knights, B.A., 1%7. Identification of plant. sterols using combined GIG/mass spectrometry. J.G.C. 5:273-282. Knights, B.A. and W. Laurie, 1%7. Application of combined gas-liquid chromatografily - mass spectranetry to the identification of sterols in cat seed. Paytochem. 6:1iO7-h16. Eobayashi, A. , 197,4. Personal cmmication. Kottke, B.A., J. Wollenweber, and C.A. Owen, Jr., 1%6. Quantitative thin-layer chrmatography of free and conjugated cholic acid in human bile and duodenal contents. J. Chranatog. 21:l+39-ld+7. Kovats, E., 1958. 206. Gas-chmtographische Charakterisierung organischer Verbindungen. Teil 1: Retentionsindices alipha- tischer Halogenide, Alkohole, Aldehyde und Ketone. Helv. Chim. Acta 1&1 :1915-1932 . Kritchevslqr,D. , 1971:. New drugs affecting lipid metabolism. Lipids 9297-102. Kritchevsky, D. and S.A. Topper, 1%1. The free and ester sterol content of various foodstuffs. J. Nutr. 7h:hh1-Mi1+. Huksis, A., 1971a. Progress in the analysis of lipids 9: Gas chromato- grapln', part 1. Fette, Seifen, Anstrich,. 73:130-138. Euksis, A., 1971b. Progress in the analysis of lipids 10: Gas chrana- tography, part 2. Pette, Seifen, Anstrichm. 73:332-312. Kuksis, A. and T.C. Huang, 1%2. Differential absorption of plant sterols in the dog. Can. J. Biochem. Physiol. h0:1h93-150h. Lindgren, B.O. and C.M. Svahn, 1%6. Separation of trimettwlsilyl ethers of triterpenes by thin-layer chromatography: Triterpenes in wood frcn Papulus tremula L. Acta Chem. Scand. 20:1763-1768 . Lisboa, B.P. , 1%9. Thin-layer chrunatography of steroids, sterols, and related compounds, ch. 1 in Methods in Ensymology, vol. 15, Steroids and Terpenoids, R.B. Clayton, ed., Academic Press, New York. Mangold, H., 1961. Thin-layer chranatography of lipids. J. Am. Oil Chem. Soc. 38:708-727. 1M6 McNair, H.M. and R.J. Bonalli, 1%8. Basic Gas Chrunatogram. vmln ”mm, "m cm, cwom.» Mange, H. , R.J. Lillie, and C.A. Denton, 1957. A chick-growth factor in egg yolk. J. Nutr. 63:199-508. Morris, L.J. , 1%3. Fractionation of cholesterol esters by thin-layer chranatografly. J. Lip. Res. 15:357-359. Morris, L.J., 1%6. Separation of lipids by silver ion chrasatograliy. Jo “Pa ”.0 7, 717-7320 Neher, L., 1%9. T18 of steroids and related canpounds, ch. L in Thin-layer Chraaatograptnr, E. Stahl, ed. , Springer-Verlag New York Inc., New York. Parkinson, T.L., 1%6. 'nie chemical composition of eggs. Jo 361. FOOd mes 173101.111. Patterson, GM. , 1971. Relation between structure and retention time of sterols in gas chromatograpty. Anal. Chm. 16211654170. Pennoch, J.R., G. Neiss, and H.R. Mahler, 1%2. Biochemical studies on the developing avian embryo. mocha. J. 85:530-537. Radin, N.S., 1%9. Preparatim of 11 d extracts, ch. M in Methods in W,.vol. Lipids, on. mountain, ed., Academic Press, New York. Randerath, H. , 1%6. Thin-layer chrulatograptw. Verlag Chunie, anH, Neinheil/Bergstr., Academic ness, New York. Remodel, R. and J.B. Sjfiwall, 1973. Canputer evaluation of gas chralatografiiic - mass spectrcnetric analysis of steroids m blamed. "two Me was 1‘5, EBB-1059. Rohrschneider, L. , 1%6. Eine Methode sur Charakterisierung von gasgln-matogramischen Trennflflssigkeiten. J. Chrcnatog. 22: -22. Rmsos, D.R., 1973. Personal cummication. Rosanski, A. , 1%6. Gas chrmatographic determination of campeaterol, p -sitosterol and etimsterol. Anal. Chm. 38:3640. Schflnheimer, R. and H. Dam, 1933. Ober die Spaltbarheit und Lleich- keit von Sterindigitmidsn. Hoppe-Seiler's z. Physiol. Ma 215 359-630 I .ri‘li‘llllli '1. I 1h? m, A. m E. m, 1972. D1. mm m m. gemischen 3 : Quantitative Bestilnung dflnnschicht- chromatographisch getrennter Sterins . Fette, Seifen, Anstrich. 71:261-268. Singh, B.A., 1972. Effect of D-thyrosine and nicotinic acid on l.c‘l'goil‘cstufirol mtabolism of laying hens. Indian J. Anim. Sci. 3 33 35s Singh, B.A., J.P. Weiss, and E.C. Neher, 1972. Effect of asasterols onsterolmetabolisminthelsyinghen. PeultrySci. 51:1:1‘9-1557. Sousa, R.J. is and H.R. Nee, 1%9. moved separation of sterols by reversed-phase thin-layer chromtogrqiw. J. Lip. Res. 10:21IO-2h3. Sperry, W.M. , 1%3. Quantitative isolation of sterols. J. Lip. Res. #:221-225. Stahl, E., 1%9. min-m Chralatogram. Spinner-Vern, New York Inc., New York. Subbiah, M.T.R., 1972. Analysis of sterols: applicatim to sterol balance studies. In. J. Clin. Nutr. 25:780-788. W, Coos, Bane m, MOI. mm, IN Jere W, 19720 Low-resolutim mass spectraneters. 1. Michigan State University, ch. 3, B in Biochmical Applications of Mass Spectranetry, G.R. Waller, ed., Wiley-Interscience, a Div. of John Wiley 8: Sons, Inc., New York. Thaysen, Th. E.H., 191,4. Beitrlge sur plysiologischen Chemie des Cholesterins und der Cholesterinester. I. Die Digitonin- methode zur quantitativen Bestimung dss Cholesterins und der Cholesterinester. Biochm. z. 62:89-130. 'menot, J.P. and E.C. Homing, 1972. some derivatives of Inseam urinary steroids for CC and GC-TB studies. Anal. Letters 53219330 manpson, R.H., Jr., 1971:. finD. Thesis in preparation. Michigan State University. TBkes, L., G. Jones, and C. Djerassi, 1%8. Mass spectranetry in structural and stereochemical problems . 161. Elucidatim of the course of the characteristic ring D framtation of “OM“. Jo “a Ma 8“. ”35165-51‘770 Tu, C., W.D. Powrie, and 0. remains, 1970. Steroids in egg yolk. . J. Food Sci. 35:601-6w. I’ll! MI!” I‘ .l.‘ I 1.. II! .III' 1 III 1&8 Turk, D.R. and B.D. Barnett, 1972. Diet and egg cholesterol content. Poultry Sci. 51:1881, (Abstract). Vanden Heuvel, W.J.A., E.O.A. Haahti, and E.C. Homing, 1961. A new liquid phase for gas chromatographic separations of steroids. J. Am. Chm. SOC. 83:1513-151’4'. Vanden Hem-cl, R.J.A. and E.C. Homing, 1962. A study of retention- time relationships in gas-chromatography in terms of the structure of steroids. Biochim. BiOphys. Acta 61+:hl6-h29. Weiss, J.F., E.C. Naber, and R.M. Johnson, 1961+. Effect of dietary fat and other factors on egg yolk cholesterol. Arch. Biochem. Biopkrys. 105521-526. Windaus, A., 1%9. Uber die Entgiftung der Saponine durch Cholesterin. Ber. Dto Chm. 688. “2:238-2‘46. Windaus, A., 1925. Uber die Reindarstellung des Digitonins und fiber die Nebensaponine des Digitalisssmens. HOppe Seiler's z. Physiol. Chem. 150:205-210. Wortmann W., M. Kasparow, and J .C. Touchstone, 1972. Determination of specific activity of isot c materials by thin-layer chromatography. Lipids 7: 204425. Wotiz, R.H., and S.J. Clark, 1969. Gas-liquid chranatographic methods for the analysis of steroids and sterols, ch. 2 in Methods in Enzymology, vol. 15, Steroids and Terpenoids, R.B. Clayton, ed., Academic Press, New York. Wyllie, 8.6. and C. Djerassi, 1968. Mass spectrometry in structural and stereochemical problems. 1146. Mass spectrometric fragmentations typical of sterols with unsaturated side chains. Je Org. Chane 333305-313. 4 . 'I'Oal. rl... — . .. A. _ HICH TQT N V. L I ”I 1 RARIES WHEN I 437 IGRN S E U "WWI“ HUIIWII 9 1 03 312 3 ll” 7