8/3/2019 J. Lipid Res. 1969 Andersen 316 9

1/4

Preparative thin-layer and column

chromatography of prostaglandinsNIELS H. ANDERSEN*

Converse Laboratory, Harvard University, Cambridge, Massachusetts 02138

ABSTRACT Analytical and preparative chromatographicmethods for monounsaturated prostaglandins are described.Th e systems were developed specifically for separation of thevarious hydroxy epimers of prostaglandin E l and F1 but also

offer superior separations for some of the known naturalprostaglandins.

.SUPPLEMENTARY KEY WOR DS visualization methods *

oxygenated fatty acids . stereochemical designations .diastereomers

CONTINUINGtudies on th e synthesis of diastereomericprostaglandins (1, 2) have forced us to reexamine theTLC systems suggested' in the literatu re. Many of thesesystems do not differentiate between PGE, PGF, an d

PGA compounds differing only in configuration at (2-11and (or) C-15. Thus the PGE-like compounds, withrelative R, 1.15 and 1.32 (PGEl, relative R, = 1.00) inthe A-system of Nugteren, Vonkeman, and van Dorp(4), ha t can be isolated from the nonenzymic cyclizationof eicosatrienoic acid are probably not, as Nugterenet al. suggested, hydroxyl diastereomers of the naturally-occurring PGEI.

The recent reexamination (5) of the only synthesis ofPGFla (and traces of PGEl methyl ester) published prior

The abbreviated nomenclature used for the prostaglandins inthis article is explained in the Appendix. Abbreviations: TLC,

thin-layer chromatography; PGA, prostaglandin A; PGB, pro-staqlandin B, etc; THF, tetrahydrofuran.*

Postdoctoral fellow at H arvar d University, 1968 (NIH GrantSF2-GM-33, 327-02). Present address: Depart ment of Chemistry,University of Washington, Seattle, Wash. 98105.

No attempt at a complete literature survey is implied. Theexcellent review of the prostaglandins compiled by the Worcesterand Upjohn groups (3) was available to us prior to publication andserves as an excellent starting point. In addition, several suggestionsof W. P. Schneider of the Upjohn Co. were extremely helpful.Some of the T LC systems which we developed are closely allied oridentical to those used by other workers.

to ours (1, 2) again shows the need for better TLC sys-tems. The materials isolated by the method of Just andSimonovitch (6) appear to be epimeric with the naturalhormones.

Several TLC systems with high resolving power aredescribed here. Knowledge of the chromatographicbehavior of prostaglandins in these systems may facilitateidentification of synthetic prostaglandins prepared byother workers, and the methods devised should proveuseful to biochemists working with the natural prosta-glandins as well. Some representative formulas are shownopposite.?

EXPERIMENTAL PROCEDURES ANDRESULTS

Thin -Laye r ChromatographyStandard and unknown samples [dissolved, about 1pg/pl, in acetone, dioxane, or tetrahydrofuran (THF)]were applied as spots 1-2 mm in diameter on 5 X 10 or5 X 15 cm plates with a fluorescent indicator. The plateswere developed in glass tanks, 18 cm high, lined withfilter paper. Th e development distance was 7-8 cm forneutral silica and alumina plates and 10 cm for acidicsilica plates. Fur ther development was prevented byscraping away a 1 mm strip of the adsorbent across thewidth of the plate.

The plates were viewed under UV radiation (PGBand, with heavy loading, PGA were visible as dark spotson the fluorescent background) and then made perma-nently visible by one of the following methods: ( u ) heat-ing to 120C on an electric hot plate and immediatedipping into a 1% solution of SbCls in CCl4-CHzClz 5 : 1and then further heating, ( b ) vanillin-HoPO4-ethanolspray, ( c ) acidic ceric sulfate spray, (d)phosphomolybdic

2 The bracketed notations for these prostaglandins are explainedin the Appendix.

316 JOURNAL OF LIPIDRESEARCH OLUME0, 1969

8/3/2019 J. Lipid Res. 1969 Andersen 316 9

2/4

or molybdic acid-ethanol spray, (e) exposure to iodinevapor followed by char ring at 200-250C, or (f) %cupric acetate in 15 % aqueous phosphoric acid sprayfollowed by heating at 120C. The characteristic colorsobtained with the vanillin spray ( b ) have been reported(3). The SbCl5 reagent ( u ) also gives different colorswith various prostaglandins: PGF, initial reddish cast

becomes dark gray-brown; PGEI, brown with slightreddish cast; PGA, brown with initial green cast; andPGB, lemon yellow. The acidic cupric acetate spray (f )also gives characteristic colors: green for PGA and PGE,yellow for PGB, and violet for PGF. Methods a, 6, d , andf are the most sensitive, giving excellent results with 2-5pg of prostaglandins and usable results with less. Methodsa and f are the best, but the spots obtained by method afade noticeably with time.

Adsorbents. Commercial adsorbent-coated glass plateswere used throughout. The plates coated with neutralsilica and alumina (Merck adsorbents) were supplied byBrinkmann Instruments Inc., Westbury, N . Y . Th eacidic silica plates were Mallinckrodt ChromAR 4GF

9(!yw%~sH,l HH'gHll

OH O H OHH

plates. The plates with 2-nim layers of silica were alsosupplied by Brinkmann. Acidic silica for columnchromatography was 100-200 mesh MallinckrodtSilicAR CC-4.

P-11 is the superficially dried 3 organiclayer from ethyl acetate-hexane-water-methanol-aceticacid 4 : 2 : 2 1 : 1. C-I is chloroform-THF-acetic acid

10 : 2 : . N-I is hexane-THF-methylene dichloride1 : 1 : 1. Th e H-series consists of hexane-methylenedichloride-THF-acetic acid in ratios as follows: H-I .6:2:2:1; H-11, 30:10:3:3; and H-IV, 1O:lO:lO:l.Th e D-series consists of benzene-dioxane-acetic acid inratios as follows: D-I, 3:2:0; D-11, 40:lO:l; D-111,20:lO:l; and D-IV, 20:20:1. D-IV is the same asGrten and Samuelsson's A- I system (7). The F-series isbased on ethyl acetate and more polar additives asindicated: F-I, ethyl acetate-formic acid 100 : 1 F-EL-.ethyl acetate-formic acid 400 : 5; F-V, ethyl acetate-ethanol-acetic acid 100 : 1 : 1 ; F-VI, ethyl acetate-acetone-acetic acid 90 : 10 : 1 and F-VII, cyclohexane-ethyl acetate-acetic acid 60 : 40 : 2.

These are shown in Table 1 asRf X 100 values. Most of the difficult separations wereaccomplished by multiple development, and since R ,values are not reproducible after three or more de-velopments, the R, X 100 values given should be usedonly as an indication of relative mobilities.

Preparative TLC. The first consideration here is to

Solvent Systems.

ApParent Rf Values.

have nondestructive visualization. Iodine vapor andwater spraying have been recommended (3 , 7 ) . butneither was satisfactory, particularly on the commercialplates carrying 2-mm layers of silica. PGB and PG-4 (a t

sufficiently high loading) offer no difficulties, since theyquench fluorescence. PGF and PGE diastereomers canbe located easily by exposing the layer to a glowing-hot

PGEi [ = PG(Eaa)i] PGFi, [ = PG(aaa)il

OH

15-epi-PGA1 [ = PG(AAS)11 Dried by brief shaking with crystalline sodium chloride.

TABLE 1 APPARENT f X 100 FOR SOMEPROSTAGLANDINS

Neu tral Silica Acidic Silica AI#&

P G F qPGFI,PGEi11-epi-PGEl15-epi-PGEl11 lS-epi-PGElPGAI15,-epi-PGAlPGBiCHa-PGEICHI-1 1-epi-PGE1CHI-15-epi-PGEiCHA-11 IS-epi-PGEiCHI-PGAICHa-PGBi

18 5 17 23 15 1 2 10 14 3 15 822 12 26 30 25 23 18 24 7 25 1734 25 24 38 15 15 42 3 7 32 57 37 24 3 7 32 39 10 38 2 7 3935 29 31 48 17 21 66 47 29 49 45 54 11 52 3940 37 36 56 21 28 56 46 48 76 57 36 58 54 62 15 58 48 4242 40 41 59 2 2 29 76 56 34 59 55 64 19 60 5057 78 69 8 2 50 69 7 8 57 62 76 50 79 7 7 79 62 75 81 6259 80 72 84 57 73 80 58 77 51 83 81 86 68 81 8 7

77 74 79 76 78 60 76 856 79 67 81 5047 38 34 49 20 2 7 42 45 38 30

57 51 5163 58 6465 57 63

64 84 79 69 58 65 76 67 8263 85 79 58

~ ~ ~

ANDERSEN Chromatography of Prostaglandins 317

8/3/2019 J. Lipid Res. 1969 Andersen 316 9

3/4

wire for 1-2 sec. The hot wire is touched to the pla te ina series of lines perpendicular to the solvent front. Afterthis treatment the PGF and PGE zones quench fluor-escence. Further exposure leads to charring but isdefinitely not necessary. This treatment is essentiallynondestructive since only the very top layer in thenarrow exposed zones is partially converted to the

material containing the chromophore responsible forthe observed quenching . T he exposed areas of the zonesare eluted with the entire zone and in no case haveartifacts been detectable by TLC of the eluted materials.15-150 mg of prostaglandins ca n be separated o n onelayer that is 100 mm square and 2 mm thick; the hotwire visualization works well in this entire range.

The free acids are eluted from the powdered zoneswith methanol. The residue from the methanol solutionis partitioned between water and ether (containingadded ethyl acetate or methylene chloride) after theaddition of several drops of acetic acid. The organicphase is washed with several portions of water andthen with saturated sodium chloride solution, driedover anhydrous sodium sulfate, filtered, and concen-trated on a rotary evaporator.

PGE isomers purified in this way are invariably con-taminated with traces of PGA resulting from dehydra-tion du ring chromatography and reisolation. Th e degreeof contamination is reduced significantly when aceticacid is used in the extraction instead of the mineralacids usually employed. Oxalic acid can also be used.PGE diastereomers from preparative TLC are bestpurified on small columns of acidic silica (see below).

Smaller quantities (1 pg-10 mg) are best separated on

a ChromAR 4GF plate (0.25 mm layer). Th e free acids,eluted with acetone in nearly quantitative yield, do notcontain dehydration products when nonacidic systems(generally cyclohexane, ethyl acetate-acetone) or weaklyacidic systems (1-2% acetic acid) ar e used. Th e pro-staglandin samples obtained in this way need not besubjected to reisolation from acidic aqueous solution.Th e only further purification needed (if any) is removalof fine particles by filtration or centrifugation afterthe sample has been dissolved in chloroform, ether, orethyl acetate. As an example, the PGFla and PGFlPmixture obtained from 10.6 mg of PGEl by the methodof Bergstrom, Krabisch, Samuelsson, and Sjovall (8)was separated on a 20 X 20 cm ChromAR 4GF plate.The plate was developed to 8 cm and then 16 cm withF-VI. PGF la (4.5 mg, mp 97-100C) and PGF$ (6 mg,mp 127-128C) were obtained from th e two zones(R , for PGF la 0.5-0.6 and for PGF$ 0.32-0.48) byelution with acetone.

Column Chromatography on Acidic Silica

PGE diastereomers contaminated with PGA were puri-

fied on 5 X 60 mm columns of acidic silica (packed incyclohexane) by elution with ethyl acetate-cyclohexane1 2 after prior elution with the same solvent in the pro-portions 3 : 2. Such columns can be used for as much as20 mg of contaminated PGE obtained by preparativeTLC.

The same method can be used directly for separation

of the classes as well as separation of t he various di -astereomers. The column is packed with SilicAR CC-4(100-200 mesh) as a cyclohexane slurry. The pro-staglandin mixture is applied to the top in a smallvolume of ethyl acetate or cyclohexane-ethyl acetateand eluted with cyclohexane-ethyl acetate mixtures:PGA and PGB are eluted with a 2:l mixture, PGEdiastereomers with a 1 : 1 mixture, and PGF diastereo-mers with pure ethyl acetate (sometimes acetone wasadded). PGEI, PG(EPa)1, and PG(Ea0)l can be sepa-rated from each other in this way. A 100 mg sample ofPGE mixture can be separated on a 17 X 300 mm col-umn with more than 80% of the material eluted in pur eform. Th e mixed fractions can be rechromatographed toimprove the yields of separated components. Recoveryof PGE is greater than 97y0 as measured by UV absorp-tion of PGB formed on base treatment, and conversionto PGA is insignificant ( - P ad < @PP)-( P 4 ) < ( a d

R, values for PGE diastereomers:

Rf values for PGA diastereomers:

NOMENCLATURE APPENDIX

A new, convenient extension of the presently used abbrevia-tions for the prostaglandins has been used in this article. Inthis extension, the configurations at C-9, C-11, nd C-15 (theusual positions for oxygen functional groups) are indicated, in

4 N. H. Andenen, unpublished observations.

318 JOURNALF LIPIDRESEARCHVOLUME 0, 1969

8/3/2019 J. Lipid Res. 1969 Andersen 316 9

4/4

OH OH1 a,l5a-dihydroxy-9-0~0-

13-trans-prostenoic acid

PG (Eaa)

.... C H 2 ) 6 C 0 2 H

C5 H I

OHO H

naturally derived 1 5-epi-PGEI= PG (Ea@)

OHOHantipode of PGEl

= ent-PG(Eaa)I

OH

OH15-keto-1 -epi-PGFla

= PG(a@K)I

0,

OH OH9a,lla,l5a-trihydroxy-l3-trans-

prostenoic acid

PG(aaa)1

OHH

naturally derived 8-epi-PGE1= iso-PG(Eaa)I

0

15-keto-dihydro-PGEl= PG(EaK)o

OH

antipode of PGBI= ent-PG(B-a)l

that order, by a or /3 within parentheses. The hydroxyl groupsin natural PGEl an d P GF la are a-oriented in the projectionmost commonly used for the prostaglandins.

This system retains the use of subscripts to indicate t he degreeof unsaturation in the corresponding E-type prostaglandin.When no configuration assignment is required, the designa-tions are: K = keto, A = unsaturation, and - (dash) for nosubstituent. In the case of %oxo compounds, these are furtherdistinguished as E-, A-, or B-type by using these letters inplace of K. Prostaglandins having cis-oriented side chains aredesignated by prefixing iso- to the abbreviation used for theC-8 epimer. Further, we designated the antipodes and race-mates by the prefixes ent- (or enantio) and rac-, respectively.

We have found these configuration-indicating designationsparticu larly useful in discussing the effects of stereochemist ryon pharmacological activities, relative mobilities, and relativerates of biological degradation of these substances.

The examples below, and in the text and table above,illustrate the system.

Manuscript received 75 July 1968; accepted 10 February 1%9.

REFERENCES

1. Corey, E. J. , N. H. Andersen, R. M. Carlson, J. Paust, E.

Vedejs, I. Vlattas, and R . E. K. Winter. 1968 .3. Amer. Chem.Soc. 90: 3245.2. Corey, E. J. , I. Vlattas, N. H. Andersen, and K. Harding.

1968. 3. Amer. Chem. SOC.0: 3247.3. Ramwell, P. W., J. E. Shaw, G. B. Clarke, M. F. Grostic,

D. G. Kaiser, and J . E. Pike. 1968. In Progress in theChemistry of Fats and O ther Lipids. R. T. Holman, editor.Pergamon Press, New York. 9 (Pt. 11): 231-273.

4. Nugteren, D. H., H. Vonkeman, and D. A. van Dorp. 1967.Rec. Trau. Chim. Pays-Bas. 86: 1237.

5. Holden, K. G., B. Hwang, K. R. Williams, J . Weinstock,M. Harman, and J. A. Weisbach. 1968. Tetrahedron Lett.1569.

6. Just, G., and C. Simonovitch. 1967. Tetrahedron Lett. 2093.7. Green, K., and B. Samuelsson. 1964. 3 . Lipid Res. 5: 117.

8. Bergstrom,S.,

L. Krabisch, B. Samuelsson, and J. Sjovall.1962. Acta Chem. Scand. 16: 969.

ANDERSEN Chromatography of Prostaglandins 319