Identification and quantitation of nucleosides, bases and other UV-absorbing compounds in serum, using reversed-phase high-performance liquid chromatography

.I. &ROMATGGRAPHiC

METHODOLOGY

RKHAFtDA. HAE2TWICK, SEBASTIANP. ASSENZA Department

of chemistry,

Urhersz-ty of Ride

and PHYLLIS

R BROWN

trccnz, iCin@otz, R.Z. 02881(USA.)

A comprehensiveinvestigationof the high-performanceliquidcbromatographic separationof nucleosides,their bases and other low-mokcular-weightW-absorbing compounds that might be found in serum is reported. A buffer-methanolgradientwas used in conjunctionwith chemicallybonded, microparticulatecolumns to .separatemany of the biologicallyimportantcompounds under study in minimal time with maximal resolution. Retention data, absorbance ratios(280/254MI) and fb~omscenceresponsesare reportedfor 86 nuckosides, bases, nucleotides and other W-absorbing compounds commonly encountered in biological studies.

Recenteffortsin our laboratoryhave been directedtowardsdeterminingthe pro&s of W-absorbing, low-molec~ar-weig~rn~~~ in serumusing reversedphase high-pe~rformance liquid chromatography (HPLC). Of particular interest were alterationsof the serumnucleosideand base profilesin patientswith neoplastic diseases, as concentrationsof several mod&d nucleosides and bases have been found to be elevatedin the urine of patientstiering from various types of malignancies.One phase of our long-termstudyinvolvesthe identificationand quantitation of compounds in serumprofilesof normal subjectsand the determinationof the range of normal values. In orderto accomplishthis study, a comprehensiveinvestigationof the separationof nuckosides, bases and other W-absorbing compounds that might be found in serum was rxv. Ion exchange was originally used in the separationby EipLC of nucleosides and bases’-“, as well as nucleotides=sr3; however, recentinvestigationshave shown that the raveme&phasemode is parScula.rlysuitable for the separationof most. of these wmpoundsxi~~. Isocratic elution on a reversed-phasecolumn has been used for the separation of ribonuckosides in miner6 and tissuesz7_Gradient dution is usehI for the s&k-step determinationof the~overallprofilesof both the nuckhides

648

RkKARTwraQS.P.ASSENZR.

P. R

BROWN

and other UV-absorbing compounds”. Studies on uremic patient9 have indicated that componentsother than nucleosidesin serum includethe aromatic amino acids, their me%abolites,tnk&an~us organic acids and nitrogenouscompounds. Becauseof improvementsin column stabilityand ef%iency, it is now possible to derive equations using empiricalconstants, with which predictedreteotiontimes of co~pouuds on reversed-phase packingmaterialscan be calculated.Theseequations based on the work of Janderaand Chur5&k’g_” and other+. can be used to reduce the laboratory time needed in the developmentof an optimized, gradient elution se~tiOKF. The data IlXessary for such calculations are reported for 56 compounds. In addition, with the constantsused in the equations, the retentionbehavior cf relatedcompoundscan be evaluatedin a quantitativemanner,thus facilitatingobservationsconcerningstructure-retentionrelationships.Although structure-retention correlationshave beenobservedon ionexchange systemsfor the nucleosidesand bases by Katz and Bur-tislfand othersU_25, studieshave not been reportedfor the nucleosides and bases on reversed-phasesystems. Systematicstudies of a series of model compounds, such as the purinesand their analogs,can yield insight into the nature of the separation process on reversed-p*hase columns in aqueous and organic systems.

imtrumentation A Waters ALC 204 liquid chrcmatograph equipped with a Model M660 solvent.programmerand a Model U6K sampleinjectorwas used. A Model 440 dualwavelengthdetector(280 and 254 nm) was used to obtain absorbanceratios on an Omniscribedual-pen recorder(Houston Instruments,Houston, Texas, U.S.A.). The fluoressnce of the various compounds was monitored using an on-line fluorescencedetector(Model GM770, Kratos, SchoeffelInstrumentDivision, Westwood, N-J., U.S.A.). An excitationwavelengthof 285 nm, with a cut-off filterof 320 nm, proved to be useful settingsfor most of the compounds studied. Integrationsand retentiontimeswereobtainedusinga Hewlett-PackardModel 3380A electronicintegrator(Hewlett-Packard,Avondale, Pa., U.S.A.).

Stainless-steelcolumns (30 cm x 4.6 mm I.D.) were pre-packed by the manufacturerwith lO-pm totaliy porous silica support, utilizingan octadecyl (C& chemicallybonded stationaryphase (Waters Assoc., MiEord, Mass.. U.S.A.). A precolumn.(Whatman, Clifton, N-J., U.S.A.) consistingof a short stainless-steelcolumn (5 cm x 4.6 mm I.D.), packed with pellicularreversed-phasematerial, was used to profong the He of the analyticalcolumn_This pre~lumn was periodicallym-packed (usingthe”tap-~Yechnique) in the laboratoryafterapproximately2030 separations. Chemic~ and chromatogmphI-cstaR&r& . Potassiumdihydrogenphosphate&naJyticaJ-reagent grade)was obtainedfrom Ma&n&rod: (St_Louis, MO., U.S.A.). Methanol (spectralquality)was obtainedfrom Bmclick and Jackson (Muskegon, M&h_, U.S.A.).

Chromatographic standards were obtained from Sigma (St. Louis, MO., U.S.A.) and were of the highest grade available. The nucleoside N2,NQlimethyE guanosinewas obtainedfrom Vega-Fox Biochemkals (Tucson, Ark, U.S.A.). All standards were made up as approximately1 mM stock solutions, and were prepared by dissolving appropriate masses of the dried standards in- 0.01 mol/l KH~F04 solution @II 6.0). These stock solutions were stored at -20”. Working standardswerepreparedby appropriatedilutionsof the stock solutions.A convenient working conceutration for the standards was found to be about L- lOwa mol/l. Severalof the standardsare difficultto dissolve @articula.rIyguanine and xanthine) and sonic&ion and adjustmentof the pH are helpful in dissolvingsuch compounds. Preparation of eluents The low-strength eluent consisted of a solution 0.02 mol/l in KEI~PO,, the

pH of whichwas adjustedto 5.6 by theaddition of potassiumhydroxidesolution.The high-strengtheluent(methanol-water,3 :2) was preparedby addingdistilled,deionized water to the appropriatevolume of methanol- The specificgravity of the degassed solvent was 0.904. The high- and low-strength eluents were degassed prior to use to prevent bubble formation in the pump chambers.The low-strengtheluentwas degassedby a water-aspiratorvacuum. Methanol solutions were degas& using a stream of helium gas (60 mI/min) for approximately30 set- A discarded line-titer from a Waters M6000 pump inlet servesas an efficientaspiratorfor the helium gas. Chromatographic conditions The final chromatographicconditions which were adopted are as follows.

The gradientwas linear (curve 6 on the M660 solvent programmer),runningfrom 0 to 100% of the high-strengtheluent in 87 min, producing a gradient slope of 0_69”/,/min. A flow-rate of 1.50 ml/min (measured volumetrically) was used at ambienttemperature.It should be noted that any combinationof final concentration and gradienttimes that will produce a gradientslope of 0.69% can be used_ Equations

The equationsused for the calculationof predictedretentiontimes have been reported prtviously”. They were derived for a linear gradient, assuming a linear relationshipbetweenIn k’ and the percentageof organicmodifierin the mobile phase. The assumption of linearitybetween In k’ and organic modifier concentrationhas been shown to be valid up to about 15% of methanol for most of the compounds studied.

OptimaE conditions The chromatogramin Fig. 1 shows the separationof 0.1-0.5 nmol each of 29

nucleosides,bases, nucleotides,amino acids and tryptophan metabolites using the chromatographicconditionsdescribedunderExperimental.

-

TIME

Fig. I.

l

LSpratioon of 0.1-0.5 run01of 28 nuckosides,

metabam.

chew

(mid

injection volume:

banded

rev&-phase

bases, nuckotides, aromatic aminoacids and 1 - 10”” molfl in e2tcI.l standard. calumn: itspm totdly porous SiIica support. EIuents: bw-

40 J.41of a solution

(C,)

011

streogth,0.02 moIP M-@‘O,, @Z S-6; higEt_ (O-60% methanolin 87 min), Iinar. Tkmpaata,

Sax m&anoL Gradktx: slow O_cs9o/dmi;l ambient; Bow-rate,1.5 m&&.t.

The effectsof pH, ionic strength and type of btier on the reversed-phase separation of nuckosides and bases have been dim previ~usly~*~~~~.T&e effe& of eluent pH can be predicted qalitatively from a knowledge of the pK, and p& vzdues ofthe axnpounds under study. With a pH ofS.6, good resolution was achieved for most of the compounds of interest. A slightly .lower pH of 5.5 wiU improve the

separationof creatinineand uric acid but will decreasethe resolutionof xanthosine from inosine. Ehent &gassing AMtough a &atex~~pirator producedvacuum is zm efkctive and easy method

of solvent degassing*compositionchangescan occur with organi~aqueous mixtures. A decrease of 2bo1S 0.5 "/I&I was observed for bo’th 3Q% and 60 % methanol solutions. However, when heliumwas bubbled through the solutionsfor up to 5 min at a flow-rateof 60 ml/mm, no changesin the methanolconcentrationswereobserved. Both methods are effectivein preventingbubble formation in the pump and in the detectorcells. Stan&rd retention times The retentiondata of 84difEerent compoundsweredetermined,togetherwiththe

absorbance ratios and fluorescencedata (Table I). The standard retention times reportedwere obtained on a single column for consistency.The constant A and the antilogarithmof the intercept,k& from plots of In k’ V~KSKS percentageof organic moditler for 56 nucleosides,bases and amino acids are also reported in Table I. The validity of these equationspredictingretentiontimes under gradientconditions is shown from the close agreement between the predicted and observed values (Table I). Although there are occasional discrepancies in the elution order betweencloselyelutingcompounds, most of thesedifferencesare withinexperimental error. Peak-heigkt ratios mdjluorescence The peak-heightratios (280/254 nm) are listed in Table I. These ratios help

in the routine identikation or verifkation of peak identities.Although they are not unique for each compound, they can help to rule out possible identitiesfor compounds elutingwith similarretentiontimes. Compounds that elutewith similarretention times will often have dissimilarratios. Most of the nucleosidesare not fluorescent,the exceptionbeing the 7-methyl nucleosides(Table I)_ However, several important compounds in serum fluoresce strongly,especiallytryptophanand metabolitescontainingthe indolering. Fluorescent behavior can be very useful both for identi@ing specificcompounds and, in the absence of a fluorescentresponse, for ruling out possible identitiesof peaks. ReproducibiIity The overall precisionof the retentiontimes was studied with respectto NQto-run precision,day-to-day precisionand ovemll column-to-c&mu reproducibility. The results of these studies are presentedin Tables II and III. For simplicity,the data are presentedfor five compounds that span the time involved for the overall separation. The run-to-run precision for five runs within a single day averaged 1.4 % (relative standard deviation, RSD), while the day-to-day precision over a 2-week period using a -singlecolumn averaged slightly higher, at about 2.2% (RSD). The resultsof the column-tocolumn reproducibilitystudiesshow that the absolute retention times for the five compounds averagedabout 7 % (RSD). The four cohmms used

-...l-.-------..-~Observed retetrllorr rhe f 2”

..-_p--y_

Cytosino(CYT) : Uridine diphoaphoglucoso(UDIW) Uridinc monophosphate (UMP) Orotidinc (ORTD) Nicotinamidcmononuclcotidc(NMN) Crcntininc(CRT) Uric Acid (URIC A.) Uracil(URA) Pscudouridino(1%URD) I-Amino~5~imidazolocarboxnmidc (4.A.5.IC) 5-Fluorouracil(S-FU) Guanosinc monophosphnta(GMP) Xanthosinemonophosphnto(XMP) Adcnosincmonophosphotc(ADP) , L-Tyrosino(LTY R) l=Mcthyladcninc(l-M-ADO) Cytidinc(CYD) Ribofluvin(RIBFLVN) Hypoxanthino(HYP) l~hlcthylcytidinc(I.M-CYD) Ouanina (WA) Uridlnc (URD) Xanthinc(XAN) S-Aminoimldazolocarboxamidcribosidc (5mAICAR) Thymidinomonophosphnto(TMP) Nicotinamidcodcninc dinuclcotidophosphate(NADP) Tbymine (THY) ‘I-Mcthylinosinc(7=MJNO) Adonosinomonophosphtc (AMP)

. ..-mu--

I_--__-_---___-“.-_---..Contpo14rMl

2,28 (0.04) 2.35 (0.06) 2S4 (0.06) 3,25 (0.03) 3,41(0.03) 3,59(0.06) 431 (0,14) 413 (0.04) 4,31 (0.14) 4,31 (OJ3) 4.36 (0.40) 4,52 (0.16) 4,75 (0.26) 5,04(0.15) 5,40(0.07) 5.77(0.10) 5.85(0.10) 6-84(0,32) 7.31(0.12) 7.25(0,36) 7,56 (0.06) 8,27 (0.19) 8.53(0,OS) 8,77(0.17) 8,89 (0.35) 8.91(0,40) 9.30 (0.14) 9.46(0.19) 9,53(0.28) ;.30(1.63)

7,43 (0.15) 7.67 (0.72) 8.04 (OJ6) 9.07 (0.90)

693 (0,27)

5.89(0.18) 5.26(0.39) 5.96 CO,14)

4.38 (OJ7)

452 (Orn)

3.92(0,24)

_. 2,72 (0.02)

2,OE(0,03)

Predicted relerrtlon rittre rl: Z

.----. Peuk-height rurto

(SBD.)

2:lGo(0,104O) 0.118(0,018O) 0.272(080153) 0.621(0,014O) 0.450(0.0058) 0.349(0,oOlO) 0.458(0.0326) 0,084(0,0042) 2.47 (0.181) 0.194(0.0640) 0.683(0,024l) 0.758(0,040) 0.010’ 1.330(0.0698) 0.532(0.0140) 0.215(0.0120) 0.459(0.0010) 0.529(0.0370) 0.575(0.0288) 0.369(0.0542) 0.369(0.0542) 0.286(0.0130) 0.073(0.0135)

0,437(0.0280) 0,188(0,026O) 0,189(0.0154) 0,516(0.0240) g.~;,co.o150)

_

ST

WK

Flrrorested”

CO926(0,003) 0.1120(0,017) 0.0975(0,003) OJ300 (0,019)

;0786 (0,007)

0,0844(0,002) 0,0973(0,014) 0,0984(0,004)

0.0646(OaOO7)

0,0653(0,003)

OS0565 (0,OlI)

0,033o(0,ooo)

O,OS80 (0,001)

A (S.D.) -.

-

- 3659 (0,083) 3,80 (0.435) 4,04 (0,0!)2) 498 (0,638)

3,23 (0,145)

-2,55 ,(0,038) 2,16 (OW) 260 (0,076)

-1.71 (0,034)

-1.35 (0,105)

-0.63 (0,002)

0,250(0,002)

k;(S,D,)

SUMMARY OF ODSERVEDAND PREDImED RETENTION TIMES, PEAK.HEICIHT RATIOS (260/254nm), PLUORESCBNCORESPONSE AT 285.nm EXCITATION, 320-m CUT-OFF, THE ABSOLUTEVALUE OF THE SLOPE A, THE ANTILOQARITHM OF THE INTERCEIW; (SEE TEXT), AND STANDARD DEVIATIONSAND ERROR LlMlTS (90% CONFIDENCE LEVEL) FOR FIVE SEQUENTIAL ANALYSES’ Chromatographic conditions arc drwibcd in tha lcbcnd to Fig. 1.

TAIILE I

F 7

,vr ?;I

P

I4

P

8

9,82(0,23) D,L-Kynurcninc(DL.KYN) 9.98(0,22) Flavin adcninc dinuclcotidc(FAD) Alpha~nicotinamidcudcnincdinuclcotidc(ALPI+NAD) 10.15(0,OS) Adcnosinc5’.diphosphoribosc(ADPR) 10.26(0.45) x.=Phcnylolaninc (LPH.ALA) 10.40(0416) 10.52(0,21) Allopurinol (ALL0 PUR) 7~Mcthylxanthosinc(‘I-MmXAO) 10.69(0,21) 10.76(0425) 3’,S’Cyclic cytidina monophosphatc(GCMP) 10.79(OJ9) S.Mcthylcytidinc(5.MCYD) Purinc (PURINE) l1,20(0,31) Anthranilhc Acid (ANTMA.) 11,23(OJ6) 7=Mcthyl~uanosinc (‘I-M-GUO) 11.82(0.16) Pyrimidinc(PRMD) 11,92(0,25) Xanthosino(XAO) 12,40(0859) Nicotinnmide(NCI’MD) 12,55(0,13) 12,86(0,07) 3’,5’-Cyclicuridinc monophosphntc(GUMP) Bctn-nicotinamidoadenine dinucleotidc(BETA.NAD) l3,OO(OJ6) Inosinc (INO) 13,52(0.16) Adcnhic (ADE) 13,54(OJ5) Gunnosinc (GUO) 14,40(OJ8) 7~Mcthylguaninc(7.M.GUA) 14.65(OJ7) 1512 (0.07) 3’,5’-Cyclicgunnoshrcmonophosphatc(CGMP) Nz-Mcthyl~uanine(N1.MGUA) 1522 (0.22) 2’.Dcoxyinosine(2’.DJNO) 15.29(0.15) 15.92(0.06) J’,S’Cyclic inosincmonophosphnte(C-IMP) 15.98(0.05) Mippurk acid (H1PP.A) ?‘_Deoxyguanosinc(2’.DGUO) 16.86(0.20) 2’-Deoxythymidine(D.THYD) 17,29(0,ll) L.Tryptophan (L-TRP) 17.47(0,16) l.Methylinosine(I-M-IN@ 18.22(0.14) 6.Mcthylpurinc(G-M-PUR) 18.98(0.19) 2~Methyladenine(2.M-ADE) 19.13(O&l) NI~Mcthylguanosine(N,.M-GUO) 19.22(0.18) Indole3ktic acid (I-3LCT A.) 19.28(0,43) Purinc ribosidc (PUR RIB) 19.69(OJ6) Tubcrcydin (TUBR) 19.73(0.15) 19.51(0.49) 19.50(1.15)

17.31(1.09) 17.35(0.11) 18.65(0.40) 20.43(1.32) 20.34(0.39) 19.63(0.61)

u.59 (0.03) 15.70(0.50)

12844(1.16) 12.4OG.91) 13.53(0.72) 13.61(0.45) 14.49(0,68) 14.14(0.89) IS.10(0.52) 14,89(1.13)

Il.85 (0.62) 12.33(1.49) 12.52(l.08)

9,93(0.71) 10.30(0.85) IO,34(1.06) 10.48(0.03) 10.14(0.77) 11631(0.16)

IO,14(0.73) 0,036(0.0014) 0.010(0.0010) 0,132(0.0300) 0,116(0.0052) 0.001 WK 0,023(0.0105) 0,707(0.0080) WK 0,630(0.0320) 1.130(0.0280) 0,148(0,007O) 0,241(0.0520) ST 0.572(0.0050) MD 0,319(0.0250) 0,512(0.0215) O.OG4 (0.0025) 0,171(0,0040) 0,120(0.0060) 0,092(OJOG7) 0.080(0.0390) 0.373(0.0020) 1,120(0.0460) 0,397(0.0412) 0.535(0,017O) 0,089(0.0076) 0.113(0.0380) 0,001 1 0,356(0.0178) 0,565(0.0260) 1,430(0.0480) ST 0.171(0.0880) 0.063(0.0020) 0,129(0.0050) 0,373(0.0140) 1.680(0.0523) ST 0.099(0.0208) 0.968(0.0090) 0,169O(0.008) 0.1610(0,018)

0.1550(0,017) On1230 (0.001) 0.1930(0,007) 0.1240(0,012) 0.260 (0.085) 0.1910(0,011)

;1790 (0,001) 0,109o(0.007)

illslO (0,087) 0.2120(0.042) 0.1600(0.013) 0.1110(0.006) O.lG20(0,012) 0.1290(0.014) 0.1690(0,009) 0.1330(0.017)

0.1890(0.013) 0.1050(0.022) 0.1670(0,021)

0.0759(0.011) 0.1070(0.015) 0.1590(0.022) 0‘1510(0.001) 0,131o(0.015) 0.1010(0.002)

0.0938(0.012)

;:g;

(1,302) (0.058) (2.900) (0.930) (0,416) (0.991)

(2.066) (0.150) (1.257) (2.508) (0,739) (2.145) 2S.Oi (1,356) 23,83 (2.943)

77.46 15.20 25,49 21,37 21,42 29.01

:5,26 (0.053) 11,93 (0.535)

11.76 (1,391)

(0.988) 13.69 lo,55 (0.864)

9.27 8.82 lo,14 10.60 12.26 9,37

7,85 (Li61)

5897 (0.601) (0.022) 6.49 6,79 (0,114) -8.74 (0,771)

;:;

-5.40 (0.407)

5,69 (0,477)

l-

B c F

3 irk

E w

8 fj *

1

;

0 6

.--..---VI

28,38(1,32)

24,47(0,31) 25,09(1.39)

23.05(0,26)

23,52(1,03)

20,09(0,OG) 20,38(1,44) 21.4G(0,17) 22.55(0,SS)

--_---__

retetrtion the f z

19.92(0,17) 20.23(0,17) 20.80(OJ6) 21.77(0,22) 21.99(0,33) 22.51(OJ3) 23.23(0824) 23.47(0,31) 23.55(0.81) 24.42(OJS) 24,90(OJ5) 27.20(0.18) 27.58(OJ8) 29.11(0.18) 30.86(0.27) 31.41(0,78) 34.66(0,27) 35,08(0.27) 41.11(0.65)

Predlctcrl

Ohwed reterrtlorl t/me f z**

_.

-___-

Fluorescemc”’

0.126(0,OllO) 0.44l(O,O4SO) 0.099(0,0101) WK 0,084(0,0048) 1.730(00433) 1.270(0,029O) 0.001~ 0.125(0,014O) 0.079(0,oOGl) 0.476(0,007O) 0.465(0,006O) 1.010(0,0105) 0.184(OaO120) 1.400(0,0532) 0.640(0,047O) 1.500(0,075O) 1.730(0,0521) 1,260(0,056l) 1.410(000365)

Penk&l&t ratio (SD.)

;2510 (O,Ol7)

0.1550(0,003) on1450(0,019)

0,188O(0.004)

0.1520(0.011)

0x40 (om) 0.1910(0.025) 0,129a(0,002) 0.1560(0,009)

A (SD.)

35.44(I-87) Caffcinc(CAF) 0,142O(0,017) 3~Indolc~nnthrnnillic ncid (J-I-ARA,) -__-” Lognrith~ data valid over the range of O-15% mcthnnol in 0,Ol mole/l KHIPO, pH 5,G. ” Z =I S.D. * t/d/t, u = 0.05, “” Fluorcsccnca symbols:WK = weaklyfluorescent;MD = modcrntclyfluorcsccnt; ST = stron&y fluorcacent, 1Numbers too small to be dctcrmincdnccurntcly,

‘I-Mcthyladcninc (7.M-ADE) N1=Mcthylguonosinc (I$-M-GUO) Kynurenicncid(KYNA) Adcnosiao(ADO) Indole3accticacid(L3ACTA) Thcobrominc(THD) S=Mchylpyriaiidinc (5.M-PRMD) 3’,5’CycllcadcnoshlaIronophosphntc(C-AMP) 2’-Dcoxyadcnosino (2’.D-ADO) NI,Na=Dimcthylguonosinu (N$M-GUO) 6-Mothylndcninc (6-M-ADE) Thcophyllinc(THP) Flnvinmononuclcotidc (FMN) Dyphyllinc(DDN) (EMclhylndcnosinc (G-M-ADO) IndolcJactnmido (13mACMD) Indolod-propionic acid(I-3PRA.)

.-_---w-T.-*

ConJpolulrl

TABLE I (corrrlnrred)

(3.678)

(0.102) (5.575) (0,364) (3.030)

(36,450) 116,80(17,520)

k30

k4 (1,2j9) 40.9?) (5,255)

;5,80 (1.356)

k.42

20-61 32,30 24.59 34,Ol

-_

ho (SD.).

.“o P

I

p yl to

J

z

P

p

TABLE Li REPRODUCXBILFJX BASES

OF RETENTION

Prh??a%kR

TTMES FOR Reteldiolz

COmpormd

Run-to-run precision (SingIecolumn) -i&in 1 day

HypoXi3IlthCl~

l&sine Guamsine nyptopkaa Caflkine

Day-to-day precision (sir&e colmrq 14 days)

Hypoxanthke hosine Guanosine

Tiyptophm ca&ine

SELECTED tii7l.z

NUtXJZOSIDES

AN-D

S.D.

trrrin)


EF

~2=S.D.]

7.47 14.2 15.1 17.6 35.7

0.112 0.190 0.201 0.263 0.527

1.50 1.34 1.33 l.ti 1.47

7.25- 7.69 13.8 -14.6 14.7 -15.5 17.1 -18.1 34.6 -37.8

7.40 14.1 15.0 17.4 35.5

0.252 0.315 0.319 0.322 0.417

3.46 2.23 2.13 1.85 1.17

6.90-7.90 13.5 -14.7 14.4 -15.6 16.8 -18.0 34.7 -36.3

Stunmary of xtention timesfor purinecompoundson a reversed-phase methaiiol-water systemusing the chromatographic conditionsdescribed in Fig. 1.

for this comparison were at differentstages of their useful lifetimes, ranging from factory fresh to one used for about 100 runs with serum samples. The lower part of Table 111 shows the precision for the same compounds, using the retention times relative to a standard (inosine). The relative retention times were calculatedfrom

where T, is the elution 4tie of an unretainedcompomd and TRr and TRcare the retentiontimesof the unknowncompound and the referencecompound, respectively. It is evident that the error was considerablyreducedfor compounds with retention times similar to that of inosine. The error for Trp is reducedfrom 6.5 % to 2.0 % (RSD) while that for Guo is reducedfrom 7-L% to 0.43 % (RSD). The error for Hyp is only slightly reduced, from 7.6% to 6.6% (RSD), while the error for Caf is increasedfrom 3.7 oAto 5.3 % (RSD) when the relativeretentiontime is used rather than the absolute retentiontime. In practice, the relative retentiontimes are most TABLE

m

REPRODUClBlLlTY EASES

OF

Parame?er

RElXNTlON

TIMES

FOR

SELECIED

NUCLEOSiDES

Compound

AND

Raw= 2 SD.)

(r C&mm-tocoham reproducibility: Hypoxanthhe absohte retentiontime lmsine Guanosine Tryptophan caffeine Retention times relativeto inosine

Hypoxanthine kmsiae Guanosine Tryptophan caffeine

720 13.7 14.6 16.9 34.8 0.419 1.00 1.08 1.29 289

1.03 1.10 1.28

7.63 7.24 7.05 6.51 3.69

0.0277 om467 0.0251 0.153

6.61 0.434 1.95 5.29

0.550 0.992

6.1 11.7 12.5 14.7 32.2 -

- 8.3 -15.7 -16.7 -19.1 -37.4

o-364- 0.474 l-.07 - 1.09 1.24 - 1.34 2.58 - 3.20

TABLE IV ZXRUCKJECE- RETENTION

RELATIONSHIPS:

EFFECX. OF BASE SIRUCXURE

PlEinf2

6.79

Hypmzdbke

3.23

Guanine

2-Amine-6-one

xanthosinc

f6-mane

A-

6-AIIliIlO

FMethyIguanis~~

2-Ancrtw, 7-znetilyl 2-Methykmimd-one 6-MethyIamko

N=MethyIglaIlk

N6-MethyIadenine

3.68 3.28 3.43 2.29 3.63 0.825 2.74 4.83

25.0 10.6 12.3 9.27 34.0 8.74 32.3 198

3.59 4.w 9.37 10.6 11.8 41.0

useful for compounds eluting within 2-5 min of the reference compound. The reference~compoundscan either be added to the sample (internalstandard),or the sample can be internahynormalizedwith respectto one or more endogenous compounds that are well separatedand are found in every sample. This procedure of internal normalization can be very useful when an internal standard cannot be added owing to unknown interferences in the region of interest. Elrrton

order

The parameterk& used to cakuiate the predictedretentiontimes in Table I, representsthe ki of a compound eluted isocraticaliywith no organic m&et in the mobile phase. It offers a convenientquantitativemeasure of the retentionof compounds Using the limited structure-retentiondata available, it is possib!e to make some :preliminaryobservationsconcerning trends in retention behavior within the groups of puriineand pyrimidinecompounds studied. Tables IV and V summarize the ki valuesof several of the compounds in Table I, grouped according to molecular structure. Although the data are too limited to make broad generaiizations,neverthelesssome interestingtrends are apparent. The effect of the addition of functionalgroups to the purinering is shown in = the k; by about 40 o/O over Table 17. An amino group in position6 (Ado/A&) increase_ the purin~purine riboside pair. The addition of oxygen to the purines to produce the iactam structureof Ino/I-Iyp, Guo/Gua and Xao/Xan produces a decrease in TABLE V STRtiCTI.JR&RETENTION Methyl position

RELATIONSHIPS:

EFFECT OF METI-XYILATION N~ckos&.s

&sc?s A-

Grcanine

Purine

AoT.emmk

G-

k?kosk

Pareni l2d

9.37 2.16 21.4 41.0

3.59 11.8 -

6.79 21.4

34.0 198

123 32.3 -

10.6 25.5 -

7-

20.6

10.6

-

-

8.74

5.40

HPLC OF NUCLEOSIDES,

EASES, ETC. EN SERUM.

I.

657

retentiontimes over the unsubstitutedpurinering. The additionofan amino group in ffie Z-position(Guo!Gua) in the presenceof otherlactam groupsdoes not signiticantly increasethe elution times, in contrastto the additionof an amino group alone to the ring (Ade/Ado). The effect of the ribosyl group was determinedby comparing eight of the purinebaseribonucleoside pairs. For most of the pukes, theadditionof a 9-ribosyl group increases the k; of the parent base 3.3-4.8-fold. The exception to this is 7-m-Guo, in which charge formation occurs. For the pyrimidines,the pair Ura/Urd shows a similarincreasein k; (2.2 times that of the base), although Cyd shows a IOfold increasein its k; value over the base Cyt_ Thesetrendsshowthattheadditionof theribosylgroup(exceptin thepresenceof chargeformation)increasesthe retentionof a purinebase by a fairlyconstantamount. Although the systematicadditionof methylgroups to variouspurinering positions for all of the purinederivativeshas not yet been studied,the data for the base Ade wereobtained.Methylatedadenineswereavailablein whichthe methylgroup was substitutedat the 1,2,7- positionson the ring and the 6-positionson the amino group. Methyl groups in the 2- and 7-positions both increasedthe k; values of Ade about 2.3-fold, while the addition of a methyl group to the 6-amino group increasedthe retentionby 4.4 timesthat of Ade. Methylationof the l-position of Ade decreasedthe retentionconsiderably.This can be exp!ained by the observationthat I-m-Ade has been shown to exist in the rare imino form=. I-m-Ade thus has a pK, value of 7.2’q whichresultsin chargeformationat the eluentpH of 5.6. The effectof methylpositions for other purinecompoundswas similarto that of Ade, althoughthesedata are not as complete. The addition of a methyl group increasesthe retentionof the purinecompounds on the reversed-phasecolumns used, but the degreeof increaseis not always predictable.Chargeformation, as with the 7-methylnucleosides,resultsin anomalous behavior_ More data will be necessarybefore any generalizedrules caa be formulated that might be used to predict retentiontimes from basic structure_However, it is encouragingthat some trendsare consistentlyobserved.It is hoped that eventuallyit may be possibleto predictthe retentionbehaviorof classesof compoundssuch as the purineson reversed-phasecolumns from molecularstructurealone. DISCUSSION

With the reversed-phaseseparationdescribeda wide varietyof serumconstituents can be separated.These constituentsincludenot only the nucleosidesand bases, but also othercompoundssuchas severalamino acidsand theirmetabolites.Therefore, reversed-phaseHPLC profilesof these compounds in serumcan serveas an excellent tool in studiesof normalmetabolismand aberationswhichoccur duringdiseasestates. To regeneratethe cohunns, periodicpurgingwith 60% methanol for about 30 min is usuallysufiicientto eliminateghost peaksand driftingbaselines.It is dif&ult to make generalizationsconcerningabsolutecolumn lifetimes.However,in our work the lifetime of our columns has been about 200-250 analysesof biological samples. The predictiveequations= are helpfulin developinggradientseparationsof the nucleosides,basesand other UV-absorbing compoundspresentin serumand in minimizingthetimeandinmaximizm g the resolutionof selectedcompounds of interest_

658

The use of purine and pyrimidinecompounds as model systems to study column behaviorandretention mechanismsappearsto be promising.A fargenumber of purine dogs ate ~~LEMercially available OT are easily synthesized and much information on electronicand conformationalstructuresis available. As the purpose of this researchwas the developmentand evaluation of an EWLC separationfor serum nucleosides,only a limited number of compounds were studied_However, a comprehensiveproject involvingretcntion-structnrcrelationshipswithinthe purinesand pyrimidincsis currentlyin PiOm in this laboratory. The limited trends Obzie~ed here should be used only in a qualitativemanner, yet they seem to indicate that additivityrulesmay exist withina class ofcompounds that may allow one to estimate retentionbehavior based on strnctures. ACECXOWLEDGEMENTS

The authors acknowledgeDr. Raymond Panzica, Department of Medicinal Chemistry,and Dr. Lecn Goodman, Departmentof Chemistry,Universityof Rhode Island for their advice and discussions_We also thank Klaus Lohse (Schoeffel instruments, Kratos Inc.) Paul Champlin (Waters Assoc.) and Fmci Rabcl (Whatman) for their assistance. This work was supportedby an American ChemicalSociety, Division of AnaIyticai Chemistry, Fellowship sponsored by the Society of Analytical Chemists of Pittsburgh,and by United States Public Health Grant CA-GM-17603-03.

1 H. J. Breter. G. Seibert and R K. Zahn. J_ Chromatogr., 140 (1977) 251. 2 A. Floridi, C. A. Faherini uld C. Fini, f_ Chromatogr_, 138 (1977) 203. 3 P_ R Brown, S_ Bobick and F. L. Hauley, J. Chromtuogr_, 99 (1974) 587. 4 K. W. St&l, E. Schlimme and D. Bojanowski, J. CZzromarogr_,83 (1973) 395. 5 R. P. Sin&d and W. E. Cohn, Biochemistry,12 (1973) 1532 6 H_ J. Breter acd R K. zahn, bzzZ_ Bbchem., 54 (1973) 346. 7 R. P. Sin&al and W. E. Cohn, AR&_ Bi~chenz_, 45 (1972) 585. 8 J. J. KirkIand, J. Chronrutogr. Sci., 8 (1970) 72. 9 F. Muakami, S. Rokushika and H. Hataao, J. Chromzrogr., 53 (1970) 584. 10 C. Horvsth 2nd S. IL Lipsky. Annl. C&m., 41 (1969) 1227. 11 M. Uziel, C. Koh and W. E. Cohn, &i_ Biachem., 2.5 (1968) 77. 12 R. A. H%rtwicckand P. R Brown, J. Ciiromurogr_. 112 (1975) 651. 13 J. X Khysn, J. Ciuonmogr., 97 (1974) 277. 14 C. W. Gehrke, K, C. Kuo, G. E. Davis, R D. Suits, T. P. Waakes and E. Bore& 1. CZwomptogr., 150 (1978) 455. 15 R. A. Hzrtwic!c and P. R Brom J. Chromafogr., 126 (1976) 679. I6 G. E. Davis, R. D. Suits, K. C. Kuo, C. W. Gehrke, T. P. Waalkes and E. Bmxk, C&h Ckm., 23 (1977) 1427. 17 F- S. Anderson md R. C. Murphy, 1. Chrormfogr.., 121 (1976) 251. 18 F. C. Senftieber, A. G. Haike, H. Veenig and D. A. Dayton, Clin C&m., 22 (1976) 1522. 19 P. Jandera and J. Chur%ek, J. Chromcltopr., 91 (1974) 223. 20 P. Jandera 2nd J. Ch~&&k, J. Chromatogr., 93 (1974) 17. 21 P. J. Schoenmakz, H. A H. Billie& R. Tijssen and L. de G&n, J. Ckomfogr., 149 (1978) 519. 22 R A. Flartwic!c,C. M. Grill znd P. R Brown. Anel. Chem., 51 (1979) 34. 23 S. Katz and C. A. Butis, i. Ciuotimo~., 40 (1969) 270. 24 M. Uzkl, C. K. Koh and W. E Co&n, Anal. Biochem., 25 (1968) 77. 25 D. Henry, J. H. Block, J. L. Anderson and G. R. Caikmt, i. ikfeci.C&m.., 19 (1976) 619. 26 B. PcIlmm and A. Pullman, kfvmz. ~eremqcZ. C/km.. 13 (1971) 115.