EVIDENCE FOR THE NATURAL OCCURRENCE OF ZEATINAND DERIVATIVES: COMPOUNDS FROM MAIZE

WHICH PROMOTE CELL DIVISION*BY CARLOS 0. MILLER

DEPARTMENT OF BOTANY, INDIANA UNIVERSITY, BLOOMINGTON

Communicated by Ralph E. Cleland, August 12, 1965Three laboratories previously have obtained from Zea mays kernels (milk stage)

one or more 6-(substituted)aminopurines capable of greatly promoting cell divisionin plant tissue cultures.1-3 At least two of the laboratories have worked on a singlecompound4 which has been identified as 6-(4-hydroxy-3-methylbut-2-enyl)-aminopurine.5 The name zeatin has been applied to this particular compound.6Only a small portion of the total cell-division activity in crude extracts from kernelsseems to be due to zeatin'-3 and its actual existence in the kernel has been ques-tioned.7 We now present evidence strongly indicating that zeatin does occur nat-urally in its unsubstituted form and also in nucleoside and nucleotide forms.Assay.-The active materials in the maize extracts were detected by using the

soybean (Glycine max, var. Acme) tissue test.3' 8 This tissue requires a substancelike zeatin for continued cell division and has proved to be very sensitive to zeatinand similar compounds. For example, in recent tests we have obtained meas-urable growth responses to as little as 5 X 10-11 M zeatin. Most assays were per-formed with small quantities of extracts which contained rather low concentrationsof active materials. With such low concentrations, the soybean test shows con-siderable variation of fresh weights for individual pieces in a particular assay but stilla high degree of reproducibility from test to test. Repeats of all experiments re-ported here have given consistent results.

Stock cultures of the soybean tissue have been maintained on a medium con-taining (mg/liter): KH2PO4, 300; KNO3, 1000; NH4NO3, 1000; Ca(NO3)2 .4H20,500; MgSO4-7H20, 71.5; KCl, 65; MnSO4 .4H20, 14; NaFe ethylenediaminetetra-acetate, 13.2; ZnSO4 7H20, 3.8; H3BO3, 1.6; Cu(NO3)2.3H20, 0.35; (NH4)6Mo7024 *4H20, 0.1; i-inositol, 100; nicotinic acid, 0.5; pyridoxine HCI, 0.1;thiamin * HCl, 0.1; a-naphthaleneacetic acid, 2; kinetin, 0.5; sucrose, 30,000;and Bacto-agar, 10,000. The pH was adjusted to 5.8 (NaOH). A preparation tobe tested for cell-division activity was added to this medium with the kinetin omit-ted, the pH was adjusted to 5.8 with NaOH or HCl, and sterilization was achievedby autoclaving. Each of four 125-ml Erlenmeyer flasks containing 50 ml ofhardened medium was planted with four pieces of the stock soybean tissue. Aftergrowth for 28 days at 27C and constant fluorescent lighting at about 40 ft-c, thepieces were weighed individually. The fresh weight averages indicate relativeamounts of cell division.Extracts.-Frozen kernels (milk stage) of Zea mays, var. Golden Cross Bantam,

were extracted by macerating in either cold or boiling 95 per cent ethanol. Thefinal concentration of ethanol was adjusted to 70 per cent (assuming the weight ofkernel was due entirely to water). The extract was cooled and the filtered pre-cipitate discarded. The filtrate was used in all of the experiments.

Chromatography.-Ethanol solutions or other preparations were streaked onsheets of Whatman #1 filter paper. For development, one of the following mix-

1052

VOL. 54, 1965 BOTANY: C. 0. MILLER 1053

tures was used (without equilibration) as the ascending solvent system: (I) 3 t-butanol: 1H20: 1 NH40H, (II) 3 t-butanol: 2H20, (III) 2 n-butanol: 1 benzene:1 methanol: 1H20, (IV) water-saturated n-butanol, (V) water-saturated iso-butanol, and (VI) water-saturated sec-butanol. Usually, an amount of ethanolicextract equivalent to 0.83 gm of kernels could be successfully resolved when streakedon an origin about 20 cm long and the solvent allowed to run 18 cm beyond theorigin. For assay, a chromatogram was thoroughly dried and cut into strips whichwere eluted by placing them in the basal medium at the time the agar was dissolvedby a short period of heating. A strip ordinarily included one tenth of the distancefrom the origin to the solvent front (some exception to this in Fig. 4); the stripdesignated 0.1, however, included the origin as well as the paper up to R. 0.1.Each average fresh weight of the tissue pieces has been graphed against the R,corresponding to the middle of the strip being tested.

Resolution of Crude Extract.-The active materials in the crude extract were re-

-

~~~~~~~~~~~~~~~~~~Layer1

Face1-Soybeantissue > so 0 0LO~~~~~~~~z-asyof a chroinatogram de- ' d\ ~ a \ 009w Water0.5 1.0 -Hot 0 50 Layer0 0.5 00 H I - -FIG. 1.-Soybean tissue W /veloped with water-saturated . .n-butanol. A cold ethanol 0 0.5 0 FIG. 3.-Soybean tissueextract from 0.83 gm of assays of chromatograms de-kernels was used. Average FIG. 2.-Soybean tissue veloped with water-saturatedfresh weight of the control assays for chromatograms de- sec-butanol. A crude aque-(without cell-division factor) veloped with water-saturated ous extract of maize was ex-pieces was 3 mg. (Averages sec-butanol. Hot and cold tracted with n-butanol, andof 16 pieces. Beginning ethanol extracts of maize the two fractions were chro-weights were 3 mg. The each equal to 0.83 gm of matographed separately.same is true for the other kernels were used. The The control pieces averagedfigures.) control pieces averaged 6 mg. 4 mg.

solved into two major fractions with all six of the above solvent systems. The re-sults with solvents IV and VI are depicted in Figures 1 and 2, respectively. In ad-dition to the two major bands, a small amount of activity which moved to aboutRr 0.5 in solvent VI was sometimes detected but usually seemed to be absent. Fig-ure 2 also indicates that hot and cold extracts are qualitatively the same, but thatthe hot extraction may be more efficient. In all six systems, the fast-moving bandof activity included the region to which synthetic zeatin migrated. The R., valuesobtained for the pure compound were 0.83 with solvent I, 0.80 with II, 0.74 withIII, 0.68 with IV, 0.82 with V, and 0.88 with VI. For any one system, rechroma-tography of the fast band (henceforth designated band B) in the same solventgave activity at the same location. The band, therefore, represents a true separa-tion from the slower-moving material. Rechromatography of the slower band(called A) in the same solvent, however, gave a little activity at the higher R,, al-though most of the activity again was located at the original position of the slowband. Both A and B were quite active in the soybean test when sterilization was

1054 BOTANY: C. 0. MILLER PROC. N. A. S.

achieved either by autoclaving or by putting the eluates through a Millipore filter;the active compounds therefore are not formed by heating.

Since materials A and B were so distinctly separated by the above solvent sys-tems, a larger-scale separation of the two was attempted with n-butanol. Ethanolwas evaporated from the crude extract and the remaining water solution extractedwith four successive equal volumes of n-butanol. A portion of the combined bu-tanol fractions and an equivalent portion of the water layer were chromatographedwith solvent VI. Materials A and B can be separated nicely by this procedure(Fig. 3). The results support the conclusion that A and B are different chemicalentities.

Resolution of Band B.-Since various purine bases and their respective nucleo-sides tend to run closely together in the solvent systems thus far described, thepossibility that band B might contain the nucleoside of zeatin was investigated.Band B was eluted from a previous chromatogram (solvent VI) of the crude extract,

~~~~~~~90-~~~~~~~~~~~~~~~~~~~9

o90 soson

E - Alk. PhosEX~~~~~~~~~~~~~~~~~~~~~~~6E/ Control

RtGo 6o 0 - /W X CD3 - - S2W40-X - 40 -

la 30 to0 0

05~~~~~~~~2

FIG. 4.-Soybean tissue f FIG. 6.-Soybean tissueassay of a chromatogram FIG. 5.-Soybean tissue assay of chromatograms de-developed with distilled assay of a chromatogram de- veloped with water-saturatedwater. The chromato- veloped with 0.03 M boric sec-butanol. Control repre-graphed material had been acid adjusted to pH 8.4 sents band A eluted from apreviously eluted from the with NaOH. The chro- previous chromatogram.fast band of a chromatogram matographed material had Alk. Phos. represents band Asuch as represented in Fig. 2. been previously separated eluted from a previous chro-The average fresh weight of from other active materials matogram and then treatedcontrol pieces was 8 mg. by butanol extraction (see with chicken intestine al-The Rf of synthetic zeatin text and Fig. 3). The con- kaline phosphatase. Con-was 0.50. trol pieces averaged 4 mg. trols averaged 3 mg.

and the eluate was rechromatogaphed with distilled water which separates the basesfrom the more rapidly moving nucleosides. Assay of the chromatogram indicatedtwo active bands, the slower of which (B2) moved to the position of synthetic zeatin(Fig. 4). The faster band (B1) was roughly in the region where one would expect anucleoside of zeatin. A chromatogram prepared in the same way, but developedwith 0.03 M boric acid adjusted to pH 8.4 with NaOH, gave a more distinct separa-tion of B1 and B2. This same pattern of resolution was observed when a n-butanolextract (see earlier) was chromatographed with the borate solvent (Fig. 5). In theborate system, the position of B2 was but little changed from its position when wa-ter was used as the solvent and again its Rf corresponded with that of syntheticzeatin (center of spot at Rf 0.51). B1, however, showed a greater migration whenthe borate was present; this would be expected of a nucleoside.9

Nature of Band A.-When band A was rechromatographed, it usually yielded asmall amount of B, thereby indicating that the latter may arise from A. Since Bactually consists of two factors, one of' which may be a nucleoside, it seemed that Acould be a phosphate derivative of B1. This possibility was tested by treating a

VOL. 54, 1965 BOTANY: C. 0. MILLER 1055

sample of A, which had been eluted from a previous chromatogram, with chickenintestine alkaline phosphatase (Worthington Biochemicals) for 2 hr at 320C. Thereaction mixture included 0.01 M MgCl2, 0.1 M trishydroxymethylaminomethane(Tris) at pH 8.2, and the enzyme at a concentration of 1 mg/ml. The product anda control sample of A which had been incubated under similar conditions in theabsence of the phosphatase were chromatographed using solvent VI (Fig. 6).Treatment with the phosphatase destroyed A but yielded B. Additional chroma-tography of the hydrolysate in the borate solvent revealed that B, and not B2 isproduced by the treatment with the phosphatase; that is, phosphatase attacks A andyields a compound which may be a nucleoside. Identical results were obtainedwhen calf intestine alkaline phosphatase (Calbiochem) was used, but no hydrolysisoccurred when potato acid phosphatase (Calbiochem) was employed. The knownspecificity of the chicken enzyme indicates that perhaps A is a monophosphate nu-cleotide.

Factor A remained in the supernatant when a cold aqueous solution containing itwas adjusted to pH 9.0 and saturated with barium acetate (pH 9.0 maintained).The factor was precipitated, however, when 95 per cent ethanol (4 vol) was addedto this saturated solution. Addition of the ethanol in the absence of the bariumacetate did not bring down the factor. Factor A was recovered by dissolving theprecipitate in acid and removing the barium with sulfate.

Further data concerning the properties of A, B1, and B2 were obtained by chroma-tographing the crude maize extract on the anion exchanger Dowex 1.10 After re-moving the ethanol, the remaining aqueous extract was adjusted to pH 7.8, filtered,and applied to a 1.2 X 18-cm column of Dowex 1 (formate, X2, 200-400 mesh).The resin was washed with 200 ml of distilled water. The combined effluent andwash contained B, and B2. The column was developed with a modified convexgradient of formic acid by running first 60 ml of 1 M HCOOH into a mixing flaskreservoir containing 50 ml distilled water and then adding quantities of 4 MHCOOH. An active factor as detected by the soybean tissue test was eluted onthe trailing shoulder of adenylic acid (about 45 ml 4 M HCOOH), and another inmuch smaller quantity came off after uridylic acid (about 250 ml 4 M HCOOHadded). The latter factor occurred in such small quantity that nothing furtherhas been accomplished with it. The active substance coming off after adenylicacid, however, migrated as factor A on the paper chromatograms and yielded B,when treated with alkaline phosphatase.The regenerated material from the ethanol-barium precipitate discussed above

also has been subjected to chromatography by the Dowex 1 method. The activematerial was held by the resin and was largely eluted from the column on the trail-ing shoulder of adenylic acid. This material chromatographed on paper as A andwas converted to B, by alkaline phosphatase; it also moved very close to adenylicacid when chromatographed on DEAE cellulose paper (Whatman DE-81) with0.05 M formic acid as the solvent." A small amount of activity also eluted fromthe Dowex 1 column shortly after uridylic acid.

All of this information is consistent with the idea that A is a nucleotide of zeatin.To test this idea further, A was treated by the method of Khym and Cohn."2 Inthis method, a nucleotide may be hydrolyzed to a free base and a sugar phosphateby a short heat treatment. Two ml of an aqueous solution of A (obtained by bar-

1056 BOTANY: C. 0. MILLER PROC. N. A. S.

ium precipitation and subsequent Dowex 1 chromatography) was mixed with 1 mlof Dowex 50 (H+) and the mixture heated in a boiling water bath for 150 see andimmediately cooled. The surrounding liquid was filtered off and the resin washedwith 5 ml distilled water. This combined filtrate and wash contained no mate-rial active in the soybean cell-division test. The resin was next converted to theNH4+ condition by washing with an excess of 6 N NH40H. The eluate was freedof ammonia and chromatographed on paper with the borate solvent. Activity wasdetected only at the location to which synthetic zeatin migrates (Rf 0.51). A simi-lar chromatogram of A not treated with Dowex 50 but nevertheless exposed to NH4-OH yielded only an active band at Rf 0.8-1.0. The Dowex 50 hydrolysis thereforeproduces a compound which may be zeatin.

If A really contains zeatin, it should be inactivated immediately by a weak neu-tral solution of potassium permanganate. The double bond in the allyl group ofzeatin is immediately broken and oxidized by such treatment, and the inactive N-(purin-6-yl)glycine is formed.'3"16 A solution of A therefore was evaporated to dry-ness and a 0.01 per cent aqueous solution of KMnO4 was added to the residue untilthe permanganate color just began to persist for more than a few seconds. An ex-cess of ethanol was then added to complete the decomposition of the permanganate.After evaporation to dryness, the preparation was assayed and showed a completeloss of activity. A run in which an even greater quantity of permanganate wasdecomposed with ethanol and the water washings of the residue added to a controlcontaining kinetin showed no effect.

Nature of B,.--B, sometimes produced very small quantities of B2, especiallywhen the particular experiment involved prolonged procedures. B2 was also formedfrom B. by treatment with Dowex 50 (H+) as described for A. In addition to this,B2 was produced when a small quantity of 0.1 per cent sodium metaperiodate9was mixed with B, for 5 min and the mixture subsequently brought to a boil for 1 miisin 4 N NH40H. Treatment of B, with boiling NH40H alone did not produce B2.The activity of both B, and B2 was destroyed by the potassium permanganate treat-ment.

Discussion.-Both B2 and zeatin have been obtained from the same vegetablesource, show the same type of biological activity, migrate to the same positions inthe eight solvent systems investigated, are very sensitive to permanganate, aresoluble in n-butanol, do not exchange onto Dowex 1 resin (at pH used) but do ex-change onto Dowex 50. These similarities all strongly suggest that B2 is zeatin.The facts that B, may be converted to B2 (and therefore presumably to zeatin) bybreakdown during handling and by the Dowex 50 hydrolysis, that it is highly sensi-tive to permanganate, that its chromatographic behavior is strongly influenced byboric acid (which complexes with cis-hydroxyl groups), that it is sensitive to peri-odate (which attacks adjacent hydroxyl groups), that it is formed fromA by alkalinephosphatase treatment, and that it is soluble in n-butanol strongly support but donot absolutely prove the interpretation that B, is a nucleoside of zeatin. Greaterquantities of purified B, are needed to check the nature of the sugar.

Since A is converted to B, by chicken intestine alkaline phosphatase or by pro-longed handling, is precipitated by barium in the presence of ethanol, acts as ananion by exchanging onto Dowex 1 resin, elutes from Dowex 1 much as does ade-nylic acid (zeatin is a substituted adenine), and is converted to B2 by the Dowex 50

VOL. 54, 1965 BOTANY: C. 0. MILLER 1057

heat treatment, the probability seems great that A is a monophosphate nucleotide ofzeatin.The small amount of active material precipitated by barium ions and eluted from

the Dowex 1 column after uridylic acid may be a nucleotide of zeatin with morethan one phosphate group, but there is no real proof of this, other than the fact that itelutes much as one would expect of similar nucleotides of adenine. In any case,the likelihood that at least a major portion of the cell-division activity in maizekernels is due to zeatin or its derivatives seems strong indeed. One would antici-pate that the same or a similar situation might be found in other plant materials.When the nature of zeatin was considered, this array of close derivatives was antici-pated.3 Furthermore, the array was to be expected from the results obtained whenradioactive compounds such as benzyladenine were supplied to plant materials;apparent glycosides (nucleosides) and nucleotides were recovered.9 14, 15Some published discussion has dealt with the possibility that zeatin as such does

not occur naturally. Kefford7 has pointed out pitfalls of using a resin such as Dowex50 and has stated that zeatin is unlikely to be a native substance. Miller andWitham"6 held to the opposite view with respect to natural occurrence of the com-pound, since they were able to demonstrate that active material in cold ethanol ex-tracts of maize migrated with zeatin; however, it is now evident that the nucleosidealso moves closely with free zeatin in the solvent systems which they used. In thepresent work, B2, which appears to be zeatin, was obtained without any use of theDowex 50. The chances are substantial, however, that some zeatin is freed fromthe nucleoside and nucleotide forms when resins such as Dowex 50 are used for itsisolation. Quantitative estimates of zeatin separated from other plant materialsin this way must be subject to close scrutiny; working at low temperatures and pre-venting localized heating, such as can occur when elution is made with NH40H,should lessen the error. One must also consider the possibility that zeatin is formedfrom B1 even with the gentle procedures employed in the present work. Theamount of B2 obtained from rechromatography of the B1 produced by alkalinephosphatase, however, has always been quite small compared to that found in then-butanol extracts or the eluates from the B band of chromatograms of the crudeethanol extracts. Furthermore, chromatography of cold ethanol extracts with theboric acid system has yielded activity at the position of zeatin. Thus, the evidencefor the natural existence of the free compound is considerable. Of course, the pointof critical importance is that zeatin probably occurs naturally not only as the freebase but also as the active core of several derivatives.Many unsuccessful attempts to find an in vitro action of zeatin and especially of

the related compound kinetin (6-furfurylaminopurine) have been made. Most ofthese attempts have been with the free compound. In view of the fact that mostpurines or synthetic analogues show activity only after they are incorporated intomore elaborate forms, the use of the more complicated derivatives of zeatin or kine-tin in the attempts to find an in vitro reaction seems warranted.Summary.-The materials extracted from maize kernels and active in promoting

cell division in plant tissue cultures have been resolved into at least four activefractions. One of these seems to be zeatin, another behaves as a nucleoside of zea-tin containing cis-hydroxyl groups, and a third by several criteria appears to be amonophosphate nucleotide of zeatin. Evidence for the natural existence ofzeatin is presented.

1058 PHYSIOLOGY: TWENTE AND TWENTE PROC. N. A. S.

The author wishes to thank Mr. Robert A. Brookshire for his excellent assistance during thecourse of this work, and H. Rickenberg and G. Williams for valuable suggestions.

* Supported by grants G 24052 and GE 3612 from the National Science Foundation.1 Beauchesne, G., M. Leboeuf, and R. Goutarel, in Regulateurs Naturels de la Croissance Vegtale

(Paris: Centre National de la Recherche Scientifique, 1964), p. 119.2 Letham, D. A., in Regulateurs Naturels de la Croissance Vegetale (Paris: Centre National de

la Recherche Scientifique, 1964), p. 109; and ref. 6 (below).3Miller, C. O., these PROCEEDINGS, 47, 170 (1961); and refs. 13 and 16 (below).4Letham, D. S., and C. 0. Miller, Plant Cell Physiol., in press.6 Letham, D. S., J. S. Shannon, and I. R. McDonald, Proc. Chem. Soc., 1964, 230.6 Letham, D. S., Life Sci., 2, 569 (1963).7 Kefford, K. P., Science, 142, 1495 (1963).8 Miller, C. O., in Modern Methods of Plant Analysis (Berlin: Springer-Verlag, 1963), vol. 6,

p. 194.9 McCalla, D. R., D. J. Moore, and D. Osborne, Biochim. Biophys. Acta, 55, 522 (1962).

10 Hurlbert, R. B., H. Schmitz, A. F. Brumm, and V. R. Potter, J. Biol. Chem., 209, 23 (1954).11 Jacobson, K. B., Science, 138, 515 (1962).12 Khym, J. X., and W. E. Cohn, J. Am. Chem. Soc., 76, 1818 (1954).13 Miller, C. O., Plant Physiol., 37, xxxv (1962).14 Loeffier, J. E., and J. Van Overbeek, in Regulateurs Naturels de la Croissance V~ggtale (Paris:

Centre National de la Recherche Scientifique, 1964), p. 77.15 Fox, J. E., Plant Physiol., 39, xxxi (1964).16 Miller, C. O., and F. H. Witham, in Rlgulateurs Naturels de la Croissance V~ggtale (Paris:

Centre National de la Recherche Scientifique, 1964), p. I (erratum).

REGULATION OF HIBERNATING PERIODS BY TEMPERATURE*BY JOHN W. TWENTE AND JANET A. TWENTE

UNIVERSITY OF UTAH, SALT LAKE CITY

Communicated by Henry Eyring, August 5, 1965

Hibernation of golden-mantled ground squirrels (Citellus lateralis) is characterizedby periods of dormancy which are interrupted at intervals by arousal with a returnto the homeothermic state.1 We have shown2 that the duration of hibernatingperiods of this species is related to the body temperature achieved; hibernatingperiods at 40C averaged approximately twice as long as those at 11 'C.

These observations suggested that the following should be investigated: overwhat temperature range is hibernation possible; is the temperature dependencyof the duration of hibernating periods continuous over the possible range; and ifthe temperature dependency is continuous, what function does the time/tempera-ture relationship best approximate.Methods.-Adult golden-mantled ground squirrels of both sexes were caged singly and main-

tained undisturbed in two dark, relatively soundproof, constant-temperature rooms. Maximumtemperature fluctuations at the specific places in cages where hibernation occurred (the micro-environmental temperature) were less than 0.50C.

Thirty-gauge, iron-constantan, Teflon-coated thermocouples (0.050C) were implanted sub-dermally 1-2 months prior to experimentation with the sensing element tied to a lower rib in themidventral region.2 When ground squirrels were curled in the hibernating posture, temperaturesrecorded from the rib area were not different from heart and liver (core) temperatures as recordedby a hypodermic thermocouple probe. Temperature records were printed at 6-min intervals on