PlantPhysiol.

(1988) 88, 315-320

0032-0889/88/88/0315/06/$0 1.00/0

Productionin Phalarisaquatica
N,N-Dimethyltryptamine
Seedlings
A MATHEMATICALMODEL FOR ITS SYNTHESIS
Receivedfor publicationDecember31, 1987and in revisedform April26, 1988

JOSEPHP. G. MACK, DAWN P. MULVENA, AND MICHAEL SLAYTOR*

ChemistryDepartment,Universityof MarylandBaltimoreCounty,Catonsville,Maryland21228
(J.P.G.M.),and Departmentof Biochemistry,Universityof Sydney,
Sydney,N.S. W.,2006, Australia(D.P.M., M.S.)

ABSTRACI grasses, particularlyunpalatabilityto sheep, which has been

linked in P. aquaticato the graminecontent (4), a considerable
The activitiesof threeenzymesandthe concentrationof intermediates
amount is known of the factorsinvolvedin alkaloidproduction.
involvedin the synthesisof N,N-dimethyltryptamine (DMT) fromendog- For example,genetic evidence in P. arundinaceaindicatesthat
enous tryptophan(TRP) have been measuredin vitro in seedlings of gramine synthesis is independent of DMT' and 5MeODMT
PhalarisaquaticaL. cv AustralianCommercialover 16 days afterplant- formation(1 1). This is furthersupportedby the findingthat the
ing. The activitiesof tryptophandecarboxylaseand the two N-methyl- N-methyltransferase activitiesinvolved in graminesynthesisare
transferasesincreasedrapidlyto maximalrates of substrateconversion different from those involved in the synthesis of DMT and
at day 5 of 95, 1000, and 2200 micromolesper hour per milliliter, 5MeODMT(13).
respectively.Afterthese maximalrates, the activitiesdecreasedrapidly. Feeding experimentswith P. aquatica have establishedthe
The concentrationof intermediatesincreasedrapidlyfrom zero in the biosynthesisand turnoverof DMT and SMeODMT(1), while
seeds to maximalvaluesof 25 and53 micromolarat day 5 for tryptaminethe biosynthesisof graminehasbeen studiedin the closelyrelated
(T) andN-methyltryptamine (MT), respectively,1000 micromolarat day
reedcanarygrass,P. arundinacea(9). Enzymesfrom P. aquatica
6 for TRP, and650 micromolarat day 8 for DMT. The concentrationof that catalyzethe in vitroformationof DMT have been partially
DMT and of all the intermediatesin its synthesisdeclinedrapidlyafter
characterized,namely TDase (2) and the two SAM-dependent
the maximal value had been reached. A mathematicalmodel of the N-methyltransferases, PIMaseand SIMase(10), which methylate
pathwayfromTRP to DMT using these enzymescorrectlypredictsthe T and MT, respectively.Becauseof the smallnumberof enzymes
concentrationsof T andMT, intermediateswhoseconcentrationis deter- involved,the synthesisof DMT from TRP offersthe possibility
minedonly by the pathway,and confirmsthat these three enzymesare of completelycharacterizingan alkaloid-synthesizingsystemand
responsiblefor the in vivosynthesisof DMT. Kineticstudiesare reported
of testingthe hypothesisthat the enzymesare responsiblefor the
for these enzymes.Tryptophandecarboxylaseuses pyridoxalphosphate in vivo synthesis of DMT.
(PALP) as a coenzymeand has the followingkineticconstants:KJFALP= Thispaperreportsthe activitiesof enzymesand concentrations
2.5 micromolar, KIRP = 200 micromolar, KIT = 5 millimolar, and of reactantsof the DMT synthesizingpathwayas a function of
KMT = 4 millimolar.The N-methyltransferases
use S-adenosylmethio- age of P. aquaticaseedlingsfor the first 16 d of seedlinggrowth.
nine (SAM) as substrate;S-adenosylhomocysteine (SAH) is assumedto These data togetherwith the kinetic constants of the enzymes
be the product.The mechanismof secondaryindolethylamine-N-meth- have been used to model the DMT synthesizingsystem mathe-
yltransferase,determinedby initialvelocitystudies,is rapidequilibrium maticallyduringthis initial 16 d of seedlinggrowth.
randomwith formationof both dead end complexes.Secondaryindol-
ethylamine-N-methyltransferasemethylatesbothMT and5-methoxy-N-
methyltryptamine (5MeOMT).The kineticconstantsforthe methylation
of MT are:KMT = 40 ? 6, KSAM = 55 ? 15, KDMT = 60, KSAH = 4.3 ?
MATERIALS AND METHODS
0.4 micromolarwith unity interactionfactors.The kineticconstantsfor Labeled Compoundsand Chemicals. L-['4C-Methylene]tryp-
the conversion of 5MeOMT to 5-methoxy-N,N-dimethyltryptaminetophan (["4C]TRP) and ['4C-methyl]SAM (['4C]SAM)were pur-
(5MeODMT)are KsMeOMT = 40 ? 10, KSAM = 90 ? 40, andKSAH = 2.9 chased from AmershamAustraliaPty Ltd. PALP and the sub-
? 0.3 micromolarwith unity interactionfactors, except for SAM- strateswere obtainedfrom SigmaChemicalCo., St Louis, MO.
5MeODMT= 2.0 ? 0.9 andSAH-5MeOMT= 0.45 ? 0.25. The kinetic [I4C]SAMusedwas assayedfor SAH contentby chromatography
constantsfor primaryindolethylamine N-methyltransferaseare RT.= 20, on Whatman 3MM paper using butanol/acetic acid/water
= 40, KMT = 450 micromolar with the substrates binding inde-
KSmAM (12:3:5).The nucleosideswere detectedwith Cd-ninhydrin(3),
pendently. elutedwith methanol,and quantitatedby absorbanceat 500 nm.
The SAM used in the assay contained0.024% SAH, which, at
the highestconcentrationof SAM used (200 ztM), gives an SAH

'Abbreviations: DMT, N,N-dimethyltryptamine;5MeODMT, 5-

methoxy-N,N-dimethyltryptamine;TDase, tryptophan decarboxylase;
SAM,S-adenosylmethionine;PIMase,primaryindolethylamineN-meth-
Phalarisspp., notably Phalarisaquaticaand Phalarisarundi- yltransferase;SIMase,secondaryindolethylamineN-methyltransferase.
nacea, are major pasture grasses in many parts of the world. T, tryptamine;MT, N-methyltryptamine;
TRP, tryptophan;PALP,pyr-
wrt, with respectto.
Because there are undesirable features associated with these idoxalphosphate;SAH, S-adenosylhomocysteine;
315
316 MACK ET AL. PlantPhysiol.Vol. 88, 1988
concentrationin the assayof 0.05 aM. This is small comparedto The trendswere the same in each experiment,but there was a
KSAH for the enzymesand the concentrationof SAH used in the variationby up to half a day of the time at which the plumule
inhibitorbindingstudies. emerged through the sand. As the maximal activities of the
Preparationof EnzymeExtracts. Seedlingsof Phalaris aqua- enzymesand the concentrationsof intermediateswererelatedto
tica L. cv AustralianCommercialwere grown as describedpre- this event, it was felt it would be more appropriateto presentthe
viously (10). Whole seedlings(i.e. all plant materialincluding resultsfrom a singleexperimentratherthan as SE means.
the testa for pre-emergentseedlings)(5 g) were homogenizedin Total Nitrogen Determination.Forty mg of seedlings were
the assaybuffer(1:2, v/w) with sand. An additional1%(v/v) of digestedwith 3.0 mL of 18 M H2SO4, 1.5 g of K2SO4, and 0.1 g
mercaptoethanolwas addedto the homogenate,which was then of HgO at 300?C;NH3 producedwas estimatedby the phenol/
centrifugedat 5000g for 5 min. The supernatant(1 mL) was hypochloritemethod (7).
passedthrougha SephadexG-25 (fine) column (1.5 x 6 cm) to Estimationof SAM. Five g of seedlingswere homogenized
separatethe proteinsfrom endogenousalkaloidsand other low with TCA (10%;15 mL) and werecentrifugedat 2000g, and the
mol wt compounds. The excluded protein fraction (3 mL), supernatantwas washed with diethyl ether (3 x 15 mL). The
referredto as the enzyme extract, was collected and used for extractwas made 15 mmwrt to HCIand then chromatographed
enzyme assays.It behavedsimilarlywith respectto linearitywith as previouslydescribed(5).
time and enzyme concentrationand gave the same kinetic con-
stants as the original preparation.No differences in kinetic RESULTS
constantswereseen for preparationsfromleavesor whole plants.
The kineticswere investigatedat constant pH and the effect of Changesin TryptophanandAlkaloidConcentrations. Cellfluid
hydrogenion concentrationwas not investigated. weight increasedfrom 8 to 90% of the fresh weight in the first
16 d of seedlinggrowth.Over the 16 d of the experiment,the
EnzymeAssays. TDase. The assay mixture consisted of en-
nitrogencontent of the seedlingsremainedconstantat 105 ? 10
zyme extract(300 ,uL),['4C]TRP(1.5 mM;specificactivity0.07
,uCi/mmol),PALP (25 gM), Na-phosphate(0.1 M; pH 7.6) in a ,umol/100 seedlings.No freeindolecompoundscouldbe detected
in the dry seeds, and the amount of TRP in the seed proteins
total volume of 400 ,uL.This was incubatedfor 2 h at 25?Cin a was 5 ,umol/100 seeds. The concentrationof SAM at d 6, the
scintillationvial. The reactionwas stoppedby adding1 M H3BO3- day at which TRP concentrationwas maximal,was 30 ,gM.The
Na2CO3(2 mL; pH 10.0), and T was extractedinto toluene earliest time that any of the 3 compounds involved in the
scintillantas previouslydescribed(10). Activity is measuredas synthesisof DMT, namely TRP, T, and MT, could be detected
nmol T h-1 mL-' of cell fluid. Cell fluid was calculatedfrom the was 3 d afterplanting(TableI and Fig. 1). The concentrationof
differencebetweenfreshand dried(100?Cfor 1 h) weightsof the TRP increasedrapidlyto 1000 ,M in seedlingsat d 6 (Table I).
seedlings. The plumule emerged through the sand between d 6 and 7.
PIMase and SIMase. The assaymixtureconsistedof enzyme Thereafter,the concentrationof TRP decreasedsteadily.Simi-
extract(60 ,L), T (for estimatingPIMaseactivity)(1 mM),MT larly,the concentrationsof T and MT increasedrapidlyfrom d
(for estimating SIMase activity) (1 mM), [4C]SAM (0.4 mM; 3, the earlieststage at which they could be detected,until d 5,
specific activity 0.5 mCi/mmol), and Tris buffer (0.25 M; pH and then they decreasedrapidly(Fig. 1). The maximumconcen-
8.5) in a total volume of 100 ,uL.This was incubatedfor 30 min trations of T and MT were 25 and 53 IAM,respectively.The
at 25?Cin a scintillationvial. Activity is measuredas nmol of maximum concentrationof DMT was 650 ,uM at d 8, thereafter
"4C-methylfrom SAM incorporatedper h into the tryptamine decliningquickly.
alkaloids per mL of cell fluid. The reactionswere stopped by Activities of TDase, PIMase, and SIMase. None of the en-
adding 1 M H3BO3-Na2CO3 (2 mL;pH 10.0),and MT (or DMT) zymes, namely TDase, PIMase,and SIMase,could be detected
was extractedinto toluene scintillant as describedpreviously in P. aquatica before the 3rd d after sowing (Table I); they
(10). increasedrapidlyuntil just before the seedlingsemerged from
Extractionand Estimationof Tryptaminesand TRP. T, MT, the sand (d 6) and then declined until d 14. The maximum
DMT, and TRP wereextractedfrom plant materialas described activity of TDase was 95 Amol h-'mL-'. The profiles of the
previously(14). T, MT, and DMT wereseparatedand estimated changesin concentrationof the two N-methyltransferases were
by two-dimensionalTLC (14). This method did not separate similarand decreasedat a slowerratethan TDase.
TRP from 5-methyoxytryptophan. Accordingly,the extractwas The maximum activitiesfor PIMase and SIMasewere 1000
evaporatedto dryness, dissolved in 1 mL of 0.2 M (wrt Na+)
citratebuffer,pH 2.2 (12), and estimatedon a JLC 6AH amino Table I. Variationwith Time of the Concentrationof TRPand the
acid analyser(JEOLCo. Ltd, Tokyo) using LCR-2 cation ex- Activitiesof TDase,PIMase,and SIMase in P. aquatica
changeresinon a column (45 cm x 8 mm) at 53?Cat a flow rate Day TRP TDase PIMase SIMase
of 0.84 mL/min. Threecitratebuffers(12) wereused sequentially
for differinglengthsof time, namely(wrtNa+)0.2 M buffer(pH MM Mmol-'mL-1
3.3) for 110 min, 0.2 Mbuffer(pH 4.46) for 50 min, and 0.35 M 3 333 6.77 222 583
buffer (pH 5.40) for 130 min. Under these conditions, TRP 4 444 55.6 687 1630
separatedfrom 5-methoxytryptophan and the otheramino acids 5 733 94.7 1000 2190
in the extractand was eluted at 69 mL after the bufferchange 6 1000 55.9 767 1500
from pH 4.46 to 5.40. 7 767 26.7 606 1370
The TRP in the proteinsin the seedswas estimatedas follows: 8 826 17.4 601 1210
100 mg of pulverizeddry seeds were incubatedfor 18 h at 40?C 9 728 12.0 440 982
with pronasesolution (1.0 mL; 5.5 mg/mL in 0.2 M phosphate 10 630 6.67 381 851
buffer, pH 7.4). TRP in aliquots (0.1 mL) of the incubation 11 533 2.78 330 737
mixturewas estimatedby boilingthe aliquotwith 1.0 mL of van 12 436 1.16 286 639
Urk-Salkowskireagent(14) for 10 min, cooling, and readingthe 13 340 0.72 217 490
absorbanceat 600 nm. Appropriatecontrols were included to 14 267 0.45 164 376
ensurethat the pronasewas not self-digesting. 15 217 0.27 157 343
Reproducibilityof Results. The concentrationsof intermedi- 16 167 0.0 150 315
ates and the activities of enzymes were measured three times.
 ALKALOID SYNTHESIS IN PHALARISAQUATICA
50
40-
30-
0 The inhibitionpatternsfor SIMaseare for a rapidequilibrium
F-20
10
4 8 12
Day
FIG. 1. Variationwith time of T and MT in P. aquaticaseedlings.
Resultspresentedfor T (0) and MT (a) are fromone experiment;values
calculatedfromthe numericalsimulationare T (0) and MT (I).
and 2200 ,mol h-'mL-', respectively.The high levels of PIMase
and SIMaserelativeto that of TDase could easily convertthe T
producedby TDase to DMT and resultedin a low concentration
of the intermediatesT and MT.
Kinetic Analysis. Inhibitiontype nomenclatureand interac-
tion factorswereused accordingto Segel(15). Interactionfactors
are the ratio of Kd-,,from the ternarycomplex to Kd,sSfrom the
binaryenzyme complex.
Kineticconstantsweredeterminedusinginitialvelocityexper-
iments, and the data were analyzedby a least squarefitting of
intersectingstraightlines to Lineweaver-Burkplots using the
program Labtech Notebook (LaboratoryTechnologies Corp,
Cambridge,MA) run on an IBM PC computer. Data for the
Lineweaver-Burkplots were weighted inversely to the Line-
weaver-Burky(= I v) value to reflectequalweightingfor each of
the originaldata points. Resultswere not significantlydifferent
with an unweightedfit, indicatingno bias in the data at high or
low reactionvelocity.
The mechanismsof TDase and PIMasewere not determined,
and kineticconstantsused steadystate nomenclatureof Kmand
Ki. The binding of TRP to TDase, which was tested at concen-
trationsof 5 and 25 ,M of PALP,had a smallslope and intercept
effect of about equal size giving KP,ALP, 2.5 (,M)). MT and
317
DMT are competitiveinhibitorsof TDase. K' for MT and DMT
for the E-PALPcomplex was high relativeto the concentrations
of inhibitorused (1 mM), giving KiMT 5 mM and KiDMT 4
mm. These Ki values were large comparedto KmTRP, consistent
with the inabilityof these productanalogsto forman imine with
PALP.
Kineticconstantsderivedfrom initialvelocityexperimentsare
listed for PIMase in Table II, for the methylation of MT by
SIMasein Table III, and for the methylationof 5MeOMT by
SIMasein TableIV.
random bi-bi mechanism with formation of both dead end
complexes.The productinhibitionplots intersectnear the axes.
The bindingpairs,DMT-MT and SAH-SAM,were good fits to
a competitive inhibition mechanism, and, on the assumption
that these pairs of ligands bind mutually exclusively,the data
werefittedto a Lineweaver-Burk plot constrainedto intersecton
the 1/v axis. The fits to the graphsfor the DMT-SAMand SAH-
MT pairswere not constrainedfor the position of intersection.
The binding constant for any ligand should be independentof
the other ligandsthat are used in the experiment.For example,
KSAM shouldbe the same whetherit was determinedfor experi-
ments with MT or 5MeOMT as the variedligand and with or
without any inhibitorspresent.Kinetic constantsvaryby up to
a factor of 2 (Table III) indicating that the partiallypurified
enzyme preparationsused in any particularexperimentwere not
alwaysidentical.Thus, the magnitudesof the Kdi,svaluescannot
be regardedas beingmore accuratethan a factorof 2. Inspection
of interactionfactors shows that with the normal number of
methyl groupsfor the reactionat the active site of SIMase,i.e.
SAM and MT or 5MeOMT,the ligandsbind independently.An
extra methyl group, i.e. SAM and DMT, decreasesthe binding
of the second ligand by a factor of 3, or, with 5MeODMT, a
factorof 2. One less methyl group, i.e. SAH and MT, increases
the binding of the second ligand by a factor of 5, or with 5
MeOT, a factorof 2.
MathematicalModel for DMT Synthesis.To determineif the
synthesis of T, MT, and DMT could be described in terms of
the enzymes TDase, PIMase, and SIMase, a model for the
synthesisof DMT from TRP was constructedfrom the data in
Table I and the kinetic constantsin Scheme I. These constants
are derivedfrom the kinetic data in Tables II to IV. Velocities
for the three enzymes for the numericalsimulationwere calcu-
latedusingthe rateequationforthe mechanism.The mechanism
of SIMaseis rapidequilibriumrandombi-bi with both dead-end
complexes formed. Because of the many similaritiesbetween
PIMaseand SIMase,the mechanismfor PIMasewas assumedto

Table II. KineticConstantsfor TDaseand PIMase

Enzyme VariedLigand Inhibitor tion Constant Interaction

AM
TDase PALP,TRPb KmTRP 200
KmPALP 2.5 1.0
TRPc MT C KTMT 5000
TRVc DMT C K?DMT 4000
PIMase T, SAMd KMT 17 ? 3 1.2 0.4
K,nSAM 38 ?7 1.?4
Te DMT C K1DMT 450
a b Initialvelocitieswere measuredat five concentrationsof
TRP (I100-1000AM) in the
C, competitive.
presence of two concentrationsof PALP (5 and 25 ,M). CInitial velocities were measuredat eight
concentrationsof TRP (40-500 AM) in the presence of two concentrationsof MT or DMT (0 and 1
mM). d Initialvelocitiesweremeasuredat five concentrations of T (0.01-0.075 mM) in the presenceof four
concentrationsof SAM (5-40 gM). e Initial velocities were measuredat five concentrationsof T (0.01-
0.075 mM)and two concentrationsof DMT (0 and 1 mM).
318 MACK ET AL. Plant Physiol. Vol. 88, 1988

Table III. KineticConstantsfor SIMase withMT and SAM as Substrates

VariedLigand Inhibitor T ipeo Constant Factor

,uM

SAM, MTb KMT 40 ? 6 1.15 ? 0.20

KSAM 55 ? 15
MTC DMT C KMT 120 ? 20
KDMT 105 ? 10
SAMd DMT NC KSAM 130 ? 20 3.1 ? 0.9
KDMT 35 ? 10
MTe SAH NC KMT 120 ? 40 0.18 ? 0.08
KSAH 25 ? 10
SAM' SAH C KSAM 45 ? 5
KSAH 4.3 ? 0.4
aC, competitive;NC, noncompetitive. b Initial velocitieswere measuredat five concentrationsof MT
(100-2000 gM) in the presence of four concentrationsof SAM (20-200 ,M). CInitial velocities were
measuredat four concentrationsof MT (100-1000 ,uM)in the presenceof five concentrationsof DMT (0-1
mM).Concentrationof SAM = 200 ,uM. d Initialvelocitiesweremeasuredat 3 concentrationsof SAM(40-
200 Mm)in the presenceof fiveconcentrationsof DMT (0- 1 mM).Concentrationsof MT = 1000AM. e Initial
velocitieswere measuredat four concentrationsof MT (100-1000 AM)in the presenceof five concentrations
of SAH(0- 1 mM). Concentrationof SAM= 200 gM. 'Initialvelocitiesweremeasuredat fourconcentrations
of SAM (20-200 gM)in the presenceof six concentrationsof SAH (0-20 AM).Concentrationof MT = 1000
MM.

Table IV. KineticConstantsfor SIMase with5MeOMTand SAM as Substrates

VariedLigand Inhibitor Inhibition Constant Factors

SAM, 5MeOMTb K5MeOMT 40 ? 10 0.85 ? 0.15

KSAM 90 ? 40
5MeOMTC 5MeODMT C K5MeOMT 75 ? 10
K5MeODMT 83 ? 6
SAMd 5MeODMT NC KSAM 66 ? 5 2.0 ? 0.8
K5MeODMT 40 ? 10
5MeOMTe SAH NC K5MeOMT 60 ? 25 0.45 ? 0.25
KSAH 7? 4
SAM' SAH C KSAM 25 ? 4
KSAH 2.9 ? 0.3
b
aC, competitive; NC, noncompetitive. Initial velocities were measured at five concentrationsof
SMeOMT(100-2000 four
Mm)in the presenceof concentrationsof SAM (20-200 Mm). CInitialvelocities
were measuredat four concentrationsof 5MeOMT(100-1000 AM) in the presenceof five concentrationsof
5MeODMT (0-1 mM). Concentrationof SAM = 200 AM. dInitial velocities were measuredat four
concentrationsof SAM (20-200 Mm) in the presence of five concentrationsof SMeODMT (0-1 mM).
Concentrationof SMeOMT = 1000 pM. e Initial velocities were measuredat four concentrationsof
SMeOMT(100-1000 Mm)in the presenceof five concentrationsof SAH (0-1 mM).Concentrationof SAM =
200 Mm. 'Initial velocitieswere measuredat five concentrationsof SAM (20-200 AM) in the presenceof
six concentrationsof SAH (0-20 pM). Concentrationof SMeOMT= 1000 MM.

be the same as the SIMase.For TDase, PALP was assumedto
be saturatingand the enzyme was treatedas a uni-uni enzyme.
Inspection of the experimentallydeterminedconcentrations
of intermediatesT and MT (Fig. 1) shows that their pool size is
small comparedto the amount of DMT producedand that the
rate of change of concentrationof these intermediatesis small
comparedto the flux of the pathway.For all practicalpurposes,
all TRP consumed by the pathwayappearsas DMT, and the
concentration of the indolethylamineintermediatesis steady
state. The worst case approximationoccurs on day 4, when the
pathwaybeginsto function,when the ratio of the rateof change
of concentrationto flux is 1:20.The profileof the changesin the
concentrationsof DMT (Fig. 2, inset)was very similarto that of
TRP (Table I) indicatingthat DMT was being furthermetabo-
lized and was only presentwhen being producedfrom TRP.
The behaviorof this systemwas analyzedby numericalmeth-
ods usingthe Eulermethod (6), which gives the exact calculated
concentrationsfor all the intermediates(Figs. 1 and 2), and, by
assumingthe concentrationof T and MT is small and steady
state,derivingapproximateanalyticequations,which show how
the parametersof the model affectits results.
The followinginformationand assumptionswere made about
reactantsthat interactwith outside pathwaysand hence are not
controlledby the alkaloidpathway:
1. TRP: TRP was determinedexperimentally(TableI).
2. PALP:There are no data for the in vivo concentrationof
PALP;k1,aAeLp is low (-2 ltM), and it was assumedthat PALP
was alwayssaturating.
 ALKALOID SYNTHESIS IN PHALARISAQUATICA 319

5
where, V3 = activityof SIMasesaturatedwith substratesand VI
= activityof TDase saturatedwith substrates.
Prime superscriptsrepresentthe reduced concentration of
ligands, i.e. the concentrationof the ligand divided by the Km,
4
Ki, or KdisSfor the enzyme.
The approximationused to derive Equation 1 is reasonable
0.6 betweendays and 5 and 12, while at day 4 this approximation
givesresultslow by a factorof 2 comparedto the exactnumerical
E method. For SAM and MT not saturating,Equations 1 and 2
1- _ ?0.3
show that SAM has no effect on the flux of the pathway,only
on the concentrationof intermediates.Thus, a decreasein the
concentrationof SAMleadsto an initialdecreasein methyltrans-
ferasevelocity followedby a compensatingincreasein the con-
1- /s8 16 centration of MT. Since,exceptat d 6, the concentrationof SAM
is unknown,this propertyof SIMasewas used to test two simple
hypothesesabout the time course of SAM concentration.SAM
concentrationwhen assumed to be constant led to calculated
4 8 12 16 MT concentrationswhich, relativeto concentrationsat d 6 were
too low, the disparityincreasingthe furtherthe age from d 6.
Day This indicatedthat the concentrationof SAM used in the simu-
lation is too high at days other than d 6. The concentrationof
FIG. 2. Variationwith time of DMT in P. aquatica seedlings.Values
SAM, when allowed to follow the time curve of TRP, gave
calculatedfrom the numericalsimulation;inset, resultspresentedfrom calculatedMT concentrationswhich stayed in step with their
one experiment. experimentallydeterminedconcentrations.This latterhypothesis
3. Binding of MT and DMT to TDase: KJ0aSfor MT and is reasonable,in that both TRP and SAM are being producedby
DMT are large comparedto the concentrationof these ligands the same metabolicmachineryduringmobilizationof the seed
and their bindingwas ignoredin the simulation. storagematerial,and it was used for the resultspresentedhere.
4. SAH:No informationwasavailableaboutthe concentration
in vivoof SAH. Since it is a stronginhibitorof the methyltrans- DISCUSSION
ferases and mechanisms for its removal in other systems are The concentrationsof the intermediatesT and MT are deter-
known (8), the concentrationof SAH was assumedto be zero. mined solely by the pathwayand are used as the measure of
5. DMT: DMT is turned over in seedlings of P. aquatica (1), worthof the model. From Equation 1 the errorsin determining
and the data presentedhere show that the alkaloidsynthesizing the concentrationof MT are mainly due to three kinetic con-
pathwayproducesmore DMT than is found in vivo.The model stantswhichhave errorsof a factorof 2 (errorsin concentrations
continuallyremovedDMT so that the concentrationof DMT in or activitiesare only about 10%).In the worst case, where all
the model neverexceededthat in vivo.Thus, the enzymesin the three kinetic constants contribute equally and independently,
simulation are presentedwith experimentallydeterminedcon- the final error is a factor of 3. Estimates of errors for the
centrationsof TRP and DMT. concentrationof T are difficultto determineas, unlike SIMase,
6. SAM:The analyticexpressionfor the concentrationof the PIMaseis not mostly bound to DMT. In this case the expected
intermediateMT can be simplifiedfor SIMasewhich is mostly errors for MT are used as an estimate for the error of the
bound to DMT giving concentrationof T. The data of Figure 1 show that the ratio
M Flux. DMT' (1) calculated/experimentalfor the concentrationsof T and MT is
MT' = ) less than a factor of 3, showingthe pathwayas modelled to be
V3 SAM' an acceptablefit to the data.
in which TDaseactivityis the ratelimitingstepin the pathway,resulting
in bound concentrationsof the intermediatesT and MT. Rerun-
Flux = activity of TDase = V,TRP'/( 1+TRP') (2) ning the simulationby increasingonly the activityof TDase gave

SAM SAM

TRP A TDase PT PMase SIMase DMT

TRP
TTP SM A
RKmP= 200jM = 4OM
KKSAM K,AM= 6Q1M
PALP saturating m = 2OpM KmT = 4C,jM
= 450pM
KDMT KDMT= 6OpM

SAH = OPM
SAH = OpM
1. Kineticconstantsused in the simulationof the synthesisof DMT in P. aquaticaseedlings.
SCHEME
320 MACK ET AL.
bound concentrationsfor the intermediatesuntil TDase reached
0.1 to 0.5 times the activityof SIMase,when the concentrations
of T and MT increasedwithoutlimit. The highestTDase/SIMase
ratio found experimentallyis 0.05 on d 5 indicatingthat the
pathwayis alwaysmore than a factorof two inside the require-
ments for a bound solution for the intermediates.Within the
time of the simulation,TDase activityvariesover a factorof 350
and TDase/SIMasevaries over a factor of 65. These numbers
are largecomparedto the marginof 2 requiredfor maintenance
of TDase as the rate limiting step in the pathway, and it is
reasonableto suppose that the TDase is rate limiting through
selective pressure, rather than coincidence. TRP is close to
saturatingat the time when the amount of TDase is highest
(aroundday 6) indicatingthata changein the pathwayto produce
unboundconcentrationsof intermediateswould only come from
increasingthe amount of TDase, rather than increasingthe
concentrationof TRP.
The affinity of DMT for the enzymes is in the order
SIMase>PIMase>TDase,decreasingby a factorof about 10 for
each enzyme. The very weak binding of DMT to the PALP-
TDase complex is expected from the chemistryof the TDase
reaction,where an imine cannot be formed.The weak (relative
to SIMase)binding of DMT to PIMase,when consideredwith
the similarchemistryof the two enzymes,indicatesthat PIMase
is designed to preclude binding of DMT. DMT only binds
strongly to SIMase because it is the enzyme product. Larger
amounts of SIMase, relative to the other two enzymes in the
pathway,arethus requiredto overcomeinhibitionby DMT. The
enzymesof the DMT-synthesizingpathwaythus arenot designed
to use DMT as a feedback-regulating element for the pathway,
and the ratio of enzyme activities of the enzymes (in reverse
orderof their affinitiesfor DMT) reflectsa requirementfor low
concentrationsof the indolethylamineintermediatesin the pres-
ence of DMT.
The two criteriafor the acceptanceof simulation,the ratio of
the calculated to experimentallyfound concentration of the
intermediatesand that the concentration of intermediatesis
PlantPhysiol.Vol. 88, 1988
bound,both showthat the model is an acceptablerepresentation
of the actualDMT synthesizingsystem.We concludethatTDase,
PIMase,and SIMaseare the enzymes responsiblefor the in vivo
synthesisof DMT.
Acknowledgment-We thank Mr. L. Higginbottom for the tryptophan estima-
tions.
LITERATURE CITED
1. BAXTER C, M SLAYTOR 1972 Biosynthesis and turnover of N,N-dimethyltryp-
tamine and 5-methoxy-N,N-dimethyltryptamine in Phalaris tuberosa. Phy-
tochemistry 11: 2767-2773
2. BAXTER C, M SLAYTOR 1972 Partial purification and some properties of
tryptophan decarboxylase from Phalaris tuberosa. Phytochemistry 1 1: 2763-
2766
3. COWARD JK, EP SLIsz, FY Wu 1973 Kinetics studies on catechol 0-methyl-
transferase. Product inhibition and the nature of the catechol binding site.
Biochemistry 12: 2291-2297
4. CULVENOR CCJ, R DAL BON, LW SMITH1964 The occurrenice of indolealky-
lamine alkaloids in Phalaris tuberosa L. and P. arundinacea L. Aust J Chem
17: 1301-1304
5. GLAZERRI, AL PEALE1978 Measurement of S-adenosyl-L-methionine levels
by SP Sephadex chromatography. Anal Biochem 91: 516-520
6. GROVEWE 1986 Brief Numerical Methods: An Introduction to Elementary
Numerical Methods Emphasizing Iterative Solutions. Prentice Hall, New
York, p 98
7. KAPLANA 1965 Urea nitrogen and urinary ammonia. In S Meites, ed, Standard
Methods of Clinical Chemistry, Vol 5. Academic Press, New York, pp 245-
256
8. KERRSJ 1972 Competing methyltransferase systems. J Biol chem 247: 4248-
4252
9. LEETEE, ML MINICH1977 Biosynthesis of gramine in Phalarisarundinacea.
Phytochemistry 16: 149-150
10. MACKJPG, M SLAYTOR 1979 Indolethylamine N-methyltransferases of Phal-
aris tuberosa, purification and properties. Phytochemistry 18: 1921-1925
11. MARUMP, AW HOvIN, GC MARTEN1979 Inheritance of three groups of
indole alkaloids in reed canary grass. Crop Sci 19: 539-544
12. MOORES, DH SPACKMAN, WH STEIN1958 Chromatography of amino acids
on sulfonated polystyrene resins. An improved system. Anal Chem 30: 1185-
1190
13. MULVENADP, M SLAYTOR1983 N-Methyltransferase activities in Phalaris
aquatica. Phytochemistry 22: 47-48
14. MULVENADP, M SLAYTOR1982 Separation of tryptophan derivatives in
Phalaris aquatica by thin-layer chromatography. J Chromatogr 245: 155-
157
15. SEGELIH 1975 Enzyme Kinetics. Wiley Interscience, New York, pp 274-346