Iodothyronine 5'-deiodinase in rat kidney
microsomes. Kinetic behavior at low substrate
concentrations.
A Goswami, I N Rosenberg
J Clin Invest.
1984;74(6):2097-2106. https://doi.org/10.1172/JCI111634.
The thiol-activated enzymatic outer-ring monodeiodination of iodothyronines by rat kidney
microsomes at low (nanomolar) substrate concentrations shows an apparently sequential
reaction mechanism and is further characterized by insensitivity to inhibition by dicoumarol,
a moderate sensitivity to inhibition by propylthiouracil (Ki = 100 microM) and iopanoic acid
(Ki = 0.9 mM), responsiveness to 5 mM glutathione (GSH), and a thermal activation profile
that is concave downward with a Td of approximately 20 degrees C. In contrast, the activity
at high (micromolar) substrate concentrations shows a ping-pong reaction mechanism, is
inhibited by micromolar concentrations of propylthiouracil, iopanoic acid and dicoumarol, is
unresponsive to 5 mM GSH, and shows a concave upward thermal activation profile.
Analysis of the microsomal deiodinase reaction over a wide range of 3,3',5'-triiodothyronine
(rT3) concentrations (0.1 nM to 10 microM) suggested the presence of two enzymatic
activities, with apparent Michaelis constants (Km) of 0.5 microM and 2.5 nM.
Lineweaver-Burk plots of reaction velocities at nanomolar substrate concentrations in presence of 100
microM propylthiouracil also revealed an operationally distinct enzymatic activity with Km's
of 2.5 and 0.63 nM and maximum velocities (Vmax's) of 16 and 0.58 pmol/mg protein per h
for rT3 and thyroxine (T4), respectively. These findings are consistent with the presence of a
low Km iodothyronine 5'-deiodinase in rat kidney microsomes distinct from the well
characterized high Km […]
Research ArticleFind the latest version:
lodothyronine
5'-Deiodinase
in
Rat
Kidney Microsomes
Kinetic
Behavior at LowSubstrate
Concentrations
AjitGoswami and Isadore N. Rosenberg
DepartmentofMedicine, Framingham UnionHospital,
Massachusetts 01701; Department ofMedicine, Boston University Schoolof Medicine, Boston, Massachusetts 02118
Abstract.
The thiol-activated enzymatic
outer-ring monodeiodination of
iodothyronines by
ratkidney
microsomes
atlow
(nanomolar)
substrate
concentrationsshows
anapparently sequential
reaction
mechanism andis further characterized by
insensitivity
toinhibition
by
dicoumarol,
amoderate
sensitivity
to inhibitionby
propylthiouracil
(Ki
=100
MAM)
and
iopanoic
acid
(K1
=
0.9
mM),
responsiveness
to5
mMglutathione (GSH),
and a thermal
activation profile
that is
concavedown-ward
with
a Td
of
-20'C.
In contrast,
the
activity
at
high (micromolar)
substrate
concentrations
shows a
ping-pong
reaction
mechanism,
is inhibited
by
micro-molar concentrations of
propylthiouracil, iopanoic
acid
and
dicoumarol, is unresponsive
to5 mM
GSH,
and
shows
aconcave
upward
thermal activation
profile.
Analysis of
the
microsomal
deiodinase reaction
over awide
range
of
3,3',5'-triiodothyronine (rT3)
concentrations
(0.1
nM
to10
uM)
suggested
the presence of
twoenzymatic
activities, with
apparent
Michaelis
constants(Kin)
of
0.5 ,M and 2.5 nM.
Lineweaver-Burk plots of
reaction velocities
at
nanomolar substrate
concentrations
in presence
of 100
,M
propylthiouracil
also revealed an
operationally
distinct
enzymatic activity
with Km's of
2.5 and 0.63 nM and
maximum velocities (Vmax'S) of
16 and
0.58 pmol/mg
protein
per h
for
rT3
and
thyroxine
(T4),
respectively.
These
findings
are
consistent
with the
presence
of
a
low
Km
iodothyronine 5'-deiodinase
in rat
kidney microsomes distinct from
the well
characterized
high
Km
enzyme and
suggest that at
circulating levels of
free
T4 the postulated low
K.
enzyme could be
physio-logically important.
Receivedfor publication 5 March 1984 and in revised form 26 July 1984.
Introduction
Whereas the
thyroid
gland is the sole source ofcirculating
thyroxine
(T4),'
3,5,3'-triiodothyronine (T3),
thebiologically
active form of the
thyroid hormone,
islargely
generated by
5'-monodeiodination
of T4extrathyroidally.
Theenzyme system responsible for this conversion has been identified in the microsomal fractionsof the rat liverandkidney
andpartially
purified
(1, 2). The microsomal enzyme has an apparent Michaelis constant(Ki)
of -5 MM for T4, is activatedby
thiols, especially
dithiols,
in a two-step transfer(ping-pong)
mechanism, is inhibited by
propylthiouracil
and can usereverseT3
(3,3',5'-triiodothyronine,
rT3)as analternatesubstrate(Km
- 0.5AM)
(3).AlowKm
(10-9
Mfor T4) iodothyronine 5'-deiodinase has recently been described in rat brain (4), pituitary (5), and brownadiposetissue (6). Theenzyme(designatedtypeII) also usesrT3 as an alternate substrate with high affinity(K.
- 3 X10-9
M), butisdifferent from thehighK.
(type I)hepatic and renal enzyme in being insensitive to propylthiouracil (PTU)andinits thiol activation mechanism, which is believed to be sequential rather than ping-pong (5). This type IIenzymatic
pathwaymay beparticularly
importantinregulation of intracellular, as opposed to circulating triiodothyronine levels, but may also play a role in the generation of that fraction of extrathyroidally produced T3 that is not blocked byPTUin athyreoticT4-replaced rats (7). Based on inhibition analysis with PTU, Silva and co-workers (7) also concluded thatthe type IIpathway,while ofmajor importance in brain and pituitary, wasof little, ifany, quantitative significance in the ratliver.Inthe present study, we have
examined
thekinetic behavior ofrat kidney microsomal iodothyronine5'-deiodinase
at low (nanomolar) concentrations of the substrates, T4 and rT3. The experiments revealed some properties of the enzyme at low substrate concentrations that differed significantly from those at high substrate concentrations and, indeed, resembled in1. Abbreviations used in this paper: DHL, reduced lipoamide; DTT, dithiothreitol; GSH, glutathione; PTU, propylthiouracil; rT3, 3,3',5'-triiodothyronine; T3,3,5,3'-triiodothyronine; T4, thyroxine.
2097 RatKidney Iodothyronine 5'.Deiodinase
J.Clin. Invest.
© The American Society for Clinical Investigation, Inc. 0021-9738/84/12/2097/10 $1.00
some,butnotall, respectsthe characteristics described forthe
type II enzyme in the cerebral cortex, pituitary, and brown adipose tissue. We report here some observations on kidney microsomal 5'-deiodination of T4 and rT3 at low substrate
concentrations and raise the possibility, on the basis of kinetic and other data, that rat kidney microsomes may contain both alowand a high Km iodothyronine 5'-deiodinase.
Methods
Reagents. L-[3'- or
5'-'25-I]
rT3 and T4 (750-1,250gCi/Mg)
were obtained from New England Nuclear (Boston, MA) and were further purified just before use by high voltage paper electrophoresis (3). Dihydrolipoamide (DHL) was prepared by borohydride reduction of theoxidized form (obtained from Sigma Chemical Co., St. Louis, MO) by the method of Reed et al. (8) as modified by Silver (9). lopanoic acid (Telepaque) was a product of Winthrop Laboratories Co., New York; a stock solution was made by alkaline extraction of the tablets andneutralized before use.Assay ofmicrosomaliodothyronine S'-deiodinase. Rat renal micro-somes were prepared by methods as described earlier (10). The incubation medium contained, in a total volume of 200
;l:
microsomal protein, 50 Mg; K-phosphate buffer (pH 7.0), 20 Mmol; EDTA, 200 nmol; and varying amounts of dithiothreitol (DTT), rT3, or T4 (containing -50,000 cpm of '25I-labeled rT3 or T4), as indicated in legends. The incubation times were 3 min for rT3 (except in the experiment of Fig. 1) and 60 min for T4. The incubations were terminated by the addition of 0.1 ml of 2% bovine serum albumin (BSA), followed immediately by 0.5 ml of 10% ice-cold trichloroacetic acid. '25I- liberated was measured after separation by ion-exchange chromatography on Bio-Rad Ag5OWX8(H+) resin (BioRad Labora-tories,Richmond,CA)asdescribed previously (10). The reaction rates wereexpressed as picomolesofrT3 degraded per milligram protein per hour. The products were frequently checked by descending paper chromatography (solvent system-hexane/t-amyl alcohol/ammonia (2 N) 1:10:11).Results
Studies
of
microsomal 5'-deiodination over a wide rangeof
substrate concentrations. Fig. 1 showsrT3deiodinaseactivities ofkidney microsomes, at5mMDTT,from agroupof normal
rats and from rats given intraperitoneal PTU (2
mg/l00
g body wt) 15 min before killing. A plot of enzyme reactionvelocities vs.-log
(S]
for normal microsomeswassigmoid (Fig. 1 A), withan inflection at M0.5gM
andamaximumvelocity of -58 nmol/mg protein per h, thus reflecting thewell-documented characteristics of the kidney 5'-deiodinating
en-zyme. In a plot ofthe activities at substrate levels <10 nM
and
extending
down to 0.1 nMon a 25-fold expanded scale, however, adeviationwasnoted between theobserved reactionvelocities andthose
predicted
bythe Michaelisequation, using valuesforKmandV.,.
of 0.5 ,uMand 58 nmolrT3 degraded/ mg protein perh,respectively,
asobtained from experiments athigher substrateconcentrations for this microsomal prepa-rationjust before thisexperiment.
Asshown inFig.
1 A, theobserved reaction velocities over the range of substrate
con-p[S]
Figure 1.Semilogarithmic
110 B plots of reaction velocities
100 vs. rT3concentrations
ob-100 < tainedwith renal
micro-80 somesfromnormal and
> \ PTU-treated (2 mg/100 mg
60
60 bodywt, i.p., 15
min
be-40- fore killing)rats.The
incu-bation medium contained,
20- in a total volume of 200 Ml: ° 77 8 9 10_L_ microsomal
protein,
50Mg;
[
9J
K-phosphate
buffer(pH
7.1), 20Mmol;
EDTA, 200 nmol; DTT, Igmol
andvaryingamountsofrT3(containing 50,000cpmof'251-rT3),asindicated. The incubation wasfor I min andwas
terminated bytheaddition of 0.1- ml of 2%BSA,followed immedi-atelyby0.5 mlof 10% TCA. I-liberatedwasmeasuredbyion
exchange,
chromatographyonBio-RadAG50WX8resinasdescribed (10). (A)Activities in normal(-)and PTU-treated(o)microsomesasafunctionoflog
l/[rT3].
The lower part ofAshows the activitiesatnanomolar substrateconcentrationson a25-foldexpandedactivity
scale. Also shown(dashedline)isatheoreticalplotof theactivities,
atlow(nanomolar)substrateconcentrations, extrapolatedfrom the kinetic parameters observedathigh(micromolar)substrate
concen-trations. The inset shows doublereciprocal plotsof theactivity profileat0.75-20-nM substrate range.(B)Difference between
reac-tionratesobservedatlow
[S]
and thoseprojected
from the kinetic parametersathigh 1S],plottedas afunction oflogl/[rT3].
Thereactionrateswerecalculatedaspicomoles rT3 degradedper milli-gramproteinper hour. The dataarefromone
experiment,
done induplicate,andarerepresentativeof three similar
experiments.
centrations from 0.25 to 10 nM (solid line)were consistently greaterthan the extrapolated values (dotted line) and showed a sigmoidal profile.2 The Lineweaver-Burk plots also showed a discontinuity at the 5-10-nM substrate range, as shown in the inset. A difference curve, constructed by plotting the difference between observed andpredictedvalues of enzymatic activity at the low substrate concentrations, also showed a sigmoidal profile (Fig. 1 B) suggesting the presence ofa low Km enzymatic activitywith aKm of -2.5 nM and a
V.
of-1 10 pmol rT3 degraded/mg proteinper h.
Thisanomalousbehavior of the enzymecannotbeexplained on the basis of increased fractional binding of substrate to
microsomal proteins at low substrate
concentrations,
as the reaction velocity wouldbeexpected todecrease,
notincrease, because ofdiminished availability of substrate tothe enzyme under such conditions. Moreover, under the conditions used in these experiments, binding ofiodothyronine
substrate tomicrosomal proteinswasfoundtobe
relatively slight
(5-10%) and valuesfor percentagebinding
werewithin the same range atboth micromolar and nanomolarsubstrateconcentrations.3 The activity profile of the microsomes from the PTU-treated rats showed the expected sharp decline inactivity
atsubstrateconcentrations above 100 nM,
confirming
the findings of Leonard and Rosenberg (3), with relatively less inhibition (35-40%)atlower(0.5-2.5 nM) substrate concentrations. The plotofthePTU-inhibited activity wasdoublysigmoid(Fig. 1 A) and, while the activity at thehigher
range of substrate concentrations had aKm/
V.
ratio similar to that obtained with normal microsomes at high substrate concentrations, at the lowsubstrateconcentration avalueforKm (-2.5 nM)was observedthatwassimilartothatderivedfromtheuninhibited profileatthisrange, butwithalower V,, .Whiletheinhibition pattern at high substrate concentrations is as expected from the well knownuncompetitive
characterofPTUinhibition (3, 11, 12)of therenal andhepatic enzyme, theinhibition profile at low substrateconcentrations,
characterized by lesser PTU-sensitivityand an unalteredK.,
indicates eithersome unusual2Reproducibility
of data. For the points in Fig. 1 A at which thesubstrate concentrations were <1
IMM,
the observed values for rT3degradationranged between 11.7 and 34% of the total and the absolute differences between duplicate determinations were usually <1 and in
nocase >2.8; the deviations of the duplicates from the means averaged 4%. Forthe 1-, 2.5-, and 5-MM substrate concentrations, the mean percentagesof rT3 degraded were 8.1, 4.2, and 2.2%, and the duplicates differed by 1.0, 0.4, and 0.6, respectively.
3Thebindingstudies were performed by incubation of renal microsomes (50Mgprotein) with 0.5 nM and 5AM labeled rT3 or T4 at 41C for 15 min in 200 Ml of0.1 M potassium phosphate buffer (pH 7.0) followed by ultrafiltration through Millipore HAMK filters (Millipore Corp., Bedford, MA), with analysis of the radioactivity in both the filterand the filtrate. Nonspecific binding was determined after incu-bation in presence of a 100-foldexcess of nonlabeled rT3 or T4 and
wasbetween 2 and 3% of total radioactivity. Under these conditions
the specific binding for rT3 and T4 was 8 and 6 fmol at the low and
17 and 11 pmol at the high substrate concentrations, respectively.
kineticsofa PTU-modified enzyme atlow substrate
concen-trationsorthe presence ofa separate lowKm enzyme with a
different mechanism of inhibitionbyPTU.
ThevaluesforKmand
V,,,,
asdetermined from the linear portionsof theLineweaver-Burk plots at low substrateconcen-trations in presence of 100 AM PTU were 2.5 nM and 16 pmol/mg proteinper h for rT3 and 0.63 nM and 0.58 pmol/ mg protein per h forT4, respectively, at 2.5 mM DTT. The DTTconcentration of 2.5 mM was a compromise to achieve
a level of enzyme activity that was detectable by the assay procedure used, yet was not high enough to overcome the inhibitory effects of PTU, especially at high substrate concen-trations(10-12). Theproducts ofthe enzymatic reactionwere
identified by paper chromatography as 3,3'-diiodothyronine and I-, and T3 and I-, in equimolar proportions, from rT3 and T4, respectively. The reaction rates were proportional to the amounts of microsomal protein and the enzymaticactivity wascompletelyeliminated by heat treatment
(80'C
for 1 min) orby pronase digestion of the microsomes before incubation. Athigh substrate concentrations(range, 0.1 to 10 ,uM), in the absence of PTU, the kinetic parameters for T4 and rT3, as determined by Lineweaver-Burkplots,were:Km(rT3), 0.5 AM; Km(T4), 5AM;
Vmax(rT3), 60 nmol; andV.
(T4), 2.1 nmol/ mg protein per h, respectively, in reasonable agreement with published values for kidney microsomes (3); these activities werecompletely inhibited at 100AM
PTU.The doubly sigmoid substrate concentration curves, ob-served in both normal and PTU-treated microsomes (Fig. 1, AandB), indicate the presence inrenal microsomesof either (a) a single enzyme withnegativelycooperative ligandbinding, or (b) two enzymes with different Km values acting on the samesubstrate (13). This question was then addressed experi-mentallybyinhibition and activation analyses of the enzymatic activities at micromolar and nanomolar substrate
concentra-tions.
Inhibition analysis. PTU inhibited the renal microsomal enzyme, when assayed at 0.5 nM rT3, by 30% at 100
AsM;
this was considerably greater than one would expect from the inhibition constant(Ki)
(1 MM) reported, and experimentally confirmed, for the uncompetitive inhibition ofthe high Km enzyme. As shown in Fig. 2, the inhibition was characterized bypartial inhibition kineticswith -90% inhibitionofthe total activity at 1 mM PTU, and very little further inhibition at higher concentrations of the inhibitor. The data suggest satu-ration ofa separateinhibitor-bindingsite.The residual activity in presence of excess of inhibitor was near the limits of experimental detectionandwas not furtheranalyzedkinetically. In contrast to PTU, iopanoate and dicoumarol inhibited -65%of theactivityat 10MM.4The residual activity, however,was more resistant to these inhibitors, requiring -2.5 mM 4Thisdegree ofinhibition is close to what one would expect from 10 MMiopanoate (K, =9.2MM)if itwereinhibiting onlythat component of the enzymatic activity attributable to the high Km enzyme as determined by extrapolation(e.g.,as inFig. I A).
60-
(50-C
5,
40-030
20-
10-0*
0.2 0.4 0.6 0.8 1.01.5 2.5 3.0
[11mM
Figure 2. Effects ofvarious concentrations of PTU(.),iopanoate (o), anddicoumarol (a)onmicrosomal 5'-deiodinationof rT3. The
ab-scissareferstomillimolar concentrations of the inhibitor. The micro-somes wereincubatedfor3 min in absence or in presence of indi-catedconcentrationsofspecified inhibitors under conditions as de-scribed inFig. l. rT3 was 0.5 nM and DTT at 5 mM. The values are from one of three closely agreeing experiments.
iopanoateforcompleteinhibition, while dicoumarol was much less active (40% inhibition at 2.5 mM). The results indicate thepresenceofaseparateenzymewith different sensitivity to these inhibitors.
Lineweaver-Burkplots at nonsaturating PTU concentrations (Fig. 3 A) showed noncompetitive, in contradistinction to uncompetitive, kineticsover the rangeoflower rT3 concentra-tions (0.25-2.5 nM). A replot ofthe vertical axis intercepts against the inhibitor concentrationswaslinear (Fig. 3 A,inset) andrevealeda
Ki
of- I00MM, comparedwithaK, of 1;M
reported for the uncompetitive inhibition by PTU at high substrate concentrations (14). The inhibition by PTU was competitive with DTT (Fig. 3 B),withconsiderable releaseof theinhibition
at high DTT concentrations. For example, at 550 MM PTU,although
the inhibitionwas -87% at2.5 mM DTT,itwasonly -56% at 20 mM DTT. Arelativeinsensitivity was also observed withiopanoic
acid, a potent inhibitor of microsomal deiodination ofiodothyronines
at high substrate concentrations (15).The inhibitionby iopanoatewas compet-itive withrT3(Fig.
4),asisthecasewiththehigh
Km enzyme, but the observedKi (0.9 mM)
was over 100-fold greater than thatreported for thehigh
Km enzyme(16).
Similarinhibitory
patterns [K;
(PTU),
100MM; Ki (iopanoate),
1.38mM]
werealso obtained withthe 5'-deiodination of
T4,
whenassayed
atnanomolar substrate concentrations, and the
similarity
oftheK,'s
suggestedthat theouterring
deiodination of both substratesN
.5 06
01
E
N,
01
a
*0
-0.06
-600
rT3(nM)-'
.C
80.4
E /0.5
0.3/
id 21-0 2
E 02
0.1 0.2 0.3 0.4.
DTTimMr'
Figure 3. Doublereciprocalplots of PTU inhibitionof
rT3-deiodi-naseactivitywith(A) rT3asa variable substratewithDTTat5mM,
and(B)DTT as avariable substrate withrT3at0.5nM.Incubation was for 3 min.ThePTU concentrationschosenwerenonsaturating
(compareFig. 2), and,inadditiontothe indicated
concentrations,
allincubationswereperformedat abasal PTU level of 50 MM.An
intercept replot(A,-inset)waslineargivingaKiof 100MMfor the inhibition. The dataarefromasinglerepresentativeof three
closely
agreeingexperiments.
at these concentrations was mediated
by
the same enzyme. The renal enzyme,assayed
at low substrateconcentrations,
thus differs from the type
II,
low Kmpituitary
and brainlopanate
(mM) op1.0
c ._
0
E
0
0
0
l-0t
0 2 4
K;=0.9mM rT3(nM)-'
6 8 1U
Figure4.Competitiveinhibition ofrT35'-deiodinaseby iopanoic acidatnanomolar substrateconcentrations. Incubationwasat50 uM
PTU and 5 mM DTT. Incubation timewas3min.AKi of 0.9 mM
wascalculated for the inhibition. The dataarefromoneof three
similarexperimentswhere thevaluesagreedwithin 10% of each
other.
enzymes in being moderately, ifincompletely, PTU-sensitive
andthismayaccountfor thefailure of Silva et al.(7)to detect alowKmenzymebyinhibitionanalysisalone.
Progress curves. Fig. 5 shows the progress curves of the
enzymatic activity at 0.5 nM rT3 in the presence of 50 and
150
gM
PTU. The profiles resembled that reported for thehighKmenzyme(3),composedofaninitialrapid phase during the first minute, followed by a slower phase of enzymatic activity. The activity at 150 MM PTU was -40% inhibited
from the levels obtained at 50 MM PTU. This inhibition, however, was observed in both the early and later phases of
theprogress curve,in contrast tothehighKmprogress curve,
where the rapid early phase, presumably representingthe first
catalytic cycle, has beenreportedtobe insensitivetoPTU(3). These findings indicate that the 5'-deiodination of rT3 at
nanomolarconcentrations is steady andsustained during the entire incubationperiodandthat the relativePTU-insensitivity
cannotbeattributedtotheenzymeactivity duringthe brief(1 min) period of PTU-insensitive catalytic activity.
Reaction mechanisms. Lineweaver-Burk plots of reaction velocities vs. rT3 concentrations in the nanomolar range at
fixed DTT concentrations and in presence of 50 MM PTU
generatedaseriesoflinesthatintersected atthe velocityaxis (Fig. 6 A), in sharp contrast to the parallel lines observed at
micromolar substrate concentrations (inset). A similarplot of
reaction velocitiesvs.DTTconcentrationsatfixedrT3 concen-trations yielded a series of lines intersecting in the second
Time(min)
FigureS. Progresscurvesof microsomalrT35'-monodeiodinasein 50
MM (o)and 150 MMPTU(o)at0.5 nMrT3.The microsomal preparations and theincubation conditionswere asdescribed inFig.
1.Thedataarefromoneoftwoexperimentsincloseagreement.
quadrant above the I/Saxis (Fig. 6 B) and a secondary plot (inset) of the slopes of these lines against l/rT3 yielded a
straight line passing through the origin (intercept zero). The
data suggestanorderedsequentialmechanismfor the deiodinase
reactionand wouldseemtoindicate the formation ofaternary
complex in which rT3 is the second substrate (17, 18), in contrast tothe well documentedping-pongmechanismathigh
substrate concentrations.
Thermal activationprofiles. The effectsoftemperature on
the
V.
of the enzymatic reaction, when measured at low(nanomolar)substrateconcentrationsinpresenceofPTU, also
differedsignificantlyfromthose observedathigher(micromolar)
substrateconcentrations in absence of PTU. As shown in Fig.
7, while the Arrhenius plots of both activities showed
discon-tinuities (Td)at '20'C, the display was concavedownwards whenmeasuredatalow substrate concentration (20nM)with
activationenergies of 5.98 and 11.8 kcal/mol above and below
the Td,respectively. This wasin sharpcontrast to theprofile
obtained with 5 ,uM rT3, which was concave upwards with
activation energies of 2.14 and 7.5 kcal/mol, respectively,
above and belowthe Td.
Activation energiesareregardedasintrinsic characteristics
of enzymes rather than substrates (19), and the different
thermal activation profilesobservedathigh and lowsubstrate
concentrations can be taken as indirect -evidence for two separate enzymes operating at these two substrate levels. Itshouldbenoted, however, thatthereaction velocitiesatthe
lowsubstrate levelweremeasuredinpresenceofPTU andthe
possibility that the activation profile seen at this substrate
concentration was that ofa PTU-modified form ofa single
2101 RatKidneyIodothyronine S'-Deiodinase
C
-W 0
E
IN 0 0 0
rT3,20nM
1.5
1.4
1.3
1.2
1.1 1.0
0.9
0.8
3.2 3.3 3.4 3.5 3.6
1OOO/T
3.2 3.3 3.4 3.5 3.6
1OO0/T
Figure 7. Temperature activation profiles ofmicrosomalrT3
5'-mon-odeiodinationatmicromolar and nanomolarsubstrate
concentra-tions. Microsomes (50 jigprotein)wereincubated with5jiM(A)or
20 nM (B) rT3 inpresenceof5 mM DTTattheindicated
tempera-turesfor3 min, and theratesof deiodinationweremeasured by
methodsdescribed in Fig. 1.In B the incubation mediumalso
contained 50 juM PTU. The reaction velocitieswerecalculatedas
picomoles rT3 degradedpermilligram proteinperhour. Thedataare
fromarepresentative of three closely agreeing experiments.
rT3 (nM)-'
E)TT (mM)-l
Figure6.Renalmicrosomal rT35'-deiodinaseactivitiesatnanomolar
substrate concentrationswith(A)DTT and(B) rT3asfixed
sub-strates.Incubationtimewas3 min.InA,activity patternsobserved
atmicromolarrT3concentrationshavebeenincluded for
compari-son.In B,aslope replot (inset)waslinearpassing throughtheorigin. All incubationswereperformedin50AMPTU.The data arefroma
representativeof threecloselyagreeing experiments.The values
agreedtowithin10%of eachother.
enzyme cannot be entirely excluded. The linearity of the
Arrheniusplotwithin thetemperaturerangesunder observation
would, however, suggest thattheoverallvelocityiscontrolled
byasingleratelimitingstepthroughoutthetemperaturerange (20) and that its thermal activation coefficient isunaffectedby interaction withPTU.
Relative
substrate preference athigh and low[S].
Table I shows the5'-deiodinase
activities of rat kidney microsomes with0.5 nM rT3 and T4 in presence of DTT and DHL. In the presence of5 mM DTT and DHL, while the rT3 deiodinaseactivities
were only slightly abovethe values calculated from thekinetic parameters derived from experimental values ob-served with high substrate concentrations, theT4
deiodinase activitieswere about fivefold higher. The preferred substrate for thedeiodinating
reaction at low substrate concentrations thus appears tobe T4ratherthanrT3. This isalso borne outTableI.
S'-Monodeiodination
ofT4andrT3(0.5 nM) by KidneyMicrosomes:Effects of
Various ThiolsY-Monodeodinaseactivity
(pmolsubstratedegraded/ngproteinperh)
Thiol T4 rT3
DTT 1.1±0.15 (0.21) 93.4±7.5(60)
DHL 1.4±0.18 (0.21) 104.8±9.7(60)
Eachfigurerepresentsmean±SE ofdata from sixexperimentsin whichrT3was substrate and four in whichT4wassubstrate,each done induplicate.Assay conditionswereasdescribedin thetext.
Thefiguresinparenthesesrepresenttheoreticalvalues calculated
fromthefollowing kineticconstants(seetext):Km(rT3),0.5
AM;
Km(T4),
5AM; V (rT3),60; andV.,,,
(T4), 2.1nmol/mg
proteinper h.2102 A. GoswamiandI.N. Rosenberg.
5.0
4.8 x
* 4.6
E
>4.4
0
4.2
4.0
3.8
A B
by the lower Km for T4 compared with rT3 observed at low substrate concentrations (see above) and is a reversal ofthe relativesubstrate preference seen athigh substrate concentra-tions.
Thiol activation profiles. The thiol activation of enzymatic activity at nanomolar substrate concentrations was further investigatedover awiderange ofconcentrations ofglutathione (GSH), DTT, and DHL, with results as shown inFig. 8. The previously reported 10-fold greater potency of DHL than of DTT (14) instimulating deiodinating activityathighsubstrate concentrationswasalsofoundatnanomolarconcentrationsof substrate. In addition, while the enzyme, in agreement with published data (14), was almost totally unresponsive to 1-5 mM GSH at high substrate levels, it responded to these concentrations ofGSH when rT3 was 0.5 nM, at -20-25% of the level with DTT.Although thenatureof this differential GSH-responsiveness athigh andlow substrate concentrations is unclear, a possible explanation may be that rat kidney microsomes may contain an operationally distinct GSH-sen-sitive lowKm enzyme.
A summary of the kinetic parameters ofthe enzymatic activityathigh andlow substrate concentrations is presented inTableII.
Discussion
When assayed at low (nanomolar) substrate concentrations, rat kidney microsomal iodothyronine 5'-deiodinase displays somekinetic characteristicsthatdiffer significantlyfrom those observedathigh (micromolar) substrate concentrations. Some of these
differences,
notablytherelativePTUinsensitivityand asequential reaction mechanism observedat low(nanomolar) substrate concentrations asopposedtothehighPTU-sensitivity
and thetwo-steptransfer (ping-pong) mechanism, whichhave been observed athigh (micromolar) substratelevels, are rem-iniscent ofthe low and high
Km
enzymes described in theiff rT3, 0.5 nM A rT3, 0.5uM
2 100 DHL
E
E DiT
so
60
C
.2
40-S~ ~ ~ ~ ~~~~~S
20
0
5 4 3 2 17 6 5 4
p[-SHJ p[-SH]
cerebralcortex
(4), pituitary (5),
and brown fat(6),
and would suggest the presence also of a distinct lowKm
enzyme inkidney
microsomes. Thereare,however,
severalconsiderations,
as discussedbelow,
that must becritically
evaluated before suchaninterpretation
canbe made.Inhibition kinetics.
(a)
PTU inhibition of the enzyme at micromolarsubstrate concentrations ischaracterizedby parallel
upward displacements
of doublereciprocal plots (3).
From the kinetic expression of such uncompetitive inhibitions (21,22):
v =Vma,/[(Km/S)
+(1
+i/Ki)],
where i andKi
refer to the inhibitorconcentration and theinhibitorconstant,respectively,
it is clearthat inhibition at any fixed concentration of PTU would become progressively lowerassubstrateconcentrations
arelowered becausethe
i/Ki
termin thedenominator becomes small in relationtoKnrS.
Forexample,
ifK, (PTU)
= 1,M,
Km
(rT3)
=0.5,uM,
andVm,>(rT3)
= 60nmol/mg protein
perh,
although
theactivityat0.5AM
rT3will be -85% inhibitedat 10
,tM PTU,
the activity at 0.5 nM rT3 will be only 1% inhibited(and only
50%inhibitedat 1 mMPTU).
Conclusionsregarding
enzymeidentificationbasedsolely
onPTUinhibitionaretherefore dangerous since even a single enzyme's
suscep-tibility
to inhibitionvaries withexperimental conditions(e.g.,
substrate
concentration). (b)
Inmostpublished studies,partic-ularly
thosedealing
with the brain and pituitary deiodinases (4,5), high
DTT concentrations (e.g., 20-100 mM)have been used whichmight
beexpectedto cause asubstantiallessening
of PTU inhibition since PTU andDTT arecompetitive
(Fig.
3 B).(c)Furthermore, underconditionswherePTU is inhibi-tory, the inhibited enzyme would display much lower Km's and
V,,,,'s (depending
onPTUconcentrations) becauseoftheuncompetitive
nature of the inhibition with respect to theiodothyronine
substrate(uncompetitive
inhibition being char-acterizedby
aproportional decreaseofKm
andVm.x
withKm/
V,,.
remaining
constant). A critical analysis of the kinetic data is therefore imperative in ascertaining whether what isbeing
measuredis the residual activityof thehighKm
enzymeat low substrate concentrations in an environment of
high
B
- DHL
o DTT
Figure 8. Thiol activation of5'-monodeiodinationof
GSH rT3
by
renalmicrosomes.Theincubationconditions were as described in Fig. 1, except that the incubation , timewas3 min. Each point is the mean of three 3 2 1 determinations, each done in duplicate, from threedifferentmicrosomal preparations.
TableII.Kinetic Characteristics of Renal Microsomal Iodothyronine5'-Deiodinaseat Low (Nanomolar) and High (Micromolar) Substrate Concentrations
Low[S] High
[S]
0.1-10nM 0.1-10)M
Response to GSH ++ 0
(1-5mM)
Inhibition by
PTU Ki= 100MM KiJ=1 tM
(Noncompetitive) (Uncompetitive) lopanoate K,=0.9 mM K.=9.21M
(Competitive) (Competitive)
Dicoumarol Minimal* Ki=7.5,M
(Competitive)
Mechanism Sequential Ping-pong
(ordered)
Km
rT3 2.5nM* 0.5 IM
T4 0.63nMt 5 ,M
rT3 16pmolt 60 nmol
T4 0.58pmolt 2.1 nmol
Temperature Concave Concave upwards
activationcurve downwards
*Assayedat50AMPTUand 5 mM DTT; Assayed at100uM PTU and 2.5 mMDDT; §mgprotein'hV'.
PTUand DTT,orthe activity ofadistinctly differentlow
Km
enzyme.
Several distinctive features emerge from analysis ofthe kineticdataobtained with the kidneyenzyme atlow substrate
concentrations:
(a) Although therT3-deiodinating
activity at low[SJ
differed only modestly from that projected from the Km and V,, obtained athigh
substrate concentrations,T4-deiodinating
activityat low[SI
wasappreciably
higher than that extrapolated from the kinetic parameters derived from studiesathigh
[SI
and, incontrast tothe observationsathigh
[S],
the enzymeshowedamuch lowerKmforT4
than forrT3.
Suchareversal of substrate
preference
forT4
than forrT3 has also beenreported
with the type II enzyme of the brain andpituitary;
but unlike the brain andpituitary
type II enzyme, where the VI,,Km
ratio forT4 washigh
and where therT3
deiodination
washalf-maximallyinhibitedbyvery low(2nM)
concentrations
of T4(5),
thekidney
enzyme atlow substrateconcentrations
showedamuch smallerV.t
forT4
andrequired
muchhigher
concentrations
of T4(200 nM)
forhalf-maximal inhibition of rT3 deiodination(data
not shown). The reasonfor this limited
ability
ofT4
to suppressrT3
deiodination,
despitethe lower Km for
T4
than forrT3,
isnotentirelyclear. Itshould be noted, however,that inatwo-substratesequential
mechanism,
theK,
ofanalternate substrate(e.g., T4
inrespect torT3)isnot a truedissociationconstant, but isasteady-state
constant (23)thatcould assumea high value when the rate of degradation oftheternary complex formedby thecompetitive inhibitor is slow, asappearsto be the case with T4relativeto rT3 in the kidney microsomes. (b) At 5 mM DTT, 1 mM PTU surprisingly inhibited the enzyme at low
[S]
by >90% (Fig.2), a levelof inhibitionthat wasconsiderably higherthan that calculated from the kinetic parameters at high[S];
this suggests the presence of a separatePTU-sensitive,
lowKm
enzyme.The renal low
Km
enzymethusappearstodifferfrom the type II enzyme described in the cerebral cortex and pituitary (4, 5) in being moderately sensitiveto inhibition by PTU. (c) The PTUinhibition at low[S]
wasnoncompetitive with aK,
considerably larger(100 AM)than theKi
(1
AM) for the uncompetitive inhibition observed at high[5].
(d)Iopa-noate,acompetitive inhibitorof rT3 deiodination
(15),
was a much weaker inhibitorat low than at high[S].
If the highKm
enzyme werethesolerenal microsomal outer-ring deiodinase, one would have expected, given the nature of competitive inhibitory kinetics, that a fixedconcentration of the inhibitor would be more,rather thanlesseffective at low[S].
Thus the observation thattheKi
of iopanoate is-100
times greaterat nanomolarthan at micromolar[S]
isadditional evidence for the presence ofaseparatelowKm
deiodinase, lesssensitive to iopanoate because of its higher affinity for the substrate. Dicoumarol,apotent competitive inhibitorat high[S],
was a weak inhibitor at low [S]; the structural unrelatedness of dicoumarol andiodothyronine
raises the possibilitythat dicou-marol could be amodifier interacting withaseparate binding siteon thehighKm enzyme alone,perhapsassumingcompetitive
inhibitory kineticsas aK-system effector (24).
To eliminate the activity ofthe
conventional
highKm,
PTU-sensitive
renal microsomal5'-iodothyronine
deiodinase, we added PTU(50-100,uM)
in almost allexperimentsat low substrateconcentrations
(except for theexperiments
inFigs.
1, A and B, 2, and 8). Actually, the degree of inhibition by PTU ofa true high Km enzyme
operating
atsubnanomolar
substrate
concentrations
isnegligible,
for reasonsexplained
above. Nevertheless, the
possibility
that theproperties
ofthe low Km enzyme arereally those ofaPTU-modified
high Km
enzyme needstobe
considered.
Wethink itunlikely
thatthis is thecase because(a)aPTU-modifiedenzymewouldprobably
not be PTU
sensitive,
as the lowK.
enzyme appears tobe;
(b)
PTU-modification
of thehigh
Km
enzyme would be unlikely to alteriopanoate binding
withoutsimilarly altering
substrate
binding;
and,(c)
the difference betweenpredicted
andobserved reaction rates
(Fig.
1,
A andB)
atlowsubstrateconcentrations
were obtainedinthe absenceofPTU.
Reaction mechanisms. The
essential
feature oftheping-pong
mechanism
is that the enzyme reacts with one ofthe substrates toformaMichaeliscomplex
that isfollowed
by
the dissociation of the firstproduct
leaving
a modified enzyme form usuallycontaining
amoiety
ofthe first substrate. The second substrate interacts withthemodifiedenzymeonly
afterdissociation
ofthe firstproduct.
Basedonkineticdataobtainedat high substrate
concentrations,
a model has beenproposed
forthehepaticand renalhighKmenzymein whichanoxidized form of theenzyme(possiblyasulfenyl
iodide)
is first generated during substrate conversion,and reducediodothyronine product is released; DTT is involved only in the reconversion of the oxidized enzyme to the native form with release ofiodide:DT1
E +rT3 EErT3 - T2+
EO,*I
- E + I- +DTTO,
(1) (2) (3)
(4),T
Eio.
PTU +1-.The ping-pong reaction mechanism is
usually diagnosed
by findingparalleldouble reciprocal plotsof initialrate data and has been experimentally verified for thekidney
and liver enzyme in a number oflaboratories (10, 11),including
our own(3).It mustbeemphasized, however,
asCleland hasdone (25,26), that theparallel initial velocitypatterns in ping-pong mechanisms are observed only when the modified enzyme form generated after the release of the first product(in
this caseEO,
*I) iskinetically
stableduring
the timeperiod
ofthe reaction. At very low concentrations of the iodothyronine substrate and high concentrations of DTT, this may not be the case, since the rate offormation ofthe enzyme-substrate complex might conceivably be small as compared with its degradationandrapidequilibrium
conditionsmight
not prevail. Furthermore, ifE.,*
I is theintermediate
enzyme form and themechanism ofPTUinhibition is by formation ofamixed disulfide withE.,*
I [asenvisaged
by Leonard andRosenberg (27) and byVisseret al.(28)],then inpresence of excess PTU, the reaction will be diverted towards the formation of the PTU-complex (step4) and theapparentfirst-orderrateconstant forthereactionof enzyme andsubstratetogive iodide willbe nearly zero(ignoring
the minute amount of iodide generated duringthe first catalytic cycle). RaisingDTTlevels will divert the reaction flux from the dead-end formation of the PTU-complex (step 4) (or the spontaneous hydrolysis oftheE..
* I complex)tothiol-dependent
production of iodide(step 3) and thus increase the apparent rate of iodide generation (withoutaffecting
the true rateof combination betweenEandS).This would decrease the slopeof thereciprocalplot5
and intersecting rather than parallel lines would be observed on Lineweaver-Burk plots, even if the reaction mechanism isbasically ping-pong. This becomes particularly relevant at low substrate concentrations wherethePTU-inhibitionissmall and measur-ablereactionrates can beobtainedeven atfairly high concen-trations of PTU. Considerable caution must therefore be exercised in interpretingtheintersecting Lineweaver-Burk plots at very low iodothyronine concentrations (as observed in the5The
lack of slope effect in double reciprocal plots of ping-pong mechanisms is due to a constancy of the ratio of free enzyme [E] to the central complex (F), irrespective of the concentration of the secondsubstrate, because of a lack of reversible connection between E and F
(23).
present studiesand in those ofLarsenand co-workers
[4-6]),
asevidencefora
separate
(sequential)
kinetic mechanismand,therefore,
foraseparate
enzyme.Whether the reaction mechanisms observed at
high
and lowsubstrate concentrationsaredifferentor not,other mani-festationsof theenzymatic activity
atlow substrateconcentra-tions, notably
the differential response toGSH,
differential response to bothuncompetitive
andcompetitive
inhibitors anddissimilar Arrheniusdisplays
are,indeed, suggestive
of the presence ofaseparatelowKm
enzyme inratkidney
microsomes. Such a lowKm
enzyme could bequantitatively
important
atphysiological
concentrationsofcirculating T4
and could makea
significant
contributiontotheT3 production
rate.Anestimate,
based on our
experimentally
observed renal andhepatic
mi-crosomalyields
inrats(assuming
similar kineticproperties
for the enzyme activities in these twoorgans),
indicates that atcirculating
levelsof freeT4 [3
X10-"1
M(29)],
-38and 80pmol
ofT3/100
gbody
wt could begenerated daily by
thehigh
and lowKm
enzymes,respectively,
in these two organs,thus
accounting
forapproximately
half of the totaldaily
production
rateofT3
inrats, estimatedby
AbramsandLarsen(30)
tobe -250pmol/100 g/body
wtd. Thefindings
ofLumetal.
(31)
that theserumT3
autoregulatory
systemineuthyroid
humans with lowcirculating T4
concentrations isminimally
affectedby
PTU
andiopanoate
might
beexplained
on the basis ofthe relativeinsensitivity
ofthe lowKm
enzyme tothesetwoinhibitors.Therelative
insensitivity
toPTU and thepartial inhibitory
kinetics observed with this inhibitor when enzymeactivitiesaremeasuredatlowsubstrateconcentrations may alsoaccount for theobservations ofSilva et al.(6)
that,
in PTU-treated rats,
circulating T3
concentrations cannot be lowered below aboutone-third of the normallevels (evenwitha
dosage
ashigh
as 1mg/
100gbody wt).
Acknowledgment
This workwassupportedbygrant A-2585 from the National Institute ofArthritis, Diabetes, DigestiveandKidneyDiseases.
References
1. Fekkes, D., E. Van Overmeeren, G. Hennemann, and T. J. Visser. 1980. Solubilization and partial characterization of rat liver
iodothyroninedeiodinases. Biochim. Biophys. Acta. 613:41-51.
2. Leonard, J. L.,and I. N. Rosenberg. 1981. Solubilization ofa
phospholipid-requiringenzyme, iodothyronine5'-deiodinase, fromrat
kidneymicrosomes. Biochim.Biophys.Acta. 659:205-218.
3. Leonard, J. L., andI. N. Rosenberg. 1980. Iodothyronine 5'-deiodinase fromratkidney:substrate specificity and the5'-deiodination
ofreversetriiodothyronine. Endocrinology. 107:1376-1383.
4. Visser, T.J.,J. L. Leonard, M. M. Kaplan, and P. R. Larsen. 1982. Kinetic evidences suggesting two mechanisms foriodothyronine
5'-deiodination in rat cerebral cortex. Proc. Natl. Acad. Sci. USA. 79:5080-5084.
5. Visser, T. J., M. M. Kaplan, J. L. Leonard, and P. R. Larsen. 1983. Evidence fortwopathways of iodothyronine 5'-deiodinationin
rat pituitary that differ in kinetics, propylthiouracil sensitivity, and response to hypothyroidism. J. Clin. Invest. 71:992-1002.
6.Leonard, J. L., S. A. Mellen, and P. R. Larsen. 1983. Thyroxine
5'-deiodinase activityinbrown adipose tissue. Endocrinology. 112:1153-1156.
7.Silva, J. E., J. L. Leonard, F. R. Crantz, and P. R. Larsen. 1982. Evidence for two tissue-specific pathways for in vivo thyroxine
5'-deiodinationin the rat. J. Clin.Invest. 69:1176-1184.
8. Reed,L.J., M.Koike, M. E.Levitch, and F. R. Leach. 1958.
Studieson the nature and reaction of protein-bound lipoic acid. J. Biol. Chem. 232:143-158.
9. Silver, M. 1979. Thin layer chromatography of lipoic acid,
lipoamide and persulfides.Methods Enzymol.62:135-137.
10. Goswami, A., J. L. Leonard, and I. N. Rosenberg. 1982. Inhibition by coumadin anticoagulants of enzymaticouterring
mon-odeiodination ofiodothyronines. Biochem. Biophys. Res. Commun. 104:1231-1238.
11. Chopra, I. J., S. Y. Wu, Y. Nakamura, and D. H. Solomon. 1978. Monodeiodination of 3,5,3'-triiodothyronine and 3,3',5'-triio-dothyronineto3,3'-diiodothyronineinvitro.Endocrinology.
102:1099-1106.
12.Visser, T. J. 1979.Mechanism ofiodothyronine5'-deiodinase.
Biochim. Biophys.Acta. 569:302-308.
13.Dixon,H. B.F., and K F.Tipton.1973.Negatively cooperative
ligandbinding. Biochem.J. 133:837-842.
14. Goswami, A., and I. N. Rosenberg. 1983. Stimulation of iodothyronine outerring monodeiodinase by dihydrolipoamide.
En-docrinology. 42:1180-1187.
15. Chopra, I. J., D. H. Solomon, U. Chopra, S. Y. Wu, D. A.
Fisher and Y. Nakamura. 1978. Pathways of metabolism ofthyroid
hormones. RecentProg. Horm. Res. 34:521-567.
16. Fekkes, D., G. Hennemann, and T. J. Visser. 1982. One enzyme for the 5'-deiodination of 3,3',5'-triiodothyronine and
3',5'-diiodothyroninein rat liver. Biochem. Pharmacol. 31:1705-1709. 17.Dixon, M.,and E.C. Webb. 1979.Enzymes. Academic Press, New York. 90-91.
18.Rudolph,F. B., and H. J. Fromm. 1979.Plottingmethodsfor
analyzingenzymeratedata. MethodsEnzymol.63:138-159.
19. Sizer,I. W. 1943. Effects of temperature on enzyme kinetics. Advances Enzymol. 3:35-62.
20. Wong, J. T.-F. 1975. Kinetics of Enzyme Mechanisms. Academic Press, New York. 219-223.
21.Dixon, M., and E. C. Webb. 1979. Enzymes. Academic Press, NewYork. 342.
22.Cleland,W. W.1963. The kinetics of enzyme-catalyzed reactions with two or more substrates or products.II.Inhibition: nomenclature andtheory. Biochim. Biophys.Acta.67:173-187.
23. Dixon, M., and E. C. Webb. 1979. Enzymes. Academic Press, New York. 105-109.
24. Frieden, C. 1964. Treatmentofenzyme kinetic data. 1. The
effect of modifiers on the kinetic parameters of single substrate enzymes. J. Biol. Chem. 239:3522-3531.
25. Cleland, W. W. 1970. Steady statekinetics. In The Enzymes (Student Edition). Vol. II. P. D. Boyer, editor. Academic Press, New York. 1-65.
26.Cleland,W.W. 1963. Thekinetics ofenzyme-catalyzedreactions with two or more substratesor products. I. Nomenclature andrate
equations. Biochim.Biophys.Acta.67:104-137.
27. Leonard, J. L., and I. N. Rosenberg. 1980. Characterizationof essentialenzymesulfhydrylgroups ofthyroxine-5'-deiodinase fromrat kidney.Endocrinology. 106:444-451.
28.Visser, T. J., D. Fekkes, R. Docter, and G. Hennemann. 1978. Sequential deiodination of thyroxine in rat liver homogenates. Biochem.
J. 174:221-229.
29.Refetoff, S.,N. I.Robin,and V. S. Fang. 1970. Parameters of
thyroid functioninserumof 16 selected vertebratespecies:astudyof
PBIserumT4, free T4, and the pattern of T4 and T3bindingto serum
proteins. Endocrinology. 86:793-805.
30. Abrams, G. M., and P. R. Larsen. 1973.Triiodothyronineand
thyroxinein theserum andthyroid glandsofiodine-deficientrats.J.
Clin.Invest. 52:2522-2531.
31. Lum, S. M.C.,R. Morris,C. A. Spencer,and J. T. Nicoloff. 1983. Dualperipheral 5'-deiodinase (51))systemsaffecting circulating
T3 levels in man.65thAnnu.Meet. EndocrineSoc.,266.(Abstr.)