6870
`
`Biochemistry 1984, 23, 6870-6876
`
`Studies on the Polyglutamate Specificity of Thymidylate Synthase from
`Fetal Pig Liver
`
`Yong-Zhi Lu,~ Patrick D. Aiello, and Rowena G. Matthews*
`
`AaSrP, ACr: Thymidylate synthase has been purified 1700-fold
`idues, but derivatives with two to seven glutamyl residues all
`from fetal pig livers by using chromatography on Affigel-Blue,
`bind at least 30-fold more tightly than the monoglutamate.
`DEAE-52, and hydroxylapatite. Steady-state kinetic mea-
`When CH~-H4PteGlu~ is used as the one carbon donor for
`surements indicate that catalysis proceeds via an ordered se-
`thymidylate biosynthesis, the order of substrate binding and
`quential mechanism. When 5,10-methylenetetrahydro-
`product release is reversed, with binding of CH~-H4PteGlu4
`pteroylmonoglutamate (CH2-H4PteGIu~) is used as the sub-
`preceding that of dUMP and release of dTMP preceding
`release of H2PteGlu4. Vmax and Km values for dUMP and
`strate, dUMP is bound prior to CH2-H4PTeGIul, and 7,8-
`dihydropteroylmonoglutamate (H2PteGlul) is released prior
`CHz-H4PteGlu. show relatively little change as the poly-
`to dTMP. Pteroylpolyglutamates (PteGlun) are inhibitors of
`glutamate chain length of the substrate is varied. Comparison
`thymidylate synthase activity and are competitive with respect
`of the kinetic data obtained in these studies with earlier studies
`to CH~-H4PteGlu~ and uneompetitive with respect to dUMP.
`on methylenetetrahydrofolate reductase from pig liver
`Inhibition constants (Ki values), which correspond to disso-
`[Matthews, R. G., & Baugh, C. M. (1980) Biochemistry 19,
`2040-2045] leads us to predict that the partitioning of limiting
`ciation constants for the dissocation of PteGlun from the en-
`zyme-dUMP-PteGlun ternary complex, have been determined
`concentrations of CHz-H4PteGlun between the reactions
`for PteGlun derivatives with one to seven glutamyl residues:
`catalyzed by thymidylate synthase and methylenetetra-
`hydrofolate reductase will vary with polyglutamate chain
`PteGluL, 10 ~zM; PteGlu2, 0.3 #M; PteGlu3, 0.2 uM; PteGlu4,
`length, with hexaglutamyl substrates preferentially being re-
`0.06/~M; PteGlus, 0.10 ~zM; PteGlu6, 0.12 ~tM; PteGlu7, 0.15
`~tM. Thus, thymidylate synthase from fetal pig liver prefer-
`duced to methyltetrahydrofolate.
`entially binds pteroylpolyglutamates with four glutamyl res-
`
`Intracellular folate derivatives are present mainly as ptero-
`ylpoly-’y-glutamates with 2-10 glutamyl residues (Brown et
`al., 1974; Eto & Krumdieck, 1981; Foo & Shane, 1982).
`There is evidence to suggest that the distribution of pteroyl-
`polyglutamates is species dependent (Priest et al., 1981) and
`may differ with the nature of the folate derivative (Eto &
`Krumdieck, 1981) and with metabolic conditions inside the
`cell (Foo & Shane, 1982; Eto & Krumdieck, 1982). Synthesis
`of a polyglutamate "tail" requires considerable expenditure
`of cellular energy and presumably results in compensatory
`advantages to the cell. However, the rationale for formation
`of long-chain pteroylpolyglutamates remains unclear.
`Our laboratory has been examining the binding properties
`of a series of folate-dependent enzymes for pteroylpoly-
`glutamate derivatives. In addition to the present study on
`thymidylate synthase, we have examined methylenetetra-
`hydrofolate reductase (Matthews & Baugh, 1980), serine
`hydroxymethyltransferase (Matthews et al., 1982), and me-
`thylenetetrahydrofolate dehydrogenase (Ross et al., 1984).
`These studies have utilized pteroylpolyglutamate derivatives
`as inhibitors of these enzymes and have determined dissoci-
`ation constants by kinetic or thermodynamic measurements
`for a series of inhibitors which differ only in the number of
`glutamyl residues. All studies have been performed with
`enzymes isolated from pig liver, to eliminate species differences
`in polyglutamate specificity. Once the specificity of each
`enzyme for the length of polyglutamate tail is known and the
`binding energies associated with the interaction of each glu-
`
`* From the Biophysics Research Division and the Department of Bio-
`logical Chemistry, The University of Michigan, Ann Arbor, Michigan
`48109. Received May 17, 1984. This work was supported in part by
`National Institutes of Health Grant GM 30885 (R.G.M.) and by an
`NIH Short-Term Training Grant for Students in Health Professional
`Schools (T-35-NS-07197) (P.D.A.).
`t Permanent address: Fujian Normal University, Fuzhou, Fujian,
`People’s Republic of China.
`
`tamyl residue with the enzyme have been determined, we can
`also determine how significantly binding of the polyglutamate
`tail affects the kinetic parameters associated with catalysis
`utilizing pteroylpolyglutamate substrates.
`Such studies provide base-line information which can be
`used to predict the flux of folate metabolites through com-
`peting pathways. They are also helpful in predicting the
`pharmacological and physiological effects that poly-
`glutamylation will have on the inhibition of individual enzymes
`by folate and antifolate derivatives.
`
`Experimental Procedures
`Purification of Thymidylate Synthase. Fetal pigs were
`obtained from a local slaughterhouse. The highest thymidylate
`synthase activity was associated with pigs weighing less than
`225 g. Enzyme preparation was initiated within 6 h of
`slaughter, since enzyme activity in the pigs, dissected livers,
`or homogenates was not stable to storage at either 4 or -20
`oC. The livers were dissected out and washed with 0.9% NaC1.
`They were homogenized in a small Waring blender in 50 mM
`potassium phosphate buffer, pH 7.5, containing 0.1 M NaC1,
`50 mM 2-mercaptoethanol, and 100 ~zM dUMP.~ The ratio
`of buffer to dissected livers in the homogenate was 1 mL of
`buffer/g of liver. The homogenate was centrifuged at 30000g
`for 1 h.
`The supernatant was decanted and assayed for thymidylate
`synthase activity (see below). Affigel-Blue beads (Bio-Rad),
`
`Abbreviations: CHz-H4PteGlun, 5,10-methylenetetrahydropteroyl-
`polyglutamate with n glutamyl residues; H:PteGlu~, 7,8-dihydro-
`pteroylpolyglutam~te with n glutamyl residues; CH3-H4PteGIu~, 5-
`methyltetrahydropteroylpolyglutamate with n glutamyl residues; dTMP,
`thymidylate; dUMP, deoxyuridylate; HTP, hydroxylapatite; DTT, di-
`thiothrei~ol; CH~-H~folate, 5-methyltetrahydrofolate; CH:-H~folate,
`5,10-methylenetetrahydrofolate; H~fotate, 7,8-dihyrofolate; FdUMP,
`5-fluorodeoxyuridylate; Tris, tris(hydroxymethyl)aminomethane; EDTA,
`ethylenediaminetetraacetic acid.
`
`0006-2960/84/0423-6870501.50/0
`
`© 1984 American Chemical Society
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1111-0001
`
`

`
`POLYGLUTAMATE SPECIFICITY OF THYMIDYLATE SYNTHASE
`
`VOL. 23, NO. 26, 1984 6871
`
`previously equilibrated with the homogenizing buffer and then
`allowed to settle in the buffer, were added to the supernatant
`in the proportion 100 mL of settled beads per 0.22 unit of
`activity (in/~mol min-~). The suspension was stirred overnight
`at 4 °C. The beads were collected by filtration in a Buchner
`funnel, and the filtrate was discarded. The beads were re-
`suspended in 50 mM phosphate buffer, pH 7.5, containing 0.1
`M NaC1, 50 mM 2-mercaptoethanol, and 100 ~M dUMP and
`collected by filtration. Rinsing was repeated until the ab-
`sorbance at 280 nm was less than 0.1 when measured vs. a
`buffer blank. A slurry of beads in the same buffer was poured
`into a 2.5-cm diameter column. The enzyme was eluted with
`1 M NaCI in 50 mM phosphate, pH 7.5, containing 50 mM
`2-mercaptoethanol and 100 gM dUMP. The elution was
`performed at 4 °C, and the column was pumped at a flow rate
`of 16 mL/h. Fractions, 4 mL, were collected and analyzed
`for thymidylate synthase activity and for dii~ferential absor-
`bance at 280 nm. The active fractions (8-19) were pooled and
`concentrated in an Amicon concentrator with a PM30 mem-
`brane and were then dialyzed overnight in 10 mM phosphate
`buffer, pH 7.5, containing 10 mM 2-mercaptoethanol.
`The dialyzed enzyme was applied to a 2.5 × 10 cm column
`of DEAE-52 (Whatman), previously equilibrated with 50 mM
`Tris-HC1 buffer, pH 7.8, 20% glycerol, 10 mM 2-mercapto-
`ethanol, and 1 mM EDTA.l The column was eluted with a
`linear gradient of 0-0.1 M KC1 in the same buffer, and 5-mL
`fractions were collected. Active fractions (39-47) were col-
`lected, concentrated, and dialyzed overnight in 10 mM
`phosphate buffer, pH 7.5, containing 10 mM 2-mercapto-
`ethanol.
`The enzyme was then applied to a 1.5 x 7 cm Bio-Rad HTP
`column previously equilibrated with glass-distilled water. The
`adsorbed enzyme activity was eluted with a 250-mL linear
`gradient of 0-0.1 M phosphate buffer, pH 7.5, containing 5
`mM DTT, and 4-mL fractions were collected. Active fractions
`(24-38) were pooled and concentrated to an activity of 0.026
`umol min-~ (mL of enzyme solution)-1. The enzyme solution
`was clarified by centrifugation, brought to 20% in glycerol,
`and then stored at -70 °C. Enzyme solutions prepared in this
`manner showed little or no loss of activity after storage for
`several weeks.
`Methods for Assay of Thymidylate Synthase Activity.
`Enzyme activity was monitored during the purification of the
`protein by measuring tritium release from [5-3H]dUMP in
`the presence of (6RS)-CH2-H4folate as initially described by
`Lomax & Greenberg (1967). Assay mixtures, 0.2 mL, con-
`tained 0.1 mL of enzyme and 0.1 mL of an assay cocktail [100
`mM Tris-HC1 buffer, pH 7.4/50 mM MgC12/30 mM form-
`aldehyde/200 mM 2-mercaptoethanol/2 mM EDTA/1.25
`mM (6RS)-H4folate/0.2 M NaF/125 uM [5-3H]dUMP
`(94 000 dpm/nmol)]. The mixture was incubated for 15 min
`at 37 °C and then quenched by addition of 1 mL of activated
`charcoal in 4% perchloric acid (20 g/100 mL). The quenched
`mixture was incubated for 5 rain at 37 °C with frequent
`shaking and then centrifuged in an Eppendorf microfuge at
`15600g for 3 rain. An aliquot of the supernatant, 0.3 mL, was
`added to 5 mL of aqueous counting scintillant (Amersham).
`The dpm values detected in each sample were corrected for
`the dpm detected in a control assay, in which the charcoal/
`perchloric acid quench mixture was added to the enzyme prior
`to addition of the assay cocktail. Enzyme units are micromoles
`of 3H released per minute.
`For kinetic studies utilizing the purified enzyme a spec-
`trophotometric assay was used. The assay involves mea-
`surement of the absorbance changes at 340 nm accompanying
`
`the conversion of CH~-H4folate to H2folate (Wahba &
`Friedkin, 1961). In calculating activities a molar extinction
`coefficient of 6152 M-1 cm-~ was used for the absorbance
`change at 340 nm associated with the conversion of CH2-
`H4folate to H2folate. This molar extinction coefficient was
`calculated from the ratios of 340-nm absorbance to peak ab-
`sorbance of our preparations of CH2-H4folate (340/297 =
`0.045) and H~folate (340/282 -- 0.267) by using 32 000 M-~
`cm-~ as the molar extinction coefficient for CHz-H4folate at
`297 nm (Blakley, 1960a) and 28 400 M-1 cm-~ as the molar
`extinction coefficient for Hzfolate at 282 nm (Blakley, 1960b).
`Assay mixtures, 1.0 mL, contained approximately 2.6 × 10-3
`unit of thymidylate synthase (as measured by the tritium
`release assay)/0.1 M phosphate buffer, pH 6.8/100 uM
`dUMP/20/~M (6RS)-CHz-H4folate. The CH2-H4folate was
`prepared as a mixture of 0.05 M NaHCO3/2 mM (6RS)-
`H4folate/1.3 mM formaldehyde/50 mM 2-mercaptoethanol
`and was stored at -20 °C under nitrogen prior to use. Such
`solutions were stable for several weeks. The assay mixture
`(with CH2-H4folate and enzyme omitted) was equilibrated
`with nitrogen in a 1-mL cuvette for several minutes, CH2-
`H~folate was added, and equilibration with nitrogen was
`continued for two more minutes. The cuvette was sealed with
`parafilm and equilibrated in a 25 °C bath for several minutes,
`and then the blank rate was measured at 340 nm in a spec-
`trophotometer. The assay was initiated by addition of enzyme.
`Measurements were made on a recorder with an expanded
`scale (0.1 absorbance full scale) and a spectrophotometer with
`an optical offset.
`
`Kinetic parameters were evaluated graphically from dou-
`ble-reciprocal plots by using linear regression analysis.
`Preparation of Folate Substrates and Inhibitors. Pter-
`oylpolyglutamates (PteGlu,) were prepared by solid-phase
`synthesis (Krumdieck & Baugh, 1969, 1982), purchased from
`Dr. Charles M. Baugh, and used without further purification.
`The purity of the pteroylpolyglutamates was checked by
`high-pressure liquid chromatography using a modification of
`the procedure described by Schilsky et al. (1983) for the
`separation of polyglutamate analogues of methotrexate. Ap-
`proximately 1.5 nmol of PteGlun was applied to an Ultrasphere
`ODS column (0.46 x 25 cm) equilibrated with 30% aceto-
`nitrile/70% 5 mM tetrabutylammonium phosphate (Waters
`PIC A) in glass distilled water. The samples were eluted at
`1 mL/min along gradients of 30-50% acetonitrile and 3.5-2.5
`mM tetrabutylammonium phosphate in water over 30 min.
`The column was then eluted isocratically with 50% aceto-
`nitrile/50% 5 mM tetrabutylammonium phosphate in water
`for 15 min. The eluate was monitored at 254 nm. Under these
`conditions each PteGlun derivative chromatographed as a single
`major peak, and this peak comprised 85-90% of the 254-nm-
`absorbing material eluting from the column. In particular,
`contamination by PteGlu~_l was always less than 10%, based
`on peak height ratios. Under these conditions the following
`retention times were observed: PteGlun, 7.0 min; PteGlu2, 12.5
`min; PteGlu3, 18.6 min; PteGlu4, 22.0 rain; PteGlu5, 24.5 rain;
`PteGlu6, 26.4 min; PteGluT, 27.9 rain.
`For studies with CH:-H4PteGlu~, (6RS)-H4PteGlu~ was
`prepared by catalytic hydrogenation of a neutral aqueous
`solution of PteGlul and purified as previously described (Ross
`et al., 1984). Alternatively, (6S)-H4PteGlu~ was prepared by
`enzymatic reduction of PteGluI using dihydrofolate reductase
`from L. casei (Matthews et al., 1982). (6S)-H4PteGIun de-
`rivatives were prepared from PteGlu~ derivatives in the same
`manner. H2PteGlu. inhibitors were prepared by dithionite
`reduction of the corresponding PteGlun derivatives as described
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1111-0002
`
`

`
`6872 BIOCHEMISTRY
`
`LU, AIELLO, AND MATTHEWS
`
`Table I. Purification of Thymidylate Synthase from Fetal Pig Livera
`
`step
`
`supernatant from 30000g centrifugation
`eluate from AffiGel Blue
`eluate from DEAE-Sephadex
`eluate from Bio-Rad HTP
`
`activity
`(unit)b
`
`0.23
`0.045
`0.066
`0.066
`
`protein
`(mg)c
`
`12 420
`325
`45.2
`2.1
`
`volume
`(mL)
`
`180
`50
`58
`38
`
`sp act.
`(unit!rag)
`
`1.85 × 10-~
`1.38 × 10-4
`1.46 × 10-3
`3.14 × 10-2
`
`yield
`(%)
`
`100
`20
`29
`29
`
`purification
`(x-fold)
`
`1
`7.5
`79
`1700
`
`QThis preparation utilized 225 g of fetal pig liver, obtained from approximately 30 fetal pigs. bThe units are micromoles of 3H released per minute
`at 37 °C. cProtein determinations were made by using Bio-Rad protein assay, according to the manufacturer’s directions, and using bovine serum
`albumin as a standard.
`
`Table II: Product Inhibition of the Thymidylate Synthase Reaction With CH2-H4PteGIula as Folate Substrateb
`
`variable substrate
`
`fixed substrate (#M)
`
`inhibitor
`
`inhibition pattern
`
`Ki (gM)
`
`dUMP
`CH2-H4PteGIu1
`
`CH2-H4PteGluI (7.2)
`dUMP (5)
`
`CH2-H4PteGIu~
`
`dUMP (100)
`
`dTMP
`dTMP
`
`dTMP
`
`competitive
`noncompetitive
`
`noncompetitive
`
`CH~-H4PteGIu~
`
`dUMP (10)
`
`H2PteGlu~
`
`noncompetitive
`
`CH2-H4PteGlu~
`
`dUMP (100)
`
`H2PteGlul
`
`noncompetitive
`
`dUMP
`dUMP
`
`CH2-H4PteGIut (7.2) H2PteGluI noncompetitive
`CH2-H4PteGIu~ (72) H2PteGlu~ uncompetitive
`
`Kiq = 20
`Ki(slope) = 60
`Kiq = 20
`Ki(slope) = 1100
`Kiq = 20
`Ki(intercept) = 135
`Kip = 115
`Ki(intercept) = 97
`Kip ---- 95
`Ki(intercept) = 13
`K~p = 90
`
`° Enzymatically reduced (6R)-CHz-H4PteGlu] was used for these experiments. With (6R)-CH:-H~PteGlu~ the Km for dUMP is 1.7 uM, and the
`Km for CH:-H4PteGlu~ is 5.2 ~M, while when racemic (6RS)-CH2-H~PteGlul is used, the Km for dUMP is 8 #M, and the Km for (6R)-
`CH2H4PteGIuI is 8 ~M. We infer that (6S)-CH~-H4PteGluI may have some affinity for the dUMP binding site. bPlots of the primary data are
`available to the interested reader on request.
`
`by Matthews & Baugh (1980).
`Source and Preparation of Other Reagents. [5-3H]dUMP
`was purchased from Amersham and purified on Dowex AG
`l-X8 formate columns as described by Lomax & Greenberg
`(1967). Purified dihydrofolate reductase from L. casei was
`the generous gift of Professor Bruce Dunlap, University of
`South Carolina. dUMP and dTMP were purchased from
`Sigma and used without further purification.
`
`Results
`
`Purification of Thymidylate Synthase. The purification
`procedure utilized to prepare enzyme for kinetic studies is
`shown in Table I. Certain aspects of this procedure deserve
`comment. The thymidylate synthase activity in the initial
`homogenate was very labile if 0.1 M NaC1 was omitted from
`the homogenizing buffer and disappeared with a half-time of
`about 1 h. With 0.1 M NaC1 in the homogenizing buffer,
`activity in the homogenate was considerably stabilized, with
`a half-time for inactivation of about 4 h. Addition of phe-
`nylmethanesulfonyl fluoride to the homogenate did not protect
`the enzyme against inactivation. Once the enzyme had been
`purified by adsorption on Affigel Blue, its lability was greatly
`decreased and NaC1 could be omitted without loss of activity.
`However, activity disappeared on storage at 4 °C for 4 days
`unless 5 mM DTT and 20% glycerol were present in the buffer.
`Inclusion of these agents permitted storage of the enzyme for
`several weeks at this stage of purification without loss of ac-
`tivity at either 4 or -70 ~C. Chromatography on DEAE-52
`could be carried out conveniently in 20% glycerol at 4 °C
`provided that wide, short columns were used. However, rea-
`sonable flow rates during chromatography on Bio-Rad HPT
`at 4 oC could be achieved only if glycerol was removed from
`the protein prior to application to the column. At this stage
`of the purification, glycerol was only required to protect the
`enzyme against inactivation due to freezing.
`Characterization of the Kinetic Mechanism Associated with
`Use ofCH2-HaPteGlu1. Figure 1 shows a kinetic analysis of
`the thymidylate synthase reaction using (6RS)-CH:-
`H,PteGlu~ as the one-carbon donor. It can be seen that the
`
`"7 6-~ -I~
`
`-, 5-
`
`-0.1 0 ~ 0.1 0.29,,,"
`
`Kr.~ = 8~,
`
`I
`
`-0.t0 0.00 0.~0 0.20
`[(6R)-CHz-H4PteGlu~ ]’l (#M
`
`FIGURE 1: Steady-state kinetic measurements with CH2-H4PteGlu~
`as substrate. A double-reciprocal plot of velocity vs. [(6R)-CH~-
`H4PteGlu~] is shown. The folate substrate was added as the racemic
`mixture, (6RS)-CHz-H~PteGIul. The dUMP concentrations were
`(13) 5, (,~) 7.5, (X) !0, (~) 20, and (O) 60 uM. Velocities were
`determined by using the spectrophotometric assay described under
`Experimental Procedures.
`
`double-reciprocal plots converge, indicative of a sequential
`mechanism. The K~ for dUMP is 8 #M, as is the K~ for
`(6R)-CH~-H4PteGluv The order of addition of substrates and
`of release of products was determined by product inhibition
`studies, and these experiments are summarized in Table II.
`The product inhibition studies are indicative of the following
`kinetic pathway:
`
`dUMP
`
`CH~-H4PteGluI
`
`HzPteGluI
`
`dTMP
`
`Measurement of the Inhibition Constants Associated with
`PteGlu, Inhibitors. The Ki values associated with PteGlu,
`inhibitors were measured by using (6RS)-CHz-H4PteGIu~ as
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1111-0003
`
`

`
`POLYGLUTAMATE SPECIFICITY OF THYMIDYLATE SYNTHASE
`
`VOL. 23, NO. 26, 1984 6873
`
`Table III: Inhibition Constants for PteGlun Inhibitorsa
`
`inhibitor
`
`PteGlu~
`
`PteGlu2
`
`PteGlu3
`
`PteGlu4
`
`Ki (~M)
`
`15
`6
`0.38
`0.24
`0.30
`0.13
`0.060
`0.060
`
`inhibitor
`
`PteGlu~
`
`PteGlu6
`
`PteGlu~
`
`Ki (gM)
`
`0.100
`0.099
`0.12
`0.12
`0.19
`0.10
`
`aAll assays were conducted with (6RS)-CH2-H4PteGIu~ as the fo-
`late substrate, under the assay conditions described under Experimen-
`tal Procedures and employing 100 uM dUMP.
`
`Table IV: Kinetic Parameters Associated with CH2-H~PteGlun
`Substratesa
`
`substrate
`
`CH2-H4PteGIu~
`CH2-H4PteGIu:
`CH~-H4PteGIu3
`CH2-H4PteGIu4
`CH~-H4PteGIu~
`CH2-H~PteGIu6
`CH2-H~PteGIu7
`
`rel
`Vmax
`1.00
`0.56
`0.39
`0.36
`0.38
`0.43
`0.37
`
`(dUMP)
`(~M)
`
`(CH:-H4PteGlu,)
`(~M)
`
`1.7
`1.7
`1.7
`1.7
`2.3
`2.5
`2.6
`
`5.2
`2.0
`1.9
`1.9
`1.6
`1.6
`2.1
`
`aAll assays utilized enzymatica!ly prepared (6R)-CH2-H4PteGIun
`derivatives as the folate substrates. As noted in the footnote to Table
`II, somewhat different Km values for dUMP and for CH~-H4PteGIu~
`are obtained when racemic (6RS)-CH2-H~PteGIu~ is used as the sub-
`strate. Assay conditions are described under Experimental Procedures.
`Double-reciprocal plots of velocity vs. [CH2-PteGlu~] and of velocity
`vs. [dUMP] intersected on the x axis. Thus in all cases, Ki~ = K~ for
`the first substrat¢ bound.
`
`the folate substrate in each experiment. A typical experiment
`is shown in Figure 2, with PteGlu4 as the inhibitor. The
`inhibitor pattern is linearly competitive with respect to
`CH2-H4PteGlul and linearly uncompetitive with respect to
`dUMP. Such inhibition patterns are expected if PteGlu4
`competes with CH2-H#PteGlu~ for the folate binding site on
`the enzyme-dUMP binary complex. Similar patterns were
`observed for each PteGlu~ inhibitor, and the Ki values that were
`calculated from replots such as is shown in the inset to Figure
`2A are listed in Table III. In each case, the Ki values cor-
`respond to dissociation constants for dissociation of PteGlun
`inhibitor from enzyme-dUMP-PteGlu~ ternary complexes.
`Measurement of the Kinetic Parameters Associated with
`CH2-H~PteGlu~ Substrates. Table IV summarizes the relative
`values for Vm~x and the Km values for dUMP and CH~-
`H~PteGlu~ obtained from kinetic analyses of the series of
`(6R)-CH2-H4PteGlu~ substrates. These experiments were all
`run on the same day with the same preparation of enzyme.
`Despite the very large differences in affinity of the enzyme
`
`FIGURE 2: Inhibition of thymidylate synthase by PteGlu4. (A)
`Double-reciprocal plots of velocity vs. [(6R)-CH2-H4PteGIu~], in the
`presence of 100 ~M dUMP. The folate substrate was added as the
`racemic mixture (6RS)-CHz-H4PteGIu~. Concentrations of PteGlu~
`inhibitor present were (~) 0, (×) 0.065, (1~) 0.087, (O) 0.108, and
`(~,) 0.152 ~M. (B) Double-reciprocal plots of velocity vs. [dUMP],
`in the presence of 10 #M (6R)-CH2-HaPteGIu~ added as the racemic
`mixture (6RS)-CH2-H~PteGlux. Concentrations of PteGlu~ inhibitor
`were (O) 0, (×) 0.07, (el) 0.12, and (~,) 0.17 gM.
`
`for PteGlu, inhibitors, the variation in V/K with polyglutamate
`chain length is very small. One possible explanation for these
`discrepancies would be that longer chain polyglutamate sub-
`strates result in a change in order of addition of substrates or
`of release of products so that the rate constants contained in
`these kinetic parameters are not comparable for all substrates.
`In order to look for changes in order of addition of substrates
`and!or of product release, we have performed product inhib-
`ition studies using CH~-H~PteGlu4 as the folate substrate and
`H:PteGlu4 and dTMP as product inhibitors. The results of
`these studies are summarized in Table V. These inhibition
`studies are consistent with the following mechanism:
`
`CH2-H4PteGI,,4 dUMP
`
`dTMP
`
`H2PteGlu4
`
`/ts
`
`Comparison with the kinetic scheme obtained by using
`CHrH~PteGlu~ as the folate substrate indicates that use of
`the polyglutamate substrate has resulted in changes in the
`order of addition of substrates and of release of products.
`Similar product inhibition studies with CH2-HaPteGlu: as
`substrate using H2PteGlu2 and dTMP as product inhibitors
`suggest that with this substrate also the folate substrate is
`bound first and the folate product is released last.
`When the inhibition patterns associated with inhibition by
`PteGlu4 are examined with CH2-H4PteGlu4 as the folate
`substrate, PteGlu4 is linearly competitive with respect to
`CH2-H4PteGlu4 but exhibits a noncompetitive pattern with
`respect to dUMP. These inhibition patterns are also consistent
`with CH2-H4PteGlu4 binding prior to dUMP. Furthermore,
`
`Table V: Product Inhibition of the Thymidylate Synthase Reaction with CH:-H~PteGlu~ as Folate Substratea’b
`
`variable substrate
`
`fixed substrate (/~M)
`
`inhibitor
`
`inhibitor pattern
`
`Ki (#M)
`
`CH:-H~PteGlu4
`dUMP
`
`dUMP (10)
`CH2-H4PteGIu~ (7.2)
`
`H2PteGlu~
`H2PteGlu4
`
`competitive
`noncompetitive
`
`Kiq = 3.3
`Ki(slope) = 15
`Kiq = 3.1
`Ki(slope) = 45
`Kiq = 1.2
`Ki(intercept) ~ 400
`Kip = 320
`Ki(intercept) = 222
`Kip = 220
`Ki(intercept) = 50
`dUMP (10)
`CH~-H4PteGIu~
`Kip = 320
`dUMP (100)
`CH~-H~PteGIu4
`aEnzymatically reduced (6R)-CH~-H~PteGlu4 was used for these experiments. ~Plots of the primary data are available to the interested reader on
`request.
`
`dUMP
`
`dUMP
`
`dUMP
`
`CH2-H4PteGIu4 (72)
`
`H~PteGlu4
`
`noncompetitive
`
`CH2-H4PteGlu~ (7.2)
`
`CH2-H~PteGlu~ (64)
`
`dTMP
`
`dTMP
`
`dTMP
`dTMP
`
`noncompetitive
`
`noncompetitive
`
`noncompetitive
`uncompetitive
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1111-0004
`
`

`
`6874 BIOCHEMISTRY
`
`dTMP
`
`dUMP~HzfOlOte
`
`@
`
`X
`
`r,4
`ADPH ~
`
`(~)
`
`biosynthesis
`
`NADP~ "CH3--H4fol0~e ---~ A~oMet
`
`FIGURE 3: Semilogarithmic plots of A(AG) associated with the binding
`of folate inhibitors vs. the number of glutarayl residues on the folate
`inhibitor. (1) For methylenetetrahydrofolate reductase, A(AG) values
`were calculated from Ki values determined kinetically for inhibition
`of enzyme by H2PteGlun inhibitors in the presence of saturating
`NADPH and varied CH2-H4PteGIul (Matthews & Baugh, 1980).
`(2) For thymidylate synthase, A(AG) values were calculated from
`the data in Table III. (3) For serine hydroxymethyltransferase, A(AG)
`values were calculated from spectrophotometric determinations of
`the Ka values for dissociation of CH3-H4PteGlun from E-CH3-
`H4PteGlun-glycine ternary complexes (Matthews et al., 1982). (4)
`For methylenetetrahydrefolate dehydrogenase, A(AG) values were
`calculated from Ki values determined kinetically for inhibition of the
`enzyme by PteGlu, inhibitors in the presence of saturating NADP+
`and varied CH2-H4PteGIu~ (Ross et al.. 1984).
`
`the measured Ki for PteGlu4 binding to the free enzyme
`(measured with CH2-H4PteGlu4 as substrate and 100 ~M
`dUMP) is 0.35 ~M, which is substantially higher than the K~
`associated with PteGlu4 binding to the enzyme-dUMP binary
`complex (0.06 ~,M, measured with CHrH4PteGlul as sub-
`strate and 100 ~M dUMP). Clearly PteGlu~ binds both free
`enzyme and the enzyme-dUMP binary complex. However,
`the relatively simple inhibition patterns we observe with this
`inhibitor suggest that such inhibition studies are relatively
`insensitive to a small amount (less than 20%) of inhibitor
`binding to an alternate enzyme form.
`
`Discussion
`The values for inhibition constants of PteGlun inhibitors
`shown in Table III can be used to calculate the change in free
`energy A(AG) associated with the binding of each of a series
`of polyglutamate inhibitors. In Figure 3, these free energy
`changes associated with the binding of folylpolyglutamates to
`thymidylate synthase are compared with similar profiles ob-
`tained for other folate-dependent enzymes from pig liver in
`earlier studies from our laboratory. We note that the fo-
`late-dependent enzymes differ markedly in their affinity for
`folylpolyglutamates as compared to folylmonoglutamates, and
`they differ also in the chain length of folylpolyglutamate which
`is preferentially bound. Such comparisons suggest that the
`chain length of folylpolyglutamate derivatives may indeed play
`a role in determining how these derivatives will be metabolized,
`as originally suggested by Baggott & Krumdieck (1979). Two
`of the enzymes that have been studied, serine hydroxy-
`methyltransferase and methylenetetrahydrofolate de-
`hydrogenase, catalyze reactions that are thought to be main-
`tained at equilibrium in the cytoplasm and presumably play
`little role in the direction of one-carbon units to various
`metabolic pathways. The other two enzymes, thymidylate
`synthase and methylenetetrahydrofolate reductase, determine
`the flux of one-carbon units through their respective pathways,
`e.g., into thymidylate biosynthesis or regeneration of AdoMet
`
`LU, AIELLO, AND MATTHEWS
`
`Table VI: Partitioning of Limiting Concentrations of
`CHrH4PteGlu. Substrates between the Reactions Catalyzed by
`Methylenetetrahydrofolate Reductase
`
`substrate
`
`CH2-H4PteGIu~
`CH2-H4PteGIu2
`CH2-H~PteGIu~
`CH2-H~PteGlu~
`CH2-H4PteGIu~
`CHrH~PteGIu6
`CH:-H4PteGIu7
`
`relative flux
`
`methylenetetra-
`hydrofolate reductasea
`
`thymidylate
`synthaseb
`
`1
`2.4
`7.2
`15.5
`11.6
`27.1
`5.5
`
`l
`1.4
`1.1
`1.0
`1.2
`1.3
`0.9
`
`flux
`ratio
`
`1
`1.7
`6.5
`15.5
`9.7
`20.8
`6.1
`
`aData calculated from Matthews & Baugh (1979) assuming a cy-
`toplasmic NADPH concentration of 200 ~,M. OData calculated from
`Table IV of this paper assuming a cytoplasmic dUMP concentration of
`60 ~M (Jackson, !978),
`
`via CHrH4folate and methionine, and catalyze reactions that
`are essentially irreversible in vivo. In Table VI we have at-
`tempted to calculate the effect of the polyglutamate chain
`length of CH2-H4folate on the relative rates of incorporation
`of the methylene group into dTMP or into CHrH4folate.
`These calculations assume a cytoplasmic concentration of
`NADPH of about 200 #M (Conway et al., 1983) and a cy-
`toplasmic dUMP concentration of about 60 ~tM [calculated
`from the data of Jackson (1978) assuming 0.68 mL of
`water/109 cells]. The calculations further assume limiting
`concentrations of CHrH4folate (V/Kconditions). Under these
`conditions, it can be seen that the ratio of fluxes through these
`two competing pathways is quite dependent on the poly-
`glutamate chain length of CHrH4PteGlu,, so that if the flux
`ratio through the two pathways is set equal to 1 for the
`monoglutamate, it increases to 20.8 for the hexaglutamate,
`with polyglutamates being preferentially reduced to CH3-
`H4folate. Thus, under conditions of limiting CH2-H4folate,
`chain length is expected to be an important determinant of
`flux.
`Our studies also suggest that one should be very cautious
`about drawing inferences about substrate affinities from
`comparisons of V~,~/K~ with folate substrates of varying
`polyglutamate chain length, particularly for ordered sequential
`mechanisms. Our observation that increased affinity of thy-
`midylate synthase for folylpolyglutamates leads to reversals
`in the order of substrate binding and product release means
`that the kinetic constants contained in Vm~x/Km change with
`polyglutamate chain length and hence that these values can
`not be directly compared.
`Previous studies of thymidylate synthase from a number of
`sources have suggested an ordered mechanism in which dUMP
`binding precedes CH2-H4PteGIu~ binding (Lorenson et al.,
`1967; Dolnick & Cheng, 1977; Danenberg & Danenberg,
`1978; Daron & Aull, 1978; Bisson & Yhorner, 1981). Evi-
`dence in support of an ordered mechanism in which dUMP
`binding precedes CH~-H4PteGIu~ binding was also obtained
`from equilibrium dialysis measurements with the enzyme from
`L. easel (Galivan et al., 1976b). Studies have also been
`performed with enzyme from a variety of sources that exam-
`ined the effect of substrate polyglutamate chain length on V~ax
`and the K~ values for CHz-H4PteGlu,. Thus, the yeast enzyme
`exhibits a 10-fold decrease in K~, when CH_,-H4PteGIus is
`compared with CH2-H4PteGIu~ (Bisson & Thorner, 1981);
`the enzyme from human blast cells shows a 15-fold decrease
`in Km when the same two substrates are compared (Dolnick
`& Cheng, !978), and the enzyme from L. casei shows de-
`creases in both Vmax (3-fold) and the Km for CH2-H4folate
`(! 5-fold) (Kisliuk et al., 1981). The enzyme from calf thymus
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1111-0005
`
`

`
`POLYGLUTAMATE SPECIFICITY OF THYMIDYLATE SYNTHASE
`
`VOL. 23, NO. 26, 1984 6875
`
`Scheme I
`
`ks[CH2-H4PteGIun]
`
`kcat
`E .CH2-H4PteGIun ~ ~ E. dUMP ¯ CH2-H4PteGIun,~ E + dTMP + H2PteGlun
`k8
`
`shows no change in Km with polyglutamate chain length
`(Dwivedi et al., 1983). Thus, in the cases which have been
`examined relatively small changes in kinetic parameters have
`been observed as the polyglutamate chain lengths of CH2-
`H4folate substrates were varied, in agreement with