Ultraviolet Spectrophotometric Characterization of a Glutamate Dehydrogenase-reduced Coenzyme-a-Ketoglutarate Complex*

SUMMARY Glutamate dehydrogenase binds a-ketoglutarate and TPNH in a ternary complex such that both ligands are more tightly bound than in binary complexes. TPNH, bound in this ternary complex, absorbs maximally at 332 nm with &32 = 5.2 rnM-’ cm-l in contrast to the absorbance of “free” TPNH (I&39 = E,m,, = 6.2 m~-l cm-l) and TPNH bound in either a binary or in a glutamate dehydrogenase-TPNH-L-glutamate dead-end ternary complex (Esd8 = E,,, = 4.8 mM-’ cm-l). One molecule of TPNH per 47,000 to 54,000 molecular weight glutamate dehydrogenase (approximately one per peptide chain) was found to be bound in either the ol-ketoglutarate or L-glutamate ternary complex. The spectral shifts associated with ternary complex formation were not observed when the coenzyme and substrate analogs, 3-acetylpyridine DPNH or oc-ketobutyrate, were used. There-fore, an intact amide on the reduced nicotinamide moiety and a gamma-carboxyl of the dicarboxylate substrate must be required

DALLAS G. CROSS SUMMARY Glutamate dehydrogenase binds a-ketoglutarate and TPNH in a ternary complex such that both ligands are more tightly bound than in binary complexes.
One molecule of TPNH per 47,000 to 54,000 molecular weight glutamate dehydrogenase (approximately one per peptide chain) was found to be bound in either the ol-ketoglutarate or L-glutamate ternary complex. The spectral shifts associated with ternary complex formation were not observed when the coenzyme and substrate analogs, 3acetylpyridine DPNH or oc-ketobutyrate, were used. Therefore, an intact amide on the reduced nicotinamide moiety and a gamma-carboxyl of the dicarboxylate substrate must be required for observable complex formation.
The demonstration of the high stability of the glutamate dehydrogenase-TPNH-oc-ketoglutarate complex and the correlation of the binding characteristics of the ligands involved in this complex with the effects of these ligands on catalytic events warrants the inclusion of this complex as either a reactive intermediate or a dead-end complex in all general mechanisms proposed for glutamate dehydrogenase catalysis.
The binding of TPNH to glutamate tloh\,drogel~:rsc ill t,he absence and in the presence of L-glutamate was sho\\rn ill this laboratory (1) to be characterized by a rctl-shifted diffcrc~llcespectrum of the reduced nicotinnmide absorbance. Since that, time both red-and blue-shifted reduced nicotinamide spcct,rn have been observed during the osidative tlcaminntioll of Lglutamate by glutamate dehydrogenase and Tl'N+ with stopped flow spectrophotometry (2,3). This blueshiftjed si)c:c:icls of TPNH was tentatively assigned by Fisher et al. (2) to either a * This work was supported in part by United States Natiolrtll Institutes of Health Grant GM15188 and by United Stales National Science Foundation Grant GR20923 to Harvey F. Fisher in whose laboratory these experiments were performed. transitory GJ~I11-1'l'NI-I-NT14+~otE<(~ complex or :I C;l)H-'I'I'NH-QKG complex OIL the basis of :I dcutcriurn isotope effect. .Just prior to the complet,ioll of this inrestigation, di I~'ranco ant1 lwa,t subo (3) reported the: (X)11-TPNH-otE<G complex to po~ss a blue-shifted spectrum usill, 0 stopl)etl flow displacement, studies. 'I'his complex, however, 11:~s not been coIlfirmed as obligatory for catalysis.
Iwatsubo and I'antaloni (6), usitlg stopped flow evidence, suggest that, the rate-limiting step of t,lr~ osidativcl rc:tction with L-glutamate is the release of rcducetl cocllz)-me suhscquent to hydride transfer. This report Ijrescnts the absolute spcctrrun of TPNII in the glutamate dchydrogel~asC~Tl'NII~~-k(:togllltaratc complcz obi nined with direct, ultjrnT-iolct, spcctrol)hotometric techniques, shows the formatiolr or this coml)lcz to involve an aromatic amino acid residue 011 the clnzymc, h (Tires the stoichiometr>-of c~omplcs formation, tlcscribes some of the fulrctional group requirenrcnts on the suljstrate and coenzyme nccessnr~ for con~plc~ I'ormation, coml)arcs this complex to free TPXH and TPNII irl other glutamate dehydrogcnase complexes, and relates the characteristics of this caomplex to some of the propcrtics of ~lrIt:nnat,c: tlehydrogenasc: catalysis.
TI'NII was also a product of Sigma. The 3-acetylpyridillc analog of I>PNI T was obtainctl from Calbiochem.

785
Ovarian model 620/i computer was used to collect all spectral data. Each spectrum presented in this study represents an average of four to five spectra collected from consecutive runs.
The absorbance values and peak positions given in the text were determined with the numerical values of each averaged spectrum. All spectra mere recorded as follolvs: an averaged base line was obtained with equal amounts of glutamate dehydrogenase in the sample and reference cells mixed with equal concentrations of the substrates as indicated in the text. Coenzyme, made up in an identical concentration of substrate, was added to the sample cell with Lang-Levy pipets and an equal volume of substrate, without coenzgme, was added to the reference cell. The spectra were recorded after mixing the contents of both cells with nonwettable polypropylene stirrers and waiting for 10 to 15 min. Since coenzyme is present only in the sample cells and enz)-me is present in both the sample and reference cells, this arrangement results in the spectra of free and bound coenzyme and any difference spectra resulting from the perturbation of enz) me aromatic amino acid absorption usually observable in the 280. to 300.nm region and superimposed, of course, on any absorbance contributions of the coenzyme. This method also results in a slight dilution of the enzyme during the titration (each of the Fcven to ten additions was 0.01 ml to an initial volume of 3.00 ml). For calculations in the text this dilution is compensated by assuming that the binding signal5 were proportional to enzyme concentration. The error introduced in the calculations with this assumption is less than 3?{. 1. Spectra of reduced nicotinamide absorbance (300 to 390 nm) and difference spec*trx of glut:Lmate dehydrogenase aromatic amino acid pert,urbation (280 to 300 nm) resulting from the titration of enzyme and dicarboxylete substrate with TPNH. A, 32 rn~ L-glutamate and 2.2 mg per ml of glutamate dehgdrogenase titrated with TPNH. B, 5.5 m>r a-ketoglutarate and 1.85 mg per ml of glut,nmate dehydrogennse titrated with TPNH. Other experimental conditions are to be found in the text and in Fig. 2. The cl ashed line indicates absorption maxima as determilled from the llllmerical values of absorption for each titration.
The dashed lines in both figures indicate the changing position of the maximum absorbance during the titrations. Fig. 1, A and 13, shows that for the first few titrations TPNH binds in the presence of glutamate exhibiting a spectrum with maximum absorbance at 348 nm and in the presence of a-ketoglutarate it binds exhibiting a spectrum with maximal absorbance at 332 nm. Both titrations show the expected shift of the maxima toward that, of free TI'NH at 339 nm upon accumulation of free TPNH in the system. In addition to the change in the reduced nicotinamide absorption spectrum, small feature5 in the 280-to 300.nm region are difference spectra attributable to the perturbation of enzyme aromatic amino acid absorption. These same spectral features are also evident when TPNH binds to the enzyme in a binary complex (1, 9, 7) and have been previously reported as features of a difference spectrum resulting from the binding. of glutamate to a glutamate dehydrogenase-DPKH complex (10). Theso spectral features are assigned to aromatic amino acid perturbation since neither the solvent perturbation difference spectra of reduced coenzyme, the adenine chromophore, nor that of reduced nicotinamide mononucleotide showed the 292 nm peak and 288 nm trough depicted in Fig. 1, A and B (11). It is conceivable that the coenzyme absorbance is so altered as to show the 292 to 288 nm feature normally attributed to changes in aromatic amino acid absorbance (12, 13) but no such changes have been reported.
The extinction coefficient, of binding is the same for this feature whether measured for the binding of coenzymc in a GDH-TPNH binary complex or in the GDH-TPNH-G or GDH-TPNH-olKG ternary complexes. Fig. 2  TPNII-G and GDI-I-TPNH-orKG complexes were calculated from the differences in the spectra shown in Fig. 1 in the regions which exhibited linear binding in Figs. 2 and 3. Thus, for GDH-TPNH-G formation EFax = 4.8 mM-1 cm-1 and for GDH-TPNH-aKG formation EFs:"," = 5.2 rnM-1 cm-l. These contrast with the 6.2 mM extinction coefficient of "free" TPNH at 340 nm. Using stopped-flow spectrophotometry, di Franc0 (16) describes the spectrum of the GDH-TPNH-aKG complex as having a h,,,, = 335 nm and E340 = 5.3 rni+r1 cm-l, whereas the X,,, reported herein was at 332 nm and EaJO is 5.0 mM1 cm-l.
The spectra of 1 InM TPNH in a GDH-TPNH-aKG complex, in a GDH-TPNH-G complex (both superimposed on enzyme perturbation in the 280-to 300-nm region), and as a free species in soluOion are shown in Fig. 4. It is apparent that the spectra of both complexes are hypochromic to that of free TPNH.
A similar hypochromism is also exhibited when TPNH binds to glutamate dehydrogenase in a binary complex (9). Possible causes of hypochromism in this system include an orientation of the nicotinamide chromoyhore to another chromophore or an orientation of nicotinamide to a charged group (17). It has been shown previously that hyperchromism of the 340-nm band of TPNH absorption results when the native conformation of coenzyme in solution is changed with a concomitant disorientation of the coparallel, coplanar orientation of the nicotinamide and adenine rings (11). Thus, the hypochromism evidenced in both the GDH-TPNI-I-G and GDH-TPNH-aKG complex formation is most likely due to an orientation of the nicotinamide ring to either the adenine portion of the coenzyme, to an enzyme aromatic amino acid chromophore or possibly to a chsrge on the enzyme or coenzyme.
13ccause hypochromism is present in the spectra of both GDII-'TPNH-G and GDII-TI'NH-aKG, it appears that the orientations of the reduced nicotinamide rings are relatively equivalent for both species.
In addition to the hypochromism exhibited upon coenzyme binding in the two ternary complexes, the binding of TPNH in the GDH-TPNH-aKG and GDI-I-TPNH-G complexes produces blue-and red-shifted spectra, respectively.
Causes of spectral shifts seen when reduced coenzyme associates with the various dchydrogenases have been discussed at length (9), but they arc briefly: spectral shifts due to changes in the polarizability of the general medium surrounding the chromophorc, the effect seen in solvent perturbation experiments, and spectral shifts due to changes in the electronic configuration of a chromophore caused by specific interactions such as by hydrogen bond formation or breaking.
To ascribe the specific causes of the red and blue shifts observed above is, of course, hazardous but in view of the constant hypochromicity and the large (8 nm) absorbance shifts seen in the spectra of the GDH-TPNH-G and GDH-TPNI-I-aKG complexes, it now appears likely that these shifts are a result of specific changes in the electronic configuration of the nicotinamide ring rather t'han the result of general medium effects arising from changes in coenzyme conformation or the burying of the ring in a hydrophobic pocket.
A survey of the dehydrogenase-reduced coenzyme binding difference spectra (11) has shown that all dehydrogenases which stereospecifically transfer hydrogen to the A side of the nicotinamide ring exhibit blue shifts of the reduced nicotinamide absorbance and that most B-stereospecific dehydrogenases produce red shifts. Since both red-and blue-shifted reduced nicotinamide spectra have been demonstrated for two complexes 7  ,  I  I  I  I  I  I  I  I  I  I X h-n) FIG. 4. Spectra of 1 rnnf TPNH (---), bound to glutamate dehydrogenase in the presence of ar-ketoglutarate (--), or L-glutamate (e-e. ). The latter two spectra are superimposed, in the 280-to 300-nm region, on difference spectra resulting from the perturbation of enzyme absorption.
FIG. 5. Partial reaction scheme for the reduction of TPN+ by glutamate dehydrogenase.
of glutamate dehydrogenase, a B-stereospecific dehydrogenase, the relationship of dehydrogenase stereospecificity with the direction of the spectral shift now appears to be less direct.
3-Acetylpyridine DPNf is an excellent coenzyme analog for the glutamate dchydrogenase reaction (18). The steady state rat,es using this analog are, in fact, higher than when either DPN+ or TPPU'+ are used. It has been shown that the 363.nm maximum absorbance of this reduced coenzyme analog remains unshifted when bound to glutamate dehydrogenase (7j. When 3-acetylpyridine DPNH was substituted for TPTH in titrations similar to those described above, there was no observable red or blue shift of the 363 nm peak even though the concentrations of glutamate and oc-ketoglutarate were increased lo-and 50fold over what was used to demonstrate the TPNH complexes. It has been previously shown that the red shift of the reduced nicotinamide absorption of TPNH was a consequence of the binding of the nicotinamidc moiety to a subsite which forms a part of the coenzyme binding site for both TPXH and DPNH on the enzyme. This subsite IT-as designated Subsite I (7). In order for coenzgme to bind the nicotinamide moiety at Subsite I it must have an intact amide group; which is not the case with 3.acetylpyridine DPNH.
Since neither red nor blue shifts were evident in the experiments using 3-acetylpyridine DPNH, the appearance of the shifted spectra associated with GDH-TPNH-aKG as well as GDH-TPNH-G formation also requires the binding of the reduced nicotinamide moiety to Subsite I on the enzyme.
The binding of glutamate to the glutamate dehydrogenasecoenzyme complex has been characterized by Caughey From these data the subsites which bind the alphaand gamma-carboxyl groups have been designated as Xubsites V-alpha and V-gamma and from other evidence the subsite which binds the amino group of glutamate was designated Subsite IV (7). A consequence of substrate binding to these subsites is the mutually enhanced binding of TPNH and dicarboxylate substrate since coenzyme and substrate are both afforded additional means of attachment to the enzyme surface.
It is well known that glutamate dehydrogenase can utilize a large number of monocarboxylic amino-and keto-substrates (20, 21). L-oc-Aminobutyrate and oc-ketobutyrate4 are both es&lent glutamate dchydrogenase substrates. Therefore, these monocarboxylate substrates were examined to determine whether they formed ternary complexes with enzyme and TPNH. Base lines lvere recorded with 1.52 mg per ml of glutamate deh>-drogenasc and 40 /.LM TPNH present in both the sample and reference compartments of the spectrophotometer and additions of up to 25 rnrvr oc-ketobutyrate and 77 rntir Lu-aminobutyrate were made t,o the enzyme-coenzyme mixture in the sample compartment.
Any shift of the reduced nicotinamide absorbance would bc easily observable by the production of a difference spectrum upon addition of the monocarboxylate substrates but none was seen. The experimental conditions used to attempt to observe moilocarboxylate complex formation were the same as those described for ternary complex formation by the dicarboxylate substrates which showed gross difference spectjra using t,his same technique.
This lack of enhanced TPNH bindilrg in the presence of high concentrations of the monocarboxylat,e substrates, cr-ketobutyrate and a-aminobutyrnte, shows that the gamma carboxyl moieties of the dicarbosylnte substrates, cr-ketoglutarate and L-glutamate, are required for the mutually enhanced binding of TPNH and substrate observed in the formation of the blue-shifted GDH-TPSIS-crKG coml~lex and the red-shifted GDII-TPNH-G complex.
Thus, the gamma carboxyl of the dicarboxylate substrates and the amide of the reduced coenzymc are both necessary for the formation of obhervable GDH-TPNH-crKG and GDII-TPKI-I-G complexes. A partial reaction scheme for catalysis by glutamate dehydrogenase, using part of a random mechanism scheme suggested by Engel and Dalziel (5) and including a dead-end GDII-TI'XII-G complex is shown in Fig. 5. Frieden (4) and Iwatsubo and Pantaloni (6) have shown that the release of reduced coenzyme from the enzyme is the ratelimiting step of the reaction using TI'N+ and L-glutamate. Iwatsubo and Pantaloni (6) also conclude that the rate-limiting step of the reaction, using monocarboxylic amino acid substrates, is not t.he release of coenzyme from glutamate dchydrogenase. This kinetic consequence is a result of the lack of a gammacarbosyl group on the substrate and is consistent with the observed lack of enhanced TPNII binding when amounts of 01aminobutyrate or ac-ketobutyrate , sufficient for catalytic activity, are added to a solution of glutamate dehydrogenase and TPNH. That is, the rate of release of coenzyme, using monocarboxylate substrates, is not lowered by formation of tight GDH-TPNH-aKG or deacl-end (:DH-TPNH~G complexes.
If it is assumed that the lack of euhnnced TPNH binding is iudicative of a