Bioconlugate Chem. 1881, 2, 142-147
142
Identification of Peptides in the Adenine Ring Binding Domain of Glutamate and Lactate Dehydrogenase Using a-Azido-NAD+ Hyuntae Kim and Boyd Haley' University of Kentucky, Department of Biochemistry, College of Medicine, Lexington, Kentucky 40536. Received November 20, 1990 Previously, the synthesis and validation of [32P]2N3NAD+as an active site directed photoaffinity probe for glutamate dehydrogenase (GDH) was reported (8). This report shows that 2N3NAD+ is also an effective probe for the NAD+ binding site of lactate dehydrogenase (LDH). With the appropriate photolabeling procedures and immobilized boronate column chromatography the active site peptides of GDH and LDH involved in the adenine base binding domain have been isolated and sequenced. With both GDH and LDH a single photolabeled peptide, which contained the majority of the photoinserted radiolabel, was isolated. Additionally, these peptides had UV spectra that were markedly different from the nonphotolabeled peptides. The modified peptide from GDH corresponded to through Lys289. Both sequencing and compositional analysis identified G ~ as uthe ~site ~of photoinsertion. ~ Sequencing of this peptide aborted at G ~ after u five ~ rounds ~ ~of analysis, indicating that insertion was blocking further progress. Compositional analysis showed that the entire sequence from residues 270 to 289 was present except that the single Glu residue was missing. This is interpreted as indicating that the photoinsertion is into the polypeptide backbone a t the Glu site. The peptide isolated from LDH corresponded to Asps2through Argw. Sequencing of this peptide could be completed throughout with only the round at Tyt83 giving no identifiable residue. Compositional analysis of this peptide was in agreement with the peptide from Aspa2to Argw with the exception that the single Tyr residue was missing. This indicates that the photoinsertion is into the tyrosine side chain. This data was found to be in agreement with X-ray crystallographic results identifying the NAD+-binding domains.
INTRODUCTION Conservation of a three-dimensional domain responsible for coenzyme binding has been demonstrated for a number of dehydrogenases by X-ray crystallography at a high resolution (1-4). A comparison of the structures of those dehydrogenases revealed that the coenzyme domain consists of several short stretches of parallel ,&pleated sheets flanked by helices. The NAD+-binding domain of GDH has been proposed by combination of secondary structure predictions and sequence comparisons with other dehydrogenases of known structure (5, 6). Crystallographic structure determination of GDH would be desirable to confirm the existence of a similar domain. To date the crystallization of mammalian GDHs in a form suitable for X-ray analysis has not been accomplished. However, the crystal structure of glutamate dehydrogenase from Clostridium symbiosum at 0.6-nm resolution has been reported (7). The similarity between the electron micrographs for the bovine enzyme and the X-ray structure of the bacterial enzyme support a similar three-dimensional structure of the two enzymes. This indicates that there may exist structural homology between NAD+-binding domains of these dehydrogenases, although a highresolution structure analysis would be necessary to confirm the relationship. In previous reports 2N3NAD+ was shown to be a valid active-site probe for bovine liver GDH and IL-2 (8,9).By using enzyme photolabeled with this probe it should be
* To whom correspondence should be addressed.
+ This work was supported by research grant GM-35766 from
the National Institutes of Health. t Abbreviations used: 2N3NAD+, nicotinamide 2-azidoadenosine dinucleotide; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis;GDH, glutamate dehydrogenase; LDH, lactate dehydrogenase; EDTA, ethylenediaminetetraacetic acid.
possible to isolate and identify the active site peptides of dehydrogenases and other proteins that interact with the adenine ring of NAD+. This could be used to confirm the general binding site assignments obtained from X-ray crystallography and support the contention that the crystallizedstructure is very similar to the solubilizedform. Also, it would verify the use of 2N3NAD+ as a general probe for the active site peptides that interact with the adenine ring of NAD+. A major problem encountered when trying to identify the peptides photolabeled with nucleotide photoaffinity probes is the lability of the photoinserted group (10).This is particularly true when photolabeled peptides are subjected to HPLC techniques where it is most often observed that the photoinserted radiolabel is lost on initial binding to the column and elutes in the void volume (11). Consequently, very few photolabeled peptides have been purified by HPLC that retain a convincing percentage of the initial photoinserted radiolabel to unequivocally allows their identification as the major photolabeled site in the protein from which they were derived. To resolve this problem our laboratory has concentrated on procedures that take advantage of the incorporated, highly negatively charged photoprobe by using such techniques as anionexclusion chromatography (11). HPLC techniques are avoided except in the final step to verify the purity of the isolated peptide(& In this study the availability of cishydroxyls on the two ribose moieties of the photoinserted NAD+ analogue indicated that the use of immobilized boronate column chromatography would aid in the isolation of the photolabeled peptides. This would be of considerable value since many of the photolabeled peptides our laboratory has worked with lose the photoinserted radiolabel when subjected to HPLC. The techniques described herein were very effective for isolating the photolabeled peptides in pure form using only one column procedure. Surprisingly, after purification on boronate columns, the 0 1991 American Chemical Society
Bioconlugate Chem., Vol. 2, No. 3, 1991 t43
Photoafflnity-Labeled ActivaSRe Peptides
majority of the photoinserted radiolabel remained attached during HPLC analysis. Also, since only the photolabeled peptide was isolated, the amino acid residue where photoinsertion occurred was easily identified. This technique should have good general application for investigating the NAD+ binding sites of several enzymes. EXPERIMENTAL PROCEDURES
Materials. Bovine liver glutamate dehydrogenase was obtained from Sigma as a solution in 50% glycerol and used without further purification. The protein concen= 8.9 (121, and tration was determined by using E2hm1% a molecular weight of 56 100 for the monomer was used in the calculations. Porcine muscle lactate dehydrogenase was purchased from Boehringer/Mannheim Corp. as a solution in 50% glycerol and stored at 4 "C. E z m l % = 14.5 was used for the protein concentration. TPCKtrypsin was purchased from Sigma. [(Dihydroxybory1)phenyl]succinamyl derivatives of aminoethyl polyacrylamide (Affi-gel 601) were obtained from Bio-Rad Laboratories. Phenyl boronate-agarose Matrex Gel PBA30 was from Amicon. 2-Azido-NAD+ and [32P]-2-azidoNAD+ were synthesized as previously described (8). Methods. Analytical Photoaffinity Labeling of LDH. Five micrograms of enzyme were photolyzed in 30 pL of the buffer 50 mM Tris-acetate, pH 7.0, in the presence of 50 mM oxalate. Samples were incubated at room temperature for 5 min, then placed on ice and irradiated for 3 min with a Spectroline medium-wave ultraviolet lamp (302 nm, 6.4 mW/cm2) a t a distance of 5.5 cm. The reaction was stopped by addition of a halfvolume of a protein-solubilizing mixture which consisted of 6 % SDS, 30 % glycerol, 0.2 M Tris-HC1, pH 6.8,1.5 % 2-mercaptoethanol, and 0.003 7% bromophenol blue. Aliquots of sample (2 pg of protein) were analyzed by 10% SDS-PAGE. The protein bands were cut from the gel and the radioactivity was measured by liquid scintillation counting as previously reported (8). Photoaffinity Labeling of GDH and Isolation of Photolabeled Peptide. Bovine liver GDH (1 mg/mL, 17.8 pM subunit) was incubated with 50 pM [32P]-2-azidoNAD+ on ice in the presence of 12 mM glutarate and 1 mM GTP in lOmM Tris-acetate buffer, pH 8.0, containing 10 pM EDTA. In each experiment, 10 mg of protein was photolabeled in a volume of 10 mL placed in a plastic weighing boat that gave a solution depth of less than 0.3 mm. After 5 min of incubation, the enzyme solution was irradiated for 3 min with the UV lamp as described above. Completion of photolysis of the photoprobe was determined by UV spectroscopy. The photoaffinity-labeling reaction was quenched by addition of dithiothreitol to a final concentration of 25 mM. The photolabeled protein was precipitated with ammonium sulfate (80%saturation) to remove excess unbound ligand. The resulting precipitate was desalted by gel filtration on Sephadex G-25 equilibrated with 0.1 M ammonium bicarbonate. Urea was added to 2 M to the protein sample followed by digestion with 2% TPCK-trypsin (w/w) at room temperature overnight. The tryptic digest was applied to a 3-mL Affi-gel 601 column, equilibrated with 0.1 M ammonium acetate buffer, pH 8.8,and washed with the same buffer to elute the unmodified peptides. When the UV absorbance reached a minimum, a pH gradient (8.85.5) of 200 mL of 0.1 M ammonium acetate was started to elute the photolabeled peptide retained on the column. The location of the photolabeled peptide was followed by monitoring the UV absorbance and the radioactivity profile of the eluant.
Photoaffinity Labeling of LDH and Isolation of Photolabeled Peptide. LDH (1 mg/mL, 28.6 pM subunit) was incubated with 50 pM [32P]-2-azido-NAD+in 100 mM Tris-acetate, pH 7.0, and 50 mM oxalate at room temperature for 5 min. The protein sample (10 mg) was placed on ice and photolyzed as above. After addition of dithiothreitol to 25 mM concentration, 100 pg of TPCKtrypsin was added to the protein solution and digestion was continued overnight at room temperature. The total digest was applied onto a 10-mL PBA-30 column, equilibrated in 0.1 M ammonium acetate, pH 8.8. After elution of photolabeled peptide with a pH gradient as described above, the radioactive pool was lyophilized and further analyzed by HPLC. Separationof PhotolabeledPeptides by HPLC. The pooled fractions which contained the radioactive photolabeled peptide were analyzed by reverse-phase HPLC (LKB, equipped with a diode-array spectral detector) with a Brownlee Lab Aquapore RP 300 C-8 column (4.6 mm X 25 cm) developed with a gradient of increasing acetonitrile concentration in 0.1 ?4 trifluoroacetic acid. Analysis of Isolated Peptides. The photolabeled peptides were characterized by analysis of amino acid composition on a Beckman 6300 high performance amino acid analyzer. The amino acid sequence of the peptides was determined on an Applied Biosystem 477A pulse liquid protein sequencer with on-line identification of phenylthiohydantoin amino acids. Each experiment reported in this paper has been repeated a minimum of three times. RESULTS
Purification of PhotolabeledPeptides of GDH. To minimize the extent of nonspecific photoinsertion and at the same time to optimize the specific labeling of the enzyme, 50 pM [32P]2N3NAD+was used, which is the concentration at which photoincorporation approaches saturation. In addition, 1mM GTP and 12 mM glutarate were included in the reaction mixture, since they were shown to increase the binding affinity of NAD+ and 2N3NAD+ to GDH (8). Under these conditions, 4040% of the enzyme incorporated photolabel as determined by both a decrease in enzyme activity and incorporation of radiolabel. In the previous report it was demonstrated that 2N3NAD+specificallyphotolabeled the coenzyme binding site of bovine liver glutamate dehydrogenase. To identify the site of photoinsertion, enzyme was digested with trypsin to produce peptides. Taking advantage of the existence of the two cis-hydroxyl groups on the ribose groups of the photoincorporated probe, the photolabeled peptide was isolated by using immobilized boronate column chromatography. The UV and radioactivity profiles of tryptic peptides from the boronate column are shown in Figure 1. All of the unlabeled peptides were eluted with ammonium acetate buffer, pH 8.8,in the void volume as determined by HPLC analysis (data not shown). However, photolabeled peptides were selectively retained on the column. These radioactive peptides were eluted with a pH gradient of 8.8-5.5 and subjected to reversephase HPLC (Figure 2). A major peak in both the UV and radioactivity profile at 33 min was obtained. The radioactive flow-through fraction contains unbound probe including any decomposition product of photoadduct produced as peptide binds to the HPLC column matrix. These flow-through fractions, covering 4-7 min, were subjected to analysis and no significant amounts of amino acids were detected. The major peak containing peptide and radiolabel also showed an unusual spectrum with an absorbance maximum a t 264 nm, presumably due to the
144
Bkconlugete Chem,, Vol. 2,
Kim and Heley
No. 3, l B B l
' 3
10
0
20
40
30
Fraction Number
Figure 1. Purification of photolabeled peptide from GDH on a boronate column. %tic digests of 10 mg of 2-azido-NAD+labeled GDH were applied to a 3-mL boronate column, equilibrated in 0.1 M ammoniumacetate buffer. Unmodified peptides were eluted with the same buffer and photolabeled peptide was eluted with a 200-mL gradient of pH 8.8-5.5 0.1 M ammonium acetate. Fractions (3mL) were monitored for absorbance at 220 nm and for radioactivity. 0.5
E $
0.4
-
0.3
-
0.2
-
0.1
.
6 "
h
B z x
f
L
E
200
220
240
260
280
300
320
Wavelength (nm)
Figure 3. UV absorption spectra of the purified photolabeled peptide from GDH (A) and a typical unmodified GDH tryptic peptide (B). Table I. Amino Acid Analysis of the Isolated GDH Peptides amino predeteramino predeteracid dicted mined0 acid dicted mined" ASX 4 4.34 (4)* Met 0 0.03 (0) Thr 0 0.07 (2) Ile 2 1.83 (2)
h
Ser Glx Pro GlY Ala Val
1/
3
0.0
I
0
2 1
2 3 1
2
1.78 (2) 0.22 (0) 2.41 (2) 3.02 (3) 1.00 (1) 1.85 (2)
Tyr Phe
Hie
Lys Arg Trp
0 0 0 1 0 1
0.04 (0) 0.21 (0) 0.00 (0) 0.98 (1) 0.04 (0) N/Dc
Normalized values. * Nearest integer. Not detected. Destroyed
during hydrolysis. 0
10
20
30
40
50
60
Time (min)
Figure 2. C-8 HPLC purification of the GDH peptide obtained after the boronate column chromatographystep. The volume of pooled peak fractionswas reducedto 1mL by evaporation.Sample was injected into a C-8 HPLC column and the profile was developedwith a linear gradient of solvent B (70% CHsCN) into solvent A (0.1 % trifluoroaceticacid)(10-60% B for 60min). The flow rate was 1.0 mL/min, and fractions were collected every 1.0 min and levels of a 2 p determined by liquid scintillation counting in aqueous phase. covalently attached NAD+ photoprobe. This is shown in Figure 3A, where it is compared to the spectrum of a typical unmodified GDH peptide with an absorbance maximum a t 276 nm (Figure 3B). The shift from the expected 280 nm maximum may be due to some solvent interference. Amino acid sequence analysis revealed that the peak fraction contained the amino acid sequence X-Val-AlaVal-Gly, which was identified as residues 270-274 of the protein as judged by comparison with the known amino acid sequence of bovine liver GDH (13). The sequence data gave the following picamole amounts of the indicated residue a t each cycle: 1; no detected residue (Cysexpected), 2; 2612 of Val, 3; 1939 of Ala, 4; 2043 of Val, 5; 112 of Gly, 6; 102 of Gly 7; 100 of Gly. After cycle 5 only Gly was seen and it slowly bled down to where at cycle 20 the amount of Gly was 40 pmol. Since these are tryptic digests, it was
expected to produce a sequence ending with Lys-. However, the peptide could not be sequenced beyond Gly274.The amino acid composition of the major peptide revealed that the peptide had a composition that was compatible with that of the tryptic peptide spanning residues 270-289 with the exception of Glu, where a significant reduction of recovery was seen (Table I). Trp and Cys could not be quantified because they are destroyed during acid hydrolysis. The amino acid composition data implied that the entire sequence from 270 to 289 was present but that sequencing aborts after Glyn4. This is supported by the rapid reduction of yields of phenylthiohydantoin amino acid residues observed from cycle 4 to 5 (2043 pmol to 112 pmol). Cycles 6-20 gave Gly a t about 100-40 pmol per cycle. No other residues were released. Thus, sequencing of the peptide was prevented beyond Gly274and this Gly residue bled off slowly presumably because modifications by 2NsNAD+ occurred in the vicinity of Gly274. The result from the amino acid analysis suggests G11.1275is the site of modification with the exact site of modification being in the peptide backbone between G1y2T4and G1u275. The side chain of G ~ is an u unlikely ~ ~ ~ place of modification because modification a t such a position should not abort sequencing of Gly2". Since the picamole drop from cycle 4 to 5 was about 95% it is also
8hn"uQate the" Vol. 2, No. 3, 1991 145
Photoaffinity-Labeled Active-She Peptkles 0.3 0.4
I
E,
0,
@l N
P)
0.3
0
X
0.2
g
0.1
0
50
100
150
NAD+ concentration &M)
Figure 4. Protection against photolabeling of LDH by 2-azido-NAD+in the presence of increasing concentrations of NAD+. The concentration of [aP]-2-azido-NAD+was 10 pM. Each photolabelingreaction was performed in 50 pL containing 50 mM HEPES, pH 7.0, buffer in the presence of 50 mM oxalate and 5 pg of enzyme. The photolabelingprocedure and methods of analysis are described in Materials and Methods.
.5
y
15
t '0
x
10
2
.>
%
.-8
5
LL 0
0 10
20
40
30
50
Tlme (mln)
unlikely that CysZ"3or TrpBl was also modified, and their absence is due to their known lability to the procedures used. Specific Photolabeling of LDH. To demonstrate specific active site labeling of LDH, the enzyme was photolabeled with [32P]2NsNAD+in the presence of increasing NAD+concentrations. The presence of the natural ligand should prevent the photoincorporation of 2N3NAD+ into the active site at physiologically relevant concentrations. It was found that the inclusion of oxalate, which serves as a cosubstrate analogue, greatly increases the affinity of LDH for NAD+ and 2NsNAD+ improves the efficiency of photolabeling of LDH. The results of the competition experiments are shown in Figure 4. The apparent Kd of NAD+ for LDH is estimated to be about 5 NM in the presence of 50 mM oxalate. Without oxalate, NAD+binds to muscle LDH with a Kd of about 600 pM (14). The efficiency of 2NsNAD+ photoinsertion into LDH was determined to be over 50 ?6 by using previously reported procedures (8). Purification of Photolabled Peptides of LDH. Similar procedures were followed to obtain the photolabeled peptide from LDH, as was done for GDH. The exception was that the ammonium sulfate precipitation step used to remove unbound radiolabel was omitted. Reverse-phase HPLC analysis demonstrated the purity of photolabeledpeptide obtained (Figure 5 ) . The increase of the radioactivity in the flow-through fractions was due to the presence of the excess unbound photoprobe that was not removed prior to the trypsin digestion step by the ammonium sulfate precipitation step. Amino acid analysis confirmed the absence of any peptide in the flow-through fractions. Again, the photolabeled peptide exhibits a unique UV spectrum with an absorbance maximum at 264 nm similar to the one observed in Figure 3A (Figure 6).
From amino acid sequence determination and comparison with the known amino acid sequence of pig muscle LDH (15),the isolated peptide was identifiedas a fragment from AspBzto Argm. However, at cycle 2, where the derivative of T f l was expected, no standard PTH derivative was observed. Cycles 3 through 7 gave Asn, Val, Thr, Ala, and Asn at 491-, 241-, 90-, 37-, and 10-pmol levels, respectively. Accordingly, T f l is presumed to be the residue which is covalently modified by 2NsNAD+. In this case the exact site of modification is probably the aromatic ring of the tyrosine rather than the peptide backbone since the sequencing did not stop at this residue
Figure 5. C-8 HPLC profile of LDH peptide obtained after a PBA-30column chromatography. The peak fractionselutedfrom PBA-30 column were applied to a C-8 reverse-phase column at a flow rate of 1.0 mL/min. The profile was developed with a linear gradient of solvent B to solvent A (0-10min, 0%B; 10-60 min, 040% B). Fractions were collected every 1.0 min and "P levels determined by liquid scintillation in the aqueous phase.
0.2
0.1 Q
u
O
f0
0.0
Ll
a
0.2
0.1
0.0 200
220
240
260
280
300
320
Wavelength (nm)
Figure 6. UV absorption spectra of the purified photolabled peptide from LDH (A) and a typical unmodified tryptic peptide (B) also from LDH.
as occurred at the Glu residue in GDH. In order to further demonstrate that T y P was the site of modification, the amino acid composition analysis was performed and the result is shown in Table 11. The absence of a tyrosine residue was evident in the analysis data while other residues were recovered nearly quantitatively. Thus, both the amino acid sequencing and composition data are compatible with the interpretation that T f l in porcine LDH is the site of covalent photoinsertion of 2NsNAD+. None of the derivatives recovered after gas-phase sequencing were detectably radioactive in either GDH or LDH. Most of the loaded radioactivity remained bound to the sequencing filter. The high polarity of azido-NAD+
146 B h n j u p t e Chem., Vol. 2, No. 3, 1991
Table 11. Amino Acid Analysis of the Isolated LDH Peptide predeteramino predeteramino acid dicted mined0 acid dicted mined0 Asx 3 3.14 (3)* Val 1 1.12 (1) Thr Met 0 0.17 (0) 1 0.91 (1) Ser 1 1.01 (1) Ile 0 0.05 (0) Glu 0 0.16 (0) Leu 0 0.06 (0) Pro Tyr 1 0.00 (0) 0 0.00 (0) Lys 0 0.04 (0) G~Y 0 0.34 (0) Ala Arg 1 1.00 (1) 1 l.OS(1) 0 Normalized values. * Nearest integer.
may prevent extraction of the modified derivatives from the filter by organic solvents. It has been observed that the PTH derivative of amino acid residues modified by photo-cross-linking would not be detected during gas-phase sequencing (16). DISCUSSION
NAD+ is an important nucleotide that is utilized by a wide variety of enzymes to meet the synthetic, regulatory, and energetic needs of a cell. The fluctuation of levels of such enzymes,their structures,cellular location, membrane sidedness, etc. is therefore of importance in many areas of cell research. With two important NAD+-utilizing enzymes the validation of 2NsNAD+ as an active site directed probe has been established. This probe can be used to detect NAD+-binding proteins in crude homogenates or to determine the peptide regions and amino acid residues located in the adenine-bindingdomain. The latter was accomplishedby using the following procedures. After trypsin digestion of GDH and LDH photoaffinity labeled with [32P]2N3NAD+,the modified peptides were isolated by using immobilized boronate column chromatography. The high specificity of the immobilized boronate ligand for cis-diol containing compounds has been utilized for chromatographic separations of a great number of biological substances. It has been shown that immobilized boronates can be used to fractionate adenine and pyridine nucleotides (17,18).They also have been successfully used to isolate nucleosidyl peptides (19-21). By using a pH gradient elution, a single homogeneous radiolabeled peptide suitable for amino acid sequence determination was obtained from each enzyme. The purified photolabeled peptides exhibited a highly unusual UV spectrum with a maximum absorption of 260-264 nm and a shoulder near 280 nm, which verifies that an adduct of the NAD+ photoprobe is still covalently attached to the peptide. The active site peptide of GDH identified in this work is located within the proposed NAD+-bindingdomain of bovine liver GDH spanning the residues 245-280 (5,221. A sequence ( C y ~ ~ ~ ~ - is L shown y s ~ ~to) be photolabeled by 2-azido-NAD+. Therefore, this sequence is expected to interact with the adenine ring of the coenzyme. A part of this sequencespanning from Cys270to Gluns is predicted to form a &pleated sheet (bB) which constitutes one of the six strands of parallel sheet found in the NAD+-binding domain. On the basis of information obtained on the amino acid sequence determination and composition analysis of the isolated labeled peptide, we propose that the attachment site of 2-azido-NAD+ is located at the peptide backbone between Gly274 and Glu276. Others have reported that the invariant, functional residue, Asp, is found in the bB region of four known dehydrogenases and is conservatively changed to Glu27sin GDH (5). This is based on the three-dimensional alignment of the common nucleotide-binding structure in dehydrogenases, on primary sequence homology. Crystal-
Kim and Haley
lographic studies suggest that Asp in LDH and other dehydrogenases is involved in forming a hydrogen bond with the 0-2’ hydroxyl of the adenine ribose and its potential to interact with the N-3 of adenine is also suggested (23-25). It is known that GDH utilizes either NAD+ or NADP+ as a substrate with almost equal efficiencies whereas LDH uses only NAD+. It is therefore unlikely that G ~ in uGDH ~ plays ~ ~a role in forming a hydrogen bond with the 0-2’ atom, as Asp does in LDH, because the carboxylate moiety of G ~ moving u ~close ~ to~ the 2’-phosphate would lead to charge repulsion. Therefore, G ~ is more u ~likely ~ to~interact with the adenine N-3 position and it may do this through the amide linkage region. This may imply an important distinction between the two classes of dehydrogenases represented by GDH and LDH, respectively. It is also conceivable that the nitrene generated by the 2-azido group effectively inserts into this residue at the amide linkage when it is positioned close to the N-3 atom. The presence of an additional NAD+ binding site for GDH has been suggested by a number of investigators (26-28). This second site has a low affinity for NAD+, is unaffected by cosubstrate addition, and is suspected to be a regulatory site. The possibility of photolabeling such a site was eliminated in this study because a t the concentration of 2N3NAD+ used, in the presence of cosubstrate, there is approximately one coenzyme molecule bound per subunit according to the equilibrium dialysis data and this is at the catalytic site (29). The validity of the results obtained with GDH was further supported by identification of the NAD+-binding site of porcine muscle LDH, whose X-ray structure has been determined a t a high resolution. The site of photolabeling identified was located within the well-characterized coenzyme binding domain Asn2“Cysla of dogfish LDH (23). The sequence photolabeled in porcine LDH, AspsLArgSO,is located at the connection between the two parallel strands of @-sheetstructure, bC and bD. Furthermore, the determined site of attachment, T e , is thought to be one of the amino acid residues lining the domain involved in adenosine binding. In porcine LDH TyrBcorresponds to Tyre of dogfish LDH. T f l of dogfish LDH is proposed to interact with the N-1 of adenine through a hydrogen bond (251,which places it adjacent to the azide at the 2-position. Therefore, it is not contrary to expectation that TyrB of porcine LDH was photolabeled instead of Asps3, which corresponds in threedimensional orientation to G ~ of uGDH, ~ considering ~ ~ the different roles in binding of Asp and Glu in these enzymes and the position of TyrB. The results with porcine LDH show that Tyres is the site of modification by 2NsNAD+ and this corresponding residue is proposed to be involved with hydrogen-bond formation with the N-1 position of adenine in dogfish LDH (25). It is therefore in very close proximity, being approximatelyonly one carbon-nitrogen bond length away. The situation with GDH is not quite as clear but does give some very interesting information. First, the GluZ7S that is modified corresponds to an Asp residue that is highly conserved in several other dehydrogenasesthat will not utilize NADP+ effectively, e.g. LDH, LADH, and GADPH. With these dehydrogenases Asp is proposed to be involved in hydrogen-bond formation with the 0-2’ atom of the ribose group which could not be occurring, due to charge repulsion, with NADP+. This probably accounts for the inability of NADP+ to be a substrate for these dehydrogenases whereas it is a good substrate for GDH. This concept is supported by the observation that
Photoafflnitylabeled Active-Stte Peptides
the equivalent position of G ~ in uNADP+ ~ ~specific ~ dihydrofolate reductase is replaced by a positively charged Arg residue. It is likely that the reason for the Glu replacement for Asp in GDH is to produce a different active site conformation that allows both NAD+ and NADP+ binding. The positioning of Gluns near the N-3 position of the adenine ring (7, 25) would place it approximatelyone carbon-nitrogen bond length away from the nitrene produced on 2NsNAD+. Apparently, the orientation of the binding site in GDH is such that efficient photoinsertion occurs into Glu at the peptide backbone. It is very likely that the photoinsertion site is at the nitrogen of the amide linkage since (i) compositionalanalysisshows totally intact Gly but no Glu residues (ii) and sequential analysis shows Gly being slowlyreleased with no Glu being released at all. The data obtained from the sequence analysis of the 2NsNAD+ photolabeled peptides support the selectivity of photoinsertion of this probe into GDH and LDH as validated by saturation and protection experiments (8). It also demonstrates that appropriately designed photoaffinity-labeling experiments combined with additional techniques, which take advantage of the properties of the photoinserted probe, will allow the isolation and sequencing of peptides in the binding domain and, at times, the determination of the exact amino acid residues involved. For NAD+-binding proteins it is suggested that the technique used in this research will be of general utility. Additionally, we found that 8N3GTP and 2N3ADP mimic the allosteric properties of GTP and ADP in GDH with regard to NAD+ binding. They also effectively photoinsert into GDH and are currently being used to determine the binding domains of GTP and ADP, respectively. LITERATURE CITED (1) Adams, M. J.,Ford, G. C., Koekoek,R.,Lentz,P. J., McPherson, A., Rossmann, M. G., Smiley, I. E., Schevitz, R. W., and Wonacott, A. J. (1970) Nature (London) 227, 1098-1103. (2) Hill, E., Tsernoglou, O., Webb, L., and Banaszak, L. J. (1972) J. Mol. Biol. 72, 577-591. ( 3 ) Eklund, H., Nordstrom, B., Zeppezauer, E., Soderlund, G., Ohlson, I., Boiwe,T., Soderberg,B. O.,Tapia, O., Branden, C. I., and Akeson, A. (1976) J. Mol. Biol. 102, 27-591. (4) Buehner, M., Ford, G. C., Moras, D., Olsen, K. W., and Rossmann, M. G. (1974) J. Mol. Biol. 90,15-49.
Bloconlugate Chem., Vol. 2, No. 3, 1991
147
(5) Rossmann,M. G., Moras,D., and Olsen, K. W. (1974)Nature (London) 250,194-199. (6) Wo otton, J. C. (1974) Nature (London) 252, 542-546. (7) Rice, D. W., Baker, P. J., Farrants, G. W., and Hornby, D. P. (1987) Biochem. J. 242, 789-795. (8) Kim, H., and Haley, B. E. (1990) J.Biol. Chem. 265,36363641. (9) Campbell, S., Kim, H., Doukas, M., and Haley, B. (1990) h o c . Natl. Acad. Sci. U.S.A. 87, 1243-1246. (10) Lewis, C. T., Haley, B. E., and Carlson, G. M. (1989) Biochemistry 28, 9248-9255. (11) King, S. M., Kim, H., and Haley, B. (1991) Methods Enzymol. 196, 449-466. (12) Egan, R. R., and Dalziel, K. (1971) Biochim. Biophys. Acta 250,47-50. (13) Moon, K., Piszkiewicz, D., and Smith, E. L. (1972) h o c . Natl. Acad. Sci. U.S.A. 69, 1380. (14) Stinson, R. A., and Holbrook, J. J. (1973) Biochem. J.131, 719-728. (15) Kiltz, H. H., Keil, W., Griesbach, M., Petry, K., and Meyer, H. (1977) Hoppe-Seyer’s 2.Physiol. Chem. 358,123-127. (16) Knight, K. L., and McEntee, K. (1985) J. Biol. Chem. 260, 10185-10191. (17) Gonzalez, R. A., Salinas, E. L., Jacobson, E. L., and Jacobson, M.K. (1983) Anal. Biochem. 135,69-77. (18) Moore, E. C., Peterson, D., Yang, L. Y., Yeung, C. Y., and Neff, N. F. (1974) Biochemistry 13, 2904-2908. (19) Schmidt, J. A., and Coleman, R. F. (1984) J. Biol. Chem. 259,14615. (20) Bates, S. P., and Coleman, R. F. (1986) Biochemistry 25, 3508. (21) Bailey, J. M., and Coleman, R. F. (1987) J.Biol. Chem.262, 12620. (22) Wooton, J. C. (1974) Nature (London) 252, 542-546. (23) Rossmann, M. G., Liljas, A., Branden, C.-I., and Banaszak, L. J. (1975) The Enzymes, 3rd ed., Vol. XI, pp 61-102, Academic Press, New York. (24) Grau, U. M. (1982) The Pyridine Nucleotide Coenzymes, pp 135-187, Academic Press, New York. (25) Holbrook, J. J., Liljas, A., Steindel, S. J.,and Rossmann, M. G. (1975) The Enzymes, 3rd ed., Vol. XI, pp 191-292, Academic Press, New York. (26) Bayley, P. M., and O’Neil, K. T. (1980) Eur. J. Biochem. 112,521-531. (27) Smith, T., and Bell, J. E. (1982) Biochemistry 21,733-737. (28) Banjeree, A., Levy, H. R., Levy, G. C., LiMuti, C., Golstein, B. M., and Bell, J. E. (1987) Biochemistry 26, 8443-8450. (29) Dalziel, K., and Egan, R. R. (1972) Biochem. J. 126,975984.
