Biochimica ~,~ et Biophysica AFta ELSEVIER
Biochimica et BiophysicaActa 1252 (1995) 203-208
Arginine-239 in the beta subunit is at or near the active site of bovine pyruvate dehydrogenase Devayani Eswaran a, M. Showkat Ali a,b, Bhami C. Shenoy a, Lioubov G Korotchkina Thomas E. Roche c, Mulchand S. Patel a,b,*
a Departme,~t of Biochemisto', Case Western Reserve University School of Medicine, Cleveland, OH 44106, USA b Department of Bioc.~emisto, State Universio, of New York at Buffalo, 140 Farber Hall, 3435 Main Street, Buffalo, NY 14214, USA c Department of Biochemistry, Kansas State UniversiO', Manhattan, KS 66506, USA
Received 21 December 1994; revised 8 May 1995; accepted 23 May 1995
We have modified bovine, pyruvate dehydrogenase (El), the first catalytic component of the pyruvate dehydrogenase complex, with pyreneglyoxal. Treatment of E l with pyreneglyoxal resulted in the loss of enzyme activity. Pyruvate plus thiamin pyrophosphate (TPP) afforded approximately 80% protection against this inactivation and protected two arginine residues per mol of E 1 tetramer (c~2/32) from modification. Circular dichroism spectral analysis indicated absence of any gross structural changes in the enzyme as a result of modification. Comparison of the peptide maps, monitored at 345 nm of unprotected and pyruvate plus TPP protected Els after V8 digestion revealed that a peptide in the protected enzyme was labeled by pyreneglyoxal to a lesser extent than its counterpart in the unprotected enzyme. Sequence analysis of the peptide demonstrated that it corresponded precisely to amino-acid residues 235 to 246 in the human E1/3 sequence, with arginine residues at positions 239 and 242. Since Arg-239 is conserved in the /3-subunit of all presently known sequences of the pyruvate dehydrogenase complex and branched-chain a-keto acid dehydrogenase complex, it is strongly suggested that Arg-239 in the human E1/3 sequence is at or near the active site of bovine El. Keywords: Pyruvate dehydrogenasecomplex; a-Keto acid dehydrogenasecomplex; Active site; Chemical modification;Arginine
and regulation of the PDC. It catalyzes the following two partial reactions in the total PDC reaction sequence.
The pyruvate dehydrogenase complex (PDC) catalyzes the oxidative decarboxylation of pyruvate forming acetylCoA and CO 2. Mammalian PDC is composed of multiple copies of three catalytic components [pyruvate dehydrogenase (El), dihydrolipoyl acetyltransferase and dihydrolipoamide dehydrogenase], two regulatory components [El kinase and phosphoE1 phosphatase] and protein X [1,2]. The E1 component plays a major role in the functioning
CH3COCOOH + E1-TPP
Abbreviations: El, Pyruvate dehydrogenase;PDC, pyruvate dehydrogenase complex; TPP, thiamin pyrophosphate; HPLC, high-performance liquid chromatography; CD, circular dichroism; BCKADC, Branchedchain a-keto acid dehydrogenase complex; HETPP, 2-(a-hydroxyethyl)thiamin pyrophosphate. * Tel: + 1 (716) 8292727; fax: + 1 (716) 8292725. 0167-4838/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0167-4838(95)001 19-0
C H 3 C ( O H ) =TPP-E1 + CO 2
C H 3 C ( O H ) = TPP-E1 + [lipS 2 ]-E 2 -~ [CH3CO S lipSH] E 2 q- E1-TPP
El, an c~2/32 tetramer, has a molecular mass of about 154 kDa. Very little is known about the structure of the active site(s) of this important enzyme component. Chemical modification studies have shown that cysteine , histidine , arginine , tryptophan  and lysine  residues are essential for E1 activity. The specific localization of the essential cysteine (Cys-62El a )  and tryptophan (Trp-135E1/3)  residues have recently been identified. Previously it has been reported that arginine residues play an important role in the binding of pyruvate to the
D. Eswaran et al. / Biochimica et Biophysica Acta 1252 (1995) 203-208
pigeon breast holo-El . However, the identification of this residue(s) has not yet been established. We undertook a chemical modification study to identify the exact location of the arginine residue(s) in bovine El. Using pyreneglyoxal [10,11], a fluorescent analog of phenylglyoxal, and pyruvate and thiamin pyrophosphate (TPP) as protective agents, it was possible to specifically label the arginine residue which caused the loss of enzyme activity upon reaction. Substrate and cofactor protection studies revealed that arginine residue(s) is required for the binding of pyruvate to bovine holo-El and hence at least one of these residues is present at or near the active site(s). The peptide that contained the protected arginine residue was isolated by reverse-phase high-performance liquid chromatography (HPLC) and sequenced. We have suggested this residue to be Arg-239 in human E1 /3.
2. Materials and methods
p-Nitrophenylglyoxal, L-arginine, pyruvic acid and V-8 proteinase were purchased from Pierce. TPP was obtained from ICN. Pyreneglyoxal was from Molecular Probes. HPLC grade acetonitrile was purchased from Fisher. 2(~-Hydroxyethyl)thiamin pyrophosphate (HETPP) was synthesized as described by Holzer et al. .
on a Jasco Model J-600 Spectropolarimeter calibrated with d-10 camphor sulfonic acid. Measurements were made in a quartz cell with a path length of 1 mm. E1 (0.1 m g / m l ) in 50 mM potassium phosphate buffer (pH 7.5) was treated with 3 mM phenylglyoxal for 30 min at 25°C. The reaction was stopped by adding excess arginine and the samples were desalted on G-25 Sephadex columns. The recorded absorbance was corrected for solvent and background noise using an automated computer program. 2.5. Proteinase digestion
E1 (6.5 nmol) samples were modified with 0.2 mM pyreneglyoxal for 10 min at 25°C both in the presence and absence of pyruvate (0.1 mM) plus TPP (0.1 mM). The reactions were terminated by the addition of a 10-fold excess of arginine and reaction mixtures were then desalted by passing it consecutively over two G-25 Sephadex columns. The enzyme preparations in 50 mM potassium phosphate buffer (pH 7.5) were then digested with V8 protease (1:5 w / w proteinase/protein) for 24 h at room temperature in the presence of 2 M guanidinium chloride, the digestion terminated by the addition of trifluoroacetic acid, the sample filtered through Millipore Ultrafree MC Centrifugation filters and injected onto a HPLC column. All reactions were carried out in the dark. 2.6. R e v e r s e - p h a s e H P L C pyreneglyoxal-modified peptides
2.2. Assay of overall activity of the reconstituted PDC
Bovine kidney PDC was purified as described previously by Roche and Cate  and its components were resolved according to Linn et al. . The overall activity of the PDC was determined by reconstituting E1 with E2-X and E3 according to the method of Roche and Reed [ 15]. Protein was quantitated using Bio-Rad protein reagent and bovine serum albumin as the standard. 2.3. Modification of E1
E1 (78 pmol) in 50 /~1 of 20 mM potassium phosphate buffer (pH 7.5) was incubated with varying concentrations of pyreneglyoxal in the dark for 30-40 min at 25°C. Aliquots (5 /zl) were removed at different time intervals and immediately quenched with excess arginine (10 mM) in the assay solution to measure E1 activity as described above. In protection experiments, E1 was preincubated with TPP (0.1 mM), pyruvate (0.1 mM), pyruvate plus TPP, and HETPP (0.1 raM), respectively prior to modification with pyreneglyoxal.
The pyreneglyoxal labeled peptides generated by V-8 proteinase digestion were separated by a Shimadzu HPLC system (Model LC 600) equipped with a Synchropack C-4 reverse-phase column (4.6 × 25 mm) equilibrated with 0.1% trifluoroacetic acid in triply distilled water. The peptides were eluted in a stepwise gradient of increasing concentrations of acetonitrile containing 0.1% trifluoroacetic acid developed over 4 h at a flow rate of 0.9 ml/min. Absorbance was monitored simultaneously at 220 nm (for peptide absorption) and 345 nm (for pyreneglyoxal absorption). The pyreneglyoxal labeled peptides of interest were collected and dried in a Speed Vac Concentrator. The dried material was redissoived in 50 ~1 of 50% ( v / v ) acetonitrile and the peptides were sequenced using an Applied Biosystems Model 470A/120A automatic gasphase sequencer with an on-line phenylthiohydantoin analyser .
3. Results and discussion
3.1. Modification of arginine residues in E1 2.4. Circular dichroism
The circular dichroism (CD) spectra between 190-260 nm of native and phenylglyoxal modified El were recorded
Phenylglyoxal [ 17] and its analogs pyreneglyoxal [ 10,11 ] and p-nitrophenylglyoxal  are known to modify arginine residues in proteins. Incubation of bovine E1 with
D. Eswaran et al. / Biochimica et Biophysica Acta 1252 (1995) 203-208 120. 100
2o 0 0
~--. . . . . . 24
Fig. 1. Time-courseof inactivation of El by pyreneglyoxal.E1 (78 pmol) in 20 mM potassium phosphate buffer (pH 7.5) was incubated with different concentrations of pyreneglyoxal (0 to 200 /zM) for 30 min at 25°C. At the times indicated aliquots were removed for each concentration of pyreneglyoxal for measurement of residual activity. The concentrations of pyreneglyoxalwere (11) 0/xM; (0) 20/xM; (O) 60/xM and (zx) 200 /xM. increasing concentrations of pyreneglyoxal over a 40 min period resulted in inactivation of E 1 in a concentration and time dependent manner. A typical time-course of inactivation using fixed concentrations of pyreneglyoxal is shown in Fig. 1. Nearly complete loss of El activity was achieved with 0.2 mM pyreneglyoxal in 20 min while under similar conditions untreated E1 had no loss of enzyme activity. 3.2. Substrate protection c f E1 with pyreneglyoxal
Pyruvate, TPP, pyruwtte plus TPP or HETPP were tested for their ability to protect the enzyme against inactivation by pyreneglyoxal ([Zig. 2). Pyruvate (0.1 mM) alone afforded no protection against inactivation by pyreneglyoxal. TPP (0.1 mM) afforded about 20% protection against inactivation while pyruvate (0.1 mM) plus TPP (0.1 mM) protected the enzyme from inactivation by pyreneglyoxal by approx. 80%. Less than 10% inactivation was noted during modification when the enzyme was pretreated with HETPP (Fig. 2). On the basis of the protective effect of pyruvate plus TPP  and the effects of different pyruvate analogs (hydroxypyruvate, bromopyruvate and pyruvamide) in combination with TPP on inactivation of pigeon breast El, Severin et al.  suggested that the positively charged arginine residue in El might be involved in the interaction with anionic carboxyl of pyruvate and hence might be localized in the substrate binding site. As pyruvate protected the enzyme only in the presence of TPP, it was also suggested that the substrate binding site is formed only upon holoenzyme formation, i.e., E1-TPP . Our results on bovine ]31 are in close agreement with those observed with pigeon breast muscle El suggesting a similar role to the arginine residue(s) in bovine El. We did find that TPP protected the bovine enzyme from inactivation by about 20% while there was no protective effect
seen in the case of E1 from pigeon breast muscle. This 20% protection against inactivation can be interpreted by reasoning that the critical arginine residue(s) may be present very close to the TPP binding site and hence may only be partially exposed to the inhibitor. However, in marked contrast to pigeon breast El, HETPP afforded about 90% protection against inactivation of bovine E 1 (Fig. 2) showing that essential arginine residue is protected by this substrate. The reason for these conflicting findings with E1 proteins is not clear. HETPP is thought to be rapidly deprotonated to form hydroxyethylidene ( H E = T P P ) intermediate which is tightly bound to El. One possibility is the arginine side chain helps to stabilize the carbanion from the H E = T P P intermediate. Alternatively, the 2-carbon fragment in HETPP may interact with the substrate-binding site and hence afford protection from inactivation. The difference between the bovine and pigeon E1 may be due to a differential influence of the amino acids surrounding the critical arginine residue. These hypotheses cannot be tested because the primary amino-acid sequences of bovine and pigeon E ls are not known. 3.3. Determination o f the number o f critical arginine residues in E1
The number of arginine residues modified during El inactivation was determined using p-nitrophenyglyoxal by the method of Yamasaki et al.  instead of pyreneglyoxal, as the compound itself is highly colored, pNitrophenylglyoxal (0.2 mM) also produced 95% inactivation of El, and a combination of TPP plus pyruvate afforded about 77% protection against this inactivation
0 ...................... .......................
............................. 0 ............... 0
i 4o '
Fig. 2. Substrate and cofactor protection of El from inactivation by pyreneglyo×al. E~ (78 pmol) in 20 mM potassium phosphate buffer was preincubated in the presence of pyruvate, TPP and pyruvate plus TPP for 10 min at 37°C and then treated with 0.2 mM pyreneglyoxal. The treatments were: () El alone; (0) E1 treated with 0.2 mM pyreneglyoxal; (A) E1 preincubated with 0.1 mM HETPP and then treated with pyreneglyoxal;(©) El preincubated with 0.1 mM pyruvate plus 0.1 mM TPP and then treated with pyreneglyoxal;(A) El preincubated with 0. l mM TPP and then treated with pyreneglyoxal, and (Ill) E1 preincubated with O.1 mM pyruvate and then treated with pyreneglyoxal.
D. Eswaran et al. / Biochimica et Biophysica Acta 1252 (1995) 203-208
Table 1 Modification of arginine residues of E1 with p-nitrophenylglyoxal in the presence and absence of pyruvate plus TPP Treatment
Arg modified/El Activity (mol/mol) (%)
El 0 El + p-nitrophenylglyoxal 2.5 E1 + pyruvate + TPP + p-nitrophenylglyoxal 0.5
100 5 77
Bovine El was preincubated with pyruvate (0.1 mM) plus TPP (0.1 mM) in 20 mM potassium phosphate buffer (pH 7.5) for 10 min prior to incubation with 0.2 mM p-nitrophenylglyoxal for 20 min. The details for determination of the modified arginine residues and enzyme activity are described in Section 2.
(Table 1). It was found that 2.5 arginine residues were modified per mol of unprotected El tetramer after 30 min, whereas 0.5 mol of arginine residue was modified in E1 pretreated with TPP plus pyruvate. These results indicate that modification of two arginines per a2/32 tetramer causes loss of activity and these arginines are specifically protected by pyruvate plus TPP. These observations corroborate the results reported by Nemerya et al.  on the modification of two arginine residues in pigeon breast El.
HPLC and the eluent was monitored at both 220 and 345 nm. Typical chromatograms at 345 nm for peptides from both unprotected and substrate-protected enzymes are shown in Fig. 3. This chromatogram detected only the peptides containing the arginine residue(s) modified by pyreneglyoxal. In three separate experiments, the chromatograms showed that 3 peptides were labeled by pyreneglyoxal in both protected and unprotected El s. These peaks are designated 1, 2, and 3 in the chromatograms. Of these, peak 1 was found to be that of the residual pyreneglyoxal-arginine adduct. This was confirmed by passing free arginine modified by pyreneglyoxal on a reverse-phase HPLC under identical conditions (results not shown). Moreover, peak 1 did not have a corresponding peptide in the 220 nm chromatogram. Peaks 2 and 3 were consistently differentially labeled by pyreneglyoxal. These peptides were 2- to 3-times more heavily labeled when El was treated with pyreneglyoxal in the absence of TPP plus pyruvate versus in the presence of these compounds. Also, these peaks had characteristic peptide absorbance at 220 nm. This indicated that these peptides contained modified arginine residue(s). Since these peptides were still labeled to some extent in
3.4. Inactivation caused by phenylglyoxal is not due to gross structural changes in E1 0.20
The CD spectra of native and phenylglyoxal-modified Els, monitored from 190 to 260 nm, were analyzed to determine whether the reaction of arginine residue(s) with phenyglyoxal resulted in gross changes in the secondary structure of the enzyme. In this experiment phenyglyoxal (3 mM) was used instead of pyreneglyoxal because pyreneglyoxal had high absorbance in the 190-260 nm wavelength and interfered significantly with the CD spectra of El. The CD spectra of the unmodified and phenylglyoxal-modified E ls were identical (results not shown) indicating that the overall secondary structure of the phenylglyoxal modified enzyme remained unchanged, and hence the loss of activity during phenylglyoxal treatment was not correlated to changes in secondary structure. Therefore, it is concluded that the inactivation of E1 upon reaction with phenylglyoxal (and may be extended to pyreneglyoxal) is due to the modification of essential arginine residue(s) at or near the active site rather than to modification-induced gross structural changes. 3.5. Identification of the essential arginine residue(s) at or near the active site
To locate the essential arginine residue(s) in bovine El, differential peptide mapping was employed after pyreneglyoxal modification of the enzyme both in the presence and absence of pyruvate plus TPP. The peptides generated by V8 proteinase digestion were separated by reverse-phase
0.16 0.12 e~
~ ~ [ ~
0.08 0.04 0.00 0
T i m e (rain)
Fig. 3. Reverse-phase HPLC chromatograms of the labeled El peptides generated by V8 proteinase digestion. Absorbance was monitored at 345 nm to identify the pyreneglyoxal-labeled peptides. The figure shows the chromatograms of El modified with 0.2 mM pyreneglyoxal (A) and El modified with pyreneglyoxal (0.2 mM) in the presence of pyruvate (0.1 rnM) plus TPP (0.1 mM) (B). The three peaks obtained at this wavelength have been numbered as 1, 2 and 3. For both the unprotected and the protected holo-E1 proteins, the peptides were eluted in a stepwise gradient over 4 h of increasing concentrations of 0.1% trifluoroacetic acid in acetonitrile.
D. Eswaran et al. / Biochimica et Biophysica Acta 1252 (1995) 203-208
the substrate protected holo-enzyme, either these peptides also contained arginine residue(s) that were not protected by substrate and cofactor or that treatment only afforded partial protection. Peptides 2 and 3 were collected and processed for sequence determination. It was found that peptide 3 was 16 amino-acid residues in length with a sequence V Q N M R T I R P M D M E T I E, while peptide 2 was a smaller peptide containing 11 amino-acid residues with the sequence V Q N M R T I R P M D, demonstrating it was a part of peptide 3. Thus, this peptide was a result of incomplete proteinase cleavage. Based on the sequence of human El, not only the internal cleavage site but the terminal sites of cleavage for the larger peptide were as predicted for V8 cleavage sites.
3.6. Sequence comparison of peptides 2 and 3 with sequences f r o m other species Both peptides 2 and 3 of bovine E1 contained 2 arginine residues. The sequence of bovine E1 a or/3 has not been determined yet, therefore the sequence of peptide 3 was compared to the deduced sequences of human E1 a and /3. The results indicated that peptide 3 spanned amino-acids between 235 and 250 in human E1/3 (Fig. 4). The two arginine residues in peptide 3 corresponded to Arg-239 and Arg-242 in the human E1/3 sequence. Since it was seen earlier that two arginine residues per tetramer (~2/32) were protected by pyruvate and TPP, and since peptides 2 and 3 contained arginine residue(s) that were not protected by substrate plus cofactor, a likely prospect
PDC (Peptide 3)
PDC PDC PDC PDC PDC BCKADC BCKADC BCKADC PO PD TK TK
Human Rat Pig Nematode Bacillus stearothermophillus Human Bovine P s e u d o m o n a s putida Escherichia coil Zymomonas mobilis Saccharomyces cerevisae Human
is that either Arg-239 or Arg-242 is the essential arginine residue but not both. As adducts of pyreneglyoxal plus arginine undergo decomposition during sequencing , it was difficult to identify which of the two arginine residues was labeled after protection with pyruvate plus TPP. To identify the essential arginine residue the sequence of peptides 2 and 3 were compared to those of the E1 component of the PDC and branched-chain a-keto acid dehydrogenase complex (BCKADC) isolated from several species. BCKADCs are analogous to PDC in terms of having a tetrameric ( a 2 f12) structure for their respective El components. Obviously, the difference between PDC and BCKADC lies in their substrates specificity. There is considerable sequence homology among the E1 sequences of PDC and BCKADC from various species . The similarity of the reaction mechanism, sequences, substrates and cofactors of PDC and BCKADC indicate that their active sites might be homologous. Fig. 4 shows the comparison of the sequence of Peptide 3 with the amino-acid sequences from PDC isolated from human, rat, pig, nematode and Bacillus stearothermophillus, and with those from BCKADC isolated from human, bovine and Pseudomonas putida species . It was found that Arg-239 is conserved in the sequence of E1 fl from both PDC and BCKADC from all species while Arg-242 was not. When this sequence was compared with the sequences of other TPP-dependent enzymes, it was observed that Arg-239 is conserved in pyruvate oxidase in Escherichia coli and replaced by lysine residue in transketolase from Saccharomyces cerevisiae (Fig. 4). In contrast, Arg-242 is not conserved in
Sequence 239 242 V Q N M [~ T I R P M D M E T I
V V V V
I I I I
N N N N
M M M L
~] ~ ~ ~
T T T C
I I I V
R R R R
P P P P
M M M L
D D D D
M M M F
E I E Q
T E T T
I I I V
E E E K
V V V V K V
V I I I S I
D D D D T L
L L L L L F
~ ~ ~ ~ ~ V
T T T S A G
V I I L L D
Q I L W L G
P P P P L
L W W L L Q
D D D D V L
I V V L E T
E D D D E V
T T T T K Q
I I V I A E
I C C V D I
L S P D K T C H N W K V A V A L V L D P F T I K P L D M E T I E
Fig. 4. Comparisonof the sequence of Peptide 3 with the sequences of El/3 of PDC, BCKADCand other TPP-dependentenzymes(PO, pyruvateoxidase; PD, pyruvate decarboxylaseand TK, transketolase) from several species. These sequence comparisionswere obtained using the Bestfit program on the GCG softwarepackage.
D. Eswaran et aL / Biochimica et Biophysica Acta 1252 (1995) 203-208
these two enzymes. Additionally, there is no conservation of amino-acid residues at Arg-239 and Arg-242 in human transketolase and Zymomonas mobilis pyruvate decarboxylase (Fig. 4). We therefore strongly suggest that the arginine residue corresponding to Arg-239 in the human E1/3 is the essential arginine in bovine El. The crystallographic studies on three TPP-requiring enzymes, namely transketolase, pyruvate oxidase and pyruvate decarboxylase showed that TPP is bound between two subunits so that pyrophosphate binding domain belongs to one subunit and aminopyrimidine domain belongs to another subunit . Using the X-ray crystal structure of Saccharomyces cerevesiae transketolase as template, Robinson and Chun  suggested that a hydrophobic region of the /3 subunit of the E1 (c~2/32) enzymes is likely to contain a binding site for the thiazolium ring of TPP and to provide a key catalytic histidine residue for the active site. From chemical modification studies, a specific cysteine residue in El ot  and a tryptophan residue in El/3  have been identified at or near the active site of mammalian El. Based on the sequence similarities between El/3 subunit of the branched-chain a-keto acid dehydrogenase and transketolase from several species, Zhao et al.  proposed that the El/3 subunit may participate in TPP-binding. The present study further supports the role of El/3 subunit in El catalysis. Thus, the amino-acid residues from both the subunits (oe and /3) constitute the active-site structure of the enzyme.
Acknowledgements This work is supported by the US Public Health Service Grant DK20478 (to M.S.P.) and DK18320 (to T.E.R.). We are thankful to Drs. Joyce E. Jentoft and Ganesh K. Kumar for helpful discussions and critical reading of the manuscript. We are indebted to Gary A. Radke for preparation of bovine pyruvate dehydrogenase complex and resolution of this complex.
References  Reed, L.J. (1974) Acc. Chem. Res. 7, 40-46.  Patel, M.S. and Roche, T.E. (1990) FASEB J. 4, 3224-3233.  Khailova, L.S., Aleksandrovich, O.V. and Severin, S.E. (1983) Biochem. InL 7, 223-233.  Khailova, L.S., Korochkina, L.G. and Severin, S.E. (1989) Ann. N.Y. Acad. Sci. 573, 36-54.  Nemerya, N.S., Khailova, L.S. and Severin, S.E. (1984) Biochem. Int. 8, 369-376.  Korochkina, L.G., Khailova, L.S. and Severin, S.E. (1984) Biochem. Int. 9, 491-499.  Stepp, L.R. and Reed, L.J. (1985) Biochemistry 24, 7187-7191.  Ali, M.S., Roche, T.E. and Patel, M.S. (1993) J. Biol. Chem. 268, 22353-22356.  Ali, M.S., Shenoy, B.C., Eswaran, D., Andersson, L.A., Roche, T.E. and Patel, M.S. (1995) J. Biol. Chem. 270, 4570-4574.  Hiratsuka, T. (1987)J. Biochem. 101, 1457-1462.  Tyagi, S.C. and Simon, S.R. (1990) Biochemistry, 29, 9970-9977.  Holzer, H., Goedde, C.M. and Ulrich, B. (1961) Biochem. Biophys. Res. Commun. 5, 447-451.  Roche, T.E. and Cate, R.L. (1977) Arch. Biochem. Biophys. 183, 664-677.  Linn, T.C., Pelley, J.W., Pettit, F.H., Hucho, F., Randall, D.D. and Reed, LJ. (1972)Arch. Biochem. Biophys. 148, 327-342.  Roche, T.E. and Reed, L.J. (1972) Biochem. Biophys. Res. Commun. 48, 840-846.  Hunkapiller, M.W., Hewick, R.M., Dryer, W.J. and Hood, L.E. (1983) Methods Enzymol. 91,399-413.  Means, G.E. and Feeney, R.E. (1973) Chemical Modification of Proteins, Holdes-Day, San Franscisco, CA, pp. 86-87.  Yamasaki, R.B., Shimer, D.A. and Feeney, R.E. (1981) Anal. Biochem. 111,220-226.  Severin, S.E., Khailova, L.S. and Gomazkova, V.S. (1986) Adv. Enzyme Regul. 25, 347-375.  Haining, R.L. and McFadden, B.A. (1990) J. Biol. Chem. 265, 5434-5439.  Wexler, I.D., Hemalatha, S.G. and Patel, M.S. (1991) FEBS Lett. 282, 209-213.  Muller, Y.A., Lindqvist, Y., Furey, W.W., Schulz, G.E., Jordan, F. and Schneider, G. (1993) Structure 1, 95-103.  Robinson, B.H. and Chun, K. (1993) FEBS Lett. 328, 99-102.  Zhao, Y., Kuntz, MJ., Harris, R.A. and Crabb, D.W. (1992) Biochim. Biophys. Acta 1132, 207-210.