ANALYTICAL
109, 32-40
BIOCHEMISTRY
Modification
(1980)
of Available Arginine Residues p-Hydroxyphenylglyoxa11~2
R. BRYAN YAMASAKI, of Food
Drpurtmrnt
Scier1c.r
ANTONIO
und Trt~hnolopy,
p-Hydroxyphenylglyoxal which
reacts
can
be
quantitated
mild conditions. pH of N”-citraconyl-L-arginine of other determined
7-9
at pH previously
phenylglyoxal
and
the
work
losses
in
arginines
was
supported
in
part
Consejo Spain.
de Investigaciones constitutes part
Superior material
Slope Biochemical Reno. Nevada. the
whom
by
National
26031 and HL 18619. by a grant from the Cientificas. of the thesia
completion
Conference, of
the
June
present
correspondence
should
0003.2697/80/170032-09$02.00/O Cupynght All right,
up with
reagent were
determined
those of containing
to give
reaction
9.5616
a single
prod-
place
under
takes
to complete less than 5%,
of modification coefficient of
of excess arginines, were
of
25-37.
research.
a
publication (I) appeared describing the USC‘ of 4hydroxy-3-nitrophenylglyoxal as a chromophoric reagent for arginine residues in one enzyme, creatine kinase. Advantages and disadvantages of this reagent are delineated in that article. (( To
extent
with but
R.B.Y. submitted to the Graduate Division of the University of California at Davis in partial fulfillment of the requirements for a Ph.D. degree in Biochemistry. Portions of this work were presented at the
Z After
Ctrl~wnicr
1.83
modification modification
in a protein x 10’ Mm'
can cm-’
be for
by gel filtration. Several modified by p-hydroxyspectrophotometrically.
previous investigators. an essential lysine.
was
Rhea rela-
unaffected.
Grants AM supported
Pacific 1979.
conditions 60 min
removal essential
close agreement without arginines
Institutes of Health A.V. was financially This
in proteins The
these within
The chemical modification of the guanidyl group of arginine has proved valuable in understanding structure-function relationships in proteins. Through chemical modification techniques it was shown that arginine residues. largely protonated at physiological pH, serve as recognition sites for anionic ligands in many proteins (2.3). When using such techniques, however. it is
1 This
residues
side chains. The the molar absorption
9.0 and 25°C following shown to have
Davis.
3 1. 1980
arginine
and 25°C. Under was obtained
These results were in ovomucoid. a glycoprotein tively
with
oj’C~~/iJ~wnitr,
spectrophotometrically.
common amino acid at 340 nm using
the product proteins,
March
by
E. FEENEY’~
AND ROBERT
University
Received
uct
VEGA.
in Proteins
‘$ I’)80 by Academc Pre\\. Inc. of rcnmductwn m anv form rewrvcml
be addressed.
important to know the extent of modification, especially when the essentiality of a residue is being assessed. Although several reagents have been developed for the modification of arginine residues in proteins, only a limited number allow the direct determination of the extent of modification. The well known Sakaguchi reaction. employing a-naphthol and sodium hypochlorite. has been used in the direct calorimetric determination of arginine (4). 9, IO-Phenanthrenedione, a fluorometric reagent for the quantitation of arginine (.5), has also been reported in the calorimetric determination of arginine (6). These reagents, however, generally require conditions for development that are too harsh for most biologically active proteins, rendering activity data difficult. if not impossible, to interpret. In nearly all other cases, the extent of modification or quantitation of the arginines in a protein is determined by amino acid analysis following acid hydrolysis. This process is relatively time consuming, especially when a large number of samples are to be analyzed.
AVAlLABLEARGlNlNERESlDUESBYp-HYDROXYPHENYLGLYOXAL,
Takahashi (7,8) has shown that phenylglyoxal reacts highly specifically with the guanidyl group of arginine and arginine derivatives under mild conditions to give a stable single product. He has also shown a high degree of success with this reagent on proteins (7,9). p-Hydroxyphenylglyoxal. a chromophoric analog of phenylglyoxal, was found to react specifically with arginines under mild conditions (pH 7-9 and 25°C) and behaved similarly to phenylglyoxal. Quantitation of the arginines modified, however, can now be determined spectrophotometrically after removal of unreacted reagent by gel filtration. The problem of long analysis time has been greatly shortened by this reagent and. because of the mildness of conditions used for modification, the biological activity of a large number of proteins can be subsequently assayed. EXPERIMENTAL
PROCEDURES
Mutevials
p-Hydroxyphenylglyoxal was prepared from p-hydroxyacetophenone (Aldrich) by the method of Fodor and Kovacs (IO). p-Hydroxyphenyl[2-‘“Clglyoxal was prepared from sodium [I-‘“Clacetate (New England Nuclear) and phenol (J. T. Baker) by the following method: 0.1 mmol sodium [ 1-“Clacetate (2.4 mCi/mmol) was dissolved in 17 mmol distilled acetyl chloride (Aldrich) and the solution was allowed to stand overnight at room temperature. The reaction mixture was centrifuged and to the supernatant was added 14 mmol distilled phenol in small portions with evolution of gas. A small amount of acetyl chloride was added and the reaction mixture was allowed to stand at room temperature for 1 h. The reaction mixture was placed under water aspirator vacuum and the residue was shown to be phenyl acetate by proton NMR. p-Hydroxyacetophen[‘“C]one (melting point 106- 107°C) was prepared from the labeled phenyl acetate by the Fries re-
33
action (11,12) and treatment of the labeled p-hydroxyacetophenone according to Fodor and Kovacs (10) yielded 100 mg of analytically pure p-hydroxyphenyl[2-‘“C)glyoxal monohydrate (1.12 x 10’ cpmimmol). (YAmino acids were purchased from Nutritional Biochemicals Corp. t.-Norleucine, N”-acetyl-t--lysine, r.-arginine methyl ester dihydrochloride, I-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride, bovine pancreatic ribonuclease A, soybean trypsin inhibitor. and bovine pancreatic carboxypeptidase A (treated with diisopropylfluorophosphate to eliminate traces of tryptic and chymotryptic activity) were purchased from Sigma. Citraconic anhydride was purchased from Pfaltz and Bauer. Antifreeze glycopeptide 8 was isolated from the blood serum of polar cod (Borro~adus stride) by the method of Osuga and Feeney (13). Chicken, duck, and rhea ovomucoids were isolated from the egg whites of chicken (Grrl1lr.s gcrlllrs). duck (Arm platvrh~nc,lto.s), and rhea ( Rhcrr tltncricu/lo ), respectively, according to Liu or trl. (14). Chicken ovotransferrin was isolated from the egg white of chicken (Gclllll.s ,~~r/llr.s) according to the method of Rogers et al. (15). All other materials used were purchased as reagent grade.
Bovine trypsin inhibitory activity was determined at 25°C by the spectrophotometric method of Rhodes et (11. ( 16). Ferric ion-binding activity was determined at 25°C by the spectrophotometric titration method of Rogers er (11. (15). The enzymatic activity of bovine pancreatic ribonuclease A toward yeast ribonucleic acid was determined at 37°C and pH 5.0 as described by Kalnitsky et rrl. (17). The enzymatic activity of bovine pancreatic carboxypeptidase A toward hippuryl-L-phenylalanine was done at 25°C according to Folk and Schirmer (18). All pH measurements were done on a Radiometer Model 26 pH meter. Ultraviolet and visible spectra were re-
34
YAMASAKI,VEGA,ANDFEENEY
corded on a Cary Model 118C spectrophotometer. A Beckman LS-250 liquid scintillation system was used to measure radioactivity. Typically, 0.40 ml of aqueous sample was mixed with 5.0 ml of phase combining system liquid scintillation fluid (Amersham) and kept in the dark for 5 min before taking counts. Complete acid hydrolysis of polypeptide and protein samples was achieved with 6 M HCI at 110°C for 22 h. Hydrolysates were analyzed on a Technicon Auto Analyzer. Losses in amino acids were calculated relative to L-norleucine internal standard. Rrrrction o~.p-hgdro.uvphrrl~l~l~~~.K~~l it>ith crmino clcids. The a-amino groups of the
common amino acids found in proteins were blocked by modification with citraconic anhydride according to Dixon and Perham (19). To 0.20 ml of the citraconylated amino acid solution (50 mM) was added 0.80 ml of buffer followed by 0.10 ml of p-hydroxyphenylglyoxal solution (400 mg/ml adjusted to the pH of the buffer with dilute NaOH). The pH of the reaction mixture was adjusted, when necessary, to that ofthe buffer, and the solution was incubated in the dark at 25°C. The buffers used were 0.1 M sodium phosphate, pH 7.0; 0.1 M sodium phosphate, pH 8.0: 0.1 M sodium pyrophosphate, pH 9.0; and 0.1 M sodium bicarbonate-carbonate. pH 10.0. At intervals, aliquots of 0.20 ml were removed and quenched with 0.80 ml of 0.1 M citric acid (final pH -3.3). The acidified aliquot was incubated at 25°C for 6 h to remove the citraconyl blocking groups. Control experiments showed complete blocking and removal of blocking groups under the conditions employed. Amino acid analysis was performed directly on the acidified samples. The citraconylation step was omitted for glycine and N”-acetyl-L-lysine. Rccrction c?f’p-llvdro.~~phrn~l~l~o.rnl (I rmmourginyl peptidr . Antifreeze
u.ith
glycopeptide 8, a glycopolypeptide containing no arginines (13). was dimethylated at the
a-amino group of the amino terminus with formaldehyde according to Means and Feeney (20). The dimethylated glycopolypeptide was reacted with L-arginine methyl ester in the presence of I-ethyl-3-(3-dimethylaminopropyl)carbodiimide coupling agent according to Hoare and Koshland (2 1). Complete coupling of arginine methyl ester was achieved after 20 min of incubation at 25°C. Purification of the modified glycopolypeptide by CM-cellulose chromatography yielded pure antifreeze glycopeptide 8 whose (Yamino group at the amino terminus was dimethylated and whose carboxyl terminus was extended by an arginine methyl ester residue. The reaction mixture, containing 10 mg of the arginylglycopolypeptide and 15 mg of p-hydroxyphenylglyoxal in 0.5 ml of 0.1 M sodium pyrophosphate buffer, pH 9.0, was incubated for 75 min in the dark before desalting on Sephadex G-IO. The product was immediately lyophilized. The same general procedure was followed when using radioactive p-hydroxyphenylglyoxal. Reaction of p-hydro.~~phrnyl~l-vo.~crl btjith proteins. In modifications where characteri-
zation of the products was desired, 40-80 mg of protein was reacted with 20-50 mg ofp-hydroxyphenylglyoxal in 2.5 ml of 0. I M sodium pyrophosphate buffer, pH 9.0. The reaction mixture was incubated at 25’C in the dark. At various intervals. 0.5-ml aliquots were removed for gel filtration on Sephadex G-25. The protein. eluted with deionized water, was lyophilized and assayed for activity. The arginines remaining were assayed by use of the molar absorption coefficient for the product (see Results), ‘“C incorporation, and by the calorimetric method of Tanabe et al. (6). The same basic procedure was followed when radioactive reagent was used. Small amounts of protein can also be modified and analyzed. When bovine pancreatic carboxypeptidase A was modified, 0.2 M sodium bicarbonate-carbonate/l.5 M NaCl buffer. pH 9.5, was used for the re-
AVAILABLE
ARGlNlNE
RESIDUES
action buffer. The reaction mixture, containing 4 mg of protein plus 2 mg of phydroxyphenyfgfyoxaI/mf reaction buffer, was incubated at 25°C in the dark. Aliquots (IO ~1) were removed from the reaction mixture at various intervals and assayed for activity directly. Aliquots (0.6 ml) were simultaneously removed for gel filtration on Sephadex G-25. The eluant was 0.2 M sodium bicarbonate/l.5 M NaCl buffer, pH 9.0. Arginines were determined directly on the eluate. The number of essential arginines of each protein was determined by the statistical analysis of Tsou (22). The number of essential groups is determined from quantitation of the number of groups modified and the residual activity in samples of partially modified protein. If all of the arginines in a protein are equally reactive toward the modifying reagent, but only i are essential, then the fraction of the i essential arginines remaining unmodified is equal to the fraction, x, of the total arginines remaining unmodified. The fraction of the activity remaining, (I, after partial modification is therefore given by CI”’ = x with the assumption that modification of any one of the essential arginines leads to complete loss of activity. The value of i. normally a small integer, is found from the plot of II’!’ versus x, which gives the best straight line. The Tsou method also treats the case where not all of the arginines react at the same rate. The general relation is then given by ((‘I’ 1
tz,y - (I? - p - s) P
where II is the total number of arginines in the protein, s arginines react most rapidly, none of which are essential, followed by the slower modification of p arginines, i of which are essential, and 01 - p - s) arginines are unreactive. Again, the number of essential arginines, i, may be obtained from the best straight line of the plot N Iii versus x,
BY p-HYDROXYPHENYLGLYOXAL
3.5
RESULTS A m ino Acids The relative rates of reaction between [I-hydroxyphenylglyoxal and the side chain groups of the common amino acids found in proteins were studied at pH’s 7.0, 8.0. 9.0, and 10.0. Arginine reacted much faster than did the other amino acids at all pH values. The rate of reaction with arginine increased with pH, resulting in a loss of 47% at pH 7.0. 93% at pH 8.0. and 100% at pH’s 9.0 and 10.0 after 60 min. At pH 9.0 or less. reaction with the other amino acids resulted in a loss of less than 5% after the same time period. Appreciable side reaction with glycine. asparagine. and histidine was found at pH 10.0. The p-hydroxyphenylglyoxal-arginine product appears unstable to conditions of acid hydrolysis (6 M HCl at 110°C for 22 h). Approximately 35% of the modified arginine is converted back to arginine under these conditions.
At pH 5.0. p-hydroxyphenylglyoxal absorbs maximally at 285 nm. with a molar absorption coefficient of 1.28 x lo-” M-’ cm-‘, while at pH 9.0 it absorbs maximally at 333 nm with a molar absorption coefficient of 1.68 x 10“ M-’ cm-‘. The phenolic proton has a pK,, of approximately 7.9 at 25°C. Phenylglyoxal does not demonstrate a shift in absorbance maximum in this pH range. The glycopolypeptide, containing only one arginine residue and prepared from antifreeze glycopeptide 8. was modified with p-hydroxyphenylglyoxal. Control experiments showed no modification of the carbohydrates or amino acid residues other than arginine. Polyacrylamide gel electrophoresis showed complete modification of the arginylglycopolypeptide to a single
36
YAMASAKI,
VEGA,
TIME 1. Bovine
ribonuclease phosphate via
absorption
pancreatic
ribonuclease
at 340
OF
A modified
A (I .4 mM) was reacted with buffer. pH 9.0, at 25°C. Activity coefficient
FEENEY
50
0
FIG.
AND
REACTION
100 (min)
withp-hydroxyphenylglyoxal.
p-hydroxyphenylglyoxal towards yeast
Bovine (57
ribonucleic
mM) acid
pancreatic
in 0.1 M sodium pyro(0): arginines calculated
nm (0).
derivative. Modification with p-hydroxyphenyl[2-‘4C]glyoxal showed 2.0 mol of p-hydroxyphenylglyoxal incorporated per mole of guanidyl group. An absorbance maximum for the modified arginylglycopolypeptide is found at 340 nm with a molar absorption coefficient of 1.83 x 10” Mm’ cm-’ at pH 9.0 and 25°C. The concentration of the modified arginylglycopolypeptide used in calculating the molar absorption coefficient was determined by amino acid analysis and sugar analysis according to Ashwell (23). The two methods gave results that differed by less than 3%. Duck ovomucoid, a protein containing only one arginine (24), was also modified by p-hydroxyphenylglyoxal and yielded an absorbance maximum at 340 nm with a molar absorption coefficient of 1.86 x IO-” M-l cm-‘. in close agreement with the absorption coefficient determined from the modified arginylglycopolypeptide.
Bovine pancreatic ribonuclease A, bovine pancreatic carboxypeptidase A, chicken ovotransferrin. soybean trypsin inhibitor, chicken ovomucoid, and rhea ovomucoid were treated with p-hydroxyphenylglyoxal. The results for bovine pancreatic ribonuclease A, a typical example, are shown in Fig. 1. The enzyme was inactivated when its arginines were modified. and analysis of the data by the Tsou Chen-Lu plot (Fig. 2) indicated the essentiality of a single arginine residue. Analysis of chicken ovotransferrin, soybean trypsin inhibitor, and chicken ovomucoid in the same manner also indicated a best fit for i = 1 (i.e.. a single essential arginine residue in each active site). It is necessary to remember that essentiality, as defined here, means that the modification of such a residue causes loss of activity.
AVAILABLE
ARGININE
RESIDUES
BY L>,-HYDROXYPHENYLGLYOXAL
37
FIG. 2. Tsou Chen-Lu plots of p-hydroxyphenylglyoxal-modified bovine pancreatic ribonuclease A. a = Fraction of activity toward yeast ribonucleic acid remaining: x = fraction of arginines remaining:i = number of essential arginines: i = I (0) and i = 2 (0). For further explanation see Experimental Procedures
With bovine pancreatic carboxypeptidase A (results shown in Fig. 3), nearly complete inactivation was obtained when only
0 TIME
, 15 OF REACTION
-a 30
.
(min)
FIG. 3. Bovine pancreatic carboxypeptidase A modified with p-hydroxyphenylglyoxal. Bovine pancreatic carboxypeptidase A (0.12 mM) was reacted with p-hydroxyphenylglyoxal (12 mM) in 0.2 M sodium bicarbonate-carbonate/l.5 M NaCI buffer, pH 9.5, at 25°C. Activity toward hippuryl-L-phenylalanine (0): arginines calculated via absorption coefficient at 340 nm (0).
two arginine residues were modified. The rest of the arginines in carboxypeptidase A appear unreactive. Riordan (2.5) has also found nearly complete inactivation with loss of two arginines when carboxypeptidase A is modified with butanedione in the presence of borate. Only one arginine appears to be essential for activity. However, the Tsou Chen-Lu plot of the carboxypeptidase A data did not give a straight line for either i = 1 or i = 2, indicating that the rate of modification of one arginine is significantly but not markedly different from the rate of the other (22). Tsou also treats this case but the analysis assumes first-order kinetics. Treatment of the carboxypeptidase A data did not show firstorder kinetics, precluding analysis by Tsou Chen-Lu plots and Koshland plots (26). It should be noted, however, that analysis by the Tsou Chen-Lu method does not require first-order kinetics if the rates of the different categories of reacting groups (see Experimental Procedures) are markedly dif-
38
YAMASAKI,
VEGA,
TIME
FIG. mM)
4. was
Chicken reacted
9.0, at 25°C. 340 nm (0);
Bovine arginines
ovomucoid with
modified
with
OF
FEENEY
REACTION
(min)
p-hydroxyphenylglyoxal.
p-hydroxyphenylglyoxal trypsin inhibitor calculated via
AND
(57 activity (0): l”C incorporation
mM)
Chicken
in 0. I M sodium
arginines (A).
calculated
via
ovomucoid
(0.85
pyrophosphate
buffer,
absorption
coefficient
pH at
ferent. Such appears to be the case with the other protein examples used in this study. The results for chicken ovomucoid are shown in Fig. 4. The protein was inactivated when its arginines were modified and the loss in arginine calculated by the molar absorption coefficient (at 340 nm) of 1.83 x lo4 M-’ cm-’ at pH 9.0 agreed well with the loss in arginine calculated by “‘C incorporation.’ Some of the results calculated by the molar absorption coefficient at 340 nm were checked by the calorimetric method of Tanabe ef al. (6) and these results are shown in Table 1. There is reasonable agreement between all three methods of arginine quantitation. Rhea ovomucoid, a glycoprotein with no arginines but containing an essential
lysine (14), was relatively unaffected by treatment with p-hydroxyphenylglyoxal. No loss in bovine trypsin inhibitory activity was found for any of the time points. The gain in absorbance at 340 nm after 60
4 Calculation of the loss in arginine according lJC incorporation assumes 2 mot of p-hydroxyphenylglyoxa) per mole of arginine.
M ‘Cm-’ arginines. ” Used
TABLE
PERCI.,TVT~G~ ARGININF RI~MAINI~G p-HYDROXYPH~NYLt;IYOXAIMODIFIED PROIIJINS
IN
By calorimetric Protein Chicken
Soybean inhibitor
” Used to
I
ovomucoid
trypsin
molar at pH the
analysis”
BY A:uu”
absorption 9.0 and 25°C
method
74
70
56
54
94 79
VI 75
.54
60
coefficient of for quantitating
of Tanabe
~‘1 rrl.
t6).
1.83 x IO’ modified
AVAILABLE
ARGININERESIDUES
TABLE? PERCENTAGE AKGININE PHENYI.CLYOXAL-MODIFIED RFTER DIAI YSIS
Dialyzed against/ dialysis time (h)
REMAINING IN p-HYOROYW PROTEINS BEFORE .&ND
Chicken ovomucoid Before dialysis
After dialysis
Soybean trypsin inhihitor Before dialysis
After dialysi\
water’1 water/25 pH 7”/48 pH 9”/48 ” 0. I M sodium ‘) 0. I M sodium
phosphate. pH 7. pyrophosphate. pH 9.
min was less than 2% of that for chicken ovomucoid. and the amount of reagent introduced into the protein (by 14Ccounts) was 0.13 mol per mole or less than 2YSthat for chicken ovomucoid. Absorbance spectra of p-hydroxyphenylglyoxal-modified proteins (except rhea ovomucoid) at pH 9.0 showed an absorbance maximum at 340 nm in addition to the native protein absorbance maximum near 280 nm. No observable absorption was found at 340 nm for any of the native proteins studied. The absorbance at 340 nm, pH 9.0. of the modified proteins, obeyed Beer‘s law in the range of 5-50 FM with respect to modified arginine residues. The stability of the p-hydroxyphenylglyoxal-modified arginine is reflected in Table 2. Very little reversal of modification takes place after dialyzing against water for I h at ?S”C. However, significant amounts of reversal may be obtained when dialyzing for longer periods under neutral or alkaline conditions. DISCUSSION
The results of this study show that phydroxyphenylglyoxal can be used to modify arginine residues in proteins rapidly. with high specificity, under mild conditions. A maximum rate of arginine modification with
BY/>-HYDROXYPHENYLGLYOXAL
39
maximum specificity is obtained between pH 8 and 9. The absorbance at 340 nm (molar absorption coefficient of I.83 x JO’ M-’ cm-’ at pH 9.0) allows a quick estimation of the extent of modification and because of the mildness of conditions employed, the biological activity of a large number of proteins can be reliably assayed. Determination of p-hydroxyphenylglyoxal-modified arginine residues in proteins by amino acid analysis does not appear to be feasible due to the instability of the modified arginine under acid hydrolysis conditions. The necessity of removing the excess reagent by molecular exclusion requires an additional step for the quantitation. and a simpler method would be possible if this step were unnecessary. This step is, however, neither difficult nor time consuming. Under the conditions employed. the product is stable and reversibility of the reaction is negligible. p-Hydroxyphenylglyoxal, easy to prepare inexpensively and in pure form, both unlabeled and “C labeled. also has the potential for generating a reporter group because of the titratable phenolic proton. The results in this investigation indicating numbers of essential arginines agree closely with those of previous investigators. Chicken ovomucoid ( 14). soybean trypsin inhibitor ( 14). bovine pancreatic ribonuclease A (7.27). chicken ovotransferrin (28). and bovine pancreatic carboxypeptidase A (25) have been shown to contain a single arginine residue that cannot be modified with retention of activity. The method of Tsou (22) has been successfully used by Paterson and Knowles (319) and Rogers ct 01. (38). ACKNOWLEDGMENTS The authors thank David Sherman for preparing p-hydroxyphenylglyoxal and p-hydroxyphenyl[Z-“C]glyoxal, David Friend and Dorothy Shimer for technical assistance, David T. Osuga for advice, Chris Howland for editorial assistance. and Clara Robison for typing the manuscript.
40
YAMASAKI,
VEGA,
REFERENCES 1. Borders, C. L., Jr., Pearson, L. J.. McLaughlin. A. E.. Gustafson, M. E.. Vasiloff. J.. An, F. Y., and Morgan, D. J. (1979) Bioclritn. Biuphx.s. Acru 568, 491-495. 2. Riordan. J. F., McElvany. K. D.. and Borders, C. L., Jr. (1977) Science 195, 884-886. 3. Levy, H. R.. Ingulli, J., and Afolayan, A. (1977) J. Bid. Chrm. 252, 3745-3751. 4. Sakaguchi. S. ( 1925)J. Bkxhrm. (Tok.w) 5,25-31. 5. Yamada, S., and Itano. H. A. 11966) Eiochim. Biophys. Acta 130, 538-540. 6. Tanabe, S., Oya, T.. and Sakaguchi. T. (1975) Chem. Phormuc~ol. Bull. (Tokyo) 23, 16571663. 7. Takahashi. K. (1968) J. Bid. Cfrrm. 243, 61716179. 8. Takahashi, K. (1977) J. Biochem. tTo!,yo) 81, 395-402. 9. Takahashi, K. (1977)J. Biochrm. fToXpoj 81.403414. IO. Fodor. G.. and Kovacs, 0. ( 1949) J. Ampr. Chem. Sot. 71, 1045- 1048. 11. von Rosenmund, K. W.. and Schnurr. W. (1928) J~tsfrts Lichigs Arm. Chrrn. 460, 56-98. I?. Blatt, A. H. (1942) in Organic Reactions (Adams. R.. ed.), Vol. 1. pp. 342-369, New York. 13. Osuga. D. T., and Feeney. R. E. (1978) /. Bid. fhem. 2.53, 5338-5343. 14. Liu, W.-H., Feinstein, G.. Osuga. D. T., Haynes. R.. and Feeney, R. E. (1968) Biochemistry 7, 2886-2892.
AND
FEENEY
IS. Rogers, T. B.. Gold, R. A., and Feeney, R. E. (1977) Biwhemistry 16, 2299-2305. 16. Rhodes, M. B., Bennett, N.. and Feeney, R. E. (1960) J. Bicd. Chem. 235, l686- 1693. 17. Kalnitsky. G.. Hummel. J. P.. and Dierks. C. (1959j.1. Bid. Chum. 234, 1512-1516. IX. Folk. J. E.. and Schirmer. E. W. (1963) J. Bid. C-hem. 238, 3884-3894. 19. Dixon, H. B. F.. and Perham, R. N. (1968) Eiochrm. J. 109, 312-314. 20. Means. G. E.. and Feeney, R. E. (1968) Biochrm;.\try 7, 2 192-2201. 21. Hoare. D. G.. and Koshland. D. E.. Jr. ( 1967) J. Biol. C/rem. 242. 2447-2453. 22. Tsou Chen-Lu (1962) SC,;. Sin. 11, 1535-1558. 23. Ashwel, G. (1966) itr Methods in Enzymology (Neufeld. E. F., and Ginsburg, V.. eds.), Vol. 8. pp. 85-95. Academic Press, New York. 14. Feeney, R. E.. and Allison, R. G. (1969) irr Evolutionary Biochemistry of Proteins. Homoogous and Analogous Proteins from Avian Egg Whites. Blood Sera, Milk. and Other Substances. p. 209, Wiley. Interscience. New York. 25. Riordan, J. F. (1973) Biochemisrry 12, 3915-3923. 26. Koshland. D. E., Jr., Ray. W. J.. Jr.. and Erwin, M. J.( 1958) Fed. Proc. 17, 1145- 1150. 27. Patthy, L.. and Smith. E. L. (1975) J. Birjl. Chem. 2.50, 565-569. 28. Rogers, T. B.. Borresen. T.. and Feeney, R. E. (1978) Biochemisrry 17, 1 IOS- 1109. 29. Paterson, A. K.. and Knowles. J. R. (1972) Fur. J. Biochem. 31. 510-517.