ARCHIVES
OF BIOCHEMISTRY
AND BIOPHYSICS
Vol. 297, No. 2, September, pp. 199-204, 1992
Characterization and Partial Purification from Neurospora crassal Hemanta
K. Kole, Deborah
R. Smith,
and John
of an lnsulinase
Lenard2
Department of Physiology and Biophysics, University of Medicine and Dentistry, (at Rutgers), 675 Hoes Lane, Piscataway, New Jersey 08854-5635
Received November
Robert Wood Johnson Medical School
5, 1991, and in revised form April 2, 1992
An insulin-binding metal- and thiol-dependent proteinase has been purified 1491-fold from high speed cytosolic fractions of the fungus Neurospora crassa. This enzyme resembles insulin-degrading enzymes (insulinases) present in mammalian cells and in Drosophila melanogaster in the following ways: (i) it degrades radiolabeled insulin with a specificity similar to that of rat muscle insulinase, as demonstrated by HPLC analysis of the degradation products; (ii) it is inhibited by bacitracin, EDTA, l,lO-phenanthroline, and the sulfhydryl-reactive compounds N-ethylmaleimide and p-chloromercuribenzoate, but not by inhibitors of serine proteases or by lysosomal protease inhibitors. Cross-linking with 12’I-insulin labels a band of ca. 120 kDa, and several smaller bands which may represent degradation products. The N. crassa insulinase is stimulated by Mn2+ and strongly the enzyme inhibited by Zn2+; Mn2+ can also reactivate after inhibition by EDTA, but Zn2+ is ineffective. The N. crassa protein differs in this regard from mammalian and insect insulinases which are generally activated by both Mn2+ and Zn2+. This finding extends the apparent evolutionary conservation of these metal- and thiol-dependent proteases into the microbial realm. 0 lsea Academic Press, Inc.
The insulinase family of enzymes consists of metaland thiol-dependent proteases that have previously been found in the cells of mammals (1) and of Drosophila melanogaster (2). As purified and cloned from both sources, insulinase (insulin-degrading enzyme) is a llO-kDa cytoplasmic protein bearing homology to an Escherichia coli periplasmic protease named protease III (3-5). The purified protein from both sources degrades insulin with i Supported by Grant DK 39502 from the National Health. * To whom correspondence should be addressed. 0003~9861/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Institutes
of
characteristic specificity (6). It is thought to play a role in regulating the magnitude or extent of insulin-induced signals by inactivating cell-associated insulin, although conclusive proof of this is lacking. Recent reports from this laboratory have provided extensive evidence that the filamentous fungus Neurospora crassa possesses a functional insulin-induced signal transduction pathway. Addition of mammalian insulin to N. crassa cells results in stimulation of glucose metabolism, as measured by increased production of COZ, ethanol, alanine, and other metabolites, as well as enhanced glycogen synthesis (7-10). Stimulation of glycogen synthesis was shown to arise from the activation of the enzyme glycogen synthase from the glucose B-phosphatedependent form (D) to the independent form (1)-a conversion that results from insulin-induced dephosphorylation in mammalian cells (10). A specific, high affinity insulin-binding protein was purified from N. crassa membranes and was proposed to be the insulin receptor (11). Insulin-dependent protein phosphorylation was found to occur in N. crassa cells at serine, threonine, and tyrosine residues (12). Earlier reports described the presence of an insulin-like material in N. crassa medium and cells (13), and of effects of insulin on adenylate cyclase levels in N. crassa membranes (14). During the course of purification of the putative insulin receptor from N. crassa membranes, a contaminating insulin-binding protein which proved to be a bacitracininhibitable insulin-degrading protease was encountered (11). We now report further characterization and partial purification of this protein and discuss its resemblance to previously described insulinases. EXPERIMENTAL
PROCEDURES
insulin and iz61-Bz6-human insulin were Materials. ?-An-porcine purchased from NEN and Amersham, respectively. Ultrapure ammonium sulfate was obtained from BRL (Bethesda, MD) and HPLC grade acetonitrile from Baxter (Muskegon, MI). Disuccinimidyl suberate was purchased from Pierce (Rockford, IL). All other chemicals were pur199
200
KOLE,
SMITH,
chased from Sigma Chemical Co. (St. Louis, MO). MonoQ and Superose6 columns were obtained from Pharmacia and a Zorbax C8 column was from DuPont. Rat muscle insulin protease was a generous gift from Dr. Frederick G. Hamel, University of Nebraska Medical Center (Omaha, NE). Cells. N. crossa strain FGSC 4761 is a rapidly growing variant of the wall-less “slime” mutant FGSC 1118 (15). Cells were maintained and grown as described previously (12). Insulin degradation assay. Insulin-degrading activity was determined by the trichloroacetic acid (TCA)3 precipitation method as described (11) with modifications. Briefly, the enzyme was incubated in a 200-~1 reaction mixture containing 0.1 M Na,HPO, buffer (pH 7.5), 100 pg bovine serum albumin (BSA), and 800 pM “‘I-insulin. After 15 min at 37°C the reaction was stopped with 200 ~1 of 10% TCA and processed as described (11). During studies with divalent cations, 100 mM Hepes buffer, pH 7.4, was used. Enzyme activity was linear with respect to time and amount of enzyme protein added. Determination of optimal pH for insulin degradation. Optimal pH for insulin degradation was determined using Universal buffer (16) at different pH values. Purification of the insulin-degrading enzyme. The separation and purification were performed at 4”C, if not otherwise indicated. Cells were harvested, broken by Dounce homogenization, and 100,OOOgsupernatants (cytosol) were prepared as previously described (11) except that in this case we saved the cytosolic fraction insteadof the membrane fraction. The supernatant was fractionated by adding ammonium sulfate to 40% saturation, and the solution was stirred for 30 min and then centrifuged for 30 min at 15,000g. The pellet was discarded and the supernatant was adjusted to 65% saturation with ammonium sulfate, stirred, and centrifuged as above. The pellet was resuspended in 5 ml of 0.01 M potassium phosphate buffer, pH 6.5, at 23°C and dialyzed against the same buffer at 4°C for 18 h. The dialyzed solution was applied to a Mono& HR 5/5 anion-exchange column pretreated with 0.1 M TrisHCl buffer, pH 7.4. The column was washed with 50 ml of the same Tris-HCI buffer and the enzyme activity eluted with 30 ml of a linear gradient from 0 to 0.4 M NaCl in Tris-HCl buffer at a flow rate of 1 ml/ min. Active fractions were pooled, concentrated in a Centriprep 10, and loaded onto a Superose-6 column which was preequilibrated with 10 mM potassium phosphate buffer, pH 6.5. Enzyme activity was eluted with the same phosphate buffer and loaded directly onto the MonoQ column for a second time, as above. The chromatogram was developed using a 40-ml linear gradient (O-O.3 M NaCl) in the same Tris buffer at a flow rate of 0.5 ml/min. Fractions (0.5 ml) were collected in glass tubes containing 150 ~1 of 100% glycerol. Protein was determined according to Bradford (17). Affinity labeling of proteins with ‘251-insulin. Samples from the different purification steps were incubated with 2 ng of 1251-BZG-insulin for 1 h at 4°C. Disuccinimidyl suberate (0.4 mM) was added to the reaction mixture and incubated for another 15 min. The reaction was stopped by the addition of denaturing buffer and the samples were separated on a 10% SDS-polyacrylamide gel. Following SDS-PAGE, gels were dried and analyzed by autoradiography as previously described (11). Insulin-degrading activity on polyacrylamide gel. Samples from different steps of purification were fractionated by discontinuous polyacrylamide gel electrophoresis at pH 8.9 in gels containing 7% polyacrylamide, as described by Dewald et al. (18), omitting the detergent from the gel and electrode buffers. The gels were run at 4°C for 3 h at 5 mA constant current. At the end of the run, the gels were equilibrated with 0.1 M sodium phosphate buffer (pH 7.5) containing 1 mg/ml BSA. The gels were then sliced and each l-cm section was incubated with
3 Abbreviations used: TCA, trichloroacetic acid, BSA, bovine serum albumin; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PMSF, phenylmethylsulfonyl fluoride; NEM, N-ethylmaleimide; PCMB, p-chloromercuribenzoate.
AND
LENARD
1251-A’4-insulin at 37°C for 30 min. Degradation by the TCA solubility method.
of ‘251-insulin was assayed
HPLC analysis of “s1-insulin degradation products. ‘251-B26-human insulin was incubated with either rat muscle or N. crassa insulinase at 37°C in Tris-HCl buffer, pH 7.5, so as to reach a level of degradation of ca. 20% as measured by TCA precipitation. HPLC analyses of the insulin degradation products were carried out both before and after sulfitolysis following the method of Duckworth et al. (6). Fractions (0.5 ml) were collected in siliconized glass vials containing 10 pg of unlabeled insulin and counted directly in a LKB CompuGamma CS counter to determine the elution profile of radioactivity. Recoveries of the injected radioactivity ranged from 80 to 95% in the experiments shown.
RESULTS
We have previously reported (11) that N. crassa membranes possessed insulin-degrading activity which was purified away from a membrane-associated high affinity insulin-binding protein. When the distribution of this activity in a total N. crassa homogenate was investigated, however, it was found that 85% of the insulin-degrading activity was present in the 100,OOOgsupernatant, while only 15% remained associated with the membrane preparation. Furthermore, the membrane-associated activity was not enriched by purification or subsequent solubilization of the membranes. It was therefore concluded that the N. crassa insulin-degrading activity arises from a cytosolic enzyme which contaminated membrane preparations to some extent. The cytosolic insulin-degrading activity was partially purified in several steps, as shown in Table I. Modest enrichments of specific activity were obtained from a 4065% ammonium sulfate cut (step 2, Table I) and from MonoQ ion-exchange chromatography, with activity eluting between 0.19 and 0.29 M NaCl (step 3, Table I). Superose-6 chromatography (step 4, Table I) resulted in some further purification. Application of the resulting material to a second Mono& column (step 5, Table I) afforded major purification, with activity eluting at 0.1870.207 M NaCl (Fig. 1). Purification of 1491-fold was attained, with 18.9% yield. Attempts to purify the enzyme further using an insulin-affinity column were unsuccessful. The purified material did not cross-react with antiIDE antibody prepared against the Drosophila enzyme (kindly provided by Dr. Marcia Rich Rosner, University of Chicago, Chicago, IL; data not shown). Cross-linking of purified insulinase with 1251-insulin consistently labeled a polypeptide of ca. 120 kDa and showed several smaller bands that varied between experiments. Labeling of these bands was specifically competed by unlabeled insulin (data not shown). Analysis by disc electrophoresis showed that insulinase activity was restricted to a single, well-defined region of the gel, which remained invariant through several stages of purification (Fig. 2). Specificity of insulin cleavage by N. crassa insulinase was compared with that of a previously characterized insulin protease from rat muscle (kindly provided by Dr. F.
INSULINASE
IN Neurospora TABLE
Purification Total protein (mgl
Steps 1. cytoso1 2. 40-65% Ammonium sulfate 3. Mono& ion-exchange column, first time 4. Superose 6 5. MonoQ ion-exchange column, second time
of N.
crassa Insulinase Total activity (fmol)
63.4 306
2.52 0.69 0.0102
a fmole insulin degraded/min/mg
I
Specific activity”
80.5 15.75
201
crassa
Relative purification
5104 4820
1.0 4.8
1,168 2,756
2943 1902
18.4 43.5
94,502
964
Yield (%) 100 94 57.7 37.3
1491
18.9
protein.
Hamel, University of Nebraska Medical Center, Omaha, NE). Results are shown in Figs. 3 and 4. Extensive similarities are evident in the degradation products generated by the N. crassa and rat muscle insulinases, analyzed both before (Fig. 3) and after (Fig. 4) insulin chain separation by sulfitolysis. In order to characterize the N. crussa insulinase, a number of physical and kinetic parameters were determined. The optimal pH for the cytosolic N. crassa insulinase was around 7.5 (Fig. 5), similar to that found for the insulinases isolated from human fibroblasts (19), rat islets (20), and Drosophila (21). The degradation rate increased with increasing temperature up to 37°C (not shown) which is the standard assay temperature for mammalian and Drosophila insulinases (19, 21, 22). Effects of various agents on insulin degradation by intact N. crassa cells and by the purified enzyme are listed in Table II. Bacitracin was found to inhibit essentially all of the activity at 1 mg/ml, both of intact cells and of purified insulinase. Excess porcine insulin (10 PM), as expected, inhibited degradation of labeled material by both
purified insulinase and intact cells. Excess glucagon, on the other hand, was much less effective than excess insulin. The purified enzyme was unaffected by the serine protease inhibitor PMSF, soybean trypsin inhibitor, aprotinin, or the cathepsin inhibitor leupeptin (Table II). Chloroquine had no effect on the purified enzyme, and actually stimulated insulinase activity in intact cells. The N. crassa insulinase was strongly inhibited by 10 mM EDTA (Table II) and by l,lO-phenanthroline (Table III), suggesting that it is a metal-dependent protease. The effects of adding various divalent cations to N. crassa insulinase are shown in Fig. 6. The profile is quite unusual, and differs from other insulinases in that Mn2+ stimulated the activity while Zn2+ inhibited strongly. Lower concentrations of Zn2’ are very effective in inhibiting insulinase activity (Fig. 6, inset). Ca2+ and Mg2+ stimulated activity slightly. Consistent with these results, Mn2+ was the most effective of the four divalent cations tested at reversing inhibition by EDTA (Table IV). Ca2+ was partly effective,
I 70
70 I
Ammonium
adfat*
20
- 60 3
- 50 = /\
Protein
c
-40 :
: : :
’
m J - 30 .g
\ \ \
i
-20 \
‘.
\ .L’\-.
25
E
E .sJ
15
: e
10
: 0 5
- 10 0
30
Fraction
FIG. 1. Elution profile of insulin-degrading activity from the second MonoQ column (step 5, Table I). Twenty-microliter aliquots of each fraction were tested for degrading activity by the TCA-precipitation method, as described under Experimental Procedures.
0
0
3
5
5
10
Gel Slices
FIG. 2. Localization of insulin-degrading activity on a polyacrylamide disc gel. Samples from the ammonium sulfate and first and second Mono Q steps were fractionated by discontinuous polyacrylamide gel electrophoresis at 4°C and assayed for degrading activity as described under Experimental Procedures.
202
KOLE,
SMITH,
AND
19,
19 -
17 -
17 -16
-
I
15 H -
3 @ 's LI
10 -
13 -
1211 10 -
f!
9-
Q
a-
.G 't'
7 6-
-
g;-
1
5 1
E
4-
Q q
!j4-
3-
=
3-
15
20
25
30
35
40
45
50
55
60
65
70
75
-
14 -
IB E
9f
-
Q '5
8 8 a
14 t
B
16 -
A
16 -
-16
LENARD
60
0
5
1015202530354045505560657075601
Elutlon Time (min) FIG. 3. HPLC elution profile of degradation rat muscle insulinase (B).
5
Elutlon Tlmr (mln)
products of iz51-Bz6-human insulin produced by partial
digestion with N. crassa insulinase
(A) or
B A 21 -
24 4
s:
18 g17
I;;-
-
t
16 15 -
8
15 -
= 1 5 l's
14 -
12 -
s Ir: = 1 $ c 6
E
11 -
E
11 -
e
10 -
8
10 -
13 -
Q92 z 5 $6u a
a
6-
f
7-
3-
12 -
9-
g
6-
8 B 0
a
4-
13 -
d
=g
5-
14 -
7-
654-
1. :x ; ! LA!! a
f
2 .q 1 0
0
5
10
15
20
i LI
30
35
40
45
Xl
55
60
65
70
75
Elutlon Time (mln) FIG. 4. HPLC elution profile after sulfitolysis insulinase (A) or rat muscle insulinase (B).
60
30
’ 35
t 40
1 45
I 50
I 55
60
65
70
75
60
Elutlon Tlme (mln) of degradation
products of ‘261-B2G-human insulin produced by partial
digestion with N. crassa
INSULINASE
IN Neurosporu
crassa
203 TABLE
III
Effect of Sulfhydryl Inhibitors and 0-Phenanthroline on N. crassa and Rat Muscle Insulinase %Inhibition Inhibitors NEM 0.1 8
4
N. crassa insulinase
1.2 4 14 41
mM
0.5 mM
12
1.0
mM
5.0 mM
PH
FIG. 5. pH-dependence of insulin degradation by N. crossa insulinase. Insulin degrading activity of 8 pg of Superose 6 eluate (step 4, Table I) was determined in Universal buffer (16) adjusted to the indicated pH values.
PCMB 0.1 mM
72 95
0.5 mM 1.0 mM l,lO-phenanthroline
0.05mM while Mg2+ and Zn2+ were inactive. The N. crassa insulinase thus behaves as a Mn2+-requiring, Zn2+-inhibitable metal-dependent protease. The effect of sulfhydryl inhibitors N-ethylmaleimide (NEM) and p-chloromercuribenzoate (PCMB) on N. crassa and rat muscle insulinases were compared, as shown in Table III. In agreement with previous reports, both inhibitors gave >80% inhibition of the rat muscle insulinase at 0.1 mM, the lowest concentration tested. For N, crassa insulinase, on the other hand, higher concentrations of both inhibitors were required; NEM inhibited only 41% at 5 mM, the highest concentration tested, but complete inhibition was achieved with 1 mM PCMB (Table III). The N. crassa insulinase thus appears to be a thioprotease, but the essential sulfhydryl group is substantially less reactive than that in rat muscle insulinase. Analysis of the degradation of porcine insulin by the last Mono& eluant showed maximal activity at 280-400
TABLE
II
of degradation
0.5 mM
pmol/min/mg
protein,
f 1.7 f2 k3 +2
80 +- 1 97 t 2 99 * 1 100
f3 *2 100
19 48 86 97
0.1 mM 0.25 mM
Rat muscle insulinase
100 100 100
+3 29 + 0.4 fl
with 50% activity
at ca. 150 nM
insulin. DISCUSSION
The insulinase from N. crassa that is described in this report bears several significant similarities to insulinases purified from other sources: it is a bacitracin-inhibitable metal- and thiol-dependent protease (Tables II, III, IV, Fig. 6), probably of ca. 120 kDa, which cleaves insulin with
characteristic
specificity
(Figs. 3, 4). It differs
200 ,
1
Agents on Insulin Degrading Activity of N. crussa Cells and of Purified N. crassa Insulinase
Effect
from
previously reported insulinases in its specific metal requirement, and in its unusually low sensitivity to sulfhydryl inhibitors. A literature search failed to reveal any published examples of a Mn2+-stimulated, Zn2+-inhibited
of Different
% Degradation Cells
Inhibitors None Bacitracin, 1 mg/ml Porcine insulin, 10 pM EDTA, 10 mM Chloroquine, 1 mM PMSF, 1 mM Soybean trypsin inhibitor, Aprotinin, 0.5 mg/ml Leupeptin, 0.5 mg/ml Glucagon, 10 pM Dithiothreitol, 1 mM ’ Not determined.
1 mg/ml
49 0 12 + 3 31 + 3 56 k 3 69 k 2 ND” ND ND ND ND
Insulinase 36 -c 2 0 1 + 0.2 4 + 0.6 35 * 3 36 + 3 35 rt 3 36 + 2 34 -+ 1 22 f 2 37 AZ1
0
1
,
m* 0 05
2
3
.I .I5 .2 IZn *+ I. mM 4
.25 5
[M2+l, mM FIG. 6. Effect of divalent cations on insulin degradation by N. crassa insulinase. Insulin degrading activity was determined in the presence of different divalent cations as indicated, using 0.1 fig of the second MonoQ eluate (step 5, Table I). Inset, effect of lower concentrations of Zn2+ on insulinase activity.
204
KOLE. TABLE
SMITH,
None 12.5 mM EDTA 10 mM MnClr 10 mM CaCi, 10 mM MgClz 10 mM ZnCI,
LENARD
REFERENCES
IV
Reactivation of EDTA-Inhibited N. crassa Insulinase by Divalent Cations Additions
AND
1. Duckworth,
% Degradation
48 + 0.4 6.4 510.3 52 +2 25 k3 10 12
*1 * 0.4
protease; most other insulinases are stimulated by both Mn2+ and Zn2+ (1). Previous reports from this laboratory have provided evidence for the presence in N. crassa of an insulin-induced signal transduction system including a putative membrane-bound receptor (ll), numerous insulin-induced changes in protein phosphorylation at serine, threonine, and tyrosine residues (12), and enhancement of glucose metabolism, including activation of glycogen synthase (7, 8, 10). The presence in the same cells of an enzyme resembling the insulinases of other insulin-responsive cells raises the question of its functional role. Although no function has been established for insulinase in any cell type, several observations have raised the possibility that it might have a role in insulin regulation. Treatment of human hepatoma cells (23) or opossum kidney epithelial cells (24) with bacitracin or other inhibitors of insulinase resulted in reduced degradation of added insulin and other growth factors. Insulin degradation in hepatoma cells was decreased 18-54% by microinjection of antibodies directed against insulinase (25), and insulin could be cross-linked to insulinase in these cells (26). These findings indicate that insulinase can act on insulin added to intact cells in culture, but provide no evidence as to the functional significance of this action. It has been reported that insulinase varies in amount at different developmental stages in D. melanogmter (27), and that insulinase might be involved in differentiation of muscle cells (28), but the relationship of these properties to insulin degradation has not been demonstrated. The finding of a metal- and thiol-dependent protease with properties of insulinase in microbial cells extends further the evolutionary range in which these enzymes are known to exist, and may open possibilities for genetic studies of function that are not feasible in other systems. This finding is also consistent with a recent suggestion that a superfamily of insulinase-like enzymes, which include the Mn2+-activated mitochondrial processing proteases (29), exists.
W. C. (1988) Endocr. Reu. 9,319-345.
2. Rosner, M. R. (1990) Mol. Reprod. Dev. 27, 54-59. 3. Atfholter, J. A., Fried, V. A. and Roth, R. A. (1988) Science 242, 1415-1418.
4. Affholter,
J. A., Hsieh, C. L., Francke, Mol. Endocrinol. 4, 1125-1135.
U., and Roth, R. A. (1990)
5. Kuo, W. -L., Gehm, B. D., and Rosner, M. R. (1990) Mol. Endocrinol. 4,1580-1591.
6. Duckworth,
W. C., Garcia, J. V., Liepnieks, J. J. Hamel, F. G. Hermodson, M. A., Frank, B. H., and Rosner, M. R. (1989) Biochemistry
28,2471-2477. 7. McKenzie, M. A., Fawell, S. E., Cha, M., and Lenard, J. (1988) Endocrinology 122,511-517. 8. Greenfield, N. J., McKenzie, M. A., Adebodun, F., Jordan, F., and Lenard, J. (1988) Biochemistry
27,8526-8533.
9. Greenfield,
N. J., Cherapak, C. N., Adebodun, F., Jordan, F., and Lenard, J. (1990) B&him. Biophys. Actu 1025, 15-20.
10. Fawell, S. E., McKenzie, M. A., Greenfield, N. J., Adebodun, Jordan, F., and Lenard, J. (1988) Endocrinobgy 122,518-523. 11. Kole, H. K., Muthukumar,
F.,
G., and Lenard, J. (1991) Biochemistry
30,682-688. 12. Kole, H. K., and Lenard, J. (1991) FASEB J. 5,2728-2734. 13. LeRoith, D., Shiloacb, J., Roth, J., and Lesniak, M. A. (1980)
PFOC.
Natl. Acad. Sci. USA 77,6184-6188. 14. Flawia, M. M., and Torres, H. N. (1973) J. Biol. Chem. 248,4517-
4520. 15. Scarborough, G. A. (1985) Exp. Mycol. 9, 275-278. 16. Davison, R. M. C., Elliot, D. C., Elliot, W. H., and Jones, K. M. (Eds.) (1978) Data for Biochemical Research, 2nd ed., pp. 485, Clarendon, Oxford. 17. Bradford, M. M. (1976) Anal. B&hem. 72,248-254. 18. Dewald, B., Dulaney, J. T., and Touster, 0. (1974) in Methods in Enzymology (Fleischer, S., and Packer, L., Eds.), Vol. 32, pp. 8291, Academic Press, San Diego. 19. Stentz, F. B., Harris,
H. L., and Kitabchi,
E. (1985) Endocrinology
116,926-934. 20. Bhatnena,
S. J., Timmers, K. I., Oie, H. K., Voyles, N. R., and Recant, L. (1985) Diabetes 34, 121-128. 21. Garcia, J. V., Fenton, B. W., and Rosner, M. R. (1988) Biochemistry
27,4237-4244. 22. Duckworth, W. C., Heinemann, 23. 24. 25. 26.
M. A., and Kitabchi, A. E. (1972) PFOC.Natl. Acad. Sci. USA 60, 3698-3702. Gehm, B. D., and Rosner, M. R. (1991) Endocrinology 128, 16031610. Dahl, D. C., Tsao, T., Duckworth, W. C., Frank, B. H., and Rabkin, R. (1990) Diabetes 39, 1339-1346. Shii, K., and Roth, R. A. (1986) Proc. N&Z. Acad. Sci. USA 83, 4147-4151. Hari, J., Shii, K., and Roth R. A. (1987) Endocrinobgy 120, 829-
831. 27. Stoppelli,
M. P., Garcia, J. V., Decker, S. J., and Rosner, M. R. (1988) PFOC.Natl. Acad. Sci. USA 85, 3469-3473. 28. Kayalar, C., Wong, W. T., and Hendrickson, L. (1990) Biochemistry
44,137-151. 29. Rawlings, N. D., and Barrett, 391.
A. J. (1991) Biochem. J. 275,
389-
OF BIOCHEMISTRY
AND BIOPHYSICS
Vol. 297, No. 2, September, pp. 199-204, 1992
Characterization and Partial Purification from Neurospora crassal Hemanta
K. Kole, Deborah
R. Smith,
and John
of an lnsulinase
Lenard2
Department of Physiology and Biophysics, University of Medicine and Dentistry, (at Rutgers), 675 Hoes Lane, Piscataway, New Jersey 08854-5635
Received November
Robert Wood Johnson Medical School
5, 1991, and in revised form April 2, 1992
An insulin-binding metal- and thiol-dependent proteinase has been purified 1491-fold from high speed cytosolic fractions of the fungus Neurospora crassa. This enzyme resembles insulin-degrading enzymes (insulinases) present in mammalian cells and in Drosophila melanogaster in the following ways: (i) it degrades radiolabeled insulin with a specificity similar to that of rat muscle insulinase, as demonstrated by HPLC analysis of the degradation products; (ii) it is inhibited by bacitracin, EDTA, l,lO-phenanthroline, and the sulfhydryl-reactive compounds N-ethylmaleimide and p-chloromercuribenzoate, but not by inhibitors of serine proteases or by lysosomal protease inhibitors. Cross-linking with 12’I-insulin labels a band of ca. 120 kDa, and several smaller bands which may represent degradation products. The N. crassa insulinase is stimulated by Mn2+ and strongly the enzyme inhibited by Zn2+; Mn2+ can also reactivate after inhibition by EDTA, but Zn2+ is ineffective. The N. crassa protein differs in this regard from mammalian and insect insulinases which are generally activated by both Mn2+ and Zn2+. This finding extends the apparent evolutionary conservation of these metal- and thiol-dependent proteases into the microbial realm. 0 lsea Academic Press, Inc.
The insulinase family of enzymes consists of metaland thiol-dependent proteases that have previously been found in the cells of mammals (1) and of Drosophila melanogaster (2). As purified and cloned from both sources, insulinase (insulin-degrading enzyme) is a llO-kDa cytoplasmic protein bearing homology to an Escherichia coli periplasmic protease named protease III (3-5). The purified protein from both sources degrades insulin with i Supported by Grant DK 39502 from the National Health. * To whom correspondence should be addressed. 0003~9861/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Institutes
of
characteristic specificity (6). It is thought to play a role in regulating the magnitude or extent of insulin-induced signals by inactivating cell-associated insulin, although conclusive proof of this is lacking. Recent reports from this laboratory have provided extensive evidence that the filamentous fungus Neurospora crassa possesses a functional insulin-induced signal transduction pathway. Addition of mammalian insulin to N. crassa cells results in stimulation of glucose metabolism, as measured by increased production of COZ, ethanol, alanine, and other metabolites, as well as enhanced glycogen synthesis (7-10). Stimulation of glycogen synthesis was shown to arise from the activation of the enzyme glycogen synthase from the glucose B-phosphatedependent form (D) to the independent form (1)-a conversion that results from insulin-induced dephosphorylation in mammalian cells (10). A specific, high affinity insulin-binding protein was purified from N. crassa membranes and was proposed to be the insulin receptor (11). Insulin-dependent protein phosphorylation was found to occur in N. crassa cells at serine, threonine, and tyrosine residues (12). Earlier reports described the presence of an insulin-like material in N. crassa medium and cells (13), and of effects of insulin on adenylate cyclase levels in N. crassa membranes (14). During the course of purification of the putative insulin receptor from N. crassa membranes, a contaminating insulin-binding protein which proved to be a bacitracininhibitable insulin-degrading protease was encountered (11). We now report further characterization and partial purification of this protein and discuss its resemblance to previously described insulinases. EXPERIMENTAL
PROCEDURES
insulin and iz61-Bz6-human insulin were Materials. ?-An-porcine purchased from NEN and Amersham, respectively. Ultrapure ammonium sulfate was obtained from BRL (Bethesda, MD) and HPLC grade acetonitrile from Baxter (Muskegon, MI). Disuccinimidyl suberate was purchased from Pierce (Rockford, IL). All other chemicals were pur199
200
KOLE,
SMITH,
chased from Sigma Chemical Co. (St. Louis, MO). MonoQ and Superose6 columns were obtained from Pharmacia and a Zorbax C8 column was from DuPont. Rat muscle insulin protease was a generous gift from Dr. Frederick G. Hamel, University of Nebraska Medical Center (Omaha, NE). Cells. N. crossa strain FGSC 4761 is a rapidly growing variant of the wall-less “slime” mutant FGSC 1118 (15). Cells were maintained and grown as described previously (12). Insulin degradation assay. Insulin-degrading activity was determined by the trichloroacetic acid (TCA)3 precipitation method as described (11) with modifications. Briefly, the enzyme was incubated in a 200-~1 reaction mixture containing 0.1 M Na,HPO, buffer (pH 7.5), 100 pg bovine serum albumin (BSA), and 800 pM “‘I-insulin. After 15 min at 37°C the reaction was stopped with 200 ~1 of 10% TCA and processed as described (11). During studies with divalent cations, 100 mM Hepes buffer, pH 7.4, was used. Enzyme activity was linear with respect to time and amount of enzyme protein added. Determination of optimal pH for insulin degradation. Optimal pH for insulin degradation was determined using Universal buffer (16) at different pH values. Purification of the insulin-degrading enzyme. The separation and purification were performed at 4”C, if not otherwise indicated. Cells were harvested, broken by Dounce homogenization, and 100,OOOgsupernatants (cytosol) were prepared as previously described (11) except that in this case we saved the cytosolic fraction insteadof the membrane fraction. The supernatant was fractionated by adding ammonium sulfate to 40% saturation, and the solution was stirred for 30 min and then centrifuged for 30 min at 15,000g. The pellet was discarded and the supernatant was adjusted to 65% saturation with ammonium sulfate, stirred, and centrifuged as above. The pellet was resuspended in 5 ml of 0.01 M potassium phosphate buffer, pH 6.5, at 23°C and dialyzed against the same buffer at 4°C for 18 h. The dialyzed solution was applied to a Mono& HR 5/5 anion-exchange column pretreated with 0.1 M TrisHCl buffer, pH 7.4. The column was washed with 50 ml of the same Tris-HCI buffer and the enzyme activity eluted with 30 ml of a linear gradient from 0 to 0.4 M NaCl in Tris-HCl buffer at a flow rate of 1 ml/ min. Active fractions were pooled, concentrated in a Centriprep 10, and loaded onto a Superose-6 column which was preequilibrated with 10 mM potassium phosphate buffer, pH 6.5. Enzyme activity was eluted with the same phosphate buffer and loaded directly onto the MonoQ column for a second time, as above. The chromatogram was developed using a 40-ml linear gradient (O-O.3 M NaCl) in the same Tris buffer at a flow rate of 0.5 ml/min. Fractions (0.5 ml) were collected in glass tubes containing 150 ~1 of 100% glycerol. Protein was determined according to Bradford (17). Affinity labeling of proteins with ‘251-insulin. Samples from the different purification steps were incubated with 2 ng of 1251-BZG-insulin for 1 h at 4°C. Disuccinimidyl suberate (0.4 mM) was added to the reaction mixture and incubated for another 15 min. The reaction was stopped by the addition of denaturing buffer and the samples were separated on a 10% SDS-polyacrylamide gel. Following SDS-PAGE, gels were dried and analyzed by autoradiography as previously described (11). Insulin-degrading activity on polyacrylamide gel. Samples from different steps of purification were fractionated by discontinuous polyacrylamide gel electrophoresis at pH 8.9 in gels containing 7% polyacrylamide, as described by Dewald et al. (18), omitting the detergent from the gel and electrode buffers. The gels were run at 4°C for 3 h at 5 mA constant current. At the end of the run, the gels were equilibrated with 0.1 M sodium phosphate buffer (pH 7.5) containing 1 mg/ml BSA. The gels were then sliced and each l-cm section was incubated with
3 Abbreviations used: TCA, trichloroacetic acid, BSA, bovine serum albumin; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PMSF, phenylmethylsulfonyl fluoride; NEM, N-ethylmaleimide; PCMB, p-chloromercuribenzoate.
AND
LENARD
1251-A’4-insulin at 37°C for 30 min. Degradation by the TCA solubility method.
of ‘251-insulin was assayed
HPLC analysis of “s1-insulin degradation products. ‘251-B26-human insulin was incubated with either rat muscle or N. crassa insulinase at 37°C in Tris-HCl buffer, pH 7.5, so as to reach a level of degradation of ca. 20% as measured by TCA precipitation. HPLC analyses of the insulin degradation products were carried out both before and after sulfitolysis following the method of Duckworth et al. (6). Fractions (0.5 ml) were collected in siliconized glass vials containing 10 pg of unlabeled insulin and counted directly in a LKB CompuGamma CS counter to determine the elution profile of radioactivity. Recoveries of the injected radioactivity ranged from 80 to 95% in the experiments shown.
RESULTS
We have previously reported (11) that N. crassa membranes possessed insulin-degrading activity which was purified away from a membrane-associated high affinity insulin-binding protein. When the distribution of this activity in a total N. crassa homogenate was investigated, however, it was found that 85% of the insulin-degrading activity was present in the 100,OOOgsupernatant, while only 15% remained associated with the membrane preparation. Furthermore, the membrane-associated activity was not enriched by purification or subsequent solubilization of the membranes. It was therefore concluded that the N. crassa insulin-degrading activity arises from a cytosolic enzyme which contaminated membrane preparations to some extent. The cytosolic insulin-degrading activity was partially purified in several steps, as shown in Table I. Modest enrichments of specific activity were obtained from a 4065% ammonium sulfate cut (step 2, Table I) and from MonoQ ion-exchange chromatography, with activity eluting between 0.19 and 0.29 M NaCl (step 3, Table I). Superose-6 chromatography (step 4, Table I) resulted in some further purification. Application of the resulting material to a second Mono& column (step 5, Table I) afforded major purification, with activity eluting at 0.1870.207 M NaCl (Fig. 1). Purification of 1491-fold was attained, with 18.9% yield. Attempts to purify the enzyme further using an insulin-affinity column were unsuccessful. The purified material did not cross-react with antiIDE antibody prepared against the Drosophila enzyme (kindly provided by Dr. Marcia Rich Rosner, University of Chicago, Chicago, IL; data not shown). Cross-linking of purified insulinase with 1251-insulin consistently labeled a polypeptide of ca. 120 kDa and showed several smaller bands that varied between experiments. Labeling of these bands was specifically competed by unlabeled insulin (data not shown). Analysis by disc electrophoresis showed that insulinase activity was restricted to a single, well-defined region of the gel, which remained invariant through several stages of purification (Fig. 2). Specificity of insulin cleavage by N. crassa insulinase was compared with that of a previously characterized insulin protease from rat muscle (kindly provided by Dr. F.
INSULINASE
IN Neurospora TABLE
Purification Total protein (mgl
Steps 1. cytoso1 2. 40-65% Ammonium sulfate 3. Mono& ion-exchange column, first time 4. Superose 6 5. MonoQ ion-exchange column, second time
of N.
crassa Insulinase Total activity (fmol)
63.4 306
2.52 0.69 0.0102
a fmole insulin degraded/min/mg
I
Specific activity”
80.5 15.75
201
crassa
Relative purification
5104 4820
1.0 4.8
1,168 2,756
2943 1902
18.4 43.5
94,502
964
Yield (%) 100 94 57.7 37.3
1491
18.9
protein.
Hamel, University of Nebraska Medical Center, Omaha, NE). Results are shown in Figs. 3 and 4. Extensive similarities are evident in the degradation products generated by the N. crassa and rat muscle insulinases, analyzed both before (Fig. 3) and after (Fig. 4) insulin chain separation by sulfitolysis. In order to characterize the N. crussa insulinase, a number of physical and kinetic parameters were determined. The optimal pH for the cytosolic N. crassa insulinase was around 7.5 (Fig. 5), similar to that found for the insulinases isolated from human fibroblasts (19), rat islets (20), and Drosophila (21). The degradation rate increased with increasing temperature up to 37°C (not shown) which is the standard assay temperature for mammalian and Drosophila insulinases (19, 21, 22). Effects of various agents on insulin degradation by intact N. crassa cells and by the purified enzyme are listed in Table II. Bacitracin was found to inhibit essentially all of the activity at 1 mg/ml, both of intact cells and of purified insulinase. Excess porcine insulin (10 PM), as expected, inhibited degradation of labeled material by both
purified insulinase and intact cells. Excess glucagon, on the other hand, was much less effective than excess insulin. The purified enzyme was unaffected by the serine protease inhibitor PMSF, soybean trypsin inhibitor, aprotinin, or the cathepsin inhibitor leupeptin (Table II). Chloroquine had no effect on the purified enzyme, and actually stimulated insulinase activity in intact cells. The N. crassa insulinase was strongly inhibited by 10 mM EDTA (Table II) and by l,lO-phenanthroline (Table III), suggesting that it is a metal-dependent protease. The effects of adding various divalent cations to N. crassa insulinase are shown in Fig. 6. The profile is quite unusual, and differs from other insulinases in that Mn2+ stimulated the activity while Zn2+ inhibited strongly. Lower concentrations of Zn2’ are very effective in inhibiting insulinase activity (Fig. 6, inset). Ca2+ and Mg2+ stimulated activity slightly. Consistent with these results, Mn2+ was the most effective of the four divalent cations tested at reversing inhibition by EDTA (Table IV). Ca2+ was partly effective,
I 70
70 I
Ammonium
adfat*
20
- 60 3
- 50 = /\
Protein
c
-40 :
: : :
’
m J - 30 .g
\ \ \
i
-20 \
‘.
\ .L’\-.
25
E
E .sJ
15
: e
10
: 0 5
- 10 0
30
Fraction
FIG. 1. Elution profile of insulin-degrading activity from the second MonoQ column (step 5, Table I). Twenty-microliter aliquots of each fraction were tested for degrading activity by the TCA-precipitation method, as described under Experimental Procedures.
0
0
3
5
5
10
Gel Slices
FIG. 2. Localization of insulin-degrading activity on a polyacrylamide disc gel. Samples from the ammonium sulfate and first and second Mono Q steps were fractionated by discontinuous polyacrylamide gel electrophoresis at 4°C and assayed for degrading activity as described under Experimental Procedures.
202
KOLE,
SMITH,
AND
19,
19 -
17 -
17 -16
-
I
15 H -
3 @ 's LI
10 -
13 -
1211 10 -
f!
9-
Q
a-
.G 't'
7 6-
-
g;-
1
5 1
E
4-
Q q
!j4-
3-
=
3-
15
20
25
30
35
40
45
50
55
60
65
70
75
-
14 -
IB E
9f
-
Q '5
8 8 a
14 t
B
16 -
A
16 -
-16
LENARD
60
0
5
1015202530354045505560657075601
Elutlon Time (min) FIG. 3. HPLC elution profile of degradation rat muscle insulinase (B).
5
Elutlon Tlmr (mln)
products of iz51-Bz6-human insulin produced by partial
digestion with N. crassa insulinase
(A) or
B A 21 -
24 4
s:
18 g17
I;;-
-
t
16 15 -
8
15 -
= 1 5 l's
14 -
12 -
s Ir: = 1 $ c 6
E
11 -
E
11 -
e
10 -
8
10 -
13 -
Q92 z 5 $6u a
a
6-
f
7-
3-
12 -
9-
g
6-
8 B 0
a
4-
13 -
d
=g
5-
14 -
7-
654-
1. :x ; ! LA!! a
f
2 .q 1 0
0
5
10
15
20
i LI
30
35
40
45
Xl
55
60
65
70
75
Elutlon Time (mln) FIG. 4. HPLC elution profile after sulfitolysis insulinase (A) or rat muscle insulinase (B).
60
30
’ 35
t 40
1 45
I 50
I 55
60
65
70
75
60
Elutlon Tlme (mln) of degradation
products of ‘261-B2G-human insulin produced by partial
digestion with N. crassa
INSULINASE
IN Neurosporu
crassa
203 TABLE
III
Effect of Sulfhydryl Inhibitors and 0-Phenanthroline on N. crassa and Rat Muscle Insulinase %Inhibition Inhibitors NEM 0.1 8
4
N. crassa insulinase
1.2 4 14 41
mM
0.5 mM
12
1.0
mM
5.0 mM
PH
FIG. 5. pH-dependence of insulin degradation by N. crossa insulinase. Insulin degrading activity of 8 pg of Superose 6 eluate (step 4, Table I) was determined in Universal buffer (16) adjusted to the indicated pH values.
PCMB 0.1 mM
72 95
0.5 mM 1.0 mM l,lO-phenanthroline
0.05mM while Mg2+ and Zn2+ were inactive. The N. crassa insulinase thus behaves as a Mn2+-requiring, Zn2+-inhibitable metal-dependent protease. The effect of sulfhydryl inhibitors N-ethylmaleimide (NEM) and p-chloromercuribenzoate (PCMB) on N. crassa and rat muscle insulinases were compared, as shown in Table III. In agreement with previous reports, both inhibitors gave >80% inhibition of the rat muscle insulinase at 0.1 mM, the lowest concentration tested. For N, crassa insulinase, on the other hand, higher concentrations of both inhibitors were required; NEM inhibited only 41% at 5 mM, the highest concentration tested, but complete inhibition was achieved with 1 mM PCMB (Table III). The N. crassa insulinase thus appears to be a thioprotease, but the essential sulfhydryl group is substantially less reactive than that in rat muscle insulinase. Analysis of the degradation of porcine insulin by the last Mono& eluant showed maximal activity at 280-400
TABLE
II
of degradation
0.5 mM
pmol/min/mg
protein,
f 1.7 f2 k3 +2
80 +- 1 97 t 2 99 * 1 100
f3 *2 100
19 48 86 97
0.1 mM 0.25 mM
Rat muscle insulinase
100 100 100
+3 29 + 0.4 fl
with 50% activity
at ca. 150 nM
insulin. DISCUSSION
The insulinase from N. crassa that is described in this report bears several significant similarities to insulinases purified from other sources: it is a bacitracin-inhibitable metal- and thiol-dependent protease (Tables II, III, IV, Fig. 6), probably of ca. 120 kDa, which cleaves insulin with
characteristic
specificity
(Figs. 3, 4). It differs
200 ,
1
Agents on Insulin Degrading Activity of N. crussa Cells and of Purified N. crassa Insulinase
Effect
from
previously reported insulinases in its specific metal requirement, and in its unusually low sensitivity to sulfhydryl inhibitors. A literature search failed to reveal any published examples of a Mn2+-stimulated, Zn2+-inhibited
of Different
% Degradation Cells
Inhibitors None Bacitracin, 1 mg/ml Porcine insulin, 10 pM EDTA, 10 mM Chloroquine, 1 mM PMSF, 1 mM Soybean trypsin inhibitor, Aprotinin, 0.5 mg/ml Leupeptin, 0.5 mg/ml Glucagon, 10 pM Dithiothreitol, 1 mM ’ Not determined.
1 mg/ml
49 0 12 + 3 31 + 3 56 k 3 69 k 2 ND” ND ND ND ND
Insulinase 36 -c 2 0 1 + 0.2 4 + 0.6 35 * 3 36 + 3 35 rt 3 36 + 2 34 -+ 1 22 f 2 37 AZ1
0
1
,
m* 0 05
2
3
.I .I5 .2 IZn *+ I. mM 4
.25 5
[M2+l, mM FIG. 6. Effect of divalent cations on insulin degradation by N. crassa insulinase. Insulin degrading activity was determined in the presence of different divalent cations as indicated, using 0.1 fig of the second MonoQ eluate (step 5, Table I). Inset, effect of lower concentrations of Zn2+ on insulinase activity.
204
KOLE. TABLE
SMITH,
None 12.5 mM EDTA 10 mM MnClr 10 mM CaCi, 10 mM MgClz 10 mM ZnCI,
LENARD
REFERENCES
IV
Reactivation of EDTA-Inhibited N. crassa Insulinase by Divalent Cations Additions
AND
1. Duckworth,
% Degradation
48 + 0.4 6.4 510.3 52 +2 25 k3 10 12
*1 * 0.4
protease; most other insulinases are stimulated by both Mn2+ and Zn2+ (1). Previous reports from this laboratory have provided evidence for the presence in N. crassa of an insulin-induced signal transduction system including a putative membrane-bound receptor (ll), numerous insulin-induced changes in protein phosphorylation at serine, threonine, and tyrosine residues (12), and enhancement of glucose metabolism, including activation of glycogen synthase (7, 8, 10). The presence in the same cells of an enzyme resembling the insulinases of other insulin-responsive cells raises the question of its functional role. Although no function has been established for insulinase in any cell type, several observations have raised the possibility that it might have a role in insulin regulation. Treatment of human hepatoma cells (23) or opossum kidney epithelial cells (24) with bacitracin or other inhibitors of insulinase resulted in reduced degradation of added insulin and other growth factors. Insulin degradation in hepatoma cells was decreased 18-54% by microinjection of antibodies directed against insulinase (25), and insulin could be cross-linked to insulinase in these cells (26). These findings indicate that insulinase can act on insulin added to intact cells in culture, but provide no evidence as to the functional significance of this action. It has been reported that insulinase varies in amount at different developmental stages in D. melanogmter (27), and that insulinase might be involved in differentiation of muscle cells (28), but the relationship of these properties to insulin degradation has not been demonstrated. The finding of a metal- and thiol-dependent protease with properties of insulinase in microbial cells extends further the evolutionary range in which these enzymes are known to exist, and may open possibilities for genetic studies of function that are not feasible in other systems. This finding is also consistent with a recent suggestion that a superfamily of insulinase-like enzymes, which include the Mn2+-activated mitochondrial processing proteases (29), exists.
W. C. (1988) Endocr. Reu. 9,319-345.
2. Rosner, M. R. (1990) Mol. Reprod. Dev. 27, 54-59. 3. Atfholter, J. A., Fried, V. A. and Roth, R. A. (1988) Science 242, 1415-1418.
4. Affholter,
J. A., Hsieh, C. L., Francke, Mol. Endocrinol. 4, 1125-1135.
U., and Roth, R. A. (1990)
5. Kuo, W. -L., Gehm, B. D., and Rosner, M. R. (1990) Mol. Endocrinol. 4,1580-1591.
6. Duckworth,
W. C., Garcia, J. V., Liepnieks, J. J. Hamel, F. G. Hermodson, M. A., Frank, B. H., and Rosner, M. R. (1989) Biochemistry
28,2471-2477. 7. McKenzie, M. A., Fawell, S. E., Cha, M., and Lenard, J. (1988) Endocrinology 122,511-517. 8. Greenfield, N. J., McKenzie, M. A., Adebodun, F., Jordan, F., and Lenard, J. (1988) Biochemistry
27,8526-8533.
9. Greenfield,
N. J., Cherapak, C. N., Adebodun, F., Jordan, F., and Lenard, J. (1990) B&him. Biophys. Actu 1025, 15-20.
10. Fawell, S. E., McKenzie, M. A., Greenfield, N. J., Adebodun, Jordan, F., and Lenard, J. (1988) Endocrinobgy 122,518-523. 11. Kole, H. K., Muthukumar,
F.,
G., and Lenard, J. (1991) Biochemistry
30,682-688. 12. Kole, H. K., and Lenard, J. (1991) FASEB J. 5,2728-2734. 13. LeRoith, D., Shiloacb, J., Roth, J., and Lesniak, M. A. (1980)
PFOC.
Natl. Acad. Sci. USA 77,6184-6188. 14. Flawia, M. M., and Torres, H. N. (1973) J. Biol. Chem. 248,4517-
4520. 15. Scarborough, G. A. (1985) Exp. Mycol. 9, 275-278. 16. Davison, R. M. C., Elliot, D. C., Elliot, W. H., and Jones, K. M. (Eds.) (1978) Data for Biochemical Research, 2nd ed., pp. 485, Clarendon, Oxford. 17. Bradford, M. M. (1976) Anal. B&hem. 72,248-254. 18. Dewald, B., Dulaney, J. T., and Touster, 0. (1974) in Methods in Enzymology (Fleischer, S., and Packer, L., Eds.), Vol. 32, pp. 8291, Academic Press, San Diego. 19. Stentz, F. B., Harris,
H. L., and Kitabchi,
E. (1985) Endocrinology
116,926-934. 20. Bhatnena,
S. J., Timmers, K. I., Oie, H. K., Voyles, N. R., and Recant, L. (1985) Diabetes 34, 121-128. 21. Garcia, J. V., Fenton, B. W., and Rosner, M. R. (1988) Biochemistry
27,4237-4244. 22. Duckworth, W. C., Heinemann, 23. 24. 25. 26.
M. A., and Kitabchi, A. E. (1972) PFOC.Natl. Acad. Sci. USA 60, 3698-3702. Gehm, B. D., and Rosner, M. R. (1991) Endocrinology 128, 16031610. Dahl, D. C., Tsao, T., Duckworth, W. C., Frank, B. H., and Rabkin, R. (1990) Diabetes 39, 1339-1346. Shii, K., and Roth, R. A. (1986) Proc. N&Z. Acad. Sci. USA 83, 4147-4151. Hari, J., Shii, K., and Roth R. A. (1987) Endocrinobgy 120, 829-
831. 27. Stoppelli,
M. P., Garcia, J. V., Decker, S. J., and Rosner, M. R. (1988) PFOC.Natl. Acad. Sci. USA 85, 3469-3473. 28. Kayalar, C., Wong, W. T., and Hendrickson, L. (1990) Biochemistry
44,137-151. 29. Rawlings, N. D., and Barrett, 391.
A. J. (1991) Biochem. J. 275,
389-
