Professional Documents
Culture Documents
Markland 1971
Markland 1971
I . Introduction . . . . . . . . . . . . . 5 6 2
I1. Historical Background and Development . . . . . .562
I11. Physical. Chemical. and Stability Properties of the Subtilisins . 564
A . Subtilisin Carlsberg . . . . . . . . .564
B . Subtilisin BPN' . . . . . . . . . . 5 8 5
C . Subtilisin Novo . . . . . . . . . .566
D . Subtilisin Amylosacchariticus . . . . . . . 5 6 6
IV . Primary Structure of the Subtilisins . . . . . . . 567
A . General Comparison . . . . . . . . .567
B . Subtilisin B P N . . . . . . . . . .567
C . Subtilisin Novo . . . . . . . . . . 569
D . Subtilisin Carlsberg . . . . . . . . . .569
E . Subtilisin Amylosacchariticus . . . . . . .569
F. Comparison of Sequences . . . . . . . .571
V. Active Site Studies . . . . . . . . . . .575
A . Serine . . . . . . . . . . . . .576
B. Histidine . . . . . . . . . . . . 580
VI . Substrate Specificity and Enzymic Properties . . . . . 5 8 4
A . Protein and Peptide Substrates . . . . . . .584
B . Synthetic Substrates . . . . . . . . .586
C . Transesterification and Transpeptidation . . . . .593
D . Mechanism of Action of the Subtilisins . . . . . 593
VII . Chemical Modification Studies . . . . . . . . .596
A . Lysine Modification . . . . . . . . .596
B. Methionine Modification . . . . . . . .598
C . Tyrosine Modification . . . . . . . . . 599
VIII . Inhibitors: Dye Binding . . . . . . . . . . 602
IX . Other Bacillus Alkaline Proteinases . . . . . . .605
561
562 F. S. MARKLAND, JR., AND E. L. SMITH
1. Introduction
Since the last review o i this subjcct appeared in this treatise (1) con-
siderable progress has been made in our understanding of both the struc-
ture and properties of bactcrial proteinases. This chapter will review
only the subtilisins: the diisopropylfluorophosphate (DFP)-sensitive,
extracellular, alkaline proteinaees from Bacillus s u b t i h and other species
of Bacillus. Since most o f the recent work has been on the subtilisins,
major emphasis will be placed on these enzymes with mention of other
alkaline proteinases as applicable. A separate chapter ( 2 ) reviews the
X-ray structure of subtilisin BPN. Other bacterial proteinases are dis-
cussed separately (3).
A. SUBTILISINCARLSBERG
TABLE I
OF PHYSICOCHEMICAL
COMPARISON PROPERTIES
OF THE SUFITILISINS
Amylosac-
BPN' Carlsberg Novo chariticus
Propertmy (18-20) (7, 17, 21,221 (9,141 (10,23, 24)
17. M. Ottesen and C. G. Schellman, Compt. Rend. Trav. Lab. Carlsberg, Ser.
Chim. 30, 157 (1957).
18. H. Matsubara, C. B. Kaaper, D. M. Brown, and E. L. Smith, JBC 240, 1125
( 1965).
16. SUBTILISINS 565
TABLE I1
AMINOACID COMPOSITIONS OF SUBTILISINS
Carlsberg. Amylosacchariticusc
Amino acid BPN’(I (19) (a1) Novob (14) (94.6)
Lysine 11 9 11 8
Histidine 6 5 6 6
Arginine 2 4 2 4
Tryptophan 3 1 3 3
Aspartic acid 11 9 11
26
Asparagine 17 19 17
T hreonine 13 19 13 16
Serine 37 32 37 40
Glutaniic acid 4 3 4
15
Glutamine 11 7 11
Proline 14 9 14 13
Glycine 33 35 33 33
Alanine 37 41 37 30
Valine 30 31 30 25
Methionine 5 5 5 4
Isoleucine 13 10 13 16
Leucine 15 16 16 15
Tyrosine 10 13 10 12
Phenylalanine 3 4 3 3
Cysteine, cystine 0 0 0 0
Total 275 274 275 275
B. SUBTILISIN
BPN’
and other salts (11). Thc proteinase was shown to be a single polypep-
tide chain of molecular weight 27,600 having an isoelectric point of 7.80
(18).The amino acid composition is given in Table I1 (19, 22) and the
amino- and carboxyl-terminal sequences of the D I P enzyme were identi-
fied as NH,--Ala-Glx and Ala-Gln-COOH, respectively ( 2 0 ) . These re-
sults agree with carlier reports by the Japanese workers who showed that
the enzyme had a molecular weight of about 30,500 with alanine as the
amino-terminal residue (25).
C. SUBTILISINNovo
D. SUBTILEISAMYLOSACCHARITICUS
A. GENERAL
COMPARISON
B. SUBTILISIN
BPN’
The sequence of subtilisin BPN’ was deduced from studies of the tryp-
tic ($53,chymotryptic (36),peptic (37), and cyanogen bromide digests
(19). The complete sequence is shown in Fig. 1.
The protein consists of a single polypeptide chain of 275 residues de-
void of any disulfide bridges, There is no apparent homology with the
sequences of the pancreatic proteinases (16). The serine residue reactive
Th r Ile Pro Leu Asp Lvs Val Gln Ala PheLys Ala
-
NH -Ala-Gln-Ser-Val-Pro-Tyr-Gly-Val-Ser-Gln~lle-Lys-Ala-Pro-Ala-Leu~His-Ser-Gln-Gly-Tyr-Thr-Gly-Ser-Asn-Val~Lys-Val-Ala-Val~
2 10 a 30
Leu Thr Gln Ala Am Val Phe Ala Gly Ala Tyr Asn Thr
Ile-Asp-Ser-Gly-l le-Asp-Ser-Ser-His-Pro-Asp-Leu-Lys-Val-Ala-Gly-Gly-Ala-Ser-Met-Val-Pro-Ser-Glu-Thr-Pro-Asn~Phe~Gln-Asp-
40 m m
Gly Gly ASP lhr Thr Val Ser
Asp-Asn-Ser-His-Gly-Thr-His-Val-Ala-Gly-Thr-Val-Ala-Ala-Leu-Asn-Asn~Ser-l le-Gly-Val-Leu-Gly-Val-Ala-Pro-Ser-Ser-A~-~u-
70 m 9Ll
A m Ser Ser Ser Gly Val Ser Thrlhr Gly
Ty~-Ala-Val-Lys-VaI-Leu-Gly-Asp-Ala-Gly-Ser-Gly-Gln-Tyr-Ser-Trp-Ile-I le-Asn-Gly-l le-GI uTrp-Ala-lle-Ala-Asn-Asn-hret-Arp-
Im 110 120
Ala Thr Me1 Gln Am Tyr Arg
Val-I le-Asn-Met-Ser-Leu-Gly-Gly-Pro-Ser-Gly-Ser-Ala-Ala-Leu-Lys-Ala-Ala~Val-Asp-Lys-Ala-Val-Ala-Ser-Gly-Val-Val-Val-Val~
130 la im
Ser Asn Ser Thr Asn Ile Ala Asp
Ala-Ala-Ala-Gly-Asn-Glu-Gly-Ser-Thr-Gly-Ser-Ser-Ser-Thr-Val-Gly-lyr-Pro-6ly-Lys-Tyr~Pr~Ser-Val-l le-Ala-Val-Gly-Ala-Val-
1M 170 180
Asn Ser Asn Ala Glu Ala Gly Val Tyr 1~r
Asp-Ser-Ser-Asn-Gln-Arg-Ala-Ser-Phe-Ser-Ser-Val-Gly-Pro-Glu-Leu-Asp-VaI-Met-Ala-Pro-Gly-Val-Ser-l le-GIn-Ser-Thr-Leu-Pro-
190 200 210
Thr Thr Ala Thr Leu
Gly-Asn-Lys-Tyr-Gly-Ala-Tyr-Asn-Gly-Thr-Ser-Met-Ala-Ser-Pm-His-Val-Ala-Gly-AIr-Ala-Ala-Leu-I le-Leu-Ser- Lys-His-Pro-Asn-
220 rm 24l
l e u Ser Ala Ser Asn Arg Ser Ser Ala Tyr Ser
Trp-Thr-Asn-lhr-Gln-Val-Arg-Ser-Ser-Leu-Gln-Asn~hr~Thr-Thr-Lys-Leu-Gly-As~-Ser-Phe-Tyr-Tyr-Gly-Lys-Gly-Leu- Ile-Asn-Val-
2% 2M 210
Glu
Cln-Ala-Ala- Ala-GlnCOOH
71%
FIG.1. Comparison of amino acid sequence of subtilisins BPN' and Carlsberg
(42).The continuous sequence is that of subtilisin BPN. The residues which differ
in subtilisin Carlsberg are given above the corresponding residue. The dash at
residue 56 indicates that the corresponding residue is lacking in the Carlsberg en-
zyme. The active Ser 221 is denoted by asterisk.
with DFP is a t position 221 and the sequence around that serine (Thr-
Ser+-MetrAla), as originally pointed out by Sanger and Shaw (38),is
different from that around the reactive serine in the mammalian pan-
creatic proteinases (see Section V,A) .
There are several sequences in the polypeptide chain which appear to
be repeated. These may have some relationship to the evolutionary his-
tory of the subtilisins and will be discussed in Section IV,F.
Although the possible significance is presently unknown there are
many di-, tri-, and, in one case, tetrapeptide (residue 147 through 150,
Fig. 1) repetitions of the same residue. The residues appearing in these
repeating sequences are Ala (six Ala-Ala sequences, three of which are
Ala-Ala-Ala) , Val (one Val-Val-Val-Val sequence), Ser (six Ser-Ser
sequences, including one Ser-Ser-Ser) , Thr (one Thr-Thr-Thr se-
quence), Gly (two Gly-Gly sequences), Ile (one I l e I l e sequence), and
C. SUBTILISINNovo
D. SUBTILISINCARLSBERG
E. SUBTILISIN
AMYLOSACCHARITICUS
Ala,Ala,Ala.Gly.Asn,Glu,Gly.Ser.Ser.Gly.Ser.Ser.Ser.Thr~Val.Gly,Tyr~Pro.~)Lys(Tyr.Pro,Ser~Val~Ile..~la~Thr~Gly~Ala.Yal.
Ala-Ala-Ala-Gly-.~n-Glu-Gly-Ser-Thr-Gly-Ser-Ser-Ser-Thr-Val-Gly-Tyr-Pro-Gly-Lys-Tyr-Pro-Ser-~al-Ile-.4la-Val-GIy-.4la-Yal-
160 170 180 r
Asp,Ser.Ser,.4sn,Gln)Arg(Ala,Ser,Phe.Ser.Ser.Ser,Gly,Ala.Glu.Leu.ilsp,Yal)Jlet
(.Ua.Pro.Gly )(Yal,Ser)Ile-Gln-Ser-Thr-1,eu-Pro- P
~aySer-Ser-~n-Gln-~g-Ala-Ser-Phe-Ser-Ser-V~-Gly-R~Gl11-I~eu-.4sp-Val-.~et-Ala-Pro-Gly-~al-~~r-Ile-~~In-Ser-Thr-I.eu-Pro-
190 200 210
~Val.Ala.Gly,Ala.Ala,Ala.Leu)(Ile,Leu,Ser)Lys(His.Pro.Asn.
Gly-Gly-Thr-Tyr-Gly-.4Ls-Tyr(Asn,Gly.Thr,Ser*)~~et(Ala.Thr.Pro.His)
Gly- Asn-Lys-Tyr-Gly-Ala-Tyr-Asn-Gly-Thr-Ser*-~let-Als-Ser-Pro-His-Val-Ala-Gly-Ala-A1a-.4la-l~eu-Ile-I~u-Ser-I,y~-His-Pro-Asn-
220 230 240
Trp)(Thr.hsn ) (Ala.Gln,Val)Arg(Asn,Arg)
(Leu,Gln,Ser,Thr,Ala,Thr.?lr)
(Leu.GIy.hsp)(Ser.Phe.Tyr.Tyr.C.ly)I.ys(Gly.I.eo. Ile..bn.Val
Trp-Thr- ~bn-Thr-Gln-Val-Ara-Ser-Ser-Leu-Gln-Asn-Thr-Thr-Thr-Ly~l~ii-(;ly-.~p-Ser-Phe-Tyr-Tyr-C.Iy-l.ys-(;ly-Leu-Ile-.49n-Ya1-
250 260 270
Gln)Ala-Ala-Ah-GlnCOOH
Gln-Ala-Ala-Ala-GlnCOOH
275 m
FIG.2. Comparison of amino acid sequence of subtilisins RPN’ and Amylosacchariticus (944).The lower se- r
quence is that of BPN. Residues in subtilisin Amylosacchariticus whose sequences have not been deter-
mined are enclosed in parentheses. However, due to the extensive homology the arrangement of many oi
the residues within the parentheses is probably correct. The active His 64 and Ser ‘221 are denoted by aster- x
isks. Apparent differences are indicated in boldface. The two enzymes apparently differ by at least 28 residues.
16. SUBTILISINS 571
F. COMPARISON
OF SEQUENCES
46. F. S. Markland, E. Shaw, and E. L. Smith, Proc. Natl. Acad. Sci. U. S . 61,
1440 (1968).
47. C. S. Wright, R. A. Alden, and J. Kraut, Nature 221,235 (1969).
48. D. M. Blow, J. J. Birktoft, and B. S. Hartley, Nature 221, 337 (1969).
49. E. L. Smith, “The Enzymes,” 3rd ed., Vol. I, Chapter 6,1970.
572 F. S. MARKLAND, JR., AND E. L. SMITH
TABLE I11
SEQUENCES
AROUND THE ASPARTYLRESIDUESIN ACTIVE SITESO
Enzyme Sequence'
4 Residue numbers are below the sequence using the numbering of chymotrypsino-
gen A for the mammalian proteinases. The active aspartic acid residue is marked by
an asterisk.
* For identificationin mammalian enzymes,see Hartley el al. (34,48);in the subtilisins,
see Smith et u2. (4S),Wright et al. (47), and Fig. 2.
TABLE IV
AMINOACID SUBSTITUTIONS
IN SUBTIUSINS
Subtilisin BPN'
which show no substitution. One of these occurs around the reactive Ser
221 extending from residues 218 through 240, and another is around the
reactive His 64 extending from residues 64 through 75. As noted above,
there is a conservative substitution of Ser 224 by Thr in subtilisin
Amylosscchariticus.
Comparison of the sequences of subtilisins B P N and Carlsberg shows
the presence of similar sequences in different portions of the molecules.
These are obvious in either subtilisin individually but are even more ap-
parent when both protein sequences are considered together (Fig. 3 ) . In
addition to those shown in the figure, there are a number of tripeptide
(A)
67 75
Hie-Val-Ala-Gly-T hr-Val-Ala-Ala-Leu
His-Val-Ala-Gly - Ala-Ala-Ala-Leu
226 229 230 233
(B)
82 Val-Ser
Leu-Gly-Val-Ala-Pro-Ser-Ser-Ala-Leu-Tyr-Ala- Val-Lya
Leu-Gly - Gly-Pro-Ser-Clly-Ser-Ala - Ala-Leu-Lya
126 127 128Ala Thr 134 Met 136
(C)
88
77 Thr Thr Val
Aan-Ser-Ile-Gly-Val-Leu-Gl y-Val- Ala-Pro-Ser-Ser
Ser-Thr- Val-Gly-Tyr-Pro-Gly-Lys-T yr-Pro-Ser-Val
Aan Ile Ala Asp 174
163
(D) (E)
45
39 42 Val 49
HbPro-Asp-Leu Ala-Gl y-Gly-Ala-Ser
HbPro-Am-Trp Leu-Gly-Gly-Pro-Ser
238 Leu 126 Ala 130
241
(F) (GI
73 A8p 77 167 Ah 171
Ala-Ala-Leu-Aan-Aen Tyr-Pro-Gly - Lye-Tyr
Ala-Ala-Val-AspLys Leu-Pr o-Gly- Am-Ly a-T yr
Gln A8n Tyr Thr Thr 214
137 141 209
FIG.3. Segments of the sequences of subtilisins B P N and Carlsberg which ap-
pear to involve repetition of similar sequences (4.9). Residues shown in boldface are
identical; residues shown in italics would appear to represent conservative replace-
ments. The continuous sequences are BPN' or both subtilisins; residues above or
below these sequences represent differences in subtilisin Carlsberg.
16. SUBTILISINS 575
A. SERINE
the same sequence around the reactive serine as the mammalian protein-
ases, whereas proteinases from other organisms and plants resemble the
subtilisins ; e.g., those from Aspergillus oryzae ( 6 3 ) , Aspergillus flavus
(64), and a caseinase from French beans ( 6 5 ) . Indeed, other sequences
have been reported for other DFP-sensitive proteinases, e.g., a proteinase
from baker's yeast and phaseolin of French beans possess the sequence
Glu-Ser"-Val (65) and a proteinase from Arthrobacter has the sequence
Ser-Ser"-Gly (66).Table V shows the presently known sequences around
the reactive serine residue of various DFP-sensitive proteinases.
In view of the substantial variations in these sequences, it would
appear that the residues in the sequences near the active serine residue
may have little influence on the reaction mechanism, the reactivity of
the serine hydroxyl, or in determining the specificity of the enzyme;
however, these sequences are useful in classifying the various types of
enzymes.
TABLE V
AROUND REACTIVE
SEQUENCES SERINERESIDUES"
Source Sequenceb
ported for acetylcholine esterase (68) and for liver aliesterase (69).
e
0
Subtilisin-SH
+ HIO subtilisin-& -CHs
- CHGOz-
Titration with p-mercuribenzoate (7.9) or 5,5’-dithiobis-2-nitrobenzoic
acid ( 7 4 ) , as well as amino acid analysis, indicated the formation of
0.7-0.9 mole of cysteine per mole subtilisin. No other changes could be
detected by sedimentation analysis, starch gel electrophoresis, fluores-
cence spectroscopy, ultraviolet absorption, or tryptophan reactivity.
However, a small change in the optical rotatory dispersion pattern was
observed (76) .
Study of thiolsubtilisin showed that it had about 33% of the initial
activity of native subtilisin toward p-nitrophenylacetate ( 7 1 ) .Thiolsub-
tilisin also hydrolyzes N-trans-cinnamoylimidazole (76) and the p-ni-
trophenyl esters of several amino acid derivatives although at drastically
reduced rates compared to the native enzyme (75). No activity has
been detected for the thiol enzyme against other ester, amide, or peptide
substrates of subtilisin.
Several findings suggest that the reactions catalyzed by thiolsubtilisin
may be enzymic in character: (1). The intermediate in the hydrolysis
of N-trans-cinnamoylimidazole has the spectral characteristic of a thiol
enzyme ; (2) the hydrolysis of p-nitrophenylacetate follows Michaelis-
Menten kinetics; (3) the hydrolysis of the p-nitrophenyl esters is inhib-
68. D. C. Shaw, unpublished work (1963), cited by F. Sanger, Proc. Chem. SOC.
p. 76 (1963).
69. H. S. Jansz, C. H. Posthumus, and J. A. Cohen, BBA 33,396 (1959).
70. L. PolgBr and M. L. Bender, JACS 88,3153 (1966).
71. K. E. Neet and D. E. Koshland, Jr., Proc. Natl. Acad. Sci. U . S. 56, 1606
( 1966).
72. D. E. Fahrney and A. M. Gold, JACS 85,997 (1963).
73. P. D. Boyer, JACS 76,4331 (1954).
74. G. L. Ellman, ABB 82, 70 (1959).
75. K. E. Neet, A. Nanci, and D. E. Koshland, Jr., JBC 243,6392 (1968).
76. L. Polgtir and M. L. Bender, Biochemistry 6,610 (1967).
16. SUBTILISINS 579
reacts only with the cysteine in thiolsubtilisin (80)and does not bridge
to the active His 64 (46) that, a t least in subtilisin BPN’, is close to
the active serine ( 4 7 ) .
B. HISTIDINE
a
0 700 -@)
60 64 68 72 76 80
PH
FIG.4. Variation of VmaTwith pH for (A) the hydrolysis of benzoyl-L-arginine
ethyl ester by subtilisin Novo and for (B) the hydrolysis of tosyl-L-arginine methyl
ester by subtilisin Amylosacchariticus. Reactions were at 37” with automatic addi-
tion of 0.02 N base in a pH stat. Initial velocities were determined from the first
5% of the hydrolysis and VmaXwas obtained from the Lineweaver-Burk plot (86).
A molecular weight of 27,600 was assumed. Data were taken from Myers and Glazer
(87). Results similar to those with subtilisin Novo were also obtained with sub-
tilisins BPN’ and Carlsberg (39).
residue, direct chemical evidence was lacking. Specific reagents used for
labeling the histidine residues of trypsin (89) and chymotrypsin (90)
were found to be ineffective in inactivating the subtilisins ( 1 6 ) . Further-
more, other alkylating agents such as bromopyruvate, bromoacetate,
iodoacetate, and iodoacetamide were without effect over a broad range
of pH (24, N u ) .
Shaw and Ruscica (91) synthesized a highly reactive phenylalanine
derivative similar to the chymotrypsin inhibitor, tosyl-L-phenylalanine
chloromethyl ketone. The new reagent, benzyloxycarbonyl-L-phenylala-
nine bromomethyl ketone (ZPBK), was shown to react. stoichiometrically
89. E. Shaw, M. Mares-Guia, and W. Cohen, Biochemistry 4, B19 (1965).
90. G. Schoellmann and E. Shaw, Biochemktry 2, 252 (1963).
90a. E.L. Smith, F. S. Markland, and A. N. Glazer, in “Structure-Function Re-
lationships of Proteolytic Enzymes” (P. Desnuelle. H. Neurath, and M. Ottesen,
eds.), p. 160. New York, 1970.
91. E.Shaw and J. Ruscica, JBC 243, 6312 (1988).
582 F. S. MARKLAND, JR., AND E. L. SMITH
TABLE VI
AROUND REACTIVE
SEQUENCE HISTIDINE
RESIDUES"
Enzyme Sequenceb
a The residue numbers are under the sequence; for the mammalian proteinases the
numbering system is that of chymotrypsinogen A. The reactive histidine residue is
marked by an asterisk.
b For review of references for the pancreatic proteinases, see Smillie and Hartley (92)
and for the subtilisins, see Markland el al. (46).
Time (hours)
Since the discovery of the subtilisins, it has become evident that these
enzymes manifest a broad specificity as proteinases. During purification
of the Carlsberg enzyme esterase activity was observed on methyl
butyrate, as well as proteinase activity on casein, hemoglobin, oval-
bumin, and gelatin. The pH optimum for casein digestion was in the
range from p H 10 to 11 ( 7 ) . Graae (93) showed that the esterase and
proteinase activity was associated with the same enzyme.
Okunuki et al. (94, 95) reported that subtilisin BPN’ also hydrolyzed
a variety of proteins, including casein, hemoglobin, gelatin, and oxidized
lysozyme. Cleavage appeared to occur mainly a t the NH,- and COOH-
terminal sides of neutral and acidic amino acid residues ( 9 4 ) . In addi-
tion, subtilisin BPN’ was shown to have esterase activity with a pH
optimum around 8.5-9.5 (11).
A. PROTEIN SUBSTRATES
AND PEPTIDE
with the specificity for aromatic or apolar residues obtained with small
ester and amide substrates (@,100).
Reinvestigation of the subtilisin-catalyzed cleavage of insulin under
restricted conditions yielded more definitive information concerning the
specificity of subtilisin. Morihara and Tsuzuki (100) found that after
digestion of the oxidized insulin B chain for 2-3 hr with 0.1% by weight
of subtilisin B P N a t room temperature and pH 9 cleavage occurred
only a t peptide bonds 6 5 (Gln-His) , 9-10 (Ser-His), 15-16 (Leu-Tyr),
16-17 (Tyr-Leu), and 25-26 ( P h e T y r ) . This is consistent with the
findings of these authors that hydrolysis of small ester and amide sub-
strates such as acetyl-X-ethyl ester and carbobenzoxy (Cbz)-Gly-X-
NH, occurs a t the carboxy-terminal side of residue X (where X rep-
resents aromatic or apolar residues such as L-tyrosine, L-phenylalanine,
and L-leucine) . Earlier, Ottesen and co-workers (101) compared the
specificities of subtilisins Carlsberg and Novo after hydrolysis of the B
chain of oxidized insulin a t 30' and pH 8 for 2 hr a t very low enzyme-
to-substrate ratios. Digestion by subtilisin Carlsberg a t an enzyme to
substrate ratio of 1 :3960 produced 12 peaks after separation of the pro-
ducts of digestion on Dowex 50-X4. The two major peptides, however,
obtained in about 50% yields were composed of residues 1-15 and 16-30,
indicating that the major point of cleavage was a t the Leu-Tyr bond
(residue 15-16) of the oxidized insulin B chain. When the enzyme con-
centration was increased, peptide segments 1-15 and 16-30 disappeared
and were replaced by fragments representing hydrolysis a t residues
4-5 (Gln-His), 9-10 (Ser-His), and 11-12 (Leu-Val) for fragment
1-15, and a t residues 16-17 (Tyr-Leu), 17-18 (Leu-Val), and 26-27
(Tyr-Thr) for fragment 16-30.
Similar results were obtained with subtilisin Novo, but the relative
rates of hydrolyses were different. Thus peptide bonds 9-10 (Ser-His)
and 26-!27 (Tyr-Thr) were hydrolyzed faster by subtilisin Novo than
by subtilisin Carlsberg while the opposite was the case with bonds 4.5
(Gln-His) and 11-12 (Leu-Val) .
Ottesen and gstergaard (102)also observed slight differences between
subtilisins Novo and Carlsberg in the conversion of ovalbumin to plak-
albumin. At an enzyme-to-protein ratio of one to l0,OOO a t 30" and pH
8 both subtilisins Novo and BPN' acted in identical fashion on oval-
bumin, as judged by the solubility of the plakalbumin formed, but
100. K. Morihara and H. Tsuruki, ABB 129, 620 (1989).
101. J. T. Johansen, M. Ottesen, I. Svendsen, and G. Wybrandt, Compt. Rend.
Truv. Lab. Curlsberg 36, 365 (1968).
102. M. Ottesen and B. Ostergaard, Compt. Rend. Trav. Lab. Carlsberg 34, 187
(1964).
586 F. S. MARKLAND, JR., AND E. L. SMITH
B. SYNTHETIC
SUBSTRATES
103. F. M. Richards, Compt. Rend. Trav. Lab. Carlsberg, Ser. Chim. 29, 329
( 1955).
104. F. M. Richards and P. J. Vithayathil, JBC 234,1459 (1959).
105. F. M. Richards, Compt. Rend. Trav. Lab. Carlsberg, Ser. Chim. 29, 322
(1955).
106. M. Ottesen and M. Szhkely, Compt. .Rend. Trav. Lab. Carlsberg 32, 319
(1962).
107. G. Gordillo, P. J. Vithayathil, and F. M. Richards, Yale J . Bwl. Med. 34,
582 (1962).
108. M. 5. Doscher and C. H. W. Hirs, Biochemistry 6, 304 (1987).
109. E. Gross and B. Witkop, Biochemistry 6, 745 (1967).
16. SUBTILISINS 587
110. J. T.Johansen, R. W.A. Oliver, and I. Svendsen, Compt. Rend. Trav. Lab.
Carkrberg 37, 87 (1969).
TABLE V I I
KINETICCONSTANTS
FOR SUBTILISIN-CATALYZED
HYDEOLYSIS ACID ESTERS~
OF N-ACETYUMINO
Subtilisin
Substrate
N-Acetyl-btyrosine ethyl ester 0.146 548 0.07 731 0.09 1316 0.0222 383.3
N-Acetyl-ctyrosine methyl ester 0.046 281 0.09 1560 0.07 1930
N-Acetyl-btryptophan methyl ester 0.031 149 0.09 415 0.05 820
N-Acetyl-cphenylalanine methyl ester 0.044 93.9 0.06 415 0.03 765
N-Benzoyl-carginine ethyl ester' 0.012 19.8 0. 007b 3.9 0.007 16.1 0.010 3.1
N-Tosyl-barginine methyl ester' 0.042 17.7 0.07 15.5 0.044 68.6
N-Acetyl-bvaline methyl ester 0.28 28 0.19 23 0 F
N-Acetylglycine ethyl ester 0 m
N-Acetyl-calanine methyl ester 0.123 72.0
N-Acetyl-bnorvaline ethyl ester 0
N-Acetyl-bleucine methyl ester 0.0666 57.5
N-Acetyl-cphenylalanine ethyl ester 0.0166 30.6
N-Acetyl-btryptophan ethyl ester 0.0238 35.8
N-Acetyl-clysine methyl ester 0.0909 47.4
4
pl
a The reaction mixture contained, initially, 0.0025 to 0.25 M substrate in 5.0 ml of 0.1 M KC1 containing 8% dioxane of volume.
The suhtilisin concentration was 0.0028 to 0.4 mg/ml. Titrations were performed a t pH 8.0, 37", with 0.02 M base as titrant.
b Data from Myers and Glazer (87).
c Data from Glazer (39) and Bare1 and Glazer (40).
Data from Morihara and Tsuzuki (100).Conditions differed as follows: measurements at 30Oand pH 7.5 using 0.05 N NaOH as titrant.
Reaction mixtures contained initially from 0.2 to 35 mM substrate in 5.0 ml of 0.1 M KCl containing 4% ethanol. Subtilisin concentra-
tion was adjusted to the range where the rate of reaction was proportional to the enzyme concentration.
0 Calculated per mole of enzyme, based on a molecular weight of 27,600.
The finding (39) that certain aromatic compounds, e.g., phenol, indole,
hydrocinnamate, and indolepropionate, are competitive inhibitors of the
hydrolysis of N-acetyl-L-tyrosine ethyl ester but are noncompetitive in-
hibitors with N-a-benzoyl-L-arginine ethyl ester suggests that these two
ester substrates are bound in a somewhat different manner at the active
sites. Both modes of binding obviously lead to productive complexes, a
fact which may in part explain the broad range of specificity of the
subtilisins. Thus different substrates have different productive modes of
binding and the subtilisins are therefore able to cleave peptide bonds
in a wide variety of natural and synthetic substrates.
Although subtilisin Carlsberg has a higher V,,, for the acetylamino
acid esters than either subtilisins Novo or Amylosacchariticus, the K,
values for the substrates (Table VII) and the KI values for the inhibitors
(Table VIII) are similar for all of these enzymes indicating that, despite
differences in amino acid sequences they possess similar substrate binding
sites. It has been observed (39), however, that subtilisin Carlsberg has
higher V,,, values with the N-acetylamino acid esters than subtilisin
NOVO,but that both have similar V,,, values with the free amino acid
esters (40). This suggests that the binding site of Carlsberg is less polar
than that of the Novo enzyme and thus more sensitive to the polar
a-amino group in the free amino acid esters (40).
The broad specificity exhibited by the subtilisins is in marked contrast
to the high degree of specificity shown by trypsin and chymotrypsin
toward synthetic substrates (111). Furthermore, the subtilisins have
TABLE VIII
COMPETITIVE
INHIBITORS
OF HYDROLYSES OF ACETYL-L-TYROSINE
&HYL
ESTERBY SUBTILISIN“
The compounds listed in the table were tested as competitive inhibitors of subtilisin-
catalyzed hydrolysis of acetyl-btyrosine ethyl ester in 0.1 M KCI containing8% dioxane
by volume at 37” and at t,he pH and inhibitor concentration indicated.
* Data from Glazer (39).
111. N . M. Green and H. Neurath, in “The Proteins” (H. Neurath and K.
Bailey, eds.), 1st ed., Vol. 2, Part B, p. 1057. Academic Preas, New York, 1954.
590 F. S. MARKLAND, JR., AND E. L. SMITH
much higher K,,, values with their best ester substrates than do chymo-
trypsin and t.rypsin and also act extremely slowly on acetyl-L-amino
acid amides (39, 100). It would appear that high K , values for simple
ester substrates may be associated with proteinases of low specificity
such as papain and ficin (112) as well as various bacterial and fungal
enzymes (100).
Morihara (113) and Morihara and Tsuzuki (100) have studied the
action on synthetic substrates of a variety of alkaline and neutral pro-
teinases from various bacteria and molds. The neutral proteinases hydro-
lyze mainly at peptide bonds involving the amino group of hydrophobic
amino acid residues (113,114), whereas the alkaline proteinases appear
to hydrolyze mainly a t peptide bonds linking the carboxyl group of
hydrophobic amino acid residues. The data in Table I X indicate that
subtilisin B P N (as do the other alkaline proteinases studied) hydrolyzes
synthetic substrates of the type Cbz-Gly-X-NH, where X is Ala, Leu,
Phe, Tyr, Ser, and Trp (the last two show less extensive hydrolysis).
Benzoyl-Tyr-NH, is hardly hydrolyzed at all by B P N in distinction
to a-chymotrypsin where this amide is hydrolyzed much more rapidly
than Cbz-Gly-Tyr-NH,. To investigate the effect of neighboring groups
on the hydrolysis of synthetic substrates a variety of leucine derivatives
was studied (Table I X ) . The hydrolysis of Cbz-Gly-Leu-NH, is greatly
reduced when the Cbz group is replaced by acetyl and there is almost
no hydrolysis when the amino group is unsubstituted. When L-leucine
is replaced by D-leucine there is no hydrolysis. With various acylamino
acid esters, Morihara and Tsuzuki (100) found, as did Glazer (39) and
Bare1 and Glazer (@), that subtilisin B P N has higher activity when
aromatic amino acids are present, but there is also activity with esters
of the basic amino acids and certain neutral aliphatic amino acids (Table
VII) .
Additional work by Morihara et al. (114a) with leucine substrates of
the type Cbz-A- (Gly),,-Leu-NH, and Cbz-Leu- (GIST),,-B (where A or B
are various D or L amino acid residues; n equals 0, 1, and 2 ; and cleavage
occurs a t the COOH-side of the leucyl residue) have shown that hydrol-
ysis was influenced by the three amino acid residues on the NH,-terminal
side and the two amino acids on the COOH-terminal side of the sensitive
bond. Furthermore, these effects on hydrolysis were mainly related tc
catalysis rather than substrate binding.
112. E. L. Smith and J. R. Kimmel, “The Enzymes,” 2nd ed., Vol. 4, p. 133
1960.
113. K. Morihara, BBRC 26, 656 (1967).
114. K. Morihara, H. Tsuzuki, and T. Oka, ABB 123, 572 (1968).
114a. K. Morihara, T. Oka, and H. Tsuzuki, ABB 138,515 (1970).
16. SUBTILISINS 591
TABLE IX
HYDROLYSES SYNTHETICSUBSTRATES
OF VARIOUS BPN'.
BY SUBTILIBIN
-~
Hydrolysisb Hydrolysis*
20 16 20 16
min hr min hr
Substrate (%I (%) Substrate (%) (%I
Cbz-Gly-Gly-NHz 0 Cbz-Gly-Pro-D-Leu-G1y-Pro-oH 0
Cbz-Gly-Ala-NH2 11 >95 Cbz-Gly-Pro-Leu-Gly-OH 20 >95
T T
Cbz-Gly-Ser-NH2 0 22 Cbz-Gly-Pro-Leu-OH 0
t Cbz-Gly-Pro-NHS 0
Cbz-Gly-Val-NHzc <5 Cbz-Gly-PheNHsC 6 70
Cbz-G ly-Ile-N Hp 0 t
Cbz-Gly-Leu-NHZ 18 >95 (Cbz-Gly-Tyr-NHf 32 >95
T t
Cbz-Gl y-D-Leu-NHz 0 Cbz-Gly-TrpNHnc 0 13
Acety1-Gly-Leu-NH? 11 T
T Benzoyl-Tyr-NHn <5
H-Gly-Leu-NHz <5 Benzoyl-Arg-NHz <5
Cbz-Leu-N Hz <5
Cbz-Gly-Pro-Leu-Gly-Pro-OH 37 95
t I
All reactions were performed in 0.1 M tris buffer (pH 8) for 20 min and for 16 hr at
40". The percentage of hydrolysis was determined by ninhydrin analysis. Substrate con-
centration was 4 mM and tha enzyme concentration was 50 pg/ml of reaction mixture.
The arrows show the points of cleavage.
* Data from Morihara and Tsuzuki (100).
The reaction mixture contained 4% methanol because of the low solubility of the
substrate.
It was also found that hydrolysis was inhibited when the terminal
a-amino and a-carboxyl groups of the peptide substrate were unblocked,
if the former group was within four residues on the NH,-terminal side
and the latter group within one residue on the COOH-terminal side of
the cleavage point. Finally, these authors reported that the side chain
specificity a t the residue being cleaved is extremely stringent compared
with those for each of the five amino acid residues surrounding the sen-
sitive residues, and that bulky residues on either side of the sensitive
bond inhibit hydrolysis.
Morihara and Tsuzuki (100) have called the alkaline proteinases
chymotrypticlike in their specificity. However, in comparing the size and
specificity of the active site of a-chymotrypsin and subtilisin BPN' by
using a variety of synthetic peptides, Morihara et al. (116) found con-
115. K. Morihara, T. Oka, and H. Tsuzuki, BBRC 35, 210 (1969).
592 F. S. MARKLAND, JR., AND E. L. SMITH
ever, retains its normal stereospecificity when the related open chain
ester iV-benzoyl alanine methyl ester is the substrate, hydrolyzing the
k
L and not he D isomers. Since a-chymotrypsin behaves in an identical
manner wit these substrates, both enzymes have similar specificity with
respect to the configuration of their substrates (117).
C. TRANSESTERIFICATION
AND TRANSPEPTIDATION
D. MECHANISM
OF ACTION OF THE SUBTILISINS
now well documented and a role for Asp 32 has been postulated ( 4 7 )
similar to the involvement of a specific aspartic acid in the mechanism
of a-chymotrypsin and other pancreatic proteinases ( 4 8 ) .
Keizer and Bernhard (125) noted that the protonic equilibrium ac-
companying acylation of both subtilisin Novo and a-chymotrypsin with
N-P- (3-indole) acryloylimidazole is identical, indicating that the in-
volvement of proton diseociable groups in the mechanism of catalysis
must be similar for these two enzymes. Thus, hydrolysis presumably
occurs via acylation ( k L ) and deacylation (Ic,) steps according to
Eq. (3):
EH + RCOX ek-i
kl
EH.RCOX
kg
k-2
ECOR + HX
(3)
ECOR + HnO e IlCOiH + EH
ks
k-,
An interesting observation that was made during the studies with thiol-
subtilisin (Section V,A) may shed sonic light on the general mccha-
nism of the serine proteinases. The high reactivity of the serine hydroxyl
at the active site of serine proteinases has been generally attributed to
hydrogen bonding between this residue and a nearby histidine residue
(122) which in turn is hydrogen bonded to an aspartic acid residue
(47, 48). However, Polg6r and Bcndcr (126, have presented a contrary
view. By measurement of the kinetic parameters it was shown that sub-
tilisins Novo and Carlsberg and their thiol derivatives in the acyl-
enzyme form possess an ionizable group with a pK, of about 7.0 in the
active site. This group, presumably the active histidine residue, also is
present in the unsubstituted enzymes but not in the free thiol derivatives
(76, 7 8 ) . The p l i has been shifted to 6.15 in free thiolsubtilisin Carls-
berg and somewhat lower than pH 5.5 in thiolsubtilisin Novo. To explain
this they postulate that a hydrogen bond is present between the thiol
group and the nearby histidine residue in the free thiol enzymes, whereas
in the serine enzymes, since there was 110 shift in the pK between the free
and aeyl enzymes, it appeared that no hydrogen bond is present. Kinetic
studies with D 2 0 supported the above conclusions. I t is known that the
velocity of reactions with a rate limiting protori transfer decrease to
about one-third the normal rate in D,O. The formation of a hydrogen
bond is coiisidcrctl a partial proton transfer; however, if a hydrogen bond
were formed between the catalytic groups during the formation of acyl
enzyme the rate would riot I)c markedly affected in D,O. The results in-
dicate that there was a D,O effect for deacetylation with p-nitrophenyl-
125. J. Keizcr nnd S. .I.Bcrnhard. B i o c h ~ m i ~ t r5,y 4127 (1966)
126. I,. Polgtir and M. L. Bender. Proc. Natl. Acnd. Sci. U . S. 64, 1335 (1969).
16. SUBTILISINS 595
0
Oh
C -I Asp
OdH
FIG.6. Possible mechanism of the formation (a), and the hydrolysis (b) of the
ncyl enzyme in the catalysis by serine proteases; X represents the leaving group.
From Polgir and Bender (126).
596 F. S. MARKLAND, JR., AND E. L. SMITH
A. LYSINEMODIFICATION
B. METHIONINE
MODIFICATION
C. TYROSINE
MODIFICATION
128a. K. Ohtsuki, C. L. Liu, and H. Hatano, J . Biochem. (Tokgo) 66, 863 (1969).
129. J. T. Johansen, M. Ottesen, and I. Svendsen, BBA 139, 211 (1967).
130. I. Svendsen, CompL. Rend. Trau. Lab. Carlaberg 36, 347 (1968).
600 F. S. MARKLAND, JR., AND E. L. SMITH
(24, 132) has shown that 5 tyrosine residues in subtilisin Novo react
readily with tetranitromethane and an additional 3 or 4 react a t least
partially. Isolation of the peptides containing nitrotyrosine from com-
bined tryptic and cyanogen bromide digests has allowed precise identi-
fication of the nitrated residues (Table X ) . Tyrosine residues 6,21,104,
217, and 262 are fully nitrated, whereas residues 91 and 263 are partially
nitrated (see Fig. 1 ) . The tryptic peptide containing tyrosine residues
167 and 171 contains nitrotyrosine; however, it is not clear as yet
whether both residues are partially nitrated or only one is nitrated fully.
Presumably the 5 readily nitrated residues are the same as those that
titrate normally and those that are partially nitrated may be the resi-
dues with slightly higher pK,,,.
Nitration of subtilisin Novo with a fivefold molar excess of tetra-
nitromethane yields only one major nitrated residue which was identi-
fied as Tyr 104 (2.4).Nitration of this single residue did not fully pro-
mote activity on clupeine indicating that either a combination of several
nitrated tyrosine residues is necessary for full increase of activity or that
the nitration of a single tyrosine residue controls the activation but that
it is not the first to be nitrated. Weber and Kraut (133) have reported
that the first residue iodinated in subtilisin BPN’ is also T y r 104. Other
TABLE X
REACTIVITY
OF TYROSINE I N SUBTILISIN
RESIDUES BPN’ (OR Novo)
6 Full
21 Full Monoiodo
91 Partial
104d Full Diiodo
167 None or ? One is monoiodo
171 None or ?
2 14 None
217 Full Diiodo
262 Full
263 Partial
Of the 10 tyrosine residues, 5 titrate normally with pK.,, = 9.74, 3 with pK.,, =
11.25, and 2 with pK.,, > 12.5.
b Data from Markland (S4, 13%’).
c Data from Wright et al. ( 4 7 ) .
d This is the first residue to be iodinated or nitrated (94,f33).
138. A. N. Glazer. Proc. Natl. Acnd. Sci. U.S. 50, 996 (1968).
604 F. S. MARKLAND, JR., AND E. L. SMITH
E I +
fast
(E1)acti”e -
slow
(E1’)inactive (4)
From the pH dependence of inhibition of the subtilisins by the phenyl-
arsonates it would appear that the binding of the inhibitor requires the
presence of a protonated histidine residue at the active site. A possible
scheme for the interaction of the phenylarsonates with the subtilisins is
presented in Fig. 7. The initial reaction (Step 1) is envisaged as depend-
ing on an electrostatic interaction between the phenylarsonate anion and
a protonated histidine residue, as well as a hydrophobic interaction in-
volving the aromatic ring. Subsequently (Step 21, ester formation may
take place with the active site serine residue. This mechanism is con-
Step 1
CHI-OH
-NH-CH-CO-
I
Q HC=C-CH,-CH
I I
I
NH
I
I
co
H O / t b H+Nbc/NH
H I
Step 2
HN,
+ ‘6 ,NH CO
I
-NH-CH-CO-
FIO.7. Possible mechaniam for the formation of an arsonic acid ester with the
reactive serine in the subtilisins. From Glazer (199).
sistent with all the available data but awaits definitive proof from
X-ray crystallography.
As already noted (Section 11), Keay and Moser (IS) were able to
group the highly purified subtilisins into two classes on the basis of
amino acid composition, serological cross-reactions, and the ratio of
esterase to proteinase activity. It remains to be determined whether
other alkaline proteinases will fit this classification.
There have been some reports describing the partial characterization
of alkaline proteinases from a variety of different strains of Bacillus.
Among these are B . cereus, strain K, 931 which produces three alkaline
proteinases ( 1 4 ) . Bacillus licheniformis culture filtrates also produce
three proteinases, one of which is certainly of the subtilisin type since
it had high esterase activity with N-acetyl-L-tyrosine ethyl ester and
much lower activity with N-benaoyl-L-arginine ethyl ester, exhibited an
alkaline pH optimum, and was inhibited by both DFP and phenyl-
methane sulfonyl fluoride (141).
A transforniable strain of B. subtilis (strain 168 indole-) produced two
different proteinases (designated basic and acidic based on their binding
to DEAE-cellulose) at different times in the growth cycle (14.9). Both
were inhibited by DFP and phenylmethane sulfonyl fluoride, although
the basic one did not possess appreciable activity on benzoyl-L-tyrosine
ethyl ester. Amino acid analysis showed significant differences between
the two proteinases although the basic proteinase had a composition
that was not unlike that of subtilisin Carlsberg. More recently it was
shown that the basic proteinase appeared similar kinetically and im-
munologically t o subtilisin Nova as well as having some antigenic
regions in common with subtilisin Carlsberg whereas the acidic pro-
teinase appeared altogether different from the other subtilisins (143).
Under conditions for production of the acidic proteinase a high molecular
weight proteinase was also present; this enzyme waa devoid of esterase
activity on benzoyl-L-tyrosine ethyl ester. Both the acidic and basic
proteinases were acid labile and had similar p H optima around pH 7.5-
140. Y. Furukawa, Y. Fujii, and H. Takahashi, Agr. Biol. Chem. (Tokyo) 32, 832
(1968).
141. F. F. Hall, H. 0.Kunkel, and J. M. Preacott, ABB 114, 146 (1960).
142. H. W. Boyer and B. C. Carlton, ABB 128,442 (1988).
143. J. H. Hageman and B. C. Carlton, ABB 139,67 (1970).
606 F. S. MARKLAND, JR., AND E. L. SMITH
A. PROTEIN
SEQUENCE
Probably the first practical use for subtilisin was in the determination
of the amino acid sequence of oxytocin ( 9 9 ) .Subtilisins have also been
used for the hydrolysis of a variety of other proteins and peptides.
Among these were the limited digestions, already mentioned, of oval-
bumin (147) and ribonuclease (104).Other uses may be noted briefly:
the digestion of native hemoglobin ( l o ) , the determination of the
amino-terminal sequence of a-crystallin (I@), the identification of
citrulline in the medulla protein of the quill of the African porcupine
(150),the determination of the sequence of the nonapeptide phyllocae-
rulein from the skin of a South American amphibian (161), as an aid
in distinguishing the differences in sequence between beef and pig insulins
( 9 8 ) ,and as an aid in determining the positions of the disulfide bridges
linking the two chains of insulin ( 9 7 ) .Foster and co-workers (156, 153)
have used subtilisin BPN’ to probe the structure of bovine plasma
albumin by limited digestion of an albumin-sodium dodecyl sulfate
complex. Hill (154) has presented an extensive discussion of the hydrol-
ysis of a variety of proteins by the subtilisins, and the articles by Smyth
(155) described the procedures to be employed in using subtilisin for
sequence work.
Because of their broad specificity, the subtilisins have been useful
for studying the sequences of small peptides which are not hydrolyzed
by more highly specific proteinases.
A water-insoluble active derivative of subtilisin Novo has been pre-
pared by cross-linking with glutaraldehyde. Amino acid analysis in-
dicates that lysine residues are probably involved in the intermolecular
cross-linking reaction with glutaraldehyde. This preparation, when mixed
with cellulose powder in a chromatographic column, has shown some
promise in digesting proteins (166).
B. USE IN DETERGENTS