protein Science (1993), 2, 1383-1390. Cambridge University Press. Printed in the USA.

Copyright 0 1993 The Protein Society

~~~ ~~~~ ~ ~~~~ ~~

~~~

Conformational instability of the

N- and C-terminal lobes of porcine
pepsin in neutral and alkaline solutions

XINLI LIN,' JEFFREY A. LOY,' FREDY SUSSMAN,' AND JORDAN TANG','

' Protein Studies Program, Oklahoma Medical Research Foundation, and 'Department of Biochemistry
and Molecular Biology, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma 73104
(RECEIVEDApril 13, 1993; REVISEDMANUSCRIPT
RECEIVEDJune 9, 1993)

Abstract
Pepsin contains, in a single chain, two conformationally homologous lobes that are thought to have been evolu-
tionarily derived by gene duplication and fusion. We have demonstrated that the individual recombinant lobes
are capable of independent folding and reconstitution into a two-chain pepsinor a two-chain pepsinogen (Lin,
X . , et al., 1992, J. Biol. Chem. 267, 17257-17263). Pepsin spontaneously inactivates in neutral or alkaline solu-
tions. We have shown in this study that the enzymic activity of the alkaline-inactivated pepsin was regenerated
by the addition of the recombinant N-terminal lobe but not by the C-terminal lobe. These results indicate that
alkaline inactivation of pepsin is due to a selective denaturation of its N-terminal lobe. A complex between re-
combinant N-terminal lobe of pepsinogen and alkaline-denaturedpepsin has been isolated. This complexis struc-
turally similar to a two-chain pepsinogen, but it contains an extension of a denatured pepsin N-terminal lobe.
Acidification of the complex is accompanied by a cleavage in the pro region and proteolysis of the denatured
N-terminal lobe. The structural components that are responsible for the alkalineinstability of the N-terminal lobe
are likely to be carboxyl groups with abnormally high pK, values. The electrostatic potentialsof 23 net carboxyl
groups in the N-terminal domain (as compared to 19 in the C-terminal domain) of pepsin were calculated based
on the energetics of interacting charges in the tertiary structure of the domain. The groups most probably caus-
ing the alkaline denaturation are Asp", Asp"', Glu4, Glu13, andAsp"'. Especially, the partially buried Asp",
which interacts with Asp159, could cause one of these two groups have to an abnormally high pK, and the other
an abnormally low pK, value. Thus, the ionization of Asp'' at ahigh pH may place two negatively charged resi-
dues in close vicinity. This unfavorable situation may be the trigger for the denaturation of the N-terminal lobe
of pepsin.
Keywords: carboxyl ionization; denaturation; domain structure; electrostatic potentials; porcine pepsin

The fact that some aspartic proteases, such as HIV pro- Pepsin is a particularly good model for the study of as-
tease and renin, areinvolved in human diseases has stim- partic proteases because detailed informationis available
ulated much interest recentlyin the understanding of the on its structure (Abad-Zapateroet al., 1990; Cooper et al.,
structure-function relationships of this group of enzymes. 1990; Sielecki et al., 1990), enzymic properties (Fruton,
1976), and zymogenactivation(Tang & Wong, 1987;
~ ~~~~

Hartsuck et al., 1992). Like other eukaryotic aspartic pro-

Reprint requests to: Jordan Tang, Protein Studies Program, Okla- teases, pepsin is a single-chain protein, and its tertiary
homa Medical Research Foundation, 825 N.E. 13th Street, Oklahoma
City, Oklahoma 73104. structure consists of two homologous lobes. The substrate
Abbreviations: bis-Tris, bis(2-hydroxyethyl)-iminotris(hydroxy- binding cleft and thecatalytic aspartic residues are located
methy1)methane; FPLC, fast protein liquid chromatography (chromato- between these lobes (for review, see Davies [1990]). On
graph); HIV, human immunodeficiency virus; PAGE, polyacrylamide
gel electrophoresis; pep, the NH,-terminal lobe of pepsin; pep.sin, two- the other hand, the retroviral aspartic proteases,includ-
chainpepsin;pepsin,,alkaline-inactivatedpepsin;pro,propeptide; ing the HIV protease, are homodimerswith the substrate
propep, the NH2-terminal lobe of pepsinogen; propep.pepsin,, com- binding cleft and the active-site aspartic residues equally
plex between propep and alkaline-inactivated pepsin; propep.sin, two-
chain pepsinogen; SDS, sodium dodecyl sulfate; sin, the C-terminal lobe contributed by the two monomers(Foundling et al., 1989;
of pepsinogen or pepsin; Tris, tris(hydroxymethy1)aminomethane. Lapatto et al., 1989; Miller et al., 1989; Wlodawer et al.,
1383
1384 X . Lin et al.

1989; Fitzgerald et al., 1990). It has been proposed that

the internally homologous lobes of pepsin and other eu-
karyotic aspartic proteases are evolutionarily derived by
-1 I

gene duplication and fusion, and that a primordial aspar-

tic protease would have a homodimeric structure (Tang
et al., 1978). The close relationships of pepsin and retro-
viral aspartic protease active-site structures and their over-
all folding patterns (Rao et al., 1991) suggest that the
retroviral protease is closely related to a primordial form
of aspartic protease. It is therefore interesting to compare
the individual domains of the eukaryotic aspartic prote-
ase with the monomersof the retroviral aspartic protease
(Kinemage 1). Pepsin + Propep Pepsin t SIN

We have reportedrecently that the individually ex- Fig. 1. Regeneration of activity of alkaline-inactivated pepsin by recom-
pressed recombinantN-terminallobeofpepsinogen, binant N-terminal lobe, propep. Pepsin (0.1 mg/mL) was incubated in
called propep, and the C-terminal lobe,called sin, are ca- 20 mM Tris-HCI ( p H 8.0) at4 "C for 20 h, and then 1 pL was taken for
pable of individual folding and reconstitution into active enzyme assay using ['4C]hemoglobin as substrate (left column).In the
center column, 39 pL of partially purified recombinant propep (by S-
two-chain pepsinogen, propep.sin (X. Lin et al., 1992).
300 Sepharose gel column; X . Lin et al., 1992) was incubated with 1 pL
Upon activation in acid, propep.sin can yield two-chain of the denatured pepsin at20 "C for 20 min and then assayed for pro-
pepsin, pep-sin. Additionally, either propep or pep can teolytic activity. In the right column, denatured pepsin was preincubated
self-associate to form proteolytically active homodimer. with purified recombinant sin prior to the enzyme assay. Under the ex-
These results suggest that the N- and C-terminal lobes of perimental conditions, recombinantpropep showed virtually no activ-
ity. The bars at the top of the center column represent the standard
pepsin have retained some of the features of the homo-
deviation from three determinations.
dimeric enzymes, especially in the ability of individual
lobes to fold independently and in interlobal recognition.
The ability for folded domains of pep and sin to rec-
ognize each other and togenerate activity provides an op- ( p e p ) of pepsin but retain the conformation of the
portunity to distinguish the conformationally folded lobes C-terminal lobe, sin.
from theunfolded lobes in pepsin. It has long been known
that pepsin is spontaneously inactivated in neutral and al-
Purification of propep .pepsin, complex
kaline solutions. By adding the individually folded recom-
binant lobes in an activity-reconstituting assay, we have The restoration of the activity of alkaline-inactivated pep-
now demonstrated that the alkaline inactivation of pep- sin, pepsin,, by recombinant fragment propep must have
sin is caused by the denaturation of theN-terminal lobe, resulted from thebinding of the conformationally intact
pep. Structural components possibly responsible for the sin in pepsin, to propep to form acomplex, propep-pep-
alkaline denaturation of the pep domain have also been sin,. For further studies, this complex was prepared in a
analyzed and identified by energy calculations. larger scale by diluting a urea solution of recombinant
propep into a solutionof alkaline-inactivated pepsin (see
Materials and methods). This procedure took advantage
Results
of the presence of pepsin, during the refolding ofpropep
and thusgave a higher yield of the complex than from the
Alkaline sensitivity of the two lobes of pepsin
mixing of components refolded separately. The lower
An experiment was designed to determine the relative sen- yield in the procedure where refolding took place sepa-
sitivity of two lobes in pepsinto inactivation at different rately was probablycaused by the self-associationof
pH values. The results in Figure 1 demonstrated that pep- propep after refolding. Propep-pepsin, complex was pu-
sin was completelyinactivated at pH 8 (left column). rified by chromatography on Sephacryl s-300 (not shown)
When prefolded recombinant sin was added to the inac- and QAE columns. The latter gave rise to a single pro-
tivated pepsin, no activity was generated (right column). tein peak (Fig. 2). Neither the refolded propep nor pep-
The addition of prefolded recombinant propep resulted sin, forms a peak on QAE-column chromatography. The
in a dramatic restoration of activity (center column). The purity of the complex was also analyzed by FPLC on a
intrinsic enzyme activity of propep (X. Lin et al., 1992) Superose 12 column, which resulted in a single peak with
is not detectable in a control under the experimental con- a shoulder on thelower molecular weight side (Fig. 3A).
ditions (results not shown). So the regenerated activity The shoulder probably does notrepresent chemical het-
must have come from the recombinationof propep with erogeneity and is likely to be due to the lack of confor-
alkaline-inactivated pepsin. These results suggest that al- mational uniformity in the denatured propep domain.
kaline solutionsspecifically denature theN-terminal lobe Some heterogeneity in the conformation of refolded
Alkaline denaturation of pepsin 1385

1.6 -1 I A
1.6 - .0.0
.OB
1.4.

E
= 1.2.
.0.7
0
N
e
m 1.0.
.0.6
u
m
-E.
e 0.8. .0.5
6: B
U

I OB-
- I
0.4
I
0.4 - .0.3 I

0.2 I
i .0.2 I 1
5 10 15 2250 30 35 40

./I'
I I I
~ 01
Elution Time (min) B
5 10 15 20
Elution Time (min)

Fig. 2. FPLC fractionation of propep.pepsin, on a MonoQ column.

Active fractions from SephacrylS-300 chromatography (not shown)were
applied to a MonoQ column in a Pharmacia FPLC, which had been
equilibrated with 20 mM bis-Tris, 0.8 M urea, pH6.0. Elution was ef-
fected with a linear gradient of I M NaCl in the same buffer.

recombinant propep (X.Lin et al., 1992) may also con-

tribute to this chromatographic behavior. The conforma-
tional but not the chemical origin of the shoulder is
supported by the results from SDS-PAGE (Fig. 4, lane 3),
which revealed that the purified complex consists of a
pepsin band (itsmobilityusuallycorresponds to near
40 kDa) and a propep band (near 25 kDa). Elution Time (min)

Fig. 3. FPLC gel filtration of propep.pepsin, and its acid-activation

Acid-activation of propep .pepsin, complex products. A: Complex propep.pepsin, (0.68 mg in 120 pL) from the
MonoQ peak fraction was applied to an FPLC Superose 12 column,
Like pepsinogen, the complex propepepepsin, is capable which was preequilibrated with20 mM bis-Tris, 0.8 M urea, and0. I M
of activation in acid with a concomitant cleavage in the NaCI, pH 6.0. The sample was eluted with the same buffer atflow
a rate
p r o peptide region. Acid-treated complex eluted from a of 0.5 mL/min. B: The same amount of propep.pepsin, as in A was
acidified to pH 2.0by the additionof 20 pL of 0.5 M Na citrate, pH2.0.
molecular weight-calibrated gel-filtration column(Y. Lin The mixture was allowed to incubate at 20°C for I O min, and the
et al., 1992) at an eluentvolumecorrespondingto a insoluble aggregate was pelleted by microcentrifuge in an Eppendorf
smaller size (Fig. 3B) than that of the untreated complex centrifuge for8 min at 4 "C. The clear supernate was subjected to chro-
(Fig. 3A). The apparentsize of the purified complex is be- matography asin A . The protein peak was eluted at 25.7 min, compared
tween 150 and approx. 200 kDa (Fig. 3A). (The high ap- to 22.2 min for the peak in A.

parent size is caused by the presence of a denatured

N-terminal lobe in pepsin,.) After acidification, the ap-
parent size was reduced to 53 kDa (Fig. 3B). On SDS-
PAGE, the propep band in the acid-treated (15 min at by acid activation of propep .sin (X. Lin et al., 1992).
pH 2) complex is smaller (22 kDa; Fig. 4, lane 4) than Pepsin, in the complex appears to have been degraded
from the untreated sample (25 kDa; Fig. 4, lane 3). The during acid activation, as its band has decreased to 33
22-kDa band represents an intermediate in acid activation kDa (Fig. 4, lane 4) from the untreated 40 kDa (Fig. 4,
of propep in the propep.pepsin, complex, since it is lane 3). These results indicated that, in an acid solution,
larger than thepep band (20 kDa; Fig. 4, lane 5) produced the complex is activated and proteolytically cleaved in the
1386 X . Lin et al.

M 1 2 3 4 5 6 7 9-10 is the main difference between these two two-chain

zymogens.
9L-
67-
L3- Relative stability of propep, pep,and
sin lobes at different pH values
30-
Pepsin, and the recombinant and individuallyfolded
20.1- lobes propep and sin were incubated separately in solu-
tions of different pH values for defined time in order to
1L.L-
find out their relative stability. The potentials for resto-
ration of activity were assayed after the additions ofre-
combinant propep (for preincubated pepsin, and sin) or
sin (for preincubated propep). Figure 6 shows that recom-
Fig. 4. SDS-PAGE electrophoresis patterns of propep.pepsin and its
binant propep loses its potential activity aspH increases
acid-activation products. Lane I , pepsinogen (1 p g ) ; lane 2, pepsin, from 6 to 11. On the other hand, recombinant sin loses
(2 p g ) ; lane 3, purified propep.pepsin, (2 p g ) ; lane 4, acid-activated partial activity only between pH 8 and 9, and does not
propep.pepsin, (2 p g ) ; lane 5 , acid-activated propep.sin (resulting in lose further activity above pH 9. The loss of potential
pep.sin, I p g ) ; lane 6, untreated propep.sin (3 p g ) ; lane 7, recombinant regeneration of activity of pepsin is similar to that of sin
propep in cell lysate ( 5 p g ) used as a marker for the position of propep.
between pH 8 and 9, but further declines above pH 9.
The difference between sin and pepsin above pH 9 is
probably dueto the ability of sin to refold in the activity-
pro region of the propep fragment and in the denatured regenerating assay (at pH 6). In pepsin, the presence of
pep region of pepsin,. denatured pep lobe may have prevented the refoldingof
its sin lobe. These results also confirmthat sin is more al-
kaline stable, atleast under the experimental conditions,
Relative alkaline stability of propep .pepsin,
than is propep.
to propep -sin,pepsin, and pepsinogen
Complex propep. pepsin,, two-chain pepsinogen, pepsin,
Discussion
and pepsinogen were compared for their stability in a
20-min incubation at different pH values ranging from The observed events from this study are schematically
6.0 to 10.5. Figure 5 shows the well-known instability of illustrated in Figure 7. The results indicate that the alka-
pepsin above pH6 and of pepsinogen above pH 10 (Fru- line sensitivity of pepsin is due to thespecific denatura-
ton, 1971). Two-chain pepsinogen and complex propep.
pepsin, are unstable above pH 8.5. The greater resistance
to inactivation for propep-pepsin, in the pH range of

+Pepsin
+Propep Sin

6 7 a 9 10 il
pH

Fig. 6 . Stability of the potential regenerating activity of propep, sin,

PH and pepsin. Pepsin and recombinant propep and sin were preincubated
individually in different pH solutions. Excess of recombinant propep
Fig. 5. Inactivation of propep.pepsin,, propep.sin,pepsin, and pep- was added to the preincubated pepsin and sin, while excess of sin was
sinogen in solutionsof different pH values. The enzyme and zymogen added to the preincubated propep. The activity of the solutionwas then
solutions were preincubated in different pH solutions before assays. assayed.
Alkaline denaturation of pepsin 1387

Fig. 7. Schematic presentation of observa-

tions made in this study. An alkaline solu-
7
Cleaved
r - propep tion denatures thepep lobepepsinof to cause
Tb an inactivation. The addition of recombi-
nant (r-) propep to the alkaline-denatured

@ Akali enzyme (pepsin,) results in the formation of

complex propep.pepsin,, which resembles
a two-chain pepsinogen (X. Lin et al., 1992).
When the complex is in an acidic solution,
Pepsin the activation is accompanied by the cleav-
+
Denatured pep
h p e p pepsina T age of the pro peptide and the denatured
domain of pepsin,.
pep

Cleaved pep I

tion of its N-terminal lobe propep. This conclusion is vation kinetics of propep.pepsin, complex. The appar-
supported by the regeneration of activityof the alkaline- ent first-order activation constant is at least 100 times
inactivated pepsin by the addition of propep but not by slower (0.106 min" in pH 2.0 at 29 "C) than that of the
sin (Fig. 1). The alkaline stabilities of the individual re- native pepsinogen or two-chain pepsinogen (results not
combinant lobes (Fig. 6) also support this conclusion.Re- shown). This is in part due to competition from the sec-
combinant sin is stable at pH <8 but at high pH it may ond proteolytic event (digestion of the denaturedpep do-
be in reversible denaturation (Fig. 6). It has long been main). This competitive inhibition is analogous to the
known that pepsin undergoes irreversible denaturation inhibition of first-order activation of pepsinogen in the
above pH 6 (Fruton, 1971), which is independentof presence of hemoglobin substrate (Marciniszyn et al.,
buffer used. Therefore, we consider the results obtained 1976). A second factor for theslow activation rate is the
from the current studies to be a general model of alka- hindrance of the denaturedpep domain on the movement
line pepsin denaturation. of the pro region, which is required in the activation
The propep.pepsin, complex is sufficiently stable to mechanism (Marciniszyn et al., 1976).
purification. This complex spontaneously activates in acid In alkaline solution, complex propep.pepsin,is more
with a concomitant cleavage in thepro region, illustrat- stable than pepsin, less stable than pepsinogen, and sim-
ing that the added propepmoiety and the lobe sin from ilar in stability to two-chain pepsinogen (Fig. 5). The in-
the alkaline-treated pepsin must form a complex whose creased stability of propep.pepsin, in the pH range of
conformation is very near that of native pepsinogen 9-10, however, is somewhat surprising. This is possibly
(Fig. 7). During acid activation, at least two different due to thenonspecific interaction between the denatured
types of cleavage take place. The first event is the cleav- pep domain inpepsin,with theaddedrecombinant
age of thepro peptide normally associatedwith the acti- propep, which may have slowed the rate of denaturation.
vation of pepsinogen. The fact that the activation does We have used the known three-dimensional structure of
not reduce propep to thesize ofpep indicates that the ob- pepsin to generate information on the groups that may be
served activation product of propep (Fig. 4)is an activa- responsible for thealkaline denaturation of the N-terminal
tion intermediate. The incomplete cleavage of the p r o lobe of pepsin (Kinemage 2). Pepsin is a very acidic pro-
peptide is apparently due to the much reduced rate of ac- tein with an isoelectric point below pH 1 (Tiselius et al.,
tivation of this complex as compared to that of the na- 1938). This can be attributed to large
a number of nega-
tive zymogen (see below). When porcine pepsinogen is tively charged groups, 23 in thepep domain and19 in the
activated in acid, thefirst bond to be cleaved is between sin domain. Theonly two positive groupsin thepep lobe
P16 and P17 (Ile-Leu) in the pro region (Dykes & Kay, are thecy-amino group at theN-terminus and His5'. Both
1976). This is the probable site of the cleavage during ac- of these groups are located on the outside of the three-
tivation of the complex. The second event is the digestion dimensional structure. Thus theloss of charges of these
of the alkaline-denatured pep domain in pepsin, by the groups above theirpK, values is unlikely to cause the de-
activated complex (Fig. 7). As a result, activation also naturation of thepep lobe. Therefore, the most reasonable
produced a pepsin, fragment (Fig. 4). This is not surpris- explanation of the denaturation of thepep domain above
ing since the denaturation of pep domain in pepsin, ex- pH 7 is the ionization ofsome carboxyl groups with pK,
poses some hydrophobicresidues normally buried in the values in the neutral and alkaline range. Normally, car-
tertiary structure of pepsin, thus providing specificity sites boxyl groups of aspartic and glutamic side chains have
for cleavage. It is known that in acid the pro region is se- pK, values near pH 4. However, in a special environment,
lectively "denatured" in pepsinogen; thus it is capable of such as buriedin the tertiary structureof the protein or in-
substrate hydrolysis (Marciniszyn et al., 1976). The ini- teracting with other charged groups, thepK, of these car-
tial autodigestion of propep.pepsin, must have come boxyl groupscan begreatlyelevated.Figure 8 and
from a similar mechanism. We have measured the acti- Kinemage 2 show all residues in pepsin with partially or
1388 X . Lin et al.

Fig. 8. Ribbon model of pepsin with resi-

dues potentially important for alkaline dena-
turation of the enzyme. Superimposed on the
ribbon view of pepsin are the acidic titrat-
able side chains (with residue numbers),
which are not fully exposed to the solvent
and hence could have abnormal pK, values.
Criteria for solvent exposure were taken from
the structural work of Sielecki et al. (1990).
Eight out of the 10 residues marked in the
figure are located in the N-terminal lobe
(residues 1-172). These groups could con-
tribute to the pH-dependent denaturation of
the N-terminal lobe.

completely buried carboxyl groups. It can be seen that most Materials and methods
of these residues are located in the N-terminal lobe, pep
(residues 1-172). Thus they are potential contributors to Materials
the alkaline denaturation. In order to investigate the effects Recombinant propep and sin, purified and refolded, are
of closely positioned carboxyls on their ionizations, we prepared as previously described (X. Lin et al., 1992).
have used Mehler distance dielectric dependence (Mehler Porcine pepsin was purchased from Sigma. Other re-
& Eichele, 1984)to calculate the electrostatic potentials of
agents are the highest quality obtained commercially.
these groups based on their interaction with other nega-
tively charged groups in the vicinity. As shown in Table 1,
Asp" and Asp'59are among several with the largest po-
tential observed. It is interesting that Asp" is partially
buried and forms a hydrogen-bonded dyad with AspJ5',
also located in the N-terminal lobe (Fig. 9; Kinemage 2).
This relationship is analogous to that of the active-site
Asp32and Asp2". The former also has a high electrostatic
potential (Table 1). We have reported that the pK, values
of the active-site dyad are 1.57 and 5.02 (Y. Lin et al., 1992) . .. \
in native pepsin. Because Asp" is much more buried than
the active-site residues, its pK, is likely to be higher than 5 .
Characteristically, the counterpart of the dyad, AspIS9,
may have a pK, lower than Asp". Thus, the ionization of
Asp'' at neutral or alkaline solutions would place two
negative charges in close vicinity. This is an unfavorable
interaction for these two aspartyl groups and may play an
important role in the denaturation of the pep domain in
pepsin. Other residues with high electrostatic potentials
(Table 1) that may be important in the alkaline denatur- Fig. 9. The environment of residues Asp" and AspIs9. Asp" and
Aspis9 form a hydrogen bond dyad similar to that of the active-site
ation of the N-terminal lobe are Glu4, GluI3, and Asp'l8, Asp32 and Asp2". Asp" is only partially accessible to the solvent. It
all high in electrostatic potentials (Table 1). All these is hypothesized that Aspis9and Asp", respectively, have abnormally
groups are buried in the tertiary structure of pepsin. Thus, low and high pK, values for carboxyl groups. The ionization of both
their ionization may induce conformational changes in or- groups in neutral and alkaline solutions in close vicinity is disruptive to
the conformation of the N-terminal lobe. Other groups that may mod-
der for these side chains to gain access to and be stabilized
ulate the pK, values of this dyad are Ams, which forms a hydrogen
in an aqueous environment. The role of these residues in bond with Asp", and the backbone peptide amide group of AspIS9,
pepsin denaturation in alkaline solutions may be tested in which forms a hydrogen bond with its side chain. Carbon, oxygen, and
future experiments using site-directed mutagenesis. hydrogen atoms are shown in green, red, and white, respectively.
Alkaline denaturation of pepsin 1389

Table 1. Electrostatic potential on acidic residues previously for thepurification ofpropep-sin (X. Lin et al.,
of the N-terminal (pep) lobe of pepsin 1992).
~- ~

~ _ _ _ _ _ -~

Potential energy
Residue pH stability of proteolytic activity
Asp 3 -3.7 Ten microliters of a solution containing either pepsino-
Glu 4 -5.5 gen, propep.sin, propep.pepsin, (in 20 mM Tris-HC1,
Glu 7 -4.9
Asp 11 -6.3
pH 8.0), or pepsin (in 20 mM bis-Tris, pH 6.0) was added
Glu 13 -5.5 to a series of 40 pL of buffers (bufferA or B) ranging in
Asp 26 -3.7 pH from6.0 to 10.5. The solution was left standing at 20 "C
Asp 32 -6.5 for 20 min and then added to a premixed solution con-
Asp 52 -3.0 taining 50 pL of synthetic substrate (KPAEFF(NO,)AL,
Asp 59 -3.8
Asp 60 -4.4
2 mg/mL), 110 p L of HzO, and 1 mL of 0.2 M sodium
Glu 65 -2.5 citrate, pH 2.0, preequilibrated at 37 "C. The hydrolysis
Glu 70 -1.0 rate was monitored at A300nmin an HP8452A diode ar-
Asp 87 -4.2 ray spectrophotometer, and theinitial rate was taken as
Asp 96 -3.7
the enzyme activity.
Glu 105 -2.7
Glu 107 -4.5
Asp 118 -5.4
pH stability of the potentialregenerating activity
Asp 138 -0.6
Asp 142 -3.7 Recombinant fragments propep andsin were refolded at
Asp 149 -2.8
pH 6 and 8, respectively, as previously described (X. Lin
Asp 159 -8.0
Asp 160 -3.8
et al., 1992). Alkaline-inactivated pepsin, pepsin,, was
Asp 171 -2.1 prepared by direct dissolving of pepsin in a20 mM Tris-
HCl buffer, pH 8.0.

Stability of propep
Buffers To 10-pL aliquots of propep solution (10 mg/mL in
20 mM bis-Tris, pH 6), 90 pL of buffer of different pH
Buffers used forpH-dependentexperimentsincluded values were added and incubated at20 "C for 20 min. Ali-
buffer A, pH 6-9.6, 50 mM potassium phosphate, 50 mM quots of 20 pL of 1 M bis-Tris,pH 6, and2 pL of sin so-
bis-Tris, 50 mM Tris, and 100 mM NaCl, and buffer B, lution (7 mg/mL in 20 mM Tris-HC1, pH 8.0) were added
pH 9.8-1 1.O, 50 mM cyclohexylaminopropane sulfonic separately. The samples were assayed at pH 2 using I4C-
acid in buffer A. hemoglobin as substrate (Lin et al., 1989).

Proteolytic assays Stability of sin

Aliquots of sin solution (as for propep) were diluted
Proteolytic activitywas assayed using either radiolabeled
five times individually with buffers of different pH val-
bovine hemoglobin substrate (Lin et al., 1989) or synthetic
ues and incubated at 20 "C for 20 min. Two-microliter ali-
peptide KPAEFF(NOz)AL, (F(NOz)is p-nitrophenylal-
quots of this solution were mixed separately with 1 p L of
anine), as described previously (Fusek et al., 1990).
1 M bis-Tris, pH 6, and then37 pL of apropep solution
(as for propep). Afterincubation at 20 "C for 10 min, the
Purification of the complex of propep to alkaline- samples were assayed as for propep.
inactivated pepsin (propep .pepsin,)
Stability of pepsin,
Pepsin (150 mg) was dissolved in 4 L of20 mM Tris-HC1,
pH 9.0. This solution was kept at 20 "C for 30 min to Aliquots of pepsin, solution were diluted individually
completely inactivate the pepsin. A 400-mL urea-solubi- 200-fold with buffers of different pH values and incu-
lized recombinant propep solution (X. Lin et al., 1992) bated at 20 "C for 20 min. Two microliters of each sam-
was then diluted dropwise into the inactive pepsin solu- ple were treated in the same manner as sin (above).
tion with rigorous stirring. The final solutionwas kept at
20 "C for 15 min, and then 80 mL of 1 M bis-Tris, pH 6.0,
Calculation of electrostatic potential energies
was added slowly with stirring. After adjusting thepH to
6.2 with 1 M HCI, this solution was kept overnight at To determine the interactionenergies among charges we
4 "C. The subsequent purification procedures for active employed a simple coulombic expression witha sinusoi-
complex of propep. pepsin, was the same as described dal dielectric constant introduced by Mehler and Eichele
 1390 X . Lin et al.

(1984). In this approach, the shift in pK, due to thepres- D.H. (1989). Crystal structure of a retroviral proteinase from avian
ence of other charges is given by myeloblastosis associated virus in viral proteinases
as target for che-
motherapy.In CurrenlCommunications in MolecularBiology
(Krausslich, H.-G., Oroszlan, S., & Wimmer, E., Eds.), pp. 181-
191. Cold Spring Harbor Publications, New York.
Fruton, J.S. (1971). Pepsin. In TheEnzymes, Vol. III (Boyer, P.D., Ed.),
pp. 119-164. Academic Press, New York.
where Kb is Boltzmann's constant, Tis temperature, and Fruton, J.S. (1976). The mechanism of the catalytic action of pepsin and
is the potential at charge i , related acid proteinases. Adv. Enzymol. 44, 1-36.
Fusek, M., Lin, X.L., &Tang, J. (1990). Enzymic properties of ther-
mopsin. J. Bioi. Chem. 265, 1496-1501.
Hartsuck,J.A.,Koelsch, G., & Remington,S.J. (1992). The high-
resolution crystal structureof porcine pepsinogen. Proteins Struct.
Funct.Genet.13, 1-25.
Here R, is the distance between charges qi and q j , and Lapatto, R., Blundell, T.L., Hemmings, A., Overington, J., Wilderspin,
A.W., Woods, S.P., Merson, J., Whittle, P., Danley, D.E., Geoghe-
E,;is the sinusoidal distance-dependent dielectric, which gan, K.F.,Hawrylik, S., Lee,S.E.,Scheld,K.,Hobart,P.M.
is given by (1989). X-ray analysis of HIV-I proteinaseat 2.7 A resolution con-
firms structural homology among retroviral enzymes.Narure 342,
299-302.
cl;(R;j)= A + B / [ 1 + k exp( -XBR,)] , Lin, X.L., Lin, Y.Z., Koelsch, G., Gustchina, A,, Wlodawer, A,, &
Tang, J. (1992). Enzymic activities of two-chain pepsinogen, two-
whereB=cH2,-A,A=-20.9,k=3.5,andX=0.0018. chain pepsin, and the amino-terminal lobe of pepsinogen. J. Biol.
Chem. 267, 17257-17263.
The dielectric constant of water is All basic and Lin, X.L., Wong, R.N.S., & Tang, J. (1989). Synthesis, purification,
acidic titratable groups are taken into account in the and active site mutagenesis of recombinant porcine pepsinogen. J.
calculation. Bioi. Chem. 264, 4482-4489.
Lin, Y.Z., Fusek, M., Lin, X.L., Hartsuck, J.A., Kezdy, F.J., &Tang,
J. (1992). pH dependenceof kinetic parametersof pepsin, rhizopus-
pepsin, and their active-site hydrogen bond mutants. J. Biol. Chem.
Acknowledgments 267, 18413-18418.
Marciniszyn, J., Jr., Huang, J.S., Hartsuck, J.A., & Tang, J . (1976).
This work is supported by NIH grant DK-01107 to J.T. and GM- Mechanism of intramolecular activation of pepsinogen. J. Biol.
49553 to F.S. Chem. 251, 7095-7102.
Mehler, E.L. & Eichele, 0. (1984). Electrostatic effects in water-acces-
sible regions of proteins. Biochemislry 23, 3887-3891.
References Miller, M., Schneider, J., Sathyanarayana, B.K., Toth, M.V., Marshall,
G.R., Clawson, L., Selk,L., Kent, S.B.H., & Wlodawer, A. (1989).
Abad-Zapatero, C., Rydel, T.J., & Erickson, J. (1990). Revised 2.3 A Structure of complex of synthetic HIV-I protease with a substrate-
structure of porcine pepsin: Evidence forflexible
a subdomain. Pro- based inhibitor at 2.3 A resolution. Science 246, 1149-1 152.
teins Struct. Funct. Genet. 8, 62-81. Rao, J.K.M., Erickson, J.W., & Wlodawer, A. (1991). Structural and
Cooper, J.B., Khan, G., Taylor, G., Tickle, I. J., & Blundell, T.L. (1990). evolutionary relationships between retroviral and eucaryotic aspar-
X-ray analyses of aspartic proteinases 11. Three-dimensiona! struc- tic proteinases. Biochemistry 30, 4663-4671.
ture of the hexagonal crystal form of porcine pepsin at 2.3 A reso- Sielecki, A.R., Fedorov, A.A., Boodhoo, A., Andreeva, N., & James,
lution. J. Mol. Bioi. 214, 199-222. M.N.G. (1990). Molecular*and crystal structures of monoclinic por-
Davies, D.R. (1990). The structure and functionof aspartic proteases. cine pepsin refined at 1.8 A resolution. J. Mol. Bioi. 214, 143-170.
Annu. Rev. Biophys. Biophys. Chem. 19, 189-215. Tang, J., James, M.N.G., Hsu, I.-N., Jenkins, J.A., & Blundell, T.L.
Dykes, C.W. & Kay, J. (1976). Conversion of pepsinogen into pepsin (1978). Structural evidence for gene duplication in the evolution of
is not a one-step process. Biochem. J. 153, 141-144. the acid proteases. Nature 271, 618-621.
Fitzgerald, P.M.D., McKeever, B.M., van Middlesworth, J.F., Springer, Tang, J. & Wong, R.N.S. (1987). Evolution in the structure and func-
J.P., Heimbach, J.C., Leu, C.-T., Herber, W.K., Dixon, R.A.F., & tion of aspartic proteases. J. Cell. Biochem. 33, 53-63.
Darke, P.L. (1990). Crystallographic analysis of a complex between Tiselius, A , , Henschen, G.E., & Svensson, H. (1938). Electrophoresis
human immunodeficiency virus type 1 protease and acetyl-pepsta- of pepsin. Biochem. J. 32, 1814-1818.
tin at 2.0-A resolution. J. Bioi. Chem. 265, 14209-14219. Wlodawer, A., Miller, M., Jaskolski, M., Sathyanarayana, B.K., Bald-
Foundling, S.I., Salemme, F.R., Korant, B., Wendoloski, J. J., Weber, win, E.T., Weber, I.T., Selk, L.M., Clawson, L., Schneider, J., &
P.C., Treharne, A.C., Schadt, M.C., Jaskolski, M., Miller, M., Kent, S.B.H. (1989). Conserved folding in retroviral proteases:Crys-
Wlodawer, A,, Strop, P., Kostka, V., Sedlacek, J., & Ohlendorf, tal structure of a synthetic HIV-I protease. Science 245, 616-621.