Cloning of a virulence factor of Entamoeba histolytica: Pathogenic strains possess a unique cysteine proteinase gene

Abstract

Cysteine proteinases are hypothesized to be important virulence factors of Entamoeba histolytica, the causative agent of amebic dysentery and liver abscesses. The release of a histolytic cysteine proteinase from E. histolytica correlates with the pathogenicity of both axenic strains and recent clinical isolates as determined by clinical history of invasive disease, zymodeme analysis, and cytopathic effect. We now show that pathogenic isolates have a unique cysteine proteinase gene (ACP1). Two other cysteine proteinase genes (ACP2, ACP3) are 85% identical to each other and are present in both pathogenic and nonpathogenic isolates. ACP1 is only 35 and 45% identical in sequence to the two genes found in all isolates and is present on a distinct chromosome-size DNA fragment. Presence of the ACP1 gene correlates with increased proteinase expression and activity in pathogenic isolates as well as cytopathic effect on a fibroblast monolayer, an in vitro assay of virulence. Analysis of the predicted amino acid sequence of the ACP1 proteinase gene reveals homology with cysteine proteinases released by activated macrophages and invasive cancer cells, suggesting an evolutionarily conserved mechanism of tissue invasion. The observation that a histolytic cysteine proteinase gene is present only in pathogenic isolates of E. histolytica suggests that this aspect of virulence in amebiasis is genetically predetermined.

Full text

Cloning
of
a
Virulence
Factor
of
Entamoeba
histolytica
Pathogenic
Strains
Possess
a
Unique
Cysteine
Proteinase
Gene
Sharon
Reed,
*
Jacques
Bouvier,
Anna
Sikes
Pollack,
*
Juan
C.
Engel,t
Margaret
Brown,t
Ken
Hirata,
*
Xuchu
Que,*
Ann
Eakin,"
Per
Hagblom,1I
Frances
Gillin,*
and
James
H.
McKerrowt'
*Departments
of
Pathology
and
Medicine,
University
of
California,
San
Diego,
California
92103-8416;
Departments
of
$Pathology
and
§Pharmaceutical
Chemistry,
University
of
California,
San
Francisco,
California
94143;
the
tIDepartment
of
Microbiology,
University
of
Uppsala,
Uppsala,
Sweden;
and
the
'Department
of
Veterans
Affairs
Medical
Center,
San
Francisco,
California
94121
Abstract
Cysteine
proteinases
are
hypothesized
to
be
important
viru-
lence
factors
of
Entamoeba
histolytica,
the
causative
agent
of
amebic
dysentery
and
liver
abscesses.
The
release
of
a
histoly-
tic
cysteine
proteinase
from
E.
histolytica
correlates
with
the
pathogenicity
of
both
axenic
strains
and
recent
clinical
isolates
as
determined
by
clinical
history
of
invasive
disease,
zymodeme
analysis,
and
cytopathic
effect.
We
now
show
that
pathogenic
isolates
have
a
unique
cysteine
proteinase
gene
(ACP1).
Two
other
cysteine
proteinase
genes
(ACP2,
ACP3)
are
85%
identi-
cal
to
each
other
and
are
present
in
both
pathogenic
and
non-
pathogenic
isolates.
ACP1
is
only
35
and
45%
identical
in
se-
quence
to
the
two
genes
found
in
all
isolates
and
is
present
on
a
distinct
chromosome-size
DNA
fragment.
Presence
of
the
ACP1
gene
correlates
with
increased
proteinase
expression
and
activity
in
pathogenic
isolates
as
well
as
cytopathic
effect
on
a
fibroblast
monolayer,
an
in
vitro
assay
of
virulence.
Analysis
of
the
predicted
amino
acid
sequence
of
the
ACP1
proteinase
gene
reveals
homology
with
cysteine
proteinases
released
by
acti-
vated
macrophages
and
invasive
cancer
cells,
suggesting
an
evolutionarily
conserved
mechanism
of
tissue
invasion.
The
ob-
servation
that
a
histolytic
cysteine
proteinase
gene
is
present
only
in
pathogenic
isolates
of
E.
histolytica
suggests
that
this
aspect
of
virulence
in
amebiasis
is
genetically
predetermined.
(J.
Clin.
Invest.
1993.
91:1532-1540.)
Key
words:
amebiasis.
proteinases
*
cathepsins
*
pathogenicity
*
cytopathic
Introduction
Entamoeba
histolytica
infects
more
than
500
million
people
worldwide
(1).
Almost
from
the
time
of
its
discovery,
it
was
observed
that
although
E.
histolytica
most
often
causes
mild
or
asymptomatic
infections,
-10%
of
patients
develop
severe
dysentery
and
life-threatening
invasive
and
extraintestinal
dis-
ease
(2).
Whether
these
distinct
clinical
courses
represent
in-
fection
by
two
different
species
of
Entamoeba
(2)
or,
con-
versely,
whether
any
strain
of
E.
histolytica
can
potentially
cause
disease
in
the
presence
of
certain
environmental
factors
has
been
an
area
of
active
debate
(3,
4).
Distinct
isoenzyme
patterns
(zymodemes)
(5)
and
differences
in
restriction
frag-
ment-length
patterns
between
pathogenic
and
nonpathogenic
Address
correspondence
to
James
McKerrow,
Ph.D.,
M.D.,
Depart-
ment
of
Veterans
Affairs
Medical
Center,
San
Francisco,
Anatomic
Pathology
Service-
11
3B,
4150
Clement
St.,
San
Francisco,
CA
94121.
Received
for
publication
7
May
1992
and
in
revised
form
30
Oc-
tober
1992.
The
Journal
of
Clinical
Investigation,
Inc.
Volume
91,
April
1993,
1532-1540
strains
(6-10)
strongly
support
the
premise
that
the
potential
to
cause
invasive
disease
is
genetically
determined.
If
virulence
is
genetically
determined,
we
would
expect
a
gene
encoding
a
virulence
factor
that
correlated
with
pathogenicity
to
be
pres-
ent
only
in
pathogenic
strains.
We
tested
this
hypothesis
using
the
ameba
cysteine
proteinase.
Factors
associated
with
the
virulence
of
E.
histolytica
in-
clude
surface
lectins
(1
1,
12),
cytolytic
ion
channel-forming
proteins
(
13-15),
phospholipases
(
11
),
and
proteinases
(10,
16-18).
The
major
proteinase
of
E.
histolytica
is
a
cysteine
proteinase
biochemically
similar
to
cathepsin
B
(
17-20).
Ex-
perimental
evidence
supporting
the
role
of
this
cysteine
pro-
teinase
in
the
pathogenesis
of
amebiasis
includes
its
ability
to
degrade
fibronectin,
collagen,
and
basement
membrane
matrix
and
to
activate
the
third
component
of
complement
(
17-19,
21).
Purified
proteinase
reproduces
the
cytopathic
effect
of
pathogenic
amebae
(
17-20),
and
a
specific
irreversible
inhibi-
tor
of
the
proteinase
prevents
destruction
of
cell
monolayers
by
live
axenic
amebae
(22).
Increased
expression
and
excretion
of
the
proteinase
correlates
with
virulence
of
both
axenic
labora-
tory
strains
of
E.
histolytica,
and
fresh
clinical
isolates
(
16,
19,
20).
Antibodies
to
the
proteinase
were
detected
in
83%
of
pa-
tients
with
invasive
disease
but
were
not
detected
in
patients
with
noninvasive
infections
(
16).
Eukaryotic
cysteine
proteinases
have
long
been
recognized
as
particularly
potent
proteinases
but
were
best
known
for
their
role
in
intracellular
protein
digestion
(23).
More
recently,
evi-
dence
has
accumulated
that
a
subset
of
invasive
cancers
excrete
active
or
activatable
cysteine
proteinases.
A
cathepsin
L-like
cysteine
proteinase
is
the
major
excreted
protein
of
trans-
formed
fibroblasts
(24-26).
Levels
of
cathepsin
B
were
found
to
correlate
with
the
metastatic
potential
of
melanoma
vari-
ants
(24).
To
determine
whether
the
cysteine
proteinase
released
by
pathogenic
E.
histolytica
is
structurally
related
to
those
re-
leased
by
cancer
cells
and
activated
macrophages,
we
used
poly-
merase
chain
reaction
to
identify
and
amplify
genes
encoding
the
cysteine
proteinases
of
E.
histolytica
using
primers
based
on
conserved
structural
motifs
of
eukaryotic
cysteine
protein-
ases
(27).
We
identified
and
sequenced
three
cysteine
protein-
ase
genes
and
now
show
that
one
gene
is
unique
to
pathogenic
isolates
of
E.
histolytica.
Methods
Strains
of
E.
histolytica.
Axenic
strain
HM-
1IMSS
was
obtained
from
the
American
Type
Culture
Collection
(Rockville,
MD)
and
cultured
in
TYI-S-33
media
(28).
Nine
pathogenic
and
12
nonpathogenic
clini-
cal
isolates
were
cultured
directly
from
stools
or
liver
abscesses
into
Robinson's
media
as
previously
described
(29).
Three
strains
(SAW
760,
SAW
1453,
and
SAW
1734)
were
a
gift
from
Peter
Sargeaunt
(London
School
of
Hygiene
and
Tropical
Medicine),
and
five
strains
1532
Reed
et
al.

(FAT
957,
FAT
967,
FAT
973,
FAT
1
0
10,
and
FAT
10
14)
were
agift
from
T.H.F.G.
Jackson
(Research
Institute
for
Diseases
in
a
Tropical
Environment,
Durban,
South
Africa).
One
strain
was
cultured
from
a
patient
in
a
refugee
camp
in
Costa
Rica
(REF
29
1).
The
other
12
strains
(pathogenic:
SD
4,
53,
92,
135,
and
136;
nonpathogenic:
SD
1,
107,
116,
130,
137,
143,
and
147)
were
isolated
by
the
Microbiology
Laboratory
at
the
University
of
California
Medical
Center,
San
Diego.
Strains
were
assigned
to
zymodemes
by
the
method
of
Sargeaunt
et
al.
(5).
DNA
and
RNA
purification.
DNA
was
isolated
by
a
modification
of
the
method
of
Huber
et
al.
(30).
Cell
pellets
from
clinical
isolates
were
first
incubated
with
proteinase
K
(1
mg/ml)
for
2
h
at
560C.
Total
RNA
was
isolated
from
l0'
to
108
trophozoites,
according
to
the
method
of
Chomczynski
and
Sacchi
(3
1).
PCR
amplification
and
sequencing.
Degenerate
oligonucleotide
primers
based
on
active
site
sequences
conserved
in
all
eukaryotic
cys-
teine
proteinases
were
used
to
amplify
cysteine
proteinases
from
geno-
mic
HM-l
strain
DNA
by
the
PCR
under
conditions
previously
de-
scribed
(27).
A
new
5'
primer
(GCC
GAA
TTC
GCT
GCT
CCA
GAA
TCA
GTT
GAT
TGG
AGA)
based
upon
the
amino-terminal
sequence
of
ACP1I
was
used
with
the
original
3'
primer
to
amplify
and
isolate
the
two
additional
proteinase
gene
fragments
(ACP2
and
ACP3).
The
5'
sequence
of
the
mature
ACP2
proteinase
was
then
confirmed
using
a
new
3'
primer
(AAA
GGA
TCC
ACA
TGA
TCC
GCA
TTG
TC/
GC
TT).
The
remaining
sequence
of
ACP2
was
obtained
using
an
F.
histoly-
tica
cDNA
library
in
XgtlI
I
(strain
H-302:NIH;
a
gift
from
Dr.
Bruce
Torian)
(32).
DNA
was
purified
from
bacteriophage
plaques
by
bind-
ACP2
ing
to
DEAF-cellulose
(DE52;
Whatman
Inc.,
Clifton,
NJ).
The
5'
sequence
was
obtained
by
PCR
using
a
Xgt
11I
reverse
sequencing
primer
(Promega
Corp.,
Madison,
WI)
and
a
second
primer
corre-
sponding
to
a
seven-amino
acid
sequence
at
residues
18-25
of
the
mature
ACP2
enzyme
(5'-AGA
GTC
GAC
TGT
ATA
ACA
TGA
TCC
ACA
TTG
TCC
TTG
ATC).
After
PCR
amplification,
a
32
1-bp
fragment
that
encoded
the
entire
pro-sequence
was
obtained.
The
car-
boxy-terminal
sequence
was
similarly
obtained
by
PCR
using
the
for-
ward
sequencing
primer
for
Xgt
11I
(antisense)
and
a
sense
primer
made
to
the
same
seven-amino
acid
sequence.
Sequencing
was
performed
by
the
dideoxy
method
with
the
Sequenase
Kit
(U.S.
Biochemical
Corp.,
Cleveland,
OH)
or
by
automated
dye
terminator
sequencing
at
the
Biomolecular
Resource
Center,
University
of
California
at
San
Fran-
cisco,
using
an
ABI373A
instrument.
Genomic
library.
An
E.
histolytica
genomic
library
of
EcoRI
*-di-
gested
DNA
fragments
was
prepared
in
Lambda
ZAP
(Stratagene,
Inc.,
San
Diego,
CA).
The
ACP
I
probe
(452
bp)
was
labeled
by
the
random
primer
method
(Bethesda
Research
Laboratories,
Gaithersburg,
MD)
and
used
to
screen
2
x
106
plaques.
Five
clones
were
selected
after
tertiary
screening
and
sequenced.
Southern
and
Northern
analysis.
For
Southern
blot
analysis,
5-
10
jog
of
DNA
was
digested
with
EcoRI
and
subjected
to
electrophoresis
on
0.8%
GTG
agarose
gels
(FMC
Corp.,
Rockland,
ME)
and
trans-
ferred
to
Gene
Screen
Plus
nylon
membranes
(New
England
Nuclear
Research
Products,
Boston,
MA).
Hybridizations
were
performed
under
high
stringency
by
the
method
of
Church
and
Gilbert
(33)
in
7%
SDS,
0.5
M
NaHP04,
pH
7.2,
1
mM
EDTA,
and
1%
bovine
serum
albumin
at
650C.
Ethidium
bromide
staining
of
the
gel
and
hybridiza-
-89
gct
gca
A
A
-69
ACPI
T
H
N
K
V
F
A
N
R
A
E
-87
aca
cat
aac
aaa
gta
ttt
gct
aat
aga
gct
gaa
gga
att
cgg
att
gca
agt
gct
att
gat
ttc
aat
aca,
tgg
gct
tct
aaa
aac
aat
aaa
cac
ttc
aca
gca,
att
gaa
aag
ctt
aga
aga
ACP2
G
I
R
I
A
S
A
I
D
F
N
T
W
A
S
K
N
N
K
H
F
T
A
I
E
K L
R
K
-58
ACP1
Y
L
Y
R
F
A
V
F
L
S
N
K
K
F
V
E
A
N
A
N
TEL
N
V
F
G
D
M
tat
ctt
tac
aga
ttt
gct
gtt
ttc
tta
gac
aac
aaa
aaa
ttt
gtt
gaa
gct
aat
gct
aat
act
gaa
ctt
aat
gtt
ttt
ggt
gat
atg
aga
gct
atc
ttC
aat
atg
aat
gct
aaa
ttc
gtt
gat
agt
ttc
aat
aaa
att
ggt
tca
ttc
aaa
tta
tca
gta
gat
gga
cca
ttt
gct
ACP2
KR
A
I
F
N
M
N
A
K
F
V
D
SF
N
K
I
G
S
F
K
L S
V
S S
P
F
A
-29
-1
ACP1
T
HN
E
E
F
I
Q
T
H
L
G
M
T
Y
E
V
P
E
T
T
S
N
V
K
A
A
V
K
A
act
cac
gaa
gaa
ttc
atc
caa
act
cat
ctt
998
atg
act
tat
gaa
gtt
cca,
gaa
act
act
tct
aat
gtt
aaa
gct
9cc
gtt
aaa
gct
gct
atg
act
aat
gaa
gaa
tac
aga
act
ctt
ctt
aaa
tct
aaa
aga
act
act
gaa
gaa
aat
gga
caa
gtt
aaa
tat
ttq
aat
atc
caa
ACP2
A
MN
T
N
E E
Y
R
T
L
L
K
S
K
R
TT
E
EI
N G
S
V
K
Y
L
N
I
Q
ACP1
A
P
E
5
V
D
W
R
S
S
M
N
-
-
P
A
K
SD
Q
G
Q
C
G
S
C
W
T
F
C
gct
cca
gaa
tca
gtt
gat tgg
aga
agt
att
atg
aat
cca
gct
aaa
gat
caa
gga
caa
tgt
ggt
tca,
tgt
tgg
act
ttc
tgt
gca
cca
gaa
tca
gta
gat
tg9
aga
aaa
gaa
998
aaa
gta
act
cca
ctt
aga
gat
caa
gca
caa
tgc
gga
tca
tgt
tat
aca
ttt
ggt
ACP2
A
P
E
5
V
D
W
K
KE
G
K
V
T
P
L
R
SD
QA
Q
C
G
S
C
Y
T
F
G
28
ACP1
T
T
A
V
L
E
G
R
V
N
K
S
L
G
K
L
Y
S
F
S
E
Q
-
Q
L
V
S
C
S
aca,
act
gca
gtt
ctt
gaa
gga
aga
gtt
aac
aaa
gat
ctt
9ga
aaa
ctt
tac
tca
ttc
tct.
gaa
caa
caa
tta
qtt
gat
tgt
gat
tca
ctt
gca
gct.
ctt
gaa
gga
aga
tta
tta
att
gaa
aaa
gga
ggt
gat
gct
aat
aca
ctc
gat
ctt
tca
gaa
gaa
cat
atg
caa
tgc
ACP2
S
L
A A
L
E
GR
L
L
I
E
K
G
G
S
A
NT
L
S
L
S
E
E H
M
5
C
56
ACP1
A
S
D
N
-
-
-
G
CE
R
G
P
-
S
N
S
-
L
K
F
I
Q
E
N
N
G
L
G
gct
tct
gat
aat
gga
tgt
gaa
cga
gga
cca
tct
aac
tca
ctt
aaa
ttc
atc
caa
gaa
aat
aat
998
tta
998
aca
aga
gat
aat
9ga
aat
aat
9ga
tgt
aat
gga
gga
ctt
998
tca
aat
gtc
tat
gat
tac
att
att
gaa
cac
998
gtt
gct
ACP2
T
RN
S
N
G
N
N
G
C
N
G
S
L
G
S
N
V
Y
DI
Y
S
I
E
-
H
G
V
A
ACP1
L
E
S
S
Y
P
Y
K
A
V
A
G
T
C
K
-
K
V
K
N
V
A
T
V
T
G
S
R
N
tta
9aa
agc
gat
tat
cca
tat
aaa
gct
gtt
gct
ggt
act
tgc
aag
aaa
gtt
aaa
aac
gtt
gct
act
gtt
act
ggt
tct
aga
aga
aaa
gaa
agt
gat
tat
cca
tac
act
gga
agt
gat
tct
aca
tgc
aaa
act
aat
gta
aaa
tca
ttt
cgt
aaa
att
act
gga
tat act
aaa
ACP2
K
EK
S
S
Y
P
Y
T
S
S
D
S
T
C
K
T
N
V
K
S
F
N
K
I
T
G
Y
T
K
108
ACPI
V
T
S
G
S
E
T
S
L
5
T
I
S
A
E
N
S
P
V
A
V
S
M
S
A
S
N
P
5
gtt
act
gat
g98
agt
gaa
act
gga
ctt
caa
act
att
att
gct
gaa
aac
998
cct
gtt
gct
gtt
ggt
8tg
gat
gct
agc
898
cca
tca
gtc
cca
aga
aac
aat
gaa
gct
gaa
ctt
aaa
gct
gca
ctt
tca
caa
ggt
ctt
ctt
gat
gtt
tca
att
gat
gtc
tca
tct
gctaa
ACP2
V
P
R
N
N
I
A
E
L K
A
A
L
-
S
Q
G
L
L
S
V
S
S
S
V
S
S
A
K
137
ACP1
F
Q5
L
Y
K
K
S
T
I
Y
S
S
T
K
C
R
S
KR
N
M
N
H
C
V
T
A
V
ttc
caa
tta
tat
aag
aaa
gga
act
atc
tat
tct
gat
act
aaa
tgt
aga
tca
aga
atg
atg
aat
cac
t9t
gtt
act
gct
gtt
ttc
caa
tta
tac
889
agc
gga
gct
tat
act
Hat
act
aaa
tgc
aag
aat
aac
tac
ttt
gct
ttg
aat
cac
gaa
gtt
tgt
gct
gtt
ACP2
F
5
L
Y
K
S
G
-
A
Y
T
S
T
K
C
K
N
N Y
F
A
L
N
H
E
V
C
AV
ACP1
ACP2
164
G Y
S
S
N
S
N
S
K
Y
N
I
I
R
N
S
N
G
S
A
S
Y
F
L
L
Y
F
L
L
ggt
tat
ggt
tca
aat
agt
aat
ggt
aaa
tat
tgg
att
att
aga
aac
tca
tgg
998.
aca
tca
tgg
gga
gat
gct
998
tac
ttc
ctt
ctt
gga
tat
ggt
gtt
gtt
gat
998 888
988
tgt
tgg
ata
gtt
898
aac
tca
tgg
998
aca-tca
tgg
998
gat
888
998
tac
att
aat
atg
G
Y
G
V V
S
G
K
E
C
N
I
V
K
N
S
N
G
T
5
W
S S
K
S
Y
S
N
N
193
ACP1
A
N
S
S
N
N
M
C
S
S
G
N
S
S
N
Y
P
T
S
V
K
L
I
STOP
gct
898
gac
tcc
aac
aac
8tg
tgt
ggt
att
998
898
gat
tct
aac
tat
cca
acc
998
gtc
aag
tta
att
taa
gtt
att
988
998
aat
acc
-
tgt
9gt
gtt
gct
aca
gat
cca
ctt
tat
cca
act
ggc
gtt
caa
tat
ctt
tga
ACP2
V
I
E
S
N
T
-
C
G
V A
T
S
P
L
Y
P
T
G
V
Q
Y
L
STOP
Figure
1.
Nucleic
acid
and
predicted
protein
sequences
of
cysteine
proteinase
genes
from
F.
histolytica
HM-lI
strain.
ACP1I
and
ACP2
are
shown.
The
ACP3
gene
fragment
was
identical
to
a
cDNA
sequence
(cEh-CPp)
previously
published
(10);
the
amino
acid
sequence
of
ACP3
is
shown
in
Fig.
6.
Se-
quence
data
for
ACP
1
is
from
a
genomic
clone
and
includes
207
bp
of
proenzyme
se-
quence
5'
to
the
amino
terminus
of
the
ma-
ture
proteinase.
Residue
+1I
is
the
amino
ter-
minus
of
mature
proteinase.
Unique
Cysteine
Proteinase
Gene
of
Pathogenic
Entamoeba
histolytica
1533

1
2
A
12
3
1.9
Mb
220
Kb
co
G)
E
0
0
E
0
L-
(-)
1900
__-
1200
a-
550
_
-
_1
.
L-
C
U)CA3
I
I
C]
3
tion
of
the
blot
with
32P-labeled
E.
histolytica
actin
cDNA
(34),
a
gift
of
Dr.
Isaura
Meza,
confirmed
the
presence
of
equivalent
amounts
of
DNA
from
pathogenic
and
nonpathogenic
strains.
For
Northern
analy-
sis,
5-10
gg
of
total
RNA
was
subjected
to
electrophoresis
on
a
1%
agarose-2.2
M
formaldehyde
gel
in
3-(N-morpholino)propanesul-
*s
1900
fonic
acid
buffer
and
transferred
to
Gene
Screen
Plus
nylon
mem-
-F
1200
branes.
Hybridizations
were
performed
under
identical
conditions.
200
Ribonuclease
protection
assay.
Antisense
RNA
transcripts
were
synthesized
from
Bluescript
(pKS)
vectors
containing
the
ACPI
gene
cut
with
SmaI
and
ACP2
cut
with
EcoRI
to terminate
transcription
of
the
insert.
a-[32P]CTP-labeled
transcript
was
synthesized
with
the
T3
bacteriophage
polymerase
according
to
the
Maxiscript
kit
instructions
(Ambion,
Inc.,
Austin,
TX)
and
the
DNA
template
was
digested
with
RNase-free
DNase.
Aliquots
of
5-10
Iug
of
total
RNA
were
added
to
a
fourfold
molar
excess
of
labeled
probe
and
hybridized
at
420C
over-
night
using
the
RPAII
kit
(Ambion,
Inc.).
Unhybridized
RNA
was
digested
with
RNase
A,
RNA
hybrids
were
precipitated
and
pelleted,
-
550
and
the
fragments
were
separated
on
an
8
M
urea,
5%
acrylamide
gel.
The
intensity
of
the
resulting
bands
on
autoradiographs
was
compared
by
scanning
with
a
Quikscan
(Helena
Laboratories,
Beaumont,
TX).
Field
inversion
gel
electrophoresis
(FIGE).
Trophozoites
were
har-
vested
from
early
stationary
phase
cultures
and
washed
twice
by
centrif-
ugation
(10
min
at
200
g)
in
TSE
buffer
(100
mM
NaCI,
50
mM
EDTA,
20
mM
Tris
base,
pH
8.0).
An
equal
volume
of
melted
1.2%
InCert
Agarose
(FMC
Corp.)
in
TSE
was
added
to
the
parasite
pellet.
The
mixture
was
poured
into
63-,ul
wells
of
a
Hexa-A-Field
agarose
plug
mold
(Bethesda
Research
Laboratories)
and
allowed
to
gel
at
4°C.
The
agarose
blocks
were
transferred
to
a
tube
containing
10
ml
of
TSE
plus
1%
N-lauroylsarcosine,
1%
Nonidet
P-40,
and
incubated
for
3
h
with
gentle
agitation
at
4°C.
The
buffer
was
then
replaced
with
fresh
TSE,
1%
N-lauroylsarcosine
(TSE-Sarkosyl)
and
stored
at
4°C
for
a
minimum
of
18
h.
5
h
before use
the
agarose
plugs
were
transferred
to
a
new
tube
containing
TSE-Sarkosyl
with
2
mg/ml
Proteinase
K
and
incubated
at
45°C.
FIGE
was
performed
in
a
horizontal
electrophoresis
unit
(model
HE
100
SuperSub
Hoefer,
Scientific
Instruments,
San
Francisco,
CA)
connected
to
a
PC
750
pulse
controller
(model
PC
750;
Hoefer
Scien-
tific
Instruments).
The
agarose
plugs
were
loaded
into
wells
of
a
1.2%
agarose
gel
(Seakem
agarose;
FMC
Corp.)
made
with
1.Ox
TBE
buffer
(89
mM
Tris
base,
89
mM
boric
acid,
2
mM
EDTA,
pH
8.0).
DNA
molecules
were
separated
using
two
cycles
at
a
constant
voltage
of
125
V
at
8°C
in
l
x
TBE
buffer.
The
first
cycle
was
run
with
a
beginning
forward
pulse
of
2.4
s,
a
reverse
pulse
of
0.8
with
a
1.6
ramp
value
for
22-24
h.
The
second
cycle
lasted
20-24
h
with
an
initial
forward
pulse
of
3.6
s
and
1.2
reverse
pulse
with
a
1.6
ramp
value.
Lambda
DNA
ladder
and
Saccharomyces
cerevisiae
chromosomal
DNA
(FMC
Corp.)
were
included
to
estimate
the
molecular
size
of
the
E.
histolytica
chromosomes.
The
DNA
was
depurinated
with
0.2
N
HCI
for
10
min
and
denatured
with
0.5
N
NaOH
for
45
min
before
it
was
transferred
to
nylon
membranes.
Hybridizations
were
performed
in
30%
formamide,
6x
SSPE
(sodium
chloride,
sodium
phosphate,
EDTA),
5X
Den-
hardt's,
0.2%
SDS,
and100
jg/ml
transfer
RNA
(tRNA)
at
45°C
over-
night.
The
filters
were
washed
as
described
in
Sambrook
et
al.
(35)
at
a
final
temperature
of
68°C
for
30
min.
Proteinase
purification
and
peptide
sequencing.
HM-
1
trophozoites
were
washed
and
suspended
at
a
concentration
of
107/ml
in
PBS
and
48.5
KB
Figure
2.
(A)
Southern
blot
of
ACP1, ACP2,
and
ACP3
with
HM-l
strain
DNA.
Southern
hybridization
of
EcoRIl-digested
DNA
from
E.
histolytica
axenic
strain
HM-l
with
a
probe
corresponding
to
the
ACP
1
gene
(lane
1),
ACP2
gene
(lane
2),
and
ACP3
gene
(lane
3).
(B)
Southern
blot
after
FIGE
of
DNA
from
HM-1
strain
hybridized
with
ACP
1
and
ACP2
genes.
Lane
1,
ACP1;
lane
2,
ACP2
hybridized
to
same
blot;
lane
3,
both
probes
simultaneously
hybridized
to
second
blot.
ACP3
gave
identical
pattern
to
ACP2
(not
shown).
Note
unique
location
of
ACP
I
gene
copy
on
<
550-kb
"chromosome."
(C)
Ethi-
dium
bromide-stained
FIGE
gel
used
for
Southern
blot
transfer
in
B.
1534
Reed
et
al.

ACP
2
ACP
3
W
Path----
r--Nonpath--
,
P
H
_-
x
Uv§_
wo
;
IFe
ide
membrane
(Immobilon;
Millipore
Corp.,
Bedford,
MA)
by
the
method
of
Matsudaira
(36).
Peptide
sequencing
of
the
amino
terminus
was
performed
at
the
Biomolecular
Resource
Center
at
University
of
California
at
San
Francisco,
using
a
gas
phase
sequencer
(Applied
Bio-
systems
Inc.,
Foster
City,
CA).
Monolayer
assay
for
cytopathic
effect.
Trophozoites
were
purified
from
xenic
media
as
previously
described
(
16)
and
resuspended
in
PBS
containing
20
mM
cysteine,
0.
15
mM
CaCl2,
and
0.5
mM
MgCl2
at
a
concentration
of
107/ml.
After
incubation
for
3
h
at
370C,
the
super-
natants
were
separated
and
passed
through
a
0.2-PM
filter
to
remove
remaining
bacteria
from
the
medium.
This
procedure
did
not
affect
the
proteinase
activity
but
was
necessary
to
remove
any
residual
bacterial
flora,
which
is
different
for
each
clinical
isolate.
Proteinase
activity
was
quantified
by
the
cleavage
of
a
synthetic
peptide
substrate,
Z-Arg-Arg-
AMC
(benzyloxycarbonyl-arginine-arginine-4-amino-7-methylcou-
marin)
(Enzyme
Systems
Products,
Livermore,
CA),
as
previously
de-
scribed
(
17),
and
recorded
as
the
initial
velocity
of
cleavage
of
the
fluorescent
4-amino-7-methylcoumarin/
10
zd.
Proteinase
activity
was
inhibited
by
preincubating
with
Z-Phe-Arg-CH2F,
an
irreversible
cys-
teine
proteinase
inhibitor
(10
gM
for
30
min
at
370C),
which
has
no
effect
on
the
viability
of
the
monolayer
(Enzyme
Systems
Products).
24-well
culture
plates
were
seeded
with
2
x
105
cells
from
a
foreskin
epithelial
cell
line
(HFS
1)
and
incubated
overnight
in
DME
with
10%
fetal
calf
serum
in
CO2.
The
medium
supplemented
with
serum
was
removed
and
the
monolayer
washed
twice
with
MEM
without
sera.
Amebic
supernatants
(
100
gl)
were
added
in
triplicate
to
400
ul
of
MEM
and
incubated
for
3
h
at
37°C
with
CO2.
Detached
cells
were
removed
by
washing
twice
with
PBS.
The
remaining
cells
were
fixed
(4%
Formalin)
and
quantified
by
staining
with
methylene
blue
dye
(0.1%
in
0.1
M
borate,
pH
8.7)
and
measuring
the
extracted
absor-
bance
at
660
nM
(22).
Cytopathic
effect
(CPE)'
was
calculated
as
the
absorbance
of
control
wells
(MEM
alone)
minus
the
absorbance
of
sample
wells
divided
by
control
X
100%.
To
control
for
variations
between
different
experiments,
the
values
were
standardized
against
the
CPE
caused
by
strain
HM-1,
which
was
taken
as
1.00
(rela-
tive
CPE).
Results
Figure
3.
(A)
Southern
blot
of
ACPI
cysteine
proteinase
gene
with
DNA
of
clinical
isolates.
Southern
hybridization
of
EcoRI-digested
DNA
from
pathogenic
(Path)
and
nonpathogenic
(Nonpath)
clinical
isolates
with
ACP1
at
65°C.
Examples
of
three
of
nine
pathogenic
and
three
of
nine
nonpathogenic
isolates
are
shown.
(B)
Southern
blot
of
ACP2
and
ACP3
cysteine
proteinase
gene
with
DNA
of
clini-
cal
isolates.
Southern
hybridization
of
EcoRI-digested
DNA
with
ACP2
(three
of
six
pathogenic
and
three
of
five
nonpathogenic
strains
shown)
and
ACP3
(one
of
three
pathogenic
and
one
of
three
non-
pathogenic
strains
shown)
at
55°C.
incubated
for
3
h
at
37°C.
Cysteine
proteinase
activity
was
identified
and
the
enzyme
purified
from
the
supernatant
by
fast
protein
liquid
chromatography
as
previously
described
(
17
).
Both
the
high
molecular
mass
(56
kD)
and
low
molecular
mass
(27
kD)
forms
of
the
proteinase
were
purified.
The
purified
proteinase
was
subjected
to
electrophoresis
by
SDS-PAGE
on
a
10%
gel
and
transferred
to
a
polyvinylidenedifluor-
To
isolate
the
cysteine
proteinase
genes
of
E.
histolytica,
we
amplified
cysteine
proteinase
gene
fragments
by
PCR
using
primers
based
on
conserved
structural
motifs
identified
in
the
eukaryotic
cysteine
proteinase
family
(27).
A
450-bp
fragment
was
initially
amplified
from
DNA
isolated
from
the
pathogenic
axenic
strain
HM-1,
as
previously
reported
(27).
The
450-bp
fragment
was
subsequently
used
to
isolate
a
genomic
clone
(ACPl
)
containing
the
entire
coding
region
of
the
mature
pro-
teinase
(Fig.
1).
A
second
set
of
primers
based
on
the
amino
terminus
of
this
gene
and
the
region
around
the
active
site
asparagine
was
then
used
to
amplify
two
550-bp
genomic
frag-
ments
of
the
second
and
third
cysteine
proteinase
genes
(ACP2
and
ACP3)
(Fig.
1).
The
5'
end
of
ACP2,
representing
the
amino
terminus
of
the
mature
proteinase,
was
then
confirmed
by
sequence
from
genomic
DNA
using
a
3'
to
5'
primer
based
on
adjacent
downstream
sequence
(see
Methods).
The
5'
pro
sequence
and
the
carboxy-terminal
sequence
of
ACP2
were
obtained
by
PCR
from
DNA
purified
from
a
cDNA
library
(see
Methods).
The
sequence
of
the
ACP3
fragment
was
found
to
be
identical
to
the
sequence
of
a
cDNA,
cEh-CPp,
previously
published
by
Tannich
et
al.
(10).
To
confirm
that
these
genes
encoded
E.
histolytica
protein-
1.
Abbreviations
used
in
this
paper:
CPE,
cytopathic
effect;
FIGE,
field
inversion
gel
electrophoresis.
Unique
Cysteine
Proteinase
Gene
of
Pathogenic
Entamoeba
histolytica
1535
A
23.1-
9.1-
6.6-
4.4-
2.3-
2.0-
0.6-
B
23.1
-
9.1
-
6.6
-
4.4
-
2.3-
2.0
-
0.6
-
I

ases
and
were
expressed
in
the
HM-
1
strain,
we
purified
pro-
teinase
from
the
culture
supernatant
of
this
strain.
Microse-
quencing
of
the
first
eight
amino
acids
of
the
amino
terminus
of
the
purified
27-kD
enzyme
revealed
heterogeneity
at
posi-
tions
3
and
4
in
an
approximately
equimolar
ratio
(2.1:1.8),
A-P-E/K-S/A-V-D-W-R.
One
sequence
(APES
.
.)
was
iden-
tical
to
that
predicted
from
ACP
1
and
ACP2
and
also
identical
in
seven
out
of
eight
residues
with
the
amino-terminal
se-
quence
of
a
cysteine
proteinase
(called
"histolysin")
purified
by
Luaces
and
Barrett
(
18)
from
HM-
1
trophozoites.
The
sec-
ond
sequence
(APKA
.
.)
was
identical
to
the
first
eight
resi-
dues
of
the
cEh-CPp
proteinase
predicted
from
a
cDNA
clone
isolated
from
HM-
1
amebae
by
Tannich
et
al.
(10).
The
nu-
cleotide
sequence
of
cEh-CPp
is
identical
to
ACP3.
Southern
blot
analysis
showed
that
the
ACP
1
gene
frag-
ment
hybridized
with
a
single
restriction
fragment
in
EcoRI
(2.2
kb)-
(Fig.
2
A)
or
BglII
(3
kb)-
(not
shown)
digested
HM-
1
genomic
DNA
whereas
ACP2
and
ACP3
each
hybrid-
ized
to
two
larger
(
10
and
12
kb,
respectively)
EcoRI
fragments
(Fig.
2
A).
Southern
blot
analysis
after
FIGE
of
HM-
1
strain
(Fig.
2
B)
showed
that
the
ACP
1
gene
fragment
hybridized
with
two
DNA
bands
of
a
relative
size
of
-
1
Mb
and
550
kb.
In
contrast,
the
ACP2
and
ACP3
gene
fragments
hybridized
with
a
single
DNA
band
of
relative
size
1.9
Mb.
FIGE-derived
Southern
blots
hybridized
with
mixtures
of
ACP
1
and
ACP2
or
ACP3
gene
fragments
(Fig.
2
B,
lane
3)
showed
three
hybridiza-
tion
sites:
1.9
Mb
corresponding
to
ACP2/ACP3
genes,
and
1
Mb
and
550
kb
corresponding
to
the
ACP1
gene
as
described
above.
These
data
suggested
there
was
a
distinct
chromosomal
location
for
the
ACP
1
gene
versus
ACP2
and
ACP3.
The
detec-
tion
of
two
sites
of
hybridization
of
the
ACP1
probe
on
FIGE
suggested
more
than
one
gene
copy
was
present.
The
presence
of
multiple
gene
copies
was
also
supported
by
repeating
South-
ern
blot
analysis
under
conditions
of
varying
EcoRI
digestion
(0.1,
0.5,
2,
4,
and
8
U).
Six
hybridization
bands
were
detected
ranging
in
six
regular
increments
from
2.2
to
10
kb,
suggesting
multiple
gene
copies
are
present,
at
least
six
of
which
may
be
in
tandem.
This
pattern
was
reminiscent
of
the
multiple
tandem
copies
of
the
cysteine
protease
gene
of
Trypanosoma
cruzi
(37).
To
determine
whether
the
presence
of
any
of
these
genes
was
correlated
with
amebic
virulence,
DNA
was
extracted
from
nine
pathogenic
and
nine
nonpathogenic
clinical
isolates
clas-
sified
by
their
zymodeme
patterns
and
the
clinical
syndrome
of
the
patients
(5).
In
Southern
blot
analysis
of
EcoRI-digested
DNA
from
all
nine
pathogenic
isolates,
ACP1
hybridized
to
a
2.1-2.3-kb
fragment,
similar
to
that
seen
in
the
digest
of
HM-
I
DNA.
It
did
not
hybridize
to
any
DNA
fragment
from
any
of
the
nine
nonpathogenic
isolates
(Fig.
3
A).
A
portion
of
the
cysteine
proteinase
gene
from
one
pathogenic
clinical
isolate
(SAW
1453)
was
amplified
by
PCR. The
450-bp
amplified
sequence
from
this
isolate
was
identical
to
that
of
the
HM-
1
axenic
strain
(ACP
1)
shown
in
Fig.
1.
However,
there
was
slight
but
detectable
variation
in
size
(2.1-2.3
kb)
or
intensity
of
the
EcoRI
fragment
in
DNA
from
other
pathogenic
isolates,
suggesting
the
presence
of
some
sequence
heterogeneity
at
this
locus
among
different
isolates.
In
contrast,
ACP2
hybridized
to
2.2-,
4.0-,
and
5.5-kb
frag-
ments
and
ACP3
hybridized
to
2.2-, 2.9-,
and
3.6-kb
fragments
in
EcoRI-digested
DNA
from
six
pathogenic
and
five
nonpath-
ogenic
clinical
isolates
(Fig.
3
B).
Because
complete
genomic
clones
of
ACP2
and
ACP3
were
not
available,
it
is
difficult
to
A
kb
9.5
-
7.5
-
4.4-
2.4
-
1.4
-
0.24
-
kb
9.5
-
7.5
-
4.4-
2.4
-
1.4
-
0.24
-
Ax
r-Pa-
iNP
i
B
Ax
r-
P
---ir-
NP
-ir--B---,
I
.I._
Figure
4.
(A)
Northern
blot
with
ACP1.
Northern
blot
of
pathogenic
(P),
nonpathogenic
(NP),
and
bacteria
(B)
RNA
hybridized
with
ACP1
gene
probe.
Examples
of
three
of
five
pathogenic
and
two
of
five
nonpathogenic
isolates
are
shown.
(B)
Northern
blot
with
ACP2.
Same
blot
as
A
with
ACP2
gene
as
probe.
(C)
RNase
protection
assay.
Total
RNA
from
one
axenic
(HM-
1;
Ax),
three
pathogenic
(Path),
and
four
nonpathogenic
(Nonpath)
strains
was
hybridized
with
anti-
sense
RNA
transcripts
from
ACP1
and
ACP2.
assign
specific
map
positions
to
the
fragments
in
Fig.
3
B.
It
is
possible
some
may
be
the
result
of
partial
digestion
by
the
re-
striction
enzyme.
None
of
the
ameba
gene
probes
hybridized
to
1536
Reed
et
al.
L

Ax
r-Path-
I--
Nonpath-,
*i4
40
_ -
_,4
DNA
from
the
bacteria
in
the
cultures
of
clinical
isolates.
Northern
blots
probed
with
ACP1
revealed
a
strong
signal
at
1.2
kb
with
RNA
from
HM-l
and
five
of
five
pathogenic
strains,
but
only
a
faint
band
in
five
of
five
nonpathogenic
strains
(Fig.
4
A).
When
the
same
blot
was
hybridized
with
the
ACP2
probe,
a
band
of
equivalent
intensity
and
identical
mo-
lecular
weight
was
observed
with
all
strains.
The
ACP3
probe
also
hybridized
to
a
band
of
identical
size
in
all
three
patho-
genic
and
four
nonpathogenic
RNA
samples
tested
(not
shown).
These
Northern
blots
suggested
that
ACP2
and
ACP3
were
expressed
in
all
isolates
whereas
ACP1
was
expressed
only
in
pathogenic
isolates
but
crosshybridized
weakly
with
the
other
genes
(faint
bands
in
nonpathogenic
lanes
in
Fig.
4
A).
To
confirm
this
result
with
a
more
sensitive
measure
of
poten-
tially
low
abundance
mRNAs,
ribonuclease
protection
assays
were
performed
using
antisense
RNA
transcripts
synthesized
from
ACP1
and
ACP2.
A
1.2-kb
band
was
detected
in
all
iso-
lates
with
the
ACP2
probe
whereas
the
ACP1
probe
hybridized
strongly
only
with
transcripts
from
the
pathogenic
isolates
(Fig.
4
C).
To
directly
correlate
the
presence
of
the
ACP1
gene
in
pathogenic
amebae
with
enhanced
proteinase
activity
and
a
quantitative
assay
of
virulence,
the
total
cysteine
proteinase
activity
of
amebic
supernatants
was
determined
and
correlated
in
the
same
ameba
isolates
with
the
cytopathic
effect
on
a
fibro-
blast
cell
monolayer.
The
proteinase
activity
(initial
velocity/
10
Al)
of
five
pathogenic
strains
containing
ACP1
was
signifi-
cantly
greater
than
that
of
four
nonpathogenic
strains
lacking
it
(P
<
0.03)
(Fig.
5).
The
relative
CPE
(corrected
to
that
of
HM-1
as
1.00)
by
pathogenic
strains
was
also
significantly
greater
(P
<
0.001
)
and
was
completely
inhibited,
as
demon-
strated
previously
(22),
by
preincubation
with
100
,AM
Z-Phe-
Arg-CH2F,
an
irreversible
and
specific
cysteine
proteinase
in-
hibitor.
Furthermore,
the
enhanced
proteinase
activity
corre-
lated
with
elevated cysteine
proteinase
mRNA
in
the
pathogenic
versus
nonpathogenic
isolates
(Fig.
5).
In
sum-
mary,
Fig.
5
shows
that
the
presence
of
an
additional
cysteine
proteinase
gene
(ACPl
)
in
pathogenic
isolates
correlates
with
increased
proteinase
mRNA,
increased
proteinase
activity,
and
[1.2
kb
-1.2kb
J
Figure
4.
(Continued)
icreased
CPE
that
is
inhibited
by
Z-Phe-Arg-CH2F,
a
cysteine
roteinase
inhibitor.
The
coding
sequence
of
ACP1
predicted
an
amino
acid
quence
with
only
35-45%
identity
to
the
other
two
genes.
In
ntrast,
ACP2
and
ACP3
were
85%
identical
to
each
other
in
Total
Proteinase
Activity
(VO
q1)
0.5
Total
RNA
Densitometer
Units)
Path
Nonpath
ACP
I
+
ACP
2
+
+
ACP3
+
+
Figure
5.
Proteinase
activity
(measured
as
initial
velocity,
Vo/
10
ml)
of
supernatants
of
clinical
isolates
(
1O0
amebae/ml)
compared
with
the
relative
CPE
(measured
as
percent
of
monolayer
destruction/
percent
destruction
by
axenic
strain
HM-
1)
and
total
cysteine
pro-
teinase
mRNA.
Proteinase
activity
is
the
mean±SEM
for
13
mea-
surements
of
five
pathogenic
strains
and
18
measurements
of
four
nonpathogenic
strains.
CPE
is
mean±SEM
for
five
pathogenic
and
four
nonpathogenic
isolates.
Sum
of
mRNA
for
ACP1
and
ACP3
assayed
by
densitometer
scan
of
signals
from
RNase
protection
assays
of
three
pathogenic
and
four
nonpathogenic
isolates.
+
or
-
indicates
presence
or
absence
of
each
gene
by
Southern
blot
analysis
(see
Fig.
2).
Relative
CPE
Unique
Cysteine
Proteinase
Gene
of
Pathogenic
Entamoeba
histolytica
1537
c
ACPI
4I04**
I
2
ACP
2
_i_
_
w.
_'_
w

predicted
amino
acid
sequence
and
90
and
100%
identical
at
the
nucleic
acid
level
to
the
two
corresponding
genes
(cEh-
CPnp
and
cEh-CPp)
isolated
previously
by
Tannich
et
al.
(
10
).
All
three
genes
were
also
homologous
to
the
major
excreted
protein
of
transformed
fibroblasts
(26)
as
well
as
other
members
of
the
cysteine
proteinase
family
of
enzymes
(Fig.
6).
Specifically,
all
three
amino
acid
residues
of
the
catalytic
triad
of
cysteine
proteinases
(cysteine,
histidine,
and
asparagine)
were
present,
and
structural
motifs
flanking
these
residues
were
highly
conserved
(Fig.
6).
Discussion
The
potential
multifactorial
roles
of
cysteine
proteinases
in
in-
vasion
of
pathogenic
amebae
are
well
documented
(
16-22).
They
include
degradation
of
host
extracellular
matrix
and
mu-
coproteins,
dislodgment
of
epithelial
cells
and
degradation
of
epithelial
basement
membrane,
and
possibly
recruitment
of
inflammatory
cells
to
sites
of
ameba
invasion
by
activation
of
complement.
We
have
now
shown
that
the
enhanced
expression
and
re-
lease
of
cysteine
proteinase
activity
by
virulent
laboratory
strains
(
17,
20)
and
pathogenic
clinical
isolates
(
16)
correlates
with
the
presence
of
a
unique
cysteine
proteinase
gene
(ACP
1),
which
was
not
detected
in
Southern
blots
of
DNA
of
nonpatho-
genic
amebae.
Tannich
et
al.
(
10)
also
identified
restriction
fragment-length
polymorphisms
between
pathogenic
and
nonpathogenic
strains
by
hybridization
with
two
other
cysteine
proteinase
genes.
One
of
these
genes
(cEh-CPnp)
may
corre-
spond
to
ACP2
and
the
other
(cEh-CPp)
is
identical
to
ACP3.
However,
our
examination
of
a
larger
set
of
clinical isolates
suggests
that
neither
of
these
latter
two
closely
related
genes
is
unique
to
nonpathogenic
or
pathogenic
amebae
(Fig.
2).
We
speculate
that
ACP2
and
ACP3
arose
from
copies
of
the
same
gene
and
remain
closely
linked
on
one
chromosome.
Their
presence
in
both
pathogenic
and
nonpathogenic
isolates
of
E.
ACP1
ACP2
ACP3
MEP
CB
p
GCERGHP
SNSLKFIQENNGLGLESDYPYKAVAGTCKKVKNVATVTGSRRVTD
GCNGGLG
SNVYDYIIE
NGVAKESDYPYTGSDSTCKTNVKSFRKITGYTKVP
GCNGGLG
SNVYNYIME
NGIAKESDYPYTGSDSTCRSDVKAFAKIKSYNRVA
GCNGGLM
DYAFQYVQD
NGGLDSEESYPYEATEESCKYNPKYSVANDTGFDIPKQE
GCNGGYP
AEAWNFWTR
KGLVSGGYRSHVGCRPYSIPPCEHHVNGSRPPCTGEGDTRKCSKICEP
GCNGGYP
WSALQLVAQ
YGIHYRNTPYYEGVQRYCRSREKGPYAAKTDGVRQVQPY
66
100
histolytica
suggests
they
may
play
a
role
in
basic
metabolism
of
amebae,
most
likely
intracellular
protein
degradation.
ACP
1
is
quite
divergent
in
sequence
from
ACP2
and
ACP3,
and
copies
of
the
ACP
1
gene
are
present
on
different
chromo-
some-sized
DNA
fragments
than
ACP2
and
ACP3
(Fig.
2
B).
Nevertheless,
the
data
from
both
studies
presented
here
to-
gether
with
those
reported
by
Tannich
et
al.
(
10)
are
consistent
with
the
presence
of
an
additional
cysteine
proteinase
gene
in
pathogenic
strains,
resulting
in
increased
cysteine
proteinase
mRNA
and
consequently
higher
levels
of
extracellular
proteo-
lytic
activity.
Our
observation
that
amino-terminal
sequencing
of
the
E.
histolytica
cysteine
proteinase
yields
more
than
one
signal
(APE
[S]
K[A]
.
.)
also
supports
the
hypothesis
that
the
enhanced
production
and
release
of
cysteine
proteinases
by
virulent
E.
histolytica
is
due
to
the
presence
and
expression
of
multiple
genes.
Although
our
Southern
blot
analysis
(Fig.
3
A)
suggested
that
the
ACP1
gene
was
unique
to
pathogenic
iso-
lates,
Orozco
(38)
has
offered
an
alternative
explanation
for
these
results.
She
speculates
that
gene
amplification
may
lead
to
enhanced
expression
of
virulence
factors
in
pathogenic
ame-
bae.
We
cannot
exclude
the
possibility
that
our
hybridization
conditions
did
not
identify
a
single
copy
of
ACP
I
in
nonpatho-
genic
isolates
and
that
the
gene
is
highly
amplified
in
patho-
genic
DNA,
leading
to
enhanced
cysteine
proteinase
expres-
sion.
In
fact,
the
results
of
varying
EcoRI
digestion
conditions
(0.
1-8
U
of
enzyme)
suggested
that
multiple
tandem
copies
of
ACP
1
were
present.
FIGE
also
suggested
that
there
were
at
least
two
distinct
chromosome
locations
for
ACP
1
gene
copies
(Fig.
2
B).
In
either
case,
we
speculate
that
in
pathogenic
isolates
the
expression
of
the
ACP
1
gene(s)
in
addition
to
expression
of
ACP2
and
ACP3
genes
results
in
extracellular
proteinase
activ-
ity
exceeding
a
critical
threshold,
which,
coupled
with
other
virulence
factors
(
1
1-18
),
may
lead
to
tissue
lysis,
invasion
of
the
bowel
wall
by
trophozoites,
and
disseminated
infection.
There
are
striking
parallels
in
the
release
of
cathepsin
L
and
B
by
invasive
cancer
cells
or
activated
macrophages
and
cys-
GSETGLQTIIAENGPVAVGMDASRPSFQLYK
KGTIYSDTKCRSR
MMN
H
CVTAV
RNNEVELKAALSQGLLDVSIDVSSAKFQLYK
GGAYTDTKCKNNYEALN
H
QVCAV
RNNEVELKAAISQGLVDVSIDASSVQFQLYK
SGAYTDTQCKNNYFALN
H
EVCAV
KALMKAVATVGPISVAID
AGHESFLFYK
EGIYFEPCDSSE
DMD
H
GVLVV
GYSPTYKNGPVAF
SVYSDFLLYK
SGVYQHVTGEM
MCG
H
AIRIL
NQGALLYSIANQPVSWLQ
AAGKDFQLYR
GGIFVGPCGN
KVD
H
AVAAV
125
159
GYG
SNSNG
KYWIVKNSWGTSWGDAGYFL
LARDSNNM
CGIGRDDSNYPTGVKLI
GYG
VVDGKEC
WVVRNSWGTSWGDKGYINMVIEGNT
CGVATDPLYPTCVQYL
GYG
VADGKEC
WIVRNSWGTGWGEKGYINMVIEGNT
CGVATDPLYPTGVEYL
GYG
FESTESDNN
KYWLVKNSWGEFWGMGGYIK
IAKDRDNH
CGLATAASYPVVN
GWG
VENGTP
YWLVANSWNTDWGNGFFK
ILRGQDH
CGIESEWAGIPRTD
GYN
PG
YILVANSWGTGWGENGYIRIKRGTGNSYGV
CGLYTSSFYPVKN
primer
2
Figure
6.
Alignment
of
predicted
sequences
of
the
cysteine
proteinases
of
E.
histolytica
(ACP1,
ACP2,
and
ACP3)
to
the
secreted
cysteine
proteinase
(cath-
epsin
L)
of
transformed
fibroblasts
(MEP),
human
cathepsin
B
(CB),
and
papain
(P).
Active
site
resi-
dues
marked
by
asterisk.
Conserved
regions
used
for
PCR
primers
are
indicated.
Sequence
of
ACP3
deter-
mined
from
this
study
is
from
primer'
to
primer2
site;
the
remaining
sequence
shown
is
that
published
for
the
identical
gene
cEh-CPp
(10).
ACP1
APESVDWR
SIMN
PAKDQGQCGSCWTF
CTTAVLEGRVNKDLGKLYSF
SEQQLVDCDASDN
ACP2
APESVDWR
KEGKVT
PIRDQAQCGSCYTF
GSLAALEGRLLIEKGDANT
LEEHMVQCTRDNGNN
ACP3
APKAVDWR
KKGKVT
PIRDQGNCGSCYTF
GSIAALEGRLLIEKGGDSETLDL
SEEHMVQCTREDGNN
MEP
APRSVDWR
EKGYVT
PVKNQGQCGSCWAF
SATGALEGQMFRKTGRLISL
SEQNLVDCSGPQGNE
CB
LPASFDAR
EQWPQCPTIKE
IRDQGQSCGSCWAF
GAVEAISDRICHIHTNVSVEV
SAEDLLDCCGIQCGD
P
IPEYVDWR
QKGAVT
PVKNQGSCGSCWAF
SAVVTIEGIIKIRTGNLNQY
SEQELLDCDRRSY
primer
3
primem
1
1
25
50
ACP
1
ACP2
ACP3
MEP
CB
p
ACP
1
ACP2
ACP3
MEP
CB
p
1538
Reed
et
al.

teine
proteinase
release
by
pathogenic
amebae.
In
each
case,
cysteine
proteinases
are
found
both
in
intracellular
digestive
organelles
as
well
as
released
extracellularly
in
multiple
molecu-
lar
forms,
including
active
or
activatable
proenzymes
(
17,
24-
26).
Like
the
targeting
of
cysteine
proteinases
to
lysosomes
or
endosomes
in
mammalian
cells,
the
cysteine
proteinases
of
both
pathogenic
and
nonpathogenic
strains
of
E.
histolytica
appear
to
be
targeted
at
least
in
part
to
endosome-like
cytoplas-
mic
vacuoles
where
endocytosed
bacteria
and
cells
are
de-
graded
(39).
Default
secretion
due
to
alterations
in
mannose-
6-phosphate
containing
signal
moieties
or
transcriptional
over-
expression
of
specific
proteinase
genes
are
two
mechanisms
by
which
transformed
mammalian
cells
are
thought
to
release
lyso-
somal
cysteine
proteinases
into
the
extracellular
milieu
(26).
In
amebae
there
is
a
correlation
between
elevated
rates
of
phagocytosis
and
virulence
(40),
and
phagocytosis-deficient
mutants
release
less
extracellular
cysteine
proteinase
(22).
The
adaptive
advantage
for
the
pathogenic
amebae
of
an
additional
cysteine
proteinase
gene
and,
consequently,
higher
levels
of
extracellular
cysteine
proteinase
may
therefore
be
related
to
enhanced
phagocytosis
and/or
extracellular
digestion
of
bacte-
ria
and
cells.
At
this
time
we
cannot
exclude
the
possibility
that
the
dif-
ferences
in
the
coding
sequence
of
the
three
cysteine
proteinase
genes
of
E.
histolytica
might
reflect
distinct
transport
or
bio-
chemical
properties.
However,
at
least
the
specificity
for
syn-
thetic
peptide
substrates
among
the
three
gene
products
ap-
pears
to
be
identical
(17-20)
and,
unlike
their
mammalian
counterparts,
none
of
the
ameba
cysteine
proteinases
contain
asparagine-linked
carbohydrate
addition
sites
for
mannose-
6-phosphate
lysosome
targeting
signals
(Fig.
6)
(10).
All
previous
studies
of
virulence
have
used
axenized
patho-
genic
strains
that
have
become
attenuated
to
varying
degrees
during
cultivation.
A
direct
comparison
to
nonpathogenic
strains
has
not
been
possible
because
these
strains
have
never
been
axenized.
On
the
basis
of
the
observations
reported
here,
we
can
now
confirm
that
there
are
quantitative
differences
in
CPE
between
authentic
pathogenic
and
nonpathogenic
clinical
isolates,
which
in
turn
correlate
with
the
presence
of
the
ACP1
gene
in
pathogenic
strains
and
enhanced
proteinase
release
(Fig.
5).
Noteworthy,
however,
is
the
observation
that
CPE
is
not
totally
absent
in
nonpathogenic
isolates
and
therefore
can-
not
be
used
alone
as
an
assay
of
pathogenicity,
as
was
suggested
by
its
historical
use
with
axenic
strains.
We
also
confirm
pre-
vious
studies
(22)
that
CPE
is
inhibited
by
a
specific
inhibitor
of
cysteine
proteinases.
It
is
likely
that
host
immune
status,
nutritional
status,
and
associated
intestinal
bacteria
influence
susceptibility
to
coloni-
zation
and
invasion
by
E.
histolytica.
However,
our
identifica-
tion
of
a
cysteine
proteinase
gene
that
is
specific
to
pathogenic
strains
and
correlates
with
increased
extracellular
proteinase
activity
argues
strongly
that
the
potential
for
pathogenicity
is
an
intrinsic
property
of
the
organism.
Acknowledgments
This
work
was
funded
in
part
by
the
Lucille
P.
Markey
Foundation,
National
Institutes
of
Health
grants
AI-28035
(S.
Reed)
and
DK-
35108
(S.
Reed),
and
the
Department
of
Veterans
Affairs
(J.H.
McKerrow).
S.
Reed
is
a
Lucille
P.
Markey
Scholar.
J.
Bouvier
was
a
Swiss
National
Science
Foundation
Fellow.
References
1.
Walsh,
J.
A.
1988.
Prevalence
of
Entamoeba
histolytica
infection.
In
Ame-
biasis.
Human
Infection
by
Entamoeba
histolytica.
J.
I.
Ravdin,
editor.
John
Wiley
and
Sons,
New
York. 93-105.
2.
Brumpt,
E.
1925.
Etude
sommaire
de
l'Entamoeba
dispar
n.
sp.
amibe
a
kystes
quadrinuclees,
parasite
de
l'homme.
Bull.
Acad
Nati.
Med.
(Paris).
94:942-952.
3.
Mirelman,
D.,
R.
Bracha,
A.
Wexler,
and
A.
Chayen.
1986.
Changes
in
isoenzyme
patterns
of
a
cloned
culture
of
nonpathogenic
Entamoeba
histolytica
during
axenization.
Infect.
Immun.
54:827-832.
4.
Sargeaunt,
P.
G.
1987.
The
reliability
of
Entamoeba
histolytica
zymodemes
in
clinical
diagnosis.
Parasitol.
Today.
3:40-43.
5.
Sargeaunt,
P.
G.,
J.
E.
Williams,
and
J.
D.
Greene.
1978.
The
differentia-
tion
of
invasive
and
noninvasive
Entamoeba
histolytica
by
isoenzyme
electropho-
resis.
Trans.
R.
Soc.
Trop.
Med.
Hyg.
72:519-521.
6.
Tannich,
E.,
R.
D.
Horstmann,
J.
Knobloch,
and
H. H.
Arnold.
1989.
Genomic
DNA
differences
between
pathogenic
and
nonpathogenic
Entamoeba
histolytica.
Proc.
Natl.
Acad.
Sci.
USA.
86:5118-5122.
7.
Garfinkel,
L.
I.,
M.
Giladi,
M.
Huber,
C.
Gitler,
D.
Mirelman,
M.
Revel,
and
S.
Rozenblatt.
1989.
DNA
probes
specific
for
Entamoeba
histolytica
possess-
ing
pathogenic
and
nonpathogenic
zymodemes.
Infect.
Immun.
57:926-931.
8.
Que,
X.,
and
S.
L.
Reed.
1991.
Nucleotide
sequence
of
a
small
subunit
ribosomal
RNA
(I
6s-like
rRNA)
gene
from
Entamoeba
histolytica:
differentia-
tion
of
pathogenic
from
nonpathogenic
isolates.
Nucl.
Acids
Res.
19:5438.
9.
Edman,
U.,
M.
A.
Meraz,
S.
Rausser,
N.
Agabian,
and
I.
Meza.
1990.
Characterization
of
an
immuno-dominant
variable
surface
antigen
from
patho-
genic
and
nonpathogenic
Entamoeba
histolytica.
J.
Exp.
Med.
172:879-888.
10.
Tannich,
E.,
H.
Scholze,
R.
Nickel,
and
R.
D.
Horstmann.
1991.
Homolo-
gous
cysteine
proteinases
of
pathogenic
and
nonpathogenic
Entamoeba
histoly-
tica.
J.
Biol.
Chem.
266:4798-4803.
1
1.
Ravdin,
J.
I.
1986.
Pathogenesis
of
disease
caused
by
Entamoeba
histoly-
tica:
studies
of
adherence,
secreted
toxins,
and
contact-dependent
cytolysis.
Rev.
Infect.
Dis.
8:247-260.
12.
Petri,
W.
A.,
R.
D.
Smith,
P.
H.
Schlesinger,
C.
F.
Murphy,
and
J.
I.
Ravdin.
1987.
Isolation
of
the
galactose-binding
lectin
that
mediates
the
in
vitro
adherence
of
Entamoeba
histolytica.
J.
Clin.
Invest.
80:1238-1244.
13.
Gitler,
C.,
E.
Calef,
and
I.
Rosenberg.
1984.
Cytopathogenicity
of
Enta-
moeba
histolytica.
Philos.
Trans.
R.
Soc.
Lond.
Biol.
Sci.
B307:73-85.
14.
Young,
J.
D.,
T.
M.
Young,
L.
P.
Lu,
J.
C.
Unkeless,
and
Z.
A.
Cohn.
1982.
Characterization
of
a
membrane
pore-forming
protein
from
Entamoeba
histolytica.
J.
Exp.
Med.
156:1677-1690.
15.
Leippe,
M.,
E.
Sebastian,
0.
L.
Schoenberger,
R.
D.
Horstmann,
and
H.
J.
Mtuller-Eberhard.
1991.
Pore-forming
peptide
of
pathogenic
Entamoeba
histoly-
tica.
Proc.
Natl.
Acad.
Sci.
USA.
88:7659-7663.
16.
Reed,
S.
L.,
W.
E.
Keene,
and
J.
H.
McKerrow.
1989.
Thiol
proteinase
expression
correlates
with
pathogenicity
of
Entamoeba
histolytica.
J.
Clin.
Micro-
biol.
27:2772-2777.
17.
Keene,
W.
E.,
M.
G.
Pettit,
S.
Allen,
and
J.
H.
McKerrow.
1986.
The
major
neutral
proteinase
of
Entamoeba
histolytica.
J.
Exp.
Med.
163:536-549.
18.
Luaces,
A.
L.,
and
A.
J.
Barrett.
1988.
Affinity
purification
and
biochemi-
cal
characterization
of
Histolysin,
the
major
cysteine
proteinase
of
Entamoeba
histolytica.
Biochem.
J.
250:903-909.
19.
Scholze,
H.,
and
E.
Werries.
1986.
Cysteine
proteinase
of
Entamoeba
histolytica.
I.
Partial
purification
and
action
on
different
enzymes.
Mol.
Biochem.
Parasitol.
18:103-112.
20.
Lushbaugh,
W.
B.
1988.
Proteinases
of
Entamoeba
histolytica.
In
Ame-
biasis:
Human
Infection
by
Entamoeba
histolytica.
J.
I.
Ravdin,
editor.
John
Wiley
and
Sons,
New
York.
219-231.
21.
Reed,
S.
L.,
W.
E.
Keene,
J.
H.
McKerrow,
and
I.
Gigli.
1989.
Cleavage
of
C3
by
a
neutral
cysteine
proteinase
of
Entamoeba
histolytica.
J.
Immunol.
143:189-195.
22.
Keene,
W.
E.,
M.
E.
Hidalgo,
E.
Orozco,
and
J.
H.
McKerrow.
1990.
Entamoeba
histolytica:
correlation
of
the
cytopathic
effect
of
virulent
tropho-
zoites
with
secretion
of
a
cysteine
proteinase.
Exp.
Parasitol.
71:199-206.
23.
Barrett,
A.
J.,
and
H.
Kirschke.
1981.
Cathepsin
B,
Cathepsin
H,
and
Cathepsin
L.
Methods
Enzymol.
80:535-561.
24.
Sloane,
B.
F.,
J.
R.
Dunn,
and
K.
V.
Honn.
1981.
Lysosomal
Cathepsin
B:
correlation
with
metastatic
potential.
Science
(Wash.
DC).
212:1151-1153.
25.
Mort,
J.
S.,
M.
Leduc,
and
A.
D.
Recklies.
1981.
A
latent
thiol
proteinase
from
ascitic
fluid
of
patients
with
neoplasia.
Biochim.
Biophys.
Acta.
662:173-
180.
26.
Troen,
B.
R.,
D.
Ascherman,
D.
Atlas,
and
M.
M.
Gottesman.
1988.
Cloning
and
expression
of
the
gene
for
the
major
excreted
protein
of
transformed
mouse
fibroblasts:
a
secreted
lysosomal
protease
regulated
by
transformation.
J.
Biol.
Chem.
263:254-261.
27.
Eakin,
A.
E.,
J.
Bouvier,
J.
A.
Sakanari,
C.
S.
Craik,
and
J.
H.
McKerrow.
1990.
Amplification
and
sequencing
of
genomic
DNA
fragments
encoding
cys-
teine
proteases
from
protozoan
parasites.
Mo/.
Biochem.
Parasitol.
39:1-8.
28.
Diamond,
L.
S.,
D.
R.
Harlow,
and
C.
C.
Cunnick.
1978.
A
new
medium
Unique
Cysteine
Proteinase
Gene
of
Pathogenic
Entamoeba
histolytica
1539

for
the
axenic
cultivation
of
Entamoeba
histolytica
and
other
Entamoeba.
Trans.
R.
Soc.
Trop.
Med.
Hyg.
72:431-432.
29.
Robinson,
G.
L.
1968.
The
laboratory
diagnosis
of
human
parasitic
amoe-
bae.
Trans.
R.
Soc.
Hyg.
Trop.
Med.
Hyg.
62:285-294.
30.
Huber,
M.,
L.
Garfinkel,
C.
Gitler,
D.
Mirelman,
M.
Revel,
and
S.
Ro-
zenblatt.
1987.
Entamoeba
histolytica:
cloning
and
characterization
of
actin
cDNA.
Mol.
Biochem.
Parasitol.
24:227-235.
31.
Chomczynski,
P.,
and
N.
Sacchi.
1987.
Single-step
method
of
RNA
isola-
tion
by
acid
guanidinium
thiocyanate-phenol-chloroform
extraction.
Anal.
Bio-
chem.
162:156-159.
32.
Torian,
B.
E.,
B.
M.
Flores,
V.
L.
Stroeher,
F.
S.
Hagen,
and
W.
E.
Stamm.
1990.
cDNA
sequence
analysis
of
a
29-kDa
cysteine-rich
surface
antigen
of
patho-
genic
Entamoeba
histolytica.
Pro.
NatL.
Acad.
Sci.
USA.
87:6358-6362.
33.
Church,
G.
M.,
and
W.
Gilbert.
1984.
Genomic
sequencing.
Proc.
Nati.
Acad.
Sci.
USA.
81:1991-1995.
34.
Edman,
U.,
I.
Meza,
and
N.
Agabian.
1987.
Genomic
and
cDNA
actin
sequences
from
a
virulent
strain
of
Entamoeba
histolytica.
Proc.
NatL.
Acad.
Sci.
USA.
84:3024-3028.
35.
Sambrook,
J.
C.,
E.
F.
Fritsch,
and
T.
Maniatis.
1989.
Molecular
Cloning:
A
Laboratory
Manual.
2nd
ed.
C.
Nolan,
editor.
Cold
Spring
Harbor
Press,
Cold
Spring
Harbor,
New
York.
36.
Matsudaira,
P.
1987.
Sequence
from
picomole
quantities
of
proteins
elec-
troblotted
onto
polyvinylidene
difluoride
membranes.
J.
Biol.
Chem.
262:10035-10038.
37.
Eakin,
A.
E.,
A.
A.
Mills,
G.
Harth,
J.
H.
McKerrow,
and
C.
S.
Craik.
1992.
The
sequence,
organization,
and
expression
of
the
major
cysteine
protease
(cruzain)
from
Trypanosoma
cruzi.
J.
Bio.
Chem.
267:7411-7420.
38.
Orozco,
E.
1992.
Pathogenesis
in
amebiasis.
Infectious
Agents
and
Dis-
ease.
1:19-21.
39.
De
Meester,
F.,
E.
Shaw,
H.
Scholze,
T.
Stolarsky,
and
D.
Mirelman.
1990.
Specific
labeling
of
cysteine
proteinases
in
pathogenic
and
nonpathogenic
Entamoeba
histolytica.
Infect.
Immun.
58:1396-1401.
40.
Orozco,
E.,
M.
A.
Rodriguez,
and
F.
Hernandez.
1988.
The
role
of
phago-
cytosis
in
the
pathogenic
mechanism
of
Entamoeba
histolytica.
In
Amebiasis:
Human
Infection
by
Entamoeba
histolytica.
J.
I.
Ravdin,
editor.
John
Wiley
and
Sons,
New
York.
326-338.
1540
Reed
et
al.