Content uploaded by Elzbieta Szczesna Skorupa
Author content
All content in this area was uploaded by Elzbieta Szczesna Skorupa on May 11, 2018
Content may be subject to copyright.
Eur.
J.
Biochem.
73,
515-520
(1977)
Membrane-Dependent Cleavage
of
the Human Placental Lactogen Precursor
to
Its Native Form
in
Ascites Cell-Free Extracts
Irving BOIME, Elzbieta SZCZESNA and Donna SMITH
Department
of
Obstetrics and Gynecology, and Department
of
Pharmacology,
Washington University School
of
Medicine, St Louis, Missouri
(Received October
5,
1976)
Messenger RNA derived from term placenta directs the synthesis of a precursor to human placental
lactogen (prelactogen, molecular weight
25
000)
in an ascites cell-free system containing ribosome-
free supernatant and preincubated purified ribosomes. The processing of prelactogen to native
lactogen (molecular weight 22
200)
was only observed when a microsomal membrane preparation
was added prior to the synthesis of complete protein,
i.e.
before release.
Placental mRNA directed the synthesis of prelactogen in a system containing free polysomes,
whereas in a comparable system containing membrane-bound polysomes both prelactogen and
lactogen were synthesized. The prelactogen synthesized in the latter system could be cleaved by the
addition of membranes at the start of incubation.
Preprotein cleavage activity was inhibited 100
%
by
0.04
%
Triton X-100, while protein synthesis
was inhibited only about 30
%.
Using Triton to block cleavage specifically at intervals after mRNA
and membrane additions, it was determined that the overall cleavage reaction required about
15
min.
When the ascites system was incubated with charged initiator [35S]methionyl-tRNAyet prelactogen
was formed. The labeled prelactogen was processed when membranes were added during the first
few minutes of incubation, but
no
processing occurred when membranes were added after 60 min
of
incubation.
The results indicate that prelactogen is the primary gene product, and cleavage activity is appar-
ently associated only with the membrane-bound ribosomal fraction of the cell.
It is now apparent that in cell-free systems secretory
proteins are synthesized as larger forms. For example,
mRNAs encoding for immunoglobulin light chain
[
1
-
51, parathyroid hormone [6], human placental
lactogen (lactogen [7- 9]), growth hormone [lo],
proinsulin
[ll
-
131 and prolactin
[14-
161 direct the
synthesis in various cell-free systems of heavier forms
of these proteins which contain authentic sequences
found in their corresponding finished products. It
seems that these proteins are in fact physiological
precursors based on the following observations
:
(a) a precursor
-
product relationship has been dem-
onstrated
in
vitro
[17,18]; (b) the precursor forms have
been detected in polysome 'run-off systems [2,17,18],
where the initiation step in the mRNA-dependent
system is avoided; (c) the extra protein portion,
located at the N-terminal region of these proteins,
contains similar hydrophobic amino-acid sequences
[6,13,19-211; and (d) the light-chain and growth
Abbreviations.
Human placental lactogen, lactogen; precursor
to placental lactogen, prelactogen.
hormone precursor forms have been observed when
tissue slices are incubated with various protease
inhibitors [10,22].
This last point had been puzzling; except with the
use of protease inhibitors, they have not been observed
in vivo.
However, experiments
in
vitro
have shown
that cleavage occurs during synthesis and that once
the protein is released from the ribosome it is insensi-
tive to cleavage [17,18,26]. Thus, the appearance of
intact precursors either in the tissue or in the blood
would not be expected.
Because of its transient existence and hydrophobic
nature the extra amino acid sequence in these precur-
sors has been a candidate to mediate binding to the
endoplasmic reticulum, which would account for the
formation of membrane-bound polysomal structures
that are apparently specific for the synthesis of secre-
tory proteins [2,17,23,24]. In addition myeloma
microsomes and polysomes derived from them syn-
thesized native light-chain and pre-light-chain im-
munoglobulin respectively, in a run-off cell-free system
[2,17]. Thus one would expect that membrane-bound
516
Preprotein Cleavage in
Ascites
Tumor Cell-Free Extracts
polysomes should contain essentially all of the pre-
protein cleavage enzyme activity of the cell.
An important aspect of this cleavage process is
the kinetics of the reaction and its relationship to the
growing peptide chain. For example, when the nascent
chain is of the appropriate length is there a time lag
before scission of the precursor occurs? Also, what is
the relationship between the size of the nascent chain
and the extent of cleavage? Furthermore, if these
precursors are truly initial products of translation and
not derived from a still larger protein, they should
represent the primary gene product. Here we show for
the case of lactogen mRNA that it directs the syn-
thesis of prelactogen in an ascites tumor cell-free
system composed of free polysomes, whereas, in a
comparable system containing membrane-bound poly-
somes, cleaved lactogen was observed. Further data are
presented showing kinetics of this cleavage and that
prelactogen is the primary gene product.
EXPERIMENTAL PROCEDURE
[35S]Methionine (specific activity
300
Ci/mmol)
was obtained from Amersham/Searle. Human placen-
tal lactogen (95% pure) was purchased from Nutri-
tional Biochemicals. Sucrose was purchased from
Schwarz/Mann. Purified [35S]methionyl-tRNA!'et was
a gift from Dr Dolph Hatfield at the National
Institutes of Health.
Isolation of Cell-Free Extracts
Ribosomes and ribosome-free supernate were
prepared from Krebs ascites tumor cells by placing a
preincubated
30000xg
supernate on a layer of
1
.O
M
sucrose containing buffer A
(30
mM Tris-HC1,
pH 7.5, 120
mM
KCl, 5 mM magnesium acetate,
7
mM 2-mercaptoethanol) and centrifuged for 5 h
at
50000
rev./min in a Spinco
60
Ti rotor [18]. The
non-membrane-bound and membrane-attached poly-
somes were prepared from ascites tumor cells accord-
ing to Swan
et al.
[l].
Term placental mRNA was
isolated as previously described
[8].
The ascites membrane fraction was prepared from
preincubated extracts by collecting the material that
accumulated at the ribosome-free supernate/l
.O
M
sucrose interphase, diluting it with
5
volumes of
buffer
A
and centrifuging at 200
000
x
g
for 90 min.
The pellet was resuspended to give a concentration
of 20
-
40
A260
units/ml.
Protein Synthesis Assay
Protein synthesis was assayed in 0.18 ml reaction
mixture containing
30
mM Tris-HC1 (pH 7.5), 90 mM
KCl, 2.6 mM magnesium acetate, 2.5 mM dithiothrei-
tol,
0.3
mM GTP,
0.6
mM CTP, 1.5 mM ATP, 1
.O
mM
creatine phosphate, 0.12 mg/ml creatine kinase, 40 pM
each of non-radioactive amino acids minus methio-
nine, 1.4 pM [35S]methionine, and term placental
RNA. The amounts of extracts are noted in the
experiments. Incubation was at
30
"C for 90 min
and samples were subsequently treated as previously
described
[8].
The products synthesized
in vitro
were examined
on 20 polyacrylamide slab gels
[8].
RESULTS
Membrane-dependent cleavage of pre-immuno-
globulin light chain [26,17] and prelactogen
[18]
to
their corresponding native circulating forms
in
vitro
occurs on the growing nascent chains, but upon
release of the proteins they are insensitive to cleavage.
One question that arises concerns the kinetics of the
cleavage reaction. How much of prelactogen has to
be synthesized before it is cleaved? Also, does cleavage
occur immediately or is there a lag time before the
finished product is observed? In terms of the synthesis
of prelactogen it takes about 15 min to synthesize
the complete protein (Fig.
1
C). Since cleavage is an
early event during synthesis, the relative protein
length required for cleavage might be determined if the
membrane fraction is added to the ascites cell-free
system at various times after initiation. Reaction mix-
tures containing lactogen mRNA, ribosome-free
su-
pernate, and preincubated ribosomes were incubated
for 5,
10,
and 15 min and then membranes and
1
pM
pactamycin were added together. (Pactamycin was
added to prevent the appearance of labeled lactogen
synthesized by contaminating ribosomes associated
with the membrane fraction. Pactamycin at this
concentration blocks placental mRNA-dependent ini-
tiation in the ascites system without affecting the
elongation rate [18].) The incubation was then contin-
ued for a total of 90 rnin and samples were taken for
gel analysis (Fig. 1). The addition of membranes
after 5 min results in the synthesis of primarily
lactogen; very little prelactogen was synthesized.
After 10 and 15 rnin of initial incubation there was a
progressive increase in the amount of prelactogen
synthesized despite the addition of the membrane
fraction. In contrast when pactamycin alone was
added, only prelactogen was synthesized. Thus, even
though the membrane fraction was added prior to
the synthesis of the entire protein, complete cleavage
was not observed
;
apparently when prelactogen
reaches a certain chain length it is less sensitive to
cleavage.
To examine the kinetics of cleavage
it
is necessary
to
stop the cleavage reaction after various periods
of
incubation. It was observed that 0.04% Triton com-
517
I.
Boime,
E.
Szczesna and
D.
Smith
A
8
C
preHPL
-
HPL
-
0
5
10
15
5
10
15
5
12
15
\
J
*
--Y-
-
+pacta
+
rnernb
+
pacta Total incubation time (rnin)
Total incubation
90
rnin
Fig.
1.
Sodium doderylsulphute gel electrophoresis
of
proteins synthesized
in
response to plucental
RNA.
Following incubation
at
the
indicated times, additions were made
of
1.0
pM
pactamycin (pacta) and membranes (memb) (A)
or
pactamycin alone
(B).
After additions,
incubation was continued for a total
of 90
min. In experiment
(C)
aliquots from scaled-up reaction mixtures containing RNA and
no
membranes were taken at the indicated times and processed for gel analysis. The radioactivity (counts/min) added was the following:
(A)
0
min, 25000; 5 min, 40000; 10 min, 50000; 15 min, 70000; and
(B)
5 min, 35000;
10
min, 70000; 15 min, 80000; and
(C)
5 min,
15000;
12
min, 45000; 15 min,
60000.
preHPL, prelactogen; HPL, lactogen
4
25000
-
22200
-
-
Tri
+
Tri
+
Tri
+
Tri
+
Tri
0
2
5
10
Incubation time (rnin)
Fig.
2.
Time course
of
proteins synthesized
in
the presence
of
0.04
%
Triton.
A
scaled-up reaction mixture containing ribosome-free supernate,
ribosomes, and mRNA was incubated
for
5
min; pactamycin and membranes were then added and aliquots were taken at the indicated
times and treated with Triton (Tri). The Triton-treated samples were then incubated
for
a total of
90
min. The lanes contained about
60000
counts/min, except
0
min
+
Tri, which contained about 40000 counts/min
pletely eliminated cleavage activity (Fig.
2),
while
reducing mRNA-dependent protein synthesis only
about
30%.
For studying the kinetics
of
the cleavage
reaction a scaled-up reaction mix containing pre-
incubated ribosomes and mRNA was incubated for
5
min. After this initial incubation, pactamycin and
membranes were added and aliquots were taken
0,
2,
5,
and 10 min after the addition, and placed in tubes
containing Triton. The reactions were then incubated
for a total of
90
min (Fig.2). It is apparent that the
cleavage is essentially complete
15
rnin following
the beginning of the mRNA-dependent reaction,
i.r.
at about the time of release
of
the protein.
The previous experiments were performed with
ribosomes that were obtained from preincubated
extracts. During preincubation the levels of endoge-
nous mRNA available for reinitiation are apparently
reduced
;
and thus ribosomes would be released from
free polysomes as well as membrane-bound poly-
soma1 complexes. Therefore, to examine the point
that the cleavage activity may be associated only with
membrane-bound polysomal complexes, term pla-
518
Preprotein Cleavage in Ascites Tumor Cell-Free Extracts
-PreHPL
-HPL
free Free
MB
MB
MB
-
RNA
-
RNA
+Tri
Fig.
3.
Sodium dodecylsulphate gel electrophoresis ofproteins synthe-
sized in ascites extracts containing membrane-bound
(MB)
polysomes
or
free
polysomes
in
the presence
of
term placental mRNA.
Where
indicated, 24
pg
free polysomes and
20
pg
membrane-bound poly-
soma were used, Approximately
65000
counts/min (free poly-
somes) and
15
000
counts/min (membrane-bound polysomes) syn-
thesized in the absence of RNA were applied. About
140000
counts/
min (free polysomes) and
55
000
counts/min (membrane-bound
polysomes) synthesized in the presence of RNA were applied to the
gel. Where indicated
0.04
Triton was added to a reaction mixture
containing membrane-bound polysomes.
PreHPL,
prelactogen;
HPL,
lactogen
1
2
3
4
-
preHPL
-
HPL
(25
000)
(22
200)
+
RNA +RNA -RNA
-Mernb +Mernb
+
Mernb
Fig. 4.
Sodium dodecylsulphate gel electrophoresis
of
placental-
mRNA-dependent proteins synthesized
in
the presence ofmemhrane-
bound polysomes.
Where indicated, 25
pg
membranes was added.
Approximately
40000
counts/min
(-
RNA), 140000 counts/min
(f
RNA,
-
Memb), and
200000
counts/min
(+
RNA,
+
Memb)
were added to the gel
5
6
-
preHPL
-HPL
+rn
+m
-rn
+rn
+rn
[35S]Met
Incubation
time
(min)
-
RNA
0
5
60
Fig.
5.
Protein synthesis in response toplacental RNA
in
thepresence
of
[35Sjmethionyl-tRNA~".
At the time points indicated, membranes (m)
were added.
Total
incubation time was
90
min, except for the reaction mixture receiving membranes after
60
min; this reaction was
incubated an additional
60
min after membrane addition. [35S]Met refers to proteins synthesized in ascites extract with free [35S]methionine;
no
[35S]methionyl-tRNA was added. The amount
of
radioactivity (counts/min) added to each lane was the following:
-
RNA,
+
m
0
min,
+
m
5
min,
5000;
-
m,
+
m
60
min,
12000;
[3SS]Met,
60000
I.
Boime,
E.
Szczesna and
D.
Smith
519
cental RNA was translated in cell-free systems com-
posed of ascites ribosome-free supernatant and either
free polysomes or membrane-bound polysomes pre-
pared from non-preincubated ascites extracts (Fig.
3).
In the presence of free polysomes, term placental
RNA directs the synthesis of only prelactogen whereas
the synthesis of both lactogen and prelactogen is
observed in the system containing membrane-bound
polysomes. The addition of 0.04
%
Triton to the reac-
tion mix containing membrane-bound polysomes
inhibited the cleavage of prelactogen. Furthermore, the
presence of membrane-bound polysomes was con-
firmed by the appearance of polysomes on a 15
-
40
%
sucrose gradient following deoxycholate treatment
of
the membrane-bound fraction. That the fraction of
prelactogen synthesized by membrane-bound poly-
somes was not insensitive to cleavage was examined by
the addition of ascites membranes prepared from
preincubated extracts to the cell-free system (Fig.
4).
There is a marked conversion of prelactogen to lacto-
gen in the presence of the added membrane fraction.
Therefore, essentially all
of
the preprotein cleavage
enzyme activity is associated with the membrane-
bound ribosome fraction.
An important point that follows from the process-
ing data is whether prelactogen is the primary product
of translation or if it arises from a cleavage of a larger
protein. If prelactogen is the primary gene product,
then it should initiate with methionine. The predic-
tion can be tested by labeling the proteins with the
initiator tRNA charged with [35S]methionine. Ac-
cordingly placental mRNA was incubated in the
ribosome-free supernatant/ribosome system contain-
ing purified charged [35S]methionyl-tRNAEE" obtained
from rabbit reticulocytes, and
10
mM unlabeled
methionine (Fig. 5). Non-radioactive methionine was
added to dilute out any discharged [35S]methionine
released during incubation and thus avoiding the
insertion of [3sS]methionine into the internal posi-
tions of the proteins. It
is
clear that an mRNA-depend-
ent protein was synthesized which comigrated with
prelactogen synthesized from free [35S]methionine. In
addition, as predicted from the previous experiments,
this labeled protein was not observed when mem-
branes were added at zero time or after
5
min of
incubation but remained when membranes were
added following 60 min of incubation. These data
show that prelactogen is the primary gene product and
that preprotein cleavage activity is exerted at the amino
terminal region of the protein during its synthesis
on
the ribosome.
DISCUSSION
The data presented support for lactogen, and
by
inference for othdr secreted proteins, the hypothesis
that the precursor forms of secretory proteins syn-
thesized in various cell-free systems represent physio-
logical intermediates to the corresponding extracel-
lularly released mature forms. First, it is clear that
prelactogen is processed during synthesis by a mem-
brane fraction that probably included membranes
derived from the endoplasmic reticulum. If, as sug-
gested [2,17], the extra amino acid sequence represents
an evanascent intracellular signal for the formation of
membrane-bound ribosomal complex specific for
secretory proteins, then, as observed, polysomes
bound to membranes should contain preprotein
cleavage enzyme activity. Free polysomes by contrast
are devoid of this cleavage activity. The data are thus
consistent with the notion that these precursors are
involved in the formation of membrane-dependent
secretory complexes.
It also follows from the model that the membrane
contains receptor sites for binding to polysomes
attached to nascent secretory proteins. Lactogen
mRNA directed the synthesis of both prelactogen and
lactogen in the cell-free system containing non-
preincubated membrane-bound polysomes. The ob-
servation that cleavage is not complete may indicate
that membranes attached to polysomes bearing endog-
enous ascites mRNA were not available
to
the nascent
lactogen chains undergoing synthesis on a population
of free ribosomes, or on ribosomes that were released
during incubation. Therefore, some nascent lactogen
chains might be too long before they would be acces-
sible
to
membranes and thus would be less sensitive
to
cleavage. Consistent with this possibility is the observa-
tion that membranes prepared from preincubated ex-
tracts and added to the system containing mRNA and
membrane-bound polysomes resulted in a marked con-
version of prelactogen to lactogen. Also, more con-
version is seen when preincubated membranes were
used than when membranes collected from non-
preincubated extracts were employed
(E.
Szczesna
and
I.
Boime, unpublished observation).
There are approximately 215 amino acids residues
in prelactogen and it takes 15 min to synthesize it.
Therefore, the polymerization rate in the ascites
extracts is about 14 amino acids/min. Thus after
5
min, the largest nascent chains would be 70 amino
acids long and they are cleaved effectively
in
vitro.
However, after 5 min the larger nascent chains are
apparently not cleaved. This may result from the time
required (10- 15 min) for the overall cleavage reac-
tion; thus the larger nascent chains may have been
released before scission of the protein could occur.
Alternatively, as previously suggested for the myeloma
system [17], the larger nascent chains may prevent
interaction of the nascent-chain . ribosomal complex
with the membrane, and thereby not be cleaved.
The lag period probably does not reflect preprotein
cleavage activity
per
se
but indicates a sequence of
events; the synthesis of the chain to the appropriate
520
1.
Boime,
E.
Szczesna and
D.
Smith: Preprotein Cleavage in Ascites Tumor Cell-Free Extracts
length, the binding
of
the polysome to the membrane
and subsequent interaction of the protein with the
membrane, and finally cleavage.
If the initiation codon is part of the 'signal' then
the precursor form should represent the primary gene
product. Evidence that this is the case for prelactogen
is the demonstration of a major protein synthesized
in the presence
of
[35S]methionyl-tRNA";1"'
that co-
migrated with prelactogen synthesized with free [35S]-
methionine. Further proof that the label was at the
amino terminus was the loss of the label when mem-
branes were added at zero time or after
5
min
of
incubation, time periods when prelactogen is cleaved.
However, little loss
of
label occurred when membranes
were added after
60
min
of
incubation when prelacto-
gen is known to be refractory to cleavage. The lack of
a band at
22000
(lactogen) shows that no detectable
incorporation of [35S]methionine into internal posi-
tions occurred. Using [35S]methionyl-tRNAyet, me-
thionine labelling of the amino terminus of the light-
chain precursor
[2]
and preproparathyroid hormone
[25] has also been observed. Therefore, prelactogen
and presumably similar 'pre' secretory proteins re-
present the primary gene products.
This work was supported by a grant from the Population
Council
of
the Rockefeller University (M 75.47) and by a grant
from the National Institutes of Health (AM-16865). The authors
are grateful for the excellent typing assistance
of
Betsy Duhn.
REFERENCES
1.
Swan, D., Aviv, H.
&
Leder,
P.
(1972)
Proc. Nail Acad. Sci
2. Milstein, C., Brownlee, G., Harrison, T.
&
Mathews,
M.
(1972)
U.S.A.
69, 1967-1971.
Nut. New Biol.
239. 117-120.
3. Mach, B., Faust, C.
&
Vassalli, P. (1973)
Proc. Nut1 Acad. Sci.
4. Schechter,
I.
(1973) Proc.
Natl Acad. Sci. U.S.A.
70, 2256-
5. Blobel, G.
&
Dobberstein, B. (1975)
J.
Cell Biol.
67, 852-862.
6. Kemper, B., Habener,
J.,
Ernst, M.
D.,
Potts,
J.
T.,
Jr
&
Rich,
7. Boime,
I.,
Boguslawski, S.
&
Caine,
J.
(1975)
Biochem. Bio-
8.
Boime,
I.,
McWilliams, D., Szczesna,
E.
&
Camel, M. (1976)
9. Cox,
G.,
Weintraub, B. D., Rosen,
S.
W.
&
Maxwell,
E.
(1976)
10.
Sussman,
P.
L., Tushinski, R. S.
&
Bancroft, F. C. (1976) Proc.
11.
Permutt, M. A.
&
Boime,
I.
(1975)
Diabetes,
24,
suppl.2, 405.
12. Lomedico,
P.
T. (1975)
Diabetes,
24,
suppl. 2,405.
13. Chan, S.
J.,
Keim, P. L.
&
Steiner, D.
F.
(1976)
Proc. Nut1
U.S.A.
70,451 -455.
2260.
A. (1976)
Biochemistry,
15,
15-19.
phys.
Res.
Commun.
62, 103- 109.
J.
Biol. Chem.
251, 820-825.
J.
Biol. Chem.
251, 1723-1730.
Nail Acad. Sci. U.S.A.
73, 29
-
33.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Acad. Sci. U.S.A.
73, 1964-1968.
2801
-
2807.
2842- 2847.
Res. Commun.
70, 1180-1189.
Mauer, R. A,, Stone,
R.
&
Gorski,
J.
(1976)
J.
Biol. Chem.
251,
Evans, G. A.
&
Rosenfeld, M. G. (1976)
J.
Biol. Chem.
251,
Dannies, P. S.
&
Tashjian, A.
H.,
Jr
(1976)
Biochenz. Biophys.
Blobel, G.
&
Dobberstein, B. (1975)
J.
Cell
Biol.
67, 835-851.
Szczesna,
E.
&
Boime,
I.
(1976)
Proc. Nail
Acad.
Sci. U.S.A.
73,
Schechter,
I.
&
Burstein, V. (1976)
Biochem. Biophys. Rrs.
Devillers-Thiery, A., Kindt, T., Scheele, G.
&
Blobel,
G.
(1975)
Birken,
S.,
Smith, D., Canfield, R.
&
Boime,
I.
(1977)
Biochem.
Schmeckpepper, B.
J.,
Adams,
J.
M.
&
Harris, A. W. (1975)
Harrison,
T.
M., Brownlee,
G.
G.
&
Milstein,
C.
(1974)
Eur.
J.
Mechler, B., Vassalli, P. (1975)
J.
Cell Biol.
67, 25-37.
Kemper, B., Habener,
J.
F.,
Potts,
J.
&
Rich, A. (1976)
Bio-
Harrison,T. M. (1973) Ph. D. Thesis, University of Cambridge.
1179- 1183.
Commun.
68,489
-
496.
Proc. Nail Acad. Sci. U.S.A.
72, 5016-5020.
Biophys. Res. Commun.
74,
106-112.
FEBS Lett.
53, 95-98.
Biochem.
47,613
-
620.
chemistry,
15, 20- 25.
I.
Boime and
D.
Smith,
Department
of
Obstetrics and Gynecology, Washington University School of Medicine,
491
1
Barnes Hospital Plaza, Saint Louis, Missouri, U.S.A. 63110
E. Szczesna,
Instytut Biochemii
i
Biofizyki, P.A.N.,
PL-02-532 Warszawa,
u.
Rakowiecka 36, Poland
Note added in
Proof(February 9, 1977). Amino-acid-sequence
analyses
of
the amino-terminal region of prelactogen and lactogen
have shown that the ascites membrane fraction cleaves prelactogen
to native lactogen [21]; no intermediate precursor containing an
extra amino-terminal hexapeptide, such as that seen for the pro-
parathyroid hormone [6] was formed. Since it appears that there
is
no similar pro-form for the immunoglobulin light chain [19] the
data suggest that only a certain class of secretory proteins are
processed through a pro-intermediate form.
