Glycobiology vol. 7 no. 4 pp. 565-570, 1997
Structural and biological roles of glycosylations in pulmonary angiotensin
I-converting enzyme
Bruno Baudin1'2'3, Nathalie Alves1, Antoine Pilon1,
B£n6dicte Be'ne'teau-Burnat1'2 and Jacqueline Giboudeau1
'Laboratoire de Biochimie A, HOpital Saint-Antoine, 184, Rue du Fbg
Saint-Antoine, 75571 Paris Cede* 12, France and 2Laboratoire de Biochimie
Generale et de Glycobiologie, UFR Pharmacie, University Rent Descartes
(Paris V), 4 Avenue de l'Observatoire, 75006 Paris, France
'To whom correspondence should be addressed at: Laboratoire de
Biochimie A, H6pital Saint-Antoine, 184, Rue du Fbg Saint-Antoine, 75571
Paris Cedex 12, France
We enzymatically deglycosylated pig lung angiotensin
I-converting enzyme (ACE) to study the involvement of its
glycanic chains in its physicochemical and catalytic properties. The effects of endoglycosidases F 2 and H, and of
N-glycanase were assessed by ACE mobility in SDS-PAGE.
N-Glycanase only was completely effective with or without
previous denaturation, leading to a shift in ACE M r from
172 to 135 kDa; endoglycosidase F 2 produced the same shift
but only without previous denaturation. Deglycosylated
ACE had the same kcat as native ACE for the substrate
hippuryl-histidyl-Ieucine, and an identical Stokes radius as
measured by size-exclusion high performance liquid chromatography. Neuraminidase had no effect on ACE Stokes
radius but slightly decreased its kcat which could be related
to variations in ionization of the active site. The isoelectric
point of ACE, as, determined by isoelectric focusing, increased from 4.5-4.8 to 5.0-53 after either endoglycosidase
F3 or neuraminidase digestion, but still with microheterogeneities which thus did not seem to be related to ACE
glycans. Deglycosylated ACE did not bind onto agaroselectins in contrast to native ACE which bound strongly to
concanavalin A showing interactions involving oligomannosidic or biantennary and sialylated N-acetyl-lactosaminic
isoglycans. Finally, tunicamycin, an inhibitor of Nglycosylation, did not modify ACE secretion by endothelial
cells. Thus, ACE glycans have no drastic effects on structural and biological properties of the protein, but they may
have a functional role on intracellular targeting of both
secreted and membrane-bound ACE isoforms, also for the
protection of the soluble plasma form against hepatic
lectins and the maintenance of its hydrosolubility.
Key words: angiotensin I—converting enzyme (peptidyldipeptidase EC.3.4.15.1 )/endothelium/glycosidases/lectins
Introduction
Angiotensin I-converting enzyme (ACE) (or peptidyldipeptide
hydrolase EC.3.4.15.1) is a key enzyme in renin-angiotensin
and kallikrein-kinin systems, by removing the carboxyterminal dipeptides of angiotensin I and of bradykinin, thereby
activating the former into angiotensin II, a vasopressor, and
© Oxford University Press
degrading the latter, a vasodilator. ACE is a zinc-containing
glycoprotein which is expressed in various tissues under different molecular forms, but always with one polypeptide chain.
The somatic isoenzyme exhibits two molecular forms, a membrane-bound form and a soluble form. The membrane-bound
form is mainly located in vascular endothelial cells as an ectoenzyme (Ryan et al., 1975; Baudin et al, 1996), but also on the
brush-border of renal tubular and intestinal epithelial cells
(Bruneval et al., 1986) and in monocytes and macrophages
(Friedland et al, 1978). Testicular cells synthesize a germinative isoenzyme with a lower molecular mass, around 100 kDa
versus around 160 kDa for the somatic enzyme (Iwata et al.,
1982; Soubrier et al, 1988). The larger somatic isoenzyme has
two highly homologous domains, each bearing a catalytic site,
whereas the smaller isoenzyme from germinal cells has a single
domain and therefore only one active site and would represent
the ancestral, nonduplicated form of the ACE gene (Lattion et
al., 1989). A circulating soluble form of ACE is found in blood
plasma and is most likely released by the vascular endothelial
cells (Das et al., 1977; Beldent et al, 1993).
We have previously purified ACE from pig and human lung
and blood (Baudin et al, 1991), compared the physicochemical
properties and the chemical compositions of both the pulmonary and serum forms (B.Baudin, personal communication),
and demonstrated the elongated shape of the pig pulmonary
enzyme in aqueous solution (Baudin et al, 1988). In particular,
lung ACE appeared to be a glycoprotein containing about 10%
(w/w) of sugars, and only those which are characteristic of
asparagine N-glycosylation. The aim of the present study was
to investigate the involvement of glycosylation in the physicochemical, catalytic, and biosynthetic properties of the enzyme.
We carried out comparative analyses before and after digestion
with glycosidases.
Results and discussion
Effect of deglycosylation on ACE Mr and size
ACE was purified from pig lung to electrophoretic homogeneity as shown in Figure 1; typically, 2 mg was isolated from 300
g of pulmonary tissue. Several protocols for enzymatic deglycosylation of ACE were compared; N-glycosidases were chosen because pulmonary ACE contains N-asparagine bound glycans only (Hartley and Soffer, 1978; Baudin et al., 1988).
Endoglycosidase H was not effective for ACE deglycosylation under the conditions used (not shown). It was not possible
to decrease the pH below 5 because ACE precipitated around
its isoelectric point (B.Baudin, personal communication). In
contrast, both endoglycosidase F 2 , at pH 5.6, and N-glycanase,
at pH 8.0, were effective, decreasing ACE mobility in SDSPAGE from 172 ± 4 kDa to 135 ± 4 kDa. Curiously, previous
denaturation in SDS did not allow the deglycosylation by endoglycosidase F 2 whereas N-glycanase was effective with or
565
B.Baudin et al
Mr
X103
ACE
-*
45
8
Fig. 1. SDS-PAGE of pig lung ACE before (2, 4) and after digestion with either neuraminidase (1), N-glycanase (5), endoglycosidase F 2 (6), or
endoglycosidase H (7), without previous SDS denaturation. High molecular weight standards (3, 8).
without previous denaturation of ACE (Figure 1). The lower
Mj of deglycosylated ACE is near to that deduced from endothelial cDNA which does not take account of glycan mass
(Soubrier et al, 1988). Neuraminidase treatment had no effect
on the molecular mass of ACE (Figure 1). It is known that
endoglycosidase H hydrolyses more particularly oligomannosidic glycans, that the endoglycosidases F have a greater specificity for N-acetyl-lactosaminic glycans, and that N-glycanase
is not specific for a particular type of glycans; thus, our data
with these glycosidases show that the glycans of pulmonary
ACE are not of one or the other type but rather a mixture of
both types or hybrid structures. Previous data with the establishment of sugar molar ratio already predicted a mixture of
complex and lactosaminic glycans, some of them being sialylated or fucosylated (Baudin et al, 1988). Moreover, the
fractionation of ACE glycans has shown at least two types of
oligosaccharides, one being rich in mannose and with one fucose residue, the another being rich in sialic acid (unpublished
data). In the former study, we also demonstrated that the ACE
molecule is clearly elongated in aqueous solution. It could be
then hypothesized that deglycosylation would modify the ACE
conformation leading to a decrease in its size. Therefore, we
compared the ACE Stokes radius (Rs) before and after endoglycosidase F 2 digestion. By plotting the Rs of standard globu566
lar proteins versus (-log Kav)1/2, Kav being the constant of
diffusion available on Superose 12 HPLC, ACE Rs was found
to be 6.0 ±0.2 run for the native enzyme as well as for deglycosylated ACE which does not support the postulated hypothesis (Figure 2). Nevertheless, the glycans could be important
for the hydrosolubility of ACE and for its stability in aqueous
solution.
Effect of deglycosylation on IEF pattern of ACE
As judged on IEF, with Coomassie blue staining or specific
zymography, ACE pi was 4.5-4.8 with 4-5 bands focused in
the enzymatically active area. The deglycosylation of ACE by
endoglycosidase F 2 or its desialylation by neuraminidase led to
a shift of ACE pi from 4.5-^.8 to 5.0-5.3, both still with
microheterogeneity as judged on protein staining and the same
displacement on zymography (Figure 3). These results obviously show that the acidic pi of ACE is partly related to its
sialic acid content, the residual acidity being due to dicarboxylic amino acids since they are preponderant in pig lung ACE
(Baudin et al, 1988) as well as in other species (Das et al,
1975; B6neteau-Burnat et al., 1994). Deglycosylation and desialylation increased the ACE pi almost identically, which indicates that, within ACE sugar moiety, only N-acetylneuraminic acid provides electric charges. Moreover, the
Glycosyiations of angjotensln-converting enzyme
maintenance of microheterogeneity after deglycosylation
shows that it is not related to heterogeneity in ACE glycans,
but to abnormalities in the primary sequence or to the existence
of modified amino acids.
Effect of deglycosylation on ACE affinity for lectins
Pulmonary ACE interacted strongly with concanavalin A as
shown by its complete binding onto ConA-Sepharose 4 B and
by the incomplete elution from the column using ot-methylmannoside alone. A more complete elution was achieved in 0.2
M a-methylmannoside plus 0.5 M NaCl (Figure 4). In contrast,
ACE binding on WGA-Ultrogel was incomplete and specific
elution in 0.2 M N-acetylglucosamine was easy. Endoglycosidase F2-treated ACE did not bind onto either of the lectins,
demonstrating the role of ACE glycans in the interactions with
these lectins. These data are in favor of the predominance in
pulmonary ACE of oligomannosidic or hybrid structures.
Fig. 2. Plot of Stokes radius in run (Y) versus (-log Kav)'« (X), Kav being
measured on Superose 12 HPLC. •, Rs standards; O, native ACE; A,
endoglycosidase F2-treated ACE.
.B-
-A.
2.8
•
3.75
•
4.15
•
4.55
•
5.2
•
Effect of deglycosylation on ACE catalytic activity
ACE activity was radiometrically determined on synthetic substrate; neither endoglycosidase F 2 - nor N-glycanase-treatment
significantly modified ACE kcat in comparison with native
pH
in situ
3L
5.5
5.85
•
6.55
»*
6
8
9
10
Fig. 3. IEF of pig lung ACE before (2, 8) and after desialylation by neuraminidase (3, 4, 9) or after deglycosylation by endoglycosidase F2 (5, 6, 10). pi
standards (1, 7). 2, 4, 5: 20 p.g of ACE; 3, 6, 8, 9, 10: 10 (ig of ACE. Coomassie blue staining (A) and specific ACE zymography (B).
367
B.Baudin et al
WGA
CON-A
0.05 _
£
£
0
£
CM
b
I
o
aci
280
z
£
6
•
ro
z
1
AI
£
M
1
0.01 _
40
10
I
10
30
ml
0.0 5
I
B
I
0.01
10
40
30
ml
Fig. 4. Chromatography on agarose-bound lectins of pig lung ACE before (A) and afteT endogtycosidase F2-treatment (B). OLMM, a-methylmannoside; GN,
N-acetylglucosamine. All protein peaks contained ACE activity.
568
GlycosyUtions of angiotensin-converting enzyme
ACE incubated in the respective buffers but without the glycosidases (Table I). These results correlate with those of IEF
where ACE zymography was possible even after the deglycosylations, but not after N-glycanase treatment after denaturation of ACE which completely abolished its activity. Nevertheless, desialylation decreased ACE kcat by 30% without
clear explanation. Thus, ACE glycans do not seem indispensable for its enzymatic activity, although sialic acid could have
a significant importance on the catalytic events in ACE active
site.
Table IL Effect of tunicamycin on ACE biosynthesis and secretion by
porcine pulmonary artery endothelial cells: incubation with (+) or without
(-) 10 jig/ml tunicamycin
Culture
supernatant
(U/ml)
Cells
(U/mg proteins)
24 h
0.523 ± 0.09*
2.197 ±0.48
NS b
NS
2.129 ±0.32
0-506 ±0.10
48 h
Effect of tunicamycin on ACE biosynthesis and secretion by
endothelial cells
Endothelial cells were grown in vitro from the pig pulmonary
artery; 48 h after confluency, i.e., when the expression of ACE
is maximal (Baudin et al., 1996), primary cultures were incubated for 24 or 48 h with 10 u-g/ml tunicamycin, an inhibitor of
cellular N-glycosylation. This treatment did not modify ACE
secretion in the culture medium or its cellular biosynthesis
since ACE activity in both media did not vary (Table II). A
higher tunicamycin concentration was toxic for the cells, therefore we could not be sure that tunicamycin completely blocked
N-glycosylation. However, glycosylation of endothelial ACE
does not seem important for its secretion by the endothelium.
It is noteworthy that glycans are indispensable for testicular
ACE since the blockage of glycosylation results in rapid intracellular turnover of the underglycosylated protein (Kasturi et
al., 1994). On the other hand, intact glycosylated chains are
necessary for thrombospondin secretion by human endothelial
cells (Vischer et al., 1985). Nevertheless, ACE glycans could
be important in its intracellular targeting, in particular for the
expression of the membrane-bound form and the secretion of
the soluble form, maybe also for the protection of the soluble
form against hepatic clearance, since ACE contains both galactose and sialic acid residues; some of them could be exposed
so as to recognize endogeneous lectins.
Materials and methods
Reagents
ACE was purified to electropboretic homogeneity from pig lung using a protocol that we have previously detailed (Baudin et aL, 1991). Endoglycosidase
F 2 (endo-P-N-acetylglucosaminidase F 2 , EC.3.2.1.96) from Flavobacterium
meningosepticum, and recombinant N-glycanase (peptide: N-glycosidase F,
EC.3.5.1.52) were from Genzyme (Cambridge, MA). Neuraminidase (acetylneuraminylhydrolase, EC.3.2.1.18) from Clostridium perfringens was provided by Sigma (St. Louis, MO). Recombinant endoglycosidase H
(EC.3.2.1.96) was from Oxford (Abingdon, UK).
N-Acetylglucosamine, a-methylmannoside, Nonidet P-40, Triton X-100,
phenylmethylsulfonyl fluoride (PMSF), N-ethylmaleinimide (NEM), ophthaldialdehyde, and tunicamycin were from Sigma. Dithiothreitol (DTT) and
sodium dodecyl sulfate (SDS) were from Bio-Rad (CA). All other reagents of
analytical grade were provided by Merck (Darmstadt, Germany) or Prolabo
(Paris, France).
0.510 ±0.05
1.09 ±0.26
NS
0.526 ± 0.08
NS
1.25 ±0.26
•Mean ± SD (n = 5).
"T^S, nonsignificant (p > 0.05), Mann-Whitney U test
Gtycosidases digestion
Endoglycosidase F 2 was used at 5 mU per mg of purified ACE in 0.3 M Na
citrate/phosphate, pH 5.6 buffer with 1% Nonidet P^M), 1 mM PMSF and at
37°C. Endoglycosidase H was used in the same conditions but with 0.9 M
Na-citrate/phosphate, pH 5.2 buffer. N-Glycanase was used at 37°C and at 17
U per mg of ACE in 0.05 M Tris, 0.1 M NaCl, pH 8 buffer plus 1% Nonidet
P^40, 1 mM PMSF, and 2 mM NEM. All these enzymes were also tested after
denaturation of ACE (5 min at 100°C) in 0.5% SDS and 50 mM DTT. Neuraminidase digestion was carried out at 37°C, at 10 U per mg of ACE, in 0.05
M Na-citrate/phosphate, 20 mM CaCl2, 1 mM PMSF, 2 mM NEM, pH 5.6
buffer. After enzymatic hydrolysis, the solutions were dialyzcd in Microdialyzer System 500 from Pierce (Rockford, IL) (cut-off 1000 Da) against 0.05 M
Tris, 0.1 M NaCl, pH 8 buffer.
Electrophoresis
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was
performed on 6.5% gels using Bio-Rad materials, dithiothreitol as a reducing
reagent and followed by Coomassie blue R-250 staining. Gels were calibrated
for molecular mass (MJ determination using SDS-PAGE standards "High
Molecular Weight" from Bio-Rad. Isoelectric focusing was carried out on
LKB system (Pharmacia, Uppsala, Sweden) in the pH range 4.0-6.5 (2.2%
Ampholine); ACE was revealed by both Coomassie blue staining and a specific fluorimetric zymogram which we have developed from the method of
Friedland and Silverstein (1976) using hippuryl-histidyl-leucine as the substrate (Bachem, Bubendorf, Switzerland) and o-phthaldialdehyde to promote
the fluorescence of histidyl-leucine (Baudin et aL, 1986). Gels were calibrated
for isoelectric point (pi) determination using the low pi isoelectric focusing
calibration kit from Pharmacia; the pH was also measured in situ using a
Microprocessor Ionalyser model 901 from Orion Research (Mettler, France).
Ckromatography
Size-exclusion HPLC was performed on Superose 12 HR 10/30 column, from
Pharmacia, calibrated with gel-filtration standards, from Bio-Rad, and using
Gilson material as previously described (Baudin et aL, 1991). Lectin-affinity
chromatography was carried out in 2 ml columns fitted with either Concanavalin A-Sepharose 4 B (Pharmacia) or wheat germ agglutinin-Ultrogel
(Biosepra, ViUeneuve-la-Garenne, France), at 15 ml/h and in 0.1 M sodium
phosphate, 0.1 M NaCl, 1 mM CaCl2, 1 mM MgCl2, 1 mM MnCl2, pH 7.4
buffer.
Culture of endothelial celts
Table I. Effect of glycosidase treatment on ACE catalytic activity as judged
on the hydrolysis of hippuryl-histidyl-leucine
Control*
Glycosidase
N-Glycanase
EndoF2
Neuraminidase
19.4
9.8
9.1
9.8
6.8
•Native ACE dissolved in the respective buffers, then dialyzed like for the
aliquots treated by the glycosidase.
b
In units per mg of protein (kcat), mean of three determinations.
Endothelial cells were cultured from pig pulmonary arteries according to a
method which we have previously detailed (Baudin et aL, 1996). Primary
cultures only were used. After reaching perfect confluency, the cells were
rinsed and incubated with 1 ml of serum-free medium containing 0 to 10 M-g/ml
tunicamycin, for 24 or 48 h, at 37°C in 5% CO 2 , 95% air atmosphere. Culture
supernatants were collected at the end of the incubation; they contained ACE
secreted by the cell monolayers in the time interval. Then the monolayers were
washed twice with serum-free medium and were lysed for 15 min at 37°C with
1% Triton X-100 in phosphate-buffered saline containing 2 mM PMSF and 5
mMNEM.
569
B.Baudin et aL
ACE assay
l4
ACE activity was radiometrically determined using ( C-glycine)hippurylhistidyl-leucine (New England Nuclear, Boston, MA) in an assay that we
previously developed (Baudin et aL, 1990). Enzyme activity is expressed as
micromoles of substrate hydrolyzed per minute per milligram of protein, ACE
protein being measured using e ^ ^ , = 1.5 x 105 M~' cm"1 calculated on
UV-spectrum (Baudin et aL, 1995), and total proteins in cell lysates using
Markwell's method (1978) based on Lowry's reaction but avoiding the interference due to the detergent which is still in the cell extracts.
Acknowledgments
We thank Ms. Pascale Jue for the perfect typing of the manuscript and Paul
Fradet for taking photographs of the gels.
Abbreviations
ACE, angiotensin I-converting enzyme (peptidyldipeptidase EC.3.4. 15.1);
SDS—PAGE, sodium dodecyl sulfate—polyacrylamide gel electrophoresis;
HPLC, high performance liquid chromatography; IEF, isoelectric focusing; pi,
isoelectnc point; Con A, concanavaUn A; WGA, wheat germ agglutinin.
References
Baudin.B., Beneteau-Bumat3., BaumaruvF.C. and GiboudeauJ. (1990) A reliable radiometric assay for the determination of angiotensin I-converting
enzyme activity in urines. J. Clin. Chem. Clin. Biochem., 28, 857—861.
BaudinJB., DrouetJL, Legrand,Y. and BaumannJ'.C. (1986) Application of
IEF to follow the purification of angiotensin-converting enzyme. In Peteers,H., (ed.), Protides of the Biological Fluids, Vol. 24. Pergamon Press,
Oxford, pp. 889-893.
Baudin3., Timmins,P.A., DrouetJ.., Legrand.Y. and BaumannJ'.C. (1988)
Molecular weight and shape of angiotensin I-converting enzyme. A neutron
scattering study. Biochem. Biophys. Res. Commun., 154, 1144-1150.
BaudinJ3., Tahraoui^A., BaumannJ\C, RobicJX, DrouetJ-. and Legrand.Y.
(1991) Purification and analysis of lung and plasma angiotensin I-converting enzyme by high performance liquid chromatography. Protein Expr. Purif., 2, 412-419.
Baudin,B-, Mario,N., Gaba,S. and GiboudeauJ. (1995) Spectroscopic studies
on angiotensin I-converting enzyme. Enzyme Protein, 48, 265-274.
Baudin^., Berard,M., CarrierJ.L., Legrand,Y. and DrouetJL. (1996) Vascular
origin determines angiotensin I-converting enzyme expression in endothelial cells. Endothelium, in press.
Beldent,V., MichaudA-, WeiJL, Chauvet,M.T. and CorvolJ'. (1993) Proteolytic release of human angiotensin-converting enzyme. LocaUzation of the
cleavage site. J. BioL Chem., 268, 26428-26434.
B6n6teau-Bumat3i Tahraoui.A., Barbut.F., GiboudeauJ. and Baudin,B(1994) Physicochemical and immunological comparisons between angiotensin I—converting enzymes purified from different mammalian species.
Comp. Biochem. PhysioL, 109B, 623-635.
BrunevalJ1., Hinglaisj^., Alhenc-GelasJ'., Tricottet.C, Corvol^1., MenardJ.,
CamilleriJ.P. and BarietyJ. (1986) Angiotensin I-converting enzyme in
human intestine and kidney. Ultrastructural immunohistochemical localization. Histochemistry, 85, 73-SO.
Das^L, HartleyJ.L. and Soffer^L. (1977) Serum angiotensin-converting enzyme. Isolation and relationship to the pulmonary enzyme. /. BioL Chem.,
252, 1316-1319.
Das,M. and Soffer,R.L. (1975) Pulmonary angiotensin-converting enzyme:
structural and catalytic properties. J. BioL Chem., 250, 6762-6768.
Friedland J . and Silverstein,E. (1976) A sensitive fluorimetric assay for serum
angiotensin-converting enzyme. Am. J. Clin. PathoL, 66, 116-124.
FriedlandJ., Setton,C. and SUverstein,E. (1978) Induction of angiotensinconverting enzyme in human monocytes in culture. Biochem. Biophys. Res.
Commun., 83, 843-849.
HartleyJ.L. and Soffer,R.L. (1978) On the oligosaccharide moiety of angiotensin-converting enzyme. Biochem. Biophys. Res. Commun., 83, 1545—
1552.
IwataJC., Lai.C, El-Dorry,H.A. and Soffer,R.L. (1982) The NH2 and COOHterminal sequences of the angaotensin-converting enzyme from rabbit lung
and testis. Biochem. Biophys. Res. Commun., 107, 1097-1103.
Kasturi.S., JabbarJ^LA., Sen.G.C. and SenJ. (1994) Role of glycosylation in
the biosynthesis and activity of rabbit testicuUr angiotensin-converting enzyme. Biochemistry, 33, 6228-6234.
Lartion^A.L., SoubrierJ3., AllegriniJ., Hubert.C, Corvol.P. and Alhenc-
570
Gelas,F. (1989) The testicular transcript of the angiotensin I-converting
enzyme encodes for the ancestral, non-duplicated form of the enzyme. FEBS
Lett. 252, 99-104.
Markwell,M.A., Haes^Jvl, BieberJLL. and Talbert,N.E. (1978) A modification of the Lowry procedure to simplify protein determination in membrane
and lipoprotein samples. AnaL Biochem., 87, 206-210.
RyaiU.W., Ryan,U.S., SchultzJJJL, Whitaker.C, Chung,A. and Dorer,F.E.
(1975) Subcellular localization of pulmonary angiotensin-converting enzyme (kininase II). Biochem. J., 146, 497-499.
Soubrier,F., Alhenc-GeUs,F., Hubert,C, AllegriniJ., JohruM., Tregear.G. and
CorvolJ1. (1988) Two putative active centers in human angiotensin Iconverting enzyme revealed by molecular cloning. Proc. NatL Acad. ScL
USA, 85, 9386-9390.
Vischer,P., Beeckji. and Voss3. (1985) Synthesis, intracellular processing
and secretion of thrombospondin in human endothelial cells. Eur. J. Biochem., 153, 435-443.
Received on September 18, 1996; revised on October 22, 1996; accepted on
October 22, 1996