Ultrastructural Immunolocalization of Lysyl Oxidase
in Vascular Connective Tissue
Herbert M. Kagan,* Charles A. Vaccaro,* Rebecca E. Bronson,*
Shiow-Shih Tang,* and Jerome S. Brody*
* Department of Biochemistry and ~ Pulmonary Center, Department of Medicine, Boston University School of Medicine,
Boston, Massachusetts 02118
Abstract. The localization of lysyl oxidase was exam-
ined in calf and rat aortic connective tissue at the ul-
trastructural level using polyclonal chicken anti-lysyl
oxidase and gold conjugated rabbit anti-chicken im-
munoglobulin G to identify immunoreactive sites.
Electron microscopy of calf aortic specimens revealed
discrete gold deposits at the interface between ex-
tracellular bundles of amorphous elastin and the
microfibrils circumferentially surrounding these bun-
dies. The antibody did not react with microfibrils
which were distant from the interface with elastin.
There was negligible deposition of gold within the
bundles of amorphous elastin and those few deposits
seen at these sites appeared to be associated with
strands of microfibrils. Lysyl oxidase was similarly
localized in newborn rat aorta at the interface between
microfibrils and nascent elastin fibers. Gold deposits
were not seen in association with extracellular colla-
gen fibers even after collagen-associated proteoglycans
had been degraded by chondroitinase ABC. However,
the antibody did recognize collagen-bound lysyl oxi-
dase in collagen fibers prepared from purified collagen
to which the enzyme had been added in vitro. No
reaction product was seen if the anti-lysyl oxidase was
preadsorbed with purified lysyl oxidase illustrating the
specificity of the antibody probe. The present results
are consistent with a model of elastogenesis predicting
the radial growth of the elastin fiber by the deposition
and crosslinking of tropoelastin units at the
YSYL oxidase plays a pivotal role in the biosynthesis of
collagen and elastin by oxidizing peptidyl lysine to
peptidyl a-aminoadipic-5-semialdehyde, the precur-
sor to the covalent crosslinkages in these connective tissue
proteins. The lysine-derived crosslinkages that evolve from
this aldehyde stabilize the fibrous structures of elastin and
collagen, thus providing anchoring points important to the
tensile and/or elastic properties exhibited by these proteins
It appears likely that lysyl oxidase functions in vivo in the
extracellular space and, indeed, the enzyme is secreted into
the growth medium by cultured fibroblasts (13, 24) and
smooth muscle cells (8). Little is known about the localiza-
tion of lysyl oxidase in either the intracellular or extraceUular
compartments, however. Siegel et al. (22) had demonstrated
at the level of the light microscope that the enzyme appears
to be associated with extracellular fibers of collagen in
fibrotic liver, using fluorescently labeled anti-lysyl oxidase
as a probe. Its localization has not been probed at the ultra-
structural level, however, nor has it been examined in an
elastin-rich tissue. In the present report, we describe the lo-
calization of lysyl oxidase in the extracellular matrices of bo-
vine calf aorta and in newborn and adult rat aorta using ultra-
structural immunocytochemical methods and a polyclonal
antibody to bovine aortic lysyl oxidase that we have recently
Materials and Methods
Preparation and Assay of Enzyme
Lysyl oxidase was isolated from 4-M urea extracts of bovine aorta by a
modification (27) of the published procedure (n). The modified procedure
substitutes chromatography through Cibacron Blue-Sepharose (Pharmacia
Fine Chemicals, Piscataway, NJ) for the N,N'-DEAE cellulose step origi-
nally described, thus providing a method which co-purifies the four ionic
variants of lysyl oxidase normally resolved by DEAE cellulose chromatog-
raphy. Prior studies had established that the peptide maps of proteolytic
digests of each species were remarkably similar as were the molecular
masses of each variant (32,000 D) and the substrate and inhibitor profiles
ofcach (25). Because the four species display such structural and catalytic
similarities, use of the co-purified mixture for certain immunological, cata-
lytic, and physical studies seems warranted especially since co-purification
provides larger quantities of enzyme.
Preparation of Antibody
Lysyl oxidase purified as described appeared homogeneous by SDS PAGE
performed according to Laemmli (12). The enzyme was freed of possible
contaminants by excising the band corresponding to iysyl oxidase from the
acrylamide gel for use as the antigen. A narrow strip of the gel slab was
cut out and separately stained with Coomassie Brilliant Blue and used as
a guide for excising the section of the unstained gel corresponding to the
electrophoretic migration position of iysyl oxidase at 32,000 D. This gel-
purified antigen was then used to raise antibodies in chickens as described
(2). Animals were bled at bi-weekly intervals and serum samples were
screened for anti-lysyl oxidase activity by a modification of the dot blot pro-
cedure of Hawkes et al. (9). The co-purified 32,000-D lysyl oxidase antigen
© The Rockefeller University Press, 0021-9525/86/09/1121/8 $1.00
The Journal of Cell Biology, Volume 103, September 1986 1121-1128 1121
(5 ~tg/dot) was immobilized on nitrocellulose, the blot was then incubated
with dilutions of the immune serum and then with rabbit anti-chicken anti-
body coupled to horseradish peroxidase (Miles-Yeda, Inc., Elkhart, IN).
The blot was thoroughly washed as described by Hawkes et al. (12) before
and after incubations with the first and second antibody, respectively. Im-
mobilized immune complexes were visualized by incubation of the blot with
the peroxidase substrate, 4-chloronaphthol, to develop visible color. Freshly
obtained immune serum developed a titer of 1:16,000 by this technique.
Preimmune chicken serum yielded negative reactions in this procedure, as
did controls from which primary antibody or the lysyl oxidase antigen had
Immunoprecipitation of uSI-Labeled Antigen
Purified bovine aortic lysyl oxidase was iodinated with Nat25I in the pres-
ence of 1,3,4,6,-tetrachloro-3o.,6ct-dipbenylglycouril (Iodogen; Pierce Chem-
ical Co., Rockford, IL) as described (7) and then incubated at 26"C with
a 1:30 dilution of chicken anti-lysyl oxidase for 90 rain. Immune complexes
were precipitated by further incubation with rabbit anti-chicken IgG (Miles
Laboratories, Inc., Elkhart, IN) for 120 min. The precipitate was isolated
by centrifugation through a cushion of 1 M sucrose at 4300 g for 7 rain,
resuspended in 50 mM Tris, (pH 7.6) 1 M NaC1, 1% sodium deoxyeholate,
and 1% Triton X-100 and reisolated by centrifugation. The pellet was dis-
solved and heated at 100°C for 2 rain in gel electrophoresis sample buffer
containing 6 M urea, 62.5 mM Tris (pH 6.8), 4% 2-mercaptoethanol, 3%
SDS, and 2% brornophenol blue, and electrophoresed in 10% crosslinked
polyacrylamide slab gels, according to Laemmli (12).
Both anti-lysyl oxidase serum and preimmune serum were partially
purified before use as ultrastructmal probes. Each serum was dialyzed
against 0.02 M Tris HCI (pH 8.0) and applied to columns (1 x 1.5 cm) of
DEAE-Atti-Gel Blue (Bio-Rad Laboratories, Richmond, CA). The IgG
components of the sera passed through this column and were completely
eluted by washing with 0.02 M "Iris HC1 (pH 8.0). Anti-lysyl oxidase activ-
ity was recovered in the IgG eluate as determined by the dot/blot titration
Generation of Enzyme-Collagen Complexes In Vitro
Type I collagen was isolated from fetal calf skin by an adaptation of the
method of Monson and Bornstein (15) as previously described (6). The acid
extraction step was avoided to preserve the N- and C-terminal telopeptides
intact. Collagen types in the neutral salt extract were separated by precipita-
tion with specific concentrations of NaCI. The type I product was chro-
matograpbed over DEAE cellulose under conditions that should resolve
type I tropocollagen from procollagen species and from acidic proteogly-
cans. The purity of the product as type I tropocollagen was established by
amino acid analysis and by SDS PAGE, each of which indicated the absence
of procollagens in this preparation. The product was found to be free of con-
tamination by type BI collagen by the delayed reduction gel electrophoresis
method of Sykes et al. (26). Fibrils of type I collagen were prepared by add-
ing 50 ttl of a 5 mM acetic acid solution of collagen (0.57 mg mi-t) to 1.5
ml polyethylene tubes on ice and diluting to 0.5 ml with 5 mM acetic acid.
To this was added 0.5 ml of a buffer composed of 30 m_M N-tris(hydroxy-
methyl)methyl-2-aminoethane sulfonic acid, 30 mM NazHPO4, and 135
mM NaCI (pH Z5). Fibrils were formed by incubating this mixture at 26"C
for 3 h. Lysyl oxidase in 6 M urea was dialyzed against 5 raM potassium
phosphate (pH 7.5) before use in these studies. Aliquots of 50 ttl containin~
2 ttg of the dialyzed enzyme were added to tubes containing the preformed
collagen fibrils, the mixtures were incubated at 37"C for 10 rain and cen-.
trifuged at 10,000 g for 1.5 min. The supernatants were removed and the
pellets were gently suspended in the fibril-forming buffer and reisolated by
centrifugation to remove unbound enzyme. The pellet of the enzyme-colla-
gen complex was then fixed by treatment with 4% paraformaldehyde for 2 h
on ice and processed as above for electron microscopy.
Aortas of 2-wk-old calves, newborn rat pups and adult Sprague Dawley rats
were freshly obtained and random pieces of aortic tissue were finely minced
and fixed in 4 % paraformaldehyde in 0.1 M sodium phosphate buffer (pH
7.2) for 2 h at 0*C. The tissue was washed four times in this phosphate buffer
in the absence of fixative and then processed into Lowicryl K4M using the
low temperature embedding method of Carlemalm et al. (3). The embedded
samples were polymerized by exposure to ultraviolet light and thin (60-70
nm) sections were cut and mounted on collodion-coated nickel grids. A two-
step indirect immunogold procedure (16) was used to localize antigen in the
tissue specimens reacting with chicken anti-lysyl oxidase. Mounted thin
sections were etched in 10% I-I202, washed in distilled deionized water,
and incubated in a 1:30 dilution of normal goat serum for 30 rain. The grids
were transferred directly into chicken anti-lysyl oxidnse diluted 1:1, 1:10, or
undiluted and incubated in a moist chamber for 1-24 h at 4"C. The sections
were then washed four times in a buffer of 0.1 M Tris and 0.1% BSA'(pH
7.2), and incubated for 1 h with a 1:6 dilution of rabbit anti-chicken IgG
conjugated with 10-urn particles of colloidal gold (E. Y. Laboratories, Inc.,
San Mateo, CA). The grids were extensively washed in filtered Tris-BSA
buffer, rinsed twice in distilled deionized water, air-dried, and stained with
uranyl acetate and lead citrate. Control sections were incubated with preim-
mune chicken serum, chicken anti-lysyl oxidase preincubated with excess
purified lysyl oxidase before addition to tissue sections or gold-conjugated
rabbit anti-chicken IgG in the absence of anti-lysyl oxidase.
After immunogold labeling of lysyl oxidase, some grids were stained
with a 1% palladium chloride solution for 5-15 min, rinsed thoroughly, and
counterstained with uranyl acetate and lead citrate before observation in the
electron microscope (14).
To account for the possibility that proteoglycans might mask the im-
munoreactivity of lysyl oxidase associated with collagen fibers, freshly iso-
lated calf aorta was finely minced and incubated with 1 U m1-1 of chun-
droitinase ABC (Miles Laboratories, Inc.) in 0.1 M Tris buffer (pH 8.0)
containing 0.5% BSA for 1 h at 37"C (20). Control specimens were in-
cubated under these conditions but in the absence of chondroitinase ABC.
After incubation, the tissue preparations were each divided into two ali-
quots, one of which was fixed and embedded for immunolabeling as de-
scribed above, while the other was stained with 1% ruthenium red in McI1-
vaines buffer, (pH 5.6) for 30 min at room temperature. The tissue was then
fixed overnight at 4"C in a solution of 3 % glutaraldehyde and 1% ruthenium
red in Mcllvaines buffer. The next day, the tissue was quickly rinsed four
times with buffer containing ruthenium red, postfixed for 2 h in 1% osmium
tetroxide, and then dehydrated through asoending concentrations of acetone
and embedded routinely in Epon for electron microscopy.
Reactivity of Antibody with Lysyl Oxidase
SDS gel electrophoresis of the purified preparation of bovine
aortic lysyl oxidase used as antigen in the present study
demonstrated that it consisted of an apparently homogeneous
protein species of 32,000 D (Fig. 1 A ). The reactivity of the
chicken anti-lysyl oxidase against this protein was demon-
strated by immunoprecipitation of a preparation of purified
lysyl oxidase that had been radioactively labeled with usI
(Fig. 1 B). Studies in progress have demonstrated that the
antibody also precipitates considerably larger isotopically
labeled proteins obtained from the cell matrix of calf aortic
smooth muscle cells pulsed with radioactive amino acids in
culture (1). It is assumed that these proteins are precursors
of lysyl oxidase or are otherwise related to the 32,000-D
product since the immunoreactivity with these species is
competitively prevented by preincubation of the chicken
anti-lysyl oxidase with the unlabeled, purified 32,000-D
species (1). The antibody is unreactive against peptides solu-
bilized by hot oxalic acid extraction of bovine ligament
elastin, is not reactive toward bovine fibronectin, and does
not precipitate products which are susceptible to bacterial
collagenase from cultured calf smooth muscle cells (data not
shown). The assumption that the antibody is specific for
forms of lysyl oxidase in cellular and tissue specimens is sup-
ported by the competing effect of the purified antigen.
Bundles of amorphous elastin in a section of calf aorta are
shown in Fig. 2, A-C, with microfibrils most evident at the
The Journal of Cell Biology, Volume 103, 1986 1122
periphery of the bundles. Gold deposits signifying the pres-
ence of lysyl oxidase are also present at the periphery of the
elastin bundle in association with microfibrils. Microfibrils
that are not at the elastin interface do not bind the antibody.
The distribution of lysyl oxidase is continuous around the
bundles. Fig. 3, A and B, represent presumed earlier stages
of elastin formation in newborn rat aorta where microfibrils
predominate over amorphous elastin. In addition to apparent
complexes of lysyl oxidase with microfibrils at the periphery
of the developing elastin bundle (Fig. 3 A), there are linear
deposits of lysyl oxidase running through the bundle in as-
sociation with thin strands of microfibrils. Fig. 3 B shows a
bed of microfibrils with extremely small areas of amorphous
elastin. Lysyl oxidase appears in linear, seemingly periodic
deposits within the microfibrillar bed. This figure gives the
impression of a microfibril-lysyl oxidase template within
which linear bands of elastin will ultimately develop.
Adult rat aorta, reacted with anti-lysyl oxidase and gold-
labeled second antibody (Fig. 4), reveals a thin ring of
microfibrils delineating amorphous elastin. Microfibrils are
fewer in number and much less dense than in the newborn
aorta. Gold labeling of the elastin-microfibril complex is
considerably less than in developing elastin, suggesting that
much less lysyl oxidase is available for binding by the
Gold deposits are not seen in control specimens prepared
using chicken anti-lysyl oxidase that had been preadsorbed
by excess purified bovine aortic lysyl oxidase before incuba-
tion with fixed tissue (Fig. 5), thus illustrating the specificity
of the gold labeling patterns seen in this study. Completely
negative results were also obtained with preimmune chicken
serum or if the chicken anti-lysyl oxidase but not the gold-
labeled second antibody was omitted from the immunolocal-
ization protocol described in Materials and Methods. The
gold deposits seen in bovine calf, rat pup, and adult rat aortae
thus appear to localize epitopes found in purified 32,000-D
bovine aortic lysyl oxidase.
Gold deposits are not seen on the longitudinal or trans-
verse sections of collagen fibrils in the aortic sections exam-
ined (see Fig. 2), nor were gold deposits clearly visible
within cell structures in the process of secretion or transport
to the extracellular microfibrillar loci. The apparent lack of
gold deposits on collagen fibers was somewhat surprising
since prior studies have shown that purified lysyl oxidase
binds to native fibers of type I collagen in vitro (6). Gold
deposits were also not seen on collagen fibers of tissue speci-
mens that had been preincubated with chondroitinase ABC
to remove proteoglycans associated with collagen before
incubation with the first and second antibodies (data not
shown). To further investigate this result, complexes of
purified lysyl oxidase with native fibers of type I collagen
were generated in vitro as described and examined by the
anti-lysyl oxidase immunolocalization technique. As shown
(Fig. 6) gold deposits are seen in association with these col-
lagen fibers. Collagen fibers to which the enzyme had not
been added in vitro did not exhibit gold deposits.
The present report identifies immune-reactive product in as-
sociation with microfibrillar material found at the periphery
of elastic fibers in the extracellular space. The absence of
Figure 1. (A) SDS PAGE of lysyl oxidase (left lane) purified from
bovine aorta. The molecular mass of the purified lysyl oxidase is
32,000 D. Standard proteins of the molecular masses (xl0 -3) indi-
cated are in the right lane. (B) SDS PAGE of the immunoprecipi-
tared uSI-labeled derivative of the purified lysyl oxidase shown in
A. [t251]Lysyl oxidase was precipitated from solution by incubation
with chicken anti-lysyl oxidase and then with rabbit anti-chicken
IgG. The immunoprecipitate was dissolved and heated at 100°C in
the electrophoresis sample buffer before electrophoretic resolution
on the gel. The radioactive band was visualized by autoradiography.
gold deposit in specimens treated with preimmune serum,
in specimens treated with immune serum that had been
pretreated with the 32,000-D aortic lysyl oxidase antigen that
had been purified to apparent homogeneity, and in those
specimens in which the primary antibody was omitted argues
in favor of the conclusion that the discrete gold deposits iden-
tify lysyl oxidase protein at these sites. The complete inhibi-
tion of the deposition of gold by preincubation of anti-lysyl
oxidase with purified lysyl oxidase is consistent with the con-
clusion that the antibody both recognizes and is specific for
lysyl oxidase. The fact that gold is not deposited on all
microfibrils but is only deposited on those at the interface
with elastin indicates that the antibody is not directed against
microfibrils themselves or against a component common to
all microfibrils. Moreover, the distribution of gold deposits
is not consistent with that expected of fibronectin since col-
lagen fibers in tissue specimens were not reactive with
anti-lysyl oxidase in the present study but would be expected
to react with antibodies directed against fibronectin (28).
The presence of lysyl oxidase in association with micro-
fibrillar material surrounding elastin fibers is of particular in-
terest. These results suggest that elastogenesis is character-
ized by accumulation of microfibrillar structures adjacent to
the surface of the fibrogenlc cell; lysyl oxidase associates
with these microfibrils, after which elastin fibrils appear
within the microfibril-lysyl oxidase aggregates. The micro-
fibrillar material remains in association with the periphery
of the elastin component but is not evident within the elastin
fiber which remains structurally amorphous. A model for the
Kagan et al. Ultrastructural Localization of Lysyl Oxidase
Figure 2. (,4) Low power view
of elastic fibers from bovine
calf aorta. Gold particles (ar-
rowheads) demonstrate lysyl
oxidase localized to the elas-
tin-microfibril interface sur-
rounding the amorphous elas-
tin bundle (EL). Col, collagen.
Bar, 0.2 Ixm. (B) Edge of
elastin bundle from bovine
calf aorta. Amorphous elastin
(EL) is bordered by micro-
fibrils (MF). Lysyl oxidase
antibody labeled with 10-nm
gold particles (arrows) is
present at the interface of
elastin and microfibrils in an
irregular distribution. Not all
microfibrils are associated with
lysyl oxidase. Collagen (COL)
appearing in longitudinal and
cross section displays no lysyl
oxidase. Bar, 0.1 lan. (C)
Edge of elastin bundle from
bovine calf aorta also stained
with palladium chloride. The
outline of the amorphous elas-
tin (EL) can be more readily
appreciated in this section. Ly-
syl oxidase antibody labeled
with gold (arrowheads) is
present at the elastin-micro-
fibril interface. Those micro-
fibrils (MF) away from the
elastin bundle do not bind the
antibody. Note occasional gold
deposits on the few micro-
fibrils seen within the amor-
phous elastin. Bar, 0.1 ~m.
The Journal of Cell Biology, Volume 103, 1986 1124
Figure 3. (a) Newborn rat aorta; developing bundles of elastin. Bundles of amorphous elastin (EL) surrounded by extensive array of
microfibrils (MF), some of which appear to run through the bundles. Aggregates of 10-nm gold particles localizing lysyl oxidase (arrow°
heads) appear at the interface of elastin and microfibrils. (b) Bed of microfibrils (MF) with little amorphous elastin (EL). Aggregates of
lysyl oxidase (arrowheads) appear in a pattern that may define sites of subsequent elastin deposition. Bars, 0.1 ~m.
Kagan et al. Ultrastructural Localization of Lysyl Oxidase
Figures 4-6 (Fig. 4) Adult rat aorta with amorphous elastin (EL) and microfibrils (MF). Gold particles (arrowheads) identify sparse local-
ization of lysyl oxidase associated with adult fibers. (Fig. 5) Bovine calf aorta; incubated with excess purified exogenous lysyl oxidase before
application of anti-lysyl oxidase. Lack of gold labeling signifies specific inhibition of antibody binding to elastin (EL), microfibrils (MF),
and collagen (COL). (Fig. 6) Isolated pellet of native type I collagen (COL) complexed with exogenous lysyl oxidase. Note binding of
antibody (arrowheads) to collagen fibers. Bars, 0.1 lain.
morphogenesis of elastin fibers has been proposed that is
consistent with the present observations. This model sug-
gests that microfibrils are first formed in the extraceUular
space.in the pericellular environment and then may serve as
a scaffolding for elastin deposition (4, 5, 17-19). Noting the
continuing association of microfibrils with the developing
elastin fiber, Cleary et al. (5) had speculated that the
microfibrillar material may be associated with lysyl oxidase.
The present findings substantiate that possibility. Indeed, the
present observation that lysyl oxidase is associated with
microfibrillar deposits before and after the appearance of
amorphous elastin within the area circumscribed by the
microfibrils is consistent with the hypothesis that this com-
plex plays an essential role in elastin fiber biosynthesis. The
The Journal of Cell Biology, Volume 103, 1986 1126
paucity of microfibfiUar material and gold deposits in the in-
terstices of the amorphous elastin fiber is consistent with a
model predicting the radial growth of the elastin fiber by the
deposition and crosslinking of newly accreting tropoelastin
precursors of elastin at the fiber-microfibril interface. It
seems possible that the microfibrillar material may actively
participate in this process by providing a matrix for the align-
ment of enzyme, tropoelastin and the preexistent elastin
fiber, thus facilitating radial fiber growth.
The association of lysyl oxidase with only some of the
microfibrils suggests selectivity of the elastogenic process.
Indeed, the linearity of the localization of lysyl oxidase on
microfibrils where little amorphous elastin is visualized
(Fig. 3 B) further suggests that this complex may act as a
template for deposition of the elastic fiber. It is also evident
that lysyl oxidase remains associated with the microfibrillar
components of mature elastic bundles as revealed by the
micrographs of the enzyme localization in the adult rat aorta.
The persistance of enzyme in the adult tissue may signal that
the synthesis and/or repair of elastin fibers continues as the
animals age. It is important to note, however, that the im-
munological marker for lysyl oxidase does not differentiate
between catalytically functional and nonfunctional enzyme.
Notably, gold deposits were not found in association with
extracellular collagen fibers in tissue specimens, although
the enzyme was seen in association with collagen when lysyl
oxidase was added to collagen fibers in vitro. In contrast with
the present results, collagen fibers in the extracellular ma-
trices of fibrotic rat liver and chick tendon were labeled in
a prior study that used immunofluorescently labeled anti-
body directed against lysyl oxidase purified from chick carti-
lage (22). The purification method used for the isolation of
the chick cartilage enzyme involved chromatography of urea
extracts of cartilage on DEAE cellulose and on colla-
gen-Sepharose columns. In contrast, the isolation of bovine
aorta lysyl oxidase used as antigen in the present study em-
ployed DEAE cellulose, Cibacron Blue Sepharose and gel
exclusion chromatographic steps (27). It has been our ex-
perience that gel exclusion chromatography in 6 M urea is
required to free the enzyme product of traces of fibronectin
and other high molecular weight contaminants (27). We had
previously observed this also to be necessary with enzyme
purified by DEAE cellulose and collagen-Sepharose (11).
Because specific information was not given about the
specificity of the antibody raised against the chick cartilage
enzyme (22), it appears possible that the immunolocali-
zation seen in the earlier study (22) may in part reflect a
reaction of the antibody with other collagen-bound macro-
molecules such as fibronectin. Moreover, although collagen-
or elastin-specific forms of lysyl oxidase have not been
identified by assays of isolated enzymes in vitro, it also re-
mains possible that the discrepancy between the im-
munolocalization results of the present study with those of
Siegel et al. (22) may reflect such enzyme specificity differ-
ences between the chick cartilage and bovine aortic enzymes
It is unlikely that collagen-associated proteoglycans pre-
vented access of the antibody to lysyl oxidase since preincu-
bation of tissue specimens with chondroitinase ABC did not
alter the distribution of the antibody binding sites. It is possi-
ble, however, that the extracellular collagen fibers seen in the
tissue specimens are fully crosslinked products from which
bound lysyl oxidase had previously dissociated. Indeed, re-
cent studies have noted that collagen-lysyl oxidase com-
plexes formed in vitro do slowly dissociate (6). Further, al-
though the purified enzyme used as antigen in the present
study oxidizes both collagen and elastin substrates in vitro,
the possibility has not been excluded that collagen-specific
forms of lysyl oxidase may exist in vivo that are not recog-
nized by the antibody directed against the bovine aortic
This research was supported by National Institutes of Health Grants HL-
19717, HL-13262, and AM-18880.
Received for publication 5 December 1985, and in revised form 29 May
1. Bronson, R., D. Reiss, G. Sonensbein, and H. Kagan. 1985. Biochemical
and immunological characterization of lysyl oxidase in cultured bovine aortic
smooth muscle cells. Fed. Proc. 44:663.
2. Burbeio, P. D., H. M. Kagan, and C. O. Chicbester. 1985. Immunolog-
ical characterization of bovine lysyl oxidase. Comp. Biochem. Physiol. 81B:
3. Carlemalm, E., R. M. Garavilo, and W. Villinger. 1982. Resin develop-
ment for electron microscopy and an analysis of embedding at low temperature.
J. Microsc. (Oxf.). 126:123-143.
4. Cleary, E. G., and M. A. Gibson. 1983. Elastin-associated microfibriis
and microfibrillar proteins. Int. Rev. Connect. Tissue Res. 10:97-209.
5. Cleary, E. G., J. C. Fanning, and I. Prosser. 1981. Possible roles of
microfibrils in elasmgenesis. Connect. Tissue Res. 8:161-166.
6. Cronlund, A. L., B. D. Smith, and H. M. Kagan. 1985. Binding oflysyl
oxidase to fibrils of type I collagen. Connect. Tissue Res. 14:109-119.
7. Fraker, P. J., and J. C. Speck. 1978. Protein and cell membrane iodina-
tions with a sparingly soluble chloroamide, 1,3,4,6-tetraehloro-3a,6a-diphe-
nylglycouril. Biochem. Biophys. Res. Commun. 80:849-857.
8. Gonnerman, W. A., R. Ferrara, and C. Franzblan. 1981. Measurement
of medium lysyl oxidase activity in aorta smooth muscle cells. Effects of multi-
ple medium changes and inhibition of protein synthesis. Biochemistry. 20:
9. Hawkes, R., E. Niday, and J. Gordon. A dot-immunobinding assay for
monoclonal and other antibodies. 1982. Anal. Biochem. 119:142-147.
10. Kagan, H. M. 1986. Characterization and regulation of lysyl oxidase. In
Biology of the Extraceilular Matrix. R. P. Mecham, editor. Vol. 1. Academic
Press, Orlando, Florida. 321-398.
11. Kagan, H. M., K. A. Sullivan, T. A. Olsson, and A. L. Cronlund. 1979.
Purification and properties of four species of lysyl oxidase from bovine aorta.
Biochem. J. 177:203-214.
12. Laemmli, U. K. 1970. Cleavage of structural proteins during the assem-
bly of the head of bacteriophage T4. Nature (Lond.). 227:680-687.
13. Layman, D. L., A. S. Narayanan, andG. R. Martin. 1972. The produc-
tion of lysyl oxidase by human fibroblasts in culture. Arch. Biochem. Biophys.
14. Monson, J. M., and P. Bornstein. 1973. Identification of a disulfide-
linked procollagen as the biosynthetic precursor of chicken collagen. Proc.
Natl. Acad. Sci. USA. 70:3521-3525.
15. Morris, S. M., P. J. Stone, W. A. Rosenkrans, J. D. Calore, J. T. AI-
bright, and C. Franzblau. 1978. Palladium chloride as a stain for elastin at the
ultrastructural level. J. Histochem. Cytochem. 26:635-644.
16. Probert, L., J. DeMay, and J. Polack. 1981. Distinct subpopulations of
enteric p-type neurons containing substance P and vasoactive intestinal peptide.
Nature (Lond. ). 294:470--471.
17. Robert, B., M. Szigeti, J. Derouette, L. Robert, H. Bouisson, and M. T.
Fabre. 1971. Studies on the nature of the microfibrillar component of elastic
fibers. Fur. J. Biochem. 21:507-516.
18. Ross, R. 1973. The elastic fiber. A review. J. Histochem. Cytochem.
19. Ross, R., and P. Bornstein. 1969. The elastic fiber. I. The separation and
partial characterization of its macromolecular components. J. Cell BioL
20. Salto, H., T. Y. Yamagata, and S. Suzuki. 1968. Enzymatic methods for
determining small quantities of isomeric chondroitin sulfates. J. Biol. Chem.
21. Siegel, R. C. 1979. Lysyl oxidase. Int. Rev. Connect. Tissue Res.
22. Siegel, R. C., H. K. Chen, J. S. Grcenspan, and J. M. Aguiar. 1978.
Biochemical and immunocbemical study of lysyl oxidase in experimental he-
patic fibrosis in the rat. Proc. Natl. Acad. Sci. USA. 75:2945-2949.
23. Siegel, R. C., and J. C. C. Fu. 1976. Collagen erosslinking. Purification
and substrate specificity of lysyl oxidase. J. Biol. Chem. 251:5779-5785.
Kagan et al. Ultrastructural Localization of Lysyl Oxidase
24. Starcher, B. C., J. A. Madaras, and A. S. Tepper. 1977. Lysyl oxidase
deficiency in lung and fibroblasts from mice with hereditary emphysema. Bio-
chem. Biophys. Res. Commun. 78:706-712.
25. Sullivan, K. A., and H. M. Kagan. 1982. Evidence for structural similar-
ities in the multiple forms of aortic and cartilage lysyl oxidase and a catalytically
quiescent aortic protein. J. Biol. Chem. 257:13520-13526.
26. Sykes, B. C., B. Puddle, M. Francis, and R. Smith. 1976. The estimation
of two collagens from human dermis by interrupted gel electrophoresis. Bio-
chem. Biophys. Res. Commun. 72:1472-1480.
27. Williams, M. A., and H. M. Kagan. 1985. Assessment of lysyl oxidase
variants by urea gel electrophoresis: evidence against disulfide isomers as bases
of the enzyme heterogeneity. Anal. Biochem. 149:430-437.
28. Yamada, K. M. 1983. Cell surface interactions with extracellular
materials. Annu. Rev. Biochem. 52:761-799.
The Journal of Cell Biology, Volume 103, 1986 1128