(Received for publication, February 24, 1997)
From the Committee on Cell Physiology and the Department of
Pharmacological and Physiological Sciences, University of Chicago,
Chicago, Illinois 60637 and the Department of Anatomy and
Cell Biology, University of Michigan,
Ann Arbor, Michigan 48109-0616
Members of the L1 family of homophilic neural
cell adhesion molecules are thought to play an important role in
nervous system development and function. It is also suggested that L1
is a direct target of ethanol in fetal alcohol syndrome, since ethanol
inhibits the aggregation of cultured cells expressing L1 (Ramanathan,
R., Wilkemeyer, M. F., Mittel, B., Perides, G., and Charness, M. E. (1996) J. Cell Biol. 133, 381-390). If ethanol acts
directly on the homophilic adhesive function of the L1 molecule, then
inhibition of aggregation by ethanol should be observed in any cell
type that expresses L1. Here we examined the effect of physiologically relevant concentrations of ethanol on the aggregation of
Drosophila S2 cells that expressed either neuroglian (the
Drosophila homolog of L1) or human L1. The aggregation of
these S2 cells is known to be solely dependent on the homophilic
interactions between L1 or neuroglian molecules. Neither cell adhesion
molecule was affected when cell aggregation assays were carried out in
the presence of 38 mM ethanol. The recruitment of
membrane skeleton assembly at sites of cell-cell contact (a
transmembrane signaling function of human L1) was also unaffected by
the presence of ethanol. Thus the previously described inhibition of
cell adhesion by ethanol in L1-expressing cells cannot be explained by
a simple direct effect on the adhesive activity of L1 family
members.
Human L1 and related neural cell-adhesion molecules in other vertebrates are thought to play an important role in the development and function of the nervous system. Mutations in the human L1 gene are associated with several phenotypes including hydrocephalus, macrocephaly, mental retardation, agenesis of the corpus callosum, aphasia, and adducted thumbs (1, 2). Studies of the L1 molecule in vitro have revealed some of the mechanisms that are likely to be responsible for its important roles in vivo. First, L1 behaves as a homophilic cell adhesion molecule that promotes aggregation of L1-transfected fibroblasts (3) or latex beads that display purified L1 (4). Second, L1 adhesion transduces a signal that promotes neurite outgrowth from neurons grown on an L1 substrate (5). The pharmacological properties of L1-mediated signaling suggest that it may act through a second messenger system that involves the fibroblast growth factor receptor (6, 7). Third, L1-mediated adhesion induces ankyrin assembly specifically at sites of cell-cell contact (8, 9). There is a direct interaction between the cytoplasmic domain of L1 family members and ankyrin (8-10), and the regulation of this interaction by cell adhesion may contribute to the adhesion-based signal for neurite outgrowth.
Fetal alcohol syndrome is a clinical manifestation of alcohol abuse in pregnant women (11). Charness et al. (12, 13) noted that some of the neurological defects associated with fetal alcohol syndrome overlap with the phenotype of human L1 mutations. Their studies revealed that the aggregation of cultured fibroblasts expressing human L1 is significantly inhibited by the presence of ethanol in the culture medium at concentrations comparable to blood alcohol levels reached after consuming alcoholic beverages. In one study, transfected fibroblasts constitutively expressing a cDNA-based human L1 minigene were tested for their ability to aggregate in the presence of 0-50 mM ethanol (13). Control untransfected cells did not aggregate under any conditions. The L1-expressing cells aggregated extensively in the absence of ethanol, but their aggregation was greatly reduced in the presence of as little as 5 mM ethanol. Half-maximal inhibition was observed at 7 mM ethanol. Based on these results, it was suggested that ethanol acts directly on the L1 molecule to inhibit its adhesive activity. One interesting implication of these results is that the L1 molecule may be a target of ethanol in vivo, thereby suggesting a mechanism through which ethanol could exert some of its pathological effects.
The Drosophila cell adhesion molecule neuroglian is related to vertebrate L1 in domain structure and sequence (14). Like L1, neuroglian exhibits homophilic cell adhesion when expressed in Drosophila S2 tissue culture cells (15). Neuroglian also behaves as a signal-transducing molecule that recruits assembly of the membrane cytoskeleton (ankyrin and spectrin) at sites of cell-cell contact (8). S2 cells expressing a recombinant human L1 gene under control of the metallothionein promoter exhibit similar adhesive properties as well as the ability to recruit ankyrin to cell contacts (9). Thus S2 cells provide a useful experimental system in which to further examine the properties of the L1/neuroglian family of cell adhesion molecules. Here we used transfected S2 cells to examine the effects of ethanol on the activities of neuroglian and human L1.
Drosophila S2 cells were grown at 25 °C in Schneider's Drosophila medium supplemented with 15% fetal calf serum (Life Technologies, Inc.) and 50 units/ml penicillin/streptomycin. S2 cells were stably transfected with either a Drosophila neuroglian (15) or a human L1 transgene (9). A complete human L1-CAM cDNA including the alternatively spliced RSLE motif in the cytoplasmic domain was kindly supplied by Dr. Vance Lemmon (16). The L1 cDNA was subcloned as a 4.5-kilobase pair EcoRI fragment into the pRmHa3 vector as described by Hortsch et al. (9). Transgene expression was induced by the addition of copper chloride to the growth medium at a final concentration of 0.7 mM for 20-24 h.
Homotypic Cell Aggregation AssayNeuroglian or L1 transfected cells were seeded at 1 × 106 cells/ml in 4.5-cm2 12-well tissue culture plates and incubated for 20 h at 25 °C. Cultures were gently rocked for an additional 4 h to promote cell aggregation. Cells were grown in culture medium alone, medium containing 100 mM ethanol (Aaper Alcohol & Chemical Co., Shelbyville, KY), or medium diluted with an equal volume of Drosophila Ringer's solution (17). Evaporation of ethanol was minimized by sealing the plates with parafilm. For quantitation of aggregation, cultures were seeded at 5 × 106 cells/ml under the above conditions. Cells were carefully resuspended at 24 h and diluted 50-fold, and single cells were counted on a hemacytometer.
Measurement of EthanolOne ml of medium containing induced neuroglian cells seeded at 1 × 106 cells/ml, medium alone, medium containing 50 mM ethanol, or medium containing 100 mM ethanol were placed in 12-well tissue culture plates, sealed with parafilm, and incubated for 24 h at 25 °C. A sample of the medium was sealed in a microcentrifuge tube and stored at 4 °C for 24 h as a control. A fresh sample of ethanol-containing medium was also prepared immediately before the ethanol assay as a further control for evaporation. Ethanol concentration was measured in duplicate samples with an ethanol assay kit (Sigma) according to the manufacturer's instructions.
Antibodies, Immunofluorescent Staining, and MicroscopyThe primary antibodies (used at 1:500) were mouse monoclonal anti-human L1 (5G3, Ref. 18), mouse monoclonal anti-neuroglian (3F4, Ref. 15), and affinity-purified rabbit anti-ankyrin antibody (19). Secondary antibodies (used at 1:1000) included affinity-purified fluorescein isothiocyanate-conjugated goat anti-rabbit Ig (Zymed Laboratories, Inc.), Texas Red-conjugated goat anti-mouse Ig (Jackson ImmunoResearch Laboratories, Inc.), and alkaline phosphatase-conjugated goat anti-mouse Ig (Zymed Laboratories, Inc.). Alcian blue-coated microscope slides were used to immobilize cells for staining. Cells were fixed, permeabilized, and stained as described previously (8). Primary and secondary antibodies were diluted in Tris-buffered saline supplemented with 5% newborn calf serum (Life Technologies, Inc.) and 0.1% Tween 20 (Sigma) and then allowed to react with cells for 2 h at room temperature. After each incubation cells were rinsed three times with Tris-buffered saline containing Tween 20. Cells were mounted in Testog fluorescein isothiocyanate guard (Testog, Inc.), viewed, and photographed with an ausJena microscope using a 50× 0.95 NA Plan Apo objective.
For direct analysis of aggregation, cells were transferred to microscope slides and viewed and photographed with differential interference contrast optics using a 12.5× objective. Fields were photographed on TMax 400 film and digitized with a Polaroid Sprintscan scanner.
Cell Extractions and ElectrophoresisNeuroglian levels in induced cells grown in medium alone, medium containing 100 mM ethanol, and medium containing Drosophila Ringer's solution (as above) were compared in Western blots of total cellular proteins probed with anti-neuroglian antibody. Cultures were seeded at 1 × 106 cells/ml, induced with copper chloride, and incubated for 20 h at 25 °C. Cells were centrifuged, rinsed twice in Drosophila Ringer's, solubilized in Laemmli sample buffer, and boiled for 2-3 min. Sample loadings were normalized according to protein absorbance at 280 nm. After electrophoresis, samples were transferred electrophoretically to nitrocellulose and reacted with anti-neuroglian antibody 3F4 and alkaline phosphatase-conjugated secondary antibody as described previously (20).
Drosophila S2 cells ordinarily grow as a single cell
suspension with no stable aggregate formation. S2 cells expressing
recombinant neuroglian under control of the Drosophila
metallothionein promoter (15) form large stable aggregates within
24 h of induction with 0.7 mM copper chloride (Fig.
1A). Here we examined the effect of ethanol
on the aggregation of neuroglian-expressing cells. Aggregate formation
was not visibly altered when cells were induced in the presence of up
to 100 mM ethanol (Fig. 1B). Preformed cell aggregates were also unaffected by the addition of 100 mM
ethanol to the culture medium for 30 min (not shown). To quantify the extent of aggregation in control and ethanol-treated cultures, unaggregated cells remaining in the culture were counted with a
hemacytometer 24 h after neuroglian induction. Similar numbers of
single cells were detected whether or not ethanol was present in the
culture medium during the aggregation assay (Table
I).
|
A dramatic decrease in neuroglian-mediated cell aggregation occurred when cells were grown in culture medium diluted 1:2 with Drosophila Ringer's solution (Fig. 1C). In initial experiments, ethanol was diluted first in Drosophila Ringer's solution and then added to the neuroglian-expressing cell culture. Subsequent control experiments revealed that the inhibition of aggregation (Fig. 1C) was due to the Ringer's solution and not the ethanol. Thus neuroglian-mediated cell adhesion was sensitive to the culture conditions but not to the presence of ethanol in the culture medium.
The synthesis of neuroglian was monitored in Western blots of induced
cell cultures grown in the presence and absence of ethanol or
Drosophila Ringer's solution. High level expression of
neuroglian was observed in response to induction with copper chloride
(Fig. 2), whether ethanol was present (lane
2) or absent (lane 1). In contrast, neuroglian
synthesis was significantly suppressed (~2-fold) in cultures treated
with Drosophila Ringer's solution (Fig. 2, lane
3). However, Ringer's-treated cells did form aggregates during the second day of induction (not shown) once neuroglian accumulated to
a sufficient level. Thus the inhibition of adhesion caused by Ringer's
solution was simply due to the slowed rate of neuroglian synthesis and
not to a direct effect on neuroglian activity.
One possible explanation for the lack of an ethanol effect on neuroglian-mediated adhesion is the loss of ethanol from the growth medium during prolonged incubation. The previous demonstration of an ethanol effect on human L1-mediated aggregation was carried out over a short time course (30 min; Ref. 13) whereas the present experiments were carried out over a 24-h time course since aggregation required induction of the neuroglian gene. The ethanol concentration during these experiments was determined by a colorimetric assay (Table II). There was a significant drop in the ethanol concentration over time from a starting concentration of 50-100 mM ethanol to a final level of 19-38 mM, respectively, after 24 h of incubation. The drop was not due to metabolism of ethanol by the S2 cells since their presence or absence did not affect the final concentration of ethanol in the medium. However, the amount of ethanol remaining after 24 h (38 mM) produced a near-maximal effect on the adhesion of L1-expressing fibroblasts in the previous study (13). Therefore, the lack of an effect of ethanol in the aggregation experiments described here was not due to the loss of ethanol from the medium.
|
Neuroglian and human L1 are related to one another in domain
organization, amino acid sequence (28.5% overall identity; Refs. 14
and 16), and homophilic adhesive behavior (15). To test the possibility
that the ethanol insensitivity of neuroglian was due to its divergence
from the human L1 protein, we examined the effect of ethanol on the
aggregation of S2 cells that expressed an inducible human L1 minigene
(9). The L1-expressing S2 cells formed aggregates (Fig.
3B) under the same conditions as the
neuroglian-expressing cells described above, although the aggregates
were generally smaller after 24 h. L1-expressing cells aggregated
normally when the assay was carried out in the presence of 100 mM ethanol (38 mM final concentration, Fig.
3C). Quantitative analysis also failed to detect a
significant effect of ethanol on the aggregation of L1-expressing cells
(Table I).
A further test of L1 sensitivity to ethanol was carried out by staining
cell aggregates with antibodies against ankyrin and human L1. Both
neuroglian and L1 recruit ankyrin assembly at sites of cell contact in
response to cell adhesion (8, 9). A representative field of
L1-expressing cells grown in the presence of 100 mM ethanol is shown in Fig. 4 (A, stained for ankyrin;
B, for human L1). Similar random fields of L1-expressing
cells were scored for the presence of brightly stained aggregates and
weakly stained single cells (Table III). The bright
staining of ankyrin and L1 at cell contacts made it possible to
accurately distinguish small cell aggregates from single cells. Once
again, no difference was observed between cells grown in the presence
or absence of 100 mM ethanol. Ankyrin was recruited
specifically to sites of cell-cell contact. L1 also appeared
concentrated at contacts, although control experiments showed that much
of the total L1 was extracted by detergent during cell permeabilization
(8, 9).
|
The above results show that the ability of two members of the L1 protein family to promote aggregation of Drosophila S2 cells is insensitive to physiologically relevant concentrations of ethanol. Adhesive activity was measured using an established assay for cell aggregation in response to the inducible expression of Drosophila neuroglian (15) or human L1 (9) in Drosophila S2 tissue culture cells. The robust aggregation exhibited by these cells was not inhibited by the presence of ethanol. Control experiments revealed that the level of ethanol in the growth medium remained physiologically significant throughout the duration of the assays.
Differences in assay conditions between the present study and previous studies do not explain the lack of an ethanol effect on cell aggregation. The cells used by Ramanathan et al. (13) constitutively expressed human L1, but were prevented from aggregating with each other because they grew as adherent cells in culture. Once released from their substrate, the suspended L1-expressing cells formed aggregates within 30 min. In contrast, the Drosophila S2 cells used here expressed neuroglian or human L1 under the control of an inducible promoter. Since S2 cells grow in suspension, cell adhesion commences once the induced adhesion molecule accumulates to a sufficient level (at ~16 h). Ethanol was added to the cultures at the time of adhesion molecule induction, and the effect of ethanol was scored 24 h later after cell aggregates were allowed to form. Despite these differences in culture conditions and regulation of L1 expression, both experimental systems ultimately tested the ability of L1-expressing cells to aggregate in the presence of ethanol.
There are several possible explanations for the inhibition of L1-mediated adhesion by ethanol in one cell system, but not another. First, ethanol may somehow act on the fibroblasts used in the previous study to produce an indirect effect on L1-mediated adhesion. The inhibitory effect of Ringer's solution in the present study illustrates that culture conditions can indirectly influence the outcome of a cell aggregation assay without affecting the activity of the adhesion molecule. However, ethanol-induced degradation of human L1 in fibroblasts was ruled out by the finding that the effects of ethanol were rapidly reversible (13). A pleiotropic effect on cell adhesion was also ruled out by the finding that N-CAM-mediated cell adhesion was not inhibited by ethanol (13). Second, there may be a post-translational modification of L1 (e.g. a difference in glycosylation) that causes L1 to become ethanol-sensitive in mammalian fibroblasts, but not in Drosophila S2 cells. It is possible that without this putative modification, L1 and neuroglian exhibit ethanol-insensitive adhesion, but whether or not there is a modification of L1 in fibroblasts, there is no evidence for an ethanol-sensitive modification of L1 function in the nervous system. Third, there may be interacting protein(s) found in fibroblasts but not S2 cells that confer ethanol sensitivity on the aggregation of L1-expressing cells. For example, heterophilic ligands that interact with L1 (2) may have contributed to the cell aggregation reported by Ramanathan et al. (13). If so, perhaps it is the heterophilic partner rather than L1 that is responsible for the ethanol effect. Nevertheless, the present results demonstrate that the homophilic adhesive activities of two L1 family members, as well as their cytoplasmic interaction with ankyrin, are not detectably altered by the presence of ethanol. Future studies should address the possibility that an indirect mechanism is responsible for the ethanol sensitivity of adhesion in L1-expressing fibroblasts.
We thank Dr. Allan Bieber for providing anti-neuroglian antibody 3F4, Dr. Ralph Reisfeld for providing anti-human L1 antibody 5G3, and Drs. Michael Charness and Vance Lemmon for discussing unpublished studies of the effect of ethanol on aggregation of L1-expressing cells.