Adhesion mutants of Dictyostelium discoideum lacking the saccharide determinant recognized by two adhesion-blocking monoclonal antibodies

DEVELOPMENTAL

BIOLOGY

Adhesion

109,

111-117

(19%)

Mutants of Dictyostelium discoideum Lacking the Saccharide Determinant Recognized by Two Adhesion-Blocking Monoclonal Antibodies

WILLIAM

F.LOOMIS,*

Departments

of

STEVE A. WHEELER,*

WAYNE R. SPRINGER,~AND

SAMUEL H. BARONDES~

*Biology and tP.sychiatry, University of California at San Diego, La Jolla, Cal?&rnia and tveterans Administratimz Medical Center, San Diego, California 92161 Received

August

9, 1984;

accepted

in revised

form

November

92093,

28, 1984

A mutant of Dictyostelium diswideum, strain HL260, was isolated based on its failure to bind d-41, a monoclonal antibody that blociks developmentally regulated cell-cell adhesion. The mutant fails to normally acquire cell-cell adhesion as assayed with cells shaken in 10 mM EDTA, but aggregates and constructs fruiting bodies. Other mutant strains, HL216 and HL220, previously shown to have impaired cell-cell adhesion, also lack the determinant that binds d-41. The three strains all carry mutations in a gene designated mod B, which directs a post-translational modification of several developmentally regulated D. discwideum glycoproteins. Diploids formed between independent mod B mutant haploid strains also lack this determinant and show marked impairmant of cell-cell adhesion in EDTA, indicating Ithat mutations in mod B, rather than other mutations not shared by the haploid strains, are related to the adhesion defect. The results are consistent with other evidence that an oligosaccharide carried on several developmentally regulated glycoproteins plays an essential role in EDTA-resistant cell-cell adhesion in D. disctridewn. However, this type of adhesion is not essential for morphogenesis in that the only defect detected thus far in mod B mutamt strains is that they construct relatively smaller fruiting bodies that contain fewer spores. 0 1985 Academic

Press, Inc.

B locus on linkage group VI (Murray et al, 1984). We now show that strains carrying either mod B501 or mod B502 mutations also fail to bind monoclonal antibody d-41. Diploids formed between these strains and strain HL260 lack the determinant recognized by d-41 and E28D8 and do not develop EDTA-resistant adhesion, suggesting that these properties are correlated and are not the result of other unrelated mutations. Thus, these results support a direct relationship between the mod B-dependent determinant and developmentally acquired adhesion.

1NTR:ODUCTION

Monoclonal antibody d-41 blocks EDTA-resistant cell-cell adhesion of Dictyostelim discoideum and reacts with several glycoproteins (Springer and Barondes, 1983, 1984) including one which migrates with gp80, a glycoprotein previously implicated in adhesion (Miiller et aZ., 1979). The determinant it recognizes is sensitive to periodate oxidation but resistant to exhaustive digestion by Pronase (Springer and Barondes, 1985). These and other results suggest that an oligosaccharide found on these glycoproteins may be directly involved in cell-cell adhesion. To test this hypothesis, we sought to isolate a mutant that lacks this oligosaccharide. We screened a mutagenized population for clones that failed to bind monoclonal d-41 when difl’erentiated to the aggregation stage. We here describe such a mutant strain, HL260, and show that it has impaired cell-cell adhesion of the type characteristic of aggregating cells, when measured in a standard in vitro assay in 10 mM EDTA. Mutations resulting in the lack of the oligosaccharide recognized by monoclonal antibody E28D8 have been previously isolated and shown to be impaired in EDTAresistant adhesion (Murray et al., 1984). These mutations were found to be recessive and to lie in the mod

METHODS

Organism. D. discoideum strain HL80 was grown in association with Klebsiella aerogenes and allowed to develop in Millipore filters (Murray et ab, 1983). Strains HL216 (mod B501) and HL220 (mod B502) have been previously described (Murray et al., 1984). Strain AX3 was grown axenically in HL5 medium (Loomis, 1971). Genetic analysis. Mutants were selected from a population of strain AX3 that was treated with N-methylN’-nitro-N-nitrosoguanidine as previously described (Murray et aZ., 1983) except that the cells were allowed to develop directly in the wells of multitest plates. HL-

111

0012-1606185 Copyright All rights

$3.00

0 1985 by Academic Press. Inc. of reproduction in any form reserved.

112

DEVELOPMENTAL

BIOLOGY

5 medium was shaken out of the multitest plates and replaced with 20 mM phosphate buffer, pH 6.4. The cells were allowed to develop for 18 hr at 22°C before the buffer was shaken out and the cells fixed to the bottom of the wells. The wells were screened by an ELISA test using monoclonal antibody d-41 (Springer and Barondes, 1983) and alkaline phosphatase-coupled goat anti-mouse IgG antibodies. Reactive clones were recognized by hydrolysis of 4-umbelliferyl phosphate. Nonreactive clones were noted and picked from replica plates for retesting. A temperature-sensitive derivative of strain HL260 was isolated from a mutagenized population by replica picking 250 surviving clones to plates that were incubated at 22 or 27°C (Loomis, 1969). Strain HL261 fails to grow at 27°C. Diploids were selected from crosses by plating lo5 cells on plates spread with Bacillus subtilis and incubating at 27°C (Williams and Newell, 1976). Diploids were checked for ploidy by spore size and segregation of recessive markers following growth at permissive temperatures. Immunological analysis. Western blots of proteins separated electrophoretically on one and two-dimensional acrylamide gels were stained with monoclonal antibodies (10 pg in 15 ml) as previously described (Murray et al., 1983). Bound antibody was recognized by rabbit antiserum against mouse kappa light chains and lz51-labeled staphylococcal protein A. Monoclonal antibodies E28D8 (Murray et al., 1983) and d-41 (Springer and Barondes, 1983) have been previously described. Adhesion analysis. Cells that had developed for 13 hr at 22°C were dissociated in 10 mM EDTA, pH 6.4, and gyrated at 200 rpm for 10 or 30 min as indicated in the text before determining the number of remaining single cells with a Coulter counter (Murray et ab, 1983; McDonough et al., 1980). Where indicated, adhesion was evaluated with a phase-contrast microscope. Adhesion blocking by the monoclonal antibodies was determined by incubating dissociated cells in the presence of various concentrations of the antibody for 30 min at 0°C before determining the degree of adhesion; Fab fragments of goat antibodies against mouse IgG (GAMFab) were added after incubation with the monoclonal antibody to block agglutination (Springer and Barondes, 1980). We found that the potency of various preparations of GAMFab against monoclonal antibodies varied greatly so that each preparation had to be titrated in terms of the concentration necessary to produce maximum inhibition. The titer for each class of monoclonal IgG also varied presumably due to the different amounts of antibodies specific for these classes.

VOLUME

109. 1985

Antibody binding. Binding of the monoclonal antibodies to whole cells was determined using 1251-labeled goat anti-mouse IgG Fab and counting the precipitable material (Springer and Barondes, 1983). RESULTS

Isolation

of Mutant

Strain

HL%O

To screen for mutants that lacked the determinant that bound monoclonal antibody d-41, we used the same technique as that previously employed successfully to isolate mutants that failed to bind other monoclonal antibodies (Murray et ah, 1984; Knecht et al., 1984). A thousand mutagenized clones were grown up in multitest wells and replicated to fresh medium, and one set was screened for the presence of the determinants that bound d-41. Five clones were picked as putative mutants that failed to bind the monoclonal antibody and, upon rescreening, strain HL260 was found to lack this determinant (Fig. 1).

g P80’

FIG. 1. Strain HL260 lacks determinants that bind d-41. Aggregating cells of strains HL260 and AX3 were solubilized and the extracts separated by electrophoresis in the presence of SDS. Proteins were transferred to nitrocellulose and immunostained with monoclonal antibody d-41. Bound antibody was detected by fluorography after reaction with rabbit anti-mouse IgG and ‘%I-protein A.

LOOMIS

COMPARISON

ET AL.

TABLE 1 OF EDTA-RESISTANT ADHESIONIN HAPLOIDS AND DIPLOIDS

Strain

Adhesion

modB+

m.od Bf bsg b500 m.od B503 mod B503 (h7)

HL260 HL261 Diploids DL117 Parentals: HL80 and HL261 DL118 Parentals: HL216 and HL261 DLl19 Parentals: HL220 and HL261

mod B503/+ bsg B500/+ +/(tw) mod B50l/mod bsg A5/+ +‘(tsg) mod B502/mod bsg A5/+ +‘(tw)

AND modB

Percentage single cells”

Genotype

Haploids HL80

Mutants

12 + 1 55 f 3 57 f 3

19 + 1

B503

74 f 4

B503

72 -+ 4

a Cell-cell adhesion was measured in cells that had developed for 12 hr on filters. The number of single cells remaining after gyration for 30 min at 200 RPM in the presence of 10 mM EDTA was determined with a Coulter counter. Cells that have developed cellcell adhesion form clumps and fewer remain single. The percentages of single cells are given with their SE.

Strain HL260 Lacks Cell-Cell Adhesion

Dez)elopmentaEly

Regulated

Cell-cell adhesion of strain HL260 was measured by gyration under standard!ized conditions in the presence or absence of 10 mM EDTA and compared with that of strain HL80. When vegetative, both strains showed adhesion in the absence of EDTA and virtually no adhesion in its presence, which is characteristic of undifferentiated D. discoideum (Gerisch, 1961; McDonough et al., 1980). When adhesion was measured in 10 mM EDTA after 12 hr of starvation, at which time both strains were form.ing aggregates, the wild-type strain showed substantial adhesion with only 12% of the cells remaining single. In contrast HL260 showed impaired adhesion in tlhat 55% of the cells failed to become associated and remained as single cells (Table 1). In other experiments in which cells were studied at a variety of times betwieen 8 and 16 hr after onset of starvation, and adhesion was determined after 10 or 30 min of gyration, HL260 showed impaired adhesion. In some instances there was virtually no detectable adhesion (95% single cells) but the data shown in Table 1 are typical and indicate that the loss, though consistent, may not be complete. Similatities

of Monoclonal

Antibodies

d-41 and E28D8

In previous studies, we described a monoclonal antibody, E28D8, that reacts with a periodate-sensitive

Lacking

Saccharide

Determinant

113

determinant on D. discoideum glycoproteins (Murray et ab, 1983). We used this monoclonal antibody to screen a mutagenized population from which we isolated strains HL216 and HL220, which lack this determinant (Murray et aZ., 1984). These mutants have impaired cell-cell adhesion. Given the present findings, we sought to determine the relationship between the determinants recognized by monoclonal antibodies d41 and E28D8. Initial studies showed that mutants HL216 and HL220, which fail to bind monoclonal antibody E28D8, also fail to bind monoclonal antibody d-41. Conversely, the new mutant, HL260, which lacks the determinant recognized by monoclonal antibody d-41, also fails to bind monoclonal antibody E28D8. Furthermore, immunoblots with d-41 or E28D8 of crude extracts taken throughout development of D. discoideum were indistinguishable (Fig. 2). In addition, immunoblots of the acidic region of two-dimensional gels of a crude extract of aggregating cells were also indistinguishable when prepared using either d-41 or E28D8 (Fig. 3). In both instances, the series of spots in this portion of the gel migrate exactly with purified gp80 (Murray et al., 1981), indicating that both these monoclonal antibodies react with a determinant present on various forms of this glycoprotein, but also shared with others. Given the apparent similarities between d-41 and E28D8, we compared their effect on cell-cell adhesion of aggregating D. discoideum and their binding to the surface of these cells. Monoclonal antibody d-41 saturated the cell surface when incubated with the cells at concentrations of about 25 pg/ml, whereas E28D8 required about 10 times as much to reach saturation (Fig. 4). Maximal inhibition of cell-cell adhesion with d-41, in this experiment about 70%, was also achieved at a concentration of about 25 pg/ml, whereas with E28D8 there was little inhibition at 25 pg/ml, but complete inhibition was found at 250 pg/ml (Fig. 4). The higher concentrations of E28D8 required to block cell-cell adhesion may explain the failure of this monoclonal antibody to inhibit cell-cell adhesion in previous experiments (Murray et al., 1983). The concentration of goat anti-mouse Fab used in previous experiments, 250 pg/ml, also was apparently suboptimal since, in the present experiments, 3 mg/ml was needed. The finding that there appear to be only about half as many determinants on the cell surface that bind E28D8 compared with d-41 (Fig. 4) might reflect some difference in the determinant bound by these two monoclonal antibodies. However, since this binding assay actually measures the amount of ‘251-labeled goat anti-mouse Fab bound to the primary monoclonal antibody on the cell surface, it could also reflect differences in the reactivity of the Fab for these two different

114

DEVELOPMENTAL BIOLOGY

VOLUME 109,1985

B 2 4 6

D 41

8 10 1214

1618

20 22

E28D8

FIG. 2. E28D8 and d-41 react with the same bands on Western blots. Cells of strain HL80 that had developed on filters were collected at the times indicated (hours). Extracts of 2 X lo5 cells were electrophoretically separated on 10% polyacrylamide and transferred to nitrocellulose. Proteins were immunostained with monoclonal antibody d41 (A) or with monoclonal antibody E28D8 (B). Bound antibody was detected as in Fig. 1.

immunoglobulins. At saturation, 5.8 X lo6 molecules of lz51-labeled goat anti-mouse Fab bound to aggregating cells reacted with d-41 and 3.0 X lo6 molecules bound to the cells reacted with E28D8. The number of copies of each determinant on the cell surface would be expected to be somewhat less, since more than one goat anti-mouse Fab would be expected to bind to each monoclonal antibody molecule bound to the surface.

D41

Construction of Diploids and Mapping of the Mutations

The mod B mutants were isolated from mutagenized populations of D. discoideum and can be expected to carry inadvertent mutations in other genes, some of which might affect development. However, diploids constructed between haploid mutants isolated from separate populations should carry wild-type alleles for

E28D8

FIG. 3. E28D8 and d-41 bind to gp80. 500 ng of purified gp80 (Murray et al, 1981) was subjected to nonequilibrium electrophoresis toward the acid end (left) followed by SDS-polyacrylamide electrophoresis in the vertical dimension. Protein to nitrocellulose and immunostained with monoclonal antibody d41 or E28D8. Bound antibody was detected as in Fig. 1.

pH gradient was transferred

LOOMIS

ET AL.

Adhesion

Mutants

Lacking

Saccharide

Determinant

115

HL220 and diploids were selected for growth on B. subtilis at 27°C. These diploids and the parental strains were allowed to develop to tight aggregates (12 hr of development) and then analyzed on immunoblots using monoclonal antibody E28D8 or d-41 and for cell-cell adhesion (Table 1). Diploid strain DL117 was formed from a cross of strains HL261 and HL80 and contained the same proteins that reacted with E28D8 or d-41 in the wild-type strain HL80. Since DL117 and HL80 showed similar degrees of adhesion (Table l), the mutation in HL261 is recessive. In contrast, no specific bands reactive with E28D8 or d-41 were found in extracts of strains HL261, DL118 (a diploid formed from strains HL261 and HL216), or DL119 (a diploid formed from strains HL261 and HL220) even afer prolonged exposure, and all these strains showed markedly impaired adhesion in 10 mM EDTA (Table 1). The lack of complementation of HL261 in diploids carrying either mod B501 or mod B502 shows that HL261 carries an independent mutation in the mod B locus, which we now refer to as mod B503. Failure of the diploids to develop both the determinant recognized by the adhesion-blocking monoclonal antibodies and EDTA-resistant cell-cell adhesion shows that they are related. In the absence of EDTA, mutant cells were as adhesive as wild-type cells when examined with a microscope (Fig. 5) or with the usual assay.

2 oo-y

0,5 ANTIBODY

CONCENTRATION

(mg/ml)

FIG. 4. Comparison of E28D8 and d-41 binding to cells and effect on adhesion. Antibodies E28D8 (0) and d-41 (0) bound to D. discoideum cells at 4°C. The cells were washed and the amount of ‘%I-GAMFah which bound to these cells was determined as described previously (Springer and Barondes, 1980). Effect of the two antibodies on cell-cell adhesion in 10 mM EDTA was determined as described previously (Gerisch, 1961).

each of the other mutattions. The availability of the three independent mutant strains, HL260, HL216, and HL220, which fail to bind either d-41 or E28D8 and show impaired cell-cell adhesion, permitted us to evaluate the correlation between these phenomena by the construction of stable diploids between independently isolated mutants. HL216 contains a mutation previously designated mod B501 :and HL220 contains another mutation in the same locus, designated mod B502 (Murray et al., 1984). These strains are haploid derivatives of diploids constructed from the original mod B mutants, and both carry a mutation on linkage group III (bsg A5) that limits their growth on B. subtilis without affecting growth on gram-negative bacteria. For diploid isolation, a derivative of HL260 that is unable to grow at 27°C was first isolated. This tsg strain, HL261, was crossed with strains HL216 and

Phenotype

of mod B Mutants

Although the three mod B mutants all showed impaired cell-cell adhesion when measured by gyrating dissociated cells in suspension in the presence of 10 mM EDTA, they all were capable of aggregating and constructing fruiting bodies that contain viable spores. This raised the question of the biological function of the developmentally regulated adhesion measured in the presence of 10 mM EDTA. This adhesion has received considerable attention because it appears during differentiation of D. discoideum cells as they aggregate (Gerisch, 1961; McDonough et al., 1980). We, therefore, carefully examined the diploid mod B mutants for any evidence of abnormality. Only the mod B locus is mutant in both parental strains of the diploids DL118 and DL119, so the effects of the particular locus on development should be apparent from examination of these strains. Within 8 hr of development on filters, cells of strains DL118 and DL119 started to aggregate. By 12 hr they had formed tipped aggregates. Thus, the timing and gross appearance of early aggregation and the formation of multicellular structures were indistinguishable in mod B+ and mod B mutant strains. When aggregates of these strains were examined microscopically in

116

DEVELOPMENTAL BIOLOGY

WITH

EDTA

VOLUME 109.1985

WITHOUT

EDTA

DL117

FIG. 5. Cell-cell adhesion in the presence or absence of EDTA. Cells of strains DL117 and DL118 were developed to the aggregation stage, dissociated,and suspended in phosphate buffer with (left) or without (right) 10 mM EDTA. The cells were observed with a phase-contrast microscope. The bar indicates 50 pm.

phosphate buffer, clumps of cells of mod B mutant strains were seen to be as resistant to shear (moving the coverslips) as those of wild-type strains. These observations, although crude, suggest that the adhesion that is missing in mod B cells is not essential for holding the cells together in aggregates in viva. Slugs of strains DL118 and DL119 migrated normally, were phototactic, and proceeded through the stages of morphogenesis to give fruiting bodies at the same rate as did the mod B+ strains DL117 or HL80. Culmination initiated at 20 hr of development on filters and was complete by 24 hr. However, the slugs and fruiting bodies formed by strains DL118 and DL119 were about half the size of those formed by strain DL117. The fruiting bodies of the mutants contained about half the number of spores as wild-type fruiting bodies (Table 2). Therefore, despite clear evidence of a marked impairment of cell-cell adhesion in DL118 and DL119 when measured by gyration of dissociated cells in 10 mM EDTA, the only obvious abnormality in the morphogenesis of these mutants under the laboratory conditions in which they were examined is in the size and spore content of their fruiting bodies. We also found that when equal numbers of wild-type and mutant cells were allowed to develop, the number of spores recovered from the mutant was one third that of wild-type.

DISCUSSION

The results show that D. discoideum mutants which lack the saccharide epitope that binds adhesion-blocking monoclonal antibodies are themselves deficient in cellcell adhesion. They therefore support the immunological evidence in the previous paper (Springer and Barondes, 1985) that some component of the saccharide, whose synthesis is dependent on the mod B gene, may directly participate in the adhesive bonds between cells. Since the saccharide epitope is found on several molecules, one or all might be involved in adhesion. However, neither the present nor the previous results conclusively implicate the saccharide in the adhesion process, since alternative interpretations are possible. For example, the protein portion of one of the molecules that carries the saccharide, such as gp80, may actually mediate adhesion but its display in the membrane may be critically dependent on the saccharide. Despite their defective EDTA-resistant cell-cell adhesion, the mod B mutants can aggregate and form fruiting bodies under usual laboratory conditions. This was examined in detail with the diploids DL118 and DL119 constructed from independently isolated mod B haploids and presumed to be homozygous mutant only at the mod B locus. Both diploid strains showed a

LOOMIS

ET AL.

COMPARISON

modB Time of tipped aggregates (hr) Time of culmination (hr) Phototactic migration Size of fruiting bodies height (mm) Spores per sorus Adhesion without EDT.A Adhesion with EDTA

Adhesimz

Mutants

Lacking

Saccharide

TABLE 2 OF THE PROPERTIES OF WILD-TYPE

117

Determinant

AND mod B MUTANTS

HLSO

DL117

DL118

DL119

+ 12 20 + 2.5 + .l 5.5 x lo4 + +

B503/+ 12 20 + 2.6 f .1 3.6 X lo4 + +

B503/B501 12 20 + 1.8 f .05 2.3 X lo4 + -

B503/B502 12 20 + 1.9 + .l 1.8 X lo4 + -

Note. All sori were counted and collected from a small counted in a hemocytometer. Two separate determinations estimated using measured marks on a capillary pipet.

area of a filter on which they had developed. The spores were dispersed varied by less than 10% and were averaged. The height of fruiting bodies

normal morphogenetic sequence but formed smaller fruiting bodies with fewer spores than those of a diploid strain, DL117, heterozygous wild type at the mod B locus. There may be a selective disadvantage to the smaller fruiting bodies, but the dramatic result is that the developmentally acquired adhesion system in which they are impaired is not essential for formation of a multicellular organism. Furthermore, the fact that the mod B mutant cells differentiate to form stalk and spore cells indicates that the impaired adhesion system does not play an indispensible role in the regulation of postaggregation genes. To the extent that cell-cell adhesion is critical for this developmental process, another adhesion system may be sufficient. The phenotype of mod B mutants was characterized under laboratory conditions. Even in these simplified environments fewer spolres were formed and the fruiting bodies were significantly smaller. In the wild there might be other consequences to loss of the mod B gene product, and specific laboratory conditions might be required to detect them. For example, no deleterious effects were initially found in mutants containing very low levels of discoidin I (Alexander et al., 1983), but recently they were shown to be defective in substrate attachment and aggregation under submerged conditions (Springer et al., 1984). Thus, a wide variety of conditions should be considered in assigning the roles of a gene product in development. This work was supported by a grant from the National Science Foundation (PCM79-02698) to William F. Loomis and by grants from the USPHS (HD13542), the McKnight Foundation, and the Veterans Administration Medical Center to S. H. Barondes. We thank Tom Jongens for help in selecting and characterizing strain HL260. REFERENCES ALEXANDER, S., SHINNICK, T. M., and LERNER, R. A. (1983). Mutants of Dictyostelium discoideum blocked in expression of all members of the developmentally regulated discoidin multigene family. Cell 34,467-4’75.

and was

GERISCH, G. (1961). Zellfunktionen and zellfunktion wechsel in der entwichlung van Dictyostelium discoideum. Exp. Cell Res. 25, 535554. KNECHT, D. A., DIMOND, R. L., WHEELER, S., and LOOMIS, W. F. (1984). Antigenic determinants shared by lysosomal proteins of Llictyolstelium discoideum: Characterization using monoclonal antibodies and isolation of mutations affecting the determinant. J. BioL Chem. 259, 10633-10640. LOOMIS, W. F. (1969). Temperature-sensitive mutants of Dictyostelium disco-ideum. J. Bactetiol 97, 1149-1154. LOOMIS, W. F. (1971). Sensitivity of Dictyostelium disctideum to nucleic acid analogues. Exp. Cell Res. 64, 484486. MCDONOUGH, J. P., SPRINGER, W. R., and BARONDES, S. H. (1980). Species specific cell adhesion in cellular slime molds. Exp. Cell Res. 125, l-14. MORRISEY, J. M., DEVINE, K. M., and LOOMIS, W. F. (1984). The timing of cell-type specific differentiation in Dictyostelium discoideum. Dev. BioL 103,414-424. MtiLER, K., GERISCH, G., FROMME, I., MAYER, H., and TSUGITA, A. (1979). A membrane glycoprotein of aggregating Dictyostelium cells with properties of contact sites. Eur. J. B&hem. 99,419-426. MURRAY, B. A., NIMAN, H. L., and LOOMIS, W. F. (1983). Monoclonal antibody recognizing gp80, a membrane glycoprotein implicated in intercellular adhesion of Dictyostelium discoideum. Mol. Cell. BioL 4, 863-870. MURRAY, B. A., WHEELER, S., JONGENS, T., and LOOMIS, W. F. (1984). Mutations affecting a surface glycoprotein, gp80, of Dictyostelium diswideum. MoL Cell BioL 4, 514-519. MURRAY, B. A., YEE, L. D., and LOOMIS, W. F. (1981). Immunological analysis of a glycoprotein (contact sites A) involved in intercellular adhesion of Dictyostelium discoideum. J. SupramoL Stmxt. Cell. Biochem 17, 197-211. SPRINGER, W. R., and BARONDES, S. H. (1980). Cell adhesion molecules: Detection with univalent second antibody. J. Cell BioL 87, 703-707. SPRINGER, W. R., and BARONDES, S. H. (1983). Monoclonal antibodies block cell-cell adhesion Dictyostelium discoideum J. BioL Chem. 258,4698-4701. SPRINGER, W. R., and BARONDES, S. H. (1985). Protein-linked oligosaccharides implicated in cell-cell adhesion in two Dictyostelium species. Dev. BioL 109, 102-110. SPRINGER, W. R., COOPER, D. N. W., and BARONDES, S. H. (1984). Discoidin I is implicated in cell-substratum attachment and ordered cell migration of Dictyostelium discoideum and resembles fibronectin. Cell 39, 557-564. WILLIAMS, K. L., and NEWELL, P. C. (1976). A genetic study of aggregation in the cellular slime mold Dictyostelium discoideum using complementation analysis. Genetics 82, 287-307.