The Cyclic Nucleotide Specificity of Three cAMP Receptors in Dictyostelium

,


$4
To whom correspondence should be addressed.
points in the developmental program of Dictyostelium discoideum.During early aggregation, intermittent stimulation with cAMP coordinates the accumulation of individual amoebae to form organization centers (Devreotes, 1982) and to regulate the expression of various early genes (reviewed by Kessin (1988)).At the mound stage, induction of prespore and the early stages of prestalk gene expression require persistent exposure to micromolar concentrations of cAMP (reviewed by Gerisch (1987)).Cell differentiation into at least four cell types (Williams et d., 1989) results in the final multicellular structure, the fruiting body.
Cell surface cAMP binding sites, which are present throughout the development cycle of Dictyostelium, are most abundant during early aggregation (Schaap and Spek, 1984;  Schenk et al., 1991).At this stage, the cell surface receptors are coupled to G-proteins which, when stimulated, activate a variety of effector enzymes.Ligand stimulation of cAMP receptors initiates a signal transduction cascade to cause increases of second messengers, such as intracellular cAMP and cGMP, and permit cell-cell signaling.In addition, cAMP stimulation causes cytoskeletal changes, such as actin polymerization and myosin phosphorylation, which enable chemotaxis (reviewed in Devreotes (1982)) Van Haastert (1991)).
A cAMP receptor (cAR1)' has been cloned and like other G-protein-coupled receptors found in mammals and yeast, its coding sequence predicts a protein with seven putative transmembrane domains and a cytoplasmic C terminus (Klein et al., 1988).Recently, three additional cAMP receptors (cAR2, cAR3, and cAR4) have been cloned and sequenced.Members of this family of receptor subtypes share approximately 60% identity within their transmembrane and loop regions but have distinct C-terminal domains (Saxe et al., 1991a(Saxe et al., , 1991b)).
The developmental regulation of the major mRNA of each CAR is unique, but there is some overlap between each.CAR1 expression is low during growth, peaks during early aggregation, and then subsides (Klein et aL, 1987).cAR2 mRNA, which is enriched in prestalk cells, is expressed after 15 h of development while cAR3 mRNA is detected earlier at approximately 10 h of development (Saxe et al., 1991a).Cells which lack CAR1 as a consequence of antisense RNA expression (Klein et al., 1988;Sun et al., 1990) or gene disruption (Sun and Devreotes, 1991) do not enter the developmental program and remain as individual amoebae.cAMP derivatives have been used to determine the analog specificity of cell surface cAMP receptors (Van Haastert and Kien, 1983) and other CAMP-binding proteins in Dictyoste- lium (De Wit et al., 1982; Van Ments-Cohen and Van Haastert, 1989).These studies have shown that cAMP binds to surface receptors in aggregation stage cells in a manner dis- tinct from that of intracellular cAMP dependent protein kinase and cell surface phosphodiesterase.The pharmacological specificity for chemotaxis (Van Haastert, 1983), activation of guanylate (Van Haastert and Kien, 1983) and adenylate cyclase (Theibert et al., 1986), induction of gene expression (Oyama and Blumberg, 1986; Haribabu and Dottin, 1986;  Gomer et al., 1986), and cell-type differentiation (Schaap and Van Driel, 1985) have all been demonstrated to match that for surface cAMP receptors.
The presence of multiple CAR subtypes during development suggests that different cAMP receptors mediate separate physiological responses or signal transduction pathways.The affinity and cyclic nucleotide specificity of each CAR subtype may help to distinguish those functions that each receptor controls.Interactions between cAMP and each CAR subtype may vary and thereby provide insight into how the ligand is oriented in each binding pocket.It also may be possible to identify cAMP analogs which specifically activate or block one receptor subtype.
We have expressed three cAMP receptor subtypes, cAR1, cAR2, and cAR3, in growing Dictyostelium cells and examined their biochemical and pharmacological properties.Since there are few endogenous receptors present during growth, each individually expressed CAR can be examined without interference from other receptor subtypes.Cells expressing CAR1 during growth have been described previously and found to have similiar biochemical characteristics to the endogenous receptors in aggregation stage cells (Johnson et al., 1991).In this paper, we demonstrate that cAR1, cAR2, and cAR3 represent a group of similar CAMP-binding proteins which have subtle differences in their interaction with CAMP.Furthermore, it now should be possible to distinguish each receptor subtype during development on the basis of its relative cyclic nucleotide specificity.

EXPERIMENTAL PROCEDURES
Materials-The names and structures of the cAMP derivatives are shown in Fig. 1  ); w-0-CAMP, 3'-NH-cAMP, 5'-NH-cAMP, cBIMP, and PuRMP were synthesized; the synthesis of these analogs has been described previously (Jastorff and Freist, 1974; Morr et al., 1974;  Murayama et al., 1971; Yagura et al., 1980, Baraniak et al., 1979)  Conditions for Growth and Development-AX-3 cells were maintained in HL-5 (Watts and Ashworth, 1970) in shaking culture.Transformants were maintained on Petri dishes in HL-5 with 20 pg/ ml G418 but transferred to shaking cultures for experiments.All cells were harvested during late log phase growth and washed once in 10 mM KH2P04/Na2HP04 buffer (PB), pH 6.5.For development, AX-3 cells were shaken in PB for 4 h at 2 X 107 cells/ml as described (Devreotes et al., 1987).
Construction and Transformation of Expression Vectors-The creation of cells overexpressing CAR1 has been described (Johnson et al., 1991).A full-length cAR2 clone was isolated from a sheared, sizeselected (2-5-kb range) genomic Dictyostelium library (Lambda Zap, Stratagene; gift of Dr. H. Innis).'This 2-kb clone contains 158 bp of 5'-and about 900 bp of 3"untranslated sequence and was shuttled into the EcoRI site of Bluescript KS+ (Stratagene).A full-length cAR3 clone, GR-6, was isolated from a partial Sau3A Dictyostelium genomic library (PAT plasmid, gift of Dr. R. Firtel).3This 1.7-kb clone, which contains 35 bp of 5'-and 40 bp of 3"untranslated sequence, was isolated from the parent plasmid by digesting with XbaI and SmaI.The inserts of cAR2 and cAR3 were filled in with Klenow, BamHI linkers added, and cloned into the BamHI site of separate Bluescript vectors.cAR2 or cAR3 were then cloned into the expression construct, pB18 (gift of Dr. R. Firtel), by digesting with BamHI and ligating them into the BglII site of pB18 in the sense orientation.These vectors or the parent construct, pB18, were transformed into AX-3 cells by electroporation as described (Dynes and Firtel, 1989).Stable transformants were selected by resistance to 10 or 20 pg/ml G418 in HL-5.Total transformants (cAR3) or clones (cAR2) were examined for CAR expression by their ability to bind [3H]cAMP.CAR1 A208 (thymidine auxotroph) cells (Sun and Devreotes, 1991) were cotransformed with cAR2 or cAR3 expression plasmid with pGEM 26-6 (gift of R. Firtel) in a 3:l (pg:pg) ratio and selected for growth in unsupplemented HL-5 media.Transformant clones were screened by immunoblot.Cells expressing high levels of cAMP binding sites were used for further experiments.
cAMP Binding Assays-CAMP binding was performed in the absence and presence of ammonium sulfate (AS) as described (Van Haastert, 1985a).In brief, 8 X lo6 cells were added to PB containing 10 mM dithiothreitol, 10 nM [3H]cAMP, and various concentrations of cAMP or cyclic nucleotide analog in a 100-pl volume at 0 "C.Cells were incubated 1 min and then centrifuged for 2 min at 10,000 X g.To determine binding in AS, 850 pl of 3 M AS was included in the above assay, and after adding cells, 50 p1 of 10 mg/ml bovine serum albumin was added.Cells were incubated 5-7 min and then centrifuged for 3 min.For both assays, the supernatants were carefully aspirated and the cells resuspended in 80 pl of 0.1 M formic acid.One ml of scintillation fluid (Emulsifier, Packard) was then added and radioactivity determined.Nonspecific binding was determined by adding excess cAMP to the incubation mixture at a final concentration of 1 mM (PB) or 0.1 mM (AS).Scatchard binding curves were best fit using the computer modeling programs, LIGAND (Munson and Rodbard, 1980) and Pfit (Elsevier).For analog studies, 3 different concentrations centering around the IC6,, of each analog was used with data points taken in duplicate.The ICso of each analog was tested in two to three independent experiments.Correlation matrix values were obtained using linear regression analysis.
Zmmunoblotting-Membranes were prepared by solubilizing 1 volume of cells with 9 volumes of a lysis buffer containing 1.5% CHAPS and pelleting at 10,000 X g for 20 min (Klein et al., 1985).The pellet was resuspended with 10 volumes of lysis buffer without CHAPS and centrifuged as above.This pellet was suspended in Laemmli's sample buffer (Laemmli, 1970), and 50 p1 of sample was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotted as described (Klein et al., 1987).The blot was probed with a polyclonal antiserum (1:lOOO) raised against the peptide, KREPE-PERFEKYC," a sequence found in the cytoplasmic loop between the putative transmembrane domains 111 and IV of all three CARS (Klein et aL, 1988).2,3

FIG. 2. Expression of CAR proteins in Dictyostelium cells.
Growing cells transformed with either B18, cAR1, cAR2, or cAR3 expression vectors (lanes 1-4, respectively) were examined for protein expression by immunoblotting with a CAR common antiserum.CAR1 migrates at 40 kDa, cAR2 at 39 kDa, and cAR3 at 62 kDa.

RESULTS
Expression and Affinity of CAMP Receptor Subtypes-Each of the three cAMP receptor subtypes, cAR1, cAR2, and cAR3, were expressed in growing Dictyostelium cells.Since at this stage, cells express only a low number of cAMP binding sites (Klein et al., 1987), each individual receptor can be expressed and studied without interference from the endogenous receptors.The expression construct chosen utilizes the actin 15 promoter which is constitutively active during growth and early development (Knecht et al., 1986).Cells overexpressing CAR1 (denoted CAR1 cells) have been previously characterized (Johnson et al., 1991).cAR2 and cAR3 expression constructs were created in a similar fashion and transformed into AX-3 cells.Transformants were selected and screened for their ability to bind [3H]cAMP.One clone (cAR2 cells) or mass culture (cAR3 cells) expressing high levels of cAMP binding sites were examined further.
The presence of each exogenously expressed CAR in growth stage transformants was verified by an immunoblot (Fig. 2).Membranes were prepared from whole cells and immunoblotted with a polyclonal antiserum developed against a common peptide sequence present in all CARS.^Cells transformed with either cAR1, cAR2, or cAR3 (lanes 2-4) had an apparent molecular mass of 40, 39, and 62 kDa, respectively.Each of the CAR cells expressed a similar amount of their respective receptor protein, while control cells ( l a n e 1 ) transformed with the parent vector, expressed very low levels of endogenous CAR1 protein and undetectable levels of cAR2 and cAR3.Receptor affinity was determined by the binding of [3H]cAMP to cells in phosphate buffer in the presence of increasing amounts of CAMP.The units for bound/free (y-axis) and bound (x-axis) are nM/ sites/cell X 1000 and sites/cell X lo5, respectively.See Table I1 for binding parameters.
Bands present at 45 and 29 kDa are probably nonspecific proteins and unrelated to the transformed plasmids since they appear in each of the cell lines including the vector control, which contains a low number of cAMP binding sites (Klein et al., 1988).
We determined the affinity and number of cAMP binding sites for each of the CAR cells.[3H]cAMP binding to cells under physiological conditions (phosphate buffer, PB) was determined and the data analyzed by Scatchard plots (Fig. 3 and Table 11).Growing CAR1 cells and developed vector control cells each have two binding sites of similar affinities of approximately 30 and 300 nM (Johnson et al., 1991).CAR1 cells expressed over 3 x lo5 sites/cell, which is about 30-fold higher than growing and 4-fold higher than developed B18 cells.cAR3 cells displayed two binding affinities which were lower than cAR1.Of the approximately 2 x 10' sites/cell expressed, the majority constituted a low affinity site of about 500 nM while about 5% were of 20 nM.Surprisingly, little cAMP binding in PB was detected in two independent cAR2 transformants under physiological conditions.The low level of cAMP binding detected did not differ significantly from that of control vector cells, which contained some binding sites derived from endogenous CAR1 expression.10 mM Ca2+, which has been shown to reveal additional binding sites in aggregation state cells (Juliani and Klein, 1977;Van Haastert, 1985b), did not increase the level of cAMP binding in the cAR2 cells.In addition, filter lysis of cellular membranes did not expose additional cAMP binding sites (data not shown).
The cAMP binding sites in cAR2 cells and the high affinity sites in the cAR3 cells could not be attributed to the endogenous CAR1 present during growth.To demonstrate this, a CAR1 null cell line (designated A208, Sun and Devreotes (1991)), was transformed with the cAR2 or cAR3 expression construct?Transformants were screened for protein expression by immunoblot, and clones which expressed high levels of cAR2 or cAR3 were then assayed for cAMP binding.During growth, A208 cells have no detectable cAMP binding (Sun and Devreotes, 1991).Since the A208 cells transformed with cAR2 (denoted A208/cAR2 cells) showed a small amount of cAMP binding (490 f 370 sites/cell at 10 nM [3H]cAMP), these binding sites must be attributed to the presence of cAR2.The CAR1 null cells expressing cAR3 (denoted A208/ cAR3 cells) had binding characteristics similiar to that of the cAR3 cells with the wild-type background (Table 11).Scatchard analysis of these cells demonstrate that both the high and low affinity components result from cAR3 expression (Fig. 3).
Previous experiments in developed wild-type cells have shown that 3 M AS increases both the affinity and number of detectable cAMP binding sites (Janssens and Van Driel, 1984;Van Haastert, 1985a).The effect of AS on cAMP binding was tested on all three CAR cells and was found to vary for each (Fig. 4 and Table 11).We have shown previously that AS enhances receptor affinity in both growing CAR1 and developed B18 cells by over 30-fold (Kd = 4 nM) and increases the number of binding sites (Johnson et al., 1991).For the cAR3 or the A208/cAR3 cells, ammonium sulfate enhanced both high and low affinity binding by different extents and increased the number of cAMP binding sites.The high affinity sites increased about &fold to 4 nM, while the low affinity sites increased by approximately %fold to 200 nM.In the ' R. Johnson, R. Gundersen, J. Milne, S. Turgendreich, and P.
Devreotes, manuscript in preparation.cAR2 cells, AS greatly increased the levels of cAMP binding by exposing over 2 x lo5 binding sites/cell of a single affinity of 11 nM.
CAMP Derivatives-The 14 cAMP analogs used in this study test various interactions between the ligand and receptor such as hydrogen and ionic bonding and hydrophobic interactions (Fig. 1 and Table I).W-0-CAMP, 6-Cl-PuRMP, 7-CH-cAMP, 2'-H-cAMP, 3'-NH-cAMP, and 5"NH-cAMP have modifications which prevent potential hydrogen bonds.To test for stereoselective interactions in the phosphate moiety, (S,)-CAMPS and (RJ-CAMPS replace the exocyclic oxygens (axial or equatorial, respectively) with a negatively charged sulfur atom (Frey and Sammons, 1985).In solution, cAMP equally favors either a syn-or anti-conformation (Hemmes et al., 1976).Since 8-Br-CAMP exists primarily in a syn-conformation (Schweizer and Robins, 1973), one may infer the conformation of cAMP when bound to the receptor from its relative affinity.Finally, derivatives cBIMP, PuRMP, cIMP, and cGMP differ in their degree of polarity (cIMP > cGMP > PuRMP > cAMP > cBIMP) (Van Haastert et al., 1983).
Cyclic Nucleotide Specificity of CAR Subtypes-Each of the CAR cells was tested in ammonium sulfate for their ability to bind 14 cAMP derivatives.
values were determined by measuring the concentration of derivative which inhibited 50% of [3H]cAMP binding to the receptors.The data are presented as KO, derivative/K0, cAMP ratios (Table 111) and 6AG values (Table IV).6AG values are derived from the following equation to compare these results to previous studies (Jastorff et al., 1979).
6AG values are expressed in kJ/mol and represent the derivative's reduction of binding energy when compared with the binding of CAMP.
The interactions of all three CARS with cAMP share some common features which have been previously noted in studies on the endogenous receptors in developed wild-type cells (Van Haastert and Kien, 1983).The low affinity of 6-Cl-PuRMP and 3"NH-cAMP indicates that hydrogen bonds are formed between the receptor and cAMP at the 03' position in the ribose ring and the N6 position in the adenine moiety in all three CARS.In addition, since 8-Br-CAMP is primarily in the syn-conformation (Schweizer and Robins, 1973), the greatly reduced affinity of this derivative suggests that cAMP is in an anti-conformation when bound to the receptors.However, the poor affinity of this analog may result from the bromine's effect on the electron distribution in the purine ring or steric hindrance as well.
Receptor interactions varied at the exocyclic oxygens in the Receptor affinity was determined by the binding of [3H]cAMP to cells in 3 M ammonium sulfate in the presence of increasing amounts of CAMP.The units for bound/free (y-axis) and bound (x-axis) are nM/sites/cell x 1000 and sites/cell X lo', respectively.See Table I1 for binding parameters.

TABLE I11
Specificity of CAR subtypes in ammonium sulfate and phosphate buffer K,, derivative/Kw cAMP ratios for WT-D were derived from Van Haastert and Kien (1983).Ratios for CARS were determined as described under "Experimental Procedures." e Value differs from that of Van Haastert and Kien (1983).Data from one experiment.
phosphate moiety of CAMP.CAR1 bound (S,)-CAMPS and derivatives relative to cAMP suggests that there are important (R,)-CAMPS, which replace an axial or equatorial oxygen interactions at both exocyclic oxygens.In contrast, both cAR2 respectively with a sulfur atom, with approximately equal and cAR3 bound (S,)-CAMPS with 6-and 30-fold higher affinity.The loss of 8-10 kJ/mol binding energy for these two affinity, respectively, than cAR1, and both bound (R,)cAMPS See Fig. 1 and Table I.
e 6AG value differs from that of Van Haastert and Kien (1983).
Data from one experiment.
with affinities similar to that of CAR1 (Table 111).In the A208/cAR3 cells, where the low levels of CAR1 were absent, cAR3 bound (S,)-CAMPS as well as or better than cAMP (Fig. 5).This suggests that both cAR2 and cAR3 lack a stereoselective interaction at the axial exocyclic oxygen that is present in cAR1.The nature of this interaction is probably not ionic, since a loss of about 25 kJ/mol binding energy would be expected.Steric disruption is a more likely explanation since a thio-substitution of an exocyclic oxygen would occupy more space.
The hydrophobic cleft which binds the adenine ring (Van Haastert and Kien, 1983)  three receptors.Fig. 6 plots the relative binding energy of five derivatives (cIMP,cGMP,PuRMP,cBIMP) as a function of their relative polarity in comparison with CAMP.As shown previously in developed wild-type cells, the polarity of these derivatives is negatively correlated with binding energy.In addition, these compounds are missing the N6 amino group in the adenine ring.The loss of this amino group raises the binding energy by about 15 kJ/mol relative to CAMP.When this energy increment is subtracted away, the binding energy of W-0-CAMP, which has the N6 amino group but is very polar, fits this correlation well.Hence the adenine ring is thought to rest in a hydrophobic pocket.
All three CARS bound these analogs similarly in that the loss of the N6 amino group contributed an increase of about 15 kJ/mol in binding energy, However, the nature of the hydrophobic cleft differed for each CAR as reflected in the slope of the lines.Both CAR1 and cAR3 have large negative slopes (-1.78 and -1.96, respectively), whereas cAR2 has a slope that is 3 times smaller (-0.622).These data suggest that the adenine moiety is bound in a cleft of the receptor which is more hydrophobic for CAR1 and cAR3 than for cAR2.The loss in hydrophobicity in cAR2 may be caused by changes of amino acid residues in the cleft from a nonpolar to polar nature.
The analog specificity of CAR1 and cAR3 was also examined in phosphate buffer (Tables I11 and IV).cAR2 cells were not included in these studies because of the low number of cAMP binding sites detected in PB.In comparision with specificity studies performed in ammonium sulfate, both CAR1 and cAR3 maintain the general order of analog specificity in phosphate buffer with two exceptions.7-CH-CAMP showed increased affinity, while cBIMP had reduced affinity relative to the other cAMP analogs.Interestingly, the relative binding affinities of the derivatives were not influenced by ammonium sulfate for CAR1 but were enhanced for cAR3.In addition, while adenine ring polarity negatively correlated with binding affinity, the slopes were less steep (data not shown).W-O-CAMP, however, does not fit this correlation as well; its high polarity cannot account for all of the loss in binding energy.
Some analogs were also tested on the A208/cAR3 cells which lack CAR1 (Tables I11 and IV).Since cAR3 is a low affinity receptor, the low levels of the endogenous, higher

FIG.
6.The dependence of binding affinity on the polarity of some cAMP derivatives varies between CAR subtypa.For each graph, filled squares represent binding data in ammonium sulfate from Table IV for N-0-CAMP (2), cIMP (13), cGMP (14), PuRMP (12), cAMP (I), 6-C1-PuRMP (3), cBIMP (11) in order from left to right.Linear regression analysis of the top solid line shows the strong influence of polarity on binding affinity for CAR1 and cAR3 (slope = -1.67,r = -0 .986and slope = -2.17,r = -0.95,respectively) but less so for cAR2 (slope = -0.652,r = -0.86).These derivatives are all missing the N6 amino group which contributes about 16, 15.3, and 14 kJ/mol, respectively, to the binding energy of cAR1, cAR2, and CAM.Subtraction of this binding increment yields the open squares.Linear regression analysis of the bottom dashed line (which includes N -O -CAMP) gives a similar slope for each CAR (cAR1, slope = -1.78,r = -0.988;cAR2, slope = -0.622,r = -0.887;CAM, slope = 1.96, r = -0.946).In AS, the relative nucleotide specificity of cAR3 appears similiar in both sets of cells, but because the ICso for cAMP is larger in the A208/cAR3 cells, the relative bAG values are smaller.These same analogs were tested on the A208/cAR3 cells in phosphate buffer and gave similiar results to the cAR3 cells (data not shown).

Comparison of cARs with Other CAMP-binding Proteins-
The analog specificities of the three cARs were compared with each other, the endogenous receptors in aggregation stage cells (WT-D) and two CAMP-binding proteins in Dictyostelium: the regulatory subunit of cAMP dependent protein kinase A (CAK) and extracellular phosphodiesterase (ePDE).As shown in Table V, the pharmacological specificity of the three CAR subtypes were highly correlated and formed a group of CAMP-binding proteins.Similar to previous results (Van Ments-Cohen and Van Haastert, 1989), cAMP receptors were distinct from other CAMP-binding proteins in Dictyostelium since there was no correlation with either CAK or ePDE.As expected, CAR1 was most similar to the endogenous receptors in developed wild-type cells.The developmental expression of CAR1 protein correlates with the increase of cAMP binding sites during early development and the major band photoaffinity labeled with 32P-8-N3-cAMP at 6 h of development is CAR1 (Klein et al., 1987).However, (S,)-CAMPS was of higher affinity in CAR1 cells than in developed wild-type cells.In addition, our studies showed 8-Br-CAMP to have 3.3 kJ/mol less binding energy in developed wild-type cells than previously reported by Van Haastert and Kien (1983) (Table IV).

DISCUSSION
We have examined the affinity and cyclic nucleotide specificity of three cAMP receptor subtypes by expressing each individually in growing Dictyostelium cells.Each receptor has a different affinity for cAMP in phosphate buffer and the binding parameters of each is uniquely influenced by ammonium sulfate.The cyclic nucleotide specificity indicates that all three receptors comprise a family of CAMP-binding proteins, but each CAR interacts with cAMP in a slightly different manner.
The affinity of the three CAR subtypes in phosphate buffer vary greatly.CAR1 affinity is similiar to the endogenous receptors in developed AX-3 cells (Johnson et al., 1991).The majority of cAR3 binding, however, has an affinity of approximately 500 nM that is about 2-fold lower than cAR1.cAMP binding sites of low affinity, termed C sites, have been detected in wild-type NC-4 cells (Van Ments-Cohen et al., 1991).The exposure of aggregation competent cells to UM levels of cAMP for several hours depletes the cells of CAR1 mRNA

FIG. 4 .
FIG. 4. Scatchard analysis of CAR cells in ammonium sulfate.Receptor affinity was determined by the binding of [3H]cAMP to cells in 3 M ammonium sulfate in the presence of increasing amounts of CAMP.The units for bound/free (y-axis) and bound (x-axis) are nM/sites/cell x 1000 and sites/cell X lo', respectively.See TableI1for binding parameters.
of derivatives of CAMP binding proteins in Dictyostelium bAG values for WT-D, CAK, and ePDE were derived from VanMents-Cohen and Van Haastert (1989). in growing wild-type cells may affect the measurement of the true nucleotide specificity of cAR3.

TABLE I1 CAR binding pclrameters in phosphate buffer and ammonium sulfate
Values for B18 and CAR1 were retrieved from Johnson et al.(1991).Values for cAR2 and cAR3 were determined as described in "Experimental Procedures."

TABLE IV Specificity of CAR subtypes in ammnium sulfate and phosphate buffer 6AG values
Haastert and Kien (1983)re derived from VanHaastert and Kien (1983).8AG values for CARS were determined as described in "ExDerimental Procedures."