Eur. J. Biochem. 210, 785-791 (1992)
0FEBS 1992
Purine-cytosine permease of Saccharomyces cerevisiae Effect of external pH on nucleobase uptake and binding Daniel BRETHES, Christian NAPIAS, Elisabeth TORCHUT and Jean CHEVALLIER Institut de Biochimie Cellulaire du Centre National de la Recherche Scientifique (ERS 0022), Bordeaux, France (Received July 31/September 30, 1992) - EJB 921109
The cloned FCY2 gene (strain pAB4) of the purine-cytosine permease (PCP) of Saccharomyces cerevisiae and the cloned allele fcy2-21 (strain pAB25) introduced into an S. cerevisiae strain carrying a chromosomal deletion at the FCY2 locus [Weber, E., Rodriguez, C., Chevallier, M. R. & Jund, R. (1990) Mol. Microbiol. 4, 585-5961 were studied. The influence of external pH (varying over 3.5-6) has been analysed on the uptake of adenine, hypoxanthine and cytosine (KtaPp,apparent Michaelis constant and V,) and on the binding constants of these three solutes (Kd,,,, apparent half-saturation constant and B,,,, total binding sites) determined on plasma membranes. For pAB4, the variations of Ktm,, and V , were the same for the three bases, i.e. an increase in Kt,,, when the pH increased and a maximum V , around pH 5. For pAB25, Kt,,, values varied in the same way and were significantly higher for the three bases than those found in pAB4. There was almost no variation of V , for adenine, and there was a continuous decrease when the pH increased in the V, of hypoxanthine and cytosine. Equilibrium binding measurements were performed for the three bases with plasma membrane isolated from pAB4 and pAB25. One single class of binding sites was detected. For pAB4, the affinity increased when the pH decreased for the three bases. The affinity of PCP for adenine was always greater than for cytosine or hypoxanthine. For pAB25, the same phenomenom was observed. However, the curves showing the variation of Kdapp, as a function of pH were shifted towards more acidic pH values. A model was used to fit the experimental binding data obtained with hypoxanthine for the calculation of the dissociation constants of its binding to PCP and to determine the ionization constants of an amino acid involved in ligand binding. For pAB4, at acid pH, the dissociation constant was 1.7+0.4 pM. An amino acid displaying a pK of 3.8 was determined; this value was shifted to pK 4.8 when hypoxanthirie was bound. For pAB25, the main effects of the mutation were a large decrease in the affinity of PCP for hypoxanthine (Kd of 14.4k4.3 pM) and a shift in the pK of the amino acid towards a more acidic pH (about 2.9). The pK of this group remained similar to the value obtained with pAB4 when hypoxanthine was bound. From these data, it is proposed that the binding of hypoxanthine and H f is a random process.
In the yeast Saccharomyces cerevisiae, the purine-cytosine permease (PCP) mediates cotransport through the plasma membrane of proton and purine bases (adenine, hypoxanthine and guanine) or a pyrimidine base (cytosine) [l-31. The utilization of a transmembrane electrochemical-potential difference in the proton leads to accumulation of bases in the cytoplasm [4 - 61. In an attempt to study this secondary activetransport mechanism, a pH-dependence kinetic study of the Correspondence to J. Chevallier, Institut de Biochimie Cellulaire du Centre National de la Recherche Scientifique, 1 rue Camille SaintSaens, F-33077 Bordeaux Cedex, France Abbreviations. PCP, purine-cytosine permease; Ktapp,apparent Michaelis constant of transport; Kda,,, apparent half-saturation constant of solute binding; B,,,, maximal amount of binding sites; PMF, plasma membrane fraction; PMF-pAB4, plasma membrane fraction isolated from pAB4 strain; PMF-pAB25, plasma membrane fraction isolated from pAB25 strain.
uptake of nucleobases into whole cells of a wild-type strain of S. cerevisiae showed that the base-transport activity parameters (K, and V,) were dependent on the concentration of the cosubstrate H [7]. In addition, the apparent affinity for base accumulation increased in the presence of higher concentrations of proton. Whether the external proton concentration has an effect on the binding of the nucleobase to the permease, the first step of the translocation process, remains to be established. Comparison between data obtained by kinetic measurements in vivo and those obtained by equilibrium-bindingexperiment measurements in vitro might allow better understanding of the translocation mechanism [8]. The question of whether or not the PCP mechanism is an ordered binding model can be studied by measuring the binding properties of the nucleobases as a function of proton concentration. This approach was made possible since the FCY2 gene encoding the PCP has been cloned into a pJDB207 plasmid, leading to +
786 the amplification of gene expression and an increase of the amount of functional PCP in the plasma membrane [9, 101. Various strains of PCP mutants have been obtained which display modifications in their uptake activity [ l l , 121. One has been cloned and incorporated in a multicopy plasmid allowing in-vitro studies of a mutated form of PCP [13]. In this paper, we analyze the variation of the transport activity parameters as a function of external pH in vivo on a strain carrying a multicopy plasmid encoding the permease gene. We also study the equilibrium binding parameter variations in vitro on plasma membrane-enriched fractions (PMF). These parameters were determined for hypoxanthine, adenine and cytosine. In addition, with the only strain carrying a multicopy plasmid encoding for a mutated permease gene available at that time [13], we describe the influence of mutations on the behaviour of the permease.
A
3
5
4
6
PH I
I
EXPERIMENTAL PROCEDURES Materials Adenine, hypoxanthine and cytosine were purchased from Aldrich. [U-14C]Adenine (9.99 GBq . mmol- '), [S-l"C] hypoxanthine (1.85 GBq . mmol- ') were purchased from Amersham. [2-'4C]Cytosine (1.9 GBq . mmol-') was purchased from Moravek. Yeast nitrogen base without amino acid was from Difco. Scintillation cocktail was Ready Value from Beckman. All other reagents were of analytical grade from Merck. Strains Isolation and characterization of the strains used in this work have already been published [lo]. Strain NC233lOB(pAB4) is a permease-proficient strain carrying a multicopy plasmid encoding the FCY2 gene. The mutated strain NC233-l0B(pAB25) is a permease-proficient strain carrying a multicopy plasmid encoding a mutated FCY2 gene. In this mutated strain, four amino acid residues have been replaced (Met192, Ile371 and Ile375 changed to Val; Asn377 changed to Gly) [ll]. The Metl92Val mutation is a silent amino acid replacement [12]. In the text, the NC233-10Btransformed clones are referred to only in terms of the plasmid they harbor, i.e. pAB4 and pAB25. Strains were grown and harvested as described [13].
3
5
4
6
PH Fig. 1. Variation of purine and cytosine uptake parameters as a function of external pH for pAB4 strain. The initial uptake rates were obtained as described under Experimental Procedures. Experiments were performed with concentrations of bases varying over 0.2-75 pM (0) adenine, (0)cytosine, (A) hypoxanthine. 12 different solute concentrations were used to determine the apparent Michaelis constant of uptake (Ktapp;A) and the maximal rate of uptake (V,,,; B). For a non-amplified wild-type strain (FL loo), the rate of cytosine uptake was 1.2kO.1 nmol ' m n - ' . lo7 cells-' and Kta,, 1.8k0.2 pM at pH 5 [15]. The average is for two different sets of experiments.
desired pH containing 0.1 M NaCl. The radioactive ligands adenine (780 MBq . mmol-'), hypoxanthine (680 MBq . Initial rates of uptake (adenine, cytosine and mmol-') and cytosine (620 MBq . mmol-') were included at hypoxanthine) were obtained by a filtration technique after various concentrations (0.07 - 170 pM). Determination of the incubation of cells harvested in the exponential phase (5 - amount of specifically bound ligand was made by measuring 7 x l o 7 cells . ml-l) in 50 mM sodium citrate, at different pH, the difference between the amount of ligand bound in the containing 2% glucose at 30°C [13]. Cell concentrations of absence (total binding) and the presence (non-specific binding) 5 - 7.5 x lo6 cells . ml- were used. Uptakes were started by of an excess of adenine (2mM). B,,, (maximal amount of addition of either [U-14C]adenine (157 MBq . mmol-'), [S- specifically bound ligand) and Kda,,(apparent half-saturation 14C]hypoxanthine (161 MBq . mmol- ') and [2-14C]cytosine constant of solute binding) were calculated by non-linear re(125 MBq . mmol-'). Total solute concentrations in the as- gression analysis of data. It was not possible to measure bindsays varied over 0.2 - 75 pM. Uptake rates were constant up ing at pH values below 3.5 since PMF started to floculate to 10 s of reaction under our experimental conditions [14]. V,,, under these conditions. (maximal rate of uptake) and the apparent Michaelis constant of transport (Ktdp,)were calculated by non-linear regression Miscellaneous analysis of the initial uptake rates versus solute concenThe cell concentration was determined by measuring the trations. Ligand-binding measurements were performed by a turbidity at 550 nm (an absorbance unit corresponding to centrifugation technique [13] in 50 mM sodium citrate at the 2 x lo7 cells . ml-I). Plasma membrane fractions (PMF) were
Activity measurements
'
787 Table 1. Variation of half-saturation constant of purines and cytosine on PMF-pAB4 as a function of external pH. Binding measurements were performed as described in Experimental Procedures at a PMF-pAB4 protein concentration of 100 - 120 pg . ml- ' with the following concentrations of radioactive ligands: [U-14C]adenineover 0.05 - 30 KM,specific radioactivity 780 MBq . mmol- I , [8-'4C]hypoxanthine over 0.12-40 pM, specific radioactivity 680 MBq . mmo1-l and [2-'4C]cytosine over 0.12-30 pM, specific radioactivity 620 MBq . mmol-'. The non-specific binding measured for each ligand was not pH dependent and corresponded to 13 f 1 pmol . mg protein-' for I KM free ligand a t equilibrium. Calculations were made by non-linear regression analysis of the saturation curves [12]. B,,, maximal amount of specifically bound ligand; adenine = 1054- 1390 pmol . mg protein-', hypoxanthine = 915- 1020 pmol . mg protein-', cytosine = 860- 1000 pmol . mg protein-'. Ten different ligand concentrations were used for each determination. The average is for three sets of experiments. n.d., not determined. Ligand
Hypoxanthine Adenine Cytosine
Half-saturation constant at pH 3.6
3.8
4.0
4.2
4.5
4.8
5.0
6.0
n.d. 0.87 k 0.06 4.87f0.87
3.8 f 0 . 3 n.d. 3.6k0.52
3.17 k0.08 n.d. 6.32+0.18
5.2 k0.5 0.42f0.04 n.d.
7.13 k0.8 0.46k0.04 4.38 & 0.73
n.d. 0.37+0.05 5.26 f 0.66
13.9 k 0 . 4 0.6k0.03 n.d.
17.2 k 2 . 6 1.17k0.05 13.2 5 3.7
isolated as described [15], stored at -20°C and used within three weeks. Protein concentration was determined by the method of Lowry [16] with bovine serum albumin as a standard.
RESULTS Effects of external pH on apparent kinetic parameters of uptake for wild strain Accumulation of adenine, hypoxanthine and cytosine by pAB4 strain was studied at pH values of 3.5 - 6.0. lnitial rates were measured after 4, 8 and 10 s of incubation. For pAB4, the variations of K,,,, and VmaXas a function of the external pH were about the same for all bases, i.e. an increase in Kta,, by a factor of 2.5-4.8 over pH 3.5-6 and a slight variation of V,,, with an optimum pH around pH 4.5 and pH 5.5 (Fig. 1). These parameter variations were similar to those already obtained with a wild-type strain [7]. The effect of permease amplification was only an increase in V,,, (by a factor of about 5 at pH 5.5) without any modification in Kt,,, values [15]. Measurement of equilibrium binding constant on plasma membrane isolated from wild strain as a function of external pH The maximal amount of specifically bound ligand was similar for all ligands (B,,, = 1000- 1300 pmol . mg PMF protein-') whatever the pH value. Variations in Kdap,as a function of pH are displayed in Table 1. At all pH values, the affinity of PCP for adenine was better than for hypoxanthine and cytosine and an increase in Kda,, was noticeable for all ligands when the proton concentration was lowered in the external medium. Variations in solute binding as a function of pH for a given concentration of ligand (adenine 0.4 pM, hypoxanthine 2.5 pM and cytosine 1.0 pM) are displayed in Fig. 2. The amount of hypoxanthine bound to PMF decreased continuously when the pH varied over 3.5-6. In contrast, maximum binding was noted for adenine over pH 4.2-4.6 and for cytosine over pH 4.5 - 5. This difference could be explained by considering the pK values of ligands, which are 2 , 8.9 and 12.1 for hypoxanthine, < 1, 4.1 and 9.5 for adenine and 4.5 and 12.2 for cytosine [17]. Thus, it is considered that in the
A 7
a
0
3
5
4
8
PH
Fig. 2. Variation in binding of purine and cytosine on plasma membraneenriched fractions from pAB4 strain as a function of external pH. Specific binding measurements were performed as described in Experimental Procedures and in the legend of Table 1 at a given ligand Concentration. (0)Adenine 0.48 pM; (A)hypoxanthine 2.5 pM; (El) cytosine 1.0 pM. The average is for two sets of experiments.
pH range studied, the variation in KdaP.,for hypoxanthine cannot be ascribed to a change in its ionization. In contrast, this parameter should be taken into account for adenine and cytosine. Variations in uptake and binding parameters as a function of external pH for a mutated strain pAB25
As shown in Fig. 3, the uptake parameters for all bases were modified in Kt,app, which were significantly higher than those determined in the pAB4 strain for cytosine and hypoxanthine (Fig. 1). The variations as a function of external pH were similar, with a continuous increase (Fig. 3A). However, with this strain, the K,,,, variation as a function of pH was significant only with hypoxanthine. The situation was more complex for V,,, since their variations were different with the three solutes, i.e. with adenine almost no variation, but with hypoxanthine and cytosine a continuous decrease
788 100
i
I
I
t
80 h
a
60
M
fi 2 a
v
n d"
Y
40
400 -
Y
a r:
20
-g 200 75
3
5
4
c
4
6
0
m
PH 25 i
I
9
1
0
3
4
5
6
PH
B
Fig. 4. Variation in purine and cytosine binding on plasma membraneenriched fractions from pAB25 strain as a function of external pH. Specific binding measurements were performed as described in Experimental Procedures and in the legend of Fig. 2 at a given ligand concentration. (0)Adenine 2.7 pM; (A)hypoxanthine 55 pM; (0) cytosine 25 pM, concentrations chosen after Kdappdetermination at different pH values as described in Experimental Procedures (not shown). Kdapp= 55.8t-8 pM at pH 3.6 for hypoxanthine; Kdapp= 5.5k0.8 p M a t p H 5.0foradenineandKdapp= 40t-17 p M a t p H 5.0 for cytosine.
>
0
3
5
4
6
cytosine transport. The translocation involves two substrates; the proton and the nucleobase (the latter being more or less Fig. 3. Variation in purine and cytosine uptake parameters as a function protonated). Determination of Kla,, at different pH values of external pH for pAB25 strain. Experimental conditions were similar showed that, when the proton concentration increased (i.e. to those described in Fig. 1 . (A) Apparent Michaelis constant of when the protonation of PCP increased), the Kla,, decreased. uptake (Ktapp).(B) Maximun rate of uptake ( Vmax). (0)Adenine, (0) Owing to the inherent complexity of the uptake mechanism, cytosine and (A)hypoxanthine. The average is for two sets of exper- this result alone cannot account for this process. As already iments. stated [8], there is now a need for the analysis of nucleobase binding at equilibrium on membrane preparations as a function of pH in the bulk phase. In order to explain the results of these binding studies, the following two assumptions have to be made: (a) the proton when pH increased from pH 3.5 to pH 5 (Fig. 3B). Above concentration at the membrane surface is similar to that meapH 5, V,, was almost constant. In addition, it should be sured in the bulk phase; (b) ligand protonation on the memnoted that, with cytosine, a twofold increase in V,, compared brane should be almost identical to that in water. Therefore, to the pAB4 strain was observed above pH 5. with hypoxanthine, for which there is no ionization variation As with the plasma membrane fraction isolated from pAB4 within the range of the external pH studied, only a modifistrain (PMF-PAB~),only one class of specific binding site cation of PCP amino acid residue side-chain protonation may was detected for the plasma membrane fraction isolated from explain the observed changes in Kdapp. pAB25 strain (PMF-pAB25), and the amounts of specifically The binding of H + and base to PCP may occur in three bound solute were similar for all ligands (1000 pmol . mg-' different ways; either two ordered mechanisms (H + first, then . protein-', data not shown). Variations in the amount of base or vice-versa), or a random process with the two ligands specifically bound solute at a given concentration of ligand as bound independently. Our first assumption was that a function of external pH are shown in Fig. 4. As already hypoxanthine and the cotransported proton are bound on observed with PMF-pAB4 (Table 1 and Fig. 2), Kdappin- PCP (assuming that it is a 1 : l stoechiometric process), creased with pH. Compared to PMF-pAB4 (Fig. 2), a shift of through the more general random process shown in scheme the curves towards a more acidic pH was observed for the 1, where K1 and K2 are the dissociation constants of the three ligands. proton binding to free and liganded PCP respectively, and Kd and Kh are the dissociation constants for hypoxanthine binding to protonated and free PCP, respectively. Calculation of DISCUSSION these constants was using the following equations. (1) Lb = Bmax ' [L1/(Kd,,, + [L1>> Varying the extracellular proton concentration modifies the apparent Michaelis constant of adenine, hypoxanthine and Lb being the amount of ligand bound. mg protein-' at a given PH
789 201
L
15
i
'
'
'
'
'
'
'
'
. '
.
I
Kd (PCP)-H
(PCP)-LH
7 L
Scheme 1. Ligand binding to plasma membrane-enrichedfractions. It is assumed that in the pH range considered, only the PCP ionization would be modified and hypoxanthine ionization would be unchanged. K1 and K2 are the dissociation constants of proton binding on PCP and (PCP)-L respectively, and Kd and Kb the dissociation constants for the ligand binding on (PCP)-H and PCP, respectively. Results in Figs 2 and 4 show that (PCP)-L still existed at pH 6. PCP is assumed to be in the form of dispersed material with no sideness in the plasma membrane preparations.
0
E
free ligand concentration [L], B,, the maximum amount of ligand bound and Kda,, the apparent half-saturation constant of binding with
+
Kda,,= Kd(K1 [Hf1)/(K2+ [Hfl) and
(2)
(3) K1.& = K2.Kd. It was shown that B,,, was not pH dependent. The variations in Kda,, for PMF-pAB4 as a function of proton concentration [H'] in the bulk phase are given in Fig. 5A. In this figure, two kinds of data are displayed; those taken from Table 1 and those calculated using Eqn (1) from the data in Fig. 2. The curves Kda,, versus [ H f ] were fitted according to Eqn (2). Values of the calculated constants are displayed in Table 2. With this random model, it is clearly shown that the protonated form of the permease (PCP)H displayed a better affinity for hypoxanthine than the non-protonated form (PCP) (K)d= 1.68 pM and K'd = 17.4 pM). If it is assumed that proton binding occurs before the hypoxanthine-binding process (an ordered mechanism with the proton first), we should obtain no bound ligand when H f approaches 0, and a value of Kda,, approaching infinity. This does not occur since Kda,, can be precisely determined at high pH (Table l), and since the Kda,, curve approaches a constant value while the proton concentration tends to zero (not shown and Table 2). In another set of experiments, detectable amounts of ligand were measured at more alkaline pH (pH 7.5, results not shown). Therefore, this ordered binding process, with the proton binding first, can be ruled out. Fitting of the experimental data is possible, assuming that hypoxanthine is bound before the proton (ordered mechanism, ligand first) and gives fairly decent values for K > (16k1.3 pM) and K2 (33k5.3 pM with pK2 = 4.6f0.1). With this mechanism, we should obtain a Kda,, variation towards 0 when [H'] tends to 0. However, at lower pH the experimental values of Kda,, are systematically higher than those calculated using this model. Therefore, at this point, since we could not perform binding experiments with the membranes at a pH below 3.5 to decide which model was correct, we considered that the random mechanism more accurately indicated the binding of proton and hypoxanthine to plasma membrane fractions.
100 150 200 [H+l (wM)
50
250
300
400
3. v s P
s
200
O
f
0
.
I
50
,
I
100
,
I
150
,
I
200
,
I
250
,
!
300
Fig. 5. Variation in hypoxanthine apparent dissociation constant on PMF-pAB4 and PMF-pAB25 as a function of proton concentration in the hulk phase. Kda,, measured as described in Table 1 ( 0 )or calculated from data displayed in Fig. 2 (for PMF-pAB4) and Fig. 4 (for PMF-pAB25) using Eqn (1) (0)were plotted versus the proton concentration. (A) pAB4, B,, = 1000 pmol . mg protein- '. (B) pAB25, B,, = 980 pmol . mg protein-'. The curves were obtained by fitting Eqn (2) to the experimental data. The calculated dissociation and ionization constants are diplayed in Table 2.
These results also show that an ionizable group with an apparent pK1 of 3.8 was evidenced for the wild-type PCP. The binding of hypoxanthine shifted this pK by about one pH unit towards a more alkaline pH (pK2 = 4.8). The question arises as to how hypoxanthine binding induces such a pK variation of a PCP amino acid side chain. In order to explain this, we assume that the environment of this group was modified by ligand binding. This pK modification can be explained either by a direct interaction between the ligand bound close to (or near) the residue or by an indirect ligand effect through a conformational change modifying the residue environment. The nature of the amino acid involved in this step remains to be elucidated but, in view of the calculated pK, it should correspond to a carboxyl group. This hypothesis is in agreement with the intervention of a carboxyl group of an amino acid side chain for the specific recognition of purine and pyrimidine bases [18]. As yet, it cannot be concluded that this group is the same as that already identified by kinetic studies of hypoxanthine uptake with a wild-type strain [7], although it seems likely.
790 Table 2. Hypoxanthine binding parameters on plasma membrane-enrichedfractions. Dissociation constants KI, K2 and Kd were obtained from the data displayed in Fig. 5A (for PMF-pAB4) and Fig. 5B (for PMF-pAB25) and calculated by fitting Eqn (2) to the data using a nonlinear regression Turbo Pascal program. K &was calculated using Eqn (3). The correlation coefficients of the fits were 0.9933 and 0.9923 for PMFpAB4 and PMF-pAB25, respectively. Since it was not possible to perform binding experiments at low pH (below pH 3.5), K1 values were determined with a very high standard deviation. Strain
Binding parameter
pA B4 pAB25
1.68 f 0.4 14.4 f 4.3
17+ 5 570 f 200
170f 70 850 f 800
The same kind of calculations were performed both for PMF-pAB25 and for PMF-pAB4. Results are presented in Table 2 and in Fig. 5B. As for the pAB4 strain, the random model gives a better description of the proton dependence on the ligand binding. The mutation has two effects on hypoxanthine binding; a dramatic increase in the dissociation constants Kd (10-fold increase) and K &(30-fold increase) and the pK value of the ionizable group of the non-liganded PCP was more acidic than that calculated for the wild-type strain (pK1 around 2.9) as expected from the data in Figs 2 and 4. It is to be noted that the pK of this group remained the same as that calculated for the wild-type strain when hypoxanthine was bound (Table 2, pK2 = 4.7). In summary, this multiple mutation seems to modify both proton and ligand binding to PCP. Since adenine and cytosine display changes in their ionization state in the pH range where measurements were performed, scheme 1 has to be modified in order to analyze Kda,, variation as a function of external proton concentration, and to calculate the true dissociation and ionization constants. Attempts have been made to fit experimental data obtained with adenine and cytosine binding with a model taking into account different ligand ionized forms. These assume that the ligand-binding process was similar for these two ligands compared to hypoxanthine. Our experimental data were not accurate enough to obtain a good fit and to calculate the binding parameters as was performed for hypoxanthine. However, two comments may be made here: (a) the importance of the protonation state of the nucleobase was clearly evidenced since the binding of the protonated form of adenine and cytosine to the permease was apparently less efficient than binding to the unprotonated forms; (b) as observed for hypoxanthine, a shift of the binding curves towards a more acidic pH occurred in PMF-pAB25 when compared to PMFpAB4 (Figs 2 and 4). CONCLUSION The data in this paper demonstrate that the binding of ligands to PCP depends on the external proton concentration. In view of studies dependent on coupling cation concentrations of some transport carriers in Escherichia coli, an ordered binding mechanism has been proposed as the energycoupling mechanism for these secondary transport systems [8, 19, 201. For PCP in yeast, according to our analysis with hypoxanthine, the binding process could be a random mechanism, the proton and base being bound independently. Although the situation seems relatively simple when considering hypoxanthine, the system has a much more complicated
16.3 4 21 f 6.5
3.8 f 0.2 2.9
4.8 f 0.1 4.7 f 0.2
behaviour when the bases are protonated differently, such as adenine and cytosine. Zn-vivo and in-vitro studies of simpler mutants of PCP will allow a more precise approach to the molecular mechanism of this permease. For hypoxanthine, the main effect of mutations in the region with the cluster of amino acids changed at positions 371, 375 and 377 is an increase in dissociation and ionization constants. Therefore, it cannot be stated that this region of the polypeptidic chain is specifically and exclusively involved in the solute recognition process. The occurrence of a possible ‘double-mutational effect’ [21], also modifying the cotransported proton recognition process, should be considered. One of the ways to study this and to understand its precise involvement in the hypoxanthine transport pathway, is to built new strains bearing multicopies of the FCY2 gene mutated on only one amino acid. Such mutant strains are being developed. We gratefully acknowledge Prof. 0.Viratelle (Centre de Recherche Paul Pascal, Bordeaux) for constructive theoretical discussion and Dr M. R. Chevallier (Znstitut de Biologie Moliculuire et Celluluire, Strasbourg) for the gift of the mutant strain and advice. We also acknowledge A. Bocquier for skillful technical assistance and R. Cooke for improving the language.
REFERENCES 1. Polack, A. M. & Grenson, M. (1973) Eur. J . Biochem. 32,276282. 2. Reichert, U. & Winter, M. (1974) Biochim. Biophys. Actu 356, 108- 116. 3. Chevallier, M. R., Jund, R. & Lacroute, F. (1975) J . Bucteriol. 12, 629-641. 4. Reichert, U., Schmidt, R. & ForCt, M. (1975) FEBS Lett. 52, 100- 103. 5. Reichert, U. & ForEt, M. (1977) FEBS Lett. 83, 325-328. 6. Hopkins, P., Chevallier, M. R., Jund, R. & Eddy, A. A. (1978) FEMS Microbiol. Lett. 49, 173- 177. 7. Forit, M., Schmidt, R. & Reichert, U. (1978) Eur. J . Biachenz. 82, 33-43. 8. Yamato, I. (1992) FEBS Lett. 298, 1 - 5. 9. Schmidt, R., Manolson, M. F. & Chevallier, M. R. (1984) Proc. Nut1 Acad. Sci. USA 81,6276-6280. 10. Weber, E., Rodriguez, C. Chevallier, M. R. & Jund, R. (1990) Mol. Microbiol. 4 , 585 - 596. 11. Weber, E., Jund, R., Silve, S., Tsapis, R., Pallares, C., Rodriguez, C. & Chevallier, M. R. (1989) in Highlightsofmodern biochemistry (Kotyk, A,, Skoda, J., Paces, V. & Kostka, V., eds) vol. 1, pp. 761 -770, VPS International, Utrecht. 12. Bloch, J. C., Syhcrova, H., Souciet, J. L., Jund, R. & Chevallier, M. R. (1992) J . Mol. Microbiol., in the press. 13. Brethes, D., Chirio, M. C., Napias, C., Chevallier, M. R., Lavie, J. L. & Chevallier, J. (1992) Eur. J . Biochem. 204, 699-704.
79 1 14. Chirio, M. C. (1990) PhD These Universitt de Bordeaux 2. 15. Chirio, M. C., Brethes, D., Napias, C., Grandier-Vazeille, X., Rakotomanana, F. & Chevallier, J. (1990) Eur. J . Biochem. 196,293 -299. 16. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1961) J. Biol. Chem. 193, 265-275. 17. Dawson, R. M., Elliot, D. C., Elliot, W. H. &Jones, K. M. (1986) in Data for biochemical research, Oxford Science Publications (Clarendon Press).
18. 19. 20. 21.
Htlene, C. (1977) FEBS Lett. 74, 10-13. Yamato, I. & Rosenbush, J. P. (1983) FEBS Lett. 151, 102- 104. Yamato, I. & Anraku, Y. (1989) Biochem. J. 258, 389-396. King, S. C. &Wilson, T. H. (1990) Molecular Microbiol. 4,1433 1438.
0FEBS 1992
Purine-cytosine permease of Saccharomyces cerevisiae Effect of external pH on nucleobase uptake and binding Daniel BRETHES, Christian NAPIAS, Elisabeth TORCHUT and Jean CHEVALLIER Institut de Biochimie Cellulaire du Centre National de la Recherche Scientifique (ERS 0022), Bordeaux, France (Received July 31/September 30, 1992) - EJB 921109
The cloned FCY2 gene (strain pAB4) of the purine-cytosine permease (PCP) of Saccharomyces cerevisiae and the cloned allele fcy2-21 (strain pAB25) introduced into an S. cerevisiae strain carrying a chromosomal deletion at the FCY2 locus [Weber, E., Rodriguez, C., Chevallier, M. R. & Jund, R. (1990) Mol. Microbiol. 4, 585-5961 were studied. The influence of external pH (varying over 3.5-6) has been analysed on the uptake of adenine, hypoxanthine and cytosine (KtaPp,apparent Michaelis constant and V,) and on the binding constants of these three solutes (Kd,,,, apparent half-saturation constant and B,,,, total binding sites) determined on plasma membranes. For pAB4, the variations of Ktm,, and V , were the same for the three bases, i.e. an increase in Kt,,, when the pH increased and a maximum V , around pH 5. For pAB25, Kt,,, values varied in the same way and were significantly higher for the three bases than those found in pAB4. There was almost no variation of V , for adenine, and there was a continuous decrease when the pH increased in the V, of hypoxanthine and cytosine. Equilibrium binding measurements were performed for the three bases with plasma membrane isolated from pAB4 and pAB25. One single class of binding sites was detected. For pAB4, the affinity increased when the pH decreased for the three bases. The affinity of PCP for adenine was always greater than for cytosine or hypoxanthine. For pAB25, the same phenomenom was observed. However, the curves showing the variation of Kdapp, as a function of pH were shifted towards more acidic pH values. A model was used to fit the experimental binding data obtained with hypoxanthine for the calculation of the dissociation constants of its binding to PCP and to determine the ionization constants of an amino acid involved in ligand binding. For pAB4, at acid pH, the dissociation constant was 1.7+0.4 pM. An amino acid displaying a pK of 3.8 was determined; this value was shifted to pK 4.8 when hypoxanthirie was bound. For pAB25, the main effects of the mutation were a large decrease in the affinity of PCP for hypoxanthine (Kd of 14.4k4.3 pM) and a shift in the pK of the amino acid towards a more acidic pH (about 2.9). The pK of this group remained similar to the value obtained with pAB4 when hypoxanthine was bound. From these data, it is proposed that the binding of hypoxanthine and H f is a random process.
In the yeast Saccharomyces cerevisiae, the purine-cytosine permease (PCP) mediates cotransport through the plasma membrane of proton and purine bases (adenine, hypoxanthine and guanine) or a pyrimidine base (cytosine) [l-31. The utilization of a transmembrane electrochemical-potential difference in the proton leads to accumulation of bases in the cytoplasm [4 - 61. In an attempt to study this secondary activetransport mechanism, a pH-dependence kinetic study of the Correspondence to J. Chevallier, Institut de Biochimie Cellulaire du Centre National de la Recherche Scientifique, 1 rue Camille SaintSaens, F-33077 Bordeaux Cedex, France Abbreviations. PCP, purine-cytosine permease; Ktapp,apparent Michaelis constant of transport; Kda,,, apparent half-saturation constant of solute binding; B,,,, maximal amount of binding sites; PMF, plasma membrane fraction; PMF-pAB4, plasma membrane fraction isolated from pAB4 strain; PMF-pAB25, plasma membrane fraction isolated from pAB25 strain.
uptake of nucleobases into whole cells of a wild-type strain of S. cerevisiae showed that the base-transport activity parameters (K, and V,) were dependent on the concentration of the cosubstrate H [7]. In addition, the apparent affinity for base accumulation increased in the presence of higher concentrations of proton. Whether the external proton concentration has an effect on the binding of the nucleobase to the permease, the first step of the translocation process, remains to be established. Comparison between data obtained by kinetic measurements in vivo and those obtained by equilibrium-bindingexperiment measurements in vitro might allow better understanding of the translocation mechanism [8]. The question of whether or not the PCP mechanism is an ordered binding model can be studied by measuring the binding properties of the nucleobases as a function of proton concentration. This approach was made possible since the FCY2 gene encoding the PCP has been cloned into a pJDB207 plasmid, leading to +
786 the amplification of gene expression and an increase of the amount of functional PCP in the plasma membrane [9, 101. Various strains of PCP mutants have been obtained which display modifications in their uptake activity [ l l , 121. One has been cloned and incorporated in a multicopy plasmid allowing in-vitro studies of a mutated form of PCP [13]. In this paper, we analyze the variation of the transport activity parameters as a function of external pH in vivo on a strain carrying a multicopy plasmid encoding the permease gene. We also study the equilibrium binding parameter variations in vitro on plasma membrane-enriched fractions (PMF). These parameters were determined for hypoxanthine, adenine and cytosine. In addition, with the only strain carrying a multicopy plasmid encoding for a mutated permease gene available at that time [13], we describe the influence of mutations on the behaviour of the permease.
A
3
5
4
6
PH I
I
EXPERIMENTAL PROCEDURES Materials Adenine, hypoxanthine and cytosine were purchased from Aldrich. [U-14C]Adenine (9.99 GBq . mmol- '), [S-l"C] hypoxanthine (1.85 GBq . mmol- ') were purchased from Amersham. [2-'4C]Cytosine (1.9 GBq . mmol-') was purchased from Moravek. Yeast nitrogen base without amino acid was from Difco. Scintillation cocktail was Ready Value from Beckman. All other reagents were of analytical grade from Merck. Strains Isolation and characterization of the strains used in this work have already been published [lo]. Strain NC233lOB(pAB4) is a permease-proficient strain carrying a multicopy plasmid encoding the FCY2 gene. The mutated strain NC233-l0B(pAB25) is a permease-proficient strain carrying a multicopy plasmid encoding a mutated FCY2 gene. In this mutated strain, four amino acid residues have been replaced (Met192, Ile371 and Ile375 changed to Val; Asn377 changed to Gly) [ll]. The Metl92Val mutation is a silent amino acid replacement [12]. In the text, the NC233-10Btransformed clones are referred to only in terms of the plasmid they harbor, i.e. pAB4 and pAB25. Strains were grown and harvested as described [13].
3
5
4
6
PH Fig. 1. Variation of purine and cytosine uptake parameters as a function of external pH for pAB4 strain. The initial uptake rates were obtained as described under Experimental Procedures. Experiments were performed with concentrations of bases varying over 0.2-75 pM (0) adenine, (0)cytosine, (A) hypoxanthine. 12 different solute concentrations were used to determine the apparent Michaelis constant of uptake (Ktapp;A) and the maximal rate of uptake (V,,,; B). For a non-amplified wild-type strain (FL loo), the rate of cytosine uptake was 1.2kO.1 nmol ' m n - ' . lo7 cells-' and Kta,, 1.8k0.2 pM at pH 5 [15]. The average is for two different sets of experiments.
desired pH containing 0.1 M NaCl. The radioactive ligands adenine (780 MBq . mmol-'), hypoxanthine (680 MBq . Initial rates of uptake (adenine, cytosine and mmol-') and cytosine (620 MBq . mmol-') were included at hypoxanthine) were obtained by a filtration technique after various concentrations (0.07 - 170 pM). Determination of the incubation of cells harvested in the exponential phase (5 - amount of specifically bound ligand was made by measuring 7 x l o 7 cells . ml-l) in 50 mM sodium citrate, at different pH, the difference between the amount of ligand bound in the containing 2% glucose at 30°C [13]. Cell concentrations of absence (total binding) and the presence (non-specific binding) 5 - 7.5 x lo6 cells . ml- were used. Uptakes were started by of an excess of adenine (2mM). B,,, (maximal amount of addition of either [U-14C]adenine (157 MBq . mmol-'), [S- specifically bound ligand) and Kda,,(apparent half-saturation 14C]hypoxanthine (161 MBq . mmol- ') and [2-14C]cytosine constant of solute binding) were calculated by non-linear re(125 MBq . mmol-'). Total solute concentrations in the as- gression analysis of data. It was not possible to measure bindsays varied over 0.2 - 75 pM. Uptake rates were constant up ing at pH values below 3.5 since PMF started to floculate to 10 s of reaction under our experimental conditions [14]. V,,, under these conditions. (maximal rate of uptake) and the apparent Michaelis constant of transport (Ktdp,)were calculated by non-linear regression Miscellaneous analysis of the initial uptake rates versus solute concenThe cell concentration was determined by measuring the trations. Ligand-binding measurements were performed by a turbidity at 550 nm (an absorbance unit corresponding to centrifugation technique [13] in 50 mM sodium citrate at the 2 x lo7 cells . ml-I). Plasma membrane fractions (PMF) were
Activity measurements
'
787 Table 1. Variation of half-saturation constant of purines and cytosine on PMF-pAB4 as a function of external pH. Binding measurements were performed as described in Experimental Procedures at a PMF-pAB4 protein concentration of 100 - 120 pg . ml- ' with the following concentrations of radioactive ligands: [U-14C]adenineover 0.05 - 30 KM,specific radioactivity 780 MBq . mmol- I , [8-'4C]hypoxanthine over 0.12-40 pM, specific radioactivity 680 MBq . mmo1-l and [2-'4C]cytosine over 0.12-30 pM, specific radioactivity 620 MBq . mmol-'. The non-specific binding measured for each ligand was not pH dependent and corresponded to 13 f 1 pmol . mg protein-' for I KM free ligand a t equilibrium. Calculations were made by non-linear regression analysis of the saturation curves [12]. B,,, maximal amount of specifically bound ligand; adenine = 1054- 1390 pmol . mg protein-', hypoxanthine = 915- 1020 pmol . mg protein-', cytosine = 860- 1000 pmol . mg protein-'. Ten different ligand concentrations were used for each determination. The average is for three sets of experiments. n.d., not determined. Ligand
Hypoxanthine Adenine Cytosine
Half-saturation constant at pH 3.6
3.8
4.0
4.2
4.5
4.8
5.0
6.0
n.d. 0.87 k 0.06 4.87f0.87
3.8 f 0 . 3 n.d. 3.6k0.52
3.17 k0.08 n.d. 6.32+0.18
5.2 k0.5 0.42f0.04 n.d.
7.13 k0.8 0.46k0.04 4.38 & 0.73
n.d. 0.37+0.05 5.26 f 0.66
13.9 k 0 . 4 0.6k0.03 n.d.
17.2 k 2 . 6 1.17k0.05 13.2 5 3.7
isolated as described [15], stored at -20°C and used within three weeks. Protein concentration was determined by the method of Lowry [16] with bovine serum albumin as a standard.
RESULTS Effects of external pH on apparent kinetic parameters of uptake for wild strain Accumulation of adenine, hypoxanthine and cytosine by pAB4 strain was studied at pH values of 3.5 - 6.0. lnitial rates were measured after 4, 8 and 10 s of incubation. For pAB4, the variations of K,,,, and VmaXas a function of the external pH were about the same for all bases, i.e. an increase in Kta,, by a factor of 2.5-4.8 over pH 3.5-6 and a slight variation of V,,, with an optimum pH around pH 4.5 and pH 5.5 (Fig. 1). These parameter variations were similar to those already obtained with a wild-type strain [7]. The effect of permease amplification was only an increase in V,,, (by a factor of about 5 at pH 5.5) without any modification in Kt,,, values [15]. Measurement of equilibrium binding constant on plasma membrane isolated from wild strain as a function of external pH The maximal amount of specifically bound ligand was similar for all ligands (B,,, = 1000- 1300 pmol . mg PMF protein-') whatever the pH value. Variations in Kdap,as a function of pH are displayed in Table 1. At all pH values, the affinity of PCP for adenine was better than for hypoxanthine and cytosine and an increase in Kda,, was noticeable for all ligands when the proton concentration was lowered in the external medium. Variations in solute binding as a function of pH for a given concentration of ligand (adenine 0.4 pM, hypoxanthine 2.5 pM and cytosine 1.0 pM) are displayed in Fig. 2. The amount of hypoxanthine bound to PMF decreased continuously when the pH varied over 3.5-6. In contrast, maximum binding was noted for adenine over pH 4.2-4.6 and for cytosine over pH 4.5 - 5. This difference could be explained by considering the pK values of ligands, which are 2 , 8.9 and 12.1 for hypoxanthine, < 1, 4.1 and 9.5 for adenine and 4.5 and 12.2 for cytosine [17]. Thus, it is considered that in the
A 7
a
0
3
5
4
8
PH
Fig. 2. Variation in binding of purine and cytosine on plasma membraneenriched fractions from pAB4 strain as a function of external pH. Specific binding measurements were performed as described in Experimental Procedures and in the legend of Table 1 at a given ligand Concentration. (0)Adenine 0.48 pM; (A)hypoxanthine 2.5 pM; (El) cytosine 1.0 pM. The average is for two sets of experiments.
pH range studied, the variation in KdaP.,for hypoxanthine cannot be ascribed to a change in its ionization. In contrast, this parameter should be taken into account for adenine and cytosine. Variations in uptake and binding parameters as a function of external pH for a mutated strain pAB25
As shown in Fig. 3, the uptake parameters for all bases were modified in Kt,app, which were significantly higher than those determined in the pAB4 strain for cytosine and hypoxanthine (Fig. 1). The variations as a function of external pH were similar, with a continuous increase (Fig. 3A). However, with this strain, the K,,,, variation as a function of pH was significant only with hypoxanthine. The situation was more complex for V,,, since their variations were different with the three solutes, i.e. with adenine almost no variation, but with hypoxanthine and cytosine a continuous decrease
788 100
i
I
I
t
80 h
a
60
M
fi 2 a
v
n d"
Y
40
400 -
Y
a r:
20
-g 200 75
3
5
4
c
4
6
0
m
PH 25 i
I
9
1
0
3
4
5
6
PH
B
Fig. 4. Variation in purine and cytosine binding on plasma membraneenriched fractions from pAB25 strain as a function of external pH. Specific binding measurements were performed as described in Experimental Procedures and in the legend of Fig. 2 at a given ligand concentration. (0)Adenine 2.7 pM; (A)hypoxanthine 55 pM; (0) cytosine 25 pM, concentrations chosen after Kdappdetermination at different pH values as described in Experimental Procedures (not shown). Kdapp= 55.8t-8 pM at pH 3.6 for hypoxanthine; Kdapp= 5.5k0.8 p M a t p H 5.0foradenineandKdapp= 40t-17 p M a t p H 5.0 for cytosine.
>
0
3
5
4
6
cytosine transport. The translocation involves two substrates; the proton and the nucleobase (the latter being more or less Fig. 3. Variation in purine and cytosine uptake parameters as a function protonated). Determination of Kla,, at different pH values of external pH for pAB25 strain. Experimental conditions were similar showed that, when the proton concentration increased (i.e. to those described in Fig. 1 . (A) Apparent Michaelis constant of when the protonation of PCP increased), the Kla,, decreased. uptake (Ktapp).(B) Maximun rate of uptake ( Vmax). (0)Adenine, (0) Owing to the inherent complexity of the uptake mechanism, cytosine and (A)hypoxanthine. The average is for two sets of exper- this result alone cannot account for this process. As already iments. stated [8], there is now a need for the analysis of nucleobase binding at equilibrium on membrane preparations as a function of pH in the bulk phase. In order to explain the results of these binding studies, the following two assumptions have to be made: (a) the proton when pH increased from pH 3.5 to pH 5 (Fig. 3B). Above concentration at the membrane surface is similar to that meapH 5, V,, was almost constant. In addition, it should be sured in the bulk phase; (b) ligand protonation on the memnoted that, with cytosine, a twofold increase in V,, compared brane should be almost identical to that in water. Therefore, to the pAB4 strain was observed above pH 5. with hypoxanthine, for which there is no ionization variation As with the plasma membrane fraction isolated from pAB4 within the range of the external pH studied, only a modifistrain (PMF-PAB~),only one class of specific binding site cation of PCP amino acid residue side-chain protonation may was detected for the plasma membrane fraction isolated from explain the observed changes in Kdapp. pAB25 strain (PMF-pAB25), and the amounts of specifically The binding of H + and base to PCP may occur in three bound solute were similar for all ligands (1000 pmol . mg-' different ways; either two ordered mechanisms (H + first, then . protein-', data not shown). Variations in the amount of base or vice-versa), or a random process with the two ligands specifically bound solute at a given concentration of ligand as bound independently. Our first assumption was that a function of external pH are shown in Fig. 4. As already hypoxanthine and the cotransported proton are bound on observed with PMF-pAB4 (Table 1 and Fig. 2), Kdappin- PCP (assuming that it is a 1 : l stoechiometric process), creased with pH. Compared to PMF-pAB4 (Fig. 2), a shift of through the more general random process shown in scheme the curves towards a more acidic pH was observed for the 1, where K1 and K2 are the dissociation constants of the three ligands. proton binding to free and liganded PCP respectively, and Kd and Kh are the dissociation constants for hypoxanthine binding to protonated and free PCP, respectively. Calculation of DISCUSSION these constants was using the following equations. (1) Lb = Bmax ' [L1/(Kd,,, + [L1>> Varying the extracellular proton concentration modifies the apparent Michaelis constant of adenine, hypoxanthine and Lb being the amount of ligand bound. mg protein-' at a given PH
789 201
L
15
i
'
'
'
'
'
'
'
'
. '
.
I
Kd (PCP)-H
(PCP)-LH
7 L
Scheme 1. Ligand binding to plasma membrane-enrichedfractions. It is assumed that in the pH range considered, only the PCP ionization would be modified and hypoxanthine ionization would be unchanged. K1 and K2 are the dissociation constants of proton binding on PCP and (PCP)-L respectively, and Kd and Kb the dissociation constants for the ligand binding on (PCP)-H and PCP, respectively. Results in Figs 2 and 4 show that (PCP)-L still existed at pH 6. PCP is assumed to be in the form of dispersed material with no sideness in the plasma membrane preparations.
0
E
free ligand concentration [L], B,, the maximum amount of ligand bound and Kda,, the apparent half-saturation constant of binding with
+
Kda,,= Kd(K1 [Hf1)/(K2+ [Hfl) and
(2)
(3) K1.& = K2.Kd. It was shown that B,,, was not pH dependent. The variations in Kda,, for PMF-pAB4 as a function of proton concentration [H'] in the bulk phase are given in Fig. 5A. In this figure, two kinds of data are displayed; those taken from Table 1 and those calculated using Eqn (1) from the data in Fig. 2. The curves Kda,, versus [ H f ] were fitted according to Eqn (2). Values of the calculated constants are displayed in Table 2. With this random model, it is clearly shown that the protonated form of the permease (PCP)H displayed a better affinity for hypoxanthine than the non-protonated form (PCP) (K)d= 1.68 pM and K'd = 17.4 pM). If it is assumed that proton binding occurs before the hypoxanthine-binding process (an ordered mechanism with the proton first), we should obtain no bound ligand when H f approaches 0, and a value of Kda,, approaching infinity. This does not occur since Kda,, can be precisely determined at high pH (Table l), and since the Kda,, curve approaches a constant value while the proton concentration tends to zero (not shown and Table 2). In another set of experiments, detectable amounts of ligand were measured at more alkaline pH (pH 7.5, results not shown). Therefore, this ordered binding process, with the proton binding first, can be ruled out. Fitting of the experimental data is possible, assuming that hypoxanthine is bound before the proton (ordered mechanism, ligand first) and gives fairly decent values for K > (16k1.3 pM) and K2 (33k5.3 pM with pK2 = 4.6f0.1). With this mechanism, we should obtain a Kda,, variation towards 0 when [H'] tends to 0. However, at lower pH the experimental values of Kda,, are systematically higher than those calculated using this model. Therefore, at this point, since we could not perform binding experiments with the membranes at a pH below 3.5 to decide which model was correct, we considered that the random mechanism more accurately indicated the binding of proton and hypoxanthine to plasma membrane fractions.
100 150 200 [H+l (wM)
50
250
300
400
3. v s P
s
200
O
f
0
.
I
50
,
I
100
,
I
150
,
I
200
,
I
250
,
!
300
Fig. 5. Variation in hypoxanthine apparent dissociation constant on PMF-pAB4 and PMF-pAB25 as a function of proton concentration in the hulk phase. Kda,, measured as described in Table 1 ( 0 )or calculated from data displayed in Fig. 2 (for PMF-pAB4) and Fig. 4 (for PMF-pAB25) using Eqn (1) (0)were plotted versus the proton concentration. (A) pAB4, B,, = 1000 pmol . mg protein- '. (B) pAB25, B,, = 980 pmol . mg protein-'. The curves were obtained by fitting Eqn (2) to the experimental data. The calculated dissociation and ionization constants are diplayed in Table 2.
These results also show that an ionizable group with an apparent pK1 of 3.8 was evidenced for the wild-type PCP. The binding of hypoxanthine shifted this pK by about one pH unit towards a more alkaline pH (pK2 = 4.8). The question arises as to how hypoxanthine binding induces such a pK variation of a PCP amino acid side chain. In order to explain this, we assume that the environment of this group was modified by ligand binding. This pK modification can be explained either by a direct interaction between the ligand bound close to (or near) the residue or by an indirect ligand effect through a conformational change modifying the residue environment. The nature of the amino acid involved in this step remains to be elucidated but, in view of the calculated pK, it should correspond to a carboxyl group. This hypothesis is in agreement with the intervention of a carboxyl group of an amino acid side chain for the specific recognition of purine and pyrimidine bases [18]. As yet, it cannot be concluded that this group is the same as that already identified by kinetic studies of hypoxanthine uptake with a wild-type strain [7], although it seems likely.
790 Table 2. Hypoxanthine binding parameters on plasma membrane-enrichedfractions. Dissociation constants KI, K2 and Kd were obtained from the data displayed in Fig. 5A (for PMF-pAB4) and Fig. 5B (for PMF-pAB25) and calculated by fitting Eqn (2) to the data using a nonlinear regression Turbo Pascal program. K &was calculated using Eqn (3). The correlation coefficients of the fits were 0.9933 and 0.9923 for PMFpAB4 and PMF-pAB25, respectively. Since it was not possible to perform binding experiments at low pH (below pH 3.5), K1 values were determined with a very high standard deviation. Strain
Binding parameter
pA B4 pAB25
1.68 f 0.4 14.4 f 4.3
17+ 5 570 f 200
170f 70 850 f 800
The same kind of calculations were performed both for PMF-pAB25 and for PMF-pAB4. Results are presented in Table 2 and in Fig. 5B. As for the pAB4 strain, the random model gives a better description of the proton dependence on the ligand binding. The mutation has two effects on hypoxanthine binding; a dramatic increase in the dissociation constants Kd (10-fold increase) and K &(30-fold increase) and the pK value of the ionizable group of the non-liganded PCP was more acidic than that calculated for the wild-type strain (pK1 around 2.9) as expected from the data in Figs 2 and 4. It is to be noted that the pK of this group remained the same as that calculated for the wild-type strain when hypoxanthine was bound (Table 2, pK2 = 4.7). In summary, this multiple mutation seems to modify both proton and ligand binding to PCP. Since adenine and cytosine display changes in their ionization state in the pH range where measurements were performed, scheme 1 has to be modified in order to analyze Kda,, variation as a function of external proton concentration, and to calculate the true dissociation and ionization constants. Attempts have been made to fit experimental data obtained with adenine and cytosine binding with a model taking into account different ligand ionized forms. These assume that the ligand-binding process was similar for these two ligands compared to hypoxanthine. Our experimental data were not accurate enough to obtain a good fit and to calculate the binding parameters as was performed for hypoxanthine. However, two comments may be made here: (a) the importance of the protonation state of the nucleobase was clearly evidenced since the binding of the protonated form of adenine and cytosine to the permease was apparently less efficient than binding to the unprotonated forms; (b) as observed for hypoxanthine, a shift of the binding curves towards a more acidic pH occurred in PMF-pAB25 when compared to PMFpAB4 (Figs 2 and 4). CONCLUSION The data in this paper demonstrate that the binding of ligands to PCP depends on the external proton concentration. In view of studies dependent on coupling cation concentrations of some transport carriers in Escherichia coli, an ordered binding mechanism has been proposed as the energycoupling mechanism for these secondary transport systems [8, 19, 201. For PCP in yeast, according to our analysis with hypoxanthine, the binding process could be a random mechanism, the proton and base being bound independently. Although the situation seems relatively simple when considering hypoxanthine, the system has a much more complicated
16.3 4 21 f 6.5
3.8 f 0.2 2.9
4.8 f 0.1 4.7 f 0.2
behaviour when the bases are protonated differently, such as adenine and cytosine. Zn-vivo and in-vitro studies of simpler mutants of PCP will allow a more precise approach to the molecular mechanism of this permease. For hypoxanthine, the main effect of mutations in the region with the cluster of amino acids changed at positions 371, 375 and 377 is an increase in dissociation and ionization constants. Therefore, it cannot be stated that this region of the polypeptidic chain is specifically and exclusively involved in the solute recognition process. The occurrence of a possible ‘double-mutational effect’ [21], also modifying the cotransported proton recognition process, should be considered. One of the ways to study this and to understand its precise involvement in the hypoxanthine transport pathway, is to built new strains bearing multicopies of the FCY2 gene mutated on only one amino acid. Such mutant strains are being developed. We gratefully acknowledge Prof. 0.Viratelle (Centre de Recherche Paul Pascal, Bordeaux) for constructive theoretical discussion and Dr M. R. Chevallier (Znstitut de Biologie Moliculuire et Celluluire, Strasbourg) for the gift of the mutant strain and advice. We also acknowledge A. Bocquier for skillful technical assistance and R. Cooke for improving the language.
REFERENCES 1. Polack, A. M. & Grenson, M. (1973) Eur. J . Biochem. 32,276282. 2. Reichert, U. & Winter, M. (1974) Biochim. Biophys. Actu 356, 108- 116. 3. Chevallier, M. R., Jund, R. & Lacroute, F. (1975) J . Bucteriol. 12, 629-641. 4. Reichert, U., Schmidt, R. & ForCt, M. (1975) FEBS Lett. 52, 100- 103. 5. Reichert, U. & ForEt, M. (1977) FEBS Lett. 83, 325-328. 6. Hopkins, P., Chevallier, M. R., Jund, R. & Eddy, A. A. (1978) FEMS Microbiol. Lett. 49, 173- 177. 7. Forit, M., Schmidt, R. & Reichert, U. (1978) Eur. J . Biachenz. 82, 33-43. 8. Yamato, I. (1992) FEBS Lett. 298, 1 - 5. 9. Schmidt, R., Manolson, M. F. & Chevallier, M. R. (1984) Proc. Nut1 Acad. Sci. USA 81,6276-6280. 10. Weber, E., Rodriguez, C. Chevallier, M. R. & Jund, R. (1990) Mol. Microbiol. 4 , 585 - 596. 11. Weber, E., Jund, R., Silve, S., Tsapis, R., Pallares, C., Rodriguez, C. & Chevallier, M. R. (1989) in Highlightsofmodern biochemistry (Kotyk, A,, Skoda, J., Paces, V. & Kostka, V., eds) vol. 1, pp. 761 -770, VPS International, Utrecht. 12. Bloch, J. C., Syhcrova, H., Souciet, J. L., Jund, R. & Chevallier, M. R. (1992) J . Mol. Microbiol., in the press. 13. Brethes, D., Chirio, M. C., Napias, C., Chevallier, M. R., Lavie, J. L. & Chevallier, J. (1992) Eur. J . Biochem. 204, 699-704.
79 1 14. Chirio, M. C. (1990) PhD These Universitt de Bordeaux 2. 15. Chirio, M. C., Brethes, D., Napias, C., Grandier-Vazeille, X., Rakotomanana, F. & Chevallier, J. (1990) Eur. J . Biochem. 196,293 -299. 16. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1961) J. Biol. Chem. 193, 265-275. 17. Dawson, R. M., Elliot, D. C., Elliot, W. H. &Jones, K. M. (1986) in Data for biochemical research, Oxford Science Publications (Clarendon Press).
18. 19. 20. 21.
Htlene, C. (1977) FEBS Lett. 74, 10-13. Yamato, I. & Rosenbush, J. P. (1983) FEBS Lett. 151, 102- 104. Yamato, I. & Anraku, Y. (1989) Biochem. J. 258, 389-396. King, S. C. &Wilson, T. H. (1990) Molecular Microbiol. 4,1433 1438.
