Isolation and functional reconstitution of a phosphate binding protein of the cyanobacterium Anacystis nidulans induced during phosphate-limited growth.

Adaptation of the blue-green algae Anacystis nidulans to phosphate-deficient growth leads to the expression of two membrane proteins, which appear as major constituents after separation by gel electrophoresis. One of these proteins, referred to as high affinity phosphate binding protein, has been isolated and its function reconstituted in liposomes. Partial sequencing showed no significant homologies to other proteins. The binding capacity of the proteoliposomes could be inhibited by arsenate but not by sulfhydryl reagents. Scatchard plot analyses of phosphate binding to reconstituted proteoliposomes suggested the existence of two different binding sites, one with a dissociation constant below micromolar and the other in the micromolar range.


Adaptation of the blue-green algae Anacystis nidu2ans
to phosphate-deficient growth leads to the expression of two membrane proteins, which appear as major constituents after separation by gel electrophoresis. One of these proteins, referred to as high affinity phosphate binding protein, has been isolated and its function reconstituted in liposomes. Partial sequencing showed no significant homologies to other proteins. The binding capacity of the proteoliposomes could be inhibited by arsenate but not by sulfhydryl reagents. Scatchard plot analyses of phosphate binding to reconstituted proteoliposomes suggested the existence of two different binding sites, one with a dissociation constant below micromolar and the other in the micromolar range.
In oligotrophic lakes the phosphate concentration is extremely low and usually does not exceed the nanomolar range (1). In consequence, incorporation of this nutrient has to proceed against a concentration gradient of about 4 or 5 orders of magnitude in Anucystis nidulans (2). The molecular components that catalyze this process have so far not been characterized, and very little is known about the energy-consuming mechanism underlying the translocation of phosphate across the cell membrane. Among other prokaryotes only the phosphate uptake system of Escherichia coli has been characterized in detail at the molecular level (for review see Refs. 3 and 41, including the structural elucidation of the periplasmic binding protein (5). However, since E. coli lives in a completely different environment, where phosphate usually does not become a limiting nutrient, it can be expected that the properties of its uptake system might not be representative for an organism that has to adapt to nutrient fluctuations at or below the nanomolar range.
Using the cyanobacterium A. niduZuns we have investigated the membrane proteins that are induced during phosphate-* The financial support of the Austrian Fonds zur Forderung der Wissenschaftlichen Forschung is gratefully acknowledged. The costs of charges. This article must therefore be hereby marked "aduertisement" publication of this article were defrayed in part by the payment of page in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
11 To whom correspondence should be addressed.
limiting growth conditions. The present report deals with a high affinity phosphate binding protein (HAPBPI1 that is different from the binding protein in E. coli. It could not be released by cold osmotic shock and revealed a different phosphate binding behavior in reconstitution experiments. Furthermore, the results of partial sequencing of the HAPBP showed no homologies with any characterized protein. EXPERIMENTAL PROCEDURES Materials and Growth Conditions-A. nidulans (Synechococcus sp.) Drouet (strain 1402-1, Algal Culture Collection Gottingen) was grown photoautotrophically in medium D of Kratz and Myers (6) at a temperature of 37 "C and supplied with air enriched with 5% COz. Non-phosphate-deficient algae were cultivated on 3 m M external phosphate in batch cultures. For phosphate-deficient growth algae were cultivated in a 20-liter vessel on 2.5 VM total phosphorus in a discontinuous mode; every 2nd day 10 liters of the suspension was removed for preparation, and the culture was refilled with 10 liters of fresh medium, again containing 2.5 p~ K,HPO,, which was incorporated by the starved culture within less than half an hour. The state of deficiency of these algae was monitored by measuring the uptake activity and the threshold value as described previously (7).
Electrophoresis-Analytical gel electrophoresis was carried out on 8-22% gradient polyacrylamide minislab gels according to the procedure of Matsudaira and Burgess ( 8 ) in the buffer system of Laemmli (9). Protein samples were supplemented with SDS sample buffer (final concentrations: 2.5% SDS, 1% p-mercaptoethanol, 7% glycerol, 0.001% bromphenol blue, 62.2 m M TriskICl, pH 6.8) and boiled for 1 min in a water bath. Gels were stained with Coomassie Brilliant Blue R-250 and destained in 10% acetic acid.
Protein Purification-Cells were harvested by centrifugation at 4,000 x g for 6 min, washed with distilled water by centrifugation, and resuspended in 40 ml of medium I that contained 20% sucrose, 5 m M NaCI, and 10 m M HEPESIKOH, buffered at pH 7.6. EDTA (2 m) was added to remove the outer membrane, and the cells were incubated at 37 "C with 40 mg of lysozyme to digest the cell walls. After 2.5 h the spheroplasts were centrifuged at 6,000 x g for 6 min, washed with medium I1 containing 10 m M NaCl, 2.5 m M MgClZ, 1 m M phenylmethylsulfonyl fluoride, and 20 m M HEPESIKOH, pH 7.6, and then incubated for 10 min in about 5 ml of medium I1 with additional 0.1% Triton X-100 and 300 m M Na2S04 (or 500 m M Na2S04 without Triton). The spheroplasts were centrifuged at 20,000 x g for 10 min, and (NH,),SO, was added to the supernatant to 3 M final concentration. The precipitated proteins were collected by centrifugation and redissolved in medium 11. After further centrifugation at 100,000 x g for 10 min to remove all cell fragments (crude protein extract), 500 pl of the protein suspension was applied to an FPLC Superose 12 column equilibrated in medium 11. Proteins were chromatographed at a constant flow rate of 0.3 ml.min", and fractions of 1 ml were collected for gel electrophoresis. The phosphate binding activity of individual fractions was assessed in the reconstitution experiments described below.
Protein BLotting and Amino Acid Sequencing-Protein peak fractions from the FPLC column containing the high affinity phosphate binding protein (see Fig. 2) were run on acrylamide minislab gels, and stained bands at around 30 kDa corresponding to the phosphate binding protein were excised with a scalpel and pooled. The gel slices were incubated in Tris/HCl, pH 6.8, three times for 15 min for equilibration and loaded onto a 15% acrylamide gel with 1.5-mm spacers containing a 4% stacking gel constituting about two-thirds of the gel length. The concentrated gel band was electroblotted onto a polyvinylidene difluoride membrane Ummobilon P, Millipore) as described previously 110, 11). The blot was air-dried and stored in aluminum foil at 4 "C.
The excised polyvinylidene difluoride membrane was cleaved in situ with trypsin. Liberated peptides were separated by reversed phase high pressure liquid chromatography on a Bakerbond C18 column in 0.1% trifluoroacetic acid developed with a linear gradient from 0 to 70% The abbreviations used are: HAPBP, high affinity phosphate binding protein; FPLC, fast protein liquid chromatography.
5509 of Cyanobacterium A. nidulans acetonitrile in water a t 30 "C. Peaks were sequenced in an Applied Biosystems 477A automated protein sequenator according to the manufacturer's instructions.
Reconstitution and Binding Experiments-The liposomes used for the reconstitution experiments were prepared by the method previously described (12). The lipids employed in this procedure had been extracted from the membranes ofA. nidulans (13) in order to reconstitute the proteins in a natural lipid environment.
The reconstitution followed a modified freeze-thaw sonication method (12). An aliquot of the protein suspension (50-500 pl, either crude protein extract or purified protein) was added to 50-250 pl of liposomes that contained 0.05% Triton X-100. The solution was then diluted with 1-5 ml of medium I1 below the critical micelle concentration of Triton X-100 (0.02% (cf. Ref. 14)) and rapidly frozen in liquid nitrogen. After 5 min the sample was thawed in a water bath at room temperature, kept on ice, and sonified twice for 15 s each (W-220 F, Ultrasonics Inc.). The resulting proteoliposomes were centrifuged in an Eppendorfcentrifuge, and the pellet was resuspended in medium I1 and finally analyzed by gel electrophoresis and used for binding experiments.
Protein was determined by the procedure of Bradford (cited in Ref. 15) or following the bicinchoninic acid method (Sigma), which is less disturbed by the presence of detergents.
Binding experiments with the reconstituted system were camed out a t room temperature under equilibrium conditions; 5 pl of the respective ["2Plphosphate solution was added to 80 p1 of proteoliposomes and incubated for about 10 min. For each phosphate concentration double samples were taken.
After incubation the samples were centrifuged, 50 pl of the supernatant and an aliquot of the resuspended pellet was taken, and the radioactivity was measured in water in a scintillation counter.
Arsenate is a competitive inhibitor of phosphate transport (16) that does not enter the cell, as shown by its lack of effect on photophosphorylation in A. niduluns. 2 We used this inhibitor to block phosphate binding to proteoliposomes. In the presence of this inhibitor only a small amount of tracer was found in the proteoliposome pellet, independent of whether or not the inhibitor was added before or after the ligand, indicating the absence of phosphate transport into the proteoliposomes (17). Independent experiments with tritium water to determine the pellet water confirmed that the remaining "phosphate binding" in the presence of arsenate corresponded to tracer unspecifically trapped by the centrifuged proteoliposomes. The specific binding was therefore defined as the total binding without arsenate minus the unspecific binding of the proteoliposomes determined in the presence of 5-10 mM Na2HAs0,.
The yield of reconstituted proteoliposomes obtained from the starting volume of a 10-20-liter algal culture was too low to allow both binding studies and protein determination. Therefore binding was related to the suspension volume. The dissociation constant obtained in Scatchard plots (18) is, however, unaffected by the reference system. Fig. 1 shows that the synthesis of two proteins in high amount is induced within a few hours after the onset of phosphate-limiting growth conditions (lanes 1 3 , bands I and ZI).

Changes in Protein Pattern Caused by Phosphate-deficient Growth-
Lane I shows the proteins of phosphate-unlimited algae. Lanes 2 and 3 represent the protein pattern in the transition state between static and phosphate-limited growth. The algae applied to lane 4 were obtained from a culture adapted to permanent phosphate-limited growth conditions as described under "Experimental Procedures." In this physiological state the proteins of bands I and I1 are predominantly expressed in A. nidulans; whereas the protein of band I showed no phosphate binding or transport activity, the protein of band I1 with an apparent molecular mass of 28-35 kDa showed high phosphate binding activity in liposome reconstitution experiments (see below). We have therefore designated this protein as the HAPBP ofA. nidulans. This protein was also found to be present in the so-called cytoplasmic membrane fraction (19) and could easily be solubilized from spheroplast preparations by either 200 mM Na2S04 and 0.05% Triton X-100 or 500 mM G. Falkner, unpublished data. Na2S04 without Triton, strongly indicating its location in the cytoplasmic membrane. The HAPBP could, however, not be released by the cold osmotic shock procedure so that the protein seems not to be a periplasmic binding protein (20) but a (peripheral) constituent of the plasma membrane that is loosely associated with it.
Elution profiles of this protein suggested a mainly globular appearance and gave no indications of self-aggregation. Initial sequencing revealed a blocked NH2 terminus of the HAPBP, therefore prohibiting the determination of the NH2-terminal sequence. After in situ digestion with trypsin and reversed phase chromatography of the liberated peptides two peaks were sequenced giving rise to two non-overlapping sequences, SFVNFYLQNA and VYAPGTDSGTYDYFNEAILNK. Both sequence stretches showed no significant match with any other protein in homology searches in the PIR and SWIS-SPROT data base.
Reconstitution-Treatment with 0.1% Triton X-100 and 300 mM Na2S04 did not result in lysis of the spheroplasts. Electrophoretic analysis of the crude protein extract and the remaining membranes showed that this low concentration of detergent removed only the HAPBP quantitatively from the membranes.
This protein solution could be successfully used for reconstitution experiments. Fig. 1, lane 6, shows that only the HAPBP was present in the reconstituted proteoliposomes (the protein band of low molecular weight is due to a contamination with lysozyme, previously used in the preparation of spheroplasts). Although many other proteins were present in the crude extract, apparently none of them associated with the liposomes in significant amount. Fig. 2 shows the elution profile of the FPLC fractions and the corresponding phosphate binding activity of these fractions reconstituted in proteoliposomes. The fraction that contained the main portion of HAPBP (18 in this experiment; see inset) exhibited the highest activity. Measurements of the binding capacity of fractions flanking the peak showed that the activity specifically followed the presence of HAPBP and was not associated with any of the contaminating proteins in fractions 18 and 19. Binding Studies with the HAPBP-Binding was investigated using proteoliposomes reconstituted either with the crude protein extract or the purified HAPBP. These preparations showed a significantly high phosphate binding capacity, whereas binding by protein-free liposomes was negligible. In a representative experiment the binding activity of the control (protein free) was 0.03 pmol of Pi.liter-l liposome suspension, while that of the reconstituted system was 0.49 pmol of P,.liter-' at 1 p~ phosphate. The phosphate binding process could not be inhibited by N-ethylmaleimide or mersalyl acid a t concentrations between 0.1 and 2 mM, indicating that sulfhydryl groups are not essential for binding. However, binding was completely abolished in the presence of 5 mM arsenate.
The binding parameters of HAPBP were determined by Scatchard plots (18). Fig. 3 shows an experiment with proteoliposomes containing the purified HAPBP. It can be seen from the inset that the binding data could not be fitted by a straight line but followed a curved relation that suggested the existence of more than one binding site.
The dissociation constants were calculated using a computer program based on the assumption that this behavior reflects the existence of two binding sites with different affinities for phosphate. For the data shown in Fig. 3 this treatment gave a high affinity dissociation constant of 0.3 p~ and a lower one of 10 PM phosphate. Hence the high affinity binding constant is somewhat lower than that of the periplasmic phosphate binding protein of E. coli (1 p~ (4)).
Some of the proteoliposome preparations did not exhibit the high affinity binding site in the submicromolar range. In these cases the data described a straight line that only reflected the low affinity binding site with dissociation constants between 2 and 10 V M phosphate (cf. Fig. 4, inset, with a dissociation constant of about 4 p~) , possibly due to the inactivation of the high affinity binding during preparation. This may indicate that the protein in its native state may have an even higher binding affinity than we were able to measure in the best case.
We suggest that during phosphate-limited growth the synthesis of a phosphate binding protein is induced in A. nidulans. This protein is different from the polypeptide of E. coli in that it shows no sequence homology and seems to be a constituent of the cytoplasmic membrane. The finding of a phosphate binding protein in these algae opens a way to new investigations in the phosphate uptake process.