- Email: [email protected]
Isolation and characterization of a 50 kDa testosterone-binding protein from Pseudomonas testosteroni
J. sreroidBiochem.Vol. 32, No. lA, pp. 27-34, 1989 Printedin Great Britain. All rights reserved
Copyright 0
0022-4731/89 $3.00 + 0.00 1989 Pergamon Press plc
ISOLATION AND CHARACTERIZATION OF A 50 kDa TESTOSTERONE-BINDING PROTEIN FROM PSEUDOMONAS
TESTOSTERONZ
JAMESE. THOMAS*,ROGAYAHCARROLL,LUISA PO SY and MAMORUWATANABE Department of Medicine, Faculty of Medicine, The University of Calgary, Calgary, Alberta, Canada T2N 4Nl
(Received 18 March 1988)
Summary-A testosterone-binding protein (M, = 50,500) has been isolated from the Gram-negative bacterium Pseudomonas testosteroni. The protein was partially purified by a combination of ion exchange chromatography and chromatofocusing. Final purification was achieved by electroelution of the 50 kDa protein from SDS-polyacrylamide gels. Following renaturation from a diluted solution of guanidineHC1, specific binding of [3H]testosterone to the purified protein was observed. The native protein has a pI of 6.8. It appears to contain 428 amino acids, 39% of which are hydrophobic. There is only one cysteine residue. Both chymotrypsin and VS protease were used to produce peptide maps of the protein for use in future identification. The first 10 amino acids situated at the N-terminal of the protein were Ser-Pro-Phe-Asp-Leu-Arg-Pro-Leu-Se+Gly. Testosterone binding to the protein was saturable at _ 3.8 nmol/mg protein; the binding constant was -25 nM. Unlabelled testosterone, androstenedione, Sa-dihydrotestosterone and Sfi-dihydrotestosterone were able to compete for [3H]testosterone bound to the protein; 17/I-estradiol also competed for [3H]testosterone but to a lesser degree. Neither progesterone nor desoxycorticosterone competed for the testosterone-binding site. Binding of testosterone to the protein was stable at pH’s ranging from 5.5 to 9.0 and at various temperatures ranging from 4 to 30°C. The protein was unable to metabolize testosterone in either the presence or absence of the cofactor NAD.
active transport system [5] consisting of both periplasmic [6] and membrane-integrated [5,7,8] protein components. A periplasmic component has been partially purified by ammonium sulfate precipitation [9]. It possesses at least one high affinity testosterone-binding protein (& = 9.4 nM) and has a Svedberg coefficient of 4S [lo]. A membraneassociated steroid-binding protein has also been identified [ 111. This 34 kDa cationic testosteronebinding protein is released from. Kaback membrane vesicles [12] during centrifugation at 45,OOOg.It binds testosterone (& = 39 nM), and has 278 amino acids, 44% of which are hydrophobic. It also possesses 9 cysteine residues. In this paper, we report on a 50 kDa testosteronebinding protein which appears to be membraneassociated. Because of its location, high Kd value and inability to metabolize testosterone, we suggest that it is one of the protein components involved in an osmotic shock sensitive steroid transport system indigenous to P. testosteroni.
INTRODUCrION Gram-negative bacteria make use of a number of binding-protein systems in order to actively transport extracellular metabolites into the cell. These systems can be subdivided into those which are sensitive to osmotic shock and into those which are insensitive. Transport systems sensitive to osmotic shock require three or more proteins, at least one of which is a binding protein located in the periplasmic space of the cell [l]. Energy to drive these systems appears to come from hydrolysis of phosphate bonds associated with either ATP, acetyl phosphate or a similar compound [l-3]. In contrast, shock-insensitive transport systems consist of membrane bound carrier proteins energized by a chemiosmotic ion gradient similar to that proposed by Mitchell[4]. P. testosteroni is a Gram-negative bacterium capable of growth on certain Cl9 and C21 steroids. The mechanisms involved in transporting free extracellular steroid across the cell wall and membrane of the bacterium to the inside of the cell are essentially unknown. In order to elucidate these mechanisms, the proteins involved in the transfer process must first be identified. Work from this laboratory has shown that access to extracellular steroid occurs via an
EXPERIMENTAL Materials P. testosteroni, ATCC 11996, was obtained from American Type Culture Collection, Rockville, MD. Unlabelled steroids were obtained from Steraloids, Wilton, NH and from Sigma Chemical Co., St Louis, MO. Labelled [7-‘HItestosterone (25 Ci/mmol) was
*Present address: J. E. Thomas, Department of Biological Sciences, The University of Lethbridge, Alberta, Canada TIK 3M4.
Lethbridge, 21
28
JAMESE. THOMASet al.
obtained from New England Nuclear Coloration, Boston, MA. DEAE-Trisacryl is a product of LKB, Bromma, Sweden. Polybuffer Exchanger 94 and Polybuffer 96 are products of Pharmacia, Uppsala, Sweden. All other chemicals were purchased from either Sigma Chemical Co., St Louis, MO or Fisher Scientific, Edmonton, AB.
is directly proportional to the amount of steroid nonspecifically bound to the protein. Thus specific binding is calculated from the difference between total binding and nonspecific binding of steroid to the protein sampIe. Enzyme activity
Samples were assayed for their ability to metabolize testosterone, i.e. for the presence of vesicles Al-dehydrogenase and 3(17)fl-hydroxysteroid dehyas described drogenase. Fractions consisting of about 3-5 pg of testosteroni was grown P. column previously [ 131for 40 h at which point 0.5 g of testos- protein taken from the chromatofocusing terone (17~-hydroxy-~androsten-3-one} was added and 40-5Opg of protein obtained from the vesicle as a powder per liter of culture in order to induce the supematant were tested for the enzyme activity. The formation of a steroid transport and metabolic sys- latter was used as a control containing known dehytem within the bacteria. After an additional 40 h drogenase activity [15]. For each assay, protein was SOpCi period, cells were harvested by concentration with a added to a reaction mixture containing Millipore Pellicon Cassette System equipped with 10 [3H]testosterone mixed with 10 pg unlabelled testosmembranes each possessing an exclusion limit of terone, 0.5 M Tris-HCl pH = 9, in a final volume of 300,000 mol. wt. The cells were washed in 10mM 1 ml. Samples were incubated at 22°C for l-3 h and then assayed for steroid metabolites using the method of Tris-HCl, pH = 8 at 4”C, and the residual medium removed by cent~fugation at 12,OOOg for 10min. Francis and Watanabe[ 161. A need for cofactors was Membrane vesicles were prepared from whole cells tested by addition of 1 mM nicotinamide adenine dinuusing a method initially outlined by Kaback[l2] but cleotide (NAD) to samples as indicated in the text. as modified and described by Francis and General Watanabe[l4]. Under these conditions the protein of Protein concentrations were assayed by the Cointerest appeared to be released sporadically during osmotic shock in lysozyme-EDTA. Thus, in order to omassie blue method described by Bradford[l7], using bovine serum albumin as standard. facilitate recovery of the testosterone-binding proVertical slab gel el~trophoresis was performed in tein, EDTA was omitted from all solutions used dur0.2% (w/v) SDS using 7.5% (w/v) polyacrylamide ing preparation of the vesicles. Membrane-associated gels according to the method of Laemmli[lll]. proteins, including the testosterone-binding protein, Isoelectric focusing was performed on an LKB were found in the supernatant fraction recovered Multiphor 2117 using Ampholine PAG Plates, pH during final pelleting of the vesicles at 45,000g. 3.5-9.5 (LKB). Methods were as described for the kit. Binding assays The pH gradient was determined from p1 standard proteins obtained from Pharmacia. When radioBinding assays were performed by equilibrium labelled samples were run on an isoelectric focusing dialysis using a method modified from Watanabe et gel, each lane from the unfixed gel was cut into 2 mm af.[lO]. Assays were performed in a buffer consisting sections. Gels were solubilized in 250~1 of a 19:l of 30 mM Tris-HCl, pH = 7, containing 1 M KC1 in order to facilitate protein solubility. Each assay was (v/v) mixture of 50% hydrogen peroxide: 0.88% ammonium hydroxide and left overnight at 37°C. performed by placing no more than 1 ml of protein solution in each of two dialysis bags. One bag was The following day, 5000 U of catalase were added to each sample. After 30min (37”(Z), 250~1 of 10% added to 15 ml of buffer made 0.23 p M in unlabelled (w/v) ascorbic acid were added, the samples mixed testosterone plus 5 /tCi [3H]testosterone. The second with 5 ml of Ready-Solv HP and counted in a liquid dialysis bag was added to 15 ml of buffer made 23 p M in unla~lled testosterone ( IOO-fold excess) scintillation spectrometer. Amino acid analysis was performed on a Beckman plus 5 PCi [3H]testosterone. Solutions within each of 6300 amino acid analyzer. Samples were subjected to the two bags were allowed to dialyse to equilibrium methanesulfonic acid hydrolysis [191in order to idenfor 20 h at 4°C at which point IO&200 ,ul aliquots were removed from both inside and outside of the tify trypt~phan; performic acid oxidation [20] in order to identify methionine and cysteine residues; and bags, mixed with 5 ml Ready-Sok HP (Beckman Instruments, Inc., Palo Alto, CA) and counted in a sequential hydrochloric acid hydrolysis for 24,48 and liquid scintillation spectrometer. The difference in 72 h in 6 N HCl, 0.1% phenol, 0.1% thiodiglycolic acid at I lo”C, in vacua, in order to identify the dpms between the recovery of radioactivity from inside and from outside of the first dialysis bag is remaining amino acids. The amino acid analysis did equal to the total binding and is proportional to the not differentiate between either asparagine and aspartotal amount of steroid bound by the protein. In tic acid or glutamine and glutamic acid. Protein contrast, the difference between the recovery of radiosamples were sequenced using the method of McKay activity inside and outside of the second dialysis bag et af.[21]. Growth of organism and preparation
of membrane
Steroid-binding
protein
in P. testosteroni
29
-ao?
J+
I- 7sti. 4
- 7.0
xl
Fig. 1. Elution profile of the chromatofocusing column. The column was prepared and eluted as described each consisting of 5 ml, were collected and both the pH and absorbance at 280 nm determined for each. Aliquots of 100~1 were taken from each tube, mixed with 5 ml of Ready-Solv HP and counted in a liquid scintillation spectrometer. Typically, the testosterone-binding protein was eluted in the text. Fractions,
at a pH = 7.4, in fractions 4!j-50. RESULTS
Previous work from this laboratory has shown that the 45,000 g supematant obtained during membrane vesicle preparation contains a number of testosterone-binding proteins [11J. This fraction was precipitated in 3.9 M ammonium sulfate for 45 min (4°C) and then centrifuged at 37,000g for 30 min. The pellet was dissolved in N 50 ml of 5 mM EDTA, 25 mM Tris-HCl, pH = 7.0 (Buffer A) and dialysed into Buffer A. In order to remove anionic contaminants from the sample, it was applied to a column (2.5 x 45 cm) of DEAE-Trisacryl equilibrated with Buffer A. A nonbinding fraction containing testosterone-binding proteins was eluted with 500 ml of Buffer A [14]. This fraction was pr~ipitated in 3.9 M ammonium sulfate for 45min (4”C), and then centrifuged for 30 min at 37,000g. The pellet was dissolved and then dialysed into 25 mM ethanolamin*HCl, pH = 9.0 (Buffer B). In order to identify testosterone-binding proteins within this
Purification
fraction, proteins were incubated in lO@i of [‘H]testosterone overnight (4°C). After this, the solution was added to a chromatof~using column (1 x 40 cm) of Polybuffer Exchanger 94 ~uilibrated with Buffer B and eluted with 415ml Polybuffer 96-HCl, pH = 7.0, diluted 1: 12 with deionized water (Fig. 1). Free steroid was eluted from the column in peak 1, typically in tubes 5 through 15, whereas the steroidbinding protein was eluted as a single peak when the pH of the column reached -7.4. Protein from this peak was precipitated in 5.3 M ammonium sulfate for 45 min (4”(Z), then centrifuged at 37,000 g for 30 min. The pellet was dissolved and then dialysed into Buffer A. This fraction constituted the semi-purified protein fraction used in all subsequent experiments. The testosterone-binding protein from this fraction was purified N IOO-fold (Table 1). Protein fractions obtained during the purification procedure were examined by SDS-polyacrylamide gel electrophoresis (Fig. 2). Following chromatofocusing one major protein band M, = 50,500 rf: 3,600 (n = 10) was recovered. Major contaminants of this fraction had a M, less than 30,000.
Table 1. A typical purification table for the 50 kDa testosterone-binding protein Specificactivity Total Tota activity (pm01 testosterone protein (pm01 testosterone Purification
Recovery (%)
step’
(mg)
bound)?
bound/mg protein)
Vesicle su~matant 80% (NH&SO, saturation DEAE-Trisacryi Chromatofocusing
346.4
18,543
52
1
100
280.9
16,968
59
1
89
21.1 0.7
9,855 3,386
462 5,121
9 95
50 18
‘The (NH&SO, precipitate examined in step 2 was resuspended and dialysed into 25 mM Tris-HCI, pH = 7.0 before doing a steroid binding assay. The non-binding fraction obtained from the DEAE-Trisacryl column was precipitated in 3.9 M (NH&SO,, then resuspended and dialysed into 25 mM ethanolamine-HCI, pH = 9.0, before being assayed for total binding activity. The fraction obtained from the chromatofocusing column was precipitated in 5.3 M (NH&SO,, then resuspended and dialysed into 25 mM Tris-HCI, pH = 7.0, before assaying total binding activity. +Binding assays were performed as described in the Experimental section. All assays were performed using saturating amounts of testosterone.
JAMES E. THOMAS et al.
30
Mr (kDol
92.5 -
66.2 -
Fig. 2. S~S-polyacrylamide gel eleetrophoresis of various fractions obtained during puri~cation of the 50 kDa testosterone-binding protein. Purification of the 50 kDa protein was carried out as described in the text. Fractions were electrophoresed in 0.2% SDS on a 7.5% polyacrylamide gel. Gels were stained for 30 min in 0.4% (w/v) Coomassie blue R-250 made up in 40% (v/v) methanol, 5% (v/v) glacial acetic acid in deionized water and destained in 10% (v/v) methanol, 15% (v/v) glacial acetic acid in deionized water. Lanes 1 and 6: Bio-Rad low mol. wt standards; lane 2: 45.0001: vesicle supernatant: lane 3: (NH&SO, precipitate; lane 4: cationic fraction from the DEAE-Trisacryl column; lane 5: testosterone-binding fraction from the chromatofocusing column.
In order to show that the 50 kDa protein binds testosterone, it was purified by SDS-polyacrylamide gel electrophoresis and renatured after the method of Hagar and Burgess[22]. Semi-purified samples from the chromatofocusing column were run on 3 mm preparatory SDS-polyacrylamide gels, rinsed in I mM dithiothreitol (DTT) prepared in deionized water (4°C) and stained in 0.25 M KC1 containing 1 mM DTT (4 C). Under these conditions the 50 kDa protein appears as an opaque band on the gel after about 5 min. This band was cut out of the gel, rinsed with deionized water containing 1 mM DTT (4”‘C) and homogenized in 0.1% SDS, 10 mM Tris-acetate, pH = 8.6. The protein was electroeluted from the gel for 4 h at 3 Watts, constant current, in an Isco Electroelution Apparatus using 0. I % SDS, 40 mM Tris-acetate, pH = 8.6 in the buffering wells. After this, the sample was precipitated with 4 vol of acetone (-20‘ C) for 30 min in 15 ml glass Corex tubes and the precipitate centrifuged at 37,000 g for 30 min. The pellet was dissolved in 100 ~1 of 6 M guanidine-HCI made up in a dilution buffer consisting of 20% glycerol (v/v), 0.15 M NaCI, 1 mM DTT, 0.1 mM EDTA, 50 mM Tris-HCl, pH = 8.0, and left at room temperature for 20 min, at which point it was diluted 50-fold with dilution buffer and left at room temperature overnight. The solution was concentrated to I ml in an Amicon Stirred Cell concentrator using a Millipore PTTK membrane with a 30,000 mol. wt
exclusion limit and then assayed for testosterone binding by equilibrium dialysis. Specific binding of testosterone to the purified protein was I .7 nmol/mg protein. Only one SOkDa band could be detected following SDS-polyacrylamide gel electrophoresis and silver staining after the method of Wray et a!.[23].
The 50 kDa testosterone-binding protein was electroeluted from SDS-polyacrylamide gels as described above and its amino acid composition determined (Table 2). The protein appears to consist of 428 amino acids, 39% of which are hydrophobic. Only one cysteine residue could be detected in the protein. Based on this composition, it has an apparent mol. wt of 45,408. This discrepancy in molecular weight with results obtained by SDS-gel electrophoresis may result from secondary modifications to the protein. In two separate experiments, the N-terminal amino acid sequence for the protein was shown to be Ser-ProPhe-Asp-~eu-Arg-Pro-Leu-Ser-Gly. Only one protein was detected in these experiments confirming that the electrophoreticaily extracted 50 kDa protein was purified to homogeneity. Peptide mapping
The 50 kDa testosterone-binding protein was peptide-mapped after the method of Cleveland et a1.[24]. Treatment of the protein with various concentrations of both V8 protease and chymotrypsin produced distinct degradation profiles (Fig. 3). Note that multiple banding of the untreated 50 kDa protein appeared to result from binding of various amounts of SDS to the protein in a 0.1% SDS gel. This phenomenon was not observed when gels were run in 0.2% SDS. Characterization
Isoelectric focusing was performed on a sample of semi-purified protein labelled overnight with Table 2. Amino acid composition
of the 50 kDa testosterone-binding protein
Amino ASX Thr Ser Glx Pro GlY Ala cys Val Met Ile Leu TY~ Phe His Lys TW
A&
acid
Amino acid ratio 4.481 3.113 3.311 5.073 3.522 9.400 3.847 0.129 3.272 0.74x 2.917 3.831 2.126 2.617 0.832 2.407 0.763 2.856
4028.1 2426.4 2264.6 5034.9 262 I .7 4161.0 2133.0 103.2 2480.0 787.2 2603.6 3396.0 2611.2 2944.0 823.2 2433.8 1117.2 3436.4
Amino acid composition* 35 24 26 39 27 73 30 25 6
23 30 16 20 19 22
by assuming that the protem conraIned 1 cysteine residue.
*The amino acid composition
testosterone-binding
Projected amino acid formula weight (kDa)
was determined
Steroid-binding protein in P.
31
testosteroni
Mr 92.5k 66.2
31.0
21.5
74.4
1
2
3
4
5
6
7
8
9
10
Fig. 3. Peptide maps of the 50 kDa testosterone-binding protein. The binding protein was peptide mapped in 0.1% SDS on a 15% polyacrylamide gel after the method of Cleveland et u1.[24].Characteristic peptide profiles were obtained for both ~hymotrypsin (lanes S-IO) and V8 protease (lanes 46). Lane 1: Bio-Rad low mol. wt standards; lane 2: non-digested protein; lane 3: 0.5 pg V8 protease; lanes 4-6: protein mixed
with either 0.5, 0. I or 0.025 pg V8 protease; lane 7: 0.5 pg ch~otrypsin; lanes &IO: protein mixed with either 0.5. 0.1 or 0.025 pg chymotrypsin. [3H]testosterone (Fig. 4). While several protein bands were observed after staining with Coomassie blue (Fig. 4B), only one band at pH = 6.8 was radiolabelled (Fig. 4A). The second peak represents unbound [3H]testosterone which did not migrate from the point at which the sample was initially placed on the gel. This peak could be moved anywhere on the gel without affecting the position of the radiolabelled protein peak (data not shown). This shows, (1) that the semi-purified protein contains only one species of testosterone-binding protein and, (2) that the p1 of the native 50 kDa testosterone-binding protein is 6.8 as opposed to 7.4 as indicated by the chromatofocusing column. Steroid specificity was determined by competition between [3H]testosterone and various other nonlabelled Cl 8, Cl9 and C21 steroids for the protein binding site (Table 3). When the semi-purified protein fraction was labelled with [3H]testosterone, all unlabelled Cl9 steroids tested competed for the labelled binding site; testosterone was most efficient in competing for the labelled steroid while 5adihydrotestosterone, androstenedione and SFdihydrotestosterone were able to compete for the binding site, but to a lesser degree. The Cl8 steroid, 17/3-estradiol, also competed for the binding site, but at a lower level. The C21 steroids, progesterone and desoxycorticosterone, did not compete for the labelled binding site. Binding of testosterone to the 50 kDa protein was saturable (Fig. 5) and exhibited a Kd of 16.7nM (insert) and 33.3 nM based on two separate experiments. The Scatchard plot [25] (Fig. 5, insert) shows
that the 50 kDa protein is capable of binding -4.5 nmol testosterone/mg protein under the assay conditions used. This value correlates with that seen for protein obtained from the chromatofocusing column (Table 1); on average - 5.1 nmol testosterone was bound per mg protein obtained by chromatofocusing. This value represents about 0.2-0.3 mol of testosterone bound per mole of 50 kDa protein. Based on this observation, only a small proportion of the purified protein is capable of binding steroid under the assay conditions used. This will be discussed later. However, because the Scatchard plot (Fig. 5, insert) appears linear, we suggest that only one high affinity binding site for testosterone is located on the 50 kDa protein. Binding of [3H]testosterone to the 50 kDa protein was stable at pH’s ranging from 5.5 to 9.0 (Table 4). No steroid binding was observed at a pH of 1I .2. The binding of [3H]testosterone to the protein was also stable at temperatures ranging from 4 to 30°C (Table 4). However, at 37°C a 47% decrease in binding of [3H]testosterone to the protein was observed. This result is consistent with obervations which show that growth of the bacterium at 37°C is inhibited f13]. Incubation of vesicle supernatant protein at 22°C with 13H]testosterone resulted in metabolism of the testosterone to androstenedione and A1.4androstadienedione. Within 1 h, about 41% of the steroid was converted to androstenedione and 50% to A’*4androstadienedione. These results show that both Al-dehydrogenase and 3(17)P_hydroxysteroid dehydrogenase are present in the vesicle supernatant fraction. Francis and Watanabe[ 141have shown that
JAMESE. THOMASet al.
32
w
-
6.65
-
8.45
Wa@W -
6.15
-
7.35
-
6.65
-
6.55
).
Binding + protein
- 5.65 - 5.20
DISTANCE
FROM CATHODE
(cm)
Fig. 4. Isoelectric focusing of the semi-purified testosterone-binding protein. The 50 kDa protein obtained by chromatofocusing, was labelled overnight with [3H]testosterone and applied to an LKB-PAG plate, pH 3.5-9.5. Isoelectric focusing was performed as described in the Experimental section. Only one testosterone-binding protein (Panel A) was observed, 4.5cm from the cathode; the free radiolabelled steroid seen l-3 cm from the cathode did not migrate from the position on the plate where the sample was placed at the beginning of the run. The binding protein had a p1 of 6.8 (Panel B) when compared to pH standards run simultaneously.
A 1-dehydrogenase is removed from the vesicle supernatant fraction by use of anion exchange chromatography as performed here. When fractions eluted from the chromatofocusing column were tested for the presence of 3(17)8-hydroxysteroid dehydrogenase, we found that a peak in activity was eluted from the column when the pH of the elution buffer reached about 7.6 (data not shown). In the presence of both [3H]testosterone and 1 mM NAD, samples taken from this peak were able to convert 96% of the [3H]testosterone to androstenedione. No specific binding of testosterone to this protein fraction was obtained using the assay conditions outlined in the Experimental section. In contrast, when protein purified by electrophoresis and subsequently renatured as described in the text was mixed with [3H]testosterone and 1 mM NAD no degradation products of testosterone were found. This indicates that the 50 kDa protein has no enzymatic activity directed towards the breakdown of testosterone.
DISCUSSION
P. testosteroni
is a Gram-negative
bacterium
capa-
ble of growth on a number of Cl9 and C21 steroids. In this study, we have purified a Sb kDa protein which appears to possess a single high affinity site, specific for androgens. Testosterone, androstenedione, Scr-dihydrotestosterone and Sfi-dihydrotestosterone were all able to compete for this site on the protein. In contrast, neither of the C21 steroids, progesterone or desoxycorticosterone, exhibited competition for the testosterone-binding site supporting previous observations [IO] that a separate class of C21 steroid binding proteins exist within the bacterium. The amino acid composition indicates that about 72% of the protein is uncharged, consisting of both hydrophobic and neutral amino acids. This is consistent with data obtained by isoelectric focusing which indicates that the protein has a p1 = 6.8; near neutral. This lack of charge on the native protein may be the cause of
Steroid-binding protein in P. fesiosteroni
33
Table 3. Steroid specificity Steroid concentration fnM)
Steroid
nmol testosterone bound/ma urotein
4 69 341 66 331 69 344 69 344 64 318 61 303 73 367
Testosterone
Androstenedione Sa-Dihydrotestosterone 5~-~hyd~testosterone Progesterone Desoxycorticosterone 17/I-Estradiol
% Competition
8.41
0
0.33 0.08 1.48 1.15 I .20 0.00 6.19 2.00 9.33 10.60 10.02 9.25 7.37 6.36
96 99 83 86 86 100 27 76 -10 -25 -18 -9 13 25
Steroid specificity was determined by competition binding between radio-labelled testosterone and various unla~~1~ C18, Cl9 and C2l steroids to the 50 kDa testosteronebinding protein. Bach soiution contained 4nM of labeiled testosterone (5 &i [>H]testosterone in 15 ml solution). Competition for various steroids was performed by adding unlabelled steroid to 15 ml of the labelled solutions, each containing the 4 nM labelled testosterone. Equilibrium dialysis was performed at 4°C for 16-18 h.
I
1
I
I
20
I
I
I
40
I
80
TESTOSTERONE
CONCENTRATION
1
I
I
100 (nM)
Fig. 5. [3HFestosterone binding to the 50 kDa protein. Binding assays were performed at increasing t~tosterone con~ntrations as described in the Experimental section. (Insert) A Scatchard plot of testosterone binding by the 50 kDa protein. The plot has a correlation coefficient of 0.94 as determined by linear regression and a Kd of 16.7 nM.
Table 4. Effect of pH and temperature on binding of testosterone to the 50 kDa protein A.
B. PH
5.6 6.0 7.0 7.5 z 9:o 11.2
Binding (nmol/mg protein) Sample 1
Sample 2
1.2 1.3 1.2 1.2 t.2 1.3 1.3 0
1.3 I.3 I.3 1.2 I.3 1.2 I.2 0
Temperature (“C)
4 IO $ 37
Binding (nmol/mg protein) Sample 1
Sample 2
1.3 1.2 1.4 1.3 0.5
I.2 1.3 1.3 1.3 0.8
The binding assays were performed as described in the Experimental section except that 30mM sodium acetate-acetic acid buffer was used at pH 5.6; 30mM potassium phosphate buffer was used at PI-I 6.0 to 8.0; 30mM Tris-HCI was used at pH 8.5 and 9.0; and 30 mM potassium pho~~te-s~~urn hydroxide was used at pH 11.2.
JAMES E. THOMAS et al.
34
its early elution from the chromatofocusing
column.
Binding of steroid to the protein was not efficient; only 0.24.3 mol of testosterone were bound per mole of 50 kDa protein under the assay conditions used. We believe that this lack of efficiency in steroid binding may result from removal of the protein from its native environment within the double membrane of the bacterium. Such an environment would most likely influence both secondary and tertiary structure of the protein and consequently access of steroid to its binding site. Whether changes in this environment normally regulate binding of steroid to the 50 kDa protein is the subject of future studies. At least one other cationic testosterone-binding protein can be found in the 45,OOOg supernatant recovered during preparation of membrane vesicles [l 11. While similar in steroid specificity, charge and hydrophobicity, this 30 kDa protein is smaller in size than the 50 kDa protein reported here and has a different amino acid composition. Most notable, the 30 kDa protein possesses 8 cysteine residues whereas the 50 kDa protein reported in this study possesses only one. The presence of other steroid-binding proteins in crude extracts obtained from bacteria can influence our interpretation of the purification table (Table 1) obtained for the 50 kDa protein. While the numbers seen in the table typify those obtained during
purification, a low recovery of the 50 kDa protein may not only reflect loss of sample during manipulation of the protein, but purification of the protein away from other steroid binding proteins. Loss of steroid-binding activity due to purification of the protein away from its hydrophobic environment may also contribute to variations in specific activity of a purified sample. From these observations, we suggest that the relative abundancy of the 50 kDa protein is considerably less than indicated by the recovery data. Acknowledgements-This work was supported by a grant MT4425 from the Medical Research Council of Canada. We would like to thank Dr D. J. McKay for sequencing the protein. Also we wish to thank Mrs C. Friend for help in preparation of the manuscript. REFERENCES 1. Ames
2.
3.
4.
5.
G. F.-L. and Higgins C. F.: The organization, mechanism of action, and evolution of periplasmic transport systems. Trends Biochem. Sci. 8 (1983) 97-100. Hobson A. C., Weatherwax R. and Ames G. F.-L.: ATP-binding sites in the membrane components of histidine pe-rrnease, a periplasmic transport system. Proc. narl Acad. Sci. U.S.A. 81 (1984) 7333-7337. Higgins C. F., Hiles I. D., Whahey K. and Jamieson D. J.: Nucleotide binding by membrane components of bacterial periplasmic binding protein-dependent transport systems. EMBO J. 4 (1985) 1033-1040. Scarborough G. A.: Binding energy, conformational change, and the mechanism of transmembrane solute movements. Microbial. Rev. 49 (1985) 214-231. Watanabe M. and PO L.: Testosterone uptake by membrane vesicles of Pseudomonas testosreroni. Biochim. biophys. Acta 345 (1974) 419429.
6. Watanabe
M. and Watanabe H.: Periplasmic steroidbinding proteins and steroid transforming enzymes of
Pseudomonas testosteroni. J. steroid Biochem. 5 (1974) 439446. 7. Lefebvre Y., PO L. and Watanabe M.: The involvement of the electron transport chain in uptake of testosterone by membrane vesicles of Pseudomonas restosleroni. J. steroid Biochem. 7 (1976) 867-868. M.: The effect of carbonyl 8 Culos D. and Watanabe cyanide m-chlorophenylhydrazone on steroid transport in membrane vesicles of Pseudomonas tesrosteroni. J. steroid Biochem. 19 (1983) I12771 133. 9 Watanabe M., PO Sy L., Hunt D. and Lefebvre Y.:
Binding of steroids by a partially purified periplasmic protein from Pseudomonas testosferoni. J. steroid Biothem. 10 (1979) 2077213. IO Watanabe M., Phillips K. and Chen T.: Steroid-receptor in Pseudomonas iestosieroni released by osmotic shock. J. steroid Biochem. 4 (1973)613-62 1. 11 Francis M. McD. and Watanabe M.: Purification and characterization of a membrane-associated testosterone-binding protein from Pseudomonas festosreroni. Can. J. Biochem. Cell Biol. 61 (1983) 307-3 12. 12. Kaback H. R.: Bacterial membranes. In Methods in Enzvmologv (Edited bv W. B. Jakobv). Academic Press. New York; Vol. XXII (1971) pp. 99-120. 13. Watanabe M., Phillips K. and Watanabe H.: Induction of steroid-binding activity in Pseudomonas tesrosteroni. J. steroid Biochem. 4 (1973) 6233632. M.: Membrane-associated 14. Francis M. and Watanabe steroid-binding proteins of Pseudomonas tesfosteroni. Can J. Microbial. 27 (1981) 129G-1297. 15. Lefebvre Y. A., Schultz R., Groman E. V. and Watanabe M.: Localization of 3 and 17fi-hydroxysteroid dehydrogenase in Pseudomonas iestosteroni. J. steroid Biochem. 10 (1979) 5233528. M.: Partial purification and 16. Francis M. and Watanabe characterization of a membrane-associated steroidbinding protein from Pseudomonas lesrosteroni. Can. J. Biochem. 60 (1982) 7988803. 17. Bradford M. M.: A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analyr. Biothem. 72 (1976) 248 -254. proteins during 18. Laemmh U. K.: Cleavage of structural the assembly of the head of bacteriophage T4. Nature 227 (1970) 680-685. M. R. and Liu T.-Y.: Com19. Simpson R. J., Neuberger plete amino acid analysis of proteins from a single hydrolysate. J. biol. Chem. 251 (1976) 19361940. of cystine as cysteic acid. 20 Hirs C. H. W.: Determination In Methods in Enzymology (Edited by C. H. W. Hirs). Academic Press, New York, Vol. XI (1967) pp. 59-62. 21 McKay D. J., Renaux B. S. and Dixon G. H.: The amino acid sequence of human sperm protamine PI. Biosci. Rep. 5 (1985) 3833391. 22 Hagar D. A. and Burgess R. R.: Elution of proteins from sodium dodecyl sulfate-polyacrylamide gels, removal of sodium dodecyl sulfate, and renaturation of enzymatic activity: Results with sigma subunit of Escherichia coli RNA polymerase, wheat germ DNA topoisomerase, and other enzymes. Analyr. Biochem. 109 (1980) 76-86. 23. Wray W., Bouhkas T., Wray V. P. and Hancock R.: Silver staining of proteins in polyacrylamide gels. Analyt. Biochem. 118 (1981) 1977203. 24. Cleveland D. W., Fischer S. G., Kirschner M. W. and Laemmh U. K.: Peptide mapping by limited proteolysis In sodium dodecyl sulfate and analysis by gel electrophoresis. J. biol. Chem. 252 (1977) 1102~1106. G.: The attraction of proteins for small 25. Scatchard molecules and ions. Ann. N. Y. Acad. Sri. 51 (1949) 56&672.
Copyright 0
0022-4731/89 $3.00 + 0.00 1989 Pergamon Press plc
ISOLATION AND CHARACTERIZATION OF A 50 kDa TESTOSTERONE-BINDING PROTEIN FROM PSEUDOMONAS
TESTOSTERONZ
JAMESE. THOMAS*,ROGAYAHCARROLL,LUISA PO SY and MAMORUWATANABE Department of Medicine, Faculty of Medicine, The University of Calgary, Calgary, Alberta, Canada T2N 4Nl
(Received 18 March 1988)
Summary-A testosterone-binding protein (M, = 50,500) has been isolated from the Gram-negative bacterium Pseudomonas testosteroni. The protein was partially purified by a combination of ion exchange chromatography and chromatofocusing. Final purification was achieved by electroelution of the 50 kDa protein from SDS-polyacrylamide gels. Following renaturation from a diluted solution of guanidineHC1, specific binding of [3H]testosterone to the purified protein was observed. The native protein has a pI of 6.8. It appears to contain 428 amino acids, 39% of which are hydrophobic. There is only one cysteine residue. Both chymotrypsin and VS protease were used to produce peptide maps of the protein for use in future identification. The first 10 amino acids situated at the N-terminal of the protein were Ser-Pro-Phe-Asp-Leu-Arg-Pro-Leu-Se+Gly. Testosterone binding to the protein was saturable at _ 3.8 nmol/mg protein; the binding constant was -25 nM. Unlabelled testosterone, androstenedione, Sa-dihydrotestosterone and Sfi-dihydrotestosterone were able to compete for [3H]testosterone bound to the protein; 17/I-estradiol also competed for [3H]testosterone but to a lesser degree. Neither progesterone nor desoxycorticosterone competed for the testosterone-binding site. Binding of testosterone to the protein was stable at pH’s ranging from 5.5 to 9.0 and at various temperatures ranging from 4 to 30°C. The protein was unable to metabolize testosterone in either the presence or absence of the cofactor NAD.
active transport system [5] consisting of both periplasmic [6] and membrane-integrated [5,7,8] protein components. A periplasmic component has been partially purified by ammonium sulfate precipitation [9]. It possesses at least one high affinity testosterone-binding protein (& = 9.4 nM) and has a Svedberg coefficient of 4S [lo]. A membraneassociated steroid-binding protein has also been identified [ 111. This 34 kDa cationic testosteronebinding protein is released from. Kaback membrane vesicles [12] during centrifugation at 45,OOOg.It binds testosterone (& = 39 nM), and has 278 amino acids, 44% of which are hydrophobic. It also possesses 9 cysteine residues. In this paper, we report on a 50 kDa testosteronebinding protein which appears to be membraneassociated. Because of its location, high Kd value and inability to metabolize testosterone, we suggest that it is one of the protein components involved in an osmotic shock sensitive steroid transport system indigenous to P. testosteroni.
INTRODUCrION Gram-negative bacteria make use of a number of binding-protein systems in order to actively transport extracellular metabolites into the cell. These systems can be subdivided into those which are sensitive to osmotic shock and into those which are insensitive. Transport systems sensitive to osmotic shock require three or more proteins, at least one of which is a binding protein located in the periplasmic space of the cell [l]. Energy to drive these systems appears to come from hydrolysis of phosphate bonds associated with either ATP, acetyl phosphate or a similar compound [l-3]. In contrast, shock-insensitive transport systems consist of membrane bound carrier proteins energized by a chemiosmotic ion gradient similar to that proposed by Mitchell[4]. P. testosteroni is a Gram-negative bacterium capable of growth on certain Cl9 and C21 steroids. The mechanisms involved in transporting free extracellular steroid across the cell wall and membrane of the bacterium to the inside of the cell are essentially unknown. In order to elucidate these mechanisms, the proteins involved in the transfer process must first be identified. Work from this laboratory has shown that access to extracellular steroid occurs via an
EXPERIMENTAL Materials P. testosteroni, ATCC 11996, was obtained from American Type Culture Collection, Rockville, MD. Unlabelled steroids were obtained from Steraloids, Wilton, NH and from Sigma Chemical Co., St Louis, MO. Labelled [7-‘HItestosterone (25 Ci/mmol) was
*Present address: J. E. Thomas, Department of Biological Sciences, The University of Lethbridge, Alberta, Canada TIK 3M4.
Lethbridge, 21
28
JAMESE. THOMASet al.
obtained from New England Nuclear Coloration, Boston, MA. DEAE-Trisacryl is a product of LKB, Bromma, Sweden. Polybuffer Exchanger 94 and Polybuffer 96 are products of Pharmacia, Uppsala, Sweden. All other chemicals were purchased from either Sigma Chemical Co., St Louis, MO or Fisher Scientific, Edmonton, AB.
is directly proportional to the amount of steroid nonspecifically bound to the protein. Thus specific binding is calculated from the difference between total binding and nonspecific binding of steroid to the protein sampIe. Enzyme activity
Samples were assayed for their ability to metabolize testosterone, i.e. for the presence of vesicles Al-dehydrogenase and 3(17)fl-hydroxysteroid dehyas described drogenase. Fractions consisting of about 3-5 pg of testosteroni was grown P. column previously [ 131for 40 h at which point 0.5 g of testos- protein taken from the chromatofocusing terone (17~-hydroxy-~androsten-3-one} was added and 40-5Opg of protein obtained from the vesicle as a powder per liter of culture in order to induce the supematant were tested for the enzyme activity. The formation of a steroid transport and metabolic sys- latter was used as a control containing known dehytem within the bacteria. After an additional 40 h drogenase activity [15]. For each assay, protein was SOpCi period, cells were harvested by concentration with a added to a reaction mixture containing Millipore Pellicon Cassette System equipped with 10 [3H]testosterone mixed with 10 pg unlabelled testosmembranes each possessing an exclusion limit of terone, 0.5 M Tris-HCl pH = 9, in a final volume of 300,000 mol. wt. The cells were washed in 10mM 1 ml. Samples were incubated at 22°C for l-3 h and then assayed for steroid metabolites using the method of Tris-HCl, pH = 8 at 4”C, and the residual medium removed by cent~fugation at 12,OOOg for 10min. Francis and Watanabe[ 161. A need for cofactors was Membrane vesicles were prepared from whole cells tested by addition of 1 mM nicotinamide adenine dinuusing a method initially outlined by Kaback[l2] but cleotide (NAD) to samples as indicated in the text. as modified and described by Francis and General Watanabe[l4]. Under these conditions the protein of Protein concentrations were assayed by the Cointerest appeared to be released sporadically during osmotic shock in lysozyme-EDTA. Thus, in order to omassie blue method described by Bradford[l7], using bovine serum albumin as standard. facilitate recovery of the testosterone-binding proVertical slab gel el~trophoresis was performed in tein, EDTA was omitted from all solutions used dur0.2% (w/v) SDS using 7.5% (w/v) polyacrylamide ing preparation of the vesicles. Membrane-associated gels according to the method of Laemmli[lll]. proteins, including the testosterone-binding protein, Isoelectric focusing was performed on an LKB were found in the supernatant fraction recovered Multiphor 2117 using Ampholine PAG Plates, pH during final pelleting of the vesicles at 45,000g. 3.5-9.5 (LKB). Methods were as described for the kit. Binding assays The pH gradient was determined from p1 standard proteins obtained from Pharmacia. When radioBinding assays were performed by equilibrium labelled samples were run on an isoelectric focusing dialysis using a method modified from Watanabe et gel, each lane from the unfixed gel was cut into 2 mm af.[lO]. Assays were performed in a buffer consisting sections. Gels were solubilized in 250~1 of a 19:l of 30 mM Tris-HCl, pH = 7, containing 1 M KC1 in order to facilitate protein solubility. Each assay was (v/v) mixture of 50% hydrogen peroxide: 0.88% ammonium hydroxide and left overnight at 37°C. performed by placing no more than 1 ml of protein solution in each of two dialysis bags. One bag was The following day, 5000 U of catalase were added to each sample. After 30min (37”(Z), 250~1 of 10% added to 15 ml of buffer made 0.23 p M in unlabelled (w/v) ascorbic acid were added, the samples mixed testosterone plus 5 /tCi [3H]testosterone. The second with 5 ml of Ready-Solv HP and counted in a liquid dialysis bag was added to 15 ml of buffer made 23 p M in unla~lled testosterone ( IOO-fold excess) scintillation spectrometer. Amino acid analysis was performed on a Beckman plus 5 PCi [3H]testosterone. Solutions within each of 6300 amino acid analyzer. Samples were subjected to the two bags were allowed to dialyse to equilibrium methanesulfonic acid hydrolysis [191in order to idenfor 20 h at 4°C at which point IO&200 ,ul aliquots were removed from both inside and outside of the tify trypt~phan; performic acid oxidation [20] in order to identify methionine and cysteine residues; and bags, mixed with 5 ml Ready-Sok HP (Beckman Instruments, Inc., Palo Alto, CA) and counted in a sequential hydrochloric acid hydrolysis for 24,48 and liquid scintillation spectrometer. The difference in 72 h in 6 N HCl, 0.1% phenol, 0.1% thiodiglycolic acid at I lo”C, in vacua, in order to identify the dpms between the recovery of radioactivity from inside and from outside of the first dialysis bag is remaining amino acids. The amino acid analysis did equal to the total binding and is proportional to the not differentiate between either asparagine and aspartotal amount of steroid bound by the protein. In tic acid or glutamine and glutamic acid. Protein contrast, the difference between the recovery of radiosamples were sequenced using the method of McKay activity inside and outside of the second dialysis bag et af.[21]. Growth of organism and preparation
of membrane
Steroid-binding
protein
in P. testosteroni
29
-ao?
J+
I- 7sti. 4
- 7.0
xl
Fig. 1. Elution profile of the chromatofocusing column. The column was prepared and eluted as described each consisting of 5 ml, were collected and both the pH and absorbance at 280 nm determined for each. Aliquots of 100~1 were taken from each tube, mixed with 5 ml of Ready-Solv HP and counted in a liquid scintillation spectrometer. Typically, the testosterone-binding protein was eluted in the text. Fractions,
at a pH = 7.4, in fractions 4!j-50. RESULTS
Previous work from this laboratory has shown that the 45,000 g supematant obtained during membrane vesicle preparation contains a number of testosterone-binding proteins [11J. This fraction was precipitated in 3.9 M ammonium sulfate for 45 min (4°C) and then centrifuged at 37,000g for 30 min. The pellet was dissolved in N 50 ml of 5 mM EDTA, 25 mM Tris-HCl, pH = 7.0 (Buffer A) and dialysed into Buffer A. In order to remove anionic contaminants from the sample, it was applied to a column (2.5 x 45 cm) of DEAE-Trisacryl equilibrated with Buffer A. A nonbinding fraction containing testosterone-binding proteins was eluted with 500 ml of Buffer A [14]. This fraction was pr~ipitated in 3.9 M ammonium sulfate for 45min (4”C), and then centrifuged for 30 min at 37,000g. The pellet was dissolved and then dialysed into 25 mM ethanolamin*HCl, pH = 9.0 (Buffer B). In order to identify testosterone-binding proteins within this
Purification
fraction, proteins were incubated in lO@i of [‘H]testosterone overnight (4°C). After this, the solution was added to a chromatof~using column (1 x 40 cm) of Polybuffer Exchanger 94 ~uilibrated with Buffer B and eluted with 415ml Polybuffer 96-HCl, pH = 7.0, diluted 1: 12 with deionized water (Fig. 1). Free steroid was eluted from the column in peak 1, typically in tubes 5 through 15, whereas the steroidbinding protein was eluted as a single peak when the pH of the column reached -7.4. Protein from this peak was precipitated in 5.3 M ammonium sulfate for 45 min (4”(Z), then centrifuged at 37,000 g for 30 min. The pellet was dissolved and then dialysed into Buffer A. This fraction constituted the semi-purified protein fraction used in all subsequent experiments. The testosterone-binding protein from this fraction was purified N IOO-fold (Table 1). Protein fractions obtained during the purification procedure were examined by SDS-polyacrylamide gel electrophoresis (Fig. 2). Following chromatofocusing one major protein band M, = 50,500 rf: 3,600 (n = 10) was recovered. Major contaminants of this fraction had a M, less than 30,000.
Table 1. A typical purification table for the 50 kDa testosterone-binding protein Specificactivity Total Tota activity (pm01 testosterone protein (pm01 testosterone Purification
Recovery (%)
step’
(mg)
bound)?
bound/mg protein)
Vesicle su~matant 80% (NH&SO, saturation DEAE-Trisacryi Chromatofocusing
346.4
18,543
52
1
100
280.9
16,968
59
1
89
21.1 0.7
9,855 3,386
462 5,121
9 95
50 18
‘The (NH&SO, precipitate examined in step 2 was resuspended and dialysed into 25 mM Tris-HCI, pH = 7.0 before doing a steroid binding assay. The non-binding fraction obtained from the DEAE-Trisacryl column was precipitated in 3.9 M (NH&SO,, then resuspended and dialysed into 25 mM ethanolamine-HCI, pH = 9.0, before being assayed for total binding activity. The fraction obtained from the chromatofocusing column was precipitated in 5.3 M (NH&SO,, then resuspended and dialysed into 25 mM Tris-HCI, pH = 7.0, before assaying total binding activity. +Binding assays were performed as described in the Experimental section. All assays were performed using saturating amounts of testosterone.
JAMES E. THOMAS et al.
30
Mr (kDol
92.5 -
66.2 -
Fig. 2. S~S-polyacrylamide gel eleetrophoresis of various fractions obtained during puri~cation of the 50 kDa testosterone-binding protein. Purification of the 50 kDa protein was carried out as described in the text. Fractions were electrophoresed in 0.2% SDS on a 7.5% polyacrylamide gel. Gels were stained for 30 min in 0.4% (w/v) Coomassie blue R-250 made up in 40% (v/v) methanol, 5% (v/v) glacial acetic acid in deionized water and destained in 10% (v/v) methanol, 15% (v/v) glacial acetic acid in deionized water. Lanes 1 and 6: Bio-Rad low mol. wt standards; lane 2: 45.0001: vesicle supernatant: lane 3: (NH&SO, precipitate; lane 4: cationic fraction from the DEAE-Trisacryl column; lane 5: testosterone-binding fraction from the chromatofocusing column.
In order to show that the 50 kDa protein binds testosterone, it was purified by SDS-polyacrylamide gel electrophoresis and renatured after the method of Hagar and Burgess[22]. Semi-purified samples from the chromatofocusing column were run on 3 mm preparatory SDS-polyacrylamide gels, rinsed in I mM dithiothreitol (DTT) prepared in deionized water (4°C) and stained in 0.25 M KC1 containing 1 mM DTT (4 C). Under these conditions the 50 kDa protein appears as an opaque band on the gel after about 5 min. This band was cut out of the gel, rinsed with deionized water containing 1 mM DTT (4”‘C) and homogenized in 0.1% SDS, 10 mM Tris-acetate, pH = 8.6. The protein was electroeluted from the gel for 4 h at 3 Watts, constant current, in an Isco Electroelution Apparatus using 0. I % SDS, 40 mM Tris-acetate, pH = 8.6 in the buffering wells. After this, the sample was precipitated with 4 vol of acetone (-20‘ C) for 30 min in 15 ml glass Corex tubes and the precipitate centrifuged at 37,000 g for 30 min. The pellet was dissolved in 100 ~1 of 6 M guanidine-HCI made up in a dilution buffer consisting of 20% glycerol (v/v), 0.15 M NaCI, 1 mM DTT, 0.1 mM EDTA, 50 mM Tris-HCl, pH = 8.0, and left at room temperature for 20 min, at which point it was diluted 50-fold with dilution buffer and left at room temperature overnight. The solution was concentrated to I ml in an Amicon Stirred Cell concentrator using a Millipore PTTK membrane with a 30,000 mol. wt
exclusion limit and then assayed for testosterone binding by equilibrium dialysis. Specific binding of testosterone to the purified protein was I .7 nmol/mg protein. Only one SOkDa band could be detected following SDS-polyacrylamide gel electrophoresis and silver staining after the method of Wray et a!.[23].
The 50 kDa testosterone-binding protein was electroeluted from SDS-polyacrylamide gels as described above and its amino acid composition determined (Table 2). The protein appears to consist of 428 amino acids, 39% of which are hydrophobic. Only one cysteine residue could be detected in the protein. Based on this composition, it has an apparent mol. wt of 45,408. This discrepancy in molecular weight with results obtained by SDS-gel electrophoresis may result from secondary modifications to the protein. In two separate experiments, the N-terminal amino acid sequence for the protein was shown to be Ser-ProPhe-Asp-~eu-Arg-Pro-Leu-Ser-Gly. Only one protein was detected in these experiments confirming that the electrophoreticaily extracted 50 kDa protein was purified to homogeneity. Peptide mapping
The 50 kDa testosterone-binding protein was peptide-mapped after the method of Cleveland et a1.[24]. Treatment of the protein with various concentrations of both V8 protease and chymotrypsin produced distinct degradation profiles (Fig. 3). Note that multiple banding of the untreated 50 kDa protein appeared to result from binding of various amounts of SDS to the protein in a 0.1% SDS gel. This phenomenon was not observed when gels were run in 0.2% SDS. Characterization
Isoelectric focusing was performed on a sample of semi-purified protein labelled overnight with Table 2. Amino acid composition
of the 50 kDa testosterone-binding protein
Amino ASX Thr Ser Glx Pro GlY Ala cys Val Met Ile Leu TY~ Phe His Lys TW
A&
acid
Amino acid ratio 4.481 3.113 3.311 5.073 3.522 9.400 3.847 0.129 3.272 0.74x 2.917 3.831 2.126 2.617 0.832 2.407 0.763 2.856
4028.1 2426.4 2264.6 5034.9 262 I .7 4161.0 2133.0 103.2 2480.0 787.2 2603.6 3396.0 2611.2 2944.0 823.2 2433.8 1117.2 3436.4
Amino acid composition* 35 24 26 39 27 73 30 25 6
23 30 16 20 19 22
by assuming that the protem conraIned 1 cysteine residue.
*The amino acid composition
testosterone-binding
Projected amino acid formula weight (kDa)
was determined
Steroid-binding protein in P.
31
testosteroni
Mr 92.5k 66.2
31.0
21.5
74.4
1
2
3
4
5
6
7
8
9
10
Fig. 3. Peptide maps of the 50 kDa testosterone-binding protein. The binding protein was peptide mapped in 0.1% SDS on a 15% polyacrylamide gel after the method of Cleveland et u1.[24].Characteristic peptide profiles were obtained for both ~hymotrypsin (lanes S-IO) and V8 protease (lanes 46). Lane 1: Bio-Rad low mol. wt standards; lane 2: non-digested protein; lane 3: 0.5 pg V8 protease; lanes 4-6: protein mixed
with either 0.5, 0. I or 0.025 pg V8 protease; lane 7: 0.5 pg ch~otrypsin; lanes &IO: protein mixed with either 0.5. 0.1 or 0.025 pg chymotrypsin. [3H]testosterone (Fig. 4). While several protein bands were observed after staining with Coomassie blue (Fig. 4B), only one band at pH = 6.8 was radiolabelled (Fig. 4A). The second peak represents unbound [3H]testosterone which did not migrate from the point at which the sample was initially placed on the gel. This peak could be moved anywhere on the gel without affecting the position of the radiolabelled protein peak (data not shown). This shows, (1) that the semi-purified protein contains only one species of testosterone-binding protein and, (2) that the p1 of the native 50 kDa testosterone-binding protein is 6.8 as opposed to 7.4 as indicated by the chromatofocusing column. Steroid specificity was determined by competition between [3H]testosterone and various other nonlabelled Cl 8, Cl9 and C21 steroids for the protein binding site (Table 3). When the semi-purified protein fraction was labelled with [3H]testosterone, all unlabelled Cl9 steroids tested competed for the labelled binding site; testosterone was most efficient in competing for the labelled steroid while 5adihydrotestosterone, androstenedione and SFdihydrotestosterone were able to compete for the binding site, but to a lesser degree. The Cl8 steroid, 17/3-estradiol, also competed for the binding site, but at a lower level. The C21 steroids, progesterone and desoxycorticosterone, did not compete for the labelled binding site. Binding of testosterone to the 50 kDa protein was saturable (Fig. 5) and exhibited a Kd of 16.7nM (insert) and 33.3 nM based on two separate experiments. The Scatchard plot [25] (Fig. 5, insert) shows
that the 50 kDa protein is capable of binding -4.5 nmol testosterone/mg protein under the assay conditions used. This value correlates with that seen for protein obtained from the chromatofocusing column (Table 1); on average - 5.1 nmol testosterone was bound per mg protein obtained by chromatofocusing. This value represents about 0.2-0.3 mol of testosterone bound per mole of 50 kDa protein. Based on this observation, only a small proportion of the purified protein is capable of binding steroid under the assay conditions used. This will be discussed later. However, because the Scatchard plot (Fig. 5, insert) appears linear, we suggest that only one high affinity binding site for testosterone is located on the 50 kDa protein. Binding of [3H]testosterone to the 50 kDa protein was stable at pH’s ranging from 5.5 to 9.0 (Table 4). No steroid binding was observed at a pH of 1I .2. The binding of [3H]testosterone to the protein was also stable at temperatures ranging from 4 to 30°C (Table 4). However, at 37°C a 47% decrease in binding of [3H]testosterone to the protein was observed. This result is consistent with obervations which show that growth of the bacterium at 37°C is inhibited f13]. Incubation of vesicle supernatant protein at 22°C with 13H]testosterone resulted in metabolism of the testosterone to androstenedione and A1.4androstadienedione. Within 1 h, about 41% of the steroid was converted to androstenedione and 50% to A’*4androstadienedione. These results show that both Al-dehydrogenase and 3(17)P_hydroxysteroid dehydrogenase are present in the vesicle supernatant fraction. Francis and Watanabe[ 141have shown that
JAMESE. THOMASet al.
32
w
-
6.65
-
8.45
Wa@W -
6.15
-
7.35
-
6.65
-
6.55
).
Binding + protein
- 5.65 - 5.20
DISTANCE
FROM CATHODE
(cm)
Fig. 4. Isoelectric focusing of the semi-purified testosterone-binding protein. The 50 kDa protein obtained by chromatofocusing, was labelled overnight with [3H]testosterone and applied to an LKB-PAG plate, pH 3.5-9.5. Isoelectric focusing was performed as described in the Experimental section. Only one testosterone-binding protein (Panel A) was observed, 4.5cm from the cathode; the free radiolabelled steroid seen l-3 cm from the cathode did not migrate from the position on the plate where the sample was placed at the beginning of the run. The binding protein had a p1 of 6.8 (Panel B) when compared to pH standards run simultaneously.
A 1-dehydrogenase is removed from the vesicle supernatant fraction by use of anion exchange chromatography as performed here. When fractions eluted from the chromatofocusing column were tested for the presence of 3(17)8-hydroxysteroid dehydrogenase, we found that a peak in activity was eluted from the column when the pH of the elution buffer reached about 7.6 (data not shown). In the presence of both [3H]testosterone and 1 mM NAD, samples taken from this peak were able to convert 96% of the [3H]testosterone to androstenedione. No specific binding of testosterone to this protein fraction was obtained using the assay conditions outlined in the Experimental section. In contrast, when protein purified by electrophoresis and subsequently renatured as described in the text was mixed with [3H]testosterone and 1 mM NAD no degradation products of testosterone were found. This indicates that the 50 kDa protein has no enzymatic activity directed towards the breakdown of testosterone.
DISCUSSION
P. testosteroni
is a Gram-negative
bacterium
capa-
ble of growth on a number of Cl9 and C21 steroids. In this study, we have purified a Sb kDa protein which appears to possess a single high affinity site, specific for androgens. Testosterone, androstenedione, Scr-dihydrotestosterone and Sfi-dihydrotestosterone were all able to compete for this site on the protein. In contrast, neither of the C21 steroids, progesterone or desoxycorticosterone, exhibited competition for the testosterone-binding site supporting previous observations [IO] that a separate class of C21 steroid binding proteins exist within the bacterium. The amino acid composition indicates that about 72% of the protein is uncharged, consisting of both hydrophobic and neutral amino acids. This is consistent with data obtained by isoelectric focusing which indicates that the protein has a p1 = 6.8; near neutral. This lack of charge on the native protein may be the cause of
Steroid-binding protein in P. fesiosteroni
33
Table 3. Steroid specificity Steroid concentration fnM)
Steroid
nmol testosterone bound/ma urotein
4 69 341 66 331 69 344 69 344 64 318 61 303 73 367
Testosterone
Androstenedione Sa-Dihydrotestosterone 5~-~hyd~testosterone Progesterone Desoxycorticosterone 17/I-Estradiol
% Competition
8.41
0
0.33 0.08 1.48 1.15 I .20 0.00 6.19 2.00 9.33 10.60 10.02 9.25 7.37 6.36
96 99 83 86 86 100 27 76 -10 -25 -18 -9 13 25
Steroid specificity was determined by competition binding between radio-labelled testosterone and various unla~~1~ C18, Cl9 and C2l steroids to the 50 kDa testosteronebinding protein. Bach soiution contained 4nM of labeiled testosterone (5 &i [>H]testosterone in 15 ml solution). Competition for various steroids was performed by adding unlabelled steroid to 15 ml of the labelled solutions, each containing the 4 nM labelled testosterone. Equilibrium dialysis was performed at 4°C for 16-18 h.
I
1
I
I
20
I
I
I
40
I
80
TESTOSTERONE
CONCENTRATION
1
I
I
100 (nM)
Fig. 5. [3HFestosterone binding to the 50 kDa protein. Binding assays were performed at increasing t~tosterone con~ntrations as described in the Experimental section. (Insert) A Scatchard plot of testosterone binding by the 50 kDa protein. The plot has a correlation coefficient of 0.94 as determined by linear regression and a Kd of 16.7 nM.
Table 4. Effect of pH and temperature on binding of testosterone to the 50 kDa protein A.
B. PH
5.6 6.0 7.0 7.5 z 9:o 11.2
Binding (nmol/mg protein) Sample 1
Sample 2
1.2 1.3 1.2 1.2 t.2 1.3 1.3 0
1.3 I.3 I.3 1.2 I.3 1.2 I.2 0
Temperature (“C)
4 IO $ 37
Binding (nmol/mg protein) Sample 1
Sample 2
1.3 1.2 1.4 1.3 0.5
I.2 1.3 1.3 1.3 0.8
The binding assays were performed as described in the Experimental section except that 30mM sodium acetate-acetic acid buffer was used at pH 5.6; 30mM potassium phosphate buffer was used at PI-I 6.0 to 8.0; 30mM Tris-HCI was used at pH 8.5 and 9.0; and 30 mM potassium pho~~te-s~~urn hydroxide was used at pH 11.2.
JAMES E. THOMAS et al.
34
its early elution from the chromatofocusing
column.
Binding of steroid to the protein was not efficient; only 0.24.3 mol of testosterone were bound per mole of 50 kDa protein under the assay conditions used. We believe that this lack of efficiency in steroid binding may result from removal of the protein from its native environment within the double membrane of the bacterium. Such an environment would most likely influence both secondary and tertiary structure of the protein and consequently access of steroid to its binding site. Whether changes in this environment normally regulate binding of steroid to the 50 kDa protein is the subject of future studies. At least one other cationic testosterone-binding protein can be found in the 45,OOOg supernatant recovered during preparation of membrane vesicles [l 11. While similar in steroid specificity, charge and hydrophobicity, this 30 kDa protein is smaller in size than the 50 kDa protein reported here and has a different amino acid composition. Most notable, the 30 kDa protein possesses 8 cysteine residues whereas the 50 kDa protein reported in this study possesses only one. The presence of other steroid-binding proteins in crude extracts obtained from bacteria can influence our interpretation of the purification table (Table 1) obtained for the 50 kDa protein. While the numbers seen in the table typify those obtained during
purification, a low recovery of the 50 kDa protein may not only reflect loss of sample during manipulation of the protein, but purification of the protein away from other steroid binding proteins. Loss of steroid-binding activity due to purification of the protein away from its hydrophobic environment may also contribute to variations in specific activity of a purified sample. From these observations, we suggest that the relative abundancy of the 50 kDa protein is considerably less than indicated by the recovery data. Acknowledgements-This work was supported by a grant MT4425 from the Medical Research Council of Canada. We would like to thank Dr D. J. McKay for sequencing the protein. Also we wish to thank Mrs C. Friend for help in preparation of the manuscript. REFERENCES 1. Ames
2.
3.
4.
5.
G. F.-L. and Higgins C. F.: The organization, mechanism of action, and evolution of periplasmic transport systems. Trends Biochem. Sci. 8 (1983) 97-100. Hobson A. C., Weatherwax R. and Ames G. F.-L.: ATP-binding sites in the membrane components of histidine pe-rrnease, a periplasmic transport system. Proc. narl Acad. Sci. U.S.A. 81 (1984) 7333-7337. Higgins C. F., Hiles I. D., Whahey K. and Jamieson D. J.: Nucleotide binding by membrane components of bacterial periplasmic binding protein-dependent transport systems. EMBO J. 4 (1985) 1033-1040. Scarborough G. A.: Binding energy, conformational change, and the mechanism of transmembrane solute movements. Microbial. Rev. 49 (1985) 214-231. Watanabe M. and PO L.: Testosterone uptake by membrane vesicles of Pseudomonas testosreroni. Biochim. biophys. Acta 345 (1974) 419429.
6. Watanabe
M. and Watanabe H.: Periplasmic steroidbinding proteins and steroid transforming enzymes of
Pseudomonas testosteroni. J. steroid Biochem. 5 (1974) 439446. 7. Lefebvre Y., PO L. and Watanabe M.: The involvement of the electron transport chain in uptake of testosterone by membrane vesicles of Pseudomonas restosleroni. J. steroid Biochem. 7 (1976) 867-868. M.: The effect of carbonyl 8 Culos D. and Watanabe cyanide m-chlorophenylhydrazone on steroid transport in membrane vesicles of Pseudomonas tesrosteroni. J. steroid Biochem. 19 (1983) I12771 133. 9 Watanabe M., PO Sy L., Hunt D. and Lefebvre Y.:
Binding of steroids by a partially purified periplasmic protein from Pseudomonas testosferoni. J. steroid Biothem. 10 (1979) 2077213. IO Watanabe M., Phillips K. and Chen T.: Steroid-receptor in Pseudomonas iestosieroni released by osmotic shock. J. steroid Biochem. 4 (1973)613-62 1. 11 Francis M. McD. and Watanabe M.: Purification and characterization of a membrane-associated testosterone-binding protein from Pseudomonas festosreroni. Can. J. Biochem. Cell Biol. 61 (1983) 307-3 12. 12. Kaback H. R.: Bacterial membranes. In Methods in Enzvmologv (Edited bv W. B. Jakobv). Academic Press. New York; Vol. XXII (1971) pp. 99-120. 13. Watanabe M., Phillips K. and Watanabe H.: Induction of steroid-binding activity in Pseudomonas tesrosteroni. J. steroid Biochem. 4 (1973) 6233632. M.: Membrane-associated 14. Francis M. and Watanabe steroid-binding proteins of Pseudomonas tesfosteroni. Can J. Microbial. 27 (1981) 129G-1297. 15. Lefebvre Y. A., Schultz R., Groman E. V. and Watanabe M.: Localization of 3 and 17fi-hydroxysteroid dehydrogenase in Pseudomonas iestosteroni. J. steroid Biochem. 10 (1979) 5233528. M.: Partial purification and 16. Francis M. and Watanabe characterization of a membrane-associated steroidbinding protein from Pseudomonas lesrosteroni. Can. J. Biochem. 60 (1982) 7988803. 17. Bradford M. M.: A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analyr. Biothem. 72 (1976) 248 -254. proteins during 18. Laemmh U. K.: Cleavage of structural the assembly of the head of bacteriophage T4. Nature 227 (1970) 680-685. M. R. and Liu T.-Y.: Com19. Simpson R. J., Neuberger plete amino acid analysis of proteins from a single hydrolysate. J. biol. Chem. 251 (1976) 19361940. of cystine as cysteic acid. 20 Hirs C. H. W.: Determination In Methods in Enzymology (Edited by C. H. W. Hirs). Academic Press, New York, Vol. XI (1967) pp. 59-62. 21 McKay D. J., Renaux B. S. and Dixon G. H.: The amino acid sequence of human sperm protamine PI. Biosci. Rep. 5 (1985) 3833391. 22 Hagar D. A. and Burgess R. R.: Elution of proteins from sodium dodecyl sulfate-polyacrylamide gels, removal of sodium dodecyl sulfate, and renaturation of enzymatic activity: Results with sigma subunit of Escherichia coli RNA polymerase, wheat germ DNA topoisomerase, and other enzymes. Analyr. Biochem. 109 (1980) 76-86. 23. Wray W., Bouhkas T., Wray V. P. and Hancock R.: Silver staining of proteins in polyacrylamide gels. Analyt. Biochem. 118 (1981) 1977203. 24. Cleveland D. W., Fischer S. G., Kirschner M. W. and Laemmh U. K.: Peptide mapping by limited proteolysis In sodium dodecyl sulfate and analysis by gel electrophoresis. J. biol. Chem. 252 (1977) 1102~1106. G.: The attraction of proteins for small 25. Scatchard molecules and ions. Ann. N. Y. Acad. Sri. 51 (1949) 56&672.