- Email: [email protected]
Adsorption of proteins from the Spiroplasma citri cell membrane by magnesium lauroyl-sarcosinate crystals
BIOCHIMIE, 1981, 63, 177-186.
Adsorption of proteins from the Spiroplasma citri cell membrane by magnesium lauroyl-sarcosinate crystals. Henri WR()BLEWSKI *~, Rende B U R L O T * and Daniel T H O M A S ** (Refu le 2-8-1980, acceptd le 10-12-1980).
* Laboratoire de Biologie Cellulaire ** and Laboratoire de Cytologie expdrimentale, Universitd de Rennes, Complexe de Beaulieu, A venue du Gdndrat Leclerc, 35042 Rennes COdex (France).
R~sum~.
Summary.
Les interactions entre le Sarkosyl (lauroyl sareosinate de sodium), les ions M g 2÷ et la membrane de Spiroplasma cirri ont ~td analysdes au moyen de techniques microscopiques et de techniques glectrophordtiques. Cette dtude a dtd rdalisde dans des conditions oh les protdines membranaires n'dtaient apparemment pas extraites par le d~tergent (rapport molaire de M g C l J S a r k o s y l = 0,5).
Interactions between the ionic detergent Sarkosyl (sodium lauroyl sarcosinate), M g ~+ions, and the Spiroplasma citri cell membrane were analyzed microscopically and electrophoreticalIy. Studies were performed under conditions where membrane proteins were apparently not released from the membrane by the detergent (molar ratio of MgCle/ Sarkosyl = 0.5).
Bien que la membrane de S. citri interfbre dans une certaine mesure avec le phdnomkne de cristallisation, les cristaux de Sarkosyl-Mg 2+ se /orment inddpendemment de la sdquence d'addition des trois composants. Ceci s'accompagne d'une dd~intdgration de la structure membranaire et de I'adsorption de composants membranaires par les cristaux. La fraction de protdines lide aux cristaux contient la ma]oritd des polypeptides prdsumds intrinsbques, dont la spira~ine (une protdine amphiphile), et plusieurs polypeptides extrinskques. Les analyses dlectrophordtiques et immunodlectrophordtiques indiquent que les bandes-M ([ractions de cristaux chargds de matdriel membranaire), prdpardes ?t partir de cellules et celles prdpardes ~ partir de membranes isoldes de S. citri, possbdent des compositions similaires en protdines. Ces rdsu!tats montrent que la membrane de S. citri, contrairement aux membranes bactdriennes, n'est pas protdgde par les ions M g 2+ des ef]ets du Sarkosyl.
Although the S. citri membrane interfered with the crystallization phenomenon to some extent, the formation of Sarkosyl-Mg 2- crystals occurred regardless to the sequence o/ addition of the three components. Concomitantly the structure of the membrane disintegrated and membrane components were adsorbed to the crystal sur[aces. The membrane protein fraction bound to the crystals was composed of the ma]ority o / t h e putatively intrinsic polypeptides, including the amphiphilic protein spiralin, and several extrinsic polypeptides. The polypeptide compositions of M-bands (crystal fractions loaded with membrane material) prepared from S. citri cells" and ]rom isolated S. citri membranes were similar, as shown by sodium dodecyl-sulfate electrophoresis and crossed immunoelectrophoresis. These results show that, the S. citri cell membrane, in contrast to bacterial membranes, is not protected from the effect of Sarkosyl by Mg 2+ ions.
Abbreviations :
Introduction.
CIE, crossed immunoelectrophoresis ; CM-bands, Mbands prepared from cells; EDTA, ethylene diamine tetraacetate ; I, ionic strength ; MM-bands, M-bands prepared from isolated membranes ; SDS, sodium dodecylsulfate; SDS-PAGE, sodium dodecyl-sulfate polyacrylamide gel electrophoresis. <> To whom all correspondence should be addressed.
For practical reasons which are self-evident, a low sedimentation coefficient is the most c o m m o n ly used criterion of the solubilization of membrane proteins, e.g. by a detergent [1]. Specifically, after centrifugation for about 1 h at 100,000 to 14
H. Wr6blewski and coll.
178
2 0 0 , 0 0 0 × g in a m e d i u m with a density close to that of water, proteins recovered in the supernatant are considered as those which are solubilized or at least extracted (see ref. [1] for a criticism of the terminology). Conversely, those recovered in the pellet are considered to be the proteins which are n o t solubilized u n d e r the conditions of the particular experiment. T h e exclusive use of the above m e n t i o n e d criterion m a y be misleading in some circumstances. F o r example, w h e n the m e m b r a n e of the helical m y c o p l a s m a Spiroplasrna citri [2] is treated with the ionic detergent Sarkosyl (sodium lauroyl-sarcosinate) i n the presence of M g 2+ ions, some proteins are recovered in the residue a n d the r e m a i n d e r in the s u p e r n a t a n t . T h e ratio of the two fractions is a f u n c t i o n of the Mg2+/Sarkosyl ratio. T h u s when this ratio is ~ 0.5, all the proteins are quantitatively recovered in the residue [3]. It becomes tempting to conclude, as was done for some bacterial m e m b r a n e s [4], that M g 2+ ions protect the m e m b r a n e against the effect of Sarkosyl. This would necessarily m e a n that the structure of the m e m b r a n e is preserved u n d e r these conditions. We have, however, suggested a n o t h e r plausible explan a t i o n : the proteins are extracted from the m e m b r a n e a n d sediment u p o n centrifngation as proteincrystal complexes [3]. A c c o r d i n g to this hypothesis the m e m b r a n e would, of course, be m o r e or less disintegrated.
duced to 5 per cent (v/v). The cells were harvested by centrifugation at 15,000 × g for 15 min at 4oc, dispersed in 0.1 M Tris-HC1 buffer (pH 8.0) and disrupted by sonication at 20 kcycles for 2 × 1 rain at 0°C. The membranes were pelleted by centrifugation at 40,000 x g for 1 h at 4°C and extensively washed with 0.1 M TrisHC1 buffer (pH 8.0).
T h e purpose of the present investigation was to resolve this problem, in particular b y a m o r p h o logical analysis of the material. Electrophoretic a n d i m m u n o e l e c t r o p h o r e t i c studies were perform e d to characterize the S. citri m e m b r a n e p r o t e i n fraction recovered in <>. This preparative procedure is of great practical i m p o r t a n c e , since it has b e e n used to purify n u c l e o i d m e m b r a n e complexes from prokaryotic cells [5-9]. It has also b e e n exploited for the isolation of a n u c l e a r m e m brane-associated D N A complex in m a m m a l i a n cells [10].
SDS-PAGE was performed in 0.1 M Tris-HC1 buffer (pH 8.0) containing 10 mM SDS. Compositions of the lower gel and of the upper (pre-separaing) gel were respectively T = 6 per cent and C = 5 per cent, and T = 4 per cent and C = 2 per cent. The nomenclature of Hjert~n [12] is used to describe gel composition : percentage T is the total monomer (acrylamide + N, N'-methylene bisacrylamide) concentration, and percentage C is the percentage of N, N'-methylene bisacrylamide relative to the total monomer concentration.
Material
and Methods.
Chemicals. Sodium lauroyl-sarkosinate (Sarkosyl), acrylamide, and N-N'-methylene bis acrylamide were purchased from Sigma (U.S.A.). Sodium dodecyl sulfate (SDS) was from Merck (Germany), and agarose from Bio-Rad (U.S.A.).
M-band #actionation [5]. Membranes or intact cells were treated with Sarkosyl in the presence of Mg e+ ions. To 0.5 ml of Tris-HCl buffer (pH 8.0) containing 10 mg of membranes (5 mg of membrane protein) were added 0.25 ml of 0.4 M Sarkosyl and 0.25 ml of 0.2 M MgC12, both in distilled water. After mixing, the solutions were left for 1 h at 4°C. Each sample was then layered on top of a discontinuous sucrose gradient : 2 ml of 60 per cent sucrose, 13 ml of 40 per cent sucrose, and 13 ml of 15 per cent sucrose in 50 mM Tris-HC1 buffer (pH 8.0) containing 50 mM MgC1e. The gradients were centrifuged at 32,000 × g in a Spinco SW 25 rotor for 40 rain at 4°C. M-bands containing the Mge+-Sarkosyl crystals were visible at the interface of the 15 per cent and 40 per cent sucrose solutions. The material containing M-bands was dialyzed for 5 days at 4°C against 10 mM Tris-HCl buffer (pH 8.0) containing 20 mM Sarkosyl and 1 mM EDTA. The buffer was changed daily. The sample was lyophilized and again solubilized with 0.1 M Tris-HCl buffer (pH 8.0) in order to obtain a 0.1 M Sarkosyl solution containing 5 mg protein/ml.
SDS-polyacrylamide gel electrophoresis (SDS-PAGE).
Crossed immunoelectrophoresis (CIE). CIE [13, 14] was performed in the presence of detergent [15-18] in 1 mm thick agarose gels cast on 85 × 100 mm glass plates. The agarose concentration was 1 per cent (w/v) in veronal buffer (pH 8.6), I = 0,03 ; 17 mM Sarkosyl was included in the catholyte and in the gel of the first direction in order to prevent protein precipitation [3,26]. Antiserum containing antibodies against the S. cirri membrane was prepared as described [11]. Other details concerning CIE were described in a previous publication [19].
Light microscopy. Culture o] cells and membrane preparation. Spiroplasma cirri strain C189 (ATCC 27665) was grown as previously described [11] except that the concentration of foal serum in the culture medium was re-
BIOCHIMIE, 1981, 63, n ° 3.
Crystal suspensions were observed without any treatment under cover slips in a light microscope and were photographed. Phase contrast and dark field optics were used.
Membrane protein adsorption by sarkosyl-Mg ~÷ crystals. Electron microscopy. Sarkosyl-Mg 2+ crystals, either alone or loaded with membrane proteins, were pelleted by centrifugation (180,000 X g, 60 min, 4°C), fixed with 2.5 per cent glutaraldehyde, and postfixed with 2 per cent Os 04 in 0.1 M phosphate buffer (pH 7.2). The pellets were then dehy-
179
drated through a graded series of acetone and embedded in an araldite-epon mixture. Ultra-thin sections were stained with uranyl acetate and lead citrate [20]. For negative staining, drops of crystal suspensions were placed on carbonated collodion coated copper grids for 3 rain. After removal of excess fluid, the material was
B
,,L
FIG. 1 . -
Light microscope views o/ Sarkosyl- Mg2"crys tals. A) Crystals obtained by adding MgCI2 (50 mM final concentration) to a solution of 0. I M Sarkosyl in 50 m M Tris-HC1 buffer, pH 8.0. B) Preformed crystals mixed with S. citri membranes (10 mg protein/ml) and observed after incubation for 6 hours at room temperature. C) Crystals obtained after addition of MgC1._. (50 mM final concentration) to S. citri membranes (I0 mg protein/ml) treated for 1 hour with 0.1 M Sarkosyl in 0.05 M TrisHC1 buffer pH 8.0. The crystals were observed after incubation for 6 hours at room temperature. Phase contrast optics were used to observe the crystals in all samples. B I O C H I M I E , 1981, 68, n ° 3.
H.
180
Wr6blewski
stained for 20 sec with 2 per cent sodium phosphotungstate (pH 7.2). Excess fluid was again removed and the preparations were left to dry overnight at 32°C before observation.
,
500
a n d coll. Other techniques. Sarkosyl was titrated as a detergent-methylene blue complex extracted with chloroform, as described fer the
nm
FIG. 2. - - Electron mtcroscope views o] S. citri membranes and Sarkosyl-Mg 2+ crystals negatively stained with 2 per cent sodium phosphotungstate. A) Membrane vesicles obtained from S. cirri cells lysed by sonication and washed with 0.1 M Tris-HC1 buffer pH 8.0. B) Crystal of Sarkosyl-Mg2+ in a mixture of preformed crystals and S. citri membranes (see fig. 1 B). C) A crystal of Sarkosyl-Mg'-'+ from a suspension devoid of membrane material (see fig. 1 A).
BIOCH1MIE, 1981, 63, n ° 3.
Membrane protein adsorption by sarkosyl-Mg ~ crystals.
determination of SDS [21]. Protein concentrations of the membrane samples were determined by the procedure of Lowry et al. [22l after solubilization of the membrane with SDS. Crystalline serum albumin was used as a standard.
Results. 1. Light microscopy.
Figure 1 A shows the features of the flaky white crystals obtained by the addition of MgCI.:, to a Sarkosyl solution. The crystal suspension was heterogeneous with respect to crystal size and shape. Some crystals had a nearly rhomboid shape, which is in agreement with an earlier description [5], while some were much more elongated. There were also many crystals lacking a well defined form. Most crystals, regardless of their shape, had an irregular outline. When the crystals were incubated for several hours in the presence of S. citri membranes, similar results were obtained (fig. 1 B). The crystals were, however, approximately two to four times smaller than those visualized in figure 1 A. This difference was even greater when crystallization was enhanced with Mg 2+ ions after extensive solubi!ization of S. citri membranes with Sarkosyl. The number of crystals per microscope field was much greater and the crystals were also much smaller and essentially elongated (fig. 1 C).
181
3. Electron microscopy : thin sections.
Thin sections of isolated S. citri membranes, observed in the presence or absence of Sarkosy!-Mg ~ crysta'=s, are compared in figure 3. In agreement with an earlier investigation [23], the pure membrane preparation seemed to consist essentially of empty vesicles and only a few unlysed cells could be detected (fig. 3 A). When observed under suitable magnification, the membranes exhibited the trilaminar structure which is typical of fixed and stained thin biomembrane sections. Pure Sarkosyl-Mg 2+ crystals could not be visualized as thin sectioned specimens in the electron microscope. This was probably due to the fact that this material was neither electron dense nor stained by uranium and lead salts. Conversely, it was possible to visualize the crystals when they were mixed with S. citri membranes. Figure 3 B shows that the contour of the crystals was underlined by an electron dense material of membrane origin. Membrane vesicles were no longer detected and the electron dense material bounding the surface of the crystals had a unilaminar structure. These observations show that the vesicles were disrupted and that the native structure of the S. cirri membrane was destroyed. The crystals observed when crystallization was enhanced with Mg 2÷ ions after solubilization of the membranes with the detergent (fig. 3 C) were similar to those described above although more elongated.
2. Electron microscopy : negative staining.
4. SDS-PAGE.
Figure 2 A shows a collection of membrane vesicles obtained from S. citri cells lysed by sonication. Most of the vesicles were rounded, but some exhibited a tubular shape, strongly reminiscent of the helical morphology [2] of the cells from which they were prepared.
We have previously shown that the membrane material which cosedimented with Sarkosyl-Mg 2÷ crystals during high speed centrifugation contained almost all the membrane proteins, including the major component spiralin [3]. The problem thus became the identification of those membrane polypeptides which were adsorbed to the crystals. The protein-crystal complexes were purified by the Mband technique [5], proteins were solubilized with SDS and were analyzed by SDS-PAGE. Unfractionated S. citri membranes solubilized with SDS were used as a reference system in this experiment. M-bands were prepared from isolated membranes (MM-bands) and from whole cells (CM-bands),
The vesicles were no longer detected after incubation of the membranes with preformed SarkosylMg 2+ crystals (fig. 2 B). This shows that the membranes were disintegrated under the conditions of the experiment. The crystal surfaces had an appearance different from that of pure crystals (compare with figure 2 C), suggesting that material of membrane origin was adsorbed to their surface. When crystallization was enhanced with Mg 2* ions after extensive solubilization of membrane components with the detergent, the crystals which were obtained (not illustrated) did not differ from those visualized in figure 2 B, within the limits of the present investigation. BIOCH1MIE, 1981, 63, n ° 3.
After solubilization with SDS and reduction of disulfide bonds with 2-mercaptoethanol, it was possible to separate S. citri membrane polypeptides bv SDS-PAGE into 29 bands with apparent molecular weights between 12 000 and approximately 170 000. Major and minor components
182
H.
Wr6blewski
were l a b e l e d with D and d respectively and there were 7 m a j o r b a n d s : D5 (spiralin, M r 26,000, see ref. [19]), D 7 ( M r 39,000), D9 (Mr 51,000), D 1 3 (Mr 69,000), D 1 4 ( M r 75,000), D 1 6 (Mr 89,000), and D 1 7 (Mr 95,000). F u r t h e r analysis showed t h a t b a n d D 1 4 p r o b a b l y c o n t a i n e d two distinct p o l y p e p t i d e s , which we called D 1 4 a a n d D 1 4 b .
a n d coll.
These observations indicate that at least 30 distinct p o l y p e p t i d e s could be d i s c r i m i n a t e d by SDSP A G E in the S. citri m e m b r a n e . It was n o t possible to detect the release of any p r o t e i n f r o m the S. citri m e m b r a n e with a high ionic strength buffer ( l M KC1 in 0.1 M T r i s - H C l
Fro. 3. - -
Thin
sections o!
S. citri membranes and SarkosylM g ~+ crystals.
A) Untreated membrane vesicles (sarle material as in figure 2 A); note the trilaminar structure of the material observed at a higher magnification. B) A mixture of preformed crystals and S. citri membranes observed after incubation for 6 hours at room temperature (same material as in figures 1 B and 2 B) ; note the unilaminar structure of the material observed at a higher magnification. C) Crystals obtained by the addition of MgCI.o to S. citri membranes treated with 0.1 M Sarkosyl (see figure 1 C). BIOCH1MIE, 1981, 63, n ° 3.
Membrane
protein adsorption
buffer, p H 8.0). Polypeptides corresponding to b a n d s d6, D l l , D14a, d l 5 , d18, d22, d25, d26, and d27 were, however, released by washing the
i
2
183
b y s a r k o s y l - M g '+ crystals.
m e m b r a n e with a high p H and low ionic strength buffer c o n t a i n i n g 2 - m e r c a p t o e t h a n o l a n d E D T A (5 m M 2 - m e r c a p t o e t h a n o l and 1 m M E D T A in 5 m M g l y c i n e - N a O H buffer, p H 9.05). This indicates that the polypeptides corresponding to these bands, which are n o longer present in the electro-
3
2
1
•5 :
....
d22 d 2 1 ~-~ d20 ~
:
~
~,
,
:
2 ~
.....
~.i•
~~'~
F1~5. 5. - - S D S - P A G E o] S. cirri membrane proteins recovered in M-bands.
i. . . . Fro. 4 . -
K@
S D S - P A G E oJ S. citri membrane solubilized with SDS.
proteins
1) Membranes washed with 0.1 M Tris-HCl buffer, pH 8.0. 2) Membranes washed with 5 mM glycine-NaOH buffer, pH 9.5, containing 1 mM EDTA and 5 mM 2-mercaptoethanol. The starting material was 300 p~g (dry weight) of membranes solubilized with 0.1 M SDS in the presence of 0.6 M 2-mercaptoethanol. Dimensions of the gel : 100 (width) × 120 (height) X 4 ram. Current, 90 m A ; duration of the run, 9 h. Proteins were stained with Amido Black 10 B. Major and minor bands are labeled with D and d, respectively. D5: spiralin. BIOCHIM1E, 1981, 6.3, n ° 3.
1) Unfractionated membranes solubilized with 0.1 M SDS (same sample as in figure 4, lane 1). 2) Proteins in M-bands prepared from S. citri cells. 3) Proteins in Mbands prepared from S. cirri isolated membranes. The starting material was 150 ~g of membrane protein for all samples. Dimensions of the gel: 120 (width) × 130 (height) × 3 mm. Current, 80 mA ; duration of the run, 9 h. Proteins were stained with Amido Black 10 B.
phoregram of lane 2 in figure 4, are probably extrinsic m e m b r a n e components. This water soluble protein fraction represented approximately
184
H. W r 6 b l e w s k i and coll.
7.5 per cent of membrane dry weight, a figure in agreement with the highest values reported for other mycoplasmal membranes [24, 25]. The remaining polypeptides, including all the major species except D14a, shouid be thus regarded, at least as a first approximation (see discussion), as intrinsic polypeptides. Figure 5 illustrates the SDS-PAGE comparison of the polypeptide compositions of S. citri membranes (lane 1), CM-bands (lane 2), and MMbands (lane 3). CM-bands contained all the membrane polypeptides which could be detected by this technique, except d6, dl0, DII~ and d12. The
was not possible to obtain as many immunoprecipirates as expected [26J, since Sarkosyl inhibited the formation of several of them [3, 26]. The advantage of the method over SDS-PAGE is that it offered the possibility of a quantitative comparison of the components. Figure 6 shows that up to 7 antigens could be demonstrated in the two types of M-bands in the presence of Sarkosyl. The area under immunoprecipitate N ° 5 was 5 times greater in the immunoelectrophoregram corresponding to CM-hands (fig. 5 A) than in that corresponding to MM-bands (fig. 5 B). This means that antigen N ° 5 w~s
F16. 6. - - Crossed immunoelectrophoresis o] S. cirri n embrane components recovered ]rom M-bands. A) M-bands prepared from S. cirri cells. B) M-bands ~repared from isolated S. citri membranes. The amount of protein in both samples corresponded to 25 ~g of unfra~t[onated membrane protein as the starting material. The amount of gel was 100 M.cm-e. The amount of antime~ brane serum was 3.5 M.cm-2. Detergent : 17 mM Sarkosyl in the first directional catholyte and gel. Size ef the ge s : 85 × 100 × 1 mm. Ist direction : current, 15 mA/ plate ; duration of the run, 1.5 h ; temperature, 5°C. 2rd direction : current, 2.5 mA/plate ; duration of the run, 19 h ; room temperature. The immunoprecipitates were stained with Amido Black 10 B. Immunoprecipitate N ° 7 corresponds to spiralin. polypeptide composition of MM-bands was rather similar to that of CM-bands, since only minor differences were observed. MM-bands, contrary to CM-bands, were indeed devoid, of po!ypeptides d8, D14a, and d15. Otherwise, components d23, d26, and d27 were represented to only a slight extent. 5. Crossed immunoelectrophoresis.
The compositions of M-bands were also compared by crossed immunoelectrophoresis in the presence of Sarkosyl to prevent the precipitation of amphiphilic proteins. Under these conditions, it BIOCH1MIE, 1981, 63, n ° 3.
approximately 5 times more abundant in CMbands than in MM-bands, since in quantitative immunoelectrophoresis the area under an immunoprecipitate is proportional to the quantity of antigen, provided the amount o[ antibody remains constant [27]. The concentrations of other antigens, including spiralin (antigen N ° 7), were approximately the same in the two types of Mbands. 6. Free Sarkosyl in S a r k o s y l - M g ~" crystal suspensions.
Aqueous suspensions of crystals containing 0.1 M Sarkosyl and 50 mM MgCle were filtered
Membrane protein adsorption by sarkosyl-Mg ~+ crystals.
on paper and subsequently on cellulose nitrate filters with 10 nm pores. The recovered fluids contained about 2.5 mM Sarkosyl, which was assumed to be the concentration of free detergent in these crystal suspensions.
Discussion. The membranes used in the present investigation were prepared from S, cirri cells lysed by sonication. Observations of thin sectioned or negatively stained specimens showed that these membranes had a vesicular morphology similar to that of membranes prepared by osmotic lysis of cells [23]. The vesicles were no longer detected in mixtures containing membrane material and Sarkosyl-Mg 2+ crystals. It was also shown that the membranes had lost their trilaminar appearance in thin sections and were adsorbed to the surface of the crystals, These observations indicate that the structure of isolated S. cirri membranes was extensively disintegrated by Sarkosyl, even under conditions where apparent protection was maximal (Mg2+/ Sarkosy! = 0 5 [3]). Consequently. from a practical standpoint, sedimentation shou!d not be used as the only criterion for the structural integrity of membranes. A morphological characterization, e.g. by electron microscopy is also essential. Among the 30 polypeptides which were characterized by SDS-PAGE, 10 (bands d6, dl0, D l l , D14a, d15, d18, d22, d25-d27) were re!eased from the S. citri membrane with a detergent-free buffer. These polypeptides are thus hydrophilic and should be considered as extrinsic membrane components. Since intrinsic membrane proteins are amphiphilic [30], it is tempting to assume that the polypeptides which were not released from the membrane with detergent-free buffers are the intrinsic polypeptides of the S. citri membrane. This criterion may, however, be misleading since an extrinsic membrane protein bound to the inner surface of a closed membrane vesicle cannot be separated from the membrane fraction without disrupting the vesicle [5]. Though spiralin [191 is the only protein of the S. citri membrane whose amphiphilic nature has been demonstrated [31], it seems reasonable to infer that the protein fraction mentioned above is predominantly composed of intrinsic polypeptides. This conclusion is supported by the fact that the surface of the SarkosylMg 2÷ crystals is probably entirely hydrophobic and that the polypeptides recovered in CM-bands and BIOCHIMIE, 1981, 6], n ° 3.
185
MM-bands were essentially those that could not be released from the membrane with detergentfree buffers. Spiralin, in particular, was demonstrated in these M-bands. If the surface of the crystals was really totally hydrophobic, as was assumed since they rapidIy entered the organic phase when shaken in the presence of ether [5], then the extrinsic polypeptides which were recovered in M-bands were probably those which were the most tightly bound to intrinsic membrane components. It sould be noted that only minor differences were detected in the polypeptide compositions of the two types of M-bands. These differences were mainly quantitative, a fact which was more conclusively shown for one component by crossed immunoelectrophoresis (see fig. 6). It would have been interesting to correlate electrophoretic data with immunoelectrophoretic ones by establishing a reciprocal correspondence between the two systems, i.e., for a given component, to know which immunoprecipitate in crossed immunoelectrophoresis corresponds to a given band in SDS-PAGE and vice versa. This cannot be done simply at the present time, since purification of the different proteins is necessary for unambiguously establishing this correspondence [32]. In the case of tbe S. citri membrane, this was possible on!y for spiralin (Band D5 and immunoprecipitate n ° 7) since it is the only protein which has so far been purified from this membrane [19]. Tremblay et al. [5] have shown that it was possible to prepare Mbands with isolated Bacillus rnegaterium membranes. This is consistent with the fact that Mbands were obtained from isolated S. citri cell membranes. These authors also showed that the formation of the protein-crystal complex was not part of the process of crystallization [5] : some free detergent was necessary for cell lysis to occur. The small amount of free Sarkosyl (2.5 mM) which was detected in suspensions of SarkosylMg 2. crystals, was probably sufficient to lyse S. citri cells, but also to at least initiate the disintegration of the cell membrane structure. Membranes interfered with the formation of the crystals since these were much smaller when crystallization was enhanced with Mg z~ ions after extensive sotubilization of membrane components with Sarkosyl. Since the formation of the proteincrystal complexes occurred regardless of the sequence in which the differend components (membranes, Sarkosyl, Mg 2÷) were mixed, however, it is likely that homogeneous Sarkosyl-Mg 2÷ crystals containing surface-bound membrane proteins was the more stable arrangement and thus the most
186
H. Wr6blewski and coll.
favored. The fine structure of the complexes, i.e. the precise relative arrangement of Mg 2+ ions and Sarkosyl molecules in the crystals and the arrangement of proteins (and probably lipids) on the surface of the crystals, remains to be determined.
The M-band technique [5] has proved useful for the purification of DNA-membrane complexes, especially from bacterial cells. The ultimate goal is the purification of those membrane proteins to which DNA is specifically attached. Since these proteins probably represent a quantitatively minor fraction of the membrane, the purification procedure requires a precise optimization. It should be possible to improve the method by a systematic investigation of the binding selectivity of membrane proteins by the crystals, expressed as a function of Sarkosyl/protein and Mg2+/Sarkosyl ratios. The use of detergents other than Sarkosyl, but which are also able to cocrystallize with Mg 2+ ions (e.g. bile salts), could prove useful. Most investigations in this field have been carried out with bacterial cells. Mycoplasmas could also prove a good experimental model, not only because they are surrounded by a plasma membrane alone, but because of the smaller size of their genome [33]. It is likely that cells containing less genetic information possess simpler membranes (membranes exhibiting a smaller diversity of proteins), which is a great advantage for purifying membrane proteins. This assumption is supported experimentally by the fact that, to date, fewer antigens have been seen in crossed immunoelectrophoregrams of mycoplasmal membranes [16, 26, 33, 34] compared to those from bacterial plasma membranes [35, 36, 37]. Acknowledgments, This work was supported by the Centre National de la Recherche Scientifique (L.A. 256, Contrat C.N.R.S.Universite'). W e wish to thank Dave Grant ]or constructive criticism, Anne-Marie Touzalin and Annie Caval&r ]or skill]ul technical assistance and Franfoise de Sallier Dupin for typing the manuscript.
REFERENCES. 1. Maddy, A. H. 8¢ Dunn, M. J. (1976) In <> (Maddy, A. H., Ed.), pp. 177-196, Champman and Hall, London. 2. Cole, R. M., Tully, J. G., Popkin, T. J. & Bov6, J. M. (1973) J. Bacteriol., 115, 367-386. 3. Wrdblewski, H., Burlot, R. & Johansson, K.-E. (1978) Biochimie, 60, 389-398. BIOCH1MIE, 1981, 63, n ° 3.
4. Filip, C., Fletcher, G., Wulff, J. L. & Earhart, C. F. (1973) J. Bacteriol., 115, 717-722. 5. Tremblay, C. Y., Daniels, M. J. & Schaechter, M. (1969) J. Mol. Biol., 40, 65-76. 6. Firskein, W. (1972) J. Mol. Biol., 70, 383-397. 7. Ballesta, J. P., Cundliffe, E., Daniels, M. J., Silverstein, J. L., Susskind, M. M. & Schaechter, M. (1972) J. Bacteriol., 112 195-199. 8. Harmon, J. M. ~ Taber, H. W. (1977) J. Bacteriol., 129, 789-795. 9. Harmon, J. M. & Taber, H. W. (1977) 1. Bacteriol., 130, 1224-1233. 10. lnfante, A. A., Firskein, W., Hobart, P. ~ Murray, L. (1976) Biochemistry, 15, 4810-4817. 11. Wr6blewski, H. (1975) Biochimie, 57, 1095-1098. 12. Hjert6n, S. (1962) Arch. Biochem. Biophys. suppl. 1, 147-151. 13. Laurel/, C. B. (1965) Anal Biochem., 10, 358-361. 14. Clarke, H. G. M. 8~ Freeman, T. A. (1967) In Protides oJ the Biological Fluids (Peeters, H., Ed.) vol. 14, pp. 503-509, Elsevier, Amsterdam. 15. Bjerrum, O. J. 8~ Lundahl, P. (1974) Bixchim. Bit;phys. Acta, 342, 69-80. 16. Johansson, K.-E. ~ Hjert6n, S. (1974) J. Mol. Biol., 86, 341-348. 17. Bjerrum, O. J. e Bog-Hansen, T. C. (1976) Ill <> (Maddy, A. H., Ed.) pp. 378-426, Chapman and Hall, London. 18. Owen, P. & Smyth, C. J. (1977) In <> (Sa!t~n, M. R. H., Ed.) pp. 147-202, John Wiley and Szns Inc., New York. 19. Wr6blewski, H., Johansson, K.-E. ~ Hjert~n, S. (1977) Biochim. Biophys. Aeta, 4,~5, 275-289. 20. Reynolds, E. S. (1963) J. Cell Biol., 2A5, 208-212. 21. Takagi, T., Tsujii, K. s~ Shirahama, K. (1975) J. Bi:chem., 77, 939-947. 22. Lowry, O. H., Rosebrough, N. J., Farr, A. L. e, Randall, R. J. (1951) J. Biol. Chem., 193, 265-275. 23. Razin, S., Hasin, M., Ne'eman, Z. a Rottem, S. (1973) J. Bacteriol., 116, 1421-1435. 24. Ne'eman, Z., Kahane, I. & Razin, S. (1971) Bigchim. Biop~ys. Acta, 2~:9, 169-176. 25. Razin, S. (1975) In <
Adsorption of proteins from the Spiroplasma citri cell membrane by magnesium lauroyl-sarcosinate crystals. Henri WR()BLEWSKI *~, Rende B U R L O T * and Daniel T H O M A S ** (Refu le 2-8-1980, acceptd le 10-12-1980).
* Laboratoire de Biologie Cellulaire ** and Laboratoire de Cytologie expdrimentale, Universitd de Rennes, Complexe de Beaulieu, A venue du Gdndrat Leclerc, 35042 Rennes COdex (France).
R~sum~.
Summary.
Les interactions entre le Sarkosyl (lauroyl sareosinate de sodium), les ions M g 2÷ et la membrane de Spiroplasma cirri ont ~td analysdes au moyen de techniques microscopiques et de techniques glectrophordtiques. Cette dtude a dtd rdalisde dans des conditions oh les protdines membranaires n'dtaient apparemment pas extraites par le d~tergent (rapport molaire de M g C l J S a r k o s y l = 0,5).
Interactions between the ionic detergent Sarkosyl (sodium lauroyl sarcosinate), M g ~+ions, and the Spiroplasma citri cell membrane were analyzed microscopically and electrophoreticalIy. Studies were performed under conditions where membrane proteins were apparently not released from the membrane by the detergent (molar ratio of MgCle/ Sarkosyl = 0.5).
Bien que la membrane de S. citri interfbre dans une certaine mesure avec le phdnomkne de cristallisation, les cristaux de Sarkosyl-Mg 2+ se /orment inddpendemment de la sdquence d'addition des trois composants. Ceci s'accompagne d'une dd~intdgration de la structure membranaire et de I'adsorption de composants membranaires par les cristaux. La fraction de protdines lide aux cristaux contient la ma]oritd des polypeptides prdsumds intrinsbques, dont la spira~ine (une protdine amphiphile), et plusieurs polypeptides extrinskques. Les analyses dlectrophordtiques et immunodlectrophordtiques indiquent que les bandes-M ([ractions de cristaux chargds de matdriel membranaire), prdpardes ?t partir de cellules et celles prdpardes ~ partir de membranes isoldes de S. citri, possbdent des compositions similaires en protdines. Ces rdsu!tats montrent que la membrane de S. citri, contrairement aux membranes bactdriennes, n'est pas protdgde par les ions M g 2+ des ef]ets du Sarkosyl.
Although the S. citri membrane interfered with the crystallization phenomenon to some extent, the formation of Sarkosyl-Mg 2- crystals occurred regardless to the sequence o/ addition of the three components. Concomitantly the structure of the membrane disintegrated and membrane components were adsorbed to the crystal sur[aces. The membrane protein fraction bound to the crystals was composed of the ma]ority o / t h e putatively intrinsic polypeptides, including the amphiphilic protein spiralin, and several extrinsic polypeptides. The polypeptide compositions of M-bands (crystal fractions loaded with membrane material) prepared from S. citri cells" and ]rom isolated S. citri membranes were similar, as shown by sodium dodecyl-sulfate electrophoresis and crossed immunoelectrophoresis. These results show that, the S. citri cell membrane, in contrast to bacterial membranes, is not protected from the effect of Sarkosyl by Mg 2+ ions.
Abbreviations :
Introduction.
CIE, crossed immunoelectrophoresis ; CM-bands, Mbands prepared from cells; EDTA, ethylene diamine tetraacetate ; I, ionic strength ; MM-bands, M-bands prepared from isolated membranes ; SDS, sodium dodecylsulfate; SDS-PAGE, sodium dodecyl-sulfate polyacrylamide gel electrophoresis. <> To whom all correspondence should be addressed.
For practical reasons which are self-evident, a low sedimentation coefficient is the most c o m m o n ly used criterion of the solubilization of membrane proteins, e.g. by a detergent [1]. Specifically, after centrifugation for about 1 h at 100,000 to 14
H. Wr6blewski and coll.
178
2 0 0 , 0 0 0 × g in a m e d i u m with a density close to that of water, proteins recovered in the supernatant are considered as those which are solubilized or at least extracted (see ref. [1] for a criticism of the terminology). Conversely, those recovered in the pellet are considered to be the proteins which are n o t solubilized u n d e r the conditions of the particular experiment. T h e exclusive use of the above m e n t i o n e d criterion m a y be misleading in some circumstances. F o r example, w h e n the m e m b r a n e of the helical m y c o p l a s m a Spiroplasrna citri [2] is treated with the ionic detergent Sarkosyl (sodium lauroyl-sarcosinate) i n the presence of M g 2+ ions, some proteins are recovered in the residue a n d the r e m a i n d e r in the s u p e r n a t a n t . T h e ratio of the two fractions is a f u n c t i o n of the Mg2+/Sarkosyl ratio. T h u s when this ratio is ~ 0.5, all the proteins are quantitatively recovered in the residue [3]. It becomes tempting to conclude, as was done for some bacterial m e m b r a n e s [4], that M g 2+ ions protect the m e m b r a n e against the effect of Sarkosyl. This would necessarily m e a n that the structure of the m e m b r a n e is preserved u n d e r these conditions. We have, however, suggested a n o t h e r plausible explan a t i o n : the proteins are extracted from the m e m b r a n e a n d sediment u p o n centrifngation as proteincrystal complexes [3]. A c c o r d i n g to this hypothesis the m e m b r a n e would, of course, be m o r e or less disintegrated.
duced to 5 per cent (v/v). The cells were harvested by centrifugation at 15,000 × g for 15 min at 4oc, dispersed in 0.1 M Tris-HC1 buffer (pH 8.0) and disrupted by sonication at 20 kcycles for 2 × 1 rain at 0°C. The membranes were pelleted by centrifugation at 40,000 x g for 1 h at 4°C and extensively washed with 0.1 M TrisHC1 buffer (pH 8.0).
T h e purpose of the present investigation was to resolve this problem, in particular b y a m o r p h o logical analysis of the material. Electrophoretic a n d i m m u n o e l e c t r o p h o r e t i c studies were perform e d to characterize the S. citri m e m b r a n e p r o t e i n fraction recovered in <
SDS-PAGE was performed in 0.1 M Tris-HC1 buffer (pH 8.0) containing 10 mM SDS. Compositions of the lower gel and of the upper (pre-separaing) gel were respectively T = 6 per cent and C = 5 per cent, and T = 4 per cent and C = 2 per cent. The nomenclature of Hjert~n [12] is used to describe gel composition : percentage T is the total monomer (acrylamide + N, N'-methylene bisacrylamide) concentration, and percentage C is the percentage of N, N'-methylene bisacrylamide relative to the total monomer concentration.
Material
and Methods.
Chemicals. Sodium lauroyl-sarkosinate (Sarkosyl), acrylamide, and N-N'-methylene bis acrylamide were purchased from Sigma (U.S.A.). Sodium dodecyl sulfate (SDS) was from Merck (Germany), and agarose from Bio-Rad (U.S.A.).
M-band #actionation [5]. Membranes or intact cells were treated with Sarkosyl in the presence of Mg e+ ions. To 0.5 ml of Tris-HCl buffer (pH 8.0) containing 10 mg of membranes (5 mg of membrane protein) were added 0.25 ml of 0.4 M Sarkosyl and 0.25 ml of 0.2 M MgC12, both in distilled water. After mixing, the solutions were left for 1 h at 4°C. Each sample was then layered on top of a discontinuous sucrose gradient : 2 ml of 60 per cent sucrose, 13 ml of 40 per cent sucrose, and 13 ml of 15 per cent sucrose in 50 mM Tris-HC1 buffer (pH 8.0) containing 50 mM MgC1e. The gradients were centrifuged at 32,000 × g in a Spinco SW 25 rotor for 40 rain at 4°C. M-bands containing the Mge+-Sarkosyl crystals were visible at the interface of the 15 per cent and 40 per cent sucrose solutions. The material containing M-bands was dialyzed for 5 days at 4°C against 10 mM Tris-HCl buffer (pH 8.0) containing 20 mM Sarkosyl and 1 mM EDTA. The buffer was changed daily. The sample was lyophilized and again solubilized with 0.1 M Tris-HCl buffer (pH 8.0) in order to obtain a 0.1 M Sarkosyl solution containing 5 mg protein/ml.
SDS-polyacrylamide gel electrophoresis (SDS-PAGE).
Crossed immunoelectrophoresis (CIE). CIE [13, 14] was performed in the presence of detergent [15-18] in 1 mm thick agarose gels cast on 85 × 100 mm glass plates. The agarose concentration was 1 per cent (w/v) in veronal buffer (pH 8.6), I = 0,03 ; 17 mM Sarkosyl was included in the catholyte and in the gel of the first direction in order to prevent protein precipitation [3,26]. Antiserum containing antibodies against the S. cirri membrane was prepared as described [11]. Other details concerning CIE were described in a previous publication [19].
Light microscopy. Culture o] cells and membrane preparation. Spiroplasma cirri strain C189 (ATCC 27665) was grown as previously described [11] except that the concentration of foal serum in the culture medium was re-
BIOCHIMIE, 1981, 63, n ° 3.
Crystal suspensions were observed without any treatment under cover slips in a light microscope and were photographed. Phase contrast and dark field optics were used.
Membrane protein adsorption by sarkosyl-Mg ~÷ crystals. Electron microscopy. Sarkosyl-Mg 2+ crystals, either alone or loaded with membrane proteins, were pelleted by centrifugation (180,000 X g, 60 min, 4°C), fixed with 2.5 per cent glutaraldehyde, and postfixed with 2 per cent Os 04 in 0.1 M phosphate buffer (pH 7.2). The pellets were then dehy-
179
drated through a graded series of acetone and embedded in an araldite-epon mixture. Ultra-thin sections were stained with uranyl acetate and lead citrate [20]. For negative staining, drops of crystal suspensions were placed on carbonated collodion coated copper grids for 3 rain. After removal of excess fluid, the material was
B
,,L
FIG. 1 . -
Light microscope views o/ Sarkosyl- Mg2"crys tals. A) Crystals obtained by adding MgCI2 (50 mM final concentration) to a solution of 0. I M Sarkosyl in 50 m M Tris-HC1 buffer, pH 8.0. B) Preformed crystals mixed with S. citri membranes (10 mg protein/ml) and observed after incubation for 6 hours at room temperature. C) Crystals obtained after addition of MgC1._. (50 mM final concentration) to S. citri membranes (I0 mg protein/ml) treated for 1 hour with 0.1 M Sarkosyl in 0.05 M TrisHC1 buffer pH 8.0. The crystals were observed after incubation for 6 hours at room temperature. Phase contrast optics were used to observe the crystals in all samples. B I O C H I M I E , 1981, 68, n ° 3.
H.
180
Wr6blewski
stained for 20 sec with 2 per cent sodium phosphotungstate (pH 7.2). Excess fluid was again removed and the preparations were left to dry overnight at 32°C before observation.
,
500
a n d coll. Other techniques. Sarkosyl was titrated as a detergent-methylene blue complex extracted with chloroform, as described fer the
nm
FIG. 2. - - Electron mtcroscope views o] S. citri membranes and Sarkosyl-Mg 2+ crystals negatively stained with 2 per cent sodium phosphotungstate. A) Membrane vesicles obtained from S. cirri cells lysed by sonication and washed with 0.1 M Tris-HC1 buffer pH 8.0. B) Crystal of Sarkosyl-Mg2+ in a mixture of preformed crystals and S. citri membranes (see fig. 1 B). C) A crystal of Sarkosyl-Mg'-'+ from a suspension devoid of membrane material (see fig. 1 A).
BIOCH1MIE, 1981, 63, n ° 3.
Membrane protein adsorption by sarkosyl-Mg ~ crystals.
determination of SDS [21]. Protein concentrations of the membrane samples were determined by the procedure of Lowry et al. [22l after solubilization of the membrane with SDS. Crystalline serum albumin was used as a standard.
Results. 1. Light microscopy.
Figure 1 A shows the features of the flaky white crystals obtained by the addition of MgCI.:, to a Sarkosyl solution. The crystal suspension was heterogeneous with respect to crystal size and shape. Some crystals had a nearly rhomboid shape, which is in agreement with an earlier description [5], while some were much more elongated. There were also many crystals lacking a well defined form. Most crystals, regardless of their shape, had an irregular outline. When the crystals were incubated for several hours in the presence of S. citri membranes, similar results were obtained (fig. 1 B). The crystals were, however, approximately two to four times smaller than those visualized in figure 1 A. This difference was even greater when crystallization was enhanced with Mg 2+ ions after extensive solubi!ization of S. citri membranes with Sarkosyl. The number of crystals per microscope field was much greater and the crystals were also much smaller and essentially elongated (fig. 1 C).
181
3. Electron microscopy : thin sections.
Thin sections of isolated S. citri membranes, observed in the presence or absence of Sarkosy!-Mg ~ crysta'=s, are compared in figure 3. In agreement with an earlier investigation [23], the pure membrane preparation seemed to consist essentially of empty vesicles and only a few unlysed cells could be detected (fig. 3 A). When observed under suitable magnification, the membranes exhibited the trilaminar structure which is typical of fixed and stained thin biomembrane sections. Pure Sarkosyl-Mg 2+ crystals could not be visualized as thin sectioned specimens in the electron microscope. This was probably due to the fact that this material was neither electron dense nor stained by uranium and lead salts. Conversely, it was possible to visualize the crystals when they were mixed with S. citri membranes. Figure 3 B shows that the contour of the crystals was underlined by an electron dense material of membrane origin. Membrane vesicles were no longer detected and the electron dense material bounding the surface of the crystals had a unilaminar structure. These observations show that the vesicles were disrupted and that the native structure of the S. cirri membrane was destroyed. The crystals observed when crystallization was enhanced with Mg 2÷ ions after solubilization of the membranes with the detergent (fig. 3 C) were similar to those described above although more elongated.
2. Electron microscopy : negative staining.
4. SDS-PAGE.
Figure 2 A shows a collection of membrane vesicles obtained from S. citri cells lysed by sonication. Most of the vesicles were rounded, but some exhibited a tubular shape, strongly reminiscent of the helical morphology [2] of the cells from which they were prepared.
We have previously shown that the membrane material which cosedimented with Sarkosyl-Mg 2÷ crystals during high speed centrifugation contained almost all the membrane proteins, including the major component spiralin [3]. The problem thus became the identification of those membrane polypeptides which were adsorbed to the crystals. The protein-crystal complexes were purified by the Mband technique [5], proteins were solubilized with SDS and were analyzed by SDS-PAGE. Unfractionated S. citri membranes solubilized with SDS were used as a reference system in this experiment. M-bands were prepared from isolated membranes (MM-bands) and from whole cells (CM-bands),
The vesicles were no longer detected after incubation of the membranes with preformed SarkosylMg 2+ crystals (fig. 2 B). This shows that the membranes were disintegrated under the conditions of the experiment. The crystal surfaces had an appearance different from that of pure crystals (compare with figure 2 C), suggesting that material of membrane origin was adsorbed to their surface. When crystallization was enhanced with Mg 2* ions after extensive solubilization of membrane components with the detergent, the crystals which were obtained (not illustrated) did not differ from those visualized in figure 2 B, within the limits of the present investigation. BIOCH1MIE, 1981, 63, n ° 3.
After solubilization with SDS and reduction of disulfide bonds with 2-mercaptoethanol, it was possible to separate S. citri membrane polypeptides bv SDS-PAGE into 29 bands with apparent molecular weights between 12 000 and approximately 170 000. Major and minor components
182
H.
Wr6blewski
were l a b e l e d with D and d respectively and there were 7 m a j o r b a n d s : D5 (spiralin, M r 26,000, see ref. [19]), D 7 ( M r 39,000), D9 (Mr 51,000), D 1 3 (Mr 69,000), D 1 4 ( M r 75,000), D 1 6 (Mr 89,000), and D 1 7 (Mr 95,000). F u r t h e r analysis showed t h a t b a n d D 1 4 p r o b a b l y c o n t a i n e d two distinct p o l y p e p t i d e s , which we called D 1 4 a a n d D 1 4 b .
a n d coll.
These observations indicate that at least 30 distinct p o l y p e p t i d e s could be d i s c r i m i n a t e d by SDSP A G E in the S. citri m e m b r a n e . It was n o t possible to detect the release of any p r o t e i n f r o m the S. citri m e m b r a n e with a high ionic strength buffer ( l M KC1 in 0.1 M T r i s - H C l
Fro. 3. - -
Thin
sections o!
S. citri membranes and SarkosylM g ~+ crystals.
A) Untreated membrane vesicles (sarle material as in figure 2 A); note the trilaminar structure of the material observed at a higher magnification. B) A mixture of preformed crystals and S. citri membranes observed after incubation for 6 hours at room temperature (same material as in figures 1 B and 2 B) ; note the unilaminar structure of the material observed at a higher magnification. C) Crystals obtained by the addition of MgCI.o to S. citri membranes treated with 0.1 M Sarkosyl (see figure 1 C). BIOCH1MIE, 1981, 63, n ° 3.
Membrane
protein adsorption
buffer, p H 8.0). Polypeptides corresponding to b a n d s d6, D l l , D14a, d l 5 , d18, d22, d25, d26, and d27 were, however, released by washing the
i
2
183
b y s a r k o s y l - M g '+ crystals.
m e m b r a n e with a high p H and low ionic strength buffer c o n t a i n i n g 2 - m e r c a p t o e t h a n o l a n d E D T A (5 m M 2 - m e r c a p t o e t h a n o l and 1 m M E D T A in 5 m M g l y c i n e - N a O H buffer, p H 9.05). This indicates that the polypeptides corresponding to these bands, which are n o longer present in the electro-
3
2
1
•5 :
....
d22 d 2 1 ~-~ d20 ~
:
~
~,
,
:
2 ~
.....
~.i•
~~'~
F1~5. 5. - - S D S - P A G E o] S. cirri membrane proteins recovered in M-bands.
i. . . . Fro. 4 . -
K@
S D S - P A G E oJ S. citri membrane solubilized with SDS.
proteins
1) Membranes washed with 0.1 M Tris-HCl buffer, pH 8.0. 2) Membranes washed with 5 mM glycine-NaOH buffer, pH 9.5, containing 1 mM EDTA and 5 mM 2-mercaptoethanol. The starting material was 300 p~g (dry weight) of membranes solubilized with 0.1 M SDS in the presence of 0.6 M 2-mercaptoethanol. Dimensions of the gel : 100 (width) × 120 (height) X 4 ram. Current, 90 m A ; duration of the run, 9 h. Proteins were stained with Amido Black 10 B. Major and minor bands are labeled with D and d, respectively. D5: spiralin. BIOCHIM1E, 1981, 6.3, n ° 3.
1) Unfractionated membranes solubilized with 0.1 M SDS (same sample as in figure 4, lane 1). 2) Proteins in M-bands prepared from S. citri cells. 3) Proteins in Mbands prepared from S. cirri isolated membranes. The starting material was 150 ~g of membrane protein for all samples. Dimensions of the gel: 120 (width) × 130 (height) × 3 mm. Current, 80 mA ; duration of the run, 9 h. Proteins were stained with Amido Black 10 B.
phoregram of lane 2 in figure 4, are probably extrinsic m e m b r a n e components. This water soluble protein fraction represented approximately
184
H. W r 6 b l e w s k i and coll.
7.5 per cent of membrane dry weight, a figure in agreement with the highest values reported for other mycoplasmal membranes [24, 25]. The remaining polypeptides, including all the major species except D14a, shouid be thus regarded, at least as a first approximation (see discussion), as intrinsic polypeptides. Figure 5 illustrates the SDS-PAGE comparison of the polypeptide compositions of S. citri membranes (lane 1), CM-bands (lane 2), and MMbands (lane 3). CM-bands contained all the membrane polypeptides which could be detected by this technique, except d6, dl0, DII~ and d12. The
was not possible to obtain as many immunoprecipirates as expected [26J, since Sarkosyl inhibited the formation of several of them [3, 26]. The advantage of the method over SDS-PAGE is that it offered the possibility of a quantitative comparison of the components. Figure 6 shows that up to 7 antigens could be demonstrated in the two types of M-bands in the presence of Sarkosyl. The area under immunoprecipitate N ° 5 was 5 times greater in the immunoelectrophoregram corresponding to CM-hands (fig. 5 A) than in that corresponding to MM-bands (fig. 5 B). This means that antigen N ° 5 w~s
F16. 6. - - Crossed immunoelectrophoresis o] S. cirri n embrane components recovered ]rom M-bands. A) M-bands prepared from S. cirri cells. B) M-bands ~repared from isolated S. citri membranes. The amount of protein in both samples corresponded to 25 ~g of unfra~t[onated membrane protein as the starting material. The amount of gel was 100 M.cm-e. The amount of antime~ brane serum was 3.5 M.cm-2. Detergent : 17 mM Sarkosyl in the first directional catholyte and gel. Size ef the ge s : 85 × 100 × 1 mm. Ist direction : current, 15 mA/ plate ; duration of the run, 1.5 h ; temperature, 5°C. 2rd direction : current, 2.5 mA/plate ; duration of the run, 19 h ; room temperature. The immunoprecipitates were stained with Amido Black 10 B. Immunoprecipitate N ° 7 corresponds to spiralin. polypeptide composition of MM-bands was rather similar to that of CM-bands, since only minor differences were observed. MM-bands, contrary to CM-bands, were indeed devoid, of po!ypeptides d8, D14a, and d15. Otherwise, components d23, d26, and d27 were represented to only a slight extent. 5. Crossed immunoelectrophoresis.
The compositions of M-bands were also compared by crossed immunoelectrophoresis in the presence of Sarkosyl to prevent the precipitation of amphiphilic proteins. Under these conditions, it BIOCH1MIE, 1981, 63, n ° 3.
approximately 5 times more abundant in CMbands than in MM-bands, since in quantitative immunoelectrophoresis the area under an immunoprecipitate is proportional to the quantity of antigen, provided the amount o[ antibody remains constant [27]. The concentrations of other antigens, including spiralin (antigen N ° 7), were approximately the same in the two types of Mbands. 6. Free Sarkosyl in S a r k o s y l - M g ~" crystal suspensions.
Aqueous suspensions of crystals containing 0.1 M Sarkosyl and 50 mM MgCle were filtered
Membrane protein adsorption by sarkosyl-Mg ~+ crystals.
on paper and subsequently on cellulose nitrate filters with 10 nm pores. The recovered fluids contained about 2.5 mM Sarkosyl, which was assumed to be the concentration of free detergent in these crystal suspensions.
Discussion. The membranes used in the present investigation were prepared from S, cirri cells lysed by sonication. Observations of thin sectioned or negatively stained specimens showed that these membranes had a vesicular morphology similar to that of membranes prepared by osmotic lysis of cells [23]. The vesicles were no longer detected in mixtures containing membrane material and Sarkosyl-Mg 2+ crystals. It was also shown that the membranes had lost their trilaminar appearance in thin sections and were adsorbed to the surface of the crystals, These observations indicate that the structure of isolated S. cirri membranes was extensively disintegrated by Sarkosyl, even under conditions where apparent protection was maximal (Mg2+/ Sarkosy! = 0 5 [3]). Consequently. from a practical standpoint, sedimentation shou!d not be used as the only criterion for the structural integrity of membranes. A morphological characterization, e.g. by electron microscopy is also essential. Among the 30 polypeptides which were characterized by SDS-PAGE, 10 (bands d6, dl0, D l l , D14a, d15, d18, d22, d25-d27) were re!eased from the S. citri membrane with a detergent-free buffer. These polypeptides are thus hydrophilic and should be considered as extrinsic membrane components. Since intrinsic membrane proteins are amphiphilic [30], it is tempting to assume that the polypeptides which were not released from the membrane with detergent-free buffers are the intrinsic polypeptides of the S. citri membrane. This criterion may, however, be misleading since an extrinsic membrane protein bound to the inner surface of a closed membrane vesicle cannot be separated from the membrane fraction without disrupting the vesicle [5]. Though spiralin [191 is the only protein of the S. citri membrane whose amphiphilic nature has been demonstrated [31], it seems reasonable to infer that the protein fraction mentioned above is predominantly composed of intrinsic polypeptides. This conclusion is supported by the fact that the surface of the SarkosylMg 2÷ crystals is probably entirely hydrophobic and that the polypeptides recovered in CM-bands and BIOCHIMIE, 1981, 6], n ° 3.
185
MM-bands were essentially those that could not be released from the membrane with detergentfree buffers. Spiralin, in particular, was demonstrated in these M-bands. If the surface of the crystals was really totally hydrophobic, as was assumed since they rapidIy entered the organic phase when shaken in the presence of ether [5], then the extrinsic polypeptides which were recovered in M-bands were probably those which were the most tightly bound to intrinsic membrane components. It sould be noted that only minor differences were detected in the polypeptide compositions of the two types of M-bands. These differences were mainly quantitative, a fact which was more conclusively shown for one component by crossed immunoelectrophoresis (see fig. 6). It would have been interesting to correlate electrophoretic data with immunoelectrophoretic ones by establishing a reciprocal correspondence between the two systems, i.e., for a given component, to know which immunoprecipitate in crossed immunoelectrophoresis corresponds to a given band in SDS-PAGE and vice versa. This cannot be done simply at the present time, since purification of the different proteins is necessary for unambiguously establishing this correspondence [32]. In the case of tbe S. citri membrane, this was possible on!y for spiralin (Band D5 and immunoprecipitate n ° 7) since it is the only protein which has so far been purified from this membrane [19]. Tremblay et al. [5] have shown that it was possible to prepare Mbands with isolated Bacillus rnegaterium membranes. This is consistent with the fact that Mbands were obtained from isolated S. citri cell membranes. These authors also showed that the formation of the protein-crystal complex was not part of the process of crystallization [5] : some free detergent was necessary for cell lysis to occur. The small amount of free Sarkosyl (2.5 mM) which was detected in suspensions of SarkosylMg 2. crystals, was probably sufficient to lyse S. citri cells, but also to at least initiate the disintegration of the cell membrane structure. Membranes interfered with the formation of the crystals since these were much smaller when crystallization was enhanced with Mg z~ ions after extensive sotubilization of membrane components with Sarkosyl. Since the formation of the proteincrystal complexes occurred regardless of the sequence in which the differend components (membranes, Sarkosyl, Mg 2÷) were mixed, however, it is likely that homogeneous Sarkosyl-Mg 2÷ crystals containing surface-bound membrane proteins was the more stable arrangement and thus the most
186
H. Wr6blewski and coll.
favored. The fine structure of the complexes, i.e. the precise relative arrangement of Mg 2+ ions and Sarkosyl molecules in the crystals and the arrangement of proteins (and probably lipids) on the surface of the crystals, remains to be determined.
The M-band technique [5] has proved useful for the purification of DNA-membrane complexes, especially from bacterial cells. The ultimate goal is the purification of those membrane proteins to which DNA is specifically attached. Since these proteins probably represent a quantitatively minor fraction of the membrane, the purification procedure requires a precise optimization. It should be possible to improve the method by a systematic investigation of the binding selectivity of membrane proteins by the crystals, expressed as a function of Sarkosyl/protein and Mg2+/Sarkosyl ratios. The use of detergents other than Sarkosyl, but which are also able to cocrystallize with Mg 2+ ions (e.g. bile salts), could prove useful. Most investigations in this field have been carried out with bacterial cells. Mycoplasmas could also prove a good experimental model, not only because they are surrounded by a plasma membrane alone, but because of the smaller size of their genome [33]. It is likely that cells containing less genetic information possess simpler membranes (membranes exhibiting a smaller diversity of proteins), which is a great advantage for purifying membrane proteins. This assumption is supported experimentally by the fact that, to date, fewer antigens have been seen in crossed immunoelectrophoregrams of mycoplasmal membranes [16, 26, 33, 34] compared to those from bacterial plasma membranes [35, 36, 37]. Acknowledgments, This work was supported by the Centre National de la Recherche Scientifique (L.A. 256, Contrat C.N.R.S.Universite'). W e wish to thank Dave Grant ]or constructive criticism, Anne-Marie Touzalin and Annie Caval&r ]or skill]ul technical assistance and Franfoise de Sallier Dupin for typing the manuscript.
REFERENCES. 1. Maddy, A. H. 8¢ Dunn, M. J. (1976) In <
4. Filip, C., Fletcher, G., Wulff, J. L. & Earhart, C. F. (1973) J. Bacteriol., 115, 717-722. 5. Tremblay, C. Y., Daniels, M. J. & Schaechter, M. (1969) J. Mol. Biol., 40, 65-76. 6. Firskein, W. (1972) J. Mol. Biol., 70, 383-397. 7. Ballesta, J. P., Cundliffe, E., Daniels, M. J., Silverstein, J. L., Susskind, M. M. & Schaechter, M. (1972) J. Bacteriol., 112 195-199. 8. Harmon, J. M. ~ Taber, H. W. (1977) J. Bacteriol., 129, 789-795. 9. Harmon, J. M. & Taber, H. W. (1977) 1. Bacteriol., 130, 1224-1233. 10. lnfante, A. A., Firskein, W., Hobart, P. ~ Murray, L. (1976) Biochemistry, 15, 4810-4817. 11. Wr6blewski, H. (1975) Biochimie, 57, 1095-1098. 12. Hjert6n, S. (1962) Arch. Biochem. Biophys. suppl. 1, 147-151. 13. Laurel/, C. B. (1965) Anal Biochem., 10, 358-361. 14. Clarke, H. G. M. 8~ Freeman, T. A. (1967) In Protides oJ the Biological Fluids (Peeters, H., Ed.) vol. 14, pp. 503-509, Elsevier, Amsterdam. 15. Bjerrum, O. J. 8~ Lundahl, P. (1974) Bixchim. Bit;phys. Acta, 342, 69-80. 16. Johansson, K.-E. ~ Hjert6n, S. (1974) J. Mol. Biol., 86, 341-348. 17. Bjerrum, O. J. e Bog-Hansen, T. C. (1976) Ill <