Gas vesicles isolated from Halobacterium cells by lysis in hypotonic solution are structurally weakened

Abstract

Analysis of pressure-collapse curves of Halobacterium cells containing gas vesicles and of gas vesicles released from such cells by hypotonic lysis shows that the isolated gas vesicles are considerably weaker than those present within the cells: their mean critical collapse pressure was around 0.049–0.058 MPa, as compared to 0.082–0.095 MPa for intact cells. The hypotonic lysis procedure, which is widely used for the isolation of gas vesicles from members of the Halobacteriaceae, thus damages the mechanical properties of the vesicles. The phenomenon can possibly be attributed to the loss of one or more structural gas vesicle proteins such as GvpC, the protein that strengthens the vesicles built of GvpA subunits: Halobacterium GvpC is a highly acidic, typically “halophilic” protein, expected to denature in the absence of molar concentrations of salt.

Halobacterium, Gas vesicles, GvpC

1 Introduction

Gas vesicles that increase the buoyancy of cells are found in different groups of prokaryotes, Archaea as well as Bacteria. The properties of these gas vesicles and the advantage that the possession of gas vesicles may bestow on the organism that produce them have been extensively studied and documented [1]. The groups in which gas vesicles have mostly been studied are the cyanobacteria (e.g., Anabaena, Pseudanabaena, Microcystis) and the halophilic Archaea of the family Halobacteriaceae.

Production of gas vesicles has been described in four Halobacteriaceae species: Halobacterium salinarum, Haloferax mediterranei, Halorubrum vacuolatum, and Halogeometricum borinquense [2]. The genetics and regulation of gas vesicle production have been investigated in-depth in Hbt. salinarum [3,4] and in Hfx. mediterranei [5–8]. Although few ecological studies have been performed that may confirm the advantage of the production of gas vesicles in halophilic Archaea, it is generally assumed that the main benefit to the cells may be the migration toward the oxygen-rich interface between the brine and the atmosphere [6,9].

Intact gas vesicles have been isolated from cyanobacteria by breaking the cell wall by rapid addition of sucrose (for Anabaena) or by lysozyme treatment (for Microcystis) [10,11]. However, to isolate and purify gas vesicles from halophilic Archaea a different protocol has been developed, based on the lysis of the cell wall in low-salt solutions. The method introduced by Simon [12], involving lysis of the cells at a low salt concentration, treatment with DNAse to reduce the viscosity of the solution, followed by low-speed centrifugation to collect the vesicles by flotation, is generally used in studies on gas vesicles in the Halobacteriaceae, e.g. [5,13–15]. Alternative methods for the isolation of gas vesicles from Halobacterium have been described, based on dialysis against distilled water [16] or cell lysis at high pH [17]. It is always silently assumed that the properties of the gas vesicles thus isolated reflect those of the vesicles within the living cells.

We here provide evidence that gas vesicles isolated from halophilic Archaea by hypotonic or alkaline lysis are substantially weaker than those present in the intact cells, and we propose a possible explanation for this observation.

2 Materials and methods

2.1 Strains and culture conditions

Halobacterium NRC-1 (ATCC 700922) (now classified as a strain of H. salinarum) [18] was grown with shaking at 35 °C in 250 ml Erlenmeyer flasks containing 150 ml portions of medium (g l⁻¹): NaCl, 250; MgSO₄.7H₂O, 20; KCl, 2; tri-Na-citrate, 3, Bacto tryptone, 5, and Bacto yeast extract, 3; pH 7 [19]. Portions of 0.1 ml of 3-day old liquid cultures (OD₆₀₀= 2.0) were spread on plates of the above medium, solidified with 15 g l⁻¹ agar. Alternatively, we grew H. salinarum in liquid medium containing (g l⁻¹): NaCl, 250; MgCl₂.6H₂O, 5; KCl, 5; NH₄Cl, 5, and Bacto yeast extract, 10; pH 7. Cultures (150 ml in 250 ml Erlenmeyer flasks) were grown with shaking at 35 °C for 3 days to maximize production of gas vesicles in the early stationary phase.

2.2 Preparation of cell suspensions and isolated gas vesicles

Cells grown on agar plates for 10 days at 35 °C were suspended in a solution containing the salts of growth medium, buffered at pH 7 with 5 mM HEPES. Liquid cultures were left to stand at room temperature without shaking for at least 24 h. Cells accumulating at the surface were collected by means of a Pasteur pipette, and they were further concentrated by ‘accelerated flotation': cell suspensions were centrifuged at 20 °C in 50-ml tubes in a Sorvall RC-5B centrifuge using an SS-34 rotor over a period of 90 min, during which the rotation velocity was gradually increased from 500 to 1000 rpm. For experiments with intact cells, cells were suspended to an OD₆₀₀ between 0.4 and 0.8 in a solution containing the salts of the growth medium without yeast extract, buffered to pH 7 with 5 mM HEPES. Isolated gas vesicles were obtained in the following ways: (a) Cells were washed from each agar plate with 15 ml of 1 mM MgSO₄ containing 10 μg ml⁻¹ DNAse I (bovine pancreas, Sigma). The suspension was incubated at 37 °C for 3 h. NaCl was then added to a final concentration of 100 g l⁻¹, and the suspension was filtered through tissue paper into a 30 ml Correx glass centrifuge tube. Each tube was overlaid with 10 ml of 50 g l⁻¹ NaCl and the tubes were centrifuged for 16 h in a swing-out rotor at 60 g. The gas vesicles floating to the top of the tube were recovered using a Pasteur pipette, and suspended in 50 g l⁻¹ NaCl [12]. (b) Alternatively, we isolated gas vesicles by diluting concentrated cell suspensions diluted with a 20-fold volume of distilled water or 5 mM HEPES, pH 7, to achieve lysis of the cells. Some of the lysate was used directly in the pressurization experiments. Gas vesicles were purified from the remainder by treating with 0.1 mg ml⁻¹ DNAse for 15 min, followed by accelerated flotation by centrifugation in 30 ml glass tubes for 2.5 h, with gradual increase of the rotation velocity from 500 to 1200 rpm. Gas vesicles accumulating at the surface and at the meniscus of the tubes were collected and resuspended to OD₆₀₀ values between 0.25 and 0.5 in 5 mM HEPES buffer, pH 7, or in buffered solutions of different salts (NaCl, KCl) as indicated in the experiments. (c) We also attempted the procedure described by Larsen et al. [17], based on cell lysis at high pH, followed by neutralization and filtration.

2.3 Pressure-collapse experiments

Portions of 1.2 ml of suspensions of cells or isolated gas vesicles in 15 ml glass test tubes were placed in an anaerobic jar converted to a pressurizing vessel by connecting the gassing port in the lid to a nitrogen gas cylinder. The pressure, as monitored with a pressure gauge, was increased to the desired value. The pressure gauge was calibrated by connecting a U-shaped tube filled with mercury to the gassing inlet and recording the height of the mercury column. After pressurizing for at least 30 s the pressure was released, the jar was opened, and the OD₆₀₀ of the suspensions was measured against water. Experiments were performed in triplicate.

The percentage of intact gas vesicles was calculated on the basis of the optical density before pressurization (100% vesicles intact) and after exposure to at least 0.18 MPa (all gas vesicles collapsed). The mean critical collapse pressure was calculated from the pressure-collapse curves [1,20].

3 Results and discussion

Comparison of the pressure-collapse curves for intact Halobacterium cells and for gas vesicles isolated from these cells shows that considerably less pressure is required for the collapse of isolated gas vesicles than for gas vesicles in intact cells. Representative results are shown in Fig. 1. For cells grown on agar plates (procedure (a) in Section 2.2), the mean critical collapse pressure was 0.095 ± 0.005 MPa (mean ± SD, n= 3 independent experiments), while 0.058 ± 0.003 MPa (n= 3) was sufficient for the collapse of half of the isolated vesicles. Slightly lower pressures were required to cause collapse of the gas vesicles in cells collected from liquid cultures (procedure (b) mean critical collapse pressure of vesicles in intact cells 0.082 ± 0.010 MPa (n= 3), and 0.049 ± 0.007 MPa (n= 3) for isolated vesicles). A similar weakening was found in gas vesicles isolated by alkaline lysis of the cells in the presence of high salt concentrations (procedure (c)). The values obtained for intact cells were in the same range as those earlier reported for Halobacterium [1,14,20].

1

Typical critical pressure distribution of gas vesicles from Halobacterium NRC-1 grown on agar plates (upper panel) or in liquid medium (lower panel). Pressure-collapse curves were prepared of intact cells (•) and of vesicles purified by enhanced flotation and suspended in water (□), in 4 M NaCl (δ), and in 4 M KCl (□). Values were calculated from OD₆₀₀ data so that the OD of unpressurized cells or vesicles = 100% intact vesicles, and fully collapsed = 0%. The OD₆₀₀ of suspensions of pressurized cells was typically about 30–35% of that of intact cells; pressurization of preparations of isolated gas vesicles led to a turbidity loss of about 95%.

When we initiated these experiments we did not expect to find significant differences. In cyanobacteria, which maintain a turgor pressure intracellularly, considerably more pressure is required to achieve collapse of isolated gas vesicles, or gas vesicles in cells in which the turgor pressure had been abolished by suspension in concentrated sucrose solutions, than in cells living in their normal environment [1,20]. As halophilic Archaea do not maintain a measurable turgor pressure [1,20–22], the mean critical collapse pressure of gas vesicles inside and outside the cells should theoretically be identical.

It cannot be expected that the fragile glycoprotein subunit cell wall of Halobacterium can maintain a difference in pressure inside and outside the cells when pressure is applied. The apparent weakening of isolated gas vesicles could also not be attributed to damage by the accelerated flotation procedure used for their isolation: the same weakening of the gas vesicles was observed in lysates prior to centrifugation. Therefore, the explanation for the observed phenomenon should be sought in the properties of the gas vesicles themselves.

Gas vesicles are known to contain two major structural proteins, GvpA and GvpC. GvpA is the major constituent of the vesicles. This extremely hydrophobic protein constitutes about 90% of the mass of the vesicles. Its primary structure is one of the main factors that determine the strength of the vesicles [14]. GvpC is a hydrophilic protein that has 4–7 repeats of about 33 amino acids. Extensive research on cyanobacterial gas vesicles has shown that this protein is located at the outer surface of the vesicles. It is generally assumed that GvpC strengthens the structure of the vesicles from the outside, the repeated sequences spanning several ribs made of GvpA [23–25]. Treatment of cyanobacterial gas vesicles with 2% sodium dodecyl sulfate or 6 M urea led to the detachment of GvpC from the vesicles, with a concomitant decrease in their critical collapse pressure [23–26]. Thus, Anabaena gas vesicles stripped of GvpC by urea treatment had a mean critical collapse pressure of 0.190 MPa, as compared to 0.557 MPa of native vesicles [24].

The true function of GvpC in the gas vesicles of halophilic Archaea is less clear. gvpC is not among the 8 out of the 14 identified genes involved in gas vesicle formation found to be essential for the synthesis of functional gas vesicles in Hfx. mediterranei [6]. However, GvpC was shown to be required for the production of gas vesicles of constant diameter [15]. The intramolecular repeats within the GvpC protein are much less clear than in cyanobacterial GvpC, and sequence homology between the GvpCs is low [3,13]. Using immunoblotting with antibodies raised against either recombinant Halobacterium GvpC or a LacZ-GvpC fusion protein, presence of GvpC in gas vesicle-containing cells and/or in isolated from Hfx. mediterranei and Hbt. salinarum has been ascertained [4,13].

If indeed GvpC strengthens the Halobacterium gas vesicles from the outside, weakening of the vesicles isolated by hypotonic lysis of the cells can be explained on the basis of the halophilic nature of the GvpC protein. Most proteins of halophilic Archaea are highly salt-dependent, and denature when the salt concentration is reduced. Such proteins are characterized by a large excess of acidic over basic amino acids, giving them a large negative charge at physiological pH values [2,27,28]. Only a few Halobacterium proteins do not show such a dependency on salt, notable exceptions being the retinal proteins bacteriorhodopsin and halorhodopsin and the structural gas vesicle protein GvpA. GvpC is not one of these exceptional cases: halobacterial GvpC shows a great excess of acidic amino acids, and accordingly has a much lower isoelectric point than corresponding proteins of non-halophiles. The pI values (calculated using the algorithm given in http://www.embl-heidelberg.de/cgi/piwrapper.pl) of GvpC of Halobacterium NRC-1 (chromomsomal and plasmid-encoded) and Hfx. mediterranei are 3.53, 4.06 and 3.77, respectively (GenBank Accession Nos. P24574, Q9HHT0, and Q02228), while the pI values for GvpC proteins of the cyanobacteria Microcystis aeruginosa, Planktothrix rubescens, Planktothrix agardhii, Anabaena flos-aquae, Anabaena lemmermannii, and Fremyella diplosiphon are 9.97, 8.56, 4.60, 5.22, 5.22, and 5.16, respectively (GenBank Accession Nos. CAE11901, CAD41964, CAB59524, P09413, AAX37438, and P08041). It is therefore probable that halophilic GvpC will denature in the presence of molar concentrations of salt. It will then become detached from the gas vesicles, which will become weaker in the process. If some GvpC can be lost even during the preparation of gas vesicles from cyanobacteria [23], this will be true even more in the case of the halophiles. Adding molar concentrations of salt (KCl, NaCl) to the H₂O-purified vesicles from cells (procedure (b) in Section 2.2) led to a slight increase (0.005–0.009 MPa) in the mean critical collapse pressure (Fig. 1, lower panel). A possible explanation is that not all GvpC had become detached during isolation and purification of the gas vesicles, and that reconstitution is to some extent possible. That at least some GvpC is still present in purified gas vesicles is also proven by the positive reaction of such gas vesicles with antibodies raised against GvpC [13]. Alternatively, the effect may be due to salt-dependent conformational changes on the structural proteins of the vesicles. No significant strengthening was found by adding NaCl or KCl to vesicles isolated using procedure (a) (Fig. 1, upper panel).

A number of other proteins (GvpF, GvpG, GvpJ, GvpL, and GvpM) were recently found to be associated with intact gas vesicles of Halobacterium NRC-1 [29]. To what extent these proteins may be involved in the strengthening of the vesicles is not yet known. The same explanation brought forward above based on detachment of GvpC can theoretically be applied to any of these proteins if indeed their presence contributes to the mechanical properties of the vesicles. Comparison of the calculated pI values of these minor proteins of Halobacterium NRC-1 with equivalents from the genome of Nostoc sp. PCC 7120 shows that most of the Halobacterium proteins are more acidic than those of Nostoc [pI values for GvpF 3.97 vs. 4.97; for GvpJ 3.59 vs. 4.53; for GvpL 4.16 vs. 6.53 (GenBank Accession Nos. VNG7022, all2248, VNG7018, all2250, VNG7016 and all2245, respectively); the gvpG equivalent of Nostoc (all2247) is more acidic (pI 3.69) than that of Halobacterium (VNG7021) (pI 4.07); no equivalent for Halobacterium GvpM (VNG7015) was reported in Nostoc].

No protocol has yet been developed for the isolation of intact gas vesicles of halophilic Archaea that does not involve removal of salt or extremely alkaline conditions that can otherwise influence associations between proteins. Lysis of the cells by bile acids [30,31] destroys the gas vesicles as well. There is also no simple way to assess what fraction of the GvpC is still present in the purified gas vesicles. Conventional methods of protein electrophoresis cannot be used to assess the relative abundance of GvpC and GvpA as the extremely hydrophobic GvpA does not enter polyacrylamide gels. The lack of phenylalanine in cyanobacterial GvpA and the great difference in glycine content of the two proteins has enabled the estimation of the GvpC/GvpA ratio in gas vesicles isolated from cyanobacteria to be about 1:33 [11]. The amino acid composition of halophilic archaeal GvpA is less distinctive, so calculations based on the abundance of amino acids that occur predominantly or exclusively in one protein or the other cannot be used for Halobacterium gas vesicles.

The results presented in this paper show that the procedures commonly used to isolate and purify gas vesicles from halophilic Archaea do not yield vesicles with mechanical properties equivalent to those in the intact cell. Whether the effect is due to loss of GvpC or any minor structural components of the gas vesicles remains to be ascertained.

Acknowledgements

We are grateful to A.E. Walsby (University of Bristol) and to two anonymous reviewers for helpful comments. We further thank the staff of the Interuniversity Institute for Marine Sciences of Eilat for logistic support. This study was supported by the Israel Science Foundation (Grant No. 504/03).