ABSTRACT
The outer membrane of heterotrophic Gram-negative bacteria plays the role of a selective permeability barrier that prevents the influx of toxic compounds while allowing the nonspecific passage of small hydrophilic nutrients through porin channels. Compared with heterotrophic Gram-negative bacteria, the outer membrane properties of cyanobacteria, which are Gram-negative photoautotrophs, are not clearly understood. In this study, using small carbohydrates, amino acids, and inorganic ions as permeation probes, we determined the outer membrane permeability of Synechocystis sp. strain PCC 6803 in intact cells and in proteoliposomes reconstituted with outer membrane proteins. The permeability of this cyanobacterium was >20-fold lower than that of Escherichia coli. The predominant outer membrane proteins Slr1841, Slr1908, and Slr0042 were not permeable to organic nutrients and allowed only the passage of inorganic ions. Only the less abundant outer membrane protein Slr1270, a homolog of the E. coli export channel TolC, was permeable to organic solutes. The activity of Slr1270 as a channel was verified in a recombinant Slr1270-producing E. coli outer membrane. The lack of putative porins and the low outer membrane permeability appear to suit the cyanobacterial autotrophic lifestyle; the highly impermeable outer membrane would be advantageous to cellular survival by protecting the cell from toxic compounds, especially when the cellular physiology is not dependent on the uptake of organic nutrients.
IMPORTANCE Because the outer membrane of Gram-negative bacteria affects the flux rates for various substances into and out of the cell, its permeability is closely associated with cellular physiology. The outer membrane properties of cyanobacteria, which are photoautotrophic Gram-negative bacteria, are not clearly understood. Here, we examined the outer membrane of Synechocystis sp. strain PCC 6803. We revealed that it is relatively permeable to inorganic ions but is markedly less permeable to organic nutrients, with >20-fold lower permeability than the outer membrane of Escherichia coli. Such permeability appears to fit the cyanobacterial lifestyle, in which the diffusion pathway for inorganic solutes may suffice to sustain the autotrophic physiology, illustrating a link between outer membrane permeability and the cellular lifestyle.
INTRODUCTION
The outer membrane of Gram-negative bacteria plays the role of a selective permeability barrier that prevents the entry of toxic compounds into the cell while allowing the passage of nutrients essential for cell proliferation (1). The porin channels are responsible for nonspecific diffusion of small hydrophilic compounds across the outer membrane. Classical porins, such as OmpF and OmpC of Escherichia coli, form a transmembrane β-barrel structure and allow the passage of hydrophilic compounds with Mr values of <600 (1). This type of channel is widely distributed among heterotrophic Gram-negative bacteria, providing a diffusion route for nutrients (which are often small and hydrophilic) across the outer membrane. We recently demonstrated the presence of porin-type diffusion channels also in bacteria belonging to the class Negativicutes, a taxon classified as Gram positive based on nucleotide sequence-based phylogeny but known for having Gram-negative cell surface structure (2, 3).
The outer membrane permeability and the existence of classical porins in cyanobacteria are less well understood and are largely under debate. Electrophysiological studies demonstrated the presence of inorganic ion-permeable channels in the cyanobacterial outer membrane but did not afford a straightforward suggestion about the permeability (4–6). Channel conductance of the major outer membrane proteins SomA and SomB of Synechococcus sp. strain PCC 6301 was first reported as 5.5 nS in 1 M KCl (5) but was later corrected to 0.4 to 0.9 nS by the same group (6). The outer membrane of Anabaena variabilis contains two pore-forming units, with conductances of 0.2 nS and 3.5 nS in 1 M KCl (4). Similarly, conductance values of 0.3 nS and 2.2 nS in 1 M KCl were reported for the outer membrane of Synechocystis sp. PCC 6714 (7). The presence of multiple ion-conducting units led to conflicting interpretations, arguing for the possible presence (4) or absence (6, 8) of porins on the basis of comparisons with the conductance of porin channels. Notably, the conductance value of 2 to 3 nS in 1 M KCl for trimeric porin channels was sometimes used as a reference (4, 6), rather than the single-channel conductance of ∼0.7 nS in 1 M KCl (1, 9). Moreover, it became clear that channel conductance does not always correlate with pore size and channel permeability, e.g., the channel conductance value of a mutant OmpF channel with an enlarged pore size was reduced (9). On the other hand, the existence of classical porins in Anabaena sp. strain PCC 7120 was proposed based on the ethidium bromide uptake rate of intact cells (10). However, the uptake rate does not directly represent the porin-mediated diffusion rate, because it incorporates both permeation through the lipid bilayer of the outer membrane and that across the cytoplasmic membrane. The existence of a cyanobacterial porin remains elusive.
To characterize the permeability of the cyanobacterial outer membrane, permeability measurements with small organic nutrients appear necessary; however, such studies have not yet been conducted. In the current study, using small carbohydrates, amino acids, and inorganic ions as permeability probes, we determined the outer membrane permeability of Synechocystis sp. strain PCC 6803 (referred to as PCC 6803) in an assay using intact cells and proteoliposomes reconstituted with the outer membrane proteins. We found that the permeability of the PCC 6803 outer membrane was more than 20-fold lower than that of E. coli. The most abundant proteins, Slr1841, Slr1908, and Slr0042, were permeable to inorganic ions but not to organic nutrients. Only the less abundant membrane protein Slr1270 was permeable to organic solutes. None of the proteins exhibited porin-like properties.
RESULTS
Outer membrane permeability of intact cells: measurements with radiolabeled dextran and sucrose.Porin-mediated diffusion allows the flux of small substrates across the outer membrane while restricting the penetration of large compounds. This molecular sieving activity can be assessed by quantifying the penetration of small hydrophilic substrates (i.e., small carbohydrates) into the periplasm and comparing that with the penetration of large, outer membrane-impermeable compounds (i.e., dextran), as described by Decad and Nikaido (11). We used [14C]sucrose and [3H]dextran (Mr of ∼10,000) as permeability probes. Intact PCC 6803 or E. coli K-12 cells were incubated with those substrates, and the penetration of the substrates into the periplasm was determined and compared (Fig. 1). The extent of penetration was quantified by calculating the volume of permeable space for each substrate across the outer membrane (see Materials and Methods). In E. coli, the penetration of sucrose was ∼1.6-fold greater than that of dextran, in good agreement with a previous study (11). In contrast, no significant difference in penetration was observed for PCC 6803 cells, indicating that the PCC 6803 outer membrane does not function as a sieve for sucrose.
Relative permeable space values for [14C]sucrose and [3H]dextran in E. coli K-12 and PCC 6803 cells. The data are presented as means ± standard deviations (SDs) from triplicate measurements. The statistical significance was evaluated with the Student t test.
Permeability measurements with proteoliposomes reconstituted with outer membrane proteins, revealing low permeability of the PCC 6803 outer membrane.A possible explanation for the lack of sieving function described above was low permeability of the PCC 6803 outer membrane that restricted the influx of both sucrose and dextran. To test this assumption and to evaluate the outer membrane permeability of other substrates, we prepared proteoliposomes reconstituted with the outer membrane proteins of PCC 6803. The protein composition of the outer membrane preparation was examined by SDS-PAGE (Fig. 2A). Five proteins were readily detectable by Coomassie brilliant blue (CBB) staining, and they were identified by time of flight tandem mass spectrometry (TOF MS/MS) analysis. Slr1841, Slr1908, and Slr0042 proteins are homologs of SomA and SomB of Synechococcus sp. PCC 6301 (6). These proteins contain an N-terminal S-layer homologous (SLH) domain and a C-terminal transmembrane region that is predicted to form a β-barrel structure. In our previous study, we demonstrated that the SLH domains of these proteins bind to the polysaccharide moiety of the peptidoglycan, providing a physical linkage between the outer membrane and the peptidoglycan, which presumably plays a role in maintaining the structural stability of the outer membrane (12). Slr1270 is a homolog of the TolC export channel of E. coli, which is widely distributed in the outer membrane of Gram-negative bacteria and is involved in the export of a wide variety of substances (13–15). Finally, Sll1951 is an S-layer protein of PCC 6803 that is attached to the outer membrane via noncovalent interactions (16). This protein cannot be removed from an outer membrane preparation unless it is treated with a chaotropic agent (16). The overall protein profile of the outer membrane preparation was essentially consistent with previously reported findings (17, 18). We roughly estimated the relative abundance of these five proteins by quantifying the intensity of protein bands on SDS-PAGE gels. The relative abundance ratio of Sll1951, Slr1841, Slr1270, Slr1908, and Slr0042 proteins was 1.5:5.5:0.4:2.3:0.3.
Permeability of the outer membrane proteins of PCC 6803 and E. coli cells, as measured by a liposome swelling assay. (A and B) SDS-PAGE analysis of the solubilized outer membrane proteins of PCC 6803 (A) and E. coli (B). The E. coli proteins were analyzed on a 10% gel containing 6 M urea to facilitate the separation of OmpF and OmpC. PCC 6803 proteins were analyzed on a 10% gel without urea. The gels were stained with CBB. M, molecular mass standard. (C and D) Channel activities of the outer membrane proteins of PCC 6803 (C) and E. coli (D). Various amounts of proteins were reconstituted with the liposomes, and the permeation rates for arabinose and NaCl were determined. The data are presented as means ± SDs from three independent experiments.
The permeability of the proteoliposomes was measured by the liposome swelling assay, in which the swelling of the proteoliposomes caused by the influx of tested solutes is monitored by measuring the reduction of the optical density at 400 nm (OD400) of the reaction mixture (19). We first used arabinose as a substrate and examined its rates of permeation into proteoliposomes reconstituted with different amounts of proteins (Fig. 2C). As a reference, we used proteoliposomes reconstituted with the E. coli outer membrane proteins (Fig. 2B and D). As expected, the permeability of the PCC 6803 proteoliposomes was more than 20-fold lower than that of E. coli proteoliposomes, i.e., the permeation of arabinose or NaCl produced by 20 μg of PCC 6803 proteins was less than that produced by 1 μg of E. coli proteins. The low permeability was not caused by the inefficiency of reconstitution, since we confirmed that ∼90% of the proteins were successfully reconstituted into the liposomes (data not shown). On the other hand, rapid permeation was observed when NaCl was used as a substrate, confirming the presence of ion-permeable channels shown in previous studies (4, 6). It should be noted that the NaCl permeation rates were underestimated by this assay, because a 2-fold lower concentration of NaCl, compared with arabinose, had to be used to adjust the osmotic balance in the reaction mixtures. Essentially the same results were obtained with other substrates, such as sucrose, several amino acids, and other inorganic ions, e.g., NaHCO3 (data not shown). These experiments demonstrated that the outer membrane of PCC 6803 is permeable to inorganic ions but is markedly less permeable to small organic compounds, providing the explanation for the lack of sucrose penetration of intact cells (Fig. 1).
Characterization of the modest permeability of the PCC 6803 outer membrane to small organic nutrients.The results described above suggested that porin channels were probably not present in the outer membrane of PCC 6803. However, we observed modest permeation of organic compounds when large amounts of proteins (≥20 μg) were reconstituted into the liposomes (Fig. 2C). To characterize the observed permeability, we prepared proteoliposomes reconstituted with 30 μg of the PCC 6803 outer membrane proteins, and then we tested the permeability of a wide variety of substrates (Fig. 3). Proteoliposomes reconstituted with 1 μg of E. coli outer membrane proteins were used again as a reference. No obvious substrate specificity was observed, and the permeability seemed to solely depend on the size (molecular weight) of the substrates. This suggested the existence of a nonspecific diffusion channel in the PCC 6803 outer membrane preparation, although the protein level may be quite low. The expected pore size of this diffusion channel, however, seemed to be quite different from that of E. coli porins, whose pore radii are ∼0.58 nm (19), because the PCC 6803 channel allowed more rapid permeation of large solutes (i.e., lactose, sucrose, raffinose, and stachyose) than did porins.
Relative rates of permeation of sugars and amino acids into proteoliposomes reconstituted with outer membrane proteins of PCC 6803 (30 μg) and E. coli (1 μg). The permeation rates are presented as relative values, compared to the rate for arabinose. The substrates were as follows: 1, Ala (Mr of 89.09); 2, l-Pro (Mr of 115.13); 3, l-Ile (Mr of 131.17); 4, l-Asn (Mr of 132.12); 5, arabinose (Mr of 150.13); 6, l-Phe (Mr of 165.19); 7, glucose (Mr of 180.16); 8, N-acetylglucosamine (GlcNAc) (Mr of 221); 9, lactose (Mr of 342.3); 10, sucrose (Mr of 342.3); 11, raffinose (Mr of 594.53); 12, stachyose (Mr of 666.8). The data are presented as means from three independent experiments.
Responsibility of Slr1270 for the modest permeability of small organic nutrients.While attempting to identify the PCC 6803 diffusion channel permeable to organic nutrients, we noticed that the activity of this channel could be enriched in a detergent-extracted fraction of the outer membrane. When the outer membrane preparation was treated with 2% lithium dodecyl sulfate (LDS) at 37°C, the majority of proteins remained insoluble but some portion was solubilized and extracted to the supernatant (Fig. 4A). The diffusion channel activity was enriched in this supernatant, as revealed by the liposome swelling assay using 2 μg of proteins from the supernatant or the insoluble fraction (Fig. 4B). The relative protein abundance in the supernatant was roughly estimated by comparing the intensity of protein bands on SDS-PAGE gels. The relative abundance ratio of Sll1951, Slr1841, Slr1270, and Slr1908 was 3.6:2.6:2.7:1.1. Compared with the outer membrane preparation (Fig. 2A), Slr1270 was highly enriched in supernatant, by about 7-fold. This finding led us to focus on Slr1270 as a plausible candidate for the diffusion channel.
Enrichment of organic-nutrient-permeable protein channels of PCC 6803 in detergent-extracted supernatant. The outer membrane proteins were treated with 2% LDS at 37°C, and the supernatant (sup) and insoluble fraction (ppt) were separated by centrifugation at 20,000 × g for 20 min. The protein composition and channel activity were examined by SDS-PAGE (A) and the liposome swelling assay (B), respectively. The gel was stained with CBB. M, molecular mass standard. The permeability was determined using proteoliposomes reconstituted with 2 μg of proteins, with arabinose as a substrate. The data are presented as means ± SDs from three independent experiments. ND, not detected.
We examined the function of Slr1270 by performing permeability measurements with (i) the outer membrane preparation of a PCC 6803 Δslr1270 mutant and (ii) the outer membrane of E. coli LA51FCA producing the recombinant Slr1270 protein. LA51FCA is a mutant strain that lacks major porins (OmpF and OmpC) and a multidrug efflux pump (AcrAB) (20). As expected, the proteoliposomes reconstituted with the Δslr1270 mutant outer membrane were not permeable to arabinose (Fig. 5). However, because the Δslr1270 mutation is accompanied by loss of the S-layer, which is mainly composed of Sll1951 (14), we went on to check whether Sll1951 was responsible for the channel activity. We purified the Sll1951 protein and, using a liposome swelling assay, confirmed that Sll1951 did not possess channel activity (Fig. 6).
Outer membrane permeability of the Δslr1270 mutant. (A) SDS-PAGE analysis of the outer membrane proteins of the Δslr1270 mutant. The gel was stained with CBB. M, molecular mass standard. (B) Permeability measurements using a liposome swelling assay with proteoliposomes reconstituted with 20 μg of the outer membrane proteins of PCC 6803 wild-type (WT) and Δslr1270 mutant cells. Arabinose was used as a substrate. The data are presented as means ± SDs from three independent experiments.
Permeability measurements using purified Sll1951 protein. SDS-PAGE analysis of the purified Sll1951 protein (left) and results of a liposome swelling assay (right) with proteoliposomes reconstituted with 1 μg of protein, with arabinose as a substrate, are shown. The gel was stained with CBB. M, molecular mass standard. The data are presented as means ± SDs from three independent experiments.
The presence of the Slr1270 protein in the LA51FCA outer membrane was examined by SDS-PAGE (Fig. 7A). Although the produced Slr1270 was not highly abundant, an ∼60-kDa protein band was clearly visible, and this band was verified as Slr1270 by TOF MS/MS analysis. We first examined the permeability of the Slr1270-containing outer membrane by the liposome swelling assay (Fig. 7B); the permeability of arabinose was more than 5-fold higher than that of the outer membrane lacking Slr1270. We confirmed the increased permeability by a benzylpenicillin (PEN) influx assay using intact cells that produced or did not produce Slr1270 (Fig. 7C). The PEN influx rate was ∼4-fold higher in cells with Slr1270 than in cells lacking Slr1270 (Fig. 7C). These results verified that Slr1270 acts as a nonspecific diffusion channel.
Channel activity of a recombinant Slr1270 protein produced in LA51FCA cells. (A) SDS-PAGE analysis of the outer membrane proteins of a Slr1270-producing LA51FCA strain and a LA51FCA strain harboring an empty vector. The gel was stained with CBB. M, molecular mass standard. (B) Permeability measurements using a liposome swelling assay with proteoliposomes reconstituted with 10 μg of the outer membrane proteins of the Slr1270-producing LA51FCA strain and the LA51FCA strain harboring an empty vector. Arabinose was used as a substrate. The data are presented as means ± SDs from three independent experiments. (C) PEN influx assay using intact cells of the Slr1270-producing LA51FCA strain and the LA51FCA strain harboring an empty vector. The influx rate (Vin) was measured using 200 μM PEN. The data are presented as means ± SDs from three independent measurements.
Lack of effect of deletion of slr1270 on the inorganic solute permeability of the outer membrane.We then examined whether the ion permeability of the outer membrane was affected by deletion of the slr1270 gene. We used proteoliposomes reconstituted with the Δslr1270 mutant outer membrane proteins, which contain Slr1841, Slr1908, and Slr0042 as main components (Fig. 5A), and we tested the permeability of various inorganic solutes by the liposome swelling assay (Fig. 8). No significant differences were observed between the wild-type and Δslr1270 mutant preparations. Thus, outer membrane proteins other than Slr1270, i.e., Slr1841, Slr1908, or Slr0042, were suggested to be responsible for ion permeation in PCC 6803. This appeared plausible, because Slr1841, Slr1908, and Slr0042 are all homologs of SomA and SomB, whose ion-permeating function was demonstrated by Hansel and Tadros (6).
Inorganic ion permeability of the outer membrane of a Δslr1270 mutant. The permeability was measured using a liposome swelling assay with proteoliposomes reconstituted with 20 μg of the outer membrane proteins of PCC 6803 wild-type (WT) and Δslr1270 mutant strains. The data are presented as means ± SDs from three independent experiments.
DISCUSSION
In the current study, we showed that the permeability of the PCC 6803 outer membrane relies on two different types of channels (Fig. 9). (i) The predominant proteins Slr1841, Slr1908, and Slr0042, which are SomA and SomB homologs, are not permeable to organic nutrients but allow the diffusion of inorganic ions; these protein channels are highly abundant in the outer membrane and are estimated to represent ∼80% of all proteins of the outer membrane preparation. (ii) Slr1270, a TolC homolog, allows the diffusion of small organic nutrients; the relative protein abundance of this channel in the outer membrane preparations is quite low (∼4%). The predominance of inorganic-ion-permeable channels renders the outer membrane relatively permeable to inorganic ions but markedly less permeable to organic nutrients. The overall permeability of the outer membrane is >20-fold lower than in E. coli. Our results are consistent with previously published electrophysiological studies that demonstrated the presence of two pore-forming units in the cyanobacterial outer membrane (4, 7); because the channel properties of Slr1841/Slr1908/Slr0042 and Slr1270 proteins differ significantly, these channels most likely constitute two different ion-conducting units in electrophysiological measurements.
Graphical summary of the permeability characteristics of the PCC 6803 outer membrane. The most abundant outer membrane proteins, Slr1841, Slr1908, and Slr0042 (representing ∼80% of the total outer membrane proteins), are not permeable to organic nutrients but allow the diffusion of inorganic ions. Slr1270 allows the diffusion of small organic nutrients; however, its relative abundance in the outer membrane is low (∼4%). The overall permeability of the PCC 6803 outer membrane is >20-fold lower than that of E. coli.
The lack of classical porins and the intrinsically low outer membrane permeability appear to fit well with the cyanobacterial lifestyle proposed by Hoiczyk and Hansel (8), which might not require the uptake pathway for organic nutrients, with channels only large enough to allow the diffusion of small solutes (e.g., inorganic ions) sufficing to sustain the photoautotrophic physiology. The low permeability of the outer membrane would be advantageous, as it would protect the cell from toxic compounds. For instance, the development of antibiotic resistance in the nosocomial pathogen Klebsiella pneumoniae is often accompanied by the mutational loss of its porins, which results in the production of an outer membrane highly impermeable to antibiotics (21). The low permeability may also facilitate the retention of essential metabolites within the cell. Indeed, Nicolaisen et al. (22) reported that impaired outer membrane biogenesis in Anabaena sp. PCC 7120 resulted not only in increased susceptibility to harmful compounds (e.g., antibiotics) but also in increased uptake of small sugars and amino acids from the external environment. The authors proposed that the outer membrane plays the role of a permeability barrier preventing the leakage of small nutrients. Our results are in good agreement with those observations.
It is well established that TolC of E. coli functions as an outer membrane channel component of various transenvelope exporter complexes (13). In cyanobacteria, a phenotypic analysis of the PCC 6803 Δslr1270 mutant strain by Oliveira et al. (14) suggested that Slr1270 is involved in the secretion of a wide variety of substances. Furthermore, a TolC homolog of Anabaena sp. strain PCC 7120, HgdD, was reported to participate in a tripartite complex of a resistance-nodulation-cell division-type efflux transporter and to be involved in multidrug resistance (23). One specific difference between Slr1270 and TolC revealed by this study is that Slr1270 shows diffusion channel activity for small organic compounds; this is in contrast to TolC, which exists as a closed channel unless it is assembled into an exporter complex where it exists in an open state (24, 25). In TolC, a hydrogen bond network composed of T152, D153, Y362, and R367 residues located in the periplasmic coiled-coil region is mostly responsible for constricting the channel and creating the closed state. In contrast, this hydrogen bond network is not likely to be present in Slr1270 because of poor conservation of these residues (i.e., T152F, Y362F, and R367G). This may constitute one explanation for the open state of Slr1270. However, the proportion of the Slr1270 protein population in intact cells that plays the role of a diffusion channel in addition to being an exporter requires further elucidation. In E. coli, only a small fraction of the intracellular TolC population is suggested to be involved in multidrug efflux transporters (26), but such studies have not been performed in cyanobacteria.
Heterotrophic Gram-negative bacteria generally possess multiple porin channels with slightly different properties (e.g., pore size or substrate specificity), such as OmpF and OmpC in E. coli (1, 20). In this regard, it would be interesting to assess the differences (if any) in the channel properties of Slr1841, Slr1908, and Slr0042 proteins. However, our attempts to generate the corresponding deletion mutants were hindered by the fact that the PCC 6803 genome contains an additional three paralogs of slr1841, slr1908, and slr0042; these paralogs are usually cryptic but are induced upon deletion of the slr1841, slr1908, or slr0042 genes; for example, one of the paralogs, Sll1550, is induced when the slr1908 gene is deleted (data not shown). Our attempts to isolate these proteins from the outer membrane preparations and our attempts to produce heterologous gene expression in E. coli were unsuccessful. Further studies, utilizing a different approach, are required for a detailed understanding of the channel properties of these proteins.
Lastly, we propose that elucidation of the outer membrane permeability characteristics of cyanobacteria not only will lead to a detailed understanding of the cyanobacterial lifestyle but also will shed light on the evolutionary process of chloroplast generation. It is well accepted that chloroplasts originated from an endosymbiotic cyanobacterium (27). We recently discovered that the diffusion channel proteins CppS and CppF are the dominant proteins of the outer membrane of a primitive chloroplast of Cyanophora paradoxa; strikingly, these proteins are apparently derived from a noncyanobacterial lineage (12). In contrast, the majority of cyanobacterial outer membrane proteins are not conserved in an extant plant lineage. Establishment of the diffusion route for various substances across the outer membrane could have constituted an essential step in the conversion of an endosymbiotic cyanobacterium to a chloroplast, if we can assume low permeability of the outer membrane in the ancestral cyanobacterium.
MATERIALS AND METHODS
Strains and culture conditions.PCC 6803 and an Δslr1270 mutant (14) were cultured at 30°C in BG11 medium, under continuous light. E. coli K-12 (20), DH5α, and LA51FCA (20) strains were grown at 37°C in LB medium supplemented with 5 mM MgCl2.
Measurement of outer membrane permeability in intact cells using radiolabeled dextran and sucrose.The protocol was essentially the same as described by Decad and Nikaido (11). E. coli and PCC 6803 cells were grown in 15 ml of the appropriate medium; the cells were harvested at an OD600 of 1.6 for E. coli and an OD730 of 1.2 for PCC 6803, by centrifugation at 2,500 × g for 10 min at room temperature (RT). The cell pellets were washed once with 50 mM potassium phosphate buffer (pH 7.0) supplemented with 5 mM MgCl2 (K-P buffer). The cells were then suspended in 0.5 ml of 25 mM potassium phosphate buffer (pH 7.0) containing 0.3 M NaCl, 5 mM MgCl2, 10 mM sucrose, 20 mM raffinose, and radiolabeled substrates (0.5 μCi of [14C]sucrose or 0.4 μCi of [3H]dextran). The suspensions were vortex-mixed for 5 s. The mixtures were then centrifuged at 4,500 rpm for 10 min at RT. After the supernatants (supernatant I) were removed, the leftover liquid on the walls of the tubes was removed using a tissue and the cell pellets were suspended in 1 ml of K-P buffer. After vortex mixing for 5 s, the suspensions were centrifuged at 2,500 × g for 10 min at RT, and the supernatants (supernatant II) were collected. The radioactivity of supernatants I and II was determined by liquid scintillation counting. The permeable space volume (in microliters) for each substrate was calculated as follows: permeable space = (radioactivity in 400 μl of supernatant II/radioactivity in 100 μl of supernatant I) × 250.
Preparation of PCC 6803 outer membrane proteins.The outer membrane was isolated according to the method of Weckesser and Jürgens (28). The cells were grown in 200 ml of medium until the cultures reached an OD730 of ∼1.0. The cells were harvested by centrifugation at 5,000 × g for 10 min at RT and were washed once with 20 mM Tris-HCl buffer at pH 7.5 (Tris buffer). The cells were then disrupted using a French pressure cell press at 14,000 lb/in2, and unbroken cells were removed by centrifugation at 5,000 × g for 5 min at 4°C. The crude envelope preparations were sedimented by centrifugation at 20,000 × g for 30 min at 4°C; they were then loaded onto a sucrose density gradient (60%, 55%, 50%, 45%, and 40% [wt/vol] sucrose in Tris buffer) and centrifuged at 20,000 rpm for 4 h at 4°C. The outer membrane was collected from the 60% sucrose fraction and the pellets at the bottom of the tubes and was washed twice with Tris buffer. To solubilize the proteins, the outer membrane preparation obtained was digested overnight at 37°C with 100 μg/ml lysozyme in Tris buffer. The peptidoglycan-digested preparation was collected by centrifugation at 20,000 × g for 30 min at RT. The outer membrane proteins were solubilized for 70 min at 37°C with 2% LDS in Tris buffer containing 0.5 M LiCl and 10 mM EDTA. The supernatant was collected after centrifugation at 20,000 × g for 30 min at RT. To remove excess LiCl, the solubilized outer membrane proteins were dialyzed against 10 mM Tris-HCl buffer (pH 7.5) overnight at 4°C. Protein composition was analyzed by SDS-PAGE, and relative protein abundance was estimated by quantifying the intensity of protein bands on SDS-PAGE gels using ImageJ software.
Preparation of E. coli outer membrane proteins.The outer membrane was prepared as described previously (20), and the proteins were solubilized as described above.
Protein identification by mass spectrometry.Protein bands from SDS-PAGE gels were identified by TOF MS/MS analysis as described previously (12).
Liposome swelling assay.To determine the permeability of the outer membrane proteins, the solubilized proteins were reconstituted into liposomes containing 6.2 nmol of phosphatidylcholine and 0.2 nmol of dicetylphosphate, according to the protocol described by Nikaido and Rosenberg (19). The efficiency of reconstitution of proteins into liposomes was analyzed by the method of Yoshimura et al. (29). The reconstituted proteoliposomes were suspended in NAD-imidazole buffer (pH 6.1), and the assay was performed as described previously (2). To determine the permeability of the purified Sll1951 protein and of the outer membrane of Slr1270-producing LA51FCA cells, the proteins were reconstituted into liposomes containing 2.4 μmol of phosphatidylcholine and 0.15 μmol of dicetylphosphate. The proteoliposomes were suspended in Tris-HCl buffer (pH 8.0) containing 15% dextran (Mr of ∼40,000), and the assay was performed according to the method described by Nikaido et al. (30).
Purification of the Sll1951 protein.The solubilized PCC 6803 outer membrane proteins were treated with 75 mM dithiothreitol for 15 min at RT and then incubated with 0.5 M (final concentration) iodoacetamide for 15 min at RT. The Sll1951 protein was purified by gel filtration–high-performance liquid chromatography. The resolution conditions were as follows: column, Superdex 200 Increase 10/300 GL; eluent, 10 mM Tris-HCl (pH 7.5) containing 0.1% LDS and 0.4 M LiCl; flow rate, 0.5 ml/min. The purified Slr1951 protein was collected at an elution time of ∼18 min.
Production of the Slr1270 protein in a heterologous E. coli host.To construct an slr1270 expression vector, the slr1270 coding sequence was amplified from the chromosomal DNA of PCC 6803. It was then fused by overlapping PCR with the upstream region of the E. coli ompF gene (base positions −213 to +66), harboring the promoter and signal sequences. The construct was then inserted into the pHSG575 vector. The expression of the slr1270 gene (and production of the Slr1270 protein) was under the control of the ompF promoter, as described previously (20).
PEN influx assay with intact cells.The Slr1270-producing LA51FCA strain was cultured in LB medium, and its outer membrane permeability was examined by measuring the PEN influx rate by coupling the periplasmic hydrolysis of PEN by AmpC β-lactamase. The experimental procedure was the same as described previously (31). For this assay, we used freshly transformed cells (cells transformed with Slr1270-expressing vector 1 day before the experiments and never exposed to a low temperature), because old transformants sometimes gave scattered data for unknown reasons.
ACKNOWLEDGMENTS
We are grateful to Paulo Oliveira for kindly providing the PCC 6803 Δslr1270 mutant strain.
This work was supported by a JSPS Grant-in Aid for Young Scientists (B) (grant 15K20860) to S.K. and JSPS Grant-in-Aid for JSPS Research Fellow (grant 17J03811) to H.K.
FOOTNOTES
- Received 6 June 2017.
- Accepted 30 June 2017.
- Accepted manuscript posted online 10 July 2017.
- Copyright © 2017 American Society for Microbiology.
