Channel-Forming (Porin) Activity in Herpetosiphon aurantiacus Hp a2

Abstract

Detergent extracts of cell envelopes of the gliding bacterium Herpetosiphon aurantiacus formed channels in lipid bilayers. Fast protein liquid chromatography across a HiTrap-Q cation-exchange column demonstrated that a 45-kDa protein forms the channel. The observation of a channel-forming protein suggests that Herpetosiphon aurantiacus Hp a2 has a permeability barrier on its surface.

Herpetosiphon aurantiacus is a nonphototrophic, strictly aerobic, gliding bacterium. It is composed of unbranched filaments consisting of cylindrical single cells. Together with Chloroflexus and Thermomicrobium it forms the 8th phylum of bacteria (the green nonsulfur bacteria) (19, 20, 27). A characteristic of H. aurantiacus is the deep orange color caused by a pigment of the highly folded cytoplasmic membrane (13, 22). Little is known about the structure and composition of the cell wall of H. aurantiacus. An early study, however, demonstrated the absence of an outer membrane (21). Biochemical attempts to study the composition of a prospective outer membrane did not reveal any indication of the presence of such a layer. Interestingly, the cell wall lacked lipopolysaccharides (LPS) and contained l-ornithine instead of meso-diaminopimelic acid, both of which indicate that H. aurantiacus possesses a gram-positive-like cell wall (10).

So far it has been an open question if H. aurantiacus contains a permeability barrier on its surface similar to the outer membrane of gram-negative bacteria. However, there exists emerging evidence that certain gram-positive bacteria also contain permeability barriers on the surface of the peptidoglycan layer. Prominent examples represent members of the mycolata (7, 9). The mycolic acid layer of Mycobacterium chelonae (24), Mycobacterium smegmatis (25), and Corynebacterium glutamicum (15) contains hydrophilic channels. Similarly, cell wall channels exist in Streptomyces griseus and Micromonospora purpurea (11, 12). Channel-forming activity in the cell wall of the 8th phylum of bacteria has not been observed so far. In this study, we describe the isolation, purification, and biophysical characterization of an ion-permeable channel from H. aurantiacus Hp a2.

H. aurantiacus Hp a2 (DMSZ 589, Deutsche Sammlung von Mikroorganismen und Zellkulturen, Braunschweig, Germany) was cultivated in a modified Hp 74 medium (1% Na-glutamate, 0.2% MgSO₄ · 7H₂O, 0.2% yeast extract, 1.19% HEPES, pH 7.2) according to reference 21. The cells were harvested by centrifugation and washed once in a Tris buffer (20 mM Tris-HCl, 10 mM MgCl₂, pH 8.0). The cells were suspended in the same buffer and broken with a vibratory cell breakage mill (type Vi 2; Bühler, Tübingen, Germany). The cell homogenate was centrifuged at 200,000 × g at 4°C for 60 min (Beckman Ti60; Beckman), and the pellet was washed and treated with 0.4% LDAO (NN-dimethyldodecylamine N-oxide) extraction buffer (buffer A; 0.4% LDAO, 20 mM Tris-HCl, 10 mM MgCl₂, pH 8.0) at 30°C for 30 min. After centrifugation (200,000 × g, 4°C for 60 min), the supernatant was divided into 1.5-ml aliquots and frozen at −20°C. The LDAO extract was inspected for channel-forming activity with the lipid bilayer assay. Membranes were formed as described earlier (3) across a hole in a Teflon cell from a 1% solution of diphytanoyl phosphatidylcholine (PC; Avanti Polar Lipids, Alabaster, Ala.) dissolved in n-decane. The electrical measurements were performed with Ag/AgCl electrodes (with salt bridges) connected in series to a voltage source and a current amplifier. Zero current membrane potentials were measured by establishing a fivefold KCl gradient across membranes containing 100 to 1,000 channels (4).

The 0.4% LDAO supernatant was reduced to [1/4] of its volume with a Speedvac (Maxi Dry Lyo; Heto-Holton AIS, Allerod, Denmark) and applied to a 1-ml HiTrap-Q column (Amersham Pharmacia Biotech, Freiburg, Germany). The column was washed first with buffer A and eluted with buffer A supplemented with NaCl. Linear NaCl gradients between 0 and 1 M NaCl were applied to the column, and 80 fractions of 1 ml were collected. The protein concentration in the fractions was detected at 280 nm. To increase the protein concentration, protein samples were precipitated by 65% trichloroacetic acid or concentrated by two-phase separation (26). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to Laemmli (14). The gels were stained with Coomassie brilliant blue or silver (6).

The supernatant of the LDAO extraction showed channel-forming activity in the lipid bilayer assay. Total channel-forming activity was released from the cell envelope when it was treated with 0.4% LDAO. Purification of channel-forming activity was achieved by fast protein liquid chromatography (FPLC) on a HiTrap-Q column. Figure 1, lane 3, shows the protein composition of the LDAO extract that was applied to the column. The column was first washed with buffer and then eluted with buffer supplemented with increasing concentrations of NaCl. First a linear gradient from 0 to 0.09 M NaCl solution was applied. Then the increase in NaCl concentration was stopped, and the NaCl concentration was kept constant at 0.09 M during the elution of 5 ml. The eluted protein peak (fractions 29 to 34) contained a protein with a molecular mass of 45 kDa, which showed a very high channel-forming activity (Fig. 1, lane 4).

FIG. 1.

SDS-PAGE (7% polyacrylamide), performed according to Laemmli (14), as part of the purification procedure for the 45-kDa channel-forming protein of H. aurantiacus Hp a2. The gel was stained with Coomassie brilliant blue, and all protein samples were treated for 10 min at 100°C. Lane 1, low-molecular-mass markers of 36, 45, and 66 kDa; lane 2, high-molecular-mass markers of 36, 45, 55, 66, 84, 97, 116, and 205 kDa; lane 3, 15 μl of the 0.4% LDAO extract of the cell envelope fraction dissolved in 10 μl of sample buffer; lane 4, 5 μg of protein of fraction 32 of the Hitrap-Q FPLC column dissolved in 15 μl of sample buffer.

Figure 2 shows a single-channel recording of a PC membrane in the presence of the 45-kDa protein, which was added in a concentration of 10 ng/ml to a membrane. The single-channel recording demonstrates that the protein formed defined channels with an average single-channel conductance of about 800 pS in 1 M KCl. Only a minor fraction of channels with other conductance was observed (data not shown). The channels formed by the 45-kDa protein of H. aurantiacus Hp a2 had a long lifetime (mean lifetime of at least 5 min) similar to that detected previously for porins of gram-negative (2) and gram-positive (25) bacteria at small transmembrane potential. No voltage-dependence closure was observed in KCl solution up to voltages of 100 mV. The only exception of long channel lifetime was found in experiments at low pH, where transient channels were observed. A typical example for channel closure represents the recording of Fig. 3 (NH₄Cl at pH 5; applied voltage, 10 mV). Under these conditions, the channels closed in one step, indicating that the channel of H. aurantiacus Hp a2 could be formed by a protein monomer (Fig. 3). Voltage-dependent closure of OmpF of E. coli occurs in three distinct steps corresponding to the closing of the three individual channels in a trimer (23). OmpG of Escherichia coli is obviously active as a monomer, in contrast to most porins of enteric bacteria, as has been demonstrated by SDS-PAGE and electron microscopic analysis of two-dimensional crystals (1). Voltage-dependent closure or gadolinium-induced block of OmpG channels occurs in single steps similar to the channel of H. aurantiacus Hp a2, which suggests both could be monomeric channels (8).

FIG. 2.

Single-channel recording of a PC/n-decane membrane in the presence of pure 45-kDa protein of H. aurantiacus Hp a2. The aqueous phase contained 1 M KCl (pH 6) and 10-ng/ml protein. The applied membrane potential was 20 mV (temperature of 20°C).

FIG. 3.

Single-channel recording of a PC/n-decane membrane in the presence of pure 45-kDa protein of H. aurantiacus Hp a2. The aqueous phase contained 1 M NH₄Cl (pH 5) and 10-ng/ml protein. The applied membrane potential was 10 mV (temperature of 20°C). Note that the channels were transient under these conditions and open and closed in a single step, indicating that the channel is formed by a protein monomer.

We performed single-channel experiments with salts other than KCl to obtain information on the size of the channels formed by the 45-kDa protein of H. aurantiacus Hp a2 and its ion selectivity. The results are summarized in Table 1. The conductance sequence of the different salts within the channel was NH₄Cl ≈ KCl > CsCl > NaCl > LiCl > K acetate > Tris-Cl, which means that the single-channel conductance follows the bulk aqueous conductivity of the different salts. The influence of the anion and cations of different mobility on the conductance in 1 and 0.1 M salt solutions was moderate (Table 1), suggesting only a low selectivity of the channel. Table 1 shows also the average single-channel conductance, G, as a function of the KCl concentration in the aqueous phase. The relationship between conductance and KCl concentration was linear, which is characteristic for wide and water-filled channels without point net charges and/or binding sites for ions (2). Its size is presumably similar to that of OmpF or OmpC of E. coli because of the comparable single-channel conductance of about 800 pS in 1 M KCl compared to 1.5 to 2.0 nS for the porins from enteric bacteria (5). This is not a contradiction because the conductance of enteric porins is that of a trimer, whereas the channel of H. aurantiacus Hp a2 is presumably formed by a monomer.

TABLE 1.

Average single-channel conductivity of the 45-kDa protein from H. aurantiacus Hp a2 in different salt solutions of concentration c^a

Salt	Concn (M)	Single-channel conductance (G [pS])
LiCl	0.1	52
	1.0	500
NaCl	0.1	63
	1.0	630
KCl	0.03	37
	0.1	78
	0.3	210
	1.0	800
	3.0	2,700
NH4Cl	0.1	66
	1.0	840
CsCl	0.1	73
	1.0	730
Tris-HCl	0.1	26
	1.0	260
KCH3COO (pH 7)	0.1	78
	1.0	550

^aThe membranes were formed of PC dissolved in n-decane. The aqueous solutions were unbuffered and had a pH of 6 unless otherwise indicated. The applied voltage was 20 mV, and the temperature was 20°C. The average single-channel conductance (G) was calculated from at least 80 single events.

The ion selectivity of the channel formed by the 45-kDa protein of H. aurantiacus Hp a2 was studied by zero-current membrane potential measurements in presence of KCl gradients. A fivefold KCl gradient (100 versus 500 mM) across a membrane containing about 100 to 1,000 channels resulted in an asymmetry potential of about 12 mV on the more dilute side (mean of five measurements). This result indicated little preferential movement of potassium ions over chloride through the channel at neutral pH. The zero-current membrane potentials were analyzed by using the Goldman-Hodgkin-Katz equation (3, 5). The ratio of the potassium permeability, P_K, divided by the chloride permeability, P_Cl, was about 2, which indicated a low selectivity of the channel formed by the 45-kDa protein of H. aurantiacus Hp a2.

For localization of the channel-forming activity, the cell envelope fraction (2 ml) was applied to a sucrose step gradient of 20, 40, 50, and 60% sucrose (wt/vol) and centrifuged at 120,000 × g at 4°C for 16 h. Fractions were washed twice with Tris-HCl buffer (20 mM Tris-HCl, 10 mM MgCl₂, pH 8). They were treated with LDAO extraction buffer (0.4% LDAO, 10 mM Tris-HCl, pH 8) and checked for channel-forming activity. The highest activity was found in the fraction between 40 and 50% sucrose (fraction I, light rose color), about 50% less activity was found in the 50% sucrose fraction (fraction II, red color), and very low activity was found in 60% sucrose (fraction III, dark red color comprising the cytoplasmic membrane with the dark red pigment) (22). All fractions revealed a 45-kDa protein band on SDS-PAGE. The intensity of this band decreased corresponding to the decreasing channel-forming activity of the sucrose gradient fractions I to III, suggesting that the channel-forming activity is localized in the light fraction of the cell envelope. This result is in some contrast to sucrose-density centrifugations of the cell envelope of enteric gram-negative bacteria and also of the mycolata, where the outer membrane together with the porin activity is found in the heavy fractions (16, 18, 24). Nevertheless, it is clear that the cytoplasmic membrane of H. aurantiacus Hp a2 has a higher density than other constituents of the cell wall, because the dark red color of fraction III is caused by the red pigment that is almost exclusively found in the cytoplasmic membrane (22). This means presumably that the putative outer membrane of H. aurantiacus has a low density, presumably because this bacterium and Chloroflexus aurantiacus do not contain LPS (10, 17), although the unfinished genome of C. aurantica contains genes for LPS biosynthesis (available at http://www.ncbi.nlm.nih.gov/genomes, NZ_AAAH00000000). It is noteworthy that a similar result has been obtained in a recent study with S. griseus, where the cell wall channel was found in a fraction of the sucrose-step density centrifugation that had a lower density than the cytoplasmic membrane (11). The high channel-forming activity of the pure 45-kDa protein rules out the possibility that we are dealing here with an unknown contaminant protein responsible for channel formation. Clearly, these channels can only be present in the cell wall of H. aurantiacus Hp a2 and not in its cytoplasmic membrane. This means that we demonstrated the presence of a channel in the cell envelope of H. aurantiacus Hp a2, which belongs to the 8th phylum of eubacteria, the green nonsulfur bacteria (19, 20). However, the exact localization of the 45-kDa protein in the cell envelope of H. aurantiacus Hp a2 and the structure of its cell wall are largely unknown and need further elucidation.