INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The use of biotransformations and environmental bioremediation of hydrocarbons is expanding (13, 43). However, such processes are often hindered by the toxic effects of organic solvents on whole cells. A major target for organic solvents that are toxic to microorganisms is the cell membrane (35). These compounds affect the membrane structure and membrane fluidity, resulting in inadequate membrane function, aspecific permeabilization and, ultimately, loss of physiological functions and cell death (8, 9, 36). Nevertheless, there are variations in the solvent tolerance of microorganisms, and in recent years, several bacterial strains which are able to tolerate high concentrations of organic solvents have been isolated (7, 19, 33).

Several possible mechanisms of solvent tolerance have been proposed in the past few years, among them, an energy-dependent active efflux system for solvent such as toluene in Pseudomonas species (22, 34), as well as adaptive alterations of membrane fatty acid and phospholipid head group composition (21, 34, 39). One such key reaction is the recently observed isomerization of cis to trans membrane unsaturated fatty acids in Pseudomonas strains and Vibrio species (4, 15, 29, 34, 40), which occurs without a shift in double-bond position and is independent of de novo fatty acids or lipid synthesis (3, 11). It is now clear that this response is induced by environmental stress factors. Elevated temperatures and increased salt concentration as well as growth phase changes can affect the cis/trans ratio of membrane lipids (16, 21), and organic solvents, in particular, have been found to drastically increase the production of trans-unsaturated lipids in Pseudomonas species (4, 11, 15, 34, 40).

Since an increase of trans-unsaturated fatty acids decreases membrane fluidity, it has been postulated that bacteria use the cis-trans isomerization of unsaturated lipids to counter the increase in fluidity caused by membrane-perturbing conditions, thereby protecting the cells against toxic compounds such as solvents (3, 10, 21). This idea is supported by recent findings of Holtwick et al. (17) and Ramos et al. (34), who found that mutants of Pseudomonas putida strains unable to carry out the cis-trans isomerization of unsaturated lipids are sensitive to increased temperatures and hypersensitive to solvents, such as toluene, while wild-type cells survive such conditions.

Recently the structural gene for the unsaturated membrane fatty acid cis-trans isomerase, cti, was cloned, analyzed, and transferred to Escherichia coli, which normally does not form trans-unsaturated fatty acids (17). The cti recombinant was able to carry out a solvent-induced cis-trans isomerization of membrane unsaturated fatty acids, analogous to the isomerization seen in wild-type pseudomonads (17).

However, not much is known about this isomerase, and it is still not understood how cells sense the presence of organic solvents to change the ratio of trans- to cis-unsaturated fatty acids. Since CTI is expressed constitutively (11), it is unlikely that the rapid isomerization of membrane lipids is caused by induction of the isomerase. Possible alternatives are that the isomerization of fatty acids is induced by activation of CTI or that stress factors such as organic solvents enable the isomerase to access the double bond of membrane phospholipid unsaturated fatty acyl chains.

To gain more information about the biochemistry and function of the cis-trans isomerase, we have purified the enzyme and determined its cellular localization. We have found it to be located in the periplasmic space. Analysis of the substrate specificity in vitro showed that the isomerase is fully active and does not require activation by the addition of organic solvents. Interestingly, it acts only on free fatty acids. Pure phospholipids, even in the presence of organic solvents, failed to serve as substrate for the isomerase. However, when isolated membranes from Pseudomonas oleovorans cells or E. coli cells were used as phospholipid sources and exposed to organic solvents, cis-trans isomerization was detectable. Clearly, additional membrane components are essential for controlling and enabling the cis-trans conversion of membranes by the periplasmic isomerase in response to organic solvents.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and culture conditions. P. oleovorans GPo1 is a strain of the P. putida type (1) which is able to use alkanes as a carbon source (23). P. oleovorans GPo12 is an OCT plasmid-cured variant of wild-type GPo1. The cells were cultivated at 25°C in a 1-liter Erlenmeyer flask containing 250 ml of E2 medium (25), with 1% (wt/vol) glucose as a carbon source. E. coli AM1095 (14) was cultured at 25°C in 1-liter Erlenmeyer flasks containing 250 ml of Luria-Bertani broth. Cell dry weight was determined as described previously (41).

Formation of spheroplasts. P. oleovorans and E. coli cells were converted to spheroplasts by the procedure of Witholt et al. (42), but no mild osmotic shock was used. About 200 mg of cells (dry weight, exponential phase) were suspended in 20 ml of buffer A (100 mM Tris-HCl, 20% [wt/vol] sucrose [pH 8]) and incubated for 5 min in the hypertonic sucrose-Tris solution. Subsequently, 20 ml of buffer B (100 mM Tris-HCl, 20% [wt/vol] sucrose, 5 mM EDTA [pH 8]) and 2 mg of lysozyme were added. The resulting cell suspension containing 100 mM Tris-HCl, 20% (wt/vol) sucrose, 2.5 mM EDTA (pH 8), and 50 µg of lysozyme/ml was incubated at room temperature (RT) for 25 min.

Preparation of the periplasmic, cytoplasmic, and membrane fractions. The generated spheroplasts were centrifuged at 10,000 × g for 20 min, and the supernatant was used as the periplasmic fraction. Pelleted spheroplasts were resuspended in 15 ml of buffer C (100 mM Tris-HCl, 20 mM MgCl₂ [pH 8]) to which DNase (0.01 mg/ml) was added. The spheroplasts were disrupted by two passages through a French press, and cell debris was removed by centrifugation at 5,000 × g for 15 min. Total membranes were collected by ultracentrifugation at 250,000 × g for 120 min, and the resulting supernatant was used as the cytoplasmic fraction. The crude membrane pellet was resuspended in 5 ml of buffer D (50 mM KP_i [pH 7.2]) and stored at 20°C. The periplasmic and the cytoplasmic fractions were dialyzed against 5 liters of buffer D for 15 h, brought to 80 ml with buffer D, and stored at 20°C.

Purification of the cis-trans isomerase. P. oleovorans cells (12 g of dry cells, early stationary phase) were resuspended in 250 ml of buffer A and subsequently mixed with 250 ml of buffer B. Lysozyme was added to a final concentration of 50 µg/ml, and after incubation for 25 min at RT the spheroplasts were collected by centrifugation (10,000 × g, 20 min). The supernatant or periplasmic fraction (450 ml) was mixed with 550 ml of saturated ammonium sulfate (55% saturation), and after 1 h of incubation the suspension was centrifuged at 10,000 × g for 20 min. The resulting precipitate was dissolved in 30 ml of buffer D and dialyzed against 5 liters of buffer D for 15 h. The dialysate was ultracentrifuged (250,000 × g, 2 h), and the supernatant (brought to 100 ml with buffer D) was incubated with 30 ml of anion exchanger material (Fractogel EMD DEAE-650 (S); Merck, Darmstadt, Germany) equilibrated in buffer D. After 5 min of incubation, the suspension was separated by centrifugation at 10,000 × g for 20 min, and the supernatant was mixed with 30 ml of cation exchanger material (Fractogel EMD SO₃-650 (S); Merck) equilibrated in buffer D. After further incubation for 5 min, the mixture was centrifuged (10,000 × g, 20 min), and the supernatant was applied to a hydroxylapatite column (1.6 × 15 cm) equilibrated with buffer D. The proteins were eluted with an increasing gradient of 50 to 500 mM KP_i (pH 7.2) and assayed for isomerase activity. The fractions containing activity were supplemented with (NH₄)₂SO₄ to a final concentration of 1 M and applied at 25°C onto a Fractogel EMD Phenyl I 650 (S) (Merck) column (1 by 8 cm) equilibrated with 1 M (NH₄)₂SO₄ in 50 mM KP_i (pH 7.2). The proteins were eluted at 25°C with a linear gradient from 1 to 0 M (NH₄)₂SO₄, and the fractions containing isomerase activity were combined and concentrated with a Biomax-50 filter (Millipore). The concentrated protein solution (200 µl) was passed through a Superdex 200 gel filtration column (1.6 × 60 cm) equilibrated with buffer D at a flow rate of 1 ml/min. The pooled fractions containing activity were concentrated with a Biomax-50 filter (Millipore) and stored at 20°C until further use. All operations were carried out at 4°C unless stated otherwise.

Preparation of phospholipids. Whole-cell phospholipids were prepared from exponentially growing P. oleovorans or E. coli cells cultured at 25°C in the media described above. After centrifugation of the cells, the lipids were extracted and purified as described by Kates (20). The resulting chloroform phase was evaporated under vacuum to a minimal volume (0.5 to 1 ml) and mixed with 10 volumes of acetone. The resulting phospholipid precipitate (freed of neutral lipids) was dissolved in chloroform and stored at 4°C. Fatty acids of the purified phospholipids from P. oleovorans and E. coli cells contained about 50% cis-unsaturated fatty acids (palmitoleic acid and vaccenic acid). The pure phospholipid substrates used in the assay were present as sonicated dispersions or multilamellar liposomes made as described by Taylor and Cronan (37).

Enzyme and protein assays. The isomerase activity was routinely measured by using an assay mixture which contained 50 mM KP_i (pH 7.2), 1 mM palmitoleic acid, and enzyme solution in a total volume of 1 ml. After incubation at 30°C for 60 min the reaction was stopped by addition of 2 ml of 6 N HCl in methanol. Subsequently, the mixture was heated for 30 min at 90°C, and the methyl esters were extracted with 1 ml of hexane and analyzed by capillary gas chromatography as described by Chen et al. (5). When crude membranes or pure phospholipids were used as substrate for the isomerase, the assay solution was first subjected to saponification and methylation as described by Marvin et al. (26). Subsequently the methyl esters were extracted and analyzed as described above. pH dependence was investigated with the following buffers: 50 mM BisTris-HCl (pH 5 to 6.5), 50 mM K₂HPO₄/KH₂PO₄ (pH 6.5 to 8), and 50 mM Tris-HCl (pH 7 to 9). Glucose-6-phosphate (G-6-P) dehydrogenase was assayed by established procedures (9). One unit of activity was defined as the amount of enzyme that catalyzes the conversion of 1 µmol of substrate to product per minute. Protein concentrations were determined as described by Bradford (2), with bovine serum albumin as the standard.

Electrophoresis and molecular mass determination. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on slab gels was performed by using a separate gel of 12% acrylamide and a stacking gel of 4% acrylamide, according to the method of Laemmli (24). Gels were stained by using Coomassie brilliant blue. The isolectric point was determined by isoelectric focusing by using the Pharmacia Phast Gel system (PhastGel IEF 3-9). The pI was estimated from the position of the protein relative to the position of the bands of the marker proteins (broad pI calibration kit, 3.5 to 10; Pharmacia). The molecular mass of the purified isomerase was determined under native conditions by gel filtration on a Superose 12 gel filtration column (1 by 30 cm) equilibrated with buffer D. The column was calibrated by using the following references (Sigma): B-amylase (200 kDa), alcohol dehydrogenase (150 kDa), albumin (66 kDa), carbonic anhydrase (29 kDa), and cytrochome c (12.4 kDa).

Amino acid sequence analysis. Amino acid sequencing was performed by automated Edman degradation. For NH₂-terminal sequencing, 20 µg of purified isomerase was subjected to SDS-PAGE and electroblotted to a polyvinylidene fluoride membrane as described by Matsudaira (27). Since the blotted isomerase band was visible as a yellow spot without staining, it was directly subjected to on-membrane amino acid analysis.

Chemicals. Fatty acids, D-glucose 6-phosphate, and NADP were purchased from Sigma.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Localization of the cis-trans isomerase. Recently, Okuyama et al. (28) localized the main activity of a cis-trans-unsaturated fatty acid isomerase (CTI) in the membrane-free fraction of Pseudomonas sp. strain E-3, suggesting that CTI resides either in the periplasm or in the cytoplasm. To localize CTI in more detail, we fractionated P. oleovorans cells and isolated the periplasmic, cytoplasmic, and membrane fractions.

Purification of the cis-trans isomerase. The periplasmic fraction prepared from P. oleovorans cells was used as the starting material for the purification of the cis-trans isomerase (Table 2). At pH 7, the enzyme failed to bind to either anion or cation exchangers. We used this property to remove contaminating proteins which did adsorb to these materials. Thus, an ammonium sulfate precipitate was dialyzed and exposed to anion and cation exchangers. The supernatant was subsequently chromatographed on hydrophobic, hydroxylapatite, and gel filtration columns.

View larger version (33K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE analysis of the cis-trans isomerase from P. oleovorans GPo12. Lane A, purified cis-trans isomerase; lane B, molecular size standards (from top to bottom): phosphorylase b (97.4 kDa), serum albumin (66.2 kDa), ovalbumin (45 kDa), carbonic anhydrase (31 kDa), trypsin inhibitor (21.5 kDa), and lysozyme (14.4 kDa).

Molecular characteristics of the cis-trans isomerase. The relative molecular mass of the pure enzyme under native conditions was determined to be 70 kDa by gel filtration, suggesting a monomeric structure of the enzyme. The isoelectric point was determined to be 7.1 by isoelectric focusing.

NH₂-terminal amino acid sequence. The purified isomerase was electroblotted onto a polyvinylidene fluoride membrane, and the first 21 NH₂-terminal residues were determined. The sequence for this isomerase from P. oleovorans GPo12 is identical to that deduced for the cloned cis-trans isomerase gene from P. putida P8 between residues 21 and 41 (17) (Fig. 2), indicating that the nucleotide sequence encodes a precursor protein which is cleaved during export to the periplasmic space of Pseudomonas strains.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   (A) Deduced NH2-terminal amino acid sequence for the isomerase gene product of P. putida P8. (B) NH2-terminal amino acid sequence of the purified isomerase from P. oleovorans GPo12. Identical residues are shaded.

Optimal pH and temperature conditions for enzyme activity. The activity of the purified cis-trans isomerase was tested over a range of 20 to 55°C and was found to be highest at about 30°C, the optimum growth temperature of P. oleovorans. The isomerase retained more than 70% of its activity after exposure for 60 min to 40°C. However, no activity was detectable after incubation of the isomerase for 10 min at 55°C. The pH activity profile of the enzyme resembled a bell-shaped curve, with a pH optimum at about 7.

Catalytic properties. The substrate range of the cis-trans isomerase was determined by incubating the purified enzyme with various fatty acids and determining the rate of product formation (Table 3). Highest activities were found for 9-cis-hexadecenoic acid (palmitoleic acid), 11-cis-octadecenoic acid (vaccenic acid), and 13-cis-eicosenoic acid. The isomerase evidently prefers unsaturated fatty acids having seven carbon atoms after the double-bond position, indicating that the substrate specificity of the enzyme is based on the double-bond position rather than on the length of the acyl chain.

Isomerization of membrane unsaturated fatty acids by CTI in vitro. When isolated crude membranes from Pseudomonas or E. coli cells were used as a phospholipid source, the purified isomerase increased the amount of trans-unsaturated fatty acids. However, a significant cis-trans conversion of lipids occurred only in the presence of organic solvents (Table 4). Note that octanol, which is known to cause a strong increase in the trans/cis ratio in vivo (4, 34), is also a very strong activator of the isomerization of membrane unsaturated fatty acids in vitro.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The cis-trans isomerase from P. oleovorans GPo12, which catalyzes the cis-trans conversion of unsaturated membrane fatty acids, is an 80-kDa monomeric periplasmic protein. Although 80 kDa is unusually large for periplasmic proteins, the cell fractionation data are unequivocal. Moreover, the determined N-terminal amino acid residues of the purified isomerase from P. oleovorans GPo12 corresponded to the deduced sequence spanning residues 21 to 41 of the cis-trans isomerase gene from P. putida P8, indicating that a 20-amino-acid signal peptide was cleaved from a pre-CTI precursor in vivo, as expected for a periplasmic protein (32).

Analysis of the isomerase showed that there was no cis-trans isomerization of esterified fatty acids such as pure phospholipids or simple methyl esters in vitro, even in the presence of organic solvents. The purified isomerase did, however, act on free unsaturated fatty acids, and in contrast to the in vivo cis-trans conversion, this in vitro isomerization of free fatty acids did not require the addition of organic solvents. Determination of the substrate range indicates that CTI prefers free unsaturated fatty acids having seven carbon atoms after the double-bond position. These two specific determinants suggest that the negatively charged carboxy group is a key group in binding to a presumably positively charged site on the isomerase, while the distal C7 alkyl chain presumably binds in a hydrophobic pocket, with the isomerization taking place at the proximal end of this C7 alkyl segment. The alkyl segment between the carboxy group and the double bond can vary from at least 8 to 12 methylene groups, suggesting that these may be bound rather unspecifically to a hydrophobic pocket.

Okuyama et al. (31) have isolated a cis-trans-unsaturated fatty acid isomerase from the membrane-free fraction of Pseudomonas sp. strain E-3. This enzyme resembles the CTI of P. oleovorans GPo12, with an estimated molecular mass of about 80 kDa. Also in agreement with our data, it was found that purified CTI from Pseudomonas sp. strain E-3 acts only on free fatty acids. However, the CTI from Pseudomonas sp. strain E-3 has a substrate specificity which differs from that of the CTI of P. oleovorans GPo12. Okuyama et al. (31) found no activity for unsaturated fatty acids with a chain length of more than 17 carbons (e.g., vaccenic acid), in accord with the recent finding that although E-3 lipids contain palmitoleic acid and vaccenic acid, only palmitelaidic acid was detected as a membrane trans-unsaturated fatty acid (30). Evidently, the CTI from Pseudomonas sp. strain E-3 cannot isomerize vaccenic acid.

Our finding that the CTI from P. oleovorans, a P. putida strain (1), does isomerize vaccenic acid is in good agreement with several earlier studies. Thus, the in vivo and in vitro cis-trans conversion of membrane unsaturated fatty acids was investigated in several P. putida strains, and it was shown that these cells do have an activity catalyzing the isomerization of palmitoleic and vaccenic acid (3, 11, 15, 34, 40). Moreover, recently Holtwick et al. (17) cloned the structural gene for the cis-trans isomerase from P. putida P8 and found that recombinant E. coli cells carrying cti are able to convert vaccenic acid to the corresponding trans isomer. Clearly, P. putida strains do have a CTI catalyzing the isomerization of unsaturated long-chain fatty acids such as vaccenic acid.

The fact that the isomerase catalyzes the conversion of unsaturated fatty acids in phosphate buffer indicates that no external cofactor or energy is needed during isomerization. This result is in accord with the previous work of Diefenbach and Keweloh (11), which showed that the cis-trans conversion is detectable even when energy-dependent mechanisms cannot function. The findings that metal ions did not influence the cis-trans conversion of free fatty acids and that the isomerase was fully active in the presence of EDTA suggest that CTI is metal independent or that the enzyme may contain very tightly bound metal ions which cannot be removed by chelating agents.

We and others have previously found that the in vivo cis-trans conversion of membrane lipid unsaturated fatty acids in Pseudomonas depends on activation with solvents (4, 10, 11, 15, 34, 40). Similarly, transformation of E. coli, which normally does not form trans-unsaturated fatty acids, with the cti gene conferred in vivo cis-trans isomerization activity to the host in response to organic solvents (17).

We initially concluded from these early observations that the isomerase is the only enzyme required for the solvent-induced cis-trans conversion of membrane lipids. However, since in membranes cis-unsaturated fatty acids are esterified to phospholipids in bacteria, which are not substrates of CTI in vitro, it is now clear that the in vivo activity of the isomerase must depend on one or more solvent-activated components not present in in vitro experiments with pure phospholipids. When crude membranes from Pseudomonas or E. coli cells were used as a phospholipid source in the in vitro cis-trans assay, there was significant isomerization, but only in the presence of organic solvents. Thus, the solvent activation which was not seen in the in vitro experiments with pure phospholipids did occur in isolated membranes from Pseudomonas or E. coli.

Based on the result that CTI acts only on free fatty acids, it can be postulated that possible candidates for solvent-activated membrane component could be membrane phospholipases which, known to be strongly triggered by organic solvents (12, 38), hydrolyze phospholipids to free fatty acids that can subsequently serve as substrates for the periplasmic isomerase. In general, membrane-bound phospholipases of bacteria do have Ca²⁺ as an essential cofactor (18, 38). This would explain why, although the isomerase is not affected by calcium ions and EDTA, the solvent-induced in vitro isomerization of Pseudomonas and E. coli membrane lipids was strongly inhibited in the presence of EDTA and why it could be reconstituted by adding calcium ions to the cis-trans isomerization assay.

In previous studies it has been clearly shown that isomerized membrane unsaturated fatty acids of Pseudomonas species do contain trans-unsaturated fatty acids mainly esterified to phospholipids (11, 21, 28). Therefore, if membrane lipolytic enzymes do have an essential role in the isomerization of unsaturated lipids, free fatty acids should be reincorporated in lyso-phospholipids. Another possibility to produce trans-unsaturated phospholipids is that specific interactions of a membrane component(s) with membrane lipids and/or CTI allow the direct isomerization of esterified fatty acids without producing free fatty acid (28). However, thus far there is no indication as to how the periplasmic isomerase might bind to or react with the membrane factor(s) to produce trans-unsaturated lipids.

Future work should focus on the isolation and characterization of the membrane component(s) needed to enable the cis-trans conversion of membrane unsaturated fatty acids by the periplasmic isomerase. Since the isomerization of lipids can also take place in E. coli membranes, this microorganism and mutants lacking membrane components (e.g., lipolytic enzymes) might offer us the tools to study this unique reaction in more detail, in vitro as well as in vivo.