Carbohydrate Research Center, University of Georgia Complex, Athens, Georgia 30602,1 Department of Microbiology and Biotechnology, University of Tübingen, D072076 Tübingen, Germany,2 Department of Plant and Soil Sciences, University of Delaware, Newark, Delaware 197173
Received 23 September 2002/ Accepted 16 December 2002
| ABSTRACT |
|---|
|
|
|---|
-1 hydroxy fatty acid, 27-hydroxyoctacosanoic acid (27OHC28:0). The 27OHC28:0 moiety is common in lipid A's among members of the Rhizobiaceae and also among a number of the facultative intracellular pathogens that form chronic infections, e.g., Brucella abortus, Bartonella henselae, and Legionella pneumophila. In this paper, a mutant of R. leguminosarum was created by placing a kanamycin resistance cassette within acpXL, the gene which encodes the acyl carrier protein for 27OHC28:0. The result was an LPS containing a tetraacylated lipid A lacking 27OHC28:0. A small amount of the mutant lipid A may contain an added palmitic acid residue. The mutant is sensitive to changes in osmolarity and an increase in acidity, growth conditions that likely occur in the nodule microenvironment. In spite of the probably hostile microenvironment of the nodule, the acpXL mutant is still able to form nitrogen-fixing root nodules even though the appearance and development of nodules are delayed. Therefore, it is possible that the acpXL mutant has a host-inducible mechanism which enables it to adapt to these physiological changes. | INTRODUCTION |
|---|
|
|
|---|
-1 fatty acid, 27-hydroxyoctacosanoic acid (27OHC28:0), which doubles in amount in R. leguminosarum bv. viciae lipid A during pea symbiosis (26). Also during symbiosis, and during growth under conditions that mimic those that occur during symbiosis, both the LPS and the entire bacterial cell become more hydrophobic (26). Thus, the increase in 27OHC28:0, together with increased O-chain methylation, may be responsible for this increase in hydrophobicity of the LPS and bacterial cell. The initial stages of symbiosis include root hair deformation and the stimulation of root cortical cell division, resulting in nodule formation. These events are induced via an exchange of signal molecules; flavonoids produced by the host plant induce expression of nodulation (nod) genes in the Rhizobium strain (for recent reviews, see references 14 and 44). The nod gene products produce an acylated chitin oligosaccharide, the Nod factor, which stimulates these initial events. Rhizobia adhere to the emerging root hairs of the host, and an infection conduit known as the infection thread is formed by invagination of the plasma membrane and deposition of plant extracellular matrix material. Bacteria spread into the nascent nodule tissue though these infection threads. This is followed by intracellular invasion of the root nodule tissue in an endocytosis-like process, resulting in an intracellular compartment known as a symbiosome, in which the invading Rhizobium cells are surrounded by a plant membrane. In this process, the infection thread penetrates the nodule cortical cell and expands, forming an infection droplet that contains a number of symbiont bacteria. Bacteria are released from the droplet into the host cell cytoplasm.
With indeterminate nodules, such as those formed by pea, the bacteria become singly surrounded by the host-derived membrane known as the symbiosome membrane via a synchronous bacterium-symbiosome membrane division process. Electron micrographs suggest that during this process, there are actual points of contact between the bacterial membrane and the symbiosome membrane (8). Inside the symbiosome, Rhizobium differentiates into a bacterial form known as the bacteroid, which produces nitrogenase.
It has been hypothesized that the presence of 27OHC28:0 in lipid A is required for maintaining the stability of the bacterial membrane during the endocytotic invasion process and also for survival of the bacterium within the plant-derived symbiosome compartment. Increased membrane stability could be provided by the 27OHC28:0 residue because the alkyl chain of this fatty acid is long enough to span the entire outer membrane bilayer.
The structure of the lipid A's from R. etli (formerly R. leguminosarum biovar phaseoli) and R. leguminosarum bv. viciae (5, 39, 40) is shown in Fig. 1 together with the structure of Escherichia coli lipid A. The R. leguminosarum bv. viciae and R. etli lipid A's have identical structures with unique features compared to those of E. coli lipid A. These unique structural features include the replacement of the 4' phosphate with a galacturonosyl residue, the oxidation of the proximal glucosaminosyl residue to 2-amino-gluconate, the lack of acyloxyacyl lauryl and myristyl residues, and the presence of a very long chain acyloxyacyl moiety, 27OHC28:0. A similar lipid A occurs in the LPS from Rhizobium sp. strain Sin-1 (which nodulates the roots of Sesbania spp.) but differs slightly from that of R. leguminosarum bv. viciae in that it lacks the galacturonic acid residue and, while containing 27OHC28:0, has an altered fatty acylation pattern (22).
|
The focus of the work presented in this paper is the unusual 27OHC28:0 acyl residue of the R. leguminosarum bv. viciae lipid A because it increases in the lipid A during symbiosis between R. leguminosarum bv. viciae and pea and its presence in the lipid A's of the Rhizobiaceae as well as in the lipid A of several intracellular pathogens suggests a functional importance in symbiosis as well as in pathogenicity. As a first step towards elucidating the function of this remarkable acyl component of the lipid A, a mutant of R. leguminosarum bv. viciae that is unable to incorporate 27OHC28:0 fatty acid into its lipid A was prepared. In order to eliminate the presence of 27OHC28:0 in R. leguminosarum bv. viciae LPS lipid A's, it was decided to target acpXL for mutational inactivation. The acpXL gene was cloned and partially sequenced as orf* by Colonna-Rommano et al. (12), and the AcpXL protein was characterized by Brozek et al. (10). As described by Colonna-Rommano et al. (12), the gene is localized on cosmid pCS115, and this plasmid provided the starting point for the generation of acpXL::kan R. leguminosarum bv. viciae mutants. Here we report the construction and characterization of the acpXL::kan mutant of R. leguminosarum bv. viciae 3841.
| MATERIALS AND METHODS |
|---|
|
|
|---|
|
To inactivate acpXL, a kanamycin resistance (kan) cassette from pUC4K (Amersham BioLabs) was isolated by digestion with EcoRI, and the resulting 1.2-kb kan cassette was inserted into pVV1 digested with EcoRI to generate pVV2. Plasmid pVV2 was transformed into E. coli XL1 Blue, and transformants were selected on LB agar plates containing ampicillin and kanamycin. The plasmid DNA from each transformant was analyzed for orientation of the kan cassette by restriction analysis (EcoRI) of their plasmid DNA.
A 4.8-kb fragment containing the interrupted gene from pVV2 was released by partial digestion with XhoI and cloned into suicide plasmid pJQ200uc1 at a unique XhoI site in the multiple cloning site of this vector. Plasmid pJ200uc1 carries the conditionally lethal gene sacB, allowing discrimination between integration of the vector and double recombination events (38). The resulting plasmid, pVV4, was transformed into E. coli JM109, and transformants were selected on LB agar plates containing kanamycin. The construction of the pVV4 plasmid is schematically depicted in Fig. 2A.
|
A complement of the acpXL::kan mutant was constructed to confirm that any observable physiological differences in the mutant were a result of the acpXL::kan mutation. For this purpose, a 2.5-kb fragment containing the paren-tal acpXL gene was isolated from pVV1 (see above) with the restriction enzyme KpnI. This fragment was cloned into pBBR1MCS3, a broad-host-range vector (38) that had been digested with KpnI. The resulting construct, pVV5, was transformed into E. coli JM109, and transformants were selected on LB plates containing tetracycline. This strain was used as the donor for triparental mating into R. leguminosarum bv. viciae 22 with E. coli pRK600 as the helper strain. Transconjugants were selected on TY plates containing tetracycline and kanamycin. Confirmation of acpXL (plasmid-borne) and acpXL::kan (chromosomal) in the complemented strains was done by PCR with primers to acpXL (described above) with plasmid and chromosomal DNA of the complements, respectively.
LPS extraction and purification.
The wild-type and mutant LPSs were first extracted by the TEA/EDTA/
procedure as previously described (43). For each strain, the bacterial pellet (approximately 10 g wet weight) from 8 liters of culture was extracted with 40 ml of TEA/EDTA/
(0.25 M EDTA containing 5% phenol and titrated to pH 6.9 with triethylamine [TEA]) with constant stirring at 37°C for 1 h. The extract was then centrifuged at 10,000 rpm for 1 h, and the supernatant was collected and dialyzed (2,000 molecular weight cutoff; Spectrapor) against deionized water. This material was further purified by polymyxin affinity chromatography with polymyxin B-Sepharose (Pierce Chemical Company) as previously described (20, 43, 46). Briefly, after sample application in 50 mM NH4HCO3, the column (10-ml bed volume) was sequentially eluted with 30 ml of a solution of 0.3 M TEA adjusted to pH 6.4 with acetic acid plus 10% ethylene glycol, followed by 30 ml of a solution of 2.0 M urea in 0.1 M NH4HCO3 to elute the non-LPS components. The LPS was eluted from the column with 20 ml of 2.5% deoxycholate in 0.1 M NH4HCO3. The LPS fractions were pooled and dialyzed extensively (2,500 molecular weight cutoff) against a solution of 50 mM TEA brought to pH 8.5 with acetic acid and containing 10% ethanol and then against deionized water, after which the fractions were lyophilized.
Analytical procedures. Gel electrophoresis in the presence of deoxycholate was performed as described before (42). Typically, 1 µl of LPS sample (this volume includes dye, loading buffer, and 2 µg of LPS) was loaded onto the gels. The gels were silver stained as described previously (42). Analysis of the lipid A fatty acids was accomplished as previously described (5, 53) by methanolysis of the purified LPS samples in methanolic 1 M HCl, followed by trimethylsilylization of the hydroxyl groups and analysis by combined gas chromatography-mass spectrometry (GC/MS) with a 30-m DB-5 column from J&W Scientific.
Lipid A was isolated from the LPS preparation by mild acid hydrolysis (11). The LPS was dissolved in 1% sodium dodecyl sulfate in 20 mM sodium acetate, adjusted to pH 4.5 with 4 M HCl, and then placed in an ultrasound bath until the sample was dissolved. This solution was then heated at 100°C for 1 h, followed by lyophilization. The sodium dodecyl sulfate was removed by washing the lyophilized residue with a solution of 2:1 deionized H2O-acidified ethanol (100 µl of 4 M HCl in 20 ml of 95% ethanol). The residue was collected by centrifugation, washed again with 95% ethanol (nonacidified), and collected by centrifugation (200 x g for 15 min). The washing and centrifugation steps were repeated. Lastly, the residue was lyophilized to give a white, solid, fluffy lipid A preparation.
Matrix-assisted laser desorption ionization time of flight-mass spectrometry (MALDI-TOF MS) was performed with a Kratos Analytical Kompact SEQ MALDI-TOF spectrometer system (Kratos, Manchester, United Kingdom) in the negative-ion reflectron mode with a 337-nm nitrogen laser operating at a 20-kV extraction voltage and with time-delayed extraction. Approximately 2 µl of a 1-mg/ml lipid A solution in chloroform-methanol (3:1, vol/vol) was mixed with 1 µl of trihydroxyacetophenone matrix solution (
93.5 mg of trihydroxyacetophenone per ml of methanol) and applied to the probe for mass analysis. Spectra were calibrated externally with E. coli lipid A (Sigma).
Study of physiological properties. Growth rates of the wild-type and mutant rhizobia were compared by standard methods at 30°C (spectrophotometer readings of liquid cultures in TY/calcium broth at 600 nm and plating dilution series). Three replicate flasks of bacterial cultures were initiated with starter cultures, and samples were collected at multiple time points between 16 and 96 h for spectrophotometric analysis and serial dilution plating. For both the mutant and the parent strain, 5-ml cultures from single colonies were prepared. In the case of the R. leguminosarum bv. viciae 3841 parent, 50 µl of the 5-ml culture having an optical density at 600 nm of 0.103 (1.96 x 106 cells/ml, as determined by colony counting of serial dilutions) was placed in 500 ml of medium, and the growth curve was monitored by periodic removal of 1-ml samples for measurements of both optical density and cell count. In the case of the R. leguminosarum bv. viciae 22 mutant, 50 µl of the 5-ml culture, which had an optical density at 600 nm of 0.168 (2.09 x 106 cells/ml), was placed in 500 ml of culture, and the growth curve was monitored as for R. leguminosarum bv. viciae 3841.
In order to assess the osmotic tolerance of the mutant, R. leguminosarum bv. viciae 22, its complement, and R. leguminosarum bv. viciae 3841 were grown at 30°C on TY/calcium agar containing various concentrations of NaCl (0.0%, 0.25%, 0.5%, 0.75%, and 1.0%). Sensitivity to pH was determined by growth on TY plates at various pHs (7.0, 6.0, 5.5, 5.0, and 4.8), with PIPES (piperazine-N,N'-bis[2-ethanesulfonic acid]) and piperazine as buffering agents. PIPES was used at a final concentration of 40 mM to buffer the medium at pH 7, and a final concentration of 40 mM piperazine was used in the medium while adjusting the acidity of the medium to the various lower pHs.
Symbiotic phenotype characterization. The effect of the R. leguminosarum bv. viciae acpXL::kan mutation on nodulation of pea, Pisum sativum bv. Early Alaska, was examined by inoculating pea seedlings with 0.5 ml of late-log/early-stationary-phase liquid cultures of wild-type and mutant strains. The seedlings were grown in Erlenmeyer flasks with defined, nitrogen-free plant agar (Fahraeus medium) in growth chambers. The temperature was maintained at 22°C with 14 h of daylight, at 18°C with 10 h of dark, and at 55% humidity. Nodule formation was monitored throughout the growth period.
Nitrogenase activity was measured with the acetylene reduction assay as previously described (51). Three weeks after inoculation, roots with attached nodules were transferred into glass bottles (20 ml) with screw caps and fitted rubber membranes. Then 1 ml of acetylene was added to the roots and incubated for 1 h at room temperature. The amounts of ethylene generated from the nitrogenase were measured through GC analysis (isotherm at 70°C, with a 50 m by 0.53 mm, 10-µm Al2O3-KCl column [Chrompack]).
| RESULTS |
|---|
|
|
|---|
Isolation of R. leguminosarum bv. viciae 3841 acpXL::kan mutants. The acpXL::kan mutants were isolated as described in Materials and Methods (depicted in Fig. 2A). Figure 3 shows the PCR DNA fragments isolated from the parent strain 3841 and the two mutants, R. leguminosarum bv. viciae 22 and 26, with acpXL-specific primers. For comparison, DNA from R. etli CE3 was also included.
|
LPS isolation and characterization.
We employed two standard protocols for LPS extraction, hot phenol-water and TEA/EDTA/
. With both methods, the yield of LPS from the mutant was less than that obtained from the R. leguminosarum bv. viciae 3841 parent; however, the LPS yields from both parent and mutants were increased with TEA/EDTA/
extraction. In the case of the phenol-water extraction, the LPS from the R. leguminosarum bv. viciae 3841 parent extracted into both the phenol and water phases, as described previously (26). However, with the acpXL::kan mutants, the LPS largely extracted into the water phase, with only small amounts found in the phenol phase. This result suggested that the content of hydrophobic LPS in the mutant was much less than that found in the R. leguminosarum bv. viciae 3841 parent. For chemical analysis, the TEA/EDTA/
extraction method was used because this procedure resulted in increased yields of LPS compared to hot phenol-water extraction. The mutant LPS yield per gram (wet weight) of bacterial pellet from the TEA/EDTA/
procedure was approximately one-third of that obtained from the R. leguminosarum bv. viciae 3841 parent.
In order to determine if mutation of acpXL resulted in an LPS that lacked 27OHC28:0, the LPS preparations from the two acpXL::kan mutants, 22 and 26, were analyzed for their fatty acid content and compared with that of the parent strain. The GC/MS profiles for the fatty acids of the lipid A from the parent and mutant 22 are shown in Fig. 4. The parent LPS contained 27OHC28:0 as well as a small amount of 29OHC30:0. The latter fatty acid is often observed in small amounts in this R. leguminosarum bv. viciae lipid A, and its presence is consistent with the report that AcpXL exists in an acylated form in which the acyl group can be either 27OHC28:0 or 29OHC30:0 (1). The presence of the N-acylated GlcN residues (N-3OHC14:0-, N-3OHC16:0-, N-3OHC18:1-, and N-3OHC18:0-GlcN) is consistent with previous reports (5) showing the heterogeneity of the N-acyl substituents and the resistance of the N-acyl residues to the methanolysis procedure (2).
|
|
-galacturonosyl residue and is fatty acylated with ß-hydroxy fatty acyl moieties at the 2, 3, 2', and 3' positions (e.g., 3OHC14:0, 3OHC16:0, 3OHC14:0, and 3OHC16:0, respectively), with the 27OHC28:0 moiety present as a secondary acyloxyacyl residue ester linked to the hydroxyl group of the 3'-ß-hydroxy fatty acyl residue.
|
As previously reported (22, 40), this structure is most likely an artifact of the lipid A isolation procedure, which promotes both lactonization of the 2-aminogluconate residue and the subsequent acid-catalyzed ß-elimination of the 3-hydroxy fatty acyl moiety (22). The most intense ion in this cluster, m/z 1,653.5, is due to this structure, which lacks a ß-hydroxybutyrate group, and the ion at m/z 1,738.9 is due to the molecule containing the ß-hydroxybutyrate group. Again, the other ions in this cluster are due to variations of this structure as a result of various fatty acyl chain lengths. The structures for the parent R. leguminosarum bv. viciae 3841 lipid A molecules are shown in Fig. 6A and B.
|
The ions in the minor ion cluster, ranging from m/z 1,732.1 to 1,786.0, are approximately 238 mass units larger than the corresponding ions in the m/z 1,493.2 ion cluster, i.e., 1,493.2 + 238 = 1,731.2, 1,521.7 + 238 = 1,759.7, and 1,549.3 + 238 = 1,787.3. The additional 238 mass units could be accounted for by the addition of a palmitoyl residue to the mutant lipid A. Small amounts of palmitic acid were observed in both the parent and mutant lipid A preparations, with increased amounts in the mutant lipid A. However, this fatty acid is normally one of the major components of phospholipids and therefore its presence in the lipid A preparation is usually due to some contamination by phospholipids. Thus, confirmation of palmitic acid as a component of the mutant lipid A will require enrichment of that minor lipid A species and further structural characterization.
The LPSs from the mutants and the parent strain were also compared by deoxycholate-polyacrylamide gel electrophoresis (PAGE) analysis to determine if there were any differences in the LPS banding patterns. The banding patterns for the R. leguminosarum bv. viciae 3841 parent and R. leguminosarum bv. viciae mutant 22 (R. leguminosarum bv. viciae mutant 26 LPS gave the same results as shown for R. leguminosarum bv. viciae mutant 22) are shown in Fig. 7. Some alterations in both intensities and mobilities of the R. leguminosarum bv. viciae mutant 22 LPS bands were observed compared to R. leguminosarum bv. viciae 3841 LPS. The LPS I banding region of the mutant LPS was missing one band, and the LPS II of the mutant was somewhat greater in intensity, with a slightly faster electrophoretic mobility than that of the parent LPS II, a result indicating a slightly lower molecular weight for the mutant LPS II. However, it is somewhat surprising that the mobility of the mutant LPS II, while slightly increased compared to that of the parent, is still so similar, because if the mutant lacks 27OHC28:0, its LPS II should have a molecular weight that is 422 mass units less than that of the parent LPS II. It is possible that an answer to this apparent discrepancy lies in some slight structural difference in the core oligosaccharide of the molecule. Further LPS structural work is in progress.
|
Despite considerable gaps in our understanding of the physiological conditions inside nodules, it is increasingly clear that rhizobia are required to adjust to changes within the nodule microenvironment. Besides the low oxygen conditions of the central nodule tissue and cells, increased osmolarity is likely to exist in the infection threads and the symbiosomes (31). In addition, acidic conditions are likely to occur within the infection thread and symbiosome microenvironments. The rhizobial symbiont would have to adapt to these changes. Therefore, the acpXL::kan mutants were examined for their sensitivity to salt and acidic pHs. The growth of the mutant was monitored on solid medium at added salt concentrations of 0.0%, 0.25%, 0.5%, 0.75%, and 1.0%. Its growth was impaired at 0.25% and abolished at 0.5% salt. Figure 8B shows the results at 0.5% salt. Both the parent and the complemented mutant grew well in 0.5% salt. The pH sensitivity was examined by growing the strains on solid medium at pH 7.0, 6.0, 5.5, 5.0, and 4.8. The mutants were unable to grow at pH 5.0 and below, while the parent and the complemented mutant grew well at pH 4.8 and above. The results at pH 5.0 are shown in Fig. 8C.
|
The rhizobia recovered from mutant nodules were genetically stable, and no revertants were observed, as indicated by the presence of expected antibiotic resistance and agarose gel electrophoretic analysis of the acpXL::kan PCR fragments (data not shown).
| DISCUSSION |
|---|
|
|
|---|
There is a great deal of precedent in the literature for the ability of bacteria to modify their lipid A in response to a changing environment, which illustrates the functional importance of lipid A changes in the adaptation of bacteria to their changing environments. It is known that enteric bacteria modify their lipid A in response to changes in environmental conditions such as those within the host intracellular "spacious phagosomes," i.e., low pH and/or low Mg2+ concentration (for reviews, see references 18 and 41). These modifications, such as the addition of 4-aminoarabinose, phosphoethanolamine, 2-hydroxylation of myristic acid, and the addition of a palmitoyl residue, are mediated by the global regulatory gene products PhoP and PhoQ and are important for protecting the bacteria from the host defenses.
In symbiotic bacteria, the reported doubling of 27OHC28:0 in the lipid A of R. leguminosarum bv. viciae during symbiosis (26) probably reflects one of the ways in which this bacterium adapts to the changing environment of its host. However, the exact function of this structural modification has not yet been determined. For other intracellular pathogens, e.g., Brucella abortus, which survive within phagosomes and cause chronic infections, it is not known whether the lipid A 27OHC28:0 component contributes to virulence.
The first step that was necessary to characterize the symbiotic function of this unusual lipid A 27OHC28:0 fatty acyl component was to isolate a Rhizobium mutant that was unable to incorporate this very long chain fatty acid into its lipid A. In this report, we described the preparation and partial characterization of such a mutant. In summary, we have shown that (i) an acpXL mutant of R. leguminosarum bv. viciae 3841 has been generated, (ii) this mutant does not incorporate significant levels of 27OHC28:0 into its lipid A, resulting in a tetraacylated lipid A molecule, (iii) the mutant has altered growth properties (increased lag time and lower stationary-phase cell density), (iv) the mutant is sensitive to both salt and low pH, (v) it forms nodules more slowly than the parent but these mutant nodules are able to fix nitrogen, and (vi) when the mutant is complemented with normal acpXL, it is corrected in both its lipid A structure and its physiological characteristics.
It is not yet known if the acpXL::kan mutation and resulting lack of 27OHC28:0 on the lipid A affects the biosynthesis of the polysaccharide portion of the R. leguminosarum bv. viciae 3841 LPS. It is clear from the deoxycholate-PAGE analysis that the acpXL::kan mutant is able to produce LPS that contains O-chain polysaccharide, because it has a banding pattern that, while slightly altered, includes the presence of LPS I, which contains O-chain polysaccharide. Detailed structural work on the carbohydrate portions of the LPSs from the parent and mutant is in progress.
Analysis of the mutant lipid A showed that it has the same structure as the parental lipid A, the major difference being the lack of 27OHC28:0. As stated in the introduction, it has been shown that R. leguminosarum bv. viciae contains the same enzymes that convert UDP-GlcNAc into a common lipid A precursor known as Kdo2lipid-IVA as E. coli, that R. leguminosarum bv. viciae has unique enzymes that process this precursor into its unique lipid A, and that AcpXL is required (as described in this paper) for acyloxyacylation of Kdo2lipid-IVA with 27OHC28:0. The fact that the acpXL::kan mutant is able to produce a lipid A structure that is the same as that of the parent except for the missing 27OHC28:0 residue shows that the addition of this fatty acid is not required for processing the Kdo2lipid-IVA precursor into the rhizobial structure, i.e., 27OHC28:0 acylation is not required for removal of the 4'- and 1-phosphates, addition of the galacturonosyl residue, or oxidation of the proximal glucosaminosyl residue to 2-aminogluconate.
The mutant lipid A preparation has a minor molecular species in which the mass of the molecular ion indicates that the 27OHC28:0 residue has been replaced with a palmitoyl residue. The addition of a straight-chain fatty acyl moiety to the rhizobial lipid A would be unusual because it has been reported that the transfer of straight-chain fatty acyl substituents to R. etli or R. leguminosarum bv. viciae lipid A does not occur (10), and, in fact, no such fatty acyl residues have been observed on these lipid A molecules (5, 39, 40). Changes in lipid A fatty acylation do occur in enteric bacteria. For example, under certain conditions, it has been reported that Salmonella spp. can transfer a palmitoyl residue from glycerophospholipids to its lipid A via the action of PagP, a protein that is under PhoPQ regulation (7). However, there is no evidence of a PagP homolog in either the Mesorhizobium loti or Sinorhizobium meliloti genome sequence (www.kazusa.or.jp/en/, http://sequence.toulouse.inra.fr/meliloti.html). The R. leguminosarum bv. viciae 3841 genome sequence is not yet finished; however, there is no PagP homolog in the sequence that is currently available (www.sanger.ac.uk/Projects/R_leguminosarum/). It will be necessary to isolate this minor lipid A species in larger amounts in order to determine its structure and verify whether or not it contains a palmitoyl residue and, if so, the location of that residue.
The increased lag phase and the lower LPS yield of the R. leguminosarum bv. viciae acpXL::kan mutant may be due, indirectly, to the production of tetraacylated lipid A. In the case of an E. coli lpxL mutant, tetraacylated lipid A accumulates at the inner membrane (49, 50), suggesting inefficient transport. It may be that the reduced yield of LPS from the R. leguminosarum bv. viciae 3841 acpXL::kan mutant and, perhaps, the increased lag time are due to poor transport of the tetraacylated lipid A. The inefficient transport of the tetracyl lipid A in the E. coli lpxL mutant can be suppressed by overexpression of the ABC transporter MsbA (15). Further investigation is required to determine if the R. leguminosarum bv. viciae 3841 acpXL::kan mutant accumulates tetraacyl lipid A at the inner membrane.
During symbiosis, the bacterial symbiont must adapt to the environment of its plant host. These environmental changes likely require that the bacteria adjust to changes in osmolarity, acidity, and O2 tension. In view of these necessary adjustments, it is interesting that the R. leguminosarum bv. viciae acpXL::kan mutant is able to form nitrogen-fixing nodules even though it is unable to grow at pH 5.0 or in 0.5% NaCl. These results suggest that the mutant, in planta, is capable of compensating, at least in part, for the loss of acpXL. In a recent review by Geiger and Lopez-Lara (21), the presence of multiple types of acyl carrier protein in various bacterial species was discussed. Some members of the Rhizobiaceae have several specialized types of acyl carrier protein, including AcpXL and NodF (required for acylation of the acylated chitin oligomers known as Nod factors) (21). Thus, it may be that an additional acyl carrier protein comes into play in the in planta environment.
In the case of S. meliloti, Geiger and Lopez-Lara reported that an acyl carrier protein is present on the symbiotic plasmid, and other workers have shown that various long-chain
-1 hydroxy fatty acyl residues (C18 to C26) can be incorporated into the Nod factors and that this incorporation is regulated by nodD3 (13, 21). Thus, it may be that R. leguminosarum bv. viciae has an alternative mechanism for the addition of 27OHC28:0 to lipid A which is host inducible. The fact that the lipid A from R. leguminosarum bv. viciae 3841 isolated from bacteria grown under low O2 has increased levels of 27OHC28:0 supports the possibility of such an additional and, perhaps, inducible mechanism for 27OHC28:0 acylation.
Further examination of this R. leguminosarum bv. viciae acpXL::kan mutant may also facilitate understanding of the virulence mechanism of several gram-negative pathogens. As mentioned in the introduction, long-chain
-1 hydroxy or oxo-fatty acyl residues are present in a variety of gram-negative bacterial species in addition to members of the Rhizobiaceae. The presence of these lipid A long-chain
-1 fatty acyl residues and/or acpXL or lpxXL homologs has been reported in a number of gram-negative facultative intracellular organisms, including Brucella abortus (4), Brucella melitensis (1), Legionella pneumophila (54, 55), Bartonella henselae (4), and Rickettsia prowazekii (1). Thus, the presence of the 27OHC28:0 in the lipid A of such pathogens may play a role in their ability to survive within the host cell.
As with the proposed function for 27OHC28:0 in Rhizobium-legume symbiosis, this very long chain fatty acyl component may increase the stability of the bacterial membrane. It is also possible that this fatty acyl residue results in a low endotoxic response in the host. In fact, the R. etli LPSs and the Rhizobium sp. strain Sin-1 LPS, which has a lipid A structure similar to that of R. etli, do not induce cytokines in human macrophages (monomac 6) cells and inhibit the ability of enteric LPSs to induce cytokines by interfering with endotoxin binding to both LPS binding protein and the cellular differentiation protein CD14 (47). It would be very instructive to characterize the effects of acpXL::kan mutations on the virulence of these intracellular pathogens.
| ACKNOWLEDGMENTS |
|---|
This work was supported by NIH grant GM89583 to R.W.C. and, in part, by DOE grant DE-FG02-93ER20097 to the CCRC and Sonderforschungsbereich 323 to E.L.K. at the University of Tübingen.
| FOOTNOTES |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| ||||||||||||||||||||