The Pea Nodule Environment Restores the Ability of a Rhizobium leguminosarum Lipopolysaccharide acpXL Mutant To Add 27-Hydroxyoctacosanoic A
http://www.100md.com
《细菌学杂志》
Complex Carbohydrate Research Center, University of Georgia, 315 Riverbend Road, Athens, Georgia 30602,Department of Plant and Soil Sciences and the Delaware Biotechnology Institute, University of Delaware, Newark, Delaware 19711,Department of Microbiology and Biotechnology, University of Tubingen, D072076 Tubingen, Germany
ABSTRACT
Members of the Rhizobiaceae contain 27-hydroxyoctacosanoic acid (27OHC28:0) in their lipid A. A Rhizobium leguminosarum 3841 acpXL mutant (named here Rlv22) lacking a functional specialized acyl carrier lacked 27OHC28:0 in its lipid A, had altered growth and physiological properties (e.g., it was unable to grow in the presence of an elevated salt concentration [0.5% NaCl]), and formed irregularly shaped bacteroids, and the synchronous division of this mutant and the host plant-derived symbiosome membrane was disrupted. In spite of these defects, the mutant was able to persist within the root nodule cells and eventually form, albeit inefficiently, nitrogen-fixing bacteroids. This result suggested that while it is in a host root nodule, the mutant may have some mechanism by which it adapts to the loss of 27OHC28:0 from its lipid A. In order to further define the function of this fatty acyl residue, it was necessary to examine the lipid A isolated from mutant bacteroids. In this report we show that addition of 27OHC28:0 to the lipid A of Rlv22 lipopolysaccharides is partially restored in Rlv22 acpXL mutant bacteroids. We hypothesize that R. leguminosarum bv. viciae 3841 contains an alternate mechanism (e.g., another acp gene) for the synthesis of 27OHC28:0, which is activated when the bacteria are in the nodule environment, and that it is this alternative mechanism which functionally replaces acpXL and is responsible for the synthesis of 27OHC28:0-containing lipid A in the Rlv22 acpXL bacteroids.
INTRODUCTION
Rhizobium leguminosarum cells have an envelope similar to that of other gram-negative bacteria. Lipopolysaccharide (LPS) is the primary component of the bacterial outer leaflet and is comprised of three structural regions: the O-chain polysaccharide, the core oligosaccharide, and lipid A. The lipid A region is anchored in the bacterial outer membrane, and the carbohydrate portion projects from the outer surface into the surrounding milieu and is the primary immunogenic determinant. Correlating LPS structure with function has been difficult as LPS is a very complex molecule and LPS preparations consist of structurally heterogeneous mixtures of molecules. There are pronounced variations in LPS structure from strain to strain, and even within a strain there are different sizes and compositions of LPS (23). Interestingly, there are marked differences between LPSs from free-living cultures and LPSs from nitrogen-fixing bacteroids in terms of size, composition, and antigenic properties (10, 32, 33). It has been shown that the rhizobial LPS undergoes structural modifications during the formation of bacteroids and that there are composition differences between the bacterial and bacteroid LPSs (10, 32).
Variation in LPS structure due to environmental changes has been studied in cultured rhizobia by altering the growth conditions, such as lowering the oxygen level, lowering the pH, altering the carbon source, or adding plant-derived compounds (2, 14, 21, 23, 27). Such studies have shown that cues from the environment play an important role in LPS composition. The results led to the hypothesis that the bacterial LPS structure inside legume root nodules is probably controlled, to a large extent, by the in planta microenvironmental conditions. In the study of Kannenberg and Carlson (22), Rhizobium leguminosarum bv. viciae 3841 (named here Rlv3841) was cultured under various growth conditions, and the LPS structural modifications were analyzed chemically and immunochemically. It was observed that the LPS extracted from nodule bacteria or from laboratory-grown bacteria cultured under low-oxygen conditions was much more hydrophobic than the LPS from bacteria grown under normal laboratory conditions. Chemical analysis of the LPS derived from bacteria grown under low-oxygen conditions indicated that changes occurred in both the polysaccharide and lipid A portions of the LPS; the polysaccharide was affected in the extent of methylation and acetylation, while the lipid A showed an increase in a unique very-long-chain fatty acyl component, 27-hydroxyoctacosanoic acid (27OHC28:0). This suggested that the low-oxygen conditions in nodule cells may cause similar structural changes to the LPS.
The functions of LPS structural changes that occur during symbiosis are not known. These changes may be needed for the increased membrane stability and barrier properties that are required for the bacteroid to persist and function within the symbiosome compartment. One of the LPS components that may be essential in the symbiotic process is the very-long-chain fatty acyl component 27OHC28:0 in lipid A. This fatty acyl component is present in the LPS of members of the Rhizobiaceae (5, 11, 23). In addition, a number of facultative intracellular pathogenic bacterial species that cause chronic infections also contain this lipid A fatty acyl residue or orthologs of acpXL or lpxXL' genes that are required for its synthesis and transfer to lipid A. These species include Brucella abortus (3), Brucella melitensis (1), Legionella pneumophila (39, 40), and Bartonella henselae (3). Thus, it is possible that the 27OHC28:0 residue may be required for endosymbiotic rhizobia and these intracellular pathogens to persist and function within their host cells.
In order to examine the symbiotic function of the 27OHC28:0 lipid A component, we prepared and characterized an LPS mutant that is defective in the acyl carrier protein (ACP) required for its synthesis, AcpXL (34, 35). Laboratory-grown cultures of this mutant, Rlv22, produced an LPS that did not contain 27OHC28:0 in its lipid A and was unaffected in its O-chain polysaccharide and core oligosaccharide structures (35). The Rlv22 mutant was unable to grow under laboratory conditions at a low pH (pH 5.0) or with 0.5% NaCl added to the medium, and nodule development was delayed, although eventually nitrogen-fixing nodules were formed. Similar results were reported by Sharypova et al. for an acpXL mutant of Sinorhizobium meliloti (31). The ability of Rlv22 to form nitrogen-fixing nodules in spite of its inability to adjust to changes in osmotic strength or pH ex planta suggested that it may adapt in a specific manner to the in planta conditions (34, 35). Microscopic examination revealed that the Rlv22 mutant formed large irregularly shaped bacteroids and that multiple bacteroids were often surrounded by a single symbiosome membrane (34). However, it was also observed that a significant number of normal bacteroids were present, and, as previously observed, nitrogen-fixing nodules formed, although the level of nitrogenase in Rlv22-induced nodules was significantly lower than the level in normal nodules (34). Thus, it was concluded that depletion of 27OHC28:0 in the lipid A of the Rlv22 mutant disrupted bacteroid development and the synchrony between bacterial and symbiosome membrane division. However, the eventual appearance of some normal bacteroids and nitrogen-fixing nodules suggested that the mutant, once it was in its host, was able to partially compensate in some manner for the loss of 27OHC28:0 in its lipid A. Investigation of the nature of this compensation required isolation and characterization of the lipid A from mutant bacteroids. In this paper, we report the results of this analysis.
MATERIALS AND METHODS
Bacterial strains and culture conditions. The strains used are listed in Table 1. Strains were cultured on solid or in liquid tryptone yeast extract (TY) medium (Difco Laboratories), as previously described (35). Cultures were supplemented with kanamycin (Kan) (50 μg ml–1) and streptomycin (Str) (100 μg ml–1), when appropriate. The osmotic tolerance of bacterial isolates was studied by growing two strains on TY containing 0.5% NaCl at 30°C, along with the appropriate controls. The pH tolerance of the bacterial isolates was examined at pH 5, using piperazine-buffered TY plates as described by Vedam et al. (35).
Growth of peas. Peas (Pisum sativum cv. Early Alaska) were surface sterilized by soaking them in 95% ethanol for 30 s, rinsing them three times with sterile water, washing them with bleach for 3 min, and rinsing them 10 times with sterile water. Seeds were placed in sterile petri dishes with a small amount of sterile water and allowed to imbibe and germinate in the dark for 48 h. Pea seedlings were sown in "caissons" and supplied with Lullien's nutrient solution (25) lacking ammonium nitrate. Caissons are aeroponic growth systems that provide controlled conditions for the growth of peas. After 1 week of germination and stabilization of the seedlings, the peas were inoculated with either Rlv3841 (wild type) or Rlv22 (acpXL mutant). The pea plants were allowed to nodulate for 3 weeks, and the caissons were periodically refreshed with nutrient solution. At the end of the growth period, the nodules were harvested for preparation of bacteroids.
Extraction of bacteroids from pea nodules. Nodules were harvested from pea plants that had been inoculated with either Rlv3841 or Rlv22 into an ice-cold solution of 0.5 M sucrose in 50 mM Tris-HCl (pH 7.4) plus a 1:100 dilution of protease inhibitor cocktail (Sigma P9599). For (bio)chemical analysis, bacteroids were isolated from these nodules and purified by a procedure involving a sucrose step gradient, as described previously (13). Briefly, pea nodules were washed with cold Tris-HCl/sucrose buffer (0.5 M sucrose-50 mM Tris-HCl [pH 8.0] at 4°C containing dithiothreitol, proteinase inhibitor, and polyvinylpolypyrrolidone), suspended in the same buffer, and ground up using a mortar and pestle. The initial steps of the isolation protocol were done with sucrose-containing buffer to osmotically stabilize symbiosomes and bacteroids. To remove tissue and cell debris, the crushed nodule material was filtered through miracloth and rinsed with the same solution. The rinse solution was added to the filtrate, and the resulting suspension was centrifuged for 1 min at 10,000 x g. The pellet, containing the symbiosomes, was resuspended in Tris-HCl/sucrose buffer. Portions of the suspension were distributed in several microcentrifuge tubes, overlaid onto sucrose cushions (composed of 1.5 M sucrose and 50 mM Tris-HCl [pH 8.0] at 4°C), and centrifuged at 5,000 x g for 30 s. The top phases of the different tubes, strongly enriched in the bacteroid-containing symbiosomes, were transferred to a clean tube and centrifuged at 10,000 x g for 90 s to collect the symbiosome fraction in the pellet. The supernatant was discarded, and the pellet was resuspended in Tris-HCl/sucrose buffer and again overlaid onto a sucrose cushion as described above; however, this time the preparation was centrifuged at 10,000 x g for 5 min, which sedimented the symbiosomes with the bacteroids in the pellet. To remove the peribacteroid membrane by osmotic shock from the symbiosomes and wash the bacteroids, the pellets were repeatedly (two or three times) resuspended, with vigorous mixing, in 500 μl of Tris-HCl buffer without sucrose (0.50 mM Tris-HCl [pH 8.0] with dithiothreitol and proteinase inhibitor) and centrifuged; the final pellet was suspended in 1,600 μl of the same buffer and stored at –20°C.
During this work, individual colonies obtained from Rlv22-induced nodules were examined for antibiotic resistance and the presence of acpXL::kan. Resistance to Kan and Str was measured for about 500 colonies using growth on solid agar with and without the antibiotics. The presence of acpXL::kan (or acpXL) was measured for 12 of the colonies described above using PCR and primers for acpXL (GAGGGGGTTTAAATAGTCA and AGGCTTGGCCGCTTTGA), as previously described by Vedam et al. (35).
To rule out the possibility that a small number of possible mutant revertants escaped the analysis described above and occupied nodules undetected, a PCR screening analysis was performed directly with bacteroid preparations from nodules using the primers specific for acpXL described above in order to determine if any of the wild-type acpXL gene was present. Preparations of bacteroids isolated from nodules of 10 pea plants induced by Rlv3841 or Rlv22 were PCR screened. Aliquots of the bacteroid preparations (2 to 5 μl from an approximately 500-μl dense suspension of bacteroids) were pipetted into the PCR mixture and lysed during an initial 15-min hot start in the assay. The PCR products were analyzed using 1% agarose gel electrophoresis.
Preparation of bacterial isolates from Rlv22-induced nodules (Rlv22 EN isolates). In order to fully examine the effect of passage of Rlv22 through the plant, it was necessary to isolate and characterize in more detail mutant isolates obtained from pea root nodules. Rlv22 ex nodule (EN) isolates were analyzed for antibiotic resistance, the presence of acpXL::kan or acpXL, and sensitivity to growth at a low pH (pH 5.0) and in the presence of 0.5% NaCl, and their LPSs were extracted and analyzed as described below. Seeds were germinated and grown in Erlenmeyer flasks containing solid Fahraeus medium. After surface sterilization (see above) and transfer into the flasks, pea seeds were immediately inoculated with the Rlv22 mutant, and plant growth and nodulation were evaluated as described by Brewin et al. (9). Root nodules from several pea plants were removed, briefly washed in 95% ethanol, and, by size, surface sterilized in diluted bleach for 30 to 60 s. The nodules were transferred through a series of washes with sterile water and finally crushed to squeeze the occupying bacteria out onto TY agar plates. The plates were developed, and three single colonies from each plate were streaked onto fresh plates. Care was taken to pick colonies of different sizes (small, medium, and large) in case the size differences represented different variants of the original mutant. Two or three nodules per plant were analyzed in this way. From the Rlv22 EN colonies (kept on TY plates), a total of 23 clones (derived from four plants and eight nodules) were analyzed for antibiotic resistance (resistance to Str and Kan). A random selection of 16 clones (two clones per nodule) were tested further for sensitivity to salt (0.5% NaCl) and to an acidic pH (pH 5). As a control and for comparison, strains Rlv3841 and Rlv22 were included in the antibiotic, saline, and low-pH sensitivity tests. Since all Rlv22 EN isolates behaved similarly in these tests, two random Rlv22 EN isolates, EN2 and EN4, were selected, and we verified that they contained the acpXL::kan mutation using PCR and the primers for acpXL as previously described by Vedam et al. (35).
Lipid A purification. LPSs were extracted by the triethylamine (TEA)/EDTA/phenol procedure as previously described (25). Briefly, for each strain, the LPS was extracted from the bacterial or bacteroid pellet using 3 volumes of TEA/EDTA/phenol (0.25 m EDTA, 5% phenol, titrated to pH 6.9 with TEA with constant stirring for 1 h at 37°C). The extract was centrifuged at 13,000 x g for 1 h, and the supernatant was collected and dialyzed (molecular weight cutoff, 10,000; Spectrapor) against deionized water. The bacterial and bacteroid LPSs were lyophilized for analysis. Lipid A was isolated from the LPS preparations by mild acid hydrolysis (12). Briefly, the LPS was dissolved in 1% sodium dodecyl sulfate in 20 mM sodium acetate, the pH was adjusted to 4.5 with 4 M HCl, and then the preparation was placed in an ultrasound bath until the sample was dissolved. The solution was then heated to 100°C for 1 h, which was followed by lyophilization. The sodium dodecyl sulfate was removed by washing the lyophilized residue with a 2:1 solution of deionized H2O and acidified ethanol (100 μl of 4 M HCl in 20 ml of 95% ethanol). The residue was collected by centrifugation, washed with 95% ethanol (nonacidified), and collected by centrifugation (200 x g for 15 min). The washing and centrifugation steps were repeated. Finally, the residue was lyophilized to obtain a white, solid, fluffy lipid A preparation.
Analytical procedures. The composition was determined by gas chromatography-mass spectrometry (MS) analysis of trimethylsilyl derivatives of methyl glycosides and hydroxy fatty acid methyl esters as described previously (4, 38). Matrix-assisted laser desorption ionization—time of flight (MALDI-TOF) MS was performed in the negative-ion reflectron mode with a 337-nm nitrogen laser operating at an extraction voltage of 20 kV 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/ml of methanol) and applied to the probe for mass analysis. Spectra were calibrated externally using E. coli lipid A (Sigma).
RESULTS
Stability of the acpXL mutation in Rlv22 during nodule development. In order to rule out the possibility of genetic reversion of the acpXL mutation in planta, bacterial colonies obtained from nodules infected by Rlv22 were tested for resistance to kanamycin by using an antibiotic marker cassette introduced into acpXL (35). Approximately 500 colonies were screened, and there was no loss of the resistance marker; i.e., all 500 colonies were resistant to both Kan and Str. In addition, genomic DNA from 12 of these colonies was tested by PCR, using primers for acpXL. Each of the 12 colonies showed only the presence of acpXL::kan, indicating that the mutation was intact. Furthermore, chromosomal DNA from two of these colonies was analyzed by Southern hybridization using amplified acpXL as the labeled probe, and the results showed that both colonies retained the mutation. In addition, bacteroids obtained from Rlv22-induced nodules were directly examined by PCR using primers for acpXL. As a control, PCR analysis was also performed with bacteroids from Rlv3841-induced nodules. The results (Fig. 1) showed that bacteroids from Rlv22-induced nodules contained only acpXL::kan and that there was no detectable evidence of intact acpXL. This finding confirmed the stability of the acpXL mutation of Rlv22 during nodule development.
Analysis of the lipid A from Rlv22 bacteroids. The lipid As from LPS preparations of parent Rlv3841 and mutant Rlv22 laboratory-grown bacteria and from bacteroids isolated from root nodules were analyzed to determine their fatty acid compositions (Table 2) and by MALDI-TOF MS (Fig. 2).
The composition analysis revealed that the lipid A from Rlv3841 laboratory-grown cells contained the same fatty acids at the same ratio that was observed for the lipid A from Rlv3841 bacteroids. In contrast, the Rlv22 laboratory-grown cells produced lipid A in which no long-chain 27OHC28:0 could be detected. However, the lipid A from Rlv22 bacteroids contained 27OHC28:0 at a level that was about 50% of the level observed in the Rlv3841 parent lipid A. In addition, the Rlv22 lipid A preparations contained noticeably increased levels of palmitate (C16:0) and stearate (C18:0) compared to the levels in the Rlv3841 lipid A samples.
The mass spectrum of the lipid A from laboratory-grown Rlv3841 (Fig. 2A) shows that there were two clusters of ions. The first cluster of ions ranged from m/z 1887.6 to 2058.0, with the most intense ion at m/z 1914.0, and the other cluster of ions ranged from m/z 1625.7 to 1738.9, with the most intense ion at m/z 1652.0. The structures corresponding to the lipid A ions for the laboratory-grown Rlv3841 (structures I and II) have been reported previously (4, 35) and are shown in Fig. 2. The ion at m/z 1914.0 is consistent with the previously published R. leguminosarum or Rhizobium etli lipid A structure (4, 28, 29, 35), in which the lipid A has a disaccharide backbone consisting of a distal glucosaminosyl residue that is -1,6 linked to a proximal 2-aminogluconate (GlcNonate) residue. At the 4' position of the distal glucosamine there is an -galacturonosyl residue substitution, and the -glucosaminosyl-(16)-GlcNonate disaccharide is acylated with -hydroxy fatty acids at the 2, 3, 2', and 3' positions. The 27OHC28:0 lipid A moiety is present as a secondary acyloxyacyl residue and is ester linked to the hydroxy group of the 3'--hydroxy fatty acid residue. The ions at m/z 2001.4 and above are due to molecules in which the 27-OH group of the 27OHC28:0 molecule is esterified with -hydroxybutyrate. Other ions in this cluster are due to structural variants resulting from different fatty acyl chain lengths. In the second ion cluster, the most intense ion is the ion at m/z 1652.0 and is likely due to a structure (structure II) caused by elimination of the -hydroxy fatty acyl residue from position 3 of the GlcNonate residue forming a 2,3-unsaturated 2-aminoglucono-1,5-lactone residue; this is a reaction which may be an artifact of the lipid A isolation procedure (20, 35).
The mass spectrum of the lipid A preparation from laboratory-grown mutant Rlv22 (Fig. 2C) also shows that there were two main ion clusters, one centered around the ion at m/z 1493.0 and the other centered around the ion at m/z 1230.0. A third minor cluster is centered around the ion at m/z 1758.0. The two main ion clusters represent structures that are devoid of 27OHC28:0 or 27O(-hydroxybutyryl)C28:0 (structures III and IV). These structures are identical to those previously reported for the mutant lipid A (35). We also reported previously (35) that the minor cluster of ions centered around m/z 1758.0 was due to replacement of 27OHC28:0 with a palmitoyl (C16:0) residue. These ions were observed again in the current study, and this result, together with the composition data (Table 2), could have been due to structures in which 27OHC28:0 was replaced with a C16:0 or stearoyl (C18:0) residue.
Figures 2B and 2D show the mass spectra of the lipid A preparations from Rlv3841 and Rlv22 bacteroids, respectively. The mass spectrum of the Rlv3841 bacteroid lipid A preparation (Fig. 2B) is identical to that of the lipid A from the laboratory-grown culture (Fig. 2A) described above. The spectrum of Rlv22 bacteroid lipid A has four major ion clusters (Fig. 2D). Three of these ion clusters are identical to those observed for the lipid A from laboratory-grown Rlv22 (Fig. 2C). The fourth ion cluster is centered around the m/z 1914.0 ion and is identical to the second ion cluster for the lipid As from both the Rlv3841 bacteroids and laboratory-cultured cells. This result, together with the composition data described above, shows that the 27OHC28:0 moiety was present in some of the structures in the mutant bacteroid lipid A preparation. Also present in the Rlv22 bacteroid lipid A preparation were ions that were consistent with structures that lack 27OHC28:0 but contain an added C16:0 or C18:0 residue. In summary, the lipid A preparation from the Rlv3841 bacteroids appeared to have the same structures as the preparation from the Rlv3841 laboratory-grown cells. In contrast, Rlv22 laboratory-grown cells lacked 27OHC28:0 in their lipid A (Fig. 3C), but bacteroids from Rlv22-induced nodules produced some lipid A molecules that contained 27OHC28:0 (Fig. 3D).
Properties of EN acpXL mutant isolates from Rlv22-infected pea plants. In order to determine if passage through the plant had selected for an Rlv22 variant that constitutively expressed an alternative mechanism for synthesis of 27OHC28:0, we isolated and examined a number of ex nodule mutant isolates. We showed previously that the Rlv22 acpXL::kan mutant was unable to grow in the presence of 0.5% NaCl or at pH 5.0 (35). Sixteen Rlv22 EN isolates were grown under the same test conditions as the original mutant. We observed that all of the Rlv22 EN isolates were able to grow in the presence of 0.5% NaCl (Fig. 3). On plates with acidic growth conditions (pH 5), the Rlv22 EN isolates showed stronger growth than the original Rlv22 mutant, although the growth was not as good as the growth of wild-type strain Rlv3841 in the same experiment. PCR analysis to detect either acpXL or acpXL::kan was performed with the genomic DNA from two Rlv22 EN isolates, EN2 and EN4, using primers specific for acpXL. The results were the same as those obtained for the bacteroid preparations shown in Fig. 1 and showed that the PCR products obtained from the Rlv22 EN isolates were each 1.56 kb long, which is consistent with the size of acpXL::kan, and were the same size as the PCR product from the original Rlv22 strain. A PCR product indicating that there was an intact acpXL gene was not detected in either the EN2 or EN4 isolate.
In order to determine the presence of 27OHC28:0 in the lipid A from the two Rlv22 EN mutant isolates, lipid A was prepared from these two strains grown under normal laboratory conditions or in the presence of 0.5% NaCl. The lipid A samples were then analyzed by MALDI-TOF MS to determine their fatty acyl compositions. The original Rlv22 mutant was also included in this study as a control. The results are shown in Fig. 4. The mass spectrum of the lipid A from the original Rlv22 mutant (Fig. 4A) was the same as the mass spectrum described above and shown in Fig. 2C. The mass spectra of the lipid As from EN2 and EN4 grown under normal laboratory conditions (Fig. 4B) or in the presence of 0.5% NaCl were identical to one another and very similar to the mass spectrum of the lipid A from the original Rlv22 mutant in that all of the mass spectra lacked the 27OHC28:0 residue. The minor difference between the spectra of EN2 and EN4 lipid A preparations and the spectrum of the Rlv22 lipid A was that the third minor ion cluster, likely due to replacement of 27OHC28:0 with a palmitoyl or stearoyl residue, was significantly less intense in the EN2 and EN4 lipid A preparations.
DISCUSSION
We previously showed that R. leguminosarum acpXL mutant Rlv22 is unable to add 27OHC28:0 to its lipid A, exhibits delayed nodulation and nitrogen fixation, forms many aberrantly shaped bacteroids, exhibits disrupted synchrony of bacterium and symbiosome membrane division, and, under laboratory culture conditions, is unable to grow at a low pH (pH 5.0) or in the presence of 0.5% NaCl (35). Sharypova et al. (31) showed that an acpXL mutant of S. meliloti also lacked 27OHC28:0 in its lipid A, was sensitive to deoxycholate, exhibited delayed nodulation of Medicago sativa, and had reduced competitive ability. However, nodules elicited by this mutant on roots of M. sativa and Medicago truncatula had normal morphology and fixed nitrogen. Sharypova et al. did not describe an effect of the acpXL mutation on S. meliloti bacteroid development. The aberrant bacteroids that we observed for the Rlv22 mutant (34) support the idea that the 27OHC28:0 moiety of lipid A plays an important role in bacteroid development and in the synchronous division of bacteria and the symbiosome membrane. However, the ability of the mutant to eventually form nitrogen-fixing nodules in a process in which it is likely to encounter acidic pH conditions and changes in osmolarity raised the question of whether Rlv22 adapts in some manner to the loss of 27OHC28:0 within the host plant. In this study, we addressed this question by examining the lipid A isolated from Rlv22 bacteroids and from ex nodule acpXL mutant isolates EN2 and EN4 grown under laboratory conditions at normal NaCl concentrations and in the presence of 0.5% NaCl. We found that, unlike laboratory cultures, Rlv22 bacteroids are able to add 27OHC28:0 to their lipid A. Furthermore, we observed that the acpXL::kan mutation was maintained in Rlv22 bacteroids and in the EN isolates of Rlv22 and that, when grown under laboratory conditions, the EN isolates were still unable to add 27OHC28:0 to their lipid A. These results indicate that in Rlv3841 an alternative mechanism is induced within the nodule that can functionally replace the inactivated AcpXL in the acpXL::kan mutant. The presence of an alternative to acpXL for the synthesis of 27OHC28:0 that is induced within the host suggests that this fatty acyl residue in lipid A may be essential for the symbiotic interaction between R. leguminosarum and peas.
As mentioned above and consistent with our results with Rlv3841, recent work with S. meliloti also showed that an acpXL mutant lacked the 27OHC28:0 residue in its lipid A (31). A subsequent study (15) demonstrated that the S. meliloti acpXL mutant contained another lipid A species, in which a C18:0 residue replaced the 27OHC28:0, a result that is consistent with our results (21; this study), which shows that the Rlv22 mutant can produce a lipid A molecule in which the 27OHC28:0 moiety is replaced by a C16:0 or C18:0 residue. Since replacement of 27OHC28:0 by C18:0 in the S. meliloti acpXL mutant was prevented by a second mutation in the gene encoding the 27OHC28:0 acyltransferase (lpxXL), it was concluded that LpxXL is able to transfer either C18:0 or 27OHC28:0 to the acyloxyacyl position of S. meliloti lipid A (15). This report (15) also showed that the S. meliloti acpXL mutant, an lpxXL mutant, and an lpxXL acpXL double mutant were able to form, in a delayed manner, nitrogen-fixing nodules. The latter two mutants were particularly sensitive to stressful conditions, such as detergents or increased salt concentrations. While these results also suggest that in S. meliloti there may be alternative mechanisms that are activated within the host that can functionally replace the mutated acpXL, as well as lpxXL, there has been no analysis of the lipid A from S. meliloti acpXL or lpxXL mutant bacteroids.
The nature of this alternative host-induced mechanism for synthesizing 27OHC28:0 is unknown. However, it has been reported that S. meliloti possesses multiple ACPs. Apart from the four known major ACPs in rhizobia, genomics has predicted the existence of additional ACPs (19). The complete sequence of S. meliloti indicates that there are at least two novel ACPs. One of the ACP genes is located on the Sym plasmid (it is located in a cluster of four genes, all closely linked, perhaps belonging to one operon), and the other is located on the chromosome. The Mesorhizobium loti genome also contains an operon similar to the operon identified in S. meliloti (19). Since rhizobia possess multiple ACPs, it is possible that R. leguminosarum contains an additional ACP that could be activated in planta, thereby compensating for the disrupted acpXL in Rlv22 during symbiosis. In this regard we have located in the Rlv3841 genome sequence (www.sanger.ac.uk), which has not been fully annotated yet, two such possible acp gene candidates. One acp candidate is located on the chromosome, and the second is located on the symbiotic plasmid, pRL10. These two putative acp genes encode ACPs that are very similar to one another and to ACPs from Agrobacterium and a number of Burkholderia strains, particularly strains of the pathogens Burkholderia mallei and Burkholderia pseudomallei. The Rlv3841 ACP gene region in pRL10 is shown in Fig. 5. The acp gene is next to a gene encoding an acyl-ACP dehydrogenase. It is preceded by DNA sequence binding motifs for NifA and for alternative sigma factor 54 in a DNA region that exhibits sequence similarity to a region that precedes the fixWABC genes in R. leguminosarum UPM791 (26). This location suggests that expression of the acp gene may be regulated by the O2 status of the bacterial cell. The possibility that the other ACPs may be involved in the in planta ability of the acpXL::kan mutant to synthesize 27OHC28:0 is under investigation.
27OHC28:0 in the lipid A or orthologs of acpXL or lpxXL have been found in a number of bacterial species, including several pathogens that cause chronic intracellular infections. In all cases in which the gene sequences can be compared, there is a very similar arrangement of genes encoding fatty acyl-ACP synthase orthologs (three or four genes) flanked by acpXL on one side and by lpxXL on the other side. The arrangement of this gene region is shown for several bacterial species in Fig. 6. There is some variability in this gene region; some species do not contain the putative fatty acid ACP dehydratase (Orf1) gene, while others do not contain the dehydrogenase (Orf4) gene. In all cases, genes encoding proteins with high levels of similarity to AcpXL, LpxXL, Orf2 (a putative 3-oxoacyl ACP synthase), and Orf3 (a putative second 3-oxoacyl ACP synthase) are present. The only exception to the high level of sequence similarity in this gene region is the region from Legionella pneumophila. However, the arrangement of the genes in L. pneumophila is identical to the arrangement of this region in the other bacterial species shown. Therefore, it is highly likely that this gene region in all of these bacterial species consists of a cassette of five to six genes that are specifically required for the synthesis of very-long-chain -hydroxy or -oxo fatty acids and their transfer to lipid A.
In the case of the brucellae, their commonality with rhizobia is that they are both endocytosed into their host cells and are able to persist and function within host-derived membrane compartments (symbiosomes and phagosomes, respectively) (7, 8, 30). Thus, it is possible that the 27OHC28:0 lipid A component in both of these taxa plays an important role in their virulence and persistence within their hosts. Recently, another connection between Brucella abortus and rhizobia has been reported with regard to a protein called BacA and its importance for producing lipid A with 27OHC28:0 (16). The BacA protein was found to be necessary for infection of alfalfa by S. meliloti (17) and was also shown to be present in B. abortus and necessary for this pathogen to maintain a chronic infection in its host (24). The BacA protein exhibits sequence similarity with a peroxisomal membrane protein family (16) that is thought to play a role in the transport of long-chain fatty acids (6, 18, 36, 41). BacA mutants of S. meliloti and B. abortus both produce lipid A preparations in which the level of 27OHC28:0 is reduced to about 50% of the normal level (16). Unlike AcpXL, which is specifically required for the synthesis of 27-OHC28:0, it is likely that the BacA protein has multiple functions that affect symbiotic infection, including a function which affects the level of 27-OHC28:0 in lipid A and another function which is essential for infection.
In summary, we showed that the acpXL mutation is partially suppressed by a possible host-activated alternative mechanism for the synthesis of 27OHC28:0. This result suggests that the presence of 27OHC28:0 in rhizobial lipid A is essential for symbiosis. Further investigation is in progress to (i) identify the alternative mechanism for 27OHC28:0 synthesis that occurs in planta and (ii) prepare and characterize the symbiotic phenotypes of mutants (e.g., deletion of the entire acpXL-lpxXL region) that are unable to synthesize 27OHC28:0 both ex planta and in planta. In addition to increasing our understanding of the molecular basis of the Rhizobium-legume symbiosis, determining the role of 27OHC28:0 in the symbiotic process should also provide information regarding the virulence mechanism of the brucellae and possibly several other pathogen species that cause chronic intracellular infections.
ACKNOWLEDGMENTS
We thank Thomas Hrtner (University of Tubingen) for his help with isolating ex nodule rhizobia.
REFERENCES
Basu, S. S., M. J. Karbarz, and C. R. H. Raetz. 2002. Expression, cloning and characterization of the C28 acyltranferase of lipid A biosynthesis in Rhizobium leguminosarum. J. Biol. Chem. 277:28959-28971.
Bhat, U. R., and R. W. Carlson. 1992. Chemical characterization of pH-dependent structural epitopes of lipopolysaccharides from Rhizobium leguminosarum biovar phaseoli. J. Bacteriol. 174:2230-2235.
Bhat, U. R., R. W. Carlson, M. Busch, and H. Mayer. 1991. Distribution and phylogenetic significance of 27-hydroxy-octacosanoic acid in lipopolysaccharides from bacteria belonging to the alpha-2 subgroup of Proteobacteria. Int. J. Syst. Bacteriol. 41:213-217.
Bhat, U. R., L. S. Forsberg, and R. W. Carlson. 1994. The structure of the lipid A component of Rhizobium leguminosarum bv. phaseoli lipopolysaccharide. A unique non-phosphorylated lipid A containing 2-amino-2-deoxy-gluconate, galacturonate, and glucosamine. J. Biol. Chem. 269:14402-14410.
Bhat, U. R., H. Mayer, A. Yokota, R. I. Hollingsworth, and R. W. Carlson. 1991. Occurrence of lipid A variants with 27-hydroxyoctacosanoic acid in lipopolysaccharides from the Rhizobiaceae group. J. Bacteriol. 173:2155-2159.
Braiterman, L. T., P. A. Watkins, A. B. Moser, and K. D. Smith. 1999. Peroxisomal very long chain fatty acid -oxidation activity is determined by the level of adrenodeukodystrophy protein (ALDP) expression. Mol. Genet. Metab. 66:91-99.
Brewin, N. J. 1991. Development of the legume root nodule. Annu. Rev. Cell Biol. 7:191-226.
Brewin, N. J. 1998. Tissue and cell invasion by Rhizobium: the structure and development of infection threads and symbiosomes, p. 415-429. In H. P. Spaink, A. Kondorosi, and P. J. J. Hooykaas (ed.), The Rhizobiaceae. Kluwer Academic Publishers, Dordrecht, The Netherlands.
Brewin, N. J., J. G. Robertson, E. A. Wood, B. Wells, A. P. Larkins, G. Galfre, and G. W. Butcher. 1985. Monoclonal antibodies to antigens in the peribaceroid membrane from Rhizobium-induced root nodules of pea cross-react with plasma membranes and Golgi bodies. EMBO J. 4:605-611.
Brewin, N. J., E. A. Wood, A. P. Larkins, G. Galfre, and G. W. Butcher. 1986. Analysis of lipopolysaccharide from root nodule bacteroids of Rhizobium leguminosarum using monoclonal antibodies. J. Gen. Microbiol. 132:1959-1968.
Carlson, R. W., B. L. Reuhs, L. S. Forsberg, and E. L. Kannenberg. 1999. Rhizobial cell surface carbohydrates: their structures, biosynthesis, and functions, p. 53-90. In J. B. Goldberg (ed.), Genetics of bacterial polysaccharides. Ann Arbor Press, Ann Arbor, MI.
Caroff, M., A. Tacken, and L. Szabo. 1988. Detergent-accelerated hydrolysis of bacterial endotoxins and determination of the anomeric configuration of the glycosyl phosphate present in the "isolated lipid A" fragment of the Bordetella pertussis endotoxin. Carbohydr. Res. 175:273-282.
Catalano, C. M., W. S. Lane, and D. J. Sherrier. 2004. Biochemical characterization of symbiosome membrane proteins from Medicago truncatula root nodules. Electrophoresis 25:519-531.
Duelli, D. M., and K. D. Noel. 1997. Compounds exuded by Phaseolus vulgaris that induce a modification of Rhizobium etli lipopolysaccharide. Mol. Plant-Microbe Interact. 10:903-910.
Ferguson, G. P., A. Datta, R. W. Carlson, and G. C. Walker. 2005. Importance of unusually modified lipid A in Sinorhizobium stress resistance and legume symbiosis. Mol. Microbiol. 56:68-80.
Ferguson, G. P., A. Datta, J. Baumgartner, R. M. Roop II, R. W. Carlson, and G. C. Walker. 2004. Similarity to peroxisomal-membrane protein family reveals that Sinorhizobium and Brucella BacA affect lipid-A fatty acids. Proc. Natl. Acad. Sci. USA 101:5012-5017.
Ferguson, G. P., R. M. Roop II, and G. C. Walker. 2002. Deficiency of a Sinorhizobium meliloti bacA mutant in alfalfa symbiosis correlates with alteration of the cell envelope. J. Bacteriol. 184:5625-5632.
Footitt, S., S. P. Slocombe, V. Larner, S. Kurup, Y. Wu, T. Larson, I. Graham, A. Baker, and M. Holdsworth. 2002. Control of germination and lipid mobilization by COMATOSE, the Arabidopsis homologue of human ALDP. EMBO J. 21:2912-2922.
Geiger, O., and I. M. Lopez-Lara. 2002. Rhizobial acyl carrier proteins and their roles in the formation of bacterial cell-surface components that are required for the development of nitrogen-fixing root nodules on legume roots. FEMS Microbiol. Lett. 208:153-162.
Jeyaretnam, B., J. Glushka, V. S. K. Kolli, and R. W. Carlson. 2002. Characterization of a novel lipid-A from Rhizobium species Sin-1. A unique lipid-A structure that is devoid of phosphate and has a glycosyl backbone consisting of glucosamine and 2-aminogluconic acid. J. Biol. Chem. 277:41802-41810.
Kannenberg, E. L., and N. J. Brewin. 1989. Expression of a cell surface antigen from Rhizobium leguminosarum 3841 is regulated by oxygen and pH. J. Bacteriol. 171:4543-4548.
Kannenberg, E. L., and R. W. Carlson. 2001. Lipid A and O-chain modifications cause Rhizobium lipopolysaccharides to become hydrophobic during bacteroid development. Mol. Microbiol. 39:379-392.
Kannenberg, E. L., B. L. Reuhs, L. S. Forsberg, and R. W. Carlson. 1998. Lipopolysaccharides and K-antigens: their structures, biosynthesis, and function, p. 119-154. In H. P. Spaink, A. Kondorosi, and P. J. J. Hooykaas (ed.), The Rhizobiaceae: molecular biology of model plant-associated bacteria. Kluwer Academic Publishers, Dordrecht, The Netherlands.
LeVier, K., R. W. Phillips, V. K. Grippe, R. M. Roop II, and G. C. Walker. 2000. Similar requirements of a plant symbiont and a mammalian pathogen for prolonged intracellular survival. Science 287:2492-2493.
Lullien, V., D. G. Barker, P. DelaJudie, and T. Hugeut. 1987. Plant gene expression in effective and ineffective root nodules of alfalfa (Medicago sativa). Plant Mol. Biol. 9:469-478.
Martinez, M., J. M. Palacios, J. Imperial, and T. Ruiz-Argueso. 2004. Symbiotic autoregulation of nifA expression in Rhizobium leguminosarum bv. viciae. J. Bacteriol. 186:6586-6594.
Noel, K. D., D. M. Duelli, H. Tao, and N. J. Brewin. 1996. Antigenic change in the lipopolysaccharide of Rhizobium etli CFN42 induced by exudates of Phaseolus vulgaris. Mol. Plant-Microbe Interact. 9:180-186.
Que, N. L. S., S. H. Lin, R. J. Cotter, and C. R. H. Raetz. 2000. Purification and mass spectrometry of six lipid A species from the bacterial endosymbiont Rhizobium etli—demonstration of a conserved distal unit and a variable proximal portion. J. Biol. Chem. 275:28006-28016.
Que, N. L. S., A. A. Ribeiro, and C. R. H. Raetz. 2000. Two-dimensional NMR spectroscopy and structures of six lipid A species from Rhizobium etli CE3—detection of an acyloxyacyl residue in each component and origin of the aminogluconate moiety. J. Biol. Chem. 275:28017-28027.
Roop, R. M., B. H. Bellaire, M. W. Valderas, and J. A. Cardelli. 2004. Adaptation of the brucellae to their intracellular niche. Mol. Microbiol. 52:621-630.
Sharypova, L. A., K. Niehaus, H. Scheidle, O. Holst, and A. Becker. 2003. Sinorhizobium meliloti acpXL mutant lacks the C28 hydroxylated fatty acid moiety of lipid A and does not express a slow migrating form of lipopolysaccharide. J. Biol. Chem. 278:12946-12954.
Tao, H., N. J. Brewin, and K. D. Noel. 1992. Rhizobium leguminosarum CFN42 lipopolysaccharide antigenic changes induced by environmental conditions. J. Bacteriol. 174:2222-2229.
VandenBosch, K. A., N. J. Brewin, and E. L. Kannenberg. 1989. Developmental regulation of a Rhizobium cell surface antigen during growth of pea root nodules. J. Bacteriol. 171:4537-4542.
Vedam, V., J. G. Haynes, E. L. Kannenberg, R. W. Carlson, and D. J. Sherrier. 2004. A Rhizobium leguminosarum lipopolysaccharide lipid A mutant induces nitrogen-fixing nodules with delayed and defective bacteroid formation. Mol. Plant-Microbe Interact. 17:283-291.
Vedam, V., E. L. Kannenberg, J. G. Haynes, D. J. Sherrier, A. Datta, and R. W. Carlson. 2003. A Rhizobium leguminosarum acpXL mutant produces lipopolysaccharide lacking 27-hydroxyoctacosanoic acid. J. Bacteriol. 185:1841-1850.
Verleur, N., E. H. Hettema, C. W. T. Roermund, H. F. Tabak, and R. J. A. Wanders. 1997. Transport of activated fatty acids by the peroxisomal ATP-binding-cassette transporter pxa2 in a semi-intact yeast cell system. Eur. J. Biochem. 249:657-661.
Wood, E. A., G. W. Butcher, N. J. Brewin, and E. L. Kannenberg. 1989. Genetic derepression of a developmentally regulated lipopolysaccharide antigen from Rhizobium leguminosarum 3841. J. Bacteriol. 171:4549-4555.
York, W. S., A. G. Darvill, M. McNeil, T. T. Stevenson, and P. Albersheim. 1985. Isolation and characterization of plant cell walls and cell wall components. Methods Enzymol. 118:3-40.
Zahringer, U., Y. A. Knirel, B. Lindner, J. H. Helbig, A. Sonnesson, R. Marre, and E. T. Rietschel. 1995. The lipopolysaccharide of Legionella pneumophila serogroup I (strain Philadelphia 1): chemical structure and biological significance. Prog. Clin. Biol. Res. 92:113-139.
Zahringer, U., B. Lindner, and E. T. Rietschel. 1999. Chemical structure of lipid A: recent advances in structural analysis of biologically active molecules, p. 93-114. In H. Brade, S. M. Opal, A. M. Vogel, and D. C. Morrison (ed.), Endotoxin in health and disease. Marcel Dekker, Inc., New York, NY.
Zolman, B. K., I. D. Silva, and B. Bartel. 2001. The Arabidopsis pxa1 mutant is defective in an ATP-binding cassette transporter-like protein required for peroxisomal fatty acid beta-oxidation. Plant Physiol. 127:1266-1278.(Vinata Vedam, Elmar Kanne)
ABSTRACT
Members of the Rhizobiaceae contain 27-hydroxyoctacosanoic acid (27OHC28:0) in their lipid A. A Rhizobium leguminosarum 3841 acpXL mutant (named here Rlv22) lacking a functional specialized acyl carrier lacked 27OHC28:0 in its lipid A, had altered growth and physiological properties (e.g., it was unable to grow in the presence of an elevated salt concentration [0.5% NaCl]), and formed irregularly shaped bacteroids, and the synchronous division of this mutant and the host plant-derived symbiosome membrane was disrupted. In spite of these defects, the mutant was able to persist within the root nodule cells and eventually form, albeit inefficiently, nitrogen-fixing bacteroids. This result suggested that while it is in a host root nodule, the mutant may have some mechanism by which it adapts to the loss of 27OHC28:0 from its lipid A. In order to further define the function of this fatty acyl residue, it was necessary to examine the lipid A isolated from mutant bacteroids. In this report we show that addition of 27OHC28:0 to the lipid A of Rlv22 lipopolysaccharides is partially restored in Rlv22 acpXL mutant bacteroids. We hypothesize that R. leguminosarum bv. viciae 3841 contains an alternate mechanism (e.g., another acp gene) for the synthesis of 27OHC28:0, which is activated when the bacteria are in the nodule environment, and that it is this alternative mechanism which functionally replaces acpXL and is responsible for the synthesis of 27OHC28:0-containing lipid A in the Rlv22 acpXL bacteroids.
INTRODUCTION
Rhizobium leguminosarum cells have an envelope similar to that of other gram-negative bacteria. Lipopolysaccharide (LPS) is the primary component of the bacterial outer leaflet and is comprised of three structural regions: the O-chain polysaccharide, the core oligosaccharide, and lipid A. The lipid A region is anchored in the bacterial outer membrane, and the carbohydrate portion projects from the outer surface into the surrounding milieu and is the primary immunogenic determinant. Correlating LPS structure with function has been difficult as LPS is a very complex molecule and LPS preparations consist of structurally heterogeneous mixtures of molecules. There are pronounced variations in LPS structure from strain to strain, and even within a strain there are different sizes and compositions of LPS (23). Interestingly, there are marked differences between LPSs from free-living cultures and LPSs from nitrogen-fixing bacteroids in terms of size, composition, and antigenic properties (10, 32, 33). It has been shown that the rhizobial LPS undergoes structural modifications during the formation of bacteroids and that there are composition differences between the bacterial and bacteroid LPSs (10, 32).
Variation in LPS structure due to environmental changes has been studied in cultured rhizobia by altering the growth conditions, such as lowering the oxygen level, lowering the pH, altering the carbon source, or adding plant-derived compounds (2, 14, 21, 23, 27). Such studies have shown that cues from the environment play an important role in LPS composition. The results led to the hypothesis that the bacterial LPS structure inside legume root nodules is probably controlled, to a large extent, by the in planta microenvironmental conditions. In the study of Kannenberg and Carlson (22), Rhizobium leguminosarum bv. viciae 3841 (named here Rlv3841) was cultured under various growth conditions, and the LPS structural modifications were analyzed chemically and immunochemically. It was observed that the LPS extracted from nodule bacteria or from laboratory-grown bacteria cultured under low-oxygen conditions was much more hydrophobic than the LPS from bacteria grown under normal laboratory conditions. Chemical analysis of the LPS derived from bacteria grown under low-oxygen conditions indicated that changes occurred in both the polysaccharide and lipid A portions of the LPS; the polysaccharide was affected in the extent of methylation and acetylation, while the lipid A showed an increase in a unique very-long-chain fatty acyl component, 27-hydroxyoctacosanoic acid (27OHC28:0). This suggested that the low-oxygen conditions in nodule cells may cause similar structural changes to the LPS.
The functions of LPS structural changes that occur during symbiosis are not known. These changes may be needed for the increased membrane stability and barrier properties that are required for the bacteroid to persist and function within the symbiosome compartment. One of the LPS components that may be essential in the symbiotic process is the very-long-chain fatty acyl component 27OHC28:0 in lipid A. This fatty acyl component is present in the LPS of members of the Rhizobiaceae (5, 11, 23). In addition, a number of facultative intracellular pathogenic bacterial species that cause chronic infections also contain this lipid A fatty acyl residue or orthologs of acpXL or lpxXL' genes that are required for its synthesis and transfer to lipid A. These species include Brucella abortus (3), Brucella melitensis (1), Legionella pneumophila (39, 40), and Bartonella henselae (3). Thus, it is possible that the 27OHC28:0 residue may be required for endosymbiotic rhizobia and these intracellular pathogens to persist and function within their host cells.
In order to examine the symbiotic function of the 27OHC28:0 lipid A component, we prepared and characterized an LPS mutant that is defective in the acyl carrier protein (ACP) required for its synthesis, AcpXL (34, 35). Laboratory-grown cultures of this mutant, Rlv22, produced an LPS that did not contain 27OHC28:0 in its lipid A and was unaffected in its O-chain polysaccharide and core oligosaccharide structures (35). The Rlv22 mutant was unable to grow under laboratory conditions at a low pH (pH 5.0) or with 0.5% NaCl added to the medium, and nodule development was delayed, although eventually nitrogen-fixing nodules were formed. Similar results were reported by Sharypova et al. for an acpXL mutant of Sinorhizobium meliloti (31). The ability of Rlv22 to form nitrogen-fixing nodules in spite of its inability to adjust to changes in osmotic strength or pH ex planta suggested that it may adapt in a specific manner to the in planta conditions (34, 35). Microscopic examination revealed that the Rlv22 mutant formed large irregularly shaped bacteroids and that multiple bacteroids were often surrounded by a single symbiosome membrane (34). However, it was also observed that a significant number of normal bacteroids were present, and, as previously observed, nitrogen-fixing nodules formed, although the level of nitrogenase in Rlv22-induced nodules was significantly lower than the level in normal nodules (34). Thus, it was concluded that depletion of 27OHC28:0 in the lipid A of the Rlv22 mutant disrupted bacteroid development and the synchrony between bacterial and symbiosome membrane division. However, the eventual appearance of some normal bacteroids and nitrogen-fixing nodules suggested that the mutant, once it was in its host, was able to partially compensate in some manner for the loss of 27OHC28:0 in its lipid A. Investigation of the nature of this compensation required isolation and characterization of the lipid A from mutant bacteroids. In this paper, we report the results of this analysis.
MATERIALS AND METHODS
Bacterial strains and culture conditions. The strains used are listed in Table 1. Strains were cultured on solid or in liquid tryptone yeast extract (TY) medium (Difco Laboratories), as previously described (35). Cultures were supplemented with kanamycin (Kan) (50 μg ml–1) and streptomycin (Str) (100 μg ml–1), when appropriate. The osmotic tolerance of bacterial isolates was studied by growing two strains on TY containing 0.5% NaCl at 30°C, along with the appropriate controls. The pH tolerance of the bacterial isolates was examined at pH 5, using piperazine-buffered TY plates as described by Vedam et al. (35).
Growth of peas. Peas (Pisum sativum cv. Early Alaska) were surface sterilized by soaking them in 95% ethanol for 30 s, rinsing them three times with sterile water, washing them with bleach for 3 min, and rinsing them 10 times with sterile water. Seeds were placed in sterile petri dishes with a small amount of sterile water and allowed to imbibe and germinate in the dark for 48 h. Pea seedlings were sown in "caissons" and supplied with Lullien's nutrient solution (25) lacking ammonium nitrate. Caissons are aeroponic growth systems that provide controlled conditions for the growth of peas. After 1 week of germination and stabilization of the seedlings, the peas were inoculated with either Rlv3841 (wild type) or Rlv22 (acpXL mutant). The pea plants were allowed to nodulate for 3 weeks, and the caissons were periodically refreshed with nutrient solution. At the end of the growth period, the nodules were harvested for preparation of bacteroids.
Extraction of bacteroids from pea nodules. Nodules were harvested from pea plants that had been inoculated with either Rlv3841 or Rlv22 into an ice-cold solution of 0.5 M sucrose in 50 mM Tris-HCl (pH 7.4) plus a 1:100 dilution of protease inhibitor cocktail (Sigma P9599). For (bio)chemical analysis, bacteroids were isolated from these nodules and purified by a procedure involving a sucrose step gradient, as described previously (13). Briefly, pea nodules were washed with cold Tris-HCl/sucrose buffer (0.5 M sucrose-50 mM Tris-HCl [pH 8.0] at 4°C containing dithiothreitol, proteinase inhibitor, and polyvinylpolypyrrolidone), suspended in the same buffer, and ground up using a mortar and pestle. The initial steps of the isolation protocol were done with sucrose-containing buffer to osmotically stabilize symbiosomes and bacteroids. To remove tissue and cell debris, the crushed nodule material was filtered through miracloth and rinsed with the same solution. The rinse solution was added to the filtrate, and the resulting suspension was centrifuged for 1 min at 10,000 x g. The pellet, containing the symbiosomes, was resuspended in Tris-HCl/sucrose buffer. Portions of the suspension were distributed in several microcentrifuge tubes, overlaid onto sucrose cushions (composed of 1.5 M sucrose and 50 mM Tris-HCl [pH 8.0] at 4°C), and centrifuged at 5,000 x g for 30 s. The top phases of the different tubes, strongly enriched in the bacteroid-containing symbiosomes, were transferred to a clean tube and centrifuged at 10,000 x g for 90 s to collect the symbiosome fraction in the pellet. The supernatant was discarded, and the pellet was resuspended in Tris-HCl/sucrose buffer and again overlaid onto a sucrose cushion as described above; however, this time the preparation was centrifuged at 10,000 x g for 5 min, which sedimented the symbiosomes with the bacteroids in the pellet. To remove the peribacteroid membrane by osmotic shock from the symbiosomes and wash the bacteroids, the pellets were repeatedly (two or three times) resuspended, with vigorous mixing, in 500 μl of Tris-HCl buffer without sucrose (0.50 mM Tris-HCl [pH 8.0] with dithiothreitol and proteinase inhibitor) and centrifuged; the final pellet was suspended in 1,600 μl of the same buffer and stored at –20°C.
During this work, individual colonies obtained from Rlv22-induced nodules were examined for antibiotic resistance and the presence of acpXL::kan. Resistance to Kan and Str was measured for about 500 colonies using growth on solid agar with and without the antibiotics. The presence of acpXL::kan (or acpXL) was measured for 12 of the colonies described above using PCR and primers for acpXL (GAGGGGGTTTAAATAGTCA and AGGCTTGGCCGCTTTGA), as previously described by Vedam et al. (35).
To rule out the possibility that a small number of possible mutant revertants escaped the analysis described above and occupied nodules undetected, a PCR screening analysis was performed directly with bacteroid preparations from nodules using the primers specific for acpXL described above in order to determine if any of the wild-type acpXL gene was present. Preparations of bacteroids isolated from nodules of 10 pea plants induced by Rlv3841 or Rlv22 were PCR screened. Aliquots of the bacteroid preparations (2 to 5 μl from an approximately 500-μl dense suspension of bacteroids) were pipetted into the PCR mixture and lysed during an initial 15-min hot start in the assay. The PCR products were analyzed using 1% agarose gel electrophoresis.
Preparation of bacterial isolates from Rlv22-induced nodules (Rlv22 EN isolates). In order to fully examine the effect of passage of Rlv22 through the plant, it was necessary to isolate and characterize in more detail mutant isolates obtained from pea root nodules. Rlv22 ex nodule (EN) isolates were analyzed for antibiotic resistance, the presence of acpXL::kan or acpXL, and sensitivity to growth at a low pH (pH 5.0) and in the presence of 0.5% NaCl, and their LPSs were extracted and analyzed as described below. Seeds were germinated and grown in Erlenmeyer flasks containing solid Fahraeus medium. After surface sterilization (see above) and transfer into the flasks, pea seeds were immediately inoculated with the Rlv22 mutant, and plant growth and nodulation were evaluated as described by Brewin et al. (9). Root nodules from several pea plants were removed, briefly washed in 95% ethanol, and, by size, surface sterilized in diluted bleach for 30 to 60 s. The nodules were transferred through a series of washes with sterile water and finally crushed to squeeze the occupying bacteria out onto TY agar plates. The plates were developed, and three single colonies from each plate were streaked onto fresh plates. Care was taken to pick colonies of different sizes (small, medium, and large) in case the size differences represented different variants of the original mutant. Two or three nodules per plant were analyzed in this way. From the Rlv22 EN colonies (kept on TY plates), a total of 23 clones (derived from four plants and eight nodules) were analyzed for antibiotic resistance (resistance to Str and Kan). A random selection of 16 clones (two clones per nodule) were tested further for sensitivity to salt (0.5% NaCl) and to an acidic pH (pH 5). As a control and for comparison, strains Rlv3841 and Rlv22 were included in the antibiotic, saline, and low-pH sensitivity tests. Since all Rlv22 EN isolates behaved similarly in these tests, two random Rlv22 EN isolates, EN2 and EN4, were selected, and we verified that they contained the acpXL::kan mutation using PCR and the primers for acpXL as previously described by Vedam et al. (35).
Lipid A purification. LPSs were extracted by the triethylamine (TEA)/EDTA/phenol procedure as previously described (25). Briefly, for each strain, the LPS was extracted from the bacterial or bacteroid pellet using 3 volumes of TEA/EDTA/phenol (0.25 m EDTA, 5% phenol, titrated to pH 6.9 with TEA with constant stirring for 1 h at 37°C). The extract was centrifuged at 13,000 x g for 1 h, and the supernatant was collected and dialyzed (molecular weight cutoff, 10,000; Spectrapor) against deionized water. The bacterial and bacteroid LPSs were lyophilized for analysis. Lipid A was isolated from the LPS preparations by mild acid hydrolysis (12). Briefly, the LPS was dissolved in 1% sodium dodecyl sulfate in 20 mM sodium acetate, the pH was adjusted to 4.5 with 4 M HCl, and then the preparation was placed in an ultrasound bath until the sample was dissolved. The solution was then heated to 100°C for 1 h, which was followed by lyophilization. The sodium dodecyl sulfate was removed by washing the lyophilized residue with a 2:1 solution of deionized H2O and acidified ethanol (100 μl of 4 M HCl in 20 ml of 95% ethanol). The residue was collected by centrifugation, washed with 95% ethanol (nonacidified), and collected by centrifugation (200 x g for 15 min). The washing and centrifugation steps were repeated. Finally, the residue was lyophilized to obtain a white, solid, fluffy lipid A preparation.
Analytical procedures. The composition was determined by gas chromatography-mass spectrometry (MS) analysis of trimethylsilyl derivatives of methyl glycosides and hydroxy fatty acid methyl esters as described previously (4, 38). Matrix-assisted laser desorption ionization—time of flight (MALDI-TOF) MS was performed in the negative-ion reflectron mode with a 337-nm nitrogen laser operating at an extraction voltage of 20 kV 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/ml of methanol) and applied to the probe for mass analysis. Spectra were calibrated externally using E. coli lipid A (Sigma).
RESULTS
Stability of the acpXL mutation in Rlv22 during nodule development. In order to rule out the possibility of genetic reversion of the acpXL mutation in planta, bacterial colonies obtained from nodules infected by Rlv22 were tested for resistance to kanamycin by using an antibiotic marker cassette introduced into acpXL (35). Approximately 500 colonies were screened, and there was no loss of the resistance marker; i.e., all 500 colonies were resistant to both Kan and Str. In addition, genomic DNA from 12 of these colonies was tested by PCR, using primers for acpXL. Each of the 12 colonies showed only the presence of acpXL::kan, indicating that the mutation was intact. Furthermore, chromosomal DNA from two of these colonies was analyzed by Southern hybridization using amplified acpXL as the labeled probe, and the results showed that both colonies retained the mutation. In addition, bacteroids obtained from Rlv22-induced nodules were directly examined by PCR using primers for acpXL. As a control, PCR analysis was also performed with bacteroids from Rlv3841-induced nodules. The results (Fig. 1) showed that bacteroids from Rlv22-induced nodules contained only acpXL::kan and that there was no detectable evidence of intact acpXL. This finding confirmed the stability of the acpXL mutation of Rlv22 during nodule development.
Analysis of the lipid A from Rlv22 bacteroids. The lipid As from LPS preparations of parent Rlv3841 and mutant Rlv22 laboratory-grown bacteria and from bacteroids isolated from root nodules were analyzed to determine their fatty acid compositions (Table 2) and by MALDI-TOF MS (Fig. 2).
The composition analysis revealed that the lipid A from Rlv3841 laboratory-grown cells contained the same fatty acids at the same ratio that was observed for the lipid A from Rlv3841 bacteroids. In contrast, the Rlv22 laboratory-grown cells produced lipid A in which no long-chain 27OHC28:0 could be detected. However, the lipid A from Rlv22 bacteroids contained 27OHC28:0 at a level that was about 50% of the level observed in the Rlv3841 parent lipid A. In addition, the Rlv22 lipid A preparations contained noticeably increased levels of palmitate (C16:0) and stearate (C18:0) compared to the levels in the Rlv3841 lipid A samples.
The mass spectrum of the lipid A from laboratory-grown Rlv3841 (Fig. 2A) shows that there were two clusters of ions. The first cluster of ions ranged from m/z 1887.6 to 2058.0, with the most intense ion at m/z 1914.0, and the other cluster of ions ranged from m/z 1625.7 to 1738.9, with the most intense ion at m/z 1652.0. The structures corresponding to the lipid A ions for the laboratory-grown Rlv3841 (structures I and II) have been reported previously (4, 35) and are shown in Fig. 2. The ion at m/z 1914.0 is consistent with the previously published R. leguminosarum or Rhizobium etli lipid A structure (4, 28, 29, 35), in which the lipid A has a disaccharide backbone consisting of a distal glucosaminosyl residue that is -1,6 linked to a proximal 2-aminogluconate (GlcNonate) residue. At the 4' position of the distal glucosamine there is an -galacturonosyl residue substitution, and the -glucosaminosyl-(16)-GlcNonate disaccharide is acylated with -hydroxy fatty acids at the 2, 3, 2', and 3' positions. The 27OHC28:0 lipid A moiety is present as a secondary acyloxyacyl residue and is ester linked to the hydroxy group of the 3'--hydroxy fatty acid residue. The ions at m/z 2001.4 and above are due to molecules in which the 27-OH group of the 27OHC28:0 molecule is esterified with -hydroxybutyrate. Other ions in this cluster are due to structural variants resulting from different fatty acyl chain lengths. In the second ion cluster, the most intense ion is the ion at m/z 1652.0 and is likely due to a structure (structure II) caused by elimination of the -hydroxy fatty acyl residue from position 3 of the GlcNonate residue forming a 2,3-unsaturated 2-aminoglucono-1,5-lactone residue; this is a reaction which may be an artifact of the lipid A isolation procedure (20, 35).
The mass spectrum of the lipid A preparation from laboratory-grown mutant Rlv22 (Fig. 2C) also shows that there were two main ion clusters, one centered around the ion at m/z 1493.0 and the other centered around the ion at m/z 1230.0. A third minor cluster is centered around the ion at m/z 1758.0. The two main ion clusters represent structures that are devoid of 27OHC28:0 or 27O(-hydroxybutyryl)C28:0 (structures III and IV). These structures are identical to those previously reported for the mutant lipid A (35). We also reported previously (35) that the minor cluster of ions centered around m/z 1758.0 was due to replacement of 27OHC28:0 with a palmitoyl (C16:0) residue. These ions were observed again in the current study, and this result, together with the composition data (Table 2), could have been due to structures in which 27OHC28:0 was replaced with a C16:0 or stearoyl (C18:0) residue.
Figures 2B and 2D show the mass spectra of the lipid A preparations from Rlv3841 and Rlv22 bacteroids, respectively. The mass spectrum of the Rlv3841 bacteroid lipid A preparation (Fig. 2B) is identical to that of the lipid A from the laboratory-grown culture (Fig. 2A) described above. The spectrum of Rlv22 bacteroid lipid A has four major ion clusters (Fig. 2D). Three of these ion clusters are identical to those observed for the lipid A from laboratory-grown Rlv22 (Fig. 2C). The fourth ion cluster is centered around the m/z 1914.0 ion and is identical to the second ion cluster for the lipid As from both the Rlv3841 bacteroids and laboratory-cultured cells. This result, together with the composition data described above, shows that the 27OHC28:0 moiety was present in some of the structures in the mutant bacteroid lipid A preparation. Also present in the Rlv22 bacteroid lipid A preparation were ions that were consistent with structures that lack 27OHC28:0 but contain an added C16:0 or C18:0 residue. In summary, the lipid A preparation from the Rlv3841 bacteroids appeared to have the same structures as the preparation from the Rlv3841 laboratory-grown cells. In contrast, Rlv22 laboratory-grown cells lacked 27OHC28:0 in their lipid A (Fig. 3C), but bacteroids from Rlv22-induced nodules produced some lipid A molecules that contained 27OHC28:0 (Fig. 3D).
Properties of EN acpXL mutant isolates from Rlv22-infected pea plants. In order to determine if passage through the plant had selected for an Rlv22 variant that constitutively expressed an alternative mechanism for synthesis of 27OHC28:0, we isolated and examined a number of ex nodule mutant isolates. We showed previously that the Rlv22 acpXL::kan mutant was unable to grow in the presence of 0.5% NaCl or at pH 5.0 (35). Sixteen Rlv22 EN isolates were grown under the same test conditions as the original mutant. We observed that all of the Rlv22 EN isolates were able to grow in the presence of 0.5% NaCl (Fig. 3). On plates with acidic growth conditions (pH 5), the Rlv22 EN isolates showed stronger growth than the original Rlv22 mutant, although the growth was not as good as the growth of wild-type strain Rlv3841 in the same experiment. PCR analysis to detect either acpXL or acpXL::kan was performed with the genomic DNA from two Rlv22 EN isolates, EN2 and EN4, using primers specific for acpXL. The results were the same as those obtained for the bacteroid preparations shown in Fig. 1 and showed that the PCR products obtained from the Rlv22 EN isolates were each 1.56 kb long, which is consistent with the size of acpXL::kan, and were the same size as the PCR product from the original Rlv22 strain. A PCR product indicating that there was an intact acpXL gene was not detected in either the EN2 or EN4 isolate.
In order to determine the presence of 27OHC28:0 in the lipid A from the two Rlv22 EN mutant isolates, lipid A was prepared from these two strains grown under normal laboratory conditions or in the presence of 0.5% NaCl. The lipid A samples were then analyzed by MALDI-TOF MS to determine their fatty acyl compositions. The original Rlv22 mutant was also included in this study as a control. The results are shown in Fig. 4. The mass spectrum of the lipid A from the original Rlv22 mutant (Fig. 4A) was the same as the mass spectrum described above and shown in Fig. 2C. The mass spectra of the lipid As from EN2 and EN4 grown under normal laboratory conditions (Fig. 4B) or in the presence of 0.5% NaCl were identical to one another and very similar to the mass spectrum of the lipid A from the original Rlv22 mutant in that all of the mass spectra lacked the 27OHC28:0 residue. The minor difference between the spectra of EN2 and EN4 lipid A preparations and the spectrum of the Rlv22 lipid A was that the third minor ion cluster, likely due to replacement of 27OHC28:0 with a palmitoyl or stearoyl residue, was significantly less intense in the EN2 and EN4 lipid A preparations.
DISCUSSION
We previously showed that R. leguminosarum acpXL mutant Rlv22 is unable to add 27OHC28:0 to its lipid A, exhibits delayed nodulation and nitrogen fixation, forms many aberrantly shaped bacteroids, exhibits disrupted synchrony of bacterium and symbiosome membrane division, and, under laboratory culture conditions, is unable to grow at a low pH (pH 5.0) or in the presence of 0.5% NaCl (35). Sharypova et al. (31) showed that an acpXL mutant of S. meliloti also lacked 27OHC28:0 in its lipid A, was sensitive to deoxycholate, exhibited delayed nodulation of Medicago sativa, and had reduced competitive ability. However, nodules elicited by this mutant on roots of M. sativa and Medicago truncatula had normal morphology and fixed nitrogen. Sharypova et al. did not describe an effect of the acpXL mutation on S. meliloti bacteroid development. The aberrant bacteroids that we observed for the Rlv22 mutant (34) support the idea that the 27OHC28:0 moiety of lipid A plays an important role in bacteroid development and in the synchronous division of bacteria and the symbiosome membrane. However, the ability of the mutant to eventually form nitrogen-fixing nodules in a process in which it is likely to encounter acidic pH conditions and changes in osmolarity raised the question of whether Rlv22 adapts in some manner to the loss of 27OHC28:0 within the host plant. In this study, we addressed this question by examining the lipid A isolated from Rlv22 bacteroids and from ex nodule acpXL mutant isolates EN2 and EN4 grown under laboratory conditions at normal NaCl concentrations and in the presence of 0.5% NaCl. We found that, unlike laboratory cultures, Rlv22 bacteroids are able to add 27OHC28:0 to their lipid A. Furthermore, we observed that the acpXL::kan mutation was maintained in Rlv22 bacteroids and in the EN isolates of Rlv22 and that, when grown under laboratory conditions, the EN isolates were still unable to add 27OHC28:0 to their lipid A. These results indicate that in Rlv3841 an alternative mechanism is induced within the nodule that can functionally replace the inactivated AcpXL in the acpXL::kan mutant. The presence of an alternative to acpXL for the synthesis of 27OHC28:0 that is induced within the host suggests that this fatty acyl residue in lipid A may be essential for the symbiotic interaction between R. leguminosarum and peas.
As mentioned above and consistent with our results with Rlv3841, recent work with S. meliloti also showed that an acpXL mutant lacked the 27OHC28:0 residue in its lipid A (31). A subsequent study (15) demonstrated that the S. meliloti acpXL mutant contained another lipid A species, in which a C18:0 residue replaced the 27OHC28:0, a result that is consistent with our results (21; this study), which shows that the Rlv22 mutant can produce a lipid A molecule in which the 27OHC28:0 moiety is replaced by a C16:0 or C18:0 residue. Since replacement of 27OHC28:0 by C18:0 in the S. meliloti acpXL mutant was prevented by a second mutation in the gene encoding the 27OHC28:0 acyltransferase (lpxXL), it was concluded that LpxXL is able to transfer either C18:0 or 27OHC28:0 to the acyloxyacyl position of S. meliloti lipid A (15). This report (15) also showed that the S. meliloti acpXL mutant, an lpxXL mutant, and an lpxXL acpXL double mutant were able to form, in a delayed manner, nitrogen-fixing nodules. The latter two mutants were particularly sensitive to stressful conditions, such as detergents or increased salt concentrations. While these results also suggest that in S. meliloti there may be alternative mechanisms that are activated within the host that can functionally replace the mutated acpXL, as well as lpxXL, there has been no analysis of the lipid A from S. meliloti acpXL or lpxXL mutant bacteroids.
The nature of this alternative host-induced mechanism for synthesizing 27OHC28:0 is unknown. However, it has been reported that S. meliloti possesses multiple ACPs. Apart from the four known major ACPs in rhizobia, genomics has predicted the existence of additional ACPs (19). The complete sequence of S. meliloti indicates that there are at least two novel ACPs. One of the ACP genes is located on the Sym plasmid (it is located in a cluster of four genes, all closely linked, perhaps belonging to one operon), and the other is located on the chromosome. The Mesorhizobium loti genome also contains an operon similar to the operon identified in S. meliloti (19). Since rhizobia possess multiple ACPs, it is possible that R. leguminosarum contains an additional ACP that could be activated in planta, thereby compensating for the disrupted acpXL in Rlv22 during symbiosis. In this regard we have located in the Rlv3841 genome sequence (www.sanger.ac.uk), which has not been fully annotated yet, two such possible acp gene candidates. One acp candidate is located on the chromosome, and the second is located on the symbiotic plasmid, pRL10. These two putative acp genes encode ACPs that are very similar to one another and to ACPs from Agrobacterium and a number of Burkholderia strains, particularly strains of the pathogens Burkholderia mallei and Burkholderia pseudomallei. The Rlv3841 ACP gene region in pRL10 is shown in Fig. 5. The acp gene is next to a gene encoding an acyl-ACP dehydrogenase. It is preceded by DNA sequence binding motifs for NifA and for alternative sigma factor 54 in a DNA region that exhibits sequence similarity to a region that precedes the fixWABC genes in R. leguminosarum UPM791 (26). This location suggests that expression of the acp gene may be regulated by the O2 status of the bacterial cell. The possibility that the other ACPs may be involved in the in planta ability of the acpXL::kan mutant to synthesize 27OHC28:0 is under investigation.
27OHC28:0 in the lipid A or orthologs of acpXL or lpxXL have been found in a number of bacterial species, including several pathogens that cause chronic intracellular infections. In all cases in which the gene sequences can be compared, there is a very similar arrangement of genes encoding fatty acyl-ACP synthase orthologs (three or four genes) flanked by acpXL on one side and by lpxXL on the other side. The arrangement of this gene region is shown for several bacterial species in Fig. 6. There is some variability in this gene region; some species do not contain the putative fatty acid ACP dehydratase (Orf1) gene, while others do not contain the dehydrogenase (Orf4) gene. In all cases, genes encoding proteins with high levels of similarity to AcpXL, LpxXL, Orf2 (a putative 3-oxoacyl ACP synthase), and Orf3 (a putative second 3-oxoacyl ACP synthase) are present. The only exception to the high level of sequence similarity in this gene region is the region from Legionella pneumophila. However, the arrangement of the genes in L. pneumophila is identical to the arrangement of this region in the other bacterial species shown. Therefore, it is highly likely that this gene region in all of these bacterial species consists of a cassette of five to six genes that are specifically required for the synthesis of very-long-chain -hydroxy or -oxo fatty acids and their transfer to lipid A.
In the case of the brucellae, their commonality with rhizobia is that they are both endocytosed into their host cells and are able to persist and function within host-derived membrane compartments (symbiosomes and phagosomes, respectively) (7, 8, 30). Thus, it is possible that the 27OHC28:0 lipid A component in both of these taxa plays an important role in their virulence and persistence within their hosts. Recently, another connection between Brucella abortus and rhizobia has been reported with regard to a protein called BacA and its importance for producing lipid A with 27OHC28:0 (16). The BacA protein was found to be necessary for infection of alfalfa by S. meliloti (17) and was also shown to be present in B. abortus and necessary for this pathogen to maintain a chronic infection in its host (24). The BacA protein exhibits sequence similarity with a peroxisomal membrane protein family (16) that is thought to play a role in the transport of long-chain fatty acids (6, 18, 36, 41). BacA mutants of S. meliloti and B. abortus both produce lipid A preparations in which the level of 27OHC28:0 is reduced to about 50% of the normal level (16). Unlike AcpXL, which is specifically required for the synthesis of 27-OHC28:0, it is likely that the BacA protein has multiple functions that affect symbiotic infection, including a function which affects the level of 27-OHC28:0 in lipid A and another function which is essential for infection.
In summary, we showed that the acpXL mutation is partially suppressed by a possible host-activated alternative mechanism for the synthesis of 27OHC28:0. This result suggests that the presence of 27OHC28:0 in rhizobial lipid A is essential for symbiosis. Further investigation is in progress to (i) identify the alternative mechanism for 27OHC28:0 synthesis that occurs in planta and (ii) prepare and characterize the symbiotic phenotypes of mutants (e.g., deletion of the entire acpXL-lpxXL region) that are unable to synthesize 27OHC28:0 both ex planta and in planta. In addition to increasing our understanding of the molecular basis of the Rhizobium-legume symbiosis, determining the role of 27OHC28:0 in the symbiotic process should also provide information regarding the virulence mechanism of the brucellae and possibly several other pathogen species that cause chronic intracellular infections.
ACKNOWLEDGMENTS
We thank Thomas Hrtner (University of Tubingen) for his help with isolating ex nodule rhizobia.
REFERENCES
Basu, S. S., M. J. Karbarz, and C. R. H. Raetz. 2002. Expression, cloning and characterization of the C28 acyltranferase of lipid A biosynthesis in Rhizobium leguminosarum. J. Biol. Chem. 277:28959-28971.
Bhat, U. R., and R. W. Carlson. 1992. Chemical characterization of pH-dependent structural epitopes of lipopolysaccharides from Rhizobium leguminosarum biovar phaseoli. J. Bacteriol. 174:2230-2235.
Bhat, U. R., R. W. Carlson, M. Busch, and H. Mayer. 1991. Distribution and phylogenetic significance of 27-hydroxy-octacosanoic acid in lipopolysaccharides from bacteria belonging to the alpha-2 subgroup of Proteobacteria. Int. J. Syst. Bacteriol. 41:213-217.
Bhat, U. R., L. S. Forsberg, and R. W. Carlson. 1994. The structure of the lipid A component of Rhizobium leguminosarum bv. phaseoli lipopolysaccharide. A unique non-phosphorylated lipid A containing 2-amino-2-deoxy-gluconate, galacturonate, and glucosamine. J. Biol. Chem. 269:14402-14410.
Bhat, U. R., H. Mayer, A. Yokota, R. I. Hollingsworth, and R. W. Carlson. 1991. Occurrence of lipid A variants with 27-hydroxyoctacosanoic acid in lipopolysaccharides from the Rhizobiaceae group. J. Bacteriol. 173:2155-2159.
Braiterman, L. T., P. A. Watkins, A. B. Moser, and K. D. Smith. 1999. Peroxisomal very long chain fatty acid -oxidation activity is determined by the level of adrenodeukodystrophy protein (ALDP) expression. Mol. Genet. Metab. 66:91-99.
Brewin, N. J. 1991. Development of the legume root nodule. Annu. Rev. Cell Biol. 7:191-226.
Brewin, N. J. 1998. Tissue and cell invasion by Rhizobium: the structure and development of infection threads and symbiosomes, p. 415-429. In H. P. Spaink, A. Kondorosi, and P. J. J. Hooykaas (ed.), The Rhizobiaceae. Kluwer Academic Publishers, Dordrecht, The Netherlands.
Brewin, N. J., J. G. Robertson, E. A. Wood, B. Wells, A. P. Larkins, G. Galfre, and G. W. Butcher. 1985. Monoclonal antibodies to antigens in the peribaceroid membrane from Rhizobium-induced root nodules of pea cross-react with plasma membranes and Golgi bodies. EMBO J. 4:605-611.
Brewin, N. J., E. A. Wood, A. P. Larkins, G. Galfre, and G. W. Butcher. 1986. Analysis of lipopolysaccharide from root nodule bacteroids of Rhizobium leguminosarum using monoclonal antibodies. J. Gen. Microbiol. 132:1959-1968.
Carlson, R. W., B. L. Reuhs, L. S. Forsberg, and E. L. Kannenberg. 1999. Rhizobial cell surface carbohydrates: their structures, biosynthesis, and functions, p. 53-90. In J. B. Goldberg (ed.), Genetics of bacterial polysaccharides. Ann Arbor Press, Ann Arbor, MI.
Caroff, M., A. Tacken, and L. Szabo. 1988. Detergent-accelerated hydrolysis of bacterial endotoxins and determination of the anomeric configuration of the glycosyl phosphate present in the "isolated lipid A" fragment of the Bordetella pertussis endotoxin. Carbohydr. Res. 175:273-282.
Catalano, C. M., W. S. Lane, and D. J. Sherrier. 2004. Biochemical characterization of symbiosome membrane proteins from Medicago truncatula root nodules. Electrophoresis 25:519-531.
Duelli, D. M., and K. D. Noel. 1997. Compounds exuded by Phaseolus vulgaris that induce a modification of Rhizobium etli lipopolysaccharide. Mol. Plant-Microbe Interact. 10:903-910.
Ferguson, G. P., A. Datta, R. W. Carlson, and G. C. Walker. 2005. Importance of unusually modified lipid A in Sinorhizobium stress resistance and legume symbiosis. Mol. Microbiol. 56:68-80.
Ferguson, G. P., A. Datta, J. Baumgartner, R. M. Roop II, R. W. Carlson, and G. C. Walker. 2004. Similarity to peroxisomal-membrane protein family reveals that Sinorhizobium and Brucella BacA affect lipid-A fatty acids. Proc. Natl. Acad. Sci. USA 101:5012-5017.
Ferguson, G. P., R. M. Roop II, and G. C. Walker. 2002. Deficiency of a Sinorhizobium meliloti bacA mutant in alfalfa symbiosis correlates with alteration of the cell envelope. J. Bacteriol. 184:5625-5632.
Footitt, S., S. P. Slocombe, V. Larner, S. Kurup, Y. Wu, T. Larson, I. Graham, A. Baker, and M. Holdsworth. 2002. Control of germination and lipid mobilization by COMATOSE, the Arabidopsis homologue of human ALDP. EMBO J. 21:2912-2922.
Geiger, O., and I. M. Lopez-Lara. 2002. Rhizobial acyl carrier proteins and their roles in the formation of bacterial cell-surface components that are required for the development of nitrogen-fixing root nodules on legume roots. FEMS Microbiol. Lett. 208:153-162.
Jeyaretnam, B., J. Glushka, V. S. K. Kolli, and R. W. Carlson. 2002. Characterization of a novel lipid-A from Rhizobium species Sin-1. A unique lipid-A structure that is devoid of phosphate and has a glycosyl backbone consisting of glucosamine and 2-aminogluconic acid. J. Biol. Chem. 277:41802-41810.
Kannenberg, E. L., and N. J. Brewin. 1989. Expression of a cell surface antigen from Rhizobium leguminosarum 3841 is regulated by oxygen and pH. J. Bacteriol. 171:4543-4548.
Kannenberg, E. L., and R. W. Carlson. 2001. Lipid A and O-chain modifications cause Rhizobium lipopolysaccharides to become hydrophobic during bacteroid development. Mol. Microbiol. 39:379-392.
Kannenberg, E. L., B. L. Reuhs, L. S. Forsberg, and R. W. Carlson. 1998. Lipopolysaccharides and K-antigens: their structures, biosynthesis, and function, p. 119-154. In H. P. Spaink, A. Kondorosi, and P. J. J. Hooykaas (ed.), The Rhizobiaceae: molecular biology of model plant-associated bacteria. Kluwer Academic Publishers, Dordrecht, The Netherlands.
LeVier, K., R. W. Phillips, V. K. Grippe, R. M. Roop II, and G. C. Walker. 2000. Similar requirements of a plant symbiont and a mammalian pathogen for prolonged intracellular survival. Science 287:2492-2493.
Lullien, V., D. G. Barker, P. DelaJudie, and T. Hugeut. 1987. Plant gene expression in effective and ineffective root nodules of alfalfa (Medicago sativa). Plant Mol. Biol. 9:469-478.
Martinez, M., J. M. Palacios, J. Imperial, and T. Ruiz-Argueso. 2004. Symbiotic autoregulation of nifA expression in Rhizobium leguminosarum bv. viciae. J. Bacteriol. 186:6586-6594.
Noel, K. D., D. M. Duelli, H. Tao, and N. J. Brewin. 1996. Antigenic change in the lipopolysaccharide of Rhizobium etli CFN42 induced by exudates of Phaseolus vulgaris. Mol. Plant-Microbe Interact. 9:180-186.
Que, N. L. S., S. H. Lin, R. J. Cotter, and C. R. H. Raetz. 2000. Purification and mass spectrometry of six lipid A species from the bacterial endosymbiont Rhizobium etli—demonstration of a conserved distal unit and a variable proximal portion. J. Biol. Chem. 275:28006-28016.
Que, N. L. S., A. A. Ribeiro, and C. R. H. Raetz. 2000. Two-dimensional NMR spectroscopy and structures of six lipid A species from Rhizobium etli CE3—detection of an acyloxyacyl residue in each component and origin of the aminogluconate moiety. J. Biol. Chem. 275:28017-28027.
Roop, R. M., B. H. Bellaire, M. W. Valderas, and J. A. Cardelli. 2004. Adaptation of the brucellae to their intracellular niche. Mol. Microbiol. 52:621-630.
Sharypova, L. A., K. Niehaus, H. Scheidle, O. Holst, and A. Becker. 2003. Sinorhizobium meliloti acpXL mutant lacks the C28 hydroxylated fatty acid moiety of lipid A and does not express a slow migrating form of lipopolysaccharide. J. Biol. Chem. 278:12946-12954.
Tao, H., N. J. Brewin, and K. D. Noel. 1992. Rhizobium leguminosarum CFN42 lipopolysaccharide antigenic changes induced by environmental conditions. J. Bacteriol. 174:2222-2229.
VandenBosch, K. A., N. J. Brewin, and E. L. Kannenberg. 1989. Developmental regulation of a Rhizobium cell surface antigen during growth of pea root nodules. J. Bacteriol. 171:4537-4542.
Vedam, V., J. G. Haynes, E. L. Kannenberg, R. W. Carlson, and D. J. Sherrier. 2004. A Rhizobium leguminosarum lipopolysaccharide lipid A mutant induces nitrogen-fixing nodules with delayed and defective bacteroid formation. Mol. Plant-Microbe Interact. 17:283-291.
Vedam, V., E. L. Kannenberg, J. G. Haynes, D. J. Sherrier, A. Datta, and R. W. Carlson. 2003. A Rhizobium leguminosarum acpXL mutant produces lipopolysaccharide lacking 27-hydroxyoctacosanoic acid. J. Bacteriol. 185:1841-1850.
Verleur, N., E. H. Hettema, C. W. T. Roermund, H. F. Tabak, and R. J. A. Wanders. 1997. Transport of activated fatty acids by the peroxisomal ATP-binding-cassette transporter pxa2 in a semi-intact yeast cell system. Eur. J. Biochem. 249:657-661.
Wood, E. A., G. W. Butcher, N. J. Brewin, and E. L. Kannenberg. 1989. Genetic derepression of a developmentally regulated lipopolysaccharide antigen from Rhizobium leguminosarum 3841. J. Bacteriol. 171:4549-4555.
York, W. S., A. G. Darvill, M. McNeil, T. T. Stevenson, and P. Albersheim. 1985. Isolation and characterization of plant cell walls and cell wall components. Methods Enzymol. 118:3-40.
Zahringer, U., Y. A. Knirel, B. Lindner, J. H. Helbig, A. Sonnesson, R. Marre, and E. T. Rietschel. 1995. The lipopolysaccharide of Legionella pneumophila serogroup I (strain Philadelphia 1): chemical structure and biological significance. Prog. Clin. Biol. Res. 92:113-139.
Zahringer, U., B. Lindner, and E. T. Rietschel. 1999. Chemical structure of lipid A: recent advances in structural analysis of biologically active molecules, p. 93-114. In H. Brade, S. M. Opal, A. M. Vogel, and D. C. Morrison (ed.), Endotoxin in health and disease. Marcel Dekker, Inc., New York, NY.
Zolman, B. K., I. D. Silva, and B. Bartel. 2001. The Arabidopsis pxa1 mutant is defective in an ATP-binding cassette transporter-like protein required for peroxisomal fatty acid beta-oxidation. Plant Physiol. 127:1266-1278.(Vinata Vedam, Elmar Kanne)