Skip to main content
PMC home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2001 May 1;98(10):5910–5915. doi: 10.1073/pnas.101037998

Two enzymes of diacylglyceryl-O-4′-(N,N,N,-trimethyl)homoserine biosynthesis are encoded by btaA and btaB in the purple bacterium Rhodobacter sphaeroides

Rouven M Klug 1, Christoph Benning 1,*
PMCID: PMC33312  PMID: 11331765

Abstract

Betaine lipids are ether-linked, nonphosphorous glycerolipids that resemble the more commonly known phosphatidylcholine in overall structure. Betaine lipids are abundant in many eukaryotes such as nonseed plants, algae, fungi, and amoeba. Some of these organisms are entirely devoid of phosphatidylcholine and, instead, contain a betaine lipid such as diacylglyceryl-O-4′-(N,N,N,-trimethyl)homoserine. Recently, this lipid also was discovered in the photosynthetic purple bacterium Rhodobacter sphaeroides where it seems to replace phosphatidylcholine under phosphate-limiting growth conditions. This discovery provided the opportunity to study the biosynthesis of betaine lipids in a bacterial model system. Mutants of R. sphaeroides deficient in the biosynthesis of the betaine lipid were isolated, and two genes essential for this process, btaA and btaB, were identified. It is proposed that btaA encodes an S-adenosylmethionine:diacylglycerol 3-amino-3-carboxypropyl transferase and btaB an S-adenosylmethionine-dependent N-methyltransferase. Both enzymatic activities can account for all reactions of betaine lipid head group biosynthesis. Because the equivalent reactions have been proposed for different eukaryotes, it seems likely that orthologs of btaA/btaB may be present in other betaine lipid-containing organisms.

Polar lipids are essential components of all biological membranes. Most common are glycerolipids containing a diacylglycerol moiety to which a polar head group is attached. A head group can be a carbohydrate moiety as in the very abundant plant galactolipids or a phosphorylester as in the glycerophospholipids, the most common lipid class in animals. Betaine lipids represent a third class of glycerolipids in which a quaternary amine alcohol is bound in an ether linkage to the diacylglycerol moiety (1, 2). The overall structure of betaine lipids (Fig. 1) resembles to some extent that of the glycerophospholipid phosphatidylcholine (PC). Although the phase transition temperature for betaine lipid was found to be slightly higher compared with PC with identical fatty acid composition, the physical phase behavior of betaine lipid in mixtures with water is similar to that of PC (3). The betaine lipid diacylglyceryl-O-4′-(N,N,N,-trimethyl)homoserine (DGTS) (4) and a closely related isoform diacylglyceryl-O-2′-(hydroxymethyl)(N,N,N-trimethyl)-β-alanine were discovered in the unicellular alga (Chrysophyceae) Ochromonas danica. Soon after, it became apparent that betaine lipids are abundant and widespread in nonseed plants and alga (5–13). Furthermore, they are common in fungi including edible mushrooms and human pathogens (14–16) and are also present in amoebae (17). Although betaine lipids are abundant in primitive vascular plants such as ferns, they are absent from all tested seed plants (11, 14). Recently, the betaine lipid DGTS also has been discovered in the photosynthetic purple bacterium Rhodobacter sphaeroides (18) and the plant-nodule-forming bacterium Sinorhizobium meliloti (19). It has been pointed out repeatedly (2, 14) that there is a peculiar inverse relationship between the presence of betaine lipids and PC in different organisms, indicating that betaine lipids may substitute for PC. This hypothesis is supported by the accumulation of DGTS in the two bacteria mentioned above only when these are grown under phosphate-limiting conditions, under which the relative amounts of glycerophospholipids including PC are decreased. In S. meliloti the regulator encoded by phoB has been implicated in the control of DGTS biosynthesis after phosphate deprivation (19).

Figure 1.

Figure 1

Open in a new tab

Structures of PC and DGTS. R1 and R2 represent the hydrocarbon chains of the respective acyl groups.

Feeding experiments with methionine labeled at specific carbon atoms in algae and moss consistently suggest that S-adenosylmethionine (SAM) donates the 3-amino-3-carboxypropyl portion of the DGTS head group as well as the three methyl groups giving rise to the quaternary ammonium (20–23). Because similar results were obtained with R. sphaeroides (24), it seems plausible that the pathway for DGTS biosynthesis is universal in eukaryotes and prokaryotes. For this reason we chose R. sphaeroides as the initial model organism to identify the genes and enzymes essential for DGTS biosynthesis. A similar approach was successfully used to clone the first genes essential for the biosynthesis of the sulfolipid sulfoquinovosyldiacylglycerol from R. sphaeroides (25, 26) and subsequently the orthologs from cyanobacteria and the plant Arabidopsis thaliana based on sequence similarity to the genes from R. sphaeroides (27, 28).

Materials and Methods

Strains, Plasmids, Media, and Growth Conditions.

Bacterial strains and plasmids (26, 29–32) are shown in Table 1. Cells of R. sphaeroides were grown photoheterotrophically in modified Sistrom's medium (33, 34) containing 50 mM Hepes-KOH, pH 6.8, and different potassium phosphate concentrations as described (18, 26). If required 25 μg/ml kanamycin, 100 μg/ml rifampicin, or 0.8 μg/ml tetracycline was added. Strains of Escherichia coli were grown in Luria broth. Antibiotics were added at concentrations of 100 μg/ml ampicillin, 50 μg/ml kanamycin, or 10 μg/ml tetracycline.

Table 1.

Description of strains and plasmids used in this study

Strain or plasmid Description or construction Source or ref. no.
R. sp. 2.4.1 Wild type ATCC 17023
R. sp. RKL3 DGTS-deficient btaA MNNG-induced mutant This study
R. sp. btaB-dis DGTS-deficient btaB disruption mutant This study
E. coli HB101 F− Δ(mcrC-mrr) leu supE44 ara14 galK2 lacY proA2 rpsL20 (Strr) xyl-5 mtl-1 recA13 30
E. coli XL-10 Gold TetrΔ(mcrA)183 Δ(mcrCB-hsdSMR-mrr) 173 endA1 supE44 thi-1 recA1 gyrA96 relA1 lac Hte [F′ proAB lacIqZDM15 Tn10 (Tetr) Amy Camr] Stratagene
E. coli DH10B F−mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80dlacZDM15 ΔlacX74 deoR recA1 endA1 araD139 Δ(ara, leu)7697 galU galKl-rpsL nupG GIBCO/BRL
E. coli MM294 F−endA1 hsdR17 (rInline graphicmInline graphic) supE44 thi-1 relA1 29
pBluescript II SK(+) Ampr Stratagene
pRK2013 Kanr Tra+ RK2-ColE1rep 31
pCHB500 Tcr; expression vector for R. sphaeroides 26
pUC4K Kanr Nmr; contains neomycin phosphotransferase gene of Tn903 Pharmacia
pSUP202 Ampr Cmr Tcr pBR325rep 32
pRKL301 Cosmid clone complementing RKL3 This study
pRKL323 Smallest subclone of pRKL301 complementing RKL3 This study
pbtaA Nucleotides 436–1834 of pRKL323 in pCHB500 This study
pbtaANT Nucleotides 633–1834 of pRKL323 in pCHB500 This study
pbtaB Nucleotides 1814–2625 of pRKL323 in pBS II SK(+) This study
pbtaB-dis btaB inactivation cassette in pSUP202 This study
Open in a new tab

Mutant Isolation Procedure and Complementation Assay.

Cells of R. sphaeroides wild type were mutagenized with N-methyl-N′-nitro-N-nitrosoguanidine (26). Colonies were grown under phosphate-limiting conditions and screened for polar lipid mutants by analyzing lipid extracts on high-performance TLC plates (HPTLC 60, Merck) essentially as described (26). However, polyhydroxybutyrate, which accumulates in phosphate-stressed cells, was removed from the extracts by precipitation after the addition of 4 vol of hexane. The TLC plates were developed with chloroform, acetone, methanol, acetic acid, and water (50:20:10:10:5 by volume), and the lipids were visualized with iodine vapor. Cosmids and other transferable clones were conjugated into the mutant RKL3 by triparental mating as described (26), and exconjugants were tested for complementation by analyzing lipid extracts as mentioned above.

Quantitative Lipid Analysis.

Cultures (50 ml) of mutant or wild-type strains were grown photoheterotrophically under phosphate-limiting conditions to late logarithmic phase. Cells were harvested by centrifugation, and lipid extracts were prepared as described (35). Lipids were separated by two-dimensional TLC (18), and fatty methyl esters were prepared from each lipid and quantified by gas chromatography as described (26). From these data the mol % fraction of the analyzed polar lipids was calculated for each lipid.

DNA Manipulations.

Standard DNA manipulations were performed according to Sambrook et al. (36) or as suggested by manufacturers of enzymes and kits. The complete insert of pRKL323 was cloned into pBluescript II SK(+) (Stratagene) and sequenced on both strands by the Michigan State University Sequencing Facility. Database searches were conducted at the web site of the National Center for Biological Information (http://www.ncbi.nlm.nih.gov/) using blast algorithms (37). For the construction of pbtaA and pbtaANT, btaA and btaANT (bta N-terminally Truncated) were amplified from pRKL323 by standard PCR using the oligonucleotides 5′-CTTCTAGACGAGGCGAGGCAACGACAGG-3′ and 5′-CTTCTAGACGAAGGACTGATGGAGCGGATGT-3′ as forward primers, respectively, containing an XbaI site and 5′-TCGAATTCGGTAGGTCGCGTCCATCAGC-3′ as reverse primer containing an EcoRI site. The PCR products were digested with XbaI and EcoRI and inserted into the corresponding restriction sites of pCHB500. The btaB inactivation cassette of pbtaB-dis (btaB disrupted) was constructed by deleting the bases 245–397 of btaB and inserting the neomycin phosphotransferase gene from pUC4K. For this purpose, btaB was amplified by PCR using 5′-GCGAATTCTACGGCGGCTTCCACCTCTACC-3′ and 5′-GAGAATTCCTCTGCAACACCGGCTCCACACC-3′ as primer pair and cloned into the EcoRI site of pBluescript II SK(+), giving rise to plasmid pbtaB. The deletion in btaB was introduced by inverse PCR of the whole plasmid using the following primer set: 5′-ATAGATCTAGGCGGCGCTTCATCTCGTG-3′ and 5′-ATAGATCTAGGCCAGCATCTCCTGCGAG-3′, both containing a BglII site. The construction of the inactivation cassette was completed by ligating the 1.3-kb BamHI neomycin phosphotransferase fragment of pUC4k with the PCR product cut with BglII. The complete inactivation cassette was cut with EcoRI and cloned into the corresponding restriction site of pSUP202. The resulting plasmid pbtaB-dis was used to inactivate the respective wild-type gene in R. sphaeroides wild-type cells as described (35). For Southern analysis, probes were labeled by random priming using a kit from Amersham Pharmacia. DNA fragments were separated on agarose gels blotted onto Hybond-N+ nylon membranes (Amersham Pharmacia) by the alkaline transfer method according to the manufacturer's instructions. The DNA was hybridized according to the method of Reed and Mann (38).

Structural Elucidation.

Fast atom bombardment mass spectrometry measurements were done at the Michigan State University Mass Spectrometry Facility. Standard 1H-NMR spectra were recorded with a Varian VXR500 spectrometer (500 MHz for protons) at 25°C using CD3OD/CDCl3 (1:1, vol/vol) as solvent. For each spectrum 200 acquisitions at a recycle delay of 0.2 s were measured. The signal of the deuterated methanol at 3.30 ppm was used as internal standard.

Results

Isolation of DGTS-Deficient Mutants.

To isolate genes essential for betaine lipid biosynthesis we used a genetic approach. Based on our previous experience with the isolation of sulfolipid- and phosphatidylcholine-deficient mutants of R. sphaeroides (35, 39) and the fact that DGTS is only conditionally present in this bacterium, we rationalized that the loss of DGTS may not be lethal and, therefore, the respective mutant could be readily isolated. To enhance the mutation frequency, we treated R. sphaeroides with a mutagenic compound such that the fraction of surviving cells was 1–2% and the occurrence of rifampicin-resistant colonies was increased more than 10-fold. Mutagenized cells were plated on Sistrom's medium that contained 0.1 mM phosphate, leading to phosphate deprivation, a condition inducing the synthesis of DGTS in R. sphaeroides (18). Screening only 288 clones of this mutagenized population by extraction of lipids from individual colonies and analysis of lipid extracts by one-dimensional TLC, six putative mutants with altered lipid composition were identified. Among these was one, RKL3, that apparently lacked DGTS. A thin-layer chromatogram developed under the same conditions as used for screening is shown in Fig. 2, depicting lipid extracts of the wild-type and the RKL3 mutant lines (Fig. 2, left two lanes).

Figure 2.

Figure 2

Open in a new tab

Comparison of lipid extracts from different strains of R. sphaeroides. All cells were grown under phosphate-limited conditions at an initial Pi concentration of 0.1 mM. A one-dimensional thin-layer chromatogram stained by iodine vapor is shown. The indicated strains and plasmids are described in Table 1 and in the text.

Mutant Line RKL3 Lacks the Betaine Lipid DGTS.

To reveal the full complexity of the mutant lipid spectrum, lipids from phosphate-deprived cultures of wild type and RKL3 were compared by using a two-dimensional chromatography system. The spot for DGTS was clearly missing from the RKL3 chromatogram. Instead, the spot for phosphatidylethanolamine was more pronounced and a new faint spot appeared. Because we initially suspected that the new compound present in the RKL3 might be related to the defect in DGTS biosynthesis, we isolated this lipid and subjected it to 1H-NMR analysis. The 1H-NMR spectrum of this lipid (data not shown) was in agreement with a spectrum published for monomethylphosphatidylethanolamine (40). Furthermore, the exact mass determination by fast atom bombardment mass spectrometry provided a value of 758.5716 m/z, very close to the predicted mass for this compound (758.5700 m/z) containing two vaccenic acyl groups (18:1 cis Δ11). We observed this lipid on occasion in wild-type extracts and also noticed it to be absent in more rare cases from RKL3 extracts. Therefore, the accumulation of monomethylphosphatidylethanolamine was not related to a deficiency in DGTS biosynthesis and was affected by the culturing conditions in ways we did not further explore. Nevertheless, we included monomethylphosphatidylethanolamine in the quantitative analysis as shown in Table 2. No traces of DGTS were detectable in the extract of the RKL3 mutant strain, which also showed an increase in relative amounts of the two glycolipids monohexosyldiacylglycerol and glucosylgalactosyldiacylglycerol (Table 2). Furthermore, phospholipids were not as strongly reduced in the RKL3 mutant compared with wild type under the used phosphate-limited growth conditions. Taken together, these data identified RKL3 as a DGTS-deficient mutant of R. sphaeroides.

Table 2.

Lipid composition of R. sphaeroides wild type and RKL3 after phosphate deprivation

Lipid Wild type, mol % RKL3, mol %
MHDG 1.3  ± 0.2 12.7  ± 1.6
DGTS 15.9  ± 1.7 n.d.
MPE n.d. 0.9  ± 0.2
GGDG 32.6  ± 2.4 36.7  ± 2.8
PE 1.5  ± 0.1 7.5  ± 0.4
OL 21.2  ± 3.6 15.2  ± 0.1
PG 4.7  ± 0.6 7.5  ± 0.4
SQDG 17.5  ± 0.9 14.5  ± 0.7
PC 0.1  ± 0.1 1.1  ± 0.6
PL 5.3  ± 0.5 3.7  ± 0.2
Open in a new tab

Mean values from three independent cultures (0.1 mM Pi) and standard errors are shown. GGDG, glucosylgalactosyldiacylglycerol; MHGD, monohexosyldiacylglycerol; MPE, N-monomethylphosphatidylethanolamine; n.d., not detected; OL, ornithine lipid; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PL, undefined phospholipid; SQDG, sulfoquinovosyldiacylglycerol. 

DGTS Biosynthesis Is Restored by a Wild-Type DNA Fragment Containing a Two-ORF Operon.

The isolation of the DGTS-deficient RKL3 mutant provided the basis for the cloning of the corresponding wild-type gene by genetic complementation. For this purpose, individual cosmids containing 20- to 30-kb fragments of wild-type DNA (26) were transferred by conjugation into the RKL3 mutant line. Exconjugants were analyzed for restoration of DGTS biosynthesis by TLC analysis of lipid extracts as in the original mutant screen. Examining 150 clones, one cosmid was identified, RKL301, that rescued the mutant phenotype. A small subclone library was prepared by partial digestion of this cosmid in the expression vector pCHB500, and the subclones were tested again for their ability to restore the DGTS defect in RKL3. Among 384 subclones, 13 complementing clones with inserts of variable length were identified and seven were further characterized (Fig. 3A). The complementing clone, pRKL323 (Fig. 2, lane 3), with the smallest insert of ≈3 kb (Fig. 3A) was submitted to sequence analysis. Close examination of the DNA sequence revealed two adjacent ORFs overlapping by 2 bp that were tentatively designated btaA and btaB (Fig. 3B). Furthermore, a putative promoter with −35 and −10 consensus sequences was identified upstream of the btaA ORF (Fig. 3B). Downstream of the btaB gene a putative Pro-tRNA gene was located. Taken together, these sequence features of the complementing DNA fragment suggested that it contains a small two-ORF operon involved in DGTS biosynthesis.

Figure 3.

Figure 3

Open in a new tab

Cloning and analysis of the DGTS operon of R. sphaeroides. (A) Partial restriction map of cosmid pRKL301 complementing RKL3 and complementing subclones of pRKL301. The region of pRKL301 common to all subclones is indicated by solid line. B, BamHI; P, PstI; S. SalI. (B) Structure of the insert of pRKL323 derived from sequence analysis. The two ORFs associated with DGTS biosynthesis, btaA and btaB, are drawn as gray and open boxes, respectively. In addition, a putative promoter (filled box) with consensus sequences (bold) and a putative proline-tRNA gene (Pro-tRNA) are indicated. Three potential start codons (bold) and ribosome binding sites (bold) are shown. Numbers refer to nucleotides of the sequence deposited at GenBank (accession no. AF329857).

Expression of btaA in RKL3 Restores DGTS Biosynthesis and Leads to the Accumulation of Diacylglyceryl-O-4′-Homoserine (DGHS).

Three potential initiation codons with upstream ribosome binding site consensus sequences were present at the 5′ end of the putative btaA ORF (Fig. 3B). To determine whether btaA or btaB was affected in RKL3 and to examine which of the potential initiation codons of btaA was actually used, a long version of the btaA ORF containing the furthest upstream initiation codon (Fig. 3B, nucleotide 544) as well as a shorter version containing only the second and third initiation codons (Fig. 3B, nucleotides 652 and 709) were inserted into the expression vector pCHB500 under the control of the cytochrome c promoter, which is not affected by phosphate availability. The introduction of the construct containing the longer ORF (pbtaA) into RKL3 resulted in restoration of DGTS biosynthesis whereas the truncated construct (pbtaANT) failed to rescue the mutant phenotype (Fig. 2, lanes 4 and 5). Therefore, btaA starting at the first possible initiation codon (Fig. 3B, nucleotide 544) encodes a protein essential for DGTS biosynthesis in R. sphaeroides. It is also the gene mutated in RKL3. Upon closer examination of lipid extracts from RKL3 harboring the plasmid pbtaA, we discovered the appearance of two other lipid spots [Fig. 4A, diacylglyceryl-O-4′-(N-monomethyl)homoserine (DGMS) and DGHS]. The material from both spots was isolated and analyzed by 1H-NMR and mass spectrometry. One spot (Fig. 4, DGMS) represented a mixture of divaccenic acylglyceryl-N-monomethylhomoserine (736.6058 m/z; expected 736.6091 m/z) and divaccenic acylglyceryl-N-dimethylhomoserine (750.6248 m/z, expected 750.6248 m/z), and the other divaccenic acylglycerylhomoserine (DGHS; 722.5923 m/z, expected 722.5934 m/z). Chemical shift (coupling constants) assignments in ppm (Hz) for the 1H-NMR spectrum of DGHS were as follows: Homoserine H-2, 3.58,(J1,2a 7.0, J1,2b 3.5); H-3a, 2.14; H-3b, 2.03; H-4, 3.65. Glycerol H-1a, 4.33 (J1a,1b 12.2, J1a,2 3.5); H-1b, 4.11 (J1b,2 6.5); H-2, 5.2 (J2,3 5.6); H-3, 3.55. Fatty acids -CH3- 0.8–0.9; -CH2- 1.2–1.4; -βCH2, 2.29–2.31; -βCH2, 1.59; =CH-5.3; -CH2-CH = 1.99. Growing of RKL3 harboring pbtaA under phosphate- sufficient condition led to the formation of small amounts of DGHS but not DGTS (data not shown). Taken together, these observations suggested that btaA may be involved in the formation of DGHS, a postulated precursor of DGTS biosynthesis (24).

Figure 4.

Figure 4

Open in a new tab

Lipid phenotype of RKL3 containing pbtaA (A) and a mutant disrupted in btaB (B). The cells were grown under phosphate-limited conditions at an initial Pi concentration of 0.1 mM. Abbreviations are as defined in the footnote and the legend to Table 2.

Inactivation of btaB Causes DGTS Deficiency and Accumulation of DGHS Under Low-Phosphate Growth Conditions.

The predicted amino acid sequence of btaA revealed only weak similarity to a bacterial methyltransferase and no further clues toward its possible function after database searches. However, the predicted amino acid sequence of the second ORF in the operon, btaB, showed high similarity to methyltransferases and contained clear protein motifs characteristic for SAM binding sites of methyltransferases (41, 42). This observation suggested that btaB encodes an SAM-dependent N-methyltransferase postulated to be involved in DGTS biosynthesis in R. sphaeroides (24). To test this hypothesis, we inactivated btaB in R. sphaeorides wild type by replacing an internal fragment with a kanamycin resistance cassette (Fig. 5A). For confirmation, Southern analysis was conducted by digesting genomic DNA with BglII/HinDIII and probing with the insert of pbtaB. The radiogram of the btaB disruption line, R. sp. btaB-dis, indicated that a ≈9-kb fragment diagnostic for the wild-type chromosome was completely missing (Fig. 5B) in agreement with a loss of all wild-type copies of btaB. Instead, a ≈2-kb fragment predicted for the disrupted btaB gene was present in R. sp. btaB-dis. A second fragment, for which the size could not be predicted due to the lack of genomic sequence information, was expected for R. sp. btaB-dis in this experiment. However, the unexpected third fragment (Fig. 5B) may have been the result of incomplete digestion. Analysis of the lipid extract from R. sp. btaB-dis revealed complete DGTS deficiency (Figs. 2 and 4B). Therefore, btaB as shown for btaA above, is essential for DGTS biosynthesis in R. sphaeroides. Moreover, the closer examination of the lipid extract from R. sp. btaB-dis grown under phosphate limitation revealed the accumulation of DGHS (Fig. 4B, DGHS) and to some extent of partially methylated DGHS (Fig. 4B, DGMS).

Figure 5.

Figure 5

Open in a new tab

Disruption of the btaB gene. (A) Maps of the inactivation construct pbtaB-dis and the genomic DNA surrounding the inactivated btaB gene in R. sphaeroides btaB-dis. The kanamycine resistance cassette is drawn as a black box. Numbers refer to nucleotides of the btaB ORF (open box). (B) Genomic DNA blot of R. sphaeroides and btaB-dis strain probed with the btaB wild-type gene. Genomic DNA was digested with BglII and HinDIII. The blot was probed with the insert of pbtaB. Diagnostic hybridizing fragments visualized by exposure to x-ray film and their approximate length are indicated.

Discussion

During evolution, living organisms adopted a wide range of amphipathic molecules that can serve as polar lipids. Membranes of some bacteria, such as R. sphaeroides, contain a rich complement of polar lipids. Its biochemical resourcefulness allows R. sphaeroides not only to change the membrane lipid composition under adverse conditions, e.g., phosphate starvation (18, 35); it also has the capability to incorporate xenobiotic compounds such as the buffer substance Tris(hydroxymethyl)aminomethane (Tris) to form phosphatidyl-Tris (43). Because R. sphaeroides genetics is straightforward, this bacterium represents a very useful model organism for the identification of genes that encode enzymes essential for the biosynthesis of some of the less well-known membrane lipids that are, nevertheless, very common in large groups of organisms.

Screening only a few hundred heavily mutagenized colonies of R. sphaeroides, we were able to identify a mutant, RKL3, unable to produce DGTS when grown on low-phosphate medium. Brute force complementation cloning using RKL3 yielded a wild-type DNA fragment containing a putative two-gene operon. Expression of the first gene, btaA, under the control of the Rhodobacter capsulatus cytochrome c promoter of pCHB500 (26) in RKL3 restored DGTS biosynthesis and led to the accumulation of DGHS as well as partially methylated DGHS under low-phosphate growth conditions (Fig. 4A). Apparently, btaA is overexpressed in this strain and more DGHS is produced than can be processed by the DGHS methylase. The expression of btaA under the control of the cytochrome c promoter is not affected by phosphate deprivation. When RKL3 containing pbtaA was grown on high-phosphate medium, no DGTS was formed and DGHS accumulated (data not shown), presumably because the phosphate-induced DGHS methylase activity was not present under these conditions. The database search using the predicted btaA gene product revealed only weak similarity to a bacterial methyltransferase and two other proteins of unknown function. Considering labeling data obtained by Hofmann and Eichenberger (24), which indicated that the homoserine moiety in DGTS is derived from SAM, the accumulation of the intermediate DGHS in RKL3/pbtaA strongly suggests that btaA encodes an SAM:diacylglycerol 3-amino-3-carboxypropyl transferase catalyzing the first step of DGTS biosynthesis in R. sphaeroides (Fig. 6). Although rare, SAM-dependent 3-amino-3-carboxypropyl transferases are known to participate in other bacterial syntheses, e.g., in the biosynthesis of the antibiotic nocardicine (44).

Figure 6.

Figure 6

Open in a new tab

Hypothesis for the function of btaA and btaB in DGTS biosynthesis in R. sphaeroides. DAG, diacylglyceryl; 5′-MTA, 5′-methylthioadenosine; S-AHC, S-adenosylhomocysteine.

The metabolite DGHS has been proposed to be the intermediate of DGTS biosynthesis in R. sphaeroides (24) that becomes N-methylated in an SAM-dependent reaction. Interestingly, the protein predicted to be encoded by btaB shows similarity to SAM-dependent methyltransferases. Two of the three characteristic amino acid sequence motifs (amino acids 47–55, motif I and amino acids 137–146, motif II) are clearly recognizable (41, 42), whereas motif II (amino acids 110–117) is less obvious. The strong accumulation of DGHS following the disruption of btaB in R. sp. btaB-dis grown on low-phosphate medium (Fig. 4B) is consistent with a defect in N-methylation leading to the accumulation of the DGTS precursor DGHS. However, assuming that btaB is completely inactivated in this line, another methyltransferase must be present under these conditions that is able to methylate DGHS with low efficiency, because a fairly small amount of partially methylated DGHS is present in the R. sp. btaB-dis mutant. Furthermore, this methyltransferase seems only capable of transferring one or two methyl groups to DGHS because DGTS is clearly not formed in the btaB disruption line. Therefore, it seems highly probable that btaB encodes the SAM:DGHS methyltransferase required for DGTS biosynthesis in R. sphaeroides (Fig. 6). The pmtA encoded SAM:phosphatidylethanolamine methyltransferase of R. sphaeroides catalyzes all three N-methylations during phosphatidylcholine biosynthesis (39). However, in yeast two distinct enzymes are involved in the sequential N-methylation of phosphatidylethanolamine (45). At the present time, we cannot fully exclude the possibility that the btaB gene product primarily catalyzes the later methylation steps during DGTS biosynthesis and that another methyltransferase catalyzes the first methylation reactions.

Acknowledgments

We thank Beverly Chamberlain and Doug Gage at the Michigan State University mass spectrometry facility for the analysis of lipids. This work has been supported in parts by grants from the National Science Foundation (MCB-9807993) and the Michigan State University Center for Plant Products and Technologies.

Abbreviations

PC

phosphatidylcholine

DGHS

diacylglyceryl-O-4′-homoserine

DGMS

diacylglyceryl-O-4′-(N-monomethyl)homoserine

DGTS

diacylglyceryl-O-4′-(N,N,N,-trimethyl)homoserine

SAM

S-adenosylmethionine

Footnotes

This paper was submitted directly (Track II) to the PNAS office.

Data deposition: The sequence reported in this paper has been deposited in the GenBank database (accession no. AF329857).

References

  • 1.Dembitsky V M. Prog Lipid Res. 1996;35:1–51. doi: 10.1016/0163-7827(95)00009-7. [DOI] [PubMed] [Google Scholar]
  • 2.Sato N. Bot Mag Tokyo. 1992;105:185–197. [Google Scholar]
  • 3.Sato N, Murata N. Biochim Biophys Acta. 1991;1082:108–111. doi: 10.1016/0005-2760(91)90306-3. [DOI] [PubMed] [Google Scholar]
  • 4.Brown A E, Elovson J. Biochemistry. 1974;13:3476–3482. doi: 10.1021/bi00714a009. [DOI] [PubMed] [Google Scholar]
  • 5.Araki S, Eichenberger W, Sakurai T, Sato N. Plant Cell Physiol. 1991;32:623–628. [Google Scholar]
  • 6.Eichenberger W, Boschetti A. FEBS Let. 1977;88:201–204. [Google Scholar]
  • 7.Eichenberger W. Plant Sci Lett. 1982;24:91–95. [Google Scholar]
  • 8.Eichenberger W. Plant Physiol Biochem. 1993;31:213–221. [Google Scholar]
  • 9.Evans R W, Kates M, Wood G W. Chem Phys Lipids. 1982;31:331–338. [Google Scholar]
  • 10.Sato N, Furuya M. Plant Cell Physiol. 1983;24:1113–1120. [Google Scholar]
  • 11.Sato N, Furuya M. Phytochemistry. 1984;23:1625–1627. [Google Scholar]
  • 12.Sato N, Furuya M. Plant Sci. 1985;38:81–85. [Google Scholar]
  • 13.Sato N. Plant Cell Physiol. 1991;32:819–825. [Google Scholar]
  • 14.Künzler K, Eichenberger W. Phytochemistry. 1997;46:883–892. doi: 10.1016/s0031-9422(97)81274-5. [DOI] [PubMed] [Google Scholar]
  • 15.Yamada T A, Nozawa Y. Biochim Biophys Acta. 1979;574:433–439. doi: 10.1016/0005-2760(79)90239-x. [DOI] [PubMed] [Google Scholar]
  • 16.Vaskovsky V E, Khotimchenko S V, Benson A A. Lipids. 1991;26:254–256. [Google Scholar]
  • 17.Furlong S T, Leary J A, Costello C E, Dawidowicz E A. J Lipid Res. 1986;27:1182–1189. [PubMed] [Google Scholar]
  • 18.Benning C, Huang Z H, Gage D A. Arch Biochem Biophys. 1995;317:103–111. doi: 10.1006/abbi.1995.1141. [DOI] [PubMed] [Google Scholar]
  • 19.Geiger O, Röhrs V, Weissenmayer B, Finan T M, Thomas-Oates J E. Mol Microbiol. 1999;32:63–73. doi: 10.1046/j.1365-2958.1999.01325.x. [DOI] [PubMed] [Google Scholar]
  • 20.Sato N. Plant Physiol. 1988;86:931–934. doi: 10.1104/pp.86.3.931. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Sato N, Kato K. Plant Sci. 1988;55:21–25. [Google Scholar]
  • 22.Sato N. Plant Cell Physiol. 1991;32:845–851. [Google Scholar]
  • 23.Vogel G, Eichenberger W. Plant Cell Physiol. 1992;33:427–436. [Google Scholar]
  • 24.Hofmann M, Eichenberger W. J Bacteriol. 1996;178:6140–6144. doi: 10.1128/jb.178.21.6140-6144.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Benning C, Somerville C R. J Bacteriol. 1992;174:6479–6487. doi: 10.1128/jb.174.20.6479-6487.1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Benning C, Somerville C R. J Bacteriol. 1992;174:2352–2360. doi: 10.1128/jb.174.7.2352-2360.1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Essigmann B, Güler S, Narang R A, Linke D, Benning C. Proc Natl Acad Sci USA. 1998;95:1950–1955. doi: 10.1073/pnas.95.4.1950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Güler S, Seeliger A, Härtel H, Renger G, Benning C. J Biol Chem. 1996;271:7501–7507. doi: 10.1074/jbc.271.13.7501. [DOI] [PubMed] [Google Scholar]
  • 29.Bachmann B J. In: Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. Neidhardt F C, Ingraham J L, Low K B, Magasanik B, Schaechter M, Umbarger H E, editors. Washington, DC: Am. Soc. Microbiol.; 1987. pp. 1190–1290. [Google Scholar]
  • 30.Boyer H W, Roulland-Dussoix D. J Mol Biol. 1969;41:459–472. doi: 10.1016/0022-2836(69)90288-5. [DOI] [PubMed] [Google Scholar]
  • 31.Figurski D, Helinski D R. Proc Natl Acad Sci USA. 1979;76:1648–1652. doi: 10.1073/pnas.76.4.1648. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Simon R, Priefer U, Pühler A. Bio/Technology. 1983;1:784–791. [Google Scholar]
  • 33.Sistrom W R. J Gen Mircrobiol. 1960;22:778–785. doi: 10.1099/00221287-22-3-778. [DOI] [PubMed] [Google Scholar]
  • 34.Sistrom W R. J Gen Microbiol. 1962;28:607–616. doi: 10.1099/00221287-28-4-607. [DOI] [PubMed] [Google Scholar]
  • 35.Benning C, Beatty J T, Prince R C, Somerville C R. Proc Natl Acad Sci USA. 1993;90:1561–1565. doi: 10.1073/pnas.90.4.1561. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Sambrook J, Fritsch E F, Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd Ed. Plainview, NY: Cold Spring Harbor Lab. Press; 1989. [Google Scholar]
  • 37.Altschul S F, Gish W, Miller W, Myers E W, Lipman D J. J Mol Biol. 1990;215:403–410. doi: 10.1016/S0022-2836(05)80360-2. [DOI] [PubMed] [Google Scholar]
  • 38.Reed K C, Mann D A. Nucleic Acids Res. 1985;13:7207–7221. doi: 10.1093/nar/13.20.7207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Arondel V, Benning C, Somerville C R. J Biol Chem. 1993;268:16002–16008. [PubMed] [Google Scholar]
  • 40.McGuigan C, Anson M S, Swords B. J Chem Soc Perkin Trans 1. 1992;16:2075–2078. [Google Scholar]
  • 41.Joshi C P, Chiang V L. Plant Mol Biol. 1998;37:663–674. doi: 10.1023/a:1006035210889. [DOI] [PubMed] [Google Scholar]
  • 42.Kagan R M, Clarke S. Arch Biochem Biophys. 1994;310:417–427. doi: 10.1006/abbi.1994.1187. [DOI] [PubMed] [Google Scholar]
  • 43.Schmid P C, Kumar V V, Weis B K, Schmid H H O. Biochemistry. 1991;30:1746–1751. doi: 10.1021/bi00221a003. [DOI] [PubMed] [Google Scholar]
  • 44.Reeve A M, Breazeale S D, Townsend C A. J Biol Chem. 1998;273:30695–30703. doi: 10.1074/jbc.273.46.30695. [DOI] [PubMed] [Google Scholar]
  • 45.Nikoloff D M, Henry S A. Annu Rev Genet. 1991;25:559–583. doi: 10.1146/annurev.ge.25.120191.003015. [DOI] [PubMed] [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES