TEXT

Top Abstract Text References

The biosynthetic steps of the ubiquinone (coenzyme Q₈ [CoQ₈]) pathway in Escherichia coli have been known for many years (11, 16). However, the identities of some ubiquinone biosynthetic genes remained hidden, even after the complete nucleotide sequence of E. coli was published. This was due partly to a lack of sequence data on ubiquinone biosynthetic genes in other organisms and partly to the multiplicity of open reading frames at the suspected locations. Thus, at 86.7 min on the E. coli chromosome, where the ubiD gene was mapped (3), there were no fewer than seven unidentified, hypothetical genes. A systematic search of these some years ago failed to locate the ubiD gene (R. Meganathan, personal communication). This was likely due to the fact that the some of the early gene assignments were incorrect. Recently the identities of the ubiE (8) and ubiF (7) genes were published, leaving ubiD the last of the known ubi genes to be identified.

View larger version (5K): [in this window] [in a new window]   FIG. 1.   Composition of the 2,595-bp-long E. coli chromosomal fragment in p522, showing the location of the ubiD gene.

We prepared a PUC 18 plasmid library which contained chromosomal fragments of E. coli K-12 strain AN256, the isogenic ubiD⁺ strain of the ubiD mutant strain AN66. Strain AN66 (thr-1 leuB6 ubiD410) (3) was obtained from the E. coli Genetic Stock Center, New Haven, Conn., and strain AN256 (thr-1 leuB6) (9) was from C. F. Clarke's laboratory. Chromosomal fragments, 3 to 12 kDa in size, were obtained by partial digestion with Sau3AI restriction enzyme. Competent AN66 cells were electroporated in the presence of this plasmid library, and ampicillin-resistant transformants that could grow on plates which contained minimal medium (6), ampicillin (100 µg/ml), leucine and threonine (20 µg/ml each), and succinate (3 mg/ml) were isolated.

In our hands, AN66 cells spontaneously acquired the capacity to utilize succinate, at a frequency of 0.01%. Thus, when competent cultures of AN66 were transformed with the plasmid pBR322, approximately 300 transformed colonies which grew on succinate were obtained. These transformed revertant colonies (which grew faster on succinate plates than the ubiD⁺ strain AN256) seriously interfered with the identification of ubiD gene-harboring transformants.

Therefore, several cycles of transformations were carried out to enrich the transformant population with ubiD gene-containing plasmids. This was done by recovering all transformed colonies, growing them together in Luria-Bertani medium with ampicillin, and extracting their plasmids. Competent cells were transformed by this preparation, and the procedure was repeated again. After two cycles, besides the 300 or so transformed revertant colonies, a strong haze was also seen on the succinate-containing selection plate.

Cells from this haze were cultured, and their plasmids were extracted. This plasmid preparation produced 1.3 × 10⁵ transformed colonies that were able to grow on succinate as the sole carbon source. One of these was isolated and named AN66p522. Colony sizes of transformed cells on succinate plates were comparable to those of the ubiD⁺ strain AN 256.

The ubiquinone contents of strains AN256, AN66, and AN66p522 were determined by a method described earlier (17). Cell preparation included the growth of a 500-ml culture of strain AN256 in Luria-Bertani medium plus glucose (0.3%, wt/vol) and identical volumes of AN66 and AN66p522 in brain heart infusion broth plus glucose (0.3%, wt/vol). At an A₆₀₀ of 0.9 to 1.0, the cells were harvested, washed with distilled water, and lyophilized. We found 0.29 nmol of CoQ₈ per mg of dried cells of strain AN256, 0.05 nmol of CoQ₈ per mg of dried cells of strain AN66, and 0.73 nmol of CoQ₈ per mg of dried cells of strain AN66p522. Thus the chromosomal fragment on plasmid p522 fully complemented the ubiquinone deficiency of AN66.

Sequencing of this chromosomal fragment showed that it was 2,595 bp long. It started near the end of the open reading frame yigW_2, 146 bases downstream from the end of the rfaH gene (a regulatory gene of lipopolysaccharide, sex factor, and hemolysin genes, oriented in the opposite direction from yigW_2) and ended 252 bases into the fre gene. The only other open reading frame located between rfaH and fre was yigC. It is our contention that this 1,491-base segment, immediately upstream from the fre gene at 86.7 min on the E. coli chromosome, is the ubiD gene (Fig. 1).

Until recently, the fre gene, coding for NAD(P)H flavin oxidoreductase, was designated ubiB (5). However, the true ubiB gene is now shown to be the former open reading frame yigR, approximately 6 kb downstream from its previous location (12).

The yigC segment and its upstream region were isolated by PCR from the AN256 chromosome (primers used were as follows: in the forward direction, 5'-GATCATCGGTGCCAGGCAATTCACAGCC-3', in the reverse direction, 5'-TCAGGCGCTTTTACCGTTGTTAAAA-3'). It was cloned into a pNoTA/T7 shuttle vector (manufactured by 5 Prime  3 Prime Inc.), and this construct was transformed into AN66 cells. This plasmid, designated p613, complemented the ubiD mutant trait of AN66 cells to the same extent as the larger plasmid, p522.

Based on its nucleotide sequence, the product of ubiD gene is a 497-amino-acid protein, its molecular mass is 55,603.7 Da, and its theoretical isoelectric point is 5.31. The ubiD gene product is one of two enzymes (3-octaprenyl-4-hydroxybenzoate carboxy-lyase) which catalyze the decarboxylation of 3-octaprenyl-4-hydroxy benzoate to 2-octaprenylphenol. Earlier work with this enzyme suggested that it is a membrane-associated protein, although during cell fractionation much activity was found in the cytoplasmic fraction (9). Analysis of its amino acid sequence for transmembrane helices indicated zero (13), one (positions 215 to 235) (K. Hofmann and W. Stoffel, Biol. Chem. Hoppe-Seyler 347:166, abstr. MF C-35), or two (positions 226 to 232 and 334 to 340) (4) such regions, depending on which program was used. This enzyme's molecular mass by gel filtration measurement was reported to be approximately 340,000 Da (9). This suggests that it is a hexameric protein in vivo.

We isolated the ubiD gene from strain AN66 by PCR and sequenced it, for the purpose of locating the site of mutation. The long gene was sequenced in overlapping segments, and the last fragment was sequenced in the reverse direction as well. (The following primers were used: 1, 5'-ATGGACGCCATGAAATATAACGATT-3'; 2, 5'-GCGTGGCGATGGGCATGGGGCAGG-3'; 3, 5'-GCATTCCCATTATGACCTGCTGGCCGG-3'; 4, 5'-GGTGCCGATCCCGCCACGATTCTCGG-3'; and 5, 5-GGGCGTCCGCCAGATGAGCCCGCGGCGGTG-3' [all forward direction] and 5'-TCAGGCGCTTTTACCGTTGTTAAAA-3' [reverse direction].) Comparison of the results with the published nucleotide sequence showed a single deviation. Codon 452, GGG (glycine), appeared in the mutant as AGG (arginine). A homology study (see below) shows that this glycine is a consensus residue.

A comparison (14) between the amino acid sequences of the ubiX and yigC (ubiD) gene products showed no significant similarities. A Blast search of the entire protein database of 525,243 sequences yielded 86 hits of similarities with yigC (ubiD). One of these hits was a 29% sequence identity and 48% similarity with a 4-hydroxybenzoate decarboxylase from Clostridium hydroxybenzoicum.

In the Swiss-Prot database yigC is listed as a member of the family UPF (uncharacterized protein family) 00096, with a taxonomic range of archaea, eubacteria, and eukaryota. Table 1 lists these proteins, and Fig. 2 shows regions of homology among them, as assigned by the Multaline program (2). Clearly not all of these proteins can be functional homologues of ubiD. Two proteins are from the same organism (Aeropyrum pernix). The gram-positive organisms Bacillus subtilis and Streptomyces sp. synthesize menaquinones, and the cyanobacterium Synechocystis sp. makes plastoquinone instead of ubiquinone (1). Helicobacter pylori also utilizes menaquinone instead of ubiquinone (10). Archaebacteria also employ a variety of ubiquinone analogues (15). However, due to the extensive homology between these proteins, it is reasonable to expect that they all function in some membrane-associated decarboxylation process.

View larger version (262K): [in this window] [in a new window]   FIG. 2.   Homology between the ubiD gene product and 18 members of the UPF 00096 family of proteins, as arranged by the Multaline program. The identity of each protein is shown in Table 1. Lowercase letters designate consensus amino acids in 50 to 90% of the sequences, and boldface capital letters indicate 90 to 100% homology.