Genetic Characterization of the Carotenoid Biosynthetic Pathway in Methylobacterium extorquens AM1 and Isolation of a Colorless Mutant

Abstract

Genomic searches were used to reconstruct the putative carotenoid biosynthesis pathway in the pink-pigmented facultative methylotroph Methylobacterium extorquens AM1. Four genes for putative phytoene desaturases were identified. A colorless mutant was obtained by transposon mutagenesis, and the insertion was shown to be in one of the putative phytoene desaturase genes. Mutations in the other three did not affect color. The tetracycline marker was removed from the original transposon mutant, resulting in a pigment-free strain with wild-type growth properties useful as a tool for future experiments.

Methylotrophic bacteria are capable of growth using C₁ compounds such as methanol as their only carbon and energy source, and therefore they could potentially serve as biocatalysts for the conversion of methanol to useful products. Methanol is a good candidate for an alternative feedstock, as it is abundant, soluble, and low in cost. The pink-pigmented facultative methylotroph Methylobacterium extorquens AM1 is known to generate a pink carotenoid (6, 11). Since carotenoids have a variety of industrial uses (4), the carotenoid biosynthesis pathway is a potential target for metabolic engineering of this bacterium. In this work, genomic reconstruction and mutagenesis were used to characterize the carotenoid biosynthetic pathway in M. extorquens AM1.

Reconstruction of the carotenoid biosynthetic pathway from the genome sequence.

The biosynthetic pathway for carotenoids has been elucidated through extensive study of various plants and bacteria (3) and is outlined in Fig. 1. The first part of the pathway, common to all isoprenoids, is the production of isopentenyl pyrophosphate (IPP). In most bacteria, this occurs through a pathway that begins with glyceraldehyde-3-phosphate and proceeds through 1-deoxy-d-xylulose-5-phosphate (DXP), as shown in Fig. 1 (13). This is referred to as the mevalonate-independent pathway. Although the details of the conversion to IPP are not fully understood, some of the genes in these intermediate steps have recently been identified (3, 12). Protein sequences of the enzymes involved in the upper pathway were obtained from the GenBank (http://www.ncbi.nlm.nih.gov) database and used to search the M. extorquens AM1 partial genome sequence (http://www.integratedgenomics.com/genomereleases.html#list6). Putative gene function was further evaluated by a BLAST search of the corresponding translated sequence against the National Center for Biotechnology Information database (http://www.ncbi.nlm.nih.gov/BLAST). The results of this search are listed in Table 1. Genes potentially encoding all of the steps of the pathway were found, with the exception of the IPP delta isomerase (idi gene). Since the production of IPP is essential for all organisms, it is likely that this gene is either divergent or located in the unsequenced portion of the genome. The other pathway for IPP synthesis, uncommon in bacteria, proceeds through mevalonic acid as an intermediate (3). No genes predicted to encode enzymes of this pathway were identified in the M. extorquens AM1 genome sequence.

FIG. 1.

Summary of bacterial carotenoid biosynthesis pathways and associated genes compiled from references 1, 3, 8, 12, and 13.

TABLE 1.

Summary of putative genes for the non-mevalonate pathway for IPP production in M. extorquens AM1^a

Gene(s)	Protein in GenBank database with lowest E value	BLASTP E value	% Identity	Accession no.
dxp	R. capsulatus DXP synthase	<1e-199	56	P26242|DXS_RHOCA
dxs	Zymomonas mobilis DXP reductoisomerase	2e-70	44	Q9X5F2|DXR_ZYMMO
ispD-2ispF	R. capsulatus IspD-IspF bifunctional protein	6e-47	44	Q08113|ISDF_RHOCA
ispE	Z. mobilis IspE	1e-28	35	Q9X3W5|ISPE_ZYMMO
idi	IPP delta isomerase gene not found

^aEnzymes used for functional assignments are those found by a BLAST search of the translated gene sequence yielding the lowest E value.

IPP and its isomer dimethylallyl pyrophosphate are condensed to various degrees to form a wide variety of isoprenoid compounds, some of which are essential for growth. The first step unique to carotenoid synthesis is the condensation of four such units to make geranylgeranyl diphosphate (GGPP) via the enzyme GGPP synthase. Two of these molecules are then condensed by phytoene synthase to form phytoene, a precursor of all carotenoids (3). Phytoene desaturases then introduce a variable number of desaturations to create molecular diversity. The Rhodobacter capsulatus phytoene desaturase introduces three desaturations to produce neurosporene, which is yellow, while phytoene desaturases from many other bacteria introduce four desaturations to produce the red compound lycopene (7). Beginning with these compounds, further diversity in carotenoids is introduced by modifications such as desaturation, methylation, and hydroxylation. Examples include β-carotene, zeaxanthin, and astaxanthan (3).

Identitysearches were performed by using translated sequences of known R. capsulatus (1) and Erwinia herbicola (8, 15) genes to locate putative carotenoid biosynthetic genes in the M. extorquens AM1 partial genome sequence (Table 2), which were further evaluated as described above by using BLAST. The predicted carotenoid biosynthesis genes are not tightly clustered, as they are located on seven different contigs in the genome sequence, separated by intervening DNA. Genes predicted to encode all steps for the formation of the first colored carotenoid (e.g., lycopene) from IPP were identified in the genome. In addition, three candidates for genes involved in further modification were found, with the highest identity to the hydroxyspheroidene pathway of R. sphaeroides (1). No candidates for the Erwinia β-carotene-zeaxanthan pathway (8) were found. Multiple candidates were identified for some carotenoid biosynthesis genes: two genes for phytoene synthase (crtB) and four for phytoene desaturase (crtI).

TABLE 2.

Summary of putative carotenoid biosynthesis genes found in the M. extorquens AM1 genome sequence^a

Gene(s)	Protein in GenBank with lowest E value	BLASTP E value	% Identity	Accession no.
crtB	R. capsulatus phytoene synthase	1e-50	46	P17056|CRTB_RHOCA
orf1	R. capsulatus phytoene desaturase (dehydrogenase)	1e-121	47	P17054|CRTI_RHOCA
zdsA	Anabaena sp. zeta-carotene desaturase	3e-66	32	D26095
crtZ	R. capsulatus phytoene desaturase	2e-61	31	P17054/CRTI_RHOCA
Gap of 4 putative ORFs
crtI	Erwinia herbicola phytoene desaturase	2e-50	31	AAA64981
crtB	Targetus erecta (marigold) phytoene synthase	5e-22	37	AAM45379
orf2	Streptomyces coelicolor putative phytoene dehydrogenase	0.003	24	NP_630834
crtD (incomplete sequence)	Rubrivax gelatinosus methoxyneurosporene dehydrogenase	3e-70	44	AAC44798
crtE	R. sphaeroides GGPP synthase	7e-51	42	P54976|CRTE_RHODSH
crtF	R. sphaeroides methoxyneurosporene methyltransferase	4e-49	39	T50751
ispD-ispF	IspD-IspF (see Table 1)
ddsA	Paracoccus zeaxanthinifaciens decaprenyl diphosphate synthase (quinone biosynthesis)	4e-84	54	CAD24417
ispA	R. sphaeroides FPP synthase	6e-43	46	BAA96457
crtC (incomplete sequence)	R. sphaeroides hydroxyneurosporene synthase	8e-12	42	QO167055|CRTC_RHOSH

^aAbbreviations are given in Fig. 1. Genes are grouped as they are clustered in the genome, with empty lines separating clusters. Enzymes used for the functional assignment are those found by a BLAST search of the translated gene sequence yielding the lowest E value.

Isolation and genetic characterization of a colorless mutant of M. extorquens AM1 and generation of unmarked strain AM1-W.

Transposon mutagenesis using the mini-Tn5 derivative ISphoA/hah-Tc (5) was performed as previously described (10) in order to determine whether a single insertion mutation could result in loss of carotenoid production. Three white transposon mutants were isolated out of approximately 7,000 screened visually. The chromosomal site of insertion in each of these strains was determined as previously described (10). The sequences obtained were all mapped to the same open reading frame (ORF) in the M. extorquens AM1 genome sequence: one of the four genes predicted to encode phytoene desaturase (phytoene dehydrogenase). This gene is shown in bold in Table 2 and will be referred to as crtI.

For future genetic screening of this strain with colorimetric substrates, it would be useful to have a colorless, unmarked mutant. In order to generate such a pigment-free mutant, the tetracycline resistance gene of crtI::ISphoA/hah-Tc mutant S234-13 was removed by using cre expression plasmid pCM158 (9). The resulting unmarked, white mutant of M. extorquens AM1, AM1-W, was confirmed by analytical PCR assay. Under all standard laboratory conditions, growth of AM1-W was indistinguishable from that of the pigmented wild-type strain (data not shown).

To verify that crtI is responsible for the loss of color in the mutants, as opposed to a polar effect, gene complementation was performed. A 1.5-kb region containing crtI was amplified by PCR, cloned directly into pCR2.1 (Invitrogen), and subcloned as a BamHI-EcoRI fragment into pCM66 and pCM160 (10) to produce pCAR12 and pCAR14. The resulting plasmids contain crtI expressed from Escherichia coli P_lac or M. extorquens AM1 P_mxaF, respectively. Both plasmids were introduced by conjugation with E. coli S17-1 (16) into AM1-W, resulting in pink colonies. In contrast, colonies of AM1-W containing either pCM66 or pCM160 remained white. These results demonstrate that the M. extorquens crtI gene is essential for formation of the pink pigment and confirm previous suggestions (6, 11) that the pigment was due to the presence of a carotenoid.

Surprisingly, crtI was not the most likely candidate to encode an active phytoene desaturase, on the basis of the information in Table 2. The gene labeled orf1 is located immediately downstream of a putative phytoene synthase gene and has greater identity to a known phytoene desaturase. In contrast, crtI is surrounded by putative phospholipid and cell wall biosynthesis genes and a GMP synthase. Deletion mutants with changes in two of the remaining phytoene desaturase gene candidates, orf2 and orf3, were prepared by using PCR products of these ORFs and the allelic exchange vector pCM184 (9). These mutants were confirmed to be double crossovers by PCR analysis, indicating that they have null mutations in the corresponding gene. Both mutants remained pink, suggesting that these genes do not play a major role in the synthesis of colored carotenoids in M. extorquens AM1.

Isolation of carotenoid from wild-type M. extorquens AM1.

Carotenoids were extracted from M. extorquens AM1 by using a modification of the procedure commonly used for photosynthetic bacteria (17). Cell pellets obtained from approximately 100 ml of cell culture grown in mineral salts medium (2) containing 125 mM methanol were resuspended in 1 ml of methanol at 65°C and vortexed for 1 min. A 0.4-ml volume of water and 0.3 ml of chloroform were added, and the sample was vortexed for an additional 2 min. The bottom organic layer, containing the carotenoid, was extracted into clean tubes. A 1-ml volume of methanol and 0.4 ml of water were added, and the organic layer was extracted again. The extract from wild-type cells was pink, whereas that from mutant strain AM1-W was colorless. A 2-ml volume of acetone was added to the extract (about 0.15 ml) in order to remove phospholipids and glycolipids (14), and the sample was placed at −20°C overnight. The sample was centrifuged, and the supernatant was evaporated to dryness and redissolved in 0.1 ml of chloroform. The visible absorption spectrum of the sample, with maxima at 480, 510, and 540 nm, matched that of the extract from a related pink-pigmented methylotroph, previously referred to as Protaminobacter ruber(14). Furthermore, as observed with the carotenoid from P. ruber (14), a 9-nm red shift in the absorption spectra was observed following hydrolysis of the carotenoid in 0.2 M HCl-methanol. These results suggest that the pink carotenoid in M. extorquens AM1 is similar to that in P. ruber, which has been shown to be a carotenoid with an attached sugar moiety that is lost during hydrolysis. The ¹H nuclear magnetic resonance spectrum of the P. ruber extract indicated the presence of a carbonyl group and about 13 conjugated double bonds, and thin-layer chromatography data suggested that the carotenoid has a bacterioruberin- or oscillaxanthin-like structure (14).

Nucleotide sequence accession number.

The sequence of M. extorquens AM1 crtI has been deposited in the GenBank database (accession number AY331188).