Rickettsial metK-Encoded Methionine Adenosyltransferase Expression in an Escherichia coli metK Deletion Strain

Abstract

The obligate intracellular bacterium Rickettsia prowazekii has recently been shown to transport the essential metabolite S-adenosylmethionine (SAM). The existence of such a transporter would suggest that the metK gene, coding for the enzyme that synthesizes SAM, is unnecessary for rickettsial growth. Genome sequencing has revealed that this is the case for the metK genes of the spotted fever group and the Madrid E strain of R. prowazekii, which contain recognizable inactivating mutations. However, several strains of the typhus group rickettsiae possess metK genes lacking obvious mutations. In order to determine if these genes code for a product that retains MAT function, an Escherichia coli metK deletion mutant was constructed in which individual rickettsial metK genes were tested for the ability to complement the methionine adenosyltransferase deficiency. Both the R. prowazekii Breinl and R. typhi Wilmington metK genes complemented at a level comparable to that of an E. coli metK control, demonstrating that the typhus group rickettsiae have the capability of synthesizing as well as transporting SAM. However, the appearance of mutations that affect the function of the metK gene products (a stop codon in the Madrid E strain and a 6-bp deletion in the Breinl strain) provides experimental support for the hypothesis that these typhus group genes, like the more degenerate spotted fever group orthologs, are in the process of gene degradation.

Rickettsia prowazekii, the causative agent of epidemic typhus, is an obligate intracellular bacterium that grows directly in the cytoplasm of its eukaryotic host cell rather than within a host cell vacuole. To exploit this intracellular niche, R. prowazekii has evolved specialized transporters for the complex metabolic intermediates found in the host's cytoplasm (4, 5, 14, 16, 17). The ability to transport the end products of biosynthetic pathways has permitted the rickettsiae to undergo the reductive evolution of biosynthetic genes, resulting in genomes that contain numerous pseudogenes and a high proportion of noncoding DNA (3, 10, 11).

The product of the metK gene, methionine adenosyltransferase (MAT), synthesizes S-adenosylmethionine (SAM). SAM is one of the most versatile compounds found in living cells. Not only does it function as the primary methyl donor in a number of biosynthetic reactions, but it also serves as a substrate in the polyamine biosynthetic pathway (8, 9, 13). Thus, SAM is an essential compound in the growth of bacteria, including R. prowazekii (12). The essential nature of SAM is supported by the fact that a deletion of the E. coli metK gene cannot be obtained unless a second functional metK gene is present (15). Examination of the genome sequence of the Madrid E strain of R. prowazekii revealed that the metK gene is a pseudogene in that there is an interruption of the coding sequence (2, 3). However, in contrast to the more disrupted metK pseudogenes of the spotted fever group rickettsiae, the R. prowazekii Madrid E strain contains a single nonsense mutation near the middle of the coding region (1, 2, 3, 11). Interestingly, the virulent Breinl strain of R. prowazekii and the Wilmington strain of R. typhi possess metK genes with complete open reading frames (1, 10).

If SAM is essential, then the fact that R. prowazekii Madrid E metK contains a stop codon is a strong indicator that R. prowazekii has evolved a transport system for SAM. This hypothesis led to the discovery of the first bacterial SAM transporter in R. prowazekii (Madrid E and Breinl strains) and R. typhi (Wilmington strain) (14). This raises the question of whether the Breinl and Wilmington typhus group strains synthesize an active MAT enzyme as well as transport SAM. In this study, we constructed an E. coli metK deletion strain that is entirely dependent on SAM transport for growth in order to examine the functionality of rickettsial MAT proteins.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The bacterial strains and plasmids used in this study are described in Table 1. All are derivatives of E. coli strain BW25113 (lacI^q rrnB_T14 ΔlacZ_WJ16 hsdR514 ΔaraBAD_AH33 ΔrhaBAD_LD78) (6). Plasmids containing the metK genes from E. coli, R. prowazekii Madrid E, R. prowazekii Breinl, and R. typhi were constructed using a PCR approach. In an attempt to ensure comparable expression, primers for the amplification of metK coding sequences were engineered to contain the ribosomal binding site of the E. coli metK gene. PCR products were cloned into pBluescript (Stratagene, La Jolla, Calif.) to simplify identification of recombinant plasmids, digested with appropriate restriction enzymes to release the metK fragment, and subsequently ligated into similarly digested pBAD33. This placed each gene under the tight control of the araBAD promoter and an efficient ribosomal binding site. Proper identification and construction of the metK plasmids were confirmed by sequencing the entire open reading frame and ligation junctions of each recombinant. DNA sequencing was performed by the DNA Sequencing and Synthesis Facility, Iowa State University. Growth assays were performed in Luria-Bertani (LB) medium at 37°C. Arabinose at a final concentration of 5 mM was included in the LB medium to ensure expression from the araBAD promoter. When appropriate for selection of individual strains, antibiotics were added to final concentrations of 50 μg/ml for ampicillin, chloramphenicol, and rifampin and 25 μg/ml for kanamycin. Bacterial growth was followed by measuring optical density at 600 nm (OD₆₀₀).

TABLE 1.

Strain and plasmid descriptions

Strain	Plasmid(s)	Resistance markersa	metK location	SAM plasmid
BW25113	pKD46	Amp	Chromosome	−
MOB1469	pKD46, pMW1464	Amp and Cm	Chromosome and plasmid (E. coli)	−
MOB1467	pMW1464	Kan and Cm	Plasmid (E. coli)	−
MOB1473	pMW1464, pMW1402	Kan, Cm, and Amp	Plasmid (E. coli)	+
MOB1490	pMW1402, pMW1484	Amp, Rif, and Kan	None	+
MOB1535	pMW1402, pMW1527	Amp, Cm, and Kan	Plasmid (R. typhi)	+
MOB1507	pMW1402, pMW1504	Amp, Cm, and Kan	Plasmid (R. prowazekii, Breinl)	+
MOB1509	pMW1402, pMW1506	Amp, Cm, and Kan	Plasmid (R. prowazekii, ME)	+
MOB1516	pMW1402, pMW1514	Amp, Cm, and Kan	Plasmid (R. prowazekii, BreinlRepaired)	+
MOB1540	pMW1402, pBAD33	Amp, Cm, and Kan	None	+

^aAmp, Ampicillin; Cm, Chloramphenicol; Kan, Kanamycin; Rif, Rifampin.

Construction of an E. coli metK deletion strain.

Deletion of the E. coli metK gene in strain BW25113 was accomplished using the λ-red recombinase targeted deletion technique and a rescue plasmid, pMW1464, containing the E. coli metK gene. Plasmid pMW1464 was transformed into the BW25113 E. coli strain which contains plasmid pKD46 encoding the λ-red recombinase system. These plasmids were maintained in the newly generated strain (MOB1469) by selection with ampicillin and chloramphenicol. To generate a metK chromosomal deletion, electrocompetent MOB1469 bacteria were electroporated with a PCR product containing a kanamycin resistance gene flanked on one side by a homologous metK sequence and on the other side by a sequence found downstream of the metK coding sequence. The PCR product was generated using forward primer DW598 (5′-CATGGTTTTAGTTGGCGGCGAAATCACCAC CAGCGCCTGGGTAGACATCGAAGAGATCACGTGTAGGCTGGAGCTGCTTC-3′) and reverse primer DW642 (5′-CAGTTGTGATGATAATCTGCGGTGAAACGTGATCAGGAA GAATGATGTTATTGTGCATTCCGGGGATCCGTCGACC-3′). The underlined regions are homologous to plasmid pKD13, which contains the kanamycin resistance gene. By targeting a sequence outside of the metK coding region (DW642), we were able to preclude recombination of our kanamycin resistance cassette into the metK coding sequence found on the plasmid. This resulted in a strain with a deletion of the chromosomal metK gene. The λ-red recombinase pKD46 plasmid was then eliminated by shifting the culture to 37°C, the nonpermissive temperature for pKD46 replication, and selecting for an isolate that was resistant only to chloramphenicol and kanamycin. The rickettsial SAM transporter, contained on plasmid pMW1402, was then introduced into the strain. Finally, the plasmid-located E. coli metK gene was replaced with an incompatible pBAD-based plasmid coding for rifampin resistance (pMW1484). The resulting strain (MOB1490) contained a chromosomal deletion of metK (Kan^r), plasmid pMW1402 encoding SAM transport (Amp^r), and pMW1484, a pBAD derivative plasmid (Rif^r). For assaying the functionality of metK genes, plasmids incompatible with the pBAD rifampin resistance plasmid and containing the metK genes were introduced into the deletion strain via electroporation. Selection for chloramphenicol resistance expressed by the metK plasmids rather than rifampin resistance expressed by the resident plasmid ensured the isolation of a strain containing a metK plasmid.

Western blot analysis.

Bacterial strains were grown to an OD₆₀₀ of 0.3 in LB medium supplemented with 5 mM arabinose. Samples were normalized to an OD₆₀₀ of 1.0 and stored at −80°C. Aliquots (100 μl) were pelleted, washed twice with deionized water, suspended in 100 μl of water, and diluted 1:1 with Laemmli buffer (7). For immunoblot analysis, 30 μl of each experimental sample was loaded onto a 10% polyacrylamide gel and electrophoresed for 3 h at 90 V. Only 3 μl of the positive control was examined due to the intensity of the E. coli metK band when analyzed by immunoblot. It is unknown whether this is due to higher expression of the E. coli protein or the specificity of the E. coli antibody used in the assay. Proteins were transferred to a polyvinylidene difluoride membrane and blocked for 2 h with Tris-buffered saline (TBS) buffer containing 1% casein and 0.1% Tween 20 (Buffer A). Buffer was removed and replaced with 40 ml of Buffer A containing a 1:3,000 dilution of Guinea pig anti-MAT serum generously provided by George Markham. After 16 h of incubation at room temperature, the membrane was washed twice for 10 min with TBS. Anti-guinea pig immunoglobulin G-alkaline phosphatase (AP) conjugate (Sigma, St. Louis, MO) was diluted 1:10,000 in Buffer A. To visualize the markers, an AP-conjugated anti-streptactin antibody (Bio-Rad, Hercules, Calif.) was added at a final dilution of 1:5,000. After 2 h of room-temperature incubation, the membrane was washed twice with TBS and developed using the Bio-Rad AP-Conjugate Substrate kit following the manufacturer's recommendations.

RESULTS

Construction of an E. coli strain that requires SAM.

To measure the functionality of rickettsial MAT in an E. coli background, it was necessary to obtain an E. coli strain devoid of MAT activity. Before deleting the E. coli metK gene of strain BW25113 using the λ-Red recombinase system, this lethal metK mutation was rendered innocuous, first by complementation with a plasmid-borne E. coli metK gene, followed by complementation with a plasmid containing the rickettsial SAM transporter and elimination of the metK plasmid. PCR analysis using primers that flanked the metK gene demonstrated the deletion of the metK gene and replacement by the larger kanamycin resistance gene (Fig. 1). This new strain, MOB1490, requires the addition of SAM (17.5 μM or greater) to the medium for optimum growth (Fig. 2). To our knowledge this is the first complete knockout of the E. coli metK gene. Isolation of an E. coli strain that is dependent on extracellular SAM for growth should be useful in the study of SAM-dependent systems.

FIG. 1.

PCR analysis of the metK gene deletion. Oligonucleotide primers that would amplify the wild-type metK gene and are located outside of the deletion region were used in PCRs with template DNA from both wild-type and mutated strains. Size markers are indicated in kilobases. Amplification of wild-type would yield a 2.0-kb product, while the deletion mutant would yield a larger product of 2.3 kb. Lane 1, wild-type E. coli DNA template; lane 2, metK deletion mutant (MOB 1490) template DNA.

FIG. 2.

Growth dependence of E. coli strain MOB1490 on extracellular SAM. Bacteria were grown in LB medium containing 5 mM arabinose. ▵, No SAM; X, 0.35 μM; ▪, 0.7 μM; ▴, 3.5 μM; ○, 17.5 μM; •, 35 μM. Points are an average of OD₆₀₀ for three independent experimental values at each time point, ± standard errors.

Complementation of an E. coli metK deletion with rickettsial metK genes.

The isolation of an E. coli metK knockout provided an appropriate background to assay the functionality of the metK gene products from the two R. prowazekii strains and the R. typhi strain that exhibited complete open reading frames. The metK genes were amplified from each rickettsial strain and cloned into pBAD33. The same competitive replacement technique used to construct MOB1490 was used to eliminate the resident pBAD plasmid and substitute a plasmid containing a rickettsial metK gene (or E. coli metK as control). A comparison of the rickettsial MAT amino acid sequences is presented in Fig. 3. In order to confirm expression of the rickettsial MAT proteins, a Western blot was performed using antibodies raised against E. coli MAT (kindly provided by George Markham) (Fig. 4). While these antibodies cross-reacted with a number of proteins in the crude preparations, unique proteins not expressed in the strain containing only the vector could be identified. All of the rickettsial full-length MAT proteins were of the predicted size and expressed at comparable levels. After ensuring induction of a rickettsial MAT by the addition of arabinose, the growth of the strains, in the absence of SAM in the medium, was monitored (Fig. 5). The metK genes from R. prowazekii Breinl and from R. typhi Wilmington were able to complement the metK deletion. As might be expected, the Madrid E metK nonsense mutant failed to complement. It was observed that the strain containing the R. prowazekii Breinl metK gene demonstrated a different growth pattern from the R. typhi or E. coli constructs. This may be due to the existence of a small two amino acid deletion in the Breinl MAT (Fig. 3) that was not detected when this gene was originally sequenced (1 and Gregory Dasch, personal communication). To determine if this affected the activity of the enzyme, we constructed a metK gene that contained the upstream portion of Madrid E metK (no deletion) with the downstream portion of Breinl metK (no stop codon). Growth of an E. coli strain containing this recombinant (Breinl_Repaired) was comparable to that of strains expressing the E. coli and R. typhi metK genes (Fig. 5).

FIG. 3.

Sequence comparison of MAT enzymes from R. prowazekii Breinl, R. prowazekii Madrid E, and R. typhi Wilmington. The sequence of Breinl MAT is shown. Amino acid changes found in R. typhi MAT are shown below the Breinl sequence. Differences found in the Madrid E sequence are shown above the Breinl sequence. Stop codons are indicated by an asterisk. The numbering system refers to the R. typhi sequence.

FIG. 4.

Western blot analysis of proteins expressed by the E. coli metK deletion strain containing plasmid-borne rickettsial metK genes. With the exception of the E. coli control lane (lane 1), which contains 10-fold less, each lane contains proteins extracted from an equal number of cells. M, molecular mass markers indicated in kilodaltons. Lane 1, E. coli metK (41.96 kDa); lane 2, vector control; lane 3, R. prowazekii Madrid E metK (15.17-kDa truncated MAT, not retained on gel); lane 4, R. prowazekii Breinl metK (41.91 kDa); lane 5, R. typhi metK (42.34 kDa); lane 6, R. prowazekii Breinl_Repaired metK (42.16 kDa). The location of full-length MAT proteins is indicated by a bracket.

FIG. 5.

Growth of E. coli strain MOB1490 (ΔmetK) and metK-containing strains in the absence of SAM. Strains were grown in LB medium containing 5 mM arabinose. Strains examined were MOB1490 derivatives containing pBAD33 (X) and pBAD33 recombinants expressing E. coli metK (▵), R. prowazekii Breinl metK (□), R. prowazekii Madrid E metK (⧫), R. typhi metK (○), and a repaired R. prowazekii Breinl metK (•). Points are an average of OD₆₀₀ for three independent experimental values at each time point, ± standard errors.

DISCUSSION

The ability to transport the end products of biosynthetic pathways has permitted the rickettsiae to undergo reductive evolution of biosynthetic genes, resulting in relatively small genomes that contain numerous pseudogenes and a high proportion of noncoding DNA. What appeared to be a recent example of this process was the identification of a single nonsense mutation in the R. prowazekii Madrid E metK gene, which encodes the enzyme responsible for the synthesis of the essential metabolite SAM (3). To obtain SAM, the rickettsiae possess a transport system eliminating the need for SAM synthesis (14). However, two of the typhus group rickettsiae, R. prowazekii Breinl and R. typhi Wilmington, possess metK genes that lack the internal termination codon, suggesting that the typhus group rickettsiae retain SAM synthesis capability. This is indeed the case. Both R. prowazekii Breinl and R. typhi MAT enzymes exhibit sufficient activity to complement an E. coli metK deletion strain and support its growth in the absence of SAM. The deletion strain used to assay rickettsial MAT function is, to our knowledge, the first metK knockout in E. coli. An E. coli strain that is dependent upon extracellular SAM for growth should be useful in studies examining SAM-dependent processes.

The sequencing of the R. prowazekii genome and the discovery of numerous pseudogenes, such as the one resulting from a nonsense mutation in the metK gene of the Madrid E strain, led Andersson et al. to propose that this bacterium is undergoing genome degradation (1, 3). In this study we have demonstrated that, as expected, the metK gene of the Madrid E strain codes for a nonfunctional gene product. However, we have also shown that while the metK gene of the Breinl strain has undergone a deletion, the gene product remains functional. These mutations are intermediates in the gene degradation process and exemplify the gradual nature of genome reduction described by Ogata et al. (11).

The conversion of a gene to a pseudogene, followed by its eventual loss from the organism's gene repertoire, is possible only if the gene codes for a product that is no longer required by the cell. In the case of metK and SAM, it is obvious that a SAM transporter must be present before the loss of MAT activity. Our results demonstrate that a fully functional SAM transporter evolved or was acquired prior to loss of SAM synthetic ability. All three strains examined have comparable, high-level SAM transport capability. However, strains that contain complete metK open reading frames (Breinl and R. typhi Wilmington) retain SAM synthetic ability. The appearance of the stop codon in the Madrid E strain and the 6-bp deletion in the Breinl strain, which affected its ability to complement in our assay system, provide experimental support for the hypothesis that this typhus group gene, like the more degenerate spotted fever group orthologs, is in the process of gene degradation.