High-Resolution Physical Map of the Sinorhizobium meliloti 1021 pSyma Megaplasmid

Abstract

To facilitate sequencing of the Sinorhizobium meliloti 1021 pSyma megaplasmid, a high-resolution map was constructed by ordering 113 overlapping bacterial artificial chromosome clones with 192 markers. The 157 anonymous sequence tagged site markers (81,072 bases) reveal hypothetical functions encoded by the replicon.

The symbiotic soil bacterium Sinorhizobium meliloti forms nitrogen-fixing nodules on the roots of leguminous host plants and displays a complex genome consisting of a 3.7-Mb chromosome and two megaplasmids, pSyma (1.4 Mb) and pSymb (1.7 Mb) (8, 27, 41). Genes required for symbiosis are located on all three replicons (18, 20), but are more frequently found on the megaplasmids. Genes involved in nodulation and nitrogen fixation are located on pSyma (21, 22, 39), whereas those essential for extracellular polysaccharide synthesis and other symbiotic functions are located on pSymb (16, 44). These two genetic elements have both chromosome-like and plasmid-like features: both are 3 orders of magnitude larger than many cloning vectors and carry some copies of housekeeping genes, such as groESL, and genes associated with other metabolic functions (33). On the other hand, the megaplasmids can be mostly or entirely cured without affecting growth and reproduction (at least in permissive conditions) (13, 24; M. Hynes, personal communication). Moreover, the megaplasmids can be transferred to and maintained in at least one heterologous genus, Agrobacterium (25, 43). Maps for the chromosome and pSymb of strain 1021 exist (9, 10, 12, 20, 23), but concerning pSyma, only three markers outside the 250-kb region containing symbiotic genes have been identified (3). As part of the international effort to sequence the entire S. meliloti genome, we constructed a high-resolution physical map of the pSyma megaplasmid, using PCR-based screening and assembly of recombinant bacterial artificial chromosome (BAC) clones. In addition to providing a valuable tool for the total genome sequencing project, the data reported here provide new insights into the genetic information contained on pSyma.

A high-resolution map of the S. meliloti pSyma megaplasmid was constructed by using the same materials and methods that were successfully used for chromosome mapping (9) except that additional random sequences from a pSyma-enriched library were incorporated into the BAC screening (methods are described at http://cmgm.stanford.edu/∼mbarnett/syma.htm). After screening 192 megaplasmid clones, we identified 88 pSyma clones. Additional clones from the total genomic BAC library were similarly screened to fill gaps in poorly represented regions of the pSyma contig. Thus, we assembled 113 BAC clones into a circular contig (Fig. 1A) encompassing the entire 1.4-Mb pSyma replicon, using a total of 192 markers, including 157 sequence tagged sites (STSs) (9) and 33 gene markers (Table 1) representing 14 individual genes, 15 operons (52 genes), and four insertion sequences available in the GenBank (7) and EMBL (42) databases. Assuming a random distribution of markers, average spacing was estimated at 7 kb, with a tiling path of 6.8 colinear BAC clones per marker. No region is represented by only one BAC, and we detected only five chimeric clones between pSyma and pSymb. All of the 108 other assembled BAC clones show an exact colinear distribution of markers, and pSyma is covered by a set of 19 BAC clones with minimal overlap. Assuming a map density of 7 kb, deletions or rearrangements on some clones should be smaller than 7 kb, if they do exist.

FIG. 1.

(A) High-resolution map of the pSyma megaplasmid of S. meliloti 1021. The map is presented in three linear and contiguous parts of approximately 500 kb for convenience. Identified S. meliloti genes (genetic database or BLASTX results) are indicated on the left side while anonymous STS markers are located on the right side. The positions of underlined genes were deduced from mapped genes in the operon; i.e., no PCR primers were designed. Some genes are listed more than once because several sets of primers were used. Black rectangles indicate pairwise invertable markers. ‡, partial similarity. The minimum set of BAC clones covering the replicon is also presented. (B) Simplified map oriented according to the Honeycutt et al. map (23) showing STS markers mentioned in the text, selected genes, and the minimum set of BAC clones. Lengths of BAC inserts are shown relative to the sizes determined by field inversion gel electrophoresis. Genes previously mapped by Honeycutt et al. are marked with asterisks. Corresponding BAC insert sizes are as follows: BAC01, 110 kb; BAC02, 75 kb; BAC03, 110 kb; BAC04, 75 kb; BAC05, 55 kb; BAC06, 85 kb; BAC07, 120 kb; BAC08, 65 kb; BAC09, 80 kb; BAC10, 75 kb; BAC11, 60 kb; BAC12, 110 kb; BAC13, 100 kb; BAC14, 140 kb; BAC15, 80 kb; BAC16, 25 kb; BAC17, 100 kb; BAC18, 80 kb; and BAC19, 60 kb.

TABLE 1.

Previously identified S. meliloti genes mapped on the pSyma megaplasmid

Gene(s)	Encoded function or product
adhA	Alcohol dehydrogenase
fixABCX	Putative electron transport chain to nitrogenase
fixGHIS	Putative cation transport complex
fixJ2T2-fixK2	Transcriptional activators
fixKorf151	Transcriptional activator
fixLJT1	Hemoprotein kinase; transcriptional activator
fixNOQP	Putative bacteroid oxidase
groESLa	Chaperonin
nifABfdxNfixU	Nitrogen fixation regulatory protein; ferredoxin-like protein
nifHDKE	Nitrogenase reductase
nifN	FeMo-cofactor biosynthesis
nodABCIJ	Acyltransferase; N acetylase; chitin synthase; transporter of nod factors
nodD1	nod gene activator
nodD2	nod gene activator
nodD3	nod gene activator
nodFE	Acyl carrier protein; β-ketoacyl synthase
nodG	Putative dehydrogenase
nodH	Sulfotransferase
nodLnoeAB	O-Acetyltransferase
nodMnolFGHInodN	Glucosamine synthase; transport
nodPQ	ATP sulfurylase APS kinase
nolQS	Unknown function; similar to a thiamine biosynthetic enzyme
nosRZDFY	Nitrous oxide reduction proteins
maturase	Reverse transcriptase/maturase
rhbF	Siderophore biosynthesis in rhizobactin regulon
syrA	Increases exopolysaccharide abundance
syrB	Negatively affects syrM expression
syrM	nod gene activator

The relative positions of the nodulation and nitrogen fixation genes are consistent with previous mapping data (22, 23, 39), in particular the (i) presence of two fixJ loci flanking the nod-nif region (5, 35), (ii) organization of nodulation genes (39), (iii) orientation between nos and fix gene clusters (11), (iv) location of syrA (4), and (v) location of groESLa (33). All of these genes are clustered in a well-known symbiotic region. Based on the insert size of the clones covering this region, the total length was estimated to be between 250 and 300 kb, which is also in agreement with Renalier et al. (35). The remaining 1.1 Mb of the replicon does not contain any known symbiotic genes, except syrB. We also positioned several previously unmapped genes: adhA, rhbF, and a gene encoding a maturase. Concerning insertion sequences, we detected one copy of ISRm1 (46), at least five copies of the widespread ISRm2011-2 (40), five copies of ISRm3 (45), and one copy of ISRm5 (28), compared with five copies on the chromosome (9). We did not obtain PCR products with primers designed from ISRm2, ISRm4, ISRm6, ISRm7, ISRm8, and ISRm9.

Each STS was analyzed by BLASTX (1) comparison with the nonredundant protein database from the National Center for Biotechnology Information, and results are available at the website http://www-recomgen.univ-rennes1.fr/meliloti. STS match results were divided into four categories (Fig. 2A), and the most significant homologies were divided into functional groups according to Riley's classification for orthologous Escherichia coli genes (36, 37) (Fig. 2B).

FIG. 2.

(A) Distribution of BLASTX results among four categories of significance. 1, strong similarity with S. meliloti proteins (E ≤ 1e⁻⁶; identity, ≥85%); 2, strong similarity with sequences available in the databases (E ≤ 1e⁻⁶); 3, local or weaker similarity with sequences available in the databases (1e⁻² ≤ E ≤ 1e⁻⁶); 4, no similarity with sequences available in the databases. (B) Classification of the most significant matches (E ≤ 1e⁻⁶) using Riley's classification (34, 35).

This distribution shows many STS markers containing genes involved in the metabolism of small molecules, such as (i) 4-deoxy-l-threo-5-hexosulose-uronate-ketol-isomerase and succinate-semialdehyde dehydrogenase (encoded by gabD), which is involved in carbohydrate (C₄-to-C₆) degradation (32), and (ii) serine hydroxymethyltransferase, which is the key enzyme of C₁ and C₂ compound assimilation and is necessary for the formation of effective nodules in Bradyrhizobium japonicum (38). The next largest group of STSs is made up of those possibly involved in cellular processes (chemotaxis and transport); no matches with cell division proteins or general housekeeping genes were detected other than those for the previously reported groESLa (33). We also found matches to genes involved in nitrogen metabolism: the periplasmic nitrate reductase precursor of Paracoccus and Pseudomonas, the NifX-like protein of Rhizobium sp. strain NGR234, an arginine deiminase of Rhizobium etli (15), and the NifL nitrogen fixation regulator of Klebsiella and other bacteria (31), previously unknown in S. meliloti. Also, there are several less stringent intriguing matches: an STS identical to stage IV sporulation protein FB of Bacillus subtilis, required for spore formation (14), and a marker identical to protein AttB of the plant pathogen Agrobacterium tumefaciens, required for the attachment of bacteria to plant cells (30). One STS is similar to VirB4 from Agrobacterium and TraB from E. coli, both of which are required for DNA transfer and may represent part of a region involved in conjugative transfer of the pSyma replicon. We also detected one sequence similar to the adducin-like protein AddA of the obligate intracellular parasite Rickettsia prowazekii and to alpha-adducin, which promotes the assembly of the spectrin-actin network in eukaryotic cells (2, 26). In addition, some STSs have matches to transcriptional regulators from the LysR family of transcriptional regulators, the AraC family activators, the GntR family regulators, the trp repressor, and a repressor of the TetR-AcrR family. We did not find any STSs with matches to regulators of two-component systems.

The most relevant comparison for the pSyma sequence will be with that of the closely related Rhizobium sp. strain NGR234, which has a complex genome (17), including a symbiosis plasmid of 536 kb, for which the complete nucleotide sequence has been established (19). Given that the pSyma megaplasmid of S. meliloti is almost three times the size of the pSym megaplasmid of NGR234, it will be interesting to determine how related they are. In this regard, we noted that seven of the S. meliloti pSyma STS markers had a match with the pSym of NGR234, while 150 of the S. meliloti pSyma markers did not (for E ≤ 1e⁻⁴).

The elucidation of the S. meliloti total genome sequence will aid greatly our understanding of the ancestry and behavior of the pSyma replicon as well as provide insight into the genomic plasticity, the presence of multicopy genes, and the relative involvement of each replicon, both in symbiotic and free-living bacteria.