J Mol Evol (1999) 48:142–150

© Springer-Verlag New York Inc. 1999

Physical and Genetic Map of the Genome of Buchnera, the Primary Endosymbiont of the Pea Aphid Acyrthosiphon pisum Hubert Charles,* Hajime Ishikawa Department of Biological Sciences, Graduate School of Science, University of Tokyo, Hongo, Tokyo 113-0033, Japan Received: 20 May 1998 / Accepted: 1 July 1998

Abstract The genome of Buchnera, an endosymbiotic bacterium of the pea aphid Acyrthosiphon pisum, was characterized by pulse-field gel electrophoresis (PFGE) as a circular DNA molecule of 657 kb. The enzymes I-CeuI, CpoI, ApaI, SmaI, NaeI, SacII, MluI, FspI, and NruI were used to cleave the DNA of Buchnera into fragments of suitable size for PFGE analysis. A physical map of the Buchnera genome, including restriction fragments from seven of these enzymes, was constructed using double cutting, partial digestion, and hybridization with linking fragments, and 29 genes and operons were localized on the map. In addition, the genomic map of Buchnera was compared with those of Escherichia coli and Haemophilus influenzae. The gene order in Buchnera is more similar to that of E. coli than to H. influenzae. The dramatic shrinkage of the Buchnera genome compared with those of other members of the closely related Enterobacteriaceae family is discussed in terms of evolution under the influence of the intracellular symbiotic association. Key words: Buchnera — Acyrthosiphon pisum — Pea aphid — Pulse-field gel electrophoresis — Genome size — Genetic map — Symbiosis

* Present address: Laboratoire de Biologie Appliquee, INSA 406 UAINRA 203 SDI-CNRS 5128, 20 Av. A. Einstein, 69621 Villuerbanne Cedex, France Correspondence to: H. Ishikawa; e-mail: [email protected]

Introduction Buchnera species are intracellular bacteria found in association with most aphids (Homoptera, Aphididae) (Buchner 1965; Baumann et al. 1995). They are found, surrounded by a perisymbiont membrane, in the cytoplasm of bacteriocytes, specialized cells within the body cavity of aphids. Buchnera are Gram-negative cells 1–2 ␮m in diameter and belong to the ␥3 group of the Proteobacteria, closely related to the Enterobacteriaceae family (Munson et al. 1991). They are maternally transmitted to eggs and embryos through host generations (Baumann et al. 1995). Phylogenetic analysis, based on 16S rDNA sequences of Buchnera from numerous aphid species, has revealed that the symbiotic relationship was established 200–250 million years ago and led to cospeciation of the hosts and their endosymbionts (Moran and Baumann 1994). During their cospeciation, both the partners developed intimate genetic and physiological interactions, and the association became so obligate that neither partner could reproduce independently (Houk and Griffith 1980; Ishikawa 1989). Buchnera are implicated in the supplementation of the aphids’ diet, and there is now strong evidence that they provide essential amino acids to host insects (Sasaki and Ishikawa 1995; Douglas 1998). Aphids feed exclusively on phloem sap, a medium rich in carbohydrates but poor in nitrogenous and other compounds. One of the roles of Buchnera could be to equilibrate and maintain the pool of free essential amino acids available for the host (Febvay et al. 1995). The genome of Buchnera seems to have undergone dramatic changes because of their long symbiotic relationship with aphids. The rRNA operon is split

143

into two transcription units, rrs and rrl–rrf, and only one copy of each unit is present in the genome. This arrangement is rare in eubacteria and specific to slow-growing bacteria (Baumann et al. 1995). Genes of the leucine and tryptophan biosynthetic pathways are relocated and amplified in specific plasmids (Bracho et al. 1995; Lai et al. 1994). The G+C content of the Buchnera genome is approximately 30% (Ishikawa 1987), probably due to a strong AT pressure (Sueoka 1988) imposed upon most intracellular bacteria (Heddi et al. 1998; Moran 1996). To date, more than 100 genes of Buchnera have been sequenced, representing about 110 kb of nonredundant sequence in the GenBank database. Most of these sequences are from the work on Buchnera aphidicola (Baumann et al. 1995; Clark et al. 1998), the type species of the genus Buchnera, from the aphid Schizaphis graminum (Munson et al. 1991). The results indicate that the genome of Buchnera contains many essential genes involved in DNA synthesis, transcription and translation, the heat shock response, and various other metabolic pathways. These genes are arranged in a compact genetic organization roughly similar to that of E. coli and H. influenzae (Clark et al. 1998). To characterize the Buchnera genome further, in the present study we estimated the genome size of Buchnera from the pea aphid Acyrthosiphon pisum using pulsefield gel electrophoresis (PFGE). While we previously estimated the genome size of the same Buchnera to be about five times larger than that of E. coli (Ishikawa 1987), the estimate was based on reassociation kinetics of the DNA fragments, an indirect method that is prone to give erroneous results due to contamination by foreign nucleic acids and proteins. One of the purposes of the present study was to examine the previous result employing the direct method. A physical map that indicates restriction sites of seven enzymes was also constructed, and the location of 29 genes and operons was determined.

Materials and Methods Insect Rearing. A long-established parthenogenetic clone of pea aphids, Acyrthosiphon pisum (Harris) was maintained on young broad bean plants, Vicia faba (L.), at 15°C with a 16-h photoperiod. Isolation of Symbionts. Buchnera were purified from pea aphids by an adaptation of the procedures described by Sasaki and Ishikawa (1995) and Heddi et al. (1991). One to five grams of aphids was collected, washed in buffer A [25 mM KCl, 10 mM MgCl2, 250 mM sucrose, 35 mM Tris–HCl, pH 7.5 (Ishikawa 1982)], and lightly crushed in 50 ml of the same buffer. The homogenate was passed through nylon mesh with a pore size of 90 ␮m and centrifuged (1500g, 10 min, 4°C), and the pellet was suspended in 20 ml of buffer A. The suspension was successively filtrated through nylon mesh, with pore sizes of 20 and 10 ␮m, and Isopore membranes, with pore sizes of 5 and 3 ␮m (Milipore Co). The filtrate was centrifuged (1500g, 10 min, 4°C) and the pellet was suspended in 1 to 5 ml of buffer A. The suspension was centrifuged (12,000g, 15 min, 4°C) on a 27–

Fig. 1. PFGE separation of digested genomic DNA of Buchnera. (␭) Molecular weight marker, (1) CpoI, (2) ApaI, (3) SmaI, (4) NaeI, (5) SacII, (6) MluI, (7) NruI and (8) FspI. Pulse time was ramped from 7 to 30 s (A) and 3 to 10 s (B). (C) Conventional electrophoresis was performed at 80 V over 6 h for the separation of small DNA fragments.

70% Percoll gradient [27–70% Percoll in buffer A, 5% PEG 6000, 1% bovine serum albumin (BSA), 1% Ficoll, 250 mM sucrose] in order to eliminate mitochondria (upper phase) and host small nuclei (pellet). The Buchnera cells, localized in a green band, were collected with a glass pipette and washed twice in buffer A. Cells were counted in a Thoma cell, and the concentration was adjusted to 1–2 × 109 cells/ml. After this procedure, less than 1% contaminant particules were observed in the samples. Preparation of High Molecular Weight Genomic DNA. To prepare intact genomic DNA, the cell suspension was mixed with 1 vol of 1% low-melting point agarose (FMC), as described previously (Smith et al. 1988). The agarose plugs were suspended in EC solution [6 mM Tris– HCl (pH 7.6), 100 mM EDTA, 1 M NaCl, 0.5% Brij 58, 0.2% deoxycholate, 0.5% N-lauroylsarcosine, 1 mg/ml lysozyme, 20 ␮g/ml RNAse (DNAse free)] overnight at 37°C, with gentle shaking. The EC solution was then replaced by solution ESP [0.5 M EDTA (pH 8), 1% Nlauroylsarcosine, 1 mg/ml proteinase K] and incubated twice for 24 h at 50°C. The plugs were stored in ESP at 4°C. Restriction Digestion. The agarose plugs containing the genomic DNA were treated overnight with 1 mM phenylmethylsulfonyl fluoride and washed three times (2 n) in at least 10 vol of TE [10 mM Tris–HCl (pH 7.5), 1 mM EDTA]. Portions of the plugs (20 ␮l) were preincubated for 30 min at 4°C in 100 mg/ml BSA and 1 × digestion buffer in a final volume of 100 ␮l. The buffer was then renewed and 2 to 20 U of restriction enzyme was added. The plugs were incubated for 5 h at 30 or 37°C for digestion. PFGE and Southern Hybridization. Electrophoresis was performed in a Contour-clamp Homogeneous Electric Fields Mapper Pulsed Field

144 Table 1. DNA fragments detected by PFGE from the genome of Buchnera, after digestion with selected restriction endonucleases: Fragments are numbered in order of decreasing size (kb) I-CeuI (rrl gene)

CpoI [CGG(A/T)CCG]

ApaI (GGGCCC)

SmaI (CCCGGG)

NaeI (GCCGGC)

SacII (CCGCGG)

MluI (ACGCGT)

NruI (TCGCGA)

FspI (TGCGCA)

670

280 (C1) 266 (C2) 108 (C3)

290 (A1) 230 (A2) 77 (A3) 60 (A4)

174 (Sm1) 174 (Sm2) 145 (Sm3) 98 (Sm4) 45 (Sm5) 11 (Sm6) 6.5 (Sm7)

172 (N1) 112 (N2) 106 (N3) 98 (N4) 72 (N5) 48 (N6) 33 (N7) 11 (N8) 8b

181 (Sa1) 181 (Sa2) 75 (Sa3) 55 (Sa4) 50 (Sa5) 46 (Sa6) 32 (Sa7) 26 (Sa8) 8 (Sa9)

170 (M1) 130 (M2) 96 (M3) 71 (M4) 70 (M5) 47 (M6) 38 (M7) 31 (M8) 3.7 (M9) 3.5c

135 110 52 37 (2) 33 30 27 (2) 23 22 18 (2) 15 (2) 12 (2) 8 7.5 5.5 4.5 3 (2)

670

654d

657

654

652

656

657

83 62 57 53 50 48 45 30 (3)a 25 21 (2) 18 17 16 14 10 6.5 5.5 5 4.5 2 654

654

a

Multiple fragments. leu plasmid. c trp plasmid. d Total size. b

Electrophoresis System (Bio-Rad). The enzyme-digested DNA was separated on 1% agarose gel using 0.5 × TBE [1 × TBE 0.89 M Tris–HCl, pH 8.3, 0.89 M boric acid, 20 mM EDTA] as running buffer at 6 V/cm and various ramping pulse times depending upon experimental conditions. Phage ␭ multimeric DNA (Biolabs) was used as molecular mass marker. For Southern hybridization, the gels were stained first with ethidium bromide and treated three times for 15 min each in solutions A (0.25 M HCl), B (0.5 M NaOH, 1.5 M NaCl), and C [0.25 M Tris–HCl (pH 7.5), 1.5 M NaCl]. Gels were blotted overnight to nylon membranes (Hybond-N; Amersham) in 20 × SSC. DNA was fixed by baking the membranes at 80°C for 2 h. Prehybridization and hybridization were performed at 55 and 42°C, respectively, and the membranes were then washed twice (20 min) at 58°C in 0.5× SSC. The membranes were finally air-dried, exposed to IP films, and analyzed using a BAS 2500 imaging analyzer (Fujifilm).

and Methods. The main contaminant in our sample appeared to be host eukaryotic DNA, which smeared the area around 50 kb, when electrophoresed without any digestion. To test whether this contaminant generated some additional bands in our experiments, we isolated bacteriocytes manually and carefully removed their large nuclei by filtering through a 5-␮m-pore size mesh. Plugs, containing purer Buchnera, were digested with the enzymes FspI and NruI and electrophoresed. As a result, manually isolated Buchnera gave the same electrophoretic pattern as before (data not shown), suggesting that the sample we used was of suitable purity for the present purpose.

Preparation of Probe DNA. For labeling of particular DNA fragments, the fragment required was excised from the agarose gel (low melting point agarose, FMC). DNA was extracted with ␤-agarase I (Biolabs) and concentrated using a 30,000 dialysis membrane (Millipore) with centrifugation at 5000g for 30 min. DNA was then labeled by random priming using the BcaBEST kit (Takara).

Restriction Analysis

Results Purity of Buchnera DNA Buchnera cells were isolated from whole aphids to obtain DNA in sufficient amounts to be compatible with the genome mapping by PFGE, as described under Materials

Buchnera DNA was screened for cleavage with different restriction enzymes to determine the restriction fragment patterns suitable for mapping the genome. The low G+C (30%) of the Buchnera DNA suggests that restriction enzymes with recognition sequences rich in G and C would cleave the genome infrequently. While the restriction enzymes NotI and SfiI, with 8-bp recognition sequences, did not cut the Buchnera chromosome at all, nine other enzymes were found to cut the genome into a number of fragments suitable for analysis (Fig. 1): I-CeuI (1 site), CpoI (3 sites), ApaI (4 sites), SmaI (7 sites), NaeI (8 sites), SacII (9 sites), MluI (8 sites), NruI (23 sites), and FspI (23 sites).

145 Table 2. Double-cutting and partial digestion products of the Buchnera genome generated by the enzymes I-CeuI, CpoI, ApaI, SmaI, NaeI, SacII, and MluI Double-cutting products (hybridized with CpoI or ApaI fragments) CpoI

ApaI

I-CeuI

Linking fragments and partial digestion

A1 A2 A3 A4

240 (C2) + 48 (C1)a 95 (C1) + 108 (C3) + 26 (A2) A3 (C1)c A4 (C1)

— — — —

A1 A2 42 + 35 A4

A1–A2–A3–A4 (P)b

C1 C2 C3

— — —

95 (A2) + 77 (A3) + 60 (A4) + 48 (A1) 240 (A1) + 26 (A2) C3 (A2)

138 + 142 C2 C3

Sm1 Sm2 Sm3 Sm4 Sm5 Sm6 Sm7

20 (C2) + 105 (C3) + 50 (C1) 135 (C2) + 35 (C1) Sm3 (C1) Sm4 (C2) Sm5 (C1) Sm6 (C2) Sm7 (C1)

Sm1 (A2) Sm2 (A1) 77 (A3) + 60 (A4) +7 (A1) Sm4 (A1) 43 (A2) + 4 (?)e 8 (A1) + 3 (?) Sm7 (A1)

Sm1 Sm2 100 + 45 Sm4 Sm5 Sm6 Sm7

Sm1–Sm5 (M3)d Sm2–Sm4 (Sa3) Sm3–Sm7–Sm2 (N5)

N1 N2 N3 N4 N5 N6 N7 N8

N1 (C2) N2 (C1) 97 (C1) + 10 (C3) 64 (C3) + 36 (C2) 64 (C1) + 8 (?) N6 (C2) N7 (C3) N8 (C2)

N1 (A1) 77 (A3) + 35 (A4) + 1 (?) N3 (A2) 92 (A2) + 8 (A1) 48 (A1) + 24 (A4) N6 (A1) N7 (A2) N8 (A1)

N1 72 + 40 N3 N4 N5 N6 N7 N8

N1–N6 (Sm4) N2–N5–N8 (P) N3–N2 (M2) N7–N3 (M3)

Sa1 Sa2 Sa3 Sa4 Sa5 Sa6 Sa7 Sa8 Sa9

95 (C1) + 86 (C2) 108 (C3) + 46 (C1) + 24 (C2) Sa3 (C2) Sa4 (C2) Sa5 (C1) Sa6 (C1) Sa7 (C1) Sa8 (C2) Sa9 (C1)

120 (A1) + 60 (A4) 180 (A2) + 1 (?) Sa3 (A1) Sa4 (A1) Sa5 (A2) 36 (A3) + 10 (A4) Sa7 (A3) Sa8 (A1) 4 (?) + 4 (argA)

Sa1 Sa2 Sa3 Sa4 Sa5 43 + 3 Sa7 Sa8 Sa9

Sa1–Sa3 (P) Sa2–Sa5 (M3) Sa3–Sa4–Sa8 (Sm4) Sa5–Sa9–Sa7–Sa6 (M2)

M1 M2 M3 M4 M5 M6 M7 M8 M9

M1 (C2) M2 (C1) 68 (C1) + 28 (C3) M4 (C1) 36 (C2) + 34 (C3) 38 (C2) + 9 (C1) M7 (C2) M8 (C3) M9 (C1)

M1 (A1) 77 (A3) + 37 (A2) + 17 (A4) M3 (A2) 40 (A4) + 35 (A1) 60 (A2) + 10 (?) M6 (A1) M7 (A1) M8 (A2) M9 (A1)

M1 70 + 60 M3 M4 M5 M6 M7 M8 M9

M4–M9–M6 (N5) M8–M3–M2–M4 (P)

a

A1, digested by CpoI, generates two fragments (240 and 48 kb) that hybridize with C2 and C1, respectively. The order A1–A2–A3–A4 was determined by partial digestion with ApaI. c A3 is not cleaved by CpoI and hybridizes with C1. d Sm2 and Sm5 are linked by the M3 fragment. e (?) Theoretical fragment not revealed by Southern blotting. b

The restriction fragments were numbered on the basis of size in descending order, except for NruI and FspI, as summarized in Table 1. Fragments of the same length, i.e., Sm1-2 and Sa1-2, were arbitrarily numbered and distinguished from each other by hybridization and double cutting. Different ramping pulse and electrophoresis times were used so that each restriction fragment

size could be determined under the optimal electrophoresis conditions. Smaller fragments were separated by conventional electrophoresis using a large amount of DNA in each lane (Fig. 1C). The average genome size, presumed on the sum of the sizes of the restriction fragments, was between 652 and 657 kb. The size of the linearized chromosome of Buch-

146 Table 3.

Probes used to locate genes on the physical map of the genome of Buchnera

Probe

Gene(s)

Fragments hybridized

Ref.a

TE PL potsk1 KJ4 prH2 16S 23S ftsH ftsZ F1 F2 F3 F4 F5 F6 F7

trpE, trpG (anthranilate synthase) leuA (2-isopropylmalate synthase) groES, groEL (Hsp10 and Hsp60) dnaK, dnaJ (Hsp70 and DnaJ) rpoH (heat shock sigma factor 32) rrs (16S ribosomal RNA) rrl (23S ribosomal RNA) ftsH (cell cycle FtsH protein) ftsZ (cell septation protein) thrB (homoserine kinase) argA (N-acetylglutamate synthase) sucA (s-oxyglutarate dehydrogenase) gyrB (DNA gyrase subunit B) sodA (superoxide dismutase) atpE (ATP synthase c chain) ribH (riboflavin synthase ␤ chain)

1 2 3 4 5 6 7 8 9 10 10 11 12 11 12 13

F8 F9 F10 F11 F12 F13 F14 F15 F16

deoB (Phosphopentomutase) tgt (queuine tRNA-ribosyltransferase) pth (peptidyl-tRNA hydrolase) rpsH (30S ribosomal protein S8) rpsM (30S ribosomal protein S13) yche (19.7-kDa hypothetical protein) yheL (10.7-kDa hypothetical protein) b2511 (hypothetical GTPb protein) ychF (hypothetical GTPb protein)

3.5-kb trp plasmid (MluI) 8-kb leu plasmid (NaeI) C1, A1, Sm2, N5, Sa1, M4 C2, A1, Sm2, N1, Sa3, M1 C1, A1, Sm2, N5, Sa1, M9 C2, A1–2, Sm6, N4, Sa2, M5 C1, A3, Sm3, N2, Sa6, M2 C1, A2, Sm1, N3, Sa2, M3 C2, A1, Sm4, N6, Sa4, M1 C2, A1, Sm4, N1, Sa4, M1 C1, A3, Sm5, N2, Sa9, M2 C3, A2, Sm1, N4, Sa2, M5 C1, A1, Sm2, N5, Sa1, M4 C2, A1, Sma4, N1, Sa4, M1 C1, A1, Sm2, N5, Sa1, M4 C1, A4, Sm2, N2, Sa1, M2, C1, A1, Sm3, N5, Sa1, M4 C1, A3, Sm3, N2, Sa6, M2 C2, A1, Sm2, N1, Sa3, M1 C2, A1, Sm4, N1, Sa4, M1 C1, A3, Sm3, N2, Sa6, M2 C1, A3, Sm3, N2, Sa6, M2 C2, A2, Sm1, N4, Sa2, M5 C1, A3, Sm3, N2, Sa6, M2 C2, A1, Sm7, N5, Sa1, M4 C2, A1, Sm4, N6, Sa4, M1

11 11 11 11 11 11 11 11 11

a

1, Lai et al. (1994); 2, Silva (submitted for publication); 3, Ohtaka et al. (1992); 4, Sato and Ishikawa (1997a); 5, Sato and Ishikawa (1997b); 6, Unterman et al. (1989); 7, Rouhbakhsh and Baumann (1995); 8, Sato (personal communication); 9, Baumann and Baumann (1998); 10, Nakabachi and Ishikawa (1997); 11, Nakabachi (personal communication); 12, Clark et al. (1998); 13, Nakabachi and Ishikawa (1998).

nera, estimated after I-CeuI digestion, was about 670 kb. This overestimation could be explained by the high A+T content of the Buchnera DNA, which interferes with the PFGE mobility of large DNA fragments, as described previously for Mycoplasma sp. (Pyle et al. 1988).

plasmid. The overall size of this plasmid, i.e., the number of tandem repeats, was not determined in this study, but was previously reported to vary between 18.5 and 37 kb (5 to 10 tandem repeats) in Buchnera from A. pisum (Rouhbakhsh et al. 1996).

Chromosomal Shape and Extrachromosomal Elements

Physical Mapping of the Buchnera Genome

Partial digestion with the enzymes CpoI (three sites) and ApaI (four sites) produced 7 and 13 fragments, respectively, suggesting an endless restriction pattern and a circular shape for the chromosome of Buchnera. A circular chromosome was also suggested by the single fragment generated by the enzyme I-CeuI that has one site (data not shown). When electrophoresis was performed without any endonuclease digestion, no extrachromosomal band was visualized on the gels after staining by ethidium bromide. In the meantime, the two plasmids containing trpEG (Rouhbakhsh et al. 1996) and leuABCD genes Bracho et al. 1995), which were previously detected in Buchnera, were identified by Southern blotting using specific probes (Table 3). The leu plasmid was linearized by the enzyme NaeI, giving rise to a single band at 8 kb on the gel (Table 1). The trp plasmid was cut by the enzyme MluI within the trpE gene, resulting in a single fragment of 3.5 kb on the gel (Table 1), which corresponded to a unit of the trpEG tandem repeat of this

We first constructed a physical map of the Buchnera genome in terms of restriction sites of the three enzymes I-CeuI, CpoI, and ApaI that generate very few restriction fragments (i.e., one, three, and four, respectively). The three CpoI fragments were arbitrarily assigned in the order, C1–C2–C3, with zero positioned at the ICeuI restriction site, as determined by double digestion with CpoI and I-CeuI. Partial digestion with ApaI and double digestion with ApaI and CpoI allowed us to map the four ApaI restriction fragments on the CpoI map. Restriction fragments from other enzymes, SmaI, NaeI, SacII, and MluI, were then placed on the CpoI map using the double-digestion technique and hybridization with overlapping C1, C2, and C3 fragments as probes. The same experimental approach, i.e., double digestion and hybridization with A1, A2, A3, and A4, was also adopted to place the same fragments on the ApaI map. However, it was impossible to assign all the frag-

147

Fig. 2. Physical and genetic map of the Buchnera genome using the enzymes I-CeuI, CpoI, ApaI, SmaI, NaeI, SacII, and MluI. Restriction fragments are numbered on the basis of size (Table 1). Genes marked with an asterisk were used as probes in this study (Table 3). Other genes were placed based on the previous results (Baumann et al. 1995; Clark et al. 1998) on Buchnera from the aphid Schizaphis

graminum. The gene rpoH has been attached to the rho-39kDa fragment, about 6 kb downstream from groEL (Sato and Ishikawa 1997b). Two sets of fragments were found to hybridize with the ribH probe, suggesting the presence of two copies of this gene in the Buchnera genome, although nonspecific hybridization may also occur. Restriction sites: C, CpoI; A, ApaI; Sm, SmaI; N, NaeI; Sa, SacII; M, MluI.

ments unambiguously, particularly those with no CpoI and ApaI site, in this manner. Analysis of the partial digestion products generated by the four enzymes, SmaI, NaeI, SacII, and MluI, and hybridization with unassigned fragments as probes were performed to surmount this difficulty. Finally, some specific linking fragments, i.e., Sm4, Sa3, N4, N5, M2, and M3, were

used as probes to complete the physical map (Table 2). The chromosomal locations of 29 genes and operons, listed in Table 3, were placed on the map by Southern hybridization. Results of hybridization were also used to confirm the constructed physical map. The physical and genomic map of the Buchnera genome is summarized in Fig. 2.

148

Fig. 3. Comparison of the genetic map of Buchnera with that of E. coli (A) and H. influenzae (B). Arrow origins indicate the location of Buchnera genes and arrow points the homologues of E. coli or H. influenzae. The locations of E. coli genes are given as minutes, as referred to in the E. coli database of the NCBI FTP server (ncbi.nlm. nih.gov). The locations of homologous geneswere compared

in each pair of concentric circles on a scale of 100 U of the total length. The solid lines (and the boldface genes) correspond to pairs of homologous genes whose locations are different from each other by more than 20 U. The maps were synchronized by iterative simulations in order to minimize the number of boldface genes in A and B, independently.

Discussion

cation since its divergence from Buchnera seems unlikely, as has been denied in explaining the difference in genome size between the more distantly related H. influenzae and E. coli (Watanabe et al. 1997). Consistent with these results, genome shrinkage has also been reported for symbiotic bacteria of the weevil Sitophilus oryzae (Charles et al. 1997) and the ciliate Parauronema acutum (Soldo et al. 1983), and for the cyanelle of Cyanophora paradoxa (Herdman and Stanier 1977), as well as parasitic bacteria such as Mycoplasma, Chlamydia, and Rickettsia (Krawiec and Riley 1990). To explain the genome shrinkage of intracellular bacteria, an old rule of evolutionary biology and natural selection, ‘‘use it or lose it,’’ seems plausible (Maniloff 1996). The three smallest of the fully sequenced genomes to date are Mycoplasma genitalium, M. pneumoniae, and Borrelia burgdorferi, all of which are parasitic bacteria. These genomes equally lack genes necessary for biosynthesis of amino acids, fatty acids, nucleotides, etc., and in addition, none of the enzymes of the citric acid cycle are found in these bacteria (Fraser et al. 1997). If an organism enjoys an abundant supply of nutrients from its host, then it seems predictable that the machinery responsible for their synthesis would decay. As a result, the genome would tend to lose the relevant gene sequences, leading to reduction of the genome size. However, this could not be the case with Buchnera. Buchnera species are quite different from the parasitic bacteria in that these endosymbionts not only retain the

The genome of an endosymbiotic bacterium, Buchnera, from the pea aphid A. pisum was characterized by PFGE to be a circular DNA molecule of 657 kb. Assuming a mean gene length of 1 kb and the proportion of coding DNA as 90% (Clark et al. 1998), the total number of Buchnera genes can be estimated at about 600. The genome of Buchnera is hence similar in size to that of Mycoplasma genitalium (580 kb), the smallest freeliving bacteria described to date (Fraser et al. 1995). Buchnera is closely related to members of the Enterobacteriaceae family, whose genome sizes are usually between 4 and 5 Mb. Clark et al. (1998) have previously noted a good conservation of gene arrangement between small DNA fragments from Buchnera and E. coli. Construction of the genetic map of Buchnera allowed us to extend this observation to the whole genome. Although the genome of E. coli is about seven times larger than that of Buchnera, it was possible to correlate the two genetic maps, at least partially (Fig. 3A). In contrast, this was not possible with the maps of Buchnera and H. influenzae, a more distantly related ␥-proteobacterium (Fig. 3B). These results suggest that the Buchnera genome could have evolved by deletion of large DNA regions from an ancestor that possessed a genomic organization similar to that of the extant Enterobacteriaceae. The alternative hypothesis that a small ancestral genome of E. coli could have increased in size by genome dupli-

149

ability to synthesize amino acids, but also provide them to the host insects (Sasaki and Ishikawa 1993, 1995; Douglas 1998). Moreover, differential cDNA display suggests that in Buchnera many other metabolic pathways, including the citric acid cycle, are active as well (Nakabachi and Ishikawa 1997). Thus, it is intriguing to know what caused the genome shrinkage in Buchnera and which kinds of genes are missing from their genome. One characteristic feature of Buchnera is an overall A and T enrichment of their genomic DNA (Ishikawa 1987; Ohtaka and Ishikawa 1993). Genetic recombinations are strongly limited in Buchnera populations, which are isolated in the host bacteriocytes and pass through a bottleneck with each new host generation. In addition, the effects of mutational bias could be enhanced by the diminishing selective pressure in the intracellular environment. A combination of these factors, increasing the substitution rate in both coding and noncoding DNA, could be partially responsible for the overall A and T accumulation in the Buchnera genome (Moran 1996). A negative correlation between genome size and AT content has been described in many bacterial genomes in general (Heddi et al. 1998) and also in pseudogenes of eukaryotes that are smaller and more AT rich than their functional homologues (Gu and Li 1995; Mouchiroud, personal communication). Assuming that the loss of DNA stability due to AT accumulation tends to favor the deletion process rather than insertion, the reduction in size of the Buchnera genome could be considered a neutral consequence of the accumulation of A and T. Acknowledgments. We would like to thank Drs. T. Kuroiwa, T. Sasaki, and Y. Rahbe for their cooperation during the study. We are also grateful to Drs. S. Sato and F. Silva and Mr. A. Nakabachi for providing probes of Buchnera. This work was supported by a grant from the Program for Promotion of Basic Research Activities for Innovation Biosciences (ProBRAIN) of the Bio-oriented Technology Research Advancement Institution.

References Baumann L, Baumann P (1998) Characterization of ftsZ, the cell division gene of Buchnera aphidicola (Endosymbiont of aphids) and detection of the product. Curr Microbiol 36:85–89 Baumann P, Baumann L, Lai C-Y, Rouhbakhsh D, Moran N, Clark MA (1995) Genetics, physiology, and evolutionnary relationships of the genus Buchnera: Intracellular symbionts of aphids. Annu Rev Microbiol 49:55–94 Bracho AM, Martinez-Torres D, Moya A, Latorre A (1995) Discovery and molecular characterization of a plasmid localized in Buchnera sp. bacterial endosymbiont of the aphid Rhopalosiphum padi. J Mol Evol 41:67–73 Buchner P (1965) Endosymbiosis of animals with plant microorganisms. Interscience, New York Charles H, Condemine G, Nardon C, Nardon P (1997) Genome size characterization of the principal endocellular symbiotic bacteria of the weevil Sitophilus oryzae, using pulse field gel electrophoresis. Insect Biochem Mol Biol 27:345–350

Clark MA, Baumann L, Baumann P (1998) Sequence analysis of a 34.7kb DNA segment from the genome of Buchnera aphidicola (endosymbiont of aphids) containing groEL, dnaA, the atp operon, gidA and rho. Curr Microbiol 36:158–163 Douglas AE (1998) Nutritional interactions in insect-microbial symbiosis: Aphids and their symbiotic bacteria. Annu Rev Entomol 43:17–37 Febvay G, Liadouze I, Guillaud J, Bonnot G (1995) Analysis of energetic amino acid metabolism in Acyrthosiphon pisum: A multidimensional approach to amino acid metabolism in aphids. J Insect Biochem Physiol 29:45–69 Fraser CM, Gocayne JD, White O, Adams MD, Clayton RA, Fleishmann RD, Bult CJ, Kerlavage AR, Sutton GG, Kelley JM, et al. (1995) The minimal gene complement of Mycoplasma genitalium. Science 270:397–403 Fraser CM, Casjens S, Huang WM, Sutton GG, Clayton RA, Lathigra R, Ketchum KA, Dodson R, Hickey EK, et al. (1997) Genomic sequence of a Lyme disease spirochete, Borrelia burgdorferi. Nature 390:580–586 Gu X, Li WH (1995) The size distribution of insertions and deletions in human and rodent pseudogenes suggest the logarithmic gap penalty for sequence alignment. J Mol Evol 40:464–473 Heddi A, Lefebvre F, Nardon P (1991) The influence of symbiosis on the respiratory control ratio (RCR) and the ADP/O ratio in the adult weevil Sitophilus oryzae (Coleoptera, Curculionidae). Endocytobios Cell Res 8:61–73 Heddi A, Charles H, Khatchadourian C, Bonnot G, Nardon P (1998) Molecular characterization of the principal symbiotic bacteria of the weevil Sitophilus oryzae: A peculliar G+C composition of an endocytobiote DNA. J Mol Evol (in press) Herdman M, Stanier RY (1977) The cyanelle: Chloroplast or endosymbiotic prokaryote? FEMS Microbiol Lett 1:7–12 Houk EJ, Griffith GW (1980) Intracellular symbiotes of Homoptera. Annu Rev Entomol 25:161–187 Ishikawa H (1982) Isolation of the intracellular symbionts and partial characterizations of their RNA species of the elder aphid, Acyrthosiphon magnoliae. Comp Biochem Physiol 72B:239–247 Ishikawa H (1987) Nucleotide composition and kinetic complexity of the genomic DNA of an intracellular symbiont in the pea aphid Acyrthosiphon pisum. J Mol Evol 24:205–211 Ishikawa H (1989) Biochemical and molecular aspects of endosymbiosis in insects. Int Rev Cytol 116:1–45 Krawiec S, Riley M (1990) Organization of the bacterial chromosome. Microbiol Rev 54:502–539 Lai C-Y, Baumann L, Baumann P (1994) Amplification of trpEG: Adaptation of Buchnera aphidicola to an endosymbiotic association with aphids. Proc Natl Acad Sci USA 91:3819–3823 Maniloff J (1996) The minimal cell genome: ‘‘On being the right size.’’ Proc Natl Acad Sci USA 93:10004–10006 Moran N (1996) Accelerated evolution and Muller’s rachet in endosymbiotic bacteria. Proc Natl Acad Sci USA 48:644–649 Moran N, Baumann P (1994) Phylogenetics of cytoplasmically inherited microorganisms of arthropods. Trend Ecol Evol 9:15–20 Munson MA, Baumann P, Kinsey MG (1991) Buchnera gen. nov. and Buchnera aphidicola sp. nov., a taxon consisting of the mycetocyteassociated, primary endosymbionts of aphids. Int J Syst Bacteriol 41:566–568 Nakabachi A, Ishikawa H (1997) Differential display of mRNAs related to amino acid metabolism in the endosymbiotic system of aphids. Insect Biochem Mol Biol 27:1057–1062 Nakabachi A, Ishikawa H (1998) Provision of riboflavin to the host aphid, Acyrthosiphon pisum, by endosymbiotic bacteria, Buchnera. J Insect Physiol (in press) Ohtaka C, Ishikawa H (1993) Accumulation of adenine and thymine in a groE-homologous operon of an intracellular symbiont. J Mol Evol 36:121–126 Ohtaka C, Nakamura H, Ishikawa H (1992) Structure of chaperonins

150 from an intracellular symbiont and their functional expression in Escherichia coli groE mutants. J Bacteriol 174:1869–1874 Pyle L, Corcoran LN, Cocks BG, Bergemann AD, Whitley JC, Finch LR (1988) Pulsed-field electrophoresis indicates larger than expected sizes for Mycoplasma genomes. Nucleic Acid Res 16:6015– 6024 Rouhbakhsh D, Baumann P (1995) Characterization of a putative 23S5S rRNA operon of Buchnera aphidicola (a prokaryotic endosymbiont of aphids) unlinked to the 16S rRNA-encoding gene. Gene 155:107–112 Rouhbakhsh D, Lai C-Y, von Dohlen CD, Clark MA, Baumann L, Baumann P, Moran NA, Voegtlin DJ (1996) The tryptophan biosynthetic pathway of aphid endosymbionts (Buchnera): Genetics and evolution of plasmid-associated anthranilate synthase (trpEG) within the Aphididae. J Mol Evol 42:414–421 Sasaki T, Ishikawa H (1993) Nitrogen recycling in the endosymbiotic system of the pea aphid, Acyrthosiphon pisum. Zool Sci 10:779– 785 Sasaki T, Ishikawa H (1995) Production of essential amino acids from glutamate by mycetocyte symbiont of the pea aphid, Acyrthosiphon pisum. J Insect Physiol 41:41–46

Sato S, Ishikawa H (1997a) Structure and expression of the dnaKJ of Buchnera, an intracellular symbiotic bacteria of aphids. J Biochem 122:41–48 Sato S, Ishikawa H (1997b) Expression and control of an operon from an intracellular symbiont which is homologous to the groE operon. J Bacteriol 179:2300–2304 Smith CL, Klco SR, Cantor CR (1988) Pulsed-field gel electrophoresis and the technology of large DNA molecules. In: Davies KE (ed) Genome analysis: A practical approach. IRL Press, Oxford, p 41 Soldo AT, Brickson SA, Larin F (1983) The size and structure of the genome of symbiote xenosome particule in the ciliate Parauronema acutum. J Gen Microbiol 129:1317–1325 Sueoka N (1988) Directional mutation pressure and neutral molecular evolution. Proc Natl Acad Sci USA 85:2653–2657 Unterman BM, Baumann P, McLean DL (1989) Pea aphid symbiont relationships established by analysis of 16S rRNAs. J Bacteriol 171:2970–2974 Watanabe H, Mori H, Itoh T, Gojobori T (1997) Genome plasticity as a paradigm of eubacterial evolution. J Mol Evol 44 (Suppl 1):S57– S64