APPLIED AND ENVIRONMENTAL MICROBIOLOGY,0099-2240/00/$04.0010

Oct. 2000, p. 4503–4509 Vol. 66, No. 10

Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Detection and Identification of Bacterial Endosymbionts inArbuscular Mycorrhizal Fungi Belonging to the

Family GigasporaceaeVALERIA BIANCIOTTO,1 ERICA LUMINI,2 LUISA LANFRANCO,2 DANIELA MINERDI,2

PAOLA BONFANTE,1,2* AND SILVIA PEROTTO1

Centro Studio Micologia del Terreno-CNR1 and Dipartimento di Biologia Vegetale dell’Universita,2 10125 Turin, Italy

Received 5 April 2000/Accepted 17 July 2000

Intracellular bacteria have been found previously in one isolate of the arbuscular mycorrhizal (AM) fungusGigaspora margarita BEG 34. In this study, we extended our investigation to 11 fungal isolates obtained fromdifferent geographic areas and belonging to six different species of the family Gigasporaceae. With theexception of Gigaspora rosea, isolates of all of the AM species harbored bacteria, and their DNA could be PCRamplified with universal bacterial primers. Primers specific for the endosymbiotic bacteria of BEG 34 couldalso amplify spore DNA from four species. These specific primers were successfully used as probes for in situhybridization of endobacteria in G. margarita spores. Neighbor-joining analysis of the 16S ribosomal DNAsequences obtained from isolates of Scutellospora persica, Scutellospora castanea, and G. margarita revealed asingle, strongly supported branch nested in the genus Burkholderia.

Arbuscular mycorrhizal (AM) fungi are obligate biotrophsthat belong to the order Glomales and develop in close rela-tionship with the roots of about 80% of land plants. Fossil andmolecular data have demonstrated that AM fungi are veryancient, dating back to 350 to 400 million years ago (28, 30).The success of AM fungi in evolution is mainly due to theircentral role in the capture of nutrients from the soil (29).Despite recent breakthroughs in our knowledge of the molec-ular basis of plant-fungus interactions (1, 12), many aspects ofthe biology of AM fungi, particularly their genomes, are stillobscure due to their biotrophic status, their multinuclear con-dition, and an unexpected level of genetic variability (13, 15,17).

A further level of complexity is due to the presence ofcytoplasmic structures initially termed bacterium-like organ-

isms (BLOs) that have been found in different AM fungalspecies (Glomus calidonium, Acaulospora laevis, Gigasporamargarita) by electron microscopy (7, 18, 21, 27). A combinedmorphological and molecular approach has shown that BLOsin the spores of G. margarita (isolate BEG 34) are true bacteria(6). Amplification of bacterial 16S RNA genes from total sporeDNA followed by direct sequencing indicated a homogeneousbacterial population closely related to the genus Burkholderia(6). Attempts to isolate and grow these endobacteria fromspores have been unsuccessful so far.

To determine whether intracellular bacteria occur sporadi-cally in individual AM fungal isolates or are a common featurein the family Gigasporaceae, we investigated using morpholog-ical and molecular approaches, two more isolates of G. mar-garita, derived from distant geographic areas, and nine isolates

TABLE 1. AM fungal isolates analyzed in this study

Species Origin Isolatea Supplier(s)

G. margarita Becker & Hall New Zealand BEG 34 V. Gianinazzi-PearsonG. margarita Becker & Hall West Virginia INVAM WV 205A J. MortonG. margarita Becker & Hall Brazil Personal collection J. DoddG. rosea Nicolson & Schenck Unknown DAOM 194757b D. Douds and G. BecardG. rosea Nicolson & Schenck United States BEG 9 J. DoddG. rosea Nicolson & Schenck United States BEG 9 V. Gianinazzi-PearsonG. rosea Nicolson & Schenck Florida INVAM FL 185 D. DoudsG. gigantea Gedermann & Trappe Pennsylvania HC/F E30 D. DoudsG. decipiens Hall & Abbott Unknown BEG 45 C. LeyvalS. persica (Schench & Nicol.) Walker & Sanders Porto Caleri (Rovigo), Italy HC/F E28 V. BianciottoS. persica (Schench & Nicol.) Walker & Sanders Migliarino (Pisa), Italy HC/F E09 V. BianciottoS. castanea Walker France BEG 1 Biorize

a BEG, European Bank of Glomales; DAOM, Department of Agriculture, Ottawa, Mycology; INVAM, International Culture Collection of Arbuscular andVesicular-Arbuscular Mycorrhizal Fungi; HC/F, Herbarium Cryptogamicum Fungi, Department of Plant Biology, Turin Italy.

b This isolate was originally received as G. margarita, but 18S rDNA and internal transcribed spacer sequence analyses demonstrated that it belongs to G. rosea (5;V. Bianciotto, E. Lumini, J. Morton, L. Lanfranco, and P. Bonfante, unpublished data)

* Corresponding author. Mailing address: Dipartimento BiologiaVegetale, Universita di Torino, V. le Mattioli 25, 10125 Turin, Italy.Phone: (39) 011 6502927. Fax: (39) 011 6707459. E-mail: p.bonfante@csmt.to.cnr.it.

4503

on May 30, 2015 by guest

http://aem.asm

.org/D

ownloaded from

belonging to five other AM species in the genera Gigasporaand Scutellospora (Table 1). Spores were picked with forceps,rinsed five times with sterile filtered distilled water, surfacesterilized with 4% chloramine T and 0.04% streptomycin for 30min, sonicated five times, and then rinsed five times (10 mineach) with sterile filtered distilled water. To eliminate thepossibility that contaminating bacteria were present on thefungal surface at the end of the sterilization procedure, sporesfrom each of the isolates were stained with a Live/Dead Bac-Light bacterial viability kit (Molecular Probes) as previously

described (6) and were observed without prior crushing. In allcases, the spore surface was completely free of bacterial con-taminants (Fig. 1a and b).

Localization of endobacteria in AM fungal spores by fluo-rescence and in situ hybridization. Fungal cytoplasm was re-leased by crushing spores between a microscope glass slide anda coverslip. Staining with the fluorescent BacLight dye showedthat intracellular bacteria were present in 7 of 11 fungal iso-lates (Fig. 1). The four isolates that did not contain bacteria allbelonged to the species Gigaspora rosea (Fig. 1c). The en-

FIG. 1. Bacterial endosymbionts in the cytoplasm of manually crushed spores of six fungal isolates (e through h) stained with the Live/Dead Baclight kit andobserved by using a Nikon Optiphot-2 microscope with a View Scan DVC-250 confocal system (Bio-Rad, Hemel Hempstead, United Kingdom). Living bacteriafluoresce bright yellow-green under blue light, while dead bacteria fluoresce red under green light. (a and b) No contaminating bacteria were observed on the externalsurfaces of sterilized and sonicated spores of G. rosea (a) and G. margarita (WV 105A (b). Bars, 100 mm (a) and 50 mm (b). (c) No bacterial endosymbionts weredetected in the cytoplasm of G. rosea BEG 9. Only the fungal nuclei (red masses) are visible. Bar, 10 mm. (d) Cytoplasm of a G. margarita WV 205A spore containingmany living rod-shaped bacteria that fluoresce green (arrows) and fungal nuclei (red masses). Bar, 10 mm. (e) Cytoplasm of S. persica HC/F E09 containing numerousrod-shaped bacteria (arrows). The nuclei are broken, and red filaments of chromatin are visible. Bar, 10 mm. (f) Appearance of bacteria in S. persica HC/F E28. Bar,7 mm. (g) S. castanea BEG 1 cytoplasm. Numerous living bacteria are present between the fungal nuclei. Bar, 7 mm. (h) Cytoplasm of G. gigantea containing a veryhigh number of living bacteria that are smaller and rounder than the bacteria in the other isolates. Bar, 7 mm. N, nuclei.

4504 BIANCIOTTO ET AL. APPL. ENVIRON. MICROBIOL.

on May 30, 2015 by guest

http://aem.asm

.org/D

ownloaded from

dobacteria mostly fluoresced as green, rod-shaped spots (Fig.1d, e, f, and g), indicating that they were alive. The bacteriawere less numerous in Scutellospora persica HC/F E28 than inthe other S. persica isolate (Fig. 1f). A very high number ofendobacteria that were more round and smaller were found inthe cytoplasm of Gigaspora gigantea (Fig. 1h).

To confirm the identity and location of endobacteria in AMfungal spores, in situ hybridization experiments were per-formed with three isolates, G. margarita BEG 34 and WV 205Aand G. rosea BEG 9. Oligonucleotide probes targeted to 16SrRNAs have been used successfully to detect and identify en-

vironmental nonculturable prokaryotes (3) and bacterial endo-symbionts of insects (8, 9). The specific protocol described byFukatsu et al. (9) using digoxigenin (DIG)-labelled probes wasfollowed. A positive signal was obtained with G. margaritaBEG 34 and WV 205A after hybridization with the oligonu-cleotide ribosomal DNA (rDNA) sequence BLOr, specificallydesigned for the bacterial endosymbiont of G. margarita BEG34 (6). Blue rod-shaped spots were especially visible when theygrouped together close to the fungal lipid droplets (Fig. 2a).Their shape and size (about 1 mm) corresponded well to thoseof endobacteria revealed with the fluorescent dye on unfixed

FIG. 2. In situ hybridization of intracellular symbiotic bacteria in spores of G. margarita BEG 34 and G. rosea BEG 9. BLOr (59-GTCATCCACTCCGATTATTTA-39) (6) hybridizes specifically with the 16S rRNA of the G. margarita endosymbiont and was used as probe. (a) In G. margarita a large number of blue rod-shapedspots (diameter, about 1 mm) (arrows) were visible in the fungal cytoplasm. They were especially visible when they grouped together in the cytoplasm. Bar, 10 mm. (b)The shape and position of spots correspond well with those of endobacteria revealed after Baclight kit staining of unfixed spores from the same fungal isolate. Bar,7 mm. (c) No hybridization signal was obtained when the DIG-labelled probe was omitted. Bar, 10 mm. (d) No hybridization signal was obtained with the cytoplasmof G. rosea when the DIG-labelled BLOr probe was used. Bar, 10 mm. W, spore wall; L, lipid masses.

VOL. 66, 2000 BACTERIAL ENDOSYMBIONTS IN AM FUNGI 4505

on May 30, 2015 by guest

http://aem.asm

.org/D

ownloaded from

spore sections from the same fungal isolates (Fig. 2b). Similarsignals were also obtained by using as a probe oligonucleotideEUB338 (2), designed to bind to bacterial 16S rDNA (data notshown). No signal was found in control experiments in whichthe DIG-labelled probe was omitted (Fig. 2c). None of theprobes gave hybridization signals in the cytoplasm of G. rosea(Fig. 2d).

The in situ hybridization results provide important confir-mation of the nature and topology of endobacteria in AMfungi. In fact, AM fungi and bacteria interact at different levelsof cellular integration, ranging from apparently loose associa-tion through surface attachment to intimate and obligatoryendosymbiosis (23). Therefore, the simultaneous presence ofbacteria outside and inside the fungal cell requires carefulexperimental procedures to make sure that PCR amplificationis targeted to the endosymbiotic bacterial DNA. The develop-ment of in situ protocols should also result in an important toolfor investigating bacterial functions related to the expression ofspecific genes, some of which have been already characterizedin the endosymbiotic bacteria of G. margarita BEG 34 (25).

Amplification of endobacterial 16S rDNA with universaland specific primers. For crude DNA preparation, 10 sporesamples were surface decontaminated as described above. Ex-treme care was taken to avoid subsequent bacterial contami-nation, and all steps were carried out in a laminar flow hood.

DNA was extracted by the protocol described previously (17).Two sets of primers were used: the universal eubacterial 704f-1495r primer pair and the BLOf-BLOr primer pair, specificallydesigned for the bacterial endosymbiont of G. margarita BEG34 (6). PCR amplifications were performed in a Hybaid Om-nigene thermal cycler with the following parameters: 3 min at95°C (one cycle); 45 s at 92°C, 45 s at 50°C, and 45 s at 72°C (40cycles); and 5 min at 72°C (one cycle).

Universal bacterial primers 704f and 1495r were first used toinvestigate the presence of bacteria inside the AM fungalspores. These primers amplified a DNA fragment of the ex-pected size (about 790 bp) from most isolates (Fig. 3a) Onlythe DNA in four G. rosea isolates could not be amplified withthe universal bacterial primers, although their DNA were suc-cessfully amplified with primers for the fungal rDNA genes(data not shown). In addition, the DNA in G. rosea isolates andG. gigantea could not be amplified with primers BLOf andBLOr, whereas an amplified DNA fragment about 400 bp longwas obtained from all other DNA samples (Fig. 3b).

Endobacteria are not a sporadic phenomenon in the Giga-sporaceae. Morphological and molecular analyses demon-strated the presence of endobacteria in the cytoplasm of five ofsix different fungal species in the genera Gigaspora and Scute-llospora. Isolates belonging to the same species were found tobe similar in terms of the presence or absence of endobacteriaand bacterial number, shape, and 16S rDNA sequences, evenwhen they were derived from distant geographic areas. Thiswas well documented for G. rosea and G. margarita.

The Gigasporaceae comprise two genera and 33 species(http://invam.caf.wvu.edu/Myc_Info/Taxonomy), and our anal-ysis is not representative of the whole family mainly due todifficulties in obtaining spore samples for all species. However,four of the five species in the genus Gigaspora were studied inthis investigation. In this genus, different species could havequite distinct features. In Gigaspora, two extreme cases are G.margarita and G. rosea; the former harbors an estimated250,000 bacteria per spore (6), and the latter harbors none, asalso reported by Hosny et al. (14) on the basis of PCR results.G. gigantea also contains endosymbiotic bacteria, but they aredifferent from those found in other Gigaspora species bothbecause of their round shape and because total spore DNAcould not be amplified by the specific primers that amplifiedbacterial DNA from G. margarita and Gigaspora decipiens.These endobacteria are currently under investigation.

In Scutellospora, both species investigated contained rod-shaped endobacteria whose 16S rDNA was amplified withthe BLO primers, although some variability in bacterialnumber was found in the two isolates of S. persica. Success-ful amplification of bacterial DNA from Scutellospora cas-tanea BEG 1 and Scutellospora gregaria by the BLO primerswas reported by Hosny et al. (14). The genus Scutellosporacomprises almost 30 species, and analysis of a wider range ofspecies is needed to elucidate if endobacteria are commonin this genus.

Endobacterial phylogeny. Total spore DNA was amplifiedby using universal bacterial primers 27f and 1495r to obtainmost of the 16S ribosomal gene. Amplified fragments about1,500 bp long were obtained from S. persica HC/F E28 andHC/F E09, S. castanea BEG 1, and G. margarita WV 205A.They were cloned into the pGEM-T vector, and three differentclones were sequenced for each isolate as described by Lan-franco et al. (17). The 16S rDNA sequences obtained in thisstudy were aligned with those of the G. margarita BEG 34endobacteria (6) and of closely related bacterial species ob-tained through a BLAST search in which the endobacterialsequences were used as queries.

FIG. 3. PCR experiments designed to reveal the presence of endobacteria inspores of different AM fungal isolates when two pairs of primers were used. (a)Agarose (1.2%) gel electrophoresis of PCR products amplified with bacterialprimers 704f and 1495r when the following templates were used: G. margaritaWV 205A (lane 1), G. margarita Brazil isolate (lane 2), G. rosea BEG 9 (lane 3),S. persica HC/F E09 (lane 4), S. castanea BEG 1 (lane 5), G. decipiens (lane 6),G. gigantea (lane 7), and no DNA (lane 8). Lane M contained a 1-kb DNA ladder(Gibco BRL). (b) Agarose (1.2%) gel electrophoresis of PCR products amplifiedwith primers BLOf and BLOr specific for the endobacteria of G. margarita BEG34 (6) when the following templates were used: G. margarita WV 205A (lane 1),G. margarita Brazil isolate (lane 2), S. persica HC/F E09 (lane 3), G. rosea BEG9 (lane 4), S. castanea BEG 1 (lane 5), G. decipiens (lane 6), G. gigantea (lane 7),and no DNA (lane 8). Lane M contained a 1-kb DNA ladder (Gibco BRL).

4506 BIANCIOTTO ET AL. APPL. ENVIRON. MICROBIOL.

on May 30, 2015 by guest

http://aem.asm

.org/D

ownloaded from

All new sequences clustered together with the G. margaritaBEG 34 endosymbiont (Fig. 4) in a single, well-supportedbranch. This endosymbiont was originally classified as a sistergroup of Burkholderia cepacia (6) in the beta subdivision of thedivision Proteobacteria (32). A more recent comparison of the16S rDNA sequences of Burkholderia strains has revealed aclear separation into two branches, and a number of speciespreviously assigned to this genus have been reassigned to thegenus Ralstonia (16, 33). This taxonomic rearrangement, aswell as identification of novel species of Burkholderia, has againraised basic questions concerning the taxonomic position ofthe endosymbiotic bacteria of the Gigasporaceae. The neigh-bor-joining tree in Fig. 4 suggests that the closest relatives of

the endobacteria that have been sequenced are members ofthe genus Burkholderia, as they form a well-supported branchnested in this genus that is well separated from Ralstonia.

Some hypotheses concerning the establishment of symbiosisbetween AM fungi and their endobacteria. Intracellular sym-bioses raise fascinating questions about the acquisition of theendosymbionts, the transmission of the endosymbionts, andthe evolution of reciprocal adaptations (10). The presence ofendobacteria or BLOs in all glomalean families suggests thatthe ability of AM fungi to establish this type of associationappeared very early in evolution.

Margulis and Chapman (19) have discussed the importanceof endosymbiosis as an evolutionary mechanism and distin-

FIG. 4. Neighbor-joining tree obtained from alignment of the 16S rDNA of the endosymbionts of G. margarita, S. castanea, and S. persica isolates with the closestbacterial sequences retrieved by a BLAST search. Sequences were aligned by using the ClustalX program (31), and the alignment was edited with GeneDoc (22).Neighbor-joining analysis was performed with the ClustalX program using Kimura’s distance method. The branch comprising species in the genus Pseudomonas wasused as an outgroup. Branches are shown only when the percentage of bootstrap support (1,000 trials) exceeded 70%.

VOL. 66, 2000 BACTERIAL ENDOSYMBIONTS IN AM FUNGI 4507

on May 30, 2015 by guest

http://aem.asm

.org/D

ownloaded from

guished between permanent and cyclical endosymbioses; theformer remains stable over time, and the latter involves regularreassociation events. The type of relationship between AMfungi and their endobacteria remains an open question thatwill be more properly addressed by analysis of a wider range ofspecies. However, the observations made so far with membersof the Gigasporaceae suggest at least two possible and oppositescenarios.

Related endobacteria were found by sequencing DNA indifferent isolates belonging to the Gigasporaceae from verydistant geographic areas. This situation may be the result ofrare bacterial acquisition events during evolution, followed bystrictly vertical transmission of endosymbionts (permanentsymbiosis) through generations. The asexual reproduction typ-ical of AM fungi and the coenocytic nature of the mycelium inthe zygomycetes (26) are factors that could facilitate this typeof transmission.

However, an alternative scenario can also be envisaged.The complex situation observed in the Gigasporaceae couldbe derived from more temporary but frequent associationsof AM fungi with free-living soil bacteria (cyclical symbio-sis), with AM fungal species selecting their bacterial symbi-onts from the environment. Different free-living Burkhold-eria species have been identified in the rhizosphere and thehyphosphere of AM fungi (4) and may represent a reservoirof potential endosymbionts for the AM species harboringthis group of bacteria. A physical constraint that wouldmake endocytosis, and thus acquisition of bacteria from theenvironment, a very rare event in fungi is the cell wall thatsurrounds the fungal hyphae. However, zygomycetes mayrepresent a special case among fungi since Geosiphon pyri-forme, a zygomycete which is ancestral to the Glomales (11),is the only known fungus able to establish cyclical endosym-biotic associations with cyanobacteria (20).

In conclusion, we demonstrated by sequencing that at leastthree different fungal species in the two genera of the Gigas-poraceae harbor in their cytoplasm endosymbiotic bacteriarelated to each other and closely related to the genus Burk-holderia. The occurrence of related endobacteria in differentgeographic isolates of the same AM fungal species may arise asa result of either permanent or cyclical endosymbiosis based onspecific recognition mechanisms. The pattern of distribution ofendobacteria in different AM species, together with the recentfinding that isolates of G. margarita and G. rosea influenceplant growth and plant mineral content to different extents(24), raises intriguing questions about the biological role ofthese endosymbionts.

Nucleotide sequence accession numbers. The nucleotide se-quences determined in this study have been deposited in theGeneBank database under accession numbers AJ251634 (S.persica HC/F E28), AJ251635 (S. persica HC/F E09), AJ251636(S. castanea BEG 1), and AJ251633 (G. margarita WV 205A).

We are grateful to G. Becard, J. Dodd, D. Douds, C. Leyval, and J.Morton for AM fungal spore samples and to J. Morton for morpho-logical identification of some isolates. We thank M. Girlanda for crit-ical reading of the manuscript and C. Bandi for initial help withsequence alignment.

This research was funded by the EU IMPACT2 project (BIO-CT96-0027) and by the Italian National Council of Research (CNR).

REFERENCES

1. Albrecht, C., R. Geurts, and T. Bisseling. 1999. Legume nodulation andmycorrhizae formation; two extremes in host specificity meet. EMBO J.18:281–288.

2. Amann, R. I., N. Springer, W. Ludwig, H. D. Goertz, and K. H. Schleifer.

1991. Identification in situ and phylogeny of uncultured bacterial endosym-bionts. Nature 351:161–164.

3. Amann, R. I., W. Ludwig, and K. H. Schleifer. 1995. Phylogenetic identifi-cation and in situ detection of individual microbial cells without cultivation.Microbiol. Rev. 59:143–169.

4. Andrade, G., K. L. Mihara, R. G. Linderman, and G. J. Bethlenfalvay. 1997.Bacteria from rhizosphere and hyphosphere soils of different arbuscular-mycorrhizal fungi. Plant Soil 192:71–79.

5. Bago, B., S. P. Bentivenga, V. Brenac, J. C. Dodd, Y. Piche, and L. Simon.1998. Molecular analysis of Gigaspora (Glomales, Gigasporaceae). New Phy-tol. 139:581–588.

6. Bianciotto, V., C. Bandi, D. Minerdi, M. Sironi, H. V. Tichy, and P. Bon-fante. 1996. An obligately endosymbiotic mycorrhizal fungus itself harborsobligately intracellular bacteria. Appl. Environ. Microbiol. 62:3005–3010.

7. Bonfante, P., R. Balestrini, and K. Mendgen. 1994. Storage and secretionprocesses in the spore of Gigaspora margarita Becker & Hall as revealed byhigh-pressure freezing and freeze substitution. New. Phytol. 128:93–101.

8. Fukatsu, T., and N. Nikoh. 1998. Two intracellular symbiotic bacteria fromthe mulberry psyllid Anemoneura mori (Insecta, Homoptera). Appl. Environ.Microbiol. 64:3599–3606.

9. Fukatsu, T., K. Watanabe, and Y. Sekiguchi. 1998. Specific detection ofintracellular symbiotic bacteria of aphids by oligonucleotide-probed in situhybridization. Appl. Entomol. Zool. 33:461–472.

10. Futuyma, D. J., and M. Slatkin. 1983. Coevolution, p. 155. Sinauer, Sunder-land, Mass.

11. Gherig, H., A. Schußler, and M. Kluge. 1996. Geosiphon pyriforme, a fungusforming endocytobiosis with Nostoc (Cyanobacteria) is an ancestral memberof the Glomales—evidence by SSU rRNA analysis. J. Mol. Evol. 43:71–81.

12. Harrison, M. J. 1999. Molecular and cellular aspects of the arbuscularmycorrhizal symbiosis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 50:361–389.

13. Hijri, M., M. Hosny, D. Van Tuinen, and H. Dulieu. 1999. Intraspecific ITSpolymorphism in Scutellospora castanea (Glomales, Zygomycota) is struc-tured within multinucleate spores. Fun. Genet. Biol. 26:141–151.

14. Hosny, M., D. van Tuinen, F. Jacquin, P. Fuller, B. Zhao, V. Gianinazzi-Pearson, and P. Franken. 1999. Arbuscular mycorrhizal fungi and bacteria:how to construct prokaryotic DNA-free genomic libraries from the Gloma-les. FEMS Microbiol. Lett. 170:425–430.

15. Hosny, M., M. Hijri, E. Passerieux, and H. Dulieu. 1999. rDNA units arehighly polymorphic in Scutellospora castanea (Glomales, Zygomycetes). Gene226:61–71.

16. Kersters, K., W. Ludwig, M. Vancanneyt, P. De Vos, M. Gillis, and K. H.Schleifer. 1996. Recent changes in the classification of the pseudomonads: anoverview. Syst. Appl. Microbiol. 19:465–477.

17. Lanfranco, L., M. Delpero, and P. Bonfante. 1999. Intrasporal variability ofribosomal sequences in the endomycorrhizal fungus Gigaspora margarita.Mol. Ecol. 8:37–45.

18. MacDonald, R. M., and M. R. Chandler. 1981. Bacterium-like organelles inthe vesicular-arbuscular mycorrhizal fungus Glomus caledonius. New Phytol.89:241–246.

19. Margulis, L., and M. J. Chapman. 1998. Endosymbioses: cyclical and per-manent in evolution. Trends Microbiol. 6:342–345.

20. Mollenhauer, D., R. Mollenhauer, and M. Kluge. 1996. Studies on initiationand development of the partner association in Geosiphon pyriforme (Kutz.) v.Wettstein, a unique endocytobiotic system of a fungus (Glomales) and thecyanobacterium Nostoc punctiforme (Kutz.) Hariot. Protoplasma 193:3–9.

21. Mosse, B. 1970. Honey-coloured sessile Endogone spores. II. Changes in finestructure during spore development. Arch. Mikrobiol. 74:146–159.

22. Nicholas, K. B., H. B. J. R. Nicholas, and D. W. Deerfield. 1997. GeneDoc:analysis and visualization of genetic variation. EMBNEW. News 4:14.

23. Perotto, S., and P. Bonfante. 1997. Bacterial associations with mycorrhizalfungi: close and distant friends in the rhizosphere. Trends Microbiol. 5:496–501.

24. Ruiz-Lozano, J. M., and P. Bonfante. Intracellular Burkholderia strain has nonegative effect on the symbiotic efficiency of the arbuscular mycorrhizalfungus Gigaspora margarita. Plant Growth Regulat., in press.

25. Ruiz-Lozano, J. M., and P. Bonfante. 1999. Identification of a putativeP-transporter operon in the genome of a Burkholderia strain living inside thearbuscular mycorrhizal fungus Gigaspora margarita. J. Bacteriol. 181:4106–4109.

26. Sanders, I. R. 1999. No sex please, we’re fungi. Nature 399:737–739.27. Scannerini, S., and P. Bonfante. 1991. Bacteria and bacteria-like objects in

endomycorrhizal fungi, p. 273–287. In L. Margulis and R. Fester (ed.),Symbiosis as a source of evolutionary innovation: speciation and morpho-genesis. MIT Press, Cambridge, Mass.

28. Simon, L., J. Bousquet, R. C. Levesque, and M. Lalonde. 1993. Origin anddiversification of endomycorrhizal fungi and coincidence with vascular landplants. Nature 363:67–68.

29. Smith, S. E., and D. J. Read. 1997. Mycorrhizal symbiosis. Academic Press,London, United Kingdom.

30. Taylor, T. N., W. Remy, H. Hass, and J. Kerp. 1995. Fossil arbuscularmycorrhizae from the early Devonian. Mycologia 87:560–573.

4508 BIANCIOTTO ET AL. APPL. ENVIRON. MICROBIOL.

on May 30, 2015 by guest

http://aem.asm

.org/D

ownloaded from

31. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G.Higgins. 1997. The ClustalX Windows interface: flexible strategies for mul-tiple sequence alignment aided by quality analysis tools. Nucleic Acids Res.24:4876–4882.

32. Woese, C. R., O. Kandler, and M. L. Wheelis. 1990. Towards a natural systemof organisms: proposal for the domains Archea, Bacteria, and Eucarya. Proc.

Natl. Acad. Sci. USA 87:4567–4579.33. Yabuuchi, E., Y. Kosako, I. Yano, H. Hotta, and Y. Nishibuchi. 1995. Trans-

fer of two Burkholderia and an Alcaligenes species to Ralstonia gen. nov.:proposal of Ralstonia pickettii (Ralston, Palleroni and Doudoroff, 1973)comb. nov. and Ralstonia eutropha (Davis, 1969) comb. nov. Microbiol.Immunol. 39:897–904.

VOL. 66, 2000 BACTERIAL ENDOSYMBIONTS IN AM FUNGI 4509

on May 30, 2015 by guest

http://aem.asm

.org/D

ownloaded from