Glomalean Fungi from the Ordovician

Dirk Redecker, et al.

Science 289, 1920 (2000);
DOI: 10.1126/science.289.5486.1920

The following resources related to this article are available online at

www.sciencemag.org (this information is current as of February 21, 2007 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/cgi/content/full/289/5486/1920

This article cites 8 articles, 4 of which can be accessed for free:

Downloaded from www.sciencemag.org on February 21, 2007

http://www.sciencemag.org/cgi/content/full/289/5486/1920#otherarticles

This article has been cited by 125 article(s) on the ISI Web of Science.

This article has been cited by 28 articles hosted by HighWire Press; see:
http://www.sciencemag.org/cgi/content/full/289/5486/1920#otherarticles

This article appears in the following subject collections:

Paleontology
http://www.sciencemag.org/cgi/collection/paleo

Information about obtaining reprints of this article or about obtaining permission to reproduce
this article in whole or in part can be found at:
http://www.sciencemag.org/help/about/permissions.dtl

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
c 2000 by the American Association for the Advancement of Science; all rights reserved. The title SCIENCE is a
registered trademark of AAAS.
 REPORTS
extant spores, the loss of wall layers by
Glomalean Fungi from the microbial degradation has been frequently
reported (17 ).

Ordovician Our findings push back the time of the

fossil record of glomalean fungi by 55 to 60
million years and also suggest that these
Dirk Redecker,1* Robin Kodner,2 Linda E. Graham2 fungi were present before the first vascular
plants arose. Although hepatics and horn-
Fossilized fungal hyphae and spores from the Ordovician of Wisconsin (with an worts lack roots that would allow them to
age of about 460 million years) strongly resemble modern arbuscular mycor- form a “mycorrhiza” in a strict sense, AM-
rhizal fungi (Glomales, Zygomycetes). These fossils indicate that Glomales-like like colonization of their thallus has long
fungi were present at a time when the land flora most likely only consisted of been known (2). Evidence that a hornwort
plants on the bryophytic level. Thus, these fungi may have played a crucial role forms an association with a defined AM
in facilitating the colonization of land by plants, and the fossils support mo- fungus was only obtained recently (3). We
lecular estimates of fungal phylogeny that place the origin of the major groups have no evidence that these Ordovician
of terrestrial fungi (Ascomycota, Basidiomycota, and Glomales) around 600 fossil fungi were associated with plants, but
million years ago. fossil evidence for Ordovician land plants
(18) and the well-documented capability of

Downloaded from www.sciencemag.org on February 21, 2007

Arbuscular mycorrhizal (AM) symbiosis be- ible. Where spores are not entangled in hy- some extant bryophytes for AM-like inter-
tween modern plants and fungi is ubiquitous phae, a single hyphal connection per spore actions suggest this possibility. Alterna-
and contributes to nutrient acquisition by can be discerned (Fig. 1, C and E to G). We tively, they may have formed symbioses of
most vascular plants (1). It has also been found no single spore that did not retain its the Geosiphon type or have been saprobes.
reported for hepatics and hornworts (2, 3). subtending hypha. The spore wall apparently Geosiphon is a nonmycorrhizal ancestral
Arbuscular mycorrhizae may have played an consists of a single layer. In their shape, size, member of the Glomales, which forms an
important role in the success of early terres- and hyphal connection, the spores resemble endosymbiosis with cyanobacteria (19) and
trial plants (4, 5). those of modern glomalean fungi in the genus evolved at approximately the same time as
Because of the poor preservation of Glomus (Fig. 1, D and H). Similar loose spore the deeply divergent mycorrhizal lineages
most fungal structures, it has been difficult clusters are frequently formed by present-day (20). It also forms spores of the Glomus
to interpret the fossil record of fungi (6 – 8). AM fungi but are unknown from algae, other type (21).
Hyphae, the vegetative body of fungi, bear fungi, or Oomycetes. The fossils reported here support time
few distinctive morphological characters, Species determination in the Glomales estimates of fungal phylogeny derived from
and organisms as diverse as cyanobacteria, depends on the layer structure of the spore ribosomal small subunit sequences (22)
eukaryotic algal groups, or Oomycetes can wall (17 ). No such layers could be dis- (Fig. 2). Using a modified data set from a
be easily mistaken for them. The most con- cerned in the fossils. Layers may have de- previous study (22), we included the Glo-
vincing evidence for the presence of fungi graded before they were deposited, over males-like fungi from the Ordovician and
is fungal fossils associated with plants, in- time, or during the acid treatment used to some other recently reported fossils to date
cluding arbuscular mycorrhizae (9, 10) and dissolve the rock. Our experiments with their divergences within the tree (23). Some
ascomycetous or chytrid parasites (11, 12). living extant spores showed that the spore of these fossils provide useful calibration
The previously reported oldest fungal fos- wall layers are acid stable, but for dead points for the tree (Fig. 2, points a and d),
sils originate nearly exclusively from a sin-
gle site, the Devonian Rhynie Chert [400
million years ago (Ma)] (9, 10). Possible
earlier ascomycetes were reported from the
Silurian of Sweden (13).
The fossils reported here were found in
the Guttenberg Formation, mid-Ordovician
dolomite of Wisconsin, which was deposited
in a presumed shallow marine setting be-
tween 460 and 455 Ma (14 –16 ). The fossil-
ized material consists of entangled, occasion-
ally branching, nonseptate hyphae and spores
(Fig. 1, A to C and E to G). The hyphae have
a diameter of 3 to 5 ␮m and bear globose to
subglobose terminal spores that are 40 to 95
␮m in diameter. These spores are connected
to cylindrical to slightly flared (Fig. 1, C and
G) or recurved (Fig. 1E) subtending hyphae.
No specialized attachment structures are vis- Fig. 1. (A to C and E to G) Fossil hyphae and spores from the Ordovician and (D and H) spores
formed by extant glomalean fungi. (A and B) Overviews of the fossilized material. (C, E, F, and G)
Fossil spore details. (C) Detail of (B). (D) A spore of present-day Glomus sp. S328 with layered wall
1
Department of Plant and Microbial Biology, 111 Kosh- structure. In (G), the arrow shows walls of a subtending hypha in connection with the spore wall.
land Hall, University of California, Berkeley, CA 94720, (H) A spore of present-day Glomus leptotichum, a member of the deeply divergent glomalean
USA. 2Department of Botany, University of Wisconsin, lineages. Images were obtained by light microscopy (28) of the specimens in air (A, C, F, and G),
Madison, WI 53706, USA. differential interference contrast microscopy of the specimens in polyvinylalcohol-lactoglycerol (D,
*To whom correspondence should be addressed. E- E, and H), and confocal laser scanning microscopy with the autofluorescence of the material (B). All
mail: [email protected] scale bars are 50 ␮m.

1920 15 SEPTEMBER 2000 VOL 289 SCIENCE www.sciencemag.org

 REPORTS
22. M. L. Berbee and J. W. Taylor, Systematics and Evo-
lution, Part B, vol. VII of The Mycota, D. J. McLaughlin,
E. G. McLaughlin, P. A. Lemke, Eds. (Springer-Verlag,
New York, 2000), pp. 229 –246.
23. A subset of 20 taxa from the original data set was
combined with newly added sequences from Mor-
tierella polycephala and nine glomalean species,
among them four representatives of the deeply
divergent lineages (Geosiphon pyriforme, the di-
morphic fungus Acaulospora gerdemannii Glomus
leptotichum, A. trappei, and Glomus occultum). A
neighbor-joining tree from these data was ob-
tained by the Kimura two-parameter method in
PAUP* (26) with a gamma shape parameter of 0.5
for among-site variation, which was then normal-
ized under the maximum likelihood criterion with
the molecular clock enforced. The maximum like-
lihood settings were as follows. The transition/
transversion ratio (2.17), base frequencies, and the
proportion of invariant sites were estimated by
maximum likelihood. A gamma shape parameter of
0.5 was assumed for among-site variation.

Downloaded from www.sciencemag.org on February 21, 2007

24. Taylor, Remy, and Hass (7) reported an Allomyces-
like fossil from the Devonian. As the Chytridiomy-
cetes probably originated much earlier than the
Ascomycete/Basidiomycete/Glomales clade (22,
27), it would not be useful as a calibration point. It
was not included in the sequence alignment, be-
Fig. 2. The phylogenetic tree of the fungi derived from small subunit ribosomal sequences. cause the 18S ribosomal DNA of Allomyces causes
Letters a to e assign tree branching points to fungal fossils. The respective geological times are extremely long branches in phylogenetic trees.
given as numbers on the tree. Triangles indicate that all fossils could also have been deposited 25. This scenario is more compatible with calibration
later in the history of each clade, allowing the origins of the clades to be shifted back in time. point d in Fig. 2 than assigning the 460-Ma fossils to
Point a indicates the glomalean fossils reported in the present study; b indicates the fossil the separation point of the Glomales/Ascomycete/
arbuscular mycorrhizae from the Rhynie Chert (400 Ma) (9); c indicates a fossil clamp Basidiomycete clades.
connection (earliest evidence of Basidiomycetes, 290 Ma) (29); d indicates the Ascomycetes 26. D. L. Swofford, PAUP*, Phylogenetic Analysis Using
from the Rhynie Chert (400 Ma) (30); e indicates the gilled mushroom in amber (90 Ma) (8); Parsimony (* and Other Methods), version 462 (Si-
and f indicates the arbuscular mycorrhizae of the Gigasporaceae type from the Triassic (⬃240 nauer, Sunderland, MA, 1999).
Ma) (31). The outgroups were Diphanoeca grandis (Choanoflagellate), Ichtyophonus hoferi, and 27. C. J. Alexopoulos, C. W. Mims, M. Blackwell, Introduc-
Dermocystis salari [early divergences of the animals (32)]. Branches that are not supported by tory Mycology ( Wiley, New York, ed. 4, 1996).
at least 50% bootstrap analyses are shown in gray. 28. In Fig. 1, C, F, and G were obtained with a Nikon
compound microscope equipped with an Image
12. W. I. Illman, Mycologia 76, 545 (1984). Explorer system (Hitachi Genetic Systems, Ala-
whereas others apparently were formed af- meda, CA) that was used to focus through the
ter the origin of the clade (Fig. 2, points b, 13. M. A. Sherwood-Pike and J. Gray, Lethaia 18, 1
(1985). specimens and combine the images of the different
c, and e) (24 ). The assumption that the 14. W. D. Huff, S. M. Bergstrom, D. Kolata, Geology 20, focusing planes to a composite with a very large
fossil organisms existed at the time of, or 875 (1992). depth of field. Figure 1B was obtained with a
after, the divergence of the deep glomalean 15. R. Sloan, Minn. Geol. Surv. Rep. Invest. 35, 7 (1987). Molecular Dynamics laser scanning confocal micro-
16. To eliminate possible problems of modern contam- scope (Molecular Dynamics, Sunnyvale, CA). The
lineages (Fig. 2, point a) places the origin ination, we removed 50 cm of the exposed rock excitation wavelength was 530 nm, and the
of the Ascomycete/Basidiomycete clade at before sampling. Samples were taken from the autofluorescence of the specimens above 600 nm
600 to 620 Ma (25), in good agreement center of the outcrop, contained no weathered was detected. This method allows similar optical
with the date of 600 Ma from Berbee and surfaces, and had not been penetrated by any sectioning as the method described above, but it
modern plant. To eliminate ancient contamination, also allows sectioning within optically dense areas
Taylor (22). No lineage before the Gloma- we used a rock saw to cut off the edges of the of the specimen that are not transparent enough
les developed plant symbioses, but numer- samples, which showed no inclusion or any evi- for light microscopy.
ous later diverging clades did, with global dence of past penetration. There were no obvious
29. R. L. Dennis, Mycologia 62, 578 (1970).
bedding planes in the rocks. If the rocks were
impacts on terrestrial ecosystems. contaminated some time after deposition, we 30. The earliest Ascomycetes from the Devonian show
would see other more common, similar-sized, re- features [Perithecia or Pseudothecia (11)] that are
sistant organic materials representative of later found in the present-day Ascomycete group of
References and Notes
1. S. E. Smith and D. J. Read, Mycorrhizal Symbiosis periods (e.g., vascular plant–type spores or pollen Pyrenomycetes and Loculoascomycetes, respec-
(Academic Press, London, ed. 2, 1997). grains, root fragments, and diatoms). Whole sam- tively (27). The Pyrenomycetes are the clade of the
2. M. Stahl, Planta 37, 103 (1949). ples of rock were subsequently submerged in two tree containing Neurospora, Hypomyces, and
3. A. Schü␤ler, Mycorrhiza 10, 15 (2000). baths of concentrated HCl in new plastic contain- Ophiostoma. Therefore, these fossils were assigned
4. K. A. Pirozynski and D. W. Malloch, BioSystems 6, 153 ers. Sludge and acid from the first bath were to the major radiation of Euascomycetes (Fig. 2,
(1975). discarded. The sediment from the second bath and point d).
5. L. Simon, J. Bousquet, R. C. Levesque, M. Lalonde, occasionally remaining rock were examined under 31. C. J. Phipps and T. N. Taylor, Mycologia 88, 707
Nature 363, 67 (1993). a dissecting microscope to separate the organic (1996).
6. T. N. Taylor and E. L. Taylor, The Biology and Evolution matter from the inorganic sediments. Specimens 32. M. A. Ragan et al., Proc. Natl. Acad. Sci. U.S.A. 93,
of Fossil Plants (Prentice Hall, Englewood Cliffs, NJ, shown in Fig. 1, B and C and E to G, were deposited 11907 (1996).
1993). at the University of California Museum of Palae- 33. We thank J. Gray for valuable suggestions. We also
7. T. N. Taylor, W. Remy, H. Hass, Nature 367, 601 ontology, Berkeley (UCMP 151983 and 151984).
thank T. Bruns for suggestions and continuous
(1994). 17. J. B. Morton, Mycotaxon 32, 267 (1988).
support of D.R., J. Taylor and F. W. Byers for advice
8. D. S. Hibbett, D. Grimaldi, M. J. Donoghue, Nature 18. J. Gray, D. Massa, A. J. Boucot, Geology 10, 197
on the manuscript, S. Lee for introducing and giving
377, 487 (1995). (1982).
access to the Image Explorer System, S. Ruzin and
9. W. Remy, T. N. Taylor, H. Hass, H. Kerp, Proc. Natl. 19. H. Gehrig, A. Schü␤ler, M. Kluge, J. Mol. Evol. 43, 71
D. Schichnes for helping with the confocal laser
Acad. Sci. U.S.A. 91, 11841 (1994). (1996).
20. D. Redecker, J. B. Morton, T. D. Bruns, Mol. Phylo- scanning microscope, and F. Landis for establishing
10. T. N. Taylor, W. Remy, H. Hass, H. Kerp, Am. J. Bot.
82, 92 (1995). genet. Evol. 14, 276 (2000). communication.
11. T. N. Taylor, H. Hass, H. Kerp, Nature 399, 648 21. A. Schü␤ler, D. Mollenhauer, E. Schnepf, M. Kluge,
(1999). Bot. Acta 107, 36 (1994). 2 June 2000; accepted 26 July 2000

www.sciencemag.org SCIENCE VOL 289 15 SEPTEMBER 2000 1921