Lichen-Like Symbiosis 600 Million Years Ago

See allHide authors and affiliations

Science  13 May 2005:
Vol. 308, Issue 5724, pp. 1017-1020
DOI: 10.1126/science.1111347


The fossil record of fungi and lichens is scarce. Here we report the discovery of lichen-like fossils, involving filamentous hyphae closely associated with coccoidal cyanobacteria or algae, preserved in marine phosphorite of the Doushantuo Formation (between 551 and 635 million years old) at Weng'an, South China. These fossils indicate that fungi developed symbiotic partnerships with photoautotrophs before the evolution of vascular plants.

Fungi are a major eukaryote kingdom and perform critical ecological roles in nutrient recycling. Many living fungi maintain facultative or obligate interactions with marine and terrestrial photoautotrophs (1, 2). However, the fossil record of fungi is poor and includes Ordovician [460 million years ago (Ma)] glomaleans (3) and microfossils interpreted as probable fungi dating to >720 Ma (4). Fossil evidence for fungal interactions (such as cyanolichenization, mycoparasitism, and vesicular arbuscular mycorrhizal symbiosis) with other organisms comes from the ∼400-million-year-old Rhynie chert in Scotland, which also preserves a diverse fungal assemblage, including chytridiomycetes and ascomycetes (5). In addition, some Ediacara fossils (575 to 542 Ma) have been interpreted, on the basis of taphonomic observations, as fungi (6) and lichens (7).

Here we describe three specimens of lichen-like fossils occurring in thin sections of two phosphorite samples from the upper Doushantuo Formation at Weng'an, South China (8) (fig. S1). The samples were collected from a 0.5- to 5-m-thick unit of black bituminous phosphorite immediately above a karstification surface in the middle Doushantuo Formation (9). This unit was probably deposited in a shallow subtidal environment and contains abundant algal fossils (10, 11). The Doushantuo Formation in the Yangtze Gorges area is bracketed by U-Pb ages between 635 ± 1 and 551 ± 1 Ma (12), and direct Pb-Pb dating of upper Doushantuo phosphorite at Weng'an indicates that the fossils described here are probably 599 ± 4 million years old (13); however, Condon and colleagues argue that the fossiliferrous upper Doushantuo Formation may be between 580 and 551 million years old (8, 12).

The lichen-like fossils are completely phosphatized. They consist of two closely associated components: coccoidal cells and thin filaments (Figs. 1 and 2). The coccoid cells are 6 to 15 μm in diameter (average = 9 μm, SD = 2 μm, n = 25 cells) and are usually clustered (Figs. 1A, 2A, and 2C). They typically consist of an opaque central body surrounded by a hyaline envelope 1 to 2 μm thick (Fig. 2E). In some, the remains of organic sheaths are visible in the hyaline envelope. These coccoidal cells are interpreted as sheathed cyanobacteria (similar to modern Gloeocapsa, Entophysalis, and Chroococcus) or possibly green algae (similar to modern colonial chlorococcaleans).

Fig. 1.

Thin-section photomicrographs of two better-preserved specimens. (A) Coccoidal thallus divided by dense filaments in the middle. Further compartmentalization of coccoidal thallus by less densely packed filaments is visible at higher magnification (Fig. 2, A to C). (B) Coccoidal thallus with filaments (not discernable at this magnification; see magnified views of arrowed areas in Fig. 2, D to G) in the left part, but not the right part, of this specimen. Scale bars, 100 μm.

Fig. 2.

(A) Detail of the upper center part of Fig. 1A. (B) Detail of the upper right quarter of (A), showing filament tract. Several filaments have dark swollen terminal structures. Arrows indicate terminal structures with multiple filamentous appendages. (C) Detail of the center part of (A). Two packets of coccoidal cells (center and lower center) are surrounded by a few filaments. The arrow points to a filament that appears to envelop a coccoidal cell. The arrowhead indicates a filament with a terminal loop (close-up in inset). (D) Detail (white arrow in Fig. 1B) showing hyphae with funnel-shaped terminal structures (arrowheads), the distal parts of which were probably dehisced along a transverse split. (E) Detail (white arrowhead in Fig. 1B) showing funnel-shaped terminal structures (black arrowheads), clustered coccoidal cells with hyaline envelopes (upper right quarter), dark terminal structures connected to multiple or branching filaments (black arrows), and a filament with a terminal structure and a laterally borne intercalary vesicle (white arrowheads). (F) Detail (black arrow in Fig. 1B) showing terminal structure (arrow) with subtending filament. (G) Detail (black arrowhead in Fig. 1B) showing branching filament (arrow) in close association with coccoidal cells. Scale bars, 20 μm.

The filaments are about 0.5 to 0.9 μm wide (average = 0.6 μm, SD = 0.1 μm, n = 20 filaments). They are up to 50 μm long, although they may be longer, because the 30-μm-thick thin section captures only a segment of the filaments. It is unclear whether they are septate, because they are opaque. Some filaments branch dichotomously (Fig. 2, E and G). Many bear opaque, pyriform terminal structures (Fig. 2, B and D to F) that are smaller than the coccoidal cells described above, about 3 to 6 μm in maximum dimension (average = 5 μm, SD = 1 μm, n = 6 terminal structures) and 2 to 4 μm in minimum dimension (average = 3 μm, SD = 1 μm, n = 6 terminal structures). Some terminal structures show evidence of possible transverse splits (Fig. 2, D to E). A number of filaments appear to envelop coccoidal cells or are arranged in loops (Fig. 2C). In some cases, a single filament connects two pyriform structures, or a single pyriform structure is connected to multiple filamentous appendages. The filaments lack hyaline sheath-like envelopes that characterize filamentous cyanobacteria, and can be distinguished from pseudoparenchymatous multicellular algae preserved in the same deposit (Fig. 3) (10, 11). In one specimen (Fig. 1A), which was probably fragmented during postphosphatization reworking, the filaments can be found throughout the entire specimen. In another (Fig. 1B), the filaments occur on only one side of the specimen. However, because the specimens were found in thin sections, it remains impossible to reconstruct the three-dimensional structure of the coccoid/filament association.

Fig. 3.

Doushantuo algal thalli with no filamentous symbionts. (A) Pseudoparenchymatous red alga (Thallophyca corrugata) from the same horizon (9, 11). (B) Thallus from the same thin section where lichen-like fossils were found. Emerging cell islands (ci) are indicated. Scale bars, 20 μm.

We interpret these filaments as fungal hyphae and the pyriform terminal structures as resting spores, reproductive structures, or some type of fungal vesicle. Alternative interpretations (such as filamentous cyanobacteria) are inconsistent with the combination of morphological features (thin filaments, dichotomous branching, pyriform terminal structures, and absence of sheaths). The diameter of the hyphae may have been reduced during phosphatization (14), but modern marine fungal hyphae can be <1 μm in diameter (1). The pyriform terminal structures are similar to, although smaller than, modern and fossil glomalean spores or vesicles (2, 3, 15). Furthermore, glomalean (such as Entrophospora) hyphae can bear terminal sporiferous saccules and lateral spores (2), which are similar to those illustrated in Fig. 2E (white arrowheads).

It is unlikely that the fungal hyphae were saprophytic or were accidentally preserved with the coccoidal cells. In all three specimens, the hyphae are associated only with coccoidal thalli; they do not occur in pseudoparenchymatous red algae in the same deposit (Fig. 3A) (10, 11), which would be expected if they were saprophytic. Furthermore, the coccoidal cells would be expected to show a greater degree of decomposition if the fungal hyphae were saprophytic; instead, the preservation of coccoidal cells is not inferior to that of the fungal hyphae. Third, the hyphae appear to be structurally (and not accidentally) associated with the cyanobacterial coccoids; the coccoid clusters are distinctly compartmentalized and surrounded by abundant hyphae (Figs. 1A, 2A, and 2C) similar to the hyphal nets described in the Devonian cyanolichen (16, 17). This structural association make the coccoidal clusters appear different from structures described as “cell islands” in Doushantuo multicellular algae (10); cell islands (Fig. 3B) are surrounded by ellipsoidal cells rather than hyphae. In addition, some hyphae are in close contact with coccoid cells (Fig. 2, C and G), suggesting that there was direct physiological interaction between them.

The association between coccoidal cells and fungal hyphae is interpreted to be symbiotic, not parasitic. The coccoidal thalli show no evidence of host reaction to mycoparasitism. Neither do the coccoid cells in contact with hyphae show morphological abnormality. On the other hand, there are numerous similar coccoidal thalli in the same deposit that are not associated with fungal hyphae (Fig. 3B). Thus, the coccoidal thalli may have functioned as facultative photobionts that could form loose lichen-like or lichenoidal (1) association with filamentous mycobionts.

Terrestrial lichens, involving ascomycetes or basidiomycetes as mycobionts and cyanobacteria or chlorophytes as photobionts, have affected global weathering since the Devonian (5). Modern marine fungi (mostly ascomycetes) also form a wide range of interactions with cyanobacteria, chlorophytes, phaeophytes, and rhodophytes. These interactions can be loose lichenoidal association with microscopic photobionts, mycophycobiosis with macroscopic algae, mycoparasitism, or obligate lichen association (1). Lichenized fungi are phylogenetically widespread within the Dikaryomycota (Ascomycota + Basidiomycota), which suggests that fungal lichenization may have evolved multiple times (18-20). However, the broadly defined symbiotic life-style (including arbuscular mycorrhizal symbiosis) has a broader phylogenetic distribution and characterizes the Symbiomycota (Glomeromycota + Dikaryomycota) (21, 22). Although most glomerocycetes are arbuscular mycorrhizal fungi with vascular plants, Geosiphon pyriforme (a basal glomeromycete) is symbiotic with cyanobacteria (23). The ease with which symbionts can be gained, lost, and switched in fungal/photoautotroph associations (24, 25), the fungal phylogenetic tree that is basally populated by aquatic chytrids (22), and probably fungal fossils from Proterozoic marine deposits (4) indicate that the early steps toward fungal/photoautotroph symbiosis may have begun as facultative interactions with aquatic cyanobacteria or algae. The Doushantuo lichenoidal fossils suggest that these early steps may have occurred long before the colonization of land by vascular plants, in a shallow marine ecosystem where a large number of free-living cyanobacteria, algae, and fungi were in close association—a necessary step in the evolution of symbiosis. Thus, this and other fossil evidence (4) join molecular data (26, 27) to support a deep history of fungi and lichen-like symbiosis.

Supporting Online Material

SOM Text

Fig. S1


References and Notes

Stay Connected to Science

Navigate This Article