Epulopiscium fishelsoni, Big Bug Baffles Biologists!
Imagine the lessons to be learned if bacteriologists could work with pure cultures
of a prokaryote that is: larger than many eukaryotic cells; structurally and ecologically
complex; highly motile and covered with bacterial flagella; undergoing marked
circadian cycles in physiological processes; and found only as a gut symbiont
of a vertebrate host.
traits are characteristic of gigantic (to at least 600 x 60 µm) eubacterial
gut symbionts of tropical surgeonfishes (Family Acanthuridae; "tangs").
Although only one form has been named (Epulopiscium fishelsoni; Montgomery
and Pollak 1988a), 10 structural "morphs" of these symbionts (differing
in maximum size, shape and reproductive traits) have been recorded in surgeonfishes
from Hawaii, Guam, Japan, French Polynesia, the Great Barrier Reef, Tuvalu,
South Africa and the Caribbean (e.g., see Clements et al. 1989). We use the
term epulo to distinguish these organisms from other types and taxa
Epulos possess a distinct nucleoid separated from the cytoplasm by an unidentified,
thin, membranous structure, and the cytoplasm is packed with an intricately-folded
system of what appear to be membranes (Fishelson et al. 1985; Montgomery and
Pollak 1988a; Clements et al. 1989; Bresler et al. 1998). Condensed DNA is arranged
in elongate, chromosome-like structures with cross-striations similar to those
seen in dinoflagellates, so-called "mesokaryotes". Cells are highly
motile, driven by a covering of bacterial flagella (Fishelson et al. 1985; Clements
and Bullivant 1991) yet they do not display the typical run and tumble
behavior of most flagellate bacteria. Instead, they spiral back and forth, much
like the single-celled protozoan, Paramecium.
the day, as the host fish feeds, Epulopiscium fishelsoni grows (size
increases from 30-50 µm to > 200 µm; Fishelson et al. 1985, Montgomery
and Pollak 1988a). At night, when fish are inactive, E. fishelsoni undergoes
recurring reproductive events that return cell size to its initial small range.
Lengths of E. fishelsoni span from (at least) 10-15 µm to >
500 µm. DNA quantity is proportional to cell volume over cell lengths
of 30 µm to >500 µm (a >2000-fold difference in DNA quantity;
Bresler et al. 1998). Cell division involves production of two nucleoids within
a cell, deposition of cell walls around expanded nucleoids, and emergence of
daughter cells from the parent cell. Formation of daughter nucleoids and cells
occurs both during diurnal periods of host feeding and epulo cell growth and
during nocturnal periods of host inactivity when mean cell size declines. Epulos
also form some cells with unusually heavy cell walls (spores?) at night (Bresler
et al. 1998).
Although we had originally suggested that epulos were protists due to their
large size, structural complexity and unusual life cycle (Montgomery and Pollak
1988a), molecular work by Esther Angert (now at Cornell), Kendall Clements (University
of Auckland) and Norman Pace (now at Colorado; Angert et al. 1993) identified
epulos as the largest Eubacteria. Their paper also brought scientific attention
to apparent violations by epulos of long-standing generalizations about cell
size in bacteria, specifically to the structural and physiological phenomena
previously thought to limit cell size, and to assignment of fossil unicells
as prokaryotes or eukaryotes on the basis of size alone (Sogin 1993). These
cells are large enough to allow use of many light microscopic techniques as
well as manipulative and invasive techniques (e.g., microprobes, microinjection)
to study the physiology of individual bacteria. Research on epulos could address:
factors which may limit cell size and cell morphology, or support gigantism,
in prokaryotes; subcellular compartmentalization in prokaryote cells; mechanisms
of cell division, daughter cell formation and encystment; vertebrate host-symbiont
interactions in an open system (the gut is continuous with the environment);
and bacterial dispersal and evolution in widespread but patchily distributed
habitats (surgeonfishes on tropical reefs throughout the world).
As noted, structural features such as size, shape, and mode of daughter cell
formation distinguish a variety of epulo morphs (Clements et al.
1989), but 16s rRNA gene sequence divergence has apparently not paralleled structural
divergence (Treutner, et al. in prep.). Angert et al. (1993) used 16S rRNA gene
sequences to place epulos among the low G+C gram positive bacteria related to
Clostridium spp., and later (Angert et al. 1996) identified Metabacterium
polyspora from guinea pigs as a sister group of epulos. We have since examined
epulos from 17 specimens of 5 host fishes from the Red Sea, Japan and the Hawaiian
Archipelago (Treutner, et al. in prep). Epulos represent a distinct, monophyletic
clade with two genetically distinct subdivisions within the epulo clade. The
subdivisions do not appear to relate to epulo morph, locality of collection
or identity of host species, and both subdivisions can be found in a single
At this time, we do not know how epulos are transmitted to the host. Epulopiscium
fishelsoni occupies the mid-intestine of the fish during the day, follows
food materials into the posterior intestine as these materials are evacuated
at night, and remains in the posterior intestine overnight (Fishelson et al.
1985, Montgomery and Pollak 1988a). We found neither epulos nor clearly identifiable
resting stages in the first fish feces of the day, on the algae that the fish
eat, in the water, or in fish eggs. Fish that were starved to eliminate epulos,
released back upon the reef, and recaptured in 10-12 days were not reinfected
although the sample size was small (n = 2).
We also do not know the nature of the symbiosis between epulos and their host
fishes. Anecdotal observations suggest a mutualism (two organisms living in
a close, mutually beneficial relationship). A tang that became ill and died
in an aquarium in Tel Aviv lacked epulos. Epulopiscium fishelsoni suppresses
local gut pH by 1 pH unit during the day, but not at night (Montgomery and Pollak
1988b), reflecting a significant metabolic shift from day to night that parallels
the hosts feeding activity. Epulopiscium fishelsoni from the Red
Sea enhances lipid digestion by the host (Pollak and Montgomery 1994), but smaller
morphs from Australian fish may enhance carbohydrate digestion (Clements 1997).
These unique bacteria remain an intellectual oddity rather than a laboratory
tool for one primary reason: they cannot be maintained in a culture where they
exhibit the complex life cycle and dynamic cell and population growth seen in
nature. All knowledge about these unique bacteria comes from preserved cells
or live cells within 1 hour of collection. This necessitates repeated travel
and long stays at tropical reefs to collect the hosts and microbes.
Successful culture of epulos will require collaboration between ichthyologists
trained to collect and dissect the fish and microbiologists skilled in handling
and maintaining bacteria. In May 1999, a German graduate student and microbiologist
(Silke Treutner) in our laboratory attempted culture of epulos from surgeonfishes
collected in Hawaii. Previously, epulos had been kept alive for only a few hours
(personal observation; K. Clements, pers. commun.). In contrast, our epulos
remained alive for several months. Unfortunately, only small cells survived
and neither individual cell growth nor significant population growth occurred.
Thus, while the first and perhaps most significant step in culturing epulos
has been accomplished (maintaining live cells for an extended period), we still
cannot support the complex patterns of growth and reproduction seen in nature.
Early in 2001, we will begin research to extend Treutners important work
and, hopefully, find a way to culture epulos.
your question for Dr. Pollak and/or Dr. Montgomery.