Download The Development of Cooperative Associations Between Animals

AMER. ZOOL., 38:593-608 (1998)
The Development of Cooperative Associations Between Animals and
Bacteria: Establishing Detente Among Domains'
MARGARET J. MCFALL-NGAI 2
University of Hawaii, Pacific Biomedical Research Center, Kewalo Basin Laboratories,
41 Ahui Street, Honolulu, Hawaii 96813
SYNOPSIS.
detente = n.(fr.) the relaxation of tensions between two nations, usually through
cooperation and negotiation. [The Random House Dictionary, 2nd Edition, 1987]
Despite the ubiquitous occurrence of cooperative associations between animals and
bacteria, there is little understanding of how these interactions arose, how they
evolved, and how they persist. Thus, an extensive database concerning the influence
of bacteria on developmental pathways is not yet available. However, in much the
same way that mutually beneficial liaisons are created between nations with vastly
different histories and cultures, it is likely that highly refined developmental mechanisms exist in which a type of detente is created to retain the integrity of the
associations between the partners, both within and between generations. These
developmental pathways would be responsible for insuring that a balance of cell
growth is established and maintained among the community members, comprised
of animal and microbial cells, such that neither form of pathogenesis, i.e., overgrowth (war) or aposymbiosis (isolation), ensues. This contribution examines aspects of how alliances with prokaryotes may have been integrated into the mechanisms and patterns of host animal developmental programs.
(Margulis, 1970). It is against this backdrop
of
cooperating cells that the Domain EuBiologists have long known that animal
karya
evolved. Thus, when animals radiatevolution, including the evolution of their
ed,
two
possibilities presented themselves:
developmental pathways, occurred in envianimals
could develop mechanisms to inronments ecologically dominated by bactehibit
interactions
with bacterial cells; or,
ria. The members of the Domain Bacteria
they
could
take
advantage
of the already expreceded the animals with a long evolutionisting
trend
of
interaction
and foster coloary history in which almost all known metabolic pathways were derived (Gottschalk, nization by specific subsets of bacteria,
1986); prokaryotes evolved mechanisms for availing themselves of bacterial metabolic
the use of a wide variety of organic and diversity.
Most likely it was a combination of these
inorganic energy sources available in their
environments, forming interacting commu- two pressures that resulted in the formation
nities of mutually dependent individuals of sets of complex prokaryotic communities
(Madigan et al, 1997). The formation of in association with animals, which were difthe eukaryotic cell gave rise to the predict- ferent in composition from the bacterial
able persistence of subsets of interacting communities in the surrounding environcommunities, but did not divorce these ment. Until recently, however, biologists
units from specific interactions with less in- have focused primarily on the evolution of
tegrated prokaryotes in the environment mechanisms by which animals would inhibit interactions with prokaryotes. This viewpoint has been largely driven by:
INTRODUCTION
1
From the Symposium The Evolution of Development: Patterns and Process presented at the Annual
Meeting of the Society for Integrative and Comparative Biology, 26-30 December 1996, at Albuquerque,
New Mexico.
2
E-mail: [email protected]
1) our concepts of the nature of the immune system;
2) the emphasis on bacteria as animal
pathogens; and,
593
594
MARGARET J. MCFALL-NGAI
3) the technical difficulties associated with
exploring the establishment and maintenance of cooperative animal-bacteria
relationships.
Nonetheless, both recent changes in our
conceptual framework and advances in
technology have brought about reevaluations in all three of these areas. First, over
the past five years, there has been growing
support for a paradigm shift in immunology
in which the animal immune system rather
than being considered a non self recognition system (i.e., "self" vs. "non-self"), is
viewed as a system by which threats (i.e.,
"danger" vs. "non-danger") are detected
(Matzinger, 1994; Duncan and Edberg,
1995; Matzinger and Fuchs, 1996). This
change in viewpoint was primarily derived
as an attempt to understand autoimmune
disease, where a subset of "self" is treated
by the immune system as "danger." However, it also presents a very important
change in perspective for biologists who
study animal-bacterial mutualisms. This
shift in viewpoint allows for incorporation
of the idea that, under certain circumstances, animal tissues tolerate, and even promote, the presence of biotic "non-self." In
so doing, this new view reconciles the longstanding problem in immunology of whether to classify the bacterial symbionts as
"self" or "non-self." It has been recognized since the 1970s that indigenous microorganisms induce the production of host
antibodies poorly or not at all, especially
when the microbes are restricted to the enteric tract (Foo and Lee, 1974; Morishita
and Mitsuoka, 1973); and, in addition, intestinal bacteria can even share antigens
with the mucosa (Foo and Lee, 1974).
These observations suggested that indigenous microbiota evolved in a close immunological relationship with their host (Savage, 1977). Thus, a healthy animal can be
denned as a community comprised of interacting, yet widely divergent, genomes.
Taken in this context, the immune system
is seen to function to provide defense
against danger presented by cells both within and outside of the community. Any cell
within that community can become a threat,
such as miscommunicating cells of the host
(e.g., autoimmune disease or cancer) or opportunistic pathogens (e.g., overgrowing intestinal Escherichia coli); or, by the more
traditional view, the community itself can
be threatened by interlopers from the outside, which would span the range from viruses to grafts of cells from conspecifics.
Secondly, in addition to our historical
view of the immune system, the traditional
emphasis in microbiology on human bacterial pathogenesis has skewed our perspective (Garrett, 1994). The fact that human
history has been so tremendously affected
by outbreaks of bacterial disease (McNeill,
1977; McKeown, 1988) has promoted the
study of this type of animal-bacterial interaction, and incidentally overshadowed investigations of the normal, essential interactions of the healthy individual (Lee,
1985). An eventual outcome of this restricted view was the "antibiotic" solution to
microbial disease (Levy, 1992). Unfortunately, during this early phase of antibiotic
development, the plasticity of the bacterial
genome was not appreciated, and now bacterial pathogens with multiple antibiotic resistances have arisen and present a threat as
never before (Cohen, 1992; Neu, 1992). Yet
we continue to focus primarily on how
these pathogens affect animal tissues, rather
than how they interrupt the complex functions of the normal animal-bacterial community. It is clear that efforts must be directed toward exploring the nature of these
"normal" interactions.
Finally, there have been few experimental systems by which to approach the study
of the complex interactions between prokaryotes and animals. Advances in the application of molecular genetics to diverse
bacteria, the examination of nonculturable
bacteria within complex communities using
molecular probes (Hugenholtz and Pace,
1996), and the development of model systems, such as the Euprymna/Vibrio association described in the last section of this
contribution, are offering avenues by which
biologists can study cooperative interactions between animals and microbes.
These shifts in emphasis in the fields of
immunology and microbiology, and the development of new technology, have paved
the way for a changing perspective on an-
DEVELOPMENT OF COOPERATIVE SYMBIOSES
imal-bacterial relationships. Once one entertains the idea that evolution fostered the
creation of complex communities (=host +
microbes), the door is opened for exploration of how the interactions among the partners affect all levels of the biology of these
communities, including the evolution of the
developmental mechanisms of the animal
hosts. Thus, the development of an animal
must not only be concerned with the fidelity
of cell divisions and differentiations within
the single genome of that animal, but must
also incorporate mechanisms by which to
foster with a similar fidelity the cell divisions and differentiations of microbial cells
of a number of different genomes.
The field of biology that is concerned
with the influence of bacteria on animal development, as well as the development of
animal-bacterial community structure that
comprises the unit, has really only been
considered in any detail by the bacteriologists. To incorporate this new avenue of research, animal biologists must not only define experimental programs to examine
these interactions, but also create a strong
theoretical framework within which to formulate questions. In this paper I address
some of the issues that are relevant to a
consideration of the influence of bacteria on
animal development. Specifically, I consider key aspects of the relationships between animals and bacteria that have bearing on host development, discuss the state
of our current database in this field, and finish by summarizing what is known for one
model, the Euprymna-Vibrio symbiosis.
THE NATURE OF COOPERATIVE
ASSOCIATIONS
Specific characteristics of cooperative
animal-bacterial associations, sometimes
termed "mutualisms," are likely to affect
animal host developmental processes. Some
of the more conspicuous influences are discussed below.
Community structure
In most cooperative associations between
animals and bacteria, the animal host provides a nutrient-rich environment that promotes bacterial growth in exchange for the
products of specific bacterial metabolic
595
pathways (such as those mediating chitin or
cellulose breakdown, or vitamin synthesis;
Madigan et al., 1997). As such, the most
conspicuous and widespread type of animal-bacterial interaction is between hosts
and their specific consortia of enteric tract
microbiota, which usually are the primary
cell type; e.g., in a typical omnivorous
mammal, there is a 10:1 ratio of bacterial
to animal cells (Savage, 1977, 1986; Duncan and Edberg, 1995). The enteric consortium usually consists of two principal components: the resident, indigenous microbiota, which are believed to have coevolved
with the host, and nonspecific "tourist" microbes, which have been introduced from
the environment and do not form permanent
associations (Savage, 1977).
While most animals have, and may require, specific interactions with microorganisms in their enteric tracts, occasionally
in evolutionary history, specializations have
occurred that drive the structure of the community, or a portion of the community, in
one of two ways: toward a highly distinctive consortium driven by unusual diets,
such as in the rumen (Hungate, 1966) or
termite hindgut (Breznak, 1982); or, toward
organs that involve a two-partner symbiosis, i.e., one species of bacterium in a specialized host organ, such as a light organ
(McFall-Ngai and Toller, 1991; Haygood,
1993). Both of these types of derived conditions provide convincing illustrations of
how interactions with microorganisms can
profoundly influence both the evolution of
specific animal groups and the ecology of
their associated biosphere (Margulis and
Fester, 1991). For example, the radiation of
the ruminant mammals must have been dependent on their association with the rumen
microbiota (McFarland et al., 1979; Stahl,
1985); and, the evolution and ecology of
coral reef (Stoddard, 1969; Newell, 1972)
and deep sea hydrothermal vent communities (Grassle, 1986; Tunnicliffe, 1988; Childress and Fisher, 1992) appear to have been
driven by associations of animals with autotrophic microorganisms.
Maintenance of community integrity over
time—transmission of microbes between
host generations
Whether the community defining the
symbiosis is consortial or two-partner, there
596
MARGARET J. MCFALL-NGAI
is a growing body of evidence indicating Lee and Ruby, 1994). In both direct and
that bacteria influence of the development indirect cyclic transmission, the bacteria are
of the animal tissues with which they as- not actually present during embryogenesis.
sociate (Gordon and Pesti, 1971; McFall- Nonetheless, during this period, the animal
Ngai and Ruby, 1991; Schwemmler and must develop tissues to promote establishGassner, 1989; Saffo, 1992). How and ment of the association, and often rudiwhen these influences express themselves is ments of the more mature organ. For exdependent on mechanisms by which the as- ample, over evolutionary time selection
sociations are maintained between genera- must have occurred to incorporate into the
tions.
host's developmental program the differThe principal modes of transmission entiation of cells with surface properties inhave been termed transovarian (or vertical) volved in recognition and adherence of the
and cyclic (or horizontal) (Douglas, 1994). appropriate environmental bacteria. The
In transovarian transmission, the bacterial first interactions of the host with such cysymbionts are translocated to the ovary and clically transmitted bacterial partners occur
are transmitted in or on the egg. As such, at either hatching or birth. These interacthey participate directly in the process of tions of the animal and bacteria during postembryogenesis of the host, as well as in embryonic development can have profound
postembryonic development. For example, conspicuous effects on the maturation of
at least 11 % of all insect species examined the tissues of the host (Gordon and Pesti,
vertically transmit specific symbionts that 1971; McFall-Ngai and Ruby, 1991).
In addition to transmission between genreside in a tissue called the bacteriome (formerly, mycetome). These symbionts pro- erations, the dynamics of bacterial comvide vitamins and/or other nutrients that al- munities may be crucial to the normal prolow these insects to survive on nutritionally grams that occur within and between depoor diets (Douglas, 1989). In cyclic trans- velopmental stages in the life history of an
mission, the bacterial partners are acquired individual, particularly in species with comfrom the environment anew each genera- plex larval life histories. In more simple life
tion. While vertical transmission is restrict- histories, such as that of mammals, changes
ed to a limited number of taxa, cyclic in- in gut microbiota are well documented in
oculation with specific microbiota that be- the early postnatal development (Savage,
come associated with the enteric tract and/ 1977). Colonization by pioneer species is
or integument probably occurs in all ani- followed by an ecological succession, a
progression that eventually results in a stamals (Savage, 1977, 1986).
Cyclic transmission between generations ble, climax community of interacting mican be further divided into direct and indi- crobes. The community structure is then alrect. In direct transmission, the transfer of tered as a result of a dramatic dietary
bacterial symbionts between generations is change, such as that occurring at weaning
facilitated by the adult host, and is often a (Gordon and Pesti, 1971; Savage, 1977).
hallmark of animal/bacterial communities While these processes have been studied
with special metabolic capabilities. For ex- extensively in mammals, very little inforample, termite juveniles are fed symbiont- mation is available on the nature of the baccontaining feces of the adults (Breznak, terial communities associated with inverte1982), and cows inoculate their calves with brate larvae, and the extent to which these
the rumen microbiota during grooming bacteria influence the often complex devel(Dehority and Orpin, 1988). With indirect opmental programs of their hosts. However,
cyclic transmission, the most prevalent several examples have been reported in
form of transmission between generations, which host species have been demonstrated
the young acquire their inoculum from the to require bacteria for proper settlement and
broader environment without the interces- metamorphosis (Sheltema, 1974; Hadfield
sion of conspecifics, although the adult et al., 1994). For example, the larvae of
population may be more or less responsible many sessile benthic species assess the nafor seeding the environment (Drasar, 1974; ture of substrate biofilms, and only settle
DEVELOPMENT OF COOPERATIVE SYMBIOSES
597
association, the bacterial partner may be superficial (e.g., on the skin), or invested in
Superficial
the microvilli of a simple tissue (e.g., the
enteric epithelium) or of a complex organ
(e.g., squid and fish photophores). Alterdevelopmental/
evolutionary progression
natively, the bacterial symbionts can be
within the host cell, either surrounded by a
vacuolar membrane or free in the cytoplasm. For the intracellular symbioses that
Intracytoplasmic
are cyclically transmitted, not only does this
progression of intimacy between the bacterial and animal cells obviously occur during establishment of the mature relationship, but also it is believed to represent an
[cslt/vacuolB membrane]
evolutionary progression; i.e., intracellular
symbioses are presumed to be more derived
surface contact
than extracellular symbioses, although there
is little evidence to support this idea (Smith,
exchange of gene products
1979).
The location of the bacterial symbiont in
relation
to the animal cell does not neceshorizontal gene transfer
•'>i\
sarily correlate with the degree to which the
partners can and do exercise control over
FIG. 1. Intimacy of host and microbial cells as defined by position and mechanisms of reciprocal sig- one another's metabolism. Instead, the innaling. A. From top to bottom, progression of micro- tensity of the dialogue, as reflected in
bial cells from an extracellular, superficial location on changes in gene expression of the partners,
the host cell, to free within the host cytoplasm, i.e., should be more meaningful in defining the
without a vacuolar membrane. In intracellular, cyclic
transmission this progression occurs developmentally, impact of the symbiosis on their biology
and may also represent an evolutionary trend toward (Fig. IB). Such communication can be meincreasing interdependence of the symbiotic partners. diated remotely through surface compoB. Modes of signal exchange between host and micro- nents of the plasma or vacuolar membranes,
bial symbiont [see text for details], a = host genes; b
and second messenger pathways may be re= microbial genes.
sponsible for transmitting the information
to the partner's genomes. Similarly, the
and metamorphose when bacteria with pre- cells can exchange gene products or metabscribed characteristics are present (e.g., olites, which either directly or indirectly inHofmann and Brand, 1987; Fitt et al., 1990; fluence gene expression within the recipient
Maki et al., 1990). Yet how prevalent such cells. For example, during the interaction of
a dependency might be is an area that re- Vibrio cholerae with mammalian intestinal
mains to be explored.
cells, the bacterium produces the protein
cholera toxin (Spangler, 1992). The A subLocation of the cells of the partners and
unit of this protein is introduced into the
its relationship to mechanisms of
host cell, where it perturbs the metabolism
reciprocal signaling
of that cell, which leads to changes in host
Between species interactions.—Different cell gene expression (Lycke and Strober,
bacteria colonize a variety of locations on 1989; Spangler, 1992). In contrast to animal
or in the cells of their host. (Fig. 1A) The cells, bacteria do not typically import protopography of these associations may be teins, so animal cells usually influence bacexpected to influence the extent to which terial gene expression through smaller sigthe bacteria participate in the differentiation nal molecules. For example, leguminous
of host cells as well as the overall mor- plants release flavenoids into the soil that
phology of the tissues and organs that house are taken up by specific bacteria termed
the microbiota. Thus, in a given cooperative "rhizobia" (Long, 1989; Hirsch, 1992).
—
- ».
K--—
v
598
MARGARET J. MCFALL-NGAI
These flavenoids induce expression of bacterial genes that promote the development
of the symbiosis between the legume and
its specific species of rhizobia, which ultimately leads to formation of the mature, nitrogen-fixing root nodule (Appleby, 1984;
Gussin et al., 1986).
While these more indirect mechanisms of
host/microbe exchange have been recognized for decades, a grasp of the implications associated with direct introduction of
a genetic element into the cell of a symbiotic partner, either historically or during the
ontogeny of a given symbiotic association,
represents a frontier area in biology. The
genetic makeup of an individual can be
thought to consist of four types of elements:
(i) the principal components of the nuclear
genome that share a single, continuous evolutionary history and that obey the laws of
Mendelian genetics; (ii) genes in organelles,
such as the mitochondria and plastids,
which coevolved with the nuclear genome
of the eukaryotic cell; (iii) those associated
genetic elements that may be derived from
that genome, but have become independent,
such as retroviruses in eukaryotes and
phage or plasmids in prokaryotes; and, (iv)
elements with a completely independent
evolutionary origin that are, or have been,
inserted into the genome—a phenomenon
referred to as horizontal gene transfer (Amabile-Cuevas and Chicurel, 1993). The best
documented example of horizontal gene
transfer and its influence on host development occurs in the Agrobacterium tumefac/ens-dicotyledonous plant pathogenesis
called crown gall disease. The virulence
genes of this bacterium are encoded on satellite DNA, the Ti plasmid, which is transferred into susceptible cells of the plant host
(Raineri et al, 1993; Madigan et al, 1997).
Genes on this plasmid are incorporated into
the plant's nuclear genome, where they
cause changes in host gene expression and
protein synthesis that result in the formation
of the gall, a hypertrophied tissue that houses the bacterial symbionts (Hooykaas,
1994). In animal-bacterial associations, few
such striking examples have been described, although convincing molecular evidence suggests that such transfers have occurred in the past; i.e., chromosomal genes
have been identified in both bacteria and
animals that have sequences more closely
related to the genes of a distantly related
organisms than to those of the nuclear genome of that individual (Dyer, 1989; Amabile-Cuevas and Chicurel, 1992; Smith et
al, 1992; Amabile-Cuevas and Chicurel,
1993). Further, recent studies on symbionts
of mammalian intestines have shown that
bacterial symbionts can donate stable, transcribable pieces of DNA into host eukaryotic cells (Courvalin et al., 1995). Although
these latter genes are not directly associated
with the germ line of the host, the degree
of mobility of a particular genetic element
and the susceptibility of the recipient cell
are likely to be heritable traits that may influence developmental patterns in the partners of the symbiosis.
Within species interactions.—The intimate interactions described above between
symbiotic cells can be transmitted to the
other cells comprising the populations both
within the host and among the symbiotic
bacteria themselves, a process which often
determines whether a symbiosis will be established, the nature of the interaction
(pathogenic or mutualistic), and the course
of development of the association. As in the
communication between partners, there exist examples of both indirect and direct influences on the genome. Indirect effects that
have been extensively studied in the host
include the induction of cortical cell division in the root by interactions with rhizobia at the plant root hair (Long, 1989;
Hirsch, 1992) and remote changes in the
animal immune system that result from antigen sampling in the intestine (Kagnoff and
Kiyono, 1996). Among bacteria, one of the
best described examples of within-species
communication is a process called "quorum
sensing" (Fuqua et al., 1994, 1996). In recent years this form of cell-cell interaction,
first described in marine luminous bacteria
(Nealson et al., 1970), has been discovered
in a rapidly growing list of bacterial species
that, while phylogenetically diverse, share
the characteristic of being either pathogenic
or cooperative symbionts of certain animals
or plants. Briefly, each of these bacterial
species secretes one of a class of structurally related Af-acyl-homoserine lactone mol-
DEVELOPMENT OF COOPERATIVE SYMBIOSES
ecules (autoinducers). If cells of that particular species are in abundance, their specific
autoinducer will accumulate in the surrounding environment. These bacteria then
sense this elevated concentration, which
signals the presence of a "quorum" of conspecifics, and leads to the induction of certain genes whose products are required or
useful only when the bacteria are colonizing a host tissue. Well studied examples of
such regulated activities are luciferase synthesis in symbiotic luminous bacteria (Nealson et al., 1970), and the production of the
enzyme elastase by pathogenic pseudomonads (Passador et al., 1993). The prevalence
of this form of signaling between bacterial
cells in host tissues, as well as the homology of the systems as revealed by biochemical and genetic comparisons, indicate its
antiquity, ecological plasticity, and success
in bacterial symbiosis, and encourage an
examination of other host-microbe interactions for its presence.
In addition to such molecular signaling
between symbiotic cells, direct gene transfer within a bacterial species has been reported for certain bacteria during initiation
of a pathogenesis. (Mel and Mekalanos,
1996). In these cases, a set of "virulence"
genes are carried by only a small subset of
the bacterial population present outside of
the host. Interaction with the host induces
the transfer of these genes, which are carried either on bacteriophages or on plasmids, and renders the recipient capable of
expressing required virulence determinants.
For example, only a small proportion of the
free-living population of the plant pathogen
Agrobacterium tumifaciens carries the virulence Ti plasmid when outside of an association with a plant. However, the plant
provides a confined space for the accumulation of an autoinducer that promotes the
transfer of the plasmid to other conspecifics
rendering them virulent (Farrand, 1993).
Another example has recently been demonstrated in the V. cholerae association
with mammals. Several of the cholera virulence genes are carried on a bacteriophage
that is transferred to conspecifics under
conditions encountered in the host tissues
(Waldor and Mekalanos, 1996).
599
THE EXISTING DATABASE
Studies of cooperative associations of
hosts with bacteria have produced many elegant studies on a wide array of associations (reviewed in Saffo, 1992; Douglas,
1994; Losick and Kaiser, 1997). However,
there are three specific research programs
that have provided extensive experimental
bases and that together are of heuristic value for developmental biologists seeking to
understand how bacteria may influence general developmental programs of animals.
They are research on: (i) germfree and gnotobiotic animals, (ii) bacterial pathogens,
and (iii) the plant/nitrogen-fixing bacteria
(rhizobia) symbioses. Rather than reviewing in detail what is known about the developmental biology of each system, I discuss below what each system might contribute to our understanding of the effects
of microorganisms on animal developmental processes, as well as some caveats.
With the exception of research on the rumen, studies of germfree and gnotobiotic
animals were until recently the principal
source of information on how extracellular
relationships with cooperative bacteria, the
most prevalent form of symbiosis, affect the
biology of their hosts (Heneghan, 1973;
Coates and Gustafsson, 1984; Wostmann et
al., 1984). The results of this area of research have revealed that not only do germfree animals fail to thrive without intervention (Gordon and Pesti, 1971), but the tissues with which they associate and the immune system do not mature in the absence
of interactions with the natural complement
of bacteria (Gordon and Pesti, 1971; Woolverton et al., 1992; Duncan and Edberg,
1995).
However, using germfree and gnotobiotic
animals to understand the interactions of
hosts with their indigenous gut microbiota
has met with certain technical problems.
Firstly, although many experiments have
compared these animals with their normal
conspecifics, the nature of what actually
constitutes the "normal" animal state is
still poorly understood; i.e., even aspects as
fundamental as the composition of the microbial community have not been determined for any animal species, including
600
MARGARET J. MCFALL-NGAI
Homo sapiens, in which it has been extensively studied (Savage, 1977, 1986; Duncan
and Edberg, 1995). Certain species within
the community have been difficult to designate as either indigenous or nonspecific,
because of sampling problems. Further, it is
not well understood how either changes in
the diet of an adult host or the maintenance
of the host in captivity affect community
composition. Finally, despite recent advances in the ability to culture specific types of
bacteria, particularly anaerobic species, a
significant portion of the gut microbiota
consists of bacteria that have not yet been
grown in the laboratory (Duncan and Edberg, 1995). Newly developed molecular
methods for cataloguing unculturable microbes (Hugenholtz and Pace, 1996) may
help to define more fully the community.
However, even once the community is more
precisely defined, extensive experimental
studies with germfree and gnotobiotic systems will continue to be difficult, because
the multispecies community of microbes
certainly works as a unit, and the results of
studies attempting to isolate and experimentally manipulate one component in the absence of the others would be difficult to interpret.
In spite of these inherent limitations of
using germfree and gnotobiotic animals to
study the dialogue between normal microbiota and their hosts, genetic engineering of
specific members of the intestinal bacterial
community is providing insight into the
"messages" that pass between the host and
bacterial cells during the onset of the association. For example, by exposing germfree mice to either wild-type or a genetically altered strain of the indigenous enteric
bacterium, Bacteroides taiotaomicron, Bry
et al. (1996) showed that production of normal levels of glycoconjugates by intestinal
epithelia requires specific interactions with
bacteria. Continued studies in this area
promise to provide further insight into how
bacteria participate in specific aspects of enteric remodeling during development.
Perhaps the most robust database on animal-bacterial interactions has resulted from
research into the nature of pathogenic associations. Numerous studies of bacterial
pathogens and their activities have revealed
the diversity of mechanisms by which such
microbes interact with host cells, as well as
the diversity of host cellular processes that
are involved in responses to pathogens (Iglewski and Clarke, 1990; Salyers and Whitt,
1994; Cossart et al, 1996). The degree to
which fully defining these interactions
might lead to a better understanding of all
types of animal/bacterial interactions, intracellular to extracellular, cooperative to
pathogenic, remains to be determined.
However, one might suspect that using
pathogens to understand a normal condition
would be similar to trying to understand the
cooperative workings of a complex, multicultural city by examining the behavior of
wartime occupation forces. With this limitation recognized, studies of pathogens
have revealed to developmental biologists
the extent to which bacteria are capable of
altering the morphology, biochemistry and
molecular biology of the cells and tissues
with which they associate (Salyers and
Whitt, 1994).
Studies of bacterial pathogenesis have revealed that often normal host cell processes
are incorporated into responses to pathogens, and pathogens themselves have a
complex set of genes and gene products that
have been associated with the mediation of
their interactions with the host (Cossart et
al., 1996). However, whether host and bacteria are using the same cellular responses
that would be used to mediate cooperative
associations is not well defined. Further, an
inherent anthropocentric view has labeled
as pathogens bacterial species that cause
disease in humans, although their primary
ecological niche might be to occur as cooperative symbionts in other hosts. In such
associations, the same molecular "currency" of exchange may be used, with modifications resulting in a benign association,
rather than a pathogenesis. For example, recently it was reported (Weis et al., 1996)
that a halide peroxidase, erstwhile associated only with tissues responding to pathogens, occurs in high concentrations in the
symbiotic organ of Euprymna scolopes, a
sepiolid squid that supports a cooperative
association with the luminous bacterium
Vibrio fischeri. Similarly, on the bacterial
side, cholera-toxin-like molecules and their
DEVELOPMENT OF COOPERATIVE SYMBIOSES
601
regulators have been found in a number of coordinate interplay between cells with two
luminous bacteria that are either enteric or very different genomes.
light organ symbionts (Reich and SchoolDespite the fact that the hosts in the lenik, 1994, 1996; Reich et al, 1997). These gume/rhizobia symbioses are plants, studies
two examples suggest that certain mole- of root nodule development illustrate to ancules that have traditionally been associated imal developmental biologists just how
with pathogenesis may have evolved in- complex a dialogue can be required that restead for a variety of interactions of bacteria sults in a stable, cooperative association bewith animals, and it is the identity of the tween a microbe and a multicellular organhost and its response that determines the na- ism.
ture of the symbiosis.
THE EUPRYMNA/VIBRIO COOPERATIVE
An understanding of how cooperative
ASSOCIATION
microorganisms can influence the developDetailed reviews of the basic development of a multicellular organism has been
best studied in plant/bacterial associations, mental program of the Euprymna scolopes
most notably the symbiosis between legu- light organ have recently appeared in Amerminous plants and nitrogen-fixing bacteria ican Zoologist (McFall-Ngai, 1994; Mont(rhizobia) (Long, 1989; Nap and Bisseling, gomery and McFall-Ngai, 1995). Thus, for
1990; Brewin, 1991; Hirsch, 1992; van the purposes of this contribution, I will simRhijn and Vanderleyden, 1995). Because ply provide an overview of the symbiosis,
these symbioses involve one host and one and then focus on how this symbiosis inmicrobial species, both of which can be terfaces with the existing, above-described
grown independently in culture, they have databases and how it provides new opporoffered an unparalleled, manipulable exper- tunities to developmental biologists.
Euprymna scolopes is a sepiolid (bobtail)
imental system for the study of the establishment and maintenance of a symbiotic squid endemic to the Hawaiian archipelago
association. Free-living rhizobia invade root (Berry, 1912). This nocturnal predator burhairs of the plant, and induce the formation ies in the sand of the shallow waters of the
of an infection thread in the root (van Rhijn backreef during the day, and emerges just
and Vanderleyden, 1995; Madigan, 1997). after dusk to forage in the water column
This structure acts as a conduit through (Moynihan, 1983). The conspicuous light
which the bacteria are transported from the organ in the center of the mantle cavity of
root surface to its cortex, where the mature E. scolopes houses a culture of the bactenitrogen-fixing nodule will form. The de- rium Vibrio fischeri (McFall-Ngai and
velopmental program of the root nodule, Montgomery, 1990). It is hypothesized that
which spans a few weeks, requires that the the luminescence of these bacteria is used
host and microbes be able to recognize their as a camouflaging mechanism by the squid,
specific partner, as well as induce responses where ventral luminescence is emitted of
that orchestrate the morphogenetic process the same quality (color and intensity) as
(Long, 1989; Hirsch, 1992). Decades of downwelling moonlight or starlight so that
study of these associations have revealed the squid casts no shadow. This behavior,
that dozens of genes in each partner, ex- called counterillumination, could be used
pressed in a specific, reciprocal order, are by the squid to avoid being detected by eiessential for completion of development ther predators or prey (McFall-Ngai, 1990).
The gross morphology of the symbiotic
(van Rhijn and Vanderleyden, 1995; Denarie et al., 1996). The volley of signaling light organ can be observed by ventral disbetween the host and bacteria is similar to section of the squid, which reveals a bithe communication that occurs between tis- lobed organ associated with the ink sac
sues within animal organs, such as the eye, (McFall-Ngai and Montgomery, 1990).
during their proper development. However, Histological cross sections of the organ
the salient difference lies in the fact that in show that it consists of a series of tissues
the plant/bacteria interaction, there exists that together render it similar in architecture
to the vertebrate eye (McFall-Ngai and
602
MARGARET J. MCFALL-NGAI
Montgomery, 1990; Montgomery and
McFall-Ngai, 1992). In the center of the organ is a network of epithelial cells with
which the extracellular V. fischeri are directly associated. Surrounding this photogenic tissue is a thick reflector, which
serves to direct luminescence ventrally out
of the body of the squid. The reflector is in
turn surrounded by diverticula of the ink
sac, which act as both choroid and iris analogues, i.e., to prevent stray light from
passing dorsally out of the animal, and to
rotate dynamically around the ventral portion of the light organ to modulate the intensity of light emission. The entire ventral
surface of the light organ is covered with a
thick, transparent lens, which diffuses the
point-source luminescence over the ventral
surface of the animal (Montgomery and
McFall-Ngai, 1992; Weis et al, 1993).
Sepiolids with light organ symbioses exhibit cyclic transmission of their symbionts.
Adult females of E. scolopes lay clutches
of eggs (ranging in number from 50 to 450)
on hard substrates in their environment (M.
McFall-Ngai, personal observation). At
around 20 days post-fertilization, following
no parental care by the adults, the juvenile
squid hatch from their egg cases (Montgomery and McFall-Ngai, 1993). The juveniles emerge just after dusk {i.e., light inhibits hatching; M. McFall-Ngai, personal
observation) with light organs that are devoid of bacteria, and if these juveniles are
not exposed to V. fischeri, a common constituent of the free-living bacterioplankton,
the light organ remains sterile (McFall-Ngai
and Ruby, 1991). However, under normal
conditions, V. fischeri is abundant (approx.
200 cells/ml) in the seawater where adults
are found (Lee and Ruby, 1994), most likely because of a diel behavior characteristic
of the adults in which they release 90% of
their bacterial culture into the water column
each day at dawn (Boettcher et al., 1996).
The juvenile E. scolopes ventilate symbiont-rich water into their mantle cavity and
inoculation of the juvenile light organ occurs. In laboratory experiments using a similar density of bacteria as that occurring in
the field, inoculation of the organ occurs in
minutes to hours after hatching, a process
that can be monitored by measuring light
production of the growing symbiont culture
(Ruby and Asato, 1993).
During embryogenesis, when no bacterial symbionts are present, the squid host
develops an organ anatomy and morphology that appears to promote efficient colonization of the light organ (Montgomery
and McFall-Ngai, 1993). Superficially, each
lateral face of the organ is covered with a
complex, ciliated, microvillous epithelial
field, composed of an anterior and posterior
appendage and an area surrounding three
pores. Observations with high speed cinematography showed that the cilia surrounding the pores and on the appendages, which
form a ring, entrain bacteria-ladened water
toward the pores (M. McFall-Ngai and R.
Emlet, personal observation). The pores
lead by long ducts to epithelia-lined crypts
that form sequentially through the embryonic period, presumably through invagination (Montgomery and McFall-Ngai, 1993).
In normal development, once the bacterial culture has become established, the superficial ciliated, microvillous field regresses within days (Fig. 2 and Fig. 3), and the
light organ goes on to elaborate those tissues associated with function of the mature
light organ, i.e., light modulation (McFallNgai, 1994). This process, which occurs
from about Day 4 to Day 21 post-hatching,
involves a dramatic thickening of the reflector, development of the ink sac diverticula, elaboration of the lens, and an overall
remodeling of the rounded hatchling organ
into the bilobed morphology characteristic
of the adult (McFall-Ngai, 1994). In addition to these superficial changes in the organ, during normal symbiotic development
the crypt cells associated with the bacteria
swell to a volume four times that of the
hatchling crypt cells, and the microvillar
density increases four fold (Fig. 2).
This developmental program strikingly
resembles that of the legume/rhizobia symbioses both in pattern and timing. One principal difference is the presence in E. scolopes of a group of cells in the hatchling,
i.e., the superficial ciliated, microvillous
field, whose sole purpose appears to be the
mediation of the inoculation process. Cells
of rhizobia infect the already existing, functioning root hair, and induce creation of the
DEVELOPMENT OF COOPERATIVE SYMBIOSES
Initiation
•
Colonization
603
• Persistence
FIG. 2. Early development of the squid/vibrio symbiosis. The diagrams represent a ventral view of the right
side of the juvenile light organ. At the time of initiation (upper, left-most diagram), the ciliated field functions
to promote infection by flagellated, V. fischeri in the environment. The bacteria enter the hatchling organ through
pores on its surface, travel down ciliated ducts and come to reside in crypts lined with columnar epithelia (lower,
left-most diagram). Bacterial colonization of the light organ (middle diagram) leads to massive cell death of the
superficial, ciliated field, beginning at about 12 hr post-initiation at the tips of the appendages and at the edge
of the field (depicted as small circles on the organ). This cell death program results in the eventual loss of the
entire superficial, ciliated field (upper, right-most diagram), which characterizes the persistence stage of the
symbiosis, i.e., that period when a stable association has been established. In addition, the host cells swell and
the microvillar density of the brush border increases 4-fold over the first 4 days of the symbiosis. Bacterial cells
lose flagellation and reduce in size in response to the symbiotic state.
infection thread, which mediates the inoculation of the root nodule. As in the squid/
vibrio symbiosis (McFall-Ngai, 1994), the
maturation of the functional, i.e., nitrogenfixing, organ is an elaborate process that
follows a transformation from the inoculation morphology and requires a few weeks
after infection to complete (Long, 1989).
In addition to being similar to the legume/rhizobia symbioses in overall developmental pattern, the squid/vibrio symbiosis offers developmental biologists similar
opportunities to dissect the process of symbiotic organ morphogenesis. Each of these
types of associations, in contrast with consortial symbioses, has one host and one microbial species, both of which can be maintained in culture independent of the other
partner. Further, both of these symbioses involve a prokaryotic partner, so that molecular genetics are readily applied to their
study (Long, 1989; Ruby, 1996). In contrast
with the legume/rhizobia symbioses no genetics is as yet available in the squid host.
However, development of a genetic system
may be possible; maintenance of a colony
of only 12 female squid yields between
50,000 and 100,000 eggs per year (M.
McFall-Ngai, personal observation), and
techniques for raising the squid through
their life cycle have been recently refined
at the Marine Biological Laboratories in
Woods Hole, Mass. (R. Hanlon, personal
communication). However, manipulation of
the host genome, using antisense probes or
viral vectors may be more promising than
conventional breeding approaches to genetics.
Experimental manipulation of the squid/
vibrio symbiosis have shown that the partners have a profound influence on each other's development. Thus far, we have focused on the early developmental events,
i.e., those that occur within the first four
days of the symbiosis, a period which can
be divided into stages of initiation, colonization and persistence (Fig. 2). By maintaining hatchlings in the presence (symbi-
604
MARGARET J. MCFALL-NGAI
BACTERIAL
DEVELOPMENTAL
PATHWAY
SYMBIOSIS
DEVELOPMENTAL
PATHWAY
HOST
DEVELOPMENTAL
PATHWAY
(embryogenesis)
Hatching
I
• expulsion behavior
- cell death signalling
- brush border
remodeling
- /uxgenes induction
- full colonization
- flagella loss
- size decrease
- catalase induction (?)
1 day
- CMS regression
- crypt cell swelling
- oxidatlve stress
- competitive exclusion
- iron acquisition
- cyclic recolonization
4 days
- lens formation
(ALOH induction)
- reflector elaboration
- ink sac remodeling
>10 days
FIG. 3. Reciprocal signaling during the development of the squid/vibrio symbiosis. Host and bacterial development are characterized by changes in behavior, morphology, and biochemistry in response to interactions with
one another. Molecular genetic studies of the bacteria, and responses of the host to mutant strains, indicate that
a dynamic dialogue occurs between the two. The result of a successful dialog is the symbiosis developmental
pathway, which leads to the establishment of the stable, mature association. ALDH = aldehyde dehydrogenase,
the principal protein of the light organ lens; CMS = ciliated, microvillous surface; LPS = lipopolysaccharide
surface of bacterial symbionts; lux = descriptive of genes associated with the bacterial operon that encodes light
production capability; ROS = reactive oxygen species; Sym = symbiosis.
otic) and absence (aposymbiotic = exposed
to other bacterioplankton, in the absence of
V. fischeri), we have been able to determine: 1) what components of the developmental program require interaction with
symbionts; and, 2) which of these components require persistent influence of symbionts, and which require only transient
interaction with V. fischeri cells.
In the absence of the symbionts, the su-
perficial epithelial field of the hatchling
light organ does not regress (Montgomery
and McFall-Ngai, 1994), the crypt cells do
not swell (Montgomery and McFall-Ngai,
1994) and the brush border does not change
in microvillar density (L. Lamarcq and M.
McFall-Ngai, personal observation). The
loss of the superficial field occurs as a result
of bacteria-induced cell death (Montgomery and McFall-Ngai, 1994), which pro-
DEVELOPMENT OF COOPERATIVE SYMBIOSES
605
gresses over the first four days of the sym- fischeri has resulted in the description of
biosis. Despite the fact that this process re- several classes of mutant strains that have
quires four days to complete, the presence revealed bacterial genes that are essential
of bacterial symbionts is only required for for proper colonization of the host organ.
approximately 12 hr (Doino and McFall- Ruby (1996) has divided these mutants into
Ngai, 1995); i.e., antibiotic treatment to re- three categories, initiation, accommodation
move bacteria from the light organ after this and persistence phenotypes. Initiation muearly stage still results in the complete re- tants are incapable of any colonization of
gression of the superficial field (Doino and the light organ; accommodation mutants
McFall-Ngai, 1995). Thus, the bacteria never colonize to levels characteristic of
must be delivering a signal that produces wild type; and, persistence mutants are able
an irreversible cascade of developmental to enter the light organ, may or may not
events in the host. In contrast, the changes colonize fully, but are not able to persist in
in the brush border require a continued in- the light organ past the first day or two. The
teraction with the symbionts; antibiotic cur- behavior of several of these mutants suging at any time results in a return to hatch- gests that the symbiont is responding to the
ling or aposymbiotic levels of microvillar change in oxidative environment of the
density (L. Lamarcq and M. McFall-Ngai, crypts.
personal observation). The extent to which
These experimental studies of the squid/
cell swelling requires persistent interaction vibrio symbiosis, while still in their infancy,
remains to be determined.
are beginning to reveal the elaborate diaDuring this early period in the initiation logue that goes on between E. scolopes and
of the squid/vibrio symbiosis profound V. fischeri during the establishment and
changes are also seen in the morphology of eventual maintenance of the symbiotic
the bacterial symbiont (Fig. 3). Between 12 state. As in the legume/rhizobia symbioses,
and 24 hr after initiation of colonization, there are indications that the dialogue is rethe bacteria have lost the tuft of sheathed, ciprocal; i.e., the progression through depolar flagella that is characteristic of this velopment requires a volley of gene exbacterial species in its free-living niche pression between the partners (Fig. 3).
In addition to providing insight into the
(Ruby and Asato, 1993). In addition, they
have decreased approximately 8-fold in vol- mechanisms by which bacteria may influume to dimensions of approximately 1 mm ence the normal development of the epitheby 0.5 mm as differentiated symbionts lial tissues with which they associate, the
(Ruby and Asato, 1993). In addition, the squid/vibrio system promises to help biolbacterial symbionts induce luminescence ogists understand how cooperative associ1,000-fold (Ruby and Asato, 1993), and ations differ from pathogenic ones. The
their growth rate is curtailed from an initial finding of host halide peroxidases in the
doubling time of 0.5 hr to one of between light organ (Tomarev et al., 1993; Weis et
ah, 1996) and genes associated with vibrio
5 and 6 hr (Lee and Ruby, 1994).
Several biochemical and molecular pathogenesis in the bacterial symbiont
changes in one or the other of the partners (Reich and Schoolnik, 1994, 1996) suggest
correlate with these morphological and an- that some of the language between the partatomical modifications. Most notably, a sig- ners is conserved between the two very difnificant modulation of indicators of oxida- ferent types of associations.
In summary, the study of the bacterial
tive stress in the host light organ occurs
over the first few days of the association; influence on animal development is a field
specifically, there is some evidence of a on the frontier. As both the abiotic and bitransient increase in toxic oxygen species, otic environments of animals become afperhaps the result of a respiratory burst of fected by human intervention, it becomes
the crypt cells, followed by a decrease in increasing important to understand how anthe levels of these molecular species (Small imals form and maintain stable associations
with environmental microbes, and the conand McFall-Ngai, 1993).
Application of molecular genetics to V. ditions that determine whether such asso-
606
MARGARET J. MCFALL-NGAI
ciations become cooperative or pathogenic.
Just as certain nations, such as Great Britain
and Germany, have gone through periods of
history as either allies or enemies, it is likely that the delicate balance that is created
in cooperative associations has been historically, and will continue to be, susceptible
to forces that threaten to compromise detente.
ACKNOWLEDGMENTS
Thanks to Mark Martindale and Billie
Swalla for organizing the symposium, and
to Edward Ruby and the members of my
laboratory for numerous helpful discussions
and comments on the manuscript. Research
reported in this contribution on the squid/
vibrio symbiosis is supported by NSF grant
IBN 96-01155, ONR grant N00017-911347, and NIH R01RR12294.
REFERENCES
Amabile-Cuevas, C. F. and M. E. Chicurel. 1992. Bacterial plasmids and gene flux. Cell 70:189-199.
Amabile-Cuevas, C. E and M. E. Chicurel. 1993. Horizontal gene transfer. Amer. Scientist 81:332-341.
Appleby, D. A. 1984. Leghemoglobin and Rhizobium
respiration. Annu. Rev. Plant Physiol. 16:443478.
Berry, S. 1912. The Cephalopoda of the Hawaiian Islands. Bull. U. S. Bur. Fish. 32:255-362.
Boettcher, K. J., E. G. Ruby, and M. J. McFall-Ngai.
1996. Bioluminescence in the symbiotic squid Euprymna scolopes is controlled by a daily biological rhythm. J. Comp. Physiol. 179:65-73.
Brewin, N. J. 1991. Development of the legume root
nodule. Annu. Rev. Cell Biol. 7:191-226.
Breznak, J. A. 1982. Intestinal microbiota of termites
and other xylophagous insects. Annu. Rev. Microbiol. 36:323-343.
Bry, L., P. R. Falk, T. Midtvedt, and J. I. Gordon. 1996.
A model of host-microbial interactions in an open
mammalian ecosystem. Science 273:1380—1383.
Childress, J. J. and C. R. Fisher. 1992. The biology of
hydrothermal vent animals: Physiology, biochemistry, and autotrophic symbioses. Oceanogr. Mar.
Biol. Annu. Rev. 30:337-441.
Coates, M. E. and B. E. Gustafsson. 1984. The germfree animal in biomedical research. Laboratory
Animals, Ltd., London.
Cohen, M. L. 1992. Epidemiology of drug resistance:
Implications for a post-antimicrobial era. Science
257:1050-1055.
Cossart, P., P. Boquet, S. Normark, and R. Rappuoli.
1996. Cellular microbiology emerging. Science
271:315-316.
Courvalin, P., S. Goussard, and C. Grillotcourvalin.
1995. Gene transfer from bacteria to mammalian
cells. Comptes Rendu de L'Academie des Sci-
ences Serie Ill-Sciences de la Vie-Life Sciences
318:1207-1212.
Dehority, B. A. and C. G. Orpin. 1988. Development
of, and natural fluctuations in, rumen microbial
populations. In P. N. Hobson (ed.), The rumen microbial ecosystem, pp. 151—184. Elsevier Applied
Science, London.
Denarie, J., F Debelle, and J.-C. Prome. 1996. Rhizobium lipo-chitioligosaccharide nodulation factors:
Signaling molecules mediating recognition and
morphogenesis. Annu. Rev. Biochem. 65:503535.
Doino, J. A. and M. J. McFall-Ngai. 1995. A transient
exposure to symbiosis-competent bacteria induces
light organ morphogenesis in the host squid. Biol.
Bull. 189:347-355.
Douglas, A. E. 1989. Mycetocyte symbiosis in insects.
Biol. Rev. 64:409-434.
Douglas, A. E. 1994. Symbiotic interactions. Oxford
Science Pubs., Oxford.
Drasar, B. S. 1974. Human intestinal flora. Academic
Press, London.
Duncan, H. E. and S. C. Edberg. 1995. Host-microbe
interaction in the gastrointestinal tract. Crit. Rev.
Microbiol. 21:85-100.
Dyer, B. D. 1989. Symbiosis and organismal boundaries. Amer. Zool. 29:1085-1093.
Farrand, S. K. 1993. Conjugal transfer of Agrobacterium plasmids in bacterial conjugation. In D. B.
Clewell (ed.) New York-Plenum Press pp. 255285.
Fitt, W. K., S. L. Coon, M. Walch, R. M. Weiner, R.
Colwell, and D. B. Bonar. 1990. Settlement behavior and metamorphosis of oyster larvae (Crassostrea gigas) in response to bacterial supernatants. Mar. Biol. 106:389-394.
Foo, M. C. and A. Lee. 1974. Antigenic cross-reaction
between mouse intestine and a member of the autochthonous microflora. Infect. Immun. 9:10661069.
Fuqua, C, S. C. Winans, and E. P. Greenberg. 1994.
Quorum sensing in bacteria: The LuxR/LuxI family of cell density-responsive transcriptional regulators. J. Bacteriol. 176:269-275.
Fuqua, C , S. C. Winans, and E. P. Greenberg. 1996.
Census and consensus in bacterial ecosystems:
The LuxR-LuxI family of quorum-sensing transcriptional regulators. Annu. Rev. Microbiol. 50:
727-751.
Garrett, L. 1994. The coming plague. Penguin Books,
New York.
Gordon, H. A. and L. Pesti. 1971. The gnotobiotic animal as a tool in the study of host microbial relationships. Bacteriol. Rev. 35:390-429.
Gottschalk, G. 1986. Bacterial metabolism. SpringerVerlag, New York.
Grassle, J. F. 1986. The ecology of deep sea hydrothermal vent communities. Adv. Mar. Biol. 23:
301-362.
Gussin, G. N., C. W. Ronson, and F. M. Ausubel. 1986.
Regulation of nitrogen fixation genes. Annu. Rev.
Genet. 20:567-591.
Hadfield, M. G., C. C. Unabia, C. M. Smith, and T.
M. Michael. 1994. Settlement preferences of a
DEVELOPMENT OF COOPERATIVE SYMBIOSES
ubiquitous fouler Hydroides elegans. In M. E
Thompson, R. Nagahushanan, R. Sarojini, and M.
Fingerman (eds.), Recent developments in biofouling control, pp. 65—74. Oxford University
Press, Oxford, and IBH Publ. Co., New Delhi.
Haygood, M. G. 1993. Light organ symbioses in fishes. Crit. Rev. Microbiol. 19:191-216.
Heneghan, J. B. (ed.) 1973. Germfree research. Biological effect of gnotobiotic environments. Academic Press, New York.
Hirsch, A. M. 1992. Tansley review no. 40. Developmental biology of legume nodulation. New Phytol. 122:211-237.
Hofmann, D. K. and U. Brand. 1987. Induction of
metamorphosis in the symbiotic scyphozoan Cassiopea andromeda: role of marine bacteria and
biochemicals. Symbiosis 4:99-116.
Hooykaas, P. 1994. The virulence system of Agrobacterium tumifaciens. Annu. Rev. Phytopathol. 32:
157-179.
Hugenholtz, P. and N. R. Pace. 1996. Identifying microbial diversity in the natural environment—a
molecular phylogenetic approach. Trends Biotechnol. 14:190-197.
Hungate, R. E. 1966. The rumen and its microbes. Academic Press, New York.
Iglewski, B. and V. Clarke, (eds.) 1990. Molecular basis of bacterial pathogenesis, vol. XI. Academic
Press, New York.
Kagnoff, M. E and H. Kiyono. (eds.) 1996. Essentials
of mucosal immunology. Academic Press, New
York.
Lee, A. 1985. Neglected niches. The microbial ecology
of the gastrointestinal tract. In K. C. Marshall
(ed.), Advances in microbial ecology, Vol. 8, pp.
115-162. Plenum Press, New York.
Lee, K.-H. and E. G. Ruby. 1994. Effect of the squid
host on the abundance and distribution of symbiotic Vibrio fischeri in nature. Appl. Environ. Microbiol. 60:1565-1571.
Levy, S. B. 1992. The antibiotic paradox. Plenum
Press, New York.
Long, S. R. 1989. Rhizobium-iegume nodulation: Life
together in the underground. Cell 56:203-214.
Losick, R. and D. Kaiser. 1997. Why and how bacteria
communicate. Sci. Amer. 276:68—73.
Lycke, N. and W. Strober. 1989. Cholera toxin promotes B cell isotype differentiation. J. Immunol.
142:3781-3787.
Madigan, M. T, J. M. Martinko, and J. Parker. 1997.
Brock biology of microorganisms, 8th edition.
Prentice-Hall Press, Englewood Cliffs, New Jersey.
Maki, J. S., D. Rittschof, M.-O. Samuelsson, U. Szewzyk, A. B. Yule, S. Kjelleberg, J. D. Costlow, and
R. Mitchell. 1990. Effect of marine bacteria and
their exopolymers on the attachment of barnacle
cypris larvae. Bull. Mar. Sci. 46:499-511.
Margulis, L. 1970. Origin of eucaryotic cells. Yale
University Press, New Haven, Connecticut.
Margulis, L. and R. Fester, (eds.) 1991. Symbiosis as
a source of evolutionary innovation. MIT Press,
Cambridge.
607
Matzinger, P. 1994. Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12:991-1045.
Matzinger, P. and E. J. Fuchs. 1996. Beyond "self"
and "non-self": Immunity is a conversation, not
a war. J. NIH Res. 8:35-39.
McFall-Ngai, M. J. 1990. Crypsis in the pelagic environment. Amer. Zool. 30:175-188.
McFall-Ngai, M. J. 1994. Animal-bacterial interactions
in the early life history of marine invertebrates:
The Euprymna scolopeslVibrio fischeri symbiosis.
Amer. Zool. 34:554-561.
McFall-Ngai, M. J. and M. K. Montgomery. 1990. The
anatomy and morphology of the adult bacterial
light organ of Euprymna scolopes Berry (Cephalopoda:Sepiolidae). Biol. Bull. 179:332-339.
McFall-Ngai, M. J. and E. G. Ruby. 1991. Symbiotic
recognition and subsequent morphogenesis as early events in an animal-bacterial mutualism. Science 254:1491-1494.
McFall-Ngai, M. J. and W. W. Toller. 1991. Frontiers
in the study of the biochemistry and molecular
biology of vision and luminescence in fishes. In
P. W. Hochachka and T. P. Mommsen (eds.), Biochemistry and molecular biology of fishes, Vol. I,
Phylogenetic and biochemical perspectives, pp.
77-108. Elsevier, New York.
McFarland, W. N., F H. Pough, T. J. Cade, and J. B.
Heiser. 1979. Vertebrate Life. Macmillian Publ.
Co., New York.
McKeown, T. 1988. The origins of human disease. Oxford Press, Oxford.
McNeill, W. H. 1977. Plagues and peoples. Doubleday, New York.
Mel, S. F. and J. J. Mekalanos. 1996. Modulation of
horizontal gene transfer in pathogenic bacteria by
in vivo signals. Cell 87:795-798.
Montgomery, M. K. and M. J. McFall-Ngai. 1992. The
muscle-derived lens of a squid bioluminescent organ is biochemically convergent with the ocular
lens. J. Biol. Chem. 267:20999-21003.
Montgomery, M. K. and M. J. McFall-Ngai. 1993.
Embryonic development of the light organ of the
sepiolid squid Euprymna scolopes Berry. Biol.
Bull. 184:296-308.
Montgomery, M. K. and M. J. McFall-Ngai. 1994.
Bacterial symbionts induce host organ morphogenesis during early postembryonic development
of the squid Euprymna scolopes. Development
120:1719-1729.
Montgomery, M. K. and M. J. McFall-Ngai. 1995. The
inductive role of bacterial symbionts in the morphogenesis of a squid light organ. Amer. Zool. 35:
372-380.
Morishita, Y. and T. Mitsuoka. 1973. Antibody responses in germ-free chickens to bacteria isolated
from various sources. Jpn. J. Microbiol. 17:181 —
187.
Moynihan, M. 1983. Notes on the behavior of Euprymna scolopes (Cephalopoda:Sepiolidae). Behavior 85:25-41.
Nap, J.-P. and T. Bisseling. 1990. Developmental biology of a plant-prokaryote symbiosis: The legume root nodule. Science 250:948-954.
Nealson, K. H., T. Platt, and J. W. Hastings. 1970.
608
MARGARET J. MCFALL-NGAI
Cellular control of the synthesis and activity of
the bacterial luminescent system. J. Bacteriol.
104:313-322.
Neu, H. C. 1992. The crisis in antibiotic resistance.
Science 257:1064-1073.
Newell, N. D. 1972. The evolution of coral reefs. Sci.
Amer. 226:54-65.
Passador, L., J. M. Cook, M. J. Gambello, L. Rust, and
B. H. Iglewski. 1993. Expression of Pseudomonas
aeruginosa virulence genes requires cell-to-cell
communication. Science 260:1127-1130.
Raineri, D. M., M. I. Boulton, J. W. Davies, and E. W.
Nester. 1993. VirA, the plant-signal receptor, is
responsible for the Ti plasmid specific transfer of
DNA to maize by Agrobacterium. Proc. Natl.
Acad. Sci. U.S.A. 90:3549-3553.
Reich, K. A. and G. K. Schoolnik. 1994. The light
organ symbiont Vibrio fischeri possesses a homolog of the Vibrio cholerae transmembrane transcriptional activator ToxR. J. Bacteriol. 176:30853088.
Reich, K. A. and G. K. Schoolnik. 1996. Halovibrin,
secreted from the light organ symbiont Vibrio fischeri, is a member of a new class of ADP-ribosyltransferases. J. Bacteriol. 178:209-215.
Reich, K. A., E. G. Ruby, and G. K. Schoolnik. 1997.
Distribution of halovibrins and ADP-ribosyltransferase activity in luminous bacteria from different
light organ symbioses. Abst. Gen. Meet. Am. Soc.
Microbiol. 97:58.
Ruby, E. G. 1996. Lessons from a cooperative bacterial-animal association: The Vibrio fischeri-Euprymna scolopes light organ symbiosis. Annu.
Rev. Microbiol. 50:591-624.
Ruby, E. G. and L. M. Asato. 1993. Growth and flagellation of Vibrio fischeri during initiation of the
sepiolid squid light organ symbiosis. Arch. Microbiol. 159:160-167.
Saffo, M. B. 1992. Invertebrates in endosymbiotic associations. Amer. Zool. 32:557-565.
Salyers, A. A. and D. D. Whitt. 1994. Bacterial pathogenesis. A molecular approach. ASM Press,
Washington.
Savage, D. 1977. Microbial ecology of the gastrointestinal tract. Annu. Rev. Microbiol. 31:107-133.
Savage, D. 1986. Gastrointestinal microflora in mammalian nutrition. Annu. Rev. Nutr. 6:155-178.
Schwemmler, W. and G. Gassner. (eds.) 1989. Insect
endocytobiosis: Morphology, physiology, genetics, evolution. CRC Press, Boca Raton, Florida.
Sheltema, R. S. 1974. Biological interactions deter-
mining larval settlement of marine invertebrates.
Thalassia Jugosl. 10:263-296.
Small, A. L. and M. J. McFall-Ngai. 1993. Changes in
the oxygen environment of a symbiotic squid light
organ in response to infection by its luminous bacterial symbionts. Amer. Zool. 33:61 A.
Smith, D. C. 1979. From extracellular to intracellular:
The establishment of a symbiosis. Proc. Roy. Soc.
Lond. B 204:115-130.
Smith, M. W., D.-F. Feng, and R. F. Doolittle. 1992.
Evolution by acquisition: The case for horizontal
gene transfers. Trends Biochem. Sci. 17:489—493.
Spangler, B. D. 1992. Structure and function of cholera
toxin and the related Escherichia coli heat-labile
enterotoxin. Microbiol. Rev. 56:622-647.
Stahl, B. J. 1985. Vertebrate history: Problems in evolution. Dover Publ., Inc., New York.
Stoddard, D. R. 1969. Ecology and morphology of coral reefs. Biol. Rev. 44:433-498.
Tomarev, S. I., R. D. Zinovieva, V. M. Weis, A. B.
Chepelinsky, J. Piatigorsky, and M. J. McFallNgai. 1993. Abundant raRNAs in the squid light
organ encode proteins with a high similarity to
mammalian peroxidases. Gene 132:219—226.
Tunnicliffe, V. 1988. Biogeography and evolution of
hydrothermal vent fauna in the eastern Pacific
Ocean. Proc. Roy. Soc. B 233:347-366.
van Rhijn, P. and J. Vanderleyden. 1995. The Rhizobium-p\ant symbiosis. Microbiol. Rev. 59:124142.
Waldor, M. K. and J. J. Mekalanos. 1996. Lysogenic
conversion by a filamentous phage encoding cholera toxin. Science 272:1910-1914.
Weis, V. M., M. K. Montgomery, and M. J. McFallNgai. 1993. Enhanced production of ALDH-like
protein in the bacterial light organ of the sepiolid
squid Euprymna scolopes. Biol. Bull. 184:309—
321.
Weis, V. M., A. L. Small, and M. J. McFall-Ngai.
1996. A peroxidase related to the mammalian antimicrobial protein myeloperoxidase in the Euprymna-Vibrio symbiosis. Proc. Natl. Acad. Sci.
U.S.A. 93:13683-13688.
Woolverton, C. J., L. C. Holt, D. Mitchell, and R. B.
Sartor. 1992. Identification and characterization of
rat intestinal lamina propria cells—consequences
of microbial colonization. Vet. Immunol. Immunopathol. 34:127-138.
Wostmann, B. S., J. R. Pleasants, M. Pollard, B. A.
Teah, and M. Wagner. 1984. Germfree research.
Microflora control and its application to the biomedical sciences. Alan R. Liss, Inc., New York.
Corresponding Editor: Gregory A. Wray