Santa Rosalia revisited: Why are there so many species of bacteria? Antonie Van Leeuwenhoek 73:25-33
Abstract
The diversity of bacteria in the world is very poorly known. Usually less than one percent of the bacteria from natural communities can be grown in the laboratory. This has caused us to underestimate bacterial diversity and biased our view of bacterial communities. The tools are now available to estimate the number of bacterial species in a community and to estimate the difference between communities. Using what data are available, I have estimated that thirty grams of forest soil contains over half a million species. The species difference between related communities suggests that the number of species of bacteria may be more than a thousand million. I suppose that the explanation for such a large number of bacterial species is simply that speciation in bacteria is easy and extinction difficult, giving a rate of speciation higher than the rate of extinction, leading to an ever increasing number of species over time. The idea that speciation is easy is justified from the results of recent experimental work in bacterial evolution.
3 Figures
Antonie van Leeuwenhoek 73: 25–33, 1998. 25
c
1998 Kluwer Academic Publishers. Printed in the Netherlands.
Santa Rosalia revisited: Why are there so many species of bacteria?
Daniel E. Dykhuizen
Department of Ecology and Evolution, State University of New York at Stony Brook, Stony Brook, New York,
11794-5245,U.S.A.
Received 24 January 1997; accepted 17 October 1997
Key words: bacterial species diversity, DNA hybridization, experimental evolution, species definitions, speciation
Abstract
Thediversityofbacteriainthe worldisverypoorlyknown.Usuallyless thanone percentof thebacteriafromnatural
communities can be grown in the laboratory. This has caused us to underestimate bacterial diversity and biased
our view of bacterial communities. The tools are now available to estimate the number of bacterial species in a
communityand to estimate the differencebetween communities. Using whatdataare available,I haveestimatedthat
thirty grams of forest soil contains over half a million species. The species differencebetween related communities
suggeststhat the number of species of bacteria may be morethan a thousandmillion. I supposethat the explanation
for such a large number of bacterial species is simply that speciation in bacteria is easy and extinction difficult,
giving a rate of speciation higher than the rate of extinction, leading to an ever increasing number of species over
time. The idea that speciation is easy is justified from the results of recentexperimentalwork in bacterial evolution.
Introduction
One can imagine the American ecologist, G. Evelyn
Hutchinson, sitting next to a pool just outside the
church of Santa Rosalia on a hill overlooking the city
of Palermo. In this pool teem vast numbers of two
different species of water-bugs of the genus Corixa.
One can imagine him asking the questions: Why are
there just two species in this pool? Why not a single
species? Why are there such an enormous number of
animal species? Why are a majority of these insects?
In his presidential address to the American Soci-
ety of Naturalists on December 30, 1958, Hutchinson
triedtoanswersomeofthesequestionsintermsof food
chains and niche requirements. One of his conclusions
is that there will be many more species of small organ-
isms than of large ones. Given that bacteria are much
smaller than insects, howmany million species of bac-
teria might there be? Surely many more than the three
to four thousand well described species. In this paper
I would like to provide an estimate of the number of
species of bacteria there might be in the world and
provide reasons why there are so many.
Estimation of number of bacterial species in the
world
Species definitions
I will use a definition of species which in this context
is both useful and conservative. The formal definition
is as follows: ‘The phylogeneticdefinition of a species
generally would include strains with approximately
70% or greater DNA-DNA relatedness and with 5 C
or less t . Both units must be considered.’ (Wayne
et al 1987). Thus two strains are different species if
less than 70% of their DNA will reassociate and the
decrease in melting temperature of the DNA that does
reassociate is morethan 5 C( t 5 C). This defi-
nition is conservativebecause species defined this way
are likely to be separate species by any criteria, while
groups that are different species by other criteria can
be defined as within the same species by this criterion.
For example, NeisseriagonorrhoeaeandN. menin-
gitidis are considered separate species eventhough the
percentage of DNA that reannealed was 93% (Hoke
and Vedros, 1982), well above the 70% criterion used
to distinguish species. By an ecological criterion, i.e.
26
niche differentiation, these are separate species – one
livesintheurogenitaltractofmanand the otherlivesin
the upper respiratory tract. By a genetic criterion, they
are separate species – a dendogram of the glnA gene
sequences from eleven N. gonorrhoeae strains and
eight N. meningitidis strains separates the strains from
the two species onto two different branches (Spratt et
al., 1995). These species are distinct by both an eco-
logical criterion, they occupy different niches, and a
genetic criteria, there is little or no genetic exchange
between them, but are considered part of the same
speciesbythecriterionofDNAhybridization.ByDNA
hybridization, not only would N. gonorrhoeae and N.
meningitidis be considereda single species, but N. lac-
tamica and N. polysaccharea would also be included
inthisspecies (Hokeand Vedros1982, MaynardSmith
1995).
Itis clearthatspeciesas definedbyDNAhybridiza-
tion are certainly separate species given other defini-
tions of species. A decrease of the melt temperature of
5 C impliesthattheDNAhas a sequencedifferenceof
more than 7–8% (Caccone et al., 1988). This amount
of divergence indicates that genetic transfer between
the species is rare to non-existent. The more divergent
two strains are, the larger the effect any recombination
has on homogenizingthem (Cohan 1994, 1995). Also,
the rate of recombination between homologous DNA
decreasesas the divergenceincreases (Shen andHuang
1986) such that as the divergence increases, any resid-
ual recombination will become less and less likely.
This sequence difference can be translated into
a time difference. Ochman and Wilson (1988) have
estimated that synonymous sites change at a rate of
1%/million years and nonsynonymous sites about 20
fold slower. Since most of the homologous DNA in
bacterial genomes is coding sequence, one can esti-
mate that the rate of change is about 1%/2.7 million
years. This implies that two lineages that have a T
of 5 C will have been separated for 20 million years.
If these two species occupy the same niche and conse-
quently not different species by the ecological criteri-
on, one would have eliminated the other over a period
of 20 million years by competative exclusion. Thus
we do not expect two species to be different by DNA
homology,butnot to be differentspecies by anecolog-
ical or genetic criteria. Consequently, defining species
in terms of DNA homology is conservative, i.e. it will
underestimate the number of species.
Defining species in terms of DNA homology can
be used to estimate the number of species within
an ecosystem without characterizing isolated strains.
Only about 0.1% to 1% of the cells from most natural
ecosystems can be grown in the laboratory and these
represent a biased sample of the species in the ecosys-
tem. Thus, any estimation of the number of species of
bacteria in any ecosystem must be indirect.
Estimation of number of species in a community
Reassociation kinetics can be used to estimate species
diversity (Torsvik et al., 1990a). The rate of the reas-
sociation of single stranded DNA with its homologue
depends upon the number of differenttypes of DNA in
a solution or on the complexity of the DNA. The more
complex the DNA, the longer it will take homologous
strands topair. Thereassociationkineticsaremeasured
in terms of the concentrationof DNA in moles per liter
(C
0
)timesthetimein seconds(t).Thisis theCotvalue.
If the concentration of DNA is held constant, then the
number of molecules of each unique sequence in the
genomedecreasesasgenomesize increases. Forexam-
ple, if the concentration of DNA is 12 pg, a solution
will contain 4000 copies of a genome of 0.003 pg but
only 4 copies of a genome of 3 pg. In this example, it
will take on average 1000 times longer for the DNA in
the large genome to find its homologue,since there are
manyfewercopiesofthelargergenome.The Cotvalue
when half the DNA is associated gives an estimate of
the genome size (Figure 1). If we think of the bacteria
in a natural community as a single species of bacte-
ria, how large would its ‘genome’ be? The number of
species in the community can be estimated by divid-
ing this ‘genome’ by the average size of a bacterial
genome.
Torsvik, Goksøyr and Daae (1990a) isolated 30
grams of top soil from a beech forest north of Bergen,
Norway. This soil contained 1.5 10
10
bacteria/gm
of dry soil by microscope observation, but the colony
count was only 4.3 10
7
. As usual, the colony count
was less than 1% of the observablecells. The bacterial
fraction was isolated from the soil and DNA extract-
ed from the bacteria. The DNA was sheared, single
stranded DNA removed, and the DNA melted. The
reassociation was done at 25 ClessthanT ,where
T is temperature at which half of the native DNA
melted. As seen in Figure 1, the reassociation does
not follow the same kinetics as E. coli or calf thymus.
These two curves in Figure 1 are the same; one is just
movedto the right of theother. The differencebetween
the reassociation of theDNAfrom the soilbacteria and
from the calf thymus shows that the various species in
27
Figure 1. The reassociation of soil bacterium (X), Calf thymus ( )
and E. coli B( ) DNA. A curve was drawn through the calf thymus
DNA and then shifted to the right such that the curve passes through
the soil bacterium when 50% is reassociated. This figure is derived
from Figure 3 of Torsvik et al., (1990a).
the soil are at different frequencies. The reassociation
of the bacterial DNA is about ten times slower than
calf thymus DNA where the two curves cross the dot-
ted line inFigure 1. The estimated heterogeneityof the
bacterial DNA at this point is 2.7 10
10
bp (Torsvik
et al., 1990a). If the average size of the genome of
soil bacteria is 6.8 10
6
bp (Torsvik et al., 1990b),
then there are about 4,000completely differentbacter-
ial genomes in the sample. Figure 1 shows that the first
hybridization is taking place at about the same time
as the calf thymus. The complexity of calf thymus is
equivalent to about 500 bacterial genomes. If this first
reassociation is due to the most common species in the
soil, then less than one percent of the individuals or
3 10
7
bacteria/gram of soil belong to this species.
This is a very different picture of diversity than the
one obtained from 16s RNA cloning and sequencing.
Usually at least one species is isolated more than once
within a sample of 100 or less.
However, the definition of species does not require
that the genomes be completely independent and non-
hybridizing – it requires that only 30% of the DNA
does not hybridize. The melt curve of the DNA that
did hybridize shows that much of hybridizing DNA
is from separate species. The T is about 18 C
(Torsvik et al., 1990a). The curve suggests that all but
about10% of the reassociated DNAmeltedby thetime
the temperature had risen to 5 C less than the T of
the native DNA. Thus 90% of the reassociating DNA
belongsto differentspecies and we haveunderestimat-
ed the number of species by nine fold. This implies
that the number of species is 36,000.
Fortunately, there is another way to estimate this
correction. Two hundred and six strains were isolated
from the same soil (Torsvik et al., 1990a). While these
were not checked by the species definition above, the
phenotypic diversity of the population was high. The
rehybridizationofthesestrainswhentheywereinequal
frequencygave an estimate of 20.6 separate species. If
we assume that these are all different species, then
the reassociation estimate is a ten fold underestimate,
which is close to the nine fold estimate as determined
above. I feel the assumption to treat these 206 strains
as separate species is valid, even though many of the
strains are phenotypically identical by the tests used
(Torsvik et al., 1990b). Jimenez (1990) has shown that
strains from subsurfacesoil that are phenotypicallythe
same by the tests used are often different species by
percentDNA hybridizationand G+C content. Thus we
will estimate 40,000 species in the sample. Clearly,
this conversion factor will have to be estimated with
more precision if this method is to be used regularly.
In natural ecosystems, species are not equally
frequent. Some are common with many individuals
present and other are rare with only a few individuals
present. The reassociation kinetics of the DNA from
soil bacteria do not follow the typical second order
reaction kinetics. Rather the reaction rate slows faster
than expected as the reassociated progresses. This is
exactlythe patternexpectedifthespeciesarenotequal-
ly frequent. The estimate of the number of species
is made from the 50% of the individuals belonging
to the most common species, with the other 50% of
the individuals belongingto rare species. Thus 40,000
would be a considerable underestimate of the number
of species present. Animal and plant ecologist have
enumerated the number of each species in a sample of
individualsfrom a community. If we dividethe sample
into equal numbers of individuals with the individuals
from common species in one part and the individuals
from rare species in the other, how many rare species
are there for every common species? Table 1 shows
this value for various studies. This value is not con-
stant but covers a wide range. With the high values,
one finds one or two species that are excessively com-
mon. About half of the individuals belong to these
very common species. With the low values one finds
fewer rare species than expected. Presumably this is
because species of large animals with few individuals
are likely to go extinct. With the soil bacteria con-
sidered here, we clearly do not have an excessively
28
Table 1. The ratio of rare to common species within diverse communities
Taxonomic group No. of species No. of individuals Ratio (rare/common) Reference
Birds 50 365 6.8 Preston 1948
Birds 80 14,353 12 Saunders, 1936
Insects and mites 52 822 25 Pielou and Matthewman 1966
Moths 240 15,609 30 Fisher et al 1943
Moths 277 87,110 150 Preston 1948
common species (shown above). On the other hand, as
explained below, we would not expect high extinction
rates for bacteria even when rare. Thus we will take 25
as the factor we will use since it is the modal estimate.
Clearly, this correction could be much too large if the
numberof individuals per species are more evenlydis-
tributedthan with animals. Or, this correction could be
much too small if there are many rare species within
eachcommunity.Thiscorrectionfactoris unsatisfacto-
ry because one is dividinga large number (the number
of rare species) by a small number (the number of
common species). I checked this ratio using data from
Patrick (1968). In 1965 she set up eight semi-natural
populations of diatoms and in 1966 an additional four
populations. In 1966, the ratios ranged from 15 to 25
with an average of 19. Thus the ratio is fairly constant
for the same group of organisms in the same ecosys-
tem. However, in 1967, the ratios ranged from 66 to
76 with an average of 69. The difference between the
yearsisthat in1966therewasonecommonspeciesand
in 1967 there were two common species. Clearly, this
correction factor is unsatisfactory, and a more robust
one will have to be devised.
But if we accept for the time being a correction
factor of 25, we estimate that there are about 20,000
common species and 500,000 rare species in a small
quantity of soil or about a half million species. I
contend that even this is an underestimate. Animal
and plant ecologists deal with samples of hundreds or
thousands of individuals, not numbers in the range of
10,000,000,000. The larger the population, the larger
the chance of rare species being present, the more rare
species, and the larger the number of rare species for
each common species.
Estimation of the number of bacteria communities
Bacteria are found everywhere. There are ten times
more individual bacteria on you than there are cells in
your body. There are communities of bacteria at 600
meters down into the earth. There are bacteria found
in the rocks in Antarctica where they are frozen except
for less than one day a year. There are bacteria that
live in hot springs. Bacteria are everywhere, perhaps
evenon Mars. Howmany differentcommunitiesmight
there be?
The measurement of differences between commu-
nities is much easier than attempting to measure the
number of species within a community. For example,
using rRNA sequence diversity, one can recover sim-
ilar types from the same or similar habitats, thereby
defining different ecosystems (Stahl 1995). For exam-
ple, the two types of Fibrobacter succinogenes isolat-
ed from the intestine of a horse were clearly different
from the three types isolated from the rumen of a cow
(Lin and Stahl, 1995). While these five types are con-
sidered subspecies because of no clear physiological
differentiation, genetically they are different enough
to be classified as separate species, perhaps separate
genera (Lin and Stahl, 1995).
DNA-DNA hybridization has been used to define
differentbacterialcommunities.Thismethodmeasures
both species composition and relative diversity. The
DNA from one community is bound to a filter and
hybridized with radioactivity labeled DNA from the
same and different communities. The similarity index
is then the comparison of the amount of bound labeled
DNA from the different community normalized by the
bound labeled DNA from the same community. The
reciprocal cross-hybridization(DNA from community
one as a probe hybridizing to DNA from communi-
ty two compared to DNA from community two as a
probe hybridizing to DNA from community one) will
give the same results only when the two communi-
ties are equally diverse. Otherwise, the more diverse
community used as probe will give a higher similarity
value.
Griffiths et al. (1996) tested four local soils. The
order of the diversity in the communities from most to
least diverse was sandy loam, sandy clay loam, loamy
sand, andclay loam. Therewasconsiderablesimilarity
29
or shared species between clay loam and sandy clay
loam, between sandy clay loam and sandy loam and
between sandy loam and loamy sand. But, there was
little three way overlap. For example, those species
in sandy clay loam that are also in clay loam are not
found in sandy loam. These data suggest that the four
soil types represented at least two or three different
non-overlappingcommunities.
Lee and Fuhrman (1990) showed that bacteria iso-
latedfromacoralreeflagoon,fromLongIslandSound,
from the Caribbean Sea, and from the Sargasso Sea
are all different communities. There was overlap only
between the communities from the Caribbean and Sar-
gasso Sea.
Another approach is to isolate a group of bacte-
ria from diverse locations and then to compare the
divergence of these strains using dot blot hybridiza-
tion. Holm et al., (1996) isolated 1,199 pure cultures
ofHyphomicrobiumfromasewagetreatmentplantand
its adjacent receiving lake. These strains were divided
into six morphologically different types. Of the 755
strains which could be used for dot hybridization (the
436isolatesfrommorphologicaltype6grewtooslowly
tobeused), 671 isolateswereassignedto 30hybridiza-
tion groups. The dot blot hybridization were done at a
temperature 10 C lower than the melting temperature
of the DNA. Thus, strains which belong to different
hybridizationgroups belong to differentspecies by the
DNA homology definition of species. Only one of the
hybridization corresponded to one of the 14 known
species tested. Of the 30 groups, 12 were found in the
sewage plant, 14 in the lake and only 4 groups found
in both. Two of these four groups were found most-
ly in the sewage plant with only a single isolate from
the lake. These single isolates probablyrepresent tran-
sients from the effluent from the sewage plant. Thus
the overlap between these communities represent only
two out of thirty species.
We can conclude there are many different com-
munities of bacteria. Clearly, much more work will
have to be done before we know how many. Are the
bacterial communities in the soil of a beech forest dif-
ferentfromthecommunitiesinanoak forest ora maple
forest? How closely related are the bacterial commu-
nities as one descends into the subsurface soil? How
closely related are communities in different types of
desert soils? How closely related are the communi-
ties in beech forests in Norway and Michigan? Then
we have all the communities in fresh water and in the
ocean. We have communities on every living person.
How close are the communities on a person and a
robin? Clearly, I do not have an answer to these ques-
tions, but I will take a guess that there are more than
a 2,000 communities that are different from each oth-
er and as complex as the Norwegian forest soil. This
would imply there are more than a thousand million
species of bacteria and this is probably a considerable
underestimate. I would guess that the number of bac-
terial species might be as high as a trillion given all the
underestimations that I have done.
Why are there so many species of bacteria?
One simple answer is that newspecies of bacteria form
faster than old ones disappear by extinction. Even in
metazoans, the rate of speciation normally appears to
be faster than the rate of extinction, since, between
periods of mass extinction, there seems to be a gradual
accumulation of metazoan diversity (Sepkoski 1984).
Low extinction rates
Iwouldexpectthe extinctionrates of bacteriawouldbe
lower than metazoans. Bacteria don’t starve to death
as do you or I. Bacteria don’t die of old age as do
you or I. Bacteria don’t need sex to reproduce as do
you or I. Bacteria are much smaller than you or I and
so there are a lot more of them than you or I. When
conditions are bad, bacteria can sit and do nothing
as they do in stab tubes. The range of environments
that bacteria can live in far exceeds the range of envi-
ronments metazoans can live in. Bacteria are able to
stand harsher conditions. All this might lead to a lower
extinction rate over geological time. If bacteria avoid
mass extinctions,they could haveundergonecontinual
expansion in the number of species for 3 billion years.
However, the large number of species is likely to be
due to an unusually high speciation rate rather than an
unusually low extinction rate.
High Speciation Rates
Laboratorystudiesin experimentalevolutionhavepro-
vided evidence that speciation is easy and likely in
bacteria. I am changing the definition of species and
speciation away from a DNA based definition to an
ecological definition. Two strains are different species
if they occupy different niches and the same species
if they occupy the same niche. This definition implies
that different populations of the same species can eas-
ily replace each other, but that populations of differ-
30
Figure 2. Evolution of one species into three. The width of the uptake and excretion arrows shows the evolutionary changes in the derived
species. The dark circles at the bottom give the approximate proportion of each type in the mixed culture. This figure is derived from the data
of Helling et al. (1987) and Rosenzweig et al. (1994).
ent species can not: each species has an advantage on
certain parts of the resource base and consequently
remains extant for some extended period of time. If
this time is long enough, the species become geneti-
cally distinct enough that one can apply a DNA based
rule to define species.
Itis oftenassumedthatspeciesmultiply tofill avail-
able niches, so that there can be no more species than
there are available niches. By this assumption, nich-
es are given in the environment and then species fill
them. However, niches are not outthere in the environ-
ment waiting to be filled. Originally an environment
is undivided, but as species develop, they divide the
environment into separate niches. Species create nich-
es. Differentniches havedifferentamounts of resource
and are differentially stable. If a new niche occupied
by an incipient species is too small or too unstable, the
species will go extinct and the niche will disappear. Its
resources will be incorporated into other niches occu-
pied by other species. This will be clarified by the
discussion of the experiments discussed below.
Three niches from one; three species from one
Helling, Vargas and Adams (1987) did a simple exper-
iment. They grew a single strain of Escherichia coli
in a glucose-limited chemostat for a long time. The
chemostat provides a homogeneous constant environ-
ment with only one nutrient limiting. The cells are
competing only for glucose, and all other nutrients are
in excess. The cultures transferred periodically into
fresh chemostats to make sure that the glass walls of
the chemostatdid notprovideanother niche. Thus they
grew a single strain in a zero dimensional niche, the
least complicated environment anyone could imagine.
A zero dimensional niche can be thought of as a point
31
in environmental space. In this niche there is no envi-
ronmental variation, such as changing temperature, or
environmentalgradient as a pH gradient which can be
divided to form multiple niches.
Theyplannedtoinvestigatehowanisolated lineage
would evolve over time, but the experiment did not
work as planned. Instead of one strain evolving over
time, the one strain evolved into at least three strains
whichcoexist. The one niche was partitioned intothree
niches. As shown in Figure 2, just adding the bacteria
complicated the environment. E. coli, while metabo-
lizing glucose, excretes acetate and a little glycerol.
These compounds are later taken up again and used.
Overthe eight hundred generations in the chemostat, a
strain evolved which was a specialist on glucose. This
strain excreted a lot of acetate and some glycerol, but
could no longer take up and metabolize either of these
compounds.Twootherstrainsarose eachofwhich spe-
cializedon these compoundswhileretainingtheability
to utilize any glucose they could acquire (Rosenzweig
et al., 1994).Thus thebacteriahavecreated three nich-
es where beforeonly one existed and the three coexist-
ing strains, none of which can become extinct within
this system, can evolveinto separate species over time.
Onemight evenconsiderthemseparatespecies already
by the ecological definition of species.
Environmental partitioning
Bennett, Lenski and Mittler (1992) have shown that
bacteria can partition the temperature dimension of
the niche very finely. An ancestor E. coli culture was
adapted for 2,000 generations to 37 C (Lenski et al.,
1991). This culture was divided into six replicate lines
for each of three temperatures, 32, 37, 42 C. The
18 cultures were then adapted for another 2,000 gen-
erations. At each temperature, there was significant
adaptation to that temperature, but no adaptation or
loss of adaptation to another temperature (Bennett et
al., 1992). For example, the culture evolved at 32 C
was significantly fitter than the ancestor at 32 C, but
not significantly different in fitness from the ancestor
at either 37 or 42 C. Whenthe adaptationsweretested
across a range of temperatures, it was found that cul-
tures adapted to the three temperatures partitioned the
niche dimension of temperature into three parts (Ben-
nett and Lenski, 1993). Clearly, one could partition
E. coli into at least three species along a temperature
dimension.
Natural variation for niche partitioning
When naturally occurring samples of the lac oper-
on of E. coli are extracted from natural populations,
different operons have different fitness on various -
galactosides (Silva 1992, Dean 1995). For example,
the lac operon from ECOR16 (Ochman and Selander
1984) is fitter than the operon from K12 on lactose,
galactosyl-glycerol and methyl-galactoside while the
reverse is true on lactulose and galactosyl-arabinose.
When two sugars are mixed in a chemostat, the rela-
tivefitnessof twostrainscanbepredictedforanygiven
mixture by the equation:
1
2
1
0
0 0
1
0
0 0
(1)
where is the relative fitness on the mixture of sug-
ars,
1
and
2
are the growth rates of the two strains,
0
and
0
are the concentration of the two sugars in
the reservoir, and
1
and
1
are the relative fitness
on and respectively (Dykhuizen and Dean 1994).
Figure 3 shows an experimental conformation of this
equation. Silva (1992) has shown that this applies not
only to sugars that are used by two differentpathways,
but also by sugars that are used by the same pathway
(Figure 4). As shown in this figure, the lac operon
from ECOR16 has a selective advantage when the lac-
tose concentration is higher than about 60%, but the
other operon from K12 has the advantage when the
lactose concentration is below this level. At a mixture
of lactose and lactulose of about 60/40, one expects a
stable equilibrium. If the numbers of cells with the lac
operonfromECOR16drops,the relativeconcentration
of its preferred substrate, lactose, will increase in the
chemostat and there will be selection for ECOR16. If
the other strain becomes rare, the relative concentra-
tion of lactulose will increase and it will be selected
for. Thus the system is stable and, over time, these two
strains could evolve to specialize on their preferred
substrate and form separate species. Thus the genetic
variation is available in populations to allow special-
ization and speciation given that the environment is
suitable for partitioning.
In the laboratory, niche division and specialization
are easy and happenquickly. Sexuality or lateral trans-
fer of DNA between lineages and diploidy will slow
theprocessdescribedabove. Sincebacteriaare haploid
and generally have lower rates of gene exchange that
sexually reproducing metazoans, the speciation rate
is expected to be faster in bacteria. While it is usually
assumedthat the environmentismuchmore variablein
32
Figure3. Fitness on mixed sugars. The competition is between TD2
(lacI Z Y A ) and TD4 (lacI Z
mut
Y A ). The lacZ mutation
in TD4 has 1.1% of the activity of wild type. This figure is derived
from Figure 4 of Dykhuizen and Dean (1994).
Figure 4. Fitness on a mixture of two -galactosides. The compe-
tition is between two strains isogenic except for the lactose operon.
The two lactose operons are both wild type, having been isolated
from nature. This figure is derived from Figure 6.2 of Silva (1992).
nature than in the laboratory and niche division will be
much more difficult, these examples suggest that even
short periods of environmental stability could lead to
speciation. Speciation is easy.
Epilogue
In this paper I want to suggest that there are a tremen-
dous number of species of bacteria in the world. The
numbers estimated in this paper are clearly best guess-
es for now. But it is clear that as the techniques are
refined and much more work is done comparing the
species composition of various communities, much
better estimates of the number of bacterial species will
be made. If bacteria communities have structure like
metazoancommunities(Preston1962,Sugihara1980),
then a much more accurate estimate of the numbers
of species in a community can be derived from the
deviation of the Cot curve from the expected curve as
shown in Figure 1. Ifthe numberof species is anything
like I suggested, the cloning and sequencing 16s RNA
types from nature will never give an adequate picture
of the species diversity. This method will only provide
information on some of the common species of the
community.
While we might not want to define and name each
andeveryspeciesaswehavedonewithbirdsand mam-
mals, we need to have an idea of world-wide bacterial
diversity and good ways of assessing this diversity to
make sure that the microbial world does not become
unstable as man changes the environment. The meth-
ods are available and now need refinement and use.
Acknowledgements
I thank Amalia Karagouni for her hospitality and for
providing me with a reason to put my thoughts on
bacterial speciation into a coherentwhole. This is con-
tribution no. 1007 from Graduate Studies in Ecology
and Evolution, State University of New York at Stony
Brook.
References
Bennett AF, Lenski RE& Mittler JE(1992) Evolutionary adaptation
to temperature. I Fitness responses of Escherichia coli to changes
in its thermal environment. Evolution 46: 16–30
Caccone A, DeSalle R & Powell JR(1988) Calibration in the change
in thermal stability of DNA duplexes and degree of base pair
mismatch. J. Mol. Evol. 27: 212–216
Cohan FM (1994) The effects of rare but promiscuous genetic
exchange on evolutionary divergence in prokaryotes. Am. Nat.
143: 965–986
Cohan FM (1995) Does recombination constrain neutral divergence
among bacterial taxa? Evolution 49: 164–175
Dean AM (1995) A molecular investigation of genotype by environ-
ment interactions. Genetics 139: 19–33
Dykhuizen DE & Dean AM (1994) Predicted fitness changes along
and environmental gradient. Evol. Ecol. 8: 1–18
Fisher RA, Corbert AS & Williams CB (1943) The relation between
the number of individuals and the number of species in a random
sample of an animal population. J. Anim. Ecol. 12: 42–58
Griffiths BS, Ritz K & Glover LA (1996) Broad-scale approaches to
the determination of siol microbial community structure: Appli-
33
cation of the community DNA hybridization technique. Microb.
Ecol. 31: 269–280
Helling RB, Vargas C & Adams J (1987) Evolution of Escherichia
coli duringgrowthin a constant environment. Genetics 116: 349–
358
Hoke C & Vedros NA (1982) Taxonomy of the Neisseriae: Deoxyri-
bonucleic acid base composition, interspecific transformation,
and deoxyribonucleic acid hybridization. Int. J. Syst. Bacteriol.
32: 57–66
Hutchinson GE (1959) Homage to Santa Rosalia or whyare there so
many kinds of animals? Am. Nat. 93: 143–159
Jim
´
enez L (1990) Molecular analysis of deep-subsurface bacteria.
Appl. Environ. Microbiol. 56: 2108–2113
Lee S & Fuhrman JA (1990) DNA hybridization to compare species
composition ofnatural bacterioplankton assemblages. Appl.Env-
iron. Microbiol. 56: 739–746
Lenski RE & Bennett AF (1993) Evolutionary response of
Escherichia coli to thermal stress. Am. Nat. 142: S47–64
Lenski RE, Rose MR, Simpson SC & Tadler SC (1991) Long-term
experimental evolution in Escherichia coli. I. Adaptation and
divergence during 2,000 generations. Am. Nat. 138: 1315–1341
Lin C & Stahl DA (1995) Taxon-specific probes for the cellulolytic
genus Fibrobacter reveal abundant and novel equine-associated
populations. Appl. Environ. Microbiol. 61: 1348–1351
Maynard Smith J (1995) Do bacteria have population genetics? In:
Baumberg S, Young JPW, Wellington EMH and Saunders JR
(eds.) Population Genetics of Bacteria (pp1-12). Cambridge Uni-
versity Press, Cambridge
Ochman H & Wilson AC (1987) Evolution in bacteria: Evidence for
a universal substitution rate in cellular genomes. J. Mol. Evol. 26:
74–86
Ochman H & Selander RK (1984) Standard reference strains of
Escherichia coli from natural populations. J. Bacteriol. 157: 690–
693
Patrick R (1968) The structure of diatom communities in similar
ecological conditions. Am. Nat. 102: 173–183
Pielou EC & Matthewman WG (1966) The fauna of Fomes fomen -
tarius (Linnaeus ex Fries) Kieckx. growing on dead birch in
Gatineau Park, Quebec. Can. Ent. 98: 1308–1312
Preston FW(1948) The commonness, and rarity, of species. Ecology
29: 254–283
Preston FW (1962) The canonical distribution of commonness and
rarity. Ecology 43: 185–215, 410–432
Rosenzweig RF, Sharp RR, Treves DS & Adams J (1994) Microbial
evolution in a simple unstructured environment: Genetic differ-
entiation in Escherichia coli. Genetics 137: 903–917
Saunders AA (1938) Ecology of the birds of Quaker Run Valley,
Allegany State Park, New York. New York State Museum Hand-
book 16. Albany, N. Y.
Sepkoski JJ, Jr (1984) A kinetic model of Phanerozoic taxonomic
diversity III. Post-Paleozoic families and mass extinction. Paleo-
biology 10: 246–267
Shen P & Huang HV (1986) Homologous recombination in
Escherichia coli: Dependence on substrate length and homology.
Genetics 112: 441–457
Silva PJN (1992) Natural selection of the lac operon of Escherichia
coli. Ph.D. Thesis. State University of New York at Stony Brook.
153 pp
Spratt BG, Smith NH, Zhou J, O’Rourke M & Feil E (1995) The
population genetics of pathogenic Neisseria. In: Baumberg S,
Young JPW,Wellington EMH and Saunders JR (eds.) Population
Genetics of Bacteria (pp 143–160). Cambridge University Press,
Cambridge
StahlDA(1995)Applicationofphylogeneticallybasedhybridization
probes to microbial ecology. Molec. Ecol. 4: 535–542
Sugihara G (1980) Minimal community structure: An explanation
of species abundance patterns. Am. Nat. 116: 770–787
Torsvik V, Goksøyr J & Daae FL (1990a) High diversity of DNA in
soil bacteria. Appl. Environ. Microbiol. 56: 782–787
Torsvik V, Salte K, Sørheim R & Goksøyr J (1990b) Comparison of
phenotypic diversity and DNA heterogeneity in a population of
soil bacteria. Appl. Environ. Microbiol. 56: 776–781
Wayne LG, Brenner DJ, Colwell RR, Grimont PAD, Kandler O,
Krichevsky MI, Moore LH, Murry RGE, Stackebrant E, Starr
MP & Tr
¨
uper HG (1987) Report of the ad hoc committee on
reconciliation of approaches to bacterial systematics. Int. J. Syst.
Bacteriol. 37: 463–464
- CitationsCitations232
- ReferencesReferences51
- Biodiversity research on soil microorganisms is conducted throughout the world (Ramette and Tiedje 2007;Tedersoo et al. 2012). This interest is triggered by the pivotal contributions of microorganisms to ecosystem functioning (Torsvik and Øvreås 2002;van der Heijden et al. 2008), and the vast diversity of bacterial and fungal species (Curtis et al. 2002;Dykhuizen 1998;O'Brien et al. 2005). In depth analysis of microbial communities is realized by high-throughput sequencing generating millions of nucleic acid reads using next generation sequencing (NGS) platforms (Caporaso et al. 2012;Shokralla et al. 2012;Taberlet et al. 2012).
[Show abstract] [Hide abstract] ABSTRACT: Aims Soil sample preservation is a challenging aspect in molecular studies on soil microbial communities. The demands for specialized sample storage equipment, chemicals and standardized protocols for nucleic acid extraction often require sample processing in a home laboratory that can be continents apart from sampling sites. Standard sampling procedures, especially when dealing with RNA, comprise immediate snap freezing of soils in liquid nitrogen and storage at −80°C until further processing. For these instances, organizing a reliable cooling chain to transport hundreds of soil samples between continents is very costly, if possible at all. In this study we tested the effect of soil sample preservation by freeze-drying with subsequent short-term storage at 4°C or ambient temperatures compared to −80°C freezing by comparative barcoding analyses of soil microbial communities. Methods Two grassland soil samples were collected in Central Germany in the Biodiversity Exploratory Hainich-Dün. Samples were freeze-dried or stored at −80°C as controls. Freeze-dried samples were stored at 4°C or ambient temperature. Investigated storage times for both storage temperatures were 1 and 7 days. Total DNA and RNA were extracted and bacterial and arbuscular mycorrhizal (AM) fungal communities were analyzed by amplicon 454 pyrosequencing of the 16S (V4-V5 variable region) and 18S (NS31-AM1 fragment) of ribosomal RNA (rRNA) marker genes, respectively. Important Findings Bacterial communities were sufficiently well preserved at the rDNA and rRNA level although storage effects showed as slightly decreased alpha diversity indices for the prolonged storage of freeze-dried samples for 7 days. AM fungal communities could be studied without significant changes at the rDNA and rRNA level. Our results suggest that proper sampling design followed by immediate freeze-drying of soil samples enables short-term transportation of soil samples across continents.- Nevertheless, we have credible theoretical estimates (Curtis et al., 2002;Hughes et al., 2001;Whitman et al., 1998), which place the total count in the vicinity of 10 30 and the number of species at 10 9. Theoreticians estimate that more than half a million bacterial species are present in thirty grams 30 g of forest soil because speciation is easy and extinction is difficult for this group of organisms. This implies an implausibly infinite increase in the number of species over time, with no real impact of natural selection, except perhaps under narrowly defined local situations (Dykhuizen, 1998). Achieving precision in these estimates is neither possible nor useful for practical purposes, except for the fact that our knowledge of Earth's biosphere and its potential for sustainability is remarkably inadequate, and will not improve without understanding prokaryotic diversity and its myriad linkages to fundamental ecosystem functions (Ogunseitan, 2005).
[Show abstract] [Hide abstract] ABSTRACT: Bacteria and Archaea represent the two domains of prokaryotes – unicellular organisms that lack membrane-bound internal organelles such as the nucleus. Bacteria are considered the most abundant and diverse organisms on Earth. Investigations of bacterial diversity seek to understand the causes and outcomes of variability in phenotype, genotype, and ecological functions within the microbiome. Such variability is presumed to be the outcome of selection, evident in bacterial populations challenged with natural or anthropogenic environmental pressures. Such selections are exploited, for example, in public health and food preservation, where measurable reduction in diversity and increased fitness of selected varieties are beneficial.- Wartość T m wskazuje na temperaturę, przy której 50% DNA znajduje się w stanie denaturowanym[39,64,82]. Hybrydyzacja genomowego DNA jest podstawą do klasyfikacji mikroorganizmów do rangi gatunku[64,82]. Do jednego gatunku genomowego zaliczamy szczepy o wartości hybrydyzacji 70% i więcej oraz charakteryzujące się ΔT m mniejszą niż 5°C[20,82,94,97]. Metodą hybrydyzacji oceniono różnorodność drobnoustrojów w 1 gramie różnych gleb i oszacowano ją na poziomie od 500 tysięcy do nawet 10 milionów genomów organizmów prokariotycznych[33]. Obecnie stosowanych jest kilka technik hybrydyzacji DNA-DNA, m.in.
[Show abstract] [Hide abstract] ABSTRACT: Biodiversity and the identification of new important features of microorganisms is crucial for the development of biotechnology. The current knowledge about microbs in natural environments is limited, thus the analysis of the microbial diversity in nature is not an easy task. So far, only a small percentage of prokaryotic microorganisms has been identified. It is believed that the soil environment is one of the richest reservoirs of microorganisms, as approximately 2 000 to 18 000 prokaryotic genomes can be isolated from one gram of soil. In this publication the selected methods used to identify microorganisms are presented. The first molecular marker used in the genetic identification of soil microorganisms was the analysis of the G+C base content, sincemicroorganisms exhibit differences in the (G+C)/(A+T) relative factor. Another method used to identify bacteria is the nucleic acid hybridization. This technique involves a determination of the degree of similarity of DNA-DNA between two organisms. One of the most frequently used-hybridization technique is FISH - fluorescent in situ hybridization. The most precise method for analyzing the nucleic acids is sequencing, i.e. determining the order of nucleotides which form the genetic information of the microorganism studied. Very often in molecular studies the 16S rDNA molecule is subjected to sequencing.- Although microbial communities represent the most diverse type of communities on the Earth (Curtis et al., 2002; Dykhuizen, 1998.), the estimation of their biodiversity in natural communities and the factors controlling them is still a challenge. Established analyzing methods determining functional and structural compositions of natural communities (Forney et al., 2004) are hitherto only capable to detect the most dominating microbial groups (e.g.
[Show abstract] [Hide abstract] ABSTRACT: Most natural environments are characterized by frequent changes of their abiotic conditions. Microorganisms can respond to such changes by switching their physiological state between activity and dormancy allowing them to endure periods of unfavorable abiotic conditions. As a consequence, the competitiveness of microbial species is not simply determined by their growth performance under favorable conditions but also by their ability and readiness to respond to periods of unfavorable environmental conditions. The present study investigates the relevance of factors controlling the abundance and activity of individual bacterial species competing for an intermittently supplied substrate. For this purpose, numerical experiments were performed addressing the response of microbial systems to regularly applied feeding pulses. Simulation results show that community dynamics may exhibit a non-trivial link to the frequency of the external constraints and that for a certain combination of these environmental conditions coexistence of species is possible. The ecological implication of our results is that even non-dominant, neglected species can have a strong influence on realized species composition of dominant key species, due to their invisible presence enable the co-existence between important key species and by this affecting provided function of the system.- We will refer to such a matrix as an interaction matrix. While one can argue that in reality the entries of the interaction matrix should be expressed by the probabilities p i j ∈ [0, 1] at which strain i can kill strain j, it would be very difficult to derive these probabilities in laboratory measurements and therefore experimentalists confine themselves to discrete 0/1 values. In this chapter, we will confine ourselves to this simplification as well, bearing in mind that the situation considered is a special case of the general case.
[Show abstract] [Hide abstract] ABSTRACT: Huge numbers of microbes coexist in almost all habitats of our planet. Their interactions are governed by complex mechanisms, where both competition for resources and toxin production play important roles. Our goal is to understand key mechanisms that lead to coexistence. In this chapter we study many possible scenarios of microbial interactions and we analyze whether or not they can lead to coexistence of species. To achieve this we implemented agent-based models that mimic local dynamics of microbes; initially well mixed microbes from different species interact in a grid with a regular structure. Among others, we show that the coexistence rate is negatively correlated with the number of neighbors of each cell in the grid. Another observation is that the order of selection of focal cells in the grid influences the coexistence rate.- However, we soon faced another challenge, i.e, taxonomically describing these species, as the current taxonomic tools and validation processes are extremely long and disconnected from the acceleration of clinical microbiology research. Of the estimated 10 million different bacterial species on Earth [69,70], only 12,000 have been identified to date, including a little more than 2200 that are associated with humans [71]. Therefore, the implementation of new strategies enabling us to deal with the currently available high throughput genomics and MALDI-TOF-MS has become necessary.
[Show abstract] [Hide abstract] ABSTRACT: By diversifying culture conditions, in a strategy named culturomics, we were able in a short time to grow 124 new bacterial species from human stools, including 39 strict anaerobes. To describe these microorganisms, we use genome sequencing and MALDI-TOF mass spectrometry. Both tools have been major breakthroughs in clinical microbiology over the past decade, have previously been used for taxonomic purposes, and have the advantage over chemotaxonomic methods and DNA-DNA hybridization, to exhibit an excellent intra- and inter-laboratory reproducibility. We developed a polyphasic taxonomic strategy including MALDI-TOF MSand genomic analyses to describe new bacterial species associated with human beings. This strategy, that we have named taxono-genomics, was used to propose the description of 48 new species, the names of 13 of which have officially been validated. In this manuscript, we briefly reviewed the pros and cons of the currently validated taxonomic tools and propose that genomic sequencing and MALDI-TOF mass spectrometry may be incorporated in the taxonomic classification of prokaryotes.
Article
Chemostats are open systems in which nutrients are continually added at a constant rate and spent medium plus cells removed at the same rate, such that a constant volume is maintained. Thus, they are designed to provide a constant, homogeneous environment in which the cells grow at a constant rate. Chemostats are used in two ways to study evolution. The first approach attempts to characterize... [Show full abstract]
Article
Four genes of Escherichia coli whose products are needed to reduce biotin-d-sulfoxide to biotin have been mapped: bisA next to chlA, bisB next to chlE, bisC linked to xyl, and bisD next to chlG. A defective lambda transducing phage, lambdadbis5, which carries all the bacterial genes between the lambda attachment site and chlE, was isolated and shown to have lost the phage genes from int through Q.
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.
This publication is from a journal that may support self archiving.
Learn more