91av

During the 1960s and 70s biologists were apt to remark, assigning the
belief to someone else of course, that what is true of the bacterium Escherichia
coli is true of the elephant. The witticism, which no one really believed,
mocked researchers at the cutting ege of molecular biology who worked almost
exclusively with the E coli but applied their conclusions to all living
things. In truth, elephants and E coli do have a lot in common, inasmuch
as all organisms share fundamental pathways of cellular metabolism – such
as protein synthesis or carbohydrate breakdown – as well as the genetic
code. When it comes to genetic flexibility, however, the star performer
is quite definitely E coli.

During the years when the elephant jibe was rife, microbial genetics
was at the forefront of molecular biology. Yet, ironically, it is at the
genetic level that the most striking molecular differences are found between
E coli (representing bacteria in general) and elephants (or any higher organisms
that have cells with nuclei). True enough, these two groups use the same
basic genetic code. But they have different ways of retrieving the information
coded in their DNA and translating it into synthesis of cell material. Moreover,
unlike higher organisms, bacteria can actually exchange DNA with other bacterial
species or genera. E coli, for example, has three ways of transferring or
receiving genes: transduction, transformation and conjugation.

To a certain extent, transduction and transformation are fortuitous.
Conjugation, however, is part of the everyday biology of E coli. It is the
bacterial world’s equivalent of sex: ‘male’ microbes donate fertility genes
to receptive ‘female’ microbes. When the fertility genes (about 30) reside
in E coli’s chromosome, parts of the adjacent chromosomal DNA are often
transferred with them during conjugation. This extra DNA may include genes
useful to the recipient.

More often, though, the fertility genes reside on a small, circular
piece of genetic material, or ‘plasmid’. While E coli carries its main genetic
archive on conventional chromosomes, plasmids equip the microbe with small
libraries of supplementary information. Completely separate from any central
chromosome, they carry a broad palette of bacterial genes and provide genetic
engineers with a useful means of introducing new genes into microbes. Past
research has revealed numerous different E coli plasmids scattered among
different strains of the bacterium. No strain carries a full set, however,
so the chromosomal gene pool of the species as a whole is much larger than
its gene pools of plasmid-borne information.

All this may seem a curious way for a species to handle its genes. But
it has a certain logic. The main habitat of E coli, the lower intestine
of mammals, is very stable. (E coli also lives, albeit transiently, in soil
and fresh water, which it reaches in faeces. Here it is usually overwhelmed
by indigenous microbes). Current thinking is that the chromosome of E coli
meets all the microbe’s ordinary domestic needs, which rarely change, and
that plasmid-borne information comes into play only when the microbe encounters
a stress – a change nutrient, an antibiotic or a competitor. Only a minority
of the population would have the plasmids needed to combat such a stress
but at least the species, if not a majority of individuals, would survive.

This implies that E coli’s chromosomal information ought to be stable.
In fact the opposite is true. For one thing, when fertility genes abscond
with neighbouring DNA during conjugation, E coli’s chromosomal information
fluctuates. For another, many of the genes carried by plasmids – such as
those specifying resistance to the antibiotics kanamycin or penicillin –
are flanked by special DNA which enables them to jump from plasmid to chromosome
and back, or from one plasmid to another. Again, these jumping genes, or
‘transposons,’ cause chromosomal fluctuations. Moreover, the chromosomes
of some strains of E coli contain enigmatic lengths of DNA (‘insertion sequences’)
which code for no known product but can move about the chromosome during
bacterial multiplication, activating or silencing genes. They are present
in plasmids, too.

Finally, the chromosome is not even intrinsically stable – it mutates.
Mistakes are made when DNA is copied during replication, and environmental
mutagens such as background radiation also cause errors. Although E coli
has enzymes which detect and repair such faults, a minority escape repair
and produce true mutants. For the last 20 years researchers have been able
to calculate genome sizes and mutation rates. Elementary arithmetic shows
that the collective daily mutability of the chromosomes of all E coli in
humans is astronomical. Add in the E coli living in other mammals and the
number becomes even larger.

By far the shakiest part of the calculation is the average mutation
rate. For the E coli chromosome has both ‘hot spots’, which mutations are
more frequent than average, and stable zones where they are less frequent.
Moreover, mutation rates seem to change with the physiological state of
the organism. Even so, the margin for error is so huge that the upshot of
the calculation remains valid: among the world’s population of E coli, every
possible mutation is occurring an enormous number of times a day.

The net impression is one of tremendous genetic fluidity, brought about
by plasmids, transpositions, insertions, transduction, transformation and
mutation. Why, then, are microbiologists still able to identify E coli from
its genetic make-up when they examine faeces or environments polluted with
sewage? The answer lies in the microbe’s primary habitat. The mammalian
gut has been stable for many millennia and acts to constrain the flexibility
of E coli’s genome. Though many genetic fluctuations do occur, most die
out because they fail to confer any survival advantages on gut-dwelling
microbes.

Different guts will impose marginally diffeffent genotypes; no doubt
that is why biochemical families of E coli exist, differing in the fine
detail of their enzymes. More practically, it is probably one reason why
the E coli of distant lands, where different diets provide it with significantly
distinct primary habitats, are apt to give the intrepid traveller diarrhoea.

Of course, when a microbiologist takes E coli out of its usual habitat
and grows it in a laboratory, the strain eventually mutates. But this is
a problem which can be lived with because mutations in a test-tube population
come in hundreds rather than thousands of billions. In any case, most laboratory
mutations either die out or have no obvious effect, or merely improve the
growth and manageability of the strain.

Despite its apparent mutability, E coli does have the power to protect
its genetic integrity. If alien DNA enters a cell – by transformation or
as a bacteriophage, for example – E coli can recognise and destroy the unwanted
DNA using restriction enzymes. These chop the DNA up into pieces which the
cell can degrade and excrete (restriction enzymes have proved immensely
useful for manipulating DNA in the laboratory). Plasmids from other E coli
are not alien, of course.

A close relative of E coli is salmonella, a bacterium whose primary
habitat is the intestines of birds (notoriously chickens and ducks) and
which, like E coli, inhabits soils and fresh water transiently. The chromosomes
of E coli and salmonella are so similar that there are long stretches of
identical DNA in the two organisms.

Salmonella is also a match on genetic flexibility. Like E coli it undergoes
transduction, transformation and conjugation. The two organisms readily
exchange plasmids, too. But there is a barrier to the exchange of chromosomal
DNA. If chromosomal DNA from E coli is introduced into Salmonell a typhimurium,
for example, the DNA repair system of the salmonella recognises that something
is wrong and ‘tidies up’ the alien DNA. Somehow the salmonella knows it
is not E coli even though it would not require many mutations for it to
become E coli, or the reverse. Indeed, the global mutation rates of the
genomes of both species are such that, given the right selection pressure,
their convergence could be rapid.

So what stops this happening? Evolutionary history provides some clues.
The structure of the ribosomal RNA of bacteria provides a useful molecular
marker for assessing phylogenetic similarities (see 91av, 21 January
1989). Using this marker, Howard Ochman and Allan Wilson of the University
of California, Berkeley, concluded that the two species diverged about 130
million years ago, ‘intriguingly’ (they wrote) about the time mammals originated.
Because the bird and mammalian habitats of the two organisms have barely
changed since, salmonella and E coli have remained distinct, yet recognisably
close relatives.

Do similar principles apply to microbes with less constant habitats?
in the last decade microbiologists have realised that, though what is true
of E coli may not walsy be true of the elephant, it is generally true of
other eubacteria. (Arachaebacteria, which inhabit extreme environments such
as hot springs, salt pans and oxygen-free decomposing detritus, ane biologically
distinct, (see 91av, 11 August 1990). In particular, the genetic
flexibility shown by E coli seems to be common. transformation was first
discovered in pneumococci and largely worked out in the genus Bacillus.
Its disdovery in E coli is relatively recent, and it now seems that most
bacteria can be transformed by raw DNA. Transduction of genetic information
by bacteriophages occurs in many groups of bacteria. And other mechanisms
have come to light, such as cell fusion among bacilli and an enigmatic gene
transfer agent is some photosynthetic bacteria.

Perhaps most significant, microbiologists now realise that plasmids,
once thought to be exceptional, are so common as to be almost the rule among
bacteria. Many are ‘cryptic,’ meaning that their discoverers have no idea
what their DNA codes for. Rhizobia, the mitrogen-fixing symbionts of leguminous
plants, usually have a few, some of them huge (up to 30 per cent of the
size of the chromosome). In R leguminosarum, which colonises peas, genes
both for recognising the right species of host plant and for fixing nitrogen
reside on plasmids.

But not all plasmids are good news for plants. The plant pathogen Agrobacterium,
for example, carries a plasmid which makes plant cells cancerous. Some microbiologists
believe that Agrobacterium’s chromosome is identical to that of the friendly
Rhizobium, and that the two microbes different solely in their plasmid complement.
Azotobacter chroococcum, a free-living nitrogen fixer, usually has between
five and seven plasmids (none of which carries the fixation genes). The
staphylococcus family boasts a wide repertoire of plasmids, too. And most
soil and water bacteria carry at least one plasmid.

One common group of soil bacteria, the psuedomonads, carry plasmids
enabling them to metabolise exotic chemicals such as toluene, camphor or
oil hydrocarbons, as well as plasmids coding for drug resistance. Some pseudomonads
carry fertility genes which enable them to mate among themselves and pass
plasmids to unrelated, or distantly related, bacteria. Armed with transposons
and other genese, these ‘promiscuous’ plasmids are not sensed as alien by
other bacteria. Quite the opposite: they are willingly accepted by a diverse
range of microbes, from Proteus and Azotobacter to E coli. And any of these
recipients can usually pass the plasmids on again. In fact, many of the
plasmids used in research on E coli were first found in pseudomonads or
salmonella; it can be difficult to tell in which species a plasmid first
originated.

How important are these gene transfer processes in the natural environment?
Transduction seems to be too drastic and to involve fragments of DNA too
small to be of serious ecological importance. But transformation may well
be common in environments with a high microbial turnover, such as decomposing
organic matter. Examples of plasmid transfer in the external world are well
established; for instance, it is the mechanism by which drug resistance
spreads in hospitals or intensive farming units.

It is likely, too, that the chromosomes of all eubacteria are as mutable
as that of E coli. Although some yield mutants less readily in the laboratory
than does E coli, this is because they are better equipped to repair mutations.
So, in principle, the calculation in Box 3 applies to all eubacteria. A
slow-growing soil bacterium, good at DNA repair, might require weeks, even
a couple of months, for its global gene pool to undergo as many mutations
as afflict the world’s E coli; but even so, the potential of eubacteria
for rapid mutation is phenomenal.

The late Bob Hedges, of the Royal Postgraduate Medical School at Hammersmith,
a pioneer of plasmid research, suggested almost 20 years ago that bacterial
evolution has not been linear, as in higher organisms, but rather a patchwork,
with organisms drawing from a communal gene pool. Advances in molecular
genetics have reinforced that view. Restriction enzymes and DNA repair systems
confer a certain degree of integrity on bacterial genomes. but these systems
mutate too. The speed at which bacteria can mutate and their readiness to
pass around packages of genetic information means that bacteria are poised
to react to selection pressure with rapid and substantial genetic changes.

It is safe to conclude that at any stage of this planet’s history the
world of bacteria has been overwhelmingly conditioned by the state of the
biosphere. Anxiety over global environmental change has perhaps made us
more aware of the converse idea: that the activitires of microbes largely
determine the state of the biosphere, and in particular the evolution of
the Earth’s atmosphere. But both propositions are equally true. Just as
laboratory bacteria are artifacts of the culture media, so the bacterial
world can be viewed as an artefact of the rest of the biosphere. Yet in
the face of environmental stress, a malleable genone has given the microbe
the edge on the elephant.

Finally, reflect, if you will, on how drastically humanity has changed
the biosphere during its brief strut on the terrestrial stage; one wonders
how much of today’s microbial world we have ourselves created.

* * *

1: How Escharichia coli pass their genes around

Transduction: Some bacterial viruses (bacteriophages), after they have
infected an E coli cell, combine with some of their host’s DNA and make
it part of themselves. As the virus DNA multiplies, the piece of the host
DNA multiplies along with it. In due course the host dies and releases into
the environment virus carrying fragments of its DNA. When the hybrid virus
attacks and enters a new cell, it carries in that DNA. Most of the cells
attacked will succumb to the infection, but a few will be resistant, and
these will integrate the alien DNA into their own genomes. If the new DNA
includes DNA sequences which the resistant host can use, the host will do
so and its genotype will thus have been changed. The pieces of DNA transuced
by bacteriophages may carry one or two whole genese, but they are often
smaller fragments which may nevertheless modify genes already present in
the recipient.

Transformation: E coli at a certain stage of their growth cycle can,
after treatment with chemicals such as calcium chloride or rubidium chloride,
take up raw DNA (for example, DNA purified in the laboratory) and, if it
contains appropriate sequences, will incorporate them into their own genomes.
Transformation can involve pieces of DNA comprising dozens of genese, such
as purified plasmids.

Conjugation: E coli only conjugate when one of the cells possesses fertility
genes and the other does not. The two organisms come together, a tube grows
betweem them, pulling them close, and the fertile strain donates fertility
genes to the recipient, which then becomes fertile itself. Conjugation may
involve the transfer of anything from about 30 to over 100 genes.

* * *

2: Plasmids

These are mini-chromosomes found in bacteria. They are distinct from
the cell’s main store of DNA, the bacterial chromosome, yet still multiply
during cell growth. In E coli plasmids range from about 3 to 20 per cent
of the size of the chromosome. Most often there is one copy of a given plasmid
per chromosome, but with the small ones there may be several copies. Sometimes
two or more kinds of plasmid are present in a given strain of E coli, though
plasmids seem to belong to groups, not all of which are compatible with
each other.

Plasmids that do not carry fertility genese are unable to move from
one bacterium to another, but they can be mobilised by plasmids, which do
have them. Fertility genes enable a plasmid’s genetic information to be
transferred from a donor to a recipient strain.

Plasmid genes can determine a wide variety of bacterial properties,
including resistance to antibacterial substances, ability to make an antibiotic
or toxin, or to metabolise the sugar sucrose. As an example, a plasmid known
as R100 comprises 90,000 base pairs of DNA (compared with the 4 million
base pairs of the chromosome) and has genes which make its host cell fertile
and resistant to five antibacterial substances.

* * *

3: Global mutations

The bacterium E coli grows at a phenomenal rate. An average human might
discharge 200 grams of faeces a day. The E coli count of human faeces is
surprisingly constant at about 10 ⁸ cells per gram. Thus one
human produces about 2 x 10 ¹⁰humans. So the global growth rate
of our intestinal E coli is about 10 ¹⁹ cells per day. The E
coli genome comprises about 4 x 10 ⁶ base pairs of DNA. Genes
differ in length, averaging about 10 ³ base pairs. So E coli
has about 4 x 10 ³ genes.

Spontaneuous nonlethal mutations (to drug resistance, to a new nutritional
requirement, etc) occur in the genomes of multiplying E coli at frequencies
in the range 1 per 10 ⁴ to 1 per 10 ⁹ new progeny.
For the sake of argument, say 1 in 10 the power of 7. Each one significes
an altered gene. Therefore more than 10 ¹² E coli genes mutate
daily inside humanity. Which means that every E coli gene of the E coli
genome traversing the human intestine mutates at least 2.5 x 10 ⁸
times daily.

John Postgate FRS is emeritus professor of microbiology at the University
of Sussex.