Most fossils are colourless. Ancient organisms were probably just as
colourful as many of the birds and animals we see today, but the fossil
record does not show it because the creatures’ soft tissues were destroyed
during the course of geological time. A few fossil shells do show traces
of colour, but there was no way of telling whether these were the true colours
of the living creature or were caused by other geological processes such
as iron mineralisation. New biochemical information about the molecules
responsible for the colour of shells is now revealing not only how colour
could survive for millions of years, but also that these molecules could
be a key to a better understanding of ancient environments.
Much of what we now know about fossil colour has come from investigations
of the shells of marine invertebrates called brachiopods – also known as
lamp shells because they resemble early Roman oil lamps. One of the most
abundant life forms, brachiopods appeared 550 million years ago and are
of interest to paleontologists because of their continuous presence in the
fossil record. About 400 species of brachiopod live in today’s oceans and
their shells show a variety of colours, from bright red to pink and brown.
Could such colours be preserved during fossilisation?
Interest in the colour of fossil brachiopod shells dates back to the
first discoveries in the 19th century of shells bearing pale remnants of
colour patterns. But if colour is always associated with soft tissues, which
are never preserved, how could these patterns have survived? Our research
team at the University of Glasgow began to answer this question in 1989,
during an investigation of the molecules inside the hundreds of thousands
of tiny calcite crystals that make up the brachiopod shell. Embedded, or
‘occluded’, within these crystals are other molecules, such as proteins,
lipids and carbohydrates. Extracting such intracrystalline molecules is
a complicated task. It involves separating the crystals by destroying the
proteinaceous membranes which hold them together and dissolving any remaining
surface molecules in an organic solvent, leaving the intracrystalline molecules
which must then be carefully separated from any traces of solvent.
Advertisement
In 1990 we used this method to isolate three different proteins from
a living brachiopod shell. We checked that the molecules came from inside
the crystals by making antibodies against them. The antibodies did not react
until the crystals were dissolved, so the target of the antibodies must
be inside the crystals rather than on the surface. We did not set out to
look for colour molecules. But our sample was bright red, so it was clear
that it included the molecule responsible for the bright red colour of the
shell. When we isolated it, using a mixture of techniques taken from analytical
chemistry, it turned out to be a protein with 70 amino acids and two attached
carotenoids. Carotenoids are long-chain hydrocarbons, and are a common source
of orange or red colour in animals and in plants – tomatoes and carrots,
for example.
Since our original study we have found the same caroteno-protein in
six living red brachiopods from as far apart as Japan and New Zealand. The
amino acid part of the molecule is so similar in each species that we concluded
it must be the same molecule. In contrast, we found no trace of the protein
or the carotenoids in 10 other colourless brachiopods.
The position of the caroteno-protein within, rather than between, the
crystals of the shell gave us a clue to how such colour patterns could have
survived for hundreds of millions of years. In his studies of fossil brachiopods
in 1984, Alwyn Williams of the University of Glasgow showed that the calcite
crystals can be seen with an electron microscope even in some of the earliest
fossils, which are more than 500 million years old. It seems that the crystals
have acted as sealed time-capsules to protect caroteno-proteins and other
molecules, while other soft tissues have perished. Matthew Collins, working
at the Dunstaffnage Marine Laboratory near Oban in Scotland, has shown that
molecules located between the crystals are destroyed within a year when
bacteria penetrate the shell. But bacteria cannot easily penetrate the calcite.
In 1988, Art Boucot of the University of Oregon and Gertruda Biernat
of the University of Warsaw discovered colour patterns in brachiopods that
were 400 million years old. Usually the shells show only the positions of
the original colour pattern, the colour itself having faded to shades of
brown or black. This change is probably caused by the breaking of carbon-carbon
double bonds and other structural alterations to the pigment which change
the way it absorbs light. But the fact that the pattern is preserved at
all suggests that remnants of the original pigment survive within the biocrystal.
Provided they are not extensively altered, it may be possible to recover
and identify them, and perhaps the original colour.
No one has tried to extract carotenoids from fossil brachiopods. But
as long ago as 1965 Max Blumer of the Scripps Institution of Oceanography
in La Jolla, California showed that it is possible to extract colours from
fossils. Blumer unravelled the complex degradation chemistry of the pigments
associated with preserved colour patterns in a Jurassic sea lily about 170
million years old. Even though they had been partially protected, these
pigments had decayed into new compounds. Palaeontologists are now beginning
to extend such studies.
The central lowlands of Scotland provide a rich source of well-preserved
fossils, including brachiopods. Our team from the University of Glasgow
has collected many 350-million-year-old Lingula brachiopods. At first sight
these did not appear to be coloured, except from the colour of the rock
showing through the thin shell. But reflected light revealed strong concentric
bands of colour which follow precisely the growth bands in these shells.
This adds another dimension to our studies because these shells are made
of apatite – an even more robust geological material than calcite.
Other molecules we have found inside the crystals of brachiopods include
lipids and carbohydrates. These remnants of the long-dead organisms provide
perhaps the most interesting aspect of these studies: the possibility of
reconstructing ancient environments. Steve Macko of the University of Virginia
and Mike Engel of the University of Oklahoma have developed a method for
measuring accurately the isotopic composition of fossil molecules from shells
and bones. From these measurements they could derive information about the
diet of the fossil .
Carotenoids are good candidates for this technique because animals,
including brachiopods, cannot manufacture them, but obtain them from their
food. Brachiopods have a mechanism for filtering food particles from sea
water passing through their shells. This food contains large quantities
of algae – microscopic plants which are capable of producing a range of
pigments, including carotenoids. Most algae produce some carotenoids, but
their orange or red colour is hidden by dominant green chlorophylls. Algal
pigments collect light energy and convert it into chemical energy during
photosynthesis.
It seems that brachiopods harvest carotenoids from the algae, attach
them to a protein and then secrete the composite molecules from shell-secreting
cells in such a way that they become enclosed within the calcite crystals.
The carotenoids will have an isotopic composition characteristic of the
algae that produced them. Researchers can then retrieve this information
from the fossil record to reconstruct a profile of which algae existed during
the life of the fossil. Because prevailing environmental conditions – temperature,
for example – affect the stable isotopic composition of molecules, researchers
can work backwards from the fossil molecules to derive environmental information.
The same applies to intracrystalline amino acids, lipids and carbohydrates
– provided their isotopic signal is first calibrated in living brachiopods.
This work is now under way, based on the many modern brachiopods which have
long and well-documented fossil records.
Gordon Curry is a Royal Society Research Fellow in the Department of
Geology and Applied Geology at the University of Glasgow.
* * *
FOSSIL FOOD FOR THOUGHT
You are what you eat. This is the idea behind a method of extracting
information from fossils, used by Steve Macko of the University of Virginia,
Peg Ostrom of Michigan State University and Michael Engel of the University
of Oklahoma. They believe that by measuring carbon and nitrogen isotopes
in fossil brachiopod shells they can find out whether the pigment molecules
these contain once formed part of the phytoplankton eaten by the brachiopods.
The researchers have already measured the ratio of nitrogen and carbon
isotopes from amino acids in various species of fossil mollusc. From this
information they found out where these animals fitted into food chains,
by analogy with similar isotopic measurements from modern organisms whose
position in food chains is known. They and other researchers have already
established that the ratio of carbon-13 to carbon-12 and of nitrogen-15
to nitrogen-14 from organic material in modern fossils varies in a predictable
way going up the food chain, increasing slightly.
The accuracy of these results depends on the ‘indigenity’ of the samples.
That is, the researchers must be sure that they are measuring original material,
unchanged by geological processes. This notion is still controversial: not
everyone accepts that the material can have survived. The researchers gained
evidence that it does survive by looking at the amino acids in fossil shells.
Amino acids exist in two mirror-image forms or ‘enantiomers’, labelled
L and d. Life on Earth, with the exception of bacteria, uses the L form.
But when an organism dies, some of the L form is gradually converted into
the d form until the two forms exist in equal amounts – a process called
racemisation.
Macko and his colleagues found in their analyses of organic matter from
the fossil shells that most of the amino acids were the L form, instead
of the expected mixture of d and L. They say that this suggests the molecules
are indigenous – that is, the researchers are studying the remnants of proteins
that are between 100 000 and 20 million years old. ‘We think they are altered
somewhat, but only a little bit,’ says Macko. They have also found unexpectedly
large amounts of L amino acids in oyster shells from Florida that are 100
000 to 1 million years old, clam shells from Chesapeake Bay near Washington
DC, and land snails from the Negev desert in Israel.
But other researchers, including Jeff Bada of the Scripps Institution
of Oceano-graphy in La Jolla, California, and Ed Hare of the Carnegie Institution
in Washington DC have analysed for fossil amino acids and found large amounts
of the d form. Macko says that this is because they measured all the amino
acids, including those not bound to any mineral. He says these would indeed
racemise, to produce the d form, because racemisation involves the rupturing
and reforming of bonds, as happens in solution and with free amino acids.
Steve Weiner at the Weizmann Institute of Science in Israel obtained
unusually large amounts of L amino acids from a high molecular weight material
isolated from molluscs as long ago as 1976, but his findings seemed to contradict
the established fact that racemisation does occur to give a mixture of enantiomers.
The key to Macko, Engel and Ostrom’s technique, and a possible explanation
for Weiner’s anomalous results, is to isolate the largest molecules from
the organic material extracted from fossil bones and shells. For this they
use a dialysis method, which works in an analogous way to a kidney. They
lose some good material in the process but claim also to eliminate contaminants.
In the case of shells from Israel and the US, they looked at protein fragments
that were more than about 50 amino acids units long, reckoning that these
would be too large to migrate or dissolve, and so are more likely to be
preserved in their original form.
Amino acids containing carboxylic acid groups – which bind to the shell
mineral calcite – are most likely to have survived. Macko, Engel and Ostrom
found that the dicarboxylic glutamic and aspartic acids were dominant in
the shells. Using a new and faster continuous flow technique they were able
to isolate and identify these amino acids individually and demonstrate that
they were the L isomer while measuring their isotopic compositions. Previously
this would have taken them weeks, rather than hours, for each amino acid.
The researchers say they have ruled out the possibility of contamination
because the sediments that surrounded the fossils have fewer and different
combinations of amino acids. ‘The similarity in isotope between modern
and fossil food webs is strong evidence for the retention of an indigenous
(isotopic) signal from the high-molecular-weight material isolated from
these fossils,’ they conclude.