Evidence for the unique function of docosahexaenoic acid (DHA) during the evolution of the modern hominid brain

Abstract

The African savanna ecosystem of the large mammals and primates was associated with a dramatic decline in relative brain capacity associated with little docosahexaenoic acid (DHA), which is required for brain structures and growth. The biochemistry implies that the expansion of the human brain required a plentiful source of preformed DHA. The richest source of DHA is the marine food chain, while the savanna environment offers very little of it. Consequently Homo sapiens could not have evolved on the savannas. Recent fossil evidence indicates that the lacustrine and marine food chain was being extensively exploited at the time cerebral expansion took place and suggests the alternative that the transition from the archaic to modern humans took place at the land/water interface. Contemporary data on tropical lakeshore dwellers reaffirm the above view with nutritional support for the vascular system, the development of which would have been a prerequisite for cerebral expansion. Both arachidonic acid and DHA would have been freely available from such habitats providing the double stimulus of preformed acyl components for the developing blood vessels and brain. The n-3 docosapentaenoic acid precursor (n-3 DPA) was the major n-3-metabolite in the savanna mammals. Despite this abundance, neither it nor the corresponding n-6 DPA was used for the photoreceptor nor the synapse. A substantial difference between DHA and other fatty acids is required to explain this high specificity. Studies on fluidity and other mechanical features of cell membranes did not reveal a difference of such magnitude between even alpha-linolenic acid and DHA sufficient to explain the exclusive use of DHA. We suggest that the evolution of the large human brain depended on a rich source of DHA from the land/water interface. We review a number of proposals for the possible influence of DHA on physical properties of the brain that are essential for its function.

Full text

Evidence for the Unique Function of Docosahexaenoic
Acid (DHA) During the Evolution of the Modern
Hominid Brain.
Crawford MA1, Bloom M2, Leigh Broadhurst C3, Schmidt WF3, Cunnane
SC4, Galli C5, Ghebremeskel K1, Linseisen F2., Lloyd-Smith J2 and
Parkington J6
Lipids 2000, 34: S39-S47
1. Institute of Brain Chemistry and Human Nutrition, University of North London,
London N7 8DB UK.
2. Dept of Physics, University of British Columbia, Vancouver V6T 1Z1. 3. USDA
3 United States Department of Agriculture, Environmental Chemistry
Laboratory, NMR Facility, Beltsville, MD 20705, USA.
4. Dept Nutritional Sciences, Faculty of Medicine, University of Toronto, Ontario M5S
3E2, Canada.
5. Institute of Pharmacological Sciences, Milan 20133, Italy.
6. Archaeology Department, University of Capetown, South Africa.

1
EVIDENCE FOR THE UNIQUE FUNCTION OF DHA DURING THE
EVOLUTION OF THE MODERN HOMINID BRAIN.
Crawford MA 1, Bloom M2, Broadhurst CL3, Schmidt WF3, Cunnane SC4, Galli C5
Gehbremeskel K1, Linseisen F2, Lloyd-Smith J2 and Parkington J6,
1. Institute of Brain Chemistry, London N7 8DB UK. 2. Dept of Physics, University of
British Columbia, Vancouver V6T 1Z1. 3. USDA Beltsville, Environmental Chemistry
Laboratory, MD 20705, USA. 4. Dept Nutritional Sciences, University of Toronto,
Ontario M5S 3E2, Canada. 5. Institute of Pharmacological Sciences, Milan 20133, Italy.
6. Archaeology Department, University of Capetown South Africa.
ABSTRACT
The African savanna ecosystem of the large mammals and primates was associated with a
dramatic decline in relative brain capacity. This reduction happened to be associated with a
decline in docosahexaenoic acid (DHA) from the food chain. DHA is required for brain
structures and growth. The biochemistry implies that the expansion of the human brain
required a plentiful source of preformed DHA. The richest source of DHA is the marine
food chain while the savannah environment offers very little of it. Consequently H.
sapiens could not have evolved on the savannahs. Recent fossil evidence indicates that the
lacustrine and marine food chain was being extensively exploited at the time cerebral
expansion took place and suggests the alternative that the transition from the archaic to
modern humans took place at the land/water interface.
Contemporary data on tropical lake shore dwellers reaffirms the above view.
Lacustrine habitats provide nutritional support for the vascular system, the development of
which would have been a prerequisite for cerebral expansion. Both arachidonic acid (AA)
and DHA would have been freely available from such habitats providing the double
stimulus of preformed acyl components for the developing blood vessels and brain.
The ω3 docosapentaenoic acid precursor (ω3DPA) was the major ω3 metabolite in
the savanna mammals. Despite this abundance, neither it or the corresponding ω6DPA
were used for the photoreceptor nor the synapse. A substantial difference between DHA
and other fatty acids is required to explain this high specificity. Studies on fluidity and
other mechanical features of cell membranes have not revealed a difference of such
magnitude between even α-linolenic acid (LNA) and DHA sufficient to explain the
exclusive use of DHA. We suggest that the evolution of the large human brain depended
on a rich source of DHA from the land/water interface. We review a number of proposals
for the possible influence of DHA on physical properties of the brain that are essential for
its function.
Key words: arachidonic, brain, blood vessels, docosahexaenoic, evolution, lacustrine,
marine foods, membranes, nutrition, Nyasa, Turkana, Australopithecus, H. erectus, H.
sapiens.

2
I THE INVERTBRATES, REPTILES AND MAMMALS
THE ORIGIN OF AIR BREATHING ANIMALS :
For the first 2.5 billion years of life on the planet, the blue-green algae dominated the proto
oceans. The photosynthesis of the algae produced complex molecules including proteins,
carbohydrates and lipids which were rich in ω3 fatty acids. Based on the explosion of the
phyla in the fossil record, oxidative metabolism became predominant about 600 million
years ago. Thus animals, visual, nervous systems and brains evolved in a DHA rich
environment (1). DHA is the most prominent essential fatty acid used for the structures
and functions of the photoreceptor and synaptic junction. So the question addressed in
this paper is why, and what are the evolutionary implications of the abundance of DHA in
the marine food chain compared the relatively paucity in land ecosystems.
MAMMALS.
The dominance of ω3 fatty acids in the early oceans was associated with fish and reptiles
requiring ω3 fatty acids for their reproduction. This dominance persisted until the end of
the Cretaceous period, 70 million years ago. In the wake of the extinction of the giant
reptiles, cycads, ferns and their allies, the flowering plants appear in the fossil record.
They stored lipids, for energy during germination, containing seed oils rich in
ω
6 fatty
acids. Then, a new set of species, the mammals, evolved: it may not be a coincidence that
they required ω6 fatty acids for their reproduction.
Mammalian brain size is larger in relation to body size compared to the previous egg laying
amphibians, reptiles and fish. The difference could be explained by the evolution of the
placenta. The placenta enables nutrients and energy to be focused continuously on the
development of one or a small number of progeny throughout the critical time of brain
development. In the human, 70% of the calories transferred by the placenta to the fetus is
devoted to brain growth. The placenta is a rapidly growing vascular system with a high
requirement for ω6 fatty acids especially AA. In 42 species so far studied, AA and DHA
are major acyl constituents with the precursors being virtually absent. So the emergence of
the ω6 fatty acids may have added the missing biochemical link, liberating genetic
potentials for vascular development and hence the evolution of the placenta, mammary
gland and the larger brains of the mammals.
The difference between species is not the chemistry but the extent or size to which the brain
is developed (1). The manner of these differences is so large as to imply the availability of
AA and DHA were limiting factors in the evolution of the brain (1,2). Indeed, the need for
both ω6 and ω3 fatty acids for development and health of the vascular system and brain
has long been recognised (3.).

3
II EVIDENCE FROM PALEONTOLOGY
THE EVOLUTION OF THE MODERN HUMAN BRAIN :
The accepted dogma regarding the evolution of Homo sapiens is that he was originally a
hunter and gatherer on the African savanna. A study of savanna and other African species
show that as they evolved larger and larger bodies, the relative size of the brain diminished
logarithmically with increase in body weight (1,4). A cebus monkey of 0.9 kg body
weight has 2.3% of its body weight as brain, a 60 kg chimpanzee 0.5%. The larger gorilla
at 110 kg has only 0.25% brain which is physically smaller than the chimpanzee’s brain.
At the extreme, the one ton rhinoceros has <0.1% with its brain weighing only 350g. It
reaches that massive one ton body weight at four years of age.
Why does size and velocity of growth matter? The reason is that the biosynthesis of AA
and DHA is relatively slow, and may not be able to keep pace with body growth in fast-
growing animals. Rats and mice desaturate and chain elongate the parent essential fatty
acids to produce larger amounts of AA and DHA than their precursors. Stepping up in
size from the guinea pig to the wild pig the impact of velocity of growth results in a
progressive decline in AA and DHA whilst the precursors linoleic and α-linolenic acids
become more dominant in liver lipids (1). Instead of DHA, the
ω
3DPA is now the major
metabolite of
α
-linolenic acid (5).
So the faster an animal grows, the larger it becomes and the greater is the constraint of the
biosynthesis of AA and DHA. The large savanna, mammals of Africa all shared the same
fate namely, DHA and brain capacity declined as body size accelerated. The important
issue is that desaturation and elongation in these large mammals peters out at the ω3DPA
with relatively little DHA being synthesized. What little DHA is synthesized is used in the
brain and photoreceptor. The abundant ω3DPA is not found here. Brain size was
sacrificed not brain DHA (1, 6). This fact raises two issues:
• The savanna food chain on which Homo sapiens is supposed to have evolved is fairly
devoid of DHA so how did Homo evolve a large brain?
• Why was the more readily available ω3DPA not used for the brain instead of DHA?
M odern human intellect and brain-specific nutrition .
Australopithecus spp. are unremarkable in their apparent encephalization throughout their
evolutionary history as far as can be deduced from the fossil record . No australopithecine
has a cranial capacity much over 500 cm3 (7 ), despite the existence of the genus for over 3
Myr. Contrast this to genus Homo, whose cranial capacity doubled from H. erectus to H.
sapiens in a span of at most 1 Myr (Table 1). The Homo spp. fossil evidence and
encephalization quotient (EQ ) values do not support a slow, linear Darwinian progession
towards modern intelligence, but rather a sudden, exponential growth of relative brain size
in the last 200,000 years or so.
The earliest evidence for modern H. sapiens is found in Africa. Homo spp. in general are
associated with lake shore (lacustrine) environments in the East African Rift Valley, while
Australopithecines are associated more with forested areas (8,9). Thus far, evidence for
precocious cultural development of Homo sapiens is exclusively confined to lacustrine and

4
coastal marine environments. Lakeshore sites in the Rift Valley have yielded fairly
sophisticated stone tools as old as 260 kyr associated with H. sapiens remains. The
implications of this land/water habitat providing brain specific nutrients has largely been
overlooked.
Table 1. Mean brain volumes and encephalization quotients (EQ) for selected hominoid
species. EQ1 from the calculations of Martin (4); EQ2 from the calculations of Harvey
and Clutton-Brock (47).
Species Brainvolume (cm3) EQ1 EQ2
A. afarensis 384 1.23 1.45
A. africanus 420 1.31 1.62
A. boisei 488 1.37 1.72
A. robustus 502 1.49 1.92
H. habilis 579-597 1.74-1.79 2.10-2.29
H. rudolfensis 709 1.41 2.11
H. erectus 820-844 1.59-1.63 2.38-2.44
H. sapiens 1250 3.05 4.26
P. troglodytes 410 1.25 1.57
Fossils from coastal sites on the southern Cape of South Africa are widely regarded as the
earliest modern human fossils (10,11,12). At numerous sites along the Cape, hominid
occupation is evidenced dating from 120 to 60 kyrs before now. Modern human fossils
dating to about 100 kyr have been recovered at Klasies River Mouth and Border Cave are
found associated with incontrovertible evidence for the consumption of seafoods dating to
the time of rapid cerebral expansion (10,13,14,15).
Parkington (15) points out that in coastal hunter-gatherer cultures, women are responsible
for collecting shellfish. So stone age women could have easily provided themselves with a
plentiful source of brain-specific nutrition, even when strength/mobility are compromised
during pregnancy and lactation. Children would have naturally participated in exploitation
of, at that time, this extremely rich resource. Early consumption of shellfish is also present
in the archaeological record on the Mediterranean coast of Africa. It is likely to have
occurred elsewhere as well, however most possible coastal sites which could be
investigated have been obliterated by the higher sea levels of the current interglacial era.
Successful e arly Homo spp. were Tropical Coastal Migrants
Both H. erectus and H. sapiens sucessfully colonized areas outside of Africa. There is all
but unanimous agreement amongst paleoanthropologists that H. sapiens originated in
Africa and then spread throughout the world (16 ,17, 18,19). Recently, stone tools of 0.8-
0.9 Myr have been found on the island of Flores, one of the Wallacean islands lying
between Java and Timor in Indonesia (20). The antiquity of the specimens suggests they
were manufactured by H. erectus, not H. sapiens. Although Java and Bali were
periodically connected to the mainland during the Pleistocene glaciation, even at times of

5
lowest sea level, reaching Flores would have required a sea crossing of at least 19 km.
This implies that at least in Indonesia, H. erectus had already reached the cognitive
capability to build and use watercraft repeatedly.
Previous to Morwood et al.’s (20) recent discovery, the earliest evidence for the use of
watercraft dates from about only 40 kyr or slightly earlier with the migration of H. sapiens
from the Wallacean Islands to Australia. That initial colonization of Australia, Tasmania
and New Guinea was accomplished by modern H. sapiens. Similar to the movements of
H. erectus, these early migrants are considered to have followed a tropical coastal route.
Therefore, both the earliest occurrence of modern H. sapiens, the earliest use of watercraft
and successful colonization of Southeast Asia were intimately associated coastal migrations
and with the utilization of food resources from the marine food chain.
We consider this association not accidental nor coincidental, but a reflection of the dramatic
influence of brain specific nutrition on the evolutionary process. We do not accept the
postulate that H. sapiens a priori evolved a large, complex brain, then began to hunt in
order to maintain it--the brain must come first.
Our thesis is that there must have been enough long chain polyunsaturated fatty acids (LC-
PUFA) available in the diet to:
1. Provide many generations of hominids with fuel for fetal/infant development as well as
childhood and adult needs for the cardio-vascular system and the brain.
2. Allow for substantial amounts of 18 carbon polyunsaturated fatty acids (PUFA) which
would have been oxidised for energy requirements (21,22 ).
3. Explain and allow for our inefficient conversion of LA to AA and LNA to DHA (which
is illustrated by preferential incorporation of DHA in the infant brain (23) and
improved problem solving in infants fed DHA which persisted beyond the period of
supplementation (24)).
The evidence on the extensive coastal and lacustrine exploitation implies LC-PUFA were
consistently abundant in the food supply as we evolved. Homo did not however, go as far
as the obligate carnivores in which the desaturation process is barely detectable (25). If H.
sapiens had developed his intellect by evolving into a primate which can make heroically
efficient use of 16 and 18 carbon PUFA from vegetarian sources, we would see an
obvious signature in our current PUFA metabolism, since we are only a 300-100 kyr old
species. Instead we see the opposite. We might hypothesize that Australopithecus spp.
could not mount this heroic metabolic effort either which explains why their brain capacity
was constrained by their land based diet at 400 -500 cc for 3 Myr and explains why
exploitation of coastal foods fits with the rapid and recent cerebral expansion to 1.3 kg after
3 Myr of a static brain size (9,15).

6
III CONTEMPORARY EVIDENCE.
The human brain requires a rapidly developing heart and vascular system to meet the
prodigious energy and nutrient demand during its development. The vascular system itself
has a high requirement for AA (26). Hence the principles of vascular development were
sine qua non vital for the final evolution of the large human brain.
If we now examine the contemporary evidence on cardiovascular disease we find that land
based animal fats have been causaly linked to heart disease as revealed by the Seven
Countries study of the 1950s and even earlier (27). Saturated fats and vascular
degeneration would be incompatible with CNS expansion. Also, there is increasing
evidence that cardio-vascular disease has its origin in poor fetal nutrition (28) consistent
with our hypothesis of long term, multi-generation effects operating on vascular and CNS
evolution. Those forces can act for expansion or degeneration.
Worldwide diets and cardiovascular risk factors show that marine fats, especially DHA, are
cardioprotective (29,30, 31). It is well known that people living on sea foods have low
cardiovascular risk factors. The diet of contemporary populations beside East African
lakes (Lake Nyasa and Lake Turkana) is still largely based on fish rich in ω3 and ω6 LC-
PUFA. From Table 2, calculated intakes of ω3 LC-PUFA are 1-4 g/d and AA 0.5-1.0
g/d, compared with ω3 LC-PUFA 0.2-0.3 g/d and AA 0.1-0.2 g/d for populations
consuming Western diets. Total dietary fat in the African populations is similar at 10-15%
of the dietary energy (32).
Table 2. The FA composition data of the fish from Lake Nyasa and Turkana (wt% of fatty
acids): Turkan
Tilapia
Turkan
Perch Nyasa
Mfui Nyasa
Kambale Nyasa
Carp Nyasa
Mbelele
cat fish
Fat g% wet
weight 2.3 2.6 1.1 1.8 4.9 10.3
20:4ω6 8.4 7.7 8.0 5.8 5.8 4.3
20:5ω3 2.8 1.8 3.1 2.2 1.8 4.2
22:5ω3 3.2 3.8 3.7 5.2 5.0 1.8
22:6ω3 15.7 18.1 19.1 13.3 7.8 8.6
We have compared East African lake shore, mainly vegetarian and Europeans
cardiovascular risk factors (Table 3). Blood cholesterol, blood pressure, lipoproteins
(Lp(a)) are lower in the contemporary Africans living on lake shores of Turkana and Nyasa
compared to their vegetarian cousins and Europeans. Plasma AA, eicosapentaenoic acid
and DHA are highest in the fish eating, lake shore people and least in the vegetarian or
omnivorous inland cousins. Furthermore, comparison of children from the lake shore
versus European children living in East Africa showed that the two populations can be
separated on the basis of blood cholesterol at the age of 6 years! Whilst the European
children’s blood pressure and cholesterol continues to rise the Africans remain stable
illustrating the compatibility of the lacustrine diet with good cardiovascular performance
and the needs of fetal brain expansion. It is of interest that the Turkana have the highest
mitochondrial DNA diversity of any ethnic group. In fact 36 Turkana people have a higher
diversity than the world-wide population database. The simplest interpretation is that
humans date back to the East African Rift Valley (33).

7
Table3. Comparison of cardiovascular risk fatty factors, plasma and fish fatty acid
composition of lake shore, fish eating vegetarian and European, communities in the Rift
Valley and East Africa.
_____________________________________________________________________
Populations Largely vegetarian Lake shore Significance of
Differences ___
Plasma lipids (mg/dL) p
Plasma T C 136. ± 39.8 n=686 122 ± 30.9 n=622 < 0.05
Plasma TG 105 ± 53.1 80.6 ± 40.7 <0.001
Lp(a) 32.3 ± 22.4 19.9 ± 17.7 <0.001
Blood pressure (mmHg)
Nyasa Systolic 135 ± 20.4 120 ± 15.1 <0.001
Diastolic 77.6 ± 10.6 70.5 ± 8.9 <0.001
Turkana Systolic European El Molo
Age yrs 0-3 82 ± 14 n=15 85 ± 9.7 n=6 ns
6-10 98 ± 22 n=16 87 ± 17 n=16 ns
16-20 90 ± 28 n=24 119 ± 28 n=14 <0.001
25-45 31± 34 n=265 105 ± 30 n=24 <0.001
Cholesterol 0-3 102 ± 24 n=15 97 ± 18 n=12 ns
mg/dl 6-8 167 ± 35 n=18 112 ± 32 n=16 <0.01
25+ 228 ± 44 n=145 147 ± 49 n=24 <0.001
Plasma FA (wt %)
Lake Nyasa
LA 14.8 ± 4.30 n=53 23.9 ± 4.37 n=53 <0.002
LNA 0.60 ± 0.20 0.31 ± 0.14 <0.001
AA 8.26 ± 1.94 9.85 ± 2.68 <0.005
EPA 0.72 ± 0.22 2.48 ± 1.35 <0.001
DHA 1.48 ± 1.04 5.93 ± 1.77 <0.001
Lake Turkana , El Molo* n=32 Bantu n=98 European n=124
LA 9.3 ± 3.0 22.8 ± 4.8 19 ± 4.9
DHLA 1.9 ± 0.7 3.5 ± 1.3 2.4 ± 1.1
AA 12.2 ± 3.8 5.1 ± 2.7 7.0 ± 2.4
EPA 4.7 ± 1.3 1.6 ± 0.8 0.5 ± 0.2
DPA 2.6 ± 0.9 3.2 ± 1.2 2.1 ± 0.9
DHA 9.3 ± 3.3 3.5 ± 1.3 5.6 ± 2.2
* p < 0.001 for LA, AA & DHA in El Molo cfd all others.
Legend to table 3
N, number of subjects; TC = Total Cholesterol; TG, Triglycerides. LA, Linoleic
Acid; LNA, alpha Linolenic Acid; AA, Arachidonic Acid; EPA, Eicosapentaenoic
acid; DHA, docosahexaenoic acid. Values are the average ± SD. Adapted from ref.
(32): El Molo live on a lava desert which runs down to the eastern shore of Lake
Turkana, N. Kenya (48), The Bantu were Baganda and Bunyoro people of central
Uganda, The Europeans were living in East Africa mainly, Kampala Uganda (data

8
from 49). The slow conversion of linoleic to AA and α-linolenic to DHA is illustrated
in the equilibrium of the higher linoleic acid content of the plasma lipids and the lower
AA and DHA in the vegetarian and European plasmas compared to the fish eating
people where preformed AA and DHA is consumed in the duiet and appears as higher
levels in the plasma. Such data reflects the rate limitations of the conversion process
especially ∆-6 desaturase which is involved twice in the synthesis of DHA (50). The
higher circulating levels of AA and DHA would favour their incorporation into the
developing brain where their incorporation is an order of magnitude greater than their
synthesis from precursors (51).
These unique conditions of the Rift Valley lake shores replicate the contrast in the high
mortality from vascular disease and high prevalence of mental disease amongst US and
Europeans versus the low risk of Japanese, Greenland and Inuit Indians living on a rich
fish and sea food diet (34 ). Similarly, comparison of fish eating populations in the
Faroes compared to their genetically similar mainland Denmark contemporaries, shows that
the high intake of fish and seafoods results in higher birthweights and a lower proportion
of preterm deliveries (35). The advantage of longer gestations is the greater exposure of
the fetus and its developing vascular system and brain, to the placental biomagnification of
AA and DHA (36). The conclusion is that land based diets led increasingly in this century
to cardiovascular disease being no. 1 killer in Western consumers which would have made
cerebrovascular expansion in utero difficult, and been incompatible with expansion of the
hominid brain.
Experimental support for the above case came in an unexpected result from studies on
diabetes by our colleagues, Professor Lucilla Poston and others at St Thomas’s Hospital
Medical School. Pregnant rats were subjected to high saturated fat diets similar to those
consumed in Western countries and blood vessel function tested in mothers and newborn
offspring. The results from small vessel myography described arterial dysfunction
specifically associated with the high, fat Western type diet. Vascular function tests on the
15 day old pups from mothers on the high (30%) fat diet exhibited reduced vascular
endothelium dependent relaxation to acetylcholine (ACh) with evidence of constrictor
responses to noradrenaline and the thromboxane mimetic U46619. The vascular
dysfunction was still observable at 120 days of age despite rearing on a normal diet. Thus
the high fat diet fed to the mother changed the intrauterine milieu which caused persistent
vascular dysfunction in the newborn animals without any genetic predisposition. Diabetes
imposed on the high fat diet made vascular function worse. Biochemical analysis of the
tissues from the control low fat and high saturated fat animals revealed the high fat diet
selectively depressed liver DHA of the pups. Here is experimental evidence of the negative
influence of land based animal fats fed to the mother on the next generation, emphasising
the importance of long term nutritional forces (37).

9
IV THE SPECIFICITY OF DHA.
The question which arises from this discussion is what is so special about DHA? Why has
DHA been chosen so overwhelmingly for photoreceptor and synaptic membranes, despite
the availability of similar molecules which would be less difficult to obtain, and are less
vulnerable to oxidative damage (38,39)? In particular, what advantage does it convey
relative to the very closely related ω3 and ω6 DPAs, each of which differs from DHA only
in the absence of one double bond (between carbons 4 -5, and 19-20, respectively)?
As described above, Nature’s preference for DHA in the brain is strikingly demonstrated in
large land mammals, in which DPA is the dominant ω3 metabolite yet neural membranes
still retain the DHA-rich composition observed in other species (possibly at the expense of
gross brain size, since DHA is in such limited supply). Significant quantities of the ω6
form of DPA are observed only in situations of artificial ω3 deficiency, yet even here brain
membranes are resistant to decreases in their DHA levels. Nature is thus highly sensitive
to the slight difference between DHA and DPA molecules; the presence of DHA’s full
complement of six double bonds is for some reason an important priority in neural
membranes.
What is the cause of such specificity in membrane composition? It is understood that
biological membranes, while always having the form of a fluid lipid bilayer, have detailed
distributions of lipid and protein molecules that reflect the interactions between lipids and
integral membrane proteins (40). It seems that the one missing double bond in DPA
species renders them unsuitable for whatever lipid-protein interaction favours DHA’s
inclusion in membranes of the brain.
Tight regulation of membrane lipid composition extends to differentiation between
polyunsaturated species. We recently investigated the relationship between DHA and AA
levels in plasma of red cell membranes of maternal and fetal blood samples. While these
studies revealed only a modest correlation in levels of the two PUFA species in plasma
choline phosphoglycerides (r=0.62, p<0.001, n=74), a strong positive correlation was
revealed between DHA and AA in the maternal red cell membrane (r=0.85, p<0.0001,
n=74), and a still tighter relationship in the red cells of preterm infants (r=0.88,
p<0.00005, n=24) (41). Bearing in mind the very different dietary origins of these two
PUFAs, such significant correlations indicate that some powerful mechanism exists to
regulate their relative abundances in the membrane.
It is possible that specific esterification processes could explain the correlations. The
ethanolamine (PE) and serine phosphoglycerides (PS) have the highest content of DHA.
In the brain there is an active base-exchange reaction for serine and ethanolamine. Other
ω3 fatty acids do not esterify easily with PE and PS. So specificity of composition could
be brought about by DHA-PS or DHA-PE formation. None-the-less the double bonds in
positions 4-5 and 19-20 would still have to be relevant for the esterification to explain why
the ω6 and ω3 DPAs might not match these conditions.
Nuclear magnetic resonance (NMR) and fluorescence studies have attempted to
differentiate the membrane properties conferred by PUFAs. In another paper (42), we
discuss some of the constraints of such approaches and review the results obtained to date.
Holte et al., (43) have conducted a thorough NMR investigation of the effects of

10
polyunsaturation on lipid acyl chain orientational order, which revealed significant changes
as the number of double bonds increased from one to three, but little difference as further
double bonds were introduced. Ehringer et al. (44) directly compared the effects of 18:3
and 22:6 on membrane physical properties, and observed considerably higher permeability
and perhaps vesicle fusability in the samples containing DHA.
In summary, a number of studies have been conducted on the physical effects of
polyunsaturation on membranes, in which DHA has been compared to a range of other
unsaturated chains having from one to five double bonds. Thus far, however, all
differences that have been measured have been matters of degree, and none provide a
compelling explanation for the striking specificity with which DHA is selected for
membranes of the eye and brain. In addition, to our knowledge no study has compared
DHA to either species of DPA to search for whatever property it is that causes neural
membranes to discriminate so clearly between these seemingly similar molecules. The
minimized energy structures presented here (see below) represent a preliminary step in this
direction.
Where, then, can we hope to find an explanation of DHA’s preferred status in neural
membranes? An obvious starting point is in protein-lipid interactions: some way in which
DHA favourably affects any of the myriad integral membrane proteins which are so
important to neural membrane function. Such an effect could conceivably involve either a
specific, molecular interaction between lipid and protein, or some modulation of bulk
properties of the bilayer which alters protein function. We believe that specific binding
interactions between lipid and protein molecules in a biological membrane are unlikely,
since the membrane’s fluid state means that individual lipid molecules will be undergoing
rapid translational diffusion within the bilayer, and thus will never be in prolonged contact
with any one protein. Furthermore, Brown’s studies (45) on the rod photoreceptor outer
segment membrane revealed that specific chemical-type interactions could not be the cause
of DHA’s established role in supporting rhodopsin function. It was found that full
rhodopsin efficiency could be obtained by substituting other lipid mixtures designed to
mimic the bulk mechanical properties of the physiological, DHA-rich membrane. This
gave rise to a model in which DHA’s role was to promote mechanical conditions in the
membrane suitable to stabilize certain critical conformational changes undergone by
rhodopsin in the course of photoactivation. These models do not fully reconstitute the
structure of the photoreceptor cell and its synaptic function, the ten thousand fold adaptive
capability of which is still unexplained. However, should this model be valid to
conditions in vivo it could potentially be extended to other G-protein systems elsewhere in
the CNS.
Applying this reasoning to the problem of distinguishing DHA from DPA, we must find a
way in which the difference of one double bond might have a large enough impact on some
bulk membrane property. The simulated structures shown in the figures are encouraging in
this respect, as they show considerable differences between the minimized energy
conformations of di-DHA PE and di-DPA PE (perhaps the first results which show a
difference of sufficient magnitude to account for Nature’s longstanding, clear
discrimination between the two). It must be stated, though, that these simulations have
been carried out in vacuum and report only the lowest energy state; their applicability to
lipid molecules in a fluid bilayer at physiological temperatures is thus open to question.
A more speculative, possibility is that DHA in vivo plays a more direct role in
neuronal signalling, in which some special properties conferred on the membrane

11
by DHA chains exert an influence on membrane electrical phenomena. These might
include distinctive dielectric or polarizability properties arising from the unique
periodic and symmetric arrangement of double bonds in the DHA chain (which
would be disrupted in DPAs). It is conceivable that some polarization of π-electron
clouds might occur, and perhaps even be transmitted from one double bond to
another (either within a given chain, or between neighbouring chains in the
membrane). It must be emphasized that this model is strictly speculative, and there
is no evidence, experimental or theoretical, to support it. An experiment to measure
the dielectric response of lipid systems at a broad range of applied frequencies is
currently being developed. In a similar vein, Penrose (46) has postulated that some
brain functionality may arise due to quantum coherence in the microtubules of
neurones (46); it may be worthwhile to look for a similar phenomenon in
membranes containing DHA.

12
Legend to figures:
Global three dimensional energy minimized structures of DHA, n-3 DPA, n-6 DPA and
various phospholipids containing these LC-PUFA were constructed with MOPAC
software (Alchemy 2000 v. 2.0, Tripos Inc., St. Louis, MO). MOPAC (molecular orbital
pacakage) which calculates the steric energy and energy minimized configuration of a given
molecule by successive approximation, and is considered to be reasonably accurate as
compared to known structures. The free fatty acids (FFA) DHA, n-3 DPA, n-6 DPA are
shown in Figures 1 to 3, respectively. The sixth ethylenic bond in DHA changes the
character of the FFA, completing the methylene interupted sequence along the carbon
chain, and conferring a folded, slightly spiral nature to the molecule. In n-3 DPA, the side
of the chain closest to the terminal methyl is essentially ethylenic, while the other side is
essentially saturated. The opposite is seen in the n-6 DPA, where the side of the chain
closest to the methyl group is saturated, and the other side unsaturated. The DPAs lack the
full methylene interupted sequence of double bonds throughout the carbon chain, which
could be the basis of why they apparently do not have the functionality required by retinal
and brain tissue.
The energy minimized ethanolaminephosphoglyceride structures with DHA (Fig. 4)
and n-3 DPA (Fig. 5) dramatically illustrated the significance of the missing double bond in
DPA vs. DHA. The final (C19) double bond in DHA constrains the position of the
phosphoethanolamine head group, pulling it in and maintaining the spiral structure. This
reduces the molecular volume, and may facilitate communication between the head group
and the esterfied lipid chains. In contrast, the head group in n-3 DPA is far less
constrained, and in fact moves away from the lipid ester chains. This structure would be
more typical of phospholipids in general since most FA are less polyunsaturated than
DHA. The cell membrane is in constant fluid motion so these structures only represent the
preferred orientations of the molecules.x
Figure 1: 3D energy-minimized structure of docosahexaenoic acid (DHA). This and
following figures energy minimized and drawn with MOPAC as described in text.

13
Figure 2: N-3 docosapentaenoic acid (n-3 DPA).
Figure 3: N-6 docosapentaenoic acid (n-6 DPA).

14
a b
Figure 4: a: 3D energy-minimized structure of phospholipid with ethanolamine,
DHA, DHA. Side view. b: Ethanolamine, DHA, DHA. End view, note position of
phosphate group.

15
a
b
Figure 5a: a: 3D energy-minimized structure of phospholipid with ethanolamine, n-3
DPA, n-3 DPA. Side view. b: Ethanolamine, n-3 DPA, n-3 DPA. End view, note
position of phosphate group.

16
V CONCLUSION
There is much still to be learned about the physical properties of membranes containing
DHA. The extremely high degree of specificity with which it is selected for membranes of
the brain (and has been, since very early evolutionary times) cannot be explained on the
basis of the conventional measurements that have been made thus far. DHA’s special role
may relate either to undefined interactions with integral membrane proteins or, more
speculatively, to some role in neuronal signalling arising from unusual electrical properties.
Nature’s sharp discrimination between DHA and the nearly identical DPA species may give
guidance to further inquiries into DHA’s putative role, by focusing attention on the
importance of the full complement of six periodic double bonds.
Acknowledgements:
We are grateful for financial support from the Mother and Child Foundation and Martek
Biosciences especially for travel expenses for meetings to finalise this paper.

17
REFERENCES
1. Cr awf or d MA, Cu nna ne SC a nd Har bi ge L S (19 93) A n ew theo ry o f evo lu ti on : qu ant um
t h eo r y . II I r d In te r n a t io n a l Co n g re s s o n es s en ti a l fa t t y ac id s an d e ic o s an o i d s , A m . Oi l C h e m .
S o c. e d A. J . S i n c la ir , R. G i bs o n , Ad e l a id e, 87 - 9 5 .
2. C ra wfo rd MA. a nd Si nc l ai r AJ (1 97 2) Nut r it io na l in fl ue nce s in t he e vo lu ti on o f th e
ma mmal ia n br ai n. I n Li pi ds, m al nut ri t io n and t he d ev el opi ng b ra in: 267 -29 2. El l io t, K. and
Kni gh t, J. (Eds .) . A Cib a Fou nd at io n Symp osi um (19-21 O ct obe r, 1 971 ). A mst er da m,
E l s e v i er .
3. FAO/WHO (1978) The role of dietary fats and oils in humane nutrition. Nutrition
report no 3, FAO, Rome
4. Martin, R D (1983). Human Brain Evolution in an Ecological Context. Fifty-second
James Arthur Lecture on the Evolution of the Human Brain. New York: American
Museum of Natural History.
5. Cra wfo rd MA, Gal e MM a nd Wo odf or d MH ( 19 69) L in ol ei c ac id a nd lino le ni c ac id
el on ga ti on pr od uct s i n mus cl e ti ss ue o f Syn ce rus c af fe r an d ot her rum in ant spe ci es . Bi och em.
J. 115 : 25 -27 .
6 . W i ll i a m s G an d C r a w fo r d MA (1 9 87 ) Co m p ar i s o n o f th e fa tt y ac id c o m p o n e n t i n
st r uct ur al l i pi ds from d ol phi ns , ze br a and g ir af fe : po ss ibl e e vol ut io na ry impl ic at io ns . J. Z ool .
Lo nd. 2 13 : 673 - 684 .
7. Conroy GC, Weber GW, Seidler H, Tobias PV, Kane A, Brunsden B (1998).
Endocranial capacity in an early hominid cranium from Sterkfontein, South Africa.
Science 280 1730-1731.
8. Sikes NE (1994). Early hominid habitat preferences in East Africa: paleosol carbon
isotopic evidence. Journal of Human Evolution 27, 25-45.
9. Broadhurst CL, Cunnane SC & Crawford MA (1998). Rift Valley lake fish and
shellfish provided brain specific nutrition for early Homo. British Journal of Nutrition 79,
3-21.
10. Rightmire GP and Deacon HJ (1991). Comparative studies of Late Pleistocene human
remains from Klasies River Mouth, South Africa. Journal of Human Evolution 20, 131-
156.
11. Churchill SE, Pearson OM, Grine FE, Trinkaus E, Holliday TW (1996).
Morphological affinities of the proximal ulna from Klasies River main site: archaic or
modern? Journal of Human Evolution 31, 213-237.
12. Pfeiffer S and Zehr MK (1996). A morphological and histological study of the human
humerus from Border Cave. Journal of Human Evolution 31,49-59.

18
13. Brauer G and Singer R (1996). The Klasies zygomatic bone: archaic or modern?
Journal of Human Evolution 30, 161-165.
14. Shreeve, J. (1995). The Neandertal Enigma: Solving the Mystery of Modern Human
Origins. New York: William Morrow.
15. Parkington JE (1998). The impact of the systematic exploitation of marine foods on
human evolution. Colloquia of the Dual Congress of the International Association of the
Study of Human Paleontology and the International Study of Human Biologists, Sun City,
South Africa, 28 June to 4 July, 1998, in press.
16. Stringer CB (1992). Reconstructing recent human evolution. Philosophical
Transactions of the Royal Society, London B 337, 217-241.
17. Foley RA (1994) Speciation, extinction, and climatic change in hominid evolution.
Journal of Human Evolution 26, 275-289.
18. Tishkoff SA, Dietzsch E, Speed E, Pakstis J, Kidd JR, Cheung K et al. (15 authors)
(1996). Global patterns of linkage disequilibrium at the CD4 locus and modern human
origins. Science 271, 1380-1387.
19. Stringer CB and McKie R (1997). African Exodus. New York: Henry Holt and
Company.
20. Morwood, MJ, O’Sullivan, PB, Aziz F and Raza, A. (1998). Fission track ages of the
stone tools and fossils on the East Indonesian island of Flores. Nature 392, 173-176.
21. Chen Z-Y, Menard CR and Cunnane SC (1995). Moderate selective depletion of
linoleate and a-linoleate in weight-cycled rats. American Journal of Physiology 268,
R498-R505.
22. Cunnane SC and Anderson MJ (1997). The majority of dietary linoleate in growing
rats is b-oxidized or stored in visceral fat. Journal of Nutrition 127, 146-152.
23. Farquharson J, Jamienson EC, Abbasi KA, Patrick WJA, Logan RW and Cockburn F
(1995) Effect of diet on the fatty acid composition of the major phospholipids of infant
cerebral cortex. Arch Dis Childhood, 72: 198-203,
24. Willats P, Forsyth JS, DiMudongo MK, Varmia S and Colvin M (1998) Effect of
long chain polyunsaturated fatty acids in infant formula on problem solving at 10 monthgs
of age. Lancet 352: 688-691.
25. Ri ve rs J PW, Si nc la ir A J and C ra wfo rd MA (1 975 ). " In ab il it y of t he c at to de sa tu ra te
es se nt ia l fa t ty a ci ds. " Na tur e 28 5: 1 71-17 3.
2 6 . C ra w f o r d M. A , Gh e b r e m e s k el K an d Ph y l a ct o s AC ( 1 99 7 ) A r e de f i ci t s of ar a c h i d o n i c
an d doc os ah exa en oi c aci ds r es po nsi bl e fo r the ne ur al a nd va sc ul ar c omp li ca ti on s of pr et er m
b a b i es ? Am J. C l i n . N u t r . 66 : 1 032 S-1 04 1S.
27. Keys A (1997) Coronary heart disease in seven countries. 1970. Nutrition 13:250-252

19
28. Barker DJ and Clark PM (1997) Fetal undernutrition and disease in later life. Rev
Reprod 1997 May;2(2):105-112
29. Albert CM, MD, Hennekens CH, O'Donnell CJ, Ajani A, Carey VJ, Willett WC,
Ruskin JN and Manson JE, (1998) Fish Consumption and Risk of Sudden Cardiac
Death JAMA. 1998;279:23-28
30. Ha rr is WS (198 9) Fi sh o il s an d pl as ma li pi d an d li po pro te in m et abo li sm i n hum ans , a
cr i ti ca l re vi ew. J. Li pid R es . 30: 7 85- 807 .
31. Weylandt KH, Kang JX and Leaf A (1996) Polyunsaturated fatty acids exert
antiarrhythmic actions as free acids rather than in phospholipids. Lipids 1996
Sep;31(9):977-982
32. Pauletto P, Puato M, Caroli MG, Casiglia E, Munhambo AE, Cazzolato G, Bittolo
Bon G, Angeli MT, Galli C, and Pessina A. (1996) Blood pressure and atherogenic
lipoprotein profiles of fish-diet and vegetarian villagers in Tanzania : the Lugalawa Study.
Lancet, 348:784-88.
33. Vigilant L, Stonemaking M, Harpending H, Hawkes K, and Wilson AC (1991)
African populations and the evolution of human mitochondrial DNA, Science, 253, 1503-
1507.
34. Dyerberg J (1994) The epidemiology of omega 3 fatty acids World Rev Nutr Diet
76:133-136
35. Ol se n FS an d Sec he r NJ (1 990 ) A pos si bl e pr ev ent i ve ef fe ct of low-do se f is h oi l on e ar ly
d e li v e r y a n d pr e - e c la m p s ia : in d i c a ti o n s fr o m a 50 y ea r ol d co n t r o l le d tr i al . B. J . N u t r . , 64 :
59 9-60 9.
36. Cr awf or d M.A, Hass am AG, Wi l li ams G a nd Whi te ho use WL ( 19 76) E ss ent i al f at ty
ac id s an d fe ta l br ai n gr owt h. La nc et ( i) : 4 52-45 3
37. Kou kk ou E, Bi t sa ni s D, Ghe br eme ske l K, L owy C, P ost on L a nd Cr awf or d M.A (1 99 7)
B o t h d i a b e t es a n d m a t e r n a l d ie t s r i c h in s a t u r a t ed f at t y a c i d s a l t er s fe ta l l iv er l ip i d c omp osi t io n
an d va s cu l a r re a c ti v i t y. I n Pr o ce e d i n g s of t h e I Vt h In t e r n a t i o n a l C o n g re s s on E s s en t i a l F a t t y
Ac id s an d t h e i r Ei c o s a n o id s . J. P r o s t a g l a n d i n s , Leu k o t r i e n e s an d E F As , , 57 : 268
38. Rodriguez de Turco EB, Deretic D, Bazan NG and Papermaster DS (1997)
Post-Golgi vesicles cotransport docosahexaenoyl-phospholipids and rhodopsin during frog
photoreceptor membrane biogenesis. J Biol Chem 1997 Apr 18;272(16):10491-10497
39. Su zuk i, H., Ma na be, S ., Wa da O an d Cra wf ord , MA. , (1 997 ) Rap id i nc or por at i on of
d o c o s ah ex aen o i c ac i d f r o m d i e ta ry s o u r ce s in t o b r a i n m i c r o s o m a l, s y n ap t os o m a l an d
m i to c h o n d r i al m e m b r an es i n ad u l t mic e. I n te r n a t . J. Vit . R e s . , 67 : 272 -2 78
40. Mouritsen OG. and Bloom M. (1993) Models of Lipid-Protein Interactions in
Membranes. Ann. Rev. Biophys. Biomol. Struct. 22, 145-171.

20
41. Ghe br eme sk el K, D ubo wit z L, Th oma s, B. , Mi n Y, Gol f et to I , Low y, C, a nd Cr awf or d
M. A. (1 998 ) Relationship between arachidonic and docosahexaenoic acid in plasma and
red cell membranes, submitted for publication..
42. Bloom M, Linseisen F, Lloyd-Smith J and Crawford MA (1998) "Insights from NMR
on the Functional Role of Polyunsaturated Lipids in the Brain" Enrico Fermi International
School of Physics: "Magnetic Resonance and Brain Function - Approaches from Physics"
in Varenna, Italy.
43. Holte LL, Peter SA, Sinnwell TM and Gawrisch K. (1995). 2H Nuclear FMagnetic
Resonance Order Parameter Profiles Suggest a Change of Molecular Shape for
Phosphatidylcholines Containing a Polyunsaturated Acyl Chain. Biophys. J. 68, 2396-
2403.
44. Ehringer W, Belcher D, Wassall SR and Stillwell W (1990). A comparison of the
effects of linoleic (18:3Ω3) and docosahexaenoic (22:6Ω3) acids on phospholipid
bilayers. Chem. Phys. Lipids, 54: 79-88
45. Brown MF (1994). Modulation of rhodopsin function by properties of the membrane
bilayer. Chem. Phys. Lipids 73, 159-180.
46. Penrose R (1997) The large, the Small and the Human Mind, Cambridge University
Press.
47. Harvey, PH and Clutton-Brock TH (1985). Life history variation in
primates. Evolution 39, 557-581.
48 . B ro wn, RE , Sha ff er R .D, H ans en I L, Ha nse n HB an d Cra wfo rd MA (196 6) He al th
s u r v ey of th e El Mo l o . E. Af r . M e d . J . 43 : 48 0 -4 8 8 .
4 9 . C ra w f o r d , M. A , R i v e r s , J . P . W . , an d H a s s a m , A . G . ( 1 9 7 8) E s s en t i a l f at t y a c i d s a n d t h e
vu ln er abi l it y of t he a rt er y du ri ng gr owt h. P ost Gra d. Me d. J . 54: 149 -15 3.
50. Spr ec her H ( 19 93) In te rc ov er si ons b et wee n 20- and 2 2- c ar bon n -3 an d n- 6 fa tt y ac id s
vi a 4-d es at ur as e in dep end an t pa th way s. II Ir d In t. C ong re ss on e ss ent i al f at t y a ci ds a nd
ei c o s a n o i d s . Am . O i l C h e m . S o c . Ed . A. J S i n c l ai r , R . Gi b s o n , Ad e l ai d e , 18 - 22 .
51. Si nc la ir A J (1 975 ) Lon g ch ai n pol yu nsa tu ra te d FAs i n th e mamm al ia n br ai n. Pr oc Nu tr
So c. 3 4: 28 7- 291 .
52. Sponheimer M., Lee-Thorpe JA (1999) Isotopic evidence for the diet of an early
hominind. Science: 283: 368-370.