Start trading with as little as 1$ and open your account free, get $5 reward

Orthodox and Recalcitrant Seeds


Orthodox and
Recalcitrant Seeds

P A T R I C I A B E R J A K and
Plant Cell Biology Research Unit,
School of Life Sciences
University of Natal, Durban
4041 South Africa

The seeds of many species cannot be classified as ortho-
dox, and this is particularly so in the case of tropical tree seeds.
The views presented here favor a continuum of seed behavior
that is based on a variety of characteristics. A suite of mecha-
nisms or processes is discussed that embodies the properties
currently thought to promote the acquisition of desiccation
tolerance and to ensure survival of the desiccated condition in
orthodox seeds. These include: cellular and intracellular phys-
ical characteristics; intracellular de-differentiation; the
“switching off” of metabolism; the presence and efficient
operation of antioxidant systems; the accumulation and roles
of putatively protective molecules, including late embryogenic
accumulating/abundant proteins (LEA’s), sucrose, and certain
oligosaccharides; deployment of amphipathic molecules; an
effective peripheral oleosin layer around lipid bodies; the
occurrence and operation of repair mechanisms during rehy-
dration; and others yet to be identified. The presence of some
of the mechanisms/processes, or their absence or partial
expression, is considered in the context of the varied respons-
es to dehydration shown by nonorthodox seeds. The factors
that determine distinct variations in the behavior of recalci-
trant seeds of individual species under the same conditions is
given attention, with the effects of drying rate (i.e. the rate of
water loss from tissues of desiccation-sensitive seeds) being
stressed. Two different factors are distinguished in this regard:

(1) damage that occurs at low water contents when nonfreez-
able water, which is held to stabilize intracellular structures
and macromolecules, is removed, which is desiccation damage
in the strict sense; and (2) damage that occurs during slow
dehydration, when metabolic imbalances are proposed to
cause the generation of damaging chemical species, e.g. free
radicals, which is termed metabolic damage. Desiccation dam-
age, in the strict sense, is attributed to the lack or inadequate
operation of the processes/mechanisms held to protect desic-
cation-tolerant seeds in the dry state, while metabolic damage
is considered in the context that nonorthodox seeds (especial-
ly those that are truly recalcitrant) do not possess the suite (or
full suite) of mechanisms/processes that facilitate the acquisi-
tion and maintenance of desiccation-tolerance as exhibited by
maturing and mature orthodox seed-types.


Orthodox seeds (Roberts 1973) acquire desiccation tolerance
during development and may be stored in the dry state for pre-
dictable periods under defined conditions. Unless debilitated
by zero-tolerant storage fungi, orthodox seeds should maintain

Chapter 4: Orthodox and Recalcitrant Seeds

high vigor and viability at least from harvest until the next
growing season (Berjak and others 1989) or for many decades
at -18 °C (IBPGR 1976). Generally, such seeds undergo a peri-
od of drying during their maturation and are shed at low water
content which is in equilibrium with the prevailing relative
humidity (r.h.). The equilibrium water content at any particu-
lar r.h. is determined by seed composition, but all orthodox
seeds can withstand dehydration to around 5 percent (0.053 g
H2O g-1 dry material [g g-1 ]), even when maturation drying is
not completed prior to shedding. Any seed that does not
behave this way is not orthodox, and, in fact, the seeds of a
great number of tropical species may accordingly be nonortho-
dox. Nonorthodox seeds have so far been described as either
being recalcitrant (Roberts 1973) or intermediate (Ellis and
others 1990a) according to their storage behavior.
Recalcitrant seeds are those that undergo little, or no,
maturation drying and remain desiccation sensitive both during
development and after they are shed. The situation is, howev-
er, far more complex than this because of the wide range of
variability among recalcitrant seeds of different species and,
indeed, of individual species under different conditions (Ber-
jak and Pammenter 1997). Such seeds are shed hydrated, but
the water content can generally be anywhere in the range from
0.43 to 4.0 g g-1, which is 30 to 80 percent on a wet mass basis
(wmb). Shedding water content is partly species characteristic,
depending on the degree of dehydration that occurs late dur-
ing seed development; this has been suggested to be correlat-
ed with the degree of desiccation tolerance developed by indi-
vidual species (Finch-Savage 1996).
Recalcitrant seeds are not equally desiccation sensitive,
in that variable degrees of dehydration are tolerated depend-
ing on the species. This implies that the processes or mecha-
nisms (see below) that confer desiccation tolerance are vari-
ably developed or expressed in the nonorthodox condition. As
diverse mechanisms have been suggested to be involved in the
acquisition of desiccation tolerance and maintenance of the
integrity of dehydrated orthodox seeds, it should be appreci-
ated that any one of these may be absent, or present but inef-
fective, in recalcitrant seeds. Another important consideration
is that desiccation tolerance is probably controlled by the inter-
play of mechanisms or processes, and not by any one, acting in
isolation. Thus, the absence or incomplete expression of any fac-
tor proposed to confer dehydration tolerance could have pro-
found consequences on the ability of a seed species to withstand
a measure of dehydration below a particular level of hydration.
Differential desiccation sensitivity among recalcitrant
seeds of various species is clearly shown by their different
responses when subjected to the same drying regime–those of
some species tolerating only a slight degree of dehydration,
but others surviving to far lower water contents. There are also

marked differences in the rates at which water will be lost from
seeds of various species under the same dehydrating condi-
tions (Farrant and others 1989). Other factors, too, influence
the postharvest responses of recalcitrant seeds, e.g. develop-
mental status (Berjak and Pammenter 1997, Berjak and others
1992, Berjak and others 1993, Finch-Savage 1996, Finch-Sav-
age and Blake 1994) and chilling sensitivity (Berjak and Pam-
menter 1997).
In terms of desiccation sensitivity alone, therefore, it is
not merely that a seed species is recalcitrant, but rather, how
recalcitrant it is. This fact led to the proposal of a continuum of
recalcitrant seed behavior, from species that are highly desic-
cation–and probably also chilling–sensitive, to those that will
tolerate drying to the lowest water content still commensurate
with recalcitrant seed behavior and will also tolerate relatively
low temperatures (Farrant and others 1988).
The concept of a continuum of postharvest seed behav-
ior (that is, dependent on preshedding developmental events)
extends beyond the category of recalcitrant seeds. The contin-
uum grades from extreme desiccation-sensitive types through
the minimally recalcitrant types, to the intermediate seed
species that do not react adversely to low temperatures,
through those that are chilling sensitive when dehydrated
(Hong and Ellis 1996), and finally to orthodox seeds that will
tolerate less or more extreme dehydration (Vertucci and Roos
1990). It is also possible that there are seed species that behave
in a manner that characterizes them as lying between the hith-
erto-defined categories: recalcitrant, intermediate, and ortho-
dox. The idea of an extended continuum of seed behavior
from the most desiccation tolerant of orthodox species, to the
recalcitrant species that are most sensitive to even slight water
loss, embodies many properties of seeds and their responses
(Berjak and Pammenter 1994, 1997). It has its foundations in
an appreciation of the physiological status of seeds at various
water potentials (Vertucci 1993, Vertucci and Farrant 1995,
Vertucci and Roos 1990) and the properties of water at the var-
ious hydration levels corresponding to specified water poten-
tial ranges (Vertucci 1993, Vertucci and Farrant 1995). It is
more meaningful to consider seed responses to dehydration in
terms of water potential rather than water content, but, since
these two measures can be loosely correlated (Vertucci and
Farrant 1995), the more familiar water content terminology is
used here.
According to Vertucci and Farrant (1995): “Discrete
changes in metabolic activity with moisture content are hypoth-
esized to be associated with discrete changes in the physical
properties of water… Thus upon the loss of water with certain
properties, an essential function provided by [that] water is no
longer possible. A tissue that is not damaged by the removal of
a certain type of water has developed mechanisms to tolerate


Part I—Technical Chapters

or avoid that particular stress.” While the discussion that fol-
lows is not dependent on the reader’s appreciation of the dif-
ferences in the types of intracellular water, the basis of the
arguments presented is that sequential removal of water with
specific properties will have particular damaging effects on
seed tissues that are not possessed of the appropriate mecha-
nisms or processes to counteract that damage. We will, how-

additional properties that contribute to the ability of seeds to
withstand extreme dehydration are likely to be elucidated.


Vacuolation and Reserve Deposition

ever, focus on the mechanisms or processes themselves.
In 1957, Iljin had already identified one of the major require-
ments of cells of desiccation-tolerant plant material: the abili-
ty to withstand mechanical stress. Vacuole volume reduction,


It is most expedient to consider the processes or mechanisms
listed below, which might confer protection against desicca-
tion, and their deficiency or absence, which could contribute
to the relative degrees of desiccation sensitivity.

• Intracellular physical characteristics such as
— reduction of the degree of vacuolation,
— amount and nature of insoluble reserves
— integrity of the cytoskeleton,
— conformation of the DNA, chromatin, and
nuclear architecture.
• Intracellular de-differentiation, which effectively results in
the minimization of surface areas of membranes and proba-
bly also of the cytoskeleton.
• “‘Switching off” of metabolism.
• Presence, and efficient operation of, antioxidant
• Accumulation and roles of putatively protective molecules,
including late embryogenic accumulating/abundant proteins
(LEA’s), sucrose and certain oligosaccharides, or galactosyl
• Deployment of certain amphipathic molecules.
• An effective peripheral oleosin layer around lipid bodies.
• The presence and operation of repair mechanisms during

In the discussion that follows comparisons are made, as
far as is possible, between desiccation-sensitive and orthodox
seeds, and of the status of the processes or mechanisms that
have been suggested to contribute to desiccation tolerance.
Although the interrelationships among them are far from being
resolved, these processes or mechanisms are those that have
been implicated to date in the acquisition and maintenance of
desiccation tolerance. However, it is important to realize that

whether by the shrinkage of the space occupied by these usu-
ally fluid-filled organelles or by their becoming filled with
insoluble reserve material, is one of the mechanisms that
would contribute to increased mechanical resilience of cells to
dehydration. This aspect was examined by Farrant and others
(1997) for (1) Avicennia marina, the highly recalcitrant seeds
of which can withstand very little dehydration either before or
after they are shed; (2) Aesculus hippocastanum, a temperate
recalcitrant species, the seeds of which overwinter in the hydrat-
ed condition during which the necessary stratification occurs
to facilitate germination the following spring; and (3) Phaseo-
lus vulgaris, a typical orthodox seed that attains a low water
content prior to shedding and is long-lived in this condition.
Avicennia marina seeds lose no water during develop-
ment, and are as sensitive to dehydration before shedding as
after abscission (Farrant and others 1992b). These seeds, at
best, are unable to surive water contents lower than 0.5 g g-1
(33 percent wmb). The vacuoles ultimately occupy almost 60
percent on average of the volume across the cells of all axis tis-
sues, and 90 percent of the cotyledonary cells when mature. At
no stage do either the axial or cotyledonary vacuoles contain
insoluble reserves, the little insoluble reserve material occur-
ring as plastid starch. Seeds of A. hippocastanum naturally
undergo a measure of dehydration during development,
accompanied by an increase in relative desiccation tolerance
(Tompsett and Pritchard 1993). The mature seeds are more
desiccation tolerant than those of A. marina, being able to
withstand dehydration to water contents in the range of 0.42
to 0.25 g g -1 (30 to 20 percent wmb). Vacuoles ultimately con-
stitute only a small fraction of the intracellular volume, partic-
ularly in the axis cells at maturity. The cotyledonary cells con-
tain many large, starch-filled plastids and protein bodies and
are considerably less vacuolate than those of A. marina. In P.
vulgaris seeds, which are orthodox and able to tolerate low
water contents, vacuolar volume is reduced to an insignificant
proportion in axis cells, and vacuoles in cotyledonary cells
accumulate an amorphous, presumably insoluble, material.
The differential degree of vacuolation and insoluble reserve
deposition among the three species, in both developing and

Chapter 4: Orthodox and Recalcitrant Seeds

mature seeds, correlates with their degree of desiccation sensi-
tivity. This is in accord with the concept that a high degree of
vacuolation can lead to lethal mechanical damage upon dehy-
dration (Farrant and others 1997).

Reaction of the Cytoskeleton

The cytoskeleton, the major components of which are micro-
tubules and microfilaments, is not only an integrated intracel-
lular support system, it also plays a major role in imposing
organization on the cytoplasm and also the nucleus. Micro-
tubules consist of polymerized a -tubulin, and microfilaments
are composed of F-actin, which is a polymer of G-actin. We
are presently investigating the status of the actin microfila-
ments in hydrated and variously dehydrated embryonic axes of
seeds of Quercus robur, a temperate recalcitrant species. In the
hydrated state, there is an extensive microfilamentous network
in the cells of the root tip, which becomes dismantled as the
seeds are increasingly dehydrated–a feature that is expected
for orthodox seeds as well. In such desiccation-tolerant seeds,
orderly reassembly of the elements of the cytoskeleton accom-
panies imbibition, but once the water content falls to dam-
agingly low levels in Q. robur, the microfilaments are not
reassembled when the seeds are subsequently rehydrated
(Mycock and others 2000). The resultant lack of the intracel-
lular support and structural organization afforded by the
cytoskeleton would obviously be a major damaging factor upon
rehydration of recalcitrant seeds. Additionally, certain cytoma-
trical (cytoplasmic) enzyme systems exist as multienzyme par-
ticles in plant cells (Hrazdina and Jensen 1992), the formation
of which could occur because of the binding of key or anchor
enzymes to the microfilaments of the cytoskeleton, as illustrat-
ed for glycolysis by Masters (1984). Thus, failure of the
cytoskeleton to reassemble following deleteriously low levels
of dehydration would have physiological as well as structural
consequences in the cells of desiccation-sensitive seed tissues.

DNA, Chromatin, and
Nuclear Architecture Conformation

Maintenance of the integrity of the genetic DNA material in
the desiccated condition in orthodox seeds, and/or its rapid
repair when seeds are rehydrated, is considered to be a funda-
mental requirement for desiccation tolerance. There is, how-
ever, little information on which to draw. DNA assumes dif-
ferent conformational states depending on water activity and,
although this has not yet been demonstrated for seeds, it is
considered that as water is lost (i.e. water activity is lowered)
such conformational changes will occur (Osborne and Boubri-
ak 1994). According to information reviewed by those authors,

there is an increase in the number of base pairs per turn of the
DNA helix as water is lost from the individually hydrated
phosphate groups, and water bridges are formed instead as the
conformation changes from the B to the Z form. Osborne and
Boubriak (1994) have suggested that protein glycation (i.e. the
nonenzymic addition of reducing sugars to [i.a. ] histone pro-
teins) is likely to occur, which could increase the incidence of
DNA conformations appropriate to the dehydrated state.
Those authors also discuss the possibility of nonenzymic
methylation of cytosine occurring, which would favor the Z-
form of the DNA.
However, besides the postulated necessity of conforma-
tional changes in the DNA occurring as desiccation-tolerant
material is dehydrated, the structure of the chromatin itself
must also be stabilized. The highly condensed state of the
chromatin in dry, orthodox seeds (e.g. Crévecoeur and others
1976, Sargent and others 1981), which is reversed at the stage
in germination when desiccation sensitivity ensues (Deltour
1985), is thought to be a visible manifestation of its stabilized
condition. A major factor in chromatin stabilization in the dry
state in orthodox seeds might be the change in the H1-his-
tone:nucleosome ratio to 2:1 from the 1:1 ratio that typifies the
hydrated condition (Ivanov and Zlatanova 1989).
Nuclear architecture is a further factor that is probably
involved in chromatin stability. The structural framework of
the nucleus has been convincingly demonstrated for plant cells
and is based on intermediate-type filaments called lamins
(Moreno Díaz de la Espina 1995). The nucleoskeleton, organ-
ized into the lamina (underlying and connected to the inner
surface of the nuclear envelope) and matrix (ramifying through-
out the nucleus) is suggested to support and localize the chro-
matin in discrete domains, imposing the topological organiza-
tion and coordination of intranuclear processes (Moreno Díaz
de la Espina 1995). It is implicit that during dehydration and
in the desiccated state of orthodox seeds, orderly reorganiza-
tion of the nucleoskeleton should occur with its restitution as
a functional framework upon rehydration.
While little is known about the effects of dehydration on
the DNA, chromatin, and nuclear architecture in desiccation-
sensitive seeds, their stability in the dehydrated state clearly
must be a prerequisite for desiccation tolerance. Maintenance
of the integrity of the nucleus as a whole, and the genome in
particular, may be imperfectly expressed, or the ability for this
may even be totally lacking, in recalcitrant seeds. (For a fuller
account of some of these aspects, see Leprince and others
(1995) and Pammenter and Berjak (1999)). What is equally
likely is that DNA repair mechanisms themselves are inade-
quate to restitute damage caused by dehydration of desicca-
tion-sensitive seeds (see below).


Part I—Technical Chapters


De-differentiation, a characteristic of maturing desiccation-
tolerant seeds, is essentially a means by which intracellular
structures are simplified and minimized (reviewed by Vertuc-
ci and Farrant 1995), which strongly suggests that membranes
and cytoskeletal elements are vulnerable to dehydration. This
phenomenon is reversed in orthodox seeds when water is
taken up early during germination (Bewley 1979, Dasgupta
and others 1982, Galau and others 1991, Klein and Pollock
1968, Long and others 1981).
An examination of the quantitative and qualitative status
of mitochondria in seeds of Avicennia marina, Aesculus hip-
pocastanum, and Phaseolus vulgaris showed that the propor-
tion of cell volume occupied by these organelles was highest in
A. marina, which is very desiccation sensitive, and substantial-
ly less for A. hippocastanum, which is in keeping with its less
recalcitrant nature. In P. vulgaris, mitochondria occupied a sig-
nificantly smaller proportion of the cell volume, even preced-
ing the onset of maturation drying (Farrant and others 1997).
Also, the mitochondria occupied a far greater proportion of
the cell volume in the axis meristems of the two recalcitrant
species than in the orthodox species, P. vulgaris. There were
also marked differences in the structural complexity of the
mitochondria among these three species: A. marina and A. hip-
pocastanum, had well-developed cristae and a structure that
was generally typical of an active, hydrated plant tissue; while
in P. vulgaris, the mitochondria were almost completely de-dif-
ferentiated even at tissue water contents comparable to those
of the recalcitrant species at shedding (Farrant and others
1997). It thus seems that retention of organelles in the highly
differentiated state is a major factor in the desiccation sensi-
tivity of recalcitrant species, whereas the ability for ordered
de-differentiation is, in fact, a prerequisite for seed survival in
the dehydrated state.
There has long been uncertainty as to whether dehydra-
tion causes de-differentiation, or this intracellular minimiza-
tion actually precedes the initiation of maturation drying (e.g.
Vertucci and Farrant 1995). However, the observations on P.
vulgaris reported by Farrant and others (1997), indicating that
mitochondrial de-differentiation occurs, and that respiratory
rate declines markedly (see also below) before maturation dry-
ing, support the idea that substantial qualitative and quantita-
tive change actually occurs in advance of water loss.


Electron transport, albeit at a low level, has been recorded for
dehydrated plant tissues, and respiration is measurable even at

seed water contents as low as 0.25 g g-1 [20 percent, wmb]
(Vertucci 1989, Vertucci and Farrant 1995). However, in the
water content range 0.45 to 0.25 g g (30 to 20 percent
[wmb]), unbalanced metabolism may lead to the generation,
and essentially uncontrolled activity, of free radicals (Finch-
Savage and others 1994a, Hendry 1993, Hendry and others
1992, Leprince and others 1990b, Vertucci and Farrant 1995).
It is therefore imperative that, during maturation drying, des-
iccation-tolerant seeds be able to pass through this water con-
tent range with the minimum of damage. The efficient opera-
tion of antioxidant systems (Leprince and others 1993, Pun-
tarulo and others 1991), as well as the “switching off” of
metabolism, would reduce such damage. Rogerson and
Matthews (1977) recorded that a sharp decline in respiratory
substrates precedes, and presumably causes, the fall in respi-
ratory rate which, they suggested, is an essential event enabling
an orthodox seed to withstand rapid loss of water. The obser-
vations of Farrant and others (1997), indicating that a decline
in respiratory rate occurs while mitochondria become sub-
stantially de-differentiated prior to maturation drying in the
orthodox seeds of Phaseolus vulgaris, support the data and
suggestions of Rogerson and Matthews (1977).
In desiccation-sensitive seeds, lethal damage occurs in
the water content range 0.45 to 0.25 g g1 (Vertucci and Farrant
1995) and, in some species, at considerably higher levels
(Pammenter and others 1993). Death of relatively hydrated
recalcitrant seeds (at c. 0.7 g g-1 , or higher [ 40 percent, wmb])
occurs when water is lost slowly. However, rapid dehydration
rates allow survival to lower water contents (Farrant and oth-
ers 1985). This observation led initially to the use of relatively
rapid air-drying of excised embryonic axes to facilitate cryos-
torage (Normah and others 1986, Pritchard and Prendergast
1986) and later to the development of the flash-drying tech-
nique (Berjak and others 1990), by which the axes are dehy-
drated much more rapidly.
Flash-dried axes are not desiccation tolerant; on the con-
trary, they will not survive for longer than a day or two at best,
under ambient conditions (Walters and others 2001) although
they may be cryostored successfully (Wesley-Smith and others
1992). The desiccation sensitivity of recalcitrant material is the
outcome of the fact that the axes (seeds) are actively metabol-
ic, and the success of very rapid dehydration is that it mini-
mizes the effects of this metabolism. This important point
about drying rate is discussed in detail later.
Damage occurring in conjunction with unbalanced
metabolism at these relatively high water contents should not
be confused with desiccation damage in the strict sense. The
latter describes the damage that occurs when water that is
required to maintain the integrity of intracellular structures is
removed (Walters and others 2001). Desiccation damage sensu

Chapter 4: Orthodox and Recalcitrant Seeds

stricto is the consequence of removing (any, or some, depend-
ing on the species) structure-bound, nonfreezable water (Pam-
menter and others 1991, Walters and others 2001). Lethal
damage occurs upon loss of this water, even if flash-drying has
successfully maintained axis viability to, or close to, this level
of hydration (Pammenter and others 1991).
Another critical aspect of ongoing metabolism is cell
cycling. The cell cycle describes the nuclear DNA content as
2C in cells that are not preparing for nuclear division, and as
4C in cells in which DNA replication has occurred, where the
constant, C, denotes the DNA content of the haploid condi-
tion. During the cell cycle four distinct phases can be identi-
fied, viz. the G1 phase (2C), which is followed by the S phase,
during which DNA replication occurs; after this the cells enter
the G2 phase, during which the amount of DNA remains dou-
bled (i.e. 4C) as a result of events in the S phase, and this is fol-
lowed by the phase known as G2M, when mitosis reduces the
DNA content to the 2C level typical of somatic cells in the
next G1 phase. Brunori (1967) found that in orthodox Vicia
faba seeds, most of the cells were arrested in G1, and that
DNA replication was one of the first events to be curtailed as
the embryo cells lost water. S-phase replication is resumed
only after several hours of imbibition, when water again
becomes available to postharvest, orthodox seeds, as shown by
Sen and Osborne (1974) for Secale cereale (rye): as soon as
replication to 4C values occurs and the cells enter G2M, des-
iccation tolerance is lost.
In the the highly recalcitrant seeds of Avicennia marina,
there is only the most transient arrest of DNA replication in
root primordia (meristems) of Avicennia marina lasting no
more than 24 hours around shedding. This is the time when
the seeds (although highly desiccation sensitive) are relatively
most tolerant of water loss and least active. Ongoing cell
cycling is associated with marked desiccation sensitivity of the
DNA. When only 16 to 18 percent of the total water is lost
from the A. marina material, there is a reduction of 70 to 80
percent in the nuclei that will incorporate thymidine, and after
a 22-percent water loss, damage of the DNA cannot be
repaired even when water is made freely available. Ongoing
cell cycling, therefore, is another manifestation of the fact that
metabolism is not “switched off,” at least in these highly recal-
citrant seeds, which is considered to be a major factor account-
ing for their desiccation sensitivity. In related work on the tem-
perate recalcitrant species Acer pseudoplatanus, however, cell
cycling was found to be arrested, with over 60 percent of the
cells in the 2C state (Finch-Savage and others 1998). Howev-
er, seeds of A. marina are poised for immediate germination,
while those of A. pseudoplatanus are dormant, requiring cold
stratification before they will germinate. For seeds of
Azadirachta indica, recorded as showing intermediate behav-

ior, the 2C DNA level has been reported as occurring to the
virtual exclusion of 4C (Sacandé and others 1997). These dis-
parate results on the status of the cell cycle in three nonortho-
dox seed species serve to highlight the fact that different fac-
tors may contribute to the nature, and differing degrees, of
desiccation sensitivity.


A range of antioxidant processes operate in orthodox seeds
(e.g. Hendry 1993, Leprince and others 1993), and the role of
such processes under conditions of water deficit and desicca-
tion stress in plants has been reviewed by McKersie (1991) and
Smirnoff (1993). As discussed above, it is particularly in the
water content range from 0.45 to 0.25 g g -1 (30 to 20 percent,
wmb), that unregulated metabolic events resulting in the first
wave of free-radical generation are likely to occur (Vertucci
and Farrant 1995). This implies that antioxidant systems (i.e.
free-radical scavenging systems) should be maximally effective
during maturation drying of orthodox seeds, and again when
seeds take up water upon imbibition.
Reviews of metabolic damage associated with dehydra-
tion of recalcitrant seeds highlight the idea that free-radical
generation may well be a major injurious factor (Berjak and
Pammenter 1997; Côme and Corbineau 1996a, 1996b; Smith
and Berjak 1995), particularly because protective mechanisms
appear to become impaired under conditions of water stress
(Senaratna and McKersie 1986, Smith and Berjak 1995). Rapid
formation of free radicals and decreasing activity of antioxi-
dant systems have been reported as occurring during dehydra-
tion of the seeds of the temperate recalcitrant species Quercus
robur (Finch-Savage and others 1993). Lipid peroxidation,
which is a major consequence of uncontrolled free-radical gen-
eration, with the ultimate accumulation of a stable free radical
in the embryonic axes, has been shown to accompany dehy-
dration of the seeds of three temperate, recalcitrant species—
-Q. robur, Castanea sativa, and Aesculus hippocastanum (Finch-
Savage and others 1994a)–and free radical formation has been
reported to accompany viability loss in seeds of the highly
recalcitrant, tropical species Shorea robusta (Chaitanya and
Naithani 1994). While hydroperoxide formation has been
shown to accompany dehydration at a range of temperatures
of the recalcitrant seeds of Zizania palustris, significantly more
was produced at 37 °C than at 25 oC, and tetrazolium tests
revealed that viability was severely affected by water loss at the
higher temperature (Ntuli and others 1997).
From the evidence reviewed above, there is no doubt
that damage ascribable to uncontrolled free-radical generation


Part I—Technical Chapters

occurs during dehydration in the recalcitrant seeds of a range
of species that show differing degrees and manifestations of
nonorthodox behavior. This implies not only that free radicals
are produced as a consequence of water stress in these desic-
cation-sensitive seeds, but also that antioxidant systems are
ineffective at curbing them. Together, then, these factors must
be seriously considered as constituting one of the major caus-
es of desiccation sensitivity.


Late Embryogenic Accumulating/Abundant
Proteins (LEA’s)

LEA’s (Galau and others 1986) comprise a set of hydrophilic,
heat-resistant proteins associated with the acquisition of des-
iccation tolerance in developing orthodox seeds (Galau and
others 1991 reviewed by Bewley and Oliver 1992, Kermode
1990, Ried and Walker-Simmons 1993). Their synthesis appears
to be associated with the high ABA levels that peak during the
later stages of seed development (Kermode 1990). The charac-
teristics of LEA’s and the conditions under which they appear
have led to suggestions that they function as protectants, per-
haps stabilizing subcellular structures in the desiccated condi-
tion (Close and others 1989, Dure 1993, Lane 1991).
The position of LEA’s (or dehydrin-like proteins, as they
may be termed) in nonorthodox seeds appears at first sight to
be anomalous, as some species do not express these proteins
while others express them to variable extents. Seeds of Avi-
cennia marina, which are extremely desiccation sensitive,
appear not to express LEA’s at all (Farrant and others 1992a).
In contrast, seeds of Zizania palustris (North American wild
rice), which are recalcitrant (Vertucci and others 1994) but
show differential responses to dehydration depending on tem-
perature (Kovach and Bradford 1992a, Ntuli and others 1997),
do express this type of protein (Bradford and Chandler 1992,
Still and others 1994). Dehydrin-like proteins were shown to
be expressed in a range of temperate, recalcitrant species
(Finch-Savage and others 1994b, Gee and others 1994), but
the absence of such proteins correlated with low ABA levels
was found to characterize the mature, recalcitrant seeds of 10
tropical, wetland species (Farrant and others 1996). Those
authors showed the presence of dehydrin-like proteins in other
temperate and tropical recalcitrant (nonwetland) species, and
suggested that their occurrence may be habitat-related, per-
haps also providing protection against low-temperature stress.
In a comparative study on mature seeds of two tropical tree
species, neither of which occurs in wetlands, dehydrin-type

proteins were absent in Trichilia dregeana, while accumulating
in Castanospermum australe (Han and others 1997). The
immature seeds and the seedlings of these two species were
shown to differ in terms of production of such proteins in
response to stresses imposed by dehydration, application of
ABA, or exposure to cold, with T. dregeana not responding by
the production of these putatively protective proteins (Han
and others 1997).
Thus, it seems that the ability to express LEA’s or dehy-
drin-type proteins cannot be taken as an indication that the
seeds of a particular species will or will not withstand dehy-
dration. This indicates clearly that desiccation tolerance must
be the outcome of the interplay of more than one (and proba-
bly many) mechanisms or processes. Details of this, particu-
larly pertaining to LEA’s/dehydrins, sugars, and various stress-
es, have been reviewed by Kermode (1997). However, the vari-
able expression of LEA’s/dehydrins in recalcitrant seeds on a
species basis may, in association with the presence or absence
of other factors, account for the degree of nonorthodox behav-
ior exhibited under a particular set of circumstances.

Sucrose, Oligosaccharides, or
Galactosyl Cyclitols

The possible role(s) of nonreducing sugars in relation to des-
iccation tolerance in seeds has been extensively reviewed (e.g.
by Berjak and Pammenter 1997, Horbowicz and Obendorf
1994, Obendorf 1997, Vertucci and Farrant 1995). Accumula-
tion of nonreducing sugars, particularly of the raffinose series
(Blackman and others 1992, Koster and Leopold 1988, Lep-
rince and others 1990a) and/or galactosyl cyclitols (Horbowicz
and Obendorf 1994, Obendorf 1997) has been implicated in
the acquisition and maintenance of the desiccated state in
orthodox seeds, generally in two major ways. These are in
terms of the “Water Replacement Hypothesis” (Clegg 1986,
Crowe and others 1992) and vitrification, otherwise referred to
as glassy state formation (Koster and Leopold 1988, Leopold
and others 1994, Williams and Leopold 1989).
Orthodox seed maturation invariably seems to be
accompanied by the accumulation of nonreducing oligosac-
charides which coincides with the reduction of monosaccha-
rides, and maintenance of the desiccated state is associated
with high levels of sucrose and other oligosaccharides. Evi-
dence for the replacement of membrane-associated water (the
Water Replacement Hypothesis, i.e. the replacement of water
by sucrose to maintain lipid head-group spacing, thereby pre-
venting gel-state transformation) is equivocal, and a recent cri-
tique questions its relevance in the desiccated state of ortho-
dox seeds (Hoekstra and others 1997). However, the role of
sucrose in the formation of intracellular glasses (i.e. vitrifica-

Chapter 4: Orthodox and Recalcitrant Seeds

tion) is more convincing. The metastable, glassy state occurs at
low water contents in seeds, when sucrose and certain oligosac-
charides or galactosyl cyclitols form high-viscosity, amor-
phous, super-saturated solutions (Obendorf 1997). The occur-
rence of glasses is held to impose a stasis on intracellular reac-
tivity, protecting macromolecules against denaturation and
possibly preventing or minimizing liquid crystalline gel phase
transformations of the lipid bilayer of membranes (e.g.
Leopold and others 1994).
Walters and others (1997) have suggested that a signifi-
cant proportion of the sugars may be tightly associated with
LEA’s–these complexes acting to control and optimize the rate
of water loss during dehydration of orthodox seeds. It should
be noted, however, that this should not obviate the participa-
tion of either the LEA’s or the sugars in the maintenance of
orthodox seed viability in the desiccated state.
The formation of intracellular oligosaccharides occurs at
the expense of monosaccharides, and confers the advantage
that immediately available respiratory substrates are removed
(Koster and Leopold 1988, Leprince and others 1992, Roger-
son and Matthews 1977). This would serve to reduce the spec-
trum of damaging reactions that can occur as orthodox seeds
pass through critical water content ranges favoring unbalanced
metabolism, during maturation drying (see “Switching off” of
metabolism, above).
Whatever the role(s) of sucrose and oligosaccharides or
galactosyl cyclitols may be in orthodox seeds, seeking parallels
for desiccation-sensitive seeds is entirely inappropriate. While
sucrose and other oligosaccharides are produced in some of
the few recalcitrant seed species that have been assayed (Far-
rant and others 1993, Finch-Savage and Blake 1994), glass for-
mation will occur only at water contents well below the lethal
limit. When recalcitrant seeds are dehydrated under ambient
conditions (which is what would occur in the natural habitat),
they lose viability at relatively high water contents–in the
region of 0.7 g (or more) water per g dry mass [ 40 percent,
wmb] (Pammenter and others 1991), which are far higher that
those required for glass formation to occur (Bruni and
Leopold 1992, Leopold and others 1994, Sun and others 1994,
Williams and Leopold 1989). The same argument holds if
water replacement by sugars is an operative phenomenon in
orthodox seeds; this too would occur only at water contents of
0.3 g per g dry material (Hoekstra and Van Roekel 1988),
which is well below the lethal limit for slowly drying recalci-
trant seeds.
The one involvement of sugars in the variable desicca-
tion sensitivity of recalcitrant seeds might be via the mecha-
nism suggested by Walters and others (1997) for maturing
orthodox seeds, viz. the modulating effect of sugar/LEA com-
plexes on dehydration rate. Very marked variability occurs in

the rate at which recalcitrant seeds of different species lose
water under the same conditions (Berjak and Pammenter
1997, Farrant and others 1989) and it is possible that the sig-
nificance of sugars and LEA’s in embryos of recalcitrant seeds
of some species lies in the modulation of the drying rate by
complex formation. Walters and others (1997) have also sug-
gested that LEA proteins in temperate recalcitrant seeds may
play a role in their survival during overwintering.


It has been suggested that partitioning of endogenous amphi-
pathic molecules (amphipaths) into membranes upon water
loss may be a prerequisite for desiccation tolerance (Golovina
and others 1998). Those authors have presented evidence of
the movement during dehydration of both introduced, apolar
spin probes and endogenous amphipaths into the bilayer of
desiccation-tolerant pollen. This process, which was complete
after dehydration to the relatively high water content of 0.6 g
per g dry mass (37 percent, wmb), was reversed during rehy-
dration, when the amphipaths repartitioned to the cytomatrix
(aqueous cytoplasm). This reverse movement was suggested to
account for the transient leakage that is invariably observed
when desiccation-tolerant material (pollen and seeds) is
imbibed from the dry state (Golovina and others 1998).
The partitioning of amphipathic molecules into the
bilayer was suggested by those authors as serving to maintain
the integrity of membranes in the dry state in desiccation-tol-
erant organisms, by substantially lowering the water content at
which the phase change of membrane lipids occurs. Liquid
crystalline to gel phase changes in membranes are well docu-
mented in response to dehydration, but the essential property
for desiccation tolerance is that they must be reversible,
reestablishing the membranes in a functional condition upon
rehydration (Hoekstra and others 1992). This demands that
integral membrane proteins retain their position in the desic-
cated state, a role that might also be ascribed to the amphi-
pathic molecules.
If the partitioning of amphipaths into membranes is
established as a universal phenomenon occurring during dehy-
dration of orthodox seeds, it is possible that they are absent or,
if present, incompletely functional or nonfunctional in desic-
cation-sensitive seeds. Dehydration of the embryos from
recalcitrant Camellia sinensis seeds was found to induce a
phase change in membrane lipids, which was reversible, but
the proteins were irreversibly affected (Sowa and others 1991).
It may be significant that at a water content of 0.6 g g-1 , when
amphipath partitioning has been observed to be complete
(Golovina and others 1998), slowly dried recalcitrant seeds,


Part I—Technical Chapters

and even the flash-dried axes of certain species, will have lost
viability (Pammenter and others 1991, 1993; also see below).
In highly desiccation-sensitive recalcitrant seeds, it is possible
that phase changes of the membrane bilayers might not be
reversible, for example, if nonbilayer structures or hexagonal
phases result (reviewed by Vertucci and Farrant 1995). Parti-
tioning of endogenous amphipaths into the bilayer upon dehy-
dration is unlikely to act in isolation; thus, even if such mole-
cules are present in cells of recalcitrant seeds, they may well
depend on another mechanism or process to achieve their
reversible migration.


The term oleosin refers to a unique protein type that surrounds
the lipid (oil) droplets in plant cells (Huang 1992). Oleosins
have a central, hydrophobic domain that interacts with the
periphery of the lipid, and an amphipathic N-terminal domain
that, with the C-terminal domain, facilitates interaction with
the aqueous cytomatrix. The oleosin boundary of lipid bodies
allows these hydrophobic masses to be accommodated as dis-
crete entities in the aqueous cytomatrix under hydrated con-
ditions, and it has been suggested that their role during dehy-
dration prevents the bodies from coalescing in desiccation-tol-
erant seeds (Leprince and others 1997).
Leprince and others (1997) recorded a lack (or inade-
quate amount) of oleosins in desiccation-sensitive seeds of
some species, and although little obvious change in the integri-
ty of the bodies as a consequence of dehydration was
observed, rehydration appeared to have deleterious effects on
their stability. Coalescence of lipid bodies is a common abnor-
mality accompanying deterioration, even in cells of stored,
orthodox seeds (Smith and Berjak 1995). Although the effects
of fungi associated with both stored orthodox and recalcitrant
seeds in bringing about lipid body coalescence cannot be ruled
out, the occurrence of this phenomenon could well be, at least
partly, a consequence of some deficiency in desiccation-sensi-
tive seeds. In view of the findings of Leprince and others
(1997), the deficiency of an adequate oleosin sheath around
the lipid bodies may underlie the inherent instability of these
organelles during rehydration following damaging levels of
desiccation of some recalcitrant seeds. However, it must be
stressed that the presence of fully functional oleosins cannot,
in itself, account for desiccation tolerance. Rather, it must be
viewed as one of the mechanisms contributing to the spectrum
of properties necessary if orthodox seeds are to survive
extreme dehydration.


There is both indirect and direct evidence that repair mecha-
nisms do come into play when dry orthodox seeds are rehy-
drated. For example, seeds that have been stored under
adverse conditions, but are still 100-percent viable, typically
show a lag before there are visible signs of germination, during
which it is commonly accepted that repair processes are taking
place. Ultrastructural studies on maize seeds have provided
evidence supporting this contention, where mitochondrial
repair was observed during the lag period (Berjak and Villiers
1972). Studies on rye seeds have shown that even in the dry
state there is progressive deterioration of the DNA as a result
of endo- and exonuclease activity during storage (Elder and
others 1987), which cannot be repaired until the seeds are
rehydrated (Boubriak and others 1997).
Much of the evidence for the operation of repair process-
es during rehydration comes from osmopriming experiments
on low-vigor seeds. This process involves controlled rehydra-
tion to the end of phase II, which achieves a hydration level
that facilitates repair but precludes germination proper (Bray
1995, Bray and others 1993). Those authors have shown that
replacement of damaged rRNA occurs, and lesions in the DNA
and protein-synthesizing systems are repaired, during priming.
It is generally agreed that free radical generation (see
above) continues in air-dried orthodox seeds during storage
(reviewed by Smith and Berjak 1995) and the ensuing damage
obviously must be repaired on rehydration, arguing strongly
for the presence and efficient operation of antioxidant systems
at this stage. During dehydration of desiccation-sensitive seeds
and seedlings, however, such systems have been shown to fail
(Hendry and others 1992, Leprince and others 1992) and are
assumed to remain ineffective when water is once again pro-
vided (Côme and Corbineau 1996a, 1996b).
When recalcitrant seeds or axes excised from such seeds
are subjected to nonlethal dehydration, it is generally observed
that there is an increase in the time taken for the onward
growth of germination, which might be interpreted as facili-
tating repair. However, this is likely to be strictly limited; pres-
ent studies have shown that after 22 percent of the water is lost
from hypocotyl tips of Avicennia marina, dehydration-associ-
ated DNA damage can no longer be repaired when water is
once again provided. DNA instability to dehydration is also
shown by seedlings produced from orthodox seeds, once they
have reached the stage when desiccation tolerance has been
lost (Boubriak and others 1997).
Very little work that targets the aspect of possible repair
of mature, dehydration-damaged recalcitrant seeds has yet

Chapter 4: Orthodox and Recalcitrant Seeds

been done. It is presently tacitly assumed that the necessary
repair systems are present, but are themselves damaged by
dehydration beyond certain limits–limits that might vary among
seed species of markedly differing desiccation sensitivity. How-
ever, this aspect requires considerable investigation to obtain
both qualitative and quantitative data to clarify the situation.


We now know that much confusion has occurred in compara-
tive work on individual species of recalcitrant seeds because of
conflicting data regarding “critical water contents,” below
which viability will be lost. This is because the dimensions of
the time taken for water to be lost, or the temperature at which
the drying experiments were carried out, have been ignored.
While the effects of temperature will not presently be dis-
cussed, there are several publications focused on the seeds of
Zizania spp. which show that this parameter can have very
marked effects on the outcome of drying regimes and/or opti-
mal storage water contents (Kovach and Bradford 1992b;
Ntuli and others 1997; Vertucci and others 1994, 1995). The
effect of the maturity status of the seeds–which is often
extremely difficult to ascertain for recalcitrant types–also has
significant effects on the degree of dehydration that will be tol-
erated (reviewed by Berjak and Pammenter 1997, Finch-Sav-
age 1996) but also will not be taken further.
The aspect of the time taken for water to be lost is a vari-
able that has been identified as having profound effects on the
degree of dehydration that desiccation-sensitive seed material
will tolerate. The more rapidly dehydration can be achieved,
the lower is the water content to which the seeds or axes can
be dried without damage accumulation that culminates in via-
bility loss. This is particularly marked when excised axes are
dried (Berjak and others 1993; Normah and others 1986; Pam-
menter and others 1991, 1993). Very rapid drying of excised
recalcitrant axes (flash-drying) facilitates nonlethal dehydra-
tion to water contents in the region of 0.4 to 0.25 g g-1 dm,
which is close to the hydration level where all the water is non-
freezable (generally structure-associated), although tolerance
to such low water contents is not invariably the case (Pam-
menter and others 1993). It must be noted, however, that such
rapid drying does not mean that the seed tissues are potential-
ly desiccation tolerant; rather, the faster dehydration can be
achieved, the less the time during which the axes are in the

water content range that permits damaging, potentially lethal,
aqueous-based reactions to occur. As discussed below, these
are the processes that, given sufficient time, will cause viabili-
ty loss at relatively high water contents when the tissues are
dehydrated slowly (Berjak and others 1989, 1993; Pammenter
and others 1998; Pritchard 1991). Far from actually being des-
iccation tolerant, axes from recalcitrant seeds will survive only
for very short periods (hours to a day or two), at the lowest
water contents attainable (Walters and others 2001).
Marked effects of drying rate on whole seeds are gener-
ally harder to attain, because seed size often prevents the
achievement of suitably rapid dehydration. However, not all
recalcitrant seeds are too large, or lose water too slowly, to
facilitate the achievement of very different drying rates. The
ability to achieve lower water contents while retaining viabili-
ty has been recorded for whole seeds of Avicennia marina
(Farrant and others 1985) and Quercus rubra (Pritchard 1991).
We have recently carried out studies to ascertain the effects of
drying rate on whole seeds of Ekebergia capensis, a tropical,
meliaceous, recalcitrant species) for which markedly different
drying rates can be achieved (Pammenter and others 1998).
The results obtained illustrated the effects of drying rate dra-
matically: viability loss was already apparent in slowly dried
seeds at high axis water contents [ 1.25 g water per g dry mate-
rial ( 55 percent, wmb)] while those that were dehydrated rap-
idly showed unimpaired vigor and full germinability at an axis
water content of 0.7 g g-1 (40 percent, wmb). Seeds dried at an
intermediate rate retained viability to the intermediate axis
water content level of c. 1.0 g g-1 (50 percent, wmb). Ultra-
structural observations suggested that different damaging
mechanisms bring about intracellular damage, depending on
the drying rate. Advanced degradation of membranes, partic-
ularly of the plastids, and an abnormality of the lipid bodies
occurred in axes from slowly dried seeds at water contents in
the region of 1.1 g g-1 (52 percent, wmb) when viability had
declined to 37 percent. The damage became steadily worse
with slow drying to lower water contents, until, at 0.6 g g-1 (37
percent, wmb), only fragments of intracellular components
remained. At a water content of 0.57 g g-1 (36 percent, wmb),
axes from rapidly dried seeds (viability 80 percent) showed lit-
tle signs of intracellular damage; it was only at considerably
lower axis water contents that signs of deterioration were
noted, which coincided with declining viability. At no stage
did the extensive degradation that characterized axis cells
from slowly dried seeds occur, supporting the proposal that if
desiccation-sensitive material can pass quickly enough through
water content ranges at which lethal reactions are prevalent,
then it is possible to dry the material down to a far lower
hydration level (see Vertucci and Farrant 1995 for discussion
of the various hydration levels).


Part I—Technical Chapters

There will be a water content at which rapidly dried
material that is desiccation sensitive will sustain injury, and,
while the value varies from species to species, it is usually near
the range where only structure-associated (nonfreezable)
water remains (Pammenter and others 1991, 1993; Pritchard
1991). Damage occurring at such relatively low water contents
is defined as desiccation damage in the strict sense (Pam-
menter and others 1998; Walters and others 2001) and is sug-
gested to coincide with the perturbation of the nonfreezable
water (Pammenter and others 1991). In contrast, desiccation-
tolerant material can withstand the removal of a considerable
proportion of this water (Pammenter and others 1991, Vertuc-
ci and Farrant 1995).
Slowly dried desiccation-sensitive material sustains dam-
age at relatively high water contents, certainly those where
solution (i.e. freezable) water prevails. This damage is suggest-
ed to result from aqueous-based, degradative reactions that are
the result of unbalanced metabolism (Pammenter and others
2001; Walters and others 2001). Recalcitrant seeds (and,
indeed, probably all nonorthodox types) are hydrated and
metabolically active when shed (Berjak and others 1989, Ber-
jak and Pammenter 1997). As water is slowly lost, metabolism
will continue, but when the seeds are still at relatively high
water contents, metabolism will become unbalanced or out-of-
phase as a result of internal water stresses (Senaratna and
McKersie 1986, Smith and Berjak 1995, Vertucci and Farrant
1995). A likely consequence of this unregulated metabolism
will be the generation of free radicals and accompanying
oxidative damage (Finch-Savage and others 1994a, Hendry
1993, Hendry and others 1992, Leprince and others 1990b).
The severity of this type of damage, which is being termed
metabolic damage (Walters and others 2001), is predicted to

increase in inverse proportion to the drying rate, with viabili-
ty loss occurring at increasingly high water contents.


It is proposed that nonorthodox seed behavior is a consequence
of the lack of some, or perhaps all, of the suite of protective
mechanisms or processes that together confer desiccation tol-
erance on orthodox seeds. There is likely to be a gradation in
the presence and/or efficacy of the proposed processes/mech-
anisms among seeds of nonorthodox species, accounting for
the variability of the responses to stresses, particularly that
imposed by dehydration. The most desiccation-sensitive recal-
citrant seeds are probably those that lack virtually all the pro-
tective and restitutional factors that facilitate the acquisition
and maintenance of desiccation tolerance in orthodox seeds.
Two major factors are proposed to contribute to the loss
of viability of recalcitrant seeds: (1) the consequences of unbal-
anced metabolism during dehydration [and possibly also when
such seeds are stored in the hydrated condition (Smith and
Berjak 1995)]; (2) desiccation damage in the strict sense, which
occurs when water that is essential for the integrity of intracel-
lular structures is removed; in recalcitrant seeds, this equates
with nonfreezable water (Pammenter and others 1991).
We will probably be unable to account satisfactorily for
nonorthodox seed behavior, particularly that of truly recalci-
trant seeds, until complete understanding is gained of the
apparently numerous interacting factors that enable desicca-
tion-tolerance to be achieved.

Chapter 4: Orthodox and Recalcitrant Seeds