Critical periods as fundamental events in life

ARTICLE IN PRESS

Theory in Biosciences 123 (2004) 17–32 www.elsevier.de/thbio

Critical periods as fundamental events in life G. Nissim Amzallag* The Judea Center for Research and Development, Carmel 90404, Israel

Abstract Development is not a continuous phenomenon. Rather, phenophases are interspaced with short critical periods. This phenomenon reflects an alternance between stabilization (during a phenophase) and dismantling (during a critical period) of a network of between-organ relationships generating the organism. Networks of relationships may be compared to dissipative systems in physics. In this context, a critical period represents a transient phase of isolation of the systems enabling its evolution towards equilibrium. As suggested here, this transition from dissipative to isolated system represents the source of newly emerging dissipative structures in which environmental or developmental perturbations are adaptively integrated. In contrast to non-living systems, an endogenous control of the transition towards critical period seems to exist during development. By extension to other scales of biological organization, it is suggested that the capacity to self-define its status (dissipative or close-toequilibrium) represents the key property of living systems. This asks for a reconsideration of some basic notions about life, such as the role of genes in normal development, in physiological adaptation, and even in the emergence of evolutionary novelty. r 2004 Published by Elsevier GmbH. Keywords: Adaptation; Critical period; Definition of life; Dissipative systems; Non-linear dynamics; Phenophase; Response to stress

La vie comporte en elle-m#eme des e! tats critiques qu’elle surmonte par des cre! ations qui constituent pre! cise! ment la vieyCes e! tats critiques ordinairement nous e! chappent parce que l’ organisme y a de! ja" re! pondu.1 Maldiney (2001, p. 83)

*Tel.: +972-2-9960061; fax: +972-2-9960061. E-mail address: [email protected] (G.N. Amzallag). 1 Life comprises critical stages that it overcomes by creations which precisely represent lifey . These critical stages usually escape us because the organism has already responded to them. 1431-7613/$ - see front matter r 2004 Published by Elsevier GmbH. doi:10.1016/j.thbio.2004.03.003

ARTICLE IN PRESS 18

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

Introduction During the last decades, accumulation of data in genetics and in molecular biology led to the representation of development as a process controlled by the genome. Consequently, the main objective of biological research was to investigate the causal link supposed to exist between the ‘‘genotype’’ and the ‘‘phenotype’’. In this context, development has been understood as an informatively closed process, and signals (endogenous or environmental) were interpreted as simple triggers for expression of the genetic pre-existing programs. This deterministic relationship between genotype and phenotype has been denounced as an abusive simplification for a long time. Canalization (as defined by Waddington, 1961) and plasticity (as defined by Bradshaw, 1965) are two antagonistic characteristics of development (Stearns, 1982). However, it remains impossible to restrict canalization to the expression of genetic information since, in many cases, phenotypic and genetic variability are unrelated (Cock, 1966; Barton and Turelli, 1989; Perfectti and Camacho, 1999). Being frequently adaptive, phenotypic plasticity cannot be restricted to the expression of a genetic information (Sultan, 1992, 1995), though the latter is clearly involved (Cheplick, 1995; Pigliucci, 1996). Hence, self-emergent and genetic components appear both contrasted and interacting during development (Via et al., 1995). For example, ‘developmental’ and ‘genetic’ factors are so intricately related in control of phyllotaxy that it ‘‘seems to be a contradiction between the determination of the phyllotaxy by genetic factors and by self-organization’’ (Douady and Couderc, 1996). The aim of this paper is to show that this obscure situation results from a misleading representation of development as a continuous phenomenon. Discrete stages (also termed phenophases) have been observed during the development of multicellular organisms (Brink, 1962; Zhirmunsky and Kuzmin, 1988; Slafer and Rawson, 1994). They have commonly been interpreted as changes in the pattern of gene expression (Poethig, 1990). Assuming that development is an informatively closed process, periods of transition between two discrete stages have been considered as short time-intervals devoid of biological significance. However, as emphasized by Gerhart et al. (1982) ‘‘development proceeds by a series of stages, with the morphology of one stage providing the ‘initial conditions’ within which morphogenetic mechanisms must operate’’. This suggests that the time interval of transition between two phenophases, defined here as a critical period, may be involved in emergence of the new phenophase.

The dissipative nature of living systems and its quantification A continuous dissipation of chemical energy is required for preservation of the cellular organization. For this reason, cells have been considered as dissipative systems. Such a representation relates the developmental buffering and robustness to typical behavior of dissipative systems (Prigogine and Wiame, 1946; von Bertalanffy,

ARTICLE IN PRESS G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

19

1950; Atlan, 1972; Conrad, 1983). From this analogy, Glandsdorff and Prigogine (1971, p. 289) concluded that ‘‘life no longer appears as an island of resistance against the second law of thermodynamics, or as the deed of some Maxwell demons. It would appear now as the consequence of the general laws of physics, appropriate to specific chemical kinetics and to far from equilibrium conditions. These specific kinetic laws are those which permit the flow of energy and matter to build and maintain functional and structural order in open systems.’’ Dissipative systems are characterized by long-range correlations between microscopic units. For this reason, the linkage between homologous subunits should be considered as the expression of the ‘‘dissipative dimension’’ of a biological system. In ecosystems, the relationship between the density of population of two species has been recently reconsidered as the level of linkage within a food web. This connectivity determined for food webs has been analyzed in a physical perspective. This reveals that the level of stability of an ecosystem is proportional to its connectivity (Martinez, 1993; Fonseca and John, 1996; Murtaugh and Kollath, 1997; Schneider, 1997). Moreover, a parallel between the rate of regeneration of an ecosystem and connectivity of its food webs has also been observed (Law and Blackford, 1992). These observations confirm theoretical considerations about the relationship (within a specific range) between connectivity of a system and its stability (Gardner and Ashby, 1970, 1972). They also confirm the validity of interpretation of connectivity as an expression of the dissipative nature of a biological system. Between-organs interactions generating an organism are not similar to the trophic links within an ecosystem. However, a quantification of connectivity may also be performed at the organism level. The idea of measuring the level of linkage of a character with other homologous characters was first suggested by Terentjev (1931). Based on calculation of the correlation coefficient for linear regression, this analysis enables to cluster characters in correlation pleiades. By the use of such a method, it was shown that constancy in morphogenesis of insect-pollinated flowers (as compared to high variation in wind-pollinated flowers) corresponds to a high level of relationship between reproductive characters (Berg, 1959, 1960). Similarly, the stability of patterns on butterfly wings was found to be proportional to the level of linkage between motives generating the pattern (Paulsen and Nijhout, 1993; Paulsen, 1994). Some authors observed that connectivity varies during development. For example, an increase in the relationship between morphological characters is observed towards the adult stage in mammals (Herbert et al., 1979; Atchley and Rutledge, 1980; Atchley, 1987) and in lizards (Aleksic and Turic, 1994). In Sorghum bicolor, the relationship between parameters of consecutive leaves also increases towards the late vegetative stage (Amzallag, 1999a, 2002a). In developmental studies, connectivity has been estimated by calculation of the correlation coefficient (r) for linear regression between two variables. A global quantification of connectivity between a series of characters implies to consider the r coefficients (calculated between each couple of characters) as

ARTICLE IN PRESS 20

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

parametric values. However, the r coefficients are not normally distributed. This is why, out of ecological studies (in which connectivity has been quantified as the conditional probability of the presence of a species when the other is observed), the concept of connectivity remained mainly qualitative. But this problem may be resolved by transforming the r coefficients (non-normally distributed) into z values (approximately normally distributed, excepted for r values very close to 1 or 1). According to Sokal and Rohlf (1981, pp. 583–587), the z-coefficient calculated for each r value (z=0.5 Ln[(1+r)/(1 r)]) may be considered as a quantification of the level of relationship. Thus, connectance of a character may be defined as the mean of the absolute value of the z-transformed r coefficients for its correlation with the other (more or less homologous) characters considered (Amzallag, 2000a). Through this quantification, it becomes possible to verify if properties described for connected food webs are also observed at the organism level, and if stability and adaptedness during development are related to variations in the dissipative nature of the organism during its development. In Sorghum bicolor, connectance calculated between reproductive characters has been found to be inversely proportional to the level of phenotypic variability (Amzallag, 2000a). This confirms the assumption that connectivity is a central component of stability in expression of a character. In this case, a domain of developmental instability has also been found for high connectance values (Amzallag, 2000a). But this instability inherent to very high levels of connectivity does not invalidate the parallel with dissipative systems. Rather, it aims for the loss of redundancy (networkness) in interactions, so that inherent variations are not buffered (Amzallag, 2001a).

The theoretical requirement for critical periods in development Spontaneous elaboration of an organized structure has been observed for a long time in social insects (Deneubourg, 1977; Grasse! , 1984; Millonas, 1992; EdelsteinKeshet et al., 1995) as well as in microorganisms (Keller and Segel, 1970; Goldbeter and Segel, 1980; Dujardin and Walbaum, 1985; Shapiro and Hsu, 1989). In all these cases, the structure is generated by progressive extension from a nucleation event spontaneously emerging in the initially homogeneous population. These observations correspond to models of emergence of dissipative systems in physics (Glandsdorff and Prigogine, 1971). The ‘physical nature’ of this process is confirmed by the fact that, under artificial conditions, coherent and reproducible superstructures are also observed in micro-organisms for which a ‘social dimension’ is unknown in the wild (Slutsky et al., 1985; Shapiro, 1985, 1987). In this latter case, no genetic information may explain emergence of the superstructure. In general, the environmental anisotropy dissipated by a biological system is a gradient of the concentration of a substance rather than heat. In the case of selfaggregation of micro-organisms, the already aggregated cells, through stimulation

ARTICLE IN PRESS G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

21

and coordination of their secreting activity, represent the source of the chemical anisotropy. But, in development, the chemical gradient is frequently generated by secretion from cells already organized in tissues. Such a diversity of sources of anisotropy increases considerably the variety of structures that may emerge from dissipation of a chemical gradient (Turing, 1952). Moreover, it also provides a precise control (in time and in space) of their emergence. One of the main characteristics of dissipative structures is stability. But morphogenesis is not a single-step process of differentiation in a homogeneous population of egg-derived cells. Rather, emergence of new structures and organs occurs frequently until achievement of development. This reveals that, in spite of their high stability, the dissipative structures may be modified after their emergence. Two hypotheses may explain such a phenomenon: (i) the dissipative system is emerging slowly, so that it completes progressively towards the adult stage; (ii) the dissipative system emerges rapidly in response to anisotropy and it is later modified as a whole. The functional integration within an organism depends on the level of connectivity between its organs (Chauvet, 1993). Thus, according to the first hypothesis, one has to expect a lower level of physiological coordination in the larva, as compared to that of the adult stage. However, this phenomenon is not observed, since juvenile and adult stages display a similar capacity of regulation. Even the ability to cope with a stress is frequently higher during the larval stage than during the adult stage (Saliba and Krzyz, 1976; Pascoe and Beattie, 1979; Fraser, 1980; Dixon and Sprague, 1981a, b; Kerver and Rotman, 1987). In plants, adaptive phenotypic plasticity is generally expressed for organisms exposed to an environmental perturbation during the very early stage of development (Sultan, 1995; Amzallag, 1999c). Accordingly, the physiological networks should not be considered as incompletely developed during the juvenile stage of development. A dissipative structure is characterized by its high stability. For this reason, modification of an already existing dissipative system (as suggested by the second hypothesis) implies a transition through a critical period during which connectivity is reduced (Ashby, 1962). This phenomenon transforms the development in a succession of discrete phases of stability (phenophase) interspaced by short critical periods of their (partial) decomposition.

Evidences towards critical periods in development Because of its redundant structure, a network is able to buffer variations from endogenous and/or environmental origin. Thus, changes in variability in expression of characters during development may reveal modifications of the network involved regulation of morphogenesis (see Amzallag, 2001a). Fluctuations in variability have been observed during plant development. Concerning the leaf shape, the level of variability depends on the stage of development in Nicotiana (Paxman, 1956; Sakai and Shimamoto, 1965) and in

ARTICLE IN PRESS 22

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

Clarkia temblorensis (Sherry and Lord, 1996). Fluctuating asymmetry may also be suddenly modified at specific periods of development (Sherry and Lord, 1996; Swaddle and Witter, 1997). These observations represent the first indications concerning sudden modifications of the network involved in morphogenesis. They have been confirmed by direct measurements of connectivity. For example, in the rock lizard Lacerta oxycephala, Aleksic and Turic (1994) observed a sudden drop in what they termed ‘phenotypic correlations’ about 30 days following hatching. Similarly, in Sorghum bicolor, the relationship between leaf size and growth decreases suddenly at unfolding of the fifth leaf (Amzallag, 2002a). In this case, the transient drop in connectance for leaf morphogenesis has been related to the emergence of a new type of organ : the adventitious roots (Amzallag, 1999a). This latter event induces a physiological perturbation at the whole plant level by introducing a new source of informative molecules (the tip of adventitious roots) with its own regulations (Amzallag, 2001b, c). Accordingly, the transient drop in connectance (measured during unfolding of the fifth leaf, after emergence of the first adventitious roots in Sorghum) has been related to integration of a new type of organ in the whole-plant regulation network (Amzallag, 2002b). This assumption is strengthened by the fact that, in Sorghum, plants may adapt to environmental stresses during critical period. For example, an exposure to a moderate concentration of NaCl (150 mM) induced, within three weeks, a capacity to grow at 300 mM NaCl, a concentration lethal for non-treated plants (Amzallag et al., 1993). This adaptive response (that is not restricted to exposure to NaCl, see Amzallag, 2002c) is expressed only for plants exposed to NaCl from unfolding of the fifth leaf (Amzallag et al.,1993). This precise timing reveals the link existing between adaptation to environmental changes and ‘developmental adaptation’ (the physiological adjustments to new morphogenesis events occurring during development). In this case, a large variability has been observed in the population of adapted plants (Amzallag, 1999c), including strong differences in the regulation of physiological parameters (such as osmotic regulation and shoot ion content) (Amzallag, 2001c). This phenomenon is difficult to explain in the context of expression of a genetic program in a genetically homogeneous population. In annual plants, adaptive phenotypic plasticity is also conditioned by environmental stimuli during the early development (embryo maturation, germination or early vegetative development) (Diggle, 1994; Lejeune and Bernier, 1996; Bechoux et al., 2000; Callahan and Waller, 2000; Amzallag, 2002c). In most of these cases, phenotypic plasticity was interpreted as an environmentally induced stimulation of one of the developmental trajectories controlled by genes selected for their fitness. However, such an interpretation cannot explain how the adaptive dimension of phenotypic plasticity fits the precise environmental conditions encountered (Sultan, 1992; Amzallag, 2002b; Trewavas, 2003). It cannot explain the sudden increase in variability inherent to the physiological changes (Belayev, 1978; Amzallag, 1999b; Trewavas, 2003). In contrast, these phenomena may be clearly integrated in a non-linear perspective of development for which adaptive changes occur specifically during critical periods.

ARTICLE IN PRESS G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

23

About emergence of critical periods Emergence of a critical period results from a sudden dismantlement of the dissipative structure, bringing back the system to its equilibrium status. This phenomenon strongly resembles death (Dillon and Root-Bernstein, 1997). However, critical periods differ from death by reversibility of the change in status of the system. A sudden drop in connectivity of the dissipative system may have three origins: (i) a spontaneous breakdown of the network following its deformation, (ii) an elimination of the anisotropy generating the superstructure, (iii) a decrease in cell sensitivity to the existing chemical gradient. Mechanisms (i) and (ii) may be considered as exogenous modes of regulation of connectivity, while the third mechanism represents an endogenous mode. A disappearance of the chemical gradient (case (ii)) permits only the reemergence of a dissipative structure in the context of reduced anisotropy. In this case, development would generate adult organisms less organized than the juvenile form. Except simplification observed for the adult stage of some parasitic organisms, this case is generally not observed. A complete breakdown of the superstructure due to a deformation (case (i)) is an all-or-none process. In this case, the threshold for this sharp transition is a function of the number of units generating the superstructure and of intensity of the deformation affecting the network (Gardner and Ashby, 1970; May, 1972). By complete recovery of the equilibrium status, this mode of regulation induces reversibility in the self-emergent process. Thus, formation of dissipative structures becomes similar for all the critical periods (the main difference remaining the number, nature and position of the units generating the superstructure). Accordingly, even though this exogenous mode of regulation of breakdown probably exists, it cannot relate the progressive increase in complexity during development. It seems, therefore, that an endogenous mode of control of expression of critical periods (case (iii)) probably exists, and that it induces only a moderate (controlled in its intensity) breaking of the networks. In this context, a change in sensitivity to the molecules building the gradient may induce a transient dissipation of the environmental anisotropy, shifting the system as a whole towards an equilibrium status even though the chemical gradient is maintained. Hormones play a central role in expression of the relationship between entities building the organism. Thus, a drop in connectivity of endogenous origin implies a transient inactivation of the hormonal pathway of communication. This process, during which cells determine their own sensitivity to informative molecules, has been extensively described for animal development and termed hormonal imprinting (reviewed by Csaba, 1994). Similarly, two modes of hormone action have been observed in plants. The first, dose–response effect, corresponds to the widespread scheme of hormone action described in animal physiology. The second mode of action has been characterized as a change in sensitivity of the target cell to the hormone (Trewavas, 1991; Weyers et al., 1995). However, these two modes of action

ARTICLE IN PRESS 24

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

are mutually exclusive: dose–response represents a regulation from the outside, while change in sensitivity is an expression of autonomy of the target cells from the message produced by the source. This contradiction is the source of the ‘‘hormone dilemma’’ in plant physiology (Guern, 1987). But this obscure situation may be clarified in a non-linear perspective of development. Dose–response requires a stable relationship between the source and the target entities, whereas change in sensitivity is determined by the target cell according to its own equilibrium. Thus, dose– response may be the normal response of a cell integrated in a network of physiological regulations, while autonomous determination of sensitivity may correspond to the phase of its self-emergence. A sudden switch from dose–response to change in sensitivity becomes sufficient for inducing a sudden drop in connectivity of endogenous origin. In Sorghum bicolor, a change in sensitivity to hormones has been observed during the developmental window for salt-adaptation, also identified as a critical period in development (Amzallag, 2001b, c). In this case, the transient decrease in connectance induces a resetting of the hormonal imprinting enabling insertion of the adventitious roots in the whole plant network of regulation (Amzallag, 2001b). A concerted change in sensitivity at the level of the organism is required for emergence of a critical period. This implies the existence of a factor of synchronization of the switch towards change in sensitivity. In animals, tyroxin and steroids (belonging to the same superfamily of informative compounds) are known both as stress hormones and as factors involved in control of abrupt transitions in development (such as metamorphosis, transition from juvenile to reproductive phase of development). Thus, it is likely that some of these hormones are involved in synchronization of cell differentiation and hormonal imprinting. In plants, changes in sensitivity to hormone have also been reported following treatments with brassinosteroids (Khiprach et al., 1999; Ephritikhine et al., 1999; Amzallag, 2001a). Moreover, in Sorghum bicolor, treatments with phytosteroids modified the time and duration of the critical period emerging during the early vegetative development (Amzallag, 2002a, b). As for animals, the effect of phytosteroids is not constant during development (Takeuchi et al., 1995; Arteca and Arteca, 2001), suggesting that other factors (such as developmental stage, photoperiod, environmental changes) are also involved in the control of emergence of a critical period.

Critical periods as fundamental events A transient (and partial) dismantlement of the dissipative structure induces a spontaneous evolution of the system towards its lowest free-energy level. In development, such a phenomenon generates new levels of complexity, a functional integration of newly emerging organs and physiological adjustment to suboptimal conditions. By its adaptive dimension, this sudden scale change has been considered by Conrad (1985) to be ‘‘the most fundamental mechanism of problem solving in nature’’.

ARTICLE IN PRESS G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

25

It seems that the whole entity is able to define its nature (as a dissipative or an equilibrium system) independent of the strength of the chemical gradient which structures the network and before reaching the critical limit for stability of the network. I suggest considering this capacity for a superstructure to self-define its status as the key property of living systems, because it enables emergence of new levels of complexities after the formation of the first dissipative structure in the egg-derived community of cells. This perspective invites one to reconsider some well-known concepts in biology. (a) Variability and individuality. A complex system of interrelations between microscopic entities emerges by progressive extension of a nucleating event in a close-to-equilibrium system. In consequence, (i) the nucleating phase of emergence is extremely sensitive to environmental fluctuations, as compared both to dissipative structures and to systems at equilibrium; and (ii) this initial sensitivity is progressively reduced during emergence of the new dissipative structure, due to the buffering capacity of complex (redundant) networks. In such a context, variability in expression of a character does not always have the same significance. During a phenophase, variability is the consequence of a perturbation affecting the existing network. During a critical period, variability is generated by the fact that, at its lowest level of connectivity (the lowest free-energy level), a system becomes conditioned by all the factors influencing the spontaneous emergence of a symmetry-breaking event and its evolution towards a new dissipative status. In such a context, variability does not reflect an environmental perturbation in expression of a pre-existing (genetic) program of development, nor a constitutive heterogeneity (genetic or environmental). It is the expression of a process of individuation (Trewavas, 1999). (b) Structure of a population and its significance. The nucleation-induced process of individuation is generally buffered by the high level of redundancy inherent to regulations existing in a mature network (Amzallag, 2001a). However, the level of redundancy in networks may be considerably reduced in suboptimal conditions (Amzallag, 2001a). In these conditions, a part of the individuality inherent to the first stages of emergence of the phenophase may be maintained. Thus, it may condition the last stages of maturation of the new regulation network. This phenomenon is able to modify the structure of the population. During a phenophase, a disturbing noise is more or less buffered by the network regulating the character analyzed. In these conditions, and for a homogeneous population initially exposed to optimal conditions, the distribution frequency of a parameter remains monomodal, even though the variation around the mean is amplified. During a critical period, the environment becomes a source of information influencing evolution of the system towards maximal dissipation of the anisotropy (see above). Thus, the distribution frequency becomes plurimodal, even in an initially homogeneous population. Such a phenomenon has been described in Sorghum bicolor exposed to salinity during expression of the critical period (Amzallag, 1998). Accordingly, a plurimodal distribution frequency observed in a population does not imply necessarily an initial heterogeneity (genetic or macroenvironmental).

ARTICLE IN PRESS 26

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

In biology, all the statistical tools used for comparison of populations (t-tests, ANOVA) are restricted to comparison of parameters normally distributed in a population. This indicates that the transformations occurring during a critical period (especially for populations exposed to a suboptimal environment) are automatically hidden by the use of these statistical methods. But the phenomenon of individuation prevents the a priori use of statistical tests for comparison populations. Moreover, the sudden emergence of plurimodal distributions in populations exposed to suboptimal conditions may serve as a ‘phenotypic marker’ of emergence of a critical period in development. (c) The role of genes in development. In a non-linear representation of development, the effect of an environmental stimulus is not constant. Being a disturbing factor during a phenophase, an environmental factor may become a source of information during a critical period (see above). In this context, introduction of a new gene product during a critical period represents a new source of information for the emerging dissipative structure. In contrast to exogenous sources of information, this change is well controlled in time, space and intensity by the mechanisms underlying gene expression. Such an accurate regulation enables the formation of concentration gradients generating a dissipative structure. This also enables a precise control of the biochemical environment conditioning its self-emergence. It is not surprising, therefore, that gene expression plays a central role in the reliability of developmental processes. However, this reality does not correspond to a stringent cause-to-effect control. Gene products may hold a transient significance when expressed during a critical period, even for genes previously expressed during a phenophase. This property generates a diversity of functions (morphogenetic, structural and/or metabolic) for the same gene product. Moreover, the notion of pre-existing information cannot be restricted to the gene-expression level. The elements existing before the critical period and maintained after the breakdown characterizing the critical period have many constraints (such as mechanical or metabolic stress) that also condition self-emergence of the new structure (Beloussov et al., 1994). In this context, the notion of pre-existing information remains relative to each stage of development and to each level of complexity. (d) The dual response to stress. According to the phase of development (phenophase or critical period), the significance of an environmental stimulus oscillates from perturbation to source of information. As a consequence, one may expect the coexistence of two types of responses to stress. A resistance response (the expression of mechanisms of homeostasis and/or protection of the initial status) induced during a phenophase, and an adaptation response (including an adaptive resetting of the regulation network) typically expressed during a critical period. Although generally ignored, physiological adaptation has been described in plants (see Amzallag, 1999c; 2002b). In animals, an increase in tolerance to a toxic substance following a pretreatment (during a short period of early development) has also been reported (Saliba and Krzyz, 1976; Pascoe and Beattie, 1979; Fraser, 1980; Dixon and Sprague, 1981a, b; Kerver and Rotman, 1987). In most of these cases, the type of response (resistance or physiological adaptation) cannot be completely

ARTICLE IN PRESS G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

27

identified. However, both the existence of a developmental window for adaptation (opened during the early development) and an increase in variability have been generally reported.

Conclusion Until now, the non-linear dimension of development has been generally ignored or minimized. Our attention was focused on adult biology instead of physiology of the early development, on analysis of expression of pre-existing mechanisms of response to stress instead of adaptive plasticity, on homeostatic regulations instead of hormone imprinting, and on isotropic variability instead of processes of individuation. This constant bias has strengthened the illusion of a simple deterministic link between genotype and phenotype, dismissing something fundamental about life. However, stability and relative autonomy of the biological structures appear here as generated by their nature of dissipative systems. In such a context, the progressive increase in complexity in development and its adaptedness result from the capacity of living systems to induce transition phases before their spontaneous occurrence. In contrast to physical systems, organisms (at least during their development) display an ability to self-define their physical status, from a close-to-isolated to a dissipative system. The dissipative nature of biological systems may also be observed at other scales of organization. For example, sensorial inputs generate a network of synaptic communications (Castelo-Branco et al., 2000). It seems that cerebral representations result from self-emergence of networks of interactions between modules of neurons stimulated by the sensorial perception (Freeman, 1991; Maunsell, 1995; Rodriguez et al., 1999). At the cellular level, pathways of signal transduction are organized in labile networks (Trewavas, 2002). Redundancy and extensive feedbacks in regulatory pathways (generating a network of metabolic interactions) are now considered as the basis of stability and reliability in cell differentiation processes, cellular physiology and cell-to-cell interactions (Fell, 1997; Igamberdiev, 1999; McAdams and Arkin, 1999). Moreover, the metabolic network appears to be organized in small highly connected topological modules combined in a hierarchical manner (Ravasz et al., 2002). Even at the genome level, gene expression is also organized in hierarchically related networks, a phenomenon considered as fundamental for control of cell differentiation (Thieffry et al., 1998; Davidson et al., 2002). In this case, modifications of structure of the network also require transient changes in connectivity (Guet et al., 2002). According to these observations, an organism may be considered as a series of dissipative systems organized in a hierarchical order. From such a reality, it would be very surprising that the non-linear processes inherent to transformations of the networks may concern only the whole organism, the latest level organization. This suggests that the concept of critical period may probably be transposed at the sub-cellular level of organization. By their self-emergent nature, the entities generated during a critical period are conditioned by environmental conditions towards maximum dissipation of the

ARTICLE IN PRESS 28

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

perturbation. This indicates that, in spite of the diversity generated by such a process, all the emergent structures are more or less equally adapted to their environment. Thus, the preferential representation, within a population, of some patterns does not necessarily result from natural selection. Rather, the structure of a population appears to be a consequence of preferential orientation of the selfemergent process towards specific attractors as a function of their stability (Conrad, 1990; Amzallag, 2000b). This observation asks for a reconsideration of the units of evolution. Rather than genes or gene families, the modular entities self-emerging during development become the units of evolution (as previously suggested by Schlosser, 2002). They also become the sources of emergence of novelty (see Muller and Wagner, 1991) through a process very similar to that occurring during critical periods in development (Stadler et al., 2001). This is why a self-organized component independent of the mutation/selection tandem probably operates in evolution (Moseley and Jan, 1997). From all these considerations, it appears that integration of the non-linear phenomenon of critical periods does not invalidate the current paradigms in biology. However, it restricts their validity to phases of stability of the existing structures (phenophases), while the crucial events related to their emergence are issued from another horizon.

Acknowledgements My thanks go to the anonymous reviewer for his fruitful advices and his patience. His efforts contributed to clarify the concepts presented here and their formulation.

References Aleksic, I., Turic, N., 1994. Changes in the component of phenotypic variance and covariance among traits during ontogeny in the sharp-snouted rock lizard (Lacerta oxycephala). Amphibia-Reptilia 15, 47–61. Amzallag, G.N., 1998. Induced modifications in reproductive traits of salt-treated plants of Sorghum bicolor. Israel Journal of Plant Science 46, 1–8. Amzallag, G.N., 1999a. Adaptive nature of the transition phases in development: the case of Sorghum bicolor. Plant Cell and Environment 22, 1035–1042. Amzallag, G.N., 1999b. Individuation in Sorghum bicolor: a self-organized process involved in physiological adaptation to salinity. Plant Cell and Environment 22, 1389–1399. Amzallag, G.N., 1999c. Plant evolution: towards an adaptive theory. In: Lerner, H.R. (Ed.), Response of Plants to Stress: From Phytohormones to Genome Reorganization. Marcel Dekker Inc., New York, pp. 171–245. Amzallag, G.N., 2000a. Connectance in Sorghum development: beyond the phenotype–genotype duality. BioSystems 56, 1–11. Amzallag, G.N., 2000b. Canalization as a non-genetic source of adaptiveness during development. BioSystems 57, 95–107. Amzallag, G.N., 2001a. Data analysis in plant physiology: are we missing the reality? Plant Cell and Environment 24, 881–890. Amzallag, G.N., 2001b. Maturation of integrated functions during development. I. Modification of the meristem network during transitory periods in Sorghum bicolor. Plant Cell and Environment 24, 337–345.

ARTICLE IN PRESS G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

29

Amzallag, G.N., 2001c. Developmental changes in effect of cytokinin and gibberellin on shoot K and Na accumulation in salt-treated Sorghum plants. Plant Biology 3, 1–7. Amzallag, G.N., 2002a. Brassinosteroids as metahormones: evidences from specific influence during critical period in Sorghum development. Plant Biology 4, 656–663. Amzallag, G.N., 2002b. The adaptive potential of plant development: evidence from the response to salinity. In: L.auchli, A., Luttge, . U. (Eds.), Salinity: Environment—Plant-Molecules. Kluwer, The Netherlands, pp. 291–312. Amzallag, G.N., 2002c. Dual response to nutrients in Sorghum bicolor: the role of critical periods in expression of adaptive plasticity. Israel Journal of Plant Science 50, 87–93. Amzallag, G.N., Seligmann, H., Lerner, H.R., 1993. A developmental window for salt-adaptation in Sorghum bicolor. Journal of Experimental Botany 44, 645–652. Arteca, J.M., Arteca, R.N., 2001. Brassinosteroid-induced exaggerated growth in hydroponically grown Arabidopsis plants. Physiologia Plantarum 112, 104–112. Ashby, W.R., 1962. Principles of the self organizing system. In: von Foester, H., Zopf, G.W. (Eds.), Principles of Self Organization. Pergamon Press, Oxford, pp. 255–278. Atchley, W.R., 1987. Developmental quantitative genetics and the evolution of ontogenies. Evolution 41, 316–330. Atchley, W.R., Rutledge, J.J., 1980. Genetic components of size and shape. I. Dynamics of components of phenotypic variability and covariability during ontogeny in the laboratory rat. Evolution 34, 1161–1173. Atlan, H., 1972. L’organisation Biologique et la Th!eorie de l’Information. Hermann, Paris. Barton, N.H., Turelli, M., 1989. Evolutionary quantitative genetics: how little do we know? Annual Review of Genetics 23, 337–370. Bechoux, N., Bernier, G., Lejeune, P., 2000. Environmental effects on the early stages of tassel morphogenesis in maize. Plant Cell and Environment 23, 91–98. Belayev, D.K., 1978. Destabilizing selection as a factor in domestication. Journal of Heredity 70, 301–308. Beloussov, L.V., Saveliev, S.V., Naumidi, I.N., Novoselov, V.V., 1994. Mechanical stresses in embryonic tissues: patterns, morphogenetic role, and involvement in regulatory feedback. International Review of Cytology 150, 1–34. Berg, R.L., 1959. A general evolutionary principle underlying the origin of developmental homeostasis. American Naturalist 93, 103–105. Berg, R.L., 1960. The ecological significance of correlation pleiades. Evolution 14, 171–180. Bradshaw, A.D., 1965. Evolutionary significance of phenotypic plasticity in plants. Advances in Genetics 13, 115–155. Brink, R.A., 1962. Phase change in higher plants and somatic cell heredity. Quarterly Review of Biology 37, 1–22. Callahan, H.S., Waller, D.M., 2000. Phenotypic integration and the plasticity of integration in the amphicarpic annual. International Journal of Plant Sciences. 161, 89–98. Castelo-Branco, M., Goebel, R., Neuenschwander, S., Singer, W., 2000. Neural synchrony correlates with surface segregation rules. Nature 405, 685–689. Chauvet, G.A., 1993. Hierarchical functional organization of formal biological systems: a dynamical approach. I. The increase of complexity by self-association increases the domain of stability of a biological system. Philosophical Transactions of the Royal Society of London Series B 339, 425–444. Cheplick, G.P., 1995. Genotypic variation and plasticity of clonal growth in relation to nutrient availability in Amphibromus scabrivalvis. Journal of Ecology 83, 459–468. Cock, A.G., 1966. Genetical aspects of metrical growth and form in animals. Quarterly Review of Biology 41, 131–190. Conrad, M., 1983. Adaptibility. The Significance of Variability from Molecules to Ecosystems. Plenum Press, New York. Conrad, M., 1985. Scale change and the emergence of information processing primitives. Journal of Theoretical Biology 112, 741–755. Conrad, M., 1990. The geometry of evolution. BioSystems 24, 61–81. Csaba, G., 1994. Phylogeny and ontogeny of hormone receptors: origin and development of hormone receptors. International Review of Cytology 155, 1–48. Davidson, E.H., Rast, J.P., Oliveri, P., Ransick, A., Calestani, C., Yuh, C.H., Minokawa, T., Amore, G., Hinman, V., Arenas-Mena, C., Otim, O., Brown, C.T., Livi, C.B., Lee, P.Y., Revilla, R., Rust, A.G., Pan, Z.J., Schilstra, M.J., Clarcke, P.J.C., Arnone, M.I., Rowen, L., Cameron, R.A., McClay, D.R., Hood, L., Bolouri, H., 2002. A genomic regulatory network for development. Science 295, 1669–1678.

ARTICLE IN PRESS 30

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

Deneubourg, J.L., 1977. Application de l’ordre par fluctuations a" la description de certaines e! tapes de la construction du nid chez les termites. Insectes Soctaux 24, 117–130. Diggle, P.K., 1994. The expression of andromonoecy in Solanum hirtum: phenotypic plasticity and ontogenetic contingency. American Journal of Botany 81, 1354–1365. Dillon, P.F., Root-Berstein, R.S., 1997. Molecular complementarity II: energetic and vectorial basis of biological homeostasis and its implication for death. Journal of Theoretical Biology 188, 481–493. Dixon, D.G., Sprague, J.B., 1981a. Acclimation-induced changes in toxicity of arsenic and cyanide to rainbow trout, Salmo gairdneri. Journal of Fish Biology 18, 579–589. Dixon, D.G., Sprague, J.B., 1981b. Acclimation to copper by rainbow trout (Salmo gairdneri)—a modifying factor in toxicity. Canadian Journal of Fisheries and Aquatic Sciences 38, 880–888. Douady, S., Couderc, Y., 1996. Phyllotaxis as a dynamical self organizing process. Part III: the stimulation of the transient regimes of ontogeny. Journal of Theoretical Biology 178, 295–312. Dujardin, L., Walbaum, S., 1985. Apport de la th!eorie des catastrophes a" la description de la morphogen"ese de Candida albicans. Physiologic V!egetale 23, 309–320. Edelstein-Keshet, L., Watmough, J., Ermentrout, G.B., 1995. Trail following in ants: individual properties determine population behaviour. Behavioral Ecology and Sociobiology 36, 119–133. Ephritikhine, G., Fellner, M., Vannini, C., Lapous, D., Barbier-Brygoo, H., 1999. The sax1 dwarf mutant of Arabidopsis thaliana shows altered sensitivity of growth response to abscisic acid, auxin, gibberellins and ethylene and is partially rescued by exogenous brassinosteroid. Plant Journal 18, 303–314. Fell, D., 1997. Understanding the Control of Metabolism. Portland Press, London. Fonseca, C.R., John, J.L., 1996. Connectance: a role for community allometry. Oikos 77, 353–358. Fraser, J., 1980. Acclimation to lead in the freshwater isopod Asellus aquaticus. Oecologia 45, 419–420. Freeman, W.J., 1991. The physiology of perception. Scientific American 264, 78–85. Gardner, M.R., Ashby, W.R., 1970. Connectance of large dynamic (cybernetic) systems: critical values for stability. Nature 228, 784. Gerhart, J.C., Berking, S., Cooke, J., Freeman, G.L., Hildebrandt, A., Jokusch, H., Lawrence, P.A., Nusslein-Volhard, C., Oster, G.F., Sander, K., Sauer, H.W., Stent, G.S., Wessells, N.K., Wolpert, L., 1982. The cellular basis of morphogenetic change. In: Bonner, J.T. (Ed.), Evolution and Development. Springer, Berlin, pp. 87–114. Glandsdorff, P., Prigogine, I., 1971. Thermodynamics of Structures, Stability and Fluctuations. Wiley, London. Goldbeter, A., Segel, L.A., 1980. Control of developmental transitions in the cyclic AMP signalling system of Dictyostelium discoideum. Differentiation 17, 127–135. Grass!e, P.P., 1984. Termitologia. II: Fondation des Soci!et!es, Construction. Masson, Paris. Guern, J., 1987. Regulation from within: the hormona dilemma. Annals of Botany 60, S75–S102. Guet, C.C., Elowitz, M.B., Hsing, W., Leibler, S., 2002. Combinatorial synthesis of genetic networks. Science 296, 1466–1470. Herbert, J.G., Kidwell, J.F., Chase, A.B., 1979. The inheritance of growth and form in the mouse. IV. Changes in the variance components of weight, tail length and tail width during growth. Growth 43, 36–46. Igamberdiev, A.I., 1999. Foundations of metabolic organization: coherence as a basis of computational properties in metabolic networks. BioSystems 50, 1–16. Keller, E.F., Segel, L.A., 1970. Initiation of slime mold aggregation viewed as an instability. Journal of Theoretical Biology 26, 399–415. Kerver, W.J.M., Rotman, G., 1987. Development of ethanol tolerance in relation to the alcool dehydrogenase locus in Drosphila melanogaster. The influence of phenotypic adaptation and maternal effect on survival on alcohol supplemented media. Heredity 58, 239–248. Khiprach, V.A., Zhabinskii, V.N., de Groot, A.E., 1999. Brassinosteroids. A New Class of Plant Hormones. Academic Press, San Diego. Law, R., Blackford, J.C., 1992. Self-assembling food webs: a global viewpoint of coexistence of species in lokta-volterra communities. Ecology 73, 567–578. Lejeune, P., Bernier, G., 1996. Effect of environment on the early steps of ear initiation in maize. Plant Cell and Environment 19, 217–224. Maldiney, H., 2001. Existence: Crise et Cr!eation. Encre Marine, Foug"eres (France). Martinez, N.D., 1993. Effects of resolution on food web structure. Oikos 66, 403–412. Maunsell, J.H.R., 1995. The brain’s visual world: representation of visual targets in cerebral cortex. Science 270, 764–769. May, R.M., 1972. Will a large complex system be stable? Nature 238, 413–414.

ARTICLE IN PRESS G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

31

McAdams, H.H., Arkin, A., 1999. It’s a noisy business! Genetic regulation at the nanomolar scale. Trends in Genetics 15, 65–69. Millonas, M.M., 1992. A connectionist type model of self-organized foraging and emergent behavior in ant swarms. Journal of Theoretical Biology 159, 529–552. Moseley, L.L., Jan, N., 1997. Essential evolution. Theory in Bioscience 116, 140–144. Muller, G.B., Wagner, G.P., 1991. Novelty in evolution: restructuring the concept. Annual Review of Ecology and Systematics 22, 229–256. Murtaugh, P.A., Kollath, J.P., 1997. Variation of trophic fractions and connectance in food webs. Ecology 78, 1382–1387. Pascoe, D., Beattie, J.H., 1979. Resistance to cadmium by pretreated rainbow trout alevins. Journal of Fish Biology 14, 303–308. Paulsen, S.M., 1994. Quantitative genetics of butterfly wing color patterns. Developmental Genetics 15, 79–91. Paulsen, S.M., Nijhout, H.F., 1993. Phenotypic correlation structure among elements of the color pattern in Precis coenia (Lepidoptera). Evolution 47, 593–618. Paxman, G.J., 1956. Differentiation and stability in the development of Nicotiana rustica. Annals of Botany 20, 47–58. Perfectti, F., Camacho, J.P.M., 1999. Analysis of genotypic differences in developmental stability in Anona cherimola. Evolution 53, 1396–1405. Pigliucci, M., 1996. How organisms respond to environmental changes: from phenotypes to molecules (and vice versa). Trends in Ecology and Evolution 11, 169–173. Poethig, R.S., 1990. Phase change and the regulation of shoot morphogenesis in plants. Science 250, 923–930. Prigogine, I., Wiame, J.M., 1946. Biologie et thermodynamique des ph!enom"enes irr!eversibles. Experentia 2, 451–453. Ravasz, E., Somera, A.L., Mongru, D.A., Oltvai, Z.N., Barabasi, A.L., 2002. Hierarchical organization of modularity in metabolic networks. Science 297, 1551–1555. Rodriguez, E., George, N., Lachaux, J.P., Martinerie, J., Renault, B., Varela, F.J., 1999. Perception’s shadow: long-distance synchronization of human brain activity. Nature 397, 430–433. Sakai, K.I., Shimamoto, Y., 1965. Developmental instability in leaves and flowers of Nicotiana tabacum. Genetics 51, 801–813. Saliba, L.J., Krzyz, R.M., 1976. Acclimation and tolerance of Artemia salina to copper salts. Marine Biology 38, 231–238. Schlosser, G., 2002. Modularity and the units of evolution. Theory in Bioscience 121, 1–80. Schneider, D.W., 1997. Predation and food web structure along a habitat duration gradient. Oecologia 110, 567–575. Shapiro, J.A., 1985. Scanning electron microscope study of Pseudomonas putida colonies. Journal of Bacteriology 164, 1171–1181. Shapiro, J.A., 1987. Organization of developing Escherichia coli colonies viewed by scanning electron microscopy. Journal of Bacteriology 169, 142–156. Shapiro, J.A., Hsu, C., 1989. Escherishi coli K12 cell-cell interactions seen by time-lapse video. Journal of Bacteriology 171, 5963–5974. Sherry, R.A., Lord, E.M., 1996. Developmental stability in flowers of Clarkia temblorienis. Journal of Evolutionary Biology 9, 911–930. Slafer, G.A., Rawson, H.M., 1994. Sensitivity of wheat phasic development to major environmental factors: a re-examination of some assumptions made by physiologists and modelers. Australian Journal of Plant Physiology 21, 393–426. Slutsky, B., Buffo, J., Soll, D.R., 1985. High-frequency switching of colony morphology in Candida albicans. Science 230, 666–669. Sokal, R.R., Rohlf, F.J., 1981. Biometry, 2nd Edition. Freeman, San Francisco. Stadler, B.M.R., Stadler, P.F., Wagner, G.P., Fontana, W., 2001. The topology of the possible: formal spaces underlying patterns of evolutionary change. Journal of Theoretical Biology 213, 241–274. Stearns, S.C., 1982. The role of development in the evolution of life histories. In: Bonner, J.T. (Ed.), Evolution and Development. Springer, Berlin, pp. 237–258. Sultan, S.E., 1992. Phenotypic plasticity and the neo-Darwinian legacy. Evolutionary Trends in Plants 6, 61–71. Sultan, S.E., 1995. Phenotypic plasticity and plant adaptation. Acta Botanica Neerlandica 44, 363–383. Swaddle, J.P., Witter, M.S., 1997. On the ontogeny of developmental stability in a stabilized trait. Proceedings of the Royal Society of London B 264, 329–334.

ARTICLE IN PRESS 32

G.N. Amzallag / Theory in Biosciences 123 (2004) 17–32

Takeuchi, Y., Omigawa, Y., Ogasawara, M., Yoneyama, K., Kpnnai, M., Worsham, A.D., 1995. Effects of brassinosteroids on conditioning and germination of clover broomrape (Orobanche minor) seeds. Plant Growth Regulation 16, 153–160. Terentjev, P.V., 1931. Biometrische untersuchungen uber die morphologishen merkmale von Rana ridibunda Pall. Biometrika 23, 23–51. Thieffry, D., Huerta, A.M., Perez-Rueda, E., Collado-Vives, J., 1998. From specific gene regulation to global regulatory networks: a characterization of Escherichia coli transcriptional network. BioEssays 20, 433–440. Trewavas, A.J., 1991. How do plant growth substances work? Plant Cell and Environment 14, 1–12. Trewavas, A.J., 1999. The importance of individuality. In: Lerner, H.R. (Ed.), Plant Response to Stresses: From Phytohormones to Genome Reorganization. Marcel Dekker Inc., New York, pp. 27–50. Trewavas, A.J., 2002. Plant signal transduction: the emerging phenotype. Plant Cell (Suppl. 2002), S3–S4. Trewavas, A.J., 2003. Aspects of plant intelligence. Annals of Botany 92, 1–20. Turing, A.M., 1952. The chemical basis of morphogenesis. Philosophical Transactions of the Royal Society of London Series B 237, 37–72. Via, S., Gomulkiewicz, R., De Jong, G., Scheiner, S.M., Schlichting, C.D., van Tienderen, P.H., 1995. Adaptive phenotypic plasticity: consensus and controversy. Trends in Ecology and Evolution 10, 212–216. von Bertalanffy, L., 1950. The theory of open systems in physics and biology. Science 111, 23–29. Waddington, C.H., 1961. Genetic assimilation. Advances in Genetics 10, 257–293. Weyers, J.D.B., Paterson, N.W., A’Brook, R., Peng, Z.Y., 1995. Quantitative analysis of the control of physiological phenomena by plant hormones. Physiologia Plantarum 95, 486–494. Zhirmunsky, A.V., Kuzmin, V.I., 1988. Critical Levels in the Development of Natural Systems. Springer, Berlin.