Transmission of acquired adjustments to salinity in Sorghum bicolor

ELSEVIER

BioSystems 40 (1997) 257-261

Transmission

of acquired adjustments

to salinity in Sorghum

bicolor Hervk Seligmann* Department

oJ’Evolution,

Systematics

and Ecology,

The Hebrew

University of’Jerusalem, 91904 Jerusalem, Israel

Received 21 March 1996

Abstract Previous studies showed that exposure of eight-day-old Sorghum bicolor for three weeks to sublethal salinity induces an increase in salinity tolerance, called physiological adaptation (A). During A, plants of a same population differ in reaction and tolerance to salinity. Tolerance levels of the reaction types depend on environmental conditions besides salinity. Reactions observed most frequently in an experiment have generally highest tolerance levels. This phenomenon is defined adaptive determinism (AD). In this study, the relationship between a potential source of the information subjacent to AD and AD itself is analysed in plants first exposed to salt-inducing A. When the reaction types are close variations of one reaction mode, AD is highest. This relationship is inversed in progeny of adapted plants. Results suggest that information relevant to AD is transmitted to the progeny of adapted plants, and that adaptive information is created during A in plants first exposed to adaptation inducing treatment. Copyright 0 1997 Elsevier Science Ireland Ltd Keywords:

Transmittance

of acquired

character;

Adaptation;

1. Introduction

Amzallag et al. (1990) showed that a three-week exposure of S-d-old seedlings to 150 mM NaCl induces the capacity to grow at lethal salinity in Sorghum bicolor (S610). The phenomenon was termed adaptation (A). Plants not exposed to this

*Corresponding author. Tel.: 972 2 6585874; Fax: 972 2 6584741; E-mail: [email protected]

Adaptive

determinism;

Sorghum

hicoior

pretreatment do not survive at 300 mM NaCl. The NaCl environment induces A, but the induction is dependent on the developmental stade of the plant, as in Amzallag et al. (1993). Because close to 90% of the plants survive exposure to lethal salinity, A does not result from a selection on pre-existing, within-genotype variability. The induction of increased salinity tolerance, a physiological adaptation, is genotype dependent (Amzallag and Lerner (1994)). In Amzallag et al. (1995), the amplitude of A is shown to be propor-

0303-2647/97/$17.00 Copyright 0 1997 Elsevier Science Ireland Ltd. All rights reserved PII SO303-2647(96)01653-X

Salinity;

258

H. Seligmann / BioSystems 40 (1997) 2.57-261

tional to a general increase in variability of the morphogenic characters within the genotype during the period of A. This phenomenon suggests that A is not pre-defined, and that A does not only result from the expression of pre-existing information. Seligmann et al. (1993) described a perturbation of the expansion of some of the leaves which unfold during the three weeks of maturation of A. This malformation, termed DPL, was observed on some of the fifth to the ninth leaves. The upper part of the developing leaf is crinkled, and it does not unroll. Sometimes, the next developing leaf is trapped by the rolled part of the older leaf. Because the number of DPLs on plants relates positively to their growth capacity at 300 mM NaCl (Amzallag et al., 1993; Seligmann et al., 1993), DPLs are morphological markers of the amplitude of A. Even within a sample of plants exposed to homogenous conditions, plants differ in the serial order of development of leaves with a perturbed development, as well as in the total number of DPLs. This serial order defines DPL patterns, which are different reaction types (RT) to salinity. The determinism index is the correlation coefficient between the frequencies of RTs in a sample and an index of salinity tolerance of these plants (senescence of old leaves). The frequencies of the RTs vary widely among experiments, and even among replicates of a same experimental set. However, there was generally a positive correlation between the frequency of RT and its tolerance, as estimated by a senescence index at the end of A. This indicates that in most conditions, in a population of adapting plants, RTs with the higher tolerance levels to the precise environmental conditions are most likely to develop, a property termed by Seligmann and Amzallag (1995) as adaptive determinism (AD). The origin of the information underlying AD is unknown. Seligmann and Amzallag (1995) consider as unlikely that all the adaptive information subjacent to AD, precisely tuned according to the encountered environment, is pre-existing in the plant. RT and senescence are linked, and this link is a plausible cause of AD. Therefore, I expect that the amount of variation in senescence explained

by RTs should be correlated to the determinism index. Because Amzallag (1994) showed that the progeny of adapted plants (Al) have a higher level of pre-adaptation to salinity than their parent (AO) (fertilization in Sorghum is by selfing), we expect that high AD is observed in populations where variation in senescence is explained by RT, and that this correlation will be stronger in Al than AO.

2. Materials

and methods

Populations of 30 plants of Sorghum bicolor (variety MP610, inbred line, male parent of the commercial variety 610) were transfered to hydropony in half-strength, modified Hoagland solution, eight days following germination, as described in Amzallag et al. (1990). After the development of the third leaf, they were exposed to 150 mM NaCl by a daily increase of 25 mM NaCl. Part of the leaves which develop after exposure of NaCl are developmentally perturbed (DPL). DPLs are observed on the fifth to the ninth leaf. All the leaves are not developmentally perturbed. Groups of plants are defined according to which leaf is a DPL. All combinations of DPLs can be observed on plants, even when exposed to uniform conditions. Within homogenous populations of 30 plants growing in the same container, plants with the same pattern of DPL (i.e., with DPL occurring on the same leaves) were pooled as subpopulations, or reaction types to salinity, RTs. For each RT, I determined the mean number of dead leaves each day during three weeks of exposure to 150 mM NaCl. The tolerance of a RT is considered inversely correlated to senescence. For each population of 30 plants, the frequencies of the RTs were plotted versus their mean number of dead leaves. The correlation coefficients of these plots quantify a determinism index (DI) of the frequency of RT in the population in relation to the senescence of RT in the environment of the sample. The similarity in DPL reactions between pairs of plants is expressed as the percentage of leaves with similar development (with or without DPL),

H. Se&maim / BioSystems 40 (1997) 257-261

for all combinations of plants in a sample of 30 plants. The correlation between these similarity indexes and the physical distance between the plants in the sample growing unit was tested by Mantel tests, in 13 samples among A0 samples, and five Al samples. The proportion of RT-related senescence variation was estimated by the F statistic, the ratio between variation in senescence between RTs and the total amount of senescence variation in the whole sample. A low F reveals uniformity in leaf senescence between RTs. High F reveals strong linkage between leaf senescence and RTs. Plants exposed for the first time to salt-adaptation, termed AO, were grown and selfed in the presence of 150 mM NaCl until complete seed maturation. An equal number of seeds was taken from each plant, in order to prevent over-representation due to individual fertility differences. The seeds were pooled, germinated, and exposed to the same adaptation treatment as for their parents. The offspring were termed Al plants.

259

growing unit. Therefore, I tested the similarity in DPL of pairs of plants as explained by the physical distance between the two plants. In none of the 18 samples tested were the correlations to distance between plants close to significance, not even at the 5% significance level. This verification

80

A .

40 0 -40 -80

3. Results The DI in most samples of A0 and Al was negative. DI was plotted versus F for samples of A0 and Al (Fig. 1A and lB, respectively). A significant positive correlation (Spearmann rank correlation, P < 0.01) appears for AO. Conversely, a significant negative correlation (Spearmann rank correlation, P < 0.05) is observed in Al populations (Fig. IB). DI was recalculated in all Al and 30 A0 samples, after exclusion of RTs with extreme senescence values. These exclusions increase the homogeneity in senescence of the sample, and cause a decrease of F. In AO, AD showed a significant tendency to decrease (P < 0.05, sign test, AD decreased in 17, was unchanged in 7, and increased in 6 samples). In Al, AD increased in 7 replicates, in 3 was unchanged and it decreased in 3 others. Calculation of AD involves the supposition that samples react as unities, and that intra-sample variations (in DPL, for example), are not related to environmental micro-variations inside a sample

F Fig. 1. Relationship between DI and differentiation between RTs (F). 1A: Relationship for A0 (six independent experiments). IB: Relationship for progeny from A0 plants (Al plants, two independent experiments). 37 populations of 30 plants of Sorghum bicolor (male parent of the commercial cultivar 610) were exposed to 150 mM NaCl from day 8 following germination. For each population, F and DI were calculated three weeks after beginning of salinization. RTs are defined by their DPL patterns. DI is the least-square r correlation coefficient between RT frequency and the mean senescence of each RT. F was calculated as detailed in Text, and indicates the divergence in senescence between the RTs in a sample. It was predicted that in case large differences between RTs exist (large F), AD would be observed (negative DI values). Spearmann rank correlation rS = 0.61, P < 0.01 for A0 plants. Spearmann rank correlation, rS = - 0.55, P c 0.05 for Al plants. The lines represent the least square logarithmic and linear regression equations (Fig 1A and IB, respectively) for illustrative purposes.

260

H. Seligmann / BioSystems 40 (1997) 257-261

indicates that the plants inside samples were probably grown in homogenous conditions, and that within a sample, inter-plant variation has to be explained by causes other than micro-environmental variation.

4. Discussion Seligmann and Amzallag (1995) defined adaptive determinism as a negative correlation between RT frequency and RT senescence during A. F estimates the link between RT and senescence. The development of DPLs, which define RTs, is intimately related to A. In the samples where inter-RT variation in senescence is low (low F), the different RTs are part of one senescence mode. The interpretation of results on variation in AD between different samples already indicates that adaptive determinism is observed when the environmental factors to which the plants react during adaptation involve effects above a minimal threshold. In cases where the levels of the factors are intermediate, the low ADS which are obseved might result from a lack of homogeneity in the conditions to which the plants react, as a result of intermediate levels of the environmental factors. Because the analyses for correlation between reaction type and plant position did not show any correlation, it seems that the lack of homogeneity in the conditions to which plants react are not due to spatial heterogeneity. It is also possible that in the cases where there is no adaptive determinism, the physiological characteristics of the different RTs are closely related. This situation makes likely that at the beginning of the adaptation process, each individual plant might have developed into any of these very close RTs, because all the RTs are variations of a same mode. When F is high, RTs express more than one mode of senescence, and it is likely that at the beginning of the process, some variation between plants already exist, so that not any plant might have developed into any RT. DI is mostly negative, in both A0 in Al, which confirms that the more frequent RTs are those with the lowest senescence rate. It was expected

that in cases there are large differences in senescence between RTs, the information underlying the link between senescence and RTs also adaptively directs the development of the individual plant towards RTs with low senescence, resulting in AD. The assumption which underlies this expectation is that the origin of the differences in senescence causes AD, the process which directs plant development towards RTs with low senescence. In A0 the prediction of this model was not confirmed by experimental results (Fig. 1A). The results suggested that adaptive determinism exists at the condition that all RTs are very similar. It is in the cases that RTs form a homogenous group that we observe high AD. I suggest that AD observed in these conditions is due to the fact that (1) each individual plant is able to develop into any RT, matching the condition of homogeneity, and that (2) the probability of developing into an RT is increased by the tolerance level of the RT in the precise environmental conditions. In the case that condition (1) is not met, RTs are not variations of one senescence mode, DI increases, because part of the plants are directed towards RTs with highest tolerance within one mode, while another part of the population is directed towards other RTs, within another mode of senescence. In these conditions, still all the plants might be adapted; however, AD, a negative correlation between senescence of RTs and RT abundance, will not be observed, because the sample is not homogenous. This model is confirmed by a computational experiment where DI is calculated after exclusion of RTs with extreme values. F drops, as a result of the exclusion of RTs with extreme senescence. DI often decreases after this ‘homogenization’ of RTs in AO. Plants seem, by some unknown way, to predict the future senescence related to RTs, and to preferentially develop into RTs with the lower senescences, measured by low AD values. Because the phenomenon exists especially in cases where senescence in RTs is similar, the plants are sensitive to very minute differences. In Al, in contrast to AO, the predictions are met: there exists a positive correlation between F and AD. In Al, the higher frequency of RTs with lower senescence is observed when options between separated RTs exist (high J’). This indicates

H. Seligmann

/ BioSystems

that AD is due to pre-existing information on modes coping differently with the environment, and not to the ability of plants to compare between expected senescences of closely related options inside one mode. The recalculations of DI after exclusion of RTs with extreme senescences confirm this: DI does not decrease, and even tends to increase. Amzallag et al. (1995) showed that salinity tolerance of Al is higher than of AO. Such phenomena of transmittance of adaptive information have already been described in other plant systems, such as, for example, in the reaction of Lactuca scaricola and of clipper barley to day length and hormones, as reported by Guttermann et al. (1975) and Cottrell et al. (1982) respectively, and in bean to temperature, by Moss and Mullet (1982). The relationship between DI and F in Al indicates that the relationship between senescence and DPL is adaptive, affecting positively AD (Fig. 1B). The comparison between A0 and Al confirms that part of the adaptive response to salinity induced in A0 was accurately transmitted to Al, and that AD might result from de novo, environmentally induced and directed, transmittable adaptive modifications of AO. AD is observed under most combinations of environmental conditions with salinity. Therefore, Seligmann and Amzallag (1995) suggested that not all information underlying the phenomenon is pre-existing, because it is unlikely that precise adaptive information for any combination of secondary factors and salinity exist in the glycophyte Sorghum bicolor. The positive inter-genotype correlation between the increase in morphogenic character variation and A (Amzallag et al., 1995)

40 (1997) 257-261

261

is a further, independent indication that the adaptive response of Sorghum bicolor to salinity is not entirely pre-programmed. Such evidences suggest emergence of information during A. References Amzallag, G.N., 1994, Influence of parental NaCl treatment on salinity tolerance of offspring in Sorghum bicolor. New Phytol. 128, 7155723 Amzallag. G.N. and Lerner, H.R., 1994. Adaptation versus pre-existing resistance, an inter-genotype analysis of the response of Sorghum bicolor to sanity. Israel J. Plant Sci. 42, 1255141 Amzallag, G.N., Lerner, H.R. and Poijakoff-Mayber. A., 1990, Induction of increased salt tolerance in Sorghum to high salinity. J. Exp. Bot. 41, 29-34. Amzallag. G.N., Seligmann, H. and Lerner, H.R.. 1993, A developmental window for salt-adaptation in Sorghum bicolor. J. Exp. Bot. 44, 6455652. Amzallag, G.N., Seligmann, H. and Lerner, H.R., 1995, Induced variability during the process of adaptation in Sorghum hicolor. J. Exp. Bot. 46, lOl7- 1024. Cottrell, J.E., Dale, J.E. and Jeffcoat, B., 1982, The effect of day length and treatment with gibberellic acid on spikelet initiation and development in clipper barley. Ann. Bot. 50, 57-68. Guttermann. Y., Thomas, T.H. and Heydecker, W., 1975, Effect on the progeny of applying different daylength and hormone treatments to parent plants of Lactuca scariola. Physiol. Plant. 34, 30-38. Moss, G.1. and Mullet, J.H., 1982, Potassium release and seed vigour in germinating bean seed as influenced by temperature over previous five generations. J. Exp. Bot. 33. 1147 1160. Seligmann, H. and Amzdllag, G.N., 1995, Adaptive determinism during salt-adaptation in Sorghum bidor. BioSystems 36, 71 -77. Seligmann, H., Amzallag G.N. and Lerner H.R., 1993, Perturbed leaf development in Sorghum bicolor exposed to salinity: a marker of transition towards adaptation. Aust. J. Plant Physiol. 20, 2433249.