Journal of Experimental Botany, Vol. 51, No. 342
MP Special Issue, pp. 41–50, January 2000
Using stable isotope natural abundances (d15N and d13C)
to integrate the stress responses of wild barley (Hordeum
spontaneum C. Koch.) genotypes
D. Robinson1, L.L. Handley, C.M. Scrimgeour, D.C. Gordon, B.P. Forster and R.P. Ellis
Scottish Crop Research Institute, Dundee DD2 5DA, UK
Received 3 February 1999; Accepted 10 May 1999
Abstract
To integrate the complex physiological responses of
plants to stress, natural abundances (d) of the stable
isotope pairs 15N/14N and 13C/12C were measured in 30
genotypes of wild barley (Hordeum spontaneum C.
Koch.). These accessions, originating from ecologically
diverse sites, were grown in a controlled environment
and subjected to mild, short-term drought or
N-starvation. Increases in total dry weight were
paralleled by less negative d13C in shoots and, in
unstressed and droughted plants, by less negative
whole-plant d13C. Root d15N was correlated negatively
with total dry weight, whereas shoot and whole-plant
d15N were not correlated with dry weight. The difference in d15N between shoot and root varied with stress
in all genotypes. Shoot–root d15N may be a more sensitive indicator of stress response than shoot, root or
whole-plant d15N alone. Among the potentially most
productive genotypes, the most stress-tolerant had the
most negative whole-plant d15N, whether the stress
was drought or N-starvation. In common, controlled
experiments, genotypic differences in whole-plant d15N
may reflect the extent to which N can be retained
within plants when stressed.
Key words: Hordeum spontaneum, d13C, d15N, stress,
drought, nitrogen.
Introduction
There is a continuing search for crops tolerant of harsh
environments. One strategy is statistical, and involves
examining plants from contrasting habitats, correlating
their stress responses to habitat characteristics and—in
suitable mapping populations—to molecular markers on
the genome ( Forster et al., 1997; Ellis et al., 1997; Handley
et al., 1997). This approach can reveal genotypic and
phenotypic facets of stress tolerance, and their
interactions. In principle, genes associated with specific
physiological processes involved in stress tolerance can
be identified and introgressed into breeding lines
(Holmberg and Bülow, 1998).
‘Stress tolerance’ comprises many physiological processes which vary quantitatively rather than qualitatively
( Yeo, 1998; Zhang et al., 1999). It is often impractical to
measure each process individually on many plants. One
solution is to measure surrogate variables which integrate
many physiological processes. Some of the most useful
of these surrogates are the natural abundances (denoted
as d) of biologically important stable isotope pairs, e.g.
13C/12C and 15N/14N.
d13C has been used to screen C genotypes for potential
3
water use efficiency ( Ehleringer et al., 1993). A robust
theory is available ( Farquhar et al., 1982) with which to
interpret d13C variations among C plants in terms of
3
measurable physical and physiological processes. Plant
d13C reflects mainly the extent to which primary CO
2
assimilation is limited by carboxylation and/or CO
2
diffusion in leaves. Whole-plant d13C is dominated by
these processes. Internal partitioning and metabolism of
primary assimilate may produce differences in d13C among
plant organs (Hubick and Gibson, 1993) and chemical
groups (Gleixner et al., 1993; Brugnoli et al., 1998;
Schmidt and Kexel, 1998). Environmental stresses (e.g.
drought) modify d13C in largely predictable ways, explicable ultimately via effects on the balance between stomatal
conductance and carboxylation.
In contrast, d15N has been used much less extensively
in this way. Plant d15N reflects the potentially variable
1 To whom correspondence should be addressed: Fax: +44 1 382 562426. E-mail: D.Robinson@scri.sari.ac.uk
© Oxford University Press 2000
42 Robinson et al.
d15N values of external N sources and 15N/14N fractionations which occur during the assimilation, transport and
loss of N. The use of d15N in plant ecophysiology is
currently at the ‘pattern generation’ or ‘hypothesis development’ stage. Taxonomic and environmental variations
in d15N are being explored and documented in natural
and controlled environments (Handley and Scrimgeour,
1997; Handley et al., 1998), despite there being no theory
able to explain these variations mechanistically. A theory
has been proposed for d15N (Robinson et al., 1998)
which, despite being restricted to NO−-grown plants, still
3
demanded information about the d15N values of external
and internal N pools, information that is difficult to
obtain routinely. The ‘decoding’ of plant d15N into underlying mechanisms promises to be a non-trivial problem.
An alternative is to find statistical associations between
plant d15N, growth and specified environmental conditions, and to establish testable hypotheses about the main
cause(s) of such associations. Genotypic and environmental variations in plant d15N exist. For example,
Handley et al. (1997) showed that shoot d15N varied by
up to 2.4‰ among wild barley (Hordeum spontaneum C.
Koch.) genotypes grown on a common N source. Salinity
caused shoot d15N to become, on average, 2‰ more
negative. But to begin interpreting plant d15N physiologically, data for whole plants, not just shoots or roots, are
required.
The purpose here is to explore further the utility of
d15N as a physiological integrator. Specifically, the aims
are to (1) measure the variations in shoot, root and
whole-plant d15N in genotypes of one species (H. spontaneum) in relation to experimentally imposed environmental stresses and to site-of-origin conditions; (2)
correlate these measurements with stress tolerance; and
(3) assess the potential usefulness of d15N as an integrator
of stress responses.
Materials and methods
Caryopses of H. spontaneum plants collected from 30 sites
(Table 1) in the Fertile Crescent, the centre of diversity for this
species, were bulked under glasshouse conditions. Surfacesterilized caryopses were germinated on moist filter paper in
Petri dishes on 13 September 1996. Three days later, when
roots were 5–6 mm long, seedlings were transplanted into openended Sarsted tubes filled with 0.8% agar containing 200 mg l−1
benzimidazole to suppress fungal pathogens. To minimize root
damage and to prevent hypoxia around the embryo, a 3 mm
diameter core of agar was removed from each tube, into which
a seedling was inserted, and its roots covered immediately with
cold agar extruded via a syringe. Seedlings were grown in a
controlled environment (4 °C, lit by high-pressure sodium lamps
at 300 mmol m−2 s−1, 8 h daylength) for 7 weeks’ vernalization,
during which time the base of each tube was kept in water.
Vernalized seedlings were transferred, on 4 November 1996,
into a glasshouse hydroponic system. Air temperature was
maintained between 16–24 °C, and the glasshouse was ventilated
with outside air to ensure steady [CO ] and d13C of source CO
2
2
(c. −8‰). Plants were illuminated by natural daylight,
supplemented by sodium lamps. The hydroponic system consisted of three 80 l troughs containing aerated half-strength
Hewitt’s nutrient solution (Hewitt, 1966) with additional
NaSiO (Epstein, 1994), changed weekly. N, as Ca(NO ) and
3
3 2
KNO , was supplied at 6 mol m−3 with an initial mean d15N
3
value in solution of +1.4±0.2‰. This value became more
positive between solution changes, reaching +2.2 to +4.5‰.
The causes of the gradual 15N enrichment are unknown, but
could reflect the loss from roots of partly assimilated, 15Nenriched N (Robinson et al., 1998), partial denitrification of
NO− in the non-sterile solution (Robinson and Conroy, 1999),
3
or both. Mean solution temperature was 16±0.1 °C; mean pH
was 6.0±0.1. Solution [O ] at the end of the experiment (when
2
O depletion would have been greatest) was 92±0.3% satura2
tion. During the experiment, the mean outdoor solar radiation
receipt was ~3 MJ m−2 d−1 (DKL Mackerron, personal communication); about half this amount would have reached the
plants growing in the glasshouse.
Three experimental treatments were established: controls, in
which plants were maintained in the nutrient solution throughout; drought, in which plants were, 7 d after transfer to the
hydroponic system, raised out of the solution to expose their
roots to air for 3 h daily (Hendry, 1993); and N starvation, in
which plants were deprived of all external N after 7 d growth.
In the N starvation treatment, Ca2+ and K+ were supplied as
CaCl .6H O and K SO , respectively, giving both cations a
2
2
2 4
uniform concentration of 2 mol m−3 in all treatments. The
drought and N starvation treatments were deliberately mild,
intended to reveal the diversity of sub-lethal stress responses
among the H. spontaneum genotypes. Genotypes were replicated
four times per treatment, and arranged in randomized blocks
in each trough. Each trough contained 170 plants, of which 50
were guard plants of H. vulgare cv. Derkardo.
Plants were harvested 16 d after transfer to the nutrient
solution, long after seed C and N (and their d values) had been
trivialized in the whole plants (on average, 8 and 0.4 mg C and
N per seed, versus at least 90 and 7 mg C and N per harvested
plant). Shoots were separated from roots. Plant material was
oven-dried (60 °C for 48 h), weighed and milled. Concentrations
of total C and N, and d13C and d15N were determined on
subsamples (c. 1 mg dry wt) of shoots and roots using
continuous-flow isotope ratio mass spectrometry ( Europa
Scientific Ltd., Crewe, UK ), as described (Handley et al., 1993;
Scrimgeour and Robinson, 1999). d values (‰) were calculated
as
R
–R
(1)
d= sample standard×1000
R
standard
where R is the ratio of 13C/12C or 15N/14N.
Data were subjected to a two-way ANOVA with Genotype
and Treatment as factors. It was unnecessary to transform data
to homogenize variances. Statistical analyses were done with
Genstat v. 5 (Genstat 5 Committee, 1993) and StatisticaTM v.
5.1 (StatSoft, Norman, Oklahoma) software.
Whole-plant d15N (‰) was calculated as an average of shoot
and root d15N weighted by the total N contents (mg) of shoots
and roots:
Whole-plant d15N
=
(Shootd15N×ShootN )+(Rootd15N×RootN )
ShootN+RootN
(2)
A dimensionless ‘stress index’ (SI ) was calculated for each
genotype to account for innate size differences among genotypes
Stress responses of wild barley
43
Table 1. Sites of origin, locations and code numbers of the Hordeum spontaneum genotypes
Meteorological data for the sites are given in Table 1 of Pakniyat et al. (Pakniyat et al., 1997).
Site of origin
Site description
Longitude
(°E )
Latitude
(°N )
Altitude
(m)
Code
Iran
Gates of Zagros
Gawdar (18 km W )
Gawdar (29 km W )
Gawdar (2 km E )
Ilam (35 km SW )
Mehran
Pol-e Dokhtar (7 km S)
Shahabad-e Gharb
Symarch (8 km S)
High mountains
Mountains, humid
Mountains, semi-humid
Mountains, humid
Foothills, semi-arid
Foothills, dry/warm
Hills, arid, dry/warm
Mountains, humid
Mountains, semi-humid
46.12
46.39
46.33
46.47
46.24
46.28
47.73
46.49
47.22
34.35
34.00
33.70
34.04
33.17
33.42
33.39
34.05
33.59
900
1700
1300
1300
600
500
900
1400
1100
9
22
11
29
19
30
21
24
4
Basalt plain, semi-arid
Coastal plain, humid
Mountains, humid
Mountains, semi-arid
Mountains, humid
Hills, semi-dry, humid
Hills, semi-humid, shaded,
deep soil
Hills, semi-humid, open,
deep soil
Hills, semi-humid, open,
shallow soil
Hills, semi-dry
Hills, semi-humid
Mountains, semi-dry
35.70
34.90
35.00
35.75
35.40
35.02
35.11
32.78
32.50
31.50
33.28
33.05
32.43
32.44
325
10
800
1530
1150
60
150
8
6
15
18
23
5
25
35.11
32.44
150
28
35.11
32.44
150
1
35.53
35.53
35.23
32.90
32.90
31.75
35
35
800
2
27
7
Mountains
Mountains
Basalt plateau,
Basalt plateau,
Basalt plateau,
Basalt plateau,
High plateau
Basalt plateau
Hills
41.72
40.25
39.83
40.75
37.40
36.82
39.44
39.25
38.36
38.07
37.84
37.86
38.12
37.09
37.18
37.91
37.70
37.20
780
860
1200
730
800
730
950
620
600
20
10
17
3
26
16
12
14
13
Israel
Afiq
Caesarea
Canada Park
Mt. Hermon
Mt. Meron
Nahal Oren
Neve Ya’ar (a)
Neve Ya’ar (b)
Neve Ya’ar (c)
Tabigha (basalt)
Tabigha (terra rossa)
Talpiyot
Turkey
Bitlis (60 km S)
Diyarbakir (10 km S)
Diyarbakir (38 km W )
Diyarbakir (50–60 km NE road)
Gaziantep (3 km W )
Gaziantep (50 km W )
Siverek (20 km E )
Siverek (9 km E )
Urfa (7 km E )
semi-arid
semi-arid
semi-arid
semi-arid
in their stress responses:
W
−W
stressed
(3)
SI= unstressed
W
unstressed
where W
and W
are the mean dry weight per plant
unstressed
stressed
(mg) of unstressed and stressed plants, respectively. The SI can
range from 0 to 1. SI values A 0 represent little stress, plants
becoming increasingly stressed—in terms of the effect of the
environment on growth—as SI A 1.
Results
Plant growth
Total dry weight per plant varied significantly (P<0.001)
with both Treatment and Genotype, but there was
no Genotype×Treatment interaction (Table 2). Twelve
genotypes showed significant reductions in total dry
weight, relative to controls, in response to drought; five
genotypes responded significantly to N starvation
(Fig. 1). Of those genotypes which did not respond to
either stress, almost all had relatively slow growth rates
(as indicated by their ranking in Fig. 1, i.e. Genotypes 15
to 11). Exceptions to this were Genotypes 7 and 8 which,
notably, were unaffected by N starvation despite being
two of the most productive genotypes.
Whole-plant d15N
There were highly significant effects of Genotype and
Treatment on whole-plant d15N (P≤0.001; Table 2).
Variations in whole-plant d15N were dominated by the
Treatment main effect, but the interaction was significant.
Whole-plant d15N in controls ( Fig. 2) ranged from
−0.9‰ (Genotype 12) to +0.3‰ (Genotype 3). Given
that source NO− had a d15N value >+1‰, the plants
3
clearly discriminated against 15N (where discrimination
~ source d15N–whole-plant d15N ). However, discrimination could not be quantified because of the temporal
variability in source d15N (see Materials and methods).
Whole-plant d15N responded significantly (P<0.05) to
drought in 16 of the 30 genotypes; 10 genotypes responded
significantly to N starvation ( Fig. 2). When drought or
44 Robinson et al.
Table 2. Summary analyses of variance of the measured characters of the H. spontaneum genotypes (Geno.) in the three treatments (Trt.)
Geno.
Trt.
Geno.×Trt.
Total dry weight
Whole-plant d15N
Shoot d15N
Root d15N
F
P
F
P
F
P
F
12.7
39.4
1.22
<0.001
<0.001
0.234
2.73
53.2
1.80
<0.001
<0.001
0.001
2.38
32.6
1.78
<0.001
<0.001
<0.001
7.08
1120
3.18
Fig. 1. Mean whole-plant dry weight of control (&), droughted (#)
and N-starved (6) H. spontaneum genotypes. Genotypes are ranked in
order of increasing total dry weight of controls. Code numbers ( Table 1)
of genotypes which responded significantly (P<0.05, LSD) to drought
are underlined on the right-hand axis; those which responded significantly to N starvation are underlined on the left-hand axis. For clarity,
error bars have been omitted from data points. SEs were, on average,
0.15, 0.11 and 0.13 g per plant for control, droughted and N-starved
plants, respectively.
N starvation had a significant (P<0.05) effect on wholeplant d15N, this almost always became more negative
than in controls (Fig. 2); the only exception to this was
Genotype 5. The mean difference in d15N between control
and stressed plants was 0.6 (±0.1 SE) and 0.3 (±0.1 SE)
‰ for the drought and N starvation treatments, respectively. The largest such difference under drought was 1.4‰
in Genotype 15 and 1.1‰ when plants were N-starved
(Genotype 26).
Shoot–root d15N
Whole-plant d13C
P
F
P
F
P
<0.001
<0.001
<0.001
4.94
830
2.80
<0.001
<0.001
<0.001
13.6
360
1.71
<0.001
<0.001
0.003
Fig. 2. Mean whole-plant d15N of control (&), droughted (#) and N
starved (6) H. spontaneum genotypes. Genotypes are ranked in order
of increasing whole-plant d15N of controls. Code numbers ( Table 1) of
genotypes which responded significantly (P<0.05, LSD) to drought are
underlined on the right-hand axis; those which responded significantly
to N starvation are underlined on the left-hand axis. For clarity, error
bars have been omitted from data points. The SE was, on average,
0.3‰ in all treatments.
Shoot and root d15N
Shoot d15N, root d15N, and the difference between them,
varied significantly with Genotype and Treatment, and
there was a significant interaction between these (Table 2).
Drought and N starvation caused shoot d15N to
become, on average, significantly more negative than in
controls, N starvation having a larger effect ( Table 3).
Drought caused root d15N to become, on average, 2.1‰
Stress responses of wild barley
45
Table 3. Mean (±SE) d15N and d13C values (‰), and total N
concentrations (% dry weight) of roots and shoots in control,
droughted and N-starved H. spontaneum plants
Data comprise measurements of all 30 genotypes. Means in the same
row followed by different letters are significantly different (LSD,
P<0.001, n=30).
Shoot d15N
Root d15N
Shoot d13C
Root d13C
Shoot [N ]
Root [N ]
Control
Drought
N starvation
−0.1±0.1 a
−0.8±0.1 a
−31.9±0.1 a
−30.8±0.1 a
5.5±0.04 a
4.2±0.05 a
−0.4±0.1 b
−2.9±0.1 b
−30.2±0.1 b
−30.3±0.1 b
3.8±0.07 b
2.6±0.03 b
−0.6±0.1 c
−0.1±0.1 c
−32.0±0.1 a
−30.7±0.1 a
2.4±0.05 c
1.5±0.02 c
more negative than controls. This trend was reflected in
all genotypes except one (Genotype 2) whose root d15N
was unaffected by drought (data not shown). In contrast,
N starvation caused root d15N to become significantly
less negative than the controls.
All genotypes responded significantly (P<0.05) to one
or both stresses in terms of the shoot–root d15N difference
(Fig. 3). In most genotypes, the difference in shoot–root
d15N was slightly positive in controls (mean 0.7±0.1‰),
i.e. shoots were more 15N-enriched than roots. In response
to drought, shoot–root d15N usually increased (mean
2.5±0.1‰) compared with the control. By contrast,
shoot–root d15N usually decreased under N starvation
(mean −0.5±0.1‰). When the absolute shoot–root
difference in d15N between control and drought treatments
was plotted against that between control and N starvation
treatments, there was a significant, inverse relation
between them (Fig. 4).
There were no significant correlations between shoot
d15N and shoot N content or concentration, nor between
root d15N and root N content or concentration (data
not shown).
d13C
Whole-plant d13C varied significantly (P<0.001) with
Genotype and Treatment, and there was a strong interaction (Table 2). As with whole-plant d15N, Treatment
had the dominant effect on d13C. In controls, whole-plant
d13C varied by 1.8‰, ranging from −32.4‰ (Genotype
14) to −30.6‰ (Genotype 6). Shoot and root d13C were
both affected significantly by drought, but not N starvation ( Table 3). Under drought, d13C became less negative
than in controls and this effect was more pronounced in
shoots than roots.
There were no significant correlations between d13C
and d15N in shoots or roots (data not shown).
Correlations between plant dry weight and d values
The heaviest plants had the least negative shoot d13C
values (P<0.05; Table 4). Root and whole-plant d13C
Fig. 3. Mean shoot–root differences in d15N of control (&), droughted
(#) and N starved (6) H. spontaneum genotypes. Genotypes are
ranked in order of increasing shoot–root d15N of controls. Code
numbers ( Table 1) of genotypes which responded significantly (P<0.05,
LSD) to drought are underlined on the right-hand axis; those which
responded significantly to N starvation are underlined on the lefthand axis.
values varied similarly with total dry weight in control
and droughted plants, but not when plants were
N-starved. Shoot d15N was not correlated with total dry
weight, although the heaviest plants had the most negative
root d15N values (P<0.05; Table 4). Absolute shoot–root
d15N differences in droughted or N-starved plants were
correlated positively with total dry weight (P<0.05;
Table 4).
Whole-plant d15N was not correlated with total dry
weight in any treatment ( Table 4). Many genotypes,
however, showed no dry weight response to either stress
( Fig. 1). Excluding the least productive and least responsive genotypes (i.e. Genotypes 15 to 11 in the drought
treatment, and 15 to 5 in the N starvation treatment:
Fig. 1), revealed significant positive correlations between
whole-plant d15N and the stress index, SI (P<0.05, n=
18 for drought and n=8 for N starvation; Fig. 5). The
least stressed of those plants (SI0) had the most negative
whole-plant d15N. In droughted plants, this relation was
due to shoot d15N, root d15N showing no correlation with
the SI. In N-starved plants, however, both shoot and
46 Robinson et al.
Fig. 4. Absolute differences in shoot–root d15N between control and
N-starved plants in relation to those between control and droughted
plants. Genotype numbers (Table 1) are shown against each symbol.
The regression y=2.43–0.679x is significant at P<0.05 (r=−0.72,
n=30).
Table 4. Correlations (Pearson product-moment coefficients, r)
between total dry weights in control, droughted or N-starved
plants and d13C and d15N values, and total N concentrations
measured in those treatments
n.s.=Non-significant (P>0.05) correlations.
Total dry weight per plant
Shoot d13C
Root d13C
Whole-plant d13C
Shoot d15N
Root d15N
Whole-plant d15N
Shoot–root d15N
Shoot [N ]
Root [N ]
Control
Droughted
N-starved
0.57
0.42
0.55
n.s.
−0.49
n.s.
n.s.
n.s.
0.53
0.52
0.50
0.53
n.s.
−0.38
n.s.
0.51
−0.38
0.48
0.48
n.s.
n.s.
n.s.
−0.42
n.s.
0.42
−0.59
0.53
root d15N were significantly (P<0.05) correlated with the
SI. In contrast, neither shoot nor root d13C was correlated
with the SI (data not shown).
Correlations between plant dry weight and total N
concentrations in shoots and roots
Total dry weight of control plants was always correlated
positively (P<0.05) with the concentration of total N in
root dry matter, but not in shoots ( Table 4). When plants
were stressed, however, total dry weight became negatively correlated with total N concentration in shoots: the
Fig. 5. Mean whole-plant d15N values under (a) drought and (b) N
starvation in relation to the stress index (SI) for those treatments.
Genotype numbers ( Table 1) are shown against each symbol. Genotypes
were those which, when unstressed, grew at least as fast as the slowestgrowing genotype which responded to stress (see Fig. 1). For (a), the
regression y=1.17+1.14x is significant (r=0.47, P<0.05, n=18). For
(b), the regression y=−0.984+1.85x is significant (r=0.93, P<0.05,
n=8).
heaviest plants in the stress treatments had the smallest
total N concentrations in their shoots.
Correlations between isotope natural abundances and
long-term meteorological averages
Of the long-term meteorological averages available for
the sites-of-origin (Pakniyat et al., 1997), only mean
humidity at 14.00 h local time was correlated consistently
with any isotopic data ( Table 5). The more humid the
site-of-origin, the less negative were the experimentally
determined d13C values of genotypes from those sites.
Absolute shoot–root d15N values of control and
N-starved plants were also correlated positively with
humidity at site-of-origin, as was shoot d15N under N
starvation. Only in droughted plants was d13C correlated
with mean annual and mean January temperatures. Mean
January temperature was correlated positively with shoot,
root and whole-plant d15N in the N starvation treatment.
The only isotopic measurement with which mean annual
rainfall was correlated was root d15N of controls.
Stress responses of wild barley
47
Table 5. Correlations (Pearson product-moment coefficients, r) between d13C and d15N
values and long-term meteorological averages for sites-of-origin (Table 1 in Pakniyat
et al., 1997)
Blank cells indicate non-significant correlations (P>0.05). MAT, MAuT and MJaT: mean annual,
August and January temperatures (°C ), respectively; MAR: mean annual rainfall (mm). Humid:
mean humidity (%) at 14.00 h local time.
MAT
Control plants
Shoot d13C
Root d13C
Whole-plant d13C
Shoot d15N
Root d15N
Whole-plant d15N
Shoot–root d15N
Droughted plants
Shoot d13C
Root d13C
Whole-plant d13C
Shoot d15N
Root d15N
Whole-plant d15N
Shoot–root d15N
MAuT
MJaT
MAR
Humid
0.49
0.67
0.59
−0.39
−0.37
−0.56
0.41
0.46
0.50
0.49
N-starved plants
Shoot d13C
Root d13C
Whole-plant d13C
Shoot d15N
Root d15N
Whole-plant d15N
Shoot–root d15N
Discussion
Variations in whole-plant d15N in relation to drought and N
starvation
The 1.3‰ range in whole-plant d15N in controls ( Fig. 2)
indicates the extent to which genotype may influence
15N/14N fractionations in H. spontaneum when plants
have access to a common N source. By comparison,
whole-plant d13C varied by 1.8‰ among controls, a larger
range than that found for well-watered genotypes of
H. vulgare (Acevedo, 1993); greater variability is to be
expected among individuals from wild populations than
among those from genetically narrower breeding lines.
Whole-plant d15N may vary significantly, therefore, for
reasons unconnected to changes in the d15N values of
external N source(s).
15N/14N discriminations between whole plants and an
external N source can have only one general cause: the
loss from plants of some isotopically altered N. There is
no experimental or theoretical evidence that N uptake
itself fractionates 15N/14N. Possible mechanisms of N loss
include: the shedding of senescent plant parts; the loss of
N volatiles (e.g. NH , NO , amines, HCN: Wetselaar and
3
x
Farquhar, 1980) from leaves into the atmosphere; and
the loss of soluble N (e.g. NO−, amino acids: Jones and
3
Darrah, 1993; Van der Leij et al., 1998) from roots into
0.45
0.38
0.45
0.42
0.41
0.43
0.48
0.57
0.51
0.48
0.52
0.52
0.43
0.38
the rooting medium. Each of these is considered as a
possible explanation for the data in Fig. 2.
Senescence can be discounted: the young H. spontaneum
plants showed little visible leaf senescence. In an annual
species such as H. spontaneum, there is little root senescence until synchronous mortality occurs towards the end
of an individual’s life, as in Triticum aestivum L. ( Van
Vuuren et al., 1997).
NH -N lost from organic matter can have d15N values
3
down to −40‰ (Handley et al., 1996, 1999). Such a net
loss of 14N will be reflected by an increase of whole plant
d15N in proportion to the fraction of total plant N lost,
is most likely to occur via stomata, and to increase with
stomatal conductance (which should, in turn, cause shoot
d13C to become more negative). Therefore, if N volatilization is significant, shoot d15N should increase as shoot
d13C becomes more negative. No such relation was found
for H. spontaneum, and N volatilization was unlikely to
have caused the variations in whole-plant d15N shown in
Fig. 2. A similar conclusion was reached by other authors
(Bergersen et al., 1988; Robinson and Conroy, 1999) who
were unable to attribute variations in whole-plant d15N
of Glycine max (L.) Merrill or Panicum coloratum L.,
respectively, to N volatilization.
The loss of soluble N from roots does occur, as does
its partial resorption (Jones and Darrah, 1993). As with
48 Robinson et al.
gaseous N losses, little is known of their magnitude or
d15N values relative to whole-plant d15N (Schmidt and
Kexel, 1998). Robinson et al.’s theory suggested that the
loss from roots of organic-N which was less 15N-enriched
than total N was consistent with the d15N values measured
for other N pools (e.g. root and shoot total N and
NO−) ( Robinson et al., 1998). This is not conclusive
3
evidence but, pro tempore, N exudation seems the most
likely determinant of variations in whole-plant d15N
among the H. spontaneum genotypes under the conditions
of the experiment. Direct tests of this possibility are now
being done.
In a common, controlled environment in which source
N is defined and for which a good estimate of its d15N
value exists, genotypic differences in whole-plant d15N
values reflect the extent to which plants retain N in their
tissues. Agronomic interest in stress tolerance does not
usually concern genotypes whose tolerance involves slow
growth rates and, probably, small yields. Rather, the aim
is to identify genotypes which have the potential to grow
well should conditions allow, and to produce economically and nutritionally acceptable yields when conditions
are unfavourable. In Fig. 5, the genotypes which were
most productive and stress tolerant (i.e. were heaviest in
comparison with their potential growth when unstressed,
as indicated by their small SI values) were those which
probably retained most N. Yet, those genotypes expressed
the largest discriminations against 15N, i.e. they had the
most negative whole-plant d15N values. According to
isotope mass balance arguments, if those genotypes lost
only small amounts of N, that N must have had ‘exotic’
d15N values significantly different from total plant N
( Yoneyama, 1995; Schmidt and Kexel, 1998). It may be
that such plants, when stressed, restrict the loss of N
from their roots to only one or two amino acids, say,
which happen to have ‘exotic’ d15N values (NO− exuda3
tion from the roots of N-starved barley is negligible: Van
der Leij et al., 1998).
Conversely, genotypes expressing the smallest discriminations against 15N were smaller, contained less N and,
perhaps, lost relatively more N from their tissues. They
may have less capacity to restrict N loss from their roots
when stressed than did more stress-tolerant genotypes.
Lost N would then comprise a diverse mixture of N
compounds with a correspondingly wide range of d15N
values. The average d15N of lost N would then be closer
to that of the total N, resulting in smaller whole-plant
discriminations against 15N. These intriguing possibilities
also require explicit testing.
Variations in d values in relation to conditions at sites-oforigin
No consistent or strong associations were found between
plant d15N measured under the experimental conditions
and habitat characteristics ( Table 1) or long-term meteorological averages (Table 5). d13C values, by contrast,
varied consistently with site-of-origin humidity. The association between less negative d13C (measured at a common
ambient vapour pressure deficit; vpd) and greater site-oforigin humidity is opposite to that found between ambient
vpd and 13C discrimination in several C species
3
(Madhavan et al., 1991; Masle et al., 1993). As commonly
reported for other plant species, Handley et al. found
that shoot d13C was most negative in H. spontaneum
genotypes from sites receiving the least rainfall annually
(Handley et al., 1994). That correlation was not found
in this hydroponic experiment.
A strong inverse relation (r=−0.59, P<0.001) has
been found between mean annual rainfall and siteaveraged foliar d15N for a wide range of ecosystems
(Handley et al., 1999). Such was the variability in d15N,
however, that many samples were required to reveal
significant correlations with environmental factors.
Discrepancies between samples collected from the field
and those produced in common, controlled environments
have been well-documented for d13C (Condon and
Richards, 1993); similar constraints will, no doubt, apply
to research with d15N.
Some genotypes which originated from one locality
expressed similar d15N responses to stress, while others
from another site showed very different responses. For
example, the Tabigha genotypes (2 and 27) showed
opposite shoot–root d15N responses to drought and N
starvation (Fig. 4), despite there being no significant
difference in their growth or whole-plant d15N whether
stressed or unstressed (Figs 1, 2). By contrast, the three
Neve Ya’ar genotypes (1, 25 and 28) had very similar
shoot–root d15N responses in the two stress treatments
( Fig. 4). While Genotypes 1 and 28 grew similarly (Fig. 1)
and had similar whole-plant d15N values ( Fig. 2),
Genotype 25 differed from these in shoot–root d15N when
droughted, but not when N-starved.
The utility of plant d15N as a physiological integrator
Although the statistically significant correlations between
d15N and total plant dry weight were not particularly
strong ( Table 4), they were of similar magnitude to some
which have been measured between total dry weight and
d13C (Condon and Richards, 1993). The correlations
between whole-plant d15N and SI, especially for N starvation, but less convincingly for drought (Fig. 5), suggest
that causal links exist between d15N and stress tolerance.
The possibility that, in controlled experiments, wholeplant d15N may reflect net N retention suggests that
whole-plant d15N could be used to screen plants for this
agriculturally and ecologically important trait. It will not
be easy to apply such an idea in field settings because of
the uncertainties which continue to surround the identity
Stress responses of wild barley
and d15N value(s) of plant-available N species in soil
(Handley and Scrimgeour, 1997; Handley et al., 1999).
Whole-plant d15N in droughted or N-starved H. spontaneum was always more negative than in controls,
confirming the report for shoot d15N of hydroponically
grown plants in response to salinity (Handley et al.,
1994). The whole-plant d15N of many genotypes did not
respond to either drought or N starvation, whereas
shoot–root d15N did. Shoot–root d15N may, therefore, be
a more sensitive indicator of incipient stress than shoot,
root or whole-plant d15N.
Acknowledgements
The Scottish Crop Research Institute receives grant-in-aid from
the Scottish Office Agriculture, Environment and Fisheries
Department. Winnie Stein, Sigrun Holdus and Richard Keith
provided technical help, Professor Eviatar Nevo the
H. spontaneum genotypes, Dr Donald Mackerron solar radiation
data, and two anonymous referees helpful comments.
References
Acevedo E. 1993. Potential of carbon isotope discrimination as
a selection criterion in barley breeding. In: Ehleringer JR,
Hall AE, Farquhar GD, eds. Stable isotopes and plant carbon–
water relations. London: Academic Press, 399–417.
Bergersen FJ, Peoples ML, Turner GL. 1988. Isotopic discriminations during the accumulation of nitrogen by soybeans.
Australian Journal of Plant Physiology 15, 407–20.
Brugnoli E, Scartazza A, Lauteri M, Monteverdi MC, Máguas
C. 1998. Carbon isotope discrimination in structural and
non-structural carbohydrates in relation to productivity and
adaptation to unfavourable conditions. In: Griffiths H, ed.
Stable isotopes. Oxford: Bios Scientific Publishers, 133–146.
Condon AG, Richards RA. 1993. Exploiting genetic variation in
transpiration efficiency in wheat: an agronomic view. In:
Ehleringer JR, Hall AE, Farquhar GD, eds. Stable isotopes
and plant carbon–water relations. London: Academic Press,
435–450.
Ehleringer JR, Hall AE, Farquhar GD. (eds) 1993. Stable
isotopes and plant carbon–water relations. London:
Academic Press.
Ellis RP, Forster BP, Waugh R, Bonar N, Handley LL, Robinson
D, Gordon DC, Powell W. 1997. Mapping physiological traits
in barley. New Phytologist 137, 149–157.
Epstein E. 1994. The anomaly of silicon in plant biology.
Proceedings of the National Academy of Sciences, USA
91, 11–17.
Farquhar GD, O’Leary MH, Berry JA. 1982. On the relationship
between carbon isotope discrimination and the intercellular
carbon dioxide concentration in leaves. Australian Journal of
Plant Physiology 9, 121–137.
Forster BP, Russell JR, Ellis RP, Handley LL, Robinson D,
Hackett CA, Nevo E, Waugh R, Gordon DC, Keith R, Powell
W. 1997. Locating genotypes and genes for abiotic stress
tolerance in barley, a strategy using maps, markers and the
wild species. New Phytologist 137, 141–147.
Genstat 5 Committee. 1993. Genstat 5 Release 3 Reference
manual. Oxford: Clarendon Press.
Gleixner G, Danier H-J, Werner RA, Schmidt H-L. 1993.
Correlations between the 13C content of primary and
49
secondary plant products in different cell compartments and
that in decomposing basdiomycetes. Plant Physiology 102,
1287–1290.
Handley LL, Austin AT, Robinson D, Scrimgeour CM, Heaton
THE, Raven JA, Schmidt S, Stewart GR. 1999. The 15Nnatural abundance (d15N ) of ecosystem samples reflects
measures of water availability. Australian Journal of Plant
Physiology 26, 185–199.
Handley LL, Brendel O, Scrimgeour CM, Schmidt S, Raven JA,
Turnbull MH, Stewart GR. 1996. The 15N natural abundance
patterns of field-collected fungi from three kinds of ecosystems. Rapid Communications in Mass Spectrometry 10,
974–978.
Handley LL, Daft MJ, Wilson J, Scrimgeour CM, Ingleby K,
Sattar MA. 1993. Effects of the eco- and VA-mycorrhizal
fungi Hydnagium carnaeum and Glomus clarum on the d15N
and d13C values of Eucalyptus globulus and Ricinus communis.
Plant, Cell and Environment 16, 375–382.
Handley LL, Nevo E, Raven JA, Martı́nez-Carrasco R,
Scrimgeour CM, Pakniyat H, Forster BP. 1994. Chromosome
4 controls potential water use efficiency (d13C ) in barley.
Journal of Experimental Botany 45, 1661–1663.
Handley LL, Robinson D, Forster BP, Ellis RP, Scrimgeour
CM, Gordon DC, Nevo E, Raven JA. 1997. Shoot d15N
correlates with genotype and salt stress in barley. Planta
201, 100–102.
Handley LL, Scrimgeour CM. 1997. Terrestrial plant ecology
and 15N natural abundance: the present limits to interpretation for uncultivated systems with original data from a
Scottish old field. Advances in Ecological Research 27,
133–212.
Handley LL, Scrimgeour CM, Raven JA. 1998. 15N at natural
abundance levels in terrestrial vascular plants: a précis. In:
Griffiths H, ed. Stable isotopes. Oxford: Bios Scientific
Publishers, 89–98.
Hendry GAF. 1993. Drought tolerance. In: Hendry GAF,
Grime JP, eds. Methods in comparative plant ecology. London:
Chapman and Hall, 53–54.
Hewitt EJ. 1966. Sand and water culture methods used in the
study of plant nutrition, 2nd edn. Farnham: Commonwealth
Agricultural Bureaux.
Holmberg N, Bülow L. 1998. Improving stress tolerance in
plants by gene transfer. Trends in Plant Science 3, 61–66.
Hubick KT, Gibson A. 1993. Diversity in the relationship
between carbon isotope discrimination and transpiration
efficiency when water is limited. In: Ehleringer JR, Hall AE,
Farquhar GD, eds. Stable isotopes and plant carbon–water
relations. London: Academic Press, 311–325.
Jones DL, Darrah PR. 1993. Influx and efflux of amino acids
from Zea mays L. roots and their implications for N nutrition
and the rhizosphere. Plant and Soil 155/156, 87–90.
Madhavan S, Treichel I, O’Leary MH. 1991. Effects of relative
humidity on carbon isotope fractionation in plants. Botanica
Acta 104, 292–294.
Masle J, Shin JS, Farquhar GD. 1993. Analysis of restriction
fragment length polymorphisms associated with variation of
carbon isotope discrimination among ecotypes of Arabidopsis
thaliana. In: Ehleringer JR, Hall AE, Farquhar GD, eds.
Stable isotopes and plant carbon–water relations. London:
Academic Press, 371–386.
Pakniyat H, Powell W, Baird E, Handley LL, Robinson D,
Scrimgeour CM, Nevo E, Hackett CA, Caligari PDS, Forster
BP. 1997. AFLP variation in wild barley (Hordeum spontaneum C. Koch) with reference to salt tolerance and
associated ecogeography. Genome 40, 332–341.
50 Robinson et al.
Robinson D, Conroy JP. 1999. A possible plant-mediated
feedback between elevated CO , denitrification and the
2
enhanced greenhouse effect. Soil Biology and Biochemistry
31, 43–53.
Robinson D, Handley LL, Scrimgeour CM. 1998. A theory for
15N/14N fractionation in nitrate-grown vascular plants. Planta
205, 397–406.
Schmidt H-L, Kexel H. 1998. Metabolite pools and metabolic
branching as factors of in vivo isotope discrimination by
kinetic isotope effects. Isotopes in Environmental and Health
Studies 34, 19–30.
Scrimgeour CM, Robinson D. 1999. Stable isotope analyses and
applications in soil and environmental science. In: Smith KA,
Cresser MS, eds. Soil and environmental analysis. New York:
Marcel Dekker (in press).
Van der Leij M, Smith SJ, Miller AJ. 1998. Remobilisation of
vacuolar stored nitrate in barley root cells. Planta 205, 64–72.
Van Vuuren MMI, Robinson D, Fitter AH, Chasalow SD,
Williamson L, Raven JA. 1997. Effects of elevated atmospheric
CO and soil water availability on root biomass, root length,
2
and N, P and K uptake by wheat. New Phytologist
135, 455–465.
Wetselaar R, Farquhar GD. 1980. Nitrogen losses from tops of
plants. Advances in Agronomy 33, 263–302.
Yeo A. 1998. Molecular biology of salt tolerance in the context
of whole-plant physiology. Journal of Experimental Botany
49, 915–929.
Yoneyama T. 1995. Nitrogen metabolism and fractionation of
nitrogen isotopes in plants. In: Wada E, Yoneyama T,
Minagawa M, Ando T, Fry B, eds. Stable isotopes in the
biosphere. Kyoto: Kyoto University Press, 92–102.
Zhang J, Nguyen HT, Blum A. 1999. Genetic analysis of
osmotic adjustment in crop plants. Journal of Experimental
Botany 50, 291–302.