Nitrogen in cell walls of sclerophyllous leaves accounts for little of the variation in photosynthetic nitrogen-use efficiency.

Abstract

Photosynthetic rate per unit nitrogen generally declines as leaf mass per unit area (LMA) increases. To determine how much of this decline was associated with allocating a greater proportion of leaf nitrogen into cell wall material, we compared two groups of plants. The first group consisted of two species from each of eight genera, all of which were perennial evergreens growing in the Australian National Botanic Gardens (ANBG). The second group consisted of seven Eucalyptus species growing in a greenhouse. The percentage of leaf biomass in cell walls was independent of variation in LMA within any genus, but varied from 25 to 65% between genera. The nitrogen concentration of cell wall material was 0.4 times leaf nitrogen concentration for all species apart from Eucalyptus, which was 0.6 times leaf nitrogen concentration. Between 10 and 30% of leaf nitrogen was recovered in the cell wall fraction, but this was independent of LMA. No trade-off was observed between nitrogen associated with cell walls and the nitrogen allocated to ribulose 1.5-bisphosphate carboxylase/oxygenase (Rubisco). Variation in photosynthetic rate per unit nitrogen could not be explained by variation in cell wall nitrogen.

Full text

Page 1

Nitrogen in cell walls of sclerophyllous leaves accounts
for little of the variation in photosynthetic
nitrogen-use efficiency
MATTHEW T. HARRISON1,2, EVERARD J. EDWARDS1*, GRAHAM D. FARQUHAR1, ADRIENNE B. NICOTRA2&
JOHN R. EVANS1
1Environmental Biology Group, Research School of Biological Sciences,The Australian National University, Canberra,ACT
2601,Australia, and2School of Botany and Zoology,The Australian National University, Canberra,ACT 0200,Australia
ABSTRACT
Photosynthetic rate per unit nitrogen generally declines as
leaf mass per unit area (LMA) increases.To determine how
much of this decline was associated with allocating a greater
proportion of leaf nitrogen into cell wall material, we com-
pared two groups of plants.The first group consisted of two
species from each of eight genera, all of which were peren-
nial evergreens growing in the Australian National Botanic
Gardens (ANBG). The second group consisted of seven
Eucalyptusspeciesgrowinginagreenhouse.Thepercentage
of leaf biomass in cell walls was independent of variation in
LMA within any genus,but varied from 25 to 65% between
genera.The nitrogen concentration of cell wall material was
0.4 times leaf nitrogen concentration for all species apart
from Eucalyptus,which was 0.6 times leaf nitrogen concen-
tration.Between 10 and 30% of leaf nitrogen was recovered
in the cell wall fraction, but this was independent of LMA.
No trade-off was observed between nitrogen associated
with cell walls and the nitrogen allocated to ribulose 1·5-
bisphosphate carboxylase/oxygenase (Rubisco). Variation
in photosynthetic rate per unit nitrogen could not be
explained by variation in cell wall nitrogen.
Key-words:cell wall nitrogen;leaf mass per unit area;nitro-
gen allocation; Rubisco; structural nitrogen.
INTRODUCTION
The photosynthetic capacity of a leaf is generally well-
correlated with leaf nitrogen content. Although this rela-
tionship varies between species, much of the variation is
related to another leaf parameter, specific leaf area (SLA),
the projected leaf area per unit leaf dry mass. Thus, there
exists a global function, regardless of life-form or location,
which can predict photosynthetic capacity per unit leaf dry
mass from nitrogen concentration and SLA (Reich,Walters
& Ellsworth 1997; Wright et al. 2004). Photosynthetic rate
per unit nitrogen [photosynthetic nitrogen-use efficiency
(PNUE)] tends to decrease as SLA decreases (Poorter &
Evans 1998; Hikosaka 2004). Because a smaller SLA is
associated with greater leaf longevity, Field & Mooney
(1986) suggested that there may be a trade-off between
investing nitrogen in photosynthetic proteins such as ribu-
lose 1·5-bisphosphate carboxylase/oxygenase (Rubisco)
versus compounds required for longevity.
This hypothesis languished for lack of measurements of
structural nitrogen in leaves.However,Onoda,Hikosaka &
Hirose (2004) and Takashima, Hikosaka & Hirose (2004)
developed methods for extracting detergent-soluble pro-
teins from leaf material. They assumed that the nitrogen
that remained behind represented cell wall protein. The
comparison between evergreen and deciduous Quercus
species (Takashima et al. 2004) revealed a clear trade-off
between nitrogen invested in Rubisco and cell wall pro-
teins.Leaves from evergreen Quercus had greater leaf mass
per unit area (LMA,the reciprocal of SLA) and allocated a
greater proportion of leaf nitrogen to cell wall protein than
leaves from deciduous Quercus. Leaves of Polygonum cus-
pidatum also allocated a greater proportion of leaf nitro-
gen to cell walls as LMA increased (Onoda et al. 2004).
However, for a given LMA, Polygonum allocated a smaller
proportion of nitrogen to cell walls than Quercus. While
both of these genera provide support for the hypothesis put
forward by Field & Mooney (1986),the maximum LMA for
leaves from both of these studies was only 60 g m-2.This is
at the lower end of the range reported by Reich et al.
(1997), and so may not be representative of sclerophyllous,
long-lived leaves.
Ellsworth et al. (2004) analysed leaf photosynthesis
from 16 species with LMA ranging from 50 to 300 g m-2.
They calculated that the proportion of nitrogen allocated
to Rubisco declined as LMA increased, and suggested that
this was related to the need for greater investment in struc-
tural nitrogen.Clearly,there is a need for more data on cell
wall nitrogen. Therefore, our first objective was to sample
leaves from species representing a broad range of LMA to
see whether the proportion of nitrogen allocated to cell
walls was related to LMA. Two sampling strategies were
used. Firstly, pairs of species from each of eight genera
Correspondence: J. R. Evans. Fax: +61 2 6249 4919; e-mail:
john.evans@anu.edu.au
*Present address: CSIRO Plant Industry, Private Mail Bag,
Merbein,Vic. 3500,Australia.
Plant, Cell and Environment (2009) 32, 259–270 doi: 10.1111/j.1365-3040.2008.01918.x
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd
259

Page 2

growing in the Australian National Botanic Gardens
(ANBG) were chosen on the basis of contrasting LMA.
These allowed phylogenetically independent contrasts to
be made (Felsenstein 1985). Secondly, seven Eucalyptus
species were grown and measured in a greenhouse to
enable more comprehensive analyses of a single genus over
a fourfold range in LMA.
Another feature of the relationship between PNUE and
LMA is that for a given LMA, PNUE varies by an order of
magnitude. It is likely that the nitrogen concentration in
leaf structural biomass varies between species,and that this
could account for some of the scatter.Therefore,our second
objective was to assess how much of the variation in PNUE
was associated with variation in the proportion of leaf nitro-
gen allocated to cell walls.
MATERIALS AND METHODS
Morphological and physiological measurements were
obtained from two independent investigations:a field study
comparing species pairs from eight genera and a green-
house experiment that examined seven species of the genus
Eucalyptus.
Plant material
A field study was conducted using two C3species from each
of eight perennial Australian evergreen genera growing
in the ANBG, Canberra (35°12″S, 149°04″E). The genera
were selected so as to provide a wide range of LMA, and
thus included a variety of growth forms including vines,
shrubs and trees (Table 1).
The greenhouse study was conducted at the Australian
National University, Canberra, using 12-month-old seed-
lings of Eucalyptus bridgesiana, Eucalyptus elata, Eucalyp-
tus mannifera, Eucalyptus moorei, Eucalyptus pauciflora,
Eucalyptus polyanthemos and Eucalyptus rossii,which were
purchased from theYarralumla Nursery (Canberra).At the
nursery,seedlings were grown in potting mix that contained
a controlled-release fertilizer (18% nitrogen), applied at a
rate of 3 kg m-3. All species were grown in full sunlight on
the same site. Seedlings were transplanted from nursery
tubes into large plastic pots (180 ¥ 180 ¥ 240 mm3; length ¥
width ¥ depth) filled with a sterilized sand/peat/perlite
mixture on 1 May 2006.
Growth conditions and experimental design of
the greenhouse study
Five blocks containing 14 eucalypts were arranged in the
greenhouse, with each block containing two replicates of
each species. Greenhouses were maintained at 22–25 °C
during the day and 15–18 °C at night. Supplementary light-
ing [280 mmol photosynthetically active radiation (PAR)
photons m-2s-1] was provided by six 150 W flood lamps
between 0500–1000 and 1700–2100 h in order to extend day
length.Midday PAR measured with a quantum light sensor
(Li-Cor Inc., Lincoln, NE, USA) averaged 500 mmol PAR
photons m-2s-1on sunny days. Seedlings were watered to
field capacity, twice daily. Rorison’s nutrient solution
(Hewitt 1966) was applied twice per week to each seedling,
100 mL from 1 May to 16 June,and 625 mL from 17 June to
1 July, to increase the size and growth rate of young leaves.
For five of the replicates of each species, one per block, the
4 mm Ca(NO3)2in the Rorison’s solution was replaced with
4 mm CaCl2,providing plus- and minus-nitrogen treatments,
respectively. To distinguish between new and pre-existing
leaves, white tags were attached to the youngest petiole on
the main stem prior to the start of the experiment.Seedlings
were periodically sprayed with chemicals for control of
psyllids and powdery mildew.
Gas exchange measurements
Measurements of CO2assimilation rate per unit area (Aa)
in theANBG were made at saturating irradiance (Table 1),
which was determined for each species by first measuring a
light response curve. Measurements were made on leaves
for all species except Acacia, where phyllodes were used,
from 28 March 2006 to 24 April 2006, using an infrared
gas analyser (IRGA) (LI-6400, Li-Cor) open gas exchange
system.Where possible,a second LI-6400 was used on adja-
cent leaves of the same plant, allowing cross-checks for
consistency in measurement. Steady-state measurements
were made on similar fully expanded young leaves at
an ambient CO2 concentration (Ca) of 375 mmol mol-1,
between 0900 and 1500 h on sunny days. Leaf temperature
was allowed to follow ambient conditions, which ranged
between 15 and 32 °C.
Photosynthetic light response curves were measured for
all seven greenhouse-grown eucalypt species. A standard
irradiance of 1800 mmol PAR photons m-2s-1was adopted
for all species during the measurement of CO2 response
curves, with the block temperature maintained at 22 °C, a
flow rate of 500 mmol s-1and the relative humidity of air
entering the leaf chamber maintained between 70 and 80%.
Following equilibration, Aawas measured at ambient CO2
(375 mmol mol-1) and a series of nine consecutive CO2con-
centrations from 50 to 1300 mmol mol-1.As with theANBG
measurements, two LI-6400s were used.
At the time of measurement,only E.elata,E.bridgesiana,
E. mannifera and E. polyanthemos plants had grown fully
expanded, new leaves.There were no new leaves on minus-
nitrogen E. polyanthemos plants at this time. To allow
contrasts between all species, at least four pre-existing,
non-necrotic leaves on different plants from the plus-
nitrogen treatment were measured.
Scaling photosynthesis to a common Ci value
The model developed by Farquhar, von Caemmerer &
Berry (1980) was used to scale all Aa measurements to
a common intercellular CO2 partial pressure (Ci) of
300 mmol mol-1(Aa300).This approach was carried out using
260
M. T. Harrison et al.
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 3

Table 1. Leaf attributes of the congeners studied in the Australian National Botanic Gardens (ANBG)
Family
Genus
Species1
Growth form
LMA (g m-2)
Nm(mmol g-1)
Aa(mmol m-2s-1)
PNUE
[mmol (mol N)-1s-1]
Ci/Ca
D (‰)
Fabaceae
Acacia
beckleribTindale
Shrub
212 ? 28*
1.71 ? 0.19 ns
9.1 ? 2.3 ns
50 ? 13 ns
0.52 ? 0.05*
22.3 ? 0.7**
implexabBenth.
Tree
126 ? 6
1.72 ? 0.05
9.1 ? 0.5
46 ? 3
0.72 ? 0.03
25.9 ? 0.3
Proteaceae
Banksia
blechnifoliafF. Muell.
Shrub
491 ? 17**
0.64 ? 0.04 ns
17.5 ? 4.1*
102 ? 18 ns
0.46 ? 0.05 ns
19.2 ? 0.4*
serrataaLinn. f.
Tree
166 ? 6
0.54 ? 0.07
5.6 ? 1.0
137 ? 10
0.37 ? 0.02
20.7 ? 0.3
Myrtaceae
Eucalyptus
paucifloraaSieber ex Spreng.
Tree
246 ? 10**
0.79 ? 0.04 ns
6.1 ? 0.5 ns
63 ? 6**
0.39 ? 0.01 ns
21.8 ? 0.2*
radiataaDC.
Tree
114 ? 2
0.89 ? 0.02
6.3 ? 0.3
115 ? 10
0.43 ? 0.02
22.9 ? 0.2
Proteaceae
Hakea
browniiaMeisn.
Shrub
524 ? 12**
0.31 ? 0.03*
11.6 ? 1.3*
80 ? 7*
0.74 ? 0.03**
20.2 ? 0.2**
salicifoliaa(Vent.) B.L. Burtt
Shrub
129 ? 4
0.70 ? 0.08
7.5 ? 1.0
113 ? 11
0.59 ? 0.02
21.3 ? 0.1
Fabaceae
Hardenbergia
comptonianac(Andrews) Benth.
Vine
75 ? 5**
2.41 ? 0.16**
16.0 ? 1.4*
123 ? 6 ns
0.61 ? 0.02*
22.0 ? 0.2*
violaceac(Schneer.) Stearn
Vine
116 ? 5
1.38 ? 0.07
10.4 ? 0.8
104 ? 13
0.54 ? 0.01
20.2 ? 0.4
Byttneriaceae
Lasiopetalum
discolorbHook.
Shrub
179 ? 15**
0.89 ? 0.01 ns
11.6 ? 0.7*
129 ? 8 ns
0.53 ? 0.03 ns
20.0 ? 0.3*
schulzeniia(F. Muell.) Benth.
Shrub
87 ? 6
1.04 ? 0.06
6.7 ? 1.1
151 ? 13
0.42 ? 0.05
21.1 ? 0.3
Rhamnaceae
Pomaderris
apetaladLabill.
Shrub
81 ? 5**
0.77 ? 0.06 ns
7.3 ? 0.8 ns
144 ? 4 ns
0.66 ? 0.03 ns
22.5 ? 0.5*
eriocephalaaN.A.Wakef.
Shrub
196 ? 10
0.62 ? 0.05
7.2 ? 2.4
84 ? 25
0.53 ? 0.05
19.3 ? 0.5
Byttneriaceae
Rulingia
magnifoliabF. Muell.
Shrub
85 ? 4 ns
1.16 ? 0.03*
12.9 ? 0.3 ns
162 ? 7 ns
0.67 ? 0.01**
21.5 ? 0.3 ns
salvifoliabBenth.
Shrub
78 ? 4
1.34 ? 0.06
12.8 ? 0.7
164 ? 7
0.61 ? 0.01
20.6 ? 0.5
Values are mean ? standard error for at least four oven-dried leaves.
1PPFD applied during the measurement of Aa: a, 1200; b, 1400; c, 1500; d, 1600; e, 1700; f, 1800 mmol PAR quanta m-2s-1.
Nm, leaf nitrogen concentration; Aa, CO2assimilation rate per unit area; PNUE,Aanormalised to Ciof 300 mmol mol-1divided by leaf nitrogen content per unit area;Ci/Ca, ratio of intercellular
to atmospheric CO2; D, carbon isotope discrimination.
Asterisks indicate statistical significance for t-tests (two tailed and assuming unequal variance) for a given attribute within each genus,where **P < 0.001,*P < 0.05 and ns = not significant.(For
additional parameters, see Supporting Information Table S1).
Cell wall nitrogen, sclerophylly and PNUE
261
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 4

Eqn 1, assuming the in vivo maximum carboxylation activ-
ity of Rubisco per unit area (Vcmax) was limiting photosyn-
thesis (Aa).
A
VC
(
(
CK O K
R
a
cmax
+
i
ico
d
*
=
−
)
+
)−
Γ
1
(1)
The kinetic constants of Rubisco (Kc and Ko, the
Michaelis–Menten constants for CO2 and O2, respectively,
and G*,the CO2compensation point in the absence of dark
respiration, Rd) were adopted from von Caemmerer et al.
(1994) assuming infinite internal conductance, with their
temperature dependence functions given in von Caem-
merer (2000).Rdwas assumed to be 0.1 of Aafor theANBG
data.These assumptions were validated from CO2response
curves measured on the eucalypt leaves (data not shown).
Morphological measurements
When gas exchange measurements were completed each
day, leaves were detached and the segment used for mea-
surement of photosynthesis was cut out and weighed. The
segment area was determined using a leaf area meter (Li-
Cor L3100, Li-Cor Inc.). Lamina thickness (T) was mea-
sured between the midrib and leaf edge with a Mitutoyo
(Japan) analogue thickness gauge (precision ? 20 mm). T
was calculated as the mean of four measurements. All
ANBG leaf segments were dried for a minimum of 48 h at
80 °C then reweighed, allowing calculation of leaf dry mass
per unit area (LMA). Water content (WC, mol m-2) was
computed as the change in leaf mass caused by drying,
divided by leaf area. A duplicate set of leaves, matching
those used for photosynthesis measurements, was sampled
for cell wall nitrogen measurement, being snap-frozen in
liquid nitrogen, then stored at -80 °C until used. Leaves
were then freeze-dried at -45 °C, 64 mT for at least 3 d,
using a Microprocessor Controlled Bench-Top Lyophilizer
(FTS Systems, Inc., Stone Ridge, NY, USA). All sampled
greenhouse eucalypt leaf segments were freeze-dried.
Leaf nitrogen measurements
Total leaf nitrogen
All dried leaves were ground separately in a ball mill.
The nitrogen concentration of the photosynthetic segment
was assayed using an elemental analyser (EA 1110 CHN-
O; Carlo-Erba Instruments, Milan, Italy) with a typical
machine precision of ?0.02% N. Approximately 1.2 mg of
each segment was analysed.
Cell wall mass and nitrogen
A protocol was adapted from Lamport (1965) and Onoda
et al. (2004) to remove soluble protein from the milled leaf
material. Approximately 10 mg of freeze-dried leaf was
extracted in 1.5 mL of buffer (50 mm tricine, pH 8.1)
containing 1% PVP40 (average molecular weight 40 000,
product no. 1407; Sigma Chemical Company, St Louis, MO,
USA). The sample was vortexed, centrifuged at 12 000 g
for 5 min (Eppendorf AG 5424, Hamburg, Germany) and
the supernatant was carefully removed. The pellet was
resuspended in buffer without PVP containing 1% sodium
dodecyl sulphate (SDS), incubated at 90 °C for 5 min, then
centrifuged at 12 000 g for 5 min. This was repeated, and
then two washes with 0.2 m KOH, two washes with deion-
ized water and finally two washes with ethanol were carried
out. The tube containing the pellet was then oven-dried at
80 °C. The remaining dry mass of pellet was assumed to
represent the leaf structural biomass,and the N content was
determined on 2–5 mg of material using the elemental
analyser as above.
The fraction of leaf nitrogen in cell wall material,NCW/NL,
was calculated using Eqn 2:
N
N
N
M
N
CW
L
CW
L
CW
CW
L
L
=××
M
M
M
(2)
where the fraction of cell wall material (MCW) recovered
from the total leaf biomass (ML) was multiplied by the
nitrogen concentration of cell wall material (NCW/MCW)
divided by the leaf nitrogen concentration (NL/ML).
Attempts were made to extract Rubisco for several of
the species using the method which has routinely been used
for Nicotiana tabacum (Mate et al. 1993).We tried to grind
fresh leaves, or leaves frozen in liquid N2, using mortar and
pestle, a Ten Broeck homogenizer or a Polytron, and we
also tried to extract freeze-dried leaf material that was
ball milled. None of our attempts yielded adequate soluble
protein or Rubisco, presumably because we were unable to
successfully rupture the mesophyll cells.
Calculation of Rubisco nitrogen and PNUE
The fraction of nitrogen allocated to Rubisco (NR/NL) was
calculated from Vcmax (mmol CO2 m-2s-1, derived from
Eqn 1) as follows:
N
N
N
n
N
R
L
cmax
V
R
cat
R
R
a
=××[]÷
M
k
(3)
where MRis the molecular mass of Rubisco, 0.55 g Rubisco
(mmol Rubisco)-1, kcat is the catalytic turnover number at
25 °C, 3.5 mol CO2 (mol Rubisco sites)-1s-1(von Caem-
merer et al. 1994), nR is the number of catalytic sites per
mole of Rubisco, 8 mol Rubisco sites (mol Rubisco)-1and
[NR] is the nitrogen concentration of Rubisco, 11.4 mmol N
(g Rubisco)-1and Na is the nitrogen content per unit leaf
area (mmol N m-2).This provides a minimum estimate as it
assumes full Rubisco activation.
PNUE [mmol CO2(mol N)-1s-1] was calculated by divid-
ing the CO2 assimilation rate per unit area scaled to a
common Ci (Aa300), by the nitrogen content per unit leaf
area (Na).
We also inferred the fraction of leaf nitrogen allo-
cated to Rubisco from the global relationship between
262
M. T. Harrison et al.
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 5

photosynthetic rate per unit leaf nitrogen and LMA {PNUE
[mmol CO2 (mol N)-1s-1= 587 ¥ LMA (g m-2)-0.435]} (Hiko-
saka 2004;Wright et al. 2004) using Eqn 4, as follows:
N
N
LMA
N
n
LMA
R
L
c max
V
A
a
R
cat
R
R
=××××[]×=
−−
5871 5.
0 435.
250
0 435.
M
k
(4)
where from Eqn 1, a Vcmax of 100 mmol m-2s-1is required
for an assimilation rate per unit leaf area, Aa250 of
19.7 mmol m-2s-1when leaf temperature is 25 °C and inter-
cellular CO2concentration is 250 mmol mol-1.The factor 1.5
was necessary to scale the fit to published NR/NLdata and
may indicate that either the average C3species kcatis only
2.33 mol CO2 (mol Rubisco sites)-1s-1, or Rubisco is not
fully active.
Carbon isotope composition measurements
The carbon isotope composition (d13C) of leaf material
sampled after gas exchange measurement was determined
using an elemental analyser (EA 1110 CHN-O;Carlo-Erba
Instruments) coupled to an isotope ratio mass spectrometer
(VG Isochrom; Fisons Instruments, Manchester, UK). d13C
values were obtained using approximately 0.1–0.2 mg of
milled leaf. Typical machine precision was ?0.2‰ d13C.
Composition values were converted to discrimination (D)
values using Eqn 5 (Farquhar & Richards 1984):
Δ =
−
+
⎛
⎝⎜
⎞
⎠⎟×
δ
1
δ
δ
ap
p
1000‰
(5)
where da is the carbon isotope composition of the air
(-8‰ = -8 ¥ 10-3), and dp is the measured carbon isotope
composition of the plant material.
Statistical analysis
All statistics were determined using SPSS v12.0 (SPSSAus-
tralasia,Sydney,Australia).Student’s t-tests were applied to
determine whether a given attribute of the two species of
each genus was significantly different.Regressions,linear or
non-linear least square and associated P and adjusted R2
values were calculated using the curve estimation function.
All nominal variables (Ci/Ca, D, the fraction of leaf biomass
in cell wall and the fractions of nitrogen allocated to cell
walls or Rubisco) were arcsine square root transformed
to meet the normality assumptions of parametric tests;
however, graphs of these variables have been left untrans-
formed for ease of interpretation. All figures were drawn
with species means, including one standard error of the
mean. Significant results are those with P < 0.05 unless
otherwise stated.
RESULTS
Morphology, chemistry and physiology
LMA differed significantly between each species pair from
the ANBG, with exception of the genus Rulingia (Table 1).
There was a sevenfold variation in LMA across ANBG
species (75–524 g m-2, Table 1) and a fivefold variation
across greenhouse eucalypts (50–240 g m-2, Table 2). The
greater spread of LMA with theANBG species was prima-
rily caused by the very high LMA values for Banksia blech-
nifolia and Hakea brownii. Similar variation was observed
for total leaf nitrogen concentration (Nm), with approxi-
mately eightfold (0.3–2.4 mmol g-1) and fourfold (0.6–
2.5 mmol g-1) ranges measured
greenhouse eucalypts, respectively.A threefold range in Aa
was found in both surveys (5–18 mmol m-2s-1). Some
species pairs had similar photosynthetic rates (e.g. Acacia),
while others (e.g.Banksia) differed significantly.Leaf thick-
ness was positively related to LMA [T (mm) = 1.6 ¥ LMA
(g m-2) + 380, R2= 0.44, n = 173, P < 0.001 (slope) for the
ANBG, T (mm) = 2.3 ¥ LMA (g m-2) + 92, R2= 0.89, n = 71,
P < 0.001 (slope) for the greenhouse Eucalyptus leaves
sampled from the ANBG followed the greenhouse regres-
sion and were generally thinner than other species for a
given LMA]. WC of leaves was positively related to LMA
[WC (mol m-2) = 0.033 ¥ LMA + 4.47, R2= 0.83, n = 177,
P < 0.001 (slope) for the ANBG (with the exception of
Acacia beckleri which had twice the WC (20.5 mol m-2) for
its LMA of 200 g m-2compared to all the other species],WC
(mol m-2) = 0.068 ¥ LMA + 3.5, R2= 0.88, n = 72, P < 0.001
(slope) for the greenhouse data.TheWC of leaves from the
greenhouse Eucalyptus was about 50% more than for
leaves from other genera from the ANBG for any given
LMA with the exception of A. beckleri. The ratio of CO2
concentration in the intercellular space (Ci) to that in the
IRGA cuvette (Ca) was smaller and more variable for the
ANBG species (0.37–0.74; Table 1) than the greenhouse
Eucalyptus (0.78–0.88;Table 2).The relationship between D
and Ci/Ca deviated from the expected relationship for the
ANBG data (Table 1). Small Ci/Ca ratios were not associ-
ated with small D values, possibly reflecting the dry condi-
tions during the gas exchange survey. For the greenhouse
Eucalyptus, D values were consistent with the measured
Ci/Ca, and although the correlation was not strong
(R2= 0.17), it was significant (slope P = 0.005, n = 41;
Table 2).The nitrogen treatment applied to the Eucalyptus
plants significantly increased nitrogen concentration per
unit leaf dry mass and increased photosynthetic rate, both
per unit leaf area and per unit leaf nitrogen.
inthe ANBG and
Relationships between cell wall biomass, cell
wall nitrogen and LMA
The fraction of leaf biomass recovered as cell walls was
independent of LMA within Eucalyptus andANBG genera
(Fig. 1). In general,ANBG species had greater proportions
of biomass in cell walls. The highly sclerophyllous genus
Banksia had around 0.65 of leaf biomass in cell walls com-
pared to 0.30 for Eucalyptus. For Eucalyptus leaves, the
proportion of leaf biomass in cell walls for a given LMA
was not affected by leaf age, N treatment or whether the
plants were grown in the ANBG or greenhouse.
Cell wall nitrogen, sclerophylly and PNUE
263
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 6

Table 2. Leaf attributes of Eucalyptus species grown in a greenhouse (see Methods)
Species
Old/Young
N treatment
LMA (g m-2)
Nm(mmol g-1)
Aa(mmol m-2s-1)
PNUE
[mmol (mol N)-1s-1]
Ci/Ca
D (‰)
Eucalyptus bridgesianaa
O
-N
80.8 ? 3.1
1.13 ? 0.10
11.1 ? 1.8
119.1 ? 14.2
0.82 ? 0.02
24.6 ? 0.4
E. bridgesianaa
Y
-N
54.5 ? 2.8
1.32 ? 0.10
9.0 ? 0.7
125.2 ? 7.0
0.86 ? 0.02
25.2 ? 0.3
E. bridgesianaa
Y
+N
49.5 ? 2.4
1.85 ? 0.05
13.7 ? 1.3
153.3 ? 12.9
0.84 ? 0.02
24.6 ? 0.2
Eucalyptus elatab
O
-N
63.3 ? 5.5
0.72 ? 0.04
5.1 ? 0.4
111.4 ? 4.9
0.88 ? 0.01
22.6 ? 0.8
E. elatab
Y
+N
57.0 ? 2.3
1.02 ? 0.02
10.2 ? 0.8
179.2 ? 10.9
0.86 ? 0.02
26.2 ? 0.2
Eucalyptus manniferac
O
-N
94.8 ? 10.4
0.90 ? 0.09
15.9 ? 2.9
172.5 ? 15.5
0.86 ? 0.02
25.6 ? 0.8
E. manniferac
Y
-N
62.2 ? 3.1
1.21 ? 0.13
9.6 ? 1.3
125.2 ? 11.4
0.91 ? 0.02
26.0 ? 0.1
E. manniferac
Y
+N
57.5 ? 11.7
1.41 ? 0.17
12.6 ? 1.8
162.2 ? 8.8
0.81 ? 0.05
25.6 ? 0.5
Eucalyptus mooreid
O
+N
241.0 ? 10.1
0.59 ? 0.05
15.7 ? 0.3
110.0 ? 9.2
0.78 ? 0.02
21.6 ? 0.4
Eucalyptus pauciflorae
O
+N
172.9 ? 6.9
0.64 ? 0.06
16.3 ? 1.3
140.1 ? 9.7
0.78 ? 0.03
23.7 ? 0.4
Eucalyptus polyanthemosf
O
+N
121.5 ? 6.7
1.14 ? 0.07
16.4 ? 0.6
117.0 ? 5.8
0.86 ? 0.02
23.5 ? 0.3
E. polyanthemosf
Y
+N
53.0 ? 3.2
2.51 ? 0.32
17.7 ? 1.8
138.9 ? 7.0
0.85 ? 0.02
24.4 ? 0.1
E. rossiig
O
+N
181.4 ? 8.3
0.69 ? 0.08
18.3 ? 1.3
145.4 ? 7.4
0.82 ? 0.02
23.7 ? 0.9
aR.T. Baker,bDenh.,cA. Cunn. ex Benth.,dMaiden & Cambage,eSieber ex Spreng.,fSchau.,gR.T. Baker & H.G. Sm.
Abbreviations: Nm, leaf nitrogen concentration; Aa, CO2assimilation rate per unit area; PNUE, Aanormalised to Ciof 300 mmol mol-1divided by leaf nitrogen content per unit area; Ci/Ca,ratio
of intercellular to atmospheric CO2; and D, carbon isotope discrimination.
Values for LMA,Nm,Aaand Ci/Caare mean ? standard error of at least four leaves;values for D are mean ? standard error of three leaves.For all species,Aavalues were measured at a PPFD
of 1800 mmol quanta m-2s-1.
All leaves were freeze-dried after photosynthetic measurements (see Supporting Information Table S1).
264
M. T. Harrison et al.
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 7

The nitrogen concentration of cell wall material was
roughly 0.4 times leaf nitrogen concentration for all the
species sampled from the ANBG, and 0.6 times leaf nitro-
gen concentration for Eucalyptus (Fig. 2). For Eucalyptus,
the two species sampled from the ANBG fell within
the greenhouse data. Leaves from the leguminous species
(Acacia and Hardenbergia) had noticeably higher leaf
nitrogen concentrations than did the non-leguminous
species.
The fraction of leaf nitrogen recovered in cell walls
was independent of LMA, both within and across genera
(Fig. 3). The fraction of nitrogen allocated to cell walls
for the ANBG Eucalyptus species fell within the range
observed for the greenhouse-grown Eucalyptus species.
Although there was an increase in the fraction of nitrogen
associated with cell walls with increasing LMA between the
two ANBG Eucalyptus species, the difference was not
significant. Indeed, there was no overall tendency for an
increasing proportion of leaf nitrogen to be allocated to cell
walls with increasing LMA in the greenhouse Eucalyptus.
The only other significant increase in the fraction of cell
wall nitrogen with increasing LMA in the ANBG species
was for the genus Lasiopetalum. There was a threefold
spread (0.1–0.3) in the fraction of nitrogen recovered in cell
walls for leaves with an LMA of 150 g m-2.
Relationship between PNUE and LMA
PNUE calculated with a Ci value of 300 mmol mol-1
was weakly but negatively associated with LMA for the
ANBG species [PNUE = 131 (?7) - 0.12 (?0.03) ¥ LMA,
R2= 0.12, n = 93, slope P < 0.001, Fig. 4].A similar relation-
ship was apparent for Eucalyptus, although this relation-
ship was not significant [PNUE = 145 (?7) - 0.08 (?0.06) ¥
LMA, R2= 0.70, n = 71, slope P = 0.2]. Indeed, the variation
apparent in PNUE for a given greenhouse Eucalyptus
LMA was nearly as great as that across the 10-fold range in
LMA of the ANBG species. The Acacia species had the
lowest PNUE values, whereas Banksia species had among
the highest despite having the greatest fraction of leaf
biomass recovered in cell walls. No general relationship
existed between PNUE and the fraction of nitrogen in cell
walls (Fig. 5). Eucalyptus was the only ANBG genus that
had significantly reduced PNUE associated with greater
fraction of nitrogen in cell walls, and while a negative rela-
tionship was also observed for the greenhouse Eucalyptus
0 100200300 400500 600
0.0
0.2
0.4
0.6
0.8
1.0
Acacia
Banksia
Eucalyptus
Hakea
Hardenbergia
Lasiopetalum
Pomaderris
Rulingia
E. bridgesiana
E. elata
E. mannifera
E. moorei
E. pauciflora
E. polyanthemos
E. rossii
Cell wall biomass fraction
LMA (g m–2)
Figure 1. The fraction of leaf biomass recovered in cell walls as
a function of leaf mass per unit area (LMA).Australian National
Botanic Gardens (ANBG) species pairs have solid symbols
joined by solid lines. Eucalyptus species grown in the greenhouse
have hollow symbols. Error bars denote one standard error of
the mean.
0.0 0.51.01.5 2.02.53.0
0.0
0.5
1.0
1.5
2.0
Cell wall N concentration (mmol g
–1)
Leaf N concentration (mmol g–1)
Figure 2. Cell wall nitrogen concentration versus leaf nitrogen
concentration.Australian National Botanic Gardens (ANBG)
species (solid symbols as shown in Fig. 1), regression [NCW] = 0.41
[NL], R2= 0.76, n = 16, P < 0.001. Eucalyptus species grown in the
greenhouse (hollow symbols as shown in Fig. 1), regression
[NCW] = 0.59 [NL], R2= 0.90, n = 13, P < 0.001.
0 100200300 400500600
0.0
0.1
0.2
0.3
0.4
0.5
Cell wall N fraction
LMA (g m–2)
Figure 3. The fraction of leaf nitrogen present in cell walls as a
function of leaf mass per unit area (LMA).Australian National
Botanic Gardens (ANBG) species pairs have solid symbols
joined by solid lines; Eucalyptus species grown in the greenhouse
have hollow symbols, as defined in Fig. 1.
Cell wall nitrogen, sclerophylly and PNUE
265
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 8

[PNUE = 155 (?9) - 70 (?29) ¥ [NCW/NL], n = 61, slope
P = 0.02], it was very weak (R2= 0.08). Therefore, the
decline in PNUE as LMA increases is not associated with
an increased fraction of nitrogen in the cell walls.
Relationships between the fractions of leaf
nitrogen in Rubisco or cell walls
As we were unable to extract Rubisco from the leaves, we
calculated the fraction of nitrogen in Rubisco, assuming a
constant set of kinetic parameters for all species (Eqn 3).
The fraction of nitrogen in Rubisco decreased as LMA
increased for each of the eight species pairs sampled
fromthe ANBG[[NR/NL] = 0.13
(?3.1 ¥ 10-5) ¥ LMA, R2= 0.08, n = 92, slope P = 0.003,
Fig. 6].There was no equivalent relationship evident for the
Eucalyptus data.However,published data from many other
species where Rubisco protein content has been directly
measured make the overall inverse relationship as a func-
tion of LMA more obvious (see the hollow circles in Fig. 6).
The solid curve is calculated from the relationship between
PNUE and LMA in the Glopnet database (Wright et al.
2004); see Eqn 4. There is broad overlap between values
based on Rubisco protein assays and those calculated from
gas exchange measurements. Overall, the fraction of nitro-
gen in Rubisco increases rapidly as LMA is reduced below
100 g m-2and decreases more gradually as LMA increases
above 100 g m-2.
By calculating the nitrogen cost of Rubisco, it is possible
to directly compare any trade-off between nitrogen allo-
cated to Rubisco versus cell walls (Fig. 7). Nitrogen alloca-
tion to Rubisco significantly decreased as more nitrogen
was allocated to cell walls for only three of the species
pairs (Banksia, Lasiopetalum, Eucalyptus). Out of these,
cell wall nitrogen allocation only differed significantly
between the two Lasiopetalum species.Moreover,there was
(?0.01) - 9.4 ¥ 10-5
0 100200300 400 500600
0
50
100
150
200
PNUE (mmol mol
–1 s
–1)
LMA (g m–2)
Figure 4. Relationship between photosynthetic nitrogen-use
efficiency (PNUE) and leaf mass per unit area (LMA).
Photosynthesis was measured at 1200–1800 mmol PAR photons
m-2s-1for Australian National Botanic Gardens (ANBG) genera
and 1800 mmol PAR photons m-2s-1for the greenhouse species,
then adjusted to a common Ciof 300 mmol mol-1.ANBG species
pairs have solid symbols joined by solid lines; Eucalyptus species
grown in the greenhouse have hollow symbols, as defined in
Fig. 1.The regression for Eucalyptus (dashed line) was
PNUE = 157.7 (?12) - 0.234 (?0.095) LMA, r2= 0.31, n = 15.
0.00.10.2 0.30.4
0
50
100
150
200
PNUE (mmol mol
–1 s
–1)
Cell wall N fraction
Figure 5. Relationship between photosynthetic nitrogen-use
efficiency (PNUE) and the fraction of leaf nitrogen present in cell
walls.Australian National Botanic Gardens (ANBG) species pairs
have solid symbols joined by solid lines;Eucalyptus species grown
in the greenhouse have hollow symbols,as defined in Fig. 1.
0 100 200300 400500 600
0.0
0.1
0.2
0.3
0.4
Rubisco N fraction
LMA (g m
–2)
Figure 6. The fraction of leaf nitrogen in Rubisco in relation to
leaf mass per unit area (LMA).Australian National Botanic
Gardens (ANBG) species pairs have solid symbols joined by
solid lines (as defined in Fig. 1); Eucalyptus species grown in the
greenhouse (?). Published data where Rubisco has been assayed
biochemically (?): Ethier et al. (2006) (Pseudotsuga menziesii),
Evans et al. (1994) (Nicotiana tabacum), Evans & Seemann
(1984) (Triticum aestivum), Gleadow et al. (1998) (Eucalyptus
cladocalyx), Hikosaka et al. (1998) (Quercus mysinaefolia,
Chenopodium album), Katahata et al. (2007a,b) (Daphniphyllum
humile), Kuppers (1996) (Acacia and Eucalyptus spp.), Onoda
et al. (2004) (Polygonum cuspidatum), Poorter & Evans (1998) 10
species,Takashima et al. (2004) (Quercus spp.),Warren,Adams &
Chen (2000) 10 species.The solid curve is calculated from the
power relationship between PNUE and LMA (Hikosaka 2004;
Wright et al. 2004), NR/NL= LMA-0.435, see methods.
266
M. T. Harrison et al.
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 9

no significant relationship between nitrogen allocated to
Rubisco and that to cell walls for the eight Eucalyptus
species measured in the greenhouse study. The fraction of
nitrogen allocated to Rubisco in E. pauciflora was much
lower in leaves sampled from the ANBG (0.08) than from
the greenhouse (0.15),but similar fractions of nitrogen were
recovered in their cell walls (0.2). The only significant
increase in both the fraction of nitrogen in Rubisco and in
cell walls was for the Acacia species pair.Overall,these data
did not suggest that an increased allocation of nitrogen to
cell walls occurred at the expense of nitrogen allocation to
Rubisco.
DISCUSSION
Consistent with expectations based on worldwide observa-
tions (Reich et al. 1997; Hikosaka 2004;Wright et al. 2004),
the rate of photosynthesis per unit leaf nitrogen, PNUE,
was negatively related to LMA for the ANBG leaves
sampled in this study (Fig. 4).The negative correlation was
also apparent in the Hakea and Eucalyptus congeneric com-
parisons (Table 1). Our sampling strategies tried to maxi-
mize the variation in LMA between species that typify the
sclerophyllous evergreen leaves of vegetation in temperate
Australia. If LMA is causally related to characters that
affect PNUE,then the relationship should be evident in the
10-fold range in LMA between our samples. We focused
on measuring nitrogen allocated to leaf structure to explain
variation in PNUE because few previous studies have done
so.A possible reason for this has been the lack of simple and
robust methods for separating cell walls and their bound
protein from the remainder of the cell. The trade-off
between nitrogen allocated to Rubisco and cell walls seen
in Quercus (Takashima et al. 2004) and Polygonum (Onoda
et al. 2004) was not confirmed with the Australian species
examined here.
Relationship between the fraction of leaf
nitrogen in cell walls and LMA
The fraction of leaf nitrogen in cell walls can be factored
into three components: (1) the fraction of leaf biomass in
cell walls; (2) the nitrogen concentration of cell walls; and
(3) the nitrogen concentration of the whole leaf.While the
nitrogen concentration of leaf material has been routinely
measured, there are few data available for components 1
and 2.The fraction of leaf biomass in walls ranged from 0.20
to 0.65, and varied between genera.A considerably greater
fraction of leaf biomass was recovered in the cell wall mate-
rial from Banksia and Hakea leaves compared to Acacia
and Eucalyptus leaves with similar LMA (Fig. 1). The
fraction of leaf biomass in cell walls was independent of
changes in LMA for the nine Eucalyptus species and all
other species pairs except Hardenbergia.
Plant cell walls contain between 2 and 10% protein
(Brett & Waldron 1996), which equates to 0.2–1.1 mmol N
(g cell wall)-1. Our values were consistent with this, ranging
from 0.1 to 0.9 mmol N (g cell wall)-1. Cell wall nitrogen
concentration was about 0.4 times leaf nitrogen concentra-
tion for all genera apart from Eucalyptus where the factor
was 0.6 (Fig. 2).Consequently,variation in cell wall nitrogen
concentration was cancelled out by equivalent changes in
leaf nitrogen concentration (see Eqn 2).
We carried out extensive comparisons between methods
forextractingleafmaterial.Ourinitialextractionsweredone
without PVP as we found that it adhered to some samples,
and because it contains 12% N, minor contamination easily
distorted the value derived for cell walls in these leaves with
low nitrogen concentrations.The inclusion of PVP required
multiple washing steps, including the use of 0.2 m KOH,
to completely remove the ‘soluble’ PVP from the cell wall
material. The difference between including PVP or not
was most evident with the Eucalyptus species, and the data
obtained without PVP are available on request.
In contrast to the strong positive relationship between
the fraction of leaf nitrogen in cell walls and LMA in four
Quercus species (Takashima et al. 2004) and Polygonum
(Onoda et al. 2004), these parameters were independent in
our study (Fig. 3). We conducted additional measurements
with leaves collected from mature trees of two oak species
to test the cell wall extraction methods. The evergreen
Quercus suber had greater LMA than the deciduous
Quercus acutissima, but a slightly smaller fraction of nitro-
gen in cell walls (0.106 ? 0.004 and 0.113 ? 0.012, with
LMA values of 140 and 116 g m-2, respectively). By con-
trast, the evergreen Quercus acuta and Quercus glauca had
much smaller LMA values of 40–60 g m-2, but similar cell
wall nitrogen fractions between 0.1 and 0.2 (Takashima
0.0 0.1 0.20.30.4
0.0
0.1
0.2
0.3
Rubisco N fraction
Cell wall N fraction
Figure 7. The fraction of leaf nitrogen in ribulose
1·5-bisphosphate carboxylase/oxygenase (Rubisco) in relation to
the fraction of leaf nitrogen in cell walls. Nitrogen in Rubisco
was calculated from measurements of photosynthesis (see
Methods).Australian National Botanic Gardens (ANBG) species
pairs have solid symbols joined by solid lines, and Eucalyptus
species grown in the greenhouse have hollow symbols, as defined
in Fig. 1.The shaded zone represents the upper bound assuming
that on average, Rubisco represents one quarter to one-third of
the nitrogen in soluble and thylakoid protein (Evans & Poorter
2001).
Cell wall nitrogen, sclerophylly and PNUE
267
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 10

et al. 2004).The fraction of nitrogen in cell walls is indepen-
dent of LMA when one considers all of the Quercus data
together.About 0.06 of leaf nitrogen was recovered in cell
walls from both sun and shade leaves of Lindera umbellata,
although they differed 2.5-fold in LMA (Yasumura, Hiko-
saka & Hirose 2006).Therefore, it is not valid to generalize
from the results of Onoda et al. (2004) and Takashima et al.
(2004). When more species are included, one sees that the
fraction of leaf nitrogen in cell walls is independent of
LMA.
Variation in PNUE reflects changes in nitrogen
allocated to Rubisco
The majority of variation in light-saturated PNUE between
species can be attributed to factors affecting the supply of
CO2 to the sites of carboxylation (e.g. stomatal conduc-
tance,internal conductance) or the Rubisco activity per unit
leaf nitrogen (because of nitrogen allocation, kinetic prop-
erties and activation state of Rubisco) (Pons, van der Werf
& Lambers 1994; Hikosaka et al. 1998; Poorter & Evans
1998; Ripullone et al. 2003; Warren & Adams 2004). To
facilitate the analysis of the underlying causes, we normal-
ized the rates of CO2assimilation to a common intercellular
CO2 mole fraction of 300 mmol mol-1(Aa300). Some of the
remaining variation in PNUE could still be associated with
variation in internal conductance,which influences the mole
fraction of CO2in the chloroplast, Cc. None of the fourfold
variation in PNUE observed by Lloyd et al. (1992) between
four tree species was related to variation in Ci–Cc. Two
reviews comparing sclerophyllous and mesophytic leaves
also showed no difference in Ci–Cc if the comparison was
restricted to the range where the photosynthetic rates over-
lapped (Evans 1999; Warren 2008). While there is a ten-
dency for leaves with greater LMA to have lower internal
conductances (Flexas et al. 2008), as they also tend to have
lower photosynthetic capacities,the draw down Ci–Ccis not
significantly related to LMA.A comparison between seven
Banksia species ranging in LMA from 130 to 480 g m-2also
found that the draw down Ci–Ccwas independent of LMA
(Hassiotou, personal communication).Although it was not
feasible to measure internal conductance in the present
study, the values we observed for the carbon isotopic com-
position,D,and the extensive literature reviews suggest that
it is unlikely that internal conductance accounts for the
observed variation in PNUE.
Instead, nitrogen allocation to Rubisco, its kinetic prop-
erties and activation state are the probable causes for varia-
tion in PNUE. While there are several methods used to
measure Rubisco content of leaves,it is not always possible
to reliably extract Rubisco from some species. Conse-
quently, an alternative approach is to derive Rubisco activ-
ity from gas exchange data using the Farquhar et al. (1980)
model of photosynthesis. This can be converted to an
amount of protein or nitrogen by assuming a catalytic turn-
over number, kcat. Assuming constant values for Rubisco,
kinetic parameters provide a convenient way of approxi-
mating Rubisco content, but conceal the variation which
undoubtedly exists between species (kcat values range
between 2 and 6 mol CO2 (mol Rubisco sites)-1s-1
(Seemann, Tepperman & Berry 1981; Seemann & Berry
1982; Makino, Mae & Ohira 1988; Evans 1989; Poorter &
Evans 1998; Sage 2002; Ghannoum et al. 2005). There are
several examples demonstrating good agreement between
direct quantification of Rubisco and that calculated from
gas exchange, such as for Phaseolus vulgaris (von Caem-
merer & Farquhar 1981), Triticum aestivum (Evans &
Seemann 1984), Nicotiana tabacum (von Caemmerer et al.
1994) or Pseudotsuga predicted using Rubisco kinetic
parameters from Nicotiana (Ethier et al. 2006). However,
Rubisco derived from gas exchange should be regarded as a
minimum nitrogen cost because it is usually calculated
assuming that all the Rubisco is fully activated, which was
not the case for older Pseudotsuga leaves (Ethier et al.
2006). Assembling published data for different species
reveals that the fraction of leaf nitrogen allocated to
Rubisco declines curvilinearly as LMA increases (Fig. 6).
Trade-off between the fraction of leaf nitrogen
in cell walls and Rubisco
The fraction of leaf nitrogen in cell walls was independent
of LMA when examined across species and showed no
consistent pattern within genera (Fig. 3). Therefore, it can-
not be contributing to the negative relationship between
PNUE and LMA.However,it is appealing to think that the
allocation of nitrogen between different pools within a leaf
involves competitive trade-offs and that variation in the
allocation of nitrogen to cell walls could impact on nitrogen
allocated to photosynthesis. The majority of nitrogen in a
leaf is directly associated with photosynthesis (Evans &
Seemann 1989; Makino & Osmond 1991; Pons et al. 1994;
Hikosaka & Terashima 1995; Evans 1996; Niinemets &
Tenhunen 1997; Poorter & Evans 1998). There are two
major pools,the proteins of the chloroplast thylakoid mem-
branes involved in light capture, electron transport and
photophosphorylation, and the soluble proteins of the
Calvin and photorespiratory cycles. To a first approxima-
tion,the sum of thylakoid and soluble protein gives a guide
to the nitrogen cost of photosynthesis for a leaf. This cost
was quite similar across 10 species representing herbaceous
and woody plants, being four times the nitrogen content of
Rubisco (Evans & Poorter 2001). However, for pea leaves,
the sum of chloroplast plus mitochondrial nitrogen was
three times the nitrogen content of Rubisco (Makino &
Osmond 1991) and for Eucalyptus cladocalyx, total protein
nitrogen was three times the nitrogen content of Rubisco
(Gleadow,Foley &Woodrow 1998).This cost (three to four
times Rubisco N) allows one to scale the nitrogen trade-off
between photosynthesis versus leaf structure. The shaded
zone in Fig. 7 represents the upper bound of this trade-off.
For points falling within the shaded zone,all of leaf nitrogen
would be accounted for by photosynthesis and cell wall
material,leaving none for other cellular functions.Our data
fall below this zone, suggesting that little soluble protein
268
M. T. Harrison et al.
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 11

inadvertently stuck to the cell wall pellet, leaving 5–10% of
leaf nitrogen unaccounted for.
Increased allocation of nitrogen to structure was accom-
panied by a reduced investment in Rubisco for both
Polygonum (Onoda et al. 2004) and Quercus (Takashima
et al. 2004) leaves. However, for Polygonum, the slope of
the relationship (-1.5) greatly exceeded a direct trade-off
(-0.25 to -0.33). Compared to deciduous Quercus, ever-
green Quercus leaves increased nitrogen allocated to
cell walls and decreased nitrogen allocated to Rubisco
(Takashima et al. 2004), but the slope (-0.56) again
exceeded a direct trade-off.A third study sampled leaves of
Lindera umbellata throughout a growing season (Yasumura
et al. 2006).This revealed that allocation of nitrogen to cell
walls increased early in the season without any concomitant
change to Rubisco and then during autumn, nitrogen
released from Rubisco degradation was also not associated
with any change in cell wall nitrogen. Clearly, there was no
internal trade-off in nitrogen allocation between cell wall
and Rubisco through the lifespan of these leaves. The
spread of the Eucalyptus data revealed no correlation
between Rubisco and cell wall nitrogen (Fig. 7), and the
Acacia species pair from the ANBG actually had greater
nitrogen allocation to Rubisco as the fraction of nitrogen
allocated to cell walls increased.Considering all of the data
together reveals that the 23 species populate a large part of
the space below the shaded zone. The striking trade-off
between nitrogen allocated to Rubisco versus cell walls
observed for deciduous and evergreen Quercus is therefore
unlikely to hold as a general rule. Although we were only
able to assay cell wall nitrogen and had to calculate Rubisco
nitrogen from gas exchange, our results refute the general-
ity of the claim that increasing amounts of cell wall nitrogen
in leaves with greater LMA necessarily result in a reduction
in nitrogen allocation to Rubisco.
ACKNOWLEDGMENTS
We would like to thank the ANBG for allowing us to
sample from the gardens, and Stephanie McCaffery for her
painstaking analyses.
REFERENCES
Brett C.T. & Waldron K.W. (1996) Physiology and Biochemistry of
Plant Cell Walls, 2nd edn. Chapman & Hall, London, UK.
von Caemmerer S. (2000) Biochemical Models of Leaf Photo-
synthesis, Vol. 2. CSIRO Publishing, Collingwood, Victoria,
Australia.
von Caemmerer S. & Farquhar G.D. (1981) Some relationships
between the biochemistry of photosynthesis and the gas
exchange of leaves. Planta 153, 376–387.
von Caemmerer S., Evans J.R., Hudson G.S. & Andrews T.J.
(1994) The kinetics of ribulose-1,5-bisphosphate carboxylase/
oxygenase in vivo inferred from measurements of photosynthe-
sis in leaves of transgenic tobacco. Planta 195, 88–97.
Ellsworth D.S., Reich P.B., Naumburg E.S., Koch G.W., Kubiske
M.E. & Smith S.D. (2004) Photosynthesis,carboxylation and leaf
nitrogen responses of 16 species to elevated pCO2across four
free-air CO2 enrichment experiments in forest, grassland and
desert. Global Change Biology 10, 2121–2138.
Ethier G.J., Livingston N.J., Harrison D.L., Black T.A. & Moran
J.A. (2006) Low stomatal and internal conductance to CO2
versus Rubisco deactivation as determinants of the photosyn-
thetic decline of ageing evergreen leaves. Plant, Cell & Environ-
ment 29, 2168–2184. doi:10.1111/j.1365-3040.2006.01590.x.
Evans J.R. (1989) Photosynthesis and nitrogen relationships in
leaves of C3plants. Oecologia 78, 9–19.
Evans J.R. (1996) Developmental constraints on photosynthesis:
effects of light and nutrition. In Photosythesis and the Environ-
ment (ed N.R. Baker), pp. 281–304. Kluwer, Dordrecht, the
Netherlands.
Evans J.R. (1999) Leaf anatomy enables more equal access to light
and CO2between chloroplasts. New Phytologist 143, 93–104.
Evans J.R. & Poorter H. (2001) Photosynthetic acclimation of
plants to growth irradiance: the relative importance of specific
leaf area and nitrogen partitioning in maximizing carbon gain.
Plant, Cell & Environment 24, 755–767.
Evans J.R. & Seemann J.R. (1984) Differences between wheat
genotypes in specific activity of ribulose-1,5-bisphosphate car-
boxylase and the relationship to photosynthesis. Plant Physiol-
ogy 74, 759–765.
Evans J.R. & Seemann J.R. (1989) The allocation of protein nitro-
gen in the photosynthetic apparatus: costs, consequences, and
control. In Photosynthesis (ed W.R. Briggs), pp. 183–205. A.R.
Liss, New York, NY, USA.
Evans J.R.,von Caemmerer S.,Setchell B.A. & Hudson G.S. (1994)
The relationship between CO2 transfer conductance and leaf
anatomy in transgenic tobacco with a reduced content of
Rubisco. Australian Journal of Plant Physiology 21, 475–495.
Farquhar G.D. & Richards R.A. (1984) Isotopic composition of
plant carbon correlates with water-use efficiency of wheat geno-
types. Australian Journal of Plant Physiology 11, 539–552.
Farquhar G.D.,von Caemmerer S. & Berry J.A. (1980) A biochemi-
cal model of photosynthetic CO2 assimilation in leaves of C3
species. Planta 149, 78–90.
Felsenstein J. (1985) Phylogenies and the comparative method.The
American Naturalist 125, 1–15.
Field C. & Mooney H.A. (1986) The photosynthesis–nitrogen rela-
tionship in wild plants.In On the Economy of Form and Function
(ed. T.J. Givnish), pp. 25–55. Cambridge University Press,
Cambridge, UK.
Flexas J., Ribas-Carbo M., Diaz-Espejo A., Galmes J. & Medrano
H. (2008) Mesophyll conductance to CO2: current knowledge
and future prospects. Plant, Cell & Environment 31, 602–621.
Ghannoum O.,Evans J.R.,Chow W.S.,Andrews T.J.,Conroy J.P. &
von Caemmerer S. (2005) Faster Rubisco is the key to superior
nitrogen-use efficiency in NADP–malic enzyme relative to
NAD–malic enzyme C4 grasses. Plant Physiology 137, 638–
650.
Gleadow R.M., Foley W.J. & Woodrow I.E. (1998) Enhanced CO2
alters the relationship between photosynthesis and defence in
cyanogenic Eucalyptus cladocalyx F. Muell. Plant, Cell & Envi-
ronment 21, 12–22.
Hewitt E.J. (1966) Sand and water culture methods used in the
study of plant nutrition. Commonwealth Bureau of Horticulture
and Plantation Crops, East Malling, Technical Communication
Number 22.
Hikosaka K. (2004) Interspecific difference in the photosynthesis–
nitrogen relationship:patterns,physiological causes,and ecologi-
cal importance [review].Journal of Plant Research 117, 481–494.
Hikosaka K. & Terashima I. (1995) A model of the acclimation
of photosynthesis in the leaves of C3 plants to sun and shade
with respect to nitrogen use. Plant, Cell & Environment 18, 605–
618.
Cell wall nitrogen, sclerophylly and PNUE
269
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270

Page 12

Hikosaka K., Hanba Y.T., Hirose T. & Terashima I. (1998) Photo-
synthetic nitrogen-use efficiency in leaves of woody and herba-
ceous species. Functional Ecology 12, 896–905.
Katahata S.-I., Naramoto M., Kakubari Y. & Mukai Y. (2007a)
Photosynthetic capacity and nitrogen partitioning in foliage of
the evergreen shrub Daphniphyllum humile along a natural light
gradient. Tree Physiology 27, 199–208.
Katahata S.-I., Naramoto M., Kakubari Y. & Mukai Y. (2007b)
Seasonal changes in photosynthesis and nitrogen allocation
in leaves of different ages in evergreen understory shrub Daph-
niphyllum humile. Trees – Structure and Function 21, 619–629.
Kuppers B.I.L. (1996) Nitrogen and Rubisco contents in eucalypt
canopies as affected by Acacia neighbourhood.Plant Physiology
and Biochemistry 34, 753–760.
Lamport D. (1965) The protein component of primary cell walls.
Advances in Botanical Research 2, 151–218.
Lloyd J., Syvertsen J.P., Kriedemann P.E. & Farquhar G.D. (1992)
Low conductances for CO2diffusion from stomata to the sites of
carboxylation in leaves of woody species. Plant, Cell & Environ-
ment 15, 873–899.
Makino A. & Osmond B. (1991) Effects of nitrogen nutrition on
nitrogen partitioning between chloroplasts and mitochondria in
pea and wheat. Plant Physiology 96, 355–362.
Makino A., Mae T. & Ohira K. (1988) Differences between wheat
and rice in the enzymic properties of ribulose-1,5-bisphosphate
carboxylase/oxygenase and the relationship to photosynthetic
gas exchange. Planta 174, 30–38.
Mate C.J., Hudson G.S., Voncaemmerer S., Evans J.R. & Andrews
T.J. (1993) Reduction of ribulose bisphosphate carboxylase
activase levels in tobacco (Nicotiana-tabacum) by antisense
RNA reduces ribulose bisphosphate carboxylase carbamylation
and impairs photosynthesis. Plant Physiology 102, 1119–1128.
Niinemets U. & Tenhunen J.D. (1997) A model separating leaf
structural and physiological effects on carbon gain along light
gradients for the shade-tolerant species Acer saccharum. Plant,
Cell & Environment 20, 845–866.
Onoda Y., Hikosaka K. & Hirose T. (2004) Allocation of nitrogen
to cell walls decreases photosynthetic nitrogen-use efficiency.
Functional Ecology 18, 419–425.
Pons T., van der Werf A. & Lambers H. (1994) Photosynthetic
nitrogen use efficiency of inherently slow- and fast-growing
species: possible explanations for observed differences. In A
Whole Plant Perspective on Carbon–Nitrogen Interactions (eds J.
Roy & E. Garnier),pp. 61–77.Academic Publishing,The Hague,
the Netherlands.
Poorter H. & Evans J.R. (1998) Photosynthetic nitrogen-use
efficiency of species that differ inherently in specific leaf area.
Oecologia 116, 26–37.
Reich P.B., Walters M.B. & Ellsworth D.S. (1997) From tropics to
tundra – global convergence in plant functioning.Proceedings of
the NationalAcademy of Sciences of the United States ofAmerica
94, 13730–13734.
Ripullone F., Grassi G., Lauteri M. & Borghetti M. (2003)
Photosynthesis–nitrogen relationships: interpretation of differ-
ent patterns between Pseudotsuga menziesii and Populus x
euroamericana in a mini-stand experiment. Tree Physiology 23,
137–144.
Sage R.F. (2002) Variation in the k(cat) of Rubisco in C-3 and C-4
plants,and some implications for photosynthetic performance at
high and low temperature. Journal of Experimental Botany 53,
609–620.
Seemann J.R. & Berry J.A. (1982) Interspecific differences in the
kinetic properties of RuBP carboxylase protein. Carnegie Insti-
tution of Washington Yearbook 81, 78–83.
Seemann J.R., Tepperman J. & Berry J.A. (1981) The relationship
between photosynthetic performance and the levels and kinetic
properties of RuBP carboxylase–oxygenase from desert winter
annuals. Carnegie Institution of Washington Yearbook 80, 67–
72.
Takashima T., Hikosaka K. & Hirose T. (2004) Photosynthesis or
persistence: nitrogen allocation in leaves of evergreen and
deciduous Quercus species.Plant,Cell & Environment 27, 1047–
1054.
Warren C.R. (2008) Stand aside stomata, another actor deserves
centre stage: the forgotten role of the internal conductance to
CO2transfer. Journal of Experimental Botany 59, 1475–1487.
Warren C.R. & Adams M.A. (2004) What determines rates of
photosynthesis per unit nitrogen in eucalyptus seedlings? Func-
tional Plant Biology 31, 1169–1178.
Warren C.R., Adams M.A. & Chen Z.L. (2000) Is photosynthesis
related to concentrations of nitrogen and Rubisco in leaves of
Australian native plants? Australian Journal of Plant Physiology
27, 407–416.
Wright I.J.,Reich P.B.,Westoby M.,et al. (2004) The worldwide leaf
economics spectrum. Nature 428, 821–827.
Yasumura Y., Hikosaka K. & Hirose T. (2006) Seasonal changes in
photosynthesis, nitrogen content and nitrogen partitioning in
Lindera umbellata leaves grown in high or low irradiance. Tree
Physiology 26, 1315–1323.
Received 10 September 2008; received in revised form 3 November
2008; accepted for publication 3 November 2008
SUPPORTING INFORMATION
Additional Supporting Information may be found in the
online version of this article:
Table S1. Leaf morphological, gas exchange and biochemi-
cal mean values for the congeneric species sampled in the
Australian National Botanic Gardens and for the Eucalyp-
tus species grown in the greenhouse.
Please note: Wiley-Blackwell are not responsible for the
content or functionality of any supporting materials
supplied by the authors. Any queries (other than missing
material) should be directed to the corresponding author
for the article.
270
M. T. Harrison et al.
© 2009 The Authors
Journal compilation © 2009 Blackwell Publishing Ltd, Plant, Cell and Environment, 32, 259–270