Limnol. Oceanogr., 36(S), 1991, 1772-1782 0 1991, by the American Society of Limnology and Oceanography, Inc.

Effect of iron limitation on photosynthesis in a marine diatom

Richard M. Greene Oceanographic and Atmospheric Sciences Division, Brookhaven National Laboratory, Upton, New York 11973

Richard J. Geider College of Marine Studies, University of Delaware, Lewes 19958

Paul G. Falkowski Gceanographic and Atmospheric Sciences Division, Brookhaven National Laboratory

Abstract The response of the marine diatom Phaeodactylum ,tricornutum to Fe deficiency was evaluated

in the context of fundamental physiological models of growth and photosynthesis. Fe deficiency induced chlorosis, which decreased Chl a : C ratios and Chl a-specific light-saturated photosynthesis (P,“). In contrast to Pme, (Ye was slightly increased under Fe deficiency, and photosynthesis in Fe- deficient cells became light-saturated at lower irradiances than in Fe-replete cells grown at the same irradiance. Fe deficiency increased the in vivo absorption cross section normalized to Chl a (a*), but decreased the maximum quantum yield of photosynthesis (4,). Thus, the product a* &,,, which equals the Chl a-specific initial slope of the photosynthlesis-irradiance curve (@), was less sensitive to Fe limitation than was a* or 4, alone. Using a pump-and-probe fluorometer, we found that Fe deficiency reduced the maximum fluorescence yield (A@,,), which is consistent with the reduction in &, but increased the absorption cross section of photosystem 2 (gps2). Immunoassays of proteins separated electrophoretically indicated that the reduction in maximum fluorescence yields was accompanied by a reduction in the relative abundance of D 1, the photosystem 2 reaction center protein. Light-harvesting chlorophyll proteins (LHCP) and the large and small subunits of ribulose bisphosphate carboxylase were not affected by Fe deficiency. Changes in the abundance of Dl relative to LHCP suggest an increase in the fraction of nonfunctional reaction centers under Fe- limited conditions. Fe-deficient cells, growing at ~20% of their maximum growth rate, had reduced cellular C, N, and P contents, but maintained C : N : P ratios at the Redfield proportions. These results imply that C : N : P ratios do not provide an unequivocal index of relative growth rate.

The underlying goal in a variety of phy- toplankton growth models is to provide a general and internally consistent “mecha- nistic” description of phytoplankton growth, photosynthesis, and respiration (Bannister 1979; Shuter 1979; Kiefer and Mitchell 1983; Laws et al. 1983, 1985; Sakshaug et al. 1989). Although varying in form, all the models are based on an energy balance for phytoplankton growth that equates the gross photosynthesis rate to the sum of growth and respiration (Geider et al. 1986), assum- ing other terms such as excretion of dis-

Acknowledgments We are grateful to Z. Kolber and K. Wyman for

technical assistance and discussions, D. Henry for pu- rification of LHCP, and J. Hirschberg for the gift of the Dl antibody.

This research was supported by the U.S. DOE con- tract DE-AC02-76CHOOO 16 to P.G.F. and by NSF grant OCE 89-15084 to R.J.G.

solved organic material are minimal. The energy balance can be written simply as

p + .R = P,” 8 tanh( -aBZ/PmB) (1)

where p is the growth rate (d-l), R the res- piration rate (d-l), 8 the Chl a : C ratio, P,” the Chl a-specific, light-saturated photosyn- thesis rate, (xtl the Chl a-specific initial slope of the PI curve, and Z the irradiance. It is sometimes informative to simplify Eq. 1 and consider cases of light limitation (Eq. 2) or saturation (Eq. 3) of photosynthesis and growth:

p + R = 9aBZ (2) pi- R=8P,“. (3)

14pplications of Eq. 1 most often consider phytoplankton growth in response to limi- tation by light, temperature, or nutrient (particularly N03-) availability. Recent ex- perimental evidence, however, suggests that Fe may limit phytoplankton production in

1772

Fe limitation of photosynthesis 1773

the open ocean. Martin and coworkers (Martin and Fitzwater 1988; Martin et al. 1990) showed that Fe addition to natural phytoplankton samples stimulated Chl a synthesis and N03- utilization compared to control samples with no added Fe. Because of the inherent problems associated with long-term, nutrient-addition bottle experi- ments, the evidence for Fe limitation is still controversial (Banse 1990). Here we ex- amine the fundamental parameters of the phytoplankton growth model (Eq. 1) as they relate to the question of Fe limitation of phytoplankton growth.

We are also concerned with developing a means of identifying and interpreting inter- nal physiological processes or “diagnostic” signals which reflect an external, environ- mentally imposed limitation (i.e. Fe). An environmental constraint or limitation which displaces a phytoplankton species or community away from its maximum spe- cific growth rate is, in effect, a physiological stress. The stress may trigger biophysical, biochemical, and(or) molecular processes unique to the limiting factor. Diagnostic markers or signals should allow one to in- terrogate, without experimentally manipu- lating, natural phytoplankton communities. We consider this approach attractive, par- ticularly in the case of Fe limitation, because it circumvents the need for long-term bottle incubations supplemented with various sus- pected limiting factors.

Materials and methods The diatom Phaeodactylum tricornutum

was grown in &liter unialgal batch cultures in artificial seawater medium (ASW, Gold- man and McCarthy 1978) enriched with f/2 nutrients and trace metals (Guillard and Ryther lg62), except for Fe which was add- ed separately. The cultures were grown in continuous light (250 pmol quanta m-2 s-l) at 18°C with constant aeration and mixing. Fe-replete cultures were supplemented with 1 PM Fe and 1 PM EDTA. Media were ster- ilized by autoclaving. Fe-replete, exponen- tial-phase cells were kept optically thin for several days by daily transfer into fresh me- dia. The growth rate in Fe-replete cultures was 1.5 d-l. Fe-replete cultures were har- vested when the cell density reached - lo6

ml-l, which corresponded to - 10% of the maximum yield under Fe-replete condi- tions.

Eight-liter polycarbonate bottles contain- ing ASW media with no added Fe were in- oculated from an exponential phase culture to obtain Fe-deficient cells. Trace amounts of Fe, sufficient to support a yield of - lo6 cells ml-‘, were introduced through impu- rities in the salts used to prepare our arti- ficial media and during autoclaving. Growth continued in the Fe-limited cultures at a rate of - 0.2 d-l prior to harvesting for physi- ological and biochemical investigations. Addition of 1 PM Fe-EDTA to these cul- tures induced rapid chlorophyll synthesis and cell division indicating that growth was Fe limited. Physiological and biochemical measurements (described below) were con- ducted over three successive days on Fe- replete and on Fe-limited cultures. Cell den- sities were determined with a hemacytom- eter.

The minimum cell Fe quota under Fe- limited conditions was determined by mea- suring cell yield as a function of total added Fe. An Fe-EDTA solution was added to ali- quots of a stationary phase culture contain- ing -6 x lo5 cells ml-l to obtain enrich- ments ranging from 5 to 40 nM added Fe. In vivo fluorescence, measured with a Tur- ner Designs lo-005 fluorometer, and cell yield were determined when the cultures reached stationary phase. All plasticware was thoroughly acid washed and then sterilized in a microwave oven by bringing distilled water to boiling (for 10 min). In this exper- iment, the culture media was filter sterilized (0.2-pm pore size). We calculated a back- ground Fe concentration in the culture me- dia of 6 nM by extrapolation of the linear regression of cell yield vs. added Fe to the x-axis (i.e. zero cell yield).

Chl a and c and carotenoids were mea- sured spectrophotometrically as described by Dubinsky et al. (1986). Cells were filtered on Whatman GF/C glass-fiber filters and homogenized in a glass tissue grinder in 90% acetone. The homogenate was clarified by centrifugation and pigment concentrations were calculated with the equations of Jeffrey and Humphrey (1975). Total carotenoid concentration was calculated using an ex-

1774 Greene et al.

tinction coefficient of 2,500 mM-l at 480 nm (Davies 1976).

The in vivo spectral absorption cross sec- tion (a*,, X = 375-750 nm) was determined on cell suspensions in the scattering sample accessory of an Aminco DW 2C spectro- photometer with ASW media in the refer- ence cuvette. The spectrally averaged in vivo absorption cross section normalized to Chl a (a*), was calculated from convolution of LE*, with the emission spectrum for the lamp used in our photosynthesis measurements (Dubinsky et al. 1986).

Photosynthesis-irradiance relationships were measured with a YSI 533 1 Clark-type O2 electrode in a thermostatted PVC cham- ber with a flat-glass surface as described by Dubinsky et al. (1987). A collimated light beam from a quartz-halogen source was fil- tered through a heat filter and attenuated with neutral density filters. Irradiance at the rear of the sample chamber was measured with a Lambda LiCor 185 cosine quantum sensor, and the mean irradiance was cor- rected for attenuation by the cell suspen- sion. Before measurements, the cell suspen- sion was gently bubbled with N2 to reach 70-80% 0, saturation. Dark respiration rates were measured initially, and the pho- tosynthetic parameters P,” and Zk were cal- culated from nonlinear least-squares fit to the hyperbolic tangent equation PB = Pm” tanh(Z/ZJ. The initial slope (a”) was cal- culated as P,,,B/Z,, and the quantum require- ment of photosynthesis (I/&,) was calculat- ed as a*laB.

Oxygen flash yields (PSIJ-0,) were mea- sured with a Rank Brothers O2 electrode at flash frequencies ranging from 5 to 50 s-l as described by Falkowski et al. (198 1) and Dubinsky et al. (1986). Chl a : PToO ratios were determined from light-induced ab- sorption differences at 697 nm relative to 720 nm in 0.05% Triton X- 100 cell extracts (Falkowski et al. 198 1). Chl a : cytochrome 66/f ratios were determined from hydro- quinone-reduced minus ferricyanide-oxi- dized difference spectra (Bendall et al. 197 1). The cellular concentrations of PToO and cy- tochrome be/f were calculated from cellular Chl a content and Chl a : P,,,O and Chl a : cyt b6/f ratios, respectively. Total cellular C and N were measured on cells collected

on precombusted GF/C glass-fiber filters with a Perkin Elmer 240B elemental ana- lyzer.

Changes in fluorescence yield (A$) were determined with the pump-and-probe flu- orescence technique (Falkowski et al. 1986; Kolber et al. 1990). Cell suspensions were dark adapted for 30 min before fluorescence meiasurements. The fluorescence yield of a weak probe flash was measured before (F,) and after (Fs) a saturating pump flash. The change in the maximum fluorescence yield (A&J was calculated as (F, - FJIF,. Flash intensity saturation curves for A$ were gen- erated by changing the intensity of the pump flash from very weak to very intense flash energy. The absorption cross section of pho- tosystem 2 (a,,,) was calculated from flash intensity saturation curves by nonlinear re- gression to a cumulative one-hit Poisson distribution,

WA&,, = 1 - exp(- flps21) where I is the flash intensity.

(4)

Samples for protein analysis were con- centrated by centrifugation and washed in 0.01 M Tris, 0.0 1 M EDTA, and 0.4 M NaCl (pH 8). After centrifugation, the sam- ples were resuspended in a solution con- taining 8% sodium dodecyl sulfate (SDS), 0.2 M CC&, and 200 PM phenylmethyl sul- fonyl fluoride (PMSF) and subjected to three 30-s cycles of sonication on ice with a Kontes microprobe sonicator at maximum energy output. The samples were centrifuged for 5 min at 14,000 rpm in an Eppendorf micro- centrifuge and the pellet containing cell de- bris was discarded. Total protein content was determined with the Pierce BCA pro- tein assay on subsamples taken from the supernatant. The extract was diluted with an equal volume of 0.2 M dithiothreitol and two volumes of 4% SDS, 15% glycerol, and 0.05% Bromthymol Blue. At this point, the salmples were stored at. - 80°C until further analysis.

Protein samples were thawed on ice, heat- ed in a boiling water bath for 2 min, and centrifuged for 5 min at 14,000 rpm in an Eppendorf microcentrifuge. Proteins were loaded onto 15% SDS-polyacrylamide slab gels with a 7% stacking gel and separated elcctrophoretically at 125 V by means of

Fe limitation of photosynthesis 1775

Table 1. Light-harvesting pigments and electron transport components of Fe-replete and Fe-deficient cultures of Phaeodactylum tricornutum. Values represent the mean and relative standard deviation (RSD) of 3-4 replicate samples.

Chl u, pg cell-’ Chl c/a, mol mol-* Carotenoid : Chl a, mol mol-’ PSU-02, Chl a: 0, (mol mol-I) Chl a: RC2, mol mol-’ Chl a : PToO, mol mol-’ Chl a : b6/f, mol mold1 RC2 cell-‘, mol x 10-l” RCl cell-l, mol x lo-l9 Cyt b,/‘cell-I, mol x lo-l9 RC2 : b,/f: RC 1 Chl a : C, mg g-l Chl u : N, mg g-l

Fe-replete

0.250 (8%) 0.160 (5%) 0.62 (2%)

1,193 (8%) 298 (8%) 879

KOZ 114 9.5 (152) 3.3 (23%)

25.3 (29%) 2.9 : 7.9 : 1

19.4 (6%) 102.0 (7%)

Fe-deficient

0.057 (14%) 0.256 (6%) 0.84 (10%)

2,049 (26%) 512

2,115 ‘,:Y 113 (12)

1.3 (19%) 0.30 (17%) 5.9 (25%) 4.2 : 19 : 1 6.0 (24%)

33.2 (17%)

Laemmli’s (1970) buffer system. Samples were loaded on an equal protein basis (25 pg of protein per lane). Gels were stained with Coomassie Blue or electrophoretically transferred to nitrocellulose (Towbin et al. 1979). The proteins on nitrocellulose were probed, using the Western blot procedure (Bennett et al. 1984), with polyclonal anti- bodies raised against the light-harvesting chlorophyll (LHC) protein of the hapto- phyte Zsocrysis glabana, the photosystem 2 reaction center protein (Dl) of the chloro- phyte Amaranthus hybridus, and the large (LSU) and small (SSU) subunits of ribulose bisphosphate carboxylase (Rubisco) of I. glabana.

Results and discussion Fe limitation and phytoplankton growth

energetics-The C-specific rate of gross

photosynthesis (Eq. 1) is determined by ir- radiance, the Chl a : C ratio (e), PmB, and &. Because many Fe-containing proteins are involved in mediating photosynthetic elec- tron transport and Chl biosynthesis (Raven 1988, 1990), Fe deficiency is expected to directly affect the terms on the left-hand side of Eq. 1. In fact, we measured a 3-fold de- crease in the Chl a : C ratio (Table l), a 2- fold reduction in PmB, and 1.3-fold increase in cy” (Table 2, Fig. 1) between Fe-replete cells growing at 1.5 d-l and Fe-deficient cells growing at 0.2 d-l. Because Zk (=PmElaB) decreased from 4 12 to 133 pmol quanta mm2 s-l (Table 2) between Fe-replete and Fe- deficient cells, the growth irradiance in the culture vessels (250 pmol quanta m-2 s-l) was within the light-saturated portion of the PI curve of Fe-deficient cells, but within the light-limited region of the PI curve of Fe-

Table 2. Photosynthesis and fluorescence yield parameters of Fe-replete and Fe-deficient cultures of Phaeo- dactylurn tricornutum. The values represent the mean and relative standard deviation (RSD) of 3-4 replicate samples.

Fe-replete Fe-deficient

Pm, pm01 0, cell-’ min-’ x lo-lo P,,,n, mol 0, (mol Chl)-’ min-l (Y, pm01 0, cell-’ min-’ (pm01 quanta m-* s-‘)-I cxR, mol 0, (mol Chl)-’ min-’ (pm01 quanta m-* s-‘)-I Ik, pm01 quanta m-* s-l R, pmol 0, cell-’ min-’ x lO-‘O a*, m* (mg Chl a)-’ CZECH, m* (mg Chl a)-] 7, ms l/4,, quanta 0,-l CPS2, A* quanta-’ AA,,

28.2 (15%) 10.0 (5%) 0.068 (23%) 0.024 (16%)

412 (9%) 4.21 (17%) 0.0080 (8%) 0.0215 (12%) 5.03 (6%)

18 (12%) 214 (12%)

1.55 (5%)

2.9 (20%) (16%)

E-20 (17%) 0.032 (11%)

133 (10%) 1.38 (17%) 0.0170 (10%) 0.0481 (13%) 7.0 (17%)

29 ( 12%) 484 (13%)

0.48 (8%)

1776 Greene et al.

. +Fe

0 0

I (pm01 quanta m -2 s -‘)

Fig. 1. Representative curves of cell-specific (A) and Chl u-specific (B) photosynthesis vs. irradiance for Fe- replete and Fe-deficient cultures of Phaeodactylum tri- cornutum.

replete cells. Thus, the growth energy bal- ance (Eq. 1) is dominated by 0 and P,,,” in our Fe-deficient cells and by 0 and CY~ in our Fe-replete cultures.

We calculated C-specific gross photosyn- thetic rates of 3.80 mol0, (ma1 C)-l d-l in Fe-replete cultures and 0.54 in Fe-deficient cultures which were, in both cases, about twice the net growth rate (given a photo- synthetic quotient, PQ, of 1.3, calculated from C : N ratios, Falkowski et al. 1985). In Fe-replete cells there was an excess of pho- tosynthesis over the maximum energy de- mands for cell growth because, at the growth irradiance, growth rates were light saturated but photosynthesis was not. The gross pho- tosynthetic rate (PB) of Fe-replete cells at the growth irradiance was 2.3 mol O2 (mol C)--l d-l, - 60% of P,” and 18% greater than the growth rate (given PQ = 1.3). The dif- ference between p and P,” can be attributed to respiration, which accounted for - 14% of P,” in Fe-replete cells.

Gross photosynthesis is assumed to be partitioned between growth and respiration in Eq. 2. Despite a decline in C-specific res- piration rates from 0.57 to 0.26 mol O2 (mol

C)- l d-l between Fe-replete and Fe-deficient conditions, respiration accounted for a gre,ater proportion of gross photosynthesis under Fe-deficient conditions (- 50%). We cannot ascertain whether the decline in res- piration was due to an impact of Fe avail- ability on the size or integrity of the mito- chondrial electron transport chain or to a decline in substrate availability associated with the reduction in gross photosynthesis. The measured respiration rate in Fe-defi- cient cells was well above the 0.02 d-l min- imum maintenance metabolic rate estimat- ed for light-limited P. tricornutum (Geider and Osborne 1989). Maintenance metabo- lism in cells grown in saturating irradiance may exceed that calculated for the limit as irradiance approaches zero because of the potential additional demand for repair from light-induced damage. Our respiration rates (measured after a 30-min dark period), and by implication calculated gross photosyn- thesis rates, may be underestimates if res- piration is stimulated in the light (Weger et al. 1989).

Fe limitation and variations in PmB and @--Fe limitation reduced P,” by a factor of two and slightly increased (x~ (Fig. 1, Ta- ble 2). Despite large efforts devoted to mea- suring PI parameters in marine waters, the environmental determinants of P,” and (x~ have proven difficult to identify. We now investigate, within the context of current models of photosynthesis, the fundamental biophysical mechanisms for Fe-induced changes in P,” and aB.

I’,” is determined by the concentration of photosynthetic units (n) and the minimal turnover time (T) required for an electron to pass from water to CO, (Herron and Mauzerall 19 72):

en” = n/r (5) where n is the ratio of O2 to Chl a (l/PSU- 0,) determined experimentally from mea- surements of O2 flash yield. In Fe-deficient cells, the reduction in P, per cell resulted from a large decrease in the number of PS2 reaction centers (RC2, Table 1). The de- crease in Pm”, as well as the decrease in P, calculated per RC2, in Fe-deficient cells can be explained by the increase in T (Table 2). Additionally, Fe deficiency reduced the cel-

Fe limitation of photosynthesis 1777

Wavelength (nm)

Fig. 2. Representative spectra showing the in vivo absorption cross section normalized to Chl a (a*) be- tween 375 and 750 nm for Fe-replete and Fe-deficient cultures of Phaeodactylum tricornutum.

lular concentration of PSl reaction centers (RCl) and cyt b,/f (Table 1). Fe limitation has been shown to reduce photosynthetic rates and RC concentrations in other species of microalgae as well (Glover 1977; Sand- mann 1985).

The initial slope of the photosynthesis vs. irradiance curve (cuB) is determined as

aB = a%& (6) where a* is the in vivo absorption cross section normalized to Chl a, and 4, the quantum yield of photosynthesis (0, quan- ta-l). Fe deficiency in P. tricornutum, as well as N deficiency in other microalgae (Herzig and Falkowski 1989; Chalup and Laws 1990), increases a* but reduces 4, (Fig. 2, Table 2). Although cyB is slightly higher in Fe-deficient cells, it is clear that the product a* $, is less sensitive to Fe limitation than are the individual parameters alone. The increase in a* is attributable to an overall decrease in light-harvesting pigment con- tent, which reduces the “package” effect (Berner et al. 1989). The minimum quan- tum requirement of photosynthesis ( l/&J, calculated from Eq. 6, increased from 18 quanta 02-1 in Fe-replete to 29 quanta 0,-l in Fe-deficient cells (Table 2). The Fe-in- duced changes in l/4, suggest that one or more processes involved in light absorption and excitation energy transfer are affected by Fe deficiency.

Fe limitation and changes in fluorescence yield-Estimates of the fluorescence yield of Chl have been a valuable tool not only for probing the primarv bionhvsical events of

1.6

1.2

4s 0,s

0.4

0.0

0.10 1 .oo 10.00

Relatfve flash intensity

Fig. 3. A. Flash intensity saturation curves for changes in fluorescence yield (A4) measured with the pump-and-probe fluorometer for Fe-replete and Fe- deficient cultures of Phaeodactylum tricornutum. B. Normalized flash intensity saturation curves for A@.

photosynthesis (Mauzerall 1972; Falkowski and Kiefer 1985) but also as a means of estimating phytoplankton biomass in the field (Lorenzen 1966; Falkowski and Kiefer 1985). We examined the effect of Fe limi- tation on the changes in the quantum yields of fluorescence with the pump-and-probe fluorescence technique (Falkowski et al. 1986, 1988). The theoretical basis for the pump-and-probe technique has been re- viewed elsewhere (Falkowski and Kiefer 1985).

Maximum fluorescence yields were ap- proximately 3 times lower in Fe-deficient than in Fe-replete cultures (Fig. 3A, Table 2). The reduction in A&,, is consistent with the reduction in &,., and P,“. When nor- malized to the maximum yield, the resulting curves saturated at lower flash intensities in Fe-deficient cells than in Fe-replete cells (Fig. 3B). The absorption cross section of PS2 (a,,,) calculated from Eq. 4, was more than 2 times higher in Fe-deficient than in Fe- replete cells (Table 2). The increase in cps2 in Fe-deficient cells, like the increase in a*, can be attributed to a reduction in the pack- age effect and an increase in accessory pig- ments relative to Chl a (Table 1). Because Chl c and fucoxanthin are associated solelv

1778 Greene et al.

with the light-harvesting complex of PS2 (Owens 1986), the change in pigment ratios indicates an increase in light-harvesting ca- pability in Fe-deficient cells.

Similar to our results for Fe-limited cells, N limitation, but not light limitation, re- duces A&,, and increases ups2 in various phytoplankton species (Kolber et al. 1988). A4 in our Fe-replete, as well as N-replete, cells (Kolber et al. 1988), increases faster than the calculated fit to the Poisson dis- tribution (Fig. 3B). The cause for this de- viation may be due to the transfer of exci- tation energy from closed to open reaction centers (Paillotin 1976).

The minimum quantum requirement of photosynthesis (l/4,) can be expressed as

14JIn = ~PSU&PS2 (7)

where cpsu-02 is the effective optical cross section per molecule of O2 produced [A2 (molecule 0,))‘1, and ups2 is the absolute absorption cross section of PS2 (A2 quan- ta-l) calculated from flash intensity satu- ration curves of fluorescence yield (Ley and Mauzerall 1982; Dubinsky et al. 1986). ~psu-02 is calculated as

UPSU-02 = PSU-02a*bXE u-9 where PSU-O, is the 0, flash yield [mol Chl(mo1 0,)) ‘1, a*bx the in vivo absorption cross section, normalized to Chl a, within the optical region of the blue-xenon flash lamp of the pump-and-probe fluorometer (Table 2), and E a dimensional conversion factor. From Eq. Y and 8, we calculated that l/+, increased from 18 quanta 0,-l in Fe- replete cells to 30 in Fe-deficient cells. These values are identical to those calculated from PZ parameters (Eq. 6, Table 2) and provide a completely independent measure of the minimum quantum requirement of photo- synthesis.

Fe limitation and cellular proteins-The response of P. tricornutum to Fe limitation involves major changes in physiological and bio-optical parameters. Using immunoas- says on proteins separated by SDS-PAGE (Fig. 4A), we examined the effect of Fe lim- itation on the relative abundance of specific cellular proteins involved in photosynthe- sis.

When examined on an equal protein ba-

sis, Fe deficiency caused a pronounced re- duction in the relative abundance of D 1, the PS2 reaction center protein (Fig. 4B). The loss of Dl is consistent with the re- duction in functional RC2 content (Table 1). Fe deficiency seems to affect other PS2 proteins as well. Our preliminary results suggest a reduction in the relative abun- dance of CP43 and CP47, two proteins which mediate excitation energy transfer from the antenna to the reaction center (Bassi et al. 1987). However, Fe deficiency had little affect on the relative abundance of the nuclear-encoded LHC proteins (Fig. 4B). A decrease in the abundance of Dl relative to ‘LHCP would imply an increase in the fraction of nonfunctional reaction centers (i.e. PS2 complexes lacking 01). It appears that the apoproteins of the LHC, which are encoded in the nuclear genome, may be synthesized under low Fe conditions even though light-harvesting pigment syn- thesis is greatly reduced. This may indicate that some LHCPs are not coupled to reac- tion centers. The loss of critical PS2 proteins in our Fe-deficient cells, as well as in N-lim- ited cultures (Falkowski et al. 1989), reduces excitation energy transfer and results in a marked decrease in A$,, (Kolber et al. 1988).

Similar to the LHC proteins, Fe deficien- cy h.ad no effect on the chlorop’last-encoded large and small subunits of Rubisco (Fig. 4B). This deficiency contrasts N limitation, in which chloroplast-encoded proteins are reduced preferentially over those synthe- sized by the nuclear genome (Kolber et al. 1988; Falkowski et al. 1989). Thus, the ra- tios of Dl to LHCPs and of Dl to Rubisco LSIJs are useful for differentiating between N and Fe limitation. In addition, two pro- teins of unknown function, one at 24.4 kDa which cross-reacts with the LHCP antibody (Fig. 4B) and one at 25 kDa, appear in great- er relative abundance in Fe-deficient cells tha:n in Fe-replete cells (Fig. 4A); both are potentially useful diagnostic markers of Fe limitation.

Fe limitation and elemental composition: Implications for use of C: N:P ratios as to indicate relative growth rate-The cellular content of C, N, and P were lower in Fe- deficient than in Fe-replete cells (Table 3), indicating that in addition to a reduction in

A

kDa

Fe limitation qfphotosvnthessis

B

-Fe +Fe

1779

-Fe +Fe

Fig. 4. A. Comptison of protein banding patterns between Fe-deficient and Fe-replete Phaeodactylum rricornutum. Proteins were separated by electrophoresis on 15% SDS-polyacrylamide gels and stained with Coomassie Blue. Proteins were loaded to give equal protein concentrations for each lane. B. Western blots of total proteins in Fe-deficient and Fe-replete P. tricornutum probed with polyclonal antibodies for the photosystem 2 reaction center protein (Dl), the light-harvesting chlorophyll (UK) protein. and the large and small subunits (LSU and SW) of Rubisco.

C assimilation, Fe deficiency also decreased both N and P assimilation. The C : N : P ra- tio, however, remained close to the Redfield proportions in both Fe-deficient and Fe-re- plete cells (Table 3).

The minimum cell Fe quota, determined from the slope of a linear regression of cell yield vs. added Fe (Fig. 5), ranged from 0.005 to 0.01 fmol Fe cell-‘. Given a cell C quota of 0.78 pm01 cell-’ (Table 3), we calculate a maximum C : Fe ratio for Fe-deficient cells of 78,000: 1, which may overestimate the C: Fe ratio if the added Fe was not com- pletely incorporated. However, our C : Fe ratios agree with those of Anderson and Morel (1982), who reported a 100,000: 1 C : Fe ratio in Fe-deficient cells compared to 10,000: 1 in Fe-replete cells. Given Fe requirements of the various electron trans- port components ranging from 12 Fe : PS 1 to 6 Fe: PS2 and 5 Fe :cyt be/J: our mea- surements indicate that 40% of intracellular

Fe in Fe-deficient cells can be accounted for by the PSI, PS2, and cyt b,/f complexes, with the cyt b,/fcomplex alone accounting for 30% of cellular Fe.

Our results demonstrate that the C : N ra- tio remains low in Fe-limited cells growing at <20X of their maximum growth rate, which is consistent with independent ob- servations (Glover 1977; Rueter and Ades 1987) showing that the fraction of ‘%Z in-

Table 3. Elemental composition (pg cell ‘) and mo- lar clemental ratio of Fe-replete and Fe-deficient cul- tures of Phaeodactylum tricornufum. Values represent the mean and relative standard deviation (RSD) of 3 or 4 replicate samples.

C 12.81 (10%) 9.32 (11%) N 2.45 (9%) 1.68 (7%) P 0.314 (18%) 0.243 (19%) C:N:P 109:18:1 103:16:1

Greene et al.

10 20 30

Added Fe (pt4)

40 50

2.5 ?

El 0

2.0 . 1.5 * r/’

1.0 -

0.5 -

XT-cl

0 0

0.0 0 loo 200 300 400 500

xl O4 cells ml-’

Fig. 5. A. Yield of Phaeodactylum tricornutum in stationary phase cultures as a function of added Fe- EDTA. B. Relation between in vivo fluorescence and cell density in stationary phase cultures illustrated in panel A.

corporated into TCA-insoluble products (i.e. the protein fraction) remains high in Fe- limited cells at low growth rates and implies that the C : N ratio does not provide an un- ambiguous measure of relative growth rate under Fe-deficient conditions. These results have important implications for the use of elemental composition (Goldman 1980) and biochemical fractionation of photosynthet- ically fixed 14C (DiTullio and Laws 198 6) as indices of relative growth rate in oligo- trophic ocean regions. Goldman (1980), on the basis of the observed Redfield ratio of particulate matter, and Laws et al. (1987, 1989), on the basis of the relative incor- poration of 14C into protein, have suggested that open-ocean phytoplankton are often growing at near maximal rates (> 0.7 p/pm). However, recent evidence based on mea- surements of A&, with a pump-and-probe fluorometer indicates that phytoplankton are growing at <0.2 p/p, in some oligotrophic ocean regions (Falkowski et al. 199 1). Al- though spatial and temporal variability may

account for the contradictory conclusions, our evidence for Redfield ratio elemental composition, and low A&,t, suggests that phytoplankton growth may be limited by Fe availability more o,ften than is currently suspected.

Conclusion Oceanographers have long contended that

Fe, or some other trace metal elements, may be ‘biolimiting for marine primary produc- tivity in the open ocean (Hart 1934; Menzel and Ryther 196 1). Our results have impli- cations for understanding the impact of Fe deficiency on phytoplankton physiology and provide potential diagnostic tools for elu- cid,ating Fe limitation in the ocean.

Fe-replete and Fe-deficient cells exhibit large variations in optical, photosynthetic, and biochemical characteristics, all of which affect the energy balance in phytoplankton growth models. The reduction in cell-spe- cific photosynthesis in Fe-deficient cells can be *attributed to a reduction both in cell Chl a content and in Chl a-specific light-satu- rated photosynthesis. In addition, fluores- cence per unit Chl is higher in Fe-deficient than in Fe-replete cells (Sakshaug and Holm- Hansen 1977; Rueter and Ades 1987). Fe deficiency induces chlorosis, which increas- es both the in vivo absorption cross section (a*) and the absorption cross section of PS2 (up&. Thus, Fe-limited cells become light saturated at lower irradiances than Fe-re- plete cells. Evidence for Redfield ratio ele- mental composition in our Fe-deficient cells suggests that, although the metabolic de- mands for growth are reduced, Fe-deficient cells are a’ble to meet their nutrient require- ments and maintain low C : N ratios. Thus, “Fe-deficient cultures otherwise give the impression that their chemical composition and appearance is close to that of healthy, exponential cells, except for being chlorot- ic”’ (Sakshaug and Holm-Hansen 1977, p. 27).

‘We have identified some unique bio- physical and molecular markers in Fe-de- ficient P. tricornutum which, if ubiquitous, may provide evidence for physiological lim- itation in natural phytoplankton commu- nities by Fe availability. Let us consider the nutrient-rich areas of the open ocean, where

Fe limitation of photosynthesis 1781

macronutrient (N and P) concentrations ex- ceed requirements for phytoplankton growth (Martin et al. 1990). If phytoplankton pho- tosynthesis were limited by Fe availability in these regions, one might expect to mea- sure low cellular Chl a, low A&l, and low C : N and Chl a : C ratios. Questions of light limitation, particularly in the Southern Ocean, can be evaluated in a similar manner because A4sat is independent of growth ir- radiance (Kolber et al. 1988) and cellular Chl a is inversely related to growth irradi- ance. On a molecular level, Fe limitation reduces the relative abundance of Dl, but has little or no effect on LHC proteins and Rubisco subunits. Although the latter two characteristics are unique to Fe-deficient cells, the former also occurs in NO,--defi- cient cells (Kolber et al. 1988). We suggest that evaluating the biophysical and molec- ular signals in phytoplankton populations may provide a better understanding of the factors that limit primary productivity in the ocean.

References ANDERSON, M. A., AND F. M. M. MOREL. 1982. The

influence of aqueous iron chemistry on the uptake of iron by the coastal diatom Thalassiosira weiss- jlogii. Limnol. Oceanogr. 27: 789-8 13.

BANNISTER, T. T. 1979. Quantitative description of steady state nutrient-saturated algal growth, in- cluding adaptation. Limnol. Oceanogr. 24: 76-96.

BANSE, K. 1990. Does iron really limit phytoplankton production in the offshore subarctic Pacific? Lim- nol. Oceanogr. 35: 772-775.

BASSI, R., G. HOYER-HANSEN, R. BARBATO, G. M. GIA- COMETTI, AND D. J. SIMPSON. 1987. Chlorophyll- proteins of the photosystem II antenna system. J. Biol. Chem. 262: 133-141.

BENDALL, D. S., H. E. DAVENPORT, AND R. HILL. 197 1, Cytochrome components in chloroplasts of the higher plants. Methods Enzymol. 23A: 327-344.

BENNETT, J., G. I. JENKINS, AND M. R. HARTLEY. 1984. Differential regulation of the accumulation of the light-harvesting chlorophyll a/b complex and ri- bulose bisphosphate carboxylase/oxygenase in greening pea leaves. J. Cell Biochem. 25: 1-13.

BERNER, T., Z. DUBINSKY, K. WYMAN, AND P. G. FAL- KOWSKI. 1989. Photoadaptation and the “pack- age” effect in DunalielIa tertiolecta (Chlorophy- ceae). J. Phycol. 25: 70-78.

CHALUP, M. S., AND E. A. LAWS. 1990. A test of the assumptions and predictions of recent microalgal growth models with the marine phytoplankter Pavlova lutheri. Limnol. Oceanogr. 35: 583-596.

DAVIES, B. H. 1976. Carotenoids, p. 38-166. Zn T. W. Goodwin [ed.], Chemistry and biochemistry of plant pigments. Academic.

DITULLIO, G. R., AND E. A. LAWS. 1986. Diel peri- odicity of nitrogen and carbon assimilation in five species of marine phytoplankton: Accuracy of methodology for predicting N-assimilation rates and N/C composition ratios. Mar. Ecol. Prog. Ser. 32: 123-l 32.

DUBINSKY, Z., P. G. FALKOWSKI, A. F. POST, AND V. H. VAN HES. 1987. A system for measuring phy- toplankton photosynthesis in a defined light field with an oxygen electrode. J. Plankton Res. 9: 607- 612.

- AND K. WYMAN. 1986. Light har- vekting and utilization by phytoplankton. Plant Cell Physiol. 27: 1335-l 349.

FALKOWSKI, P. G., Z. DUBINSKY, ANDK. WYMAN. 1985. Growth-irradiance relationships in phytoplank- ton. Limnol. Oceanogr. 30: 3 1 l-32 1.

- AND D. A. KIEFER. 1985. Chlorophyll a flu- orescence in phytoplankton: Relationship to pho- tosynthesis and biomass. J. Plankton Res. 7: 7 15- 731.

-, Z. KOLBER, AND Y. FUJITA. 1988. Effect of redox state on the dynamics of photosystem II during steady-state photosynthesis in eucaryotic algae. Biochim. Biophys. Acta 933: 432-443.

-, T. G. OWENS, A. C. LEY, AND D. C. MAUZERALL. - 1981. Effects of growth irradiance levels on the

ratio of reaction centers in two species of marine phytoplankton. Plant Physiol. 68: 969-973.

- A. SUKENIK, AND R. HERZIG. 1989. Nitrogen limitation in Zsochrysis galbana (Haptophyceae). 2. Relative abundance of chloroplast proteins. J. Phycol. 25: 471-478.

-, K. WYMAN, A. C. LEY, AND D. C. MAUZERALL. 1986. Relationship of steady-state photosynthe- sis to fluorescence in eucaryotic algae. Biochem. Biophys. Acta 849: 183-l 92.

-, D. ZIEMANN, Z. KOLBER, AND P. K. BIENFANG. 199 1. Nutrient pumping and phytoplankton re- sponsc in a subtropical mesoscale eddy. Nature 352: 55-58.

GEIDER, R. J., AND B. A. OSBORNE. 1989. Respiration and microalgal growth: A review of the quanti- tative relationship between dark respiration and growth. New Phytol. 112: 327-34 1.

-, T. PLATT, AND J. A. RAVEN. 1986. Size de- pendence of growth and photosynthesis in dia- toms: A synthesis. Mar. Ecol. Prog. Ser. 30: 93- 109.

GLOVER, H. E. 1977. Effects of iron deficiency on the physiology and biochemistry of Zsochrysisgalbana (Chrysophyceae) and Phaeodactylum tricornutum (Bacillariophyceae). J. Phycol. 13: 208-2 12.

GOLDMAN, J. C. 1980. Physiological processes, nu- trient availability, and the concept of relative growth rate in marine phytoplankton ecology, p. 179-194. Zn Primary productivity in the sea. Brookhaven Symp. Biol. 3 1. Plenum.

-, AND J. J. MCCARTHY. 1978. Steady state growth and ammonium uptake of a fast-growing marinc diatom. Limnol. Oceanogr. 23: 695-703.

GUILLARD, R. R. L., AND J. H. RYTHER. 1962. Studies on marine planktonic diatoms. 1. Cyclotella nana Hustedt and Detonula confervacae (Cleve) Gran. Can. J. Microbial. 8: 229-239.

1782 Greene et al.

HART, T. J. 1934. On the phytoplankton ofthe south- west Atlantic and Bellingshausen Sea, 1929-3 1. Discovery Rep. 8: 1-268.

HERRON, H. A., AND D. MAUZERALI,. 1972. The de- velopment of photosynthesis in a greening mutant of Chlorella and an analysis of the light saturation curve. Plant Physiol. 50: 141-148.

HERZIG, R., AND P. G. FALKOWSKI. 1989. Nitrogen limitation in Zsochrysis glabana (Haptophyceac). 1. Photosynthetic energy conversion and growth efficiencies. J. Phycol. 25: 462-47 1.

JEFFREY, S. W., AND G. F. HUMPHREY. 1975. New spectrophotometric equations for determining chlorophylls a, b, cl and c2 in higher plants, algae and natural phytoplankton. Biochem. Physiol. Pflanz. 167: 191-194.

KIEFER, D. A., AND B. G. MITCHELL. 1983. A simple, steady state description of phytoplankton growth based on absorption cross section and quantum efftciency. Limnol. Oceanogr. 28: 770-776.

KOLBER, Z., K. WYMAN, AND P. G. FAI.KOWSKI. 1990. Natural variability in photosynthetic energy con- version efficiency: A field study in the Gulf of Maine. Limnol. Oceanogr. 35: 72-79.

--, J. ZEHR, AND P. G. FALKOWSKI. 1988. Effects of growth irradiance and nitrogen limitation on photosynthetic energy conversion in photosystcm II. Plant Physiol. 88: 923-929.

LAEMMLI, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685.

LAWS, E. A., G. R. DITULLIO, P. 13. BETZER, D. M. KARL, AND K. L. CARDER. 1989. Autotrophic production and elemental fluxes at 26”N, 155”W in the North Pacific central gyre. Deep-Sea Res. 36: 103-120.

_.-- -, AND D. G. REDAWE. 1987. High phy- toplankton growth and production rates in the North Pacific subtropical gyre. Limnol. Oceanogr. 32: 905-9 18.

-----, D. R. JONES, K. L. TERRY, AND J. A. HIROTA. 1985. Modifications in recent models of phyto- plankton growth: Theoretical developments and experimental examination of predictions. J. Thcor. Biol. 114: 323-341.

---, D. G. REDAIJE, D. M. KARL, AND M. CHALUP. 1983. A theoretical and experimental examina- tion of the predictions of two recent models of phytoplankton growth. J. Theor. Biol. 105: 469- 491.

LEY, A. C., AND D. C. MAUZERALL. 1982. Absolute absorption cross-sections for photosystem II and the minimum quantum requirement for photo- synthesis in Chlorella vulgar-is. Biochem. Biophys. Acta 680: 95-106.

LORENZEN, C. J. 1966. A method for continuous mea- surement of in vivo chlorophyll concentration. Deep-Sea Res. 13: 223-227.

MARTIN, J. H., AND S. E. FITZWATER. 1988. Iron de-

ficiency limits phytoplankton growth in the north- east Pacific subarctic. Nature 331: 34 l-343.

--, R. M. GORDON, AND S. E. FITZWATER. 1990. Iron in Antarctic waters. Nature 345: 156-l 58.

MAIJZERALL, D. 1972. Light-induced changes in Chlorella, and the primary photoreaction for the production of oxygen. Proc. Natl. Acad. Sci. 69: 1358-1362.

MENZEL, D. W., AND J. H. RYTHER. 196 1. Nutrients limiting the production of phytoplankton in the Sargasso Sea, with special reference to iron. Deep- Sea Res. 7: 276-28 1.

OWENS, T. G. 1986. Light-harvesting function in the diatom Phaeodactylum tricornutum. Plant Phys- iol. 80: 739-746.

PAII,LOTIN, G. 1976. Movement of excitations in the photosynthetic domain of photosystem II. J. Theor. Biol. 58: 237-252.

RAVEN, J. A. 1988. The iron and molybdenum use efficiencies of plant growth with different energy, carbon and nitrogen sources. New Phytol. 109: 279-287.

--. 1990. Predictions of Mn and Fe use efficien- cies of photoautotrophic growth as a function of light availability for growth and C assimilation pathway. New Phytol. 116: l-l 8.

RUIITER, J. G., AND D. R. ADES. 1987. The role of iron nutrition in photosynthesis and nitrogen as- similation in Scenedesmus quadricauda (Chloro- phyceae). J. Phycol. 23: 452-457.

SAKSHAUG, E. K., K. ANDERSEN, AND D. A. KIEFER. 1989. A steady state description of growth and light absorption in the marine planktonic diatom Skeletonema costatum. Limnol. Oceanogr. 34: 198- 205.

-- , ANI) 0. HOLM-HANSEN. 1977. Chemical composition of Skeletonema costatum (Grev.) Cleve and Pavlova (Monochrysis) lutheri (DROOP) Green as a function of nitrate-, phosphate-, and iron-limited growth. J. Exp. Mar. Biol. Ecol. 29: l-34.

SANDMANN, G. 1985. Consequences of iron deficien- cy on photosynthetic and respiratory electron transport in blue-green algae. Photosynth. Res. 6: 261-271.

SI-IIJTER, B. 1979. A model of physiological adapta- tion in unicellular algae. J. Theor. Biol. 78: 5 19- 552.

TOWBIN, H., T. STAEHELIN, AND J. GORDON. 1979. Electrophoretic transfer of proteins from poly- acrylamide gels to nitrocellular sheets. Procedure and some applications. Proc. Natl. Acad. Sci. 76: 4350-4354.

WEGER, H. G., R. HERZIG, P. G. FALKOWSKI, AND D. H. TUR.PIN. 1989. Respiratory losses in the light in a marine diatom: Measurements by short-term mass spectrometry. Limnol. Oceanogr. 34: 1153- 1161.