I N T RO D U C T I O N

Oceanic phytoplankton are known to fix CO2 primarilythrough the activity of the enzyme ribulose-1,5-bisphos-phate carboxylase/oxygenase (Rubisco), the major carb-oxylating enzyme from the Calvin–Benson cycle inmarine microalgae (Hellebust and Terborgh, 1967;Paasche, 1971; Morris and Farrell, 1971). However, thereaction catalysed by Rubisco is not the only carboxyla-tion reaction in aquatic phototroph cells, as �-carboxyla-tion also may fix inorganic carbon. These two pathwaysoccur both in photoautotrophic (Glover and Morris, 1979;Appleby et al., 1980) and mixotrophic (Raven, 1974;Levedahl, 1978; Falkowski and Raven, 1997) cells.

Carboxylase activity measurements provide an assess-ment of carbon assimilation processes in primary produc-tion, as opposed to classical 14C incorporationmeasurements such as described by Steemann-Nielsen,which include not only fixation terms but also loss terms,such as respiration, excretion and grazing (Steemann-Nielsen, 1952). Although the information on the netcarbon fixation driving cell growth is important, themethod used may introduce serious biases inherent incontainment protocols (Leftley et al., 1983; Glover, 1989)and does not provide any information on the biochemicalprocess involved (Glover and Morris, 1979). On the otherhand, measurements of CO2 fixation by �-carboxylaseactivities (via the �-carboxylation pathway) and by

© Oxford University Press 2001

Size-fractionated carboxylase activitiesduring a 32 h cycle at 30 m depth in thenorth-western Mediterranean Sea afteran episodic wind eventERIC FOUILLAND1, CLAUDE COURTIES2 AND CHANTAL DESCOLAS-GROS3

UNIVERSITÉ MONTPELLIER II, UMR CNRS , LABORATOIRE D’HYDROBIOLOGIE MARINE ET CONTINENTALE, CC, PLACE EUGÈNE BATAILLON, MONTPELLIER CEDEX , FRANCE

PRESENT ADDRESSES: 1DUNSTAFFNAGE MARINE LABORATORY, P.O. BOX , OBAN, ARGYLL, PA AD, UK, 2LABORATOIRE MODÈLES EN BIOLOGIE

CELLULAIRE, UMR CNRS , OBSERVATOIRE OCÉANOLOGIQUE, 66551 BANYULS-SUR-MER, FRANCE, 3UNIVERSITÉ MONTPELLIER II, UMR CNRS ,ISEM, CC , PALÉOENVIRONNEMENTS ET PALYNOLOGIE, PLACE EUGÈNE BATAILLON, MONTPELLIER CEDEX , FRANCE

Using size-fractionation filtration (1 µm), we associated carboxylase activities (Rubisco, b-carboxy-

lases) and chlorophyll measurements with cell enumeration by flow cytometry at a permanent site of

the central Ligurian Sea in the north-western Mediterranean Sea (73°25�N–7°51�E). The analy-

ses were carried out over a day/night cycle (at 30 m depth) following a strong wind event, during

the transition period from spring mesotrophic to summer oligotrophic conditions. The highest values

of Rubisco activity and b-carboxylase activity per chlorophyll a (Chl a) for >1 µm cells were observed

during the light period of the cycle, reaching 18.9 and 4.3 nmol CO2 (µg Chl a)–1 h–1, respectively.

This higher activity is assumed to be correlated with a dominance of nanoflagellates in the phyto-

plankton community. Such phytoplankton species generally had higher b-carboxylase activity,

expressed as a percentage of Rubisco activity (the bC/R ratio), than diatoms. Using flow cyto-

metry analysis to enumerate those cells <1 µm in size, we followed the values of Rubisco activity

and pigment content expressed per cell, for picophytoplankton cells. The photoautotrophic activity,

measured as the in vitro Rubisco activity for small picoeukaryote cells, was higher than for cyanobac-

teria cells with lower apparent cell size. These results suggested an optimum of CO2 assimilation

reached by the pico- and nano-phytoplankton in accordance with the cell size and growth rates from

previous observations in the literature.

JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒

fouilland (2188) 7/6/01 12:27 pm Page 623

Rubisco activity (via the Calvin–Benson cycle) allow us todistinguish between the two inorganic carbon fixationpathways used. The relative importance of each pathwaymay be influenced by species composition (Glover andMorris, 1979; Descolas-Gros and Oriol, 1992), environ-mental conditions and the physiological state of cells(Glover and Morris, 1979; Descolas-Gros and Fontugne,1985; Mortain-Bertrand et al., 1988; Descolas-Gros andFontugne, 1988).

Various studies have revealed that different size frac-tions of phytoplankton have different specific ecologicalproperties in photosynthesis (Chavez, 1989). In oligo-trophic environments, picophytoplankton [organismsbetween 0.2 and 2 µm in width (Sieburth et al., 1978)] maybe responsible for up to 70% of the chlorophyll a and upto 80% of the primary production (Krupatkina, 1990).Several studies in the Mediterranean Sea, reviewed byMagazzù and Decembrini (Magazzù and Decembrini,1995), demonstrated that the picophytoplanktonic contri-bution to primary production varied from 31% to 92%.Additionally, the results indicated that the fractionmeasuring from 0.5 to 1.0 µm was much more productivethan the 1.0 to 2.0 µm fraction. It is generally consideredthat small cells have higher photosynthetic efficienciesthan large cells, due to a high ratio of surface area tovolume (Malone, 1980a, b; Krupatkina, 1990).

This paper presents the simultaneous measurements ofdifferent photosynthetic activities during a relatively highfrequency sampling over a 32 h period. Samples were col-lected at the DYFAMED time-series station site (JGOFS-France) located in the central part of the northwesternMediterranean Sea (Ligurian area) during a transitionalperiod from mesotrophic conditions in spring to summeroligotrophy, under the influence of successive wind events(Andersen and Prieur, 2000; Vidussi et al., 2000). Duringthe 32 h cycle, at a fixed depth, we focused on the smal-lest members of the picophytoplankton by measuringchlorophyll, carboxylase activities and phytoplanktoniccell enumeration by flow cytometry in the 1 µm size frac-tion. The flow cytometry method relies on the auto-fluorescence of pigments for detection and enumerationof phototrophs in the small plankton community (Yentschet al., 1983; Olson et al., 1985; Wood et al., 1985). Flowcytometry analyses offer the potential for added speci-ficity, since discrete components of the community can befollowed, thus reducing the potential for interference bychanges in the community composition (Olson et al.,1991). The combination of methods allows us to deter-mine the contribution of very small phytoplankton to thetotal carboxylase activities. It also enables us to follow thein situ variations of the carboxylase activities by size classesand, at a cellular scale, by expressing Rubisco activity percell.

M E T H O D

Sampling station

The present work was part of a more general investigationof biogeochemical processes (DYNAPROC cruise,JGOFS-France, May 1995) conducted mainly at a time-series station in the open north-western MediterraneanSea, on board the R.V. Le Suroit and Théthys II (Ander-sen and Prieur, 2000). The study was located at the per-manent DYFAMED (Buat-Ménard and Lambert, 1993)site, 28 miles off Cap Ferrat (France) in the central zone ofthe Ligurian Sea (73°25�N–7°51�E).

Fixed depth (30 m) samplings were performed fromMay 14 at 17:30 h to May 16 at 01:00 h (local time), usingtwo different, but simultaneous, sampling methods: Niskinbottles for size-fractionation filtrations and underwaterpumps for total filtrations. Samples were collected every 4 h during a 32 h cycle on board the R.V. Théthys II. Thefixed sampling depth was chosen based on the depth ofthe chlorophyll maximum as observed at the beginning ofthe DYNAPROC Cruise (May 2 1995), which was locatedat 30 m (Vidussi et al., 2000). The chlorophyll maximumwas shifted up to about 20 m after a wind event on May13 and remained at about 20 m during our study period.Although our sampling depth was just below the chloro-phyll maximum depth, the maximum of primary produc-tion performed in situ from 14 h incubations, was locatedat 30 m on May 14 and 15 (Vidussi et al., 2000).

Sample collection by underwater pumps

Sea water was collected by underwater pumps (Chal-lenger Oceanic INSU-CNRS) every 4 h during the 32 hcycle at 30 m depth, which permitted large volumes ofwater (50–100 l h–1) to be filtered on 142 mm diameter(0.7 µm pore size) glass fibre filters (Gelman). The under-water sampling was carried out less than 1 h after theNiskin bottle sampling for size-fractionation.

Size-fractionation filtrations

At each sampling time, 5 l from Niskin bottles weresampled for two simultaneous size-fractionation filtra-tions. All samples were pre-filtered through a 200 µmnylon mesh to remove larger zooplankton assemblages.Filtrations and size-fractionations were carried out with 1µm Nuclepore and GF/F Whatman filters within 1 h ofsampling (Figure 1).

Three litres of water were filtered (with pressure <230mmHg) for size-fractionated carboxylase activities (see[A] and [B] in Figure 1) and 2 l were filtered for size-frac-tionated chlorophyll measurements (see [C] and [D] inFigure 1). Subsamples were taken for flow cytometricanalysis from each fraction during the carboxylase

JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒

fouilland (2188) 7/6/01 12:27 pm Page 624

size-fractionation filtration: the first subsample was takenfrom the bulk seawater sample (see [1] in Figure 1); thesecond was taken after Nuclepore filtration (see [2] inFigure 1); and the last subsample was taken after GF/FWhatman filtration (see [3] in Figure 1). The picophyto-plankton cell concentrations were calculated as the differ-ence in the number of cells measured in filtrate [2] andfiltrate [3] (see Figure 1). Two filters (Nuclepore and GF/FWhatman) for carboxylase assays and two filters forchlorophyll (a, b and c) analysis were stored in the dark inliquid nitrogen. Samples for flow cytometric analysis werefixed with formaldehyde (2% final concentration) andstored in liquid nitrogen (Troussellier et al., 1995).

Chlorophyll concentrations

The concentrations of chlorophyll a, b and c were deter-mined using the spectrofluorimetric method described inNeveux and Lantoine (Neveux and Lantoine, 1993).

Flow cytometry analysis

Flow cytometry analyses were run with a FacsCalibur flowcytometer (Becton Dickinson, San Jose, CA) equippedwith an air-cooled argon laser (15 mW). Phytoplanktoncells were detected according to their right-angle lightscatter (related to cell size and structure) and to theirorange and red fluorescence emissions (due to phycoery-thrin and chlorophyll pigments), collected through585/42 nm band pass and 650 nm long pass optical filtersrespectively. One micrometre beads (Polysciences, War-rington, PA), used as internal standard, were systemati-cally added to samples in order to normalize cellfluorescence emissions and to compare the differentphytoplankton groups. Each group was characterized bytheir red fluorescence relative to the 1 µm beads(expressed in bead fluorescence units, bfu) and their right-angle scatter relative to the 1 µm beads (expressed in bead

E. FOUILLAND ET AL. SIZE-FRACTIONATED CARBOXYLASE ACTIVITY

Fig. 1. Size-fractionation protocol.

fouilland (2188) 7/6/01 12:27 pm Page 625

scatter units, bsu). Flow cytometry allowed us to differen-tiate phytoplankton with sizes ranging between 0.5 µmand 50 µm (flow cytometer injector size).

Carboxylase activity measurements

The in vitro carboxylase activity measurements allowed us toquantify the in vitro Rubisco activity and �-carboxylaseactivities (phosphenolpyruvate carboxylase [PEPC], phos-phoenolpyruvate carboxykinase [PEPCK], pyruvate car-boxylase [PYRC]). The carboxylase activities weredetermined by measurements of the incorporation ofradioactive bicarbonate into stable products. All carboxy-lase assays were made on the same extract. Extraction wascarried out at 0ºC and pH 8, following the protocoldescribed in Descolas-Gros and De Billy (Descolas-Grosand De Billy, 1987). For the Rubisco assay, the reactionmixture contained 50 mM Tricine buffer (pH 8), 5 mM

dithiothreitol (DTT), 10 mM MgCl2, 20 mM NaH14CO3and 100 µl of the crude extract. For the other three assays,the reaction mixture contained 50 mM PIPES buffer (pH6.8) instead of Tricine buffer. Moreover, 5 mM MnCl2 and5 mM ADP were added for the PEPCK assays, and 5 mM

MgCl2 and 5 mM ATP were added for the PYRC assays. Allreactions were initiated after a pre-incubation period of 10min by adding substrates: 2 mM RuBP for the Rubisco assay,5 mM PEP for the PEPC and PEPCK assays and 5 mM

pyruvate for the PYRC assay. All mixtures were incubatedat 25ºC for 20 min. The reactions were stopped by theaddition of 6 M HCl. More details on the protocol are givenby Descolas-Gros and colleagues (Descolas-Gros andFontugne, 1985; Descolas-Gros and De Billy, 1987; Desco-las-Gros and Oriol, 1992). Every carboxylase assay wascarried out in triplicate and was automated with a Gilsonauto-analyser (222 Autosampler Injector interfaced with aDilutor 401). Carboxylase activity is expressed in nmol CO2l–1 h–1, with an assay error <15%, or normalized perchlorophyll a unit (nmol CO2 µg Chl a–1 h–1). The �-carboxylase activity (sum of PEPC, PEPCK and PYRC),expressed as a percentage of Rubisco activity (�C/R in %)is used to reflect the relative contribution of the differentassimilation pathways (Descolas-Gros and Oriol, 1992). Alow ratio (<40%) indicates a significant contribution ofphotoautotrophy relative to the �-carboxylation pathwayand is characteristic of strictly autotrophic cells in goodphysiological state (Descolas-Gros and Fontugne, 1990;Descolas-Gros and Oriol, 1992).

Values for the picophytoplanktonic contribution toRubisco activity, chlorophyll concentration and cellnumbers were obtained from the size-fractionated filtra-tions as described above. Enumeration by flow cytometryconcerned phytoplankton with cell size less than 50 µm,thus the Rubisco activity and chlorophyll are expressed asper cell for the <1 µm phytoplanktonic fraction only.

R E S U LT S

Comparison of results from underwaterpumps and size-fractionation filtrations

The variations in carboxylase activity per litre (Rubiscoand �-carboxylase activities) from underwater pumpingand from the sum of size-fractionated values (>1 µm + <1µm carboxylase activities) were very similar (Figure 2). Theaverages of Rubisco and �-carboxylase activities were 15.0± 7.2 and 3.5 ± 2.6 nmol CO2 l–1 h–1 respectively, forsamples from the underwater pumps (n = 9), and 15.3 ±8.9 and 3.0 ± 2.1 nmol CO2 l–1 h–1, respectively, for thesum of the size fractions (n = 8). Therefore, we assume thatthe values of carboxylase activity from size-fractionationfiltrations are representative of the in situ photosyntheticactivity and that the shipboard filtrations did not changethe measures of the in vitro carboxylase activities. Similarobservations have been made from in situ values sampled inthe Indian sector of the Southern Ocean (Fouilland et al.,2000). The largest differences between the values from thetwo sampling methods, observed especially at night, couldbe due to some disparity in sampling times.

Flow cytometry detection: comparison ofthe different <1 µm cell groups

According to their light scatter and fluorescence proper-ties, different cell types were identified and numerically

JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒

Fig. 2. Rubisco and �-carboxylase activities per litre from samplesfrom underwater pumps (UP) and from size-fractionation filtrations (S-F: sum of >1 µm + <1 µm fraction) during the 32 h cycle from May 14to May 16 at 30 m depth.

fouilland (2188) 7/6/01 12:27 pm Page 626

followed as cyanobacteria cells, small eukaryotes and largeeukaryotes.

Unidentified cells, presumably small eukaryotes, werefound to pass through the 1 µm filter (Figure 3). Somelarge eukaryote cells, with large apparent size and high redfluorescence, were present in the filtrate after 1 µm filtra-tion, with very low cell numbers (<30 cells ml–1), consti-tuting <0.2% of total cell numbers after 1 µm filtration(Figure 3). In the <1 µm filtrate, a population ofcyanobacteria was identified according to its typicalorange and red fluorescence (Figure 3). The relative con-tribution of cyanobacteria cells to the total number of<1 µm cells ranged from 66% to 90% during the cycle(Table I).

The red fluorescence for each cell group in the <1 µmfraction was summed and compared with the <1 µmchlorophyll a+b+c concentrations. A positive linearregression was observed between these two measures (R2

= 0.86, n = 7). The total red fluorescence was thus directlycorrelated with the chlorophyll pigment concentrationsmeasured in the small phytoplankton fraction and therelationship was used to define the <1 µm apparentchlorophyll. During the 32 h cycle, the contribution of theeukaryote cells to the <1 µm apparent chlorophyll was fol-lowed (Table I). The contribution of eukaryote cells to the<1 µm apparent chlorophyll varied from 28% to 77% andwas highest during the light period (≥75%).

Size-fractionated results expressed per litre

The chlorophyll concentrations, as well as the carboxylaseactivities, from both size classes and <1 µm cell numbers

are reported in Table I. Small phytoplankton (<1 µm) con-tributed about 23% of the total phytoplankton chlorophylland about 19% of the total Rubisco activity. The �C/Rratios averaged 18 ± 6% for the >1 µm fraction and 33 ±16% for the <1 µm fraction, and were generally less than40%, indicating that the in vitro Rubisco activity (photoautotrophic CO2 fixation) was dominant. The >1µm Rubisco activity and �-carboxylase activity seemed to

E. FOUILLAND ET AL. SIZE-FRACTIONATED CARBOXYLASE ACTIVITY

Table I: Temporal variations of size-fractionated chlorophyll (a+b+c), Rubisco and b-carboxylase

activities, <1 µm total cell numbers and cyanobacteria cell numbers during the 32 h cycle from May

14 to May 16 at 30 m depth. Contribution of eukaryote cells to the < 1 µm apparent chlorophyll

Chlorophyll Rubisco activity �-carboxylase activity

<1 µm cell numbers Contribution of eukaryote

Sampling date

(µg l–1) (nmol CO2 l–1 h–1) (nmol CO2 l–1 h–1)

(cells ml–1) cells to the total <1 µm

and local time >1 µm <1 µm >1 µm <1 µm >1 µm <1 µm Total cells cyanobacteria cells (%) apparent chlorophyll (%)

14/05/95 17:30 h 1.44 ND 16.2 2.4 4.3 1.0 16 672 13 411 (80) 53

14/05/95 21:00 h 1.93 0.43 12.0 2.5 2.2 1.6 22 762 20 490 (90) 28

15/05/95 01:00 h ND ND 10.5 3.5 1.3 0.5 19 654 17 184 (87) 33

15/05/95 05:30 h 1.00 0.45 9.1 2.2 2.2 0.6 15 522 12 019 (77) 60

15/05/95 09:30 h 2.11 0.88 29.1 5.0 6.6 1.8 20 678 14 498 (70) 77

15/05/95 13:00 h 1.38 0.53 15.1 2.7 3.3 0.7 15 091 10 034 (66) 75

15/05/95 17:30 h 0.61 0.09 ND 1.1 ND 0.4 4770 3602 (76) 65

15/05/95 21:00 h 0.55 0.18 4.7 1.9 0.3 0.2 8246 6521 (79) 53

16/05/95 01:00 h 0.64 0.12 4.6 1.2 0.6 0.5 5875 4491 (76) 63

ND, no data; data in bold denote a period of darkness.

Fig. 3. Flow cytometric characterization of the <1 µm picophyto-planktonic assemblage according to cell properties (red fluorescenceequivalent to the apparent chlorophyll and light scatter equivalent toapparent cell size, normalized by 1 µm beads) during the 32 h cycle fromMay 14 to May 16 at 30 m depth.

fouilland (2188) 7/6/01 12:27 pm Page 627

follow a day/night cycle, with higher values measuredduring the light period, at 17:30 h on May 14 and at 09:30 h and 13:00 h on May 15 (Table I). The highestvalues of sized-fractionated chlorophyll concentrations,Rubisco and �-carboxylase activities were generallyobserved on May 15 at 09:30 h and the lowest values weremeasured on May 15 at 17:30 h or 21:00 h (Table I).

Size-fractionated carboxylase activitiesexpressed per chlorophyll a unit

The Rubisco activity averaged 12.8 ± 3.8 nmol CO2 (µgChl a)–1 h–1 and 10.5 ± 3.8 nmol CO2 (µg Chl a)–1 h–1 andthe �-carboxylase activity averaged 2.5 ± 1.3 nmol CO2(µg Chl a)–1 h–1 and 3.4 ± 1.7 nmol CO2 (µg Chl a)–1 h–1

for the >1 µm and the <1 µm fractions, respectively.During the 32 h cycle, a strong day/night variation ofRubisco and �-carboxylase activity per Chl a wasobserved for the >1 µm phytoplankton, with the highestvalues being measured during the day and the lowest atnight (Figure 4). A similar pattern was observed for the >1µm Rubisco activity and �-carboxylase activities per litre(Table I). For the <1 µm fraction, an increase in Rubiscoand �-carboxylase activities per Chl a unit occurredduring the latter hours of May 15. This was especiallynotable for Rubisco activity per Chl a, which increasedtwofold (Figure 4) and may be explained by the strongerdecrease of <1 µm chlorophyll per litre relative to thedecrease of <1 µm Rubisco activity per litre on May 15from 17:30 h (Table I).

Rubisco activity and chlorophyll expressedper cell for <1 µm cells

During the 32 h cycle, the chlorophyll and Rubiscoactivity expressed per cell for <1 µm phytoplankton variedfrom 19 to 43 fg Chl cell–1 and from 0.11 to 0.24 fmol CO2cell–1 h–1, respectively. The highest values for these twovariables were observed during the light period (Figure 5).During the second dark period, the Rubisco activity percell remained high while the chlorophyll per celldecreased (Figure 5).

D I S C U S S I O N

Influence of species succession on thevariation of >1 µm carboxylase activitiesduring the 32 h cycle

The particulate organic matter at 30 m depth during this32 h cycle mainly originated from photosynthetic pro-cesses, as suggested by the good relation obtained betweenparticulate organic carbon (POC) and Chl a concentra-tions and by the POC/Chl a ratio being lower than 200(Bentaleb et al., 1999). From the present results, larger phy-toplanktonic cells (>1 µm) showed a close diel variation ofcarboxylase activity per Chl a during the 32 h cycle, withthe highest values measured during the day and the lowestmeasured during the night (Figure 4). The co-variation of�-carboxylase and Rubisco activities per Chl a during thiscycle showed that these two C assimilation pathways werestrongly interdependent (Falkowski and Raven, 1997).

JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒

Fig. 4. Temporal variations of Rubisco and �-carboxylase activitiesexpressed per chlorophyll a (R/Chl a and �C/Chl a, respectively) duringthe 32 h cycle from May 14 to May 16 at 30 m depth.

Fig. 5. Variations of small picophytoplankton (<1 µm) chlorophyllconcentration and Rubisco activity (expressed per cell ) during the 32 hcycle from May 14 to May 16 at 30 m depth.

fouilland (2188) 7/6/01 12:27 pm Page 628

Nevertheless, these diel variations could reflect a changein the >1 µm species composition of the phytoplankton at30 m depth during the cycle. From high-performanceliquid chromatography pigment detection (data providedby F. Vidussi), an equivalent Chl a for chromophytenanoflagellates (identified by 19�-hexanoyloxyfucoxan-thin + 19�-butanoyloxyfucoxanthin pigments) anddiatoms (identified by fucoxanthin pigment) was estimatedthrough an empirical regression equation. Coefficients forthe relationship at 30 m depth were determined by VanWambeke and Vidussi (Vidussi et al., 2000). Higher con-tributions of chromophyte nanoflagellates to the total esti-mated chlorophyll a (TChl a) relative to the diatomcontribution were observed in association with the highestRubisco and carboxylases activities (Figure 6). The chro-mophyte nanoflagellates, mainly prymnesiophytes(Vidussi et al., 2000), dominated on May 14 at 17:30 h andon May 15 at 09:30 h, contributing >30% of the TChl a(Figure 6). The peak of Rubisco and �-carboxylase activi-ties per Chl a observed during the light period of the cycle(Figure 4), may be due both to the higher contribution ofnanophytoplankton to the TChl a and to the light regu-lation of Rubisco activity. This latter regulation couldexplain the measurement of the highest Rubisco activityper Chl a on May 15 at 09:30 h, whereas the highest con-tribution from nanoflagellates was observed on May 14 at17:30 h. Moreover, the highest �C/R ratios weremeasured (≥20%) during the period of nanoflagellatedominance (Figure 6). In the prymnesiophyte group, �-carboxylase activity relative to Rubisco activity appearedto be greater than for temperate diatoms (Descolas-Grosand Oriol, 1992). This observation could confirm thehypothesis of such a prymnesiophyte dominance in the>1 µm production at 30 m depth during the 32 h cycle.Thus, these phytoplankton species could possess a CO2fixation metabolism with a higher �C/R ratio and ahigher in vitro carboxylase (Rubisco and �-carboxylase)activities per Chl a than Mediterranean diatoms. As thesampling depth was below the chlorophyll maximum,such high potential C metabolism appeared to be welladapted to the low irradiance conditions, which concurswith the classical view that small phytoplankton cells areusually more efficient at using low light intensities forphotosynthesis due to a smaller package effect in photonabsorption [reviewed in (Raven, 1998)].

Dependence of cellular chlorophyll andRubisco activity variations on thecyanobacteria and eukaryote cellproportions

The Rubisco and �-carboxylase activity per Chl a for the<1 µm fraction did not show a diel variability during the32 h cycle. However, a strong increase occurred at the end

of the cycle, which was explained by a strong decrease of<1 µm biomass (Table I). Likewise, no diel variation ofthe Rubisco activity per cell was observed during thecycle (Figure 5). No relationship between the cellularRubisco activity and the environmental conditionsinduced by the episodic wind event was clearly shown.On May 13, a high wind event has been reported in thestudy area (Andersen and Prieur, 2000). The eventinduced an upward shift of the nitracline, accompaniedby an enrichment of nitrate and phosphate, and adecrease of water temperature at 30 m depth (Andersenand Prieur, 2000). The water temperature at 30 m depthdecreased from May 11 to May 15 and then became rela-tively constant. Conversely, nitrate and phosphate con-centrations increased from 2 µM to 6 µM and from 0.05µM to 0.25 µM, respectively. Ammonium concentrationremained constant (0.01 µM) over the sampling period.The maximal values of the surface light irradiance calcu-lated by the Klein model (Chifflet, 1996) varied from 500to 900 µmol m–2 s–1. Based on nitrogen (nitrate + nitrite)values and phosphorus concentrations, the calculatedN/P ratio was about 50 before May 15 and decreased toabout 25 after this date at 30 m depth. Despite the lowambient ammonium concentration and unfavourableN/P ratio, evidence of a negative influence of theenvironmental conditions on diel cellular Rubisco

E. FOUILLAND ET AL. SIZE-FRACTIONATED CARBOXYLASE ACTIVITY

Fig. 6. Comparison between the contribution (%) of chromophytenanoflagellate pigments (19�HF + 19�BF) and diatom pigments (fucox-anthin) to the total pigments (TChl a) for the whole phytoplanktonassemblage as measured by HPLC analysis (Vidussi et al., 2000) and the�-carboxylase activity expressed as a percentage of Rubisco activity(�C/R) for the >1 µm fraction during the 32 h cycle from May 14 toMay 16 at 30 m depth.

fouilland (2188) 7/6/01 12:27 pm Page 629

activity cannot be demonstrated. Nevertheless, cellularRubisco activity variation and chlorophyll content werepartially explained by a species change during the 32 hcycle. The averages of <1 µm chlorophyll per cell andRubisco activity per cell varied with the contribution ofthe eukaryote cells to the <1 µm apparent chlorophyll asshown in Figures 7(a) and 7(b), respectively. During the 32h cycle at 30 m depth, the lowest values of chlorophyll percell (19 fg Chl cell–1) and Rubisco activity per cell (0.11and 0.18 fmol CO2 cell–1 h–1) were measured for a 30%contribution of the eukaryote cells to the <1 µm appar-ent chlorophyll, which is equivalent to a 70% contri-bution for cyanobacteria cells. The highest values ofchlorophyll per cell (35 and 43 fg Chl cell–1) and Rubisco

activity per cell (0.18 and 0.24 fmol CO2 cell–1 h–1) wereobserved for a 75% contribution of the eukaryote cells tothe <1 µm apparent chlorophyll. The relative differencein cellular chlorophyll content and cellular Rubiscoactivity between the two cell groups (for 70% contri-bution of cyanobacteria cells and for 75% of eukaryotescells) was about twofold. The lowest and highest means ofapparent size of <1 µm phytoplankton measured by flowcytometry were 0.07–0.08 bsu and 0.17–0.19 bsu. Thesewere reported when cyanobacteria cells and eukaryotecells dominated the apparent chlorophyll, respectively.The apparent cell size provided by flow cytometry analy-ses is determined by the cell size, volume and shape [seereview (Collier, 2000)]. The differences in cellular contentbetween the two types of cells were thus explained by thetwofold difference in apparent cell size for small eukary-otes compared to cyanobacteria, consistent with the ratioobtained for chlorophyll content. Only a few referenceshave made comparisons of these two small cell groups interms of CO2 fixation per cell. Glover reported values ofspecific cyanobacterial Rubisco activity between 1 and 3fg C cell–1 h–1 for the 0.6–1 µm fraction, with Synechococ-

cus spp. accounting for more than 93% of autofluoresc-ing cells (Glover, 1989). These values are close to thoseestimated from our study. This author used a similarcombination of methods (i.e. an enzymatic method anda combination of epifluorescence microscopy and size-fractionation techniques) for the study of samples fromthe Atlantic Ocean. CO2 uptake was estimated forautotrophic prokaryote and eukaryote picoplanktonusing a compilation of biovolume data, which was con-verted into carbon per cell combined with specific growthrate data from the literature (Li, 1986). Li showed thatpicoeukaryote CO2 uptake was 2.3 times greater thanthat of cyanobacteria, which is consistent with the obser-vations in this study. The same difference between thetwo cell groups was noticed in Lake Kinneret (Israel),with values of maximum cell-specific photosyntheticrates of C fixation by <3 µm picoeukaryotes being 2.5times more than values for pico-cyanobacteria cells (Malinsky-Rushansky et al., 1997). Nevertheless, similarvalues of CO2 fixation per cell were shown for these twocell groups (<1 µm cell diameter) using flow cytometricsorting of 14C-labelled cells in the surface waters of thenorth Atlantic Ocean (Li, 1994). Such an agreement incell-specific CO2 fixation is probably explained by thevery similar apparent cell size of these two cell groups, asmeasured by flow cytometry (Li, 1994). The differencebetween cyanobacteria and small picoeukaryote cellsreported in the present study seemed to be related to cellsize. Dorsey et al. reported 14C measurements per cell,calculated from total sample uptake and divided by thenumber of cells in samples from different sizes of

JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒

Fig. 7. Variations of the <1 µm chlorophyll per cell (a) and the <1 µmRubisco activity per cell (b) as a function of the contribution of eukary-ote cells to the <1 µm apparent chlorophyll fraction during the 32 h cyclefrom May 14 to May 16 at 30 m depth.

fouilland (2188) 7/6/01 12:27 pm Page 630

phytoplankton cultures (Dorsey et al., 1989). Theyobtained a positive linear regression between cell size and14C uptake per cell (1–10 µm size cell). In contrast, forlarger cells (102–107 µm3 cells), as carbon per cell is pro-portional to volume of cell, the light-saturated photosyn-thesis rate is negatively correlated to cell size (Taguchi,1976). This suggests an optimum CO2 uptake per cellrelated to its size. A size-dependence has also been shownfor diatom growth rates [reviewed in (Geider et al.,1986)], suggesting an evident relationship betweenphotosynthesis providing the necessary energy for cellgrowth and the growth rate itself. Due to their tiny size,the growth rates of picophytoplankton were poorlystudied, especially in the field. Nevertheless, themaximum growth rate reported in cultures by Throndsen(Throndsen, 1976) for Micromonas pussilla, a picoflagellatecommonly observed in the Mediterranean Sea, washigher than values for cyanobacteria cells such as Syne-

chococcus spp. in cultures [reviewed by (Furnas, 1990)].Moreover, using an experimental protocol with dialysisbags to study the effect of protozoan presence, Ferrier-Pagés and Rassoulzadegan observed generally highergrowth rates for autotrophic picoflagellates than forcyanobacteria sampled from surface sea water outside theBay of Villefranche-sur-Mer, close to the permanentDYFAMED site (Ferrier-Pagés and Rassoulzadegan,1994). Such a difference in growth rates could beexplained by higher cellular photoautotrophic capabilityfor picoeukaryotes than for cyanobacteria cells.

Thus, in vitro carboxylase activities for both large andsmall size phytoplankton assemblages during the sam-pling period were more dependent on a change in thedominant cell species or group rather than on theday/night cycle. The highest photoautotrophic capabilitywas observed when nanophytoplankton and picoeukary-otes dominated in the >1 µm and <1 µm fractions,respectively. This suggests that these small cells are poten-tially more efficient at using light as an energy source thanthe larger phytoplankton, explaining their high contri-bution to primary production in unfavourable environ-ments (low light and/or nutrient). However, it is difficultto show evidence of a clear pattern in cellular photoau-totrophic capability using data such as cell abundance,which depend not only on the growth rates but also onphysical phenomena, i.e. advection (Jacquet et al., 1998)and grazing rates. The combination of enzymatic assayand cell enumeration by flow cytometry on size-fraction-ation filtrations was the first step to estimate and followcellular CO2 fixation in the field. However, to gain furtherinsight into the diel cellular cycle of phytoplankton, newmethods, based on single-cell in vivo Rubisco activityadapted for a phytoplanktonic assemblage in the field,must be developed.

AC K N OW L E D G E M E N T S

This work was supported by the Centre National de laRecherche Scientifique-Institut National des Sciences del’Univers (CNRS-INSU) and was part of a more generalinvestigation, DYNAPROC, headed by Valérie Andersen.We are grateful to the officers and crew of the Téthys IIfor their assistance at sea and to our colleagues Ilhem Ben-thaleb, Michel Denis, Valérie Martin and Danielle Martyfor their help during the sampling period. We would alsolike to thank Virginie Pons for providing chlorophyll dataand for helping with the carboxylase assays, andFrancesca Vidussi and Jean-Claude Marty for helpingwith HPLC pigment concentrations. This is ContributionISEM 2001–027.

R E F E R E N C E SAndersen, V., and Prieur, A. (2000) One-month study in the open NW

Mediterranean Sea (Dynaproc experiment, May 1995): overview ofhydrobiogeochemical structures and effects of wind events. Deep-Sea

Res., 47, 397–422.

Appleby, G., Colbeck, J., Holdsworth, E. S. and Wadman, H. (1980) �-carboxylation enzymes in marine phytoplankton and isolation andpurification of pyruvate carboxylase from Amphidinium carterae (Dino-phyceae). J. Phycol., 16, 290–295.

Bentaleb, I., Grimalt, J. O., Vidussi, F., Marty, J.-C., Martin, V., Denis,M., Hatté, C. and Fontugne, M. (1999) The C37 alkenone record ofseawater temperature during seasonal thermocline stratification. Mar.

Chem., 64, 301–313.

Buat-Ménard, P. and Lambert, C. E. (1993) Overview of theDYFAMED program. Ann. Inst. Océanogr. Paris, 69, 101–105.

Chavez, F. P. (1989) Size distribution of phytoplankton in the central andeastern tropical Pacific. Global Biochem. Cycles, 3, 27–35.

Chifflet, M. (1996) Modelisation physique-biologie unidimensionnellede l’écosystème pélagique marin. Application à l’étude des processusrapides. DEA Sciences de l’Environnement Marin. Université de laMéditerranée Aix-Marseille II, France, 56 pp.

Collier, J. L. (2000) Flow cytometry and the single cell in phycology. J.

Phycol., 36, 628–644.

Descolas-Gros, C. and Fontugne, M. (1985) Carbon fixation in marinephytoplankton: carboxylase activities and stable carbon-isotope ratios;physiological and paleoclimatological aspects. Mar. Biol., 87, 1–6.

Descolas-Gros, C. and De Billy, G. (1987) Temperature adaptation ofRuBP caboxylase: kinetic properties in marine Antarctic diatoms. J.

Exp. Mar. Biol. Ecol., 108, 147–158.

Descolas-Gros, C. and Fontugne, M. (1988) Carboxylase activities andcarbon isotope ratios of Mediterranean phytoplankton. Oceanol. Acta,Special Issue, 245–250.

Descolas-Gros, C. and Fontugne, M. (1990) Stable carbon isotope frac-tionation by marine phytoplankton during photosynthesis. Plant Cell

Environ., 13, 207–218.

Descolas-Gros, C. and Oriol, L. (1992) Variations in carboxylase activityin marine phytoplankton cultures. �-carboxylation in carbon fluxstudies. Mar. Ecol. Progr. Ser., 85, 163–169.

Dorsey, J., Yentsch, C. M., Mayo, S. and McKenna, C. (1989) Rapid

E. FOUILLAND ET AL. SIZE-FRACTIONATED CARBOXYLASE ACTIVITY

fouilland (2188) 7/6/01 12:27 pm Page 631

analytical technique for assessment of cell metabolic activity inmarine microalgae. Cytometry, 10, 622–628.

Falkowski, P. G. and Raven, J. A. (1997) Aquatic Photosynthesis. BlackwellScience, Oxford.

Ferrier-Pagés, C. and Rassoulzadegan, F. (1994) Seasonal impact of themicrozooplankton on pico- and nanoplankton growth rates in thenorthwest Mediterranean Sea. Mar. Ecol. Prog. Ser., 108, 283–294.

Fouilland, E., Courties, C. and Descolas-Gros, C. (2000) Variations incarboxylase activity for size-fractionated phytoplankton of IndianSouthern Ocean. J. Plankton Res., 22, 1185–1201.

Furnas, M. J. (1990) In situ growth rates of marine phytoplankton:approaches to measurement, community and species growth rates. J.

Plankton Res., 12, 1117–1151.

Geider, R. J., Platt, T. and Raven, J. (1986) Size dependence of growthrate and photosynthesis in diatoms: a synthesis. Mar. Ecol. Prog. Ser., 30,93–104.

Glover, H. E. (1989) Ribulosebisphosphate carboxylase/oxygenase inmarine organisms. Int. Rev. Cytol., 115, 67–137.

Glover, H. E. and Morris, I. (1979) Photosynthetic carboxylatingenzymes in marine phytoplankton. Limnol. Oceanogr., 24, 510–519.

Hellebust, J. A. and Terborgh, J. (1967) Effects of environmental con-ditions on the rate of photosynthesis and some photosyntheticenzymes in Dunaliella tertiolecta. Limnol. Oceanogr., 12, 559–567.

Jacquet, S., Lennon, J. F., Marie, D. and Vaulot, D. (1998) Picophyto-plankton population dynamics in coastal waters of the northwesternMediterranean Sea. Limnol. Oceanogr., 43, 1916–1931.

Krupatkina, D. K. (1990) Estimates of primary production in oligo-trophic waters and metabolism of picoplankton: a review. Mar. Microb.

Food Webs, 4, 87–102.

Leftley, J. W., Bonin, D. J. and Maestrini, S. Y. (1983) Problems in esti-mating marine phytoplankton growth, productivity and metabolicactivity in nature: an overview of methodology. Oceanogr. Mar. Biol.

Annu. Rev., 21, 23–66.

Levedahl, B. H. (1978) Heterotrophic CO2 fixation in Euglena. InBuetow, D. E. (ed.), The Biology of Euglena. Academic Press, New York,pp. 85–96.

Li, W. K. W. (1986) Experimental approaches to field measurementsmethods and interpretation. Can. Bull. Fish. Aquat. Sci., 214, 251–286.

Li, W. K. W. (1994) Primary production of prochlorophytes, cyanobac-teria, and eucaryotic ultraphytoplankton: measurements from flowcytometry sorting. Limnol. Oceanogr., 39, 169–175.

Magazzù, G. and Decembrini, F. (1995) Primary production, biomassand abundance of phototrophic picoplankton in the MediterraneanSea: a review. Aquat. Microb. Ecol., 9, 97–104.

Malinsky-Rushansky, N. Z., Berman, T. and Dubinsky, Z. (1997) Sea-sonal photosynthetic activity of autotrophic picoplankton in LakeKinneret, Israel. J Plankton Res., 19, 979–993.

Malone, T. C. (1980a) Size-fractionated primary production of marinephytoplankton. In Falkowski, P. (ed.), Primary Productivity in the Sea.Plenum Press, New York, pp. 301–320.

Malone, T. C. (1980b) Algal size. In Morris, I. (ed.), Physiological Ecology

of Phytoplankton. University of California Press, Berkeley, pp. 433–464.

Morris, I. and Farrell, K. (1971) Photosynthetic rates, gross patterns ofcarbon dioxide assimilation and activities of ribulose diphosphate carboxylase in marine algae at different temperatures. Physiol. Plant,25, 372–377.

Mortain-Bertrand, A., Descolas-Gros, C. and Jupin, H. (1988) Pathwayof dark inorganic carbon fixation in two species of diatoms: influenceof light regime and regulator factors on diel variations. J. Plankton Res.,10, 199–217.

Neveux, J. and Lantoine, F. (1993) Spectrofluorometric assay of chloro-phylls and phaeopigments using the least squares approximation tech-nique. Deep-Sea Res., 40, 1747–1765.

Olson, R.J., Vaulot, D. and Chisholm, S. W. (1985) Marine phyto-plankton distributions measured using shipboard flow cytometry.Deep-Sea Res., 32, 1273–1280.

Olson, R. J., Zettler, E. R., Chisholm S. W. and Dusenberry, J. A. (1991)Advances in oceonography through flow cytometry. In Demers, S.(ed.), Particle Analysis in Oceanography. NATO ASI Series, G27,pp. 351–397.

Paasche, E. (1971) Effect of ammonia and nitrate on growth, photosyn-thesis, and ribulose diphosphate carboxylase content of Dunaliella

tertiolecta. Physiol. Plant, 25, 294–299.

Raven, J. A. (1974) Carbon dioxide fixation. In Stewart, W.O.P. (ed.),Algal Physiology and Biochemistry. University of California Press, Berke-ley, pp. 434–455.

Raven, J. A. (1998) The twelfth Tansley Lecture – Small is beautiful: thepicophytoplankton. Funct. Ecol., 12, 503–513.

Sieburth, J. McN., Smetacek, V. and Lenz, J. (1978) Pelagic ecosystemstructure: heterotrophic compartments of the plankton and theirrelationship to plankton size fractions. Limnol. Oeanogr., 23,1256–1263.

Steemann-Nielsen, E. (1952) The use of radioactive carbon (14C) formeasuring organic production in the sea. J. Cons. Int. Explor. Mer, 18,117–140.

Taguchi, S. (1976) Relationship between photosynthesis and cell size ofmarine diatoms. J Phycol., 12, 185–189.

Throndsen, J. (1976) Occurrence and productivity of small marine flagellates. Norw. J. Bot., 23, 269–293.

Troussellier, M., Courties, C. and Zettelmaier, S. (1995) Flow cytomet-ric analysis of coastal lagoon bacterioplankton and picophytoplank-ton: fixation and storage effects. Estuar. Coast. Shelf Sci., 40, 621–633.

Vidussi, F., Marty, J.-C. and Chiavérini, J. (2000) Phytoplankton pigmentvariations during the transition from spring bloom to oligotrophy inthe northwestern Mediterranean Sea. Deep-Sea Res., 47, 423–445.

Wood, A., Horan, P. K., Muirhead, K., Phinney, C. M., Yentsch, C. M.and Waterbury, J. B. (1985) Discrimination between pigment types ofmarine Synechococcus spp. by scanning spectroscopy, epifluorescencemicroscopy, and flow cytometry. Limnol. Oceanogr., 30, 1303–1315

Yentsch, C. M., Horan, P. K., Muirhead, K., Dortch, Q., Perry, M. J.,Phynney, D. A., Pomponi, S. A., Spinar, R. W., Wood, M., Yentsch, C.S. and Zahuranec, B. J. (1983) Flow cytometry and cell sorting ofaquatic particles. Limnol. Oceanogr., 28, 1275–1280.

Received on August 15, 2000; accepted on February 22, 2001

JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒

fouilland (2188) 7/6/01 12:27 pm Page 632