All experiments were performed by using Chlorella sorokiniana Shihira & Krauss, strain 211/8K (CCAP of Cambridge University). The algae were grown in batches placed in a thermostatic chamber at 35˚C, continuously stirred and illuminated by fluorescent lamps (150 µmol photons m⁻²∙s⁻¹). The cultures were insufflated with air containing 5% CO₂ at a flow rate of 80 - 100 l∙h⁻¹. Three different types of medium were used to grow cells: basal, nitrate-free and sulphate-free medium. The composition of the basal medium was previously reported [10] . S- and N-starved cells were obtained collecting batch grown cells by a low speed centrifugation (4000 r for 5 min), washed two times and re-suspended in a N-free or S-free medium as previously described [3] [11] . Algal growth was evaluated by absorbance (OD₅₅₀). Culture purity was monitored by optical microscopy.

The packed cell volume (PCV) was estimated by centrifuging 10 ml of cell suspension in a haematocrit tube at 4000 r for 5 min. At the end of centrifugation the PCV value can be directly read from the calibrated tube.

Intracellular glutamine concentration was tested in sufficient and in S- or N-starved cells. Cell suspensions (10 ml) were collected by centrifugation (4000 r for 5 min), the packed cells mixed with a solution containing 1 ml cold 80% ethanol and 1 mM γ-aminobutyric acid, left for 15 min and then centrifuged (4000 r for 5 min). The supernatant was filtered through Waters Sep-Pak® Cartridges to remove chlorophyll. The glutamine was determined by HPLC as previously described [12] . For cysteine content determination the packed cells were treated with 2 ml of extraction buffer as previously described [12] . Cysteine content was quantified by reversed-phase HPLC after derivatization with monobromobimane according to Carfagna et al. [12] .

To prepare crude extract for GS activity determination, aliquots of 100 ml of algal culture were harvested by centrifugation at 4000 r for 5 min; the pellets were re-suspended in 5 ml of extraction buffer ( 10 mM Tris-HCl pH 7.2, 2.5 mM MgCl₂, 1 mM dithiothreitol, 5 mM ethylenediaminetetraacetic acid, 15% glycerol) and the cells were lysed by a passage at 11,000 psi through French pressure cell (Aminco). To 1 ml of mix reaction ( 40 mM imidazole-HCl pH 7.0, 30 mM glutamine, 20 mM K-arsenate pH 7.0, 120 mM NH₂OH pH 7.0, 0.4 mM adenosine diphosphate, 3 mM MnCl₂) 1 ml of crude extract was added. The whole mixture was incubated at 30˚C and after 10 min, 2 ml of stop mixture (4% FeCl₃, 2.4% trichloroacetic acid, 0.6 M HCl) was included. Subsequent to a low speed centrifugation the absorbance of the mix was measured at 540 nm. One enzyme U is defined as 1 µmol of γ-glutamylhydroxamate min⁻¹∙mg⁻¹ protein. To determine OASTL activity, aliquots of 100 ml of cultures were harvested by centrifugation (4000 r for 10 min) and the pellets were re-suspended in the extraction buffer ( 50 mM phosphate-buffer pH 7.5, 10 μM pyridoxal phosphate and 1 mM dithiothreitol). The cells were broken by passing twice through a French pressure cell (11,000 psi). The lysates were clarified by centrifugation at 15,000 rpm for 15 min at 4˚C. The supernatant represented the crude extract. The enzymatic activity of OASTL was measured according to Gaitonde method [13] modified as described by Carfagna et al. [14] . One enzyme U is defined as 1 µmol cysteine min⁻¹∙mg⁻¹ protein. Protein amounts were determined using the Bio- Rad protein assay based on the Bradford method [15] with bovine serum albumin as the standard.

Cells were collected by a low speed centrifugation (4000 r for 10 min). The packed cells were extracted twice with 80% ethanol at 80˚C for 15 min. After cooling, the pellets were re-suspended twice in distilled water and then centrifuged. The pellets were washed twice in 50 mM acetic acid-NaOH buffer (pH 4.8) by centrifugation and then autoclaved for 30 min at 120˚C. The extracts were cooled and total starch was determined as glucose derived from its hydrolysis as previously described [12] .

Experimental data analyses were made using Sigmaplot 12 software. All data are expressed as the means ± SE for 5 to 9 determinations. The statistical analysis was performed by one-way ANOVA with a Tukey post-hoc test to determine differences between sufficient and S or N starved algae, P < 0.05 and P < 0.001 as significant. If necessary, the data were log + 1 (x) transformed before the analysis. Data expressed in percentages were trans-

formed by arcsin transformation (y′ = arcsin,) and then analyzed.

In order to define and compare the effects of N- or S-starvation on algae growth, we performed OD₅₅₀ determination. When cells were in the exponential phase of growth (OD₅₅₀ about 0.8), they were starved of N- or S-nu- trient and the growth monitored for the following 48 h of nutrient deprivation (Figure 1 and Figure 2). Cells were incubated in culture media with a starting cell density of approximately 414.062 cells/ml.

Some previous papers reported reduction in microalgal growth rate and increase of cell volume as a consequence of long-term N- and S-starvation [1] [16] .

We observed that cellular growth is more affected in N-starved cells if compared to S-starved ones (Figure 2(b)).

S-deprivation implies a reduction of cell growth that is less evident than in N-starvation: the fact that in S- starved cells the rate of growth has a slowdown delayed probably corresponds to a period of cellular sulphur recycling. As previously reported in Chlorella sorokiniana cells [14] , upon S-deprivation, the degradation of intracellular glutathione has the purpose to maintain Cys homeostasis. Besides, this growth discrepancy between N- and S-starved cells is likely caused by the relative importance and abundance of N compared with S: while S can be obtained from intracellular stores as glutathione [14] , N must be supplied constantly for adequate growth because it cannot be sufficiently compensated by recycling pathways. In fact, nitrogen occurs in every protein and most metabolites and total N content of dry C. sorokiniana biomass is almost 10-fold greater than the total S content. As such, a much greater mass of N ( 10 mM KNO₃) is necessary to facilitate growth of C. sorokiniana, while a comparatively lesser amount of S ( 1.2 mM MgSO₄) is sufficient. In 24 h of nutrient shortage, significant changes in the average cell volume did not occur (Figure 1). The slight cell enlargement caused by N- depriva- tion is likely due to the starch store occurring in N-starved cells (Figure 1 and Table 1). S-starved cells appear small and grouped suggesting a likely presence of mucilage material (Figure 1).

Our studies showed that N-starvation generally yielded similar effects as S-starvation, but the impacts on cell growth and total protein were much more severe.

Total soluble protein concentration in algal samples decreased drastically in response to nutrient starvation. A 65% decrease in protein content was observed after 24 h of S-deprivation and a decrease of over 80% was observed after 24 h of N-starvation (Figure 3). Typically, in microalgae, the decrease in protein content was simultaneous to the increase of starch (Table 1) and lipid upon nutrient starvation [16] [17] .

Strikingly, S-deficiency also affects free amino acids pool, which appeared greatly increased compared to sufficient algae [12] , while S-containing amino acids like Cys appeared at low level [14] . Among amino acids, glutamine, the first organic compound deriving from N-assimilation, increased following S-starvation and obviously resulted reduced in N-deprived cells (Figure 4). Noticeably, GS activity, which resulted enhanced in N- deprived cells, was not affected by S-deprivation (Table 2). The glutamine increase in S-starved cells could indicate an imbalance of N-assimilation and protein synthesis caused by the inadequacy of Cys or methionine. Furthermore, N-starvation affect intracellular Cys content. A reduction of 50% in cysteine intracellular level was observed after 24 h of N-starvation (Figure 4). In fact, the backbone for cysteine synthesis is O-acetylserine (OAS) which in turn derives from serine, a product of N-assimilation. The enzyme OASTL catalyses the synthesis of cysteine from OAS and sulphide and represents an important point of connection between the nitrogen and sulphur assimilation [7] [18] . Indeed, our results show that OASTL activities were strongly decreased in N- starved cells (Table 2). The enzyme OASTL exists in several isoforms that are localised into different compartments of the plant cell [19] [20] . S-deprivation induced OASTL activities increase in C. sorokiniana cell [7] [11] , similar to those observed in a number of higher plants [21] . The increase in OASTL activity during S-depriva- tion could be ascribed to an increase in the cytosolic OASTL isoform [7] [11] .

N- or S-deprivation causes a loss of photosynthetic capacity and a decrease in the cells chlorophyll content. As previously reported, chlorophyll content strongly decreased upon both S- and N-starvation [3] [22] . Several studies indicate the degradation of PSII as responsible for this decrease [23] . Chlamydomonas reinhardtii is known to restructure its photosynthetic machinery upon S-deprivation to minimize oxidative stress that may occur in conditions of reduced photosynthesis [1] . As such, the decrease in chlorophyll content could be ascribed to the deactivation and disassembly of photosynthetic complexes as a response to oxidative stress resulting from S-starvation. As well, protein degradation that occurs in N-starved cells does not save the enzyme ribulose-1,5-

bisphosphate carboxylase oxygenase [1] [24] and the depletion of this protein may compromise the mechanism of photosynthesis, leading to the decrease in chlorophyll content observed in N-deficient Chlorella cells.

Our previous results [3] [22] suggest that a global shutdown in energetic functions may occur upon N- or S- deprivation: photosynthesis and chlorophyll levels drop considerably, but also respiration resulted decreased.

Our recent investigation showed that activation of antioxidant enzymes in S-deprived C. sorokiniana occurred in a period of 24 h of S-starvation and that these increases correspond to the rapid raise in Reactive Oxygen Species (ROS) occurring in the first hour of starvation [25] .

In this paper we show that total C is reduced by 10% in N-deprived and 16% in S-deprived cells. Another interesting result is that in cells S-deprived the total N is reduced by only 23%, whereas in cells N-deprived the total S is reduced by 47%. These data seem to confirm that N-deprivation strongly affects S-assimilation (Figure 5).

It has also been shown that N-deprivation increases the C/N ratio (6.4), probably due to reduced synthesis of amino acids and proteins and to the storage of starch. In S-starved cells the C/N ratio slightly increased (4.7) respect to that of sufficient cells (4.3).