Formation of thionates by freshwater and marine strains of sulfate-reducing bacteria

The formation of thionates (thiosulfate, trithionate and tetrahionate) during the reduction of sulfate or sulfite was studied with four marine and four freshwater strains of sulfate-reducing bacteria. Growing cultures of two strains of the freshwater species Desulfovibrio desulfuricans formed up to 400 M thiosulfate and 100 M trithionate under conditions of electron donor limitation. Tetrathionate was observed in lower concentrations of up to 30 M. Uncoupler-treated washed cells of the four freshwater strains formed thiosulfate and trithionate at low electron donor concentrations with sulfite in excess. In contrast, only one of four marine strains formed thionates. The freshwater strain Desulfobulbus propionicus transformed sulfite almost completely to thiosulfate and trithionate. The amounts produced increased with time, concentration of added sulfite and cell density. Tetrathionate was detected only occasionally and in low concentrations, and was probably formed by chemical oxidation of thiosulfate. The results confirm the diversity of the sulfite reduction pathways in sulfate-reducing bacteria, and suggest that thiosulfate and trithionate are normal by-products of sulfate reduction.Abbreviations CCCP carbonyl cyanide m-chlorophenylhydrazone     

Adenylylsulfate reductase in some new sulfate-reducing bacteria

Four recently described species of new genera of sulfate-reducing bacteria, Desulfobulbus propionicus, Desulfobacter postgatei, Desulfococcus multivorans and Desulfosarcina variabilis were examined with respect to adenylylsulfate reductase. All of the species examined contained the enzyme in sufficient concentrations to account for dissimilatory sulfate reduction.Adenylylsulfate reductase was enriched 17.1-fold from Desulfobulbus propionicus by ammonium sulfate fractionation, ion exchange chromatography and gel filtration. The molecular weight was 175,000 and the enzyme contained 1 mol of flavin, 8 mol of non heme iron and 8 mol of labile sulfide per mol enzyme. Either ferricyanide or cytochrome c could be used as electron acceptors; the pH optimum was 7.7 with ferricyanide and 8.8 with cytochrome c. K _m values for AMP and sulfite were 90 M and 1.3 M with ferricyanide and 91 M and 71 M with cytochrome c as electron acceptor. K _m values for ferricyanide and cytochrome c were 89 M and 21 M, respectively. The properties of the enzyme were compared with those of purified adenylylsulfate reductases from other microorganisms.Non-common abbreviation APS adenylylsulfate     

Protonmotive force in freshwater sulfate-reducing bacteria,and its role in sulfate accumulation in Desulfobulbus propionicus

The protonmotive force in several sulfate-reducing bacteria has been determined by means of radiolabelled membrane-permeant probes (tetraphenyl-phosphonium cation, TPP⁺, for , and benzoate for pH). In six of ten freshwater strains tested only the pH gradient could be determine, while the membrane potential was not accessible due to nonspecific binding of TPP⁺. The protonmotive force of the other four strains was between –110 and –155 mV, composed of a membrane potential of –80 to –140 mV and a pH gradient between 0.25 and 0.8 (inside alkaline) at pH_out=7. In Desulfobulbus propionicus the pH gradient decreased with rising external pH values. This decrease, however, was compensated by an increasing membrane potential. Sulfate, which can be highly accumulated by the cells, did not affect the protonmotive force, if added in concentrations of up to 4 mM. The highest sulfate accumulation observed (2500-fold), which occurred at external sulfate concentrations below 5 M, could be explained by a symport of three protons per sulfate, if equilibrium with the protonmotive force was assumed. At higher sulfate concentrations the accumulation decreased and suggested an electroneutral symport of two protons per sulfate. At sulfate concentrations above 500 M, the cells stopped sulfate uptake before reaching an equilibrium with the protonmotive force.Abbreviations CCCP carbonyl cyanide m-chlorophenylhydrazone - MOPS morpholinopropanesulfonic acid - TPP⁺ tetraphenylphosphonium cation - EDTA ethylenediaminetetraacetic acid - pH transmembrane pH gradient (pH_in-pH_out) - transmembrane electrical potential difference     

Chemolithotrophic growth ofDesulfovibrio desulfuricans with hydrogen coupled to ammonification of nitrate or nitrite

Two of nine sulfate reducing bacteria tested,Desulfobulbus propionicus andDesulfovibrio desulfuricans (strain Essex 6), were able to grow with nitrate as terminal electron acceptor, which was reduced to ammonia. Desulfovibrio desulfuricans was grown in chemostat culture with hydrogen plus limiting concentrations of nitrate, nitrite or sulfate as sole energy source. Growth yields up to 13.1, 8.8 or 9.7 g cell dry mass were obtained per mol nitrate, nitrite or sulfate reduced, respectively. The apparent half saturation constants (K _s) were below the detection limits of 200, 3 or 100 mol/l for nitrate, nitrite of sulfate, respectively. The maximum growth rates {ie63-1} raised from 0.124 h^-1 with sulfate and 0.150 h^-1 with nitrate to 0.193 h^-1 with nitrite as electron acceptor. Regardless of the electron acceptor in the culture medium, cell extracts exhibited absorption maxima corresponding to cytochromec and desulfoviridin. Nitrate reductase was found to be inducible by nitrate or nitrite, whereas nitrite reductase was synthesized constitutively. The activities of nitrate and nitrite reductases with hydrogen as electron donor were 0.2 and 0.3 mol/min·mg protein, respectively. If limiting amounts of hydrogen were added to culture bottles with nitrate as electron acceptor, part of the nitrate was only reduced to the level of nitrite. In media containing nitrate plus sulfate or nitrite plus sulfate, sulfate reduction was suppressed.The results demonstrate that the ammonification of nitrate or nitrite can function as sole energy conserving process in some sulfate-reducing bacteria.     

Product inhibition during the feeding phasein fed-batch ethanol fermentation of sugar-cane blackstrap molasses

Summary The following equations represent the influence of the ethanol concentration (E) on the specific growth rate of the yeast cells () and on the specific production rate of ethanol () during the reactor filling phase in fed-batch fermentation of sugar-cane blackstrap molasses: = ₀ - k · E and v = v ₀ · K/(K +E) Nomenclature E ethanol concentration in the aqueous phase of the fermenting medium (g.L^–1) - E_m value of E when = 0 or = 0 (g.L^–1) - F medium feeding rate (L.h^–1) - k empirical constant (L.g^–1.h^–1) - K empirical constant (g.L^–1) - M_as mass of TRS added to the, reactor (g) - M_cs mass of consumed TRS (g) - M_e mass of ethanol in the aqueous phase of the fermenting medium (g) - M_s mass of TRS in the aqueous phase of the fermenting medium (g) - M_x mass of yeast cells (dry matter) in the fermenting medium (g) - r correlation coefficient - S TRS concentration in the aqueous phase of the fermenting medium (g.L^–1) - S_m TRS concentration of the feeding medium (g.L^–1) - t time (h) - T temperature (° C) - TRS total reducing sugars calculated as glucose - V volume of the fermenting medium (L) - V₀ volume of the inoculum (L) - X yeast cells concentration (dry matter) in the fermenting medium (g.L^–1) - filling-up time (h) - specific growth rate of the yeast cells (h^–1) - ₀ value of when E=0 - specific production rate of ethanol (h^–1) - ₀ value of when E=0 - density of the yeast cells (g.L^–1) - dry matter content of the yeast cells     

Inoculation of sugar mill by-products compost with N2-fixing bacteria

Inoculation of sugar mill by-products compost with N₂-fixing bacteria may improve its quality by increasing total N and available P. Compost was inoculated with Azotobacter vinelandii(ATCC 478), Beijerinckia derxii (ATCC 49361), and Azospirillumsp. TS8, each alone and all three together. Numbers of all N₂-fixing bacteria in compost declined from an initial population of 5×10⁵cellsg^–1 during incubation. The population of Azotobacter declined to approximately 2×10²cellsg^–1 and the population of Beijerinckia and Azospirillum declined to approximately 9×10³ and 3.5×10⁴cellsg^–1 respectively, at day 50. Inoculation with N₂-fixing bacteria increased acetylene reduction, total N by 6–16 and available P by 25–30% in comparison to the uninoculated control. Increasing the N content and P availability of compost increases its value and there may be additional benefit from providing N₂ fixing bacteria.     

Concentrations of fumonisin B1 in feeds associated with animal health problems

Ninety-eight samples of feeds associated with 44 cases of equine leukoencephalomalacia (ELEM) and 83 samples of feed associated with 42 cases of a porcine pulmonary edema syndrome (PPE) were analyzed for fumonisin B₁ (FB₁). For comparison purposes, 51 feed samples not associated with PPE or ELEM were also analyzed. Feed associated with ELEM contained FB₁ ranging from less than 1 g/g to 126 g/g with 75% of the cases having at least 1 sample above 10 g/g. Feeds associated with PPE ranged from less than 1 g/g to 330 g/g with 71% of the cases having at least 1 sample greater than 10 g/g. Quantitation was by high performance liquid chromatography (HPLC)/fluorescence using the fluorescamine derivative with confirmation by thin layer chromatography (TLC) and/or gas chromatography/mass spectroscopy (GC/MS).     

Role of mitochondria in ethanol tolerance of Saccharomyces cerevisiae

The presence of active mitochondria and oxidative metabolism is shown to be essential to maintain low inhibition levels by ethanol of the growth rate (), fermentation rate (v) or respiration rate () of Saccharomyces cerevisiae wild type strain S288C. Cells which have respiratory metabolism show K _i (ethanol inhibition constant) values for , v and , higher (K _i>1 M) than those of petite mutants or grande strains grown in anaerobiosis (K _i=0.7 M). In addition, the relationship between or v and ethanol concentration is linear in cells with respiratory metabolism and exponential in cells lacking respiration. When functional mitochondria are transferred to petite mutants, the resulting strain shows K _i values similar to those of the grande strain and the inhibition of and v by increasing ethanol concentrations becomes linear.     

Plant regeneration via somatic embryogenesis in Styrian pumpkin: cytological and biochemical investigations

Somatic embryo formation was induced from cotyledon explants of Styrian pumpkin (Cucurbita pepo L. subsp. pepo var. styriaca Greb.) by using a solid MS medium supplemented with 16.11M NAA and 4.44M BA or 26.85M NAA and 13.32M BA. The callus proliferation was more efficient on medium supplemented with 26.85M NAA and 13.32M BA. In contrast, the embryogenic response was higher on medium with lower concentrations of growth regulators (16.11M NAA and 4.44M BA). The time needed for embryo induction did not depend on medium composition. Embryos in globular stage were transferred to three different maturation media, containing 2.89M GA₃ in combination with 0.54M NAA, 11.42M IAA and growth regulator-free medium. The germination rate was the highest when embryos were cultured on medium with 11.42M IAA. Plantlets grown on this medium achieved maturity suitable for transplantation into soil within 9 to 10weeks. The regenerated plants were successfully transferred into field and developed fertile flowers and set fruits. Biochemical analysis showed significant lower total glutathione levels among in vitro grown plantlets compared to seedlings grown in soil. When the plantlets were transferred into soil, they reached a normal size within a month and the glutathione concentration was comparable to seed-derived plants at the same developmental stage. Transmission electron microscopy was used to investigate possible differences in the ultrastructure of cells from callus cultures, and leaf cells of regenerated and seed-derived plants. Differences in the ultrastructure were found within chloroplasts which contained only single thylakoids, large starch grains and small plastoglobuli in callus cells in comparison to leaf cells, which possessed a well developed thylakoid system, small starch grains and large plastoglobuli.     

Methylobacterium rhodesianum MB 126 possesses two acetoacetyl-CoA reductases

Two constitutive acetoacetyl-CoA (AcAc-CoA) reductases were purified from Methylobacterium rhodesianum MB 126, an NADPH-linked d(-)--hydroxybutyryl-CoA forming reductase (enzyme A) and an NADH-and NADPH-linked l(+)--hydroxybutyryl-CoA forming reductase (enzyme B). Enzyme A and B give apparent K _m values of 15 M and 30 M for AcAc-CoA, 18 M for NADPH and 30 M for NADH, respectively. They are inhibited by AcAc-CoA at concentrations higher than 25 M and 50 M, respectively. The contribution of the two reductases to poly--hydroxybutyrate synthesis is discussed.     

Control of yeast fed-batch process through regulation of extracellular ethanol concentration

At high growth rates, the biomass yield of bakers yeast (Saccharomyces cerevisiae) decreases due to the production of ethanol. For this reason, it is standard industrial practice to use a fed-batch process whereby the specific growth rate, , is fixed at a level below the point of ethanol production, i.e., _crit. Optimally, growth should be maintained at _crit, but in practice, this is difficult because _crit is dependent upon strain and culture conditions. In this work, growth was maintained at a point just above _crit by regulating ethanol concentration in the bioreactor. The models used for control design are shown, as are the experimental results obtained when this strategy was implemented. This technique should be applicable to all microorganisms that exhibit an overflow type metabolism.     

Desulfovibrio brasiliensis sp. nov., a moderate halophilic sulfate-reducing bacterium from Lagoa Vermelha (Brazil) mediating dolomite formation

A novel halotolerant sulfate-reducing bacterium, Desulfovibrio brasiliensis strain LVform1, was isolated from sediments of a dolomite-forming hypersaline coastal lagoon, Lagoa Vermelha, in the state of Rio de Janeiro, Brazil. The cells are vibrio-shaped and 0.30 to 0.45 m by 1.0 to 3.5 m in size. These bacteria mediate the precipitation of dolomite [CaMg(CO₃)₂] in culture experiments. The strain was identified as a member of the genus Desulfovibrio in the -subclass of the Proteobacteria on the basis of its 16S rRNA gene sequence, its physiological and morphological properties. Strain LVform1 is obligate sodium-dependent and grows at NaCl concentrations of up to 15%. The 16S rRNA sequence revealed that this strain is closely related to Desulfovibrio halophilus (96.2% similarity) and to Desulfovibrio oxyclinae (96.8% similarity), which were both isolated from Solar Lake, a hypersaline coastal lake in the Sinai, Egypt. Strain LVform1 is barotolerant, growing under pressures of up to 370 bar (37 MPa). We propose strain LVform1 to be the type strain of a novel species of the genus Desulfovibrio, Desulfovibrio brasiliensis (type strain LVform1 = DSMZ No. 15816 and JCM No. 12178). The GenBank/EMBL accession number for the 16S rDNA sequence of strain LVform1 is AJ544687.     

Problems in estimating growth rates of marine phytoplankton from short-term14C assays

Growth rate estimates () of phytoplankton populations that were sampled from nitrogen-limited continuous cultures and then incubated for short durations in batch culture with added¹⁴C-HCO₃ ^– were significantly different than steady-state growth rates () for 3 of 5 marine phytoplankton species. Two diatoms,Thalassiosira weissflogii andChaetoceros simplex, displayed virtually identical growth rates (=) over a wide range of, whereas for a third diatom,Phaeodactylum tricornutum, was overestimated by an average of 40% compared to. In contrast, was underestimated by the¹⁴C technique for the two remaining species: up to 40% at a steady-state of 1.0 day^–1 for the chlorophyteDunaliella tertiolecta and up to 100% at of 1.4 day^–1 for the haptophytePavlova lutheri. For the latter two species the divergence between and appeared to increase with increasing steady-state. A simple model of labeled and total carbon flow between the aqueous phase and cellular biomass was constructed to demonstrate that respiration was negligible when=, but was significant when>. In the cases in which<, a rapid physiological alteration presumably took place once the steady state was disturbed and cells were placed in the incubation chambers, which perhaps was related to the nutritional state of the cultures at the time of sampling. Questions thus are raised regarding our ability to measure accurately primary productivity from shipboard experiments with confined samples of phytoplankton from nutrient-impoverished waters that probably are less hardy than the laboratory cultures used in these studies.     

Different Ks values for hydrogen of methanogenic bacteria and sulfate reducing bacteria: An explanation for the apparent inhibition of methanogenesis by sulfate

Desulfovibrio vulgaris (Marburg) and Methanobrevibacter arboriphilus (AZ) are anaerobic sewage sludge bacteria which grow on H₂ plus sulfate and H₂ plus CO₂ as sole energy sources, respectively. Their apparent K_s values for H₂ were determined and found to be approximately 1 M for the sulfate reducing bacterium and 6 M for the methanogenic bacterium. In mixed cell suspensions of the two bacteria (adjusted to equal V _max) the rate of H₂ consumption by D. vulgaris was five times that of M. arboriphilus, when the hydrogen supply was rate limiting. The apparent inhibition of methanogenesis was of the same order as expected from the different K_s values for H₂. Difference in substrate affinities can thus account for the inhibition of methanogenesis from H₂ and CO₂ in sulfate rich environments, where the H₂ concentration is well below 5 M.     

Effects of light on nitrate-limited Oscillatoria agardhii in chemostat cultures

The effects of the average light irradiance (I) on growth and nitrate uptake kinetics of the cyanobacterium Oscillatoria agardhii, in nitrate-limited chemostat cultures, were studied. Light was nonsaturating for I <9.4 Wm^–2, for all growth rates () studied. However, was throughout limited by the availability of nitrate. Under light-saturating conditions the kinetics of nitrate-limited growth could be adequately described by both the Monod and Droop equations. Under light-non-saturating conditions the internal nitrogen content (Q) was a function of both and I, for which new formulas were derived. The high uptake capacity (V _max) of nitrate-limited cells was independent of , but was significantly increased for cells growing at I <9.4 Wm^–2. The half-saturation constant for nitrate uptake (K _s ^u ) increased with increasing , but was independent of the prevailing light conditions. The effects of light during nitrate-limited growth were associated with the regulation in the nitrogen-containing pigments.The results reported herein have important consequences for the use of Q, K _s ^u and V _max values as indicators of nutrient-deficiency of natural populations.     

Effects of GABA antagonists,SR 95531 and bicuculline,on GABAA receptor-regulated chloride flux in rat cortical synaptoneurosomes

Synaptoneurosomes isolated from cerebral cortices of male Sprague-Dawley rats were used for studying GABA_A receptor-regulated chloride influx. The in vitro effects of GABA antagonists, SR 95531 (a pyridazinyl GABA derivative) and bicuculline, on pentobarbital-stimulated, muscimol-stimulated or flunitrazepam-enhanced, muscimol-stimulated chloride uptake were studied. The chloride uptake was determined at 30°C, for 5 sec. Pentobarbital and muscimol produced a maximal stimulation of chloride uptake in cortical synaptoneurosomes at 500 M and 50M, respectively. SR 95531 as well as bicuculline had no effect on the basal uptake of chloride. Whereas, SR 95531 (0.3–30 M) and bicuculline (0.1–100 M), when added 5 min before muscimol (50 M), produced a significant concentration-dependent inhibition of muscimol (50 M)-stimulated chloride uptake (IC₅₀ s of 0.89±0.11 M and 13.45±2.10M, respectively). In studies of the inhibitory effects of SR 95531 and bicuculline on pentobarbital (500 M)-stimulated chloride uptake, the IC₅₀ s were 0.81±0.12 M and 3.86±1.14 M, respectively. SR 95531 exhibited a more potent inhibitory effect than bicuculline on flunitrazepam-enhanced, muscimol-stimulated chloride uptake. The results revealed that SR 95531 has a more potent antagonistic effect than bicuculline on GABA_A-regulated chloride flux.     

Binding of [3H]muscimol to calf cerebrocortical synaptic membranes and the effects of sulphur-containing convulsant and non-convulsant compounds

Endogenous and xenobiotic sulphur-containing convulsant and non-convulsant compounds containing structural moieties of, or bearing a structural resemblance to, GABA and homocysteine were tested in binding studies for their potency in displacing the GABA-mimetic [³H]muscimol from specific, high-affinity sites (K _d=3.6 nM;B _max=3.94 pmol/mg protein) on freeze-thawed, Triton-treated calf-brain synaptic membranes. The xenobiotic convulsants, 4-mercaptobutyric acid (MBA), 3-mercaptopropionic acid (3-MPA) and 2-mercaptopropionic acid (2-MPA) were found to be two-site competitive inhibitors exhibiting apparent inhibition affinity constants (K _i ^app ) of 5000 M, 3750 M, and 4800 M, respectively; while homocysteic acid (K _i ^app =4800 M) was shown to be a one-site partial competitive inhibitor. Intermediary metabolites of methionine: S-adenosyl-l-homocysteine,l-cysteine, the convulsantl-homocysteine, and its non-convulsant disulphide oxidation product, homocystine, were found to be one-site partial competitive inhibitors exhibitingK _i ^app values of 5750 M, 8350 M, 5000 M, and 510 M, respectively. The endogenous anticonvulsant neuroeffector, taurine, and the tripeptide, reduced glutathione (GSH) were shown to be, respectively, one-site (K _i=20 M) and two-site (K _i ^app =4300 M) competitive inhibitors of [³H]muscimol binding. These findings are discussed with regard to a previously proposed mechanism for the convulsant action of homocysteine.     

Image processing technique for measuring specific growth rates of Gibberella fujikuroi

Summary Two methods were compared for estimating of Gibberella fujikuroi grown with different proportions of glucose and starch. They were, =Ln2 (V_r/Le) on Petri dish (V_r= rate of tip extension and L_e= mean hyphal length) and, ₂=d (LnX)/dt in stirred fermenter. Values of ₁ and ₂ were in close agreement with each other.     

Distance-constraint approach to protein folding. I. Statistical analysis of protein conformations in terms of distances between residues

A statistical analysis of protein conformations in terms of the distance between residues, represented by their C atoms, is presented. We consider four factors that contribute to the determination of the distanced _i,i+k between a given pair ofith and(i+k)th residues in the native conformation of a globular protein: (1) the distancek along the chain, (2) the size of the protein, (3) the conformational states of theith to(i+k)th residues, and (4) the amino acid types of the and(i+k)th residues. In order to account for the dependence on the distancek along the chain, the statistics are taken for three ranges, viz., short, medium, and long ranges (k8; 9k20; andk21; respectively). In the statistics of short-range distances, a mean distanceD _k and its standard deviationS _k are calculated for each value ofk, with and without taking into account the conformational states of all residues fromi toi+k (factors 1 and 3). As an Appendix, the relations for converting from the distances between residues into other conformational parameters are discussed. In the statistics of long-range distances, a reduced distanced* _ij (the actual distance divided by the radius of gyration) is used to scale the data so that they become independent of protein size, and then a mean reduced distanceD _l (a, a) and its standard deviation _l (a, a) are calculated for each amino acid pair (a, a) (factors 2 and 4). The effect of the neighboring residues along the chain on the value of the distanced* _ij is explored by a linear regression analysis between the actual reduced distanced* _ij and the mean value over theD _l for all possible pairs of residues in the two segments of the (i–2)th to the (i+2)th and the (j–2)th to the (j+2)th residues. The effect is assessed in terms of the tangentA _l (a, a) of the calculated regression line for each amino acid pair (a, a). In the statistics of medium-range distances, only factors 1 and 4 are considered, to simplify the analysis. The scaled distanced _i,i+k =(d _i,i+k -D _k )/S _k is used to eliminate the dependence onk, the distance along the chain. The propertiesD _m (a, a), _m (a, a) andA _m (a, a) corresponding toD _l (a, a), _l (a, a), andA _l (a, a), and also calculated for each amino acid pair (a, a). The results are interpreted as follows: the smaller values ofD _l (a, a) andD _m (a, a) indicate a preference of the pair (a, a) for a contact (e.g., pairs between hydrophobic amino acids, and pairs of Cys with aromatic amino acids), and the larger values of these quantities indicate a preference for distant mutual location (e.g., pairs between strong hydrophilic amino acids); the smaller values of _l (a, a) and _m (a, a) indicate a strong preference for either contact or noncontact (e.g., pairs between hydrophobic amino acids, and pairs between strong hydrophobic and hydrophilic amino acids, respectively), and the larger values of these quantities indicate the ambivalent/neutral nature of the preference for contact and noncontact (e.g., pairs containing Ser or Thr); the smaller values ofA _l (a, a) andA _m (a, a) indicate that the distance of an (a, a) pair is determined independently of the amino acid character of the neighboring residues along the chain (e.g., some pairs of Cys or Met with other amino acids) and the larger values of these quantities indicare that such amino acid character contributes strongly to the determination of the distance (e.g., pairs containing Ser or Thr, and pairs between amino acids with small side chains). The difference between the statistics for the long- and medium-range distances is also discussed; the former reflect the difference between the hydrophobic and hydrophilic character of the residues, but the latter cannot be easily interpretable only in terms of hydrophobicity and hydrophilicity. The data analyzed here are used in the optimization of an object function to compute protein conformation in a subsequent paper.     

Somatic embryogenesis in callus cultures of Rosa hybrida L. cv. Landora

Callus was initiated from immature leaf and stem segments of rose (Rosa hybrida cv. Landora) and subcultured every four weeks on a basal medium of half-strength Murashige & Skoog (1962) salts plus 30 g l^-1 sucrose (1/2 MS) and supplemented with 2.2 M BA, 5.4 M NAA and 2.2–9.0 M 2,4-D. Embryogenic callus and subsequently somatic embryos were obtained from 8-week-old callus culture on 1/2 MS+2.2 M BA+0.05 M NAA+0.3 M GA₃+200–800 mg l^-1 L-proline. Long-term cultures were established and maintained for up to 16 months by repeated subculture of embryogenic callus on L-proline deficient medium. About 12% of cotyledonary stage embryos taken from cultures cold-stored at 8±1°C for 4 days germinated on 1/2 MS+2.2 M BA+0.3 M GA₃+24.7 M adenine sulphate.Abbreviations BA benzyladenine - NAA -naphthaleneacetic acid - 2,4-D 2,4-dichlorophenoxyacetic acid - GA₃ gibberellic acid