Characterization of arsenic compounds formed by Daphnia magna and Tetraselmis chuii from inorganic arsenate.

Experiments to grow Tetraselmis chuii (a marine alga) and Daphnia magna in the presence of inorganic arsenate are described. The algae incorporate arsenic rather efficiently and form a lipid-soluble organic arsenic compound. T. chuii has been successfully mass cultured in a medium containing 10 ppm arsenic as arsenate. Daphnia magna was cultured in a medium containing 74As-labeled H3AsO4 and 1 ppm Na2HAsO4 expressed as arsenic. The arsenic metabolites were extracted with a chloroform-methanol solution and isolated by using column and thin-layer chromatography. TLC analysis of the metabolites revealed the presence of a 74As-containing product which migrated with phosphatidylethanolamine. This product was hydrolyzed with the phospholipases A, C, and D. The experimental results are not inconsistent with the presence of an arsenocholine moiety in the lipids. Arsenocholine, arsenobetaine, and acetylarsenocholine have been synthesized and will serve as reference substances in the chromatography experiments. The preparation of arsenocholine-containing lipids is in progress.


Introduction
Arsenic has been shown to be incorporated into both marine and fresh water organisms (1,2) in the form of water-soluble and lipid-soluble organic arsenic compounds (2)(3)(4). Although fish and higher members of the aquatic food chain seem to be cap- tAuthor to whom all inquiries should be addressed. To avoid the difficulties inherent in judging the relative importance of the contributions of each participant in this interdisciplinary effort usually expressed in the sequence of names appearing on this paper, the position of each name was determined in an unbiased manner with the aid of a table of random numbers. able of synthesizing organic arsenic compounds by themselves, organisms which are at the lower end of the food chain are, however, very likely the prime source of these compounds (2,5,6). The arsenic containing, lipid soluble compounds appear to be closely related to phospholipids (3), especially in their ability to produce hydrolysis products including arsenate (as compared to ahosphate produced by phospholipids (7), to yield a positive ninhydrin reaction (6), and to migrate to Rf values which correspond to values characteristic of known phospholipids (7).
Because of the chemical similarities between arsenic and phosphorus, arsenic could perhaps replace phosphorus and form, for instance, an arsenolipid. Since arsenate esters are, in contrast to August 1977 phosphate esters, very easily hydrolyzed, arsenolipids are expected to be isolated only when all extractive operations are carried out under anhydrous conditions. It has been reported that hydrolysis of arsenic containing lipid samples extracted from cod and mackerel with chloroform-methanol (2: 1 v/v) yield two types of stable organic arsenic compounds: a water-soluble compound stable to 6.6N HCI (8) and a less polar compound which migrated in TLC experiments with the fatty acids (9). The identities of these lipid-and water-soluble organic arsenic compounds remain to be determined.
The investigations which are presently in progress at Texas A&M University and at the Agricultural Environmental Quality Institute, have as their goal the identification and characterization of the organic arsenic compounds formed by Daphnia magna and Tetraselmis chuii from inorganic arsenate. The two organisms were chosen because they are easy to grow, accumulate arsenic rather efficiently and are at the base of the food pyramid. Experiments with crayfish and lobsters were terminated because it was too difficult to maintain stable populations.
Our research group is presently engaged in three areas of activities, which in a synergistic manner should lead to the identification and isolation of the organic arsenic compound(s) formed by Daphnia and Tetraselmis chuii: large-scale culturing of Tetraselmis in the presence of arsenate to isolate gram quantities of the arsenic compound(s), smallscale experiments with Daphnia, and efforts to prepare arsenocholine, arsenobetaine, and arseniccontaining lipids.

Experiments with Tetraselmis chuii
The work of Lunde (2, 6) with marine algae made it possible to embark on a large-scale experiment to grow several hundred kilograms of algae in a salt water medium containing inorganic arsenate, extract the arsenic compound(s) with the lipids and separate the arsenic compound(s) from the lipids.
Twelve species of marine algae were first grown in test tubes in von Stosch and in Instant Ocean medium at arsenate levels (expressed as As) ranging from 0.5 to 50 ppm. Tetraselmis chuii was selected for large-scale experiments because of the organism's ruggedness and ability to efficiently take up arsenic. The growth behavior of T. chuii in the presence of arsenate (Na2HAsO4 * 7H20) is shown in Figure 1. The organism grew well in media with arsenic concentrations as high as 50 ppm. The lag phase increased linearly with the arsenic concentra-tion. Preliminary experiments employing media containing Na2HAsO4 spiked with gamma-active 74As-H3AsO4 proved that T. chuii incorporates considerable amounts of arsenic into the cell (Fig.  2). If a similar arsenic accumulation occurs in the large scale growth experiments, 5000 liters of algae cultures will produce several grams of the arsenic compound(s). . Growth of Teti-aselmis c-huii in arsenate-containing von Stosch medium: (0) 50 ppm As as arsenate:, (0) 30 ppm As; (0) 20 ppm As; (@) 10 ppm As; (0) 500 ppb As. When a sample of T. chuiii from a culture in the stationary phase in a low arsenic medium was transferred into a higher arsenic medium, the algae proceeded without much impairment through the three growth phases. By repeated transfer of algae into media of increasingly higher arsenic concentrations, organisms were obtained, which multiplied and thrived at 1000 ppm of arsenic. Transfer of these arsenic-tolerant algae into an arsenic-free medium caused the demise of the culture. Similarly, algae grown in arsenic-free medium did not survive when transferred into a 1000 ppm arsenic culture solution. Additional experiments in this area are underway.
Tetr-aselmiiis chIiiii was successfully mass cultured in four batches of 300 liters each in a medium containing 10 ppm of nonradioactive arsenic as arsenate. After 6 days, six liters of algae were separated by centrifugation, frozen, and stored for later processing. A room with appropriate environmental controls has been secured and four growing tanks, each holding 1000 liters, have been constructed. All necessary equipment for preparing artificial seawater and processing the algae has been installed. Large-scale culturing of algae will commence soon.
The methods of separating the lipids from the algae and isolating the arsenic compound(s) from the lipids are similar to the ones described for the Daphnia experiments. Analytical and preparative high-pressure chromatography will be used extensively in the final purification of the arsenic compounds. Preliminary results obtained with the arsenic compounds from algae suggest that they are similar to the one found in Daphnia.

Experiments with Daphnia magna
The Daphnia were cultured in an arsenic-free nutrient solution containing a thick growth of Euglena. Daphnia samples, estimated to weigh between 5 and 8 g were then removed and placed into another culture medium containing NaH2AsO4 (1 ppm As) and 74As-H:As04. After 3 days the Daphnia were harvested, washed with distilled water, air dried, and then treated with cold trichloroacetic acid-water (5:95, v/v) to remove low molecular mass compounds (10). The separated solids were washed with distilled water and air dried. The air-dried sample was extracted (11) with CHC13/CH:tOH (2: 1, v/v). The chlor6form layer, which contains the lipids was analyzed by column chromatography (silica gel) with CHC13/CH30H/H2O/CHiCOOH (65:25:4:6v/v) as eluent and by dry column chromatography (12). Figure 3 summarizes these operations in a flow-chart.  The lipid-containing chloroform layer accounted for approximately half of the 74As activity present in the Daphnia. A dry column chromatographic separation of this extract produced a yellow-brown band with Rf 0.58 which contained most of the 74As activity. An appreciable amount of radioactivity remained at the top of the column.
Another sample of the chloroform layer was loaded on a silica gel column and eluted with the CHCI3/CH30H/H20/CH3COOH mixture. The yellow-brown eluate was further separated by TLC with the above solvent mixture. The chromatogram after spraying with ninhydrin solution revealed two violet spots, one at Rf 0.57 (62%) and another of less intensity of Rf 0.29 (15%). With dichlorofluorescein under uitraviolet light, four more spots at Rf 0.10 (5%), 0.34 (7%), 0.67 (8%), and 0.79 (13%) were detected. The percentages in parentheses represent the fraction of the total 74As activity on the TLC plate found with the particular spots. The ninhydrin-positive spot at Rf 0.57 contains most of the activity and has the same Rf value as the yellow-brown band on the dry column.
Various phospholipids and a sample of the yellow-brown eluate were simultaneously chromatographed two-dimensionally, first with CHCI3/CH30H/NH3 and then with the CHCl:3/CH30H/H20/CH3COOH system. The ninhydrin-positive spots with Rf 0.57 and Rf 0.29 migrated in this experiment to the same Rf values as phosphatidylethanolamine and lysophosphatidylethanolamine, respectively.
When the TLC analysis of the yellow-brown fraction was performed immediately after elution only the spot with Rf 0.57 was detected. A thin layer chromatogram of a 3-day-old sample of the eluate even when stored under nitrogen at 0°C showed all six spots described above. Storage of the sample under nitrogen at -12°C prevented the formation of the compounds responsible for the additional five spots.
The yellow-brown fraction was subjected to enzymatic hydrolysis by phospholipase A, C, and D (13)(14)(15). The hydrolysis mixture was extracted with diethyl ether. Both, the organic and the aqueous phase were characterized by TLC and counted for 74As-activity.
The following 74As-activity distribution between the aqueous and the ether phase was found [phospholipase (PL), % activity in H20, % activity in ether]: PL A 98%, 2%; PL C, 3%, 97%; PL D 43%, 57%. Upon total hydrolysis and methylation with CH30H/H2S04, 7% of the activity was found in the ether phase and 93% in the aqueous layer.
When the lipids were cleaved by PLA, the reaction mixture concentrated and a sample chromatographed, a ninhydrin-positive, 74As-active TLC spot was observed. The ether extract contained palmitic acid but little radioactivity.
Hydrolysis by PL C and extraction produced an ether layer containing 97% of the activity. TLC analysis of the ether layer produced a ninhydrinnegative, 74As-active spot which tailed from the The arsenic-containing materials in the PL D hydrolysis mixture partitioned about equally between the ether and the water layer. The TLC of the ether layer gave a ninhydrin-negative, 74As-active spot at Rf 0.39, a ninhydrin-positive, but inactive spot at Rf 0.67, and a ninhydrin-negative, inactive spot with Rf '-, 1.0. The material in the aqueous layer, which was ninhydrin-positive and radioactive, did not migrate.
A sample of the aqueous layer obtained upon complete hydrolysis of the yellow-brown eluate spotted on TLC and developed with CHC13/CH30H/NH3 similarly did not migrate. The spot was radioactive and ninhydrin-positive. Based on these very preliminary results, suggestions as to the nature of the organic arsenic compounds formed by Daphnia can be made. The further experiments, which are now being carried out, will show whether or not these suggestions have any merit.
Structures 1-VI suggest how arsenic could be incorporated into lipids. It is of course possible that arsenic does not appear as part of a lipid molecule but has been transformed into an arsenic compound, whose properties allow it to be separated together with the lipids. At the moment, it seems to be expedient to let the arsenic atom become part of a lipid molecule and take the place of the nitrogen atom in the choline group bonded to glycerine via a phosphate group. Considering the sites of attack of the phospholipases as shown in VII, an attempt can be made to explain the experimental results on the basis of the structures and likely properties of the arsenic-containing lipids I-V I. Let us assume then, that the arsenic is present in the lipid molecule as a replacement for nitrogen in choline but not in ethanolamine. The AsH2 group is extremely oxygen sensitive and toxic. Organisms probably prefer methylated arsenic groups. Such a lipid could have formula II or perhaps I. PL A hydrolysis will remove the /8-carboxylic acid in the choline-containing lipid and perhaps in the arsenocholine-containing lipid. Since the N-and As-lipids seem to have similar TLC properties, a 74As-active, ninhydrin-positive (from lysophosphatidylethanolamine) spot would be expected. The experiment produced such a spot with a Rf value of 0.29. PL C cleaves the phosphoethanolamine group. It is conceivable that a compound with arsenic in the place of nitrogen is for steric reasons not able to serve as a substrate for the enzyme. The arseniclipid would not be hydrolyzed. It would move into the ether layer and not give a ninhydrin reaction as observed.
PL D gave a hydrolysis mixture, half of whose activity partitioned into the ether layer. The phospholipid will lose its ethanolamine group, which will be soluble in the aqueous phase. Perhaps the en-zyme does interact with, for instance compound II, but at a rate slower than that characteristic for the nitrogen derivative. The arsenocholine would go into the aqueous phase, the intact arsenic lipid would stay in the ether phase as observed. Add total hydrolysis in pursuit of our third activity in this area.
All these experimental results are not incompatible with the presence of an arsenocholine group in a lipid of structure II. The data available at this time allow, of course, no definite statement at all about the mode of occurrence of arsenic in lipids. Our preliminary, tentative and intriguing "arsenocholine" hypothesis does, however, provide some guidance in our search for the elusive organic arsenic compounds formed by algae and Daphnia.
We have recently prepared arsenocholine and arsenobetaine and are working on the synthesis of arsenic lipids of the type I, 11, IV, and V. Enzymatic hydrolysis reactions will be carried out with these compounds. They will also be used as reference derivatives in the chromatographic investigations.
The frequently mentioned possibility of arsenate replacing phosphate in biologically important molecules leads to the question: could lipids be formed in which arsenate replaced phosphate as in compound III? Such a derivative could be stored only under anhydrous conditions and would certainly not survive the extraction procedures which August 1977 rL t-rL L) are not at all anhydrous. Arsenate esters are extremely easily hydrolyzed. Several of the aqueous hydrolysis solutions were chromatographed simultaneously with arsenate and the chromatogram sprayed with silver nitrate. Arsenate was not detected in the hydrolysis mixture. No radioactivity was found on the chromatogram of the hydrolyzate corresponding to the arsenate region. The replacement of phosphate by arsenate in lipids is, therefore, considered to be unlikely.
Since phosphonolipids do occur in nature, arsonolipids could exist as well. These compounds IV and VI and the arsinolipid V have all an arsenic-carbon bond. As long as there is not a hydroxyl group in fl-position to the arsenic atom the derivatives should be stable. Hydroxyethyl arsenic compounds have a tendency toward arsenic-carbon bond cleavage under relatively mild conditions. If an arsonoor arsinolipid (IV, V) had been present, PL D would have had no effect on the molecule.
The situation becomes, of course, more complex when a compound of type VI is considered, especially if arsenic also replaces the nitrogen atom in this lipid molecule.
We believe that the nature of the lipid-soluble organic arsenic compounds will be known soon. Whether arsenocholine is one of these compounds remains to be seen. These investigations are supported by an NIEHS grant, Grant Number ES 01125-02. The partial support for construction of the algae growing facilities through Texas A&M University's Fund for Organized Research is acknowledged. We express our appreciation to Dr. Neal and Mrs. Ross of the Gulf Coastal Fisheries Service Laboratories in Galveston, Texas for their help and advice and making their facilities available for the initial algae mass culture experiments.