Young leaves and peduncles of Prunus yedoensis and P. lannesiana trees grown in the Yokohama National University, were used directly for induction of suspension cultures, or leaves were used for the sandwich method after treatment without or with 0.5% NaClO aqueous solution for 12 min, and after drying at 60˚C for 18 hrs. Lactuca sativa (lettuce) seeds (Great Lakes) were sterilized with 1.5% NaClO aqueous solution for 15 min, washed with autoclaved water three times and sown on 0.8% agar medium. Cotyledons of 6-to 12-day-old seedlings, which were grown aseptically under a light condition at 25˚C, were used for protoplast isolation [6] .

The sandwich method of dried leaves of P. yedoensis and P. lanesiana was performed as reported [2] [18] . Briefly, 10 or 50 mg dry weight of leaves were sandwiched between two layers of 5 mL of 0.5% agar (gelling temp. 30˚C - 31˚C, Nacalaitesque Co. Ltd. Kyoto Japan) in six multi-well dish (Nunc). Five seeds of lettuce were put on per well. After 3 days of incubation in the dark at 20˚C, length of hypocotyls and roots was measured. Data were averaged of three wells and described as % growth of the control without leaves with standard deviation (SD).

Leaves of P. yedoensis were sterilized with a neutral detergent and 1% NaClO aqueous solution for 12 min and peduncles of P. lannesiana and P. yedoensis were sterilized with 2.5% NaClO aqueous solution for 30 min, and washed with autoclaved water three times. Two 3 mm sections were put in a 1 mL liquid medium of Murashige and Skoog’s (MS) [19] basal medium containing 10 μM each of naphthalene acetic acid and benzyladenine (BA) and 3% sucrose in flat-bottomed 10-mL tubes as reported previously for petiole and peduncle of P. yedoensis [16] . Maintenance of suspension culture was performed in 20 mL of the same medium composition in a 100 mL flask. They were incubated in the dark on a shaker at 100 rpm speed at 27˚C. Sub-cultured 14-day-old suspension cells were used for protoplast isolation.

Leaf-derived suspension cells of P. yedoensis were filtered on 80 μm sized mesh and digested for 3 hrs by the enzyme combination of 1% Cellulase RS, 1% Driselase 20, and 0.25% Pectolyase Y-23 in 0.6 M mannitol, and peduncle-derived suspension cells of P. yedoensis were digested overnight by 1% each of Cellulase RS, Hemicellulase, and Macerozyme R10. Peduncle-derived suspension cells of P. lannesiana were digested by the enzyme combination of 1% each of Cellulase R10, Hemicellulase, Driselase 20, and Macerozyme R10 in 0.6 M mannitol. Each enzyme combination was preliminary selected from 24 combinations tested using six cell wall degrading enzymes in 0.4 - 0.6 M mannitol solution.

Isolated Prunus protoplasts were purified by passing through 63 μm sized mesh and washed three times with mannitol solution by centrifugation at 100 × g for 5 min, and used for protoplast culture directly. Isolated suspension cell protoplasts of P. yedoensis originating from leaves were also stored at -70˚C after precipitation in a 1 mL glass tube for ABA measurement as reported [15] [16] .

Lettuce cotyledon protoplasts were isolated for 20 to 24 hours in 0.6 M or 0.8 M mannitol solution containing 1% each of Cellulase RS and Macerozyme R10 and purified with mannitol solution by centrifugation at 100 × g for 5 min as reported [6] .

Protoplast co-culture method was performed as preciously described [6] [7] using cotyledon protoplasts of recipient lettuce. Protoplasts of different densities were put in a 50 μL medium per well in a 96-well culture plate. Basal medium was MS containing 3% sucrose, 1 μM of 2,4-dichlorophenoxyacetic acid (2,4-D) and 0.1 μM of BA and 0.6 M mannitol. The non-spherically enlarged or divided lettuce protoplasts were counted under an inverted microscope (Olympus CK40) after 4 to 12 days of co-culture in the dark at 28˚C in a humid condition (CO₂-incubator without the supply of CO₂, APC-30DR, ASTEC Co. Ltd.). The percentage of control without Prunus protoplast was calculated. The % values were averaged with standard errors (SEs) at different lettuce protoplast densities (5 × 10³/mL - 100 × 10³/mL).

Coumarin was dissolved in ethanol and a small volume (less than 2.5 μL) was put in the medium after sterilization with a filter as reported [8] . Basal medium was the same as that used for the co-culture except for 0.8 M mannitol. Non-spherically enlarged or divided lettuce protoplasts were counted under an inverted microscope after 6 days of culture. Percentage of control without coumarin was calculated at different lettuce protoplast densities (5 × 10³/mL - 100 × 10³/mL), and the average % values with SEs were obtained.

ABA was dissolved in ethanol as described for coumarin. Protoplasts of leaf- origin suspension cells of P. yedoensis was cultured in the same way as the co-culture. After 4 or 12 days of culture, the numbers of non-spherically enlarged and divided protoplasts were counted. The percentage of control with neither ABA nor coumarin was calculated at each Prunus protoplast density. The % values were averaged with SEs at different Prunus protoplast densities (5 × 10³/mL - 25 × 10³/mL).

Procedures were the same as described previously [15] [16] . Briefly, protoplasts of leaf-origin suspension cells of P. yedoensis were extracted with 80% methanol in a 1-mL glass tube. After evaporation to dryness, a small scale partition was performed between water (pH 2.5 with HCl) and methylene chloride. Purification by silica gel TLC was performed. The ABA fraction was eluted and assayed by Enzyme-Linked Immuno Sorbent Assay (ELISA) using the ABA-ELISA-Kit (Sigma) according to the procedure of the manufacturer, though the ABA standard solution was made separately.

Leaf-origin suspension cells of P. yedoensis were filtered on 80 μm sized mesh and 30.64 g fresh weight of them were extracted by methanol and evaporated to dryness (0.997 g). Coumarin standard and methanol extracts were TMS-de- rivatized with N,O-bis (trimethylsilyl) trifluoroacetamide, and were analyzed by a gas chromatograph-mass spectrometer (TRACE DSQ GC-MS, Thermo Scientific). Coumarin in the methanol extracts were quantified with a calibration curve made by using a coumarin standard. The conditions for GC-MS analysis were as follows: ion source temp., 250˚C; inlet temp., 250˚C; m/z range, 60 - 650; column, DB-1 (15 m × 0.25 mm I.D., film thickness 0.25 µm (Agilent J&W)); carrier gas, He; flow rate, 1.5 mL/min; column temp. program, 100˚C (1 min hold) → 5˚C /min → 250˚C (15 min hold).

Effects of dried leaves of Prunus species on the growth of lettuce seedlings were tested using the sandwich method. As shown in Table 1, highly inhibitory effects of 50 mg of leaves of both P. yedoensis and P. lannesiana, which was less than 3% growth of control (more than 97% inhibition), were obtained on both hypocotyl and root growth of lettuce. Ten mg of leaves were similarly very inhibitory. Sandwich method using 10 and 50 mg dry weight of leaves were first developed using leaf leachates [2] [18] , which was based on the amount of yearly leachates of broad-leaved deciduous trees is about 3 tons/ha, corresponding to 30 mg dry weight/10 cm²/10 mL [3] . Compared with the previous data of many plant species using the sandwich method [2] [18] , Prunus in this report is in the group of very strong allelopathic activities, though young leaves of Prunus were used in this report. Stronger activity was also obtained in the younger leaves of Mucuna gigantea [9] . With the sandwich method, bacterial or fungal contamination may occur during the 3 days of incubation. However, the results obtained with leaves of P. lannesiana sterilized with NaClO aqeous solution before drying the leaves was not different from those obtained without the treatment (Table 1). Such sterilization treatment can be used to exclude the possible effect of fungal or bacterial contamination in the sandwich method.

Compared with the inhibitory effects of leaves of several plant species investigated using the sandwich method (Derris indica [8] ; Mucuna gigantea, Leucaena

^aLeaves were sterilized before drying. Data were described as % of control without Prunus leaves, and averaged with SDs.

leucocephala [9] , Bamboo species [10] ), more than 90% inhibition by 50 mg leaves is very strong. Especially, no growth of lettuce root obtained with leaves of P. yedoensis was the strongest of all. Leaves of an invader plant, Leucaena leucocephala, inhibited the root growth by 90% in the sandwich method [9] .

In this report, suspension cultures of leaves and peduncles of Prunus species were induced and sub-cultured well using naphthaleneacetic acid and a cytokinin, BA, instead of N-(2-chloro-4-pyridyl)-N’-phenylurea used in the previous report on petiole and peduncle of P. yedoensis [16] . Selected enzyme combinations containing Macerozyme R10 for peduncles-derived suspension cells in the present study, could not be employed in the previous report, in which the 0.5% Pectolyase Y23 was used [16] . Therefore, selection method of 24-combinations of cell wall degrading enzymes in different osmotic conditions was applicable in all materials. In this paper, all protoplast cultures were performed at 28˚C. This temperature was the same as that for bamboo species and co-culture with lettuce [10] . The temperature employed for the protoplast cultures of several mangrove tree species and for the co-culture with lettuce was 30˚C [7] . The same range of temperatures and osmotic conditions (0.4 - 0.8 M mannitol or sorbitol) were also selected for both protoplast co-culture and test of allelochemicals and related metabolites. In lower osmotic conditions, the lettuce protoplasts grew vigorously in short days [8] [9] [11] .

Figure 1 shows the results obtained by the protoplast co-culture method using protoplasts of Prunus and lettuce. The growth of lettuce protoplasts was inhibited depending on the protoplast density of all three Prunus suspension cells. Protoplasts from peduncle-derived suspension cells of P. yedoensis and P. lannesiana were also inhibitory but less inhibitory than protoplasts of P. yedoensis suspension cells derived from leaves. Addition of 2 × 10⁴/mL of protoplasts of leaf-origin P. yedoensis almost totally inhibited the lettuce protoplast growth,

which was the strongest of the three kinds of cells. This value is similar to that of Sasa kurilensis, which showed the strongest inhibition among the four bamboo species [10] , and is between the Derris indica [8] and the invader plant, Leucaena leucocephala protoplasts, which totally inhibited the lettuce protoplast growth at 10⁴/mL [9] . Protoplasts of peduncles-origin P. yedoensis and P. lannesiana were less inhibitory than leaf-origin P. yedoensis. However, the nearly total inhibition at 10⁵/mL indicated very strong activity.

Recently, the allelopathic activity measured by the sandwich and protoplast co-culture methods has been reported to differ in bamboo species. Some stimulation at low protoplast densities was observed in suspension cell protoplasts of two bamboo species, which showed moderately strong allelopathic activities in the sandwich method [10] .

However, no stimulation was observed even at low protoplast densities in three Prunus cells (Figure 1). Young growing leaves and vigorously growing suspension cells of leaves were used in our study. The same strong inhibitory activities by leaves of P. yedoensis were observed with both sandwich and protoplast co-culture methods.

Prunus protoplasts can grow slowly in the medium employed for co-culture, i.e., MS basal medium containing 2,4-D and BA, and colony developed at lower concentrations than 10 μM of plant hormones after one month of culture [20] . Such colony development was also observed in the co-culture with bamboo protoplasts [10] . However, in early culture, only the growth of lettuce protoplasts can be well distinguished under an inverted microscope.

Coumarin was studied as a strong allelochemical of Gliricidia sepium with lettuce as a recipient plant [14] . Recently, allelopathic effects of coumarin on adult plants of Arabidopsis thaliana was investigated [21] . However, its role as an allelochemical in Prunus species has not been well investigated. Though allelopathic activity of coumarin has not been discussed in P. yedoensis, content of coumarin in young leaves of P. yedoensis was reported to be high [13] .

Effects of coumarin was investigated on the growth of lettuce protoplasts (Figure 2). Inhibition at higher than 10 μM was clearly observed along the concentrations of coumarin. One mM of coumarin totally inhibited the non-spherical enlargement nor divisions of lettuce protoplasts. Such inhibition is stronger than those of another allelochemical, caffeine, on lettuce protoplasts growth [12] .

Lettuce protoplasts can develop colonies without allelochemicals after 20 days of culture (Figure 3(a)). Figure 3(b) shows the spherical protoplasts of lettuce by addition of 1 mM coumarin. Such a spherical enlargement without yellow color was also observed indicating the strong inhibitory effect of nicotinic acid, while an alkaloid trigonelline, synthesized from it, had no inhibitory effect in the same lettuce protoplast culture system [11] . Lettuce protoplasts can develop yellow colonies (Figure 3(a)) after cell wall formation and divisions. Such yellow substance of lettuce was a carotenoid and applicable for quantitative bioassay of allelopathy with digital image analysis [10] [22] .

Though the inhibitory effects of ABA on the growth of plants cells and protoplasts are well known [15] [16] [17] , a different phenomenon was reported in the lettuce protoplast culture. Stimulation of growth was observed by ABA in the range of 0.1 to 10 μM, while antagonistic plant growth regulator, GA₃ inhibited the lettuce protoplast growth [6] . Such stimulation effect of ABA was also observed in the protoplast cultures of highly salt-tolerant mangrove tree species [23] [24] . However, the role of ABA on protoplast cultures as an allelochemical in Prunus species has not been studied.

Figure 4 shows the effects of coumarin and ABA on the growth of protoplasts isolated from suspension cells of P. yedoensis originating from leaves after 4 days of culture.

Coumarin was not inhibitory at up to 100 μM, but showed strong inhibition at the highest concentration, 1 mM. While ABA showed strong inhibition at even 0.1 μM. Compared with the strong inhibitory effects on the growth of lettuce protoplasts (Figure 2), coumarin had little inhibitory effect on P. yedoensis protoplasts. Similar differences between coumarin and ABA were obtained after 12 days of culture when numbers of colonies composed of 4 cells or more were counted (data not shown). In addition, a highly inhibitory effect of ABA on P. yedoensis was inverse of the stimulatory effect on lettuce protoplast reported [6] .

Non-spherical protoplasts observed without allelochemicals, which is the sign of cell wall formation before cell divisions (Figure 5(a)), were totally inhibited at 10 μM of ABA (Figure 4, Figure 5(b)). Red spherical protoplasts of P. yedoensis were prominent at 10 μM of ABA (Figure 5(b)).

Exogenously supplied ABA strongly inhibited the growth of P. yedoensis protoplasts at low concentrations, 0.1 - 10 μM (Figure 4), but stimulated the growth of lettuce protoplasts [6] . The difference in ABA content in P. yedoensis might cause such different effects of ABA in protoplast co-culture. ABA content of protoplasts has been reported for P. yedoensis suspension cells derived from peduncle and petiole, but not from leaves, using small scale extraction and purification steps and monoclonal antibody measurement method [16] . We examined the ABA content in leaf-derived suspension cells using the same protoplasts method after partition purification steps and ELISA. Though ELISA is possible with protoplast samples [15] . The content of ABA in protoplasts of leaf-origin suspension cells of P. yedoensis was 34.3 pmol/10⁷ protoplasts, 2.9 pmol/mg dry weight. These values are similar to those of petiole-derived

suspension cells measured previously (43.9, 5.0 respectively) and higher than peduncle-derived suspension cells [16] . As the diameter of P. yedoensis protoplasts is 50 μm, volume of a protoplast is 39.3 pL, the calculated concentration of ABA was 0.09 μM in a P. yedoensis protoplast. This value is inhibitory on P. yedoensis protoplasts themselves (Figure 4) but stimulatory on lettuce protoplasts [6] . Addition of 2 × 10⁴/mL P. yedoensis protoplasts totally inhibited the lettuce protoplast growth (Figure 1). Even when all of the endogenous ABA was eluted from the P. yedoensis protoplasts, this value is much lower than expected for inhibition of growth as an allelochemical. ABA is not likely the allelochemical of P. yedoensis. However, rapid increase of ABA content during co-culture and the secondary effect of endogenous ABA in P. yedoensis might be considered.

As coumarin at low concentrations does not affect P. yedoensis, and it inhibits the growth of protoplasts and seedlings of the recipient lettuce, coumarin is a candidate allelochemical of P. yedoensis. By using GC-MS analysis, the coumarin content of 3.26 μg/g fresh weight was obtained in the suspension cells of P. yedoensis originating from leaves. This value corresponds to 22 μM, at which concentration did not inhibit the growth of P. yedoensis protoplasts (Figure 4), and inhibited strongly the growth of lettuce (Figure 2) when exogenously supplied in the media. In the protoplast co-culture among P. yedoensis and lettuce, 2 × 10⁴/mL of P. yedoensis (1000 protoplasts per 50 μL medium) totally inhibited the growth of lettuce protoplasts (Figure 1). As suspension cells contain cell wall portions, the coumarin content in the cytoplasm of cells might be higher than that calculated. Coumarin is more likely the allelochemical of P. yedoensis protoplasts in co-culture with lettuce protoplasts.