STUDIES ON THE NATURAL HYDROPHOBIC BINDERS OF FLABELLIFERINS AND THEIR EFFECT ON SOME BIOACTMTIES

The fruit pulp and flour of Palmyrah (Borassus MATERIALS AND METHODS Flabellifer L.) contain flabelliferins (steroidal saponins) which are found naturally bound to hydrophobic molecules which palmyrah fruit pulp (PFP): Pulp from fruits from g v e them a blue fluorescence in ultra violet (W) light. These hydrophobic molecules have been isolated and have been different was extracted with ~ found to be carotenoids, most probably phytofluene and water (1: 2)e1 phytoene. The presence of these binders modulate the bioactivity of flabelliferins. Tha t is, they lower ATPa"' SeDaration of Flabelliferins:Flabelliferins were inhibition activity of flabelliferin-I1 (F-11) and increase the isolated by the methods described previouslyl and anti-microbial activity of flabelliferin-B (F-B) against Escherichia coli (ATCC 25922). the binder was removed by eluting in methano1:n


INTRODUCTION
Palmyrah (Borassus flabellifer L.) fruit pulp and flour contain steroidal saponins, flabelliferins'.These flabelliferins gave a blue fluorescence under the W light and it was due to another compound binding the p-sitosterol moiety as the latter is not f l u o r e s ~e n t .~ The binder can be removed by a toluene/methanol gradient in Medium Pressure Liquid Chromatography2 (MPLC) and by a gradient of methano1:n-propanol (1: 1) mixture in a chr~matotron.~Flabelliferins are known to be bioactive.For example,flabelliferin-I1 (F-II), a tetraglycoside reduces weight gain2 and glucose uptake4 in mice.Flabelliferin-B (F-B), a triglycoside, inhibits yeast and a range of b a ~t e r i a .~ Therefore, t h i s study was carried out to investigate the nature of the binder and to Instruments for analysis: The instruments used were MPLC6 with a solvent .gradient2,chromatotron4 for separating free and bound carotenoids and flabelliferins.UV absorbance was measured by a Shimadzu automatic reading double beam spectrophotometer model W 1601.
Densitometry was conducted on a Shimadzu CS 9301 PC which is a n UV N I S variable dual wavelength flying spot scanning densitometer together with a fluorescent attachment.
High-Pressure Liquid Chromatography (HPLC) analysis was carried out using Waters HPLC system.This was used with Waters symmetry C,,, 5 pm, 3.9~150 mm column, 515pump system and a 2487 dual absorbance detector.
The optical density of the microbial studies was measured by using the Thermo spectronic spectrophotometer (spectronic educator) , used as a nephliometer.determine its effect on binding F-I1 and F-B and Preparation intestinal mucosal samples: Mice their respective bioactivities.
were anesthetized using diethyl ether.
Immediately after, the jejunum of each mouse was excised and was washed with cold saline (50 ml).
Each jejunum was cut into 5 cm segments and were placed in cold saline.Each of the segments were then cut open, the mucosal surfaces were blotted by pressing between two sheets of filter paper and were then scraped using a glass slide.The scrapings were collected into 2.5 ml cold sucrose-EDTA medium.T h e mixture was homogenized and its volume was adjusted to 50 ml with more of the sucrose-EDTA medium.This medium contained intestinal Na+/K+ ATPase activity .7 Efect of bitter F-11 on the intestinal Na+ activity8: Samples containing NaL/K+ ATPase activity were prepared as described above.Test reaction mixture (1.0 ml) was incubated with ' MgC1, (5 mM), NaC1,(100 mM), KC1 ( l G mM), ATP (3 mM), Tris-HC1 (pH 7.4) and F-I1 a t doses of 0.1 mg, 0.2 mg and 0.3 mg and 0,15 mg, 0.3 mg and 0.6 mg for F-I1 (without binder) with binder respectively.The control reaction mixture contained the same reactants but with no F-11.
The resulting inorganic phosphate was converted to phosphomolybdate with ammonium molybdate reagent (1 ml, 50 gL).This was reduced with aminonapthosulphonic (ANS) acid (0.5 ml, 2 g L ) to produce a blue colour (molybdous acid), which was measured a t a il of 680nm.Inorganic phosphate concentration was calculated from a standard curve (r2 = 0.9896) and an IC,, value was determined.
Microbial studies: Escherichia coli (E.coli) -ATCC 25922 was obtained from the Medical Research Institute (MRI) of Sri Lanka and was maintained on agar slopes containing 1.5% agar, 1% glucose, 0.4% peptone and 0.2 % beef extract, by subculturing every two weeks.F-B trials in liqdid medium were prepared containing the same nutrient composition as above.Varying quantities of F-B (with and without the UV binder) 5-40 mg, were used.Test a n d control samples were prepared with an equal volume of sterilized water.
The total volume was 5 ml.This was introduced under sterile conditions into sterilized test t u ~e s (with cotton wool stoppers).The test tubes used were specific to the spectrophotometer, which was used for absorbance m e a s u r e m e n t s .The absorbances were measured daily a t a il of 700 nm (scatter peak) using a medium blank.The growth \Nas followed for five days.
The Bauer-Kirby disc methods was also used.
After 24 h, the inhibition zones of the test samples (with 50th equal weight and equal molar amounts of F-B with a n d without t h e h;nder) were measured.

Nature of the binder
Studies using extracts of PFP ~i t n MPLC have shown t h a t a white compound ~i t h blue fluorescence eluted before the yellow carotenoids.I t had a Amax of 331 nm compatible w ~t h that of phytofluene.1°On isolation of thc flabelliferin blnder complex and subjecting it to elution on a chromatotron with methanokn-p~opanol(l:l), the fractions collected contained a blue fluorescent substance.This on dissolving in acetonitri1e:methanol: tetra hydrofuran (58:35:7), a carotenoid HPLC mobile phase which was used as it was known that fluorescent carotenoid were y e s e n t in PFP, gave a yellow pre~ipitate.This which is most likely oxidized-polymerised ~a ~o t e n o i d s , was separated by filtration to yield a colourless sclution with intense blue fluorescence.
0-1 subjecting this to Reverse Phase Highpressure Liquid Chromatography (RP-HPLC), a single peak was obtained prior to the retention time of hydrophobic coloured carotenoids on a C,, reverse phase column.The r.r.t.corresponded to s t a n d a r d phytofluene a n d h a d t h e same absorption spectrum with a Amax a t 331 nm.Other spectroscopic instruments (for example Mass Spectronleter) were not available to confirm identity.
Subjecting crude flabelliferins to reparative TLC yielded a blue mild fluorescent band, which gave a Amax of 280 nm.Flabelliferins bound to this compo~nd gaTre a ilmU of 278 nm, showing t h a t there is some distortion to the molecule on binding.Me? suring hbsorbance with a variable wavelength spectrophotometer and densitometer gave a good absorbance around 280 nm but no peak a t 330 nm.Using a dmsitometer with fluorescent detection with excitation a t 330 nm gave only small peaks (high fluorescence was expected), using a 300-400 nm filter indicating that phytofluene was a minor component.From the strong absorbance a t 280 nm on t h e spectrophotometer, it seems likely that the major component could be the carotenoid phytoene.This is supported by further evidence that a natural hydrophobic binder separated from F-I1 by the chromatotron, gave a Am at 279 nm and not at 331 nm.

Inhibition of ATPa"
The IC,, of F-I1 with and without the carotenoid is given in Table 1.This shows that the presence of the binder leads to inhibitory activity.

Antimicrobial activity
F-B without the UV binder gave an IC,, value of 31 pmol/L in solution.However, with the nonpolar binder F-B was less soluble in water and gave a cloudy solution to which the bacteria adhered to and gave a precipitate within 24 hours.Thus the IC,, of F-B with the binder could not be calculated.At concentrations ofwhich F-I3 showed full inhibition, an agar plate streaking showed no bacterial growth.However, the precipitate with F-Bhinderhacteria, on streaking on an agar plate showed growth.
Therefore, the Bauer-Kirby method had to be used.This gave larger inhibition zones for the F-Bhinder complex compared to pure F-B.This is despite diffusional and solubility limitations.This was so when comparing bo.th equal weight and equi-molar quantities (Table 2).F-I1 was not used in antimicrobial studies as it had been previously shown that its inhibitory properties are much lower than F-B.5 F-I1 0.5 Changes in properties F-I1 and binder 0.9 The binder changes the following physical Note: Tests were conducted in duplicate after plotting the Pi properties of F-B and also all other flabelliferins.released versus t h e concentration i n each case a n d determining the 50%inhibition value.
2. Increased solubility in methanol.Liberation of Pi was measured over 5 min in the presence and in the absence of the binder 3. Increased Rfin TLC using varying amounts (0.15 mg to 0.26 mg) of 4. The TLC spot is observed as disc shaped (not flabelliferin and graphs of Pi formed versus rounded like in the case of pure flabelliferins).concentrations were plotted.F-B was not used for full inhibition studies as it did not show any DISCUSSION inhibition in preliminary studies and this is confirmed by the finding of a previous s t u d 9 that The evidence indicates that, it is very likely that F-B in PFP had no effect on glucose uptake by t h e blue-fluorescent naturally occurring mice.
hydrophobic binder of F-I1 is mainly phytofluene.It is probable that the other 280 nm binder is phytoene from UV-spectrum.loPhytoene is present only in minor quantities in PFP and has not been isolated from flabelliferins.The fact that F-B with binder is fluorescent points to phytofluene as the major binder in F-B.Further Jayaweera and co-workersll have shown by molecular modeling t h a t phytoene and phytofluene, unlike other carotenoids, can form a stable complex with B-sitosterol (the aglycone of flabelliferins) by getting inserted into the cavity between the tail of the steroid and its ring system.This makes the entity more hydrophobic and will explain its solubility and TLC properties.
The lowering of inhibition of ATPme can be explained by two factors.(1) The presence of t h e carotenoid binder distorts t h e active carbohydrate moiety of F-I1 (Uluwaduge and Jayaweera, 2004 unpublished results), making it more difficult to bind to the inhibition site.(2) The insertion of the carotenoids increases the effective radius of the inhibitory molecule11, which also results in the lowering of inhibition of ATPa".On the contrary, the decreased inhibition ofE.coli was not clear as the opposite effect was expected.However, two factors need to be considered.Firstly, saponins inhibit microbes by penetrating the phospholipid bilayer (PLB) and causing a formation of a pore in the membrane from which K+ leaks out.12This contributes to the cause of death of the bacteria.Migration through the PLB will be facilitated by the presence of the non-polar carotenoid binder and this increases t h e inhibition.Secondly, there are reports that 0 glucose linked a-1,4 and a-1,2 to rhamnose is a potent group for microbial inhibition, no matter what the parent aglycone is.13,14,15 This group is present in FB2.It appears feasible that having effectively penetrated the PLB with the aid of the hydrophobic binder, this active carbohydrate moiety could interact with some other yet unknown vital biochemical system causing lethality at low concentrations.

Table 2 : Inhibition zones of F-B with and without the binder
* Results are an average of three readings by the disc method.