THE OF BIOLOGICAL Vol. No. Issue November 5, pp. …THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 264, No. 31,

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 264, No. 31, Issue of November 5, pp. 18608-18617,1989 Printed in U. S. A.

Rubber Elongation Factor from Hevea brasiliensis IDENTIFICATION, CHARACTERIZATION, AND ROLE IN RUBBER BIOSYNTHESIS*

(Received for publication, June 2, 1989)

Mark S. Dennis and David R. Light$ From the Department of Medicinal and Biomoleculur Chemistry, Genentech, Inc., South San Francisco, California 94080

The presence of a protein, rubber elongation factor (REF), which is tightly bound to serum-free rubber particles purified from Hevea brasiliensis latex, is necessary for prenyltransferases from a number of sources to add multiple cis-isoprene units to rubber molecules. These prenyltransferases show normal farnesyl pyrophosphate synthase activity (two trans additions of isopentenyl pyrophosphate to dimethylallyl pyrophosphate) in the absence of REF bound to rubber particles. REF bound to rubber molecules can be highly purified from all other proteins in whole latex by treatment of rubber particles with low concentrations of detergent. Treatment of rubber particles with trypsin which hy- drolyzes bound REF, removal of REF with high concentrations of various detergents, or treatment of whole latex with polyclonal antibodies specific for REF all prevent prenyltransferase from adding [‘“C]isopentenyl pyrophosphate to rubber molecules. However, we have not been successful using detergent-solubilized REF in the reconstitution of in vitro rubber biosynthesis with either REF-depleted rubber particles or allylic pyrophosphate primers.

REF has a molecular mass of 14,600 Da and is associated specifically with rubber particles in whole latex. It makes up between 10-60% of the total protein in whole latex but is absent in C-serum, the supernatant fluid obtained when rubber particles are removed by centrifugation. The amount of REF in whole latex is proportional to the rubber content. Based on a number average molecular mass of 500,000 Da for rubber and the content of rubber and REF in whole latex or serum- free rubber particles, the stoichiometry of REF molecules to rubber molecules is 1:l in both cases. There is sufficient REF to form a monomolecular protein layer coating large rubber particles (700-1,000 nm).

In the electron microscope, serum-free rubber particle preparations contain particles with diameters from 800 to as small as 10 nm. In the presence of 1% sodium dodecyl sulfate no particles smaller than 100 nm are observed. We suggest that the smaller particles may be mainly composed of REF molecules.

In the previous two papers (3,4), we reported the purification and characterization of a prenyltransferase from Heuea brasiliensis that makes multiple additions of isopentenyl py-

*This paper is the third in a series on the biochemistry of cis- polyisoprene rubber synthesis. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

4 To whom correspondence should be addressed.

rophosphate (IPP)’ with cis stereochemistry to serum-free rubber particles. Surprisingly, prenyltransferases from three different sources (1-3) which make t,t-FPP from IPP and DMAPP were shown to add IPP to serum-free rubber particles to make cis-polyisoprene in a manner similar to the Hevea prenyltransferase (4). All three prenyltransferases displayed a different activity and stereospecificity of addition depending on whether the soluble, low molecular weight allylic pyrophosphate, DMAPP or insoluble, rubber particle associated rubber molecules were the allylic substrate. Therefore, we studied the properties of serum-free rubber particles in order to understand the stereochemical switch and the termination override.

Many carrier proteins or binding proteins are known in biosynthetic and biological transport systems. Such proteins bind to large hydrophobic biomolecules, often isoprenoids, with low capacity and high affinity. For the sake of simplicity these proteins can be divided into two groups according to their function: 1) those with a role of facilitating the transfer of the hydrophobic molecule across cell membranes or between tissues and 2) those that directly participate in the enzymatic bioconversion of the molecules they bind. The first group includes retinol-binding protein (21 kDa), which trans- ports vitamin A from the liver to retinol-dependent tissues after secretion into the plasma (5,6) and transcortin (52 kDa) which carries cortisol and corticosterone in the blood (7).

Members of the second group accelerate the activities of the enzymes that operate on the substrates they bind. Some of the better characterized members include: sterol carrier protein 2 (13 kDa) which participates in a number of transfer and enzymatic conversion steps of cholesterol and lanosterol in the liver and steroidogenic tissue (8) and sterol carrier protein or fatty acid binding protein (12 kDa) which accounts for 2-8% of the cytosolic protein in the liver, binds the majority of long chain fatty acids in the cytosol, and participates in sterol and lipid transport and metabolism (9-14). In addition, the activity of the liver enzyme, 4-hydroxybenzoate:polyprenyltransferase, which catalyzes the prenylation of 4-hydroxybenzoate by all-trans-nonaprenyl pyrophosphate during the biosynthesis of ubiquinone, is stimulated by a soluble polyprenyl pyrophosphate carrier protein (15, 16). Thus various carrier or binding proteins share a functional

The abbreviations used are: IPP, isopentenyl pyrophosphate; REF, rubber elongation factor; UTL, untapped tree latex or latex tapped from uncultivated trees which are not tapped on a regular basis; CTL, continuously tapped latex or latex from trees which have been tapped every 3 days for at least 3 months before samples are collected; FPP, farnesyl pyrophosphate; t,t-FPP, all-trans-FPP; DMAPP, dimethylallyl pyrophosphate; NPP, neryl pyrophosphate; GPP, geranyl pyrophosphate; CSsPP, undecaprenyl pyrophosphate (8 cis, 2 trans); SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; STI, soybean trypsin inhibitor; ELISA, enzyme- linked immunosorbent assay; M,,, number average molecular mass; HPLC, high performance liquid chromatography.

18608

Rubber Elongation Factor Characterization 18609

similarity by allowing insoluble molecules to move through and interact with components in the aqueous environment.

A number of other multicomponent systems catalyze prenyltransferase activities. While many prenyltransferases are aa dimers (3, 17), two have been characterized which are composed of two nonidentical subunits. It has not yet been determined whether or not these enzymes are simple ap dimers or whether the active complex has a more complicated stoichiometry. A low molecular mass component (20 kDa) is essential for the activity of hexaprenylpyrophosphate synthase from Micrococcus luteus B-P 26. The activity is fractionated into a 20- and a 60-kDa component by hydroxylapa- tite chromatography. In this case the FPP synthase activity in M. luteus B-P 26 can be separated from both proteins that make up the long chain prenyltransferase (18, 19). Similarly, ion exchange chromatography on DE-52 separates heptapren- ylpyrophosphate synthase from Bacillus subtilis into two components, each having a molecular mass of about 30 kDa. This prenyltransferase activity is also distinct from FPP synthase, undecaprenylpyrophosphate synthase and geranylgeranylpy- rophosphate synthase found in the same organism (20). In neither of these cases has either protein component been found to have activity alone.

In the case of rubber biosynthesis it is reasonable to expect a binding protein or carrier protein to play a functional role at the interface between the hydrophobic, rubber particle- bound, high molecular mass (1000 kDa) cis-polyisoprenylpy- rophosphate and the prenyltransferase in the C-serum. Rub- ber particles are known to be coated with a proteinaceous film of unknown function (21). Natural rubber produced from H. brasiliensis contains about 1% (w/v) protein and efforts have been made to remove this residual protein to improve the quality of the rubber product (22, 23). We find that if serum-free rubber particles are carefully prepared they retain a single protein that accounts for 3-4% of the mass of the rubber. In the following paper (25), we begin to address its role in rubber biosynthesis.

MATERIALS AND METHODS

Reagents-Bovine pancreatic trypsin was obtained from Worthing- ton Enzymes; soybean trypsin inhibitor (STI), Nonidet P-40, and n- octyl-8-D-glucopyranoside were obtained from Sigma; all other reagents were described previously (3,4).

SDS-Polyacrylamide Gels-SDS-polyacrylamide gels (15% acryl- amide, 4% stacking gel) were used according to standard methods and stained with Coomassie Blue (Bio-Rad) or silver-stained (24). Samples containing rubber particles were prepared by extraction with 1 or 2% SDS. The floating pellet of extracted rubber particles was removed carefully after centrifugation (5 min in a benchtop microcentrifuge). If necessary to decrease the sample volume of the soluble phase, dissolved proteins were precipitated by 2 volumes of ice-cold acetone, incubated 20 min on ice, pelleted by centrifugation (5 min in a benchtop microcentrifuge), and redissolved in a minimal volume of sample buffer. The following sets of molecular weight markers were used set 1 (Bio-Rad): phosphorylase B, bovine serum albumin, ovalbumin, carbonic anhydrase, soybean trypsin inhibitor, lysozyme; and set 2 (myoglobin and CNBr fragments, Pharmacia LKB Biotech- nology Inc.): myoglobin, fragment 1-131, fragment 56-131, fragment 1-55, fragment 132-153.

Preparation of Latex from Continuously Tapped and Untapped Trees-Rubber trees from a H. brasiliensis grove were selected and tapped every 3 days by Charles Hunter, Granjas Tropicales S. A., San Jose, Costa Rica. This latex was discarded. After 3 months of continuous regular tapping, samples of latex were obtained from these continuously tapped trees (CTL) as well as untapped trees (UTL) from the same grove. For each sample, 700 ml of latex was diluted with 300 ml of 333 mM Tris, pH 7.5, 0.33% NaN3, and then quickly frozen on dry ice and shipped to this laboratory where it was stored at -70 "C until needed.

Purification of REF Bound to Particles-A liter bottle of diluted frozen latex was thawed quickly by breaking it into small chunks and

warming it in a room temperature waterbath. After thawing, the latex was centrifuged at 4,080 X g for 5 min to remove coagulated latex. Triton X-100 was added to 0.05% to reduce coagulation in the latex and also to prevent nonspecific protein association with the rubber particles described below, A 50-ml portion of latex was centrifuged (50,500 x g, 20 min) to separate or cream the rubber particles from the protein in the serum (29). The floating pellet of washed rubber particles (creamed rubber particles) was carefully removed with a pipette and resuspended in 10 ml of 50 mM Tris, pH 8.0, and 0.01% Triton X-100. This 10 ml of resuspended CTL particles was then loaded on a 5 X 50-cm Sepharose S-300 column equilibrated in 50 mM Tris, pH 7.5,0.01% Triton X-100. A sharp peak containing REF bound to serum-free rubber particles was collected in the void volume.

Extraction and Estimation of REF-Serum-free rubber particles in Triton X-100 were extracted with 2% SDS; the sample was clarified by centrifugation to remove visible rubber particles. The supernatant fluid was then precipitated with an equal volume of cold acetone. The protein was redissolved in SDS-PAGE sample buffer and fractionated by 15% SDS-PAGE. Some rubber that still remained in the loaded sample was retained at the top of the gel. The protein was examined in the gel with 1 M KCI, and the gel slice electroeluted inside a SpectroPor dialysis bag (molecular weight cut off 2000) in 100 mM ammonium bicarbonate, pH 8.0, 0.1% SDS. Starting with 50 ml of serum-free rubber particles purified in Triton X-100 (12 mg of rubber/ ml), we recovered between 3-5 mg of pure REF with a yield of 15- 20% (Table IV).

Estimation of the amount of REF per mg of latex particles was achieved by reverse phase HPLC. The protein was extracted from particles that had been quantitated for rubber content (3). The rubber was removed by centrifugation and filtration. The samples were injected on a RP4 column (Synchropak Inc.). REF was eluted in a gradient from 30 to 40% 1-propanol in 0.1% trifluoroacetic acid. REF was estimated by integration of the elution curve and comparison to a known amount of REF as determined by amino acid analysis of purified REF (25). In order to calculate the molar ratio of REF to rubber, the molecular mass of REF (14.6 kDa) and the number average molecular mass (M.) of rubber (500 kDa) were used.

Preparation of anti-REF ZgG-Rabbit antibodies to REF were prepared by injection of 2 ml of complete Freund's adjuvant containing 1.05 mg of SDS-PAGE purified REF (described above); 1 ml was used for an intramuscular injection of 1 ml divided among four subcutaneous sites on the back of the animal. Booster injections using incomplete Freund's adjuvant were made every 3 weeks. During intervening weeks, beginning week 4,25-ml bleeds were obtained and the titer tested by ELISA. After 3 months, a final bleed was made and the serum stored at -20 "C. Immunizations and animal care were performed at E. L. Labs, Soquel, CA.

In order to purify polyclonal anti-REF IgG, total IgG from rabbit serum was precipitated with ammonium sulfate and chromatographed on an immobilized protein A column (Bio-Rad) by published procedures (40). A column specific for anti-REF IgG was made by coupling 3.2 mg of SDS-PAGE-purified REF with 1 g (dry) of 1 mM HCl washed CNBr activated Sepharose 4B (Sigma) in phosphate-buffered saline overnight at 4 "C. Excess sites were blocked with 0.9 M etha- nolamine, pH 8.0, and the resin washed with 50 mM Tris, pH 7.5. Protein A-purified IgG was loaded in the same buffer and cycled through the column 19 times (0.5 ml/min, overnight). Protein A column-purified IgG that did not bind to the REF affinity column was used as control IgG (Fig. 4). Anti-REF IgG was eluted with 2 M KSCN in the same buffer, and both IgG fractions were dialyzed against 50 mM Tris, pH 7.5, and 0.2% NaN3. The column was run several times to purify 33 mg of anti-REF IgG and 18 mg of control IgG. IgG was quantitated from UV/Vis spectra (Ezaoo.lw = 1.36). Anti- REF and control IgG were compared by reducing SDS-PAGE, and both were identical with a single band at 50 kDa and multiple bands around 24 kDa.

UTL and CTL were brought to 2% SDS, disrupted 5 min in a bath Laser Light Scattering Measurement of Particle Size-Samples of

sonicator to break up aggregates, and diluted to a concentration of 0.03 mg of rubber/ml in a solution of 0.2% SDS that had been prefiltered through a 0.22-pm membrane (Nalgene Inc.). Dynamic light scattering measurements were made on a Lazer Particle Sizer, Model 200, Nicomp Instruments by Dr. Michael Mulkerrin, Genen- tech, Inc. Samples were compared to a standard sample of synthetic rubber particles (mean diameter = 91 nm; variance = 0.10).

Electron Microscopy of Rubber Particles-Electron micrographs were taken by Mei Lie Wong and Dr. Robert Stroud at University of California, San Francisco. Serum-free rubber particles were freshly

18610 Rubber Elongation Factor Characterization

prepared on a Sephacryl S-300 column. SDS-washed particles were prepared by diluting particles 10-fold in 1% SDS and 0.1 M Tris, pH 8.0; centrifuging in an airfuge for 10 min; carefully removing the floating pellet and washing a second time in 1% SDS buffer by the same method. Drops of samples were placed on coated grids for 60 s; washed three times with phosphate-buffered saline; washed two times with deionized H20; and negatively stained with filtered uranyl acetate for 5 s. They were examined in a Phillips EM-400 electron microscope.

RESULTS

Comparison of Latex from Tapped and Untapped Trees- When commercial rubber trees are tapped they demonstrate a “wound response” and replenish their latex a t such a rate that they can continue to be productively tapped every 2-3 days (26). An expected component of this wound response is the induction of the rubber biosynthetic pathway after tapping. The latex that was used as a source of Hevea prenyltransferase was obtained from trees in Costa Rica that were previously untapped (3). A comparison of Costa Rican whole latex from untapped trees (UTL) and continuously tapped trees (CTL), prepared as described under “Materials and Methods,” is shown in Fig. 1 and Table I. When latex was prepared in this manner UTL had at least three times the rubber content of CTL. The usual losses due to coagulation, although not quantitatively determined, were similar by in- spection for both samples. The average particle diameter was measured by dynamic laser light scattering and rubber particles in CTL (245 nm, variance = 0.55) were only about 15% smaller than particles in UTL (285 nm; variance = 0.61).

The two most striking differences between the two latex samples were that UTL had about 5% of the rubber biosynthetic activity of CTL (Table I), and the major 14.6-kDa

1 2 - 92.5 - 66.2 -45.0

” c

-31.0 - W

24kDa- - -21.5

REF- - I) ”14.4 ..”e e : ,

FIG. 1. UTL (lune I ) versus CTL (lune 2) on SDS-PAGE. Samples of latex from Table I matched for total protein content (20 pg each) by loading the equivalent of 6.9 p1 of CTL and 4.4 pl of UTL treated with 2% SDS and 10% 2-mercaptoethanol. Molecular masses of marker proteins (kDa) are shown.

protein band was about three times more prominent in UTL than CTL (Fig. 1). As is shown below, REF, the 14.6-kDa protein, was tightly bound to rubber particles and therefore correlated with the higher rubber content in UTL. This rubber was not as reactive a substrate as the rubber in CTL, however. When normalized to the weight of rubber in the assay the rubber particles in the CTL in the experiment shown in Table I were 20 times more active than UTL, and this high ratio of activities was consistent in four separate comparisons (20 & 9; mean k S.D.). The marked difference between CTL and UTL could not be attributed to protein content, average particle diameter, or prenyltransferase content, as measured by the FPP synthase activity since these parameters were comparable. Further, the addition of pure prenyltransferase to either UTL or CTL did not increase the rate of rubber biosynthesis.

The biosynthesis of rubber is irreversible, since there is no evidence that the plant catabolizes the polymers (27,28), thus the average age of the rubber particles in UTL is expected to be much greater than in CTL obtained from trees continuously challenged to make new rubber. The lower biosynthetic activity in older UTL latex is likely due to the hydrolytic loss of productive allylic pyrophosphate end groups with age. Serum-free rubber particles bound to REF could be success- fully isolated from both types of latex and the differences in the biosynthetic activities were exploited in attempted reconstitution experiments described below.

Effect of Detergents on Rubber Biosynthesis-Serum-free rubber particles prepared by chromatography in a 50 mM Tris, pH 7.5, buffer were used to monitor the purification of the Hevea prenyltransferase (3). When we extracted these particles with 2% SDS and fractionated them by SDS-PAGE, we observed two proteins, a minor 24-kDa band and a major 14.6-kDa band we have called rubber elongation factor or REF (Fig. 2). We also found both proteins associated with rubber particles that are prepared by the method of creaming whole latex (29) described under “Materials and Methods.” In C-serum that was carefully clarified to removed visible rubber particles by centrifugation and filtration only traces of REF remained (Fig. 2, A and B). Thus REF is specifically associated with the rubber particle fraction of whole latex.

When serum-free rubber particles were extracted with high concentrations of detergents to remove the tightly associated proteins and then assayed for the ability to incorporate [“C] IPP, no activity was observed. This result was complicated, however, by the fact that detergents inhibited the rubber biosynthesis assay at concentrations lower than needed to removed the proteins. All the detergents tested had some inhibitory effect on the incorporation of IPP into serum-free rubber particles as can be seen in Fig. 3; nonionic detergents, especially n-octylglucoside, appear to be least inhibitory. Con- centrations of detergents that were sufficient to remove REF

TABLE I Comparison of latex samples from untapped trees (UTL) and trees tapped continuously every 3 days

for 3 months (CTL) Latex samples from the two groups of trees were tapped into buffer, frozen, shipped, thawed, centrifuged, and

assayed in parallel as described previously (3) and under “Materials and Methods.” Because of the marked difference in rubber biosynthetic activity, serial dilutions of the samples were first assayed in order to compare samples with similar total radioactivity incorporated into rubber. All data are averages of duplicate determinations and deviations from the mean.

Source of latex Rubber Protein content content FPP synthase assay Rubber elongation assay

w l m l dpmlpg protein dpmlpg protein dprnlmg rubber Untapped trees (UTL) 95 -e 5 2.9 & 0.1 105 f 14 12 f 0.6 365 f 20 Tapped trees (CTL) 30 Ifr 3 4.6 Ifr 0.4 97 * 19 67 & 10 10,300 f 1,700

Rubber Elongation Factor Characterization

A. B. C. D. 1 2 1 2 3 1 2 3 1 2 3 4 5

3"- - ..

- 6 - e-- - -

E - #; - - - 92.5 -66.2

-45.0 . " -

18611

-31.0 0

- -

- 2.6

FIG. 2. SDS-PAGE analysis (15% gels, Coomassie-stained) of rubber particle-associated proteins. A, lane 1, C-serum centrifuged twice to remove residual rubber particles (30 pg of protein); lane 2, SDS extract of 1.8 mg of serum-free rubber particles prepared without detergent (3). Both samples derived from UTL. B, lane I, whole latex extracted with 2% Triton X-100; lane 2, C-serum (centrifuged twice); lane 3, serum-free rubber particles prepared without detergent (1.8 mg of rubber extracted with 2% Nonidet P-40). All samples derived from UTL. C, lane I, whole latex (UTL), lane 2, serum-free rubber particles prepared with 0.01% Triton X-100; lane 3, serum- free rubber particles prepared without detergent. Serum-free rubber particles were prepared from UTL on matched columns as described in Table 111. Samples were extracted with SDS from 2 mg of rubber particles. D, proteins extracted from large scale preparation of serum-free rubber particles with 0.01% Triton X-100. Rubber particles from a 50-ml sample of CT latex were concentrated about 5-fold by creaming in 0.05% Triton X-100 and deproteinated on an S-300 column (5 X 100 cm, 0.5 ml/min) in 50 mM Tris, pH 8.0, and 0.01% Triton X-100. Lane I, 1.5-mg particles; lane 2,3-mg particles; lane 3,6-mg particles; lane 4,9-mg particles; lane 5, whole latex proteins extracted from 2 mg of rubber. Molecular masses of marker proteins and peptides (kDa) are shown.

1 0 0

.t a0

2 c 60

.- c >

C a)

2 40

X )

0 0.0001 0.001 0.01 0.1 I 0

Detergent Concentration (%I FIG. 3. Concentration dependence of detergent inhibition of

rubber elongation. The rubber elongation assay was performed as previously described (2) with purified Heuea prenyltransferase and serum-free particles isolated on a Sepharose S-300 column without detergent. The following detergents were included in the assay at the indicated concentrations: SDS (A), Nonidet P-40 (A), Triton X-100 (O), Tween 20 (a), n-octylglucoside (0).

completely were completely inhibitory; however, inhibition could be observed a t levels of detergent that did not remove a significant amount of REF. Detergent-extracted rubber particles that were washed with detergent-free buffer to remove bulk detergent were still unreactive in the rubber elongation assay; however, it is almost certain that detergents remained bound to the particles. Thus, the removal of REF by detergents was not a definitive demonstration of the role of REF in rubber elongation because of the additional ability of detergents to inhibit the rubber elongation assay a t low concentrations.

Effect of Trypsin on Rubber Biosynthesis-Since low concentrations of detergents inhibited the rubber elongation assay and high concentrations were necessary to remove REF,

we tried proteolytic removal of the serum-free rubber particle associated protein. Proteolysis with trypsin removed both REF, 14.6 kDa, and the 24-kDa protein from serum-free rubber particles. Serum-free rubber particles preincubated with excess soybean trypsin inhibitor (STI) for 0.5-2 h prior to assay, in the presence or absence of trypsin served as control substrates for Hevea prenyltransferase or avian prenyltransferase. However, if active trypsin was present during the preincubation, the rubber particles were no longer substrates for prenyltransferase-catalyzed ['4C]IPP incorporation into polyisoprene. No incorporation of radioactivity above background nonspecific incorporation observed in buffer controls could be measured (Table 11). These trypsin- treated rubber particles lacked REF when analyzed by SDS- PAGE; whereas, the particles protected by ST1 retained 100% of REF intact. By adjusting the amount of trypsin and the time before ST1 addition it was possible to produce rubber particles with decreasing amounts of intact REF, and in these experiments decreasing rubber biosynthetic activity correlated with the amount of remaining REF when compared to control particles by SDS-PAGE. As a further control against hydrolysis of the allylic pyrophosphates through possible con- tamination by phosphatases, it was determined that the trypsin preparations used above do not catalyze hydrolysis of p - nitrophenylphosphate, a phosphatase substrate, at a rate higher than observed in control buffer. Thus, since ST1 completely protected the rubber particles against trypsin inacti- vation we attribute the inhibition to protein hydrolysis.

Purification of REF Bound to Rubber Particles-Active REF bound to serum-free rubber particles was purified by exclusion chromatography in 0.01% Triton X-100 on Sepharose S-300, a modification of the method used to produce particles for the assay (3). Although this method worked starting from UTL or CTL, CTL is recommended owing to the higher reactivity of the rubber particles it contains. This procedure affected a nominal &fold concentration of the rubber particles, but only 40-60% of the particles could be recovered since a significant

18612 Rubber Elongation Factor Characterization TABLE I1

Effect of trypsinization on rubber biosynthetic activity with deproteinated rubber particles Deproteinated rubber particles were prepared in 50 mM Tris, pH 8.0, and preincubated with trypsin or trypsin

bound with an excess of soybean trypsin inhibitor (STI) under the conditions indicated. During the second stage of preincubation excess ST1 was added to trypsin tubes and trypsin bound to ST1 added to indicated control tubes. No active trypsin was present during the rubber biosynthesis assay. Prenyltransferases, ['4C]IPP, dithiothreitol (10 mM), and MgC12 (1 mM) were added to initiate the assay stage. The concentration of trypsin and ST1 used is indicated in parentheses (mg/ml). All data are averages of duplicate determinations and deviations from the mean.

Additions to assay tubes at indicated stage of experiment ["CIIPP in-

Assay corporation First preincubation Second preincubation (10 min)

d m Experiment 1 (2 h, 37 "C) (4 h)

Particles alone Buffer Hevea enzyme 9600 k 180 Particles alone Buffer No prenyltransferase Particles + trypsin (0.12)

530 f 30 ST1 (0.56) No prenyltransferase

Particles alone 400 f 20

ST1 (0.56) + trypsin Hevea enzyme 8800 k 80

Particles alone ST1 (0.11) + trypsin Hevea enzyme 9800 f 120

Particles + trypsin (0.12) ST1 (0.56) Heuea enzyme 370 -+ 25 Particles + trypsin (0.012) ST1 (0.11) Hevea enzyme 540 f 55

Particles alone Buffer Hevea enzyme 4090 f 30 Particles + ST1 (0.59) + trypsin Buffer Hevea enzyme 4700 f 100

Particles alone Buffer No prenyltransferase 500 f 30 Particles + trypsin (0.059) ST1 (0.59) Hevea enzyme 460 f 45 Particles alone Buffer Avian enzyme 4200 f 50 Particles + trypsin (0.059) ST1 (0.59) Avian enzyme 220 f 20

(0.12)

(0.012)

Experiment 2 (1 h, 30 "C) (2.5 h)

(0.059)

amount of the creamed rubber coagulated into lumps which were discarded.2

The 24-kDa protein was selectively removed when the particles were creamed and then chromatographed in the presence of 0.01% Triton X-100 as determined by SDS-PAGE (Fig. 2, C and D). These particles from which the 24-kDa protein has been removed are at least 85% as reactive as serum-free rubber particles which are not purified in detergent (Fig. 2C and Table 111). Although all of the REF in whole latex was associated with nondetergent washed rubber particles (Fig. 2, A and B ) , the 24-kDa protein was found both in the serum fraction and the rubber fraction (Fig. 2, A and B) . Further, the ratio of the 24-kDa protein to REF varied from preparation to preparation of serum-free rubber particles when Triton X-100 was not included in the column buffer (Fig. 2, A-C; and data not shown). Therefore the 24-kDa protein is not required for serum-free rubber particles to serve as substrates for the Heuea prenyltransferase.

When the Sepharose S-300 column was eluted with Triton X-100 at low flow rates (3) and assessed on SDS-PAGE gels overloaded with extract, no protein bands other than REF were detected (Fig. 2 0 ) . In addition, particles were subjected to amino acid analysis before and after SDS extraction to verify that there were no proteins bound to the particles that could not be extracted by SDS. By amino acid analysis, at least 75% of the protein associated with these particles was SDS-extractable and yielded a composition identical to that found for purified REF (25).

Characterization of REF after Extraction from Serum-free Rubber Particles-REF was denatured and removed from serum-free rubber particles for physical studies by extraction

We tried a number of methods to concentrate rubber particles and all led to major losses due to coagulation. During the preparation of this manuscript Archer and Audley (30) reported an improved method for concentrating rubber particles which consisted of adding dry Bio-Gel P-6 (Bio-Rad) to the suspension of rubber particles followed by recovery with a syringe.

TABLE 111 Rubber biosynthetic activity of prenyltransferase with normal

deproteinated rubber particles and Triton X-100 deproteinated rubber particles that lack the 24-kDa protein

Two matched 2.5 X 100-cm Sepharose S-300 columns were equilibrated in the indicated buffers and run in parallel at 0.4 ml/min. Samples of 2.5 ml of whole latex (UTL; rubber content: 170 mgjml; Fig. 2C, lane 1 ) were deproteinated and 1.0-mg samples of the result- ing rubber particles (Fig. 2C, lanes 2 and 3) were compared as substrates in the rubber biosynthesis assay (3) using 1.2 pg of purified Heuea prenyltransferase. Control incubations of 1.0-mg samples with 25 mM EDTA to inhibit rubber biosynthesis allowed the measurement of 170 f 15 dpm of nonspecific incorporation of radiolabel. All data are averages of duplicate determinations and deviations from the mean.

Deproteination buffer for Rubber elongation assay

S-300 column No added prenyl- Plus purified transferase prenyitransferase

dpm 50 mM Tris, pH 7.5 320 rt 15 2090 f 10 50 mM Tris, pH 7.5 + 300 rt 10 1780 f 45

0.01% Triton X-I00

with 2% SDS, acetone precipitation, followed by SDS-PAGE and electroelution as described under "Materials and Meth- ods." From SDS-PAGE, the molecular mass is 14.6 kDa. When 3 nmol of the purified REF was subjected to NH2- terminal sequence analysis, we obtained no sequence, sug- gesting a blocked NH2 terminus. The sequence of REF has been completely determined and it does not contain cysteine, histidine, tryptophan, or methionine (25).

The amount of REF per mg of latex particles was quantitated by reverse phase HPLC. We determined the amount of REF on the latex particles by comparison to a standard curve of a REF sample quantitated by amino acid analysis. With serum-free rubber particles, UTL, and CTL samples, REF was determined to be present at 1.5, 1.0, and 1.6 pg/Azso, respectively. A gravimetric determination of rubber showed that 40-60 pg of rubber corresponds to 1 Azm depending on


whether deproteinated particles or CTL latex are used (3). This determination was not changed significantly by pretreat- ing the rubber with SDS. Rubber content was converted from milligrams to moles by using a number average molecular mass of 500,000 Da (13). The results shown in Table IV for whole latex and serum-free rubber particles purified in detergent strongly suggest a stoichiometric relationship of one molecule of REF per molecule of cis-polyisoprene in Heuea rubber particles.

During these studies we tried alternative methods for ex- tracting REF. A wide variety of solvents, detergents, salts, and denaturants were tested for their ability to remove REF from serum-free rubber particles eluted from the Sepharose S-300 column. The rubber particles were extracted with a given agent and then the rubber particles were removed by centrifugation and filtration. The extract was then evaluated by SDS-PAGE. REF could be removed by SDS, propanol, and high concentrations of various nonionic detergents in- cluding Tween 20, Triton X-100, Nonidet P-40, and n-octylglucoside (Fig. 2). REF was not readily extracted by high salt, ammonium sulfate, acid, base, or EDTA. 3 M guanidine HC1 was effective in the removal of REF, but 1 M urea was not. Since REF was not easily removed, we concluded that it is tightly associated with the rubber particles.

Reconstitution and initiation with REF-We pursued three general strategies to reconstitute rubber biosynthesis with REF: 1) reversible removal of REF from reactive rubber particles with mild detergents, 2) initiation of new rubber by solubilized REF and low molecular weight primers, and 3) a shift of REF from low activity particles (made from UTL) to unreactive particles that were prepared from highly reactive particles (from CTL) by the removal of REF. Although many solvents and reagents were tested to remove REF, non-ionic detergents, such as n-octylglucoside, were expected to be the mildest method for extraction. The reversibility of extraction of active REF by nonionic detergents was examined by incu- bating whole latex with a concentration of detergent that was previously shown to remove REF. The sample was then diluted to a lower detergent concentration and assayed. A sample at the lower detergent concentration was assayed as a control. The results of several repeated experiments in which various protocols were explored indicated that disruption of the system by nonionic detergents was less than 10% reversible at best. The activity of the particles that were treated with the high detergent concentration and diluted had less activity than control particles in dilute detergent and only slightly higher activity than particles at the high concentration that were not diluted.

We attempted several times to reconstitute rubber biosynthesis with a soluble preparation of REF. From 50 ml of serum-free rubber particles purified in Triton X-100 (A28O I. 300), 3-5 mg of REF was recovered by extraction with 1%

Triton X-100 or 10% n-octylglucoside. We used purified Heuea prenyltransferase to test the detergent-solubilized REF with 10% n-octylglucoside washed rubber particles, trypsin- treated rubber particles, DMAPP, GPP, NPP, t,t-FPP, and C55PP. The detergent-washed rubber particles were further treated with a combination of SM-2 BioBeads and dialysis to removed residual detergent before use. Potential products of allylic pyrophosphate primers were analyzed by molecular exclusion chromatography in toluene or tetrahydrofuran (3). However, we were unable to demonstrate an incorporation of [14C]IPP into any isoprene product greater than 10% over the appropriate controls under any of the conditions tested.

The observation that rubber particles from trees that have been continually tapped are up to 20-fold more reactive that particles from untapped trees led us to attempt the transfer of REF from low reactivity particles (UTL) to high reactivity (CTL) particles stripped of REF. CTL particles were treated with detergents or trypsin to remove the protein. Complete removal was confirmed by SDS-PAGE. These particles did not support the incorporation of [’4C]IPP when mixed with purified rubber transferase or serum. When mixed with UTL, less than 10% increase over the activity of UTL alone was observed. Attempts to affect transfer of REF by sonic oscil- lation or subinhibitory detergent concentrations did not result in high percentage increases in the final observed activity. Since the activity of the UTL was not changed, the conditions used did not inhibit the assay and successful reconstitution should have been detected by these experiments.

Inhibition of Rubber Biosynthesis by anti-REF IgG-We purified polyclonal antibodies to REF in order to determine whether antibody binding to REF in whole latex could inhibit the rubber elongation assay. REF used to raise antibodies was extracted from serum-free rubber particles under denaturing conditions with 2% SDS and isolated by electroelution from a preparative SDS-PAGE gel as described above. To isolate REF specific IgG, this same highly purified REF was coupled to CNBr-activated agarose. With this column, REF-specific IgG was purified from REF-challenged rabbit serum. In a second preparation, REF-specific IgG was removed from protein A-purified IgG from the same source by using the REF affinity column. The fraction of protein A-purified IgG which did not bind the REF affinity column was used as a control. Western analysis of the two IgG samples confirmed that the IgG that bound to the REF affinity column bound to REF, whereas the IgG that did not bind to the column did not bind REF.

The two IgG fractions were tested for their ability to inhibit rubber biosynthesis in whole latex as described under “Ma- terials and Methods” and Fig. 4. The REF-specific IgG inhibited rubber biosynthesis to a much greater extent than control IgG when matched parallel dilutions of the two antibody preparations were compared. At higher concentrations of the

TABLE IV Stoichiometry of REF to rubber

Diluted samples of latex (CTL) and deproteinated rubber particles (without detergent) were assayed gravimet- rically for rubber content (3) and additional samples were extracted with SDS for determination of REF by reverse phase HPLC column. Sample injections were compared to a standard curve constructed from injections of serial dilutions of a standard of purified REF which has been quantitated by amino acid analysis (15). In order to calculate a molar ratio the weight average molecular weight of natural rubber (M, = 1 x IO6) was converted to a number average molecular mass (M, = 5 X lo5) assuming a normal polymer distribution (13). Data are averages of four determinations & S.D.

Sample Rubber Extracted REF Ratio of REF to rubber

mgf ml 26 f 1 0.69 -C 0.07 Whole latex (CTL)

Deproteinated rubber particles 1.9 k 0.1 0.075 +- 0.008 0.039 & 0.006 0.027 k 0.004 0.93 -+ 0.1

1.3 & 0.2

wfw mol/mol


IgG ( m o l )

FIG. 4. Affinity-purified anti-REF IgG inhibition of rubber elongation in whole latex. Polyclonal IgG in immune serum from a rabbit challenged with pure REF was first purified on a protein A affinity column and further fractionated on a REF affinity column into anti-REF IgG (0) and control IgG (O), the latter of which did not bind to the REF affinity column. The IgGs were dialyzed to 0.15 M NaCl, 50 mM Tris, pH 7.5. IgG samples were scanned, matched by their absorbance at 280 nm, and assessed by SDS-PAGE (not shown) as described under “Materials and Methods.” The indicated amounts of IgG were preincubated with 0.1-ml samples of whole latex (CTL; 0.6 mg of rubber containing 14 pg or 1 nmol of REF) for 6 h at 4 “C; after which 0.2 ml of the standard [“CIIPP reaction mixture (without dithiothreitol) was added. Samples were incubated for 4 h and the rubber elongation assay performed as described (1).

control IgG there was some inhibition of the rubber elongation assay (Fig. 4); however, inhibition could also be demonstrated by high concentrations of bovine serum albumin (0.25-1%).

Electron Microscopy of Rubber Particles-Most of the particles observed in electron micrographs of serum-free rubber particles (with Triton X-100) had diameters of 250-500 nm, consistent with the average diameter measured by laser light scattering (Fig. 5A) . However, particles as large as 600-800 nm could be found. Because variable amounts of rubber were lost owing to coagulation during the freezing and thawing associated with the shipping of whole latex, the proportion of larger particles could be even greater in whole latex if large particles are more readily lost to coagulation. Also present in these serum-free rubber particles purified in detergent were numerous particles with a diameter of 100 A or less (arrows, Fig. 5 A ) . These small diameter particles were no longer present when the rubber is washed with 1% SDS (Fig. 5B).

DISCUSSION

We have identified REF, a low molecular weight protein bound to rubber particles from which serum has been removed by molecular exclusion chromatography in buffer containing Triton X-100. Although REF could simply be a binding protein that functions to prevent coagulation of rubber particles in an aqueous serum, we present data that are consistent with it playing a larger role in rubber biosynthesis. Thus, we have named this protein REF or rubber elongation factor.

A number of prenyltransferases make long chain, cis-poly- prenols from IPP and an appropriate allylic pyrophosphate (17,31,32). Before beginning this work, it seemed reasonable to expect that the Hevea prenyltransferase would be one of this type, but would lack ability to terminate the growing chain precisely. In the preceding paper (4), we reported that yeast and avian FPP synthase as well as Hevea prenyltransferase are able to add IPP to elongate cis-polyisoprene. In

addition, Heuea prenyltransferase makes all-trans-FPP from DMAPP and IPP. Thus, Heuea prenyltransferase which is purified as a rubber transferase appears to be a typical FPP synthase. We looked at serum-free rubber particles, the second component of the assay system, for other factors respon- sible for both the switch in stereospecificity of addition of IPP to polyisoprene from trans to cis and for the override of the termination by FPP synthase which is normally seen after two or at most three prenyl transfers.

Preparation of Rubber Particles Bound to REF and Removal of REF by Detergents-Serum-free rubber particles washed with SDS yielded two proteins, 24 and 14.6 kDa, when analyzed by SDS-PAGE. The 24-kDa protein is a major protein in whole latex and in serum (Fig. 2), and the amount associated with the serum-free rubber particles purified in the absence of detergent varied between preparations. The majority of the 24-kDa protein remained in the serum. In contrast, the 14.6-kDa protein, REF, was only associated with rubber particles and was always present in particles that were reactive in the rubber elongation assay. Latex particles with only REF could be made by exclusion chromatography in 0.01% Triton X-100. These particles were still reactive even though the 24-kDa protein had been removed (Fig. 2C and Table 111), thus the 24-kDa protein is not required for serum-free rubber particles to serve as substrate.

When these serum-free rubber particles were washed with SDS or nonionic detergents (>1%), rinsed repeatedly in buffer by centrifugation to remove most of the detergent, diluted and assayed with Hevea prenyltransferase, no rubber biosynthetic activity was observed coincident with the removal of REF. However, since low detergent concentrations were inhibitory in the rubber elongation assay (Fig. 3), we were concerned that residual detergent bound to the rubber particles could account for the observed inhibition.

Removal of REF by Proteolysis and Inhibition by anti-REF ZgG-We chose to remove REF with the serine protease trypsin since we could readily control the activity of the protease with excess soybean trypsin inhibitor. Treatment of rubber particles with trypsin inhibited by ST1 had no effect on their ability to serve as substrate for Hevea or avian prenyltransferases. However, when ST1 was added after a pretreatment with trypsin, rubber particles become totally incapable of incorporating [‘4C]IPP catalyzed by either prenyltransferase (Table 11). The proteolytic removal of REF as monitored by SDS-PAGE was correlated with loss of reactivity and thus strongly suggests that this protein is required for rubber biosynthesis.

Rubber biosynthesis in whole latex could be specifically inhibited by anti-REF IgG purified on a REF affinity column in contrast to the control IgG that was the fraction of protein A affinity-purified IgG that did not bind to the REF affinity column (Fig. 4).

Stoichiometry of REF to Prenyltransferase-In a represent- ative assay in which the concentration of prenyltransferase is saturating, there is 1 mg of serum-free rubber particles and 1 pg of purified prenyltransferase. Thus, the stoichiometry of prenyltransferase to REF is 0.026 nmol of prenyltransferase monomer to 39 pg or 2.7 nmol of REF. As discussedpreviously, we do not know whether the prenyltransferase monomer or dimer binds to REF to form an active elongation complex (3).

We have no reason to postulate an additional low stoichiometry, high affinity, mobile elongation factor at present. However, it could be argued that when anti-REF antibodies inhibit rubber biosynthesis, they bind to the protein coat on the rubber particle and sterically inhibit a second distinct elongation factor at the site of rubber biosynthesis. About


1“‘ I

1”

b ”

FIG. 5. Electron microscopy of serum-free rubber particles prepared in 0.01% Triton X-100 before ( A ) and after ( B ) treatment with 1% SDS. Both samples were-negatively stained with uranyl acetate (see “Materials and Methods”); the magnification is 97,740 and 1,000 A bar is shown. A, large serum-free rubber particles are slightly irregular spheres with diameters ranging from 200 nm up to 700 nm or more. Patches on larger particles are due to incomplete staining at the grid interface. Numerous small particles, some as small as 10 nm, are also present and are indicated by arrows. B, particles washed with 1% SDS are rounded and no particles less than 100 nm diameter remain in the sample.

0.01% of the rubber molecules isolated from latex (UTL) were maximum number of growing ends is less than 1% of the competent for addition of IPP (3); and, even if this number number of rubber molecules. Thus, a second low molecular were to increase by a factor of 100 in CTL or other prepara- mass protein or even residual 24-kDa protein could remain tions with higher activity on a per weight basis (30), the bound to particles deproteinated in the presence of Triton X-


100. In order to be stoichiometric with added prenyltransferase it would only need to be present at less than 1% of the concentration of REF and would escape detection. The low molar ratio of prenyltransferase to REF leaves open the possibility of a protein cofactor present at 0.1% of the level of REF even in the cleanest preparations of serum-free rubber particles purified in detergent. However, the simplest inter- pretation is that the REF-specific IgG is acting directly to inhibit rubber biosynthesis by binding to and inhibiting REF.

Stoichiometry of REF to Rubber-Quantitative estimation of REF and rubber in whole latex and serum-free rubber particles purified in detergent demonstrated that there is one molecule of REF per molecule of rubber (Table IV). Small, low capacity, high affinity binding proteins or carrier proteins of large, hydrophobic biomolecules serve in the transport, storage, and in some cases the enzymatic modification of the ligands they carry. Several examples are discussed in the introduction. The stoichiometric relationship between REF and rubber molecules is consistent with REF coating rubber particles and serving in the synthesis and storage of rubber in latex. At first it seems remarkable to think that each rubber molecule should have its own protein associated with it even though REF at 14,600 Da is much smaller than an average 500,000-Da molecule of rubber. When compared to other binding proteins, the role of REF is not so unusual. For example, liver fatty acid binding protein is a low capacity binding protein that binds 60% of the long chain fatty acids in liver cytosol (9). Since REF represents between 10-60% of the total protein in whole latex, there must be a tremendous increase in the synthesis of this protein during the 2-3 day wound response after the Heuea rubber tree is tapped (26). However, this is no more remarkable than liver FABP which represents up to 8% of the total protein in liver and undergoes daily changes in concentration on the order of 7-10-fold over a 12 h period (13, 14).

In our experience, the problem of rubber particle coagulation exists in whole latex and in serum-free rubber particles or rubber particles that have been washed free of REF and resuspended in buffer. However, we have not yet attempted to quantitate the rate or extent of coagulation and such measurements would be complicated by the detergents used in our procedures. A reliable and sensitive coagulation assay could address the role of REF or other less tightly associated proteins, such as the 24-kDa protein, in the stabilization of the latex emulsion.

A Coating of REF on Rubber Particles-When viewed in the electron microscope, serum-free rubber particles purified in Triton X-100 that still have bound REF and the same particles treated with 1% SDS to remove REF appeared to be fairly similar except that small particles of less than 100-A diameter were missing (Fig. 5). Archer and Audley (30) en- riched for small diameter particles that remain in C-serum after most of the bulk rubber was removed by centrifugation. They concentrated the C-serum by reconstituting freeze-dried serum and compared washed rubber particles from this preparation to washed rubber particles from whole latex (30). Washed rubber particles were prepared by exclusion chromatography in detergent as are the serum-free rubber particles used in our current studies. The particles that remained in the C-serum fraction had a diameter of 50-80 A in the electron microscope (34) and were thus similar in size to the small diameter particles removed by SDS treatment (Fig. 5A, arrows).

The actual rubber content of C-serum is very low. Archer and Augley (30) report that there are 5 mg of rubber/g of freeze-dried solids. If solids are mainly protein (this assump-

tion yields a conservatively high estimate), then C-serum that contains 0.5% protein (3, 27) contains 0.0025% rubber. Com- pared to 33% rubber in whole latex (3, 27), this number is consistent with our assertion that REF is present in whole latex, but not C-serum (Fig. 2, A and B ) . In fact, Western analysis of C-serum proteins (data not shown) did indicate that trace amounts of REF remain in C-serum consistent with the 10‘-fold difference in their rubber contents. Archer and Augley (30) found that these low diameter particles were 75 times more reactive in a rubber biosynthesis assay than particles from whole latex on a weight basis (30).

We model REF containing rubber particles in a monolayer. We assume REF is a spherical globular protein with a molecular mass of 14,600 Da and a hydrated specific volume of 0.93 ml/g which leads to a diameter of 35.1 A. A number average molecular mass of 500,000 Da and a specific volume of 1.1 ml/g are used for rubber (3). We assume hexagonal packing of spheres of REF on the surface of the rubber particle and calculate tkat when a particle diameter of 550 nm is reached the 35.1 A thick monolayer will increase to reach 100% occupancy. Given these assumptions, larger particles are not covered by a simple monolayer of REF but would begin to form a bilayer or other structure? In these calculations the molecular mass of rubber is always held to be 500,000 Da. Based on the observed one to one stoichiometry of REF to rubber, the maximal particle diameters that are coated by a monolayer are consistent with those observed in whole latex (Fig. 5; 29, 35-39). However, we do not infer that a causal relationship exists by which the upper limit to rubber particle diameter is somehow limited by the ability of REF to form a monolayer.

By changing the model slightly, we allow the molecular weight of rubber to vary with particle size. We make the assumption that the coating is a hexagonal close packed monolayer of 100% occupancy for all particle sizes. Again, we retain the observed one to one correspondence of REF to rubber in the model and assume the ends of the rubber molecules will be able to remain complexed with a REF molecule on the surface of the particle. Thus, particles with a diameter of 1100 nm coated by REF have sufficient volume in the rubber core for each molecule to have a molecular mass of 1.05 X lo6 Da. We do not mean to suggest that the molecular weight of rubber is uniformly proportional to particle size but to demonstrate that the one to one stoichiometry of REF to rubber is entirely consistent with the known properties of natural latex.

If we still assume the particles are completely coated by REF at the low end of particle size, the volume of the protein coat begins to be the major portion of the particle. Thus, 1,000-A particles are modeled to have about 20% of the volume in the protein shell which contains 2,800 REF molecules. The same number of rubber molecules in the core would have a molecular mass of 84,000 Da each. Particles with a diameter of 350 A have equal volumes in the rotein shell and the rubber core. Finally, in this model, 100- x particles would have over 95% of the volume in the protein shell composed of 14 molecules of REF surrounding a “rubber” core of 14 isoprene molecules with a molecular mass of 555 Da. This size isoprene molecule contains 8-9 isoprene units and is slightly smaller than the isoprene portion of undecaprenyl pyrophosphate. By way of comparison, the diameter of a single spherical molecule of 500,000 Da rubber is 120 A.

Archer et al. (34) suggest that the small diameter (50-80 A) Other assumptions about the geometry can be used, for example,

the largest particle diameter consistent with a monolayer of densely packed cubes of REF is approximately 800 nm.


highly reactive particles in reconstituted freeze-dried serum may be single rubber particles of approximately 100,000 Da. Based on the stoichiometry of REF to rubber, we suggest a different model for the biosynthesis of rubber in Hevea latex. Small diameter particles which are mainly protein would be expected to vanish when treated with 1% SDS, conditions which we have shown solubilize REF (Fig. 5). A better understanding of the structure of REF and rubber particles must await the careful fractionation and analysis of rubber particles.

Attempts to Dissociate Reversibly or Reconstitute REF-A series of experiments to reconstitute REF partially or fully in a rubber elongation assay is described under “Results.” We attempted to 1) reversibly dissociate REF from rubber particles with detergents, 2) cause REF to move from low reactivity particles to previously high reactivity particles that were stripped of REF by detergents, 3) add detergent-dissolved REF to rubber particles stripped of REF by treatment with protease or concentrated detergents or 4) add dissolved REF to the small molecule allylic pyrophosphates DMAPP, GPP, NPP, t,t-FPP, or C55PP. No evidence of reconstitution greater than 10% of controls was seen. Products were assessed in the rubber elongation assay (3) or by drying down the reaction mixture, dissolving the pellet in toluene or tetrahydrofuran and fractionating by exclusion chromatography (3). Two of the low molecular weight primers used, t,t-FPP, and Cs6PP, have the requisite w-trans,trans linkage detected by 14C NMR in goldenrod cis-polyisoprene (33). This laboratory (3,4) and Archer and Audley (30,34) have demonstrated stimulation of rubber biosynthesis by these low molecular weight primers.

The methods we used to extract REF may have caused it to be irreversibly denatured; the native conformation may require a particular combination of detergent concentration or ionic strength to bind allylic pyrophosphates productively. If REF is associated with a single rubber molecule during its entire lifetime it may be that it associates with a primer as it folds during translation from its mRNA. Irreversible association of native REF is not consistent, however, with the ability of low molecular weight primers to stimulate synthesis of high molecular weight rubber when added to mature rubber particles (30). Phospholipids present on rubber particles may also be required to make an active elongation complex, but in a limited number of tests with phospholipid vesicles we saw no positive effect on reconstitution. Thus, the understanding of the nature of the complex of REF and rubber remains a challenge.

REFERENCES

1. Eberhardt, N. L., and Rilling, H. C. (1975) J . Biol. Chem. 250,

2. Poulter, C. D., and Rilling, H. C. (1981) in Biosynthesis of Isoprenoid Compounds (Spurgen, S. L., and Porter, J. W., eds) Vol. 1, pp. 163-224, John Wiley & Sons, New York

3. Light, D. R., andDennis, M. S. (1989) J. Biol. Chem. 264,18589- 18597

4. Light, D. R., Lazarus, R. A., and Dennis, M. S. (1989) J. Biol. Chem. 264,18598-18607

5. Laurent, B. C., Nilsson, M. H. L., BHvik, C. O., Jones, T. A., Sundelin, J., and Peterson, P. A. (1985) J. Biol. Chem. 260,

863-866

11476-11480

6. Newcomer, M. E., Jones, T. A., Aqvist, J., Sundelin, J., Eriksson, U., Rask, L., and Peterson, P. A. (1984) EMBO J. 3, 1451- 1454

7. Westphal, U. (1983) J. Steroid Biochem. 19, 1-15 8. Pastuszyn, A., Noland, B. J., Bazan, J. F., Fletterick, R. J., and

Scallen, T. J. (1987) J. Biol. Chem. 262, 13219-13227 9. Gordon, J. I., Alpers, D. H., Ockner, R. K., and Strauss, A. W.

(1983) J. Biol. Chem. 258,3356-3363 10. Chan, L., Wei, C-F., Li, W-H., Yang, C-Y., Ratner, P., Pownall,

H., Gotto, A. M., Jr., and Smith, L. C. (1985) J. Biol. Chem. 260,2629-2632

11. Dempsey, M. E. (1985) Methods Enzymot. 111,293-303 12. Heuckeroth, R. O., Birkenmeier, E. H., Levin, M. S., and Gordon,

13. McGuire, D. M., Olson, C. D., Towle, H. C., and Dempsey, M. E.

14. Glatz, J. F. C., Baerwaldt, C. C. F., Vearkamp, J . H., and Kempen,

15. Gupta, A., Paton, B. C., Ranganathan, S., and Rudney, H. (1984)

16. Gupta, A., and Rudney, H. (1985) Methods Enzymol. 110, 327-

17. Muth, J. D., and Allen, C. M. (1984) Arch. Biochem. Biophys.

18. Fuiii. H.. Kovama, T., and O a r a , K. (1985) Methods Enzymol.

J. I. (1987) J. Biol. Chem. 262,9709-9717

(1984) J. Biol. Chem. 259,5368-5371

H. J. M. (1984) J. Biol. Chem. 269,4295-4300

Biochem. Biophys. Res. Commun. 119, 1109-1115

334

230,49-60

i io, i92-ig8 . . - .

19. Fuiii. H.. Kovama. T.. and Oeura. K. (1982) J. Biol. Chem. 257. i46101146i2

, , - , . .

20. Sagami, I., Fujii, H., Koyama, T., and Ogura, K. (1985) Methods Enzymol. 110, 199-205

21. Archei, B. L. (1980) Encycl. Plant Physwl., New Ser. 8,309-327 22. Chin, P. S., Chang, W. P., Lau, C. M., and Pong, K. S. (1974)

Proceedings of Rubber Research Institute of Mahysia Planters’ Conference, pp. 252-262, Rubber Research Institute of Malay- sia, Kuala Lumpur

23. Planter’s Bull. Rubber Res. Inst. Malaysia (1976) 144, 57-59 24. Morrissey, J. H. (1981) Anal. Biochem. 117.30-310 25. Dennis, M. S., Henzel, W. J., Bell, J., Kohr, W., and Light, D. R.

26. Hurley, P. E. (1980) Natural Rubber News August, 3-9 27. Archer, B. L., and Audley, B. L. (1973) in Phytochemistry (Nord,

F. F., and Miller, L. P., eds) Vol. 2, pp. 310-343, Van Nostrand- Reinhold, New York

(1989) J. Bioi. Chem. 264,18618-18626

28. Backhaus, R. A. (1985) Zsr. J . Bot. 34, 285-293 29. Archer, B. L., Audley, B. G., Cockbain, E. G., and McSweeney,

30. Archer, B. L., and Audley, B. L. (1987) Bot. J. Linn. SOC. 94,

31. Takahashi, I., and Ogura, K. (1982) J. Biochem. (Tokyo) 92,

32. Adair, W. L., Jr., Cafmeyer, N., and Keller, R. K. (19&1) J . Bioi.

33. Tanaka, Y., Sato, H., and Kageyu, A. (1983) Rubber Chem.

34. Archer, B. L., Audley, B. G., and Bealing, F. J. (1982) Plast.

35. Ho, C. C., and Ng, W. L. (1979) Colloid & Polym. Sci. 257,406-

36. Lucas, F. F. (1938) I d . Eng. Chem. 30, 146-153 37. Schoon, T. G. F., and van der Bie, G. J. (1955) J. Polym. Sci. 16,

38. Stavely, F. W., Biddison, P. H., Forster, M. J., Dawson, H. G., and Binder, J. L. (1961) Rubber Chem. Technot. 34,423-432

39. McMullen, A. I., and McSweeney, G. P. (1966) Biochem. J. 101, 42-47

40. Tijssen, P. (1985) in Practice and Theory of Enzyme Immunoas- says, Laboratory Techniques in Biochemistry and Molecular Biology (Burdon, R. H., and van Knippenberg, P. H., eds) Vol. 15, pp. 96-108, Elsevier Science Publishing Co., New York

G. P. (1963) Biochem. J. 89,565-574

181-196

1527-1537

Chem. 259,4441-4446

Technol. 56, 299-303

Rubber Int. 7,109-111

412

63-88

Documents

THE OF BIOLOGICAL Vol. No. Issue November 5, pp. …THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 264, No. 31,