Rapid induction by fungal elicitor of the synthesis of cinnamyl-alcohol dehydrogenase, a specific enzyme of lignin synthesis

Transcription

1 Eur. J. Biochem. 169, (1987) c) FEBS 1987 Rapid induction by fungal elicitor of the synthesis of cinnamyl-alcohol dehydrogenase, a specific enzyme of lignin synthesis Claude GRAND'. ', Farid SARNI' and Christopher J. LAMB' Plant Biology Laboratory, Salk Institute for Biological Studies, San Diego, California Centre de Physiologie Vegetale, Universite Paul Sabatier, Toulouse (Received June 9, 1987) - EJB A fivefold increase in the extractable activity of cinnamyl-alcohol dehydrogenase, an enzyme of phenylpropanoid metabolism specific for lignin synthesis, was observed within 10 h of treatment of cell-suspension cultures of bean (Phaseolus vulgaris L.) with a high-molecular-mass elicitor preparation heat-released from mycelial cell walls of the bean pathogen Colletotrichum lindemuthiunum. Elicitor caused a rapid, marked but transient increase in the synthesis of cinnamyl-alcohol dehydrogenase with maximum rates 2-3 h after elicitation, concomitant with the phase of rapid increase in enzyme activity. There is a close correspondence between increased polysomal mrna activity encoding cinnamyl-alcohol dehydrogenase, as measured by incorporation of [35S]methionine into immunoprecipitable enzyme subunits in vitro, and the stimulation of enzyme synthesis in vivo in response to elicitor. This marked increase in polysomal mrna activity represents an increase as a proportion of total cellular mrna activity, indicating that elicitor does not stimulate synthesis of this enzyme by selective recruitment from the total pool of cellular mrna. Elicitor stimulation of cinnamyl-alcohol dehydrogenase activity and enzyme synthesis is more rapid than previously observed for other proteins involved in inducible defense mechanisms, such as enzymes of phytoalexin biosynthesis or the apoproteins of cell-wall hydroxyprolinerich glycoproteins. Lignin is a major plant cell-wall structural polymer elaborated by oxidative polymerization of cinnamyl alcohols during the differentiation of specific cell types such as xylem elements [l]. Lignins are highly resistant to attack by microorganisms and lignified cell walls are an effective barrier to pathogen ingress and spread [2]. Many plants respond to microbial attack by deposition of lignin and other wall-bound phenolic material at the point of attack and this may occur in peripheral tissues, such as epidermis and parenchyma, which in uninfected plants do not contain lignin [3]. Depohion of lignin can also be stimulated by treatment of tissue with elicitor molecules derived from fungal cell walls and culture fluids [4, 51. Stimulation of lignin synthesis is one of a number of inducible defense responses to microbial attack [3]. Other examples include synthesis of phytoalexins, accumulation of cell-wall hydroxyproline-rich glycoproteins and increases in the activity of certain hydrolytic enzymes such as chitinase and glucanase [6-91. Disease resistance is an active process dependent on RNA and protein synthesis [7,9]. Recent studies have shown that fungal elicitor or pathogen attack causes massive changes in the pattern of host RNA synthesis including transcriptional activation of defense genes encoding chitinase, cell-wall hydroxyproline-rich glycoproteins and socalled pathogenesis-related proteins as well as enzymes involved in the synthesis of phenylpropanoid defense products Correspondence to C. Grand, Centre de Physiologie VkgCtale, Universite Paul Sabatier, Unit& Associke aucnrs no. 241,118 Route de Narbonne, F Toulouse-Cedex, France Enzyme. Cinnamyl-alcohol dehydrogenase (NADP') (EC 1.I.I.-). [lo, 131 (and unpublished results). Accumulation of transcripts of these defense genes leads to marked stimulation of the synthesis of the encoded proteins and expression of the corresponding defense responses. Lignin monomers are elaborated from phenylalanine and the first three steps, which lead to the synthesis of 4-coumaroyl-CoA, are common to the synthesis of a wide range of phenylpropanoid products including, in addition to lignin, flavonoid pigments and ultraviolet protectants, furanocoumarin and isoflavonoid phytoalexins as well as hydroxycinnamic acid esters (Fig. 1) [14]. 4-Coumaroyl-CoA is a key intermediate from which branch pathways originate which are specific for the synthesis of different classes of phenylpropanoid natural product. Production of lignin monomers involves two steps catalyzed by cinnamoyl-coa reductase and cinnamyl-alcohol dehydrogenase [14]. These enzymes together with the polymerization catalyzed in the cell wall by peroxidase are specific to the synthesis of lignin and related wall-bound phenolic material. Elicitor treatment of bean (Phuseolus vulgaris L.) cells causes a marked stimulation of the transcription of genes encoding phenylalanine ammonia-lyase, the first enzyme of the common, central pathway of phenylpropanoid biosynthesis and genes encoding chalcone synthase and chalcone isomerase, the first two enzymes of the branch pathway specific to the synthesis of flavonoid pigments and isoflavonoid phytoalexins [I 1, 121 (and unpublished results). Transcriptional activation results in a marked induction of synthesis of these enzymes with maximum rates about 4 h after addition of elicitor, correlated with the phase of rapid increase in enzyme activity and onset of the accumulation of

2 74 Fig. 1. Metabolic role of cinnamyl-alcohol dehydrogenase in lignin biosynthesis. The enzyme catalyzes the conversion of 4-coumaraldehyde to 4-coumaryl alcohol (R, = Rz = H); coniferaldehyde to coniferyl alcohol (R, = H, R2 = OCH3) and sinapaldehyde to sinapyl alcohol (R, = R2 = OCH3), thereby generating lignin precursors phytoalexins and wall-bound phenolic material [15, 161. In the present paper we extend this analysis to examine the regulation of cinnamyl-alcohol dehydrogenase, an enzyme of phenylpropanoid synthesis specific to the production of lignin and related wall-bound phenolic material [I 71. MATERIALS AND METHODS Plant cell cultures Cell-suspension cultures of bean (Phaseolus vulgaris cv. Canadian Wonder) were grown as previously described [18] except that cultures were maintained in total darkness. All experiments were conducted with 7-10-day-old cultures, the medium of which exhibited a conductivity between 2.4 mho and 2.8 mho. After treatment with fungal elicitor, cells were harvested by vacuum filtration, frozen in liquid nitrogen and stored at -70 C. Fungal elicitor preparation The source, maintenance and growth of Colletotrichum lindemuthianum in batch liquid culture were as previously described [19]. An elicitor preparation was obtained from the high-molecular-mass fraction released from isolated mycelial walls by heat treatment [20]. Elicitor was applied to cell cultures to give a final concentration of 60 pg glucose equivalent/ml. Enzyme extraction and assay Cinnamyl-alcohol dehydrogenase activity was assayed in crude cell extracts. Frozen cells (2 g fresh weight) were homogenized at 4 C in 10 mlo.1 M Tris/HCl buffer (ph 7.5) containing 0.5% poly(ethyleneglyco1) (M, 6000) and 15 mm 2-mercaptoethanol using a pestle and mortar. Following filtration through muslin, the extract was centrifuged at xg for 30 min. The supernatant was assayed for cinnamyl-alcohol dehydrogenase activity by a spectrophotometric procedure which measured the formation of coniferaldehyde from coniferyl alcohol [17]. Protein was estimated by the Bio-Rad procedure [21]. Immunotitration of cinnamyl-alcohol dehydrogenase Cells were labeled with [35S]methionine 2.5 h after elicitor treatment and harvested 30 rnin later as previously described [22], the labeling period corresponding to the phase of maximum enzyme synthesis and increase in enzyme activity. After cell homogenization, extracts containing 5.2 pkat radiolabeled cinnamyl-alcohol dehydrogenase and between 0 and 5 pl anti-(poplar cinnamyl-alcohol dehydrogenase) serum were incubated in a final volume of 180 p1 Tris/HCl buffer (0.1 M, ph 7.5) containing 0.5% poly(ethy1ene-glycol) (M, 6000) and 15 mm 2-mercaptoethanol. After addition of 20 p1 1.5 M NaCl, the reaction was incubated for 30 min at 28 C and then 3 h at 0 C under agitation. Immunoprecipitated material was collected by centrifugation for 5 min in an Eppendorf desk-top centrifuge and radioactive subunits of cinnamyl-alcohol dehydrogenase were analyzed by SDS/poIyacrylamide gel electrophoresis [22]. The supernatant was assayed for enzyme activity. Protein synthesis in vivo After labeling of cells with [35S]methionine for 30min as described above, at various times after elicitor treatment, cinnamyl-alcohol dehydrogenase was separated from other labeled proteins in cell extracts by indirect immunoprecipitation with antiserum followed by SDS/polyacrylamide gel electrophoresis. Enzyme subunits were located by fluorography and [35S]methionine incorporation determined as described [22]. The rate of enzyme synthesis is defined as the incorporation of [35S]methionine into enzyme subunits as a percentage of incorporation into total protein [22]. RNA extraction Polysomal RNA was isolated by a modification [23] of the method of Palmiter [24]. Total cellular RNA was isolated from cells homogenized directly in a phenol/o.l M Tris (ph 9.0) emulsion [25]. Further purification of phenol-extracted total cellular RNA was according to the procedure used for polysomal RNA [24]. RNA was assayed spectrophotometrically at 260 nm. Protein synthesis in vitro Isolated polysomal RNA or total cellular RNA was translated in vitro in the presence of [35S]methionine using an mrna-dependent rabbit reticulocyte lysate translation system [26] and incorporation of [35S]methionine into total protein was measured as described [27]. Cinnamyl-alcohol dehydrogenase subunits were separated from other translation products by indirect immunoprecipitation with specific antiserum and protein-a-sepharose followed by SDS/poly-

3 75 A 0 L, I TIME [hl Fig. 2. Elicitor stimulation of cinnamyl-alcohol dehydrogenase enzyme activity. ( ) The kinetics for elicitor stimulation of phenylalanine ammonia-lyase, chalcone synthase and chalcone isomerase activities acrylamide gel electrophoresis [27]. mrna activity is defined as the incorporation of [35S]methionine into immunoprecipitable enzyme subunits as a percentage of incorporation into total protein. RESULTS Elicitor stimulation of enzyme activity Treatment of bean cell-suspension cultures with an elicitor heat-released from mycelial cell walls of C. lindemuthiunum caused a rapid, fivefold increase in the extractable activity of cinnamyl-alcohol dehydrogenase. Enzyme activity reached maximum levels 10 h after addition of the elicitor (Fig. 2). In the absence of elicitor, cinnamyl-alcohol dehydrogenase activity remained at a low level throughout the experiment TIME (h) Fig. 3. Elicitor stimulation of cinnamyl-alcohol dehydrogenase synthesis in vivo. (A) Incorporation of radioactivity into immunoprecipitable enzyme subunits following pulse-labeling with [35S]methionine for 30 min immediately prior to harvest at various times after elicitor treatment. (B) Kinetics for elicitor stimutation of cinnamyl-alcohol dehydrogenase synthesis (-0-). Rates of synthesis at each time point are calculated from the incorporation of radioactivity into immunoprecipitable subunits (A) as a proportion of incorporation of radioactivity into total protein, and are presented as a percentage of the maximum rate of synthesis. The kinetics for elicitor stimulation of the synthesis of phenylalanine ammonia-lyase, chalcone synthase and chalcone isomerase (- ----) are included for comparison Immunoprecipitation of enzyme subunits The anti-(cinnamyl-alcohol dehydrogenase) serum used in this study was raised against the enzyme purified to apparent homogeneity from young stems of Populus euramericana [17]. In view of the difference in species, the immunoprecipitation of cinnamyl-alcohol dehydrogenase from elicitor-treated bean cells was first characterized. Treatment of extracts from elicitor-treated bean cells labeled in vivo with [35S]methionine with antiserum immunoprecipitated a radioactive polypeptide which migrated in SDS/ polyacrylamide gel electrophoresis with an apparent subunit molecular mass of 65 kda (Fig. 3). The identity of this polypeptide as the bean cinnamylalcohol dehydrogenase subunit was confirmed by immunotitration experiments. Aliquots of a crude cell extract from cells labeled in vivo with [35S]methionine were incubated with various amounts of anti-(cinnamyl-alcohol dehydrogenase) serum. The loss of enzyme activity from supernatants was directly correlated with the amount of [35S]methionine present in the immunoprecipitated polypeptide of 65 kda (Fig. 4). Complete inhibition of enzyme activity was obtained with 5 p1 anti-(cinnamyl-alcohol dehydrogenase) serum/pkat enzyme activity. These data indicate that the polypeptide of 65 kda, immunoprecipitated from extracts of elicitor-treated cells by anti-(poplar cinnamyl-alcohol dehydrogenase) serum, represents the cinnamyl-alcohol dehydrogenase subunit of bean. ANTISERUM (i0-3rnlfpkat enzyme) Fig. 4. Imrnunotitration of enzyme activity and immunoprecipitation of [35Sjmethionine-labeled enzyme subunits in extracts of bean cells. Cinnamyl-alcohol dehydrogenase activity remaining in the supernatant (-0-) and radioactivity in the immunoprecipitated enzyme subunits (-O-) following treatment with various amounts of anti-(poplar cinnamyl-alcohol dehydrogenase) serum Elicitor stimulation of cinnamyl-alcohol dehydrogenase synthesis Changes in the rate of enzyme synthesis in elicitor-treated cells were measured by immunoprecipitation of labeled enzyme subunits in extracts from cells exposed to [35S]methionine for 30 min immediately prior to harvest. Elicitor caused a marked, rapid but transient increase in the rate of synthesis of cinnamyl-alcohol dehydrogenase concomitant with the phase of rapid increase in enzyme activity (Fig. 3). The maximum rate of enzyme synthesis was attained 2-3 h after addition of elicitor and there was subsequently a rapid decay to a low basal rate of synthesis. In untreated

4 76 < Z U E... I TIME Ih) Fig. 5. Elicitor stirnulation of translatable mrna activity encoding cinnamyl-alcohol dehydrogenase present in the polysomal RNA fruction. (A) Immunoprecipitable [35S]methionine-labeled subunits synthesized in vitro by translation of polysomal RNA isolated from cells at various times after elicitor treatment. (B) Kinetics for elicitor stimulation of cinnamyl-alcohol dehydrogenase polysomal mrna activity. mrna activities at each time point are calculated from the incorporation of radioactivity into immunoprecipitable subunits (A) as a proportion of incorporation of radioactivity into total protein and are presented as a percentage of the maximum mrna activity attained. The kinetics for elicitor stimulation of cinnamyl alcohol dehydrogenase synthesis in vivo (....) and polysomal activities of mrnas encoding phenylalanine ammonia-lyase, chalcone synthase and chalcone isomerase ( ) are included for comparison control cells the rate of cinnamyl-alcohol dehydrogenase synthesis remained at a low level. Eticitor stimulation of cinnamyl-alcohol dehydrogenase mrna activity Radio-labeled cinnamyl-alcohol dehydrogenase subunits of identical size (65 kda) were isolated by indirect immunoprecipitation using the antiserum in conjunction with protein-a-sepharose, from the products of in vitro translation of mrna isolated from elicitor-treated cells. Changes in cinnamyl-alcohol dehydrogenase mrna activity were measured by immunoprecipitation of [35S]methionine-labeled subunits synthesized in vitro by translation of polysomal RNA or total cellular RNA isolated from cells at various times after elicitor treatment. In both the polysomal fraction and total cellular RNA from control, untreated cells, only very low levels of cinnamyl-alcohol dehydrogenase mrna activity could be detected. Elicitor caused a rapid increase in mrna activity encoding cinnamyl-alcohol dehydrogenase within 30 min, with maximum activity levels 2.5 h and 1.5 h after addition of elicitor in the polysomal RNA fraction and total cellular RNA population respectively (Figs 5,6). In both cases this was followed by a rapid decay to low basal levels. The kinetics for stimulation of cinnamyl-alcohol dehydrogenase mrna activity in the polysomal RNA fraction in vitro are closely correlated with the kinetics for stimulation of enzyme synthesis in vivo. DISCUSSION The present data demonstrate that fungal elicitor stimulates the synthesis and hence activity of cinnamyl-alco- / I a I TIME [hl Fig. 6. Elicitor stimulation of translutable mrna activity encoding cinnamyl-alcohol dehydrogenase in the pool of total cellular mrna. (A) Tmmunoprecipitable [35S]methionine-labeled subunits synthesized in vitro by translation of total cellular RNA isolated from cells at various times after elicitor treatment. (B) Kinetics for elicitor stimulation of cinnamyl-alcohol dehydrogenase mrna activity as a proportion of total cellular mrna activity. mrna activities at each time point are calculated from the incorporation of radioactivity into imrnunoprecipitable subunits (A) as a proportion of incorporation of radioactivity into total protein and are presented as a percentage of the maximum mrna activity attained. The kinetics for stimulation of polysomal mrna activity encoding cinnamyl-alcohol dehydrogenase (....) and the activities of mrnas encoding phenylalanine ammonialyase, chalcone synthase and chalcone isomerase as a fraction of total cellular mrna activity ( ) are included for comparison hol dehydrogenase, an enzyme of phenylpropanoid biosynthesis specific to the production of lignin monomers. This is part of a massive activation of phenylpropanoid biosynthesis which also involves induction of enzymes of the common, central pathway and the flavonoid/isoflavonoid branch pathway [ll, 12, 15, 16, 281. There is a close temporal correlation between the stimulation of cinnamyl-alcohol dehydrogenase synthesis, as measured by in vivo labeling or translation of polysomal RNA in vitro, and the rapid increase in enzyme activity. Hence, elicitor stimulation of enzyme synthesis is a major factor regulating the level of cinnamyl-alcohol dehydrogenase activity. Moreover, the marked stimulation in polysomal mrna activity reflects an increase in the fraction of total cellular mrna activity devoted to synthesis of cinnamyl alcohol dehydrogenase, indicating that elicitor does not stimulate dehydrogenase synthesis by selective recruitment from the total pool of cellular mrna. A particularly striking feature is the very rapid stimulation of enzyme synthesis and activity in response to elicitor. Thus, marked increases in mrna activity were observed within 30 min with maximum levels h after elicitation. In contrast, maximum synthesis of phenylalanine ammonialyase, chalcone synthase and chalcone isomerase is not observed until about 4 h after addition of elicitor (Figs 3, 5) [28] and transcripts encoding cell-wall hydroxyproline-rich glycoproteins do not reach maximum levels until about 12 h after elicitation [29]. Likewise cinnamyl-alcohol dehydrogenase enzyme activity increases at near maximum rates from the onset of the response to elicitor, whereas rapid increases in

5 77 the extractable activities of other enzymes of phenylpropanoid biosynthesis are observed only after a short lag (Fig. 2) [228]. Interestingly, synthesis of chitinase is also very rapidly induced in response to elicitor, with kinetics similar to those observed here for cinnamyl-alcohol dehydrogenase (unpublished results). The rapidity of elicitor stimulation of cinnamyl-alcohol dehydrogenase synthesis and enzyme activity implies that this is not an indirect effect, but rather an early component in the causally related sequence of events between elicitor binding to a putative plant cell receptor and expression of inducible defense responses. The present data taken together with the previously observed effects of elicitor on the overall pattern of RNA synthesis and transcription of other defense genes [lo- 131 suggest that elicitor likewise stimulates cinnamylalcohol dehydrogenase gene transcription. Molecular cloning of cdna sequences encoding cinnamyl alcohol dehydrogenase will allow direct analysis of the mechanisms underlying elicitor stimulation of enzyme synthesis. Different kinetics for stimulation of (a) cinnamyl-alcohol dehydrogenase and chitinase, (b) phenylalanine ammonia-lyase, chalcone synthases and chalcone isomerase and (c) cell-wall hydroxyproline-rich glycoproteins may reflect distinct stimuli or a single stimulus leading to either sequential effects or divergent signal pathways. Moreover, these different temporal patterns for elicitor stimulation in cultured cells may reflect different spatial patterns of induction in intact tissue in response to localized mechanical damage or infection. Fungal elicitor causes significant modifications of the cell wall including accumulation of characteristic hydroxyprolinerich glycoproteins and increases in the amount of esterified phenolic compounds [16]. However, elicitor does not cause the accumulation of lignin in bean cell walls as measured by a spectrophotometric assay after removal of esterified cell wall components by treatment with alkali (unpublished results). Hence, stimulation of cinnamyl-alcohol dehydrogenase may be related to the formation of an extracellular complex containing lignin-like material such as that recently described in soybean cell-suspension cultures [30]. Chitinase, which is induced concomitantly with cinnamylalcohol dehydrogenase, releases elicitor-active oligomers by breakdown of the cell walls of invading fungi [31]. Hence the very rapid elicitor induction of chitinase may represent a signal amplification system leading to further release of elicitor-active molecules in the initial stages of the response. Interestingly the lignin precursors, dehydrodiconiferyl glucosides, have recently been shown to exert regulatory effects in plants cells, exhibiting cytokinin-like activity [32, 331. Hence the very rapid elicitor stimulation of cinnamylalcohol dehydrogenase may likewise be a component in the generation of intercellular signals. We thank Carole L. Cramer for providing mrna samples from elicitor-treated cells. This research was supported by a grant to C. J. L. from the Samuel Roberts Noble Foundation and a grant to C. G. from the J. Aron Foundation. C. G. is a French Government Fellow. REFERENCES 1. Sarkanen, K. V. & Hergert, H. L. (1971) in Lignins: occurrence, formation, structure and reactions (Sarkanen, K. V. & Ludwig, S., eds) pp , Wiley, New York. 2. Ride, J. P. (1983) in Biochemicalplant pathology (Callow, J. A., ed.) pp , Wiley, London. 3. Vance, C. P., Kirk, T. K. & Shenvood, R. T. (1980) Annu. Rev. Phytopathol. 18, Ride, J. P. (1986) Hoppe-Seyler s Z. Physiol. Chem. 367, suppl Grisebach, H. (1981) in The biochemistry of plants. A comprehensive treatise, vol. 7 (Stumpf, P. K. & Conn, E. E., eds) pp , Academic Press, New York. 6. Darvill, A. & Albersheim, P. (1984) Annu. Rev. Plant Physiol. 35, Dixon, R. A., Dey, P. M. & Lamb, C. J. (1983) Adv. Enzymol. Relat. Areas Mol. Biol. 55, Ebel, J. (1986) Annu. Rev. Phytopathol. 24, Sequeira, L. (1983) Annu. Rev. Micrabiol. 37, Chappell, J. & Hahlbrock, K. (1984) Nature (Lond.) 311, Cramer, C. L., Ryder, T. B., Bell, J. N. & Lamb, C. J. (1985) Science (Wash. DC) 227, Lawton, M. A. & Lamb, C. J. (1987) Mol. Cell. Biol. 7, Somssich, I. E., Schmelzer, E., Bollmann, J. & Hahlbrock, K. (1986) Proc. Nut1 Acad. Sci. USA 83, Robertson, B. (1986) Physiol. MoI. Plant Pathol Dixon, R. A. & Lamb, C. J. (1979) Biochim. Biophys. Acta 586, Bolwell, G. P., Robins, M. P. & Dixon, R. A. (1985) Eur. J. Biochem. 148, Sarni, F., Grand, C. & Boudet, A. M. (1984) Eur. J. Biochem. 139, Dixon, R. A. & Bendall, D. S. (1978) Physiol. Plant Pathol. 13, Bailey, J. A. & Deverall, B. J. (1971) Physiol. Plant Pathol. I, Anderson-Prouty, A. J. & Albersheim, P. (1975) Plant Physiol. 56, Bradford, M. M. (1976) Anal. Biochem. 72, Lawton, M. A,, Dixon, R. A., Hahlbrock, K. & Lamb, C. J. (1983) Eur. J. Biochem. 129, Schroder, J., Betz, B. & Hahlbrock, K. (1976) Eur. J. Biochem. 67, Palmiter, R. D. (1974) Biochemistry 13, Haffner, M. H., Chin, M. B. &Lane, B. G. (1978) Can. J. Biochem. 56, Pelham, H. R. B. & Jackson, R. J. (1976) Eur. J. Biochern. 67, Lawton, M. A., Dixon, R. A., Halbrock, K. & Lamb, C. J. (1983) Eur. J. Biochem. 130, Cramer, C. L., Bell, J. N., Ryder, T. B., Bailey, J. A,, Schuch, W., Bolwell, G. P., Robbins, M. P., Dixon, R. A. & Lamb, C. J. (1985) EMBO J. 4, Showalter, A. M., Bell, J. N., Cramer, C. L., Bailey, J. A,, Varner, J. E. & Lamb, C. J. (1985) Proc. Natl Acad. Sci. USA 82, Farmer, E. E. (1985) Plant Physiol. 78, Hadwiger, L. A. & Beckman, J. M. (1980) Plant Physiol. 66, Lynn, D. G., Chen, R. H., Manning, K. S. &Wood, H. N. (1987) Proc. Natl Acad. Sci. USA 84, Binns, A. N., Chen, R. H., Wood, H. N. & Lynn, D. G. (1987) Proc. Natl Acad. Sci. USA 84,