Substrate Specificity and Kinetic Studies of Nodulation Protein NodL of Rhizobium leguminosarumt

@ Abstrat 12712 Biohemistry 1995, 34, 12712-12720 Substrate Speifiity and Kineti Studies of Nodulation Protein NodL of Rhizobium leguminosarumt Guido V. Bloemberg,* Ron M. Lagas,$ Steven van Leeuwen,g Gijs A. Van der Marel,$ Jaques H. Van Boom,$ Ben J. J. Lugtenberg,* and Herman P. Spaink*%$ Institute of Moleular Plant Sienes, Leiden University Wassenaarseweg 64, 2333 AL Leiden, The Netherlands, and Gorlaeus Laboratories, Leiden University, Department of Organi Chemistry, Postbox 9502, 2300 RA Leiden, The Netherlands Reeived April 3, 1995; Revised Manusript Reeived July 25, 1995@ ABSTRACT: All lipo-hitin oligosaharides identified from Rhizobium leguminosarum arry an 0-aetyl moiety on C6 of the nonreduing terminal N-aetylgluosamine residue. Previously, we have shown that purified NodL protein, using aetyl-oa as aetyl donor, in vitro aetylates N-aetylgluosamine, hitin oligosaharides, and lipo-hitin oligosaharides. In this paper, the enzymati properties and substrate speifiity of NodL protein were analyzed, using a spetrophotometri assay to quantify NodL transaetylating ativity. NodL funtions optimally under alkaline onditions. Transaetylating ativity has a broad temperature optimum between 28 and 42 "C. NodL protein is stable for at least 15 min up to 48 "C. Gluosamine, hitosan oligosaharides, terminally de-n-aetylated hitin derivatives, and ellopentaose were identified as aetyl-aepting substrates for NodL protein. Quantitative substrate speifiity studies show that hitin derivatives with a free amino group on the nonreduing terminal residue are the preferred substrates of the NodL protein. Our results strongly indiate that the nonreduing terminally de-n-aetylated hitin oligosaharides produed by the NodC and NodB enzymes are the in vivo aetyl-aepting substrates for NodL protein. The symbioti interation between rhizobia1 bateria and leguminous plants results in the formation of nitrogen-fixing root nodules. Host speifiity of nodulation is determined by signal exhange between the plant, and the baterium. In response to flavonoid signal moleules sereted by the plant the rhizobial nod or no1 genes are transribed [for a review, see Shlaman et al. (1992)l. Many of the nod gene produts mediate the synthesis of sereted lipo-hitin oligosaharides (LCOs). Purified LCOs an evoke various responses in the host plant roots, whih are indistinguishable from those observed during several stages of the nodule formation proess [for reviews, see Fisher and Long (1992) and DCnariC and Cullimore (1993)l. The noda, nodb, and nodc gene produts are minimally required for the prodution of LCOs. NodC protein has sequene homology with hitin synthases (Bulawa et al., 1991; Atkinson et al., 1992; Debelle et al., 1992) and is involved in the prodution of p- 1,4-1inked N-aetylgluosamine oligomers (Spaink et al., 1993, 1994; Geremia et al., 1994). The nodb gene produt shows homology with a hitin deaetylase (Kafetzoupoulos et al., 1993), and in vitro experiments have demonstrated that the nodb gene produt is able to remove the N-aetyl moiety from the nonreduing ' G.V.B. was supported by finanial aid from the Netherlands Organization for Chemial Researh; H.P.S. was supported by the Royal Netherlands Aademy of Art and Sienes, a Pionier Grant from the NWO, and a ontrat from the European Community (BI02-CT93-0400, DG12 SSMA). * Address for orrespondene: H. P. Spaink, Institute of Moleular Plant Sienes, Leiden University, Wassenaarseweg 64, 2333 AL Leiden, The Netherlands. Phone: 071-275055. FAX: 071-275088. E-mail: SBY3HS@rulsfb.leidenuniv.n1. t Institute of Moleular Plant Sienes. 8 Gorlaeus Laboratories. published in Advane ACS Abstrats, September 15, 1995. terminal residue of hitin oligosaharides (John et al., 1993). Reently, pentameri and tetrameri forms of terminally de- N-aetylated hitin oligosaharides have been identified as being produed in vitro and in vivo by the ombined presene of the NodB and NodC proteins (Spaink et al., 1994), whih led to the designation of these moleules as NodBC intermediates. The NodA protein is assumed to transfer an ayl moiety from a donor to de-n-aetylated hitin oligosaharides, resulting in a basi LCO moleule (Rohrig et al., 1994; Atkinson et al., 1994). Various biovar-speifi modifiations in LCO struture, whih are determined by other nod genes, determine the biologial ativity of LCOs on the host plant. For Rhizobium leguminosarum bv. viiae, it has been shown that the presene of an 0-aetyl group on the 6-OH group of the nonreduing terminal residue of LCOs is required for preinfetion thread formation and nodule meristem formation (Spaink et al., 1991; van Brussel et al., 1992). Studies of mutants revealed that the nodl gene produt is required for the presene of this 0-aetyl group (Spaink et al., 1991). Reently, we have shown that purified NodL protein aetylates LCOs, hitin oligosaharides, and N-aetylgluosamine in vitro, using aetyl-oa as the aetyl donor (Bloemberg et al., 1994). In addition, we showed by strutural analyses that NodL transfers this aetyl moiety only to the 6-OH group of the nonreduing terminal residue (Bloemberg et al., 1994). In this paper, the harateristis and substrate speifiity of the transaetylase NodL protein are desribed in more detail. Our results show that a free amino group on the nonreduing terminal residue of oligosaharide aeptors is preferred by the NodL protein. On the basis of these and other results a model for in vivo LCO synthesis is proposed. 0006-296019510434- 127 12$09.00/0 0 1995 Amerian Chemial Soiety

Substrate Speifiity of NodL Protein Biohemistry, Vol. 34, No. 39, 1995 12713 Table 1: Large Sale Purifiation of NodL Protein purifiation step volume (ml) total protein (mg) total ativity (U) speifi ativity (Ulmg) pur fab reovery (%) soluble protein 11.0 115.2 30.0 0.26 1 100 (NH&S04 preipitation 2.9 69.6 27.9 0.40 1.5 93.0 Blue B 29.5 0.41 6.3 15.37 59.1 21.0 MonoQ 5.9 0.09 1.45 16.11 62.0 4.8 Purifiation data are given for 1 L of indued BL21(DE3) ells ontaining pmp3401. Purifiation fator. EXPERIMENTAL PROCEDURES Puripation of NodL Protein. NodL protein was isolated from Esherihia oli strain B121(DE3) (Studier et al., 1990) harboring pmp3401 (Bloemberg et al., 1994). pmp3401 ontains the nodl gene of R. leguminosarum bv. viiae under the ontrol of the T7 promoter. BL21(DE3) ontaining pmp3401 was grown in Luria-Bertani (LB) medium supplemented with kanamyin to a final onentration of 0.86 mm. To ativate the T7-polymerase promoter, isopropyl-p-dthiogalatopyranoside (IPTG) (final onentration of 2.5 mm) was added to a growing ulture with an OD260 of 0.4. NodL protein was purified to homogeneity from ell ultures indued for 16 h, using a purifiation proedure onsisting of Frenh pressure ell lysis, ultraentrifugation, ammonium sulfate preipitation, Blue B affinity hromatography, and MonoQ anion exhange hromatography, as desribed previously (Bloemberg et al., 1994). After eah purifiation step, protein samples were dialysed for 16 h against 10 mm Tris- HCl, ph 8.0, at 4 C and protein onentrations were determined aording to a modified Lowry quantifiation method (Dulley & Grieve, 1975) (see Table 1). Purified NodL protein was stored at -20 C. Chemials. Saharides used in this study were as follows: N-aetylgluosamine (Sigma), N, N -diaetylhitobiose (Sigma), N,N N -triaetylhitotriose (Sigma), N,N,N,N -tetraaetylhitotetraose (Seikagaku), N,N, - N, N, N -pentaaetylhitopentaose (Seikagaku), N, N,N, - N,K,N -hexaaetylhitohexaose (Seikagaku) (for brevity, the latter five ompounds are referred to in the text as hitinbiose, hitintriose, hitintetraose, hitinpentaose, and hitinhexaose, respetively), gluosamine (Sigma), hitosanbiose (Seikagaku), hitosantriose (Seikagaku), hitosantetraose (Seikagaku), hitosanpentaose (Seikagaku), ellopentaose (Sigma), and maltopentaose (Sigma). Analysis of Radio-Labeled Reation Produts. In order to demonstrate in vitro-aetylated reation produts, 100 nmol of various saharides were inubated with 1.4 yg of NodL protein and 50 nci (0.8 nmol) of [l- 4C]aetyl-CoA (Amersham) in 10 mm phosphate buffer, ph 7.5, in a total volume of 25 pl for 3 h at 28 C. In the ase of N-aetylgluosamine (Sigma), 1 pmol was used for labeling. For de-0-aetylation, samples were taken up in 400 yl of 1:l methano1:onentrated NH40H, inubated for 18 h at room temperature, and subsequently dried under vauum and redissolved in water. I4C-labeled N-aetylgluosamine and gluosamine were obtained from Amersham. 14C-labeled hitin fragments were obtained by methods as desribed by Kamst et al. in press. Reation samples were hromatographed on NH2 thin layer hromatography (TLC) plates (Merk) using a mobile phase of aetonitri1e:water (68:32, v:v). Radioativity on TLC plates was deteted with a PhosphorImaging system from Moleular Dynamis, using the Image Quant software. Spetrophotometri Assay for Transaetylating Ativity. Transaetylating ativity of NodL protein was spetrophotometrially analyzed using the method of Alpers et al. (1965). This method is based on a disulfide interhange reation between CoA, liberated from aetyl-oa after the transfer of the aetyl group to the aeptor substrate, and dithiobis(2-nitrobenzoi aid) reagent (NBT) (Ellman, 1959). The appearane of the reation produt thionitrobenzoi aid is followed spetrophotometrially at 415 nm. Unless indiated otherwise, reation mixtures had a final volume of 75 yl and ontained 125 nmol of aetyl-oa (Sigma) and 500 nmol of NBT (Fluka) in 67 mm Tris-HC1, ph 8.0. Furthermore, eah reation mixture ontained an amount of purified NodL protein and a variable amount of a saharide. Reations were performed at 28 C, and OD415 was measured in flat bottom mirotitration plates (Greiner) using a miroplate reader (Model 3550, BioRad). To avoid evaporation, 40 UL of mineral oil was added to eah reation mixture. Determination of NodL Enzyme Charateristis. In order to determine the influene of ph on the transaetylating ativity of NodL protein, reation mixtures ontaining 125 nmol of aetyl-oa, 20 ng of NodL protein, and 25 nmol of hitosanpentaose (Seikagaku, Corp.) were inubated for 15 min in final volumes of 50 yl in 50 mm phosphate buffer, ph 5.7, 6.2, 6.7, and 7.3, in 50 nm Tris-HC1 buffer, ph 7, 7.45, 8.0, 8.3, and 9.0, and in 50 mm glyine buffer, ph 9.1, 9.6, and 10.1, respetively. Subsequently, the reation mixtures were rapidly ooled in ie-water, and the amount of free CoA was determined spetrophotometrially after addition of 500 nmol of NBT in 25 yl of 100 mm Tris- HCl, ph 8.0. The influene of the Tris onentration on NodL protein ativity was determined by the testing of standard reation mixtures, ontaining 20 ng of NodL protein and 25 nmol of hitosanpentaose, in final Tris-HC1 (ph 8.0) onentrations of 0.067,O. 134,0.201,0.268,0.335, and 0.402 M, respetively. The influene of temperature on the stability of NodL protein ativity was determined after preinubating 20 ng of NodL protein for a period of 15 min at 4, 16, 25, 28, 37, 43, 48, 56, and 62 C. The ativity of eah preinubated NodL sample was tested in a standard reation mixture ontaining 25 nmol of hitosanpentaose. After an inubation period of 15 min at 28 C the amounts of free CoA were determined. The temperature dependene of NodL protein ativity was determined by the inubation of standard reation mixtures ontaining 20 ng of NodL protein and 25 nmol of hitosanpentaose for 15 min at the temperatures listed above. The influene of mono- and divalent metal ions on the ativity of NodL protein was determined by supplementing standard reation mixtures with CaC12, CoC12, MgC12, MnC12, ZnCl2, NaC1, or KC1 to a final onentration of 10 mm. The influene of helators on NodL protein ativity was tested

12714 Biohemistry, Vol. 34, No. 39, 1995 Bloemberg et al. NHA V R 1330 3 0 % 0 NR R R a NHA OMe 8 6, NIS/at.TfOH j L, 4 R R2 = Phth, R3 = Bn ii L, 5 R = R2 = H, R3 = Bn 1 R =R2=R3=H 7 * o * o &. + 6, NIS/at.TfOH o l R 1340 ~ 4 0 ~ 4 0 OMe NR2R3 NHA NHA 12 R1 = A, R2R3 = Phth, R4 = Bn ii, 13 R = R2 = R3 = H, R4 = Bn iv L, 111 R = R2 = R3= R4 = H Reagents and onditions: i. a) H2NNH2.H20, EtOH, 90 C b) MeOH, ApO; ii. Pd(OH)2/C, EtOH, 90 C; iii. Pd(OH)& EtOH, H20, HOA; iv. MeOH, t-buok; v. PdlC, H2, i-proh, H20, HOA. FIGURE 1: Organi synthesis of 1-methyl N -aetylhitobiose (I), 1-methyl N,N,N -triaetylhitotriose (11), and 1-methyl,A - diaetylhitotnose (111). For details, see Experimental Proedures. by addition of EDTA to a final onentration of 5 mm in a standard reation mixture. Chemial Synthesis and NMR Spetrosopy of De-Naetylated Chitin Derivatives. The assembly of the requisite 1-methyl N -aetylhitobiose (I), 1-methyl N,K,N -triaetylhitotriose (11), and 1-methyl N,W-diaetylhitotriose (111) was readily aomplished, as shown in Figure 1, by N-iodosuinimide-(NIS) and atalyti trifli aid-(tfoh) mediated (Veeneman et al., 1990) elongation of the terminal aeptor 1 (Alais & David, 1990) and 2 with appropriate donors 3 and 6 (Srivastava & Hindsgaul, 1991). Thus, hydrazinolysis of 1 and subsequent ondensation with 3 gave the fully proteted dimer 4. Hydrazinolysis of 4 followed by hydrogenolysis of 5 afforded homogeneous 1-methyl N - aetylhitobiose (I). In addition, glyosylation of 1 with donor 6 yielded the fully proteted dimer 7, deaetylation of whih afforded the partially proteted dimer 8. NIS- TfOH-assisted glyosylation of the latter with donor 6 gave, after hydrazinolysis of 9 followed by seletive N-aetylation and then hydrogenolysis of 10, homogeneous 1-methyl N, N,N -triaetylhitotriose (11). On the other hand, hydrazinolysis of dimer 7, and subsequent seletive aetylation of the free amino groups, resulted in the isolation of the partially proteted dimer 11. Elongation of dimer 11 with aeptor 6 gave the fully proteted trimer 12. Hydrazinolysis followed by hydrogenolysis of 13 led to homogeneous 1-methyl N,N -diaetylhitotriose (111).

Substrate Speifiity of NodL Protein Experimental Details. Methyl 3,6-Di-O-benzyl-2-deoxy- moleular sieves in DCE (10 ml) at 20 OC (N2 atmosphere) 2-aetamido-~-~-gluopyranoside (2). Compound 1 (312 mg) dissolved in ethanol (96%, 30 ml) was heated at 80 C after hydrazine-monohydrate (0.5 ml) was added. After 16 h, the reation mixture was onentrated in vauo and the residue was dissolved in MeOH (20 ml), and at 0 C, aeti anhydride (1.0 ml) was added. After stirring for 16 h at 20 C, the reation mixture was onentrated and oevaporated with toluene (5 x 5 ml). Purifiation of the residual oil by silia gel olumn hromatography (MeOH:EtzO, 0: 100 to 10:90, v:v) afforded 2 (180 mg, 70%). 13C{ H}-NMR (CDCl3): 6 23.4 (CH3 NHA), 54.5 (CH3 Me), 56.1 (C-2), 68.7, 74.5, 75.2 (C-6, 2 x CH2 benzyl), 75.2, 74.6, 77.2 (Cwas added 12.0 ml of a solution ontaining N-iodosuinimide (900 mg) and trifluoromethanesulfoni aid (39 pl) in DCE:Et20 (1: 1, 40 ml). After 1 h, the reation mixture was neutralized with Et3N, diluted with CH2Cl2, and filtered over Celite. The filtrate was washed with 10% Na2S203, 10% NaHC03, and H20, dried (MgSO& and onentrated. The rude produt was purified by silia gel olumn hromatography (Et0A:light petroleum, 30:70 to 5050, v:v) to afford 7 (800 mg, 88%). 13C{ H}-NMR (CDC13): 6 20.7 (CH3 aetyl), 55.4 (CH3 Me), 56.1, 56.2 (C-2,2 ), 67.9, 69.2, 72.5, 73.4, 73.8, 74.3 (C-6, 6, 4 x CH2 benzyl), 72.5, 73.2, 74.3, 76.0, 76.7 (C-3, 3, 4, 4, 5, 5 ), 96.4, 98.8 (C-1, l ), 3, 4, 5), 100.5 (C-1), 127.3-128.4 (CH,,, benzyl), 137.9, 123.0-133.9 (CH,, benzyl, Phth), 131.5 (Cq Phth), 137.6-138.3 (2 x Cq benzyl), 170.6 ((2-0 NHA). 138.4 (5 x Cq benzyl), 169.5 (5 x C=O Phth, A). Methyl 4-0-(3,4,6-Tri-O-benzyl-2-deoxy-2-phthalimido-P- Methyl 4-0-(3,6-Di-O-benzyl-2-deoxy-2-phthalimido-/3-~- ~-gluopyrunosyl)-3, 6-di-0-benzyl-2-deoxyy-2-aetamido-~- D-gluopyrunoside (4). To a mixture of aeptor 2 (162 mg), donor 3 (320 mg), and powdered 4 8, moleular sieves in 1,2-dihloroethane (DCE, 7 ml) at 0 C (N2 atmosphere) was added 5.1 ml of a solution ontaining N-iodosuinimide (230 mg) and trifluoromethanesulfoni aid (10 pl) in DCE:Et20 (l:l, 10 ml). After 3 h, the reation mixture was neutralized with EbN, diluted with EtOA, and filtered over Celite. The filtrate was washed with 10% Na2S203, 10% NaHCO3, and H20, dried (MgS04), and onentrated. The resulting oil was purified by silia gel olumn hromatography (MeOH:EtzO, 15:85, v:v) to afford 4 (129 mg, 34%). 13C{ H}-NMR (CDCl3): 6 23.3 (CH3 NHA), 52.9 gluopyranosyl)-3,6-di-o-benzyl-2-deoxy-2-phthalimido-~- D-gluopyranoside (8). Compound 7 (230 mg) was dissolved in dry MeOH (5 ml), and t-buok was added to adjust the ph to 10. After stirring for 2 h at 20 C, the reation mixture was neutralized with Dowex 50 XW4 resin (H+ form), filtered, and then onentrated in uuuo. The residue was purified by silia gel olumn hromatography (Et0A:light petroleum, 0:lOO to 40:60, v:v) to yield 8 (170 mg, 78%). Methyl 4-0-[4-0-(4-0-Aetyl-3,6-di-O-benzyl-2-deoxyy-2- phthalimido-~-~-gluopyranosyl)-3,6-di-o-benzyl-2-deoxy- 2-phthalimido-~-~-gluopyranosyl]-3,6-di-O-benzyl-2-deoxy- 2-phthalimido-P-~-gluopyranoside (9). Aeptor 8 (170 mg) and donor 6 (125 mg) were dissolved in DCE (5 ml) (CH3 Me), 56.1, 56.4 (C-2, 2 ), 68.1, 68.9, 72.6, 73.3, 74.7 (C-6, 6, 5 x CH2 benzyl), 73.9, 74.5, 74.8, 77.5, 78.8, 79.4 and treated under onditions idential to those desribed for the preparation of 7, using 23 ml of the NIS-TfOH mixture. (C-3, 3, 4, 4, 5, 5 ), 96.8, 101.2 (C-1, l ), 123.2-133.8 Purifiation of the resulting oil by silia gel olumn hro- (CH,,, benzyl, Phth), 131.4 (Cq Phth), 137.9-138.5 (5 x matography (Et0A:light petroleum, 30:70 to 5050, v:v) Cq benzyl), 169.9 (3 x C=O Phth, NHA). afforded 9 (140 mg, 61%). Methyl 4-0-(3,4,6-Tri-O-benzyl-2-deoxy-2-amino-P-~-glu- Methyl 4-0-[4-0-(3,6-Di-O-benzyl-2-deoxy-2-aetamido- opyrunosyl)-3,6-di-o-benzyl-2-deoxy-2-aetamido-~-~-glu- ~-~-gluopyranosyl)-3,6-di-o-benzyl-2-deoxy-2-aetamidoopyranoside (5). Hydrazinmonohydrate (223 pl) was P-~-gluopyranosyl]-3,6-di-O- benzyl-2-deoxy-2-aetamidoadded to a solution of 4 (129 mg) in ethanol (96%, 10 ml). P-D-ghopyranoside (IO). Trimer 9 (120 mg) was treated The reation mixture was heated for 25 h at 80 C. After in a manner idential to that used for the preparation of dimer onentration in vauo, the residue purified by silia gel 5. The rude produt was purified by silia gel olumn olumn hromatography (MeOH:EtzO, 0: 100 to 10:90, v:v) hromatography (MeOH:toluene, 10:90, v:v) and LH-20 afforded 5 (59 mg, 53%). 13C{ H}-NMR (CDCl3): 6 23.4 olumn hromatography (MeOH:CH2C12, 1 :2, v:v) and (CH3 NHA), 54.5 (CH3 Me), 56.4,57.6 (C-2,2 ), 68.6,68.8, afforded pure 10 (73 mg, 76%) 73.1, 73.2, 74.5, 75.2 (C-6, 6, 5 x CH2 benzyl), 74.8, 75.0, I-Methyl N,K -Triaetylhitotriose (ZZ). Compound 74.8, 75.9,77.4, 78.4, 84.8 (C-3, 3, 4,4, 5,5 ), 100.9, 103.2 (C-1, 1 ), 127.3-128.4 (CH,,, benzyl), 137.9-138.8 (5 x Cq benzyl), 170.6 (C=O NHA). 1 -Methyl K-Aetylhitobiose (I). Pd(OH)z/C (28 mg) was added to a solution of 5 (59 mg) in EtOH:H20 (40:1, v:v, 4.1 ml) and HOA (10 pl), and the mixture was shaken under an atmosphere of hydrogen. After 48 h at 20 C, the reation mixture was filtered, onentrated, and purified by LH-20 olumn hromatography (MeOH:H20, 4: 1, v:v) and additionally Fratogel HW 40 S olumn hromatography (0.15 M TEAB in MeOH:H20, 10:90, v:v) to furnish I(15 mg, 56%). 13C{ H}-NMR (D20): 6 22.7 (2 x CH3 NHA), 55.7 (CH3 Me), 57.2, 57.6 (C-2, 2 ), 60.8, 61.2 (C-6, 6 ), 70.1, 75.2, 75.2, 76.1, 76.6, 79.0 (C-3, 3, 4,4, 5, 5 ) 102.4, 103.2 (C-1, l ), 175.2 (2 x C-0 NHA). Methyl 4-0-(4-0-Aetyl-3,6-di-O-benzyl-2-deoxy-2-phthul- Biohemistry, Vol. 34, No. 39, 1995 12715 10 (50 mg) was dissolved in i-proh:hzo (4: 1, v:v, 4.0 ml) and HOA (0.18 ml). The mixture was shaken with PdC (10%) (25 mg) under an atmosphere of hydrogen for 16 h. Workup and purifiation were idential to those used for the preparation of ompound I to give pure I1 (13 mg, 51%).13C- { H}-NMR (D20): 6 22.9, 23.0 (3 x CH3 NHA), 55.7-58.2 (C-2, 2, 2, Me), 60.8-61.4 (C-6, 6, 6 ), 70.6, 73.0, 74.3, 75.3, 75.4, 75.6, 76.7, 80.2 (C-3, 3, 3, 4, 4, 4, 5, 5, 5 ), 102.2, 102.3, 102.7 (C-1, l, l ), 175.4, 175.5 (3 x C=O NHA). Methyl 4-0-(3,6-di-0-benzyl-2-deoxyy-2-aetamido-~-~gluopyranosyl)-3,6-di-O-benzyl-2-aetamido-~-~-gluopyranoside (11). Hydrazinolysis of dimer 7 (630 mg) under the same onditions as those desribed for dimer 5 gave, after purifiation by silia gel olumn hromatography (MeOH:toluene, 10:90, v:v), pure 11 (311 mg, 63%). 13Cimido-~-~-gluopyranosyl)-3,6-di-O-benzyl-2-deoxy-2- { H}-NMR (CDCl3): 6 23.1, 23.4 (2 x CH3 NHA), 51.3 phthalimido-p-d-gluopyrunoside (7). To a mixture of (CH3 Me), 54.5, 56.4 (C-2, 2 ), 69.6, 70.6, 72.2, 73.4, 73.6 aeptor 1 (450 mg), donor 6 (650 mg), and powdered 4 8, (C-6, 6, 4 x CH2 benzyl), 73.3, 74.6, 77.7, 80.5 (C-3, 3, 4,

12716 Biohemistry, Vol. 34, No. 39, 1995 4, 5, 5 ), 99.8, 101.6 (C-1, 1 ), 127.2-128.4 (CH,,, benzyl), 138.4-138.0 (4 x Cq benzyl), 170.3, 170.6 (2 x C=O NHA). Methyl 4-0-[4-0-(3,6-Di-O-benzyl-2-deoxy-2-amino-p-~gluopyranosyl)-3,6-di-O-benzyl-2 -deoxy-2-aetamido-p-~gluopyranosyl]-3,6-di-o- benzyl-2 -deoxy-2-aetamido-@-~gluopyranoside (13). Glyosylation of aeptor 11 ( 172 mg) with donor 6 (180 mg) under the same onditions as those desribed for dimer 7 gave, after purifiation by silia gel olumn hromatography (Et0A:light petroleum, 2090, v:v), trimer 12. This resulting trimer was dissolved in EtOH (96%, 15 ml), hydrazine-monohydrate (2.0 ml) was added, and the reation mixture was heated at 90 C. After 16 h, the solution was onentrated in uauo, and silia gel olumn hromatography (MeOH:toluene, 10:90, v:v) afforded 13 (32 mg, 11%). I -Methyl N,N -diaetylhitotriose (111). Hydrogen01 ysis and subsequent purifiation of ompound 13 (32 mg) under onditions idential to those desribed for the preparation B of I1 afforded I11 (9 mg, 49%). I3CC( H}-NMR (D20): 6 22.8 (2 x CH3 NHA), 55.6, 56.2, 56.6, 57.8 (C-2, 2, 2, Me), 60.8, 60.9, 61.0 (C-6, 6, 6 ), 70.2, 72.2, 73.0, 73.3, 75.1, 75.2, 77.1, 77.4, 79.9 (C-3, 3, 3, 4, 4, 4, 5, 5, 5 ), 102.0, 102.5 (C-1, l, 1 ), 175.3, 175.4 (2 x C=O NHA). Finally, the produts I, 11, and I11 were dissolved in 75% aetonitri1e:water (v:v) and purified on a nuleosil 120-7 NH2 HPLC olumn (Mashery Nagel) using an isorati elution in 75% aetonitrile in water, giving a reovery of 30% for eah. Quantitative Analysis of Suhsti*ate Speifiity. For quantitative studies of substrate speifiity of NodL protein, standard reation mixtures ontaining 100 ng of NodL protein were inubated at 28 C and the hange in OD2,s was monitored in time. The initial reation rates of transaetylating ativity were determined using 0.3 mm saharide substrates (V[0.3 mm]). For poor substrates, onentrations were raised to 3 or 33 mm and reations rates for 0.3 mm were obtained by linear extrapolation. For several aetylaepting substrates, a Mihaelis-Menten urve was obtained after the initial reation rates were determined at various saharide onentrations. Estimates of K,, values were alulated by linear regression from Lineweaver-Burke plots. These estimates were used to alulate K, values by nonlinear least squares using Statgraphis sofware. RESULTS In Vitro Transaetylating Labeling Studies Using [I - C]- Aetyl-oA. Previously, we have shown, using TLC and strutural analyses, that, in vitro, N-aetylgluosamine, hitin oligosaharides, and lipo-hitin oligosaharides are substrates for the NodL protein (Bloemberg et al., 1994). Chitin oligosaharides onsisting of a maximum of five N- aetylgluosamine units were tested and shown to be aetylaepting substrates for NodL protein. To test whether the NodL protein an aetylate longer hitin oligosaharides, hitinhexaose was inubated with purified NodL protein and [ 1-14C]aetyl-CoA. TLC analysis of the labeled reation produts shows that only one hitinhexaose dependent spot is deteted (Figure 2A, lane 7), whih migrates slower than the 0-aetylated form of hitinpentaose (Figure 2A, lane 6), as ould be expeted from its length. In both lanes 6 and 7, a spot is present in the middle of the TLC plate whih is not Bloemberg et al. 1 2 34 5 6 7 8 9101112 FIGURE 2: Analysis and haraterization of reation produts aetylated by NodL protein using NH1 TLC. A. After inubation, reation mixtures ontaining NodL protein, [ 1 -llc]aetyl-coa, and various mono- or oligosaharides were analyzed by TLC. Lanes: 1. no saharides added to the reation mixture; 2, N-aetylgluosamine: 3. hitinbiose: 4, hitintriose; 5. hitintetraose; 6, hitinpentaose: 7, hitinhexaose: 8, gluosamine; 9. hitosanbiose; 10, hitosantriose: 1 1. hitosantetraose: 12, hitosanpentaose; 13, 1 -methyl N -aetylhitobiose; 14. 1 -methyl AI, -diaetylhitotriose. B. Mild alkaline hydrolysis of NodL dependent reation produts or referene ompounds. Shown are untreated (uneven lanes) and alkaline-treated samples (even lanes). Lanes: 1 and 2, IT-labeled hitin fragments; 3 and 4 IT-labeled N-aetylgluosamine: 5 and 6, C-labeled gluosamine: 7-12, reation produts of mixtures ontaining NodL protein. [ 1-14C]aetyl-CoA, and gluosamine (7 and 8). hitosanpentaose (9 and IO), and no saharide added (,I 1 and 12). The migration positions of hitinpentaose (V). hitintetraose (IV). and hitintriose (111) are indiated with arrowheads. saharide dependent and is assumed to be [ 14C]-aetate (Figure 2A. lane 1). For a further substrate analysis of NodL protein, several other ommerially available saharides were used. TLC analysis of the labeled produts of reation mixtures ontaining gluosamine, hitosanbiose, hitosantriose, hitosantetraose, and hitosanpentaose shows that in eah ase a single radiolabeled saharide-related spot is piodued, with different Rf values on TLC, presumably du to differenes in their lengths (Figure 2A, lanes 8-12). All reation produts migrate more slowly than the 0-aetylated

Substrate Speifiity of NodL Protein Biohemistry, Vol. 34, No. 39, 1995 12717 A phosphate o tris HCI I glyine - 0 Y 5 x 8 C Q) 20 15 a 5 10 v) 0._ - V h. a 5 V m I 0 0 I I I 5 6 7 8 9 1 0 1 1 0 0.1 0.2 0.3 0.4 0.5 PH Tris (M) FIGURE 3: Dependene of NodL transaetylating ativity on ph and Tris-HC1 onentration. A. NodL transaetylating assays were performed in phosphate buffer, Tris-HC1 buffer, or glyine buffers at final onentrations of 0.05 M. B. Using 25 nmol of hitinpentaose, reations were performed under standard onditions in Tris-HC1, ph 8.0, of different molarities, and amounts of aetylated hitosanpentaose were determined 15 min after initiation of the reation. hitin oligosaharides of the same length (Figure 2A, lanes 2-6). The aetylated reation produt of gluosamine is migrating faster than N-aetylgluosamine or gluosamine (Figure 2B, lanes 3, 5, and 7). In addition, it is shown for the reation produts of gluosamine and hitosanpentaose that the 14C-aetyl group is sensitive to mild alkaline hydrolysis (Figure 2B, lanes 7-10), showing that hitosan fragments are 0-aetylated by the NodL protein. TLC analysis shows that the spots of the hitosan reation produts are more intense than those of the orresponding 0-aetylated hitin oligosaharides, indiating that hitosan oligosaharides are aetylated more effiiently than hitin oligosaharides. Inubation of NodL with mixtures of hitosanpentaose and NodRlv-V(C18:4) in ratios of 5 nmol:5 nmol and 5 nmol: 15 nmol, respetively, yields 0-aetylated hitosanpentaose as the major produt (data not shown). In onlusion, in omparison with lipo-hitin oligosaharides or hitin oligosaharides, hitosan oligosaharides are the preferred substrates of the NodL protein. Enzymati Charateristis of NodL Protein. In order to optimize the onditions for transaetylation, the enzymati properties of the NodL protein were studied in detail. For these studies, hitosanpentaose was hosen as the aetylaepting substrate, sine it is the best ommerially available aetyl-aepting substrate for NodL (Figure 2A). For a quantitative analysis of transaetylating ativity, the spetrophotometri assay aording to Alpers et al. (1965) was optimized as desribed in Experimental Proedures. Transaetylating ativity of NodL protein was determined at various ph values in the range 5.7-10.1 using different kinds of buffering agents. Figure 3A shows that the optimal ph for the NodL protein is approximately 9.0. However, studies for getting indiations about the in vivo substrates experiments were arried out at ph 8.0 sine (i) the ativity of NodL protein is still relatively high at ph 8 (Figure 3A) and (ii) ph 8.0 is presumed to be loser to the ytosoli ph than the optimal ph that we determined. The dependene of NodL transaetylating ativity on Tris-HC1 onentration at ph 8.0 was determined (Figure 3B). Ativity shows an optimum at 0.2 M Tris-HC1. NodL protein has a broad temperature optimum, between 28 and 42 "C (Figure 4A). NodL protein appears to be very stable at high temperatures, sine inubating NodL protein samples for 15 min at temperatures up to 48 "C, prior to the transaetylating reation being performed at 28 'e, does not lead to a signifiant loss of ativity (Figure 4B). The effet of the presene of metal ions on NodL protein ativity was determined by addition of various mono- and divalent metal ions in onentrations of 10 mm. Addition of K+, Na+, MgZf, Ca2+, Mn2+, and Co2+ results in transaetylating ativities of 112, 104, 90, 83, 57, and 13%, respetively. Addition of 5 mh4 EDTA does not affet transaetylating ativity. On the basis of these results, we have identified a set of standard onditions for the NodL transaetylation reation (see Experimental Proedures). We have defined one unit of NodL as an amount of protein whih an aetylate 1 pmol of hitosanpentaose in 1 min, under these standard onditions. This unit definition was used to monitor the large sale purifiation of NodL protein (Table l), whih was used for further analysis of substrate speifiity. Quantitative Analysis of NodL Protein Substrate Speijiity. In order to get indiations about in vivo aetyl aeptor substrates for NodL, a quantitative analysis of substrate speifiity was arried out using the protool desribed in Experimental Proedures. Initial reation rates of transaetylating ativity were determined for various hitin and hitosan oligosaharides. In addition, pentameri forms of a- 1,4- linked gluose (maltopentaose) and p- 1,4-linked gluose (ellopentaose) were tested. Initial reation rates for the substrates (0.3 mm) were ompared with the initial reation rate of the model substrate hitosanpentaose (Table 2). The results show that (i) reation rates for hitosan oligosaharides are muh higher than those for hitin oligosaharides, (ii) gluosamine and N-aetylgluosamine are muh less effiiently aetylated than their oligomeri forms, (iii) no detetable aetylation of maltopentaose was observed, and (iv) ellopentaose is aetylated by NodL protein at a rate whih is higher than that for hitinpentaose. To analyze the

12718 Biohemistry, Vol. 34, No. 39, 1995 Bloemberg et al. 12, h - 0 Y $ B m 0) a m u) 0 F - 0 2. w 0) u m I 0 12 I 0.,.,. 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 temperature ("C) pre-inubation temperature ("C) FIGURE 4: Effet of temperature on NodL protein. A. Dependene of the transaetylating ativity of NodL protein on temperature. Assays were performed at different temperatures using standard onditions. Amounts of aetylated hitosanpentaose were determined 15 min after initiation of the reation. B. Influene of temperature on the stability of NodL transaetylating ativity. NodL protein was preinubated for 15 min at different temperatures. Transaetylating ativity was subsequently determined at 28 O C under standard onditions. I Table 2: Substrate Speifiity of NodL Protein ~~ ~ u0.3 mm]" relative substrate (nmollmin) vb (%I K, (10-3~) hi tosanpentaose 3.97 100 1.1 f 0.1 hitosantetraose 2.70 68.1 1.5 i: 0.2 hitosantriose 0.32 8.0 6.0 & 0.9 hitosanbiose 0.20 5.0 nd' gluosamine 2.10 x 5.3 x lo-* nd hitinhexaose 3.97 x 1.0 x lo-' nd hitinpentaose 1.59 x lo-* 4.0 x 10-I '500 hitintetraose 9.93 x 2.5 x lo-' nd hitintriose 9.13 x 2.3 x nd hitinbiose 8.34 x 2.1 x nd N-aetylgluosamine 2.30 x 5.8 x nd ellopentaose 0.19 4.7 nd maltopentaose 0 0 I-methyl N',N"- 11.76 296.2 nd diaetylhitotriose 1-methyl N,N',N"- 7.54 x 1.9 x nd triaetylhitotriose a Initial reation rates of transaetylating ativity determined for saharides at 0.3 mm. Reation mixtures with final onentrations of 0.67 mm Tris-HCI, 1.5 mm aetyl-oa, and 6.0 mm Ellman's reagent ontained 100 ng of purified NodL protein, had a final volume of 75 pl, and were inubated at 28 "C. Perentages of initial reation rates relative to hitosanpentaose. nd = not determined. substrate affinity of NodL in more detail, K, values for several aetyl aeptors were determined from Lineweaver- Burke plots using nonlinear regression (Table 2). The results show that the K, value for hitinpentaose is at least 500- fold higher than the K, value for hitosanpentaose (Table 2). Furthermore, a omparison of K, values for hitosanpentaose, hitosantetraose, and hitosantriose shows that an inrease in the length of the hitosan fragment leads to a derease in the K, value. Substrate Speifiity for Syntheti Terminally De-Naetylated Chitin Derivatives. Sine our results show that for NodL protein hitosan is the preferred substrate, when ompared to hitin oligosaharides and LCO, the NodBC metabolites are good andidates for being in vivo aetylaepting substrates. Sine NodBC metabolites produed by Rhizobium strains annot be purified in large quantities, possibly due to the instability of these ompounds (unpub- lished results), the moleules, 1-methyl N'-aetylhitobiose and 1 -methyl ",A''-diaetylhitotriose, whih are struturally similar to NodBC metabolites, were hemially synthesized (Figure l), and their strutures were onfirmed by NMR spetrosopy (see Experimental Proedures). The I-methyl group was added in order to failitate a better monitoring of produts during synthesis and beause it is expeted to stabilize the produt. TLC analysis of NodL reation mixtures ontaining the syntheti ompounds shows that in all ases only one reation produt is formed (Figure 2A, lanes 13 and 14). Quantitative substrate analysis shows that the reation rate determined at 0.3 mm for 1-methyl N',N"- diaetylhitotriose is 3 times higher than that for hitosanpentaose (Table 2). Due to the limited quantities of the hemially synthesized ompounds, we have not been able to alulate an aurate K, value. 1-Methyl N,Nf,N"- triaetylhitotriose, whih was hemially synthesized as a ontrol ompound, is not more effiiently aetylated than hitintriose (Table 2) DISCUSSION In order to analyze their funtions, in vitro studies with several Nod proteins have been performed (Shwedok et al., 1990; John et al., 1993; Sutton et al., 1994; Atkinson et al., 1994; Rohrig et al., 1994; Geremia et al., 1994). Previously, we have shown that purified NodL protein in vitro atalyses aetyl transfer from aetyl-oa to the C-6 position of N-aetylgluosamine, hitin oligosaharides, and LCOs (Bloemberg et al., 1994). In this work, substrate speifiity and other enzyme harateristis of the NodL protein have been studied in further detail. NodL transaetylating ativity was analyzed using labeling studies and after the quantitative spetrophotometri assay desribed by Alpers et al. was optimized (1965). Enzymati properties of the NodL protein have been haraterized in order to optimize reation onditions. The results show that NodL ativity has a ph optimum of approximately 9, is stable at temperatures up to 48 "C, and has a temperature optimum around 35 "C (Figures 3 and 4). These properties seem remarkable for a protein that is loated in the ytosol (Bloemberg et al., 1994) and that has an optimal growth

Substrate Speifiity of NodL Protein temperature of 29 "C. From a pratial point of view, these properties failitate handling and storage of the NodL protein. The reation onditions finally hosen for studying the NodL protein should be onsidered to represent a ompromise between the ativity optima we determined and the presumed physiologial onditions. Our optimization of the reation onditions for NodL is valuable for the prodution of 0-aetylated ompounds for labeling studies and for obtaining new oligosaharides in large amounts. The in vivo aetyl-aepting substrate(s) for NodL protein is not known. Chitin oligosaharides and LCOs ould be onsidered to be putative in vivo aeptors for the NodL protein (Bloemberg et al., 1994). Surprisingly, our in vitro studies show that NodL protein is also able to 0-aetylate gluosamine, hitosan oligosaharides, and ellopentaose (Figure 2A,B, Table 2). The results show that hitosan oligosaharides are even better substrates for NodL than hitin oligosaharides. Cellopentaose is also more effiiently aetylated than hitinpentaose, but less effiiently than hitosanpentaose (Table 2). This shows that the presene of an N-aetylgluosamine residue at the nonreduing terminus is a more important negative fator than the presene of a gluose residue. NodL protein annot aetylate maltopentaose in detetable amounts, showing its strutural speifiity for p- 1,4-linked residues. The results also show that NodL has a different speifiity for oligosaharides varying in length. NodL protein shows the highest affinities for tetrameri and pentameri oligosaharides, whih orresponds well with the lengths of the N-aetylgluosamine bakbones of the major LCOs produed by Rhizobium. Quantitative substrate speifiity studies for LCOs are extremely diffiult, beause of their very low solubility in water. However, ompetition studies using small amounts of substrate and radiolabeled aetyl-oa show that hitosanpentaose is a muh better substrate than LCO. The onlusion that the NodL protein has a muh higher speifiity for hitosan oligosaharides than for hitin oligosaharides (Table 2) gives indiations for its in vivo substrates. Reently, pentameri and tetrameri hitin derivatives were identified whih lak the N-aetyl group on the nonreduing terminal residue, whih are produed by means of the NodB (hitin deaetylase) and NodC proteins (hitin synthase) and are therefore referred to as NodBC metabolites (Spaink et al., 1994). These metabolites are assumed to be intermediates in the biosynthesis of LCOs, funtioning as ayl aeptors for the NodA protein (Figure 5). @Aetylated NodBC metabolites were produed in the presene of the nodl gene, indiating that NodL in vivo is able to aetylate the hitin bakbone of LCOs before the substitution of the ayl moiety (Spaink et al., 1994). To date, the isolation of milligram quantities of NodBC metabolites for in vitro studies has not been suessful, possibly due to their instability. As an alternative, hitin derivatives resembling NodBC intermediates in struture, up to a length of three residues, were hemially synthesized (Figure 1). Quantitative analysis shows that 1-methyl N',"'-diaetylhitintriose is the ompound most effiiently aetylated by NodL protein of all the ompounds tested. The additional 1 -methyl group loated on the reduing terminal residue has little influene on substrate speifiity, as an be onluded from a omparison of the reation rates for l-methylhitintriose and hitintriose (Table 2). In onlusion, the results show that a free amino group on the nonreduing terminal Ho,CH3 o=, /o Biohemistry, Vol. 34, No. 39, 1995 12719 - UDP-GlNa Ho Aetyl-oA CoASH NH 1 o=( 1 \CH3 -d ayl donor N&A CH3 Ho o=/ NH LCO s n = 1, 2 or 3 5: Proposed model for the synthesis of lipo-hitin oli- FIGURE gosaharides (LCOs) in vivo. The following steps in the biosynthesis of LCOs are indiated: i, NodC funtions as a hitin synthase produing hitin oligosaharides of variable length; ii, NodB de- N-aetylates the nonreduing terminal residue; iii, NodL aetylates the NodBC intermediates at the 6-OH position of the nonreduing terminal residue using aetyl-oa as the aetyl donor; iv, NodA aylates the NodBCL intermediates, transferring an ayl moiety whih an be of variable struture. residue of the oligosaharide bakbone is preferred by the NodL protein. Taken together, the results of Spaink et al. (1994) and this report strongly suggest that NodBC intermediates are the in vivo aetyl-aepting substrates for NodL rather than hitin oligosaharides or LCOs, as depited in the model presented in Figure 5. ACKNOWLEDGMENT We thank Caroline van Hoek for her help with some of the experiments and J. E. Thomas-Oates for ritial reading of the manusript. REFERENCES Alais, J., & David, S. (1990) Carbohydr. Res. 201, 69-77. Alpers, D. H., Appel, S. H., & Tomkins, G. M. (1965) J. Biol. Chem. 240, 10-13. Atkinson, E. M., & Long, S. R. (1992) Mol. Plant-Mirobe Interat. 5,439-442. Atkinson, E. M., Palo, M. M., Hindsgaul, O., & Long, R. (1994) Pro. Natl. Aad. Si. U.S.A. 91, 8418-8422. Bloemberg, G. V., Thomas-Oates, J. E., Lugtenberg, B. J. J., & Spaink, H. P. (1994) Mol. Mirobiol. 11, 793-804. Brewin, N. J. (1991) Annu. Rev. Cell Biol. 7, 191-226. Bulawa, C. E., & Waso, W. (1991) Nature 353, 710. DebellC, F., Rosenberg, C., & DCnariC, J. (1992) Mol. Plant-Mirobe Interat. 3, 317-326. DCnariC, J., & Cullimore, J. (1993) Cell 74, 951-954.

12720 Biohemistry, Vol. 34, No. 39, 1995 Downie, J. A. (1989) Mol. Mirobiol. 3, 1649-1651. Dulley, J. R., & Grieve, P. A. (1975) Anal. Biohem. 64, 136-141. Ellman, G. L. (1959) Arh. Biohem. Biophys. 82, 70-77. Fisher, R. F., & Long, S. R. (1992) Nature (London) 357, 655-660. Geremia, R. A., Mergaert, P., Geelen, D., Van Montagu, M., & Holsters, M. (1994) Pro. Natl. Aad. Si. U.S.A. 91, 2669-2673. John, M., Rohrig, H., Shmidt, J., Wieneke, U., & Shell, J. (1993) Pro. Natl. Aad. Si. U.S.A. 90, 625-629. Kafetzopoulos, D., Thireos, G., Vournakis, J. N., & Bouriotis, V. (1993) Pro. Natl. Aad. Si. U.S.A. 90, 8005-8008. Kamst, E., van der Drift, K. M. G. M., Thomas-Oates, J. E., Lugtenberg, B. J. J., & Spaink, H. P. J. Bateriol. (in press). Lerouge, P. (1994) Glyobiology 4, 127-134. Rohrig, H., Shmidt, J., Wieneke, U., Kondorosi, E., Barlier, I., Shell, J., & John, M. (1994) Pro. Natl. Aad. Si. U.S.A. 91, 3122-3 126. Shlaman, H. R. M., Okker, R. J. H., & Lugtenberg, B. J. J. (1992) J. Bateriol. 174, 5177-5 182. Shwedok, J., & Long, S. R. (1990) Nature 348, 644-647. Bloemberg et al. Spaink, H. P., Sheeley, D. M., van Brussel, A. A. N., Glushka, J., York, W. S., Tak, T., Geiger, O., Kennedy, E. P., Reinhold, V. N., & Lugtenberg, B. J. J. (1991) Nature 354, 125-130. Spaink, H. P., Wijfjes, A. H. M., Geiger, O., Bloemberg, G. V., Ritsema, T., & Lugtenberg, B. J. J. (1993) in New Horizons in Nitrogen Fixation (Palaios, R., Ed.) pp 165-170, Kluwer Aademi Publishers, Dordreht. Spaink, H. P., Wijfjes, A. H. M., Van der Drift, K. M. G. M., Haverkamp, J., Thomas-Oates, J. E., & Lugtenberg, B. J. J. (1994) Mol. Mirobiol. 13, 821-831. Srivastava, G., & Hindsgaul, 0. (1991) J. Carbohydr. Chem. 10, 927-933 Studier, F. W., Rosenberg, A. H., Dunn, J. J., & Dubendorff, J. W. (1990) Methods Enzymol. 185, 60-89. Sutton, J. M., Lea, E. J. A., & Downie, J. A. (1994) Pro. Natl. Aad. Si. USA. 91, 9990-9994. Van Brussel, A. A. N., Bakhuizen, R., van Spronsen, P., Spaink, H. P., Tak, T., Lugtenberg, B. J. J., & Kijne, J. (1992) Siene 257, 70-72. Veeneman, G. H., Van Leeuwen, S. H., & Van Boom, J. H. (1990) Tetrahedron Lett. 31, 1331-1334. BI950739Z