"Detection of Citrus Leaf and Seed Glycoproteins Using Biotinylated

Home , Blot (biology)

HORTSCIENCE 26(7):910-913. 1991. chromatography (Chicken and Sharom, 1983), to study carbohydrate structure (Chu et al., Detection of Citrus Leaf and Seed 1981) and, when conjugated to ferritin, flu- orescein isothiocyanate, Horseradish perox- Glycoproteins using Biotinylated idase, or biotin derivatives, as histochemical and cytochemical probes (Pena et al., 1981). Each lectin’s characteristic carbohydrate- Lectin Probes binding specificity can distinguish glycopro- Randall P. Niedz1, Michael G. Bausher2, and C. Jack Hearn3 teins separated electrophoretically (Gordon and Pena, 1982). Horticultural Research Laboratory, U.S. Department of Agriculture- The ABC affinity method used is a mod- Agricultural Research Service, 2120 Camden Road, Orlando, FL 32803 ification of an immunohistochemical procedure developed by Hsu et al. (1981) for Additional index words. Concanavalin A, SDS-PAGE, electroblotting studying polypeptide hormones. The ABC Abstract. Citrus leaf and seed glycoproteins were detected after sodium dodecyl sulfate method was estimated to be 20 to 40 times polyacrylamide gel electrophoresis (SDS-PAGE) and electroblotting onto polyvinyli- more sensitive than the peroxidase-antipe- dene difluoride (PVDF) membrane when probed with biotinylated lectin at 1 µg·ml-1. roxidase (PAP) method. A modified ABC Four lectins representing three carbohydrate-binding groups were used as probes. A method, using biotinylated lectins, was faster preformed avidin-biotin-complex (ABC) was used to detect the glycoprotein-bound and gave better resolution than 125I-labeled lectins and resulted in dark bands and little background staining. Concanavalin A lectins when used to detect electroblotted cell- (ConA) and wheat germ agglutinin (WGA) resulted in the darkest-staining bands. The surface glycoproteins of human skin fibro- four Citrus spp. and one related species studied had unique seed glycoprotein profiles blasts (Gordon and Pena, 1982). It had the when probed with ConA and WGA. This procedure might be useful in clarifying citrus further advantage of being nonradioactive. taxonomy, providing genetic markers, and in physiological studies involving glycopro- Shah and Stegemann (1986), by the same teins. technique, discriminated between two wheat cultivars using wheat germ agglutinin (WGA) Citrus taxonomy is an unresolved, com- and Rhodes, 1976). Barrett believes the fa- and Ulex europaeus agglutinin (UEA). These plex problem. For instance, the number of cultative apomictic reproductive system of two cultivars could not be distinguished by species reported to comprise the genus ranges Citrus is responsible for much of the con- sodium dodecyl sulphate polyacrylamide gel from 162 (Tanaka, 1977) to three (Barrett fusion over what constitutes a true species. electrophoresis (SDS-PAGE) and isoelectric Citrus spp. relationships have been studied focusing. Glycoprotein profiles from four Received for publication 30 Apr. 1990. Mention by numerical taxonomy (Barrett and Rhodes, Citrus spp. and one Citrus relative were ex- of a trademark, warranty, proprietary product, or 1976), isozyme profiles (Torres et al., 1978), amined to determine the usefulness of the vendor does not constitute a guarantee by the U.S. chemical constituents (Rouseff et al., 1987), ABC method in glycoprotein detection in Dept. of Agriculture and does not imply its ap- and Ouchterlony immunoprecipitation pat- Citrus. proval to the exclusion of other products or ven- terns (Morimoto, 1977). Leaf protein extraction. Shoots were in- dors that may also be suitable. We thank Delores This paper presents a sensitive procedure duced on 2-year-old Citrus sinensis L. Os- F. Jordan for her competent assistance and Ran- to detect specific glycoproteins based on the beck cv. Hamlin seedlings under sodium dall C. Smith for preparing the figures. The cost -1 -2 avidin-biotin-complex (ABC) method of Hsu vapor lamps (1200 µmol·s ·m , 16-h pho- of publishing this paper was defrayed in part by toperiod, 25 C). The new, just-elongated flush the payment of page charges. Under postal regu- et al. (1981) by probing electroblotted citrus lations, this paper therefore must be hereby marked leaf and seed proteins with biotinylated lec- of leaves was collected, washed with dis- advertisement solely to indicate this fact. tins. Lectins are carbohydrate-binding pro- tilled water, dried, and the midrib removed. 1Research Geneticist. teins, not immunologically derived and Leaves were frozen in liquid nitrogen and 1Research Plant Physiologist. generally of plant origin. They have been powdered with mortar and pestle. Cold (4C) 2 Research Geneticist. used to isolate glycoproteins by affinity extraction buffer [50 mM Tris·HCl pH 8.4,

910 HORTSCIENCE, VOL. 26(7), JULY 1991 Table 1. Major ConA and WGA-binding gly- or 1.0% crystallized BSA (U.S. Biochemical coproteins in four Citrus spp. and one relative. Corp., Cleveland). Excess blocking agent was removed by washing (3´, 5 min) in HBS. Biotinylated lectins Concanavalin A (Con A), pokeweed agglutinin (PWA), peanut agglutinin (PNA), and wheat germ agglutinin (WGA) were obtained from Sigma Chemical Co. (St. Louis). Lectins were dissolved in HBS containing 0.04% sodium azide and stored at – 20C. Blots were then incubated for 60 min in 5 ml HBS containing biotin- Fig. 1. Identification of lectin-binding glycopro- ylated lectin at 1-50 µg·ml. To remove the teins in C. sinensis cv. Hamlin leaf and aced unbound lectin, blots were washed (3´, 5 protein extracts on a 12% SDS-PAGE gel. (A) zSee Fig. 3. min) in HBS. Leaf proteins; (B) seed proteins. Lane 1, WGA; y+, Band present; –, band absent. An avidin-biotinylated alkaline phospha- lane 2, ConA; lane 3, PNA; lane 4, PWA. Blots xSee Fig. 4. tase complex (ABC-AP) was made using the -1 were probed with 1 µg·ml of lectin and gly- Vectastain ABC-alkaline phosphatase kit coproteins detected using the ABC technique of (Vector Laboratories, Burlingame, Calif.) Hsu et al. (1981). Molecular mass markers are and electroblotting. Leaf and seed extracts according to the manufacturer’s instructions. indicated on the left and right. were mixed 1:1 and 1:9, respectively, with To detect biotin-labeled lectins, blots were a solution containing 0.1 M Tris·HCl (pH first incubated in 5 ml ABC-AP solution for kDa 6.8), 2% (v/v) glycerol, 1% (w/v) SDS, 10% 90 min at 37C, washed (3 ´, 5 min) in HBS (v/v) 2-mercaptoethanol and 0.0002% (w/v) to remove unbound ABC-AP, and finally 66 bromophenol blue. The extract then was equilibrated for 15 min before staining in heated for 5 min at 90C and 10 µl was loaded substrate buffer (0.1 M Tris·HCl pH 9.5, 0.1 2+ per lane (»10 µg of protein) for leaf and M NaCl and 50 mM Mg ). Alkaline phos- 42 seed extracts. Electrophoresis was per- phatase was visualized by adding 1.65 mg formed in a Mini-Protean II 7 cm ´ 8 cm nitroblue tetrazolium chloride (NBT) and 0.85 31 ´ 0.75 mm slab cell (Bio-Rad, Richmond, mg 5-bromo-4-chloro-3-indolylphosphate p- Calif.) using the Laemmli buffer system toluidine salt (BCIP) (Bethesda Research (Laemmli, 1970: 0.125 M Tris pH 6.8 stack- Laboratories, Gaithersburg, Md.) to 10 ml 21 ing gel (4% T, 2.67% C), 0.375 M Tris pH substrate buffer. Blots were incubated in 5 8.8 separation gel (8% and 12% T, 2.67% ml of substrate in darkness until bands de- 14 C), and Tris-glycine (0.025 M Tris, 0.192 M veloped. The color reaction was stopped by glycine, 0.1% SDS, pH 8.3) electrode buffer. washing the blots in 20 mM Tris·HCl (pH The proteins were separated by 40 mA con- 7.5), 5 mM ethylene diaminetetraacetate 123 stant current until the dye front reached the (EDTA). Fig. 2. Control blot to determine nonspecific lectin bottom of the gel, i.e., »1 h. Bromophenol In control experiments, the blots were binding and possible endogenous alkaline phos- blue was used as the tracking dye and phos- treated as above but lectins were omitted, or phatase activity. Blot is seed protein extract from phorylase b (97 kDa), BSA (66 kDa), oval- the lectin solution contained 0.5 M of the C sinensis cv. Hamlin run on 12% SDS-PAGE. bumin (42 kDa), carbonic anhydrase (31 inhibiting saccharides for ConA and WGA, Lane 1, ConA; lane 2, ConA incubated with kDa), soybean trypsin inhibitor (21 kDa), a -methyl-D-mannoside and N-acetylglucos- 0.5 M a -methyl-D-mannoside before probing; and lysozyme (14 kDa) were used as molec- amine, respectively. Experiments testing the lane 3, no lectin. Molecular mass markers are ular mass standards. Chemical reagents and four blocking agents were carried out with indicated on the left. standards for SDS-PAGE were purchased ConA only. from Bio-Rad Laboratories. The increased sensitivity of the ABC

150 mM NaCl, 1 mM CaCl2, 1% insoluble After electrophoresis the gels were equil- method is based on the extreme affinity of polyvinylpolypyrrolidone (PVP)] was mixed ibrated for 30 min in transfer buffer: Tris/ the egg white glycoprotein avidin (70,000 with the leaf powder (0.5 ml buffer per 0.1 glycine (14 mM/11 mM), 0.0375% (w/v) SDS, MW) for biotin (dissociation constant g leaf tissue) and stirred for 15 min. This pH 8.0 and 20% (v/v) methanol. Polyvinyl 10-15M). As avidin has four binding sites solution was filtered through cheesecloth and difluoride (PVDF) membrane was prepared for biotin, a “lattice” structure is believed the filtrate centrifuged for 20 min at by a 20-sec soak in methanol, three 5-min to be formed between avidin and biotin-la- 20,000´ g. Protein concentration was de- washes in distilled water, and equilibration beled enzyme with avidin acting as a bridge termined by the Bradford dye-binding method for 15 min in transfer buffer. Proteins were between the biotin-labeled enzyme mole- (Smith, 1988) using bovine serum albumin electroblotted on PVDF (Immobilin-P, Mil- cules (Hsu et al., 1981). Because of the in- (BSA) as a standard. The supernatant was lipore, Bedford, Mass.) transfer membrane creased sensitivity, a low ligand concentration stored at – 20C. for 30 min in a field strength of 1.6 mA/cm2 can be used resulting in minimal background Seed protein extraction. Seeds from C. using the LKB NovaBlot (Pharmacia LKB staining. We found that lectin concentration aurantium L. (sour orange), C. grandis L. Biotechnology, Piscataway, N.J.). was the single most important factor con- Osbeck (pummelo), C. sinensis cv. Hamlin Probing and staining. After electroblot- tributing to background staining. Concentra- (sweet orange), Poncirus trifoliata L. Raf- ting, the blots were washed (3´, 5 min) in tions >1 µg·ml-l significantly increased the inesque (trifoliate orange), and C. paradisi Hepes buffered saline (HBS): 10 mM HBS, background. Band intensity was not im-

Macfayden cv. Conners (grapefruit) were 150 mM NaCl, pH 7.5,0.1 mM CaCl2·2 H2O, proved with higher lectin concentrations. extracted from mature fruit, air-dried, and 0.1 mM MgCl2·6 H2O and 0.01 mM MnCl2·4 Background staining was also influenced by the coats removed. Decoated seeds were H2O. Nonspecific lectin binding sites were the type of blocking agent used. It is nec- ground to a paste in an analytical laboratory blocked by incubating the blots for 30 min essary to minimize nonspecific lectin bind- mill (Chemical Rubber Co., Cleveland). Cold at 37C in HBS containing one of the follow- ing by blocking the remaining free-binding (4C) extraction buffer was mixed (1 ml/0.l ing blocking agents: 1) 0.5% or 3% liquid sites on the transfer membrane after elec- g) with the seed paste and stirred for 15 min. fish gelatin (Hipure Liquid Gelatin, Norland troblotting. Of the four blocking agents tested, Subsequent steps were the same as for leaf Products, New Brunswick, N.J.), 2) 5% de- 1% BSA resulted in less background than the protein extraction. fatted milk powder (Carnation Co., Los An- others. At lower concentrations, BSA re- SDS-polyacrylamide gel electrophoresis geles), 3) 0.1% Tween 20, 4) 0.1%, 0.5%, sulted in higher background. Nonfat milk

HORTSCIENCE, VOL. 26(7), JULY 1991 911 with ConA is shown in Fig. 2. The first con- teins. The four Citrus spp. and one relative, trol (lane 2) included the inhibitory saccha- P. trifoliata, had unique glycoprotein pro- rides mixed with the lectin before probing. files when probed with four biotinylated lec- This control tests the specificity of the lectin. tin probes. ConA and WGA-binding Other than some minor bands, there were glycoproteins stained as dark bands with lit- two heavy staining bands between the 21 kDa tle background staining. A single represent- and 31 kDa standards. The second control ative genotype from each of five species was (lane 3) omitted ConA. This was an alkaline used in this study to demonstrate the tech- phosphatase control. The two dark bands were nique. Further characterizations are needed unchanged in the second control blot. We to determine the technique’s utility in var- made no attempts to identify these bands. ietal identification and taxonomic classifi- They may represent endogenous alkaline cation. Lectin binding may provide additional phosphatase, as these bands appeared when insight into the taxonomic relationships be- Fig. 3. ConA-binding glycoprotein profiles of four NBT and BCIP were used, chromogenic tween Citrus spp. and relatives, and be use- Citrus spp. and one relative run on 8% SDS- substrates for alkaline phosphatase. Also, a ful as a detection method in the further PAGE. Lane 1, sour orange; lane 2, pummelo; reductase may have reduced NBT to the in- characterization and study of citrus glyco- lane 3, sweet orange; lane 4, grapefruit; lane soluble formazan directly. These two bands proteins. 5, trifoliate orange. Molecular mass markers are were found in all five species. ConA-binding indicated on the left. bands were located between 31 kDa and 66 Literature Cited kDa and between 21 kDa and 31 kDa for all Barrett, H.C. and A.M. Rhodes. 1976. A numer- the species. Control blots for WGA (not ical taxonomic study of affinity relationships in shown) were similar to ConA controls in that cultivated Citrus and its close relatives. System- WGA-binding bands were located between atic Botany 1(2): 105–136. 31 kDa and 66 kDa. Species comparisons of Chicken, C.A. and F.J. Sharom. 1983. The con- these two regions showed that the region from canavalin A receptor from human erythrocytes 31 kDa to 66 kDa was the most variable. To in lipid bilayer membranes; interaction with better resolve this region, 8% rather than 12% concanavalin A and succinyl-concanavalin A. gels were used. Biochim. Biophys. Acta 729:200-208. Major ConA-binding glycoproteins ap- Chu, F.K., F.M. Maley, and A.L. Tarentino. 1981. peared as prominent, darkly stained bands The use of iodinated lectins for determining the degree of deglycosylation of high-mannose gly- with distinct differences between species (Fig. coproteins by endo- b -N-acetylglucosaminidase 3, Table 1). Minor ConA and WGA bands H. Anal. Biochem. 116:152-160. were not scored. Two bands just below the Goldstein, I.J. and R.D. Poretz. 1986. Isolation, 31 kDa marker were present in each species physiochemical characterization, and carbo- Fig. 4. WGA-binding glycoprotein profiles of four (Fig. 2, lane 1). These two bands are not hydrate-binding specificity of lectins, p. 35-247. Citrus spp. and one relative run on 8% SDS- visible in Fig. 3 as they ran off the gel with In: LE. Liener, N. Sharon, and I.J. Goldstein PAGE. Lane 1, sour orange; lane 2, pummelo; use of 8% acrylamide. The region between (eds.). The lectins: properties, functions, and lane 3, sweet orange; lane 4, trifoliate orange; 31 kDa and 66 kDa contained a distinct applications in biology and medicine. Aca- lane 5, grapefruit. Molecular mass markers are banding profile for each species (Fig. 3). The demic, Orlando, Fla. indicated on the left. major bands are summarized in Table 1. Band Gordon, B.B. and S.D.J. Pena. 1982. The surface a, the slowest migrating band, was present glycoproteins of human skin fibroblasts detected after electrophoresis by the binding of powder decreased band intensity and in- in all five species. Band d was only faintly peanut (Arachis hypogaea) agglutinin and Ri- creased background significantly compared visible in sour orange, but prominent in cinus communis (castor-bean) agglutinin I. with the other three blocking agents. This pummelo. Trifoliate orange contained a large, Biochem. J. 208:351-358. reaction might be due to the presence of gly- dark staining zone that migrated at the d band Hsu, S-M., L. Raine, and H. Fanger. 1981. Use coproteins recognized by ConA. Tween 20 position. Sweet orange’s d band was closely of avidin-biotin-peroxidase complex (ABC) in and liquid fish gelatin resulted in good band sandwiched by two unique bands, c and e, immunoperoxidase techniques: A comparison definition but high background. In all sub- not observed in the other species. Band c is between ABC and unlabeled antibody (PAP) sequent experiments, 1% BSA was used. distinct because four bands were resolved procedures. J. Histochem. Cytochem. 29:577- Leaf and seed proteins were extracted from when sweet orange and grapefruit proteins 580. four Citrus spp. and one related species, run were mixed, the b band plus the three sweet Ladizinsky, G. and T. Hymowitz. 1979. Seed in SDS-PAGE, electroblotted, and probed orange bands—bands c, d, and e. Band h protein electrophoresis in taxonomic and evo- lutionary studies. Theor. Applied Genet. 54:145- with four biotinylated lectins representing was unique to grapefruit. 151. three carbohydrate-binding groups (Gold- Major WGA-binding glycoprotein profiles Laemmli, U.K. 1970. Cleavage of structural pro- stein and Poretz, 1986); ConA, a mannose/ were obtained for each species (Table 1). teins during the assembly of the head of bac- glucose-binding lectin, WGA and PWA, N- Bands were not as prominent as those ob- teriophage T4. Nature (London) 227:680-685. acetylglucosamine-binding lectins, and PNA, tained with ConA. Bands a and d were only Morimoto, J. 1977. Identification of Citrus spe- a galactosyl end-group-binding lectin. Of the present in pummelo and grapefruit (Fig. 4). cies by phytoserological methods. Proc. Intl. four lectins used, ConA resulted in the most Band a is actually two bands and band d is Soc. Citriculture 2:625-631. prominent bands (Fig. 1). Banding profiles faint, particularly in grapefruit. Band c was Pena, S.D.J., B.B. Gordon, G. Karpati, and S. of seed proteins developed rapidly (30 sec) unique to pummelo and band e is actually Carpenter. 1981. Lectin histochemistry of hu- and were much darker than those obtained two bands, and was present in sour orange, man skeletal muscle. J. Histochem. Cytochem. from leaf proteins. A taxonomic advantage pummelo, and grapefruit. Band e was de- 29:542-546. of using seed proteins is their known stability tected with WGA but not PWA though both Rouseff, R.L., S.F. Martin, and C.O. Youtsey. in spite of fluctuating environmental factors lectins have the same reported sugar-binding 1987. Quantitative survey of narirutin, narin- gin, hesperidin, and neohesperidin in Citrus. J. (Ladizinsky and Hymowitz, 1979), unlike specificity. Trifoliate orange had faint bands Agr. Food Chem. 35:1027-1030. potentially variable leaf proteins. Seed pro- that migrated to the band e position, but were Shah, V.A.A. and H. Stegemann. 1986. Lectin- teins and ConA were used in the majority of slightly more separated than those classified based characterization of glycoproteins in plants experiments. as band e. and its application to cultivar discrimination. Control blots were run to determine if there This work demonstrates that the ABC af- Nachrichtenbl. Deut. Pflanzenschutzd. 38:100- were any nonspecific lectin-positive bands. finity technique of Hsu et al. (1981) is easily 103. A blot of sweet orange seed protein probed adapted to the detection of plant glycopro- Smith, J.A. 1988. Quantitation of proteins, p.

912 HORTSCIENCE, VOL. 26(7), JULY 1991 10.1.1-10.1.3. In: F.M. Ausubel, R. Brent, R.E. in molecular biology, vol. 2. Wiley, New York. Torres, A.M., R.K. Soost, and U. Diedenhofen. Kingston, D.D. Moore, J.G. Seidman, J.A. Tanaka, T. 1977. Fundamental discussion of Cit- 1978. Leaf isozymes as genetic markers in cit- Smith, and K. Struhl (eds.). Current protocols rus classification. Studia Citrogia 14:1-6. rus. Amer. J. Bot. 65(8):869-881.

HORTSCIENCE, VOL. 26(7), JULY 1991