The Plant Cell, Vol. 3, 259-270, March 1991 O 1991 American Society of Plant Physiologists

A Nodulin Specifically Expressed in Senescent Nodules of Winged Bean 1s a Protease lnhibitor

Jean-François Manen,”’ Patrice Simon,b Jan-Christoph Van Slooten,c Magne Q)steras,c Severine Frutiger,d and Graham J. Hughesd

a Jardin Botanique, 1 ch. de I’lmpératrice, 1292-Chambésy, Genève, Switzerland Laboratoire de Physiologie Végétale, Université de Genève, 3 place de I’Université, 1205 Genève, Switzerland Laboratoire de Biologie Moléculaire des Plantes Supérieures, Universite de Genève, 1 ch. de I’lmperatrice, 1292-

Département de Biochimie Médicale, Faculté de Médecine, 1 rue Michel Servet, 1206 Genève, Switzerland Chambésy, Genève, Switzerland

Nodule senescence is one aspect of nitrogen fixation that is important to study from the perspective of improving the host-bacteroid interaction. In winged bean nodules, a 21-kilodalton protein is specifically expressed when senescence begins. Using subcellular fractionation, we observed that this plant protein interacts with the bacteroids. Microsequencing of the protein allowed us to obtain a specific oligonucleotide that was used to isolate the corresponding nodule cDNA. Sequence analysis of this cDNA revealed that the 21-kilodalton protein has all of the features of a legume Kunitz protease inhibitor. Subsequent analysis confirmed that this nodulin is indeed a protease inhibitor. lmmunocytochemical study showed that the protease inhibitor is exclusively localized i11 infected senescent cells of the nodule, particularly in disorganized bacteroids, the peribacteroid membrane, vacuole membranes, and in the vacuole fluid. The specific expression of a protease inhibitor at senescence may be of particular interest if the targeted proteolytic activity is important for the symbiotic relationship. This point is discussed in relation to the known nodule proteases.

INTRODUCTION

Nitrogen fixation in legumes is performed by endosym- biotic bacteria of the genus Rhizobium or Bradyrhizobium. By binding to the root hairs, the bacteria induce the root to produce a specific structure, the nodule, and then invade its central part through an infection thread. lnside the host plant cells, they differentiate in a new morphological and physiological form, the bacteroid. These bacteroids are enclosed in a host-cell membrane system, the peribacter- oid membrane (PBM) and fix nitrogen (Goodchild and Bergersen, 1966; Kijne and Planqué, 1979; Mellor and Werner, 1987; Long, 1989). The ability of individual in- fected cells to fix nitrogen is a transient step followed by senescence in a short period of time (Kijne, 1975). Mor- phological studies of senescent nodule cells show a dis- order in the PBM and the formation of large vacuoles containing degraded bacteroids and autophagocytized cy- toplasm (Kijne, 1975; Vance et al., 1980; Mellor and Wiemken, 1988). Within a developing nodule, all the cells are not at the same stage of development. The different stages are easily observed in indeterminate nodules but exist in determinate nodules such as winged bean nodules (Vance, 1986). In determinate nodules, the senescent cells

’ To whom correspondence should be addressed.

are localized in the central part of the nodule and radiate from the center. When pods and seeds are developing, they mobilize a large part of the assimilated carbon to the detriment of nodules, whose growth is thereby sup- pressed. All of the nodule-infected cells become senes- cent, and the nitrogen fixation decreases rapidly.

During symbiosis, the host plant expresses a certain number of proteins specific to nodule development and nitrogen fixation called nodulins. The best known among these proteins are leghemoglobin, which is found in all legumes and regulates oxygen tension in nodules (Witten- berg et al., 1974), uricase II and glutamine synthetase II, involved in ammonia assimilation (Bergmann et al., 1983; Cullimore et al., 1984), and sucrose synthase (Thummler and Verma, 1987).

Other nodule-specific host proteins, sometimes ex- pressed at stages earlier than leghemoglobin, have been characterized, but their functions remain generally un- known and are under investigation (Katinakis and Verma, 1985; Campos et al., 1987; Gloudemans et al., 1987; Jacobs et al., 1987; Dunn et al., 1988; Scheres et al.,

If senescence could be delayed, an improvement in the nitrogen value of legume crops would be possible at the

1990).

260 The Plant Cell

stage of seed formation, i.e., when the plant needs largeamounts of nitrogen. Proteases are probably involved innodule senescence (Vance, 1986; Peoples and Dalling,1988) because two thiol proteases (one with a pH optimumof 3.5 and the other with a pH optimum of 8.0) have beenpurified from French bean senescent nodules and appearto be active only during senescence. It has also beenshown that they are able to partially digest in vitro thepeptidoglycans of the bacteroid cell wall (Pladys et al.,1986; Pladys and Rigaud, 1988). A serine protease withdecreasing activity during senescence was found in func-tional nodules (Pladys and Rigaud, 1985). Moreover, serineprotease activities in senescent soybean nodules havealso been observed (Malik et al., 1981; Pfeiffer et al., 1983),but degradative changes were not detected in the bacter-oids. So far, no specific expression of other nodulins andno specific activation of enzymes have been reported insenescent nodule tissue.

By differential immunoscreening of winged bean noduleproteins, we observed a nodule-specific 21-kD protein thatappears at the beginning of senescence. Partial amino acidsequencing and oligonucleotide designing enabled us toisolate a specific cDNA clone from a nodule cDNA library.Sequencing and analysis of this cDNA showed that the21-kD protein is homologous to the plant Kunitz trypsininhibitors.

21 kO

Legh.

Figure 1. Evidence for a Late-Expressed 21-kD Polypeptide inNodule Development.

Total root or nodule proteins were extracted in hot SDS samplebuffer, run on an SDS gel, blotted on nitrocellulose, and immu-nostained with nodule-specific antibodies. Lane 0, sections ofmature roots with no apparent nodule budding (noninfected root);lane 1, root sections containing 0.25-mm nodule buds; lane 2,0.5-mm nodules; lane 3, 1-mm nodules; lane 4, 1.5-mm nodules;lane 5, 2-mm nodules; lane 6, 2.5-mm nodules; lane 7, 3-mmnodules; lane 8, 4-mm nodules; lane 9, 5-mm nodules. Nodule-specific antibodies were used as primary antibodies, and goatanti-rabbit IgG coupled to alkaline phosphatase were used assecondary antibodies. Legh., leghemoglobin.

RESULTS

Evidence for the Appearance of a 21-kD Protein VeryLate in Nodule Development

Proteins were extracted from nodules at different stagesand separated on an SDS gel. A protein gel blot was donewith nodule-specific antibodies (see Methods), and theresult is shown in Figure 1. The different developmentstages are defined by the nodule size. We observed thatmost of the labeled nodule-specific proteins appear at thesame stage as leghemoglobin, which is present whennitrogen fixation occurs. An interesting observation wasthat a nodule-specific protein of 21 kD appeared muchlater (stage 6) than leghemoglobin (stage 3). At this stage,many infected cells of the central zone of the nodules werealready senescent.

Subcellular Location of the Late Nodule-SpecificProtein

A nodule-specific protein can be expressed by the plant orby the microsymbiont genome. To determine whether the

21-kD protein was associated with bacteroids, subcellularfractionation experiments were conducted on large nod-ules. Figure 2 shows the different fractions analyzed byimmunoblotting with nodule-specific antibodies. Under theconditions used, most antibodies of the antiserum ap-peared to be directed against the bacteroid proteins rep-resented by bacteroid proteins themselves and by theproteins of the peribacteroid fluid (PBF), instead of theplant proteins represented by the soluble fraction. Thismay be due to the large amount of bacteroid proteins inthe tissue (thousands of bacteroids relative to the numberof infected cells) and/or to the greater antigenicity ofbacteroid proteins. The presence of the 21-kD polypeptidein the lane corresponding to the pure bacteroid fractionwould suggest that it is a bacteroid protein, but a largeamount of this protein is present in the soluble fraction anda small amount is present in the PBF and membranefractions. This result would favor the hypothesis that theplant protein is targeted to the bacteroids.

The faint 21-kD band observed in the PBM lane (Figure2) would reflect the transient passage of this protein acrossthe PBM from the plant cell cytoplasm to the bacteroids.However, the opposite is also possible, and it cannot beruled out that the 21-kD polypeptides seen in the bacteroidlane and in the soluble proteins lane are totally differentproteins with the same apparent molecular weight.

Protease Inhibitor in Senescent Nodules 261

•o0)<*rfIc

N o d u l e

«>0)1o

5Q. 00

0.

TJo0>*rfos

.2̂0)o

O)

0)0)

designed from polypeptide 3 (Figure 3B) and used as probefor RNA gel blot hybridization of plant mRNA. Figure 4Ashows that oligonucleotide 1 hybridized to a nodule polyA*transcript of 0.9 kb that was absent in noninfected roots.This result demonstrated that the 21-kD soluble protein isprobably encoded by a plant gene because the prokaryoticmicrosymbiont has no polyA+ mRNA, unless a prokaryoticpolyA~ transcript contaminated the polyA+-purifiedfraction.

Antibodies against the purified 21-kD protein were pro-duced and reacted with nodule subcellular fractions byimmunoblotting. Figure 4B shows that the antibodies re-acted with the 21 -kD protein of both fractions—the solubleproteins and the bacteroid proteins—supporting the hy-pothesis that the late expressed 21-kD protein is a plantnodulin interacting with bacteroids.

The Late Nodulin Is Specifically Localized inSenescent Cells

21 kD

Legh

The plant cell cytoplasm (where the protein is synthesized)and the bacteroids (with which it interacts) are separatedby the peribacteroid membrane, which acts as a barrier.To reach the bacteroid, the 21-kD protein must be specif-ically sorted out from the cytoplasm. However, direct

Figure 2. Subcellular Localization of the 21-kD Protein.

Proteins from nodule subcellular fractions (see Methods) were runon an SDS-gel, blotted on nitrocellulose, and immunostained withnodule-specific antibodies. The nodule lanes are, respectively:total, total nodule proteins extracted in SDS-sample buffer; solu-ble, nodule-soluble proteins from the first sucrose gradient super-natant; PBM, peribacteroid membrane; PBF, peribacteroid fluid;bacteroid, bacteroids from the second sucrose gradient. Thenoninfected root lane is the same as lane 0 in Figure 1. The free-living bacteria lane is total proteins from a Rhizobium sp NGR234culture. The antibodies used are the same as in Figure 1. Legh.,leghemoglobin.

(1) He Lys Pro Tyr Sec Pro Gly Ser(2) Leu Leu Phe Cys Pro Asp Gly Ser Gin Cys(3) Asn Asn Gly Glu Tyr Tyr lie Val Pro(4) Leu Val Val Ser Asn Thr Leu Pro Thr Leu

Asn Asn Gly Glu Tyr Tyr...AAU AAU GGU GAA UAU UAU...3'

C C C G C CAG

The Late Nodule-Specific Protein Is a Plant NodulinThat Interacts with Bacteroids

A different approach was used to resolve the ambiguity ofthe origin of the 21-kD polypeptide. The soluble proteinwas purified from large nodules by a procedure involvinganion-exchange chromatography on a DEAE-Sephadexcolumn followed by preparative SDS-PAGE. The purifiedprotein was digested with trypsin and the derived peptideswere separated. The N-terminal sequence of four peptideswas determined by automatic Edman degradation. Theprotein sequence data of the four tryptic peptides areshown in Figure 3A. Three oligonucleotides mixtures were

( 1 )( 2 )

( 3 )

...TTG TTG CCG CTT ATG ATG...5'C

...TTG TTG CCT CTT ATG ATG...5'C

...TTG TTG CCC CTT ATG ATG...5'C

Figure 3. Microsequencing and Specific Oligonucleotide Designof the 21-kD Protein.

(A) N-terminal sequence of four 21-kD protein oligopeptides. Thepurified 21-kD protein was trypsin digested, and the resultingpolypeptides were microsequenced by automated Edmandegradation.(B) Design of three mixtures of oligonucleotides from peptide 3(Asn-Asn-Gly-Glu-Tyr-Tyr). They are complementary to the mRNAand the choice was made according to codon use statistics inwinged bean.

262 The Plant Cell

o -oo occ z"5 3

0)5 +•*1 ga a

0.9kb

21 kD

Figure 4. The 21 -kD Protein Is a Host Protein That Interacts withthe Bacteroids.

(A) RNA gel blot hybridization of noninfected root and nodulepolyA* RNAs with the 32P-kinased oligonucleotide 1 mixture(TTGTTGCCGCTTATGATG and TTGTTGCCGCTCATGATG) de-signed from peptide 3 (Asn-Asn-Gly-Glu-Tyr-Tyr). One microgramof the polyA* RNAs was loaded on the denaturing gel.(B) Protein gel blot of the nodule subcellular fractions (the sameas in Figure 2) reacted with anti-pure 21 -kD protein antibodies.

interaction of this protein with the bacteroids is possible ifat the beginning of senescence the PBM becomes disor-ganized and permeable. To answer these questions andalso to complement the results obtained by cell fractiona-tion, we made optic and electron microscopic studies ofnodule cells.

Figure 5A shows thin sections of a 3-mm nodule ob-served by optic fluorescence microscopy. The small darkcells are noninfected nodule cells. The large, lightly flu-orescent cells are functional infected cells. The large cellscontaining fluorescent inclusions are suspected to be se-nescent infected cells. Nothing is known about the chem-ical constitution of the fluorescent material. This could bedue to protein-metal complexes such as ferric-specificligands (siderophores), which are found in Rhizobium(Bosch et al., 1988). Nevertheless, this method is able todetect senescent cells without ambiguity. Ultrathin sec-tions of the same nodule tissues were incubated withrabbit anti-21-kD protein antibodies and stained with goldparticles complexed with protein A. Figures 5B, 5C, and5D show senescent cells labeled with gold particles. (Thelabeling is not visible in Figure 5B because of low magni-fication.) These cells contain large vacuoles produced bythe fusion of vesicles containing electron-dense material.

Vance et al. (1980), when studying senescent cells ofalfalfa nodules, observed the same vesicles containing

electron-dense material and defined them as disorganizedbacteroids enclosed in PBMs. The electron-dense materialcould be residual bacteroid aggregates containing metalions such as iron and molybdenum, which are componentsof nitrogenase. It has been calculated that the concentra-tion of these metals in soybean bacteroids is 14 mM and1 mM, respectively (Verma and Nadler, 1984). These met-als would make bacteroid material electron dense. Byfusion, the vesicles make large vacuoles containing aggre-gated bacteroid materials (Vance et al., 1980). In thesecells, gold particles are associated with the senescentdisorganized bacteroids, the vacuoles, and the resultingmembranes from PBM fusion. A few gold particles, whichmust not be confused with ribosomes, were observed inthe cytoplasm, which is apparently in contradiction to theresults obtained by subcellular fractionation experiments.Upon close examination, the membranes seem to maintaintheir integrity, and nonspecific leakage of the 21 -kD proteinwould, therefore, be unlikely. Figure 5E shows that func-tional infected cells were not labeled by anti-21-kD proteinantibodies. Figure 5F shows a control experiment withpreimmune serum.

cDNA Sequence and Deduced Amino Acid SequenceAnalysis

We screened a nodule cDNA library with oligonucleotide1, which hybridized with a 0.9-kb polyA+ RNA. Two cloneswere identified out of 3000 phage plaques. Figure 6 showsRNA and DNA gel blot hybridization of one clone to plantmRNAs and genomic DNAs. The clone hybridized to a 0.9-kb transcript, which appeared late in the nodule develop-ment (Figure 6A), later than the leghemoglobin transcript(Figure 6B). DNA gel blot experiments confirmed that this0.9-kb messenger is a transcript from the plant genomeand that the bacterial microsymbiont genome does notcontain any related sequence (Figure 6C). The same re-sults were obtained with the other clone (data not shown).

The sequence of the two selected clones is identical andis shown in Figure 7. We did not succeed in obtaining thecomplete transcript. The two clones are interrupted exactlyat the same 5'-end site, suggesting repeated interruptionsof reverse transcription in the conditions used (30 min at37°C). The reading frame shown (representing 201 aminoacids with a predicted molecular mass of 22.3 kD) is longerthan that required for a 21-kD polypeptide. If we assumethat the apparent molecular weight obtained by SDS-PAGE is correct, this suggests that only a few amino acidsare missing from the N-terminal end and that a leadersequence may be encoded.

The 21-kD Nodulin Is a Protease Inhibitor

Comparison of the amino acid sequence with protein databank sequences indicates homology between our se-

Protease Inhibitor in Senescent Nodules 263

B

nic

Figure 5. Optic and Electron Microscopy of Nodule Cells.

(A) Thin sections of a 3-mm nodule observed with an optic microscope equipped with fluorescence. Bar = 20 /urn.(B), (C), and (D) Electron-immunogold labeling of ultrathin section of senescent cells with anti-21-kD protein antibodies as primaryantibodies and 15-mm gold particles coupled to protein A. Bar = 0.4 ^m in (B) and 0.2 ^m in (C) and (D).(E) The same as in (B), (C), and (D) but on ultrathin sections of functionally infected cells. Bar = 0.1 ^m.(F) Control. The same as in (B), (C), and (D) but with preimmune serum as primary antibodies. Bar = 0.2 ^m.ic, infected cell; nic, noninfected cell; sc, senescent cell; cy, cytoplasm; pbm, peribacteroid membrane; pbf, peribacteroid fluid; b, bacteroid;db, disorganized bacteroid; v, vacuole; cw, cell wall; t, tonoplast.

264 The Plant Cell

0 1 2 3 4 5C3Z

OU)

0.9kb

21 kD protein

kbp-23-9.4-6.5-4.3

-2.3-2.0

0.7kb

Leghemoglobm BamH1 EcoFM

Figure 6. Temporal Regulation of the 21-kD Protein Gene andGene Representation in the Host Genome.

(A) RNA gel blot hybridization of 1 ^g of noninfected root anddeveloping nodule polyA+ RNAs with a 32P nick-translated 21 -kDprotein cDNA clone insert. Lane 0, noninfected roots; lane 1, 0.5-mm nodules; lane 2, 1-mm nodules; lane 3, 2-mm nodules; lane4, 3-mm nodules; lane 5, 5-mm nodules.(B) The same as in (A) but with a 32P nick-translated winged beanleghemoglobin cDNA clone insert.(C) DMA gel blot hybridization of Rhizobium sp NGR234 (NCR)and winged bean (PSO) genomic DNA. Five micrograms of DNAwere digested with BamHI or EcoRI. The probe used was the 32Pnick-translated 21-kD protein cDNA clone insert.

quence and soybean (Kim et al., 1985; Jofuku et al., 1989),as well as winged bean (Yamamoto et al., 1983) seedtrypsin inhibitors. The sequence alignment is shown inFigure 8. The N-terminal and C-terminal sequence regionsare very similar. The central region is less similar, but the4 cysteine residues that make two disulfide bridges in theseed trypsin inhibitors are present at the same positions.The reactive arginine residue at position 63 in the soybeanprotein and position 64 in winged bean seed is also pres-ent. However, the arginine residue is followed by a prolineresidue that prevents trypsin modification at this site. Alysine-serine bond at position 56-57 could be the truereactive site.

By comparing the protein sequence of the mature soy-bean seed trypsin inhibitor (Kim et al., 1985) with thepredicted amino acid sequence of the cloned protein,Jofuku et al. (1989) deduced that an N-terminal signalsequence and a C-terminal sequence are cleaved to pro-duce the mature protein. By analogy, and knowing thesequence of the mature winged bean seed trypsin inhibitor(Yamamoto et al., 1983), we suggest that in the 21-kDprotein described here 11 amino acids from the C-terminalend are likely to be post-translationally cleaved, and thatonly 5 codons of the N-terminal end of the putative signalsequence are missing from our cDNA clone.

The calculated high isoelectric point (10.2) deduced fromthe amino acid composition is in agreement with the factthat this protein does not bind to DEAE-Sephadex at pH8.6 (see Methods). Although variable, the isoelectric pointof Kunitz seed inhibitor is generally lower (Yamamoto etal., 1983; Richardson et al., 1986).

To confirm the function inferred from the observed ho-mology, we tested in vitro the biological activity of the late-developing 21-kD nodulin. Bovine pancreatic trypsin wascoupled to CNBr-activated Sepharose 4B. A nodule ex-tract was loaded on the column, and the bond materialwas eluted by lowering the pH. Figure 9 shows that theacidic eluate contains exclusively a polypeptide of 21 kD(lane 3), which reacts with the anti-21-kD protein antibod-ies (lane 4).

Experiments to determine protease inhibition, shown inFigure 10, demonstrate that the 21-kD protein is indeed aprotease inhibitor and is active against trypsin and againstendogenous proteolytic activities. The results show that,with the condition used, 1.5 ^g of the 21-kD protein wasable to totally inhibit 0.5 M9 of trypsin and 30% of theendogenous proteolytic activity when the inhibitor was firstremoved from the nodule extract. This would explain thetotal disappearance of leghemoglobin in unbound material

75.CGAC

-20 -10

100 125 ISOACAAACGGTAACCAACTTAGAAACAATGGCGAATACTACATAGTGCCAGTTTCAGGAGGCGCTGGAATAGATGTTT N G N Q L R N N G E Y Y 1 V P V S G G A G I D V

10 20 30175 200 225

GTTGCAACAGGAACAGAAAAATGCCCTCTCACTATAGTGCAATCCTCAAGTAGCACCAAGTATTCAACCGTTTTC

AAATCCGAAGCAACGC

60 70

300G C

325 350 375AATCAATACAGAGGGAGACCATGGACTGTAGTGAGGGCTCTACCAGAAGGGCTCGCTCTQAAGCTTAATGGGTACN Q Y R G R P H T V V R G L P S G L A L K I . N G Y

90 100400 425 450

ACAAACACAGCACGGGTACTCAGAATCAAACCCTATTCTCCTGGCTCAAGAAACTACAAGCTTTTGTTCTGCCCA

475 500 525GATGGTAGCCAGTGTGCCAATGTTGGTTTTACACTGATGCTCAGCGAAAGAAAGCGTTTGGTTGTGTCTAATACG

140550

625 650 675GAr.Tf.AGTGAAAf;AACTGCAAGATTGAGTATGTGCTTTGATTGTAATAAATTATTATGAAATAAATAAAATATCAE •

700AAGCTTATGTTTGGTTTTAAAAAAAAAAAAAAAAAAAAAAAAAAA

Figure 7. Nucleotide and Deduced Amino Acid Sequence of the21-kD Protein cDNA Clone.

The sequences of the four microsequenced peptides are under-lined. The cDNA insert is 720 bp long (the polyA stretch is 27 bplong). The deduced protein sequence contains 201 amino acids.The polyadenylation signals AATAAA are indicated.

Protease Inhibitor in Senescent Nodules 265

-20 -10 1 10 20

(A) KKSTIFFAI.rLVCAFTISYI.PSATA-DFVLDNEGNPLENGGTYYILSDITAF-GGIR

( B I XXXXXLFTI.ri.I,SSrT-SYLPSATAQQPVYDTNCNQI.RNNGEYYIVPVSG—GAGID"* * * ** * ** * * **

( C) EPLLDSEGELVRNGGTYYLLPDRWALGGGIE

30 40 50 60 70 80

(A) AAPTGNEF C PLTWQSRNE-LOKGIGTIJSSPK P IRFIAEGHPLSLlCFDSFAVIMi

( 8 ) WATGTEF C PLTIVOSSSSTKYSTVEg|EATtP F PINYI—————TQNMAFSIKFI

( C ) AAATGTE1CPLTWRSPNE-VSVGEPLRISSOIFSGFIP——DYSWRIGFANPP:

90 100 110 120 130

(Al V-GIPTEWSWEDLPEGPAVKIGENKDAMDGWrRLERVSDDErmiYm.Vf :PQQAED

( B ) NQYRGRPWTWRGLPEGIALKLNGYTNTARVtRIKPYSPGS—RNYKH.F C PDG——**** * ** • * * *** •

( C) A-PSP-WWTWEDOPOOPSVKLSELKSTKFDYLFKFEl!VTSKrSSYKLltV C AKR——

140 ISO 160 170 180

(A) DKCGDIGISIDHDDGTRRLWSKNKP-LWQFQKLDKESLAEKNHGLSRSG

(B) SO C AHVCFTLB-LSERKRLWSNTLPTLQVOFOKYG-PRFAKDHHLRSVAE* * *•** * * * * *

( C) D1C KDIGIYRD-OKGYERLWTDENP-LWIFKK-V-ESS

Figure 8. Alignment of the 21-kD Protein Sequence with TwoPublished Kunitz Trypsin Inhibitor Protein Sequences.

Rows A, cloned soybean seed trypsin inhibitor sequence desig-nated KTi 3 (Jofuku et al., 1989); rows B, cloned winged beannodule 21-kD protein sequence; rows C, mature winged beanseed trypsin inhibitor sequence (Yamamoto et al., 1983). Theputative signal sequence and the putative C-terminal cleavedsequence are underlined. The putative N-terminal missing aminoacids are indicated by XXXXX. The 4 cysteines forming twodisulfide bridges are boxed, as is the arginine residue at position63. A suggested reactive site for trypsin, the lysine-serine bondat position 56-57, is boxed. The amino acid position numbersrefer to the nodule 21-kD protein and correspond to those ofFigure 7.

proline residue (isoleucine in soybean seed and serine inwinged bean seed), which makes it resistant to hydrolysisby trypsin. However, it actually binds to trypsin (Figure 9).As noted in Results, the Lys-Ser bond at position 56-57could be the actual reactive site. Interestingly, an Arg-llebond is also found at position 58-59 in winged bean seedtrypsin inhibitor (Yamamoto et al., 1983). The calculatedmolecular mass of the predicted N-terminal processed 21-kD protein is 20.2 kD, which is close to the 21 -kD apparentmolecular mass. The calculated molecular mass of the N-terminal and C-terminal processed protein (Figure 8) is18.9 kD, which is close to the 19.2 kD of the winged beanseed trypsin inhibitor (Yamamoto et al., 1983), but whichis much lower than our observed 21 -kD apparent molecularmass. This would probably suggest that the C-terminalprocessing is not achieved in the nodules, but this needsto be demonstrated.

The winged bean nodule protease inhibitor is also differ-ent from the winged bean seed inhibitor. The DNA gel blotstringent hybridization experiment (Figure 6C) showed thatthere are several different Kunitz protease inhibitor genesin the plant genome. There was no BamHI or EcoRI sitein our nodule cDNA, and two or three bands were ob-served. If the nodule protease inhibitor genes of wingedbean have no introns, as is the case for the soybean

(Figure 9, lane 2) from which the protease inhibitor wasremoved. These experiments confirm that the late-devel-oping 21 -kD protein is indeed a biologically active proteaseinhibitor.

DISCUSSION

Characterization of the Nodule Protease Inhibitor

We have reported on the complete characterization of asenescence-related nodulin. As indicated in the Introduc-tion, nodule senescence is a development stage that isimportant to study with the aim of improving nitrogenfixation.

The 21 -kD protein we have characterized is homologouswith the legume Kunitz inhibitors (Ryan, 1988). The great-est homology is found in the N-terminal and C-terminalregions. The arginine residue of the putative active site islocated at position 63 as in soybean. It is followed by a

21 kD

Legh.

Figure 9. Specific Binding of the Nodule 21-kD Protein onTrypsin.

SDS-PAGE results. Lane 1, nodule soluble proteins; lane 2, nodulesoluble proteins that do not bind to the trypsin-Sepharose column;lanes 3 and 4, acidic eluate of the material bound to the trypsin-Sepharose column. Lanes 1,2, and 3 are Coomassie Blue stainedand lane 4 is a protein gel blot stained with anti-pure 21-kD proteinantibodies.

266 The Plant Cell

0-0- O 2 4 6 8 10

pm of the 21-kD protein

Figure 10. Protease lnhibitor Activity of the 21-kD Protein.

The caseolytic activity of a fixed amount of trypsin (A) or of a nodule extract before (O) and after (O) passage through a trypsin- Sepharose column were inhibited with increasing amounts of the 21 -kD protein eluted from the trypsin-Sepharose column (see Figure 9). The noninhibited caseolytic activity of trypsin (0.5 pg) and of nodule extract before and after passage through the affinity column (200 pg and 175 pg, respectively) were 0.1 02, 0.030, and 0.033 A400/hr, respectively. The inhibitory activity is expressed in percent of inhibition.

trypsin inhibitor genes (Jofuku and Goldberg, 1989), each observed band (Figure 6C) would represent at least one individual gene. This is not surprising because Jofuku and Goldberg (1 989) found, under nonstringent hybridization conditions, 1 O different trypsin inhibitor-related sequences in the soybean genome.

The expression of the protease inhibitor we have char- acterized is temporally regulated and coincides with the development of senescent cells in the nodule. By what signal is this expression regulated? Jofuku and Goldberg (1 989) showed the expression of severa1 Kunitz protease inhibitors in soybean tissue. Two of them (KTil and KTi2), which are homologous to the major seed trypsin inhibitor, are expressed at low levels in seeds and in leaves and roots. If these genes are transferred to tobacco plants, they are correctly regulated in seeds and leaves but not in roots, where no messenger is present. The reason could be that in tobacco the trans-acting regulator of transcrip- tion is absent from root cells. In soybean roots, the regu- lator would be the result of the presence of Rhizobium, which is obviously impossible in tobacco roots.

lnvolvement of the Nodule Protease lnhibitor in Nodule Senescence

A protease (thermolysin) inhibitor has been observed in the PBF of soybean root nodules (Garbers et al., 1988). It is heat stable (30 min at 1 OOOC), has a molecular mass of 20 kD, and is not active against trypsin. No further char- acterization (particularly if it is specifically expressed during senescence) allowed us to compare it with the 21-kD protease inhibitor.

The involvement of the nodule protease inhibitor in se- nescence is demonstrated by two observations. First, it appears very late in nodule development, after nitrogen fixation has reached a steady state, demonstrating that it is not necessary for nitrogen fixation. Second, it is exclu- sively localized in senescent cells. Electron microscopy studies have located the nodule protease inhibitor in dis- organized bacteroids, PBM, and in resulting membranes as well as in PBF and resulting vacuole fluid. Its near absence in PBM and PBF subcellular fractions (Figure 2) could be explained by the fact that in these fractions the majority of PBM and PBF results from functional infected cells (see Figure 5A) that do not contain the 21-kD protein (see Figure 5E), and/or by the fact that senescent vesicles are more fragile than functional infected cell vesicles and are destroyed in subcellular fractionation experiments. This would also explain the high leve1 of the protein in the soluble fraction and the apparent contradiction between subcellular fractionation and electron microscopy experiments.

The steady state of interaction between the plant and Rhizobium is called symbiosis because of reciproca1 inter- ests in the exchange. It is clear that during infection of roots by the microsymbiont (which can be considered parasitic at this stage), the bacteria produce signals that divert the host's defense. Nevertheless, the observed rapid senescence could be considered as a delayed reaction of the host plant against Rhizobium. Thiol-protease activities have been observed in French bean senescent nodules (Pladys et al., 1986) and should be considered as a host- pathogen interaction. Although suggested, it has not been demonstrated that the isolated French bean proteases are expressed by the plant and not by the bacteroids. They are able to digest in vitro the bacteroid peptidoglycans. In soybean nodules, a serine proteolytic activity against azo- casein increases when nitrogenase activity is lost (Pfeiffer et al., 1983). However, this activity does not seem to degrade the bacteroids.

In this context, it would be interesting to know the function of this late-developing nodulin, which had shown protease inhibitor activity. With few data available on the senescence process of nitrogen fixation, it is difficult to speculate about the significance of a protease inhibitor in nodules. The two proteases appearing in French bean senescent nodules are thiol proteases. The inhibitor we have characterized in winged bean nodule is specific to

Protease lnhibitor in Senescent Nodules 267

serine proteases (Ryan, 1981). In French bean, it has been shown that a serine protease activity is present in func- tional nodules but disappears at senescence (Pladys et al., 1986). If a Kunitz-type protease inhibitor is also present in French bean nodules, it would be of interest to observe its interaction with the serine protease and to determine whether this protease activity decreases at senescence because of its own disappearance or of the appearance of the inhibitor.

The function of a protease inhibitor could be to regulate a protease activity. To understand this regulation, studies on the origin, as well as on the target, of nodule proteases must first be undertaken. We have localized the nodule protease inhibitor in senescent cells. We speculate that the inhibitor interacts specifically with a protease, the origin of which is still to be determined (plant or bacteroids), that is important for the maintenance of symbiosis. The steady state of nitrogen fixation is probably an equilibrium, where proteases may play a role. However, when the equilibrium is shifted (appearance of a protease inhibitor), the host- pathogen interaction is established and the plant rejects the invading microorganism.

Further studies will be conducted to understand the biological significance of this protease inhibitor within nod- ules induced by Rhizobium mutants producing Fix- phe- notypes, such as those that are characterized by lysis of bacteroids (early senescence) (Werner et al., 1984). Puri- fication and complete characterization of the specific protease inhibited by the nodule inhibitor will also be undertaken.

METHODS

Biological Material and Nodule Production

Winged bean (fsophocarpus tetragonolobus Dc cv UP599) seeds were treated for 10 min in concentrated sulfuric acid, 10 min in 10% (w/v) sodium hypochlorite, washed with sterile distilled water, dispersed on 1% agar, and allowed to germinate for 1 week at 28OC in the dark. Seedlings were then disposed in a mist-room containing nitrogen-free medium (Cutting and Schulman, 1969) and incubated at room temperature (2OOC to 25OC) under direct sunlight. When the preleaves were fully developed, the mist-room was inoculated with Rhizobium sp NGR234 (Trinick, 1980). The nodules produced were harvested according to their size, ranging from 0.25 mm (root sections with nodule budding) to 5 mm. Under these conditions, the 0.25-mm nodule buds first appeared after 8 days to 10 days after inoculation, and the 5-mm nodule size appeared 10 days to 20 days later. Although the precise timing of nodule growth was not determined, the sizes of the nodules were a good approximation of their stage of development.

Root sections with no apparent nodule budding were collected for controls (noninfected roots). The harvested nodules and the root controls were used immediately or frozen in liquid nitrogen and stored at -70°C.

Protein Extraction and Nodule Subcellular Fractionation

Total proteins from fresh or frozen nodules and uninfected roots were extracted in 50 mM Tris-HCI buffer, pH 8.0, containing 5 mM EDTA, 5 mM P-mercaptoethanol, and 1 mM phenylmethyl- sulfonyl fluoride for antibody production, or directly in hot SDS- sample buffer [125 mM Tris-HCI, pH 6.8, 1% (w/v) SDS, 0.1% (v/v) j3-mercaptoethanol, 10% (v/v) glycerol] for SDS-PAGE.

Subcellular fractions were prepared according to the method of Fortin et al. (1 985). Briefly, fresh nodules (2 mm to 5 mm) were ground in 50 mM Tris-HCI buffer, pH 8.0, containing 5 mM EDTA, 5 mM 0-mercaptoethanol, 1 mM PMSF, and 0.5 M sucrose. Bacteroid-enclosing vesicles were isolated by sucrose gradient centrifugation, washed, and broken by an osmotic shock. The broken vesicles were subsequently fractionated by a second sucrose gradient centrifugation into their three constituents: the PBM, the PBF, and the bacteroid.

To purify the specific late nodulin, frozen large nodules were extracted in 50 mM NaCI, 20 mM Tris-HCI, pH 8.6, at 4°C. The extract was separated on a DEAE-Sephadex column equilibrated with the same buffer. As monitored by protein gel blotting exper- iments, the 21-kD protein did not bind to the gel. The protein was further purified by preparative SDS-PAGE, and the 21 -kD poly- peptide was electroeluted from the gel in 50 mM ammonium bicarbonate containing 0.1 O/O (w/v) SDS. SDS was removed from the sample by repeated extraction with a mixture of acetone:triethylamine:acetic acid:water (1 7:1:1:1), according to Konigsberg and Henderson (1 983).

Microsequencing Techniques

The acetone precipitate of the 21-kD protein (about 1 mg) was redissolved in 0.5 mL of 75 mM Tris-HCI buffer, pH 8.6, containing 6 M guanidine-hydrochloride. After reduction and carboxymethy- lation, the sample was extensively dialyzed against 1 ?'O (w/v) ammonium bicarbonate and digested overnight with trypsin (en- zyme-to-protein ratio, 1 :50). Resulting peptides were separated by reverse-phase HPLC on an Aquapore RP 300 column (4 to 6 x 130 mm, Applied Biosystems, Foster City, CA) equilibrated with 0.1% (w/v) trifluoroacetic acid. Elution was performed with a gradient of 10% (v/v) to 50% (v/v) acetonitrile for 60 min. Detection was at 220 nm. Each peak was collected and subjected to Edman degradation. Amino-terminal sequence determination was carried out with a pulsed liquid phase microsequenator (model 477A, Applied Biosystems). Phenylthiohydantoin derivatives were analyzed essentially as described by Fonck et al. (1986), with the exceptions that an on-line injection to the sequenator was used and the column diameter was 2 mm.

lmmunochemistry

Antibodies were produced by standard procedure in rabbits. For total nodule protein antibodies, proteins from a mixture of nodules at different developmental stages were injected. To obtain nodule- specific antibodies, the serum was adsorbed severa1 times with root proteins coupled to CNBr-activated Sepharose 4B (Pharma- cia LKB Biotechnology Inc.) in 20 mM potassium phosphate buffer, pH 7.4, containing 0.9% (w/v) NaCl (PBS) (Fortin et al., 1985). Protein gel blotting of an SDS gel on nitrocellulose was

268 The Plant Cell

performed according to Towbin et al. (1979) with antisera diluted 1/100 in PBS as primary antibodies. lmmunostaining of the protein gel blot was obtained using Sigma alkaline phosphatase coupled to goat anti-rabbit IgG (Morgan and Manen, 1985).

RNA and DNA Gel Blot Analyses

Total RNA was isolated from nodules or macroscopically unin- fected root sections as described by Prescott and Martin (1987). PolyA' RNAs were purified according to Maniatis et al. (1982). The RNA gel blot was performed with 1 pg of denaturated RNAs on GeneScreen membranes (Du Pont-New England Nuclear) ac- cording to Khandjian (1986). Plant genomic DNA was isolated according to Graham (1978) from fresh leaves, and Rhizobium total DNA was isolated according to Stanley et al. (1987). The DNA gel blot was performed on nylon membranes (Amersham). Hybridizations with 32P-ni~k translated DNA probes were done at 65OC according to Maniatis et al. (1 982). Hybridizations with 32P- kinased oligonucleotides were done at 45°C according to Wallace and Miyada (1 987).

Construction and Screening of a Nodule cDNA Library

PolyA+ RNA was isolated from nodules as described above. A cDNA library was constructed in hUni-ZAP XR using the ZAP- cDNA Gigapack II Gold synthesis kit as directed by the supplier (Stratagene). After plating of the library, 3000 phage plaques were lifted on nitrocellulose filters that were hybridized with labeled oligonucleotides under the same conditions as for the RNA gel blot hybridization with the oligonucleotide probes. The positive clones were purified and the hUni-ZAP XR phage DNA converted into pBluescript phagemids (M13 derivative) according to the manufacturer's instructions (Stratagene) for DNA sequencing and production of nick-translated probes.

DNA Sequencing Techniques

DNA sequencing was performed by dideoxy chain termination (Sanger et al., 1977) as described in the United States Biochemi- cals protocol for the Sequenase enzyme (Tabor and Richardson, 1987) and with %-ATP as the radioactive nucleotide (Amersham). The complete sequence of the two strands of the cDNA was obtained from the two ends with standard primers (universal M13 17-mer and reverse) and from the central region of the sequence with complementary oligonucleotides as primer. Sequence data were analyzed using the program IDEAS on VAX/VMS.

Assay of Enzymatic and lnhibitory Activities

The proteolytic activity was assayed with azocasein as the sub- strate. Five hundred microliters of the reaction mixture contained 250 pg of azocasein, 0.1 M Tris-HCI, pH 8.0, 10 mM CaCI2, and a fixed amount of trypsin (0.5 pg) or of a nodule extract before and after passage through the trypsin-Sepharose affinity column (200 eg and 175 pg, respectively, measured according to Lowry et al., 1951). After incubation at 37°C for 1 hr, the reaction was stopped with 500 pL of 5% (v/v) trichloroacetic acid. The precipi- tated proteins were removed by centrifugation, and the absorb- ance of the supernatant was read at 400 nm. The 21-kD protein, purified by trypsin-Sepharose affinity chromatography, was added at increasing concentrations to the reaction mixture, and the inhibitory activity was estimated by measurement of the residual caseolytic activity. Assays without incubation at 37°C were used as controls.

Electron-lmmunogold Techniques

Three-millimeter nodules were cut into small cubes (0.5 mm) in a fixation solution [5% (w/v) paraformaldehyde, 0.5% (w/v) gluta- raldehyde in 25 mM phosphate buffer, pH 7.01 and incubated at room temperature for 1 hr and then overnight at 4OC. Dehydration and embedding (Lowicryl K4M) were performed according to Roth et al. (1 981). Grids containing ultrathin sections were blocked with 0.5% (w/v) ovalbumin in PBS, reacted with the anti-21-kD anti- serum diluted 1/50 in PBS, and washed in distilled water. They were then reacted with 15 nm protein A-gold particles (Jannsen, dilution 1/200), washed in distilled water, and dried. Sections were post-stained with standard solutions of uranyl acetate (1 O min) and lead citrate (1 min). Thin sections of the same tissue were also observed by fluorescence microscopy . The autofluorescence produced by UV excitation of the senescent cells made them easily observable.

ACKNOWLEDGMENTS

This work was supported by the Fonds National Suisse de Ia Recherche Scientifique (Contract Nos. 3.1 76-0.85 and 31 -26461 - 89). We thank Carole Ghezzi and Nicole Paquet for their excellent technical assistance. We also thank Michèle Crèvecoeur for ultra- thin nodule sections and Karin Franz and Guidon Ayala (Glaxo, Genbve) for their kind gifts of oligonucleotides.

Received October 29, 1990; accepted December 31, 1990. Purification of the Protease lnhibitor on Trypsin-Sepharose

A nodule extract was dialyzed against 0.1 M Tris-HCI, pH 8.0, containing 10 mM CaCI2 and loaded onto a bovine pancreatic trypsin-Sepharose 4B column prepared according to the supplier (Pharmacia). After washing of the column with the same buffer, the protein was eluted with 0.2 M HCI, 0.5 M KCI according to Kortt (1979). The eluate was dialyzed against 1 mM HCI and lyophilized.

REFERENCES

Bergmann, H., Preddle, E., and Verma, D.P.S. (1983). Nodulin- 35: A subunit of specific uricase (uricase II) induced and local- ized in the uninfected cells of soybean nodules. EMBO J. 2, 2333-2339.

Protease lnhibitor in Senescent Nodules 269

Bosch, I., Meidl, E. J., Hoult, M., Plessner, O., and Guerinot, M.L. (1988). lron uptake and metabolism in the Bradyrhizobium- soybean symbiosis. In Nitrogen Fixation: Hundred Years After, H. Bothe, F.J. de Bruijn, and W.E. Newton, eds (New York: Gustav Fischer), p. 652.

Campos, F., Padilla, J., Vasquez, M., Ortega, J.L., Enriquez, C., and Sanchez, F. (1 987). Expression of nodule-specific genes in Phaseolus vulgaris L. Plant MOI. Biol. 9, 521-532.

Cullimore, J.V., Gebhart, C., Saarelainen, E., Mifling, B.J., Idler, K.B., and Barker, R.F. (1984). Glutamine synthetase of Phas- eolus vulgaris L.: Organ-specific expression of a multigene family. J. MOI. Appl. Genetics, 2, 589-599.

Cutting, J.A., and Schulman, H.M. (1969). The site of heme synthesis in soybean root nodules. Biochim. Biophys. Acta 192,

Dunn, K., Dickstein, R., Feinbaum, R., Burnett, B.K., Peterman, T.K., Thoidis, G., Goodman, H.M., and Ausubel, F.M. (1988). Developmental regulation of nodule-specific genes in alfalfa root nodules. MOI. Plant Microbe Interact. 1, 66-74.

Fonck, C., Frutiger, S., and Hughes, G.J. (1 986). Solvent system for the rapid identification of phenylthiohydantoin derivatives of amino acids by high-performance-liquid-chromatography. J. Chromatogr. 370, 339-343.

Fortin, M.G., Zelechowska, M., and Verma, D.P.S. (1985). Spe- cific targeting of membrane nodulins to the bacteroid-enclosing compartment in soybean nodules. EMBO J. 4,3041-3046.

Garbers, C., Meckbach, R., Mellor, R.B., and Werner, D. (1988). Protease (thermolysin) inhibition activity in the peribacteroid space of Glycine max root nodules. J. Plant Physiol. 132,

Gloudemans, T., de Vries, S.C., Bussink, H.J., Malik, N.S.A., Franssen, H.J., Louwerse, J., and Bisseling, T. (1 987). Nod- ulin gene expression during soybean nodule development. Plant Moi. Biol. 8, 395-403.

Goodchild, D.J., and Bergersen, F.J. (1 966). Electron micros- copy of the infection and subsequent development of soybean nodule cells. J. Bacteriol. 92, 204-213.

Graham, D.E. (1978). lsolation of high molecular weight DNA from whole organisms or large tissue masses. Anal. Biochem. 85,

Jacobs, F.A., Zhang, M., Fortin, M.G., andVerma, D.PS(1987). Severa1 nodulins of soybean share structural domains but differ in their subceliular locations. Nucl. Acids Res. 15, 1271-1280.

Jofuku, K.D., and Goldberg, R.B. (1 989). Kunitz trypsin inhibitor genes are differentially expressed during the soybean life cycle and in transformed tobacco plants. Plant Cell 1, 1079-1093.

Jofuku, K.D., Schipper, R.D., and Goldberg, R.B. (1989). A frameshift mutation prevents Kunitz trypsin inhibitor mRNA accumulation in soybean embryos. Plant Cell 1, 427-435.

Katinakis, P., and Verma, D.P.S. (1985). Nodulin-24 gene of soybean codes for a polypeptide of the peribacteroid membrane and was generated by tandem duplication of a sequence resem- bling an insertion element. Proc. Natl. Acad. Sci. USA 82,

Khandjian, E.W. (1986). UV cross-linking of RNA to nylon mem- brane enhances hybridization signals. MOI. Biol. Rep. 11,

486-493.

442-445.

609-61 3.

41 57-41 61.

107-1 15.

Kijne, J.W. (1975). The fine structure of pea root nodules. 2. Senescence and disintegration of the bacteroid tissue. Physiol. Plant Pathol. 7, 17-21.

Kijne, J.W., and Planqué, K. (1979). Ultrastructural study of the endomembrane system in infected cells of pea and soybean root nodules. Physiol. Plant Pathol. 14, 339-345.

Kim, S.O., Hara, S., Hase, H., Ikeaka, T., Toba, H., Kitamura, K., and Kaizuma, N. (1985). Comparative study on amino acid sequences of Kunitz-type soybean trypsin inhibitors, Tia, Tib, and Tic. J. Biochem. 98, 435-448.

Konigsberg, W.H., and Henderson, L. (1983). Removal of sodium dodecyl sulfate from proteins by ion-pair extraction. Methods Enzymol. 91, 254-259.

Kortt, A.A. (1979). lsolation and characterization of the trypsin inhibitors from winged bean seed (Psophocarpus tetragonolo- bus (L) Dc.). Biochim. Biophys. Acta 577,371 -388.

Long, S.R. (1 989). Rhizobium-legume nodulation: Life together in the underground. Cell 56, 3203-3214.

Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. (1951). Protein measurement with the folin reagent. J. Biol. Chem. 193,265-275.

Malik, N.S.A., Pfeiffer, N.E., Williams, D.R., and Wagner, F.W. (1 981). Peptidohydrolases of soybean root nodules. Identifica- tion, separation, and partia1 characterization of enzymes from bacteroid-free extracts. Plant Physiol. 68, 386-392.

Maniatis, T., Fritsch, E.F., and Sambrook, J. (1 982). Molecular Cloning, A Laboratory Manual. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory).

Mellor, R.B., and Werner, D. (1 987). Peribacteroid membrane biogenesis in mature legume root nodules. Symbiosis 3,

Mellor, R.B., and Wiemken, A. (1 988). Peribacteroid organelles as organ-specific forms of lysosomes. In Nitrogen Fixation: Hundred Years After, H. Bothe, F.J. de Bruijn, and W.E. New- ton, eds (New York: Gustav Fischer), p. 528.

Morgan, M.R., and Manen, J.-F. (1 985). Lectin variability in Phaseolus coccineus. Phytochemistry 24, 1981 -1 985.

Peoples, M.B., and Dalling, M.J. (1 988). The interplay between proteolysis and amino acid metabolism during senescence and nitrogen reallocation. In Senescence and Aging in Plants, L.D. Noodén and A.C. Leopold, eds (New York: Academic Press, Inc.), pp. 181-218.

Pfeiffer, N.E., Torres, C.M., and Wagner, F.W. (1983). Proteo- lytic activity in soybean root nodules. Activity in host cell cytosol and bacteroids throughout physiological development and se- nescence. Plant Physiol. 71, 797-802.

Pladys, D., and Rigaud, J. (1 985). Senescence in French-bean nodules: Occurrence of different proteolytic activities. Physiol. Plant. 63, 43-48.

Pladys, D., and Rigaud, J. (1988). Lysis of bacteroids in vitro and during the senescence in Phaseolus vulgaris nodules. Plant Physiol. Biochem. 26, 179-186.

Pladys, D., Trinchant, J.-C., and Rigaud, J. (1986). Proteases from French-bean nodule host-cells: In vitro effects on bacte- roids. Physiol. Veg. 24, 697-705.

Prescott, A., and Martin, C. (1987). A rapid method for the quantitative assessment of ievels of specific mRNA in plants. Plant MOI. Biol. Rep. 4, 219-224.

75-100.

270 The Plant Cell

Richardson, M., Campos, F.A.P., Xavier-Filho, J., Macedo, M.L.R., Maia, G.M.C., andYarwood, A. (1986). Theaminoacid sequence and reactive (inhibitory) site of the major trypsin inhibitor (DE5) isolated from seeds of the Brazilian Carolina tree (Adenanthera pavonina L.). Biochim. Biophys. Acta 872,

Roth, J., Bendayan, M., Carlemalm, E., Villiger, W., and Gar- avito, M. (1 981). Enhancement of the structural preservation and immunocytochemical staining in low temperature em- bedded pancreatic tissue. J. Histochem. Cytochem. 29,

Ryan, C.A. (1 981). Proteinase inhibitors. In Biochemistry of Plants, Vol. 6, A. Marcus, ed (New York: Academic Press), pp.

Ryan, C.A. (1 988). Proteinase inhibitor gene families: Tissue specificity and regulation. In Temporal and Spacial Control of Plant Genes, D.P.S. Verma and R.B. Goldberg, eds (Vienna: Springer-Verlag), pp. 223-233.

Sanger, F., Nicklen, S., and Coulson, A.R. (1977). DNA sequenc- ing with chain-terminating inhibitor. Proc. Natl. Acad. Sci. USA

Scheres, B., Van de Wiel, C., Zalensky, A., Horvath, B., Spaink, H., Van Eck, H., Zwartkruis, F., Wolters, A.M., Gloudemans, T., Van Kammen, A., and Bisseling, T. (1990). The ENOD12 gene product is involved in the infection process during the pea- rhizobium interaction. Cell 60, 281 -294.

Stanley, J., Dowling, D.N., Stucker, M., and Broughton, W.J. (1 987). Screening costramid libraries for chromosomal genes: An alternative interspecific hybridization method. FEMS Micro- biol. Lett. 48, 25-30.

134-140.

663-671.

351-370.

74,5463-5467.

of free heme in nodules. J. Biol. Chem. 262, 14730-14736. Towbin, H., Staehelin, T., and Gordon, J. (1 979). Electrophoretic

transfer of proteins from polyacrylamid gels to nitrocellulose sheets: Procedures and some applications. Proc. Natl. Acad. Sci. USA 76,4350-4354.

Trinick, M.J. (1980). Relationships among the fast-growing rhi- zobia of Lablab purpureus, Leucaena leucocephala, Mimosa spp., Acacia farnesiana and Sesbiana grandiflora and their affinities with other rhizobial groups. J. Appl. Bacteriol. 49, 39-53.

Vance, C.P. (1986). Proteolytic enzymes of legume nodule senes- cence. In Plant Proteolytic Enzymes, Vol. 2, M.J. Dalling, ed (Boca Raton, FL: CRC Press), pp. 104-124.

Vance, C.P., Johnson, L.E.B., Halvorsen, A.M., Heichel, G.H., and Barnes, D.K. (1 980). Histological and structural observa- tions of Medicago sativa root nodule senescence after foliage removal. Can. J. Bot. 58, 295-309.

Verma, D.P.S., and Nadler, K. (1 984). Legume-Rhizobium-sym- biosis: Host's point of view. In Genes lnvolved in Microbe-Plant Interactions, D.P.S. Verma and T.H. Hohn, eds (New York: Springer-Verlag), pp. 57-93.

Wallace, R.B., and Miyada, C.G. (1 987). Oligonucleotide probes for the screening of recombinant DNA libraries. Methods Enzymol. 152,432-442.

Werner, D., Morschel, E., Kort, R., Mellor, R.B., and Bassarad, S. (1984). Lysis of bacteroids in the vicinity of the host cell nucleus in an ineffective (fix-) root nodule of soybean (Glycine max). Planta 162, 8-1 6.

Wittenberg, L.J.B., Bergersen, F.J., Appleby, C.A., and Turner, G.L. (1 974). Facilitated oxygen diffusion. The role of leghemo- alobin in nitroaen fixation bv bacteroids isolated from soybean

Tabor, S., and Richardson, C.C. (1987). DNA sequence analysis with a modified bacteriophage T7 DNA-polymerase. Proc. Natl. Acad. Sci. USA 84,4767-4771.

root nodules. j. Biol. Chem: 249, 4057-4066. Yamamoto, M., H ~ ~ ~ , s., and Ikenaka, T. (1 983). ,t,,,,ino acid

sequence of two trypsin inhibitors from winged bean seeds Thummler, F., and Verma, D.P.S. (1 987). Nodulin-1 O0 of soybean (Psophocarpus tetragonolobus (L) Dc.). J. Biochem. 94,

is the subunit of sucrose synthase regulated by the availability 849-863.

DOI 10.1105/tpc.3.3.259 1991;3;259-270Plant Cell

J F Manen, P Simon, J C Van Slooten, M Osterås, S Frutiger and G J HughesA nodulin specifically expressed in senescent nodules of winged bean is a protease inhibitor.

 This information is current as of May 17, 2021

 

Permissions 8X

https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw153229

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists