a block in the endocytosis of rhizobium allows cellular differentiation in nodules but affects the...

Plant Molecular Biology 9: 185 - 196 (1987) 0 Martinus Nijhoff Publishers, Dordrecht - Printed in the Netherlands 185

A block in the endocytosis of Rhizobium allows cellular differentiation in nodules but affects the expression of some peribacteroid membrane nodulins

Nigel Morrison and Desh Pal S. Verma* Centre for Plant Molecular Biology, Department of Biology, McGill University, 1205 Docteur Penfield Avenue, Montreal, Quebec H3A lB1, Canada (*author for correspondence; present address: Biotechnology Center, Ohio State University, Columbus, OH 43210-1002, USA)

Received 2 February 1987; accepted in revised form 7 May 1987

Key words: Glycine max, nitrogen fixation, nodule development, nodulin gene expression

Abstract

A transposon-induced mutant (T8-1) of Bradyrhizobium japonicum (61A76) was unable to develop into the nitrogen-fixing endosymbiotic form, the bacteroid. Comparison between this mutant and T5-95, an ineffective (non-nitrogen fixing, Fix’) mutant, confirmed that the process of bacteroid development is a distinct phase of differentiation of the endosymbiont and is independent of nitrogen fixation activity. The T8-1 mutant was able to induce normal-size nodules which differentiated two plant cell types and contained numerous infection threads. However, the infected cells were devoid of bacteroids. Electron m icroscopy revealed that the ends of the infection threads were broken down in a normal manner once the thread had penetrated the cells, but the mutant was not internalized by endocytosis. The lack of peribacteroid membrane (PBM) in nod-. ules induced by this mutant was correlated with a reduced level of expression of plant genes coding for PBM nodulins. These genes were expressed in the T5-95 mutant, showing that the low expression in T8-1 was not due to the lack of nitrogen fixation. One of the PBM nodulins, nodulin-26, was found at normal levels in the nodules which lack PBM, suggesting that there are at least two developmental stages in PBM biosynthe- sis. These data suggest that a coordination of plant and Rhizobium gene expression is required for the release and internalization of bacteria into the PBM compartments of infected cells of nodules.

Introduction

Nitrogen fixation in legumes is carried out by endosymbiotic bacteria in their morphologically and physiologically distinct form known as bacteroids. As the bacteria penetrate the root tissue via “infection threads”, they induce cell division in the cor- tex, providing foci which become nodule primordia [l]. The subsequent ontogeny of nodule development has been studied in a number of legumes [I, 61. In soybean, the central region of the nodule differentiates into two cell types and the infection

thread releases bacteria into one of these cell types, the infected cell. The release of bacteria occurs via dissolution of the infection thread wall, followed by the development of a “release vesicle” or “unwalled droplet” in which bacteria are in contact with the host cell plasmalemma (see refs. [31] and [34]). Subsequently, the bacteria are taken up into the plant by a poorly understood mechanism resembling endocytosis.

As the bacterial cell is released from the infection thread, it differentiates into the nitrogen-fixing bacteroid [28]. During this “uptake”, the plant cell

186

plasma membrane surrounding the infection thread encloses the bacterium and becomes the peribacteroid membrane (PBM), keeping the bacteroid outside the host cytoplasm [31]. This membrane serves as an interface between the bacterium (which exists as a nitrogen-fixing organeIle) and the host cell cytoplasm [35]. A number of nodule-specific plant proteins (nodulins) are specifically targeted to the PBM [9] and in strains where PBM breaks down, failure of symbiosis occurs [37]. Since major nodulins are not induced without the proper development of nodules [32], they can be used as markers for the differentiation of nodule cell types and the development of the peribacteroid membrane.

In Rhizobium, genes essential for the early infection process have been cloned from a number of species (see refs. [7] and [14]) and loci have been identified which affect normal infection events and lead to nodules devoid of bacteroids [8]. How- ever, genes required for efficient bacterial release

and bacteroid development have not been described, despite the obvious importance of these processes for the establishment of a nitrogen-fixing symbiosis, We describe here the molecular cloning, phenotypic and genetic characterization of a locus in Bradyrhizobium japonicum which is essential for the endocytotic release of bacteria and its con- sequence for nodulin gene expression and development of the peribacteroid membrane.

Materials and methods

Bacterial strains and microbiological methods

Bradyrhizobium japonicum (61A76) was obtained from the Nitragin Co. (Milwaukee). The mutants T5-95 and T8-1 were isolated after random Tn5 mutagenesis of strain 61A76 followed by plant screening and acetylene reduction assays [23]. The R. fredii mutant MU042 in a Tn5-induced mutant of strain USDA 191 [27]. Bradyrhizobium strains were maintained on yeast-extract mannitol agar with 0.1070 bromothymol blue indicator. R. fredii strains were grown in liquid or solid peptone-yeast extract medium [22]. Kanamycin (Kan) and streptomycin (Str) were added to media each at 200/~g/ml. E.

coli strains were grown in Luria broth at 37 °C.

Plant growth

Glycine max cv. Prize seeds were obtained from Strayer Seed Farms (Iowa) and cv. Williams was provided by Dr H. Keyser (USDA). Seeds were sterilized by treatment with 10o7o sodium hypo- chlorite [12]. Inoculum was prepared by harvesting late log cultures, washing twice in sterile water and concentrating 15 times before application to the seeds. Plants were grown in a controlled environ- ment chamber as described [12] using autoclaved, washed vermiculite either in Leonard Jar assem- blies for individual mutant screening or in sterile pots when growing plants for RNA analyses. Culti- var Williams was used with MU042 to induce pseu-

donodules as described [27]. For all other tests, cul- tivar Prize was used. Nitrogen fixation was meas- ured using the acetylene reduction assay on the whole plant root system [13].

Light and electron microscopy

Nodule tissue was cut into 1-mm pieces and fixed in glutaraldehyde-paraformaldehyde at 4°C over- night [34]. The fixed tissue was washed thoroughly in 0.1 M sodium cacodylate buffer (pH 7.2) and os- micated in lo70 osmium tetraoxide in cacodylate buffer for 1 h. Following extensive washings, the tissue was dehydrated in an ethanol series over 10 h and embedded in Spurr resin by slow infiltration over two days and polymerization at 65 °C over- night. Thick sections were prepared for light microscopy, stained with toluidine blue and pho- tographed using a Zeiss photomicroscope. Thin sections were mounted on nickel grids with carbon- coated Formvar films, stained with uranyl acetate and phosphotungstic acid and examined using a Philips-200 electron microscope. Microscopy was done on 10-, 14- and 21-day-old nodules.

RNA isolation and hybridization conditions

Total plant polysomes from nodule tissue were pre-

pared as previously described [36]. RNA was purified from resuspended polysomes by sequential ex- traction with phenol-chloroform (1:1) and chloroform-isoamyl alcohol (24:1), followed by ethanol precipitation. For dot-blot analysis [12], various RNA preparations were serially diluted and adjusted to a final concentration of 1 mg/ml by mix- ing with yeast tRNA as filler to give a total of 12/~g RNA per sample. Blots were prehybridized over- night in hybridization buffer (900 mM NaCI, 90 mM sodium citrate, 0.1 mg/ml bovine serum albumin, 0.1 mg/ml Ficol 400, 0.1 mg/ml polyvinylpyrolli- done, 0.2 mg/ml denatured salmon sperm DNA, 0.1 mg/ml poly(A), pH 7.0) at 65 °C and hybridized with 32p-labelled purified insert DNAs from nodulin cDNA clones. Filters were washed at 65 °C in 2 × SSC (300 mM NaCI, 30 mM trisodium citrate, pH 7.0) for two periods of 30 min followed by two 1-hour washes in 0.1 x SSC, 0.1°70 SDS at 65°C. Blots were exposed to preflashed Kodak XAR X-ray film for various times without intensify- ing screens. Resulting images were quantitated by densitometry in the region of linear film response using a Bio-Rad densitometer coupled to a Hewlett-Packard integrator. Spots were also cut from the filters and quantitated by liquid scintilla- tion counting. Data were then plotted as hybridized activity against RNA concentration and the ratios of slopes were used to quantitate the relative expression of nodulin genes in different tissues [12].

RNAs from free-living broth-grown B. japonicum 61A76 and from bacteroids were prepared as described [5]. Bacteroids were prepared by pelleting through two consecutive sucrose cushions [9].

Bacterial DNA preparation and recombinant DNA techniques

Bradyrhizobium strains were grown on YM medium for five days at 28 °C. Cells were harvested and washed twice in Tris-EDTA buffer, and DNA was prepared essentially as described [23]. Recom- binant plasmids, propagated in E. coli, were prepared by the method of Clewell and Helinski [4].

187

DNA sequencing and analysis

The Tn5 junction and flanking regions (XhoI/EcoRI) were subcloned in pUC18. These plasmids were subsequently cut with HpaI and a set of Bal31-generated nested deletions were subcloned in M23mp18 vectors and sequenced by the dideoxy chain intermination method [24].

Results

T8-1-induced nodules do not contain bacteroids

Nodules formed by the Tn5 mutant, T8-1, were normal in size and appeared at the same time as wildtype 61A76 on the main root of soybean. These nodules were, however, ineffective (Fix-) in nitrogen fixation. Ultrastructural studies revealed that there were almost no bacteroids inside the infected cells.

The mutant T8-1 derived from strain 61A76 is not an auxotroph since it grows at the same rate as 61A76 in Bishops' minimal medium in the presence of Kan and Str. Whereas the wildtype nodules develop a pink colour by 14 days after infection (in- dicative of leghemoglobin accumulation), the T8-1-induced nodules are a pale green colour; and by 21 days, they turn bright green inside. Spec- troscopic and gel-electrophoresis examination of the leghemoglobin (Lb) contents suggest that there is only 1007o of the normal level of apo- leghemoglobin in T8-1-induced nodules (see be- low).

Nodules formed by T8-1 mutant develop differentiated cells comparable in size to those of wildtype nodules (see Fig. 1A), "infected cells"; however, they are devoid of bacteroids. The nucleus is enlarged as in wildtype infected cells and the central vacuole is broken down into smaller ones. An- other mutant (T5-95), which also produced Fix- nodules, formed nodules that were cytologically identical to the wildtype with abundant bacteroids in the infected cells (Fig. 1B). In addition, these nodules were red in colour and contained Lb.

188

Fig. 1. Light and electron microscopy of nodules induced by mutants T8-1 and T5-95. A and B are light micrographs showing the differ- ence in morphology between nodules induced by T8-1 (A) which contain few if any bacteroids but contain almost normal-size fully differentiated cell types and (B) T5-95 nodules which have an essentially normal morphology. (C) Electron microscopy of T8-1 induced nodules shows cytoplasm containing numerous profiles of vesicles fusing with the tip of the infection thread (cf. ref. [2]). (D) T5-95 nodules show abundant bacteroids enclosed in peribacteroid membranes with an essentially normal ultrastructure. Both types of nodules contained prominent amyloplasts which are characteristic of ineffective symbiosis. Bar is 10 #m in A and B and 1 #m in C and D. uc, uninfected cells; ic, infected cells; n, nucleus; v, vacuole; b, bacteroid; pm, plasma membrane; tw, thread wall; pbm, peribacteroid membrane.

189

Fig. 2. Ultrastructure of infection thread dissolution in T8-l-induced nodules. (A) Longitudinal section through an infection thread as it penetrates the infected cell. The fibrillar structure of the infection thread wall (tw) becomes disorganized and is possibly internalized in small vesicles as proposed by Bassett et al. [2]. Towards the end of the thread, invading bacteria (b) are in direct contact with the plasma membrane (pm) around the tip of the infection thread. (B) Tangential section through a release vesicle at the end of an infection thread. Wall material is almost absent and the plasma membrane appears like 'pbm'. v, vacuole. Bar represents 1 #m.

We examined bacterial release from infection

threads in detail, since a block at this stage would

lead to the absence of bacteroids. As shown in

Figs. 1C and 2A, T8-1 formed infection threads

which appear normal in that the infection thread

wall loses continuity, bringing invading bacteria in

contact with plasmalemma. So-called "release vesi-

cles" or "unwalled droplets" are formed (Figs. 2A

and B) which allow the bacteria to come in contact

with the plant cell membrane. Mutant T8-1 fails to

proceed beyond this stage; neither the internaliza-

tion of the rhizobia nor the subsequent prolifera-

tion of bacteroids enclosed in PBM occurs. We ob-

served occasionally infected cells with a few

bacteria escaping from the infection threads, aver- againg two to four bacteria per sectioned cell, either

naked or enclosed in PBM. In comparison, several

hundred bacteroids per sectioned cell can be seen in

wildtype nodules. It is apparent from the cytology of mutant T5-95

that the development and maintenance of a popu- lation of bacteroids is not dependent on nitrogen fixation. It is only after T5-95 is released and de-

velops into bacteroids that its symbiotic lesion be-

comes apparent. The mutation carried by T8-1

seems to inactivate a bacterial function that is required for the endocytotic event.

Cellular commi tment in nodules is independent

o f bacteroid formation

The fact that the mutant T8-1 forms nodules of

normal size with two cell types indicates that the

process of nodule organogenesis is independent of

the release of Rhizobium inside the infected cell. The uninfected cell type in soybean is specialized

for nitrogen assimilation and is characterized by

the presence of large peroxisomes having nodule-

specific uricase [3, 21]. This protein, nodulin-35,

serves as a marker for this differentiated cell type.

The mutant T8-1 produces uninfected cells contain-

ing small peroxisomes (Fig. 3A) and nodulin-35 is detected in the total cytoplasmic protein (Fig. 3B

and Western blot data, not shown). Thus, the induction of nodulin-35 is independent not only of nitrogen fixation but also of the presence of bac-

teroids. This nodulin is induced concomitant with

190

Fig. 3. (A) Electron micrograph showing ultrastructure of the uninfected cell type (uc) in T8-1-induced nodules with characteristic prominent peroxisomes (p). The infected cell type (ic) is essentially devoid of bacteroids but occasionally a few bacteria escape the block (see open arrowheads). (B) SDS-PAGE of soluble protein from wildtype and mutant nodules. (1) Molecular weight markers, (2) Wildtype 3-week-old nodules (61A76), (3) T5-95-induced nodules, (4) T8-1-induced nodules, ss, sucrose synthetase (see ref. [29]); u, uricase [2];

lb, leghemoglobin.

cellular specialization during nodule morphogene- sis [30]. Nodulin-100 (sucrose synthetase, ref. [29]) and Lb also appear to follow this development since both are present in T8-1-induced nodules (al- beit at low levels) but are absent in MU042-induced pseudonodules lacking infection threads and bacteroids (Table 1 and ref. [27]).

Expression of nodulin genes in mutant-induced nodules

During normal nodule organogenesis, the development of bacteroids requires considerable membrane synthesis as many thousands of bacteroids in the infected cell are surrounded by a new membrane compartment, PBM. We investigated whether the failure of mutant T8-1 to be internalized and pack- aged in PBM was due to a reduced capacity of the plant to synthesize PBM proteins. Since the Fix- mutant T5-95 produces abundant PBM-enclosed bacteroids, it serves as a control for any effects that

Table 1. Expression of soybean nodulin genes in various Rhizobium mutant-induced nodules as a percentage of wildtype levels.

Nodulin T5-95 T8-1 MU042

Leghemoglobin (Lb) 37 8 0.0 Nodulin-23 100 26 1.0 Nodulin-24 100 10 1.0 Nodulin-26 120 90 0.1 Nodulin-27 100 13 0.0 Nodulin-26b 79 33 1.0 Nodulin-35 48 20 0.8 Nodulin-44 76 15 0.1

Increasing amounts (0.3 to 12 #g) of total polysomal RNA from nodules (wildtype and mutant) was dot-blotted (see Materials and methods) and probed with each nodulin sequence as previously described [12]. The probes for Lb, nodulin-23, nodulin-24, nodulin-27 and nodulin-44 were pLB23, pNodA25, pNodC60, pNodB45 and pNodAl5 respectively [11, 20]. cDNA clones for nodulin-26, nodulin-26b and nodulin-35 are those described recently [10, 15, 21]. Representative data for nodulin-24 and nodulin-26 at 3/~g RNA level is shown in Fig. 4.

191

Fig. 4. Representative RNA dot-blot hybridizations with two peribacteroid membrane nodulin probes. Total polysomal RNA (3 ~tg each) from nodules formed by strains (1) 61A76 (wildtype), (2) T5-95, (3) T8-1 and (4) MU042 were blotted onto GeneScreen and hybridized with (A) nodulin-26 and (B) nodulin-24 cDNA probes as described in Materials and methods.

the lack of nitrogen fixation may have on PBM and other nodulin gene expression. Another mutant, MU042, which affects nodule differentiation (Dif - ) and fails to produce infection threads and bacteroids [27] was also used for comparison.

Table 1 summarizes the results of RNA dot-blot analyses and presents the expression of nodulin genes as a percentage of wildtype levels. Figure 4 shows two representative data points at 3 #g RNA levels for nodulin-24 and nodulin-26 expression in various types of nodules. Gene expression of the nodulins examined in the Dif - mutant-induced nodules is close to background levels (about 1%), suggesting that these nodulins are not necessary for the elaboration of non-differentiated pseudonodules. The expression of two PBM nodulins, nodulin-23 and -24, is at 100% of the wildtype level in T5-95-induced nodules (which make normal PBM-enclosed bacteroids). However, the RNA levels for these nodulins are strongly depressed in mutant T8-1-induced nodules. These low levels of PBM nodulins correlate with a few bacteria being internalized in these nodules.

Nodulin-26, a recently described nodulin of PBM [10], was found to be expressed at essentially normal levels in T8-1 nodules (Table 1). These data suggest that there are at least two stages of PBM nodulin induction and that the expression of

nodulin-26 is independent of the endocytotic release of bacteria.

The levels of expression of Lb genes, 37°70 in T5-95 and 8070 in T8-1, approximately correspond to estimates of leghemoglobin apoprotein on SDS- PAGE gels (Fig. 3B). Similarly, levels of nodulin-35 detected in protein gels and immunoblots of nodule extracts from T5-95 and T8-1 (data not shown) correspond approximately to RNA levels, suggesting that the regulation of expression of these two nodule-specific genes is at the transcriptional level [21, 33].

Cloning of the locus in Rhizobium affecting endocytosis

Rhizobium genes essential for endocytotic release and bacteroid development have not yet been described. We attempted to analyse Tn5 flanking sequences from the mutant T8-1 in which the infection process is apparently blocked at the bac-

terial release stage. The Tn5-containing EcoRI fragment of T8-1 genomic DNA was cloned (pT8-1) in pBR322 and this fragment (7.9 kb) was used as a probe against a Southern blot of EcoRI-cut genomic DNA from 61A76 and T8-1. It hybridized with a 2.2 kb fragment in 61A76 and a 7.9 kb fragment in T8-1 (data not shown). Since Tn5 is 5.7 kb [20], it appeared that a single insertion had occurred in the 2.2 kb fragment of 61A76. Insert DNA from pT8-1 was used to probe a cosmid clone bank of 61A76 DNA (partially cut with HindlII) prepared in the vector pVK102 [17]. Two homologous cosmids (p16G2 and p9B4) were isolated and these had over- lapping inserts of about 18 kb. Both cosmids contained a 2.2 kb EcoRI fragment which hybridized to the pT8-1 insert. Figure 5A and B shows hybridization data with p16G2.

To verify that Tn5 insertion had indeed caused the mutant phenotype, we initially tried com- plementation in trans by transferring the cosmid (p16G2) into T8-1 and assaying for a return to an effective phenotype. However, we could not detect stable inheritance of p16G2 in T8-1 by selecting for Tet r (coded by p16G2). We therefore took the alter- native approach of reconstructing the mutation by

192

Fig. 5. Hybridization of Tn5-containing EcoRI fragment cloned from T8-1 with the wildtype 2.2 kb EcoRI fragment from the cosmid p16G2. (A) Agarose gel of (1) EcoRI-cut p16G2 DNA and (2) lambda HindIII-size markers. (B) Autoradiograph of hybridization of the gel in panel A with the Tn5-containing EcoRI fragment from pT8-1. (C) Map of the Tn5-containin'g EcoRI fragment and sequencing strategy. Subclones of the XhoI to EcoRI Tn5 junction fragments were made in pUC18. These plasmids were subsequently cut with HpaI and a set of Bal31-generated nested deletions were subcloned in M13mpl8 and sequenced. Wildtype sequence represents that derived from reconstituting flanking areas of Tn5. Arrowhead shows the position of Tn5 insertion and arrow indicates the direction and position of the larger open-reading frame.

double reciprocal recombination in the wildtype strain and assaying for a loss of effectiveness.

The T8-1 insert was recloned on the suicide vector pSUP202 and a recombinant plasmid (pRjl0) was transformed into the mobilizing strain E. coli SM10 [26]. Strain SM10 containing pRjl0 was then mated with a fresh culture of 61A76. Since Tn5 ex- presses both Kan r and Str r in B. japonicum 61A76 but only Kan r in E. coli, we used streptomycin to counterselect against the E. coli donor. After five days, a frequency of transfer of Kan r Str r to 61A76 o f 10 -4 was obtained. Control matings with E. coil SM10 were done (since this strain has a chro- mosomally integrated Kanr marker) and no transfer of Kanr was found under identical conditions. pSUP202 was used as a probe in colony hybridization tests to distinguish between single and double

crossovers among Kanr Str r exconjugants which still retained vector sequences. This was necessary since the other marker on pSUP202 (Amp r) is not phenotypically expressed in 61A76. Of the 240 colonies screened, 15 did not hybridize to the vector sequences. Of the five transconjugants assayed by Southern hybridization using pT8-1 insert DNA, three appeared to result from random transposition of Tn5 from the incoming pRjl0, since they all had the wildtype 2.2 kb EcoRI fragment and each had another hybridizing band (presumably containing Tn5) of variable size. One transconjugant did not

have the wildtype fragment but had a 7.9 kb hybridizing band and was therefore identical to TS-1, brought about by a double recombination event between pRjl0 and the 61A76 chromosome.

The reconstructed mutant (61A76 pRjl0-5) and

193

one of the random transposition derivatives (61A76 pRjl0-3) were inoculated onto soybean (20 plants each) with appropriate controls. At 14 days, acetylene reduction assays showed high nitrogen fixation activity for 61A76- and 61A76 pRjl0-3-inoculated plants while there was no activity for 61A76 pRjl0-5 and T8-1-induced nodules even after a 24-hour incubation. The effective nodules were pink inside while the reconstructed mutant bacteria and T8-1 both induced nodules with green interi- ors. Bacteroids could not be observed in the nodules of the reconstructed mutant (61A76 pRjl0-5).

These results prove that the Tn5 insertion in the 2.2 kb EcoRI fragment is the direct cause of the mutant phenotype of T8-1 and strongly suggest that gene(s), essential for endocytosis and subsequent bacteroid development are located on this re-

gion of DNA. We could not detect in Northern blots any transcript from this region of DNA during nodule development (10-21 days), indicating that the affected gene is expressed at a very low level or expressed very early in the infection process.

Since we could not detect a transcript from this DNA fragment during nodule development, the DNA sequence of a 1.4 kb region spanning the site of insertion of Tn5 in mutant T8-1 was determined to find out whether there was a potential coding region. Figure 5C shows the point of insertion of Tn5 and the sequencing strategy. The site of insertion of Tn5 was confirmed by the presence of a 9 basepair direct repeat characteristic of Tn5 insertion event [25]. Computer analysis of 1447 basepairs revealed a long open-reading frame encoding 249 amino acids (Fig. 6) spanning the site of insertion of Tn5.

10 20 30 40 50 60 70 • 80 90 100 GACGCAACCGGTCCGcATCGTTGcGACAGCGTACGGCATCCGAGCGCGCTCATTGCCTTGTCGCGcTCGCCCATCACGAGATAGGCGGACCAGCCGCAGC

I I0 120 130 140 150 160 170 180 190 200 CAGCCCTCGACGTCGTCGCCATTGGTcTTcAGCCGcGTCGCCAGCCGATcGACCATGCCTTTGATCATGGTGCCGCGATCGGcCTCGCTCATGTCCCTGG

210 220 230 240 250 260 270 280 290 300 CCGCAGcCACCGCAcTGTCGGCACGCGGTGccTGAGCCCACCGGGCGCCTGACGCTGCCGGGCAGCCTGCCGCCGACCcGCAcCAGCCGGCCTGCACCAG

310 320 330 340 350 360 370 380 390 400 CGGCCGCCACGGCGCGTC•GCCGGCGcCTTCGCCAGCAGcGGCTGCCAGAGGCCCGCGCGTCGCCTGCCTCGGCGGCAAGGCCGAGGAAATAGTTCGCCT

410-91---- I 420 430 440 450 460 470 480 • 490 500 TGGCATCGTCGCCATTCAAGCCGATTGCGCGCTCGAACTcCGcCTTGGCGTCCGCGGTGACGACGcCGCCAGCAGTCCcCATCAG•GcCTcCGAGATCGG

510 520 530 540 555 570 CGCGGCGAGTCGCGTGTCGCCGGCATAGGTG ATG GCT TGC GAT AGG CGC GGA CGG CGT CAT CGT AGC GGC CGA GGC GCG ACA

Met Ala Cys Asp Arg Arg Gly Arg Arg His Arg Ser Gly Arg Gly Ala Thr 585 600 615 630 645 GCA CGG GGC CAG CAC CGT CCA GCC GCG GCC GTC GGT CGG TTC TTT TCG AGA TGC GCC TCG ACC TGC GCG AGC CAG Ala Arg Gly Gln His Arg Pro Ala Ala Ala Val Gly Arg Phe Phe Set Arg Cys Ala Ser Thr Cys Ala Ser Gln 660 675 690 705 720 ATT GGC GAG CGG TCG GTT GCA TCG GCC ACC GCG CGC TCG GCG AGG GAA AGT CAC CGA GGC GCG GCG AGC CGA GCG l le Gly Glu Arg Ser Val Ala Set Ala Thr Ala Arg Set Ala Arg Glu Ser His Arg Gly Ala Ala Set Arg Ala 735 750 765 780 ~ ~ 795 CGA GAT AGA AGC CCG AGG CAA TCA CCG GCA ATC GCG ACC AGG CCG ACG ATC GCA TCG CCG CAG ATA GGG CTG CCT Arg Asp Arg Ser Pro Arg Gln Ser Pro Ala l le Ala Thr Arg Pro Thr l le Ala Ser Pro Gln l le Gly Leu Pro 810 V 825 840 855 870 GCA CCG GGT CTC ACG CCC GCT CTG CAT CGG CGG CGG CCA GCA GCC GCG GCT GAT CTC GAC CCG CGC GGC GTC GGC Ala Pro Gly Leu Thr Pro Ala Leu His Arg Arg Arg Pro Ala Ala Ala Ala Asp Leu Asp Pro Arg Gly Val Gly 885 900 915 930 945 CTC GGC GGC ACC GAT GAT CCC GGC CGC CAT GTC GCG GTC GAT TTC GGC GAG CTG GTC GCG ATA GAC CAC GAC CTC Leu Gly Gly Thr Asp Asp Pro Gly Arg His Val Ala Val Asp Phe Gly Glu Leu Val Ala l le Asp His Asp Leu 960 975 990 1005 1020 GCT GCC GCC GGT ATC CGC GCG CCG CCC GGG CGG CTG AGC GGC CAC AGC ACG GCA AAG ATC GCC GCG ACC GTC ATC Ala Ala Ala Gly l le Arg Ala Pro Pro Gly Arg Leu Ser Gly His Ser Thr Ala Lys l le Ala Ala Thr Val l le 1035 1050 1065 1080 1095 AGC GCG AAC ACA AAC CAC AGC GTC ATC AAG CGG TTC CGG ATG GAA CGG CGG CTT CGG CAA GCG CTT AGA ATC GCT Set Ala Asn Thr Asn His Ser Val l le Lys Arg Phe Arg Met GIu Arg Arg Leu Arg Gln Ala Leu Arg l le Ala

i i i 0 1125 1140 1155 1170 TTA CAT ACG GCG GCG AGC CGG CAA TTG ACA TTG CTG GAT CGA CCG AAG ACA AAA TCA GCG GTT TCC GGT GCC GTT Leu His Thr Ala Ala Ser Arg Gln Leu Thr Leu Leu Asp Arg Pro Lys Thr Lys Ser Ala Val Ser Gly Ala Val

1185 1200 1215 1230 1240 CCG GCG GTA TCG CAA CAG CTT ACC TGG CCA AGG CGC ACC CGA CCA AAA TAT CGA AAA CAA CCC ATG CAA AGG AGC Pro Ala Val Ser Gln Gln Leu Thr Trp Pro Arg Arg Thr Arg Pro Lys Tyr Arg Lys Gln Pro Met Gln Arg Set

1260 1275 1290 1300 1310 1320 1330 1340 1350 TGG CGG CCG CGG CGC TCA GAT TGA CCGTTGACACGTCGGGCAAATCAGCCCCACACTTCATTATTCGAAATTTCGGCAAATGCTTGCTGTGGC Trp Arg Pro Arg Ar9 Ser Asp *

1360 1370 1380 1390 1400 1410 1420 1430 1440 1447 GCAGGGCAATCGCATATGGGCTCGGTCAGcTcCAcAAACTGTCATGCCCGGCCTTGTGGCATCCACGTCTTGTCTCCGcAAcAGAAAGACGTGCATG

Fig. 6. DNA sequence of the region where Tn5 was inserted, interrupting a potential gene involved in bacterial release and development of bacteroids. The position of the Tn5 insertion is shown by a large arrowhead and is derived from sequencing the junction fragment (see Fig. 5C). Two open-reading frames on the opposite strands coding for peptides of ll4 amino acids and 102 amino acids respectively are indicated by arrows and closed circles indicating the termination of these peptides. In addition, there are other possible open-reading frames on the opposite strand.

194

In addition, two open-reading frames of 114 and 112 amino acids, starting at nucleotide positions 787 and 415 respectively, exist in the reverse orienta- tion. Search for the E. coli type or Rhizobium symbiotic gene promoter sequences did not lead to the identification of any putative promoters. No signifi- cant protein sequence homologies of these coding sequences were found with those in the protein data- bank.

Discussion

We have shown that the expression of peribacteroid membrane nodulins is differentially influenced by a mutation in B. japonicum which prevents the internalization of Rhizobium inside the host cell and the development of bacteroids. The data suggest that a gene in B. japonicum is expressed prior to or during the release event affecting endocytosis and/or bacteroid proliferation. Concomitantly, full induction of nodulins 23 and 24 (and possibly other genes) may be required for efficient "packaging" of invading bacteria into peribacteroid membranes. Nodulin-24 [16] and nodulin-23 [19] have been characterized to be nodulins of PBM [15].

Microscopic analysis showed abundant membrane structures derived from the Golgi complex in the infected cell cytoplasm of T8-1-induced nodules, suggesting that the general membrane synthesis ability of the plant is not inhibited by the mutant. It may be that key factors, such as the membrane nodulins, which are specifically targeted to the PBM, are not synthesized, causing an accumulation of vesicles which are possible precur- sors of PBM. Only nodulin-26 is expressed at essentially normal levels in nodules induced by T8-1. The expression of this nodulin is, therefore, uncom- promised by this mutation in Rhizobium while the expression of other nodulins of PBM is affected. These data suggest that there are at least two different events in PBM nodulin gene induction and that maximal expression of nodulin-26 is required for or induced by events prior to the point at which the mutation in T8-1 causes a block in nodule development.

Nodulin-27 also followed the pattern of expres-

sion of the PBM nodulins 23 and 24, being max- imally expressed in T5-95 nodules and depressed in T8-1 nodules. The induction of nodulins which have reduced expression in T5-95 nodules is proba- bly influenced by events subsequent to bacteroid development.

The mutant T8-1 did not affect the development of the two plant cell types within the nodule. Although the marker proteins for the two cell types (nodulin-35 and Lb) were expressed at reduced levels, cytological examination indicated a clear differentiation of the larger "infected" cell and the smaller "uninfected" cell which contains prominent peroxisomes. These data further suggest that the synthesis of nodulin-35 (nodule-specific uricase) begins prior to the commencement of nitrogen fixation [21]. Initial induction of nodulin-35 appears to be influenced by a process of cellular specialization in the nodule [30].

Cell commitment does not seem to be affected by a block in bacteroid formation. The ultrastructural features of the two cell types and the marker proteins are absent from the nodules induced by the Dif- mutant MU042 [27]. These mutant nodules are similar to other recently described (ndv) mutants [18] and "uncoupled" exopolysaccharide- deficient mutants of R. meliloti [8] which induce undifferentiated, bacteria-free nodules. The fact that Lb and nodulin-35 are present in nodules which have infection threads (T8-1) and are absent in nodules lacking infection threads (MU042) sup- ports the hypothesis that the differentiation of the two cell types of the normal nodule requires the physical presence of bacteria in the infection thread [27]. Since nodulin genes were not expressed above background levels in MU042-induced nodules, we can conclude that the expression of any of the known nodulin genes is not required for the elaboration of the pseudonodule structure. Whether there is a causal relationship between the presence of the infection thread and nodulin gene induction remains to be proven.

Acknowledgements

We would like to thank Christine Morrison for her

perseverance in this work, Prakash Sista for the pVK102 cosmid bank of 61A76, Marc Fortin for providing PBM cDNA clones and critical com- ments, Fritz Thummler for help with spectroscopy, Maria Neuwirth for help with sectioning and electron microscopy, Ashton Delauney for reading the manuscript and Yvette Mark for her excellent secretarial assistance. This research was supported by an Operating and a Strategic research grant from the Natural Sciences and Engineering Research Council of Canada.

References

1. Bauer WD: Infection of legumes by rhizobia. Ann Rev Plant Physiol 32:407-449 (1981).

2. Bassett FJ, Goodman RN, Novasky A: Ultrastructure of soybean nodules. I. Release of rhizobia from infection thread. Can J Microbiol 23:573-582 (1977).

3. Bergmann H, Preddie E, Verma DPS: Nodulin-35: A subunit of nodule-specific uricase (uricase II) induced and localized in uninfected cells. EMBO J 2 :2333- 2339 (1983).

4. Clewell DB, Helinski DR: Supercoiled circular DNA-protein complex in Escherichia coli: Purification and induced con- version to an open circular DNA form. Proc Natl Acad Sci USA 62:1159-1166 (1969).

5. Cooper TG, Whitney P, Magasnik B: Reaction of Lac- specific ribonucleic acid from Escherichia coli with Lac deoxyribonucleic acid. J Biol Chem 249:6548-6555 (1974).

6. Dart PJ: Infection and development of leguminous nodules. In: Hardy RWF, Silver S (eds) A Treatise on Dinitrogen Fix- ation, Sec III. Wiley-Interscience Pub, New York, pp 367-472 (1977).

7. Downie JA, Knight CD, Johnston AWB, Rossen L: Identifi- cation of genes and gene products involved in the nodulation of peas by Rhizobium leguminosarum. Mol Gen Genet 198:255-262 (1985).

8. Finan TM, Hirsch AM, Leigh JA, Johansen E, Kuldau GA, Deegan S, Walker GC, Signer ER: Symbiotic mutants of Rhizobium rneliloti that uncouple plant from bacterial differentiation. Cell 40:869-877 (1985).

9. Fortin MG, Zelechowska M, Verma DPS: Specific- targetting of membrane nodulins to the bacteroid-enclosing compartment in soybean nodules. EMBO J 4: 3041-3046 (1985).

10. Fortin MG, Morrison N, Yerma DPS: Nodulin-26, a peribacteroid membrane nodulin is expressed independently of the development of the peribacteroid compartment. Nucl Acids Res 15:813-824 (1987).

11. Fuller F, Kunstner PW, Nguyen T, Verma DPS: Soybean nodulin genes: Construction and analysis of eDNA clones reveals several tissue-specific sequences expressed in

195

nitrogen-fixing root nodules. Proc Natl Aead Sci USA 80: 2594-2598 (1983).

12. Fuller F, Verma DPS: Accumulation of nodulin mRNAs during the development of effective root nodules of soybean. Plant Mol Biol 3." 21-28 (1984).

13. Hardy RWF, Holsten RD, Jackson EK, Burns RC: The acetylene-ethylene assay for nitrogen fixation: Laboratory and field evaluation. Plant Physiol 43:1185-1207 (1968).

14. Jacobs TW, Egelhoff TT, Long SL: Physical and genetic map of a Rhizobium meliloti nodulation gene region and nucleotide sequence of nodC. J Bacteriol 162:469-476 (1985).

15. Jacobs F, Zhang M, Fortin M, Verma DPS: Several nodulin genes of soybean derived from a common ancester during recent evolution. Nucl Acids Res 15:1271-1280 (1987).

16. Katinakis P, Verma DPS: Nodulin-24 gene of soybean codes for a polypeptide of the peribacteroid membrane and was generated by tandem duplication of a sequence resembling an insertion element. Proc Natl Acad Sci USA 82: 4157-4161 (1985).

17. Knauf VC, Nester EW: Wide host range cloning vectors: A cosmid clone bank of an Agrobacterium Ti plasmid. Plas- mid 8 : 4 5 - 5 4 (1982).

18. Dyland T, Lelpi L, Stanfield S, Kasbyap L, Douglas C, Yanofsky M, Nester E, Helinski DR, Ditta G: Rhizobium meliloti genes required for nodule development are related to chromosomal virulence genes in Agrobacterium tumefa- ciens. Proc Natl Acad Sci USA 83:4403-4407 (1986).

19. Mauro VP, Nguyen T, Katinakis P, Verma DPS: Primary structure of nodulin-23 gene and potential regulatory ele- ments in the 5' flanking regions of nodulin and leghemoglobin genes. Nucl Acids Res 3 :239-249 (1985).

20. Meade HM, Long St, Ravkun GB, Brown SE, Ausubel FM: Physical and genetic characterization of symbiotic and aux- otrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis. J Bacteriol 149:114-122 (1982).

21. Nguyen T, Zelechowska MG, Foster V, Bergmann H, Verma DPS: Primary structure of the soybean nodulin-35 gene encoding nodule-specific uricase localized in peroxisomes of uninfected cells of soybean. Proc Natl Acad Sci USA 82: 5040- 5044 (1985).

22. Olson E, Sadowsky M, Verma DPS: Identification of genes involved in the Rhizobium-legume symbiosis by Mu-dI (Kan,lac)-generated transcription fusions. Bio/Technology 3:143-149 (1985).

23. Rostas K, Sista PR, Stanley J, Verma DPS: Transposon mutagenesis of Rhizobium japonicum. Mol Gen Genet 197: 230-235 (1984).

24. Sanger F, Nicklen S, Coulson AR: DNA sequencing with chain terminating inhibitors. Proc Natl Acad Sci USA 74: 5463-5467 (1977).

25. Shapiro JA: Molecular model for the transposition and replication of bacteriophage mu and other transposable ele- ments. Proc Natl Acad Sci USA 76:1933-1937 (1979).

26. Simon R, Priefer U, Puhler A: A broad host range mobiliza- tion system for in vivo genetic engineering: Transposon

196

mutagenesis in gram negative bacteria. Bio/Technology 1: 784-791 (1983).

27. Stanley J, Longtin D, Verma DPS: A genetic locus in Rhizo- bium japonicum 6fredit) affecting soybean root nodule differentiation. J Bacteriol 166:628-634 (1986).

28. Sutton WD, Pankhurst CE, Craig AS: The Rhizobium bacteroid state. In: Giles KL, Atherly AG (eds) International Review of Cytology, Supp 13, Biology Rhizobiaceae. Aca- demic Press, New York, pp 149-177 (1981).

29. Thummler F, Verma DPS: Nodulin-100 of soybean is the subunit of sucrose synthetase regulated by the availability of free heme in nodules. J Biol Chem (submitted) (1987).

30. Verma DPS: Involvement of specific host genes in symbiotic nitrogen fixation. In: Chopra V e t al. (eds) Genetics: New Frontiers. Oxford and IBH Pub, New Delhi, pp 227-239 (1984).

31. Verma DPS: Host-Rhizobium interactions during symbiotic nitrogen fixation. In: Smith H, Grierson D (eds) The Molecular Biology of Plant Development. Blackwell Pub, Oxford, pp 437-466 (1982).

32. Verma DPS, Fortin MG, Stanley J, Mauro V, Purohit S,

Morrison N: Nodulins and nodulin genes of Glycine max. Plant Mol Biol 7 :51-61 (1986).

33. Verma DPS, Haugland R, Brisson N, Legocki R, Lacroix L: Regulation of the expression of leghemoglobin genes in effective and ineffective root nodules of soybean. Biochim Bi- ophys Acta 653:98-107 (1981).

34. Verma DPS, Kazazian V, Zogbi V, Bal AK: Isolation and characterization of the membrane envelope enclosing the bacteroids in soybean root nodules. J Cell Biol 78:919-936 (1978).

35. Verma DPS, Long S: Molecular biology of legume- Rhizobium symbiosis. In: Jeon K (ed) International Review of Cytology, Supp 14, pp 211-245 (1983).

36. Verma DPS, Nash DT, Schulman HM: Isolation and in vitro translation of soybean leghemoglobin mRNA. Nature 251:74-77 (1974).

37. Werner D, Morschel E, Kort R, Mellor RB, Bassarab S: Ly- sis of bacteroids in the vicinity of the host cell nucleus in an ineffective (Fix-) root nodule of soybean (Glycine max). Planta 162:8-16 (1984).

a block in the endocytosis of rhizobium allows cellular differentiation in nodules but affects the...

Documents