QUARTERLY

The LlPRP2 gene coding for a proline-rich protein shows a high level of similarity to, as well as significant differences from the family of ENOD2 nodule-specific genes. Several sequence motifs with putative regulatory function were identified in the 5' and 3' noncoding regions of the LlPRP2 gene. Northern blot analysis revealed that the expression of the LlPRP2 gene begins 9 days after inoculation of yellow lupin roots with Bradyrhizobium sp. (Lupinus); the expression is restricted to symbiotic nodules and is not detected in other tissues or organs. Detailed hybridization analysis showed that, when expression is activated, the LlPRP2 transcript is modified so as to produce at least three bands and a continuous distribution of decay intermediates. The modification of the LlPRP2 transcript probably involves degradation from the 5'- and/or 3'-ends of the RNA molecules. Southern blot analysis indicates that only one gene is present in the yellow lupin genome. The presence of genes homologous to the LlPRP2 gene was confirmed for three cultivars of yellow lupin and for Lupinus angustifolius. However, LlPRP2 homologues were not detected in Lupinus albus cv. Bac, indicating that this plant may lack the ENOD2 sequence.

(for review see [2]).Depending on the timing of nodulin gene expression, genes are commonly classed as "early", activated during nodule formation, or "late", which are expressed in fully developed nodules [3].We are interested in studying the symbiotic interactions between the legume Lupinus luteus and its microsymbiont Bradyrhizobium sp.(Lupinus).The nodules formed on yellow lupin roots are collar-shaped, but show some similarities to indeterminate nodules in having several distinct zones [4].Our earlier studies indicated that expression of late nodulin genes is controlled in lupin by mechanisms that differ from those found in most legumes [5].Additionally, lupins also differ from other legumes in that, although they are nodulated by symbiotic bacteria, they are less frequently colonized by mycorrhizal fungi [6,7].These two symbiotic associations share some common features [8,9], including activation of the same plant genes [10,11].
Here we describe the sequence analysis and characterize the expression of an early nodulin gene from yellow lupin coding for a protein that is highly homologous to ENOD2.ENOD2 is one of the most extensively studied early nodulins.DNA sequences coding for ENOD2 proteins have been published for soybean [12,13], pea [14], alfalfa [15], Maackia amurensis [16], Sesbania rostrata [17,18] and yellow lupin [19,20].
The presumed products of ENOD2 genes are proline-rich proteins built up of two repeating pentapeptides.Based on the presence of proline-rich motifs in the sequence of the putative protein, ENOD2 nodulins have been postulated to be cell wall components [12].ENOD2 gene expression is activated between the 7th and 10th day after inoculation but the expression pattern varies from one type of nodules to another.In determinate nodules, e.g., those formed on soybean roots, ENOD2 genes are transiently expressed [21].In contrast, in indeterminate nodules, which form on pea or alfalfa roots, ENOD2 expression remains at a constant level during the de-velopment of the nodule [14].ENOD2 gene expression takes place in ineffective nodules, as well as in bacteria-free nodules of alfalfa when induced by auxin transport inhibitors or by exopolysaccharide-deficient Sinorhizobium meliloti mutant strains [22].Moreover, ENOD2 mRNAs accumulate in both S. rostrata [17,23] and alfalfa [24] roots in the absence of rhizobia in response to cytokinin treatment.
The LlPRP2 gene from yellow lupin, described in this paper, differs significantly in structure as well as expression pattern from ENOD2s described for other legumes.Additionally, we have found that in one lupin cultivar a gene homologous to ENOD2 appears to be missing.This finding might provide interesting insights into ENOD2 function in symbiotic nodule development.

MATERIALS AND METHODS
Plant materials.Yellow lupin (Lupinus luteus) cultivar Ventus plants were grown, inoculated with Bradyrhizobium sp.(Lupinus) (strain 3045 USDA, Beltsville, MD), and the nodules were collected as described earlier [25].Lupin plants (L.luteus cvs.Ventus, Parys, Popiel; L. angustifolius cv.Bar; L. albus cv.Bac) used for genomic DNA isolation were grown under the same conditions but without the inoculation step.DNA sequence analysis.DNA and protein sequences were analyzed using DNASIS and PROSIS computer programs (Hitachi Software Engineering Co.).A multiple sequence alignment was performed using the CLUSTAL W program [26].DNA and aminoacid sequences were compared with sequences present in GenBank and EMBL databases using the BLAST algorithm [27].
Southern and Northern blot analysis.Genomic DNA from 3-day-old, uninoculated lupin roots was isolated as described earlier [28].After additional purification on a cesium chloride gradient, the DNA was digested with EcoRI and/or HindIII restriction enzymes.Genomic DNA from Bradyrhizobium sp.(Lupinus) was isolated according to the protocol described in [29] and then digested with HindIII.The digested DNA was fractionated on a 0.8% agarose gel and blotted onto a nitrocellulose membrane (Schleicher and Schuell).Hybridization was performed with 32 P-labeled probes at 50°C for 48 h in a buffer containing 5 ´SSC, 1 ´Denhardt's solution, 0.1% SDS and 100 mg/ml of salmon sperm DNA.After hybridization, the filters were washed in 1 ´SSC/0.1% SDS buffer at 50°C.
Total RNA was isolated from segments of roots or from nodules at different stages of development as described by de Vries et al. [30].The poly(A) + fraction was isolated from total RNA using the Dynabeads Biomagnetic Separation System according to the manufacturer's manual (DYNAL International).RNA was separated on 1.5% agarose gel containing formaldehyde [31], transferred onto a nitrocellulose membrane and hybridized with 32 P-labeled probes.The hybridization experiments were performed at 42°C in the same buffer as the Southern hybridization except that it also contained 50% formamide.Filters were washed under high stringency conditions: 0.1 ´SSC at 55-65°C, depending on the probe.
The DNA probes, 3¢-and 5¢-end specific as well as LENOD2 cDNA, were 32 P-labeled using a "Random Primed DNA Labeling Kit" according to the manufacturer's instructions (Boehringer Mannheim).
5¢-End transcript mapping.The synthetic primer (sttr1: 5¢-GTTTGCCAGCACCGGAGG-GGT -3¢), complementary to the sequence located between 49 and 69 bp downstream from ATG codon in the LlPRP2 gene, was used in the primer extension reaction.The reaction was performed on total RNA isolated from 21-day-old nodules.The sttr1 primer was radio-labeled with [g-32 P]ATP and T4 polynucleotide kinase (Amersham) and then incubated with RNA in annealing buffer (0.4 M NaCl, 50% formamide, 10 mM Pipes, pH 6.4, and 1 mM EDTA, pH 8.0) overnight at 35°C.After precipitation, the nucleic acids were resuspended in reverse transcriptase buffer.Then, 0.5 mM 4 ´dNTPs, and 50 units of M-MLV Reverse Transcriptase (Promega) were added to the mixture.After 2 h incubation at 37°C, the RNA was degraded by ribonuclease and the sample was subjected to electrophoresis in a denaturing 8% polyacrylamide gel together with the sequencing leader initiated with the same primer.

Analysis of the LlPRP2 gene
The LlPRP2 gene was isolated from a yellow lupin genomic library using as a probe a cDNA LENOD2 that codes for a protein showing sequence similarity to ENOD2 nodulins [19,20].This gene encodes a putative protein of 103 kDa with a signal sequence at the N-terminus followed by a stretch of proline-rich pentapeptide repeating units.These units belong to two groups, appearing one after the other within the protein sequence.The sequence of the first motif, PPHEK, is identical to that of the two known motifs from the ENOD2 peptide.However, the second amino-acid motif from the LlPRP2 putative protein is less conserved, converging with the second motif characteristic for ENOD2 (PPEYQ) only in the first two proline residues.Also other motifs found in known proline-rich nodulins are absent from the amino-acid sequence of the LlPRP2 protein.A comparison of the nucleotide and deduced amino-acid sequences of the LlPRP2 gene with various nodulin sequences reveals the highest level of homology to the ENOD2 gene family (62-63% identity).However, this level of homology is lower than that between other proteins deduced from known ENOD2 sequences (72% identity between ENOD2 proteins from soybean and S. rostrata).Moreover, the deduced LlPRP2 protein is more that 2.5 times as large as those of ENOD2 described to date [20].
In the LlPRP2 promoter region, several putative regulatory motifs were found.These include: a 150 bp duplicated fragment that forms a direct repeated sequence, and a sequence characterized by a stretch of repeated (TA) 10 [20].In position -97, a 38 bp fragment was identified that has 66% identity to the OSE (Organ Specific Element) from the soybean lbc3 gene [32] (Fig. 1A).This fragment possesses both characteristic sequence motifs with single substitutions (AAAGcT and CcC-TT).These sequences are separated from each other by 5 bp as it occurs in the promoter of the pea Pslb gene.Interestingly, the localization of these motifs, -111 bp from the transcription initiation site for the CTCTT sequence, is similar to that in leghemoglobin genes [33].It is worth noting that this motif is characteristic for late nodulin genes from several legumes, but it is absent in the promoters of late nodulin genes from lupin, i.e. leghemoglobin I [34] and glutamine synthetase [35].
Our analysis of the promoter sequences of ENOD2 genes resulted in identification of a 9-bp element that immediately precedes the AGGA-box found in the ENOD2 gene promoters of soybean [12] and S. rostrata [17].This new motif, composed exclusively of A and T residues (7 and 2, respectively), is identical in the ENOD2 and LlPRP2 promoter regions (Fig. 1B).The analogous sequence was absent in the corresponding position of other prolinerich proteins encoding genes: ENOD12 and MtPRP4 [36].However, the AGGA-box, which is characteristic for ENOD2 genes, is not conserved in the LlPRP2 promoter sequence and has only 58.3% homology, compared to 91.7% between genes from soybean and Sesbania (Fig. 1B).
In the 3¢-end UTR region of the LlPRP2 gene, we found a high degree of similarity to several plant and animal sequences deposited in GenBank, with homology ranging from 60% to 70%.However, all of those sequences are located in the intron parts of these genes.Further analysis of this region revealed the presence of a putative intron (56 bp long), which covers the polyadenylation signal and the site where poly(A) is added to mRNA.Additionally, in the 3¢ untranslated region of the LlPRP2 gene, an 11-bp GT-rich element in a position similar to those found in of ENOD2 sequences was identified.This element is also present in sequences of MtPRP4 and ENOD12 from pea but is further downstream from the stop codon (Fig. 1C).It is worth noting that the GT-rich element is more conserved in ENOD12 than in ENOD2 genes from pea and that it is not present in ENOD12 sequences from other legumes.This element may be involved in stabilization of the RNA molecule [37], or may be a part of the signal for polyadenylation of mRNA [38].

Analysis of the LlPRP2 gene expression
The expression of the LlPRP2 gene was examined in different plant organs grown under different conditions.A dot-blot RNA hybridization analysis revealed the presence of LlPRP2 mRNA only in symbiotic nodules and not in other parts of the plant (Fig. 2).During nodule development, LlPRP2 transcript begins to accumulate 9 days after inoculation with Bradyrhizobium sp.(Lupinus) and its presence is detectable at a high level until the nodule senescence (Fig. 2).A more detailed analysis with total RNA separated on a gel revealed the presence of one dominant transcript of about 3000 nt in size, which is in agreement with the size of the transcribed region of the LlPRP2 gene (Fig. 3A).
However, in addition to the dominant band, a continuous distribution of smaller RNAs, in which there are at least two additional bands (of about 1600 nt and 1200 nt), was observed (Fig. 3A, arrows).The same pattern of hybridization was observed in several additional experiments using RNA isolated by two different methods (not shown).To verify that the ENOD2-hybridizing RNAs are not a product of degradation caused by the preparation method and/or manipulation of the RNA samples, several other probes including leghemoglobin I (Fig. 3B) and LlR18A were used.In all these experiments, only the LlPRP2 probe gave the characteristic hybridization pattern of several bands and an RNA smear.Figure 3B shows a non-degraded, single transcript hybridization against the same RNA with a lupine leghemoglobin I cDNA.The LlPRP2 gene is expressed earlier than the Lb gene during a symbiotic nodule development.
To exclude the possibility of non-specific hybridization of RNA with repeated sequences present in the cDNA LENOD2 clone, two specific probes, one for the 5¢-end and the other for the 3¢-end, were prepared by PCR on the LlPRP2 gene (Fig. 4).The specificity of each probe was confirmed by hybridization with HindIII-digested lupin genomic DNA, resulting in each case in single-band hybridization (not shown).
Figure 5 illustrates the results of hybridization of RNA isolated from 21-day-old nodules with the 5¢-and 3¢-end specific probes.The hybridization pattern after incubation with the 3¢-end specific probe (Fig. 5A) is similar to that observed after hybridization with the complete cDNA probe (Fig. 3A).Interestingly, after hybridization with the 5¢-end specific probe (Fig. 5B), a set of degradation products was observed, but it was less intense and consisted of numerous discrete bands.The use of such specific probes eliminated the possibility of nonspecific hybridization with those RNA species, which code for other repetitive proline-rich proteins.Both sets of hybridiza-   tion experiments clearly show that, in nodule tissue, there exists a pool of RNAs, which differ in size but possess the specific, terminal 5¢and/or 3¢-end sequences of mRNA encoded by the LlPRP2 gene.
A different pattern of hybridization was obtained when poly(A) + RNA samples from 21-day-old lupin nodules were probed.Hybridization with the 3¢-end specific probe showed only one band, which was separated by a gap from a large number of continuously migrating smaller mRNAs, in which we could not detect any distinct bands (Fig. 5C, gap in hybridization signal is indicated by a bracket).In contrast, only one band was detected after hybridization with the 5¢-end specific probe (Fig. 5D).From the Northern hybridization experiments we conclude that, among the pool of RNAs specifically hybridizing to the LlPRP2 probe, only one transcript which is polyadenylated and possesses both the 5¢-and 3¢-ends, can be detected.This transcript corresponds to the full-length mRNA of the LlPRP2 gene (Fig. 5, arrow).All other signals detected on the RNA gel blots appear to represent degradation intermediates of the LlPRP2 transcript.To additionally confirm this hypothesis, we conducted Northern hybridization experiments using single-stranded probes repre-senting sense and antisense strands of the LENOD2 cDNA.The hybridization signals were observed only after hybridization with antisense probe (not shown).The obtained hybridization data strongly suggest that specific degradation of LlPRP2 gene transcript occurs during nodule development.

Mapping of the 5¢-end of the LlPRP2 transcript by primer extension reaction
A primer extension reaction was performed on total RNA to determine the transcription initiation site of the LlPRP2 gene.The reaction was also designed to answer the question whether the transcripts are degraded from the 5¢-end by exonucleases or whether an endonuclease, which would cut-off a 5¢-end fragment of mRNA, is involved in this process.
The putative transcription initiation site (+1) was identified as the 15th nucleotide preceding the start of translation.On the autoradiogram, one dominant band was observed (not shown).Therefore, we conclude that an enzyme, which probably digests the fragment from the 5¢-end downstream from the primer alignment site (at least 63 bp), is likely to be involved in the degradation of the Vol.47 The LlPRP2 gene from yellow lupin 377 LlPRP2 transcript.If an exonuclease was involved in the degradation process, additional fragments corresponding to shortened mRNA molecules would have been generated in the primer extension reaction.

Southern hybridization with genomic DNA
The results from Northern hybridization experiments indicated that a pool of transcripts, which are specific for the LlPRP2 gene, exists in nodule tissue.However, as shown in Fig. 6A, only one hybridizing DNA fragment was detected on a genomic Southern.Genomic DNA isolated from Bradyrhizobium sp.(Lupinus) did not hybridize to the probe.Moreover, a PCR reaction on yellow lupin genomic DNA with primers surrounding the coding region of the LlPRP2 gene resulted in amplification of a single band (not shown).The PCR and hybridization results indicate that only one LlPRP2 gene is present in the genome of yellow lupin.
In addition, the results of Southern hybridization show that a gene with high homology and of a size similar to LlPRP2 gene is present in three of the four analyzed cultivars of yellow lupin (Fig. 6B).Also, we identified a positive signal in the genomic DNA isolated from L. angustifolius, but the size of hybridizing DNA fragment was smaller.Surprisingly, using the same conditions of hybridization (see Material and Methods), we were unable to detect any positive signal in the lane corresponding to genomic DNA from L. albus cv.Bac (Fig. 6B).

DISCUSSION
The LlPRP2 gene from yellow lupin encodes a protein with significant homology to the proteins encoded by the ENOD2 gene family.The gene lacks introns, and the putative protein contains a signal peptide that is almost identical to those found in ENOD2 sequences [20].Additionally, the promoter sequence shows the presence of an AT-rich element, which so far was found only in ENOD2 genes.Moreover, like in the case of ENOD2, LlPRP2 gene expression precedes activation of the leghemoglobin gene.On the other hand, one of the pentapeptide motifs found in the LlPRP2 amino-acid sequence was not homologous to any of those found in other proline-rich nodulins.Additionally, the putative protein encoded by the gene from yellow lupin is two-and-a-half-times as large as the ENOD2 putative proteins described to date [20], which makes the LlPRP2 gene distinct from members of the ENOD2 gene family.We found that LlPRP2 gene expression was restricted to symbiotic nodules.In addition, we found that a large number of RNAs specifically hybridized to the LlPRP2 probes.From the latter results we conclude that: u The LlPRP2 gene is transcribed into a full-length RNA, which is polyadenylated and to which corresponds the majority of the signal detectable after hybridization experiments.u The transcript is modified to produce shorter RNAs as soon as the expression of the LlPRP2 gene is detectable.u A comparison of results presented on Fig. 5C and 5D indicates that the LlPRP2-hybridizing transcripts are probably degraded from the 5¢-end, most likely by a mechanism involving the removal of the 5¢-end part of the transcript by a specific endonuclease, followed by exonuclease digestion.This mechanism might be similar to that proposed for degradation of oat phytochrome A mRNA [40].From the primer extension experiments, we conclude that a specific endonuclease cuts off at least 63 nt from the 5¢-end of the LlPRP2 RNA molecule (behind the primer alignment site).However, the presence of the "gap" in the signal observed after hybridization of poly(A) + fraction of RNA with the 3¢-end specific probe (see Fig. 5C, bracket) suggests that the removed part of the transcript might be much larger.non-polyadenylated transcripts.The origin of the two distinct bands of 1600 nt and 1200 nt in size is not known.The control of mRNA stability plays a fundamental role in the regulation of gene expression in plants [41].Among the many analyzed-to-date ENOD2 genes from different legumes, only the yellow lupin gene shows this unusual mechanism of posttranscriptional regulation.Moreover, in most plant mRNAs no naturally-occurring intermediates can be detected on gel blots, and only two exceptions to this rule have been well characterized: soybean ribulose-1,5-bisphosphate carboxylase small-subunit (SRS4) mRNA [42] and oat phytochrome A (PHY A) mRNA [40].
One of the characteristic features of the ENOD2 genes is their tissue-specific expression.In soybean, alfalfa, and pea, ENOD2 transcripts were localized to the nodule parenchyma (inner cortex), close to the bacteriafilled area in the central part of the nodule [14,22].In contrast to all known ENOD2 genes, the LlPRP2 gene is expressed in the infected central tissue [43], and not in the nodule parenchyma.It is possible that the expression of the LlPRP2 gene in yellow lupin nodules is up-regulated in the "young" infected cells (prefixation zone) and is later turned off (i.e. the transcript is degraded) in "older" cells.This could explain the unusual pattern of hybridization with LlPRP2-specific probes.Another interesting result was the surprising observation that a DNA fragment homologous to ENOD2 was not detected in L. albus cv.Bac.Because of the high sequence homology of the LlPRP2 gene to ENOD2, it is possible that the genome of L. albus cv.Bac lacks an ENOD2 sequence.An analogous situation was described for the alfalfa cv.Iroquois genome, where the ENOD12 gene was absent in the genome without effecting on nodulation [44].In our case, L. albus cv.Bac would be the first legume described to be lacking an ENOD2 gene.Thus, finding questions whether ENOD2 is essential for the nodulation process (the Bac cultivar is readily nodulated) or is there a protein that could take over the function of ENOD2 proteins.However, additional experiments will have to be conducted to answer those questions.
The results presented in this paper clearly show that the first full-length early nodulin gene isolated from yellow lupin has several unique sequence and expression features.Although the LlPRP2 gene shows high similarity to ENOD2 nodulins, it shows also many significant differences.The data presented here confirm the hypothesis that yellow lupin is a distinctive legume which differs from other legumes not only in its symbiotic characteristics (structure and development of the nodule and resistance to mycorrhizal colonization) but also on the molecular level.

Figure 1 .
Figure 1.Analysis of LlPRP2 DNA sequence.(A) Alignment of the OSE sequence from the lbc3 gene with the promoter fragment of the LlPRP2 gene.The conserved OSE motifs (AAAGAT and CTCTT) are shown as shaded boxes.The deviations from the lbc3 OSE sequence are written in lowercase letters.The numbers represent the distance from the transcription start site.(B) Sequence comparison of the promoter region of the LlPRP2 gene with homologous fragments of the GmENOD2A and B (acc. nos.X16875 and X16876) and the SrENOD2 (acc.no.X63339) promoters.The "AGGA" motif is shown as an open box, and the A+T-rich element is shown as a shaded box.The asterisks indicate the base pairs that are identical in all sequences.(C) 3¢-End sequence alignment of the LlPRP2 gene with analogous sequences of PsENOD2 (acc.no.X51987), SrENOD2-3A and -3B (acc.no.M23314 and M23315), SrENOD2 (acc.no.X63339), GmENOD2A and B (acc. nos.X16875 and X16876), MaENOD2 (acc.no.AF039708), PsEnod12A and B genes (acc.nos.M60585 and X57232), and MtPRP4 (acc.no.L23504).The shadowed box indicates the conservative GT-rich fragment.The numbers represent the distance from the translation termination codon.Dashes indicate gaps that have been introduced

Figure 2 .
Figure 2. Analysis of the LlPRP2 gene expression in different organs of yellow lupin.Dot-blot hybridization of RNA (20 mg) isolated from different stages of nodule development and from different plant organs of yellow lupin.LENOD2 cDNA was used as a probe.The positive signals were present only in "dots" representing RNA isolated from nodules.Numbers indicate days after inoculation with Bradyrhizobium sp.(Lupinus).

Figure 3 .
Figure 3. RNA gel blot analysis of total RNA from yellow lupin nodules.Total RNA (25 mg) was fractionated by denaturing gel electrophoresis, blotted to nitrocellulose membrane, and hybridized with the radiolabeled cDNA LENOD2 probe (A).The arrows indicate the dominant RNA band of the size corresponding to that of the transcribed region of the LlPRP2 gene (3000 nt) and additional signals of about 1600 nt and 1200 nt in size.The control hybridization of the same RNA samples with cDNA leghemoglobin I (B) shows the lack of a degradation intermediates.(C) Picture of the RNA gel used in (A) and (B) showing equal loading of RNA samples.

Figure 4 .
Figure 4. Schematic representation of all molecular probes used in hybridization experiments.The whole cDNA fragment shown covers the 3¢-end half (1550 bp) of the transcribed region of the LlPRP2 gene.The probe specific for 5¢-end of LlPRP2 gene contains 40 bp of the promoter, 5¢-end untranslated region and the sequence coding for signal peptide.The 3¢-end specific probe contains a DNA fragment (56 bp) coding for non-repetitive end of the protein and 147 bp of the 3¢-end untranslated sequence.

Figure 5 .
Figure 5. RNA gel blot analysis with the LlPRP2 5¢-and 3¢-end specific probes.RNA samples isolated from 21-day-old yellow lupine nodul were fractionated on an agarose-formaldehyde gel, transferred to nitrocellulose, and probed with 32 P-labeled 3¢-and 5¢-end specific probes.The lanes (A) and (B) represent hybridization of total RNA with the 3¢-and 5¢-end specific probes, respectively.The lanes (C) and (D) represent hybridization of the poly(A) + RNA fraction with the 3¢-and 5¢-end specific probes, respectively.The arrow indicates the RNA band appearing in all analyzed lanes, its size corresponds to the size of the LlPRP2 transcribed region.The bracket in C indicates the "gap" in the signal between the major band and the continuously migrating decay intermediates (see text).

Figure 6 .
Figure 6.Southern blot analysis of lupin genomic DNA.(A) Lupinus luteus cv.Ventus genomic DNA was digested with EcoRI (lane 1) and HindIII (lane 2) and loaded at 10 mg per lane.A 2 mg sample of Bradyrhizobium sp.(Lupinus) genomic DNA (lane 3) was digested with HindIII.The hybridizing band appeared only in lanes 1 and 2, with plant DNA.The same results were obtained after hybridization with cDNA LENOD2 and the 5¢-end specific probe (not shown).The numbers on the right show the sizes of the DNA molecular marker.(B) Lupinus luteus cv.Parys (lane 1), Popiel (lane 2), Ventus (lane 3), Lupinus angustifolius cv.Bar (lane 4) and Lupinus albus cv.Bac (lane 5) genomic DNA were digested with HindIII, loaded at 8 mg per lane and hybridized with the LENOD2 cDNA probe.The lane corresponding to L. albus cv.Bac genomic DNA lacks a hybridization signal.The numbers between the columns show the sizes of the DNA molecular marker.Picture of the genomic DNA used for hybridization in (B) shows the amount of DNA loaded in each lane.