PHYSIOLOGIA PLANTARUM 115: 563–570. 2002 Copyright C Physiologia Plantarum 2002

Printed in Denmark – all rights reserved ISSN 0031-9317

Boron requirement in the Discaria trinervis (Rhamnaceae) and Frankiasymbiotic relationship. Its essentiality for Frankia BCU110501growth and nitrogen fixation

Luis Bolanosa,†, Miguel Redondo-Nietoa,†, Ildefonso Bonillaa and Luis G. Wallb,*

aDpto. de Biologıa, Facultad de Ciencias, Universidad Autonoma de Madrid, 28049-Madrid, SpainbDpto. de Ciencia y Tecnologıa, Universidad Nacional de Quilmes, R. Saenz Pena 180, 1876–Bernal, Argentina*Corresponding author, e-mail: luis.balarios/uam.es†These authors contributed equally to this work

Received 20 September 2001; revised 11 February 2002

The essentiality of boron (B) for nitrogen fixation in hetero-cystous cyanobacteria and rhizobial symbioses has beenwidely established. However, nothing is known about thepossible involvement of the micronutrient in actinorhizalsymbioses. Therefore, the effect of boron (B) deficiency onthe establishment of the Discaria trinervis-FrankiaBCU110501 symbiosis was investigated. Nodulation wasdiminished in B-deficient D.trinervis or in plants inoculatedwith Frankia grown in the absence of B. These poorly nodu-lated plants showed a reduction of shoot and root weight andsmall size. Because depletion of the micronutrient duringgrowth of the actinomycete altered the infection capacity ofFrankia, we also studied growth, structure and nitrogen fix-

Introduction

Boron (B) is a micronutrient essential for the developmentof higher plants, diatoms and some species of algae. How-ever, it is apparently not required by fungi and bacteria(Dugger 1983, Loomis and Durst 1992). An exception arethe heterocystous cyanobacteria (blue-green algae) whichrequire B when grown under N2-fixing conditions for thestabilization of the envelope which controls the access ofoxygen to the heterocyst (Bonilla et al. 1990, Garcıa-Gon-zalez et al. 1991). B has been reported to have several im-portant functions (reviewed by Blevins and Lukaszewski1998): cell wall structure, membranes and membrane-as-sociated reactions, reproduction, pollen tube growth, andpollen germination, nitrogen fixation, etc.

Besides heterocystous cyanobacteria, we have shownthat B is essential for N2 fixation in Rhizobium-legumesymbioses during several stages of legume root nodule de-velopment: nodule cell wall and membrane structure(Bolanos et al. 1994, Bonilla et al. 1997), legume-rhizobia

Physiol. Plant. 115, 2002 563

ation of free-living Frankia BCU110501. Growth wasdelayed in B-deficient BAP media (πN cultures), and com-pletely inhibited in B-deprived N-free BAP media (–N cul-tures), suggesting that B is required to enhance growth ofFrankia and essential for the development of nitrogen fixingactivity. Ultrastructural study of B-deficient cells showed analteration of filament walls both in πN and especially in –N cultures, indicating a possible role of the microelement inthe maintenance of these structures. Moreover, the stabilityof vesicle envelopes was impaired in the absence of B and,hence, nitrogenase occurrence and nitrogen fixation were to-tally absent. The results show that B is required for bothpartners to establish an effective symbiosis.

molecular signalling (Redondo-Nieto et al. 2001), infec-tion thread development and nodule invasion (Bolanoset al. 1996), and symbiosome development (Bolanos et al.2001). All of these effects were due to B deficiency in thehost plant, and there is no evidence so far for an essen-tiality of the micronutrient for Rhizobium growth.

A nitrogen-fixing bacterium with structural and func-tional similarities to heterocystous cyanobacteria is theactinomycete Frankia. Bacteria of the genus Frankia formso-called actinorhizal symbioses with several non-leg-uminous shrubs and trees termed actinorhizal plants,wherein the endophytic form of the microsymbiont de-velops the N2-fixing activity (Wall 2000). Actinorhizalsymbioses can be clearly differentiated from rhizobial-leg-ume symbioses although a common evolutionary originand some anatomical and physiological features are com-mon to both symbioses (Pawlowski and Bisseling 1996).Similar to cyanobacteria, but different to rhizobia,

Frankia strains isolated from nodules can fix N2 when cul-tured without a nitrogen-combined source. Nitrogenase infree-living cultures or in symbiotic state is localized insidethe specialized vesicles that differentiate from some fila-ment tips (Huss-Danell 1997). The N2-fixing vesicle is inmany ways structurally and functionally analogous to theheterocyst (Zehr et al. 1998). Therefore, based on thesimilarity to heterocysts, B could be essential not only forthe development of the actinorhizal symbioses but alsofor the differentiation of N2-fixing vesicles of Frankia asin heterocystous cyanobacteria.

Frankia BCU110501 was the first infective and effec-tive isolate from field nodules of Discaria trinervis(Chaia 1998). Frankia BCU110501 infects the host rootthrough an intercellular pathway, inducing the develop-ment of typical actinorhizal nodules (Valverde and Wall1999). This strain was used to investigate the require-ment of B for the establishment of the symbiotic interac-tion with D. trinervis, and also to study the possible es-sentiality of the micronutrient for Frankia growth, struc-ture and N2-fixing vesicle development.

Materials and methods

Plant growth

Seeds of D. trinervis (Hooker et Arnot) Reiche were steril-ized by scarification (3 min immersion in 10 M H2SO4,with occasional shaking by hand), and then exhaustivelyrinsed with sterile distilled water (Chaia 1998). After-wards seeds were blotted dry with filter paper, transferredto B-free perlite moistened with modified N-free Evanssolution, diluted to 1/10 strength (1/10 E), and kept at 4æCfor 5 days. For control plants, B was added as H3BO3 at afinal concentration of 0.1 mg B lª1. For B-deficient plants,the micronutrient was never added to the medium. To en-sure B deficiency, all solutions, media and experimentswere done in plastic containers previously tested not to re-lease B even under sterilizing conditions, as previously re-ported (Mateo et al. 1986). Furthermore, media weretreated with the boron-specific resin Amberlite IRA743from Sigma Co. (St. Louis, MO, USA) (Asad et al. 1997).Prior to use, boron was determined in the media by themethod of azomethine H at pH 5.1 (Wolf 1974), and no Bwas detected (detection limit was 0.02 mg mlª1).

Germination and plant growth were carried out in aglasshouse, with mean maximum and minimum tem-peratures of 27 and 20æC, respectively, and relative hu-midity range of 65–95%. At the cotyledonary stage (12–14 days after germination), plants were aseptically trans-ferred to growth pouches (Mega International, Minnea-polis, USA) moistened with 1/10 E. Pouches were keptin glasshouse with a photoperiod of 16 h light.

Frankia cultures

Frankia strain BCU1105 was first isolated from D. tri-nervis nodules (Chaia 1998). A pure liquid culture kindlyprovided by Dr Chaia was grown axenically at 28æC in

Physiol. Plant. 115, 2002564

Fig. 1. Rate of nodulation (% nodulated plants) (A) and noduleaverage (B) of Discaria trinervis inoculated with FrankiaBCU110501 grown in the presence (πB) or in the absence (–B) ofboron in BAP glucose (πN) or N-free BAP glucose (–N) media.Closed bars: plants grown with B; open bars: B-deficient plants.There is a statistically significant difference (P Ω 0.05) betweenplants inoculated with πB–N Frankia cells and the other treat-ments. 100% in A correspond to 20 plants.

static culture in BAP glucose medium (Murry et al. 1984).Inocula for experiments were prepared from 4-week-oldcultures. Cells were collected by centrifugation, washed 6times in media without NH4Cl and without B, and the cellsuspension was then homogenized by repeated passagethrough needles of 0.8 and 0.5 mm gauge (3 times each).Inocula with approximately 9–10 mg protein mlª1 wereused for experiments with nitrogen-containing (πN) ornitrogen-free (–N) BAP medium. For B deficiency, pre-cautions as described for B-deficient plants were taken.

Nodulation assays

To study the effect of B deficiency on the establishmentof the symbiotic interaction, germinated D. trinervis weregrown in pouches as stated above for 4 weeks, and afterthat, each seedling was inoculated by dripping a volumeof homogenate of Frankia BCU110501 on the root zone(cultured in the presence or in the absence of B) corres-ponding to 20 mg of protein (Valverde and Wall 1999).

Fig. 2. Shoot length (A), shoot fresh weight (B) and root FW (C)of Discaria trinervis plants inoculated with Frankia BCU110501grown in the presence (πB) or in the absence (–B) of boron in BAPglucose (πN) or N-free BAP glucose (–N) media. Closed bars:plants grown with B; open bars: B-deficient plants. There is a statis-tically significant difference (P Ω 0.05) between plants grown with Band inoculated with πB–N Frankia cells and the other treatments.

Two months after inoculation, nodule number, length ofstems and weight of shoots and roots was determined in8–20 replicates per treatment and per experiment.

Frankia growth estimation

Growth rate of Frankia BCU110501 was measured asprotein content determined by the bicinchoninic acid

Physiol. Plant. 115, 2002 565

method (Nittayajarn and Baker 1989), using bovineserum albumin (BSA) as standard. Assays were per-formed on three replicate samples from each experimen-tal culture.

Nitrogenase activity

Nitrogenase activity of Frankia was estimated by theacetylene-reduction assay (Postgate 1971). Nitrogenlimited cultures were incubated in a 10% acetyleneatmosphere during 4 h and the production of ethylenewas measured in a gas chromatograph Shimadzu GC-8A.

SDS-PAGE electrophoresis and immunoblotting

Samples for gel separation were extracted by heatingFrankia cells in SDS buffer for 10 min at 100æC pre-viously homogenized by passage through needles. Aftercentrifugation to remove insoluble debris, the extracts(10 mg protein loaded per lane) were subjected to 12%acrylamide minigels (Laemli 1970). Gels were trans-ferred electrophoretically to membranes of nitrocellulose(Bittner et al. 1980). Blots were incubated with 5% bo-vine serum albumin) in Tris–Buffered saline (50 mMTris-HCl, pH 7.4; and 200 mM NaCl) buffer containing1% of nitrogenase component I of Rhodospirillum rub-rum rabbit antiserum. Immunostaining was visualizedwith goat antirabbit IgG conjugated with peroxidase, asa secondary antibody.

Fluorescence microscopy

Cell viability was studied by the use of acridine orangeas fluorescent indicator (Zelenin 1999). Five ml of Fran-kia culture were centrifuged at 500 g for 10 min. Thepellets were resuspended in 50 ml of 0.9% NaCl, and in-cubated for 15 min with 50 ml of a sterile solution of

Fig. 3. Effects of B deficiency (–B) on growth kinetics measured asprotein yield of Frankia BCU110501 in BAP batch culture contain-ing ammonium (πN) or none combined nitrogen source (–N).

Fig. 4. Fluorescence micrographs of Frankia BCU110501 cultures after acridine orange staining. A: πBπN, B: –BπN, C: πB–N, D: –B–N. Arrowheads indicate vesicles; double arrowheads indicate sporangia. Bar markers: 10 mm.

0.05 mg mlª1 acridine orange in 50 mM KH2PO4, pH6.8. Fluorescence of cells was observed with a combi-nation of 435–490 nm excitation.

Electron microscopy

The structure of Frankia filaments and vesicles wasstudied by transmission and electron microscopy. Fortransmission electron microscopy, cells in culture werefixed with 2% glutaraldehyde and post-fixed with 2%KMnO4. Dehydration was carried out with a series ofwater-ethanol solutions. Samples embedded in Epon 812(60æC for 24 h) were ultrasectionated and stained withuranyl acetate. For scanning electron microscopy, dehy-dration was carried out in a series of water-acetone solu-tions. A drop of each dehydrated sample in 100% ace-tone was dried and coated with gold.

Statistical analysis

All the experiments were repeated at least three times.All data were statistically analysed by the one-way -

Physiol. Plant. 115, 2002566

test. Data in figures are means plus minus standarddeviation.

Results

D. trinervis plants grown in N-free media and in thepresence or absence of B were inoculated with FrankiaBCU110501 that had previously been cultivated in B-free (–B) or B-containing (πB) media and with (πN) orwithout (–N) ammonium as nitrogen source. Twomonths after inoculation, the highest rate of infectionwas obtained in πB plants inoculated with Frankiagrown in πB–N media (Fig. 1A). Nodulation rates werereduced in B-deficient plants and when Frankia weregrown in B-deficient conditions, mainly in –B–N media,and used as inoculum. The number of developed nodulesper plant was halved in B-deficient plants and plants ino-culated with Frankia grown in πN media (Fig. 1B). Fig-ure 1A also indicates that this Frankia strain was moreinfective when grown without N (compare πBπN andπB–N inocula treatment in πB plants) suggesting thatN is probably implicated in the expression of symbiotic

Fig. 5. Transmission electron micrographs of filaments of Frankia BCU110501 grown in πBπN (A), –BπN (B), πB–N (C), and –B–N (D)media. Bar markers: 0.2 mm.

Frankia genes, which should be a subject for a differentfuture investigation.

Growth of plants was related to the behaviour ofnodulation of each treatment. The size, root weight, andshoot weight (Fig. 2) showed a significant reduction inpoorly nodulated Discaria, including πB plants inocu-lated with –B–N bacteria, where nodule number had notbeen reduced.

These results suggest that both the plant and Frankiarequire B for a correct establishment of a symbiotic re-lationship. Since the absence of the micronutrient duringgrowth of Frankia affected nodulation and led to notfunctional nodules, the effects of B deficiency on growth,structure, and nitrogen fixation activity of free-livingFrankia BCU110501 were investigated.

Growth of Frankia (as mg protein per ml of culture) inπBπN conditions showed a typical curve for batch cul-ture described for other Frankia strains (Murry et al.1984), with a lag period, an exponential growth phase

Physiol. Plant. 115, 2002 567

followed by an abrupt cease and autolysis (Fig. 3). Whenammonium was absent from the media (πB–N), growthwas slower, probably because cultures must adapt to ves-icle development and N2 fixation. Boron deficiency re-flected an almost total inhibition of protein yield in –B–N Frankia BCU110501. There was little growth duringthe early days of the experiments, probably as a resultof B derived from the initial inoculum, but growth com-pletely stopped thereafter during the time course. Behav-iour of –BπN cultures was slightly different, becausethere was a delay in growth, compared with πBπN con-ditions, and reached a content of protein similar to πB–N and old πBπN cultures at a later time of treatment(60 days).

Using fluorescence microscopy and acridine-orange asindicator for cell viability (Fig. 4), the majority of acri-dine orange-stained πB cells emitted green colour (Fig.4A,C), while most of cells from –B cultures, includingfilaments, sporangia, and vesicles of –N cultures, fluor-

Fig. 6. Immunodetection of nitrogenase component I followingSDS-PAGE and electroblotting in Frankia BCU110501 cell extractsderived from cultures after 21 days of growth in N-free BAP mediain the presence (lane 1), or the absence (lane 2) of B.

esced orange/red (Fig. 4B,D). As a control, πB and –Bcultures were killed by heat or by dehydration in ethanol,and fluorescence of πB cells turned to red (data notshown). Therefore, red fluorescence may indicate that Bdeficiency alters cell viability of Frankia.

Ultrastructural examination by electron microscopy(Fig. 5), showed that B-deficient filaments of both πNand mainly –N treatments appeared with a highly disor-ganized structure (Fig. 5B,D). Walls seemed to be weaklypacked, and cytoplasm was almost empty comparedwith filaments developed with normal B nutrition (Fig.5A,C).

The development of nitrogen fixing activity and ves-icles was investigated because Frankia was unable togrow in the absence of B in N-free media. Nitrogen fix-ation in B-normal or B-deprived Frankia cultures wasstudied both by chromatographic determination ofARA and by immunological detection of nitrogenasecomponent I (CI) on Western blots. While ARA of cellsgrowing with B after 15 days was 94.4 ∫ 25.2 nmol ethy-lene mgª1 protein hª1, activity of –B cultures was alwaysunder the detection limit of the chromatograph. Theseresults were supported by immunological identificationof nitrogenase CI that demonstrated that there was nodetectable nitrogenase protein in cell extracts derivedfrom B-deficient treatments (Fig. 6).

The study of the ultrastructure of vesicles was inagreement with nitrogenase measurements. The electronmicroscopic study of vesicles developed in B-deficienttreatments (Fig. 7B) demonstrated a very disorganizedinner compartment and an apparently smaller ‘voidarea’ compared with vesicles from πB cultures (Fig. 7A).Moreover, the normal solid pear-like structure of thevesicle from control cultures examined by scanning elec-

Physiol. Plant. 115, 2002568

tron microscopy (Fig. 7C) disappeared in B-starved cul-tures (Fig. 7D), indicating a weak vesicle envelope.

Discussion

The aim of this study was to test the possible essentialityof the micronutrient B for actinorhizal symbiosis. Evi-dence is presented here that the absence of B resulted infailure of the D. trinervis–Frankia BCU110501 interac-tion leading to a poor plant growth and development(Fig. 2), as occurred in legume symbiosis (Bolanos et al.1994).

Nodulation was diminished in Discaria grown in theabsence of B (Fig. 1). Although little is known aboutFrankia-actinorhiza molecular signalling and cell surfaceinteractions, these results reflect that B could modulatethem, as it does during the Rhizobium-legume relation-ship (Bolanos et al. 1996, Redondo-Nieto et al. 2001).

Despite these similarities between B-deficient acti-norhizal and legume symbiosis, there is a very relevantdifference. To date, no effects of the absence of B onfree-living rhizobial strains have been demonstrated. Bycontrast, Figs 1 and 2 show that the absence of B thenutrient during growth of Frankia BCU110501 reducedthe infection capacity of the bacterium. These results in-dicated that B-starvation also altered the microsym-biont. In effect, B was required for growth of this strainin BAP-glucose and absolutely essential for growth in N-free medium (Fig. 3). Moreover, fluorescence of culturesafter acridine orange staining indicated that B deficiencyprovoked changes in the cells. Shifts of acridine orangefluorescence from green to orange/red have been usuallyreported to be a test for cellular death in several organ-isms, including the death of bacterial cells (Mason andLloyd 1997). Therefore, red fluorescence of B-deficientFrankia seemed to confirm the essentiality of the mi-cronutrient for viability of this strain, and can explainthe low rate of plants nodulated by inocula developed inthe absence of B (Fig. 1).

The physiological role of B has usually been relatedto the maintenance of structures rich in carbohydratemolecules (Blevins and Lukaszewski 1998). The specialstructure of borate anions makes them able to becomeesterified with cis-diol groups (Mazurek and Perlin1963). Because of this chemical property, B is able tointeract and stabilize glycoconjugates in cell walls andmembranes (O’Neill et al. 1996), including the envelopeof the heterocyst in cyanobacteria (Garcıa-Gonzalezet al. 1991). Similarly, observations made in the ultra-structure of filaments (Fig. 5) and vesicles (Fig. 7) ofFrankia BCU110501, suggest that B could also stabilizewall and/or membrane components rich in diols, as canbe polyhydroxyl compounds detected in lipidic fractionsof Frankia cells (Tunlid et al. 1989).

Frankia BCU110501 was unable to grow and to de-velop functional N2-fixing vesicles in N-free B-deficientmedia (Fig. 3, –B–N treatments and Fig. 6,7). The pro-tection of nitrogenase activity against oxygen diffusionis attributed to the resistance properties of the lipidic

Fig. 7. Transmission (A, B) and scanning (C, D) electron micrographs of vesicles of Frankia BCU110501 developed in the presence (A, C)or in the absence (B, D) of B. Bar markers: 0.2 mm.

multilaminate vesicle wall (Parsons et al. 1987), whichcan change its thickness by modifying the number orlipidic monolayers in response to different pO2 (Harrisand Silvester 1992). The analysis of lipids showed thatvesicles have a higher content of glycolipids and neutrallipids than vegetative cells, being the major proportionlong-chain polyhydroxy fatty acids or alcohols (Tunlidet al. 1989). A very high concentration of the hopanoidbacteriohopanetetrol is also present (Berry et al. 1991).All of these constituents of the vesicle envelope are com-pounds rich in diol groups which can interact with bo-rate ions. The wrinkled appearance of B-deficient ves-icles (Fig. 7B) is similar to that reported for B-starvedheterocysts (Garcıa-Gonzalez et al. 1991), which is dueto the loss of the inner laminated layer of the heterocystenvelope. That layer is composed of glycolipids withlong-chain polyhydroxyl alcohols (Lambein and Wolk1973) stabilized by boric acid. The narrow ‘void area’inside B-deficient vesicles (Fig. 7A), which is the place

Physiol. Plant. 115, 2002 569

were the lipidic envelope is supposed to be (Torrey andCallaham 1982), inside B-deficient vesicles showed inFig. 7A also suggest a thinner laminated envelope. Re-sults reported herein might therefore indicate that Bplays a role in the stabilization of vesicle envelope, asthe micronutrient does in the heterocysts.

Overall, these results demonstrate a B requirement notonly for Discaria to establish a symbiotic relationshipwith Frankia BCU110501, but also that the micronutri-ent is needed for Frankia vegetative growth and infectioncapacity and this B is essential when the bacteria arecultivated under nitrogen limitation. Among prokar-yotes, the essentiality of B has been reported only fornormal heterocyst development in cyanobacteria (Bonil-la et al. 1990), but not for vegetative growth. The innerlaminated layer of the heterocyst envelope stabilized byB is composed of lipids not found in vegetative cells(Nichols and Wood 1968), while the vesicle envelope isenriched in lipids that are also constitutive of filaments.

This particular difference could explain why B is alsoneeded for the structure, growth and infectivity of veg-etative cells of Frankia BCU110501.

Finally, this work can increase the list of organismsthat require B. Except for heterocystous cyanobacteria,none of the tested Gram-negative microorganisms showa requirement for the micronutrient. Since this Frankiastrain needs B, future work should focus on other ac-tinomycetes and gram-positive bacteria. The study ofother Frankia isolates and plants infected via root hairsmay be of a particular importance in order to clarifywhether or not B deficiency also affects the capacity ofthe bacteria to elicit a prenodule structure in the hostplant. These studies could also help understanding notonly the role of B but also the molecular mechanism ofrecognition between the actinorhizal symbiotic partners.

Acknowledgements – Frankia BCU110501 was kindly provided byDr E. Chaia (Centro Regional Universitario Bariloche, Argentina),and nitrogenase component I antiserum by Dr N. J. Brewin (JohnInnes Centre, Norwich, UK). This work was supported by Pro-grama Sectorial de Promocion General del Conocimiento (M.E.C)no. PB98-0114-CO2-01, and Programa de Cooperacion con Ib-eroamerica (Argentina). Luis Bolanos was granted by Comunidadde Madrid. Luis G. Wall is researcher of CONICET (Argentina).

ReferencesAsad A, Bell RW, Dell B, Huang L (1997) Development of a boron

buffered solution culture system for controlled studies of plantboron nutrition. Plant Soil 188: 21–32

Berry AM, Moreau RA, Jones AD (1991) Bacteriohopanetetrol:abundant lipid in Frankia cells and in nitrogen-fixing noduletissue. Plant Physiol 95: 111–115

Bittner M, Kupferer P, Morris CF (1980) Electrophoretic transferof proteins and nucleic acids from slab gels to diazobenzyloxy-methyl cellulose or nitrocellulose sheets. Anal Biochem 102:459–471

Blevins DG, Lukaszewski KM (1998) Boron in plant structure andfunction. Annu Rev Plant Physiol Plant Mol Biol 49: 481–500

Bolanos L, Brewin NJ, Bonilla I (1996) Effects of boron on Rhi-zobium-legume cell–surface interactions and nodule develop-ment. Plant Physiol 110: 1249–1256

Bolanos L, Cebrian A, Redondo-Nieto M, Rivilla R, Bonilla I(2001) Lectin-like glycoprotein PsNLEC-1 is not correctly glyco-sylated and targeted in boron deficient pea nodules. Mol PlantMicrobe Interact 14: 663–670

Bolanos L, Esteban E, de Lorenzo C, Fernandez-Pascual M, deFelipe MR, Garate A, Bonilla I (1994) Essentiality of boron forsymbiotic dinitrogen fixation in pea (Pisum sativum)-Rhizobiumnodules. Plant Physiol 104: 85–90

Bonilla I, Garcıa-Gonzalez M, Mateo P (1990) Boron requirementin Cyanobacteria. Its possible role in the early evolution of pho-tosynthetic organisms. Plant Physiol 94: 1554–1560

Bonilla I, Mergold-Villasenor C, Campos ME, Sanchez N, Perez H,Lopez L, Castrejon L, Sanchez F, Cassab GI (1997) The ab-errant cell walls of boron-deficient bean root nodules have nocovalently bound hydroxyprolin-/proline-rich proteins. PlantPhysiol 115: 1329–1340

Chaia E (1998) Isolation of an effective strain of Frankia from nod-ules of Discaria trinervis (Rhamnaceae). Plant Soil 205: 99–102

Dugger WM (1983) Boron in plant metabolism. In: Läuchli A, Bi-eleski RL (eds) Inorganic Plant Nutrition (Encyclopedia of

Edited by J. K. Schjørring

Physiol. Plant. 115, 2002570

Plant Physiology). Springer Heidelberg, New York, NY, pp 626–650

Garcıa-Gonzalez M, Mateo P, Bonilla I (1991) Boron requirementfor envelope structure and function in Anabaena PCC 7119 het-erocysts. J Exp Bot 42: 925–929

Harris S, Silvester W (1992) Nitrogenase activity and growth ofFrankia in batch and continuous culture. Can J Microbiol 38:296–302

Huss-Danell K (1997) Actinorhizal symbioses and their N2 fixation.New Phytol 136: 375–405

Laemli UK (1970) Cleavage of structural proteins during the as-sembly of the head of bacteriophage T4. Nature 227: 680–685

Lambein F, Wolk CP (1973) Structural studies on the glycolipidsfrom the envelope of the heterocyst of Anabaena cylindrica. Bio-chem 12: 791–798

Loomis WD, Durst RW (1992) Chemistry and biology of boron.Biofactors 3: 229–239

Mason DJ, Lloyd D (1997) Acridine orange as an indicator of bac-terial susceptibility to gentamicin. FEMS Microbiol Lett 153:199–204

Mateo P, Bonilla I, Fernandez-Valiente E, Sanchez-Maeso E (1986)Essentiality of boron for dinitrogen fixation in Anabaena sp.PCC 7119. Plant Physiol 81: 430–433

Mazurek M, Perlin AS (1963) Borate complexing by five-member-ed-ring vic-diols. Vapor pressure equilibrium and NMR spectralobservations. Can J Chem 41: 2403–2411

Murry MA, Fontaine MS, Torrey JG (1984) Growth kinetics andnitrogenase induction in Frankia sp. HPF ArI5 grown in batchculture. Plant Soil 78: 61–78

Nichols BW, Wood BJB (1968) New glycolipid specific to nitrogen-fixing blue-green algae. Nature 217: 767–768

Nittayajarn A, Baker DD (1989) Methods for the quantification ofFrankia cell biomass. Plant Soil 118: 199–204

O’Neill MA, Warrenfeltz D, Kates K, Pellerin P, Doco T, DarvillAG, Albersheim P (1996) Rhamnogalacturan-II, a pectic poly-saccharide in the walls of growing plant cell, forms a dimer thatis covalently cross-linked by a borate ester. In vitro conditionsfor the formation and hydrolysis of the dimer. J Biol Chem 271:22923–22930

Parsons R, Silvester WB, Harris S, Gruijters WTM, Bullivant S(1987) Frankia vesicles provide inducible and absolute oxygenprotection for nitrogenase. Plant Physiol 83: 728–731

Pawlowski K, Bisseling T (1996) Rhizobial and actinorhizal symbi-oses: what are the shared features. Plant Cell 8: 1899–1913

Postgate JR (1971) The acetylene reduction test for nitrogen fix-ation. Meth Microbiol 6B: 342–356

Redondo-Nieto M, Rivilla R, El-Hamdaoui A, Bonilla I, BolanosL (2001) Boron deficiency affects early infection events in thepea-Rhizobium symbiotic interaction. Aust J Plant Physiol 28,819–823

Torrey JG, Callaham D (1982) Structural features of the vesicle ofFrankia sp. Cpi1 in culture. Can J Microbiol 28: 749–757

Tunlid A, Schultz NA, Benson DR, Steele DB, White DC (1989)Differences in fatty acid composition between vegetative cellsand N2-fixing vesicles of Frankia sp. strain Cpi1. Proc Natl AcadSci USA 86: 3399–3403

Valverde C, Wall LG (1999) Time course of nodule developmentin the Discaria trinervis (Rhamnaceae)-Frankia symbiosis. NewPhytol 141: 345–354

Wall LG (2000) The actinorhizal symbiosis. J Plant Growth Regula-tion 19, 167–182

Wolf B (1974) Improvement in the Azomethine H method for deter-mination of boron. Commun Soil Sci Plant Anal 5: 39–44

Zehr JP, Mellon JM, Zani S (1998) New nitrogen-fixing microorgan-isms detected in oligotrophic oceans by amplification of nitro-genase (nifH) genes. Appl Environ Microbiol 64: 3444–3450

Zelenin AV (1999) Acridine orange as a probe for cell and molecu-lar biology. In: Mason WT (ed) Fluorescent and LuminiscentProbes for Biological Activity. Academic Press, San Diego, pp117–135