microscopic evidence on how iron deficiency limits nodule ... · lupins (lupinus angustifolius l....

New Phytol. (1992), 121, 457-467

Microscopic evidence on how iron

deficiency limits nodule initiation in

Lupinus angustifolius L.

BY CAIXIAN TANG1, A. D. ROBSON1, M. J. DILWORTH3 AND J. KUO2

Soil Science and Plant Nutrition, School of Agriculture, 2 Electron Microscopy Centre, The University of Western Australia, Nedlands, WA 6009, Australia,

3 School of Biological and Environmental Sciences, Murdoch University,

Murdoch, WA 6150, Australia

(Received 9 September 1991; accepted 3 March 1992)

SUMMARY

Lupins (Lupinus angustifolius L. cv. Yandee), grown in solution culture, have been used to study the sites and

process of infection by Bradyrhizobium sp. (Lupinus) and the impairment of nodulation by iron deficiency.

Infection leading to nodulation occurred in an area of epidermal cells either lacking root hairs or with very young

root hairs at the time of inoculation. Cells aged 13 h or over appeared not to be infected. Infection was initiated

in the outer cortex. Rare, short infection threads were evident on day 4 after inoculation, 2 d after the initial

division of cortical cells resulting from the bradyrhizobial inoculation. Bacteria had been released into the

cytoplasm of cortical cells within 5 d after inoculation. Bacteroids multiplied in the cytoplasm, segregated

passively and spread in the infection zones by repeated division of the invaded cells.

Under iron deficiency, initial cell division occurred in the outer cortex of host roots, as in iron-sufficient plants

after inoculation. Iron deficiency then limited further division of cortical cells. Only a few surviving infection sites

developed nodules with normal structure but development was much slower than in iron-sufficient plants.

Key words: Iron deficiency, Bradyrhizobium infection, Lupinus angustifolius, anatomy, ultrastructure.

INTRODUCTION

In many legumes, the first step in nodule formation

involves rhizobial-induced deformation of host root

hairs, the bacteria entering the roots by forming

infection threads originating in the curled tips of root

hairs. The infection threads, which contain bacteria

and are enclosed by a cellulose wall, develop to the

base of infected root hairs and then penetrate the

root cortex. The rhizobial symbionts stimulate host

cell division to form meristems and enter newly

divided host cells through the growth and branching

of the infection threads. Bacteria are released into

cytoplasm through a degradation of the infection

thread wall or folding of host plasmamembrane

(Nutman, 1956; Dart, 1974, 1977; Dazzo, 1980;

Meijer, 1982). In some tropical legumes, e.g. Arachis

hypogaea (Chandler, 1978) and Stylosanthes spp.

(Ranga Rao, 1977; Chandler, Date & Roughley,

1982) bradyrhizobial infection occurs at the junction

of the lateral roots without the formation of infection

threads either in the root hairs or in the nodules. The

bradyrhizobia enter the cells through the structurally

altered cell wall. In some woody legumes, epidermal

infection appears likely. Sprent (1989) describes

rhizobial penetration between epidermal cells in a

manner similar to that for Frankia infection of the

non-legume plant Elaeagnus. Rhizobial infection in

Mimosa scabrella has been confirmed as being of this

type (Faria, Hay & Sprent, 1988).

Little is known regarding the site of infection and

the processes involved in nodule initiation in lupins,

no infection threads having been observed in early

studies either in root hairs (Dart, 1977; Quispel,

1983) or in nodules (Dart, 1977). However, in

developing lupin nodules, infection threads have

been subsequently found in nodule tissue, each with

wall material continuous with the cell wall of the

plant (Robertson et al., 1978).

In lupins, iron is required in a greater amount for

nodule formation than for host plant growth (Tang,

Robson & Dilworth, 1990a). Nodules were not

formed on those roots directly exposed to a solution

with deficient iron, irrespective of whether the shoot

or the rest of the root system was iron-sufficient

(Tang, Robson & Dilworth, 1990b, 1991). Iron

This content downloaded from 134.115.74.61 on Fri, 23 Sep 2016 06:23:03 UTCAll use subject to http://about.jstor.org/terms

458 C. Tang and others

Table 1. The rate of root elongation and nodulation of

seedlings of Lupinus angustifolius inoculated with

Bradyrhizobium 3, 5 and 8 d after germination

Age of seedling Rate of root Number of at the time of elongation* nodulest inoculation (d) (mm h-1) (no. plant-1)

3 1 89+005 64+4

5 186+009 85+7

8 186+006 134+6

* The average rate between the times of inoculation and harvest, means of three replicates + standard error.

t Nodules counted 12 d after inoculation; means of 15 plants + standard error.

deficiency markedly decreased the number of nodule

initials and thereafter the number of nodules. In

iron-sufficient plants, the formation of nodule initials

commenced on day 5 after inoculation and was

completed within three days (Tang et al., 1990b).

With transfer experiments, Tang et al. (1991) found

that the impairment of the nodulation process by

iron deficiency could be attributed to the prevention

of a step at day 4 after inoculation, the stage just

before nodule initials were formed. However, how

iron deficiency interferes with the events of nodu-

lation is not understood.

The present study had three aims: (i) to identify

the location of the infection site in lupins by marking

the positions of the smallest emergent root hairs and

the root tips at the time of inoculation and examining

the distribution of nodules, (ii) to examine the time

course of the infection process using light and

transmission electron microscopy and (iii) to study

how iron deficiency impairs the process of nodulation

in Lupinus angustifolius L.

MATERIALS AND METHODS

Infection site

Seeds of narrow-leafed lupin (Lupinus angustifolius

L. cv. Yandee) were germinated on a stainless steel

screen over an aerated solution of 600 aM CaCl2 and

2/tM H3BO3 for 3, 5 or 8 d. Five seedlings were transferred to 5-1 plastic pots containing an aerated nitrogen-free nutrient solution with concentrations

(aM) of KH2PO4, 20; K2SO4, 600; MgSO4, 200; CaCl2, 600; Fe"'.NaEDTA, 2-5; 13B03, 5; Na2MoO4, 0 03; ZnSO4, 0 75; MnSO4, 1 0; CoSO4,

0 2; and CuSO4, 0-2, and roots were maintained at 20-22 ?C. At the time of transplanting, the positions

of the smallest emergent root hairs and the root tips

of the plants were determined by inspection through

a dissecting microscope and marked with a water- proof pen on pieces of plastic which were attached to

the plants. The length of the taproot and the distance

between the smallest emergent root hairs and the

root tips were recorded (Fig. 1). A suspension of

Bradyrhizobium sp. (Lupinus) WU 425 (Tang et al.,

1990a) was then added to the solution in order to

provide 3 x 105 cells ml-'. The solution was left

unchanged for 3 d and then replaced with a similar

solution with added bradyrhizobia. After another

2 d, plants were placed in solutions without added

bradyrhizobia and the solution changed every second

day. The pH of the solution was adjusted daily to 5 5

with 0.1 M KOH.

Twelve days after inoculation, plants were har-

vested and nodule distribution on the taproot was

recorded. No lateral nodules were formed at this

stage. The rate of root elongation was calculated

from the increment of root length during the period

of the experiment.

Microscopy

Lupin seeds were germinated for 5 d as above. Five

seedlings were then transferred to 5-1 plastic pots

containing an iron-free nutrient solution otherwise

identical to the above experiment and roots were

maintained at 20-22 'C. Iron was added as Fe"'-

NaEDTA at concentrations of 0 05, 0 25 or 2 5 /tM. A dense suspension of Bradyrhizobium sp. (Lupinus)

WU 425 was added to the solution. After 24 h, this

was replaced by solution without added rhizobia,

which was thereafter changed daily. A point 1 cm

behind the root tips was marked at the time of

inoculation. Daily harvests were commenced on day

1 after inoculation and continued until day 17. A

1-cm segment of root below the mark was sampled

for microscopy, and the appearance of nodules was

recorded.

Small pieces of harvested root tissues were fixed in

2-5 00 (v/v) glutaraldehyde in 0025 M phosphate

buffer (pH 7) and post-fixed in 1 Go (w/v) osmium tetroxide for 2 h at room temperature. The speci-

mens were dehydrated in increasing acetone concen-

trations and embedded in Spurr's epoxy resin

(Spurr, 1969). Cross sections of root or nodule

tissues 1-2 aM thick were stained with 0 05 0 (w/v) toluidine blue 0 (G. T. Gurr, London), pH 9, and

examined under a Zeiss research light microscope.

Serial sections were also undertaken where enlarged

root hairs were associated with the infection centre.

Ultrathin sections of selected areas were mounted on

75/300 mesh naked grids, stained with saturated

aqueous uranyl acetate and lead citrate (Reynolds,

1963), and examined under a Jeol 2000FXII electron

microscope at 80 kV.

RESULTS

Iron-adequate plants

Localization of infection site. Irrespective of the age

of the seedling at the time of inoculation, the rate of

root elongation of all plants was very similar (about

1 9 mm h-1). However, the total number of nodules


Iron deficiency and nodule initiation in Lupinus 459

co

a.> 10 ?X_ _ (a) " 10 CD

H 4 8 12 16 20 24 52 56 60 (b)

" 10 K , 64 l CD 4 8 12 16 20 24 28 32(c 64 3 0

10

z 4 8 12 1620 2428 3236 4072 76

Distance from the root-shoot transition region (cm) Direction of root growth ->

Figure 1. The distribution of nodules on the taproots of seedlings of Lutpinus angustifoliuis that were (a) 3 d, (b) 5 d and (c) 8 d old when inoculated. Symbols on the horizontal

axis are: t, the position of the root-shoot transition region of the seedlings; I, the position of the smallest emergent root hairs (SERH) at the time of inoculation; +, the position of the root tips (RT) at the time of inoculation; =, the region of taproots without root hairs at the time of inoculation; >, the position of the root tips at the time of harvest. Values are the average of 15 seedlings.

on the taproot increased as the age of seedlings at the

time of inoculation was increased. Nodules appeared

1 d earlier and the number of nodules was doubled

on 8-d-old plants when compared with those which

were 3 d old (Table 1). The higher number of

nodules on older seedlings was attributed to an

increased number of nodules in the vicinity of the

root tips (Fig. 1).

Nodules on the taproots of all plants were formed

most frequently in the region between the smallest

emergent root hairs and the root tips (Fig. 1), where

no root hairs had emerged at the time of inoculation.

A relatively small number of nodules developed in

the zones that contained developing root hairs. No

nodules developed further up the root from the

smallest emergent root hair region than a distance

equivalent to that from the zone of smallest emergent

root hairs to the root tip at the time of inoculation.

Distribution of nodules declined gradually in regions

below the root tip marks and diminished after 10

units of the interval of the smallest emergent root

hairs and the root tips from the root tip marks (Fig.

1).

Anatomy of infection. Under the light microscope the

most obvious change in root tissues resulting from

inoculation with bradyrhizobia was the division of

cortical cells. It was evident that cell division

occurred as early as 2 d after inoculation, the simplest

discernible change being the division of a hypo-

dermal cell into two daughter cells. Such cell

divisions normally occurred in the outer layers of the

cortical cells with enlarged nuclei (Fig. 2). By day 4,

cortical cells in the outer layers had divided several

times and the incidence of division moved towards

the inner cortex (Fig. 3). By day 5, cell division was

evident in the inner cortex as well. A small cluster of

newly divided cells in the outer cortex had filled with

cytoplasm, as indicated by dark staining (Fig. 4);

this was the meristem of a nodule initial and was

distinguishable after roots were cleared and stained

with Brilliant Green (Tang et al., 1990 a). Bacteroids

were visible in these cells (micrograph not shown).

By days 6 to 8, repeated cell division had occurred in

both the inner and outer cortex and the meristems of

nodule initials had enlarged. Near the meristems, the

cortical cells had divided further, distorting the

alignment of their periclinal walls (Figs 5, 6). The

meristem of nodule initials had distended the

epidermis to form visible nodules after day 9. A

distinct nodule cortex formed over the nodule

meristem and the vascular bundle connecting the

nodule vascular system with the root stele had begun

to differentiate (Fig. 7). Some nodules consisted of

more than one bacteroid zone, each of which was

separated by rows of nodule cortical cells. Uninvaded

cells in bacteroid zones were seldom observed

(micrographs not shown).

The clusters of dividing cortical cells were as-

sociated with enlarged root hairs in some cases, but

curled root hairs were rarely observed. Infection

threads could not be seen within the infection centre

by light microscopy, nor in the root hairs in serial

sections.

Ultrastructure. Structures resembling infection

threads were relatively rare and short; they were

usually found between the epidermal cells and the

outermost layer of cortical cells at early stages, and

between infected cells of the nodule meristem and

neighbouring uninfected nodule cortical cells at later

stages. Infection thread-like structures were formed

by day 4 after inoculation. An infection thread-like

structure was observed beneath an epidermal cell of

a young nodule when it was just visible on the root

surface on day 9 after inoculation (Figs 8, 9). The

wall of the infection thread-like structure was thicker

than that of host cells. The wall material appeared

continuous with the wall of cortical cells and seemed

to have a similar electron density. The infection

thread-like structure did not appear to cross cells or

to produce branches.

Individual bacteroids were observed in the cyto-

plasm of infected cells as early as day 5 (data not

shown), just 1 d after infection thread-like structures

were first evident. If bacteria are released from such

threads into the cytoplasm, the process must occur

very shortly after threads penetrate a cortical cell.

The development of the infection from these cells

may take place by proliferation of this initial

infection. The infected cell may divide repeatedly to

extend the infection centre (Fig. 10); the brady-

rhizobia seemed to be largely spread by repeated

plant cell division. Bacteria became distributed



NO St:

1'>~~~~~~~~~i 2

' '.~~~~~~~~P Figures 2-7. Light micrographs of cross sections of the roots of seedlings of Lupinus angustifolius receiving 2-5 /M iron. (Bar = 150 zsm for Figs 2-6 and 300 zsm for Fig. 7).

Figure 2. 2 d after inoculation. Cells of the outer cortex have divided into two daughter cells. Single arrows indicate the wall of the original cortical cell and arrowheads indicate the new wall. Nuclei are divided by the new wall (double arrows). Figure 3. 4 d after inoculation. The inner cortex starts to show cytoplasmic staining

and cell division. Cells of outer layers of cortex divide repeatedly and contain cytoplasm and enlarged nuclei. An enlarged root hair is associated. Figure 4. 5 d after inoculation. Cell division has occurred in the inner cortex. Oblique divisions of cortical cells have occurred (arrows). The newly divided cells of the outer cortex have become isodiametric, obscuring the parental cell profiles, and have stained darkly, indicating that they

have become enriched in cytoplasm. An enlarged root hair is associated (R). The meristem of a nodule initial (arrowhead) has formed. Part of another centre of cortical cell division is shown nearby. Figure 5. 7 d after inoculation. Cells are dividing further and the darkly stained area (arrowhead) is expanded. Anticlinal cell division has begun (arrows). Figure 6. 8 d after inoculation. Activity of the meristem (arrowhead) of a nodule



within the progeny of these cells as cell division

proceeded. Bacterial cell division could occur within

the peribacteroid membrane, resulting in two or

more bacterial cells within one envelope (Fig. 11).

As a result of further bacterial multiplication, the

cytoplasm of infected cells became more fully

occupied by bacteria (Fig. 12).

Iron-deficient plants

Plant growth and nodulation. Plants grown in 0 05 /AM iron showed chlorosis of young leaves from day 5

after treatments were imposed, while those grown in

0 25 /tM iron were slightly pale after day 11. The fresh weight of shoots was not significantly (P >

0 05) affected by the iron treatments for the 17 d of

iron deficiency. Roots exposed to 25 ,aM iron had more than 40 nodules per plant at the last harvest,

while those grown in 0 05 and 0 25 aM iron had only

2 and 5 nodules, respectively (Table 2). Nodules

appeared 9 d after inoculation on roots supplied with

2 5 AM iron, but nodule appearance on roots supplied

with 0 05 or 0 25 aM iron was delayed for a further 2

or 3 d. Many small brown spots, possibly aborted

infection sites, were observed on the surface of roots

grown in 005 and 0 25 /AtM iron, but not on roots grown in- 25 AM iron.

Differences in events in the process of nodulation

in iron-deficient and iron-sufficient plants are sum-

marized in Table 2.

Anatomy. Under iron deficiency, cortical cell div-

ision occurred as early as 2 d after inoculation, and

followed a generally similar pattern to that of iron-

sufficient plants until day 4. The root cortical cells of

plants given 0 05 or 0 25 gm iron did not divide

further after day 5. Cell division ceased in the outer

cortex (Figs 13, 14), with nodule meristems and

visible nodules rarely being formed in root tissues of

these two treatments. The initial division of root

cortical cells in these plants eventually ceased and

showed up as brown points on the root surface.

Ultrastructure. Until day 4 after inoculation, the

ultrastructure of target cells was generally similar to

that of plants grown in iron-sufficient solutions. In

iron-deficient plants no infection centre was formed,

though infection thread-like structures might have

been formed occasionally (Fig. 15), similar to those

in iron-sufficient plants at early stages. Brady-

rhizobia might also be released into the cytoplasm

(Fig. 16), but infection thread-like structures and

release of bacteria were observed in only a few

sections. Moreover, these bradyrhizobia did not

appear to divide further. The cortical cell wall and

thread wall became thickened. In contrast, under

iron-sufficient conditions, bradyrhizobia actively

multiplied and normal organelles were present (Figs

11, 12).

Nodules, which were occasionally formed from

infection sites that survived in roots exposed to 0 05

or 0 25 AM iron, showed delayed development, and

the numbers of bradyrhizobia per cell were much

decreased at the earlier stages (day 14) compared

with those in roots exposed to 25 ,aM iron (Figs 17, 18). Subsequently, the number of bradyrhizobia in

the nodules of iron-deficient plants reached levels

similar to those in cells of iron-sufficient plants at a

late stage. The development process of these in-

fection sites that survived in iron-deficient roots was

generally similar to that in iron-sufficient roots. The

ultrastructure of bacteroids in nodules was not

markedly affected by iron deficiency (micrographs

not shown).

DISCUSSION

Infection site

Infection in lupins seemed to be localized in a zone

similar to that in soybean and the apical infection in

clover but, unlike clover, the development of lateral

infections in the mature zone of roots does not

appear to occur in lupins. In soybean, Bhuvaneswari,

Turgeon & Bauer (1980) found that nodules de-

veloped most frequently in the region from 20 to

800% of the distance between the root tips and the smallest emergent root hairs from the root tip marks.

From observation of infection thread formulation of

clover, Nutman (1956) concluded that apical in-

fections initiated at the tips of root hairs occurred

only in the zone of the developing root hairs.

However, lateral infections initiated at a branch near

the base of a root hair were observed in the zone of

mature fully elongated root hairs.

cortex has caused a localized swelling on the root. Figure 7. 9 d after inoculation. The nodule has appeared at the root surface. Three darkly strained centres (bacteroid zones; arrowheads) are shown in this micrograph. The nodule vascular bundle (arrows) has differentiated between the nodule meristems and the root stele. Nodule

cortex (Nc) can be recognized.

Abbreviations in Figures 2-18

A amyloplast B bacterium or bacteroid C cortical cell W cell wall

E epidermal cell IS intercellular space M mitochondrion N nucleus

Nu nucleolus Nc nodule cortex R root hair St Stele

V vesicle or vacuole



* ~~~~~~~~~~~~~~~~~~~~~~~~~~. _............................ _

Fw w w !5 k...w 0 i R | I ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.....



Table 2. Comparison of events in the nodzlation process of iron-sufficient and iron-deficient lapin roots

Iron-deficient

Days after

inoculation Iron-sufficient Aborted infection sites Surviving infection sites

2-3 Initial cell division of outer cortex As in iron-sufficient plants

of roots

4 Infection threads observed No infection threads observed 5 Bacteria released into cytoplams of No bacteria observed in cytoplasm of cortical cells. Meristems of

cortical cells. Meristems of nodule initials not formed nodule initials formed, resulting

in nodule initiation. Cell division in both outer and inner cortex

7 Expansion of infection centres. Majority of infection sites not Meristems of nodule initials Further division of cortical and developed. Brown points on the formed from a few infection

bacterial cells in the centres root surface sites 8 Cell division extended to stele, Very short infection threads and Development of nodule initials

ending the stage of nodule bacterial release into the initiation cytoplasm

9 Nodules visible on root surface Organelles in initially divided Further expansion of infection

vascular bundle differentiated cortical cells disappearing centres

11 Further development of nodules. Majority of infection sites aborted Nodules visible on the roots. Bacteroids multiply and occupy Vascular bundle differentiated

the cytoplasm of infected cells 14 Maximum bacteroid number in Further development of nodules

nodules reached 17 Development of nodules ceased Maximum bacteroid number in

nodules reached. Ultra-

structure similar to that in

nodules on iron-sufficient

plants

Over 40 nodules per plant formed Only 2 or 5 nodules formed on plants receiving 0 05 or

0 25 aM iron respectively

Many studies of infection in legumes have pro-

vided evidence that rhizobia normally enter the root

through infection threads formed in root hairs.

However, our light microscope examination of serial

sections of the infected region of Lupinus roots did

not show infection threads in root hairs, consistent

with the observations of Dart (1977) and Quispel

(1983) for this genus. In contrast in soybean,

Bhuvaneswari et al. (1980) observed that infection

threads were developed in short emergent root hairs

and thus suggested that emergent root hairs were the

infectible cells.

Infection process

The bradyrhizobial infection process in lupin roots

seems to differ from that for many other legumes.

The entry of bradyrhizobia into lupin roots may be

similar to that for Frankia infection in the non-

legume (actinorhizal) plant Elaeagnus angustifolia

(Miller & Baker, 1985) or for Rhizobium infection in

Mimosa scabrella (Faria et al., 1988). The present

study suggests that infection threads are very scarce

or absent and may not be of significance in

nodulation in lupins; bradyrhizobia may therefore

have to enter the lupin roots through gaps between

epidermal cells and then penetrate cells of the root

cortex through their walls. Further study would be

required to confirm the mechanism of infection in

lupin.

The bradyrhizobial stimulus to division of target

cells in root tissues was active prior to bacterial

invasion. The enlarged nucleus and condensed

cytoplasm were observed 2 d after inoculation,

Figures 8-12. Electron transmission micrographs of cross sections from selected areas of infected roots

receiving 2 5 /kM iron.

Figure 8. Root tissue containing infection threads beneath an epidermal cell (single asterisk). Bacteria are

surrounded by wall material continuous with the host cell wall. The double asterisk indicates a cortical cell (bar

= 4,am). Figure 9. Higher magnification of the outlined area in Fig. 8 illustrating nodule bacteria in the infection thread (bar = 1 aum). Figure 10. Portion of infection centre at day 8. An infected cell is dividing.

Newly-formed wall material (arrows) is associated with the vesicles (bar = 4 /m). Figure 11. Two (single arrows) and several (double arrows) bacteroids are within the same peribacteroid membrane (bar = 1 ,um). Figure 12. Bacteroids have occupied almost the whole cytoplasm of a cell in the infected zone of a nodule,

14 d after inoculation. At the periphery of the cell, amyloplasts (A) and other organelles are visible (bar = 2 ,m).



*. A . :i X i fXiii-';?'!::. ' ............................... . _ \ j t.. ... j\. . x ... 8 :-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. ... ....

.=~ ~ ~ _ R R _ a:i.li.M. . . ...... . u - ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ :=Fr .: .: s ::. ............. 1 .sF :: :__ B: : ............. . : s: .:.... ...... .. ........ IM

-;. ;a E R | g w ; L . - z . !t ........ ; ; E * S .. ........ ll E=-- r r E ... ...... ..

S1 ;[g.^>; l . X s ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..... .

il314~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4 j' ....... !. t;.U. ....................... -. .-~~~~~~~~~~~~~~~~~~~~~~~~~~~. . ....

S~~~~~~~~~~~~~~~~~~~~~~~~~~il V0t 14.

_w+S! * > > iitiF 00V _ ~~~~~~~~~~~~~~~~~s ....................... -.w ]



whereas the. formation of infection threads and

bradyrhizobial invasion into cortical cells occurred at

days 4 and 5 respectively. In lucerne, a sulphated

acylated oligosaccharide signal appears to be in-

volved in cortical cell proliferation (Lerouge et al.,

1990); a similar extracellular signal may operate

following addition of Bradyrhizobium to lupin roots

and bacterial proliferation in the rhizosphere. In

contrast to some other legumes (Libbenga & Harkes,

1973), no threads developed into the inner cortex of

lupin roots. The bradyrhizobia in lupin infection

threads are usually surrounded by a thin layer of

what may be rhizobial polysaccharide (Robertson et

al., 1978). In pea and some other legumes, however,

the development of root nodules begins with the

induction of cell division in inner cortical cells at

some distance from the advancing infection thread

(Libbenga & Harkes, 1973).

The expansion of infection zones in lupin roots

appears to be similar to that in A. hypogaea

(Chandler, 1978) and Stylosanthes (Chandler et al.,

1982), with the invaded cells dividing repeatedly to

form a nodule. Uninfected cells were seldom ob-

served in the infection zones in lupins whereas

uninfected cells are often found in nodules on other

legumes, e.g. pea (Newcomb, 1976), mung bean

(Newcomb & McIntyre, 1981) or soybean (Meijer,

1982) in which the infection zone is expanded by

branching of infection threads.

Effect of iron deficiency

Initial division of cortical cells resulting from

inoculation with bradyrhizobia does not seem to be

limited by iron deficiency since it occurred in both

iron-deficient and iron-sufficient plants as early as

2 d after inoculation. There was little effect on the

division of cortical cells from different iron treat-

ments until 4 d after inoculation. However, after this

time, initially-stimulated cortical cells did not divide

further to form the meristem of nodule initials if iron

supply was limited. The total number of aborted

infection sites and visible nodules in iron-deficient

roots was similar to the number of visible nodules in

iron-sufficient roots (Tang, 1991). The present study

suggests that iron deficiency inhibits the division of

cortical cells and the establishment of the meristems

of nodule initials, perhaps by decreasing the pro-

liferation of bradyrhizobia inside the roots of L.

angustifolius.

While very small numbers of infection thread-like

structures were formed in iron-deficient root tissues,

the structure did not extend. Although, in some

circumstances, bradyrhizobia may be released into

the cytoplasm under iron-deficient conditions, their

further proliferation may also be greatly inhibited.

The evidence from anatomical observations thus

supports the previous finding from transfer experi-

ments that iron deficiency impaired a step at the

stage of nodule initiation (Tang et al., 1991).

Whether the failure of nodule initiation in lupins

results from low internal iron supply or from external

iron deficiency is unclear. Attempts to increase

internal iron concentrations using different ap-

proaches such as foliar application of iron and split-

root techniques have failed (Tang et al., 1990b;

1991). Nevertheless, in lupin plants pretreated with

extremely high concentrations of iron (50 AM) before

exposure to low iron concentrations (0-05 AM),

nodule formation was not increased (Tang, 1991). If

bradyrhizobial proliferation inside the root and the

formation of nodule initials requires internal iron,

questions arise about what form of iron in the host

plant can be used and whether there is any barrier to

Figures 13, 14. Light micrographs of cross section of the roots exposed to 0.05 or 0.25 atm iron (bar = 150 am).

Figure 13. Section from a root exposed to 0 05 /aM iron. On day 8 after inoculation, cell division has occurred in the outer cortex (arrows) but without formation of the meristem of nodule initials. Figure 14. At day 9 after

inoculation, cell divisions in the cortex of roots exposed to 0.25 /tm iron had ceased at the stage of day 4 after inoculation (Fig. 3). Some divided cells are darkly stained (arrow).

Figures 15-18. Electron transmission micrographs of cross section from selected areas of infected root tissues and nodules.

Figure 15. The centre of an infection site in the root 9 d after inoculation in 0 25 gM iron. An infection thread- like structure with a bacterium (arrow) is formed beneath a root hair (single asterisk), (bar = 1 ,um). Figure 16.

Centre of an infection site of a root receiving 0 05 gM iron. Many vesicles containing fibrous materials are present. Ribosomes appear to be disappearing and other organelles are hardly distinguishable. The arrow may indicate a bacterium released into the cytoplasm, and surrounded by peribacteroid membrane (double arrowhead). Electron opaque material with regular pattern (asterisk) may be chromatin separated from the

nucleus, (bar = 0 5 ,tm). Figure 17. Thin section of a young nodule developed on a plant in 0 05 /tM iron, taken 14 d after inoculation. Bacteroids surrounded by peribacteroid membranes occupy some part of the cytoplasm.

Cells contain vacuoles and vesicles and peripherally-located amyloplasts. Many organelles, endoplasmic reticulum (arrows) and a nucleus with nucleolus are also present (bar = 2 ,tm). Figure 18. Electron

micrograph, taken 14 d after inoculation, of a thin section of a nodule developed from a root exposed to 2 5 /tM iron, shows numerous bacteroids occupying almost the entire cytoplasm. Vacuoles have become fewer. Most of the endoplasmic reticulum (arrows) and organelles are confined to the periphery of the cell (bar = 4 ,tm).



bradyrhizobia using the iron sources in root tissues.

Total iron concentration in root tissues is much

higher than the external concentration required for

maximum nodulation, and the requirement of iron

for bradyrhizobial growth in pure culture is low

(O'Hara, Boonkerd & Dilworth, 1988).

In various Rhizobiumn species, common and host-

specific nodulation genes determining infection and

nodulation of specific legumes have been identified

(Long, 1989). Cytological studies have suggested

that nodule bacteria are able to elicit, at a distance,

cortical cell differentiation and initiation of nodule

organogenesis by means of extracellular signals

(Bauer et al., 1985; Lerouge et al., 1990). The major

alfalfa-specific signal, NodRm-1, has been purified

from the culture supernatant of R. meliloti strains

(Lerouge et al., 1990). Both common nodABC and

host-specific nodH and nodQ genes of R. meliloti are

involved in the production of extracellular symbiotic

signals (Lerouge et al., 1990). Iron deficiency might

therefore alter the expression of nodulation genes

and inhibit the production of the bradyrhizobial

signal required to establish nodule meristems in

lupins. However, no work has yet been done on the

effect of iron deficiency on nodulation gene ex-

pression and signal production in nodule bacteria.

The development of nodules in lupins is less sensitive than nodule initiation to iron deficiency.

However, development was delayed under iron

deficiency because the density of bacteroids in the

invaded zone of iron-deficient nodules was lower

than that of iron-sufficient nodules at early stages of nodule development. Numbers of bacteroids in

nodules of iron-deficient plants eventually reached

similar levels to those in nodules of iron-sufficient

plants. The ultrastructure of bacteroids was not

affected by iron deficiency, unlike effects observed in

nodules of iron-deficient soybean (Hecht-Buchholz

et al., 1987). The low number of bacteroids in the

developing iron-deficient lupin nodule mainly re-

sulted from the delayed nodulation process. It

follows that the multiplication of bacteroids in

nodules is unlikely to have been limited by the low

concentration of iron in the external solution.

ACKNOWLEDGEMENTS

C. Tang is grateful to The University of Western Australia

for a postgraduate research scholarship for this work and

to Mr. M. Stevens and Mr. G. L. Husca for their excellent

technical assistance.

REFERENCES

BAUER, W. D., BHUVANESWARI, T. V., CALVERT, H. E., LAW, I. J.,

MALIK, N. S. A. & VESPER, S. J. (1985). Recognition and

infection by slow-growing rhizobia. In: Nitrogen Fixation

Researchl Progress (Ed. by H. J. Evans, P. J. Bottomley & W. E.

Newton), pp. 247-253. Martinus Nijhoff, Dordrecht.

BHUVANESWARI, T. V., TURGEON, B. G. & BAUER, W. D. (1980).

Early events in the infection of soybean (Glycine max L. Merr)

by Rhizobiuim japonicumiii. I. Localization of infectible root cells. Plant Phlysiology 66, 1027-1031.

CHANDLER, M. R. (1978). Some observations on the infection of

Ar achis hypogaea L. by Rhizobiumn. Journal of Experzimental Botan-y 29, 749-755.

CHANDLER, M. R., DATE, R. A. & ROUGHLEY, R. J. (1982).

Infection and root-nodule development in Stylosanthles species by Rhizobiurm. J7ournal of Experimental Botany 33, 47-57.

DART, P. J. (1974). Development of root nodule symbioses, the infection process. In: Biology of Nitrogen Fixation (Ed. by A.

Quispel), pp. 381-429. North Holland Publishing Co., Am- sterdam, Oxford.

DART, P. J. (1977). Infection and development of leguminous nodules. In: A Treatise on Dinitrogen Fixationz (Ed. by R. W. F. Hardy & W. S. Silver), pp. 367-472. John Wiley & Sons, New York.

DAZZO, F. B. (1980). Infection processes in the rhizobium-legume

symbiosis. In: Advan7ces in Legume Science (Ed. by R. T. Summerfield & A. H. Bunting), pp. 49-59. Royal Botanic Gardens, Kew, England.

FARIA, S. M. DE, HAY, G. T. & SPRENT, J. I. (1988). Entry of

rhizobia into roots of Mimosa scabrella Bentham occurs between epidermal cells. Journal of General Microbiology 134, 2291- 2296.

HECHT-BUCHHOLZ, C., Su, H., Wu, Y. P., CHENG, C. H. & LIN, L. P. (1987). Effect of different iron concentrations on the

regeneration of leaves, nodulation and agronomical characters

in iron deficient soybean. Research Bulletin, School of Agri- cultuire, National Taiwan University 27, 66-81.

LEROUGE, P., ROCHE, P., FAUCHER, C., MAILLET, F., TRUCHET,

G., PROME, J. C. & DENARIE, J. (1990). Symbiotic host-

specifity of Rhizobiumr meliloti is determined by a sulphated and acylated glucosamine oligosaccharide signal. Nature 344, 781-784.

LIBBENGA, K. R. & HARKES, P.EA. A. (1973). Initial proliferation of cortical cells in the formation of root nodules in Pisum

sativuim L. Planta 114, 17-28. LONG, S. R. (1989). Rhizobium-legume nodulation: life together

in the underground. Cell 56, 203-214.

MEIJER, E. G. M. (1982). Development of leguminous root

nodules. In: Nitrogen Fixation, vol. 2, Rhizobium. (Ed. by W. J. Broughton), pp. 311-331. Clarendon Press, Oxford.

MILLER, I. M. & BAKER, D. D. (1985). The initiation, development and structure of root nodules in Elaeagnus angustifolia L. (Elaeagn7aceae). Protoplasma 128, 107-119.

NEWCOMB, W. (1976). A correlated light and electron microscopic study of symbiotic growth and differentiation in Pisum sativum

root nodules. Canadian J7ournal of Botany 54, 2163-2186. NEWCOMB, W. & MCINTYRE, L. (1981). Development of root

nodules of mung bean (Vigna radiata): a reinvestigation of endocytosis. Canadian J7ournal of Botany 59, 2478-2499.

NUTMAN, P. S. (1956). The influence of the legume in root-nodule symbiosis; a comparative study of host determinations and fun-ctions. Biological Reviews 31, 109-151.

O'HARA, G. W., BOONKERD, N. & DILWORTH, M. J. (1988). Mineral constraints to nitrogen fixation. Plant and Soil 108, 93-110.

QUISPEL, A. (1983). Dinitrogen-fixing symbiosis with legumes, non-legume angiosperms and associative symbioses. In: En-

cyclopedia of Plant Physiology, new series, vol. 15 A (Ed. by A. Lauchli & R. L. Bieleski), pp. 283-329. Springer-Verlag, Berlin, Heidelberg.

RANGA RAO, V. (1977). Effect of root temperature on the infection

processes of Lotus and Stylosanthes. J7ournal of Experimental Botany 28, 241-259.

REYNOLDS, E. S. (1963). The use of lead citrate at high pH as an electron opaque stain in electron microscopy. J7ournal of Cell Biology 17, 208-212.

ROBERTSON, J. G., LYTTLETON, P., BULLIVANT, S. & GRAYSTON,

G. F. (1978). Membranes in lupin root nodules. I. The role of Golgi bodies in the biogenesis of infection threads and peribacteroid membranes. J7ournal of Cell Science 30, 129-149.

SPRENT, J. I. (1989). Which steps are essential for the formation of

functional legume nodules? Newu Ph.ytologist 111, 129-153. SPURR, A. R. (1969). A low-viscosity epoxy resin embedding

medium for electron microscopy. Jfournal of Ultrastrucsture Resear ch 26, 31-43.



TANG, C. (1991). The role of iron2 in nlodildationl anld nitrogen fixation7 by legumnes. Ph.D. Thesis, The University of Western Australia, Australia.

TANG, C., ROBSON, A. D. & DILWORTH, M. J. (1990a). The role

of iron in nodulation and nitrogen fixation in Luipinu2ls aniguisti- folli2s L. Nezv Phytologist 114, 173-182.

TANG, C., ROBSON, A. D. & DILWORTH, M. J. (1990b). A split-

root experiment shows that iron is required for nodule initiation

in Luipinuis angglstifoliuls L. New Phytologist 115, 61-67. TANG, C., ROBSON, A. D., DILWORTH, M. J. (1991). Which stage

of nodule initiation in Lulpinuiis anguistifoliuis L. is sensitive to

iron deficieincy? Nezv Phytologist 117, 243-250.


microscopic evidence on how iron deficiency limits nodule ... · lupins (lupinus angustifolius l....

Documents