An endogenous β-galactoside-specific lectin has previously been isolated from rabbit bone marrow. The quantification of extracted lectin now indicates that approximately 75% of the lectin is intracellular in marrow. Indirect immunofluorescence studies show the extracellular lectin is associated with the erythroblast cell surface and is also found in some acellular areas of the marrow stroma. At enucleation, lectin surrounds the extruded nucleus while some residual lectin is observed in the cytoplasm of circulating reticulocytes and erythrocytes.

An endogenous β-galactoside-specific lectin has previously been identified in rabbit bone marrow tissue and preliminary observations suggested that the lectin might mediate inter-erythroblast associations during maturation; the lectin was therefore termed erythroid developmental agglutinin, EDA (Harrison & Chesterton, 1980a). Subsequently a lectin indistinguishable from that isolated from the marrow has been identified in a number of adult and neonatal rabbit tissues (Harrison et al. 1984). Moreover, endogenous β-galactoside-specific lectins have been identified in a number of other vertebrate species and are probably ubiquitous in animal tissues. On the basis of their structural similarity and saccharide specificity we have proposed that these lectins be collectively termed ‘galaptins’ (Harrison & Chesterton, 1980b) and now refer to the lectin in marrow as bone-marrow galaptin.

We report here an investigation of the intra-/extracellular distribution of the marrow galaptin using biochemical techniques and also the confirmation and extension of these studies by the immunocytochemical localization of lectin in both gently disaggregated marrow cells and resin-embedded whole marrow. A preliminary report of some of this work has been published previously (Harrison & Catt, 1985).

Materials

New Zealand White rabbits (2 ·0—2 ·5 kg) were made anaemic by five daily injections of phenylhydrazine (2 ·5% in water, pH 7 ·4, 0.35 ml kg−1) and marrow was extracted from the femur, tibia and humerus 48 h after the last injection. The antiserum used had been raised in a sheep, immunized with bone-marrow galaptin purified from lactose washes of marrow cells. The antibodies’ specificity for the isoforms of bone-marrow galaptin has previously been determined (FitzGerald et al. 1984). Anti-galaptin (anti-bmg) and preimmune IgG (immunoglobulin G) fractions were prepared from sera by ammonium sulphate precipitation (50%) and subsequent dialysis against 0 ·1 M-NaCl.

Analytical techniques

Lactose extracts were analysed by sodium dodecyl sulphate (SDS)-polyacrylamide gel electrophoresis (11% T, 2 ·6% C), isoelectricfocusing on agarose gels and Western blotting, using anti-bmg followed by peroxidase-linked rabbit anti-sheep IgG as described (FitzGerald et al. 1984). The lectin content of each lactose extraction was determined within 24 h by single radial immunodiffusion (SRID) assays. Duplicate samples were assayed in triplicate and compared with standard dilutions of purified bone-marrow galaptin on the same diffusion plate. The reliability of the assay in crude lactose extracts was established by adding a known amount of purified galaptin to each sample and checking that the expected increase in apparent lectin concentration was observed. Experimental details have been reported (Harrison et al. 1984).

Immunofluorescence studies

Cytocentrifuge slides, prepared from normal and anaemic marrow, gently disaggregated in phosphate-buffered saline (PBS; 6 ·5 mM-phosphate, pH7 ·4, 0 ·15M-NaCl), and from blood cells collected into PBS and concentrated by centrifugation (1500 g for 5 min), were dried under vacuum (over P2O5) overnight. In anaemic blood 92–98% of the cells were reticulocytes, judged by the presence of intracellular RNA precipitates after staining with Cresyl Blue (1%, w/v, Brilliant Cresyl Blue in PBS, mixed 1:1 (v/v) with blood sample and viewed immediately). The cytocentrifuge slides were fixed for 10 min in methanol and rehydrated in PBS for 15 min immediately before staining. All immunological reagents were spun for 2 min in a Beckman Microfuge (approx. 9000g) just before use and all procedures were performed at room temperature. Cells, covered with a drop of preimmune or anti-bmg IgG, 0 ·1 mg ml−1 in 5% (w/v) bovine serum albumin (BSA) in PBS, were incubated for 3 h in a wet box and then washed for 3h in a minimum of 500 ml PBS, with stirring, keeping slides treated with preimmune and anti-bmg IgG separate. Cells were then covered with a drop of fluorescein isothiocyanate(FITC)-derivatized rabbit anti-sheep IgG (FIRaSh, Miles Laboratories Ltd), diluted 80 times in rabbit serum + 0 ·01% (v/v) Nonidet P40 (NP40), incubated for 30 min and then washed for 1 h in PBS. Slides were mounted in 80% glycerol in PBS containing Citifluor mountant, which retards fading of fluorescence during illumination (obtainable from Chemistry Department, City University, London, UK) and viewed on a Leitz Dialux microscope with epifluorescent illumination. Results were recorded on Kodak XP 1 or Kodacolour 400 film and printed on Ilford Multigrade II paper for black and white prints.

To prepare sections, bone marrow was chopped into approximately 5 mm cubes and fixed in 8% formaldehyde for 1 h. Fixation, embedding in JB-4 resin (Polysciences Ltd) and the cutting of sections was carried out as previously described (Catt & Harrison, 1985a), except that the high fat content of the marrow necessitated a prolonged dehydration of 30 min at 0°C in each of a graded acetone/water series and overnight dehydration in 100% acetone at 0°C. Sections were rehydrated for 30min in 5% (w/v) BSA in Tris-saline (10mM-Tris, pH7 ·2) and incubated for 2 ·3 h with anti-bmg IgG, 0’2mgml−1 in 5% BSA (w/v) in Tris-saline, which had been preincubated for 30 min before the addition of 0 ·01% (v/v) NP40. Slides were spray-rinsed for 1 min and then soaked in 600 ml Tris-saline for 30–45 min before incubation with FIRaSh, 0 ·1 mg ml−1 in 2 ·5% (w/v) BSA/50% (v/v) rabbit serum in Tris-saline, which had been preincubated for 30min before the addition of 0’01% (v/v) NP40. Slides were then spray-rinsed and soaked in Tris-saline for 4 h before mounting and viewing as described above. Control incubations were performed with 0’2 mg ml−1 preimmune IgG and with FIRaSh alone to assess non-specific binding. The staining specificity of anti-bmg IgG was checked by preincubating both the antibody and the section with excess purified lectin (0 ·2mgml−1). Results were recorded on Kodak Ektachrome (daylight) 400 ASA and plates were prepared by copying the colour slides on a Bowen’s Illumitran 3, through a Wratten no. 63 filter, onto Ilford Pan F to improve the distinction on black and white prints between the green fluorescence and the red background caused by the Evan’s Blue.

Extraction of the lectin from bone marrow

All procedures were carried out on ice or at 4°C. Extracellular lectin was isolated by gently disaggregating bone marrow in 10 ml LM/EDTA/NaCl per rabbit (0 ·3 M-lactose in 1ml I−1β-mercaptoethanol, 2mM-EDTA (pH 7 ·4), 0 ·15 M-NaCl) containing 1 mM-phenylmethylsulphonyl fluoride. Cells were pelleted at 1000g for 5 min and washed again with 10ml LM/EDTA/NaCl. Pooled washes were centrifuged at 26500 g for 15 min and the supernatants recentrifuged at 122500g for 2h. Intracellular lectin was extracted by lysis of washed bone marrow cells, resuspended in a minimum of LM/EDTA/NaCl and ground in a pestle and mortar with coarse silver sand under liquid nitrogen. After thawing, the supernatant was diluted to 10 ml and the sand removed by centrifugation at 800g for 1 min then washed in 10 ml LM/EDTA/NaCl. Lysate and wash were pooled and centrifuged as for the cell washes above. Centrifugation pellets were extracted with either 0 ·1% SDS (w/v) or 0 ·1% (v/v) Triton X-100 in LM/EDTA/NaCl. The extent of cell lysis during each procedure was estimated by comparison of the haemoglobin content of each extraction with the total tissue haemoglobin, determined by the maximal absorption of supernatants at 380–440 nm (approx. 418 nm) relative to that of a rabbit haemoglobin standard.

For the quantitative determination of intra- and extracellular lectin alternative homogenization conditions were used. Marrow was extracted, pooled and mixed by gentle stirring. Two 1-g samples were weighed and one was washed with LM/EDTA/NaCl as described above. Whole tissue and washed bone marrow cells were then pulverized in an anvil press precooled with ‘cardice’ and liquid nitrogen and then the frozen wafer of tissue was homogenized in 5 ml LM/EDTA/NaCl in an Ultra-Turrax homogenizer for four 15-s periods. Approximately 1 ml volumes of blood were either allowed to clot or collected into 5 ml LM/EDTA/NaCl and weighed. Lectin was extracted from whole blood (clot) or from cells collected into LM/EDTA/NaCl and washed once, as for the marrow cells. Haemoglobin concentrations were determined spectrophotometrically.

Biochemical studies

Extracellular lectin, purified from lactose washes of marrow cells, as previously described (Harrison et al. 1984), was identical in molecular weight, isoelectric point, lectin activity and immunological reactivity to intracellular lectin purified from lysates of lactose-washed cells. Since minor isoforms of bone-marrow galaptin have previously been identified in crude lactose washes of bone marrow and in preparations of galaptin after purification (FitzGerald et al. 1984), intracellular lectin isoforms were compared with extracellular lectin by isoelectricfocusing followed by Western blotting. Haemoglobin determinations indicated that lysis of 20–23% marrow cells occurred during lactose washing, thereby releasing a proportion of the internal lectin into the washes. Nevertheless, staining the nitrocellulose blots with anti-bmg (Fig. 1) suggested that the isoform of pI 5 ·6 was predominant both intra- and extracellularly in each of four rabbits investigated. Two other isoforms, at pI 5 ·2 and 5 ·9, were also detectable in each lactose extract. Biochemical studies therefore indicated no major difference between the internal and external lectin.

Fig. 1.

Western blot of an isoelectricfocusing gel of intra- and extracellular galaptin from the bone marrow of four anaemic rabbits. The three iaoforms of bone marrow galaptin, pl 5 ·2, 5 ·65 and 5 ·95, are similarly abundant in all preparations. 1, Bone marrow galaptin purified from lactose washes of marrow cells (1 μg): 2,4,6,8, intracellular lectin from approximately 2 mg marrow; 3,5,7,9, extracellular lectin from approximately 2 mg marrow.

Fig. 1.

Western blot of an isoelectricfocusing gel of intra- and extracellular galaptin from the bone marrow of four anaemic rabbits. The three iaoforms of bone marrow galaptin, pl 5 ·2, 5 ·65 and 5 ·95, are similarly abundant in all preparations. 1, Bone marrow galaptin purified from lactose washes of marrow cells (1 μg): 2,4,6,8, intracellular lectin from approximately 2 mg marrow; 3,5,7,9, extracellular lectin from approximately 2 mg marrow.

To quantify intra- and extracellular lectin, single radial immunodiffusion assays were carried out on the crude lactose washes and cell lysates of normal and anaemic blood and marrow. Comparative experiments showed that grinding with sand and liquid nitrogen to lyse the marrow cells depleted the extracts of lectin, but adequate lysis was not achieved by osmotic shock or freeze-thawing procedures. To extract cytoplasmic lectin washed marrow cells were therefore pulverized in a precooled anvil press and subsequently homogenized as described in Materials and Methods.

Insoluble material, pelleted on centrifugation of crude lactose extracts, was extracted with SDS or Triton X-100 and subjected to SDS-polyacrylamide gel electrophoresis. Only trace amounts of galaptin were detected, indicating efficient release of the lectin by these extraction procedures. To assess the relative distribution of the lectin, we compared the galaptin content of washed cells with that of whole undissociated marrow. Circulating erythrocytes from normal rabbits and reticulocytes from anaemic rabbits were collected in duplicate samples, and lectin levels in whole blood and lactose-washed cells were then determined as for the marrow samples. Results are presented in Tables 1 and 2.

Table 1.

Determination of total and intracellular galaptin in circulating blood cells

Determination of total and intracellular galaptin in circulating blood cells
Determination of total and intracellular galaptin in circulating blood cells
Table 2.

Determination of total and intracellular galaptin in bone marrow

Determination of total and intracellular galaptin in bone marrow
Determination of total and intracellular galaptin in bone marrow

No significant lysis of circulating blood cells occurred during lactose-washing procedures but there was inevitably some lysis of bone marrow cells on disaggregation and washing. Histological staining of cytocentrifuge slides indicated that the relative proportions of different cell types in the marrow remained constant during washing procedures, suggesting that there was not significant differential lysis of particular cell types. We therefore used haemoglobin as an intracellular marker and took the% haemoglobin lost from duplicate marrow samples during dissociation and washing as an estimate of the total cell lysis. The extent of lysis varied between 18 and 36% for normal marrow samples and 25 and 56% for anaemic marrow. This unavoidable lysis resulted in a loss of intracellular lectin from the washed cells. To account for this we expressed the lectin concentrations in both lactose extracts in terms of μg lectin mg−1 haemoglobin. This procedure permits direct comparison of the experimentally determined lectin concentrations in washed cells and undissociated tissue, and also corrects for any differences between duplicate tissue samples and any losses from either sample during preparation procedures. To compare intra- and extracellular lectin concentrations a least-squares linear regression of total lectin in undissociated marrow (μgmg−1 haemoglobin) on intracellular lectin in washed cells (figmg-1 haemoglobin) was performed. In all cases intercepts were not significantly different from zero, so slopes through the origin were estimated (Table 3). In circulating erythrocytes and reticulocytes there were no significant differences between total and intracellular lectin concentrations (i.e. slope = 1), indicating that there is very little extracellular lectin in serum or bound to the cell surface. However, in normal and anaemic marrow significant differences were observed (i.e. slopes, which are equivalent to mean ratios of total: intracellular lectin concentration, were 1 ·25 and 1 ·34, respectively). These differences indicate a partially extracellular distribution of the galaptin. In terms of the total lectin content, in normal marrow approximately 25% and in anaemic marrow approximately 20% of the total lectin is extracellular. In fact, the intercepts and slopes for regression of total and intra cellular lectin were not significantly different for normal and anaemic marrow. The combined slope through the origin (1 ·311 ±0 ·063) indicates that 76 ±4% of total lectin is intracellular in both normal and anaemic animals. This suggests that anaemia induced by phenylhydrazine treatment does not cause any major redistribution of the lectin. From these data the quantitites of intracellular lectin per erythroblast and per circulating cell can be estimated to be 80–170 fg/erythroblast, 50fg/reticulocyte and 30 fg/erythrocyte, which indicates that there are of the order of 4 ×106 molecules within an erythroblast, decreasing to 1 ·5 ×106 molecules within circulating erythrocytes.

Table 3.

Comparison of total and intracellular lectin concentration in circulating blood cells and in bone marrow, in normal and anaemic rabbits

Comparison of total and intracellular lectin concentration in circulating blood cells and in bone marrow, in normal and anaemic rabbits
Comparison of total and intracellular lectin concentration in circulating blood cells and in bone marrow, in normal and anaemic rabbits

Immunocytochemistry

Sections of resin-embedded bone marrow show the complex multicellular organization of the tissue (Fig. 2B). Large fat cells (f) are prominent and megakaryocytes, eosinophils and neutrophils can also be readily identified. Erythroid (e) and myeloid cells (m) differentiate in clusters within the marrow matrix. Eosinophils (arrowhead) and erythrocytes (ey) seen on sections were fluorescent in all incubations, even when incubated with FIRaSh alone. This non-specific fluorescence has been reported for eosinophils (Zucker-Franklin et al. 1981), which are autofluorescent and also bind the fluorochrome, FITC. The fluorescence of erythrocytes results from the fixation procedure, since circulating erythrocytes fixed with 8% formaldehyde exhibit a high degree of non-specific staining on cytocentrifuge slides whereas cytocentrifuged erythrocytes fixed in methanol show virtually no fluorescence in control incubations. All other cell types on sections showed only a very diffuse background staining in control incubations. The pattern of fluorescent staining seen with anti-bmg IgG (Fig. 2A) indicates that the galaptin is restricted to particular areas of the marrow, some of which can be morphologically identified as erythroid, though at the light-microscopic level it is not possible to classify each cell; in particular, early myeloid and erythroid blast cells cannot be distinguished. Erythroblasts at later stages of differentiation appeared to contain the lectin but recognizable myeloid cells did not. Unfortunately, cell surface lectin could not be clearly distinguished in the sections because of the high intracellular fluorescence and the close packing of the cells within the marrow matrix; however, lectin appeared to be extracellular in a Y-shaped area of cellular material, clearly stained by anti-bmg IgG (Fig. 2A, arrow) and by Toluidine Blue (Fig. 2B, arrow).

Fig. 2.

Indirect immunofluorescent staining of semi-thin section of resin-embedded normal bone marrow. Bar, 20 pm. After fluorescent staining (A) the section was counterstained with Toluidine Blue to enable identification of cell type (B);f, fat cell; arrowhead, eosinophil; ey, erythrocytes; e, predominantly erythroid area; m, predominantly myeloid area. Fluorescent staining (A) shows the galaptin is restricted to certain areas of the marrow, most of which can be identified as erythroid, e.g. e. Predominantly myeloid areas do not contain the lectin. A Y-shaped area of connective tissue (arrow), in A and B, clearly contains galaptin. Fluorescent staining of eosinophils and erythrocytes is artifactual, as discussed in the text.

Fig. 2.

Indirect immunofluorescent staining of semi-thin section of resin-embedded normal bone marrow. Bar, 20 pm. After fluorescent staining (A) the section was counterstained with Toluidine Blue to enable identification of cell type (B);f, fat cell; arrowhead, eosinophil; ey, erythrocytes; e, predominantly erythroid area; m, predominantly myeloid area. Fluorescent staining (A) shows the galaptin is restricted to certain areas of the marrow, most of which can be identified as erythroid, e.g. e. Predominantly myeloid areas do not contain the lectin. A Y-shaped area of connective tissue (arrow), in A and B, clearly contains galaptin. Fluorescent staining of eosinophils and erythrocytes is artifactual, as discussed in the text.

To identify positively those cells showing intracellular lectin cytocentrifuge slides of gently disaggregated marrow were prepared, and the cells were permeabilized by methanol fixation. Fig. 3A,B shows that the galaptin is intracellular in erythroblasts but not in myeloid cells (m). Cells treated with preimmune IgG showed no fluorescent staining (Fig. 3C,D). Occasionally erythroblastic islands were seen, containing a central macrophage surrounded by erythroblasts, and staining showed lectin within and possibly between erythroblasts but not in the cytoplasm of the macrophage nurse cell (n) (Fig. 3E,F). The fluorescent staining became concentrated around the nucleus in maturing erythroblasts until, at enucleation, a ring of fluorescence surrounded the extruded nucleus and the reticulocyte cytoplasm showed only weak diffuse staining (Fig. 4A,B). Light-microscopic techniques cannot distinguish whether the lectin is on the external surface or within the small amount of cytoplasm surrounding the extruded nucleus. Circulating reticulocytes and erythrocytes showed very faint diffuse intracellular staining when cytocentrifuge slides were prepared and stained as described above (data not shown).

Fig. 3.

Indirect immunofluorescent staining of methanol-fixed bone marrow cells from an anaemic rabbit on cytocentrifuge slides, showing galaptin in the cytoplasm of erythroid cells. Bar, 20 μm. A,B. Fluorescent staining and corresponding phase micrograph shows cytoplasmic lectin in erythroblasts at different stages of maturation. Myeloid cells (tn) do not contain the lectin. C,D. Fluorescent and corresponding phase micrograph showing that there is very little staining with preimmune IgG. E,F. An erythroblastic island, lectin is detected only in erythroblasts and not in the central macrophage ‘nurse cell’ (n).

Fig. 3.

Indirect immunofluorescent staining of methanol-fixed bone marrow cells from an anaemic rabbit on cytocentrifuge slides, showing galaptin in the cytoplasm of erythroid cells. Bar, 20 μm. A,B. Fluorescent staining and corresponding phase micrograph shows cytoplasmic lectin in erythroblasts at different stages of maturation. Myeloid cells (tn) do not contain the lectin. C,D. Fluorescent and corresponding phase micrograph showing that there is very little staining with preimmune IgG. E,F. An erythroblastic island, lectin is detected only in erythroblasts and not in the central macrophage ‘nurse cell’ (n).

Fig. 4.

Indirect immunofluorescent staining of methanol-fixed bone marrow cells from an anaemic rabbit on cytocentrifuge slides, showing concentration of galaptin around the extruding nucleus. Bar, 20 μm. Fluorescent staining (A) and the corresponding phase micrograph (B) show extensive perinuclear concentration of the lectin in late orthochromatic erythroblasts (arrowheads) and around the extruding nucleus (arrows).

Fig. 4.

Indirect immunofluorescent staining of methanol-fixed bone marrow cells from an anaemic rabbit on cytocentrifuge slides, showing concentration of galaptin around the extruding nucleus. Bar, 20 μm. Fluorescent staining (A) and the corresponding phase micrograph (B) show extensive perinuclear concentration of the lectin in late orthochromatic erythroblasts (arrowheads) and around the extruding nucleus (arrows).

In spite of their relative abundance the precise functions of lectins in situ remain uncertain. A role in intercellular recognition and adhesion has been proposed for lectins in several other systems, but contradictory evidence has also been presented. A setback has been the discovery that the majority of the lectin is located intracellularly. In aggregating slime moulds the extracellular lectin has been estimated to be as little as 0 ·2% (Springer et al. 1980) or even 0 ·02% (Barties et al. 1982) of the total cellular lectin, and Erdos & Whitaker (1983) were unable to detect any immuno-cytochemically reactive lectin on the cell surface of Dictyostelium discoideum. In the case of myoblast fusion in the chick, immunocytochemical studies have shown that the galaptin is predominantly intracellular in myoblasts and early myotubes, and is externalized only later in development (Barondes & Haywood-Reid, 1981). Furthermore, neither competing saccharides, exogenous lectin or anti-lectin antibody inhibits the fusion of embryonic myoblasts in culture (Den et al. 1976; Den & Chin, 1981). Immunocytochemical studies, however, have suggested that galaptins are extracellular in chicken intestine, where they are particularly associated with mucus secreted by goblet cells (Beyer & Barondes, 1982) and in rat lung where some lectin is associated with elastic fibres (Cerra et al. 1984).

In considering the role of the galaptin in erythropoiesis it is important to establish the precise location of the lectin in bone marrow. Immunofluorescence staining for light microscopy and immunogold techniques for electron microscopy (Harrison & Chesterton, 1980a; Catt et al. 1985) have shown that galaptin is present at the surface of erythroblasts that have been isolated from marrow tissue, but we could not rule out the possibility of lectin redistribution during disaggregation. Using quantitative biochemical techniques we have now firmly established that a significant proportion of the galaptin, approximately 25%, is indeed extracellular in rabbit bone marrow. The origin of the extracellular lectin has not been determined but it appears to be identical to that found within erythroblasts. Using resin-embedded sections to observe the overall distribution of the lectin in marrow we found that galaptin is associated with discrete areas of cells, probably the sites of erythroid differentiation and also with some acellular areas of extracellular matrix. This distribution is similar to that observed in other rabbit tissues where immunocytochemistry suggests that the lectin is both intra- and extracellular, and is particularly prominent surrounding differentiating cells, for example in intestinal crypts and hair follicles in the skin (Catt & Harrison, 1985a). Increasing evidence suggests that matrix components play an important role in the regulation of cellular differentiation in marrow (Zuckerman & Wicha, 1983; Spooncer et al. 1983). The galaptin is presumably not involved in the induction of erythropoiesis, as erythroblasts themselves seem to be major producers of the lectin, but the lectin may play a role in the maintenance and regulation of the very high rates of erythrocyte production (about 150 million erythrocytes min−1 in human marrow; Dexter, 1984), perhaps by mediating cellcell or cell-matrix associations during maturation. We are now investigating these possibilities using bone marrow cultures where differentiation can be observed and experimentally manipulated. At enucleation, when the maturing erythroblast is released from its close association with the nurse cell, a ring of fluorescence surrounds the extruded nucleus, while residual lectin in the reticulocyte remains intracellular. Though electron-microscopic studies are required to confirm that the endogenous lectin is actually associated with the plasma membrane surrounding the extruded nucleus, such a differential distribution of the lectin at enucleation would be of obvious relevance both to the phagocytosis of extruded nuclei and to the release of reticulocytes into the circulation.

We are grateful to Sally Martin and Sue Carleton for technical assistance, to Dr John Paterson for his invaluable assistance with the statistical treatment of the data and to Peter Taylor for photographic advice. J. W. Catt is a Research Fellow supported by the Cancer Research Campaign.

Barondes
,
S. H.
&
Haywood-Reid
,
P. L.
(
1981
).
Extemalization of an endogenous chicken muscle lectin with in vivo development
.
J. Cell Biol
.
91
,
568
572
.
Bartles
,
J. R.
,
Frazier
,
W. A.
&
Rosen
,
S. D.
(
1982
).
Slime mould lectins
.
In Int. Rev. Cytol
(ed.
G. H.
Bourne
&
J. F.
Danielli
), vol.
75
, pp.
61
93
.
New York
:
Academic Press
.
Beyer
,
E. C.
&
Barondes
,
S. H.
(
1982
).
Secretion of an endogenous lectin by chicken intestinal goblet cells
.
J. Cell Biol
.
92
,
28
33
.
Catt
,
J. W.
&
Harrison
,
F. L.
(
1985a
).
Selective association of an endogenous lectin with connective tissues
.
J. Cell Sci
.
73
,
347
359
.
Catt
,
J. W.
,
Peacock
,
M. A.
&
Harrison
,
F. L.
(
1985
).
Surface localization of an endogenous lectin in rabbit bone marrow
.
Histochem. J
.
17
,
189
199
.
Cerra
,
R. F.
,
Haywood-Reid
,
P. L.
&
Barondes
,
S. H.
(
1984
).
Endogenous mammalian lectin localised extracellularly in lung elastic fibres
.
J. Cell Biol
.
98
,
1580
1589
.
Den
,
H.
&
Chin
,
J. H.
(
1981
).
Endogenous lectin from chick embryo skeletal muscle is not involved in myotube formation in vitro
.
J. Biol. Chem
.
256
,
8069
8073
.
Den
,
H.
,
Malinzak
,
D. A.
&
Rosenberg
,
A.
(
1976
).
Lack of evidence for the involvement of a /3- D-galactosyl-specific lectin in the fusion of chick myoblasts
.
Biochem. Biophys. Res. Commun
.
69
,
621
627
.
Dexter
,
T. M.
(
1982
).
Stromal cell associated haemopoiesis
.
J. cell. Phys. Suppl
.
1
,
89
94
.
Dexter
,
T. M.
(
1984
).
The message in the medium
.
Nature, Land
.
309
,
746
747
.
Erdos
,
G. W.
&
Whitaker
,
D.
(
1983
).
Failure to detect immunocytochemically reactive endogenous lectin on the cell surface of Dictyostelium discoideum
.
J. Cell Biol
.
97
,
993
1000
.
Fitzgerald
,
J. W.
,
Catt
,
J. W.
&
Harrison
,
F. L.
(
1984
).
Isoforms of an endogenous lectin in rabbit bone marrow
.
Eur.J. Biochem
.
140
,
137
141
.
Harrison
,
F. L.
&
Catt
,
J. W.
(
1985
).
An endogenous /3-galactoside specific lectin is both intra- and extra-cellular during erythropoiesis
.
Biochem. Soc. Trans
.
1116
1117
.
Harrison
,
F. L.
&
Chesterton
,
C. J.
(
1980a
).
Erythroid developmental agglutinin is a protein lectin mediating specific cell-cell adhesion between differentiating rabbit erythroblasts
.
Nature, Land
.
286
,
502
504
.
Harrison
,
F. L.
&
Chesterton
,
C. J.
(
1980b
).
Factors mediating cell-cell recognition and adhesion
.
Galaptins, a new class of bridging molecules. FEBS Letts
122
,
157
165
.
Harrison
,
F. L.
,
Fitzgerald
,
J. E.
&
Catt
,
J. W.
(
1984
).
Endogenous β-galactoside-specific lectins in rabbit tissues
.
J. Cell Sci
.
72
,
147
162
.
Spooncer
,
E.
,
Gallagher
,
J. T.
,
Krizsa
,
F.
&
Dexter
,
T. M.
(
1983
).
Regulation of haemopoiesis in long-term bone marrow cultures
.
IV. J. Cell Biol
.
96
,
510
514
.
Springer
,
W. R.
,
Haywood
,
P. L.
&
Barondes
,
S. H.
(
1980
).
Endogenous cell surface lectin in Dictyostelium: quantitation elution by sugar, and elicitation by divalent immunoglobulin.^
.
Cell Biol
.
87
,
682
690
.
Zucker-Franklin
,
D.
,
Greaves
,
M. F.
,
Grossi
,
C. E.
&
Marmont
,
A. M.
(eds) (
1981
).
Atlas of Blood Cells, Function and Pathology, vol
.
1. Philadelphia, USA: Lea & Febiger
.
Zuckerman
,
K. S.
&
Wicha
,
M. S.
(
1983
).
Extracellular matrix production by the adherent cells of long-term bone marrow cultures
.
Blood
61
,
540
547
.