Lectins in the Cell Nucleus

Glycoblotogy vol. 1 no. 3 pp. 243-252, 1991

MINI REVIEW Lectins in the cell nucleus

John L.Wang, James G.Laing and Richard L.Anderson Nuclear binding of neoglycoproteins implicates lectins Department of Biochemistry, Michigan State University, East Lansing, The existence of lectin molecules in the cell nucleus was first MI 48824, USA inferred from the binding of neoglycoproteins (see Hubert et al., 1989). These neoglycoproteins were derived by coupling specific saccharide structures onto a polypeptide backbone that normally

bears no carbohydrate moiety [e.g. bovine serum albumin, Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 Key words: carbohydrate recognition/neoglycoproteins/ (BSA)]. Fluorescence microscopy and quantitative flow micro- nuclear glycoconjugates/S-type lectins/subcellular localization fluorometry were carried out with fluorescein-labelled neoglycoproteins and, in certain instances, ultrastructural analyses were performed with mannose (Man)-conjugated ferritin. When Triton-permeabilized baby hamster kidney (BHK) cells Introduction were incubated with fluorescein-labelled neoglycoprotein, fluorescence was observed in both the cytoplasm and the There is a general consensus that the potential to encode nucleus (Seve et al., 1985, 1986). Strongest binding was observed information in carbohydrate structures is enormous. Many of the for neoglycoproteins bearing a-L-rhamnose, whose structure possibilities of carbohydrate recognition in cellular function have resembles that of /S-D-galactose (Gal) (Monsigny et al., 1983). been explored for extracellular molecules and events such as Significant binding (> 3-fold over fluorescein-conjugated BSA) specific cell-cell recognition and cell adhesion to the extracellular was also observed for BSA bearing glucose (Glc), A'-acetyl- matrix. Our interest in these possibilities, however, can now be glucosamine (GlcNAc), lactose (Lac), mannose 6-phosphate and extended to the intracellular compartment as well, mainly as a fucose residues. Similarly, cryostat sections of calf tissues result of the demonstrations that carbohydrate structures such incubated with fluorescein-conjugated Lac-BSA showed cyto- as glycoproteins and glycosaminoglycans can be found in the plasmic and nuclear staining (Childs and Feizi, 1980). This cytoplasm and the nucleus (for a review, see Hart et al., 1989). staining could be competed with non-fluorescent Lac-BSA, but These advances in nuclear and cytoplasmic glycosylation have, not with Man-BSA. These results implicate the existence of in turn, stimulated interest in the intracellular localization of specific carbohydrate-binding proteins in both the cytoplasm and lectins. Defined as non-enzymatic and non-immune proteins the nucleus. which can selectively bind specific carbohydrate structures, these Karyoplasts obtained by enucleation of BHK cells also bound molecules appear to be good candidates for recognizing variations the same neoglycoproteins after permeabilization (Seve et al., in isomers, alternative linkages, sequence of saccharide units and 1986). Karyoplasts from exponentially growing cells bound a branched structures. much greater number of any given neoglycoprotein than did There are excellent reviews on glycosylation in the nucleus nuclei from density-inhibited cells. In the promyelocytic cell line and cytoplasm (Hart et al., 1989) and on the organization and HL60, in vitro differentiation into monocytes or granulocytes is functional implications of carbohydrate recognition domains in accompanied by a decrease in the binding of neoglycoproteins animal lectins (Barondes, 1988; Drickamer, 1988). The purpose to the nucleoplasm (Facy et al., 1990). These results suggest a of the present review is to summarize updated information in correlation between the nuclear localization of the lectin(s) and these areas and to discuss critically the available data pertaining the proliferation or differentiation state of the cells. Such a notion to the subcellular localization of those carbohydrate-binding is consistent with the results of comparing the level of the nuclear proteins that have been observed in the cell nucleus. lectin, carbohydrate binding protein 35 (CBP35), in mouse 3T3 As will be detailed in our discussion, all of the lectins located fibroblasts in quiescent versus proliferating cultures (Moutsatsos in the nucleus are also found in the cytoplasm. The nucleus and etai, 1987) (see below). Moreover, CBP35 has been found the cytosol, which communicate through the nuclear pores, are to be associated with the ribonucleoprotein elements (RNPs) of topologically continuous although functionally distinct. There the nucleoplasm (Laing and Wang, 1988), consistent with the are well-known examples of proteins [e.g. Drosophila heat distribution of at least a portion of the neoglycoproteins bound shock protein hsp 70 (Velazquez and Undquist, 1984)] and RNAs to the nucleus. In all of the mammalian cells studied (BHK cells, [e.g. snRNAs (Mattaj and De Robertis, 1985)] that can shuttle HeLa cells, leukaemic murine L1210 cells and rat hepatocytes), between the nucleus and the cytoplasm. What is more curious, neoglycoproteins bearing rhamnose, fucose or glucose residues however, is the considerable amount of evidence that indicates bound mostly to interchromatin spaces of the nucleoplasm and intracellular lectins also occur at the cell surface and, in some to nucleoli (Seve etal., 1985, 1986; Bourgeois etai, 1987). cases, in the extracellular medium. These latter sites belong to Condensed chromatin, revealed by the DNA-specific dye Hoechst the extracellular compartment, which is topologically continuous 33342, showed little or no binding of the neoglycoproteins. with the lumen of the endoplasmic reticulum, the Golgi stack The localization of neoglycoproteins in two non-mammalian and the secretory vesicles, but is topologically distinct from the cell types has been mapped to specific substructures within the intracellular compartment. These fascinating observations form nucleus. Nuclei of lizard ovarian follicle cells were highly labelled the basis for much of the discussion to follow. when ultra-thin sections incubated in the presence of Man-ferritin

were observed with electron microscopy (Hubert et al., 1985). source exhibit slightly different carbohydrate-binding affinities This binding can be specifically competed with Man-BSA. (Roff and Wang, 1983). In general, the polypeptide of the Labelling over nucleoli, chromatin and the outer leaflet of the L-14 group consists of a single domain, the CRD (Figure 1). nuclear envelope was particularly strong. Due to the unique In non-denaturing solvents, the native L-14 molecules are often organization of the granular component of the nucleoli in the comprised of dimers of the constituent polypeptides. specialized granulosa cells, it was possible to conclude that the In the L-30 group fall all the S-type lectins whose polypeptide Man-binding sites were associated with the dense fibrillar mol. wts have been reported to be in the range of 29 000—35 000 component of nucleolus containing the transcribing 45S rRNA (Table I). These include proteins originally isolated from various genes and with the nucleolar granules related to ribosomal sources and designated different names, but all of which exhibit precursors to be exported into the cytoplasm. In a study carried saccharide-binding activity: (i) CBP35 (Jia and Wang, 1988) from out on isolated macronuclei from the ciliated protozoan Euplotes mouse 3T3 fibroblasts; (ii) L-34 (Raz et al., 1989) from mouse eurystomus, neoglycoprotein labelling was found over nucleoli, tumour cells; (iii) Mac-2 (Cherayil et al., 1989) from mouse interchromatin spaces and, most interestingly, the macronuclear macrophages; (iv) laminin binding protein (LBP) (Woo et al., replication bands (Olins et al., 1988). Inasmuch as these repli- 1990), also from mouse macrophages; (v) IgE-binding protein cation bands represent delimited areas of DNA synthesis, this (eBP) (Albrandt et al., 1987) from rat basophilic leukaemia Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 suggests a co-localization of saccharide-binding components and cells; (vi) RL-29 (Leffler etal., 1989) and HL-29 (Sparrow sites of DNA synthesis. et al., 1987) from rat and human lung tissue, respectively. The The identity of the lectin(s) responsible for the binding of the results of Southern blot analyses of genomic DNA suggest that various neoglycoproteins in the nucleus (and in the cytoplasm) there is only one gene coding for the polypeptides of the L-30 has not been determined. Thus, it is not clear how many different group (Raz et al., 1988; Cherayil et al., 1989). For any given lectins are present in the various regions of the nucleus or whether species, the few differences in the amino acid sequences report- the various neoglycoproteins can bind to different subsites of one ed for the different proteins (e.g. mouse CBP35, L-34 and particular lectin. Extraction of nuclei or subnuclear components, Mac-2) are most probably ascribable to minor sequencing errors. followed by affinity chromatography on specific neoglycoprotein Although alternative sites of transcription initiation and/or adsorbents, will need to be carried out to purify the putative alternative splicing of the primary transcripts may lead to multiple lectin(s), and correlation of the saccharide-binding specificities mRNA species, the polypeptides of the L-30 group listed in Table of these isolated lectins with the labelling patterns of the neo- I are most likely identical. The available data indicate that the glycoproteins will need to be established. L-30 polypeptides do not form oligomers in non-denaturing solvents, in contrast to the lectins of the L-14 group. The polypeptides of the lectins in the L-30 group contain two domains; S-type lectins are found inside and outside of cells the CRD is coupled to another domain, presumably responsible for mediating some as yet unidentified effector function Two groups of S-type lectins (Figure 1). Evidence for the presence of lectins in the nucleus has also been Because lectins of both the L-14 group and the L-30 group accumulated through immunofluorescence localization and belong to the S-type family, the CRDs of all these lectins are immunoblotting of subcellular fractions using antibodies directed homologous, with conservation of 15 characteristic amino acid against polypeptides previously purified by saccharide-specific residues (Figure 1). Thus far, all of the lectins with S-type CRDs affinity chromatography. Such studies have identified Gal-/Lac- bind to lactosamine-based glycoconjugates (Barondes, 1988; specific lectins belonging to the S-type family of animal lectins, Drickamer, 1988). In addition to similar CRDs, the lectins of as classified by Drickamer (1988). Lectins are grouped into this the S-type family share another key characteristic: they are found S-type family on the basis of conserved amino acid residues within in both the intracellular as well as the extracellular compartments. a characteristic carbohydrate recognition domain (CRD), which We shall review the data concerning the subcellular localization is clearly distinguishable from the corresponding CRD of the of the L-14 group and of the L-30 group separately. calcium-dependent, C-type lectins. Some S-type lectins have been reported to depend on the presence of reducing agents (dithiothreitol or 2-mercaptoethanol) for maintaining carbo- Subcellular localization of L-14 hydrate-binding activity, but more recent studies suggest that, Comparisons of the immunofluorescence staining of live cells for certain of these lectins, the thiol dependence may be ascribed versus cells fixed with formaldehyde followed by permeabiliza- to an artefact of the extraction procedure rather than an intrinsic tion indicate that proteins of the L-14 group are found pre- requirement of the protein itself (Frigeri et al., 1990). dominantly in the intracellular compartment. Within the cells, For the purpose of the present discussion on nuclear lectins, the staining is mostly cytoplasmic. Observations and interpre- it is convenient to subdivide the S-type lectins into two (arbitrary) tations pertaining to the nuclear localization of the L-14 group groups: (i) the L-14 group and (ii) the L-30 group. The L-14 are somewhat more difficult to establish. group consists of all the S-type lectins from various species and First, there are studies that specifically show the localization various tissues whose reported polypeptide mol. wt on reducing of L-14 in the nuclei as well as the cytoplasm. Cryostat sections sodium dodecyl sulphate polyacrylamide gels (SDS—PAGE) of tissues subjected to immunofluorescence showed labelling of ranges from 10 000 to 16 000 (Table I). There is evidence that both nuclei and cytoplasm in the following cases: anti-CLL-I polypeptides of the L-14 group represent members of a multi- staining of adult chicken kidney (Beyer and Barondes, 1980) and gene family (Girt and Barondes, 1986; Sakakura et al., 1990). anti-BHL staining of calf pancreas (Quids and Feizi, 1980). A At least some of the L-14 proteins listed in Table I appear to monoclonal antibody, designated 36/8, was generated against be distinct. For example, the two chicken lectins, C-14 and C-16, BHL and stained the nuclei and cytoplasm of lymphoblastoid and have homologous, but not identical, sequences (Sakakura et al., leukaemic cells (Carding et al., 1985). This monoclonal antibody 1990). Moreover, CBP16 and CBP13.5 isolated from the same immunoblotted polypeptides of apparent molecular mass 13, 36,

244 Nuclear lectins

Table I. L-14 and L-30 groups of S-type lectins*

1 Group Name Species Tissue/cell type Mr (in reducing Reference ' SDS-PAGE)

L-14 CLL-n Chicken Intestine, liver 12 000 Beyer and Barondes, 1982 BHL Calf Heart, lung, spleen, thymus 12 000 Abbott a al., 1989 CHL Chicken Heart 13 000 Barak-Briles et al., 1979 Galaptin Rabbit Bone marrow 13 000 Fitzgerald et al., 1984 CBP13.5 Mouse, human Lung, fibroblast 13 500 Crittenden et al., 1984 C-14 Chicken Skin 14 000 Hirabayashi et al., 1987 H-14 Human Placenta, hepatoma 14 000 Abbott and Feizi, 1989 HL-14 Human Lung 14 000 Gitt and Barondes, 1986 M 3T3 L Mouse Fibroblast 14 000 Wilson et al., 1989 RL-14.5 Rat Lung, brain, intestine 14 500 Clerch et al., 1988 L-14.5 Mouse Fibrosarcoma, melanoma 14 500 Raz et al., 1988 CLL-I Chicken Muscle, liver 15 000 Beyer and Barondes, 1982 Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 C-16 Chicken Liver 16 000 Sakakura et al., 1990 CBP16 Mouse, human Lung, fibroblast 16 000 Cnttenden et al., 1984 Xenopus skin lectui Frog Skin 16 000 Bols et al., 1986 Electrolectin Electric eel Electric organ 16 000 Paroutaud et al., 1987 L-30 RL-29 Rat Lung, brain 29 000 Leffler et al., 1989 HL-29 Human Lung 29OO0 Odaef al., 1991 eBP Rat, human Basophilic leukaemia cells 31 000 Robertson et al., 1990 Mac-2 Mouse, human Macrophages 32 000 Cherayil et al., 1990 L-34 Mouse Fibrosarcoma, melanoma 34 000 Raz et al., 1989 CBP35 Mouse, human Lung, fibroblast 35 000 Jia and Wang, 1988 LBP Mouse Macrophages 35 000 Woo et al., 1990

"This table attempts to provide a comprehensive list of the S-type lectins existing under different names. Sequence data indicate that some of these lectins are identical (see text). ''Space limitation does not permit a comprehensive citation of references. For each lectin, we cite here the reference that provides amino acid/nucleotidt sequence information or the reference that is most recent. Interested readers can consult the reference cited to search for previous work.

^.MPBF-V-N-WG-C-H-FPF-O- 65, 80 and 130 kD in extracts of lymphoid cells. More recent L-14 studies have shown that the 36/8 antibody recognizes the tetrapeptide sequence Trp-Gly-Ala/Ser-Glu/Asp (Abbott et al., L-30 -(PQAYPG—),- -H-NPRF-V-N-WO-E-R-FPF-G- 1989) and, therefore, it is not clear whether the nuclear staining

E-, 01YCMC- CARSOHYORATl MCOCUcmON component is a lectin or some other polypeptide carrying this RJCH DOUAM DOHAJH epitope sequence. B. Second, in a number of studies originally performed for other objectives (e.g. to show overlap in cells expressing lectins and lactoseries glycoconjugates), the investigators provide no specific conclusion concerning the intracellular distribution of L-14. Dorsal root ganglion neurons were subjected to immunofluorescence with anti-RL-14.5 and with a monoclonal antibody directed against a lactoseries glycoconjugate (Regan et al., 1986). Although the intracellular distribution of the lectin was not discussed in that particular report, the same investigators have OO I5O 200 250 recently concluded, on the basis of the previously published data, SEQUENCE POSITION that RL-14.5 could be detected in both the nucleus and cytoplasm of the neurons (Hynes et al., 1990). Similarly, reassessment Fig. 1. Summary of the structural features of S-type lectins. A. Domain of previously published immunocytochemical studies in non- content and features of sequences within domains in the L-14 and L-30 neuronal cells has suggested the presence of RL-14.5 in nuclei groups. Shown are the 15 invariant amino acid residues that occur in a 39-residue sequence in the carbohydrate recognition domains of all of the and cytoplasm. L-14 and L-30 proteins for which sequence information is available. Also Finally, immunofluorescence and ultrastructural studies have shown is the 9-residue consensus sequence that is repeated in the proline-, glycine-ridi domain of L-30. The letter n, designating the number of also led to explicit statements that anti-L-14 antibodies failed to repeats, ranges from five in the human Mac-2 sequence to 10 in the rat label the cell nucleus. In chicken embryo fibroblasts, anti-chicken fBP sequence. A single variable residue is denoted by a hyphen (-). heart lectin (CHL) staining was observed only in cells fixed and Sequences of two or more variable residues are denoted by the symbol permeabilized, but not in unfixed cells, suggesting that most (~). B. Hydropathy plot of murine CBP35, illustrating the distinctive of the lectin was intracellular (Barak-Briles et al., 1979). The patterns that characterize the two domains of an L-30 polypeptide. The deduced amino acid sequence of CBP35 was subjected to analysis by intracellular staining was ascribed only to the cytoplasm. When the program of Hopp and Woods (1981). Positive values indicate 17-day-old chicken embryonic keratinized epidermis was stained hydrophilicity and negative values indicate hydrophobicity. with gold-labelled anti-C-14 under electron microscopy, gold

245 J.L.Wang et al.

particles were found around desmosomes, tonafilament bundles NH2-terminal domain of the polypeptide (Figure 1) contains and the intercellular space, while the cell nucleus was free of structural features also found in several of the hnRNP proteins the particles (Hirano et al., 1988). (see below). Although there is little question that the L-14 proteins are found The protein eBP was originally identified in rat basophilic within cells, there is also considerable evidence that many of these leukaemia cells as an IgE-binding protein (Albrandt et al., 1987). lectins can be found outside as well (Barondes, 1984; Catt et al., From a comparison of the amino acid sequence and binding 1987; Raz and Lotan, 1987; Hirano etal., 1988). Moreover, properties of eBP with those of CBP35, it was established that the situation need not be static; there can be a shift from an the two proteins are rat and murine homologues, respectively, intracellular to an extracellular location with differentiation or and that the eBP exhibited carbohydrate-binding activity (Laing upon application of specific stimuli. Several striking observations et al., 1989). Subcellular fractionation studies on rat basophilic by Barondes and colleagues are worth citing to illustrate this leukaemia cells showed that the majority of the eBP is intra- point. The skin of Xenopus laevis contains a lectin of the L-14 cellular, including in the nucleus (Gritzmacher et al., 1988). A group that is found in the cytoplasm of granular and mucous gland small proportion of eBP is found on the cell surface. There are cells (Bols et al., 1986). Upon injection of epinephrine, the lectin sufficient sequence data on RL-29 (Leffler et al., 1989) to indicate

is externalized directly from the cytoplasm without the involve- a close relationship, if not identity, with rat eBP. Immuno- Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 ment of secretory vesicles. In another example, CLL-I, which cytochemical studies in both non-neuronal cells and in dorsal root is concentrated intracellularly in developing chick muscle (19-day ganglion neurons have shown that RL-29 can be detected in both embryo), is exported from poly nucleated myotubes and appears the cytoplasm and the nucleus (Regan et al., 1986; Hynes et al., extracellularly with maturation (2-day-old chick) (Barondes and 1990). Haywood-Reid, 1981). The process of externalization was studied Two lectins designated as L-14.5 and L-34 were purified, on in a mouse L2 myogenic cell line (Cooper and Barondes, 1990). the basis of their carbohydrate-binding activity, from tumour Immunohistochemical localization during the intermediate stages cells such as murine melanoma cell lines B16-F1 and K-1735, of externalization suggests that the lectin becomes concentrated fibrosarcoma UV-2237-IP3 and carcinoma HeLa-S3 (Raz and in evaginations of the plasma membrane, which in turn pinch Lotan, 1987). These proteins were located, by immunofluor- off to form labile lectin-rich extracellular vesicles. Thus, a novel escence staining, at the cell surface and, after fixation and mechanism for lectin export from the intracellular to the extra- permeabilization, in intracellular pools. The presence of L-34 cellular compartment is suggested. (and of L-14.5) at the cell surface was established by lacto- peroxidase-catalysed 125I-labelling, followed by immunoprecipi- tation. It has been reported that normal rat embryonal fibroblasts express only L-14.5 but, upon neoplastic transformation, also Subcellular localization of L-30 express L-34. Among related tumour cell variants of K-1735 The identification of L-30 in the nucleus of a cell was first melanoma, the UV-2237 fibrosarcoma and the A31 angio- reported for CBP35 (Moutsatsos etal., 1986). CBP35 was sarcoma, expression is highest in those cells that exhibit the initially isolated from extracts of mouse 3T3 fibroblasts on the greatest metastatic potential. It was thus suggested that the basis of its Gal-/Lac-specific carbohydrate-binding activity. presence of L-34 at the cell surface may be related to neoplastic Although rabbit anti-CBP35 detected a small amount of the transformation and progression toward metastasis. lectin on the surface of the 3T3 cells (by immunofluorescent The Mac-2 antigen and LBP were both identified in mouse staining, by agglutination of live cells and by immunoisolation macrophages on the basis of their cell surface localization. The of l25I-labelled lectin after surface iodination), the predominant former was originally described by a monoclonal antibody that portion of the lectin was found to be intracellular by immuno- stained the cell surface of macrophages and that identified a fluorescent staining of fixed and permeabilized cells. There was protein of Mr 32 000 after surface iodination (Ho and Springer, prominent labelling of the nucleus and variable staining of the 1982). More recent pulse-chase analysis and subcellular cytoplasm. Cytoplasmic areas devoid of phase-dense intracellular fractionation studies showed that although the majority of the vesicles stained diffusely, whereas areas rich in vesicular bodies Mac-2 was intracellular, a small proportion was found in the stained in a highly reticular manner. More recently, identical extracellular medium (Cherayil et al., 1989). In these experi- results have been obtained using the monoclonal antibody ments, actin could not be detected in the medium, suggesting generated against the Mac-2 antigen (Ho and Springer, 1982) that significant leakage of the cytosolic proteins did not occur. to stain 3T3 fibroblasts. Immunoblotting of subcellular fractions Since Mac-2 is expressed on thioglycollate-elicited peritoneal identified CBP35 in the nuclear pellet (-5%), cytosol fraction macrophages at a much higher level than resident macrophages, (90%) and membranes of the postnuclear supernatant (<5%). it may be induced during the process of inflammatory activation. Several lines of evidence suggest that CBP35 is associated The major non-integrin LBP of murine macrophages was with RNPs in the nucleoplasm (Laing and Wang, 1988). First, identified as a polypeptide of Mr ~ 35 000 on affinity columns treatment of permeabilized 3T3 cells with ribonuclease A of laminin-Sepharose (Woo etal., 1990). Amino acid se- removed the immunofluorescence due to CBP35 in the nucleus, quencing of tryptic peptides of LBP revealed identities to whereas parallel treatment with deoxyribonuclease I failed to yield sequences in CBP35. It was shown that LBP could be eluted from the same effect. Second, when nucleoplasm was fractionated by laminin-Sepharose by the addition of galactose or lactose. Cell density gradient centrifugation, immunoblotting analysis localized surface iodination, followed by laminin-Sepharose affinity CBP35 in fractions with densities corresponding to those reported chromatography, showed that LBP is the predominant macro- for heterogeneous nuclear ribonucleoprotein complex (hnRNP) phage surface protein that binds to laminin with high affinity, (1.30—1.35 g/ml on caesium sulphate gradients and 40S on consistent with reports that implicate Gal residues on laminin as sucrose gradients). Third, fractionation of nucleoplasm on important determinants of cell adhesion (Runyan et al., 1988). Gal-Sepharose or anti-CBP35 polyacrylamide beads resulted The intracellular localization of LBP was not reported, but since in the binding of the protein as well as RNA. Finally, the LBP and Mac-2 are presumably the same macrophage protein,

246 Nuclear lectins the data on the large intracellular pool of Mac-2 also apply mRNAs for the L-30, including an amino-terrninal extended to LBP. sequence with an in-frame CTG translation initiation codon followed by a putative signal sequence (Cherayil et al., 1989). However, the same investigators (Cherayil etal., 1990) have Issues raised by dual intracellular and extracellular subsequently discounted this possibility; die basis for such a localization conclusion was not detailed. In general, the fraction of L-14 or L-30 externally exposed at the cell surface represents only a small portion of that lectin in Carbohydrate binding protein 35 the cell. Inside the cell, the protein appears to be in the cytoplasm and/or nucleus rather than in the lumen of membrane-enclosed As is apparent from the above discussion, the S-type lectins have vesicles. This suggests that the lectin found in the extracellular been studied from a number of perspectives. Among these, compartment did not follow the classical endomembrane pathway murine CBP35 has been pursued most intensively as a nuclear of secretion. Consistent with this notion, none of the cDNA clones lectin. On this basis, the properties of CBP35 will be highlighted. identified for the L-14 or L-30 proteins has revealed an amino- It should be kept in mind, however, that all of the L-30 proteins

terminal signal sequence for sequestration into the lumen of the listed in Table I are either identical (within a given species) or Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 endoplasmic reticulum. These observations on the S-type lectins are homologues (between species). Therefore, some of the infor- parallel similar findings for a growing list of proteins with dual mation on CBP35 to be discussed below was originally obtained intracellular and extracellular localization: (i) proteins with in a context other than as a nuclear protein. nuclear functions such as probasin (Spence et al., 1989), simian virus 40 (SV40) large T antigen (Butel and Jarvis, 1986), adenovirus El A gene product (Bellgrau et al., 1988) and La RNP Structure identified by autoimmune anti-nuclear antibodies (Bachmann Murine CBP35 is a monomer comprised of 264 amino acid et al., 1990); (ii) growth or differentiation factors capable of residues. A hydropathy analysis of the deduced amino acid binding to cell surface receptors in autocrine or paracrine target sequence (Jia and Wang, 1988) indicates that the protein consists systems such as the heparin-binding growth factors (including of two structural domains (Figure 1), an arrangement drat is also acidic and basic fibroblast growth factor) (Sano etal., 1990), found in several of the C-type lectins (Drickamer, 1988). The platelet-derived endothelial cell growth factor (Ishikawa et al., carboxyl-terminal half of the CBP35 molecule contains both 1989), interleukins la and 1/3 (Rubartelli et al., 1990), and the hydrophilic and hydrophobic regions, as is characteristic of many yeast mating hormone, a-factor (Kuchler et al., 1989); (iii) other globular proteins. This region includes a 76-residue sequence that extracellular proteins such as factor Xllla (Grundmann et al., is -35% identical with several members of the L-14 group of 1986) and ATL-derived factor (Tagaya et al., 1989). The data Gal-/Lac-specific lectins. Fifteen amino acid residues are invariant on probasin and on SV40 large T antigen will be highlighted to in every L-14 and L-30 for which sequence information is contrast and compare the mechanisms of dual localization. available (Figure 1). Consequendy, the carboxyl-terminal half Probasin is a rat prostatic protein found both in secretions and of CBP35 is assumed to harbour the CRD. In contrast, the amino- in the nucleus of prostatic epithelial cells (Spence et al., 1989). terminal half of the molecule contains neither a highly hydrophilic Only one probasin mRNA could be detected by primer extension nor a hydrophobic region. This half includes a stretch of eight and SI nuclease protection assays. In vitro translation of this contiguous 9-residue repeat units having the sequence Pro-Gly- mRNA demonstrated that a protein containing a signal sequence Ala-Tyr-Pro-Gly followed by three other residues. As a result, and a protein lacking a signal sequence were synthesized by diis stretch of the sequence is characterized by a high proportion initiation at different AUG codons. In contrast, the SV40 large of Pro and Gly. Certain of the core hnRNP proteins also exhibit T antigen contains a nuclear localization signal, but no obvious two distinct domains with a largely unequal distribution of Gly signal sequence for secretion (Butel and Jarvis, 1986). Although and Pro residues in the two halves. Over a 107-residue region, the majority of the T antigens are found in the cell nucleus, a the sequence of CBP35 is — 25 % identical with some of the core small fraction (< 5 %) is localized at the cell surface. The protein proteins of hnRNP. This degree of identity is comparable to that is glycosylated and represents an example of a nuclear glyco- found among the core hnRNP proteins themselves. Raz et al. conjugate (see below). It also undergoes post-translational (1989) have pointed out that the amino-terminal half of murine palmitylation. Palmitylation is the only known structural dif- L-34 also has 33.5% identity with the collagen al (II) chain of ference between the plasma membrane and nuclear forms of T bovine cartilage. antigen, as none of the nuclear T antigen population is modified The pi of murine CBP35 polypeptide is 8.7, as determined in this way. bodi by calculation from the deduced amino acid sequence and At present, one key question seems to be whether the experimentally by isoelectric focusing of recombinant CBP35 extracellular lectin is derived from the same population of obtained by expression of a cDNA clone in Escherichia coli molecules inside the cell. Studies on the externalization of L-14 (Cowles et al., 1990). When extracts of mouse 3T3 cells were in cultured myogenic cells would suggest that this is the case, subjected to two-dimensional gel electrophoresis and thus implicating a novel mechanism of export (Cooper and immunoblotting, two spots were observed, corresponding to pi Barondes, 1990). In several of the cases studied, however, values of 8.7 and 8.2. The pi 8.2 species represents a post- externalization appears to be induced or enhanced by some translational modification of die pi 8.7 polypeptide by the addition stimulus (e.g. epinephrine injection to release L-14 in frog skin of a single phosphate group, probably O-linked. These two or thioglycollate elicitation of inflammatory macrophages). The isoelectric variants have differential expression and localization alternative possibility is then raised that the externalized molecules in the cells (see below). arise by de novo synthesis and processing at the RNA and/or Homologues of murine CBP35 have been identified in other protein level. cDNA cloning studies on Mac-2 had suggested the species, and those from rat (Albrandt et al., 1987; Leffler et al., existence of two or more alternatively spliced macrophage 1989) and human (Cherayil etal., 1990; Oda etal., 1990;

247 J.L.Wang et al.

Robertson et al., 1990) have been sequenced. The rat and human In quiescent cultures of 3T3 fibroblasts, CBP35 is found lectins are comprised of 262 and 250 amino acids, respectively, primarily in the cytoplasm, whereas in proliferating cultures, it in contrast to the 264 amino acids that make up the mouse lectin. increases in amount and is located predominantly in the nucleus. However, the homology of the lectins among the three species The addition of serum growth factors to quiescent 3T3 cells is very high. For the carboxyl-terminal half, which represents increases the expression of CBP35; this increase occurs early the CRD, the identity is well over 80%; the amino-terminal half after stimulation, well before the onset of the S-phase of the also has highly conserved sequences, including the 9-residue first cell cycle (Moutsatsos et al., 1987; Agrwal et al., 1989). repeated sequence. Oda et al. (1991) have suggested a third These conclusions are derived from: (i) analysis at the single- domain, located at the amino-terminal end: residues 10—39 of cell level by immunofluorescence; (ii) analysis at the protein level HL-29 are 46% identical to residues 18—47 of human serum by Western blotting; (iii) analysis at the mRNA level by Northern response factor, a transcriptional activator. blotting; (iv) analysis of the transcription of the CBP35 gene in nuclear run-off experiments. At the mRNA accumulation and gene transcription levels, the increases in the expression of CBP35 Saccharide-binding specificity occur even in the presence of cycloheximide. In this respect, the In the initial isolation (Roff and Wang, 1983), it was established expression of CBP35 resembles that of c-fos, whose transcription Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 that murine CBP35 could be eluted from asialofetuin-Sepharose is also activated in response to serum growth factors, independent columns by lactose or galactose, but not by mannose, sucrose of the synthesis of other proteins (Lau and Nathans, 1987). or Af-acetylglucosamine. Lactose was more active than galactose There is a single CBP35 gene in the normal mouse genome, and concentrations as low as 250 pM caused 50% inhibition of based on sequence analysis of genomic clones (S.Jia and binding of the lectin to erythrocytes (Crittenden et al., 1984). J.L.Wang, unpublished results). This gene spans ~9 kb of More extensive analysis of the saccharide-binding specificity was genomic DNA and contains five exons and four introns. Upstream reported for RL-29 and HL-29 (Leffler and Barondes, 1986; from the initiation site of transcription, putative TATA and Sparrow et al., 1987). The binding activity of various saccharides CCAAT boxes are found at positions corresponding well with was determined by measuring the concentration that gave 50% similar sequences of other genes. There is also a putative serum inhibition of binding of the lectin to asialofetuin-Sepharose response element, a 10-nucleotide consensus sequence known to columns. The affinity of RL-29 and HL-29 for lactose was found bind serum-responsive transcription factors during the activation to be 57—100 times greater than that for galactose. The parts of c-fos and /3-actin genes (Tuil et al., 1990), - 200 nucleotides of the lactose molecule that are critical for binding include the upstream from the same transcription start site. Finally, a poly- hydroxyls at positions 4 and 6 of Gal and position 3 of Glc, since adenylation signal is found in the 3' untranslated region. These substitutions at any of these positions greatly reduce binding structural features may account for the regulation of expression activity. Substitution at positions 2 of Gal or 1 or 6 of Glc have of the CBP35 gene, particularly during serum-stimulated little effect on the binding activity. However, binding is enhanced activation of quiescent cultures of 3T3 fibroblasts. Whereas 7—11 times by the incorporation of an acetamido group at CBP35, L-34 and Mac-2 have apparently identical amino acid position 2 of Glc (i.e. to yield N-acetyllactosamine). For RL-29 sequences, the nucleotide sequence in the 5' untranslated region and HL-29, an even greater enhancement of binding affinity of the CBP35 mRNA (derived from the genomic sequence) results from the addition of GalNAcal- at position 3 of the Gal appears to be completely different from those reported for the moiety of lactose derivatives, although the binding affinity of corresponding regions of the mRNAs (sequenced as cDNAs) A'-acetylgalactosamine itself is low. These lectins also have high for L-34 (Raz et al., 1989) and Mac-2 (Cherayil et al., 1989). binding affinity for larger oligosaccharides (e.g. fetuin oligo- The significance of these differences, particularly with respect saccharide) and glycopeptides (e.g. erythrocyte lactosaminoglycan to the expression and localization of the lectin, remains to be glycopeptide) that have the lactosyl determinants discussed determined. above. At the protein level, the expression of CBP35 was further The saccharide-binding specificity of the L-14 group of lectins analysed by subcellular fractionation, two-dimensional gel electro- is similar to that of the L-30 group, but there are sufficient phoresis to separate the isoelectric variants, and irnrnunoblotting differences in specificity between these groups to suggest the (Cowles et al., 1990). The phosphorylated (pi 8.2) form of the possibility of selective interactions with cellular glycoconjugates. CBP35 is found both in the cytosol and nucleus, whereas the Within the L-30 group, the binding specificities of RL-29 and unmodified (pi 8.7) species is found exclusively in the nucleus. HL-29 are very similar but, nevertheless, distinguishable (Leffler Quiescent populations have primarily the phosphorylated form, and Barondes, 1986; Sparrow etal., 1987). but it is at a low level and is located predominantly in the cytoplasm. Serum-stimulated cells have an increased level of the phosphorylated form, both in the cytosol and the nucleus, but Regulation of expression and subcelhdar localization the amount of the unmodified (pi 8.7) form increases dramatically Crittenden et al. (1984) noted wide differences in the level of and is all nuclear. The significance and mechanism of the dif- CBP35 in various tissues of mice and, for some tissues, between ferential expression and localization of the two isoelectric variants embryonic and adult stages. Ho and Springer (1982) found that of CBP35 remain as some of the many intriguing, but unanswer- induction of cell-surface Mac-2 in mouse macrophages was ed, questions regarding this and related lectins. dependent on a specific eliciting agent such as thioglycollate. Raz and Lotan (1987) found that the expression of L-34 occurred Carbohydrate structures In the nucleus concomitantly with neoplastic transformation of normal rat embryonal fibroblasts. Taken together, these and other observa- dycoproteins containing O-linked GlcNAc residues tions indicate that the lectin is differentially expressed during There has been little or no evidence that carbohydrate recognition development, stimulation, or change in growth status or potential occurs via the nuclear lectins discussed above. For one thing, of cells. the endogenous carbohydrate ligand(s) to which the nuclear (and

248 Nudear lectins cytoplasmic) lectins can bind remains to be identified. Although In considering the function of O-GlcNAc glycosylation in the the subject of nuclear and cytoplasmic glycosylation has been cytoplasm and the nucleus, it should be noted that, to date, no reviewed elsewhere (Hart et al., 1989), it is, nevertheless, endogenous intracellular lectin that specifically recognizes such important to consider several classes of candidates for these saccharide structures has been reported. The binding of neoglyco- ligands. Just as the binding of neoglycoproteins to nuclear proteins bearing GlcNAc residues has been observed (Seve et al., structures implicated lectins, the binding of the GlcNAc-specific 1986), however, so it remains as a future challenge to isolate plant lectin, wheat germ agglutinin (WGA), to nuclear com- and characterize the lectins. ponents provided evidence for carbohydrate structures in the nucleus (Baglia and Maul, 1983; Hart et al., 1988). These WGA- binding proteins also serve as ready acceptors for the transfer of radiolabelled galactose, using UDP-[3H]Gal and galacto- Other glycoproteins and glycosaminoglycans syltransferase (Holt and Hart, 1986; Hart et al., 1988). Isolation Although the binding of plant lectins, including soybean and chemical characterization studies have led to the identification agglutinin, Ricinis communis agglutinin, concanavalin A, and of a novel carbohydrate moiety: a single residue of GlcN Ac Ulex europaeus lectin, have implicated glycoproteins other than glycosidically linked to a polypeptide through a hydroxyl group the CM31cNAc-containing polypeptides, the structures of the (designated as 0-GlcNAc). A membrane-bound cytoplasmic glycosylated moieties remain to be established. One soybean Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 GlcNAc transferase activity has been identified (Haltiwanger agglutinin-binding glycoprotein that is of particular interest is the et al., 1990); this transferase may account for the (2-GlcNAc large T antigen of SV40 (Butel and Jarvis, 1986). This multi- glycosylation of intracellular proteins without their passing purpose protein is necessary for the replication of SV40 in virus- through the endomembrane/secretory pathway. Such O-GlcNAc infected cells. Metabolic labelling experiments have demonstrated glycosylated proteins have been found in a variety of cells and that the large T antigen can incorporate both [3H]Gal and at a number of subcellular locations, including the cytoskeleton, [3H]GlcNAc (Butel and Jarvis, 1986; Schmitt and Mann, 1987; the nuclear pore complex and chromatin (Hart et al., 1988; Kelly Jarvis and Butel, 1988); this labelling is resistant to tunicamycin and Hart, 1989). treatment and endoglycosidase H digestion, suggesting that The nuclear pore complex has several OGlcNAc-bearing the saccharide is O-linked. Although originally localized in the proteins, the most abundant and well-characterized of which is nucleus, consistent with its known biological activities, a con- gp62 (Davis and Blobel, 1986; Holt et al., 1987; Schindler et al., siderable amount of effort has been expended to show rigorously 1987). WGA and monoclonal antibodies whose epitopes include that the protein is also found at the cell surface. Thus, the the O-GlcNAc moiety inhibit the import of proteins from the SV40 large T antigen is a prime example not only of a nuclear cytoplasm (Finlay et al. 1987; Yoneda et al., 1987; Featherstone glycoprotein, but one in which there is "good evidence for intra- et al., 1988) and the export of RNA from the nucleus (Baglia cellular and extracellular dual localization. From the perspective and Maul, 1983; Featherstone et al., 1988). In the presence of of our discussion on the S-type lectins, the SV40 large T antigen the lectin or antibody reagents, the binding of proteins bearing is also interesting because the disaccharide structure implicated the nuclear localization signal to the pore complex is not affected, on the basis of the available data would serve as a good ligand whereas the subsequent ATP-dependent translocation through the for the carbohydrate-binding proteins. pore is blocked. More strikingly, a system has been developed The high mobility group (HMG) proteins are abundant non- in which the 0-GlcNAc-bearing proteins can be specifically histone chromatin proteins named for their electrophoretic removed from the nuclear pore (Finlay and Forbes, 1990). These mobilities. They undergo a number of post-translational modifi- glycoprotein-depleted nuclear pores seem to maintain 'normal' cations, including methylation, acetylation, phosphorylation, morphology, but do not import proteins; readdition of the glyco- poly(ADP)ribosylation and glycosylation (Einck and Bustin, proteins fully restores transport activity. All of these results imply 1985). Metabolic labelling experiments indicate that the HMG that certain of the 0-GlcNAc-modified glycoproteins may be proteins contain Man, GlcNAc, Glc, Gal and fucose residues proximal to or may play a role in the nucleocytoplasmic transport (Reeves et al., 1981). Although it was suggested that the carbo- processes. hydrate moiety may play a role in the adhesion of these proteins RNA polymerase II transcription factors contain OGlcNAc to the nuclear matrix (Reeves and Chang, 1983), the precise moieties and may account for some of the WGA binding sites structure of the carbohydrate and the identity of the comple- associated with chromatin (Jackson and Tjian, 1988; Lichtsteiner mentary lectin remain to be established. and Schibler, 1989). In contrast, transcription factors for RNA Finally, it should be noted that a number of lines of evidence polymerase I (UBF1) and RNA polymerase UI (TF m A) do indicate that glycosaminoglycans can be found in the nucleus (see not appear to contain OGlcNAc. Transcription factors for several Hart et al., 1989). Of particular interest is the finding of a human promoters (Sp-1, CTF/NF-1, AP-1, AP-2 and AIM) and structurally unique form of heparan sulphate in the nuclei of Drosophila promoters (Zeste, GAGA and Adf-1) were labelled rat hepatocytes (Fedarko and Conrad, 1986). This nuclear with UDP-[3H]Gal and galactosyltransferase. The labelled poly- heparan sulphate is characterized by a large amount of an unusual peptides of AIM include c-fos, c-jun and a fos-related antigen. sulphated glucuronosyl residue. Generally, the uronic acids of The addition of WGA did not affect the DNA-binding properties the heparan sulphates are not sulphated until they are converted of Sp-1, but did reduce its transcriptional activation activity. into iduronic acids. Although it is not known where sulphation Recombinant Sp-1 isolated from E.coli is unglycosylated and is of the glucuronosyl residues occurs, the nuclear-specific structure at least three times less active than the glycosylated Sp-1 of HeLa of this heparan sulphate provides a strong argument that it is not cells. Finally, extension of the 0-GlcNAc moiety with Gal did a contaminant derived from the glycosaminoglycans of other not affect its ability to enhance transcription. All of these subcellular compartments. The question is now raised as to observations provide fascinating new insights on the roles of whether there are proteins that can recognize and distinguish glycosylation, including the possibility that it helps to maintain the specific carbohydrate structures on these nuclear glycos- the transcriptional enhancing activity of the polypeptide. aminoglycans.

249 J.L.Wang et al.

Perspectives Barondes.S.H. (1988) Bifunctional properties of lectins: lectins redefined. Trends Biochem. Sci., 13, 480-482. A new view of the intracellular compartment, in terms of Barondes.S.H. and Haywood-Reid,P.L. (1981) Externalization of an endo- carbohydrate recognition in cellular function, has been opened. genous chicken muscle lectin with in vivo development. J. Cell Biol., 91, Both glycoproteins and glycosaminoglycans, as well as proteins 568-572. Bellgrau.D., Walker.T.A. andCookJ.L. (1988) Recognition of adenovirus El A with carbohydrate binding activity, have been documented gene products on immortalized cell surfaces by cytotoxic T lymphocytes. in the cytoplasm and the nucleus. For all of the glycoconjugates J. Virol., 62, 1513-1519. observed in the nucleus, however, there has been no report of Beyer.E.C. and Barondes.S.H. (1980) Chicken tissue binding sites for a purified an endogenous lectin bound to the carbohydrate structure. chicken lectin. J. Supramol. Struct., 13, 219-227. Conversely, there is also no information on the carbohydrate Beyer.E.C. and Barondes.S.H. (1982) Quantitation of two endogenous lactose- mhibitable lectins in embryonic and adult chicken tissues. J. Cell Biol., 92, ligand with which the nuclear lectin(s) will complex. Therefore, 23-27. a significant challenge before us is to determine the respective Bols.N.C, Roberson.M.M., Haywood-Reid.P.L., Cerra.R.F. and Barondes.S.H. partners, if any, to the nuclear lectins and nuclear carbohydrates. (1986) Secretion of a cytoplasmic lectin from Xenopus laevis skin. /. Cell Biol., The characterization of the complementary recognition system 102, 492^99. may provide insights on the physiological significance of carbo- Bourgeois,C.A , Seve.A.P., Monsigny.M. and Hubert.J. (1987) Detection of sugar-binding sites in the fibnllar and the granular components of the nucleolus: hydrate recognition in intracellular function. an experimental study in cultured mammalian cells. Exp. Cell Res., 172, Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 In the course of our survey on nuclear lectins, it has also 365-376. become apparent that the basis for designating a protein as ButeU.S. and Jarvis.D.L. (1986) The plasma membrane-associated form of SV40 large tumor antigen: biochemical and biological properties. Biochim Biophys. intracellular or extracellular may need to be reconsidered. On Acta, 865, 171-195. the one hand, proteins with documented activities on DNA Carding,S.R., Thorpe.S.J., Thorpe.R. and Feizi.T. (1985) Transformation and replication and RNA metabolism have now been found at the growth related changes in levels of nuclear and cytoplasmic proteins anrjgenically cell surface. On the other, soluble factors with identified cell related to mammalian /3-galactoside-binding lectin. Biochem. Biophys. Res. surface receptors have also been observed to be in the nucleus. Commun., 127, 680-686. CattJ.W., Harrison.L.F. and CarletonJ.S. (1987) Distribution of an endogenous Thus, it appears that we may also be at the edge of discoveries 0-galactoside-specific lectin during foetal and neonatal rabbit development. with respect to novel mechanisms of protein externalization/ J. CellSa., 87,623-633. internalization and new views of subcellular compartments in Cherayil.B.J , Weiner.S.J. and Pillai.S. (1989) The Mac-2 antigen is a galactose- general. specific lectin that binds IgE. J. Exp. Med., 170, 1959-1972 Cherayil.B J., Chaitovitz.S., Wong.C. and Pillai.S. (1990) Molecular cloning of a human macrophage lectin specific for galactose. Proc. Natl. Acad. Sci. Acknowledgements USA, 87, 7324-7328. Childs.R.A. and Feizi,T. (1980) 0-Galactoside-binding lectin of human and bovine We thank Drs Kurt Drickamer, Gerald Hart, and Mel Schindkr for critical readings tissues. Cell Biol. Int. Rep., 4, 775. of the manuscript, and Mrs Linda Lang for her help in its preparation. The work Clerch.L.B., Whitney.P., Hass.M., Brew.K., MillerJ., Wemer.R. and carried out in the authors' laboratories has been supported by Grants GM-38740 Massaro.D. (1988) Sequence of a full-length cDNA for rat lung ^-galactoside- and GM-27203 from the National Institutes of Health. binding protein: primary and secondary structure of the lectin. Biochemistry, 27, 692-699. Cooper.D.N.W. and Barondes.S.H. (1990) Evidence for export of a muscle lectin Abbreviations from cytosol to extracellular matrix and for a novel secretory mechanism. J. Cell BHK, baby hamster kidney; BSA, bovine serum albumin; tBP, IgE-binding Biol., 110, 1681-1691. protein; CBP35, carbohydrate binding protein 35; CHL, ckicken heart lectin; Cowles.E.A., Agrwal,N., Anderson,R.L. and Wang.J.L. (1990) Carbohydrate CRD, carbohydrate recognition domain; Gal, /3-D-galactose; Glc, glucose; binding protein 35. Isoelectric points of the polypeptide and a phosphorylated GlcNAc, N-acetylglucosamine; HMG, high mobility group; hnRNP, heterogeneous derivative. J. Biol. Chem., 265, 17706—17712. nuclear ribonucleoprotein complex; Lac, lactose; LBP, laminin binding protein; Cnttenden.S.L., Roff.C.F. and Wang.J.L. (1984) Carbohydrate-binding protein Man, mannose; RNPs, ribonucleoproteinelements ; SV40, simian virus 40; WGA, 35' identification of the galactose-specific lectin in various tissues of mice. wheat germ agglutinin. Mol. Cell. Biol., 4, 1252-1259. Davis,L.I. and Blobel.G. (1986) Identification and characterization of a nuclear pore complex protein. Cell, 45, 699—709. References Drickamer.K. (1988) Two distinct classes of carbohydrate-recognition domains in animal lectins. J. BioL Chem., 263, 9557-9560. Abbott.W.M. and Feizi.T. (1989) Evidence that the 14 kDa soluble 0-galactoside- Einck.L. and Bustin.M. (1985) The intracellular distribution and function binding lectin in man is encoded by a single gene. Biochem. J., 259, 291-294. of the high mobility group chromosomal proteins. Exp. Cell Res., 156, Abbott.W.M., Mellor,A., Edwards.Y. and Feizi.T. (1989) Soluble bovine 295-310. galactoside-binding lectin. cDNA cloning reveals the complete amino acid Facy.P., Seve.A.P., Hubert,M., Monsigny.M. and HubertJ. (1990) Analysis sequence and an antigenic relationship with the major encephalitogenic domain of nuclear sugar-binding components in undifferentiated and in vitro dif- of myelin basic protein. Biochem. J., 259, 283-290. ferentiated human promyelocytic leukemia cells (HL60). Exp. Cell Res., 190, Agrwal.N., Wang.J.L. and Voss.P G. (1989) Carbohydrate binding protein 35. 151-160. Levels of transcription and mRNA accumulation in quiescent and proliferating Featherstone.C, Darby,M.K. and Gerace.L. (1988) A monoclonal antibody cells. J. Biol. Oiem., 264, 17236-17242. against the nuclear pore complex inhibits nucleocytoplasmic transport of protein Albrandt,K., Orida.N.K. and Liu,F.-T. (1987) An IgE-binding protein with a and RNA in vivo. J. Cell Biol., 107, 1289-1297. distinctive repetitive sequence and homology with an IgG receptor. Proc. Nail. Fedarko,N.S. and Conrad,H.E. (1986) A unique heparan sulfate in the nuclei Acad. Sci. USA, 849, 6859-6863. of hepatocytes: structural changes with the growth state of the cells. J. Cell Bachmann.M., Chang,S.-H., Slor.H., KulculiesJ. and MOller.W.E.G. (1990) Biol., 102, 587-599. Shuttling of the autoantigen La between nucleus and cell surface after UV Hnlay.D.R. and Forbes.D J. (1990) Reconstitution of biochemically altered nuclear irradiation of human keratinocytes. Exp. Cell Res., 191, 171-180. pores: transport can be eliminated and restored. Cell, 60, 17-29. Baglia.F. and Maul.G. (1983) Nuclear ribonucleoprotein release and nucleoside Finlay.D.R., Newmeyer.D.D., Pnce.T.M. and Forbes.D.J. (1987) Inhibition triphosphatase activity are inhibited by antibodies directed against one nuclear of in vitro nuclear transport by a lectin that binds to nuclear pores. J. Cell matrix glycoprotein. Proc. Natl. Acad. Sci. USA, 80, 2285-2289. Biol., 104, 189-200. Barak-Briles,E., Gregory.W., Fletcher,P. and Komfeld.S. (1979) Vertebrate Fitzgerald.J.E., Catt.J.W. and Harrison,F.L. (1984) Isoforms of an endogenous lectins. Comparison of properties of/3-galactoside-binding lectins from tissues lectin in rabbit bone marrow. Eur. J. Biochem., 140, 137-141 of calf and chicken. J. Cell Biol., 81, 528-537 Frigcri.L.G., Robertson.M.W. and Liu,F.-T. (1990) Expression of biologically Barondes.S.H. (1984) Soluble lectins: a new class of extracellular proteins. Science, active recombinant rat IgE-binding protein in Escherichia coli. J Biol. Chem., 223, 1259-1264. 265, 20763-20769.

250 Nuclear lectins

Gitt.M.A. and Barondes.S.H. (1986) Evidence that a human soluble j3-galactoskJe- Lichtsteiner.S. and Schibler.U (1989) A glycosylated liver-specific transcription binding lectin is encoded by a family of genes. Proc. Naxl. Acad. Sci. USA, factor stimulates transcription of the albumin gene. Cell, 57, 1179-1187. 83, 7603-7607. Mattaj.I.W. and DeRobertis.E.M. (1985) Nuclear segregation of U2 snRNA Gritzmacher.C.A., Robertson.M.W. and Liu,F.-T. (1988) IgE-binding protein. requires binding of specific snRNP proteins. Cell, 40, 111-118. Subcellular location and gene expression in many murine tissues and cells. Monsigny.M., Kieda.C. and Roche.A.C. (1983) Membrane glycoproteins J. Immunol, 141, 2801-2806. glycolipids and membrane lectins as recognition signals in normal and malignant Grundmann.U., Amann.E., Zettlmeissl.G. and Kupper.H.A. (1986) cells. Biol. Cell, 47, 95-110. Characterization of cDNA coding for human factor XIRA. Proc. Natl. Acad. Moutsatsos.I.K., Davis.J.M. and WangJ.L. (1986) Endogenous lectins from Sci. USA, 83, 8024-8028. cultured cells: subcellular localization of carbohydrate-binding protein 35 in Haltiwanger.R.S., Holt.G.D. and Hart.G.W. (1990) Enzymatic addition of 3T3 fibroblasts. J. Cell Biol, 102, 477-483. 0-GlcNAc to nuclear and cytoplasmic proteins. Identification of a undine Moutsatsos.I.K., Wade.M., Schindler.M. and Wang,J.L. (1987) Endogenous diphospho-yV-acetylglucosamine:peptide /3-W-acetylglucosaminyltransferase. lectins from cultured cells: nuclear localization of carbohydrate-binding J. Biol. Chem, 265, 2563-2568. protein 35 in proliferating 3T3 fibroblasts. Proc. Natl. Acad. Sci. USA, 84, Hart.G.W., Haltiwanger.R.S., Holt.G.D. and Kelly,W.G. (1989) Glycosylation 6452-6456. in the nucleus and cytoplasm. Annu. Rev. Biochem., 58, 841-874. Oda.Y., Leffler.H., Sakakura.Y., Kasai,K. and Barondes.S.H. (1991) Human Hart.G.W., Holt.G.D. and Haltiwanger.R.S. (1988) Nuclear and cytoplasmic breast carcinoma cDNA sequence encoding a galactoside-binding lectin glycosylation: novel saccharide linkages in unexpected places. Trends Biochem. homologous to mouse Mac-2 antigen. Gene, 99, 279-283. Sci., 139, 380-384. Olins,D.E., OlinsAL-, Seve^A.P., Bourgeois.C.A., Hubert^, and Monsigny.M. Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 Hirabayashirl., Kawasaki,H., Suzuki,K. and Kasai.K. (1987) Complete ammo (1988) Lectin-like components in the macronuclear replication bands of Euplotes acid sequence of 14 kDa ^-galactoside-binding lectin of chick embryo. eurystomus. Biol Cell, 62, 95-98. J. Biochem., 101, 775-787. Paroutaud.P., Levi,G., Teichberg.V.I. and Stromberg.A.D. (1987) Extensive Hirano.H., Akimoto,Y., Kawakami.H., Oda.Y. and Kasai.K. (1988) Localization amino acid sequence homologies between animal lectins. Proc. Natl Acad. of endogenous /S-galactoside-binding lectin in embryonic chick epidermis. In Sci. USA, 84, 6345-6348. Proceedings of the Kagoshima International Symposium of dycoconjugates Raz.A. and Lotan.R. (1987) Endogenous galactoside-binding lectins: a new class in Medicine Professional Postgraduate Service International, Tokyo, pp. 8—13. of functional tumor cell surface molecules related to metastasis. Cancer Ho,M.K. and Springer.T A. (1982) Mac-2, a novel 32,000 M, mouse Metastasis Rev., 6, 433^»52. macrophage subpopulation-specific antigen defined by monoclonal antibodies. Raz.A., Carmi,P. and Pazerirri.G. (1988) Expression of two different endogenous J. Immunol , 128, 1221-1228. galactoside-binding lectins sharing sequence homology. Cancer Res., 48, Holt.G.D. and Hart,G.W. (1986) Subcellular distribution of terminal 645-649. N-acetylglucosamine moieties. Localization of a novel protein-saccharide Raz.A., Pazerini.G. and Carmi.P. (1989) Identification of the metastasis-associated linkage, O-linked GlcNAc. J. Biol. Chem., 261, 8049-8057. galactoside-binding lectin as a chimenc gene product with homology to an Holt.G.D., Snow.C.M., Senior,A., Haltiwanger.R.S., Gerace.L. and Hart.G.W. IgE-binding protein. Cancer Res., 49, 3489-3493. (1987) Nuclear pore complex glycoproteins contain cytoplasmically disposed Reeves,R. and Chang,D. (1983) Investigations of the possible functions for O-linked W-acetylglucosamine. J. Cell Biol., 104, 1157-1164. glycosylation in the high mobility group proteins. Evidence for a role in nuclear Hopp.T.P. and Woods,K.R. (1981) Prediction of protein antigenic determinants matrix association J. Biol. Chem., 258, 679-687. from amino acid sequences. Proc. Natl. Acad. Sci. USA, 78, 3824-3828. Reeves,R., Chang,D. and Chung,S.-C. (1981) Carbohydrate modifications of Hubert.J., Seve.A.P., Bouvier.D., Masson.C, Bouteille.M. and Monsigny.M. the high mobility group proteins. Proc. Natl Acad. Sci. USA, 78, 6704-6708. (1985) In situ ultrastructural localization of sugar-binding sites in lizard granulosa Regan,L.J., DoddJ., Barondes.S.H. and Jessell.T.M. (1986) Selective expression cell nuclei. Biol. Cell, 55, 15-20. of endogenous lactose-binding lectins and lactoseries glycoconjugates in subsets Hubert.J., Seve.A.P., Facy,P. and Monsigny.M. (1989) Are nuclear lectins and of rat sensory neurons. Proc. Natl. Acad. Sci. USA, 83, 2248-2252. nuclear glycoproteins involved in the modulation of nuclear functions? Cell Robertson.M.W., Albrandt.K., KeUer.D. and Liu,F.-T. (1990) Human IgE- Differ. Dev., 27, 69-81. binding protein: a soluble lectin exhibiting a highly conserved interspecies Hynes.M.A., Gitt.M., Barondes.S.H., Jessell.T.M. and Buck.L.B. (1990) sequence and differential recognition of IgE glycoforms. Biochemistry, 29, Selective expression of an endogenous lactose-binding lectin gene in subsets 8093-8100. of central and peripheral neurons. J. Neurosci., 10, 1004-1013. Roff.C.F. and WangJ.L. (1983) Endogenous lectins from cultured cells. Isolation Ishikawa.F., Miyazono.K., Hellman.U., Drexler.H., Wernstedt,C, Hagiwara.K., and characterization of carbohydrate-binding proteins from 3T3 fibroblasts. Usuki.K., Takaku.F., Risau.W. and Heldin,C.-H. (1989) Identification of J. Biol. Chem., 258, 10657-10663. angiogenic activity and the cloning and expression of platelet-derived endothelial Rubartelli, A., Cozzolino, F., Talio, M. and Sitia, R. (1990) A novel secretory cell growth factor. Nature, 338, 557-562. pathway for interleukin-1/3, a protein lacking a signal sequence. EMBO J., Jackson, S.P. and Tjian, R. (1988) OGIycosylation of eukaryotic transcription 9, 1503-1510. factors: implications for mechanisms of transcriptional regulation. Cell, 55, Runyan, R.B., Versalovic, J. and Shur, B.D. (1988) Functionally distinct laminin 125-133. receptors mediate cell adhesion and spreading: The requirement for surface Jarvis.D.L. and Butel.J.S. (1988) Biochemical properties of SV40 large tumor galactosyltransferase in cell spreading. J. Cell Biol, 107, 1863-1871. antigen as a glycosylated protein. J. Biol. Chem., 263, 15288-15296. Sakakura.Y., Hirabayashi J., Oda,Y., Ohyama.Y. and Kasai.K. (1990) Structure Jia.S. and WangJ.L. (1988) Carbohydrate binding protein 35. Complementary of chicken 16 kDa /J-galactoside-binding lectin: complete amino acid sequence, DNA sequence reveals homology with proteins of the heterogeneous nuclear cloning of cDNA and production of recombinant lectin. J. Biol. Chem., 265, RNP. J. Biol. Chem., 263, 6009-6011. 21573-21579. • Kelly,W.G. and Hart.G.W. (1989) Glycosylation of chromosomal proteins: Sano.H., Forough.R., Maier.J.A.M., CaseJ.P., Kackson.A., Engleka.K., localization of O-linked A'-acetylglucosamine in Drosophila chromatin. Cell, Maciag.T. and Wilder,R.L. (1990) Detection of high levels of heparin binding 57, 243-251. growth factor-1 (acidic fibroblast growth factor) in inflammatory arthritic joints. Kuchler.K., Steme.R.E. and Thoma\J. (1989) Saccharomyces cerevisiae STE6 J. Cell Biol, 110, 1417-1426. gene product: a novel pathway for protein export in eukaryotic cells. EMBO J., Schindler.M., Hogan.M., Miller.R. and DeGaetano.D. (1987) A nuclear specific 8, 3973-3984. glycoprotein representative of a unique pattern of glycosylation. /. Biol Chem., LaingJ.G. and Wang.J.L. (1988) Identification of carbohydrate binding protein 262, 1254-1260. 35 in heterogeneous nuclear ribonucleoprotein complex. Biochemistry, 27, Schmitt,M.K. and Mann.K. (1987) Glycosylation of simian virus T antigen and 5329-5334. localization of glycosylated T antigen in the nuclear matrix. Virology, 156, LaingJ.G., Robertson.M.W., Grhzmacher.C.A., WangJ.L. and Uu,F.-T. (1989) 268-281. Biochemical and immunological comparisons of carbohydrate-binding protein Seve.A.P., Hubert^., Bouvier.D., Bouteille.M., Maintier.C. and Monsigny.M. 35 and an IgE-binding protein. /. Biol. Chem, 264, 1907-1910. (1985) Detection of sugar-binding proteins in membrane-depleted nuclei. Exp. Lau.L.F. and Nathans, D. (1987) Expression of a set of growth-related immediate Cell Res., 157, 533-538. early genes in Balb/c 3T3 cells: coordinate regulation with c-fos or c-myc. Seve,A.P., Hubert^., Bouvier.D., Bourgeois.C, Midoux.P., Roche.A.C. and Proc. Natl. Acad. Sci. USA, 84, 1182-1186. Monsigny, M. (1986) Analysis of sugar-binding sites in mammalian cell nuclei Leffler.H. and Barondes.S.H. (1986) Specificity of binding of three soluble rat by quantitative flow microfluorometry. Proc. Natl. Acad. Sci. USA, 83, lung lectins to substituted and unsubstituted mammalian /3-galactosides. J. Biol 5997-6001. Chem., 261, 10119-10126. Sparrow.C.P., Leffler.H. and Barondes.S.H. (1987) Multiple soluble Leffler.H., Masiaiz.F.R. and Barondes.S.H. (1989) Soluble lactose-binding ^-galactoside-binding lectins from human lung. J. Biol Chem, 262, vertebrate lectins: a growing family. Biochemistry, 28, 9222-9229. 7383-7390.

251 J.L.Wang et al.

Spence.A.M., Sheppard.P.C, DavieJ.R., Matuo.Y , Nishi.N., McKeehan.W.L., DoddJ.G. and Matusik.RJ. (1989) Regulation of a bifunctional mRNA results in the synthesis of secreted and nuclear probasin. Proc. Natl Acad. Sci. USA, 86, 7843-7847. Tagaya,Y., Maeda.Y., MitsuiA. Kondo.N., Matsui.H., HamuroJ., Brown.N., Arai.K., Yokota.T., Wakasugi.H. and YodoiJ. (1989) ATL-derived factor (ADF), an E--2 receptor/Tac inducef homologous to thMredoxin; possible involvement of dithiol-reduction in the IL-2 receptor induction. EMBO J., 8, 757-764. Tuil.D., Clergree.N., Montarras.D., Pinset.C, Kahn.A. and Phan-Dinh-Tuy.F. (1990) CC Ar GG Boxes, cis-acting elements with a dual specificity. Muscle- specific transcriptional activation and serum responsiveness. J. Mol. Biol., 213, 677-686. Velazquez J.M. and Lindquist.S. (1984) hsp 70: Nuclear concentration during environmental stress and cytoplasmk storage during recovery. Cell, 36, 655-662. WUson.T.J.G., Firth.M.N., Powell J.T. and Harrison.F.L. (1989) The sequence of the mouse 14 kDa, |S-galactoside-binding lectin and evidence for its synthesis on free cytoplasmic ribosomes. Biochem. J., 261, 847-852. Downloaded from https://academic.oup.com/glycob/article/1/3/243/610598 by guest on 27 September 2021 Woo,H.-J., Shaw.L.M., Messier.J.M. and Mercurio.A.M. (1990) The major non-integrin laminin binding protein of macrophages is identical to carbohydrate binding protein 35 (Mac-2). J. Biol. Chan., 265, 7097-7099. Yoneda,Y., Imamoto-Sonobe.N., Yamaizumi.M. and UchidaJ. (1987) Reversible inhibition of protein import into the nucleus by wheat germ agglutinin injected into cultured cells. Exp. Cell Res., 173, 586-595.

Received on January 7, 1991; accepted on January 18, 1991

252