Purified nuclei retaining a high degree of ultrastructural integrity were isolated by conventional centrifugation techniques. The cytoplasmic surface of these nuclei was iodinated using lactoperoxidase immobilized onto giant Sepharose beads; thus the outer nuclear membrane and the cytoplasmic surface of nuclear pore complexes were selectively labelled. Pore complexes in association with a fibrous lamina were isolated from these nuclei by removal of the nucleoplasm and extraction with Triton X-100. The chemical composition of the pore-lamina fraction was 93·6% protein, 6% RNA, 0.4% phospholipid. The labelling suggests that major polypeptides Ni (70000) and N2 (67000) and more than 10 other more minor polypeptides, ranging from 3 3 000 to 200 000 mol. wt, as being components of the nuclear pore complex. Polypeptide N3 (58000) is shown to be present only on the nucleoplasmic face of nuclear envelopes, probably in the fibrous lamina.
The nuclear envelope is a characteristic feature of the eukaryote cell. It comprises 2 concentric membranes, pore complexes which punctuate the membranes and an underlying fibrous lamina. The outer nuclear membrane is continuous with the inner nuclear membrane at the level of the pore complex (Watson, 1955), bears ribosomes (Watson, 1955; Palade, 1955) and provides a surface to which structural elements of the cytoplasm may attach (for references see Franke & Scheer, 1974). The inner nuclear membrane closely apposes the fibrous lamina (Fawcett, 1966; Aaronson & Blobel, 1975), which is contiguous with the pore complexes and represents the peripheral aspect of the nuclear matrix (Berezney & Coffey, 1977).
The nuclear pore complex is not a discrete structure (which is why a bulk method for its isolation has proved elusive) for it is connected both with the fibrous lamina and with the inner and outer nuclear membranes at their point of fusion. Its structure has been studied in detail for more than a decade, and the main reason it has attracted so much interest is that it is widely believed to be the principle pathway by which nascent RNA is transported from the nucleus to the cytoplasm. Despite at least 2 attempts (Aaronson & Blobel, 1975; Harris, 1977), the pure pore complex has not been isolated from mammalian cells although a highly enriched fraction has been obtained from Xenopus laevis oocytes (Krohne, Franke & Scheer, 1978,b). Analysis of the nuclear envelope’s constituent polypeptides has extended little beyond the relatively trivial establishment of its electrophoretic profile. There is a great deal of information about enzyme activities associated with nuclear envelope preparations (for refs see Franke, 1974 a, b; Harris, 1978), some of which may catalogue cross-contamination from other membrane systems, but such information loses much of its value in the absence of detail as to the location of components.
We have therefore sought to develop a suitable labelling method to study the orientation of the polypeptides. Several methods have been developed for labelling the cell surface (Hubbard & Cohn, 1976; Hynes, 1976) but none of these may be applied immediately to the labelling of the nuclear surface. The great structural complexity of the nuclear envelope and its permeability, even to macromolecules (Bonner, 1975 a, b; De Robertis, Longthorne & Gurdon, 1978; Paine, Moore & Horowitz, 1975), set problems which are without parallel in a membrane-labelling study. Simple chemical probes and soluble enzyme labelling methods are suspect since they might gain access both to the cisternal surfaces of the 2 membranes and to the nucleoplasm (via the pore complexes). We have overcome this problem by immobilizing the labelling enzyme lactoperoxidase onto giant Sepharose beads. As iodination catalysed by lactoperoxidase takes place at the enzyme surface via an indiffusible activated I -the iodinating sites are thus confined to beads several times larger than the nuclei. The beads are too large to pass through the pore complex or into the perinuclear cisternum of isolated nuclei and are restricted to the nuclear surface where their lactoperoxidase may catalyse the radio-iodination of pore complex and outer nuclear membrane polypeptides. In conjunction with a procedure for isolating pore complexes in association with the fibrous lamina, such labelling has enabled examination of pore complex and fibrous lamina polypeptides. It has also permitted the comparison of outer nuclear membrane and rough endoplasmic reticulum polypeptides of the same membrane plane, which is to be reported in a subsequent paper.
MATERIALS AND METHODS
Preparation of nuclear envelopes
Nuclei from rat livers were isolated by a dense sucrose procedure similar to that of Kay, Fraser & Johnston (1972).
Nuclear envelopes were prepared by a double DNase digestion procedure modified from Kay et al. (1972).
1st DNase digestion (pH 8 g)
A pellet of nuclei derived from 20 g of rat liver was resuspended by the addition of a few drops of glycerol and vortexing. To the suspension were added, with vigorous vortexing, 7·5 ml H,O followed by 375 µl DNase I (100 μ g/ml H2O; Sigma type DNEP) and 30 ml of a solution of 10% sucrose, 10 m M Tris.HCl, 01 mM MgCl2, and 0.1 rπM PMSF, pH 8·5. The mixture was incubated at 22 °C for 15 min with vortexing every 5 min. After 15 min the digestion was slowed by the addition of 40 ml ice-cold distilled water and the suspension centrifuged at 40000 g max,i for 15 min at 4 °C yielding a supernatant and pellet (DNA1 pell).
2nd DNase digestion (pH 7·5)
The pellet was resuspended, using a syringe and fine-gauge needle, into 7·5 ml of a solution of 10% sucrose, 10 m M Tris.HCl, 01 mM MgCls and 0.i mM PMSF, pH 7·5. To this suspension 375 µ\ DNase (10o ⁄tg/ml) were added. After incubation for 20 min at 22 °C the digestion was slowed by the addition of 9 ml ice cold distilled water and the suspension centrifuged for 10 min at 4 °C at 2 0 000 g. in a 10 × 10 titanium rotor (MSE rotor 4 3114-12s), yielding a supernatant and a pellet (DNAS pell) of nuclear envelopes.
Isolation of pore-complex lamina fraction (Modified from Dwyer & Blobel, 1976)
Triton X-10o wash of nuclear envelopes: The DNase-digested pellet (DNA2pell) was thoroughly resuspended into 2·5ml of ice-cold solution of 10% sucrose, 0·1 mM MgCl2, 10 mM Tris.HCl, pH 7·5 to which 0·25 ml 20% (v/v) Triton X-10o (British Drug Houses Scintillation grade) was added with vortex mixing. Incubation of the mixture on ice for 10 min followed by centrifugation for 10 min at 4 °C at 2ooo∞ g in the 10 × 10 titanium rotor yielded a supernatant and a pellet of crude pore laminae.
High salt extraction
The resulting pellet was gently, but thoroughly resuspended into 0.2 ml 10% sucrose, 01 mM MgCl2, 10 mM Tris.HCl pH 7·5. Homogeneous resuspension was essential. (If the pellet was resuspended directly into the high salt medium, it tended to clump and the preparation remained contaminated with nucleoplasm). To this suspension were added 2·5 ml 10% sucrose, 2·o M NaCl, 01 mM MgCl2, 10o HIM Tris.HCl pH 7·5. Incubation of the mixture for 10 min on ice, followed by centrifugation as above yielded a pellet of purified pore complexlamina fraction.
Iodination conditions for the complete system per millilitre of final solution: nuclei from 0.333 g liver, 1 µmol glucose, 33 µg lactoperoxidase coupled onto Sepharose in the ratio of 1-33 mg LPO per ml of settled beads (see below), 0.7 µg glucose oxidase (Sigma, Type V), 33 µCi Nal25I (Amersham. Radiochemicals) in 10% sucrose, 0.oooi% butylated hydroxytoluene (from a stock of 05% in ethanol), 20 µMK1”I, 1 mM glucose, 10 mM Tris.HCl, pH 7·2. Incubation was 12 min at 23 °C in a test tube rotating end-over-end at 4 rev/min. The reaction was stopped by the addition of an equal volume of ice-cold stopper buffer (10% sucrose, 0.0001% butylated hydroxytoluene (Welton & Aust, 1972) mM 3-amino, 1,2,4-triazole (Harris, 1978), 20 µM sodium sulphite, 10 m M Tris.HCl, pH 7·2). The mixture was filtered through 80-µm mesh nylon gauze to remove the Sepharose beads, under-layed with i vol. of 20% sucrose, 10 mM β-mercaptoethanol in stopper buffer, and centrifuged to pellet the nuclei (1000 g for 10 min at 4 °C in a 6 × 100-ml swing-out rotor). The supernatant was discarded, and the nuclei were washed twice in 2 vol. 10% sucrose in stopper buffer by pelleting at 7∞ g for 5 min in the same rotor at 4 °C.
Immobilization of lactoperoxidase
One gramme of CNBr-activated Sepharose 6MB (Pharmacia Fine Chemicals) was swollen in a beaker and washed for 15 min on a glass fibre filter with 1 mM HC1 (200 ml). Lactoperoxidase (Sigma, lyophilized powder) dissolved in 0. 1 M sodium phosphate buffer (pH 7·2), was mixed with the gel in a test tube, and the mixture rotated endover-end at 4 rev/min overnight at 4 °C. Unbound material was washed away with 200 ml phosphate buffer (coupling efficiency was always greater than 99·9%), and any remaining CNBr groups were reacted with 1 M glycine for 2 h at room temperature. Three washing cycles were used to remove any possible non-covalently adsorbed protein (none was ever detected), each cycle consisting of a wash in 0.2 M sodium phosphate buffer (pH 7·2) followed by a wash in I M glycine. Lastly, the beads were washed with 200 ml 10% sucrose, 1 mM MgCl2, 0· 2 mM NaHCO3 (pH 7·4) and stored for up to 4 h prior to use.
TCA-precipitated samples on glass fibre disks were counted in a Nuclear Enterprises gamma counter (efficiency about 75%).
Protein was assayed by the method of Lowry, Rosebrough, Farr & Randall (1951) with bovine serum albumin as standard.
DNA was measured by Giles and Myers modification (Giles & Myers, 1965) of the method of Burton (1956), with deoxyadenosine monophosphate as standard.
RNA was assayed by a modification of the orcinol method (Richardson, 1979).
Phospholipid phosphorous was determined according to Chen, Toribara & Warner (1956) on lipid samples extracted according to Bligh & Dyer (1959) and evaporated to dryness.
Succinate dehydrogenase activity of freeze-thawed and briefly sonicated samples was assayed by the reduction of phenazine methosulphate (Singer, 1975).
Polyacrylamide gel electrophoresis
Samples were prepared for electrophoresis by precipitation in 2 vol. ethanol at—20 °C for I 6 h in order to decrease the presence of detergent and salts. The alcohol precipitate was pelleted and resuspended into 5 vol. of a solution containing 3% w/v SDS, 5% v/v 2-mercap-toethanol, 20% v/v glycerol, 1 r∏M EDTA and 62·5 r∏M Tris.HCl (pH 6·8). The samples were then incubated for 10 min at 70 °C and for 5 min at 100 °C. Particulate material remaining after this time was removed by centrifugation at 3000 g for 5 min.
Analytical SDS/polyacrylamide-gel electrophoresis was carried out in the buffer system of Laemmli (1970) in vertical slab gels (15 × 15 × 0.2 cm) cast between glass plates, comprising a 1·5-cm stacking gel (3·75% w/v acrylamide, 0 · 1%w/v N,N′-methylene bisacrylamide) and a 13·5-cm resolving gel (16% w/v acrylamide, 0 · 09% w/v N,N′-methylene-bisacrylamide). The gels were polymerized by addition of ammonium persulphate and N,N′,N′,-N′-tetramethylethylenediamine. After electrophoresis gels were fixed and stained according to Fairbanks, Steck & Wallach (1971).
Gels were dried onto thick filter paper under vacuum at 90 °C. Dry gels were either exposed directly to X-ray film (Kodak X-Omat H film) or the film was first flash exposed, backed with an intensifying screen (Ilford Fast Tungstate) and closely apposed to the gel in a cassette at— 70 °C for approx 3 days. E. coli β-galactosidase (130000), bovine serum albumin (68000), chick brain tubulin (5 5 00 0), rabbit muscle actin (4 6 00 0) and lactate dehydrogenase (3 5 000), bovine β-lactoglobulin (175oo) and heart cytochrome C (12500) were used as standards for mol. wt determinations.
Lipid extracts were run in 2 dimensions (Zwaal & Roelofsen, 1976) on 8 × 8 cm TLC plates (Polygram SIL NHR, Camlab). The distribution of radioactivity was determined by autoradiography.
Samples were fixed with glutaraldehyde and osmium (Dwyer & Blobel, 1976), embedded in Araldite and sectioned using a diamond knife. Sections were stained with uranyl acetate (Watson, 1958) and lead citrate (Reynolds, 1963).
Electron micrographs were taken with a Philips EM 300 operating at 80 kV. Kodak ‘Estar’ sheets were used in the camera.
Morphometric determinations of nuclear integrity and membranous contamination were carried out on electron micrographs taken at magnifications between × 3 000 and × 5 000 (Richardson, 1979)· Negatives were displayed on a microfilm reader and examined at magnifications between 6·5 and 17·5 diameters. Length measurements were made at 6·5 diameters using a ‘Map Measure’ and converted to microns original membrane.
Nuclear membranes were classified as membrane profiles which were associated with nuclear chromatin in at least one site, and which contained pore complexes (Franke et al. 1976). The circumference of very small vesicles was approximated to 3 times the longest axis.
RESULTS AND DISCUSSION
Preparation of nuclei
Nuclei were prepared at a yield greater than 75% (estimated by the recovery of DNA) by chemically mild means. Electron micrographs of purified nuclei showed little cytoplasmic membrane contamination and the nuclei were rounded and showed only minor disruption of the nuclear envelope. Succinate dehydrogenase activity of the preparation was typically 0 · 5 units (µmol dichloroindophenol reduced per min per mg of protein) i.e. less than 0 · 5% of the activity of purified mitochondria. The proportion of membrane profiles which in thin section were clearly nuclear membrane profiles was determined by morphometric means (see Table 1); on average, this was 94% -a value close to that of Franke et al. (1976).
Iodination of nuclei
It was important to establish that the iodination procedure did not strip away large portions of the outer nuclear membrane and that the conditions of iodination were such as to provide a high specific activity. Morphometric determinations of the proportion of outer nuclear membrane covering the nuclear surface were made before and after the iodination reaction (Table 2). Approximately 89% of the nuclear surface was covered with outer nuclear membranes prior to iodination, and this value was not significantly different in nuclei recovered from the iodination reaction. After iodination, the nuclei were still largely rounded, showed intact pore-complexes, and retained their ribosomes on the outer nuclear membrane (Fig. i). The outer nuclear membrane occasionally showed an increased tendency toward a bleb separation from the inner nuclear membrane; a phenomenon not generally seen prior to iodination (although see Kartenbeck, Jarasch & Franke, 1973). This was not seen in all preparations and the outer nuclear membrane frequently remained closely apposed to the inner nuclear membrane after iodination.
The nuclei were iodinated to a level of 106 cmp/mg protein. Deletion of either the peroxide-generating system or lactoperoxidase resulted in less than 3·5% of the iodide incorporation achieved under standard iodinating conditions. The inclusion of carrier iodide in the reaction was essential both in order to obtain a satisfactory level of radioiodination and to ensure that the ratio of lactoperoxidase-dependent to lactoperoxidase-independent labelling was high. In the absence of carrier iodide, the ratio was about 2:1, whereas if carrier iodide was included at a concentration of 20 µM this ratio could be raised to greater than 25:1 with a more than 50-fold stimulation of 125I incorporation (cf. Hubbard & Cohn, 1976).
If the temperature of the reaction was raised from 6 ° to 23 °C, the level of iodination increased 7-fold to an incorporation efficiency of greater than 10% into trichloroacetic acid-precipitable material. Thus all iodinations were conducted at this latter temperature.
Endogenous generation of hydrogen peroxide from glucose within the reaction medium led to significantly greatly incorporation efficiencies than did the single addition of peroxide to 8 µM. The use of glucose oxidase to generate peroxide at low levels seemed preferable to the addition of concentrated peroxide which would create spatial and temporal gradients of peroxide and can result in lipid peroxidation and loss of enzyme activity (Welton & Aust, 1972). As the pattern of iodination was identical for both methods glucose oxidase does not itself contribute to the iodination pattern.
The time course of the iodination reaction was studied over a period of more than 20 min. Longer reaction times gave greater incorporation efficiencies and a greater lactoperoxidase-dependent/lactoperoxidase-independent incorporation ratio but under standardized conditions a reaction time of 12 min was found sufficient to allow adequate iodination at an acceptable lactoperoxidase-dependent/lactoperoxidase-independent ratio (~ 30:1) and was consistent with the need to ensure minimal damage to the nuclei. Sepharose 6-MB beads have the advantage of being easily and simply removed from the nuclei after iodination.
If iodination of nuclei were to occur by a non-enzymic route via oxidation of iodide to the highly reactive iodine which permeates cells, then iodination of the lipids would be expected (Hubbard & Cohn, 1976). It was thus essential to establish that iodination did not occur by such a reactive diffusible moeity, which would abolish the specificity of labelling. In 3 separate experiments less than 10% of the counts associated with washed labelled nuclei could be extracted with organic solvents by the Bligh & Dyer (1959) procedure. When this extract was chromatographed in 2 dimensions on thin layer plates, no significant portion of the radioactivity co-migrated with the 3 major nuclear phospholipids (phosphatidylcholine, phosphatidylinositol, phosphatidylethanolamine) which, together, account for approximately 93% of nuclear envelope phospholipid (Kleinig, 1970). Greater than 95% of the radioactivity ran just behind and within the 2 solvent fronts, and probably represented unbound iodide.
Pore lamina fraction
The pore lamina appears as an extensive meshwork of densely staining pore complexes connected by fine fibrillar threads (Fig. 2). Some pore complexes retain an internal structure comprising a central granule and centripetal elements but such detail is usually difficult to discern. Thus although nuclear pore complexes seated on a fibrous lamina are easily identifiable, they do not exhibit such a clear degree of organization as that seen in the micrographs of Dwyer & Blobel (1976).
The chemical composition of the preparation (93·6% protein, 6% RNA, 0.4% phospholipid, DNA not present in detectable amounts -less than 1%) differs somewhat from that of Dwyer & Blobel, containing less DNA and rather more RNA. RNA accounted for 6% of purified pore-lamina (2% in the Dwyer & Blobel study) and this is despite extraction in conditions that completely remove DNA and the outer nuclear membrane with its associated ribosomes. There is some experimental evidence to indicate that RNA is located in the pore complex (Mentré, 1969; Scheer, 1972; Franke & Scheer, 1974; Agutter, Harris & Stevenson, 1977) but whether it is RNA or some other factor that contributes to the astonishing structural stability of the pore complex is unknown.
Comparison of the Coomassie brilliant blue and autoradiograph patterns of the polypeptides from whole nuclei, of nuclear envelopes and the pore lamina fraction derived from labelled nuclei (Fig. 3) shows that the labelling of polypeptides is very selective. In particular, major iodinated bands co-migrate with 2 of the major nuclear envelope polypeptides (N1 and N2) but not with a third (N3). Both soluble and immobilized lactoperoxidase were ineffective at labelling N3 in intact nuclei and this polypeptide could only be labelled when nuclei were first broken open and the bulk of the nucleoplasm was removed. When this was done, N3 became highly labelled and the overall pattern of labelling was greatly altered (Fig. 4).
For the following reasons it is concluded that the labelling pattern in Fig. 3 is specific to the cytoplasmic surface of intact nuclei. (1) Insoluble lactoperoxidase is absolutely impermeable. (2) Labelling was dependent upon the presence of lactoperoxidase and of a peroxidase-generating system. (3) Lipid labelling was not detected, indicating the absence of diffusible I2. (4) Morphometric analysis of iodinated nuclei showed that approximately 88% of the surface of nuclei was covered by outer nuclear membrane. (5) The pattern of labelling was highly selective and dependent upon nuclei being intact. When nuclei were broken open, further polypeptides were iodinated and the overall iodination pattern was substantially altered. (6) All the polypeptides labelled in the nucleus are retained in the envelope preparation -the 2 autoradiographs are virtually identical but the Coomassie blue patterns of the 2 preparations are quite different. The pore lamina autoradiograph is very similar to the envelope. This is because the bulk of the protein of the envelope belongs to the lamina, the Tritonsoluble fraction being quantitatively small.
Morphometric analysis has indicated that 88% of the nuclear surface is covered by the outer nuclear membrane (Table 2), and that 6% of the membrane profiles exposed to lactoperoxidase beads is unidentified single membranes: 97% of radioactive counts were dependent on the presence of lactoperoxidase. Thus not less than 80% (88 × 94x 97%) of radioactive counts may with confidence be ascribed to the cytoplasmic surface of intact nuclei. It will be difficult to significantly improve upon this figure since 2 major limiting factors, purity of the preparation and integrity of the outer nuclear membrane, are almost conflicting requirements.
The iodination patterns of isolated nuclei and their subfractions (Fig. 3) reveal the selectivity of the labelling method. The labelling patterns of nuclei, and the purified pore-lamina fraction are, with the exception of 2 low-molecular-weight polypeptides, almost identical. The 2 low-molecular-weight components are coincident with histones H2b and H4 and might represent iodination of chromatin in leaky nuclei. However, when chromatin is actually made accessible, an entirely different pattern of labelling is seen where all the histones are labelled, although to varying extents (Fig. 4). The pore lamina material has been rigorously extracted in low and high salt solutions and with detergent. Triton X-1∞ extraction, which removes the outer nuclear membrane, removes only a minor portion of protein-bound label (~ 10%) although it removes more than 95% of membrane phospholipid. Although precipitation of outer nuclear membrane proteins onto the pore lamina during Triton extraction might explain the low proportion of radioactive counts removed by this procedure, we think this improbable. The evidence that Triton effectively removes the outer membrane from the pore lamina is substantial (Aaronson & Blobel, 1974, 1975; Dwyer & Blobel, 1976; Kartenbeck et al. 1973; Tata, Hamilton & Cole, 1972). It would appear therefore that the labelling procedure places label predominantly in the nuclear pore complex and to a rather lesser extent in the outer nuclear membrane. Such apparent selectivity may be explained by the fact that the pore complex sits well proud of the outer nuclear membrane and its prominence may reduce the accessibility of membrane proteins to lactoperoxidase beads. Furthermore, Triton extraction of highly purified nuclear envelopes (see subsequent paper) shows that very little protein may be extracted by Triton and that the bulk of nuclear envelope protein is associated with the pore complex and its lamina. Thus, not only does the pore complex present more protein to the nuclear surface than does the outer nuclear membrane, but it also presents it in a more accessible manner.
From the iodination pattern (Fig. 3), bands N1 (70 000 mol. wt) and N2 (67 ∞o mol. wt) of the major triplet may be identified as being externally disposed proteins of the cytoplasmic surface of the nuclear pore complex. As such, these are the first polypeptides in the mammalian cell nuclear pore complex as distinct from porelamina, to be identified. Pore-complex polypeptides may also be identified at 2 00 0 00, 160000, 118000, 97∞o, 88000, 51000, 47000, 38000, 36000, and 33000 mol. wt. The specific activity of these latter polypeptides is greater than for N1 or N2 which suggests that these are more highly exposed than Ni and N2. Significantly, N3, one of the major polypeptides of the pore-lamina fraction remains unlabelled. It could perhaps be a pore-lamina polypeptide which, buried deep within the pore complex remains inaccessible to the lactoperoxidase bead labelling system. This is unlikely to be the case however for, although N3 is labelled neither by free nor by immobilized lactoperoxidase if nuclei are intact, it is heavily labelled when lactoperoxidase beads have access to the nucleoplasmic surface of the envelope after breakage of the nuclei and removal of the bulk of the nucleoplasm (Fig. 4). Since laminal material is the main component of this surface, and because the great size of the lactoperoxidase beads would preclude their gaining access to the interior of the pore complex via the nucleoplasmic side, it is likely that N3 is a polypeptide of the fibrous lamina. It is noteworthy in this respect that this component is clearly enriched in pellets of fibrillar material detached from nuclear membranes by homogenization and centrifugation (Krohne et al. 1978 b).
The maturing amphibian oocyte contains an unusually high number of pore complexes in very close packing. Thus nuclear envelope fractions, extracted with Triton X-10∞ and high salt provide for a remarkable enrichment in nuclear pore complex material (Krohne et al. 1978 ó). Such material is greatly enriched in a polypeptide recognizable as N2 and in a polypeptide at 150000 mol. wt. N1 is apparently absent from such fractions and may be specific to preparations made from liver (Krohne et al. 1978b). Counterparts to the high mol. wt (15 0 000) component detected by the latter authors exist in the iodination pattern of nuclei labelled with lactoperoxidase beads (Fig. 3), and identified as pore complex components, at 160000 and 200000 mol. wt. We believe therefore that our data are closely compatible and complementary to that of Krohne et al. (1978 b).
Recently, a number of workers (Gerace, Blum & Blobel, 1978; Ely, D’Arcy & Jost, 1978; Krohne et al. 1978a) have eluted N1, N2 and N3 of rat liver pore-lamina from SDS polyacrylamide gels and raised antibodies to these polypeptides. Using immunofluorescence localization, they found that antibody to N1, N2 and N3 bound exclusively to the nuclear periphery. Indirect immunoperoxidase staining showed that antibodies to N1, N2 and N3 bound only the fibrous lamina and not to the pore complex (Gerace et al. 1978). From this it was concluded that these polypeptides are not present, or concentrated, in the pore complex in an immunologically reactive form; and it was suggested that N1, N2 and N3 are the major structural components of the fibrous lamina.
The absence of N1 and N2 from the pore complex is contrary to our findings, and, with regard to N2, the report of Krohne et al. (19786) as well. The binding of antibody to a site might indicate the presence of its hapten, but failure to bind does not necessarily exclude its presence. As Gerace et al. point out their antisera were raised to SDS-denatured polypeptides so that the antibodies may be directed towards determinants buried in the proteins and not exposed at the pore complex surface, and all 3 antibodies cross-reacted.
The lactoperoxidase labelling studies have indicated that Ni and N2, both major components of the nuclear envelope, are located in the nuclear pore complex (although not necessarily exclusively so) along with at least 10 other more minor, though more exposed polypeptides. It seems improbable, in view of the regular architecture of the pore complex and the high proportion of polypeptides Ni and N2 in the nuclear envelope, that these are other than skeletal components, whose gross and dynamic organization is dependent on other, quantitatively minor, envelope components. The Coomassie brilliant blue pattern of the pore-lamina fraction reveals approximately 90 bands to the naked eye (rather more than can be seen in Fig. 3) so that there is no shortage of polypeptides whose function might be to organize and control the activity of the pore complex. The radioactive peptides assigned in this paper to the porelamina fraction of the envelope are absent from the Triton-soluble fraction. The Triton-soluble polypeptides are described together with a consideration of their relationship with polypeptides of the endoplasmic reticulum in a later publication.