Isolation of Gastric Vitamin B,,-binding Proteins Using Affinity Chromatography PURIFICATION AND PROPERTIES OF HOG INTRINSIC FACTOR AND HOG NONINTRIMIC FACTOR*

SUMMARY Two vitamin Brz-binding proteins, designated hog intrinsic factor and hog nonintrinsic factor, have been isolated from hog gastric mucosa. AfEnity chromatography on vitamin Blz-Sepharose resulted in the removal of the bulk of protein present in a crude extract of hog gastric mucosa. The two vitamin Blz-binding proteins were separated subsequently by a method of “selective” affinity chromatography with an affinity adsorbent containing covalently bound derivatives of vitamin Blz that lack the nucleotide portion of the native vitamin. Under appropriate conditions hog nonintrinsic factor was adsorbed to a column of this material, while hog intrinsic factor was not. After additional purification both proteins were isolated in homogeneous form based on polyacrylamide disc gel electrophoresis, sedimentation equilibrium ultracentrifugation, and sodium dodecyl sulfate polyacrylamide gel electrophoresis. Hog intrinsic factor (24 pg) corrected vitamin Blz malabsorption when given to a patient with pernicious anemia, while hog nonintrinsic factor (49 pg) had effect.

AfEnity chromatography on vitamin Blz-Sepharose resulted in the removal of the bulk of protein present in a crude extract of hog gastric mucosa.
The two vitamin Blz-binding proteins were separated subsequently by a method of "selective" affinity chromatography with an affinity adsorbent containing covalently bound derivatives of vitamin Blz that lack the nucleotide portion of the native vitamin.
Under appropriate conditions hog nonintrinsic factor was adsorbed to a column of this material, while hog intrinsic factor was not. After additional purification both proteins were isolated in homogeneous form based on polyacrylamide disc gel electrophoresis, sedimentation equilibrium ultracentrifugation, and sodium dodecyl sulfate polyacrylamide gel electrophoresis.
Hog intrinsic factor (24 pg) corrected vitamin Blz malabsorption when given to a patient with pernicious anemia, while hog nonintrinsic factor (49 pg) had no effect.
Hog intrinsic factor binds 30.3 pg of vitamin Blz per mg of protein.
Molecular weight values of 52,300 to 58,600 were obtained by sedimentation equilibrium ultracentrifugation and amino acid and carbohydrate analysis. The protein contains 17.5 % carbohydrate which accounts for the elevated molecular weight values (66,000 to 75,000) obtained by sodium dodecyl sulfate polyacrylamide gel electrophoresis and gel filtration.
In the presence of vitamin Blz hog intrinsic factor aggregates to form dimers and higher molecular weight oligomers.
Hog nonintrinsic factor binds 25. Hog intrinsic factor and hog nonintrinsic factor differ signiflcantly from each other in their amino acid and carbohydrate composition. These differences, together with differences in other parameters, demonstrate that these two proteins are distinct and separate species.
Crude extracts of hog gastric mucosa have a vitamin B12binding activity of approximately 0.05 pg of vitamin BKJ bound per mg of protein, and these extracts also contain intrinsic factor activity since approximately 60 mg of dry material are able to correct vitamin Blz malabsorption by patients with pernicious anemia as judged by the results of Schilling tests (I). All of the vitamin Bit-binding activity in these crude extracts is not attributable to hog IF,l however, for a second vitamin Br2binding protein is present that lacks IF activity in Schilling tests (1,2). This second vitamin B12-binding protein has been referred to as hog NIP.
Studies using extracts of hog gastric mucosa or partially purified preparations of these two proteins are somewhat difficult to interpret, but they suggest that hog IF and hog NIF also differ from each other in other ways. These differences include: (a) anti-IF antibody obtained from the serum of certain patients with pernicious anemia blocks some of the vitamin Bn-binding activity present in crude extracts of hog gastric mucosa. Irvine (3) has demonstrated that the amount of anti-IF antibodymediated decrease in vitamin Blz-binding activity is a more reliable guide to the Schilling test IF activity of crude hog gastric mucosal preparations than is the level of total vitamin Blz-binding activity.
This observation suggests that the vitamin Blzbinding ability of hog IF is blocked by the antibody, while that of hog NIF is not. (b) Sera from rabbits immunized with hog leukocyte extracts contain antibodies which appear to bind to hog NIF but not to hog IF (4). (c) Hog IF appears to facilitate calcium-dependent vitamin II 12 binding by guinea pig ileal mucosal homogenates, while hog NIF does not (5). (d) Hog NIF appears to facilitate calcium-dependent vitamin Blz binding by rat liver homogenates, while hog IF does not (5). (e) Certain structural analogs of vitamin Blz are bound by hog NIF to a greater extent than they are by hog IF (1).
Numerous attempts have been made to isolate hog IF and hog NIF in homogeneous form (6). Holdsworth (2) and Ellenbogen et al. (1,7) have isolated vitamin B12-binding proteins from hog gastric mucosa in highly purified form and were able to separate the vitamin BIB-binding protein into two fractions using ion exchange chromatography.
In On the other hand, studies of molecular weight, protein aggregation, and amino acid and carbohydrate composition have failed to reveal any major and consistent differences between hog IF and hog NIF. It is thus unclear whether hog NIF is (a) a zymogen-like precursor of hog IF, (b) a limited degradation product of hog IF, or (c) structurally unrelated to hog IF. We have previously reported (8) that the hog gastric mucosal vitamin B12-binding proteins can be isolated in high yield using affinity chromatography on vitamin B&epharose which is prepared by covalently attaching monocarboxylic acid derivatives of vitamin B12 to a substituted Sepharose using a carbodiimide.
Analysis of the vitamin B12-binding protein obtained after affinity chromatography revealed that only 30% of the vitamin Bla-binding protein is IF based on assays using anti-IF antibody obtained from the serum of a pernicious anemia patient. Attempts to separate hog IF from hog NIF using ion exchange chromatography have been unsuccessful. This difficulty has been circumvented, however, and hog IF and hog NIF have been resolved using a method of "selective" affinity chromatography in which carboxylic acid derivatives of vitamin HI2 that lack the nucleotide portion of the vitamin are covalently coupled to Sepharose.
Under appropriate conditions hog NIF is adsorbed to a column of this material, while hog IF is not. This report is concerned with the isolation, separation, and properties of hog IF and hog NIF.
This material is a crude water extract of hog pyloric mucosa that was spray-dried without additional purification.
We have stored this material in the dry state at 4" for up to 1 year without noting any loss of vitamin Bla-binding activity.
All other materials were obtained as described in the accompanying paper in this series (9).

Preparation
of Xepharose-bound Derivatives of V&m& B12 Lacking the Nucleotide-Vitamin Bl2, 10.0 g, containing 2 $Zi of [57Co]vitamin B12, was dissolved in 500 ml of 11 N HCl and incubated at 70" for 1 hour in the dark.
The solution was subsequently lyophilized, dissolved in 500 ml of HQO, and lyophilized again. The dry material was dissolved in water to give a concentration of vitamin Blz derivative of 25.0 mg per ml and stored at -20".
The concentration of vitamin B12 derivatives in the coupling reaction prior to the addition of the carbodiimide was 12.5 mg per ml. The coupling reaction and subsequent washing and storage of the Sepharose substituted with vitamin Bls derivatives were performed under the same conditions used for the preparation of vitamin B&epharose (8). The washed Sepharose containing the 11 N HCI hydrolysis derivatives of vitamin Bj2 had a vitamin Bit derivative content of 0.22 pmole per ml of packed Sepharose based on measurements of radioactivity.
Paper chromatography of vitamin Blz and its derivatives was performed as previously described (8). The amount of individual components was determined by cutting out individual spots and counting them in a Packard y scintillation counter. All other methods were performed as described in the accompanying paper in this series (9).

Purification
of Hog IF and Hog NIF Step 1: Preparation of Hog Gastric Mucosal Extract-All procedures were performed at 4" unless specifically noted.
Hog intrinsic factor concentrate, 50 g, was added to 1.0 liter of 0.1 M Tris-acetate, pH 9.2, and stirred for 30 min. The suspension was centrifuged at 20,000 x g for 30 min, and the turbid supernatant was decanted.
The supernatant was filtered with vacuum suction through Celite using a Buchner funnel containing a coarse scintered glass disc. The filtrate was centrifuged at 20,000 x g for 30 min, and the supernatant was immediately subjected to affinity chromatography on vitamin Blz-Sepharose.
Step 2: Afinity Chromatography on Vitamin B&3epharose-A column (2.5 x 7.5 cm) of vitamin B12-Sepharose containing 12.0 mg of covalently bound vitamin Blz was prepared and washed with 200 ml of 0.1 M glycine-NaOH, pH 10.0, followed by 100 ml of 0.1 M potassium phosphate, pH 7.5, immediately prior to the sample application to remove any hydrolyzed vitamin B12. The flow rate was 100 ml per hour.
After the entire sample had passed onto the column, the column was washed with 25 ml of 0.1 M potassium phosphate, pH 7.5. The first 900 ml of effluent were collected in their entirety.
The column was  After 1 hour an additional 55 ml of eluate were collected, pooled with the first 10 ml, and designated as Eluate 4a. Flow was stopped for an additional 17 hours and then an additional 65 ml of eluate were collected and designated as Eluate 4b.
The starting material, initial column effluent, and each of the column eluates were assayed for vitamin B12-binding activity, IF activity, and protein content.
The results are summarized in Table I. Eluate 4a was dialyzed against 6.0 liters of distilled water for 24 hours with dialysate changes at 4 and 16 hours. The sample was concentrated with an Amicon ultrafiltrator equipped with a Diaflo UM-10 membrane and centrifuged at 20,000 X g for 10 min. The supernatant, 8.5 ml, was decanted and 1.10 ml of 1.0 M sodium acetate, pH 5.0, and 1.47 ml of 7.5 M guanidine HCl were added. The sample was brought to room temperature and immediately subjected to "selective" affinity chromatography. Step 3: "Selective" Afinity Chromalography-A column (2.5 x 18 cm) of the substituted Sepharose containing 0.30 mg per ml of covalently bound derivatives of vitamin B12 that lacked the nucleotide portion of the vitamin was prepared at room temperature. The column was washed with 300 ml of 0.1 M glycine-NaOH, pH 10.0, 120 ml of 7.5 RI guanidine HCl containing 0.1 M sodium acetate, pH 5.0, and finally with 300 ml of 1.0 M guanidine HCl containing 0.1 M sodium acetate, pH 5.0. The sample was applied by gravity at an approximate pressure of 40 cm of H20. The flow rate was 50 ml per hour.
The column was eluted with 200 ml of 1.0 M guanidine HCl containing 0.1 M sodium acetate, pH 5.0, followed by 130 ml of 7.5 M guanidine HCI containing 0.1 M sodium acetate, pH 5.0. Nine-milliliter fractions were collected and assayed for vitamin B12-binding activity, IF activity, and A280. The results are presented in Fig. 1  Appropriate fractions were assayed for: 0, Am; A, Aam; and 0, vitamin B1s-binding ability.
hours with dialysate changes at 4 and 16 hours. The sample was concentrated to 6.0 ml as described in Step 2 and centrifuged at 20,000 x g for 10 min. The supernatant was applied to a column (2.0 x 90 cm) of Sephadex G-150, fine grade, that had been equilibrated with 0.05 M potassium phosphate, pH 7.5, 0.75 M NaCl.
The flow rate was 20 ml per hour, and 4.0-ml fractions were collected.
Appropriate fractions were assayed for vitamin B12-binding activity, A280, and A320. The results are presented in Fig. 2.
Chromatography on S, S' -Diaminodipropylamine -substituted Sepharose-Fractions 27 to 32 from Sephadex G-150 chromatography were pooled and centrifuged at 20,000 x g for 10 min. The supernatant was applied to a column (2.5 x 5.0 cm) of 3,3'diaminodipropylamine-substituted Sepharose that was equilibrated with 0.05 M potassium phosphate, pH 7.5, containing 0.75 M NaCI.
The column was eluted with this same solution. The first 44 ml of effluent contained 84y0 of the vitamin Blzbinding activity applied and were subjected to repeat affinity chromatography on vitamin B:2-Sepharose. Repeat Afinity Chromatography on Vitamin B&'epharose-A column (2.5 x 2.0 cm) of vitamin B&Sepharose containing 3.52 mg of covalently bound vitamin Blz was prepared as described above, and the sample from the preceding step was applied at a flow rate of 100 ml per hour. The column was eluted with: (a) 100 ml of 0.1 M glycine-NaOH, pH 10.0, containing 0.1 M glucose and 1.0 M NaCI; (6) 50 ml of 0.1 M potassium phosphate, pH 7.5; and (c) 28 ml of 7.5 M guanidine HCl containing 0.1 M potassium phosphate, pH 7.5. Eluate 3 contained 308 pg of vitamin BL-binding activity, and only 3.6% of this activity was inhibited with anti-IF antibody. This material is hereafter referred to as hog NIF. Dialysis with Vitamin B&"Co]Vitamin B12, 360 pg, was added to the final preparation of hog IF, and 900 pg were added to the final preparation of hog NIP. Both preparations were dialyzed against 6.0 liters of distilled Hz0 for 72 hours with dialysate changes at 24 and 48 hours. Greater than 99% of 3673 unbound vitamin Blz is removed under these conditions. The proteins were stored at -20".

Purijication
of Hog IF and Hog NIF-Hog IF and hog NIF can be separated from most of the gastric proteins that do not bind vitamin Blz with affinity chromatography on vitamin Bls-Sepharose, as shown in Table I, but this technique does not result in the separation of these two proteins from each other. We have been unable to achieve satisfactory separation of hog IF from hog NIF using ammonium sulfate fractionation, ion exchange chromatography, or gel filtration on Sephadex G-150. The latter technique appeared the most promising but our preliminary studies indicated that considerable recycling would be required to achieve preparations of hog IF and hog NIF that contained less than 5% contamination with each other. We have obtained satisfactory separation of hog IF and hog NIF by the use of a method of selective affinity chromatography in which carboxylic acid derivatives of vitamin Blz that lack the nucleotide portion of the molecule are covalently coupled to 3,3'diaminodipropylamine-substituted Sepharose. Vitamin Blz derivatives that lack the nucleotide portion of the vitamin were prepared by hydrolysis of vitamin B,:, as described under "Methods." Based on paper chromatograms and the observations of Armitage et al. (lo), the lyophilized hydrolysate consisted primarily of mono-(7.9%), di-(17.3$&), tri-(32.2%) and tetra-(42.6%) carboxylic acid derivatives of vitamin Blz that lacked the nucleotide portion of the vitamin. Hydrolysis for 5 min is sufficient to hydrolyze the nucleotide from 99% of the vitamin B12 molecules (lo), but the hydrolysis time was increased to 1 hour to insure essentially comnleto nucleotide hydrolysis and to obtain a high yield of carboxyl groups which are required for coupling the vitamin B1? derivatives to the substituted Sepharose (8).
The mixture of carboxylic acid vitamin Blz derivatives was coupled to 3,3'-diaminodipropylamine-substituted Sepharose with a yield of 1.2% based on the total amount of vitamin Blz derivative present during the coupling reaction.
We have not determined what percentage of the covalently bound vitamin Blz derivatives is represented by the individual mono-, di-, tri-, and tetracarboxylic acid derivatives. It is important that selective affinity chromatography be carried out as described in Step 3 of the purification scheme (see Fig. 1) since deviations in flow rate, pH, temperature, initial guanidine concentration, and column size (i.e. total vitamin Blz derivative content) relative to sample size and content can result in decreased separation of hog IF from hog NIF.
Thus, slower flow rates, neutral PI-I, decreased temperature, lower initial guanidine concentrations, increases in column size relative to the amount of hog IF and hog NIP applied, and substitution of vitamin B&epharose for the selective affinity adsorbent all have the effect of increasing the amount of hog IF that is adsorbed to the column wit.hout affecting the adsorption of hog NIF.
Increases in flow rate and initial guanidine concentration as well as decreases in relative column size have the effect of decreasing the adsorption of both hog IF and hog NIF. The first 7.5 M guanidine eluate (Eluate 4a) from the initial vitamin B&epharose column (see Table I) has a visible yellow color and broad but significant absorption that declines gradually from 300 to 500 nm. This colored material co-chromatographs with hog NIF during selective affinity chromatography since fraction by fraction correspondence with hog NIF could be demonstrated visually and by measurements of AaZO. This colored material was separated from hog NIP by gel filtration on Sephadex G-150 (see Fig. 2) where AazO co-chromatographed with the small Azso peak that was eluted prior to the major AW peak that coincided with the single peak of vitamin Blz-binding activity.
The nature of this yellow material is unknown. A summary of the purification of hog IF and hog NIF is presented in Table II. Hog IF has been purified 920-fold with a yield of 19%. Hog NIF has been purified 320.fold with a yield of 26%.
Based on protein assays performed prior to the addition of vitamin Blz, 1 mg of the final preparation of hog IF contains 30.3 pg of vitamin B,z and has an Azso of 1.44, an AW of 0.89, and a ratio of Azso:A361 of 1.62. One milligram of the final preparation of hog NIF contains 25.1 pg of vitamin BlZ and has an A280 of 1.56, an A361 of 0.74, and a ratio of ADO prior to dialysis to remove guanidine and unbound vitamin B12, both proteins bound more vitamin Ulz than when the two proteins were diluted l:lO,OOO in guanidine-free buffer and assayed directly for vitamin B12-binding activity.
In other experiments, guanidine was first removed by dialysis prior to the addition of excess vitamin Biz and subsequent dialysis to remove unbound vitamin.
In these experiments, hog IF and hog NIF both bound the same amount of vitamin Blz per mg of protein as was observed when excess vitamin was added prior to the removal of guanidine.
This observation indicates that the presence of vitamin Brz is not required to achieve renaturation of these two proteins from guanidine and suggests that other factors such as protein concentration and the rate of guanidine removal may play important roles. Detailed studies concerning the renaturation process have not been conducted for hog IF and hog NIF.

Schilling
Tests-The results of Schilling tests performed on a single pernicious anemia patient are presented in Table III and demonstrate that 24 pg of hog IF were able to restore vitamin Blz absorption to a normal level. Hog NIF, at a dose of 49 pg, was ineffective.
Inhibition of Vitamin B12 Binding by Anti-IF Antibody-The results of experiments performed to determine the ability of anti-IF antibody to inhibit vitamin Bx~ binding by hog IF, hog NIF, and human IF are presented in Fig. 3. At a level of 200 ~1 of antibody, hog IF and human IF were inhibited approximately 98%, while only approximately 3% inhibition was noted with hog NIF. The inhibition curves obtained suggest that this particular antibody has a significantly lower affinity for hog IF than for human IF.
Interaction with Pseudo-vitamin Blz--Samples of hog IF, hog NIF, and human IF in 7.5 M guanidine HCI were diluted 1: 10,000 in 0.1 M potassium phosphate, pH 7.5, and utilized to study the ability of pseudo-vitamin Big to block vitamin Blz binding by these proteins at 4". The results are presented in Table IV  though with a lower affinity than for vitamin B12. The fact that prior incubation with pseudo-vitamin Br2 slows the rate of subsequent [67Co]vitamin Bit binding by human IF to a lesser degree than for hog IF suggests that human IF may bind pseudovitamin B,e with a lower affinity relative to vitamin Blz than does hog IF. No definite conclusion can be reached in regard to this latter point, however, since association and dissociation rates will obviously influence the time course data presented in Table IV.
Polyacrylamide Disc Gel Electrophoresis-The results of the polyacrylamide disc gel electrophoresis experiments are presented in Fig. 4. Single protein bands were observed in the absence of vitamin B12 with 25 pg of hog IF and 25 pg of hog NIF (Gels A and C, Fig. 4). When 25 pg of hog IF saturated with vitamin Blz were studied, a series of protein bands was observed that appeared closer and closer together as one approached the top of the gel (Gel B, Fig. 4  When 25 pg of hog NIP saturated with vitamin Blz were studied, the pattern (Gel D, Fig. 4) was the same as that observed for this protein in the absence of vitamin B12, except that a faint protein band was observed above the single major protein band. It is not clear whether this faint band is due to a slight degree of aggregation by hog NIF or represents trace contamination with hog IF. Gel filtration and sedimentation equilibrium ultracentrifugation studies (see below) also indicate that hog NIF does not aggregate significantly in the presence of vitamin Blz. Jlolecular Weight Determination by Sedimentation Equilibrium-when samples of hog IF and hog NIF were studied by sedimentation equilibrium ultracentrifugation in the absence of vitamin B12, straight lines were observed in both cases when In &so was plotted versus R2 (Fig. 5, A and C). Using their respective partial specific volumes of 0.714 and 0.700, calculated from the amino acid and carbohydrate analyses (see below), molecular weight values of 52,300 and 66,000 were obtained for hog IF and hog NIF.
When a sample of hog NIF saturated with vitamin Brz was studied under similar conditions, straight lines were obtained for plots of In Asso versus R2 and In A382 versus R2 (Fig. 50) Brs under the conditions in which this experiment was performed. When a sample of hog IF saturated with vitamin Blz was studied, plots of In AZ80 versus R2 and In AsG2 versus R2 revealed significant curvature (Fig. 5B). The fact that the degree of curvature is the same for both plots indicates correspondence between protein and vitamin Bi2. When portions of the two curves were used for molecular weight calculations, values ranging from 60,000 to 90,000 were obtained in each case. This range of molecular weight values indicates that the hog IFvitamin RI2 complex existed as a mixture of monomers and higher molecular weight oligomers under the conditions in which this experiment was performed.
Amino Acid and Carbohydrate Composition-The results of the amino acid and carbohydrate analyses are presented in These values are close to the molecular weight values obtained for hog IF (52,300) and hog NIF (63,500) by sedimentation equilibrium ultracentrifugation (see above) and indicate that both proteins contain single vitamin Biz-binding sites. The amino acid and carbohydrate compositions of hog IF and hog NIF are quite similar, but significant differences do exist. Hog IF contains significantly greater amounts of threonine, proline, and methionine than hog NIF, while the latter protein contains significantly more of each of the six carbohydrates that are present in both proteins.
The sulfhydryl group content of hog IF and hog NIF was assayed in 0.1 M potassium phosphate containing 7.5 M guanidine IICl.
No free sulfhydryl groups were detected (<O.l residue per mole), suggesting that all of the cysteine residues present in these proteins are involved in disulfide bonds.
Molecutar Weight Determinations by Gel Filtration-When samples of hog IF and hog NIF devoid of vitamin 1312 were subjected to gel filtration on Sephadex G-150, single peaks of vitamin B12-binding activity were observed for both proteins (Fig. 6, A and C). Based on the elution positions of the two proteins, an apparent molecular weight of 74,000 was calculated for hog IF and a value of 130,000 was calculated for hog NIF.
These values are significantly higher than the molecular weights obtained for these two proteins using sedimentation equilibrium ultracentrifugation and amino acid and carbohydrate analyses (see above).
Protein aggregation appears to be an unlikely explanation for the discrepancies since gel filtration was performed at lower protein concentrations and in the same buffer as was employed for sedimentation equilibrium ultra-Residues per mole of bound vitamin Bu This phenomenon has been reported for other glycoproteins (9,11,12). When a sample of the hog NIF-vitamin B12 complex was studied by gel filtration, the result presented in Fig. 6D was obtained.
Under these conditions, hog NIP eluted as a simple symmetrical peak with an apparent molecular weight of 128,000, which is essentially the same as the value of 130,000 that was obtained when the protein was eluted in the absence of vitamin Brz (Fig. 6C). This observation indicates that hog NIP did not aggregate in the presence of vitamin Brz under the conditions in which this experiment was performed. When the hog IF-vitamin Blz complex was studied (Fig. 6B), two peaks of [57Co]vitamin B12 were observed with apparent molecular weights of 75,000 and 160,000. This indicates that a mixture of hog IF monomers and dimers was present, and that aggregation in the presence of vitamin BLZ had occurred. The 3677 fact that a shoulder is present on the leading edge of the 160,000 molecular weight peak (Fig. 6@ suggests that trimers or tetramers of hog IF may also have been present. The experiments presented in Fig. 6, B and D were repeated under identical conditions except that the protein samples were incubated in 0.5 ml of gel filtration buffer at 37" for 8 hours prior to being cooled to 4" and placed on the Sephadex G-150 column.
In both cases, no change in the elution pattern of [57Co]vitamin Bi2 was observed.

Dodecyl Sulfate
Polyacrylamide Gel Electrophoresis-When 30 pg of hog IF and hog NIF were subjected to sodium dodecyl polyacrylamide gel electrophoresis, a single protein band was observed in each case (Fig. 7). Molecular weight estimates of 66,000 and 100,000 were obtained for hog IF and hog NIF, respectively, based on their mobilities under these conditions.
These values are similar to the values obtained for these proteins by gel filtration and suggest that hog IF and hog NIF both contain single polypeptide chains.
The molecular weight values obtained by sodium dodecyl sulfate polyacrylamide gel electrophoresis appear to be falsely elevated as has been reported for other glycoproteins studied by this technique (9,(12)(13)(14).
Equilibrium Dialysis-The results of equilibrium dialysis experiments are presented in Fig. 8. A value of 1.5 X 10'" ~-1 was obtained for the association constant, Ka, for hog IF and vitamin Biz. A value of 1.3 X 1Or0 M-' was obtained for hog NIF and vitamin 131~. These two values are not significantly different and are similar to the value of 1.5 X 10"' M-' obtained for human IF and vitamin Bi2 under t.he same conditions (9).
Absorption Spectra-The spectra of equal concentrations of the hog IF-vitamin Biz complex, the hog NIF-vitamin Br2 complex, and unbound vitamin Blz are presented in Fig. 9. When vitamin Biz is bound to either of these proteins, the 361 nm spectral maximum for unbound vitamin El2 shifts to 362 nm.
When vitamin 7312 binds to either hog IF or hog NIF, there is general enhancement of the vitamin BiB spectrum above 320 nm since the spectra of the two proteins devoid of vitamin Blz are those of typical proteins with insufficient absorption above 320 nm to account for the differences between the protein-bound vitamin Brz spectra and that of unbound vitamin Uiz (see Fig. 9) in the region above 320 nm. This observation is supported by experiments in which subsaturating aliquots of vitamin Biz were added to buffer containing hog IF and hog NIP and the increase in absorption at 361 nm was noted to be approximately 30% greater than when aliquots of vitamin Br2 were added to buffer alone. DISCUSSION Hog IF and hog NIP have been isolated from a crude extract of hog pyloric mucosa. Affinity chromatography on vitamin B&epharose served to remove most proteins that do not bind vitamin B;z, but extreme difficulty was encountered in attempts to separate the two vitamin Bi2-binding proteins from each other. Holdsworth (2) and Ellenbogen et al. (I, 7) have succeeded in at least partially separating these two proteins by the use of ion exchange chromatography, but we and other investigators (4) have not had success with this technique.
The reasons for this discrepancy are not clear, but they may involve the fact that both Holdsworth and Ellenbogen et al. employed incubation with proteolytic enzymes at early stages in their purification schemes since it is possible that limited proteolysis changes the chromatographic behavior of hog IF and hog NIF. probably attributable to the fact that previous preparations of hog IF (1, 2, 7) appear to have contained significant contamination with other proteins. This suggestion is supported by the fact that previous preparations of hog IF had lower specific activities than ours.
Under certain conditions, hog IF appears to exist as a mixture of monomers, dimers, and other higher molecular weight oligomers. This phenomenon has been observed in studies employing polyacrylamide disc gel electrophoresis, sedimentation equilibrium ultracentrifugation, and gel filtration on Sephadex G-150. The presence of vitamin Brz appears to be a requirement for oligomer formation. Similar studies have failed to provide evidence for significant oligomer formation by hog NIF. This observation appears to conflict with the observations of Ellenbogen et al. (1,7) since they obtained ultracentrifugational evidence that both hog IF and hog NIF were capable of forming oligomers in the presence of vitamin Brz. We do not have any definite explanation for this discrepancy, but it is possible that hog NIF is capable of forming oligomers only with different salt or protein concentrations than we have employed. Ellenbogen et al. employed incubation with trypsin and chymotrypsin as part of their purification scheme and it is also possible that limited proteolysis of hog NIF must occur before this protein is capable of oligomer formation.  demonstrable free sulfhydryl groups; (e) when vitamin Blz binds to either protein, there is general enhancement of the vitamin BE? spectrum above 320 nm; the spectra of the two protein-vitamin Brz complexes are indistinguishable in this region; cf) the 361 nm spectral maximum for unbound vitamin 1312 shifts to 362 nm when the vitamin is bound to either protein; (g) both proteins have molecular weights close to 60,000, although the molecular weight of hog IF is slightly less than that of hog NIP; (h) hog IF and hog NIF are both glycoproteins and contain the same kind of carbohydrate residues. Differences between hog IF and hog NIF include: (a) the vitamin Rlz-binding ability of hog IF is inhibited by antibody obtained from a patient with pernicious anemia, while that of hog NIF is not. (b) Hog IF has a lower affinity for pseudo-vitamin Blz relative to vitamin Bj2 than does hog NIF.
(c) Hog IF is able to correct vitamin ISI2 malabsorption in a patient with pernicious anemia, while hog NIF cannot.
(d) Hog IF facilitates vitamin Blz binding to homogenates of human and guinea pig distal small intestine, while hog NIF does not.2 (e) Hog IF and hog KIF have markedly different apparent molecular weights when determined by gel filtration.
Both values are falsely elevated but the degree of false elevation is greater for hog NIF. (f) The two proteins also have significantly different apparent molecular weights when determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis. The degree of false elevation in these values is also greater for hog NIF.
(g) Hog IF aggregates in the presence of vitamin Brz under conditions in which hog NIF does not. (h) The hog IF-vitamin Br2 complex contains less absorption from 250 to 290 nm than the hog NIF-vitamin Blz complex.
(i) Hog IF contains significantly greater amounts of threonine, proline, and methionine than hog NIF.
The differences observed between hog IF and hog NIF demonstrate that these two proteins are distinct entities.
The fact that hog NIP has a slightly larger molecular weight than hog IF and contains approximately twice as much carbohydrate rules out the possibility that hog IC'IF is a limited degradation product of hog IF. The fact that hog IF contains significantly greater amounts of threonine, proline, and methionine rules out the possibility that hog NIF is a zymogen-like precursor of hog IF. It is conceivable that both proteins are derived from a common, larger molecular weight glycoprotein, but we are not aware of any evidence to suggest such a possibility.
The function of hog NIF is unknown and our studies have not provided any clues concerning this question.
Aro and Grasbeck (4) have demonstrated that hog NIF reacts with antibodies prepared against crude hog granulocyte extracts and have suggested that hog NIF is a R type vitamin Br2-binding protein since members of this group of proteins have immunologic similarities and are present in granulocytes and a variety of body fluids. It is of interest in this regard that hog NIF does have many properties in common with the human granulocyte vitamin Blz-binding protein (12). Both proteins have similar amino acid and carbohydrate compositions and similar molecular weights; both proteins also give rise to the same degree of falsely high estimates of molecular weight when values are determined by gel filtration and sodium dodecyl sulfate polyacrylamide gel electrophoresis. The assignment of hog NIF as a member of the R type group of vitamin Blz-binding proteins does not, however, provide any information concerning the function of hog NIF, since specific functions have not been determined for any of the R type proteins (16). Hog IF and human IF (9) have many common properties, but the two proteins do differ in amino acid and carbohydrate composition, molecular weight, and in their interactions with anti-IF antibody and pseudo-vitamin Bls. The physiologic significance of these differences is unclear and they may reflect merely a species difference.
It is important to note, however, that hog IF was isolated from gastric mucosa, while human IF was isolated from gastric juice.
This difference could be important if IF is altered by proteolytic enzymes, or other factors, prior to, during, or after it is secreted into a gastric juice.