Académique Documents
Professionnel Documents
Culture Documents
FRACTION NUMBER
FIG.1. Purification of erythropoietin from human urine
concentrate with an immunoadsorbent column containing
monoclonal antibody against erythropoietin. 0, absorbance at
280 nm; 0, erythropoietin activity assayed with in uitro ['Hlthymidine
Mr 1 2 3 4 5
0.4 1
43K-
30K
2 OK-
.
I
20 10 0
bents column with the pH 2.5 buffer (see Fig. 1); lane 3, the second
peak from the Sephadex G-100 column chromatography (see Fig. 3A 1;
lane 4, the first peak; lane 5, the sialidase-treated second peak. The
second peak (0.4 ml) from the Sephadex G-100 column was brought
ACTIVlTY(ulml) to about pH 5.5 by adding 0.4 M acetate, pH 5.2, and incubated with
FIG.3. Sephadex 6-100chromatography of erythropoietin 0.1 unit of sialidase for 5 h at 37 "C.The sialidase was inactivated by
fraction purified withthe immunoadsorbent column. The heating at 100 "C for 5 min.
erythropoietin fraction eluted from the immunoadsorbent column
(Fig. 1) was extensively dialyzed against distilled water and lyophi-
lized. The dried material was dissolved in 2 ml of PBS, pH 6.9, and that themain component in the void fractions was MI35,000
the clear solution was loaded on aSephadex G-100 column (1.2 X 130 protein. Furthermore, it was shown that thiscomponent was
cm) equilibrated with the PBS and the column was developedat 4 "C indeed erythropoietin protein; erythropoietin activity was
with a speed of 6 ml/h. The volume of one fractionwas 2 ml. A shows
the elution pattern of protein based on absorbance a t 280 nm. The found inthe extracts from sliced gels containingthe MI35,000
horizontal lines with arrowheads indicate fractions pooled after elu- component after SDS-polyacrylamide gel electrophoresis of
tion. The pooled fractions were designated the first peak (or the void fraction 36 (not illustrated) and Western blotting revealed
fraction) and the second fraction (or the erythropoietin fraction). B the binding of this component with the monoclonal antibody
shows SDS-polyacrylamide gel electrophoresis of fractions (20 pl of against erythropoietin (Fig. 4, lane 4 ) . Interestingly, the MI
each). M,,M,standards (same as Fig. 2); lane numbers correspond to 20,000protein did not bind withthe antibody. Activity meas-
the fractions of Sephadex G-100 chromatography. C shows erythro-
poietin activity of the extracts from sliced gels. The othergel to which urement of fraction 36, in which there should be no contam-
fraction 48 was applied, was sliced into 1-mm lengths without stain- ination of erythropoietin from the second peak, showed about
ing. The gel pieces were put into test tubes containing 0.5 ml (per '/I6 of the specific activity of the secondpeak,basedon
sliced gel) of PBS, 0.1% bovine serum albumin and ground with a absorbance at 280 nm.More than 50% of the protein of
glass rod. After incubating the suspensions overnight to extract fraction 36,however, appears to be the M,35,000protein (Fig.
protein, they were centrifuged to obtain the clear supernatants. The 3B,lane 36) from the intensity of protein staining. Therefore,
activity in the supernatants was measured with the in uitm [3H]
thymidine incorporation method. it seems that erythropoietin molecules in the void fractions
have much lower activity than those in the second peak (see
"Discussion"). In spite of the Occurrenceof the undefined
silver-staining intensity. The erythropoietin preparation, erythropoietin species on Sephadex G-100 chromatography,
which we purified and used as an immunogen for raising purification of human urinary erythropoietin with an immu-
erythropoietin-directed hybridomas, also contained such noadsorbent and a Sephadex G-100 column provides pure
erythropoietin species (Fig. 4, lane 1 ). Treatment with siali- erythropoietin in a high yield (Table I).
dase converted the main MI35,000 band to one with larger Characterization of the Purified Erythropoietin-Protein
mobility (Fig. 4, lane 5 ) , which suggests that electrophoretic concentrations were determined with a Coomassie brilliant
heterogeneity of erythropoietin is, at least in part, due to its blue binding assay according to themethod of Bradford (10).
variable amount of sialic acid attached to erythropoietin pro- A value of E$&mm for pure erythropoietin was calculated to be
tein. 13.1 when a standard curve was made with ovalbumin and
The main M,20,000contaminant and othersin the eryth- 19.7 withbovine serum albumin. The formervalue was
ropoietin preparation eluted from the immunoadsorbent col- adopted here becauseof the unusual behavior of bovine serum
umn appeared in the void fractions on Sephadex G-100chro- albumin in this assay method. A value of EGnm = 12.6 was
matographv (Fig. 3B, lanes 36-38). It was rather surprising also obtained by the method of Lowry et al. (11)using bovine
2710 Isolation of Human Erythropoietin
serum albumin as standard. The specific activity of purified inaccessible to the antibody, but exposed in the presence of
erythropoietin was determined to be 88,000 units/mg of pro- SDS. It is also possible that the antibody recognizes a struc-
tein with an in vitro [3H]thymidineincorporation method and ture formed by the binding of SDS to erythropoietin. The
81,600 with an in vivo starved rat method (see Table I). Assay binding of erythropoietin to the immunoadsorbent column
with the in vitro 59Fe incorporation method gave a similar can occur only in the presence of SDS but there appears to
value. A specific activity of 70,400 units/mg of protein has be no problems concerning the eluted erythropoietin. The
been reported for erythropoietin isolated from human urine purified erythropoietin no longer binds with the antibody,
with conventional purification procedures by Miyake et al. (2) and has a specific activity comparable to the value for pure
using E;$”,,, = 8.5. erythropoietin (2). The immunoadsorbent column is so exten-
Sialidase treatment (see legend of Fig. 4) of our erythropo- sively washed before the elution of the erythropoietin that, if
ietin completely abolished in vivo activity, while in vitro SDS remains in theeluted erythropoietin,the amount would
activity increased 1.3-fold. Similar results have been reported be very small.
with erythropoietinpartially purified from the plasma of The presence of the erythropoietin molecule with a low
anemic sheep (12, 13). Loss of in vivo activity would becaused specific activity in the void fractions of Sephadex G-100
by the hepatic removal of asialoglycoproteins from the circu- chromatography is puzzling. The void fractions may contain
lation (14). associated forms of erythropoietin and those of the contami-
Thirty amino acids in the NH2-terminalportion of eryth- nant, M, 20,000 protein which has a high affinity for mouse
ropoietin were sequenced with a gas phase sequenator; the IgG (for instance, a component of compliments). Association
sequence was H2N-Ala-Pro-?-Arg-Leu-Ile-Leu-Asp-Ser-Arg- of erythropoietin may result inreducing the activity. It seems,
Val-Leu-Glu-Arg-Tyr-Leu-Leu-Glu-Ala-Lys-Glu-Ala-Glu-?- however, that the binding of the M , 20,000 protein to the
Ile-Thr-Asp-Gly-Gly-Ala. Possible candidates for the uniden- antibody column is mediated by erythropoietin, since the M ,
tified positions are cysteins involved in disulfide bonds or 20,000 protein fractionated on the gel electrophoresis was
amino acids (probably Asn, from the carbohydrates unable to bind with the monoclonal antibody (see Fig. 4, lane
Alerts:
• When this article is cited
• When a correction for this article is posted