AU4366796A - Stable n-terminally linked dtpa:protein compositions and methods - Google Patents

Description

STABLE N-TERMINALLY LINKED DTPA:PROTEIN COMPOSITIONS AND

METHODS

FIELD OF THE INVENTION

The present invention broadly relates to the field of protein modification, and, more specifically, to diethlyenetriaminepentaacetic acid (DTPA) -protein compositions wherein the chelating agent DTPA has been conjugated site-specifically to the N-terminus of the protein, thereby providing a homogenous and well-defined product capable of forming complexes with a variety of metallic radionuclides. In another aspect, the invention relates methods of conjugating DTPA to granulocyte colony stimulating factor (G-CSF) or interleukin-2 (IL-2), thereby providing a useful procedure of radio-labeling such proteins and related proteins including cytokines, while maintaining the structural and functional integrity of the protein.

BACKGROUND OF THE INVENTION

Radioactive labeling of proteins and other biological compounds is commonly achieved by iodination. Proteins may be successfully labeled with radioisotopes of iodine by a number of methods; Reogoeczi, E., Iodine- Labeled Plasma Proteins, 1, 53, (CRC Press, Boca Raton, Fla 1982), and antibodies so labeled have been used in radioi munodetection studies in which tumor localization is determined by external imaging; Keenan et al. , J^".

Nucl . Med . , 2j_, 531 (1985) . However, in the course of these investigations and others involving the use of radioisotopes of iodine, certain limitations to the use of radioiodine imaging procedures became apparent (e.g., the poor imaging characteristics of many of the radioisotopes of iodine, the involved labeling procedures, and the high degree of instability of the label in vivo) . In addition, the most common iodination methods involve oxidizing conditions in the reaction mixtures which can modify other sensitive groups and cause alteration of protein structure, and possible biological inactivation.

To avoid the difficulties encountered when trying to iodinate these proteins and other compounds, alternative methods were employed. One such method is the "bifunctional chelate" method, in which strong chelating groups are covalently attached to proteins so that the protein bound chelate can then form complexes with a variety of metallic radionuclides; Meares & Goodwin, J. Prot . Chem. , 2, 215-228, (1984), paramagnetic metal ions; Lauffer & Brady, Magn . Reson. Imag. , 3_, 11-16, (1985); Ogan et al. , Invest. Radiol . , 22., 665-671, (1987), and flourescent metals; Mukkala et al., Anal . Bioche . , 12 ., 319-325, (1989) .

The reagent most commonly used for the covalent modification of proteins with a chelating agent is the cyclic dianhydride of DTPA. The cyclic dianhydride of DTPA generally forms stronger chelates than the analogs of ethylenetriaminetetraacetic acid (EDTA) ; Perrin et al . , Organic Ligands, IUPAC Chemi cal Data Series No. 22 , (New York, Pergamon Press 1982), and involves less complicated synthesis procedures than those involved when using analogs of EDTA. In addition, the cyclic dianhydride of DTPA is stable indefinitely at room temperature, thereby providing for greater control on the conditions of coupling. Hnatowich and McGann, Int . J. Rad. Appl . Ins rum . , [B] , IA, 563-568 (1987) . Coupling of DTPA to proteins is routinely performed at pH >_.7.0, where the dianhydride reacts primarily with free amine groups (e.g., available lysine residues) to form amide bonds; Hnatowich et al. , Science, ____.. 613- 615, (1983a). - 3 -

A primary concern for one performing these covalent modifications is that there can be many possible sites on each protein where chelators can be attached. The currently existing methods provide for non-selective attachment at any reactive group, whether located within the protein, such as a lysine side group, or at the N-terminus. This results in a heterogenous population. For example, reaction of DTPA dianhydride with insulin yielded a complex mixture of several products, including cross-linked protein and acylated tyrosine residues; Maisano et al. , Bioconj . Chem. , 3_, 212-217 (1992), while reaction of albumin with DTPA dianhydride produced protein molecules with multiple chelating groups attached; Lauffer _ Brady, Magn . Reson . Imag. , 2, 11-16, (1985) .

The number of DTPA groups conjugated to the protein is often given as an average number, as sample preparations are heterogenous, each having protein with both more and less chelating groups than the average number; Hnatowich and McGann, Int. J. Rad. Appl .

Ins t rum. , [B] , ϋ, 563-568 (1987) . It is well known that proteins may be degraded by covalent attachment of chelating groups, with the extent of degradation increasing with increasing substitution; Sakahara et al., J. Nucl . Med. , 23., 750, (1985). Those protein molecules containing several chelating groups are least likely to retain their native biological properties; Meares and Goodwin, Jour, of Prot . Chem. , 2, 215-228 (1984) . From a producer's point of view, garnering regulatory approval for sale of these heterogenous therapeutic proteins may have added complexities.

The properties in vivo of chelate-tagged proteins have been reviewed; Meares et al . , Adv. Chem. , 198. 369-387 (1982) . The most general observations are that in vivo stability depends critically on the chemical nature of the protein-chelator conjugate, that the more lightly labeled proteins have longer biological half-lives, and that retention of activity is made more likely bv procedures (e.g. specific labeling) that minimize the labeling of residues involved in the active site(s) . For example, horse serum albumin (HSA) was conjugated with chelating agent, labeled with ¹¹¹In, and when injected in vivo was rapidly cleared by the liver (as compared to results following ¹²⁵I label on iodinated HSA); Leung and Meares, Biochem. Biophys . Res . Co mun . , 25., 149-155 (1977) . The chelate-conjugated HSA, at least the population with the most numerous chelating groups and thus representing a large percentage of the followed radioactivity, may have been recognized in vivo as foreign protein; Meares and Goodwin, Jour, of Prot . Chem. , 2, 215-228 (1984) . The advantage of avoiding a random and numerous distribution of products by specifically labeling a single (nonessential) site on a protein is evident.

Covalent coupling of DTPA to proteins using DTPA dianhydride has been described by several investigators. For example, Khaw et al . , Science, 209. 295, (1980) coupled DTPA to i munoglobulin G (IgG) fragments active against myosin and investigated the localization of the labeled protein in canine myocardial infarcts. Using the same method, Scheinberg et al. ,

Science, 215. 1511, (1982) prepared labeled monoclonal antibody specific for erythroleukemic cells in mice. Although these methods and others provide coupled proteins, they are invariably characterized by complicated syntheses and by low coupling efficiencies. Hnatowich et al . , Science, 220., 613-615, (1983a) .

U.S. Patent No. 4,479,930 (Hnatowich) discloses compositions comprising a dicyclic dianhydride coupled to an amine, and chelated with a radioisotope metallic cation. The compositions are reported to be stable in vivo. Methods of preparing the compositions are also disclosed. It is reported that the initial and final pH of the coupling reaction mixture is pH 7.0 in all instances, and that coupling efficiency (defined as the percentage of anhydride molecules which covalently attach to the polypeptide or protein) is high when anhydride to antibody molar ratios are held at 1:1, but decrease at pH values above or below neutrality. There is no teaching as to the distribution of the DTPA moiety on the proteins or polypeptides of the various reaction products.

Nothing can be drawn from the literature concerning the preparation of DTPA:protein conjugates which are advantageous over those previously described due to the fact that the conjugation is site-specific to the N-terminus of the protein, thereby yielding a more well-defined, homogenous composition. The compositions can be produced in large quantities and retain full in vivo bioactivity, either with or without chelated metallic radionuclide. The synthesis described in the present invention is a simple one step reaction wherein a single reactive site is created, providing a useful method for labeling proteins. The DTPA:protein conjugates of the present invention may have potential use in diagnosis, imaging, and/or treatment of leukemia and related diseases.

SUMMARY OF THE INVENTION

The present invention relates to substantially homogenous preparations of N-terminally chemically modified proteins, and methods therefor. Unexpectedly, the conjugation of the chelating agent DTPA is site- specific to the N-terminus of the protein, thereby providing a more homogenous and well-defined product as compared to other chelating agent:protein compositions. Also unexpectedly, the preferred DTPA:G-CSF conjugates - 6 -

can chelate a number of metallic radionuclides yielding a radio-labeled product, while maintaining the structural and functional integrity of the protein, and this method is broadly applicable to other proteins (or analogs thereof) , as well as G-CSF.

In one aspect, the present invention relates to a substantially homogenous preparation of DTPA:G-CSF (or analog thereof) and related methods. One working example below demonstrates that the chelating agent DTPA is conjugated site-specifically to the N-terminus of rhG-CSF, and that such compostion is capable of forming complexes with a variety of metallic radionuclides. Since the conjugation is specific to the N-terminus of the G-CSF molecule, the resulting product is a more homgenous and well-defined product than those previously described.

The present invention also relates to a method for preparing a labeled protein, said method comprising: (a) reacting a chelating agent with said protein at a pH sufficiently acidic to selectively activate the α- amino group at the amino terminus of said protein; (b) separating the conjugated protein from non-conjugated protein; (c) adding a metallic cation to said conjugate; and (d) obtaining the labeled protein. This method is described below for rhG-CSF and IL-2, and these provide for additional aspects of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGURE 1 shows the effects of initial pH and DTPA:protein molar ratio on the coupling of rhG-CSF. SDS-PAGE analysis of the following samples was performed: lane 1- MW markers; lanes 2-4, DTPA:rhG-CSF at 5:1, 50:1, and 500:1, respectively, at initial pH 6.0; lanes 5-7, DTPA:rhG-CSF at 5:1, 50:1, and 500:1, respectively, at initial pH 7.0; lanes 8-10, DTPA:rhG- CSF at 5:1, 50:1, and 500:1, respectively, at initial pH 8.0; lane 11, rhG-CSF at pH 6.0; lane 12, rhG-CSF at pH 8.0.

FIGURE 2 shows size-exclusion HPLC elution plots for rhG-CSF starting material (line 1), DTPA:rhG- CSF conjugation reaction mixture before passage through a G50 spin column (line 2), and DTPA:rhG-CSF conjugation reaction mixture after passage through a G50 spin column (line 3) . Elution was monitored for absorbance at 280 n .

FIGURE 3 shows preparative cation-exchange FPLC elution plots for rhG-CSF starting material (dashed line) and DTPA:rhG-CSF reaction mixture (solid line) . Elution was monitored for absorbance at 280 nm.

FIGURE 4 shows analytical cation-exchange HPLC analysis of rhG-CSF (lines 2 and 3) and DTPA:rhG-CSF conjugate (lines 1 and 4) preincubated with ^luIn. Elution was monitored for absorbance at 220 nm (lines 3 and 4) and for radioactivity (lines 1 and 2, inverted) .

FIGURE 5 shows analytical cation-exchange HPLC analysis of rhG-CSF (line 1), DTPA:rhG-CSF conjugate (line 2), and DTPA:rhG-CSF conjugate treated with excess InCl₃ (line 3) . Elution was monitored for absorbance at 220 nm. EDTA (ImM) was added to Buffer A.

FIGURE 6 shows silica gel TLC plate analysis of the following samples: lane 1, 0.1 nmol ^nιIn (indium with a trace of ^luIn) ; lane 2, 10 nmol ^ιnIn added to 20 nmol DTPA; lane 3, 10 nmol ^U1ln incubated with 2 nmol rhG-CSF, followed by addition of 20 nmol DTPA; and lane 4, 10 nmol In incubated with 2 nmol DTPA:rhG-CSF conjugate, followed by addition of 20 nmol DTPA. For lanes 2-4, aliquots containing 0.1 nmol ^ulIn were taken from the mixtures and loaded onto the plate.

FIGURE 7 shows the MALDI-MS spectrum of the

DTPA:rhG-CSF conjugate. The spectrum shows the multiply protonated species (1, 2, 3 and 4 protons attached) .

FIGURE 8 shows the ion-spray mass spectrum of the DTPA:rhG-CSF conjugate with chelated indium. The conjugate was preincubated with saturating InCl₃ (In:conjugate, 10:1, mol/mol) before analysis.

FIGURE 9 shows the ion-spray mass spectrum of rhG-CSF.

FIGURE 10 shows peptide mapping of the DTPA:rhG-CSF conjugate. Peptide fragments generated from the DTPA-rhG-CSF conjugate (solid line) and rhG-CSF (dashed line) by proteolysis were reduced, alkylated, and then resolved by reversed-phase HPLC. Elution was monitored for absorbance at 215 nm. The arrow indicates elution of the N-terminal peptide from the digested rhG- CSF sample.

FIGURE 11 shows an isoelectric focusing gel (pH 3-10) containing the following samples: lane 1, rhG- CSF; lane 2, DTPA:rhG-CSF conjugate preincubated with excess InCl₃ (In:conjugate, 10:1, mol/mol); lane 3, DTPA:rhG-CSF conjugate; and lane 4, isoelectric point markers.

FIGURE 12 shows circular dichroism (CD) spectra of the DTPA:rhG-CSF conjugate without { ) and with ( ) chelated indium, and of unmodified rhG-CSF

( ) and DTPA (_._.) . Samples (0.078 mg/ml protein, and 4.07 μM DTPA) were analyzed at 10°C in 20mM sodium acetate, pH 5.4. The unmodified rhG-CSF sample and DTPA:rhG-CSF sample (without indium) are identical.

FIGURE 13 shows the effects of DTPA conjugation on the in vivo activity of rhG-CSF. Activity (WBC count) was measured after subcutaneous injection of hamsters. The rhG-CSF dose was 100 μg/kg. Bars represent standard deviation (n = 8-10 for protein samples, and n = 5-6 for baseline) .

FIGURE 14 shows analytical cation-exchange HPLC analysis of IL-2 (lines 2 and 3) and DTPA:IL-2 conjugate (lines 1 and 4) preincubated with ^luIn. Elution was monitored for absorbance at 220 nm (lines 3 and 4) and for radioactivity (lines 1 and 2, inverted) .

FIGURE 15 shows analytical cation-exchange HPLC analysis of IL-2 (line 1), DTPA:IL-2 conjugate (line 2), and DTPA:IL-2 conjugate treated with excess

InCl₃ (line 3) . Elution was monitored for absorbance at 220 nm. EDTA (lmM) was added to Buffer A.

FIGURE 16 shows silica gel TLC plate analysis of the following samples: lane 1, 0.1 nmol ^nιIn (indixαm with a trace of ^ιIn) ; lane 2, 10 nmol ^nιIn added to 20 nmol DTPA; lane 3, 10 nmol In incubated with 2 nmol IL-2, followed by addition of 20 nmol DTPA; and lane 4, 10 nmol ^U1ln incubated with 2 nmol DTPA:IL-2 conjugate, followed by addition of 20 nmol DTPA. For lanes 2-4, aliquots containing 0.1 nmol ^ulIn were taken from the mixtures and loaded onto the plate.

FIGURE 17 shows peptide mapping of the DTPA:IL-2 conjugate. Peptide fragments generated from the DTPA-IL-2 conjugate (solid line) and IL-2 (dashed line) by proteolysis were reduced, alkylated, and then resolved by reversed-phase HPLC. Elution was monitored for absorbance at 215 nm. The arrow indicates elution of the N-terminal peptide from the digested IL-2 sample.

DETAILED DESCRIPTION

The DTPA:protein conjugates of the present invention are described in more detail in the discussion that follows and are illustrated by the examples provided below. The examples show various aspects of the invention and include results of biological activity testing of various DTPA:protein conjugates. Surprisingly, using the methods of the present invention, a single reactive site was created such that the resulting DTPA conjugation is site-specific to the N-terminus of the protein, yielding a well-defined and homogeneous composition capable of forming complexes with a variety of metallic radionuclides, while maintaining the structural and functional integrity of the protein.

Contemplated for use in the practice of the present invention are a variety of cytokines and related proteins. Exemplary proteins contemplated include various hematopoietic factors such as the aforementioned G-CSF, GM-CSF, M-CSF, the interferons (alpha, beta, and gamma), the interleukins (1-14), erythropoietin (EPO) , fibroblast growth factor, stem cell factor (SCF) , megakaryocyte growth and development factor (MGDF) , platelet-derived growth factor (PDGF) , and tumor growth factor (alpha, beta) .

Granulocyte colony stimulating factor (G-CSF) is a glycoprotein which induces differentiation of hemapoietic precursor cells to neutrophils, and stimulates the activity of mature neutrophils.

Recombinant human G-CSF (rhG-CSF), expressed in E. coli , contains 175 amino acids, has a molecular weight of 18,798 Da, and is biologically active. Currently, Filgrastim, a recombinant G-CSF, is available for therapeutic use. The structure of G-CSF under various conditions has been extensively studied; Lu et al. , J. Biol . Chem. Vol. 231, 8770-8777 (1992), and the three-dimensional structure of rhG-CSF has recently been determined by x-ray crystallography. G-CSF is a member of a class of growth factors sharing a common structural motif of a four α-helix bundle with two long crossover connections; Hill et al . , P.N.A. S. USA, Vol. H, 5167-5171 (1993) . This family includes GM-CSF, growth hormone, interleukin-2, interleukin-4, and interferon β. The extent of secondary structure is sensitive to the solvent pH, where the protein acquires an even higher degree of alpha helical content at acidic pH; Lu et al., Arch. Biochem. Biophys . , 286. 81-92 (1989) . In general, G-CSF useful in the practice of this invention may be a form isolated from mammalian organisms or, alternatively, a product of chemical synthetic procedures or of prokaryotic or eukaryotic host expression of exogenous DNA sequences obtained by genomic or cDNA cloning or by DNA synthesis. Suitable prokaryotic hosts include various bacteria (e.g., E. coli); suitable eukaryotic hosts include yeast (e.g., S. cerevisiae) and mammalian cells (e.g., Chinese hamster ovary cells, monkey cells) . Depending upon the host employed, the G-CSF expression product may be glycosylated with mammalian or other eukaryotic carbohydrates, or it may be non-glycosylated. The G-CSF expression product may also include an initial methionine amino acid residue (at position -1) . The present invention contemplates the use of any and all such forms of G-CSF, although recombinant G-CSF, especially E. coli derived, is preferred, for, among other things, greatest commercial practicality.

Certain G-CSF analogs have been reported to be biologically functional, and these may also be chemically modified. G-CSF analogs are reported in U.S. Patent No. 4,810,643. Examples of other G-CSF analogs which have been reported to have biological activity are those set forth in AU-A-76380/91, EP 0 459 630, EP 0 272 703, EP 0 473 268 and EP 0 335 423, although no representation is made with regard to the activity of each analog reportedly disclosed. See also AU-A-10948/92, PCT US94/00913 and EP 0 243 153.

Generally, the G-CSFs and analogs thereof useful in the present invention may be ascertained by practicing the chemical modification procedures as provided herein and testing the resultant product for the desired biological characteristic, such as the biological activity assays provided herein. Of course, if one so desires when treating non-human mammals, one may use recombinant non-human G-CSF' s, such as recombinant murine, bovine, canine, etc. See PCT WO 9105798 and PCT WO 8910932, for example.

Interleukin-2, a glycoprotein with a molecular weight of approximately 15,000 daltons, is a member of the group called lymphokines that mediate immune responses in the body. This protein is produced by activated T-cells and is known to possess various activities in vivo . For instance, IL-2 has been reported to enhance thy ocyte mitogenesis, induce T-cell reactivity, regulate gamma interferon, and augment the recovery of the immune function of lymphocytes in selected immunodeficient states. It has potential application in research and the treatment of neoplastic and immunodeficiency diseases and has been employed in therapies for the treatment of cancer. IL-2 useful in the practice of this invention may be a form isolated from mammalian organisms or, alternatively, and especially if an IL-2 analog, a product of chemical synthetic procedures or of prokaryotic or eukaryotic host expression of exogenous DNA sequences obtained by genomic or cDNA cloning or by DNA synthesis. Suitable prokaryotic hosts include various bacteria (e.g., E. coli ) ; suitable eukaryotic hosts include yeast (e.g., S. cerevisiae) and mammalian cells (e.g., Chinese hamster ovary cells, monkey cells) . Depending upon the host employed, the IL-2 expression product may be glycosylated with mammalian or other eukaryotic carbohydrates, or it may be non-glycosylated. The IL-2 expression product may also include an initial methionine amino acid residue (at position -1). The present invention contemplates the use of any and all such forms of IL-2 and its analogs, although recombinant IL-2 and analogs, especially E. coli derived, are preferred, for, among other things, greatest commercial practicality.

Methods for the preparation of IL-2 by isolation and purification of the naturally occurring form or by genetically engineered means are known. See, for instance, U.S. Patent Nos. 4,778,879; 4,908,434; and 4,925,919 (Mertelsman et al. ) ; U.S. Patent No. 4,490,289 (Stein); U.S. Patent No. 4,738,927 (Taniguchi et al. ) ; U.S. Patent No. 4,569,790 (Koths et al. ) ; U.S. Patent No. 4,518,584 (Mark et al . ) ; U.S. Patent No. 4,902,502 (Nitecki et al. ) and European patent 0136489 (Souza et al. ) .

The IL-2 receptor (IL-2R) is constitutively overexpressed in various hematologic malignancies including adult T-cell leukemia (Uchiyama, et al. , 1985), hairy cell leukemia (Trentin, et al . , 1992), chronic lymphocyte leukemia (Rosolen, et al . , 1989),

Hodgkin's disease (Strauchen and Breakstone, 1987), and non-Hodgkin' s lymphoma (Grant, et al. , 1986). Lymphocytes involved in several autoimmune diseases, including rheumatoid arthritis (Lemm and Warnatz, 1986) and allograft rejection (Waldmann, 1989) also overexpress the IL-2R. This receptor has therefore been actively pursued as a target for cytotoxic therapy. Recombinant fusion toxins have been produced in which the cell-binding domain of Pseudomonas exotoxin (Lorberboum-Galski, et al. , 1988a) or the receptor- binding domain of diphtheria toxin (Williams, et al. , 1987) have been replaced with IL-2. These fusion proteins are specifically cytotoxic to cells that express the high affinity IL-2R (Lorberboum-Galski, et al., 1988b; Williams, et al. , 1990) . A recently described Pseudomonas exotoxin/IL-4 chimeric protein may also prove useful for the treatment of autoimmune diseases, allograft rejections, and many hematologic malignancies where cells express elevated levels of IL-4 receptor (Puri, et al., 1994) . A diphtheria toxin- related human G-CSF fusion protein has also recently been constructed which may have usefulness in the study and treatment of leukemia (Chadwick, et al., 1993) . Conjugation of DTPA to IL-2, IL-4, as well as rhG-CSF may also have potential use in diagnosis, imaging and/or treatment of leukemia and related diseases. Chelatable radiometals for cytotoxic therapy may include ²¹²Bi, ²¹¹At, and ⁹⁰Y. Indeed, antibodies to the IL-2R a chain, and bearing radioisotopes ²¹²Bi and ⁹⁰Y via conjugated bifunctional chelates, are being examined by several investigators (Junghans, et al. , 1993; Parenteau, et al., 1992; Kozak, et al . , 1990) for cytotoxicity towards alloreactive T-cell lines, and for potential radiotherapy.

The DTPA useful in the conjugations of the present invention is technical grade DTPA dianhydride. In a preferred embodiment involving E. coli derived rhG-CSF, the DTPA:rhG-CSF conjugation occurs at an initial pH of 6.0, and a 50:1 DTPA:rhG-CSF molar ratio. In a preferred embodiment involving E. coli derived IL-2, the DTPA:IL-2 conjugation occurs at an initial pH of 6.0, and a 50:1 DTPA:IL-2 molar ratio.

Although the invention has been described and illustrated with respect to specific DTPA:protein conjugates and treatment methods, it will be apparent to one of ordinary skill that a variety of related conjugates, and treatment methods may exist without departing from the scope of the invention.

The following examples will illustrate in more detail the various aspects of the present invention.

EXAMPLE 1

The DTPA used for conjugation is initially the dianhydride form, and therefore there exists the potential for undesirable side-reactions such as protein:protein crosslinking; Hnatowich et al.,J^". I-πmuπo. Methods , ££, 147-157, (1983b). Reaction conditions such as initial pH and DTPA dianhydride:rhG- CSF molar ratio were therefore investigated in order to minimize the formation of such products. The rhG-CSF was produced using recombinant DNA technology in which E. coli cells were transfected with a DNA sequence encoding human G-CSF as described in U.S. Patent No.

4,810,643 to Souza. The rhG-CSF was prepared as a 2.75- 4 mg/ml solution in 100 mM sodium phosphate buffer, pH 6.0. DTPA dianhydride and tributylphosphine (TBP, technical grade) were obtained from Aldrich (Milwaukee, WI) . Preparation of DTPA:rhG-CSF Conjugate

Various amounts of DTPA dianhydride were placed in dry acid-washed (Meares, et al. , J^". Prot. Chem. , 2 215-228, (1984)) test tubes. Two milliliters of anhydrous chloroform (Aldrich Chemicals, Milwaukee, WI) was then added, and the tube vortexed under a light stream of nitrogen gas to evaporate the chloroform and form a thin film of DTPA dianhydride at the bottom of the tube. rhG-CSF at a concentration of 2.75-4.0 mg/ml in lOOmM sodium phosphate buffer, pH 6.0, pH 7.0, or pH 8.0 was added to the DTPA dianhydride-coated tubes to a final molar ratio of 5:1, 50:1, or 500:1 (DTPA:rhG-CSF) while gently swirling. Aliquots of each sample were maintained and the bulk of the sample passed through a G50 spin column as described; Penefske, H.S., Methods Enzymol . , £__, 527-530, (1979), in order to remove unconjugated DTPA.

Anaivsis of the DTPA:rhG-CSF Con-iuαate

1. SDS-PAGE Analysis.

SDS-PAGE was performed on the G50 spin column treated samples using 17-27% ISS MiniPlus gels (Nattick, MA) . Samples were diluted with nonreducing buffer, and 5μg of protein was loaded into each well. The gels were run on a discontinuous buffer system and stained with Coomassie Blue R-250 (Laemmli, U.K., Nature, 227. 680- 685, (1970) . SDS-PAGE analysis of the samples prepared above is shown in Figure 1.

The reactions conducted with an initial pH of

7.0 (lanes 5-7) or 8.0 (lanes 8-10) yielded significant amounts of higher molecular weight species in comparison to the apparent molecular weight observed for untreated rhG-CSF (lanes 11 and 12) . However, the reactions with an initial pH of 6.0 (lanes 2-4) yielded a single detectable higher molecular weight species (which corresponded to one of the species detected in the pH 7.0 and 8.0 samples). The apparent molecular weight of this species suggests it may be a DTPA-cross-linked protein dimer. With the pH 6.0 samples, this higher molecular weight band became less intense with increasing DTPA:rhG-CSF molar ratio, and was virtually absent for the 500:1 sample (lane 4) .

2. Size Exclusion HPLC.

The reaction mixture with an initial pH of 6.0, and DTPA:rhG-CSF ratio of 50:1 was further analyzed by size-exclusion HPLC. A pre-G50 spin column and post- G50 spin column sample was analyzed. HPLC was performed on a Waters Liquid Chromatograph (Milliford, MA) equipped with a WISP 717 plus auto sampler refrigerated at 5°C, and a 490E multiwavelength UV/Vis detector in line with a Raytest Ramona LS radioisotope detector

(Pittsburgh, PA) . The void volume between the UV/Vis detector and the radioisotope detector was 50 μl. For size-exclusion HPLC, samples were analyzed with an isocratic mobile phase of 0.1M sodium phosphate buffer, 0.5M NaCl, pH 6.9, on a Phenomenex BioSep S2000 column (Torrance, CA) eluted at 1.0 ml/min at 25°C. Elution was monitored for absorbance at 280 nm and recorded by Waters Millennium software on a PC computer.

As shown in Figure 2, the pre-G50 spin column sample revealed two major peaks with elution times of 8.65 and 9.53 minutes (Figure 2, line 2). The second major peak coelutes with free DTPA and was nearly eliminated in the post-G50 spin column sample (Figure 2, line 3) , indicating successful removal of unbound DTPA from the reaction mixture. The elution time of the remaining major peak was unchanged by the G50 spin column and eluted slightly before unreacted rhG-CSF (Figure 2, line 1) . This peak represents monomeric DTPA-conjugated rhG-CSF. Thus, the behavior of the DPTA-conjugated rhG-CSF on this size-exclusion column allows it to be resolved from unmodified rhG-CSF.

The analysis above shows that initial pH and DTPA:rhG-CSF molar ratio can affect formation of undesirable side reactions. By initiating the reaction in buffer with an initial pH of 6.0, the formation of products resulting from such side reactions (e.g. protein crosslinking) is greatly reduced.

EXAMPLE 2

In this example, the ability of the DTPA:rhG-CSF conjugate to chelate ^ιnIn was determined using analytical cation-exchange HPLC and thin layer chromatography. The conjugate was then further analyzed to determine the mass of the conjugate (with and without chelated indium) , the stoichiometic molar ratio of DTPA to rhG-CSF, and the location of the conjugated DTPA moiety on the rhG-CSF.

Preparation of DTPA:rhG-CSF Conjugate

In this example, analysis was conducted on a DTPA:rhG-CSF conjugate prepared using an initial pH of 6.0, and a 50:1 DTPA dianhydride:rhG-CSF molar ratio as described in Example 1. However, instead of passing the sample through a G50 spin column, preparative cation- exchange chromatography was performed using a Pharmacia Hi-Load SP-Sepharose High Performance, 16/10, strong cation-exchange column (Pharmacia, Sweden) . Separation was accomplished at 5°C by a Pharmacia FPLC system equipped with a 50 ml injection loop. The column was equilibrated in Buffer A (20mM sodium acetate, pH 5.4) and elution was carried out with a 0-40% Buffer B (20mM sodium acetate, 0.5M NaCl, pH 5.4) gradient over 180 minutes at 1.0 ml/minute. Elution was monitored for absorbance at 280 nm and recorded.

A reaction mixture (initial pH 6.0, 50:1 DTPA dianhydride: rhG-CSF molar ratio) originally containing 20 mg of rhG-CSF was diluted to 50 ml with Milli-Q water and directly applied to the Hi-Load SP-Sepharose column. A peak representing approximately 13% of the integrated peak areas (Figure 3, peak 2) coeluted with control unreacted rhG-CSF (Figure 3, dashed line) . This indicates that approximately 13% of the rhG-CSF remained unmodified. A peak eluted between 120 and 130 minutes contained approximately 84% of the total eluted protein (Figure 3, peak 1) . The shift in this material to elute at a lower salt concentration is in agreement with an increase in negative charge on the protein via conjugation with DTPA. This simple, efficient chromatography step yields a homogeneous and well-defined product with both unbound DTPA and unconjugated rhG-CSF separated from the purified DTPA:rhG-CSF conjugate.

Anaivsis of the DTPA:rhG-CSF Conjugate

1. Analytical Cation-Exchange HPLC.

Analytical cation-exchange HPLC was performed with mobile phases of Buffer A (20mM sodium acetate, pH 5.4) , and Buffer B (20mM sodium acetate, 0.5M NaCl, pH 5.4) on a Tosohaas SP-5PW, 7.5 X 7.5 mm column (Montgomery, PA) using the Waters HPLC system. The column was equilibrated with mobile phase A, and separation was performed at 25°C with a 1% B/min linear gradient over 30 minutes at 1.0 ml/minute. Separation was detected by monitoring absorbance at 220 nm, and where applicable, with the radioisotope detector. For samples containing indium, lmM EDTA was added to Buffer A before adjusting the pH to 5.4. rhG-CSF and DTPA:rhGCSF conjugate were preincubated with ^U1ln for 15 minutes. An excess of cold indium was then added (In:protein, 2:1, mol/mol), and the samples analyzed by analytical cation-exchange HPLC as described above. For the DTPA:rhG-CSF conjugate (Figure 4, lines 1 and 4, solid), 99.5% of the ^luIn radioactivity coeluted from the cation-exchange column with the protein, whereas no detectable radioactivity co-eluted with unmodified rhG-CSF (Figure 4, lines 2 and 3, dashed), indicating chelation of ^nlIn by the DTPA:rhG-CSF conjugate, and absence of ^ι In binding by unmodified rhG-CSF.

The effect of metal chelation on analytical cation-exchange HPLC analysis of the DTPA:rhG-CSF conjugate is depicted in Figure 5. Elution of rhG-CSF (line 1), the DTPA:rhG-CSF conjugate (line 2), and the conjugate preincubated with excess InCl₃ (In:conjugate, 10:1, mol/mol, line 3) was monitored for absorbance at 220 nm. The DTPA:rhG-CSF conjugate eluted at a lower salt concentration than unmodified rhG-CSF. The chelated conjugate then elutes at a slightly higher salt concentration than the non-chelated conjugate, but still at a salt concentration lower than that of unmodified rhG-CSF. The characteristic retention times of the DTPA:rhG-CSF conjugate with and without chelated metal may be used to monitor metal contamination of the conjugate preparation. Furthermore, this analysis may be used to monitor metal labeling of the conjugate. 2. Thin Layer Chromatography (TLC)

TLC was performed as previously described (Meares et al., J. Prot . Chem. , 2 215-228, (1984)) with slight modification. An indium stock solution containing InCl₃ with a trace of ^ιnIn was prepared in lOmM HCl, and was used to prepare the following samples: (1) indium added to lOOmM sodium phosphate, pH 6.0; (2) 10 nmol indium added to 20 nmol DTPA in 20mM sodium acetate, pH 5.4; (3) 10 nmol indium incubated with 2 nmol rhG-CSF at room temperature for 10 minutes, followed by addition of 20 nmol DTPA in 20mM sodium acetate, pH 5.4, and (4) 10 nmol indium incubated with 2 nmol DTPA-conjugated rhG-CSF at room temperature for 10 minutes, followed by addition of 20 nmol DTPA in 20mM sodium acetate, pH 5.4. One μl of each sample (containing 0.1 nmol indium) was spotted onto 250 μm thick silica gel (60 A) on glass backing (Whatman, Clifton, NJ) . The TLC plate was developed using 10% (w/v) ammonium acetate in distilled H 0:methanol (1:1, v/v) as the solvent. The developed plate was then analyzed using a Molecular Dynamics Phosphorlmager (Sunnyvale, CA) .

The stoichiometric molar ratio of DTPA to rhG-CSF was determined as described above. Chelation of ^ιnIn by DTPA results in migration of all radioactivity from near the solvent front (Figure 6, compare lanes 1 and 2) . Incubation of ^ιnIn (10 nmol) with the DTPA:rhG- CSF conjugate (2 nmol) , followed by addition of DTPA, resulted in retention of a portion of the radioactivity at the origin (Figure 6, lane 4) . Line graphs of the individual lanes were generated and integration of the peak areas from lane 4 revealed 18% of the radioactivity remained at the origin. The remaining unbound ^ιuln was scavenged by the added DTPA and migrated near the solvent front. Thus, approximately 1.8 nmol of ^nιIn was bound by 2 nmol of the DTPA:rhG-CSF conjugate, indicating a DTPA to rhG-CSF molar ratio of 0.9. Unmodified rhG-CSF did not retain radioactivity at the origin, indicating absence of ^ιIn binding (Figure 6, lane 3) .

3. Mass Spectrometry

Matrix-assisted laser desorption/ionization mass specrometry (MALDI-MS) was performed with a Kompact MALDI III mass spectrometer (Kratos Analytical, Ramsey, NJ) fitted with a standard 337 nm nitrogen laser. The spectra were recorded with the analyzer in linear mode at an accelerating voltage of 20 kV. A sample aliquot containing 15 pmol of protein and 1.0 μl alpha-cyano-4- hydroxycinnamic acid were mixed in the sample wells of the probe slide and allowed to air dry. The laser fluence of the instrument was set at 30 (adjustable over a relative scale of 0-100) . The mass of the DTPA:rhG-CSF conjugate was determined by MALDI-MS (Figure 7) . The acquired spectrum revealed multiply charged ions in addition to the monoprotonated species. The mass obtained by averaging the peak series was 19,171.7 (± 7.3) Da. The calculated MW for a single DTPA conjugated to rhG-CSF is 19,170.8 Da. Therefore, in general agreement with the TLC analysis, the observed mass indicated a DTPA to rhG-CSF molar ratio of 1:1 for the DTPA:rhG-CSF conjugate.

4. Ion-spray mass spectrometry.

Ion-spray mass spectrometry was performed with a Perkin-Elmer Sciex API III mass spectrometer (Norwalk, CT) equipped with an ion-spray interface by method of flow injection. Samples were diluted in - 23 -

water/acetonitrile/formic acid (50:50:0.1, v/v) and flow injected into the same solvent flowing at 25 μl/minute. The orifice was set at 70 V, and the mass spectrometer operated in Ql mode. The mass of the DTPA:rhG-CSF conjugate with chelated indium was determined as described above. Analysis of the conjugate, preincubated with saturating indium (In:conjugate, 10:1, mol/mol), yielded a series of peaks with differing m/z values. This multiply charged ion series, arising from multiple protonation of the protein, was deconvoluted to produce the MW spectrum shown in Figure 8. The measured mass of the conjugate with chelated indium was 19,286 (± 1.7) Da, which is in agreement with the calculated mass of 19,285.6 Da. For rhG-CSF, the measured mass was 18,798 (± 1.8) Da (Figure 9), in agreement with the calculated mass of 18,798.5 Da.

5. Peptide Mapping

For peptide analysis, approximately 0.5 mg of rhG-CSF or DTPA:rhG-CSF was dried in a speed vacuum, reconstituted in 100 μl of 8M Urea and sonicated for 10 minutes. After sonication, 10 μl of 1M Tris-HCl, pH 8.5 and 2.5 μg of EndoLys-C (Wako Chemicals, Richmond, VA) from a 1 mg/ml stock solution in lOmM Tris HCl, pH 8.5, was added. The total volume was adjusted to 200 μl with distilled water, and the proteolytic digestion was carried out for 7 hours at room temperature. Following the hydrolysis with EndoLys-C, the disulfide bonds were simultaneously reduced with 5 μl of 80mM TBP and alkylated with 10 μl of 40mM ABD-F (2mM final concentration) as described; Kirley, T.K., ^al . Bioche . , 1£J__. 231-236, (1989) . Immediately after reduction and alkylation, the generated peptides (200 μl) were injected directly onto a 300 A pore size C reversed-phase HPLC column (Separations Group, Vydac, Hesperia, CA) equilibrated with Solvent A (0.1% TFA in distilled water) . Peptide analysis was performed using a Waters HPLC system consisting of two 510 pumps, a WISP 712 autoinjector, and a 481LC spectrophotometer all controlled through a system interface module by the system software, Maxima. The generated peptides were eluted with a linear gradient of 3-76% Solvent B (0.1% TFA, 95% acetonitrile) over 115 minutes. Elution was monitored for absorbance at 215 nm. Individual peptides from the rhG-CSF standard peptide map were collected and identified by amino acid composition analysis and N- terminal sequencing as described; Souza et al. , Science, 222., 61-65, (1986) . Peptide fragments of unmodified rhG-CSF and

DTPA:rhG-CSF were prepared and analyzed as described above. A peak eluting from the unmodified rhG-CSF sample at 60 minutes was absent from the DTPA:rhG-CSF conjugate sample (Figure 10) . The material eluting in this peak was determined by amino acid composition analysis and N-terminal sequencing to be the N-terminal 17 residues of rhG-CSF. Thus, the corresponding N- terminal peptide fragment from the DTPA:rhG-CSF conjugate was modified, yielding a new partially split double peak eluting at 62 minutes. Analysis of peptide from each of these partially separated peaks by mass spectrometry revealed the first peak to have the expected mass of the N-terminal peptide with conjugated DTPA, while the mass of the second peak material suggested conjugated peptide contaminated with iron.

Peptide mapping therefore indicated that the conjugated DTPA group is localized to the N-terminal 17 amino acids. This N-terminal peptide contains the N-terminus, one threonine, three serine residues and a lysine residue. Cleavage of the peptide by EndoLys-C indicates that the lysine is unmodified. Acylation of threonine or serine residues is highly unlikely at pH 6.0. Undigested DTPA:rhG-CSF conjugate subjected to N-terminal sequencing revealed >99% blocked N-terminus, indicating the single DTPA moiety on the protein is conjugated to the N-terminus.

6. Isoelectric Focusing.

Isoelectric focusing was performed using Novex pH 3-10 gels (San Diego, CA) with a pi 3.5 - 8.5 performance range. Samples were diluted 1:1 with sample buffer, and 5 μg of protein was loaded into each lane. The gels were run at constant voltages of 100 V for 1 hour, 200 V for 2 hours, and then 500 V for 0.5 hour. All fixing, staining and destaining procedures were done to the manufacturer's specifications.

The DTPA:rhG-CSF conjugate revealed a single major band of pi 4.9 following isoelectric focusing (Figure 11, lane 3) . Preincubation of the conjugate with excess InCl₃ (In:conjugate, 10:1, mol/mol) shifted the band to pi 5.3 (Figure 11, lane 2). The pi values of the conjugate, both.with and without indium, were lower than that of rhG-CSF, pi 6.0 (Figure 11, lane 1) . The conjugation of DTPA, concomitant with the loss of the N-terminal free amino group, substantially decreased the pi of the rhG-CSF. Furthermore, chelated indium slightly increased the pi of the conjugate. The characteristic isoelectric points of the DTPA:rhG-CSF conjugate with and without chelated metal may also be used to monitor metal contamination and metal labeling of the conjugate preparation.

The data above demonstrate that: (1) the DTPA:rhG-CSF conjugate is able to chelate ^ιnIn; (2) the stoichiometric molar ratio of DTPA to rhG-CSF is approximately 1.0 for the DTPA:rhG-CSF conjugate;

(3) the DTPA conjugation is specific to the N-terminus of the rhG-CSF; and (4) the conjugation of DTPA to rhG-CSF decreases the pi of rhG-CSF.

EXAMPLE 3

In this example, circular dichroism analysis was used to study the effects on rhG-CSF secondary structure resulting from conjugation of a chelating group to the N-terminus of rhG-CSF. Circular Dichroism (CD) spectra were obtained with a Jasco J-720 spectropolarimeter (Japan Spectroscopic Co., LTD., Tokyo, Japan). Samples (0.078 mg/ml protein) were analyzed at 10°C in 20mM sodium acetate, pH 5.4. The CD spectra of the DTPA:rhG-CSF conjugate overlays that of unmodified rhG-CSF (Figure 12), each revealing ellipticity minima at 208nm and 222nm. Addition of excess indium to saturate all chelating sites on the conjugate (In:conjugate, 10:1, mol/mol) did not change the overall shape of the spectra, yet caused a slight (approx. 5%) reduction in alpha helicity. Therefore, the secondary structure (at pH 5.4) is shown here not to be influenced by the conjugation of a chelating group to the N-terminus.

EXAMPLE 4

In this example, the effects of DTPA conjugation and subsequent chelation of indium on the biological activity of rhG-CSF was determined. Peripheral WBC counts in hamsters by rhG-CSF, the DTPA:rhG-CSF conjugate, and the conjugate with chelated indium were evaluated after subcutaneous injection of 100 μg/kg rhG-CSF in hamsters (Figure 13) . The animals were sacrificed at the indicated time intervals, and collected blood samples were analyzed using a Sysmex F800 microcell counter. Injection of the conjugate (Figure 13, (Δ) ) induced the level of peripheral WBC counts in a manner similar to unmodified rhG-CSF (Figure 13, (o) ) . Injection of the conjugate preincubated with excess indium (In:conjugate, 10:1, mol/mol) (Figure 13, (0)) also induced a similar response, with maximal WBC levels reached at 24 to 36 hours post injection. Thus, the conjugation of DTPA and rhG-CSF did not significantly alter the observed in vivo activity of the rhG-CSF, and furthermore the activity of the conjugate is unchanged after chelation of indium.

EXAMPLE 5

In this example, the conjugation reaction described above was carried out on the related growth factor, interleukin-2 (IL-2) . The ability of the DTPA:IL-2 conjugate to chelate ^luIn was evaluated using cation-exchange HPLC as described above. In addition, the stoichiometric molar ratio of DTPA:IL-2 was determined as well as the distribution of the DTPA moiety on the IL-2.

The IL-2 was produced using recombinant DNA technology in which E. coli cells were transfected with a DNA sequence encoding IL-2 as described in European patent 0136489 (Souza et al. ) . The IL-2 was prepared as a 1.82 mg/ml solution in 100 mM sodium phosphate buffer, pH 6.0. DTPA dianhydride and tributylphosphine (TBP, technical grade) were obtained from Aldrich (Milwaukee, WI) . The conjugates were prepared as described in Example 2 above. Analysis of the DTPA:IL-2 Conjugate

1. Analytical Cation-Exchange HPLC.

Analytical cation-exchange HPLC was performed as described above. For the DTPA:IL-2 conjugate (Figure 14, lines 1 and 4, solid), 99.5% of the ^lIn radioactivity coeluted from the cation-exchange column with the protein, whereas no detectable radioactivity co-eluted with unmodified IL-2 (Figure 14, lines 2 and

3, dashed), indicating chelation of ^nιIn by the DTPA:IL- 2 conjugate, and absence of ^luIn binding by unmodified IL-2.

The effect of metal chelation on analytical cation-exchange HPLC analysis of the DTPA:IL-2 conjugate is depicted in Figure 15. Elution of IL-2 (line 1), the DTPA:IL-2 conjugate (line 2), and the conjugate preincubated with excess InCl₃ (In:conjugate, 10:1, mol/mol, line 3) was monitored for absorbance at 220 nm. The DTPA:IL-2 conjugate eluted at a lower salt concentration than unmodified IL-2. The chelated conjugate then elutes at a slightly higher salt concentration than the non-chelated conjugate, but still at a salt concentration lower than that of unmodified IL-2. As was the case with rhG-CSF analysis above, the characteristic retention times of the DTPA:IL-2 conjugate provide a useful method of monitoring metal contamination and metal labeling of the conjugate.

2. Thin Layer Chromatography (TLC) .

TLC was performed as described above in order to determine the ability of the DTPA:IL-2 conjugate to chelate ¹¹¹In, and to determine the stoichiometric molar ratio of DTPA to IL-2. As shown in Figure 16, chelation of ^U1ln by DTPA results in migration of all radioactivity from near the solvent front (Figure 16, compare lanes 1 and 2). Incubation of ^U1ln (10 nmol) with the DTPA:IL-2 conjugate (2 nmol) , followed by addition of DTPA, resulted in retention of a portion of the radioactivity at the origin (Figure 16, lane 4) . Line graphs of the individual lanes were generated and integration of the peak areas from lane 4 revealed 18% of the radioactivity remained at the origin. The remaining unbound ^luIn was scavenged by the added DTPA and migrated near the solvent front. Thus, approximately 1.8 nmol of ^U1ln was bound by 2 nmol of the DTPA:IL-2 conjugate, indicating a DTPA to IL-2 molar ratio of 0.9:1. Unmodified IL-2 did not retain radioactivity at the origin, indicating absence of ^lιαIn binding (Figure 16, lane 3) .

3. Peptide Mapping.

Peptide analysis was performed on the

DTPA:IL-2 conjugate in order to determine the location of the conjugated DTPA moiety on the IL-2. Peptide fragments were prepared as described above.

As shown in Figure 17 (arrow) , a peak eluting from the unmodified IL-2 sample at approximately 36 minutes was absent from the DTPA:IL-2 conjugate sample. The material eluting in this peak was determined.by amino acid composition analysis and N-terminal sequencing to be the N-terminal peptide of IL-2. Therefore, the corresponding N-terminal peptide fragment from the DTPA:IL-2 conjugate was modified, yielding a new partially split peak at 40 minutes. As was the case with rhG-CSF, peptide mapping indicated that the conjugated DTPA group is localized to the N-terminus. These data demonstrate that the DTPA.-IL-2 conjugate is able to chelate ^luIn, that the stoichiometric molar ratio of DTPA to IL-2 is approximately 1.0, and that DTPA is conjugated site- specifically to the N-terminus of IL-2.

Claims (20)

WHAT IS CLAIMED IS:

1. A composition comprising a chelating agent, a protein, and a metallic cation, said chelating agent bound to said metallic cation and conjugated site- specifically to the N-terminus of said protein.

2. The composition of Claim 1 wherein said metallic cation is radioactive.

3. A composition of Claim 1 or 2 wherein said chelating agent is the dicyclic dianhydride of diethlyenetriaminepentaacetic acid.

4. A composition of Claim 2 wherein said radioactive metallic cation is selected from the group consisting of gallium-67, indium-Ill and technetium-99m.

5. A composition of Claim 1 or 2 wherein said protein is selected from the group consisting of G-CSF, GM-CSF, M-CSF, the interferons (alpha, beta, and gamma) , the interleukins (1-14), erythropoietin (EPO) , fibroblast growth factor, stem cell factor, nerve growth factor, NT3, megakaryocyte growth and development factor (MGDF) , platelet-derived growth factor (PDGF) , and tumor growth factor (alpha, beta) .

6. A composition of Claim 5 wherein said protein is G-CSF.

7. A composition of Claim 5 wherein said protein is IL-2.

8. A composition comprising the dicyclic dianhydride of diethlyenetriaminepentaacetic (DTPA) ; rhG-CSF; and indium-Ill, said DTPA being bound to said indium-Ill and conjugated site-specifically to the N- terminus of said rhG-CSF.

9. A composition comprising the dicyclic dianhydride of diethlyenetriaminepentaacetic (DTPA) ; rhIL-2; and indium-Ill, said DTPA being bound to said indium-Ill and conjugated site-specifically to the N- terminus of said rhIL-2.

10. A method for preparing a labeled protein, said method comprising:

(a) reacting a chelating agent with said protein at a pH sufficiently acidic to selectively activate the α-amino group at the amino terminus of said protein;

(b) separating the conjugated protein from non-conjugated protein;

(c) adding a metallic cation to said conjugate; and

(d) obtaining the labeled protein.

11. A method according to Claim 10 wherein said pH is 6.0.

12. A method for preparing labeled G-CSF, said method comprising:

(a) reacting a chelating agent with said G-CSF at a pH sufficiently acidic to selectively activate the α-amino group at the amino terminus of said

G-CSF;

(b) separating the conjugated G-CSF from non- conjugated G-CSF;

(c) adding a metallic cation to said conjugate; and

(d) obtaining the labeled G-CSF.

13. A method according to Claim 12 wherein said pH is 6.0.

14. A method according to Claim 13 wherein said chelating agent is DTPA.

15. A method for preparing labeled IL-2, said method comprising:

(a) reacting a chelating agent with said IL-2 at a pH sufficiently acidic to selectively activate the α-amino group at the amino terminus of said IL-2;

(b) separating the conjugated IL-2 from non- conjugated IL-2;

(c) adding a metallic cation to said conjugate; and

(d) obtaining the labeled IL-2.

16. A method according to Claim 15 wherein said pH is 6.0.

17. A method according to Claim 16 wherein said chelating agent is DTPA.

18. A pharmaceutical composition comprising: (a) a substantially homogenous preparation of recombinant human G-CSF, said recombinant human G-CSF consisting of a chelating agent moiety with bound chelated metallic cation, conjugated to a recombinant human G-CSF moiety solely at the N-terminus thereof via an amide linkage; and (b) a pharmaceutically acceptable diluent, adjuvant or carrier.

19. A pharmaceutical composition comprising: (a) a substantially homogenous preparation of recombinant human IL-2, said recombinant human IL-2 consisting of a chelating agent moiety with bound chelated metallic cation, conjugated to a recombinant human IL-2 moiety solely at the N-terminus thereof via an amide linkage; and (b) a pharmaceutically acceptable diluent, adjuvant or carrier.

20. A composition of Claim 1 or 2 for use as a diagnostic imaging agent.