GB2318352A

GB2318352A - Polypeptides mimicking the activity of human erythropoietin

Info

Publication number: GB2318352A
Application number: GB9613289A
Authority: GB
Inventors: Dejan Markovic
Original assignee: Diapharm Ltd
Current assignee: Diapharm Ltd
Priority date: 1996-06-25
Filing date: 1996-06-25
Publication date: 1998-04-22
Also published as: GB9613289D0; WO1997049729A1; NO985993L; NO985993D0; EP0928292A1; AU3435397A; CA2258871A1

Abstract

Polypeptides or erythropoietin muteins, characterised by an amino acid sequence which is different from that of human erythropoietin and presents in the informational spectrum obtained by Fourier transformation according to the method of informational analysis is substantially the same frequencies of natural erythropoietin.

Description

2318352 POLYPEPTIDES MIMICKING TWIR ACTIVITY, OF HUMAN ERYTHROPOIETIN The

present invention refers to Polypeptides having a sequence obtained by the informational spectra method and the ability of stimulating the production of reticulocytes and red blood cells from bone marrow cells 5 as well as hemoglobin synthesis and iron uptake.

Human erythropoietin (EPO), an hormone playing a fundamental role in erythropoiesis, is a glycoprotein having a molecular weight of about 3438 kd, the primary structure of which is shown in Figure 1.

EPO is presently obtained by recombinant DNA techniques using eukaryotic cells which can glycosylate the expression product.

The availability of shorter polypeptides, possibly active even in unglycosylated form mimicking the biological activity of human EPO, would be an highly desirable goal allowing the convenient preparation by synthetic procedures.

It has now been found that polypeptides, having in the informational spectrum obtained by Fourier transformation according to the informational analysis method substantially the same frequencies of natural erythropoietin, have substantially the same activity of human erytropoietin.

The informational analysis method (ISM), first disclosed by Veljkovic V. et al. in IEEE Trans. Biomed. Eng. 32, 337 (1985); Cancer Biochem. Biophys. 9, 139, 1987; Phys. Rev. Lett. 29, 105, (1972) and Phys. Lett.

45A, 41, (1973) the content of which is herein 2 incorporated by reference, is based on the analysis of the information encoded in primary structure which is expressed by molecular electric oscillations propagating through polar environment. Based on the previously demonstrated strong correlation between electron-ion interagtion potential (hereinafter EIIP) [Veljkovic V., A theoretical approach to preselection of carcinogens and chemical carcinogenesis, Gordon & Breach Sci. Pub.

New York, 1980; Politzer P. and Truhlar D. G., Chemical applications of atomic and molecular electrostatic potential, Plenum Press, New York, 1981); Politzer P., Toxicol, Lett. 43, 227 (1988)] it has been proposed that information expressed by electric oscillations is encoded in protein primary structure by distribution of the values of MIP of amino acids.

According to this approach, the protein sequences are transformed into signals by assignment of numerical values to each amino acid. These values correspond to EMP [Veljkovic V. and Slavic I., Phys. Rev. Lett., 29, 105 (1972); Veljkovic V. Phys. Lett., 45A, 41 (1973)1.

The signal obtained is than decomposed in periodical function by Fourier transformation. The result is a series of frequencies and their amplitudes. The obtained frequencies correspond to the distribution of structural motifs with defined physico-chemical characteristics responsible for biological function of protein. When comparing proteins which share the same biological or biochemical function, the technique allows detection of code/frequency pairs which are specific for their common biological properties. The method is insensitive to the location of the motifs and, thus, does not require 3 previous alignment of the sequence. The ISM was successfully applied in structure/ function analysis and de novo design of peptides [Cosic I. and Nesis D., Eur.J. Biochem., 170, 247 (1988); Skerl V., and Pavlovic M., FEBS Lett., 239, 1411 (1988); Veljkovic V. and Metlas R., Cancer Biochem. Biophys., 10, 191 (1988); Cosic I. et al.., Biochemie, 71, 333 (1989); Lalovic D. and Veljkovic V., Biosys-tems, 23, 311 (1989); Cosic I., Resonant recognition model of protein-protein and proteinDNA recognition, in Bioinstrumentation and Biosensor (edited by Weis D,L.), Marcel Dekker, Inc., New York (1990); Cosic I. et al.., Eur. J. Biochem. 198, 113 (1991); Cosic I. and Hearn M.T. W., J. Mol. Recogion. , 4, 57 (1991); Veljkovic V. et al.., Biochem. Biophys. Res. Commun., 189, 705 (1992); Krsmanovic B. et al.., WO 93/17108.

An object of the invention is provided by EPO muteins having an homology degree with natural erythropoietin lower than 60% or polypeptides having from 20 to 100, preferably from 25 to 70 amino acids, characterized by informational spectrum having substantially the same frequencies as found in the informational spectrum of natural erythropoietin.

The muteins or polypeptides according to the invention may be designed so as to include appropriate 0- or N-glycosilation sites even though it has been surprisingly found that glycosilation is not always necessary for the biological activity.

More particularly, the muteins or polypeptides of the invention are characterized by the frequency component 0.312 0.004 in the informational spectrum 4 and at least one of the following frequency components: 0.023, 0.156, 0. 180, 0.185, 0.258, 0.273, 0.285, 0.363 and 0.500 determined with the accuracy of 1 0.004.

Preferred muteins of the invention are shown in Figures 2a and 2b whereas preferred polypeptides are shown in Figure 3. In Figures 1, 2a and 2b, the predicted amphipathic a helices are underlined with double line (====). Residues predicted to be on the surface are underlined. Each of the three N-linked and one of 0-linked glycosylation sites is designated by an asterisk. The mutations are designated by small letters.

The invention also refers to polynucleotide sequences coding for said muteins or polypeptides, to expression vectors comprising said polynucleotide sequences and to hosts transformed or transfected by said vectors, as well as to nucleotide sequences which hybridize to the above mentioned coding sequences.

The polypeptides sequence of the invention are determined by a procedure involving the following steps:

1. determination of the consensus characteristic frequencies for the EPO molecules; 2. derivation of a new numerical sequence of the desired length having the same characteristic frequencies using inverse Fourier transformation; 3. determination of the amino acid corresponding to each element of this new numerical sequence from values of MIP (see Table l):

------------- Amino acid WIP [Ry) L 0.0000 1 0.0000 N 0.0036 G 0.0050 v 0.0057 E 0.0058 p 0.0198 0.0242 0.0371 A 0.0373 Y 0.0516 W 0.0548 Q 0.0761 m 0.0823 S 0.0829 c 0.0829 T 0.0941 7 0.0946 R 0.0959 D 0.1263 ------------------------------------- Ry = Rydberg unit The polypeptides may be obtained by conventional methods of peptide synthesis or by known recombinant DNA techniques.

The polypeptides or muteins of the invention may be administered to humans or animals in form of suitable pharmaceutical compositions, usually but not exclusively 6 to be administered parenterally. Said compositions will contain from 1 to about 100 mg of mutein or polypeptide for the treatment of the same pathological conditions presently treated with human or recombinant EPO.

The following examples further illustrate the invention.

Example 1

Analygis of amino acid sequences by the ISM In the first step of the ISM analysis each constitutive element (amino acid) in analyzed sequence is represented by corresponding WIP value. For calculation of EIIP the following expression derived from the "general model pseudopotentiall' [Veljkovic V.

and Slavic I., Phys. Rev. Lett., 2,1, 105 (1972); Veljkovic V., Phys. Lett., A_5A, 41 (1973)l was used:

W = 0.25 Z sin (1.04 ir Z)/2 (1) where Z is the average quasi-valence number determined by: z m nizi/N (2) where Z is the number of valence electrons of the i-th atomic component, ni - the number of atoms of the i-th component, m and N - the number of atomic components and total number of atoms in the side group, respectively. The values of EIIP for side groups of amino acids calculated in accord with Eq. (1) are given in Table 1.

The numerical serie determined in this way is finite-lenght deterministic discrete signal containing information corresponding to selective longdistance interaction among biological macromolecules. In order to analyze this information, the obtained numerical sequence was subjected to discrete Fourier 7 transformation (DFT), which is defined as follows [Rabiner R., L. and old B., Theory and applications of digital processing. Prentice-Hall Inc., nglewood 1975)1:

N-1 2 X(n) x(m)e n = 1,2 N12 (3) m=0 Here x(m) is the m-th of a given numerical series, and X(n) are coefficients of DFT. The coefficients are describing the amplitude, phase, and frequency of sinusoids from which original sign al consists. The absolute value of complex DFT coefficients determines the amplitude spectrum which is in the ISM defined as informational spectrum (IS) and represented by the following equation:

S(n) = X(n)X(n) = IX(n)12, n 1, 2 N12, (4) It was assumed that points in analyzed numerical sequences are equidistant with the distance d = 1. In this case the maximal frequency in the spectrum is Fmax = 1/2d = 0. 5. It is important to note that the frequency range is independent of number of points in the sequence. The total number of points in the sequence (i.e. number of amino acids in the analyzed primary structure) influences only the resolution of IS. In the case of an N-point sequence, the resolution equals 1IN.

The minimal lenght of sequence that can be analyzed by ISM is determined by the desired resolution of the spectrum. Therefore, this number is determined by the expected number of peaks which are to be strictly separated and cannot be exactly defined. The minimal length of sequence which can be analyzed by ISM with suitable accuracy is 16 amino acids.

8 In thi's way, the information primarily defined by the sequence of symbols representing amino acids is presented in spectral form which is more suitable for' mathematical analysis. It is important to note that ISM does not influence this information and represents only a tool for its analysis (like the prism which decomposes the white light in its spectral components). Each frequency in the is represents a particular informational component encoded in the primary structure by regularly distributed structural motifs with similar electronic properties.

Example 2

Application of the technique in exa=le 1 0 ISP.

analysis of EPO Rroteins The algorithmic procedures were applied to the mammalian EPO protein sequences.

The analysis procedure comprises the following steps:

I. each amino acid sequences was converted to the numerical sequence by representing each amino acid with the corresponding value of the EIIP; 2. this numerical sequence was converted into a numerical spectrum using fast Fourier transform (hereinafter FFT); 3. spectra were mutually compared using cross-spectral analysis with the aim to extract common frequency components.

From the analysis of cross-spectra of EPO molecules from various mammalian species (mouse, rat, rabbit, ship, monkey and human) one can deduce a set of characteristic frequencies which predominates in the 9 obtained CIS. In Table 2 all characteristics frequencies in the EPO CIS with SIN > 1 are given:

Tahle 2 F1 0.023 F2 0.156 F3 0.180 F4 0.195 F5 0.258 F6 0.273 F7 0.285 F8 0.312 F9 0.363 F10 0.500 All frequency components in Table 2 are determined with accuracy of 0.004.

From the obtained results, it is possible to conclude that information which is essential for biological activity of the analyzed EPO molecules is completely determined with the set of characteristic frequencies given in Table 2.

Examiple 3 Application of the technicrue iple 1 to the analysis of EPO-EPOR interaction In order to determine, which of the characteristic frequency components from Table 2 determines the information that is essential for human EPO-EPOR interaction, the cross-spectrum between these two proteins is obtained.

The characteristic frequencies corresponding to the first 15 amplitudes in EPO-EPOR cross-spectrum are given in Table 3.

Table 3

F SIN 0.311 10.6 0.258 7.5 0.272 7.3 0.113 7.0 0.361 7.0 0.498 5.2 0.430 5.1 0.158 5.1 0.031 4.6 0.382 4.6 0.154 4.4 0.283 4.3 is 0.275 3.9 0.347 3.9 0.008 3.5 The results presented in Table 3 show that the main part of information corresponding to human EPO-EPOR interaction is determined by the frequency component 0.311. It is also important to note that, taking into account accuracy of 0.004 in the determination of frequency values, 8 of 10 characteristic frequencies from the EPO CIS (Table 2) are also contained within the first 15 characteristic frequencies in the human EPO EPOR cross-spectrum (Table 3).

Ex=Rle 4 Apnlication of the tech exaMR1e to design-QI RPO mut Once the characteristic frequencies for EPO protein family had been determined, it was possible to design EPO muteins introducing a large number of amino acid substitutions in human EPO. The main condition that must be satisfied in the design of these muteins is the conservation of IS of human EPO.

In Figure 2a, the primary structure of mutein-1 generated by substitution of 99 (56.6%) amino acids in human EPO is given. The IS of this mutein is given in Table 4 and Figure 4a. As can be seen, this IS contains all 10 characteristic EPO frequencies from Table 2, as well as other frequency components corresponding to first 15 amplitudes in IS of the human EPO.

Table

IS) IS(MUL1) IS(MEL2) 0.359 0.359 0.359 0.500 0.156 0 0.156 UW 0.156 0.195 0.285 O.M 0.258 0.258 0.259 0.295 0.195 0.195 0.203 0.203 0.273 0.273 L273 0.203 0.180 M02 0.1M 0.004 0.180 0301 0.445 0.004 0.04 0.113 0.113 0.211 0.023 0.172 0.113 0.312 0.211 0.312 0.050 (U12 0.402 12 In order to design the mutein that with higher probability will express EPO activity, also structural characteristics that are important for this biological activity must be taken into account [Boissel JP. et al.., J. Biol. Chem., 268, 15983 (1993); Wen D., et al.., J. Biol. Chem., 269, 22839 (1994)1. In Figure 2b, the primary structure of mutein-2 generated by substitution of 72 (43.4%) amino acids in human EPO is given. This mutein besides conserved IS of human EPO (Table 4 and Figure 4b) also has preserved all structural elements that are important for the biological activity of the human EPO, including all glycosilation sites [Wen D., et al.., J. Biol. Chem.

269, 22839 (1994)1. The comparison of the alpha and beta propensity, hydrophilicity, hydrophobicity, solvent accessibility, antigenicity and secondary structure of the human EPO and mutein2 are given in Figure 5-6.

13

Claims

A polypeptide mimicking the activity of human erythropoietin, having an amino acid sequence which is different from that of human erythropoletin and which presents substantially the same frequencies as natural erythropoietin in the information spectrum obtained by Fourier transformation according to the method of informational analysis.
2. A polypeptide according to Claim 1, which is an erythropoietin mutein having an homology degree with natural human erythropoietin lower than 60%.
3. A polypeptide according to Claim 1 or 2, having from 25 to 70 amino acids.
4. A polypeptide according to any one of Claims 1 to 3, having the frequency component 0.312 0.004 in the informational spectrum and at least one of the following frequency components: 0.023, 0.156, 0.180, 0. 195, 0.258, 0.273, 0.285, 0.363 and 0.500 the accuracy being 0.004.
5. A polypeptide according to any preceding Claim, having any of the primary sequences shown in Figures 2a, 2b and 3.
6. A polynucleotide coding for a polypeptide according to any preceding Claim.
7. An expression vector comprising a polynucleotide according to Claim 6.
8. A pharmaceutical composition containing a polypeptide according to any of Claims 1 to 5 in admixture with one or more suitable carriers or excipients therefor.
9. A peptide having human EPO activity but which is not natural human EPO, substantially as described herein.